ASME B89.4.

22-2004

Methods for
Performance
Evaluation of
Articulated Arm
Coordinate
Measuring
Machines

A N A M E R I C A N N AT I O N A L STA N DA R D
Intentionally left blank
 ASME B89.4.22-2004

Methods for
Performance
Evaluation of
Articulated Arm
Coordinate
Measuring
Machines

AN AMERICAN NATIONAL STANDARD

Three Park Avenue • New York, NY 10016

 Date of Issuance: August 12, 2005

This Standard will be revised when the Society approves the issuance of a new edition. There will be
no addenda issued to this edition.

ASME will issue written replies to inquiries concerning interpretation of technical aspects of this
Standard.

ASME is the registered trademark of The American Society of Mechanical Engineers.

This code or standard was developed under procedures accredited as meeting the criteria for American National Stan-
dards. The Standards Committee that approved the code or standard was balanced to assure that individuals from com-
petent and concerned interests have had an opportunity to participate. The proposed code or standard was made avail-
able for public review and comment that provides an opportunity for additional public input from industry, academia,
regulatory agencies, and the public-at-large.
ASME does not approve, rate, or endorse any item, construction, proprietary device, or activity.
ASME does not take any position with respect to the validity of any patent rights asserted in connection with any items
mentioned in this document, and does not undertake to insure anyone utilizing a standard against liability for infringe-
ment of any applicable letters patent, nor assume any such liability. Users of a code or standard are expressly advised
that determination of the validity of any such patent rights, and the risk of infringement of such rights, is entirely their
own responsibility.
Participation by federal agency representative(s) or person(s) affiliated with industry is not to be interpreted as gov-
ernment or industry endorsement of this code or standard.
ASME accepts responsibility for only those interpretations of this document issued in accordance with the established
ASME procedures and policies, which precludes the issuance of interpretations by individuals.

No part of this document may be reproduced in any form,

in an electronic retrieval system or otherwise,
without the prior written permission of the publisher.

The American Society of Mechanical Engineers

Three Park Avenue, New York, NY 10016-5990

Copyright © 2005 by
THE AMERICAN SOCIETY OF MECHANICAL ENGINEERS
All rights reserved
Printed in U.S.A.
 CONTENTS
Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv
Committee Roster . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
Correspondence With the B89 Committee . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1 Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
2 Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
3 Environmental Specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
4 Environmental Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
5 Machine Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
6 Test Equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

Figures
1 AACMM Classification Form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2 Environmental Specification Form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
3 Performance Specification Form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
4 Typical Machine of the 2-1-2 Configuration, With a-b-d-e-f deg Rotation . . . . . . . . . . . 5
5 Typical Machine of the 2-2-2 Configuration, With a-b-c-d-e-f deg Rotation . . . . . . . . . . 5
6 Typical Machine of the 2-1-3 Configuration, With a-b-d-e-f-g deg Rotation . . . . . . . . . . 5
7 Typical Machine of the 2-2-3 Configuration, With a-b-c-d-e-f-g deg Rotation . . . . . . . . 5
8 AACMM Mounted Vertically for Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
9 AACMM Mounted Horizontally for Use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
10 Mounting Stiffness Test Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
11 Three Default Locations of the Mounted Seat During the Single-Point
Articulation Performance Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
12 Isometric View of Default Articulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
13 Default Five-Point Probe Pattern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
14 Octant Numbering Scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
15 Recommended Ball Bar Locations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
16 Volumetric Performance Test Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

Tables
1 Diameter Deviations . . . . . . . . . . . . . . . . . . .................................. 10
2 Single-Point Articulation Performance Test .................................. 13
3 Recommended Artifact Positions . . . . . . . . .................................. 16
4 Volumetric Performance Test Results . . . . .................................. 16

Nonmandatory Appendices
A User’s Guide to ASME B89.4.22 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
B Thermal Environment Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
C Seismic Vibration Verification Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
D Electrical Power Monitoring Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
E Interim Testing of AACMM Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
F Ball Bar Design and Mounting Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
G Kinematic, Conical Seat, and Chamfered Hole Design Recommendations . . . . . . . . . . 36
H Determination of Thermal Error Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
I Statistics Used in Specifying AACMM Performance Evaluation . . . . . . . . . . . . . . . . . . 42
J Application of Decision Rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

iii
 FOREWORD

The ambiguity of articulated arm coordinate measuring machines (AACMM) specifications

makes comparative evaluations of performance characteristics very difficult. Because of this and
the increasing use of this class of measurement equipment, the ASME Standards Committee B89
elected to establish a USA industry standard applicable to these machines. At the October 1994
meeting, Project Team B89.4.22 was established to develop the Standard.
As far as possible, this Standard parallels ASME B89.4.1b-2001 for “conventional” coordinate
measuring machines. An attempt has also been made to make the Standard compatible with ex-
isting and emerging international standards.
This Standard addresses the performance evaluation of AACMM by supplying definitions and
test procedures. These procedures should enable users to determine if an AACMM is appropri-
ate for their specific requirements. It should also provide accurate comparison of machines from
different suppliers and provide a determination of whether an AACMM meets contractual re-
quirements without negotiations after the machine has been purchased.
The intent of this Standard is to specify the simplest methods that can be used for reasonable
performance evaluation. It is recognized that a more complex evaluation may be appropriate for
special applications. These methods, however, must be specified in adequate detail in the
AACMM specification.
This Standard was approved by the American National Standards Institute on August 9, 2004.

iv
 ASME B89 STANDARDS COMMITTEE
Dimensional Metrology
(The following is the roster of the Committee at the time of approval of this Standard.)

OFFICERS

B. Parry, Chair
D. Beutel, Vice Chair
M. Lo, Secretary

COMMITTEE PERSONNEL

D. Beutel, Caterpillar, Inc.

J. B. Bryan, Bryan Associates
T. Carpenter, USAF, Newark AF Base
T. Charlton, Jr., Charlton Associ8s
D. Christy, Mahr Federal
G. Hetland, International Institute of Geometric Dimensioning and Tolerancing
R. J. Hocken, University of North Carolina
M. Leibers, Professional Instruments Co.
M. Lo, The American Society of Mechanical Engineers
B. Parry, The Boeing Co.
S. D. Phillips, National Institute of Standards and Technology
J. Salsbury, Mitutoyo America Inc.
D. Swyt, National Institute of Standards and Technology
B. R. Taylor, Renishaw PLC

SUBCOMMITTEE 4 — COORDINATE MEASURING TECHNOLOGY

S. D. Phillips, Chair, National Institute of Standards and Technology

D. Beutel, Caterpillar, Inc.
T. Carpenter, U.S. Air Force, Newark AF Base
T. Charlton, Jr., Charlton Associ8s
T. D. Doiron, National Institute of Standards and Technology
J. L. Henry, Sheffield Measurement
R. B. Hook, Metcon
B. Parry, The Boeing Co.
R. C. Veale, National Institute of Standards and Technology
D. Wright, Renishaw PLC

PERSONNEL OF WORKING GROUP B89.4.22 — ARTICULATED ARM COORDINATE MEASURING MACHINES

B. Parry, Chair, The Boeing Co.

G. Caskey, University of North Carolina at Charlotte
S. Ilmrud, Romer CimCore Inc.
S. Raab, FARO Technologies
D. Wright, Renishaw PLC

SPECIAL ACKNOWLEDGEMENT

While not a member of the Project Team when the document was approved, we would like to acknowledge the work performed and
significant contributions made by Michael O’Laughlin, while an employee of Romer Inc. Mike was the Secretary since the inception of
the Project Team and, thanks to his diligent minutes, kept the remainder of us on track.

v
 CORRESPONDENCE WITH THE B89 COMMITTEE

General. ASME Standards are developed and maintained with the intent to represent the con-
sensus of concerned interests. As such, users of this Standard may interact with the Committee
by requesting interpretations, proposing revisions, and attending Committee meetings. Corre-
spondence should be addressed to:
Secretary, B89 Standards Committee
The American Society of Mechanical Engineers
Three Park Avenue
New York, NY 10016-5990
Proposing Revisions. Revisions are made periodically to the Standard to incorporate changes
that appear necessary or desirable, as demonstrated by the experience gained from the applica-
tion of the Standard. Approved revisions will be published periodically.
The Committee welcomes proposals for revisions to this Standard. Such proposals should be
as specific as possible, citing the paragraph number(s), the proposed wording, and a detailed de-
scription of the reasons for the proposal, including any pertinent documentation.
Interpretations. Upon request, the B89 Committee will render an interpretation of any re-
quirement of the Standard. Interpretations can only be rendered in response to a written request
sent to the Secretary of the B89 Standards Committee.
The request for interpretation should be clear and unambiguous. It is further recommended
that the inquirer submit his/her request in the following format:
Subject: Cite the applicable paragraph number(s) and provide a concise description.
Edition: Cite the applicable edition of the Standard for which the interpretation is be-
ing requested.
Question: Phrase the question as a request for an interpretation of a specific require-
ment suitable for general understanding and use, not as a request for an ap-
proval of a proprietary design or situation.
Requests that are not in this format may be rewritten in the appropriate format by the Commit-
tee prior to being answered, which may inadvertently change the intent of the original request.
ASME procedures provide for reconsideration of any interpretation when or if additional in-
formation that might affect an interpretation is available. Further, persons aggrieved by an in-
terpretation may appeal to the cognizant ASME committee or subcommittee. ASME does not “ap-
prove,” “certify,” “rate,” or “endorse” any item, construction, proprietary device, or activity.
Attending Committee Meetings. The B89 Standards Committee regularly holds meetings that
are open to the public. Persons wishing to attend any meeting should contact the Secretary of the
B89 Standards Committee.

vi
 ASME B89.4.22-2004

METHODS FOR PERFORMANCE EVALUATION OF

ARTICULATED ARM COORDINATE MEASURING MACHINES

INTRODUCTION ment ranges, and rotary axis encoder resolution. Envi-

ronmental specification includes thermal response, elec-
The primary purpose of this Standard is to clarify the trical requirements, and vibration sensitivity. Machine
performance evaluation of articulated arm coordinate performance specification includes effective diameter
measuring machines (AACMMs). A secondary purpose test, single-point articulation performance, and volu-
is to facilitate performance comparisons between ma- metric performance tests.
chines. Definitions, environmental requirements, and Within this Standard, performance values are reported
test methods are specified. This Standard defines the test as the maximum deviation, the range, and the standard
methods capable of yielding adequate results for the ma- deviation. This is done to bring the Standard more into line
jority of articulated arm coordinate measuring machines with existing national and international standards.
and is not intended to replace more complete tests that In order to clarify the use of this Standard, a short
may be required for special applications. guide on how to use it is included as Appendix A.
Productivity is an important consideration in the se-
1 SCOPE lection of an articulated arm coordinate measuring ma-
chine. There are numerous factors that affect the rela-
The scope of this Standard pertains to the perform- tive productivity of measuring systems, which include
ance evaluation of articulated arm coordinate measur- variables attributable to both the measurement system
ing machines. While any number of rotational joints can and the workpiece. This Standard does not address
be evaluated, the Standard focuses on the more com- methods to specify and evaluate productivity. Produc-
mon configurations commercially available today and tivity should be evaluated with respect to the expected
is limited to seven joints. The Standard addresses purely use of the system, including such aspects as software,
manual machines, so no motorized axes are addressed ergonomics, and the frequency of calibration.
in the current document. While the application of this
class of measuring machine continues to grow, at this 1.1 Contents and Specification Forms
point in time only contact probes are considered and op-
tical noncontact probes are specifically excluded. Any specification described as complying with this Stan-
This Standard establishes requirements and methods dard shall include, as a minimum, the following items:
for specifying and testing the performance of AACMMs. (a) a machine classification form (see Fig. 1). If no
In addition to clarifying the performance evaluation of classification is applicable, the actual configuration shall
AACMMs, this Standard seeks to facilitate performance be described in equivalent detail.
comparisons among machines by unifying terminology, (b) an environmental specification form (see Fig. 2).
general machine classification, the treatment of envi- (c) a performance specification form (see Fig. 3).
ronmental effects, and data analysis. This Standard at-
tempts to define the simplest testing methods capable 1.2 Alternatives
of yielding adequate results for most AACMMs and it
This Standard allows parts of the environmental test
is not intended to replace more complete tests that may
section to be deferred or bypassed and only the per-
be suitable for special applications.
formance tests to be carried out. This alternative is ac-
This Standard provides definitions of terms applica-
ceptable only if it is agreeable to both the user and the
ble to AACMMs. These definitions are separated into
supplier, and if deferred as specified.
two parts. The first part is a glossary covering technical
terms used throughout this and other ASME Standards.
The second part defines a number of common machine 2 DEFINITIONS
classifications.
2.1 Glossary
The actual specification is subdivided into three
sections: general machine classification, machine envi- 2.1.1 Terms. This glossary contains brief definitions
ronmental requirements, and machine performance. of the majority of technical terms used in this Standard.
Machine classification includes machine type, measure- Some of the definitions listed are used in the non-

