DESIGN OF MACHINE ELEMENTS

DESIGN OF MACHINE ELEMENTS

 There is no great concurrence between learning and
wisdom.
Other Books by VIRGIL MORING FAIRES Francis Bacon (1561-1626)

V. M. Faires ELEMENTARY THERMODYNAMICS, 3D ED.

V. M. Faires THERMODYNAMICS, 4TH ED.

V. M. Faires THERMODYNAMICS OF HEAT POWER, 2D ED.

V. M. Faires, C. M. Simmang, PROBLEMS ON THERMODYNAMICS, 4TH ED.

and A. V. Brewer

V. M. Faires and S. D. Chambers ANALYTIC MECHANICS, 3D ED.

V. M. Faires and R. M. Wingren PROBLEMS ON THE DESIGN OF

MACHINE ELEMENTS, 4TH ED.
 2
FOURTH EDITION

DESIGN OF.
LAPPEENRANNAN TEMNILLINEN KORKEAKOULU

MACHINE ELEMENTS

SS
~~

\<
\ \\
an

WK
(%
be
C

‘y Sy
\

VIRGIL MORING FAIRES

UNITED STATES NAVAL POSTGRADUATE SCHOOL

THE MACMILLAN COMPANY, NEW YORK/COLLIER-MACMILLAN LIMITED, LONDON

 PREFACE

Since this book deals with engineering, it necessarily deals with science,
© Copyright, Virgil M. Faires, 1965 because engineering is the art of applying the physical sciences to the
problems of mankind. For the most part, the science involved here is the
science of mechanics, and it is assumed that the reader is familiar with the
All rights reserved. No part of this book may be reproduced
or transmitted in any form or by any means, electronic or
contents of typical undergraduate courses in analytic mechanics and
mechanical, including photocopying, recording or by any strength of materials. Therefore, derivations common to these courses are
information storage and retrieval system, without permission not repeated, but there are extensive reminders of limitations and of how
in writing from the Publisher.
to use the resulting equations.
Eighth Printing, 1971
The amount and quality of the presentation of the science of mechanics
have been enhanced, but not at the expense of the engineering. On the
Earlier editions, copyright 1934, 1941, and 1955 by The
Macmillan Company; copyright renewed 1962 by Virgil contrary, a large percentage of the increased number of pages is attributable
M. Faires to the new engineering information included. This increased emphasis on
Library of Congress catalog card number: 65-10732
engineering is more than ever desirable as an offset to the apparently
inevitable notions of students about the infallibility of science in an
Tue MACMILLAN COMPANY
engineering situation. The most used parts of the science of mechanics are
866 Tuirp AVENUE, New York, New York 10022
precise and impeccable, but there are innumerable rough spots in the path
COLLIER-MACMILLAN CANADA, LTD., TORONTO, ONTARIO
of the practitioner. One must have an instrument (or the tools) to practice
Printed in the United States of America an art, and the machine elements serve admirably as instruments for the
beginner. While it is not necessarily true that good judgments are inevitable
when arrived at from good knowledge, it is self evident that good knowledge
sa necessary prerequisite for good judgment; and we have tried to empha-
Size this point throughout by presenting certain engineering knowledge, by

vii
viii PREFACE PREFACE ix

directing attention to references on specific points, and by including a long minor improvements in them that all together they make a significant
list of references whose titles are suggestive of sources of additional change.
knowledge. Since the point is repeatedly made that good engineering Problems are given in a separate book, Problems on the Design of
judgment is a consequence of good engineering knowledge, we are hopeful Machine Elements, 4th Ed., by Faires and Wingren. This arrangement is a
that the student who reads practically the entire text understands that good convenience, both forthe teacher and student. The book of problems includes
decisions are not made in a cerebral void, that science must be tempered many tables, charts, and formulas that increase its usefulness; in addition,
to the ‘“‘facts’’ of life, that he, the student, has just begun to acquire it contains more problems than would be feasible to present within the
engineering knowledge, and that engineering changes because we learn covers of this volume.
more but that there is no point in ignoring what we “know”? now. We As always, I shall be most grateful to those who will be so kind as to
might even be hopeful that some readers become interested enough to want report errors they have discovered or who have suggestions for improvement.
to extend mankind’s knowledge in certain areas or to correct what we
presently “‘know.’’ Considering the variability of engineering data, teachers VIRGIL M. FAIRES
are challenged to impress upon students that this variability does not Monterey, California
imply that any old guess is as good as another.
The objective added in the third edition—to wit, to make the book more
useful to the practicing engineer—has been so much appreciated by students
that the current changes in tone and content are as much due to this
objective as to the original pedagogical objective. New procedures, new
engineering information, and different points of view are so prevalent
throughout that (I am sorry to report) nearly all my solutions to the
problems must now be revised. You will observe: another large expansion
of the treatment of fatigue, more emphasis on probabilities and the natural
variability of data, a more precise treatment of the science of mechanics,
frequent references to the effects of residual stresses, a single equation easily
adaptable to design by either the maximum shear stress theory or the
octahedral shear theory (for a normal and a shear stress), computer solu-
tions of Reynolds’ equation for journal bearings, several approaches to
gear-tooth design, and more detail on contact stresses, on theories of failure,
and on the mechanics of brake shoes.
The textbook writer performs a function similar to that of a reporter.
Since there is always so much that is relevant, the writer is responsible for
choosing that which is most important to be said. Because opinions as to
what should be done in mechanical engineering design are so varied and
because machine design, the special province of mechanical engineers, deals
with myriads of machines, the choice of material that emphasizes the right
fundamentals and that allows a flexibility of approach is perhaps more
difficult for machine design texts than for most others.
It is true that for each of the previous revisions the literature has been
searched and judged as to its appropriateness for the objectives of this
text. For this edition, more than usual time has been spent on this search;
but in addition, every sentence has been examined and rejudged, with the
result that it would be difficult to find a sentence dealing with engineering
identical to one of the previous edition. Although most of the illustrations
will be familiar to long-time friends of the book, there have been so many
 ACKNOWLEDGMENTS

FIRST EDITION Michigan State College, and D. K. Wright of Case Institute of Technology.
...to Mr. T. M. Durkan of the Gleason Works for . . . suggestions... ... Among others... W. W. Austin of North Carolina State College;
on bevel gears; to Mr. M. D. Hersey . . . for reading the chapter on journal A. M. Wahl, R. E. Peterson, and John Boyd of the Westinghouse Electric
bearings...to Mr. A. M. Wahl of the Westinghouse Electric... for Co.; W. Coleman of the Gleason Works; H. G. Taylor of the Diamond
reviewing the chapter on springs; to Mr. D. T. Hamilton of the Fellows Chain Co.; R. D. Knight of American Steel & Wire; E. N. Swanson of
Gear Shaper Company for reading the chapter on spur gears; and to Mr. Brown & Sharpe Manufacturing Co.; E. Siroky of the Wagner Electric
D. F. Windenburg of the United States Experimental Model Basin for as Corp.; F. A. Votta, Jr., of the Hunter Spring Co.; W. S. ‘Worley of the
yet unpublished material on thin shells under external pressure... to Gates Rubber Co.; S. J. Needs, Kingsbury Machine Works. . . . to Professor
Professor Earle Buckingham for repeatedly rendering assistance during the P. B. Leonard of North Carolina State College for his careful work on the
preparation of the chapters on gearing and for manuscript material which line drawings. . . .
he has not yet published.
FOURTH EDITION
I am grateful to several people for their helpful interest: John Boyd for
REVISED EDITION the solutions of the journal bearing equations, F. A. Votta and W. R.
...to Professors R. M. Wingren and J. G. H. Thompson, Professors
Johnson for information on springs, O. W. Blodgett for material on the
A. H. Burr and M. L. Price...to Professor Earle Buckingham... to
design of welds, T. E. Winter and W. D. Cram anent gearing, R. M.
Mr. S. J. Needs... journal bearings.
Wingren for the many helpful comments in general. There are numerous
acknowledgments for specific help, including illustrations, in the book.
THIRD EDITION And most of all, thanks to my wife, Lucile, for her patience, understand
ing,
..to... Professors R. L. Acres of Texas A. & M. College, C. T. and valuable help during the preparation of the manuscript.
Grace of the University of New Mexico, Boynton M. Green of Stanford
University, Fred Hirsch of the University of California, L. C. Price of V.M.F.

xi
 height; a dimension; ,, minimum film thickness in journal
bearing ,
heat-transfer coefficient (transmittance)
horsepower
interference of metal
rectangular or polar moment of inertia
polar moment of inertia; geometry factor, bevel gears
radius of gyration, (J/A)*/? or (J/m)"/?; spring constant, the
load-per-unit deflection; conductivity
Wahl factor for design; Ke, factor for effect of curvature in
springs and curved beams; K,, factor for shear in spring
theoretical stress-concentration factor; K,, fatigue strength
reduction factor
shaft-design factors from ASME Code
wear factors, spur gears, worm gears, cams
kinetic energy
length; a dimension
mass in slugs (W/g)
velocity ratio
moment of a force; bending moment; M,, vertical component
SYMBOLS of the moment; M,,, mean value of the moment; etc.
angular velocity; revolutions per minute; 7,, revolutions or
cycles per second; also m,, number of cycles of fatigue
The symbols agree in general with the recommendations of the American loading
Standards Association. Some exceptions were deemed advisable. In gearing, N design factor or factor of safety; sometimes, load normal to
where symbols had not already been established, the recommendations of a surface :
the American Gears Manufacturers Association were followed. N, N., etc. N with a subscript stands for the quantity of something, as
linear acceleration; a dimension; acoustic speed number of teeth or number of threads, number of coils, etc.
area; allowance pressure in pounds per square inch

bySD
Sey eA

breath; a dimension pitch of springs, gear teeth, threads, etc.; P,, diametral pitch;
life of rolling bearings P,, circular pitch
distance from neutral axis to fiber where stress is desired, quantity of fluid; notch-sensitivity index

>rXO's
usually to extreme fiber; bearing clearance quantity of heat; sometimes a force, a constant
center distance; spring index; a number; a constant radius
YQ

Ci, C2, etc. constants reaction or resultant force; radius of the larger of two wheels;
D diameter; D,, outside diameter; D;, inside diameter; etc. ratio; roughness; Ri», vertical component of Ri; Riy,
e eccentricity of load; effective error in gear-tooth profiles; — horizontal component of Rj; etc.
efficiency Rockwell C hardness; Rp, Rockwell B hardness, etc.
E modulus of elasticity in tension Stress; s,, alternating component of the total stress; 5,,,
alternating component in shear; s,, compressive stress;
ii coefficient of friction
F a force; total load; Fi, initial force or force at 1; F,,, mean Sa, design stress; se, equivalent stress; Ses, equivalent shear
force; F,, force applied at point A; etc. stress; sy, flexural or bending stress; s,,, mean stress;
local acceleration due to gravity; g,, standard acceleration of Sms, Mean stress in shear; s’n, endurance limit; s,, endur-
09

gravity (use 32.2 fps?) ance strength; s,,, endurance strength in torsion, load
modulus of elasticity in shear or torsion from zero to maximum; S,5, endurance strength in shear,
 xiv SYMBOLS

reversed load; s,, shear stress; s;, tensile stress; s,,, ultimate
strength; s,,, ultimate shear strength; s,,, ultimate com-
pressive strength; s,, yield strength in tension; s,,, yield
strength in shear or torsion; s,, initial stress or one part of
a total stress; s4, stress at point A; see also o and rt
Sommerfeld number; centrifugal force; separating force; the
distance a body moves, displacement; scale
thickness; temperature in degrees Fahrenheit
SSG Ss

torque; tolerance; T,,, mean value; T,, alternating component

work, U;, work of friction; U,, work of spring
velocity; v,, velocity in fps; v,,, velocity in fpm
volume; shearing force in beam section
load-per-unit distance; weight-per-unit distance; mass;
weight
total weight or load; force
Lewis’ factor in gearing
=

section modulus, J/c; absolute viscosity in centipoises

NN

7
section modulus based on ‘polar moment of inertia, J/c
a (alpha) coefficient of thermal expansion; an angle; angular acceler-
ation ABBREVIATIONS
B (beta) angle of limiting friction; an angle; cam angle
y (gamma) pitch angle of bevel gears; shearing-unit strain
8 (delta) total elongation; total deflection of a beam
€ (epsilon) normal-unit strain; eccentricity ratio AFBMA Anti-Friction Bearing Manufacturers Association
efficiency of riveted or welded joint AGMA American Gear Manufacturers Association
7 (eta)
@ (theta) an angle
AISC American Institute of Steel Construction
AISI American Iron and Steel Institute
A (lambda) lead angle of worm or screw threads
ALBA American Leather Belting Association
» (mu) Poisson’s ratio; absolute viscosity in lb-sec. per sq. in.
ASA
(reyns) American Standards Association
ASLE American Society of Lubrication Engineers
v (nu) kinematic viscosity
ASM American Society for Metals
a (pi) 3.1416...
ASME American Society of Mechanical Engineers
p (tho) density; sometimes variable radius
ASTM
o (sigma) resultant normal stress in combined stresses; standard American Society for Testing Materials
AWS American Welding Society
deviation
BHN Brinell hardness number
= (sigma) shaft angle, bevel and crossed helical gears; summation sign
CE counterclockwise
7 (tau) resultant shearing stress in combined stresses; time; represents
cfm cubic feet per minute
unit of time
c.g. center of gravity
¢ (phi) angle of twist; pressure angle in gears and cams; frequency
CI. cast iron
in cycles per second or minute
CE clockwise
# (psi) helix angle in helical gearing; spiral angle
cp centipoises
w@ (omega) angular velocity in radians per unit of time
cpm cycles per minute
cps cycles per second
fpm feet per minute
fps feet per second
xvi ABBREVIATIONS

fps? feet per second-second

gpm gallons per minute
hp horsepower
ID inside diameter
ips inches per second
ips? inches per second-second
ksi kips per square inch
mph miles per hour
mr millions of revolutions
OD outside diameter
OQT oil quenched and tempered
psi pounds per square inch
psf pounds per square foot
QT quenched and tempered
rpm revolutions per minute
rps revolutions per second
SAE Society of Automotive Engineers
SCF stress concentration factor
SESA Society for Experimental Stress Analysis
WQT water quenched and tempered ~
YP yield point SELECTED CHEMICAL
YS yield strength
pin. microinch = 10-® in. SYMBOLS

aluminum Fe iron Sb antimony

boron Mg magnesium Se selenium
bismuth Mn maganese Si _ silicon
beryllium Mo molybdenum Sn tin
columbium Ni _ nickel Ta tantalum
cadmium O oxygen Ti titanium
cobalt P___ phosphorus V_svanadium
chromium Pb lead W tungsten
copper S sulfur Zn zinc

xvii
 CONTENTS

PREFACE

ES
SYMBOLS ..

CHAPTER
a STRESS ANALYSIS—SIMPLE STRESSES

1.1 Introduction. 1.2 The Machine Designer’s Respon-

sibility. 1.3 Philosophies of Design. 1.4 Theory and
Practice. 1.5 Objective of This Book. 1.6 Machine
Design Is Engineering. 1.7 Stress. 1.8 Tensile Strength
and Yield Strength. 1.9 Modulus of Elasticity. 1.10
Bending. 1.11 Mathematical Relations for Beams; Shear
Center. 1.12 Finding the Moment of Inertia. 1.13
Torsion. 1.14 Torque. 1.15 Strength of Materials.
1.16 Factor of Safety—Design Factor. 1.17 Variability
of Strength of Materials and the Design Stress. 1.18 More
on Factor of Safety and Design Stress. 1.19 Example—
Design for Torsion. 1.20 Safe Compressive Stress. 1.21
Example—Stress Analysis. 1.22 Preferred Sizes (Standard
Fractions). 1.23 Courtesy in the Written Form of Calcula-
tions. 1.24 Buckling of Beam Flange. 1.25 Thin-wall
 CONTENTS
CONTENTS

Pressure Vessel. 1.26 Example—Titanium Vessel. 1.27 with K, 4.14 Example—Piston Rod. 4.15 Example—
Contact Stresses. 1.28 Statically Indeterminate Problems. Varying Torque. 4.16 Endurance Strength for a Finite
1.29 Thermal Stresses. 1.30 For Students Only. Life. 4.17 Example—Limited Life. 4.18 Example.
4.19 Equivalent Stress. 4.20 Design Factor for Variable
Loading. 4.21 Résumé of Design for Variable Stresses.
MATERIALS AND THEIR PROPERTIES 41 4,22 Superposed Stress Raisers. 4.23 Residual Stresses.
4.24 Plate with Elliptical. Hole. 4.25 Beam with Holes.
2.1 Introduction. 2.2 Definitions. 2.3 Heat-treatment 4.26 Corrosion. 4.27 Fretting Corrosion. 4.28 Shot
Terms. 2.4 Hardness. 2.5 AISI and SAE Specification Peening and Surface Rolling. 4.29 Heat Treatments for
Numbers. 2.6 Alloy Steel. 2.7 Hardenability. 2.8 Improving Endurance Strength. 4.30 Miscellaneous
Case Hardening. 2.9 Work Hardening. 2.10 Wrought Surface Effects: 4.31 Mitigating Stress Concentrations.
Tron. 2.11 Cast Iron. 2.12 Malleable Iron. 2.13 4.32 Temperature Effects: 4.33 More on Fatigue Strength.
Nodular CastIron. 2.14 Cast Steel. 2.15 Stainless Steel. 4.34 Impact. 4.35 Elastic Energy. 4.36 Bar Axially
2.16 Copper Alloys. 2.17 Aluminum Alloys. 2.18 Loaded. 4.37 Example. 4.38 Suddenly Applied Load—
Magnesium Alloys. 2.19 Titanium. 2.20 Lead, Tin, Zero Velocity of Impact. 4.39 Member with Two or More
and Miscellaneous Alloys. 2.21 High-temperature Service. Areas. 4.40 Design for Impact Loads. 4.41 Mass of
2.22 Low-temperature Properties. 2.23 Plastics. 2.24 Rod Not Negligible. 4.42 Impact by a Body Moving
Design Hints. 2.25 Miscellaneous. 2.26 Closure. Horizontally. 4.43 Elastic Impact on Beams. 4.44
Effect of Mass of Beam. 4.45 General Remarks on
Impact. 4.46 Closure.
TOLERANCES AND ALLOWANCES 79

3.1 Introduction. 3.2 Tolerance. 3.3 Allowance. SCREW FASTENINGS AND RIVETS 155
3.4 Fits. 3.5 Example. 3.6 Interchangeability. 3.7
Press and Shrink Fits. 3.8 Stresses Due to the Interference 5.1 Introduction.’ 5.2 KindofThreads. 5.3 Definitions.
of Metal. 3.9 Natural Spread of Dimensions. 3.10 5.4 Standard Threads. 5.5 Fits of Threads. | 5.6 Design
Example—Analysis of Actual Production. 3.11 Standard of Bolts—Initial Tension Unknown. 5.7 Initial Tension
Deviation and Area under the Normal Curve. 342 and Tightening Torque. 5.8 Materials and Strength of
Statistical Distributions of Fits. 3.13 Tolerances on Threaded Products. 5.9 Elastic Analysis of Bolts for
Locating Holes. 3.14 Tolerance and Surface Finish. Joints. 5.10 Elastic Constants and Gaskets for Connected
3.15 Closure. Parts. 5.11 Example—Stud Bolts for Compressor Head.
5.12 Example—Stiff Joint. 5.13 Types of Bolts and
Screws. 5.14 Set Screws. 5.15 Depth of Tapped Hole
VARIABLE LOADS AND STRESS CONCENTRATIONS and Clearance Around Bolt Head and Nut. 5.16 Bolts and
Screws in Shear. 5.17 Thread-locking Devices. 5.18
4.1 Introduction. 4.2 Mechanism of Fatigue. 4.3 Dardelet Rivet-bolt. 5.19 Rivets. 5.20 Closure.
Endurance Limit, Endurance Strength. 4.4 Charting
Endurance Strength. 4.5 How Stresses Vary. 4.6
Representation of the Endurance Strength in Terms of SPRINGS 183
Alternating Stress. 4.7 Estimation of Endurance Strengths.
4.8 Stress Raisers. 4.9 Theoretical Stress-concentration 6.1 Introduction. 6.2 Stress in Round-Wire Helical
Factors. 4.10 Notch Sensitivity. 4.11 Effect of Surface Springs. 6.3 Design Stresses and Solid Stresses. 6.4 Scale
Condition on Endurance Strength. 4.12 Endurance of a Spring. 6.5 Deflection of Round-Wire Helical
Strength as Affected by Size. 4.13 Variable-stress Equation Springs. 6.6 Accounting for Variable Stresses. 6.7
 CONTENTS CONTENTS
Energy Absorbed by a Spring. 6.8 Solid Height and Free 8.24 Additional Comments on Fatigue. 8.25 Curved
Length. 6.9 Design of Helical Springs. 6.10 Example— Beams. 8.26 Thick-shell Cylinders. 8.27 Force and
Average Service. 6.11 Example—Indefinite Service. Shrink Fits. 8.28 Closure.
6.12 Materials Used for Coil Springs. 6.13 Factors
Affecting Fatigue Strength of Coil Springs. 6.14 Relaxa- SHAFT DESIGN 263
tion of Spring Materials. 6.15 Goodman Diagram.
6.16 Tolerances. 6.17 Surgein Springs. 6.18 Buckling 9.1 Introduction. 9.2 Bending Forces Produced by Belts
of Compression Springs. 6.19 Concentric Helical Springs. and Chains. 9.3 Design of Shafts for Strength. 9.4
6.20 Compression Coiled Springs with Rectangular Wire. Example. 9.5 Shaft Sizes and Materials. 9.6 Hollow
6.21 Extension Springs. 6.22 Torsion Springs. 6.23 Round and Square Shafts. 9.7 Vertical Shear. 9.8
Other Kinds of Springs. 6.24 Flat Springs. 6.25 Leaf Torsional Deflection. 9.9 Transverse Deflections. 9.10
Springs. 6.26 Fatigue of Leaf Springs. 6.27 General Graphical Integration. 9.11 Example—Defiection of
Remarks on Leaf Springs. 6.28 Closure. Shafts. 9.12 Vibration and Critical Speeds of Shafts.
9.13 Shafts Design by Code. 9.14 Closure.

COLUMNS, CENTRAL LOADS 211

10. KEYS AND COUPLINGS 281
7.1 Introduction. 7.2, Euler’s Formula. 7.3 Effective
Length. 7.4 ShortColumns. 7.5 Straight-line Formulas. 10.1 Introduction. 10.2 Design of Flat and Square Keys.
7.6 Transition Point between Long and Intermediate 10.3 Example—Design of Flat Key. 10.4 Stress Con-
Columns. 7.7 RadiusofGyration. 7.8 Secant Formula. centration in Keyways. 10.5 Other Types of Keys.
7.9 DesignofColumns. 7.10 Example. 7.11 Equivalent 10.6 Splined Shafts. \ 10.7 Involute Spline. 10.8 Shear
Stress for Columns. 7.12 Other Column Formulas. Pin. 10.9 Rigid Couplings. 10.10 Example—Flange
7.13 Closure. Coupling. 10.11 Flexible Couplings. 10.12 Universal
Joints. 10.13 Over-running Clutch. 10.14 Closure.

COMBINED STRESSES 221

11. JOURNAL AND PLANE-SURFACE BEARINGS 299
8.1 Introduction. 8.2 Bending and Uniform Stresses.
8.3 Example—Design of Column with Eccentric Load. 11.1 Introduction. 11.2 Types of Journal Bearings.
8.4 Eccentric Loading on an Unsymmetric Section. 8.5 11.3 Thick-film Lubrication. 11.4 Viscosity. 11.5
Coplanar Shear Stresses in More Than One Direction. 8.6 Petroff’s Equation. 11.6 Hydrodynamic Lubrication.
Combined Normal and Shear Stresses. 8.7 Principal 11.7 Geometric Relations for a Clearance Bearing. 11.8
Stresses. 8.8 Maximum Shear Stress. 8.9 Element Load Capacity and Friction, Journal Bearings. 11.9
Subjected to Two Normal Stresses and Shear. 8.10 Mohr’s Optimum Hydrodynamic Bearings. 11.10 Example—Full
Circle. 8.11 Example—Combined Tension and Shear. Bearing. 11.11 Example—Optimum Bearing. 11.12
8.12 Theories of Failure. 8.13 Design Equation for Lubricant Flow through Bearing. 11.13 Energy Increase
Maximum Shear and Octahedral Shear Theories. 8.14 of the Oil. 11.14 Permissible Minimum Film Thickness.
Example—Combined Bending, Compression, and Torsion. 11.15 Example—Partial Bearing, with Temperature Rise.
8.15 Power Screws. 8.16 Pitchand Lead. 8.17 Torque 11.16 Clearance Ratio. 11.17 Length/Diameter Ratio.
to Turn Screw. 8.18 Coefficient of Friction of Power 11.18 Heat Dissipated from a Bearing. 11.19 Example—
Screws. 8.19 Efficiency of a Square-thread Screw. 8.20 Steady-state Temperature. 11.20 Operating Temperatures.
Conditions for a Self-locking Screw. 8.21 Design of 11.21 Oil Flow for Pressure Feed. 11.22 Friction Loss
Screws. 8.22 Combining Variable Stresses. 8.23 in Cap. 11.23 Significance of Zn/p. 11.24 Thin-film
Example—Combined Bending and Torsion, Varying Stresses. Lubrication. 11.25 Construction and Lubrication.
xxiv CONTENTS CONTENTS xxv

11.26 Materials for Bearings. 11.27 Semilubricated and Phenolic Gear Teeth. 13.29 Design of Cast Teeth.
Nonlubricated Bearings. ’ 11.28 Lubricants. 11.29 13.30 Hunting Tooth. 13.31 Hubs, Metal Gears. 13.32
Thrust Bearings. 11.30 Hydrostatic Lubrication. 11.31 Arms and Webbs. 13.33 Rimand Bead.. 13:34 Unequal
Gas-lubricated Bearings. 11.32 Dynamic Loading. Addendum and Unequal Dedendum Teeth. 13.35 Internal
11.33 Closure. Gears. 13.36 Gear Trains. 13.37 Efficiency of Gears
and Thermal Capacity. 13.38 Lubrication of Gear Teeth.
13.39 Closure. ‘ I:
12. BALL AND ROLLER BEARINGS as ie un) 333

12.1 Introduction. 12.2 Stresses During Rolling Contact. HELICAL GEARS... oe A a Oe SOE
12.3 Statistical Nature of Bearing Life. 12.4 Static Load
Capacity. 12.5 Dynamic Load Capacity. 12.6 Equiv- 14.1 Introduction. 14.2 Helix Angle and Face Width.
alent Dynamic Load. 12.7 Selecting Bearings from Tables. 14.3 Pitches. 14.4 Pressure Angles. 14.5 Dynamic
12.8 Example. 12.9 Choosing Bearings When the Prob- Load, Helical Gears. 14.6 Strength of Helical Teeth.
ability of Survival Is Different from 90%. 12.10 Examples 14.7 Limiting Wear Load. 14.8 Herringbone Gears.
—Probabilities and Lives of Rolling Bearings. 12.11 14.9 Crossed Helical Gears. 14.10 Closure.
Variable Loading. 12.12 Materials and Finish. 12.13
Sizes of Bearings. 12.14 Friction in Rolling Bearings. 12.15
Types of Rolling Bearings. 12.16 Thrust Bearings. BEVEL GEARS ES is = his wa .. 407
12.17 Housing and Lubrication. 12.18 Other Ball-bearing
Devices. 12.19 Comparisons of Sleeve and Rolling Bear- 15.1 Introduction. 15.2 Bevel-gear Nomenclature. 15.3
ings. 12.20 Closure. Strength of Straight Bevel-gear Teeth. 15.4 Tooth Pro-
portions, Bevel Gears. 15.5 Form Factor. 15.6 Dynamic
Load for Bevel Gears. 15.7 Rated Strength of Bevel Gears.
13. SPUR GEARS .. Ne BF eis ois ale i) | 355 15.8 Rated Wear Load for Bevel Gears. 15.9 Example—
Horsepower for Bevel Gears. 15.10 Coniflex ® and Zerol
13.1 Introduction. 13.2 Definitions. 13.3 Base Circle ® Bevel Gears. 15.11 Spiral Bevel Gears. 15.12 Hypoid
and Pressure Angle. 13.4 Pitch. 13.5 Length of Action Gears. 15.13 Other Types of Bevel Gears. 15.14 Forces
and Contact Ratio. 13.6 Law of Gearing and the Action ona Bevel Gear. 15.15 Design Details. 15.16 Materials
of Gear Teeth. 13.7 Interference of Involute Teeth. for Bevel Gears. 15.17 Closure.
13.8 Interchangeable Involute Gear Systems. 13.9
Strength of Gear Teeth. 13.10 Stress Concentration.
13.11 Design Stress. 13.12 Face Width. 13.13 The
Transmitted Load. 13.14 Dynamic Loads on Gear Teeth. WORM GEARING _.. is ais ai 3 .. 427
13.15 Dynamic Load as a Function of Velocity Only—Metal
Teeth. 13.16 Example—Spur Gears, Intermittent Service. 16.1 Introduction. 16.2 Pitchand Lead. 16.3 Strength
13.17 Buckingham’s Average Dynamic Load for Metal of Worm-gear Teeth. 16.4 Dynamic Load for Worm
Teeth. 13.18 Service Factors. 13.19 Permissible and Gears. 16.5 Wear Loadfor WormGears. 16.6 Thermal
Expected Errors. § 13.20 Example—Buckingham’s Equa- Capacity. 16.7 Relation of Normal and Diametral
tion for Dynamic Load. 13.21 Limiting Load for Wear. Pressure Angles. 16.8 Efficiency of Worm Gearing.
13.22 Example—Wear of Cast-iron Teeth. 13.23 Gear- 16.9 Coefficient of Friction, Worm Gearing. 16.10
tooth Wear. 13.24 Materials Used for Gears. 13.25 Separating Force between Worm and Gear. 16.11 Pro-
Example—Designing Steel Gears for Continuous Service. Portions for Worm Gearing. 16.12 General Remarks on
13.26 More on the Design of Gear Teeth. 13.27 Design Design of Worm Gearing. 16.13 Design Procedure.
of Nonmetallic Gear Teeth. 13.28 Example—Laminated 16.14 Materials for Worm Gearing. 16.15 Closure.
 CONTENTS CONTENTS xxvii

17. FLEXIBLE POWER-TRANSMITTING ELEMENTS 441 20. MISCELLANEOUS PROBLEMS 523

17.1 Introduction. 17.2 Net Belt Pull and the Variation 20.1 Introduction. 20.2 Thin Cylindrical Shells under
of Stress in Belts. 17.3 Capacity of a Flat Belt. 17.4 External Pressure. 20.3 Steel Tubes Subjected to External
Belt Thickness and Width. 17.5 Coefficient of Friction. Pressure. 20.4 Flat Plates. 20.5 Cams. 20.6 Fly-
17-6 Strength of Leather. 17.7 Length of Belts. wheels. 20.7 Example—Flywheel Rim for Punch Press.
17.8 Angle of Contact. 17.9 Belt Speed. 17.10 Initial 20.8 Stresses in Flywheel Rims. 20.9 Rotating Disks.
Tension. 17.11 Rated Capacity of Leather Belts. 17.12 20.10 Closure.
Example—Flat Leather Belt. 17.13 Maintaining Initial
Tension. 17.14 Analysis of Pivoted-motor Drive. 17.15
Rubber Belting. 17.16 Flat Belt Drives for Non-parallel LIST OF REFERENCES 541
Shafts. 17.17 V-Belts. 17.18 V-Flat Drives and
Others. 17.19 Variable-speed Transmissions. 17.20 APPENDIX 561
Toothed Belts. 17.21 Roller Chain Drives. 17.22
Example—Roller Chain Drive. 17.23 Inverted-tooth Table AT 1. Properties of Sections. Table AT 2 Moments
Chains. 17.24 WireRope. 17.25 Design Considerations and Deflections in Beams. Table AT 3. Typical Properties
for Wire Rope. 17.26 Example—Wire Rope for Mine of Some Nonferrous Metals. Table AT 4. Typical
Hoist. 17.27 Traction Drives. 17.28 Fittings for Wire Properties of Some Stainless Steels. Table AT 5. Typical
Rope. 17.29 Pulleys and Sheaves. 17.30 Harmonic Properties of a Few Plastics. Table AT 6. Typical
Drive. 17.31 Closure. Properties of Cast Ferrous Metals. Fig. AF 1. Properties
of Heat-treated AISI C1040. Fig. AF 2. Properties of
18. BRAKES AND CLUTCHES .. 481 Heat-treated AISI 3140. Fig. AF 3. Properties of Heat-
treated AISI 4340, Fig. AF 4. Relation of Hardness
18.1 Introduction. 18.2 Frictional Work and Power. Numbers. Table AT 7. Typical Properties of Wrought
18.3 Computing the Energy to Be Absorbed. 18.4 Per- Ferrous Metals. Table AT 8. Typical Properties of Steel—
missible Energy Absorption and Other Design Data. 18.5 Various Sizes and Conditions. Table AT 9. Typical
Example—Drum Temperature andfhp. 18.6 Block Brake, Properties of Heat-treated Steels. Table AT 10. Mis-
Small Blocks. 18.7 Forces for Long Shoes. 18.8 cellaneous Endurance Limits and Endurance Strengths.
Internal Shoe. 18.9 Band Brakes. 18.10 Disk Frictional Table AT 11. Typical Core Properties of Carburized Steels.
Torque. 18.11 General Remarks on Disk Clutches. Fig. AF 5. Reduction of Endurance Strength of Steel.
18.12 Cone Clutch. 18.13 Braking Materials. 18.14 Table AT 12. Values of K, for Screw Threads. Table
Coefficient of Friction. 18.15 Other Types of Brakes and AT 13. Values of K, for Keyways. Fig. AF 6. K, and K,
Clutches. 18.16 Closure. for T-head and Miscellaneous. Fig. AF 7. Average Notch
Sensitivity Curves. Fig. AF 8. Flat Plate with Central
Hole. Fig. AF 9. Flat Plate with Fillets. Fig. AF 10.
19. DESIGN OF WELDED JOINTS 505| Flat Plate with Grooves. Fig. AF 11. Flat Plate with
Eccentric Hole. Fig. AF 12. Shaft with Fillet. Fig.
19.1 Introduction. 19.2 Butt Joint. 19.3 Fillet Welds. AF 13. Shaft with Radial Hole. Fig.AF14. Shaft with
19.4 Fillet Welds, Eccentric Loading. 19.5 Example— Groove. Table AT 14. Unified and American Screw
Eccentrically Loaded Fillet Weld. 19.6 Annular Fillet Threads. Table AT 15. Nominal Dimensions of Various
Weld in Bending. 19.7 Design Stresses. 19.8 Design Gages. Table AT 16. Approximate Free Lengths and
for Fatigue. 19.9 Other Types of Welds. 19.10 Mini- Solid Heights. Table AT 17. Mechanical Properties of
mum Fillet-weld Size. 19.11 Types of Welding Processes. Wire for Coil Springs. | Fig. AF 15. Stress Factors (Wahl).
19.12 Testing Welded Joints. 19.13 Other Methods of Table AT 18. Stress Factors for Curved Beams. Table
Joining Metals. 19.14 Closure. AT 19. Key Dimensions. Fig. AF 16. Typical Viscosities
xxviii CONTENTS —

of Oils. Fig. AF 17. Minimum-film Variable vs.

Sommerfeld S. Fig. AF 18. Coefficient-of-Friction
Variable vs. Sommerfeld S. Table AT 20. Dimensionless
Performance Parameters for Full Jounal Bearings. Table
AT 21. Dimensionless Performance Parameters for 180°
Bearing. Table AT 22. Dimensionless Performance Para-
meters for 120° Bearing. Table AT 23. Dimensionless
Performance Parameters for 60° Bearing. Table AT 24.
Values of Form Factor Y in Lewis’ Equation. Table
AT 25. Values of C. Fig. AF 19. Maximum Permissible
Errors in Gear-tooth Profiles. Fig. AF 20. Expected
Errors in Tooth Profiles. Table AT 26. Values of Limiting
Wear-load Factor K,. Table AT 27. Wear Factor K,, for
Worm Gears. Fig. AF 21. Transmittance, Gear Cases.
Table AT 28. Properties of Wire Rope. Table AT 29.
Design Data for Brakes. Table AT 30. Design Stresses
for Welded Joints.

INDEX .. ie ae i. ie as as - 60oy
DESIGN OF MACHINE ELEMENTS
 I. STRESS ANALYSIS—
SIMPLE STRESSES

1.1 INTRODUCTION. A new machine is born because there is a

real or imagined need for it. It evolves from someone’s conception of a
device with which to accomplish a particular purpose. From the con-
ception follows a study of the arrangement of the parts, the location and
length of links (which may include a kinematic study of the linkage), the
places for gears, bolts, springs, cams, and other elements of machines.
With all ideas subject to change and improvement, several solutions may
be and usually are found, the seemingly best one being chosen.
The actual practice of designing is applying a combination of scientific
Principles and a knowing judgment based on experience. It is seldom that
a design problem has only one right answer, a situation that is often annoy-
ing to the beginner in machine design. Even though the art of designing
machines can be learned only through years of practice, many design
Problems in Problems on the Design of Machine Elements,* hereinafter
referred to as Problems, require elementary decisions on the part of the
Student. Certainly in the beginning, some decisions will be necessary in
the absence of a suitable background of knowledge—and this is not good.
But if the reader will keep thinking of the things about which decisions are
Necessary, there will be significant improvement, even in this short study.
* é
Practice problems are in this separate book by Faires and Wingren (also published by
acmillan), which includes, for convenience, all the tabular and chart matter in the
Ppendix of the present text.
 STRESS ANALYSIS—SIMPLE STRESSES [Ch.! §1.3] PHILOSOPHIES OF DESIGN
2 3

It is true that engineers must often make decisions without full knowledge, (iii) when the design should be tested in service tests before manu-
a difference between decision making in ignorance of facture starts, (iv) when special measures should be taken to control
but let there be
knowledge that exists and decision making in ignorance of knowledge vibration and sound (and others).
that does not exist. ’ (f) Some aesthetic sense, because the product must have “cus-
Engineering practice usually requires comprom ises. Competi tion ma: tomer appeal’’ if it is to sell.
require a reluctant decision contrary to one’s best engineering judgment; (g) A knowledge of economics and comparative costs, because
production difficulties may force a change of design; etc. Watt’s famous the best reason for the existence of engineers is that thoy save
straight-line mechanism was a production compromise that resulted from money for those who employ them. Anything that increases the
the inability of the shops of his day to produce a plane surface economi= pa should be justified by, for instance, an improvement in per-
cally. Although it does not impart a true straight-line motion, Watt’s in, the addition of an attractive feature, or greater
mechanism guided the end of the piston rod closely enough to a straight
line for practical purposes at that time, and it could be economically (h) Inventiveness and the creative instinct, most important of
manufactured. | all for maximum effectiveness. Creativeness may arise because an
In a general sense, a designer may be one who designs anything ; a energetic mind is dissatisfied with something as it is and this mind
folding chair, an auto body, a pattern for dinnerware or silverware, Q is willing to act.
fountain pen, stage scenery, or ladies’ dresses. However, as the title of this Naturally, there are many other important considerations and a host
book implies, we shall limit our coverage to those topics of greates of details. Will the machine be safe to operate? Is the operator protected
importance to a mechanical engineer, whose design activity is so ofter from his own mistakes and carelessness? Is vibration likely to cause
devoted to machines and systems of machines. trouble? Will the machine be too noisy? Is the assembly of the parts
relatively simple? Will the machine be easy to service and repair? :
Of course, no one engineer is likely to have enough expert knowledge
1.2 THE MACHINE DESIGNER’S RESPONSIBILITY. A goot concerning the above attributes to make optimum decisions on every ques-
designer needs many attributes, for example: tion. The larger organizations will have specialists to perform certain
(a) A good background in strength of materials, so that th functions, and smaller ones can employ consultants. Nevertheless, the
stress analyses are sound. The parts of the machine should hay more any one engineer knows about all phases of design, the better. Biesign
adequate strength and rigidity, or other characteristics as needet 1s an exacting profession, but highly fascinating when practiced against a
(b) A good acquaintance with the properties of materials usé broad background of knowledge. Engineering is design.
in machines, which is no small achievement considering the exten
sive developments of recent years. a
(c) A familiarity with the major characteri stics and economik atin OF DESIGN. We think of an inventor as
of various manufacturing processes, because the parts that make cesta m nem and creating a new design. However, even though he
the machine must be manufactured at a competitive cost. It hap; been mac ine never before conceived, he uses ideas that have long
that a design that is economic for one manufacturing plant ma : known and, in more or less degree, he benefits from the engineering
be so for another. For example, a plant with a well-develo xperiences of one or several industries.
welding department but no foundry might find that welding is t 4 ee Be a follow a pattern familiar to and typical of an industry:
most economic fabricating method in a particular situation; whe re one, 2 el sewing machine is generally quite similar to the previous
another plant faced with the same problem might decide u Change a model automobile is similar in most respects to the old.
casting because they have a foundry (and may or may not have for tie se iy experience with the old model) are introduced either
welding department). 4 bcm: - Bees of improving the machine or to gain or maintain
an eco-
(d) A specialized knowledge in various circumstances, such } 7 si nate advantage in the market.
the properties of materials in corrosive atmospheres, at very o sad Ait : -o. approach to a particular design depends somewhat
(cryogenic) temperatures, or at relatively high temperatures. bistgee a or the kind of machine. A chemical plant, which is
(e) A preparation for deciding wisely: (i) when to use manufat bely a es em machine, may be a one-shot proposition. One plant
turers’ catalogs, buying stock or relatively available items, and whi ected gned an built. If the design is not right, the mistakes are cor-
custom design is necessary, (ii) when empirical design is justifie On the job, an expensive but necessary procedure, until
the plant
 STRESSES [Ch. 1 1.6] MACHINE DESIGN IS ENGINEERING 5
4 STRESS ANALYSIS—SIMPLE

product is that the designer needs to continue to add to his knowledge of theory in
operates as planned. The designer who works where only one order to be in a better position to judge. When it is difficult to incor ie
different,
made from the design develops attitudes or philosophies quite
the results of experience into a theoretical equation, we often ilo
for instance, from an airplane or automotive designer. the
use of practical, modifying constants until the difficulty is resolved. Hence
of utmost
In the airplane industry, light weight and reliability are if experience dictates certain portions of a design, it behoves us to use
to relatively
importance. The philosophy of the airplane designer leads him experience as a guide until a more satisfactory state of theoretical know-
worth the
high-precision (and high-cost) designs, because the results are
d under
money. Often the designed product is manufactured and operate ledge is attained. If the machine is almost entirely new and different, as
before th a rocket engine was a few years ago, look for related experience. So much
actual or simulated actual conditions, perhaps repeatedly,
the designet
design is considered acceptable. In the automotive industry, information is still uncoordinated, so much remains to be dearnes that
A sub: the student, particularly in design, should
wants to be sure that his design is suitable for mass production. adopt an agnostic attitude
in quantit ie welcoming further investigation. i
assembly design, such as the transmission, which will be made
of hundreds of thousands or in millions, will be tested under actua
operating conditions, because the “bugs’’ need to be eliminated befor
. 1.5 OBJECTIVE OF THIS BOOK. From the foregoing discussion
mass production begins.
e vessel we see that the subject of machine design is too broad to be covered in one
In heavy industries, such as the manufacture of large pressur
book without being excessively superficial. The material given in this one
the designer does not think in terms of the precision necessary in an aif
l : — from an immense body of literature, and many of the subjects
plane engine, nor is he particularly concerned about the weight. Moreove ee Me
sense. fields in which an engineer sometimes spends a lifetime
there is no mass production in the automotive
Design problems have more than one answer. Given a general stat
th Tempered by current “practical” knowledge, our major objective is to
ment of a design problem, such as design a machine to wash clothes in learn to apply basic theories of strength of materials and other pertinent
home automatically, and there will be as many differen t answers as ther
on th theories to the actual design of the most common machine elements:
are design teams—as attested by the number of washing machines
' est ae shafts, springs, etc., especially as they are affected by the
market.
ay a Ne ee (By the very nature of most machines, the loads vary.)
These brief remarks are intended, not to define the philosophies
q i ere uggest esign procedures that are expected to give safe answers,
design in each of the industries mentioned, but to show that there are
ee oo important consideration is to have minimum weight, and
different attitudes and to suggest that, in each field of design, the design a other reasons, more elaborate theories and special test results,
evolves a philosophy befitting the nature of the work he is doing. ies oe nae herein, should be searched out. The reader should
eae specialized attitudes with respect to the various machine
a re voluminous in detail. As we proceed, we shall not forget the
1.4 THEORY AND PRACTICE. If theory and practice do not agre
y requirements of § 1.2; others will be highlighted now and then.
either theory or practice is wrong. Methods of design undergo an evol
be Ht
tionary process, just as a machine invariably evolves into better and
forms. New discoveries are made each day, but, because many theori
1.6 frie
om DESIGN IS ENGINEERING.
are or become inadequate, we never know when the accepted formula W With some notion of
ei nae : ry of machine elements, we may begin the calculations. From
be discarded.
Bera wa e work done or power consumed, we compute forces on
In any derivation, we first make certain assumptions in order to simpli
Poe epee of Positions of the machine’s cycle, using the prin-
the work and to obtain a formula which appears to satisfy our req it
failv SS licasst cue Te ene each element so that it will perform
ments, but before long we often find that the formula fails. This
more of o
results in renewed study, and we usually find that one or
for a formu cr ae use the principles of strength of materials, but
original assumptions were not justified. We then search this ¢anne
respect Be e thought of as a review of those principles. Rather,
with new variables that will care for a new environment. With it is
economic to design by nie n to enaerene problems with the
aim of finding suitable
the use of theory, it is not by any means always dim a
al analysis available, a - ae ine elements. In the process the designer makes a
most comprehensive theoretical and experiment Stre.
answer the question of whether a cert# ee , deciding upon the points in the various parts subjected to
engineering judgment must
mum stress conditions (and the kind of stress).
design decision is worth $10-15 or $10,000—15,000, or what. This meat Even though only
 STRESS ANALYSIS—SIMPLE STRESSES [Ch.1
6
FIGURE 1.2
the simple stress equations F = sA, M = sljJc, and T = sJ/c, are involved,
the consequences may be involved, as we shall see.
ne
Since it is seldom possible to use any theoretical equation to determi D
e
a dimension and adopt the result unthinkingly, the important requisit
eventual
now is judgment. Computed results only provide evidence for sections is said to be in shear, and the stress is called a shear stress s,; its
decisions. Other considerations are involved. The purpose of this chapter average value is ,
stress
is to guide thought toward the designer’s point of view, to discuss
analysis, and, incidentally, to review a few of the fundamentals. Boy /
Ss; = yee or ksi,

where, for this particular case, A is twice one sectional area of the pin
1.7 STRESS. The word stress as used in this book shall always be
because both the areas, at MM and M’M’, offer resistance to the shearing
understood to mean unit stress s, measured in pounds or kips per squar e action of the load, that is, s, = F/(27D?/4) in Fig. 1.2. The pin is said to
inch (psi or ksi).* A kip is the same as 1000 Ib. (kip = contraction of
{ be in double shear. This pin is also subjected to bending and compressive
kilo-pounds).
Ly stresses. See the analysis of the pin in the example of § 1.21. Pure shear
Recall the normal stresses of tension s, and compression s,, Fig.
can be obtained only from torsion (§ 1.13).
and the corresponding stress equation for a part with an axial load (without
bending):
18 TENSILE STRENGTH AND YIELD STRENGTH. When a
F
member is subjected to the action of a force, it is deformed,
| oy

(1.1) sy = and Ss, = 7 no matter

how small the force. A test specimen subjected to increasing stress will
[FIG. 1.1(a)] [Fic. 1.1.(b)]
undergo increasing strain. Recall, by referring to Fig. 1.3, some of the
characteristics of stress-strain curves. By strain (tensile or compressive)
where A sq. in. is the area that offers tensile or compressive resistance
we mean the deformation per unit gage length,
the load F, and where we observe that the stress is an average value reveal in. per in. It is thus a
ing nothing about how it may vary in detail because of various departure
FIGURE 1.3 Comparison of Stress-strain Diagrams.
(Scale of strain beyond the dot is 10 times scale up to 150
FIGURE 1.1 Tensile
Th - dot.) The modulus lines (see Fig. 1.4) have been
7 Average Compressive Stresses.
A wn to the same scale and show relative values; the
-
E. Stress cross section may be any shaj remaining part of the curves should be considered as
” anes
Q—1- wk other than circular, but th 130 a
=
line of action of the force
‘emt qualitative but typical of certain metals. Some soft
a ire F must coincide with the ce
oo alloys elongate more than soft steel. See 120
Heat Treated
Alloy Steel
. miralty metal, Table AT 3. As the carbon content
(a) (b) (c) troidal axis of the body if the c
is from some low value, the upper and lower no
is to be no bending (§ 8.2).
enn into one YP (almost horizontal), and then sg
Fig a with increasing hardness of the steel. See 4 ud
from the ideal—an ideal consisting of a straight member of homogeneou 5 90 30% Elong.
material free from residual stresses, uniform in cross-sectional dimensions a
i 80 a Elong.
the area A is not close to the point of application of a perfectly centré
and static load. Since this ideal never exists, the actual stress distributiol & 70 Aluminum Alloy
will not be uniform, but will have a random characteristic suggested DB
g 60
Fig. 1.1(c).
a) 50
The load F, Fig. 1.2, produces a tensile stress in the parts B but tend Soft Steel
to cut across the pin on the sections MM and M’M’. The area of thes
Upper 1 Magnesium Alloy
Yield > ae ie Elong
* The student is advised to practice using kips because of the convenience of the smalle
Y 5 ng.
numbers.
— 20H | &<— Cast Iron
 STRESS ANALYSIS—SIMPLE STRESSES [Ch. 1 41.10] BENDING 9
8
from a straight line. It is indicated approximately by point P in Fig.
percentage increase (or decrease) in length. If a strain is 0.004 in. in a gage 1.4.
q Recall that for some materials, for example, cast iron Fig. 1.3 little io
length of 2 in., the unit strain is 0.002 in./in., or 0.2%. none of the stress-strain curve is straight. iF
The ultimate stress Sy, or-tensile strength, the highest point on the stress=
Below the proportional limit, stress s is proportional to strain ¢, and
strain curve, Fig. 1.3, is the maximum load divided by the original area the proportionality constant in tension is called the modulus of elasticit
before straining occurs. The stress at which a low or medium carbon steel E
the slope (s/e) of the straight part of the stress-strain curve, Fig. a
undergoes a marked elongation without an increase in load is called the :
yield point YP, Fig. 1.3. (1.2A) S = Ee psi or ksi.
a
Higher carbon steels and nonferrous metals generally do not have
characteristic yield point. For these, there is a defined yield strength, which Since the strain ¢ is dimensionless (in./in.), the units of E are the same as
is the stress for a specified deviation from the straight part of the stress- those of s. In terms of total deformation 5, where § = eZ, the normal stress
strain curve. The yield strength is determined by drawing an offset line; in simple tension or compression is
Fig. 1.4, starting at a certain offset 4, parallel to the straight part of thi
; F
s-e curve, and noting where it crosses the stress-strain curve B. The ordinat (1.2B) s= ee or aa or eee
at B is the yield strength s,. (We shall use the symbol s,, whether or no AE’
the material has a characteristic yield point.) The amount of the offset is in which L is the total length undergoing a total deformation of 8
usually 0.2°% (0.002 in./in. of strain) for steel, aluminum, and magnesiun The modulus of elasticity for ordinary steel usually falls hetween 28 to
alloys. Sometimes, the yield strength is specified for a particular tota 31 million pounds per square inch (psi), and most designers use either
extension (deformation or strain), Fig. 1.4; for example, usually 0.5%
29 x 10° or 30 x 10® psi (29,000 or 30,000 ksi). See Table AT 7. Since
extension for copper base alloys (see Table AT 3).* some alloy steels have values of E below 30 million, it may be desirable to
find more exact values where E would be significant in the design
The modulus of elasticity is a measure of stiffness. For particular values
1.9 MODULUS OF ELASTICITY. The elastic limit is the maximuf
of stress s and E, we see from equation (1.2) that there results a certain
stress to which a standardized test specimen may be subjected without strain and, therefore, a certain total deformation of an actual part.
permanent deformation. The specimen returns to its original length” It is
interesting to note that, since all grades of steel have about the saingiysins
subjected to stresses below the elastic limit. The proportional limit is, fc
but”
of E, there would be little or no reduction in deformation if a high-strength
practical engineering purposes, coincident with the elastic limit,
alloy steel were substituted for a low-strength carbon steel. From sei
precisely defined as the stress at which the stress-strain curve deviat ES of view, the unit deformation can be reduced only by reducing
* See Appendix. the
ce or by choosing a material with a higher value of E.
Bins | minimum deformation) is an important criterion in many
FIGURE 1.4: Yield Strength. At 0.2% offs eet er as those of machine tools for accurate work (lathes, milling
€ = 0.002; etc. The yield strength at 0.2% of s
m » &tc.), rotor shafts in motors and generators and turbines.
is found by drawing the line AB from the value:
0.2 at A parallel to the straight part of the s- € modulus of elasticity for most metals in compression is usually
curve, called the modulus line; the intersection
# AB with the s—e curve at B is taken as the yit
& strength. In this particular case, the yield stre ig
&o
g at 0.2% offset and at 0.4% extension is the sam
s, Kips per Square Inch

S The yield strength of nonferrous alloys is sometimt 1.10 nte

= Svi ;
ae yo (a). Bending,
<> stated in terms of percentage extension. The § or flexure, produces two kinds of
called proof stress is at the intersection of a vé 3 sion on one side of a neutral
2
f small offset line and the s—e curve, usually n the other. From strength of materials, we
oe
pl ane and com
ney
i
offset of 0.01%, as at C. The proof stress is set
z yie
Py to be closer to the elastic limit P than the
strength. Since there are different bases of (1.3 ) 2= ie
eee y=
Mc : and ie Mie psi or ksi,
mining the yield strength, a yield strength shou [BELOW x-x, FIG. 1.5] [ABOVE x-x, FIG. 1.5]
be accompanied by its basis of determination. Hoy
ever, this fact is not always made clear in i
Where: Ss t (or s,) is the stress at some point in the beam (Me/J is a maxi-
literature.
>
 g1.10] BENDING é "
wv SN oN ee eee ————

in Fig. 1.5); M in-lb. or in-kips is the bending moment at the section of

the beam that contains said point; c, is the distance from the neutral plané tensile strength, as for example cast iron (Table AT 6), there are two cases:
1. Symmetric sections—use the ultimate flexural stress (bending
to a point on the tensile side, and c, is measured to a point on the compress
modulus or modulus of rupture in bending) as a criterion for the design
sive side (if the section is symmetric, the distance of the external fibers i
both directions is c = c; = ¢,); J in.* is the centroidal moment of ine stress (see cast iron in Table AT 6);
2. Unsymmetric sections—use different design stresses in tension s4, and
(that is, J,; Fig. 1.5) of the section containing said points; the material i
(in the compression S4,, and equations (1.3). For cast iron, the usual procedure is
homogeneous; the beam is straight in the longitudinal direction
the sai¢ to design for the design tensile stress and then check the compressive stress.
unstressed state) and the neutral and centroidal axes coincide;
Recall from strength of materials that the radius of curvature r after
point is not located in the vicinity of the point of application of a force 01
the straight beam is bent is related to bending moment by
of a discontinuity of section (as at B and P, Fig. 1.5); the loading is stati
or gradually applied; there are no residual stresses (otherwise s, = Me/I it 1 M 1 Ss E
(4) == EL’ Sra
r
aE
cE
or yu =, [Fig.1.5(d)]
-r= it Position of Beam
—- under Load Neutral ~ Be where the last two forms are obtained by using M = s//c. Both equations
“Tle Wet | > 7 Plane & rie (1.3) and (1.4) are virtually true in straight beams if the maximum stress
ro Toa B \ye°
aoe le PBJ \
sth x Peake does not exceed the proportional limit, and for (1.4), r should be large
compared to c.
0 } f A,
(b) The shearing stress in a beam, distributed as shown in Fig. 1.5(c)
—T—HeutralAxist Section 4-B e t for a rectangular section, is computed from
R, p A Re
(a) (b) (c) (d)
va
(1.5) =
FIGURE 1.5 Stresses from Bending. The figure (b) shows the distribution of tensile an Ib
compressive stresses at some section AB in (a), as computed from Mc/J. At some sectio where V (lb. or kips) is the shear (from shear diagram, § 1.11) at the
such as CD, other than the section of maximum bending moment, and at some point;
other than an external fiber, the nominal or computed stress is s = Mc/I, where M = Ry
section under study, b is given in Fig. 1.6, J is the centroidal moment of
in this figure, J is the centroidal moment of inertia of the section CD with respect to th inertia as previously defined (J, in Fig. 1.6), and Q = [ y dA’ = pA’, where
neutral axis, approximately X-X, and c is the distance from the neutral axis to point P A’ is that area “‘outside”’ of a transverse line through the point whose stress
The distribution of the vertical (and horizontal) shear is suggested by (c) —and equatia is desired, partly shaded in Fig. 1.6. Thus equation (1.6), with the same
(1.6). In (d), ¢ must be very much smaller than r. restrictions as given for (1.3), gives the approximate average stress along a
line such as BB, Fig. 1.6. For the rectangular section in Fig. 1.6(a),
the change of stress due to M); the beam does not twist; flanges (if any, Q = gh(h — g)/2. In the case (b), divide the A’ into two rectangular parts
as in an H-beam) do not buckle (§ 1.24); the shear stress (vertical shear) A, and A: and get the sum of the moments of these parts for Q. It is impor-
Fig. 1.5(c), is negligible compared with the flexural stress; there is ne tant to observe that if the section is symmetrical, the flexural stress is zero
longitudinal component of the forces on the beam; and stress remaint when the vertical shear is a maximum, and in any case, the shear is zero
proportional to strain (Hooke’s law—s, < proportional limit). where the flexural stresses are maximum (external fibers). Moreover, a
The ratio J/c is called the section modulus Z (Z = I/c) and is qui metal beam has to be unusually short for the vertical shear to be significant,
convenient to use for symmetric sections. Expressions for J and Z for somi but it should be looked into for short beams (and wooden beams), and on
sections commonly met in machine design are given in Table AT 1.* If the
section is not symmetric, the centroid of the section must be located and FIGURE 1.6 Distribution of Shear Accompanying Bending. Equation (1.6) becomes more
c, and c, computed for use in equation (1.4). ee with an increase of width with respect to the depth; hence the s, distribution for
When the metal is ductile, the design stresses in tension and compres- >a Ai is undependable. Recall that the shear stress on the exposed underneath part of the
sion are usually taken the same, no matter what shape the section has. ange (at H) is zero.
When the material has significantly greater compressive strength than
* All table and figure numbers preceded by A are in the appendix at the rear, pla
together for convenient reference.
 STRESS ANALYSIS—SIMPLE STRESSES [Ch. 1 g1.11] MATHEMATICAL RELATIONS FOR BEAMS; SHEAR CENTER 13
12
stresses at
occasion, the combination (Chapter 8) of shear and normal 1.11 MATHEMATICAL RELATIONS FOR BEAMS; SHEAR
be signific ant to CENTER. Let the deflection of a straight beam be represented by as
some interior point of a beam should be checked. It may tht
, the maxim um then for small elastic deflections (the corresponding
ponding slopes
note that in a member subjected to torsion and bending sl are small
to the torsion al stress,
vertical shear at the neutral plane adds vectorially
amount of
say of the front or back fiber. The shearing stress also affects the d
signific ant only for sho (1.7) Slope 9 = = radians
deflection of a beam, and again, this effect is
beams."-1}
a few times
(c) In a beam where the maximum loading is applied only
response is
during its expected lifetime, designing for some inelastic (1.8) MomentM = Er = pe.
ive yield point (Fig. 1.3), such as dx dx?’
permissible. If the material has a distinct
stress s, first
low-carbon structural steel has, the fibers that reach this 3
stress; and while the loading is
will remain at more or less constant
shown in Fig,
(1.9) Sher ve ee
increased, additional fibers will become stressed s,, say as dx dx?”
de)
1.7(b). Here, the beam is stressed to the yield point to a depth ab (and
the beam could be designe d for this
with elastic action from 5 to c, and (1.10) pakanges A, Cie ey
design, called limit design, is based
distribution of stress. However, when Ix dx? dx*
that
on plastic action, a common assumption is that the loading is such
Thus if the loading (or shear or moment) can be expressed as a function of
x, successive integrations eventually yield the deflection y. A rational sign
a 85 convention is needed in applying these equations: the shear at a section is
[> positive if the part on the left of the section tends to move “upward” (hori-
W-

lee

FIGURE 1.7 Plastic Bend ing zontal beam) relative to the part on the right; the bending moment and curva-
>=

ture are positive when the “top” of the beam is in compression. More on
lb

4
-

Ls
Ae

obtaining deflections of beams with several loads (graphical integration) is

yi
ae a Chapter 9. For the more elementary problems that we shall be con-
(a) (b) (c)
a iBis a Pca of certain common types of beams are given
the material at the neutral plane has just been stressed s, as in Fig. 1.7(¢
see! ee ee use, recall that the bending moment at any section of a
In this case, the forces on each half of a rectangular cross section are
th te coateat : er . the moments, about an axis in the section and through
to be syA/2, which gives a resisting moment of (s,A/2)(h/2); or with
ae ‘ f a external forces acting to the right of the section (or to
applied moment M equal to the resisting moment, we have
ree ‘ Section) where the moment is desired. To determine the
bhh bh? tomede o ing ee ona beam, use the principle that the maximum
(1.6) = §,——-=S
#2 Avie
—,
tatd The ce oo where the shearing-force diagram crosses the zero axis.
Pe ae is first to draw the shear diagram. For many if not most
vs. M = sZ = sbh?/6 for all elastic stresses (Fig. 1.5), where Z =} 2)
oe 7 : : can be a freehand sketch made from actual values of the
for a rectangular section. On this assumption, the beam can tolerate fn cee patie at any section for coplanar force systems is the algebraic
Hov
maximum bending moment 50% greater than for all elastic action. ca ee ioe perpendicular to the neutral axis on either side of the sec-
that @
ever, machine parts are so often made of the higher-strength steels Barong vide M at each section where the shear diagram passes
not exhibit a true yield point (no constant stress with increasing load) ami
eugene pee one such section the moment will be a maximum. The
they are too commonly subjected to varying loading for an indefinite time aa Sk the section of maximum moment without fail,
in which case, assumption of plastic action as a basis of design is downrigl dss he ad of stress raisers elsewhere and for other reasons, he may
strue
dangerous. For more detail about limit design, see books devoted to e noKee moments at other sections. Given the loading as shown in
tural design and strength of materials.{-4:1.°* Raia be Pi par of analytic mechanics"™-®] and compute the
at the real
* Superscript numbers in brackets designate references in the list of references ibiéses 1 ni ‘2. The shear diagram can be sketched, but to determine
that thi
they indicate either that more detail is available in the indicated reference or act location of section C of maximum moment in this case, set up
factual information comes from the reference, or both.
 w \d./ft.
—— F b

51.13] TORSION ‘is

-—e

A 1B L Cc:
Ri bem l-z-2] R FIGURE 1.8 Shear and Moment Diagrams. At 4, M = by using the parallel axis theorem than to copy a formula from a table.
eae | Rix. At B, M = Rixa—Fe. While the positive direction is The procedure may be in steps as follows:
R F 1 M max customarily toward the right, it is often convenient to (a) Locate the centroid of the section. If the section is symmetric
; measure from the right and toward the left; thus, at C, M = about any axis, the centroid lies on that axis. If the section is sym-
Shear dM _| Rex—wx?*/2. ; 1
V=0 metric about two axes, the centroid lies at the intersection of the
dee Re two axes. If the section is non-symmetric, divide it into a group of
Max
rectangles, triangles, the locations of whose centroids are known,
M=0 naming them A,, Ag, Ag, etc. Let A represent the total area whose
Bending Moment
centroid is desired. Then
(b) = A® = Ayxy + Agxg + AgXg t+ °°*,
where A= A, + Ag+ Agt:'';
ae x, is the distance from the centroid of A, to any convenient moment
axis, X2 is the distance from the centroid of A, to the same moment
FIGURE 1.9 Shear Center. Beam length is perpendicular axis, etc., and X is the distance from the chosen moment axis to the
a td ce : h
pia Le Ke to paper. centroid of the composite area. Solve for x.
(b) Knowing the location of the centroid of the composite sec-
4 tion, find the moment of inertia of each “fundamental” area (A,,
(b) Ag, ete.), about the centroidal axis of the composite section, using
the parallel axis theorem.
the equation for shear measured from the right-hand end (V = Rg — wx ), (c) I, =I, + A,d,?,
equate to zero and solve for x. See the caption to Fig. 1.8 as a reminder
of the process of computing moments. More on shear and moment dia- where J, is for area A, with respect to the composite centroid, J; is
grams is given in Chapter 9. for area A, with respect to its own centroidal axis, and d, is the dis-
If the beam section is unsymmetric with respect to the plane of the tance from A,’s centroidal axis to the composite centroid. Apply
loads, as for the channel loaded as shown in Fig. 1.9(a), one must be sure equation (c) to each “fundamental’’ area and then find
that the lines of action of the loads are correctly located if twisting is to bi (d) f=h+ht+ht:::,
avoided. In short, they should pass through the shear center, which i
defined by a force analysis of the internal forces,-” and is always on any where Tis the centroidal moment of inertia of the section, the value
existing line of symmetry. For an angle section, Fig. 1.9(b), the sheai of Jin s = Me/I. (Notice also how the value of J, was obtained for
center is at the intersection of the center lines of the legs. For the chann el the H-section (5) in Table AT 1.)
with dimensions as shown, the shear center S is located by
1.13 TORSION. :
i the simple
3g
torsion equation, (a) The only y shapeshap of section n f for which
(@) © 6 + khi(gt)
(1.11) Ti
=5,2',
1.12 FINDI THE MOMENT NG OF INERTIA. _Centroidal moment
moduli may be found in handbooks (and in books on biotin is circular (hollow or solid); T (in-lb. or in-kips to
of inertia and section
for the most common sections, some of which ar Goal a the units of 55) is the applied torque or twisting moment,
analytic mechanics)
1. To save time, memorize those for circles and rec Bh deahceti ai) isa sieoune (torsional) stress, which in designing is a design
given in Table AT
tangles. However, for composite areas, that is, areas made up of two 01
beter, in,* (= aD /32 for circular area—Table AT 1) is the centroidal
H- Sie oment of inertia of the section; c in. is the distance from the neutral
more “fundamental” areas, such as T-sections (two rectangles) and
moment of inertia © the point where the stress 5, is desired (usually the maximum stress
sections (three rectangles), it is better to find the desired

14
 16 STRESS ANALYSIS—SIMPLE STRESSES [Ch. |

Z’ = J/c in.? (= a D?/16 for circular area—point —

zs .{ \
on an external fiber); and FIGURE 1.11 Shear Strain. ig \y v
on circumference) is the polar section modulus. The conditions given for \

equation (1.3) also apply to (1.11) except that the member does twist but ‘—— d ——
&

does not bend. Equation (1.11) for an external point on a solid circular
member becomes the angle y is very small, the unit strain is taken as y = 8,/L. Within the
16T ., ksi proportional limit, strain is proportional to stress,
(e) S, =D psi or ksi.
(1.12) 5S, = Gy or Y G’
(b) Since plane sections do not remain plane in twisting of noncircular
members, equation (1.11) is not exact for these."!:2:1-7] However, more where the proportionality constant G is called the modulus of elasticity in
precise analyses suggest that the use of particular values of Z’, as for shear (also modulus of rigidity and transverse modulus); Fig. 1.12.
rectangular and elliptical sections in Table AT 1, give reasonable approxi- (d) In many situations, the design of a part is defined by the permissible
mations. The Z’ for the rectangular section varies with the ratio of deformations, rather than by some safe stress. The angle of twist 6 of a
h/b;"-7 the value in Table AT 1 is good for about h/b Z 2.2. The maximum shaft between two sections M and H, Fig. 1.12, is given by
torsional stress occurs at the center of the long sides. For thin-walled
tubes, Fig. 1.10(a), equation (1.11) becomes difficult to evaluate. The
(1.13) d= Fe meiane,

iOx € t

FIGURE 1.10
which is obtained from the arc A,A, = (D/2)@ = L
equations (1.11) and (1.12). Notice that, e : Ais a
i

moment on it, 0, the angle that one section M twists through with respect
to another section H, depends on the distance L in. between sections: J in.*
(b) is the polar moment of inertia; T = FD (in Fig. 1.12) in-lb. or Sickie
3 ater, G is usually taken as 11.5 x 10® psi (11,500 ksi), sometimes
membrane (soap-film) analogy [or letting ¢<r in (1.11)] results in the a Mie ee Equation (1.13) applies to hollow or solid round shafts [but
approximate equations"™-7! Ghee an wa ‘ see equation (f)]. Pulleys and gears (and keyways) on shafts
8 . “ e of twist 6, but common practice often ignores these effects,
J d 0 = ——_, , Be. ste e 7 L between two gears, for instance, as the center distance,
(f) Ss, = an
G(2nr°t)
* anrit ae a en deflections must be computed accurately, such effects
oe cae ‘ isregarded. If a member is subjected to different torques, it is
where r = mean radius of tube wall, t = thickness of tube wall, G is the
ite e torque is constant between sections where twisting loads are
modulus of elasticity in shear (see below) and @ = angle of twist (Fig. 1.12), piled, in which case, equation (1.13) applies between these sections. If
Equation (f) says nothing about the thin wall buckling locally before the member ch : ; ;
elastic stresses are exceeded, a type of failure that needs to be checkec
diameter he anges diameter, (1.13) is applied for lengths of the same
when a significant torque is applied. The big surprise comes when the thit
wall tube is slit along its length, Fig. 1.10(b), and its loss of torsiona
capacity as compared with the solid tube is noted :@-7) FIGURE 1.12 Twisted Rod. Torque = 7 = FD.

«) aT 4 37
Ss = an = —_—:>
. ° Qart? G(2zrt®)
where the resistance is equivalent to that for a long slender rectangle d
the symbols mean the same as before.
(c) Recall that shearing stresses on an element appear as indicated im
Fig. 1.11. With respect to the base ad, the top of the element is deforme¢
an amount 8,. The unit deformation or strain is 6,/L = tan y; but since
18 | STRESS ANALYSIS—SIMPLE STRESSES [Ch. | gl.16] FACTOR OF SAFETY—DESIGN FACTOR 19

(e) The theoretical relation between the tensile and shear modulii strength of a material cannot be represented by a single number, because
(E and G) is given by (when the deformation angle = tangent of angle) ~ its ability to resist the action of loads and forces depends upon the nature
of those loads and forces, the kinds of stresses induced, and other circum-
E
(1.14) G = ——__., stances. ’ :
2(1 + #), If a member is stressed beyond its elastic limit, the permanent deforma-
where p is Poisson’s ratio. See Tables AT 3, AT 6, and AT 7 for values; tion it receives may render it unfit for further service. Thus, the elastic
» = 0.3 is often used for steel. limit is one significant criterion of strength. Instead of the elastic limit,
however, we invariably use the yield strength, a stress that is generally not
too far from the elastic limit and much easier to determine experimentally.
1.14 TORQUE. The horsepower equation is used so frequently that (§ 1.8 and Fig. 1.4.)
The ultimate stress is also an important criterion of strength because a
we had better review its origins briefly. Assume a force F, Fig. 1.13, acting
at the circumference of a circle of radius r inches. (The size of pulleys, part that has ruptured has almost certainly lost its usefulness. There are
other criteria which we shall take up later—for example, ability of a material
to absorb energy without failure, endurance strength, buckling strength,
creep strength, and excessive deflection.
All criteria of strength are modified in some way in order to obtain a
design criterion. In its simplest form, the design criterion is a design
stress, or working stress, which may also be called a safe stress or allowable
stress. The stress used in design must be a safe one to use for computations
if failure is not to occur, and such a stress is said to be allowable.

1.16 FACTOR OF SAFETY—DESIGN FACTOR. Ordinarily, the

factor of safety is a number that is divided into a criterion of strength in
order to obtain a design criterion. In the literal meaning of the words,
factor of safety would indicate by what factor the design is safe, but as
actually used this is not true. Since its meaning does not accord with the
true meaning of the words, the number would better be called the design
< (F)(2zr)(n) Frn ir Tn factor.* For the time being we shall use the design factor N or factor of
(1.15) Safety to define a design stress s4; thus, for the ultimate-stress s, and yield-
wie (12)(33,000) 63,000 63,000’
Stress s, criteria, we have
where 63,000 is approximately equal to (12)(33,000)/(27). Referring t
Fig. 1.13, we see that Fr is the turning moment or torque Tin-lb. f s s
Another useful form of the horsepower equation is obtained by lettin (h) Sq = a and Sqg= re
Um = (2mr/12)n fpm in equation (1.15), where v,, fpm (or v, fps) is the spee
with which F “‘moves through a distance.’”’ Then Inasmuch as the stress used in design is the significant number and because
for a particular design procedure, the best design stress is a particular
Fun, Fv,
(1.15A) hp number, the values of N in the foregoing equations must be different. Thus
~ 33,000 550° the design factor depends upon the criterion used for design. Suppose
If we use Tft-lb., instead of Tin-lb., hp = Tn/33,000. For T in-kips Su = 80 ksi, s, = 50 ksi, and that a good sy = 20 ksi. Then
hp = Tn/63; etc.
80
N, = 20 = 4, the design factor based on ultimate strength and

1.15 STRENGTH OF MATERIALS. The strength of a material

is *
We shall tend to favor design factor and design stress, but the most common usage is
its capacity to resist the action of applied forces. Unfortunately, the Factor of safety and working stress.
 Se oe we eeeaE os tnowewe ef
§1.17] STRENGTH OF MAIENIALS ANY IG wesw vines -

TABLE 1.1, FACTORS OF SAFETY (DESIGN FACTORS)

design factor. An unnecessarily large degree of safety means unnecessarily
The factors of safety marked with * are primarily for beginners’ use, although they are as
high cost. A stress that is computed from a stress equation, such
traditional values. They should not be used when a detailed accounting ‘is made of the
variable loading, stress concentrations, etc., Chapter 4. Acceptable for use with typical s = F/A or s = Me/I, is best called a computed or nominal stress.
strengths.

wrath ‘eras heoeiegs 1.17 VARIABILITY OF STRENGTH OF MATERIALS AND THE

Ducts warrain METALS DESIGN STRESS. Whatever design criterion is used, we must accept
KIND OF LOAD the fact: thatx it is: not a simple precise number. Although ma the properties of
Based on | Based on Based on Ultimate materials given in Tables AT 3-AT 11 look uncompromisingly fixed, they
Ultimate Yield Strength are either simply typical values or minimum values such as would be
Strength | Strength found in specifications. For example, Fig. 1.14 shows the variation in
Dead load. N = 3-4 1.5-2 5-6 7 ultimate strength of 1011 test coupons taken from hot-rolled, SAE-1020
Repeated, ole direction, gradual i 7 structural shapes. The variability of test specimens from one heat would
(mild shock),* N = 6 3 7-8 10 { be less than the variability of specimens taken from several heats and
Repeated, reversed, gradual several furnaces.
(mild shock),* N = 8 4 10-12 15
Shock,* N = 10-15 5-7 15-20 20
FIGURE 1.14 Strength of SAE 1020 Structural Steel. The test coupons, taken from the
finished rolled product, were 14 in. wide, 9 in. long, and varied in thickness from } to $ in.
50 The diagram of rectangles is called a histogram. The height of each rectangle represents
N, = 30 = 2.5, the design factor based on yield strength the number of pieces (or percentage of the total) testing within the particular range; for
example, 108 pieces (10.7%) showed an ultimate strength of between 61 and 62 ksi, step C.
The dotted curve, called a normal curve, is an idealized distribution of the ultimate strengths
If you are finding or stating a factor of safety NV, state also its basis; as, as obtained from these data. It is the curve that would be obtained from an infinite number
“factor of safety based on yield strength” or “based on ultimate strength.” of tests plotted on infinitesimal steps, the samples having been produced in a constant
environment (if some factor varies, it varies the same way ‘‘constantly’’). As the statistician
Perhaps the more basic definition of factor of safety is puts it, all tested samples are taken from the same universe. The significance of the 3c
(three-sigma) limits, A and B, of the normal curve is that it is very unlikely that values will
loading that would cause failure fall outside of these limits, which are 55.8 ksi and 71.5 ksi, which are computed by statistical
(1.16) Factor of safety = : > methods. Some of the product falls above the higher 3¢ limit at B, which suggests that some-
actual loading on part thing happened during the manufacture that tended to increase the strength of the material.
(Maybe the operators were ‘‘making sure’’ that the material would meet a minimum ultimate
at least if a single load is involved. Also this definition is used when the strength test?) The variability of the strength of heat-treated parts could be kept much less
stress does not vary linearly with force, as in some column formulas than in this illustration by adjusting the tempering temperature from heat to heat.
(Chapter 7).
3 > 8e
Table 1.1 gives some rule-of-thumb values for your guidance. In recent ci < 20
years there has been a trend toward using the yield strength as the pre- _ Average o-

Frequency in Percentage of Total No. of Pieces

ferred criterion for getting a design stress—very desirable for dead loads. - — Sig
This practice is based on the logic that failure occurs when a body ceases 157-150 A
to perform its allotted function and that most machine elements will not
perform properly after they have received a permanent deformation. Step of y
Fe g 61-62 ksi>| fe
However, the practice is questionable in machine design where loads vary,
because the endurance strength of steels is nearly proportional to the

l
o
_
ultimate strength but not to yield strength. If the loads vary in a definable
manner, the methods of design developed in Chapter 4 should be used.
Otherwise, for variable loads, it is better to use sz = s,/N with a suitable

N B Not
): |Likely
12
Tensile Strength, 1000 psi
22 STRESS ANALYSIS—SIMPLE STRESSES [Ch. g1.18] MORE ON FACTOR OF SAFETY AND DESIGN STRESS 23

Knowledge of the variability of strength affects the designer’s choice therefore worth some study, especially by way of comparisons. However, since
of a design factor. If the probable minimum strength is used to determine the sources of these data are varied and numerous, caution should be
the design stress, a lower design factor (factor of safety) may logically b exercised in drawing firm conclusions. The materials chosen for presenta-
used than if the average or mean strength were used—that is, to obtain tion are not expected to be either better or worse on average than many
the same design stress. For a factor of safety of 5, the design stress fron other similar ones not mentioned here. Do not specify materials by. the
the approximate average stress from Fig. 1.14 would be 63.7/5 = 12.74 ksi specification numbers given without referring to the source. Frequently
This design stress corresponds to a factor of safety of 55.8/12.74 = 4.48 there are classifications within an ASTM specification number, for instance.
compared with the probable minimum strength of 55.8 ksi from Fig. 1.14 For additional information about a particular material, refer to specialized
Compared with the probable maximum strength, the factor of safety i books and standard specifications. See Refs. (2.1—2.21) at the back of the
71.8/12.74 = 5.76. The difference between the maximum and minimun book. ;
values is 28.6% of 4.48, the minimum value. ;
The variability of heat-treated steel in normally controlled processe
should be materially less than in the foregoing example. Strength value 1.18 MORE ON FACTOR OF SAFETY AND DESIGN STRESS.
for another group of specimens from various heats but with each specime In a general way, the greater the uncertainties, the greater the design factor,
given the same heat treatment varied 13% from the minimum strength and the lower the design stress, should be. For this reason, factor of safety
compared to 28.6% in the previous example. The variability might hay has been.called “factor of ignorance.’’ When the loads and their manner
been less than 13 % if the tempering temperature (heat treatment) had of variation are known accurately, when the properties of materials are
varied from heat to heat in accordance with the particular characteristi¢ carefully controlled within known narrow limits, when the maximum stresses
of a heat. ! can be computed with confidence, and when the proper criterion (yield
Usually, the design factor is such as to allow for this variability — strength, endurance strength, etc.) is used, the design factor of safety may
average or typical strength values are used. When a material specificatio be relatively low (perhaps as low as. 1.2). Uncertainties and other factors
includes the minimum tensile strength, it is likely that the manufacture that affect the magnitude of the design stress are as follows.
will control his processes so that there is little probability of a low (a) Material. See § 1.17. As regards strength, it would be best to
tensile strength. He would hesitate to guarantee a higher strength than tha choose design stresses with knowledge of the variability as depicted in
at A for the case shown in Fig. 1.14. This situation accounts for typiet Fig. 1.14, but such complete information is not generally available. Ignor-
and average mechanical properties being better than the specificat ance of the limits of the properties leads to conservative design stresses.
The properties given in Tables AT 3-AT 11 are not minimum properti The possibility of internal flaws, more likely in improperly controlled cast-
unless a note so states. Where the extra expense may be justified, the pul ing procedures, suggests higher N’s. It has been traditional to use a higher
chaser may obtain closer limits than are customary in standard specif design factor for cast metal (especially cast iron, which is more brittle)
cations. y than for wrought metals, but if the processes are well controlled and stress
All mechanical properties of a particular steel vary in a pattern similé concentrations are accounted for (Chapter 4), even for dead loads, there
needs be little difference. However, brittle materials are definitely less
Suited for impulsive loading than are ductile materials. If the material is
for this variability, he may somewhat conservatively estimate the minimul oe on specifications and suitably inspected, its properties are reliably
strengths (s,,, 5,, S,) of metals that are supposedly alike but are frot nown.
untested universes, or from different universes as in Fig. 1.14, as about 107 n (b) Size Effect. Tables quoting strength values are generally based on
less than “‘typical’’ or average values, and the minimum strength of care Standard” sized specimens, commonly about }-in. diameter (sometimes
fully controlled processes, as heat-treating processes, as about 5% les 1 in.) for the ultimate tensile stress and about j-in. diameter for fatigue
than average. Reference (2.1) shows the distribution of some mechanica Strength. However, it is well known that failure of large parts occurs at
properties for a few metals. The composition also varies in a similar statistt lower computed stresses (F/A, Mc/I, Tc/J) than of small parts. See Tables
cal manner: the percentages of carbon, sulfur, silicon, and of other alloyin AT 8 and AT 9. For dead loading, the loss is not significant up to say a
elements. Moreover, the same generalities concerning variability apply 2-in, dimension (as the thickness of a flange—the flange width would not
every material. © involved), but a 10-in. axle would require a higher apparent factor of
The materials whose properties are given in the Tables AT 3-AT 1 Safety, or else experimental information must be obtained, which is expen-
and the charts of Figs. AF 1, AF 2, and AF 3 in the Appendix have been Sive for large members. In Chapter 4 we begin
making allowance for size
carefully selected to present an informative variety, and the tables are effect above }-in. size.
24 STRESS ANALYSIS—SIMPLE STRESSES [Ch.1 §1.19] EXAMPLE—DESIGN FOR TORSION 25

(c) Loading. Deciding upon the service loads for a machine is fre- than that for the shaft because its failure might then save the failure of
quently practically impossible. (Think of an automobile going over rough more expensive parts of the machine.
roads.) But some sort of estimates are essential, and in some cases, the (i) The Price Class in Which the Machine Is to Sell. Sometimes,
results are reasonably close to the truth, and can be stated as falling cheaper machines have a lower factor of safety in order to reduce the cost
between certain limits with a high probability of being right. For many of materials and manufacture.
elements where experience is available, design is based solely on experience, In the final analysis, the choice of the design factor N rests with the
When the nature of the loading is known in some detail and is accounted judgment of the designer, which in turn depends upon his experience. In
for in the design, a lower design factor is appropriate as compared with a many instances, design is necessarily an experimental procedure because so
case (such as those in Table 1.1) that covers a considerable ignorance. For little is known about the actual maximum loads. A part of a certain size
static loads and ductile materials, the yield strength is the best criterion if and material is tried. If it does not fail, it may be replaced with a smaller
a significant permanent deformation would destroy the part’s usefulness one or one made of less expensive material. If it does fail, a larger part, or
(ultimate strength if only rupture destroys usefulness). For varying loading, a stronger material, or a change in shape may be tried. Ultimately, the
the endurance strength is the best criterion. 7 factor of safety to use in this particular instance may be determined from
(d) Computed Stress. The computed average stress for a membef experience. As you proceed with this study, observe how often the method
subjected to simple tension can be viewed with considerable confidence. How: of determining the design stress is suggested for particular machine
ever, the uncertainty increases substantially for parts of complicated shapes elements; then remember that these design stresses are based on experience
In some cases of nonuniform stresses, good theoretical equations, perhap and are subject to such revision as experience may justify.*
with practical modifying constants, may be available; in other situations
suitable theoretical equations may be all but impossible to obtain. Experi ————————@

mental stress determinations on a finished part give information that doe

not exist when design begins and will show whether or not the design&
acceptable (see Fig. 4.12). Another frequent complication is the residue FIGURE 1.15 Torque on Circular Member. re
a
stresses left in the unloaded part by some manufacturing process (say, hee
treatment), which may or may not be beneficial under operating conditions
The pattern of residual stresses may not be permanent and properties may
change in service. {
(e) Environment. Some working environments introduce a com
siderable uncertainty. For example, salt water, a corrosive atmosphere 1.19 EXAMPLE—DESIGN FOR TORSION. Let the shaft in Fig. 1.15 be
etc., may result not only in a cracking of stressed material, but also in a subjected to a twisting moment by loads W = 100 lb. at a = 20 in. on each side.
actual disappearance of material, and the roughened surface is detrimenta The material is C1030, as rolled, and a design factor of 4 based on yield strength
should be sufficient (to cover the loss of strength from cutting the keyway, too).
to fatigue life. If nuclear bombardment is involved, the properties”
» Determine the shaft diameter D. (b) If the torsional deflection is to be limited
materials are changed—for some materials, in an unknown manner. A
oO 0.2 deg./ft. of length, will the diameter found be satisfactory? (c) What is the
very low temperatures (§ 2.22) and at very high temperatures (§ 2.21), meté torsional stress at a point 3 in. from the axis of the shaft found in (a)?
properties change significantly. ie Solution. (a) First find the design stress. From Table AT 8 for 1030 as
(f) Inspection. The thoroughness of inspection and the strictness ¢ * ed, we find s, 5 51 ksi (tension). Table AT7 suggests using sy, = 0.6s, =
the specifications are factors in the decision of the magnitude of the desigi -6)(51) = 30.6 ksi. For a design factor of 4, the design stress is
factor. If a rational statistical approach that insures accurate knowledgi
30.6
of the final product is used, some uncertainty is removed and a lowe Ss; = rage 7.65 ksi.
design factor may be used.
af applied moment is T = W(2a) = (100)(2)(20) = 4000 in-lb. Using equa-
T ° .

(g) The Chance of an Accidental Load. A part should be ampl On (e) for the maximum stress condition, we find
strong to withstand an accidental blow that, for example, may occur i
moving the machine about. The machine should be able to take withot
DP a 16T — ee
(16)(4000)
serious damage some overload arising from unexpected causes. Som 76507 ratte
(h) The Danger to Life or Property. Higher factors of safety af Use 13 in.
desirable if life or valuable property would be endangered in case of failure *
One might saa y facetiously
i i
that the best design stress isi the one that results i n the right
i
The factor of safety of a key that keys a pulley to a shaft should be les answer—the right answer as conclusively proved by experience. si
 26 STRESS ANALYSIS—SIMPLE STRESSES [Ch. | |

Observe in Table AT 8 that there is a minor decrease in the yield strength of a

section of this size as compared to that of a 4-in. specimen. {
(b) ForJ = 7D*/32 = n(1.375)*/32 = 0.358 in. andL = 1 ft. = 12 in., from
equation (1.13),
TL (4000)(57.3)(12)
9 = 76 = G5 x 1050358) ~ ORF FIGURE 1.17 Rod and Yoke Con-
which compared with the permissible value of 0.2° shows that the shaft must be” nection,
made larger if stiffness is the significant criterion. [Solve for D from (1.13) with
6 = 0.2/57.3.]
(c) For the shaft in (a) with c = 0.5 in.,
Te ee (4000)(0.5)
ss = = 5600 psi.
J 0.358 so nl

1.20 SAFE COMPRESSIVE STRESS. For highly ductile materials,

stress, and in any case, apply it only to ductile materials. Moreover, the
a definite ultimate compressive stress, that is, a stress above which the body
resulting design stress in bearing should generally be safely below the yield
breaks, is impossible to determine because such materials simply flatten
strength.
out under a compressive load without showing a fracture. For the purpose
Recall that projected area 1D is used for computing the surface com-
of deciding upon design stresses, the ultimate compressive strength of t
pressive stress in Fig. 1.16, the nominal value being s, = F/A = F/(tD) psi.
type of material is taken equal to the ultimate tensile strength. When tested
in compression, ductile materials usually exhibit approximately the sam
characteristics up to the yield strength as they do when tested in tension. ~ 1.21 EXAMPLE—STRESS ANALYSIS. The beginner’s first difficulty is
An illustration of a situation where experience suggests an upward in sorting out various areas subjected to different kinds of stresses; that is, recog-
modification of the design compressive stress is when the maximum st nizing bending, shear, tensile, or compressive stresses and the areas on which
exists on the surface and decreases inward with distance from the surfa they act when they exist in a single unit to be designed. The thought process
involved in this act is called stress analysis, which is best illustrated by an
example.
Design a yoke connection, similar to the one shown in Fig. 1.17, to withstand
mM’
B (| a load of F = 3000 lb. repeated in one direction, if the material is AISI C1022,

et |Z as rolled.
st

Solution. Let the first step in a design problem be to decide upon the
F design stresses. Since the load is repeated, choose a design factor for use with
FIGURE 1.16 Compressive Stress on Cylin ultimate stresses. From Table 1.1 for ductile material, use N = 6. The various
drical Surface. Loaded as shown, the rivet| ultimate stresses are found in Table AT 7 for the 1022, as rolled (look them up
also subjected to shear and bending. ) now). The corresponding design stresses are:
yp. ; 54 s
Ss= 2 = 12 ksi, Ses = 12 ksi, sae = Oksi.
In Fig. 1.16 the fibers on the surface of the hole and those on the surfacé
of the rivet are pressed together but the compressive stress due to F doe The most obvious stress is tension across the circular section A—A, Fig. 1.17.
not pervade the whole rivet nor extend far into the plates. If the compres: From F = sA, we get
sive stress exists mainly on the surface of a body, a higher design str
may be used than when the compressive stress pervades the entire bo 3000 = (12,000)(72")
as in Fig. 1.1(b). This surface type of compressive stress is sometimes
a bearing stress. The designer must decide how much higher design
call
stress
a = [:3000Ka7"" = 0.565 in.; standard 3% in.
12,0007
is permissible. As one example, the ASME Code for riveted joints permits
A “standard” fraction will be chosen for each dimension (see § 1.22).
the design surface compressive stress to be about 60% higher than the
Determining the dimensions a, b, c, e, and m is not so simple. The best thing:
design tensile stress (where N = 5 ons,). If you should use this idea, to do in most cases where several design equations are involved is first to set up

27
 -v FIGURE 1.18 Shear of Pin. FIGURE 1.20 Compression Between Pin and Yoke, and

YA
Tension Across the Hole.

JM tte = te th
Between FIGURE 1.19 Compression Between Pin and Rod, i F F Lt ey
and Rod and Tension Across the Hole in Rod.
rf Be e re pees ; Fo uF Fo ere
a |v ar aa t pa? LF?
Tension
across Hole
(a) (b) (c) (d) (e)
all the strength equations representing the various possible methods of fail re
so that they may be studied. We shall follow this plan, being sure that the areas Figure 1.21 Pin asa Beam. (a) Loose fits in yoke arms: assuming that the pin bends enough
are given in terms of the dimensions shown in Fig. 1.17. . that it rests at MM on the inside edge of the holes in the yoke; uniform load; max. M = Fc/8.
The pin may fail by shearing across the section B, Fig. 1.18. One sectiona (b) Loose fits in yoke, but support assumed at midpoint of 5; uniform load; max. M =
F(6+ c)/8, which is greater than in (a) and therefore more conservative. (c) Tight fits in all
area of the pin is 7a?/4, and the total resisting area is twice this. Thus
parts: almost a fixed end beam with uniform load, M = Fc/12; or a uniformly loaded
Ss7a? cantilever beam, M = Fb/2 (which is larger than Fc/12 for normal proportions). (d) Loads
(i) F=sA = 82) =
2 not uniformly distributed, M = Fc/4. (e) Loads not uniformly distributed, 4 = F(6+c)/4.
The maximum shearing stress computed as a vertical shear in a beam [§ 1.10(6)
will be 33% larger than the average stress as given by (i). Therefore, if shear of toward each other approaching Fig. 1.21(d). Moreover, the situation in Fig.
the pin governs its size, this difference should be considered. The compressive 1.21(e) results in the largest maximum bending moment and is therefore the most
Stress S- between the pin and the rod may be excessive. For a projected area ac, conservative shown and therefore it would perhaps be a normal choice in the
Fig. 1.19, we get absence of experimental verification, Thus for the pin in bending,

(i) F =sA =s,(ac). _, Peiaxey gs 2 nas o F= Sera

(m) M
The compressive stress between the pin and the yoke may be excessive. Fo! 4 £92) ~ 86 +0)
one side of the yoke, the projected area is ba (see Fig. 1.20) and for two sides, One other type of failure that might be checked would be the pin tearing out
2ba. This gives the end of the rod or yoke as suggested by Fig. 1.22. There is a shear on length e
and a depth c on both sides of the pin;
(k) F =sA =s,(2ba).
The rod or yoke may fail in tension across the hole for the pin, the section 0} (n) F =sA = 5,(2ce) = s,(2)(2be).
[Rop] [voxKe]
minimum area. See Figs. 1.19 and 1.20. We have
By a comparison of equations (j) and (k), we see that if 2b = c, the yoke will
(dD F=sA =s(m — a)c = sm — a)2b. : safe if the rod is safe; hence, let c = 2 in all calculations. There are still several
[Rop] [YOKE]
cisions to be made. One could assume a ratio of say a/c, solve the various
The pin is not only subjected to shear across the sections B, Fig. 1.18, b t
also to bending. There are several assumptions that might be made, some
which are shown in Fig. 1.21. This decision as to what idealization to use for FIGURE 1.22 Pin Shearing Out End of the Rod.
design purposes is an elementary illustration of the kind that the designer 18
continually making. The caption to Fig. 1.21 suggests some possible logic.
best answer would be found by comparing computed results according to one
or more ideal models with experimental measurements. On this basis, for workin
fits (Chapter 3), the assumption in Fig. 1.21(e) agrees best with actual measure-
ments,!!-8] except for relatively high loading when the points of support move

vii) # 28
30 STRESS ANALYSIS—SIMPLE STRESSES [Ch. 1
R-0
equations and use the largest values of a and c obtained. We might think in terms. F,=3000
Ib.
of a more optimum design and get proportions that would make the pin equally
strong in compression and bending or equally strong in shear and bending. Try
equal compression and bending by equating the F’s from equations (j) and (m),
s;7a3 :
(o) F = s.ac = ———— or a = 1.955ce FIGURE 1.17 (Repeated).
binary Te 4
for b +c = 1.5c, 5s. = sy. (NOTE. For a fixed connection with no relative
motion, one could safely assume the bearing stress s, ~ 1.6sy in accordance with
§ 1.20, but the resulting proportions would be even worse; of course, one does,
not know this at the start.) Solving for c from (j), we have
3 1/2
(p) F = s¢1.955c? or 0.358 in.
ig aes i
This dimension is less than d = 4, and while it would give optimum proportions
from the standpoint of utilizing the capabilities of the material, most such
point of view, let it be e = m/2 = 2 in. Summarizing, we have the dimensions:
connections have c = d. (See the design for relative motion below.) Assume then
that c = §in., which means that the compressive stress will be lower than d= fin, a=fin, c= #in., 6 = #in., e = fin, m = it in.
design value. For b + c = 1.5c = #in., equation (m) gives
The next thing to do would be to sketch this design to scale, to check the
@ = [SELOE) _ [OOSTIOT* ogy, Use dil over-all appearance and proportions. If changes are made that cause an increase
in stress, check the stress in the affected parts. Provide generous fillets and curves
Check the pin in shear, equation (i); at adjoining sections, as at B, Fig. 1.17, in order to keep the stress concentrations
at a minimum (Chapter 4). Specify all other dimensions; help on tolerances,
(r) ees ()" 1 is jee" = 0.461 in., allowances, and surface finish, is given in Chapter 3.
Solution for Relative Motion. If the ends are round with a diameter m,
Fig. 1.17, significant relative motion is possible. If relative motion occurs, the
shear. (IMPORTANT NOTE. If the load were a dead load, the shear computation co’ I permissible compressive stress must be very much smaller than before in order
be allowed to govern, because a consideration of actual rupture suggests that to avoid excessive wear on the surfaces. If the motion is small, meaning that the
failure would not occur until the pin actually sheared. However, for a repeatet heating from friction introduces no difficulties, the allowable stress may be, say,
load, failure may well occur by fatigue in bending (Chapter 4) and therefore 1000 to 4000 psi or more. Experience in a particular machine is, needed to define
cannot be ignored in this problem where the load is repeated.) the proper value. The wear rate, which also is a function of a, surface hardness
The dimension m can now be computed from (1); that has not been specified, increases with the pressure, and the life expectancy
desired would govern. We shall let the design value of s, = 4000 psi= 4 ksi and
Bcc i a gas Bc AHL + 0.875 = 1.275.
(s) m=
sc + * ~ (120.625) Present the following solution after the above pattern but without the discussion.
From equation (0),
This designer is worried about the small margin around the hole*; use m = 14 in
The computed value of e from (n) is, for c = §, 127a8
spac = 4ac = or a = 1.13¢
(t) e= ae = (290.625) = 0.266 in., 12¢

3 /
which, being less than the radius of the hole, is unreasonably small; that is, thi F= SoAc or c Pee _
=S | = 0.814;ie Use 3# in.
i

is an instance where the area needed is too small to be physically possible

Besides the assumption that such shear starts at the center line of the hole is ove!
optimistic. The dimension m and the margin distance e are roughly analogo 8(6 + c)F}}/8 1. } a3
= [ ( s | = [x : <° B129K3) =0.918; UseZin.
to a hub diameter (for gear or pulley), which is generally empirically chosen (for wT ’

steel) from 1.25D to 1.8D, where D is the hole diameter. Considering e from this

* The stress concentration effect at the hole may be surprisingly large; see Fig. AF 6
yy eae
SC
ee DA
(12)(0.8125) ;
ee :
ae ,
and ref. (4.62).
As before, use m = 1} in.; the rounded end takes care of e. There is no need to

31
32 STRESS ANALYSIS—SIMPLE STRESSES [Ch.1
51.23] COURTESY IN THE WRITTEN FORM OF CALCULATIONS 33

repeat the shear-of-pin calculation. Other suggestions are the same as before, concerned. Neatness and completeness of calculations are necessities, either
Summarizing, ;
in the school room or in the engineering office. In both cases, the design
d =3,in, a=Htin, c= 2in., b= ,in., m= Iltin. calculations are likely to be checked by another person. The time taken
by the original designer to make his work legible and complete in every
The maximum deflection of the second pin for the loading assumed i
detail will consistently be less than the time spent by a checker (or the boss!)
Fig. 1.21(e), by formula taken from Table AT 2, is
in trying to comprehend the meaning of semi-illegible figures. However, it
RES (3)(1.25)°(64) = 1.42 x 10-*in., is not only a matter of the economical use of time and of having your work
» = 48EI (48) x 10%)(n)(0.875)* so that you can quickly get the most out of it in reviewing for quizzes, it is
where L=b+c= % + # = 1.25, I = na*/64, E =3 x 10*ksi. The nex also a question of common courtesy, a concern for others. The following
question is of course what deflection is permissible. Such values for shafting an suggestions for the form of calculations should be followed except where
typically expressed in terms of deflection per inch of length between supportin; they may conflict with special directions of your instructor or of your
points, y/(b + c) in this problem. A value of 0.002 in./ft. or 0.002/12 = 0.00016 supervisor.
in./in. is relatively strict. This compares to 0.000142/1.219 = 0.0001165 in./in. fe (a) Name the calculation with a phrase or sentence, for example,
our pin. Therefore, even without prior experience in this particular case, it is Diameter of Pin for Shear.
fairly safe to assume that the deflection is not excessive.
(b) Give the equation to be used in terms of defined symbols:
By way of résumé, the general plan of attack is to write down all the strengtl
equations applying to various areas of the part in terms of letters for dimension F = s,7a?/2. Where appropriate, the symbols should be defined by
and then to study the equations for a logical method of solution, not overloo a sketch. Be liberal with sketches. It is wise to identify an equation
a chance for optimum proportions (the ideal design, like the old one-horse shay taken from a book by page number and equation number (if any).
is one in which every area fails simultaneously, but not until after the ma hin If appropriate, state the assumptions on which the equation is
has served its purpose). Compute the dimensions and study them for the based (as the pin in bending above).
practicality. Adjust as deemed desirable and check questionable stress area (c) Substitute numerals for known symbols in the same order in
The procedure from here depends on the end use and may involve building th which they appear in the equation. Be sure the units are consistent.
item and testing it, perhaps more than once. Highlight the design stresses, making it clear how they are obtained.
(d) Mathematically simplify the equation by solving for the
unknown. Many equations are so simple in form that this step is
1.22 PREFERRED SIZES (STANDARD FRACTIONS). For unnecessary. A good rule for school purposes is to have in the final
machine elements, there are standardized sizes, as bolts, keys, I-be calculations all steps in the solution except those that you can easily
which means that such sizes are more readily available on the marke make mentally.
are also cheaper. The designer always uses standard items and stand: (e) Write down slide-rule results, with units. Use separate scratch
proportions unless he feels strongly that some custom design is desirabl sheet for cancellations and juggling for slide-rule purposes.
We shall give information on standards as we proceed. If there are (f) Highlight your answer or result—by heavy underlining,
such standard sizes, let the preferred dimensions'*-1*) vary by placing it in a marginal column, etc.
Se between 4-—ds; gs between 34-353 (g) If you are computing a dimension, do not leave it in decimal
as between 33-4; $ between 3-3; form; select a standard dimension or part.
4 between 3-6; 3 above 6. (h) give pertinent conclusions, if any, derived from your compu-
tations.
In choosing a standard size from a decimal number, avoid a smaller siz (i) Reflect on your calculations and try to be sure that you have
if the change is more than some 4-5 %; take the next larger. Also observ thought of contingencies and alternatives. Compared to the other
that now and then a larger dimension results in a weakening, for exampkt parts of the problem, does the answer look reasonable? Have you
a longer beam. located points of maximum stress? Have you allowed for the fact
that a computed answer may be a bad answer, practically? Put on
record in your calculations significant reasons for designing the
1.23 COURTESY IN THE WRITTEN FORM OF CALCULATIONS way you did.
Before going too far, we should suggest that you present your problem (j) In general, avoid overcrowding and the use of small script
work in a form that will be pleasing and quickly understood by thos and letters. Work toward the desired result by using a series of
34 “ ‘ STRESS ANALYSIS—SIMPLE STRESSES [Ch.
FIGURE 1.25 Stress in Diametral Plane.
fundamental equations rather than by substituting into a more com
plicated derived equation, except for repeated calculations. j
(k) Separate the various computations clearly by drawing M
straight line after each entirely across the page. 9
as they usually are, the weakness of the joint as compared with the parent
plate is cared for by a joint efficiency 7, defined as
Minimum strength of joint
FIGURE 1.23 (v) 7= Strength of solid plate :
which would preferably be called the relative strength. Introducing y, we
L
find the required thickness of shell from

1.24 BUCKLING OF BEAM FLANGE. Some structural shapt

I-beams, etc., are designated by the nominal depth and the weight per for (1.18)
For example, 14WF34 is a wide-flange beam with a depth d = 14 in. a
it weighs 34 lb. per foot of length. If the unbraced span L, Fig. 1.23, is t The stress in a diametral section is different. The force is pxD,;?/4 and
great, the flange that is in compression may buckle, even with an otherw the area, Fig. 1.25, is closely 7Djt; substituting into F = sA, we have
safe stress level. This contingency is covered by building codes for strt 2
tures; for example, the AISC gives the permissible flange compressive stre (W) Eee: sy D,t or y= at
(s, = Me/I) as?-4} 4t

22.5 Le which, compared with equation (u), shows that the nominal stress on a
(1.17) FS [is <> a transverse section is half that on a longitudinal section. These two stresses
* 1 + (L/b)?/1800 are principal stresses (Chapter 8); the one on the longitudinal section is
where, Fig. 1.23, b = flange width and the maximum allowed L/b =4 the maximum principal stress and is the basis of design. In most if not all
For L/b < 15, the permissible s, = 20 ksi. 5 states, it is required that pressure vessels for land installations be designed
to meet code specifications,"-®-1-1°-1:1] where the factor of safety on speci-
fied material strengths is taken as 5. For thick-wall vessels, see § 8.26.
1.25 THIN-WALL PRESSURE VESSEL. A thin-wall p
vessel is one whose plate thickness is small compared to the diame’
the vessel. The pressure p psi gage, Fig. 1.24, acts in a radial dire 1.26 EXAMPLE—TITANIUM VESSEL. Helium is used to provide pres-
Summing the vertical components fp dA sin 6, we find the total Sure on the fuel and lox (liquid oxygen) in rocket motors. Let helium be stored
tending to rupture the vessel on a diametral plane, say section AA, in a cylindrical vessel, 20 x 24 in. (the first number is always the diameter), at
4000 psia. The welded joints are assumed to have an efficiency of 85% (Chapter
p(DiL)—the pressure times the projected area. Assuming uniform
19) and for the purpose of saving weight, the vessel is to be made of annealed
across thickness t and A = 2tL, we have titanium alloy, Ti6 Al4 V. For normal temperature, what thickness of plate is
D needed with a design factor of 1.4 on the yield strength?
(u) D,L = s(2tL)
pV; s(2tL or = ass
2s, and 3 = 2 ‘ Solution. In Table AT 3, we find s, = 130 ksi. (At low temperature, the
yield strength is higher, § 2.22). From equation (1.18),

_ Pdi _ (4)(20)
= 25m ~ ((130/1.4)(0.85) > 201s
FIGURE 1.24 Separation of Shell along ne
Section. use 2 in. Nore. The low factor of safety proves satisfactory here because the
ou are proof tested and closely inspected besides. If failure involves human
bea ‘. e factor of safety may be raised to about 2. It is assumed that support
see ets and fittings do not result in a significant increase in stress. Weight is
© important a factor here for designers to be free and easy about the degree of
Safety and the codes do not apply. See § 2.19.

35
36 STRESS ANALYSIS—SIMPLE STRESSES [Ch g1.28] STATICALLY INDETERMINATE PROBLEMS 37

1.27 CONTACT STRESSES. Stress equations that are usual square-root operation. The maximum shearing stress is a consequence of
omitted in the undergraduate course on strength of materials because the three principal stresses, all compressive, and is
the complications and because of the time needed for solutions are the
(1.20) Ssmax >= 0.35, max atz= 0.393w in.
giving the so-called contact stresses. The most frequently met cases
engineering are the equivalent of two cylinders in contact along an
below the surface, where
ment, Fig. 1.26, and a sphere in a groove (ball bearing). The case of|
cylinders is relatively uncomplicated, it will serve our purposes, and 4s,max (1 — uy? Lin bs)
w= + -
can be applied to gear teeth, cam and follower, roller bearings, chain poy (y)
drives, and to other elements. Theoretically, two perfect cylinders w
Ley & Alta Was ks E,
The maximum shear stress is often taken as the significant stress, because
it is thought that this stress is effective in promoting the flaking of metal
particles from the surface (the visible evidence of surface fatigue).
A sphere in contact with another surface has theoretical point contact
and therefore, for a particular force, the maximum stress is much larger
than for cylinders. While the equations for two spheres in contact are easy
to use, this case is not too often found in engineering. For detail pertaining
FIGURE 1.26 Cylinder to various shapes of surfaces in contact, see Seely,"-") where there are
Contacts. The contact are: charts to expedite solutions.
is somewhat exaggera’ od. When there is relative motion, especially sliding, the stress situation
Area After Loading 5,9 is more complicated,"-” and if the coefficient of friction is large (large
ahs y
frictional force), the maximum stresses are significantly greater than given
--+-w
by the foregoing equations. In any case, there seems to be a good correla-
Tro Stress Distribution tion between the wear life of a pair of lubricated surfaces (rolling and/or
@) sliding on one another) and the maximum stress. When the stress is above
the endurance-limit stress of the surface, an increase of stress greatly
shortens the life. Buckingham,"*-*! experimenting with two case-hardened
in Fig. 1.26(d). The original derivations were by H. Hertz (the correspo rollers, found a life of 10° cycles for a load of 12.89 kips (computed
ing stress is therefore often called a Hertz stress) on the assumption 1 Se max
= 362 ksi) and a life of 10® cycles for a load of 8.27 kips (computed
5c max =295 ksi), the life changing by a factor of 100 while the load
is elliptical, Fig. 1.26(b) and (d). The equation is decreased about one-third. However, it is an observed fact that the com-
‘a - puted contact stress may be allowed to be very much higher than that due
F(1/ry + 1/r2) to tension, bending, etc., if a reasonable finite life is expected.
(1.19) 5,2 OM
Wid — pI +o a/Es PD oF
where s, is the maximum principal stress; r, in. = radius of smaller cyli n
rg = radius of larger cylinder [r2 is negative when it wraps about sma
1.28 STATICALLY INDETERMINATE PROBLEMS. Problems are
as in Fig. 1.26(c)], b = length of cylinder in contact, E,, Ey (psi or ks!
Said to be statically indeterminate when the load on the various parts
match F lb. or kips) are the modulii of elasticity for the material " cannot be found by the principles of statics alone (LF = 0, 2M = 0).
14, Hg are Poisson’s ratios. The usual assumptions apply: elastic 2 or example, if a beam with parallel loads has two suppports, the reactions
homogeneous and isotropic materials. Poisson’s ratio does not vary g at the supports can be found by statics, but if there are three points of
with different metals (but see values given in Tables AT 3 and AT7 Support, an additional unknown is introduced and an additional condition
it is not always accurately known. Assuming a single value of » =% Must be found. Where the problem is solvable by rational methods, the
equation (1.19) reduces to xtra conditions usually relate to deformations (deflections). These prob-
a exist in wide variety, but generally some idealization can be made that
_ f0.35FU/ry + Ura)?
(x) Somax = ered psi or ksi. fads to a solution.
: For example, suppose a link A, Fig. 1.27, supports a load W and is in
Since the stress is compressive, it is often given the negative sign from1 Urn supported by slender members 1 and 2 as well as the pin B. If A has
 g1.30]) FOR STUDENTS ONLY 39

constant and, for practical purposes, it varies with the temperature. The
A FIGURE 1.27
7 values given in the Tables AT 3-AT 6 are for ambient temperatures, but
other values are available in the literature. If an unstressed body is heated
R
so that temperatures are different in its different parts, there will exist
thermal stress gradients (as well as temperature gradients), which, in many
relatively large stiffness against bending in the plane of the paper, ft cases, have led to service failures. If a body is restrained from deforming,
deformations of 1 and 2 can be considered to be the only deformati even though its temperature is uniform, it will be stressed proportionally
The deflection of the pins must also be relatively negligible. This simplifie; to the strain (s = Ee within the elastic limit), which is the difference
the problem considerably. between the unrestrained dimension and the same dimension under strain.
Thinking of A as a free body, we see that, knowing W and all th
Situations of thermal stresses are typically special problems whose solutions
dimensions, there are three unknowns, F,, F2, and R, but for a paralli
depend upon the circumstances, the ingenuity of designer, and sometimes
force system, there are only two independent conditions. First, tal
experimental stress analysis. A theoretical approach soon becomes involved,
moments about B; but it should be used where appropriate."!-1) In general, a designer as best
he can avoids mounting members so that temperature changes are met
(z) > Mp = Fe — F,a — Fob = 0,
with restraints that tend to prevent a change of dimensions (thus he uses
in which F, and F, are unknown. If A is rigid, the deflection 3 of the men expansion joints, sliding supports, etc.); but restraints are often lived with
bers 1 and 2 are proportional to their distances a and b from the pivot: and, frequently, it is impossible to avoid differences in temperatures in a
We can write this proportionality and then use 8 = eL, e =s/L, and F=s5 particular part.
[see equations (1.1), (1.2), and (1.3)] to find

a 81 €,L, 5,L,E, 5,L,E.b

-s— = SS; or So. =
b So €oL, EysqLl, : E,Le2a
1.30 FOR STUDENTS ONLY. Education without misrepresentation
is impossible. There are so many things that need to be said at once to
AogL, Eb the beginner that many statements must be simplifications in order to be
Fy 1 = A433
aed; Fg 2 = Aso2°2 = — E,L,a 1 1: intelligible. For pedagogical reasons, safe design procedures are often
given in this book—usually in too much detail. Although it is a conve-
These values of F, and F, into equation (z) give nience to a teacher to be confronted with some uniformity of approach for
grading purposes, there are likely to be other “correct”? points of view.
Agl, Eb Your teacher or supervisor may ask for another. After a few years of
Fe — Aysya — eee b=0. experience, your conclusions will rest more and more on your own back-
Lea
ground, but it is hoped that you will always be in search of better design
If the dimensions of the members 1 and 2 are known, the only unknow approaches. Since there is no single correct answer to a design problem
here is s,. Solve for s,; then for F,; then for F,; then another static eq unless the procedure is specified completely, including the design factor
brium equation will give R and all the forces are known. If the proble and the material, your instructor is more interested in how you attack a
a design problem, one first decides which member | or 2 has the maxi Problem and in the decisions you make than in the result. Therefore, work
stress. If this is s,, then s, can be a safe design stress. Perhaps for conveni Sor a good solution, not so much for an answer.
in manufacture, the rods 1 and 2 have the same cross section, A; = ; In any subject, there is a certain amount of new language to be learned.
= A; if so, solve for A. Let the reader assume that A, = A, = A, E& Since difficulty with a new subject is often synonymous with an ignorance
E, = E, and show that it would be more economical of material to rev of the language of the subject, pay close attention to new words, making
the positions of 1 and 2; that is, the shorter member should be closer toB 4 real effort to master them. Practice using them and make the subject
ae In studying the examples, work them out yourself, looking up all
os ues referred to. Comprehending study prior to working a problem is
1.29 THERMAL STRESSES. If a structure or machine elemé Tuly a time-saver, student practice to the contrary notwithstanding.
Those tables in the Appendix of immediate interest can be marked with a
undergoes temperature changes, it expands or contracts, the amoun
simple situations being defined by the coefficient of thermal expansio Paper clip for easy location.
the total linear deformation 5 = «L = o«(Af)L in. However, « is This text should be looked upon as a transition step toward real
40 STRESS ANALYSIS—SIMPLE STRESSES [Ch,

engineering work. The attitudes that you might begin to acquire have b 26
expressed in Report on Engineering Design,* from which the following|
quoted.
(1). Willingness to proceed in the face of incomplete and often contradict. °
data and incomplete knowledge of the problem. ;
*(2). Recognition of the necessity of developing and using enginee ri
judgment.
(3). Questioning attitude toward every piece of information, every specific
tion, every method, every result.
(4). Recognition of experiment as the ultimate arbiter.
**(5). Willingness to assume final responsibility for a useful result.’
Engineering is the art of applying the physical sciences to the soluti
of the problems of mankind. If, after completion of your study of f
book, you feel somewhat knowledgeable yet uncertain, our aim will ha
been at least partially accomplished. The art is never perfected. Moreoy
since uncertainty is the father of progress, only the ignorant can afford
be certain. As you will see, machine design is engineering. 4
2. MATERIALS AND
THEIR PROPERTIES

2.1 INTRODUCTION. This chapter will serve as a handy but abbre-

viated reference and it should be studied with this purpose in mind. We
hope that most of the topics have been studied elsewhere in more detail.
Since the number of available materials and the quantity of information on
their properties is staggering, one must search for more complete knowledge
when designing “‘for real.”’ The references given will be helpful. There are
many materials of interest to one designer or another that are not men-
tioned, for example: glass, asbestos, wood, concrete, cork, and of course a
number of “plastics.”
_ In making a choice of material, past experience is a good guide, so good
in fact that engineers too frequently have overlooked possibilities in new
materials. The best material is the one that will serve the desired purpose
at the lowest cost for the manufacture and operating maintenance of the
finished part. This best material is not always easy to find. It may involve
ttial and error. Sometimes the choice of material involves an intensive and
€xpensive search by a group of engineers and scientists. Since the material
often “makes or breaks” the machine, a sound decision, which calls for a
Wide background of knowledge, is important.

2.2 DEFINITIONS. See also § 2.3. For easy reference, we shall define
* Journal of Engineering Education, Vol. 51, No. 8, p. 650. tiefly some of the terms (not elsewhere defined in this book) with which
the reader should be familiar.{2-1-2-2.2-81

41
42 MATERIALS AND THEIR PROPERTIES [Ch.3 §2.2] DEFINITIONS 43

Age hardening (precipitation hardening) occurs in some metals, notably Embrittlement involves the loss of ductility because of a physical or
certain stainless steel, aluminum, and copper alloys, at ambient tempera: chemical change of the material.
ture after solution heat treatment, the process being one of a constituent Free carbon is that part of the carbon content of steel or iron that is
precipitating from solid solution. Where used, the consequences includ in the form of graphite or temper carbon.
increased strength and hardness, decreased ductility. Aging at moderatelj Hard drawn is a temper produced in a wire, rod, or tube by cold drawing.
elevated temperature expedites the process and is called artificial aging. See temper and §§ 2.16, 2.17.
Alloy is a substance with metallic properties, composed of two or more Homogeneous materials (have homogeneity) have the same structure
elements of which at least one is a metal. q at all points. (Steel consists of randomly oriented iron crystals of different
Alloying elements in steel are usually considered to be the metallic ele sizes, with other matter in between and is thus not homogeneous.
ments added for the purpose of modifying the properties. Isotropic materials have the same properties in all directions. (Wood has
Anisotropy is the characteristic of exhibiting different properties whe! a grain; rolled steel is not isotropic.)
tested in different directions (as tensile strength ‘“‘with the grain’”’ or “‘acros Izod test is a test in which a specimen, supported at one end as a canti-
the grain’). lever beam, is broken by the impact of a falling pendulum. The energy
Brittleness is a tendency to fracture without appreciable deformatior absorbed in breaking the specimen is a measure of the impact strength.
See ductility. , Impact values in the tables should be considered more qualitative than
Charpy test is one in which a specimen, supported at both ends as | quantitative because the actual variation of samples from the same uni-
simple beam, is broken by the impact of a falling pendulum. The energ verse is quite wide (see Fig. 1.14). See Charpy test.
absorbed in breaking the specimen is a measure of the impact strength ¢ Killed steel is steel that has been deoxidized with a strong deoxidizing
the metal. See Izod test. agent, such as silicon or aluminum, in order to eliminate a reaction between
Cold shortness is brittleness of metals at ordinary or low temperatures the carbon and oxygen during solidification. Ingots of killed steel are soun-
Cold working is the process of deforming a metal plastically at a ten der, containing fewer gas holes, and more homogeneous than non-killed
perature below the recrystallization temperature and at a rate to prod or rimmed steel; these are desirable characteristics for forgings and heavy
strain hardening. Cold-drawn steel is frequently used because it incre: rolled sections.
strength and machinability, and improves surface finish. It reduces ductilit} Machinability is a somewhat indefinite property that refers to the rela-
Commercial amounts of cold working of steel are of the order of 10-20%. tive ease with which a material can be cut. In the case of steels, cold-drawn
Damping Capacity is the ability of a material to absorb or damp vibr AISI B1112 being cut with a high-speed tool-steel tool and with a proper
tions, which is a process of absorbing kinetic energy of vibration owin cutting oil is usually taken as 100%. Free-cutting brass is a reference for
to hysteresis. The absorbed energy is eventually dissipated to the surroum copper alloys. Such data as in Table AT 7 are roughly relative at best,
ings as heat. At a particular stress level, cast iron is a much better dampii since the actual conditions of operation vary widely. There are significant
material than steel. A production variables, such as the sharpness and shape of the cutting tool,
Decarburization is a loss of carbon from the surface of steel, occurrif exact nature of the material, the cutting lubricant, and the use of carbide
during hot rolling, forging, and heat. treating, when the surroundin tools.
medium reacts with the carbon (as oxygen and carbon combining). Malleability is a material’s susceptibility to extreme deformation in
Ductility is that property that permits permanent deformation befo Tolling or hammering. The more malleable the metal, the thinner the sheet
fracture in tension. There is no absolute measure of ductility, but the Into which it can be formed (usually cold). Gold and aluminum are quite
centage elongation and the percentage reduction of area are used as indi malleable.
the higher these indices, the more ductile the material is said to be. Ducti Mechanical properties are those that have to do with stress and strain:
is the opposite of brittleness, but there is no sharp division line. Fe ultimate strength and percentage elongation, for example. See physical
purposes of definition, it is frequently assumed that Properties.
Ductile material > Elongation greater than 5% in 2-in. gage, Percentage elongation is the extension in the vicinity of the fracture of a
Brittle material -> Elongation less than 5% in 2-in. gage. tensile specimen, expressed as a percentage of the original gage length, as
Ductility is frequently a valuable property because, by virture of it, | 20% in 2 in.
member may take an occasional exceptionally-high load without breaking Percentage reduction of area is the smallest area at the point of rupture
Elasticity is the ability of a material to be deformed and to return to th Of a tensile specimen divided by the original area.
original shape. Stress is proportional to strain only during an elastic defor Physical properties exclude mechanical properties, and are other physi-
mation (see proportional limit, p. 8). Cal properties such as density, conductivity, coefficient of thermal
44 MATERIALS AND THEIR PROPERTIES [Ch, 2.3] HEAT-TREATMENT TERMS 45

expansion. See mechanical properties. Chemical properties include corrosi¢ Wrought steel is steel that has been hammered, rolled, or drawn in the
resistance. process of manufacture; it may be plain carbon or alloy steel.
Plasticity is the ability of a metal to be deformed considerably witho
rupture. In a plastic deformation, the material does not return to its origi
shape. See elasticity. 4 2.3 HEAT-TREATMENT TERMS.':*.2-5.2-81 Heat treatment is
Poisson’s ratio is the ratio of the lateral strain (contraction ) to t an operation or combination of operations involving the heating and cool-
longitudinal strain (extension) when the element is loaded with a longitu ing of metal or an alloy in the solid state for the purpose of altering the
nal tensile force. properties of the material. A few of the most common terms have meanings
Precipitation heat treatment brings about the precipitation of a cons as given below. See § 2.8 for case-hardening processes.
tuent from a supersaturated solid solution by holding the body at Aging (and age hardening) is a change in a metal by which its structure
elevated temperature, also called artificial aging. In some alloys, pre si recovers from an unstable or metastable condition that has been pro-
tation may also occur at ambient temperatures, a process called agi 12. duced by quenching or cold working. The change in structure, which
Proof stress is that stress which causes a specified permanent defort proceeds as a function of time and temperature, consists in precipitation
tion of a material, usually 0.01 % or less. See yield strength, § 1.8. ‘ often submicroscopic. The result is a change of mechanical and physical
Red shortness is a brittleness in steel when it is red hot. properties, a process that may be accelerated by using a temperature slightly
Relaxation, associated with creep, is the decreasing stress at a const higher than room temperature.
strain; important for metals in high-temperature service. Annealing, a comprehensive term, is a heating and slow cooling of a solid
Residual stresses are those not due to applied loads or temperat metal, usually done to soften it. Other purposes of annealing include those
gradients; they exist for various reasons, as unequal cooling rates,¢ of altering the mechanical and physical properties, producing a particular
working, etc. il microstructure, removing internal stresses (stress relieving), and removing
Rimmed steel is incompletely deoxidized steel. Ingots of this steel h gases. See normalizing below.
a surface layer quite free of slag inclusions and gas pockets, which result Critical range has the same meaning as transformation range (below).
the optimum surface on rolled sheets. Drawing is often used to mean tempering, but this usage conflicts with
Du
Solution heat treatment is the process of holding an alloy at a suif id the meaning of the drawing of a material through a die (§ 2.9) and is to
high temperature long enough to permit one or more constituents to} be avoided.
into solid solution and then cooling fast enough to hold the constitut Graphitizing, and annealing process, causes the combined carbon to
as a supersaturated solution. (Precipitation may occur with time.) transform wholly or in part into graphitic or free carbon; it is applied to
Stiffness is the ability to resist deformation. It is measured by cast iron, sometimes to high-carbon steel.
modulus of elasticity in the elastic range; the higher the modulus, Hardening is the heating of certain steels above the transformation
stiffer is the material. A Tange and then quenching, for the purpose of increasing the hardness. In
Strain hardening is increasing the hardness and strength by pla the general case, hardening is any process of increasing the hardness of a
formation at temperatures lower than the recrystallization range. } MInetal. See § 2.9.
temper. y Malleablizing is an annealing process whereby combined carbon in
Temper is a condition produced in a non-ferrous metal by mechanical white cast iron is transformed wholly or in part to temper carbon. Temper
thermal treatment: for example, annealed temper (soft), hard te carbon is free (graphitic) carbon in the form of rounded nodules, charac-
spring temper. See §§ 2.16 and 2.17. teristic forms in graphitizing and malleablizing. See § 2.12 concerning
Toughness is the capacity of material to withstand a shock load with: malleable iron.
breaking. The impact strength (see Charpy and Izod tests), though not E Bi, is the heating of an iron-base alloy to some 100°F above
absolute measure, evaluates toughness. Formerly, the energy requi ed al transformation range with subsequent cooling to below that range in
pull a standard tensile specimen in two was taken as the toughness, but 1 = air at room temperature. The purpose is to produce a uniform struc-
quantity is not representative because of the effect of the cold working
the specimen during the slow-speed test. Fr Spheroidizing is any heating and cooling of steel that produces a rounded
Transverse strength refers to the results of a transverse bend test, | globular form of carbide. Typically, it is a prolonged heating at a tem-
specimen being mounted as a simple beam; also called rupture modul Sa slightly below the transformation range, usually followed by
It is frequently applied to brittle materials, especially cast iron. Ow cooling; or, for small objects of high-carbon steel, it may be prolonged
Work hardening is the same as strain hardening. €ating alternately within and slightly below the transformation range.
 §2.5] AIS! AND SAE SPECIFICATION NUMBERS 47
46 MATERIALS AND THEIR PROPERTIES [Ch.2
Rockwell C (Rc), diamond indenter, 150-kg. load, for hard metals, as hard
Stress relieving (thermal) is the heating of a metal body to a suitable steel.
temperature (generally just below the transformation range for steel, say Rockwell A (Ra), diamond indenter, 60-kg. load, for extremely hard metals,
1100-1200°F) and holding it at that temperature for a suitable time (1 to 3 such as tungsten carbide.
hours for steel) for the purpose of reducing internal residual stresses. The Rockwell D (Rp), diamond indenter, 100-kg. load, sometimes used for case-
internal stresses may be present because the body has been cast, quenched, hardened metal.
normalized, machined, cold worked, or welded. j Rockwell E (Rg), $-in. ball, 100-kg. load, for soft metals, such as bearing
Tempering is a reheating of hardened or normalized steel to a tempera: metals and magnesium; F, (3-in. ball); H, K, L, M, P, R, S, V (with different
below the transformation range, followed by any desired rate of sizes of balls), are all used for soft materials as a substitute for the E scale; scale G
ture
(4s-in. ball) for phosphor bronze.
cooling. Quenched steel is tempered in order to reduce internal stresses,
The Rockwell superficial tester, a different machine, is used for a piece of
to restore a certain amount of ductility, and to improve toughness.
material too thin for the standard tester. The scales for this tester are N (for
time and temperature of tempering are selected in order to give the steel C-hard materials) and T (for B-hard materials). The N-scale indenter is a dia-
the desired properties. See Figs. AF 1, AF 2, AF 3, Appendix. mond, and the load may be 15, 30, or 45 kg., designated thus: 15-N, 30-N, 45-N.
If, for example, a particular steel with a particular yield strength i The T-scale indenter is a y-in. ball with loads as before and designated thus:
desired, do not specify the tempering temperature. This temperature ce 15-T, 30-T, and 45-T. ;
be varied slightly to produce closely the desired mechanical property. The Vickers tester has a square-base, diamond pyramid indenter, and
Transformation range for ferrous metals is the temperature interva the Vickers number is the load in kilograms divided by the impressed area
during which austenite is formed during heating; it is also the temperat in square millimeters.
interval during which austenite disappears during cooling. Thus, there The Shore scleroscope number is obtained by letting a freely falling
two ranges; these may overlap but never coincide. The range on heating1 hammer with a diamond point strike the object to be tested and measuring
higher than that on cooling. the height of rebound. This height is the Shore number; the higher the
rebound, the harder is the material. The Shore machine can be used on
large parts and is often used as a quick inspection aid, but it is less accurate
2.4 HARDNESS. The hardness of a material is a measure of it than the other tests. Conversions from one hardness to another may be
resistance to indentation, and is one of the most significant prope made in Fig. AF 4, which gives the approximate relationships for steel.
because, properly interpreted, it says much about the condition of The term file hard, often found in the literature, should be a hardness of
metal. The most common instruments used to determine hardness are thi perhaps 600 Brinell.
Brinell, Rockwell, Vickers, and Shore scleroscope. Increasing hardne:
numbers indicate increasing hardness. 9
2.5 AISI AND SAE SPECIFICATION NUMBERS. There are
The Brinell hardness number (BHN) is determined by a standard pres
numerous “‘standard” materials specifications."-?,?-32-6] Many large
sure (3000 kg. standard, 500 kg. for soft metals) applied to a 10-mm. ba consuming organizations and nearly all producers have some standards
which presses for 10 sec. or more on the surface of the material being tes ed of their own. The armed forces have numerous ones. However, the princi-
The load in kilograms divided by the area’ of the surface of the indentatiot pal agencies whose specifications are most widely used are the American
in square millimeters is the BHN. This hardness number is closely relates Society for Testing Materials (ASTM), the Society of Automotive En-
to the ultimate tensile stress of steel as follows:
gineers (SAE), and the American Iron and Steel Institute (AISI). The
(2.1) s, © (500)(BHN)
psi or — (0.5)(BHN) ksi. SAE and AISI specification numbers are alike for steel except that the AISI
[FOR STEEL WHEN 200 <BHN < 400] uses prefixes B, C, D, and E to indicate the method of manufacturing the
Carbon grades; see Table 2.1.
The probable range of values is (470) (BHN) < su < (530) (BHN). Us
In a general way for steel, the first digit (or the first two digits) of the
this approximation only when reliable test data are not available, and a
number represents a type of steel, for example: 1XXX is a plain carbon
not apply it to any other metal. f Steel, 11XX is a plain carbon steel with greater sulfur content for free-
The Rockwell tester, faster than the Brinell and widely used commer Cutting, 2XXX is nickel steel. The last two digits in four-digit numbers
cially, utilizes several different indenters and, in effect, measures the dept variably give the approximate or average carbon content in “points”
of the penetration by the indenter. Each indenter has an identifying symbo! or hundredths of per cent. For example, an SAE 1030 or an AISI C1030
as follows (always specify the Rockwell scale): has about 0.30 % carbon, spoken of as 30 points of carbon (nominal range
Rockwell B (Rg), z-in. ball, 100-kg. load, for medium soft metals, as fo’
many copper alloys and soft steel. ‘q
 48 MATERIALS AND THEIR PROPERTIES: [Ch. | 52.6] ALLOY STEEL 49
is 0.28-0.34%). Or in 8620, the average carbon content is close to 0.20% 2.6 ALLOY STEEL. Wrought alloy steel is steel that contains signi-
(range of 0.18-0.23 %). ficant quantitites of recognized alloying metals, the most common being
aluminum, chromium, cobalt, copper, manganese, molybdenum, nickel,
phosphorus, silicon, titanium, tungsten, and vanadium. Alloys are used to
improve the hardenability of steel (§ 2.7), to reduce distortion from heat
TABLE 2.1 SYSTEM OF SPECIFICATION
treatment, to increase toughness, ductility, and tensile strength, and to
NUMBERS FOR STEEL—AISI AND SAE improve low-temperature or high-temperature properties. See remarks on
In the AISI system, prefixes have the following meanings: B, acid bessemer steel; alloys below. With alloys, steel may be heat treated to the desired hardness
C, basic, open-hearth carbon steel; D, acid open-hearth carbon steel; E, electric-furnace
with less drastic quenching and therefore with less trouble from distortion
steel (usually alloy). Letters B or L in the middle of the number indicate that boron or
lead, respectively, has been added; as 94 B 40 and 11 L 41 (§ 2.6). An H at the end indicates and cracking. For small parts, a relatively small amount of alloy is needed
that material can be bought on hardenability specification, as 9840H (§2.7). in order for the part to respond in depth to heat treatment. Larger parts
should have greater amounts of the alloying elements for hardenability
STEEL SAE STEEL SAE purposes. Alloy steels may be classified as:
(a) Low-alloy structural steels (not heat treated). These steels (sy 2 50 ksi
Plain carbon .. es 10XX Molybdenum-chromium-nickel | 47XX
Free cutting o exe 11XX Molybdenum-nickel .. «» |48XX as rolled) were developed for structural uses where light weight is important (but
Manganese te te 13XX Chromium nh Me oe SOCK not extremely so as in aeronautics), such as in the transportation industry, but
they are also used in other structures. Phosphorous (0.03-0.15%) is an effective
ee Bi zi pi pens heat and corrosion resistant | vie strengthener, as is nickel (0.5-2%) and copper (0.2-1.25%). Copper also imparts
Nickel-chromium = 3XXX || Chromium-vanadium . «+ |OXXX resistance to atmospheric corrosion. The carbon is typically some 0.15-0.20%;
heat and corrosion resistant |303XX || Nickel-chromium-molybdenum |8XXX but more may be used on occasion. Other alloys are manganese, silicon, chro-
Molybdenum .. < 4XXX || Silicon-manganese .. «« |92XX mium, and molybdenum, but not necessarily all at once. They weld easily and do
Molybdebum-chromium 41XX Nickel-chromium-molybde- 9XXX not air-harden.
Molybdenum-chromium-nickel |43XX num (except 92XX)
(b) Low-carbon alloy steels (0.10-0.25%C), AISI steels, used chiefly for
Molybdenum-nickel .. 46XX
carburizing.
(c) Medium-carbon alloy steels (0.25-0.50% C), usually quenched and tem-
pered to hardnesses between 250 and 400 Brinell.
(d) High-carbon alloy steels (0.50-0.70% C or more), ordinarily heat treated
Examination of Table AT 7, for example, shows that in general # 2 hardnesses between 375 and 500 Brinell, for use as springs, wear resisting parts,
strength of steel increases with carbon content, while the ductility| etc, 7
creases. A brief suggestion of typical uses of plain carbon wrought steel (e) High-alloy steels, such as stainless steels.
as follows. A few brief remarks about the principal alloying elements will suggest
Carbon, 10-20 points, 10XX group. Used for tubing, forgings, pressed: ste other functions of alloys (chemical symbol in parentheses).
parts, screws, rivets, and for carburized case-hardened parts. Aluminum (Al) is an efficient deoxidizer, an alloy in nitriding steels (nitralloys),
Carbon, 10-20 points, 11XX group. Due to higher sulfur content in cert and it promotes fine grain size.
grades, it is free-cutting and good for use in automatic screw machines 1 Boron (B) in very small amounts (0.001% or less) is an economical harden-
miscellaneous parts, including screws; it also may be carburized. For case ha de ability agent in low- or medium-carbon deoxydized steels. It has no
effect on
ing, the open-hearth steels, identified by the symbol C in the AISI number, a tensile strength.
to be preferred. Higher carbon-content steels in the 11XX group, as 1141, conta Chromium (Cr) improves hardenability economically, resistance to corrosion
more manganese and are heat treatable for improved mechanical prope! a ae alloys), strength at high temperatures, and wearing properties (high
See Table AT 9. These steels are not usually welded. rdon),
Carbon, 20-30 points. General purpose grades, used for forged and machi ne Cobalt (Co) improves red hardness.
parts; screws; also for boiler plate and structural steel. Columbium (Cb) is often used to “stabilize” stainless steel (that is, it preempts
Carbon, 30-55 points. With 0.40-0.50% C, frequently used for miscelle the carbon and forestalls the formation of undesired carbides).
forged and machined parts; shafts. Frequently heat treated for improved me a pe (Cu) improves steel’s resistance to atmospheric corrosion; up to 4%
nical properties. Cold finished for shafting and similar parts. ae | increases the fluidity of the melt; it improves tensile strength and the yield
Carbon, 60-95 points. May be hardened to a good cutting edge, especially ' m in the normalized condition. Yield ratio = s,/s,. With more than 0.75%
Cu,
the higher ranges of carbon; therefore, used for tools. Also for springs. €els can be precipitation hardened.
strength, low ductility. Nearly always heat treated, say, to a Brinell hardness 0 Lead (Pb) improves machinability, but affects different alloys differently.
375 or higher.
 et are et ee ee ee ne, he ee el oer NL Fee LY ARE Se fee ne ge ae

Manganese (Mn) improves strength and increases hardenability moderately If an alloy steel is used, it should, in general, be heat treated in order to
counteracts brittleness from sulfur. Present in all steels, manganese becomes aj obtain the best properties for the purpose. See tables in the Appendix.
alloying element when its amount exceeds about 0.6%, as in the 13XX steels Illustrative of the uses of alloy steels, we have:
Medium-carbon manganese steels are subject to embrittlement at temperature AISI 2330: bolts, studs, tubing subjected to torsional stresses.
above 600°F. Austenitic manganese steel (not 13XX) typically contains 1.2% € AISI 2340: quenched and tempered shafting, connecting rods, very highly
and 12-13% Mn and responds to work hardening most readily. stressed bolts, forgings.
Molybdenum (Mo) increases hardenability markedly and economicall, AISI 2350: high-capacity gears, shafts, heavy duty machine parts.
(when Mo > Cr), tends to counteract temper brittleness, improving cree AISI 3130: shafts, bolts, steering knuckles.
strength and red hardness; it improves wear by forming abrasion-resistant AISI 3140: aircraft- and truck-engine crankshafts, oil-well tool joints, spline
ticles. It is the most effective alloy for improving strength at high temperat; shafts, axles, earth moving equipment.
(Alloys of molybdenum—not molybdenum alloys of steel—are proving to be AISI 3150: wear-resisting parts in excavating and farm machinery, gears,
effective at temperatures above about 1500°F. See § 2.21.) forgings.
Nickel (Ni) strengthens unquenched and annealed steels, toughens ste¢ AISI 3240: shafts, highly stressed pins and keys, gears.
(especially at low temperatures), and simplifies heat treatment by lesse AISI 3300 series: for heavy parts requiring deep penetration of the heat
distortion. It is the most effective element for reducing the brittleness of treatment (hardenability) and high fatigue strength per unit of weight.
at very low temperature; see § 2.22. It is one of the principal alloys for stainle AISI 4063: leaf and coil springs.
steel (§ 2.15). AISI 4130, 4140: automotive connecting rods and axles, aircraft parts and tub-
Phosphorus (P) increases hardenability, strengthens low-carbon steel: ing.
improves machinability of free-cutting steels, and improves resistance to co} AISI 4340: crankshafts, axles, gears, landing gear parts; perhaps the best
sion. general purpose AISI steel.
Selenium (Se) improves machinability of stainless steel; also added to leade AISI 4640: gears, splined shafts, hand tools, miscellaneous heavy duty
resulfurized carbon steels for the same purpose. i machine parts.
Silicon (Si) strengthens low-alloy steels and improves resistance to high AISI 8630: connecting rods, bolts, shapes; air hardens after welding.
temperature oxidation; it is a good general-purpose deoxidizer and prom AISI 8640, 8740: gears, propeller shafts, knuckles, shapes.
fine grain. Alloys with 10-20 points carbon are widely carburized (§ 2.8) in producing
Tantalum (Ta) is a stabilizer (see Columbium). pins, bolts, gears (teeth), shafts (at wearing surfaces), cams, and worm threads.
Titanium (Ti) is used for deoxidation and for stabilizing austenitic staink
steels (preventing intergranular corrosion and embrittlement); it increases
hardness and strength of low-carbon steel and improves creep strength.
Tungsten (W) increases hardenability markedly in small amounts and im 2.7 HARDENABILITY. Hardenability is the capacity of steel to
proves hardness and strength at high temperature. An expensive alloy, it is ust through-harden when cooled from above its transformation range. It is
only where a particular advantage results, as in high-speed tool steel in whie determined from a standard l-in. round specimen, Fig. 2.1, with the test,
it forms a hard, abrasion-resisting carbide. called a Jominy test, conducted according to a standard procedure. Flats
Vanadium (V) promotes fine-grain structure, improves the ratio of endurant
strength to ultimate strength of medium-carbon steels (average about 0
increases hardenability strongly when dissolved, and results in retention FIGURE 2.1 End-Quench Harden-
strength and hardness at high temperature; it is the most effective element if ability Test. The specimen is heated
retarding softening during tempering. to the proper quenching temperature,
placed in a fixture as shown in this
Since alloy steels are more costly than plain carbon steels, an allo
illustration, after which a spray of
should not be employed unless its use yields some advantage. By proper water strikes the lower end. The hard-
balancing the alloy and carbon contents, one may obtain a particula est part of the specimen will be the end
strength with higher ductility or a particular desired ductility with highé exposed to the jet of water, the most
strength than is possible without alloys. Since alloys generally improve drastic quench. Hardness will de-
Crease as the distance from the
mechanical properties, their strength/weight ratio is higher, and therefor
Sprayed end is increased, because the
their use may result in smaller parts which partly offsets the increased co rapidity of cooling decreases to prac-
per pound. It is important to note that the modulus of elasticity E (and tically that of air cooling at the far
is virtually the same for alloy steels as for carbon steels and that therefor end. Good hardenability is especially
if rigidity is the basis of design, there is no advantage in using alloy steel—am important when the entire section is
Subjected to high stresses; it is less
alloy steel deflects the same amount per unit stress as does a carbon steel
important when the high stresses are
On or near the surface. (Courtesy U.S.
Steel Corp., Pittsburgh).
 Equivalent Hardness at Center—Water Quenched
1 2 3 4 5
§2.8] CASE HARDENING 53
: Equivalent oe —< a

2.8 CASE HARDENING. Case hardening of iron base alloys is a

process of surface hardening whereby the surface or case is substantially
Rockwell C Hardness

harder than the core or inside metal. Case hardening is done by carburizing,
cyaniding, nitriding, carbonitriding, induction hardening, and flame hard-
ening. 2-1-2.3.2.6,2.8) ¥

The purpose of case hardening is usually to provide a hard, wear-

resistant surface while retaining great toughness in the core. It is also used
for increasing the endurance strength of certain machine parts and for other
purposes where high strength and hardness on the surface are advantageous.
(a) Carburizing. Carburizing is a process of adding carbon to the
lin. 2 in.
4 8 12 16 20 24 28 32
surface of steel by exposing it to hot carbonaceous solids, liquids, or
gases—above the transformation temperature. Quenching, and usually
Distance from Quenched End, 16th Inches
tempering at 300-450°F for the purpose of relieving residual stresses
FIGURE 2.2 Hardenability Curves. The materials shown were selected to show di induced by quenching, follow carburizing. The common methods of
ferences. Notice that the hardness of AISI 4340 holds up well with depth, that the h adding carbon are pack (or box) carburizing and gas carburizing. In pack
carbon 4063 has the hardest surface, that the same carbon steels, 4340 and 1340, carburizing, the part is heated in contact with solid carburizing com-
practically the same surface hardness (at 7; in.), and that the low carbon 4620 (actual 0.
pounds of various constituents, including charcoal, burned bone, charred
C) has lower hardness all the way. The individual lines are typical actual test values.
H band shown for 4140 is suitable for specifications. The top coordinates are defined by
leather, tar, and barium, sodium, and calcium carbonates, especially barium
example: a 2 in., oil-quenched piece of 4063 has a hardness at its center of R, =4 carbonate and charcoal. The depth of the case and the rapidity of the
(Courtesy Bethlehem Steel Co., Pittsburgh). r process depend in part on the soaking temperature, which is of the order
of 1650-1750°F,
are ground longitudinally on opposite sides of the cooled specimen In gas carburizing, which has been developed to an efficient and
which the Rockwell C hardness is determined at each 7-in. interval fi economic procedure, especially for large quantities, the part is heated in
the quenched end. The hardness of a quenched surface is largely de carburizing gases, such as methane, ethane, propane, and CO. The
dent on its carbon content, while the depth hardness, or hardenabilit temperatures of operation and case thicknesses obtained are much the
depends on the amount of carbon, the alloy content, and the grain size, same as those in pack carburizing. After 4 hr. at 1700°F, the thickness of
The most effective alloys in improving hardenability are boron, vam the case should range between 0.04 to 0.05 in. In liquid carburizing the
dium, manganese, chromium, molybdenum, phosphorus, tungsten, al part is immersed in a molten salt bath that imparts a case similar to that
to a lesser extent, nickel and silicon. Smaller amounts of several elemeni obtained with gas or pack carburizing except that the case is thinner, usually
increase hardenability more than does a larger amount of one alloy. Not in excess of about 0.025 in.
Typical hardenability curves are shown in Fig. 2.2, where it is seen hi For heavy duty, as in some gear teeth, a case thickness of 0.06 to 0.09 in.
the hardness at a certain point on the test specimen corresponds to # may be desired. It would seem that a safe design value of the surface hard-
hardness at the center of a certain size rod. This relationship is work hess of carburized steel would be about 600 BHN. The hardness should
out experimentally for a particular kind of quench and steel. Althou, generally fall between the limits
it is easy to harden a light piece all the way through, the material of a h
55 < Re < 65 or 560 < BHN < 730.
piece must be chosen with forethought if through hardness (or near
through hardness) is desired. Through hardness is not always wanted. 1 Carburizing steels are low carbon steels, say 0.15-0.25% carbon.
softer core may be an asset. a ae (b) Cyaniding. As in liquid carburizing, cyaniding is accomplished
Some steels, frequently designated by H as a suffix to the AISI numbé y Immersing the part in a hot (about 1550°F) liquid salt bath, sodium
(Table 2.1), may be bought on the basis of hardenability by specifying, i Cyanide (NaCN) being a common medium in both processes. The difference
example, the limits of hardness of a Jominy specimen at a certain p in the processes lies largely in the use of a catalyst in liquid carburizing that
along its length. Suppose that a hardness of Rc = 40 is desired at a ce Tesults in a more rapid penetration of carbon and a relatively small amount
depth and that this corresponds to Rc = 45 at 3§ in. in the Jominy of nitrogen in the case. Thus, the so-called cyanided case has more nitrogen,
the H band of Fig. 2.2 shows that AISI 4140 would be satisfactory in th Which is also a hardening agent. Whereas the thickness of the case of liquid-
respect. carburized parts may be somewhat greater than 0.02 in., the cyanided case
54 MATERIALS AND THEIR PROPERTIES [Ch.2
52.8) CASE HARDENING 55
is seldom thicker than 0.010 in. Low- and medium-carbon steels are usually surface hardening of steels whose carbon content is in the range of 0.35 to
used for cyaniding, and the case hardness may be of the order of that 0.55%, the carbon content with which steel readily responds to heat treat-
obtained by carburizing. ; ment. The depth of the hard case may be regulated so that, for example,
(c) Nitriding. In surface hardening by nitriding, the machined and the Rockwell hardness is C 50 or greater to depths of 0.02 to 0.17 in. The
heat-treated part is placed in a nitrogenous environment, commonly ams surface hardness may be of the order of Ro = 50 to 55 or higher (BHN =
monia gas, at temperatures much lower than those used in the previously 500 seems to be a reasonable design value); the core is of the order of R c=
described processes, say 1000°F or somewhat less. Since a nitrided part do es 30 to 35. See Fig. 2.3. After quench hardening of the heated surface, the
not need to be quenched rapidly, this process avoids the distortion that part is preferably tempered at some 400-450°F. Surface hardening through
accompanies quenching, sometimes a big advantage, especially for com induction heating is used also for cast iron and malleable iron. Since
plicated shapes. The hardening is the result of a reaction of the nitrogen
dissociated from the ammonia, with the alloying elements in the steel t
form nitrides. For maximum hardness of the case, special steels, called nitr FIGURE 2.3 Hardness of a Gear Tooth, Induction
Hardened. Observe the hardness close to the surface
alloys and containing aluminum as an alloy, are used; yet other st and near the middle of the base of the tooth (where the
notably AISI 4340, are frequently nitrided. The case hardness of AIS} lower hardness indicates a tough core). Hardness
4340, tempered at 1025°F, and nitrided at 975°F for about 40 hr., may t readings are Rockwell C. (Courtesy Ohio Crankshaft
over 600 Vickers (560 BHN) and the case depth some 0.025-0.030 Co., Cleveland).
In other situations, the nitriding time may be upwards: of 90 hr., whic
together with the control problem, accounts in part for its high cost.
The carbon content of the nitralloys falls within the approximate ranj
of 0.20 to 0.40% carbon. The case hardness of Nitralloy N, for exam
nitrided at 975°F for 48 hr. should be above Vickers 900 (equivalent Brine
of about 780), or better than Ro = 67. The case thickness may be fro)
0.010 to 0.020 in. The case is quite strong for a body subjected to bendil
or tension, so that failures that result from repeated stresses usual FIGURE 2.4 Induction-hardened Gear Teeth. If
originate in the transition region between the case and core. a gear teeth are heated in an induction hardening
(d) Carbonitriding is a process of case hardening steel by the simu machine, the heating follows the contour of the teeth
as suggested by this illustration and Fig. 2.3. The
taneous absorption of carbon and nitrogen from a surrounding hot gaseot
lighter areas were rapidly heated, after which the
atmosphere, followed by either quenching or slow cooling, as required. It gear was quenched. Besides being advantageous for
used for both batch and continuous processes. With a sufficient percen' contour hardening, as in gear teeth and cams, this
(up to 15%) of ammonia in the carburizing gas, carbon steel parts Process is well adapted to some selective hardening
emerge file-hard without quenching. (See nitriding.) Very small amoun jobs, as hardening journal surfaces while leaving the
rest of a shaft unaffected.
of ammonia (less than 1%) in the carburizing gas are sufficient to perm
attainment of maximum hardness with an oil quench. The use of sma
quantities of ammonia in combination with quenching is cheaper thé induction heating follows well the contour of the part being heated, it is
using a larger amount of ammonia. In continuous furnaces at 1500-15 appropriate for cams, gears (Figs. 2.3 and 2.4), and other irregular sur-
the case depth may range from 0.003 to 0.010 in., depending on time @ faces to be hardened.
temperature.'*-1) This process is used as a low-cost substitute for cyanidif : (f) Flame Hardening. Flame hardening, like induction hardening,
and produces a product of good quality. Some hard cases produced by th 1S a process of heating the surface of an iron-base alloy, which is preferably
method have been reported to wear many times longer than the be a or normalized, and then quenching it. Typically, neutral acetylene
cyanided or carburized cases previously used. ! ~ames are played upon the surface to be hardened, followed closely
by
(e) Induction Hardening. Induction hardening consists of hea’ Jets of water for cooling. While the process can be carried on manually,
a thin surface layer, preferably of annealed or normalized steel, above More consistent results are obtained by using especially designed machines.
transformation range by electrical induction and then cooling, as requ his method is applicable to the same metals as is induction hardening
in water, oil, air, or gas. Since this process rapidly heats a thin layer of (say 0.45 % C for steel), and while it is used for both small and large parts,
surface, leaving the core relatively cool, the process is widely used f¢ It has particular advantages for very large parts where selected surfaces
 % of Specimens
FIGURE 2.6 Cold-Drawn AISI 1117. Scatter of
yield strengths of 1 in. round bars, from 25 heats and
2 vendors. After Ref. (2.1).

70 80 90
Yield Strength, ksi

the strength of carbon steel enough that it is sometimes used instead of a

more expensive heat treatment. However, the natural spread of mechanical
properties is likely to be greater than for heat treated parts, suggesting
caution. See Fig. 2.6. Cold drawing leaves residual tensile stresses that may
be deleterious to fatigue strength; see § 4.30.
Steel is also drawn at elevated temperatures, with an even greater in-
crease in strength (up to a material temperature of about 1000°F during
drawing) as compared to cold-drawn material, stress relieved at the same
temperature. Maximum strength for AISI 1144 bars, 1-in. diameter, is
obtained when the material is drawn at about 600°F (a proprietary process*
FIGURE 2.5 Machine for Flame-hardening Gear Teeth. An example of a special fiz
called elevated-temperature drawing, ETD) and no stress relieving is
hardening machine. This machine flame-hardens the surfaces of bevel-gear teeth on
up to 24-in. The hardening operation is electroniclly controlled and automatic; ada
needed!-11_but a higher drawing temperature may be needed for im-
to spur, helical, and herringbone gears, and to straight, spiral and Zerol bevel gears. proved ductility. The benefits of this method of finishing will in some cases
hardening in general, which is especially well adapted to hardening selected surfaces on lar eliminate heat treating operations and avoid the need for alloy. For com-
parts, may also be manually controlled. (Courtesy Gleason Works, Rochester, N.Y.) i parison purposes (see Tables AT 7 and AT 10), some properties of AISI
1144 processed in this manner are: s, = 140 ksi min. (150 ksi typical),
are hardened—such as the ways of a lathe or the surface of the teeth¢ Sy = 125 ksi min., BHN = 280, elongation ~ 10%; and it has a good
large gears (Fig. 2.5). The resulting surface hardness should be of the machinability rating of 80.
of 500 Brinell for 0.45 % C steel, and the distortion may be negligible unc Additional remarks concerning cold working of other metals are found
controlled conditions. q below. The cold working of surfaces (plastic deformation limited to a
thin surface layer) by peening and rolling is taken up in Chapter 4.

2.9 WORK HARDENING. Work hardening is the result of a me

being stressed at some point into its plastic range, usually ordinary 2.10 WROUGHT IRON. Wrought iron is made by burning the car-
peratures (certainly below recrystallization temperature) ; metal cold wo: bon from molten iron and then putting the product through hammering
in this manner becomes stronger and more brittle. ‘‘Cold-finished’’ n and rolling operations. The product contains some 1-3% slag and less
terial* has had its cross section significantly reduced by cold rolling (usué than 0.1% carbon. The material is very soft and ductile and is easily forge
used for flat products) or cold drawing through a die (usually employ welded. It is used principally for rivets, welded steam and water pipes, and
to produce cold-finished rods). By 10% cold work, for example, is for general forging purposes. Its most advantageous properties are its
that the cross-sectional area is reduced 10% during the process. Typical ductility and resistance to corrosion as compared to steel.
for steel, the dimensional reduction per pass is 4; or 3 in., with a tot
reduction of 20% to 12% or less. A 12% reduction of steel (whose s, < |
ksi) results in an increase of about 20% in ultimate strength (closely 2.11 CAST IRON. Cast iron in a general sense includes white cast
straight-line variation to this point), an increase of about 70% in iron, malleable iron, and nodular cast iron, but when cast iron is used
strength (values to this point are all greater than the straight-line variat without a qualifying adjective, gray cast iron, spoken of as gray iron is
and a decrease of about 35% in percentage elongation.) The tought Meant. In general, gray iron contains so much carbon (2.6-3.6% usually)
also decreases, and machineability increases. Cold drawing improve that it is not malleable at any temperature. In gray iron, the excess carbon
is uncombined, and a fracture is gray.
* Cold finished is a term also applied to rounds that are turned or ground, but our use
the term will be as defined. y * La Salle Steel Co.

57
 MATERIALS AND THEIR PROPERTIES [Ch.2 §2.13] NODULAR CAST IRON 59
58

The ASTM, in specification A 48-46, has classified gray iron according elasticity E used when the operating stress is high. See Fig. 1.3. Typical
iro a quoted values of E would be the slope of a straight line from the origin
to minimum tensile strength.-*! See Table AT 6. Thus, a “30 gray
of the s-e diagram to a point on the curve at s,,/4.
will have a standard test strength of not less than 30,000 psi (30 ksi)
In white cast iron (the fracture is ‘“‘white’’), most of the carbon is com-
Strength is increased by reducing the carbon content (about 3.7% foi
for class 60), and the higher-strength cast irons contain bined chemically with the iron, and as a result, the metal is very hard. If
class 20, 2.8%
typical alloys.*-1-2-+4.2-15
4 the combined carbon is as much as 1.5%, the cast iron may be difficult or
Cast iron is sometimes given a heat treatment, but ordinarily it is cheape
impossible to machine. When an extremely hard surface is desired, white
to improve strength and other properties by reducing silicon and/or carbo cast iron, called chilled iron, is intentionally produced by using an iron
or by increasing alloy content than by heat treating. The histogram of Fig plate in the mold to cause rapid cooling of the surface. The rapid cooling
2.7 shows a typical distribution of tensile strength as obtained from s does not allow time enough for the carbon to be released as free carbon.
60 tests. Rigidly inspected, this lot would classify as class 35. Such a surface may be finished only by grinding and is suitable for car
Gray iron has excellent wearing properties that are improved by ce
wheels, rolls, etc.
alloys and by heat treatment (including flame and induction hardening)
2.12 MALLEABLE IRON. Malleable iron is heat-treated white
cast iron.@+1-2-17] The white cast iron is obtained not by chilling, as men-
tioned above, but by using the proper composition in the melt. The heat
treatment of the white cast iron, in which substantially all of the carbon is
combined in the form of iron carbide, is an annealing, called malleablizing
during which the white iron changes to ferrite and free (or temper) eacbiont
Malleable iron produces strong, ductile, and easily machined castings
FIGURE 2.7 Tensile Strength of Cast Iron. at low cost in quantity. Best results are obtained on relatively thin sections.
See Fig. 2.15. If the part is thicker than about 3 in., there will be difficulty in
Tensile Strength, ksi producing a white-iron casting devoid of uncombined primary graphite.
Since it is necessary that all the carbon in the original casting be combined
carbon for the best results, malleable castings are generally produced in
Thus, it is widely used for cylinder blocks, brake drums, gears, machine te
sections of from } to 2 in. thick. See grades 32510 and 35018 in Table AT 6.
ways, and in general where there is metal-to-metal contact and relat
oe hardening, as for gear teeth, may produce a surface hardness of
motion. The lower-strength varieties and all grades in the annealed st
c © 5.
are easily machined. Gray iron is more resistant to many kinds of corr
A special form, pearlitic malleable iron differs from normal malleable
than are ordinary or low-alloy steels.?-") It has much greater damp
iron in that there is a significant amount of combined carbon in the finished
capacity (for vibrations) than steel, which suggests its use under
Product. The difference is achieved by adding alloys, > changi
vibrating conditions. It is successfully used in crankshafts, as in autom: ging th the heat
hans
engines. See Fig. 2.14. Being the cheapest of metals, ordinary (low-stre treatment, or both.
gray iron is the most widely used of all cast metals. Its principal d
vantages are its brittleness and lack of toughness, but these properties é
a NODULAR CAST IRON. Nodular cast iron, also called
often not significant. Cost tends upward with specification numbers, ‘
35; the cost of melted metal for good castings fi
. e iron, has the castability (for complex forms), machinability, and
pecially above ASTM
earability of gray iron, but higher strength and ductility. It thus has
class 60 iron runs about 2.3 times that for class 30 iron,” assuming
&00d toughness (shock resistance). Instead of flakes as in gray iron, the
production; and there are other sources of increased cost.
ae has a spheroidal shape, obtained by the addition of certain alloys,
Cast iron loses strength with increase of the minimum section d
bar. bal y ee magnesia (a few hundredths of 1%) and cerium (a few
sion. A }-in., class-20 bar has a tensile strength of 28 ksi; a 4-in.
fe sandths of 1%), both because they are helpful to each
s, ~ 11 ksi, no longer class 20. For a 1-in., class-60 bar, s,, ~ 65 ksi; other. See
able AT 6, but there are other standard specifications.?:}!
a 4-in. bar, s, ~ 44 ksi.2+! Hooke’s law is not too well approximated
must be careful of the values of the modulus‘ Nodular iron is used for a wide variety of items, including casings,
cast iron, so that one
60 MATERIALS AND THEIR PROPERTIES [Ch. 2 §2.15] STAINLESS STEEL él

crankshafts, hubs, rolls, forming dies. It has a good resistance to therm al (a) Cold working, § 2.9, which is the usual way of hardening the
shock, and its “growth” at high temperature is less than that of gray iron, austenitic types because of their potent response to this treatment. Quench-
The as-cast 80-60-03, Table AT 6, is pearlitic and harder than the annealed ing these steels from about 1850°F avoids austenite transformation and
60-45-10, which is ferritic, and therefore the pearlitic has better wear pr leaves the steel soft, an effect opposite to that for ordinary steel. Cold-
perties. In general, nodular iron has wear properties comparable to those worked austenitic steels are classified according to temper as } hard (for
gray iron of the same hardness, which is good. Nodular iron (pearlit which minimum s, = 125, s, = 75 ksi), } hard (minimums, s, = 150,
responds well to flame or induction hardening, perhaps with Rg > 55, but sy = 110 ksi), } hard (minimums, s, = 175, sy, = 135 ksi), full hard (mini-
residual casting stresses should be relieved before treatment to avoid mums, s, = 185, s, = 140 ksi). Austenitic stainless is the best stainless
possible cracks, and after treatment, a stress relief at 300-400°F may be: steel for high-temperature service; its corrosion resistance is better in the
desirable. il annealed than in the work-hardened state. Uses of some types include:
301—hardware, utensils, structural; 302—strong at elevated temperature,
decorative and corrosive applications, widely used; 303—free machining
2.14 CAST STEEL. The combination of highest strength and 18-8; 321—carbon stabilized by titanium, which makes it good for welding
highest ductility in a cast ferrous metal is obtained in cast steel.(2-1.2.16 without subsequent annealing; 347—columbium (Cb) and tantalum (Ta)
When steel castings are heat treated, the carbon content generally fe stabilized for welding without annealing.
within the range of 0.25 to 0.50%. However, many steel castings (b) Age hardening, usually termed precipitation hardening with
carbon less than 0.20% are used in the as-cast condition. reference to stainless, which occurs because of the precipitation of a consti-
Steel castings may be plain carbon or alloy steels. Among the AST tuent from a supersaturated solid solution. It is generally conducted at
grades of plain carbon steel castings in Table AT 6 are (ASTM A27-58 elevated temperatures in order to increase the rate of precipitation. The
60-30, 65-35, and 70-36, where the first number, say 60, represents ih stainless steels 302, 303, 304 and 316 are subject to precipitation hardening
minimum tensile strength and the second number, say 30, represents at 800-1650°F. The 17-7PH (17% Cr, 7% Ni), Table AT 4, which contains
minimum yield strength at 0.2% offset, each in ksi. There are sey considerably less carbon, manganese, and silicon than 301, is precipitation
high-strength cast steels in Table AT 6 (ASTM A148-58) that are o hardened when mill annealed at 1950°F, air cooled and reheated to 1400°F
referred to in a similar manner, as 80-35, where the numbers have for 90 min., air cooled and, to accelerate precipitation of carbides, re-
meanings explained above. To obtain the high strength, the foundry heated to 1050°F for 90 min. Carbide precipitation is a migration of carbon
regulates the carbon and alloy content (manganese, silicon, and oth to the grain boundaries where it combines with the chromium to form
alloys) in such a way as to produce the mechanical properties stated in chromium carbide. The depletion of the chromium, the element that makes
specifications. A minimum heat treatment is annealing or normali the steel stainless, adjacent to the grain boundaries results in the boundary
Alloy cast steels respond to heat treatment in much the same way as sin material of some grades being susceptible to highly corrosive media.
wrought alloy steels do. Not all grades of cast steel are adapted to wel Corrosion of this sort is called intergranular corrosion. The lower the
but usually there is no trouble in this respect. Special alloys for corr 2 carbon content, the less carbide precipitation there is; hence, these special
resistance, heat resistance, and for other particular purposes are commol grades of stainless steels are made with carbon content less than 0.08%
(versus a normal 0.15-0.25 %), to be used, for example, where a body is to
be welded without subsequent annealing—as in the case of a body in the
2.15 STAINLESS STEEL. Table AT4 gives the mechanical field or too large for available annealing furnaces. Other grades of austenitic
perties of some of the “standard” stainless steel specifications. Sta Stainless steels have the stabilizing elements columbium plus (maybe)
steel is relatively expensive, but where the environment is signific tantalum or titanium added. These stabilizers combine with the carbon,
corrosive or at high or quite low temperatures, it provides an econon thus preventing it from combining with the chromium. The 347 grade is
answer for many problems. There are three classes: austenitic steels Stabilized with columbium and, moreover, contains less than 0.08%
and 300 series—that include 3.5 to 22% nickel for its stabilizing of a) carbon. Another treatment for precipitation hardening stainless steels
nite), martensitic steels (usually with no nickel, but some types have Involves cooling to some —100°F to transform the austenite to martensite;
maximum), and ferritic steels (no nickel) that do not harden by quenchit annealing at this stage results in a very high yield strength. Some types of
and tempering. All classes contain chromium (4 to 26 °). Methods¢ austenitic manganese steels are capable of being precipitation hardened,
hardening are: 4s well as many nonferrous alloys.
62 MATERIALS AND THEIR PROPERTIES [C ia
§2.17] ALUMINUM ALLOYS 63

(c) Quenching and tempering, as for usual steels, except that trang cold work determines the temper; the softest grade is annealed. Reductions
formations are so slow that the quenching is by air cooling for maximui for other tempers are approximately as follows :'2-1)
hardness. This is the common way of hardening the martensitic ty
(although they respond in varying degrees to cold working). The tyz TEMPER STRIP WIRE TEMPER STRIP WIRE
410 is perhaps the least expensive, and its response to heat treatment is e
than that of the 431 and some other types. The 403 is used for forged ty } hard 10.9% 20.7% Hard 37.1% 60.5%
4 hard 20.7% 37.1% Extra hard 30%. (se
bine blades; 410 is a general purpose type (screws, rods, shafts, rivet
3 hard 29.4% 50 % Spring hard 60.5% 844%
pistons, knife blades) and can be successfully cold formed; 431 is a hi Extra spring hard 68.7% 90.2%
strength stainless.
The ferritic stainless steels do not harden significantly by heat treatm Copper alloys high in zinc (over say 35 %; e.g., naval brass) cannot take the
nor do they work harden excessively. Cold working these steels raises extreme cold working because they become excessively brittle. A few
yield strength 30% or more, but the increase of ultimate strength is uses of some of the copper alloys follow.
less. Type 430 is widely used in the annealed state for auto trim, restaura Admiralty metal: condenser and other heat-exchanger tubes and plates.
fixtures, heat exchanger flues, chemical equipment, etc. Aluminum bronze: corrosion-resistant parts; marine pumps, shafts, valves;
parts where high strength, toughness, wearability, low coefficient of friction, and
damping are important, as some bearings, gears, worm wheels, cam rollers; also a
2.16 COPPER ALLOYS. Since copper is one of the oldest kn decorative metal, as in statues and costume jewelry.
metals, it has been the base of many alloys, as well as being used Beryllium copper (also called beryllium bronze): parts where high formability,
relatively pure form.-1:2-! Originally, the words brass and bronze, whi high yield, fatigue, and creep strengths, and also good corrosion resistance are
advantageous; springs, bolts and screws, firing pins, dies, surgical instruments,
have been used for hundreds of years, had fairly distinct meanings
spark resistant tools. It is solution annealed by holding at 1450°F for about 1 hr.
being an alloy of copper and zinc; bronze being an alloy of copper and
per in. or fraction of inch; quench rapidly in water.!?-4) It is precipitation har-
However, the names have become so confused that the only safe dened by starting with solution-annealed metal and then for say 4 hard, heat
know an alloy is to know its actual composition. Frequently copper: for 2 hr. at 600°F. The properties of beryllium bronze are quite attractive, but it is
alloys contain some tin or copper-tin alloys contain zinc. Moreover, some 5 times as expensive as brasses.!?-1)
other alloys for copper have come into use, notably aluminum, sili Cartridge brass: electrical parts, automotive radiator cores, pins, rivets,
beryllium, and cadmium. The tendency is to call the newer copper springs, ammunition components, tubes.
bronze, even though there is little or no tin present, as in aluminum bror Mire bronze: clutch disks, pump rods, shafts, valve stems, welding
and silicon bronze. Manganese bronze is really a high-strength brass rod.
improved mechanical properties being obtained by including small a Naval brass: condenser plates, marine hardware, propeller shafts, piston rods,
valve stems, welding rod, balls, nuts, bolts, rivets.
of aluminum, iron, manganese, and tin. Phosphor bronze is a bronze, t
Phosphor bronze: bellows, diaphragms, clutch disks, cotter pins, lock washers,
the finished product may-have only a trace of the phosphor that was a
bushings, springs, wire, welding rod, chemical hardware, wire brushes.
primarily to deoxidize the melt, a treatment that improves the mecha
Silicon bronze: hydraulic pressure lines, hardware, bolts, nuts, rivets, screws,
properties. See Table AT 3 for a few common Repeer alloys; the comp: electrical conduits, heat-exchanger tubes, welding rod.
tions given are approximate. Yellow brass: electrical fixtures, plumbing, wire, pins, rivets, screws, springs,
Copper and its alloys have characteristics that determine the advisabil architectural grillwork, radiator cores.
of their use; among these may be mentioned: electrical and thermal C
ductivity, resistance to corrosion, malleability and formability, ductil
strength, excellent machinability (especially with lead added), nonme 2.17 ALUMINUM ALLOYS. The lighter alloys are especially adap-
tic, pleasing finish, ease of being plated, and castability (sand castin ted for use where it is desired to reduce the inertia forces of moving parts
permanent-mold castings, die castings, and others). ‘I and where, in general, reduced weight is an inherent advantage, as in air-
Quenching does not harden copper alloys (except beryllium cop Plane construction and in some parts of trucks, trains, and other vehicles.
so the usual manner of increasing strength and hardness is by cold w Other characteristics of aluminum alloys that suggest their use include:
Strip and sheet are cold worked by rolling; rods, bars, shapes are ©@
high electrical and thermal conductivity; resistance to some corrosive
worked by drawing through a die. The percentage reduction of area effects (imparted by a film of oxide that forms on the surface); ease of
64 MATERIALS AND THEIR PROPERTIE 9.19] TITANIUM 6s

i rkin (hot and cold), and joining by most manu characteristics of this metal. The relative weights of steel, aluminum, and
;
a a cd niga mechanical properties in certain alloys. \ a magnesium, in that order are: 1, 0.35, 0.23. Other characteristics of mag-
including structural shapes. eee nesium include: nonsparking and nonmagnetic, good machinability, and
shapes are obtairmmable,
ed and are briefly d low modulus of elasticity. Because of its light weight, it is often found in
Temper desigr™ations have been standardiz
by 0, as 2014 ortable devices, pneumatic tools, sewing machines, typewriters; as parts
follows. The softest (annealed) temper is designated
ted, as 360-F , 3003- F. The symbo : where light weight is important, as in aircraft for blowers, nose pieces,
symbol F designates as-fabrica
wed by two digits, the seco housings, wheels, levers, brackets, etc.; and in accelerating parts where it
nates strain (work) hardening and is follo
8 being the maximum p js desired to reduce the inertia forces. Magnesium is cast as sand cast-
digit indicates the= amount of cold work, ings,!2:1-2-221 permanent mold and die castings; it is a good material for
(4 is half of 8). The
temper; H14 is storain hardened only, half hard
annealing (to reduce h extrusion.
H2x indicates str=ain hardening plus partial
hardening and then The system of designating magnesium alloys is defined briefly as follows.
a little); the symambol H3x indicates strain
sts of low-temperature The first two letters indicate the principal alloying elements (as AZ = alu-
(used when magnesium is present, and consi
longer at ambient tempe minum and zinc); the second part indicates the rounded-off amounts of
to hasten a transfeormation that takes much
T indicates oth
thus H34 is half “hard and stabilized. The symbol
these elements (as AZ61 ~ 6% Al and 1% Zn); the third part is a letter
solution heat treated, col designating the order in which the compositions became standard (as
ment: T2, annealed cast products; T3,
state; T4, solution heat treated and naturalh
naturally aged to =a stable
AZ61A is the first of this principal composition to be standardized); the
T6, solution heat tr last part indicates the temper, for which the symbols defined for aluminum
to a stable state;= TS, artificially aged only;
artificially aged; 17, solution heat treated and stabilized ; Ts apply (as AZ61A-T4 means that the alloy has been solution heat treated).
aged; T9, solution heat trea There are already a large number of more or less standard alloys. For the
heat treated, coléd worked, artificially
for some aluminum ones in Table AT 3,
ficially aged, therm cold worked. Typical uses
the principal alloying AZ61A: good for extrusions and press forgings (forging is generally done at
as follows (the fixrst digit indicates by code , slow speed).
number below).
given in parentheses after the identification AZ80A: also used for forgings and extrusions; can be heat treated.
3003 (Mn):—fmormerly 3S: vessels, tanks, tubing, cooking uten
Two characteristics should be noted particularly: magnesium alloys
ical equipment; im ee where good formability, weldability, and
are highly notch sensitive (important when a part changes section and is
i ired.
14S: aircraft fittings, truck frames, heavy duty subjected to a varying load) and, contrary to usual expectation, the com-
: ey
ss, good formability, low pressive yield strength of the wrought form is less than the tensile yield
in general, for higzh strength, high hardne
2024 (Cu)— foarm erly 24S: aircra ft struct ures, truck wheels, screw strength, each taken at an offset of 0.2%. The following comparisons of the
oie
products, rivets, hardware, miscellaneous structures. Strength/mass and stiffness/mass ratios will prove interesting; p lb./in.® is
aircra ft landi ng mats, canoe s
6061 (Mg andl Si)—formerly 61S: the density. In drawing conclusions from these index numbers, keep in
transportation equipment;
marine applicatiomms, piping, welded parts, mind the particular alloys being compared.
ty, weldab ility, and resistance to corrosion. —
good strength, foxrmabili
, aircraft struc’
1075 (Zn)—foermerly 75S: structural applications
cular. STRENGTH AND STIFFNESS COMPARISONS
360: thin-wall and intricate castings; has excellent castability and
Su(ksi) Ex 10-® Su(ksi) Ex10-®
ns, liquid-cooled cyli :
a ae hal ee bodies, air-compressor pisto
ous housi ngs; good castab ility, weldability, an MATERIAL p p MATERIAL Pp Pp
crankcases, and “vari
Magnesium (AZ61 A-F Gray iron, ASTM 40.. 150 61
es and
SE ea trakes a wide variety of pleasing finish extruded) es -» 693 100 Stainless 303, annealed 314 98
protect the base m
cluding anodize=d surfaces that additionally or slab) and su !
Aluminum (2024-T4) .. 694 108 Aluminum bronze
sion. Aluminum bonded to steel (steel ingot Steel C1020, annealed .. 200 105 (B148, cast) 290 62
corro
owt as steel with an integrated layer of aluminw Yellow brass, } hard .. 200 49
rolled comes Steel 9255, OQT 1000... 635 105 Titanium (B265, hard.) 1060 94
dipping.
aluminized steel; also obtained by hot

18 MAGNESIUM ALLOYS. i
Since i
magnesium alloys are | y TITANIUM. Since titanium is expensive, it is not used except
$I
a eaiedl as heavy as aluminum, lightness is one of the most here its properties are important—in particular in extreme-temperature
 §2.21] HIGH-TEMPERATURE SERVICE 67
The zinc alloy in Table AT 3, which has the trade name Zemak-5, may
be used for either die castings or sand castings for such articles as automo-
% of Production

tive parts, building hardware, padlocks, toys, and novelties.

2.21 HIGH-TEMPERATURE SERVICE. Ordinary steel begins to

FIGURE 2.8 Yield Strength 0 lose strength (and elasticity) significantly at about 600-700°F, which was
Ti 6 Al 4 V, annealed. After Ref. ( DL formerly an approximate widespread boundary condition. However,
modern requirements for much higher operating temperatures in petroleum
Yield Strength, ksi refining, chemical processing industries, steam power plants, gas turbines,
and now rocket engines and objects travelling at supersonic speeds in the
a e5 atmosphere, lead to the development of materials that retain significant
situations where strength is needed, especially for aeronautical strength at higher temperature levels (and also that have good corrosion
a oo am
It is also gradually finding other uses ; for example, for Fees resistance). The most advanced alloys in this respect are called superalloys
handling 60% nitric acid. Where stainless steel needed rep oe a
at “ or superstrength alloys; these alloys are some combination of nickel,
6-8 months, titanium paid for itself in the first year and is cobalt, chromium, iron, molybdenum, tungsten, columbium, titanium,
aa s ee
for 10 years. The 6 Al 4 V alloy given in Table AT 3 has and aluminum, but never containing all of these. With the advent of space
. a
above 600°F and is used for aircraft turbine blades and os exploration, the search for and development of superalloys has been in-
a eee is _ |
fittings, sheet metal, and extrusions. Figure 2.8 shows tensified and new knowledge is available every day.
Ti 6 Al 4 V with the + 20 sp q
of yield strengths of annealed Any material begins to lose strength rapidly at some temperature; as the
dicated (see § 3.9).
temperature increases, the deformations cease to be elastic and become
more and more plastic. When plastic deformations are involved, the cri-
aap There terion for design at a particular operating temperature is the creep strength
2.20 TIN, AND ore ip cmel a
LEAD,
: or the rupture stress at a specified length of time. Creep results in a per-
alloys, but we shall comment on 0
AT 3 has a lead base and manent deformation, and for a given material and stress, this deformation
ai ee 323-467, grade 8, in Table in a particular time is greater for a higher than for a lower temperatur
suitable bearing material for light and moderat e service in various e.
When constant dimensions must be maintained, measurable creep cannot
a of course be tolerated; but there are many cases where small permanent
a tin babbitt B23-49, grade 1, has a tin base andisa snes
a a changes have negligible effect. Creep strengths at a particular temperatu
bearing material, also used for die-castings. Tin base alloys are variously defined, and much of the data is so new that these
re
c
easy to bond, have excellent anti-seizure qualities, and they resist strengths
k are not always presented in a consistent and orderly manner. A common
han lead base babbitts.
of nickel, molybd enum, and nou definition of creep strength is that stress that produces a creep rate of
ey B is an expensive alloy
ee 0.0001 °% per hr. This is mathematically the same as 1% in 10,000 hr., but
that is very useful in the chemical industry because it at it does not mean that the material can maintain this rate for 10,000
boi o
rosion admirably—for example, hydrochloric acid up to its Without rupture unless it is known from other data that it can.
hr.
‘7
phosphoric acid, sulfuric acid up to 50% concentration, ce the time it takes to conduct an experiment (100,000 hr. =
Because of
and other corroding substances. Since it holds its strengt : a 11.5 yr.), test
‘ y data are most frequently found for 1000 hr. and less. If one must
temperature (e.g., 60 ksi at 1500°F), it classifies as a supera design
for 10-20 yr. or longer life with a material for which test data at said life
other grades of hastello y, some contain ing co
There are several 1S unavailable, the only thing to do is to extrapolate courageous
m. ; a ly.*
, The dangers of extrapolation are shown by the typical behavior
a is primarily an alloy of nickel and copper (67 Ni, 20 curve
of Fig. 2.9(a), Upon application of a creep producing load (some
different “‘kinds’’ of monel having small amounts of other a ie materials
a ‘reep measurably at ordinary and low temperatures—as you know
monel is used where a combination of high strength and goo from
It is also used for nonmagn etic aircraft pa aaa
resistance is needed. ~ Do not confuse a courageous extrapolation, which is done with as much knowledge
*
. . .
as
pump rods, springs, valve stems, shafts. Despite monel’s be o 1S available (unless you have this knowledge now, you will not
have time to obtain it in
ao course), with a foolhardy extrapolation, which may be a thoughtless jump to a
corrosion, its fatigue strength in adverse environments 1s sharply re Clusion, con-
about 20 ksi at 108 cycles in fresh water.
eit sl
 tT NI, io UF,
A 4 Mo,4 W,
37,1 Al

Modulus of Elasticity Ex 10-* psi

§2.22] LOW-TEMPERATURE PROPERTIES 69

$;
the designer should be in touch with the manufacturers for the best ma-
terial and its best treatment. Because so much specialized knowledge is

mR
T
17-7 Mo required for intelligent design for creep conditions, we shall leave it largely
PH Stainless for another study.
ao (15 Cr,7 Ni,2.25 Mo,
1.15 Al)
Permanent Strain,
Initial Plastic Strain Load Released at A High-Alloy Steel (5 Cr-Mo-V) 2.22 LOW-TEMPERATURE PROPERTIES. Modern science and
es 10.800 eeea
Elastic’ Strain; a
engineering are going to extremes in both directions of temperature, what
Time F
with the very low temperatures involved in cryogenic studies, in shipping
(a) (b)
and using liquified gases, etc. Although the standard test tensile strength,
FIGURE2.9 Strain-time Curve Illustrating General Characteristics of Creep and Decreaj hardness, and modulus of elasticity of steels and other metals increases
of E with Temperature. The steels in (b) are for higher-temperature service; A is a nick as the temperature decreases, ductility and toughness decrease. Over the
base alloy. years, there have been many mysterious (at the time) brittle failures of
tanks, bridges, and other structures, but the significance of the loss of
personal experience with wax), there is, as Finnie and Heller'?-?% poi toughness, as measured by an Izod or Charpy test, did not become general
out, first an initial strain, Fig. 2.9(a), partly plastic and partly elastic; th engineering knowledge until World War II experiences of vessels cracking
the initial rate of increase of strain is generally high for a while, aff in two. Sometimes the weather was heavy, sometimes not, but invariably
which there may be a period during which the creep proceeds at a mor -¢C the air and water were cold (say less than 40°F). Investigation showed that
less constant rate, called secondary creep. If the load is high enough and the low-carbon steel had a transition temperature (also true of low alloy
the time is long enough, there is a tertiary creep at an accelerating ra steels), below which it failed by a (tensile) brittle failure (typically sudden
leading eventually to fracture. Thus, a steady rate of creep as between and complete) and above which the failure is ductile (typically a shear
and C is not necessarily a safe basis of design. The time of life must bei rupture). Usually, the fracture starts at a point of high stress, a point of
cluded. Not only the strength falls off, but the modulus of elasticity d
creases as the temperature goes up, as seen in Fig. 2.9(b). FIGURE 2.10 Creep Strength and Rupture Stress. Solid curves give creep strength for a
A: proper design stress, as usual, depends on the circumstances. TI rate of 0.1% in 1000 hr. at each temperature. Dotted curves give the stress at rupture at each
life of a missile may be measured in minutes; so, accumulating experim temperature. Casting alloy HT is a nominal 35% Ni, 15% Cr alloy of iron; Incoloy is
data to fit this case is not too time consuming. On the other hand, a sui principally a Ni (32%)-Cr(20.5%) alloy of iron; Ni-Resist D2 is a ductile iron casting,
20% Ni, 2% Cr, treated with Mg to produce the spheroidal graphite; Greek Ascoloy, a
life of a steam power plant would be nearer 200,000 hr., in which e Steel alloy with 13% Cr, 2% Ni, 3% W. Data from Refs. (2.1, 2.3, 2.31).
some extrapolation of data will be necessary. (There are of course son
long-time data available.) Where relatively frequent inspection is ce
one can take bigger chances. Except where life or valuable propert 1 Greek Ascoloy
threatened, the basis of decision as to what material to use is largel:

$
economic one, and costs seem amazingly high to one accustomed only t
ordinary steel; for example, the superalloy M 252 (Ni, Cr, Co, Mo, T
etc.) costs about 170 times more than carbon steel plate. Frequent comp!
cating factors include repeated loading, thermal shock (quick tempera
changes), cycling thermal stresses (especially detrimental for brittle m

NAB OO
—
terials). For their suggestive value: the ASME Code allows a design stré

Stress, ksi

-
w
should be obtained at a particular operating temperature if the desif
stress is two-thirds of the 1% in 10,000 hr. creep. Some typical creep dat

nN
are pictured in Fig. 2.10. Since so many of the superalloys are patentee
 [Ch. 2 §2.23] PLASTICS 7
70 MATERIALS AND THEIR PROPERTIES

stress concentration (Chapter 4). The fracture for complete brittle failure o}
related to a material whose properties are well known, the purpose may be
steel has a bright faceted appearance; the ductile fracture has a silky fibrous well served. The exploration into the very-low-temperature region is so
appearance. The transition, except in pure iron, does not occur instantane: recent that engineers are still looking for suitably simple ways of specifying
the desired material properties. In general, one does not use a material
ously at a particular temperature, but rather over a temperature range wit
which the failures are partly brittle and partly ductile (Fig. 2.11). In gene better than required for the service; for example, 2.5°% Ni steel is good to
— 15°F, 3.5% Ni steel to —150°F, 5% Ni steel to —200°F, and 9% Ni
it is dangerous to use a material below its transition temperature because
it has lost so much of its capacity to absorb energy without rupture. A steel to —320°F. :
In general, alloys with high-nickel content retain toughness to quite
few materials show increasing impact strength with decreasing temperature,
notably nickel, copper, and aluminum. Nickel is the most effective alloy in low temperatures. Nickel alloy benefits cast steel similarly, especially if it is
of steel. Peening (§ 4.28) lowers the transition tem quenched and tempered. Stainless steel with nickel (as the 18-8 variety) has
increasing toughness
E good low-temperature properties, but the non-nickel bearing stainless is not
perature for the steels.
The impact test, Charpy or Izod, is currently most used as a measure
sO fe Copper and copper alloys dre also used at temperatures below
of the low-temperature suitability of steel, but we have no way of us
Work hardening steels (§ 2.9) at low temperatures often produces eco-
the data quantitatively in design. Comparing the two curves given for
nomic benefits. For example, 301 stainless steel vessels have had their
9°% Ni steel in Fig. 2.11, we observe that the V-notch Charpy test m
ultimate tensile strength increased to 260 ksi (contrast with Table AT 4) by
indicate the transition more distinctly than the keyhole Charpy. Y.
realize of course that the measured energies from a group of test sp
stretching them 13 % with cold nitrogen inside at temperatures to —300°F.
;
mens do not fall on a single curve as shown in Fig. 2.11, but within so
band. These curves are therefore merely typical and should not be u
2.23 PLASTICS. Plastics have come into use not only as decorative
in actual design unless it is known that they apply. Prior treatment
ficantly affects the transition band and the Charpy energy (different t and unstressed parts but also as load-carrying members. !2-19.2-20.2.21,2,27,2.30]
They may be divided into two main classes—thermosetting, which undergo
of specimens do not give correlated quantitative data). The 9% Nia
chemical change and harden on being heated, usually under pressure; and
and stainless steel 304L shown have been used successfully for storing
thermoplastic, which soften as the temperature rises and remain soft in the
transporting liquid N, (about —320°F); aluminum and titanium alloy
heated state. The thermoplastics may be reshaped on heating, whereas the
(to about —300°F) are used for similar service.
It is not uncommon for the specifications to state a minimum imp
thermosetting variety cannot be.
The phenolics, Table AT 5, constituting one of the largest, least expen-
strength (as 15 ft-lb., Charpy keyhole) at a particular temperature,
Sive, and most useful groups, are available in forms for molding and casting
experience suggests that this sort of specification does not necessa
when the magnitude of the energy is and in laminated forms. They are made from formaldehyde and phenolic
eliminate brittle failure. However,
bodies such as phenols (carbolic acid), cresols, or cresylic acid. Laminated
Phenolic materials machine readily, have good wearing properties, and are
To 120 ft-lb. FIGURE 2.11 Low-Temperature Toughn
used for gears, gaskets, seals, compressor-valve plates, tubing, bearings
The symbol V indicates a V-notch specimel
the symbol K, a keyhole notch. The 9% | insulators, etc. Grade X has a paper base, comes in sheet, tube, and molded
9*Ni,V
contains about 0.1% C; it is double norm forms, and is used primarily for mechanical applications. Grade XX has
and stress relieved; now available in a paper base, comes in sheet, tube, rod, and molded forms; it has good
structural shapes. The stainless steel 304L machinability, greater moisture resistance than Grade X, and is suitable
same as 304 except 0.03% C vs. 0.08% C) for electrical (insulating) applications. Grade C has a fabric base, comes in
not have a horizontal curve for the V-n'
specimen. The AISI 347 is cold drawn, + Sheets, tubes, and rods, and is used for gears. Grade A has an asbestos base
The 0.2% steel was normalized. The no Comes in sheets, tubes, and rods; it is more resistant to heat than are tidue
iron shown has higher energy properties th Previously mentioned. There are several other groups of thermosetting
many others. The gray iron or cast iron (C.I) Plastics, including the urea-formaldehyde, the melamine-formaldehyde, and
would qualify as approximately ASTM 30. | the polyester plastics.
Nodular Iron
‘ Many thermoplastics are widely used in unloaded forms, such as
andles, knobs, containers, grills, covers. Polyethylene, a relatively inex-
Nn thermoplastic, is used widely for bottles and other containers.
Gray Iron ylon (polyamide), while relatively expensive, has load carrying capacity
ai 1 i 4

—300 —200 -—100 0 100 200 300

72 MATERIALS AND THEIR PROPERTIES [ h.

up to 200-250°F; it has a low coefficient of friction, 0.05 to 0.20!?-27)

good wear resistance—hence, its use for bearings, dry or lubricated; it
machined or molded. Nylon is used for gears (quietness) and balls (valve
lightly loaded bearings for quietness). A trade name for nylon resinsfi
molding is Zytel.®
Teflon® (polytetrafluoroethylene—also called tetrafluoroethylene FIGURE 2,12 Forged Crankshaft. A
has uses similar to those of nylon, but is more expensive and would t drop forging, made in impression dies, of a
fore be used in more exacting situations; it is not molded in the conventi Jarge internal-combustion-engine crank-
manner, but more as in powder metallurgy; it makes a tough bearing shaft, ready for machining. (Courtesy Drop
Forging Association, Cleveland.)
a dry coefficient of friction of 0.1-0.2-2". Teflon, a very inert pl
solves many obdurate problems of gaskets and seals because of its a
to tolerate a continuous temperature of 500°F, higher for short pe
and because it retains some flexibility to a low temperature of —
A new relative of the original Teflon, fluorinated ethylene-propylene,
signated FEP, has a distinct advantage in that it can be processed in a mi FIGURE2.13 Upset Forging. The original
conventional manner. Engineering data on these Teflons are four bar is shown in (a); (b) and (c) are the two
Ref. (2.33, 2.35). Thin films of Teflon provide excellent corrosion resist steps in the forging; (d) is the finished
machined part. (Courtesy M. H. Harper Co.,
and make the surface nonsticking (as in cooking utensils and many i Morton Grove, III.)
trial uses). One of its recent uses is for piston rings in air compres
eliminating the need of oil lubricant, the accompanying danger of explos
and in some cases, product contamination by lubricant. ae forging is a hammering process, while upset and press forging are done
There are a number of rubbers or elastomers (buna-S, butyl, neopr by exerting pressures adequate to cause plastic metal flow into the required
buna-N, natural rubber) of engineering interest, but space does not pi shapes. In forgings shaped in dies, several steps are required to produce the
a discussion here. All problems involving plastics and rubbers show final form; for instance, the forging of a 3-ft. connecting rod for a Diesel
solved in the presence of more information than contained herein, engine is done in seven steps, including the final step of trimming off the
is intended to be suggestive only. A “‘newer and better’’ plastic mi flash. To facilitate removing the forging from the die, the sides of the die
available tomorrow. are tapered. This taper is called the draft and is about 7° for outside, and
The mechanical properties of these materials not only have a no 10° for inside, surfaces, although more or less draft may be used. Fillets
manufacturing variation, but they also vary with temperature and moi and radii joining different parts of a forging should be as large as possible
content; there is dimensional change with time; the modulus of elas’ (see Chapter 4), because small fillets increase the tendency toward forging
is not a constant; many will creep measurably under load at ambi defects, such as unfilled structure, and they increase die wear. Corner radii
temperature. Such characteristics, if not accounted for in the design,m Should also be as large as possible. Keep rib heights as small as possible
lead to an unnecessary failure. os i and rib sections as thick as possible because of the difficulty of forcing the
Plastic metal into thin, deep pockets. Metal may be distributed in a forging
In a manner that will take advantage of the “grain flow,’ inasmuch as
2.24 DESIGN HINTS. The basic dimensions of a part with kno Wrought metals are stronger under impact in the direction of the “grain”
loads on it are computed by the designer, but these dimensions 4 (direction of the rolling) than in the transverse direction. This factor may
the details of other dimensions depend to some degree on the meth need to be considered in the design phase of the job.
manufacture. Is the part machined from milled or extruded stock; i If the part is to be cast, there are many alternative materials some one
forged or pressed; is it cast; is it welded; is it sintered? of which may or may not be clearly indicated by the nature of the part.
If it is to be forged,!?-24) it may be a smith forging done with flat¢ There are cast iron (Fig. 2.14), malleable iron (Fig. 2.15), cast steel (Fig.
(manually or by machine); it may be a drop forging, Fig. 2.12, in which’ 2.17), nodular iron, nonferrous metals; any one of which may be sand
part is formed between shaped dies attached to the anvil cap and the ra Castings, permanent mold castings, die castings (Fig. 2.18), centrifugal
it may be an upset forging, Fig. 2.13, done in a machine that can forma Castings, or precision investment castings (a lost-wax process).
piece into a desired shape by pushing the piece into dies; or it may Perhaps the most important policy in the design of castings is to make
press forging, which squeezes the metal into shape between dies. the various sections of the casting as nearly as possible the same thickness.

73
 ar ee h

FIGURE 2.14 (Above) Cast-iron Crankshaft. A machined cast-iron crankshaft, 11 ft, (a) (b) (c) (a)
4 in. long, 1750 Ib., for a 5-cylinder, 2000-hp. Diesel engine. Cast of iron from an electric.
FIGURE 2.19 Joining Section Changes—Castings. Unless an important useful purpose
furnace, the shaft was annealed at a low temperature before machining in order to reliey
is served (unlikely), changing sections with sharp corners as in (a) should be avoided in all
stresses and avoid distortion after machining, and to improve the mechanical propert
design no matter what the manufacturing method may be. A radius as suggested by the
Test coupons from such shafts suggest average properties as follows: s, =68 ksi, E=23,0
proportions in (b) is acceptable; the proportions in (c) are good; and the plan in (d) is
ksi, G = 9200 ksi, BHN = 300. This company casts Diesel crankshafts up to 22 ft. I
better yet. It happens that these changes from (a) to (d) are progressively better for a part
(Courtesy Pacific Car and Foundry Co., Renton, Wash.)
made in any manner when the load is repeated (Chapter 4). The best alternative for castings
is not to change the section thickness at all (not always possible to abide by this). (After
Steel Castings Handbook'?-*°)),

FIGURE 2.20 Hot Spot at Joined Sections. The

extra mass of metal at the corner, dimension ab, means
slower cooling, possibly resulting in shrinkage cavities
or cracking. Instead of the corner at 5, the outside
radius R is preferably such that the thickness ac to
the dotted curve is somewhat less than A (not always
feasible on a stress basis). The radius r should be
about r = A or more. Another solution may be to
core a hole near the center of mass at the corner.
(After Steel Castings Handbook'?-**}),

FIGURE 2.21 Casting a ‘‘Cross’’ Section. If sec-

de
FIGURE 2.15 (Above left) Malleable Iron Swivel Hitch. Typical of many small pai
suitably made of malleable iron. (Courtesy Eastern Malleable Iron Co., Wilmington, D tions cross, as shown dotted at A, a serious hot spot
rZ1 in.
will occur at the junction. If one part of the ‘‘cross’’
is offset, as at B, the foundryman can use external chills
FIGURE 2.16 (Above right) Welded Swivel Hitch. When compared with Fig. 2.15, t iis
to prevent cavities. The minimum recommended offset
one illustrates the different ways in which a particular part may be made. (Courtesy Easter
is suggested by the dimensions shown. The fillet radius
Malleable Iron Co., Wilmington, Del.) ’
r should be between 3 and 1 in., regardless of size of the
“‘cross’’. (After Steel Castings Handbook'?-1°)), =

_
FIGURE 2.17 (Below left) Cast Steel Gear Box. Made of ASTM A 27-50, T, class 70-3¢
An example of a large complicated casting, weight about 1900 Ib. (Courtesy Steel Found
Society, Cleveland.) FIGURE 2.22 Wave Construction. The design at
(a) resulted in cracked spokes. Curved spokes as in
FIGURE 2.18 (Below right) Die Casting—Carburetor. Excellent example of comp! either the plan view (b) or end view (c) will flex under
coring and shapes obtainable with die castings. A four-jet carburetor (two carburetors i the cooling stresses, probably eliminating distortion
one), zinc bowl. Sharp corners, inside and outside, should be avoided in die castin and cracking. (From Fundamentals of Steel Casting
(Courtesy General Motors). Design by G. W. Briggs).

The tendency toward hot spots, and therefore internal stresses due to un-
€qual cooling rates, is thus reduced. The general principle is to avoid
extra mass concentrations. This can sometimes be done by using cores.
See also Figs. 2.19, 2.20, 2.21, and 2.22, and Ref. (2.16) for more infor-
mation.

75
76 MATERIALS AND THEIR PROPERTIES [Ch. §2.26] CLOSURE 7

In one situation or plant, a certain part may be fabricated by weldin; properties will be taken up later—in particular, the endurance strength
Fig. 2.16, whereas in another situation or plant, a casting, Fig. 2.15, m (Chapter 4). No doubt, you have already noted that a particular material
be a better answer. “Better” generally has to do with cost if the servic has many different “‘strengths”’; it will serve you well to master the various
conditions are met satisfactorily,the better one being the cheapest. The ur “strength” concepts as you come to them. We must proceed with our
cost will be much affected by the quantity. For one quantity, a certa study assuming that many environmental factors involving corrosion,
manufacturing process may be cheaper, but another process may be cheap embrittlement, etc. will not be overlooked; for example, hydrogen embrittle-
for another quantity. Of course, the process may be determined alme ment, sulfide cracking, stress-corrosion cracking, nitrogen embrittlement.
entirely by the part—it is difficult to think of a cheaper way to produce There are many decisions that a designer must make, but perhaps the
complicated part such as a carburetor, Fig. 2.18, if many are to be mag two crucial ones are the choice of material and the choice of design stresses.
than by die casting. There are always choices and this is where the art (engineering) of design
enters. After the decisions are made, many people (and computers) will be
able to follow an outlined routine using the mechanics-of-materials
2.25 MISCELLANEOUS. Complicated shapes and parts can equations (science) to produce certain answers. It is suggested therefore that
made from sintered metals (powder metallurgy). Their properties dep in all your problem work and in text examples and discussions you think
part on the metals used, which include iron, copper, lead, tin, silver, critically of the design stresses, whether they are specified or not, and of
sten, molybdenum (and often carbon) in various combinations. The mi the material, which is frequently specified. Since the engineer must first
are in powder form, and mixed in the desired proportions (as 88 % Cu, 1€ of all be able to solve the science side of design, our principal attention
Sn, plus iron and graphite for bearings); the bonding of the adjacents is directed to stress analysis and design stresses, with concurrent discussion
faces is accomplished by prolonged heating (sintering) at a tempe of engineering factors.
below the melting point. An additional process sometimes used,
infiltration, is to melt a layer of another metal, such as lead or copper, i
the pores of the sintered material. Since the manufacture of such p
highly specialized, a manufacturer should be consulted. Some sta
items, as bearings, can be bought off the shelf.
One never knows when a startling new engineering material
announced. An example is Pyroceram, a group of glass products
Corning Glass Works that have unusually good mechanical pro
at high temperature. This material is already in use in a number of con:
products and it has many industrial uses. A recent announcemen
cribes maraging nickel steels*, which attain yield strengths of as m
250 ksi without quenching. There are compositions of 18% Ni, 20%
or 25% Ni, together with several other alloys, all with low carbon (0.03
The 18% Ni is treated as follows: transformation to martensite
below 310°F, but the low carbon content results in soft and tough
transformed metal is held at 900°F for 3 hr. (maraging); it may be an
and cold worked before maraging for improved properties. T
properties: Ro = 50, sy = 250, su = 260, E = 27 x 10*% ksi, ¢€ =
Charpy = 16 ft-lb. at—420°F. Compare with 17-7 PH, § 2.15(b). Ar
development, recently announced by Du Pont, is a nickel alloy (about 9¢
Ni, 2% thorium oxide), called TD nickel, with unusually good cre
strength; retains useful strength to above 1800°F.

2.26 CLOSURE. There are a number of properties of engineer

materials that have not been mentioned at all in this chapter. Some of th
* International Nickel Co.
 3. TOLERANCES AND
ALLOWANCES

3.1 INTRODUCTION. It is easy for the designer to specify a certain

dimension, say 24 in., but for the workman or machine to turn it out exactly
2.50000 in. is quite another matter. In general, the closer the dimension
must be to precisely 2.5 in., which would be called the nominal size, the
more costly will be the manufacture (Fig. 3.2, p. 84).
If the designer does not know something about the natural variability
of a process, he is not in a position to specify the best tolerance. We have
seen something of the ways in which the strength ofa material of the same
specification varies. Similarly, a 2.5-in. part leaves the machine with a
dimension greater or less than 2.5 in. by some amount which may be any-
thing from less than a millionth of an inch to more than several hundredths.
How far off will it be? This question can be answered statistically, if
enough information about the process is known. Without statistics, a
satisfying answer cannot be given. For this reason, machine designers need
to interest themselves in statistical analysis applied to manufacturing
(§§ 3.9-3.13).

3.2. TOLERANCE. First, let us define tolerance, which is the stated

Permissible variation of the size of a dimension. If a particular dimension is
to fit inside of, or outside of, another dimension, as a pin in a hole, this
permissible variation or tolerance is important. If no fit is involved, as in
the external diameter of some rods—for example, dimension d, Fig. 1.17—
it would be unwise, unnecessary, and uneconomical to insist that the final

79
80 TOLERANCES AND ALLOWANCES [Ch.3 §3.4] FITS 8l

size be very close to the nominal size. Thus, if there is no fit, a liberal RC8 and RC9; loose-running fits. For use with commercial cold-rolled
tolerance of the order of +0.010 in. or + in. in machining work co shafts and tubing.
be permitted. Such tolerances are frequently cared for by a blanket note The same standard B 4.1-1955 also gives tables of limit dimensions for
on the drawing; for example, “Tolerances not specified to be +0.010 in” clearance location fits, transitional location fits (accuracy of location is
See § 3.3 et seq. for information on fitting dimensions. The tolerance may be important but can tolerate either small interference or small clearance),
(a) Bilateral, where the size of a part is permitted to be either larger o interference location fits (accuracy of location of prime importance), and
smaller than the given dimension, or 4 other information; those concerned with actual design should have a copy
(b) Unilateral, where the size of a part may be larger only, or smalle of the standard.
only, than the given dimension. For example, § The ASA fits are based on the basic hole system; the nominal size is
the minimum hole diameter. In the basic hole system, the tolerance on the
—9 £1062
1
£010; © 1.062
+0.010 hole is always positive (with zero negative tolerance), measured from the
16 32 —0.000 ; basic size, and the tolerance on the shaft is all negative for working fits.
[BILATERAL TOLERANCES] [UNILATERAL TOLERANCES] The basic hole system has an advantage over the basic shaft system in that
a

Unilateral tolerances are generally used on those dimensions involved in; the hole produced by a standard reamer may be made the minimum hole
fit such as a pin in a hole. See Fig. 3.1. size. Also, there is the matter of inspecting with go and no-go gages.
Generally, the mating parts are not so sensitive to clearance but that a
plant can adhere to some preferred tolerances on the basic hole system,
3.3. ALLOWANCE. For fitting dimensions (a fit of mating pa thereby materially reducing the number of gages needed for inspection.
involved), the tolerance, which depends in part upon service needs, she In short, the engineer should not scatter tolerances for a particular size of
be chosen with some knowledge of the natural spread of the processes t hole all over the spectrum, but should conform to company standards as
produce the mating dimensions. If a pin is to turn freely in a hole, the far as possible. Depart from such choices only for a confirmed engineering
must be somewhat smaller than the hole. In manufacture for interch reason.
able assembly, the difference in size is cared for by the allowance, whi Table 3.1 (pp. 82, 83), from the ASA Standard B 4.1-1955, defines the
for running fits, is the minimum specified difference between the dimensit details of the various classes of fits. The columns headed “‘Hole”’ are the
of the pin (male part) and the hole (female part). It is not the sar tolerance limits on the hole; the columns headed “Shaft’’ give the same for
general as the actual minimum difference. the shaft (for example, a 0.1-in. shaft of class RC 1 has a maximum diameter
When the pin is larger than the hole, the allowance is as befo of 0.1 — 0.0001 = 0.0999 in., a minimum diameter of 0.1 — 0.00025 =
difference in sizes for the tightest fit, a difference also called the interference 0.09975 in.). The allowance, as defined above, is the numerically smaller
of metali (or negative allowance). number in the “Shaft”? column (for example, the allowance for a 0.1-in.
nominal diameter, RC 1 fit, is 0.0001 in.). There is no rule that says the
values in the standard must be used, but most of the quoted values are part
3.4. FITS. There are several systems of tolerances and allowe of an ABC (American, British, Canadian) agreement. Think of a standard
Many companies have drawn up standards to be used in their own praci as what appeared to be a good engineering guide at the time it was formu-
In this country, the American Standards Association, ASA Standar lated.
B4.1-1955,-1) has classified running and sliding fits as follows: ' The standard B 4.1-1955 gives a list of over 40 preferred tolerances and
RC 1; close-sliding fits. For accurate location of fitting parts without percep! allowances; where it is apropos, we may consider the following ones as
ible play. ' our “company preferred”’ values:
RC 2; sliding fits. These fits move and turn easily but are not intended to ru
freely. With a small temperature change in the larger sizes, this fit may
0.0001 0.0006 0.0016 0.0060 0.016
RC 3; precision-running fits. For precision assemblies operating at low spec
0.0002 0.0008 0.0020 0.0080 0.020
light load, and small temperature change. ,
RC 4; close-running fits. For cases of accurate location and minimum pia}
0.0003 0.0010 0.0025 0.010 0.025
but for moderate speed, journal pressure, and temperature rise. ¥ 0.0004 0.0012 0.0040 0.012 0.030
RC 5 and RC 6; medium-running fits. Suitable for higher running speed an
heavy journal pressure. “Practical’’ tolerances, appropriate to particular manufacturing
RC 7; free-running fits. Suitable for large temperature variations and whet methods and materials, are scattered through the literature, where one
wider tolerances are permissible. Must go for detail [e.g., Refs. (2.1, 2.14, 2.15, 2.16, 2.17, 2.20, 2.23, 3.11,
 S36 Saas wom SO Ie a SENSE ee eae ee

Extracted with permission (ASME) from Table 1. ASA B 4.1-1955. The numbers given are the standard
limits in thousandths of an inch; multiply by 10 3, Limits for hole and shaft are applied algebraically bs
the basic size to obtain the limits of size of the parts. (The Standard goes to 200-in. size.)
NOMINAL CLAss RC 5 CLass RC 6 cLass RC 7 CLAss RC 8 cLass RC 9
SIZE RANGE, cLass RC 1 CLASs RC 2 cLass RC 3 CLAss RC 4
INCHES i Hole | Shaft | Hole | Shaft Hole Shaft Hole Shaft Hole| Shaft
Over To Hole Shaft Hole Shaft Hole Shaft Hole Shaft
+0.4 | —0.6 | +06 | —0.6 | +1.0 | —1.0 | +1.6 | —2.5 | +2.5 | —4.0
0.04— 0.12 +0.2 | —0.1 +0.25 | —0.1 +0.25 | —0.3 +0.4 0 —1.0 0 —1.2 0 —1.6 0 —3.5 0 —5.6
0 —0.25 0 —0.3 0 —0.55 0
+0.5 | —0.8 | +0.7 | —0.8 | +1.2 | —1.2 | +1.8 | —2.8 | +3.0 | —4.5
0.12— 0.24 +0.2 | —0.15 +0.3 | —0.15 +0.3 | —0.4 +0.5 0 —1.3 0 —1.5 0 —1.9 0 —4.0 0 —6.0
0 —0.3 0 —0.35 0 —0.7 0 +0.6 | —1.0 | +0.9 | —1.0 | +1.4 | —1.6 | +2.2 | —3.0 | +3.5 | —5.0
0.24— 0.40 +0.25 | —0.2 +0.4 | —0.2 +0.4 | —0.5 +0.6 0 —1.6 0 —1.9 0 —2.5 0 —4.4 0 —7.2
0 —0.35 0 —0.45 0 —0.9 0
+0.7 | —1.2 | +1.0 | —1.2 | +16 | —2.0 | +2.8 | —3.5 | +40 | —6.0
0.40- 0.71 +0.3 | —0.25 +0.4 | —0.25 +0.4 | —0.6 +0.7 0 —1.9 0 —2.2 0 —3.0 0 —5.1 0 —8.8
0 —0.45 0 —0.55 0 —1.0 0 +0.8 | —1.6 | +1.2 | —-1.6 | +2.0 | —2.5 | +3.5 | —4.5 | +5.0 | —7.0
0.71- 1.19 +0.4 | —0.3 +0.5 | —0.3 +0.5 | —0.8 +0.8 0 —2.4 0 —2.8 0 —3.7 0 —6.5 0 —10.5
0 —0.55 0 —0.7 0 —1.3 0
+1.0 | —2.0 | +1.6 | —2.0 | +2.5 | —3.0 | +40 | —5.0 | +60 | —8.0
1.19- 1.97 +0.4 | —0.4 +0.6 | —0.4 +0.6 | —1.0 +1.0 0 —3.0 0 —3.6 0 —4.6 0 —7.5 0 —12.0
0 —0.7 0 —0.8 0 —1.6 0
+12 | —2.5 | +1.8 | —2.5 | +3.0 | —4.0 | +45 | -—60 | +7.0 | —9.0
1.97- 3.15 +0.5 | —0.4 +0.7 | —0.4 +0.7 | —1.2 +1.2 0 —3.7 0 —4.3 0 —5.8 0 —9.0 0 —13.5
0 —0.7 0 —0.9 0 —1.9 0
+1.4 | —3.0 | +2.2 | —3.0 | +3.5 | —5.0 | +5.0 | —7.0 | +9.0 | —10.0
3.15- 4.73 +0.6 | —0.5 +0.9 | —0.5 +0.9 | —1.4 +1.4 0 —4.4 0 —5.2 0 —7.2 0 —10.5 0 —15.0
0 —0.9 0 —1.1 0 —2.3 0
+1.6 | —3.5 | +2.5 |—3.5 | +40 | —6.0 | +60 | —8.0 | +10.0| —12.0
4.73- 7.09 +0.7 | —0.6 +1.0 | —0.6 +1.0 | —1.6 +1.6 0 —5.1 0 —6.0 0 —8.5 0 —12.0 0 —18.0
0 —1.1 0 —1.3 0 —2.6 0 +1.8 | —4.0 | +2.8 |—4.0 | +45 | —7.0 | +7.0 | —10.0} +12.0} —15.0
7.09— 9.85 +0.8 | —0.6 +1.2 | —0.6 +1.2 | —2.0 +1.8 0 —5.8 0 —6.8 0 —9.8 0 —14.5 0 —22.0
0 —1.2 0 —1.4 0 —3.2 0
+2.0 | —5.0 | +3.0 |—5.0 | +5.0 | —8.0 | +8.0 | —12.0 | +12.0] —18.0
9.85-12.41 +0.9 | —0.8 +1.2 | —0.8 +1.2 | —2.5 +2.0 0 —7.0 0 —8.0 0 —11.0 0 —17.0 0 —26.0
0 —1.4 0 —1.7 0 —3.7 0

demanding about this design) and use a hole tolerance of 0.006, a shaft tolerance
of 0.004 in. Then the dimensions are (basic hole, nominal size = 0.875):
screws, cold-rolled forms, structural steel shapes, tubing, keys, pulleys,
bushings, gears, etc. fall within certain specified tolerances. + 0.006; + 0.000 in.,
Hole, 0.875 _ 9 999 shaft, 0.871 — 0.004

and the drawing is dimensioned as shown in Fig. 3.1. On the other hand, if the
3.5 EXAMPLE. In the example of § 1.21, we assumed a “working fit’ for standard is followed, for which there may be good reasons, the limit dimensions
are: hole, 0.875 to 0.880 in.; pin, 0.868 to 0.8645 in.
Assuming that the connection of § 1.21 is not a fine piece of machinery requi
a “‘just-so”’ fit, we easily decide that the tolerances can be wide with no harm don
say RC9. From Table 3.1: hole tolerance = 0.005, shaft tolerance = 0.003
3.6. INTERCHANGEABILITY.®1:3-151_ Allowances and tolerances
allowance = 0.007 in. To show that it can be done, let us decide that there 1S
nothing gained in an allowance as a as 0.007, and make it 0.004 in.; also, le ;
are a practical necessity when many mating parts are to be produced. If it
were intended to make only one of the yoke connections of § 1.21, good
results would be obtained in a small plant by describing the fit to the work-
man and leaving the rest to his judgment. But if many parts are to be made,
Ky = WY KY i SN FIGURE 3.1 Methods of Dimensioning, The interchangeable manufacture, the basic ingredient of our mass-production
a T 0008 Oars method in (b) shows the /imits, the extreme per- technique, is economically essential.
0.875 6.000 0.881 missible dimensions. The top limit in the dimension
0.871 +2 0.871 should be the one the workman reaches first— For a completely interchangeable system of manufacture, a workman
Ps a 0.867 | smaller one for holes, larger one for shafts.
 TABLE 3.2 FORCE AND SHRINK FITS (INTERFERENCE FITS)
84 TOLERANCES AND ALLOWANCES [Ch.3
Extracted with permission (ASME) from Table 5, ASA B 4.1-1955. The numbers given are the standard
limits in thousandths of an inch. Limits for hole and shaft are applied algebraically to the basic size to
with a box each of the yoke ends, rod ends, and pins may expect any pil obtain the limits of size for the parts.
to fit satisfactorily with any rod or yoke. Where engineering calls for close NOMINAL
fits than can be obtairied economically with a completely interchangea nee AO cLass FN1 | cLass FN2 | cLass FN3 ,| cLass FN4. | cLass FN 5
sea oe Mier . practiced. In this instance, the manu Over To | Hole Shaft| Hole Shaft| Hole Shaft | Hole Shaft \Hole Shaft
actured parts are classified according to two or more size groups. Then, i
: ” . Res 0.04— 0.12 0.25 +0.5 0.4 +0.85 0.4 + 0.95/4+0.4 + 1.3
there are two size groups, the larger pins would be assembled with t *O 03 *9 +0.6 bai *o +07 )-0 + 09
larger holes, smaller pins with smaller holes, etc. This procedure r
. : 0.12- 0.24 0.3 +0.6}+05 +1.0 +0.5 + 1.2 {+05 + 1.7
produce closer fits more economicall y than if the same fits were obtai as +0.4.|-0 +0.7 “i +09 {-0 + 1.2
by reducing the allowances and tolerances, even though 100% inspection 0.24— 0.40 +0.4 +0.75|+0.6 +1.4 +0.6 + 1.6 /+0.6 + 2.0
- is required. This is because the cost of obtaining small tolerances ma die e 03) OS, 0.5, 9 10 —0 +12|-0 +14
mount rapidly. See Fig. 3.2. Also, it will be cheaper on occasion to 0.40— 0.56 40.4 +0.8 | +0.7° +1.6 +0.7. + 1.8 |+0.7 + 2.3
facture some scrap than to take the steps necessary to produce all —O 405/-0 41.2 Obi rei lit [HO 6
within tolerance (§§ 3.9-3.12). RT 0.56- 0.71 +04 +09 |+0.7 +1.6 $074 1181407 + 2.5
When selective interchangeability is used, the problem of servici —O +06}]-O0 +1. —-0O +414/-0 =++418
customer who needs a replacement part should not be overlooked 0.71— 0.95 +0.5 +11 140.8 +1.9 +0.8 + 21/408 + 3.0
example will illustrate.'8-2] Suppose that the desired clearance betwe me phe Poe yl Fes ei eea + ae
1-in. shaft and bearing is between 0.002 in. and 0.005 in., giving a de: 0.95~ 119°} 405 412) 408 +191) 408 T 21] 4oe ’4 23 408 + 3.3
variation of 0.003 in. Suppose that the best tolerances that the mach Osa: FOB Oe FA rR ESTO i nctsbSitre + 2.5
will hold (determined by experience) are: 1.194 78/58 1} OOH FdS) | ENO Deh AO 26 Ovo 3.4L + 4.0
i t —0 +0.9 |—0 +1.8 | —0 + 2.0|—0 + 2.5 |-—0 + 3.0
On the shaft, 0.002 in., On the bearing, 0.003 in.,
i i 1.58- 1.97 +0.6 41.4 1)410 424/410 + 28/410 + 3.4 |+1.0 + 5.0
which add up to 0.005 in., whereas the desired variation is 0.003i Aci hra pagent te 8. ete ae AM
solution to the difficulty
scan would be to hold the bearings 8 to the limits 1.97- 2.56 | +070. +18+13le0)
[412 +27)
and 1.003 in. and divide them into three groups with limits as follows wi file 420 +121-0
Perlis
+3.2)412
aieSO.iy
+42
+3 Stee 1412
Sccdt
+ 6.2
ii5.0
i 2.56- 3.15 +0.7 41.9 |} 41.2 42.9 | +1.2 + 3.7)/41.2 4+ 4.7 [41.2 + 7.2
(A) 1.000 to 1.001; (B) 1.001 to 1.002; (C) 1.002 to 1.003. TO tha RO oe tee bee. 80] oe ch AO RO | + 60
Also make the shafts
two euiips wie to the limits 0.999 to 0.997 in. and divide them
ibe Hasites a 31S See he
—0 ee
+1.8 | —0 aa eye | +aeasteng
+2.8 | —03.5}—0 he + 5.0
enieg an
|—0 + 7.0
4 3.94— 4.73 +0.9 42.6 )4+1.4 43.9) +14 + 49/414 + 69/414 4+ 9.4
(1) 0.997 to 0.998: (2) 0.998 to 0.999, Bie ea a oe os te ee tohas tA
; f ;
pire (1) shafts may be assembled with bearings of either x 4.73- 5.52 10 +2.9 | +16 445 |+1.6 + 60/416 + 80/+16 +11.6
group (A) or *9 $22 *o +35/-0 +50/-0 + 70|-0 +100
and group (2) shafts with either (B) or (C) bearings, with the clearan
. é z ‘ : 5.52- 6.30 +10 +3.2 1.6 +50])+16 + 60)/+1.6 + 8.0/+1.6 413.6
every instance being as desired. Check it yourself. (But this does —0 +25 +o +40|/-O + 50/-0 +70/-0 +120
mean that the actual clearances as randomly assembled will be as sta’ 6.30— 7.09 +1.0 43.5 |} 41.6 +55 ]41.6 + 7.0/+1.6 + 90/416 +13.6
see § 3.12.) Whenever a customer needs another bearing, he will be 0 ZR PHO! FES TO OO Ree ee TO |. 120
bearing (B) which correctly fits either shaft. This plan does not solv 7,09— 7.88 41.2 +3.8 | +18 +62 ]41.8 + 82/4+1.8 +102 /4+1.8 +15.8
problems, but the idea is suggestive. SO Oe SO ee Oe here, eee pee oe
aie ' 7.88— 8.86 +12 +43 1418 +62 nt + oe thf rte +18 +17.8
3 40 FIGURE 3.2 Tolerance vs. Relative Cost. This it alae ig de ra edie niet
Z 3632
%
derived from a variety of data, gives a good idea 8.86- 9.85 41.2 44.3 141.8 +7.2 |} +18 + 9.2}41.8 +13.2 |+1.8 +17.8
general trend of costs for increasing or decreasing to! a) —0 +3.5 -—0 +6.0 —0 + 8.0 —0 +12.0 -—0 +16.0
32 Costs may fall well off of this curve on the favorable 9.85-11.03 +12 +4.9 | +2.0 1.2 2.0 10.2} +2.0 13.2 |+2.0 20.0
@ a because of some factor such as smart design with —0 +4.0 | —0 ; 6.0 +O + 9.0 +5 +120 ae +180
3 facturing in mind, the choice of an especially appro 11.03—12.41
as ; machine, an efficient quality control system, or qu
production with special machines. (From Bolz, R. |
: y
*o°1.2 $40|70°
9 0
$70: |70°: t’s0|70°
i z
Figo [20°H +200 d

: Production Processes, Penton Publishing Co.). / 85

 2.7 Parts in 1000 (0.27%)
of Area Outside + 3¢
86 TOLERANCES AND ALLOWANCES [Ch, 3

~ Frequency or No. of Parts

3.7 PRESS AND SHRINK FITS. When the hole is smaller than the
shaft, it will take force or pressure to put the cold parts together. When
this occurs, the allowance is said to be negative and is termed the inter.
ference of metal. The ASA Standard B 4.1-1955 gives details for five classe FIGURE 3.3 Normal Curve. Also called Gaus-
of interference fits, from which Table 3.2 is taken: FN 1, for light drive fits sian distribution. Total area is 100% ;¥ = average.
For example, area A is 34.13% of the total area
thin sections, long fits, cast-iron external member; FN 2, for medium driy under the curve. Area 2A, between +9, is 68.26%;
fits, ordinary steel parts, shrink fits on light sections, the tightest fit it i etc. These figures come from the mathematics of
advisable to use with a high-grade cast-iron external member; FN 3, fo the curve and apply to all normal curves.
heavy drive fits, heavy steel parts, shrink fits in medium sections; FN 4 an
FN 5, for force fits when the parts can safely withstand high stress. Sh
fits (heating the hub or cooling the shaft or both) can be used where press
the fit is impractical. The mating parts may be sorted into size grou
that the amount of the interference of metal does not vary greatly, obtai 99.73% of the area “‘under’”’ it lies between +3o. Suppose a nominal
a selected average interference i of metal. }-in. drill is used to drill a large number of holes under a constant set of
chance causes (the drill is maintained at the same sharpness, the machine
adjustments do not vary, etc.). All the holes will not be } in. in diameter.
3.8 STRESSES DUE TO THE INTERFERENCE OF METAL. Of a large number of drilled holes, there will be some of approximately
tight fits may produce large bursting pressures on the part with the average size, the highest point of the normal curve, Fig. 3.3; some will be
The stresses in the hub can be estimated with reasonable accuracy fro larger and some smaller than the average. If the number of measured holes
thick-cylinder equations (§§ 8.26, 8.27). For a quick check on the safe is large enough, the distribution of sizes of all holes produced will be much
one can assume that the shaft is rigid and all the deformation occurs in| as suggested by the normal curve, as shown by actual checks of production.
hub; thens = Ee = E8/L. The length L is the hole circumference, L = The total production is, in statisticians’ terms, the population or universe
and 8 is the difference: circumference of the shaft, 7(D + i), minus from which the samples are taken. (Not all industrial measurements have
circumference of the hole, 7D; or 5 = zi. Hence, 8/L = zi/(7D) = a normal distribution, but usually the assumption of normal distribution
and the corresponding stress is s = E8/L = Ei/D. For a cast-iron hub on does not result in a significant error.) The area under the curve, when it is
solid steel shaft, use E = 10,430 ksi. But see § 8.27. fitted to the graphic distribution of a particular production, represents
closely the percentage production between certain “extreme” dimensions.
For example, between +o and —o, where o is measured from the average,
3.9 NATURAL SPREAD OF DIMENSIONS.'-14.3.151 If tolerance there is over 68% of the total area; and we can logically and reasonably
set by the designer are to be respected, the process of manufacture sho’ expect that about 68% of the product from a certain process will fall
be such that the tolerances can be met, or the designer should hav between ¥ + o and X — o, where X = the average or mean value of the
incontrovertible evidence that his tolerances are essential. In one sense, tf process. (The standard deviation co must be determined from the output
the designer, but the material, the machine and tools, and the work of the process itself. See Fig. 3.5.) Since only 0.27% of the area lies outside
determine the tolerance. To gain a better understanding of this stateme of the +3c limits, it is very unlikely, 3 chances in 1000, that any of the
consider briefly a technique of the quality-control engineer—statis production falls outside of the +3o limits. On this account, we often refer
analysis. Statistical control of quality during manufacture is prim to the 3c limits as being the natural spread (NS) of the process, though
aimed at forestalling the manufacture of scrap, but there results a wealth Other values are sometimes used, such as +2.5c. The significance of the
useful information for the designer. It is not enough for the designer natural spread (sometimes called natural tolerance) is that if tolerance limits
know that the size of a part varies as it comes from a manufacturing pra are set narrower than the natural spread, the manufacture of scrap is inevit-
but he also needs to know something of how it varies and what the probab able.
limits of size are. j Thus, if the designer specifies a total tolerance of 0.002 in., corresponding
A theoretical aspect of the answer to this problem is embodied in th Say to EF, Fig. 3.3, and if the natural spread of the process is 0.006 in. and
normal curve, Fig. 3.3, which can be described in terms of the standal corresponding to DG, it would be expected that about 32% of production
deviation o (sigma). The curve extends to plus and minus infinity, Dl would not pass inspection. Therefore, unless there is an otherwise imperative
Teason, always specify tolerances greater than the natural spread, at least

87
 Aolerance as aera YOP SS AM AWASWAN WAS"

=T AREA UNDER NORMAL CURVE

| | From — oo to z measured from zero at <
| zie A zlo A zo A De Lele A

|Natural Spread| Natural Spread |Natural Spread | Natural Spread 0.00 0.5000 —0.8 0.2119 —1.6 0.0548 —2.4 0.0082
NS=6o ~ NS=60 ' NS*60 NS=6o —0.1 0.4602 —0.9 0.1841 —1.7 0.0446 —2.5 0.0062
—0.2 0.4207 . —1.0 0.1587 —1.8 0.0359 —2.6 0.0047
(b) (c) (d) —0.3 0.3821 —1.1 0.1357 —1.9 0.0287 —2.7 0.0035
—0.4 0.3446 —1.2 0.1151 © —2.0 0.0228 —2.8 0.0026
FIGURE 3.4 Tolerance and Natural Spread.
—0.5 0.3085 —1.3 0.0968 —2.1 0.0179 —2.9 0.0019
(a) T < NS. Scrap is sure to be manufactured— the minimum amount being when
—0.6 0.2743 —1.4 0.0808 —2.2 0.0139 —3.0 0.00135
average dimension of the process X falls at the midpoint of the tolerance limits.
—0.7 0.2420 —1.5 0,0668 —2.3 0.0107
percentage of out-of-tolerance parts made is represented by the shaded areas,
(b) T = NS. No scrap is manufactured in a process in control — provided the process
kept precisely centered; i.e., X falls at the midpoint of the tolerance limits. But, see (
(c) T = NS. It is difficult to center a process exactly and impossible to keep it there in from the appearance of Fig. 3.5(a), it would not be surprising to find that the
finitely; hence; when T = NS, some scrap is almost sure to follow. The amount is rep . natural spread of the process is somewhat larger than that computed (the normal
sented by the shaded area, the whole area being virtually 100%. curve), but this conclusion could be justified only by additional production and
(d) T > NS. The proper relation. A good objective is T = 1.3NS, which allows new computations of the natural spread.
process to be off center somewhat without the tolerance limits being exceeded.

3.11 STANDARD DEVIATION AND AREA UNDER THE NORMAL

a third greater,'*-*! unless you are willing to pay the price of the scrap
CURVE. The standard deviation of a. group of measurements x taken
unless you are willing for the shop to ignore the specified tolerance
from a particular population is given by 1°}
Study Fig. 3.4.

3.10 EXAMPLE—ANALYSIS OF ACTUAL PRODUCTION. The data where X is the arithmetic average (Zx/N) and N is the total number of
Fig. 3.5 were taken from actual production. The histogram, a graphical dev measurements. Tables giving the areas under normal curves are available
of tabulation by classes, shows the number of pieces falling into each cell. Cell in handbooks and statistics books. (Notice that o is a root mean square
in (a), for example, shows that 18 pieces measured between 0.2503 and 0.2505 (rms) deviation—§ 3.14.) Table 3.3 is a very abbreviated version, suitable
The facts that only one piece lies in cell 4, one in cell i, and, especially, none in for our purposes. To make the tables universal, the total area is taken as
suggest that there was some abnormal factor that caused these ‘‘unusually”’ I unity; the table value is that fraction of the total area as measured from
holes. A statistical analysis confirms this observation, since cells A and iare outsi —oo to a point located by z/c, where z, the deviation from the mean, is
of the 3c limits when all pieces are included in the computations. However, measured from zero at X, Fig. 3.3. Since the curve is symmetric, areas for
value of sigma used in drawing the normal curve in Fig. 3.5(b), shown dotted
only half the curve are sufficient. To illustrate the use of Table 3.3, suppose
(a), was obtained by omitting the cells 4 and i. Thus, the 3¢ limits of this no m
curve should predict the natural spread of the process when normal factors on
are present. The quality-control engineer calls an abnormal factor an “assignab FIGURE 3.5 Distribution of Size of } in. Reamed Holes. Refer to a book on statistical
cause,” since a complete knowledge of the events leading to a part falling outsi control [3.17] for the method of computing o, which was done in (b) from the same data as
of the 3c limits enables him to assign its nonconformity to a particular “‘cause. plotted in the histogram. Note that the spread between the largest and smallest holes is
Statistical analysis is the only rational tool that can be used to decide 0.0016 in. (0.2514 — 0.2498), but that if the process is kept in control, the probable spread
is 0.0012 in. (0.2510 — 0.2498).
practical certainty whether or not to expect pieces normally to fall into cells ha
In this case, it tells us that it is unlikely that pieces larger than 0.2510 or sm
than 0.2498 in. will be made, provided the process is kept in control. If h
outside of these limits are made, there is some unusual condition—an assignal
cause, such as worker’s carelessness or inexperience, dull tools, poor machine

(No. of Pieces)
setup, etc.

Frequency
Actually, 50 measurements, as in Fig. 3.5, are not enough to give an accura

_
o
measure of the natural spread of the process; hence, these first computed limi
should be considered as tentative. A recalculation could be made after meas

a
100 pieces and another still later. The larger the number of pieces measured,
better will be the definition of the process. However, measurement of three hund
pieces will define the limits of the controlled process quite accurately. Judg ae ——
oe ea
Size by Cells 0.251
 SS Re See LENE A VV AISED Fe a tt a
wn
equation
the lower tolerance limit on the hole of Fig. 3.5 is 0.2500. What proporti OF squares of the standard deviations, o, and og, of the variables; in
of the production will be less than 0.250 in. (scrap)? The mean ¥ = 0.2 04 form,
the deviation from the mean is z = x — ¥ = 0.2500 — 0.2504 = —0.00 04. cp = (o,? + Ga")
o = 0.0006/3 = 0.0002; hence z/s = —0.0004/0.0002 = —2; from Table (b)
3.3, read 0.0228 corresponding to z/s = —2, which means is the natural spread of the differences, 6c, is the natural spread
that ab.
2.28% of the production should normally be smaller than 0.2500 in . ie o. dimensions, ie 60, is the natural spread of the bore dimen-
the top tolerance limit is 0.2508, there would be here an equal loss—hole sions. Substituting o, = 0.002/6 and og = 0.003/6 into equation (b), we
too large (by symmetry). / find op = 0.0006 (to four decimal places). Adding 3cp = 0.0018 to the
Observe that while the preferred dimensioning practice is to give known mean of 0.0045, we get the maximum likely clearance of 0.0063 in.
dimensions, one works from the mean dimension in statistical stud for the situation as defined. Likewise, by subtraction, we find the minimum
clearance is likely to be 0.0027 in. These limits are much larger than the
-in. wance.
3.12 STATISTICAL DISTRIBUTIONS OF FITS. Given a produ oe cn the curves S and B are off center toward each other but
tion of two mating parts, one might suppose that the allowance is actu still within their respective tolerance ranges, the probable minimum fit will
the closest fit that will be obtained during the assembly of these p be greater than the allowance. Thus, let S move toward the right so that
No doubt, the allowance is sometimes mistakenly chosen on this assu point C is at C’ and let B move toward the left so that E is at E , a total
tion. However, such a fit under controlled manufacturing conditions movement toward each other of 0.001 in. The standard deviation of the
quite unlikely, and it is sometimes important for the designer to know t differences does not change, but the mean difference will now be 0.001 in.
For example, suppose the shaft tolerance has been set at 0.003 in., less (0.0045 — 0.001 = 0.0035), and the minimum clearance or fit will be
bore tolerance at 0.004 in., and the allowance at 0.001 in., as sho 0.0027 — 0.001 = 0.0017 in. (versus an allowance of 0.001 in.). It is seen
Fig. 3.6(a). Jf the tolerances are a third or more greater than the nat that it is possible to obtain working fits on an interchangeable basis even
spread of the processes and if each process is centered with respect to though the allowance is zero. There are occasions where the designer can
specified average dimension, the distribution of sizes of the shafts may b take advantage of this fact by decreasing the allowance and increasing
represented by the normal curve S, and of the bore by the normal curve the tolerance with a saving in costs and scrap.
Fig. 3.6. If a shaft is picked at random, its most likely size is 1.0000 in., Similarly, if several parts are assembled together externally, one next
size at the maximum ordinate on the normal curve S. Similarly, the m to the other, the standard deviation of the dimension of the (sum of the)
likely size of bore is 1.0045 in., which gives the most likely difference assembled parts is given by
clearance) as 1.0045 — 1.0000 = 0.0045 in. This 0.0045 dimension is th
most frequent difference and is therefore that found at the highest po (c) o = (0,2 + G2 + og2 + ° °°),
of the normal curve D, which shows the distribution of the clearan where 61, 6, dg, etc., are the standard deviations of the dimensions of the
Statistical theory defines the standard deviation op of the difference respective parts. Thus, if the tolerances are proportional to the standard
sum) of two independent variables as the square root of the sum of deviations, the overall tolerance may be

T = (T3244 Te 2 632,
(d)
FIGURE 3.6 Distribution of Clearance in Fits. If the processes should move off c fe
as they will, the effect on the normal curve of the differences in (b) is only to shift its and not the sum of the individual tolerances, as often assumed. Even though
from 0.0045 in. to some other value. For example, if S' moves leftward 0.0005 in. so : tolerances are not proportional to standard deviations, the conclusion in
is at A’ (shown dotted) and if B moves rightward 0.0005 in. so that F is at F’, the mea equation (d) has a general validity; the exact effect on actual dimensions
difference would be increased by 0.001 in. to 0.0055 in. the maximum fit would bece m
0.0073 in. (point H’), and the minimum fit 0.0037 in.
can be determined only from statistical analysis based on adequate data.
To illustrate the idea, assume that the processes involved in manufacturing
1.0000 1.0045 Table Gives 0.0045 parts 1, 2, and 3, Fig. 3.7, are centered in the tolerance range and that
a Processes ine Fraction of Area | tolerances are equal to the natural spread of the processes (the ideal
i LS\ Centered \ from —© to situation). Suppose that the desired overall tolerance is T = 0.018 in. =6c.
/ in Tolerances \ Here —=/ D
i h / \ eS
A NK \ee Bes 7 FF’ G cA 2} HH
Natural | | 8 Natural Natural Spread FIGURE 3.7. ee 30 =0.009 ;
Spread : Spread of Difference
0.0021 b= | Sh — 0,003 —m = 0.0036 —el
S| & >
0.0027 | 0.0063
1 2 3
|_ Tolerance S Tolerance
‘Shaft = 0.003 ad Bore =0.004 Min. Fit Max. Fit
(a) (b) Fs +301 Ie+302 % +303
 92 TOLERANCES AND ALLOWANCES [Ch.3 AND SURFACE FINISH 93
53.14] TOLERANCE

What tolerances must be applied to the individual parts? Let us assume th at Fig. 3.8. Considering the foregoing discussion of statistical aspects, we
T, = T, = Tz and o, = og = 63. On the basis of simple arithmetic, if recognize that this particular combination of parts is very unlikely ina
seems logical to divide the total tolerance by 3 and make each part witha random assembly operation, so unlikely for a manufacturing process in
control that it almost surely will not happen. However, by using this
combination, our conclusions will be on the “safe side.” Making the match
%+0.009
30 =0.009
even more unlikely, we shall assume the worst geometric situation, whose
parts have the minimum size of hole and the maximum diameter of bolt
permitted by the tolerance. Recall the definition of allowance A as the
1 2 3 FIGURE 3.7 (Repeated).
difference between the minimum hole Hmin and the maximum “shaft”
Re Ie +302 Bs +303 Bmax; Fig. 3.8; that is,

(e) A = Hpin — Bmax.

individual tolerances can be made significantly larger. Using equation (€
But we also see from Fig. 3.8 that'?-*)
with « = 0.018/6 = 0.003 and o, = og = o3, we find
T
o = 0.003 = 30? or o, = 0.00173 in. (f) Amin — xo ae

This corresponds to a tolerance of T, = 6c, = 0.0104 in., which, if usec

instead of 0.006 in., may mean a considerable reduction in cost. If where T is the tolerance on the hole spacing L. From (e) and (f),
some reason, the tolerance on one part, say 1, should be more or less t
on another, decide upon a ratio of tolerances, reduce the radical in equatic (g) ss =A or T = 24;
(c) or (d) to one unknown, and solve as described.
Since a little knowledge is dangerous, one would acquire more kn that is, the tolerance T on the hole spacing should be twice the allowance A
ledge of the statistical approach before use in actual engineering problem for hole and bolt.
In general, our conclusions are valid only when the production proce A similar geometric study*-*) shows that if there are more than two
are “in control,” holes, the tolerance on the spacing comes out

(h) fn on ws Oa Ae
3.13 TOLERANCES ON LOCATING HOLES. It often happei /2
that two (or more) parts are to be assembled with matching holes for b tolerances in two directions are necessary, and they are
In this case,
or screws where accuracy is important. If the holes are close to a match preferably the same.
location and somewhat undersized, the parts may be put together at asse holes located with reference to one
Where there are more than two
bly and the holes finished to size, a procedure that automatically obté another, there should be a “master hole”’ (or reference surfaces) from which
a good match. Frequently the most economical, this procedure is commo! all others are located. It is cheaper to locate two holes to, for instance,
However, if assembly on an interchangeable basis is to be made, the variot
+0.002 in. with respect to one another and the other holes to +0.010 in,
tolerances need to allow for this fact and they must be practical. than to hold all holes to a close tolerance, and it is usually just as satis-
Suppose we wish to locate a pair of holes on each of two articles w: factory. Bilateral tolerances, L + 7/2, are preferable for hole locations.
are to match at assembly. Suppose that one pair is at the minimum spacing

3.14 TOLERANCE AND SURFACE FINISH. There is necessarily

FIGURE 3.8 Tolerance on Hole Loca a close relationship between the smoothness of a surface and how fine the
The spacing of the holes is Z + 7/2. The tolerances can be made. It would be ridiculous to specify a tolerance of a
ference between L + 7/2 and L — 7/2 is much few ten thousandths of an inch for a surface that has irregularities of several
exaggerated in the illustration.
thousandths of an inch. As would be expected, the smoother the finish, the
more expensive—generally. The standard B 4.1-1955 gives the standard
tolerances shown in Table 3.4 and suggests the methods of finishing the
surface appropriate to the grade as given in the heading of the table. The
 TABLE 3.4 STANDARD
Flame Cutting-Mochine
TOLERANCES (from ASA B 4,1-1955) Rough Turning.
Contour
(This is only part of the tabulation in the Standard.) Tolerances are in thousandths of an Rough Grinding
inch; for use with standard gages. Includes values agreed to by ABC (American, British, Shoping ond
Drilling
Canadian). Not part of the standard but given as a guide: Grades 4, 5—lapping and honing; Milling— High Speed
Grades 5, 6, 7—cylindrical grinding, surface grinding, diamond turning and boring, broach- Finish
ing; Grades, 6-10 incl.—reaming; Grades 7-13—turning; Grades 8-13—boring; Grades Broaching
10-13—Anilling, planing, shaping, drilling. See also Fig. 3.9. Boring
Reaming
Commercial Grinding
NOMINAL Milling- Carbides
Geor
SIZE RANGE GRADE Borrel
INCHES Roller
Diamond
Over To 4 5 6 7 8 9 10} 11 12 | 13 Diomond & Precision
Precision Finish
0.04— 0.12 0.15 | 0.20 | 0.25] 0.4 | 0.6 | 1.0] 1.6] 2.5 4
0.12- 0.24 0.15 | 0.20] 0.3 | 0.5 | 0.7 | 1.2] 1.8] 3.0 5 Production
0.24— 0.40 0.15 | 0.25/04 |06 | 09 ] 1.4] 2.2] 3.5 6 KEY
0.40- 0.71 0.2 | 0.3 |04 10.7 | 1.0 | 16] 28] 4.0 7 GB Usual average or most
economical range
0.71- 1.19 0.25 | 0.4 | 0.5 | 0.8 | 1.2 | 2.01] 3.5] 5.0 8 7) Full ae }-Averoge Machine Finishes
1,19- 1.97 0.3) 04 | 0.6 | 1.0 | 1.6 | 2.5] 40] 6 10
sp Average Critical
1,97— 3.15 0.3/0.5 | 0.7 |1.2 | 1.8 | 30] 45] 7 12 Ports Ports & Ports”
3.15- 4.73 04 |06 |}0.9 |14 | 22 | 3.5] 50] 9 14 22
4.73- 7.09 0.5 | 0.7 | 1.0 | 1.6 | 2.5 | 4.0] 6 10 16

tin.)
7.09— 9.85 0.6 |08 | 1.2 |18 | 28 | 45] 7 12 18
9.85-12.41 0.6 | 0.9 | 1.2 | 2.0 | 3.0 | 50] 8 12 20
12.41-15.75 0.7 |10 | 14 | 2.2 | 3.5 | 6 9 14 22
15.75-19.69 0.8 | 1.0 11.6) | 2:5 ).4 6 10 16 25 40

required)
valuable information given in Fig. 3.9, while it does not agree in all pi

(in units of time

culars with the typical roughnesses mentioned in ASA 46.1-1955,'3-91

D
worthy of detailed study. It stands to reason that the material itself

Typical
be nearly flawless in order to obtain uniform highly-smooth surfaces.

Cost
Roughness is the relatively finely spaced irregularities of the surf:
Fig. 3.10. Waviness is the irregularities or departures from the nomi

Relative
surface of greater spacing than roughness. Lay is the direction of the
dominant surface pattern, Fig. 3.10, usually determined by the producti
method, as tool marks. The numbers that specify the roughness R
nN
either: m

1. The average arithmetical deviation from the mean line, in wi Surface Finish (microinches, rms)

case the area above the mean, Fig. 3.10(b), is the same as the area below
With none of the y coordinates taken as negative, the roughness is FIGURE 3.9 Surface Finish Related to Process and Tolerance. The tolerances shown at
1 L the right on the lower part are indicative and should not be used unless they are known to
(i)i R=—ae dx. apply to the manufacturing process. For a roughness of 63, the tolerance may be between
about 0.010 and 0.001 in., dependent to some extent on size. Enter this figure and check this
range for yourself. (From Bolz, R. W., Production Processes, Penton Pub. Co.).
Most instruments now in use give this value.-17)
2. The root mean square (rms) average, for which

i) Rime [> f ste] 0

FIGURE 3.10 Roughness and Waviness. The lay shown in (a) is perpendicular to the
line 4B that would represent the nominal surface.

The units are microinches or win., which is one millionth of an in I Mean Line
(10-® in.) The arithmetical mean is theoretically 11% less than the rms;
94
 96 TOLERANCES AND ALLOWANCES [Ch.: §3.15] CLOSURE 97

this is not sufficient in most cases to warrant changing the roughnes: 0.001 in.; milled threads; pressed fits; keys and keyways; rolled surfaces; cams
specifications in a shift from rms to arithmetic. Much of the literature deal and followers; surfaces for copper gaskets; ordinary journal bearings (nearer
with the rms, as in Fig. 3.9 and below. 32 rms); mating sliding surfaces; worm gears.
The symbol shown in Fig. 3.11 is used on drawings to designate 16-32 rms: specified when finish is of primary importance; fine cylindrical
desired maximum surface roughness.-*! The values preferred for spe grinding; smooth reaming; coarse hone or lap; automotive valve stems; pistons;
cylinder bores; ground screw threads; gear teeth for severe service; precision
cation purposes are those numbers on the abscissa of Fig. 3.9 and in th
summary below; that is, specify a maximum journal bearings; shaft seals; precision-mating sliding surfaces; cold drawn sur-
of 32 or 63 or 125, etc., 0 A
faces; splined shaft; brake drums and other friction surfaces.
range of acceptable values. Standard reference surfaces for these number
8-16 rms: used only when coarser finishes are inadequate; finest cylindrical
are available. For the ordinary run of shop work, a fingernail test has pro ve grinding; microhoning, honing, lapping, buffing; automotive cylinder bores;
good, but visual tests are deceptive. The fingernail test is a comparison seats for antifriction bearing races; crankpins; valve seats; valve stems; rolled
given surface with graded and standardized roughness specimens, threads (this is a low-cost production method); precision rolling surfaces; cam
obtained by rubbing the fingernail across the surfaces and matching faces.
feel. Specify the waviness when it is significant. The symbols used for 2-8 rms: produced by honing, lapping, superfinishing, very fine buffing and
are: =, parallel to the surface boundary line; | , perpendicular to the polishing; generally expensive; piston pins; pressure lubricated bearings; pre-
face boundary line; x, angular in both directions; M, multidirectiona cision tools.
1 rms: gages, micrometer anvils, mirrors.
Waviness 0,002

iach Applies 3.15 CLOSURE. The designer should remember that tolerance is
ecording 4
to Space 63f Sila! FIGURE 3.11 Symbol for Specifying th the stated permissible variation of a dimension and that the relation betweeh
Available 4 B Surface. ;
this stated variation and the actual variation is sometimes coincidental and
(a) (b) often not as the designer hopes. Tolerances should be realistic, as well as in
accord with good engineering practice. Realism in tolerances will follow
C, circular ; R, radial. The symbol in Fig. 3.11(b) is interpreted as a su fac from a knowledge of what the manufacturing department can and cannot
of maximum roughness of 63 pin., a maximum waviness of 0.002 in., an do. The “natural spread’? is sometimes called the “‘natural tolerance,”
the lay perpendicular to the boundary AB. but this latter term should be avoided because the word tolerance has a
Remember that the ideal surface finish is the roughest one that will meaning that results in confusion to those not familiar with the idea.
the job with satisfaction to all concerned. It should be observed tha Special considerations are involved for interference fits of materials that
surface finish for parts that have been forged, cast in permanent mol are stressed plastically. Quantitative data for some plastics commonly
dies, rolled, cold-drawn, or extruded, will depend to some extent o used in machines are found in Ref. (3.18). Tolerances concerning out-of-
surface finish of the die; for example, cold-rolled steel may have a roundness, straightness of holes, rods, etc. are important but beyond the
in the range 63-250; extrusions, 32-125. The summary scope of this book.
below.
3.8,3.10] will be helpful.
ie or greater rms: heavy cuts and coarse feeds; sand castings; hot-roll
steel. 4
250 rms: shows definite tool marks from rapid feed; forgings; very
coat
grinding; suitable for machined external surfaces in general; chased threads.
63-125 rms: high-grade machine work; smooth machine finish obtai
a high-speed, fine-feed, light-cut, sharp-tool pass; coarsest finished to be
use
bearing surfaces when the motion is slow and the load light; surfaces fo:
gaskets; forgings; permanent mold castings; investment castings; die or
threads; datum surfaces for tolerances over 0.001 in.; outside gear dia
mating surfaces of brackets, pads, housing fits, bases, etc.
32-63 rms: finest machined finish, obtained with carbide or diamond
medium grind, reamed, broached, burnished finish, die castings; teeth of
rat
and pawls; gear teeth for ordinary service; datum surfaces for tolerances
 4. VARIABLE LOADS
AND STRESS
CONCENTRATIONS

4.1 INTRODUCTION. It has been said that 80% of the failures of

machine parts have been fatigue failures. Whatever the true percentage, it
is a large one, which suggests that designing machine elements should
always be done from the viewpoint of the possibility of a fatigue failure.
Even where the pattern of variability of the loading cannot be predicted
and where the magnitudes of the forces are unknown, which is so often
true, consideration of the principles of design for fatigue should be
rewarding.
This point of view gives the designer considerably more to think about,
as we shall see, and often challenges him to extra effort in trying to define
the loading more completely. This chapter involves problems without
combinations of different kinds of stresses, so that the reader may acquire
a working knowledge of the design attitudes and of the language of metal
fatigue before facing involved stress analysis (as in Chapter 8). Although
we shall present a fairly straightforward procedure for design, this is an
area about which we are still fairly ignorant. What knowledge we have is
largely empirical, but the literature on the subject has grown to formidable
proportions within a single generation, a commentary on its importance.
If you practice mechanical engineering design, this is an area for further
extensive study.

4.2 MECHANISM OF FATIGUE. On a macroscopic scale, fatigue

failure starts at some point (because the repeated stress there exceeds the

99
sail
100 VARIABLE LOADS AND STRESS CONCENTRATIONS [Ch. 4 54.4] CHARTING ENDURANCE STRENGTH 101

endurance strength of the material) in the form of a minute crack that producing a primary shear stress tends to follow the plane of the principal
gradually spreads with repetitions of the excessive stress until the resisting tension. The crack from fatigue failure under a repeated compressive stress
area becomes so small that complete fracture suddenly occurs, probably follows approximately the maximum shear direction (so fatigue failure by
without warning and perhaps by now even with a small applied load. The compression only is a special case). Briefly compare the mechanism des-
fracture for quite ductile materials is without significant plastic action, cribed above with that described for creep failure (§ 2.21), which follows
Fig. 4.1; hence, such fractures are often referred to as brittle fractur from a deterioration and movement at the grain boundaries.
or brittle failures (§ 2.22). The suddenly fractured surface has a lustr In actual machine parts, the cracking usually starts at a discontinuity,
crystalline appearance characteristic of all brittle failures. Since fatig a fillet, a scratch, an inclusion or hole inside the member, a keyway (Fig.
failures are also the consequence of a spreading crack, they are even m 4.1), etc. This is because discontinuities have the effect of increasing the
aptly called progressive fractures. stress locally (in the vicinity of the discontinuity), as described in more
While there are several descriptions of the fatigue phenomenon in detail later. The fatigue failure of a smooth, polished, rotating-beam speci-
literature,'*-1-41°) we can indulge only the briefest one here, which may men without internal flaws starts at the outside surface not only because
the maximum stress is there but perhaps because the outermost crystals,
not being strengthened by other crystals on all sides, are more prone ona
statistical basis to be the first to slip or shear.
FIGURE 4.1 Shaft after Fati
Failure. The slots in the cylindrical
face are ends of keyways, where
fatigue failure started. Failure gradu: 4.3 ENDURANCE LIMIT, ENDURANCE STRENGTH. Instead of
progressed from both keyways to using the yield stress or ultimate stress as a basis of finding a design stress
center. The surface that failed early for a part subjected to a varying load, we should now use some sort of
subsequently rubbed nearly smooth
endurance strength. When one speaks of the ultimate stress, it is presumed
load. The final break is at the
ridge, about on a diametral line. O) that it is for a standard 0.505-in. specimen unless otherwise specified. It is
nearly identical lines of prog correspondingly convenient for endurance strengths to refer to standard
failure leading from both keyways specimens. We shall use endurance limit and fatigue limit s, to mean the
ward the center. (Courtesy Joseph maximum reversed stress that may be repeated an indefinite number of times
Ryerson & Son, Chicago).
on a polished, standard (nominal 0.2 to 0.4 in. diameter) specimen in
rotating bending without causing failure.* There are other sizes of ‘‘stan-
dard” specimens and other ways of testing for fatigue strength (as torsional
or axial loading). Not every material exhibits an endurance /imit, particu-
just as well because knowledge of the subject is incomplete. Fatigue failure larly true for many nonferrous metals. See the curve for aluminum alloy
is believed to start at random points as slip on shear planes of crys 2017, Fig. 4.3. We shall use endurance strength and fatigue strength s,, in
that happen to be oriented so that this can happen. Imperfections in a general sense for the strength of actual parts, specimens with notches,
crystal and other imperfections, such as the penetration of grain bout etc., and for materials without a particular /imit. Thus, because of its
daries by oxides, help to start such failures, which are soon microscopit general application, s, will be the symbol found in most equations, and
Crystal slip continues with repeating stresses until visible cracking occu the reader should learn when to use an endurance limit s’. For materials
Although shear produces the crystalline slip, the crack will spread in that do not have a limit, the number of cycles for which the given strength
direction of a plane on which there is a tensile stress, if there is a ten applies should be stated. For example, see Table AT 10.
stress. Almen'*-64! shows illustrations of cracks that are in the direc
of a plane on which there is compression from an external load, but tl
cracking is attributed to the residual tensile stress (§ 4.23) known to € : = CHARTING ENDURANCE STRENGTH. Finding the en-
Because the cracks were subjected to compression by the external lo ¢ urance strength is time consuming and expensive. The quickest way is to
they did not propagate to failure. On the other hand, once a crack ¢
in a plane subjected to tension, the high stress concentration at the ends ol
* . *
ees Is only fair to warn that some of the literature does not abide by this definition of
the crack (§ 4.24) encourages its rapid spread. The crack from a load urance limit, but it is a convenience to have such a reference.
102 VARIABLE LOADS AND STRESS CONCENTRATIONS [Ch.: 54.4] CHARTING ENDURANCE STRENGTH 103

use a machine such that the specimen can be subjected to a constant bend. pieces and for each group say “Half of this group will fail at stress s by
ing moment, then rotate the specimen. In 360° of rotation, a point on the time the median curve is reached,” we will be right 50% of the time.
specimen undergoes a complete cycle of stress from maximum tension To get a higher probability of being right, the sample size would have to
maximum compression, back to maximum tension. Typically, fatigue d be increased. If the sample size had been 10, we could make the same
are plotted on log-log or semilog paper, stress vs. number of cycles statement with the probability of being right 99.9% of the time.
failure. If done in the approved manner,'*-5*! the solid line of Fig. 4 Not all specialists in this area agree that the normal curve (Fig. 3.3)
represents the median strength as obtained from a number of tests. is the best distribution form to use, but it is seen in Fig. 4.2(b) that the
median is the middle-value of a group of measurements (or if the sa distribution of the logarithm of the number of cycles is reasonably approxi-
size is even, it is the average of the two middle values), which is not n mated by a normal curve. Hence, we may, with appropriate approximation,
sarily the same as the mean or arithmetic average; the median and m think of the failing stress distribution at any number of cycles to be distri-
values usually approach each other as the number of measured pieces (fi buted normally, with a certain standard deviation o. As previously pointed
the same population) increases. Thus, the median curve represents — out, one would expect that almost the entire population would fall
stress state at or below which 50% of the specimens failed and at wh within +30 from the mean. From Table 3.3, we find [84% survival:
50 % survived; it can be designated as the 50% survival level. For exam, (1 — 0.84)/2 = 0.08] for 0.08 that x/o = 1.4; that is, the 84% level in
if the stress is about 104 ksi for the polished (upper) specimens, we expec Fig. 4.2 represents about 1.40 from the median (strictly from the mean)—
that 50% would survive more than 10° cycles: of stress. However, we with a 50% confidence level. This is to say that the median of future groups
not certain of this and the probability of our being exactly right on from the same universe should fall above the 84% survival line 84% of the
50% depends on the number of specimens tested. If 100 specimens ° time. If one must reduce the odds of failure, a lower survival line must be
tested at each of several stress levels, the corresponding median line wi the basis of design (or the design factor must cover the gap). If the distance
be so close to the true value that the probability of 50% survival at a p is doubled, this would be about 2.86 and about 99.5 % do not fail (0.995 re
ticular point would be quite high. On the other hand, if there are 4 pi 1 — 2 x 0.0026, from Table 3.3). Details of a more complete handling of
tested at each stress level, a more likely number, we are much less the statistical aspects are given in working form in Ref. (4.56).
that the true median has been found, and the probability of a parti Good distributions for indefinite life are not easily obtained for
point representing the true 50% survival is lower. For the same r materials with a fatigue limit, because so many of the specimens in this
there is a corresponding uncertainty in the distribution of say fa’ vicinity run out without failing. Hence, extrapolations are made. Unques-
strengths at any particular number of cycles. The degree of certainty tionably, we should know something of the spread of endurance strengths
this regard is called the confidence level, which is easiest defined by examp as a means of being sure that a design is not being made within the failure
It happens that the confidence level for Fig. 4.2(a) with 4 specimens t ‘ range, but to get these data with a high confidence level is quite expensive,
at a given stress level is 50%{; this means that if we pick up 100 grou and one seldom finds reports that have included the necessary statistical
analysis. Additional information may become available in the future and
designers can be guided accordingly, but in the meantime, we shall accord
FIGURE 4.2 Endurance Strength Bands. Rotating beam tests.{**?5] The specimens|
with current custom and discuss the design from the standpoint of average
had a nominal diameter of 0.2 in. The tests, represented by the dots, were for c:
manufactured specimens, which accounts in part for the small scatter. With sufficien' Or median strengths (the literature seldom distinguishes). Stulen et
points, the bands typically narrow as shown, the boundaries tending to intersect al,'*-26.No. 13] report standard deviations o for long-life fatigue strengths
vicinity of 10° cycles at s, for ductile steel.{*-°4) In (b): the distribution of failures of AISI 4340 varying from 4.9 to 7.8 % Of Sy, (8% looks safe); for titanium
log n,; at a certain stress level; after Dolan.'**") 6 Al 4 V, c = 0.064s,; for aluminum 7076-T61, o = 0.06s,,; for beryllium
4340, OQT 860°F, Polished ae oe 0.075s,,; for Al-Ni-Bronze, 5 Ni 10 Al, o = 0.094s,. Thus
; € lower limit of fatigue strength for most metals might be estimated as
No. Failed

a. 3 x 8 = 24% below the average 5,. Steels less carefully manu-

* ri may well have greater dispersion and co. Also, working against
fin : 2 the probability of a greater spread (larger c) in the strength of the
a e Part because of less uniformity in geometry and composition. In
c 8n, it should be advisable to be as sure as possible that the operating
Tess does not exceed (1 — 0.24)s, = 0.76s,; that is, that the factor of

10° 2x10 -5x10 10° 2x10 5x10 10’

= Number of Cycles to Failure
 380

70 spectrum as in Fig. 4.4(a). In some cases, the designer can estimate the
60 number of times that some maximum loading will be imposed on a part
during its lifetime, and choose design stresses for this many repetitions,
50
§ 4.16. Whatever the pattern of variation, given enough experience
Stress in Kips per Square Inch

45
(measurements), an idealized model of the load variation can be constructed
40
as a basis of design. The most common models are sinusoidal and are given
35
in Fig. 4.5. There will be a maximum and a minimum stress, an average
30
or mean stress s,,, and a variable or alternating component of the stress sq.
25 If both kinds of normal stresses, tension and compression, are involved,
we must use algebraic signs, negative for compression. In Fig. 4.5, we see
20 that the alternating component is in each instance that stress which when
17
added to (or subtracted from) the mean stress s,, results in the maximum
(or minimum) stress. The average or mean stress s,, and the alternating
15
component Sg are
12
Smax + Smin Smax — Smin
and Sg = ————_,
10
10¢ 2x10! 5x104 10° 2x10° 5x10° 10° 2x10° 5x10° 10° (a) ae 2
Number of Cycles to Failure
where a compressive stress is a negative number. For a complete reversal,
FIGURE 4.3 Typical sn, Curves. The solid curves are typical 50% probability curves fo Fig. 4.5(a), 5_ = 0; that is, Smin = —Smax and Sg = Smax. In every case,
standard polished specimens in rotating-beam tests; the dotted curves are for polish
specimens but with some deviation—a notch, in water, and decarburized. The he (b) Smax = Sm + Sa:
portions of the curves represent the endurance limits. Note the absence of a horizontal
for the aluminum alloy 2017-T4, typical of nonferrous alloys.
A parameter used to locate the curves of Fig. 4.5 is a stress ratio R
Observe that SAE 1050 notched is weaker than as rolled 1020; and that the sloping pz defined as
of the dotted curves tend to be steeper than the corresponding solid curves; that is, that
effect of stress raisers is less for finite life (§ 4.16) than for indefinite life. Curves A R fal s min /
for SAE 4140, OQT to a hardness of 280 Brinell; C and D for 3140 hot-rolled mat
Smax
E and F for 1050, quenched and temperated at 1200°F. Despite the fact that the
burized layer in curve B was thin, it materially weakened the specimen. It is not shown 2 with stresses used algebraically; R = —1 for completely reversed stress,
but processes that leave a residual compressive stress on the surface, such as pe
Fig. 4.5(a).
§ 4.28, result in a smaller slope (as compared to polished specimens) for stresses gi
than s,’. For a particular material, these sloping lines tend to intersect at a point di
approximately by 10° to 10+ cycles and 0.9s,, to s, stress. 8

FIGURE 4.4 Spectrum of Stress Variation. $ ANd

safety used with average endurance strengths covers this eventuality aft Time
various stress raisers explained below have been accounted for. Time
(a) (b)
Figure 4.3 shows the typical stress curves for a number of differ
materials and configurations, and it should be studied. Observe that
fatigue limits are well below elastic limits. FIGURE 4.5 Sinusoidal Stress Variations. (a) Represents completely reversed stress;
rotating beam with constant moment; R = —1. (b) A repeated stress has either zero
minimum stress as shown (R = 0) or zero maximum stress (all compression R = ©).

4.5 HOW STRESSES VARY. Stresses may very well vary in a qui’
‘irregular and perhaps unpredictable manner, as the stresses in the stru C-
ture of an airplane passing through a storm, Fig. 4.4(a). An impulsive
load may set up a vibration, Fig. 4.4(b), that gradually damps out; and
vibrational stresses may be and often are imposed upon some irregular {

104
So = amplitude (c) Smaller Part
Compression
(a) Reversed (b) Repeated (or Tension) (d) Ten. (or Comp.)
106 VARIABLE LOADS AND STRESS CONCENTRATIONS [Ch. 4

4.6 REPRESENTATION OF THE ENDURANCE STRENGTH IN

TERMS OF ALTERNATING STRESS. The type of diagram in
Fig. 4.6, with variations, is frequently used as an aid for design. Its con-_
struction is indicated by the symbols s), sy, Su, etc. Straight lines are ofte re FIGURE 4.7 Soderberg Line. Recom-
used, as HAQ, but if test values are known, a curve such as HPQ can be mended for ductile materials only.
constructed. The lines HQ and JQ are thought of as the limiting (maxi-
mum and minimum) stresses for indefinite life (say 50% confidence). For
finite life, as 10° cycles, some line VQ (with a matching one for a lower
boundary) would apply. With a factor of safety NV, the maximum and
minimum stresses would fall on the heavy dotted lines. Note that the” We shall use a diagram, Fig. 4.7, in which the ordinate is the alternating
maximum stress is limited to the yield strength, so that the bounda stress and the abscissa the mean stress. Draw a line through the endurance
becomes X YZ. Conventionally, the static yield strength is used, because limit (or strength) at C and yield strength at T. This line is called the
its values are available, but the yield strength under short-duration loads Soderberg line,'*-1°-415] and points on it are assumed to represent a state
is greater than for gradually applied loads. of stress that puts the part on the point of failing after an indefinite number
As you may have inferred, we have no theory that says how the mean of alternations of sg. For example at P, a variable stress OV on a mean
and variable-stress components are related. Hence, the empirical approach, stress OM is the limiting condition. Since most actual points of failure of
polished steel specimens fall outside of this line, as suggested by the points
FIGURE 4.6 Goodman Type of Diagram. The influence of the mean stress is clea shown, the Soderberg line is a conservative basis of design. However, a
shown by this diagram. For example, if the mean stress is zero, failure occurs with a factor of safety is necessary; divide it into s, and s, to get points D and G;
maximum stress slightly over s,’ at H. At any mean stress s,, at M, the limiting range of and the line DG then represents a locus of points that in turn represents
stress is AB, a safe range is CD. Compressive ranges K defined by this diagram are overt;
conservative; so it is not uncommon to let the permissible range of stress TW at R =
safe conditions. The combination s,, and sg at B is for a design factor of V
remain constant as |s-| increases (but for indefinite life, it is probably best for the maximu (also NV = OP/OB). The equation for line DG is easily obtained by setting
|sc| = sy). Tests show that the mean compressive stress can be increased substantis up proportions for the similar triangles BED and COT; its useful form is
with no decrease in the amplitude of the variable component and without decreasing t obtained by solving for 1/N:
factor of safety (see Fig. 4.8). See Ref. (4.65) for Goodman diagrams drawn to scale
Sq Sy[/[N—Sm
() aa
numerous materials, including ones for torsion.

Su Q Sn Sy
Test Values
Limited No.of — p oe
Applications, ae epee
as 10°. z= Ps
NOY AS UES a
o}— fee A nt
de JAZ There are several other possible criteria for design, including the modi-
fied Goodman line and the Gerber line of Fig. 4.8. The working curve for
e / ! the Goodman line is BD, and by analogy with equation (4.1), its equation is
zie
SX
“N

(4.2) i Sm + Sa
3.5° 7 Min. Nivise Sass
Stress
often used for brittle materials, such as cast iron. The Gerber line is a
E “Zero Stress 8x parabola with apex at C; the corresponding failing curve is

(d) i (=) ee.

pee Stress Limited
by Yield Strength Sy Sn

The design curve is obtained by dividing s, and s, by N thus obtaining a

quadratic for N. The German practice is to use this equation, which in truth

107
 ital se GL ae
Mi ee eae Soe Typical Failure Points
<n erie mee ae 54.7] ESTIMATIONS OF ENDURANCE STRENGTHS 109
a ae - ie eS SN
Ly ae
BE Modified Goodman il
Line
compression, but as suggested by the points on the compressive side of
é sashaste i the sq axis in Fig. 4.8, the results are ultraconservative. Compressive stress
7 tA SJ ge inhibits the propagation of a crack.
7 a y CP ae aay ~<a In solving any of the foregoing equations, s,, and s, are computed as
rst | en a nominal stresses So, which are those corresponding to F/A, Mc/I, and Te/J;
Pat se el Nee that is (if there is no rotation),
~ SQA
Compression 9; - M B \Y 8m or sm |! En Myc TC
in & Bu (f) Sn = el or Sn = Bar or Sn = 7

a ©) 3 F, ie 3 M,c : T,c
g a= a a = T Oo Sa a@ =

FIGURE 4.8 Comparison of Variable-stress Criteria. The Goodman line is used

brittle materials, preferably with a higher design factor than would be used for ducti Special handling of Mc/J for a rotating member is required because the
materials. Many designers use it also for ductile materials. In this type of diagram, f stress undergoes cycles even though the moment is constant.
coordinates are frequently made dimensionless by dividing sg by Sp and S,, by Sy (or Sy).
change makes the yalue of the end points A and C unity.

4.7 ESTIMATIONS OF ENDURANCE STRENGTHS. A number

of endurances limits and strengths for rotating beam tests are given in
Tables AT 3, AT 4, AT 6, and AT 10. Since, however, the possible varia-
AISI 6150, QT
tions in composition and heat treatments are infinite, a number of empirical
formulas have been suggested,"-?#! all of which are of limited application.
For wrought steel in its more commonly met commercial forms, it is often
assumed that the average endurance limit for an average s,, (50% survival)
is
(h) s) x 0.5s,, 5}, & (250)(BHN) psi, 5, © (0.25)(BHN) ksi,
[WROUGHT STEEL, BHN < 400; Goop bucTILiTy]
does follow better the trend of actual failure points. It may be worthwh
which should be limited to a maximum Brinell of about 400. There are
to observe from Fig. 4.8 that the difference between the Goodman lit
many exceptions that suggest caution. Cazaud"*-2*) quotes values for steel
CA and the Gerber line is about a maximum at OG and below, where Ot Showing s,/s,, ratios, often called the endurance ratio, from 0.23 to 0.65.
represents a repeated stress, R = 0, Fig. 4.5. Thus, if CGA is a bett The lowest values for ordinary steels go with untempered martensite, a
approximation of failure, one may wish to use (d) in design or give spe: Structure left by rapid cooling of medium- to high-carbon steel, as in
attention (which we do in § 6.6) to situations of approximately R = quenching (but the usual heat treatment includes tempering after quench-
As shown in Fig. 4.9, the difference between the Goodman and Soderbe
ing, which results in more typical values of the endurance ratio). If the
lines decreases to a negligible amount for high-strength, heat-treated steels microstructure is pearlite or austenite, a working value of endurance ratio
where the yield strength is close to the ultimate. would better be 0.4 or less (check some of the austenitic stainless steels in
If the foregoing equations were developed for a varying shear (tor- Table AT 4); but in the plain carbon, ferrite structure (for example, mild
sional) stress, the various lines apply and the equation analogous to (4. 1); and soft steel, quite ductile), s n/Sy May be greater than 0.6. In general, for
involving yield strength, is heat treated steels, the ratio s "1S tends to decrease with decrease of temper-
1 Ss ms S, as ing temperature,
by Fig. 4.10. and it drops : substantially when BHN > 400, as suggested
(e) aeN seu Soe eh Sus
oat
where s,, is the shearing endurance strength, s,, the yield strength in shear, ‘g m
Sms the mean stress, and s,, the alternating or variable stress in shear. ¢ a
The same equations are often applied when the predominant stress is FIGURE 4.10 The static s, continues to increase, but s,’ turns down. aie
3 80
108 0!
110 VARIABLE LOADS AND STRESS CONCENTRATIONS [Ch.¢ 4.8] STRESS RAISERS tt

Probably a better general-purpose estimate of om for high-strength allo; and endurance tests similarly show that the torsional endurance strength
steel, which is used in Germany, as reported by Lessells,*-*%) is 4g of steels ranges from about 0.5 to 0.6 of s’, with 0.58 being a good typical
value. In the absence of test values, consider that the endurance strength
(i) 5, = a8y + bs,
of a standard steel specimen in reversed torsion is
where a is a function of s, and is obtained by a straight-line interpolatioy
between the following values: a = 0.2 for sy = 85.3 ksi, a = 0.4 for sy = (k) Sns = 0.65, = (0.6)(endurance strength, rotating beam).
[rorsion]
190 ksi. Similarly, b is a function of s,, and its value is obtained by a stra
line interpolation between: b = 0.45 for s, = 85.3 ksi, b = 0 for s, =
But be careful. For the cast irons, the range is closer to 0.85n < Sy, < S$.3
ksi. Where the difference between the values from (h) and (i) for alloy s
for copper, 0.45, < sy, < 0.565).
is significant, the value from (i) is probably the better. f
For cast steel, it is safer to use s; ~ 0.4s, if test values are not available
for cast iron, use s/ ~ 0.35s,; for nodular iron, use Sn % 0.45, (but 0.3;
4.8 STRESS RAISERS. Any discontinuity or change of section,
for normalized). The nonferrous metals either do not have a limit or
variation of s,/s,, for different alloys is so great that arbitrary values such as scratches, holes, notches, bends, or grooves, is a “stress raiser.”’
It will result in a concentration of stress or a localized stress, that is greater
seriously in error. The endurance strengths of aluminum alloys at 108 cy
than the average or nominal stress. In some situations the theoretical value
change little with large increases in tensile strength. The copper alloys be
suited for fatigue, in order of fatigue strength, are: beryllium cop
phosphor bronze D, C, and A, nickel silver B, and silicon bronze A
See Ref. (4.31) for a large number of numerical values of fatigue stren,
The endurance strength of steel specimens subjected only to ax
loading (no bending) is usually not as high by test as the endurance lim
This may be because of the difficulty of applying axial loads with
eccentricity, because the material is not homogeneous, and because
stress is never truly uniformly distributed. Also, the presence of a st
gradient in bending and its absence in pure axial tension may be a fac
moreover, since an axial load subjects a greater volume of material to
maximum stress than a bending load, the probability of a high stress al
flaw or of slippage beginning at a weak crystal is greater. The litera
reveals a range of from some 0.6 to over 1 for the ratio of axial fatigu
rotating beam endurance strengths. The lower values are more typical
plain carbon steels, and the higher values of alloy steels. But the evidene FIGURE 4.11 Stress Concentration at Re-entrant Corners. The stress distribution as
is inconclusive. In the absence of test values, use i" indicated by the photoelastic method. The crowding together of the fringe lines at the fillets
Shows a high stress concentration. The larger the fillet, the less the theoretical concentration.
(j) Sn max = (0.8)(s, from standard specimen, rotating beam) But see § 4.10. In terms of the dimensions in Fig. AF 9, h/d = 2.97 and r/d = 0.0792 in
[REVERSED AXIAL LOADING] this picture (not all of 4 is shown). Observe the parallel fringe lines away from the
Vicinity of the discontinuity. (Courtesy M. M. Leven, Westinghouse Research Laboratories).
The ratio for Al 2014-T6, Table AT 10, is seen to be 15/18 = 0.83. O
tests for aluminum alloys" ©?! show a higher ratio (0.82-1.06). In situa’
where the procedure is justified, axial fatigue testing simulating the 4 a stress concentration can be computed by elasticity theory, or it can
ticular problem should be conducted. Only in a few cases is s, (0.2% off : © determined by various experimental techniques. Among these techniques
< s,; see AISI 321 annealed in Table AT 4. If this should be true, s the photoelastic method, which uses transparent models of various
would wish to have the maximum stress in a satisfactory relation to § Plastics. A monochromatic light passing through a loaded model in a
(see Fig. 4.6); refer to § 4.33. Photoelastic instrument emerges as black and white lines, as suggested by
The octahedral shear theory (§ 8.12) predicts that the shearing stre: j < 4.11. The dark lines are called fringes, and the magnitude of the stress
is about 58% of the tensile strength (elastic action). We recall that stat fs any point is a function of the number of fringes. Many of the available
tests of yield and elastic strengths are in good agreement with this facto €oretical stress concentration factors (§ 4.9) have been found by
 §4.10]) NOTCH SENSITIVITY 13

FIGURE 4.12 Improving Design by Photo- of stresses experimentally either in an actual model or, for example,
elastic Study. Especially in very irregular
as in Fig. 4.12. Figures AF 5-AF 14, inclusive and Tables AT 12, AT 13
and in indeterminate structures, photoelastic
studies of models of machine elements reveal in the appendix, give a good selection from available data. See also Table
points of stress concentration and the corres: AT 18 for curved beams and Fig. AF 15 for curved torsion members. An
ponding magnitude of stress; thus they are very excellent compendium of stress concentration factors is given by Peter-
helpful to the designer. In this illustration, son;“-2) others are in the [email protected],4.15.4.18.ete.] In choosing a
maximum theoretical stress at A in the ori
design was reduced by one-third at B in the
value of K;, always note the method of computing the nominal stress s,; a
revised design, and the weight of the part particular value of K; refers to a certain resisting area, which is usually
reduced 300 Ib.—less material yet great the minimum. For example, in Fig. AF 13, J/c and J/c make an allowance
fatigue strength. (Courtesy Chapman Labo. for the hole. Also observe that K, depends on the loading; it is not the same
tories, West Chester, Pa.). for torsion as for bending.

photo-elastic analysis, and the method is now commonly used to aid in the
design of parts that are difficult to analyze mathematically, as illustrat 4.10 NOTCH SENSITIVITY. The quantitative effect of a particu-
in Fig. 4.12. lar discontinuity in “‘raising’’ the stress is different for different materials.
Stress concentrations are significant for ductile materials only when t Some materials are more notch sensitive than others. To account for these
different responses, we use a sensitivity index q, called the notch sensitivity
loads are repeated. Since a ductile material under a steady load yields
of a material, Fig. AF 7, defined by
points of high stress concentration if the stress exceeds the yield stren|
there occurs a redistribution of stress and no perceptible damage to the Kviteyl
as a whole. The yielding is local (confined to a very small area). If the (4.3) 4.3 ae,
Lies or K; =1+q(K; -— 1),
load is repeated, however, the stress at the concentration points may exceed
the endurance strength, and if so, the part eventually fails by fatigue. where K; = (actual S,,,x)/So is the estimated fatigue-strength reduction
Stress raisers include holes, grooves, tool marks, surface roughness of factor (or fatigue notch factor).* Let the subscript f suggest fatigue. For
any sort as that from corrosion or pitting, keyways, welds, clamps (equiva= steels, fatigue tests on the coarser-grain steels (normalized or annealed)
lent to a sudden change of section—introducing additional stresses on t show low values of the notch sensitivity; fine-grain steels (quenched. and
surface), wrench marks, accidental bruises, quenching cracks, blow ho! tempered) have high values of g. While there is no sharp dividing line
and inclusions in castings, small fillets instead of large ones, point where between coarse and fine grains, 200 BHN may be used if the simplifica-
thread ends.'4:3! tion of two classes is desired (see below). Lipson and Juvinall'*-®5! state
that K, is never larger than 4 and seldom exceeds 3.
Peterson discusses notch sensitivity at some length in Ref. (4.1) and
4.9 THEORETICAL STRESS-CONCENTRATION FACTORS. derives various equations based on theoretical and test considerations.
degree of stress concentration is usually indicated by a stress concentration He recommends the use of
factor. If the factor is obtained from a theoretical analysis or a photo-
1
elastic study (photoelastic results are close to theoretical ones), it is call (I) ae eoSL
a theoretical stress concentration factor K,, which is the ratio of t 4 1+ajr’
theoretical maximum local stress divided by a nominal computed stre:
Where r is the radius of curvature (Fig. AF 7) and typical values of a for
So. Thus the maximum theoretical stress due to some discontinuity in t
normal stresses, with the corresponding (s,) in parentheses, are: O&T
section is 5.x = K,s, for a reversed stress (R = —1), where s, represents
the nominal stress computed from s, = F/A, 8, = Me/I, or s, = Te/J. Steels, a = 0.0025 (122 ksi); annealed or normalized steel, a = 0.01 (63
Inasmuch as changes of sections and discontinuities in general can occur ksi); which give the equations of the curves in Fig. AF 7. For other values
in an infinite variety of ways, it is not always possible to find a stress con= *
: Petersonl#.21) defines several K factors: theoretical factor, combined factor accounting
centration factor that applies to the case at hand. Thus, the exercise of Or failure theory used, shear-stress factor, fatigue notch
actor. Available evidence suggests that q for shear, say gs, is
factor, shear-fatigue notch
judgment with considerable estimation is often essential. It is also fre= higher than that for normal
Stress. Those who wish to make a distinction may use a for shear in equation (I) as 0.6
quently necessary to let experiment proceed with design, with determinations times the values given for normal stress;{4-1] a, = 0.6a.

112
114 VARIABLE LOADS AND STRESS CONCENTRATIONS [Ch. 4 §4.12] ENDURANCE STRENGTH AS AFFECTED BY SIZE us

of s,, we find for steel:'4-1) 5, = 93,.a =.0,005; s, = 180, a = 0.0005; 4.11 EFFECT OF SURFACE CONDITION ON ENDURANCE
Peterson shows a straight line relation between s, and log a. Other values _ STRENGTH. The factors for some unpolished surfaces are given by
of a: aluminum alloy 2024-T4 bars, a = 0.008, and for sheet, a = 0.05; Karpov’s curves,-4) Fig. AF 5, Appendix, which show that the effect on
aluminum alloy 7075-T6 bars, a = 0.003, for sheet, a = 0.02.'4:57) Since fatigue strength may be quite large. These curves include the notch-
there is considerable scatter in test results, one cannot be arbitrary about sensitivity effect, and they thus provide visual evidence that as tensile
the accuracy of q; some indicated values for fine-grain alloy steels are strength and hardness increase, the fatigue strength does not increase in
q> 1. In view of the scatter and their relatively high notch sensitivity, proportion; for example, an as-forged forging of s, = 110 ksi is little
many designers use K,= .K;, (q = 1) for aluminum (n, > 10°), magne- | stronger than the same forging of 50-ksi steel when the loads vary. See
sium, and titanium-alloys. The endurance ratio s,/s, for polished titanium also Fig. 4.10. These factors can be treated as fatigue-strength reduction
alloy specimens is usually greater than 0.55.'*:62) factors, right-hand scale, and used as K,’s; or the percentage number on
Cast iron, with its flakes of graphite, is effectively saturated with “stress the left-hand scale can be multiplied by the endurance limit to get a
raisers,’ so that the addition of another discontinuity seems to have little corrected endurance strength (Fig. AF 5).
effect on its fatigue strength;'+-6! that is, g ~ 0, at least for small sizes and” Decarburized surfaces result in lower endurance strengths also, the
radii. Like steel, its notch sensitivity tends to increase as the radius of amount depending on how much decarburization has occurred (see also
a fillet or groove increases and as the size increases, with K, approaching §§ 4.23, 4.28).
K, (full notch. sensitivity). Even for small parts with sharp radii, the
designer may wish to be conservative, using q ~ 0.2. Certain evidence?4
for class 45 cast iron shows that when the stress is computed for the net 4.12 ENDURANCE STRENGTH AS AFFECTED BY SIZE. The
section, a severe stress raiser (K; = 2.25) had little effect on the static evidence is conflicting, but most of it shows that the endurance strength
tensile strength and reduced the fatigue limit by only about 15%. (Inci- per unit area tends to decrease as size increases.'*:7-4-8.4-9.4.28.4.93] For
dental intelligence: this cast iron has about the same fatigue strength up example, wire has a higher strength than say a standard 0.3-in. specimen.
to some 800°F.) Thus, a general observation®-1) is that at least for the The reasons are not fully understood. There is the statistical aspect; with
lower-strength cast irons, say less than class 45, cast iron is not very no’ more volume of stressed material and more surface, the probability in-
sensitive. Designers tend to avoid brittle materials for fatigue loadi creases of a “‘weak”’ point developing. The larger the specimen, perhaps the
especially when impulsive loading is possible, but sometimes there is less uniform are the properties. The material may have different properties
desired economic alternative; moreover, ductility per se is not too imp because of different rates of cooling, for example. There may have been
tant a property for fatigue resistance."-64] More homogeneous. brittle unfavorable residual stresses. Whatever the reasons, we find the following
materials, such as glass, exhibit full notch sensitivity under static loads, in the literature. An 11}-in. axle of SAE 1045, with s), = 55 ksi, had an
well as variable, in which case K;, is applied to both the mean and alternating endurance strength of 11 ksi.» An alloy-steel crankshaft with s, = 120 ksi,
components of the stress. Recall the long-known technique of scratching showed endurance strengths in torsion as follows:-*! 0.55 in., 36 ksi;
glass to cause it to break along a certain line. 1.18 in., 31 ksi; 1.77 in., 26.5 ksi. A 7-in. rotating beam, with s), = 33 ksi,
Explanations of why K; is sometimes very much less than K; had an endurance strength 17.5 ksi. On the other hand, a 1-in. rotating
(1) highly local yielding redistributes the stress so that the theoretical beam, with s', = 32 for a 0.273-in. specimen, had s, = 33.8. A 9.75-in.
stresses are not reached; (2) the stress gradient, which increases as t shaft (s, = 62 ksi) with a fillet radius r = 0.305 in. had a torsional s,, of
notch radius decreases, relative to the grain size is related to this phenom 13 ksi at 8 x 108 cycles; for r = 1.5in., 5,, = 19 ksi at 11.5 x 10° cycles.
non. As a generalization, the radii of notches, fillets, etc. should be Limited evidence suggests that for axial loading, the size effect is not
large as possible; although q approaches unity and K, > K;,, K;, decrea: Significant up to 1.3-in. diameter.'-3%)
as the radius increases. The better part of valor, when the significant dimension is greater than
For ductile materials, the fatigue-strength-reduction factor (FSRF) about } in. for cyclic bending or torsional stresses, would seem to be to
applied only to the variable component s,, so that we think of the maximum reduce the fatigue limit by about 15°% for sizes up to 2 in. (arbitrary point)—
stress as , and, conservatively, for axial loading. Thus'*:2!

Smax = 5m + KySq.- i”
(m) Between }-2-in. size, let s, = 0.855’ and s,,=0.855'ns »
[BENDING AND TORSION]
This s,,,x on the Goodman-type diagram, Fig. 4.6, is represented by EC (as
a safe value); thus, pictorially, MD in magnitude represents K;5q. unless test data are at hand. Since large-scale testing is costly, appropriate
 G

Lene "leat oe Goodman §4.15] EXAMPLE—VARYING TORQUE 117

N 2 Safe Stress — el Safe Design Line Solution. From Table AT 9, for AISI 8742, QT 1200°F, we find s, = 130
i zs eS a~ for =OP/OA
N=OP/
ksi. Use the approximation s, ~ 0.5s, = (0.5)(130) = 65 ksi. For s, = 130 and
| Ky8o pee
from Fig. AF 5, we get the factor 0.81 for a machined surface; since the load is
A eee Pr ae me ie igbine axial, we apply the 0.8 factor (§ 4.7). The corrected endurance strength is then
N ek (65)(0.81)(0.8) = 42.1 ksi = s, in equation (4.1). Since s, = 0 for a reversed
load, equation (4.1) reduces to 1/N = sa/sn or Sa = S,/N. For the data given,
(a) Ductile Material — Soderberg. (b) Cast Iron — Goodman in+8» Region we have a design stress of 42.1/1.75, and

FIGURE 4.13 Working Diagrams for Ductile Materials and Cast Iron. In (a) and (b), ss teh a Hac SANG ag
5, Should be adjusted for the various factors that affect fatigue strength; surface factor may : “1.75 A wD4
be included in deciding upon s, or counted with K,. The plan in (b) for cast iron 2-1) sug. or D = 1.29 in.; use 1}in. (p. 32). Some available tests for a 1.25 dimension
gests that when the mean stress is negative, the alternating component does not need to be —
suggest that there is little size effect in the case of axial loads (§ 4.12). (NOTE. Since
diminished with decreasing s,,. To some extent, this plan may be permissible for steel, but {
confirming data are not plentifal. q
it is likely that the method of attaching the ends of the piston rod to the piston
and crosshead will introduce an additional stress concentration, the answer
obtained is too low.)
data are scarce for larger sizes, and one needs to use the best engineering
judgment (which will be based on the best engineering knowledge). ‘
A size effect, not so pronounced, has been observed also in static 4.145 EXAMPLE—VARYING TORQUE. A hot-rolled rod of AISI C1035 is
tensile tests, Table AT 8; hence, appropriate adjustment of the yield stress _ to be subjected to a cyclic torque varying from +3000 in-lb. to —1000 in-lb.
s, is advisable unless one is confident that the design factor N cares for The crank is attached to the rod via a sled-runner keyway; let N = 1.8. What
this eventuality. should be the diameter of the rod?
Solution. From Table AT 7 for C 1035, as rolled, we find s, = 85 ksi and
s, = 55 ksi. Direct data on shearing strengths are not available; so they are
estimated. In accordance with a note in Table AT 7,

Sys © 0.6s, — (0.6)(55) = 33 ksi.

The endurance limit is s, ~ s,/2 = 85/2 = 42.5 ksi; by § 4.7, for a polished
mean stress. Thus, equations (4.2) and (e) tae specimen, Sns © 0.65, = (0.6)(42.5) = 25.5 ksi. This value is now reduced by
the surface factor (60% from Fig. AF 5 for s, = 85) and the size effect (85%),
1 5 ys, provided that the required size is over } in. Thus s,; in equation (4.5) is
(4.4) ania ip ae
N Sy Sn Sas = (25.5)(0.6)(0.85) = 13 ksi.
From Table AT 13, we get K;; = 1.3. To use equation (4.5), we need the nominal
mean and alternating stresses, which are found from the mean 7, and alternating
(4.5) . = ae Sms Kj Sas
T, components of the torque. Paraphrasing equations (a), § 4.5,we have
WV Sire Sas
Pe Tmax : Tmin ie 3000 + = 1000) = 1000 in-Ib.
where K, and K,, are determined from equation (4.3); s, and sys (Sus =

prefer the modified Goodman line as a criterion. Equation (4.4) is repre- Tp RE

2
ete 2 oa OO Hel
sented by the straight line GD, Fig. 4.13(a). In application, the s, used is” The corresponding nominal stresses are then
the one that includes any weakening or strengthening surface effects, of
which some have been mentioned and more are discussed later. A design Te —See
(1000)(16) (200016) _ . 32,.
Sy, =
7 ie Pal =
aos ksi, and Sas =
=p» Pl ==spe ksi,
stress pattern for cast iron is shown in Fig. 4.13(b).
where J/c = Z’ = »D®/16. Substitute the various values into equation (4.5);

4.14 EXAMPLE—PISTON ROD. A piston rod is to be subjected to a maxi- 1 _ sme, KnSes

N Hi Sys 5 Sns
mum reversed load of 31,416 1b. It is to be made of AISI 8742, OQT 1200°F,
machined to 63-125 rms (§ 3.14). What should be its diameter for a design factor
[witsaa UG (1.332)
of safety of N = 1.75 if there are no stress-raising discontinuities and no ‘“‘col- 8 (#D3\(33) (D3)(13)
umn action’’? from which D = 1.28 in.; use 14 in.
 §4.17] EXAMPLE—LIMITED LIFE 119
AND STRESS CONCENTRATIONS [Ch. 4—
118 VARIABLE LOADS

E LIFE. The en- significance of the 10° is that the knee of the sm, curve is so often in the
4.16 ENDURANCE STRENGTH FOR A FINIT nite vicinity of 108 cycles, Fig. 4.3. Factors for surface condition and other discon-
men will have indefi
durance limit is that stress below which the speci been” tinuities should be applied to the result from (m) if s,19% is taken as s’ for
ance limit has not
life. We have observed, Fig. 4.3, that an endur polished specimens. It would be nice if equations (n) and (0) were always
such materi als, the design may
found for certain non-ferrous metals. For during” reliable, but unfortunately the knee does not always occur at 10° cycles
of stress to occur
be made on the basis of some estimated cycles and the actual curve does not always intercept the ordinate at 0.9s,.
there are many cases where the
the desired life of the part. Similarly, Some evidence suggests that the knee moves to the right as specimen size
stress in the lifetime of steel parts
number of repetitions of the maximum increases; for example, to 107 cycles for a 7-in. size.'*-22] Another relation
it were known that the number of
is relatively small. If, for example, that fits some data for polished steel specimens is
be 2 x 10° and that other
applications of the maximum stress would
the endur ance limit, we could use 6, 0,09
repeated stresses would be lower than
0 cycles obtained fro s, = s,(—]
the endurance strength corresponding to 200,00 (p)
in Fig. 4.3, or otherwise from esti Ne
fatigue-strength curves such as those
in design ing for finite life, while on the and this one, also based on the knee being at 10° cycles; it gives results
tion. The use of the endurance /imit
y expensive—certainly if
safe side, may be inadvisable or unnecessaril ;
similar to those from (0).
weight is important or if very many parts are to be
made. As instances of finite rather than infinite life requirement, J. O.
In Fig. 4.3 we observe a certain parallelism of the
curves for the steels Almen‘-12) points out that the number of cycles at maximum loading is
kind. (CAUT ION: There are steels for which the 100,000 for automobile differential gears and automobile suspension springs,
without stress raisers of any
average value that m
slope of the sn, curve is quite different from any and 500,000 for clutch springs.
large percentage | The fatigue strength reduction factors are less for finite life than for
be chosen.) Lipson and Noll?-*-2) found that in a
ed part of the line (with ground, machined,
cases, the inclin
hot rollec indefinite life, perhaps because local yielding redistributes the stress and
ed by straight lines © also work hardens, and thus strengthens, the local material. Observe the
forged, etc., surfaces) was fairly well approximat
cycles to the endura steeper slopes of the curves with stress raisers in Fig. 4.3. At the extreme
log-log paper going from an ordinate of 0.9s,, at 10°
of one gradual application of the load, the stress raiser has no significant
effect on the ultimate stress; in between, its effect would be expected to be
somewhere between 1 and K, for indefinite life. As a matter of fact, the
0.9 8. Ground Surface
curves for a heat-treated 1050 steel, with and without a notch, intersect
at about 10* cycles, at which point K, = 1; other such pairs of curves
% FIGURE 4.14 Endurance Strength of Steel intersect in the vicinity of 10° cycles. An experimental but satisfactory
relation for steel is'*:?)
for Finite Life. After Ref. (4.2).
ni‘ logK pis niosk pis

(a) Kp 1QlexK, RK} >

10° 10* 10° 10°

where Ky, is the strength reduction factor for a limited life of n < 10°
Cycles
cycles, and the other symbols are as defined above. See also Peterson.)
Above 500,000 cycles, the point would be near or in the probable range of
on of such a line to an en- the endurance limit, so one might as well base the design on the endurance
strength at 10° cycles, Fig. 4.14. The equati
is limit. A part designed for a finite life may turn out to have indefinite life
durance strength s,19°, that is, at 10° cycles, is
Is because the variation of properties resulted in a certain part or parts
0.81s,2
(n) log 5, = oe ; log —* log ne, Operating below their individual endurance strengths; but there exists the
n n economic advantage of the design having been made with a higher design
Stress. Interest in designing for limited life has been increasing and con-
which, if s,10° = 5), = 5y/2, reduces to siderable help can now be found in the literature.'*-41.4.4.4.47.4
591
(0) Sy = St (108 /n,)-%85,
[CERTAIN STEELS, NO STRESS RAISER]
which s, is desired, log is 4.17 EXAMPLE—LIMITED LIFE. A forged crank, Fig. 4.15, has a constant
where 1, < 108 is the number of cycles for
the usual meanings. thickness of 2 in., a length of 25 in., and a depth of 6 in. at a section B-B where
the base 10, and the other symbols have
120 VARIABLE LOADS AND STRESS CONCENTRATIONS [Ch.4 j $4.19] EQUIVALENT STRESS (21

The moment at section B-B is M = (F)(20) = 20F. Using N = 1.7, sp = 19.1

it was found necessary to place two }-in. holes, each 2} in. from the center
ksi and s, = 0 in (4.4), we have
line of the link. It is expected that there will not be more than 100,000 cycles of J
stress during the lifetime of the link, which is made of AISI C 1035. What is the — 1 _ 2.18)(20F\(2.625)
safe reversed load F, Fig. 4.15, for a design factor of safety of 1.7? NOY eae re nyg (19.1)(28)
Solution. As usual, first decide upon the stress situation. From Table AT 10 from which F = 2.75 kips or 2750 lb., a safe reversed load. The endurance
for 1035 in air, we find strength of this link can be materially increased (say doubled) with no change of
si= 40.6 ksi, s4/su = 0.46 or su = 40.6/0.46 = 88.3 ksi, sy = 58 ksi. nominal size by certain changes of the surface, as by machining off all the de-
carburized layer, or by cold working the edges of the holes, and least expensive,
We next estimate the endurance strength for finite life. From equations (n) ang
by shot peening the as-forged surface (§ 4.28).
(p), we have

Sn = antilog
ie (0.80 68.3)! = i log (0.2) 88.) log 10° | = 50.8 ksi,
4.18 EXAMPLE. Letting all the data of the previous example be the same
= 40.6 Go ry = 50 ksi.
+09
except that F varies from +3000 to —1000 lb.; to find the corresponding factor
of safety.
Use 50, the more conservative value. Also, we elect to use a surface factor for Solution. Paraphrasing equation (a) we find the mean and variable forces
as-forged, as well as the strength reduction factor K, for the hole ($4.22), to be

Fu = 20004 (— 1009) = 1000 Ib, F, = 20 = 100) = 2000 Ib.

then, with s = 20Fc/I,

wa OO 221880 pel, ea (2. ae -625) _= 8200p
Figure 4.15.
where the values of c, J, and Ky: have been taken from the previous example.
Using equation (4.4), we have (sy = 58 ksi)
1 pes Sm Kya wi 1.88 8.2
Nal gh toga eh Bei, komen tt
assuming that one stress raiser reinforces the other. For s, = 88.3 ksi, surface or N = 2.16, a larger factor of safety than in the previous example even though
factor = 0.45 from Fig. AF 5; for size effect, use 0.85 (§ 4.12); thus, s, in equation - the maximum load is greater in this one.
(4.4) is ‘
= (50)(0.45)(0.85) = 19.1 ksi.
4.19 EQUIVALENT STRESS. It is often convenient to have an
To get K., use Fig. AF 11; d/b = 0.25/0.5 = 0.5; e/d = 2.5/0.25 = 10; fre equivalent stress, one that when divided into say s, (or s, or s/) would
which K, = 3.5. Then from Fig. AF 7, q = 9.92 for r = d/2 = }4 in. and curve
indicate the degree of safety. Also the equivalent stress idea is useful when
A for coarse-grain, as-rolled steel. Because it is possible, use g = 0.89 and get, —
vd dealing with variable combined stresses. To obtain such a stress, one needs
K, = 1+ qK: — 1) = 1 + 0.89(3.5 — 1) = 3.22. only to multiply equation (4.4) through by s, and consider s,/N = s, as
Allow for a limited life of n = 10° cycles by equation (q); the equivalent stress:
Kn st = en ogden a, 10(? (2 loglos K,)/3
K,/3 = 2.18. Ss Ss
(4.6) Se = a = Sm + (2) K;jSq-
nv
For the nominal stress s, = Me/I, Fig. AF 7 shows that / for the net section 1S”
used. Since there are two holes, the moment of inertia is closely (notice that this The analogous equation for shear is
is not the exact value)
(r) 5 ee Sys =‘5 a (4) x
(h — 2d)°tNee_ eee
(6 — 0.5)9(2) es N = ms Sigs eee
fs’as
a tien epee 4
: 12 12 ook
and the distance to the top edge of the hole, Fig. 4.15, is c = 23 = 2.625 in. Since sy, ~ (0.5 — 0.6)s, and s,s © (0.5 —0.6)s,, then Sys/Sp5 & Sy/Sns
122 VARIABLE LOADS AND STRESS CONCENTRATIONS [Ch.
4 §4.23] RESIDUAL STRESSES 123

some values of which are given in Table AT 10. Analogously, multiplyin For live loads, proceed as detailed in this chapter. The following check
equation (4.4) through by s, gives an equivalent stress based on s,; list will be a reminder of the details.
Se = S,/N = SpSp/Sy + KySq, different from that in (4.6). (1) What is the basic endurance strength for the kind of stress involved—
reversed bending, torsional, or axial loading?
(2) Has the loss (or gain—see below) of strength because of surface condi-
tion been accounted for? It is always on the safe side to include a surface factor,
4.20 DESIGN FACTOR FOR VARIABLE LOADING. gn
Fig. AF 5, as well as a fatigue strength reduction factor K;, but see § 4.22.
stresses based on experience with over-all safety factors, as in Chapter 1, (3) Has the possible size effect been overlooked?
are invaluable criteria, but they are not always the most economical, (4) Are there discontinuities requiring an estimation of stress concentration
When such factors are used with the yield or ultimate strengths and via a K; and the sensitivity index q?
loads, they are necessarily large in order to cover large ignorance. Howeve (5) Will a finite life satisfy requirements? This affects the value of s, and Ky,.
when we delve into the various factors that affect strength and life, as i (6) Has the natural spread of properties been accounted for?
this chapter, and make proper allowances, a much smaller design fac (7) Have the kind and condition of material been considered? A casting may
is permissible, as low as about 1.4, although if it is this low based on have a concealed blowhole near a stressed surface. Properties are changed by
median fatigue strength and if all other operating conditions have b heat treatment. Directional properties may be significant, as from rolling,
well evaluated, the actual factor of safety of the part may be close to unit extruding, forging.
(8) Are the surroundings corrosive? See the correction factor for salt water
because of the natural spread of properties (§ 4.4). If the uncertainties fc
surroundings in Fig. AF 5.
a particular design are so great that (or if experience suggests tha
(9) Are the residual stresses helpful or harmful? This topic and several other
N > 3.5 to 4 is advisable, a more intensive search for more knowledge pertinent ones are discussed below.
in order, especially for important designs and large production, Som In your problem work, check off stress, surface, size, SCF, and life, as a
factor of safety greater than unity must be provided in order not to h minimum.
a failure during the required life, because it is necessary to allow not onl
for variations in properties, but also for variations in manufacturing pi
cesses, deflections in housings, warpage from heat treatment, etc., for 4.22 SUPERPOSED STRESS RAISERS. It often happens that
inadequacies of theoretical stress equations, and for unpredictable dep there is a hole or other discontinuity at a machined or other kind of rough
tures from the ideal model. There are additional factors for the desig: surface, and the question arises as to whether both strength reduction
to consider, but first a résumé of the discussion to this point. factors should be used. One set of torsion tests'*-5*) indicates the follow-
ing for AISI 4140, 350 BHN: (a) when the stress-raising effects are about
the same in magnitude, both should be included at full value; (b) when
4.21 RESUME OF DESIGN FOR VARIABLE STRESSES. Obse; one effect is significantly greater than the other, the contribution of the
that there are two kinds of strength reduction factors, those that are u lesser one is small, perhaps negligible. Another group of bending tests'*-35)
to reduce the endurance limit to a realistic endurance strength and th on 4340, 410 BHN, also indicated that the net effect of two stress raisers
(K;,) that are multiplied into a nominal s, to get an estimate of an act (K, = 2,38) is about 20% less than the product of the individual values
maximum stress. Many of the surface factors are in a sense stress conc for a rough machined surface (K, = 1.52) and a hole (K,; = 1.96). Genera-
tration factors, but there are a number of surface treatments (decarb lizations are not justified on the basis of the evidence, except to say that
tion, peening) that lower or raise the endurance strength without a str it is always on the safe side to count both effects full value, but that if it
raising discontinuity being involved. To treat surface factors in the sa 1s important and if one factor is larger than the other, some value less
way at all times, we shall use them as factors to reduce (or raise) endurane than the product can be justified.*
stress. This is in accordance with the way the examples are worked,
the adjustments include those in Fig. AF 5 and those for size and a
loading. For fillets, holes, etc., the adjustment is by K, > 1, as explain
4.23 RESIDUAL STRESSES. In a part at uniform temperature
(a) Brittle Materials. We have mentioned evidence (§ 4.10) t
and not acted upon by an external load, any internal stresses that exist are
indicates that cast iron has a low notch sensitivity index g. However,
Called residual stresses. Since static equilibrium obviously exists, the forces
brittle materials in general, use the full theoretical K, on both mean
from the residual tensile stresses are balanced by those from the residual
variable stresses.
(b) Ductile Materials. For dead loads, ignore stress concentrations * Another approach is to let the overall K; = KiiKi2; then Ky = 1 + q(Ki — 1).0*°851
 124 VARIABLE LOADS AND STRESS CONCENTRATIONS [Ch.4 — §4.23] RESIDUAL STRESSES 125

compressive stresses; also the internal shear forces are in equilibrium, and the residual stress distribution is as shown dotted at EF, Fig. 4.16.
These stresses exist because of what has happened in the past history of The deformation may be such as to enlarge the ID permanently about 6%,
the part (as welding, rolling, heat treating, shot peening, work hardening, — and the OD about 1%."-1°1 Now when the cylinder is subjected to a
etc.), and they may be beneficial or harmful. If a tensile member is loaded pressure, the tangential compressive stress at E first begins to decrease,
somewhat beyond the yield point (§ 1.8), what actually happens is that becomes zero, and finally becomes tensile, with some distribution CD,
the surface material reaches the yield strength s, before the core does; Fig. 4.16, where the peak stress is lower than before. In application to
when the load is removed, the outside layers of material that received a guns, the idea was to make a higher internal pressure possible without
permanent deformation will have residual compressive stresses produced producing a permanent deformation, that is, without the maximum stress
by the residual tensile stresses in the core, which stresses may or may not exceeding the approximate elastic limit. The cold working by autofrettage
be helpful, depending on how the part will be stressed when it is in its also raises the elastic limit of the material. For vessels that are subjected
allotted place. Whatever the source of the residual stresses, they can be to repeated pressures, the residual compressive stress on the inside im-
eliminated in whole or in part by a suitable tempering or annealing, called proves fatigue strength, because it is the excessive tensile stresses that
stress relieving, which is what should be done if the residual stresses are produce the fatigue damage.
expected to be harmful. Residual stresses of the order of 32 ksi for as-cast If a member has a residual compressive stress on its surface, which is
cast-iron wheels have been reported,@-1! but such high values in this case subjected to repeated tensile stresses from an external load, its endurance
are not typical of good design and good engineering practice. iN strength may be much higher than when the surface is residual-stress-free.
More and more, residual stresses are put into a part intentionally for The only trouble is that there are no commercial (economic) means of
the good they will do. Autofrettage is a process of prestressing or over= determining directly what these stresses are and what their quantitative
stressing a hollow cylindrical member beyond the elastic range by hydrau effect on endurance strength will be. However, if the variable working load
pressure. A long-standing use of autofrettage is as applied to gun barr is always applied in the same sense, parts with a stress raiser can be

A,,

4 -
&
- | Stress Distribution
Mi
Residual with Residual
|
(a) ma (d)
FIGURE 4.16 Effect of Residual Stress. FIGURE 4.17 Overstressing, Presetting (Diagrammatic). Part (a) could represent
diagrammatically a leaf in an automotive spring; such springs are commonly prestressed.
Part (b), a C-frame, and part (c), a torsion spring, react as described for (a); but in these
Cases, there is a stress-concentration effect because of the curvature. The net stress for the
tensile member in (d) is the algebraic sum of the dotted residual and the stress distribution
Shown solid; this reduces the magnitude of the peak concentration stress at the edge of the
hole.

point of maximum normal stress, as at B, Fig. 4.16. (See § 8.26.) If the Overstressed in that same sense, generally with improvement in endurance
cylinder is subjected to a pressure that induces stresses into the plastic Strength as suggested by Fig. 4.17. For example, in Fig. 4.17(a), if F always
range at some distance through the wall thickness from the inside, the acts downward, the Mc/I stress distribution is as at section AB, where the
inside material takes on a permanent deformation. After the pressure 1s line AB represents zero stress. If a load large enough to produce plastic
released, the outer material that did not get much if any permanent defor- yielding isi applied and then removed, some pattern of residual stresses as
mation, contracts elastically exerting pressure on the material with perma- indicated at CD remains. The net stress at a particular point in a particular
nent deformation, so that the internal tangential stresses become compress Section with the working load on is: the Me/I stress algebraically plus the
 §4.24] PLATE WITH ELLIPTICAL HOLE 127

usually a residual tensile stress.-?!_ The effect of decarburization alone,

smooth specimens, is illustrated by the values for SAE 2340:"°-?! hardened
to Ro = 28, it lost 47% of its endurance strength (83 to 44 ksi); hardened
to Ro = 48, it lost 71% (122 to 35 ksi). Coupled with observable surface
(a) (b) © (a) defects, the damage is worse. The depth of decarburization seems to be
FIGURE 4.17 (Repeated).
unimportant, as it takes only a weak surface (from any cause) or one with
a residual tension to serve as a source of fatigue cracks and lower endurance
residual stress (s, negative). This is seen to reduce the resultant tensile stress strength. The loss of fatigue strength for limited life, say 10° cycles, is not
on the bottom fiber and therefore it increases endurance strength. This idea so great; the loss of strength for 1 cycle (gradually applied) is unnoticeable.
of overstressing may be used as an extra precautionary step where the opera- _ Steps taken to combat decarburization include heat treating in a non-
tion may be close to the endurance strength and when experience (testing — oxidizing atmosphere; and for forgings, etc., where such control is not
or service) indicates an appropriate strength improvement is thereby ob- — practicable, recarburization, surface rolling, peening, and hardening
tained. If F is a reversing load, Fig. 4.17, prestressing as shown results ina ~ (S§ 4.28, 4.29); and removing the decarburized layer by machining.
weaker part, because the residual s, and the s, = Mc/J will add numerically —
for one direction of F. It is reasonable and logical to presume that if the — 4.24 PLATE WITH ELLIPTICAL HOLE. This case is important
residual stress at a point is such that the residual stress plus the stress by — because it throws some light on the stress at cracks, which are roughly long,
the external loading exceeds the yield strength of the material, yielding — slender ellipses. If the load is perpendicular to the major axis of the ellipse,
results in some relaxing of the residual stress. Hence, if a part has been — Fig. 4.18, the worst stress condition exists, and the maximum stress occurs
overstressed for the purpose of inducing favorable residual stresses and ~ at the ends of the major axis where the theoretical K, is'4-24)
then if some loading (perhaps accidentally) induces stresses that remove —
part or all of the favorable residuals, the original improvement in strength © 2a
K,=1+ ze
is thereby lost.
A point that is frequently overlooked is that stresses induced in a part 4 a is half the axis that is at right angles to the load and 3 is half the other
by the assembly operation have effects similar to residual stresses. If any — axis. Observe that when the ratio of a/b is very large, tremendous stresses
such induced stresses are tensile, fatigue failure is more likely. are indicated by K;; but see § 4.10. This observation accords with the rapid
Quenching heat treatments leave residual stresses. Quenching in water —
is a more drastic cooling than quenching in oil; the part warps more and the — i)
residual stresses are much larger. The surface of a solid piece in contact © Sais bot
with the quenching medium cools first, and as the inside cools and shrinks, — |
residual compressive stresses (tangential and longitudinal) are left in the —
outer layers. (A hollow tube quenched from the outside generally has residual —
compressive stresses on the outside and tensile on the inside.) Medium — FIGURE 4.18 BF
|
carbon (or better) steels have a much higher residual stress if water quenched —
than if oil quenched; for example, 0.49% C steel, 2-in. bar, quenched from — —

1560°F; in water, tangential residual stress is about 70 ksi; in oil, about — 1

1 1
35 ksi.{-10 The size of the part naturally affects the numerical results. The ~
tempering operation reduces these stresses by an amount that depends on ~
the tempering temperature, the length of time held at this temperature, ~
and the rate of final cooling. The foregoing piece cooled in air from a tem- spread of cracks which are at right angles, or nearly so, to the load. If a crack
pering temperature of 1200°F has a tangential residual stress of s, = 6 ksi. — such as that at B, Fig. 4.18, were 100 times its width (a/b = 100), K, = 201,
Not only does quenching and tempering tend to leave a favorable Suggesting that a relatively small repeated load might cause failure when a
residual s, on the surface, but it also improves the mechanical properties — flaw equivalent to a crack exists in the material. The spreading of a crack
generally. Nevertheless, there are pitfalls in heat treatment, especially for— can often be stopped by drilling small holes at each end, as at C, Fig. 4.18.
the high-strength steels. If it is done in an oxidizing atmosphere, the surface — This step materially reduces K;,.
will be decarburized—one of the weakening factors also in hot-rolled and — On the other hand, if the ellipse or crack is parallel to the direction of
as-forged parts. Decarburization not only leaves a weaker material but the stress, so that a < b, the effect of stress concentration may be negligible.
128 VARIABLE LOADS AND STRESS CONCENTRATIONS [Ch. 4

4.25 BEAM WITH HOLES. The significance of the concentration N

of stress resulting from a hole in a beam depends upon the location of the”
(a)
hole. If the hole is on the neutral axis and not too large, Fig. 4.19(a), no
reduction whatsoever may occur in the fatigue strength of the beam. On FIGURE 4.20 Fretting Action in Force Fit. In (a), a varying torque causes the amount
the other hand, if the hole is near the external fibers, where the stress of twist to vary, and an element AC (unstressed) may take a position BC (much exaggerated)
naturally large, a stress much larger than the Mc/J stress at the extern under load. The distance A to C for a given fit, through which slipping occurs, depends on
fiber may occur at the edge.of the hole, Figs. 4.15 and 4.19. Particular ¢ the amount of torque. If the torque is large enough and if the shaft does not fail, the shaft
will turn in the hub, point C moving through C’ to C’’ and beyond. Maximum slipping in
should be taken in locating holes in curved sections of a beam that is to
apy case occurs at one end of the fit, MN here. If the hub is made less rigid at this end by,
subjected to repeated loading, for example, at a point where an axle is ben say, cutting away as indicated by the dotted lines at M and N as MP, then the hub may
The neutral axis of a curved beam does not coincide with the geome’ move more with the twisting shaft, thereby receive the torque 7 with less relative movement
axis but is nearer the inside (concave) surface (see § 8.25). Holes locate and less fretting. If the member has a varying bending moment on it, or if it is a rotating
near the geometric axis of a curved beam may result in progressive failure. shaft with a constant moment, as in (b), then a point just inside the hub will have a varying
stress which is the same as saying a varying deformation, and the change of deformation
results in a slight relative motion and possible fretting. The heavy lines at the entrance to
the fits represent a surface area that should be surface rolled for improved strength (§ 4.28).
If the shaft is subjected to a bending moment or a varying torque at both ends of the fit,
4.26 CORROSION. Even the normal atmosphere reduces the ens surface roll both ends. Grooving the hub as at Q in(b) also improves the fatigue strength.(*°°*)

cites an interesting case; a brass cup formed in a forging die of course

contained complex residual stresses. Placed in an ammonia atmosphere, it
the corrosive effect. Some materials are much more resistant to corros
broke into several parts without external load within 2.5 hr. because of
than others, and resistant coatings may provide considerable improvemen
tensile residuals; a similar part peened all over was intact after 100 hr. in
In noncorrosive environments, coatings often lower the endurance streng
of a part, but they can help more than harm in other circumstances. gaseous NH3.
on steel takes the galvanic action, and thus small corroded points whe:
corrosion-fatigue cracks can start do not appear. Other advantage
coating materials, especially for steel, include galvanized, cadmium plated, 4.27 FRETTING CORROSION. When two touching surfaces have
and enamel. a high contact pressure and when these surfaces have minute relative motion
Corrosion-fatigue cracks generally progress at right angles to the max (as in some force fits (§ 3.8), bolted and riveted connections, leaf springs,
mum tensile stress, and the combination of corrosion and fatigue loadin: bearing races, splined connections, etc.), a phenomenon called fretting
more detrimental than either alone. Some copper alloys are subject corrosion or just fretting often occurs. Even with the smoothest surfaces, a
corrosion fatigue, others are more resistant. Stainless steels deteriorat microscopic chafing causes small particles of metal to tear loose and oxidize.
this respect in chlorides, particularly in boiling magnesium chlor (Fretting occurs also in the presence of an inert gas—Fenner.":?®!) In the
Chromium is more effective than nickel in providing resistance to corrosic case of steel, the oxidized particles form a reddish brown powder (if dry),
fatigue, but all types of stainless are likely to become pitted in sea wate which, once the process has started, acts abrasively and hastens the deteri-
Compressive residual surface stresses inhibit the start of corrosion fatigi oration. Typical events at an interference fit are suggested by Fig 4.20, but
cracks.“-1! The endurance strength reduction factor K, as predicted t vibration of assemblies with joined parts is a common source of fretting.
equation (4.3) is not so large in the presence of corrosive action because ‘ The damage is greatest with dry surfaces, but no known lubricant stops the
the overwhelming effect of corrosion. All in all, materials for parts beif action entirely. The damage increases with the load, the slip, and with the
designed for corrosive surroundings must be chosen with care. Almen® total number of oscillations.*-28! The magnitude of the effective fatigue-
Strength reduction factor cannot be accurately predicted, but a study of
the literature will give one some ideas; for example: on a 7-in. as-forged
FIGURE 4.19 Stresses at Holes in Beams.
Shaft, K, = 3; the fatigue strength of an SAE 1025 steel was reduced from
B 40 ksi to 20 ksi by fretting;:-2! cold-worked aluminum alloy with an
€ndurance strength of 20 ksi at 5 x 10®cycles, had s, ~ 7 ksi with fretting."*-34)
e P. With fretting corrosion, even for steel, there may be no distinct
129
Neutral Axis
130 VARIABLE LOADS AND STRESS CONCENTRATIONS [Ch.4 §4.28] SHOT PEENING AND SURFACE ROLLING 131

endurance /imit. The difference between fatigue strengths with and without fretting, notably molybdenum disulfide MoS,. For dry steel on dry steel,
fretting is negligible when the life is low, say less than 10* cycles. fretting appeared in less than 100 cycles; with the surfaces coated with a
It is not certain whether or not fretting can be entirely suppressed. mixture of MoS, and grease, fretting did not appear until more than 1.5 x
Lower pressures and reduction of relative motion are obvious things to 10® cycles; and if a mixture of MoS, and corn syrup is rubbed on and
consider, but these factors often cannot be changed significantly. The most baked until dry, it was over 9.8 x 10® cycles before fretting.:31! Anodizing
effective help is obtained from a residual compressive stress on the surface aluminum surfaces inhibits fretting.
most prone to fretting failure. See § 4.28. ‘
Surface rolling is a means frequently used to prolong fatigue life. Steel
axles of SAE 1045 had a fatigue strength of about 14 ksi with fretting cor- 4.28 SHOT PEENING AND SURFACE ROLLING. Both of these
rosion (40 x 10® cycles), and about 33 ksi after surface rolling.":?! Une processes stress the surface, on quite small areas at a time, beyond the tensile
rolled magnesium had s, = 7 ksi at 5 x 10® cycles versus 21 ksi when yield strength, producing a local permanent deformation. On the spring-
rolled (24 ksi when regular—‘regular”’ not defined).©-?! Certain heat back, the underneath adjacent fibers that had no plastic flow tend to re-
treatments leave favorable residual stresses. Horger ‘28! shows that wher sume their original dimensions and in doing so, produce residual compres-
the material is slowly cooled (in furnace) from a temperature of 1160° F
sive stresses in the plastically deformed surface. Both processes are cold-
most of the helpful residual stresses from quenching having been relieved; working processes and in general increase the local mechanical properties.
but if water quenched from 1160°F, the residual compressive stresses are Thus, any resulting increase in endurance strength can be attributed partly to
large enough to increase strength markedly. Citing numbers, we have higher s,, but the principal benefit comes from the residual compressive
(1) For 9.5-in. OD, 0.51% C steel shafts; normalized and tempered 4i
stress.
1160°F ‘and furnace cooled; 85 x 10° cycles; s, = 91.2 ksi; an enduran
In shot peening, a rain of metallic shot (say chilled cast-iron pellets)
strength of 11 ksi (versus s, ~ sy/2 = 45.6). K
(2) For same size, same material and heat treatment, except that a
impinges at high speed, on the surface, perhaps on selected areas as ona fillet.
treatment is a water quench from 1160°F, fixing favorable residuals; endur: The plastically deformed portion extends inward from several thousandths
strength > 19 ksi, more than a 70% increase. of an inch to a few hundredths, and the amount of the cold working, as
(3) For 9.5-in. OD, 3-in. ID, 0.51% C steel shafts; quenched from 1550 would be expected, depends principally on the plastic work done by the
tempered at 1000°F; 85 x 10° cycles; s, = 125 ksi; an endurance strength oO} pellets, which in turn is dependent on the size and speed of the pellets
12.5 ksi (versus s, ~ su/2 = 62.5); residual stresses all small. (mv?/2) and the total number of impacts. This is not to say that improve-
(4) For same size, same material and heat treatment, except that a final tr ment continues with continued energy input, because there is a counter-
ment is a water quench from 1000°F, fixing favorable residuals (63 ksi on o1 acting effect in the damage that bombardment does to the surface. Besides
side); endurance strength of 18 ksi, a 44% increase. leaving the indentations, which are stress raisers to a degree (surface
(5) For the shafts in (3), except that the tempering temperature was 7
roughness of the order of 140 pin. (65-200 in.), arithmetic average—see
(residual s, = 50 ksi), s, = 22 ksi or better, showing that a tempering temp
Fig. 3.9), the impacts may cause cracks in the surface, one or more of
ture of 750°F does not remove the favorable residual stresses.
which betomes the source of early fatigue failure. The full improvement
Other means of leaving residual compressive stresses are mentione
to be obtained in fatigue strength is soon gained (say, in 2 min.), and ‘“‘deep”’
below. Repeated stressing in the vicinity of the endurance strength
either change the pattern of the residual stresses or, if thespecimen was 0
peening is not needed except for a reason, as in overcoming the effects of
nally stress free, it may induce residual stresses. The suggested explana’ decarburized surfaces. For each material and its state, there is some opti-
is that where the high stresses occur, there will be plastic flow (local yieldi mum combination (with the present state of knowledge, we do not always
Fuchs"-1! states that the yield strength for repeated loads is below know what the optimum is) of shot size and speed, and duration of peening,
static value. See Rosenthal,-1! Horger,-?! and Sigwart.%:28! Witht So that the engineer cannot simply specify ‘“peened surfaces’? and be co-
fident of the consequences, unless he knows that the production department
on three other alloy steels (various combinations of vanadium, mo
denum, nickel, and chromium), Horger’s results!*-28! suggest that s has the know-how for the particular part. Moreover, after peening, there is
steels do not have greater endurance strength or longer life if fret no commercial inspection of the actual part less expensive than sample
occurs. For example, a Ni-Cr—Mo alloy shaft as in (3) above, with s, = 1
testing to assure that the peening is proper. To test the peening operation,
ksi, had s, = 9.5 ksi with fretting. One side of a thin test strip of steel, Rp = 47, is subjected to the same stream
There has been some success with coatings on the surfaces subjectedt of shot as is the part, the strip takes a curvature, peened side convex, and
 }-— Working Loads —-
| ! Peened
Peened (a)
<=c 220ksi 3 Mr ( e178 ke ») 2
as)
P x

te
2 = —30 kei and 8, = — 140 ksi
8 =0 and & = +42ksi
(a) Presetting (b) Tensile Strain Peening (c)Compressive Strain Peening
FIGURE 4.21 During peening, the strip is firmly attached to a heavy
8 backing M, suggested by the dotted outline in (a), The residual s- results in FIGURE 4.23 See Table 4.2 (a) Presetting or overstressing. Held for 30 sec. In (b) and
Peened the strip bending as in (b) when released from M. (c), stresses 5; and 5; by M are elastic; s, = residual stress near surface that is placed in
(b) tension by the working loads, top side in illustrations.

the amount of the arc height J, Fig. 4.21, is a measure of the peening, called
the Almen intensity. Because of the annealing effect, the improvement due show that it is more effective than smooth-finishing forgings.'¢-6*) Since
to any kind of cold working begins to disappear when steel is heated above peening is a low-cost means of increasing fatigue strength, it is often used
500°F; for aluminum, above 250°F. on forgings, heat-treated items, springs, and hot-formed parts in general,
Typical distributions of the residual stresses in a high-strength stee that naturally suffer some decarburization; it is used sometimes instead of
are shown in Fig. 4.22; SAE 5147, Ro = 48. Notice the large increase of machining or grinding as a means of improving strength. Where the cost is
compressive residual stress over that left by the heat treatment, and thi justified, decarburized surfaces are recarburized, and peened too. Troubles
fact that the maximum compressive stress occurs a little below the surface. from service failures have been cured by adding peening to the regular pro-
This means that the dimples produced by peening may be machined of duction processes. Mattson'*-?6! reports that the shot used on hardened
4-58
polished off to the extent of about 0.002 in. (if it is worth the cost), steel does not need to be harder than the steel; that on hardened steel,
without losing the strengthening effect of these high residuals. Evidenth small shot with some minimum intensity will produce the optimum
the penetration of the effect varies somehow with the size of shot, and this strengthening; that for softer steel, larger shot and severe treatment tends
may be significant. If the penetration is not far enough, the fatigue failure to give optimum strengthening, perhaps because of greater penetration and
may start at some point below the surface, as 0.006 in. below for curve 2 greater response to work hardening.
Fig. 4.22. Since fatigue failure is prone to start on the surface, or close toi Mattson and Roberts'*-37] report on an interesting series of tests. The
(§ 4.2), with a suitable surface stress, failure beginning much below” specimens were of spring steel, SAE 5160, machined to 0.192 by 1.5-in.
surface is unlikely unless there is some stress raiser (inclusion) present (but cross section, heat treated to Rockwell C 48 + 2. All specimens were preset
see nitriding in § 4.29). If the curves of Fig. 4.22 are carried deeper into (prestressed), a common practice for leaf springs in the automotive industry,
the interior, balancing tensile stresses would be shown because the internal which consists of bending the spring leaf cold to some radius r, Fig. 4.23(a),
forces are in equilibrium. The effectiveness of shot peening improves that takes the external fibers into the plastic range, and thereby leaves resi-
the hardness of the steel increases. However, shot peening a decarburi dual stresses in the unloaded piece (§ 4.23). All testing was by repeated
surface brings about a good, sometimes spectacular, increase in stren bending, R = 0. The peening was done in such a way as to obtain five
provided the peening effect extends through the decarburized zone (whic different levels of residual stress, in addition to that in the preset specimen.
is thereby limited to about 0.04-in. thickness) where the material is muck If, for example, the piece is subjected to a constant moment M while it is
weaker. Peening may be better than recarburizing the surface, and test: being peened on the surface in tension, Fig. 4.23(b), the compressive residual
Stresses due to peening on this surface when the part is unloaded are aug-
mented by the elastic spring back.
FIGURE 4.22 Residual Stresses from Shot Peening.
Steel, SAE 5147; R. = 48. (After Mattson, R. L.{*79]) In one experiment, the unit tensile strain during peening was + 0.006,
/ (s = cE = 180 ksi); the residual stress reached about —175 ksi, vs.
As Heat Treated 7
about = 128 ksi for conventional peening (piece unloaded). By peening the
ag with the compressive stress, the spring back may leave a residual
av Stress there, as seen in Table 4.2 (p. 134), which gives a summary of
Residual Stress, ksi

“Stee of comparative tests. The effect of the residual stresses is so con-

ee that one feels sure that if there were an easy nondestructive way to
oe Stresses, many unexpected troubles might find a ready
Dac ton. The authors state that the improvement in endurance strength
ost entirely due to the residual stresses, not to work hardening.
We find the following in the literature. Spring steel, 0.77% C, 0.67%