1
 METHODS FOR PERFORMANCE
ASME B89.4.22-2004 EVALUATION OF ARTICULATED ARM CMMs

B89.4.22 MACHINE CLASSIFICATION FORM

Machine Classification: __________

Axis a b c d e f g
Degrees of rotation

Nominal arm length: _________________________ mm

Rotary axis encoder resolution in counts/revolution:

Axis a b c d e f g
Resolution

Special, nonstandard configurations, including such features as external rails (both horizontal
and vertical) should be indicated on the following diagram:

Nominal voltage __________________________________ V

Amperage ________________________________________ A
Frequency ________________________________________ Hz
Surge/sag _______________________________________ V

Fig. 1 AACMM Classification Form

mandatory appendices, but are included here for refer- machine. One form of this test consists of continuously
ence purposes. Omissions and clarifications should be recording the output of displacement sensors placed in
reported to ASME (see Correspondence). the position of a probe on the machine reading against
ball bar: gage consisting of two highly spherical tooling a sample workpiece over a period of time. Detailed pro-
balls of the same diameter connected by a rigid bar. A cedures for conducting drift tests on machines are given
ball bar, as used in this Standard, must be sufficiently in Appendix H.
mechanically rigid that its length is constant during the environmental temperature variation error (ETVE): estimate
course of a set of measurements. of the maximum possible measurement uncertainty in-
bias (of a measuring instrument): systematic error of the duced solely by deviation of the environment from av-
indication of a measuring instrument. erage conditions.
NOTE: The bias of a measuring instrument is normally estimated gage: mechanical artifact used either for checking a
by averaging the error of indication over an appropriate number workpiece or for checking the performance of a ma-
of repeated measurements. chine, or measuring device with a proportional range
drift: slow change of a metrological characteristic of a and some form of indicator, either analog or digital.
measuring instrument. hysteresis: as applied to a measuring system, the prop-
drift test (thermal): type of test used to measure the en- erty of that system whereby its response to a given stim-
vironmental temperature variation error (ETVE) on a ulus depends on the sequence of the preceding stimuli.

2
METHODS FOR PERFORMANCE
EVALUATION OF ARTICULATED ARM CMMs ASME B89.4.22-2004

B89.4.22 ENVIRONMENTAL SPECIFICATION FORM

Thermal Requirements
Significant mean temperature range
Min. ________ °C Max. _______ °C
Safe operating temperature range
Min. ________ °C Max. _______ °C

Gradients Temporal Spatial

Vertical ________ °C/h ________ °C/m
Horizontal ________ °C/h ________ °C/m
Thermal settling time _________________ min

Vibration Requirements
Total vibration amplitude _____________
m
Frequency range _____________ Hz

Mounting Requirements
Maximum applied forces and torques at the mounting interface:
X-direction translational force __________________________________ N
Y-direction translational force __________________________________ N
Z-direction translational force __________________________________ N
X-direction torque __________________________________________ N-mm
Y-direction torque __________________________________________ N-mm
Z-direction torque __________________________________________ N-mm

Maximum distances and displacements acceptable at the mounting interface:

X-direction translational distances ____________________________________ mm
Y-direction translational distances ____________________________________ mm
Z-direction translational distances ____________________________________ mm
X-direction torsional displacement ___________________________________________
rad
Y-direction torsional displacement ___________________________________________
rad
Z-direction torsional displacement ___________________________________________
rad

Limitations on AACMM mounting orientation (please describe fully any limitations in the space below):

Fig. 2 Environmental Specification Form

3
 METHODS FOR PERFORMANCE
ASME B89.4.22-2004 EVALUATION OF ARTICULATED ARM CMMs

B89.4.22 PERFORMANCE SPECIFICATION FORM

Probe type _________________ Probe tip diameter __________ mm

Test procedure used

5.3.2.1 ___________
5.3.2.2 ___________ Max. Diameter Deviations,
m

Effective diameter
performance test

Max. Dev.,
m 2sSPAT,
m

Single-point articulation performance test within 20% of the arm length

at AACMM location (X, Y, Z)

Single-point articulation performance test between 20%–80% of the arm length

at AACMM location (X, Y, Z)

Single-point articulation performance test outside 80% of the arm length

at AACMM location (X, Y, Z)

Max. Dev.,
m Range,
m 2 RMS,
m

Volumetric performance test

Fig. 3 Performance Specification Form

mean ambient temperature: mean temperature of the am- (X, Y, Z) when the AACMM is articulated through the
bient environment surrounding a machine as computed maximum possible range of motion for that single point.
from at least two readings taken at the center of the ma- supplier: party who contracts, or indicates readiness to
chine’s work zone during the interval required for a per- contract, to supply an AACMM to a user.
formance test. The time between the two readings
should be at least two-thirds of the test interval. thermal error index: summation, without regard to sign,
of the estimates of all thermally induced measurement
measurement point: point in the work zone of a machine errors expressed as a percentage of the working toler-
at which machine coordinates are recorded as part of a ance.
measurement.
NOTE: It should be noted that for historical reasons and to main-
performance test: any of a number of test procedures that tain compatibility with other B89 standards, the term thermal error
are used to measure machine performance. index is retained in this Standard, even though by definition it is
not an error. By definition, an error is the difference between the
probe: in this Standard, a device that establishes location measured value and the true value, and even the best estimate of
of the movable components of an articulated arm coor- the true value for the TEI test is never known. As no calibrated ar-
dinate measuring machine relative to a measurement tifact is used, it should be more correctly thought of as a thermal
point. Two types of probes are discussed: effect index. The index is a measure of the variation over a period
(a) passive (solid or hard) probe: probe that mechanically of time. If a calibrated artifact is used referenced back to the stan-
dard temperature, the error may well be considerably bigger than
fixes the movable component relative to the workpiece.
the variation.
(b) switching probe: probe that gives a binary signal as
a result of contact with or in proximity to a workpiece. thermal settling time: time interval that in the supplier’s
estimation is sufficient to allow the AACMM to meet
range: set of values bounded by the extreme indications. performance specifications after a temperature change
significant mean temperature range: range of mean ambi- of
5.0°C (
9.0°F). This parameter can be used as a
ent temperature over which the AACMM will still meet guide to estimate when measurements can commence
the performance specifications stated by the supplier for after the machine is moved between differing thermal
the tests defined in para. 5. environments.
single point articulation performance: ability of the working tolerance: maximum acceptable range in the mea-
AACMM to provide similar values of a point coordinate surements for any performance test in this Standard.

4
METHODS FOR PERFORMANCE
EVALUATION OF ARTICULATED ARM CMMs ASME B89.4.22-2004

workpiece: object to be measured. C

D
work zone: measurement volume of a machine as speci-
fied by the supplier. More than one work zone may be
specified for a given machine, and working tolerances E
may be specified separately for each work zone.
2.1.2 Stand Versus Fixture. Throughout this Stan- F
dard, the machine is mounted on a stand, while any ar-
tifact, such as ball bar, is mounted in a fixture. B

A
2.2 Machine Classifications
The following classifications of AACMMs are pro-
vided for ease of machine specification. A place is pro-
vided in the standard machine classification form, Fig.
1, that shall be used to designate the machine classifi- Fig. 5 Typical Machine of the 2-2-2 Configuration,
cation as described below. For AACMMs that do not With a-b-c-d-e-f deg Rotation
conform to these standard configurations, an equivalent
drawing as in Fig. 1 shall be provided.
An articulated arm coordinate measuring machine D
employs a series of rotating components around gener-
ally perpendicular axes. The arm can be thought of as
E
consisting of three joints, namely the shoulder, elbow,
and wrist. As an example, for the 2-1-2 configuration
F
AACMM shown in Fig. 4, the shoulder would consist
of the A and B rotary axes; the elbow, the D rotary axis; B G
and the wrist, the E and F rotary axes. A probe is at-
A
tached to the last rotating axis and is manually manip-
ulated to contact measurement points on a workpiece
mounted within the AACMM work zone. The configu-
ration of an articulated arm CMM is defined by three
numbers (e.g., 2-1-2) corresponding to the shoulder- Fig. 6 Typical Machine of the 2-1-3 Configuration,
elbow-wrist joints, respectively, that describe the rota- With a-b-d-e-f-g deg Rotation
tional degrees of freedom for that joint. Each rotating
component is further defined by the limits of rotation
about its axis (e.g., a, b, c, etc.), in degrees. Common con- C
figurations of the AACMM, along with their designa- D
tions, are given in Figs. 4, 5, 6, and 7.
E

F
D B
G
A
E

B Fig. 7 Typical Machine of the 2-2-3 Configuration,

With a-b-c-d-e-f-g deg Rotation
A

3 ENVIRONMENTAL SPECIFICATIONS
3.1 General
Fig. 4 Typical Machine of the 2-1-2 Configuration, It shall be the responsibility of the user to provide an
With a-b-d-e-f deg Rotation acceptable environment for conducting the performance

5
 METHODS FOR PERFORMANCE
ASME B89.4.22-2004 EVALUATION OF ARTICULATED ARM CMMs

evaluation testing. The environment shall be considered 3.2.2 Thermal Environment Parameters. The supplier
acceptable if the requirements of this paragraph and shall provide, as part of the machine specification, a
para. 4 are met. The user shall be responsible for con- statement of the acceptable thermal environment pa-
ducting all of the environmental tests at their facility. rameters. It should be noted that the thermal environ-
The supplier shall have the right to witness this testing. ment is not limited to the environment within the work
The supplier shall, upon request, supply test equipment zone, but includes the environment that is in contact
as specified in para. 6, as well as support for equipment with the machine, i.e., the impact of the external ther-
and tests, at a cost negotiated between the supplier and mal influences on the machine itself. The thermal pa-
the user. The user is cautioned that failure to conform rameters shall contain a specification on the range of
to the supplier’s recommendations on cleanliness and temperatures within which the specified performance
cleaning procedures can lead to significant performance will be attained, as well as the maximum permissible
degradation. For example, particulates, oils, and water vertical and horizontal temporal and spatial gradients.
can significantly degrade the machine’s performance, While not part of the thermal testing, the safe operating
increase friction, and accelerate wear. temperature range and the thermal settling time shall
also be stated.
3.2 Temperature and Humidity 3.2.2.1 Thermal Radiant Energy. Care should be
3.2.1 General. Taking measurements when not at the taken to minimize the machine exposure to direct sun-
standard temperature of 20°C can have a significant and light or other powerful radiant energy sources. Other
often misunderstood influence on the quality of those direct radiant energy sources (such as lighting) shall not
measurements. The provisions outlined in ISO/TR be, whenever possible, closer to any part of the machine
16015 form a part of this Standard, but interpretation is than twice the nominal arm length. Where this distance
provided for the application to the class of machine cov- requirement is impractical, indirect lighting designed
ered by this Standard. It is recognized that some ma- for diffuse reflection and increased path shall be used.
chines, because of their portability, will be used in
widely varying thermal environments. The portable na- 3.3 Vibration
ture of these devices ensures that they will be used for
measurements in most of the wide spectrum of manu- 3.3.1 General. The support surface (floor, founda-
facturing environments found today. ISO/TR 16015 de- tion, tooling stand, etc.) upon which the machine is
fines two alternative conditions under which a test en- mounted can have motion induced as a result of exter-
vironment is thermally acceptable. The first is that all nal forces in the surrounding area (due to other ma-
measurements are taken at the standard temperature, chines, transportation vehicles, compressors, motors,
which for all practical purposes can never be done. The etc.). This motion can range from a continuous vibra-
second condition is that the thermal error index (TEI; see tion to interrupted shock or a combination of both. Such
para. 6.5, ASME B89.6.2) is a reasonable percentage of motion, if transmitted to the machine, has a degrading
the working tolerance. This leads to the potential prob- effect on the overall performance and repeatability of
lem that errors, caused by differential thermal expan- the machine, causing relative motions between the
sion, hysteresis, etc., can be induced in machines when probe tip, the machine axes measuring system, and the
they are used at temperatures different from the tem- workpiece, resulting in increased measurement uncer-
perature at which they were aligned and calibrated. tainty. In addition, certain excessive amplitudes of mo-
Similarly, the measurement of a workpiece at a tem- tion can cause damage to the machine.
perature different from its design temperature can 3.3.2 Responsibilities. The user shall be responsible
cause the workpiece to change size or distort. How- for site selection, environmental shock and vibration
ever, it is not within the current scope of this Standard analysis, and for any additional special isolators deemed
to develop simple tests or procedures for quantifying necessary to ensure compliance with the maximum per-
these individual effects. In those cases where the ma- missible vibration levels specified by the supplier. All
chine is to be operated at temperatures outside the questions of compliance shall be determined at the in-
supplier’s specified temperature range, the user should terface between the support system provided by the
refer to the calculation of TEI in Appendix H, for a bet- user and the machine system provided by the supplier.
ter understanding of the potential effects this will have
on the machine’s performance. Acceptability of an en- 3.3.3 Environmental Vibration Parameters. The sup-
vironment that does not comply with the supplier’s plier shall provide, as part of the machine specification,
thermal parameters is therefore specified in terms of a statement of the acceptable seismic vibration spectra
its effect on the machine. at the user–supplier interface. This interface may be very
In general, nominal variations in humidity do not different, depending upon details of the contractual
have an adverse effect on the performance of the ma- arrangement between supplier and user. This statement
chine, but over a period of time may be detrimental. shall contain a complete description of the allowable to-

6
METHODS FOR PERFORMANCE
EVALUATION OF ARTICULATED ARM CMMs ASME B89.4.22-2004

tal vibration amplitude over a specified frequency Arm does not get
range. evaluated here

3.4 Electrical
3.4.1 General. The electrical power supplied to a ma-
chine can have a strong effect on its ability to perform
accurate and repeatable measurements. This is particu-
larly true when a machine uses some form of computer
for any control or readout function.
3.4.2 Responsibilities. It shall be the responsibility of
the user to provide electrical power meeting require-
ments specified by the supplier.
3.4.3 Electrical Parameters. The supplier shall pro-
vide, as part of the machine specification, a statement
of the steady state voltage requirements of the machine, Vertical axis
allowable deviations from this voltage, frequency re-
quirements, and amperage requirements. These param- Fig. 8 AACMM Mounted Vertically for Evaluation
eters are listed in Fig. 1.

supplier may prescribe limitations to mounting orien-

3.5 Mounting Stiffness and Orientation tation for their instrument. Additionally, mounting ori-
3.5.1 General. It is the nature of AACMMs that they entations other than those used during performance
are often portable and therefore can be mounted in a va- testing will result in different loading of the AACMM‘s
riety of locations and orientations. This paragraph is components (links and rotary axes) and may allow the
added to environmental specifications since the per- AACMM to reach locations in the measurement enve-
formance of the AACMM may be influenced by the lope that were not previously tested. This is shown in
mounting orientation, and the stability and stiffness of Figs. 8 and 9.
the installation. Therefore, it is strongly recommended that testing be
The AACMM is by design a manually manipulated performed in the orientation(s) in which the arm is to
and supported measuring instrument. This leads to op- be used, and that these orientations be within any pre-
erator-induced external forces and moments applied to scribed limits designated by the supplier.
the instrument that can cause deflections within the in-
strument as well as at the mounting interface and in the
3.5.2 Responsibilities. It is the responsibility of the
supplier to provide the maximum expected forces and
mount itself. Additionally, AACMMs are available in ei-
moments that the AACMM will generate, the maximum
ther counterbalanced or non-counterbalanced configu-
acceptable deformations at certification as depicted in
rations. The counterbalanced models employ restoring
the Mounting Requirements portion (see Fig. 2), and to
forces to one or more of the arm‘s links to make the in-
strument easier to manipulate and use over long peri-
ods of time. However, these restoring forces also repre-
sent another potential source for deflections.
Region not
As neither the magnitude nor direction of the deflec- previously
tions is constant, varying with operators and for a given evaluated
operator, varying with location and orientation of the
arm within its work volume, they cannot easily be
mapped and compensated. If these deflections are suf-
ficiently large (i.e., the deflection magnitude is near the
order of the AACMM single-point articulating per-
formance as defined in para. 5.3), any measurements
made with such a mounted device should be highly sus-
pect and the instability in the mounting should be cor-
rected before measurements are made.
As previously stated, the AACMM by design is in-
tended to be highly portable and adaptable to mount-
ing in a variety of locations and orientations. As the sta-
bility of the mounting can affect performance, the Fig. 9 AACMM Mounted Horizontally for Use

7
 METHODS FOR PERFORMANCE
ASME B89.4.22-2004 EVALUATION OF ARTICULATED ARM CMMs

state any limitations on the mounting orientation. It is tion and the thermal environment is suspect, the envi-
the responsibility of the user to determine the appro- ronmental tests that are part of Appendix B can be con-
priateness of the mounting and to mount the AACMM ducted in order to determine complete conformance to
within the limits specified by the vendor. the supplier’s thermal environment requirements. If any
of the parameters measured in Appendix B exceed the
3.5.3 Parameters. The supplier shall provide, as part
supplier’s specified parameters, then the performance
of the machine specification, a statement of mounting
tests are suspended until such time that the user is able
requirements, including the maximum applied forces
to correct the problem in order to conform with those
and torques at the mounting interface, the maximum
specified parameters. If the parameters so measured
distances and displacements acceptable at the mounting
meet the supplier’s specified parameters, it is the sup-
interface, and any limits to the mounting orientation.
plier’s responsibility to correct the performance of the
These parameters are to be listed (see Fig. 2).
measuring machine to meet the specified working tol-
erances.
4 ENVIRONMENTAL TESTS If the machine is frequently moved between varying
thermal environments, an alternate (or supplemental)
4.1 General
procedure is to perform the ETVE test, as defined in Ap-
It is the recommendation of this Committee that, if a pendix H, after moving the AACMM between thermal
machine is to be operated in an environment that is environments and following the supplier’s recommen-
significantly different from that used during the per- dations for thermal settling time between moves. While
formance testing, a subset of the performance tests be this test is somewhat of a compromise, in that it does
repeated in order to assess the performance of the not provide a full picture of how the AACMM will per-
AACMM in the new environment. Those tests that are form in the new environment, it does show the previ-
recommended for this mini performance evaluation are ously unaddressed effects of a rapidly changing ther-
the effective diameter test and a sampling of the Volu- mal environment on the performance of the AACMM.
metric Performance Test positions, which should in-
clude both radial and tangential artifact orientations. 4.3 Vibration Tests
These tests should only be performed after adhering to
any supplier’s stated procedures for ensuring the in- It is the position of this Committee that all tests be
tegrity of the AACMM measurements at temperatures performed in a vibration environment that conforms to
other than that at which they were calibrated. the supplier’s specification. The vibration environment
As stated previously, it is the philosophy of this Stan- specification parameters, total amplitude and frequency
dard that the operating environment is the responsibil- range, can each have a significant effect on AACMM
ity of the machine user. If the environment complies performance. However, constant monitoring of all of
with the parameters specified by the machine supplier, these parameters would represent an undue burden on
the responsibility for meeting performance specifica- either the user or the supplier. Should the machine fail
tions rests solely with the machine supplier. It is the po- to meet performance specification and the vibration en-
sition of this Committee that the performance tests in vironment is suspect, the environmental tests that are
this Standard must be performed under conditions that part of Appendix C can be conducted in order to de-
conform to the supplier’s environmental specifications. termine complete conformance to the supplier’s vibra-
As such, this Standard does not provide procedures tion environment requirements. If any of the parame-
for derating the performance due to nonconformance to ters measured in Appendix C exceed the supplier’s
these specifications. specified parameters, then the performance tests are
suspended until such time that the user is able to cor-
4.2 Temperature Tests rect the problem in order to conform with those speci-
fied parameters. If the parameters so measured meet the
It is the position of this Committee that all tests be supplier’s specified parameters, it is the supplier’s re-
performed in a thermal environment that conforms to sponsibility to correct the performance of the measur-
the supplier’s specification. The thermal environment ing machine to meet the specified working tolerances.
specification parameters, significant mean temperature
range and the thermal settling time, can each have a sig-
4.4 Electrical Tests
nificant effect on AACMM performance. However, con-
stant monitoring of all of these parameters would rep- Well-defined procedures and highly developed in-
resent an undue burden on either the user or the struments exist that enable the measurement of the pa-
supplier. Therefore, it is the requirement of this Stan- rameters characterizing the electrical power supplied to
dard that the mean ambient temperature shall be es- a machine. It is, however, the opinion of this Commit-
tablished both before and after each performance test to tee that such tests are, in the general case, an unwar-
ensure conformance to the supplier’s specification. ranted expense and shall be undertaken only in the
Should the machine fail to meet performance specifica- event that the machine does not meet performance spec-

8
METHODS FOR PERFORMANCE
EVALUATION OF ARTICULATED ARM CMMs ASME B89.4.22-2004

ifications and there is reason to suspect the electrical FZ

power. Failures due to electrical power usually show up Z
as intermittent control or readout failures that are diffi-
cult to link to mechanical causes. In the case that the MZ
power is suspect, this Standard provides Appendix D,
which describes the recommended procedure for deter-
mining the conformance of the electrical environment X FX forces (along X )
to the supplier’s guidelines. MX
moments (about X )

4.5 Mounting Stiffness Test

Y MY
It is the position of this Committee that the testing of FY
the mounting stiffness be carried out for the user-
supplied portion of the mount. For example, if the user
(a)
supplies a tooling stand, then the stiffness of the stand
is tested by the user. If, as part of a complete measure-
ment system, the supplier provides a stand which is
then mounted to a steel table, then the mounting stiff- D
ness test is performed on the steel table. This is not to
say that deflections in the supplier stand are negligible,
nor should they be ignored. However, it is the suppli-
ers’s responsibility to provide a mount, if provided as
part of the measurement system, that meets their own Force
stiffness requirements. In this situation, responsibility Displacement due to force
for any degradation in the AACMM performance due
to deflections in the supplier-provided mount is borne
by the supplier.
The primary forces applied to the mounting of
AACMMs are translational and torsional. These forces (b)
will tend to either move the entire measurement system
along an axis (or combination of axes) or cause the meas-
urement system to rotate about an axis (or combination L
of axes). These forces are illustrated in Fig. 10, illustra-
tion (a). The forces can be further described as acting D
along or about the three major axes of the coordinate
system at the mounting base. The translational forces, T D  rotation due to torque
F, along the axes and the moments, M, about the axes L
will tend to induce deflections of the base.
The deformation due to the translational force can be
measured as shown in Fig. 10, illustration (b). The user
must apply a force (N) using a calibrated load cell at the
(c)
mounting interface to the maximum level required and
measure the displacement due to the force, D (mm). Using
a calibrated torque wrench, a torque can generate the Fig. 10 Mounting Stiffness Test Setup
torsional forces or moments at the base. The torque is
applied about an axis and an indicator, mounted at a
As noted above, the deflections must not exceed the de-
distance, L, from the applied torque, measures the dis-
formation in the mounting requirements specified in
placement. The deflections can be described as a slope
Fig. 2.
(mm/mm) and can be measured as depicted in Fig. 10.
In either case, the deflections must not exceed the de-
formation in the mounting requirements specified in 5 MACHINE PERFORMANCE
Fig. 2.
5.1 General
Care must be exercised when applying the torque
with a torque wrench, as this can also induce a force. Prior to conducting the performance tests, the machine
An alternative method is to use two indicators, one po- shall be mounted in accordance with the supplier’s rec-
sitioned to the left of the center and the other to the right, ommended procedures. Additionally, the principal probe
and apply a static load with a dead weight and suitable type (i.e., hard or switching probe) must be specified and
mounting bar. In this case, the slope can be measured. shall remain the same for all subsequent tests. The selected