133
 134 VARIABLE LOADS AND STRESS CONCENTRATIONS [Ch. 4_ §4.29] HEAT TREATMENTS FOR IMPROVING ENDURANCE STRENGTH 135

Mn, 0.28% Ni, 0.22% Cr, OQT 752°F::?8! original machined finish,” If the designer is dealing with a shop knowledgeable in these processes
5, = 39 ksi; with surface polished to 7 pin. finish, 0.0025 in. removed, it would certainly seem safe to assume in design at least a 25% improvement
S, = 49 ksi. SAE 1020:'-?) as rolled, s, = 28 ksi; polished, s, = 35 ksi; in fatigue strength as a result of peening or surface rolling under circum-
as rolled and shot peened, s, = 37 ksi. Carburized Ni-Cr—Mo steel stances where it is known to be beneficial.
carburized and heat treated :'-?) surface as received, s, = 58 ksi; polished
surface, s, = 69 ksi; shot-peened surface, s, = 71 ksi; surface shot peened ”
and then honed, s, = 74 ksi. For highly stressed parts that are expected 4.29 HEAT TREATMENTS FOR IMPROVING ENDURANCE
to have a finite life, peening can be expected to prolong life significantly. STRENGTH. Besides quenching and tempering for through hardness,
For example, an unpeened spring wire failed in 105,000 cycles at a stress of which increases strength, there are several surface hardening processes that
+ 100 ksi; peened, it did not fail in 10’ cycles. Polishing of the peened significantly improve endurance strength and wear resistance. All of the
surface should further increase life; at 135 ksi, a peened steel spring had a processes below not only make the surface material stronger (s,,), but they
life of 60,000 cycles, and peened and polished, 10° cycles. leave residual tangential and longitudinal compressive stresses in the surface
Surface rolling* is a process that cold works a limited amount of layers. See § 2.8.
material, thus giving it higher strength, and it leaves a surface compressive (a) Flame Hardening. Flame hardening is widely used for local heat
stress—as does peening—a stress of the same order of magnitude, except treatments, such as fillets, bearing surfaces, gear teeth. Naturally, when the
that the compressive residual stress usually goes deeper, perhaps as much surface becomes hot, it first expands, then loses strength and yields; the
as }in. at times."-?) In general, it is more expensive than peening, but is hot layer transforms to martensite on sudden quenching. The naturally
greater volume of the martensite results in residual compressive stresses.
(If cooled slowly, the martensite would not be formed and the residual
TABLE 4.2 EFFECT OF RESIDUAL STRESSES"?! stresses would be tensile.) SAE 1045 steel, ;4-in. specimens showed the
following strengths;%-2] untreated, s, = 18 ksi; OQT 400°F, 27 ksi;
CONDITION OF SPECIMEN APPROX. SURFACE ENDURANCE 0.001-in. fillets flame strengthened, s, = 32 ksi (failure not at fillet); fillets
(ORIGINALLY Ro = 48) RESIDUAL STRESS, s, ksi | STRENGTH, ksi
and entire reduced section flame strengthened, s, = 51 ksi. Lessells'-25!
¢ = —0.006 (peened on compression side) . +42 55 found: 1-in. alloy steel bars, 60° V-groove with bottom radius of 3; in.,
¢ = —0.003 (peened on compression side) . 0 78 endurance strength doubled by flame hardening, from 40 to 80 ksi, in
Heat treated only (Rc = 48) 0 88
reversed bending. At a press fit, a 9.5-in. flame-hardened shaft had 63%
Preset only . 3 i —30 128
Shot peened while unstressed —128 140 greater endurance strength than when normalized and tempered, but surface
€ = +0.003 (peened on tension side). —140 176 rolling produced better than 100% increase (85 x 10® cycles). Fretting
€ = +0.006 (peened on tension side). —175 194 may start early for flame-hardened fits, but the large residual compressive
stresses retard the process. There are few quantitative data. Certainly
the flame should not be oxidizing.
(b) Induction Hardening. When flame and induction hardening are
convenient and appropriate for local treatments of round members, as at handled properly, the quality and properties after treatment are much the
interference fits (Fig. 4.20), fillets, and grooves. Horger'°-?} reports the Same. Induction hardening is likely to be economical only on a production
following: for a normalized and tempered, 9.5-in., SAE 1050 steel, the basis, because of the need of special machines, and is well suited to harden-
permissible design bending stress at the fit is 11 ksi without rolling and ing cylindrical surfaces, especially bearing surfaces on crankshafts, cam
22 ksi or higher with surface rolling,:?} a 100% improvement; rolling Surfaces, gear teeth, etc. A cylindrical specimen with a transverse oil hole
fillets increases the fatigue strength by 30 to 68 %. ; had s, = 10.2 ksi not induction hardened and S, = 7.7 ksi induction har-
Since polishing leaves a residual compressive stress that may be upwards dened,'*-10] which emphasizes the fact that some fine-sounding processes
of 15 ksi'*-6*)__most beneficial when polished in the direction of the load— Cannot be used blindly. The trouble is a residual tensile stress inside the
a polished tensile member gains little strength by cold working the surface. hole, not too far below the surface, which combined with a repeatedly
Approximately, a polished surface may be some 10% stronger in fatigue” applied tensile stress exceeded the endurance strength. This same piece
than a stress-free member. induction hardened and shot peened had s, = 10.9 ksi, which is a con-
Siderable improvement.
* Not to be confused with cold-rolled steel, which generally connotes a relatively large
The depth of case influences the magnitude of the residual stresses and
change of dimensions (as in cold drawn).
136 VARIABLE LOADS AND STRESS CONCENTRATIONS [Ch, 54.30] MISCELLANEOUS SURFACE EFFECTS 137

presumably the endurance strength. As the depth of a hardened layer ir js so frequently used for the principal purpose of obtaining a good wearing
creases, the residual stress at the surface increases to a peak and then turr surface, the strength often is secondary; in many of these situations, the
down. If the part is afterwards tempered, the fatigue strength decreases= part has excess strength. Selected surfaces can be carburized, by putting
from say 90 to 81 ksi for a tempering temperature of 300°F; from 90 t yarious coatings over the parts not to be carburized, for example,
61 ksi for 480°F. In fact, for some steels, 480°F may bring about a suk prushed on copper sulfate, or copper plating.
stantial relief of stresses, from 74 to 19 ksi in one case.'*:1°! Thus, if residug (d) Nitriding. This produces results similar to those from carburizing,
stresses are important, so is the tempering temperature. Depths of harder put the residual stresses are higher and the percentage increase in strength is
ing as much as 0.15 in. or more may be advantageous. ‘ generally greater. Also nitriding has an advantage of much reduced or
(c) Carburizing. Absorption of carbon into the case increases ij negligible distortion because severe quenching after the process is not
volume, the case material is transformed to martensite, and the core ¢ required. However, it is not effective on decarburized surfaces. One series
tracts last on cooling, thereby inducing high residual compressive stre of tests suggests that nitriding increases the endurance limit of 0.3-in.
on the surface. Also, the increased carbon content of the case results ini diameter pieces, 0.037-in. case, by about 20-25 ksi;!°-?! other results are
mechanical properties being increased. There seems to be some question ; as follows:
to which factor is most significant in accounting for the improved ¢
durance strength. It is not certain what the stress distribution is in
vicinity of a discontinuity, and since the stress state is surely triz SPECIMEN UNNITRIDED, S, ksi NITRIDED, 5, ksi
there may be residual tensile stresses. However, if the stress concentrati¢ Bar without notch . : : é 45 90
on the surface is localized within the carburized material, the high strengt With semicircle notch 3 Fe { 25 87
of the high-carbon steel is sufficient to result in a large improvement| With V notch . 3 5 5 a 24 80
fatigue strength. Alloy steels with their better hardenability can be ma Bar (1-in.) without stress raiser . 3 67.5 71
With fillet . , ; : - 32 67.5
hard enough by oil quenching and hence they distort less than plain carbe
steels (water quenched), so they tend to be favored for carburizing, ¢
tainly where no finishing operation occurs after treatment. Some tests ind When stressed in the vicinity of the endurance strength for indefinite
cate a very large increase in strength: a ;3,-in. AISI 2317, in rotary bendif life and failure occurs, the fatigue crack of a nitrided part typically starts in
0.05-in. case, showed s, = 48 ksi normalized, and s, = 120 ksi carburize the core just inside the case material, where the residual stress is tension;
water quenched and tempered; the same as above except that the materi when highly stressed, the failure starts at the outer surface.-1° Failures
was 2513, s, = 54,and s, = 123 ksi.!*:25) j originating in the core also occur when the other surface treatments are
Some less optimistic results are: 0.3-in. diameter, 0.03-in. case, 0.2% used, especially if the case is unusually thin. As usual, K; > 1 as the
steel, rotary bending, carburizing raised s, from 33 to 45 ksi;'-5! bars y number of cycles to failure decreases (stress increases). Notice the effective-
a radial hole, untreated, rotary bending, s, = 48 ksi vs. 62.3 ksi 2 ness of nitriding notches for Nitralloy 135, Table AT 10. If a thin strip is
carburizing on surface and in hole; same bars, same treatment, had nitrided throughout, it will increase in length about 2% for the Nitralloys
increase in reversed torsional endurance strength from 17 to 41 ksi. (about 6% for 4340) because of the diffusion of additional matter. When a
Directly comparative data are scarce; through hardened 4140 shafts thin layer on a heavier part is so treated, the change in length is prevented
in 105 to 4 x 105 cycles, whereas 4320 shafts carburized with a case by the heavy core, resulting in compressive stresses in the case.
0.04-0.05 in. failed in 4 x 10°to8 x 10° cycles, at the same stress level.
A shaft with a hole drilled after carburizing had s, = 29.9 ksi, but drille
before carburizing, s, = 62.6 ksi. 4.30 MISCELLANEOUS SURFACE EFFECTS. The effects of
The effect of carburizing depends to some extent on the case thicknes Plating, as with copper, nickel, chromium, cadmium, tin, vary considerably,
If the case is very thin, failure often starts near the junction of case am but usually the fatigue strength of a part is reduced by plating. The plating
core. Some tests show fatigue life increasing with case thickness up to Process can be adjusted so that the deposited metal has a residual compres-
in. ;'*-31] other tests on fatigue bending strength of gear teeth showed Sive stress,'*-1.4.64] in which case, the fatigue strength may be little affected.
strength increasing to a case depth of 0.008 in., then decreasing gradu Ordinarily, the plating process is such that the deposited metal has residual
as the case depth increased to 0.06 in.™-?®! Considering the informatio tensile stresses, which means that the surface is weak in fatigue. If the de-
available, it would appear that there are no comprehensive guide line Posited metal cracks through to the base metal, the tension in the deposited
for design stresses; each design is a special case. Inasmuch as carburizifi Metal tends to put the base metal in tension at the point of the crack, which
138 VARIABLE LOADS AND STRESS CONCENTRATIONS [Ch.

is also serving as a stress raiser. Other contributing factors include pene.

tration of hydrogen into the steel (hydrogen embrittlement) and the fac
that the deposited metal may be weaker than the base metal. Peening maj
be used to offset the loss of strength. For example, a steel with s), = 46 ksi
had an s, = 19 ksi when nickel plated and s, = 55 ksi when plated ang (e)
then peened; when chromium plated, s, = 38 ksi, and when peened an¢ FIGURE 4.24 Relieving Grooves Reduce Stress. When there is a single groove such as at
then chromium plated, s, = 51 ksi. While the evidence may not be con (a), the stress concentration is greater than when it is flanked by two, preferably smaller,
clusive, peening before plating tends to retain the original fatigue strength grooves G, as at (b). The relieving grooves G in (c) reduce the concentration about the
peening after plating tends to result in greater strength than the original ba S radial hole. These grooves are more effective if they are pressed in. Undercutting a sharp
internal fillet as in (d) or providing relieving grooves G in (e), or both, are effective. Under-
metal. Surface rolling produces effects of the same order as peening. ]
cutting the fillet leaves a face for locating a bearing or other element. (Battelle Memorial
operation is in a corrosive environment, proper plating that protects thi Institutel***}),
underlying steel will maintain s, the same as the plated part in air: buti
cracks in the plating permit the corrosive medium to reach the steel, s, am
life will be much reduced. Reference (0.2) has acollection of quantitatiy strength. Peening raised the endurance strength of a rough ground flat
fatigue data on various metal coatings. steel bar, reversed bending, from 42 to 82 ksi.{-:1
Cold drawing and cold rolling (§ 2.9) result in all the material bein In most cases, machining leaves residual tensile stresses in the surface
plastically compressed. When the load returns to zero, the material expands layers; exceptions include austenitic manganese steel and cast iron.
However, after the stress in the outer fibers reaches zero, the internal fiber Unexpectedly good results with carbon steel have been obtained by
are still in compression; hence, with expansion continuing until the interné shallow quenching, which means that the steel has low hardenability (§ 2.7).
forces are in equilibrium, the outer fibers end up in tension (in both longi For example, water-quenched 1046 steel will have a surface hardness ap-
tudinal and tangential directions) and some inner ones in compressior proaching 600 BHN, but because of poor hardenability, the hardness falls
The residual tensile stress, sometimes of considerable magnitude (60-12 off to about 280 BHN within 0.25 in. of the surface.) This treatment
ksi), would be expected.to affect unfavorably the endurance strength. On produces a residual compressive stress at the surface (the hardened surface
set of tests on 1.5-in. bars'?-1) shows tangential and longitudinal stresses % material tends to occupy more volume) and residual tension on the inside,
the outermost fiber of about 48 ksi tension, and at the central fiber, c a pattern that results in a large increase in fatigue strength for members in
pressive stresses of 45 and 80 ksi, respectively (1045, cold drawn 20 bending and torsion. Excellent results have been obtained in this manner for
Thus, any improvement in fatigue strength would be due to the hight automotive axles and for other heavy duty situations.
mechanical properties from work hardening. Stress relieving to remove tk
tensile residuals would also remove some or all of the work-hardenin
effects; if heated above the recrystallization temperature, all effects of col 4.31 MITIGATING STRESS CONCENTRATIONS. The designer
working are generally lost. Shot peening cold drawn surfaces can cha ig may specify a certain fillet radius, but the shop may make it smaller;
the residual stress to compression and significantly improve its endurane or the shop may inadvertently leave a stress raiser that the designer has
a frequent practice with coil springs. not counted on, such as clamp marks or the merest change in diameter
Evidence'*-®4! suggests that grinding may leave surprisingly large residui at the junction of two machining operations. Such things have to be
tensile stresses in the surface of an unstressed part and otherwise injure watched. However, the designer too may be responsible for the presence of
(grinding cracks), the residual tension resulting from the high temperatur an unnecessary stress raiser. All designs should be looked at with the ques-
induced in a thin surface layer. Carburized and nitrided steels may los tion “can this stress concentration point be eliminated?” in mind. If elimi-
endurance strength to some 35% of the unground value™-?! with impropé Ration is impracticable, then consider what can be done to reduce its effect.
grinding methods, and only a few data show any improvement in streng We have already indicated many basic considerations and we shall mention
after grinding. For case-hardened steels, the loss of strength is due to t Specific steps in connection with some machine elements. In the meantime,
substitution of a layer with residual tension for a layer that had high a few pointers are suggested by Figs. 4.24-4.27. In Fig. 4.26(a), typical
residual compression. Soft wheels and light cuts are relatively less damaging,
but more expensive in removing metal than rougher cuts. There is the posst FIGURE 4.25 Effect of Collars. Narrow collars
bility of some alloys with certain heat treatments being resistant to grinding Teduce stress concentration. (Battelle Memorial
damage,°-2] and peening and tumbling tends to restore the origina Institutel4.31),
 4.32] TEMPERATURE EFFECTS 141
than the ultimate strength, and most of the failure points at high tempera-
ture On aN SqS, diagram fall outside of a parabola (Gerber curve, Fig. 4.8)
through s, for reversed stress and the rupture stress with Sq = 0. See Fig.
(a) (b) 4.28. This is to say that equation (d), § 4.6, is a reasonable design basis,
that for a given mean stress, the material can withstand a much higher
FIGURE 4.26 Rolling Bores. alternating stress before rupture than that predicted by the Goodman (or
Soderberg) line.*-2®) Thus, if the temperature is such that creep is involved,
proportions for simple links, with the holes not rolled, s, = 12.7 ksi; witk a conservative design basis would be the straight line from the static stress
the holes rolled, s, = 12.1 ksi; surprisingly, no improvement. In Fig for a specified safe creep strain (instead of from $,/N) to s,/N, where s,
4.26(b), with the holes not rolled, s, = 17.6 ksi; with rolled holes, s, = 2€ is the endurance strength in reversed loading at the operating temperature
ksi. In (a), the collar raised on the edge of the bore by the rolling was no (and theoretically at the frequency of the actual loading). As is true of
removed; in (b), it was filled off."°-?! See Fig. AF 8 for a sample value o creep rupture strength, the fatigue strength at a particular frequency, is a
K;, for a bar being loaded by a pin in the hole. Reference (4.62) contains th function of time; that is, for example, the fatigue strength for 105 hr. is
most complete collection of K; values covering various configurations © higher than that for 10° hr. In a general way, with increased temperature,
the same basic design. Cold pressing grooves about holes, as in Fig notch sensitivity decreases and the effect of shot peening, so beneficial at
4.24(c) and 17.24, improves fatigue strength. room temperature, decreases. Any one starting design work for high tem-
perature situations will find the summary of the state of the art by Allen and

Tot Cae! Cee! Gd

Forrest'*:?8! helpful; considerable data on high temperature properties
are in Ref. (2.1), Notch effects at high temperature are not consistent with
(a) Lap Joint (b) Single Shear (c) Plain Scarf (d) Double Scarf those at room temperature.*:5°!
When a hot part is cooled suddenly by quenching, there is momentarily
FIGURE 4.27 Improving Fatigue Strength by Design. After Ref. (4.30). The load ¢ a high temperature gradient that induces a stress gradient. Some metal
(b), (c), and (d) varied from 10.7 to 21.3 kips; material, aluminum alloy 7075-T6. N parts under certain conditions crack as a result; this phenomenon may
be
that all surfaces under pressure may have fretting corrosion. Tapers are relatively 1 called a thermal-shock failure. If the change is not so severe, repetitions
expensive to machine. (a) No comparative values. (b) Failed in 42,000 cycles. (c) Fe of temperature and stress gradients in metals may be enough to cause
210,800 cycles. (d) Failed in 26,914,000 cycles. Spaulding!*:1) reports stress concent
eventual failure, a process referred to as thermal fatigue. It is reported
factors, based on the gross section, of 13 for (b), 4.1 to 8.5 for (c), 3.2 for (d). For re by
loading, it is evident that (b) should be avoided. A similar design, except a tongue and g ee — that the parameters Sn go9/Ea and KSngoo/ Et indicate
construction with double shear of the bolts, had K; = 4.1, a sizable saving. \ of heat-resisting alloys according to their ability to
withstand thermal fatigue; spgo9) = endurance strength at 800°C,
k = the
thermal conductivity, « = the coefficient of thermal expansion,
4.32 TEMPERATURE EFFECTS. As temperatures go below norm E =
modulus of elasticity. To a limited extent, local heating has
atmospheric temperature, the fatigue strength tends to increase; for€ been used to
Induce favorable residual stresses. When a transformation does
not occur
during cooling, the part that cools first is in compression.
at 10° cycles;'*:28! normalized SAE 2330 has an s) = 59 ksi at 75°F,
si, = 110 ksi at —320°F. In general, all steels show .a similar increase
endurance strength, but those without nickel alloy lose nearly all of the FIGURE 4,28 Curves of Mean Stress ys. Alternating Stress, from Tests. Adapted
from
Ref. (2.1), The A286 alloy consists of about 55% Fe, 15% Cr, 26% Ni,
plus small amounts
Of several other metals; this metal was stressed by
relation between impact strength and fatigue strength. a combination of axial and bending stress.
loading on the Stainless 403 was
With increasing temperature, the effect is generally the reverse, exce| axial,
axi Each curve represents failure b i
hr. at the temperature indicated. E Oe at Gens
that plain carbon and very low-alloy steels increase in ultimate and
durance strengths from 70°F to about 600°F or more, then strength di
creases rapidly. AISI 4340, unnotched, with s,, = 160 ksi at 70°F, has €
durance strengths for reversed loads with temperature as follows:™
70°F, 70 ksi; 600°F, 63 ksi; 800°F 60 ksi; 1000°F, 40 ksi.
142 VARIABLE LOADS AND STRESS CONCENTRATIONS [Ch. <4.34] IMPACT 143

4.33 MORE ON FATIGUE STRENGTH. If one is designing fo Also aluminum, copper, lead, and other nonferrous metals are found to
variable loading and using a material for which no appropriate fatigui have endurance strengths that change with the frequency. Space does not
data are available, one must purely gamble on the consequences or ruj permit elaboration, but if one is involved with unusual frequencies, the
tests to acquire information. For some machines, like airplanes, uninforme matter should be investigated.
design is never appropriate. Materials sometimes exhibit unexpected pect
liarities that actually seem illogical. For example, in some axial fatigu
tests with R = 0 (tension) on 17-7 PH and A-286 (a heat and corrosior
4.34 IMPACT. Suddenly applied loads, called impulsive loads, pro-
resistant steel), the K, value exceeded K, when the number of cycles wa duce such complex responses that any procedure of applicable design
10° or more.'-33]_ Whether the explanation involves the resultant effes involves considerable uncertainty. Two bodies are involved, a striking
of the biaxial and triaxial stresses that exist at the bottom of the notch, ¢ body or load that produces the effect and the struck body, or one being in-
compared to the simple uniform stress of an unnotched specimen, the res vestigated, that has a response; keep them mentally distinct. The loading
dual tensile stresses, or whatever, the fact is pertinent to the design of is considered to be impact or shock when the time taken for the response
part whose life is expected to be 10 cycles or more. 4 to reach a maximum is less than the lowest natural period of vibration of
In some cases, the actual repeated peak stress may be permitted to € the struck body. If the impact is repeated, some kind of impact fatigue
ceed the yield strength, when the number of lifetime cycles is less than som strength would be involved. If the rate of application of a varying load is
number, say 10* (but the fatigue strength for said number of cycles mu increased, the frequency of the applied load will become greater than the
not be exceeded), and when the possible change of dimensions (which mj frequency response, and the case passes into the range covered by the defini-
be unnoticeable when the highest stress is highly local) is permissible. Wht tion of impact loading. Thus, we see that the two cases, fatigue and impact,
the rate at which the load is applied exceeds some 500 cpm, 5,,,, may exces merge into each other at their extremes; but in their usual ranges, they are
Sy without actual yielding.-®°! Perhaps the normal operation is such th handled quite differently. The usual method of handling impact problems is
most of the time the peak stress does not exceed the endurance strengt by the laws of the conservation of energy and the conservation of momen-
but if the stress does exceed the endurance limit, the piece is dama tum. We have every confidence in the law of conservation of energy; as it
The damage apparently accumulates and, if expected, must be all applies to this situation,
for. '4-44 4.46]
(4.7) eae given | ie pone absorbed &
Some materials exhibit a marked change with the orientation o
fibers; for example, 4340 showed an endurance strength transverse to by loading body the loaded body
fiber length of 70-75% of that when the longitudinal dimension of t
specimen accords with the longitudinal direction of rolling.'*°*! Unk provided no significant energy is exchanged with the surroundings. How-
ever, there are some practical difficulties in calculating all of the energies
specified otherwise, quoted values of s,, are expected to be for longitudin
4 involved, and the ideal models that permit computations are oversimplified.
specimens,
After a part has been designed and a sample made, experimental methods
The rotating bending-fatigue strength for only a few cycles, say less h
that measure-42] velocity, acceleration, time, and of course stress, can be
1000, is greater than the ultimate strength; for 4340, the endurance streng
used to check; appropriate modification and redesign follow.
as computed from Mc// is approximately 1.6s, for a small number
cycles.'#331 Observe that the stress induced is in the plastic range, th The duration of an impact may be very short, even a fraction of a
thousandths of a second, and impact induces vibrations that affect the
Mc/I is therefore not the true stress, and that the stress distributi
probably some modification of that shown in Fig. 1.7. é
amount of the induced stress.'*-40-4-41] One helpful phenomenon is that the
strength (and ultimate too) increases considerably as the rate at which
In some tests, fatigue strength is increased by repeated loads just belo
oad is applied increases. When the unit strain of a certain mild steel
the normal fatigue limit, followed by small step by step increases of
ns tested is increased at the rate of 10-* per sec, the vicinity of the con-
loading, a process called coaxing. A test on Armco iron (s, = 26.2}
Bone, testing rate, the s, = 31 ksi; when the rate is 10%, sy = 79
was: an initial run of 107 cycles at 26 ksi, with about a 2% increase
107 cycles W ee this is closely where elastic action ceased. Another conservative
stress every 107 cycles; the stress at failure after nearly 13 x
. in pies is that points of support are usually assumed to be rigid,
30% greater than s,. Improving fatigue strength by understressing clo
| rs eas all actually deform to some extent, the consequence being that the
to the endurance limit or by coaxing is not a commercial way.
ual stress is lower than the computed value. We shall assume that all
For ordinary rates of applying repeated loads, steel shows no significal
‘formations are elastic in accordance with Hooke’s law, force is propor-
change of endurance limit, up to say 8000 cps. For very slow rates,
tonal to deformation.
10-100 cpm, and for very high rates, variations have been found.4+
144 VARIABLE LOADS AND STRESS CONCENTRATIONS [Ch 44.36] BAR AXIALLY LOADED 145

4.35 ELASTIC ENERGY. If an elastic body, say a spring, is d energy), and (f) the mass of the weight is large compared to the mass of the
formed an amount 6 under a force F that has increased gradually from zer rod (which makes the vibrational energy negligible in amount). Let W,
the spring’s response is also F, and the average force is F/2. (See Fig. 6.9 Fig. 4.29, fall freely from rest through a distance h, strike the stop M on
The work done on the spring (and the energy stored in the spring) is U the rod, and come to rest after the rod has stretched an amount 5. Assuming
(F/2)8, or
ort LLL.
FS ké&
(4.8) (in-Ib. or ft-Ib.),
co |
(eae
2
where 8 is the total elastic deflection at the point of application of the for
FIGURE 4.29 Impulsive Loading.
w7 He?
F and k = F/6 |b./in. (or lb./ft., depending on the units of 5, but in Rod—] h
junction with s psi, inch units are more convenient); k is a common pai M.
meter called the spring scale or spring constant. Observe that (4.8) appl
to any kind of elastic member, when F oc 8. If the member is a beam, o eas
Sooo ee
thus must have the deflection of the beam at the point of application of
see Table AT 2. 9
Recall also from mechanics that work done by a constant tore that the loss of potential energy W(h + 5) is all converted to elastic energy,
T is TO; and if the torque varies linearly from 0 to 7, the work is ( we have

t 24L
which is representative of the elastic energy stored in a cylindrical bar FS sAd_ ké?
torque on it is gradually increased to T. (4.10) Wh + 8) = — = — = — = — >
( ) 2 2 2 2E
T0 [a] [B] [c] [p] fe]
(4.9) U= in-lb. or ft-lb.,
where the symbols are as defined in the previous article. From the equality
of terms [A] and [D], solve for 6 and find
~+—( =) 1/2
crank at a radius of r, T = Fr and U = Fr6/2 is the elastic energy s (t) §6=—+ 1a Se
in the crank and bar (6 including deformation of crank). If the ¢ k k W
almost rigid, the bar stores nearly all of the energy. If a prism
in which 8 can be replaced by 8 = eZ = sL/E, W/k = 8,, is the deflection
member is in uniform tension or compression, the amount of energy ] of the rod under a static load W, called the static deflection, and h can
unit volume stored at any elastic stress s is the area under the stress-sti be replaced by its equivalent kinetic energy v2/2g,, where v is the speed of
diagram (Fig. 1.3) up to that stress. This is a triangular area, ()es = 5? W at the instant it contacts M (or from another point of view, one can use
which, if multiplied by the volume AL, takes the form of (4.8) above. 7 the kinematic relation v? = 2g,h). Conveniently, g, = 386 ips?, vips, E
energy-absorbing capacity of a material is related to its ductility, Psi, k Ib/in., s psi, etc. Useful expressions for the scale k that may be used
strength, and its strain-hardening characteristics. 3 In (t) (tensile or compressive member) are

4.36 BAR AXIALLY LOADED. Consider Fig. 4.29; assume that: (

(u) put Git Bs.
8 eL L
the support at G is rigid (no deformation), (b) the weight W and surfa
of the stop are rigid (these assumptions mean that the bar takes al If we solve for s from terms [A] and [C] in equation (4.10), we find
deformation), (c) the weight has been brought to rest (maximum deform
tion), (d) the stress s pervades uniformly the entire rod (it actually mo (v) Ss in A (; \3 + 1) IY,
waves—see below), (e) the system of weight and rod is a constant-¢
system (actually some energy is dissipated on impact, which first goes Where we see that for particular values of A, W, and h, the stress can be
raise the temperature of the parts and then heat is dissipated to the $ Teduced only by increasing 8 = sL/E, which may be done by increasing
roundings—if everything started out at ambient temperature; also fora given material, or changing material to one with a lower E. Most
of the energy in the rod is vibrational energy which means less elas Materials with E lower than that for steel are likely to have lower yield and
a LLLLL LS

54.39] TENSION MEMBER WITH TWO OR MORE AREAS 147

L
: | FIGURE 4.29 (Repeated). load W suddenly applied induces a response or stress twice that for a static
load W. If the load is applied gradually rather than instantaneously, the
maximum stress will be less than twice the corresponding static stress.
In the limit, the load can be let down so gradually that the stress never
exceeds the static stress.

endurance strengths too, so a change of material from steel is not too ofte 4.39 TENSION MEMBER WITH TWO OR MORE AREAS. When
the solution of a tough impact problem. See also § 4.39. a member with an axial load has two or more different sectional areas, the
If we solve for s from terms [A] and [E] in equation (4.10), we get deformation of each constant section is determined, and the total deforaias
(w) es “(i .: a tion 6 is the sum of the values for each section. Suppose a member saya
Ww s=—_ — ——_——— 5
bolt, has an area A, of the unthreaded part and a smaller area A a the
A A LW
threaded part. The nominal stresses due to some load F = ee SgAo
where W/A is the stress induced by a static load W, sometimes called t will a inversely proportional to the areas, s,/s, = A,/A,. Also, since
static stress s,,; also WL/(EA) = 8,;, the static deflection. As we see, a lar e = S/L,

number of equations differing in details can be obtained by virtue of t]

interrelations of laws and properties. For this reason, the student is advis (x) [Constant E]
not to use equations (t) or (w), for example, in his problem work, but
work from the more basic numbered forms; the practice in using the bai Suppose it is desired to find the stress s, in the smaller section of a two-
forms is too valuable to be lost by substituting numbers in an equation. section rod, Fig. 4.30, subjected to the impact of a falling weight W. The
total deformation is
4.37 EXAMPLE. What should be the diameter of a steel rod 6 in. long if it
to resist the impact of a weight of W = 500 Ib. dropped through a distance 5yL, + Sole
2in.? The maximum computed stress is to be 20 ksi. a (y) 5 = e,L, + €oL, => zB
Solution. From the given maximum stress, compute the correspondi
maximum deflection; Substitute s,; = Ags,/A, from (x) and find

oh AgSgLl,/A,+ Salo 52 [ Ag
(z) 5
Ceet G Bt , La).
From equation (4.10), terms [A] and [C], solve for A and find
Taking advantage of the convenience of the spring constant k = F/8, we
aD? e 2Wh + 8) it (2)(500)(2.004)
Aatte
4 58 (20,000)(0.004)’
rest LLLLLLALLLLLL.
from which D = 5.65 in.; tse 52 in. Thus, a startlingly large dimension
from rather guileless data.
Ly 1

4.38 SUDDENLY APPLIED LOAD—ZERO VELOCITY OF f 2 FIGURE 4.30 +—4

PACT. If the weight W is held in contact with but not pressing On 2
stop, Fig. 4.29, and is then released so that the load goes suddenly fr Le
zero to W, the distance dropped h = 0 and equation (v) yields
A+
h
2Ww
$= ——
A
ba!
Suddenly released, the weight vibrates as it would on a spring until y shal find the equivalent or over-all spring constant k’ for the
tensile (or
vibration damps out and the stress becomes the static stress W/A. TB Ompressive) member with two sectional areas; k’ = F/8, where 8 = 8, +
!
LLLLLL 1 LLLLLL.

Ly, 1
FIGURE 4.30 (Repeated). FIGURE 4.31 Bolts Suitable for Absorbing Energy.

(a)

8, is the total deflection, equation (z), and where F = 5,4, = S2Aq is th e for bending or twisting. The modifications of Fig. 4.31 will not change sig-
response load in the member. Thus, nificantly the static strength; hence, there is no reason to use such bolts
except for live loads.
1 ae 1 kik One group of tests indicate that the capacity of the bolt to absorb
(4.11) k’ ‘
mere I Sa aS Uae
eer energy continues to increase as the diameter of the body of the bolt is turned
to less than the diameter at the root of the threads—at least to where the
BO SE) Kee” Ka
body area is 57% of root area.-?1 But see Chapter 5.
where k, and k, are the spring constants for the parts 1 and 2 respectively,
Fig. 4.30. This rod is equivalent to two springs in series. In general, k’
4.41 MASS OF ROD NOT NEGLIGIBLE. If the end of a rod is
1/Z(1/k), where X(1/k) is the sum of the reciprocals of the individual
struck with a hammer, a compression wave is set up that travels along the
Now, the energy absorbed by the rod is U = k’5?/2, equation (4.8),
is equal to the energy given up by the falling weight, W(h + 8). This eque rod at the acoustic speed v, in the rod. From physics, recall that v, =
(E/p)*/?, in which a consistent system of units* must be used; that is, if
results in equation (t) except that k’ is in place of k. By using the value
8 in terms of s, from equation (z), we may find the maximum stress as before the pound is used for force and inches for length (E psi), then
mass and density expressed in pounds must be converted as follows:
This is a devious but illuminating route. Observe that making part 1,
4.30, larger, the deformation is decreased and therefore the part is weakene lb/(386 ips”), where g, = 386 ips, the standard acceleration of gravity.
Another phenomenon of this hammer blow is that the material under the
for handling energy loads.
hammer is actually moved, given a speed, which is taken as the common
speed v, of the hammer and adjacent material at the end of the deformation
LOADS. Faced with a shock load period. This first layer of material at speed v, imparts speed to the next
4.40 DESIGN FOR IMPACT
layer, and so on. Thus, in Fig. 4.32, the first layer moves with speed »,,
one’s first thought should be, ‘“‘Can it be eliminated ?” If not, the second is,
and the layer at any other point C later moves with speed v. If it is as-
“Can it be reduced?” as by shock absorbers on automobiles and aire ft sumed"2] that this speed varies as the distance from point D, for instance,
landing gear or by other means. If it must be lived with, the design criteri
we can estimate the kinetic energy, using v/x = v,/L. The element dx at
that should be kept in mind is that the material must have an approp
BC, Fig. 4.32, has a mass of (density times volume), dW, = pA dx/g,,
ability to absorb energy without failure, which, as we have seen, is rel
where p Ib/ft.° divided by 32.2 is p/32.2 slugs/ft.?; its kinetic energy is
to the amount of deformation. For example, suppose a bolt is subjecte
mv?/2 = pAv*dx/(2g,). (Consistent inch units are just as satisfactory here.)
an impact load. The stress on the smaller area at the root of the thread
the larger and critical stress. The unthreaded shank has a lower stre *
consistent system of units is defined by saying that it is one in which the propor-
.

(A, > Ag), and for this reason its unit strain ¢, is less than ¢, for the sm ap constant in Newton’s second law is unity; that is, it is defined by F = ma. If
area. If the unthreaded shank is turned down to, or somewhat less tha oe orce is in pounds, time in seconds, and length in feet, then from m = F/a, the units
= mass are m-lb/(ft/sec.?) = Ib-sec?/ft., which unit is called a s/ug. If the mass is in
the root diameter of the thread, or if a hole is drilled in the unthreadet spe 5. the conversion to the foregoing consistent system is w lb/g. = (w Ib.)/(32.2 fps?)
shank, Fig. 4.31, the bolt is strengthened for impact loads, because with | € conversion factor is 32.2 Ib/slug. There is no generally recognized name for the
higher stress in the unthreaded part, the deformation there is greater. Mass unit in a Consistent pound-second-inch system (Ib-sec.?/in.), as above. Suggestion:
Psin, acontraction of pounds, seconds, inch; conversion constant is then Zo = 386 lb/psin.
The method illustrated in Fig. 4.31(a) is cheaper, but the bolt is weak
for an external bending or twisting load. The idea of Fig. 4.31(b) involves
difficulties in manufacturing in long bolts, but there is negligible weakening
FIGURE 4.32 Kinetic Energy of Struck Bar. The com-
Pression (stress) wave with speed ¥, is reflected and rereflected
148
from the ends of the rod until it dies out.
 4 v ee 4
4 <— Fovccy 4

‘at
150 VARIABLE LOADS AND STRESS CONCENTRATIONS [Ch.4
Agee
4 -L—> j baie t B 4 <p
This expression can be integrated over the entire volume by substitutin, (a) (b) (©) (a)
v = xv,/L; FIGURE 4.33 Horizontal Impact.

Aap ph ADIL): We tt
“KE, == =| eee ee ft-lb., or in-lb..*
energy. Suppose, in Fig. 4.33(a), that a body W acted on by a constant
2g,L 0 685 3 280 ;
force P strikes the member A at the instant its speed is v and that P con-
from which we see that the kinetic energy of the bar KE, is as though one
third of the bar were given the translational speed v,. This portion of th tinues to act while A and W are in contact. Let P remain constant after
contact, Fig. 4.33(b). Then body A must absorb the kinetic energy that W
whole can be called the effective mass W,; W, = W,/3 for the bar with th
has at the instant of impact plus the work done by P during the deformation
axial impact loading.
of A; thus
The law of conservation of momentum applies; momentum of stri
Energy absorbed by A = KE of W + work of P
load W is equal to the momentum of load and rod the instant after contac
Equating momentums, we have F8 Wo?
(e) + PS.
Vw
2 2g,
Wo, = (W + W.)v, or
) "ee If P should be equal to W, this equation would reduce to (4.10) by use of
v? = 2gh. Suggested consistent units: vips, g, = 386ips?, Plb., Flb.,
6 in.
cancels) and W, is the equivalent mass. Thus, the kinetic energy of the
If P is not constant, if, for example, it is some small value until contact
in terms of W and »v,, is found by substituting the value of v, in (b) into @
is made, as at B in Fig. 4.33(c), and then if it increases gradually as sug-
Wee
gested, the maximum stress will be less than that obtained from (e).
(c) KE, = ft-lb. or in-lb. At the extreme is the case of the body W striking A with no force
2g(1 + W./W)? acting on it, so that the only energy that A must absorb is the kinetic energy
Now let the body W be moving vertically downward and apply the co of W, KE = Wv?/(2g,). The response of A to this impact is some maximum
servation-of-energy law to the situation at contact (W and equivalent force F at the end of its deformation period and the elastic energy is, as
move with ve) in the form: KE of W and W, plus loss of potential en before, F8/2 = sA8/2 = s?AL/(2E). Writing an energy balance with the
of W is equal to the elastic energy. Then using vy? = 2g0h or h = vy /( addition of the correction factor of § 4.41, we get
ve from (b), and W, = W,/3 in (a), we have
©) FS in s?AL Lm = 1 )
(d) Wo?c 2 Weve?
eVe qe k82
ey.
2 2E 2g, \1+ W./W
280 280 2
which simplifies to . Wv?E

(4.12) Wb + 8)
k
= —
2
ft-lb. or in-lb.
(g)
deere)
1+ W./W 2 where the units must be consistent. Comparing equation (g) with (w) of
where W, = W,/3 for an axially loaded bar of mass W,. Comparing (4.13 § 4.36, for a falling body (where v2/(2g,) =h), note that (g) is a good ap-
with (4.10), we see that the factor C = 1/(1 + W,/W) applied to h or’ Proximation of (w) where h > 8,, =WL/(AE).
v?/(2g,) can be thought of as a correction factor to care for the kir
energy of the mass of the rod. Thus, the equations previously derived 4.43 ELASTIC IMPACT ON BEAMS. If the weight that strikes the
the axially loaded rod can be easily modified by this factor wherever beam is much heavier than the beam, it is easy to adapt equation (4.10) in
v?/(2g,) appears, when it is desired to make this allowance. This sec the form (8 = y = symbol for beam deflection)
analysis has its shortcomings, but it serves the purpose of this text.
se

4.42 IMPACT BY A BODY MOVING HORIZONTALLY. in

(4.10) Wh +») = W(E +
So ag )-=
case will throw some light on the principle involved—the conservation¢ Provided it is assumed further that the deflection curve under impact is the
* For in-lb., use v, ips and g, = 386 ips?. Same as under static loading, which is to say that we can use deflection
 54.46] CLOSURE 153
ASS
<==F--—] = FIGURE 4.34 Impact on Cantilever. Length of beam = L; stress varies inversely as the volume of material. With respect to a beam of a
| Sb=y. certain sectional area, for example, if its length is doubled, its elastic
SOO

capacity to absorb energy is increased by the ratio of 1/2, or about 40%

formulas such as those in Table AT 2; ky?/2 is the elastic energy stored in (h > y), Whereas its capacity for static loads is reduced by half.
the beam. For example, for a cantilever beam, Fig. 4.34, we find y = FL/ The natural frequency of vibration ¢, can be expressed in terms of the
(3EI), so that the scale of the beam as a spring, with respect to the point of static deflection y,; or 8,;,
application of the load, is 1 Z 1/2 1 1/2 ] 1/2
¢ -=( | = 3.13 (— | = 3.13(--) ;
FSET
(h) k= Ib. /in.
(k) 2 2a Vst Vst 8s
2 AB which in turn, can be used to replace y,, with ¢, in the various foregoing
In one type of problem, one may substitute this value of k into (4.10) and equations.
solve for y. Then solve for y,, for a static load W and the correspondin;
stress S,,, aS usual;
4.46 CLOSURE. It should be evident by now that to design for
WL Mc WLe varying loads by considering the load static and trying to cover by a large
(i)i Yst By 3EI and qs T eee,
T
design (ignorance) factor on the yield or ultimate strengths is fraught with
danger or conducive to over design. It often happens that a combination
Finally, if the action is elastic, the stress is proportional to strain; so, the of extreme conditions results in failure, even when the designer thought
maximum stress under impact is he was being conservative. In spite of the necessarily empirical approach to
varying loading, the detailed thinking of and allowing for the various
(i) las. factors that affect the design lead to improvement. Eventually, we may have
y st something in the way of a theory (conceivably related to metallographic
Other types of beams may be similarly handled. structure and microstresses) that provides a more reliable approach, but
no suitable one is available now. There is scientific interest in microstresses,
getting down to the stress in a single crystal, whereas the stresses we com-
4.44 EFFECT OF MASS OF BEAM. If the value of k from (h)is pute are statistical averages, which we call macrostresses. Even for static
substituted into (4.10), one may solve for y, and use y,, = WL9/(3ED) design, one extrapolates from laboratory data to a finished structure; so
the result is 4 live-load design is not too different except in detail. Inevitably in complex
systems and machines, there are points of high stresses (weak spots) that

(4213); igi. Yall + =} the designer cannot possibly foresee; hence, this hazard is always present.
2h 1/2 v2 1/2

Vst
a5, + vo( 1 + So) st
) There are a number of factors that the designer can misjudge: lead alloy
makes for easier machining of steel, but weakens it in fatigue; rolling or
which is seen to be the same as equation (t), § 4.36; y = 5. In this form we peening can be overdone, injuring the material (surface cracks) rather than
can introduce a correction factor on h (or v?), which in terms of equiv Strengthening it; stress concentrations at points of support, as for a pres-
mass W, is the same as before, namely 1/(1 + W,/W). But the value Sure vessel, may be overlooked; distortions from heat treating; unantici-
W, are different: for the cantilever beam, Fig. 4.34, W,=33W,/ pated effects of deflections; etc. In general, a respectable percentage of
r, W, = 17W,/35; where a
for a simple beam, loaded at the cente failures can be eliminated if the engineer follows his design all the way
the mass of the beam, !-2) through manufacturing; many failures are attributable to discontinuities
(tool marks) left in processing and in wrong fillet radii that are insignificant
from the standpoint of manufacture. Be generous with the size of fillets at
4.45 GENERAL REMARKS ON IMPACT. It can be said of t Te-entrant corners, make changes in contiguous dimensions as gradual
foregoing analyses that the deflections are predicted within a reasonabl 4s possible, let residual stresses be compressive if you can, and as a last
approximation, but the computed maximum stresses may be in serious exhortation, be wary of the highly simplified generalizations—in this book
error. Also, if the rod or beam has a uniform cross-section, the induce¢ Or elsewhere.

152
 5. SCREW FASTENINGS
AND RIVETS

5.1 INTRODUCTION. There are several methods of joining the

parts permanently or semipermanently—by welding, by riveting, by the
use of screw fastenings, and by numerous special means. The screw fasten-
ing is one of the most useful elements of machines. Its design varies from
the simple extreme of some casual calculation to dependence on extensive
experiments intended to simulate a particular environment.

0° Pitch (P) Pitch Pitch

Sai be f Crest :

FIGURE 5.1 Thread Forms for uM

Screws.
; (a) V-Thread (b) Sellers (c) Whitworth

5.2. KIND OF THREADS. __ The sharp crest and root of the V-thread,
Fig. 5.4. occasionally cut on lathes are undesirable because the thin
Material is easily injured and because the concentration of stress at the
Toot of the thread is large. William Sellers proposed (1864) the form in
Fig, 5.1(b), with flat crest and root, which partially removes the inherent
weakness of the V-thread. The Sellers thread was the U.S. standard for
Many years.
156 STRESS FASTENINGS AND RIVETS [Ch.5 §5.5] FITS OF THREADS 157

The Whitworth thread (1841), Fig. 5.1 (c), featuring rounded crests and 5.4 STANDARD THREADS. There are a number of “standard”
roots, has been the standard in Britain. It has better fatigue strength, threads, some quite specialized, concerning which, refer to the standard.©-1
because of the rounded root, than the Sellers thread. Tables AT 14 and 5.1 give certain data for some widely used threads.
The current U.S. standard, shown in Fig. 5.2, is in agreement with the The coarse-thread series (designated UNC) is recommended for general
international* Unified Standard. This Standard has the 60° thread angle use; where jar and vibration are not important factors, where disassembly
the old American standard and the optional rounded root of the British of parts is frequent, and where tapped holes are in metals other than steel.
standard for an external thread; the crest may be either flat or rounded, as Always use the coarse thread unless there is a reason for using another.
shown. There are similar choices for the internal thread, Fig. 5.2(b). The fine-thread series (designated UNF) is frequently used in auto-
motive and aircraft work; especially where jar and vibration (tending to
loosen the nut) are present, where a fine adjustment is required with the
5.3. DEFINITIONS. The major diameter is the diameter of the imag- use of a castle nut, and where tapped holes are in steel (avoid UNF in brittle
inary cylinder that bounds the crests of an external thread and the roots of an materials).
internal thread; it is the Jargest diameter of the screw thread. This dimen- The extra-fine-thread series (designated UNEF) is particularly useful in
sion has been called the “outside diameter,”’ a confusing term when applied aeronautical equipment; suitable where thin-walled material is to be
threaded, where fine adjustments are required, and where jar and vibration
are excessive. Screws with these threads are likely to be made of heat-
treated alloy steel.
The 8-thread series (designated 8 UN), Table 5.1, is used on bolts for
high-pressure pipe flanges, cylinder-head studs, etc. There are several
constant pitch series, for example, 12 UN (see under Fine, Table AT 14),
16 UN, 20 UN. These threads are useful because, in the regular coarse
thread, the pitch continues to increase with size and it becomes increasingly
difficult to induce the desired initial tension in the bolt, so necessary for
(a) External Threads (b) Internal Thread tight joints. Even for the 8 UN thread, impact wrenches or very large
leverages are needed for tightening to stresses of the order of the yield
FIGURE 5.2 Unified Screw Threads.(5-1] At A and B in (a) and C and D in (b), see hoy strength.
the thread is derived from the V-thread (height H); H = 0-866P, where P = pitch. Some
dimensions and thread variations are shown. 0
5.5 FITS OF THREADS. The same types of fits as defined in § 3.4
to threads cut internally. The size of a screw is its nominal major diam are used; for threads, the defined tolerances are designated 1A, 2A, 3A
The minor diameter is the diameter of the imaginary cylinder that boun for external threads, 1B, 2B, 3B for internal threads.) Classes 1A
the roots of an external thread or the crests of an internal thread. and 1B have the widest tolerances and are used where quick and easy
dimension has been commonly called the “root diameter.” The pitch P assembly, even with bruised threads, is necessary, as in ordnance.
the axial distance from a point on a screw thread to a corresponding poimi Classes 2A and 2B, with tolerances suitable for normal production
on the adjacent thread. Practices, are the most widely used. Clearance for this fit minimizes galling
and seizing in highly tightened joints and in high-temperature applications,
1
a P (inches) = : and it also accommodates plating. Use this class unless there is a reason for
(@) oe) number of threads per inch using another. Class 2A provides a clearance when mated with any class
The lead is the distance in inches a screw thread (a helix) advances of internal thread.
axially in one turn. On a single-thread screw, the lead and pitch are identi Classes 3A and 3B have the closest fit; allowance (but not the clear-
ance—§ 3.12) is zero. Use only to meet exacting requirements. Other fits
on a double-thread screw, the lead is twice the pitch; on a triple-th
may be obtained by using internal and external threads from different
screw, the lead is three times the pitch; etc. See Fig. 8.21.
Classes, a 2 with a 3, for example. Interference fits are defined in ASA
* U.S.A., Britain, Canada. Standard B 1.12.
 BEBBIRAESL VOR B44 REDE
R BE BAUR SBN RF Oe

Unified and American Standard. See Subheading on Table AT 14

1-THREAD SERIES (8 NAND8 UN) §5.7] INITIAL TENSION AND TIGHTENING TORQUE 159
EXTRA-FINE SERIES (NEF AND UNEF) (PRIMARY SIZES)
initial stress than he will a larger one. Partly for this reason, the design
Basic Th. Minor Stress Minor Stress stress for bolts and screws should be a function of size when the computa-
Size Major In. Dia. Area Size Dia. Area tions consider only the external load. Some years ago, Seaton and Routhe-
Dia. in. (tpi) Ext. Th. | As sq. in. Ext. Th. | As sq. in.
waite®-8 proposed that the design stress in terms of the root area A, of the
12 0.2160 32 0.1777 | 0.0270 1 0.8466 | 0.606 thread be s = CA,5!/1, Considering the approximate nature of the design,
+ 0.2500 32 0.2117 | 0.0379 14 0.9716 | 0.790 we could use an exponent of 1/2 instead of 5/12, use the tensile stress area
te 0.3125 32 0.2742 | 0.0625 1} 1.0966 1.000 A, for convenience (Tables AT 14 and 5.1) instead of the root area 4,,
3 0.3750 32 0.3367 | 0.0932 13 1.2216 1.233
ts 0.4375 28 0.3937 | 0.1274 and obtain s = CA,1/*. The value of C depends upon the yield strength of
14 1.3466 1.492 the material and may conservatively be taken as C = s,/6. Thus, the de-
} 0.5000 28 0.4562 | 0.170 1§ 1.4716 1.78 sign stress for “well-tightened”’ bolts or screws may be used as
& 0.5625 24 0.5114 | 0.214 1} 1.5966 2.08
§ 0.6250 24 0.5739 | 0.268 1g 1.7216 2.41
i} 0.6875 24 0.6364 | 0.329 (b) Design tensile strength= sz = Ay
3 0.7500 20 0.6887 | 0.386 2a 1.8466 |
sates 2.77
vee
Substituting this value of sg into F = sA, we find
# | osi2s | 20 | 07512 | 0.458 24 ane ie
; 0.8750 | 20 | 0.8137 | 0.536 23 . : ¢ 5, A3l2
# 0.9375 20 0.8762 | 0.620 6 651 (5.1) F, = basi? As =— - : [D < #in.]
1 1.0000 20 0.9387 | 0.711 :i sate 7.6738
1% 1.0625 18 0.9943 | 0.799 34 3.3466 | 8.96 from which a safe external tensile load F, may be found, or the needed
5966 | 10.34 stress area A, determined for a particular external load. From a computed
14 1.1250 18 1.0568 | 0.901 38 oe
1¥s 1.1875 18 1.1193 1.009 4 3.8466 11.81 area A,, determine the nominal size of bolt from Table AT 14 or 5.1.
1} 1.2500 18 1.1818 1.123 44 4.0966 | 11.38 The constant 6 in equation (5.1) is not a factor of safety. Equation (5.1) may
lis 1.3125 18 1.2443 1.244 4} 4.3466 | 15.1 be applied to any thread series, although there are no tests to substantiate
1} 1.3750 18 1.3068 1.370 43 4.5966 16.8 its use for other than a coarse thread. AISC'5-34] specifies allowable
1% 1.4375 18 1.3693 | 1.503 5 4.8466 | 18.7 tensile stresses as follows: 0.45, in general; for A 307, 14 ksi (see § 5.8);
1} 1.5000 18 1.4318 | 1.64 53 5.0966 | 20.7 for A 325, 40 ksi; for A 354, 50 ksi; values that serve as a guide for bolt
1% 1.5625 18 1.4943 | 1.79 53 5.3466 | 22.7 sizes D = Zin.
1} 1.6250 18 1.5568 | 1.94 5 5.5966 | 24.9
1% 1.6875 18 1.6193 | 2.10 6 5.8466 | 27.1
5.7 INITIAL TENSION AND TIGHTENING TORQUE. The stress
or load induced by the tightening operation is called the initial tension,
which, with ordinary wrenches, depends upon the workman, how he is
Identification symbols for use on drawings, shop and storeroom, feeling, the length of wrench used, as well as the condition of the bolt or ~
etc., should be according to the following examples. fl screw. Where the amount of initial tension is important, a torque wrench
(a) An externally threaded part, 1-in. diameter, unified coarse thread, 8 should be used. Even so, there will be a /arge variation of the induced stress
threads per inch, class 2A tolerance is designated as follows: ¥ because of the way the threads are finished, their lubrication, and the other
variables of the application; see Fig. 5.3. The relation between the applied
1’°—8 UNC—2A.
torque T in-lb, and initial tension F, lb. proposed by Maney®-1! js
(b) An internal thread, 1-in. diameter, unified fine thread, 12 threads per
inch, class 2B tolerance, left-hand thread is designated ( As received!-1°1 C = 0.20
(5.2) T = CDF, in-lb.,
1’—12 UNF—2B—LH Lubricated®-14! C = 0.15
where D is the nominal bolt size and C, called the torque coefficient, is
5.6 DESIGN OF BOLTS—INITIAL TENSION UNKNOWN. * taken as a constant for a particular set of conditions. Equation (5.2) is
obtained froma force analysis of the thread (the same as for a worm thread,
mechanic with a typical set of wrenches will tighten a small bolt to a higher
§ 16.8), plus an allowance for friction between face of nut (or head of the
 FIGURE 5.3 Difficulties in a Bolted Connection. (1) Load not
§5.8] MATERIALS AND STRENGTH OF THREADED PRODUCTS 161
distributed over all threads. (2) Axis of internal threads not yp
pendicular to seating face of nut. (3) Surface not flat and per-
pendicular to axis of bolt. (4) Hole not perpendicular to surface (an¢
polted structures are somewhat different. In another plan, if both ends of
parallel to axis). (5) Misaligned holes. (6) Bearing surface on he the bolts are accessible for the measurement of bolt length, the initial
not perpendicular to axis. (7) Also, the manner of applying tension may be estimated from the measured total elongation 8. Usually,
external load may result in bending the bolt. And there is a torsional there is an unknown factor in this procedure, the effective strain length L.,
stress due to tightening. Considering the nature of these faults, ve
which may be defined as that length which when divided into the total
decide that rarely is the load on a bolt purely tensile.
elongation gives the unit elongation at some section of the bolt, as at the
root section; that is, « = 8/L,. Knowing «, we have s = cE. The effective
screw if the head is being turned), which is computed as in § 18.10. Thi length is closely some function of the grip and of the proportion of shank
value of C ~ 0.2 is obtained when the coefficient of friction is taken ag that is threaded.
f = 0.15; this value of C is considered typical and is recommended for
unless another value is known to apply. The “‘as-received”’ condition
cludes the remains of cutting fluid, for example, but no foreign partic 5.8 MATERIALS AND STRENGTH OF THREADED PRODUCTS.
Values of C have been observed in the literature from 0.1 to 0.34 or mo The designer can use any desired material for bolts and screws, but he
The most unpredictable cases would probably be those where the surfac would use only “ordinary” steel (ASTM A 307 and SAE Grade 1, minimum
have been cleaned and dried. Some data suggest C ~ 0.14 if the threads a 5, = 55 ksi—equivalent to about 1015, for example), bolts cold headed,
lubricated with molybdenum disulfide.-19! There is some evidence that without a justifying reason. The SAE and ASTM and several government
UNF threads, C is some 6-10% less than for coarse threads. Plating gen- agencies have standardized specifications for screw-fastening materials.
erally reduces the friction, another situation calling for different values 1 Among the most commonly used steels for bolts are;'-1! AISI 1013, 1018,
C. If the actual value of C = 0.1 and if the tightening torque used has been 1038, 1041, 1054, 1340, 4037, 4140, 4150, 50B40, 8635, 8735, 4340; but a
computed for C = 0.2, the initial tension in the bolt is twice that planned particular manufacturer’s processes may be more readily adapted to one steel
if the bolt does not yield and break. Thus, if the parts are lubricated, te than another.
may be advisable to avoid overstressing the fasteners. The value of C ter Bolt design is often done with a proof load or proof stress s,. In general,
to increase with decrease in size; an average value for }-UNC-20 i a proof load is some load agreed to by the purchaser and vendor as a
0.255, {5-141 capacity to be met. For bolts and screws, there seems to be no unique
Tightening the nut produces a shear stress in the bolt (by the frictiona definition, but the proof stress s, in the SAE specifications is usually close
torque); to 96% of the 0.2% offset yield strength (which may not be the same as the
yield strength of a standard tensile specimen). See Table 5.2.
Torque producing shear in bolt ~ (0.4) (total 7)
The torsional stress is generally ignored in design because, first, it is rela
tively small, and second, it probably mostly disappears as the mac TABLE 5.2
operates (a relaxing of the frictional forces). It is not uncommon for MINIMUM STRENGTHS (ksi)
initial tension to decrease because of seating action on contact surfaces 7 OF BOLTS (SELECTED
perhaps because of actual yielding of the materials; also of course if kis # STANDARD SPECIFICATIONS)
nut loosens. Where it is important for joints to remain tight, the tightness
GRADE SIZES, INCL. Sp Sy Sy
of the bolts should be checked later.
A torque wrench or other adjustable power tool is the most widely > SAE +t 55 69
an Grade 2 #1 52 64
means of inducing the approximate desired initial tension in smaller b annie 14-14 | 28 55
It is easy to tighten smaller bolts, say } in. or less for ordinary steel, unt Ay - <PS z&
they break. For the larger sizes, pneumatic impact wrenches are used, al
2 gt SAE +i | 85 120 88
2/308! Grade 5, #1 Ter 418 | cet
sometimes the bolts are heated when they are tightened so that the initia 40s ASTM A325 i--li | 74° 105 =—7
tension is greater after they have cooled. The turn-of-the-nut plan is: tu
bz ASTM 354 es | 424 | 80 105 83
the nut very tight in order to seat the surfaces (a step sometimes omitte 234 7s" 99. | 78
back off nut, turn nut snug (finger snug), then turn it through a #24 | 105 125 109
computed angle to induce a particular strain, which is proportional to Re takomeuiaia
desired stress. (The contact surfaces deform too.) Approved procedures il BD. dhb calidon 450 125
162 STRESS FASTENINGS AND RIVETS [Ch.5 g5.8] MATERIALS AND STRENGTH OF THREADED PRODUCTS 163

The ASTM and SAE specifications can be met by many “standard” life should be tested to prove the design if failure would lead to dire con-
steels; SAE 1041 QT, for example, can easily meet SAE Grade 5, a goo sequences. One manufacturer claims a repeated-load fatigue strength for
strength bolt material. In fact, 1041 heat treated to about s, = 100 his cap screws up to }-in. diameter of 40 ksi (with threads), stressing them
(BHN ~ 200) or a little more, will come close to giving the most hold froM Smin = 4 (to 40 ksi), using 8740 and rolled curved roots. Forsuggestions
capacity per dollar.?-1) A low-carbon steel can meet the requirements for geometric improvements of fatigue strength, see Fig. 5.4. With the
Grade 2, Table 5.2, and bolts of this grade are typically cold headed diameter reduced between “guide surfaces,” the capacity to absorb energy
The medium carbon grades must be hot headed. The resulfurized steel is also increased (§ 4.40).
11xx series, tend to be damaged by cold heading; so these are used prim
Max. Fillets
for studs to be machined from bar stock. For ultimate strengths greate
up te reD Undercut Nut,
than about 100 ksi, medium carbon alloy steels are common; such steel: to r=0.1D
are used to meet the ASTM A 354 specification, Table 5.2. Recall that a
4 Mint
size increases, increased hardenability is necessary to maintain the sai | t I
ie -
y
en 4 a i

strength, or else the strength decreases, as allowed by the specifica

One large consumer of stud bolts for automotive cylinder heads e Guide Surfaces 0.2D min.at
as Needed Thread End,
nated assembly-line trouble in the form of stretching and breaking quenche d<D,
and-tempered bolts by switching to elevated-temperature-drawn 41:
which came out with properties of s, = 150, s, = 130, s, = 120 ksi, min. FIGURE 5.4 Bolt for Repeated Loads. The undercut of the nut improves the distribution

without heat treatment; see § 2.9. q of the load on the nut threads. The radius of curvature of the groove at the end of the
threads should be at least 0.2D, preferably 0.5D, with a diameter d somewhat less than the
If the threads are manufactured by rolling (rolled threads—large defor
minor thread diameter D;. Rather than this grooye, the threads could be ended on a
mation of the material), as opposed to cutting, the resulting surface gradual taper, instead of an abrupt ending at full depth.
the order of 4-32 pin., § 3.14. When the material has s, < 80 ksi, there
little change in fatigue strength as compared with cut threads; but for hi
strength steels, rolled threads show a marked improvement in fatigue, Notice the “undercut nut’’ of Fig. 5.4. With a regular nut, the first
much as 2 or 3 times as great for s, > 200 ksi.*:®?! See Table AT 12| thread may be loaded about 1.8 times the average per thread (some 2.3
values of the strength reduction factor K,, which include the effect of 1 for fine threads), whereas the “top’’ thread takes about half the average
manufacturing method. (4 for fine).©-?) In any case, it is impossible to get a uniform distribution
To complicate matters, Sigwart'*-?*) found that the residual compress of load on the threads. With the undercut, the reduced rigidity at the bottom
stress at the root of the rolled thread may not be as great as that at the r of the nut permits the lower part to stretch more thus distributing the load
of a cut thread, especially if a dull cutter was used. The best improvem more evenly among the threads. In another design to accomplish the same
was found for ground threads that had been locally rolled at the r purpose, the nut is tapered toward the bottom; but of course these special
(minor plastic deformation—§ 4.28) after heat treatment, an expe designs are more expensive and would not be used unless needed. In general,
procedure. In any event, if beneficial residual stresses are to remain, the nut material may be somewhat weaker than that of the bolt, but speci-
rolling must be done after heat treatment, because otherwise the ce fications require that the threads not fail before the bolt fails in tension.
working effects would be lost. As usual, chromium and nickel platin; Thus, better than ordinary, perhaps heat-treated nuts (and hardened washers
reduce fatigue strength; the effect of cadmium and zinc is smaller. Fati if the material of the connected parts is not as hard) are required for the
failure nearly always occurs at the first thread in the nut, the end of higher-strength bolts.
Many screws are made of stainless steel, nonferrous metals, and plastics
threaded portion on the shank, or, especially if the threads are roll
under the head. Thus, if every advantage is being taken to impr (for example, Zytel, Teflon,'?-35!) which are used for a reason—corrosion
fatigue of bolts, the fillet at the head must be considered. It is often ro. Tesistance, high or very low temperatures, light weight, electrical conduc-
tance, insulation, etc. See Chapter 2. One manufacturer"9-24! indicates that
Besides rolling threads after heat treatment, other usual precautions &
the breaking strength of } to 3-in. aluminum bolts (2024-T 4, 6061-T 6) is
preserve or improve fatigue strength may be taken, for example, avoit
decarburization. Self-aligning nuts, Fig. 5.31, reduce the bending of about 62-64 ksi based on the stress area.
bolt mostly due to surfaces not being exactly perpendicular to the bolt In the design procedure, one might decide upon proof stress. From the
axis, and they prolong fatigue life; an item to consider if there is failure Proof stress, the next decision is the initial tightening stress. In structures, the
from fatigue and for other reasons.'-28! Bolts designed for finite fatigut tendency is to tighten the bolt to or beyond the yield, s; ~ s,, in high-
 On Foit of Upening

164 STRESS FASTENINGS AND RIVETS [Ch,

strength material (which does not have a distinct yield point). Erker'*:
found that after tightening bolts to just above the yield strength, s
plastic deformation in service reduced the clamping stress to an optimy
value. On the other hand, a review of the literature suggests that machij

Force, F
designers tend toward an initial stress somewhat less than s, or sy; ¢
finds values from 0.75s, (by some automobile manufacturers) to 0.9s.
more. Typical values for bolts to be subjected to a tensile load are
(c) 5; 0.95, or 5; © 0.85s,,
[WHEN PROOF STRESS AVAILABLE] [NO PROOF STRESS]

which values we shall assume are satisfactory, especially with no gaske

the case of a soft steel, a material with a yield point, one must be ca
not to tighten to the yield stress because of the relaxation and | Tension
s ia Compression
clamping force that follows from the yielding. The higher-strength s' Deformation, 6
may receive a permanent deformation, but relaxation does not occ
ordinary temperatures. FIGURE 5.6 Forces on a Screw Fastening. The slopes of the F-5 lines are kp = F/8p,
Having decided upon a suitable s,, the initial tightening force is k, = F/&c. Suppose, when the external load F, is applied, that plastic deformation of the
5,A,; then equation (5.2), with a proper value of C, can be used to esti: bolt occurs, idealized for a good strength material as PQ; F- is then represented by DQ,
the tightening torque (or one can determine the angle through w the effective F; is reduced to some value GK because of the resulting permanent deformation.
turn the nut to produce s,—§ 5.7).
and the force-deformation curve for the bolt is a straight line, represented
by OAM in Fig. 5.6. Also, the connected members deform (in compression),
Connected Parts
and if they too are elastic, their force-deformation curve is straight, repre-
YAN sented by CA in Fig. 5.6,.'-181 The more rigid a member, the steeper is
UI its F-8 curve, because it takes a larger force to produce a particular defor-
mation. Usually the connected members are more rigid than the bolt, as
shown in Fig. 5.6 with « > 0. The slope of CDA is negative and represents
a compressive deformation.
Produces an Suppose we stop tightening at a point A. The load on the bolt and on
rnal Loed FIGURE 5.5 Head Bolted on Vessel. Internal pressure is p. _
F, on Bolt the connected part is F,, the initial tightening load. The initial elongation
of the bolt is 5,, and the corresponding compressive deformation of the
Head
connected parts is 5,. To get the external load that would cause a joint such
as that in Fig. 5.5 to open, assume that the bolts do not bend, which is
€quivalent to assuming that the head and flange do not bend, and let an
5.9 ELASTIC ANALYSIS OF BOLTS FOR JOINTS. An oft-hi external load F, be applied. The bolt elongates more AS, say to B, Fig. 5.6,
rule of thumb when the parts to be joined are relatively rigid is tightent and the deformation of the connected parts decreases the same amount,
bolt (or screw) so that the initial tension is greater than the applied extern AS. The load on the bolt increases an amount AF,; the load on the con-
load, a rule that should yield safe designs if the bolts or screws are know nected parts decreases a greater amount AF, if they are more rigid. For
to be tightened to the required initial tension. However, an engin Clastic deformations, the bolt elongation continues along the line OM,
more comfortable with an analysis that guides him in the correct dire and the compressive deformation decreases along AC. The joint will be on
First, we shall see what load is required to open a joint, such as the 01 the point of opening when the deformation of the ‘connected parts becomes
suggested by Fig. 5.5. zero, at C, because if the bolt is stretched any further, the connected parts
As the nut is tightened, the load on the bolt increases, and the defor fan no longer expand to maintain the surfaces in contact. At the instant
tion of the bolt increases. Within the elastic range, Hooke’s law app! marked by C, the total elongation of the bolt is represented by the distance
 1? STRESS FASTENINGS AND RIVETS [Ch.! 35.9] ELASTIC ANALYSIS OF BOLTS FOR JOINTS 167

on the values of Q, k, and k,, the computed value of F; from (5.3) may be
of the joint, which is also the external load at this limiting condition. either smaller or larger than the external load F,.
Since the triangles OGA and OCM are similar, In Fig. 5.6, suppose a bolt is tightened to the F; shown and an external
Fy 5+ load F, is applied to the joint, all elastically. The bolt stretches an additional
(c) 3 or AS and the total load F; on it is marked by point B, with a change of force
RK. 4 AF, corresponding to an increase of deformation AS. The change of load
on the connected parts becomes AF, = HD = F, — AF,. From 8 = F/k,
of their spring constants from k = F/6, § 4.35, are 8; = F,/k, for the bi the change in deformation is equal to the change in force divided by the
and 5, = F,/k, for the connected parts. Using these values in (c), we hay corresponding spring constant; that is,

@: BA F(2=*)
k,
k, +k
Mh ae (-=-}:
k
kn + ke 4q
i
© Fated
te oR
Solving for AF, from the last two parts, we get

q nips Hee
In these expressions, F, is the external load that would place the joint «
the point of opening when the bolt had been tightened to a value F;; 0 ky
On Point of Opening
Then the total load on the bolt is, Fig. 5.6,

(5.4) R=k PAR eee (—2_)e.

ky tk,
Similarly, the net compressive force on the connected parts is found to be

(e) R= k-(—*-)h.
Force, F'

k, +k,
A negative answer from this equation indicates no force on the connected
parts. If the stiffness k, of the connected parts cannot be determined with
assurance, it is always on the safe side to use the term in parentheses as
N FIGURE 5.6 unity; that is, F, = F; + F,. See below. When the external load F, varies,
Co suman (Repeated). the total load F;, varies in accordance with (5.4); the mean load is F,, = F;
Cc + AF,/2 and the alternating component F, = AF,/2. The nominal stress
65 5
Tension Compression
for any Fis s, = F/A,.
————_» —_—
Examining equation (5.4), we see that if the stiffness of the bolt k,
Deformation, 6
is very large as compared to k,, the total load F, approaches F, + F,. If
ky is very small as compared to k,, the term in parentheses becomes small,
atid the total load approaches F;. Therefore, the actual load is always
tightening load that must be used. Practically, F, should be greater between the initial tension and the sum of the initial tension plus the external
this value, 1.2 to 2 times as great when there is no gasket.-19! That is, le load (provided the joint does not open). These remarks are interpreted
F, = QF, in (d), where F, is some actual external load, and find graphically in Fig. 5.7 (p. 168). Notice that if all parts are elastic, any ex-
ternal load, no matter how small, results in an increased load on the bolt.
(5.3) Ein or( =~) Ib. : The foregoing analyses are inadequate when the bolt is subjected to
Significant bending moments; hence, when the gasket is inside of the bolt
Keep in mind that a joint intended to resist leakage may be subjected tog Circle, a common arrangement for flat or ring gaskets, Fig. 5.8 (p. 168), the
hydrostatic test pressure of 1.5 to 2 times the working pressure. Depending flanges must be heavy enough that bending is small. Also the bolts should be
 AF, & 5.10] ELASTIC CONSTANTS AND GASKETS FOR CONNECTED PARTS = 169

g Fe For the connected parts, the same principle holds, but for parts of in-
fe i definite_extent (area), their deformation at some distance from the bolt is
| jess than that in the immediate vicinity of the bolt. In this event, the usual
AF. procedure is to assume some equivalent area of the connected parts A,,
and use k, = A,E,/L,. One such estimation is®-19)
F;
mD,*) 7D?
Q 8 (g) A, = ety ogee
Deformation, 6 Deformation, 6
(a) ke>ke (b) ko>ke where D is the nominal diameter of the bolt hole, D, e is an “equivalent”’
q
diameter of the plate area considered to be in compression; take it as
FIGURE 5.7 Effect of Relative Stiffness of Bolt and Connected Parts. These figu h
drawn for the same external load F,. Notice that when the connected parts are much s (h) D, = (Nut or head width across flats) + 3
than the bolt, in (a), the load F, does not cause much change AF in the load on the hk
But if the bolt should be much stiffer than the parts, in (b), a very large part of the exter,
load F, is added to the initial load. The use of a gasket changes the relationship betw h
= (Dimension A, for example, Table AT 14) + 2
F, and AF; in the direction shown in (b).

where h is the grip of the bolt, which is the total thickness of the plates
being clamped. If it is clear exactly what area is in compression, do not
use (g) and (h). See the example of § 5.12.
If the connected members are composed of two or more kinds of material
—for example, a gasket between connected parts—the spring constant for
the connection is (see equation (4.11), § 4.39)
ay

1 1 1 1
eh Mayne
lag heck
Soe apt ae 98 “FS

‘Metal Ring
Gasket Here <ul where k,, kg, kg are the spring constants for the individual components
(a) (b) being clamped; k, = A,F,/Ly, etc.
If needs can be met economically without a gasket, this is the best
FIGURE 5.8 Flanges. (Courtesy Taylor Forge and Pipe Works, Chicago). solution. Next best is the use of as thin a gasket as possible. However, there
must be enough thickness to let the gasket material flow into the roughnesses
close enough together that virtually uniform pressure is induced betweentl of the flange surfaces and to take care of some slight unparallelism of the
faces.'-22] From one code, we conclude that the spacing or pitch P of bo Mating-surfaces. The amount of pressure on the gasket material, called the
should be: P < 7D for fluid pressures less than 50 psi; P + 3.5D for lange pressure, to cause it to respond in this manner is different for each
material. For nonmetallic gaskets, it has been found that they should have a
pressures of about 200 psi, where D is the size of bolt.5-26! Observe tha’
foregoing analysis ignores the deflection of the bolt in the nut and thed certain minimum amount of compression,®-26! as 62% for a certain cork
flections of the threads, all of which would have an effect on k. NM 8asket; this compression corresponds to a certain required minimum flange
Pressure. Then the bolts must be designed and so tightened as to meet those
Tequirements. Smoley'>-25! recommends the use of an “apparent’’ flange
5.10 ELASTIC CONSTANTS AND: GASKETS FOR CONNECTE Pressure p,, which is said to be an upper probability limit such that if the
Initial bolt tension F, is computed from this pressure and then if the bolts
PARTS. The elastic constant k, is determined from equation
§ 4.36; that isk, = A,E,/L,; or if two diameters are involved on the are tightened with a torque T= 0.2DF,, equation (5.2), the required
use the equivalent spring constant from equation (4.11) of § 4.39: k, =A Minimum flange pressure is assured. Certain apparent pressures are: for
The subscript b is a reminder that each variable applies to the bolt. X, a cork-and-rubber mixture, p, = 1500 psi; for Y, a cork composition,
170 STRESS FASTENINGS AND RIVETS [Ch.5 g5.12] EXAMPLE—STIFF JOINT 171

Pg = 2700 psi; for Z, an asbestos base, p, = 3300 psi. For a first approxi: (b) The computed values of the initial tension and bolt torque that provide
mation after a suitable gasket material has been selected, one might proceeg the required compression of the gasket are
as follows: estimate the area of gasket; multiply this by its minimun
apparent flange pressure to get the total apparent load on the gasket; Fy sk Bint OP a MIO Gait

decide upon a suitable number of bolts and find the load per bolt F,; the T = 0.2DF, = (0.2)(0.625)(10.5) = 1.31 in-kips.
size of bolt is determined from A, = F,/s, where s = 0.75s,, say. When the
Notice that the initial tightening load needed for the gasket (10.5) is almost ten
apparent flange pressure with its built-in margin of safety is being used, ij
times the external load (1.1), but relatively higher fluid pressures would change
would appear reasonable to assume that F,; = F, and compute bolt tor
this relation materially. For s, = 10.5/0.226 = 46.5 ksi, we find s,/sy = 46.5/75 =
from T = CDF,. The bolts must also be capable of carrying the exte: 0.62, which, compared to the 0.75s, informally suggested in § 5.10 in connection
loading. There are many special considerations beyond the scope of with gaskets, implies that if one designed with this latter approach, a smaller
text.-26] Metal gaskets (aluminum, copper, monel, and others), with am bolt may possibly be used. [After the computations of part (c) have been made,
without sandwich filler, are frequently used, and necessarily so at tempera observe that the initial bolt tension required to compress the gasket material is
tures above some 850°F. qj much larger than the initial tension computed from equation (5.3).]
The modulus of elasticity for the nonmetallic materials is ordina (c) For a grip of 1 + 0.02 = 1.02in. and the distance A = 15/16 in. from
not constant over a wide range of stress; but if it is desired to check by Table AT 14, the equivalent diameter and area from (h) and (g) are
principles of § 5.9, some typical values of E and gasket thickness h are:
D. = 0.9375 + ot = 1.4475 in.; A, = : (1.44752 — 0.6252) = 1.34 in?,
rubber gasket, E = 7500 psi, h = 3/16in.; vellumoid gasket, E = 19,
psi, h = 1/16 in.; copper-clad asbestos gasket, E = 20,000 psi, h = 1/8 in Use k = AE/L and 1/k. = 1/k, + 1/k2 to get k..
pla 1 4 0.02
ke 1.34x 3x 10* 1.34 x 19’
5.11 EXAMPLE—STUD BOLTS FOR COMPRESSOR HEAD. Theheadc
a 10 x 12-in. air compressor is to be held on by 10 stud bolts; the maxi from which k, = 1230 k/in. Observe the large effect of the gasket on k..
(repeated) internal pressure is 140 psi. The bolts are to be made of cold-ro)
C1118 with threads cut on their entire length. (a) Determine the size of
ey on oe cee 29S eho kinsfia:,
4

predicted by equation (5.1), which assumes a well-tightened joint. (b) Let a 0.02
where the stress area A, is assumed to be reasonable for this computation. From
thick gasket, designated X in § 5.10, be used for which the “apparent”
equation (f),
pressure is p, = 1.5 ksi; assume a gasket area of A, = 70 in.? and a modulus
elasticity of E = 19 ksi. (The E values are so seldom used by the industry
they are hard to find.) Determine the initial tension, the bolt torque to be u
AR a ky + ke
aa ( 6650 oh+ 1230,
) = 0.93 kips.
and the ratio of s,/s, to obtain the desired flange pressure. (c) The thickness of thi For a mean load F, = F, + AF,/2 = 10.5 + 0.47 = 10.97kips, and F, =
steel cylinder head at the bolt holes is 1 in. Estimate the factor of safety by oe = 0.47 kips, and for s, = s,/2 = 40 ksi and K, = 1.8, Table AT 12, we
Soderberg criterion for the bolt size obtained in (a). ave
Solution. (a) The total load F; on the cylinder head is
Ay) Ses) ise 1097.0) 08 K0.47)
~ (0.226)(75) (0.226)(40)’
2
F. el = (140)(=)(100) = 11,000 lb. N Sy Sn
4 4
or N = 1.35. This design seems to be satisfactory, but it does not necessarily
The average external load on each of 10 bolts is 11/10 = 1.1 kips. From Tabli follow that equation (5.1) gives the best size in every instance. There are alterna-
AT 7 for C 1118 cold finished, we have tive Procedures for design and various codes that apply to pressure vessels and
Su = 80 ksi sy = 75 ksi BHN = 180. Pipe joints. Finally; it is entirely possible that for these relatively small alternating
Stresses s,, the Gerber equation (§ 4.6) gives a better estimate of the factor of
Equation (5.1) now yields Safety (and a somewhat higher one here). In any case, computations for the next
Smaller size should be made and studied prior to a final decision.

from which, we select the next larger size from Table AT 14; to wit, D = § in- 5.12 EXAMPLE—STIFF JOINT. A bolt of as-rolled C1118 as shown in
11 UNC, A, = 0.226 in.? if '8. 5.9 is to be subjected to an external load F, varying from 0 to 1650 Ib. It
 in STRESS FASTENINGS AND RIVETS [Ch. 95.13] TYPES OF BOLTS AND SCREWS 173

thicknes
connects the parts C which are made of aluminum 2024-T4, 2-in. total wrench or other method in order to obtain an initial tension reasonably close to
with a diameter 2.D of twice the bolt size D. What should be the diameter | this value. From equation (f),
and
g line? 4
the bolt for a design factor of 2 based on the Soderber kp (15 x 10%A,)1650
AF, = = Y
Solution. First, decide about the stresses. From Table AT 7, find su =75 k e en + zn k 10°A,(15 + 15.9)
= 800 lb.
x 10*ksi. Using s, ~ s,/2, the factor 0.8 for the axi
s, = 46ksi, and E =3
F, The alternat
corresp onding force isareF, = AF,/2 = 400 lb.; F_ = F, + 400 = 1672 1b. The
ing stresses
1: Fa 1672 ns.
Sm a = 7. Si and

The stress concentration factor for cut threads (UNC) is K, = 1.8 from Table
AT 12. For a design factor of N = 2 in equation (4.4), we have
1 Sm , KySa 1 1.672 (1.8)(0.4)
No ay ag) O aed, eae
Grip
or A, = 0.129 in.? (the force unit is in kips to keep the numbers small). From
Table AT 14, we get D = tin. (corresponding to A, = 0.1419). Since the answer
is borderline regarding use of a size factor, one should now recompute without it
FIGURE 5.9 and see if a BG bolt can be used. The load and material are not the same as in the
previous article, but some comparisons can be made, especially as related to the
stiffness kK, of the connected parts. Some light would be shed upon the
situation if you would solve this same problem with steel connected parts instead
of aluminum, and it would take a very few minutes; the answer is % UNC.
Steel F,

5.13 TYPES OF BOLTS AND SCREWS. Basically, a bolt is a screw

fastening with a nut on it; a screw is one that has no nut and turns into a
loading (the bearing surfaces of nuts and heads are almost sure not to be threaded hole. Some bolt heads and nuts are faced; that is, they have a
normal to the bolt’s axis), and 0.85 to be on the safe side for size, we get —
washer face, Fig. 5.10. A large variety of bolts and screws, standard, near
Sx = (5) (0.8) (0.85) = 25.5 ksi.
75 A standard, and special, is available; hence, the brief remarks below and the
illustrations are suggestive only.
Since the load varies, find the mean and alternating components, for which A machine bolt, an old name for an unfinished through bolt, comes with a
needed thestiffness constants knand k..Let Apy = 7D?/4; then A, =4A»—A»=3 Square (Fig. 5.11), hexagonal, or round head and with coarse or fine threads.
A coupling bolt is finished all over, usually having coarse threads. An auto-
ky = Ae
Ly
- Ae GX2 10) _ 15 x 1024, kipsjin., mobile bolt, Fig. 5.12, also finished all over, has fine UNF threads, is usually made
4 heat treated steel, and frequently has a castle nut (Fig. 5.10). A cap screw.
where L = 2 in., given. This computation assumes that effective length is a. 5.17 and 5.18, also falls into this same class, being finished all over. Cap
as the grip and that the threads extend very little beyond the nut. From . " come in a variety of heads: hexagonal, fillister, button, flat head (Fig.
AT 3, for 2024-T4 aluminum, E. = 10.6 x 10° ksi. The effective area i : i‘ pa heads such as shown in Fig. 5.18.
pression in this case should be closely the total area of the connected carriage bolt. Fig. 5.13, is distinguished by a short portion of the shank
A, = 3A,. Therefore, ge the head being square or finned or ribbed. Originally intended for use
. Sauls Cs ood, this bolt is found useful where, for example, the square part fits into
_ Acke _ (3Ao)(1 10") _
x0.6
k. 7 3 15.9 x 10°A, kips/in. ia: hole and prevents the screw from turning while the nut is tightened.
ho aa are many variations of eye bolts, Fig. 5.14, which provide a place for a
Thus, with Q = 1.5, the initial tension from (5.3) is (kips cancel) Ok for lifting parts; they have other uses, also.
15.9°X 108A.) Ea Ta bolts, Fig. 5.15, are among the most widely used types. They are threaded
F. = OF. ( k.= z) = (1.5)(1650) 54-954 155) = 12721
“hi
Si ends and can be used where a through bolt is impossible; for example, in
s 1g a head on an engine where the holes in the block are threaded. This
We shall assume that the factory and the mechanic in the field use a tore Ctice is especially desirable if the hole is in weak or brittle material because
 WME Co
ea A
SERA
.
MERON, i" >
. \\ \

FIGURE 5.10 Castle Nut. (Courtesy FIGURE 5.11 Square-head Machine Bo i FIGURE 5.20 Stove Bolt. FIGURE 5.21 Square-head Set Screw.
National Acme Co., Cleveland). (Courtesy Pheoll Mfg. Co., Chicago). (Courtesy Standard Pressed Steel Co.,
Jenkintown, Pa.).

{ a
FIGURE 5.12 Connecting-rod Bolt, Fine FIGURE 5.13 Carriage
Threads. (Courtesy Lamson & Sessions Co., Belt Co., Chicago).
Bolt. (Courtesy
ARAAARAARAAAAA
Cleveland).
GURE5.22 Hollow-head Set Screw. FIGURE 5.23 Hollow: -head Set Screw.
jong Standard Pressed Steel (Courtesy The Bristol Co., Waterbury,
Co., Jenkintown, Pa.). Conn.).

FIGURE 5.15 Stud Bolt—Coarse

(Courtesy National Acme Co., Cleveland).

FIGURE 5.14 Unthreaded Drop-forged

Eye Bolt. (Courtesy J. H. Williams &
Co., Buffalo). ih UE,
; i‘ st TLL

Type A Type Z Type U Phillips Head Type F

FIGURE FIGURE 5.24 Self-tapping Screws and Phillips Recessed Head. (Courtesy Parker-
5.17 Cap Screw. (Courtesy Kalon Corp., N.Y.).
National Acme Co., Cleveland).

FIGURE 5.16 U-Bolts. (Courtesy The FIGURE 5.18 Hollow-head Cap FIGURE
Th 5.25 Lag Screw. (Courtesy
é
Bourne-Fuller Co., Cleveland). (Courtesy Standard Pressed Steel Co., Jen ¢ Bourne-Fuller Co., Cleveland).
town, Pa.).
a

(a) Oval Head. (b) Clutch Head.

FIGURE
FIGURE 5.27 Forging and Finishing a Nut. (Courtesy H. M. Harper Co., Morton Grove,
5.19 Machine Screw. (Courtesy United Screw and Bolt Corp., Chicago).
TIL).
 | | bevy bsg fgtr eas ake hes
peal

176 STRESS FASTENINGS AND RIVETS [Ch, o ee

45° /ig 1 eas
it
the studs can be left in place when the head is removed, thus saving wear Eat SES,
R=0.1DAI bi at. aH
tear on the hole. However, there are available hardened steel inserts of v
kinds that can be left in the tapped part, including a thin spiral piece, that sci
CIR 3 2A Kap
into the tapped hole, covers the weaker threads, and stays there. Stud bolts a
(a) Oval (b) Cup (c) Flat (d) Cone (e) Half-Dog
also used as through bolts with nuts on both ends.
Point Point ; Point Point Point
A stove bolt, Fig. 5.20, is a cheap variety of bolt made in small sizes. Machin
screws, Fig. 5.19, are also made in small sizes, the numbered sizes in Table ATI FIGURE 5.29 Set-screw Points. Approximate dimensions; 0.5D<C<0.6D; H>0.6D.
and up to j-in., but they are more accurately made than stove bolts, are finish Short set screws (length = the diameter or less) with cone point should have a cone angle
all over, and have fine or coarse threads, various heads, Fig. 5.28, also the clut of 118°.
head, Fig. 5.19(b), and the Phillips head (see Fig. 5.24). The length L o
machine screw, Fig. 5.28, is the distance from the end of the screw to the mo
distant point of contact of the head with the material being clamped. loosening under vibration. Set screws generally have coarse threads and
U-bolts are in the form of a U, Fig. 5.16, and are used as holding clam DS, hardened points.
on an automobile spring. Plow bolts are widely used on farm machinery. Table 5.3 gives the capacity in pounds of tangential force for a cup-
bolts are used in railway track construction. point set screw. This point penetrates somewhat into the shaft material,
Self-tapping screws, Fig. 5.24, have proved economical in many assem which should be softer than the screw point by at least 10 points, Rockwell
operations involving plastics, die castings, and sheet metal (any kind). T: C, or else the capacity is greatly reduced.©-1*! Without spotting holes, the
and Z are applied with a screw driver; type F is made in standard thread pi relative capacities for other points, with an index of 1 for a cup point, are:
with fluted threads at the start for the purpose of cutting its own threads a: cone point, 1.07; flat or dog point, 0.92; oval point, 0.9. The cone point
moves into a drilled hole. These screws are hardened, and they are available w;
and dog points are often assembled with the point in a drilled hole, in
the Phillips recessed head, Fig. 5.24. There are many other styles. See Re
(5.14) and (5.28). which case, relative motion is a matter of shearing the point, and Table
A lag screw, Fig. 5.25, a large wood screw, is used to fasten machine 5.3 does not apply. Also, if the set screw has a slotted head, the tightening
equipment to a wooden base. On account of its size, it has a square or he: or seating torques indicated are not attained (perhaps only half the table
head allowing it to be turned with a wrench. A turnbuckle, Fig. 5.26, is capacity). Lubricating the threads, or plating them, the plating acting as a
venient device used to adjust the length of tie rods, etc. The stages in the fo rg lubricant, increases capacity because, with less friction in the threads,
and finishing of a blanked nut are shown in Fig. 5.27. i the normal force will be greater, equation (5.2). Sometimes, trouble with
set screws loosening can be remedied by enlarging the hole and using a
larger screw. Typically, the size of set screw is about } times the shaft
diameter. Two set screws on the same side of the shaft, side by side, virtually
5.14 SET SCREWS. Set screws, Figs. 5.21, 5.22, and 5.23, are us double the capacity for one screw, but at 180° apart, the increase is only
to prevent relative motion between two parts that tend to slide over 6 about 30 %5-14] because of the simultaneous loss of friction between shaft
another. They are obtainable with several combinations of points and heat
the more common point forms being shown in Fig. 5.29. An unprotect
square head should not be used on moving parts, since a workman m TABLE 5.3 HOLDING CAPACITY
catch his clothes on the projecting part and be injured. Slotted set OF CUP-POINT SET SCREWS
for a screw driver or hollow set screws, Figs. 5.22 and 5.23, are prefer
Approximate “‘ultimate”’ values of the tangential holding force at the surface of the shaft
for safety. The knurled point of the screw in Fig. 5.22 is designed to | when the tightening torque is as given; for cup points with hardness about Ry = 45-50,
on a shaft of hardness about R, = 15. A factor of safety should be applied. (Taken from
Standard Pressed Steel Co. pamphlet.)
FIGURE 5.28 Machine-screw Heads. The dimensions are for drawing purposes 0m
TIGHTEN- TIGHTEN- TIGHTEN-
SCREW ING HOLDING|SCREW ING HOLDING| SCREW ING HOLDING
SIZE TORQUE FORCE | SIZE TORQUE FORCE | SIZE TORQUE FORCE
in-lb. 1b. in-lb. Ib. in-lb, Ib,

0 0.5 50 8 20 385 4 620 3000

1 1.5 65 10 33 540 & 620 3500
2 1.5 85 + 87 1000 § 1225 4000
3 5 120 ts 165 1500 Z 2125 5000
4 2 160 # 290 2000 t 5000 6000
S 9 200 i 430 = 2500 1 7000 7000
6 9 250
 STRESS FASTENINGS AND RIVETS [Ch.5 95.19] RIVETS 179
178
and hub that exists opposite one screw; this friction thus contributes sig- 5.17 THREAD-LOCKING DEVICES. Locking nuts and ways of
nificantly to holding capacity. locking nuts have been invented in amazing numbers. Most of them depend
on friction to retard or avert the loosening of a nut under vibration. There
are other miscellaneous ways, such as nuts held in place by wiring, castellated
nuts held in place with cotter pins, Fig. 5.10. It has been found that when
5.15 DEPTH OF TAPPED HOLE AND CLEARANCE AROUN the initial tightening load is greater than the external load (close to the
BOLT HEAD AND NUT. The length of thread contact in a tapp
proof load), it is helpful in maintaining a tight connection. A few thread
hole should be a minimum of about 1.5D in. for cast iron and other brittl locking methods are mentioned below.
materials, and about D inches for steel or wrought iron (D = noming
Lock washers of the type shown in Fig. 5.32 come in four weights, are heat
size). If a tapped hole cannot go all the way through the piece, the hole tob
treated (45-53 Rockwell C), and their purpose is to maintain pressure between
tapped should be drilled at least an extra D/4 inches deep to allow toa
the threads of the bolt and nut, thus retarding loosening under vibration.
clearance at the bottom. The designer should be sure that heads or nutsf Lock nuts, to be effective, must have the upper nut, Fig. 5.33, drawn up so
be tightened have sufficient clearance for a wrench and are readily ac tightly that the nuts press on the threads in opposite directions. The “‘inside’’
cessible. nut (lower nut, Fig. 5.33) may be a jam nut, which is about 70% as thick as a
regular nut. Lock nuts are often not satisfactory because the “‘outside’’ nut is
not drawn up as stated above.
In the Flexloc locknut, Fig. 5.34, the sections at the top, separated by radial
5.16 BOLTS AND SCREWS IN SHEAR. Whenever the bolts ai slots, have been permanently deformed inward; they grip the bolt and hold the
to carry a shearing load the holes should be accurately sized and the bi nut in any position.
preferably have a close fit, which necessitates reamed holes and fini In the Lokut locknut, Fig. 5.35, the top of the nut has been deformed inward,
bolts. The tightening-up stress is usually neglected in the case of shear the idea being that the pressure of the deformed part on the bolt threads will hold
the bolts are larger than } in., but if desired, the resultant maximum s the nut in place.
The Esna insert type, Fig. 5.36, has a ring insert of fiber or nylon at the top;
with the nut on the bolt, the insert presses about and against the bolt threads,
holding the nut in place at any position.
FIGURE 5.30 End Bez The An-cor-lox nut, Fig. 5.37, has a ring of soft metal, such as soft steel or
a Shaft. This figure shows t brass, at the bottom of the nut which turns on freely until the nut strikes the part.
of dowel pins to carry the sh When the nut is pulled tight, the ring of soft metal flows about the bolt threads,
load. Dowel pins, which sho exerting considerable friction. Similar designs, perhaps with nylon, are also used.
on different diameters, are The speed nut, Fig. 5.38, provides excellent locking properties for light assem-
also to determine and blies; when pulled into a flat position, it presses against the root of the screw
the proper alignment be thread and also against the side of the thread by the nut’s ‘“‘spring”’ action; the
parts. See §8.11 for the me
resulting friction holds it. There are innumerable designs of speed nuts, which are
determining a tensile load
_ Widely used.
Dowel Pin bolts in this case.

5.18 DARDELET RIVET-BOLT. The Dardelet rivet-bolt, which has

due to combined shear and tension may be found according to the pi a self-locking thread, Fig. 5.39, is used in place of rivets in assembling
ciples explained in Chapter 8. If through bolts are used, arrange for | Steel structures. As suggested by Fig. 5.39, the thread has a wide, tapering
shear to be across a major diameter, but the shear may be across the n Toot. At the deep end of this root, the tapered crest or the nut thread has
diameter, especially for studs and cap screws. If the bolts are well tightenet Clearance, and the nut turns easily on the bolt. When the nut is drawn tight,
friction contributes considerably to the holding capacity, so that the actué the two tapered surfaces are pressed together, Fig. 5.39(b), and friction
shear stress is small. holds the nut in place.
When the location of a bolt is such that it would normally be in sheat
it is better practice to use dowel pins, Fig. 5.30 to carry the shearing loa¢
With dowels, it is not necessary to take extra precautions to get a clos 5.19 RIVETS. Most parts that can be bolted together can also be
fit for the bolts. Tiveted, but of course rivets are not used unless it is expected that the parts
 18D (Seen
0.87D
fom\ § 2
Srey
a ame
Grp
Me

!
2
cont
pat
o
SSS Ay» SAS

FIGURE 5.31 Self-aligning Nut.

(a) Cone Head = (b) Pan Head (c) Button Head (d)Countersunk (e) Semitubular
Head Rivet

FIGURE 5.40 Rivets with Some Approximate Proportions of Heads. (ASA B 18.4-1950).
Tolerances on D are of the order of +0.03 in., less for smaller sizes, more for larger ones.
In (e), the “‘heading”’ is accomplished by bending back the hollow part.

will seldom if ever be disassembled. As is true of threaded parts, there are

many styles and types of rivets. The heads in Fig. 5.40(a)-(d) are “‘regular”’
FIGURE 5.32 Spring Lock Washer,
ance Exaggerated.
styles; these rivets are inserted in the holes and then the opposite ends are
headed by machine (hammer or high pressure). The enlarging of a part of
the hole for the countersunk head may excessively weaken the connected
part; hence this type of head is used only when a flush, or nearly flush,
surface is desired. The semitubular rivet of Fig. 5.40(e), drawn enlarged,
is an example of small sizes used for thin material, as aluminum sheet.
There are a number of patented styles that can be applied and headed from
one side only of the connection,'©-12:5-141
The material of rivets for ordinary steel construction would typically
be a soft steel, as 1010, but copper, brass, aluminum, titanium, and others
FIGURE 5.34 Flexloc ® Locknut. (Courtesy FIGURE 5.35 Lokut ® Locknut. (
are used for a reason. Because of possible galvanic action, one must be
Standard Pressed Steel Co., Jenkintown, Pa.). Illinois Tool Works, Chicago).
cautious in using different materials together. ASME Boiler Code steel
will have a minimum s,, = 55 ksi; the design stress for shear is 11 ksi. Alu-
minum alloys for rivets include 2024 and 6061, Table AT 3; also in tempers
0 and H 13; minimum strengths for rivets will be somewhat less than
Table AT 3 values. For some fatigue information for lap connections,
see Fig. 4.27, which is as applicable for rivets as for bolts. Reference ( 5.30)
gives some design help for bolted and riveted connections, fasteners in
repeated shear. When the rivets fill the holes and clamp the parts tightly
SO as to induce a large frictional force, rivets that are nominally in shear are
FIGURE 5.36 Esna ® Insert Locknut. (Courtesy FIGURE 5.37 An-cor-lox ®
Elastic Stop Nut Corp., Union, N.J.). (Courtesy Schnitzer Alloy Products Strong in fatigue because the joint does not slip,'°-?! a condition that many
Elizabeth, N.J.). designers would not care to count on. Moreover, where the fasteners are in
Shear, one has to be on guard not to overlook the possibility of corrosion
and fretting. Tight clamping (bolts or rivets) puts the metal adjacent to the
hole in compression, which tends to improve the fatigue strength of the
Connected parts in tension.
The lengthwise shrinkage upon cooling of a hot-headed rivet induces a
tensile stress in it that may be close to the yield strength. Nevertheless,
tests on 3-in. rivets suggest :'5-29) hot-driven rivets are slightly stronger than
the original rod; rivets with long grips are not quite as strong as rivets
with short grips (varied from 2 to 6 in.); the ability of a rivet to resist an
FIGURE 5.38 Speed Nut ®. (Courtesy Tinner- €xternal tensile load is not reduced by the initial tension, probably because
man Products, Inc., Cleveland). of the high rigidity of the plate as compared to the rivet (§ 5.9). If a riveted
(a) Unlocked
connection is such that a repeated external loading tends
to put the ri d
in both tension and shear, the tensile loads
tend to relieve the frictions
force between the parts so that the varying shear
stress may cause troub le
in the Joint. At least, this factor must be cared
for. Connections involving
combined stresses are discussed further in Chapte
r 8. 1
5.20 CLOSURE. Variations in the forms of threaded parts
rivets are so numerous that other sources m ust be consulte d for moreif
are
descriptive detail.

6. SPRINGS

6.1 INTRODUCTION. Springs, which are common and important

machine elements, are used for many purposes: to absorb energy or shock
loads, as in automobile chassis springs and railroad bumper springs; to
act as a source of energy, as in clocks; to produce a pressure or force, as in
maintaining pressure between the friction surfaces of clutches and as in
keeping a cam follower in contact with the cam; and to absorb vibrations.
We shall discuss first, compression coil springs, which may be made of wire
with a round, square, or rectangular section.

_ 6.2 STRESS IN ROUND-WIRE HELICAL SPRINGS. Figure 6.1

shows certain forms of helical compression springs, illustrating four

FIGURE 6.1 Compression Springs. Since both ends of a particular spring are usually
alike, the variety of ends is for information purposes.

Plain End

183

(a) Coiled Right Handed. (b) Coiled Left Handed.

184 SPRINGS [Ch.6
FIGURE 6.3 Shearing Stress in Coiled Wire. This suggests the distribu-
methods of finishing the ends. The discussion of coil springs in this book ig
tion after the uniform or axial shear stress has been added to the torsional
limited to those in which the load is collinear with the axis of the spring (no
stress. Compare with Fig. 1.5, which shows the distribution of stress in a
bending of the spring as whole). Hence, for compression springs, we shall straight beam. The highest stresses occur over a small area, as at A.
have in mind ground ends. i
Take a quadrant of a coil as a free body, and imagine some ideal con:
Wahl’s factor; D,, = wire size; D,, = mean coil diameter; F = axial load
nection to it so that the load F is acting along the axis of the spring, Fig. 6.2
on the spring;
The force system external to the quadrant is then as shown, with the mom
FD,,/2 being resisted by the torque T at the section. If all the condition: 4C-—1 0.615
for T = s,J/c, equation (1.11), § 1.13, were met, the internal resistance = + —_,
would be given by s,J/c. A straight torsion bar, used as a spring ( (*) AG ae
AISI 9260, with 340 < BHN < 450), may be designed to meet the req where C = D,,/D,, and is called the spring index. Observe that the spring
conditions closely enough for engineering purposes, in which case, th index tells of the relative sharpness of curvature of the coil, a /ow index
torsional stress in a solid circular member is corresponding to high sharpness of curvature. The factor K is made up of
two factors,
(1.11) Te 16T
. sS=--—— = Ps
FR ae 8 (b) K= K,K,,
However, the peak stress in a coil spring is somewhat greater than tha where K, is the correction factor for curvature only, Fig. AF 15, and
obtained from (1.11) because: (1) the member is curved and there K, is the correction factor for the direct shear. For a spring index of C = 3,
curvature (stress-concentration) effect K, on the inside of the coil, Fig. the actual stress in the spring is some 60% (K = 1.58 in Fig. AF 15)
(2) there is a transverse shear stress on any section due to F (about 1.23 higher than that indicated by the simple torsion equation (1.11). To mini-
max.); (3) there is a compressive stress in the wire arising from mize the effect of curvature, we see from Fig. AF 15 that values of C > 5
component of F in the direction of the sloping coil; not to mention residu: are desirable; if C <5, special care in coiling will be necessary to avoid
stresses and some bending. Wahl’s'*-1) more complete stress analysis yiel cracking in some wires.
a satisfactory and practical result for closely coiled springs (within 29% Since equation (6.1) is for closely coiled springs, we must check the
more exact equations); pitch (or lead) angle. The distance between corresponding points of adjacent
coils is the pitch P (as in threads). If it is imagined that one full coil is
(6.1) 5, = aba = F Didaetiel unwrapped into a straight configuration, we note that it advances axially
wD,,° aD,,°
>

a distance equal to the pitch, Fig. 6.4. As the pitch angle A = tan~'P/(7D,,)
where K, obtained from Fig. AF 15 or from equation (a) below, is exceeds about 12°, equation (6.1) becomes less and less accurate.

6.3 DESIGN STRESSES AND SOLID STRESSES. As already

noted, the “strength” of a metal is a function of size, a variation that
becomes significant in spring design. From data available in the litera-
ture,{6-1,6.2.6.8] we find that an equation in the form

(c) 55= g ;
De
FIGURE 6.2 Quadrant of a Coil.
Seems to approximate various reported strengths, where the constants Q
and x depend upon the material. For this reason, Table AT 17 summarizes
a considerable amount of such information for wires with D,, < 0.5 in.
Note that smaller wires are in general stronger than larger ones. Column (3)

FIGURE 6.4 Pitch Angle.

 §6-9] Ne SS et Saas eee ee ee aed ee eed
eT NINO ta,

of Table AT 17 suggests design stresses for the static approach, on whie! rings, Lin. is the “active” length of wire and is nearly equal to 7DrNe
more is found in §§ 6.9 and 6.10. ai closely coiled springs, where N, is the number of active coils in the spring.
9
A compression spring should not be compressed solid (with cojj Substituting into (1.13) the proper values of
touching one another) in operation, because the surface of
the coils ma 7D,y*
be damaged. However, since compression springs are FD,
frequently c = > LweaD,N-, and J $
pressed solid in installation or maintenance, they should be 2 32
designed
possible so that such an occasional deflection will not damage them 5
inducing a permanent set. Table AT 17 gives values of the maxim
we get

“solid stresses,’’ which are in the vicinity of the torsional 9 (FD »/2)(7 Dm N-)(32)
yield stre radians.
The “‘solid stress”? should be checked for each design. If this requirems (e) eh GrD,,*
is not easy to meet, the manufacturer can help by presetting
(§ 6.13),
Larger wires, over about D,, = 3-}in., are coiled hot (hot woun The angular deflection in radians times the mean radius of the coil will give
to avoid cracks (and heat treated after coiling). One manufactu the axial deflection of the spring; 5 = @D,,/2. Therefore, multiplying both
recommendations®:?) for the permissible solid stresses in alloy
stee sides of equation (e) by D,,/2 and simplifying, we find the spring deflection
(6150, 9260) are approximated by (also use G = 10,500 ksi) OD, SFDAN,” 8FCON, wh
(6.2) $=
117 Dl ea GDy
Ss, = ———
D,,0:31 ksi, [D, > 0.375 Substitute all linear dimensions in inches in this equation. Wahl found
when equation that the actual deflection agreed well, within 1 to 2%» with values calcu-
(6.1) is used to compute the stress. For SAE 1095, us
straight-line interpolations between the following values: lated by this formula, provided that G, Table AT 17, is accurately known
; and the elastic limit is not exceeded. Note that, given the dimensions and
D,, = 0.5 in., 108 ksi; D,, = 1 in., 95 ksi; D,, = 2 in., 86 ksi. material, the scale, when constant, of a coil spring F/8 can be computed
from (6.2).
For extension springs, § 6.21, use 0.8 of these values. A fatigue stre
Sno © 70 ksi (§ 6.6) should be attainable for hot-wound springs.
6.6 ACCOUNTING FOR VARIABLE STRESSES. Because many
springs are subjected to repeated loading, it is logical to use the principles
6.4 SCALE OF A SPRING. The scale k of a spring is found accord of Chapter 4 in their design, especially when indefinite life is desired. Since
to Hooke’s law, force per unit of elastic deflection; or the average
valu springs are rarely subjected to reversing stresses, Wahl proposed a failure
F line on the 54-5, diagram that goes from B, Fig. 6.5, where the mean stress
(d) k= (usually Ib./in.) is equal to the variable stress (R = 0), to T at the yield strength s,,. See
6 Ad
Fig. 4.5(b). The maximum stress at B, Fig. 6.5, is
where F is the total force that produces the total deflection 8 in the spring
and AF is the increase (or decrease) of force corresponding (f) Ssmax = Sms + Sas = 2Sas = 28ms = Sno»
to an incre:
(or decrease) of the deflection A8. Other names for the scale applied where s,,. is the shearing endurance strength for a stress from zero to a
springs include the modulus, the rate (used especially when maximum. From (f), we see that at the failure point B, the mean stress
the scale is
constant), the spring constant, and the spring gradient. It is often a
rathel
important parameter. Sno
FIGURE 6.5 Variable Stresses in Springs. Point noe >
C is at Sas = Sns (Sms = 0), reversed shear. ir aS! ‘
~ Experimental
6.5 DEFLECTION R=-1
OF ROUND-WIRE HELICAL SPRINGS. Curve
torsional deflection
The
YY UL DIR
9 7 Sno H = Wahl Line
(1.13)
Te
@ = —radians,
o/ 3
GJ Z
ye F Dyk,
y Bas
A
D LL T
is applicable in general to round members. If this equation is applied to”
oO 45° My, Qy,

Se ee
Sys
N
Sys
 FIGURE 6.6.

particular deflection. In the absence of more specific information the

torsional yield strength s,, for no preset (§ 6.13) can be estimated as the
than one from C to T. To get a design line, divide s,, by a design factor values obtained from column (5), Table AT 17.
and lay off OD = s,,/N. Through D, draw the line DH parallel to T. Do not overlook a check for the “‘solid stress” as compared to the maxi-
Then, any point G on DH, represents a stress situation for which we s; mum permissible (yield).
the factor of safety is N. The equation of this design line HD is obtaine One often finds fatigue strengths given from some s,,;, to Some Sinaxs
from similar triangles QGD and MBT; ‘" as from a 10 ksi minimum. (See Fig. 6.9.) If so, a suitable design line for
other ranges can be found by using an actual failure point K, Fig. 6.6, whose
Sua) — Sms Sas coordinates are $/,, = (Ssmax + Ssmin)/2 and 5°, = (Ssmax — Ssmin)/2, draw
(g) - , line KT to sys, draw PQ parallel to KT such that Q is at s,,/N; then write
Sys — fale Sno|2
the equation of the line PQ by using similar triangles as before. The equation
from which analogous to (6.3), becomes

(6.3) ee =( - =) ee (6.4) ott e(i-

,

),
N Sys Sno Sys Sys Sno ;
N Sys Sas Sys
The alternating s,, and mean s,,, stresses are computed from equation (
pe 8KF.Dn 8KF,,Dm 6.7 ENERGY ABSORBED
Sas as = ———
- D,3 ms leah ya
Kn D,3
BY A SPRING. If a body with constant k
is gradually deflected and Hooke’s law is obeyed, the force F required at a
particular deflection is directly proportional to the deflection, and the elastic
where F, and F,, are the alternating and mean components of th
energy is the average force times distance, given by equation (4.8), § 4.35;
force. The reason for dividing by K, in the expression for s,,, is that
Wahl factor K includes a curvature concentration factor K, [Fig.
and equation (b)] that experience suggests (Chapter 4) is not needed (4.8) jie 2

mean stress; therefore, we use only K, = K/K,. Preferably, s,, and 2 2

experimental values for the size and kind of spring wire, but data on Sn; where kK = F/8 = AF/AS. In Fig. 6.7, we see that F8/2 is represented by
far from plentiful. The column headed “‘s,,”” in Table AT 17 gives enduran the area of a triangle OAD or OBC. Thus, if the force changes from F, to
strengths (unpeened) in terms of D,, within a maximum difference of S01 F, lb. (deflection, 5, to 5, in.), the work done on the spring between A and
3% from values by O. G. Meyers,'®-°] who determined the enduri B (or by the spring between B and 4A) is, from Fig. 6.7, p. 190,
strength for repeated (not reversed) shear by a correlation with r
bending tests. These results are not always in agreement with actual ;
(i)
F,6 F, + F; k
U, = — — sarge = (7 Je
but shearing endurance tests do not always agree with each other; _ 8;) = ger a 6,7) in-lb.

besides, one should not conclude that the stresses obtained from the expr
sions of Table AT 17 are exact. Endurance strength is discussed furth Most computations for elastic energy probably are made with (4.8) or (i).
in § 6.13. y However, for various kinds of springs, the energy can be expressed in terms
A definition of yield strength of springs used by the industry is t of certain mechanical and dimensional properties. For a round-wire, coiled
stress that results in a 2% decrease in the force exerted by the spring at Spring, the stored energy is obtained by using the value of F from equation
 §6.10]) EXAMPLE—AVERAGE SERVICE 91

FIGURE 6.7 Work Done on an Elastic Body. The area of the make several trial designs, and then choose the one that seems best. If
trapezoid ABCD is the average of the bases, (AD+ BC)/2, times the spring is in a cylindrical hole, a total clearance of D,,/2 (D,,/4 all around)
the height, 52—51.
should be sufficient in general. Special needs may dictate other values.
The simplest procedure is what we may call the “‘static approach,”
our Chapter 1 attitude. This is to classify the service as light, medium, or
severe (or something in between), and use a design stress from column (3),
Table AT 17, in equation (6.1). The other point of view uses the varying
load, our Chapter 4 attitude, and considers individually each factor that
the volume of wire if the spring is closely coiled is, for N, active coi
affects the operation, § 6.6. If a spring is not to operate an indefinite
V w& (xD,,?/4) (7D,,)N, in.?; N, active coils. An exercise for the readerj
number of times in its lifetime, it is wasteful to design it on such a basis.
to find ‘
In these cases, considering the information available, the static approach
seV 4K*GU, with light or medium service would seem appropriate. Sharp dividing lines
(j) U,= in-lb., or V= cu. in., cannot be drawn separating the types of service. Let us say that Jight service
4K?G 5,7
means that the load is applied not over 10* times; severe service corresponds
which is the volume of material to store U, in-lb. with the maximum s to indefinite fatigue life (10° or more cycles); medium service is something
5, as obtained from (6.1). Similar equations may be obtained similarly f in between, such as springs on clutches, brakes, switches. If indefinite life
other types of springs. If the spring is rarely to be asked to store U, in-l is desired, computations with the load varying would help in arriving at the
then the K may be taken as unity on the basis that the peak stress is hi final configuration; some design information for this approach is found in
local; of course, K is unity in a straight torsion bar. § 6.6, and its application is shown in § 6.11.
Many springs are made for which F is not proportional to 8 (§ 6.23 Always, the end use and the quantity to be produced affect the effort
which case the Fé curve (or equivalent) is needed in order to evaluate put into the design. If weight is a prime consideration or if production is
stored energy. We might note that the energy stored per unit volume large, it pays to spend more time on the design in search of the optimum
the stress is uniform is (§ 4.35): for a normal stress, s?/(2E); for a s answer. The following examples suggest ways in which the information
stress, s,7/(2G). The total energy stored is then, say for the normal st on compression springs can be used in design.
U = Jfs?/(2E)] dV, which could be evaluated when the stress is not unifor
The purpose of showing this relation here is to emphasize that for absorb:
the maximum energy with a particular volume of material, the s 6.10 EXAMPLE—AVERAGE SERVICE. Design a helical compression spring
should be uniform at its maximum value. Thus, a leaf spring, Fig. 6. with squared-and-ground ends to withstand a maximum force of F = 250 lb.
with a deflection of 1.5 in. The maximum number of applications of F is expected
absorbs much more energy with a given maximum stress than a solid p
to be about 10°; say, average service. The spring is to operate over a 1-in. rod,
of metal of the same shape. Members in bending and torsion inher
with a free length of 7 in. if possible. Use hard-drawn wire.
have nonuniform stress. Solution. There are somany unknowns that some assumptions must be made
' With iterations as necessary. Note that K, Fig. AF 15, does not vary greatly within
the normal range of springs; it is thus a good thing to assume; say, K = 1.3.
6.8 SOLID HEIGHT AND FREE LENGTH. The solid height ¢ Try Dn = 2in. From Table AT 17, ss¢ = 0.324s, for oil tempered, and 0.85
a coil spring is the over-all length of the spring when it is compressed times this expression for hard drawn, or ssa = (0.85)(0.324)(140)/Du°® =
all adjacent coils touch. The length of a coil spring under no load is 38.55/D,.°-1° ksi. Equating the design stress to the induced stress, equation (6.1),
the free length. Table AT 16 gives approximate values for different typ we have, for F = 0.25 kips,
ends, Fig. 6.1. 38.55 = K8FD,, es (1.3)(8)(0.25)(2) ksi
ss =
D,,°-18 aDw? 7D,,°
(TRIAL)

D,,?:** = 0,043 or Dw = 0.326 in.

6.9 DESIGN OF HELICAL SPRINGS. The design of spring From Table AT 15, the nearest W & M size is 2-0 or 0.331 in. If the original
usually involves a trial-and-error solution. In some cases, sometif assumptions are not too far off, this size may do very well; but a full check must
unfortunately, space limitations set limits for certain dimensions be made. The inside diameter of the coil is D, — Dy = 2 — 0.331 = 1.669 in.;
example, when a spring is to fit into a hole of a certain diameter. In the clearance between the spring and the 1.625-in. rod is then 0.044 in., which
event, when one or more unknowns must be assumed, the designer shoul May serve all right since the spring diameter will tend to enlarge as it is
192 SPRINGS [Ch.é §6.12] MATERIALS USED FOR COIL SPRINGS 193

compressed. The spring index is C = 2/0.331 = 6.04, for whichK = 1.25 from F; g. equation (6.3). Depending on how fortunate this assumption is, we might save
AF 15. The computed induced stress is time by following the procedure of § 6.10 in order to find the order of magnitude
_ (1.25)(8)(0.25)(2) _ ‘ of the wire size. A third way is to use the stress values in Table AT 17 and to
=
= (0.331)° = 44 ksi. assume values of K and K-, so that an equation in terms of D,, can be set up and
solved by trial and error. We shall use the third way. Assume K = 1.48 and
compared to the design (permissible) stress of K, = 1.29, corresponding to a spring index of C = 3.5, Fig. AF 15. The mean and
Ssqa =
38.55
D938 =
38.55 variable components of the load are
(0.331) = 47.5 ksi.

Checking the next smaller wire size, we find that the permissible stress is less tha:
Fein SUES Ss yay adres = ee = 79 Ib.
Then, from equations (h), the mean and alternating stresses are
obtained from (6.2); _ 8KFnDm _ (8)(1.48)(237)(1) a 691 0.691 ksi:
Sms = K.7D,,° a (1.29)7 Dw? D3 D,* >

_ 8GDw _ (1.5)(11.5 x 10°)(0.331) 8KF.Dn _ (8)(1.48)(79)(1) We ZoSmen. 0.298 , .

N. = 13 active coils.
~ 8FC3 (8)(250)(6.04)2 STE eget Dane ae eae A
From Table AT 16, the solid height is about From Table AT 17, use s,, = 87.5/D.°?® from column (5); try Sao = 30/D.°-*4
SH = D.AN. + 2) = (0.331)(15) = 4.96 in. from column (6). In equation (6.3), use Sms — Sas = 0.393/D,.° ksi and find (with
For a spring scale of k = F/8 = 250/1.5 = 167 lb./in., the force to compret kips for force)
the spring to the solid height is k(7 — 4.96) = (167)(2.04) = 340 lb. Since # 1 Sms — Sas 2Sas
stress is proportional to F, the “solid stress” is obtained by proportion Mone. che ee
5; = 44 ksi above; Ey 0.393 (2)(0.298)
13. D,(87.5/D.*) ~~ Dv*(30/Du***)
Stress at solid height = sao4 = 59.9 ksi
1 _ 0.00449 0.01985
compared to a permissible solid stress, column (5), Table AT 17, of 13. Du? ~*~ Dye
This is a reasonable form for iteration. After several trials with standard wire
Ss = sa = Daa = 86.4 ksi; sizes, we compute N = 1.37 from the right-hand side of this equation for
D, = 0.283 in., No. 1. This value of N is close to the desired 1.3 and we note that
which shows that the spring would not take on a permanent set if it were con we used the right expression for sn. to fit this wire size; but check with the proper
pressed solid. The equations are inaccurate unless the spring is closely co corresponding values of K and K-. Thus, C = Da/Dw = 1/0.283 = 3.54; this value
Checking this, first find the pitch, Table AT 16; is so close to the assumed value of C = 3.5 that we can read no significant
7-—2Dw 7 -—- 0.662 3 differences in K and K- from Fig. AF 15. However, since the computed N =
1.37 > 1.3, we try the next lower wire size, W & M No. 2, Dy = 0.2625. Using
- C = 1/0.2625:= 3.81, we compute N = 1.17, which is too low. Therefore, for
The pitch angle, Fig. 6.4, is Strength: D, = 0.283, W & M No. 1, oil-tempered wire.
Pitch 0.487 In order to find the number of coils, solid height, etc., more data are needed.
A == tan aD
Secon = tan In
ta eke = 4.5°,2 However, the procedure from here would be as before, § 6.10, including the com-
putation of the “‘solid stress” for comparison with the permissible.
which is well under the maximum of 12° (§ 6.2). The outside diameter of the spril
is D, = Dn + D, = 2.331 in. In practice, several other designs would be
perhaps with different D,, and/or grade of steel, and a final choice made. 6.12 MATERIALS USED FOR COIL SPRINGS. In general, steel
Springs are made of relatively high-carbon steel (usually more than 0.5%),
heat treated and/or cold worked to a high elastic limit. High elastic limit
6.11 EXAMPLE—INDEFINITE SERVICE. The load on a compressio)
is important in springs in order to get a large elastic deflection. Coil springs
are wound cold in sizes under about 3 to } in. and hot in larger sizes. The
tempered carbon steel, what size wire should be used? iq material may be heat treated (pretempered) before winding (in small sizes)
Solution. There is no easy way to solve for D,, directly. A diameter of fir or after winding. When heat-treated wire is coiled cold, it should be stress
can be assumed, after which the factor of safety N could be determined from Telieved for bending stresses after coiling by being heated at some 500°F
at SPRINGS [Ch, 36.13] FACTORS AFFECTING FATIGUE STRENGTH OF COIL SPRINGS 195

at
for 15 to 60 min., depending on its size. The wire sizes Dy given below
defects may not be significant. When fatigue is involved, the surface condi-
3 tion is primary. Any slight flaw, such as seams, pits, die marks, hardening
normal commercially available sizes.?-»}
wire (ASTM A227) is a low-cost material; size: cracks, inclusions, or an accidentally scratched spot, may result in fatigue
Hard-drawn spring
dimension: failure. For this reason, experimental endurance strengths of wires of a
0.028 to about 3% in.; suitable where service is not severe and
to 0.75 %C; not used for indefinit particular size often have a large natural spread. In fact, it is not at all
precision is not needed; wound cold; 0.45
tha certain that differences found in the fatigue strengths for different wire sizes
life. The quality of the surface is lower (with, say, hairline seams)
are significant; that is, fatigue is less dependent on size than on other
the other grades (§ 6.13). Cost index = quiere
Music wire (ASTM A228) is hard drawn also (80% reductio
n), but factors. For unpeened springs of high-quality carbon steel, some engineers
surface, compara ble to “‘valve-s let the maximum stress be 90 ksi with a range of 70 ksi; if peened, the
made of high-grade steel; excellent
is the be: maximum stress may be 110 ksi.'$-12! ;
quality”; 0.7 to 1.0%C; wound cold; sizes, 0.004 to 0.156 in. It
4 in. Cost index 35,5030) 4 If a member is subjected to torsion, one principal stress is tensile
material available in sizes below about
Oil-tempered spring wire (ASTM A229) is cold drawn to size (reductic (Chapter 8). Shot peening, as previously reported, § 4.28, leaves a residual
0.55-0.75 compressive stress which opposes the principal tension. This residual
50-70 4) and then hardened and tempered (pretempered);
say 4 together with the residual compression “‘covering’” the flaws especially
usually wound cold and stress relieved at low temperature,
not the best, but it is significa ntly better the at the inside of the coil where the stress is a maximum), results in a much
sizes 0.225 to 0.5. Its surface is
1.5.70) improved fatigue strength (peening may be used when D,, > 0.0625 in.);
hard-drawn wire. Cost index
hi examples:'®-1] music wire and Cr-V 6150, D,, = 0.148 ie Sno = 70 ini
Valve-spring-quality (VSQ) carbon steel (ASTM 230) is the
excellent su when unpeened, 115 ksi peened; stainless 302, D,, = 0.148 Oe oe = 45 ksi
quality oil-tempered wire; 0.60-0.75 % C. Since it has an
when unpeened, 90 ksi peened; phosphor bronze, D,, = 0.148 in., the re-
valve-spring quality is the most reliable (with music wire) for fatigu
0.375 in. spective values are 15 and 30 ksi. In a torsion bar, the most improvement in
is thus used for the most severe service; sizes, 0.093 to
¥ fatigue comes from peening the bar while it is overstressed in torsion."*-**!
Chromium-vanadium steel (ASTM 231) is oil tempered; 0.45 to 0.55
carbon steels of Since the force on springs nearly always acts in the same sense, over-
sizes, 0.28 to 0.375 in. The alloy steels are superior to
same quality above about 250°F. Cost index 4,'610] Valve-spring-que stressing springs to induce favorable residual stresses, § 4.23, is common
0.032 to 0.4 practice.[°-0%) The process of obtaining the residuals, called presetting or
Cr-V (ASTM 232) has the best commercial surface; sizes
¢ setting out, is to have the coiled spring somewhat longer than desired, then
Chromium-silicon (ASTM 401), good quality for impact loads
the pern
moderately high temperatures (perhaps to 450°F, depending on
compressing it into the plastic stress range; after which, the spring should
relaxati on). Cost index 4,'6-19) .
sible amount of
Stainless steel, type 302 (Chromium-nickel, ASTM A313) is corro
is cold drawn |
resistant and readily available; sizes, 0.009 to 0.375 in. It
grades descr
its relaxation at higher temperatures is much less than the
compara
above. Cost index 8.5.'6:10) Stainless steel wire with strengths FIGURE 6.8 Effect of Residual Stresses. A torque has in-
to or better than music wire is available. duced the plastic stresses PP in a straight bar, which, when the
are used for ce
Other materials, some not mentioned in this text, torque is released, then has the residual stresses RR. Now some
see Table AT applied torque 7 induces the elastic stresses EE, s, = Te/J,
springs for a reason, as for electrical conductivity; but the net stress is obtained by algebraic addition of the re-
for sp
Occasionally, there seems to be a reason to use a plastic or glass Sidual stresses, to get the distribution WW. See Fig. 4.17.
(§ 6.3) and flat spr
Typical steels used for hot-wound coil springs
As we know, the all
include AISI 1095, 50B60, 6150, 8660, 9260, 9850.
alloy spring st
have the better hardenability (§§ 2.6, 2.7). In general,
or stronger than carbon ste
in small wire sizes are not significantly better
their better hardenab ility may “ a suitable length with favorable residual stresses, Fig. 6.8. Reasonable
in the hot-wound larger sizes, alloys with
; i oceee a practice be set-out springs is to increase the design stress
more generally advantageous.
ene up to 50% for static loads, but more modestly for fatigue
fatiyie te coud coil springs, increase the design stress for
AFFECTING FATIGUE STRENGTH OF CO i: oads by 25 7o> but make no change for static loads. Above about
6.13 FACTORS
SPRINGS. If the number of cycles of loading is small (light servic F for steel, the improvement from peening is rapidly lost. Since
surfé
the computed stress in the wire may be relatively high, and normal
Presetting of very hard steel (R¢ = 50) sometimes results in cracks in the
196 SPRINGS [Ch.6 96.16] TOLERANCES 197

direction of the principal tensile stress, peening should be done before pre= be preset at a temperature and stress above operating values, which
setting in this case; the residual compressive surface stress counteracts the “removes the set.”” By way of example, the load loss of carbon valve spring
tensile"! and cracking damage is avoided. However, if presetting is n steel at a stress of 80 ksi is:'°-?) about 3% at 250°F, 5% at 300°F, 6% at
so drastic as to cause cracks, the endurance strength is greater when peeni 350°F, 7.5% at 400°F. Normal limiting temperatures of operation are:
follows presetting than when it precedes."*:?®! The “‘solid stress’’ for pres 400°F for carbon steel, 550°F for 18-8 stainless steel, 700°F for Inconel,
springs may be about 0.60s,, without excessive setting. Lastly, presetting 500°F for Monel, 200°F for phosphor bronze. Note that the spring’s
may result in some 5-10
% reduction in the modulus of rigidity G.{45) rate is significantly lower at high temperatures (via decrease of G, E, § 2.21).
Larger, hot-coiled springs are inevitably decarburized. These springs
are then heat treated by quenching and tempering at 400-500°F. If such
springs are peened, the effect of peening should reach through the entire 6.15 GOODMAN DIAGRAM. A Goodman type of diagram is often
decarburized layer; if not, the improvement may be minor or zero because used for spring design, an example being Fig. 6.9 for music wire. Because
of the existence of heat treatment cracks in the vicinity of the inner boundary
of the decarburized zone (Coates and Pope":2*!). Evidently, the compressi 160
stress from peening about these cracks is sufficient to inhibit their spread.
Coates and Pope report fatigue strengths s, at 10° cycles as folloy 140
(D,, = 0.5in., s, = 56 ksi, 0.9%C, Ro ~ 50, C = 5.25, OQT 750°F)
as received, s, = 11.1; preset, s, = 13.1; peened and preset, s, = 15: 120
preset and peened, s, = 19.7 ksi. Decarburization reduces the effect FIGURE 6.9 Goodman Diagram, Music
Wire. Tests of coiled springs, unpeened; 100
peening (lower residuals) in any event because of the lesser response 185 specimens; stress relieved at 500°F
the material to cold work. If the surface is recarburized, shot peening may for 1 hr.; ASTM A 228. All test values fell z
result in the failure being initiated at some point below the surface;':28 above line AD. Knowing
either the maximum 3 =
which suggests the possibility of coiling the springs from as-rolled b or minimum stress, the other permissible g
followed by recarburization and heat treatment, stress relieving value is found from the diagram.'?+1) 60
peening. Stress relieving can be done at about 400-500°F with no signi
40 Music Wire AD
cant loss of the shot peening effect. 0.032 < Dw < 0.177
If steel is used in corrosive environments, various coatings may contro! 0 4.9<C<12
the corrosion. Plating, say cadmium, will provide considerable protection
(§ 4.30), but consider the effect on strength. Normal cleaning of high- 0
strength steel in acid, preparatory for plating, is accompanied by atomic
hydrogen diffusing into the material. Plating tends to seal the hydroger
in (and some plating, such as chromium, results in considerable hydr of the scatter of results, this diagram lets the safe range of stress 25,, be the
genizing), the consequence being hydrogen embrittlement. A more rece same for all the sizes tested. With sufficient test data, such diagrams may
development is a process of “mechanical plating,’ which is done by wet be constructed for a smaller range of wire sizes.
tumbling of the specially prepared springs in a mixture of metal shot, vee 1
metal powder (say cadmium), and a chemical “promoter.”’-)
6.16 TOLERANCES. In a particular application, the designer should
use close tolerances only where such tolerances are significant. The signifi-
6.14 RELAXATION OF SPRING MATERIALS. The relaxation is” cant dimension might be the outside coil diameter, the inside coil diameter,
sometimes measured by the amount of the loss of force exerted by the spring © or the free length. Perhaps more often, the significant property is the force
at a particular deflection, sometimes by the change of deflection for a that a spring exerts at one or more deflections, and all the dimensions,
particular load. Springs relax at temperatures above ambient, the amount together with the torsional modulus of elasticity, which is not strictly
being a function of stress, temperature, and material, and the spring is” Constant, affect the force. In a spring manufactured under controlled
said to set when this happens. Spring manufacturers often have enough~ Conditions, the wire diameter varies between certain natural limits (§§ 1.17
data to predict the amount, so that it can be allowed for. Also, springs may and 3.9); the coil diameter, the precise number of active coils, the torsion
 Pee Tolerance >
=11.7 Ib. +11.7 Ib.
. A a 233.9 §6.18] BUCKLING OF COMPRESSION SPRINGS 199
| Ee Bea FIGURE 6.10 Distribution of Spring Fore of mass by 8o = 386 in./sec.? Using the approximate volume of active wire
| z ' Specifications were 234 Ib. +5% = 222.3
bil tf 245.7 Ib. at 5= 1.812 in. The value of o = 4 as found in § 6.7, the mass of the active coils is
S| Spec, as computed from the actual data, giving
z +30 § / natural (6c) spread from 221.9 to 245.9 Ib. (k) n=—
oak a
f Poa mV, \b 54 Ib-sec.?/in.,
D,,N- 2/i [=psin—footnote, p. 149]
Be 1 a | Normal mean of 233.9. The process is rather well center
+20 3 srt Curve | Each cell, which is 2 lb. wide, represents the num-
| 6 | ber of pieces falling within this 2-lb. range. where p is in lb/in.® and the other dimensions in inches. Using the value of
| -10 ' |
k from equation (6.2), the equation for ¢ becomes

:223 | (6.5)
agar 12 GDyt —-(386)(4)_ 1
«227 231 235 «9239 243 «247 ne a =i(GoaH. Smee)
Force, lb. per 1.812-in. Deflection
2.21D, /G\1?
modulus, and the free length vary similarly. Each of these variables has ~ NeDp? (*) ea
some effect on the scale of the spring and on the force the spring exerts G psi, plb/in.* For steel with G = 11.5 x 10® and p = 0.284 Ib/in.°,
certain deformation. For example, the force at a particular deflection vari we get
as the fourth power of the wire size, equation (6.2); thus a small varia 14,050D,,
of wire size affects the scale greatly.
(1) $= cps. [STEEL]
N.Dy?
Figure 6.10 shows a histogram of the actual distribution of the forces If the natural frequency is 12 or more (some say 20) times the rate at which
exerted by certain springs produced in a lot size of 250 to a specificatio the spring operates, no trouble is anticipated from surging.
234 lb. + 5% at8 = 1.812 in. Observe, by comparison with the corresp
ing normal curve for this histogram, that this tolerance is rather clos
since the process must be practically centered and always in contr 6.18 BUCKLING OF COMPRESSION SPRINGS. A compression
to meet the specification. Thus, a tolerance in this instance of somethit spring whose free length is more than four times its mean diameter should
greater than +5% would probably lower the cost. The minimum tolerant be checked for buckling. Enter Fig. 6.11 with the ratio (free length)/(mean
on force for hot-coiled as-rolled rod should be about + 10%, and for smé
wire, say under about 0.32 in., the tolerance may need to be larger
economic reasons, because the percentage variation of D,, will be lar; 0.7
0.6

Critical Deflection
0.5
6.17 SURGE IN SPRINGS. Unless the natural frequency of a sp S04
is quite different from the frequency of the applied load, some reson
may occur, in which case ‘‘waves”’ travel along the length of the spr
SIE 03
These waves are successive compressions and extensions moving from 0.2 < *
Xe, d
to coil, and they may result in deflections of adjacent coils equal to t 0.1
deflection when the spring is compressed solid. Accordingly, the spring FIGURE 6.11 Buckling Conditions.(®-1: °-21 0
then be subjected repeatedly to the stress corresponding to the stress 2 4 6 8 10
Free Length
solid compression. Since this stress is often above the endurance stre
Mean Coil Diameter
and since repeated contact of the coils will spoil the surface, surging may'
the cause of failures where the calculated stresses for normal loads
apparently safe. The lowest natural frequency (first harmonic) of a sp diameter), move vertically to the curve, then horizontally to the ordinate
is given by ¢ = (k/m)1/2/2 cycles per unit time,'®+) where k is the sp and read the value of (deflection)/(free length). If the deflection per inch of
scale and m is the mass of the active coils, k and m in consistent units. free length in the actual spring is greater than the chart value, Fig. 6.11,
order to match pounds for force, inches, seconds, the mass must have ti buckling with ends supported as stated should be expected. “Ends Fixed”
units Ib-sec.?/in. (from m = F/a), which are obtained by dividing pounds 1s for the case where the squared and ground ends are on rigid parallel
eve SPRINGS (CC Goel Tae Eee eee See

surfaces perpendicular to the spring’s axis. The middle curve is for one nc found in Fig. AF 15. Note that when the square wire is bent, it bulges on
on a rigid surface, one end hinged. The bottom curve is for the case wher the inside, the section becoming approximately trapezoidal. Wahl gives
both surfaces on which the ends rest are hinged ona pin. Ina commor more accurate means of computation in Ref. (0.3).
hinge equivalent, the spring rests on a ball, as in a ball valve.

6.21 EXTENSION SPRINGS. The foregoing equations for compres-

6.19 CONCENTRIC HELICAL SPRINGS. Two concentric (nest sion springs apply as well to extension springs, except that allowance is
springs may be useful in carrying heavy loads or in eliminating a vibrati made for the initial tension, if any. Extension springs are generally wound
such as surge. For a concentric arrangement of springs, there are with the coils pressing against each dther, and the initial tension is the force
conditions that may be, though not necessarily, approximately satisfied on the spring when the coils are on the point of separating. The amount of
first, the stress in each spring at any deflected position may be the same the initial tension can be regulated to some extent and it varies from
and second, the free heights of the springs are usually to be the same manufacturer to manufacturer, but reasonable maximum values of the
These conditions are approximately obtained in a round-wire spring j corresponding stress are as given in the following tabulation"®-*! [as com-
both springs have the same index C; 7 puted from equation (6.1) with K = 1]. Lower values should be
specified. 4)
Dna Dna
(m) C=
Dyi Dye
>

Ca 4 5 6 7 8 9 10 11 12 15
where subscripts 1 represent one spring and subscripts 2 represent the other 5, 24,000 | 22,500} 20,000 | 18,000 | 16,250 | 14,500 }13,000 | 11,600 | 10,600 | 9,700 | 7,000
For these conditions, the outer spring takes about 2 the total load w
C & 6, but the proportion is different at other spring indexes. The des
procedure may involve extensive iteration to obtain the desired balance Hooke’s law does not apply until the initial tension F, is overcome.
properties; Chandler'®*”) gives considerable help in the way of tabu After the coils are separated, the stress may be computed from (6.1) for the
matter. external load F. If the scale of the spring is estimated from (6.2), the load
F on the spring is the initial tension plus k8, F = F, + k8, where 8 is the
extension from its unloaded length and F, = 7s,D,,5/(8D,,). If there is no
6.20 COMPRESSION COILED SPRINGS WITH RECTANGULAI initial tension, the equations for compression springs apply without
WIRE. Wahl!®-1) gives the following formula for the maximum modification—provided that the coils are not extended to a pitch angle A
shearing stress in a spring of rectangular wire when the dimension b of the greater than about 12°. Extension springs should be designed to operate
rectangle is parallel to the axis of the spring (and when b/t < 3 if b is the with some extension at all times.*
larger dimension): ] The weak point in an extension spring is likely to be where the coil is
bent to make a hook, Fig. 6.12. In case of fatigue loading, the radii used
(n) 5, =
K,FDn(3b + 1.81) : [b PARALLEL TO in bending the wire to form the ends, Fig. 6.12(b), should be as large as
2671? possible. The maximum stress at a section B, Fig. 6.12(b), may be estimated
K, is obtained from Fig. AF 15 for C= Dait. For a small pitch angi from s, = 8K,FDn/(7D,,°), where K, is for C = 2rm/D, from Fig. AF 15
(<10° ), approximately . 1 * Given an extension spring, to find the initial-tension force Fi: extend it 0.1 in.,
2.45FD.3N Measure Fi; extend it to 2 = 0.2 in., measure F2; then F; = 2F: — F2. Prove this.
(0) S m e

~ “Gib — 0.561)’ FIGURE 6.12 Extension Spring. Other shapes of loops or hooks are readily available.
which will serve also for a spring of square wire (b = 1). From equation (m), (Courtesy Associated Spring Corp., Bristol, Conn.).
the stress in a square-wire spring is if
2.4FD > Length over All Including Loops rt F
(Pp) =o Ry
Lape eh e Size of Hook Openi ing, least of_.
Length Wire ae
Diameter
" , y f Diameter
where 6 is the average length of the side of the square section and K; is Diameter” AVVVFFE EY ; Dace T% <

wy
Full Round Hook Full Round Loop

(a) (b)
Bae FIGURE 6.13 Threaded Plugs.
§6.22] TORSION SPRINGS \ 203

Preferably, the working force winds up the spring, thus taking advantage
of the residual stresses. If the working loads act oppositely, the spring
(actually, both bending and torsional stresses exist in this vicinity). The should be stress relieved. For repeated loads, watch for sections of stress
flexural stress in the loop itself can be computed using K values for a curved concentration where the ends are formed.
beam (§ 8.25) and a maximum bending moment of FD,,/2. Let the design” The angular deflection of a torsion spring subjected to opposing
stress for the coils be about 20-25% less than for a compression spring. moments T = M at the ends is
There are numerous ways of finishing the ends of extension springs, among”
which is the idea shown in Fig. 6.13. id y _ ME _ MrDaNe
radians,

6.22 TORSION SPRINGS. The preceding discussion of springs where the length of active coils L ~ 7D,,N,; I is the rectangular moment
assumes in each case that the load is axial, inducing a torsional stress in th of inertia of the section of the wire taken as a beam. Because of the
spring. A torsion spring, Fig. 6.14, is one that resists a moment tending to “winding up” of the spring to a smaller diameter, the actual deflection is
wind up the spring. Due to the variety of ways in which the spring may be
loaded, the calculations considered below are merely suggestive.
If an end coil is attached to a disk or to some member that imposes a
pure torsional moment on the spring, the coils will be in pure bending.
If a force F is applied at the end of a coil, as in Fig. 6.14, then, in order for
the maximum stress to approach a bending stress only, a core or arbor
preferably with a diameter greater than 9D,,, must be used inside the sp FIGURE 6.15 Torsion Springs.
in such a position that it bears on the spring at the section B, Fig. 6.14, With the ¢ dimension as in (a), the
spring is stiffer. J =br3/12,
(a) (b)

somewhat less than this theoretical value.'°-12! In a long torsion spring,

6 may be the equivalent of several complete turns. In Fig. 6.14, the deflection
of the point of application of F isa. The work done on the spring from
F = 0 to Fis (F/2)a0 = 76/2.
FIGURE 6.14 Torsion Spring. This figure illust
Some authorities recommend design stresses about 60°% greater than
| ends of special design. (Courtesy Associated Sp
Corp., Bristol, Conn.). those defined for compression springs in column (3), Table AT 17. Basically,
with the stress being flexural, something less than the tensile yield strength
(so that there will be no permanent set under operating conditions)
represents a limiting value. Reported values of this strength in wire sizes
are not explicit: for oil tempered, use 0.8s,; music wire and hard drawn,
produce a reaction F’ = F. In this case, the torque is T = Fa = M, th
0.65s,; alloy steels, 0.9s,; stainless, A 313, 0.55s, (greater variability than
bending moment, where a is the moment arm of the couple F—F’. We the others); brass, 0.42s,,; phosphor bronze, monel, and beryllium copper,
dealing with a curved beam (§ 8.25), in which the curvature acts as a st 0.67s,,. If the s, values in Table AT 17 are used to obtain design stresses,
concentrator (shifts the neutral axis). Thus, s = Mc/J is modified by a stres note that they are specification minimums.
factor K,, obtained for a rectangular section from Fig. AF 15 (or fr
For torsion springs hot-wound from rods, the equations
Table AT 18 for other shapes); thus, the flexural stress is
(s) 120 132 t
K,Mc s= Dp 0.26 and s= Pp aa Ke [0.5 < Dy < 1.5]
(q) oi e f
I [SAE 1080] [6150 ap 9260]
I/c = wD,,3/32 for round wire and J/c = bt?/6 for rectangular wire, 8ive, within 3%, stresses that Associated Spring '-2) calls “maximum”
Fig. 6.15. In general, a force-and-stress analysis is made to fit the situation.
(say proportional limit) when the surface BHN ~ 437.
 t
204 SPRINGS [Ch. 6
F F F F

6.23 OTHER KINDS OF SPRINGS. As might be surmised from h

Fig. 6.16, springs are used in a myriad of different forms, and we have
space for treatment of only a few. Additional fairly common forms include:
Garter spring, the helical coil is wrapped into a circle forming an annular © FIGURE 6.17 Belleville Springs. (b) Series
(a) (c) Parallel
ring; used in tension to hold oil seals of leather, Tefion, etc., tight against ~
a surface; used in compression as piston-ring expanders.
Motor spring, thin flat strip wound up on itself as a plane spiral, usually In either case, a stack of belleville springs can absorb a relatively large
anchored at the inside end; an energy source to drive clocks, toys, etc. 3 amount of energy per unit of volume occupied.
Hairspring, wire or strip rolled into a plane spiral, no contact between A Neg’ator spring, a patented form, exerts virtually a constant force F
coils, a sensitive spring used in instruments, watches. ‘4 (zero rate) after a certain initial deflection. The basic idea is that flat strip
Volute spring, wide-strip material wound in a conical helix (frustum of is made into a plane spiral coil, its natural form without external loading,
a cone with long dimension parallel to axis) with strip overlapping; used so that when it is straightened, it exerts a force on account of its tendency

FIGURE 6.18
Neg’ator Springs.

(a) Motor (b)

to rewind to its original form. As a motor, Fig. 6.18(a), it exerts nearly a

constant torque on the output drum; it is made in many other forms to
exert constant force, one of which is suggested by Fig. 6.18(b). Its uses
include: exerting constant pressure on commutator brushes; in retriever
reels that are used for returning many things to a desired position, as
retracting the gasoline hose after the fuel has been delivered. The manufac-
turer has prepared engineering data for design purposes.'-15! The same
manufacturer makes a constant-force (nearly) compression spring, called
a Flex’ator, which is basically a tightly coiled helical spring loaded by an
where compactness, frictional damping of the turns rubbing on one another,
4 eccentric load that bends it. Constant force action can be obtained with
and a spring rate increasing with deflection, are advantageous.
ordinary compression and extension coil springs through intermediate cams
Conical spiral spring, also wound in the form of a conical helix of round
if designed to deliver a constant torque to the shafts on which they are
wire; not a constant rate spring. mounted, {6-211
A belleville spring is one made in the form of a dished washer, Fig. 6.17,
Rubber springs, and other materials with similar properties, are used in
and has been adapted to many uses. The theory'®-1®) is too extensive for
mountings to isolate vibrations (metal springs are too). The modulus of
presentation here, but Wahl has presented charts that reduce the computa- elasticity is not constant and the load-deflection curve depends on the
tion time materially.{°-3-°18! The load-deflection curves vary widely in” particular compound, but the material has a high capacity per unit volume
shape as the ratio h/t, Fig. 6.17, varies. For h/t ~ 0.5, the curve is approxi- for storing energy and advantageous damping properties that are absent
mately a straight line up to 8 ~ ¢/2. For h/t ~ 1.5, the load is constant — in metal springs; used in compression or shear. These materials are readily
for a considerable deflection (zero rate) after a certain initial deflection. bonded to various metals; and are widely used in connections such as
These springs are often used in stacks of two or many more. When stacked automotive spring shackles and other joints with limited relative motion.
in series, Fig. 6.17(b), a larger deflection results for a given load (lower rate).
(See problems in Slaymaker."-15})
A parallel stack, Fig. 6.17(c), supports a larger load for a given deflection. ©
205
SS Section Section
(a) (b)

FIGURE 6.19 Flat Springs.

- A
FIGURE 6.20 Cantilever of Uniform
6.24 FLAT SPRINGS. Flat springs may be in the form of a cantilever Strength. =
beam, Fig. 6.19(a), or in the form of a simple beam, Fig. 6.19(b). The stress
and deflection of springs like these are computed from formulas given
Table AT 2, just as for any ordinary beam. In the cantilever beam, t
maximum stress occurs at the point of support B in Fig. 6.19(a). Sin
the bending stress for a constant cross section diminishes from B to the
point of application of the load F, the section of the beam may be diminishe
in such a manner that the maximum bending stress on each section is t
same. The resulting beam is called a beam of uniform strength. For examp)
a beam tapered in the plan view as shown in Fig. 6.20 has the same bendi
stress in all sections, and for a particular material, it will support just
large a load with a particular maximum stress as the beam of Fig. 6.19(a), FIGURE 6.21 Simple Beam of Uniform i a
Strength.
provided the length of the beam and the dimensions at the section B are the | A
same, with a 50% saving in material. Since a shearing stress exists, the FR Fr
beam must not come to a point where the load is applied, because there en 3hL jun 3FL
would be no area to resist the shear. An important difference is that 2ohF 4” BEDI
spring of Fig. 6.20 deflects more under the same load than a beam of co:
stant width b—a distance of 6FL®/(Ebh®) compared with 4FL*/(Ebh’).
Tables AT 1 and AT 2. With a larger deflection, the energy that can
absorbed (F98/2) is greater than that which a beam of constant section
absorb, with the same maximum stress in each beam.
Figure 6.21 represents a simple beam of uniform strength. Sim FIGURE 6.22 Deriving Leaf Springs from
points of comparison may be made for the two types of simple be Beams of Uniform Strength.
Figs. 6.19(b) and 6.21, as for the two types of cantilever beams. The
nominal stresses in these beams may be computed from the bending moment
formulas = Me/I, and the nominal deflections at A of the beams of unif
(a) (b)
strength are as shown in Figs. 6.20 and 6.21, provided that the deflection
not large enough to alter the moment arm of F significantly. These sprin
follow Hooke’s law within the elastic limit and the scale is k = F/6; t _ Itis necessary to modify the springs of uniform strength described above
maximum stored energy is F8/2. ' in order for them to be practicable. For example, the simple beam of uniform
Strength becomes a semielliptic leaf spring, various details of which are
Suggested by Fig. 6.23 (p. 208) of an unloaded spring. The camber, Fig. 6.23,
6.25 LEAF SPRINGS. If the flat springs of uniform strength just 1s frequently made such that the main leaf is nearly straight under load.
described are divided as indicated by the dotted lines of Fig. 6.22, a An approximation of actual semielliptic springs which can be presented
assembled as indicated by the solid lines of Fig. 6.22, the same stress ane easily is the trapezoidal spring, Fig. 6.24 (p. 208), for which'®-1!
deflection formulas as given in Figs. 6.20 and 6.21 apply, friction betweer
the leaves neglected. (Friction results in the computations being inherently 6FL 3WL and 3- KiFEG3/1 =_ #9), 2 KWL S748
— 2)
(t) s=
less accurate than those for coil springs.) The result obtained is a leaf sp bh? bh? 3EI 6EI
with all leaves of the same thickness, b in the formula being equal to the
sum of the widths of the leaves; that is b = Nb’ where N, is the number o} where W = 2F is the load at the middle of a simple beam of length 2L
leaves. (F is the load at the end of a cantilever of length L), Fig. 6.24; b = Nb’,
206 207
 §6.27] GENERAL REMARKS ON LEAF SPRINGS 209

where b’ is the width of a leaf and N, is the number of leaves; u is Poisson’s

ratio and the term 1 — y? applied when the spring’s width b is large as
compared to its thickness h, in which case lateral expansion or contraction
of elements near the surface is hindered, resulting in a somewhat stiffer
spring than simple flexure theory predicts; J is the moment of inertia of the
leaves at section B; K, depends on the ratio of b’/b and is taken from Fig.
6.25. In (t), the equation for 5 gives the deflection of the end of a cantilever
spring (section B, Fig. 6.24, fixed) when F = load on end and L = length
of cantilever.
FIGURE 6.23 Leaf Spring. Several styles for finishing the ends of the leaves are illus-
trated; an unlikely combination in practice. The full-length leaf under the main leaf is ofte: y
square cut even though the other leaves may have egg-shaped and thinned ends. Beware of — 6.26 FATIGUE OF LEAF SPRINGS. Leaf springs often have holes
the stress concentrations in the vicinity of the clamps (U-clips). Interestingly, Almen(4-64]
points out that when the bushing is pressed into the eye as shown, the resulting deformatic .
or notches that are points of stress concentration, and the principles of
induces a tensile stress in the top surface of the leaf at this point, where the working stresses Chapter 4 apply. Even though the clamping action of the center bolt and
are also tensile. On this account, he recommends making the eye by bending, for example, U-clips reduces the bending stresses at the section of the bolt hole, Fig. 6.23,
ul
CL on the right instead of CC. i it would be on the safe side to check this section in accordance with fatigue
procedures when the load is repeated. Rubbing of the leaves is conducive
to fretting corrosion (§ 4.27). Data are too meager for generalization, but
tests of 0.5-0.6% C steel show an actual strength reduction factor of
K, ~ 1.4. Sharp edges on the leaves should be avoided in severe situations.
For leaf springs made of as-rolled material, a loss of fatigue strength of
the order indicated in Fig. AF 5, should be expected. Typically, the material
FIGURE 6.24 Trapezoidal Profile into Leaf Sp
is later cold rolled, which much improves the surface properties. Heat-
The spring may be designed so that 5’ is twice treated steel may have a decarburized surface. Chrome-vanadium steel
width of one leaf, that is, so that there are two resists decarburization during heat treatment better than silicon-manganese.
length leaves with squared ends. Thus, suitable surface treatment will greatly improve the fatigue strength
of leaf springs. Inducing a residual compressive stress on that surface which
operates with a tensile Mc/I stress, either by presetting or peening or both,
increases the fatigue strength as usual. Materials used for leaf springs are
about the same as for hot-rolled coil springs, principally SAE 1080, 1095,
5155-60, 6150-60, 9250-60. For flat springs in general, bronze, beryllium
Copper, stainless steel, Inconel, clad stainless and carbon steels are also
used for a reason. Maximum design fatigue stresses for 107 cycles, AISI 1095,
flat and leaf springs with 5,4, = 0, as a function of thickness are:!2-1)
155 ksi for t = 0.005 in.; 140 for 0.010; 130 for 0.020; 125 for 0.030; 120
for 0.040; 105 for 0.060; 100 for 0.090 in.

FIGURE 6.25 Correction Factor for Defi

6.27 GENERAL REMARKS ON LEAF SPRINGS.
Factor K,

(From A. M. Wahl, Mechanical Springs(®-1)). Prestressing

leaves, or flat springs in general, (§ 4.23) in the same direction as the
loading leaves a favorable residual stress that increases the safe capacity
of the spring (see also §§ 4.26-4.30, inclusive, and other points in Chapter 4).
A number of other expedients have been applied to leaf springs.
A common practice is to bend the leaves to different radii of curvature,
the radius decreasing on the shorter leaves, Fig. 6.26; the leaves are said
b/b
 210 SPRINGS [Ch.6

to be nipped or to have nip. Considering the main and second leaves, note —
that when the leaves are pulled tightly together (by the center bolt), the —
main leaf is bent opposite to the direction that the working load will bend
it. Thus, it is not stressed in the direction of the working load (tension on
top, Fig. 6.26) until after the working load bends it past its unloaded ©
ip

a Main Leaf
FIGURE 6.26 Spring Leaves of Different Radii.
2nd Leaf

curvature; consequently, the maximum stress due to purely vertical forces

is greater in the second and other leaves than in the main leaf. Therefore
the main leaf is left with some capacity for taking nonvertical loading, as it
must in springs for vehicles. Nip, which may be applied to the other leaves,
also serves to produce forces between the leaves that tend to keep them in
contact on rebounds, thereby keeping out dirt.'°-2?!
Another idea for improving the load capacity of the main leaf is to”
make it thinner than the others. Using the elastic curve equation of a
straight beam M = E//r and the stress equation M = si/c, we get s = Ec/r, 7. COLUMNS, CENTRAL
where r is the radius to which a straight beam is bent by the moment M
that produces a stress s. We see that the stress in a beam bent to a certain” LOADS
radius r is directly proportional to the thickness of the leaf (2c). Hence, if
one leaf is thinner than the others, it will be stressed less than the others
by a particular moment. i 7.1 INTRODUCTION. A type of failure that we have not yet dis-
A highly destructive action on an automotive spring is the rebound, cussed is one due to an instability, called buckling, though an equation for
unless it is held in check, because an unchecked rebound after a bum checking the buckling of wide flanges was given in § 1.24. The buckling to
may bend the leaves until the stress is dangerously high. Thus, shock a’ be discussed in this chapter is of centrally-loaded slender members in com-
sorbers act not only to improve the ride but also to prevent spring failu pression, called columns; a yard stick is a good illustration. The longer and
by checking the rebound. slenderer the column is, the lower the safe stress that it can stand. The
slenderness of a column is measured by a slenderness ratio L/k, whereL in.
is the length of the column and k = (//A)*/? in. is the radius of gyration
6.28 CLOSURE. The scope of this book limits the coverage of sprin: of a cross-sectional area about a centroidal axis—nearly always the /east
to primary considerations of the most common types. In many ar radius of gyration. See Table AT 1, but observe that this table does not
such as the design of coil and flat springs for precision instruments, for give the least radius of gyration in every case.
weighing scales, etc., unusual problems call for specialized knowledge that
the spring manufacturer may be able to supply. If a spring is to be produced
in quantity, it should be tested experimentally to be sure it possesses the 7.2 EULER’S FORMULA. The Euler analysis applies to a very
required properties. ; Slender column, and the formula for frictionless rounded ends (no bending
For a spring subjected to fatigue loading, the most important factor is moment at the ends), Fig. 7.1(a), is
the surface condition. Where the cost justifies, vacuum melted steels have
fewer inclusions and they emerge from the manufacturing processes with 7w?EA
an improved surface. As it has been said,"*-5! “An accidental bruise on a (a) F, = Tp’
highly stressed spring will almost certainly lead to early failure.”
where F, is the axial concentric load, called the critical load, that causes the

211
212 COLUMNS, CENTRAL LOADS [Ch.7 57.5] STRAIGHT-LINE FORMULAS 213

column to be on the point of buckling, and E is the modulus of elasticity, 7.4 SHORT COLUMNS. If L./k is below a certain value for a
The units in (a) must be consistent, say pounds or kips and inches. Observe particular material, such as the values of L,/k at points A, B, D in Fig. 7.2,
that stress s is not involved in the failure of a very slender column. If we respectively for AISI 8742, 1137, 1015, the Euler formula for F, gives a
are to be certain that failure does not occur, the actual load F on a column stress above the yield strength; that is, below this certain value, the failure
must be less than F,; that is, the factor of safety or design factor N must may well be a failure of elastic action. Actually, it is probably some com-
now be applied to the load F and is thus defined by | bination of buckling and plastic action, and designers usually apply em-
pirical equations in these cases. A popular equation with machine de-
(b) N=—. signers is one proposed at the turn of the century by J. B. Johnson (F, =
failing load);

For a particular cross section and length, the load capacity F, of a column
depends only on the modulus of elasticity E. Since there is little variation (7.2) syAl i a or = Pi st S =.
4° E 4n°E
in E among different grades of steel, there is no advantage in using an [APPROPRIATE FOR 30 < L,/k < 120, STRUCTURAL STEEL]
expensive high-strength alloy steel instead of structural steel for columns
with L/k greater than about 120. See Fig. 7.2. where s, is the equivalent stress that indicates the degree of safety for the
load F; that is, N = F,/F = s,/s,; F/A = actual nominal stress. In design,
s, is a suitable design stress. For these shorter columns and steady central
7.3 EFFECTIVE LENGTH. Euler’s equation as written can loads, factors of safety between 2 and 3.5 are common. Equation (7.2) is not
applied to a column with ends fixed in any manner if the length is taken only called the Johnson formula but also the parabolic formula, because
that between sections of zero bending moment; call this length the effecti a plot of F/A against L,/k is a parabola, s = a — b(L,/k)?. Values of the
length L,. We shall write all our column formulas in terms of L,. The m constants a and 6 other than those in Johnson’s formula are sometimes
common types of columns are as shown in Fig. 7.1. The theoretical v: chosen for particular classes of columns so that the resulting equation
of L,, followed in parentheses by the AISC®:5*) recommended design values, reasonably fits available experimental data. The Johnson formula accords
are: for fixed ends L, = L/2 (0.65L); for one end fixed, one end rounded fairly well with considerable data for steel columns.
(or guided), L, = 0.707L (0.8L); for one end fixed, one end free, L, = 2L
(2.1L). Thus, with a factor of safety NV, Euler’s equation becomes

TEA wEA wEI

(7.1) fan Eee = WL? 7
NL?’
[USE FOR STRUCTURAL STEEL WHEN L,/k > about 120]

where F is some safe central load. A typical design factor for Euler s
tural columns is N ~ 3.5,"-?] and most designers probably tend to inc
N with significant increases in L,/k. It is important to understand that,
the column remains straight and if the load F is concurrent with the cen-
troidal axis of A, the average stress on a section of the column is s, = F/A
and that local buckling at some point where the stress is considerably lowe!
than F/A =s, is what will lead to failure. Ml
Theory indicates that an Euler column with fixed ends will support fow
times as much load (L, = L/2) as a rounded end column, but since the
FIGURE 7.1 Types of Columns. (a) Rounded ends; (b) fixed ends; (c) one end fixed, the
parts to which the ends are connected are not fixed rigidly and because of
other rounded; (d) one end fixed, the other free.
other departures from ideal, designers rarely if ever use L, = L/2; the
extreme limit is more likely about L, = L/1.41, and in machine design, t
choice is almost always L, = L (rounded ends)—except of course wh 7.5 STRAIGHT-LINE FORMULAS. Another common type of
the column has a free end, the weakest type (see reference 7.6). Use of column formula, widely used for shorter columns in structures, is the
L,. = L is conservative for pin-ended columns (see § 7.7). Straight-line formula, s = a — b(L/k), where g and b are constants. In
 7" COLUMNS, CENTRAL LOADS [Ch.7
some city building codes, Chicago and elsewhere, it becomes, for structural

Ss
steel,

Bs
F L L L, 1
(ce) ae 16,000 — Ce = 16,000( 1 - 0.0044), [30 < x 120]

where F is the safe or design load for the column. j

$8
Cast-iron columns are designed with a straight-line formula (NYC
Code), f

se
F

82s
(d) —= 9000 — aie = 9000( — 0. 0044),
A k k
where the symbols have the usual meanings. These formulas are not to be

o
_
applied to a free-end column. FE) rN
p Xoo ht J 1 1

80 90 100 110 120 130 140 150 160

7.6 TRANSITION POINT BETWEEN LONG AND INTERMEDIATE Llk
COLUMNS. The intersection point of two column formulas, as the
Euler and Johnson, is where they both give the same value of F/A for a FIGURE 7.2 Johnson and Euler Curves, Different Materials.
particular L,/k; and if they are tangent, they have a common tangent.
[a(F/A)/d(L./k)|_ = [a(F/A)/a(L./k)|;. The constants in the Johnson equa-
L,/k = 100 be considered. If the Johnson formula were used, the failing
tion are such that the curve is tangent to the Euler curve and always at
load at Q would be obtained, whereas the load it would take, at R, is much
F./A = s,/2. Thus, if F,/A from each formula is equated, we get an equa=
larger.
tion that gives the corresponding value of L,/k;
(3) There is not a large difference at L,/k = 100 between the strengths
Le ane 1/2
of low-carbon steel (at V) and alloy steel (at R), but at L./k values some-
(e) : what less, the difference is significant.
Sy
(4) In the vicinity of the tangent point 7, it makes little difference
(The student may check the slope for the Euler and Johnson equations at which equation is used.
this value of L,/k and show that they are the same.) The situation with (5) In general, use the Johnson formula when L,/k is less than that at the
respect to steels 8742, 1137, and 1015, and aluminum 7075-T6 is shown tangent point; use the Euler formula when L,/k is greater.
in Fig. 7.2.
In Fig. 7.2, consider for example the curves for 1015. If we had an id
column (approachable in the laboratory) with L./k < 86 (point D), si Y
wae Pin
L,/k = 70 at J, the load on it could be increased until F/A = s, at M2 FIGURE 7.3 Buckling of Columns. 7/4 PA rh,
then a further increase of load would result in an elastic failure. Howev: Assume the pins
B fixed. Let the member FO Vy * ojos
the actual column would be expected to fail more nearly in the vic turn on them without clearance, so that ; (a)
of the stress at K. Notice that if a straight line were drawn between NV (L, there will be no ‘‘wobble.”’ Thus, in (a), B L—————-++ , 8 ¥
the column bends as a rounded-end F
= 30) and Tp(L,./k = 120), it closely approximates the curve for 10
column and, in (b), it bends as a fixed-
(straight-line formulas). A similar statement does not hold so well for t end column.
®)
8742, a warning against blind generalizations about straight-line formulas.
Observations from Fig. 7.2 include: a
(1) At point NV, for L,/k = 30, there is little difference between the
load F = s,A and F, from Johnson’s formula; at the same L,/k at point P 7.7 RADIUS OF GYRATION. For rounded-end or fixed-end
the difference is more significant. Columns, the radius of gyration would be the one corresponding to the
(2) The tangent point T of the Euler and Johnson curves should not b¢ least moment of inertia. To understand the basis of determining the proper
taken for granted. For example, let a column of 8742, OQT 1200, with Tadius of gyration, consider Fig. 7.3 and the assumptions in the caption.

215
 216 COLUMNS, CENTRAL LOADS [Ch, s7.11] EQUIVALENT STRESS FOR COLUMNS 217°

The column with pin ends may buckle in one of two directions. It may ber 1.9 DESIGN OF COLUMNS. At the beginning of the design of a
as indicated by the dotted lines in the upper view; or, as shown in the lows column, it is not known whether the column is slender (Euler) or in the
view. If buckling in the plane in Fig. 7.3(a) is being considered, the radi intermediate range. Often too, the section is complicated, for example,
of gyration should be with respect to the axis XX; in the plane of Fig. 7.3( an H-section or a built-up column. Thus, we see that the design is fre-
it should be with respect to the axis YY. As you would guess by now, t quently one of trial and error. If the section is a simple one (circular,
fixed-end assumption would not be justified; but there is an addition: annular, square, or rectangular), k and A may be expressed in terms of a
small restraint with respect to YY. Without tests to substantiate a differenc single dimension, which can be solved for. The Johnson formula is more
design for the least k (maximum L,/k) and L, = L. frequently applicable to machines.
If the section is a standard rolled section, a handbook giving the
properties (A, J, and k,pjn) of rolled sections can be used to good advantage.
7.8 SECANT FORMULA. _ If one assumes that the load F has¢ A few iterations should locate a satisfactory section.
eccentricity e, Fig. 7.4, that the material is elastic, and that the deflection First, compute A = F/s,, where s,"= s,/N is the design stress for
small, the consequent theoretical equation (see nearly any text on mech: Johnson’s equation, because any section of a particular shape that is sub-
of materials) is the so-called secant formula. If the limiting (critical) jected to column action must have a larger area than this.
is that corresponding to s,, we have in
F, 200) eed By F, ec) Le [Fe a
f Ss = =(1 + sec = = (1 + — sec — =), 7.10 EXAMPLE. A piston rod is subjected to a maximum compressive load of
) tal he QIN BAP! Rok QNvey a 31,416 lb. and is 20in. long. The material is C1040, OQT 1000°F in order to
where, in the second form we have used J = Ak?. Since s is not di obtain a desired hardness. What would be the diameter for N = 3?
Solution. From Fig. AF 1, Appendix, we finds, =71 ksi. The corresponding
proportional to F in this case, the factor of safety must be applied o:
design stress is 71/3 = 23.7 ksi. In Table AT 1, we find & = D/4 for a solid round
the force, equation (b). The design equation may then be (F, =
section, which gives L/k = 20/(D/4) = 80/D. Since the numerical value of L/k
ec L, [NF is unknown, we must take a chance on which equation, (7.1) or (7.2), to use.
V3 Ss -—(1 + — see— —): Try (7.2); use force in kips and L, = L.
a KY, BE
where the symbols have the usual meanings; c is the distance from
ee s(L/k}?
be. os [! 4nr°E |:
centroidal axis to the external fiber; ec/k? is called the eccentricity ri
31.416 a aS (71) (80?)
ITAA 4n°(3 x 10*)D?]’