9
 METHODS FOR PERFORMANCE
ASME B89.4.22-2004 EVALUATION OF ARTICULATED ARM CMMs

probe type is then to be mounted in the AACMM and cal- Table 1 Diameter Deviations
ibrated in accordance with the supplier’s recommended
procedures. In the case where an AACMM is to be used Formula Diameter Deviation,
m
with more than one probe type, the user is encouraged to D1  Dcal
repeat the performance tests with all candidate probe D2  Dcal
types. However, testing with more than one probe type D3  Dcal
is not a requirement of this Standard. Because of the porta- Max. deviation
bility of these instruments, any recommendation for cali-
bration after shipping should be agreed to prior to testing
per this Standard. It is possible, depending on the treat-
ment during transport, that some AACMMs may suffer and the calibrated diameters of the sphere is reported as
performance degradation and require recalibration after the diameter deviation. The effective diameter perform-
any transportation. ance test value is reported as the maximum deviation re-
The supplier shall have the responsibility for provid- gardless of sign. This method is used to harmonize the re-
ing a machine that meets all performance specifications sult with the B89.4.1b bidirectional length test, which is
agreed upon between the supplier and user when in- analogous to the effective diameter performance test.
stalled according to the supplier’s recommendations in A typical format for recording data is shown in Table 1.
any environment meeting the requirements of para. 3. The results may reveal a number of problems, which
A machine meeting performance specifications and may include either excessive machine or probe hystere-
other conditions agreed upon between the supplier and sis, or improper probe calibration.
user shall be accepted by the user. The criterion for meet- In cases where there appears to be a single or several
ing performance specifications shall be the satisfactory outlying diameter(s) that do not conform to specifica-
completion of all tests specified in this paragraph, ex- tion, it is recommended that the entire effective diame-
cept that any test or tests may be omitted by mutual ter test be repeated in order to ascertain whether the
agreement between the supplier and user. It should be large deviation reflects a systematic error. If, after a max-
emphasized that the performance tests for effective di- imum of three repeats of the test, the results still do not
ameter, single point articulation performance, and vol- conform to the working tolerance for the effective diam-
umetric performance that are described in this para- eter test, the test shall be discontinued, and the fault de-
graph contain several options, and that these options termined and corrected, before proceeding with any ad-
will not necessarily give exactly the same results on any ditional performance testing. Upon correction of any
given machine. identified fault(s), the entire test shall begin anew and the
results of this retest are the only data to be considered.

5.2 Effective Diameter Performance Test 5.3 Single-Point Articulation Performance Test
5.2.1 General Techniques. The requirement in the 5.3.1 General Techniques. The single-point articula-
definition of the effective diameter performance test tion performance test is intended to assess the AACMM‘s
shall be satisfied by measuring the diameter of a cali- ability to provide similar values of a point coordinate
brated sphere using nine probing points. The diameter when the AACMM is articulated through the maximum
measurement is repeated three times and the largest test possible range of motion for that single point. It should
deviation from the calibrated value is reported. be noted that the single-point articulation performance
5.2.2 Setup and Measurement Procedures. Using a test result is different from repeatability in the strict defi-
rigidly mounted calibrated reference sphere with a di- nition of the term. This test, by design, incorporates as-
ameter between 10 mm and 50 mm, conforming to the pects of both repeatability and reproducibility that are
requirements of para. 6.2.2, measure the sphere three necessary because of the unique nature of the AACMM.
times at the same location, in approximately the middle Unlike a conventional linear axis CMM, a single point may
of the reach of the arm. Measure the sphere diameter be measured from a nearly infinite number of directions
using nine points per sphere measurement with the fol- and orientations of the AACMM’s axes. Additionally, the
lowing distribution: four points equally distributed ap- construction of the AACMM does not map one-to-one
proximately on the equator, four points equally distrib- from what is physically measured by the instrument and
uted at an approximate latitude of 45 deg and rotated the parameters that define the instrument’s coordinate
at approximately 45 deg to those on the equator, and one system, i.e., there are no physical X, Y, and Z axes as in a
point at the pole. During this test, the articulation of the Cartesian system or R, , and  axes as in a spherical sys-
AACMM should be minimized for each measurement. tem, etc. Rather, the AACMM uses a series of kinematic
transformations to express the probe tip position, in one
5.2.3 Data Analysis. The nine points are used to deter- of any natural coordinate system, using the predeter-
mine a sphere using the AACMM software and the di- mined arm segment lengths and current rotary axis posi-
ameter calculated. The difference between the measured tions. As such, what this test tries to convey is not the re-

10
METHODS FOR PERFORMANCE
EVALUATION OF ARTICULATED ARM CMMs ASME B89.4.22-2004

AACMM position

1st location: 0%–20% of

the length of the arm

2nd location: 20%–80% of

the length of the arm

3rd location: 80%–100% of

the length of the arm

Default locations of the

mounted seat

Fig. 11 Three Default Locations of the Mounted Seat During the Single-Point
Articulation Performance Test

peatability of any physical attribute of the measurement (a) AACMM arm elbow to the left and down
instrument, but rather the system‘s combined ability to re- (b) elbow to the left and up
produce the coordinates of a fixed point in space. (c) elbow up
To maximize the articulation of all of the arm en- (d) elbow to the right and up
coders, the single-point articulation performance test (e) elbow to the right and down
shall be performed using an artifact placed at three dif- The same pattern is repeated with the wrist rotated
ferent locations within the working volume. Figure 11 approximately 180 deg about the axis of the probe. These
illustrates the three default measurement locations positions are shown in Fig. 12.
within the working volume for the single-point articu- For each of the ten orientations, the probe location is
lation performance test. The first location shall be within recorded, after which the probe is removed from the
a radius up to 20% of the length of the arm, centered at seat and then replaced prior to generating the next axes
the first rotational axis of the AACMM. The second lo- combination and taking the next point. This process is
cation shall be within a zone defined by two radii be- then repeated for the three test locations specified in
tween 20% and 80% of the length of the arm. The third para. 5.3.1.
location shall be outside of a radius greater than 80% of 5.3.2.2 Single-Point Articulation Performance Test
the length of the arm. Experience indicates that if the With a Switching Probe. The following procedure is re-
probe is kept in contact with the artifact, the force ap- quired when using a switching probe, but may be op-
plied by the operator tends to increase. Therefore, after
each point is taken, the physical contact between the ar-
tifact and probe shall be broken.

5.3.2 Setup and Measurement Procedures

Z
5.3.2.1 Single-Point Articulation Performance Test
Using a Hard Probe. The single-point articulation per-
formance test is performed by measuring the center co-
ordinates of the AACMM ball probe in a rigidly
mounted kinematic seat. The seat may be a chamfered
hole, a conical socket, or a trihedral socket, conforming
to the requirements of para. 6.2.4. During the test, the
supplier’s recommended default probe is placed in the
kinematic seat and the AACMM arm is articulated to
X
ten different orientations, maximizing the possible ori-
entations of the AACMM axes. As a default condition, Y
and noting the orientation of the wrist, the first five of
the ten orientations shall be performed as follows: If the
operator is facing the machine, the five arm positions
may be defined as Fig. 12 Isometric View of Default Articulations

11
 METHODS FOR PERFORMANCE
ASME B89.4.22-2004 EVALUATION OF ARTICULATED ARM CMMs

tionally used with the hard probe. However, the proce- The entire process of ten sphere center measurements
dure described in para. 5.3.2.1 is the default method for is then repeated at the remaining sphere positions
use with a hard probe. The requirement in the defini- shown in Fig. 11.
tion of the single-point articulation performance test to
sense the same quantity shall be satisfied by measuring 5.3.3 Data Analysis. The statistics used to represent
the center coordinates of a calibrated sphere. The rigidly the single-point articulation performance are the maxi-
mounted calibrated reference sphere shall have a diam- mum deviation of the points from a mean value and
eter of between 10 mm to 50 mm and must conform to twice the standard deviation of the point location. These
the requirements of para. 6.2.2. statistics provide a measure of the combined variability
The test is conducted by measuring the location of the in three mutually perpendicular axes. In this case, the
sphere ten times at three different positions within the standard deviation is actually a combination of the vari-
work zone as defined in para. 5.3.1. For each sphere posi- ances of the individual coordinates (X, Y, Z).
tion, the arm is articulated through the same ten orienta- The procedure for determining the single-point artic-
tions as in para. 5.3.2.1. For each of the ten arm orienta- ulating performance at each of the three measurement
tions, five points are measured and fit to a sphere to obtain locations is as follows. First, calculate the average coor-
the calculated sphere center coordinates. The default pat- dinates (X, Y, Z) of either the ten single-point measure-
tern for the five points shall be four points approximately ments (hard probe procedure) or the ten calculated
on a great circle and one point on a pole. A typical five- sphere centers (switching probe procedure). Using Eq.
point measurement pattern is illustrated in Fig. 13. The (1), calculate the three-dimensional deviations (i) from
arm must be kept at approximately the same orientation the average coordinates to each of the corresponding ten
for all five points during a sphere center measurement. points/sphere centers used to compute that average
The process is repeated for each of the ten sphere cen-
i 

(Xi  
Xa)2 
(Yi  
Ya)2 
(Zi  
Za)2 (1)
ter measurements, taking care to articulate the AACMM
to a significantly different set of joint orientations and where
rotations between each sphere measurement. Hence, the Xi, Yi, Zi  measured coordinates
ten sphere center measurements should be performed Xa, Ya, Za  average coordinates
with ten significantly different joint orientations and ro-
From these values, the maximum deviation can be de-
tations (including rotating the probe end joint), ensur-
termined and the 2sSPAT value calculated using Eq. (2). A
ing that the rotational axes are exercised over a signifi-
typical format for recording the data and the calculated re-
cant portion of their range.
sults for each measurement location is shown in Table 2.


2sSPAT  2 
 2i
(n  1)
(2)

5
where
n  number of measurements (in this case, n  10)
The results of the SPAT test then consist of both the
max and 2sSPAT values for each measurement location.
If the result obtained for each of the three measurement
locations is less than the working tolerance for the
single-point articulation performance test, then the
90 deg
AACMM has passed the test and the results are reported
along with the nominal locations of the test. In the case
where it is desired to report only a single set of results
1–4 for this test, the max and 2sSPAT value for the mea-
surement location with the largest max shall be reported.
If, the result for any measurement location is greater than
the working tolerance, then the entire test shall be re-
peated. If, after a maximum of three repetitions of the
test, the results still do not conform to the working tol-
erance for the single-point articulation performance test,
the test shall be discontinued and the fault determined
and corrected before proceeding with any additional
performance testing. Upon correction of any identified
fault(s), the entire test shall begin anew and the results
of this retest are the only data to be considered for the
Fig. 13 Default Five-Point Probe Pattern single-point articulation performance analysis.

12
METHODS FOR PERFORMANCE
EVALUATION OF ARTICULATED ARM CMMs ASME B89.4.22-2004

Table 2 Single-Point Articulation Performance Test

Reading X-Axis Y-Axis Z-Axis i
No. Xi Yi Zi [Eq. (1)] 2i

10

Average Xa  Ya  Za  ... ...

max  ...

[Eq. (2)] 2sSPAT 

GENERAL NOTE: Xi, Yi, and Zi are either points from hard probe measurements or calculated sphere cen-
ters from switching probe measurements.