EE from which D = 1.44 in.; use 1% in. For this diameter, L/k = 80/1.4375 = 55.6,

rei
Eee FIGURE7.4 Effective Eccentricity, Exagg which is in the range for the Johnson formula [L/k ~ 83 from equation (e)].
Therefore, the result is satisfactory against buckling, and with N = 3 should be
’ conservative enough to care for size effects (0.505-in. specimen for Fig. AF 1).
It will be interesting and instructive to refer to the example of § 4.14, p. 117,
where we found the diameter of an alloy-steel rod for a reversed load of the same
The secant formula is applied to columns with central loads beca magnitude (31,416) as in this problem. The column strength of the alloy rod will
comparing computations from this equation with experimental dat be only a little larger than that of the C1040 rod in this example; but the fatigue
note that with some “effective eccentricity”’ the results agree fairly Strength of the C1040 rod is significantly lower than for the alloy. Therefore, if
The ASCE indicates that the eccentricity ratio ec/k? = 0.25 is goodf the load in this example is reversed to approximately the same magnitude,
structural columns with central load."-3! A value of e ~ L,/400 is sometin fatigue rather than column action will determine the proper size.
recommended."-1] But we notice that a suitable eccentricity ratio to
given environment is needed. Moreover, the secant formula is awk
to use for design because of the involved way the dimensions of the co: 7.11 EQUIVALENT STRESS FOR COLUMNS. _ Since the stress in
the equation. Charts can be constructed that aid in obtaining
an axially loaded column at buckling is less than s,, it is useful to have an
enter into -1.1.3.1. 7.7.11
solution. equivalent stress s, that indicates the degree of safety (when compared say
218 COLUMNS, CENTRAL LOADS [Ch.
97.13] CLOSURE 219
7.13 CLOSURE. Buckling of columns occurs at a time when they
to sy). Such a stress is in the Johnson formula; so solve for s, from equatic
are in unstable equilibrium. The buckling of a flange in a rolled-section
(7.2) and get 3
beam is of the same family of phenomena. Another collapse type of failure
F 1 F is that of a thin-shell vessel subjected to external pressure (§ 20.2). Residual
«) rit Alaa ALR A
= | = cs
stresses, such as those left in rolling structural shapes, play a part that has
not been fully investigated. In all cases, there is usually a strong element
4n2E
of empiricism in the design; unknowns are answered only by experiment.
where « is equal to the term in the brackets. To introduce an equiv: |
stress into Euler’s equation, let N = F,/F = s,/s, in (7.1) and solve for 5

“| SALA IK). F iq
(h) ———— i? }
als aE A a
where « is equal to the term in the brackets. In the straight-line fo mu
(c) for steel, consider the 16,000 = s,/N = s, and solve for s,:

‘a H 1 F
k %e AL 1 — 0.0044L]k i
= —| ————_ | = az,

where « is equal to the part in the brackets. Thus, we have a simplif

point of view; the equivalent column stress (but not the actual)t
Se = «F/A, where « is greater than unity and is given by the brack
terms of equations (g), (h), and (i) for the respective situations. If a co

712 OTHER COLUMN FORMULAS. There are several

column formulas designed to cover the above situations."’-5] The ta
modulus formula, which the reader probably met in the study of stre
of materials, agrees well with experience but is too awkward to us
design unless one is designing expensively and close to the limit. /
finds formulas to fit a particular shape or a particular material. For exan
an equation recommended for magnesium columns is!?-1) fi

i F,— = ——_—_____ _
c psi,
: A C(L.|ke
1 dace eters
64.4 x 108

where C is a number that depends on the compressive yield strength

A situation that often must be watched in structural columns co:

of thin sections, for example, a wide-flange rolled-beam section, ani
wall tubular columns, is local buckling of the thin metal, mentio
§ 1.24 in connection with beams. See Ref. (7.5).
 8. COMBINED STRESSES

8.1 INTRODUCTION. Until now, we have considered only those

cases where the stresses could be considered as simple stresses (F/A,
Mc/I, Tc/J). This chapter covers design for some combinations of these
simple stresses.
If a normal stress is tensile, we shall give it a positive sign; if compres-
sive, a negative sign, where convenient. The reader should not be confused
because the use of the negative sign is purely conventional. Thus, if we
speak of the minimum stress, the largest negative (compressive) stress, if
' any, is intended. On the other hand, if the determining stress in design is
compression, we may speak of the maximum compressive stress as the
maximum stress. The context will make the intention clear. There are so
Many different kinds of stresses to deal with that symbolization becomes
bothersome. Introducing two new stress symbols, we shall use the following:

5, a normal stress—tension or compression—computed from

F/A, «F/A, Me[I, etc.,
5s, a shear stress, computed from F/A, Tc/J, etc.,
51, S2, $3, OF Sz, Sy, Sz, etc., Various normal stresses,
Ssi, Ss2, OF Szy, Sz2, Syz, etc., Various shear stresses,
¢, a resultant normal stress due to a combination of the above stresses,
either tensile o, or compressive o., and
7, a resultant shearing stress.
 222 COMBINED STRESSES [
42
8
FIGURE 8.2 Prism with Eccentric Compressive Load. Es
8.2 BENDING AND UNIFORM STRESSES. One of the commones
and simplest combinations of stress is a bending stress, Mc//, and a uniforr
stress, F/A. For example, when a load is not concentric with a member’
centroidal axis, Figs. 8.1 and 8.2, the bodies are said to be eccentrical
loaded, the eccentric load inducing the foregoing stresses in the bod: o would be a suitably safe stress at a point of maximum stress. Since
Breaking the member at a section CT, Fig. 8.1(b), observe that a moment
minus signs are somewhat of a nuisance and since this case is so simple that
and an axial force F must be applied to maintain the severed part in eq there is never any question, designers frequently use equation (8.1A) for
brium. Now introduce equal and opposite forces F, = F on the centro compressive eccentric loads, as well as for tensile.
line in order to replace F by a force F, and a couple Fe. Then note that If a member in compression is long (L,/k > 40) and loaded eccentri-
moment M = Fe. The uniform tensile stress at CT is s, = F/A and th cally, the secant formula (7.3) is theoretically correct. However, since this
equation is difficult to design with, the literature provides a number of
alternate procedures, whose results at times vary widely. One method,
B believed to be reasonably conservative, is to use the equivalent column
of stress «F/A, computed from either equation (g) or (h), § 7.11 [and sometimes
— Byrft
a F, 4
from the straight-line formula (i)], instead of F/A;
o ry fF

ae
F Fee
(a) ate — &—
A
+ eos
6
6 Lr FIGURE 8.1 Prism with Eccentric Tensile Load.
\
where « is introduced to care for column action (§ 7.11) and, in design,
o is an appropriate design stress. When bending occurs in the direction of
ry F|
the maximum resistance, computing « for buckling in the direction of
(a) (b) minimum resistance should be on the safe side. The secant formula can be
used for checking the design done by equation (a), but if it is important to
obtain an optimum design, one must keep in mind that all theoretical
flexural stress is s, = Mc/I = Fec/I. The bending stress is tensile at T equations must be related to the facts of life by experiment, frequently
compressive at C, Fig. 8.1. Thus on the side 7, by superposition, st by experiment that simulates some particular actual usage.
adds numerically to s,; on the side C, sz subtracts from s,; or, withtl
usual convention of signs, the resultant stress is
8.3 EXAMPLE—DESIGN OF COLUMN WITH ECCENTRIC LOAD. A
F Fee
(8.1A) 10-ft. column is to be designed to support a load of F = 20 kips, overhung
+ Sp = — + i >
A I A I 15 in. at the free end, Fig. 8.3. It is planned to be made from standard pipe, the
material of which is similar to AISI C1015 as rolled. Use a design factor of N =
where A is the cross-sectional area, / is the area moment of inertia abou 2.6. Specify the pipe size.
axis BB, Fig. 8.1(a), M is the moment at the section that contains the Solution. If one were designing many pipe columns, there would no doubt
at which the stress is desired, c is the distance from the neutral plane to be short cuts at hand (as charts and tables) for at least an initial estimate of size.
point where the stress is desired (c = 5/2 for the maximum). At the If only a few are to be designed in the absence of short cuts, the quickest way is
of maximum stress in design, « would be some safe normal stress. Ust
Po ee
positive sign in (8.1) on the tensile side; on the compressive side C,
stress o may be either tension (F/A > |Mc/I|) or compression (|Me/I| FIGURE 8.3
F/A). q
If a bending moment exists at a section where there is a simple compres

S--
sive stress F/A, the normal stress on the same plane is L=10'

F Me By kee
(8.1B) ms —— + — = —- — + — >
Anis ab A I
where the bending moment may be produced by an eccentric comp
load, Fig. 8.2. If, in design, the member is short (no column actior
24 COMBINED STRESSES [Ch.
simply to assume a standard pipe size and compute the corresponding desig;
factor; if the computed JN is not in the proper relation to the specified N = 2.6
iterate until satisfaction is achieved. We shall use the equivalent stress methog FIGURE 8.5 T-Section C-Frame. #22
equation (a). From Table AT 8, we get sy = 45.5 ksi. From a handbook,
we find the properties of pipe sections. Assume an 8-in. pipe (nominal size
schedule 60*; then

The equivalent length of a free-end column is L. = 2L = 20ft. or 240

giving L./k = 240/2.909 = 82.5. From previous discussion, we judge that
value puts this column in the Johnson range, but checking from equation @
5s, = F/A, a tensile stress at the inside due to bending of sy = Mc/I = Fec,/I,
§ 7.6, to be sure, we get the dividing point as
and a compressive stress at the outside of s, = Mc/I = Fec,/I; where I is
L. [ a=)" Be (2) (9.86) (3. x | the moment of inertia of the sectional area about the area’s centroidal axis
45.5 = 114; BB, c, and c, are as shown in Fig. 8.5, for the maximum stresses due to M.
bOUS,
use Johnson’s equation. The resultant tensile stress is
therefore,
10. 48 ‘
*o=7 ewi y u fp w20/
a = 2.58 ksi. b) Oo, = 5, e + 52 eee
F
= — + —.
Fee,
( 7
(4) (#3) 3 x 104)
4n2E

c=, .
+z = 2.58 + (20)20.58(15) = 17.2 ksi
The resultant compressive stress (if |Mc/J| > F/A) is

F Fec
For this approach N = s,/¢ = 45.5/17.2 = 2.64. The next smaller size of (c) Oo,e = 511 — S$,ec = —A — — I
gives a design factor too low. The answer being so close to the desired
suggests that preliminary iterations had already been done. Thus, an ;
schedule 60, pipe is satisfactory. Further study might include checking N = This problem is not solely one of getting a safe section, but also of getting
by the secant equation and perhaps some experimental stress studies. good proportions in the section. It is frequently possible to save much
material, and therefore weight and money, by improving the proportions.
If the eccentricity e is large, the strength is more affected by the moment
of inertia than by the area. Consequently, less material is needed when the
material used is distributed far from the neutral axis. The further away
the bulk of the material is from the neutral axis, the greater is the moment of
inertia and therefore the resisting moment. However, care must be taken
FIGURE 8.4 C-Clamp. not to make the sections so thin that they will be in danger of buckling.
A T-section is particularly appropriate for cast iron. The iron may be
Concentrated on the tensile side (since cast iron is so much weaker in tension
than in compression), and may be spared on the compressive side. The
thicknesses of adjoining sections in castings should be not very different
from one another; otherwise, unequal rates of solidification will result in
Significant residual stresses. For steels, which have about the same strength
In tension and compression, a box-section, a modified I-section, or an
8.4 ECCENTRIC LOADING ON AN UNSYMMETRIC SECTI
H-section, is better.
The principle involved being the same as before, we can explain this
: In the design of an unsymmetric section, we might assume the propor-
by discussing the stresses in a section of a C-frame. For a T-section C- Uons of a section in terms of one dimension and obtain an equation with
Fig. 8.4, a load F as shown in Fig. 8.5 induces a uniform tensile stress
Only one unknown. However, this method, even for a simple T-section,
* The schedule symbol designates the weight of the pipe (wall thickness). The outst may become so involved that an error is more probable than not unless
diameter D, (= 8.625 in. in this case) remains the same for all schedules, given D. i It is highly organized (as it would be when such designs are often repeated).

225
226 COMBINED STRESSES [Ch. §8.6) COMBINED NORMAL AND SHEAR STRESSES 227

The following procedure with the calculations organized to facilitate checkin the forces on the outside rivets, F, and W/4, acting at right angles to each
is recommended. other, Fig. 8.6(b). The resultant is

6 n-ne (YY
(a) Make a sketch and place the assumed dimensions on it.
(b) Determine the location of the centroid.
(c) Determine the centroidal moment of inertia. [F1G. 8.6 ONLY]
(d) Determine the eccentricity of the load e.
(e) Solve for the uniform stress s;.
(f) Solve for the flexural stresses s, and s-. and s, = R/A, where A is the section area of a rivet. Observe that in this
(g) Determine the resultant stresses o, and o, and compare with the discussion the equations obtained are suitable only for the case analyzed.
stresses.
(h) Repeat the preceding calculations until a suitable section is found. —
The analyses discussed above apply when the member is not curved ¢ FIGURE 8.6 Shear in Two F AW:
the section for which the stress is desired. If the section is curved, as | Directions. If, in this statically
MM, Fig. 8.5, use the flexural stress as K,Mc/I, where K, is a cury. indeterminate connection, uni-
factor, § 8.25. } form distribution of W seems F,
too optimistic, probably true
el Ww
with bolts, one could assume Roast
8.5 COPLANAR SHEAR STRESSES IN MORE THAN ONE DIR the maximum portion on one a
TION. Shear stresses in different directions at a point on a p) fastener to be W/3 or even W/2.
(a) (b)
section of a body may be added vectorially. However, it is probablys:
to find the resultant force first and then the resultant shear. Cons If the deflections are coplanar (no twisting of the plate), this conventional
Fig. 8.6. In the conventional approach, the load W is imagined to h procedure results in conservative results. Other similar connections are
been replaced by a force W through C, the centroid of the area of the rit analyzed with the same assumptions as outlined.
and a couple of magnitude Wa. The force W through C is assumed
resisted equally by the rivets, each rivet being subjected to a shear
downward direction of W/N, = W/4, where N, = number of rivets,
8.6 COMBINED NORMAL AND SHEAR STRESSES. Another
this case. Next, it is assumed that the plate is rigid, that all the deforn
frequent combination of stresses is a normal stress, tension or compression,
is taken by the rivets, that the small turning of the plate by the mome:
and a shear stress, uniform or torsion. Consider the bolt M holding the
occurs about the centroid C, and therefore, with the further assum
that the size and material of the rivets are the same, the deforma’
stresses, and forces (F,, F,) are each proportional to the distances Lj, w' Stresses
—s 1 Combine
the rivets from C; Here
M
(@) oo F
Neo Y
y
so {"

FIGURE 8.7 Bracket.

The forces shown in Fig. 8.6 are those acting on the plate. Since the plat H
4 Bf
is in equilibrium, equate moments about C to zero; F. tt by

(e) Wa — 2F,L, — 2F2L, = 0. wo f

po 4

Substitute the value of Fy in (d) into (e) and solve for F,;
bracket of Fig. 8.7. In order to approach the design in a rational manner,
Wa ie WaL,
F, i DL re Qk Ly) -02(E, 2-Day).
Wwe must first decide upon an idealized model (more completely defined in
§ 8.11). Because of the tendency of the load W to tip the bracket about
from which F, may be found if the dimensions are known. We now nav Some point C and because of the initial tightening stress, the bolt M is
 98.8] MAXIMUM SHEAR STRESS 229

principal stresses. For the stress configuration described by Fig. 8.8, two of
the principal stresses are given by equation (i); the third principal stress is
zero. The positive sign in equation (i) gives the principal stress that is the
FIGURE 8.8 Force System on Small Prism under Tensile and Shear Stresses. Thi
represents a “‘point’’ in a stressed body. It applies no matter how s, and s, are induced, maximum (tensile) stress at the point, Fig. 8.8(b). If the radical is greater
by bending and torsion. The length of the prism normal to the page is unity. Figure (b) s than Sx/2, the negative sign obtained means that the corresponding stress
a diagonal section with forces that produce equilibrium. No stress in a direction pe is compressive, and in this case, it is the minimum principal stress. If the
dicular to the plane of the paper. negative sign is used and the result from (i) is positive, then the minimum
principal stress is zero, the third one. (More frequently than otherwise, the
subjected to a normal (tensile) stress. In addition, the section of the negative sign in (i) results in a negative stress.)
at the junction of the bracket and the wall is subjected to a shearing st If the sy stress in Fig. 8.8 were reversed to compression, the results
These stresses combine to produce a resultant tensile stress greater th would be analogous to those discussed except that tensile stresses would
tensile stress on plane YY and a resultant shear stress greater than the shi become compressive stresses. Therefore, we can generalize equation (i)
on plane YY. The following discussion reviews briefly the theory of # by dropping the subscript X,
combination.
c= > + [2 i) (ar
(8.2)
8.7 PRINCIPAL STRESSES. Use an elementary prism with stresse
s, and s,, say taken from the bolt M, Fig. 8.7, along the section YY bet
the wall and the bracket; width dx (along the length of bolt), height dy. and letting s be the normal stress on a plane where the shearing stress is s,;
s is computed from F/A, «F/A, Mc/I, etc., or a combination thereof;
depth unity. The forces (= sA) acting on the prism are as show
Fig. 8.8(a). First, set up a general expression for the forces on any diag s, from F/A, Te/J, VQ/Ib; (or other appropriate equations).
If two principal stresses are zero (simple tensile member, for example),
plane, BC. Remove the upper portion and consider the lower portion
free body in equilibrium, Fig. 8.8(b). Equate the sum of the forces normal the state of stress is said to be uniaxial. If one principal stress is zero (as in
plane BC to zero and solve for ods (= sA); the foregoing discussion), the state of stress is biaxial. If all principal stresses
have finite values (an element on the inside of a pressure vessel), the system
ods = sxydy sin® + s,dy cos @ + s,dx sin 0. is triaxial.
Divide through by ds, substitute dx/ds = cos 0, dy/ds = sin @, and fin
; : 1 — cos20 4
(g) o = Sx sin?@ + 2s, sin@ cos? = sx 7 + s, sin26. 8.8 MAXIMUM SHEAR STRESS. The maximum shearing stress
on some diagonal plane, Fig. 8.8, may be found in the same manner as the
To find the angle 6 when the normal stress o is a maximum or mini maximum tensile stress, that is, by summing forces parallel to the plane, etc.
differentiate (g) with respect to 6 and equate do/d@ to zero. This give: However, it is worth recalling from the theory of strength of materials that
the maximum shearing stress 7 is one-half the algebraic difference of the
2s, : maximum and minimum principal stresses. Using equation (8.2) for the
(h) tan 26 = — z
Sx maximum and minimum normal stresses, we get

where 26 is measured from the x direction, positive counterclockw: (8.3) Reina HGnax — Smin),
Since tan 20 is negative, 2@ is in the second or fourth quadrant and |
therefore in the first or second quadrant with the two values 90° ap (8.4) wal E fr yi
Using the two values of 26, tan-1(—2s,/sx) and tan1[2s,/(—sx)], im
and simplifying, we get
the maximum shearing stress at some particular point in a body for the state
(i a= Ba [24 Cel of stress defined in Fig. 8.8, except that the normal stress may be either
tensile or compressive. The maximum shearing stress occurs on a plane at
45° with the plane of the maximum normal stress. Notice that the maximum
the shear stresses are zero; the normal stresses on these planes are Principal stress (8.2) exceeds (8.4) by the amount s/2; that is, ¢ = 5/2 +7.

228
 FIGURE 8.9
Orientation s
of Planes —-\, --
of Max. - reer y onds
\
Shear \
—— SY —amy
my
8.9 ELEMENT SUBJECTED TO TWO NORMAL STRESSES ANE
SHEAR. The more general case of a plane stress system is an elem LY oitoe_sx+sy | 8 Say | sidy
2 Z 2
with two normal stresses sy and sy plus a shearing stress s,, Fig. 8.9 Omax. = 01 aydy
this vicinity, do not confuse the stress in the y direction sy with the
(a)
strength s,.) Making a free body of a triangular portion of the elem
(of unit depth), Fig. 8.9(b), summing forces and proceeding as expla
(c)
for Fig. 8.8, we find FIGURE 8.10 Mohr’s Circle. The principal stress c, is on a plane at —@ from the x axis,
in (b); c2 is on a plane at 90—8 with x axis, in (c). Note that if the directions of the normal
4) ee se]
+ A 2 1/2
(8.5) me a + and shear stresses on elements can be determined by inspection, the planes of maximum and
minimum stresses are determined by inspection; in (b), with the shear and normal forces
both with tensile components on the diagonal plane ds, the stress o, is naturally expected
ie 2 1/2 to be larger than oz in (c) where the shear forces, s,dy and s,@x, tend to result in compression
(8.6) a (5) +54] : on the diagonal plane ds. In (b), the plane PR is at 45° with the principal plane P7;
maximum shear (for oc, and oz) is on plane PR.

in which sy and sy are algebraic (that is, use a negative sign for com:
sion). The positive sign before the radical in (8.5) gives the maximum stresses toward the left. From B, erect a perpendicular and layout s, to
principal stress o,; the negative sign gives the minimum principal str scale, locating C; point C is a point on Mohtr’s circle whose center is at J,
oo if the result is negative, otherwise the minimum principal stress’ (sx + sy)/2 from O, and whose radius is JC. The sign convention for the
Smin = %3 = 0. It is important to note that there are three principal . shearing stress is: if the s,dy shear forces form a CL moment, this is positive
stresses that can be computed from equation (8.3) by using o, — shear with s, ='BC laid out upward from the end of sy. With sketches, Fig.
6, — og, and o, — ag (instead of o,,,, — opi), the maximum value 8.10(b) and (c), the planes of maximum o, and minimum oz can be
as defined by (8.3). Each principal shear stress is on a plane at 45° with | verified; the corresponding planes of maximum shear are at 45° with the
planes of the principal normal stresses from which it is computed. planes of o; and og. Draw the circle CHDG. Note that in the triangle
planes of the principal stresses are as defined by JCB, the hypotenuse is [s,? + (sx — sy)?/4]'/?, which, by comparison with
equation (8.6), is seen to represent the principal shearing stress 7. Thus,
2s
(8.7) tape le ! ove, Se 2 JE = 7,,,, and JF represents the shear on a plane at 90° with JE; naturally,
Sx — Sy |JE| = |JF|. Since JC = JH and OJ = (sx +sy)/2, the distance OH re-
where 20 is positive measured counterclockwise from the x axis. presents the maximum principal stress o,, equation (8.5); by a like logic
OG represents the principal stress o. Since og is positive (tensile), the mini-
mum stress is og = 0. Note that on the plane of the principal shear stress
8.10 MOHR’S CIRCLE. The stresses in .any direction can be co Tmax» the normal stress is OJ = (sx + sy)/2. The x axis, from which 20 is
puted from free-body diagrams of an element, after the manner of §8 measured is JC; positive angles measured counterclockwise. Ignoring
However, since Mohr’s circle provides a simpler approach to other Signs, we see that tan 20’ = 2s,/(sy — sy), which agrees with equation (8.7).
principal stresses, its properties are briefly reviewed here. Normal str If it were desired to know the stresses on a plane at the point being investi-
are plotted horizontally, shear vertically. Let the stress system be p gated at a counterclockwise angle of ¢ from the x axis, Fig. 8.10(c), lay out
and as shown in Fig. 8.10(b), in which it is presumed that the no: ae 2¢ from the x axis = JC, Fig. 8.10(a) and read og = OM and ry =
(sx, Sy both taken as tensile) and shearing (s,) stresses have been comp
from F/A, Mce/I, Tc/J, etc. in accordance with the external loading. Cho‘ The case of simple tension (cg = o3 = 0) is shown in Fig. 8.11(a).
a scale and lay out sx and sy from the origin O,in Fig. 8.10(a), locati It is seen directly from Mohr’s circle that Ssmax = 5,/2 on a plane at 45°
points B and 4; tensile stresses are laid out toward the right, compressi¥ (26 = 90°) from the x direction. Also from the free body below, the sum

231
 [are

i e
Lo Tmax. c x

AO ce ft e

yr-e
| cco

a, SE,
8:dA/sin 0 a
| Tsyl=|7y2l=18s|=|01|=| 021

dA'=dA/sin@ sdA/sin@ (a) (b) (c) (d)

(a) Simple Tension (b)Tension and Compression (c) Equal Tension and
FIGURE 8.12 Mohr’s Circle, Pure Torsion and Biaxial Equal Tension. To lay out
with Shear Compression Mohrs’ circle, compute the torsional stress and draw a circle of this radius (to some scale);
the horizontal axis through the center of the circle gives the intercepts of principal stresses
ox and oc, numerically equal to Ssmax = Txy. Notice in (e) that the shear stress on any plane
FIGURE 8.11 Mohr’s Circles for Various Stress Patterns. The elements dx dy hay perpendicular to the paper is zero when Sx = Sy, both tensile.
unit depth.

Rivets, if not otherwise specified, are likely to be equivalent to about AISI 1015.
of the forces parallel to the diagonal plane is s,dA/sin 0 — s,dA cos 0 Considering ASME Boiler Code specifications'****"*®! and the strength of 1015,
from which a design stress of ss = 8 ksi should be conservative. (For only two brackets to
be made, overconservatism is not bad.) For ease of installing, all rivets should be
(i) S, = s,cos@sin@ = = sin 26; Ssmax = 2” the same size. Since the upper rivets deform more than the lower one, they are
stressed more (s = Ee = E8/L); hence the design is based on the load F, in a
where we have recognized that the maximum value of the sine is unit single upper rivet. The force Fz is expressed in terms of F, by the following logic:
Figure 8.11(b) shows a Mohr’s circle for an element subjected to t (k) Fo _ S2A2 _ $2 _ Endaila _ 92 _ ba
sion sy, compression sy, and shear s,. In Fig. 8.11(c), Mohr’s circle fo: Fy SiA, Si E48,/Li 8: by
special biaxial case of |sx,| = |sy,| shows that 7,,,, has the same num or F2 = b2F;/b,. This result is a consequence of certain assumptions: A: = Az;
value; |o,| = |o2| = |tmax|- Compare this arrangement with the o E, = E, (rivets of same material); ZL: = Lz (rivets same length); 82/8: = b2/b:
pure torsion in Fig. 8.12 (a)-(d) and note the same numerical equality. from similar triangles, Fig. 8.13. If any of the assumed conditions happen not to
interesting case of two equal tensions sy and sy is shown in Fig. 8.! be true, one can always proceed through the logic of (k) as far as it applies.
where Mohr’s circle becomes a point and the shear stresses are ther
zero (cg = 0). The idea of Mohr’s circle is adaptable for situatior a Fo ee Es OR)
three finite principal stresses.2-7)
The forces with the bracket as the free body are as shown in Fig. 8.13(b). Sum
moments about C and use Fz = 0.273F,;

8.11 EXAMPLE—COMBINED TENSION AND SHEAR. A bracke fo (m) We = 2Fib:; + F.b2 = 2F,b, + 0.273F1b2

supporting a shaft bearing, similar to Fig. 8.13(a), is to be riveted to a ve (6) (12) = (2) (5.5)Fi + (0.273) (1.5)Fi
surface and support a load of W = 6 kips at a distance e = 12 in. from the from which F, = 6.31 kips. We may now solve the stress problem. When the
As seen, there are two rivets above and one below, at a vertical spacing of a =
The lower rivet is b2 = 1.5 in. from the bottom. Design for the maximum sh
FIGURE 8.13 Wall Bracket. In (a), the T-slot permits locating, say, a bearing at any
stress and compute a suitable diameter of the steel rivets. Only two brackets”
distance from the wall. The amount of tipping in (b) is exaggerated.
to be installed. a
Solution. This problem illustrates an application of analytic mechan
a statically indeterminate situation. (The reader should perceive the prin
involved so that he can analyze analogous situations; the equations obtained
to the configuration of Fig. 8.13.) In indeterminate problems, as in others
sets up an ideal model for which computations can be made. The principal assun
tions here are that the bracket B, Fig. 8.13(b), is rigid and that it tips sli
about the lowest point C when the load W is applied. These assumptions
in the deformations of the rivets, 5, and 82 being proportional to their dista
from the pivot point C. Other assumptions affecting the design are pointe
as we proceed with the solution.

232
234 COMBINED STRESSES [Ch.{ 8.12] THEORIES OF FAILURE 235

rivet heads are set, an initial stress is induced in the rivet whether the rive this stress is given by equation (8.5) for a biaxial stress system. For static
headed hot or cold; this stress is ignored or covered by a generous factor of sa loading of ductile material, a logical limiting stress is the yield stress s,;
in such simple designs. It is also true, as in bolted joints (§ 5.9), that if the ¢ that is, a design stress may be determined from og.= s,/N. Theoretically,
nected parts are relatively rigid, very little additional load comes on the ri
the limit stress is the elastic limit as determined by a simple tensile test
when the external load is applied unless the connected parts actually begin
(these theories are called theories of elastic failure), but yield stresses
separate. Moreover, there is no practical way of deciding upon the initial s
are more commonly available. For static loading of a brittle material
Hence, the design assumes that F, produces a tensile stress
(neither a distinct yield point nor elastic limit), as cast iron, the limiting
) elt ide. (6.31) (4) _ 8.04
(n are a 7D? Dp? ksi, stress is taken as the ultimate stress; og = s,/N. For fatigue loading of any
material, the limiting stress is logically the endurance strength (of course,
and that a uniform shear stress is produced by W of with proper allowance for factors affecting this strength—Chapter 4);
Li, | AOA) os 25. oq = S,/N. In this sense, equation (8.5) could be said to represent the normal
@) Vo ae sep! De stress theory of fatigue failure of a member subjected to combined stresses.
Experimental data suggest that the maximum principal stress theory
from a vertical sum of the forces, and it is assumed that the load is distri gives good predictions for brittle materials, and it is therefore often so
uniformly over the rivets. Substitute the stresses in (n) and (0) into equation ( used. It is not recommended for ductile materials. This theory is represented
and get

[e+ Gy)
ll
3

oy

“Ge Gad se
es (23 2 8.04\771/? 4.75. i GiA A By

OL-a:
FIGURE 8.14 Boundary for Principal Stress Theory, Static | Oz
from which D = 0.77 in.; use 3 in. The foregoing is a conventional approac)
Criteria. Biaxial stresses.
this type of problem. The assumptions made tend toward the safe side; criti eal Deg
design may call for more realistic considerations. The principal stress oz is
pressive; hence 7 is the maximum shear. | [+s
NoTE. The AISC!5-%41 recommends the use of the interaction principle
bolts and rivets, which for this case is defined by the equation
by a rectangle, Fig. 8.14, whose boundaries are defined by the tensile yield
5, and the yield point in compression s,, as shown. When s,, = sy, point O
(2)'
Sa
+(=Ssa ys is at the center of a square. The significance is that the boundary represents
a failing condition and that inside points, such as A where the principal
where sq and ss4 are respectively the design stresses for tension and for s
Stresses are o,, and og,, or B, represent a safe state of stress according
s, is the tensile stress computed from the tensile load only, and s; is the
stress from the shear component of the load. If the maximum shear stress the
to this theory.
is used (next article), the design sg = 2ss4 = 16 ksi in the foregoing examp (b) Maximum Shear Stress Theory. This theory is most frequently
5; = 8.04/D? ksi; s, = 2.54/D? ksi. Substituting these values in the above accredited to J. J. Guest, though others proposed it independently. For the
tion, find D ~ 0.77 in. as before. But compare the interaction equation cai elastic criterion, failure under combined stresses has occurred by this theory
and in detail with equation (8.10), § 8.13, below. when the maximum shear stress, equation (8.6), exceeds the maximum
Shear stress in a tensile specimen when the principal normal stress is the
elastic-limit stress. The value of this shear stress is s,/2, as shown in
8.12 THEORIES OF FAILURE. Over the years, there have Fig. 8.11(a) and equation (j), p. 232. The limiting static shear stress is
proposed a number of theories that are designed to predict when a fa 5, = 5y/2, where s, = yield strength; and the maximum shear stress,
of a metal member would occur. Those most frequently used are expl €quation (8.3), is given by (omax — Omin)/2; hence the maximum shear
below. Stress criterion can also be expressed by Omax — Smin = Sy (yield). Thus,
(a) Maximum Principal Stress Theory. This theory is accredited t the design stress in equation (8.6) for static loads is tg = (s,/2)/N in
W. J. M. Rankine (about 1850). In effect, it presumes that when ti accordance with this theory (except that yield has been substituted for
maximum principal stress exceeds a certain limiting value failure occurs elastic limit). However, many designers commonly use the: torsional
 236 COMBINED STRESSES [Ch.

yield strength as the limit stress and find a design stress from rg = s,,/M
In a particular application, the significant number is the design stress th
is expected to yield suitably safe dimensions as computed from a partic
theoretical equation.
FIGURE 8.16 Octahedral Plane. The octahedral plane is
The maximum shear stress theory is also used in design for fatigue, in
shaded. There is a normal stress 6, on this plane, not shown.
which case the analogous design stress would be rg = (5,/2)/N, where G
is commonly taken as the fatigue strength in a standard rotating be
specimen, with proper corrections for the factors that affect fatigue stre
An octahedral plane may be obtained by passing it through the mid-points
(Chapter 4); but see §§ 8.13 and 8.15 below. The boundary representatic
of three sides of the cube, as ACE. If an element ABCDEF is taken as a
for the maximum shear stress theory is shown in Fig. 8.15 for biaxial stre
free body and the forces summed parallel to the octahedral plane, in a
When both principal stresses are positive (or both negative—third quad
manner analogous to procedures previously explained, the shearing stress
with o, = 0, as at A, the principal shear stresses are (o, — o3)/2 = @,
To for a triaxial system is found to be
and (cz — o3)/2 = o,/2. That is, in the first and third quadrants, Fig. 8.1
the maximum shear stress provides the same limits as the maximum norm (p) Tor = $[(01 — 52)? + (6, — og)? + (6g — )?]!/?,
stress; say o, = 5, when o, > og. When 9, is positive, o2 negative,
where the normal stresses are algebraic. The normal stress o, on this plane
o3 = 0, the maximum shear is (o, — o)/2 and the boundary is given|
(not shown) is equal to the arithmetic average of the three principal stresses,
0; — G2 = Sy OF o, — oy = Sy, where the yield s, is a constant for a p
but we shall not have occasion to use it. Applying equation (p) to a uniaxial
cular material. As seen in Fig. 8.15, this is the equation of the bounda
stress state, og = og = 0, we find that 74, = (4/2/3)0,, which at the point
the fourth quadrant. Thus, a safe state of stress here in accordance with
of yielding (0, = sy) becomes 7, = (1/2/3)sy = 0.4715, (axial load) where
75, is the limit value (at the point of failure). The octahedral shear theory
oy may now be stated as follows: when the octahedral shear stress 7, in a body
Ty@Oxt sy [Yield] ‘i becomes equal to the octahedral shear stress 7, existing in a tensile test
specimen (uniaxial stress) at the instant that the tensile stress s, is equal to
L—O1A a &
the elastic limit (use s,), elastic failure is on the point of occurring. An
= | FIGURE 8.15 Boundary for Shear Stress Theory, Stati analogous statement may be made for the fatigue criterion. Equation (p)
Fen : Criteria. Drawn for |sycl = |sy|. Biaxial stresses. is reduced to that for biaxial stresses 7,, by letting o, = 0; then squaring as
indicated and collecting terms, we find the octahedral shear for a biaxial
Ma | 0y=03-8, system as

2
(q) Top = Poo? si O22 aa 61 02)1/? [PLANE STRESSES]

shear stress theory must be represented by a point, such as B, that is in Equations (p) and (q) may be used in design by letting the design value of 7,
of the limit line. The limiting stress for a reversed axial load would be be 7,4 = 0.471s,/N for static loading, but more convenient equations can
endurance strength, which can be imagined to be substituted for s be obtained by setting up an equivalent normal stress o, and deriving an
Fig. 8.15. The maximum shear stress theory has been the most widely €quation in terms of conveniently computed stresses, say sx, Sy, and Sy.
one for ductile materials. From the tensile test, +, = (/2/3)5, Or 5, = (3/+/2)7,; this value of s, is the
(c) Octahedral Shear Stress Theory. The theory of mechanics© normal stress for simple tension when the octahedral shear is 7,. Generaliz-
materials shows that the results from the octahedral shear stress theory am ing, call this s, the equivalent normal stress o,, and substitute the biaxial
those from the maximum distortion-energy theory are the same."-™ Hi value of 7, = 7), equation (q), into o, = (3/+/2)r, and get
the equations given below can be and are referred to by either name.
(r) Oe = (6,7 + Gg? — o162)*/?,
the name von Mises theory, sometimes associated with the names Huber ¢
as the Hencky—Mises criterion, are used, in honor of thi for which a static design stress could be o, = S,/N. It is well to note that o,
Hencky,"-7)
pioneers who developed the theory. does not designate a particular actual stress, but an equivalent stress that
Given a triaxial stress system as shown in Fig. 8.16 with o, > og > % indicates the degree of safety when compared to a failure criterion such as

237
238 COMBINED STRESSES [Ch.8

the yield s,. Substitute into (r) the two principal stresses as given by equatio 1
(8.5)—use the negative sign for o.—and find

(8.8) O- = (Sx? + Sy? — Sxsy + 35,2)1/?; FIGURE 8.17 Boundary for Octahedral Shear Stress
Theory. Maximum shear and maximum strain boundaries
and for the most frequently encountered state of stress (sy = 0) of one are shown for comparison.
normal stress and a shear stress, Fig. 8.8,

(8.9) o, = (s? + 3s,7)'/?, Ox? +0y°—Ox Cy m8?

where 5, Sx, Sy, 5, are computed from F/A, Mc/I, Te/J, etc. When o, =
the octahedral shear stress theory is predicting incipient elastic fail
exceed e,. The strain is converted to stress by s = Ee; and the resulting
(It will help the reader to gain familiarity with this material if he will de
(8.8) in detail.) : equation for two normal stresses sx, sy, and a shearing stress s, is
The relation between the yield (and elastic) strengths in tension s, Wee
KG ts) +0 +H|(F>*) +52],
a 2 1/2

and in torsion s,, is obtained by assuming first a state of pure torsi (t) o=
o = 0, = —o2 = S,, Fig. 8.12, in which case, equation (q) reduces
(a3 = 0) where p» is Poisson’s ratio and, in design, ¢ = s,/N = a, for static loads.
This theory has been widely used for thick cylinders; it fits experimental
2 2 : 2
Top = 3 = 3°° To = 3 ou
data on brittle materials better than those on ductile materials. The boun-
daries are shown by the dot-dash line in Fig. 8.17. Notice that if the
where 7,, is the limit value when elastic failure is about to occur. The other theories are right in the vicinity of o, = a, this theory is relatively
value of the octahedral shear in terms of the tensile yield was ob’
dangerous, with point D so far out.
above for uniaxial stress at 7, = (1/22/3)sy. Equating these limit val
Top = Toy» and solving for s,,, we get
8.13 DESIGN EQUATION FOR MAXIMUM SHEAR AND OCTA-
Ss HEDRAL SHEAR THEORIES. Consider a biaxial stress state induced
(s) fy, = Vn = 0.577s,, by a normal stress s and a shear stress s, on a particular plane; the maximum
shear stress is given by equation (8.4). Elastic failure by the maximum shear
stress theory is incipient when the resultant shear + = s,/2, as explained
as compared with s,, = 0.5s, predicted by the maximum shear stress the
above. A safe shear is t = s,/(2N). Substitute sr = s,/(2N) in (8.4), divide
As previously stated, actual test values range widely from below 0.5s.
through by s,/2, and get
greater than 0.6s,; we have been arbitrarily using s,, = 0.6s,. The fi
that the ratio 0.577 is in better agreement with experiment than 0.5. Fe
(u) ~ = (=) : (“ u (2) i ( Si i
this reason and because test points tend to match the boundary defined t
the octahedral shear stress theory, there is an increasing tendency t y Sys
using this theory. A graphical portrayal of this boundary for biaxial stre where s,, is taken equal to s,/2 in accordance with the maximum shear stress
is obtained by letting o, = s, in equation (r) and squaring both sides. theory.
the yield s, is a constant, the resulting equation is that of an ellipse For the same state of stress, the equivalent stress o, for the octahedral
seen in Fig. 8.17 (where o, = o,, o2 = o,). The octahedral bound, shear theory is given by equation (8.9); a safe value is o, = s,/N. Substi-
everywhere greater than the maximum shear boundary except where OF tuting this value into (8.9), dividing through by s,, we get
principal stress is zero (also og = o, = 0).
(d) Maximum Strain Theory. This theory, due to St. Venant, is
only other one that we shall mention. At the elastic limit in uniaxial tens
the unit strain is «,; according to the maximum strain theory, ine
ihe rene Meg (ro con Sys

action begins at a point in a body when the strain at that point begins where we have used s, = +/3s,, from equation (s), in accordance with the
octahedral shear theory. The final form of this equation is seen to be the

239
 240 COMBINED STRESSES [Ch,

same as the previous one; for this reason, it will be suitable for design pur.
poses for either theory; to wit,
Iz" f lee
20"
1 s 2 Ss 2751/2

(8.10) == (-) + (=
FIGURE 8.19 Force F’ = F produces bending moment = 4
N Sy Sys a7 PB.
of 12 F; torque couple FF is 20F.
If s,/2 is substituted for s,,, the design would be by the maximum shea
stress theory; if sy, = sy/+/3 = 0.577s, is substituted for s,,, the de:
would be by the octahedral shear theory. If s,, = 0.6s,, a commo
assumed value, the design is not in accordance with either theory; in
case, the equation is an expression of the interaction principle (see note 4
end of § 8.11). a factors are applied to the alternating component s,. Cazaud':24! has found
In equation (8.10), let NV = 1, let x = s/s,, and y = 5,/s,, and squat that if the shear (torsion) and normal stresses are reversed, periodic, and
both sides; the result is seen to be the equation of a circle, Fig. 8.18, in phase, this equation agrees well with failure tests (V = 1); naturally,
corrections are necessary for stress raisers that may exist.

8.14 EXAMPLE—COMBINED BENDING, COMPRESSION, AND TOR-

tl

SION. The jack of Fig. 8.24(a), with a 2-in. Acme thread (Fig. 8.21), is rated at an
0.54 8000-Ib. load W. Assume that the load can move laterally with the top of the
screw, so that a force of F = 60 lb., applied at a torque arm of 20 in., produces
a bending moment on the screw as a cantilever beam, as well as the torque to raise
the load. Also assume that the load is not central, but has an effective eccentricity
ee FIGURE 8.18 Points shown are typical. ofe = tin. A diagrammatic representation of the force system is given in Fig. 8.19.
ie
If the screw is made of C1045 as rolled, is this loading safe? Check by both the
maximum shear stress and the octahedral shear stress theory. The maximum lift
of the jack is 12 in. (length of the cantilever beam), small enough that « may be
circle that is the predicted location of combinations of stress that prod taken as 1 (no buckling).
elastic failure. (Note that with s and s, the variables and with s, and Solution. From Table 8.1, we find the root diameter to be 1.75 in., for a
constants, (8.10) is the equation of an ellipse.) There is considera 2-in. screw. Neglecting the strengthening effect of the threads (§ 8.24), we have
evidence to support the validity of this conclusion, as suggested by _ (1.75)? 7D? _= ———
(1.75)?
typical failure points shown. For a design factor of NV, the radius 0 A; 4 = 2.4 in.?; Z=— = 0.525 in.?;
circle of safe stresses is 1/N in Fig. 8.18.
If, at the beginning of this article, we had postulated fatigue fi Z = 2Z = 1.05 in.’.
and used s,, = s,/2 for the maximum shear theory and s,, = S,/ The screw is subjected to the following stresses: a uniform compressive stress, a
0.577s, for the octahedral shear theory,* the resulting equation w bending stress due to the force on the handle, a bending stress due to the eccen-
become tricity of the load, and a shearing stress. Compute the normal stresses first.
5 Ww

aD v1 ch Genk
1 S 2 s 251/2 Uniform stress a a = 3330 psi, on all sections.

N Sn Sns

wherein, for combined stresses, one uses the equivalent stresses base
Bending g byby F S2 =Z Geo
0.525 1370 psi, at base (B, C) only.

S, ANd Sas; Se = SmSn/Sy + KySq (§ 4.19); surface, size, and other fa

‘ M 8000)(0.
Bending by W 53 = ae weorne2) = 3810 psi, on all sections, front and
* There are some test values showing lower ratios for (torsion fatigue)/(bending fa 0.525 rear fibers.
perhaps lower than 0.52, but other test values reach 0.67 or more for steel; above 0. At some Stage of rotation of the screw and handle, all of the normal stresses act
some cast irons; about 0.63 for aluminum 7075-T6.!*:54] But such data are not t€
together in the same sense at a particular point; for position shown in Fig. 8.19,
plentiful.

241
242 COMBINED STRESSES [Ch.8 §8.15] COMBINING VARIABLE STRESSES 243

they are all compressive at point C. If the yield strength in compression is the” If more than one of these stresses is present, the mean stress Sj,
same as the tensile yield (usual assumption for steel), the check is based on the is determined from the circumstances involved. For example, if
total stress at C, which is a fixed section is subjected to a steady uniform stress F/A and to
So = S1 + $2 + Ss = 3330 + 1370 + 3810 = 8510 psi. a variable bending moment, then s,, = F/A + M,c/I (an
If the pivot friction is neglected, the torsional stress is algebraic sum). If M is constant on a rotating shaft and if there is
a steady F/A, then s,, = F/A because the mean bending stress is
_ T _ (6020) _ zero.
Sq is obtained from F,/A, «F,/A, or M,c/I, etc. If more than one is
which is observed to be relatively small. From Table AT 7, we find sy = 59 si present, the value of s, is determined from the circumstances
Use sys = (0.5)(59) = 29.5 ksi in equation (8.10) for N based on maximum she
involved. If the variations of F, and M, are “in phase,”’ the two
and use sy; = (0.577)(59) = 34 ksi for the octahedral theory.
corresponding stresses are added. On a rotating shaft with

Ba Nak constant M, the variable bending stress is s, = Me/I.

Sms is obtained from F,,/A or 7,,c/J. See comments under s,, above.
8.51\? 1.14)? Sas is obtained from F,/A or T,c/J. See comments under s, above.
i. ( 59 ) ir (a5) K;, Ky, are strength-reduction factors owing to stress concentrations.
In the design procedure, the values from the foregoing equations (w) and
from which, N = 6.7 for maximum shear; and

om e-@T
(x) for equivalent stresses are substituted into (8.11)*, which is

may Co). + Ga)

1 8.51\? 1.14?
2 2571/2
from which, N = 6.75 (octahedral). If the shear had been relatively grea N n Sns
greater difference between these answers would have been obtained. Also,
that if s,, had been taken as 0.6s,, as you have been doing previously in thi
good for the case of one varying normal stress and a varying shear stress
the result would have changed little. The answers obtained suggest that the ja on the same plane, in phase, Fig. 8.8. Assume that s,,/s, = sy5/S,; if we
conservatively rated. It should be noted, however, that a jack may get some r take sy, = S,/2, S,, = 0.5s,, we can consider that we are using a
handling, that the effective eccentricity of the load may be more than } in., maximum-shear-stress theory of fatigue failure, § 8.12; using s,, = 5,/+/3
that a workman may use a longer handle and pull harder. = 0.577s, would correspond to the octahedral shear-stress theory; if s,, =
0.6s,, as frequently assumed, there is no accord with any theory, but the
results agree well with tests. The equation resulting from the manipulation
of (w), (x), and (8.11) is
8.15 COMBINING VARIABLE STRESSES. Many machine
ments are subjected to a combination of stresses where the axial
varies, the torque varies, or the bending moment varies, or where any (y) ate (= ene ) + (= ~ siren,
N Sy Sn Sys Sns
or all vary. Several procedures for combining different kinds of var [VARYING NORMAL AND SHEAR STRESSES ON A PLANE AT A POINT]
stresses have been proposed.'-!2) The one adopted below is logical, tho
but the student is urged to solve problems by following a detailed approach
approximate (as are all theories), and it is expected to produce safe de:
for ductile metals. Use the equivalent stresses obtained by multiplh to equation (8.11) [or (8.10)], as in the example below, because the logic
is lost in substituting numbers into (y). For brittle materials, the maximum
equations (4.4), 1/N = S,/sy + K;Sq/s,, and (4.5) through by s, and
normal-stress theory is generally better, and it too can be adapted to the
respectively. Dropping s,/N and s,,/N in favor of s, and s,;, we then
€quivalent-stress approach. Variations of stresses that are not in phase can
Sn be handled with some logic, but this class of problem is beyond the scope
(w) Se = —Sm + Kya of this book except as we may sometimes assume that peak stresses occur
Sy Simultaneously, an approach on the safe side. The more general case of
three finite principal stresses with the same frequency can be set up in
Sns
(x) Ses = Sms + Ky sSas terms of principal stresses and the octahedral shear theory™-14! but justify-
ys ing experimental data are not plentiful.
where 5,,5/Sys © Sp/Sy, “ The use of the equivalent stresses from equations (4.6) and (r), § 4.19, in equation
Sm is a nominal stress obtained from Fp/A, %Fm/A, or Mye/l, et (8.10) gives the same result. See the discussion on theories below equation (8.10).
 244 COMBINED STRESSES [Ch.8 §8.17] ADDITIONAL COMMENTS ON FATIGUE 245

8.16 EXAMPLE—COMBINED BENDING AND TORSION, VARYIN has changed from 7850 to 785 in-lb. (sum of the moments of the forces to the
STRESSES. Figure 8.20 represents a shaft with loads A and C being fore right of section C). At each position, the tensile stress on the top fiber of the shaft
exerted on the shaft by gears keyed on with profile keyways, and B and D bei is the same in magnitude as the compressive stress on the bottom fiber, but if we
the bearing reactions. For a particular power transmission, the maximum follow the particular fiber at point C as the shaft rotates, we see that the stress
moment occurs at C, which is therefore the section to be investigated since the variation at that point fixed in the shaft is from
torque is also transmitted through it. The load on the shaft is variable because § ny M __ 7850 te iho gi 785 £
the transmitted horsepower varies continuously from 60 hp to 6 hp in half a Fs ec an, ‘min = Z psi,
revolution, and from 6 to 60 in the next half, while the shaft rotates virtually at a [AT PosiTION 0] [180° LaTER]

=— 10” where the minimum stress is compression. The mean (smax + Smin)/2 and variable
Tension a
(Smax — Smin)/2 components of the bending stress are therefore
hy

FIGURE 8.20
P]\comp. 0533 be he 3.53 ksi aii ete 4318_.. 4.32
Sm = —Z-P z a=—Z psi or —- ksi.
Max.M-t 78.5 Ib.
Here 785 lb. (It is important to note that when the bending moment on a rotating shaft varies,
one is safer in working directly with stresses than with mean and variable bending
moments.) Using the foregoing findings in equation (w), we get
constant speed of 400 rpm. The force at D varies from 785 to 78.5 lb., changin
with the horsepower. For a material of cold-rolled AISI 1137 and a design fa Sn, Kjsa _ 3.53 Pi (2)(4.32) _ 12.15
(a) Se =
of N = 1.7, what should be the diameter of the shaft? Sy ASH VI8Z eee
Solution. From an outside source, for 1137, cold drawn, we find Substitute the values of s. and s-; from (a) and (z) into equation (8.11), with
Sns = 0.577s, for the octahedral shear theory, and get
Sn = 50 ksi, sy = 89 ksi, dP
From Table AT 13, we find the stress concentration factors for a profile keywa y w 7 (Ee Goh
K; = 2.0 and K;; ='1.6.

The torque on the shaft, from T = 63,000hp/n, is

+5 it (ey i (esmenc) | = 147.5
50Z
+ 99.517 = aed
63,000)(60 : 63,000)(6 from which Z = 0.524 = 7D°/32, or D = 1.748; use D = l}in. The next
Tmax = Set = 9450 in-lb., and Tmin = eo. = 945 in-lb.| chapter contains additional discussion of shaft design.

From these values, the average torque and the variable component are, respec
tively,
8.17 ADDITIONAL COMMENTS ON FATIGUE. Fatigue failure is
Tm = 5200 in-lb. and Tz = 4250 in-lb. the initiation of a crack and its propagation. Thus, the logical expectation
sna TM = Teaura
= 5200 | 5200
2Z
ang 5, = Ze =| OD
4250 _ $250,
2
is that the crack most likely follows a plane of maximum tensile stress.
For a member in pure torsion, fatigue cracks typically progress on a plane
Or sms = 2.6/Zand
sas = 2.125/Z ksi, where J/c = Z’ = 2Z, becauseZ’ = 7D*/I at 45° with the plane of the maximum shear stress, the plane of the principal
and Z = 7D*/32, Table AT 1. On the expectation that the size will be great tensile stresses (Fig. 8.12). For a member in compression only, cracks will
than 0.5 in., introduce a size factor (SF), § 4.12, applied to the alternating co! Not spread on the plane of the compressive stress, but typically occur in the
ponent only; assume that the strength reduction factors K;, K;, take care of tht Vicinity of the plane of the theoretical maximum shear, at 45° with the
surface condition, use equation (x) with sn;/sy; = Sn/sy = 1/1.78, and get i Compressive stress. For a member subjected to combined torsion and
bending, it has been found for mild steel'®-19! that: if ¢ax/Tmax is distinctly
ee Kiss, 2:6.) Geyetiasy S46. above 1.6, the crack propagates in thenormal stress direction; if Curee| Tex
(2) Ses iT.
pres Sky > Pree oar Te
1.6, the crack follows the shear direction; for omax/Tmax * 1.6, it might go
The bending stress would vary even if the bending moment were constan! either way. To further complicate the logical processes, Sines'*:1) presents
because the fiber C, Fig. 8.20, which first has a tensile stress on it, has a compr
sive stress on it 180° later when it has moved to P. Let’s say that hp = 60 when
data that indicate that the alternating shear stress produces the fatigue
fiber under consideration is at C; then it is 6 hp when the fiber is at P. In damage. He also shows for pure shear that the magnitude of the mean shear
meantime, the force at D has changed from 785 to 78.5 lb., and the moment at: Ssm has no effect on the magnitude of the alternating shear that produces
 BSREPRIEY Veoh SERNA
VIR SANA WR ANTE SZ RAINES

See Refs. (8.8) and (8.9). for more detail on Acme threads. The stub Acme thread has a
246 COMBINED STRESSES [Ch. 8 height of 0.3P instead of 0.5P, Fig. 8.21. The minor diameters are given to the nearest
thousandth of an inch. The standard(®-?4) does not specify a unique number of threads per
failure, provided the maximum stress 7,,,, does not exceed the torsion al inch for buttress threads. See Fig. 8.21.
yield strength. (But see Chodorowski'*:?® for contrary data.) %
Residual stresses (from the various manufacturing processes—ther or eee lobe ae)
and mechanical) apparently play a more important role in determin : A
fatigue strength than generally conceded. These stresses are trie me Th. ieee Th. ve ao Th.
complicated, and difficult to obtain or to estimate, but the conclusion : in. ; in. Dia. Dia. in.
that designers should work toward ways and means of including their eff
and of controlling them. See Mattson.":28! t 10 0.163 16 0.188 0.213
ts 14 0.241 0.270
$ 8 0.266 12 0.292 0.325
8.18 POWER SCREWS. As suggested by the example of 8. % 12 0.354 0.388
power screws are often subjected to combinations of stresses; hence,
: i cory : S R : 366 i i 2
shall include some engineering information about them at this point. S ; a bee ” ees Aes ao
they are designed to exert a force with a mechanical advantage, the thr 3 5 0.575 6 0.583 0.650 16
are somewhat different from fastening threads. The common forms are $ 44 0.681 6 0.708 0.775 16

1 4 0.781 5 0.800 0.880 12

1} 5 0.925 1.005 12
1} 34 1.000 5 1.050 1.130 10
13 4 1,125 1.225 10

1} 3 1.208 4 1.250 1.350 8

Thread se alates Buttress! Theead 12 24 1.400 4 1.500 1.600 7
oa Dh aee io) tee 2 24 1.612 4 1.750 1.850 6
FIGURE 8.21 Power Screw Thread Profiles. (a) Commonly used proportions. In at at. $08 3 hail ae '
jn ae a . peg the thread angle, and ¢ is called the pressure angle. (c) De} 24 2 2.063 3 2.167 2.300 5
ier er 2% 2 2.313 3 2.417 2.550 5
. 3 13 2.500 2 2.500 2.700 - 5
shown in Fig. 8.21, and other information is given in Table 8.1. The A 34 1 ase + se oe i
thread, due to the sloping sides, is theoretically not quite as efficie 4 1} 3.418 2 3.500 3.700 4
the square thread, but it has been shown in practice that the accuracy 44 2 4.000 4.200 4
manufacture and the condition of the thread surfaces are the import 5 2 4.500 4.700 4
determinants of efficiency. Since the Acme thread may be cut with dies, 1
easier and cheaper to manufacture. Furthermore, if a split nut is us distance the nut moves along the axis in one turn, Fig. 8.22. A single-
looseness due to wear may be eliminated by “taking up’”’ the nut. threaded screw has a lead equal to the pitch, Fig. 8.22(a). A double-threaded
buttress thread has virtually the same efficiency as the square thread, screw has two starts, and the lead is twice the pitch. A triple-threaded screw
can transmit power in one direction only.
FIGURE 8.22 Pitch and Lead. The lead angle is A.
8.19 PITCH AND LEAD. The axial pitch or pitch P is the dist
measured axially, from a point on one thread to the corresponding pou
on an adjacent thread, Fig. 8.22
1 :
P, = P= —————_—______— in.
No. of threads per inch
The Jead is the distance that a thread advances in one turn; it is”
 Lead

Ey Mr
Load W: Q

(a) Top View of Thread. (b) Center Line of Thread Un-

wrapped from Cylinder.
(a) Single Thread (b) Double Thread
FIGURE 8.23 Forces on a Body Moving Up an Incline. The force Q is horizontal and
normal to the axis of the screw.
FIGURE 8.22 (Repeated).

screw against turning under a load W. The left-hand side T is the torque
applied when the screw is on the point of turning (no other resistance), and
Q represents the force that must be exerted at the midpoint of the thread
[see Fig. 8.23(b)], to obtain this torque. However, the external force is applied at the end of a
lever, Fig. 8.24(a), or at some equivalent leverage, as in Fig. 8.24(b). Thus
Lead T = QD,,/2 = Fa, where Fis the applied force at some leverage a. Equation
(b) A.= tan~?
TD, (f) is sometimes easier to use if the expression tan(f' + A) is expanded;
WD,,(tanA + tan B) WD,,(tanaA + f)
(g)
- 211—tanBtand) 21 —ftand) ©
The force analysis of an Acme thread is the same as that for a worm. If
both sides of equation (k), § 16.8 are multiplied by D,,/2 (and let F, = W),
we can find the torque to turn the thread against the load; in the symbols
arrangement for a force analysis is in Fig. 8.23(b), which shows the mean of this application, it is
line of a thread developed into a plane. (h) brs | cos¢ tanA +7),
2 Lcos¢ —ftanaA
where ¢ & 14.5° for a standard Acme thread, Fig. 8.21(b). Strictly,
¢ should be the pressure angle in a plane normal to the thread, rather than
moving), and the normal plane reaction NV; F; and N are replaced by their the pressure angle in a diametral plane as shown, the relation being
resultant R, the total plane reaction. The angle of friction is 8 (where tan ¢ = tan 14.5° cos A. Since A is usually small for power screws, cos A is
tanB =f, and f is the coefficient of friction); and A is the lead angle.
Sum forces horizontally and vertically: i FIGURE 8.24 Screw Jacks. In (b), the pivot friction would be small because of the ball
(c) Q = Rsin(B
+ A). bearing. The long nut suggests low wear on threads and stability at maximum extension.
Notice the buttress thread and the bevel-gear drive.(Fig. 8.24 (b) courtesy The Duff-Norton
Mfg. Co., Pittsburgh).
(d) W = Rceos(B + d).
Dividing Q in (c) and Win (d) and solving for Q, we find
(e) Q= W tan(B +A).
Multiply both sides of (e) by D,,/2, where D,, is the mean or pitch diameter —
of the screw, which gives QD,,/2 = T, or q
QD,, WD», -

(f) ; T i Ue id Bie 2 9
tan(B + A). [SQUARE THREAD]

The right-hand side represents the resisting moment in the threads of the
 58.23] CONDITIONS FOR SELF-LOCKING SCREW 251

FIGURE 8.25 Ball-Bearing Screw. or friction. If the friction is zero, f and 8 are each equal to zero, and equation
approximate rolling contact between the (f) or (g) becomes
and the grooves in screw and nut. This a
made possible by providing the tubes ft WD
which the balls can circulate. ( i) : 7 = tand,
General Motors Corp., Saginaw, Mich.
where J’ represents the torque (twisting effort needed) to move the load
without friction. The efficiency of the screw, which is the effort without
friction divided by the effort necessary to turn the screw with friction, is
T’ in the above equation, divided by T in (g):
usually close to unity and the error involved is of a smaller order t
T’ tandA(1 — ftana)
in f. Equation (h) applies also to fastening threads where ¢ = 30°.
If there is friction at a collar or surface other than the thre:
(i 6 srg Rel oa
tand+f
equation for pivot friction may be used, an expression for which is If the efficiency by this equation is plotted against A with f constant, a
in § 18.10, equation (18.2). The total torque to be overcome would curve very similar to that of Fig. 16.6 as given for a worm will be found.
sum of that at the threads and at the pivot. A ball bearing at the Thus an increased efficiency may be obtained by increasing the lead. This
Fig. 8.24(b), materially reduces the friction. Also where there is a wo variation of efficiency with the lead angle should be kept in mind, and the
advantage, as in automotive steering mechanisms, ball bearings can be u knowledge applied where practicable. However, increasing the lead de-
between the nut and screw, Fig. 8.25. creases the mechanical advantage, and this may be undesirable. Further-
Recall from mechanics":®! that the work done by a total torque more, a lead angle will be reached where the screw is no longer self-locking,
JT, 40 or T,8, when the torque can be considered as constant; 6 rad. a property happily utilized in most screw applications.
angle through which the screw turns during the application of T,
is [T,dw, or T, for constant torque; hp = T,«/33,000 for 7; ft
w rad./min. Power is consumed at a constant rate if T and w are 8.23 CONDITIONS FOR SELF-LOCKING SCREW. A self-locking
otherwise it is an instantaneous value. screw is one that requires a positive torque to lower the load, or to loosen
the screw if it has been turned tight against a resistance. Figure 8.26 shows
the force arrangement on a square thread if the load W is to be lowered
8.21 COEFFICIENT OF FRICTION OF POWER SCREWS. (moved down the incline of the thread). Proceeding as we did to get the
thread surfaces are smooth and well lubricated, the coefficient of torque to move the load up the plane, and summing the forces vertically and
may be as low as 0.10, but for materials and workmanship of a horizontally, we find
quality, Ham and Ryan!®-7) recommend f = 0.125. For doubtful»
manship, f may be assumed as 0.15. Increase these values by 30-35
(From XV) W = Roos(B — Ad),
starting friction. (From XA) Q = Rsin(B — A).
On the basis of their experiments, Ham and Ryan concluded Dividing the value of Q by W, solving for Q, and then multiplying each
coefficient of friction is practically independent of the axial load; tha Side by D,,/2, we get
undergoes negligible changes due to speed for most speed ranges encounte
in practice; that it decreases somewhat with heavier lubricants; OD, cg WD,, tan(B — A)
(k)
shows little variation for different combinations of commercial mat Fics GARR EMME Rete
being lowest for steel on bronze; and that the theoretical equations|
good prediction of actual conditions. Which is the turning moment that must be exerted on the screw to /ower the

8.22 EFFICIENCY OF A SQUARE-THREAD SCREW. Us

in the design of screws, the aim is to obtain a large mechanical advé
FIGURE 8.26 Forces ona Body Moving Down an Incline. Load being
and since the power transmitted is small, the efficiency is compa lowered by a square-thread screw. Ky
unimportant. The efficiency of a screw would be 100% if there

ba!
250
 ee ee a ef re | re

load. If A is greater than 8, T will be negative, which means that no effor

is required; that is, the load will lower itself. Thus, the condition for self
locking of a square thread is that 8 must be greater than A, or that ta
(the coefficient of friction) must be greater than tan A (the tangent of
lead angle). This same test may be used to approximate the self-lock FIGURE 8.28 Photoelastic Study of a
point for Acme threads, even though the exact equation involves the thre Curved Beam. Notice the crowding of lines,
angle; the slope of the thread sides changes the result so little and f va indicating a stress concentration effect.
so much. (Courtesy T. J. Dolan).
Even though a screw is self-locking under static conditions and has a
very small lead angle, as in screw fastenings, the load may move down by the crowding of the lines of the photoelastic picture of Fig. 8.28. There
(or a nut loosen) under a vibrating condition. are several solutions to this problem; the following is the Winkler—Bach
formula.':7) The flexural stress at a point in a curved beam is

8.24 DESIGN OF SCREWS. It will be quite conservative

the design of axially loaded screws on the root area, because the threa
to bs
(8.12) ss = ae
a eee
Z(r +c)
. [BENDING STRESS ONLY]

provide a definite strengthening. One could use in design some dian

between the minor and major diameters, but if the effort is to get someth
where M is the bending moment at the section (the moment of the applied
in the vicinity of the smallest safe size, the design should be substanti force about the centroidal axis) Fe, Fig. 8.29. The bending moment M
by test. Let the discussion be on the basis of the minor diameter D,. should be given a positive sign when it acts to decrease the radius of
the example of § 8.14. curvature, a negative sign when it acts to increase this radius.
A is the area of the section in square inches.
If L,/k < 40, where L, is the equivalent length, design on the basis
r is the distance from the centroidal axis to the center of curvature of the
W = 5,A,.If L,/k > 40, use a proper column formula. Be careful of a situ
centroidal axis of the unstressed beam.
tion similar to that of an automobile screw jack; for example, if one corn c is the distance from the centroidal axis to the point at which the stress is
of an automobile is jacked up, the screw would probably not act as a fr desired. It should be given a positive sign when measured away from the
end column, but on the other hand, the restraint is not sufficient to class center of curvature, Fig. 8.29, a negative sign when measured toward the
it as a rounded end. center of curvature.
Z is a property of the section as defined by the relation
ydA
-zA={ r+y
y being the variable distance measured in the direction BB from the
centroidal axis, Fig. 8.29. See Table AT 18.
Equation (8.12) can be written as
FIGURE 8.27 Hand-lever Punch. (Courtesy Joseph T. Ryerson
Son, Inc., Chicago).
K,Mc [} 7 eat
(1) Ss, = where K,= ae

is a curvature factor related to the radius of curvature r for a particular

FIGURE8.29 Curved Beam. The sign of c PF c led

8.25 CURVED BEAMS. The straight-beam formula, s = Mc/J, do is negative when substituted in the curved- M(+)
not apply safely to a curved member subjected to bending, as for examp: beam formula, the sign of c’, positive. The Locus of N\
dotted forces produce a positive bending Neutral Axes . | =| boy,
at the curved section of the frame in Fig. 8.27. In a curved beam, the neutr:al
moment, the solid forces, negative. For the » ae
surface shifts away from the centroidal axis toward the inside of the section. forces F acting, section BB has a uniform - FIM(-) 4B Vorb
In effect, there results a stress concentration on the inside fibers, as suggeste d tensile stress F/A on it, which must be ' ci i ;
added to the tensile bending stress given by iy oo -
equation (8.12) in order to obtain the OE vo
maximum tensile stress.
 ea

x Locus of
Neutral Axes §8.27] THICK-SHELL CYLINDERS 255
\\ FIGURE 8.29 (Repeated).
ees 1B uneconomical dimensions. A more accurate expression for the tangential
stress o,, which is a principal stress, is obtained by assuming that the thick
I Centroidal Axis,
4, Section BB shell is composed of a series of thin shells of differential thickness, for each
of which the stress is uniform, and that plane transverse sections remain
plane (that is, all longitudinal deformations are the same). These assump-
tions lead to Lamé’s formulas"-7!;

point in the section. Some values of K, are given in Table AT 18 (comp (8.13) o; an Pir? eh PE ye x riro"(D; a3 Po)|r?
circular-section values with K, for springs in Fig. AF 15), and also equati ro? — rj 2
for calculating Z for two common sections.
The distance from the centroidal axis to the neutral axis of a cu where r is the radius to the point where a, is desired, as at A, and the other
bending only, measured toward the center of curvat re. symbols are defined in Fig. 8.30, say, ro, r; in., po, py, o psi. The radial
member in
stress o, is given by
Fig. 8.29, is
Po)/r?
Zr (8.14) 6, = Pir? ie Peter ; ae ; ( DP; ra
re?ro?
(m) Yn
ET gt ro? — ty
The maximum tangential stress o, is seen to occur at the inside where rr = 3
Knowledge of the location of the neutral axis may be important if,
let r = r,; and get
example, a hole must be put into a curved beam subjected to rep
loadings. Their effect as a stress raiser is minimized if the hole is on o Piro? + r;?) ak 2p oro”
the neutral axis. Holes at some distance from the neutral axis may reé (8.15) Oo,ti = aes >

in peak stresses higher than those at a smooth external fiber.

If there is a uniform stress F/A on a curved section, the total stre: Let r = r, and find the tangential stress at the outside as
usually taken as F/A + K,Mce/I, an algebraic sum. The kind of s
obtained from equation (1) is easily determined by inspection; but my 2p? — Polo? + 1:7)
(n) or
the signs for the terms in equation (8.12). Equation (8.12) is som e ro? — Wi?
unconservative for curved sections with flanges, as a T or I, because In (8.14), when r = r,, the radial stress o, = —p,; when r = 1r;, then
inevitable distortion of the flange results in stresses somewhat higher t o, = —p;; the maximum numerical value of a, is either p; or p,, whichever
those predicted by (8.12). is larger. The radial stress at an internal point is between p, and p,. These
stresses a, and o, are principal stresses. For a solid cylinder (shaft), r; = 0.
a As pointed out in Chapte If the inside pressure p; is negligible and the external pressure p, is
shell cylinders subjected to fluid pressure can be assum large (a closed cylinder in deep water), the term with p, may be safely
to be uniform. However, when the wall is thick, Fig. 8.30, the tanger dropped; o, as computed from (8.15) is at the inside and algebraic, com-
stress at the inside surface is much higher than that at the outside surf pressive if negative. Similarly, as in the case of many pressure vessels, if
and the distribution is no longer approximately uniform. A simple equa’ Po is negligible, as atmospheric pressure when the internal pressure is
5; = pjr,/t, analogous to the thin-shell equation (u), § 1.25, except that high, let p, = 0 and equation (8.15) gives the corresponding internal tan-
outside radius r, is used instead of r;, called Barlow’s formula, is sui gential stress o,. The third principal stress is the longitudinal stress, equation
for an approximation. A thickness too small is dangerous; too (0) below, which is between the values of o, and o,. Thus, the maximum
shear stress is (0, — o,)/2, equation (8.3); the maximum shear stress at the
FIGURE 8.30 Stress Distribution in a Thick Cylinder. inside surface of the cylinder is (o, + p;)/2, or
manner in which the tangential stress s, is distributed along a
AB for internal pressure only. The maximum stress is much
than the average. Two of the principal stresses are s, the tange
(8.16) = Lo
(Pi
2 = Po)
2? [INSIDE SURFACE]
Kou eee
stress and s, the radial stress. At the inside surface, 5, =
the outside surface, s, = P,. The longitudinal stress o, in a closed cylinder is computed on the assump-
tion that it is uniform and that the section in question is not close to the
ace COMBINED STRESSES §8.27] FORCE AND SHRINK FITS 257
[Ch.8

ends; external force on the ends is p,z7r,”, internal force is pyrr,7; the diffe} . a high compression, some 140 ksi in a certain test (r,/r; = 2) on 4340,'8-18!
ence of these forces is resisted by the metal wall, sA = o,n(r,? — r?) while the residual tangential stress at the outside is some 90 ksi tension.
thus Thus, the inside fibers will not be stressed in tension until the pressure is
2 2 more than enough to override the residual compression.* Autofrettage
Pili — Po%o
(0) Oo; [CLOSED E improves the fatigue strength significantly for normal working pressures
lo" —1? when the repetitions of load are some 10* or more, the improvement being
Equation (8.15) represents the maximum normal stress theory, and it greater the more the number of loading cycles. Because of thestress gradient,
often used for brittle materials; it is much more conservative for duc the material becomes less efficiently utilized as r,/r; increases. Thus, if the
materials than the other stress theories. Equation (8.16) would represent r,[ry seems excessively large, the use of a stronger material will reduce the
maximum shear theory as defined when s,, = s,/2, but there are subs ratio, and: by using less material, it may be less expensive. From another
tiating data for use of the yield stress from a torsion test in the maxim point of view, the use of autofrettage may permit the use of a less expensive,
shear equation."-161 (The materials of Ref. (8.15) had s,, = 0.555 weaker material. Each situation must be appraised. Indicative of the stress
torsion.) The maximum shear stress is the same in closed- and open: distribution, if r,/r; = 2 with p, = 0, the inside tangential stress is 2.5 times
cylinders, for a particular internal pressure, and this stress correlates that at the outside. The difference between these stresses decreases as the
with failure conditions, either static or fatigue.'®-18! wall thickness (and r,/r;) decreases, and autofrettage makes possible a more
Also good correlation is obtained with the octahedral shear str efficient use of the material. But as r,/r; approaches one, the benefits of
which can be obtained from equation (p), § 8.12, with the principal str autofrettage approach zero (the material is becoming more nearly uniformly
as defined above. After considerable algebraic manipulation for inte: stressed). A jacket shrunk on the working cylinder (§ 8.27) puts the working
pressure only,"-7! the wall thickness by this theory is cylinder in compression, a practice that has been widely used in guns, and
provides strength in much the same sense as autofrettage—by inducing
compressive stresses in the working cylinder.
(8.17) t= Ge) — 1 in.,

where s is the circumferential stress, a safe normal stress in design. 8.27 FORCE AND SHRINK FITS. These fits are used for connecting
cylinder-wall thickness ¢ can be incorporated into the other equations hubs and shafts, sometimes in addition to keys, when an especially rigid
usingr, —r =t. connection is desired; see §§ 3.7 and 3.8, and also Fig. 4.20, § 4.27. How-
If the internal pressure is static, or seldom repeated, some vessels mi ever, shrink fits become economically advantageous for vessels where,
be safe if designed so that the maximum shear stress at the inside does usually, the internal pressure is high, as explained below. Moreover, a steel
exceed the shearing yield (or more conservatively, if the tangential st jacket shrunk onto aluminum, copper, or other metal liner may be a less
does not exceed s,, but this is not rational design because this theory is expensive approach for, say, corrosion resistance.
supported by experiment for ductile material). The stress gradient, Fig. 8. Consider first the more general case of a hollow cylinder with an internal
shows that the stress decreases from the bore outward. If the pressut pressure p,, with a hoop (or hub) shrunk on, Fig. 8.31 (p. 258); the pressure at
repeated a sufficient number of times, fatigue failure becomes the li the interface is p,; the external pressure is p, = 0. The equations (8.13)-(8.16)
method, and in this case, the high tensile stress at the inside becomes give the stresses as defined for thick cylinders, but p,; must be determined.
ficant. To provide a longer life or to reduce the wall thickness for g The pressure p, at the interface depends on the interference of metal (§ 3.7),
specifications, autofrettage (§ 4.23) is practiced. If the autofrettag the various radii, and the materials of the parts, and it results in an increase
carried to 100% overstrain, all the material of the wall has been stre of the hoop radius of 8, in. (diameter increases 25,,). The radius of the inner
to the yield strength. Assuming plastic action, after the smaller-diar member decreases some amount 5,, its diameter by 25,. The interference
fibers have reached the yield point, the stress in them remains cons! of metal (§ 3.8) is therefore i = 2(|8,| + |8,|) = D, — D,, where D, is the
while the stress in the larger-diameter fibers continues to increase to Shaft diameter (outside diameter of the inner cylinder) and D, is the inner
value as the pressure increases. On this assumption, the internal pressui hub diameter, both measured in the unstrained state.
required for 100% autofrettage ™-7! is Recall from strength of materials that when there are two perpendicular

(Pp) Proo = 2Sys In(r,/r;)- * In case autofrettage is used, equation (p) may be divided by a factor of safety to get
P, = 25,4 In(r,/r,) in which p, is the working pressure and s,, = s,,/N is a design stress.
When the pressure in removed, the residual tangential stress at the insic Space is not available for the specialized information needed for this approach.
258 COMBINED STRESSES| §8.27] FORCE AND SHRINK FITS 259

normal stresses, say s,; and sp, the deformation in the direction o of metal i, Now the tangential stress in the hub is obtained by substituting
« = 5,/E — pSa/E, Poisson’s (u) effect considered. If the lon the value of p, from (t) into equation (8.15) and letting p, = 0; in equation
deformation is negligible, frequently assumed even if not strictly tr form, the result is
unit deformations e¢,, ¢, in the tangential direction of hoop and sha
respectively (D, ~ D, = D, for this purpose), (u) on = sl! re (2)].

(a) aE Sore
as en ® a tensile stress. If it is desired to have a safe maximum shearing stress in
the hub, let p, = 0 in equation (8.16) and substitute p,; from (t);
Ei
" Papa osr ena ton
28, Sts — Bsr (v) T [SOLID SHAFT OF SAME MATERIAL AS HUB]
Vapf
where pz is Poisson’s ratio. The interference of metal is Often, the computed stress is allowed to approach closely the yield
strength, because the stress decreases away from the bore. The foregoing
equations (t), (u), and (v) strictly apply when the shaft is the same length
(s) fese 4 Me AN p| Ot + BrP Pee
Ey
+
Eo
od
as the hub. The part of the shaft outside of the hub resists the compressive
deformation, which resistance results in an increased radial pressure just
used to compute p; with a known value of i; where o,, = o,, = inside the hub.
If the hub is cast iron and the shaft is steel, the following equations
may be obtained by assuming that the modulus of elasticity for steel E is
twice that for cast iron E,, E = 2E, = 30 x 10°, and that Poisson’s ratio
for steel and cast iron are virtually the same"®-18!;

si Ei{l — (D,/D,)?]
(w) i
DB +e + = 2) (D,/D,)?7
Ei[l + (D,/D,)?]
(x) Sin =
D3 +e + (lL — »)(D,/D,)7)’
where » may be taken as » = 0.27. If the shaft is hollow and if the diameter
of the hole does not exceed 25% of the shaft diameter, the error in using
the foregoing equations is about 7% or less. However, appropriate allow-
ance for a hollow shaft is made by following the procedure outlined above
but not using r, = 0, Fig. 8.31. Equations for any other combinations of
E,, +s for the material of the inner member. For the most genera materials can be similarly obtained. See Ref. (8.20) for considerable addi-
o, = %,4 in equation (8.15), but usually p, = 0; o;, is o,. obtaine tional detail.
equation (n) and is the stress on the inner member at the interface. The torque that the fit will transmit and the force needed to make the:
inner member is the liner of a pressure vessel, the internal and © fit may be estimated after p, has been calculated. If the length of fit is L
pressures are p, and p,, subscripts defined in Fig. 8.31. In the case of and the diameter D,, the contact area is 7 D,L; the normal force is p,(7D,L);
ona shaft, the internal pressure p, is likely to be zero; and the most fr the frictional force is fpyrD,L; and the moment of the frictional force about
case is a solid shaft {r; =-0). Adapting equation (s) to a solid shaft the shaft center is fp;(7D;L)(D;/2). Thus we find that the torque is
hub of the same material, E, = E, = E, ty, = bs = 4s Tilo = Dil Do
(y) Pe Spa D?L
oo
f n= s(t - (3) |. a The value of f = 0.1 is recommended for use here by Baugher.'8-18] For
from which the interface pressure can be computed for a known inter severe duty, it would be advisable to consider f ~ 0.05.
 COMBINED STRESSES §8.28] CLOSURE 261
260

The axial force required to press the hub onto the shaft is the product 8.28 CLOSURE. This chapter endeavours to relate engineering
the area (7D,L), the normal pressure p;,, and the coefficient of friction considerations with some of the more complicated stress states. An under-
Expressing the units of this axial force F in tons, we get standing of applicable theory should precede the engineering, and for more
details on theory consult books on strength of materials. Strength is often
wks Apia D,L
tons. affected by anisotropy (§ 2.2) and comprehensive data on this characteristic
(z) 2000 are seldom available.
It may be helpful to summarize modes of failure: They may occur by:
Since f, is quite variable (Baugher'®*®! found variations in fj from (1) plastic deformations, that ruin the shapes and relations of parts; this
to about 0.3), this formula will give only approximate results. is elastic failure; (2)-fracture, which is the kind of failure considered for
In practice, the engineer is interested in optimum proportions (espe brittle materials and for fatigue of ductile materials; in addition to the
true in case of a pressure vessel), in which case a decision must be mad external loading, resonant vibrations may result in fracture; (3) creep,
to what to optimize. For example, for hub and shaft of the same materi which is a subdivision of (1) and is usually associated with high-temperature
it may be assumed that the maximum tangential stresses in each part, on deformations, although some materials, notably magnesium among
inside of the cylinder in each case, are the same. If the parts are of differ engineering materials, may creep excessively at ordinary temperatures;
materials, these maximum stresses, instead of being made equal, could (4) wear, which is due to relative motion of surfaces in contact under
say, in proportion to their yield strengths. Let us assume that the maximu pressure; this failure may be totally an abrasive action that removes material
shear stresses are to be made equal for the case of a hollow shaft w or, as in the case of gear teeth, it may be that the contact stresses are so
internal pressure (p, = 0 in Fig. 8.31), with hub and shaft of the high that a “‘surface fatigue’’ occurs; (5) overheating, which is a phenomenon
material. Adapting equation (8.16) to the symbols of Fig. 8.31, we gi that speeds up failure by other modes; for example, it may result in a lubri-
2 2 cation failure with excessive wear, or the overheating may destroy needed
re Pi ro Pi
mechanical properties, leading to other kinds of failure.
T ie
nea — ry GtVnuai
fo by
—N ae It is appropriate at this time to reread §§ 1.16-1.18 on factor of safety.
[LINER] [HuB]
Observe that, given a particular machine element, one could compute its
“factor of safety’’ by each of the various theories of strength and the answers
where both shear stresses are taken with the same sign. Solve for r; and
will be different. Since failure is an engineering phenomenon, it (and the
(a) rye = Vrilo- range of the true factor of safety) can be determined only by experiment
(past experiment may be used to predict failure); hence it is better to speak
These proportions for power transmission through a hub are not necessam of a design factor, rather than a factor of safety, and relate the design factor
practical. One should not depart from recognized good proportions wi to the particular way in which it is computed.
knowing that the departure is good engineering. If the system is a pr
vessel, the amount of interference of metal should be optimized.
different approaches are found in Ref. (/.7) and (8.20), too lengt
detail here. Note that the use of shrink-fit pressure vessels is a me
stressing the material more uniformly, with a resultant saving. Fau
gives a curve comparing solid construction with shrink-fit constru
for example, suppose the solid construction requires a wall thickness
that r,/r,; = 3, corresponding to a rather high internal pressure; for
same pressure, this ratio for shrink-fit construction is about 1.8, indi
a significant saving. Carrying the idea further leads to a laminated
in |
struction, a number of relatively thin shrunk-on cylinders, which
limit of differential thicknesses provides the “‘most optimum” solu
As the number of laminations are increased, the diminishing returms
suggest a limit. The greatest percentage improvement comes from the
step, a liner and a single jacket. Do not forget autofrettage, § 8.26. ;
A hub on a rotating shaft is subjected to centrifugal forces, which shou
be accounted for when the speed is unusually high.'*.?0
 9. SHAFT DESIGN

9.1 INTRODUCTION. Although the theory needed for shaft design

is covered in Chapter 8, shafts are so ubiquitous that special treatment and
more engineering information is worthwhile. We shall also present briefly
the former ASME Code on transmission shafting.®-!?
As the term is usually-used, a shaft is a rotating member transmitting
power. Especially as used in the past, an axle is a stationary member
carrying rotating wheels, pulleys, etc. However, common usage favors the
word shaft whether the member is rotating or not. On the other hand, as a
carryover from wagon and buggy days, we speak of the ax/e on an auto-
mobile. A line shaft, or main shaft, is one driven by a prime mover; power
is taken from it by belts or chains, usually at several points along the
shaft. Shafts intermediate between a line shaft and a driven machine are
variously called countershafts, jackshafts, or headshafts. Short shafts on
machines are often called spindles.

9.2 BENDING FORCES PRODUCED BY BELTS AND CHAINS.

The force F in the horsepower equation is the net driving force. In a belt
drive, for instance, this net force is F = F, — Fy, where F, is the tension in
the belt on the tight side and F, is the tension on the slack side (§ 17.2).
Since both of these tensions pull on the pulley, the corresponding bending
force on the shaft is F, + Fa. The sum of the tensions is not constant in a
drive but depends upon the ratio F,/F2, which varies with such factors as

263
 24" Pulley
B 18"Gear
264 SHAFT DESIGN [Ch.9
Total 2 a ae
the power transmitted, the speed, the initial belt tension (hp = 0); the A
Torque
4 p_| |
—

initial tension to which the belt is subjected depends on other factors in DH

the belt drive (Figs. 17.7, 17.8).
The bending force F, + Fz can be computed by assuming F,/F, (see
| | |
60% of Torque
fr
Delivered
Fig. 17.10) or by assuming C in F, + Fy = C(F, — F,). For a flat belt,
12" 18” 10"--b— 10"
unless it is significantly overloaded, a value of C = 2, which corresponds to
F,/Fo = 3, is reasonable when the belt is mounted with a proper initial
bis 50" F,
tension (see Chapter 17 for more information on belts). If there is reason
FIGURE 9.1 The gears G and H “‘push back’? on C and E, respectively.
to be conservative, use a value of C of 2.5 or 3 or more. In the absence
of other information, assume the bending force of a flat belt to be
horizontally to the right. A 12-in. gear E delivers the remaining power downward
(a) F, + Fo = 2(F, — Fa), toward the left at an angle of 30° below the horizontal. Both gears have 20°
where F, — Fz = Fis usually obtained from the horsepower equation (1.1. involute teeth. The shaft is to be machined from annealed C1137, with profile
keyways for each gear and pulley. The load is smooth. (a) Find the diameter of
As the belt stretches in service, F, + F, decreases for a given power
the shaft for N = 1.8 when the Soderberg criterion is used for the equivalent
speed.
stresses. (b) Let there be a step-down in the size of the shaft at the bearing D and
The bending force due to V-belts may be taken as 1.5(F, — Fa); compute the diameter from D to E.
say 2(F, — F,) if there is cause for concern. The bending force exerted Solution. The torques transmitted by the pulley and gears are
chains and also by gears, is usually taken as the net driving force F; that
is, the slack-side tension F, in chain drives is taken as zero. _ 63,000 hp z (63,000)(30)
Ty = . 360 = 5250 in-lb. (on shaft between B and C).

T, = £2,00002) = 2100 in-lb. delivered.

9.3. DESIGN OF SHAFTS FOR STRENGTH. If a shaft carr 360
several gears or pulleys, different sections of the shaft will be subjected 63,000)(18 ?
different torques, because the total power delivered to the shaft is take T. = oe = 3150 in-lb. (on shaft between C and E).
piecemeal at various points. Hence one must note the amount of to
on each part of the shaft. Then study the distribution of the ben N ue 7
moment, preferably sketching (freehand is all right) the shear and bendin FIGURE 9.2 Forces on Gear Tooth.
we Co)

moment diagrams. - ‘ F
From this preliminary examination, which is a problem in mechan Pitch Circle Ws
we note the section where the bending moment is a maximum and t
section where the torque is a maximum. If these maximums occur Let the bending force produced by the belt be (ry = radius of B)
same section, the diameter needed for that section is determined—and
2T, — (2)(5250)
for the entire shaft when the diameter is to be constant. If the maxim Fy = (Fi ia) eo ag eee
o
do not occur at the same section, determine the diameter for the section
maximum torque and also for the section of maximum bending mome For the gears, the driving forces are computed as though the contact is always
and use the larger value. ; On the pitch circle (r. = pitch radius of E; etc.)
The diameter of a shaft is often varied from point to point, some
Sap ae een 25 lbs and F, = — = —— = 233 lb.
for structural reasons. In this case, check the stress or determine the
needed for each section. The designer makes certain that all sections 0 From your study of the kinematics of gearing, recall that the total load W on the
shaft are subjected to safe stresses, taking due note of fillets, holes, key Sear tooth (ignoring the frictional force) is normal to the tooth surface, with the
and other stress raisers. An example will serve as a reminder of the mech result that there is a separating force, N, Fig. 9.2, which is seen to be F tan 4,
involved. where F is the computed driving force. (See § 13.9). For ¢ = 20°, the separating
forces for C and E are
N. = F, tan 20 = (233)(0.364) = 84.8 1b.; N. = (525)(0.364) = 191 lb.
9.4 EXAMPLE. A 24-in. pulley B receives 30 hp at 360 rpm from below at| If C delivers power to the right, the force F, on C points leftward, as shown in
angle of 45°, as shown in Fig. 9.1. An 18-in. gear C delivers 40% of the pow
 Ay=S61 E,=428

266 SHAFT DESIGN [Ch.

0-1
F
12-—j-— 18 —=}- 10
the end view of Fig. 9.1. Similarly, with E delivering power as stated, the force
is upward toward the right. By analytic mechanics, those forces acting at so m By=619 Cy=84.8 Dy=285,7
distance from the center of the shaft are replaced by a force through the shaf
FIGURE 9.4 Section of Maximum Bending Moment, Vertical
axis and a couple. Thus, add and subtract forces F. through the shaft axis a
Plane. Diagram
indicated and note that there will now be a CC torsional couple F.r., where
is the pitch radius of gear E, and a bending force F, through the center of h B
shaft parallel to the original F.. This is the justification of the free bodies to b
used later.
At this stage, the easiest way to design a shaft subjected to a three-dimensiona
force system is to resolve the forces into two perpendicular coplanar syste
The horizontal forces at B, C, and E are
Sketching the shear diagram, Fig. 9.3, note that it crosses at B, which is therefore
B, = F, cos 45 = (875)(0.707) = 619 Ib. the section of maximum bending moment in the horizontal plane. The sum of
C. = F, = 233 Ib. the moments to the left of the section gives
E, = F. cos 30 — N. cos 60 = (525)(0.866) — (191)(0.5) = 359.1 lb. Mo: = (12)(581) = 6972 in-lb.
Looking at the end view of Fig. 9.1, imagine the horizontal plane rotated 90° | t The forces in the vertical plane are indicated in Fig. 9.4; the y components of
and sketch the forces in this plane as shown in Fig. 9.3. To find the bearing the force system of Fig. 9.1 are
tions at A and D, take moments about D and about A, and check the
B, = Fy cos 45 = (875)(0.707) = 619 Ib.
by =F, = 0. About A,
Cy, = Ne = 84.8 Ib.
M, = (12)(619) + (30)(233) — (50)(359) + 40D. = 0, E, = F. sin 30 + N. cos 30 = (525)(0.5) + (191)(0.866) = 427.9 lb.
The values of the bearing reactions A, and D, are determined by moments as
or D, = +88.3 lb. The positive sign for D, tells that its sense as shown is cor: before, from a free body of the y components, Fig. 9.4, and checked by =F, = 0.
or, since the clockwise was taken as positive, the positive sign for D, indicat The results are as shown. Since the shear diagram crosses zero at section B, the
maximum moment in the vertical plane is also at B, of value

E,=359 oy = (12)(561) = 6732 in-lb.

(Note. The horizontal and vertical shear diagrams do not always ‘“‘cross”’
at the same section. Sometimes, one or both cross at more than one section. If
—— 18 ———] oa the shaft is to be made a constant diameter, each section where the shear diagram
{
crosses zero must be considered to make sure that the size of shaft obtained is
C.=233 D;
FIGURE 9.3 Section of Maximum Bending safe at all sections. The maximum moment will be where at least one diagram
Horizontal Plane.
crosses.) Since both shear diagrams cross zero at B, this is the section of maximum
bending moment, which is the vector sum of the components computed above;
Mz = (Mp? + Moy?)*/? = (6972? + 67327)*/? = 9680 in-lb.
or 9.68 in-kips. We now notice that the maximum bending moment and the
maximum torque both occur at B, which is therefore the significant section.
(NoTE. If the maximum bending moment occurs at a section that is not subjected
to the maximum torque, some section of maximum torque must be investigated
as well as the section of maximum bending moment.)
Summing moments about the right bearing, we have (a) Designing by Variable-stress principles (Chapters 4 and 8), we probably
Need to include a size factor, say 0.85; there is a keyway at section B, for which
=Mp = 40A, — (28)(619) — (10)(233) — (10)(359) = 0
K; = 1.6 and K;;= 1.3 from Table AT 13 (for annealed steel); mechanical
from which A, = +581.3 Ib. Again A, was chosen in the correct sense. Sum t Properties for AISI 1137 from Table AT 8 are
forces as a check, =F, = 0.
Since we are interested in locating the point of maximum stress, we @ Ss, = 85 ksi, y = 0 KSIt use s,, = 85/2 = 42.5 ksi.
definitely interested in the sections of maximum bending moment, in each p ) i
Assume sn; = (0.6)(42.5) = 25.5 ksi and sy, = (0.6)(50) = 30 ksi. (Observe that

267
268 .
HAFT DESIGN [Ch.9 §9.5] SHAFT SIZES AND MATERIALS 269
q
i (8.11), t he results accord closely but: nat 9.55 SHAFT SIZES AND MATERIALS. Transmission shafting,
for these values of sy, and s,, ini equation varies |
Since the flexural stress used to transmit power from a prime mover to a machine, comes in the
not exactly with he octahedral stress theory.) use in, say,
stress to
through a complete cycle, we need to find the equivalent following sizes: '-5!
= 9.68/(7Q*/32) ksi; and from
equation (8.11); su = 0; therefore, se = Me/I
get tg, 1x, lye, 148, 14%, 2%, 2s, 248, Bye,
equation (w), § 8.15, we
Sn ec (1.6)(9.68)(32) 348, 4x5, 448, Sate, = S48, = 64, 7, 7h, 8.
Commercial bearings and couplings are readily obtainable for most of
= (5.250)/(7D°/16) ksi; from these sizes. Stock lengths of transmission shafting include 16, 20, or 24 ft.;
Assume a steady torque; Sos = 0, and Sms = Te|J lengths over 24 ft. can be obtained on order. Typical sizes of machinery
equation (x), § 8.15, we find
shafting, shafts that are an integral part of a machine, are (with tolerances
Sn
(25.5)(5.25)(16)
Sai pies aca i
Ses = ie + KysSas (30)(7D*) +0 for finished shafting):

on (8.11); with size fa By 7¢-in. + to 1 in. with tolerance of —0.002 in.

These equivalent stresses are substituted into equati

4-1 eT" <

increments in 1; to 2 in. with tolerance of —0.003 in.
this range 235 to 24 in. with tolerance of —0.004 in.
By 4-in. increments, 23 to 4in. with tolerance of —0.004 in.
if ( 1.6 x 9.68 x 32 y is fe x 5.25 x ip) By }-in. increments, 4} to 6in. with tolerance of —0.005 in.
0.85 x 42.5 x 7D* 25.5 x 307D* By }-in. increments, 6} to 8 in. with tolerance of —0.006 in.
1 16 +a Shafts are made in several different ways and of a wide variety of
ann =pie-735 + 0.0306)*/?,
materials. The shafting referred to above is likely to be cold-drawn carbon
i = 2; use D =2 in. This size shaft should be satisfactory steel in sizes smaller than 3? in. Shafts are also cold finished by turning and
gears, it should be determined that polishing, sometimes with a grinding operation also included. See §§ 2.9,
oe especially with Lees
i ropriate (§ 9.11). 4.23, 4.30 for effects of cold working. If the finish meets the requirements,
ie oe a is 1 in. wide (axial dimension); then, certainly ; cold-drawn shafting may be the lowest in cost because of its superior
the bearing may be —
first approximation, the mqment at the center of : f mechanical properties. However, because of the residual stresses from cold
tinuity occurs at the left side of D, where tl
design, even though the discon drawing, the shafts will warp when a keyway is cut, necessitating straighten-
= yin. Checking in Fig. AF 1 j
fillet whose radius must be assumed; say r ing. Straightening shafts and axles is usually done by cold-bending them
ve proces s must be used because r/d and D/d must
conclude that an iterati opposite to their curvature until there is plastic yielding, which invariably
resultant force at E is
known in order to find K,. Assume K; = 2. The
(359? + 428%)'/? = 560lb.; the moment at D is 10F. = 5.6 in-kips. The to leaves a residual tensile stress somewhere in*an outer fiber. A reduction of
ed above. As before endurance strength results, which may be aggravated if a tensile residual
on the shaft at D is Tz = 3.15 in-kips, comput
exists in a fillet or other stress raiser. For example, unstraightened auto-
(2)(5.6)(32) Sas (25.5)(3.15)(16)
) an Su= aan
(30) 7d motive axles had s, = 20 ksi; straightened in normal production, s, =
Se = Kjsa =
(0.85) (ad) Sys
13-16 ksi; straightened and peened, s, = 43 ksi.'*-1° Instead of bending
factor 0.85, we have
Substituting into equation (8.11), including the size the shaft to straightness, it may be straightened by peening with a hammer
5m 3.15) ‘‘
shod + Gowass)]
selectively at the proper points; in one case, this plan left a shaft with
urig~wlloassxars)
N 18
: is to a D
unchanged 5,.
Larger shafts are turned and/or polished and ground from hot-rolled
i
which i (except that ifi a ro lling bearing
d = 1.54; : use d = 14 in.
tion for the ne material; the turning removes some decarburized steel. Sizes over 5 or 6 in.
e bore). In prepara
aay the size would be adjusted to an availabl are usually forged, and turned to size. Crankshafts, as for automotive
For r/d = 0.0625/ 1.5 * e
iteration, check the value of K, assumed above. Fig. AF engines, may be forged or cast. They are cast from steel, high grades of
Fig. AF 12, K, = 2.15. From
and D/d = 2/1.5 = 1.33, we find, from cast iron, and nodular iron.
get q = 0.86; thence
; The most commonly stocked material for shafting would be carbon
K,=1+q(K: -—1)=1+ (0.86)(1.15) = 1.99;
steel with some 30 to 40 points of carbon, perhaps resulfurized (11XX
on d found should be sat
since this is close to the assumed K, = 2, the dimensi series) for finishing in automatic screw machines. But all kinds of materials,
for strength. including nonferrous metals and nonmetallics, are used for a reason.
 §9.9] TRANSVERSE DEFLECTIONS 271
270 SHAFT DESIGN [Ch.9

Materials for automotive crankshafts and axles include 1345, 8637, 8650, solid round member can be computed from equation (1.6), which gives the
3140, 4135, 4150, 5145, 4340. Higher carbon contents are used in the la’ correct value at the neutral plane;
sizes for greater hardenability. Heat treatment may be to BHN of 229.
16V AV
for crankshafts, to BHN of 300-444 for axles, the higher BHN being us (b) Ss = — [SOLID ROUND ONLY]
for the larger sizes (24 in.). When localized surface hardening is des: MdSanDeen) 36;
as by flame and induction hardening, low hardenability may be preferab)
where V lb. is the vertical shear at the section under consideration. In a
In general (in all design), no more carbon is used than is needed to ob
uniformly loaded cantilever beam, the vertical shear may govern when the
the required properties. Sometimes, a reduction in carbon may
length is less than about half the diameter. However, if the shaft is subjected
trouble with quench cracking. In one instance, a carbon-steel, 22-in. s
to torsion as well as bending, the vertical shear combines with the torsion,
with a keyway had to be induction hardened at the keyway to ob
say at the neutral plane, so that the consequences may be significant in a
sufficient endurance strength. All these shafts with 0.43% C crac
somewhat longer shaft, and especially for hollow shafts approaching the
when the carbon was reduced to 0.37 %, there was no further trouble
tubular.-1*) It is also true that the deflection (§ 9.9) due to shear becomes
cracking, and the part had sufficient fatigue strength. If parts of the s
significant in short beams (see texts on mechanics of materials), but for
are to be carburized, lower carbon, as 0.15-0.20 %, is used. The possibili
most shafts this is a negligible quantity,
of fretting corrosion if forced fits are used on the shaft, the effect of a
kind of corrosion, of oil holes, fillets, etc., on the fatigue strength shou
not be overlooked (§§ 4.21, 4.26, 4.27).

9.8 TORSIONAL DEFLECTION. Deflection is another and fre-

quently significant consideration in the design of shafts. Criteria for the
limiting torsional deflection vary from 0.08° per foot of length for machinery
9.6 HOLLOW ROUND AND SQUARE SHAFTS. Hollow re shafts 8-61 to 1° per foot '-5) or 1° in a length of 20 diameters for trans-
shafts sometimes serve a useful purpose, usually in large sizes, though mission shafting. Even short shafts become special problems in rigidity
are more expensive than solid ones. They have the advantages of b when the load is applied in impulses, as on an automobile crankshaft. The
stronger and stiffer, weight for weight, because the outer fibers are impulses produce a torsional vibration, usually compensated by torsional-
effective in resisting the applied moments, and they respond better to vibration dampers in an automotive engine. The torsional deflection of a
treatment because quenching can proceed outward as well as inwat round shaft is given by equation (1.13), @ = TL/GJ radians, where G psi is
The theory already presented applies exactly to hollow rounds, who: the modulus of elasticity in shear, Jin.* is the polar moment of inertia
section moduli are given in Table AT 1. The designer commonly ass! of the section, and L in. is the distance from the section where the torque
a relation between the hole size and the outside diameter. See § 9.13. T in-lb. is applied to the section of the resisting torque. Review also § 1.13
So-called square shafts are occasionally needed, but the edges sh now.
be, and usually are, generously beveled in order to remove sharp
which may be the source of failure. The values of Z and Z’, from Table
used for beveled square shafts should give conservative answers. The sim
torsion formula is not strictly correct for noncircular members, and| 9.9 TRANSVERSE DEFLECTIONS. Books on mechanics of
interested reader should look further into the theory for these cases.4:7™ materials explain various methods of finding deflections of beams. The
basic mathematical approach is to set up an equation for the loading (or
bending moment or M/E/) at any section, in accordance with equation
(1.10), p. 13, and then make successive integrations, equations (1.9), (1.8),
9.7 VERTICAL SHEAR. At this time, review § 1.10(b) on vertices and (1.7) until the deflection y is obtained (Table AT 2).* For one load, or
shear in a beam. As pointed out there, the shear due to bending is a max two, in addition to bearing reactions, the method is not too tedious.
mum at the neutral plane where the normal stress is zero, and the magnitud
* An algebraic method (as opposed to graphical), variously called pointed brackets,
of this shear is relatively small compared with other stresses, except 1 Curly brackets, singularity functions, that is shorter than any other algebraic approach,
short beams (a cantilever crank pin may qualify). Its maximum value or Is explained in Crandall and Dahl." -1®
ata SMAPI MESIGN (LUN. ° Saar TaBwiswis we weirs sa ==
§7.11]

Another approach is to use the principle of superposition; that is, th an M/EI diagram is laid out for a particular beam or shaft (or a moment
deflection at a certain section of a shaft caused by all the loads, F,, Fo,.. diagram if J is constant), the J(M/EI) dx between certain limits is the change
is equal to the vector sum at that section of the deflections caused by each of slope of the beam between the same limits. Since 6 = dy/dx, the next
of the loads acting alone. After one has learned the method, a graphi cal jntegration f@ dx gives deflection y.
integration procedure, described below, has certain advantages wheneve; One method of graphical integration is illustrated in Fig. 9.5.* The curve
there are many loads and changes in the diameter of the shaft. abc +++ is known. The area under it is divided into smaller areas A,, Ag,
Data on permissible values of deflections are rare, probably because th etc., usually with x, = Xx, = Xg, etc. for convenience only, so that a straight
range of values would be large and each situation has its own peculiarities line is a good approximation of the curve at the top of the areas. Points
An old rule of thumb for transmission shafting is that the deflection shoul a, b, ¢, etc. are then halfway across the areas A,, Ag, etc. Draw any
not exceed 0.01 in. per foot of length between supports; although gre convenient vertical reference line mu; draw aa’, bb’, etc.; choose any
stiffness may be desired. Preferably, on transmission shafts, the pull convenient pole O at a distance P from the reference line mu; draw Qa’,
and gears should be located close to bearings in order to minimize momer Ob’, etc.; choose an origin Q for the integration curve; draw Qa” parallel
If journal bearings with “thick film’’ lubrication (Chapter 11) are used to Oa’, a"b" parallel to Ob’, bc” parallel to Oc’, etc., giving the intercepts
deflection across the bearing width should be only a small fraction of a’, b", c",... on the ordinates of the Ax diagram; then g,a” represents to
oil-film thickness; if the slope is excessive here, there will be “bindin scale area A;, gob” represents Ag, etc. For proof, consider the similar tri-
the journal. A self-aligning bearing, Fig. 11.13, may eliminate this trou angles Oma’ and Qg,a"; g,a"/Qg, = ma'/P or
the deflection is otherwise acceptable.
On machine tools (lathes, milling machines, etc.), rigidity is a specia
concern because of its relation to accuracy. If a shaft supports a (c) oo P P P
deflection is more of a consideration than if it carries a V-belt pulley. I
general, for machinery shafts, the permissible deflection may be closer t where it is seen that (la)(pq) is equal to A, (approximately in this case).
0.002 in./ft. (rather than 0.01 for transmission shafts). Gleason'5-1! s Similarly, for triangles Omd’ and c"g,d’, note that g,d” = A,/P. The
that at bevel gears, sizes 5 in. to 15 in., the gears “should not lift or dep: conclusion is that the distances g,a",---g,d" represent to scale the areas
more than 0.003 in.”’ At a section of a spur-gear mesh (good quality ge A,,'**, Ag, etc. and that therefore the curve Qa"b"d" represents the integral
the shafts should have a deflection relative to each other of less than 0.00. of the function. The ordinate kd” represents the total area up to x = Qk.
More important perhaps for spur gears is the relative slope of the shafts¢ Since the area A, in its actual units is the integral at point q (and g;),
the mesh; Brown & Sharpe recommends that this slope be limited t the scale to which it is represented by g,a is obtained by multiplying both
0.0005 in./in. i sides of (c) by the u scale S,, and the space scale S, (that is, the x scale); thus

(d) (814")(P)(Su)(Ss) = (la)(Su)(pg)(Ss) = Ar.

9.10 GRAPHICAL INTEGRATION. Given a curve defined Therefore, the scale of the ordinate of the integral curve is given by
u = f(x); then, as seen in Fig. 9.5, a vertical strip of height u and width
is dA and the entire area under the curve (or curves) is [dA = fu dx. (e) (Pole distance)(ordinate scale)(abscissa scale) = PS,S,.
integrated through as many functions as required. By analogy, we h
equation (1.8), d? = (M/EI) dx, the integration of which gives 0. Hence, i Reference (9.2) gives an analogous numerical procedure for deflections,
and numerical help is found in Ref. (9.3).

u=f (x)
9.11 EXAMPLE—DEFLECTION OF SHAFTS. For the loading defined in
the example of § 9.4, determine the diameter of shaft for a maximum deflection
ARRARARARARALARS,

* The force-and-funicular polygon procedure of making a graphical double integration,

FIGURE 9.5 Graphical ae in previous editions of this text, is probably somewhat more accurate when the
Integration. Same scales are used. But the method described in this edition has the advantage of
giving the slope, which may be important. Any one frequently concerned with deflections
4 Space Scale,S; —= may wish to look into this other method. It can be used to get slopes by starting with the
ma Vertical shear diagram.
Scale for Ordinate
Sa=S. S; P
 Az=581 Br =619
C.=233 D,=88,3 B= 359
} Hub~«
274 SHAFT DESIGN [Ch.9 © T 1.8D XT 0.9 D
eb
of 0.003 in. For pedagogical purposes, assume that gear C has a forced fit with q
= 20 in./in¥ GVH Dy
he
12 —=}-— 18° ——-— 10° | — 190
the shaft and that the shaft size is to be reduced about 10% at bearing D.
Solution. It is recommended that hubs making forced fits with a shaft be» Sw/xr=60x 10-*in.+
considered as integral with the shaft for the purpose of determining deflections.{°-7)_ (Unity D)
Thus the configuration to be dealt with is shown at the top of Fig. 9.6, laid out to a
scale axially but not diametrally. Let the hub diameter (§ 13.31) be 1.8D, Fig. 9
One could use graphical integration with the shear diagram, but it will be mor
accurate and as easy to compute the points on the M/EI diagram (M diagram if J
we os
is constant). For the diameter D, J = 7D*/64 = 0.0491D*; at the hub, J, =
x(1.8D)*/64 = 0.515D*; along DE, I; = 7(0.9D)*/64 = 0.0322.D*. Now comp !
Syme ee te
' I { |
the moment at each load and at each change of section size, tabulating the result: fs,
=0i2 mf
(below) for ease of use. The results will be little affected if the size change I (Unity D) in.
assumed to occur at the center line of bearing D. A similar approximation wo
be appropriate at C, but we shall do this one straight for the instructional val ee tk

The results of the computations are (E in ksi; let D = 1):

i711it
UP SN apres es. hase ONE Sas oy
MGokip). . . 0. 6972 636 636 63. 5.75 "5.75 3.500
(MJEI)D* 10* 0 47.2 43.2 412 4.07 3.72 39.0 244 37.
Layout the second line to scale, Fig. 9.6, and get the M/EI diagram. If there h J
been negative moments (which tend to curve the beam concave downwar
they would have been plotted below the axis. If D is unknown, as in this exa:
think of it as unity until the deflection y curve is obtained, then reinstate it (E an
would be handled similarly if the moment M diagram were used). The scale 0 FIGURE 9.6 Slope and Deflection, Horizontal Plane. Having laid out the M/EI diagram
the M/EI diagram is chosen as 60 x 10-* in.~*. The integration of this curve to a convenient scale, choose pole O; at a convenient pole distance P from some vertical
carried out in accordance with § 9.10; the procedure is outlined in the captior reference line ab, Divide the areas under the M/EI diagram into convenient parts; the
Fig. 9.6. The pole Oz for integrating the slope diagram could be chosen on | boundaries here are dotted lines, as cd. Draw the mean ordinate of each area, as ef. Project
zero axis Q (which is not the axis of zero slope) as for the M/EI diagram. Si the various mean ordinates to the reference line ab, as fg. Choose a convenient starting
point Q for the slope curve and draw lines parallel to the rays 0,5, ...
the area under the @ curve is continuously increasing, the y diagram would th Oyg, ... Ora, as hi
is drawn parallel to Og. After the slope diagram is completed, choose a convenient pole O2
be cocked high on the right. By choosing O2 randomly as shown, the deflecti
for the @ diagram, make the graphical integration to obtain the y curve. Spot the location
curve stays closer to a horizontal. Remember what the ordinates to the @ of the bearings and mark the points, m, n. Draw the straight line mn and measure the ap-
mean; they represent the change of slope from the starting point andthe change propriate departures of the curve from this line; yex = 0.33 in. and y,x = 0.3 in., actual
@ ordinates between stations represents the change of slope between statio measure. See text for conversion to deflections.
Fig. 9.5. Except by some coincidence, the shaft (beam) will have some slope a
left bearing (support). The loading is such here that the curvature of the sh
never reverses (it probably would if force E, were reversed). polygon obtained. Thus, the smaller the areas used and the larger the scale, the
Having chosen Oz, the integration procedure is completed to obtain greater the accuracy of the results. The actual deflection in the horizontal plane
diagram. The bearings are assumed to be rigid and fixed in space, so that if th is the measured distance y times the scale S,,;
positions are spotted on the y curve, a straight line through them is the base fri
which deflections are measured (vertically—always perpendicular to axis. (f) ie 0.3($4) and) yas 0.33(=2) in.
original beam). Thus the bearings A and D are located at mand n, and a str
line mn through these points is the line of zero deflection (mk is the horizontal The scales are obtained in accordance with equation (e) above. For the @ and y
Since the rays from Oz have the slopes of the sides of the deflection polyg curves:
draw m‘n’ from Oz parallel to mn; this locates an axis n’x of zero slope; whe!
intersects the @ curve at p, drop a vertical line downward to measure y:z, the (g) So = PSmjerS. = (1)(60 x 10-*)(20) = 0.12 rad/in., [D = 1]
mum deflection between the bearings (which occurs where a line tangent to
elastic curve of the beam is parallel to the line through the points of suppo!
(h) Sys PSs) «(*2)c20) ib 4+ in/in.
0.12 2.

The true elastic curve is of course a smooth curve and would fall inside of the
where the diameter D has been reinstated in the latter scale.

275
 Ay=561 lb. By=619 C,=85 Dy =286 + Ey=428
i
276 SHAFT DESIGN [Ch.9 D2"
t GVH
The deflection curve for the vertical plane, Fig. 9.7, is found in the same is 18" —>—10' =p 10">}
manner, by plotting the following computed values of M/EI:
Sy1/21=60x 10-4 in.-"
Aen Bos AGL: MGs Col Hasta Dx, aah (Unity D)
M(in-kip). . . 0 673 5.8 58 5.7 S54 54 2.86 2.86
(M/JEDD*/10* . 0 45.7 39.4 3.75 3.69 3.5 37.4 19.1 29.6
The maximum deflections are, between the bearings and at the overhang, re
spectively 1
a es pect tS Ree pes A gS ak ties
2.4
(i) Vu = 0.2855; and Vey = 0.2924 in. Sp = 0.12784:
! mn.
——F (Unity D)+ :
If the overhung member at E had been a pulley, the amount of the deflectioi 1
within reason, would not be very important, but being a gear, it is importan
--r--
u
The resultant deflection is the vector sum of y, and y,, or
2.4 1.054 = 0.003,
Vmax = (0.337 +,0.29%)*/? DY 4d" ial \
Detlection y |
= 4.33, use 4 in. a
from which D Sym24iD
teIn. |
With the diameter now known, the slopes at the gear meshes can be chec *

an Yey
At the overhang, the approximate slope is the vector sum of qr, = 0.71
qry = 0.62 times the scale, which is 0.12/D*. Thus ‘q Yy=0.285 in.
Yey=0.29 in.
0.12
6 = (0.712 + 0.62?)*/? = 0.000309 in./in. FIGURE 9.7 Slope and Deflection, Vertical Plane. The deflection y,, is theoretically at
4.375*
the same section as Y;x.
compared to a permissible relative slope of, say, 0.0005 (§ 9.9). This means
the mating gear should not be sloping very much in a direction to increase |
angle between the gear axes. center of mass will equal the deflecting forces on the shaft; the shaft with
* Comment on Problem. We note that a very much larger shaft is ne its attached bodies will then vibrate violently, since the centrifugal force
to meet the condition specified for deflection than is needed for the stresses. changes its direction as the shaft turns. This speed is termed the critical
will often happen when gears are involved, especially on relatively long s speed. Above the critical speed, a smooth-running state of equilibrium is
Usually, machine shafts can be made rather compact, with the gears, be again reached when the body is then rotating virtually about its mass center
etc., crowded much closer together. This plan reduces bending moments as (centrifugal forces balance). High-speed turbines often operate above the
as deflections. If the dimensions of this problem were those assumed for the critical speed. Additional critical speeds, harmonics, higher than the first,
design computations, the next step would be a re-examination of the require’ are successively attained, but the amplitudes of the corresponding vibrations
hopefully with a much smaller spacing between parts. Additional detail info:
progressively decrease.
tion, some of which is in succeeding chapters, concerning dimensions of ©
Let Fig. 9.8 represent a shaft with any number of loads (three loads
bearings, gears, and pulley are needed for intelligent decisions at this p
are chosen for illustrative purposes) which deflect the shaft to some position
shown. Then, the lowest, or fundamental, critical speed , is given by"-7:9-81
9.12 VIBRATION AND CRITICAL SPEEDS OF SHAFTS.
=| 8 Wiy1 + Woyo + me
center of mass of a symmetric, rotating body does not coincide wil
center of rotation. This is because: (1) It is impossible from a pra 4 T Wy? + Woyo? + Ways?
viewpoint to get the mass uniformly distributed about the geometric cen In general the critical speed will be
of the body and (2) the shaft on which the body rotates deflects under la
thus moving the center of mass away from the true axis, which pi (i) Ne
_ 307 gE Wy) ] [go = 386 ips?]
through the center line of the bearings. Rotation may begin about 1 al DWP | ;
geometric axis, but at some speed, the centrifugal force of the displace¢ where 2Wy represents the sum of all the Wy terms and = Wy? represents

277
 sr
uP
Y
99.13] SHAFT DESIGN BY CODE 219
Wi We Ws
Their values are governed by what the designer thinks of the load and
FIGURE 9.8 Shaft in Defected Position. This is a diagrammatic arrangement showin; are taken from Table 9.1, where we observe that the minimum value of K,, for a
loads W,, W2, and W, at points along the shaft where the static deflections are respective}; rotating shaft is 1.5, which is intended to care for the reversal of stress during
Yay Yo, and yo. every revolution of the shaft.
For bending or torsion alone, the design equations become

the sum of all the Wy? terms. In both of the foregoing equations, th _KnMe _KnM ie 32K,M

acceleration of gravity g, is in inch units to accord with the units of y Sei icp aa Sy Sa = pe
[SoLipD ROUND]
that is, g, = 386 = (12)(32.2) in/sec.?; but n, rpm. In determining
values of 1, Yo, Vg, etc., the graphical method of § 9.10 is suitable; the d ua K;Tc _ K.T . 16K,T
(m) Sea = Fe Tee © Soa = 7D?’
_ tions are found for the static loads, due to weights of wheels, gears,
[SOLID ROUND]
The static deflections in inches taken from this diagram at the varion
points of application of the static loads W,, W2, Ws, etc., are the pro where D is the diameter of a solid circular shaft. For bending, axial load F, and
values of the y’s. If it is desired to account for the weight of the shaft torsion together, the normal stress used is
divide the shaft into parts, compute the weight of each part, consider KnM F
(n) s
weight of each part as a force acting through its center of gravity; an he aa
proceed as for any group of concentrated loads. When a large numbe:
where « is defined by equation (i), § 7.11, for L/k < 115, and by equation (h),
loads is involved, tabulate loads, deflections, values of Wy, of Wy? to kee
§ 7.11, for L/k >115 when the axial load F is compressive (Le = L); when the
the procedure comprehensible.

TABLE 9.1 VALUES OF K; AND K,

9.13 SHAFT DESIGN BY CODE. Several years ago, the ASME publishe
the Code for Design of Transmission Shafting,'®-1 which was withdrawn a Nature of Loading Ke K,
years ago, but which has been widely used for shaft designs of all kinds
codes in general, the results it gives are usually conservative, but there ma Stationary shafts (bending stress not reversed):
situations in which it has failed. The rational approach previously descri Gradually applied . : f ’ i - | 1.0 1.0
Suddenly applied 1.5 to 2.0 1.5 to 2.0
recommended, but we shall outline some of the Code provisions briefly.
The design stresses (probably with cold-drawn shafting in mind) are gi Rotating shafts (bending stress reversed):
Shear }) ra or Ssa = (0.3)(tensile yield strength) Gradually applied or steady | ‘ ¢ alee) 1.0
Suddenly applied minor shocks 1.5 to 2.0 1.0 to 1.5
Deven or
Suddenly applied, heavy shocks . 2.0 to 3.0 1.5 to 3.0
Stress J +g OF Ssq = (0.18)(tensile ultimate strength)
whichever is smaller (maximum shear theory). For a shaft in bending only,
Normal) og or sa = (0.6)(tensile yield strength)
axial load is tensile, as on a propeller shaft of an’airplane engine, « = 1; and the
Design \ or Shearing stress used is
Stress o4 Or Sa = (0.36)(tensile ultimate strength).
whichever is smaller. Allowance for keyway: (0) Ss = ais
Zz:

(k) (sa with keyway) = (0.75)(sa as above without keyway), In accordance with the maximum shear stress theory, equation (8.4), these values
where the keyway is at the section being designed. For combined stresses, of s and s,, with Z and Z’ for a hollow shaft (D,/D = B = internal ‘diameter
design stresses without keyway correspond to a design factor of about 2 on divided by the outside diameter) give
the Code makes further allowances by use of service factors, called shock am
fatigue factors, as follows: ©) D? Ce= payy
te16 aE ee [KT 2 + (Ka + egereneebes
«FD(1 + pean
fis
B?) )’ i

K, = “numerical combined shock and fatigue factor to be applied in y

every case to the computed torsional moment or horsepower’; where F and/or B may be zero, 7, is the Code design stress. Whether the Code is
used or not, when a number of similar designs are to be made, it pays to derive
K,, = ‘“‘numerical combined shock and fatigue factor to be applied in ever}
a formula to fit the situation and reduce it to the simplest form, as in (p); but in
case to the computed bending moment.”
the learning process, the student should avoid such formulas.
278
 Sa ae ta ae Orae

9.14 CLOSURE. Many shafts are supported by three

bearings, which means that the problem or
is statically indeterminate
on strength of materials give methods of
solving such problems. Thed
ping with the economics of a gi ituati
example, if one line shaft supported by
three or ase? seaiinbe Heel
probably would be cheaper to make conse
rvative assumptions as to mo
and design it as though it were determinat
e. The extra cost of ano : Si
shaft may be less than the extra cost of
an elaborate design analy i 4

10. KEYS AND

COUPLINGS

10.1 INTRODUCTION. Gears, pulleys, etc., may be attached to

shafts by force or shrink fits, by one or more of several kinds of keys, by
splined connections, or, occasionally, by some especially devised means.
This chapter will cover briefly some conventional design information on
keys and describe some typical couplings that are used to connect shafts.
Couplings that are readily disconnected are called clutches and are covered
in another chapter.

10.2 DESIGN OF FLAT AND SQUARE KEYS. Most keys are so-
called flat keys or square keys. Flat keys are rectangular in section with the
smaller dimension placed in a radial direction, Fig. 10.1, and they may
or may not be tapered. Square keys have a square section, b x b, Fig. 10.2,
and they may or may not be tapered. When one of these keys is in place,
the hub pushes on its upper half on one side and the shaft on its lower half
on the other side, Fig. 10.1, with the result that a couple acts to tip the key
in its seat. The amount that the key actually tips depends somewhat upon

FIGURE 10.1 Flat Key, Tapered. FIGURE 10.2 Square Key, Tapered.
Taper exaggerated. Stock lengths increase Taper exaggerated. Stock lengths increase
by increments of 2b from a minimum of 45 by increments of 26 from a minimum of 45
to a maximum of 166. to a maximum of 165.
K- b>}
Taper Yn 12" (98) o>
Paperin 12”(1:96) FEO7Y y |
bbl F
shea
281 L z eS L

Nie
282 KEYS AND COUPLINGS [Ch.| §10.4] STRESS CONCENTRATION IN KEYWAYS 283

the fit at the top and bottom, because some of the resisting moment maj dimensions 8, t of the flat or square key in accordance with ASA standards
act on these surfaces. from Table AT 19, solve for L from both (a) and (b), and use the larger
A conventional stress analysis simplifies the situation considerably, an value. Although it is not necessary, the length of hub and the length of
the strength is usually expressed in terms of the torque that is to be tr key are made about the same, but the hub length of course must be at least
mitted through the key. For shear, the resisting area is taken as a plar as much as the needed key length L. Typical hub lengths fall between
area between the hub and shaft, to wit, A = bL, where L in. is the lengt 1.25D and 2.4D, where D is the shaft diameter. |
of key in contact with the hub; the corresponding force is F, = s,A = s,bL If the needed key length is greater than about 2D, consider using two
Fig. 10.3; with a moment arm of D/2, the torque T is F,,D/2, or keys, 180° apart, or Kennedy keys, Fig. 10.10. If the load is other than
smooth, the keys should fit tightly, either by use of taper keys or by
s,bLD clamping the hub onto shaft and key. A taper key can be driven into place,
(a) T in-lb. (or in-kips).
and out too, facilitating disassembly. The matching taper is in the hub.
The stresses induced by driving the taper key into place may be dangerous,
Since one-half of the key is in the hub and ‘one-half in the shaft, ¢
but on the other hand, there results a large frictional force between the
measured at the side of the keyway, the bearing area in each is (¢/2)L.
hub and shaft that aids in transmitting the power. The friction may be high
force F,, shown acting at the surface of the shaft, is not the actual fi
enough that the stresses in the key from the transmitted torque are much
acting on either half of the key. The line of action of the force betwee
below those computed from equations (a) and (b). A tapered key is tapered
throughout its length, but the thickness b, t is that measured at a distance
b from the large end, Figs. 10.1 and 10.2.
The usual material of keys is cold-finished, low-carbon steel (0.2% C
or less), although heat-treated alloy steels are used where needed.

10.3 EXAMPLE—DESIGN OF FLAT KEY. A cast-iron pulley is to be keyed

FIGURE 10.3 Loading on Key, to a 24-in. shaft, made of 1040, and it is to transmit 100 hp at 200 rpm. A flat
key of cold-finished C 1020 is to be used. The drive is expected to be subjected to
quite minor vibrations, so that a design factor of 1.75 appears reasonable. Specify
a suitable length of key.
Solution. From Table AT 7 for C 1020, we find sy = 66 ksi; thus, since the
compressive strength of cast iron and 1040 is greater than that of the key material,
the design stresses are

hub and the key would be a little above, and the force between the s: _ (0.5)(66) _ Y tg OO yy ,
oe 18.85ksi and s, = 17s = 37.7 ksi.
and the key would be a little below, the line of action of F,,. Becaus
inherent inaccuracies in the analysis and the small difference invo transmitted torque is
the moment arm of the force is taken as D/2. Thus, F, = sA = s,tL/2
the torque capacity is given by F,,D/2, or T= ee ‘ £62,00000000), # 31,S00:inlb: or 91/Stnckips,
SotLD From Table AT 19, we choose b = $ and t = % for the 2}-in. shaft. Then, from
(b) T= in-Ib. (or in-kips) [COMPRES equations (a) and (b), we find

In design, the stress in equation (a) may be obtained from the shez L= BE
sD ~ cerOO 625)(2.5) ~ siya
(18.85)(0.D 2.14 in, for shear,
yield strength of the key material; the stress in (b) is taken according to (4)(31.5)
af D
= sad = @7.7)(0.4375)2.5)
4375)(2.5) = 3.06 in. for compression;
weakest of the three parts involved, the shaft, the key, or the hub. —
suggested that the design factor on the yield strength be about 1.5 ft use L = 3 in. with § x %-in. cross section.
smooth load, about 2 to 2.25 for minor shock loading, and up to 4.5
severe shock loads, especially when the loading reverses during opera 10.4 STRESS CONCENTRATION IN KEYWAYS. Cutting a key-
use Sy = Sy- for ductile metals and s,, as before. The usual design way in a shaft reduces its strength and stiffness out of all proportion to
cedure would be to find the shaft diameter, choose the cross-sectioné the amount of material removed. Shallow keyways reduce the strength
 264 KEYS AND COUPLINGS [Ch. 10 FIGURE 10.7 Rollpin. (Courtesy Elastic
Stop Nut Corp., Union, N.J.).
less than deep ones. We notice, in Table AT 13, that the sled-runner
keyway, cut by a circular milling cutter, has somewhat lower fatigue strength
factors K, than the profile keyway, cut by an end mill whose nominal dia-

©
meter is the width of keyway. The larger value of K; is for the ends of ie
keyway, where the change of section is quite abrupt.
Experiments show, as expected, that fillets at the internal corners along FIGURE 10.8 Saddle Key.
the length improve fatigue strength. Inco'°-"! reports impact fatigue test
on 4-in. monel shafts with a keyway for a 3%-in. square key as follo

®
average number of blows (all of the same magnitude) to break a s
without fillets at the root of the keyway, 6722; with +-in. radius fill
12,360. While these tests donot give specific information on the comparati
FIGURE 10.9 Flat Saddle Key.
stress concentration factors, they suggest that fillets at the root of the k
way may be significantly advantageous if the shaft is subjected to repeat

&
impact loads. The optimum value of the fillet radius appears to be about
the depth of the keyway (about t/4). A smaller radius is better than non
The effect may be illustrated by available information for a hollow,
treated shaft. Let r = radius of fillet and d = diameter of the concentr
hole in the shaft; for a particular keyway size, K, = 2.4 for r/d = 0.0 FIGURE 10.10 Kennedy Keys.
and K,; = 1.63 for r/d = 0.08, torsion.:?!

10.5 OTHER TYPES OF KEYS. A number of different kinds

keys are in common use, some of which are described below. J
Gib-head keys are square or flat and tapered, Fig. 10.4, with a head. configuration, it is sometimes used as a shear pin (§ 10.8). Stock sizes,
large end, are available from 0.0625 to 0.7060 in. diameter.
The head should not be left exposed so that there is danger that it
The Rollpin®, Fig. 10.7, one of several patented methods of keying, is
catch on workmen’s clothing, possibly resulting in serious injury. This
driven or pressed into a hole that is small enough to close the slit, assem-
is useful when the small end is inaccessible for tapping it out; the h
permits its easy removal. bled in a radial direction as in Fig. 10.6. The chamfered end makes it easy
to drive in, and the pressure it exerts holds it in place. Observe that the
A es key, Fig. 10.5 and 10.6 may be either straight or tapered (ta’
Rollpin will be effective in a hole made with loose tolerances, a circumstance
of } to + in. : ). It isSi uae a drive fit. Longitudinal bea! at the e
that is conducive to cost reduction. Made of heat-treated C 1095 (other
materials: 420 stainless, beryllium copper), its shear strength is as good as a
solid, mild-steel taper pin.

—_ “ie
Saddle keys are tapered, suitable for light power, and are either hollow,
L——<D Fig. 10.8, with a radius or curvature slightly smaller than the shaft radius, or
FIGURE 10.4 Gib-head Key. flat, Fig. 10.9, in which case they are assembled on a flat on the shaft. As
in Fig. 10.8, they depend entirely on friction to transmit the load; in this
type, the hub can be located in any angular position relative to the shaft.
A tight fit and the consequent friction also improve the load capacity of the
flat saddle key.
Kennedy keys are tapered square keys, with or without gib heads, as-
FIGURE 10.5 sembled as shown in Fig. 10.10. These keys are also said to be tangential
keys. Rectangular keys assembled with the diagonal dimension virtually
in a circumferential direction (Fig. 10.10) are also called tangential keys,
and there are other similar ways of assembly. The advantage of this con-
figuration is a significant increase in capacity.

FIGURE 10.6
 FIGURE 10.11 Woodruff Key. (Courtesy Standard Steel Specialty
Co., Beaver Falls, Pa.).

L/f LA B— | B.
(a) (b)
a ayi( 7 iN FIGURE 10.12 Feather Keys.

TABLE 10.1 sktRenen t ?

WOODRUFF KEY DIMENSIONS TABLE 10.2 NOMINAL DIMENSIONS OF SPLINES??:?]
ASA B 17f-1947. Key extends into hub a distance 4/2. Larger sizes available.
ALL PERMANENT NOT SLIDE SLIDE
DISTANCE NO. STANDARD FITS | FIT UNDER LOAD UNDER LOAD
SUGGESTED NOMINAL HEIGHT OF BELOW |SHEARING OF SIZES,
KEY SHAFT SIZES KEY SIZE KEY CENTER! | AREA SPLINES NOMINAL W h d h d h d
NO.
in. A x B in. Max.C | Max. D| Ein. sq. in. 4 by }-in. from ? | 0.241D | 0.075D | 0.850D | 0.125D | 0.750D
6 in. to 1};2, 23,
204 $s? &x 4 0.203 0.194 & 0.030 24, 3 0.250D | 0.050D | 0.900D | 0.075D | 0.850D | 0.100D | 0.800D
305 4 &x ¢ 0.250 0.240 ty 0.052
405 #7 4x ¢ 0.250 0.240 ab 0.072 10 Same as above,
506 BB ux 2 0.313 0.303 4 0.109 plus by }in.
from 3 to 6in. | 0.156D | 0.045D | 0.910D | 0.070D | 0.860D | 0.095D | 0.810D
507 $-# ux 0.375 0.365 Be 0.129
608 1-1#%; i x1 0.438 0.428 ae 0.178 16 by 4in. from 2
807 13-14 tx f} 0.375 0.365 * 0.198 to 6 in. 0.098D | 0.045D | 0.910D | 0.070D | 0.860D | 0.095D | 0.810D
809 13-12 4x 1} 0.484 0.475 & 0.262

810 1}-1}? 4x It 0.547 0.537 & 0.296

812 14-13 4x 14 0.641 0.631 au 0.356 10.6 SPLINED SHAFTS. Reversing torques and repeatedly applied
1012 14#-23 fs x 14 0.641 0.631 a 0.438 torques are tough on connections like those previously described. Splined
1212 14-24 $x 14 0.641 0.631 « 0.517 shafts make a stronger connection. Straight-sided splines are widely used
1 This dimension given to facilitate layout in drafting-room. in the automotive (Figs. 12.10 and 18.10) and other industries for perma-
nent fits, for fits not intended to slide under load, and for fits that will slide
under load. The nominal dimensions of 4-, 6-, 10-, and 16-spline fittings
The Woodruff key, which fits into a semicylindrical seat in the s are given in Table 10.2, as taken from the SAE standard, '-2! from which the
Fig. 10.11, is widely used in the automotive and machine tool industrit
details of the various fits may be obtained. The nominal size is the major
Since it goes deeper into the shaft, it has less tendency to tip when the diameter D, Table 10.2, and this is the maximum size of the fitting, since
is applied. Note that the construction is such that the key must firs the tolerances given in the standard are negative and the desired fit is ob-
placed in its seat and the hub pressed on. Table 10.1 gives some dimen tained by varying the dimensions on the shaft. In the machine-tool applica-
of selected sizes. Two grades of material are readily available; SAE tions, the practice is to vary the hole dimensions in order to obtain the de-
and heat treated alloy steel. For extra strength, two or more keys may D® sired fit, because there may be both a press fit and a sliding fit on the same
used—usually in a tandem arrangement. ’ shaft, in which case it is economical to grind the shaft to a uniform size.
A feather key is one that allows the hub to move along the shaft ) The machine tool industry, which has need of five classes of fits—free fit,
prevents rotation on the shaft. It is used, for example, to permit mo i sliding fit, push fit, light drive fit, and press fit'°-*]prefers the fit on the
gear into or out of engagement with its mate, or to engage and disen; minor diameter, since the corresponding mating surfaces are the only ones
a jaw clutch. The feather key may be attached to the shaft, Fig. 10.1 that can be ground economically. The design torque capacity T of a splined
or to the hub, Fig. 10.12(b). A force analysis based on certain assump connection with axial sliding is based on a side pressure of 1000 psi;'2-2) or
shows that two feather keys 180° apart are preferable to a single key.
axial force required to move a member along a shaft is about half as (c) T = (SA)rm = (1000)(AL)(r,)(N;) in-lb.,
when there are two or more feather keys as when there is one. where Lin. is the contact length, r,, in. is the mean radius (D + d)/4, and
 =< Internal }
288 KEYS AND COUPLINGS [Ch. 10

N; is the total number of splines. If there is no sliding under load, the sid :
pressure may, of course, be much greater than 1000 psi.
Major Dia.

Base Circle | D
10.7 INVOLUTE SPLINE. Involute splines, Fig. 10.13, are in the fo;
of concentric external and internal gear teeth with a pressure angle d =
FIGURE 10.13 Involute Spline Profile. See also Fig. 13.1 for additional information on
(no undercut with 6 teeth). The splines are specified by the diamet nomenclature. Note that the vertical center line divides internal and external spline forms.
pitch P, (the standard™°?! uses two pitches in a ratio, as 3/6, where tl The dotted outlines are optional. Certain dimensions are: Pitch diameter, D = N;/Pa;
numerator is the pitch, with the denominator always twice the numerato: Tooth thickness, t = 7/(2Pa); Addendum, a = 0.500/Pa; Dedendum, 5; = 0.900/Pa
Standard diametral pitches are: 1, 2.5, 3, 4, 5, 6, 8, 10, 12, 16, 20, 24 through Pg = 12. Dedendum, b = 0.500/Pzq on internal spline for major diameter fits, on
external spline for minor diameter fits.
40, and 48. Many of the rules and much of the nomenclature of gearit
apply, some of which may be reviewed in §§ 13.2-13.4. The pitch diame
D, Fig. 10.13, is obtained from the number of splines NV, and the
metral pitch Pz; D = N,/Pz, as in gearing.
The standard™°-?! provides three classes of fits; sliding fits with clearanc
on all mating surfaces; close fits, which are close on either the major d
FIGURE 10.14 Hob for Splining Shafts.
meter, the minor diameter, or on the sides of the teeth; press fits, whic (Courtesy Barber-Colman Co., Rockford, IIl.).
have interference on either the major diameter, the minor diameter, or
the sides of the teeth. Since the internal splines are cut with standard
broaches, these various fits are obtained by varying the dimensions of the
external splines. a
At a splined section, compute the strength of the shaft correspon
to the minor diameter. The length L of nonsliding splines may be comp’
on the basis of the splines being subjected to the same shearing stre
the shaft when one-fourth of the splines are in contact and when the she
is at the pitch diameter D. Thus the shearing area A is } times half
circumference 7D/2 times the length L, or A = 7DL/8; the correspon
force is F = s,A = s,7DL/8, and the torque T = Fr, forr = D/2,is This value of A, is based on one-fourth of the teeth being under load.
However, failure in compression is not likely.
Involute splines have maximum strength at the minor diameter, where
@ r= (65). needed; they are self-centering and tend to adjust to an even distribution
Equating this torque to the resistance of the shaft, T = s,7D,°/16, fin of the load, and when cut by hobbing, Fig. 10.14, they have a smooth
Surface that does not need grinding for most applications.
sD _sD?L 4 Involute serrations, which are used for permanent fits, are similar to
? involute splines except that the pressure angle is ¢@ = 45°, Fig. 10.13.
Fag 16
or Standard pitches P, are: 10, 16,24, 40, and 48 (the standard"°-3) includes
a denominator twice as large, as 10/20). When the fit is on the major or
L = —Pe IN 0S minor diameter, the nominal value of the addendum and of the dedendum
D2’
1s 0.5/P4, as in Fig. 10.13. For additional detail, see the standards.

where D, = root diameter. The compressive stress may be checked by

the projected contact area of a tooth as AL where h = 0.8/Pq =
minimum tooth height in contact; or theareaincompression A,=ALWN, 10.8 SHEAR PIN. Shear pins or breaking pins are used as couplings,
or in addition to other couplings where, in case of overload, there
is danger
 FIGURE 10.16 Flange Coupling.

FIGURE 10.15 Shear Pin.

cE =
of injury to machines or to material in process. One type, shown in F; FIGURE 10.17
10.15, is designed for a roller-chain sprocket wheel. Without the shear p
the part M would rotate on the part NV, which is keyed to the shaft. Thu
the drive ceases when the shear pin breaks. The breaking stress, in +
to l-in. pins, sizes manufactured by Link-Belt, is taken as 50,000 psi in
shear.

10.9 RIGID COUPLINGS. Shafts are directly connected to one an- FIGURE 10.18 Ribbed Compression Coupling. The
usual smooth sheet-metal covering improves the ap-
other by couplings; clutches, which are readily disconnected, are discuss pearance and makes it safer. (Courtesy Link-Belt Co.,
in Chapter 18. Rigid couplings (Figs. 10.16, 10.17, 10.18) are used when Chicago).
shafts are virtually collinear and when they should remain in a fixed ang:
relation with respect to each other (except for angular deflection). Howeve
nearly true alignment of the axes of the connecting shafts is difficult t
achieve and, after it has been accomplished, it is difficult to mainta
because of settling of foundations, unequal deflection of supports, defi
tion of the shaft under load, temperature changes, wear in bearin
effects of shocks and vibrations. Thus rigid cquplings result in stresses
unknown values, which may sometimes lead to failure, and are most s FIGURE 10.19 Flange Coupling. Observe that bolt
heads are shielded by the flanges.
cessfully used where the shaft is relatively flexible, as a line shaft, and
speed is low.
Each half of the ‘flanged-face coupling, Fig. 10.16, usually called
flange coupling, is keyed to a shaft; the faces are finished normal to
axis, the bolts and holes are accurately finished to give a tight fit; standar
sizes up to 8-in. bore. The flanged-compression coupling (Fig. 10.17
transmits the power via the frictional forces induced by pulling the flan L gi
towards each other over slotted tapered sleeves. This plan eliminates th
expense of cutting keyways; standard sizes up to a 3-in. bore. The ribbe 10.10 EXAMPLE—FLANGE COUPLING. Couplings are designed by the
compression coupling (Fig. 10.18) is a heavy-duty connection; the part manufacturer and, ordinarily, the user’s function is to choose one that serves his
are clamped on a long key that is fitted to both shafts, a plan that helps it Purpose. However, computations will not only provide a lesson in simple stress
aligning. An advantage of this type is that it can be installed and removed analysis, but will also indicate by the size of the factor of safety where experience
without disturbing the shafts; standard sizes up to 7-in. bore. has shown that unaccounted for stresses are involved. A manufacturer’s catalog
Stock flanges may be made of cast iron or steel, or of wrought stee ~ the following dimensions in inches for a flange coupling, Fig. 10.19: d = 3,
other metals may be used for a reason. = 53,L = 43,h = 4, H = 81,¢ = 1%. Let the shaft be made of cold-finished

291
 gl0.11] FLEXIBLE COUPLINGS 293

FIGURE 10.19 (Repeated). cylindrical, ~Dg; the resisting force is ssrDg = s,7(5.375)(1.0625);
s s™™J. D > and, wi
» With
moment arm of r = D/2 = 5.375/2 = 2.6875, the resisting torque is fi
T = Fr = s,n(5.375)(1.0625)(2.6875) = 71 in-kips,
from which s, = 1.47 ksi, and the nominal factor of safety is
33
N= 147 = 22.4, shear of flange.

gk
There' is no danger here
b of failure by pure shear, : even if a fati gue strength
duction factor were included, but this same section may have severe i
C 1035, the bolts and square keys of cold-drawn C 1020, and flanges of as-rolled i oo a on it if the flanges are imperfectly aligned, and they surely
C 1035. There are N, = 4 bolts equally spaced. Let a design factor of N = 3.5_ wi . We might also observe that the bolts will be subjected t i
based on the shearing yield strength cover the effect of stress concentration and i" which we hope will be small. : oe
determine the torque capacity of the shaft in pure torsion. Then, for this torque Let the side of a square
. key be b = 2in., “9 Table AT 19; ; let its le ngth be the
applied to other parts of the connection, compute the nominal factors of safety hub length, L = 4.75 in. The computed factors of safety of the key are
based on yield strengths for each conventional method of failure. 7
— 39.6 _ 39.6 bdL _ 39.6)(0.75)(3)(4.75)
Solution. The yield strengths of the materials are (s, for tension and com- = 2.98
= [SHEAR]
N
Ss 2T (2)(71)
pression; sys = 0.6s,): (3)(4.75)
N ea 55 _ 55tdL _ (55)(0.75)
ee SR emery ey 5ee
[COMPRESSION]
From Table AT 7, cold-drawn C 1020, s, = 66 ksi; sy, = 39.6 ksi; 4T (4)(71) = 2.07.
Sc

From Table AT 7, as-rolled C 1035, 5s, = 55 ksi; sy, = 33 ksi; If the shafts cannot be maintained in good alignment and if the loading induces
From Table AT 10, cold-drawn C 1035, sy = 78 ksi, sys = 46.8 ksi. relatively high stresses, eventual fatigue failure of the shaft becomes more likely.
The computed factor of safety of the flange at the hub suggests that it could with
In computing the nominal stresses, we shall find it convenient to set up
stand repeated bending if the misalignment is small. The nearer the flanges are to
strength equations in terms of the torque, as in the equations (a) and (b) for k
the bearings, the smaller will be the deflection of the shaft at that point and the
because all parts of the connection are subjected to the same torque. For sys; =
smaller the stresses induced in the flanges by this deflection. However, the slope
46.8 ksi and N = 3.5, the torque capacity of the shaft is, Fig. 10.19,
of the shaft at a bearing may be a source of excessive stressing. '
T=s,
ad® _ (46.8)(7)(3)° = 71 in-kips.
16 = (3.5)(16) 1
The bolts may fail by shearing between the flange faces, where the major 10.11 FLEXIBLE COUPLINGS. Since shaft misalignment is in-
diameter of the bolt (h = in.) is assumed to be resisting. For 4 bolts, the a evitable, rigid couplings often lead to fatigue failures, overheating bearings,
offering resistance is 4A, = 4ah?/4 = wh? = x(0.75)?; the corresponding resisting and other troubles, which can be avoided by using flexible couplings. A
force is s;4 = s,7(0.75)?; the moment arm of this resistance is r = Al2 = large variety of flexible couplings has been invented and designed, a selec-
8.25/2 = 4.125 in., Fig. 10.19. Hence the torque T is 7 tion of which are shown in Figs. 10.20-10.27 (pp. 294-295). See Ref. (0.5) for
additional descriptive material. These couplings care for small amounts of
T = Fr = s,n(0.75)7(4.125) = 71 in-kips,
angularity, end play, and axis displacement; they also serve the important
from which we find s, = 9.73 ksi. The factor of safety is functions of absorbing some shock and vibration that may appear on one
shaft and of preventing the occurrence of reversed stresses caused by the
N= ous = 4,07, shear of bolts, friction ignored. shafts deflecting at the coupling. With one half keyed to each shaft, the
feature of a flexible coupling is an intermediate member that is either
The compression area of one bolt ona flange is hg; for 4 bolts, itis A = 4hg =
s<(3.1875), flexible or floating (or sometimes both to some extent), Some couplings
(4)(0.75)(1.0625) = 3.1875; the corresponding resisting force is s-A =
whose moment arm is r = H/2 = 4.125. Hence, the torque is (sAr) need lubrication to prevent overheating, some can tolerate more misalign-
ment, intentional or otherwise, than others without ill consequences, but
T = Fr = s(3.1875)(4.125) = 71 in-kips,
all will benefit by good alignment. That is, if possible, the shafts should be
from which, s, = 5.4 ksi. With the flange strength governing (55 < 66), kept carefully aligned; let the coupling care for unintentional (settling of
floors) and unavoidable misalignment (due to temperature changes, de-
N= ey = 10.2, compression of bolts and flange. flection, bearing wear, etc.). Flexible couplings are not a guarantee against
The flange may shear at the outside hub diameter. The resisting area troubles and recommended maintenance procedures should be followed

292
 mediate elements are four rubber ‘‘biscuits’’ held in
a center piece. Each hub is attached to two opposite
biscuits, allowing angularity, which flex torsionally
and axially. No metal-to-metal contact. Good absor- FIGURE 10.24 Flexsteel Coupling. The spring
ber of vibrations, including torsional. Other couplings member, which transmits the load, is inserted inta
with nonmetallic intermediate elements are available. grooves in the hubs which widen toward the center,
(Courtesy Morse Chain Co., Detroit). When the load is applied, the spring is bent along the
arc of these grooves, whose curvature is such that the
spring stresses are safe. (Courtesy The Falk Corp.,
FIGURE 10.21 Roller Chain Coupling. The two Milwaukee).
opposing hubs are made with integral sprockets over
which a double roller chain is fitted. The drive is
through the chain. Flexibility is obtained by lateral
play in the fit of the chain over the sprocket teeth —
(no circumferential play). Relatively inexpensive type. —
Preferably installed with a casing to contain grease
lubricant. Similar couplings use silent chain. (Courtesy ~
Diamond Chain Co., Indianapolis).

FIGURE 10.25 Gear Type Coupling. The hubs have

integral external gear teeth, perhaps crowned (§13.12),
that mesh with internal teeth in the casing through
360° as in a splined connection. Flexibility is obtained
Patented Renewable Laminated Steel by play between the teeth. (Courtesy American Flex-
Bushing Flexpins ible Coupling Co., Erie, Pa.).
Side Seal’ tae a !
Bearing [§ Spring
Strips ; “

FIGURE 10.26 Flexible Coupling (Oldham Prin-

ciple). In the assembled form, the tongues on the
central piece engage the grooves on the end pieces
which are attached to the shafts to be connected. The
FIGURE 10.22 Flexpin Coupling. action of the parts is similar to that described in Fig.
The two flanges of the coupling are 10.27. (Courtesy W. A. Jones Foundry and Machine
connected through laminated steel Company, Chicago).
pins that are relatively flexible. In the
right flange, the pins are held in place
by the spring retaining ring. Freedom
of the pins to slide in the left flange
permits the coupling to care for some
endwise motion, as well as for angular FIGURE 10.27 The American Flex-
misalignment. (Courtesy Smith & ible Coupling. Operates on Oldham
Serrell, Newark, N.J.). principle. In the position shown, the
left member can slide up and down,
the right member can slide to and
FIGURE 10.23 Flexible Disk fro on the square floating center
Coupling. An all metal coupling with member. The combined action, when
the intermediate flexible elements be- the parts are assembled, produces a
ing thin steel disks. The shaft hubs flexible connection that will care for
and center pieces are connected at misalignment. The center member is
different points to the disks. Bending hollow, the cavity being filled with
of the disks permits misalignment. lubricant, which reaches the surface
Two sets of disks permit parallel of the nonmetallic strips through
misalignment. No play or backlash porous reeds in the block and felt
in the connection; no relative motion pads in the bearing strip. (Courtesy
or lubrication. (Courtesy Thomas American Flexible Coupling Co.,
Flexible Coupling Co., Warren, Pa.). Erie, Pa.).
296 KEYS AND COUPLINGS [Ch. 10

Manufacturers’ catalogs give rules for selecting a stock coupling for a— FIGURE 10.31 Overrunning Clutch. If the shaft C
drives in a counterclockwise direction, the rollers R are
particular job. If a special design is necessary, their specialized knowledge —
edged into the space S between A and B, and B is forced
will be helpful—also see Ref. (10.7). to turn. If C turns clockwise, the rollers are not pressed
against B and no drive occurs. The effect is the same if B
turns counterclockwise faster than C turns counter-
10.12 UNIVERSAL JOINTS. A universal joint is used to connect© clockwise.
g
PBX
RB
<4 6 Z
2
shafts whose axes intersect, that is, whose angular misalignment is perma-
nent. Hooke’s coupling is shown in Fig. 10.28. We see that if B, Fig. 10.28,
is held stationary, C may be swivelled about either pin D or E, which pro-
perty makes the universal-joint action possible. If the connected shafts
are at an angle, Hooke’s coupling does not transmit a constant velocity ratio,
as explained in books on engineering kinematics.“%-2! However, if two
universal joints are used with the driving and driven yokes in the same
plane, Fig. 10.29, the initial and last shafts, if they are in the same plane and FIGURE 10.32 Sprag-type Overrunning Clutch. The
sprags R, which are kept in contact with the members by
make the same angle with the intermediate part, will turn at a constant ve the pressure spring, wedge tight for one direction of drive
city ratio of one. Rather than the short coupling of Fig. 10.29, a long dri and release for the other direction. (Courtesy Formsprag
shaft may generally be safely inserted between the universal joints, as in Co., Van Dyke, Mich.).
automotive applications. For a single Hooke’s coupling, the shaft angle
should not be greater than about 15°; less, if the speed is high. 7
Some universal joints do transmit a constant velocity ratio, for example
the Bendix-Weiss universal joint in Fig. 10.30, where the drive is through
steel balls in races that are so shaped that the plane of contact between
balls and races bisects the shaft angle at all times. This position of the pla of contact is the condition for the driven yoke to turn at a constant angular
velocity when the driving yoke has a constant speed.
FIGURE 10.28 Hooke’s Universal Joint.
serve the added metal over the keyway, and
set screws that hold the key tight in its seat 10.13 OVERRUNNING CLUTCH. The overrunning clutch, illus-
prevent axial movement of the hubs. trated in Figs. 10.31 and 10.32, has many uses, among which are: as a
“free-wheeling’’ clutch on automobiles; as a brake on inclined conveyors
to lock the conveyor against unintentional backward motion; as a feed
mechanism and, in general, in place of a pawl-and-ratchet mechanism;
and as a coupling between shafts.
FIGURE 10.29 Double Universal Joint. There are variations in the construction of overrunning clutches, but
in all cases there are some members, such as R in Figs. 10.31 and 10.32,
between the driving and driven members that wedge themselves in one
direction and release in the other direction. Figure 10.32 is an example of
FIGURE 10.30 Constant-velocity Unive
heavy-duty construction for industrial use.
Joint. (Courtesy Bendix Aviation Corp., So
Bend.).

<i 10.14 CLOSURE. There are many elements of machines whose

design and development have become so specialized that usually the best
Procedure is to choose the item from a catalog in accordance with the manu-
facturers’ instructions. The couplings above are, for the most part, illus-
trative of this situation. The feature of the strength equations of this chapter
Is that the strength is expressed as a torque, a form that is often convenient.

297
 Il. JOURNAL AND
PLANE-SURFACE
BEARINGS

11.1 INTRODUCTION. Wherever machine elements move, there

must be bearing surfaces, some of which are lubricated easily, some with
difficulty, some not at all. When the load is light and the motion slight,
one extreme, the designer may be content to specify an oil hole and to
depend upon an operator to apply lubricant intermittently; though, this
practice is becoming rarer. For an intermediate category of load and speed,
the use of a dry lubricant (graphite), grease, porous bearings, synthetic
bearing materials, etc. is quite satisfactory. When the load or speed or both
are high, as often happens in modern machines, the lubrication of surface
bearings, whether by oil, air, or other fluid, must provide a fluid film that -
keeps relatively moving surfaces from touching.
The theoretical differential equation that interrelates many of the
variables in a journal bearing is credited to Osborn Reynolds (1886), who
based his work on experimental data obtained by Beauchamp Tower
(1883, 1885). A form of this equation restricted to incompressible fluids,41-®!
which is still difficult to solve (x is the direction of motion), is

es a
Ox \ pw Ox,
/(h® ap
te @
0z
(h® ap
\p az,
oh
ax
where v is the relative velocity of the sliding surfaces. The pressure gradient
ép/éz in the longitudinal direction (perpendicular to motion) is zero in an
toe
299
 gll.4]) View Mworer wt
300 JOURNAL AND PLANE-SURFACE BEARINGS [Ch. If

infinitely long bearing, the kind for which a solution was first made. The 11.4 VISCOSITY. When thick-film lubrication exists, the frictional
first such solutions suitable for engineering design were obtained graph force that resists relative motion is independent of the nature of the mating
cally,“1-9:11-10] followed a few years ago by computer solutions.{1-111 surfaces and is more influenced by viscosity than by any other variable.
Digital computer solutions have now been obtained for finite-length journal Consider an element of a fluid between two surfaces, Fig. 11.2, one of which
bearings, which allow for end leakage and the pressure gradie M is moving with a constant speed. A layer of fluid adheres to the surface
ép/az,"+-8-11-7) and it is from these solutions that we shall make our com- M and moves relative to the next layer, and so on until the layer in contact
putations—later. with the surface NV adheres to it and remains stationary. Let the bottom
surface of the element E move with speed », let the upper surface, a distance
11.2 TYPES OF JOURNAL BEARINGS. A journal bearing i
composed of two principal parts: the journal (which is the inside cylindrica Vm Um=Constant

part, usually a rotating or an oscillating member) and the bearing or sur- —— v+dv
rounding shell (which may be stationary, as on a line-shaft bearing, of FIGURE 11.2
moving, as on a crankpin). One basis of classification depends upon whethe: A
the bearing surface completely surrounds the journal, a type called a fi
MMMM
bearing, Fig. 11.5, or only partially surrounds it, a type called a pa N Stationary
bearing, Fig. 11.6. A simple type of partial bearing can be used when
load is carried on the top part of the journal and the journal dips int dh away, move with speed v + dv; thus, the difference between these speeds
oil sump at the bottom. When the line of action of the load (bearing is dv. Among his vast contributions to science, Isaac Newton (1642-1727)
action) bisects the arc of a partial bearing, the bearing is said to be centr. enunciated the law that the shearing stress F/A in a fluid is proportional
loaded; when the vector passes to one side of the center, the bearin to the velocity gradient dv/dh;
eccentrically loaded (also called an offset bearing). While we shall confin
our attention to central loading, extensive solutions of equation (a) fo 11.1) F dv pAv
eccentric loading are found in Pinkus and Sternlicht.41-®! (1 . gg wk h? [NEWTONIAN FLUID]

Journal bearings may also be classified as clearance bearings or

where A is the fluid area being sheared, and p is the proportionality constant,
bearings. In clearance bearings, the diameter of the bearing is larger t
called the absolute viscosity, or commonly just viscosity, of the fluid. The
the diameter of the journal. The difference in these diameters is called t
second form of equation (11.1) applies where the velocity gradient dv/dh
diametral clearance cz. The radial clearance c, = c,/2 is the difference
can be taken as a constant. Observe from (11.1) that the frictional force F
the radii of the bearing and the journal. The ratio of the diametral clearance
to shear the fluid increases as either the viscosity » or dv/dh increases.
to the journal diameter c,/D, which is the same as ¢,/r, where r is the journa
Viscosities are commonly given in metric units, poises or centipoises,
radius, is called the clearance ratio. A fitted bearing is one in which the radi
but the practice in bearing design in this country is to use the reyn (named
of the journal and the bearing are the same. Thus, a fitted bearing is neces
after Osborn Reynolds), which is a unit consistent with equation (11.1) in
sarily a partial bearing, whereas clearance bearings may be either full or par:
which we use inches, seconds, and pounds for force. Dimensionally and
tial bearings. Fitted bearings have a place in the scheme of things but they
with these units, we have
will not be covered in this work,“1-1.41-91 id
Fh (Force)(L) Pr _ I\b-sec.
(b) (ins a
11.3. THICK-FILM LUBRICATION. If the film of lubricant be Av (LyL/r) L? in.?
tween sliding surfaces is thick enough that there is no metal-to-metal
where P represents the pound unit, 7 (tau) the dimension of time, and L
contact, lubrication is said to be thick film or fluid film. The rougher the
length. A poise has the units dyne-sec. per sq. cm. and is equivalent to
surfaces, the thicker the film required to separate them, Fig. 11.1. If a film
100 centipoises. Convenient conversion constants are:
gradually gets thinner, first the highest peaks begin to touch; then more and
more metal is in contact as the film thins, until friction becomes excessive ‘ centipoises 9 centipoises
and a bearing failure soon becomes imminent if this manner of operation 6.9 x 10 and
reyn microreyn ’
was not intended. One of the aims in bearing design is to provide a film
thickness whose minimum value hf, is safe. where a microreyn is seen to be one-millionth of a reyn. Some viscosities

Moving Surface4
—> 5 FIGURE 11.1
SOS
VLE
 eee a EE ee wre?
ove JVUNNAL AND PLANE-SUNPTAGCE DEANINNVS [I Tt Bee eee ee

are given in Fig. AF 16. The viscosity varies with pressure,"1-5*] but we 3 11.6 HYDRODYNAMIC LUBRICATION. When the surfaces in
shall assume that it is constant at some average value. relative motion are so oriented that their relative motion causes the oil
If the lubricant is not given in Fig. AF 16, it will probably be necessary to pressure to build up to enough to support the load without metal-to-metal
convert from Saybolt Universal Viscosity (SUV), which is the commercial vis- contact, the lubrication is hydrodynamic. The basic requirement for this to
cometer reading usedinthiscountry,
to absolute viscosity. This conversion is done — occur is that the lubricant enter the bearing via a converging channel,
through another property called the kinematic viscosity, which is the absolute — Fig. 11.4. As seen from the curve DEF, the pressure builds up from ambient
viscosity of the fluid divided by its density, each expressed in the same system at E, which occurs somewhat closer to section D than to
to a maximum
of units. The basic dimensions of kinematic viscosity are L?/7. Because in the cgs
system of units, density p is numerically the same as specific gravity SG, it is
easier to find the kinematic viscosity v from the absolute viscosity in centipoises Z;
FIGURE 11.4 Hydrodynamic Film. The Pressure
(c) y= — = — = 0.227 - ” centistokes,
areas formed by the vectors that represent ve-
locity distribution, typically as shown, are
where 7 is the SUV reading in seconds and all properties are for the same tempera- necessarily equal in order to satisfy the conser-
ture t, The specific gravity of a petroleum oil at any temperature ¢ is given ap- vation of mass law for an incompressible fluid,
proximately by A since the areas are proportional to mass of
Velocity 3
(d) SG: = SGeo — 0.00035(t — 60), flow. The maximum film thickness is 41; the
Distribution B
minimum, fo. The resultant load W does not
where SGgp is the specific gravity at 60°F (about 0.89 to 0.93 for petroleum oils), pass through the point of maximum pressure. a Section

11.5 PETROFF’S EQUATION. If a journal is running in a film-—

lubricated bearing without load (or practically, with a light load and at section F. The velocity distribution must be such as to satisfy the continuity
moderate speed), the journal runs concentric with the bearing, Fig. 11.3, of mass law. Note that the velocity gradient is not constant; the fluid is
and the velocity gradient dv/dh = v/h is constant. Newton’s equation — subjected to pressure.
defining viscosity can be applied. The area being sheared at the journal is — Consider a journal with load W at rest in its bearing in metal-to-metal
aDL sq. in.; the film thickness is equal to the radial clearance, h = c, = _ contact, with the clearance space filled with oil, Fig. 11.5(a). As the journal
4/2; then let the “rubbing” speed (the peripheral speed of the journal) be — begins to rotate clockwise, there is first some rubbing of metal on metal,
Ups in./sec. The frictional torque T, = Fr on the journal, with equation — and the journal climbs upward toward the right, Fig. 11.5(b). Since the oil
(11.1), becomes adheres to the journal’s surface, an oil film is drawn between the surfaces
2.3 with the rotation, after which the journal moves to the left of bearing center
(e) ith eps BAL pu ine io SRI Pees hog O’; this is the equilibrium position, Fig. 11.5(c). Notice the wedge-shaped
h €q/2 Cy channel required by hydrodynamic theory. The pressure will build up until
there is no metal-to-metal contact (if a hydrodynamic bearing has been
[PETROFF’S EQUATION]

where the average film thickness h = c, in a centered journal, v,,, = 2arns, correctly designed). The minimum thickness of oil film is designated h,. A
and the units are consistent; » reyns, Din. = journal diameter, rin. = — little intuitive consideration [especially with equation (11.1) in mind]
journal radius, Lin. = bearing length (axially), c, and cq in.,n, = revo- suggests that (considering one variable at a time): the faster the journal
lutions per second of the journal (rps). Keep in mind that the unit second
is involved in reyns. From the torque above, the frictional horsepower fhp
FIGURE 11.5: Mechanism of Lubrication. The angle of eccentricity ¢ locates ho.
is conveniently computed from hp = 7n/63,000, § 1.14, where n = rpm.
For v,,fpm, the frictional work U; is also U; = Fv, ft-lb./min. and
fhp = Fv,,/33,000.

FIGURE 11.3 Concentric Journal.

Radial
Clearance
304 JOURNAL AND PLANE-SURFACE BEARINGS [Ch. II

rotates, the greater the amount of oil passing through the load-carrying
area and the larger h,; the greater the viscosity, the greater h,; the greater
the pressure, the smaller /,.