5.4 Volumetric Performance Test 5.4.2 Setup and Measurement Procedures. As the de-
fault is for a calibrated bar to be used, a suitable ther-
5.4.1 General Techniques. The approach to accessing
mal correction must be made prior to analyzing the data.
the AACMM‘s performance throughout its volume is
The ball length, corrected for thermal expansion, can be
necessarily different from that taken by the ASME
obtained by using a temperature sensor integrated with
B89.4.1b standard for traditional linear axis CMMs. Un-
a part-thermal-expansion-correction function in the
like a linear axis CMM, for which any point within the
AACMM software (if available) or by measuring the ball
work zone uniquely defines the position of the three ma-
bar temperature using a third-party temperature system
chine axes, there is an infinite number of arm orientations
and making the correction using the equation below.
that can result in the same location of the AACMM probe.
Unless otherwise agreed between the supplier and the
This characteristic of the AACMM means that an inde-
user, the default coefficient of thermal expansion
pendent test of the machine “scales” along three lines at
(COTE) used for correction shall be as shown in Table
a regular spacing, as done in the ASME B89.4.1b linear
H-1 and the thermometry used shall conform to the re-
displacement accuracy (LDA) test, will not reveal a great
quirements for such equipment given in para. 6. It is rec-
deal of information about the AACMM. Alternatively, re-
ommended that the temperature of the artifact be
quiring a test similar to the ASME B89.4.1b LDA but at
recorded at the start and finish of each specific test and
many more locations and orientations would be too time
an average of these two values used.
consuming. Therefore, aspects of the linear displacement
accuracy and volumetric performance tests are combined Lc  Lm [1  BB  (Tm  20°C)]
into the volumetric performance test. This test, as detailed
where
in the following paragraphs, requires an ASME B89.4.1b
Lc , Lm  corrected and measured ball bar lengths, re-
distance measurement with a calibrated artifact, similar
spectively
to the LDA, but requires length measurements at many
Tm  average temperature during the measure-
locations within the machine volume, as in the ASME
ments
B89.4.1b volumetric performance test.
BB  coefficient of thermal expansion of the ball
The general approach is to position the artifact in two
bar material
vertical orientations, ten horizontal orientations, and
eight 45 deg orientations. For each length measurement, The default length artifact is the ball bar, which is cali-
the operator must record five measurement points per brated for center-to-center length, but other artifacts such
gage feature on the length artifact. as a step gage can be used. The following details the pro-
cedure for ball bars: Two ball bars of different lengths are

13
 METHODS FOR PERFORMANCE
ASME B89.4.22-2004 EVALUATION OF ARTICULATED ARM CMMs

required. The short ball bar shall be between 50% and 75% The artifact shall be positioned in three inclinations,
of the radial length of the AACMM arm. The long ball bar namely vertical, horizontal, and 45 deg, as shown in Fig.
shall be between 120% and 150% of the radial length of 15. The distance the artifact is from the center of the
the AACMM arm. The length artifact is measured in 20 measurement volume is defined as near or far. Near is
default positions. Other positions are acceptable, provided within one-half of the arm length from the center. Far
the work zone is adequately covered. The ball bar shall is greater than one-half of the arm length from the cen-
be suitably fixtured in the positions indicated for meas- ter. The horizontal and 45 deg inclinations can be addi-
urement so that probing access to both balls is achievable. tionally defined by a direction of radial or tangential.
The fixture should be portable for easy movement, but at The radial direction is defined as outward from the cen-
the same time must be sufficiently rigid so that the ball ter. The tangential direction is defined as perpendicular
bar will not significantly deflect or vibrate while the lo- to the radial direction. In summary, there are five char-
cations of the balls are being measured. For additional in- acteristics to consider when positioning an artifact:
formation on the use of ball bar, refer to Appendix F. length of the artifact, octant, artifact inclination, artifact
For the purpose of this test, the working volume of the distance, and direction relative to the center. Table 3 de-
AACMM will be divided into eight approximately equal fines the recommended 20 ball bar positions.
octants. The working volume of the AACMM is, by na- For each length measurement, the operator must
ture, a sphere whose radius is defined by the full length record five measurement points per sphere. The points
of the arm and is centered at the first encoder joint. This should be well dispersed on the surface of the sphere.
sphere is divided into upper and lower hemispheres by The sphere centers are then used to calculate a center-to-
an equatorial plane parallel to the surface on which the center distance for the ball bar. In order to attain an ac-
machine is mounted. The equatorial plane is divided into curate center-to-center distance, it is necessary not to
four quadrants, which creates eight equal volumes: four cause deflections of the ball bar during the measurement
quadrants in the upper hemisphere and four quadrants process. By monitoring the measured diameter of the
in the lower hemisphere. The supplier specifies what is to spheres, it is possible to detect whether there was signif-
be considered the front of the AACMM. Figure 14 illus- icant deflection during measurement. Typically, any de-
trates the numbering scheme of the eight octants, for a flection will cause the measured diameter of the sphere
horizontal equatorial plane. The octants are numbered 1 to appear smaller in size than the calibrated diameter.
to 4 in the upper hemisphere, beginning with the back
right octant and working around in a counterclockwise 5.4.3 Data Analysis. The center-to-center length
direction. The octants are numbered 5 to 8 in the lower value, Li, is calculated and recorded for each ball bar
hemisphere, beginning with the back right octant and position. The data from all the length measurements
working around in a counterclockwise direction. are analyzed by preparing a simple plot or a simple

Front Front Front Front

(a) Octant #1 (b) Octant #2 (c) Octant #3 (d) Octant #4

Front Front Front Front

(e) Octant #5 (f) Octant #6 (g) Octant #7 (h) Octant #8

Fig. 14 Octant Numbering Scheme

14
METHODS FOR PERFORMANCE
EVALUATION OF ARTICULATED ARM CMMs ASME B89.4.22-2004

2/6 11 1/5 2/3 1/4

15
19 20
6
3
19 8 3 10
Y 9
7 Z 6
12 13 17 14 11 13
8 7 2 10
18
Z
X X
4
17 18 16 12 15 16
20
5 1
2
9
1
5
14
4
3/7 4/8 6/7 5/8

(a) Top View (c) Front View

3/4 1/2
9
20 19 5
12
20 19
Z 10 3
18
17
9
5 7 Z
4 12 15 7
16 14
18 4 10
3 11
Y
8
14 17
13
11 X 6 Y 15
6 1
1 2 16
8
13
2

7/8 5/6

(b) Side View (d) Isometric View

Fig. 15 Recommended Ball Bar Locations

table, as in Table 4 and Fig. 16, comprised of the length where

deviations, Di, from the calibrated ball bar value, Lcal. n  20 in this case
Di  Li  Lcal (3) In cases where there appears to be a single (or sev-
The results of this test will be reported as three quan- eral) outlying point(s) that does not conform to the gen-
tities: the maximum deviation, the range of the devia- eral trend, it is recommended that the measurements at
tions, and two times the root-mean-square of the devi- that specific position be repeated. The recommended
ations. The range and maximum deviation of the data procedure for checking the length measurement is as
are clearly indicated in Fig. 16; the range is the differ- follows: The ball bar shall be measured twice in the sus-
ence between the maximum length deviation and the pected position. If the difference between the two length
minimum length deviation. Two times the root-mean- measurements is within twice the single-point articula-
square deviation is calculated from tion performance, the first measurement shall be used
and the second discarded. If the difference between the
two length measurements is not within twice the single-
2RMS  2  D2
i
n
(4)
point articulation performance, both are discarded and
the procedure is repeated. This procedure may be re-

15
 METHODS FOR PERFORMANCE
ASME B89.4.22-2004 EVALUATION OF ARTICULATED ARM CMMs

Table 3 Recommended Artifact Positions

Position Artifact
No. Length Octants Inclination Distance Direction

1 Short 8 Horizontal Near Radial

2 Short 5 Horizontal Near Radial
3 Short 1&2 Horizontal Far Tangential
4 Short 4&7 45 deg Far Tangential
5 Short 7&3 Vertical Far Tangential
6 Short 1&5 Vertical Far Tangential
7 Short 2 Horizontal Near Radial
8 Long 2&8 45 deg Near Tangential
9 Short 3 Horizontal Near Radial
10 Long 1&7 45 deg Near Tangential
11 Short 1&6 45 deg Far Tangential
12 Short 6&3 45 deg Far Tangential
13 Short 5&4 45 deg Far Tangential
14 Short 3&8 45 deg Far Tangential
15 Short 5&2 45 deg Far Tangential
16 Short 1&8 45 deg Far Tangential
17 Short 2&7 45 deg Far Tangential
18 Long 3&4 Horizontal Near Tangential
19 Long 2&6 Vertical Far Tangential
20 Long 4&8 Vertical Near Tangential

peated three times. At the end of that time, if agreement corrected. After correction of the problem, the effective
based on the single-point articulation performance diameter test, the single-point articulation performance
value has not been obtained as defined above, then the test, and the volumetric performance test shall be rerun
test shall be discontinued and the fault determined and in their entirety.

Table 4 Volumetric Performance Test Results

Position Di
No. Li Lcal [Eq. (3)] D2i
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Maximum deviation  ...
Range (Dmax  Dmin)  ...
2RMS [Eq. (4)] 

16
METHODS FOR PERFORMANCE
EVALUATION OF ARTICULATED ARM CMMs ASME B89.4.22-2004

125

100
75

50
25

Error, m
0
25

50
Vol. perf. Vol. perf. Vol. perf.
75 range max. dev. 2RMS
100
125
1 3 5 7 9 11 13 15 17 19
Position

Fig. 16 Volumetric Performance Test Results

6 TEST EQUIPMENT 6.2.3 Ball Bar. The ends of the ball bar shall be spher-
ical to within one-fifth of the working tolerance for the
6.1 Temperature single-point articulation performance of the AACMM.
The time constant of temperature recorders shall be The ball bar shall be calibrated for center-to-center length
no more than one-tenth of the cycle time of the highest to within one-fifth of the working tolerance for the vol-
frequency component of the temperature variation of in- umetric performance test. For information on ball bar de-
terest in a test. The time constant is the time required sign recommendations, see Appendix F.
for the temperature recorder to indicate 63.2% of its fi-
nal change due to a step change in temperature. 6.2.4 Seats for Hard Probes. A single-point articula-
The resolution of temperature recorders needs to be tion performance test using a hard spherical probe may
no greater than one-tenth of the amplitude of the low- be performed on either of the two types of positioning
est amplitude component of temperature variation of in- seats detailed below. The primary requirement is that a
terest in a test. sphere placed in the seat assumes a position repeatable
Temperature recorders shall be calibrated by suitable to one-fifth of the working tolerance of the AACMM
means to an uncertainty of
0.1°C over the temperature during the repeated seating required for the test.
range of use. 6.2.4.1 Trihedral Seat. The trihedral seat is de-
signed on the premise that two non-deformable objects
6.2 Displacement will contact each other at a maximum of three points.
The trihedral design requires that the same three points
6.2.1 Indicating Gages. Indicating gages shall have a be contacted upon each seating of the ball. A seat cre-
resolution of less than one-fifth of the working tolerance
ated by three hardened spheres approaches this ideal
for the single-point articulation performance test. All
condition. See Appendix G for a more detailed discus-
gages shall be calibrated in accordance with the sup-
sion on seat design. Alternatively, a chamfered hole with
plier’s recommendations.
three circumferential sections removed from the cham-
6.2.2 Precision Reference Ball(s). The precision ref- fered edge may also approximate a true trihedral seat.
erence spheres for the effective diameter performance 6.2.4.2 Chamfered Hole Seat. A chamfered hole
test and the single-point articulation performance test may also be used as a single-point test artifact. Since, in
shall be spherical to within one-fifth of the working tol- theory, only three points of the chamfered edge will con-
erance for the single-point articulation performance test tact the ball probe, care must be taken that the three
of the AACMM. The surface finish of the reference points do not change during the course of the test. The
sphere shall be such as not to exceed one-fifth of the likelihood of such a problem occurring is minimized by
working tolerance for the single-point articulation per- giving careful consideration to the design of the artifact,
formance test of the machine. The combined sphericity the selection of the material, and the material proper-
and surface finish of the sphere shall not exceed one- ties. As this type of artifact may be custom designed, a
third of the working tolerance for the single-point ar- number of design considerations are discussed more
ticulation performance test of the machine. fully in Appendix G.

17
Intentionally left blank
 ASME B89.4.22-2004

NONMANDATORY APPENDIX A
USER’S GUIDE TO ASME B89.4.22

A-1 PURPOSE performance tests. However, many applications may

not require the full environmental testing to assure con-
This user’s guide is intended to provide a framework formance to the performance specification. Therefore, as
for applying this Standard. The guide is written in a detailed below, parts of the environmental tests may be
checklist form to make it easier for first-time users to be- deferred by mutual agreement between the supplier and
gin using this Standard. Also, cross-references are pro- the user. If the machine passes the subsequent per-
vided from the guide text to the main body of this Stan- formance tests, performance of the deferred parts of the
dard for further details on each test procedure. environmental tests is not required.