11.7 GEOMETRIC RELATIONS FOR A CLEARANCE BEARING,

The line that passes through the centers of the bearing and the journal is
called the line of centers, Fig. 11.6. This line locates the minimum film
thickness h,in, =A, provided that the angular length of the bearing is great
enough to include the point M. If the bearing extends only as far as some FIGURE 11.6 Central Partial Clear-
section x, Fig. 11.6, the minimum film thickness ,,;, is at the trailing end ance Bearing, Running Position. Geo-
of the bearing, and the separation of the circles at M is designated as h,. metric relation for any journal bearing.
The distance e= OO’ between the centers of the journal and the bearin y
called the eccentricity of the journal, is a

(f) e=—-h=c¢-h,
called the angle of eccentricity, or attitude angle, locates the position of the
L’
closest approach h, of the journal circle to the bearing circle, which is the
radial clearance c, is called the eccentricity ratio, « (also called eccentricity minimum film thickness if the bearing extends as far as this.
factor or attitude); that is,

(11.2) pee at es Sa elm sec Wr at: 11.8 LOAD CAPACITY AND FRICTION, JOURNAL BEARINGS.
There is a considerable convenience in working with dimensionless para-
meters in designing bearings. Theory, experiment, and dimensional analyses,
a geometric relation that must be satisfied for a journal clearance bearin
when both journal and bearing are smooth right cylinders with axes
If two of the quantities are known or assumed, the other must be calcula’
parallel, lead us to the conclusions that (¢ used to mean function of)
from equation (11.2).
The film thickness h at any angle @ measured in the direction of rota
(j) —=¢
hy
|=
$s
(=)
.

from the line of centers is found from Fig. 11.6 as follows: = 9(S). [ric. AF 17]
Cc, P \G
(g) h = ag — ab = aO’ + O’g — ab Also, the coefficient of friction variable f{D/cg = fr/c, is
= ecos@ + (r + c,) — V(Ob)? + (Oa)? 2
(k) 7 iif 3 =¢ |= (=) = ¢(S) [ric. AF 18]
=ecos6+r+c¢, — Vr? + e? sin?0. Ca cy P\G
But since e is quite small compared to r, the term e? sin? @ may be dropped where n, rps is the angular speed of the journal, p = W/(LD) = W/(2rL)
with negligible effect; which gives psi, called the unit load and bearing pressure, which is seen to be the bearing
load W divided by the projected area of the journal (use same expression
(h) h =ecos8 +c, =c,(e cos
@ + 1) for partial bearings), f is the coefficient of friction (the fluid frictional re-
The load-carrying length of a bearing measured in a circumferential” sistance F divided by the load W;f = F/W), » reyns is the average viscosity,
and the other symbols have the usual meanings, except that S stands for
bearing subtended by the angle 8, Fig. 11.6, the parameter in the brackets, called the Sommerfeld number, or bearing
characteristic number, dimensionless in a consistent system of units:
D
¢)) arc length = L, = ae = FB;
(11.3) hte (2) 3 he i) [DIMENSIONLESS]
where f is in radians. The length of the bearing in the P \G P \Ca
direction is called simply the length L. The angle ¢, Fig. The relations (j) and (k) have been presented in graphical form, formerly
306 JOURNAL AND PLANE-SURFACE BEARINGS [Ch.|! §11.10] EXAMPLE—FULL BEARING
30 7

for an infinitely long bearing, which is one for which the pressure gradient positioning. For the first computations, let D = 4 in. Let
L/D =1,orL = 4in.;
in the axial direction is zero (no end leakage). The procedure then was to 7 ee -17 fora discussion of considerations here, The toleran
ces aid allowances
correct the results for end (side) leakage by using appropriate factors, 11-15]
s be commercial if possible. Considering first a class RC
5 running fit
§ 3.4, for a 4-in. bearing, we get the following dimensions
Now we have solutions of the hydrodynamic equation (a) that care for end from Table 3.1:
leakage and the accompanying axial pressure gradient. Figures AF 17 Hale * 4.0000in. Shaft ) 4 — 0.003 = 3.9970 in.
and 18 in the Appendix give the relationships of equations (j) and (k) for Limits) 4.0014in. Limits) 4 — 0.0044 = 3.9956 in.
an L/D ratio of unity. Many other charts are given in the original Ref. a - - Beso gees Processes are centered (§ 3.12), give an average
(11.7), but to save space we have compromised by reproducing Tables — of c,= 0.0044 in.; Lec = 0.0022
K in.;eg ¢,/D = cJr = 0 0.0011. Obs
AT 20-23. Since various curves that can be plotted from these data are - hae sas a vellg a the smaller the Bist slope of the ana
not straight lines, it is best not to make linear interpolations in the tables — Pie ing an (Fig. 11.12);); for a closely fitted bearing,
i thisis situation
si i should be
for actual bearings; however, linear interpolation is assumed to be permis-
3 Another important decision is the value of minimum film
sible for pedagogical purposes. If the Raimondi and Boyd charts are not —_—
thickness which is
later in § 11.14; for this design based on the average clearanc
available [or others, Ref. (11.56) for example], one may interpolate graphi- o = 0.001 in; h, needs to be less than c,, a relationship
e use
cally by plotting enough points from the table data to define a particular for
or theOeminimum
ee
that should be checked
manufacturing clearance e ifif t there isi any doubt. Enter Fig.
curve. The use of the various parameters in Tables AT 20-23 will be ex-
i AF 17

plained in examples below and by additional discussion later.

For a bearing with the ratio L/D = 2, one could interpolate half-way ho
cal = ho
oe 0.001
= 0.0022 = 0.455 [ENTER CHART]
between values for L/D = 1 and L/D = oo. For a bearing with L/D = 1.5
a rough approximation would be say an interpolation one-fourth of the —
7 find § = 0.15 [0.16]. The values in brackets are interpolated
in Table AT 20
or comparison. Using this value with the expression for the
difference from L/D =1 to L/D = 0. For L/D > 1, but not exactly — equation
Sommerfeld number
(11.3), (a, = 360/60 = 6rps, p = W/A
1.5 or 2, use L/D = 1 or with relatively small error, the interpolation for c,/r = 1.1
= 807/(4 x 4) = 50.5 psi,
x 10-%) and solving for the viscosity 1, we get
1.5 or 2, whichever is closest. ,
me Sp(c,/r)? “ (0.15)(50.5)(1.1)2(10-°)
Ns 6
= 1,53 x 10-° reyns,

HYDROMATIC BEARINGS. Of the infinite . Z = (1.53 x 10-*)(6.9 x 10°) = 10.55 centipoises,

11.9 OPTIMUM
number of solutions that can be found for a design situation, Kings- which is the desired average oil viscosity at the normal, steady-o
perating 160°
bury"?-16] showed that for a particular supporting arc B there is a certain temperature. The coefficient of friction variable fi or the fore
i =
Fig. AF 18 [Table AT 20] is 3.8 [3.93] = fr/c,, or
eccentricity ratio « (or certain value of h,/c, = 1 — e) that results in a maxi- ~ heriiee ks
mum load capacity and another eccentricity ratio that results in a minimum
Cr
frictional energy loss. The corresponding bearings are called optimum f = 3.8 coe (3.8)(0.0011) = 0.00418.
bearings. Especially in bearings subjected to heavy loads or large frictional The energy loss to friction is
losses, the designer may very well care to work towards an optimum; the
various optimums are defined in the headings of Tables AT 20-23. U; = fWom = fW(nDn) = (0.00418)(807)(» + x 360) = 1270 ft-lb/min.,
vw , ‘ ‘
pein fpm is the peripheral speed of a point on the surface of the journal to
11.10 EXAMPLE—FULL BEARING. Returning to the shaft for which a € energy in ft-Ib. This energy must either be radiated and
convected from
with i
computations were made in Chapter 9, design the journal bearing A, the one ——— or the circulating oil must carry it away and dissipate it
to the sur-
find n = 360 rpm, 30 hp, A, = 581, a ings elsewhere. If calculations are now made for the
the maximum load, § 9.4. From § 9.4, we minimum likely clear-
A, = 561 lb. From § 9.11, a shaft diameter of 4% in. was found on the basis of a. oe os me . vis allennce of 0.003—§ 3.12), using
the same viscosity as
the
deflection. (a) Choose suitable dimensions and find the frictional loss in » It Is found that
: the minimum film &, ° is 1 than, the 0.001 i
in steady state operation of 160°F, assumed above, A and a decision must then be
bearing. (b) For an average oil temperature made as to wheth ; les
value
Hu is satisfactory. The fricti onal loss isi found not to
choose an oil. changeBry h Hae
minimum
inim clearance configuration is probably
Solution. The resultant bearing load is a safe on F duchllichlion stif ev Giles
thing is as assumed and computed above. 4 ir inerail Ost bak:
W = (A,2 + A,?)/? = (581? + 5617)"/? = 807 lb. (b) In Fig. AF 16, find the intersection of the line ¢ = 160°F
and the line of
There are now a number of decisions to be made. The journal diameter can safely # x 10° = 1.53 and note that the point is close t :
SAE 10 (or 10 W). p se to the curve for SAE 10 W; use
be less than the shaft diameter; also this would provide a shoulder for lengthwise
308 JOURNAL AND PLANE-SURFACE BEARINGS | [Ch. !1

11.11 EXAMPLE—OPTIMUM BEARING. What should be the diametral — FIGURE 11.7 Circumferential Pressure Distribu-
tion.{1,17] The curve labeled 0.05Z is a record of
clearance, average oil viscosity, and coefficient of friction, if the bearing in the. ©
pressures around the bearing at a distance 0.05Z from
preceding example is a minimum coefficient-of- friction optimum? What. oil
i the end of the bearing, where L is the axial length of
would be recommended for the 160°F operating temperature? the bearing. Thus 0.5Z represents the distribution of
Solution. From Table AT 20, we find that A./c, = 0.3 for a full bearing,
| pressure at the center of the bearing.
minimum frictional loss, which poses an immediate question because we must ~
change the previous decision on either the film thickness h, or the clearance c,. —
The film thickness decided upon before can be reduced, but if so, extra effort ;
should probably be made to get smooth surfaces and good alignment. In this” q \ * 4M to 0.5L
example, the dilemma is settled by the statement of the problem, which implies”
that we keep A, = 0.001 in. Therefore, c; = h./0.3 = 0.00333 in.; cq = 0.0067 —
----"|

in. (Ans.); ¢,/r = 0.00167; S = 0.078 from Fig. AF 17. Solving for the viscosity—
u from the Sommerfeld number, we get
_ Spler/r)? _ (0.078)(50.5)(1.67)?(10-®)
= 1.82 x 10-° reyns
vs Ns mo 6
Use fr/c, = 2.4 from Fig. AF 18 and get
Ff = (2.4)(0.00167) = 0.004,
compared with the previous answer of 0.00418. With the viscosity of 1.82 micro- —
reyns and 160°F, we locate a point in Fig. AF 16 as before and find it nearly FIGURE 11.8 Longitudinal Pressure Distribu-
midway between SAE 20 and SAE 10; the heavier one will cause both the fric- tion.{11-18] As seen, heavy loads may distort the
tional loss and the film thickness to be greater; the lighter one will result in a low curves considerably from the ideal parabolic form.
frictional loss and a thinner and less safe film thickness. Assuming that there
an adequate safety margin in the assumed h, = 0.001 in., we recommend SAE 10. d
Note than an optimum bearing with a lower /, (same c,) has a lower frictional —
loss than an optimum for a thicker film.

11.12 LUBRICANT FLOW THROUGH BEARING. The hydro-~

dynamic action of relatively moving surfaces has been explained (§ 11.6), uy of Bearing
and Fig. 11.7 shows typical circumferential pressure distributions at differ-
ent sections of a finite-length bearing. The flow in this direction depends on ~4 negative pressures, when conditions are such that they exist, will be only
the pumping action of the journal. For the ideal bearing, it can be computed —p a small amount below ambient pressure. The values of the various para-
from the flow variable q/(re,n,L), obtainable from Tables AT 20-23; that is, 4 meters in Tables AT 20-23 have been determined for a boundary condition
of p = 0 and no negative pressure.
r¢)) Value from table or curves (dimensionless) = [FLow IN] _ There is no side flow (end leakage) in an infinite bearing and no pres-
ren,L ea sure gradient, but there is in bearings of finite length, and therefore a pres-
where g cu. in./sec. is the needed rate of flow into the leading end of the Sure gradient. Measured pressure distributions on a 2.5 x 3.875-in. full
film to conform to hydrodynamic requirements. The pressure-distribution " — bearing, when the oil is fed into one end, is shown in Fig. 11.8. The curves
curves of Fig. 11.7 approximate the results on a } x 1-in. full bearing; the would have had better symmetry if the oil feed had been symmetric, a para-
maximum pressure occurs in the converging portion of the film; and we " bolic form in short bearings. For an estimate of the peak pressure in the
note that the pressure distribution is not symmetric about the line of action A bearing, see the column p/p,,,x in Tables AT 20-23.
of the load F. Another point to observe is the small negative pressure on ~ The side leakage is computed from the flow ratio g,/q, Tables AT 20-23;
the trailing end of the film, Fig. 11.7. According to theory, for the infinite-_
length, 360° bearing, there are negative pressures on the non-load-carrying — (m) Value from table or curve (dimensionless) = =
part of the same order of magnitude as the positive pressures; but of course q
pressures below absolute zero are impossible, and actually the measured which, multiplied into q from (1), gives the side flow gq, cu. in./sec., when
310 JOURNAL AND PLANE-SURFACE BEARINGS [Ch. I] _ 11.15] EXAMPLE—PARTIAL BEARING, WITH TEMPERATURE RISE 311

the oil enters at ambient pressure in the nonload-carrying region, as it should deflection) or distortion from thermal gradients, the larger the minimum
in a simple hydrodynamic bearing (Fig. 11.5). ; film thickness needs to be. Some operating conditions are such that the
load can be carried only if the finest surfaces are used. In the ordinary
commercial situation, thickness is usually such that the smaller particles
11.13 ENERGY INCREASES OF THE OIL. The frictional loss in” of foreign matter can pass through without serious surface damage, and
the bearing causes the temperature of the oil to increase Ar,. Assume that” of course, it should be enough to care for unpredictable variations in the
all the frictional loss goes to increasing the lubricant temperature; then load (§ 11.32).
the amount of energy into the oil is weAt,, where w = pq lb/sec. is th Data on design values of h, are not plentiful, so that in exceptional
mass rate of flow when p is the density (Ib/cu. in.), ¢ is the specific heat of cases in a particular situation, experiments may have to be conducted to
the oil.* determine safe limits. Karelitz™+-19! suggests h, = 0.0001 in. as a minimum
commercial limit, for finely bored small bronze bushings; 4, = 0.00075 in.
(n) Energy into oil = Q = wcAt, = pqcAt, in-lb/sec.
for commercial babbitted bearings. Denison™?:2"! suggests 0.0004 < h, <
with consistent units. For hydrocarbon oils (from petroleum), the specifie 0.0006 in. for 5-in. to 10-in. Diesel-engine bearings, speed 500-1200 rpm.
gravity will be close to 0.83 giving a density ofp ~ 0.03 Ib/cu. in. at usua Norton™?-12] proposes h, = 0.00025D in. as a rough rule, where D is the
temperatures; the specific heat is about c = 0.4 Btu/lb-°F (or 3734 in-Ib/ nominal diameter of the journal. Fuller'+-1) says that for medium-speed
lb-°F). Keeping the units consistent, we have (500-1500 rpm) babbitt bearings on electric motors and generators,
h, may be 0.00075 in.; for large shafts at high speed (1500-3600 rpm),
(0) U, = fW(D"n,) = weAt, = pgcAt, = (0.03)(3734)gAt, = 11294 babbitt bearings, pressure oil feed, 0.003 < h, < 0.005 in.; for automotive
[PETROLEUM] -
and aviation reciprocating engines, bearings with fine surface finish,
where it is assumed that all the oil has a temperature rise Ar,, the overal 0.0001 < h, < 0.0002 in.; but a filter to remove solid particles large enough
average, and where D” is the journal diameter in inches. Now if all the (S 0.0001 in.) to damage the surface is necessary. The surface for better
that enters the film wedge leaves the bearing, easy to arrange in a par grades of bearings may be finished to 32 vin. rms or below.™1:23)
bearing, if new oil at a temperature f,°F is continuously introduced, an
the heat loss to the surroundings per unit of lubricant passing through| 11.15 EXAMPLE—PARTIAL BEARING, WITH TEMPERATURE RISE.
negligible, equation (0) can be used to compute the average temperat A journal (D = 9 in.) turning at 1700 rpm with a load W = 20,000 lb. has a
rise of the oil. length of 9in., L/D = 1. The bearing is a 120° partial bearing, into which oil
The temperature-rise variable pcAt,/p in Tables AT 20-23 is somew enters at an average temperature of 100°F. The diametral clearance is c, = 0.006
different, as the additional assumption here is that “. . . the mean tempe in. (cr = 0.003 in., approximately average for RC 5 fit—Table 3.1). Determine
ture of the lubricant leaving the sides of the bearing g, was equal to (a) a suitable film thickness h., (b) the Sommerfeld number and frictional loss,
average of the inlet and outlet temperatures [of oil through the bear (c) the temperature rise from the temperature rise variable and from equation (0),
(d) the attitude angle ¢, (e) the grade of oil to be used.
ing].”’""1-7] On this basis,
(Note. Principal answers are chart values of the variables; answers in
pcAt, brackets [], from linear interpolations in the table for the particular h.|cr.)
(p) Value from table (dimensionless) = pe Solution. (a) From § 11.14, A. = 0.00025D = (0.00025) (9) = 0.00225 in.
(Norton’s rule). Fuller (§ 11.14) suggests 0.003 to 0.004 in. for pressure fed; it
where p = 0.03 Ib/cu. in. and c = 3734 in-lb/lb-°F as given above has not at this time been determined that the feed must be pressurized. Use
consistent units—and where Ar, is the temperature rise of the circumferer h, = 0.0025 in. for the first calculations.
from ( (b) For A, = 0.0025, enter Fig. AF 17 with h./c, = 0.0025/0.003 = 0.833
tial flow (q — g,) and is larger than the average At, computed
and find S = 1.2 [1.39]. With S = 1.2, enter Fig. AF 18 and find
therefore, it is a more conservative prediction.
fe =8.7[9.8] or f= (8.7)(0.000667) = 0.0058,
11.14 PERMISSIBLE MINIMUM FILM THICKNESS. _ The smallest where c,/r = 0.003/4.5 = 0.000667. The peripheral speed of the journal is
permissible minimum film thickness is analogous to design stress in tha 9
is a quantity that depends heavily on what experience has shown to be sate. ee eee (|)«700) = 4000 fpm, or 66.7 fps.
The rougher the surface and the greater the misalignment (and shaft
U; = fWv, = (0.0058)(20,000)(66.7) = 7740 ft-lb/sec.,
*Don’t confuse specific heat c, without subscript, with c, or c,. or 92,900 in-lb/sec., or 7740/550 = 14.1 hp.
 pmax. X10~* psi
Min. Film Thick, h.x 10-? in.

Max. Film Press.,

312 JOURNAL AND PLANE-SURFACE BEARINGS [Ch. 1/1

(c) The temperature-rise variable for S = 1.2 [h./c,; = 0.833] is 37 [41.6] = 4

pcAto/p. With p = 20,000/81 = 247 psi, j
FIGURE 11.9 Clearance ys. ho and pmax.
_ 37p _ _(7)(247) Basic data: D = 6in.; L = Gin.; nm, = 60
pc (0.033734) rps; SAE 20 oil. By permission; after Rai-
mondi and Boyd.{11.7!
The total oil flow should not be less than g from the flow variable, which is 4
ql(rensL) = 3.14 [3.14], or, for n, = 1700/60 = 28.33 rps, -005 010
q = 3.14 re,n.L = (3.14)(4.5)(0.003)(28.33)(9) = 10.8 cu. in./sec. Radial Clearance, c,
Using equation (0), we have the over-all average temperature rise
parameters over which there is a control, for example, the clearance or
~ 112g ~ (112)00. 4 clearance ratio. This can be done by assuming different values of, say, c,
which is fairly close to the previously computed 81.6° because the fractional part ~ and computing various items as suggested by Figs. 11.9 and 11.10. In Fig.
qs leaving the sides is small; g./qg = 0.14 [0.134]. a 11.9, observe that for a particular load, the minimum film thickness comes
These calculations indicate that if there is a plentiful supply of oil at the leading — to a maximum, but at different values of c, or c,/r for different loads. Also,
edge at a temperature of 100°F (given), the outlet temperature of the circumfer- — the peak pressure in the film, dotted curves in Fig. 11.9, is sometimes little
ential flow is about 181.6°F. Although it is not always possible to achieve, a
og
affected by clearance (p = 200), sometimes materially affected (at high
maximum oil temperature below 180°F is preferable (§§ 11.20, 11.28). 4 loading). For power loss, Fig. 11.10, an increase in clearance sometimes
(d) The attitude angle ¢, locating h., is 61° by chart, 63° by interpolation on the | results in a decrease, sometimes an increase; the loss is greater for the smaller
table. Note that h, occurs just a little beyond the end of the 120° bearing (the end
films h,. As would be expected, the amount of lubricant flow, Fig. 11.10,
is at 60° for central loading); therefore actual Amin is at the trailing end of the —
bearing. This bearing has a relatively low eccentricity ratioe = 1 — h./c, = 0.16
increases with clearance, and decreases as the load increases. The tem-
and it may well be decided that other proportions shall be tried (0.3 <e <0 perature rise as the oil passes through (Fig. 11.10) decreases with a greater
would be preferable, other things being equal, which they never are). q flow (and larger c,) up to a point, but significantly, is little affected after this
(e) The usual assumption is that the average temperature is (ft; + f.)/2, or q point is reached. (The curves of Figs. 11.9 and 11.10 are drawn from data
146°F in this case. From Sommerfeld number S = 1.2, we have : obtained for an ideal bearing with the negative pressures, but the general
conclusions for this bearing should be valid for the tables and curves given
_ pS(er/r)? _ (247)(1.2)(0.667)%(10-*)
= 4.65 x 10-° reyns, in the Appendix.)
SENT cag TEN: 28.33
Although the limits of the clearance and clearance ratio are largely
or 4.65 microreyns. From Fig. AF 16, the point for 146°F and 4.65 falls between determined by manufacturing criteria, one may come closer to some
SAE 30 and SAE 40, closer to 30; use SAE 30. The lower viscosity results in a
lower frictional loss, so that less heat is generated, and the temperature of the oil
film tends downward; this direction of change tends to increase viscosity, with — FIGURE 11.10 Clearance vs. fhp, g, and Film Temperature. Basic data as in Fig. 11.9.
the result that the oil film will not become as thin as one might think. (By permission; after Raimondi and Boyd.!41-7!)
Since this bearing is operating at other than the best conditions, several other
trials should be made before arriving at a decision. Considering the size and Joad- _ | Temperature | |Power Loss|
ing specified, it is evidently a part of a large and expensive machine. If the clearance—
is increased, more oil will flow through and the temperature rise will be less. If the ©
oil is supplied under pressure (which is likely for such a large bearing, perhaps —
pressurized to avoid excessive metal-to-metal friction at start-up), more oil will ij
flow through and the temperatures will again be lower. Lower temperature will

Aver. Film Temp., °F

permit the use of a lighter oil. The frictional loss fhp= 14.1, while not likely | 4

Power Loss, hp
to be a significant percentage of the total power involved, is high enough to —
make a minimum-friction optimum bearing intriguing enough to investigate.

g
The interested scholar may wish to try out some of these ideas.

11.16 CLEARANCE RATIO. _ In the design of an actual bearing of any —

importance, one should probably investigate the effects of changing some ~

100
0 005 -010 01S
 TABLE 11.1 TYPICAL DIAMETRAL
CLEARANCES, INCHES!+-24,11.251
3 1 2 34 54
Up to shaft diameter of >

Automotive crankshaft FIGURE 11.11 L/D ys. h,. Constant clearance ratio. (By
0.0015 | 0.0025 permission; after Raimondi and Boyd.!11-7!)
Babbitt lined bearing .
Cadmium silver copper . 0.002 0.003
Copper lead cS 3 3 0.0025 | 0.0035

Precision spindle, hardened,

ground, lapped into bronze | 0.00025 0.00075 | 0.0015 | 0.0025 | 0.0035
bushing; vn <500fpm, p < to to to to to
500 psi; 8-16 win. rms . . | 0.00075 | 0.0015 0.0025 | 0.0035 | 0.005

Precision spindle, hardened,

ground, lapped into bronze | 0.0005 0.001 0.002 0.003 | 0.0045
bushing; vm > 500 fpm,p > to to to to to FIGURE 11.12
500 psi; 8-16 win. rms . - | 0.001 0.002 0.003 0.0045 | 0.0065

Electric motor or generator,

ground journal in broached | 0.0005 0.001 0.0015 | 0.002 0.003
bronze or babbitt bushing; to to to to to
16-32 pin. rms. . . | 0.0015 0.002 0.0035 | 0.004 0.006

General machine practice, FIGURE 11.13 Self-aligning Bearings.

turned or cold-rolled journal | 0.002 0.0025 0.003 | 0.004 0.005
in reamed bronze or babbitt to to to to to Spherical
bushing; 32-63 pin.rms. .- 0.004 0.0045 0.005 | 0.007 0.008 Seat

Rough machine practice, of oil flow (hydrodynamically pumped) through the bearing decreases. As
turned or cold-rolled journal | 0.003 0.005 0.008 | 0.011 0.014
L/D decreases, the eccentricity ratio increases (with decrease of film thick-
in poured babbitt bearing; to to to to to
: 0.006 0.009 0.012 | 0.016 0.020 ness) and the approach of the journal surface to the bearing surface soon
63-152 pin. rms..
becomes dangerously close, Fig. 11.11, suggesting a revision in some of
the design assumptions. Fuller") gives the following L/D values as typi-
cal of practice: turbogenerators, 0.8-1.5; gasoline and Diesel engines, main
desired optimum condition with little or no extra cost if one knows what and crankpin, 0.3-0.8; generators and motors, 1.2-2.0; shafts with self-
optimum he desires. A long-standing rule of thumb is a clearance ratio aligning bearings, 3-4; machine tools, 2-4; railroad, 1.2-1.8.
0.001, but evidently this value will be the “‘best”’ value only coincidenta Longer bearings (than L/D ~ 1) may be used for their help in maintain-
The minimum permissible value of c,/r, as for h,, is dependent in some ing alignments (light loads). On the other hand, misalignment may be the
gree on the bearing surfaces. Table 11.1 may be helpful as a guide cause of bearing failure. In Fig. 11.12, for the indicated average h,, a journal
practice. Slope @ at the bearing as shown means the break down of the film at B,
metal-to-metal contact, excessive heating that starts locally and spreads
11.17 LENGTH/DIAMETER RATIO. Needs™+:15! concluded some: until failure occurs. Brown and Sharpe limits the shaft slope at a roller or
time ago that L/D ~ 1, say 0.8 to 1.3, was a good compromise for the gen journal bearing to 0.0005 in./in. (= @rad.). A logical conclusion from
eral case of hydrodynamic bearings, and nothing has happened to cha Fig. 11.12 is that the longer the bearing, the more likely that a given slope
this conclusion. Thus, ratios of L/D > 1.5 are less common today th @ results in a film breakdown, but for a given load, speed, and diameter,
formerly. However, bearings with L/D < 1 are often used for a reason, the film thickness h, increases as L increases, Fig. 11.11, and it happens that
say, the need for compactness as in a V-8 automotive engine. For a parti-. the maximum misalignment for a particular loading can be tolerated when
cular diameter and clearance with a certain load and journal speed, it can” the L/D ~ 1.[1-7) In extreme cases, or wherever it is economically feasible,
be said that as L/D decreases: the minimum film thickness decreases, use self-aligning bearings. At the left-hand bearing, Fig. 11.13, the slope
perhaps rapidly below L/D = 1, Fig. 11.11; the film temperature increases, of the bearing adjusts itself well to the slope of the shaft; at an inner bearing,
perhaps rapidly for L/ D below 0.5; the power loss decreases; and the amount as at B, there could still be trouble if a long bearing were located at such
a point as B where the slope is zero.
314
 g11.18] HEAT DISSIPATED FROM A BEARING 317

take several hours for the bearing to arrive at steady-state operation. Even
then, all parts of the bearing surface are not at the same temperature (the
FIGURE 11.14 Ring-oiled Bearing. As the shaft region in which most of the heat originates is in the vicinity of the maximum
rotates, the ring brings oil to the top of the journal. — pressure and minimum film thickness), and conduction will result in a tem-
The lubrication is expected to be effective for a dis-_ perature gradient along the shaft. Fuller''*-*) has demonstrated that the
tance of about 4 in. on each side of a ring. Ring dia- loss due to this conduction is small, so that it can be assumed that the
meters are about 1.5 to 2 times the journal diameter, frictional energy is radiated and convected from the bearing as heat.
(Courtesy Link-Belt Co., Chicago). 4
Nevertheless, so many uncertainties remain that the computations are ap-
proximate; yet, if natural dissipation of heat is to be depended on, an effort
11.18 HEAT DISSIPATED FROM A BEARING. Many bearings must be made to estimate a steady-state operating temperature. Let the
are made with an integral supply of lubricant, as for example a ring-oiled, heat loss Q be given by*
Fig. 11.14, or chain-oiled or collar-oiled, Fig. 11.15, bearing, in which the:
is a local sump for lubricant storage. This type of bearing can be designe (11.4) Q =h,,A,At, ft-lb/min.,
to deliver adequate oil for hydrodynamic needs, but it must be possible for
where h,, is the heat-transfer coefficient for the bearing surface in ft-lb/
such a bearing to dissipate all the frictional loss to the surroundings without
min-sq. in-°F, A, sq. in. is the effective area from which heat is lost, and
an excessive temperature rise. For example, if a check is made (by the pro-
At,°F is the temperature rise of the bearing surface above the ambient
cedure explained below) on the partial bearing of § 11.15, it will be found
temperature. An old rule of thumb"?:12! says to allow a heat loss in still
that a very large temperature rise is necessary to dissipate the large amount
air of 2 Btu/hr-sq.ft-°F, which, converted to the units of h,, in (11.4), is
of frictional loss. In this case, the normal arrangement is to circulate the
oil to an external heat exchanger. Observe in Petroff’s equation (e) that the h,, = 0.18 ft-lb/min-sq. in.-°F in still air.
The heat loss to moving air is greater than this, up to ten times greater
according to some reports in the literature. Karelitz'!1-1°) found

h,, = 0.516 ft-lb/min-sq. in.-°F at 500 fpm

FIGURE 11.15 Collar-oiled Bearing. (Courtesy speed of the air. The average temperature rise of the oil Ar,, will be greater
Weller Mfg. Co., Chicago).
than that At, of the bearing surface, how much greater depends on the
method of lubrication and the construction of the bearing. The following
approximations, adapted from Fuller,"+-1) may be used in connection with
equation (11.4) for temperature ranges likely to be reasonable:

Oil-ring bearings, still air, Ato, = 2 At,

Oil-bath bearings, still air, At,g = 1.3 At,
Waste-packed bearings, still air, At,, = 2.5 At.
Hence, as the size increases (implying a larger load too), and as the speed
increases with a particular load, a point is reached where natural heat loss,~ For the ring-oiled bearing in 500 fpm air, Af,, in terms of At, is some 15-
the simplest arrangement, will not be sufficient. In the vicinity of this point, - 20% greater than in still air.
excessive frictional heat can be got rid of by blowing air across the bearing What area A, to use is also questionable. For heavy construction,
or by circulating a coolant through cooling coils in the bearing (Fig. 11.16). as ring-oiled pedestal bearings, Norton"+-12] recommends A, = 25DL
Consider the case with the oil sump integral with the bearing. It may Sq. in., where DL is the projected area of the bearing. For light construction,
A, = 6DL will be safer. Fuller™!-1 recommends an average A, = 12.5DL
for simple pillow-block bearings, and-A, = 20DL sq. in. for pillow blocks
FIGURE 11.16 Water-cooled-pillow Block. This with separate shells.
style made in sizes from 1} in. to 3% in. (Courtesy ~
Dodge Mfg. Corp., Mishawaka, Ind.). ‘
* Don’t confuse h,, with the film thickness h,; h is commonly used in the sense of (11.4)
in heat-transfer circles.
318 JOURNAL AND PLANE-SURFACE BEARINGS [Ch. II gl1.21] OIL FLOW FOR PRESSURE FEED 319

Considering the basic laws of heat transfer, we are reminded of the Solution. Assume the average oil temperature as 150°F. From Fig. AF 16,
Stefan-Boltzmann law,"!-27! according to which radiated heat is propor- find » = 1.82 x 10-® for SAE 10 W. Then
tional to the absolute temperature to the fourth power. By compromises, ga hh oi “3 (1.82)(10-*)(6)
and considering that’ the temperatures do not vary widely, we arrive at = 0.179,
Dp \G (50.5)(1.12)(10-*)
(q) h, = 0.108 ft-Ib/min-sq. in.-°F 4 from which, fr/c, = 4.3 from Fig. AF 18.

as the unit rate of radiated heat (surface coefficient of radiation). Although Sf = (4.3)(0.0011) = 0.00473,
no directly applicable equations have been developed for the convected U; = fWv = (0.00473)(807)(377) = 1440 ft-lb/min.,
heat from bearings, the expression for round pipe exposed to a moving ex- the rate at which the lubricant and bearing gain the energy of the frictional loss.
ternal fluid™+-?8! is helpful. To compute the rate of dissipation, assume that the nearby pulley results in an
effective air speed over the bearing of v, = 80 fpm, for which
h,D Dpv\°®
(r) = 0.24 (—)
k ro he = 0.017 58 _= (0.017)(80%*)
= 0.135;
where h, is the surface coefficient for convection, D is the pipe diame hor = h, + h, = 0.135 + 0.108 = 0.243;
k, p, and w apply to the surrounding fluid and are respectively the cond take Atoa = 2 At,, or Afy = (150 — 90)/2 = 30.
tivity, density, and viscosity; the symbol groups are dimensionless an
Dpv/w is the Reynolds number. When the surrounding medium is air, Q =h,,-A,At, = (0.243)(200)(30) = 1460 ft-lb/min.,
equation reduces toh, = Cv,°-8/ D°-*, where Cisa function of the prope
compared with U, = 1440 ft-lb/min., the kind of agreement that gladdens
of air and experiment. From meager experimental data, we choose C the heart of any iterator. If these numbers had been significantly different,
0.017; or another oil temperature would have been assumed and the calculations repeated.
VA8 The conclusion is that the bearing as designed in § 11.10 will operate in a normal
(s) h, = 0.017 D°:4
ft-lb/min-sq. in.-°F, environment without overheating. From Fig. AF 17 and S = 0.179, we get
h./c, = 0.5, or ho = 0.5c, = 0.0011 in., somewhat larger and a little safer than
the h, = 0.001 assumed in § 11.10.
the unit rate at which heat is convected, where v, fpm is the air speed and
Be sure to distinguish between the Ar, of §§ 11.13 and 11.15 and Ar,, of this
D in. is the nominal bearing diameter. The total coefficient is then h,, =
article; Af, is the temperature rise of the oil as it passes through the bearing from
h, + h,, the value to be used in equation (11.4). If the speed of flow of any entering temperature ¢,; Afoa is the difference between the average oil tem-
perature in the bearing and the ambient temperature; which may be, and is in
this example, quite different from At,. You may compute Ar, from the temperature
assumed to have an air flow at 60 to 100 fpm. 7 variable pc At./p.
Computing an operating temperature of the bearing is a process
iteration. The basic procedure is to assume an average oil temperatur
11.20 OPERATING TEMPERATURES. Conventional design values
(given the bearing dimensions, oil, etc.), compute the frictional loss
of the oil-film temperatures are 140° to 160°F, or less. At higher tempera-
and compute the heat loss Q; if U,; = Q, the assumed temperature
tures, the oil oxidizes, more rapidly above 200°F, and temperatures of the
assumed to be the operating temperature. If U, ¢ Q, try another oil
order of 250°F would not be tolerated in industrial equipment unless there
temperature; iterate until the desired agreement is achieved.
is no reasonable alternative. In automotive practice, oil temperatures of
some 350°F occur;"!-2] thus, oxidation is a significant cause of deterioration
in this application, Some bearing materials will lose considerable fatigue
11.19 EXAMPLE—STEADY-STATE TEMPERATURE. Return to the
and yield strength at the higher temperatures.
bearing of the example of § 11.10 from a different point of view. Use SAE 10
and estimate the average oil temperature during operation for an ambien
temperature of 90°F. The data are: W = 807lb., Dx L =4 x 4in.,p =5
psi, c,/r = c,/D = 0.0011, n, = 6rps, peripheral speed v = 377 fpm. Let 11.21 OTL FLOW FOR PRESSURE FEED. Introducing the lubri-
bearing be a pillow-block type of medium mass for which the radiating area iS Cant into the bearing under pressure will serve to increase the flow. The type
taken as 12.5DL = (12.5)(4)(4) = 200 sq. in. of bearing in mind here is hydrodynamic with the oil entering the bearing,
 320 JOURNAL AND PLANE-SURFACE BEARINGS [Ch. We §11.22] FRICTION LOSS IN CAP 321

as usual, in a region of negligible pressure. Since the inlet pressure little defined; « = 1 —h,/c, = 1 — 2h,/cg; L is strictly the total axial load-
affects the amount of the flow through the minimum film area, most of the” carrying length, Fig. 11.18. To be conservative for temperature calculations
increased flow becomes end leakage. Even so, the greater flow results in the designer may wish to assume an actual flow smaller than the amount
reducing substantially the average temperature of the bearing. Equations computed from (t) or (u), say 70 % to 75 % of theoretical values [Ref. (11.24)].
for computing the oil flow under pressure are derived in several Also, since it depends on the cube of the radial clearance c,3, the amount of
works!11.2.11.4,11.5,11.13] from a free body of a particle of lubricant and the
use of Newton’s law of viscous flow, equation (11.1). The amount of flow Grooved
for a particular clearance, load, and oil depends upon the details of how ers

the oil is handled. For example, an oil hole at the mid-point of the cap may y
be the whole system, in which case, the axial flow in a 360° bearing as given
by Shaw and Macks"!-4) is 4

CrP 2
(t) q= (tan =a + 1.5e?) cu. in./sec., FIGURE 11.18 Circumferential Groove. If solid matter gets
3p into the oil and scores the bearing around the circumference, the
distribution of pressure may be affected in a manner similar to
where the units are consistent (u reyns), p, psi is the inlet gage pressu
that pictured. L oD
and the other symbols have the usual meanings (for conversion to gallons 2 2
231 cu. in./gal.). The size of hole, not included in equation (t), also affe
the rate of flow.'!!-24! If, in addition to the hole, there is a longitudi Circumferential
groove, as in Fig. 11.17, the flow will be two to three times greater thai a
given by (t), 1-241

VDSPM eee
g

Oil Hole FIGURE 11.17 Grooved Bearing. This is the half of the bearing flow is critical to variations of this dimension; the minimum flow depends
er not carrying the load. Such an axial groove is helpful also in dis- on the minimum actual c, obtained in interchangeable manufacture.
tributing the oil in oil-ring bearings. The design of a bearing with a circumferential groove must be on the
Die eres nes
WYLLLLLLLEELE basis of a bearing length of L/2, Fig. 11.18. Although there is no reason to
expect that significant bearing characteristics (h,, «, f) would be unchanged
when the oil supply is pressurized, we can safely decide that oil supply
In general, for hydrodynamic action, oil grooves are not used in under pressure will not cause the bearing to be less safe in any respect.
load-carrying area, because any marked discontinuity in the surface h
causes the film pressure to drop to virtually zero gage. For example,
circumferential groove shown in Fig. 11.18 makes the bearing into t 11.22 FRICTION LOSS IN CAP. Since the pressure gradients in
bearings with a pressure distribution something like that shown by the unloaded part are trivial, Newton’s equation (11.1) for viscosity may be
solid curves A, whereas without the groove, the pressure would be more used to estimate the frictional loss in the unloaded part. For a cap arc
like that shown by the dotted curve B. As you can see, for a given total load between 120° and 180°, the average thickness of oil film may be estimated
with a particular L x D, the peak pressure in the grooved bearing will be from*
greater (h, less) than that in the ungrooved bearing. Nevertheless, circum=
ferential grooving is common with pressure-fed oil, especially in the bearings h 2

of automotive engines, because more oil of a particular viscosity flov (v) hay = ¢(1 + 0.74 €) = ¢, lt + 0.74 (1 = =) in.,
through a particular clearance providing greater cooling. The theoretic: Cc;
axial flow through a 360° bearing with a circumferential groove is"1-*! which is the value of h to be used in (11.1). To illustrate, let « = 0.49,

(u) ie 2are,3Cy Pi
(1 + 1.5e?) cu. in./sec. * More information on the assumptions made to obtain this equation is given in the
3yuL Previous edition of this text; the path of the center of the journal in the bearing as the
load is increased from zero is assumed to be semicircular, as it nearly is, as may be seen
where the units are consistent (« reyns) and have the meanings previousl ‘ for the insert in the lower right-hand corner of Fig. AF 17 for B = 360°.
 3
322 JOURNAL AND PLANE-SURFACE BEARINGS [Ch. II —

a
Coefficient of Friction
3
c, = 0.00392 in., » = 1.52 x 10-® reyns, peripheral journal speed v = 3
4000 fpm or 800 ips, Dx L = 9 x 10 in.

o
hay = C(1 + 0.74 <?) = 0.00392 [1 + 0.74 (0.49)?] = 0.00462 in. FIGURE 11.20 Values of f versus Zn/p for
a Particular Full Bearing.{!!-32] With
Z se

o
ey

o
o
=
The area of lubricant being sheared in a 180° cap is A = wDL/2 =141.2 in. smoother bearing surfaces, the minimum f
Then from (11.1) P would occur further to the left.
0.000
0 100 200 WO 400 500 600 700
wAvp, (1.52 x 10-8)(141.2)(800) Zn
F= = 37.2 lb.
aN 0.00462 Pp

U, = Fom = (37.2)(4000) = 148,800 ft-Ib/min., surfaces more, the minimum f may occur at Zn/p < 1 (corresponding to a
typical Sommerfeld number of S < 0.0025, which is not an ordinary design
which is equivalent to a loss of 4.51 hp. It may be significant for some situa-
value). A conclusion repeatedly verified by experiment is that the smoother
tions to note that this loss can be materially reduced by increasing the clear-
the surfaces, the greater the load capacity of the bearing, in that the lubricant
ance in the cap, say as suggested by Fig. 11.19, leaving the normal clearance
film can be thinner without metal-to-metal contact—a potent argument
Boe
for breaking-in any new machine carefully with light loading as added in-
&
surance against failure."+-9°-11.311 The previous statement must be quali-
Babbitt Removed fied to the extent that the bearing materials must be strong enough, especially
in Center in fatigue if the load is repeated, not to break or deform too much. Only
FIGURE 11.19
+ Babbitt Left 7 in exceptional circumstances would a design value of Zn/p < 5 be used for
4

Here
hydrodynamic bearings; in the general case, Zn/p ~ 50 would be more
reasonable.
lets
"
at each end to avoid excessive axial flow and to take some small unexpected 11.24 THIN-FILM LUBRICATION. Many bearings operate as
change of load. Suppose the clearance in the cap were increased tenfold thin-film (boundary) bearings at all times, but of course the capacity of
hay = 0.0462 in. We see that the fhp is reduced to 0.451 hp. such bearings is not so great as when the surfaces are completely separated
by an oil film. Boundary lubrication is also involved in metal cutting, screw
threads, meshing gear teeth, piston and cylinder, etc. In bearings, thin-film
11.23 SIGNIFICANCE OF Zn/p. A few years ago when we were lubrication may occur because the motion is oscillation, the speed is low, the
learning the full significance of hydrodynamic action in a bearing, viscosity is low, the pressure is high, the bearing is subjected to frequent
parameter Zn/p was a popular criterion; Z = viscosity in centipoises, Starts and stops, and because of insufficient lubricant. Bearing design for
rpm, p = W/(DL) psi. (Recommended values for different applications thin films is primarily empirical. Yet in recent years, much has been
found in the literature.) Values of Zn/p and un,/p (dimensionless) are diff learned of the phenomena involved; anyone bedeviled by boundary lubri-
ent only because of units. Each sleeve bearing has its own (Zn/p)-f cu cation problems should search the recent literature for help. Space limita-
but they are all similar to that in Fig. 11.20, which shows what happens tions prevent mention of significant detail here.
the coefficient of friction f as Zn/p is varied by changes of Z, n, p, any Polar molecules are typically long-chain, fatty-acid unsymmetrical
all. At the higher values of Zn/p, there is hydrodynamic action and thi molecules, one end of which is a nonhydrocarbon group; this end is active
film lubrication. As Zn/p decreases from such a state, a point of minim and attaches itself tenaciously to many but not all metals, the action being
fis reached at B. If Zn/p is further decreased, the high points on the surfa adsorption or chemical reaction. These molecules, as oleic acid and stearic
begin to touch and f begins to move up rapidly (toward a value of the or acid, are inherent in vegetable, animal, and marine oils. When they are pre-
of 0.1 for an oil-wet surface). In the transition region BCD, we have bo sent in a lubricant, these molecules make a bond with surfaces, extending
dary lubrication and the relations for hydrodynamic action no longer apply- their long dimension more or less perpendicular to the surface. Several
The location of the lowest point B depends on the surface roughness; for the layers of molecules thus form a virtual carpet with reduced coefficient of
smoothest surfaces with the bearing carefully run in, which smooths the friction, as compared to a surface wetted with nonpolar (symmetric, as

323
 ee ne
Cover Plate f Inspection Opening

Upper. Half ilold- Down Bolt

324 JOURNAL AND PLANE-SURFACE BEARINGS [Ch. I! Felt Gasket
Liberal Distance Bearing Seat -+-h§ For Bearing Hous-
from Oil Groove ' ing Cover Plate
hydrocarbons) molecules."1-1:1-41 This effect is independent of viscosity to Bearing Housing
and was first called the oiliness by Kingsbury, who had noted differences in —
f with the use of animal fats in thin-film lubrication, Fuller''**! quotes —
Upper Half of Split
Bearing Sleeve
some interesting comparisons for boundary lubrication: nickel, a metal to {
which the polar molecule does not bond showed f = 0.7 for a clean surface, — Lower Half of Spht
f = 0.3 with paraffin oil, f = 0.28 with oil + 1% lauric acid (not much © Bearing Sleeve
change in f with the addition of lauric acid); copper, with which the polar —
molecular reacts strongly, showed f = 1.4 (clean), f = 0.3 (oil), f = 0.08 — ae toe Large Radiating Surface
(oil + 1% lauric acid); aluminum, which reacts somewhat but not avidly Keep Dirt Out
of :
y
= —-
Bearing Seat
=—
of Bearing Housing Plus
Large Oil Capacity of
with the acid, showed f = 1.4 (clean), f = 0.7 (oil), f = 0.3 (oil + 1% — Bearin Oil Ring— Ce -—=— —TVY Bearing Plus Rapid Flow
Extra of Air Along Bearing
lauric acid); all values at low speed, 0.01 cm/sec. On the basis of surface — Large Capacity Housing Due to Directed
roughness,"!-1) we find: for 2 uin. rms, f = 0.128 (mineral oil), f = 0.116 © Oil Reservoir Posting Utnusually
Keeps theCool
(oil + 2% oleic acid), f = 0.099 (100% oleic acid); for 20 win. rms surface, —
f = 0.360 (oil), f = 0.249 (oil + 2% oleic), f = 0.195 (100% oleic). From Bearing Bracket: &
Oil Drain-Out Plug
these and other values, we may conclude: the fatty acid results in a signifi-—
cant lowering of the coefficient of friction for the more usual bearing Coe 11.21 Construction of Ring-oiled Bearing. (Courtesy Electric Machinery Mfg.
materials, sometimes dramatically; fatty acid is of little help if it does not 0.).

react with the material; the value of fis also quite dependent on the surface «
roughness. i
When the load is on top of a journal, it may be possible for the shaft
The usual assumptions are made as for “dry” friction, to wit, F = fN_ to dip into a sump of oil, or be in contact with oil-soaked waste as in some
and F is independent of the area over which the load N acts. About the only railroad wheel bearings. In reciprocating mechanisms, or others with moving
general design advice that can be safely given is to be sure that the speed and links, it may be possible to lubricate bearings by splash lubrication, a moving
bearing pressure (say p S 50 psi) are low enough that the temperature rise member splashing oil from a sump. Systems that do or might result in
during operation will be conservative. Maximum value of pum = 50,000 boundary lubrication include: oil and grease cups, drop-feed devices, and
psi-fpm (see § 11.27). Rippel'+-2#! advises a bearing length given by if
wick feed (by capillary action).
Saas,’ Since it is insurance against a surface discontinuity in the load-carrying
(w) L= PORE CERIN. n.,
15.28 At
region, the line of action of the resultant load should fall within, say, a 60°
i
angle in the center of one of the bearing halves; pillow blocks with the split
where W Ib. = load, n rpm, At = temperature rise of the lubricant above on an incline, as in Fig. 11.22, may be helpful in this regard. In small bearings
ambient, and f = coefficient of friction. Values of f, the most problematic
unknown, are likely to be between 0.08 and 0.15, but perhaps as low as
0.02 if thereisa partial oil film. Equation (w) is based primarily on the assump=
tion that the frictional loss is dissipated to the surroundings. This is an area
of design for dependence on experiment or past experience. FIGURE 11.22 Angle Pillow Block. Manufactured
with the plane of the split at 30° or 45° with the
horizontal. (Courtesy Link-Belt Co. Chicago).
11.25 CONSTRUCTION AND LUBRICATION. Only a little space”
can be devoted to construction details. Figures 11.14, 11.15, and 11.16,
with the discussion of the kind of grooving used in hydrodynam ic lubrica~
tion, have already been mentioned; Fig. 11.21 gives additional information
for an oil-ring bearing. Experiments have been run on the amount of oil a
ring will deliver to the bearing,"19! and such a bearing can be designed with one-piece bushings pressed into place, allowance should be made
so that oil is plentiful. Observe that the lower half of the bearing with an for the fact that the size of the bearing will decrease, or the bushing should
oil ring, perhaps the load-carrying part, can be left without a discontinuity, be finished to size after it is in place. Bearings are made in a myriad of
but that a collar bearing, Fig. 11.15, is necessarily in two parts. ; ways,"1-81 but a frequent characteristic, especially in large bearings, is to
Relief t Relief g11.27] SEMILUBRICATED AND NONLUBRICATED BEARINGS 327
Bore \ 7 - Bore

ane FIGURE 11.23 Bearing Relieved at Split.

stronger and harder than babbitt. A copper-lead mixture, 25-50% lead,
y
in a layer about 0.03 in. thick, has good fatigue strength; normal load
capacity is about 3000 psi. Tin bronzes have a normal load capacity of
5000 psi.1-25)
Silver bearings for heavy duty are made by depositing a 0.02- to 0.03-in.
layer of silver on steel, followed by a lead layer of 0.001 to 0.003 in.; then
4-5% indium is electrolytically deposited and thermally diffused into the
V-groove the bearing at the slit, or to counterbore as suggested by Fig. 11.23. layer of lead.
This relieved space on the side ahead of the section where the oil reaches A cast-iron bearing with a hardened-steel journal has been found to
the load-carrying surface makes a good place for the entry of lubricant, be an excellent combination from the standpoint of wear and friction if
sometimes with most of it departing at the opposite side. If the bearing is thin-film lubrication exists. However, cast iron does not possess embedda-
pressure fed, the relieved areas should stop short of the ends of the bearings bility and other virtues of a softer material that usually outweigh its low
in order to avoid excessive side flow before the oil reaches the loaded area. cost.
Grooves in the load-carrying area of a bearing are often justified for Aluminum Blots have found favor for bearings in internal combustion
thin-film lubrication, [1.®-11.24.11-34.11.35] byt, except for the circumferential engines and in other demanding applications because of good strength,
groove, practically never in a hydrodynamic bearing. An interesting de- conductivity, corrosion resistance, and low cost; but the mating journal
velopment is the availability of rods cased in bronze (that is, the bearing — should be hardened. A thin layer of babbitt may be used to improve em-
material is on the journal); if applicable, this arrangement mayresultinmuch beddability.
longer-lasting bearings because the wear is not concentrated on a small © Elastomeric materials such as rubber, Fig. 11.24, serve excellently with
load-carrying area of the softer material.

11.26 MATERIALS FOR BEARINGS. Properties that are considered _

advantageous for bearing materials include:"1-+1-8] conformability (low
modulus of elasticity, which means a larger deformation per unit of load), —
compatibility (which includes antiwelding property with respect to steel and FIGURE 11.24 Cutless Rubber Bearing (f ~
score resistance), embeddability (soft enough to allow foreign particles too © 0.01). (Courtesy Lucian Q. Moffit, Inc., Akron,
O.).
large to pass through the minimum film to be forced into the surface and ©
thus to be removed from the oil film without scoring and wear), low shear
strength (capacity to be smoothed easily), compressive and fatigue strength
(capacity to withstand the maximum hydrodynamic pressure and to resist —
brittle fracture for repeated loading—at whatever temperature it operates),
corrosion resistance, good heat conductance, nearly the same coefficient of ©
thermal expansion as the material of the housing and journal, and, as always, _ water as the lubricant, and are widely used for shafts of ship propellers,
low cost. q hydraulic turbines, hydraulic dredges, pumps, etc. Thesoftrubber passes sand
Most used are babbitts and copper alloys, Table AT 3. The babbitts are and grit without scoring the journal."!1-8.11.36] Numer other materialsare
ous
either tin-base or lead-base babbitts, depending upon which metal is the used for bearings with and without oil or water lubrication, including wood
principal alloy. Babbitts are relatively weak in all forms, and lose strength © (lignum vitae and oil-impregnated oak), plastics (nylon, Teflon, phenolic
rapidly with increasing temperature, with the result that they are used more resins), and ceramics and ceramets (especially in unusually high temp-
and more in a thin layer (less than 0.04 in. and down to 0.002 in.) on steel erature applications).
backing. With their low fatigue strength, they are not satisfactory where the
load issevereand variable, although the thinnest layers stand up. In 0.016-in.
thickness, normal load capacity (thick-film lubrication) is about 1500 11.27 SEMILUBRICATED AND
psi. 1.25] j
NONLUBRICATED BEARINGS.
There are millions of bearings in machines that are not expected to get much
The copper alloys used for bearings are mostly bronzes which are much
 TT TTT ee
Vw wes PEINIwD [V. Ft $f beers ee ee a re ee ee

maintenance attention (especially those around the home) and others where — 11.28 LUBRICANTS. Solids, gases, and liquids are used as lubri-
maintenance may be impossible, difficult, or costly. In response to these cants, but this article treats briefly of liquid lubricants only. The most
needs, bearings have: been developed that carry enough lubricant for a jmportant are petroleum oils, which usually contain one or more pee
.
reasonable lifetime of service—or that require no separate lubricant at all, that advantageously change some property. Purposes of additives inclu
An early development in this direction is the sintered bearing, which is made to reduce the rate of oxidation (antioxidants: phosphorous, sulfur, —
by first compressing to shape a mixture of powdered metals (usually copper clean IC engine surfaces (detergents) by keeping insolub e
to maintain
and tin or iron and copper), and then sintering at a temperature between particles in suspension; to reduce corrosion (anticorrosives or rust -
attac
the melting points of the two metals. The result is a porous material wi hb hibitors), by adding compounds that have polar molecules that
voids of 10-35% by volume, which are then impregnated with an oil. Th e themselves to the surface (§ 11.24); to increase load capacity for boundary
oil comes to the surface when the bearing is subjected to a higher tempera- Jubrication (extreme pressure (EP) agents), necessary for automotive
ture or pressure. Count on boundary lubrication and a relatively high hypoid differentials, for example; to lower the pour point (pour-point
efficient of friction, say f = 0.12 as long as oil is present. Use limit val depressants); to improve the viscosity index (VI improvers); to ee
for temperature °F, bearing pressure p psi, peripheral speed »,, fpm, a formation of foam (foam inhibitors), helpful where severe churning of oi
PUm as: 150°F, 4000, 1000, 50,000; Fuller"+-1) recommends design p = 20 occurs. 41-5,41.81

psi and moderate speeds. ' Since petroleum oils are limited to a temperature range of roughly
Plastics with and without lubrication are being widely adopted, already _10°F to 250°F, synthetic liquid lubricants (see solid lubricants above)
in a large variety of forms. Nylon is improved for this usage by add have been and are being developed. Among the favored ones are oie
graphite or molybdenum disulfide (limit pv, ~ 3000). Helpful design silicones, which have a high viscosity index and have been used to — 100°F
formation on Zytel (101 Nylon resin):"1-55-11-61] coefficient of fricti and to 400°F, or better.“1:5}
without lubricant 0.15-0.33, with water 0.14-0.18, with oil 0.09-0.14,
Limit values of pv,, for continuous and intermittent use respectively are:
dry, 500, 1000; initially oiled, 2000, 2500; water, 2500, 2500; wick lubrica- 11.29 THRUST BEARINGS. Many rotating shafts, some vertical,
tion, 50,000, 70,000. f are subjected to axial forces of significant magnitude, forces that must be
Teflon filled with glass fiber (25%) is dimensionally more stable than balanced at a bearing, called a thrust bearing. The simplest type carries
other plastics, will withstand a higher temperature, is corrosion resista , the thrust on parallel surfaces (no wedge shape film), as on the end of the
and has a remarkably low coefficient of friction “dry,” 0.05 < f < 0.1 at shaft, Fig. 11.25, or on collars, Fig. 11.26. Unless the oil is introduced
low speed,"*-®) but more nearly 0.25 for speeds greater than 100 fpm and onto the bearing surfaces under a pressure sufficient to support the load
p = 50 psi."7-41) Limit pv, = 10,000 for low-speed (v_, = 10) continuous (§ 11.30), boundary lubrication should be assumed (f = 0.1 to 0.03), and
service; 20,000 for intermittent use."1-°7-11-41] [mpregnated with molyb- the bearings used for moderate operating conditions (50 < p < 200 psi,
denum disulfide, this Teflon mixture has found favor for dry bearing sur- 50 < v, < 200 fpm).
faces in space machines, ‘Plastic’ bearings can be designed for hydro-
dynamic lubrication, but one should keep in mind the dimensional varia-
tions with temperature and moisture content.
Carbon-graphite inserts are another type of dry bearing, the mixture
of carbon and graphite acting as its own lubricant. They can tolerate tem-_
FIGURE 11.25 Thrust Bearing for Vertical Shafts.
peratures of 750°F; limit pv,, = 15,000 dry, but not so much for continuous
(Courtesy Link-Belt Co., Chicago).
operation. Also, a graphite film some 0.00015 to 0.0005 in. thick, called
Electrofilm, can be deposited on bearings, gears, cylinders, splined shafts, —
etc., to reduce friction and prevent galling. The use of bonded solid lub-
ricants such as molybdenum disulfide in a binder of corn syrup, on bearing :
surfaces of various metals has proved successful for many inaccessible
bearings, where the presence of oil or grease is disadvantageous, in vacuums, _ Oil Holes
at very low temperatures (liquid hydrogen), and at high temperatures!1-40] _ a ee OD
(but MoS, decomposes into abrasive material at some temperature above _
700°F). FIGURE 11.26 Sketch of Collar-thrust Bearing.
 330 JOURNAL AND PLANE-SURFACE BEARINGS [Ch. I!

Hydrodynamic thrust bearings were designed by Albert Kingsbury in

this country and A. G. M. Michell in Australia, at about the same time
working independently, each being guided by the classical work of Rey- —
nolds.1-42] The basic elements of such bearings are a collar and a group of
FIGURE 11.28 Hydrostatic Thrust Bearing.
nonrotating segments, Fig. 11.27. The segments in the actual bearing may
be at a fixed angle of attack, they may be pivoted so that they are free to —
assume any angle of attack, or they may be mounted on flexible supports
that permit variation with minor constraint. With oil adhering to the collar,
running providing the film; where the film is maintained by oil flow due
to external oil pressure, as in the elementary hydrostatic thrust bearing
ROTATING ELEMENT
of Fig. 11.28, the lubrication is said to be hydrostatic. Large, heavy tele-
FIGURE 11.27 Basic Elements of the Kingsbury Bearing. scopes, antennas, etc. are often supported on hydrostatic films with coeffi-
qa!
<i Ui
iN Observe that the wedge formation of the oil film insures a com- — cients of friction surprisingly low. Fuller™1-! calculates a value of f =
| plete separation of the metal parts, provided of course that the
oil is suitable. for the load and the speed. (Courtesy Kingsburg —
0.00046 for one case and quotes another of f = 0.00000075. These values
Machine Works, Inc., Philadelphia). come about largely because of the low speed, inasmuch as the frictional
force to shear a fluid at zero speed is zero. Moreover, if a comparison were
to be made on the basis of total loss, the oil pump work is charged against
the bearings, which will make the equivalent coefficient of friction perhaps
which is attached to the rotating member, and with the converging wedge- of the same order as those computed for hydrodynamic action. Methods of
shaped film, hydrodynamic pressure is built up on the segments, which sup- making theoretical calculations for hydrostatic bearings are found else-
port the load. These bearings are designed by the hydrodynamic theory, where, {11-1,11.46,11.65]
an application somewhat simpler than the form applied to journal bear-
ings;"1-1) their coefficients of friction are of the same order as those for
journal bearings. The load capacity is a function of yun,/p. See Ref. (11.44) j
for charts. Fuller! notes the 96-in. diameter bearing supporting the 11.31 GAS-LUBRICATED BEARINGS. Gas, especially air, is used
2,150,000-1b. load of the rotor of a Grand Coulee Dam hydraulic turbine; as a lubricant and does fine except that its load-carrying capacity is quite
S = 0.0009. small (say 10 psi). The hydrostatic air bearing is used as a thrust bearing,
On occasion, large journals are supported by pivoted segments, three Fig. 11.28; and air bearings are also designed to operate hydrodynamically.
or more, distributed about the circumference. Such bearings can take a The solution of the Reynolds equation that we have previously used is on
radial load varying in any direction and, moreover, the segments will hold the basis of an incompressible fluid in laminar flow. Since a gas is compres-
the shaft accurately in a radial position (as for a machine tool spindle). sible, other solutions allowing for varying density are needed and are avail-
able (tables and charts in Ref. //.47)."1-11 For low Reynolds numbers and
light load, the Petroff equation (e) gives a reasonable estimate of the fric-
11.30 HYDROSTATIC LUBRICATION. A loaded journal at rest tional moment. Because of the low viscosity of gases, the frictional losses
soon makes surface contact with its bearing. Since this is so, it is evidently are only a fraction of those for liquid lubricants of any kind, and gas is
wise to start machines under no load or light load to avoid excessive wear sometimes appropriate for exceptionally high speeds and light loads, as
of the surfaces on start up, especially for machines frequently started. In gyroscopes, and instruments. Some applications: General Electric
However, in many cases, as for turbine rotors, this is impossible. Also, Teports the development of a small cryogenic system, with a weight of 50 lb.,
there are many other journal bearings that turn so slowly that hydrodynamic” that liquefies helium, whose compressor has been rotated at 350,000 rpm,
action does not build up a separating film. And finally, there are thrust and in which a small turboalternator rotating at 250,000 rpm expands the
bearings, Figs. 11.25 and 11.26, that are not inherently designed for pro-_ helium, both turning on air bearings; Boeing reports a hydrostatic air thrust
ducing hydrodynamic action. To reduce the large friction that would bearing running at 100,000 rpm; grinder bearings with air hydrostatic
otherwise exist in these situations, oil is pumped into the load carrying— lubrication turn up to 100,000 rpm; speeds up to 1,300,000 rpm have been
area at sufficient pressure to support the load. The turbine journal may be attained in the laboratory.4!:1)
“floated’’ in oil at start up, but after it is up to speed, hydrodynamic action A troublesome phenomenon is likely to be whirl or whip, which is the
may be permitted to take over. In some situations, the oil pumps are kept orbiting of the journal center with respect to but not necessarily around the

331
7s : IVVUANAL AND PLANE-SUNrTAGCE BSEANINGS (UN. tt

bearing center. This is an instability that may manifest itself in a destruc- {

tive vibration. Journal whirl also occurs in high-speed, lightly loaded, oil-
lubricated bearings, and is called half-frequency whirl.“1-241

11.32 DYNAMIC LOADING. Many bearings, as in internal com- —

bustion engines, are subjected to high impulsive loading. If the time of the —
impulse is short, it does not necessarily follow that the film thickness goes —
to zero, even though the maximum load is theoretically large enough to pre- _
dict this, simply because it takes time to squeeze the oil from the bearing. —
The same effect is of course involved in any kind of bearing. Some calcula- ;
tions can be made for this kind of situation,“1-1:41-+-41-241 spoken of as_
squeeze-film lubrication. Not only must there be a copious supply of lubri-—
cant, but oil holes or grooves in the area taking the impulse will certainly —
make surface contact more likely by reducing the attainable squeeze-film
pressure.

11.33 CLOSURE. All relatively moving surfaces in contact profit 12. BALL AND ROLLER
from some lubrication, but the subject is too large to be covered in a survey.
In general in this book, we assume that if lubrication is proper and adequate, — BEARINGS
the design as computed should perform satisfactorily—as in gear-tooth —
design. i
In designing journal bearings of importance by the theory we have out- | 12.1 INTRODUCTION. Probably the most significant advantage of
lined, the engineer is likely to make numerous solutions, perhaps in the — rolling-contact bearings is that the starting friction is not very much larger
form of curves similar to those of Figs. 11.9, 11.10, and 11.11, from which — than the operating friction (at usual speeds and contrasted with an initial
he chooses the solution that seems to balance the pros and cons. The appli- — metal-to-metal rubbing of ordinary sliding-contact bearings); that is, the
cation of theory to hydrodynamic bearings for the first time, no matter — “coefficient of friction’ (§ 12.14) varies little with load and speed, except
what traditional practice says, should lead to improvements. at extreme values. This property makes rolling-contact bearings particu-
larly suited for machines that are started and stopped frequently, especially
under load (as axle bearings on railway cars, where they are being slowly
adopted). Other characteristics include: they require little lubricant and
maintenance; they occupy less axial space but larger diametral space than
journal bearings; they are noisier than journal bearings, more expensive;
they have a limited life because the raceways are highly stressed repeatedly
as the shaft rotates, resulting in eventual fatigue failure; and several types
of bearings may take a radial load and an axial thrust.
Sliding-contact and rolling-contact bearings have advantages which
make one type more suitable in a particular application than the other,
and it cannot be said that either one is better except in reference to a given
engineering problem. The rolling bearing is a standardized, specialized,
and precision device that the machine designer does not design; from a large
variety, he selects an appropriate type, style, and size from a catalog. For
intelligent selection, he needs to understand the basic considerations affect-
ing capacity and life. Although we speak of “rolling-contact bearings,”

333
 334 BALL AND ROLLER BEARINGS [Ch. 12

this is a device for classification inasmuch as in operation both balls and

rollers do considerable sliding (sliding and rolling).

12.2 STRESSES DURING ROLLING CONTACT. The stresses

involved are quite high because of the small area of contact (a force of
1 1b. pressing a }-in. ball on a }-in. ball induces a maximum stress of about
150 ksi™1!-*}), and the Hertz equation for two cylindrical surfaces in static
contact in § 1.27 is applied to roller bearings. (Review § 1.27 before reading
further.) However, at the outset, we should recognize that the sliding
(frictional force) results in a tensile stress near the surface, and it is this
a Pillow Block

force that is thought to cause pitting.-6*! Also, since the contact stresses FIGURE 12.2 Single-row, Deep-groove Ball Bearing Mounted in Bearing Block. Since
the balls in this type of bearing are loaded by the eccentric displacement of the inner race,
are greater than the yield strength, rolling induces residual compressive the spacing is greater than when a filling notch is used. Compare with Fig. 12.6. (Courtesy
stresses in races and balls, which increase with repetitions, but there are Link-Belt Co., Chicago).
balancing residual tensile stresses just below the surface. Moreover, the q
deformations change significantly the local values of the radii of curvature
the time life of the bearing by one-half. As the speed increases to high values,
that appear in the Hertz equation. It follows then that for rolling contact as
in these kinds of bearings, the Hertz stress is more of an index than an actual centrifugal force begins to have an effect; for example, a number 206 ball
bearing rotating unloaded at about 20,000 rpm had a life of about
evaluation of the stress. Usually, fatigue cracks clearly start at pits or ©
1000 hr,@1-51
inclusions; so, it is not unexpected that bearings made from vacuum-melted
A change of the load has a much more decisive effect on the maximum
steel have a longer life.“?-%!
Similar equations for a sphere on variously shaped surfaces have been ~ stress and life. Palmgren"!?:1) found experimentally that the life B varies
inversely as the k power of the load F; B oc 1/F*, where values of k from
derived. However it is easy to understand without these equations that one ~
3 to 4 are found in the literature; the ASA recommends 3 for ball bearings,
factor affecting the magnitude of the actual stress is what is fetchingly called
10/3 for roller bearings. In equation form,
the degree of osculation. The area in contact in Fig. 12.1(a) is greater than, 4
and the peak stress is less than, the corresponding area and stress in ©
(12.1) (2)
a i faisoes
erica | or Fi
epee (3).
eed, Pip teeeneiecaey

PD 2 €
Fy Fi” B, Fy B,
FIGURE 12.1 Degree of Osculation. The greater ~ Observe that for k = 3, halving the load increases the life by 8 times; etc.
the osculation, the lower the stress; after some rela- —
tion of diameters is reached in (c), decreasing the dif- —
Environmental conditions that reduce the fatigue life of metals,
ference between radii of ball and race results in — corrosion from acid or water for instance, also reduce the life of rolling
increasing friction. bearings. The preferred unit for measuring life is millions of revolutions mr,
(a) (b) () but some catalogs are set up with a certain number of hours at various
rotative speeds. Fatigue occurs in the form of spalling or flaking of small
Fig. 12.1(b). Because the area in Fig. 12.1(c) is greater, the stress is lower particles of material from the surface of a ring or ball (or roller), because of
than that of either of the other two. Partly for this reason, the races of — the high shearing stress just below the surfaces in the vicinity of contact
ball bearings are curved to wrap about the balls, Fig. 12.2. Even with the — (§ 1.27). When a rolling bearing has “failed’’ in this way, it becomes noisy,
most favorable shapes, the peak stress at a particular point in a bearing race thereby giving warning. Continued running may eventually cause a ball or
(or on a ball or roller) as the rolling element passes across that point under roller to break, the end result of the progression of a fatigue crack.
load (even with a “light” load) is quite high. Since such a stress is repeated ~
with shaft rotation, it is natural to expect that eventual failure will be a ©
fatigue failure.“?-2?] And since any normal loading induces stresses above 12.3. STATISTICAL NATURE OF BEARING LIFE. Even for opera-
the fatigue strength involved, rolling bearings have a limited life, limited by tion under controlled conditions, fatigue tests of bearings which are
the number of repetitions of stress to cause fatigue failure. For example, © commercially identical show a range of life as great as 50 to 1. Thus there
on average, with a particular load, doubling the speed of rotation reduces is no way to predict the life of an individual bearing, and since this is so,

335
 ob
awoeon #£
ow
512.4) STATIC LOAD CAPACITY . 337
ey
apply so well; %?-11-12-21] but bearing test data have been analyzed by W.
Percentage of Occurrences

Weibull, who found that the probability P of survival without failure for a
ee —

particular life B is
=
ee

b
N

(12.2) P=e@e? or InP = -(=) ; [e=2.7183]

ee

a
co

FIGURE 12.3 Histogram Showing Life of —

Rolling Bearings. A normal curve (Figs. © [WEIBULL DISTRIBUTION]
oO

1.14, 3.3, 6.10 ete.) could be computed to ~

&

Median represent these data, but they are evidently — where In is to the Naperian base, a and b are taken as constants that can be

L eta
&»

skewed. The Weibull distribution is a better defined by experiment. We shall use equation (12.2) later to find an equiva-
representation.
wo

lent “90% life” when a greater than 90% survival is desired.

1 Bi BA Sh OTs Bi Oe) 10 Ts Ee
Life

the only logical approach is from statistics and probabilities. The histogram 12.4 STATIC LOAD CAPACITY. With contact stresses as high as
of Fig. 12.3 shows the distribution of the lengths of life, measured by inde they are, it does not take a large load to produce a permanent deformation.
numbers, as found for a particular series of tests. By way of interpretation,
These indentations are called brinells and the act of indenting is called
cell 3 shows that about 16% of the population (total number tested) had a brinelling or false brinelling. The objectionable damage is done on the loaded
bearing at rest, since during operation the deformations are distributed
life between 2.5 and 3.5 units, where the median life was 4.35 units; the
over all the race and rolling surfaces. The question is, how much brinelling
median life is the middlemost value. All the reasons for such a large spread
can there be before the bearing is ruined, and there is no simple answer
are not clear; actual defects that may result in the very short-lived ones a
not evident on examination; perhaps undetectable crystalline flaws, perhaps because it depends on the service. A deformation as small as 10 win. can
unusual and unknowable residual stresses. 4
be detected with an optical flat, but the effect of a deformation in normal
The ASA@2:8] recommends ratings of rolling bearings on the basis tha operation is not bothersome until the indentation is of the order of 100 yin.
90% of a large group of bearings in particular surroundings will survi
(0.0001 in.),"*-8! which is the basis on which the basic static load rating F,,
a specified life; this life may be designated B, 9, which is then interpreted Table 12.3, is calculated—equation (c) below. SKF™?-*) suggests that the
mean 10% failures may be expected in By revolutions (or time). However, permanent deformation in the ring and rolling element combined be less
than 0.0001 times the diameter of the rolling element. The static capacity
one must examine catalogs carefully, because other bases are in use; some
F;, Table 12.3, is a practical measure of the amount of brinelling that can
catalogs give the rated load for 50% survival Bs (median life), which is a
number quite different from Bo, as predicted by equation (12.1). A com= normally be tolerated when the bearing is to rotate. If subsequent noisy
monly assumed relationship that probably accords with different catal operation is permissible, a static load larger than the static capacity may
be allowed. Also, where the operating motion is slow, higher static capaci-
ratings in the USA is
ties can be used: the bearing loads supporting artillery pieces may be twice
(a) Median life ~ (5)(90% life); Bso + S Bio. the catalog rating; static loads on bearings for aircraft-control pulleys and
On sliding-door rollers may be four times their catalog ratings; the static
However, certain extensive tests on deep-groove ball bearings yielded fracture load will be more than 8F,."2-4)
Bso = 4.08B,o,"22-°! which may be used if appropriate. At the other end The radial static-load capacity is computed from Stribeck’s equation,
of the spectrum, one may estimate the Jongest life at 4 to 5 times the medi
life (4B;, to 5Bso). The design life varies widely, depending of course (c) F,'= C,N,D,?,
the machine and service; the values in Table 12.1 will be an aid to judgme
Which says that the static load capacity F,, called the basic static load rating,
The expression “*90% life’’ is interpreted to mean that the probability o
1S proportional to the number of balls (or rollers) N, and to the square of
particular bearing surviving the given time is 90%. To convert to revolu-
the ball diameter D,. The proportionality constant C, depends on the type
tions,
of bearing and the materials. For example, the static load capacity of a
(b) Revolutions = (Hours)(60 min/hr.)(” rpm). Single-row, deep-groove, radial ball bearing is about F, = 5000N,D,?."1-4)
Catalogs generally give values of F, so that the engineer does not need to
The distribution of Fig. 12.3 is seen not to be symmetric (contrast wit h
Compute it. If a bearing is to be chosen to withstand a combination of
Fig. 1.14), so that the equation for the symmetric distribution does not
336
338 BALL AND ROLLER BEARINGS [Ch. 12 512.7] SELECTING BEARINGS FROM TABLES 339
radial F, and thrust F, loads, an equivalent static load F,,, which causes” load. Not all manufacturers’ catalogs give the ASA procedure;?-8! we
the same deformation as the combined loads, is computed from shall cover some of its highlights for deep-groove, single- and double-row
pall bearings only. The equivalent load F, is obtained from
(d) Fes ae C.F; ot C2F,,
(f) F, = C;F, or F, = 0.56C,F, + C.F,
where C, and C,, given in catalogs, depend upon the type of bearing; [F.[CrFz = Q] [F./CrF: > Q]
for the single-row deep-groove, C, = 0.6, C, = 0.5, but F,, is never taken
less than F,. where F, is the radial load (computed from a force analysis), F, is the thrust
load, C, is a rotation factor (the evidence is incomplete but C, = 1 for
inner race rotating, C, = 1.2 for outer race rotating with respect to a
12.5 DYNAMIC LOAD CAPACITY. Since a rotating bearing stationary inner race; for a particular number of revolutions, a ball rolls
by fatigue, the dynamic capacity is different from the static. Palmgren"™?: further on the larger outer race than on the smaller inner race), C; is a
first modified Stribeck’s equation, considering the variables introduced thrust factor, obtained from Table 12.2, that applies only when F,/(C,F,) >
by rotation; then the ASA,"?-8! on the recommendation of the AFBMA, Q; Q is found in Table 12.2. We see that if the thrust load is a low enough
sanctioned the rating equation now in use, fraction of the radial load, it can be ignored. It is a common practice to
preload ball bearings with a thrust force by means of an adjusting nut or a
(e) F CN,?3(N, cos a)°-7,p18,
spring; this may be done to hold a shaft in position or to reduce vibrations.
All bearing manufacturers are eager to help with special problems.
where F, = the dynamic load capacity of the bearing (= F, for 1 mr it
Table 12.3), N, = number of balls, N, = number of rows of balls
D = diameter of a ball (D 2 1 in.), C = a constant that varies somewha’
12.7 SELECTING BEARINGS FROM TABLES. The rated loads in
with the type of ball bearing, and « locates the plane of the resultant force
Table 12.3 are on the basis that 90% will survive 10® rev. (1 mr) when the
Fig. 12.4, when there is a thrust load.
equivalent load is used for selection. If there is shock or vibration, the
equivalent load should be further modified in accordance with the designer’s
judgment. For example, for shafts connected by gears, multiply the com-
puted values of F, by a service factor as follows:"?-4!
Rotating machines, no impact; electric motors,
rotary compressors, etc., 1.1 to 1.5;
Reciprocating machines, 1.3 to 1.9;
Machines with pronounced impact, hammer mills, etc., 1.6 to 4.
FIGURE 12.4 Angular Contact Bearing. Observe that the action of thi
thrust is such as to move the surface of contact away from the center-! Having decided upon the equivalent load F,, compute the rated load F,
plane of the balls. Compare the shape of the grooves in this figure from equation (12.1). It is convenient to work with millions of revolutions
those of Figs. 12.2 and 12.6. (Courtesy SKF Industries, Inc., Philadelp!
(mr); let B, = 1 mr, the rated number for Table 12.3; k = 3 for deep-
groove ball bearings; then by (12.1),
(g) F, = (Byo)"* F,,
where By, mr is the desired number of revolutions before 10% failures have
occurred, and the bearing is chosen to have a rated load equal to or better
than F,. For B, = 1 mr, as above, this corresponding F, is called the basic
dynamic load rating. Compute Bio from equation (b) above. Ordinarily,
12.6 EQUIVALENT DYNAMIC LOAD. Ball bearings and so 1 there will be several types of bearings that meet the requirements; so a
roller bearings are subjected to both radial and thrust loads. Since decision on the kind of bearing, § 12.15, must be made. Moreover, the load
possible combinations of these loads are infinite, manufacturers rate thei Tatings of the different series of the same kind overlap, as seen in Table 12.3;
radial bearings in terms of a radial load only, their thrust bearings in terms the decision in this regard depends in some measure on the space require-
of thrust load only. Therefore, it becomes necessary to use an equivalent ments. The following example will be helpful concerning the details.
 DESIGN LIFE FOR ROLLING BEARINGS, HOURS"*-4)
HOURS 512.8] EXAMPLE 341
TYPE OF SERVICE (90% life)
F 5; ate at B, = 1 mr. Examine Table 12.3 and note that either ball bearing No. 221 or 317
ng ig use—instruments, demonstration apparatus, sliding a has a larger rated load.
ee BRE SE AU" Poti erat Feat peli aE OMe Fazio? Check 221 first; F, = 20,100 Ib., F, = 23,000 lb. Then
Aircraft engines. ; ‘ 4 2 ; 500 to 2000 F;— 700 F; 700
F, ~=-———
20,100 =0.0348,
GF.ee ~Sa
@.2x800) ~ 072%
BO;
Intermittent use, service interruptions of minor importance— Jan 2
hand tools, hand-driven ir farm machinery, cranes, which is seen in Table 12.2 to be greater than Q (0.22 <@Q <0.26) for the fore-
household machines . . «+... «. 4000 to 8000 going value of F,/F,. Interpolate in Table 12.2 for C, corresponding to F,/F, =
0.0348 and find C, = 1.92. Recompute the rated load for this value of C;;
Intermittent use, dependable operation important—work F, = (0.56)(1.2)(800) + (1.92)(700) © 1880 Ib.
moving devices in assembly lines, elevators, cranes, and less- Me 13 1/3 oh
frequently used machine tools ‘ : ; i 3 8000 to 12,000 F, = Bio'/°F, = (1890)'/°(1880) = 23,200 Ib.

8-hour service, not fully utilized—gear drives, electric motors 12,000 to 20,000 TABLE 12.3

8-hour service, fully utilized—machines in general, cranes, SINGLE-ROW, DEEP-GROOVE BALL BEARINGS
blowers, shop shafts. : , ; 5 : : 20,000 to 30,000 From Ref. (12.14). Basicdynamic load rating F, is for 1 million revolutions (mr); 90% of a
oup of bearings should last longer than 1 mr. with the rated loads given. The s limit
24-hour service, continuous operation—separators, compres- te is doshas for oil bath lubrication; for grease, use % of these values. net
sors, pumps, conveyor rollers, mine hoists, electric motors 40,000 to 60,000 in catalogs. Factors limiting speed include lubrication, fit, dynamic balance, vibration. It is
i et ' possible to exceed the specified limits.{/2-161 Higher speeds can be obtained by oil mist
ainsi deen re Chae lubrication, by circulating and cooling the oil, or by cooling the bearing. Consult manufac-
is , , , turers.
stations, pumping stations, continuous service machines
aboard ships ‘ ; : 5 : P i . 100,000 to 200,000 200 series (200-222) 300 sERiEs (300-322) 200

BRG, Balls Rated| Speed Balls Rated || Roller

, NO. Static | F,,\b.| Limit Static | F,, lb. F,
12.8 EXAMPLE. Select a deep-groove ball bearing to carry a radial loa No.| Dia.| F,,1b.| 1mr. | rpm || No.| Dia.| F,,1b.| 1 mr. |} 1 mr.
F, = 800 lb. and a thrust load F, = 700 lb. at 1750 rpm. This service is 8 hr. pi
day, but it is not continuous; design for 18,000 hr., Table 12.1. The operation 1 Z : oe 1 pe moe : : i om i.
smooth with little vibration; the outer ring rotates. , 02 8 R 790 1,320 20,000 7 % 1,220 1,960
Solution. First, convert 18,000 hr. at 1750 rpm to millions of revolutions. bp 03 8 bY 1,000 1,650 18,000 7 Pr 1,470 2,340

Bio = (18,000)(60)(1750)(10-*) = 1890 mr,

the desired life with no more than 10% failures. Not knowing what bearing wi : ‘ eee oan an i : ee see 2.980
be used, we do not at this point know the static rating F, (Table 12.3), which 06 9 ; 2 250 3. 360 11 000 || 8 ut 3340 4.350 3,970
needed to decide upon the form of equation (f) to use; assume a value of © 07 9 te 3,070 4,440 9,400 8 PY 4,020 5,750 5,900
Table 12.2, and check it later; assume C, = 1.8; C, = 1.2 for the outer ring ro’tal
ing. The corresponding equivalent load is a 08 9 #% | 3,520] 5,040] 8,400]] 8 3 | 5,020 | 7,040 || 7,670
F. = 0.56C,F; + CiF: = (0.56)(1.2)(800) + (1.8)(700) ~ 1800 Ib. 09 | 9 | 4£ | 4,010} 5,660 | 7,700)) 8 | i | 6,730} 9,120 |} 8,070
, y 10 | 10 34 | 4,450] 6,070 | 7,100]] 8 3| 8,010 | 10,700 |} 8,440
From equation (12.1) with k = 3, we get 11 |10 | & | 5,630] 7,500 | 6,500|] 8 | 4| 9,400 | 12,400 || 10,300
Byo\1/3
F,= ($*) F. = (1890)"/°(1800) = 22,200 Ib. 12 |10 | & | 6,950] 9,070 | 5,900]} 8| % | 10,900 | 14,100 || 12,600
. 13 | 10 +3 7,670 | 9,900 5,400 8 4% | 12,500 | 16,000 || 14,900
14 | 10 # 8,410 | 10,800 5,100 8 |1 14,200 | 18,000 || 14,800
TABLE 12.2 15 |} 11 # 9,250 | 11,400 | 4,800 8 | 1d | 16,100 | 19,600 || 18,200

THRUST FACTOR FOR DEEP-GROOVE, 16 | 10 2 | 10,000 | 12,600 | 4,500|] 8 | 14 | 18,000 | 21,300 |] 19,600
SINGLE- AND DOUBLE-ROW BALL BEARINGS 17 | 11 $$ | 12,000 | 14,400 | 4,200 8 | 1¥ | 20,100 | 22,900 || 22,400
ae ds. F, = basic static load rath 18 |10 | 4% | 13,600] 16,600 | 3,900|| 8 | 1% | 22,300 | 24,700 || 28,600
& tet tee bale saic toad rena) 19 | 10 4% | 15,600 | 18,800 3,700 8 | lie | 24,500 | 26,400 || 31,400
F,/F; 0.014 | 0.028 0.056 | 0.084 | 0.11 0.17 | 0.28 | 0.42 0.56
Cc; 23 1.99 1.71 L355 1.45 1.31 1.15 1.04 1.00 20 | 10 j}1 17,800 | 21,100 | 3,500}]} 8 | 1% | 29,400 | 29,900 || 34,800
Q 0.19 0.22 0.26 0.28 0.30 | 0.34 | 0.38 | 0.42 | 0.44 21 | 10 | 1¥s | 20,100 | 23,000 | 3,300]| 8 | 1} 32,000 | 31,800
| 44 na 11 AN can AA ONnN 21nn 2 a5 4297 Aan AL ANN AL ANN
 342 BALL AND ROLLER BEARINGS [Ch. 12 512.9] CHOOSING BEARINGS WHEN SURVIVAL IS DIFFERENT FROM 90% 343

vs. 23,000 Ib. in Table 12.3. This is of course borderline and the engineer would
12.9 CHOOSING BEARINGS WHEN THE PROBABILITY OF SUR-
have to decide if a slightly larger percentage of failures (than 10%) is permissible In equation (12.2), B may
VIVAL IS DIFFERENT FROM 90%.
in the 18,000 hrs. Actually, of course, the truth is not known with this accuracy,
represent the life in revolutions, mr, or hours (or any unit of time) at a
so that if 90% survival is considered satisfactory, bearing 221 is all right.
specified speed, as 3800 hours at 1000 rpm. Changing signs in equation
In checking bearing 317 in the same manner, we find the calculations the same
(12.2) and writing them for P;, = 0.90, probability of 90% survival, and for
because the two bearings coincidentally have the same static capacity and nearly
the same dynamic capacity. They are so close to the same capacity that the final Pp, any other probability, we have
choice would be made on bases other than capacity. In general, before a final j
decision is made, designs should be made or considered for all appropriate typ
of bearings. See the next two articles for further considerations with respect to a
choice.
VEG calms eked.
Divide one by the other and take the b root;
1/b

357 Linares
B In (1/P)
TABLE 12.4 DIMENSIONS OF ROLLING BEARINGS®2-4 (12.3) » [In(1/Pyo) = 1n(1/0.9) = 0.1053]
This table does not give all standard dimensions. The maximum fillet radius r is the
maximum radius at the shoulder on the shaft which is cleared by the corner radius on from which the life B for a probability P may be found when the life B,)
the bearing. Conversion factors: 0.03937 in./mm.; 25.4 mm/in. and probability P,, at 90% life (or at other state) are known; say Byy =
OUTSIDE DIAMETER | WIDTH OF RACES, 1 mr for P,) = 0.9; or the probability P for a life B may be computed.
BORE mm mm MAX. FILLET r Shube!42-21 finds a and b by using two points, the catalog 90% life and an
assumed median life; for the median life, he used both the generally
BRG. 200 | 300 |400 | 200 | 300 | 400 | 200 300 400
NO. | mm. in. | Series|Series | Series |Series \Series |Series| Series| Series\ Series recognized relation of 5 times 90% life and an experimental value of 4.08
times 90% life (§ 12.3):
00 | 10 |0.3937 | 30 35 9 11 0.024 | 0.024
01 12 | 0.4724 | 32 37 10 12 0.024 | 0.039 (h) For median life = (5)(90% life): a = 6.84, b = 1.17.
02 | 15 | 0.5906 | 35 42 11 13 0.024 | 0.039

03 17 | 0.6693 | 40 47 12 14 0.024 | 0.039 (i) For median life = (4.08)(90% life): a = 5.35, b = 1.34.
04 | 20 | 0.7874 | 47 52 14 15 0.039 | 0.039
05 | 25 | 0.9843 | 52 62 80 15 17 21 {0.039 | 0.039 | 0.059 Additionally, Harris"?441 recommends 5 = 1.125.
More frequently than in former years, it is desired to select bearings
06 | 30 | 1.1811 | 62 ie 90 16 19 23 {0.039 | 0.039 | 0.059 with a much greater probability than 0.9 that the bearing will survive a
07 | 35 | 1.3780 | 72 80 {100 17 21 25 |0.039 | 0.059 | 0.059 certain life. For example, where human life is at stake and, sometimes, just
08 | 40 | 1.5748 | 80 90 |110 18 23 27 |0.039 | 0.059 | 0.079
because the purchaser can afford it, reliabilities better than 99% may be
09 | 45 | 1.7717 | 85 |100 {120 19 25 29 |0.039 | 0.059 | 0.079 specified. The theoretical equation (12.3) predicts that no life is so short but
10 | 50 | 1.9685 | 90 {110 |130 20 27 31 {0.039 | 0.079 | 0.079 that some failures will occur if the population is large enough, and no life
11 55 | 2.1654 |100 {120 |140 21 29 33 10.059 | 0.079 | 0.079 is so long but that some bearings would run indefinitely. Experience does
Not bear out these extreme conclusions,"?-12! but tentative calculations
12 | 60 | 2.3622 |110 |130 {150 22 31 35 |0.059 | 0.079 | 0.079 can be made with equation (12.3). One could also introduce limits in terms
13 | 65 | 2.5591 |}120 |140 {160 23 33 37 |0.059 | 0.079 | 0.079
14 | 70 | 2.7559 |125 |150 {180 24 35 42 |0.059 | 0.079 | 0.098
of standard deviation, as the 3o limits of §§ 3.9-3.12; +40 almost insures
certainty. Experimental data suggest that failure proceeds in two stages;
15 | 75 | 2.9528 |130 |160 |190 25 37 45 10.059 | 0.079 | 0.098 first the inception of a crack in the subsurface and then the propagation of
16 | 80 | 3.1496 | 140 |170 26 39 0.079 | 0.079 the crack until it reaches the bearing surface.4?-11] When the probability
17 | 85 | 3.3465 | 150 {180 28 41 0.079 | 0.098 1s about 0.9 or lower, the first stage lasts much longer than the stage of
18 | 90 | 3.5433 | 160 {190 30 43 0.079 | 0.098
Crack propagation and it is to this phenomenon that the Weibull equation
19 | 95 | 3.7402 |170 {200 32 45 0.079 | 0.098 (12.2) applies. At high reliabilities (lower loading), the time for crack
20 |100 | 3.9370 | 180 {215 34 47 0.079 | 0.098 Propagation lengthens and this stage may last longer than the first. A plot
of data suggests that the actual minimum life at virtually 100% reliability
21 {105 | 4.1339 | 190 |225 36 49 0.079 | 0.098 is about 5% of 90% life.42-11] The corresponding reduction in the load for
22 |110 | 4.3307 | 200 |240 38 50 0.079 | 0.098
a given life is then predicted by equation (12.1).
344 BALL AND ROLLER BEARINGS [Ch. |2 g12.12] MATERIALS AND FINISH 345

A significant aspect of probabilities should not be overlooked. If there (c) A reliability of 0.98 (98% survival) in 10 mr. is desired for a roller bearing
are x bearings in a particular system, the probability of a failure of any one supporting an equivalent radial load of 5000 lb. What 200-series bearing should
is the product of the individual probabilities. If x bearings each have the be selected? ‘
same probability P, the chances of survival for the specified life becomes P*. Solution. The first question is: given P; = 0.98 for B: = 10 mr, what is
For example, suppose a gear box contains 10 bearings, all having exactly Bz for P2 = 0.9 (catalog basis)? From equation (12.3), with b = 1.125,
P = 0.9 at the design life; the probability of any one of them failing during Bo Be ae
In 1/0.90
In */2-128
this life is 0.9'° ~ 0.35. In short, on average for a large population, one of the B, 10 In 1/P, 1/0.98
In
bearings in the system fails at about one-third design life. This characteristic from which, Bz. = 44.3 mr; this life is now equivalent to the Bi, life for 0.9
suggests that for“‘space systems,” which are to have an excellent reliability, probability and a load of F = 5S kips. The next step is to compute the rated load
the individual reliabilities must be quite high. Some examples nevi by equation (12.1), using k = 10/3 for roller bearings;
probabilities will be helpful.
Faas
Pern a tN
ae ss = 44,39,
from which F, = 15.6 kips (for B, = 1 mr); choose roller bearing No. 215,
12.10 EXAMPLES—PROBABILITIES AND LIVES OF ROLLING BEAR-
Table 12.3, whose rated capacity is 18.2 kips. This bearing is probably overly
INGS. (a) Ina certain catalog, we find that the rated load capacity of a No. 212
conservative because of the statistical tendencies discussed in § 12.9; the chances
bearing is F; = 2550 lb. at a speed of 1000 rpm for 3800 hr., average life basis.
are good that a 214 bearing would meet the required reliability.(1?-11)
What is the corresponding basic dynamic load rating (for 1 mr)? i
Solution. Convert the life to mr;
= (3800)(1000)(60)(10-*) = 228 mr. 12.11 VARIABLE LOADING. If the load on the bearing varies, the
To estimate the 90% survival life, use 228/5 = 45.6 mr = B, for F; = 2550 Ib., practice is to use the cubic mean load (on the assumption that life is inversely
because 5 is the ratio given in most catalogs. Since the basic dynamic load ra proportional to the cube of the load). Thus, if force F, acts for n, revolu-
is F, for B, = 1 mr, use equation (12.1) with k = 3 and get tions (or fraction of a revolution), F, for ng rev., F3 for ng rev., etc., the
Bi cubic mean load F,, is given by
F, = A(= " = 2550(45.6)*/* = 9100 Ib.
F,8n, + Fo8ng + Fs2n3 + °° ] 1/3 cae
Compare with 9070 lb. from Table 12.3, same bearing. (j) Fn =|
un SCL tay é
(b) Preliminary calculations have been made just as those in the example of where Xn = ny + mg + ng + °° * is the total number of revolutions (for
§ 12.8; the equivalent load has been computed to be F, = 23,800 lb. when a constant rpm, time may be used in place of m,, mg, ...). When the
rated capacity is F, = 23,000 lb. for No. 221 bearing. The question is: do we ¢ variation of the load against number of revolutions (or time) is a smooth
risk using the 221 bearing? f curve, the curve may be approximated by a series of constant forces F,, Fo,
Solution. Many things need to be known for a final decision, one of them
etc., laid off in steps to approximate the curve. Or the curve can be inte-
being the probable number of failures in the appointed lifetime. Since F, is the
grated;
rated load for 1 mr., we can find the revolutions B. at the point of 10% fail
for B: = 1 mr. from (Fz = F. = 23,800 lb.)
(k) F,= [7 3
=) 1/3
os r= | 3 lr 1/3

Boe B(4t)’ = (eo. e in xr

2 =a = 0.901 mr.,
where Xr stands for the total time. For example, if the variation of F is
that is, 10% are expected to fail in 0.901 mr. Since statistical logic tends to become Sinusoidal from 0 to Fyax, Fm © 0.65F max from (k).
elusive, the problem may be restated now. We have for Bz = 0.901 the probability
of Pz = 0.9; for Bs; = 1, the design life, what is P,;? Using 6 = 1.34 in equation
(12.3), we get :
12.12 MATERIALS AND FINISH. The most commonly used
In(1/Ps) y (ae i ( 1 ye
material is SAE 52100, an alloy. steel of nominally 1%C and 1.5%
0.1053 Bz 0.901 . chromium, hardened to Rockwell C 58-65. Nickel and molybdenum are
from which P; ~ 0.887, the probability of a particular bearing surviving also used with chromium as alloys. Hardness is important for wear (surface
design life. fatigue); for example, the average life when the hardness is Ro = 50 is
 346 BALL AND ROLLER BEARINGS [Ch. 12 g12.15] TYPES OF ROLLER BEARINGS 347

only about half of that when Ro = 60."1-4! The operating temperature of — manufacturer’s judgment. A bearing No. 206 must be defined as to type,
SAE 52100 and similar steels is generally held to about 300°F (but 200°F © because it may be a roller bearing, a self-aligning bearing, etc. The various
is considered as the maximum value for the usual installation). Tool steel manufacturers use prefix and suffix notations to distinguish their products,
is occasionally used for rolling bearings because its temperature can be but most of them include the standardized numbers. With the exception of
permitted to rise to about 1000°F without it losing too much hardness, ~ the four smallest sizes, the bore in millimeters is equal to five times the basic
Nonferrous metals are used for rolling bearings for a reason; also phenolic bearing number, for example, 5 x 06 = 30 mm, which is the bore of the 06
plastic ball bearings (and other plastics, as nylon, Teflon) are available. — bearing.
Glass has some application for balls; and in exceptionally high temperature
situations, Pyroceram is exceptionally promising. If parts are made of
different materials the coefficients of thermal expansion become important 12.14 FRICTION IN ROLLING BEARINGS. The resistance to
with respect to clearances. motion of these bearings is a combination of rolling resistance and slid-
The diameters of the rolling elements in a particular bearing need to be” ing."?-1-12-17] Tests are generally designed to measure the total frictional
closely the same, say with a tolerance of 50 to 100 microinches, and even torque T;, and a “coefficient of friction’’ is computed from the equivalent
finer for the exacting applications, as in instruments and high-speed — frictional force F, at the bore diameter (F, = T;,/r) divided by the total
situations. When a difference in size exists, the load is not well distributed bearing load (radially loaded). This plan makes these coefficients of friction
among the elements, the larger ones taking excessive stresses. Manufac- comparable with those for sleeve bearings of the same bore. For the rolling
turer’s catalogs show the various tolerances needed by the designer. The bearings, the coefficient of friction is greater at start up than at operating
surface finish is the smoothest possible for commercial processes. See speed, it decreases sharply with increasing load to a point and soon tends
Fig. 3.9. 4 to become constant—except the deep-groove bearing, which seems to have
a minimum value at some load."-*! Stribeck found that fewer and larger
rolling elements tended to reduce the frictional loss at light loads. With the
12.13 SIZES OF BEARINGS. There are several series of bearings for exceptions noted, the following values of f are due to Palmgren,!!2-1)
which certain key dimensions have been standardized, some of which are” determined for the load that gives a life of 1000 mr, and are suitable only
shown in Table 12.4. The 200 series is called /ight, the 300 medium, and the’ for order of magnitude approximations.
400 heavy. Since this classification was adopted, bearings heavier than the Angular contact"??"# M k 4 f = 0.0032
Cylindrical roller, short rollers, flange-guided f =0.0011
Deep-groove, single-row ball bearing. f = 0.0015
Needle bearings (0.0014—0.0022)"%?"2# f = 0.0045
Self-aligning ball bearings A % , f = 0.0010
Tapered roller, spherical roller, flange-guided f = 0.0018
Thrust ball bearings a 4 ; : f = 0.0013
ania rikbuie* FIGURE 12.5 Various Series of Ball The values are higher in new bearings, and when too much lubricant is
__Light _Light_ Light Medium Bearings. All these bearings have the same_ used. Since a lubricant seal, Fig. 12.15, may increase the friction by several
Series Series Series Series basic number, as 09. (After New Depar-—
ture!12-6]),
hundred per cent,"!1-4! these effects must not be overlooked in estimating
(LL00) (L00) (200) (300)
the energy loss. Considering all effects, we observe that rolling bearings do

7 28,
not necessarily have a lower coefficient of friction than full-film sleeve
bearings.

12.15 TYPES OF ROLLING BEARINGS. We can mention only

the principal types. The deep-groove ball bearing, Fig. 12.2, with which we
heavy and lighter than the light have been developed for unusual service— have already become familiar, is one in which the balls are assembled by
rolling mills and fine instruments, for examples. We observe in Table 12.4 the eccentric displacement of the inner ring. With the inner ring in contact
that an 06 bearing has a 30-mm bore in each series. The bearing is made with the outer ring, as many balls as possible are placed in the grooves.
heavier by increasing the outside diameter, Fig. 12.5. The number and size” The rings are then centered, and the balls are kept in position by a separator
of balls or rollers in these bearings is unstandardized and left to the — Or retainer or cage. Bearings loaded in this manner are called the Conrad
 312.15] TYPES OF ROLLER BEARINGS 349

FIGURE 12.6 Single-row Ball Bearing with Shield. The balls in working load, for the purpose of maintaining a nearly constant alignment
this bearing were loaded with the aid of a filling slot. Notice their of parts by reducing the axial movement and, to some extent, the radial
spacing. The shield aids in keeping out foreign matter, important in
rolling bearings. (Courtesy Marlin-Rockwell Corp., Jamestown,
deflection under working loads.
N.Y.). Double-row ball bearings (not self-aligning) are similar to single-row
ball bearings, except that each ring has two grooves. The two rows of balls
give the bearing a capacity somewhat less than twice that of a single row.
Cylindrical roller bearings, Fig. 12.8, are made in different styles and
FIGURE 12.7 Self-aligning Ball Bearing. Self-aligning rolle weights, but the bore and outside diameters are the same as those given in
Table 12.4 for like bearing numbers. Geometrically, the contact is a line
instead of a point as in ball bearings, which results in a greater area carrying
the load and hence, for a particular size, in a larger radial capacity.
The principal function of the retainer in roller bearings is to keep the
FIGURE 12.8 Cylindrical Roller Bearing. The rollers run in 3
groove in the inner ring. (Courtesy Norma-Hoffman Bearing; roller axes parallel. If the rollers should skew, the frictional loss is greatly
Corp., Stamford, Conn.). increased. Skewing may also occur because the roller diameter is not con-
stant or because one end of the roller carries a greater share of the load due
to misalignment.
FIGURE 12.9 Needle Bearing. (Courtesy The Torrington Co. Self-aligning roller bearings, with spherical rollers running in a
Torrington, Conn.). double-grooved inner ring, have curved outer rings that look much like
the outer ring of a self-aligning ball bearing, Fig. 12.7. They can carry
relatively heavy radial and thrust loads and have the usual advantages of
the self-aligning feature. Figure 12.15 is a drawing of this type of bearing.
Cylindrical roller bearings are also made with relatively long rollers,
a popular type being called a needle bearing, Fig. 12.9; this type has no
retainer to hold rollers in alignment. Figure 12.10 shows an application to a
universal joint. If the needles run on the surface of the shaft, rather than
on an inner race, this surface should be hardened and polished for a reason-
able life expectancy. For example, if the surface is cold-rolled shafting with

type. Careful alignment is essential, say a maximum misalignment of 0.5°

This bearing will support a relatively high thrust load. 4
The filling-slot type of ball bearing, Fig. 12.6, has slots or notches tha
permit the assembly of more balls, giving a bearing of larger radial loa
capacity [equation (e)]. Since their thrust capacity is small, these beari FIGURE 12.10 Universal Joint mW [Connection
with Needle Bearings. (Courtesy 2W =
are used where the load is principally radial. cking
Spicer Mfg. Corp., Toledo, O.).
Self-aligning ball bearing, Fig. 12.7, compensate for angulé
misalignments that arise from shaft or foundation deflection or errors i
mounting; they are recommended for radial loads and moderate thrust 1
either direction. Since the outer race has a spherical shape, the shaft ma!
pass through the bearings at a small angle without causing binding.
Angular-contact bearings, so named because the line through thi
areas carrying the load makes an angle with the plane of the face of tht a Rockwell C 15, the capacity for the same life is only 3% of the capacity
bearing, Fig. 12.4, are intended to take heavy thrust loads. They are ofte! for a Rockwell C 60 surface. Needle bearings are appropriate where their
used in opposed pairs and are suited to preloading. Preloading a bearim| Smaller diametral dimensions are advantageous, where the speed is not too
consists of placing it under an initial axial load that is independent of th high, and where there is oscillating motion.
 nr
oe rs
912.17) Ea nee

be made suitable
it tends to provide a better seal against dirt; it easily can
housing
FIGURE 12.11 Tapered Roller Bearing. (Courtesy Timken
for temperatures between —70°F and 210°F. The bearing and its
:
Roller Bearing Co., Canton, O.). should preferably not be packed tight; two-thirds full is enough.
give
At higher speeds, oil becomes the desirable lubricant. (Catalogs

In tapered roller bearings, Fig. 12.11, the rolling elements are frustu ms FIGURE 12.12 Spherangular Roller
i Bearing. (Courtesy Hyatt Roller Bear-
axes intersect at a point on the axis of the shaft. They are capable of carryiny ing Co., Harrison, N.J.).

in standard sizes in English units (inches rather than millimeters) and ing
variety of styles.
The Hyatt ‘‘spherangular’’ roller bearing, Fig. 12.12, isan angular con.
tact bearing with rollers instead of balls. Since the outer race has a spherical
surface, the bearing also is self-aligning. With a large contact area, it ha:
a high load capacity. FIGURE 12.13 Thrust Ball Bearing.
(Courtesy Aetna Ball & Roller Bearing
Co., Chicago).
12.16 THRUST BEARINGS. In the rolling type of thrust beari
the rolling elements are balls, Fig. 12.13, short cylindrical rollers, tape
rollers, Fig. 12.14, or spherical rollers that run in spherical races and
therefore self-aligning. These bearings may be rigidly supported, or on
the races may be supported in a spherical seat to make it self-alig

FIGURE 12.14 Tapered Roller Thrust

12.17 HOUSING AND LUBRICATION. Cleanliness is a prime Bearing.
: (Courtesy Timken Roller Bear-
ing Co., Canton, O.).
consideration in the life of a rolling bearing. For this reason, these bear.
must be protected from airborne particles of matter, as well as from the
expected sources of dirt—during handling, for instance. Housings for this
purpose are usually designed to fit a particular application, Fig. 12.15, o
if appropriate, rolling bearings may be obtained in pillow blocks, Fig. 12.
which have shields to keep the lubricant in and dirt out. The designer should
study the designs found in various manufacturer’s catalogs. SSS Oe
Vi Ss >)
Interference fits for the rings must be tight enough to prevent relative bo \\wl 3
motion but not enough to harm the fit of the rolling elements by excessive Wet Cho

deformation of the rings. The stationary ring is also generally made a lig FIGURE 12.15 Typical Housing, with
Various Seals. Sometimes, an oil slinger
press fit in its housing. The bearing is usually mounted against a should only is sufficient, as at (a). At (b), a
whose height makes it possible to force off the bearing without the fo slinger and a seal; (c) labyrinth seal. :
being transmitted through the rolling elements. 4 (Courtesy SKF Industries, Inc. Philadel- ; Drain to Sump
At low and medium speeds, grease is the most satisfactory lubricant; Phia).

350

Wi

1|
|
(a) — (b) (c)
352 BALL AND ROLLER BEARINGS [Ch. 12

information concerning the permissible limits of grease.) The oil level in FIGURE 12.17 Ball Bushing. For axial motion. The
the housing should be kept low, below the center of the lowest rolling ele palls roll on the reciprocating surface B, circulating
Balls
through tubes. (Courtesy Thomson Industries, Inc., Circulate Out
ment, Fig. 12.15, or excessive churning of the oil at high speed will cause Manhasset, N.Y.). X of Contact into
overheating. At unusually high speeds, oil sprayed as a mist to the point Tube

of loading results in lower coefficients of friction (absence of churning),

lower running temperature. Since the coefficient of friction is nearly con- jubricant needed except where lubricant carries away heat, noise may be
stant for a particular method of lubricating, the temperature rise und objectionable, their life is limited. Sleeve bearings tolerate more misalign-
constant load is nearly proportional to the speed. Bearings that would other. ment than rolling, they occupy less radial but more axial space than rolling
wise run hot may be kept cool by circulating and cooling the oil. Sometimes bearings, and are likely to be cheaper in quantity.“1-5)
it is convenient and simple to lubricate rolling bearings by splash from
nearby moving parts. Prelubricated bearings with integral seals may be
expected to run for the life of the bearing (or machine) without attention 12.20 CLOSURE. One cannot learn what an engineer should know
Fig. 12.16, although means of relubrication are provided in some. The seal about rolling bearings simply by selecting a few bearings from a catalog—
may be on one or both sides. even though this is what he does in the end. This chapter high-lights other
The function of the lubricant in protecting the highly finished surface engineering information of an important nature and gives leads to more of
from rust and corrosion should also be kept in mind. q the same. Reference (/2.22) contains much important basic information.
While it is true that the capacities of a given size and type of bearing are not
all the same when taken from different manufacturers’ catalogs, a large
difference in answers by different catalogs suggests further checking and
more knowledge. And of course, there are the out-of-ordinary cases:
30,000 rpm in textile machinery, 40,000 rpm of woodworking spindles,
; Vg
100,000 rpm for internal grinders and more for small turbines, environ-
FIGURE 12.16 Ball Bearing with Seals. This type is relubricated b ments in space where lubrication may present several unusual problems.
means of a hypodermic-like needle through the small holes on the side
(Courtesy New Departure, Bristol, Conn.).

12.18 OTHER BALL-BEARING DEVICES. By providing means fo!

the circulation of balls, engineers have adapted the principle of rollin
contact to special situations. Application to a screw is shown in Fig. 8.2:
p. 250. Figure 12.17 shows a ball bushing, to be used for reciprocatif
motion."?-2°] A circulating ball adaptation for spline connections is als
available.?-181

12.19 COMPARISONS OF SLEEVE AND ROLLING BEARINGS

For sleeve bearings with full-film lubrication, the speed is limited
temperature rise (which in turn is a function of the lubricant), the ste
friction is high, the damping is relatively good, large amounts of lubri
are needed, noise is no problem, their life is unlimited. For rolling beari
the speed is jimited largely by dynamic considerations (vibration), tht
starting friction is low, the damping effect is poor, very small amount 0}

353
 13. SPUR GEARS

13.1 INTRODUCTION. Spur gears are toothed wheels whose tooth

elements are straight and parallel to the shaft axis; they are used to transmit
motion and power between parallel shafts. Kinematic considerations, as
well as the terminology, are essential for design decisions, but we must
assume that the reader is acquainted with these matters; a few pages of
review are included.

13.2 DEFINITIONS. See Fig. 13.1 on page 356. The pitch circle is the
basis of measurement of gears. The size of a gear is its pitch-circle diameter
in inches, called the pitch diameter. For interchangeable gear teeth, it is
intended that the pitch circles of mating gears be tangent, in which case,
these imaginary circles roll on one another without slipping. Gears
may not be so, mounted; then there may be said to, be an operating
Pitch circle and a standard pitch circle. The pitch circle is also the
trace of the pitch cylinder (pitch surface) as it intersects a plane normal
to the axis.
The pitch point of meshing gears is the point of tangency of the pitch
circles; for an individual gear, the pitch point will be located where the
tooth profile cuts the standard pitch circle.
The addendum circle (also outside circle) is the circle that bounds the
Outer ends of the teeth. The addendum cylinder encloses a gear. The

355
 133] BASE CIRCLE AND PRESSURE ANGLE 357
The flank is the surface of the tooth between the pitch and root cylinders.
The top land is the surface of the top of the tooth.
The bottom land is the surface of the bottom of the tooth space.
When two gears are in mesh, the smaller is called the pinion and the

1 jarger the gear.

The angle of action is the angle through which the gear turns from the
time a particular pair of teeth come into contact until they go out of con-
ae at ee ee tact. The are of action subtends the angle of action.
Root or DendencYim’ oF The angle of approach is the angle through which the gear turns from
the time a particular pair of teeth come into contact until they are in
contact at the pitch point.
The angle of recess is the angle through which the gear turns from the
time a given pair of teeth are in contact at the pitch point until they pass
FIGURE 13.1 Dimensions and Nomenclature, Gear Teeth. Full-fillet roots are sometin e
out of mesh. The angle of approach plus the angle of recess is equal to
used for severe duty because of the reduced stress concentration—as shown dotted at E
the angle of action.
The velocity ratio m,, is the angular velocity of the driver divided by
diameter of this circle is called the outside diameter. The addendum a i the angular velocity of the driven gear. For spur gears, this ratio varies
the radial distance between the pitch circle and the addendum circle. _ inversely as the pitch (or base-circle) diameters and as the tooth numbers.
The dedendum circle or root circle is the circle that bounds the bottom:
of the teeth. The dedendum d is the radial distance from the pitch circle te
the root circle, that is, to the bottom of the tooth space. pan antes)
1p! Monon Daas lat
» m a) Ne D, N,’
The whole depth is equal to the addendum plus the dedendum.
working depth is the radial distance from the addendum circle to where n is angular velocity in, say revolutions per minute, the subscript 1
working-depth circle which marks the distance that the mating to refers to the driver, and the subscript 2 refers to the driven gear.
projects into the tooth space; it is the sum of the addendums of mat n The gear ratio m, is the number of teeth in the gear divided by the num-
gears. ber of teeth in the pinion. When the pinion is the driver, m, = m,,.
The clearance is the radial distance between the working-depth cirel
and the root circle; it is the dedendum minus the mating addendum.
The circular thickness, also called tooth thickness, is the width of to 13.3. BASE CIRCLE AND PRESSURE ANGLE. The base circle
measured along the pitch circle. The chordal thickness is the tooth wi is the circle from which the involute is generated. In gearing, the expression
measured along the chord at the pitch circle. The width of space or to degree of involute is used to define the base circle for a particular pitch circle.
space is the space between teeth measured along the pitch circle. The line 4A, Fig. 13.2 (p. 359), is tangent to the standard pitch circle. If,
Backlash is the tooth space minus the circular thickness. When backl through the point of tangency, a line BB is drawn, the corresponding base
exists between two gears, one gear can be turned through a small a Circle is tangent to BB with its center at the center of the pitch circle. The
while the mating gear is held stationary. Backlash is necessary to care involute obtained from this base circle and used for the interchangeable tooth
errors and inaccuracies in the spacing and in the form of the tooth Profile is a ¢,-degree involute. An involute generated from the base circle tan-
provide a space between the teeth for lubricant (§ 13.38), and to allow Sent to the line at ¢. degrees with AA would produce a tooth profile of ¢o-de-
the expansion of the teeth with a temperature rise. Accurately cut gé Sree involute. The angle ¢ is called the pressure angle, because the direction of
should be mounted with a backlash from 0.03/P, to 0.04/P,, where the force acting normal to the profiles is along the line BB (for ¢,) when two
is the diametral pitch. Use the larger values for higher speeds. To pro Mvolute teeth are in contact with the pitch circles’ of the mating gears
backlash, the cutter generally is fed in somewhat deeper than standar tangent to each other. The angle ¢ is also called the angle of obliquity.
depth on the larger of the two gears. The line BB is the line of action; that is, all points of contact between
The face width is the length of teeth in an axial direction. 7 Mating teeth properly lie somewhere on this line, which is also the generating
The face of the tooth is the surface of the tooth between the pitch cylindél me. It is the nature of involute gearing that the line of action is always
and the addendum cylinder.

356
358 SPUR GEARS [Ch. 13
perpendicular to the tooth profiles at the point of contact. Two gear teeth
in mesh must have profiles with a common generating line. 3
The operating pressure angle is determined by the center distance, URE 13.2 Degree of Inyolute. . Just as two in-
because once the profiles are established, involute gears may be moved nee gears can operate at different pressure angles,
away from each other, thereby increasing the backlash, and still operate * so-called 20° cutter can cut gears of pressure angle
other than 20°.
correctly with no change of velocity ratio. Since the generating line re~
mains tangent to the base circles, the effects of increasing the cen Base Circle for
distance are to increase the pressure angle and to move the pitch circles Base Circle for Involute of @: Degrees
of the gears apart. For spur gears, 20° is the preferred pressure angle; but Involute of ¢2 Degrees
a 143° angle was most widely used for years. When we say a gear is a 20°
involute gear, we mean that the pressure angle will be 20°, provided that
in ae
the mating gears are mounted with their pitch circles tangent to each other The base pitch P, (involute gearing only) is the distance
to the nd .
and that ¢ = 20° defines the base circle from which the involute profile is measured along the base circle from a point on one tooth
obtained in accordance with Fig. 13.2. ing point on an adjacent tooth. It is also the distance me para
i
profiles on adjacent teeth measured along the generating line;
7D, 7Dcos¢ ulpolioa
13.4 PITCH. The pitch of a gear is a measure of the spacing, a (13.4) hy = —————- = c ?

usually also of the size, of the tooth. In this country, there are two pite N, g N, g
in common use, the circular and the diametral (di-am’-e-tral).
where D, is the diameter of the base circle and D is the pitch diameter.
The circular pitch P, is the distance in inches measured along the pi
In spur gearing the pitch angle is the angle subtended by an arc on the
circle from a point on one tooth to the corresponding point on an adjac
pitch circle equal in length to the circular pitch.
tooth, Fig. 13.1. Let D be the diameter of the pitch circle and NV, be
number of teeth in the gear. Then the circumference of the pitch circle
7D divided by the number of teeth N, is the circular pitch Ps URE 13.3 Sizes i of Gear Teeth of Different Diametra 1 Pitches—Full Size (144°).
if ok > 20, the teeth are classified as fine pitch, with some proportions different from
those of larger teeth. (Courtesy The Barber-Colman Co., Rockford, Ill.)
(13.1): ping Mee WN, a

The diametral pitch P, is the ratio of the number of teeth per inch of s
pitch diameter, or

(13.2) Piswects
In Fig. 13.3, it is seen that the size increases as the diametral pitch decreas
Conversion from circular to diametral pitch is frequently necessary.
From (13.1) and (13.2) we have 4

(13.3) Pipe (=) (=). opin pipe

g
a

Circular pitch is regularly used for cast-tooth gears as a convenience ae

et

the pattern maker, though cutters of many circular pitchesareavailablefrom

stock. For cut teeth, the diametral pitch is nearly always used. Therefore,
in specifying the diametral pitch, we should try to use one for which cutters:
are available from stock. One manufacturer’s list of standard pitches will
not necessarily be the same as another’s; for problem purposes select one
of the following standard pitches: 2, 2.25, 2.5, 3, 4, 5, 6, 8, 10, 12, 16.
 Involutes
360 SPUR GEARS [Ch. 13

13.5 LENGTH OF ACTION AND CONTACT RATIO. The length

of action Z is the length of that part of the line of action between the initial
and final points of contact and is given by

(b) Z = (aa? — Poa")? + (Fa17 = 79:7)*? — Csing,

[NO INTERFERENCE]

where r,; and ro are the radii of the addendum circles of the mating gears,
Tp, and rpg are the base-circle radii, and C is the center distance; this Pressure Angle

equation is applicable only when the intersections of the addendum circles Pressure Angle -
and the line of action lie between the interference points."°-2! The inter= o Involute
ference points are the points of tangency of the line of action and the base Interfering or Unused »
Portion of the Tooth VB
circles, Fig. 13.4.
The contact ratio m, can be defined in several ways: it is the ratio of the
~ -—>
angle of action to the pitch angle; the ratio of the /ength of action to th —~

base pitch; the ratio of the arc of action to the circular pitch. It may be
thought of as the average number of teeth in contact for mating gears. If thine Base Circle
possible, the contact ratio should be about 1.25 to 1.4 for best runni
conditions. A contact ratio less than unity means that one pair of teeth Ta sig Tb Td 2 Point J2 is the
Interference Point
out of contact before the next pair have reached their initial point of con: ge

tact, and the drive is not continuous. For example, two 12-tooth, full-depth
o mi \ : (a)
144° involute gears have a contact ratio slightly less than one, but for low
FIGURE 13.4 Action of Gear Teeth. The base pitch P, = NQ = N’Q’, as seen upper
speeds, the operation may be satisfactory. left in (b). The radii of curvature of the involute profiles at the pitch circles are: for the gear,
The contact ratio can now be computed from Cp; for the pinion, Jip.

Z Z
©) m, = —P, = ——_,
P,cos¢ this nonconjugate engagement, the base of the pinion tooth must be under-
cut, or the face of the gear tooth must be relieved. Both methods are used,
so that the teeth should not touch one another until point J, is reached.
Contact will cease where the addendum circle of the driving gear inter-
13.6 LAW OF GEARING AND THE ACTION OF GEAR TEE
sects the line of action, point B, Fig. 13.4, when the face of the driver is
The law of gearing states that for a pair of engaging teeth to transmit
in contact with the flank of the driven—provided that this intersection
constant velocity ratio the tooth curves must be such that the comma.
lies between the points of tangency J, and J, of the line of action and the
normal to the profiles at the point of contact will always pass through t
base circles. Since point B satisfies this provision, it marks the point where
pitch point."°-2) The law of gearing does not refer to the over-all ratio, bu
a pair of teeth leave contact.
to variations during the engagement of a pair of teeth. These momenta!
variations in the velocity ratio result in accelerations and decelerations
that are manifested in vibration and noise. Curves that satisfy the law
gearing are called conjugate curves, of which any number may be obtained. 13.7 INTERFERENCE OF INVOLUTE TEETH. The points of
Contact of a pair of teeth starts between the driver’s flank and the tangency of the line of action and the base circles, J, and J,, Fig. 13.4, are
driven’s face, where the addendum circle of the driven gear intersects the called the interference points. The nonconjugate action referred to in the
line of action—provided that this intersection lies between the points of Previous article is called interference; when the gear tooth is long enough
tangency (J, and Jp, Fig. 13.4) of the line of action and the base circles. to project inside of the base circle of the pinion, the tip of the involute
In Fig. 13.4, the addendum circle of the driven gear intersects the line of Profile overlaps the radial flank of the pinion tooth. Interference is a serious
action at M, which is outside the tangency point J,. That part of the tooth disadvantage of involute gears; it is a maximum when the smallest gear
on the pinion inside its base circle is drawn as a radial line, and it happens Mates the largest. The amount of interference decreases as the gear de-
that this line is not conjugate with the involute face of the gear. To prevent creases in size (for a particular pinion) or as the pinion increases in size

361
362 SPUR GEARS [Ch. I3_

(for a particular gear). In the 144°, full-depth system, § 13.8, a gear with FIGURE 13.5 Comparison of Tooth Profiles.
32 or more teeth will have no interference with a rack (or any other gear), Tooth profiles as they appear on 20-tpoth gears;
(a) is a 144° full depth, (b) is a 20° full depth, (a) (b)
In the 20°, full-depth system, a gear with 18 or more teeth will not have
and (c) is a 20° stub.
(c)
interference with a rack; for a 20° stub tooth, the corresponding number.
of teeth is 14. Thus, increasing the pressure angle and decreasing the height
of tooth decreases the interference problem.
Gear teeth generated by involute rack cutters are automatically under-
cut in the flank, enough of the interfering portion being removed to elimi-
nate actual interference. While this solves the interference problem, the
tooth is thereby weakened, and the contact ratio may become undesirably -
low. The best course is to avoid the condition of theoretical interference
unless there is some outweighing advantage. The 20° full-depth teeth have
the advantages of greater capacity and less interference trouble. Where
conditions warrant the higher cost, noninterchangeable teeth, unequal
addendum and dedendum, § 13.34, may be used to avoid interference.

13.8 INTERCHANGEABLE INVOLUTE GEAR SYSTEMS. A 1

(a) (b) (c)
interchangeable gear-tooth system is one wherein any gear of a particula
FIGURE 13.6 Forces on a Tooth.
tions for interchangeability are: (1) all gears have the same pitch; (2) all
gears have the same addendum, which is equal to the dedendum minus gearing, and its line of application moves from the upper (or lower) part of
the clearance; and (3) all gears are cut with the same angle of obliquity. the tooth to the lower (or upper) part; W, F,, NV, Fig. 13.6, are assumed to
The following tooth proportions are in use. | be uniformly distributed along the length of tooth. Considering the tooth
(a) Full-depth Involute System. The recommended dimensions for as a cantilever beam, we find the stress at its maximum when one tooth
full-depth interchangeable teeth are: '% carries the entire load at its tip, as in Fig. 13.6(a). However, if the contact
Addendum a =P, — Pla Working depth = 2/P, ratio is greater than one and if the teeth are geometrically accurate, another
Clearance (min.) = 0.25/P, Whole depth =| 22 00Re tooth is sharing the transmission of power when the tip of this tooth is in
Dedendum d = 1.25/P, Outside diameter = D + 2a contact. As the tooth in Fig. 13.6(a) moves through its angle of action, the
The AGMA recommended standard pressure angle is 20°; 25° is someti point of application of W moves down the profile. At some stage of this
used; most full-depth gears now in service are 144° involute. Somewhat motion, with a contact ratio less than 2, the tooth will carry the entire load
greater clearance, say 0.35/P,, may be needed for tool clearance on shave d as in Fig. 13.6(b).
or ground tooth profiles. Compare teeth in Fig. 13.5. At the point where the line of action of W cuts the center line of the
(b) Stub-tooth System. The usual tooth proportions are (ASA): tooth, Fig. 13.6, force W is replaced by its radial and tangential com-
Working depth = 1.6/P,, Clearance (min.) = 0.2/P,, ponents, NV and F,. The force N produces a uniform compressive stress over
Addenduma = 0.8/P;, Whole depth =.1.8/P,, any section of the tooth, say at VE. The component F, produces a bending
Dedendumd = 1/P, = P,/z Pressure angle = 20°. Stress: tension at E and compression at V. The uniform compression at E
Full-depth teeth are favored unless the interference problem is especially Subtracts from the bending tension at E, producing a lower, and therefore
severe—few teeth on the pinion, many on the gear—and an interchangeable Safer, net stress at E. The uniform compression at V adds to the bending
system, rather than unequal addendum and dedendum gears, is desired. Compression at V to give a higher total compressive stress. If the material
1S stronger in compression than in tension, as is cast iron, the effect of the
Compressive stress due to N is to strengthen the tooth. Since the uniform
13.9 STRENGTH OF GEAR TEETH. Without friction, the resultant Compressive stress is small compared to the bending stress, its effect on the
force W acting on the gear tooth lies on the generating line in involute Strength of tooth is usually ignored.
With F, acting through B with a moment arm hd, the bending moment at

363
 364 \ SPUR GEARS [Ch. |3 g13.10] STRESS CONCENTRATION 365

the section VE is M = F,h. With the face width bd, Fig. 13.1, the section of teeth for a particular ¢) and the point of application of the load. It is
modulus of the rectangular section at VE is Z = bt?/6. From M = sZ, known as the Lewis form factor and designated by Y; Y = 2xP,/3. Use Y
in equation (f) and get the equation known as the Lewis equation, after
(d) Fh = er Wilfred Lewis who first derived it in 1893;

The section VE is to be the one where the stress is a maximum from the sbY
(13.5) FaS or F, = sbP.y, [LEWIS EQUATION]
load F, and is located by the following reasoning. Let s and b in equation (d) © Pa
be constant and get h = (sb/6F,)t? = Ct?, which is the equation of a para-
bola. Inscribe this parabola through point B; it is tangent to the tooth | where we have used P, = 7/P, and Y = zy; y is the form factor for use
with circular pitch P,, but our Table AT 24 gives values of Y because cut
teeth usually have a standard diametral pitch that makes Y more convenient.
Values of Y with the load at the tip of the tooth and with the load near
the middle of the tooth are given in Table AT 24. For standard interchange-
able teeth, a 20-tooth gear of one pitch has the same value of Y as a 20-
tooth gear of any other pitch. If other tooth forms are used, construct the
outline of the tooth to large scale, insert the line of action of the load at the
desired point, measure x, Fig. 13.6, and then find Y = 2xP,/3. Also see
Ref. (13.1).

13.10 STRESS CONCENTRATION. Since there is a reentrant

corner at the bottoms of the teeth, where the profile joins the bottom land,
a stress concentration exists, Fig. 13.7, a phenomenon ignored by the
(a) (b) () original Lewis equation. Inasmuch as the load on the tooth is repeatedly
applied, we should expect that if breakage occurs, the failure would be due
FIGURE 13.6 (Repeated).
to fatigue. The value of the theoretical stress concentration factor is not
simply defined because of the complex geometry. If the teeth are cut by a
profile at V and E (where the tooth width is ¢) and it will be found to be generating method, the fillets are not arcs of circles, but trochoids, whose
inside the tooth elsewhere. Since the parabola outlines a beam of uniform form in turn depends on the number of teeth being cut and the radius of
strength (s = constant), the tooth is weakest at the section VE. (This
argument is not strictly true because the stress-concentration effect results
in the maximum stress being in the fillet.) The inconvenient dimensions 4
and t can be replaced by using the proportion, from the similar triangles’
BVG and GVH (corresponding sides perpendicular), x/(t/2) = (t/2)/h, or
h = t?/(4x). Substitute this value of h into equation (d) and obtain FIGURE 13.7 Stress Concentration in Gear Teeth. (Courtesy T.
J. Dolan, Univ. of Iil.).
(e) F,t? bt? 2 ne
r ae ee
= s— or = S. ‘.

Multiply and divide the right hand side of this equation by the diametral
pitch P, and find i
‘ r= >(-34) Curvature of the corner of the cutter tips for which there is no single
f) tonite Mio acstik Standard ; K;, also depends on the position of the point (line) of contact,
Fig. 13.6(a) and (b)."%-16) Because of these complications, the stress concen-
Consider the parameter 2xP,/3 with respect to Fig. 13.6 and note that its” tration factors used are ordinarily reasonable estimates of the true values.
magnitude depends upon the shape of the tooth (a function of the number Black,"3-41 ysing photoelastic technics, found 1.345 < K, < 1.47 on the
 Geer, "Sea ee eS Gee
366 SPUR GEARS [Ch. I:
note that reversed bending as produced by a reversed load does not neces-
tensile side and up to K, = 1.61 on the compression side. Most authoritie load.
sarily result in the same strength as a rotating beam with a constant
seem to agree that the fatigue crack propagates from the tensile side. Do given in our tables for the design stress in order to
At any rate, use s, as
and Broghamer"™®-5 devised the following equations that accord with thej load
agree with those recommended by Buckingham and with the dynamic
only
to be defined later. If test values are not at hand, recall that for steel
photoelastic studies:
t 0.2 t 0.4
(g) K, = 0.22 + (-) (;) for 144°, (i) s,, & (250)(BHN) psi or (BHN)/4ksi, but not over 100 ksi.
r h

t 0.2 t 0.4 The beam fatigue strength of carburized teeth, favorably affected by the
nce
(h) K, = 0.18 + (-) (;) for 20°, high surface strength and the residual compressive stresses as a conseque
+ 0.25, core ksi with a
r
of the hardening operation, may be taken as 60
strength of the core,
where r is the minimum radius of fillet, h is the height of load above th tolerance of +10%;'-) sy core ksi is the ultimate
11. Use
section VE, Fig. 13.6, and ¢ is the thickness of tooth at this section which can be estimated as BHN/2 ksi if necessary. See Table AT
the lower range of values for maximum reliability; as usual, a maximum
general, the theoretical values should be modified by the notch-sensi
factor q, § 4.10, but when the gear teeth are hardened, the value of gis value of 100 ksi is preferred. In general, the use of the core BHN in equation
unity, except for very small radii (Fig. AF 7). With a strength redu (i) is conservative practice when the fillets are carburized; and this value
factor K,, Lewis’ equation becomes may be used even if it is larger than the 60 + 0.2s, core and it is advised for
flame-hardened teeth, but for flame hardening, one must be sure that the
ae fillet areas do not become annealed and therefore weaker. Since gears are
(13.6)
sbY
;
K,Pq so often heat treated, read now §§ 2.6-2.9, 4.23, 4.33, 4.28-4.31, inclusive
also check § 13.24.
We shall use both forms of the Lewis equation, (13.5) and (13.6 For bronze and alloy cast irons, s, ~ 0.45, may be used in the absence
Buckingham states that if the load is assumed to be acting at the tip of tl
of better information. See also § 13.24. If a tooth is subjected to loading in
tooth [and comparison is made with his dynamic load (§ 13.17)], this is s both directions, as the teeth of an idler, some engineers reduce the design
conservative that no strength reduction factor K; is needed; when the va stress by 25-30%.
of Y is chosen for the load near the middle of the tooth, Table AT
K; should be included. When two gears are of the same material, the pi
tooth is weaker, Y, < Y,. When the materials are different, assume 13.12 FACE WIDTH. The derivation of the Lewis equation is based
tooth with the smaller s Y to be the weaker (§ 13.11). The value of K; sho on the assumption that the load is uniformly distributed across the face
fall within the range of 1.2 to 1.7 when the load is applied at the tip; bul width. Sometimes this is far from true, due to misalignment or warping of
K; is higher, say 1.4 to 2, when the load is applied near the middle. Th the teeth. One cause of tooth breakage is the concentration of load on one
AGMA gives a chart of values of Y/K, for 20° full depth, cut with a certaii end of a tooth, which results in higher stresses than when the load is dis-
rack cutter. tributed. To minimize this kind of trouble, the face width b should not be
too great as compared to the thickness (or pitch) of the tooth. In the absence
of special considerations, the following proportions are considered good:
13.11 DESIGN STRESS. In general, a design stress is chosen f 8 12.5
agree with experience and the choice is affected by how closely the stres: (jj) 25P,<b<4P,, oF —<b< * cut teeth.
analysis accords with the facts of life and by the assumed loading condition Pg a
A common attitude is to use the endurance limit as the design stress, whicl There are many exceptions to these proportions. For example, automotive
then corresponds to the expected maximum load, to be called the dynami transmission gears have shorter faces because of the need for a compact
load (§ 13.14). Buckingham™*-*! says to use the endurance strength for arrangement. In general, the longer the face and the more rigid the material
repeated load (R = 0) and says that this is 1.5 times the endurance limi of the teeth, the more accurate should be both the tooth profiles and the
5,3; however, he uses the endurance limit as s,/3, whereas we are takin alignment of the shaft for long life and trouble-free operation. In order to
as $,,/2. Hence, for steels, Buckingham’s, endurance strength in one di avoid a concentration of the load on one end of a tooth, spur-gear teeth
tion is the same as that which we have defined for reversed bending; bul
368 SPUR GEARS [Ch. 13 §13.15] DYNAMIC LOAD AS A FUNCTION OF VELOCITY ONLY 369

are sometimes ‘“‘crowned,” Fig. 13.8; that is, the teeth are shaved with initial accuracy of the teeth. (See § 13.38.) Buckingham points out that the
an elliptical reduction from the center of the tooth to the end of about dynamic load in this case approaches asymptotically from above a value
0.0003 in. per in. on each side.3-1 of twice the transmitted force; that is, as the speed increases, the dynamic
load peaks and then decreases with increasing speed, because the point is
reached where there is not time enough for the teeth to bounce apart as they
do at slower speeds.
The maximum value of the dynamic load depends on the masses of the
FIGURE 13.8 Crowned Tooth. The amount of crowning ig
gears and connected masses (F = ma), the speed of operation as pointed out
exaggerated in order to highlight the idea. (Courtesy Nationa
Broach and Machine Co., Detroit.) above, and the materials. Materials with a low modulus of elasticity deflect
more than high-modulus materials, other things being equal, and therefore
absorb the energy of impact with a lower peak load.
If we can compute a dynamic load that is a conservative estimate of the
maximum load, then we are justified in using the endurance strength as a
13.13 THE TRANSMITTED LOAD. In design we usually know th design stress, as explained in a previous article, when stress concentration
transmitted power and the angular speeds of the gears. The transmit effects are cared for. As stated above, the methods of determining the load,
Joad F,, the average tangential force on the teeth, is then obtained from th design stress, and the unknown dimensions must be known to result in
horsepower; or from the applied torque. Although the applied force va satisfactory designs. A number of equations have been derived for computing
a little as the point of application moves from the top to the bottom of the dynamic load,"3-3-13-12.13.13] byt first we shall review a few traditional
tooth, or vice versa, the nominal force acting at the pitch circle is used i and easier methods of estimating it.
design; thus!?-15!

33,000 hp
13.15 DYNAMIC LOAD AS A FUNCTION OF VELOCITY ONLY—
ES Um ?
METAL TEETH. After Lewis introduced his equation for the strength
of gear teeth, experiment showed that a “‘velocity factor’ needed to be
where v,, = 7Dn fpm, the pitch-line speed; n rpm; D ft.; and hp is h applied in obtaining a design stress in order to get a better agreement be-
input horsepower. In the case of a variable transmitted load, as in pune tween design computations and test results, which is the same thing as
presses, shears, etc., the maximum transmitted load should be the basis ° saying that the dynamic load is a function of velocity only. It is not, but
the design for strength. it is so nearly so over limited ranges of speed and for a particular class of
gears that velocity factors are still widely used. When the service is intermit-
tent and wear is not a factor, the use of the following values of the dynamic
13.14 DYNAMIC LOADS ON GEAR TEETH. Since gear to loads F, for metal teeth, compared with the beam fatigue strength computed
profiles are not perfect involutes (not precisely conjugate), since the to from equation (13.6) with K,, should give satisfactory designs; v, fpm =
spacing is not exactly right, since the shaft and mountings deflect un Pitch-line speed, F,= transmitted load from equation (1.15):
load, and since a load will deflect teeth out of shape even if they are in
perfect, the law of gearing is not ideally satisfied and local accelera 600 + Um Commercially cut
inevitably occur. At what are considered normal speeds, two peak lo
(k) Fa = wer ia Ib. tee = 2000 fpm
occur with each tooth engagement, the first accompanying accelerati
due to the foregoing imperfections. The driving gear slows down, the dri : HS 1200 + »'m
) arefully
Carefull cu
speeds up, and likely the teeth momentarily cease contact. Then the acting Fa Stage ele, {To00 < Um < t 4000 fpm
forces bring the teeth together again with an impact that results in a dynamit
load considerably greater than the transmitted load. Thus, the maxim 1
load acts shortly after the initial engagement and occurs near the middle B+2,/" Precision cut
(m) PaisTigges Ib. ( > 4000 fpm
of the profile.
At high speed, the phenomenon changes. The time of engagement If the speed is over 2000 fpm, carefully cut teeth should be used; if vu, >
becomes so short that there is only a single peak load, regardless of the 4000 fpm, the teeth should generally be precision cut; the corresponding
370 SPUR GEARS [Ch. 13 13.17] BUCKINGHAM’S AVERAGE DYNAMIC LOAD FOR METAL TEETH 371

production methods are described later.* Gear teeth are assumed to be (§ 13.15), so let us assume that the teeth will be carefully cut. Then the dynamic
strong enough against fatigue failure when the strength F, from (13.6) is load is estimated from equation (1) as
equal to or greater than the above dynamic load, F, 5 Fy with Y for the F, = 1200 + vm , _ (1200 + 1505)(438)
load at the tip; how much greater depends on the service (§ 13.18). = 988 Ib.
of igo. 1200
Velocity factors are still widely used, with AGMA sanction, probably
With interchangeable type teeth and the same material, the pinion tooth is
because it is so much easier to design with them than to consider the
the weaker. Part of the designer’s responsibility is to decide upon a material.
elasticity and inertia of the whole system, and there are many years of We might try a cast iron for both gears since it is an inexpensive material; say
experience with their use. For types of service and ranges of capacity fo: ASTM 25, whose endurance limit may be taken as (0.4)(25) = 10 ksi. Table AT 6
which they are known to apply, they give satisfactory results, but as is gives the typical s, = 11.5; hence, 10 ksi is conservative. In Lewis’ equation
seen from Fig. 13.9, it is not prudent to extrapolate into a region where it is (13.6), there are still four unknowns: b, Pa, Y, K;; hence, it must be solved by
not known that they apply without confirming tests. The AGMA reco m trial. Let Ky = 1.48 (this should be fairly close for the load at the tip); let b =
mends the use of equation (m) for high-precision shaved or ground spui 10/P., which is nicely within the generally desired range (§ 13.12). When equation
gears, and (k), (1), or (m) is used for the dynamic load, the traditional assumption is that
one tooth may take the full load on the tip; therefore, assume an appropriate
50 + 0,1? value of Y for this configuration. Since Y does not vary markedly, some reasonable
n Resa ee value, as Y = 0.32, Table AT 24, is suitable for the first approximate solution.
(n) . BOT aH Shaved With these various assumptions,

for commercially hobbed or shaved teeth. But in addition to a se sbY _ (10,000)(10/P.)(0.32)

F, = Fy = 988=
factor, § 13.18, there is recommended a load-distribution factor™?-1! w K;P 1.48P, z
value ranges from 1.3 to more than 2.2; also other factors. Thus, from which Pa = 4.68. If we assume that readily available cutters have a pitch
Fig. 13.9 indicates that these equations (m) and (m) tend to give low of 4 or 5, § 13.4, we must now make a choice of one of these pitches and check
mates of F,, the use of additional “factors” tends to even things u the strength; choose 5, the nearest. Perhaps as good procedure as any is to solve
should be observed that the velocity factors take no account of the di for the face width 6 that makes the strength equal to the dynamic load. For
ences of materials, whereas Buckingham’s work (§ 13.17) indicates N, = PaD, = (5)(5) = 25, we find Y = 0.34, Table AT 24, use K; = 1.48 as
for a particular quality of manufacture, the dynamic load for steel gea before;
much larger than for cast iron (Ext, > £c.1)- sbY ni (10,000)(6)(0.34)
F, = 988 =
K;Pa (1.48)(5) F
from which b = 2.15; use 6 = 2}in. Check the proportion; bP. = (2.25)(5) =
13.16 EXAMPLE—SPUR GEARS, INTERMITTENT SERVICE. Design 11.25, which is within the range of 8 to 12.5 and is therefore satisfactory. For
pair of 20° full-depth involute gears to transmit 20 hp at 1150 rpm. The dia m., = 2.5, we get N, = (2.5)(25) = 62.5; use 62 teeth in the gear. (If an exact
of the pinion is to be 5 in. and the velocity ratio m,, should be about 2.5; sm velocity ratio of 2.5 were essential, the pinion would have to have an even number
intermittent service (wear not considered). Let the dynamic load be taken 4 of teeth.) A solution therefore is
function of velocity only.
Solution. The pitch-line speed and transmitted load are Pa =5, b = 2¢in., Np = 25, N, = 62.
A single series of computations as given above is not likely to give the best design.
ye eda (5) 1150 = 1505 fpm, In this case, it had already been determined that cast iron could be used without
having too few teeth on the pinion. The designer should make a number of designs
Considering different materials and pitches and sizes, all satisfying the mathe-
matical requirements, and then choose the one that is most economic for the
Purpose.

* To those who used the previous edition: the velocity factors were originally adjuste¢
in this text to match the design stresses recommended by Buckingham. Since the velocity 13.17 BUCKINGHAM’S AVERAGE DYNAMIC LOAD FOR METAL
factors are generally quoted in their original form, we have conformed. The introductior TEETH.
of the fatigue strength reduction factor results in virtually the same results.
The dynamic load is considered to be made up of the
 sg E Suckingn c--] aa buckinghams |
Sl 6b Sree eo Ne ~ 600 Detailed Analysis
il C e=0. C L _
372 SPUR GEARS [Ch. 13 f& & Cast Iron Pr -———[4¢=0.001’, C.1,
ee 5
£ eg = ———=|

2 5
transmitted load F, plus a dynamic increment load I, the increment 2 E ie if i “\ Buckingham’s
2 CE i“ iA att a Average,
being the consequence of the various inaccuracies and the accompanying e=0.001', C.1.4
iia E S a E 1200
+ Um
accelerations. Since its magnitude depends upon the masses of the gears,
a simple equation purporting to give the dynamic load can apply only to” 3 3 E “7 er 1200
° r Be 4 Lo 50+
a certain class of gear. Thus, for gears of average mass with average gle sea
& of i of) a asin i
connected masses, F, = F; + J,
B y ee a ee i GI 78+ vow
0.05,(bC + F,) & 1 78
(13.7) F,= g
0.05v,, + (bC + F,)'/? 0 i 1 1 1 L 1

where v,, fpm is the pitch-line speed, F; lb. is the transmitted load, b in, 1 2 3 4 5 6

is the face width, and C is a function of the amount of the effective error Um x 10-* fpm
and of elasticity of the gear materials;
FIGURE 13.9 Dynamic Load by Different Equations. Buckingham’s detailed analy-
sis,"13-3] curve A, is for a constant transmitted load of F; = 1000 lb. and the following
kE,E, 5
(0) Cc = ——_, [raBte AT 25] data: 6 = 5in., 20° full depth, cast-iron gears, e = 0.001 in., Np = 30, my = 4, con-
E, +. Ep nected mass on pinion shaft and the pinion mass are 16 and 2.5 slugs, connected mass on
gear shaft and the gear mass are 40 and 13 slugs, respectively. The dotted curves B and C
where k = 0.107e for 144° full-depth teeth; K = 0.11le for 20° full-depth are for Buckingham’s average equation (13.7), other data the same. The velocity factors are
teeth; k = 0.115e for 20° stub teeth; e = effective or composite tooth erro: of course independent of the elasticity of the materials and the face width. Note that curve
which should be smaller at the higher speeds, Fig. AF 19 (choose the desig C for cast iron and e = 0.002 in. is the same as would be obtained with steel and e = 0.001
(see Table AT 25). The values of Fa from curves A, B, C have the relations shown with
value from Fig. AF 20; see § 13.20); and E, and E, are the modulii
those from C, D, E, F, only for F; = 1000; the relative positions of these curves will in
elasticity of the materials of the gear and pinion, respectively. general be different for other values of F;. Similar studies could be made with another
Gears that would not be average are, for example, aeronautical ge independent variable, as Fz plotted against power with constant speed.
where the moving masses, considering the loads, are less than average an
gears connected by short shafts to heavy flywheels where the masses
alternative is to introduce an experience factor, in essence an increase of
stiffness are greater than average. Also, small gears transmitting low pow
the “factor of safety” to cover ignorance. Let N,, represent this service
are not average; nor are high-speed, lightly-loaded gears. The predomin
factor; then F, 5 N,,F, in order for the teeth to be safe against breakage.
influence on larger gears at moderate speed is the connected masses. But fi
Both the driving machine (electric motor, gasoline engine, etc.) and the
small gears, especially on small shafts that easily twist through an ang
driven machine affect the choice of service factor. The service factor may be
equivalent to the effective tooth error, the mass of the gears themselves
taken as 1 if an electric motor drives a centrifugal blower (smooth load),
the predominant factor, and the dynamic load is much less than given |
whereas if the drive is by gasoline engine, 1.25 may be more appropriate.
equation (13.7). There is no easy way to define a dividing line, but if
The following summary from various sources may be helpful when the
shaft is as small as din. (let up to 2in. be piesa) the gear on it |
dynamic load is known with some confidence [see Ref. (13.1)]; the machines
are the driven machines, driven by electric motors:
For these small powers, equation (k) or (1), § 13.15, is recommended for 1 <N., < 1.25: uniform load without shock; centrifugal machines, hoisting
dynamic load. machinery, belt-driven machine tools, textile machinery, smoothly running
Since the origin of equation (13.7) is different from that of equati conveyors.
(k), (1), and (m), there is no reason to expect that dynamic loads comput 1.25 < N.; < 1.5: medium shock; frequent starts; reciprocating compressors
and pumps, pneumatic tools, well-drilling machinery, wire-drawing machinery,
from these various equations should check with one another except at
Portable electric tools, lobe blowers, heavy-duty conveyors, machine tools, kilns.
certain operation conditions, Fig. 13.9.
1.5 < Ns; <1.75: moderately heavy shock; dredging machinery, road
An alignment chart in Problems will be helpful in solving equation (13.7 machinery, railway motor cars, single-cylinder compressor, ore or stone crushers,
Punch press, tumbling barrels.
1.75 < Ns <2: heavy shock; rolling mill, rock crushers.
13.18 SERVICE FACTORS. Gears in actual service, like other
machine elements, are subject to such a variety of operating conditions, For increased reliability, use higher service factors; if limited life is
each of which is impossible of detailed analysis, that the only practica Satisfactory, use lower values.

373
374 SPUR GEARS [Ch. I3_
13.19 PERMISSIBLE AND EXPECTED ERRORS. The higher the
pitch-line speed of gears, the greater the dynamic reactions arising from FIGURE 13.11 Hob Cutter. The profile abcd is the
shape of a rack tooth, except as adjusted for clearance.
the composite tooth error. Moreover, the relative magnitude of the noise — This cutter generates the tooth profile on the gear being
of a running pair of metallic gears is an indicator of the magnitude of the cut. (Courtesy Brown & Sharpe Mfg. Co., Providence,
errors—in general, the louder the noise for a particular speed, pitch, and R.L).
environment, the larger the error. So, higher speeds require the greater
accuracy for satisfactory operation. The permissible error from Fig. AF 19”
is taken as the maximum error for satisfactory operation. We note in
Fig. AF 19 that for speeds greater than 5000 fpm, the tooth error should
be of the order of 0.0005 in.; Fig. AF 20 suggests that small teeth (P; > 5)
extreme care with accurately ground tools on shapers kept in the best
are therefore advantageous. q
condition. On hardnesses up to 450 Brinell, the teeth may be finish shaved,
The way to use Figs. AF 19 and AF 20 in design is first to determine
a precision finishing process. Shaving can produce the involute profile
the permissible error from Fig. AF 19 after having computed the speed.
within a tolerance of 0.0002in. and a spacing error within 0.0003 in.
This is an indicator of the needed accuracy of manufacture. Then with a
Shaving is preferably done before heat treating on material with a machin-
known or assumed pitch, enter Fig. AF 20 and decide on the quality f
ability rating of 25% or better. Under favorable conditions, the error
manufacture and use the value of e from the next lower curve in computing
may be reduced to 0.0003 to 0.0004 in. by lapping or grinding. Shaving is
C in Buckingham’s equation. In general, the teeth should not be more
the fastest process and lapping the slowest; see Fig. 13.10(b).
accurate than necessary because greater accuracy frequently means greater
When gears are heat treated after the teeth are cut, the profile error is
cost, which suggests using the permissible error in design—if the permissible
sure to increase. Under the most carefully controlled conditions, the in-
error is smaller than that corresponding to commercially cut teeth. How-
crease in error due to heat treatment has been held to as little as 0.0005 in.
ever, the most economic value can only be determined by knowing how
For less exacting treatments, the warpage may be considerably more—in
much effort (and cost) is necessary in a particular shop to obtain a particu
some cases, so much more that unground heat-treated teeth have shorter
error; and this is quite variable.
life than teeth not heat-treated.
In Fig. AF 20, the curve labeled First-class Commercial Gears represents
For exacting, high-capacity service, hardened teeth should be finish
the results that might be expected of gears cut by form cutters, Fig. 13.10(a
ground: (1) by a form grinder whose profile fits the tooth profile after the
or hobs, Fig. 13.11, if reasonable care is taken in performing the work a
manner of a form cutter, Fig. 13.10; (2) by a straight-sided grinding wheel
if the machines and tools are in good order; hardness ranges up to about
that generates the profile as does a rack cutter; (3) by a grinding worm
350 Brinell. ;
that has a thread section much like that of the hob of Fig. 13.11. Observe
The curve labeled Carefully Cut Gears represents the results that should
be obtained with accurately ground hobs or shaping cutters, Fig. 13.12,
when the finishing cut is taken on one side at a time. In quantity production,
carefully cut gears may be made virtually as cheaply as commercially cut
gears. ‘
The curve labeled Precision Gears represents the degree of accuracy
that may be expected in carefully ground or shaved teeth, or teeth cut with
FIGURE 13.12 Gear Shaper. The
Pinion cutter has teeth the same as on
a like size gear, except as adjusted
FIGURE 13.10 (a) Form Cutter. (Courtesy
for clearance. The gear being cut and
Brown & Sharpe Mfg. Co., Providence, R.I.).
the cutter rotate together as the cutter
(b) Shaving Cutter. (Courtesy National Broach”
reciprocates vertically, generating the
and Machine Co., Detroit). In (a), the profile
tooth profile on the gear being cut.
abcd is the shape of the tooth space on the gear —
(Courtesy Fellows Gear Shaper Co.,
being cut. (b) The shaving cutter has many
Springfield, Vt.).
cutting edges; produces accurate profiles.
 Master Gear
g13.21] LIMITING LOAD FOR WEAR 377
- a7 Floating
Carriage
FIGURE 13.13 Gear Tester. Helical gear the service factor is virtually unity, F,/Fs = 2400/2420. If this is not satisfactory,
mounted for test. (Courtesy George Scherr Co., and usually some margin of safety is desired, an even larger tooth may have to be
Inc., NYC). used, or a more accurate (and expensive) method of manufacture will reduce the
error and the dynamic load. On the other hand, the Buckingham equation gives
results on the conservative side. See Fig. 13.9. We do have the evidence from the
previous traditional design to substantiate this observation. Therefore, the solution
by this method of design is
P.z = 4, N>, = 20, N, = 50, b = 3 in., carefully cut.

that the surface finish of teeth is either machined (hobbed or shaped),

shaved, or ground (or lapped).
There are several machines available for testing the accuracy of the 13.21 LIMITING LOAD FOR WEAR. The kind of wear referred to
tooth profiles, one of which is shown in Fig. 13.13. If two perfect gears were here is that which occurs because of a fatigue failure of the surface material
in mesh, the dial pointer would not change position; the center distance as a result of high contact stresses, a phenomenon called pitting. Except
when they are in contact at the pitch point where pure rolling occurs for
would be the same for every point of engagement; and the floating carriage 7
would not move. This type of machine will indicate when the teeth are not an instant, tooth profiles roll and slide on one another while in contact.
The frictional force produces a tangential tensile stress, thus acting to
concentric with the bore, when the tooth profiles depart from an involute,
increase and complicate the stresses; its adverse effects are of course much
when there are variations in thickness of the teeth, and when there are
variations in the spacing. See Ref. (13.1) for considerable detail of manufac~ reduced when a lubricant is present to reduce the force tangential to the
turing processes. surface. Running-in under light load and thus smoothing the tooth profiles
increases the capacity because a thinner oil-film will keep the metal surfaces
out of contact (as in sleeve bearings). If a lubricant film exists, the area
13.20 EXAMPLE—BUCKINGHAM’S EQUATION FOR DYNAMIC LOAD.
carrying the load is increased and the film absorbs some of the shock of
The data are the same as in § 13.16; design the teeth using Buckingham’s equatio} n tooth engagement, which reduces the magnitude of the peak stress. What-
for average gears. 5 ever the offsetting factors may be, Buckingham found a good correlation
Solution. From § 13.16, we have vm = 1505fpm, F, = 438lb. For between surface fatigue failure and the Hertz contact stress when the con-
= 1505 fpm, we find the maximum permissible error from Fig. AF 19 to be tact is taken at the pitch point. According to the Hertz equation, the
e = 0.0019 in. Preliminary calculations suggest that a pitch of 5 will not giv calculated stress is greater where contact starts or ends (r, or rg may become
large enough tooth according to this method; therefore, assume P, = 4. Enter quite small), but the pitting phenomenon almost invariably occurs first
Fig. AF 20 at P, = 4 and e = 0.0019, we decide that the teeth must be carefully in the vicinity of the pitch surface, when one pair of teeth is carrying the
cut, which corresponds to an expected error of 0.00125; use e = 0.00125 in. load, thus justifying Buckingham’s assumption. If contact occurs at a base
With carefully cut teeth and the Buckingham dynamic load, we may assume circle, one r is zero:and the Hertz equation does not apply. To derive
the load is shared by two teeth until the point of application has moved to about
Buckingham’s wear equation, start with the Hertz equation (1.19) of § 1.27;
mid-profile. From Table AT 25, C = (1.25)(830) = 1037. For a face b = 3 in.
let the materials of both gears have the same value of Poisson’s ratio,
(6P4 =3 x 4 = 12), we get from equation (13.7) j
Hy = pe = 0.3; the radii of profile curvature at the pitch surface are
0.05un(6C + F;) ') = rp sing and rg = r, sin d, Fig. 13.14; the face width b is the length of
Fa =F, +
0.050n + (6C + F,)?/? Contact line. With these modifications, equation (1.19) becomes
(0.05)(1505)[(3)(1037) + 438]
= 438 + = 2420 Ib. ro 0.35F(1/rp'+ rg)
(0.05)(1505) + [(3)(1037) + 438]?/2 :
The fatigue strength reduction factor for the load applied near the middle of the’ ~ sin db(1/E, + 1/E,)’
tooth is higher than with the load applied at the tip; use K;= 1.7. From Table
AT 24, Y = 0.544 for P.D, = (4)(5) = 20 teeth and load near the middle of the iy bs? sin #(1/E, + 1/E,)
profile. The strength of the assumed tooth is then (p) —
(0.35)(2)(1/Dp+ 1/D,)
_ sbY _ (10,000)(3)(0.544)
K;Pa (1.7)(4) ie where the subscripts p and g refer to pinion and gear, respectively. Note
that the larger the pressure angle ¢, the larger the load F for a particular
We observe immediately that for these computations there is no margin of safety;

376
378 SPUR GEARS [Ch. 13 379
313.21] LIMITING LOAD FOR WEAR

stress s, which means that the 20° full-depth tooth has a greater capacity
F,, than the 144° tooth. The force F in equation (p) is strictly the normal steels heat treated to the same hardness may have significantly different
values of K,, but not enough is known to generalize. Also the results re-
force W in Fig. 13.6; F = W = F,,,/cos ¢ when the teeth are in contact at
ported by Cram suggest that some of the Buckingham values are too high;
see those with a double asterisk **, Table AT 26. See Ref. (13.1) for differ-
ent values of S,surr- For steel, the surface endurance strength may be
estimated from

FIGURE 13.14 Curvature of Tooth (5) Snsurr = (400)(BHN) — 10,000 psi or —_s = (0.4)(BHN) — 10 ksi
Profiles. r1 = Jip = radius of curvature [STEEL]
of pinion profile at pitch surface; r2 =
Iop = radius of curvature of gear profile when test values are not available. When F,, is compared with the Bucking-
at pitch surface. 4 ham dynamic load F,, no margin of safety or service factor is needed except
for extreme situations, and perhaps for the surface-hardened teeth where
there is some conflict in the literature as to whether there is a surface
endurance /imit.
As previously discussed, with the contact compressive stress, there is a
high shear stress just below the surface, § 1.27, which is about 0.3s, where
s is given by the Hertz equation (p). It is believed that surface fatigue failure
the pitch point, where F,,, is a tangential load. However, Buckingham
is a consequence of this shear stress with cracks starting in the vicinity of
compares the F in (p) with the tangential dynamic load. Thus, we let
maximum shear stress,"?-22) and it is evidenced by small particles of metal
1 D,D, Poo. being sheared away, leaving a-pitted surface. For this reason, surface
Dette Boa ee ee hardening, as carburizing, flame or induction hardening, should extend
further than double the depth to the point of maximum shear, given in
( : (“S*)( 1 i 1 § 1.27.[3:3] See also Grover."?-22) It is preferable for there not to be a
steep hardness gradient because the boundary area may then be the weak
» Fo high end Ba Wien ated part.
Use these in (p), which then reduces to The pinion teeth should be somewhat harder than the gear teeth, since
its teeth undergo more contacts than the gear teeth and the harder metal
(13.8) F,, = D,bOK,, cold works the softer, improving the surface smoothness and mechanical
where F = F,, is the limiting wear load and properties. See Fig. AF 4 for conversion of hardness numbers. A reasonable
commercial tolerance on BHN would be some 40 or more points. Some
(r) Q 20)
2D g uh 2N, g aay 2m g : designers prefer to use the computed contact stress from (p) as a criterion
Di, 3D, * Ny Np 20mg Po of fatigue, compared with limit values from tests or service.
In general, when indefinite, more-or-less continuous service is desired,
the gear ratio m, = N,/N, = D,/D,; the wear factor for gearing K, is 4 at Is necessary for both F,, 5 Fy and F, 5 F,. For intermittent or other
material factor* dependent also on the pressure angle ¢; for indefinite limited service, it is common to make F, 5 F, only; and for very short
service, s is taken as the limiting value for surface fatigue (5p gurr in Table Service, fatigue limit F, may be smaller than F,. If sufficient information
AT 26). Most of the values of K, in Table AT 26 are from “actuarial” were available, both the beam strength and the surface fatigue life could
tables by Buckingham (in a personal communication), some are the original estimated on an actuarial (probability) basis (as discussed for rolling
Buckingham values. Where you find the asterisk *, the values are dedu d AToeu See values of K, for different numbers of cycles of life in Table
from Cram"™?-22! by reducing by 15% his laboratory results for 0% sliding \ 1 26. If obtained at some cost, indefinite life is too long when a limited
(except class 20C.I.). There is some evidence???) that different alloy life will do.
Steel teeth are more likely to fail by “‘wear’’ (surface fatigue) than by
* Some give the wear factor Ky other definitions, even to the extent of varying it with the breakage; design steel teeth for wear, check strength. Design cast-iron
7
service, thus making it in part a service factor. teeth for fatigue breaking, and check for surface fatigue—with exceptions.
380 SPUR GEARS [Ch. 13 g13.24] MATERIALS USED FOR GEARS 381

13.22 EXAMPLE—WEAR OF CAST IRON TEETH. Check the limiting wear 13.24 MATERIALS USED FOR GEARS. All kinds of materials
load for the gear teeth of the example of § 13.20. The data are: Ps = 4, b = 3 in, are used for gears. The cheapest is ordinary gray iron, say ASTM 20
N, = 20, m, = 2.5, D, =Sin., hp = 20 at n, = 1150rpm, ¢ = 20°, and which is relatively good from the standpoint of wear and which should
Fa = 2420 Ib. be used if it fits the scheme of things. ASTM classes 30 and 40 are frequently
Solution. Using the conservative value from Table AT 26 for ordina "
cast iron, say class 20, and 20° involute, we get K, = 112 from Table AT 26, used ; and the high-strength irons, say class 60 or higher, are also appropriate
From equations (r) and (13.8), we get
for gears with the proper heat treatment.
Untreated wrought-steel gears are inexpensive and generally have a
2m, (2)(2.5)
carbon content of 0.30-0.50%; low wear capacity. Small gears are fre-
Ort ele Loic quently cut from solid pieces of cold-finished stock. Cast steel should be
F. = D,bQK, = (5)(3)(112)(1.43) = 2400 Ib.
annealed to avoid excessive distortion. Both the wrought and cast steels
Since F, = 2400 is virtually the same as Fa = 2420, these gears should last ir may be heat treated for improved mechanical properties. Oil quenching is
definitely if properly made, assembled, and cared for. not so drastic as water quenching and is generally used because of the
consequent lesser distortion.
Also because of less distortion, alloy steels are favored when teeth are
13.23 GEAR-TOOTH WEAR. In addition to the “wear,” to be heat treated. Of course, the blank can be heat treated before the teeth
are cut. Ordinarily, the maximum hardness for machining is about 250 BHN
are described in Ref. (13.18). We shall mention only four more, givir but successful machining is often done on alloy steels of 350 BHN, even tb
the words the meanings in accordance with ASA and AGMA; see als 450 BHN on AISI 4340 (not with ease).
Ref. (14.1). The heat treatments used are described briefly in Chapter 2. When gears
are to be through-hardened, the carbon content should be 0.35 % to about
0.6%; frequently used steels, roughly in order of cost, are: 1335, 5140
4037, 4140, 8640, 8740, 3135. The hardness depends on the carbon Eiene
abrasion due to film failure. ' the hardenability, and the heat treatment. See Tables AT 8, AT 9, and Fiss
Scoring occurs when the oil film fails, but in this case, the load and spee AF 1, AF 2, and AF 3. There is sometimes a problem of excessive brittle-
are so high that the surface metal is melted and the metal is smeared doy ness; to avoid this, we may for pedagogical purposes arbitrarily set a limit
the profile."3-281 Sometimes there is local spot welding of the teeth togeth on the lowest permissible tempering temperature, as 700°F. (This tempera-
and subsequent separation, leaving a pockmark. Scoring occurs at a lo v ture may be too low in some cases; remember the loss of fatigue strength
loading with thin oil than with heavy oil. So-called EP (extreme pressur of highly hardened pieces with stress raisers.)
lubricants, which contain an additive that forms a film on the surface: b When through-hardening does not produce hard enough surfaces for
a chemical reaction between the metal and the additive (sulfur and chl Surface fatigue strength, the surface is especially hardened. Cyaniding,
a
compounds, for example), can withstand much heavier loading wit shallow hardening process, is done on the medium carbon steels. Much
permitting scoring. This film is continuously being formed and worn, the same steels are used for flame-hardened and induction-hardened teeth
under steady operating conditions, arrives at steady-state thickn as are used for through hardening. The higher the carbon content (to a
It is thinner near the pitch surface where rolling occurs and thicker neart Point), the harder the surface may be. In the absence of more specific
tips where sliding is a maximum, leading to the deduction that the eg ee assume that the surface hardness of teeth by flame and
generated by friction (minimum at pitch surface, maximum at tip), i Sei hardening to a 500 BHN, a value that should be readily attain-
factor that brings about the reaction to form the film."%-?8! Scoring mi a the carbon is 0.4% or more. If the teeth are carburized, a minimum
be a problem in the high-surface-hardened, heavily loaded teeth, but ¢ t = 600 is a reasonable expectation; use this for design. Automotive
irons and bronzes with normal lubrication will probably have a toot oe ee are not subjected to continuous full power operation, were
breakage before a film failure. f Seals Biagio hardened. The addition of cyaniding reduced troubles
Spalling is a surface fatigue of greater extent than pitting; that is, ‘ aa ace fatigue; 5145 through-hardened took 10,000 hr. of operation
flakes are much larger. This type of failure occurs in surface-hardened teeth RG Pie the same material cyanided ran 25,000 hr. before pitting. '?-2)
Corrosion causes surface deterioration that in turn results in weakenim anit e automotive industry by carefully controlled processes to keep
of fatigue strengths, but presents no problem with clean petroleum oil ig ata minimum"-33] yses case-hardened alloys without a finish
and a clean environment. Operation after hardening (small gears); in the same experiment just quoted,
382 SPUR GEARS [Ch. |3 — g13.25] DESIGNING STEEL GEARS FOR CONTINUOUS SERVICE 383

8620 case hardened to a depth of 0.035-0.05 in. ran 75,000 hr. without— 13.25 EXAMPLE—DESIGNING STEEL GEARS FOR CONTINUOUS
pitting, and if the teeth were shaved first and then hardened, the running © SERVICE. Design the teeth for a pair of steel gears to transmit 40 hp at 1150 rpm
time was 150,000 hr. The life of case-hardened teeth improves with depth of the 5-in. pinion; m, = m, = 2.5; 20° full-depth, interchangeable teeth.
of case perhaps up to some 0.08 in. thickness; a rule of thumb for case _ Solution. The pitch-line speed is 7(5/12)(1150) = 1505 fpm (as in § 13.20);
thickness is about 0.1 times full depth of teeth. the transmitted force is*
Alloy steels for carburizing, roughly in order of cost, are!?-1! _ 33,000hp — (33,000)(40) _
F, = Aa
eupci lis hoes
Sa ic 876 lb.

5120, 4118, 8620, 8720, 4720, 4620, 4320, 4820. Checking the speed of 1505 with respect to Fig. AF 19, we decide as before that
the teeth are preferably carefully cut. As previously pointed out, steel gears
without special surface hardening are more likely to wear out than break. There-
The flame-hardening process for gear teeth needs development and fore, design for wear first. As usual, there are so many unknowns that iteration is
testing for particular cases. The heating and quenching is of one tooth at necessary. The following procedure is reasonable.
a time; this tends to anneal the fillets of the preceding tooth just hardened, Assume Pz = 4; then e = 0.00125 from Fig. AF 20, and from Table AT 25
thus reducing the beam fatigue strength. Also, the roots of induction- for steel and steel, we get C = (1.25)(1660) = 2075. Let b = 2.75 in. (6P.2 = 11,
hardened teeth may not be fully hardened and therefore the teeth may be which is satisfactory). We shall compare the limiting wear load with the
unexpectedly weak in bending fatigue. If unusual precision of surface Buckingham dynamic load; as computed from (13.7),
hardened teeth is necessary, care must be taken not to lap or grind away Fa = F; +
0.05un(bC + Fi)
too much of the case and lose the residual compressive stress on or near 0.050m + (6C + F,)*/?
the surface, and not to leave a discontinuity. Grinding tends to leave residual (0.05)(1505)(2.75 x 2075 + 867)
‘= 876 + = 4036 Ib.
tensile stresses, a (0.05)(1505) + (2.75 x 2075 + 867)*/?
In extreme service requirements, nitriding may be used at increas We make the limiting wear load F, = Fa = 4036 1b. In equation (13.8), D, =
cost. Because the process is carried out at a relatively low temperat S5in., b = 2.75in., Q = 2m,/(1 + m,) = 1.43; solve for K,, which will indicate
there is not so much trouble with warping. Other surface hardenin, something about the material needed. From F,, = D,bQK,,
processes are used (§ 2.8), even peening; but peening the fillets to improve ~ 4036 oy
the beam strength or peening to remove scale after heat treatment are the K, 206.
~ (5)(2.75)(1.43)
more likely applications of this process. 1 Look for this value of K, in Table AT 26 in the 20° column (indefinite service).
To summarize the relationships between machining and heat treating Interpolating between the total SBHN = 600, K, = 196 and XBHN = 650,
we may use gears after the following operations; (1) the blank can K, = 233, we find the needed total SBHN = 614 for K, = 206. Let the gear be
annealed or normalized, the teeth machined complete (low hardness); BHN, = 250 because carbon steel can be machined at this hardness and there
(2) the blank can be annealed and/or normalized, then heat treated to should be a significant saving on material cost (as compared with an alloy);
final hardness, the teeth machined complete (hardness limited by machi then the pinion hardness is BHN, = 364. These values are starting points for
able hardness; excellent hardenability probably required to obtain 1 consideration of the kind of steel to use. The final decisions are made by studying
desired hardness all the way through), or for small gears, one might b charts and tables giving mechanical properties, with appropriate metallurgical
suitably heat-treated bars; (3) as in (1) except that teeth are rough and heat-treating knowledge in mind. Consider through hardening first. If we
Stay within the pedagogical limit of 700°F tempering temperature and use only
then heat treated, then finish cut [can get somewhat higher hardness than
the information in the text, Fig. AF 1 shows that oil-quenched C 1040 can barely
in (2)]; (4) as in (2) except that the surface is later hardened by one of the
qualify; BHN = 248 at 700°F. Since the hardenability of plain carbon steels is
processes previously described, after which a finishing operation of grinding low, the teeth would need to be cut before the heat treatment. Tentatively then,
or lapping may be done (necessary in large gears) to improve the profil the gear is to be made of C 1040, OQT to a minimum hardness of 250 Brinell—
accuracy. There are of course numerous variations of heat-treatment! Provided the teeth prove to be strong enough in bending. (An alloy with a good
but the principal ideas are as outlined. Sintered gears find frequent appli hardenability, which could be hardened first and the teeth cut afterwards, could
tions, especially small ones as in appliances. For example, a crank geaf Prove more economical on an over-all basis.)
(automotive) with teeth is produced from a mixture of 6% Ni, 2% Cu, The pinion with Brinell of 364 should be an alloy, any one of several being
92% iron powders, double pressed and sintered. Designing the gears with Satisfactory, Table AT 9, Figs. AF 2 and 3. For this hardness, the teeth are more
a pressure angle larger than 20° improves their capacity.°-26) d * The data are the same as in § 13.20 except that the horsepower is 40.
384 SPUR GEARS [Ch. 13 513.26] MORE ON THE DESIGN OF GEAR TEETH 385

likely to be cut first and hardened later. Choose 4140, Table AT 9, and interpolate something of an expert, allows currently available knowledge to affect
for the approximate tempering temperature from 800°F, 429 BHN and 1200°F, his decisions, (3) the designer who is concerned with situations that require
277 BHN;; this indicates a tempering temperature of about 975°F (for 364 BHN), special judgment and experience, or the extrapolation of experience, as in
which should not result in trouble from brittleness. Now check the strength. automotive, aircraft, and space-systems gears. (Of course, these basic ideas
For BHN = 250, we estimate s,. = 250/4 = 62.5 ksi. For BHN = 364, we
of classification are extendable to other machine elements.)
get s,, = 364/4 = 91 ksi. For N, = 20 and the load near the middle of the profile,
So many variables are involved, including the manufacturing and
Y, = 0.544. For N, = 50, Y, = 0.694; Y,5an = (0.544)(91) = 49.5; Yosng =
(0.694)(62.5) = 43.4. Therefore, the gear is weaker. Let K, = 2, which should operating variables, that formulas for special service conditions have
be on the high side. ; proliferated. However, it is to be hoped that eventually reasonably
accurate computations of the dynamic load can be easily made, and related
(sY).b (43.4)(2.75) 3
F, KP. Qy4) 14.9 kips, | to the beam strength via suitable design stresses modified by service factors,
and to the limiting wear load (or Hertz stress). In the meantime, the full
or a strength of 14,900 lb. compared with a dynamic load of 4036; therefore,
Buckingham procedure"™$-5) (Fig. 13.9), which considers the inertia of the
breakage is quite unlikely.
connected masses and the rigidity of the connecting members and which
Discussion. We observe that there is an infinite number of solutions to this
problem by any approach, considering all trivial differences in the kind of steel
we do not have room to explain in sufficient detail to make useful, is probably
and heat treatment. The optimum solution, which we may define as the one tha the wisest for the designer working in an area where experience is limited.
satisfies the objectives at the minimum cost of manufacture, can only be decided Where applicable, the AGMA rating formulas should produce conservative
upon with a complete knowledge of the whole situation. There is no substitute results; this approach is given in Chapter 15 for bevel gears.
for an actual manufacturing plant in a particular environment (kinds of machine As seen from Fig. 13.9, the estimation of the dynamic load is not nearly
tools, heat-treating facilities, local suppliers, etc.). If the gears are to be bou so easy as the pat use of a formula suggests. In heavily loaded gears, the
from a gear manufacturer, bids could be obtained on two or more alternat deflection of the tooth is greater than the manufacturing profile error (the
specifications (perhaps leaving the choice of the particular material up to the effective error, § 13.17, is the manufacturing error plus the deflection).
supplier who may be prepared to process one material more cheaply than anothe: In lightly loaded gears, the tooth deflection may be very much less than the
One alternative to consider may be to use a pinion of about 330 Brinell and a profile error.4%:1] Lightly loaded gears operated at high speed may fail
of about 290 Brinell, with the idea of using a material of such good hardenability
because the dynamic load grows with speed until it is destructive. But see
and machineability that the teeth may be cut after heat treatment. The possibility
§ 13.38. On the other hand, Reswick"*-15! says that the dynamic load can be
that this procedure (at least, in the general case) might result in more accurate
virtually ignored in heavily (transmitted) loaded gears. A heavily loaded
profiles, therefore lower dynamic load, therefore more economical gears, should
not be overlooked. Perhaps the pinion is not necessarily 5 in. in diameter; :
tooth is one for which the deflection by the transmitted load is greater than
cheaper solution may be found in increasing or by decreasing its size. If low weight the manufacturing error. Also, Kohler observed (pp. 14-32, Ref. 13.1) that
is one of the objectives, surface hardening would significantly reduce the size of for his data, the dynamic load did not exceed three times the transmitted
tooth and of pinion needed. a load, but the Buckingham analysis predicts much higher ratios when the
The basic idea of the design procedure outlined by the foregoing solution is transmitted load is small compared to the effective inertia of the system.
to assume a pitch and face width, and then determine the corresponding required Buckingham points out that heavy dead loads (flywheels, rotors) on shafts
hardness; if this hardness is not obviously the best, assume another tooth size result in shaft vibrations doing more damage to the teeth than low inertia
and repeat the calculations; etc. Always check the beam strength; sometimes ii loading that permits the vibrations to be largely absorbed in the trans-
significant even for indefinitely continuous service. 4 Mission rather than reflected to the teeth.
The design should be studied further with respect to AGMA recommendations:
In extreme situations, as in heavily loaded, carburized, aircraft gears,
As a result, it may be found that the Buckingham equation is overly conservative.
Then there is the traditional design with velocity factors. Equation (I) gives 4
the profiles are modified so that they have more nearly perfect involutes in
a deflected position and also for the purpose of reducing the load on that
dynamic load of F, = 1980 1b. Introducing a factor for safety of 1.5, we get
F4 = 2970 lb., considerably less than that by the Buckingham equation.
Part of the profile subjected to maximum sliding velocity v,. Thus it should
lower value of Fz could mean important savings in heat treatment costs. Not be assumed that the applied-dynamic-load-wear-and-strength-resistance
Concept is necessarily the basic criterion for all situations. The parameter
Scl’s has been used as a criterion for high-speed gear capacity (against
Scoring), where s, psi is the Hertz contact stress, and v, fps
13.26 MORE ON THE DESIGN OF GEAR TEETH. Those who is the sliding
design gear teeth may be classified into three groups: (1) the casual Velocity,"3.9] applied particularly to helical gearing. The limit value of
designer who, more or less blindly, follows as printed an accepted procedure, Ses may be from 3 x 108 to 4.5 x 10°; but see § 13.38, also Kelley.{12-22
(2) the designer of gear reducers who must meet AGMA ratings, but being” Or more detail on the sliding velocity, see Ref. (13.2.)
386 SPUR GEARS [Ch. I: g13.28] EXAMPLE—LAMINATED PHENOLIC GEAR TEETH 387

13.27 DESIGN OF NONMETALLIC GEAR TEETH. Various nor under load may result in the tip of a metal gear tooth digging into the flank
metallic materials, Table AT 5, have long been used for gears, perhap: of the plastic pinion at the initial point of contact, thereby damaging it.
principally because of their quiet operation. A disadvantage that may b For this reason, the use of unequal addendum and dedendum teeth, § 13.34,
significant is the low heat conductivity of these materials. If the powe even to the extent of eliminating the angle of approach, is recommended.
involved is large enough, the increase in temperature may seriously weake
the nonmetallic gear teeth; the friction and work lost in deforming th
profile may not be dissipated to the surroundings at the rate at which
occurs until the temperature is excessive. The teeth for rawhide an 13.28 EXAMPLE—LAMINATED PHENOLIC GEAR TEETH. What should
laminated phenolics are designed as previously explained, with F, = F be the pitch, face, and number of teeth for a pair of 20° full-depth spur gears if
and in case of continuous service, F,, = Fy. Because of the low modulu they are to transmit 12 hp at 1150 rpm of the pinion; m,, = m, = 2; the load is
of elasticity of these materials, the dynamic load is smaller than for met smooth and the service is continuous. Let a Bakelite pinion drive a cast-iron gear,
ASTM 20 (but there may be trouble—see the previous paragraph).
teeth and is taken as
Solution. First decide on a suitable number of teeth; let VN, = 20 for which
there is no interference trouble; then N, = 40. Comparing sY for gear and pinion,
(200 oe Um) F
13.9 F, = i we note that the pinion tooth is the weaker. Since neither the diameter nor pitch
hist * 200
+ 09/4 are known, some iteration is indicated. (You can express velocity, transmitted
load, dynamic load, and strength load in terms of Pz; then solve for Pa; try it.)
which is compared with the strength F, = sb Y/P, when the load is appli Assume Pa = 5; this gives D, = N,/Pa = 20/5 = 4in.,
at the tip of the tooth; and design stress s = 6000 psi (no stress concen
tion factor); no change in the equation for F,,. i vn = aDn. = (5)(1150) = 1205 fpm,
Since the modulus of elasticity E is much less for these materials thé
for metals, the teeth have a greater deformation under a given load, Fie_ spate
33,000 hp
rune_ eae
(33,000)(12)
ee _Omi Ib.,
therefore the load is more likely to be distributed over two teeth, e7 . (200 + vn)F, (200 + 1205)(328)
when contact is in the vicinity of the pitch point. This fact is recog F
O
=
FOO + ve/4
=
200 + 1205/4
=
ae
in the recommended design procedure for nylon (Zytel) gears. When
From Table AT 24, find Y = 0.32 (at tip) for Np = 20, and then solve for 6 from
of Zytel, the teeth may be either molded (inexpensive in quantity) or ¢1 Lewis’ equation (13.5):
they are most commonly of fine pitch (Pg > 16); and strength is taken”
the determinant. Use the value of Y for the load near the middle of prof Fd Bieseios sbY ah (6000)(4)(0.32)
in F, = sb Y/Pq and design stresses as follows:"1-55-13-19] for 5 x 108 cyel Pa 5
s = 2.3 ksi for Py = 16, s = 2.6 ksi for Py = 20, s = 3.5 ksi for Pg from which 6 = 2.39 in., length of face needed for strength. For the wear equa-
s = 3.7 ksi for Py = 48; for 107 cycles, and Py of 16, 20, 32, 48, re tion, we have (K, = 64 from Table AT 26—only value at hand)
tively, use s as 4.2, 4.7, 5.8, 6.1 ksi; for other values of s, interpolate di 2m, __(2)(2)
as the log N, N = number of cycles. The dynamic load is recommend
be a velocity factor (VF) times the transmitted load; Fy = (VF)F;. Fy = Fa = 920 = D,bQK, = (4)(b)(1.33)(64),
gear teeth are lubricated, 16 < Py < 48, use VF as follows: from which b =2.7 in., needed for wear; bP = (2.7)(5) = 13.5, on the high side.
If not satisfactory (see E of phenolics), recheck for Pg = 4.
Molded Teeth Cut Teeth Discussion. To emphasize that design must follow a validated combination
Un < 4000 fpm, VF = 1.0 1.2 of loading, design stress, and the method of computing the stress, calculate the
Un > 4000 fpm, VF = 1.2 1.4 Stress in the tooth found above if the gear is molded Zytel; that is, follow the
design instructions for the nylon gears. Since vn < 4000, the velocity factor is 1;
Let F, = Fy for smooth loads; include a service factor for other kin take Fy = F, = F;; use Y = 0.544 for the load near the middle;
then
of loads. Nylon gears can be run without a lubricant (preferably ini SREP SOO 28S eu 20) :
wetted with one), but the design stress should be much reduced.4** ‘Lipo T Gsis@sasy 7 17/0 Pat
Use 20 or more full-depth, 20°-involute teeth, but as usual, the smalle This tooth is much larger than the fine pitches for which we have quoted justifica-
teeth that will do the job. tion in experience with nylon, but the stress is on the low side (see stresses given
Buckingham"™3-15) says that mating phenolic gears with steel ¢ nN § 13.27), leading to the conclusion that if a nonmetallic gear is needed here,
BHN < 400 leads to excessive abrasive wear. Also, the large deflecti@ nylon should not be ruled out.
388 SPUR GEARS [Ch. 13
13.29 DESIGN OF CAST TEETH. Cast teeth are comparatively
rough and inaccurate, and they are therefore more likely to be used ip
large, slow-speed (below 600 fpm) service, preferably outdoors where nois,
will not be objectionable. Use equation (k), § 13.15, for the dynamic lo d
(Um < 600) and use
(t) F, = 0.054sbP, (Courtesy Globe Stock
FIGURE 13.15 Spur Gears.
for computing the strength. Use the endurance strength for s, but include Gear Division, Philadelphia).
a service factor for any but the smoothest loads. Because cast teeth ha
relatively inaccurate outlines, their face width is generally less than fi
cut teeth, say 2P, < b < 3P,, cast teeth; circular pitch is used for th
convenience of the pattern maker.

13.30 HUNTING TOOTH. In the example of § 13.28, we have

N, = 20 teeth and NV, = 40 teeth. After every two revolutions of the pinioy
the same pair of teeth engage. If however, we use 41 teeth in the gear, h
pinion will rotate 41 times and the gear 20 before the same pair of tee
Pitch Cle f i FP
a on
engage again. This extra tooth is called a hunting tooth, and its use dis
butes the wear more evenly. The resulting velocity ratio must be permissible i
A :
|i | b
~
FIGURE 13.16 Moment on Arm, Dimensions at pitch ~ Outline 7}
circle may be made (0.7) x (dimensions at hub); or a taper : | \
of about ? in./ft. is reasonable. —|
13.31 HUBS, METAL GEARS. Hubs of gears are proportioned
empirically. The bore of the hub depends on the shaft diameter. The leng
of the hub depends fundamentally on the length of key, if a key is use
Two keys may sometimes be used to avoid excessive hub length. The leng
of the hub should not in general be made less than the face width of the gea
Hub lengths usually vary from about 1.25D, to 2D,, where D, is the bore
The following values are reasonable for the diameter of the hub: circle to the hub, Fig. 13.16, and F = F, if the design is based on fatigue.
For cast iron, diameter of hub = 2D,; The section modulus Z is in accordance with the shape of the section.
For steel, diameter of hub = 1.8D,. Shapes of arms include ellipses, rectangles (with corners generously
rounded), H section, I sections, and cross sections, Fig. 13.17. For an
For light, low-speed service, these proportions may be decreased somewhat. elliptical section with the major diameter h twice the minor diameter,
The bead at the hub, Fig. 13.15, may be allowed to depend on the size h, = h' = h/2, we get Z = 7h°/64, using Table AT 1. Solve for h from the
of the arms at the hub. i bending moment equation as explained above.
As a convenience for the shapes of arms shown in Fig. 13.17, solve for
hand Z as described above, then compute G and G, as given in this illustra-
13.32. ARMS AND WEBS. The arms, Fig. 13.15, are designed as tion.

significant stresses, it is safe to assume that the dynamic load is the maxi- — h,
FIGURE 13.17 ond
mum load and design the arms using the endurance strength as the desig :
Proportions for a h Bs S i id L
stress, M = sI/c = sZ, and to assume that the load is equally shared by Gear Arms. Pe G
ag o
IND a
Git <4ke -G Fh ©
Oval Arm H-Arm Cross Arm I-Arm
hy Gn G-% hd
G,=75G e-%,
 513.35] INTERNAL GEARS 391

FIGURE 13.18 Welded Gear Blank. An example of lar; jess 2 like amount. To match, the gear’s addendum is decreased and its
gear construction; outside diameter about 9 ft. 7 dedendum is increased Fig. 13.19. Ways of accomplishing these changes
weight, 25,000 Ib; inside hub diameter, 243 in. (Courtes are explained in engineering texts on kinematics or mechanism."%-?! The
Luckenweld Inc., Coatesville, Pa.). 4
amount of the change of addendum may be adjusted so that the pinion and
gear teeth have about the same form factor Y. If interference is otherwise
present, this change increases the contact ratio m,, with the result that the
gears run smoother and quieter. For quantity production for a particular
system, the disadvantage of noninterchangeability virtually disappears.
The principles of design are the same as already explained; the values of
the Y’s will have to be determined from large layouts of the teeth; but the
interested reader will have to review the kinematics of gear teeth for com-
Use six arms whenever possible for gears up to 120 in. in diamete plete understanding.
although four arms do well for small gears. For gears over 120 in. in siz
Broad E Narrow Base of
use eight or more arms. Very large gears may be of welded constructio; No etal Fistk, Full-Height Tooth,
No Interference Radial Flanks
Fig. 13.18. q
Whatever the computed dimensions, care must be taken to avoid abru;
changes in the thickness of adjoining parts. If a very thin section (com
tively) adjoins a very heavy section, unequal cooling of the casting i FIGURE 13.19 Unequal Addendum
case of cast gear blanks may result in severe residual stresses that materia and Dedendum Teeth.
weaken the gear.
Small gears are frequently made with webs of a thickness equal to Gear
0.5P, to 0.6P,, or the teeth may be cut from solid stock, sometimes
the shaft forging itself. For minimum stress concentration in a solid
locate the keyway on the center line of a tooth space."%-+) If the tee
integral with the shaft forging, the diameter of the dedendum circle she 13.35 INTERNAL GEARS. An internal or annular gear, Fig. 13.20,
be somewhat larger than the shaft diameter. has teeth cut on the inside of the rim instead of on the outside. The shape
of the teeth is thus the same as the shape of the space on an external gear
of the same pitch diameter, except as modified by clearance requirements.
13.33 RIM AND BEAD. For the assumption recommended for af
As for external gears, whenever a working part of a tooth extends inside
design, namely that the load is equally shared by the arms, the rim mu:
its base circle, either on the gear or pinion or both, interference may occur.
be rigid enough to accomplish this. Experienced engineers can decide qu
If the pinion is too large, as compared to the internal gear, the teeth will
well by appearance what is good proportion. Empirical values that h
“foul.”? To eliminate fouling, let the minimum differences in tooth numbers
been used are: make the rim thickness and depth of bead each about 0.5
between the internal gear and pinion, N, — N,, be:
(but as small as 0.4P, for small, high-strength, aircraft gears); a be
recommended for its stiffening value when arms are used. ad for 144° involute, full-depth, 12 teeth;
In case of speeds that produce sizeable centrifugal forces, some estimé for 20° involute, full-depth, 10 teeth; and
tion of the corresponding stresses is in order. i for 20° involute, stub, 7 teeth.
If the profiles are not to be modified at all, the difference should be greater
13.34 UNEQUAL ADDENDUM AND UNEQUAL DEDENDU! than given above.
TEETH. Because of interference troubles when pinions with smé
tooth numbers are used and because of certain operating advantages, no: FIGURE 13.20 Internal Gear. A broached internal gear, 16 pitch,
interchangeable gear teeth are frequently appropriate.“*-2*) Basically about 3.7 in. in diameter, made of 4140 steel; the ring gear in a

these systems, the dedendum on the pinion is decreased in order to redus Planetary transmission. (Courtesy National Broach & Machine Co.,
Detroit),
or eliminate interfering flank and the addendum is increased, more a
392 SPUR GEARS [Ch. 13 513.38] LUBRICATION OF GEAR TEETH 393

Internal gear-and-pinion sets are quieter than like external gears, prob- where n, = absolute turns of the arm (planet carrier), positive for CL
ably because more teeth are carrying the load which is therefore not rotation, negative sign for CC rotation,
shifted so abruptly. The combination is compact, with a short center ny = absolute turns of the first gear in the planetary train, same sign
distance. If the gear and pinion are of the same material, the gear tooth convention as for n,,
will be stronger; hence the design for strength is based on the pinion tooth, nz, = absolute turns of the last gear in the planetary train, same sign
With a concave surface on the gear tooth in contact with the convex convention, CL+, CC-,
surface of the pinion tooth, the contact area is larger than for two external e = train value when the first gear in the planetary train is taken asa
teeth (two convex surfaces), and the limiting wear load is larger (the radii driver; evaluated with the arm held stationary (ordinary gear
r, and r, in the Hertz equation are of opposite signs). An extra factor is that train) as the product of the tooth numbers of the driving gears
the distribution of the load among more teeth decreases the intensity divided by the product of the tooth numbers of the driven gears;
the stress. The limiting wear load is computed from F,, = D,bQK,, which given a positive sign if, with the arm held stationary, the first
is the same as before!9-3! except that and last gears turn in the same sense, a negative sign if they
turn in opposite directions.

(u) Cee
2N,
N,— Nz;
ee
m,—1
2m, Correct algebraic signs are essential in the use of (w). If properly handled, it
does not make any difference which end of the planetary train is taken as
the first gear; let it be the input gear. Then the gear for n,; can be any other
gear in the train with e figured to include the one considered to be the last
one.
13.36 GEAR TRAINS. When the velocity ratio is high, it is desirable As you recall from the study of kinematics, three or more planet gears
to reduce the speed in two or more steps. A high limit for one pair of spur are common. This permits designing the teeth for much less than the total
gears is about m,, = 10; a better limit is 6. Helical and herringbone gears, power or torque, although it is probably overly optimistic to assume that
Chapter 14, are most often used in heavy-duty gear boxes. the power is equally divided among the planets; something more like total
When two gears are integral in a double reduction, Buckingham suggests power divided by the number of planets plus 15% would be more prudent.
that the profile error in the second pair be assumed to be about 50% more This same idea is used on occasion with ordinary gear trains, several pinions
than the method of manufacture is expected to produce. This assumption i spaced round the gear delivering power to it.
made because the rigidity of the connection results in a transmission to the
second pair of gears of the effects of the dynamic loads on the first pait.
If, however, the two gears on the second shaft in a double reduction are 13.37 EFFICIENCY OF GEARS AND THERMAL CAPACITY. The
separated by a long length of shafting, the elasticity of the shaft absorbs loss per pair of spur, helical, or bevel gears in an ordinary train should
to a large extent the dynamic loads originating at the first pair of ge not exceed 2%, giving an efficiency of about 98 %, or more. The loss in the
Internal gears are commonly used in planetary trains. From the hors mesh of spur and helical gears should be less than 1°%. Ordinarily, the
power equation, say hp = 7n/63,000, we conclude, ignoring friction amount of the loss in spur and helical gear boxes presents no overheating
losses (hp = constant), that the torque on a particular shaft in a tra problem, but as the speed increases with more or less the same force on the
varies inversely as the angular velocity n (or w). Thus, the torque on any teeth (power increasing), the dissipation of the loss as heat to the surround-
gear can be found from ings may become a problem. It should be investigated when the pitch line
Speed is greater than about 2000 fpm,'!%-1! earlier when there are multiple
E. n
(v) aS = se
Pinions each with an input. Follow the plan outlined in § 16.6, using the
TT, my thermal coefficient from Fig. AF 21. Also, the AGMA gives rating formulas
for thermal capacity.
when the torque on another gear in the train is known, as the input torqu
after which the transmitted load F, is calculated by dividing the torque
the pitch radius. The easiest method of computing the angular velocity 13.38 LUBRICATION OF GEAR TEETH. Fortunately, there is a
of gears in a planetary train is by the equation™®-?! tendency for the hydrodynamic wedge to form between gear teeth, making
(w) ny, = eny + n,(1 — @), [PLANETARY TRAINS] - it possible to have thick-film lubrication. However, at a particular speed and
 513.39] CLOSURE 395
394 SPUR GEARS [Ch. 13
gives generally correct solutions. For the student’s sake, we regret that it
oil viscosity, the tooth load may be so high as to cause the oil film to be
cannot be said to do it in a certain way, but it is inherent that there are
squeezed out enough to allow metal-to-metal contact; if so and if the speed
question marks in every engineering answer; there are just more of them
is below some value (v,, ~ 800 fpm pitch line, in one research), abrasive
with respect to gearing than for most any other machine element. When
wear occurs (§ 13.23); if the speed is higher, scoring occurs. If the load ang
quantity production is involved, laboratory and service tests should be
lubricant are such that only the high spots on the surface are touching
used to prove the design. There is much highly specialized literature that
enough friction is likely to be generated locally to result in a lowering o}
can help. In important cases without adequate reassuring past experience,
viscosity and a consequent early scoring or abrasion; in short, the safe
the American designer may care to follow a British procedure as a
load for very smooth surfaces (20 win.) is much higher than for rou check. 18-251
surfaces (100 pin.), at all speeds. Once scoring or abrasion has occurred
The design of the gear case is no small matter in the success of the gears;
the capacity of the gears is much reduced. if
it must have enough rigidity to maintain the gears in the expected align-
Other findings include: the smaller the teeth, for the same face width
ment.
the higher the load capacity before scoring, which is explained by the fac
that the maximum sliding velocity v, for the same pitch-line speed is highe:
on larger teeth; the capacity of the oil film to stand up decreased as thi
profile error increased, but the effect of the error was not so great as pr
dicted by Buckingham’s F,; with SAE 30 oil, the load capacity decrease
with temperature increase to about 350°F, but increased somewhat as th
temperature went above 400°F because of the formation of carbonaceou
deposits on the working surfaces; a synthetic oil of the same viscosity a
a mineral oil does not have the same load carrying characteristics as tt
mineral oil, but in both cases, load capacity increases with viscosity; relie
the tip of the tooth had a beneficial effect on load capacity up to a relief
0.0008 in.; directing the oil to the initial point of the mesh gave a highe
capacity than for other points of lubrication. it
One relationship that might not be expected—the load capacity d
creased as the speed increased (up to about 4000 fpm pitch line in the tes
cited—mineral oil), and then increased moderately (to some 15,000 fpm)
after which there was a rapid increase in load capacity (maximum
speed, about 22,000 fpm). At speeds over 15,000 fpm (speeds of 30,000 f
maybe more, are nowinuse, introducing some unusual considerations"*
the scoring load for the heavier oils was never reached. The conclu
are, first, that the conventional design procedures are not applicable at v
high speeds, and that small gears can transmit great power; in the te
a pair of gears rated at 10 hp (AGMA) actually transmitted 500 hp with
surface damage. At these high speeds, the thick lubricant film rema
intact, first, because the contact time is shorter than a deformation puls
and the fluid is acting as a plastic solid, and second, because of the squeeZ
effect (it takes a definite time for the lubricant to be squeezed out). Tht
facts of this article and some of the opinions are taken from
Borsoff.'13-28.13.291 The actual speeds cited are not intended as genera
measures; they are known to apply only to the particular tests. See Electro
film, § 11.27. i
f
13.39 CLOSURE. As we have seen, there are so many variable
involved in a gear drive that no simple procedure has been evolved that
 14. HELICAL GEARS

14.1 INTRODUCTION. In accordance with our general plan, we

shall not give as detailed explanations of the engineering of other types of
gears as we have of spur gears, but most of the generalizations of the pre-
vious chapter on kinematics, materials, wear, fatigue strength, dynamic
load, etc. apply as well to any gear system. Differences and changes will of
course be highlighted.
As seen in Fig. 14.1, the elements of helical teeth are cylindrical helices;
one end of the tooth is advanced circumferentially over the other end.

FIGURE 14.1 Helical Gears. See also Fig. 14.7.

(Courtesy The Falk Corp., Milwaukee).

397
 398 HELICAL GEARS [Ch. [4 914.3] PITCHES 399
Without a modifying adjective or other inference by context, “helical gears” pase circles). Since helical gears are inherently not interchangeable—a
refer to ones on parallel shafts (see § 14.9). Since the contact on a particular right-hand helix meshes with a left-hand helix when the shafts are parallel—
tooth starts at one end, always with other teeth in contact, and then suc there are no standard helix angles. Angles of % = 15° to 25° are typical,
ceeding tooth sections make contact, the tooth picks up the load gradually, put for wide-face gears, the helix angle may be less than 15°. Since the end
The line of contact is never across the whole tip, as it is in spur gears, bul thrust increases as the helix angle increases, Fig. 14.2, the helix angle is
is always diagonal. When the midsections of perfect teeth are in contact sometimes limited for this reason. On the other hand, the larger the helix
at the pitch point, the line of contact runs from the addendum circle neaj angle for a particular face width, the more the overlap of the teeth and the
one end of the tooth to the working depth near the other end. The conse: more gradual the transfer of load. For significant advantage from the
quence is that the bending moment on the tooth is only a little over half as overlapping of tooth engagement, the face width should be"?-1)
much as it would be for the entire load at the tip. Helical teeth (paralle
shafts) thus have a greater breaking strength. Moreover, the gradual transfer (a) bE 2F,

of load results in quieter running, lower dynamic load, and higher permis (bmin = 1.15P, has been used in the past), where P, = P,/tan yin. is the
sible speeds. A commercial grade of helical gear runs about as quietly 4 axial pitch [equation (d) below]. The face contact ratio m; is the advance of
precise spur gears, at less cost. Because of these advantages, helical and her the tooth in the face width 5 tan divided by the circular pitch; m, =
ringbone gears are nearly always used in heavy-duty gear boxes, as well 2 btan %/P, = b/P,. Without the adjective, contact ratio has the meaning
in many others. i previously defined, § 13.5, and in particular is called the profile contact
ratio. Basically, the face width is determined by the loading to be carried,
but the helical gears often have large active face widths compared to spur
gears of the same pitch.

14.3 PITCHES. The pitch of a helical gear is the pitch in the dia-
metral (transverse) plane, P, for circular pitch, P, for diametral pitch. As
seen in Fig. 14.2, the distance between the teeth measured on the pitch
surface along a normal to the helix, which is called the normal circular
pitch P.,, is less than P, measured around the pitch circle:
FIGURE 14.2 Helical-gear Notation. A right-han
helix is one that slopes around the pitch cylinder
(b) P.,=P,cos and Py, = Pa
the manner of a right-hand screw thread; that is, cos fs’
tooth on the front side of the gear slopes up to th
¥ right with the axis vertical; apply the right-hand rul where the latter form is obtained by using P, = a/P, and P,, = 7/Pan-
If V, is the number of teeth in a particular gear and D is its pitch diameter,

aD cos P N,
(c) Poy, = P, cos =(es and = Pa, = —"— = ee:
N; cos & D cos
If the helical gears are manufactured with standard hobs, a common
Practice, the normal pitch P,, is standard. It follows that the pitch in the
Gear on Pinion diametral plane P, and also the pitch diameter will contain a decimal
End Thrusts Pinion on Gear.
fraction. For example, suppose N, = 20, Pg, = 6 and iy = 23°; then
Pa = Pa, cos % = (6)(cos 23°) = (6)(0.9205) = 5.523,
14.2 HELIX ANGLE AND FACE WIDTH. The helix angle # and D = N,/P4 = 20/5.523 = 3.62 in. Shaper cutters for helical gears are
Fig. 14.2, is the angle between a tangent to the pitch helix and an inter based on the pitch P, in the diametral plane. Thus helical gears may have a
secting axial element of the pitch cylinder. There are other helix angles Normal pitch that is standard or a pitch in the diametral plane that is
for example the base helix angle y,, which is the angle between a tangeml Standard.
to the base helix and an element of the base cylinder (which contains thi The axial pitch P, is the distance between corresponding points on
400 HELICAL GEARS [Ch. I4 514.6] STRENGTH OF HELICAL TEETH 401

adjacent teeth measured in an axial direction; it is seen in Fig. 14.3 that capacity and to give smoother kinematic action (as compared to a stubbed
P,/P, = tan, or tooth).

P., aD 7
(d) Pi =
tan is N, tan t P, tang 14.55 DYNAMIC LOAD, HELICAL GEARS. Equations (m), (n),
§ 13.15, and others with velocity factors, plus operating and service factors,
are used for estimating the dynamic load on these teeth, which will in
general be smaller than that on similarly loaded straight teeth. For
“average” gears, Buckingham’s equation is
0.05v,,(F; + Cb cos? #) cos #
(14.1) F,
= F, +
FIGURE 14.3 Force System of Helical 0.050, + (F; + Cb cos? ps)?
Gear. Friction is neglected; N = cp is the
total force normal to the tooth at the point where the symbols have the usual meanings. For precision helical gears
Tooth Elements running at above 5000 fpm, the dynamic load can be takenasthetransmitted
in Pitch Surface of contact, with components F; = ta
tial load, S = separating force, Fa = load.
axial thrust. 4

14.6 STRENGTH OF HELICAL TEETH. The Lewis equation is

used to compute the strength of helical teeth. The pitch is that in the normal
plane. Thus if s is the endurance strength, the limiting tooth strength is
taken as (Y is for load near middle if K, is used; §13.10)
sbY
14.4 PRESSURE ANGLES. The pressure angle in the normal plane (14.2) Pa Ib.,
¢, is different from that in the diametral plane. In Fig. 14.3, note K; dn

tan ¢, = gc/pg and that gc = ab; then multiply and divide by ap; where the symbols have the usual meanings, except that Y is chosen in
accordance with the equivalent (also called formative or virtual) number of
(e) tan¢, = — =—— = —— = tan¢cosy. teeth NV,. To understand the significance of this, imagine a plane cutting
a right cylinder obliquely, as the normal plane cuts the pitch cylinder,

angle is ¢, = 20°; and the diametral pressure angle with a 23° helix angle is
A
t 0.364
¢ = tan ( - *) = tan-} ( ) =) 21.577: Tooth
cos 0.9205, |
The interchangeable tooth proportions with 20° pressure angle are often FIGURE 14.4 ays Y
used for helical gears, even though there is no reason other than manufac
turing convenience, except when interference would occur. Where appro-
ea
priate, unequal addendums can be used. Note that for a constant pressur
Normal
_fS ™ Pitch cylinder
angle in the diametral plane, the top lands get narrower as the helix angle Plane
increases. (Sketch this to get the idea.) This leads to heat-treating difficulties;
when the teeth are nearly pointed, the tip comes out excessively hard and
brittle and causes trouble by breaking off. One action to take in this event Fig. 14.4. The shape of the cut section is an ellipse whose minor diameter
is to stub the teeth somewhat, thereby making the pinion tip blunter (whi h is the pitch diameter D, and the profile of the tooth in the plane of the ellipse
is the usual practice in spiral bevel gears, § 15.4). At least one manufacturer is the normal profile. Let the radius of curvature of the ellipse at A be r.
uses a 25° pressure angle and full-depth proportions, which tends to improve The equivalent number of teeth is the number on a gear of radius r with a
 7

FIGURE 14.4 (Repeated).

ws
Pitch Cylinder

pitch equal to the normal pitch. From analytical geometry, we find

r = Dj(2 cos? #), Fig. 14.4. The equivalent number of teeth V, = Py,(2r}
or

(f) N, = P4y = mn (S\y) ty mr

where WV is the actual number of teeth on the gear, P,D. The equivale:
number of teeth, rather than the actual number, defines the form o tl FIGURE 14.5 (left) Sikes Herringbone Gears. The large gear is manufactured from a
tooth in the normal plane. } casting. The teeth are the continuous variety; that is, they are cut up to the center of the
Since wear is usually the determinant for steel gears in continuous sé gear. The helix angle is 30°; pressure angle is 20°. (Courtesy of Link-Belt Co., Chicago).
vice, a reasonable estimate for Y is ordinarily satisfactory; for exampl
Y for NV, interchangeable type teeth in accordance with the closest app FIGURE 14.6 (right) Herringbone Gear Made froma Welded Built-up Blank. A particular
mate standard from Table AT 24, load near the middle. Otherwise machine is necessary to cut teeth to the center of the blank as shown in Fig. 14.5. The space
in the center of this gear is to allow clearance for the cutting tool. (Courtesy Lukenweld
true profile must be developed and Y found as indicated in Fig. 13.6.
Inc., Coatesville, Pa.).

14.7 LIMITING WEAR LOAD. _ Buckingham"*-*) gives the limitiny The hardness of helical and herringbone teeth cut after heat treatment
wear load for helical gears on parallel shafts as will generally fall between the limits of 210 and 300 Brinell for the gear,
with the pinion being some 40 to 50 Brinell points harder than the gear.
bD, OK, A double-reduction helical gear box is shown in Fig. 14.7.
(14.3) Ib.,
cos?
where the symbols have the same meanings as for spur gears, ¢ is the heli
angle, K, is based on the pressure angle ¢, in the normal plane, and th
actual numbers of teeth are used to find Q. Follow the same plan of desigi
as detailed for spur gears.

FIGURE 14.7 Helical Gear Box.

14.8 HERRINGBONE GEARS. All of the foregoing discussio! These teeth are hobbed after heat
treatment, unequal addendum and
applies to herringbone gears, Figs. 14.5 and 14.6, which are double helical
dedendum, full depth. Small gears
gears. The object of the right-hand and left-hand helices is to absorb th made from about 1045 forged blanks;
axial thrust within the gear, thus eliminating the necessity of provid largest gear, cast steel, annealed and
against this thrust in the bearings. In order to divide the load eq heat treated. Gear bore and shaft
between the two parts of the herringbone gear, one shaftis mounted so thatit Surfaces at the gear are ground
floats in the axial direction. Large helix angles, 30° to 45°, are used; hence, finished for close limits. (Courtesy
Westinghouse Electric Corp., Pitts-
the thrust forces are surely large enough to keep both sides in driving con: burgh).
tact. The face width b in equations (14.1), (14.2) and (14.3) should be the
active width; that is, the width of groove, Fig. 14.6, should not be included
in b; minimum b = 4P, [see equation (a) above].

402
 “
514.10] CLOSURE
driver divided by the angular velocity of the driven, is not a ratio of pitch
diameters; but it is the ratio of the tooth numbers, N,, N,. Equation (c)
applied to the driver 1 and driven 2 gives
(g) N, malt aD, 2 cos by and Ny ee Dz, 2 COS Ya

e cn ¥ cn
FIGURE 14.8 Crossed Helical Gears.
ny Wy Nz D, cos bo
(h) n= = SS SE
Ng W N, D, cos yy,
and the center distance C = (D, + D,)/2. For the usual case, 4, + #2 =
zr = 90°, Fig. 14.9; thus, cos #2 = cos (90 — #,) = sin, and cos y,
= cos (90 — #2) = sin #,. Therefore, for the shafts at right angles, equa-
tion (h) can be written
wo, Dgsind, D, tan $, Dz
(i) My = [= = 90°]
w, D, cosy, Dp, D, tan p,
14.9 CROSSED HELICAL GEARS. Helical gears mounted on non: Multiply through the middle terms by D,/D, and note that D,w,/(D,w.) =
parallel shafts are called crossed helical gears (formerly spiral gears). If we v,/vo = tany,; similarly, using the last term, we find v/v, = tan yo;
imagine such gears working together, Fig. 14.8, we see that the teeth where v, and v2 are pitch-line velocities of driver and driven, respectively.
across one another (as a worm thread rubs across a meshing gear too The rubbing speed »v, is in the tangential direction and is v, = v,/cos #, =
an action that is fundamentally different from that of gears on para v,/cos wo, Fig. 14.9.
shafts where the only sliding is up or down the profile. Contact occurs onl; The force arrangement on these gears is as described in detail for worm
at a point, theoretically, and the cross rubbing is much greater than up-anc gears, Chapter 16; therefore, the equation for their efficiency may be de-
down sliding with the result that the capacity of such gears is relati rived in just the same manner as in § 16.8. Adapted to crossed helicals at
small—light power at reasonable speed. Running-in under light load u right angles and with the same hand, it becomes (f = coefficient of friction)
a bright line is visible along the tooth length materially increases the capa
cos ¢, sin #, — fcos #, _ cos ¢, — f cot p,
by increasing the area taking the load and improving the surface for the
lubricant.
() e = cot, cos ¢, cos #, + fsin a ~ cos é, +f tan
[= = 90%]
The shafts can be at any angle &, but 90° is the most common, and the For coefficients of friction, see § 16.9. The literature often quotes
helix angles y, (driver) and x, (driven) can have innumerable values, either Buckingham’s"-3) approach to design for particular power transmission,
left hand or right hand, the relations being"?-?! to which the interested reader should turn; for considerable detail on de-
signing crossed helicals, see Ref. (14.1).
= = $, + or X= p, — pe.
[SAME HANDS] [OPPOSITE HANDS]
14.10 CLOSURE. Helical gear transmissions are the ones most
likely to be running at very high speeds; thus the discussions of the phe-
nomena associated with speed in Chapter 13 are appropriately considered
In this context. Engineers concerned with the design of commercial gear
boxes will check their designs against the AGMA ratings; see Ref. (14.4).
FIGURE 14.9 Crossed Helicals. Diagrammatic, sho’
1 and 2. The dotted outline is driver 1, imagined on top of 2+
The tangent line ab through 0 of gear 1 takes the position
mn after 180° turn of 1.
 I5. BEVEL GEARS

15.1 INTRODUCTION. Bevel gears are used to connect inter-

secting shafts, usually but not necessarily at 90°, Fig. 15.1. Bevel-gear
teeth are subjected to much the same action as spur and helical teeth; the
total maximum load on a tooth is compounded of the transmitted load
and a dynamic increment arising from profile and tooth-spacing inac-
curacies, and the maximum surface compressive stress is the principal
criterion of wear resistance. Since bevel gears are inherently not inter-
changeable, they are designed in pairs. This chapter also mentions briefly
a few of the skew gears.

15.2 BEVEL-GEAR NOMENCLATURE. Bevel-gear teeth are built

with respect to a pitch cone, rather than to a pitch cylinder as in spur gears.
The elements of the pitch cone intersect at the cone center, Fig. 15.2. When
two bevel gears are properly mounted, their cone centers are coincident.

FIGURE 15.1 Straight Bevel Gears. (Courtesy Link-

Belt Co., Chicago).

407
 BEVEL-GEAR TEETH 409
315.3] STRENGTH OF STRAIGHT

of teeth on the pinion, N, the number of teeth on the gear. The addendum
angle « and dedendum angle 6 are given by
Pah: addendum ne dedendum
a = tan-? ———_____ 6 = tan-? —_____.,
cone distance cone distance
Cone
oa
@
&5
Diameter = Do-" “="

15.3 STRENGTH OF STRAIGHT BEVEL-GEAR TEETH. A

straight bevel gear, Fig. 15.1, is one whose tooth profiles consist of straight
elements that converge to a point at the cone center. Since the force on a
pevel-gear tooth varies from point to point along the face, we may resort
to the calculus to determine the tooth strength in terms of a force that can
be computed from the horsepower and the speed. See Fig. 15.3. The differ-
a

FIGURE 15.2 Bevel-gear Nomenclature. ential force dF, acts on an increment of the face dx at a distance x from
order to eliminate fillet interference at the si the cone center O. Since dx is very small, the variation of force along this
ends of the teeth, the more recent practice i length may be neglected and the Lewis equation applied; F = sbY/P4,
Back Angle turn the face cone parallel to the root cone of th
Addendum Angle, @ Backi mating gear: hence, its elements do not inters : sY dx
Dedendum Angle, 6 zn ane
at the cone center, but at some point B, where th (a) aF, = —,
—-Mounting Distance dotted line intersects the axis. dz

where P,, is the diametral pitch at the increment dx. Multiplying both sides
of (a) by r,, the radius of the gear at the distance x from O;
The length of a pitch-cone element is called the cone distance, dimensio sYr,dx
Fig. 15.2. Top lands and bottom lands, Fig. 15.1, mean the same as befo | r, aF, = Ps
The cone formed by the elements of the top lands is called the face co dz

the cone formed by the elements of the bottom lands is called the ni The sum of the product of all elemental forces by their radii fr,dF, is
cone. The angle between a pitch element and the axis is called the pi equal to the torque 7 transmitted by the gear. Therefore,
angle (do not confuse with the pitch angle in spur gears, § 13.4). See
(b) r f° Yr, dx
15.2 for the meanings of root, face, addendum, and dedendum angles. i
Since a section of the tooth decreases in size as it approaches the con i F, dz 4

center, there is not a single value of the pitch, pitch diameter, addend Since all elements of the teeth on a straight bevel gear converge to the
dedendum, whole depth, etc. Unless otherwise indicated, however, thi cone center O, the circular thickness and, consequently, the circular pitch
names used alone refer to the largest tooth section, Fig. 15.2. The pitch vary as the distance from O. Thus P,,/P, = x/L, where P,, is the circular
a bevel gear and the pitch diameter are those at the large end. The b
cone is an imaginary cone whose elements are perpendicular to the pit
FIGURE 15.3 Strength of Bevel-gear Teeth. The value of the form factor Y depends on
cone elements at the large end of the tooth. The length of a back-cone elem the form of the tooth in the plane normal to a pitch element. The theory of equation (f) is
is the back-cone distance, r, in Fig. 15.2. off slightly when constant clearance is used.
The pitch angle y is computed as follows (2 = angle between shafts):
Normal: Profile
; Root Cone
Seen in This
1 Direction of Gear 2
x = 90° tan yy = = = Mg tan Vp = 3
Constant Clearance when
P 9 Face Cone Is Parallel
sin & m, sin & Equivalent
=<90° 8 tany, =
N,/N, + cos = 7 Bop m, cos &
The pitch angle on the pinion y, is equal to & — y,; N, denotes the number
Back-Cone
 Seen in This awe Vw

Direction von
. Constant Clearance when
; 3 3 Face Cone Is Parallel 15.5] FORM FACTOR au
Che: ~ to Root Cone 2
Equivalent
G
SAS . ™
assumed above, modern practice in bevel gearing is to plan for localized
Ton contact, as suggested by Fig. 15.9. Moreover, there are well recognized
m1 relations between b and L (§ 15.4), so that the factor (L — 5)/L can be
Ts
Back-Cone
cared for in a blanket fashion.
Element EZ ¥ :
() Gear i vie O= Apex of
p Pitch Cone
15.4 TOOTH PROPORTIONS, BEVEL GEARS. In bevel gears
FIGURE 15.3 (Repeated). that should have good conjugate action, the teeth are generated. Since
pitch at a distance x from the cone center. Since the diametral pitch vari bevel gears are not interchangeable and since the generating process is
inversely as the circular pitch, we get adaptable, unequal addendum tooth forms are used."5-1] The desired
qualities in gears are quietness, strength, and durability. Good kinematic
L design can contribute materially to quietness. The addendum on the pinion
(c)c Bf
— = —pil =>,tg or Pa; ei =P ma(=). is increased, its dedendum is decreased, § 13.34, eliminating (or perhaps
nearly so) interference conditions and undercut flanks, but keeping the
Moreover, from similar triangles in Fig. 15.3, we find the proportion _
working depth at 2/P, for straight and Zerol bevels, Table 15.1. Although
(@) r, Ng (=) bevel gears with other pressure angles can be obtained (144°, 174°, etc.),
—=-, or r, =r{—}. ¢ = 20° has been adopted as standard in the interests of simplification.
r L e L
The combination of shortened dedendums on the pinion and the 20°
Substituting the values P,, and r, from (c) and (d) into (b), we obtain — \pressure angle seems to give a reasonably good kinematic compromise.
sV¥i Recess action increases with the velocity ratio, so that a larger proportion
©) jens
PyL | x? dx, of the angle of action is recess, which contributes to quietness (smoothness).
Spiral bevel gears, § 15.11, have a working depth of 1.7/P4, because the
in which, s, Y, r, Pg, and L for a particular gear are each constant; r spiral angle reduces the width of top lands and shorter teeth counteract this
the pitch radius of the gear, P, the diametral pitch, and L the cone dis effect (§ 14.4).
Since the origin of the coordinates is O, the limits of the integral The face widths that should be observed for the design procedure to
should be from (L — 5b) to L to include the whole face b. Integration thi be described are:
gives Straight and spiral bevels: max. of L/3 or 10/P4, whichever is smaller;
recommended b = 0.3L.
sYr [x3 72 sYrb b b?
:
f
®
a
P,L? | 3
ni
5 L—b Pa Megara L 312
Zerol bevels (Fig. 15.10): max. of 0.275L
smaller; recommended b = 0.25L.
or 10/P,, whichever is

is customary to use the force F obtained from the horsepower equation

15.5 FORM FACTOR. Looking in a direction parallel to the axis
the pitch-line velocity at the large end. Dividing both sides of (f) byt
of a bevel gear, Fig. 15.3, one sees the tooth foreshortened. The value of
radius r at the large end, we find F in the form Ԥ
Y in equation (15.1) depends on the shape of the profile normal to a pitch-
. Loe
r
ee
Py Lyx)! ALAM
Cone element; if this profile is an interchangeable type, the value of Y
Would be chosen from Table AT 24 in accordance with the equivalent (also
Called formative or virtual) number of teeth N,. The equivalent gear is one
: The last term within the bracket is negligible when b = L/3; that has a radius equal to the back-cone radius r,, Fig. 15.4, with a pitch
sYb(L — b) €qual to the pitch at the large end (heel). Thus, for the pinion,
(15.1) F, = ———_.,
PiL
where F, represents the strength of the tooth, stress concentration neglec te
(h) Nig Beg Piome ae
'
Inasmuch as it is practically impossible to get the load distributed| Where we have used r, from r,/L = r,/r,, and L = (r,? + r,”)*/?, as seen
410
412 BEVEL GEARS [Ch.
5157] RATED STRENGTH OF BEVEL GEARS 43
from Fig. 15.4. The equivalent number of teeth for the gear V,, = 2r,LPa/i Ney = service factor, chosen from Table 15.2, and VF = velocity factor.
is found similarly. If the bevel gear teeth are cut with a milling cutt Use VF = 1 if the dynamic increment is negligible. For commercial straight
they may be an interchangeable type but inaccurate; in this event, let pevels, use the factor in equation (n), § 13.15, which is
dynamic load be F, = F,(600 + v,,)/600 and make F, 5 Fy, as outline
above with Y chosen for the load at the tip. Also v,, fpm at the large er 50 + v,1/2
(i) i aati . [STRAIGHT]
should be less than 1000.

For commercial-quality spiral bevels,

ae 78 + v_t/?71/2

iy
yhoo FIGURE 15.4
(i) VE = [=| . [spit]

SQ OG UY Straight and Zerol bevel gears should not

speed is greater than about 1000 fpm."5-7) Spiral bevel, Fig. 15.11 and
be used when the pitch-line

Las Hid EE hypoid gears Fig. 15.13, are recommended when the pitch-line speed
Leneae exceeds 1000 fpm or where angular speeds exceed 1000 rpm. Of course,
they may also be used for lower speeds when quietness is important. When
the pitch-line speed is above 8000 fpm, the teeth should be ground after
If, however, accurately generated, unequal-addendum teeth are use hardening. Because of the differences between spiral angles, hypoid pinions
it is appropriate to assume that the load is shared by two teeth until cont; are larger than spiral-bevel pinions of the same capacity, which makes it
is carried by one tooth in the vicinity of the middle of the profile, as ¢ possible to use a larger shaft on the hypoid pinion than on the spiral pinion.
plained for spur gears, at which point the value of Y is considerably lar;
than that when the load is at the tip, but there are other factors; see the ge
metry factor J, § 15.7. 15.7
; RATED STRENGTH OF BEVEL GEARS. The strength equa-
tion to match F, in equation (15.2) is$-1)

15.6 DYNAMIC LOAD FOR GENERATED BEVEL GEARS. (15.3) F, = al

shall let the subsequent calculations for bevel gear teeth serve to illustre P, K,K,K;
the AGMA approach, although minor rearrangements will be made where K, = size factor; for bevel gears, choose K, according to pitch:
order to conform with our dynamic-load point of view. The dynamic lo:
is computed from Ps" 1 2 3 4 6 8 10 over
(15.2) Fa = (VE)NosKnF F K,: 1.00 0.84 0.76 0.71 0.64 0.59 0.56 0.50
where the transmitted load F, = 33,000 hp/v,, lb. and v, = 7Dn fpmist K, = life factor for strength; for indefinite life, 6 x 107 cycles or more,
pitch-line speed at the large end, and K,, = load distribution factor,| use K, = 1; for 10® cycles, K, = 1.4; for 10° cycles, K, = 2.1; for 104
tached to the computation of F, for convenience. Easy means of determ cycles, K, = 3.1; for 10° cycles, K; = 4.6;
ing K,, are given"?-1.15-1] when the deflections of the gear and pinion 4
known; bevel gears are sensitive to misalignment, including that due as temperature factor; use unity if t°F < 160°F; acceptable value,
deflection. Experience has shown that the life of bevel gears increa es = (460 + 1°F)/620.
the rigidity of the mounting is increased. In the absence of detailed kno = reliability factor; for 1 failure in 3, K, ~ 0.75; for 1 in 100,
ledge, choose values as follows :"19-1.15-341 ‘ -12; for high reliability, K, = 1.5 or more; for maximum safety,
wm

Both gears straddle mounted (Fig. 15.18)

One gear straddle, one not : : ‘ . Ke =1.1 tol
J = geometry factor for strength, Fig. 15.5 for straight bevels, Fig. 15.6
for Spiral bevels with ¢ = 35°; 20°
Both gears overhung pressure angle, 90° shaft angle. J cares
 44DLE i9. LUUIH DIMENSIONS KOR LADLE 15.4 SERVICE FAULIUNRD (Vsfps BEVERY VERARINYG
STRAIGHT AND ZEROL BEVEL GEARS For speed-increasing drives, add 0.01m,? to the factors given.(*°-°)
The addendum of the gear is a, = A/P.; of the pinion, a = 2/P, —a, where 4 is taken Character-of-load examples are suggestive only. See also § 13.18.
from the tabulation below. Whole depth = 2.188P, + 0.002 in.; working depth = 2/P,.
MODERATE HEAVY
Dedendum = whole depth — addendum. ‘ i UNIFORM SHOCK SHOCK
For drawing purposes only, gear and pinion teeth may be considered of equal thick- i POWER SOURCE (centrifugal | (lobe blowers, | (ore crusher,
ness as measured halfway between the top and bottom of the tooth. Approximate yalues mach., belt multicyl. single-cyl.
for spiral bevel gears can be obtained by multiplying these values by 0.85. conveyor) pumps, compressor,
machine tools) | punch press)
GEAR RATIOS GEAR RATIOS

Son ae - ee i 2 Uniform (elec. motor, turbine) so : i is

1.00 1.00 1.000 1.42 1.45 icyl. ICE) - 1.25 ; .
0.760 ; ene ee Pas 15 1.75 2.25 +
1,00 1.02 0.990 1.45 1.48 0.750
1.02 1.03 0.980 1.48 4.52 0.740
1.03 1.04 0.970 1.52 1.56 0.730
oe a er es se pie 15.8 RATED WEAR LOAD FOR BEVEL GEARS. The equation
ra ine dead i iG ong | for the rated wear load originates with the Hertz contact-stress equation
1.08 1.09 0.930 1.70 1.76 0.690 , as does the Buckingham equation. To be compared with F, from (15.2),
1.09 1.11 0.920 1.76 1.82 0.680 it is i :
1.11 1.12 0.910 1.82 1.89 0.670 i Sea Cc,
1.12 1.14 0.900 "1.89 1.97 0.660 } (15.4) Fy = Dbl aa (=) ,
’
1.14 1.15 0.890 1.97 2.06 0.650 1 © ce "
Fibs
Liz
sie
1.19
pier
0.870
— oo. cee where D, = pitch diameter of the pinion, b = face width, K, is as defined
2.16 2.27 0.630 4 in § 15.7

oe o bf a — oon Seq = design contact stress, whose values (and s,q*) are found in Table
1.23 1.25 0.840 2.58 2.78 0.600 a 15:3. f I : is
1.25 1.27 0.830 2.78 3.05 0.590 i C, = life factor for wear; for indefinite life, 107 cycles or more)
4:27 1.29 0.820 3.05 3.41 0.580 i C, = 1; for 10° cycles, C, = 1.15; for 10° cycles, C, = 1.3; for 10* cycles,
1.29 1.31 0.810 3.41 3.94 0.570 " C=l5
1.31 1.33 0.800 3.94 4.82 0.560 mt + ae
1.33 1.36 0.790 4.82 6.81 0.550 i
1.36 1.39 0.780 6.81 co 0.540 4 TABLE 15.3 DESIGN TABLE 15.4
1.39 1.42 0.770 4 CONTACT STRESSES, sca ELASTIC COEFFICIENTS:
, For bevel gears; BHN’s are minimum. For bevel gears. Values of E ksi useq are:
, Reduce s,, by 10% unless the manu- steel (3 x 10*), cast iron (1.9 x 0%).
for: the form factor Y, the strength reduction factor K;,, effect of rad facturing and materials are highest aluminum bronze (1.75 x 10*), tin brdgze
quality29.1,19.94,18.1] (1.6 x 104). |
load, load sharing (m,), effective face width, tooth taper. !15-1)
b = face width, and P, = pitch at the large end. MATERIAL a eae
Sq = design flexural stress, as follows: MATERIAL (BHN) Sea ksi | Sea” (ksi)? Monet (psi) t
Steel, Min.BHN: 140 150 180 200 300 400 450 55-63 Stele) ae Steel and
Stel . . .| 2800 | 7840
Sq ksi: 11 12 14 15 RO 287 ay 30° Steel (300) ; . | 135 | 18,200 Castiron . . | 2450 | 6000
} Steel (360) . .| 160 | 25,600 Al. bronze . .| 2400 | 5760
R, = 55-63 is for carburized teeth. Flame or induction hardened wi Tin bronze . -| 2350 | 5530
unhardened root fillets, surface 450-500 BHN, sg = 13.5 ksi. For ¢ Steel (440) . «| 190 | 36,100 moe 4
iron: Class 20, 2.7 ksi; Class 30, 4.6 ksi; Class 40, 7 ksi. For nodular iro oa ee ie : se " oe nee
use the value for steel at the same BHN less 10%. Cast iron, ont Pir aaa "900 Al. bronze . .| 2200 | 4840
Compare (15.3) with the strength equation (15.1) and note the facto Tin bronze . - | 2150 | 4620
not involved in the strength theory. For no breakage F, S Fy. Obser' Cast iron, 30 (175) .| 50 | 2,500
that each gear has its geometry factor J; check and design for the too Cast iron, 40(200) .| 65 | 4,230 Al. bronze and
with the lower value Bronze (10-12% tin). 30 900 Al. bronze - | 2150 4620
: Al. bronze, HT 65 | 4,230 Tin bronze . + .| 2100 | 4410
a4 Nodular iron (use
0.9 x steel value at
same BHN) -
 Number of Teeth in Gear for which Geometry Factor is Desired

&°o

Number of Teeth in Pinion

“N
°o
2

30
>
sSo

20
w&
°o

20 10
.020 .040 .060 .080 -100 -120 -140
Geometry Factor - |
b160 0.200 0.240 0.280 0.320 0.360 :

Peoneky rotor—4 FIGURE 15.7 Geometry Factor J for Coniflex Straight Bevel Gears (¢ = 20°, 2 =
90°). (Courtesy Gleason Works, Rochester, N.Y.) .
FIGURE 15.5 Geometry Factor J for Coniflex Straight Bevel Gears (¢ = 20°, 2:
90°). (Courtesy Gleason Works, Rochester, N.Y.).

FIGURE 15.8 Geometry Factor J for Spiral Bevel Gears (¢ = 20°, % = 35°, Z=
90°). (Courtesy Gleason Works, Rochester, N.Y.).

FIGURE 15.6 Geometry Factor J for Spiral Bevel Gears (¢ = 20°, 4 = 35°, &
90°). (Courtesy Gleason Works, Rochester, N.Y.) .

Number of Teeth in Mate

100
8

Number of Teeth. in Pinion

Number of Teeth in Gear. for which Geometry Factor is Desired
8
3
s
8

40

30 .060 -080 -100 120 -140 -160 -180

Geometry Factor - |
20
417
Bio 0.200 0.280 0.320 0,360
 Pitch Line of Tooth
Midpoint
of Face
515.12] HYPOID GEARS ai

, Fig. 15.11, should be large enough that the arc of spiral is some 1.25P,
to 1.4P, (face contact ratio 1.25 to 1.4). Typical spiral angles are around
30° to 35°.

15.12 HYPOID GEARS. Looked at individually, hypoid gears, Fig.

15.13, appear much like spiral bevels, but their pitch surfaces are not the
same. Instead of a pitch cone, the pitch surface is a hyperboloid. When two
hyperboloidal pitch surfaces are in contact along a common element,
their axes do not intersect, which is the way in which they most obviously
FIGURE 15.11 Spiral Bevel Gears. The inset in the upper right-hand corner shows the
differ from spiral bevels. These gears are generally used in automotive
meaning of the spiral angle y. The pinion has a left-hand spiral, the gear a right-han
spiral; apply the right-hand rule. (Courtesy Gleason Works, Rochester, N.Y.). rear-axle drives because their use makes it possible to lower the center of
gravity of the car. Since a pair of mating pitch circles do not roll on one
another, there is sliding lengthwise along the tooth (analogous to crossed
helicals and worm gears) which results in considerably more heat genera-
tion. On this account, an EP lubricant (§ 13.38) is needed for a good power
capacity. The offset of the shafts may be so great that the shafts may con-
tinue past each other; this characteristic is used for providing multiple
take-offs from a single shaft on which several drive pinions are mounted.
Skew hypoid gears, Fig. 15.14, are similar to curved-tooth hypoids
except that the teeth are straight. Since generated curved-tooth hypoids are
readily manufactured, this variety is not so widely used.
Spiral bevel and hypoid gears are recommended when the pitch-line
speed exceeds 1000 fpm or where angular speeds exceed 1000 rpm. Of
course, they may also be used for lower speeds when quietness is important.
When the pitch-line speed is about 8000 fpm, the teeth should be ground

FIGURE 15.12 Machine for Cutting Spiral-bevel or Hypoid Teeth. An automati

machine. In generating a tooth profile, ‘‘a relative rolling motion takes place between
gear ... and the rotating cutter; the action is as though the gear ... were rolling
mating crown gear of which the cutter represents a tooth... . At the completion of
tooth space, the cradle and work reverse direction and the work is withdrawn and
takes place.’’ The sequence is repeated for each space. (Courtesy Gleason Works, Ri FIGURE 15.13 Hypoid Gears. (Courtesy Gleason Works,
ter, N.Y.). Rochester, N.Y.).

15.11 SPIRAL BEVEL GEARS. __ These gears, Fig. 15.11, which af

to straight bevel gears as helical gears are to straight-tooth spur gé
have much the same advantages as helical gears; smoother tooth en
ment, quietness of operation, greater strength, and higher permissible speed:
(Read § 14.1 again). The profiles are modified slightly to obtain the advar
tages of localized contact, which also contributes to the good runnifl
qualities of these gears. Since the teeth are cut with a rotary cutter, Fi
15.12, a good rate of production of accurately cut gears is attainable. TH FIGURE 15.14 Skew Hypoid Gears, Straight Teeth.
standard pressure angle is 20°; also used are 144°,16°, 174°. The spiral ang (Courtesy Gleason Works, Rochester, N.Y.).
422 BEVEL GEARS [Ch,|

after hardening. Because of the differences between spiral angles, hypo} FIGURE 15.17 Forces on a Straight Bevel-gear Tooth. ere e
The forces N and F, are components of the force W; Ncos_ ~- Direction
pinions are larger than spiral-bevel pinions of the same capacity, whic
and N sin y are components of N. The force F4 acts at A
makes it possible to use a larger shaft on the hypoid pinion than on f normal to the plane of the paper.
spiral pinion. This characteristicmay sometimes suggest the use of hypoj
gears, even though it is not necessary to offset the shafts for any oth
reason.
Skewed gears have appeared in a number of forms, another of which
the Spiroid gears,"°-2] Fig, 15.15. Virtually, the pinion has beco:
conical worm, so that this drive is appropriate for high velocity ratios. § The force N = F, tan ¢ acts normal to the pitch-cone element. The end
also Ref. (15.4).
thrust on the shaft that must be cared for isN sin y. The bending force on
the shaft in the diametral plane is the resultant of F, and N cos y, which are
forces at right angles. These forces are useful in shaft and bearing design.

(b) Axial thrust on Spiral Bevel Gear. If the driving member has a
right-hand spiral (see Fig. 15.11) and rotates clockwise, or if it has a left-
FIGURE 15.15 Spiroid Gears. (Courtesy Tix
hand spiral and rotates counterclockwise, the end thrusts are as follows:
Tool Works, Chicago). 4
an ae
Driviriving : b
member, Fp ‘ = =k 4 f an xo n cos mee]
y — —————
- cos
Tigiae:
\ —> tan ¢ sin y
Driven member, Fr= F, [saa cos y + ae
FIGURE 15.16 Angular Gears. (Courtesy Lin
Belt Co., Chicago). nq If the driving member has a right-hand spiral and rotates counter-
clockwise, or, if it has a left-hand spiral and rotates clockwise, theend thrusts
are as follows:

Tai gies
15.13 OTHER TYPES OF BEVEL GEARS. When a pair of a ——> tan ¢ sin y
Driving member, Fy, = Fy, [san (COs yt
gears of the same size are on shafts intersecting at right angles, they cos x
called miter gears. The velocity ratio is of course one, and y, = y, =
pee,
Angular gears are bevel gears mounted on intersecting shafts at 2 : <— tan ¢ sin y
of other than 90°, Fig. 15.16. Driven member, Fy, = F, [saa cos y — muPy EE
A crown gear is one in which the pitch angle, Fig. 15.2, is 90°; that:
the pitch cone has become a plane. where F, = axial thrust,
Pipe average transmitted load tangent to the pitch circle at midpoint of
the face,
¢% = spiral angle, Fig. 15.11,
15.14 FORCES ON A BEVEL GEAR. (a) Straight Bevel Gear. Ke
Y = pitch angle of the gear whose thrust is being computed, Fig. 15.2,
resultant of all the forces acting along the face of the tooth may be assume ¢ = pressure angle in a plane normal to a cone element,
without serious error to act at the midpoint of the face. The direction of f - indicates that the thrust is toward the cone center, and
total force is along the pressure line, which is in a plane perpendicular to > indicates that the thrust is away from the cone center.
pitch-cone element at that point. The pressure line makes an angle equal
the angle of obliquity ¢ with the tangent to the back-cone circle. The avera The resultant direction of F, depends upon which term is greater, > or<.
driving force F,, Fig. 15.17, is found from the horsepower equation, whe he direction of rotation is the direction to an observer looking from the
the velocity at the mean point of the face A is used. back of the member toward the cone center. Since the end thrusts may be

423
 0 Mf
SS 7
| Hp SS
eS g15.17] CLOSURE 425
A Tiara ULLTT Te,
peat-treated, alloy-steel gears whenever strength and wearing qualities are
important and the sizes of the gears are within the limits for heat treatment.
ea oe Rn <K An excellent construction is a carburized alloy steel, such as 2512, 4120,
NSE OG 3620, etc., with the surface hardened to a minimum of Rockwell C60
(see Table 15.3) and with a finish grinding in order to insure accurate pro-
a \
SS 9

files. However, excellent gears are obtained from oil hardening or flame
PSS SSSSSSSSSSSSSSS
es SS

nardening such materials as 4640 or 8640. Nitrided gears are also used.
ai] Nonferrous metals, plastics, die castings, and sintered materials are used
“7
when there is reason. The discussion of materials in § 13.24 is also appli-
(a) Straddle Mounting (b) Overhung Mounting
cable here.
FIGURE 15.18 Bevel-gear Assemblies. This figure shows the steel ring gear, with te¢
cut on it, bolted to a cast flange. Since the straddle mounting in (a) is relatively exp
the pinion is frequently overhung as in (b). The distance between the bearings on an ove 15.17 CLOSURE. Since there are many different gear forms, no one
pinion, as in (b), should not be less than 2e, where e is the distance from the point of appli book covers all of them, much less a general text on machine design. If it
tion of the load on the tooth to the nearest bearing. (Courtesy
ter, N.Y.).
Gleason Works, Re che
is desired to check a design against the Buckingham approach, see Ref.
(13.3). Observe that with the life factor, one can design for limited life.
The design procedure outlined in this chapter has been adapted to other
very large, often as much as half of the transmitted load, they must alw a types of gearing.4%-1)
be cared for in the bearing design. a

15.15 DESIGN DETAILS. The design shown in Fig. 15.18, in whie

the gear is a steel ring attached to a separate hub or center, is especia
adapted to bevel gears. A steel ring, preferably with a flat back, has the me
suitable shape for a hardening operation, and the center can be re-enforce
with ribs if necessary. However, a center having a conical shape of we
is generally best, especially for high speeds, because the casting strains 4
less, it is easier to machine and balance, and because the oil in the gea
case will not be so severely churned. a
The rim, bead, and hub proportions may be fixed according to the i
structions for spur gears, except that the hub should be long enough to g
a positive backing, Fig. 15.2. Bevel-gear mountings should include
arrangement, such as a spacing washer, for adjusting or locating the gea
correctly and should provide for lubricating and guarding both the bearinj
and the gear teeth. 4
Accurate alignment in mounting and good support against deflectior
are major requirements for bevel gears of all types. For this reason, ba
and roller bearings are recommended; they can be arranged to resist
change of position, either axially or radially. The distribution of bearin
pressure between the teeth and the amount of the backlash shouldt
checked.

15.16 MATERIALS FOR BEVEL GEARS. Although cast-iron am

ordinary steel bevel gears are used extensively, the tendency is to empl

424
 16. WORM GEARING

16.1 INTRODUCTION. Worm gearing, Fig. 16.1, is used to transmit

power between nonintersecting shafts, nearly always at right angles to each
other. Comparatively high velocity ratios may be obtained satisfactorily in
a small space, though perhaps at a sacrifice in efficiency as compared to
other types of gears. The impacting contact at engagement of spur and
other gears is absent in worm gearing. Instead, the threads slide into con-
tact with the gear teeth, an action that results in quiet running if the design,
manufacture, and operation are proper. The greater sliding sometimes
introduces a problem in heat from friction. Under extreme loading condi-
tions, any gear case may overheat. The principles discussed later in this
chapter regarding the heat dissipation of worm-gear cases may be applied
to others. We see then that for worm gears, we must check our designs not
only for strength and wear but also for overheating.
The section of a worm thread in a diametral plane is generally straight-
Sided, as is the section of an involute rack tooth. If the worm were moved
Without rotation in a straight line perpendicular to the axis of the worm
wheel, the action of the teeth in a plane through the worm axis and normal

FIGURE 16.1 Worm and Worm Gear. (Courtesy The Grant Gear Works,
Inc., Boston).

427
428 WORM GEARING [C . 516.3] STRENGTH OF WORM-GEAR TEETH 429

to the gear axis would be similar to the tooth action of a rack and ge on the worm and the pitch helix on the gear each have a helix angle and
The straight sides of the worm thread facilitate production from the star a lead angle; A, = lead angle of the gear, 4 = helix angle of gear, and A, +
points both of quantity and accuracy. The worm threads may be cut o; p= 90°. If the shaft angle is 90°, the lead angle on the worm A = ¥, the
lathe or with dies, or they may be milled, shaped, rolled, or hobbed. T helix angle on the gear; also ,, = A,. Equation (b) applies to the gear with
worm wheel should be hobbed to match, Fig. 13.11. these modifications.
Whereas power screws are generally single-threaded, worms usually
4 have multiple threads, unless mechanical advantage is more important than
16.2 PITCH AND LEAD. The meaning of the terms pitch and le efficiency. When a worm makes one turn, a point on the pitch circle of the
for a worm are the same as explained for power screws. See § 8.18 and B
8.22, p. 247, now. There tends to be a slight confusion of terms since,

{tage
the worm, the pitch is the axial pitch P,, which when the shafts are at rig
angles, is the same as the circular pitch P, on the worm gear; in this c ee ah
P, = P,, since we shall have no occasion to use the axial pitch of the gi FIGURE 16.3. Lead Angle, A.
paige ala
joint TD

worm wheel moves through an arc equal to the lead, L = N,P,, but for
computing the velocity ratio, we find it easier to use the number of threads
N, (or starts) on the worm and the number of teeth on the gear V,; m, =
N,/N,. Also equation (h) of § 14.9, which is m, = D2 cos #2/(D, cos ),
FIGURE 16.2 Triple Thread.
applies with #2 = A, the worm’s lead angle, and cos ¥, = cos (90 — A) =
sin A; hence, with the worm driving,
wy, Nz D,cosa Dg
) Mee, N,N
OD ean
where V, = 1 for a single-threaded worm, N, = 2 for a double thread, etc.
If high-efficiency power transmission is the objective, four, six, or more
As for helical gears, worms have a normal pitch P,,,, Fig. 16.2, but the le threads may be used on the worm. See § 16.8.
angle of the worm A, which is the angle between a tangent to the pitch ] Standard pitches. Consider the following axial pitches as standard:
and the plane of rotation, is more convenient than the helix angle y,, 0
4, 5, 3, 4, & 3 1, 14, 14, 13, and 2 inches. Diametral pitches are also
worm (§ 14.2). From Fig. 16.2, used in worm gearing.
(a) Pon = Pg 08.A = P6068 A,
The axial distance that a helix advances in one turn is the lead L. Whi rn 16.3 STRENGTH OF WORM-GEAR TEETH. The teeth on the
on the pitch cylinder, whose diameter is D,,, or on the addendum o: worm gear are weaker than the threads on the worm. The Lewis equation
cylinder, etc., all helices making up the thread have the same advance ‘5 (13.5) without K; is usually applied:
but the /ead angle changes. If we imagine one turn of a pitch helix unwrapp
from Fig. 16.2 into a plane, we have the diagram of Fig. 16.3, where’ (a) ae sYb _ sYbPon
see that » Up 7

tanA = Leer us7D. The normal circular pitch P,, is introduced into the equation because
(b) 7D W Circular pitch is most often used for worm gearing. When s = s,, the force
e n F, as obtained from this formula conservatively represents the bending
and that A + ¥,, = 90°; L = lead of the worm thread. The pitch hel €ndurance strength of the tooth on the gear; b is the effective face width of
430 WORM GEARING [Ch, §16.7] RELATION OF NORMAL AND DIAMETRAL PRESSURE ANGLES 431

the gear. If the sum of the teeth in the worm and worm gear is greater tha In many situations, the horsepower capacity is dictated by the radiating
forty, the following blanket values of Y may be used with safety.43-31 _ capacity of the housing, which in turn depends upon environmental con-
For: ¢, = 144°, ¥ = 0.314; $, = 25°, Y = 0.470; ditions.
¢, = 20°, Y = 0.392; ¢, = 30°, Y = 0,550,
The heat Q to be dissipated from any gear case is equal to the loss due
to friction, which is taken as the horsepower input hp, times (1 — e)
16.4 DYNAMIC LOAD FOR WORM GEARS. The dynamic loz where e is the efficiency of the drive as a fraction (§ 16.8); thus
on worm gears is often estimated from () Q=(—e)hp) hp or = (1 — e)(hp,)(33,000) ft-Ib/min.
= (1 — e)(hp,)(2544) Btu/hr., etc.
1200 + Ung
16.1 F pec l aoe ae The amount of heat that the housing dissipates by convection and
ah) ‘ ( 1200 ) : radiation depends upon the following: the area of the housing, the tem-
where F, is the transmitted load computed from the horsepower equatii perature difference of the housing and the surroundings, and the transmit-
applied to the gear; F, = 33,000 hp/vm,, and ¥,, fpm the pitch-line spe tance or coefficient of heat transfer h,,, which in turn is a function of the
the gear, where strictly hp is the output power, but since the efficienc y temperature, the velocity of air across the housing, and other variables.
unknown at the outset of design, it is conventional to use the input powe See § 11.18. Since the environmental temperatures encountered ordinarily
make F, 5 Fy. As given by (16.1), Fy is likely to be on the high side. — vary little, 4,, varies most significantly with the size of the housing and the
air velocity. Values of A, ft-lb/min-sq. in.-°F are shown in Fig. AF 21
for average conditions of natural ventilation. The heat-dissipating capacity
16.55 WEAR LOAD FOR WORM GEARS. Buckingham gives ; of the housing (gear case) is then expressed as
approximate relation for the limiting wear load F,, as'14:1)
(f) Q, = h,,A At ft-Ib/min.,
(16.2) fF, = DpbkS
where A sq. in. is the radiating area of a smooth housing (the area of the
where D, in. is the pitch diameter of the gear, b in. is the effective face base is not counted) and At is the temperature rise of the lubricant above
and K,, is a factor obtainable from Table AT 27, which depends upon the ambient temperature. The maximum temperature of the lubricant
materials used, the pressure angle, and the lead angle. The approxima should preferably not exceed 190°F (150°F in other types of gear cases).
values of Table AT 27 ignore change due to change of lead angle. If For heavy-duty worm-gear speed-reducers with integral worms, the AGMA
material of the worm is soft or rough, wear will be rapid. A hard, sm recommends a minimum area of housing exclusive of the base, flanges,
worm smooths and cold works the gear’s softer tooth surfaces, and for and fins of
reason, the surface of the worm should be made as smooth as possible.
best combination has been generally considered to be a case-hardened (g) Amin = 43.2C?-7 sq. in.,
burized) and ground worm (though polishing helps) with a chilled-br where Cin. is the center distance. In case the natural heat-dissipating
gear (but see high-test cast iron and same in Table AT 27). Recall that t capacity of the housing is not enough to maintain the temperature at a
theoretical contact stress is calculated from the normal force and that ther reasonable level, extra cooling is obtained by (1) circulating the lubricant
fore there is always a certain amount of approximation in using the and cooling it outside of the housing, (2) circulating water in cooling coils
gential force, as we have been doing, especially in the case of the lar inside of the housing, or (3) blowing air across the housing, as illustrated
lead angles common in worm drives. See § 16.11 for suggested b values. in Fig. 16.4 (p. 432); in this case fins for extra heat radiating area are helpful.
If the service is continuous, F,, should be equal to or greater than F The amount of coolant (water) circulated in coils and the surface area of coils
needed may be computed on the assumption that the coolant carries away
all the heat of friction.
16.6 THERMAL CAPACITY. If the gear case gets too hot, the Oo
may become too thin; it is then easily squeezed from the contact surface:
If this happens, friction increases; more heat is generated; and fina | 16.7 RELATION OF NORMAL AND DIAMETRAL PRESSURE
serious abrasion and scoring occur. Extreme-pressure (EP) lubrica ni ANGLES. A study of the contact forces will lead to an expression for
reduce such troubles when driver and driven are made of steel (see § 16.12) efficiency that is theoretically correct for any type of thread-fastening screws,
 SOOT SINT Ose ee eee ne R
160)
of the total plane reaction (N and fN) is
Th e horizontal component
W, = Ncos¢, sind + fN cos A,
(i)
ivi
W, isi the driving force on the worm, ot btainable from the horsepower
FIGURE 16.4 force pe pil _ )
So eion Gate applied to the worm. (W, is an axial
Worm-gear Speed Reducer.
an air-cooled model, the air circulating arov :
we find the relatio n betw
outside of the oil sump, which is ribbed to in Fiiminating N from equations (i) and (j),
heat-transfer surface. Observe the large lead g and Fi
the worm. Shaft extensions are standardized te
standard couplings with a press fit. (Courtesy cos ¢, sinA + fcos “|
Worm and Gear Co., Cleveland).
(k) an ee ¢, cosA — fsin a

ideri
sidering i
efficiency i
from the basici notion 9 f output/input, we have
h cae work as W,v,, ft-lb/min., where v, fpm is the pitch-line —
fe worm. The frictional force F, acts through the distance of the pas
power screws, or worm threads. The total plane reaction between
the we of thread, Fig. 16.3; now with v, = 7D,n, as the pitch-line speed, the
and the gear is the resultant of the force N normal to the tooth and which F;, acts .
frictional force F, = fN. Since, up to a point at least, the frictional : bbing speed v, = v,/cos A fpm, the distance through
is independent of the area, we may imagine the force system acti
fi Sci. Hence, the work of F; is F,v, = fNv,/cos A, and the output wor
point O, Fig. 16.5; N, being normal to the surface at that point, lea becomes W,v,, — fNv,/cos A. Using the value of W, from (j), we get
from the z axis in the direction of the y axis (because of the pressure ar (N cos ¢, sin A + fN cos A) v,, — fNv,/cos :
and in the direction of the x axis (because of the angle
plane rectangle abcd is perpendicular to the z axis; angle dOc is
of the helix), fiir (N cos $, sind + fN cos A) ey
¢; a
aOb is the normal pressure angle 4, (this plane is normal to the th Cancelling N and v, and manipulating the remainder, we find
O). Write tan ¢, = ab/bO, let ab = cd, multiply and divide by cO, = Ee $, — f tan _
and cos ¢, cos A — f sin 4
ue ——=t 63) ¢— = ¢, sind + fcosa cos ¢, tana +f
(h) tan gs R:
bOcO cObO rad ii
This equates also gives the theoretical efficiency of an Acme power nati
(§ 8.15) or a fastening thread; moreover, let ¢, = 0 to get a square threa
16.8 EFFICIENCY OF WORM GEARING. As explained abc
the forces N and F, are components of the total plane reaction sho
acting on the worm. The frictional force F; is tangent to the helix at 02
lies in the xz plane. However, if anything, the rectangular components
the x, y, and z directions are more useful; they are W,, which is the tangen
(to pitch cylinder) transmitted force on the worm, F,, the transmitted
fo
on the worm gear, and S, the separating force tending to push the membt
apart. Since these forces are respectively the x, z, and y components of|
resultant of N and F; (notice that F; has no y component), we shall e
F, to the sum of the z components of N and F,, and W, to the sum of f FIGURE 16.5 Forces for Worm Gearing.
x components. The component of N along line Ob is N cos ¢,; the compe Not a free body; all forces shown as acting
ent ofNV cos¢, in the z direction is N cos ¢, Cos A, acting down. On the worm.
The
component of F; is F; sind = fN sin); thus

(i) F, = Nos 4, cos A — fN sin A,

acting down in Fig. 16.5, where F, is the driving force on the worm ge:
obtainable from the horsepower equation (output) applied to the geé

432
434 WORM GEARING [Ch.

and see that the equation reduces to the form of equation (j), § 8.2
Observe that equation (16.3) does not include the bearing losses, which mg
be relatively small. Finally note that the foregoing equationsapplytocro
helicals on shafts at right angles; worm driver, $, = py, J. = ~, = A.
For the best efficiencies, the lead angle A should be greater than
For a given pitch, the smaller the worm diameter, the larger A, but to g
FIGURE 16.7 End View of Double-reduction Worm Drive. The
the highest efficiencies, it is necessary to use multiple-threaded worms wit
second worm is behind the small gear, invisible. Observe the large
up to 24 threads. However, for a particular velocity ratio and pitch, f ] increase in shaft size from input to output shaft. (Courtesy Cleve-
gear must be three times as large for a triple, for example, as for a sing land Worm and Gear Co., Cleveland).
thread; but note that the approximate limiting wear load (16.2) depen Cc

only on D,, the face width, and K,. See typical curves for the variation¢
efficiency in Fig. 16.6.

oe
poe Ree
| i
| \ pi
FIGURE 16.6 Efficiency of Worm Drive [Girdle .
| Observe that the efficiency falls off ra

poorer sya DY
f=0.
I
low lead angles; that, especially for Ie 7
| friction coefficients, the curve is relatively fi
| for 30°< A<60°. The maximum efficien
|
| occurs in the general vicinity of A = 45° ieee, ue
!
! comparison, the efficiency of the ball
|
i screw, Fig. 8.25, is said to increase to (a) Forces on Worm. F; is the end (b) Forces on Worm Gear. W; pro-
|
than 90% within a lead angle of 2°. thrust; the transmitted force W; pro- duces end thrust and bending. The
ww Ud
duces torsion and bending on the shaft, transmitted forced F; produces tor-
Lead Angle \
S produces bending; F;, being eccentric, sion and bending; the separating
also produces bending that should not force S produces bending.
be neglected.
16.9 COEFFICIENT OF FRICTION, WORM GEARING.
FIGURE 16.8 Forces on Worm and Worm Gear. This is the simplest arrangement for
with ¢, and A. As usual, the coefficient of friction varies, sometimes wi del finding bearing reactions and bending moments.
and seemingly capriciously; it is certainly affected by the condition o
surfaces and the materials, by the adequacy of the lubricant and its tem
ture, by the rubbing speed v,, and by alignment and workmanshi The efficiency tends to fall off as the velocity ratio increases, so that with
general. Not all sources of experimental data for f agree. The follo high ratios, it may pay to use a double reduction, Fig. 16.7. When it is
equations represent compromises of data in Ref. (/3.1) and should impossible for the gear to drive the worm, the drive is said to be self-
conservative values for a carburized and ground worm driving a phospho locking. This condition exists when the coefficient of friction is larger than
bronze gear when the design and manufacture is good: approximately tan A (actually also dependent on 4). This may mean a lead
angle of 6° or less; if locking is required, let X < 5°.
0.155 0.32
(16.4) f= or f= 7.0.36"
0.2
v,-* r
[3 < v, < 70 fpm] [70 < v, < 3000 fpm]
16.10 SEPARATING FORCE BETWEEN WORM AND GEAR.
where the rubbing speed v, = 7D,,n,,/(12 cos A) fpm when D,, is in inche
The driving forces W, on the worm and F, on the gear have been discussed
(Fig. 16.3). For other metals, increase these values by 25%, in which cas
in detail above and are shown in Fig. 16.8. The separating force S, Fig. 16.8,
they should be suitable also for crossed helicals well run in.
is the component of N in the y direction, radial to the worm (F; has no

435
436 WORM GEARING [Ch. I6_ s16.11] PROPORTIONS FOR WORM GEARING 437

component in the y direction). Thus S = N sin¢,. Using the values of N_ to be made; for the shell type,"°-?! D,, ~ 2.4P, + 1.1 may be taken as a
from equations (i) and (j) in this expression, we get i guide for up to quintuple (perhaps more) threads; for integral worms, try
Dy = 2.35P, +(0.4, or, if a center distance C is known,
F, sin ¢, W, sin ¢,
@ S= cos¢, cos —fsinA cosd, sind +fcosa C9.875
(m) 19 in.,
2.2
16.11 PROPORTIONS FOR WORM’GEARING. The meanings of
say, + 25%. Equation (m) is intended to give a worm size which is approxi-
certain dimensions are shown in Figs. 16.9 and 16.10; note what is meant by”
the pitch diameter D, of the gear, Fig. 16.10; D,,. = outside worm dia- mately optimum for maximum power for particular materials and centers.
meter, D,, = outside gear diameter. The choice of the pressure angle to It is easier to get high efficiency (higher A) with the integral worm than with
be used must be made with some knowledge of the size of the lead angle, the shell type. The face length of worms, Fig. 16.9, may be of the order
of 10:33
because if the lead angle is very much larger than the pressure angle, there
will be excessive undercutting of the gear tooth on the leaving side. Th 2[2a(D, — 2a)]*’?, where a = addendum or P,(4.5 + N,/50).
following limits are recommended."%-9! 4
Dudley"%:2°1 recommends allowing about 6° of lead angle per start; for
for ¢, = 144°, Anax = 16° for $, = 20°, Amax = 25°; example, a lead angle of 30° would require a 5-thread worm according
for $n, = 25°, Amax = 35°; — for dy = 30°, Amax = 45°; to this rule; but for another reason, 8 or more may be desired.
but this is not to prohibit the use of a 20° pressure angle with a lead angle The face width of the gear is partly a matter of good proportion, but
of 15°, for example. There are no standard tooth proportions used for all since the load is never uniformly distributed across the face, the value of the
worm drives; an interchangeable type of tooth with an addendum of 1/P, = peak load enters the decision. The wider the face, the larger the difference
0.3183P, has been widely used for single- and double-thread worms, between the peak load and the average load; on this account, Bucking-
an addendum of 0.3183P,,, on double, triple, etc. threads. Dudley™* ham"+-1] recommends bmax = 0.5D,.. Dudley™*-?°! suggests a value
suggests using the normal circular pitch for all worm drives that use mn, Fig. 16.10, which is defined by the intercepts of the outside circle with
interchangeable type tooth: a= 0.3183P,,, whole depth = 0.7P,, a tangent to the worm’s pitch circle. Small worm gears may be made from a
However, as the lead angle increases for a given whole depth and pitch in solid round with a hole for the shaft (and key). The proportions in Fig.
16.10 are intermediate, with a solid web. In larger gears, the web may be

l-—D, = Pitch eatin oe

|
FIGURE 16.9 Shell Worm.
FIGURE 16.10 Throated Worm Gear i
—\

the diametral plane, the top land becomes narrower and the problem is not lightened by having holes in it. Also, in the larger sizes, costs may be sig-
to have the teeth become too pointed. Since worm gears are inherently non- nificantly lowered by using a cast-iron or cast-steel spider to which is at-
interchangeable, there is no imperative reason to use an interchangeable tached a bronze rim in which the teeth are cut; or weight and inertia
type. might be reduced by making the spider of aluminum or titanium. An im-
There are two types of construction for the worm, the shell type, Fig. portant attribute of the gear should be a good lateral and rim stiffness; the
16.9, designed to be keyed or splined to its shaft, and the integral type, rim may be equal to about 0.6P,. Buckingham"™*-1) suggests “dishing” the
Fig. 16.7, which is a part of its shaft. A proper worm diameter is a decision web, and he proposed the interesting hub design of Fig. 16.11. The bores
 Shaft
Shoulder
516.13] DESIGN PROCEDURE 439
BoreA
Bore B,
Insert Press Fig. 16.12. Both of the gears in this drive have straight-sided teeth, giving
Pressed In Fit at surface contact.
Boch End FIGURE 16.11 Hub Design for Large-diameter Gear.

Web
16.13 DESIGN PROCEDURE. Since the boundary conditions vary
so widely from job to job, it is impossible to set up a single design pro-
cedure. If a certain limited space is available, this in effect limits the size~
A and B should be quite concentric, with press fits on the shaft that contri- of the gear and center distance. One could then estimate a worm diameter,
bute to the lateral stability of the fit; these bores are needed because the
splines are not adapted to good centering. Ordinary keys are often un-
satisfactory. 4
There is nothing compelling about any of the proportions proposed ~
above; some of them may be helpful as starting points. The basic factor
for the worm size is that the shaft be of the right size to carry the load, either

LD
for strength or for deflection; then for an integral worm, the root dia- FIGURE 16.12 ——
Tool Co., Detroit).

QLD
meter should be somewhat larger than the shaft. For a shell worm, a hub ANC,
diameter of about (1.8) (bore) should be adequate with a root diameter
A SG S77)
somewhat larger, Fig. 16.9.
. Wy S

16.12 GENERAL REMARKS ON DESIGN OF WORM GEARING.

Because of the curvature of the enveloping surfaces and the consequent
§ 16.11, and proceed from there. In any event, iteration is inevitable. If
shape of the contacting surfaces, tooth action is complex. The details
of the profile shape depend on the method of cutting the threads,"*"!! an significant power is being transmitted, the objectives are a limiting wear
load equal to or greater than the dynamic load, adequate gear-tooth
the gear teeth should be hobbed by a hob that matches the worm to b
used. Partly for this reason, a starting point in design is often a list of hobs.
strength, and a good efficiency (A $ 30°). Since our estimate of the dynamic
(sizes) available, followed by an endeavor to use one of these. Nonstandard
load is inherently in error, the tooth design can reasonably be made ignoring
teeth can be cut with standard hobs. For highly refined designs with all- the efficiency in computing F,, which is on the safe side. At this point, at
recess angle of action, see Ref. (14.1). least the approximate size of shaft needed for the worm should be known.
Given a certain power and speed for the worm, one may proceed somewhat
Worm drives have typically been used where a relatively large velocity
as follows.
ratio is desired, but they are being made with low velocity ratios where the
(a) Assume the number of starts (threads) for the worm and, for a given
right-angle relationship is advantageous and more power must be trans-
velocity ratio, compute Ny.
mitted than can be handled by crossed helicals (worm gears have theoretical
(b) Set up expressions for vmg, F:, and Fa in terms of Pc.
line contact versus point contact for helicals). } (c) Since wear is the probable determinant, if not heating, express F, in
terms of P:. Although there is not always an apparent reason, the face width on
worm gears is generally narrower in terms of circular pitch than on spur gears
acts opposite to the direction shown (opposite sense of rotation), which’ (§ 13.12); for the first approximation, try b = 2Pe, or less. In the end, 5 must have
requires another force analysis. a satisfactory relation with D,,.
If the worm is steel and the gear bronze, EP lubricants are not as bene- (d) Choose materials and select K,,. Equate Fa = F,, with Pc as the only un-
ficial as for steel and steel, The recommended lubricant"%-!) is a steam known. Solve for Pe (which is now a point of departure), judge its appropriate-
cylinder oil or an oil with 3% to 10% acidless tallow or other appropriate ness, choose a standard value, check to see that the various proportions are satis-
factory. This necessitates knowing the worm diameter. Note that the number of
animal fat. The customary system of lubrication is for the gear to dip into”
Starts on the worm can be changed with little effect on the wear capacity if Dy
an oil, Fig. 16.7, but if the oil must be circulated for cooling, it may just and D,, are kept about the same (no change of materials), but watch that the teeth
as well be directed at the mesh on the return. do not become too pointed.
We have considered only the case where the gear envelops the worm, (e) Other details of the iteration process are best not standardized. Compute
Fig. 16.10, but the worm may also envelop the gear, as it does in Cone gears, the efficiency (or A) early, because this is a factor, and the temperature rise with

438
440 WORM GEARING [Ch. 16 ©

natural ventilation. A complete set of computations that validate the final choices —
is necessary.

16.14 MATERIALS FOR WORM GEARING. Preferred gear mat- —

erials are the bronzes, especially the tin and nickel-tin bronzes (which may
or may not be chilled cast, a process that produces the hardest bronze
surface); but others, as leaded bronze (for high speed) and aluminum and
silicon bronzes (for heavily-loaded, low-speed gears), are often used. To .
reduce costs, epecially for large gears, use a bronze rim attached to a
cast-iron or cast-steel spider. See also § 16.5.

16.15 CLOSURE. Since the tooth dimensions cannot be said to be

standardized for worm gearing and since the hob and the worm should”
match, the designer is limited only by his knowledge of involute geometry
and of the manufacture of unequal addendum teeth; but further detail is
beyond the scope of this book. Backlash is commonly made larger on worm 17. FLEXIBLE POWER-
gearing than on spur gears, and customarily it is obtained by reducing the
thread thickness on the worm only; if there is a significant temperature TRANSMITTING
rise, investigatethe effect of the differential expansion of the different materials
on backlash. There are of course many applications of worms in which the ELEMENTS
power is not significant, as in instruments and servomechanisms, for which
the design criteria are not at all the same. a
17.1 INTRODUCTION. As is true of other machine elements,
flexible connectors for transmitting power appear in many different forms:
flat belts, V-belts, flat-V, “‘toothed’’ belts, ropes (manila, cotton, wire).
Also to be covered in this chapter are chain drives, which are much less
flexible except in the sense of having turning joints. The flexible drives have
distinct advantages that are on occasion overwhelming: they absorb
vibration and shock, tending to transmit only a minimum to the connected
shaft; they are suitable for relatively large center distances; they are quiet;
when properly maintained, they can be designed to have a long trouble-free
life.

17.2 NET BELT PULL AND THE VARIATION OF STRESS IN

BELTS. If the smaller pulley is driving clockwise, Fig. 17.1, the force
Fy on the approaching belt is greater than that F, on the receding (slack)
side, These forces produce the resisting torque (F, — Fo)r, = T,. The
difference of the forces F, — Fe is called the net belt pull. The driving
torque on the larger (driven) pulley is (F, — F,)r, = T;. The net belt pull
(or torque) is computed from the horsepower equation (1.15). With
F, — F, = F, hp = Fop/33,000 = Fv,/550; also hp = 7n/63,000 where
T in-lb. and n rpm apply to the same pulley.
We speak of belts wearing out, but the actual failure is quite analogous

441
 FIGURE 17.1 Open Belt. The tight side (larger
tension Fi) is preferably on the bottom side of the
pulleys because, as you see, the natural sag on the
slack side on top results in the actual angle of contact
being greater than the theoretical 0.

to fatigue. To trace out the variation of stress in a belt, accept first that with
a particular point on the belt following a curved path, there is a centrifugal FIGURE 17.3 Forces on an Element of Belt on
Driven Pulley.
force that induces a stress s,,, more or less uniform and so depicted in Fig,
17.2. Think of an element of belt as it makes its rounds, as at J, C, D, EL
H. At position J, Figs. 17.1 and 17.2, just as a section of belt leaves the Summing forces normal to the element, letting sin d@/2 = d0/2, and drop-
small pulley, the curvature of the belt is relatively negligible and the stress ping the product of two infinitesimals, we have
Sp is that due to the slack tension F,. Between B and C, the element moves
d0 d@
onto the larger pulley, is bent to a radius r,, which thereby induces a (a) 2g ES oa Re Ee oe dE) stare aS din FOR aD,
bending tensile stress at points outside of the neutral plane (see at J, Fig.
17.1); total stress is sg + Spo. (If the action were elastic, the stress due to
curving to radius r would be s, = Ec/r(§ 17.25), but ordinarily the strai from which dN = F dé — dS. Summing forces tangential to the element,
are not elastic.) As the element moves about the larger pulley to D, letting cos d6/2 = 1, and substituting the value of dN just found, we get
force on it due to the transmitted power increases more or less graduall d0 d0
the value F,; total stress is s; + 52. Between D and E, the bending s =F, nay Aer - (Fae
drops to a negligible amount, but as the element moves onto the smaller
pulley G to H, it is bent to its sharpest curvature making this the point of
maximum stress s; + 5,;. As the element moves from H to J, the fore
(b) = fF do) — f dS — dF =0.
due to power decreases to F,, and the cycle repeats with every complet The centrifugal force (F = ma) on a particle (the element) is its mass times
revolution of the be/t. A belt can stand so many of these applications of th the normal acceleration a, = v,?/(r/12), where the belt speed v, is in fps
peak stress, depending on the magnitude of the peak value and probably and r in. divided by 12 is r/12 ft. to match g, = 32.2 fps? to be introduced
the magnitude of the mean value. We may conclude that if the bendin next. The volume of the element for a belt thickness ¢ in. and belt width d in.
on a pulley (or sheave) is sharp enough, the belt may eventually fail fron is bt dL = btr d0 cu. in.; let the density p be lb/cu. in., so that the mass is
the flexing with no power transmitted. if p btr dé \b. or p btr dO/g, slugs (since the pound is used for force in mech-
anics by engineers, the mass must be in slugs to be in a consistent system
of units with g, = 32.2 fps”). Thus,
17.3> CAPACITY OF A FLAT BELT. Make a free body of an elemeni bir db 12pbtv,?
of belt, Fig. 17.3, whose length is dL = r d0. The analysis is for the case of (c) ———_dd = K do,
impending slipping (in order to eliminate N), so that the frictional forceii ~ g,(r/12) Bo
Ff dN, where dN is the normal force of the pulley on the element (or one Where K = 12 pbtv,2/g,, a constant, is used for convenience. Substitute the
think of f as being simply a ratio of the actual frictional force to the nor value of dS from (c) into (b), separate the variables, and integrate from the
force, and not a true coefficient). The belt pull on the slack side is F, on smaller to the larger, F, to Fy;
tight side of the element, F + dF. A centrifugal (reversed effective) force
dS acts radially outward. The sum of the forces is any direction is zero
es aa nae
1
= dé or In F, 7 K
= 0.
c

In exponential form,
Bends on FIGURE 17.2 Stress Variation in Belt.
—— | SmallSheave | Leaves A simplified model. After W. (@)
Bends on TH H| Sheave Worley.(17-31 fe = e/9,
——
Sheave Dhhis Fy ay K

Observe that for v, = 0 (K = 0), this equation reduces to F,/F, = e, a

J Tension
8 Slack Tension &——

i {4 Tension Due to Centrifugal Force

444 FLEXIBLE POWER-TRANSMITTING ELEMENTS [Ch. I7 g17.5] COEFFICIENT OF FRICTION 445

familiar form from analytic mechanics. Since the speed and mass are often needed after decisions have been made on the various other factors involved.
low enough that the effect of the centrifugal force is negligible, K may be Let the standard widths be as follows:
dropped from any ofthese equations when this is so. Invert (d), subtract The width varies: by } in. increments from $ to 1 in.,
one from each side, change signs, and solve for the net belt pull F, — Fo; by } in. increments from 1 to 4 in.,
by 4 in. increments from 4 to 7 in.,
eo — |
(17.1 A-m-G- "(= e
by 1 in. increments from 7 to 12 in.
Larger sizes are obtained by special order. Single leather belts should be
where Kis the load on the belt due to the centrifugal force, and is sometime; specified in widths not over 8 in., because this represents about the maxi-
conveniently expressed in terms of a mass per foot of length, w lb/ft. mum width of suitable leather than can be cut from a hide.
w = 12 dtp |b., where 12 in. is 1 ft., so that

12pbtv,2 wv? 17.5 COEFFICIENT OF FRICTION. Not only does the coefficient
(e) K =
&o 32:94 of friction vary widely with operating conditions and not only is it difficult
Since F, is the maximum tension, it has been the custom in flat belt desig to know what it is under various operating conditions, but if any of the
friction belts of this chapter were run continuously with the operating net
to let F, = sA = sbt, where it is recognized from the above discussi
belt tension at the point of limiting friction, their life would usually be
that s is not the real maximum stress; design stresses that have been foun
uneconomically short. This situation can be allowed for arbitrarily by
to give good designs are used. Making this substitution for F, in (17.1) anc
using a low value of f [or by using a value of the net tension ratio, equation
using K from (e), we get
(d) above, that is known to be satisfactory]. Another approach is to use a
low design stress s in equation (17.2)—S§ 17.6; and of course, one can use
(17.2) Focal ot ( a (= a *
re We Ne Oe et Bey safe values of both f and s.
The coefficient of friction varies with the amount of slip. A part of the
where e = 2.718: --. For design, use total slip is creep, which exists because the driving pulley receives a longer
p = 0.035 Ib/cu. in. for leather, (stretched) belt than it delivers and the driven pulley receives a shorter belt
p = 0.045 lb/cu. in. for flat rubber belting. than it delivers (Fig. 17.1), giving rise to relative motion between the belt
and the pulley. As the load increases, there is, in addition, slippage in the
The theory of this derivation is impeccable, but as usual, there are fi customary sense of the word. When used without qualification, s/ip or
of life that cannot be incorporated into a simple theory. (If this were not: ¢ total slip means the total amount including creep. If the slip becomes ex-
there would be little use for engineers.) We shall now discuss those facto: cessive, leather belting “‘squeals,” giving a warning. As the total slip
some in equation (17.2), some not, that affect belt design. increases, the coefficient increases, Fig. 17.4; 2°% slip is satisfactory for
leather on steel or iron. For normal design conditions for flat belts and the
design stress of the next article, use the following:
17.4 BELT THICKNESS AND WIDTH. Since the repeated flexin:
Leather on iron or steel,f = 0.3;
of the belt about the pulleys is a significant determinant of the life of th
Leather on paper pulleys,f = 0.5.
belt and since the thicker the belt, the greater the maximum stress induce
by the flexing, there should be some consideration of belt thickness as
FIGURE 17.4 Typical Friction-slip Curves 8
lated to the smaller pulley diameter. The thickness of a leather belt depend for Three Kinds of Pulleys.{17-5] 7
on the number of plies and on the thicknesses of the hides used in its manu
facture. Average values are given in Table 17.1, which also shows the recom 6
mended minimum pulley diameter to be used with each belt thickness. T g
larger pulleys are permissible and used, longer life with the same 5 and ¢ fo ES

the belt may be expected. With the qualification on pulley sizes, th S4

narrow belts are more economical than thin wide ones. When equatio! §
(17.2) is used for design, the normal action is to solve for the belt width| 3
32
1

0
444 FLEXIBLE POWER-TRANSMITTING ELEMENTS [Ch. I7 s17.5] COEFFICIENT OF FRICTION 445

familiar form from analytic mechanics. Since the speed and mass are ofte needed after decisions have been made on the various other factors involved.
low enough that the effect of the centrifugal force is negligible, K may be Let the standard widths be as follows:
dropped from any of ‘these equations when this is so. Invert (d), subtract The width varies: by }in. increments from 4 to | in.,
one from each side, change signs, and solve for the net belt pull F, — Fp; by } in. increments from 1 to 4 in.,
ef6 — by 4in. increments from 4 to 7 in.,
(17.1) F,- Fy =F - W 76 ) by 1 in. increments from 7 to 12 in.
e
Larger sizes are obtained by special order. Single leather belts should be
where K is the load on the belt due to the centrifugal force, and is sometim es specified in widths not over 8 in., because this represents about the maxi-
conveniently expressed in terms of a mass per foot of length, w lb/ft. mum width of suitable leather than can be cut from a hide.
w = 12 dtp |b., where 12 in. is 1 ft., so that
12pbtv,2 —wo,?
(e) K= 17.5 COEFFICIENT OF FRICTION. Not only does the coefficient
Zo 32:9: of friction vary widely with operating conditions and not only is it difficult
to know what it is under various operating conditions, but if any of the
Since F, is the maximum tension, it has been the custom in flat belt design
friction belts of this chapter were run continuously with the operating net
to let F, = sA = sbt, where it is recognized from the above discussion
belt tension at the point of limiting friction, their life would usually be
that s is not the real maximum stress; design stresses that have been fou n
to give good designs are used. Making this substitution for F, in (17.1) ane uneconomically short. This situation can be allowed for arbitrarily by
using K from (e), we get using a low value of f [or by using a value of the net tension ratio, equation
(d) above, that is known to be satisfactory]. Another approach is to use a
pmilegd oe ( seey(= _ *) low design stress s in equation (17.2)—§ 17.6; and of course, one can use
(17.2)
ee a Oe Bey safe values of both f and s.
The coefficient of friction varies with the amount of slip. A part of the
where e = 2.718: --. For design, use total slip is creep, which exists because the driving pulley receives a longer
p = 0.035 Ib/cu. in. for leather, (stretched) belt than it delivers and the driven pulley receives a shorter belt
p = 0.045 Ib/cu. in. for flat rubber belting. than it delivers (Fig. 17.1), giving rise to relative motion between the belt
and the pulley. As the load increases, there is, in addition, slippage in the
The theory of this derivation is impeccable, but as usual, there are fact: customary sense of the word. When used without qualification, slip or
of life that cannot be incorporated into a simple theory. (If this were no total slip means the total amount including creep. If the slip becomes ex-
there would be little use for engineers.) We shall now discuss those factors cessive, leather belting “squeals,” giving a warning. As the total slip
some in equation (17.2), some not, that affect belt design. increases, the coefficient increases, Fig. 17.4; 2% slip is satisfactory for
leather on steel or iron. For normal design conditions for flat belts and the
design stress of the next article, use the following:
17.4 BELT THICKNESS AND WIDTH. Since the repeated flexi
Leather on iron or steel, f = 0.3;
of the belt about the pulleys is a significant determinant of the life o
Leather on paper pulleys, f = 0.5.
belt and since the thicker the belt, the greater the maximum stress induc
by the flexing, there should be some consideration of belt thickness as
FIGURE 17.4 Typical Friction-slip Curves
lated to the smaller pulley diameter. The thickness of a leather belt depends

&
for Three Kinds of Pulleys.27-51
on the number of plies and on the thicknesses of the hides used in its manu

2A
facture. Average values are given in Table 17.1, which also shows the recom:

D®
mended minimum pulley diameter to be used with each belt thickness. Il

Coefficientof Friction
larger pulleys are permissible and used, longer life with the same b and ¢ fot

A
the belt may be expected. With the qualification on pulley sizes, thic

&®
narrow belts are more economical than thin wide ones. When equation
(17.2) is used for design, the normal action is to solve for the belt width |

&®
»
-
446 FLEXIBLE POWER-TRANSMITTING ELEMENTS [Ch. }j 517.8] ANGLE OF CONTACT 447

17.6 STRENGTH OF LEATHER. Since the hides of no two steer; 17.8 ANGLE OF CONTACT. For an open belt, the angles of contact
are likely to be of the same quality and since the strength of leather als are
depends upon the method of tanning, the variability is expected to be large,
The breaking strength of oak-tanned belting varies from 3 to more th R some Sf D, — D,
(17.4) 6=7 + 2sin 7 radians,

I+
6 ksi. Mineral tanned leather is stronger, say 7 to 12 ksi. To make the bel
endless, the ends must be joined: by cementing (7 = 1.00); by wire lacin,
with machine (y = 0.88); by metal hooks (n = 0.35); and by other methods from Fig. 17.5, where it is seen that « = sin-4(R — r)/C. Use the plus sign
for the Jarger pulley and the minus sign for the smaller pulley. For crossed
where 7 is the relative strength of the joint. For normal good operatin
conditions and pulley sizes larger than the minimums
belts, the angles of contact are the same on both pulleys, Fig. 17.6,
of Table 17.19
design stress of R+r
(h) 6@=2+2sin+ radians,
( Sq = 4007 [LEATHER
where R = the radius (D, = diameter) of the larger pulley, r = the radius
in equation (17.2) should result in a belt capable of a satisfactory lif
(D, = diameter) of the smaller pulley, C = the distance between pulley
Since the strength per se is not the criterion for belt capacity, the relatiy
centers, and @ = the angle of contact in radians; all dimensions in inches.
strength 7 probably has little to do with the belt’s life unless it is rather lo
(belt damaged by the connection). Reduce the design stress for unusual
long life or for unfavorable conditions (or introduce service factors
Table 17.7).

17.7 LENGTH OF FIGURE 17.5 Open Belt.

BELTS. _ Exact lengths are as detailed in Fig
17.5 and 17.6; closely, the belt lengths may be computed from

(17.3) L & 2C + 1.57 (D,

(D, — D,)?
+ D,) + TT Eae ae

[OPEN BELT, FIG. 17.5]

D, + D;)?
(g) L.2C 4 AST6Do.+ Di). a
[CROSSED BELT, FIG. 17.6]

where C is the center distance, Dz is the diameter of the larger pulley, Dy! FIGURE 17.6 Crossed Belt.
the diameter of the smaller pulley, and all dimensions are in inches.
the center distance C for a given endless belt is desired, solve for it fror
(17.3); or see equation (0), § 17.17. A crossed belt has the advantage of
large contact angle and it does very well at the lower speeds. The wear dt
to the rubbing at the point of crossing is a disadvantage. ;
Although it does not appear in equation (17.1), the length of belt is a
important engineering factor. The longer the belt for a given belt sp
the less frequently it is flexed about a pulley or sheave. Since the flexin _ Considering the term (e/? — 1)/e? in equation (17.1), we see that sig-
often the significant damaging action, doubling the center distance, nificant changes in the minimum angle of contact, as would occur for a
instance, would be expected to increase the belt’s life materially. Especi large velocity ratio and relatively short center distance or for the case of a
for flat belts on horizontal drives, a long belt (force of gravity) contrib tension pulley, Fig. 17.7, will bring about significant changes in capacity.
to maintaining the initial tension (§ 17.10). However, not only is a lo Owever, there is evidence that the actual effect is not so great as predicted
belt more expensive, but the space it occupies is costly. Other more econo by the theory, so that one could use some average applicable value for
mical means of maintaining belt tension are available (§ 17.13). (e/? — 1)/e/ for the more usual configurations. In general, the design by
Bi
517.11] RATED CAPACITY OF LEATHER BELTS 449

17.10 INITIAL TENSION. In order to transmit power, the belt

must have some initial tension (tension at rest); Taylor’s""-*! recommenda-
tion is 71 lb/in. of width. The tighter the tension, the more power can be
transmitted without excessive slipping; the initial tension may be so low
that the belt will not transmit its power. As the initial tension is increased,
FIGURE 17.7 Tension Pulley for Use the value of the tight tension F,, also increases. It not only becomes damag-
on Short-center Drives. (Courtesy Linl
ing to the belt, but since F, + F, (a bending force on the shaft) also in-
Belt Co., Chicago).
creases, the operating conditions may become dangerous for the shaft and
bearings. A relation F,/F, is often assumed for design purposes (see Fig.
17.10), say F,/F2 = 3 with 6 = 180°. The effect of initial tension is not
recognized by equation (17.1).

pulley). 17.11 RATED CAPACITY OF LEATHER BELTS. The foregoing

discussion of the details is included because an engineer is eager to consider
all factors that might effect his designs and because most of the ideas apply
17.9 BELT SPEED. The tension in the belt due to centrifugal fore to all kinds of belt drives (some are even easily interpreted for chain drives).
increases rapidly above about 2500 fpm. If every thing holds togethe: But there is no denying the many inherently involved intangible variants
equation (17.3) shows that a speed is reached where the belt is stressed t that only experience of a certain kind might successfully evaluate. Thus, for
whatever design stress s has been chosen, at which point no power can b the common case, a simple design procedure using rating tables and
transmitted without exceeding this stress. Thus, given a certain belt siz experience factors will give good answers to many problems; say, those of
and permissible stress, we find that there is a certain speed at which m Tables 17.1 and 17.2 for flat leather belts; the values in Table 17.1 are
mum power would be transmitted (see Problems for problems illustra based on the general belt equation (17.2). Interpolations may be made
this idea). Experience suggests that the most economical designs are ol between the speeds given, with the experience factors of Table 17.2 applied
tained for a belt speed of 4000 to 4500 fpm,"’-*! but of course, af as follows:
particular application may require some other speed. For leather, speeds 6
7000 to 8000 fpm are in use (but consult the vendor). Flat fabric belt
(i) hp = (hp/in., Table 17.1)(bCnCp)(Cy1Cya « « »)s
operate up to 20,000 fpm and more. On high-speed centrifugal blowers,| which is the nominal horsepower that the belt is to transmit (or the name-
has been observed that the arc of contact @ is reduced from 180° at rest & plate horsepower for an electric motor), and
110° in motion, one of the side effects of speed. b in. is the belt width,
We might well keep in mind that doubling the belt speed, for exam C,, is the correction factor for the type of drive; except for electric
doubles the number of flexes of the belt on the pulley in a given time, whie motors, use C,, =
suggests that the higher the speed the more important the minimum pulle C, isthe correction factor for the size of the smallest pulley; the
size. smaller the pulleys, the greater is the flexure,
To find the stress due to centrifugal force alone, let F, — F, = 0; tha C; is the correction factor for the environmental conditions, and
is, assume no power is transmitted. It follows then that the bracke more than one of these factors may apply; thus a vertical drive
s — 12pv,?/32.2, must be equal to zero. Equating this expression to zera in a dusty atmosphere and subjected to shock loads would cor-
assuming s = 400 psi, p = 0.035 lb/cu. in., and solving for the velocity, W respond to a total factor of C; = (0.83)(0.74)(0.71).
get v, = 10,500 fpm (nearly 2 mi/min.), which is the speed at which tf Other deviations from desirable practice that should be considered
centrifugal force alone theoretically stresses the belt to 400 psi. This com include abnormally short center distance and tight side on top. The life
putation does not mean that a belt cannot, or even should not, transml €xpectancy of belts selected in accordance with these ratings should be some
power at this, ora higher, speed, as some have thought. In the first place 5 to 7 years. Observe, at the bottom of Table 17.1, that smaller pulleys are
experiments"™7-7] show that the effect of centrifugal force is not as grea Permissible at the lower speeds, because the belt does not bend as often.
as it is indicated by equation (17.1); but even so, it would still be possibl The Strength of the belt joint is not considered in this design procedure.
to transmit power if values of s greater than 400 were used. For leather belting, mount the hair or grain side next to the pulley, because

448
 —_——ee
eC re wrasse Vvveasne & acne

INCH OF WIDTH, LEATHER BELTS®”.*1

For belt speeds over 6000 fpm, consult a leather belting manu- 517.13] MAINTAINING INITIAL TENSION 451
facturer; ¢ is the average thickness of leather belting.
the flesh side is tougher and can stand better the stretching it gets on the
SINGLE PLy Dou. e PLy TRIPLE PLy TABLE 17.2 outside.
; EXPERIENCE
BELT |t=%| 8” | 22” | 82” | #” a2 | 42. FACTORS
SPEED 17.12 EXAMPLE—FLAT LEATHER BELT. A squirrel-cage, line-starting
fpm | Med.| Heavy} Light | Medium) Heavy| Medium| Heavy Type OF Drive] Cp motor is rated at 30 hp at 1750 rpm and delivers its power to a horizontal, flat
- leather belt located in a dusty atmosphere. The motor pulley is 10 in. in dia-
600} 1.1] 1.2 1.5 1.8 2.2 25 2.8 Any except 14 meter. What width of medium double-ply belt is needed?
800} 1.4] 1.7 2.0 2.4 2.9 3.3 3.6 electric motor Solution. The belt speed is

1000) 18] 201-26) sa-fose head | 45 Squirrel cage, tn = «Dn = 2 (13)c1750) = 4580 fpm.
1200} 2.1 | 2.5 3.1 3.7 4.3 4.9 5.4 compensator 12.
1400) 2.5] 2.9 | 3.5) 43 | 49] 5.7 | 63 ee 0. Corresponding to this speed and to a medium double-ply belt, we get 11.97, say
1600) 2.8 | 3.3 4.0 4.9 5.6 6.5 7.1 Gaatord 12 hp/in., from Table 17.1. From Table 17.2, we read Cm = 0.5, Cp = 0.7, and
er uanire 0.5 C; = 0.74. Substituting these various values into (i), we get
1800] 3.2| 3.7 | 45] 54] 62] 7.3 | 8.0 Sere "I hp = (hpjin.) bCaC,C
2000] 3.5] 4.1 | 49] 60] 69] 81 | 89 tid’ ving: anda DT, Sener eee Ss
2200] 3.9] 45 | 54] 66] 76] 88 | 9.7 [| high salt ie i 30 = (12)()(0.5)(0.7)(0.74),
TA ee he ae, om eel te conane on from which 6 = 9.65; use a 10-in. belt.
2600] 4.5] 53 | 63] 7.7 | 89] 10.3 | 11.4 Pulley Size, in. | C,
Se
3000] 5.2] 5.9 | ee
7.2] 87] ee11.6 | 12.8
100] 4 or less 0.5 - 17.13 MAINTAINING
A civ s Phie INITIAL
althenekTENSION.; _ All
f beltbelts (and be chains)
the
3200) 5.4| 63 | 7.6] 9.2 | 10.6| 12.3 | 13.5 elongate with service. Thus, a though some types of belts may be short-
44 to 8 0.6 ened, it is convenient to have the motor mounted on an adjustable base. If
3400} 5.7] 66 | 7.9] 9.7 | 11.2] 12.9 | 142 a it is bolted down in slots on the base, it can be moved away from the
3600} 5.9] 69 | 83] 10.1 | 11.7] 13.4 | 14.8 9 to 12 0.7 driven member from time to time to reestablish a good initial tension. Sim-
3800) 6.2) 7.1 8.7} 10.5 | 12.2) 14.0 | 15.4 ; ilarly, the motor may be mounted on runners with an attached spring de-
Pee ee ee eee eae Ow aot: i o- signed to maintain the tension, but since a large stretch of the belt may result
4200/ 6.7| 7.7 | 9.3| 113 | 13.0] 15.0 | 16.5 17 to 30 0.9 in a spring force too low, an adjustment to take care of this effect is also
4400] 69] 7.9 | 9.6] 11.7 | 13.4] 15.4 | 16.9 necessary. f
4600) 7.1] 8.1 9.8] 12.0 | 13.8] 15.8 | 17.4 Over 30 1.0 An early idea was the tension pulley, Fig. 17.7, which not only will
4800) 7.2] 8.3 | 10.1] 12.3 | 14.1] 16.2 | 17.8 Onasdna i maintain initial tension but also greatly increases the arc of contact on the
Conditions | C, nearest pulley, preferably the smaller one. In Fig. 17.7, an arm with adjust-
pie pi . oh oe ae ee aa able weights (not visible) provides the tension; a spring may be used instead
5400| 7.6| 8.7 | 10.6| 129 | 1481 171 | 188 a or of weights. The tension pulley should not be smaller than that permitted by
5600} 7.7] 8.8 | 10.8] 13.1 15.0} 17.3 | 19.0 atmosphere 0.74

5800 7.7] 8.9 | 10.9] 13.2 | 151] 17.5 | 192 Vertical drives | 0.83
6000} 7.8] 8.9 10.9 13.2 15.2 17.6 | 19.3 4
Jerky loads | 0.83
Minimum Pulley Diameters, Inches Shocieand il
reversing
up to ina O71 FIGURE 17.7 (Repeated).
2500 | 24 3 4 5* g* 16** | 20** a
2500-4
4000 | 3 34 4h 6* 9* 1S9* 41522
4000-4
6000 | 34 4 5 TT 10* 208* || 24e*

* For belts over 8 in. wide, add 2 in. to the minimum diameters
shown.
** For belts over 8 in. wide, add 4 in. to the minimum diameters
 Direction Of
Reaction Torque

FIGURE 17.9 Pivoted-motor Drive.

(Courtesy The Rockwood Mfg. Co.,
{ndianapolis, Ind.)

Cradle Pivots
Near Here

wise torque, the motor frame exerts an equal and opposite counterclockwise torque, y
the supporting frame pivoting about the point indicated. Also used on V-belt di
(Courtesy American Pulley Co., Philadelphia).

the belt thickness, because the belt does bend about it, and moreover int
usual location, it results in a reversed bending. 4 FIGURE 17.10 Ratio of Tensions. With
paper driving pulley.
When a motor is delivering power, the torque on the stator (fra
equal and opposite to that on the rotor—action and reaction are eq
Hence, the frame tends to turn in the opposite direction to the motor sI 180 160 140 120
and this tendency is utilized to provide the belt tension by mounti Contact Angle 0
motor on a cradle which is free to move through a small arc, Fig. 1
higher the delivered torque, the greater is the effort of the motor fra
turn and the tighter it draws the belt, because the driving pulley
away slightly from the driven pulley. Thus, this system can be desig
provide automatically an adequate, but not constant, amount of
for any load. As a result, if the loading varies, the average net and FIGURE 17.11 Location of Pivot Axis.
tensions will be lower than for a constant pull device.
If the motor mounting is pivoted, Fig. 17.9, the center of gravity
motor can be so located that its weight provides belt tension, In thi:
of drive, the design may be based on an appropriate value of the rai
the tensions F,/F, say from Fig. 17.10, which are recommended
Tatnall™7-19) for pivoted drives with a paper driving pulley, Fig. 17.
Fig. 17.11, is known from the dimensions of the motor and pivoting base, the
line mn may be drawn. The lines of action of Fi and Fe are taken tangent to the
Pulleys. The line of action of the resultant weight of the motor assembly R is
17.14 ANALYSIS OF PIVOTED-MOTOR DRIVE. usually taken through the center line of the shaft of the motor, although it actually
cided upon a suitable value of F,/F2, one may solve directly for the corr Passes through the center of gravity of the assembly.
ponding location of the pivot axis."7-1°] But due to space limitations, ; (f) Now assume a location for the pivot point B on the line mn, and measure
© moment arms, a, b, and e. Then, a sum of the moments of Fi, F2, and R
about the pivot B should be equal to zero. Thus
outline:
(a) Decide upon a suitable value of Fi/Fe, Fig. 17.10. ® Re — Fia — Fxb = 0.
(b) Compute the net tension Fi — F2 from the horsepower equation, _ If this equation is not satisfied, assume another location of B and try (j) again.
(c) From the simultaneous solution of these two equations, find the valu ‘ a the pivoted drive, the tight tension F, should be between the pivot B
of Fi and Fa. f Nd the slack tension Fp, but such a location is not essential in all cases.
" general, keep the moment arm a as small as possible. Base the design on
space and velocity-ratio considerations affecting the drive. 4
(e) Lay out the pulleys in size and location to scale. Since the dimension © maximum output of the motor, which may be taken as (1.4)(rated hp

452 453
454 FLEXIBLE POWER-TRANSMITTING ELEMENTS [Ch

of motor) if the electric motor is compensator-started and (2.5)(rated hp

it is started across the line.

FIGURE 17.12 Y-Belt Construction. The minimum components in V-belt construction

17.15 RUBBER BELTING. In addition to leather, a number of of are: a rubberized, woven jacket material A; tensile members B that carry the principal load,
materials are used for belting, principally rubber, natural or synthe which is sometimes a fabric, sometimes cords; a resilient cushion or compression material
Rubber belts are made in layers (say 3 to 12), called plies, of canvas di C that provides the side-wall pressure; the tensile layer D, capable of withstanding the
repeated flexing. (Courtesy United States Rubber Co., N.Y.C.).
impregnated with rubber which is later vulcanized. A 6-ply rubber
has six layers of duck of about 32-ounce weight. Also rubber belts are mi
of cord construction which is more stretch resistant. Since excessive h (a) Intersecting Shafts. If two intersecting shafts are to be belt
causes a slow deterioration of rubber and since oil is particularly harm connected, two guide pulleys, one on the approaching side and one on the
a natural rubber belt should not be used in very hot places or wher receding side of the belt, must be used. Each guide pulley must be so located
spattered, unless rapid deterioration is acceptable. Neoprene belts, that it receives the belt from a direction perpendicular to its axis and de-
neoprene covered, less susceptible to damage by oil, are available. Since livers the belt in a direction perpendicular to the axis of the next-approached
belts are preferably endless, stock sizes should of course be chose pulley in accordance with the law of belting. See the guide pulley at the top
pedagogical purposes, let 6 vary by }in. to 2in. width, by }in. to of Fig. 17.15.
by lin. to 10in., by 2in. to 24in. Minimum recommended pully | (b) Shafts at Right Angles. Two shafts at right angles to each other
meters!"7-14] are: 3 ply, 3 in.; 4 ply, 5 in.; 5 ply, 7in.; 6 ply, 10 in.; may be connected by the quarter-turn arrangement. The planes of the
14in.; 8 ply, 18in.; 9 ply, 22 in. According to the Goodrich Co pulleys are of course perpendicular to each other. The relation of the
the permissible net belt pull is 13.75 lb/ply per inch of width, whic pulleys should be such that a center-line plane through each pulley is
a simple means for quick estimation; i tangent to the face of the other pulley. This drive satisfies the law of belting
for only one direction of rotation. If rotation is to be reversed, a guide pul-
ley may be located to direct the belt properly.
(k) F,—F,=13.75bN,, or hp=
From either of these equations, compute the width of belt 5 in.
17.17 V-BELTS. Since they are always used for various auxiliary
plies. Correction should be made for other than smooth service a
drives under the hood of an automobile, the appearance of V-belts is fami-
arcs of contact other than 180°. Choose the arc of contact factor
liar to nearly every one. Two typical internal constructions, which vary in
follows :{7-141
detail among different manufacturers, are shown in Fig. 17.12. Inasmuch
as rubber has a low strength and stretches easily, invariably there are ten-
6 220 200 180 170 160 150 140 130 3m sion members B, Fig. 17.12, usually either fabric or cord construction.
Ko 113 1.07 1.00 0.96 0.92 0.86 0.82 0.78 0. The cushion material C is natural or synthetic rubber. The grooved wheels
that V-belts run on are called sheaves (Fig. 17.38).
Examples of service factors, not necessarily the same as the rubber|
The factors that effect the capacity of a V-belt are much the same as
manufacturer’s recommendations, are found in Tables 17.2 and 1 those previously described, except that the wedging action, Fig. 17.13,
Rubber belts with cord construction have a higher capacity than those Tesults in a much larger normal force N for the same belt tension and there-
fabric ply construction. Rubber belts should be mounted with an fore a larger tangential (transmitted) force is obtainable for the same co-
tension of 15 to 20 lb. per inch-ply, which is approximated if the
efficient of friction. By the same token the reasonably attainable tension
finch per foot of belt length shorter than the steel tape measurem Fatio F,/F, is greater than for flat belts [the ratings and the arc correction
around the pulleys. Ko below are for a net tension ratio of 5, equation (d), § 17.3]. Also
because of the wedging, V-belts do well on short center distances without

17.16 FLAT BELT DRIVES FOR NONPARALLEL SHAF

Belt-connected shafts do not have to be parallel, but the Jaw of bel
must be satisfied, to wit: the approaching side must approach the pul
in a direction perpendicular to the pulley’s axis.
 Refer to
Manufacturer
456 FLEXIBLE POWER-TRANSMITTING ELEMENTS

Rpm of Small Sheave

frequent adjustment for initial tension. In case one belt in a multiple.
drive breaks, the remaining belts will carry the load without the nece
of an immediate shut down (no lost production time). In this case, i
complete newset of belts with lengths especially matched (close toler:
One belt much shorter than the others will take an excessive share of
load and soon fail. ay V
The theoretical equation (17.1) for the assumptions made in the 7 ot
belt derivation apply to V-belts if, instead of f, you insert f/sin ¢ 1 234
6 810 20 30 40 60 80100 200 300 500
groove angle). However, the practice is to use a standardized rated ho: Design Horsepower =(Transmitted Horsepower) (Service Factor)
power, equation (17.5) below. The second term in the brackets proy
an allowance for bending, the effect of which is greater for the th FIGURE 17.14 Belt Sections from Horsepower and Speed. The design horsepower is
V-belt."{7-17] The third term in the brackets is a correction for cen equal to the transmitted horsepower times the service factor. The solid diagonal lines are
action. For one belt, from ASA;"7-11] the dotted diagonal lines closely agree with more than one manufacturer’s
catalog.{17-13]
102 0.09 c v4 te
(17.5) Rated hp = [-(—) - —e | —, The horsepower from the rating equations must be corrected for length
Um K,D, 10°} 10
of belt (K,) and are of contact (K,); that is
where a, c, and e are constants for a particular belt section, D, is
diameter of the smaller sheave, K, is a small-diameter factor, obtai (n) Adjusted rated hp = KoK, (rated hp),
Table 17.4 for the given velocity ratio, and v,, fpm is the belt speed.
where K, is a correction factor for an arc of contact other than 180°.
The most commonly used V-belt sizes in general industrial
Since the arc of contact is a function of (Dz — D,)/C, as equation (17.4)
tions are designated by letters A, B, C, D, E, each with standard, nom
shows, this value is convenient to use in choosing K, from.Table 17.5, the
cross-sectional dimensions 6 and f¢, Fig. 17.13, as given in Fig.
easy way to make the correction; D, = larger diameter, D, = smaller
Premium belts with a greater capacity are available. A narrower i
diameter, C = center distance. This factor Kg is basically a correction of
designated 3V (b = 3), 5V (b = 8), and 8V (b = 1) is gaining favor,
the net tension ratio equation (d), § 17.3, because the permissible value of
there are other special-purpose belts. Equation (17.5) for one belt
this ratio decreases as 6 decreases.":15! If the sheave sizes are not known,
B section is they must be assumed, as well as an approximate center distance. If pos-
10\%°9 13,962 21 on |
() Rated hp = [4.737 (—) ie — 0.0234 2. 7 sible, stay above the minimum sheave diameters D, given at the top of the
a D, 108| 108” columns of Table 17.3.
Short belts are flexed more often about the sheaves than long belts.
and values of a, c, and e for A, C, and D sections are given in Table
The factor K, in equation (m) corrects for the effect of the length and is
below the columns of standard lengths. obtained from Table 17.6. The pitch length of a belt is computed from
Manufacturer’s catalogs have voluminous tables with the rated equation (17.3), § 17.7; choose a standard length from Table 17.3. Or a
powers already calculated, but we cannot spare the space. In any
length of belt may be known for which the center distance is desired; in
first step is to compute the design horsepower by mutliplying the this case, solve for C from equation (17.3). With B = 4L — 6.28(D2 + Dy),
power to be transmitted (or the nameplate horsepower on the mo we get (
a proper service factor N,,. Service factors are a consequence of engi
experience and judgments; those in Table 17.7 may be used here. ©) es 4/ BP Soe = DP
(m) Design hp = N,, (transmitted hp). 16
With the design horsepower and the rpm of the smaller sheave, ent If the size of the smaller sheave is assumed and the velocity ratio m, is
17.14 with these values and decide upon the cross section to be used Nown, choose a minimum center distance as"7-191
suggested by the dotted lines, which indicate a C-section. If the pé
is close to a dividing line, either section is reasonable; design for both: (p) bPur Pa + D, or C= D;, [MINIMUMs]
decide upon which to use on the basis of cost or other consideration. © 2
* At least one manufacturer claims that the manufacturing tolerance on the length
ol Whichever is larger. The center distance is made adjustable-so that the belts
belts is so close that special selection of matched lengths is not necessary.
&

457
 LADLE 1/749 DLAINWVARY Y-DELL LEINGLIITID; LENGTH CORRECTION
HORSEPOWER CONSTANTS!!”-131 FACTORS K,7-111
TABLE 17.4
belt.
See equation (I) for the constants a, c, e, for a B SMALL-DIAMETER
Minimum D, is the smallest sheave pitch diameter that should be used with that section.
STD.
FACTORS Ka!!7-111 ENGTH BELT CROSS SECTION
If a smaller sheave is used, short belt life should be expected; L in inches.
D2/D, Ka DESIG-
A SECTION B SECTION C SECTION D SECTION NATION A B Cc D E
1.000-1.019 1.00
= 3 in. Min. D, = 5.4 in. Min. D,; = 9 in. Min. D, = 13 in.
Min. D,
1.020-1.032 1.01 26 0.81
Pitch 1.033-1.055 1.02 31 0.84 ae
Belt Pitch Belt Pitch Belt Pitch Belt
Length 1.056-1.081 1.03 35 0.87 0.81
No. Length No. Length No. Length No.
1.082-1.109 1.04 38 0.88 0.83
27.3 B35 36.8 C51 53.9 D120 123.3 42 0.90 0.85
A26
A31 32.3 B38 39.8 C60 62.9 D128 131.3 1.110-1.142 1.05
36.3 B42 43.8 C68 70.9 D144 147.3 1.143-1.178 1.06 46 0.92 0.87 ae
A35
39.3 B46 47.8 C75 771.9 D158 161.3 1.179-1.222 1.07 51 0.94 0.89 0.80
A38
1.223-1.274 1.08 55 0.96 0.90 aol
43.3 B51 52.8 C81 83.9 D173 176.3 1.275-1.340 1.09 60 0.98 0.92 0.82
A42
A46 47.3 B55 56.8 C85 87.9 D180 183.3 68 1.00 0.95 0.85
ASI 52.3 B60 61.8 c90 92.9 D195 198.3 1.341-1.429 1.10
56.3 B68 69.8 C96 98.9 D210 213.3 1.430-1.562 1.11 75 1.02 0.97 0.87
AS55
1.563-1.814 1.12 80° | 1.04 se rs
A60 61.3 B75 76.8 C105 107.9 D240 240.8 1.815-2.948 1.13 81 Ke 0.98 0.89
69.3 B8l 82.8 C112 114.9 D270 270.8 2.949 and over | 1.14 85 1.05 0.99 0.90
A68
A75 76.3 B85 86.8 C120 122.9 D300 300.8 90 1.06 1.00 0.91
A80 81.3 B90 91.8 C128 130.9 D330 330.8
96 1.08 SS 0.92
A85 86.3 B97 98.8 C144 146.9 D360 360.8 97 1.02 sis
A90 91.3 B105 106.8 C158 160.9 D390 390.8 105 1.10 1.04 0.94
A96 97.3 B112 113.8 C173 175.9 D420 = 420.8 112 1.11 1.05 0.95 a
A105 106.3 B120 121.8 C180 182.9 D480 480.8 120 1.13 1.07 0.97 0.86 ;

Ali2 113.3 B128 129.8 C195 197.9 D540 540.8 128 1.14 1.08 0.98 0.87
A120 121.3 B144 145.8 C210 212.9 D600 600.8 144 1.11 1.00 0.90
A128 129.3 B158 159.8 C240 240.9 D660 660.8 158 1.13 1.02 0.92
B173 174.8 C270 270.9 TABLE 17.5 173 1.15 1.04 0.93 ele
ARC-OF-CONTACT 180 1.16 1.05 0.94 0.91
B180 181.8 C300 300.9 FACTORS, Ko!17-13)
B195 196.8 C330 330.9 195 1.18 1.07 0.96 0.92
B210 211.8 C360 360.9 D.—D, Ko 210 1.19 1.08 0.96 0.94
B240 240.3 C390 390.9 240 1.22 1.11 1.00 0.96
c VV | V-Flat 270 1.25 1.14 1.03 0.99
B270 270.3 C420 420.9 300 1.27 1.16 1.05 1.01
B300 300.3 0.00 1.00 | 0.75
Rated hp. Rated hp. Rated hp. 0.10 | 0.99 | 0.76 330 1.19 1.07 1.03
Constants: Constants: Constants: 0.20 | 0.97 | 0.78 360 1.21 1.09 1.05
a = 2.684 a = 8.792 a = 18.788 0.30 | 0.96 | 0.79 390 1.23 1.1 1.07
= 5.326 c = 38.819 e'=137.7 420 1.24 1.12 1.09
e = 0.0136 e = 0.0416 = 0.0848 0.40 | 0.94 | 0.80 480 1.16 1.12
0.50 0.93 0.81
SOME STOCK SHEAVE DIAMETERS 0.60 | 0.91 | 0.83 540 1.18 1.14
0.70 | 0.89 | 0.84 600 1.20 17
Varies by 0.2 in. Varies by 0.2 in. Varies by 0.5 in. Varies by 0.5 in.
660 1.23 .19
to 4.6; then 5, from 7 to 11 in.; from 13 to 16 in.;
from 2.6 through 0.80 0.87 | 0.85
5.2 in; then by 0.4 5.2, 5.4, 5.6, 6, then by 1 to 14; then 18, 22, 27, 0.90 | 0.85 | 0.85
by 2 to 20; then 33, 40, 48, 58 in.
to 6.4; then 7, 8.2, 6.4, 6.8, 7.4, 8.6, 1.00 | 0.82 | 0.82
9, 10:6, 42; 15, 9.4, 11, 12.4, 15.4, 24, 30, 36, 44, 1.10 | 0.80 | 0.80
18 in. 18.4, 20, 25, 30, 50 in.
38 in. 1.20 0:77 |) 0.77
130 | 0.73 | 0.73
1.40 | 0.70 | 0.70
|. 1.50 | 0.65 | 0.65
 ASAE Biel WIRED
V Bes SSR BED OT

Add 0.2 to the values given for each of the following conditions: continuous (over 16 hr/da
service; wet environment; idler in drive; speed-up drives. Subtract 0.2 if the opera 517.18] V-FLAT DRIVES AND OTHERS 461
quite intermittent or seasonal. These factors represent compromises of those found in
literature, with the most attention to Refs. (17.1, 17.11), and are primarily for V-belt; can be mounted into the grooves without harmful stretching and so that
but they may serve as a guide for other transmission elements. a the initial tension can be maintained. Sometimes idler pulleys are used on
open V-belt drives. Observe that if possible; not only standard belt lengths
DRIVING MACHINES are used, but also standard sheave sizes, approaching as closely as possible
esired velocit io; i j
Electric Motors Electric Motors ne 17.3. Of ih dale hacen 8 sizes are given at the bottom of
DRIVEN AC Split Phase AC Single Phase Series Woun Ta i s pecified sheave sizes and center distance must
i ene AC Normal Torque Squirrel | AC High Torque or Hight match a standard belt length. Now, the number of belts needed is
Cage, and Synchronous AC Slip Ring | } i
DC spon poeta AC Repulsion Induction (q) No“of belts) Design hp, equation (m)
Water Wheels AC Capacitor t Adi : :
Turbines, Steam and Water| DC Compound Wound Adlsted sated hp, equation (m)
Internal Combustion Engines | Steam Engines and Line Shaf Use the next larger whole number. More detail on a stress analysis of a
(Hydraulic Drive) Clutch on Driver or Driven Si V-belt is found in Ref. (17.17).

Agitators, liquid .
Cam Cutters .
Conveyors, package . 17.18 V-FLAT DRIVES AND OTHERS. V-belts can be used on
Drill presses, Lathes 1.1 1.2 nonparallel shafts, if the law of belting (§ 17.16) is satisfied, as seen in Fig.
otm neat p. 17.15 where a guide pulley is required (in the rear at top of illustration).
More commonly advantageous is the V-flat drive, in which the large wheel
Compressors and is a flat pulley (or flywheel), Fig. 17.16. Not only is a plain pulley less ex-
ret (rotating) 7 pensive than a grooved sheave, it often happens that an existing flywheel
cor ore, san i, mA or pulley may be utilized in a conversion job. The pulley face must be wide
Line Shafts i
Machine tools (other)
Printing machinery .
Pumps (rotating) .
Shears . 4

Ball mills .
Beaters (paper) FIGURE 17.15 V-Belts on Angled Shafts.
Circular saws . (Courtesy Gates Rubber Co., Denver).
Compressors (recip.)
Conveyors, bucket,
apron, screw, drag .
Crushers, jaw, etc. 1.4 1.6
Hammer mills
Pulverizers
Pumps (recip.)
Punches, presses .
Propellers .
Revolving screens
Tube mills.

Hoists .
Mine fans .
Positive blowers
Sl tac.- ne i BGURE 17.16 V-Flat Drive. Air com-
Tumbling barrels . gg on 50 hp at 855 rpm of 13-in.
Twisters (textile) . its ce eave; 54-in. flywheel; 8 D-section
*\Courtesy Gates Rubber Co., Denver).
RL
(tte
SOL FLEAIBLE POWER-TRANSMITIING ELEMENTS 720] ESP a ae
enough to accommodate the required number of belts. The power tr, to 2 smaller radius. These movements are simultaneous, the belt fitting
mitting capacity of V-flat drives may be computed as explained for
the sheaves at any setting.
drives except that there is a change of are correction factor Kg, Table 1 The Reeves variable-speed transmission, Fig. 17.18, operates on the same
this difference in K, is accounted for by the fact that the net tension ra rinciple with movable conical disks which can be controlled manually or
equation (d), § 17.3, should be less when the belts run over a flat pulley. automatically. On each wide side of a rubber belt are bolted hardwood
Double V-belts, designed to flex in both directions, are available; j blocks, as seen in Fig. 17.18, whose ends are beveled and tipped with leather
can be used for driving a sheave by the “back’’ side for instance. Alsoy ‘n order to improve the coefficient of friction. A variety of variable-speed
belts with multiple longitudinal V-grooves that run with matching gr. mechanisms have been developed, many of which operate on the principle
on the sheave (sometimes called ribbed belts) have been developed to of adjustable V-sheaves as described above. Some drives of this kind are
the same purposes as multiple V-belts. V-belts with steel-cable te integrally combined with motors to give in one unit a standard motor, a
members can be obtained for high-capacity needs. variable-speed control, and a speed reducer.
The operating temperature of V-belts should be less than 200°F,
ferably less than 160°F. V-belts that eliminate static electricity for us
explosive atmospheres (flour mills, for example) are available. Finall 17.20 TOOTHED BELTS. These belts, Fig. 17.19, are called timing
service factor for speed-up drives should be increased over the usual yal belts, and have other trade names. Since the usual tension members are
steel cables, they stretch very little under load and in service; therefore,
the initial tension may be low (low bearing loads) and tensioning devices
17.19 VARIABLE-SPEED TRANSMISSIONS. _ Variable-speed1 are not necessary. The backing and teeth are made of neoprene and covered
with a nylon duck facing. Since the drive is through the teeth, motions of
different shafts can be synchronized, as is so often necessary. Other notable
nealing furnaces, assembly conveyors, automatic welders, cement k characteristics include: large capacity in small space, can be designed for
inspection tables, metering systems, printing presses, pumps, stokers, tr: exceptionally light weight by using nonferrous pulleys, quiet operation at
ing bread ovens, and wire-coiling machines. Sheaves with movable low speeds, no lubrication needed, can tolerate low arc of contact (a
are used with V-belts to give a variable-speed adjustment. In Fig. 1 minimum of 6 teeth in contact is required for rated power), there is no
the upper shaft A drive the lower shaft B. The setting is for high s chordal action (as described for roller chains below); but they do transmit
the driven shaft B (small radius at B, large one at A). When th shock loads; if bent over too small pulleys, fatigue will occur."?:5:17-*4)
control operates via the member C, it can move the sides of the lowers They have been used up to 600 hp. Centrifugal action at high speed reduces
toward each other, causing the V-belt to climb to a larger radius, and
same time separate the sides of the upper sheaves, causing the belt ton

FIGURE 17.18 Reeves Variable-speed

Transmission. Constant-speed shaft is
A. Adjustment in illustration is for maxi-
mum speed of driven shaft B. (Courtesy
Reeves Pulley Co., Columbus, Ind.).

FIGURE 17.19 Timing Belt. (Courtesy

United States Rubber Co., N.Y.C.).
Oo PLEAIBLE FPUWER-IRANSMITIING ELEMENTS ([C gl7.4td AWLLEN VIAN WAVES we

the force available for transmitting power [equation (d), § 17.3], so ¢ with Newton’s law (F = ma). Moreover, note the vertical accelerations, be-
for the g- and 1}-in.-pitch belts, speeds above 8000 fpm are gen cause the chain naturally does some bouncing with this changing radius.
impracticable; higher speeds may be utilized with the small pitches. Sj Since rs = COS 6, the change of radius is r(1 — cos 0) and is seen to decrease
these belts are made endless and run on toothed wheels designed for hi as the number of teeth increases, thus reducing chordal action, which be-
one has the problem of adjusting pulley sizes, center distance, and comes almost negligible when there are 25 or more teeth.
length to accommodate stock parts if possible. For very slow speeds, the recommended minimum number of teeth on
Although toothed belts should undoubtedly be considered in any dj the smaller sprocket is N,,;, == 12} for.slow speeds. Ni, = 17; for moderate
design, we cannot afford the space to reproduce the catalog instt speeds, Nmnin = 21; for high speeds, Nin = 25: for speed increasing drives,
tions. 7-141 Ni, = 23-47-45] With odd tooth numbers on the smaller sprocket and an
suen number of pitches in the chain, the frequency of contact between a

17.21 ROLLER CHAIN DRIVES.

he particular tooth and a particular roller is a minimum, presumably better
distributing the wear.
The most widely used of
various types of chain power-drives is a roller chain, whose principal pe As for a belt, the centrifugal forces induce a tension in a chain of K =
are named in Fig. 17.20. The forces acting on the chain are much the wv.2/go, equation (e), § 17.3, and therefore the power that can be trans-
as those for the other drives of this chapter, except that significant im mitted for a particular maximum tension F, (Fz ~ 0 for chains) increases
(dynamic) load may occur when a roller makes contact with a spro¢ with speed to a peak and then decreases. Additionally for chains, the phe-
tooth and except that inertia forces arise because of chordal action,¢ nomenon that results in failure changes as the speed increases. At lower
cribed next. speeds, Fig. 17.22, failure is more likely to occur by fatigue of the link plate
'
The sprocket in Fig. 17.21 has much fewer teeth than actual sproc AB; at higher speeds, roller impact and joint \ ear will limit capacity BC;
usually have in order to accentuate the chordal phenomenon. In | finally, a point is reached where the capacity drops rapidly to zero when the
17.21(a), the roller A has just seated and the center line of the chaini load is great enough to break down the joint lubrication. The ratings, as in
a radius r,. This radius r, is smaller than the radius r after a rot Table 17.9, are for a service factor of 1 and an expected life of 15,000 hr.
through an angle @, as shown in Fig. 17.21(b). If we assume that this (where the chain length has presumably increased:3% maximum), and are
ket drives at a constant angular speed of n rpm, the speed of the center! predicated on adequate lubrication, the requirements being designated as
of the chain changes from v = 2zr,n to v = 2arn and back to the lo Type I, Type II, etc. Type I (v,,, = 300 fpm) is manual, oil being applied
speed during every cycle of tooth engagement. We recall that cha periodically with a brush, or can; in Type II (v,,, = 1300 fpm), oil is
velocity means an acceleration, and acceleration means force, in accorda’ supplied from a drip lubricator to link plate edges; Type III (v,,,, = 2300
fpm) is an oil bath or oil slinger disk, but a long length of chain should not
be immersed; in Type IV, oil is pumped and directed to the inside of the
lower strand (Fig. 17.26). Each size chain has its own speed limitations,
the limits for Type III lubrication being given in Table 17.8. The linear
(chain) speed may be computed from

5) ane a
Um = o(— a= 12 pm,

where D in. is the pitch diameter of the sprocket whose angular speed is
nrpm, P in, is the pitch, NV, is the number of teeth in the sprocket that runs
Corinecting: -
link plate atn rpm; PN, in. is the circumference of the sprocket. The rating tables are

FIGURE 17.2% Chordal Action. 1 FIGURE 17.22 Capacity vs. Speed [after Ref.
(17.21).

Power Capacity
angle 6 = 180/. t deg.; n 6 = Pit
or pitch diameter is D = P/sin 0.
ne

Joints CD
Speed

>
 §17.23] INVERTED-TOOTH CHAINS 467

corrosion-resistant metals, are available, the usual materials will be heat-

treated alloy steels. Good practices include: pins and bushings carburized

FIGURE 23 Multiple-chain Drive. Total of

strands of in. pitch, transmitting 300 hp. (Court
Whitney Co., Hartford, Conn.).

ees
FIGURE 17.24 Improving Fatigue Strength.

Va | a |

Embossed Grooves

entered with a design horsepower = (N,,)(nominal transmitted hp), and case-hardened for wear resistance; shot peened pins, rollers, link plates,
the design horsepower divided by the proper factor for multiple chain w ‘and bushings; pressing balls through the holes of the link plates for cold
(Fig. 17.23), given in the heading of Table 17.9. Since the service facto working to improve fatigue strength; and pressing around holes as sug-
Table 17.7 are in general conservative as compared with the recomm gested by Fig. 17.24. When necessary, the sprocket teeth are hardened (in-
tions of the chain catalogs, their use is permissible here. Extra capac duction, flame, carburizing).
results in longer life, which may be desired. For a single reduction, the velocity ratio should be kept below about 5,
The equations used to compute the rated power of the tab but as high as 10 has been used. Reference (2.1) has pertinent information
are :47.331 on wear in chains.

(r) hp = 0.004 -°8n?.9p3—0.07P and

17.22 EXAMPLE—ROLLER CHAIN DRIVE. A 15-hp motor, running at
[LINK PLATE FATIGUE] 1180 rpm, is to drive a 2-stage (reciprocating) hydrogen compressor. The velocity
ratio is to be about m,, = 3.2; service is nominally 8 hr/day. It has been decided
to use a 4-in. pitch (No. 40) chain (because this is the table at hand). Determine
.n = rpm of the smaller sprocket, Pin. = chain pitch, hp is treated the number of strands and sprocket sizes.
explained for the table values, and it is assumed that the limiting Solution. Use the recommended Ni = 21 teeth for the smaller sprocket,
for galling is not exceeded; K, = 17 for all chain numbers 40 to 24 and Nss = 1.4 from Table 17.7. The design horsepower is then (1.4)(15) = 21 hp.
except for No. 41, K, = 3.4, for Nos. 25 and 35 chains, K, = 29. Assume 4 strands. Divide by the factor 3.3, from Table 17.9, for 4 strands and
The center distance for chain drives may of course be relatively s get 21/3.3 = 6.36 hp/strand, rated value needed. From Table 17.9 for mi = 1180
but a minimum wrap of 120° is desirable; this condition is inevitably rpm, find the table value for 21 teeth to be 8 hp; therefore 4 strands is the answer.
when m, < 3. An average good center distance would be Dz + (A check for 3 strands shows that the chain would be about 4% overloaded.
os life a little shorter than 15,000 hr. should be satisfactory, 3 strands would do.)
where Dz is the pitch diameter of the larger sprocket, D, of the sm
or
The approximate length of chain is
_ PN.n _ (0.5)(21)(1180)
N, + No (No — Ni)? On
TNT ah 12
= 1030 fpm,
(s) Be oe + wae + pitches,
2 40C compared with vex = 1300 fpm for Type II lubrication, we see that Type II is
Satisfactory. For m,, = 3.2, Ne = 3.2 x 21 = 67.2, say 67 if this size is available
where C is in pitches (may contain a fraction). The length should b q from stock; Di = (0.5)(21)/7 = 3.34 in., D2 = (0.5)(67)/7 = 10.35 in., approxi-
even number of pitches to avoid using an offset link. There is the Mately. Without conflicting factors, use the recommended center distance; this
matter of adjusting chain length, center distance, and sprocket sizes so Computation and that for the length of chain is left to the student.
everything fits. Contrary to belting practice, the slack side of the chaim
preferred on the bottom of horizontal drives, especially for long ce
because, if the slack strand is on top, the strands may touch after the ch 17.23. INVERTED-TOOTH CHAINS. Inverted-tooth chains, com-
has lengthened in service. Lengthening occurs because of wear in the joi Monly called silent chains, are widely used for power transmission under
While chains of special materials, for example nonmagnetic” Much the same conditions as are roller chains, so that many of the remarks

466
 SABLE 11.9 FROPERTIES OF STANDARD $e eg
ROLLER CHAINS (REGULAR) b tela cr
* Rollerless. + Limiting chainspeed with oil bath (Type 111) lubri- E Bia ple st 3 fora stran ds. Especially for the higher speeds in
cation,!17.16) The fatigue strength of the chain may be taken as F,/4. le ; et 'ype IV lubrication region, er inverted-tooth chains.

S$
: 8 Veeen
ae
zelé
< SE/\S
a
3 : 5 7
aa
Mei
Revolutions Per Minute—Small Sprocket

|
| x
2] E/EES
<a
eo]
& 2
2
&
BD Bae Body Re ick s ‘
0 Pal <=] 5
BX &
| wand J %
25s* | 4 875 | 0.09 | 3500 | 0.155 | 0.19. | 0.0905 | 0.130 + | 0.030 RC Ve
35* | # | 2100 | 0.21 | 2800 | 0.231 | 0.283 | 0.141 | 0.200 | # | 0.050
cr +| 2000 (028) 2300 | 0.26 | 0.37 | 0.141 | 0.306 | 4 | 0.050 3
(%: + | 3700 | 0.42 | 2300 | 0.314 | 0:357| 0.156 | & fs | 0.060
§ | 6100 | 0-68"| 2000 | 0.398 | 0.434 | 0.200 | 0.400 | # | 0.080
60 | # | 8500] 1.00 | 1800 | 0.489 | 0.574 | 0.234 | 48 + | 0.094 |
80 | 1 | 14,500 | 1.73 | 1500 | 0.615 | 0.741] 0.312 | § §| 0.125
100 | 1% | 24,000 | 2.5 | 1300 | 0.754 | 0.882] 0.375 | 2 2| 0.156
120 | 14 | 34,000 | 3.69 | 1200 | 0.940] 1.116] % % | 1 | 0.187
140 | 12 | 46,000 | 5.0 | 1100 | 1.022} 1.210] 4 | 1 1 | 0.219
160 | 2 |58,000} 6.5 | 1000 | 1.228 | 1.383] 0.562 | 14 | 12 | 0.250
180 | 24 | 76,000 | 9.06] 950 | 1.362] 1.718] 0.687 | 1.406 | 138 | 0.281
200 | 24 | 95,000 | 10.65 | 900 | 1.546] 1.827] 0.781 | 1% | 14 | 0.312

on roller chains also apply to inverted-tooth chains. Horsepower the sprockets. For maximum quietness, use sprockets with 27 or more
ratinj
are given in manufacturers’ catalogs. The regular type of inverted-toot teeth. Both roller and silent chains may require an idler (or other means)
chains has links whose contact faces are straight, Fig. 17.25, and these fact to take up the slack.
contact a straight-tooth profile on the sprocket. i
The Link-Belt construction of the joint consists of case-hardene
bushings made in two parts that extend for the width of chain, with 17.24 WIRE ROPE. Wire ropes are made from cold-drawn wires
case-hardened pin joining the links. The Morse chain has a joint made u: that are first wrapped into strands; the strands are then wrapped into
of the two rockers seen in Fig. 17.25; as the joint works, these rockers roi helices about a core or central element, which is usually hemp or pulp,
on one another. Morse also makes a premium silent-chain drive, callec
Hy-Vo, for which the sprocket tooth has an involute profile; this plat
reduces the chordal effect materially, permitting quite high speeds when ther
are 25 or more teeth on the smaller sprocket (over 13,000 fpm for 1}-in
pitch). Some means must be provided to keep the chain on the sprockets
in Fig. 17.26, this is done with central guide links running in grooves
FIGURE 17.26 __ Silent-chain
Drive. Preferably, there is one
Stream of oil per inch of chain
FIGURE 17.25 Links for Invert: width. (Courtesy Socony-Vacuum
tooth Chain. As the chain Oil Co., N.Y.C.).
around the sprocket, the rocker
rolls on the seat pin. (Courtesy Morse 469
Chain Co., Ithaca). j
 Bd De
ST ee ne eee
ae ene Oe mene eee Cece we OD Lee g17.254 DESIGN CUNSIDERATIONS FOR WINKE nwre Pea

Fig. 17.27. Often, the central element is an independent wire rope cor
of cross sections, a few of which are seen in Fig. 17.28. The following
(IWRC), Fig. 17.28(e), which makes the rope much more resistant te
marks indicate something of the uses of the different types, any one o
crushing. Other factors that may make the IWRC preferred include: hig
which may have a wire rope core (IWRC).
temperature that may ruin a hemp core, its 7.5% greater strength (Tab 6 x 7, Fig. 17.28(a), being made of heavy wire, provides csiacioe *
AT 28), and the smaller elongation under load. The rope is made wif for
resistance to abrasion and wear; would be 7 x 7 with IWRC; used
either: a regular lay, in which the wires and strands are twisted in opposit iggi rd rails.
directions, Fig. 17.27(a); or Jang lay, in which the wires and strands a be 7280), being a compromise of flexibility and wear
twisted in the same direction, Fig. 17.27(b); the wires and strands may fi
vesistance, is one of the most popular styles; a good general purpose rope.
Special constructions, such as the Seale, Fig. 17.28(d), are designed for good
by having large wires on the outside, and for good flexibility by having
small wires in the inner layer. Uses, including all the various cross sections:
scraper and shovel cables, draglines, logging ropes, haulage, hoists.
6 x 37, Fig. 17.28(c), is an extra-flexible rope and therefore useful
(a) Regular Lay. These strands are right (b) Lang Lay. Both strands and wire are where abrasion is not severe and where relatively sharp bends must be
hand, the wires are wound left hand. right hand. tolerated. Without other designation, a 6 x 37 rope has wires all the same
diameter (as in Table AT 28); not as the 6 x 37 A of Fig. 17.28(c). Used
FIGURE 17.27 Lay of Wire Rope. (Courtesy John A. Roebling’s Sons Co., Trentor for winch lines, hawsers, overhead cranes, and hoists.
N.J.). The regular materials for wire rope are high-carbon, with most rope
made of improved plow steel (IPS), which has an ultimate tensile strength
either right-hand or left-hand helices. All wire rope was originally no) of s, between 240 and 280 ksi, the higher values applying to. he smaller
preformed, the wires and strands being twisted and bent into place, resultin wires (Chapter 6). Several other grades of steel are used :47-28] plow —
in high stresses in a straight, unloaded rope. At this time, most wire ropt (PS) (210 < s, < 240 ksi); mild plow steel (MPS) (180 <s, < ane
are preformed, the individual strands having been mechanically sha traction (180 < s, < 190 ksi); iron, with a lower carbon (about 0.1%)
ahead of time into the helical configuration they have in the rope. Preforme content (s, < 100), useful in undemanding situations; very highysicenert
ropes are more flexible and spool easier. (VHS) (280 < s, < 340 ksi), for the premium jobs, about 15 os stronger in
The size of a wire rope D, is the diameter of the circle that just cont rope form than IPS; and also used for a reason are galvanized wire ropes
the rope. In general, the greater the number of wires in a strand, the mor in various grades of steel, phosphor bronze, stainless steel.
flexible the rope; conversely, the fewer wires, the stiffer the rope. Thus The strength of the rope is always less than the sum of the strengths of
ropes made of small wires are suitable for sharp bends. However, the out the wires, and is commonly stated in tons of force to break, Table AT 28.
side wires are subjected to wear as they rub surfaces (pass over a shea Other information in Table AT 28: the weight per foot of length of rope
and small wires will wear through quicker than large ones. The construe w; the minimum permissible sheave diameter D,; the desirable minimum
is indicated by two numbers, the first giving the number of strands, tl sheave diameter; the approximate wire diameter D,, for the style of rope
second the number of wires in each strand. For example, a 6 x 19 wif in which all wires have the same diameter; the cross-sectional area of metal
rope has 6 strands each with 19 wires, Fig. 17.28(b). There are many style: in each rope A,,3 and the modulus of elasticity. E, of the rope (§ 17.25).