A-2 SUMMARY OF USAGE

A-3 ACCEPTANCE TESTING CHECKLIST
The use of this Standard may be divided into two dis-
Before proceeding to the following list, it is assumed
tinct parts. First, this Standard is used to provide a clear,
that the supplier and user are in general agreement that
common method for specifying articulated arm coordi-
the machine is properly installed and the utilities are
nate measuring machines during negotiations between
working satisfactorily.
users and suppliers. Second, this Standard provides uni-
___ Determine if temperature environment meets
form test procedures to be used during machine ac-
supplier‘s parameters, or defer subject to later testing or
ceptance to establish conformance to the specification.
later elimination (para. 4.2).
___ Measure relative vibration as appropriate for ma-
A-2.1 Machine Specification chine configuration, or accept vibration subject to later
Use para. 1.1 of this Standard to establish a clear un- discovery of vibration-induced performance problems
derstanding between supplier and user of the charac- (para. 4.3). (If the vibration effects are unacceptable due
teristics of the machine. Detailed specifications of the to excessive environmental sources, discontinue testing
machine can be itemized using the three-part specifica- until a testing location conforming to supplier’s specifi-
tion form provided as Fig. 1. The first-time user is cations can be found.)
strongly urged to refer to the technical glossary pro- ___ Accept electrical utility, subject to later discovery
vided in para. 2.1 for clarification of the terminology of electrical utility-induced performance problems
used on the specification form. (para. 4.4).
In reaching agreement on the general machine speci- ___ Validate mounting stiffness (para. 4.5).
fication, the principal mode of operation and the prin- Prior to any testing in accordance with para. 5, verify
cipal probe type must be selected. that all test equipment used conforms to the require-
ments of para. 6.
____ Perform effective diameter test (para. 5.2).
A-2.2 Machine Acceptance
____ Perform single-point articulation performance
For initial acceptance, the machine must pass all per- test, as appropriate for machine and probe configura-
formance tests in an acceptable environment. Accept- tion (para. 5.3).
ability of the environment may be demonstrated in one ____ Perform volumetric performance tests by the
of two ways: by passing the performance tests or by method previously chosen (para. 5.4).
showing compliance with the supplier’s environmental It is strongly recommended that a sketch or descrip-
parameters. The recommended procedure is to perform tion of numbered length bar positions be attached to the
all of the environmental tests before proceeding with the test results.

19
ASME B89.4.22-2004

NONMANDATORY APPENDIX B
THERMAL ENVIRONMENT TESTING

B-1 PURPOSE spanning the longest test (the use of five readings, rather
than two, for the measurement of the mean ambient tem-
The performance of articulated arm coordinate measur-
perature is justified here for diagnostic purposes).
ing machines is strongly affected by the detailed charac-
teristics of the thermal environment that surrounds them.
B-2.2 Frequency and Amplitude of Temperature Variation
Parameters of importance include cooling medium (usu-
ally, but not always, air), velocity of cooling medium, fre- The range of frequencies of temperature variation and
quency and amplitude of temperature variations of the the amplitude of those variations shall be determined by
cooling medium, mean temperature of that medium, and measuring and continuously recording the temperature at
temperature gradients within that medium. The effects of the center of the work zone over a period of time that
these parameters and others are discussed in detail in should, at a minimum, be representative of a daily cycle
ASME B89.6.2. Additionally, a summary of this concept is (i.e., 24 hr). The maximum peak-to-valley temperature
given in Appendix H of this Standard. It is the thesis of variation shall be determined from the recorded data. The
ASME B89.6.2 and of this Standard that currently it is not data shall be analyzed to determine the range of temper-
possible to specify parameters of a thermal environment ature variation for a daily cycle and an hourly cycle, sub-
that will ensure a specific value for the thermal error in- ject to the condition that isolated disturbances that are
dex (TEI). For a thorough discussion of the technical situ- shorter in duration than the minimum period (maximum
ation, the reader is referred to ASME B89.6.2. The purpose frequency) specified by the supplier shall be ignored. The
of this Appendix is, however, to specify procedures and daily variation shall be defined as the maximum range of
responsibilities for testing the thermal environment in the temperature readings in 24 hr, subject to the condition on
event the machine fails to meet performance specifications transients mentioned above. The amplitude of the super-
and the machine user contends that his/her environment imposed hourly cycle shall be defined as the maximum
meets the supplier’s parameters. For the purpose of this range of temperature variation in any 1 hr interval, sub-
Standard, these parameters include nominal mean ambi- ject to the same condition.
ent temperature, frequency and amplitude range of tem-
perature variation, and maximum spatial temperature gra- B-2.3 Thermal Gradients
dient. The following tests are designed to measure these Thermal gradients shall be determined by measuring
parameters for the purposes of assuring conformance to the temperature at the extreme corners of the machine in
the supplier’s parameters. a horizontal plane and also at the highest and lowest lo-
cations of the machine. These temperatures shall be de-
B-2 METHOD OF TEST fined as the average value of no less than five readings
over an interval of 10 min. The maximum thermal gradi-
In order to ensure that the environment itself is tested
ent shall be determined to be the difference between the
rather than any characteristic of the coordinate measuring
maximum and the minimum temperatures anywhere
machine supplied, these tests are to be conducted with the
within the machine work zone divided by the distance be-
AACMM, support computers (if supplied), and any other
tween the measurement point of these extreme tempera-
auxiliary equipment related to the AACMM turned off for
tures. These readings shall be taken over a period of at
a period of 24 hr preceding the test, to allow adequate soak-
least as long as the longest acceptance test (or 24 hr) and
out of AACMM-induced thermal gradients. Normal activ-
the greatest value of the gradient reported.
ity, however, should be continued about the machine, as
this constitutes part of the user-supplied environment. With
these constraints, the following tests should be performed. B-3 ANALYSIS
If any of the parameters measured in para. B-2 exceed
B-2.1 Mean Ambient Temperature
the supplier’s specified parameters, it is the responsibility
The mean ambient temperature shall be measured us- of the user to correct the problem in order to conform with
ing a thermometer with characteristics as specified in para. those specified parameters. If the parameters so measured
6 of this Standard. The mean ambient temperature shall meet the supplier’s specified parameters, it is the supplier’s
be the time average temperature of five readings taken at responsibility to correct the performance of the measuring
the center of the machine work zone over a period of time machine to meet the specified working tolerances.

20
 ASME B89.4.22-2004

NONMANDATORY APPENDIX C
SEISMIC VIBRATION VERIFICATION TESTS

C-1 SCOPE ful in relation to machine performance. However, ve-

locity and acceleration are more appropriate parameters
The purpose of this Appendix is to recommend vi- for measuring machine site vibration. Displacement
bration measurement instrumentation and procedures may be suitable for specific situations, but it is not rec-
for measuring vibration at machine installation sites. Vi- ommended for general vibration measurements.
bration levels shall be measured at the proposed ma-
chine site(s) to compare to allowable site vibration lim- C-3.1.2 Ordinate Units. The use of time or frequency
its established by the machine supplier. This document for the ordinate will depend on the acceptance criteria
also defines the instrumentation and measurement pro- format of the machine supplier. Time-based criteria are
cedures to establish vibration on the machine for addi- referred to as a time history that provides measurement
tional analysis. This Appendix does not address the of transient or very low-frequency vibratory events,
determination of vibration sources or reduction of vi- such as beat signals. The frequency domain allows
bration levels. This task is usually involved and requires measurement over a very short time range that pro-
the knowledge of vibration specialists. vides ability to diagnose many dynamic events.

C-3.2 Criteria Format

C-2 DEFINITIONS
As defined in para. 5.3, the supplier shall provide, as
To the extent possible, this document is intended to part of the machine specification, a statement of ac-
be self-defining. It is written for individuals with an en- ceptable vibration. The criteria should be provided by
gineering background. Definitions for specific vibration the supplier or listed as part of the machine specifica-
terminology may be found in the Institute of Environ- tion form (Fig. 2), if used. At least two criteria format
mental Sciences and Technology document IEST-RP- options are presented: (frequency) response function
CC024.1, Measuring and Reporting Vibration in Micro- and time history. The supplied acceptance criteria will
electronics Facilities. define the format to present the vibration data for ease of
comparison.
C-3 VIBRATION ACCEPTANCE CRITERIA C-3.2.1 Frequency Response Function Criteria. These
types of criteria are specified as a vibration amplitude as
The machine supplier is to provide site vibration cri-
a function at specific frequencies. The criteria are usually
teria of acceptability. Below these levels, the machine
presented as allowable vibration amplitude versus fre-
can operate successfully, and above these levels, prob-
quency, in hertz. The frequency range may vary from sup-
lems may occur. Each machine supplier has different
plier to supplier. In general, seismic vibrations are appli-
formats and levels of acceptance. The type of vibration
cable over a range of 0 (DC) to 100 Hz. Vibration levels
measurements to be taken will depend on format and
have large dynamic range, and it is sometimes helpful to
vibration units specified by the machine supplier. Based
present amplitude data in logarithmic scale. If decibels are
on the type of criteria, the vibration specialist should de-
used, the standard reference values must be used.
termine the necessary measurement units, frequency
range, measurement locations, and instrumentation. C-3.2.2 Time History. These measurements repre-
sent the vibration during the time period of interest. The
C-3.1 Criteria Units supplier should specify a maximum peak-to-peak ac-
ceptable vibration level and a time period over which it
Vibration is characterized by amplitude versus time or applies. The vibration amplitude could be in units of ve-
frequency. The amplitude can be defined in displacement, locity or acceleration.
velocity, acceleration, or power spectral density. Depend-
ing on the type of criteria, the amplitude ordinate can be
defined in either the time domain or frequency domain. C-4 INSTRUMENTATION
C-3.1.1 Amplitude Units. Since the machine is a This paragraph describes various instruments re-
measurement tool, units of displacement are most use- quired to perform on-site vibration measurements. Var-

21
ASME B89.4.22-2004 NONMANDATORY APPENDIX C

ious types of sensors, signal conditioners, recorders, fied below, will require the use of an FFT analyzer af-
computer programs, and signal analyzers are available ter the data are acquired. It is the user’s responsibility
that will acquire these data. It is not intended to single to understand the instrument, its capabilities, and its
out any particular equipment manufacturer, but to rec- limitations. The following list offers guidelines for FFT
ommend types of equipment that meet the requirements analysis configuration and specifications:
of this Standard. (a) Noise Floor. 100 dB/
Hz .
(b) A/D Resolution. The resolution of the analog-to-
digital converter should be at least 12 bits. The better
C-4.1 Transducers
analyzers will have a 16 bit A/D resolution.
Many types of transducers exist for various types of (c) Dynamic Range. The dynamic range should be at
vibration measurements. The measurements specified least 70 dB. Better spectrum analyzers will have a higher
in this document require a seismic accelerometer or a dynamic range.
specific type of velocity transducer. (d) Frequency Resolution. This parameter, as it applies to
the analyzer, is denoted in number of lines over which the
C-4.1.1 Seismic Accelerometers. The two most im-
analysis range is divided. Most analyzers can have selec-
portant requirements for the accelerometer are fre-
table resolution from 100 lines to 1600 lines. The resolu-
quency response and sensitivity. Site vibration meas-
tion is calculated by dividing the frequency range by the
urements generally require low frequency and high
number of lines. For example, a 0–100 Hz frequency range
sensitivity. The minimum frequency response linearity
acquired with a 400 line analysis will have 0.25 Hz
should be less than 1 Hz, preferably 0.5 Hz. The fre-
(100/400) resolution. The frequency resolution used must
quency response should be greater than 100 Hz. The
be compatible with the resolution of the frequency re-
sensitivity of the accelerometer should be 10 V/g or
sponse criteria. If the criteria are defined at every 1 Hz,
greater, where g is equal to 9.8 m/s2.
the data must be acquired with a 1 Hz resolution. For ex-
C-4.1.2 Velocity Transducers. These sensors are also ample, 0–100 Hz criteria defined every 1/2 Hz would need
referred to as geophones. The sensitivity of the geo- to be acquired with 200 lines of resolution. This document
phone should be 4 V/cm/s. The frequency response lin- recommends that 0–100 Hz data be acquired with 400 lines
earity requirement of the velocity transducer is the same of resolution, producing 0.25 Hz resolution data. The over-
as the accelerometer, 0.5 Hz to 100 Hz. all frequency resolution will also be dependent on the
transducer frequency response. The procedure above
should be followed and modified only when the machine
C-4.2 Amplifiers and Signal Conditioners
supplier’s specification requests otherwise.
The transducers require amplifiers and signal condi- (e) Anti-Aliasing Filter. This filter prevents incorrect
tioners. Most seismic accelerometers require an ampli- reporting of frequency components due to under-sam-
fier, but some models may have built-in electronics that pling of higher frequency signals. This filter is found on
do not require signal conditioning. Velocity transducers most (if not all) FFT analyzers. It should always be used.
may require amplification and signal conditioning, de- ( f ) Averaging. Most analyzers have this feature. It is
pending upon the sensitivity and signal-to-noise ratio. used to reduce the effects of transient events such as per-
It is the responsibility of the vibration specialist to use sonnel or vehicular activity. It is recommended that ten
the proper signal conditioners. averages be taken for all measurements. Some spectrum
analyzers have various types of averaging functions,
C-4.3 Signal Recording/Analysis Instruments such as linear, rms, peak hold, or exponential. Linear or
summation averaging should be used.
The type of instrumentation to use will depend on the (g) Window Functions. This feature is used to force a
type of criteria and format that have been provided by the generalized vibration signal into discrete time domain
machine supplier. The frequency response criteria require periods. When window functions are not used, the fre-
a Fast Fourier Transform (FFT) dynamic signal analyzer quency response of the vibratory signal is incorrectly
or digital recorder. Time history data can be acquired with distributed throughout the frequency range. There are
an oscilloscope, a digital recorder, or an FFT analyzer. many types of window functions. The most popular are
Hanning, flat top, and uniform. Other windows provide
C-4.3.1 FFT Signal Analyzers. This type of analyzer
excellent amplitude accuracy and poor frequency accu-
is the most sophisticated means of measuring vibration,
racy, and vice versa. The Hanning should be used for
by providing the greatest amount of information about
all measurements specified in this Appendix.
the vibration signal. In most cases this additional infor-
mation is necessary to understand the vibration envi- C-4.3.2 Data Recorders. For ease of gathering vi-
ronment. Many types of FFT analyzers exist, from many bration data in the field, the use of a multichannel data
different manufacturers. One- and two-channel units, recorder is found to be useful and convenient. Such an
hand-held, and PC-based are formats readily available. instrument allows for three or more channels of data to
It should be noted that using a data recorder, as speci- be recorded simultaneously, while providing a perma-