FIGURE 17.28 Cross Sections of Wire Roves. (Courtesy Jones & Laughlin Steel Corp.
Pittsburgh).
17.245 DESIGN CONSIDERATIONS FOR WIRE ROPE. There are
a number of applications where the design will be covered by code, often a
legal requirement; for example, elevators. We shall assume that if this is
true, the designer will meet the requirements of the code, and give our
brief attention to basi¢ considerations. There is of course some tensile
force F,, on the rope due to the loading on it (including inertia loading).
In addition, since the rope is often bent about a sheave, more stress is in-
curred on this account. Use the elastic curve equation M = E//r, in which
ris the radius to which a uniform, straight, homogeneous member is bent

Cheeta
 pe See TV VV EMI NANOMILtiING ELEMENTS iCn,|
g 17.25) ~ See ayia aries aeeat carte
by the moment M, equate this value of M to that in M = sI/c, and ge Tachau!?-25) found that the sum of the stresses due
s = Ec/r. If this equation is applied to a wire, for which the distance fr. on Ses and due to the load, s = F,/A,,, had only a fair
4 ei
that the
the neutral axis to the farthest fiber is c = D,,/2, bent about a sheave ¢
2 relation with fatigue failure (6 x 7 ropes not included), and
pitch radius r = D,/2, we get the approximate stress in the wire; / of this total stress was about 60 to 70 ksi. However, they found a
tl
area
() ceded correlation with the bearing pressure per square inch of projected
Ec Nes ie
Sp : D, psi or ksi # the rope on the sheave, p = 2F;,/(D,D,), Fig. 17.29, barca ie
y
le is taken as 180°. Fatigue failure seemed to be well predicte
where E = modulus of elasticity (3 x 10* ksi for steel)* and approximat ee /s,, (plotted against number of bending cycles to failure in Fig. 17.30),
factor
values of D,, are found in Table AT 28. This value of s multiplied by the ee Ss tie the ultimate strength of the wires. Introducing a design
D,D,,
of metal A,, gives the force on the rope that presumably would pro ¥, dividing both sides of the pressure equation by s, and solving for
this stress, called the equivalent bending load F, = SpAm. If the sheave we get
is the “desirable diameter’’ in Table AT 28, this equivalent force may be 2NF;
cared for by an overall design factor, but of course it increases as the sh
size decreases.
(w) Pele i Gueaee
ie
The modulus £, of the rope is much less than for the material, {7-26 where an appropriate value of p/s, is taken from Fig. 17.30. We note
because as the rope is loaded there is considerable sliding of the wires o indefinite life may be obtained when PIS < 0.0015 and that ne ee
one another. Thus, to get the extension of the wire rope under load, ui much greater than this if only a limited life is needed, but as usua
‘ fe
E, in equation (1.3) to estimate factor should be used to be on the safe side. The actual Seren poin :
fall surprisingly close to these curves, So that a large design factor oes a
(u) sL FL eee
u = = in., appear necessary if operating conditions are good. With
BO ae: desirable values of D, in terms of D, given in Table AT 28, we may ane y
with equation (w) in the form shown
where L in. is the length of the rope with a uniform stress F/Am3 see Tab investigate the effect of varying D,
AT 28. (unless the wire size is involved in F,, § 17.26).
The Roebling Handbook" "-?*! suggests minimum design factors ©}
guys, 3.5; miscellaneous hoisting equipment, 5; haulage ropes, cranes,
derricks, 6; small hoists, 7; hot ladle cranes, 8; when the factor is de
Tea
as N = F,/F,, where F, is the breaking strength, Table AT 28, and
the maximum working load on the rope (§ 17.26). However, it seems
dent to include the effect of bending the rope in the definition of N, si FIGURE 17.29
unless it is stated otherwise in the problems, use
(Breaking strength, F,) — (Equivalent bending load, F,) F, —F =
(v) N= : = F; F;
Tensile force in the rope, F; Fa
where the numerator is approximately the available strength left to carry
an external load F,; F, = NF, + F,. 4
The trouble with the conventional design procedure as described abovi 0.006
is that it is not in accord with the manner in which most wire ropes fail
It is a static approach, whereas, if the rope is continually flexed, failurt
FIGURE 17.30 Pressure Ratio
would be expected to be a fatigue failure (except as abrasion may wear oul vs. Cycles to Failure. (After 0.004
Drucker and Tachau.!!7-251)
*The reader may, if he wishes, conform to past conventional practice by using the p/su
modulus of elasticity E, of the rope (Table AT 28) in equation (t), but since in the ideal
model, all wires bend to a radius D,/2, we may as well be theoretically consistent 0.002
and use E for the wire material; especially inasmuch as this static approach tends te
underestimate the fatigue anyway.
;

0 2x10 4x10 6x10° 8x 10°

Number of Bends to Failure

474 FLEXIBLE POWER-TRANSMITTING ELEMENTS [Ch. {7

A theoretical and experimental study of the contact stresses, wire o1

wire, by Starkey and Cress"™7-27! confirms the foregoing design criterion i
that it would be normal to expect a direct relation between the pressure;
and the Hertz contact stress; the Hertz stress appears to be the significan FIGURE 17.31 Free Body of Hoist.
one, rather than tension plus bending. Starkey and Cress postulate;
failure theory: the relative motion of the wires, particularly during bending
together with the high contact stresses, causes fretting at points wher
fatigue cracks were observed to start and propagate. 4
If corrosion or other unfavorable operating conditions are involved
fatigue life will be shortened. The groove of the sheave should be wid
enough not to pinch the rope, which rests on the bottom. In general, th
softer the groove, the longer is the rope’s life; but the pressure of the rog
on the sheave may be too high. To avoid an excessive rate of wear (abrasio n
Roebling recommends limiting the pressure p for 6 x 19 rope as follow Use kips for convenience. The free-body diagram, Fig. 17.31, shows the load
Wn =2 +3 = Skips, the weight of the rope, which is long enough to be sig-
on cast-iron sheave, 500 psi; on cast steel, 900 psi; on manganese stee
nificant, as wL = (1.6)(1000)(10-%)D,? = 1.6D,2kips (Table AT 28), and F; is
2500 psi. 3
the force on the straight rope at the top. Summing forces with the direction of
Since a typical fatigue failure starts with the breaking of individu: the acceleration (up) as positive, we have =F = ma,

F-5-16D2 = P79),
5 + 1.6D,?
sive and design for fatigue may be the better solution. Of course, there a
F, = 5.62 + 1.8D,?.
situations where the cycles of loading are few enough that static streng
is sufficient. Loads suddenly applied to a slack rope can snap it in two. TI Use this value of F:, the minimum s,, = 240 ksi from § 17.24, N = 1.3, p/s, =
0.0015 for indefinite life from Fig. 17.30, and equation (w), and get
shorter the rope, the less energy it can absorb within its breaking stren
(F8/2 for elastic energy). A good operator is sure that the slack is taken 2NF, _ (2)(1.3)(5.62 + 1.8D,?)
D,D; =
gradually before applying full power. During manufacture, the rope’s ¢ (pisu)su (0.0015)(240)
saturated with a lubricant for the purposes of reducing wear from the w At this point, one can assume a value of Ds; and solve directly for D,. We see
rubbing on each other and the sheaves and of protecting the rope from that the larger Ds, the smaller D,, and since the rope is long, there may be an
and corrosion; for best service, it should be kept clean and lubricated fro economic advantage in using the desirable D; = 45D;, Table AT 28. With this
time to time. Recent tests suggest consideration of molybdenum disulfil value as the first trial, we find D, = 1.125 in. from the foregoing equation, which
as a lubricant. is exactly a 14-in. rope; Ds = (45)(1.125) = 50.6, say 50in. A check is always
desirable (usually, the standard sizes will depart further from those first computed),
So solve for N;
ae (0.0015)s.D,D; __ (0.0015)(240)(1.125)(50)
= 1.28,
17.26 EXAMPLE—WIRE ROPE FOR MINE HOIST. A skip for a 2F; ~ 2(5.62 + 1.8 x 1.125?)
shaft weighs 2000 lb. and is to lift a maximum load of 3000 lb. from a depth 0 which, as expected, is close enough to the desired 1.3. Use 14-in. rope on 50 in.
1000 ft. The maximum speed of 20 fps is attained in 5 sec. (a) What sizes i Sheave. At a mine shaft there should be room enough for this size sheave. Let the
6 x 19, IPS rope and sheave should be used for indefinite life and for N = Teader try other ratios of D;/D,,.
on the basis of fatigue? (b) What is the design factor on the static basis? (c) W (b) To get the design factor defined by (v), we have, from Table AT 28, F, =
sizes of rope and sheave are required if the number of cycles (bending and 52.6 tons or 105.2 kips, D,, ~ 0.067D;, Am ~ 0.4D;?. Then with equation (t),
bending) in the desired life is 200,000? What is the corresponding static Ree Eoin _@ x POO ESTP OEP
factor? (d) What is the elongation of the rope found in (c) if the 3000 Ib. load = 22.9 kips
added while the hoist hangs free at the bottom? /
Solution. (a) The acceleration, which introduces inertia effects, depent for D, = 1.125. Since F; = 5.62 + 1.8D,? = 7.9 kips,
on the torque characteristics of the motor (or on the brake properties if th N = (Fy — Fy)/F: = (105.2 — 22.9)/7.9 = 10.4.
maximum loading occurs during stopping as the hoist is moving down); but n
conventional approach is to use the average; that is, a = Av/At = 20/5 = 4 fp: This result indicates that if the design had been on the static basis with a more
 ete | ot) SRS er an ee Ee Segue: | Ate eeet ey So ee
oo ee Oe a §17.29] PULLEYS AND SHEAVES ne
normal design factor (say 5), it could not have been expected to withstand ; on
indefinite number of flexings. jimiting friction), with negligible centrifugal effects, as would be likely
a
(c) If we could study the operation, or if a similar operation were at hoists, elevators, etc., is
har
we could estimate the number of flexings that might be expected in some
des F,
lifetime. Certainly we should note that it would take a while to load, aie ple:
return, and therefore indefinite life may be unnecessary. Suppose
hoist, di
(x) Fy
we guess 6r
trips per hour; for 300 working days of 8 hr. per year, and a life of
7 yr. (whi
may well be too long for a mine shaft at a given level), this gives about
10° roy Compare with equation (d), § 17.3. Find F, and F, only by free bodies of
trips, or flexings. Now double this, in case means are found to speed up the o W and CW, including inertia. Some values of f in equation (x) for this
tion or in case of extra shifts at times, and get 200,000 cycles as specified
in ¢ lication are:@7-28!
problem statement. Enter Fig. 17.30 with 200,000, find p/s, ~ 0.028, substity ae or steel sheave: greasy rope, 0.07; wet rope, 0.085; dry rope, 0.12.
this value in equation (w) with D; = 45D,; Wood-lined sheave: greasy rope, 0.14; wet rope, 0.17; dry rope, 023%.
45p.2 = 2NFe_ _ (2)(1.3(5.62 + 1.8D,?) Rubber- or leather-lined sheave: greasy rope, 0.205; wet rope, 0.4; dry rope,
" “(p/su)su—S—SC«M(0.0028)(2.40) 0.495.
from which D, = 0.756; use #in.; D, = (45)(0.75) = 33.8, say
34 in. Che k
before. From Table AT 28: F, = 23.8 tons or 47.6 kips, There are a number of fittings
TINGS FOR WIRE ROPE.
F, me,= EDwAm _ms (3 X 10%(0.067D,)(0.4D,") aa i. developed for wire rope,"'7-?® of which the hook is shown
D, 34 = 9.97 kips,
in Fig. 17.33, and a thimble with clip in Fig. 17.34. The connection with the
F, = 5.62 + 1.8D,? = 5.62 + (1.8)(0.75%) = 6.63 kips, rope is a socket fitting in Fig. 17.33; the wires are separated and spread into
F.-F, 47.6 — 9.97 a conical shape in the socket after which zinc is poured in. Properly done,
N= eT weenie 5.68,
this connection should be as strong as the rope. The thimble is frequently
which is observed to be more in line with the design factors given in § 17.25. 4 used for connections; the rope may be turned over the thimble with its
(d) The elongation due to a static load of 3000 Ib. on the 3-in. rope, eq! end spliced to the main part of the rope (not illustrated), or clips, of which
(u) and Table AT 28, is approximately one is shown in Fig. 17.34, may be used.
__FL _ (3000100012)
° = AE, ~ 4 x 0.752 x 105 = 13:35 in.
17.29 PULLEYS AND SHEAVES. Various materials are used for
pulleys and sheaves, the most common being cast iron and steel. There are
17.27 TRACTION DRIVES. The friction between the sheave notable differences in the construction of the pulleys and sheaves of Figs.
rope is often used as a driving force; for example, if shaft B, Fig. 17. 17.35-17.40, which a careful examination will reveal. Wood pulleys have
driven by a motor (likely via a speed reduction), if W represents the a use but they need to be shellacked and varnished at intervals to keep out
and CW represents the counterweight, and if load W is moving up moisture. Paper pulleys, Fig. 17.36, are very popular in small sizes because
of their good coefficient of friction. Steel pulleys, Fig. 17.37, are lighter
than cast iron pulleys and can be safely run at higher speeds (without

FIGURE 17.32 Traction Drive. FIGURE 17.33 Hook and Open Socket,
Drop-forged Steel. (Courtesy John A.
Roebling’s Sons Co., Trenton, N.J.).

slowing down), then F, is the “tight”? tension and F, < F,. The drivi
frictional force is F, — F,, which can be computed: from the power C
speed at B. The maximum value of the ratio of the forces (at the point 6
FIGURE 17.34 Thimble with a Clip Con-
nection. (Courtesy John A. Roebling’s Sons
Co., Trenton, N.J.).
 §17.30] HARMONIC DRIVE 479

can be bored to different shaft diameters. Sheave grooves for wire rope
have rounded bottoms and should, in general, not pinch the rope, contrary

z
to V-belt practice. Pulleys, etc. that run at high speeds should be well
balanced dynamically as well as statically.

FIGURE 17.35 Wooden Step Cone. The dia- FIGURE 17.36 Fiber (Paper) Pulley. Notice the
meters of matching steps on step cones must be metal sleeve and metal rim plate. Pulleys a
such that each step requires the same length of belt, crowned (this one larger at the middle) to keep th
17.30 HARMONIC DRIVE. The Harmonic Drive,“7:°5] a new and
equation (17.3). (Courtesy Dodge Mfg. Corp., belt from running off.['7-31] (Courtesy The Rock= novel power-transmitting device, does not fit into any of the traditional
Mishawaka, Ind.) . wood Mfg. Co., Indianapolis). classifications. It is made in many configurations, but the basic elements are
a rigid circular spline, which, in Fig. 17.41, is the external member A, a
flexible spline B, called a “‘flexspline,’’ and a rotating elliptical member C
(or other element) whose function is to thrust the teeth on the flexspline into
mesh with the teeth on the rigid spline at two or more sectors (two in Fig.
17.41). The member C is called a wave generator. The teeth are standard
30° involute splines, § 10.7, and are the same pitch on both members.
The flexspline in its undeformed shape is circular and it has fewer teeth than
the external member A (which has internal teeth); the difference in the
number of teeth is always a multiple of the number of lobes in the wave
generator C—say, 2 or 4 in this case. To understand its operation, note that
the teeth are making contact in the vicinity of M, M, at the major diameter
of the wave generator rotating clockwise. The design may be such that either

FIGURE 17.41 Harmonic Drive. As C rotates, teeth are thrust into engagement with a
FIGURE 17.37 Split Steel Pulley. Some steel FIGURE 17.38 \-Belt Sheave. Cast iron, with minimum of sliding, forcing B to rotate if A is held stationary. (Courtesy United Shoe
pulleys are welded. Desired size of bushing in- split, tapered bushings available for different sizes Machinery Corp., Beverly, Mass.)
serted. (Courtesy Dodge Mfg. Co., Mishawaka, of shafts. (Courtesy Allis-Chalmers Mfg.
Ind.). Milwaukee).

FIGURE 17.39 Cast-iron Wire-rope Sheave. FIGURE 17.40 Roller-chain Sprocket W

(Courtesy John A. Roebling’s Sons Co., (Courtesy Link-Belt Co., Chicago).
Trenton, N.J.).

Lf
danger of flying apart from centrifugal force). The separate bushing of
Fig. 17.38 is an economic idea applied to all kinds of pulleys; the main body

478
480 FLEXIBLE POWER-TRANSMITTING ELEMENTS [Ch. 17

the rigid spline or the flexspline may. be stationary. Let the rigid spline be
stationary. As C rotates, other teeth come into mesh and the flexspline
moves. Examining the relations of the teeth between P and Q, Fig. 17.41,
note that the flexspline moves counterclockwise, or in the sense opposite
to C. The velocity ratio is
No. of teeth on the output member
(y) mM, =~ A Rea?
_ Difference in No. of teeth on the two splines
Suppose the rigid spline has 132 teeth, the flexspline 130; since the output
member is the flexspline, the velocity ratio is 130/2 = 65. The conseque
of this development is that a high velocity ratio can be obtained in a sm
*‘package.’’ Other characteristics claimed by the manufacturer inclu
high torque capacity, accurate angular positioning, high efficiency,
noise level, and collinear input and output shafts.

17.31 CLOSURE. There are a number of operating variable:

initial tension, for example, that affect the efficiency of these drives, and t
efficiency will also vary because of large differences in the amount of po
18. BRAKES AND
transmitted. Properly installed and maintained, efficiencies better than 9.
should be attained. Reflect on the discussions of this chapter and not
CLUTCHES
design generally is against fatigue failure because of the continually vary
stress, especially due to flexing (except the flexing of the chain cause:
18.1 INTRODUCTION. This chapter deals largely with brakes,
wear at the joints). Also, in each case, an undersized element results 1
which are friction devices used to regulate the motion of bodies (slowing
shorter life (not necessarily an immediate breakage), an oversized elemer
them down, holding their speed constant, holding them at rest, etc.) and
results in longer life. The percentage increase in life is generally mucel
with clutches, which are friction devices used to connect shafts (speeding
greater than the increase in cost, so it just depends on whether or not
up the driven bodies to the same angular velocity as the driving shaft).
longer life is desired. For example, a 100% increase in wear life is obtai We may classify such friction devices into types as follows: external shoe
for 20% increase in cost by going from a }-in., No. 40 roller chain t on drum, internal shoe on drum, disk on disk, cone in conical surface,
8-in., No. 50.%7-21] If, say 6 V-belts are required by normal design
wrapping bands on drums, and expanding bands on drums. Any one of
cedures, 7 belts result in 40% increased life for 15% greater cost, 5 b these fundamental types might be either a brake or a clutch, although the
result in 35% less life.“7-1*! The cost of lost production owing to a dri details of the brake design would be different from the details of the clutch
failure should not be overlooked. The manufacturers of these vari design. However, each type has a characteristic use; an external shoe on a
power transmitting elements have accumulated much engineering kn drum is most likely to be a brake, a disk on disk is most likely to be a clutch,
how that should be called upon if there are any unusual factors. And finally but the cones and disks and others are widely used for both. We shall
there are technical aspects that cannot be covered in this brief review."7 discuss a few of these friction devices, not exhaustively, but briefly in their
most likely forms, secure in the belief that the reader can make a transition
from a brake to a clutch, or vice versa, as needed.

18.2 FRICTIONAL WORK AND POWER. Frictional work is

Simply work done by a frictional force F. If a brake is used on a moving
body, it brings about a decrease of kinetic energy KE or it opposes a loss of
Potential energy PE, or both, the first consequence being an increase
482 BRAKES AND CLUTCHES [Ch.| 518.3] COMPUTING THE ENERGY TO BE ABSORBED 483

average speed. The limiting capacity of a brake is commonly expressed

contacting surfaces). This is to say that the temperatures of the bods in terms of the maximum instantaneous rate of energy absorption, Fv, or
increase, which results in the frictional work being eventually dissipated¢ fhp, but the total amount of energy involved is also vital. For example,
the surroundings as heat. Getting rid of the energy as heat takes time, an certain disk brakes on railway cars could be subjected to 1000 fhp maximum
for this reason, the capacity of a brake is often stated in terms of the amour for short-time application without damage, but to only 75 fhp for steady-
of frictional work U, it can absorb in a particular time or in terms of frig state application.“®-®! Stated another way, this brake absorbs 7 x 108 ft-lb.
tional horsepower (fhp). Sometimes a brake has to be applied steadily of energy without damage in 20 sec., but it takes some 8 min. for most of
long periods of time, in which case it must be able to radiate and cony this energy to be radiated and convected to the atmosphere.
the heat to the atmosphere at such a rate that the steady-state tempera
is below a damaging value, § 18.4. Sometimes the brake is applied
short intervals of time intermittently, with enough time between applicat 18.3 COMPUTING THE ENERGY TO BE ABSORBED. The body
for it to cool to a value close to the environmental temperature, in being braked may undergo a change of potential energy —APE,
case it may safely absorb energy at a much higher time rate, the ener
being mostly used to heat the metal parts. The rate at which heat is (c) —APE = Wh, — hy),
ducted through the metal is much greater than the rate at which where a body of weight W moves from an elevation h, to ha, each measured
radiated and convected to the atmosphere. If the human element is a prin from the same datum level. The minus sign appears before APE because
factor, there is no telling what the pattern of brake application may b the conventional interpretation of Ax is x. — x,, which convention makes
The theory of this chapter is from elementary mechanics. An impo the energy absorbed by the brake a negative number as computed from the
parameter for brakes and clutches is the frictional torque 7, (a rote change of motion of the braked body; that is, the only purpose of the
body is nearly always involved), and brake capacity is often express minus sign is to make the energy quantities positive—but —APE may be
in terms of T,. The reader will be reminded later of force analyses of seve positive or negative, depending on the direction of motion. The change of
configurations, from which T; is found. Recall now that work is equal kinetic energy (mv?/2) of a translating body of weight W Ib. is
“force times distance,”’ and for a rotating body, work is T dy, whe
radians is the angle the body turns while subjected to any torque T. If (d) —AKE = sea’ — Vsq”) ft-lb., [TRANSLATION]
constant or assumed to be some average value, the work of a fri &
force on the surface of a rotating drum of diameter D in., Fig. 18.1,
where v,, fps is the initial speed, v,. fps is the final speed, g fps? is the local
acceleration [W/g = m slugs (for g fps”) is the mass of the body]; on or near
U;
the surface of the earth, use g = 32.2 fps®. For a rotating body, the change
2nN,” of kinetic energy is
where N; is the number of turns of the drum while F and T, are actif T
(% = 27N, radians), and U, and 7, have the same units, usually e (e) —AKE = meet — w,7) ft-lb., [ROTATION]
in-lb. or ft-lb. Also, for a constant angular acceleration « rad./s
T; = Ix ft-lb., where J slug-ft.? (for feet, seconds) is the polar mo 4 where J slug-ft.2 is the moment of inertia of the body about its axis of
of inertia of rotating members about the axis of rotation. ‘a Totation, w, and we rad./sec. are, respectively, the initial and final angular
The power associated with 7; is T;w, where rad./unit of time is # velocities. For a rolling body,
angular velocity of the drum; w = 27n, where n is in rpm or rps. Sometim
a brake is used to keep a body moving downward with more or less ¢0) W I
(f) —AKE = beet — Bey”) + 5(os" — w,2) ft-lb.,
stant speed, but more often the speed changes, frequently to a state of ré
We write [ROLLING BoDy]

(T;, ft-lb.)wm_, (Ty; in-lb.)n where the bar over the symbols means that the velocities are for the center
(b) fhp = of gravity and the moment of inertia is with respect to that gravity axis
33,000 —«~6 3,000
Perpendicular to the plane of rotation. Energy is a scalar quantity; the
for w,, rad./min., nm rpm, where the frictional horsepower fhp is for som total energy U, that a brake absorbs isthedecrease of the stored mechanical
instantaneous speed w,,n; often, an average fhp is computed for @ €nergy brought about by the braking. Use equation (a) to find the
484 | BRAKES AND CLUTCHES [Ch, §18.5] EXAMPLE—DRUM TEMPERATURE AND fhp 485

corresponding frictional force or torque, and (b) to convert to hor The heating effects are naturally quite complex and local temperatures at
power. If more than one body has a significant energy change, find points of contact may be very high. That part of the total frictional energy
change for each and add algebraically to get the total energy to that is stored in the brake parts, principally in the drum or disk, has been
absorbed by the brake. variously estimated at 75% up, but it actually varies with the duration of
the application as well as with the quantity of energy. For very short appli-
cations of the brake, say a few seconds, the percentage may be nearly 100%
at the instant that the brake is released. An assumption made in brake-
18.4 PERMISSIBLE ENERGY ABSORPTION AND OTHER DESIG drum design for airplanes was that the drum would absorb 92% of the
DATA. For design values of the coefficient of friction, see Table AT2 kinetic energy and be at a temperature of 1075°F at the stop."®-2°) After
also § 18.14 for some remarks on it. a while, the energy stored in the parts by the braking will be rejected to the
The permissible normal pressures between the braking surfaces deps
sink. If long continuous applications are necessary, the drum and other
in varying degree on the brake-lining material, the coefficient of frictic adjacent parts keep rising in temperature until the amount of energy lost
and on the maximum rate at which energy is to be absorbed. The high
to the surroundings is equal to the rate at which frictional energy is
pressure, the greater is the rate of wear and the energy absorbed at a pi generated. Thus, as the magnitude of pv, increases, the duration of the
cular speed. Recommended design values of p are given in Table AT; braking must decrease in order to avoid overheating.
but the stated values are exceeded on occasion. Moreover, they ma For peak short-time requirements, it is generally assumed that all the
lower than shown when application periods may be frequent and frictional energy is absorbed by the adjacent metal in the drum wheel,
In short, experience with given materials and class of service tells wha which gives a temperature increase At of the metal as
best for design. . :
The pressure p psi is for the projected area A of the shoe; p = ©) Meee U, ft eS
where N is the total normal force (detail later). The actual rate of en Wm
dissipation is fNv, = Fu, = [pAvp ft-lb/min. at v,,fpm, from which
see that it is proportional to pv, ft-Ib/sq. in.-min. (or pv,/33,000 hp/s where w, lb. is the mass of the metal absorbing the energy U;, and
for a particular brake. On this account, design values of pu, are fo c ft-lb/Ib-°F is the average specific heat of the metal for the temperature
the literature. Typical values for industrial shoe brakes are:"®-1*! fr range (for a large temperature range, c varies markedly). Suitable for ordi-
application, pv,, = 5500; average use, pv, = 16,500; infrequent short nary calculations: for cast iron, c ~ 101; for steel, c ~ 93; for Al,c = 195.
use, pU, = 49,500. Raybestos-Manhattan"-15! states that most ma In automobile brakes (Fig. 18.5), a desirable maximum instantaneous
turers’ design values of pv,, will fall within the range of 1500 and 1 loading is about 2.2 fhp/sq. in. of brake rubbing surface on the drum.
The following values, in agreement with Hiitte, are widely quote One way to compute such an “instantaneous” value is from a known or
reproduced here for their informational value: assumed deceleration. For example, 15 fps? may be taken as a reasonable
pum = 55,000 for intermittent applications of the load, comparativel value for automobiles, which means that the speed decreases 15 fps in
periods of rest, poor dissipation of heat (wood blocks); for f = 0.25, this valu One second; the fhp loading can be computed from any top speed during,
equivalent to about 0.42 fhp/sq. in. absorbed; 3 Say, one second.
Pvm = 28,000 for continuous application of the load as in lowering operat The effectiveness of the brake may greatly decrease shortly after it
poor heat dissipation (wood blocks); equivalent to 0.21 fhp/sq. in. for f = begins to act continuously, a phenomenon called fade. This is basically due
PUm = 83,000 for continuous application of the load, good heat dissi to a significant decrease in the coefficient of friction at the high surface
(oil bath); equivalent to about 0.63 fhp/sq. in. for f = 0.25. ; temperatures induced, and it can be combated to some extent by the design
Compare these various values with some peak values found in # of the braking system; for example, by searching for a configuration such
literature: automobiles, (100 psi)(2500 fpm) = 250,000 at 600°F; ¢ that the frictional moment T, divided by the applied moment Wa (T;/Wa)
moving tractors (50 psi)(5000 fpm) = 250,000 at 800°F; railway Shows a minimum variation when plotted against the coefficient of friction
(100 psi)(10,000 fpm) = 1,000,000. 4 (see mechanical advantage below), or by more effective dissipation of heat.
For open and exposed band brakes, § 18.9, Rasmussen"®-14! recor
mends an energy adsorption capacity of 0.2 to 0.3 fhp per square inch '
brake contact area. " s EXAMPLE—DRUM TEMPERATURE AND fhp. A 3000-lb. auto-
The temperatures are limited by the properties of the materials (Tab pk. moving on a level road at 60 mph, is braked so that the deceleration is
AT 29). Calculations that purport to give the surface temperatures © d ‘stant at 20 fps? (close to a normal maximum) until it stops. (If it were going
only approximate some kind of average, but estimations must be mac ©wnhill, the loss of PE would be added to the loss of KE; if uphill, gravity would
486 BRAKES AND CLUTCHES|
help slow down the car.) Each cast-iron brake drum is D = 9 in. in diay
b = 2t in. face width, and t = ¥% in. thick. (a) Assuming that the radiat
conyected heat is negligible, compute the temperature rise of the drum.
is the average rate of power absorption (fhp) during the first second? Duri
last second? (c) What is the over-all average fhp?
Solution. (a) Since the process of conduction in iron is relatively
get a reasonable estimate of conditions by assuming that all the energy o}
is absorbed by the drum, keeping in mind that temperature differences
FIGURE 18.1 Block Brake.
to exist. From Table AT 6, we find the density of cast iron is 0.253
Thus the approximate mass of 4 drums is [V = 7Dbt + (7D2/4)1]

Wm ores= pV = (4)(0.253)| n(9)(2.25)(2u) + #(81)(3)

ae = 24.1 1b. |
18.6 BLOCK BRAKE, SMALL BLOCKS. If the block is short
(Considering attached metal, the effective mass may be greater.) Th
enough (small 6, Fig. 18.1), it is reasonably accurate to assume that the
60 mph is the same as (60)(5280)/3600 = 88 fps. The 3000-Ib. car loses ;
resultant frictional force F is tangent to the drum at the center of the block,
kinetic energy (rotational KE of wheels neglected on the assumption
an assumption often made for quick calculations. The free body of the
small part of the total)
brake lever and block, solid outline, Fig. 18.1, serves to illustrate the
=~ A KE EE,
Wo,”
2
eariomiccintanes
(3000)(88)?
54.
EEC. tienes Sa 361,000 ft-lb.
-lb. procedure for all such cases and shows all the forces: W, the force applied
to the lever shown parallel to the normal force N (its components parallel
and perpendicular to N would be the most convenient if W were at another
angle); F = fN, where f is the “kinetic” coefficient of friction; and Fy,
U; 361,000
tT Wall | CAIRO en unknown in magnitude and direction, the pin reaction, which would be
At = — = ——____ = °F,

used to design the pivot connection A. Assuming that the reader is qualified
If the ambient temperature is 100°F, the drum is at 249°F average. 4 to find F, and use it, we shall concentrate on the braking aspects, and
(b) During the first second, the velocity drops from 88 fps to 68 eliminate F, by choosing A as the center of moments;
—20 fps?), and therefore the loss of kinetic energy per second is
F

sti
—AKE = Elva? — vn?) = ss? — 68%) = 145,000 ft-Ib/sec. (h) EM a en le AR AEE Clit ete

fhp
_ 145,000 = 264 hp average for 1 sec.
; The location of the pivot A with respect to the line of action of the frictional
| oan force F is significant in brakes. For the position as given, we see that the
The approximate area of the 4 rubbing surfaces is sense of the moment of F is the same as that for W; that is, the frictional
A = (4)(7Db) = (4)(7)(9)(2.25) = 254 sq. in.; force helps to apply the brake. When this condition exists, the brake is said
to have self-actuating properties or to be self-energizing. Considering
264/254 = 1.04 fhp/sq. in. (But note that during the first fraction of
Fig. 18.1, note that if the pivot is located at the more likely point A’, the
the rate is higher than 1.04.) During the final second, the car stops from
Moment of the frictional force opposes the moment of the applied force W
of 20 fps, and fhp is — AKE/550 or
and there is no self-energizing effect; one may think of this as a negative
-AKE __Wo,? __ (3000)(20)? self-energizing. Thus, it takes a larger applied force W to produce a parti-
fhp = = 33.9 hp.,
550 — (2g)(550) (64.4)(550) cular braking force F when the pivot is at A’ than when it is at A, other
or 33.9/254 = 0.133 fhp/sq. in. Conditions remaining the same. However, if this is a double-block brake
(c) For constant acceleration, a = Av/Ar or Ar = Av/a = 88/20 = With another lever and block pivoted at A”, the more likely construction,
to stop the car. With an initial KE = 361,000 ft-lb., the over-all time We see that F’ provides self-actuation of the lower lever while F opposes the
loss is at the rate of @pplied moment on the upper lever; the effects are reversed if the rotation
_ _361,000 1S reversed. In actual brake analysis, one must make a free body of each
fhp = 149 hp.,
(4.4)(550) €ver and its block; F’ is not necessarily equal to F and the braking effects
or 149/254 = 0.587 fhp/sq. in. of rubbing surface on the drum. are generally different, as explained. For practice, the student should write

487
aa of

A N W
‘ Brake Lever.
¢ )

FIGURE 18.2 Two-shoe Brake

With Solenoid. A spring applies the
prake with force W and can be
adjusted somewhat to increase or
decrease the braking moment. When
FIGURE 18.1 (Repeated). the current is on, flowing through the
motor, it flows through the solenoid
too, and the solenoid keeps the force
off the brake; braking occurs when
current ceases to flow, as when an
elevator or hoist is stopped. In a
the equivalent of equation (h) for different locations of the pivot A andfi force analysis, consider the free
different directions of rotation. a bodies of links AFC, BHF, and CKE,
Notice in equation (h) that if b/f = e, then W = 0, a limiting (and qui as shown in the kinematic representa-
unstable) condition meaning that no force is necessary to apply the tion. See §18.7. The bolts K and B
may be considered to be tight enough ‘ne
Ife > b/f, W is negative, which means that a force is necessary to diseng
to make a rigid connection. (Courtesy i
the brake, once it is engaged, or that less than no force is required to Westinghouse Electric Corp., East
braking; consequently, the brake is self-locking. A few numbers in eq Pittsburgh).
(h) would soon demonstrate that it would be necessary to have an un
configuration (large e) to make a simple block brake self-locking; b r tap
self-locking were desired, as it sometimes is, § 18.9, other kinds of bre
are more adaptable for this purpose.
d WML
C EN
In addition to a brake’s capacity to absorb energy, its frictiona a ALT
braking torque T, is used for rating. For the single lever of Fig. 18.1, s¢
for F from equation (h), multiply both sides by the drum radius D/2,
get the braking torque 7; for an applied force W; i 18.7 FORCES FOR LONG SHOES. The error resulting from the
assumptions of uniform pressure and the frictional force tangent to rim at
FD ~ ~WDfa the center of the shoes, as for the short shoe above, becomes greater as the
@ AP oe py arc subtended by the brake-shoe surface increases. Typical values of 9,
Fig. 18.3, are in the vicinity of 90°. To improve the analysis, one may
If there are two or more blocks on the same drum, the safe procedure | s assume something that is expected to approximate actual conditions—the
find T, for each block and add the results. distribution of the normal pressure p on the shoe, for example. For the
In the actual design of a brake, one would probably draw curves, sh configurations shown in Fig. 18.3, the assumption that wear is proportional
ing how the action of the brake varies with changes in certain dime to the pressure results in a cosine distribution of pressure with respect to
and in the coefficient of friction, which is inherently variable. A dimens: the line OB; so we might as well assume this distribution, p = P cos «,
less parameter used for such studies, sometimes called the mechani where P is the proportionality constant and the value of p when cos « = 1.
advantage MA, is defined as the ratio of the braking torque 7, divided” Such a distribution is suggested by experimental evidence. Should the shoe
the applied moment; for example, in Fig. 18.1, for the single lever, MA and drum be relatively flexible, the pressure distribution would be more
T,|Wa, where T; is given by (i). if nearly uniform. Actually, the distribution changes with lining wear and
The advantage of using opposed double blocks or two-shoe brakes with the magnitude of the applied forces. The following analysis applies
that the normal forces more or less balance one another, significan to pivoted shoes, Fig. 18.3(a), where the axis OB from which « is measured
reducing the bearing loads and the bending moment on the drum sha Passes through the pin whose center is B (not as drawn), and it applies to
The details of the linkage designs vary widely, Fig. 18.2 being an illustratio Fig. 18.3(b) when the axis OB is perpendicular to OH; with the additional
since this is so, there is no recourse except for the engineer to apply # limitation that the contact surface is symmetric about OB, an unnecessary
principles of mechanics in finding the magnitudes of the various fore limitation that the reader can easily change by changing the limits of
and moments for the particular configuration involved. Integration.
 319.7] FORCES FOR LONG SHOES 491

where C is the distance from O to the line of action of F, either construction,

andit has the same units as r or D. In brake parlance, point B is called the
center of pressure. Thus, if the pivot pin is located at point B, there is no
tendency for the shoe to tip and concentrate the pressure on the toe or heel
of the lining, and the lines of action of N and F pass through point B. How-
ever, point B is ordinarily so close to the drum surface that there is not
room enough for a pin (though other designs might be used to get the pivot
virtually at B). Hence, for the construction shown in Fig. 18.3(a), the result-
ant F tends to turn the shoe CC, producing a greater pressure at the top
end of the shoe than at the bottom; and the sinusoidal distribution with
the center at OB is therefore no longer true. (If the pin, Fig. 18.3(a), is
(a) (b) fairly close to B, the equations for T;, etc., give reasonable answers for the
FIGURE 18.3 Long Shoes. In (a), there is a pivoted shoe; for p = P cos <, force Ng pivoted shoe.)
through the pin and B; force F on the shoe acts through B. Notice that for CL rota’ The normal force N is the sum of the components in the direction OB,
as constructed in (a), the resultant F has a moment about H that makes it self-er N = JdN cos x, with dN from equation (j);
and that the force F on a matching shoe on the opposite side of the drum makes it
self-energizing. In (b) with CL rotation, F as shown results in negative self-energ! +012 6 in 6
the opposite matching shoe would be self-energizing. When braking is largely for (m) N= | dncos x = Por { cos? « dx = Pbr(—= =").
tion of motion, the shoe that is not self-energizing can be made with a smaller con! —6/2
face, or the linkage system could be designed to produce a larger normal force o
negative self-energizing shoe than on the other. The brake shoe is ordinarily made 8} The frictional torque 7; is the moment of F about O, Mr,o = T; = Sr dF,
metric about the horizontal center line; in (b), it is intentionally as shown in order to ac with the value of dF from (j);
tuate the conditions for the formulas as derived.
+012 6
(n) T= fr dF =sPbr? | cos « dx = 2fPbr? sin >
Consider the area dA = b(r dx), Fig. 18.3, where 5 is the face width —6/2
the braking surface and r dx = ds. For a pressure p, the normal and fr
tional forces are where the units are in.-lb. or ft-lb., depending on the linear units used.
Equations (m) and (n) apply to both configurations in Fig. 18.3 if the pres-
(j) dN = pbrdx = Pbrcosadx and dF = fpbr dx = fPbrcosad sure distribution is sinusoidal as explained; as usual F = fN. It may. be
where we have substituted p = Pcos«. The moment of dF about sor convenient to note that 7, = FC, where C is defined by (I). The maximum
point B, Fig. 18.3, with f assumed constant, is (Og = C cos a) pressure P may be computed from any of the foregoing equations containing
it, say equation (m)—boundary conditions as defined for Fig. 18.3.
dMy)_ = (dF)e = dF(C cos « — r) = fPbr(C cos « — r)cos « da, — For the lever as a free body in Fig. 18.3(a), the forces Fand N act through
the center of the pin (assuming that the pin connection can offer no turning
(k) dMy)_ = fPbr(C cos*« — r cos «)dx. moment). However, for the rigidly attached shoe of Fig. 18.3(b), one must
Since P, b, r, and C are constants for a particular brake, the integral of tk
integrate for the moment of dF about H, or use the moment arm to H of
the F vector through B; forces W, F, N, H as shown in Fig. 18.3(b) consti-
expression (k) between the limits of —0/2 to +6/2 gives the moment Of
tute the free body for the link and shoe. If the lining does not extend as far
about some point B as =
as the line of action of N (¢2 < 90°), the actual maximum pressure is that
“ sin 2« : 9/2 at the end of the lining (P sin ¢5).
Mp jp = fbr] e(; + ) —rsin “| é
For best understanding, the reader should solve some problems
2 4 0/2
(different from the foregoing) by integration, running checks as possible
The moment of F about a point on its own line of action is zero; therefol by the stated principles. One variation consists of deriving the equations
if My,z = 0, the resulting value of C locates the line of action of F. Sint In terms of $, $9, Figs. 18.3 and 18.4, for which the sinusoidal pressure is
only the parts within the brackets can equal zero, we have P = Psin ¢. See Fig. 18.4 and show that, forp = P sin ¢,
4r sin 0/2 gi 2D sin 6/2
(0) T; = Mpio = fbr*P(cos $, — cos $2),
® @+sin0 6+sin0’
 FIGURE 18.5 Internal Shoe Brake.
(Courtesy Chrysler Corp., New
York).
L FIGURE 18.4 Unsymmetric External Shoe.

others, the shoe pivot is attached to another link and is free to move within
R | Sa
a short arc. With this considerable variation in detail, a number of analyses
(p) Myx = forP| r(e0s ¢, — COS Pd.) — “5 (ain — sin?) of internal-shoe brakes have been made.'®-2.18.3.18.41 For a fixed pivot,
the braking torque, etc., may be determined on the basis of a sinusoidal
pressure distribution (p = P cos « when « is measured from OB, Fig. 18.3,
brRP : :
the perpendicular to the radial line from the pivot to the drum center;
(q) Myjn = Ty Rea — $,) — (sin 24. — sin 2¢,)],
or p = P sin ¢, where ¢ is measured from the radial line from the pivot to
the drum center). The moment of the frictional force and of the normal
force about the fixed pivot of the shoe can be found by integration, as
wit eee by the procedure in § 18.7. Horsepower capacities are mentioned
With My,z and My, as positive numbers, the sum of the moments
in 4,
respect to H in Fig. 18.4 is
In the two-shoe automotive brake of Fig. 18.5, observe that if the car
(r) iMy = Wa + Myx ome Myx = 0. is moving forward, the frictional forces on the shoes have counterclockwise
moments about their respective pivots and that these moments help to apply
For opposite rotation, the sign of Mp, would reverse. The reaction
the brakes—that is, both shoes are self-energizing, a good arrangement for
the pin is obtained from the sums of components in the x and y directi front-wheel brakes, which do most of the braking. If the car is backing, the
for the force system in Fig. 18.4, a
moments of the frictional forces about the pivots tend to reduce the pressure
on the brake drum. The result is that the brakes are less effective (for a
(s) iF, —H, — WceosBp + | avsing + | aFcoss = 0, Particular foot force) when the car is moving backwards than when it is
moving forward. Maybe you have noticed.

(t) LF, = —H,+ W sin p — | dNvcos4 + | dF sing = 0,

ag BAND BRAKES. __ Ina band brake, the band wraps partly around
where the integrations are made as_ previously explained; He b € brake wheel or drum, and braking action is obtained by pulling the
(H,2 + H,?)". It is worth noting that the equilibriant of W and 4 and tight onto the wheel, Fig. 18.6. The braking force F is the difference
equal to the resultant of F and N (the lines of action of three forces !
equilibrium intersect at a point).

The problem of braking an automobi FIGURE 18.6 Band Brake.

18.8 INTERNAL SHOE.
has led to the development of several types of brakes with pivoted intermé
shoes. In some brakes, the shoes are on a fixed pivot, as in Fig. 18.551

492
 FIGURE 18.7 Differential Band Brake.
FIGURE 18.6 (Repeated).
EEE
- ZB YATE

between the tensions at the two ends of the band, F = F, — Fp. Th

analysis is similar to that given for belts, except that the centrifugal f
is absent. Directly from equation (d), § 17.3, we get y
F,
(18.1) wi fo,
Fy,
FIGURE 18.8 Differential Back Stop. This is a band
where F, is the tension of the tight side of the band, F the smaller tensi brake used as a stop to prevent reverse motion. Compare
and @ the angle of contact between the band and the wheel in radia with Fig. 18.7. It is used on the motor shaft or counter-
The conceptual tools for computations on band brakes include eq shaft of an elevator, hoist, or conveyor.
(18.1), F = F, — F,, and an equilibrium equation, as equation (u)
for the differential band brake; the detail is made clear by the
below. In Fig. 18.6, if the rotation is reversed, the larger tension
Oppose W, resulting in a smaller braking torque for a particular W.
In the differential band brake, Fig. 18.7, the tension on one end
brake band aids in applying the brake. Such brakes are self-ene
and may indeed be self-acting, Fig. 18.8. A force analysis, with the drum, we may multiply both sides of equation (w) by D/2, solve for FD/2 =
lever as the free body, Fig. 18.7, will suggest some characteristics 0! T,, and find
type of brake. Using the fulcrum B as the center of moments in or¢
eliminate the pin reaction at B from the moment equation, Fig. 18.7,
(x) pa ette a),
WDa(e’? — 1
2(c — bef?)
(u) 2M; = Wa + F,b — Fyc = 0.
The pressure (see Table AT 29) on a band brake is computed as
Hence F,b is in the same sense as Wa, and the externally applied force V F
assisted by the tight tension F, in putting on the brake. To get the re
of the frictional force F and the applied force W, substitute the value fA
F, = Fie" into F, — F2 = F, and find in which A is the contact area between the band and the drum. According
to Hagenbook,"®-®! suitable drum diameters may fall between (7; in-lb.),
F Fef@
(v) F, = a and then, F, = Seda 1/3 1/3
(z) =) <D< 2) :
5 4
Use these values of F, and F, in equation (u) and solve for W;
Another relationship that may help size the drum is
_ F(c — be!) (a) (60 fhp)"’* < D < (80 fhp)#,
(#) ~ aele — 1)’ Where fhp is taken as the maximum horsepower to be dissipated in any
which shows that W will be negative if be’? > c; that is, the brake fifteen-minute period. See also fhp capacity for band brakes mentioned in
self-acting and grab once the frictional force begins to act. § 18.4, Although the empirical values given by the independent equations
Since the braking torque is 7; = FD/2, where D/2 is the radius of |

494
496 BRAKES AND CLUTCHES [¢ 18.11] GENERAL REMARKS ON DISK CLUTCHES 497

(z) and (a) are presumed to be reasonable, new designs should, if possi Disk brakes have found favor in some areas of heavy duty because they
follow known successful practice. can be designed to radiate heat faster with less trouble from fading.
Chrysler claims that their self-energizing disk brakes operate at tempera-
tures 35 % lower than usual (no detail given). Self-energizing is attained by
18.10 DISK FRICTIONAL TORQUE. Disk surfaces used virtue of the wedging action of steel balls that move up ramps between the
clutches or brakes generally have an annular shape, Fig. 18.9. The theoret disks and press outward on the friction surfaces when the brake is applied.
equation for 7; may be based on either of two assumptions: (1) the p
is uniformly distributed over the surface, or (2) the wear on the surfac
uniform. Since (2) is a little more conservative than (1), and gives a sim)
result besides, we shall use uniform wear.
The wear may be expected to be proportional to the intensity
FIGURE 18.10 Plate Clutch. The release bearing A is mounted
pressure p and the rubbing speed. In turn, the rubbing speed is proporti on the shaft K to the transmission, and when the clutch pedal is
to the distance from the center of rotation p, Fig. 18.9. Thus, wear depressed, the bearing A moves toward the flywheel B, contacting
portional to pp; and if wear is uniform, pp = C, a constant. Con the inner ends of the release levers C. The release levers are pivoted
the differential area dA = p d@ dp, Fig. 18.9; the normal force on it at D, on pins mounted in the eyebolts E. The outer ends of the re-
dN = pdA = pp dé dp. Letting pp = C and integrating, we get i lease levers engage the pressure plate lugs F, pulling the pressure
plate G away from the driven plate H, compressing the several coil
To p2n springs J, and disconnecting the drive from the flywheel (which is
(b) N= i | C dé dp = 2nC(r, — r,), or C= attached to the engine shaft) to the shaft K. When the foot pressure
nv 0 is removed from the clutch pedal, the springs 7 force the pressure
plate G forward against the driven plate H, and power is trans-
mitted through two pairs of friction faces, one pair on each side of
an inside radius r;. To get the frictional torque, multiply dN by f :
the driven plate H, which is mounted on a splined connection M.
dF = f dN, the moment of which about the central axis of the are: (Courtesy Borg and Beck, Chicago).

dT; = p dF = pf dN = fpp? dd dp = fCp

dé dp.

Integrate with f constant [see Ref. (18.7) for f variable];

T, p2n
(c) T; = Cf I | p d0 dp = aCf(r,? — r;?). 18.11 GENERAL REMARKS ON DISK CLUTCHES. Think of
N% 0
equation (18.2) as applied to a clutch; it shows that for a given N and f,
Substituting the value of C from (b) into (c), we get maximum torque is transmitted when r,; is a maximum, but if r, ~ r,,
the pressure (psi) would be quite high and the clutch would rapidly wear
Nr, +
(18.2) T; = mete = fNrm [1 PAIR FAC out. Although an approximate optimum balance of wear and torque is
obtained for r, = 2r;, such an optimum may not be the best solution for a
where r,, is the mean radius of the annular surface, rm = (r, + r)/2, @ Particular application; more typical, r, + 1.3r;.®-7) For example, a certain
N the axial force between a pair of faces in contact. The total torque t b heavy tractor clutch has r, = 1.33r,;, in which the friction disks are lubri-
can be transmitted when more than one pair of faces is in contact wit cated sintered metal for which wear was not a serious factor; on the other
normal force N on each pair is the value from (18.2) multiplied by| hand, a large tractor had a dry friction disk clutch for which ro = 2.27.
number of pairs in contact. _ An automotive type of disk clutch, called a single-plate clutch, is shown
in Fig. 18.10, in which it is seen that a single-plate clutch has two pairs of
faces in contact, each with the normal force N produced by the springs J.
In the multiple-disk clutch of Fig. 18.11, all surfaces are subjected to vir-
tually the same normal force, and there are 4 pairs of clutch surfaces.
Multiple disks, while frequently needed to handle the load, have the
tisadvantage of heavier rotating masses (with correspondingly greater
FIGURE 18.9 Disk Friction. Nertia).
 518.13] BRAKING MATERIALS 499

18.12 CONE CLUTCH. In the diagrammatic representation of

Fig. 18.12, a forked lever engaging the groove G moves the part B in and
FIGURE 18.11 Miultiple-disk Clutch, out of contact with E, thus connecting and disconnecting the shafts.
(Courtesy Link-Belt Co., Chicago). Think of a free body of the male member B. The resultant normal force
on each half is shown, N/2, Fig. 18.12; the frictional force F is also dis-
tributed, F = fN. Assuming that the resultant frictional force acts at the
mean diameter, we find the frictional torque to be
FD, _INDn SDpR
(d) T= iz 2 ~ (sin a + fcos a)
To be on the safe side, design values of the coefficient of friction f where the value of N has been substituted from (e) below. The normal force
clutches should probably be somewhat lower than those for brakes, s N arises from the spring force R, which needs to be large enough to over-
about 0.2 to 0.3 for dry asbestos facing and about 0.1 or less in oil. T come the frictional force F’ during engagement, as well as to produce the
normal pressure between the disks may be of the order of 25 to 30 ps normal force necessary for the power transmission; F’ is also taken as fN.
commercial vehicles, but some 30 to 37 psi for passenger-car servi Sum forces horizontally and get
Pressures are as low as 15-18 psi for bus service and other heavy: dh
service where starts are frequent. Another manufacturer of industt (e) R=WNsina + F’ cosa = Msin« + fcos «).
clutches (pedal-operated) shows pressures varying from 42 psi for thi
While the capacity T, increases as « decreases, « should be greater than 8°
size to 26 psi for the 14-in. size, and on a type with a relatively large in
if the clutch is not to have a strong tendency to grab; typical « = 12°.
radius r,, Fig. 18.9, pressures up to 60 psi are used. In addition to var
the pressure to fit the service, apply a service factor to the computed to
if operating conditions are not favorable.
18.13 BRAKING MATERIALS. Stress relieved cast iron, with alloys
For clutches that cycle frequently, there may be a heating proble to qualify it for about class 30, makes an excellent brake drum material.
inasmuch as it takes time to dissipate the frictional energy. The work
Some manufacturers use nodular iron; steel, stainless steel, monel, and
by the frictional torque will be equal to the energy (kinetic, plus pote
aluminum are sometimes appropriate. Since the conductivity of aluminum
if a body moves upward) that the driven body gains while the clutch
alloys is some twice that for iron or steel, the frictional energy generated
have relative motion (driven parts brought up to speed); this torque ma’
at the braking surface is conducted away faster, tending toward a lower
may not be the maximum that can be delivered. When engageme
surface temperature. It is thought that this phenomenon reduces the loss
complete, the frictional work is theoretically zero and no longer a prob
of braking from fading by avoiding some reduction in f. Presuming suffi-
In addition to frictional energy, electromagnetic clutches must dissi cient strength, we may say that the life of a brake drum depends on its
the electrical losses.®-19) If the rate of frictional work during engage
Tesistance to thermal fatigue and to wear. The braking members are sub-
is high, excessive surface temperatures may result in welding of me
jected to temperature gradients, often to the limit of their capacity. The
plates, in disintegration of friction materials, warping, large variation hottest layers expand the most; if they are on the inside of a drum, the
the coefficient of friction, and early failure.“8-7) if Outside layers of the drum are in tension and the inside ones in compression.
If the highest stress in the stress gradient exceeds the yield strength at its
Operating temperature, permanent deformations and residual stresses are
the result; if yielding occurs at the hot inside, the stresses are reversed after
ambient temperature is again reached, tension inside, compression outside.
FIGURE 18.12 Cone Clutch. Forces sho Brake drums that have been subjected to severe cycling of this sort develop
are acting on the male member. The norm
evident “temperature checks,” which may lead to a fatigue failure. A
tributed; so that the resultant of that part Smooth drum surface is desirable to avoid excessive wear on the lining,
normal force on the upper half is in the but it is soon unavoidably roughened somewhat from galling.
of the paper and equal to N/2. The result Brake linings are made of many materials, but most of them are of some
frictional force F on either half will mixture of asbestos and a binder, molded or woven, with or without wire
normal to the plane of the paper.
Ze
Ye
 500 BRAKES AND CLUTCHES [Ch,| §18.15] OTHER TYPES OF BRAKES AND CLUTCHES 501

or metal chips or metal powder. Since the metal is for the purpose of impro may be associated with changes in the deposited film); after run-in, f is
ing the heat conduction, it is usually brass or copper. For higher temper, different for different metals (for examples, at a load of 40 lb., f ~ 0.4 for
tures (above 400°F) and heavier loads, sintered metal linings, say wit Al, f © 0.5 for C.I., f = 0.7 for steel).
either a copper or iron base with lead, tin, graphite, and/or silica in th Miscellaneous factors that bring about changes in f include: materials,
mixture, will absorb energy faster with less checking of the drum. Fo surface finish, surface temperature, pressure, rubbing speed, foreign matter
temperature range of 750° to 1000°F, sintered mixtures containing ceramic including the deposited film on the rubbing surface, moisture in the atmos-
are used."8-18] Any of the general classes of linings may be used either dj phere, number of cycles of operation, and changes of properties of the
or in oil. Oil is sometimes used only as a coolant; at other times, it; fluid, if any, on the braking surfaces. As in any other design, a margin of
necessary as a lubricant for other parts. When used in oil, the pressure mu: safety is needed; that is, the design is such that the designer is convinced
be greatly increased (and the lining must be able to withstand the pressure that the desired braking or clutching effect can be obtained, or that his
Table AT 29, for perhaps two reasons: first, a heavy pressure is needed t design has properties as close to the desired ones as possible. As a safety
break through the oil film between band and drum or disk, and second, measure, some designers.use f = 0.25 in design when the typical experi-
large pressure offsets the much lower coefficient of friction. mental coefficient for the usual friction materials is about 0.4.
For metal on metal, it may be more safely assumed that the coefficient
of friction is independent of the pressure; however, with significant changes
18.14 COEFFICIENT OF FRICTION. Data on the coefficient¢ of speed, f may vary widely. One series of tests for steel on steel (with
friction in general seem to be so disparate that the only safe generalizatic Fe,0, film) showed f decreasing from 0.4 at 500 fpm to 0.1 at 5000
is that the coefficient of friction varies. For clutch and brake material, tt fpm;"-22] with a film of molybdenum disulfide MoS, for the same speeds,
conventional rule for friction, to wit, that the coefficient of frictio on f varied from about 0.13 to 0.09. Even though a good frictional force is
desired, the latter combination with low f’s may be preferred in some
applications because of the lesser variation of f. Starting with quite clean
steel-on-steel surfaces, we find the coefficient varies through a large range,
often from values well above 1, to much lower values as the surface oxydizes,
°
8
Apparent Coefficient of Friction

as it does in the usual operating surroundings. With some materials (e.g.,

polyethylene, polytetrafluoroethylene), the coefficient decreases markedly
at low speeds as the speed decreases!!®-29]contrast with Fig. 18.13. And
0.15
so it goes.
FIGURE 18.13 Coefficient of Friction. After
(18.21). The variation of f for a cork-facing material f
on>) a clutch in oil (automatic transmission).
0.10
18.15 OTHER TYPES OF BRAKES AND CLUTCHES. In addition
p= 150 psi to the fact that the various basic elements previously described can be
1 1 1 L J 1 }
adapted to either brake or clutch operation, there are a number of special
So
So
ao

400 800 1200 1600

Rubbing Speed, fpm designs, often patented. Grooved wheels, much like a V-belt sheave, are
used when the wedging effect of the groove is helpful; water brakes are
indispensable in some circumstances; electrical and magnetic clutches and
brakes (dynamometer) are quite common, as are fluid couplings. In one
regarding friction [also Ref. (12.22)], found a much larger spread for df class, centrifugal force causes a friction member to move outward and press
molded material, from f ~ 2 for a load of 2 lb. to f ~ 0.6 for a load C against a drum on the driven member, the higher the speed, the greater the
80 Ib., though the change is asymtotic with a relatively small difference frictional force. Representative of this category is the friction-band clutch
some cases for loads of 20 lb. and 80 Ib. (area = 0.5 x 0.625). Some othe of Fig. 18.14, where a driving arm attached to the driving shaft tows the
findings reported by these experimenters are: the coefficient of friction i weighted friction band which is free to exert radial pressure in accordance
affected by the presence of a film deposited on the metal surface by ths with the law of centrifugal force (mv?/r = mrw?). Thus, the torque that
friction material; for the molded materials tested and the speed rang may be transmitted depends upon the weight of the friction band (with its
(40-240 fpm), f changed little with sliding velocity; the variation of f with attached material), the speed of the driving shaft, and the inside diameter of
temperature when t < 500°F was small (decrease of fat higher temperatures the drum. Interesting characteristics of centrifugal clutches include:
 Y
il, 4 &S
S
§18.16] CLOSURE 503

18.16 CLOSURE. Brakes and clutches, like some of the other machine
elements, can often be selected from the product lines of specialized manu-
facturers. Both of these elements deserve the use of a service factor to care
for extreme operating conditions. We have said nothing of the design of
the links, pins, and shafts for strength (and the design of bearings) on the
assumption that, with a knowledge of all the forces, the reader is equipped
to design these things by the principles already covered.

‘
FIGURE 18.14 Friction-band Clutch. This starter is applicable for one direction of driy
Starters designed for reversible drives are available. They may also be obtained in conju
tion with fiat-belt, V-belt, chain, and gear drives. Rigid shoes subjected to centrifugal for
are also used in a similar manner. (Courtesy J. P. Madden, Bethlehem Steel Co.)

FIGURE 18.15 “‘Airflex”” Clutch. (Court

Federal Fawick Corp., Cleveland). i

q iy

(1) the driving shaft almost reaches its operating speed before the dr.
subjected to a significant torque; (2) consequently, the size of a prime
may be smaller than would be necessary if high starting torque we:
quired at low speed; and (3) the clutch may be designed to slip befor
dangerous overload occurs.
The “‘Airflex” idea, Fig. 18.15, is used for both brakes and clu
The normal force on the friction surface is applied by air pressure ir
flexible tube. This type of clutch is made in ratings from 5 hp to 2880|
at 100 rpm. The idea of air-pressure operation is also used on disks. —
ig

502 4
f
 19. DESIGN OF
WELDED JOINTS

19.1 INTRODUCTION. Welding, rather than being a machine

element, is a manufacturing process, which reminds us that there are many
facets of design in addition to stress analysis. In fact, stress analysis and
sizing of parts likely consume only a minor part of the total design time.
In most instances, designs are affected in some significant way by manufac-
turing processes, which must be learned about elsewhere, perhaps with help
from the instructor. However, since the conventional stress analyses of
welds often take on an esoteric flavor, they are worth attention. The effect
of this manufacturing process on design is great enough to give welded
machines and machine elements a distinctive appearance (see Figs. 2.16,
§2.24, 13.18, § 13.32, and Fig. 19.1). The designer must exercise his
ingenuity in applying welding advantageously in his own designs; but there
is considerable help in the literature.
Whether to weld, cast, or forge, etc., is an economic problem that may
be answered correctly in different ways, depending upon local circumstances.
Welding may be the least expensive process where the pattern cost for
Castings would be a large percentage of the total cost or where there are
unusual machining or casting difficulties. Large weldments are built up
from easily fabricated parts, as in the gear blank of Fig. 13.18. To meet
the demand, there are available a number of special rolled shapes which are
Produced particularly for weldments, special screws and studs designed to
be welded in place, etc.
 oe T fay ep eee en we PS eae eae. Tawra

The standard full fillet weld has a section of an isosceles right triangle, as
shown, often with the legs b of the triangle equal to the plate thickness,
although they are also often less than this; moreover, one leg may be longer
FIGURE 19.1 A Frame of Welded Construction. (Cour than the other. A reinforced fillet weld, Fig. 19.3(a), is one that has a throat
Lukenweld, Inc., Coatesville, Pa.)
dimension t’ > bcos 45°. For a particular amount of metal, a fillet weld
with a concave surface (not shown) is relatively weak. However, the sharp
corner where the weld joins the surface of the plate at B, Fig. 19.3(a), and,
to a somewhat lesser extent, the junction of a standard fillet weld at C,
are points of stress concentration. If the joint is subjected to repeated
loading, it may be worth the cost to make this junction less abrupt; how-
ever, do not overlook consideration of discontinuities such as P, P’,
Fig. 19.3.
The nominal stress as computed for the configurations of Fig. 19.3 is
taken to be a shear stress, and the area assumed to be resisting the load is
always the throat area, because weld failures are more often across the
throat—but the size of the weld is its leg dimension b. Throat areas for
FIGURE 19.2 Butt Joint. In (a), a double-V groove,¥ Fig. 19.3(a) are tL at C and t’L at B; but assume that both welds are stan-
CT reinforcement on both sides; in (b), a single-V groove, reinfo re dard, each with area tL; then for Figs. 19.3(a), (b), and (d), the nominal
in (c), a square groove, reinforced; in (d), single-U, reinforce stress is computed from
removed. :
(b) F = s,(2tL) = 2s,Lb cos 45°,
where t = bcos 45°. Although, as you readily recognize by now, this
simple equation does not care for the complications of the actual stress

FIGURE 19.3 Fillet Welds. The various dotted boundaries about the welds suggest the
19.2 BUTT JOINT. The plates for butt joints, Fig. 19.2, may | penetration of weld metal. The points P designate theoretical points of stress concentration
unbeveled (square groove), beveled on one side only, or beveled on at the heel. When the base metal has melted and solidified with filler metal, the stress con-
sides; the groove is given various shapes such as V, U, J, for rea centration points become such as P’, suggested in (a), (b) and (d), which, it might be ob-
served, may become very sharp discontinuities. In (a), transverse loading, the upper weld is
associated with the economics of the job. A butt weld may be reinf standard, the lower one is reinforced; usually the welds are made alike. Fatigue loaded, the
on both sides, on one side, or not at all. When there is no reinforcemen reinforced fillet weld would not be expected to be any stronger, if as strong, as the standard
the usual practice is to build up the bead and then machine or grind weld. Lap welds are sometimes single-welded. In (b) and (c) are indicated closed T-joints,
flush. With or without reinforcement, the nominal average stress is cor one (b) with no bevel, one (c) with a single bevel. Double V and other groove shapes are used.
puted from In (b), note the sharp stress concentration point P’, which difficulty, for varying loading, can
be alleviated by double bevel, welded from both sides with the opposite welds joining each
(a) F= sL, { other. In (b), the full strength of the plate M can be attained (static loading) by welds with
5 = 0.75 h. For the load F as shown in (c), it is evident that a symmetric disposition of weld
where L is the length of weld and ¢ is the plate thickness, Fig. 19.2. Tes! metal with respect to F would be preferred, but it is not always possible to weld from both
suggest that reinforced butt welds on mild steel made with filler rod of ab o1 Sides. In (d) are shown standard fillet welds, loaded longitudinally. Drawing symbols for
the same mechanical properties as the parent metal have about the sam designating welds are found in various books."9-3]
Static strength as the plates, but it would be safer to assume a relatiy
strength (§ 17.6),
-6), or efficiency, y, of of y7 = 90°
90% or less; ; note the stress concentra Sepre pr x
tion due to the discontinuity of the bead. Lee 4 7
t L
j (b) T-Joint \\ G aA
19.3 FILLET WELDS. Fillet welds may be loaded transversely. Weld Length=L
Fig. 19.3(a) and (b), or longitudinally (parallel welds, that is parallel te t . ‘ PF ‘
the vector F), Fig. 19.3(d), or of course at some other angle with vector F,

506

(a) Lap Joint (c) T-Joint (d) Parallel Loading

 §19.4] FILLET WELDS, ECCENTRIC LOADING
\
t 509
(b) T-Joint \\ (a) Case 1, Fig. 19.4. Imagine the force F shown replaced by a force
Weld Length =L F at the weld, that produces a shear stress, and a couple Fa, which is the
t
bending moment; M = Fa. The section modulus through the throat, the
conventional approach, is tL?/6 for each weld or Z = tL?/3 for both sides.
These values substituted into s = Mc/I give the normal stress,
M 3Fa ot 3Fa 4.24Fa
(a) Lap Joint (ce) T-Joint (d) Parallel Loading (c) in
tL? bL? cos 45° BL
FIGURE 19.3 (Repeate Assuming a uniformly distributed shearing stress, we get

patterns, it is suitable enough for static loads when the recomme .d (d)
F F F 0.707F
5s = — = FO = CI
design stresses are used. 3 A 2tL 2Lbcos 45° Lb
In the lap joint, we observe that the forces F are not collinear and t
The maximum shearing stress theory then gives
therefore the weld is subjected to bending. This fact may cause the desigi

©
§\271/2 F\2 2751/2
“co. PR Ga eae aaeet)
to be somewhat more conservative in his choice of design stress, or bet
try to design a joint with the weld metal symmetric with respect to .
2 2tL, 2h
perhaps a butt joint.
The situation is equally bad for the parallel weld, Fig. 19.3(d). I from which the required length of weld L for a particular design stress +
known that the stress at the end of the weld G where the applied load a may be found, or vice versa.
is much greater than the average, one experiment'!®-7! giving a ratio (b) Case 2, Fig. 19.5. There are a number of configurations that
about 1.9 to 1 (two 4-in. lengths of 0.35-in. weld). This ratio would would be analyzed in accordance with the principles used in this case.
expected to vary materially, being about unity for very short longitud: Conventionally, we consider the welds as lines (adjacent to the plate being
welds; also it is a function of the ratio of the sectional area of the plat welded), find the centroid C of the “weld lines,” replace the eccentric
the sectional area of the strap,“9-®! Fig. 19.3(d). With some proporti force F by a force F’ = F through the centroid and a couple Fe, Fig. 19.5,
the stress tends toward a minimum near the mid-length of weld;"9-8)
others, the stress at the end H, Fig. 19.3(d), is the minimum. However,
true peak stress, if the elastic limit has not any where been exceeded
localized at a reentry discontinuity, point P or Q. If a small part of
metal here is stressed beyond the yield strength, no harm would be d
for static or nearly static loads. Nevertheless, if the weld is other t FIGURE 19.5 Eccentric Load.
“short,” a higher design stress may be used for transverse loading than
parallel, Table AT 30. f

19.4 FILLET WELDS, ECCENTRIC LOADING. _ Since there |

many ways to impose eccentric loading, the cases that follow are sugges
only of ways to make stress analyses."9-20
and then compute a nominal stress that is intended to be the maximum
,
—a —s The Stress s, due to the moment Fe at any point B of the weld, Fig.
19.5 >
1S assumed to be proportional to its distance from C; that is, s./p
= s,1/p’
where Ss; is the maximum stress, which occurs at the maximum radius
i
point H. Thus, at B, the shearing force perpendicular to p is taken as 5 dA,
Pb

and the resisting moment of this force about C is ps, dA, where dA is
FIGURE 19.4
arbitrarily taken as an infinitesimal part of the throat area. Using s, =
e
 Pes re eS By sw
s19.5] E—ECCENTRICALLY LOADED FILLET WELD ‘Sil
FIGURE 19.5 (Repeated).
The odd distribution of weld metal in Fig. 19.5 was chosen deliberately
to be sure that the principles involved were not obscured by the simplicity
of a symmetric arrangement. If a connection were to be made similar to
Fig. 19.5, it would be more like Fig. 19.6, perhaps with additional weld on
Plate a the left-end short dimension—and probably with some modification to
reduce or evade a twisting of the plate. Reports of experimental justifications
of the various equations used in designing eccentrically loaded welded
pSs1/p’ and equating the applied moment Fe to the resisting moment, y joints are scarce. In one experiment,"®-”) which was limited and therefore
inconclusive, with welds as in Fig. 19.6 and two plates to avoid twisting,
5,
(f) Fe = [ps,dA = — [ p?dA, the indication was that the actual maximum stress is higher than that pre-
p dicted by equation (j). Both theory and experiment give mostly an increasing
where we note that fp? dA is a polar moment of inertia J, of an area y stress gradient from left to right, Fig. 19.6, but the experiment indicated
respect to C. Equation (f) may therefore be written 4 ‘a relatively higher stress at B than theory, for reasons that are unclear—
possibly some twisting augmenting the stress. Otherwise, the experiment

Fe =_ Sside Fep’ could be said to agree with the theory. If care were taken in actual design
(g) 7 or Ss sl = 7:E and fabrication to apply the load F so as not to twist the plate, then the
small twist that may actually exist could be allowed for during design by
where Jc is computed as follows. The moment of inertia of a long sler arbitrarily moving the center of moments away from the load by say 5 to
area with respect to a centroidal axis O, perpendicular to the are; 8% of the theoretical e; that is, use a new center C’ (not shown) at e’ =
closely J ~ AL?/12, where L is the length of the area A and th 1.05e to 1.08e for all calculations, which is a shift in the conservative direc-
(throat) dimension is small compared to L.“-§! Let A be the throat tion.
and apply the parallel axis theorem (J = J + Ad?), Fig. 19.5, to gi The configuration of Fig. 19.7 presents the same problem as far as the
polar moment of inertia about C. For each strip of weld, theory is concerned. The maximum Fep’/J, stress is at the point of maximum
2 As? p’, but one must be certain that the resultant shear 7 used is the maximum.
(h) Fess dot a? = ae Note in Fig. 19.7 that tr, and 7, are each less than T, by virtue of the direc-
tions of the vectors.
where r is the distance between the axis O of a weld line and the axis
which marks the centroid of all the weld lines. In equation (g), J sho
the moment of inertia of all the throat areas about C, which is the s 19.5 EXAMPLE—ECCENTRICALLY LOADED FILLET WELD. In
the moments of the individual strips of welding; all moments of ini Fig. 19.6, 4 = 6in., e =24in., L = 10 in. and the welds are } in. (leg dimension).
are positive. q What would be a safe variable load F?
In addition to the stress from equation (g), which is in a directionp
pendicular to the radius p’, Fig. 19.5, a uniform shear stress is assumed
induced by the force F’ = F; [oF

co 6
i)
(i Sso
z
A > FIGURE 19.6 Observe that the larger h, the greater XX a\\
the resistance of a particular length of weld. TN ge.
\ | 2 F T oN \
where A is taken as the total throat area and the direction of the stress
downward. The stresses s,, and s,. add vectorially, or from the cosine le Q M q
Fig. 19.5, Co

) T = (s.,7 + Ss2” + 2551559 COS 9)1/?,

which is taken as the maximum shearing stress. The foregoing analysis!
not only approximate, but it presumes that there is no tendency for f
plate, Fig. 19.5, to twist. FIGURE 19.7
 iF, ”

Veer

5, is the maximum stress (at top). Substitute this value of s into the expres-
FIGURE 19.6 (Repeated). sion for dF above and get
dF = s,(sin 6)tr d0.

Multiply both sides by r sin @ to get the moment dM; then

Solution. This connection is neither parallel nor transversely loaded. (k) M = [dM = { (dF\(r sin ®) = s,tr? f sin?0 dé,
decide to use the Jennings recommendation in Table AT 30 for fillet w 22
(sa = 14) and a strength reduction factor for stress concentration of Ky = M = str? | sin? 6 d@ = s,tr?z,
because the load is varying; a safe calculated stress is then 14/1.4 = 10 ( 0
§ 19.8.) With t = 0.7076 = (0.707)(0.25) = 0.177 and A =7¢L for one
we find from (h) 4M 4M 5.66M
4 *1 TD? (bcos 45°)D?_—szb DD?”
Je = 2 + (5) | = 2S Doo” + (0.177)(103)"] = 61.4 int where s, may be taken as a normal stress. If the moment M is produced
The centroid C of the weld lines is located by symmetry; the maximum p’ is by an eccentric load, the stresses may be combined after the fashion of
2 2471/2 i § 19.4(a). (The weld may be considered as a line; that is, the moment of
ve =[(5) + G)] = 5 + 92 = s.84in if jnertia or section modulus of the line that represents the weld may be found
and used in the basic equations, with the throat dimension introduced
The stress ss1 at the weld ends due to the couple whose moment is Fe a nd later; helpful for repetitious calculations.“°-*))
uniform stress ss2 are i
_ Fep’ _ F(24)(5.84) _
Os age SAS 2.28F.
19.7 DESIGN STRESSES. The design of many structures such as
AAU ee chat cake ea and bridges and building frameworks are covered by codes, which are commonly
a legal requirement. Although the machine designer is not bound by these
Considering the directions of these stresses at the corners H, B, M, Q, Fig. 1 specifications, their permissible stresses, and other -specified practices
we conclude that they combine to give the maximum resultant at B (or
obtainable from the codes themselves, constitute evidence for design that
that equation (j) applies, with cos 9 = L/(2p’) = 5/5.84 = 0.856;
need not be ignored. Table AT 30 gives excerpts that may be helpful when
r= (sa3" + ‘Ssa* + 2551532 COS 9)1/2
“regular” welding rod (as 60XX) and base metal of less than 0.3% C are
= [(2.28F)? + (0.282F)? + (2)(2.28)(0.282)F2(0.856)]?/2 used; the welding rods have some 0.15% C."°-4! The quenching action of
or 7 = 2.53F = 10 ksi (10 ksi = the allowable stress); then F = 10/2.53 the cold base metal tends to harden the steel and will actually do so if the
3.96 kips or 3960 Ib. of carbon content is too high. For alloy steels, if hardening and the resulting
brittle zone are to be avoided, the carbon should be less than 0.2 %, unless
preheating and/or tempering after welding are employed. When a high-
19.6 ANNULAR FILLET WELD IN BENDING. One more cé Strength rod, as E 10016, (s, = 100 ksi), is used on alloy steel (it would not
will throw additional light on the method of using differential area:
Otherwise be used), a tempering at about 1150°F, usually called stress
the annular weld of Fig. 19.8 be subjected to a moment M only; r d@
relieving in this context, is necessary to obtain the benefits from the alloy.
differential length; then dA = tr d0, when the throat area is the ba
Higher carbon steels and many alloys need to be preheated (200-800°F,
the calculations. The force dF resisting the bending at this point is n depending on carbon content and thickness), in order to avoid a rapid
to the page; dF = sdA = str dé. If this stress s is assumed to be pr cooling that hardens the material in a zone that therefore becomes brittle.
tional to its distance from the neutral plane (horizontal center line on €1
Moreover, the rapid cooling and extreme temperature range of the molten
view), which is r sin 8, we can write s,/r = s/(r sin @) or s=s, sin 0, whe
Metal next to the cold base metal may result in a cracking of the weld, which
is eliminated by preheat. Manganese and silicon adversely affect the weld
rd0 and an equivalent carbon content of C + Mn/4 + Si/4 is a substitute
FIGURE 19.8
Criterion for the carbon content alone;!°2! 'C, Mn, Si, being respectively
the percentages by mass of each.
Ts ee e est eee ces ae) eee ae arma Ser an me
ere

Most electric arc welds are made with coated electrodes (cons With the butt weld flush with the plate, the strength of the joint: in fatigue
used for coatings: cellulose, feldspar, mica, potassium silicate, ca js increased. Jennings!9-1] suggests the following strength reduction factors:
carbonate, and others, in various combinations), whose function is to fo
a protective environment about the molten metal during welding, t Reinforced butt weld, Ky = 1.2 Toe of transverse fillet weld, Ky = 1.5
keeping oxygen and nitrogen away (see § 19.11). The steel electrodes h J-weld, sharp corners, Ky = 2.0 End of longitudinal fillet weld, Ky = 2.7
code numbers (by ASTM and AWS) such as E 6010, where the first tw
three) digits (60) indicate a minimum tensile strength (ksi) and the las There is a certain uncertainty attached to these numbers. Consider values
digits in effect specify other variables such as the coating, current s
position of the weld, etc. Unshielded arc welding, done with bare rod
nothing to protect the molten metal from air, results in a penetrati Or TABLE 19.1 SELECTED FATIGUE DATA
Oz and Nz that seriously impairs the quality of the weld. trengths in ksi, average values.
; a Ww. M. Wilson.!°-21 (b) Standard cylindrical test specimens, cut transverse of butt weld; E 10015
There are over 12 rods in the E 60XX series; 6 or more in the E
(Ni, Mo, V) electrode; rotating beam fatigue. (c) Weld metal as welded. (d) Tempered at 1150°F, s, at
series ;‘°-"") numerous others for special purposes as for gas welding, b 2 x 107. (e) Annealed at 1650°F. (f) Double-V butt, $-in. plate, axial load. The fatigue strength of welded
welding nonferrous metals and cast iron, corrosion resistance, unus aluminum is in about the same range for all the alloys used. (g) From Ref. (19.6); failure of weld metal.
high-strength, low-alloy-steel rod (as E 10016, a Ni, Mo, V alloy BASE
If the higher-strength electrodes are used, the designer must decide ¢ METAL ENDURANCE STRENGTH Sj, ksi
design
& stress with less help.PAO In general, it would seem that a design
sign facesfacto MATERIAL AND ad R=0
(m) N = 3.75 on ultimate strength Re Os
of electrode, steady load 4 JOINT Hite dane Ws orasehes te Novak eves x.
should be adequate. A lower value may be all right if the highest qua 2x10" 10° | 2x10? 210° 10°) 2x10" 2x.10° 10°
weld is certain. The transverse static strengths of inert-gas fillet welds BUTT JOINTS
certain aluminum alloys are:"9-75) for 3003, 16 ksi; for 5086, 37k }-in. plate, carbon
for 5154, 33 ksi. of steel(a) . «| 61.5 34.2
a With bead, as
J welded. "3 14.5 22.1 22085 36.8 53.3
19.8 DESIGN FOR FATIGUE. If a weld is subjected to vary
° : 5 Ditto, t
2eOrE ed
ah 143 21.4 ;
23.3 31.9 38.4 55
loading, the natural discontinuities, surface roughness, and residual te Bead mch. off,
stresses will encourage the cautious designer to look for points of maxim no, Hite. 17.6 28.7 28.7 48.8 44
stress and make checks as possible for fatigue strength—using the princif Bead off, 1200°F
i
previously i
explained i Chapter 4). Factors that unfavorably
(mostly in temper . 16.6 28 . 27.7 49.3 42.5
the fatigue strength of a weld include: lack of penetration of the weld m Alloy steel (b) . — . | 106(c) 97(c)
shrinkage cracks, slag and other inclusions, porosity and gas pockt Aswelded .. 58 78
undercutting (unfilled groove melted in base metal); all of which can Stress relieved (d) . 66 = 86
corrected by proper welding techniques and materials. wanes A cs a ise
As the molten metal solidifies and cools, residual stresses are ram De on tes
and their magnitude cannot be predicted with any degree of certain Al 5083-H113 (f)
The rule that the last part to cool has tensile residuals is significant. T As welded . «| 47.3 26.2 8 9.2 18 13.2 = 15.7
leads us to conclude that the weld metal in a butt joint as-welded has a la Bead removed. 16.2 16.6 24.5
residual tensile stress, in a complex pattern (balanced elsewhere by resid!
compressive stresses, since the internal forces are in equilibrium), and t FILLET WELDS
is found to be true. If a welded part is to be machined with close tolera ct
it should be stress relieved by normalizing or annealing before machini n Carbon steel (a)
otherwise the deformations as stressed material is removed will spoil th Single pass fea as et awe Pe
dimensions. In structural work, removal of the reinforcement on a bt Singeeaue ail 18 452 20.6 275 37* ;
weld would be rare indeed, but it is at least less rare in welds for machine a se
3 ASTM-AT steel (g) . | 60-72 33
Plug weld. ‘ 6S, 13 12:6" 23.4

* Very wide spread, 28-45 ksi.

516 DESIGN OF WELDED JOINTS [Ch, 19.8] DESIGN FOR FATIGUE 517

of K; that can be determined from Table 19.1. Computing the values of } following observations.(19-19] Analysis of the origination and propagation
for the first material in Table 19.1, the Z-in. plate, we find a range of K; fro of fatigue cracks showed that the prime factor causing these failures was
1.16 to 1.31; for the aluminum alloy 5083-H 113, Table 19.1, we ff the notch effect. The notch effect of the double-V weld was greater than
Ky values of 1.36 to 2.02. Not only are some of these values significan that for the single-V weld. (Since the metal in the big end of the V is the
different from 1.2, but they vary in different ways. For the steel, Ky seer last to solidify and cool, one would expect a very unbalanced residual
to increase as the number of cycles decreases, while for the aluminu m, stress pattern when the base plate is restrained from distorting.) Manual
definitely decreases as the number of cycles decreases (the normal expec welding resulted in fatigue strengths lower than those for automatic welding.
tion); moreover, these tests indicate that the aluminum is much me Porosity, oxide inclusions, and poor penetration significantly reduced
affected by the discontinuity of the bead (more notch sensitive—higher K fatigue strength; 79 % of the broken specimens had defects of which about
On the other hand, there are some tests that indicate little improvement half were revealed by X ray.
fatigue with the removal of the discontinuity. If the bead is ground off,| One must be careful about attachments welded to main bodies. For
grinding leaves a residual tensile stress on the surface that would be¢ example,(°-] a plain carbon plate, § x 14in., had a fatigue strength
pected to result in a fatigue strength lower than if the residual were f (2 x 10° cycles) of 22.8 ksi; with a piece attached on one side with trans-
present. Also the surface left exposed may have a significant tensile residt verse fillet welds (to form T), the fatigue strength of the plate was 18.9 ksi;
stress left from the progressive solidification of the weld metal. We re with pieces on each side, opposite each other, +5-in. transverse fillet welds,
that peening, which leaves a surface compressive stress, has no effect o the fatigue strength dropped to 13.1 ksi. The same experiment run with low
fatigue strength of the joint,{9-6] usually, no doubt, with the weld as weld alloy steel (sy = 78.9), no change in configurations, resulted in fatigue
in mind. Nevertheless, it is reasonable to expect that if shot peeni strengths of 26.4, 23.9, 10.1 ksi, respectively. While the principal cause of
applied to the ground surface, it would be possible to make the join fatigue-strength reduction is probably the discontinuity, the residual
tually as strong as the parent metal—if the welding rod and parent compressive stresses that balance the tensile residuals in the weld may, by
are of about the same material. Tests would show. The peening as superposition, seriously weaken a column. (NOTE. The evidence is not all in,
monly practiced on welds is hammer peening that often results in erat and there is not a unanimity of opinion concerning the effect of residual
and therefore stress-concentration points in the weld. Hammer peenit stresses for structural welds.) Besukladov et a/.[*-28] on research related to
the first or last layers of weld has been demonstrated to be harmfu ship fatigue failures, concluded that the usual steels used for hulls all had a
pressure vessel welds and is prohibited by code. If the expense of grindi fatigue of plate in the vicinity of a weld of about 10.7 ksi (te = 5 x 108),
is warranted, the additional cost of shot peening would be minor. and that the sme curve that bounded the lower limit of all the failure points
matter of principle, welds should be located as remote as possible fr is defined by
points of large bending moment. Heat treated alloy steels would natura
need special treatments. (n) n= 115% 10°(—_ = 0.054) ;
Since some weld metals become brittle after an annealing operatic s — 9.95
this method must be used with discretion as a means of reducing tens where Ne is the cycles of life to be expected for a stress of s ksi.
residuals. Also in this connection, observe the results in Table 19.1 fo Since the residual tensile stress in the direction of the weld length is of
alloy steel; tempering at 1150°F improves the fatigue strength, presumab the order of magnitude of the yield strength, another maneuver of stress
by releasing most of the residual tensile stresses without a signi relieving is sometimes practiced. A band of the base metal on each side of
reduction of mechanical properties, whereas annealing at 1650°F redu the weld is heated to some 350—400°F, by oxyacetylene torches or electrical
the fatigue strength to less than the as-welded value. This experiment @ resistance, causing the bands to tend to expand (and be in compression by
indicated that water quenching and tempering was not as effective as t the action of the adjacent unheated parts). The force of this deformation
ordinary stress-relieving heating. Available experiments suggest that st clongates the length of the weld, which, already being at yield, undergoes
relieving plain carbon steel (< 0.25% C) does not improve fatigue stre little or no additional stress, but is plastically deformed. Therefore, when
Yet temperature stress-relieving of welded pressure vessels is recommend the heated bands have cooled back to ambient temperature, the weld,
practice (required for some materials), and is important in restorl having been stretched, returns more or less elastically to a lower residual
ductility. Welding a strap across a good butt weld invariably lowers| tensile stress in the longitudinal direction; and there is evidence of a reduc-
fatigue strength. tion in the transverse residual tension also.{19-8] This process is used on
From the experiment on the aluminum alloy of Table 19.1, we have th Seams in large plates, as on ships, and on girth seams of pressure vessels.
 “we we
ESIGN OP WELMED JOINTS If
19.9 OTHER TYPES OF WELDS. The configurations in w
welding may appear are quite varied. The following are worth not
A corner joint, Fig 19.9(a), may have weld metal placed on either
inside, outside, or both. It is much cheaper to bend a plate to form a cor
but if welding is necessary, see Ref. (19.22).
4

:
t=(1.35)(a)
eC FIGURE 19.9
FIGURE 19.10 Automatic Welding Head. (Courtesy
General Electric Corp., Schenectady).

(a) Corner Joint (b) Edge Joint

An edge joint, Fig. 19.9(b), is made along the edges of two or m

parallel, or nearly parallel, plates.
If one plate with holes in it lies on a second plate, a plug weld is ob: ai resistance welding, or induction welding. Arc welding is done either with a
when the holes are filled or partially filled by weld metal fusing with carbon electrode on steel (tungsten electrode on non-ferrous) or with a metal
second plate. 4 electrode. The metal electrode supplies the filler metal as it melts and may be
Spot welds are resistance welds (§ 19.11), usually round, in the fed in manually or automatically, Fig. 19.10. When a carbon or tungsten
form as the electrodes that press the sheets together; used on thin mate electrode is used, a separate rod supplies the filler metal, either manually or
only; see Ref. (19.23). automatically.
An intermittent weld consists of short lengths (2-3 in. long) of we In addition to coated rods, inert gases, usually argon or helium or a
with space between, as 6 in. on centers. The minimum length should be mixture of both, are used to shield molten metal from atmospheric oxygen
least 4b and not less than 1 in. This plan saves expensive weld metal wl and nitrogen (by displacement). Inert-gas methods are very common for
a continuous weld is not needed. ii stainless steel and nonferrous metals; for example, aluminum and titanium,
A tack weld is an intermittent weld, lightly done to hold the memb with a nonconsumable tungsten electrode and separate filler rods as appro-
in position for assembly purposes or for the principal welding. priate.
Electroslag welding,®-121 a method of welding rather than a type In submerged arc welding, Fig. 19.11, the arc is covered with a welding
weld, is applicable to quite heavy sections; to be investigated by composition, and bare electrode wire is fed automatically. This process is
concerned with such a welding problem. excellent for automatic welding of flat welds (also girth seams) and is used
extensively on pressure vessels. Its high speed recommends it for production
jobs.
19.10 MINIMUM FILLET-WELD SIZE. If the leg dimension of
fillet weld is doubled, the amount of weld metal is increased by four ti mi FIGURE 19.11 Submerged Welding (Unionmelt). (Courtesy Linde Air Products Co.,
which suggests that a light continuous weld is preferred to a heavier inte New York).
mittent. However, even though strength requirements might permit a ve
small fillet weld, experience suggests the following minimum sizes accordi
AG. OR 0.C.+-ELECTRODE*
to plate thickness: ¢- WELDING ELECTRODE
GROUND SOLIDIFIED WELDING
COMPOSITION -WELDING GOMPOSITION
Plate thick., in. 1-3 te 8B #1 14-13 14 up OR Lue TUBE
Min. din. . é $ ie t 3 2 2 in. VEE IF REQUIRED

19.11 TYPES OF WELDING PROCESSES. Only a few of the mo

common processes will be mentioned. Electric welding may be arc welding

PLATE METAL
WELD METAL” = GRANULATED WELDING
COMPOSITIO
 319.14] CLOSURE 521
520 DESIGN OF WELDED JOINTS [Ci

19.13 OTHER METHODS OF JOINING METALS. Ordinary

In the atomic-hydrogen process, the energy from the arc is used soldering, using a mixture of lead and tin, which is the cheapest, is applied
break the hydrogen molecules into atoms rather than to melt the 1€ only where strength is unimportant. Whereas soldering is done on cold
Then the hydrogen recombines to form molecules, releasing energy at h parts, brazing is done with parts heated to a temperature above the melting
temperature (7200°F), and it is this released energy that fuses the me point of the nonferrous filler material. Silver soldering is a process of this
The hydrogen also provides the shield to avert porosity and brittleness rr type, used where the surfaces to be joined are brought into intimate contact
nitrogen and oxygen. This process is used on a wide variety of met with the silver alloy on the fluxed surfaces. Heating the base metal until the
manually and automatically. silver alloy melts and then allowing it to solidify results in a strong connec-
The welds mentioned above are called fusion welds because the me tion, especially in joining copper surfaces. Heating may be done in a furnace,
are joined by fusion. Resistance welding depends upon the resistance with an acetylene flame, by induction, etc. Alloys other than silver alloys
the flow of electricity at the points to be joined. In spot welding, two @ are used in this manner. In another method of silver soldering, the surfaces
trodes press the sheets of metal together, and at the spot where the pres: to be joined are placed close together and heated in the presence of flux and
is exerted, the resistance to the flow of current causes a heating h silver alloy, which enter the space between the surfaces by capillary action.
together with the pressure, results in a weld. If two copper rollers are u In braze welding, nonferrous filler metal is melted into holes, grooves,
as electrodes and if two sheets of metal are passed between the roller or as a fillet on the base metal; the base metal is not melted. This process
seam is welded where the rollers press the sheets together, in a resista! is often used for joining cast-iron parts, and is widely used in repair work.
welding process called seam welding. Plastic adhesives are also available for joining metal parts.
In gas welding, a hot flame and a metal rod are used. The oxyacetyl
process uses acetylene burned in oxygen. The flame heats the parts’
joined, maintaining a puddle of molten metal at the junction of th 19.14 CLOSURE. For weldment designs, there are in print many
and filler metal is added to form the joint. Gas welding is particularly s practical suggestions that represent accumulated experience, which should
to welding thin metal and is applied to many different metals. be consulted. For example, a liberal use of bent members and rolled
sections, such as angles and X shapes, is advised. Welding metal in place
costs considerably more per pound than structural steel.
Little has been said or implied about the human element in welding.
19.12 TESTING WELDED JOINTS. When practicable, tests It has been noted that manually applied welds do not average as strong as
welded joints should be made to ascertain the quality of the weld. In im} machine-laid welds, and the variability is greater too. Internal flaws, if any,
tant applications, such as pressure vessels, careful tests are made. Test are as likely to be sources of fatigue failures as the known points of stress
are welded at the same time that the joint is welded. From the concentration; hence, if there is any question concerning the quality of the
plates, specimens for tensile tests, for a bending test, and for determi weld, allowance should be made. Eliminating “‘stress raisers” by persistent
the density are made. Some of the requirements of a weld on Cla design effort and manufacturing care is probably more rewarding in in-
pressure vessels will be suggestive. The tensile strength of the joint creasing fatigue strength than using higher strength materials.
the deposited weld metal must be at least equal to the minimum
strength expected of the plate material. The elongation of the depos
weld metal is to be a minimum of 20% in 2 in. In the bend test, the surfi
are machined flush with the plate, and the specimen is bent cold un!
least elongation measured within or across approximately the entire W
on the outside fibers is 30%. To be judged satisfactory, the specimenB
show no cracks on the convex surface. The specific gravity of the ¥
metal shall be a minimum of 7.8.
Moreover, in Class J vessels, it is specified that every portion Of
longitudinal welded joints of the structure and at least 25% of the cir!
ferential joins be radiographed (X ray or gamma ray) in order to re
excessive porosity and points of defective fusion, if any. A number of ot
techniques of inspection have been developed."°-®!
 20. MISCELLANEOUS
PROBLEMS

20.1 INTRODUCTION. In this chapter, a few additional problems

that appear with some frequency are presented briefly. Additional details
on some of these topics, on some of the topics previously discussed, and
on some not mentioned at all are found in books on advanced strength of
materials and theory of elasticity.

20.2. THIN CYLINDRICAL SHELLS UNDER EXTERNAL PRES-

SURE. If the external pressure is large enough, the needed thickness of
shell will be so great that the thick-cylinder formulas apply as explained in
§ 8.26. When the needed thickness is small, the thin-shell equation, which
assumes uniform distribution of stress, § 1.25, accurately predicts the stress
Produced by the external pressure; the value of this compressive stress in
4 tangential direction is s, = pD/(2t); see equation (1.18). However, this
Stress for thin shells generally indicates nothing concerning the safety
against failure, because the failure is one of elastic instability, quite analo-
80us to long columns. Thus, there is a critical pressure that causes collapse,
given by Saunders and Windenburg!°-2) as

(20.1) Aes
2.60E(t/ D)5!
sl ae
e
L/D — 0.45(t/D)*?
for E psi; where the critical pressure P- is seen to depend on the unsupported

if fee): i i 5 Ataeeee ,
 MISCELLANEOUS PROBLEMS [Ch,
524
length L in. of a vessel whose diameter is D in. and thickness of shell fii
When collapse occurs, characteristic lobes or bulges are formed, the numb
depending on the ratios L/D and t/D. See Figs. 20.1 and 20.2. The low
the L/D and t/D ratios, the greater the number of lobes.
If the L/D ratio is above a certain value that is within the range¢
8 < L/D < 15, the collapse is a two-lobe failure, Fig. 20.1, and is i
dependent of the length. Economic employment of material is obtained1

lob FIGURE 20.3 Stiffeners for Tanks under External Pressure. This figure shows L the
FIGURE 20.1 Two-lobe Collapse. Typical of long tubes and cylinders; the
length of unsupported shell. The heads serve as stiffeners at the ends. The rivets used
to rivet
are not necessarily as regular as shown.
stiffening rings on the outside should have a diameter not less than the thickness of plate.
If a single row of rivets is used, as on an angle, the spacing or pitch should not be eae
using thin plate and short unsupported lengths, rather than a large ¢/, than 7t. If a double row of rivets is used, as on a tee or I section, the maximum pitch should

The short length is obtained by using circumferential stiffeners, the u

be 5t (¢ = plate thickness) and the rivets should be staggered. If the stiffening rings are
welded on the outside, the welding should be intermittent with the maximum spacing between
supported length L being the center distance between stiffening rinj welds of 8¢. If the stiffening rings are inside, they should be secure enough to maintain their
Figs. 20.2 and 20.3. The ASME Code"™-1°! specifies that the operat operating conditions. (From Rules for Constructi ae
Unfired position
proper under normal
Vessels Subjected to External Pressure). en
pressure p = p,/5, and in no case should the vessel be designed for a workii
pressure of less than 15 psi, corresponding to a minimum critical pressu
of 75 psi. Moreover, the computed maximum tangential stress should n unsupported length L from equation (20.1), using the value of t from calcu-
exceed s,/5. The following procedure may be used: (1) Calculate the pla lation (1). This value of L is the maximum spacing of stiffening rings if
thickness for stress from t = p,D/(2s,). For this thickness ¢, it takes appro? there is to be no collapse at a pressure p, = 5 p. For practical reasons, a
mately 5 times the working pressure to stress the plate to the yield stre closer spacing may be used; but if for any reason the spacing is increased
and this ¢ is taken as the minimum permissible value. (2) Compute above the value found in (2), then the thickness of plate must be increased
to a value ¢ that satisfies equation (20.1). The optimum economic design is
the one using the minimum permissible plate thickness f.
: Since the modulus of elasticity E is constant only to the proportional
limit, the Code contains a chart taking into account its subsequent variation.
The stiffening rings, which may be either inside or outside the vessel
Fig. 20.3, should probably be strong enough to support the load even if :
collapse occurs. The load that one ring supports depends on the length L
the smaller L, the lighter the section needed. The rings may be rolled shapes,
such as angles, tees, I-section, or rectangular sections whose properties be
found in various handbooks; they may be riveted or welded on. The
FIGURE 20.2 Thin Cylindrical Vessel with Clos moment of inertia J of the sectional area of a stiffener ring with respect to its
Spaced Internal Stiffening Rings. This shows the fort
centroidal axis parallel to the longitudinal axis of the vessel should be
tion of lobes between stiffening rings, the result ©
collapsing external pressure. If the lobes had fort
0.035.D®
entirely around the circumference, there would hay (a) T= ins
16, the theoretical number for the particular L/D ©
(Saunders and Windenburg!?°-?)).
sy a refinement of this computation, see Ref. (20.2). The amount of out-
Bes cate permissible is defined in the Code. The supports for the vessel
Should impose no concentrated loads on the shell, and, in horizontal
ee they should be placed at the heads or at stiffening rings.
re a quations for a number of other members subject
j to 0 elastic
elastic instability
i ili
 Seed EE Se ee ee ee eee ONE wae Fee Pe ee ee

half of the circular plate for analysis,"*-7! Fig. 20.4, we represent the forces
Continuing the analogy with the elastic instability of a column, we re acting on it as F, acting down through the centroid of the semicircular area,
that the strength of a very short column, L/k less than about 30, is depend and R, acting up through the centroid of the semicircle. For equilibrium,
on the yield strength of the material, a situation corresponding to a “th R = F,and the moment of the couple formed by R and Fis Fd, opposed by
cylinder,” § 8.26. The thin-cylinder discussion of § 20.2 is analogous to’ the resisting moment at the diametral section. This resisting moment is
Euler column. In between these extremes, as in the case of columns, th only approximately s//c because the stress is not uniformly distributed along
is an area where the failure depends on both the buckling strength and yj the section, but is some 25% higher than average at the center. Equating
strength. As for columns, this area for cylindrical members is handled the applied moment to the resisting moment, we get
largely empirical equations, and it is into this category that most commerc
tubes fall. The collapsing pressure for these tubes depends on the L/D 2r 4r sI 2rF sl
(d) F(= - =) =—, or —_ = —.
t/D ratios, the variations of wall thickness, the yield strength of
material, the elastic constants E and p, and the variations in roundness,"
Based on many tests, Stewart'°-5! proposed the following formula fork If the unit pressure on the surface is p psi, the load F on the semicircular
area is pmr?/2; at the diametral section, J = (2r)t8/12; c = t/2. Substituting
commercial lap-welded steel tubes, when D < 401,
these values into the above equation, we get
t\* 1007\3 r 2
(b) Pe = 50,200,000(—) = 50.2(—) psi,
(e) Cae ee Ai POR nT eS, = > p|-)) Pp psi,
where D is the outside tube diameter and ¢ the wall thickness. Accord
to charts by Jasper and Sullivan,°-*! one may conclude that if D 21 a reasonably good estimate of the average stress in a diametral section.
We have assumed bending in only one plane, an untrue assumption although
the thick-cylinder formula applies. It is worth noting that equation
warranted by experience.
of the same form as the theoretical equation for long thin-shell tubes,!2:2»
(b) Rectangular Plate Uniformly Loaded and Supported Along the
which is /
Edges. Experiments on square and rectangular plates supported along all
2Et® f four edges, uniformly loaded and made of ductile materials, indicate that
(c) Pe
~ (1 — pps
FIGURE 20.5 Rectangular Plate. The resultant reaction along
where y is Poisson’s ratio and the other symbols have the usual meanin: nly —|
each edge (Ri and Rz) is assumed to act up through the mid-
point of the edge. Since the total load on the half of the plate 1 a
under consideration is uniformly distributed, the resultant F we
20.4 FLAT PLATES. Loaded flat plates appear in innum acts through the centroid of the triangle ABC. Rup -f_ / re
configurations; Roark'-7) gives formulas for stresses and deflections esl Daan :
k==
70 cases without exhausting the possibilities, plus some tabular h (a2 + b?)1/2, B y

i
a proportionality between A and the R. 2 Up
solving the equations; see also Refs. (0.2, 20.6). Hence, the purp obtained by establishing
including this topic is to make the student aware of the problem and to g intercepts made by A on the diagonal AC.
leads to further help. q
(a) Circular Plate Supported at Edge with Uniformly Distribut
the significant stress is closely that obtained by an analysis similar to the
Load. A diametral section may be considered as the critical section. Taki
one above for the round plate (a is not very small compared to b). Proceeding
as before with help from Fig. 20.5, sum moments about the diagonal AC,
substitute the values of F, R,, and R, in terms of the uniform pressure
P psi, and get":7!

FIGURE 20.4 Circular Plate with Uniformly Distributed P

a’b?p :
(f) s=
2@ +
The total load on half the plate is uniformly distributed. Therf
the line of action
of the resultant
F is down through the centroid’
the area. The reaction of the edge is assumed to be uniformly di where s is the average stress along the diagonal (the stress at the corners A
tributed around the edge, and hence the resultant reaction R acts and C is some 10% more than the average). If b = a, the above equation
through the centroid of the semicircle. q applies to a square plate similarly loaded and supported.
+t
 528 MISCELLANEOUS PROBLEMS [Ch.?

weet
20.5 CAMS. Cams are among the most useful and important machin retical Curve
elements, especially in automatic machinery for the control of movemen (a) Parabolic (b) Harmonic (c) Cycloidal
of parts. The details of the design of the profile or shape of a cam al
FIGURE 20.7 Comparison of Actual vs. . Theoretical Accelerations. % (Court D. B.
covered in texts on kinematics or mechanism;"™*-?! there is at least of Mitchell (20.7) and E. I. du Pont de Nemours and Company). . a
comprehensive book on the subject.'?°1© Since it is assumed that the read
is familiar with this phase of the design, this article will concern principal
the contact stresses. of acceleration, which means that for this instant the rate of change of
The contact (Hertz) stress depends upon the force between the cam ai acceleration, or jerk, is quite high. Under these circumstances, it has been
follower, the shapes and sizes of the mating surfaces, and the modulii ¢ found'0-7.20-8) that the actual acceleration rises much higher than the
elasticity of the materials (also »). In turn, the force depends, sometim simple motion equations (Table 20.1) predict. Table 20.1 gives equations
predominantly, on the acceleration imparted to the follower systet for three popular motions conveniently in terms of cam angle @, which for
because the so-called inertia force (ma) varies with acceleration. T constant angular velocity of the cam, is proportional to time; @ = wr.
accelerations are determined by the design and manufacturing accuracy| Let the accelerations computed from these equations be the “theoretical
the cam profile and by the elasticities and clearances of the components. accelerations.” The comparative actual accelerations are shown in Fig. 20.7;
the system. Among the more common motions designed into cams are # the theoretical curves are visible with a hard look. Norte: that the actual
parabolic motion (a = constant), harmonic motion, cycloidal motion, a maximum values for parabolic and harmonic motions are about twice the
various polynomial motions. In any event, if an equation for the displace theoretical values because they are suddenly applied; that vibrations carry
ment can be written, then the usual differentiations can be made to obta over into the dwell periods; that the vibrations (and a and ma) are damped
velocity, acceleration, and jerk; v = dx/dr, a = dv[dr = d?x/|dr?, J so that the actual acceleration would eventually equal the theoretical if the
daldr = d°x/dr*. Jerk is seen to be the time rate of change of acceleratic duration were long enough. When the acceleration starts from zero at the
and it has a pronounced effect on the actual maximum acceleration. beginning of the motion, as for cycloidal, we see that the amplitude of the
The theoretical curves for the variation of the acceleration for certa vibrations is low enough that the actual acceleration is reasonably com-
popular motions are shown in Fig. 20.6. The curves OAB and CDE 2 parable to the theoretical. Let the index number 1 represent the theoretical
purely suggestive of the high acceleration that would be characteristi maximum acceleration for parabolic acceleration (a, = 1) then the theore-
a cam that was designed to start a follower at some constant speed; b tical maximum values for harmonic and cycloidal motions are a, = 1.23
since the change in speed is Av = Ja dr, which is proportional to the ar and a, = 1.57. However, for an undamped system, Mitchell'°-7! found the
actual accelerations to be in the ratio a,: a,: a, = 1:0.834:0.584; which
is to say that the actual maximum acceleration for the cycloidal motion in
A this case is only some 60% of that for constant acceleration, even though
P
1.27 theoretically it is the reverse relation that holds. With damping, the differ-
F Parabolic
1.00
0.81
ae i! FIGURE 20.6 Variation of Acceleration. / ences are not so great. In any event, the designer is justified, when the
M Harmonic celeration for the parabolic motion is constant inertia force is a significant part of the total load, in making adjustments
B CWE G, where it suddenly becomes negative at H of the theoretical acceleration on the basis of any available experimental
K Fig. 20.7 where this change occurs). Accelerat
data. In making the free body of a force system, there is some advantage in
Cycloidal aie for the harmonic motion, F JK/, starts at its
mum positive value and ends at its ma using the reversed effective force F,, if we say F = ma, then F,, = —ma
H I
Q negative value, J. where F,, is in a sense opposite to that of a. Lacking better evidence for
D Computing the contact stresses, multiply the reversed effective forces by a
B =~ factor q (F,. = —qma) to account for actual accelerations as follows:
6=aT Parabolic motion, F,, = —3ma,; harmonic motion, F,, = —2ma,, following
2 dwell; cycloidal and other zero acceleration starts following a dwell,
ae = —1.1ma,; where a,, Tp, and a, are the computed theoretical values.
curves is ni
under the curves of Fig. 20.6, the follower speed for these ee also Ref. (20.10). See Figs. 20.6 and 20.7 for locations of maximum a.
the follower is
nearly so large as for the others shown. Assuming that F ora particular cam, the numerical values of the accelerations increase with
in Fig. 20.6, we observe that fo
rest at the beginning of the accelerations increases in cam angular velocity.
, the motion starts with a finite valu
the parabolic and harmonic motions
 TABLE 20.1 MOTION EQUATIONS FOR CAM FOLLOWERS
Symbols: L = total movement of the follower in one stroke; x = follower displacement at any pq
of the movement; @ rad. = angular displacement of the cam during any follower movement x; B
angular displacement of the cam during the total follower movement L; v = speed of the foll
a = acceleration of follower;j = jerk of follower; = angular velocity of the cam; suggested con: FIGURE 20.8 Forces on Follower. A and B are normal forces at
units, inches and seconds, guides, fA and fB are the corresponding frictional forces; Fy =
For the sinusoidal curves, the maximum values occur when the trigonometric function is w i force of gravity (weight); @ = external load, which may not be
When motion starts from dwell with a finite acceleration, the jerk is theoretically infinite for an instan constant; F; = spring force; Fre = reversed effective (inertia)
a fact not revealed by the jerk equations. force with a factor g shown to care for actual acceleration when a
is the theoretical acceleration; P = normal force on contact sur-
MOTION
faces; friction at G and H assumed to have neglible effect on N,
Harmonic Cycloidal as does fA and fB for low friction coefficient. Curvature r-2: =
EQUATION Parabolic*
ToT.
FOR Limit: 6 = B/2

5 2L ‘GL 7 6 ie 3
Displacement . x= a” x= $(1 — cos5%) x= ee - 35m aa

Speed. 01. ei ee
4Lw
fon Lrw
ae 7
s=(1-
Lw

P FD 2p" B A
2w\? L{mw\? 7 se 2nLw? si
. = 1(=*) z= 3) —é
Acceleration
B 2\ 8 B B?
L [7w\? 7 4n?Lw? final point on the follower mechanism may have a motion quite different
Jerk . je 3 #=0 z= -() in —0 x = —_
2\B B p? from that of the contact surface, especially at high accelerations, because
of the lag in the transmission of the force due to deformations. This
* The limit of 6 = B/2 applies when it is assumed that the deceleration has the same numerical phenomenon can be allowed for,'°-1°) at least on a theoretical basis.
as the positive acceleration. The simplest way to get the values for the last half of this motion is A free-body analysis of the follower system is necessary for determining
use the same equations in a reversed image.
the contact force, and since there are innumerable systems, the only
possible general advice is to follow the principles of mechanics; Fig. 20.8
shows one arrangement. From an acceleration diagram,"*-2] one may
Do not conclude that the designer should design all cams to provi
compute the inertia forces at enough points to plot an inertia force F,,
an initial acceleration of zero. In the first place, instantaneous change: diagram; then the spring force F, = k8 (k = scale of the spring and
acceleration introduce difficulties only as the speeds become relativ 5 = deflection of the spring from its free length) can be plotted; the force
high (which is of course often desired in order to increase the output of of gravity F, is constant, as is perhaps load Q; finally, the contact force P,
machine) and in the second place, such cams are much more expens}
Fig. 20.8, perpendicular to the profiles, is determined at key points where
to manufacture because the cam curve changes so gradually at the beginni the contact stress may be a maximum. The experimental results are related
and end that only precision manufacturing methods would appro to the Hertz contact-stress equation, and by using a wear factor K, (as in
such curves satisfactorily. Thus, one uses zero acceleration starts only Wi gearing), the corresponding total force P is given by
there is a worthwhile advantage.
The load between the cam and follower arises from: the force of g Kb
(unless the relevant motion is horizontal), the spring force (if a spr
(20.2) = ——_————_lb.,
N(i/r, + 1/rg)
used), frictional forces, inertia forces (which may or may not invol
acceleration of a connected body), vibratory forces (arising from operat where r, in. = radius of the follower roller (r; = oo for flat face follower);
near resonant conditions, the periodic changes of force F = ma, imp ac rg in. is the radius of curvature of the cam surface at the point of contact
and instantaneously infinite jerk as described above). If a roller operates (use the positive sign if the surface is convex, negative sign if its is concave):
a groove (and for other positive-motion drives), impact occurs when® b in. is the effective face width (length of contact); N = design factor;
resultant force reverses direction; the magnitude of the impact can” K, is the wear-load factor (called a Joad-stress factor elsewhere—surface
fatigue as in gearing) for cylindrical rollers, but used for the design of
controlled to some extent by control of clearances, but clearances@
necessary. If a follower operates on an open-face (plate) cam, it may|
cam systems. The relation between the roller K, and gear K, wear factors is
difficult or impossible at high speeds to keep the follower in continuo
contact with some cam curves, in which case, the immediately adjace (g) |age at >

point on the follower does not perform the motion designed into the cart
Furthermore, if the follower system is elastic (long links involved), where ¢, is the pressure angle for the gear teeth; equation (g) may be verified
 es eee ave [Gre seen. - > iP ae a> a ee Se ee) Se eee ee tae ee ee ae Ae Sard
Bee

Sources: Talbourdet,'!?-3] Buckingham,"?-3! Cram.{12-22) Hardness numbers are minimum

limits. K, values are average for 10° cycles, laboratory results. Copious lubrication with
§20.6] FLYWHEELS 533
mineral oil, approximately SAE 20. (a) Through hardened to Ro = 60. (b) Class 35.
(c) Phosphate coated. (d) Case thickness = 0.045 in. Values of K.3 with 300% sliding are:
(e) 390, (f) 2000, (g) 750. frictional moment at the bearing is greater than that at the surface, there
will be a large percentage of sliding and relatively rapid wear. If a journal
Kei K.2
bearing is used, Talbourdet™*-34! recommends a roller diameter of 3 times
MATERIALS Rolling 9% sliding
the bore diameter in order to obtain a moment of the contact frictional
Gray cast iron and same, class 20, BHN = 130(e) ._ - 1,300 1,050 force adequate to insure mostly rolling.
Ditto except class 30, austempered, BHN = 270(f) . 4,200 3,400 For a radial follower with roller, Fig. 20.8, the curvature r, of the cam
Nodular iron, 80-60-03 and same, BHN = 207(g) . .- 3,400 1,850
1,000 900
surface at any displacement x in. of the follower from its “lowest’’ position
Tool steel (a), and gray cast iron 20, BHN= 140...
Tool steel (a), and gray cast iron, BHN = 225(b) . . 2,300 2,100 is given by:!20-101
Tool steel (a), and austempered class 30, BHN = 255. 3,100 2,500
Tool steel (a), and SAE 4150, OQT to 270 BHN(c)._- 9,000 6,700 [(%> + x)? + (v,/w)?]°?
Ditto, Parco-Lubrite coated : . | 12,000 7,900 (i) ie =
Tool steel (a) and carburized 1020, case Re = 50 () . | 13,000 8,500 (ry + x)? + 2(v,/w)? — (ry + x)a;/w?
Tool steel (a) and SAE 4340 induction h’dn’d to Re = 50 . | 13,000 9,000
Tool steel (a) and SAE 65 phosphor bronze, BHN = 67 . 1,000 where r, in. = radius of the base circle, Fig. 20.8, v, ips = * = speed of
Tool steel (a) and laminated phenolic, gadeL . . 880 830 the follower at displacement x, w rad./sec. = angular speed of the cam,
SAE 39 cast Al, BHN = 60, and cast iron, OQT to 340 BHN 300 and a, ips? = # = acceleration of the follower at displacement x, and
a, is to be considered algebraic, negative when in the opposite sense to v,;
on the outward motion. Equation (i) may be used for estimation purposes
on other than radial roller followers.'2°-1°
by comparing the derivations of equations (13.8) and (20.2). Some value
Cam shafts, linkages, bearings, etc., may be designed in accordance
‘of K, obtained under laboratory conditions are given in Table 20.2, wher
with methods previously discussed.
it is observed that the load capacity decreases materially as sliding increase
because the friction significantly increases the contact stress. The percenta;
sliding in this table is defined by:"?-??}
20.6 FLYWHEELS. A flywheel is a rotating member that acts as a
v1 — v2 storage reservoir for energy when work is not “consumed” at as fast a
a a eat?
(h)
v2
rate as the power is supplied. When the work being done is greater than
the work input, the flywheel gives up some of its stored energy to supply
where v, is the peripheral speed (ips, say) of the driving roller and v, is th the deficiency. The kind of energy stored in the flywheel is kinetic energy,
for driven roller. We have found no values of K, for flat-face followers which for a rotating member is KE = Jw?/2, from mechanics; kinetic
which v, = 0, but it is seen that design values may be of the order of energy is all available, 100% convertible into work without friction.“1-27)
to 30% of those for rolling. For rollers expected to roll, it would be or Thus, to serve its purpose, the angular velocity of the flywheel necessarily
servative to use Kz values for 9% sliding with a minimum design fact varies, the amount of the fluctuation for a given variability of the power
of perhaps N = 1.15; those who prefer to live more dangerously may input and consumption being a function of its moment of inertia J.
satisfaction from the use of K,, values, especially when rolling-type bea Reciprocating prime movers are characterized by power pulses; in
are used for the roller, Fig. 20.9. If the alignment and deflections are not k Diesel engines, for example, most of the power to be delivered is generated
most favorable, use higher values of N with K, from Table 20.2. Shortly after the fuel ignites, but a properly designed flywheel keeps the
There are two frictional moments on the roller, the one from the fi i speed fluctuation within suitable limits, and makes it possible for the engine
tional force at the bearing and the one at the point of contact. If to deliver work at almost a constant rate. Also, many driven machines
require power in fluctuating amounts. The work to drive a reciprocating
compressor varies for much the same reasons as explained for reciprocating
prime movers. The variable consumption of power is obvious in many
other machines, as punches, shears, machine tools. In the case of machines,
FIGURE 20.9 Cam Follower Roller. An ex
ample of stock rollers—with a needle bez
such as punches, that have a high peak power requirement for a fraction of
Diameters from 0.5 in. to 4 in.; widths fi on a revolution, a smaller driving motor can be used, so small in fact that the
$in. to 2} in. (Courtesy McGill Mfg. flywheel may be the major source of immediate energy needed during the
Valparaiso, Ind.) working part of the cycle (§ 20.7). The prime mover then again stores
 534 MISCELLANEOUS PROBLEMS [Ch.; §20.7] EXAMPLE—FLYWHEEL RIM FOR PUNCH PRESS 535

energy during the remainder of the cycle by bringing the flywheel back to t the fluctuations of speed will be slightly less than if the total J were used.
original speed. a With this simplification, AKE = I(w,? — w,)/2.can be reduced as explained
The amount of speed fluctuation that is permissible is an enginee n next.
decision that depends upon the application. Companies making certait If the rim thickness ¢ is small compared with the diameter D of the
kinds of machinery have their experience as a guide, and the conventiong flywheel, the mean radius of the rim is nearly equal to the rim’s radius
way of specifying the fluctuation is by the coefficient of fluctuation C, of gyration k. Let v, be the average speed of a point at the radius k and Vs1
which is defined by and U2 the limiting speeds at this radius; then v, = rw = kw fps (for
8. = 32.2 fps”) and
(j) Cys ae aR ; mk? w W(Vs1 — Vsa)(Vs1 + Vg)
(k) AKE = (wy? — a9") = gyn tal) = *
28
where the subscript 1 designates the maximum and subscript 2 designate
oO

the minimum speed; the units of each term in a particular expression mj Is ft-lb. for units as defined. Using v,, — v,. = Cv, from equation (j) and
Vs = (Vg, + Vg9)/2 in (k), we get
32.2 AKE
TABLE 20.3
@ w aioe Ii,
COEFFICIENTS OF FLUCTUATION, Cyv,?
FLYWHEELS
which is the mass of flywheel rim required to care for a change of kinetic
Typical values, taken from various sources.
energy AKE ft-lb. with a coefficient of fluctuation C, at an average speed
DRIVEN MACHINE Cy v, fps.
D-C generators, direct drive 0.002
Computation of the variation of energy AKE is sometimes a lengthy
A-C generators, direct drive 0.0035 problem. For a reciprocating engine, a force diagram plotted for a revolution
Punching, shearing, pressing machines 0.05-0.1 showing the resulting tangential force on the crankpin, in which both fluid
Stamp mills, crushers . . . 0.2 pressures and inertia forces from the accelerations are accounted for, is
Reciprocating pumps, compresso 0.03-0.05 needed. This work can be shortened significantly if average values of the
Machine tools, looms, paper mills 0.025
Spinning mills, fine to coarse thread . 0.01-0.02
work-output fluctuation of a particular type of engine is known; some such
Geared drives 0.02 values are foundin the literature. Some other situations may be satisfactorily
estimated as illustrated in the next example. When a machine has a rotor
whose moment of inertia is significant, as perhaps anelectric motor, allowance
for this in deciding on the size of the flywheel is appropriate.
be the same and we shall use w rad./sec., m rpm, and », fps, the linear spee
of some chosen point on the flywheel; the speeds , n, v, in the denominato Ir
are the average values, as w = (w, + wg)/2. Values of C, in Table 20: 20.7 EXAMPLE—FLYWHEEL RIM FOR PUNCH PRESS. Determine the
may be used as a guide; also see mechanical engineering handbooks. 1 size and mass of the rim of a flywheel for a punching machine which is to punch
The change of kinetic energy of the flywheel as a positive number I a maximum of one j-in. hole in C1020 plate, normalized, }-in. thick. The mean
AKE = I(w,? — w,?)/2 ft-lb. for J slug-ft.? [(Ib-sec./ft.)(ft.”)]. Since 1 speed of the flywheel is to be 150 rpm during the punching.
mechanics we use the pound for force, we must use the mass in slugs fo Solution. Experiments show that the variation of the force exerted during
the punching operation will be similar to the solid surve of Fig. 20.10. The area
consistent units (with feet and seconds). The moment of inertia for any
included in this curve ABC, which represents the work done (j F, dx), is given
mass may be written J = mk?, in this case with respect to the axis of rotation approximately by the triangular area ADE. Thus, the energy to punch this hole
where k ft. is the radius of gyration, and m = w/g, slugs for a mass w Ib “ for an ultimate shearing stress of 54 ksi, Table AT 7, is
the standard acceleration g, + 32.2 fps? becomes also a conversion constant
32.2 lb./slug. Since a large proportion of the mass of a flywheel may be in it 1, 14, — (54,000)[x(3/4)(1/2)(1/2)) _ c
aft = st OT Sy ken 1325 ft-lb.
rim (but see § 20.9) and since the rim is farthest from the axis of rotation,
the contribution of the hub and arms to the moment of inertia may be on y In a machine of this kind, the motor is large enough to overcome frictional losses
some 5% to 10% of the whole /.%-®! On this account, engineers frequently and to bring the flywheel back to speed soon after each punching operation. Since
neglect the effect of the hub and arms, an approximation that means that the time to punch is usually short compared with the time required for a complete
 lA oy Ip
Work Area ~~ _“B §20.9] ROTATING DISKS 537
<— Distance
CO Paree Diagram FIGURE 20.10 rotating ring is often a satisfactory criterion."-2- V°l- ™ On this basis, take
a) Plate half of the rim as a free body, ignoring the arms, and note that the force
——

E system consists of F acting down on each side, and the centrifugal force
t
S = mo,?/7, where 6, fps (for g, = 32.2 fps?) is the velocity of the center of
cycle, the work supplied by the motor during punching will be small, and itmigh t gravity of the body, 7 ft. is the radius to the c.g., and m = w/32.2 slugs
be assumed that all of the energy for punching comes from the loss of i (and w Ib.) is the mass of the half rim. If the rim thickness t, Fig. 20.11, is
energy of the flywheel; that is, AKE = 1325 ft-lb. In terms of the dimensions 0 : small enough, the center of gravity of a half rim may be taken to be in the
the flywheel, any number of solutions is possible. Suppose space limitations are same position as the centroid of a semicircular arc;"!-61 7 = D/a from the
such that a flywheel with a mean rim diameter of 42 in. (k * 21 in.) can be used. center. Thus, 0, = 7w = Du/w and S = wD*w?/(ag,D) = wDw?/(zg,).
Then From the sum of the vertical forces, the force F is F = 4wDw?/(zg,).
The mass of half the rim w = Vp = Azn(D/2)p, where p Ib/ft.? is the density
v; = 7Dn, = =(2)(Z) = 27.5 fps.
of the material. Then from s = F/A, we get the average stress as
Using Cy = 0.1 from Table 20.3 and equation (1), we get An(D/2)pDw? D?wp 7,2 v,2p
oAKE (32.2)(1325) _ (m) s= = = psf or psi,
Cpads| [email protected])* 2ng,A 4g, 80 144g,
ich is the mass required for a 42-in. mean diameter. The dimensions sho 1 where 0, = Dw/2 fps is taken as the velocity of a point on the mean
nila Baca Assume the width of rim 5 as 5 in. (If the flywheel is us < circumference of the rim.
as a pulley, this width may be governed by the width of belt needed.) The appro c Rules of thumb from experience specify the conventional limits of
mate volume of the material in the rim is V = 7Dbt = 7(42)(5)t cu. in.; he operation; for examples, 6000 fpm for cast iron and 10,000 fpm for cast
corresponding mass is pV =0.284V, where the density p = 0.284 Ib/cu. in. is taki a steel are often given as limiting speeds for these materials. For the cast-iron
from Table AT 7. Equating the mass of the rim obtained from its dimensions to limit speed, the hoop stress from (1) is 5 = (1007)(0.26)(12)/g, = 970 psi,
the mass required for energy purposes, we have which appears to be unreasonably low. Yet if the casting is not stress
w = (0.284)(7)(42)(5)t = 565, relieved, there are undoubtedly severe tensile residual stresses, and some
from which the thickness of the rim ¢ = 3.02 in.; use 3 in. The outside diameter unbalance is sure to exist; therefore, while much higher limiting speeds are
of the flywheel is then 42 + 3 = 45 in. often used, one should observe the foregoing restrictions unless it is known
that higher speeds are safe in a particular situation. If the flywheel is made
in halves, the strength of the connections for the rim sections is or may be
20.8 STRESSES IN FLYWHEEL RIMS. The centrifugal force that the determinant of the safe operating speed.
is produced by the rim of a flywheel may be large at high speeds, and the e- If the flywheel acts as a sheave or pulley, the bending moment on all
fore its effect should be checked unless it is known to be negligible. The the arms [(total torque)/(no. of arms) on each] is approximately equal to
deformation of the rim under the action of this load, Fig. 20. 11, results in a the torque transmitted to or from the shaft; hence, the torque on the shaft
tensile load on the arms and bending moments at the sections where the is sometimes used as the bending moment on the arms for design purposes
arms join the rim; but the stress calculated on the assumption of a free whether or not the flywheel is serving as a pulley. However, if there are
momentary high accelerations (as from a sudden stop), the inertia of the
BAS tim results in a tangential reversed effective force of approximately ma,
b
mo AS SAC where m slugs is the mass of the rim and a fps? is the tangential acceleration,
which may induce a greater bending moment than the torque. Also, see
~ i Gi,
handbooks for empirical approaches.

We
. Als a
FIGURE 20.11 Flywheel. The size
\i/ 20.9 ROTATING DISKS.
F a Eh of a flywheel is its outside diameter Rotating disks or cylinders are common
elements of machines, and are used as flywheels. The moment of inertia
9 -— +} j—-—*-- in inches. The dimension D shown is
used in computations. Centrifugal of a hollow cylinder about its axis is J = m(r,? + r;*)/2, where r, is the
1X
Fa \ force tends to deform the rim some- Outside radius, r; the inside radius, Fig. 20.12, and m is the mass of the
ify \ what as exaggerated ABC. hollow cylinder (slugs for r ft. and g, fps?); then the kinetic energy of the
 §20.10] CLOSURE 539

& FIGURE 20.12 If the rotating member is subjected to forces other than centrifugal, the
stresses induced by these other forces are superposed on those described
above. For example, the stress at the inner perimeter arising from a force
fit is computed as explained in §§ 8.26 and 8.27.

20.10 CLOSURE. Since there are innumerable special problems in

disk is Jw?/2. If the shaft on which the disk or cylinder is mounted ro at machine design, this book could be extended indefinitely. We hope that
it would certainly be satisfactory to let r,; = 0, and perhaps includet answers of one kind or another can be found herein for the most common
moment of inertia of other attached bodies. engineering questions related to the application of the science of mechanics
to machine design. A number of related topics, such as vibration and its
stress at the perimeter of the hole, given by®:7-1:2! control, corrosion, and others, are so involved that a brief discussion may
w
be more misleading than informative; also true of some items that have
(n) Stmax = a + pr? + (1 — wr? psf, been discussed. Besides, all things must eventually come to an end.

where the density p Ib/ft.°, g, fps”, r,, r;, ft. result in s, psf; » is Po
ratio; if all linear units are inches (g, = 386 ips”), the stress is then s; p
If the hole becomes so small that r; ~ 0, the maximum stress from (
becomes 4
_ pw*ro(3 + #) ps3 + #)
(0) St
4g, 4g,

Fig. 20.12. The maximum stress from centrifugal action in a rotating so

cylinder or disk is at the center and is™-?! :

_ pvs'(3 + #)
(Pp) Se >

8g,
from which, by comparison with (0), we conclude that the stress is dou
by putting a pinhole in the center of a solid disk. As Timoshenk
points out, the central region of large forgings, where the stress is a
mum, is likely to contain defects (stress raisers); hence, it is comm
practice to drill or bore a central hole, which serves the additional pu
of permitting inspection for defects. If the disk or cylinder is made of d
material, it can be run at overspeeds to stress the metal adjacent to the
beyond the yield strength, thus resulting in a residual compressive st
the phenomenon being as described for autofrettage, § 4.23, except th
centrifugal force is producing the overstressing.
We should note that the maximum stress, from (0) or (p), increases |
the square of the size r, of the disk, and as the square of its angular velo
If the disk is of uniform thickness t, Fig. 20.12, the tangential stress
creases toward the outer periphery. For this reason, rotors are 0
tapered in the outward radial direction to approach more closely a unifort
stress condition. See Refs. (/.2) and (0.7).
 LIST OF REFERENCES

This is not a bibliography. It includes books and papers for which acknowledg-
ment to specific material could be made, and it also includes works that should be
helpful to those who wish to pursue further study. The references whose first
digit is zero are general, many of which should be at hand in a design office.

0.1 ASME (S. L. Hoyt, ed.), Metals Properties, McGraw-Hill.

0.2 ASME (O. J. Horger, ed.), Metals Engineering—Design, McGraw-Hill.
0.3 ASME (J. Huckert, ed.), Engineering Tables, McGraw-Hill.
0.4 ASME (R. W. Bolz, ed.), Metals Engineering—Processes, McGraw-Hill.
0.5 Baumeister, T., ed., Marks’ Mechanical Engineers Handbook. 6th ed.,
McGraw-Hill.
0.6 Carmichel, C., ed., Kent’s Mechanical Engineers Handbook, 12th ed.,
Wiley.
0.7 Roark, R. J., Formulas for Stress and Strain, 3rd ed., McGraw-Hill.
Proceedings of Society for Experimental Stress Analysis.
Transactions of the American Society of Mechanical Engineers.

11 Timoshenko and MacCullough, Elements of Strength of Materials, Van

Nostrand.
12 Timoshenko, S., Strength of Materials, 2 vols., Van Nostrand.
1,3 Higdon, Ohlsen, Stiles, Mechanics of Materials, Wiley.
1.4 Beedle, L. S., Plastic Design of Steel Frames, Wiley.
15 Grinter, L. E., Theory of Modern Steel Structures, Macmillan.
1.6 Faires and Chambers, Analytic Mechanics, 3rd ed., Macmillan.
 REFERENCES 543

1.7 Seeley and Smith, Advanced Mechanics of Materials, 2nd ed., Wiley. 2.18 Heat Treatment of Stainless Steels, pub. by Rustless Iron & Steel Corp.,
1.8 Slupek, Main, and Fugenschuh, Diameter and Length of a Pin for ¢ Baltimore.
Pin-Bushing Type, Reprints from General Motors Engineering Journal 2.19 Simonds, Weith, Bigelow, Handbook of Plastics, Van Nostrand.
1959. 2.20 Plastics Engineering Handbook, Soc. of the Plastics Industry, Reinhold.
1.9 ASME, Boiler Construction Code. 2.21 Smith, P. C., Plastics for Production, Chapman & Hall. \
1.10 ASME, Rules for Construction of Unfired Pressure Vessels. 5 2.22 Lennie, A. M., Designing for Magnesium, Metals Handbook, 1948, p.
1.11 Maulbetsch and Hetényi, Stresses in Pressure Vessels, ASME Desigi 999.
Data, Book 1, p. A 29. 2.23 Die Casting for Engineers, pub. by New Jersey Zinc Co.
1.12 Boley and Weiner, Theory of Thermal Stresses, Wiley. q 2.24 Metal Quality, pub. by the Drop Forging Assoc., 605 Hanna Bldg.,Cleve-
1,13 Manson, S. S., Thermal stresses, Machine Design, Vol. 33, Nos. 23-26 land.
incl. % 2.25 Evans, Ebert, Briggs, Fatigue properties of cast and comparable wrought
1.14 Marin, Joseph, Mechanical Behavior of Engineering Materials, Prentice steels, Proc. ASTM, vol. 56.
Hall. 2.26 Hausner, H. H., ed., Modern Materials, 2 vol., Academic Press.
1.15 Slaymaker, R. R., Mechanical Design and Analysis, Wiley. 2.27 Kinney, G. F., Eagthdering Properties .. . of Plastics, Wiley.
1.16 van den Broek, J. A., Theory of Limit Design, Wiley. q 2.28 Symposium on Basic Mechanisms of Fatigue, ASTM Sp. Tech. Pub. No.
1.17 Hall, Holowenko, Laughlin, Theory and Problems of Machine De ign 237.
Schaum Pub. Co. A good study aid. 2.29 Finnie and Heller, Creep of Engineering Materials, McGraw-Hill.
1.18 Crandall and Dahl, The Mechanics of Solids, McGraw-Hill. 2.30 Mantell, C. L., Engineering Materials Handbook, McGraw-Hill.
1.19 Hinkle, R. T., Design of Machines, Prentice-Hall. q 2.31 ASTM Special Technical Publications Nos. 151, 170, 199 (high tempera-
1.20 Miller and Wright, Contact stresses, Machine Design, Vol. 35 No. 17 ture).
p. 185. q 2.32 ASTM Special Technical Publitation No. 158 (low temperature).
1.21 Asimow, M., Introduction to Design, Prentice-Hall. 2.33 Gillespie, Saxton, ahd Chapman, New design data for FEP and TFE,
1.22 Vidosic, Bogardus and Durden, Curved beams with eccentric boundarie: Machine Design, Vol. 32, No. 2, p. 126.
ASME Trans., Vol. 79, p. 1317. 2.34 Vanden Berg, R. V., Anodic coatings, Machine Design, Vol. 34, No. 6,
1,23 Vidosic, J. P., Machine Design Projects, Ronald. p. 155.
1.24 Myatt, D. J., Machine Design Problems, McGraw-Hill. 2.35 Plastics Book, Machine Design, Vol. 36, No. 22.
2.36 Gaston, S. P., Cold-drawn parts, Machine Design, Vol. 35, No. 10, p. 142.
2.1 Metals Handbook, 8th ed., by Am. Soc. for Metals, Metals Park, Novel} 2.37 Briggs and Parker, Alloy steels, Machine Design, Vol. 35, No. 13, p. 153.
O. (The most complete single reference on metals for mechanical¢ 2.38 The Ferrous Metals Book, Machine Design, (updated periodically).
gineers.) 2.39 Nonferrous Metals Book, Machine Design, Vol. 35, No. 22.
2.2 Soc. Auto. Eng. Handbook.
Publications of the International Nickel Co., Inc. 3.1 ASA standard B 4.1—1955, Preferred Limits and Fits for Cylindrical
2.4 Molybdenum Steels, pub. by Climac-Molybdenum Co. Parts.
2.5 Vanadium Steels and Irons, pub. by Vanadium Corp. of Am. 3.2 Burgess, A. R., Selective interchangeability, Texas Industrial Engineer,
2.6 Modern Steels and Their Properties, pub. by Bethlehem Steel Co. Vol. 2.
2.7 Carilloy Steels, pub. by U.S. Steel. 3.3 Shainin, D., Quality control methods, their use in design, a series of nine
2.8 Publications of Am. Soc. for Testing Materials. ‘ articles in Machine Design, beginning Vol. 24, No. 7. Good for an
2.9 Mechanical Properties of Metals and Alloys, Circular No. C 447, Bur. ¢ introduction.
Standards. q Mallett & Lundberg, Tolerances and allowances for interchangeable
2.10 Steel Handbook, Republic Steel Corp. assembly, Product Engineering, Vol. 15, p. 477.
2.11 Stainless Steel Handbook, pub. by Allegheny Ludlum Steel Corp. Mikelson, W., Determining surface roughness, Mech. Eng., Vol. 69, p. 391.
2.12 Reynolds Aluminum Data Book, Reynolds Metals Co. Sawyer, J. W., Surface finish literature, Machine Design, Vol. 24, No. 9,
2.13 Publications of Aluminum Company of America. F p. 147.
2.14 Publications of Am. Foundrymen’s Assoc., 222 W. Adams St., Chicago Bolz, R. W., Production Processes, Penton Pub. Co. This book contains
2.15 Publications of Gray Iron Founders’ Society, 33 Public Square, Clevelan considerable information on manufacturing processes of value to a
13, O. = designer.
2.16 Steel Casings Handbook, by Steel Founders’ Soc. of Am., 920 Midlan Broadston, J. A., Standards for surface quality and machine finish designa-
Bldg., Cleveland. tions, Product Engineering, Vol. 15, p. 622.
2.17 Am. Malleable Iron, by Malleable Founders’ Soc., Cleveland. ASA Standard, B46.1—-1955, Surface Roughness, Waviness, and Lay.
 REFERENCES REFERENCES 545

Engineering,
3.10 Trowbridge, R. P., Surface finish and the designer, Product 4.16 rise a. ee factors around a central circular
Vol. 21, No. 9, p. 122. ae Plate loaded through a pin in the hole, ASME Trans., Vol. 62,
ring, Vol. 19, Nos.
3.11 Nielsen, L. M., Shop run tolerances, Product Enginee
5 and 6. 4.17 Peterson,ie
ques ; R. re E. Stress
= We concent
entration
t phenomena ini fatigue
i of metals,
Die Forgings, Drop
3.12 Standard Practices and Tolerances for Impression
Forging Assoc. 4.18 Peterson and Wahl, 5 Two- - and thr ee-dime
i i
ring and Manu-
3.13 ASA Standard, B 4.1-1947, Limits and Fits for Enginee “an nse ASME Trans., Vol. 58, p. A-15. Sink Ob caaasnd tet
facturing. a ASME
rocht, enna
M. M., Factors e of stress
Ae concentrati:
ration photoelastically
i determined,
i
3.14 ASTM, Manual of Quality Control of Materials.
-Hill. Good for a
3.15 Grant, E. L., Statistical Quality Control, McGraw ke Neugebauer, , G. G. H., H., Stress concentrati ion facto i i
beginner on this subject. Product Engineering, Vol. 14, p. 82. aes ane Siig
Machine Design, Vol. 4.21
3.16 Brooks, K. A., Statistical dimensioning program, ae R. E., Stress Concentration Design Factors, John Wiley. A fine
33, No. 19, p. 140. a sth oe highly recommended for all design offices
surface roughness,
3.17 Hagen and Linberg, Instrumentation for determining Y mai ger an eifert, ce Shot peening ig” to to improve
i: fatigue resistance, Proc.
18.
G. M. Engineering Journal, Vol. 1, No. 7, p.
ne Design, Vol.
3.18 Geuder and Ebeling, Press fits with plastic parts, Machi 4.23. ae
Moore, H. F., , Size
Si effect and notch sensitivity,
itivi’ Proc: ASTM, Vol. 45, p.
34, No. 9, p. 121.
e Design, Vol. 23, pepe igue of Metals, translated by A. J. Fenner, Chapman
3.19 Ring, W. A., Determining practical tolerances, Machin 4.24 Cazaud, R., Fatigue &
No. 3, p. 121.
from Machine Design,
3.20 Bolz, R. W., Standard fits and tolerances, reprinted mo Seat ce . M., Strength and Resistance of Metals, Wiley.
January 1955. i ulen,
ates Cummings, Schulte, a series of arti icles in
i Machine
i Design,
i Vol.
3.21 Bowker and Lieberman, Engineering Statistics, Prentice-Hall.
ope — Manual for Fatigue Testing, Pub. No. 91.
E roceedings, International Conference on Fati,
41 Sines and Waisman, ed., Metal Fatigue, McGraw-Hill. Inst. of Mech. Eng., London. eneiae Me ae
summary of the
Contains a wealth of detailed references; a good recent ye ue ae on Fretting Corrosion, Pub. No. 144
state of the art. Also Ref. (4.62). ; emplin, R. L., Fatigue of aluminum, H. W. Gill i
ctured Parts, McGraw-
4,2 Lipson, Noll, Clock, Stress and Strength of Manufa = (ASTM), distributed by Alcoa. Gesu gM Ty
Hill. Many charts in this one. 31 peek Gordon, and Jackson, Fatigue of Metals and Structures
Metals, John Wiley.
4.3 Battelle Memorial Institute, Prevention of Fatigue of YWEPS 00-25-534) Supt. Doc., US Govt. Print. Office., Washing-
Metals and Alloys,
4.4 Karpov, A. V., Fatigue problems in structural designs, re ps et ze phere ae of experimental fatigue strengths
December 1939. l. ve r, Horger, Sanders, Lo comotive
i axial
i testing,
i ASME Trans.,
4.5 Moore and Kommers, The Fatigue of Metals, McGraw-Hill.
steels,
4.6 McDowell, The fatigue endurance of killed, capped, and rimmed me ASTM, Metals, Sp. Pub. No. 196.
Metals and Alloys, January 1940. a
\. Guhse, ; D. E., ; Effect of Two Stress Rais: ers
mice rs A i thesis,i USN
Acting at a Point,
axial testing, ASME Trans.,
4.7 Buckwalter, Horger, Sanders, Locomotive
Vol, 60, p. 335. 5 Bridge,
ridge, J. A., , Jr., Jr, Fatigue Stre ngth Reduction
i Caused b i.
on fatigue strength, Proc. Fe tinuities at a Point, thesis, USN Postgraduate School.
4.8 Morkovin and Moore, Effect of size of specimen oa
ASTM, Vols. 43 and 44. pp ee R:; oe Behavior of Engineering Structures, Wiley.
ASST, Vol. 42.
4.9 Peterson, R. E., Fatigue tests of small specimens, Trans. . he eiller
oes an tube, ed., Internal Stresses and Fatigue
ti inir Metals,
Handbo ok of Experi mental Stress Analysi s, John Wiley.
4.10 Hetényi, M., et al.,
Wiley.
4.11 Murray, W. M.., ef al., Fatigue and Fracture of Metals, John 38 Pope, + J. J. A., A,, ed., ed., Metal Fati;
atigue, Chapman and Hall.
4.12 Moore, H. F., et al., Surface stressing of metals, ASM. a des M. M.., Photoelasticity, Wiley.
ASME Trans.,
4.13 Soderberg, C. R., Factor of safety and working stresses, J5 rankland,
pe age J. M.,
oy 2 i
impact i
on simple elastici structures, Proc.
Vol. 52., p. 13.
ed variable-load
4.14 Hinkle, R. T., A simple method of presenting the combin . Hudson, G. E., A method of estimati i i i
ex elastic structures, Proc. SESA, Vol. VE Ne:
equations, Jour. ASEE, Vol. 41, p. 409. et ee ea cna
ed by holes and notches, Dohrenwen ‘i
d and 99:M ehaffy, Dynamic loading ine design, Machine Design,
4.15 Wahl and Beeuwkes, Stress concentration produc yodstas Hens
ASME Trans., Vol. 56, p. 617.
f 6 F
Hagenbook, L. D., Impact loads by Brinell techniques, Product Engineerin 5.3 Class 5 Interference-Fit Thread, ASA B1.12, Tentative, ASME.
Vol. 14, p. 300. 5.4 Unified Miniature Screw Threads, ASA B1.10-1958, ASME.
Marco and Starkey, A concept of fatigue damage, ASME Trans., Vol. 76 5.5 Acme Screw Threads, ASA B1.5-1952, ASME.
p. 627. 5.6 Stub Acme Screw Threads, ASA B1.8-1952, ASME.
4.45 Tavernelli and Coffin, Experimental support for generalized equat 5.7 Buttress Screw Threads, ASA B1.9-1953, ASME.
predicting low cycle fatigue, ASME paper 61-WA-199. 5.8 Seaton and Routhewaite, Marine Engineer’s Pocket Book.
4.46 Gatts, R. R., Cumulative fatigue damage with random loading, AS f 5.9 Fasteners Data Book, Industrial Fasteners Institute, Cleveland 15, Ohio.
paper 61-WA-31. 5.10 Maney, G. A., Predicting bolt tension, Fasteners Data Book.
4.47 Dieter, Horne, Mehl, Statistical study of overstressing steel, NACA Te o 5.11 Crane Catalog, No. 53, Engineering Data Section.
Note No. 3211. , 5.12 Laughner and Hargan, Handbook of Fastening and Jointing of Metal Parts,
4.48 Grinding Stresses, collected papers, pub. by Grinding Wheel Institu McGraw-Hill.
Keith Bldg., Cleveland. { 5.13 Wilhelm, J., Torque-tension standards, Machine Design, Vol. 34, No. 4,
4.49 ASTM, Effect of Temperature on the Brittle Behavior of Metals, Sp. Tee p. 159.
Pub., No. 158. 5.14 The Fasteners Book, pub. by Machine Design; also contains numerous
4.50 Vitovec, F. H., Superalloys under combined stress at elevated temper eet showing a variety of products and suppliers; Vol. 35,
tures, Machine Design, Vol. 34, No. 11, p. 157. 0. 7.
4.51 Langer, B. F., Design of pressure vessels for low-cycle fatigue, ASN 5.15 Almen, J. O., Tightening is a vital factor in bolt endurance, Machine
paper 61-WA-18. Design, Vol. 16, No. 2, p. 158.
Gwinn, J. T., Jr., Stop over-designing for impact loads, Machine Desig 5.16 Dolan and McClow, The influence of bolt tension and eccentric tensile
Vol. 33, No. 16, p. 105. 4 loads on the behavior of a bolted joint, Proc. SESA, Vol. VIII, No. 1, p. 29.
Ellis and Karbowniczek, Absorbing shock loads, Machine Design, Vo 5.17 Pickel, W. F., Tightening characteristics of nut and stud assemblies, Product
34, No. 6, p. 176. 7 Engineering, Vol. 20, p. 98.
4.54 ASTM, Symposium on Statistical Aspects of Fatigue, Sp. Tech. 5.18 Dolan, T. J., Load relations in bolted joints, Mechanical Engineering, Vol.
No. 121. 64, p. 607.
4.55 ASTM, Symposium on Basic Mechanisms of Fatigue, Sp. Tech. Pui 5.19 Radzimovsky, E. I., Bolt design for repeated loading, Machine Design,
No. 237. Vol. 24, No. Il, p. 135.
ASTM, Tentative Guide for Fatigue Testing and the Statistical 5.20 Moslander, K. D., Machine fastenings in design, Machine Design, Vol.
of Fatigue Data, Sp. Tech. Pub. No. 91-A. 13, No. 1, p. 39.
4.57 Peterson, R. E., Analytical Approach to Stress Concentration Effect 5.21 Hetényi, M., A photoelastic study of bolt and nut fastenings, ASME Trans.
Fatigue of Aircraft Materials, U.S Air Force Sc. paper No. 10-0508 Vol. 65, p. A-93.
4.58 Fuchs and Hutchinson, Shot peening, Machine Design, Vol. 30, No. 5.22 Wesstom and Bergh, Effect of internal pressure on stresses and strains in
p. 116. sg bolted flanged connections, ASME Trans., Vol. 73, p. 553.
4.59 Henry, D. L., A theory of fatigue-damage accumulation in steel, ASN 5.23 Roberts, I., Gaskets and bolted joints, ASME Jour. App. Mech., Vol. 72,
Trans., Vol. 77, p. 913. ‘ p. 169.
4.60 Miscellaneous articles, in Materials Research and Standards, Vol. 3 5.24 Rossheim and Markl, Gasket loading constants, Mech. Eng., Vol. 65,
No. 2, ASTM. ’ p. 647.
4.61 Coffey, R. L., Effect of superposition of stress raisers on axial-load fatigt 5.25 Armstrong Cork Co., Gasket Design Manual.
thesis, USN Postgraduate School. é 5.26 The Seals Book, pub. by Machine Design.
4.62 Heywood R. B., Designing against Fatigue of Metals, Reinhold. ; 5.27 Oest, L., Selecting and applying tapping screws, Machine Design, Vol. 32,
4.63 Danforth and Starkey, The effects of high-stress-amplitude fatigue dama No. 5, p. 115.
on the endurance properties of high-performance stainless-steel and titani 5.28 Ollis, Ray Jr., Self-aligning nuts, Machine Design, Vol. 34, No. 14, p. 176.
alloys, ASME paper, 62-WA-230. a 5.29 Wilson and Oliver, Tension Tests of Rivets, Bull 210., Eng. Exp. Sta.,
4.64 Almen and Black, Residual Stresses and Fatigue in Metals, McGraw-Hi Univ. of Illinois.
4.65 Lipson and Juyinall, Handbook of Stress and Strength, Macmillan. 5.30 Crum, R. G., Fatigue in metal joints, Machine Design, Vol. 33, No. 7,
p. 108.
5.31 Doughtie and Carter, Bolted assemblies, Machine Design, Vol. 22, No. 2,
5.1 Unified Screw Threads, ASA B1i.1-1960, ASME. p. 127.
5.2 Nomenclature, Definitions, and Letter Symbols for Screw Threads, 5.32 Creech, M. D., Screw-thread torque, Machine Design, Vol. 35, No. 12,
B1.7-1949, ASME. p. 173.
 REFERENCES 549

5.33 Pfeiffer, W., Bolted-flange assemblies, Machine Design, Vol. 35, 6.27 Weiss, K., Helical compression springs, Machine Design, Vol. 34, No. 28,
p. 193. p. 173 (graphical solution).
AISC, Steel Construction, 6th ed.
71 Stephenson and Cloninger, Stress Analysis and Design of Steel Columns,
6.1A Wahl, A. M., Mechanical Springs, Penton Pub. Co. Tex. Eng. Exp. Sta. Bull. No. 129, College Station. Contains help in
6.1B Wahl, A. M., Mechanical Springs, 2nd ed., Wiley (appeared¢ the form of tables and charts for solving the secant column formula.
chapter was written). 1.2 Young, D. H., Rational design of steel columns, ASCE Trans., Vol. 101,
Associated Spring Corp., Mechanical Springs. p. 442.
6,3 The Mainspring, a publication by the Associated Spring Corp., B Bri 1.9 Merriam, K. G., Dimensionless coefficients applied to the solution of
Conn. column problems, Jour. Aero. Sciences, Vol. 7, p. 478.
6.4 Bittner, E. T., Alloy spring steels, Trans. ASM, Vol. 40. 7.4 Hoeschel, H. G., Simplified column design, Machine Design, Vol. 31,
Tatnall, R. R., Fatigue of helical springs, Mech. Eng., April 1940. No. 7, p. 135.
6.6 Votta, F. A. Jr., The theory and design of long-deflection constant 7.5 Salmon, E. H., Columns, Hodder & Stoughton.
spring elements (Neg’ator), Trans. ASME, Vol. 74, p. 439. 1.6 Guide to Design Criteria for Metal Compression Members, by Column
6.7 Nolde, G. V., Designing leaf springs for precision timing and. in Research Council, Crosby Lockwood and Son, Ltd., London.
devices, Machine Design, Vol. 24, No. 6, p. 125. 7.7 Newman and Forray, Beam columns, Machine Design, Vol. 35, No. 5,
6.8 Johnson and Crooks, Helical springs for high temperature, } Mi: p. 145.
Design, Vol. 35, No. 19, p. 127. 1.8 Dow, James, Columns loaded between supports, Machine Design, Vol. 33,
6.9 Meyers, O. G., Working stresses for helical springs, Machine E No. 4, p. 167.
Vol. 23, No. 11, p. 135.
6.10 Chandler, R. D., Compression spring design, Machine Desiguil 8.1 Nadai, A., Theories of strength, ASME Trans., Vol. 55, p. 111.
No. 22, p. 138. 8.2 Nadai and Wahl, Plasticity, McGraw-Hill.
6.11 ASTM Standards, Part 1, Ferrous Metals. 8.3 Van den Broek, Elastic Energy Theory, John Wiley.
6.12 Personal communication, W. R. Johnson, Associated Spring 8.4 Den Hartog, J. P., Advanced Strength of Materials, McGraw-Hill.
6.13 Technical Bulletin, T-35, International Nickel Co. 8.5 Soderberg, C. R., Working Stresses, ASME Trans., Vol. 55, p. 131.
6.14 Atterbury and Diboll, Effect of presetting helical compa Sp 8.6 Hot Rolled Carbon Steel Structural Shapes, U.S. Steel Corp.
ASME J. Eng. for Ind., Vol. 82, p. 41. y 8.7 Ham and Ryan, An Experimental Investigation of the Friction of Screw
6.15 Hunter Spring Company, Spring Design Data. Threads, Bull. 247, Eng. Exp. Sta., Univ. of Ill.
6.16 Hinkle and Morse, Design of helical springs for minimum weight, 00 8.8 ASA Standard, B 1.5-1952, Acme Screw Threads.
and length, ASME J. of Eng. for Ind., Vol. 81, p. 37. 8.9 ASA Standard, B 1.8-1952, Stub Acme Screw Threads.
6.17 Chandler, R. D., Double-nested compression springs, Machine | 8.10 Findley, W. N., Fatigue of metals under combinations of stresses, ASME
Vol. 34, No. 19, p. 177. Trans., Vol. 79, p. 1337.
6.18 Almen and Laszlo, The uniform section disk spring, ASME Tra 8.11 Findley, Mitchell, Strohbeck, Effect of range of stress in combined bending
58, p. 305. and torsion fatigue tests of 25 S-T6 aluminum alloy, ASME Trans.,
6.19 Weibel, E. E., The correlation of spring-wire bending and tors Vol. 78, p. 1481.
tests, ASME Trans., Vol. 57, p. 501. 8.12 Findley, W. N., A theory for the effect of mean stress on fatigue of metals
6.20 Wang and Worley, Load-deflection behavior of conical spiral compt under combined torsion and axial load or bending, ASME J. Eng. for Ind.,
springs, ASME Trans. Paper No. 61-WA-128. 4 y Vol. 81, p. 301.
Groesberg, S. W., Zero-gradient spring system, Machine Design, ¥ 8.13 Cox and Field, The initiation and propagation of fatigue cracks in mild
No. 2, p. 143. steel pieces of square section, Aeronaut. Quarterly, Vol. 4, part I, p. 1.
Heldt, P. M., Motor Vehicles and Tractors. Bic 8.14 Murphey, Glenn, Advanced Mechanics of Materials, McGraw-Hill.
6.23 Smith, J. F. D., Rubber mountings, ASME Trans., Vols. 60 and 61 8.15 Crossland, Jorgensen, and Bones, The strength of thick-walled cylinders,
6.24 Brown, R. W., Engineering properties of rubber in compression, SAB’ ASME Paper 58—-PET-20.
Vol. 47, p. 432. 8.16 Davidson, Eisenstadt, and Reiner, Fatigue characteristics of open-end
6.25 Votta, F. A., Constant-force springs, Machine Design, Vol. 35) thick-walled cylinders under cyclic internal pressure, ASME Paper 62-
p. 102. { WA-164.
6.26 Krotz, A. S., Elastomeric torsion springs, Machine Design, Vol. 8.17 Sessler and Weiss, Low-cycle fatigue damage in pressure-vessel materials,
p. 130. ASME Paper 62-WA-233.
 REFERENC REFERENCES 551

ns of press and shrii 11.8 The Bearings

i Book, Machine Design, Vol. 35, No. 14.
nsmission of torque by mea
Baugher, J. W., Jr., Tra 11.9 Howarth, »H.
H.A.SA.S., The loadi icti
ASME Trans., Vol. 57, p. 169. and friction of thrust and journal bearings,
8.18
53, p. 85 MSP.
fits, ASME Trans., Vol. d assemblies, Desi
and Nel son , Des ign of press- and shrink-fitte 11.10 Howarth, » H. H. A A. S., Graphical analysi.
8.19 Hor ger
by ASME.
Data and Methods, pub. ign, Vol. 26, No. 1, a ida Vols. 45, 46, 47, pp. 21, is ea
Des ign ing for shrink fits. Machine Des Ln eo a
8.20 Faupel, J. H., aimondi, Applying bearing teary to the analysis and desi sign
p. 114. ASME. Ha
of journal beariings, ASME Jour. of Applied Mech., Vol. 73, pp. 298 and
Buttress Screw Threads,
$.21 ASA Standard, B1.9-1953. Eng . Ed., Vol. 38, No. 9, P- 653.
ss factors, J.
8.22 + Vidosic, J. P., Design stre a
11.12 a aint‘ay ae McGraw-Hill.
ca , E.1,, Lubrication of Bearii
ission Shafting, ASME.
91 Code for Design of Transm ms, Machine Desi Hersey, M. D., Theory of Lubrication, Wiley.
shafts and nonuniform bea 11.15 Needs, , S. S. J., J., Eff Effects of side leakage
ny
9.2 Hopkins, R. B., Stepped i: fF
Vol. 33, No. 4, p. 159. 17, P- 185. bearings, ASME Trans., Vol. 56, pnt “wllge hehe oto
hine Design, Vol. 34, No. tgrriee
9.3 Griffel, Wm., Beams, Mac shaf ts, Machine Desi 11.16 Kingsbury, A., “> Opti;fi
ptimum conditions in journal bearings, ASME Trans
d of multiple loaded Vol. 54, p. 123. >

Grif fel, Wm. , Crit ical spee

9.4
Vol. 34, No. 11, p- 153. 11.17 McKee and McK
3, Shafting and stock keys
. ASME Trans., Vol.ee, 54,
Pressure
p. 149. dist.stribution in; oil films of journal bearings,
9.5 ASA Standard, B 17.1-194
14th ed.
9.6 Machinery’s Handbook, Van Nostrand. 11.18 Bradford,
»
L. J
L. J., Oil film pressures i in
n Problems in Engineering,
Timoshenko, S., Vibratio McGraw-Hill. oe
9.7
n Prevention in Eng ine eri ng,
= 0. 14, Penna. State Col.
B., Performance of oll-ring beariings, ASME Trans., Vol. 52 :
Kimball, A. L., Vibratio Ii 11.19 asl
Analysis by Mobility and
_ G.52~5-57.
9.8
9.9 Thorn and Church, Simp
lified Vibration
Pub. Co. ical Met
pedance Methods, Penton n, Wiley. aralitz
ts of Mechanical Vibratio
.
, Ele men . Kenyon,
viscous ‘ia rication, Oil-film thickness of transiti
9.10 Freberg and Kemler ion , McG raw -Hi ll. ASME Trans., Vol. 59, p 239, Ce earner
ical Vibrat
Den Hartog, J. P., Mechan 11.21 Dennison, > E. % S +> Film lub
9.11 ics of Vibration, Wiley. ls i eng eM
i

9.12 Hansen and Che nea , Mec han

ASME Trans., Vol. 58, p. herrea et
l Vibrations, Wiley.
Church, A. H., Mechanica Vol. 32, No. 11.22 Howarth, » H.A
9.13 design, Machine Design,
H.A.S., Current [1934] pri
ula r shaf t
Tsien, V. C., Tub and lubrication of oil-film iaalitess eae
*

9.14
e
meni ig val . Geen
i
56, p. 891. an
p. 128. the bending fatig 11.23 Burwell, » Ka Kaye, van Nymegen
metrical design factors on ASME Trans., Vol. 63, p. ser Morgan, Effects of surface finish,
9.15 Kano, C., Influence of geo 134 .
ASME paper 62-WA-
strength of crankshafts, 11.24 Rippel,
,
H. C i
of inf ormative
H. C., A series :
articles j ine Desi
tent
kel Co. 1135 fue m ri October 1, October 29, November 12, oo eee
p. 10, International Nic
INCO, Vol. 23, No. 2, Pa
10.4 s.
, Involute Spline 11.26 McKee ee i. be Bronze Co. New Castle,
10.2 ASA Standard B 5.15-1950 cKee, Frictionof j . a
, Involute Serrations. re
10.3. ASA Standard B 5.26-1950
Armitage, J. B., Straig
ht-sided splines, Mech.
Eng., Vol. 70, p- 738.
Vol. 34, No. 29.
say thd length, ASME Trans, vost ten Mien
pub. by Machine Design,
10.4
Mechanical Drives Book, 11, pub. by Tex:
11.28 Moh dab: W Thermodynamics, Macmillan.
10.5
10.6 Flexible Couplings, in Lub
rication, Vol. 47, Nos.
10,
1129 ibdaahie-
Mais and Sisson.
Mieke » Fra Wet PE itn ainde
Journalce i friction the
the region of thin-film
Desig
et
Inc. couplings, Machine
to design flexible McKee
10.7. Gensheimer, J. R., How 11.30 , S. A. Journal beari
‘
m load:
September 14, Vol. 33,
p. 154.
33, No. 24, p- | and operating temp eratures, National Bur. St’d. Research pap ring design as related to maximu er i
don , W.T ., Uni ver sal joints, Machine Design, Vol. p. 491.
10.8 Con
ion for Engineers, Wiley 11.31 McKee , 8S. A.,Beckuiber
Mech. Eig , The effect 0, i :
oe running in on journal-bearing performance,
and Practice of Lubricat
11.1. Fuller, D. D., Theory , Wiley.
ring Lubrication Analysis
11.2 Slaymaker, R. R., Bea raw -Hill. d McKee, I TICHON
s and Lubrication, McG
O, ‘ournal beari ‘S as in luenced 4 clear ‘ance
Clower, J. 1, Lubricant ings, McGraw-Hi
11.33 Baudry. ,
and length.
11.3 ion of Bear ASME T rans., V aa
ysis and Lubricat Z
41.4 Shaw and Macks, Anal ion, McGraw-Hill. R. A., Some ther:
Wilcock and Booser, Bea
ring Design and Applicat aa Trans. , Vol. 67, No. 2, p. ee effects in oil-ring journal bearings, ASME
11.5 Lubr ication, McGraw- ‘
ory of A)'ydrodynamic ; arelitz, G. B. i
Pinkus and Sternlicht, The
11.6
Raimondi and Boyd, A
solution for the fini te jour nal bearing and |
E Trans., Vol. 11.35 McKee and White, Oil tole arings in machines , ASME Trans., 1929 ? .

ASL
and grooves in plain journal bearings, ASME
11.7. s,
design, in three part
? é:

application to analysis and Trans., Vol. 72, p. 1025.

No. 1, p. 159-
 REFERENCES
582
11.60 Du Bois, Ocvir k, and Wehe, Experimental investigation of
soft-rubber bearings, ASME Trans., eccentricity
11.36 Busse and Denton, Water-lubricated Te frictio
ratio, WOn,Eand oil flow of long and short journa
i l bearings ..., NACA
Vol. 54.
data for FEP and TFE, Machine
11.37 Gillespie, Saxton, Chapman, New design 11.61 Halberstadt,
13, p. 181.
L., » Pol: 5
Polyurethane bearings, Machine Design, Vol. 35, No.
Design, Vol. 32, No. 4, P- 156. 21
Machine Design, Vol. 32, No. 11.62 Ruffini i
Carlyon, G., Plastic sleeve bearings,
11.38
p. 168.
Machine Design, Vol. 34, No. 23
11,63 Boas
; tha Hicaa n, codesOutgass i acteristi eg ak
11.39 Rentschler, L. M., Plastic bearings,
11.64 Kla
ina acum, Machine Design, Vol. 35, Sucka
p. 156. Design, Vol. 35, » &.E,, New developments in liquid lubri: ‘ i \
Machine
11.40 Walker, G., Bonded solid-lubricant
coatings, 11.65 SME Paper Gd MEL.
bearings, Machine ‘es
No. 2, p. 182. ah H.
be: icles, C., Hydrostatic
starting in Vol. 35, No. 18 Mee eceiee<of nine
of Teflon fluorocarbon res
O’Rourke, Lewis, and Lewis, Performance 11.66 owen, Boes and
Self-lubricati McDowell,
11.41
as bearing materials, ASME Paper
61-WA-334.
theory of fluid-film lubrication, with an
Machine: Design,
f
Vol. 35, No.20,p.139.. 0% mpaste. materials,
Kingsbury, A., On problems in the
11.42
Vol. 53, p. 59-
experimental solution, ASME Trans.,
development of theories of dry and boundary
11.43 Tichvinsky and Courtel, On the 12.1 Palmgren, 4 Ass Ball and Rolle Beari Engineeri ub. by
friction, ASME Paper 63-LUB-7. F In

theory to the analysis and design

11.44 Raimondi and Boyd, Applying bearing 12.2 Styri > H.
> Friction torque
77, p. 287.
m ball and
Vol. roller bear
of pad-type bearings, ASME Trans.,
ing. ‘Ss > M echanical Engineer. =

of longitudinal and transverse profile on

11.45 Raimondi, A. A., The influence
Trans. ASLE, Vol. 3, No. 2
Styri H., Fatigue strength a, ba bear ?
the load capacity of pivoted pad bearings,
tr

p. 265. a a General Catalog and Engineering Data

of orifice- and capillary-compensate
11.46 Raimondi and Boyd, An analysis January, 1957.
a Timken Engineering Journal.
Engineering,
hydrostatic journal bearings, Lubrication na = prs ee General Catalog.
A numerical solution for the gas lubricated full jour nee
11.47. Raimondi, A. A., tandard, Method of evaluating
i load ratings of annular ball
Vol. 4, p. 131.
bearing of finite length, Trans. ASLE,
Wiley.
11.48 Gross, W. A., Gas Film Lubrication, 12.8
129 ASA
Shube,Standard
Eugene, B 3 -ll- 1959, Load Ratings
‘i for Ball and Roller Bearings.
to the study of gas-lubricateé
11.49 Katto and Soda, Theoretical contributions Ball-beari
journal bearings, ASME Paper 61-LUB-8.
study of externally pressurizet
p. 158. aring survival, Machine Design, Vol. 34, No. 17,
11.50 Lemon, J. R., Analytical and experimental 12.10 Jones, » A. A. B., B., Ba Wl motion and slidi Sriction in ball bearings, ASME
ASME Paper 61-LUB-15. J.
air lubricated journal bearings, ee oe Eng., Vol. 81, No. 1, p. 2h
gas step bearings, Machine Design, Vol. 3
11.51 Adams, C. R., High-capacity arris, T. A., Predicting bearing reliability, Mach
No. 5, p. 118. No. 1, p. 129.
ine Design, 2 Vol. ? 35 ’
of dynamically loaded sleet
’
11.52 Burwell, J. T., The calculated performance 12.12 AS
A Standard B 3.7-1960, Terminology and De}
“

bearings, ASME Trans., Vol. 73,

p. 393. Roller Bearings. finitions for Ball and
Swindells, Measurements of the combined frictio 12.13
11.53 McKee, White, and AS 5-
lubrication, ASME Trans., Ve
7 rT ir e arings
and thermal behavior in journal bearing 12.14 Daniels, Cc M., High-load oscillating bearin
gs, Machine Design, Vol 32,
70, p. 409. No. 15, Pp. 136.
film viscosity in heavily loaded bearing
Needs, S. J., Influence of pressure on 12.15 Baniak and Kohl, Fati i 5
11.54
ASME Trans., Vol. 60, p. 347. No. 12, p19. nee” Wh of ball bearings, Machitie Design, Vol, 33
11.55 Designing with Zytel nylon resins, a
bulletin of the Du Pont Co.
sok
12.16 Accinelli > J. B., Grease lubricati .
oach to lubrication problems and the bearings, ASLE Trans., Vol. a el Eu ar eon
11.56 Hays, D. F., A variational appr Vol. 81, No.
E J. Basic Eng., 12.17
tion of the finite-journal bearing, ASM Dunk and Hall, Resistance to rolling and
sliding, ASME Paper 57-SA-9
of note. 12.18
p. 13. Contains chart s wort hy
oximation for full journal bearing
3Rowland
tee D. R., Ball-bearing splines, Machine Design, Vol. 33, No. 15,
11.57 Ocvirk, F. W., Short bearing appr
NACA Tech. Note 2808. r, J. H., Cle
11.58 Du Bois and Ocvirk, Experimental
investigation of eccentricity rai No. 12, p. 145.00 olla bearings, Machine Design, . Vol. 33 ’
friction, ... NACA Tech. Note 2809. Havemeyer, H. R.
11.59 Du Bois and Ocvirk, Analytical derivation
and experimental evaluatit 1957. » Linear ball bushings, Machine Design, August 22
ee approximation for full journ al bearings, NACA ref
12.21 Bidwell, J. B., ed., Rolling Contact Phenomena,
of Elsevier Pub. Co.
1157.
12.22 Lipson and Colwell, ed., Handbook of Mechanical Wear, Univ. of M 13.31 Talbourdet, G. J., Progress report on surface endurance limits of |engineering
Press.
materials, ASME Paper 54-LUB-14.
13.32 Rideout, T. R., High-speed gearing, Machine Design, Vol. 35, No. 19,
13.1 Dudley, D. W., ed., Gear Handbook, McGraw-Hill.
p. 162; also ASME Paper 63—-MD-46.
13.2 Faires, ¥V. M., Kinematics, McGraw-Hill. 13.33 Hense and Buswell, Economic factors affecting the steel selection and
Buckingham, Earle, Analytical Mechanics of Gears, McGraw-Hill. heat treatment for automotive gears, General Motors Eng. Jour., Vol. 2,
13.4 Black, P. H., An investigation of relative stresses in solid spur gea No. 5, p. 2.
the ‘liatoetaatle method, Bull. No. 288, Eng. Exp. Sta., Univ. of Ill, 13.34 Lipson, Charles, Wear considerations in design, a series of articles in
Dolan and Broghamer, A photoelastic study of stresses in gear too)
Machine Design, beginning Vol. 35, No. 25, p. 156.
profiles, Bull. No. 335, Eng. Exp. Sta., Univ. of Ill. 13.35 Wellauer, E. J., Applying new gear strength and durability formulas to
13.6 Straub, J. C., Shotpeening as a factor in the design of gears, ¥ design, Paper given to 1964 ASME Design Conference. (Changes in
Engineering, 1953 Handbook.
some details of AGMA calculations after we went to press.)
13.7 Connell, C. B., High-speed industrial gear drives, presented to AGMA
1948. NOTE: Most of the references in chapter 13 are appropriate also for
13.8 Tuplin, W. A., Gear Load Capacity, Wiley. chapters 14, 15, or 16.
13.9 Botstiber and Kingston, High-capacity gearing, Machine Deslea) Vv 14,1 Buckingham and Ryffel, Design of Worm and Spiral Gears, Industrial
24, No. 12, p. 129. Press.
13.10 ASME Research Report, Dynamic Loads on Gear Teeth. , 14.2 Welch and Boron, Thermal instability in high speed gearing, ASME
13.11 Candee, A. H., Introduction to the Kinematic Geometry of Gear Te Paper 59-A-118.
Chilton Co. 14,3 Glover, J. H., Nonstandard crossed helical gears, Machine Design, Vol.
13.12 Attia, A. Y., Dynamic loading of spur gear teeth, ASME Paper 58-S
33, No. 12, p. 156.
13.33 Reswick, J. B., Dynamic loads on spur and helical gear teeth, 14.4 Wellauer, E. J., Helical and herringbone gears, Machine Design, Vol. 33,
Trans., Vol. 77, p. 635. No. 28, p. 125.
13.14 AGMA publications.
13.15 Buckingham, Earle, Gear Tooth Loads, pub. by Mass. Gear and To ol C 15.1 The Gleason Works, Bevel and Hypoid Gear Design.
13.16 Jacobson, M. A., Bending stresses in spur gear teeth, Proc. I. 15.2 Coleman, Wells, Improved method for estimating fatigue life of bevel and
Vol. 169, p. 587.
hypoid gears, SAE Trans., Vol. 6, No. 2.
13.17 ASA Standard, B 6.10-1954, Gear Nomenclature. 15.3 Coleman, Wells, Designing bevel gears, Machine Design, Vol. 33, No. 24,
13.18 ASA Standard B6.12-1954, Nomenclature of Gear Tooth Wear; p. 127.
Failure. 5 jt i 15.4 Baxter, M. L., Jr., High-reduction hypoids, Machine Design, Vol. 33,
13.19 Hall and Alvord, Zytel spur gears, Mechanical Engineering, Vol. No. 9, p. 142
No. 5, p. 50. ; 15.5 Spear, King, Baxter, Helixform bevel and hypoid gears, ASME J. of Eng.
13.20 Dudley, D. W., Practical Gear Design, McGraw-Hill. i
for Ind., Paper 59—-A-90.
13,21 Fellows Gear Shaper Co., The Involute Curve and Involute Gearing. .
13.22 Gatcombe and Prowell, Rocket motor—gear tooth stress a S 16.1 Larson, G. A., Helical gear and worm sets, Machine Design, Vol. 35,
Paper 59-A-256. No. 6, p. 170.
13.23 Gatcombe, E. K., Lubrication characteristics of involute oe get
16,2 Nelson, W. D., Spiroid gearing, Machine Design, Vol. 33, Nos. 4 and 5,
ASME Trans., Vol. 67, p. 177. pp. 136 and 93.
13.24 Kinsman, F. W., Designing nonstandard spur gears, Machine Desi
Di 16.3 Popper and Pessen, The twinworm drive—a self- locking worm-gear trans-
Vol. 27, No. 6, p. 195. mission of high efficiency, ASME J. of Eng. for Ind., Paper 59-A~-75.
13.25 Tuplin, W. A., Gear Load Capacity, Wiley. 16.4 Vanick, J. S., The Ni-Vee bronzes, pub. Tnternaitionat Nickel.
13.26 Martin, L. D., Powder-metal gears, Machine Design, Vol. 33, No.
p. 200. 7A Power Transmission Handbook, pub. by Power Publishing Co., Cleveland.
13.27 Borsoff, Accinelli, and Cattaneo, The effect of oil viscosity on the, pom 17.2 Greenwood, D. C., Mechanical Power Transmission, McGraw-Hill.
transmitting capacity of spur gear, ASME Trans., Vol. 73, p. 687. 17.3 Wormey, W.S., Design of V-belt drives for mass-produced machines,
13.28 Borsoff, V. N., On the mechanism of gear lubrication, ASME J. B
Product Engineering, September, 1953, p. 154.
Eng., Vol. 81, No. 1, p. 79. 17.4 Taylor, F. W., Notes on belts, ASME Trans., Vol. 15, p. 204.
13.29 Borsoff, V. N., Gear lubrication, Machine Design, Vol. 34, No. 16, p. 1 17.5 Sawdon, W. M.., Tests of the Transmitting Capacities of Different Pulleys
Fellows Gear Shaper Co., The Internal Gear. in Leather Belt Drives.
 REFERENCES
556
18.2 Frehse, A. W., Fundamentals of brake design, SAE Jour., Vol. 27.
17.6 Bolz, R. W., ef al., High speeds in design, Machine Design, Vol. 22, No.
; ' a 18.3 von Mehren, O., Internal shoe clutches and brakes, ASME Trans.,
. 148.
Vol. 69, p. 913.
17.7 Noel C. A., High-Speed Belt Drives, Ohio State Eng. Exp. Sta. Br
18.4 Rasmussen, A. C., Internal friction, blocks and shoes, Product Engineering,
No. 83.
Vol. 18, No. 3, p. 133. Also other articles in same volume, May, July,
17.8 Norman, C. A., Tests on V-belt drives and flat belt crowning. ASME Tra
aU October, December.
Vol. 71, p. 335.
; j 18.5 Gould, G. G., Determination of the dynamic coefficient of friction for
17.9 Publications of American Leather Belting Association. ;
transient conditions, ASME
17.10 Tatnall, R. R., The pivoted-motor drive, Mechanical Engineering, Vol. ; Trans., Vol. 73, p. 649.
. 287. | ] 18.6 Hagenbook, L. D., Design of brakes and clutches of the wrapping band
type, Product Engineering, Vol. 16, p. 321.
17.11 ASA. Standard B 55.1-1961, Specifications of Multiple V-Belt Dri
. 18.7 Jania, Z. J., Friction clutch transmissions, Machine Design, Vol. 30, Nos.
Pub. by ASME.
23, 24, 25, 26.
17.12 Specifications for narrow V-belts, pub. by The Rubber Mfg. Asso.,
a
4
18.8 Eksergian, C. L., High speed braking, ASME Trans., Vol. 73, p. 935.
Madison Ave., NYC.
18,9 Lowey, F. J., Powdered-metal friction material, Mechanical Engineering,
17.13 Catalog of Gates Rubber Co.
17.14 Catalogs of U.S. Rubber Co.
18.10
Vol. 70, p. 869.
Loewenberg, F., Automotive brakes with servo action, ASME Trans.,
17.15 Catalog of Link-Belt Co. ne
q Vol. 52(1), p. 185 APM.
17.16 log of Diamond Chain Co.
18.11 Rasmussen, A. C., Heat radiation capacity of clutches and brakes, Product
17.17 rae. and Hornung, A quantitative investigation of the fae
Engineering, Vol. 2, p. 529. Temperature calculations for clutches and
which influence the fatigue life of a V-belt, ASME Paper 59-SA~184
brakes, Product Engineering, Vol. 3, p. 282.
17.18 ASA Standard B 29.1-1950, Transmission Roller Chains and Sprot
- 18.12 Siroky, E., Wagner Electric Corp., personal communications.
Teeth. 18.13 Engineering Information, Bull. 501, Raybestos Manhattan, Inc.
17.19 seams W. K. Jr., Dynamic loading of chain drives, ASME Tr
18.14 Borchardt, H. A., Designing external-shoe brakes, Machine Design, Vol.
Vol. 73, p. 655.
17.20 Gerla, M. K., Improving fatigue life of roller chain link plates,
4 18.15
32, No. 13, p. 163.
Rabins and Harker, The dynamic frictional
Design, Vol. 25, No. 1, p. 171. F characteristics of molded
friction materials, ASME Paper 60-WA-35.
17.21 ta tiewaowe s Rating Charts, Bull. 462, Diamond Chain Co., Ind
3 18,16 Fazekas, G. A. G., Graphical shoe-brake analysis, ASME Trans., Vol.
apolis, who may be willing to furnish copies for school use.
Chains and Sprocket T 79, p. 1322.
17.22 ASA Standard B 29.2-1950, Inverted Tooth 18.17 Huntress, H. B., four articles in Machine Design: Friction fundamentals,
17.23 John A. Roebling’s Sons Co., i
;
Vol. 27, No. 7, p. 151; Sintered-metal friction materials, Vol. 27, No. 11,
17.24 Jones & Laughlin Steel Corp., Wire Rope. is
p. 187; Friction brakes and clutches, Vol. 28, No. 7, p. 113; Multiple-disk
17.25 Dricker and Tachau, A new design criterion for wire rope, ASME Tré
clutches and brakes, Vol. 28, No. 5, p. 82.
Vol. 67, p. A-33. 18.18
na Harker and Rabins, Torque capacity of friction devices with nonlinear
17.26 Howe, J. F., Stresses in wire ropes, ASME Trans., Vol. 40. frictional elements, ASME Trans., Paper 62-WA-77.
17.27 Starkey and Cress, An analysis of critical stresses and mode of fai 18,19 Kotnik, R. L., Electromagnetic disc clutches, Machine Design, Vol. 32,
; e, ASME Paper 58-A-63. , , a
No. 16, p. 113.
17.28 Hilsher, J. E., Wire-rope assemblies, Machine Design, Vol. 33, No,
. 88. \ a 18.20 Walls, F. J., Brake drum materials, distributed by International Nickel
Vol. Co.
17.29 Rudoiph and Imse, Designing sprocket teeth, Machine Design, 18.21
: : 5 Armstrong Cork Co. pamphlet.
No. 3, p. 102. 18.22 Bette, A. J. Friction materials, Machine Design, Vol. 32, No. 20, p. 141.
Vol. 35, ee F
17.30 Lavoie, F. J., Tensioning V-belts, Machine Design, 18.23 Westover and Vroom, A variable-speed frictionometer Sor plastics, rubbers,
No. a
17.31 Renner, E. J., Belt pulley crown, Machine Design, Vol. 32, metals, and other materials, ASME Paper 62—WA-321.
belts, ASME Paper
17.32 Fazekas, G. A. G., On the lateral creep of flat 18,24 Courtel and Tichvinsky, ... Friction, Mechanical Engineering,
WA-54, J. Eng. for Ind. Vol. 85,
Nos. 9, 10, pp. 55, 33, respectively.
17.33 Jackson and Moreland, Roller and Silent Chains, pub. by Am. §
Chain Mfg. Asso. . ; 4
Design, Vol. 35, No. 17, Pp. 19,1 Jennings, C. H., Welding design, ASME Trans., Vol. 58, p. 497.
17.34 Kerr, R. W., Plastic roller chains, Machine 19.2 Wilson and Wilder, Fatigue tests of butt welds in structural steel plates,
17.35 Publications of the United Shoe Machinery Corp.
Univ. of Ill. Bull. No. 42.
Ill. E Procedure Handbook of Arc Welding Design and Practice, 11th ed., The
18.1 Schrader, H. J., The friction of railway brake shoes, Univ. of Lincoln Electric Co.
No. 72, Vol. 25.
 REFERENCES
559
19.4 Jefferson and Mackenzie, The Welding Encyclopedia, 14th ed., McGray 20.8 Hrones, J. A., An analysis of the dynamic forces
Hill.
> >

ASME Trans., Vol. 70, p. 473.

i

19.5 Udin, Funk, and Wulff, Welding for Engineers, Wiley. 20.9 ic gram ene
pone. D. A., Polydyne cam design, Machi
19.6 Phillips, A. L., ed., Welding Handbook, Am. Weld. Soc. ne Design, Vol. 25, Nos
> #, and 5,
19.7 Young, C. H., Experimental Study of Stresses in Welded Joints, Thesi 20.10 Rothbart, H. A., Cams, Wiley.
USN Postgraduate School. 20.11 Kloomok and Muffly, Determination
19.8 Smith, J. H., Stress-strain characteristics of welded joints, Jour.
of radius of curvature for radial and
soll Pitta follower cam systems, ASME Paper 55-SA-29
Weld. Soc., September, 1929. i , W. D., Cam design, Proceedings of Third Mechanisms Conference
19.9 Efficient Machine Design in Welded Steel, Lincoln Electric Co. a Purdue Univ., 1956.
19.10 Churchill and Austin, Weld Design, Prentice-Hall. 4 13 a é
we C. H., Cam radius of curvature, Machine
19.11 Clauser, H. R., Welding electrodes and rods, Materials and Metho Design, Vol. 34, No. 18
p. 123.
December, 1950. fale
19.12 Paton, B. E., ed., transl. from Russ., Electroslag Welding, Reinhold.
19.13 Nagaraja Rao and Tall, Residual stresses in welded plate, Weld. J., V
40, No. 10, p. 468s.
19.14 Yao and Munse, Low cycle fatigue of metals, Weld. J., Vol. 41, Now
p. 182s.
19.15 Alcoa Structural Handbook, Alum. Co. of Amer. t
19.16 Blodgett, O. W., Design of Weldments, James F. Lincoln Arc Wel
Found.
19.17 Cornwall, E. P., Welded joints for hard-vacuum systems, Machine Dest
Vol. 35, No. 19, p. 135. 2]
19.18 Franks and Wooding, Some dynamic mechanical properties of he
treated low-alloy weld deposits, Weld. J., Vol. 35, No. 6, p. 291s.
19.19 Mindlin, H., Fatigue Properties of Welded Joints of Aluminum-Magnesi
Alloys, Reynolds Metals.
19.20 Sparagen and Claussen, Fatigue strength of welded joints, Weld. J., V
16, p. 1.
19,21 Sparagen and Rosenthal, Fatigue strength of welded joints, Weld.
Vol. 21, p. 297s.
19.22 Toprac, A. A., An investigation of corner connections loaded in tens
Weld. J., Vol. 40, No. 11, p. 521s. 1
19.23 Choquet, Krivobok, and Welter, Effects of prestressing on fatigue stret
of spot-welded stainless steels, Weld J., Vol. 33, No. 10, p. 509s. ;
19.24 Aluminum... . Product Information, Kaiser Alum. & Chem. Co.
I

Windenberg, D. F., Vessels under external pressure, Mech. Eng., Aug

1937,
Saunders and Windenberg, Strength of thin cylindrical shells ut
external pressure, ASME Trans., Vol. 53, p. 207 APM.
Jasper & Sullivan, The collapsing strength of steel tubes, ASME Tra
Vol. 53, p. 219 APM.
Bryan, G. H., Application of the energy test to the collapse of a long’
pipe under external pressure, Cambridge Phil. Soc. Proc:, Vol. VI, p.2
20.5 Stewart, R. T., Collapsing pressure of Bessemer steel lap-welded ti
ASME Trans., Vol. 27, p. 730.
20.6 Timoshenko, S., Theory of Plates and Shells, McGraw-Hill.
20.7 Mitchell, D. B., Tests on dynamic response of cam-follower syste
Mechanical Engineering, Vol. 72, p. 467. .
 APPENDIX

The order of arrangement

.» is the same as that of
the text’s subject matter.

icaihtellaoes'g wobec
ll IRAN
oh eae = tir 61 Sie ORD ore. ts ape a 8 ope Ue a
 AT 1 TABLE AT1 PROPERTIES OF SECTIONS
the axis x-x, J=polar moment of inertia about the cen-
rectangular section modulus, about x-x, Z’ = J/c = polar
section modulus,
k = V I/area = radiusof gyration.

The ability to understanda question from all sides

meant one was totally unfit for action. Fanatical
enthusiasm was the mark of the real man.
Thucydides on the Athenian mood during
H
i
fi

NSO
SRW
s* we Yy
LZ. Zz

Centroid
aH *+ bt
a=
2aH+ bt)’ a=H—o,

Be (Boe +Sah? + (abe

LaA5+
Area=Bi+a(H-t); k= 7A

©
ED = Oneal ee 9 oo ee eees
Fb. = ice = pounds per inch of length; F = wL, where L in. = t
ee Ale
@ radians = slope. For other beams of uniform strength, see § 6.24.

UNIFORM LOAD

rr a Wowk a
Yaw =F Eo) — M Vp Yow SET v P Ye

bie : oe eee te
—I Shear } yp Mew =F F Prec te pens at eaeas
F Shear Diagram ie L nl Fa_|p, M.=—Fa,
4 ‘ op L a<a<atb: R= Fk R=F

a To z=a: | M= Mos.
ea 3— iei ; R oresears Shear = Fa. Shear F
Bending eee | Mau Mon vent jae | Le ashumenmae
Moment Me=Fr 7, eek: 4. =a. 520 Oe |
z=: Mew. =—FL z=0: 0=—— ; 2 2 M, = 2% i a i |
2EI =—”_w (y'—4]D2 +31) L Morax.
F
¥=—aT (e—31’2+2L) v" aE” —* ‘oment
Fo?
Yoo = Sey (a+ a | Moment /

Fo_[a(L+B)]*
You = 3277 | 3 | a>b $= ie [3L4/4—a? ] , center relative to ends

FL
Yuu, —— Ti] 0< Rigi *
48EI ate
ser os BP as
¥—z)

be Fs a at i A holy UNIFORM STRENGTH, CANTILEVER UNIFORM STRENGTH, SIMPLE BEAM

L
> Moe. =
FL= z=0: O= — TET
FL fie cae
and: Mow = Fh Oe z=0: @=— ; fe Vertex é Vertex re eas
Towed, yah (alte—40" y= 2 5-2 +2) “3 Tie ie
Z End View Parabole End View eal ae

FIXED ENDS L——, UNIFORM

LOAD : Le =a M.=Fr= ©!. For
For e= s=C, F R, wep7 Ms att.
5 tt= Fors gol (et
1 Use FL yy — FP-2
4+ —+4 Z aa 6F__
Goren Pa 3F_ 2x
—r=—h?

Toz=> L F
wast ME :, — Tiss
At w= 0: Yom AtEmD: Yam —Fa
. 2 er - sheet
Viex2 3 V.=
wh ‘Wr
v
7 - f wl?
Fz FL_y.. 7
8 2
M,=—— + —

a foment i2 Moment

wl? , whe
_ wat
Mn ae
 OF SOME NONFERROUS METALS®:?:2:3.2-8-2-9.2.121
AT 3 For Aluminum alloys, let Poisson’s ratio » = 0.33; torsional ult. = 0.65s,; strength in reversed torsion, 25 ksi. (i) « in./in.—°F,
yield str. = 0.55s,. Extruded forms over j in. will have s, about 15% greater. (i) Varies with size of test specimen. (k) Yield point in compression;
alloy AZ91C-T6, 19 ksi;
For Magnesium. alloys, \et the flexural strength (symmetric sections) be the ave: AZ614-F, 19 ksi; AZ80A-TS, 28 ksi, (1) Chill cast. (m) In compression, at
0.125%, set.
of the tensile and compressive strengths. See note (k). Let » = 0.35. (n) At 1200°F, after water quench and aging. (0)
tural shapes. (q)
Estimated. (p) Used for rolled struc-
Abbreviations: H, hard; }H, } hard; H14, temper designation meaning 4 hard; For manganese gear bronze, use $, = 17ksi. (r)
For phosphor gear
heat treated; T4, temper designation meaning solution heat treated. bronze, SAE 65, use s, = 24 ksi. (s) Normal temp.;
see § 2.21 (t) Sheet.
. (u) About 17.5 in
Notes: (a) At 0.5% total elongation under load. (b) Cold reduction of 11%. (c) compression. (v) Water quenched and aged at 975°
After 1 yr. (y) Pure platinum. (z) Die castings.
F. (W) Suc = 87 ksi, Charpy = 48. (x)
0.2% offset. (d) BHN. (e) BHN with 500kg. load. (f) Minimum. (g) Flat 0.04in. thick, (h)
=
MATERIAL CONDITION ULT. STR., TEN YD. END. STR., MOD. MOD, ELONG. %| ROCK. |DENSITY |« x 10°® PER’
(ASTM No.) ksi Sy ksi 5, ksi at EL. EL.SH. | 2in. |marp. | ibjin.® | (i) CANSAGES OF SLERAINTS
Su Sus No. cycles E x 10-*|G x 10-* Vj)
zz
‘| COPPER ALLOYS
Admiralty brass (B111) Annealed 53 22(a) 18 at 107 15 5.8 65 F75 | 0.308 | 11.2 71 Cu, 28 Zn, 1 Sn
Aluminum bronze (B150-2) Extruded rod (b) | 120 70(a) 32 at 3 x 108 16 6.5 12 B100} 0.274 | 9 81.5 Cu, 9.5 Al, 5 Ni, 2.5 Fe, 1 Mn
Aluminum bronze (B150-1) .| Annealed, 800°F | 100 28 at 8 x 107 15 6.5 25 | B90] 0.274| 92 |o1cusal ti‘
Beryllium copper (B194) . HT 175 130 35 at 10°(h) 19 7.3 5 C37 | 0.297 | 9.3 1.9 Be, 0.2 Ni or Co
Cartridge brass (B134-6) . .| $H, rod 70 42:«| «= «52(a) 22 at 5 x 10" 16 6 30 B80 | 0.308 | 11.1 70 Cu, 30 Zn
Commercial bronze (B134-2) .| Spring H(g) 72 42 | 62(a) 21 at 15 x 10° 17 6.4 3 B78 | 0.318 | 10.2 90 Cu, 10 Zn
Free-cutting brass (B16) . 2” Rod, }H 55 32_-«| «44(a) 20 at 10° 14 5.3 32 B75 | 0.307 | 11.4 61.5 Cu, 35.5 Zn, 3 Pb
Manganese bronze (B138-A) .| $H, rod 75 48 | 65(a) 30 at 10°(g)(q) 16 6 25 B80 | 0.302 | 11.8 58 Cu, 39 Zn, plus
Naval brass (B124-3) }H, rod 70 = 43:—«| :« 48(a) 15 at 3 x 10 15 5.6 25 B80
Nickel silverB . . - H(g) 100 85(a) 19 18 3 Bol || 0.304
0314 || 11.8
93 5$60 Cu,
Cu, 2739.25Zn, Zn,18 0.75
Ni Sn
Phosphor bronze (B139C) }H, rod 80 65(a) 31 at 10°(r) 16 6 33 B85 | 0.318 | 10.1 92 Cu, 8 Sn
Silicon bronze (B98-B) H, 1” rod 65 45 35(a) 25 at 10° 17 6.4 10 B80 | 0.316 | 9.9 97.7 Cu, 1.5 Si, plus
Yellow brass (B36-8) 3H, 1” rod 55 36 | 40(a) 11 at 10°(z) 15 5.6 48 BSS | 0.306 | 11.3 65 Cu, 35 Zn
Yellow brass (B36-8) 3H (g) 61 40 | 50(a) 15 5.6 23 B70 | 0.306 | 11.3 65 Cu, 35-Zn
ALUMINUM ALLOYS {
3003-H14 (M1A) Strain hard. 22. «14 | 21) 9 at 5 x 10° 10 3.85 16 40(e)| 0.099 | 12.9 1.0'Mn; others
2014-T6 (CS41A) HT, aged 70 42 «|| 60(c) 18 at 5 x 10° 10.6 4.0 13° | 135(e)| 0.101 | 12.8 3.9 Cu, 0.5 Si, 0.4 Mn, 0.2 Mg
2024-T4 (CG42A)_. HT, aged 68 41 | 47(c)(p) | 20at5 x 108 10.6 4.0 20 | 120(e)| 0.098 | 13.0 3.8 Cu, 1.2 Mg, 0.3 Mn
6061-T6 (p) (GS11A) HT, aged 45 30 | 40(c) 14 at 5 x 10° 10 3.75 17 95(e)| 0.100 | 12.7 0.15 Cu, 0.8 Mg, 0.4 Si
7075-16 (ZG62A) HT, aged 82 48 | 72(c) 23 at 5 x 10° 10.4 3.9 10 | 150(e)| 0.101 | 12.9 5.1 Zn, 2.1 Mg, 1.2
SB: ets Die casting 40 27 | 24(c) 17 at 5 x 10° 10.3 3.85 1.8 | 70(e)| 0.095 | 11.7 9 Si, 0.4 Mg Seen
355-T6 . Sand casting S5:28_1..25Ke) 9 at 5 x 10° 10.3 3.85 3 80(e)| 0.098 | 11.7 1 Cu, 4.5 Si, 0.4 Mg
MAGNESIUM ALLOYS 0 3
AZ61A-F . Extruded bar 45 20 | 33(c,k) | 17 at 5 x 10® 6.5 2.40 16 E72 | 0.065 | 14.4 6 Al, 1 Zn, 0.2 Mn
AZ80A-TS . Forged, aged 50 23 | 36(c,k) | 16at5 x 108 6.5 2.4 6 E82 | 0.065 | 14.4 8.5 Al, 0.5 Zn, 0.15 Mn
AZ9IC-T6 . Sand casting 40 19(c,k) | 14 at5 x 10°@ 6.5 2.4 5 E77 | 0.066 | 14.4 9 Al, 0.7 Zn, 0.2 Mn
LEAD AND TIN ALLOYS 4 4
Babbitt (B23-46T-8) At 68°F (I) 10 sy. = 3.4 (m) | 3.9 at2 x 10 2 5 20(e)} 0.36 | 13.3 80 Pb, 15 Sb, 5 Sn
Babbitt (B23-46T-8) . At 212°F (1) 54 Sye=1.7 (m) s , 27 10(e)| 0.36 | 133 80 Pb, 15 Sb, 5 Sn
Tin babbitt (B23-49-1) At 68°F (I) 9.3 Sy =44 (m) | 3.8 at2 x 107 2 2 17%(e)| 0.265 91.80, 45'8b 4.5 Ca
MISCELLANEOUS 26
Hastelloy B Sand cast 90 50(c) | 66 at 10°(n) sei 10 B93 | 0.334 | 5.55
7.8
| 6266 Ni,
Ni, 2928 Cu,
Mo, 3 5 AlFe
K Monel Cold dr., aged 140 98 | 100(c) 42 at 10° 20 C30 | 0.306 |
Platinum alloy Annealed 45 35 90(e)} 0.722 | 5.0(y) | 10 Rhodium

Titanium (B265, gr. 5) (t) Annealed (s) 135 130(c) 42 (f) ae 12 0.160 | 5.8 6AL4V
Titanium (B265, gr. 5) (©) Hardened (s)(v) | 170 158(c) | 61 (f)) (u) 7 0.160 | 5.8 { Hi. temp. aero. service
Zinc(AC41A) - - - Die cast (w) 4716 38 8 at 10° 2(x) 7 91(e)| 0.24 | 15.2 4 Al, 1 Cu, 0.04 Mg
AT4 OF SOME STAINLESS STEELS ?-1:2-3-2-22.2.18)

Notes: (a) Coef. of thermal expan room temp., « in./in.-"F. (b) Approx.
nearsion g of about 160 ksil-18] (f) Varies some with condition: annealed, cold work: stress
values of ult. strength of 403, 410, and 416 are given by s, = 5 + 0.465 (BHIN) in ksi. relieved. In shear, for cold drawn spring wires, G ~ 10.6 x 10° pal, (g) 0.050-n. ship
(c) Varies with details of heat treatment and cold working. (d) Cold worked, full 1 r
ened; Republic Steel TH 1050;
guaranteed Sumin=170 ksi; sn at 10°.
(e) Endurance limits for stainless steels may be estimated at 0.4s,, up to tensile st eng

ULT. STR.| TEN. YD.| END. | MOD. EL. | ELONG.| RE - Se eb

MATERIAL Sy ksi | Sy ksi | LM. s, |E x 10-* | 2in.y% (aver.)| Density | ft-lb. | « x 108 REMAR
AISI No. (©) (o) ) i) (o) () | bin? | () (a) or

301,fhard . .
.
.|
.|
125(h) | 75(h) | 30)|
90 | 37 | 34 2828 | 25(h)
(57 ~ 0.286 | 90 |
150 || 0.286
260 9.4 || (17%
9.6 Austenitic. Ni) Generalby use;
Cr, 7%Hardenable work. structural
cold trim,
302,anneled .
302,dhard . . .|125(g(h)| 75(h) | 70)| 28 12(h) 260 | 0.286 9.6 | 302, 303 are 18-8 stainless steels
303, annealed <1 90; 1).38) | "98 28 50 160 | 0.286 | 80 | 9.6 | Austenitic. Hardenable by cold work.
304,annealed . . .| 85 | 35 28 50 150 | 0.286 | 110 | 9.6: | Austenitic. Hard enabl
316, : cold worked (i). -| 90 | 60 | 40 28 45 190 | :
0.286 ois ustenitic. duane odie
|tuaaieniel) 2a are eae
321,annealed . . -| 87 | 35 38 28 50 150 | 0.290 | 110 | 9.3 | Stabilized by Ti.
347,annealed - . -| 9%” 40 39 28 50 160 | 0.286 | 100 9.3 | Austenitic. Hardenable by cold work.
403, 410, heat treated (b) | 110(h) | 85(h) | 58 29 20 225 | 0.279 | 75 | 5.7 | Martensitic. Hard
410, cold worked (b) .| 100(h)| 85 | 53 29 17 ms | ox | ao. || 59) | sketaunk. Wee teenie
416, annealed (b) . .| 75 | 40 | 40 29 30 155 | 0.278 | 70 | 5.7 | Martensitic. Hardenable by HT.
430,annealed . . .| 75 | 45 | 40 29 25 160 | 0.277 | 35 | 5.8 | Ferritic. Not hardened by HT.
431, OQT 1000(b) - -| 150 | 130 29 18 325 | 0.28 50 | 65 | Martensitic. Hardened by HT to high strength.
A REPRTOOLD! X98 joy HEISE HERES) TES 2 390 | 0.276 5.6 | (17% Cr, 7% Ni, 1.15% Al) Solution annealed,
etc.

AT 5 TABLE ATS TYPICAL PROPERTIES

OF A FEW PLASTICS ?2:19:2-23.2-271

Notes: (a) TS, thermosetting; TP, thermoplastic. (b) National Electrical Mfg. A: (j) 48-hr. immersi (k) L, laminated;
on. M, molded. (1) When used for gears, let s, = 6000
grades. (c) Flatwise. (d) For } to 1-in. dia. Reduce 15% for sizes 1 to 2 in. (e) For $ psi. (m) Yield strength. (n) Unplasticized. (p) General purpose. (q) At 1% def. (r) At
dia. Reduce 10% for sizes 1 to 2 in. (f) Min. values. (g) Bending strength, sy : failure. (s) At 73°F. (t) At 5% strain. (u) Rupture.
sections. (h) Specific gravity. (i) Aver. water absorption, 24-hr., j-in. thickness, per cé

MATERIAL TYPE | CONDITION | ULT. STR.,| COMP. ULT. ELONG. 1zop, | %H,0
(a) (k) sy ksi STR., Sue | Sy % |ROCK.| MOD. EL.,|sP.GR.| ff-lb.| Ab- A FEW TRADE NAMES®
me. (r) |HaRD.|E x 10-5| (A) (f) | sorp.(i)

Phenol-formaldehyde
GradeX(b)() «= - «| TS | L. sheet | 14 35 M100] 4-20 | 1.35 | 1.3(c)] 1.4
Grade XX (6) (I) TS | L.rod 8.5(d) | 20 M100 | 4-20 | 1.35 | 1.0(c)} 0.65 || Bakelite, Durez, Formica
C (b) (1) .
Grade Ts | L.rod 7.5(d) | 20 M100 | 3.5-15 | 1.35 | 3.2(c)] 1.2 | Textolite, Micarta, Synthane
GradeA(b)() . «| TS | L.rod | 6d) | 15 M90 | 3.5-15 | 1.65 | 1.8(c)| 0.65 || Durite. ; :
Wood flour filler (p) TS M 6(f) 24 0.4-0.8 |M100| 10 14 | 04 0:8
Urea-formaldehyde . . | TS M 9 25 0.6|M1I8| 15 1.45 | 0.24 | 0.4 | Beetle, Sylplast, Plaskon
paiyviel chloride
Polyvinyl device Gute . .| TP | MM 88 1013 30 |Res | 3 1.2 |0.8 | 0.05 | Geon, Vinylite, Marvinol
10 |mM70 | 8 1.41104 | 0.1 | Exon, Pliovic, Ultron. —
Polymethyl Methacrylate . | TP | M 8 14 8 |Mioo] 4 1.16 | 0.4 | 0.3G)| Lucite, Plexigiass, Perspex.
Polystyrene (f) . +. ~- | TP M 5 115 1.2|M85
60(s) | R118
| 05
3.5 | 0.9%(s)|| 0.03
106]0.2
| 1.14 1.5 ho ee Oe Plioli
yrene, , Pliolite.
Polyamide(m) - ~. ~| TP M 11.8(s) 4.9(q)
Cellulose Acetate... . | TP M 4.5 200.4(m) 20 | |R100|
30-500 R11 0.15 1.27
2 | 0.92 |4 | 1.5-2.9]
0.01 | Dylan, Celanese,
Plastacele,Alathon, Kodapak :
Orizon,
Polyethelene(f) . . . | TP M 1.7
Polytetrafluoroethylene (m) TP M 3.8(u) 1.8(t) 100-200} R20 0.6 22 |24 | no
Polyvinylidene chloride . | TP M 5 2.4 200; fuente25
|RIO| |
iy
21 |4
fo? | none
ol tae
| Kel-F, Fluorothene.
Polychlorotrifluoroethylene TP M 6 5 200
AT 6 popcan Re RRL hrs naioeer lienOF |
gray iron, 0.4s, < S; < 0.6s,. (j) The number indicat
Notes: Approximate coefficients thermal expansion, in./in.-°F are: gray iron, 5.6. es minimum propertie
s; e.g.,
80-60-03 indicates 5, = 80 ksi, s, = 60 ksi (0.2% offset), and 3%
fron, 6.6 x Ferg iron, 6.7 x 10-°; cast steel, 6.5 * 10-° ( elongation, minimum,
ae ee
(nin):male
varies significantly with composition). Poisson’s ratio: gray andiron,SAE0.211spies
in approximately 1-in. section. (k) 0.3% C, max. (1) N&T, normali
zed and tempered. Pro-
iron, 0.265; nodular iron, 0.16; cast steel, 0.27. (a) ASTM ao m
identical. (b) Machinability, relative values, AISI B1112 = 100%. (c) .
supports. (d) Test results suggest that the flexural strength of cast arc Pde S6T. (r) ASTM A396-58. (s) For design, assume compress
sections, computed from s; = M/Z, is about 1.9s, to 2s,. Use 1.9s,. (e) ee ive ultimate and yield strengths
of nodular iron to be the same as s, and s,, respectively.
a (t) Common commercial grades.
Minimum values. Typical values may range 10-40% higher. (g) ASTM 35 (a) Tempered at 1200°F. (vy) Typical s, ~ 96 ksi, 5, = 73 ksi
considered to be high-strength, and are definitely more expensive. (h) For cast iron, 25 when WQT 1200. (w)
General purpose type.
of ult. stress; varies with section size and chemical analysis. (i) Reversed bending. F
ksi TRANSV. |END. | TEN. MOD. SHEAR
TERIAL, SPEC. NO. sc sass: STRENGTH | LIM. | YD., ELAS., psi | MOD., psi | BHN 1zOD DENSITY | MCH.
sy ks
REC. MIN.
ip ‘ ; Su San Sus | Tors.| 1b. (c) | Say ksi Ex 10-* |G x 10-¢ St-lb. Ibjin.® (6) WALL THICK.
@
() i (bh)
GRAY IRON (g) (as cast) (d) (d)
9.6(f) 3.9(f) 156 0.253
abit oe 20(f) 83 32 | 26 1850 10 j t=} in.
i‘ 11.5(f) 4.6(f) 174 0.253
25 5-5 7Hie 25(f) 97 35 | 32 2175 ee t=} in.
a 13(f) 5.2(f) 201 23 0.254 80
30 ill 30(f) 109 41 | 40 2525 t= Zin
14.5(f) 5.8(f) 212 25 0.257 65 t =
35(f) 124 49 | 48.5 2850 f Zin.
16(f) 6.4(f) 235 31 0.262 55
on a 40(f) 140 52 | 57 3175 18.5 t= Zin.
18.8(f) 7.2(f) 262 65 0.269 50
50(g) 50(f) 164 64 | 73 3600 21.5 t = in.
20.4(f) 7.8(f) 302 75 0.269
60(g) : 60(f) 187 60 | 88.5 3700 24.5 t= Jin.
12(f) 145 100 t = fin.
Ni-Resist, Inco K-6 25(f) 100(f) 12(f) 190 t= }-Tin.
Meehanite, (w) 35(f)
MALLEABLE IRON ELONG. RED.
ASTM Grade 2in., % | Area, %
pos
25 10.7 120 12 0.262 120 12.5
A47-52 32 510 52 (0) 48 | 58
25 10.7 130 16 0.262 120
A47-52 35 018 55 (0) 43 | 58 20

NODULAR CAST IRON (j) "

23 9.5 160 | 9-20(n) 0.26
60-45-10 (annealed) (q). | 70 (s) : - 5 18
ye a 23 9.9 230 | 2-8(n) 0.26 6
80-60-03 (as cast)(p) . | 88 (s) ;
g 23 9.9 270 | 2-6(n) 5
100-70-03 (heat treat) (r) | 110 (s) 88(e
MAX. CARBON i
CAST STEEL AND HEAT TREAT.
STM SAE(a) 30 11.5
C,
60(f) | 0.3% eat = 120 | 18(m) 0.284 55 30 50
peer 3 ths 60(f) 30
re 11.5 130 | 23(m) 0.284 60 30
A27-58(t) 0030(k) 65(f) 65(f) | 0.3% C, Normal
30 53
11.5 140 | 19(m) 0.284 65 26
A27-58 . 70(f) 70(f) | 0.35% C, Normalized 40
30 11.5 30(m) 0.284
A27-58 70(£) 70(£) | 0.25% C, Normalized

A148-58 080 -— nae ee 5 :0 30 11.5 160 22(m) 0.284 70 27 42

30 11.5 170 | 30(m)(u) | 0.284
A148-58 Z p 28(u) 68(u)
6 30 11.5 190 20(m)(u) 0.284
090 90(f)(v)| 90(F) Sea ay 70 24 50
Al148-58 30 11.5 235
WwQ a 28(m) 0.284 60 18 42
Al148-58 0105 105(f) 105(f)
30 11.5 269 | 25(m) 0.284 14(f)
Al48-58 . : 120(f) 120(f) WQT = ; )
30 30(f)
Ge »
11.5 310 0.284 9(f) 22(f)
0150 150(f) 150(f) aie
Al48-58 30 11.5 390 | 12(m) 0.284 8
Al48-58 0175 175(f) 175(f) WwQ 15

571
 size Acy 1360°F. |Ar, 1268'F, A me ;
stan reste | acy 1420°F. | Ar, 1200°F.| 249 ast ee" Kaos
Ac; 1340° F. Ars 1350°F. We
.505"Rd. SIZE TESTED | Ac, 1445°F. Ar; 1250" F. 1.20 08 | 6-8
BRIN.
‘SHORE
ROCK,
IzoD.

230000
220000
210000
200000
190000
180000
170000
160000
150000 21
140000 200000
130000 190000
120000 180000
110000 70000
100000 60000
90000 150000
40000
130000
120000
110000
100000
90000

|
80000
70000 0% |
60000
soo'r. 500°F, G00°F, 700°F. 800°F. 900°F, 1000°F, 1100°F, 1200°F, 40°F, SOO'F. GOO'F. 700°. 800'F. 900F, 1000'F, 1100°F. 1200'F, 1300'F,
AT 1650°F., REHEATED TO 1575'F., QUENCHED IN AT 1600 F., REHEATED TO 1525'F., QUENCHED IN AGITATED
on,
FIGURE AF 1 Properties of Heat-treated AISI C1040. Abscissa is the FIGURE AF2 Properties of Heat-treated
AISI 3140. Single heat results. (Draw =
perature. Average values. Such charts as these are guides to mechanical Temper.) Notice the specified heat treatment
may be expected when the diameter is, say, { in. to 1} in. See Tables AT 8
and the size of the specim The en.
ultimate
Observe that the BHI increases with tensile strength and that 5. ~ widely used for heat-treated parts.
For Re = 28, decarburization of surface reduces endurance
so accurate at very low tempering temperatures. (Courtesy Bethlehem Steel Co., strength by 50%. For R. = 48,
reduces endurance strength by 75%, from
Pa.). but the percentage is unusually high. See about r=
Fig. AF 3. (Courtesy alia s Sarco
Bethlehem, Pa.).
 OlL QUENCHED NHVISC Maldhess

FL. Lb,
impact,
20 30 40 50 70 80 90
19 | 22 24 28 45 | 55) 65 | 75|85) 95
z: gAge 100
& wo «3 4
J z 8 1s0

3 - :

Tensile Strength —-Thousand Pounds per Square Inch

37 i .

L oe

200 400 600 800 1000 1200 1400 °

Tempering Temperature - °F

Brinell Hardness
STEEL NO. 4340
Ol QUENCHED 260,000

=~

So
S
SIZES 240,000
aT
‘Tensile Stiength and Yield Point, 1000 Lb. per Sq. In.

oOo
oS
Elong. in 2" and Red. of Area, Per Cent

600
SB. 8 $ &

700

AFE3 Properties of Heat-treated AISI 4340. Another manner in which

FIGUR ec!
D
properties might be depicted; charts A, B, and C are tensile properties; chart
purpose alloy. Other miscellane ous
sional properties. An excellent general
strengths of this steel follow. For su ~ 270 ksi:
; 800
Sur decarburize
notfac ed, sn = 89 ksi. Surface decarburized to 0.03 in., sn = 4 -10 0 10 20 30 40 50 60 70
Decarburized surface, shot peened, sn = 95 ksi. OQT 1075, 0.625 dia., Rockwell “C” Hardness
ksi.
sn = 120 ksi. Test specimen from rolled stock, transversely, sn = 45-70
FIGURE AF 4 Relation of Hardness Numbers. (Courtesy International Nickel
For su © 160 ksi; Temp. Reversed s,, R = —1 (ksi) Repeated, R = Co., N.Y.)
variation of 10°F 70 uy
endurance strengths 600°F 61 "2 ,
with temperature. 800°F 59 = q
1000°F 39 65

(Courtesy International Nickel Co., N.Y.)

 BS onereses 6B SSR AWSRAS BROWS BURN SEE WE

AT 7 WROUGHT FERROUS METALS'!-1:2-1,2.8.2.6,2.10]

(See also charts for C1040, 3140, 4340; and Tables AT 8-AT 10, incl.).
Density is about 0.284 Ib/in.° (0.28 Ib/in.° for wrought iron).
For all wrought steels: Coefficient of thermal expansion (linear) is 0.000 007 in./in.-°F (0.000 0006 5 f
Modulus of elastinicity tension or compression, E = 30 x 10° psi (For Wrought iron, — Varies significantly with large temperature change. See § 2.22 for ak pers
E = 28 x 10° psi).
or torsion, G = 11.5 x 10° psi (For wrought iron, G =
Modulus of elasticity in shear
Notes: B preceding AISI No. indicates Bessemer, as B1113; C preceding indicates o
hearth, as C1020. (a) Minimum values. (b) Annealed. (c) QT 1000 stands for sebdacead
10 x 10° psi).
0.6s,. and tempered at 1000°F,”’ etc. (d) Ultimate stress in shear has been arbitrarily taken as
Yield strength in torsion (shear) frequently falls between 0.5s, and 0.6s,. Use
specimen, reversed bending, approximat ely s,/2; Table A’ (.75) (tensile ultimate); except starred * values which are test values. (e) 1-in. specimen.
Endurance limit of a polished
(ff) Torsion. (g) Mill annealed. (h) Cold drawn. (i) See § 2.2 for definition. (j) Charpy V-
Endurance limit in reversed torsion, approximately 0.6s,; see § 4.7.
notch, 70°F. (k) Properties depend on amount of cold work.
Poisson’s ratio is about 0.25 to 0.33. Use 0.3 for steel.
IZOD | MACHIN-
MATERIAL CONDITION ULT. STR., ksi_| TEN. YD.| ELONG.| RED.
ROCK. | ft-lb. | ABILITY SOME TYPICAL USES; REMARKS
sd) | sy, ksi | 2in. %| Area %
AISI No. (oc) ie
@
Wrought Iron} As rolled 48(a) | 36 25(a) | 35 B60 50(h) | ASTM A85-36, A41-36

Wrought Steel
C1010 (k) Cold drawn 67 50 55 25 57 i 50 Bars, strips, sheet, plate. Cold drawn sha:
> ’ 3 < pes.

C1015 (k) Cold drawn 77 58 63 25 63 3 137(j)| 50 Bars, sheets. Table AT 8. For carburizing: Table AT 11.
C1020 As rolled 65 49 48 36 59 79 64 64 Structural steel; plate, sheet, strip, wire.
Normalized 64 54* 50 39 69 B74 72 Carburizing grade, Table AT 11.
C1020
C1020 Annealed 57 43 42 36.5 66 Ee 80 General purpose.
Cold drawn 78 58 66 20 59 62 Misc. machine parts are cold forged; bars
C1020 (k)
C1022 As rolled 72 54 52 35 67 B81 60 70(h) | General purpose. . ;
C1030 As rolled 80 60 51 32 56 B88 55 60 Machinery parts. Table AT 8.

C1035 As-rolled 85 64 55 29 B91 45 57 Machine parts. . Ma: May be heat treated.

C1045 As rolled 96 72 59 22 45 B96 30 51 Large shafts. fener
C1095 Normalized 141 105 80 8 C25 3 39 Tools, springs. Usually heat treated. Table AT 9.
B1113 (k) Cold finished 83 62 72 14 B87 135 Free cutting; high sulfur.

B1113 As rolled 70 45 25 ps : Free cutting; high sulfur.

C1118 As rolled 75 56 46 32 82 Free cutting; not usually welded. Carburized, T:
C1118 (k) Cold drawn 80 60 75 16 110(j)| 85 Table AT 8 for C1117. ea:
C1144 OQT 1000 118 88 83 19 C22 36 65(b) | Free cutting. High sulfur. Tables AT 8 and AT 9 for C1137.

1340 OQT 1200 113 84 92 21 ome 95 45(g) | (1.75% Mn). Manganese steel.
13B45 OQT 800 187 140 175 16 1345 with boron for improved hardenability.
2317(e) OQT 1000 106 79 71 27 B97 85 55(h) | (34% Ni)—Gears, pump liners, etc.
2340(e) OQT 1000 137 103 120 22 C30 50 31 (34% Ni)—Gears, etc.

3150 OQT 1000 151 113 130 16 C32 46 ‘

(1.25% Ni, » 0.8%
0. Cr.)A Gears, , b bolts, shafts, etc.
3250(e) QT 1000 166 122 146 16 C36.5 30 55(b) | (1.85% Ni, 1.05% Cr) Gears, etc.
4063 OQT 1000 180 135 160 14 pe 59 (0.25% Mo) Shafts, bars, etc.
4130 WOQT 1100 127 95 114 18 25 85 65(b) | (0.95% Cr, 0.20% Mo) Shafts, forgings, pins, aircraft tubes.

4130(e) Cold drawn 122 91 105 16 45(g) | (1.85% Ni, 0.8% Cr, 0.25% Mo) General pu Fi
4340(e) Cold drawn 122 91 105 15 c33. | 41 | 5506) | (1.85% Ni, 0.25% Mo). Bape
4640(e) OQT 1000 152 104 130 19 C32 60(b) | (0.80% Cr) Gears, shafts, pins, etc.
OQT 1000 150 113 128 19 60(g)
5140(e)
Cold drawn 105 79 88 18 i (0.55% Ni, 0.5% Cr, 0.2% Mo). Table AT 9.
5140(e)
Cold drawn 10% | 115 86 100 22 o 36 60(b) | (0.55% Ni, 0.50% Cr, 0.20% Mo).
8630
OQT 1000 160 120 150 16 Be 19 50(b) | (0.55)% Ni, 0.50% Cr, 0.25% Mo) Tools, gears, bolts.
8640
OQT 800 220 165 200 12 36 7 45(b) | (2.00% Si, 0.82% Mn) Springs, chisels, tools.
8760
9255 OQT 1000 180 135 160 15 C33 73 | 60(b) | (0.45% Ni, 0.4% Cr, 0.11% Mo)
9440 OQT 1000 152 104 135 18 C37 50 | 50(b) | (1% Ni 0.8% Mn, 0.8% Cr. 0.25% Mo) Heavy duty;
OQT 1100 180 135 158 15 general.
9850

ox
577
 SSRRREs SRR OO ZR SR ANSE FERNS SUNS SEN WR VASES
HEAT-TREATED STEELS®-3.2-51
AT8 VARIOUS SIZES AND CONDITIONS®-*)
(a) Turned. (b) 10%. (c) Inconsistent—from different mill.
Values in this table have been taken from
the strength or Brinell number for any
charts such as those of Figs. AF 1-AF
3. To get
other tempering (drawing) temperatur
straight line interpolation between the e, make a
ROD DIA.|ULT. STR.) TEN YD.| ELONG. | RED. | BHN 0 values given. Extrapolation to lower tempe
might sometimes give a reasonable estima ratures
a CONDITION in. Syksi | s,ksi| 2 in., %| Area, % l te, but extrapolation should not be relied
(a) Do not interpolate using this value. upon.
C1015 As rolled 4 61 45.5 39 61 126 b AISI NO.
Annealed 1 56 42 37 69.7 | 111 83. TEM-
(Quenching SIZE | PERED |ULT. STR.|
Normalized 4 63 48 38.6 71 126 §3) TEN.YD. BHN ELONG. IZOD,
medium) AT, °F Su ksi Sy ksi 2 in. % ft-lb.
Normalized 1 61.5 47 37 69.6 | 121 85.
Normalized 2 60 44.5 37.5 69.2 | 1 7 C1035 1 600 118 87 240 11
Normalized 4 59 41.8 36.5 67.8 | 11 (water) i 40
1000 102 73 200 22 57
i 1300 85 37
led 70.6 44.3 33 63 143 60 170 29 93
ri! ie : 62 40.5 32.8 58 121 a C1095 4 800 176 112 363 11
Normalized $ 69.7 45 34.3 61 \ y (oil) 4 6
1100 145 88 293
Normalized 2 67 41.5 33.5 64. 7 a 17 6
4” 1100 130 65
Normalized 4 63.7 35 34.3 64.7 | 1 4 262 17 5
C1137 4 700 135 115 277 12
C1030 As rolled $ 80 51 32 4 a79 555 (oil) 4" 1000
13(a)
111 88 229 23
Annealed 1 67 49 31 57. , oe 61
1000 105 63 217
Normalized 4 77.5 50 32 pte pe a 23 31
Normalized 4 72.5 47 29.7 56.2 } 2330 a 600 210 195 429 13
WOQT 1000 1 88 68 28 68.6 | 179 Nickel 4 39
1000 135 126 277
Steel 20 77
4 1300 107 91
C1137 As rolled $ 93 55 26 i aw (water) 217 26 109
4” 1000 105 85
Annealed i 85 50 27 pi 7 207 26 87
Normalized 4 98 58 25 e a 4140 a 500 270 241
Cr-Mo 534 11 8(a)
Normalized 2 96 49 22 . a. 4 800 210 195 429 15 21
Cold drawn 1 103 93 15 yl (oil) Pd 1200 130 115 277 21
4’ 83
1200 112 83 229
C1045 Annealed 1 90 55 27 : a 23 87
Normalized 1 99 61 25 pd 4150 4" 800 228 215 444 10
Hot rolled(a) 1 87 54 27 : ne 5 Cr-Mo (oil) 4” 12(a)
1200 159 141 331
Cold drawn(b)} 2 100 85 19 16 53(a)
5150 4" 800
WOQT 1000 4 130 110 16 56 a 210 195 415 1l
Chromium 4 17(a)
WOQT 1200 $ 110 84 23 61 oa 1000 160 149 321
(oil) 15 39
Pd 1200 127 117
WOT 1000 2 110 70 23 50 aa 269 21 59
WOQT 1200 2 98 64 26 58 6152 4" 700 246 224 495 10
Cr-V + 9(a)
WOQT 1000 4 94 59 25 49 » 1000 184 173 375
(oil) 12 30
WOQT 1200 4 93 55 28 55 18 Pd 1200 142 131 293 18 65
2° 1200 12] 94 241 21 4) 45(a)
C1050 As rolled 4 102 58 18 : Z .9
Annealed 1 92 53 23. . e 8630 i 800 185 174 375 14
Normalized 4 111 62 21.5 ee lio Ni-Cr-Mo 3 38
1100 137 125 285
(water) 20 95
Normalized 4 100 56 21.7 : 4’ 1100 96 72 197 25 104
Cold drawn 1 113 95 12 - a 8742 IY 700 226 203 455
OQT 1100 $ 122 81 22.8 a Ni-Cr-Mo 11 14(a)
1” 1200 130 110
WOQT 1100 4 119 88 217, 60 te (oil) 262 21 67(a)
4” 1200 118 91
OQT 1100 2 112 68 23 55.6 235 22
9261 4 800
T 1100 2 119 78.5 23 61 235 259 228 514 10
Si-Mn 4 900 12
Mi 1100 4 101 58.5 25 54.5 se 215 192 429 11
(oil) 4 1200 13
WOQT 1100 4 112 68 2357 55.5 147 124 311 17 35(a)
9840 if 700 237 214 470 11
578 a Ni-Cr-Mo i 1200 10(a)
140 120 280 19
(oil) 6” 65(a)
1000 TS 131 302 16

579
 [2.1,2.3,2.5,2.9,2.12,2.16]
AND ENDURANCE STRENGTHS
AT 3. En-
Specimens 0.5 in. or smaller. See also Tables AT 3, AT 4, AT 6, AT 7, Fig.
as size of section increases, to as low as 0.35 for 6 in. dimen-
durance ratio s,/s, decreases
sion in cast steel.
specified.
Notes: (a) Manganese steel. (b) Number of cycles is indefinitely large unless
of cycles.
Cy. = cycles. (c) By analogy (not a test value). (d) Depends on the number
(e) Permanent mold.
Sn KSI AT Sn Sy Sy _ Sn KSI AT Ss 5
MATERIAL CONDITION NO. OF CY. Su KSI ne MATERIAL CONDITION NO. OF CY. Su a =
(6) (d) 4 (6) (d)
23 0.49 | 28 1 Stainless steel 403 Annealed bar
Wrought iron Longitudinal Stainless steel 403
19 0.55 | 25(c)} 1.3 Bars, heat treat. to Rs = 97 Pn rae . to
Wrought iron Transverse
0.38 (
Stainless steel 410. | Bars, OQT to Re = 97 58 osz| as |ia7
Cast iron ASTM 30 12
16 at 10° ay Stainless steel 410 ditto, except at 850°F 43
Cast iron ASTM 30
ASTM 30 21 at 10* i eerie sas oe OQT 1200 75 0.54 | 108 1.43
Cast iron
aa ainless steea eee and cold drawn;
Annealed
Cast steel, 0.18% C . | As cast 31.5 0.45 | 36 | 1 46 0.61 | 50 ;
1.09
. | Cast and annealed 34.5 0.45 | 37 1,
Cast steel, 0.18% C 45 | 1. uminum ‘. Wrought, T3 18 at 5 x 10° | 0
Cast steel, 0.25% C . | Cast and normalized 35 0.46} Aluminum 2014 .
48 0.49 | 61 Ge Wrought, T4 aot $010 Ina le 210
Cast steel, 1330(a) Cast, N&T 1200 Aluminum 2014 . Wrought, T6 18 at 5 x 10° 10.26 | 60 | 3.
Cast steel, 1330(a) , . | Cast, WQT (269 BHN) 58 0.48 |106 | 1 Aluminum 2014 . Ditto 30 at 10° oas eo | 200
64 0.40 | 148 2
Cast steel, 4340 . , ’. | Cast, WQT 1100 i
Aluminum 2014 . Ditto, A 500°F S5at5
54 0.49 | 85 e .
Cast steel, 8630 . Cast, N&T 1200 Aluminum 2014. T6, reversed axial
65 0.47 | 125 15 at 5 2 ie oot > ie
Cast steel, 8630 .° Cast, WQT (286 BHN) Aluminum $052 . Cold worked, H32 iPass oat bios . (16s
Wrought Steel 1015
{ Aluminum 5052 . Cold worked, H36 i9at5 x 10°lo47| 35 11 a
Cold drawn (10% work) 40 0.57 | 63 I
1015. 3 Aluminum 6063 . Wrought, 5 T5
45 at 10* 48 1 i 10 at 5 8
1020 . As rolled
As rolled 40 at 10° 48 1 Aluminum 7079. . | Wrought, T6 at i oase co a 2,
1020 i Aluminum, 142 alloy . | Sand casting, T77
As rolled 33 at 10° 48 10.5 at 5 x 108 0.35 23 ai
1020 Aluminum, 142 alloy . | Casting, T61(e)
E 9.5 at 5 x 108 0.20 42 ae
Cold drawn 46(c) 0.50} 78
1035 Alum. bronze (10%) Extruded, fp heat treated
In air 40.6 0.46 | 58 3 4 at7 ze
1035 Alum. bronze (10%) |
In brine 24.6 58 Sand cast, annealed 28 8 io" oa mae ne
FOSSN oes Cartridge brass (70-30) |
In sulfur 10.6 58 0.08” spring wire 22 at 108 0.17 65 or
LOSSp i8 is
Cartridge brass (70-30) | Half hard, 1” rod 22 at 5 x 107 0.31 —- a.
Cold drawn (10% work) 54 0.54 | 85
1040
Cold drawn (20% work) 59 0.5 92 Free-cutting brass Half hard, 2” rod., SAE 72 e ;
1040
68 Commercial bronze 0.08” hard wire a : #9 - er a AH
s Cold drawn 40(c) 0.50(c)}}
IGT Leaded tin bronze Sand cast, alloy 2A(Navy M) 11 at 10°
50 0.46 | 90 | | 0.29 is te
1141 . Cold drawn
68 0.54 | 112 ‘J Low brass (80-20) Spring hard, 0.04” strip 24 at 2 x 107 |0.26 | 65 | 2.70
13B45 . OQT 1100
Elevated temp. drawn (ETD) 72 0.48] 140 | | Low brass (80-20) 0.08” spring wire 26 at 108
1144.
0.61 | 50 Manganese bronze Sand cast, alloy 8A 21.2 . i oie oe =
52
2317 In air
31.6 50) Manganese bronze. | Sand cast, alloy 8C 25 at 10° Oo4.) gna fee.
2317 In brine Silicon bronze, type A | Half-hard rod 30 at 3 x 108 0.39 45
23.9 $0 im
2317 In sulfur
48 0.50 | 51 Silicon bronze, type B | Hot rolled 19 at 5 x 10’
2320 Hot rolled rod
0.53 |} 140 | Silicon bronze, type B | Extruded 20 at 5 x 107} 0.29 | 55(c)| 2
2320 Carburized, case hardened 90
0.64 | 100 | Silicon bronze, type B | Cold drawn, 72% reduction 30 at 3 x 108 0.32 69 : .
3120 Carburized, case hardened 90
Silicon bronze, type B | 0.08” hard wire 25 at 10° 0.28 a” 368
At 1000°F (OQT 1150) 40
4340
0.46 | 190 Magnesium (AZ63A). | Cast, TS 8
Heat treated 96
6150 Magensium (AZ31B). | Extruded bar a ‘ ae a > ie
(20%) 62 0.51 | 107
8630 . Cold drawn Inconel (Ni-Cr) . Cold drawn 40 at 108
70 0.51 | 119 0.38 - ot
94B40 . OQT 1100 Inconel . _. | As forged or hot rolled 38 at 10° 0.42 35 aes
.| Nitrided 124 0.65 | 180
Nitralloy N . ‘ Monel (67 Ni, > 30 Cu) | Annealed rod 31 ? . :
135, modified Un-nitrided 45
Nitralloy 0.66 | 140 E oe F : - | Cold drawn rod 42 re = ae 7 a
modified | Nitrided 90
Nitralloy 135, onel Annealed. In brackish water 21 at 108 0.28 30 tas
modified | Notched and un-nitrided 24
Nitralloy 135, 0.59 | 140
| Notched and nitrided 80 K-Monel (3 Al) - | Cold drawn, age hardened
Nitralloy 135, modified
38 0.37} 35 Titanium (5 Al, 2.5Sn) | Formed; es finish . _ Sete as ey pre
Stainless steel 316 Annealed bar

581
 Pine PalUNness

80 120 160 200 240 280 320 360 400

f

Factor
AT 1 1 TABLE AT11 TYPICAL CORE PROPERTIES OF 4

58
_
CARBURIZED STEELS?:*?

® b
Concentration
He
Carburizin g is done at about 1700°F. A tempering temperature of 300°F produces maximum
i
case hardness; 450°F results in improved impact strength.
specimen. (d) 4-in,

oN
Notes. (a) Nominal size of specimen, 1 in. (b) 4-in. specimen. (c) 2-in.

So
at 450°F; “DWOT
specimen. (e) Abbreviations: ““SOQT 450,” single oil quench and temper
(f) Of the order of

an
300,’ double water quench and temper at 300°F; Q, quench; P, pot.

Equivalent
mo ow
time of car! i
other hardnesses shown. (g) Case thickness depends on temperature and
in.; at 1700°
for example, at 1700°F for 4 hr., the case should be of the order of 0.05
not hard and fast rules.
for 8 hr., about 0.06 in. As seen from the values given, these are

o
_
+

CORE PROPERTIES CASE

120,000
130,000

S 160,000
¥# 170,000

5* 190,000
2 180,000

210,000
> 900,000
220,000
E 150,000
Ult. | Ten. |Elon.| Red.|BHN| Izod | Rock. | Thiel
oS
AISI CONDITION —

5" 140
Ll
NO. (e) Str.| Yd.| % |Area ft-lb.) Hard. in,
Ultimate Tensile Strengt
Ro | (hr.

a
sy ksi) 2” %

“
Sy ksi)

| SWQT 350 73, | 46 | 32 71 | 149 | 91 |C62 04 48

C1015(b) FIGURE AF 5 Reduction of Endurance Strength of Steel.{*-41 Some effects are stress
C€1020(a) | DWQT 300 85:5] 55) 33 65} 170 (f) | () raisers, some are strength reducers. See § 3.14 for the approximate roughnesses of the
C1020(a) | SWQT 300 80 | 50 | 30 60 | 160 (f) | (g) corresponding surfaces. A ground surface is not expected to have a roughness greater than
| SWQT 350 96; i) S923 53 | 192] 33 |C6S 0. : 100 microinches. The curve for ‘‘as forged,’’ adapted from Lipson, ef al.(4-2), assumes
C1117(b)
decarburization of the surface. The machined surface is a good one, light cut, fine
2115(a) | DO(or W)QT 300 90 | 60 | 30 | 70| 185) 70 | (f) | feed. Polishing leaves a residual compressive stress, helpful against fatigue.
951} 60 || 35 65| 195| 85 | (f) | @)
2317(a) | DOQT 300
2317(a) | DWQT 300 100 65 | 30
170 | 130 | 14
60 | 210}
50| 352]
70 |
40 |
(f)
(f)
g)
AT 12 AT 13
2515(a) | DOQT 300
TABLE AT12 VALUES OF K; FOR TABLE AT 13 VALUES OF K;
100 70 | 25 §5| 212) SS |) (g)
3115(a) | DOQT 300 SCREW THREADS “".?! For tension or bending. Not FOR KEYWAYS."-2) See § 10.4
3215(a) | SOQT 300 141 | 110 | 17 50 45 | (f) | &
E3310(b) | SOQT 450 180 | 149 | 14.5] 58| 363] 57 |CS57.5)0.04 Be 0
E3310(b) | DOQT 300 177 | 143 | 15.3] 58| 352] 47 |C61 |0.0:
ANNEALED HARDENED Su eetr ies Be ;
95 | 18 52| 285) 55) GS KIND OF THREAD | Rolled Cut | Rolled Cut me a
3415(a) | SOQT 300 130 | yway
135 | 105 | 19 55} 300} 60] (f) (
3415(a) | DOQT 300 Sellers, Amer. 22 2.8
4320(b) | Direct OQ from P300} 217 | 159 | 13 50] 429] 32 |C60.5)06 Nat'l., Sq. Th. 3.0 3.8 ee ANNEALED HARDENED
4320(b) | DOQT 450 145 | 94 | 21.8] 56| 293) 48 |}C59 |0.0
. 1.4 18 26 33 KEYWAY Bena-
i he - Bend- iE Tor =
Whitwort h
4620(b) | DOQT 300 122 | 77 | 22 | 56| 248] 64 |C62 |0.0 eg ie oon
62| 235] 78 |C59 |0€ Rounded
4620(b) | DOQT 450 115 | 77 | 22.5]
52| 415] 44 |Cél 10.0 Roots Profile 16 1.3 2.0 1.6
4820(b) | SOQT 300 207 | 167 | 13.8]
4820(b) | SOQT 450 205 | 184 | 13 53 | 415] 47 |C57.5|0. Dardelet 1.8 23 26 33 Sled-runner P3003 16 1.6

388] 26 |C64 |0: Semicircular

8620(b) | SOQT 300 188 | 149 | 11.5] 51]
167 | 120 | 14.3} 53] 341] 29} C61 |0«
8620(b) | SOQT 450
133 83 | 20 56| 269| 55 | C64 |0.
8620(b) | DOQT 300
179 | 144 | 15.3} 59) 375) 57 C59.5) O.
£9310(b) | Direct OQ from P300|

173 | 135 | 15.5] 60| 363] 61 |C62 0.0

E9310(b) | SOQT 300
un Os

174 | 139 | 15.3} 62] 363| 54 |C60.5)0.

E9310(b) | DOQT 300
159 | 122 | 15.5| 57} 321] 68 | (f) (
E9310(a) | SOQT 300
145 | 108 | 18.5| 66] 293 | 93 | (@) f
E9310(c) | SOQT 300 Approx. K; = 2.8. smax = KiSo at AB inside
hole.
136 94 | 19 62 | 277| 93 | (f)
>

E9310(d) | SOQT 300 INTERFERENCE FITS

Ks; =1.5 to 4, Examples: Cold-rolled shaft
Ky; = 1.9. Heat-treated shaft Ky = 2.6.7-in as
582 forged Ky = 3.
aa ARAL a f4c01
Tey, ies , ; i TT ’ ; eee . = c > 4
eves ’ Notch Radius rT ie RNa NT: StU eee NE tp Pee

; oe 0.02 0.04 0.06 0.08 0.10 0.12 O14 0.16 0.18

a= 0.0025 LE S200
e T 1 4.

0.9 : <—leeT oooh t rs —sd=se=o

a
preps T T 4
a+ 16
ee / A Annealed tee ¢
£07 Normalized : x {|
é i ie« r g | HN
¢ 06
ed
7
ae a a
Aluminum r Ser : poets Coes ij
3 Alloy t om 3

a | N el 8
ete
3 es
ue
_italr 7] r { ——4 ‘
ne /
Ky=qKi—))
i icin
+1 For t/r Less Than 3 a $ Solid Cu
~ SON, ~
See eS 1 rves, N

i 7 itp Bending in SY
‘6|— Plane of Paper, |
0 ula
32
paki
16
3
32
cng
8 32
a 3
16 ie
a
= 6M
:
Notch Radius r, Inches pat R lL as Tali
0 0.1
FIGURE AF7 Average Notch Sensitivity Curves. Applicable particularly to no rid
stresses; used also for shear stresses. (After R. E. Peterson)!4-1-4-571
FIGURE AF 9 Flat Plate with Fillets. The tensile load is central. For 4/d = 1.1, tensile

ed lane AF 8 Flat Plate with Central Hole—Tension and Bending.'*-?-*-*6 and bending curves are very close together down to r/d = 0.04. (After R. E. Peterson){4-21]
Use solid curve for bending in plane of paper; and for rod (D = h) with a hole (d)in
Symmetric loading. (Some evidencel4-61] that values for a tensile rectangular body ¢
overly conservative.) FIGURE AF 10 Flat Plate with Grooves.
2.8
Z.0RTITTITIT TITTY TTT TTT TTT IT TT TTT
ES \ VW A 0-35 ef
2.95 N \ 1 65 2.6
= IN 1% N <— | —t qj
28 x bs eat rl ry— d —+ |F—6.0) ae 2.4
= A rok L a oe
Theoretical SCF, K; (Solid Curve)

27
=
feOe
\
OT
\
eg ] te
si
3
i
, 38 Dotted Curves,
ye = NES ‘ = ‘ah x Tension, so=
. = re ios F, =F,
= = F, So (h-apt
8%--—- 458e
ta ee 5 a 2.0 —
és

— NX x qo ie = ~es~
2.5 ee re Dotted Curves, Loaded 4.5 % 5 1g
E » \, by F, ona Pin in the = & @ 18
nab \ Hole, Margin b per 3
rie= Solid Curve,4 y ty > == ee i F 1.6
23 Force Acts E ay Scale Here}. 2
:
bi:
on Part. spk ~
ai=—> oa
=

2.26
2 “~~ Seale ___ Seale Sah Ps | Patho,
— ° 4 Solid Curves,
im Dotted Curves “hp 37> s = 1 2 Bending, $= SM
25 . td?
21 c= or a ee)
ihe ote TS
SpbkLy |S
Fe Sh 1 ou Ca Pa a a BT tH 8 eirdon Nee
; 0.1 0.2 0.3 0.4 0.5
r/d
d/h
 er Sane nn) ee ar AEST E yr 5 eel, wea tv Vil Pe Poe ae ae e ee e Fp OV e,
SoA Sail EN

i 5 BSAb i ‘
- +.
eee
ba 7 aes a
& | trace ras we an
mes 7 z
5.0 = t

is ts sae ee: oe 4
i Tension. / ‘4
N\ :
45 w
7 4
% og NM For Tension, See Fig.; 64 my

x cg Tension, ye Wu
cy fa eer i e & ee: Solid Curve, ee
8o- th=ayt Y ‘a ZR (pat Bending (This Plane)
Bh
en
Zz

|
:
3S 2.2
Ls ry Ce ee
_ Me I_ xp dD? —

Ss ae Bending, at A
obe ss —_ Mc ys s Iad
3 a
8)=——
I s §eq”20 iin ———
EBos = 73
pa lew*t x 7 4
~S Bee
3.5 Sa ee Wes i
y ae q —0-5 or More ri 18 TA Torsion—K, at B, c~D/2
A : T
F l= r$=0.3 | =
ee , ee
89 =————— (Shear
ri ea | be Solid Curves 16 6
$0 ig aie / | for Bending 4] 1.6 —

r eu. h caao
ra a 0.1 ($>10)
l | _t— [oceans piec fret | | 6D Oe Dh EO oly ei peel ee
a6 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0 0.1 0.2 0.3
a/b d/D
FIGURE AF 13 Shaft with Radial Hole—Bending and Torsion. Maximum torsional
FIGURE AF 11 Flat Plate with Eccentric Hole.{*-?:4-2°) For tension, if h < 20¢
stress for this case falls slightly inside the shaft diameter on the inside surface of the hole at
is somewhat smaller than shown. (After R. E. Peterson).{*:?") some point B. (After R. E. Peterson).(4-21)

FIGURE AF 12 Shaft with Fillet. The tensile load is central. Torsion curve Did =
[FIGURE AF 14 Shaft with Groove. Use the solid curve for D/d = 1.01 for tension as well
approximates the bending curve for D/d = 1.01; torsion curve D/d = 2, approxima’
as bending (approximate).
bending curve for D/d = 1.02 (down to r/d ~ 0.04). Bending curve D/d = 1.1 app
mates the tensile curve D/d = 1.1. (After R. E. Peterson). ~
aC a
27 TTT. GRE
‘ 4
Ean
\
ak Tite lake
Semicircular
al it hinoinieth deka aret
:
2.6 rt x \ \\|Groove, Bending ‘a
241-44 \—4 4
2.4
75 Mis ee
ae
Ven Neen N-=—\a
\
.
ZI
\

se Dotted Curves,
7

\ \ XK XN. <a ‘ 4F
Dotted re R ‘ Tension, s, =—— 4

WX Ss
2.2

Theoretical SCF, K:
<
»
Tension,
5
8 oe
24
2.0 ‘ Sb ae
io ‘ ~ :
5 ee Solid Curves, ‘ > O7>\ Bending, s.= sal

Z3 ~> Bending, 8% = 32Ma 1.8 * he

SN~ Ler,“a=2 op aay ad

$ 1.8 SL be Ie ~<a Mon>~ a a

~ 1.6 A e> iba
3
Se < ory ze =e

& 1.6
: M. es
14 \ Y a PLsoy Li
: Yd “—~<0r More id
1.4 a Torsion,?)~~—_ “101 (for Tent Soc ee eA
2P 26? fee = 10n Too)
1.2 Torsion
(B ad? D/d=1.01 eS
16T Vy esae a ed
cog =—— (Sh tShear) ae 0.1 0.2 0.3
1.0
 num ALAR VINITIDY ALY ALWIONLUALN

AT 14 SCREW THREADS—COARSE AND FINE

Selected values abstracted from ASA B1.1-1960,!°-1] which should be referred to for de! TABLE AT15 NOMINAL DIMENSIONS OF VARIOUS GAGES
AT 15
on proportions and tolerances, and for other thread series; dimensions for a class3
The Washburn and Moen (W & M) gage, called also the steel-wire gage, is used for steel
external thread. The minor diameter for the internal thread is not quite the same as for
wire. The American Wire or Brown and Sharpe (B & S) gage is used for monel, bronze,
external thread. The fensile stress area is that area corresponding to a diameter
copper, aluminum, and brass wires. Standard wire sizes other than those in the "table
approximately the average of the pitch and minor diameters; detail in Ref. (5.1).
include multiple of 4; upsto in. The tendency is to specify the decimal size of the wire.
values of 12-thread series above 14-in. size.
There is also a music-wire gage. And much smaller sizes than those listed are available.
COARSE (UNC) FINE (UNF), AND 12 UN WIDTH
BASIC FLATS, REG' WIRE DIAMETER, IN.
SIZE | MAJOR Th. Minor Stress Th. Minor Stress UNFINISHED; GAGE NO. W&M B&S PLATE THICKNESS, IN.
DIA. oe Dia. Area, As — Dia. Area, As Ain. Ferrous Nonferrous U.S. Standard
in. In.) ext. Th. | sq. in. In. | Ext. Th.| — sq.in. | (Selected fror 7-0 0.4900
ASA B18.2- 0.500
6-0 0.4615 0.469
1952.)
5-0 0.4305 0.438
4-0 0.3938 0.460 0.406
3-0 0.3625 0.401 0.375
A i

S 4
2-0 0.3310 0.365 0.344
0 0.3065 0.325 0.313
1 0.2830 0.289 0.281
2 0.2625 0.258 0.266
0 0.0600 80 0.0447 0.00180 For square
3 0.2437 0.229 0.250
1 0.0730 64 0.0538 0.00263 72 0.0560 0.00278 heads and
2 0.0860 56 0.0641 0.00370 64 0.0668 0.00394 nuts, A is diff- 4 0.2253 0.204 0.234
3 0.0990 48 0.0734 0.00487 56 0.0771 0.00523 erent in the 5 0.2070 0.182 0.219
4 0.1120 40 0.0813 0.00604 48 0.0864 0.00661 smaller sizes” 6 0.1920 0.162 0.203
(below 7/8) 7 0.1770 0.144 0.188
5 0.1250 40 0.0943 0.00796 44 0.0971 0.0083 from __ thes 8 0.1620 0.128 0.172
. 6 0.1380 32 0.0997 0.00909 40 0.1073 0.01015 values. ee 9 0.1483 0.114 0.156
8 0.1640 3a: 0.1257 0.0140 36 0.1299 0.01474 detail in the 10 0.1350 0.102 0.141
10 0.1900 24 0.1389 0.0175 32 0.1517 0.0200 standard. ' 11 0.1205 0.091 0.125
12 0.2160 24 0.1649 0.0242 28 0.1722 0.0258 , 12 0.1055 0.081 0.109
Head Nut 13 0.0915 0.072 0.094
4 0.2500 20 0.1887 0.0318 28 0.2062 0.0364 te te4 14 0.0800 0.065 0.078
vs | 0.3125 | 18 | 0.2443 0.0524 24 | 0.2614 | 0.0580 4 4 15 0.0720 0.057 0.070
3 0.3750 16 0.2983 0.0775 24 0.3239 0.0878 rd t 16 0.0625 0.051 0.063
ve 0.4375 14 0.3499 0.1063 20 0.3762 0.1187 3 = 17 0.0540 0.045 0.056
$ | 0.5000 | 13 | 0.4056 0.1419 20. | 0.4387 | 0.1599 3 18 0.0475 0.040 0.050
5 0.5625 12 0.4603 0.182 18 0.4943 0.203 43 $ j
& | 0.6250 | 11 | 0.5135 0.226 18 | 0.5568 | 0.256 BR Oe
3 | 0.7500 | 10 | 0.6273 0.334 16 | 0.6733 | 0.373 14 1k
& | 0.875 9 | 0.7387 0.462 14 | 0.7874 | 0.509 18, If
1 1.0000 8 0.8466 0.606 12 0.8978 0.663 1} 1 . AT 16
1h | 15125 7 0.9497 0.763 12 1.0228 0.856 144 144
14 | 1.2500 7 | 1.0747 0.969 12 | 1.1478 | 1.073 1% 1%
13 1.375 6 1.1705 1.155 12 1.2728 1.315 235 2a TABLE AT 16 APPROXIMATE FREE LENGTHS AND SOLID HEIGHTS
13 1.5000 6 1.2955 1.405 12 1.3978 1.581 24 2} 7 (P = pitch of coils, N. = number of active coils, D,, = diameter of wire)

13 1.7500 5 1.5046 1.90 12 1.6478 2.1853 28 28 ' TOTAL SOLID

2 2.0000 4} 1.7274 2.50 12 1.8978 2.8892 3 3 a TYPE OF ENDS FREE LENGTH | COILS HEIGHT
2} | 2.2500 | 44 | 1.9774 3.25 12 | 2.1478 | 3.6914 33S
33 it h Plain 5
24 | 2.5000 | 4 | 2.1933 4.00 12 | 2.3978 | 4.5916 Plain
PN. + Dw N. DyNe + Dw
4h count: 3 PN. N- DN.
623 2.7500 4 2.4433 4.93 12 2.6478 5.5900f
Squared . ‘ PN. + 3Dy | Ne +2 |DwNe + 3Dw
6.6865 4h 44 : Squared and ane a PN, + 2Dy | No +2 | DuN-e + 2Dy
3 3.0000 4 2.6933 5.97 12 2.8978
3} | 3.2500 4 2.9433 7.10 12 3.1478 7.8812 4g i
34 | 3.5000 4 3.1933 8.33 12 3.3978 9.1740 589
32 | 3.7500 | 4 | 3.4433 9.66 12 | 3.6478 | 10.5649 5
14 |4.0000 | 4 | 3.6933 | 11.08 12 | 3.8978 | 12.0540 6
 MIN. TENSILE, MAX, “SOLID” END. STR” we ee
E es
G eleSTRESS,
DESIGN MIN. TENSILE, MAX aeay — ge TH.
Y
E G DESIGN STRESS,
MATERIAL x 107° |x 107° Syq ksi. 5, ksi. 5, Sno ksi.
(Approx. s,;) (R = 0) psi psi Light Load (Uncoiled) (Approx. s,.) (R= 0)
psi psi Light Load (Uncoiled)

(2) (3) (4) (5) (6) Column No. | (1) | (2) ent (4) (5) (6)
Column No. —| (1)
Oil Tempered;
29 11.5k Q
(a) Dow
146
(b) 6.
0.6s, (c) ~47
Carbon Steel
VSQ(s):
Sena) | aL.ne ®
a182
pt ©
ie
teeeexcal 1] pos OO
ASTM A229
0: D,,<90.5
<D,,
[0.032 =87.5,x=0.19 Noosi w <p,
<0.5] | [Q=87.5,x ASTM A230 [0.093 < D,, <0.25]| [0.093 < D,, <0.25]}|[0.093 <D,, <0.25]}

Hard Drawn; Use 0.85 times 140 » (d) SenaQ (s): 30 | 115 (a) aig 06s, « (c) mame as( for)
ASTM A227. | 29 | 11-5 | constants in pois 05s, (c) or Date A230

ASTM A232 [0.032 < D,, <0.437]|[Q = 100,x =0.166]][0.028 < D,, <0.5]
Note (a) D, <0.625]| [0=70,x=0.19]
[0.028'< “9

190 05s, (c) 50 , Cr-Si

ASTM Steel
AaOl 29 | 11.5 (a) 202
p00 (b) 0.6s, Pee wo
Music Wire;
ASTM A228 Waheed? (a) pois () | (Q=95,x=0.154]| pose
[0.032 <D,, <0.375] [Q = 121,x =0.107]|[0.032<D,, =
[0.004<D,, <0.192]| [0.03 <D,, <0.192;|[0.018<D,<
190 ksi max.] 92 ksi ma:
Stainless Steel 6 10 170
pees
(Cr-Ni) (i) D>
—, (b)(b 0.47 s, (h) Dt
—— (d)(q)

AT 17 [0.01 <D,,<0.13] | [(@=80,x=0.14] |[0.01 Sa

st
[0.13<D,, <0.375]|[Q=45.6,x=0.41]
TABLE AT17 MECHANICAL PROPERTIES OF WIRE FOR
SPRINGS
Beryllium 18.5 7 Use 0.8 times 35
For extension springs, use 0.8 times the value in column (5) for the maximum occasioi Copper ; values for 160-200 O.5s, (h) D2 (d)
stress. A229 eit
[0.09<D,,<0.5;
The stress is s = Q/D,,* wherever this form appears; to be used with equation 56 ksi max.]
always include the curvature factor except for the mean stress in fatigue design. The
given apply when the spring is not preset and not peened, except as stated. For live Spring Brass Use 0.35
. times , 42 j 11.5
stress values for the steels may be increased by 25% for shot-peened coils. For preset ne values for pit (tk) pow 4) po
springs, the static stress and ‘‘solid stress’’ may be some 40-50% greater than git
~ [0.03<D,,<0.20; | [0.08<D,<0.5; | [0.09<D,<05;
columns (3) and (5). See notes (f) and (n) below. The computed design stresses are 125 ksi max.] 68 ksi max.] 19 ksi max.]
be interpreted as being exact values. Reduce design stresses 50% for shock loads (
to hammer blows). Where a maximum stress is given, use it for wire sizes smaller Phosphor Use 0.5 times 106
specified limits. The limits given for D,, apply only to the equations; for many ma‘ Bronze 14.5 6 values for D008 (i) 0.45s, (h) aa (4)
wire sizes smaller or larger than the limits shown are frequent.
Notes: (a) For light service, use design S54 = 0.405s,,. For average service, use
A229 [D,<05; |{(Q=47.5;x=0.08]] [0.09<D, <0.5;
145 ksi max.] 28 ksi max.]
0.324s,,. For severe service, use 5,4 = 0.263s,. These results agree closely with Westing
recommendations, as reported by Wahl. (b) Equations for approximate minimum
(c) Agrees closely with Alco recommendations; sinc Monel (m) 129
strength as specified by ASTM. 24.5 9.3 ox (h(i) aon 0.4s, (h)(n)
D, D D.°2 (d)(n)
are higher than the stresses recommended by some authorities, a small factor of safe w [Q=51.5;x=0.1] w
be advisable, unless the spring manufacturer agrees. (d) Derived from Hunter Sp [0.058 < Dy <0.625; [D,, < 0.625; [D,, < 0.625; [D,,< 0.625;
data.[5-15] Value for indefinite life from 0 to max.; for 10° cycles for steel wire 70 ksi max.] 170 ksi max.] 68 ksi max.] 29 ksi max.(p)]
stainless), multiply this value by 1.4, for example. Use minimum N = 1.15. (e) Use b
expressions for oil tempered; for hard-drawn wire, multiply by 0.9. (f) By analog
M51 95. 1. gpa prea Le seems. ML GUAR AORG, areas
K-Monel (0) 75ksi (h)(I) 158
|
music wire. Also, Associated Spring recommends the safe design range for valve
6.9, for D,, < 0.207 and unpeened; by tr
quality as defined by the triangle ABO, Fig. [D,,>0.058] [Q=63, x =0.048] |[29 ksi max. (p)]
INCO
CBO when peened. (g) Conservative in the larger sizes. (h) In accordance with [D,,
< 0.625; [D,,< 0.625;
(i) Light service, use s, = 0.32s,,3; average service, 0.26s,,; severe service, 0.21s,. (j) 180 ksi max.] 72 ksi max.]
sy =
from Associated Spring data. (*-2) (k) A straight-line interpolation between
by 0.8 for aver
for D,, = 0.5 in. and formula limit is probably satisfactory. (1) Multiply
(m) Stress relieved; decrease 10% if as-drawn. (n) Increase 10% if
service.
peening wires larger than 0.062 in.{®-12] (0) Age hardened. (p)
25-35%, for shot
(r) INCO
gives 20 ksi for 10° cycles. (q) Multiply by 1.33 for 10° cycles, 0-max.
is somewhat stronger than the Monel, but detail I:
suggest that this K-Monel
(s) Valve spring quality.

591
 LE EOE ENE I ENR PHS RSENS RNID SAN VUE VE? DESAVED
Extracted with permission of publisher, John Wiley & Sons, Inc., from Seely and Smith. (1:71
For a hollow circular section, the values for a solid section may be used with little error
when r/c = 1.8; K., for point on inside of curvature: K., for outside.

< Center of Curvature

rie>1.2 14 16 1.8 2.0 3.0 4.0 6.0 8.0 10.0

Key 3.41 2.40 1.96 1.75 1.62 1.33 1.23 1.14 1.10 1.08
Keo 0.54 0.60 0.65 0.68 0.71 0.79 0.84 0.89 0.91 0.93
Yolr 0.22 0.15 0.11 0.08 0.07 0.03 0.02 0.01

Key 2.89 2.13 1.79 1.63 1.52 1.3 1.2 1.12 1.09 1.07
Keo 0.57 0.63 0.67 0.70 0.73 0.81 0.85 0.9 0.92 0.94
Yolr 0.31 0.20 0.15 0.11 0.09 0.04 0.02 0.01

Key 3.63 2.54 2.14 1.89 1.73 1.41 1.29 1.18 1.13 1.10
VA
LL Keo 0.58 0.63 0.67 0.70 0.72 0.79 0.83 0.88 0.91 0.92
Yolr 0.42 0.30 0.23 0.18 0.15 0.07 0.04 0.02 0.01

1.7

16
Pa ah “Hbalog.(r + c1) + (t — Blog. r+ ce)
15
+ (6 — tlog.(r — cs) — blog.(r — cz)].
For a T-section let cg = c, and b, = t.
Stress Factor

For an J-section, let b; = b. Also, if the front

and rear flanges are the same in thickness, cz will
&

be equal to c; and cz will be equal to cq.

For a box-section (dotted outline), Z is the
same as for an I-section, each side panel of the
box being t/2 inches thick.
11
Flanges in compression should not be so thin
as to result in local buckling.
1.0
BON ing Big ig PE Ra ee gn ae
C=Spring Index== or De.
Penur4 aaorlle + 4 ie ca)|
FIGURE AF 15 Stress Factors (Wahl). Use D,,/D,, as the spring index for x log. ( r+ Ce
)-@- a}
o
springs, Dm/t for rectangular-wire springs, where f is the dimension perpendicular Em Es,
grea’
of the spring. It is advisable that D,/D,. be not less than 3, preferably
AT 19
TABLE AT19 KEY DIMENSIONS
ASA standard
See Figs. 10.1 and 10.2 for b and ¢. From
on ¢ may be
B17.1-1943. Other sizes available. The tolerances
keys, positive on Temperature, ¢ °C
numerically the same as given, negative on plain
“ Be i ee, ha 9 00
tapered keys. 200 ES
150
SHAFT 800
es400
DIAMETER TOLERANCE 100
CInclusive) b t on 5, in.

—0.0020 50
, rp —0.0020
0 300
i L fe
T6-1d t ts —0,0020 200
*
@ 20
$ —0.0020 Ee
ce r 4 —0,0020 100
i
» §5
=
143-24 4 g —0.0025
z 108 m 3
—0.0025
.
x
30 8
g qs
—0.0025
§ 7
ae i 4 48° N5
33 -32 & —0.0030 za& 5 ‘cont
x 2 g
= 4
2 3.5 ie
1 —0.0030
a ns 1} ; —0.0030 20 z
53 -6 1} 1 —0.0030
3
5
a
,

5
F
<4
Ba
g
2

“is
9
8
7
0.9 ;
6
08
0.7 5
45
0.6
80 «690 100 110 120 130 140 150 160 170 180 190 200 210 220
Temperature, t °F

FIGURE AF 16 Typical Viscosities of Oils. For ring-oiled bearings, usually SAE 20 (or
the equivalent) or lighter. SAE 70 and chart paper by courtesy of Westinghouse Electric
Corp. Dotted curve U is for a high viscosity-index oil, Uniflow—typical test values,
Standard Oil of N.J. Other data from The Texas Co. On average, an SAE 10W-30 oil
has a viscosity a little lower than SAE 30 at 210°F, a little higher than SAE 10 at 100°F.

594
Minimum-Film-Thickness Variable, ho/ c, (dimensionless)
0.9

0.8
TABLE AT20 DIMENSIONLESS PERFORMANCE PARAMETERS
0.7 FOR FULL JOURNAL BEARINGS WITH SIDE FLOW
Courtesy Raimondi and Boyd"*!-7) and Westinghouse Electric. Values
0.6
of h,/c, for optimum
bearings, maximum load and minimum friction, respectively: for
L/D = cO, 0.66, 0.60;
for L/D = 1, 0.53, 0.30; for L/D = 0.5, 0.43, 0.12; for L/D = 0.25, 0.27, 0.03.
05

0.4 LID| he Ss ¢ bey ON, yes Mi Mg Ree 10 LP,

Ge Cr re,nsL q P Pmax
0.3
o 10 1.0 0 (70.92) co 7 0 co —
0.1 0.9 0.240 69.10 4.80 | 3.03 0 19.9 | 0.826
0.2 0.8 0.123 67.26 2.57 | 2.83 0 11.4 | 0.814
0.4 0.6 0.0626 61.94 1.52 | 2.26 0 8.47 | 0.764
0.6 0.4 0.0389 54.31 1.20 | 1.56 0 9.73 | 0.667
0.06 .08 .10 0.20 0.40 1.0 2.0 0.8 0.2 0.021 42.22 0.961 | 0.760 | 0 15.9 | 0.495
0.9 0.1 0.0115 31.62 0.756 | 0.411 | 0
2un ; :
23.1 0.358
0.97 | 0.03 _ _ _—
Bearing Characteristic No., S = (Zz) a (dimensionless) _— 0 — _
1.0 0 0 0 0 0 0 co 0
FIGURE AF 17 Minimum-Film Variable vs. Sommerfeld S (with Side Flow).
1 {0 1.0 oo (85)
Raimondi and Boyd"!:?! and Westinghouse Electric). ; <2 7 0 0c —
0.1 0.9 1.33 79.5 26.4 3.37 0.150 106 0.540
0.2 0.8 0.631 74.02 12.8 3.59 0.280 52.1 0.529
0.4 0.6 0.264 63.10 5.79 | 3.99 0.497 24.3 | 0.484
FIGURE AF 18 Coefficient-of-Friction Variable vs. Sommerfeld S (with Side 0.6 0.4 0.121 50.58 3.22 «|4.33 0.680
(Courtesy Raimondi and Boyd'!!-?! and Westinghouse Electric). 0.8 0.2 0.0446 36.24
14.2 | 0.415
1.70 | 4.62 0.842 8.00 | 0.313
0.9 0.1 0.0188 26.45 1.05 | 4.74 0.919 5.16 | 0.247
0.97 | 0.03 0.00474 | 15.47 0.514 | 4.82 0.973 2.61 | 0.152
1.0 0 0 0 0 _ 1.0 0 0

3 Io 1.0 co | (88.5) o | « 10 o |
0.1 0.9 4.31
88

81.62 85.6 3.43 0.173

Coefficient-of-Friction Variable, fr/c, (dimensionless)

343.0 | 0.523
0.2 0.8 2.03 74.94 40.9 3.72 0.318 164.0 | 0.506
0.4 0.6 0.779 61.45 17.0 4.29 0.552 68.6 | 0.441
&§

0.6 0.4 0.319 48.14 8.10 | 4.85 0.730 33.0 | 0.365

0.8 0.2 0.0923 33.31 3.26 | 5.41 0.874 13.4 | 0.267
0.9 0.1 0.0313 23.66 1.60 | 5.69 0.939 6.66 0.206
8

0.97 | 0.03 0.00609 | 13.75 0.610 | 5.88 0.980 2.56 | 0.126

1.0 0 0 0 0 — 1.0 0 0
J

+ 10.0 1.0 co
ane

(89.5) co 7 | 0 00 —
0.1 0.9 16.2 82.31 | 322.0 3.45 0.180 | 1287.0 | 0.515
0.2 0.8 7.57 75.18 | 153.0 3.76 0.330 611.0 | 0.489
0.4 0.6 2.83 60.86 61.1 4.37 0.567 245.0 | 0.415
0.6 0.4 1.07 46.72 26.7 4.99 0.746 107.0 | 0.334
0.8 0.2 0.261 31.04 8.80 | 5.60 0.884 35.4 | 0.240
0.9 0.1 0.0736 21.85 3.50 | 5.91 0.945 14.1 0.180
0.97 | 0.03 0.0101 12.22 0.922 | 6.12 0.984 3.73 | 0.108
1.0 0 0 0 0 —_ 1.0 0 0

q in.3/sec. p = 0.03 Ib/in.* ¢ = 3734 in-Ib/lb—°F. pe = 112,

0.04 0.06 .08 .10 0.20 0.40 0.60 0.80 1.0 2.0 40 6.080
597
2
mw enareses

180° BEARING,
SRR BFALIVARUIWAVINOD

CENTRALLY LOADED,
TERE UNIVAIICE

WITH SIDE FLOW

PADNAIOLENO PUN
120° BEARING, CENTRALLY LOADED, WITH SIDE FLOW. __
Courtesy Raimondi and Boyd!"*-7] and Westinghouse Electric. Values of h,/c, for optimum
Courtesy Raimondi and Boyd!?-7] and Westinghouse Electric. Values of h,/c, for optimum bearings, maximum load and minimum friction, respectively: for L/D = co, 0.53, 0.5;
bearings, maximum load and minimum friction, respectively: for L/D = 0, 0.64, 0.6; for L/D = 1, 0.46, 0.4; for L/D = 0.5, 0.38, 0.28; for L/D = 0.25, 0.26, 0.06. When
for L/D = 1, 0.52, 0.44; for L/D = 0.5, 0.42, 0.23; for L/D = 0.25, 0.28, 0.03.
1 — h,/c, «, the trailing end of the bearing does not reach h, as defined in Fig. 11.6; that
is, h, in this table is hin.
h, r q qs pcAt,
re cr : $ ak rensL q P |D| « Fo s $ Ly q qs pcAt, P
Cr Cr re,nsL q Pp Pmax

co} 0 1.0 © 90.0 (oe) 7 oo o

0.1 0.9 0.347 72.90 3.55 | 3.04 0 14.7 co 10 1.0 © 90.0 © 7 0 oo —
0.2 0.8 0.179 61.32 2.01 | 2.80 0 8.99 0.1 | 0.9007 | 0.877 66.69 6.02 | 3.02 0 25.1 | 0.610
0.4 0.6 0.0898 49.99 1.29 | 2:20 0 7.34 0.2 | 0.8 0.431 52.60 3.26 | 2.75 0 14.9 | 0.599
0.4 | 0.6 0.181 39.02 1.78 | 2.13 0 10.5 | 0.566
0.6 0.4 0.0523 43.15 1.06 | 1.52 0 8.71
0.8 0.2 0.0253 33.35 0.859| 0.767 | 0 14.1 0.6 | 0.4 0.0845 32.67 1.21 1.47 0 10.3 | 0.509
0.9 0.1 0.0128 25.57 0.681] 0.380 | 0 22:5 0.8 | 0.2 0.0328 26.80 0.853 | 0.759 | 0 14.1 | 0.405
0.97 | 0.03 0.00384 | 15.43 0.416] 0.119 | 0 44.0 0.9 | 0.1 0.0147 21.51 0.653 | 0.388 | 0 21.2 | 0.311
1.0 0 0 0 0 0 0 © 0.97 | 0.03 0.00406 | 13.86 0.399} 0.118 | 0 42.4 | 0.199
10 |0 0 0 0 0 0 oO 0

1 |0 1.0 © 90.0 _ 7 0 © 1 |0 1.0 © 90.0 © 7 0 (oe) _

0.1 0.9 1.40 78.50 14.1 3.34 0.139 57.0 0.1 | 0.9024 | 2.14 72.43 14.5 3.20 0.0876 59.5 | 0.427
0.2 0.8 0.670 68.93 715 | 3.46 0.252 29.7 0.2 | 0.8 1.01 58.25 7.44 | 3.11 0.157 32.6 | 0.420
0.4 0.6 0.278 58.86 3.61 | 3.49 0.425 16.5 0.4 | 0.6 0.385 43.98 3.60 | 2.75 0.272 19.0 | 0.396
0.6 0.4 0.128 44.67 2.28 | 3.25 0.572 12.4 0.6 | 0.4 0.162 35.65 2.16 | 2.24 0.384 15.0 | 0.356
0.8 0.2 0.0463 32.33 1.39 | 2.63 0.721 10.4 0.8 | 0.2 0.0531 27.42 1:27) 1Y 487 0.535 13.9 | 0.290
0.9 0.1 0.0193 24.14 0.921| 2.14 0.818 9.13 0.9 | 0.1 0.0208 21.29 0.855 | 1.11 0.657 14.4 | 0.233
0.97 | 0.03 0.00483 | 14.57 0.483} 1.60 0.915 6.96 0.97 | 0.03 0.00498 | 13.49 0.461 | 0.694 | 0.812 14.0 | 0.162
1.0 0 0 0 0 _— 1.0 0 1.0 |0 0 0 0 — 1.0 0 0

4] 0 1.0 © 90.0 © 7 0 © 4 10 1.0 © 90.0 © 7 0 _ _—

0.1 0.9 4.38 79.97 44.0 3.41 0.167 177.0 0.1 | 0.9034 | 5.42 74.99 36.6 3.29 | 0.124 149.0 | 0.431
0.2 0.8 2.06 72.14 21.6 3.64 0.302 87.8 0.2 | 0.8003 | 2.51 63.38 18.1 3.32 | 0.225 77.2 | 0.424
0.4 0.6 0.794 58.01 9.96 | 3.93 | 0.506 42.7 0.4 | 0.6 0.914 48.07 8.20 | 3.15 0.386 40.5 | 0.389
0.6 0.4 0.321 45.01 5.41 | 3.93 0.665 25.9 0.6 | 0.4 0.354 38.50 4.43 | 2.80 0.530 27.0 | 0.336
0.8 0.2 0.0921 31.29 2.54 | 3.56 0.806 15.0 0.8 | 0.2 0.0973 28.02 2.17 | 2.18 0.684 19.0 | 0.261
0.9 0.1 0.0314 22.80 L3SuesAay 0.886 9.80 0.9 | 0.1 0.0324 21.02 1.24 1.70 0.787 15.1 0.203
0.97 | 0.03 0.00625 | 13.63 0.581| 2.62 0.951 5.30 0.97 | 0.03 0.00631 | 13.00 0.550 | 1.19 0.899 10.6 | 0.136
1.0 0 0 0 0 _— 1.0 0 1.0 |0 0 0 0 —_— 1.0 0 0

; jo 1.0 co 90.0 2 7 0 © —
4|0 1.0 co 90.0 0 a | 0 oo
0.1 0.9044 | 18.4 76.97 | 124.0 3.34 0.143 502.0 | 0.456
0.1 0.9 16.3 81.40 | 163.0 3.44 0.176 653.0
0.2 | 0.8011 8.45 65.97 60.4 3.44 0.260 254.0 | 0.438
0.2 0.8 7.60 73.70 719.4 3.71 0.320 320.0
0.4 | 0.6 3.04 51.23 26.6 3.42 0.442 125.0 | 0.389
0.4 0.6 2.84 58.99 35.1 4.11 0.534 146.0
0.6 | 0.4 1.12 40.42 13.5 3.20 0.599 75.8 | 0.321
0.6 0.4 1.08 44.96 17.6 4.25 0.698 79.8
0.8 | 0.2 0.268 28.38 5.65 2.67 0.753 42.7 | 0.237
0.8 0.2 0.263 30.43 6.88 | 4.07 0.837 36.5
0.9 | O.1 0.0743 20.55 2.63 2.21 0.846 25.9 | 0.178
0.9 0.1 0.0736 21.43 2.99 | 3.72 0.905 18.4
0.97 | 0.03 0.0105 12.11 0.832 | 1.69 0.931 11.6 | 0.112
0.97 | 0.03 0.0104 12.28 0.877] 3.29 0.961 6.46
10 |0 0 0 0 _ 1.0 0 0
1.0 0 0 0 0 _— 1.0 0

599
TABLE AT 23 DIMENSIONLESS PERFORMANCE PARAMETERS FOR OF FORM SALTOR aN Eee a EON
60° BEARING, CENTRALLY LOADED, WITH SIDE FLOW FD = full depth.
Courtesy Raimondi and Boyd"?:7) and Westinghouse Electric. Values of h,/c, for optimum LOAD NEAR ' LOAD NEAR
bearings, maximum load and minimum friction, respectively: for L/D = co, 0.25, 0.23; LOAD AT TIP MIDDLE LOAD AT TIP MIDDLE
for L/D = 1, 0.23, 0.22; for L/D = 0.5, 0.2, 0.16; for L/D = 0.25, 0.15, 0.1. When NO. 144°] 20° | 20° | 144°] 20° No. | 144° 20° =. 20° | 144° 20°
1 — h,/c, * ¢, the trailing end of the bearing does not reach 4, as defined in Fig. 11.6; that TEETH | FD | FD | Stub| FD | FD |TeetH| FD FD Stub| FD FD
is, h, in this table is h,.,n-
10 | 0.176 | 0,201 }0.261 32 | 0.322 0.364 0.443 |0.547 0.617
11 | 0.192 | 0.226 |0.289 33 | 0.324 0.367 0.445 |0.550 0.623
5a hy a $ ty q q pcAt, P 12 | 0.210] 0.245 |0.311 |0.355 ]0.415 | 35 | 0.327 0.373 0.4491|0.556 0.633
|D) « a M remL | Ge pis. 4 Peas 13 | 0.223 | 0.264 |0.324 |0.377 |0.443 | 37 | 0.330 0.380 0.4541|0.563 0.645
14 | 0.236] 0.276 |0.339 |0.399 10.468 | 39 | 0.335 0.386 0.457|0.568 0.655
a6ihO 1.0 Bs 90.0 e af 410 15 | 0.245 | 0.289 |0.349 |0.415 |0.490 | 40 | 0.336 0.389 0.459|0.570 0.659
0.1 | 0.9191 | 5.75 65.91 | 19.7 |3.01 Jo 16 | 0.255 | 0.295 |0.360 |0.430 |0.503 | 45 | 0.340 0.399 0.46810.579 0.678
0.2 | 0.8109 | 2.66 48.91 | 10.1 273 10 17 | 0.264 | 0.302 |0.368 |0.446 |0.512 | 50 | 0.346 0.408 0.474]0.588 0.694
0.4 | 0.6002 | 0.931 31.96 4.67 | 2.07 | 0 18 | 0.270 | 0.308 |0.377 |0.459 |0.522 | 55 | 0.352 0.415 0.480]0.596 0.704
0.6 | 0.4 0.322 23.21 2.40 11.40 |0 19 | 0.277] 0.314 |0.386 |0.471 |0.534 | 60 |0.355 0.421 0.484|0.603 0.713
0.8 | 0.2 0.0755 | 17.39 1.10 | 0.722 | 0 20 | 0.283 | 0.320 |0.393 |0.481 |0.544 | 65 | 0.358 0.425 0.488|0.607 0.721
0.9 | 0.1 0.0241 | 14.94 0.667 | 0.372 | 0 21 | 0.289} 0.326 |0.399 |0.490 |0.553. | 70 | 0.360 0.429 0.493|0.610 0.728
0.97 | 0.03 0.00495 | 10.58 0.372 | 0.115 | 0 22 | 0.292] 0.330 |0.404 |0.496 10.559 | 75 10.361 0.433 0.496|0.613 0.735
1.0 | 0 0 0 0 0 0 23 | 0.296] 0.333 |0.408 |0.502 |0.565 | 80 | 0.363 0.436 0.499|0.615 0.739
24 | 0.302] 0.337 | 0.411 | 0.509 |0.572 | 90 | 0.366 0.442 0.503|0.619 0.747
1 oi aah Je a a aes 25 | 0.305] 0.340 |0.416 | 0.515 |0.580 | 100 | 0.368 0.446 0.506|0.622 0.755
02 | 0.8133 | 3.92 50.96 | 148 | 2.82 26 | 0.308 | 0.344 |0.421 | 0.522 |0.588 | 150 | 0.375 0.458 0.518|0.635 0.779
04 | 06010 | 134 33.99 661 | 2.22 27 | 0.311] 0.348 | 0.426 | 0.528 |0.592 | 200 | 0.378 0.463 0.524|0.640 0.787
28 | 0.314] 0.352 |0.430 | 0.534 |0.597 | 300 | 0.382 0.471 0.534|0.650 0.801
oF aa ae need ie ise 29 | 0.316] 0.355 |0.434 | 0.537 |0.602 | Rack | 0.390 0.484 0.550] 0.660 0.823
09 | 0.1 0.0309 | 15.33 0.822 | 0.519 30 | 0.318 | 0.358 | 0.437 | 0.540 | 0.606
0.97 | 0.03 0.00584 | 10.88 0.422 | 0.226
1.0 | 0 0 0 0 =
3 }0 1.0 oo 90.0 oo 7
0.1 | 0.9223 | 14.2 69.00 | 48.6 | 3.11
0.2 | 0.8152 | 6.47 52.60 | 24.2 | 2.91
0.4 | 0.6039 | 2.14 37.00 | 10.3 | 2.38
0.6 | 0.4 0.695 26.98 4.93 | 1.74
0.8 | 0.2 0.149 | 19.57 | 2.02. | 1.05 TABLE AT 25 VALUES OF C FOR e = 0.001 in.
0.9 | 0.1 0.0422 | 15.91 1.08 | 0.664
0.97| 0.03 0.00704 | 10.85 0.490 | 0.329 For other values of ¢, multiply the value given by the number of thousandths that e is; for
10 10 0 0 0 ee example, for cast iron and cast iron, 144° full depth, and e = 0.004 in., C = (4)(800) =
3200. For other materials, use C = kE,E,/(E, + E,). Values of C for bronze are virtually
t]o 1.0 ss 90.0 = os the same as for the gray iron (modulii of elasticity about the same). FD = full depth.
0.1 | 0.9251 | 35.8 71.55 |121.0 | 3.16 é R 2
0.2 | 0.8242 | 16.0 58.51 | 58.7 | 3.04 cats Gk ey! en (tee ore le eat
0.4 | 0.6074 | 5.20 41.01 | 24.5 | 2.57 s Sa oP fi Ena.
0.6 | 0.4 1.65 30.14 | 11.2 | 1.98 Gray iron and grayiron . . 800 830 860
0.8 | 0.2 0.333 21.70 4.27 | 1.30 :
0.9 | 0.1 0.0844 | 16.87 2.01 | 0.894 Grayironand steel . . . 1100 1140 1180
0.97 | 0.03 0.0110 | 10.81 0.713 | 0.507
1010 0 0 0 mee Steel and steel : ‘ ‘ : 1600 1660 1720

600 601
 on
7 = ae SS ee ee Se

TABLE AT26 VALUES OF LIMITING WEAR-LOAD FACTOR K,

Specified BHIN’s are minimums. Values are for indefinite life unless otherwise
indicated.
Straight line interpolations on the sum of BHN’s are permissible when the difference
in
Permissible Error, ¢

BHN’s is less than 100 points.

0.004
Snsurt K,
COMBINATIONS OF MATERIALS (BHN) AND LIFE ksi 144° 20°
0.003

S
BOTH GEARS STEEL:
Sum of BHN = 300, 10° cycles
0.002 63 86
Ditto
™
10? cycles a] nes 40 54
Ditto 4 x 107 cycles or more 50 30 41
Sum of BHN = 350 od 60 43 58
Sum of BHN = 400, 10° cycles
119 162
Ditto 107 cycles se tirdte 1B 715 102
Ditto 4x10" cyclesormore . .
1 1 1 L 1 1L L 4 1 L
70 58 79
1000 =. 2000 3000 «604000 «= 5000s 000—«si«‘i=7000sis« 9000 10000 Sum of BHN = 450
80 76 103
Sum of BHN = 500
Pitch Line Velocity,
feet per minute 90 96 131
Sum of BHN = 550
100 119 162
Sum of BHN = 600, 10° cycles
292 400
FIGURE AF 19 Maximum Permissible Errors in Gear-tooth Profiles. Extreme qu Ditto 107 cycles fo") Sliven aie
will require a smaller error than indicated by this curve. 185 252
Ditto 4x 10’ cyclesormore . . 110 144 196
Sum of BHN = 650
120 171 233
Sum of BHN = 700
130 196 270
Sum of BHN = 750 2 Doe | OMS SRE e 140 233 318
Sum of BHN = 800 Goer ly a aon lee ae 150 268 366
Steel (500) and steel (350) . : 145 250 342
Steel (450) and same ae gi Ee Re 170 344 470
Steel (500), induction hardened, and same, 107 cycles
880 1190
Ditto 10° cycles 670 920
Ditto 10° cycles 405 555
Steel (600), carburized case hardened, and same, 107 cycles
1230 1680
Ditto 10° cycles 940 1280
Ditto 10'° cycles . 550 750
Steel (150) and cast iron =
FIGURE AF 20 Expected Errors in Tooth Profiles. (Data courtesy
Earle Buckingham). _ 50 44 60
Steel (250) and Nicast iron, HT . ale 90 150 205
0.005 Steel (630) and SAE 65 phosphor bronze (7)a
T tT T T T T qT qT T T
53 72
Steel (250 and over) and chilled phosphor bronze .
83 128 175
Steel (630) and laminated phenolic* .
46 64
Cast iron, class 20, and same* .
0.004 81 112
Cast iron and same, 10° cycles .
Expected Error in Inches,e

376 515
Ditto 107 cycles .
Ditto 212 290
4 x 107 cyclest
150 205
0.003 Cast iron with steel scrap and same . wT Ve 170 230
Cast iron, class 30, austempered (270) and same* .
G. M. Meehanite (190) and same* 224 306
e Bie 104 142
First Class Commercial Gears Nodular iron casting, 80-60-03 (210) and same* .
0.002 180 248
Cast iron and phosphor bronze. .. . Bie coe al 83 170 234
Cast iron, class 30 (340) and cast aluminum, SAE 39 (60)* . 16 22
i Carefully Cut Gears
0,001
| | a * These values adapted from Cram.!2?-22]
i Precision Gears ~ahe ee + These values are not consistent with those from
0.0005
Cram, and probably should be discounted.

i i i L 1 2. i L i i 1
603
ll 10 9 8 7 6 5 4 3 2 1
TABLE AT 27 WEAR FACTOR Ky FOR WORM GEARS TABLE AT 28 PROPERTIES OF WIRE ROPE
Approximate weight of rope is w Ib. per ft.; D, = diameter of rope; D, = diameter of
Taken from Buckingham and Ryffel,'**-1] with permission of the publisher, The Industrial
sheave; A,, = cross-section area of metal, sq. in.; E, = modulus of elasticity of the rope;
Press.
I.P.S. stands for improved plow steel; P.S. for plow steel; M.P.S. for mild plow steel.
THREAD ANGLE ¢, All values are for rope with fiber core. It may occasionally be useful in preliminary computa-
MATERIALS
tions to estimate the ultimate strengths of terms in D,; for example, for 6 x 19 rope:
Worm Gear 1449 20° 2Sian 30% VHS, F, ~ 48D,? tons; IPS, F, + 42D,2; PS F, ~ 36D,2; MPS F, ~ 32D,? tons.
Multiply the values given by 1.075 to obtain strength of IWRC. For 6 x 19 traction steel,
{Hardened steel Chilled bronze 90 125 150 180 F,, = (0.87)(F,, for MPS) will be safe.

60 80 100 120 6 x 7 Wire Rope

{Hardened steel Bronze 6 x 19 Wire Rope 6 x 37 Wire Rope
50 60 72 w = 1.52D,? lb/ft. w 1.6D,? lb/ft. w = 1.55D,?
Steel, 250 BHN (min.) Bronze 36
Min. D, = 42D, in. Min. D, = 30D, in. Min. D, = 18D, in.
High-test C.I. Bronze 80 115 140 165 DIA. | Desirable D, = 72D,in. | Desirable D, = 45D, in. Desirable D, =
OF D, ~ 0.111D, D, ~ 0.067D, 27D, in.
*Gray iron Aluminum 10 12 15 18 ROPE An © 0.38D,? sq. in. A, © 0.4D,? sq. in. Dy & 0.048D, in.
180 Diss E, = 13 x 10° psi E, ~ 12 x 10° psi A, © 0.4D,?
*High-test C.I. Gray iron 90 125 150
in. E, ~ 12 x 108
High-test C.I. Cast steel 22 31 37 45 IWRC: IWRC:
w * 1.76D,? lb/ft. w w 1.71D,?
High-test C.I. High-test C.I. 135 185 225 270

47 64 80 95 NOMINAL BREAKING STRENGTH IN TONS OF 2000 LB., Fy

Steel 250 BHN (min.) Laminated phenolic

Gray iron Laminated phenolic 70 96 120 140 LP.S. P.S. | M.P.S.| LP.S. PS. M.P.S. IP.S. P.S.

* For steel worms, multiply given values by 0.6. + Over 500 BHN surface. i 2.64 2.30 2.00 2.74 2.39 2.07 2.59 2.25
a5 4.10 3.56 3.10 4.26 3.71 3.22 4.03 3.50
3 5.86 5.10 4.43 6.10 5.31 4.62 wile 5.02
as 7.93 6.90 6.00 8.27 7.19 6.25 7.82 6.80

4 10.3 8.96 7.79 10.7 9.35 8.13 10.2 8.85

~5 | 13.0 11.3 9.82 13.5 11.8 10.2 12.9 11.2
3 15.9 13.9 12.0 16.7 14.5 12.6 15.8 13.7
2 22:7 19.8 17.2 23.8 20.7 18.0 22.6 19.6
20
$ 30.7 26.7 23.2 32.2 28.0 24.3 30.6 26.6
18 1 39.7 34.5 30.0 41.8 36.4 31.6 39.8 34.6
# 110 ¥ 1} 49.8 43.3 37.7 52.6 45.7 39.8 50.1 43.5
1} 61.0 53.0 46.1 64.6 56.2 48.8 61.5 53.5

12 73.1 63.6 55.3 Te 67.5 58.8 74.1 64.5

3 03 14 86.2 75.0 65.2 92.0 80.0 69.6 87.9
.3 8g
62
13 107.0 93.4 81.2 103.0
76.4
89.3
12 124.0
E
108.0 93.6 119.0 103.0
ss

1 141.0 123.0 107.0 136.0 118.0

2
2 160.0 139.0 121.0 154.0 134.0
0.25 0.30 0.35 0.40 0.45 0.50 23 179.0 156.0 173.0 150.0
24 200.0 174.0 193.0 168.0
Transmittance hc ft-lb. per min- sq. in.“F

a ; 24 244.0 212.0 236.0

FIGURE AF 21 Transmittance, Gear Cases. To use: Determine the area of the 205.0
is th : 22 292.0 254.0 284.0 247.0
enter chart with this area, move rightward to curve, downward to abscissa, which
transmittance 4,,—as shown by dotted lines.

604
TABLE AT 29 DESIGN DATA FOR BRAKES AND CLUIUHES

From mixed references, mostly (18.8, 18.11, 18.13, 18.22). Unless otherwise indicated,
values of f are for dry surfaces; if greasy or wet, they are much lower. If f at the lower end
of the range (or below) is used in design, this in effect introduces a design factor. Use the
lower pressure P given as the design maximum pressure where possible. The drum tempera-
tures are the maximum values for steady operation. See manufacturers’ catalogs for more
detail.

DRUM MAX.
MATERIAL MAX, f, S P pst VEL.
Se (max.) Spm

Metalon metal . = S é 4 0.2 to 0.25 | 150(250)

Wood on metal . ; ‘ ; ‘ : 150 0.2 to 0.25 | 50(90)
Leather on metalorwood . . . «| 150 0.3 to 0.4 15 (40)
Cork omiron: <a) ae Ns epic ade 0.35 10 (15)
Molded blocksinve Sf “tens at 1650 150 7500

Asbestos in rubber compound, compressed,

pnmetalincm oe | «| | ah liseaniy.s,| 400 0.3 to 0.4 75 (100)
Asbestos in resin binder, molded, on metal.| 500 0.3 to 0.4 75 (100) | 5000
in oil BO Foe Tea ee le Nice 0.10 (600)

Asbestos, flexible woven, on metal . . 300 0.35 to 0.45| 50

AEROS orm ters Ste t we ie Bud os 0.12
Sintered metal oncastiron. . «. -|>400 0.20 to 0.40| 400
in oil (as in automatic transmissions) . 0.05 to 0.08
TABLE AT30 DESIGN STRESSES FOR WELDED JOINTS
(a) Recommendations by Jennings.!*9-] (b) AWS Buil Code. (c) AWS 5
(d) Lincoln Electric.{®-9] (e) Adapted from AWS a. Code a US. ieeg irre
Blodgett;{'®-15) for a common class of structural steel or the equivalent; for n, = 2 x 108
cycles and n, = 10° cycles; R = stress ratio = S,.:/Smax, § 4.5. Maximum values for
structural steel, butt welds, beads left on, say 18 ksi; for alloy steel, 54 ksi; for fillet welds,
12.5 and 37 ksi, respectively. The source does not mention stress concentration factors.
“ - eo lg class ramones butt. (g) Ultimate strength of aluminum alloy welded
Lae — =“
on shie! ed. 9.15 U!
se factor of safety. (h) AISC Building Code,!*-34]
5.34

KIND OF WELD
SHIELDED WELDING | UNSHIELDED WELDING FATIGUE DESIGN STRESSES(€)
AND STRESS CODES, ©
Steady Reversed | Steady Reversed Struct. Steel Q&T Alloy Steel
2 x 10° 10° 2 x 10° 105

BUTTWELDS
16(a) 8(a) 13(a) 5(a) 16 18 16.5 31
Tension alc we Ee ere EO 11@) 1—0.8R 1—0.5R 1—0.8R 1-—0.6R
8(f)

Compression PS
18(a) g(a) | 15(a) 5(a) 6
1—R
Ts
1—0.5R
18(c)

Shears -—i0-4ite se aon) sar eee

10(a)
15(d)
5(a) 8(a) aia Hee
1—0.5R 1—0.5R
Bending.
FILLET WELDS
14(a) 5(a) 11.3(a) 3(a)
All z
11.3(g)
Parallel .

16(d)
Transverse 12.118) 72, 12.5 9 20.5
1—0.5R 1—0.5R 1—0.8R 1—0.6R
 INDEX

A Almen intensity, 132

Aluminum alloys, 63
Abbreviations, summary, xXvii-xix identification numbers, 64
Acceleration, cams, 528, 530 in steel, 49
calculated vs. actual, 529 mechanical properties, 566, 583
Acme thread, 246 relative strength-mass, 65
Addendum, 356 stress-strain diagram, 8
Admiralty metal, 63 temper designation, 64
Age hardening, 42, 45, 64 Aluminum bronze, 63, 566, 583
stainless steel, 61 relative strength-mass, 65
Aging, 42, 45, 64 American screw thread, 156
Airflex clutch, 502 Angle, action 35
AISI specification numbers, 47, 48 contact of belts, 447
uses of steel, 51 eccentricity, 303, 305
Alclad, 64 Angular deflection, 17
Allowance, 80 Angular gears, 422
negative, 80, 85 Anistropy, 42
table, 82 Annealing, 45
versus fit, 82, 85 Appendix, 561-609
Alloy, 42 Arc length of bearings, 304
Alloy steel, 49 Arc welding, 518, 519
mechanical properties, 573, 574, 578, Arithmetic deviation, 94
581, 582 Arms, spur gears, 388
stress-strain diagram, 7 ASA allowances and tolerances (table),
structural steel, 49 82, 85, 94
Alloying elements for steel, 42, 49 ASME Codes, 26
effect of, 49 shaft design, 278
uses, 51 Atomic-hydrogen welding, 520

igh echtt om $ iuleeite ta ioce oh HBG {

610 611

Bearing pressure, 309 rated wear load, 415 see Mechanical properties for particular
Attitude angle, 305
Bearing stress, 26, 29, 31 strength, 409 material
Autofrettage, 124
Bearings, ball, 337, 348 Bevel gears, 407 conversion chart, 575
thick cylinders, 256, 257
gas-lubricated, 331 design details, 424 relation to strength of steel, 46
Axial pitch, 246
hydrostatic, 330 face width, 411 tolerance, 379
helical gear, 399, 400
journal, 299 materials, 424 vs. fatigue strength, 109
worm, 428
partial, 311 mountings, 424 Brinelling, 337
Axles, 263
roller, 341, 349, 350 nomenclature, 407-409 Brittle failure, 100
self-aligning, 315 straight, 407 Brittle material, 42
thrust, 329 types of, 419-422 Brittleness, 42
Belting, 441 Biaxial stresses, 229 low temperature, 69
Babbitt, 66 creep, 445 Bilateral tolerance, 80 Bronze, 62
mechanical properties, 566 density, 444 Block brake, 487 see Copper alloys
Back cone, 408 for nonparallel shafts, 454, 455 Bolt heads, dimensions, 588 Buckling, columns, 211
Backing, 424 law of, 454 Bolted connections, 140, 164 beam flange, 34
Backlash, 356 pivoted-motor drive, 452 load to open, 166 springs, 199
Ball bearing, 333 pulley sizes, 450 Bolts, automobile, 173 Buna-S, 72
coefficient of friction, 347 rated capacity, leather, 449 deformation, 147 Butt weld, 506
design life, 340 reversed bending, 452 design of, 158 Buttress thread, 246
dimensions, 342 rubber, 454 effective strain length, 161
dynamic capacity, 338 slip, 445 endurance strength, 162
equivalent load, 338 for absorbing energy, 149
Cc
stress variation, 442
for axial motion, 353 toothed belts, 463 for repeated loads, 163 Cable, 470
friction, 347 V-belts, 455 in shear, 178 see Wire rope
housing, 350 Belts, angle of contact, 447 inaccuracies in, 160 Camber, 207
life expectancy, 335, 336 coefficient of friction, 445 loading on, 159 Cams, 528
lubrication, 345 crossed, 447 materials, 161 accelerations, 528, 529
materials, 345 force on shaft, 263, 264 shock loads, 147 contact stress, 531
preloading, 348 general formula, 444 strength of materials for, 161 curvature, 533
probability of survival, 342 length, 446 types of, 170-176 equations of motion, 530
rated capacity, 341 net belt pull, 441 Boron, 49 forces, 531
screw, 250 open, 442, 447 Brakes, 481 jerk, 528
sizes, 346 speed, 448 band, 493 wear factors, 532
speed, 341, 353 stresses, 446 block, 487 Carbonitriding, 54
static load capacity, 337 tables for design, 450 capacity, 484, 485, 606 Carburizing, 53
stresses, 334 tension ratio, 453 center of pressure, 491 endurance strength, 136
tables, 339-341 width, 444 coefficient of friction, 500, 606 gear teeth, 382
types of, 347 Bending, 9 disk, 496, 497 mechanical properties, 381,
variable loading, 345 and column action, 223 dissipation of heat, 484 of core, 584
vs. journal bearings, 352 and torsion, 244 energy to be absorbed, 481-484 Cartridge brass, 63
Band brake, 493 and uniform stress, 222 fade, 485 Case hardened steel, 53
differential, 495 forces, by belt and chain, 263 frictional work, 481 see Carburizing, Flame hardening, etc.
heat capacity, 484 moment, 10, 564 internal, 492 Cast iron, 57
pressure on, 495 of belt, 442 long shoe, 489 chilled, 59
Barlow’s formula, 254 of wire rope, 472 pivoted shoe, 490 distribution of strength, 58
Base circle, gears, 357 plastic, 12 pressure, 489, 491, 606 mechanical properties, 570
Base pitch, 359 stress at a point, 10 self-energizing, 487, 494, 497 relative strength, stiffness, 65
Basic hole system of fits, 81 Bending moments, 9, 13, 266, 267 self-locking, 488, 494 stress-strain diagram, 7
Bead of gears, 390 graphical solution, 14 surface materials, 499 white, 59
Beam, curved, 252 Beryllium copper, 62, 63 temperature, 485 Cast steel, 60
impact loading, 151 Bevel gear teeth, 408 Braking torque, 482 mechanical properties, 570
maximum bending moment, 13, 264- dimensions, 414 band brake, 495 Cast teeth, design, 388
267, 275, 564 dynamic load, 412 long shoe, 491 Casting, 73-75
shear, 11 form factor, 411 small block, 488 die, 74
slope of, 13 generating, 420 Brass, 62 hints on design, 75
straight, 9 geometry factor, 413, 415-417 see Copper alloys malleable, 74
stresses, 9 method of design, 412-419 Brazing, 521 steel, 74
uniform strength, 206, 564 proportions, 411 Brinell hardness, 46 vs. welding, 76
612 INDEX

Castle nut, 174 bending and uniform, 222 Creep, of belts, 445 Design stress, 19
Center of gravity, 15 normal and shear, 227-230: secondary, 68 see Design factor
Centipoise, 301 shears in different directions, 226 strength, 67, 69 bevel gear teeth, 414, 415
Centrifugal clutches, 501 varying loading, 242 tertiary, 68 bolts, 159
Centroids, 15 Compression springs, 184 vs. time at high temperature, 68 shafts (Code), 278
Chain drives, 464, 467 buckling, 199 Critical load, columns, 211 springs, 185, 590
see Roller chain concentric, 200 shells under external pressure, 523 spur gears, 366
Charpy test, 42, 70 deflection, 186 tubes under external pressure, 526 welds, 513
tabular values, 571 design, 190 Critical range, 45 Diametral clearance, 300
Chemical symbols, xix design stresses, 185, 590 Critical speed, shafting, 276 Diametral pitch, 358
Chordal action, chains, 464 energy absorbed, 189 ¥ Crossed helical gears, 405 helical gears, 399
Circular pitch, 358 factors affecting fatigue strength, 194 velocity ratio, 406 standard values, 358
helical gears, 399 free length, 190, 589 Rat Crown gear, 422 Die casting, 74
standard values, worms, 429 materials, 193, 590 wa Crowned tooth, 368 Differential back stop, 495
Clearance, bearings, 300, 304, 312, 314 rectangular wire, 200 “ Curved beam, 252 Differential band brake, 494
diametral, 300 scale, 186 i curvature factor, 593 Dimensions, preferred values, 32
gearing, 356, 362 solid height, 190, 589 ie stress, 253 Disk clutch, 496
radial, 300 solid stress, 185, 590 a Cyaniding, 53 coefficient of friction, 498, 606
ratio, 300 spring index, 185 gear teeth, 381 frictional torque, 496
Clutches, 497-502 stresses, 184 Cycloidal motion, 530 plate, 497, 498
stress factors, 592 Cylinders, subjected to external pressure, Disk friction, 496
coefficient of friction, 500, 606
surge, 198 thick, 254 Distortion energy theory, 236
Coarse thread, 157
tolerances, 197 thin, 523 Draw, drawing (see Temper), 45
Coaxing, 142
types of ends, 183 Cylindrical roller bearings, 348 Ductile iron, 59
Coefficient of friction, ball bearings, 347 variable stresses, 187 Ductile material, 42
belts on pulleys, 445 Wahl factors, 592 Ductility, 42
brakes, 500, 606
wire sizes, 589 D Dynamic loads on gear teeth, 368, 385
clutches, 498, 500, 606 Compressive stress, 6 Buckingham, 372
force fits, 259, 260 contact, 36, 334 Damping capacity, 42 cast teeth, 388
journal bearings, 305, 323, 324, 328 ultimate (tables), 566, 568, 570 y Dardelet rivet bolt, 179
power screws, 250
curves of comparison, 373
Concentration of stress, 111 ee Decarburization, 42 helical teeth, 401
roller bearings, 347 see Stress concentration endurance strength, 115, 126
thrust bearings, 330, 331 nonmetallic teeth, 386
Concentric springs, 200 Dedendum, 356 straight bevels, 412
wire rope in sheave, 477 Cone center, 407 Deflection, bending, 13, 271
worm gears, 434
velocity factors, 369
Cone clutch, 499 graphical solution, 272-276 worm gears, 430
Coefficient of thermal (linear) expansion, Coniflex gears, 419 impact, 145, 147, 150-152
38 Conjugate curves, 360 normal stress, 9
see Mechanical properties for particular Conrad ball bearing, 347 of beams (table), 364, 564
material Consistent system of units, 149 shafts, 272, 273
E
Cold finish, 56, 130, 131 impact load formulas, 149 springs, 186 Eccentric loading, 222
Cold shortness, 42 Contact ratio, 360 torsional, 17, 271 bearing, 300
Cold working, 42, 56, 61 helical gears, 399 Deformation, 7, 9 on columns, 223
to improve endurance strength, 138 Contact stress, 36, 334 see Deflection Eccentricity factor, journals, 304
Collar bearings, 316 bevel gear teeth, design, 415 Degree of involute, 357 Eccentricity ratio, 304
Columns, 211 cam systems, 528 Density, 65 Effective length of columns, 212
cast iron, 214 endurance strength, 379 see Mechanical properties for particular Efficiency, power screws, 250
curves comparing formulas, 215 gear teeth, 377 material riveted joints, 35
design, 217 Copper alloys, 62 belting, 444 spur gears, 393
effective length, 212 as alloy in steel, 49 oil, 310 welded joints, 35
equivalent stress, 217 mechanical properties, 566, 583 Design, engineering, 40 worm gearing, 432
Euler formula, 211 Corrosion, on endurance strength, 128 of castings, 74, 75 Elastic constants, connected parts, 168,
factor of safety, 212 fretting, 129 philosophies of, 3 170
formula for magnesium, 218 gear teeth, 380 Design factor, 19, 261 effect of relative stiffnesses, 164, 168
J. B. Johnson formula, 213 Couplings, 290-296 see element involved effect on bold load, 167
secant formula, 216 clamp, 291 columns, 212, 213, 216 Elastic energy, 144
straight line formulas, 213 flange, 290, 291 considerations affecting, 23 see Energy
with eccentric load, 223 flexible, 293 table, 20 ; Elastic limit, 8
Combined stresses, 221 over-running clutch, 297 variable loads, 122 Elasticity, 9
614 INDEX INDEX 615

Elasticity (cont.) Extension springs, 201 circular, supported at edges, 526 load on, 368-370, 372
definition, 42 Extreme-pressure lubricant, 380, 430, 438 rectangular, supported at edges, 527 lubrication, 393
shear, 17 Flat springs, 206 nomenclature, 355-357
Elevated-temperature drawn, 57 Flexible couplings, 293 nonmetallic, 386
mechanical properties, 580 F types, 294, 295 pressure angle, 357
Elongation, 7, 9 Flexure (see Bending), 9 profile errors, 374
Face contact ratio, 399
see Mechanical properties for particular Flywheels, 533 sizes, 359
Face width, bevel gears, 411
material coefficient of fluctuation, 534 strength, 363-366
cast spur gears, 388
Embrittlement, 43 disk, 537 stress concentration, 365
cut spur gears, 367 ‘
End quench test, 51 rim stresses, 536 wear, 377, 380
helical gears, 398° 7
Endurance limit, 101 worm gears, 430, 437 ; Force-and-funicular polygons Gear tester, 376
see Endurance strength Factor of safety, 19, 23, 261 (footnote), 273 Gear trains, 392
Endurance strength, 101-104 see Design factor Force fits, 86 Gears, bevel, 407
see Mechanical properties for particular columns, 212 action in, 129 see Spur gears
material definition, 20 cast iron, 259 helical, 397
axial loads, 110 Failure, theories, 234 force to assemble, 260 spur, 355
beyond yield, 142 modes of, 261 stresses in, 86, 259 worm, 427
bolts, 162, 167 Fatigue limit, 101 table of limits, 85 Geometry factor, bevels, 413, 415-417
butt weld, 515, 516 see Endurance strength torque, 259 Gerber line, 107, 141
chart of, 102, 104 mechanism of, 99 Forging, 72, 73 Gleason geometry factors, 416, 417
cold rolling, 138 of metals, 99, 245 Form factor, gearing, 365 Goodman diagram, 106
combined stresses, 245 of wire rope, 473 bevel gearing, 411, 413 line, 108
contact surface, 379 roller-chain plate, 466 helical gearing, 401 spring wire, 197
corrosion effect, 128 strength, 101, 194 table, spur teeth, 601 Graphical integration, 272
effect of plating, 137 worm gearing, 430 Graphitizing, 45
welded joints, 514, 515
effect of surface, 130-134, 137-139 Fatigue-strength reduction factor, 113 Formative number of teeth, bevel gears, Gray iron, 57
estimations of, 109 interference fits, 585 411 see Cast iron
finite life, 118 helical gears, 401 Grip, 161
keyways, 585
gear teeth, 366 screw threads, 585 Free carbon, 43
limit, 101 surface effects, 585 Fretting corrosion, 129
means of improving, 123-127, 131-140 welded joints, 515 springs, 209
peening effect, 131, 134 Feather key, 286 Friction, bearing cap, 321 Hard drawn, 43
rolling bearings, 336 Fillet weld, 506 brakes, 482 Hardenability, 51
size effect, 115 design stress, 513, 515 journal bearings, 305, 307, 308, 311 alloys, 48, 52
springs, 194, 209 eccentric load, 509-513 rolling bearings, 347 curves, 52
standard deviation of, 103 reinforced, 507 Funicular polygon, 273 Hardening, 45
surface finish effects, 115 size, 507 Fusion welding, 520 case, 53
table, 580 Film thickness, bearings, 305 for endurance, 135
temperature effects, 140 average value in cap, 321 gear teeth, 367, 379, 381, 403
torsion, 111, 117 Norton’s rule, 311 G troubles, 400
welded joints, 515 permissible values, 311 Gages for wire and plate, 589 worm gears, 430
wire rope, 473 Fine screw thread, 157 Gas carburizing, 53 Hardness, 46
Energy, absorbed by spring, 189 Fits, 80 case, 53-56
Gas-lubricated bearings, 331
elastic, 144 basic hole system, 81 conversion chart, 575
Gaskets, 168, 170
from torque, 144 distribution of, 88, 90 gear tooth, 55
Gaussian distribution, 87
kinetic, 483 of threads, 157 numbers, 46
Gear cases, heat loss, 430
potential, 483 probable limits, 86 Gear ratio, 357 Hardness numbers, relation of, 575
Engineering, definition, 40 spline connections, 288 Gear shaper, 375 see Mechanical properties for particu-
Equivalent bending load, wire rope, 472 stress in, 86 Gear teeth, action, 361 lar material
Equivalent load, rolling bearings, 338 running, 80-82 curvature of profiles, 378 vs. ultimate strength, 46
Equivalent number of teeth, bevel gears, Fitted bearing, 300 cutting, 374, 375 Harmonic drive, 479
411 Flame hardening, 55 design stress, 366 velocity ratio, 480
helical gears, 401 endurance strength, 135 face width, 367 Harmonic motion, 530
Equivalent stress, columns, 217 gear teeth, 382 force on shaft, 265 Hastelloy, 66
varying combined stresses, 242 Flange coupling, 290 form factor, 365 mechanical properties, 566
varying stresses, 121 Flat belts, 442 heat treatments, 381, 382 Heat dissipation, bearings, 317
Error in gear teeth, 374 see Belts, Belting interchangeable systems, 362 brakes, 484
Euler’s formula, 211, 212 Flat plates, 526 interference, 361 gear cases, 431
 617
616
permissible film thickness, 310 mechanism of, 303
Heat transfer coefficient, bearings, 317, on body of two areas, 147
pressure distribution, 309 thin film, 323
318 roller chains, 465, 466
gear cases, 604 stress, 149 pressure feed, 319
Heat treatment, definition of terms, 45 Impulsive loading, 143 semilubricated, 327
side leakage, 309 M
effect on steel, 51-56 Increment load, gear teeth, 372
for improving fatigue strength, 135-137 Indeterminate, statically, 37 steady-state temperature, 318 Machinability, 43
gear teeth, 381 Induction hardening, 54, 55 tables and charts, 595-600 tabular values, 571, 577
welds, 516 endurance strength, 135 temperature increase of oil, 310 Machine screws, 176
Helical gears, 397 gear teeth, 382 types of, 300 heads, 176 q
dynamic load, 401 Initial tension, belts, 451 vs. rolling bearings, 352 Magnesium alloys, 64
forces on, 400 bolts, .159 water cooled, 316 column formula, 218
limiting wear load, 402 extension springs, 201 mechanical properties, 566, 583
notation, 398 vs. tightening torque, 159 strength-stiffness comparison, 65
K stress-strain diagram, 7
on non-parallel shafts, 405 Interchangeability, 83
pitch, 399 selective, 84 f Kennedy keys, 285 Major diameter, 156
pressure angles, 400 Interchangeable gear-tooth systems, 362 Keys, 281
Malleability, 43
strength, 401 Interference of metal, 80, 86, 258 i design factor, 282 Malleable iron, 59
Helical springs, 183 stress concentration factors, 585 dimensions, 594 mechanical properties, 570
see Springs stresses due to, 86, 259 materials, 283 Malleablizing, 45, 59
Helix angle, 398, 404 Interference of involute teeth, 361 types, 284
Manganese bronze, 62,63
relation to lead angle, 428 Internal gears, 391 Keyway, effect on design stress, 278 fatigue strength gear teeth, 567
Herringbone gears, 402 wear load, 392 stress concentration factors, 283, 585 steel, 50
Hertz equation, 36, 334 Internal shoe brake, 492 Killed steel, 43 Maraging, 76
gearing, 377 Inverted-tooth chain, 467 Kinematic viscosity, 302 Materials, 41 ;
High-temperature service, 67 Involute, 357 Kinematics of gearing, 355-362 (Look under name of item)
Histogram, 21 gear teeth, 361, 362 Kinetic energy, 483 charts and tables of mechanical
cast iron, 58 serrations, 289 Kingsbury thrust bearing, 330 properties, 566-582
cold drawn steel, 57 spline, 288 Kip, 6 steel specification numbers, 48
distribution of spring force, 198 Isotropic material, 43 Maximum principal stress theory, 234
rolling bearing life, 336 Izod test, 43, 70 Maximum sear stress theory, 235
structural steel, 21 Maximum strain theory, 238
tabular values, 569, 571, 579, 580, eL
titanium yield strength, 66 581, 584 Mechanical advantage, brakes, 485, 488
Hobs, gear teeth, 375 Lamé’s formulas, 255 Mechanical properties, 43
for splines, 289 Lap joint, 140, 507 aluminum alloys, 566, 583
Homogeneous material, 43 Law of belting, 454 cast iron, 570
Hooke’s coupling, 296 J Law of gearing, 360 cast steel, 570
Hooke’s law, 10 Jerk, 528, 530 Lay, surface finish, 94 chart for C1040, 572
Horsepower, equations for, 18 Johnson, J. B., column formula, 213 wire rope, 470 chart for 3140, 573
bearing friction, 311, 313 Jominy test, 51, 52 Lead, 156, 428 charts for 4340, 574
brake friction, 482 Journal bearings, 299 angle, 248, 428 copper alloys, 566, 583
leather belt, 450 attitude angle, 305, 312 square thread, 246, 247 definition, 43
roller chains, 466 worm, 428 Hastelloy, 566
capacity, 305
V-belts, 456 clearances, 300, 312-314 Leaf springs, 206 Inconel, 583
Hubs, gears, 388 concentric, 302 automotive, 208 K-Monel, 566, 583
Hunting tooth, 388 construction, 324 materials, 209 lead and tin alloys, 566
Hydrodynamic lubrication, 303 Lewis’ equation, 365 magnesium alloys, 566, 583
dynamic loading, 332
charts, 596 Lewis’ factor, 365 malleable iron, 570
end leakage, 309
side flow, 309 friction, 305 see Form factor Meehanite, 570
tables for, 597-600 Line of action, 357 Monel, 583
geometric relations, 304
temperature increase, 310 length of, 360 Ni-resist, 570
grooved, 320
Hypoid gears, 421 heat loss, 316 Line of centers, bearings, 304 Nodular cast iron, 570
Locknuts, 179 plastics, 568
hydrodynamic lubrication, 303
Lubricants, bearings, 329 spread of, 21, 57
length/diameter ratio, 306, 314
I lubricant flow, 308 extreme pressure, 380, 430, 438 stainless steel, 568, 582
viscosity, 301 tables of, 566-584
Impact, 143 materials, 326
effect of mass, 149, 152 oil grooves, 320, 321 Lubrication, of journal bearings, 316, titanium alloys, 566, 583
horizontal motion, 150 optimum, 306 320, 325, 328 wrought iron, 578
on beams, 151 partial, 311 gear teeth, 393, 394 wrought steel, 578-582, 584
618 INDEX 619

Mechanical properties (cont.) variability of fits, 88 Physical properties, 43 straight bevels, 411
see Wrought steel variability of spring scale, 198 see Mechanical properties worm gearing, 430
zinc, 566 variability of strength of steel, 21 Pillow block, 316, 325 Pressure distribution in journal bearing,
Meehanite, 570 Normal pitch, 399 Pinion, 357 309
Minor diameter, 156 worms, 428 Pitch, axial, 246, 399, 428 Pressure lubrication, flow of lubricant,
Miter gears, 422 Normal pressure angle, helical, 400 base, 359 320
Modulus line, 8 worm gearing, 430, 431 circular, 358 Pressure vessel, external pressure, 523
Modulus of elasticity, tension, 9 Normal stress, 6 diametral, 358 thick wall, 254
see Mechanical properties for particu- with shear, 227 helical gears, 399 thin wall, 34
lar material Normalizing, 45 normal, 399, 428 titanium, 35
compression, 9 Notch sensitivity, 113 roller chain, 464, 468 Prestressing, 124-126
gasket materials, 170 curves, 586 screw threads, 156, 246 leaf springs, 209
relation to shear, 18 of cast iron, 114 springs, 185 Principal stresses, 228, 230, 231
shear, 17 Nut locking, 179 worm threads, 428 failure theory, 234
wire rope, 605 Nylon, 71, 386 Pitch angle, bevel gears, 408 related to octahedral shear, 237
Modulus of spring, 186 springs, 185 Probable limits, 86
Mohr’s circle, 230-233 spur gears, 359 of fatigue strength, 102
Molybdenum disulfide, 160 oO Pitch circle, 355 of rolling bearing life, 335, 336
vs. fretting, 131 Pitch cone, 407 Problems, 1
Moment diagram, 14, 266, 267, 564 Obliquity, angle of, 357
Pitch helix, 428 Profile contact ratio, 399
Moment of inertia, 14 Octahedral shear, 236
Pitting, 377 Progressive fracture, 100
computing, 15 design equation for, 239
Pivoted-motor drive, 452 Proof load, 161
in columns, 211, 216 theory, 237
Pivoted shoe brake, 490 Proof stress, 8
table, 563 varying stresses, 243
maximum pressure (table), 606 bolts, 161, 164
Moments and deflections in beams, 13 Offset, for yield strength, 8
Planetary gear trains, 392 definition, 43
(table), 564 Offset link, 466
Plasticity, 43 Properties of steel
Monel, 66, 583 Oil grooves, 320
Plastics, 71 see Mechanical properties
mechanical properties of K-Monel, Oils, viscosities (chart), 595
Optimum bearings, 306, 308 adhesives, 521 Proportional limit, 8
566, 583 Pulleys, 477, 478
tables, 597-600 mechanical properties: cellulose ace-
Osculation, degree of, 334 tate, phenol-formaldehyde, polya- minimum diameters, 450
mide, polychlorotrifiuoroethylene, Pyroceram, 76
Over-running clutch, 297
N polyethylene, polymethyl methacry-
late, polystyrene, polytetrafluoro-
Natural spread of dimensions, 87
P ethylene, polyvinyl chloride, poly- Q
Natural tolerance, 87 vinylidene chloride, urea-formalde- Quarter turn belt, 456
Naval brass, 63 Palmgren’s equation, 338 hyde, 568 Quenching and tempering, 46, 62
Needle bearing, 349 Parabolic motion, 530 Plate clutch, 497
Neg’ator springs, 205 Partial bearing, 300, 311 Poisson’s ratio, 18
Neoprene, 72 friction in cap, 321 see Mechanical properties for particu-
Net belt pull, 263, 441 Peening, 131 lar material Radial clearance, 300
Newtonian fluid, 301 and endurance strength, 130, 134 definition, 43 Radiation, bearing, 317
Nickel, 50 effect on transition temperature, 70 Polyethylene, 71 gear case, 431
at low temperature, 70 of springs, 196, 209 Power screws, 246 Radius of gyration, 211, 215
maraging steel, 76 power chains, 467 coefficient of friction, 250 table, 564
TD, 76 residual stress from, 132 conditions for self-locking, 251 Rate of spring, 186 :
Nip, 210 test for, 132 dimensions, 247 Rated capacity, ball bearings, 341
Nitriding, 54 Percentage elongation, 8, 42, 43 Precipitation heat treatment, 43, 61 leather belts, 450
endurance strength, 137 tabular values, 567, 568, 576, 578, 579 Preferred sizes, 32 roller chains, 466, 469
gear teeth, 382 Percentage reduction of area, 42, 43 Preloading ball bearings, 348 V-belts, 456
Nodular cast iron, 59 tabular values, 568, 576, 578 Presetting, 125, 134 Red shortness, 43
mechanical properties, 570 Petroff’s equation, 302 fatigue of springs, 195, 209 Reduction of area, 43
Nominal size, 79 Phenolics, 71 Press fit, 86 see Mechanical properties for particu-
Nominal stress, 21, 109 table of properties, 568 see Interference of metal lar material
Nonmetallic gear teeth, 386 Phosphor bronze, 62, 63 Pressure angle, 357 References, list of, 541-559
Normal curve, 21, 86 fatigue strength, gears, 567 helical gears, 400 Relative strength, 35
area under, 89 Photoelastic study, 111, 112 normal, 400, 432 of belting, 446
definition, 87 curved beam, 255 spiral bevels, 420 Relaxation, 43
variability of dimensions, 89 gear teeth, 365 spur gears, 357 of spring materials, 196
620 INDEX 621

Secant formula, 216 bearings, 328 size, 355

Residual stresses, 43
Section modulus, 10 Skew bevel gears, 421 welded construction, 390
effects on endurance strength, 134
related to varying stresses, 123 table, 563 Slenderness ratio, 211 Square thread, 246
rolling bearings, 334 Sections, properties of, 563 Slip of belts, 445 see Power screws
springs, 195, 209 Self-aligning bearings, 315, 348, 349 Slope of beams, 13, 275, 277 force analyses, 248, 251
Self-energizing, 487 table, 564 proportions, 247
welds, 514
Self-locking, brakes, 488, 494, 495 Soderberg line, 107 Stainless steel, 60
Resistance welding, 519
threads, 251 vs. Goodman, 108 at low temperature, 70
Reyn, 301
worm threads, 435 Soldering, 521 ferritic, 62
Rim for gears, 390
Self-lubricated bearings, 328 Solution heat treatment, 44, 64 mechanical properties, 568, 580, 581
Rimmed steel, 43, 44
Sellers threads, 155 Sommerfeld number, 305 Standard deviation, 89
Ring-oiled bearing, 316
Service factor, bevel gears, 415 charts and tables, 596-600 of endurance strength, 103
construction, 325
leather belting, 450 Spalling, 380 Standard fractions, 32
Ring seams, stress, 35
rolling bearings, 339. Specific gravity of oil, 302 Steel, see Wrought steel, Alloy steel, or
Rivets, 179
spur gears, 372 Specific heat, cast iron, 485 Cast steel
heads for, 181 oil, 310
in tension, 181 V-belts, 460 Step cone, 478
Set screws, 175, 176 steel, 485 diameters of steps, 478
Rockwell hardness, 46
holding capacity of, 177 Specific weight (see Density) Stiffness, 9
conversion chart, 575
types of, 175, 177 Spheroidizing, 45 definition, 44
scales for, 47
Setting out, 195 Spiral angle, 420 Stiffness-mass ratio, 65
Roller bearings, 349 Spiral bevel gears, 420
Roller chain, 464 Shafts, 263 Straight bevel gears, 409-418
ASME Code, 278 geometry factors, 413, 415-417 forces on, 422
chain coupling, 294 thrust, 423
dimensions, 468 bending force due to belt, 263, 264 Straight line column formulas, 213
commercial sizes, 269 velocity factor, 413
failure phenomena, 465 Strain, 7
critical speeds, 276 Spiral gears, 404
length, 466 bodies of varying area, 147
deflection, 271 Spiroid gears, 422
materials, 467 maximum, theory, 238
hollow, 270 Spline fittings, dimensions, 287-289 shear, 17
permissible speeds, 465
rated capacity, 466, 469 kinds of, 263 Sprag-type overrunning clutch, 297 unit, 8
sprocket, 478 materials, 269 Spring constant, 144, 186 Strain hardening, 44, 56
Root mean square, 94 service factors, 279 in series, 148 aluminum alloys, 64
Roughness, surface (see Surface finish), square, 270 Spring gradient, 186 Strength, 7
94 tolerances, 269 Spring index, 185
and design factor, 22
vertical shear, 270 Spring materials, 193, 209
Rubber bearings, 327 endurance, 101
Shallow quenching; 139 properties of, 196, 590
Rubber belts, 454 of materials, 18
Shear center, 14 relaxation of, 196
Rubber springs, 205 tensile, 7
Shear diagram, 14, 266, 267 Spring scale, 144, 169, 186
Rupture strength, 67 ultimate, 8
Shear pin, 289 Springs, 183 variation of, 21
Shear stress, 7 see Compression springs, etc. yield, 8
diagram, 14 compression, 184
Ss in more than one direction, 227 Strength-mass ratios, 65
concentric, 200 Stress, 6
SAE specification numbers, 47 maximum, 229, 230 extension, 201
octahedral shear, 237, 240 analysis, 27
Safe stress, 19 fatigue strength, 187, 194 bearing, 26
Scale of spring, 144, 186 resultant, 227, 229, 230 flat, 206
theory, 235 bending, 9, 11
Scleroscope hardness, 47 leaf, 206
thick cylinders, 255 combinations, 221
Scoring, 380 materials, 193
ultimate, 566, 570, 577 compressive, 6, 26
Screw fastenings, 155 miscellaneous kinds, 204 computed, 21, 24
elastic analysis, 164 vertical (beams), 13 torsion, 202
Sheaves, V-belts, 455 concentration, 111
fatigue strength reduction factors, 585 variable stresses, 187 contact, 36
threads for, 155, 156, 588 stock sizes, 458 Sprocket, 478 design, 19
unified standard, 156, 588 Shore hardness, 46 Spur gears, 355 elastic limit, 8
Screws, 173 conversion chart, 575 at extreme speeds, 394 flexural, 9
cap, 173 Shot peening (sce Peening), 131 design of arms, rim, etc., 388 nominal, 21
hole for, 178 Shrink fits, 86 efficiency, 393 principal, 228, 230
lag, 176 Silent chains, 467 gear case, 395 relieving, 46
machine, 176 Silicon bronze, 63 materials, 381 ; shear, 7, 13, 15
power, 246 Simple beam, 564 nomenclature, definitions, 355-359 tensile, 6
self-tapping, 176 uniform strength, 207 service factors, 372 theories of failure, 234-239
set, 176, 177 Sintered metals, 76
 INDEX 623
Stress (cont.) T standard, 94 maximum stress, 114
torsion, 16 Tapped hole, depth, 178 tables, 82, 85, 94 springs, 187
ultimate, 8, 20 Teeth (see Gear teeth) unilateral, 80 stress ratio, 105
variation in belting, 442 Teflon, 72 vs. cost, 84 welded joints, 514
yield point, 8, 20 Temper, 44 vs. natural spread, 88 V-belts, 455
Stress concentration, 111 copper alloys, 63 Toothed belts, 463 center distance, 457
and variable loading, 116 hard drawn, 43 Torque, 15, 18, 282 design factors (tables), 459
brittle materials, 122 stainless steel, 61 brake, 482 force on shaft, 264
cast iron, 114 Tempering, 46 disk friction, 496 lengths (table), 458
keyway, 283 Tensile strength, 7 force fit, 259 rated power, 456
reducing effects of, 139 Tension pulley, 451 from horsepower, 18 service factors, 460
springs, 185 T-head, stress concentration values, 583 on square-thread screw, 248, 251 sheave, 478
Stress concentration factors, 112 Thermal capacity, gear case, 430 to tighten bolt, 159 sizes, 456
actual, 113 Thermal expansion, 38 Torsion, 15 stress variation in, 442
as affected by material, 113 (see Mechanical properties) bar, 184 variable speed drive, 462
curved beams, 595 stresses, 38 deflection from, 16, 17 V-flat drives, 461
elliptical hole, 127 Thermal fatigue, 141 design for, 25 Velocity factors, 369, 370
equations with, 116 Thermal shock, 141 hollow round, 16 bevel gears, 413
eyebar, 585 Thermoplastic, 71 square sections in, 16 nonmetallic teeth, 386
finite life, 119 Thermosetting, 71 Torsion springs, 202 Velocity ratio, 357
gear teeth, 365, 366 Thick-film lubrication, 300 curvature factors, 593 see particular item
interference fits, 585 in cap, 321 deflection, 203 Vertical shear, 11, 13, 270
keyway, 585 Thick-shell cylinder, 254 Torsional endurance strength, 111 Vickers hardness, 46
plate with fillets, 587 interference fits, 257 Toughness, 44 Virtual number of teeth, 401, 411
plate with groove, 587 Thin-film lubrication, 323 at low temperature, 70 Viscosity, 301
plate with hole, 586, 588 lubricants for, 323 Traction drive, 476 chart of oils, 595
screw threads, 585 Thin-wall pressure vessel, 34 Transformation range, 46 index, 595
shaft with fillet, 588 plate thickness, 35 ‘i Transmittance, bearings, 317 kinematic, 302
shaft with groove, 589 Thin-wall shells under external pressure, — gear cases, 431, 604 units, 301
shaft with hole, 589 523 Transmitted load, gearing, 368 von Mises-Hencky theory, 236
springs, 594 stiffeners, 525 , worm, 432, 433
superposed, 123 Thread angle, 156, 252 i Transverse modulus, 17
surface finishes, 115, 585 Thread locking devices, 179 wi Transverse strength, 44 Ww
theoretical, 112 Threads, 156 table, 570 Wahl factors, 592
T-head, 585 Dardelet, 179 Triaxial stresses, 229 Waviness, 94
welded joints, 515 fits of, 157 Tubes subjected to external pressure, 526 Wear
Stress raisers, 111 for bolts and screws, 157 Turnbuckle, 175 bevel gear, 415
see Stress concentration standard, 156 cam surfaces, 531
Stress relieving, 124 stress-concentration factor, 585 helical gears, 402
Stress-strain diagrams, 7 symbol for, 158 U limit load, 378
Stribeck equation, 337 tables of properties, 158, 588 on gear teeth, 380
Ultimate strength, 8
Suddenly applied load, 146 Thrust bearing, 329 spur gears, 377
charts and tables, 566-582
Superalloy, 67 ball, 351 worm gears, 430
Unequal addendum and dedendum gears,
Superfinish, 97 collar, 329 Weibull distribution, 336, 344, 345
390
Surface effects, 115 hydrostatic, 330 Weight, specific (see Density)
Uniaxial stress, 229
machining, 139 Kingsbury, 330 Welded joints, 505 -
Unified screw thread, 156
peening, 131, 134 roller, 351 Unilateral tolerances, 80 computed stresses, 506-513
Timing belts, 463 design stresses, 513, 515, 607
plating, 137 Universal joints, 296
Tolerance, 79 endurance strength, 515, 517
rolling, 130, 131, 134 with needle bearings, 349
bilateral, 80 peening, 516
Surface finish, 93 Universe, statistical, 87
cold-finished shafting, 269 testing, 520
bearings, 310, 311 keys, 594 types, 506-508, 518
designation on drawing, 96 location of holes, 92 variable loads, 514
effect on endurance strength, 115, 585 Vv
over-all, 91 Welding, rolled shapes for, 521
relation to manufacturing processes, 95 preferred values, 81 Variable speed devices, 462 vs. casting, 76
relation to tolerance, 93 related to surface roughness, 93, 95 Varying stress, 99, 104 Welds, bend test, 520
uses, 96 shafting, 269 combined stresses, 242 fatigue data, 515
Surge in springs, 198 springs, 197 factor of safety, 107 minimum size, 518
624

Welds (cont.) efficiency, 433, 434

processes for, 518-520 forces, 433, 435
rod, 514 housing area, 431
stress relieving, 516, 517 materials, 438, 440
X-ray, 520 pitch and lead, 428
Whip, 331 pressure angle vs. lead angle, 436
Whirl, 331 standard pitches, 429
White cast iron, 59 thermal capacity, 430
Whitworth threads, 156 velocity ratio, 429
Wire rope, 469 Wrought iron, 57
bending stress, 472 mechanical properties, 578
fatigue, 473 Wrought steel, 45
fittings, 477 at low temperature, 69
lay, 470 mechanical properties, A
materials, 471 carbon steels, various treatments, 580.
modulus of elasticity, 472, 605 carburized steels, 584
preformed, 470 heat-treated steels, 581
properties of (table), 605 miscellaneous, 578, 582
sheave, 478 transition temperature, 69
traction drive, 476 uses, 48
types, 470, 471
Woodruff key, 286
Work hardening, 44 Y
by cold work, 56
Working stress, 19 Yellow brass, 63
see Design stress and Design factor relative strength-mass, 65
Yield point, 8
Worm gear, 427
Yield strength, 8
design procedure, 439
dynamic load, 430
see Mechanical properties for particu-
lar material
face, 437
in torsion, steel, 576
form factor, 430
spread of, 57
pressure angle, 430, 431, 436
strength, 429
throated, 437
wear load, 430 Z
Worm gearing, 427 Zerol bevel gears, 419
coefficient of friction, 434 tooth dimensions, 414
Cone, 438 Zinc alloy, 67
double-reduction, 435 Zytel, 72, 386