22
NONMANDATORY APPENDIX C ASME B89.4.22-2004

nent record for archives and verbal data annotation dur- will be operating when the machine is expected to be
ing specific events. Additionally, the recorder allows a used should be running during the vibration testing. A
record of the real-time response, which can be most use- written test log or voice channel on a data recorder
ful. The data can then be processed at a later date using should be maintained by the individual performing the
in-house data reduction techniques such as FFT ana- test so that any abnormal events during the test may be
lyzers specified in the paragraph above. The recorder recorded. A test should be repeated if abnormal events
format must be digital and use digital audio tape (DAT) occur. Normal vehicular traffic should not be excluded.
because of its excellent signal-to-noise ratio and dy- When the environmental conditions are satisfactory, the
namic range, as compared to analog tape. data should be recorded on tape, saved to memory, and
C-4.3.3 Oscilloscopes. This piece of general labora- printed or manually recorded.
tory equipment may be easily obtained to make an ini- C-5.4.1 Time History. For time history criteria, sim-
tial set of time history readings. Most facilities have an ply compare the measured peak-to-peak vibration lev-
oscilloscope and personnel who can operate the equip- els to the permissible level. The machine supplier may
ment, which allows a user to take baseline readings. The provide vertical and horizontal criteria. It is important
oscilloscope is also useful for viewing beat signals, tran- to compare the acquired data to the criteria in the ap-
sient events, and hourly and daily vibratory changes. propriate direction.
The oscilloscope should be set to AC-coupled and free-
C-5.4.2 Frequency Response Function. Comparison
run triggering. Viewing the signal, determining the
of frequency response function criteria to frequency do-
peak-to-peak voltage amplitude, and using the trans-
main vibration data can require more effort than taking
ducer sensitivity for converting to appropriate ampli-
the data. If the criteria have the same level at all fre-
tude units, determine the vibration amplitude.
quencies (straight line) or little changes in amplitude, it
will be easy enough to draw the criteria over the printed
C-5 TEST PROCEDURES vibration levels. If the criteria are not constant or uni-
form, it may be easier to compare data and criteria with
The procedures for making vibration measurements
various software programs. This involves digitizing the
are fairly simple, once the appropriate analysis equip-
criteria, which in some cases requires entering levels at
ment is selected and configured as required.
1 Hz increments. The vibration data stored on the FFT
analyzer must be downloaded into a PC. This requires
C-5.1 Calibration different steps, depending on the analyzer manufac-
At a minimum, a qualified lab, traceable to NIST, turer. Using a spreadsheet, math, graphing, or special
should have calibrated the vibration measurement program, the vibration data and criteria are combined
equipment in the past 12 months. Site calibration of the into a single graph. Once the data are in a software for-
transducers at the start of the testing is required. mat, they can be manipulated, graphed, and analyzed
in a usable format.
C-5.2 Transducer Mounting
For all measurements, the transducers should be C-6 CRITERIA ASSESSMENT
mounted directly and firmly to the floor or a common C-6.1 Measured Vibration Below Criteria
interface for measuring three mutually orthogonal axes.
Such mounting arrangements are referred to as triaxial. If the vibration levels measured by the procedure
Some accelerometers incorporate three triaxial trans- above are within the supplier’s criteria, no additional
ducers in one device. When this mounting arrangement work is required. It is the sole responsibility of the sup-
is used, all three channels should be acquired simulta- plier to maintain the performance of the machine in or-
neously. Time-independent triaxial measurements der to meet specifications.
should not be performed.
C-6.2 Measured Vibration Above Criteria
C-5.3 Measurement Location If the vibration levels exceed the supplier’s specifica-
In general, the transducers should be mounted in the tions, it is the responsibility of the user to isolate the vi-
general area where the machine will rest. This area bration in order to conform to the specification, or else
should encompass the outer envelope of the machine accept a performance derating. Again, this Appendix
plus 3 m (approximately 10 ft) beyond this footprint. does not provide information on how to reduce exces-
sive vibration levels, but vibration isolation will reduce
the levels. Before the vibration levels can be reduced,
C-5.4 Acquiring/Recording Data
the source of the vibration must be determined. It may
Vibration measurements should be made during nor- be easy to do this with the above equipment. Shock and
mal operations of the facility. Nearby equipment that vibration isolator suppliers who specialize in low fre-

23
ASME B89.4.22-2004 NONMANDATORY APPENDIX C

X in

Y in Digital
recorder FFT analysis Printer

Z in

X in

Y in FFT
analysis Printer

Z in

Z- axis
Y amp Y- axis
conditioned output

Y- axis
To option
X amp X- axis A or B
conditioned output

X- axis

Z amp Z- axis
conditioned output

Fig. C-1 Diagram of Sensor Arrangement and Instrumentation Configuration

quency vibration attenuation should be contacted if vi- (c) calibration information

bration isolation or a vibration survey is required. (d) description and diagram of test setup
(e) procedure
(f) analysis
C-7 REPORT (g) summary
A report shall be issued by the vibration specialist It is important to note that the report should serve to
and shall include all backup information and ana- archive the baseline vibration data for later review, if
lyzed data with a comparison to the machine specifi- problems arise after machine installation.
cation. The report shall include the following as a
minimum:
C-8 FIELD INSTRUMENTATION DIAGRAM
(a) title
(b) dates (issued and when data were taken) A diagram of the instrumentation is given in Fig. C-1.

24
 ASME B89.4.22-2004

NONMANDATORY APPENDIX D
ELECTRICAL POWER MONITORING TESTS

D-1 PURPOSE plete shift. In the three-shift plant, complete 24 hr mon-

itoring is required. Additionally, care should be taken
The purpose of this Appendix is to specify test pro- that the power line monitoring occurs over a represen-
cedures for analyzing the electrical power supplied to tative period that includes all normal or even intermit-
the machine and its support equipment in the event that tent electrical activity within the plant that could affect
the electrical power is suspected to be causing inade- the machine. (As an obvious example, consider the case
quate machine performance. when arc welding is done only a few days a week at a
location that uses the same feeder as the machine. In this
D-2 TEST EQUIPMENT case, the power line monitoring should include a typi-
cal arc welding sequence.)
The parameters describing the electrical power sup- For making these measurements, an approved, cali-
plied to a machine can be measured by a variety of in- brated power line monitor of the type discussed previ-
struments (voltmeters, oscilloscopes, and the like). ously should be used. Appropriate thresholds (sag,
However, it is the recommendation of this Standard that swell, and transients) should be set at the values corre-
a power line disturbance analyzer be used for these tests sponding to those levels set by the supplier in the ma-
because of the excessive labor required when other in- chine specification. Monitoring should continue for a
struments are used. These units are designed to moni- sufficient period to ensure that all of the effects men-
tor a wide range of power line disturbances and are tioned are included.
capable of continuous, unattended operation. Typical
measured parameters include sags, swells, transients,
and line frequency. IEEE 1159-1995, Recommended D-4 ANALYSIS
Practice in Monitoring Electric Power Quality, is an ac- Typical power line monitors provide printouts of both
cepted standard for defining types of electric power the levels and times at which deviations from the
phenomena. accepted thresholds occur. If the monitor is set with
the thresholds described above, any such deviations
recorded shall constitute nonconformance with the sup-
D-3 METHOD
plier’s specifications, and it shall be the responsibility of
In order to ensure proper monitoring, the power sup- the user to correct such power line defects. If no devia-
ply to the machine should be monitored for a period tions from specifications occur, then it is the responsibil-
that includes the normal cycle of machine tool opera- ity of the machine supplier to correct the performance of
tion. In the one-shift plant, this should include a com- the machine in order that machine specifications are met.

25
ASME B89.4.22-2004

NONMANDATORY APPENDIX E
INTERIM TESTING OF AACMM SYSTEMS

E-1 INTRODUCTION similar to those stated in para. 6 of the Standard, which

typically require the uncertainty of an artifact to be less
The goal of AACMM interim testing is to identify and than 20% of the AACMM stated performance.)
rapidly remove from service defective AACMMs before Thermal properties of the artifact are also important for
significant numbers of good parts are rejected or bad workpieces measured at a temperature other than 20°C.
parts are accepted. The frequent application of interim In general, the user should select an artifact that has a
testing will increase confidence in AACMM perform- thermal expansion coefficient that is similar to that of the
ance between AACMM calibrations. Interim testing is workpieces commonly measured with the AACMM. The
not a substitute or replacement for AACMM calibration, uncertainty in the thermal expansion coefficient of the ar-
and is not normally diagnostic in nature. Rather, tifact must also be considered, as discussed in para. 4.2.
it checks the validity of the calibration by detecting If the user commonly applies a correction for the thermal
common AACMM performance failures. It is recom- expansion of the workpiece, then a thermal compensa-
mended that users regularly apply interim testing to tion should be applied to the interim artifact. This will al-
their AACMMs. An effective interim test checks the low testing of the thermal compensation system as a part
AACMM measurement system including subsystem of the interim test procedure. Note that the temperature
components that are used in the normal operation of sensors are a part of the thermal compensation system
the AACMM. This may include such components as and are subject to damage and drift. Since environmen-
probes, probe heads, temperature compensation sys- tal conditions may affect the performance of an AACMM,
tems, and rotary tables. This document assists AACMM it is advisable to record the temperature and other envi-
users by providing information on efficient interim ronmental parameters during an interim test, particularly
AACMM testing. if unusual conditions are present.
It is important that the artifact be dimensionally sta-
E-2 GENERAL INTERIM TESTING GUIDELINES ble between interim tests, so that the measurements ob-
tained during an interim test can be compared to those
Limited time is available for performing interim test- from previous interim tests and, if available, to the ar-
ing; hence an efficient test must concentrate on sources of tifact’s known length. Certain materials are dimension-
performance degradation that commonly occur. The goal ally unstable and may change in length by many mi-
is to test for as many errors as possible with a minimum crometers per meter over one year. It is important that
number of measurements. If the test fails, additional ac- the dimensional stability (including any possible dam-
tions are needed. These might involve further diagnostic age) of the artifact be substantially less than the small-
testing or involve AACMM servicing and calibration. est AACMM error of significance to the user. The in-
AACMM subsystem components need to be included in terim artifact should be securely located on the AACMM
the interim test to broaden its scope and insure that the table to prevent any possible rocking or slippage dur-
entire measurement system is operating correctly. Each ing the measurement procedure. To compare interim
user has special needs, so interim testing procedures and test results to one another, it is advisable to locate the
artifacts may vary from user to user; however, the fol- artifact in approximately the same position and orien-
lowing may provide some guidance. tation for all tests. Additionally, the inspection plan,
AACMM errors, whether systematic or random, re- such as the number of probing points taken on the ar-
veal themselves as deviations from known lengths or as tifact, should be kept constant for all tests. Widely dis-
variations of several measurements of a fixed (perhaps tributing the probing points over the gaging surface will
unknown) length. The use of a known length artifact aid in producing consistent interim testing results.
supplies additional useful information from the test. If
a known length artifact is used, the uncertainty in its
length determination should be small compared to the E-3 INTERIM TESTING STRATEGIES
threshold level at which the interim test fails. Similarly,
the form and surface finish of the artifact should not sig- There are several different strategies for choosing an
nificantly affect the measurement. (These conditions are interim artifact, depending upon the application of the

26
NONMANDATORY APPENDIX E ASME B89.4.22-2004

user. For discussion purposes, we will consider two cat- of calibrated ball or hole plates may allow more than
egories: those strategies that employ an artifact that rep- one such body diagonal to be measured in each orien-
resents a typical workpiece (the artifact may be an ac- tation.
tual workpiece from the production line) and those An artifact specifically designed for interim testing
strategies that employ an artifact specifically designed should be sensitive to common AACMM errors. The
for AACMM testing. For all strategies, it is recom- AACMM probe should be checked to ensure it is in good
mended that ten consecutive interim testing runs be con- working order. This may involve an explicit probe test
ducted immediately after the AACMM is calibrated. The that checks the directional sensitivity of the probe, i.e.,
mean of these ten measurements can be used to estab- probe lobing, or may be incorporated into part of the gen-
lish a baseline value for the interim artifact, and the eral AACMM test, such as measuring a long gage block
range of values indicates the typical variation that may that is oriented in several different directions. Similarly,
be expected under these conditions. Additional factors, to test the probe calibration, which involves the accuracy
such as thermal conditions or different operators, may of the AACMM probe calibration artifact (typically a
further expand the range of interim testing results. If, sphere with a calibrated diameter), a true bidirectional
upon recalibration of an AACMM, the new interim base- measurement of a known length is required.
line measurements differ significantly from the previ-
ous baseline, then the interim artifact or the AACMM
E-4 INTERIM TESTING EXAMPLE
calibration (or both) may be suspect and further inves-
tigation is warranted. In this example, a user with an AACMM employing
Some AACMMs are dedicated to measurements of a a hard probe has selected an interim testing artifact, con-
single type of workpiece or a family of similar work- sisting of a 25 mm diameter sphere connected at one
pieces. In this situation an actual workpiece may be used end of a rod, and a conical socket at the other end. The
as the interim testing artifact. This type of artifact will be rod length is roughly 75% of the AACMM length. The
sensitive to errors that are important to actual workpiece sphere is calibrated for form and diameter. Similarly,
measurements. An additional benefit is that the user is the length of the rod from the sphere center to the hard
familiar with the required workpiece measurements and probe tip center (when placed in the conical socket) is
consequently may have an AACMM program available also calibrated. The sphere is mounted in a three-point
which can be used for the interim testing. The selected kinematic magnetic mount that allows the rod to rotate
workpiece and the measured features on that workpiece about the sphere center, sweeping out an arc of constant
should span the largest volume of the AACMM work radius.
zone that is encountered during actual workpiece mea- The user measures the sphere while it is in the kine-
surements, to insure that the relevant volume of the matic mount with many points well distributed over its
AACMM is tested. For users measuring many small surface. The form, diameter, and center location are
workpieces located all over the AACMM work zone, it computed using a least squares fit. The reported form
is suggested that the small interim test artifact be mea- error is an indication of short-range errors of the
sured at several different locations to insure that an ad- AACMM, especially those due to repeatability prob-
equate region of the work zone is tested. It is not neces- lems. The diameter of the sphere is a feature of size, i.e.,
sary to measure every feature on the test workpiece; a bidirectional length measurement, that checks the abil-
rather, a representative group should be selected (both ity to correctly account for the probe tip size.
for feature type and location) for the interim testing pro- The point-to-point length from the center of the
cedure. The tolerance of these selected features should sphere to the center of the probe tip (located in the con-
be set comparable to those of the tightest tolerances ical socket) checks the ability to measure large lengths.
found in the actual production workpieces. In general, Rotating the bar into different orientations (leaving the
the interim artifact should be treated, fixtured, and mea- sphere end in the kinematic mount) allows many dif-
sured in a manner similar to that of the actual workpiece ferent orientations to be rapidly checked. At least four
to reflect the actual measurement situation. Although the different lengths (each in a different octant) should be
use of a test workpiece as an interim artifact has merit, checked and compared against the calibrated rod length.
it is important to note that the testing results are valid If the AACMM is equipped with temperature compen-
only for workpieces of a similar design and may not in- sation capability, this should also be used when mak-
dicate the errors present when measuring a workpiece ing the rod length measurements.
significantly different from the test artifact. The user records the ball diameter error, the calcu-
An artifact specifically designed for interim testing lated sphere form error, and the largest point-to-point
should be sensitive to common AACMM errors. On ar- rod length error of the interim test. If desired, these three
tifacts that produce several lengths upon measurement values can be further combined (in a root sum of squares
(e.g., ball plates), the longest length present will provide manner) to yield a single value indicative of the
the greatest sensitivity to errors. Alternatively, the use AACMM performance.

27
ASME B89.4.22-2004 NONMANDATORY APPENDIX E

E-5 TESTING FREQUENCY bad workpiece or rejecting a good one. It may be use-
ful to consider the interim testing interval as a percent-
The frequency of interim testing is highly user- age of total AACMM operating hours. Some users with
dependent. An AACMM being operated three shifts a high value and/or safety critical workpieces may elect
day with multiple operators in a harsh environment is to perform daily tests, where other users might test
likely to experience many more problems than the same weekly or monthly. Additionally, interim testing should
machine being used one shift a day by a single opera- be conducted after any sort of significant event, such as
tor in an excellent environment. The frequency of test- an AACMM collision, replacement of a subsystem com-
ing is also strongly affected by balancing the cost of in- ponent, or the occurrence of abnormal temperature vari-
terim testing against the consequences of accepting a ations or gradients.

28
 ASME B89.4.22-2004

NONMANDATORY APPENDIX F
BALL BAR DESIGN AND MOUNTING RECOMMENDATIONS

F-1 PURPOSE stant distance between the balls while positioning the
ball bar in different orientations and to not deflect dur-
This Appendix contains information regarding ball ing probing. The bar is frequently made of tubing to in-
bars and ball bar mounting. crease its natural frequency of vibration and at the same
time reduce its weight.
F-2 GENERAL To minimize the effect on repeatability due to the
static deflection, it is highly recommended that the bar
It is recognized that the length of the ball bar becomes
should be supported as close as possible to the ends,
quite large for the evaluation of the volumetric per-
or in the case of the double-sphere ball bar, under the
formance of some models of AACMMs. In these cases,
inboard spheres. When used for the volumetric test,
the mounting of the bar becomes particularly important.
the ball bar will be mounted both horizontally and ver-
While the bar can be made from any cross section and
tically. Some ball bar system designs incorporate mag-
be of any size and material, for a number of reasons the
nets in the mounting sockets. These magnets provide
most commonly used ball bars are made from either a
sufficient force to restrain the bar from falling. For the
solid cylinder or a round tube, frequently made from
case where magnets are not used, the deflections
steel or invar. Invar is an iron–nickel alloy, with trace
shown in Table F-1 are for a ball bar that is assumed
elements of silicon and manganese, that has a very low
to be clamped at one end and, while horizontal, sim-
coefficient of thermal expansion.
ply supported at the other. When in a vertical orienta-
tion, the bar will be supported at the top and will, in
F-3 BALL BAR DESIGN RECOMMENDATIONS
theory, simply hang straight down. If the support is
For ball bars conforming to this Standard, the fixed considered rigid, then the measuring force will cause
length bar must be rigid and stable to maintain a con- zero deflection of the bar.

Table F-1 Maximum Static Deflection of a Ball Bar Occurring at 0.5785L

Length, mm
Bar
Case Dimensions 500 1 000 1 500 2 000 2 500

1 20 mm dia. 0.005 0.085 0.429 1.355 3.307

2 25 mm dia. 0.003 0.054 0.274 0.867 2.118
3 40 mm dia. 0.001 0.021 0.107 0.339 0.827
4 20 mm OD, 0.004 0.068 0.343 1.083 2.645
10 mm ID
5 25 mm OD, 0.002 0.033 0.167 0.529 1.290
20 mm ID
6 40 mm OD, 0.001 0.015 0.077 0.243 0.594
25 mm ID
7 20 mm dia. 0.008 0.124 0.627 1.982 4.839
8 25 mm dia. 0.005 0.079 0.402 1.269 3.099
9 40 mm dia. 0.002 0.031 0.157 0.495 1.210
10 20 mm OD, 0.006 0.099 0.502 1.585 3.871
10 mm ID
11 25 mm OD, 0.003 0.048 0.245 0.774 1.888
20 mm ID
12 40 mm OD, 0.001 0.022 0.113 0.356 0.870
25 mm ID

29
ASME B89.4.22-2004 NONMANDATORY APPENDIX F

For the assumed mounting condition, the bar, loaded

under its own weight, will have a maximum static de-
flection of WL3/185EI that will occur at 0.5785L from the
clamped end, where W is the total weight of the ball bar.
Fig. F-1 Single-Ended Ball Bar
The deflection at the center of the bar, i.e., at L/2, will
be WL3/192EI.
Table F-1 shows the maximum deflections, in mil-
limeters, for various bar cross sections and lengths. As
the density of steel (7 530–8 140 kg/m3) and the density Fig. F-2 Double-Ended Ball Bar
of invar (8 080 kg/m3) are so similar, the weights will,
for practical purposes, be the same for either material.
However, as the modulus of elasticity for steel (205 GPa) L
is larger than that for invar (148 GPa), the deflection of
the invar bars will be larger than that for the steel bars.
In Table F-1, Cases 1 through 6 are for steel, while Cases
7 through 12 are for invar bars of the same cross section.

Fig. F-3 Ball Bar Mounted With Overhang

F-4 ERROR SOURCES at Each End
There are various sources of error that can contribute
to the total error of a ball bar system. A brief discussion
of these errors follows. The total static deflection when the bar is mounted
horizontally and points are being taken around the
equator is then
F-5 MOUNTING CONFIGURATION ERRORS
L3
total , mm  1  2   (8F  3qL)
Two types of bar are commercially available, one with 24EI
two spheres at each end and one with a single sphere
Note that the deflection is a function of the length
at each end. These are shown in Figs. F-1 and F-2.
cubed or to the fourth power.
As the spheres are being probed, a significant difference
Consider the following example of a ball bar mounted
in the force used by different operators may exist. A more
as shown in Fig. F-3. For a 20 mm diameter steel ball
thorough treatment of measurement errors in free-stand-
bar, the second moment of area is found from
ing ball bars is presented in a technical paper, Properties of
Free Standing Ball Bar Systems, published in the Journal of d4
Precision Engineering, January 1993, Vol. 15, No. 1. I    7 854 mm4
64
As noted in para. F-3, the bar should be supported as
close to the end as possible. If this is not done, as the The resultant end deflection (ignoring the weight of
support is moved further away from the end, the over- the ball itself) as a function of overhang distance for var-
hanging portion can be modeled as a cantilever. The ex- ious forces is shown in Fig. F-4 for a 20 mm diameter
cessive deflection will cause an apparent shortening of bar. Figure F-5 is the same plot for a 25 mm OD, 20 mm
the ball bar. In this case, the static deflection at the end ID invar tube. For this cross section,
of the bar due to the probing force can be found from

FL3  
I   D4  d4  11 321 mm4
64
1, mm  
3EI The fixture must also be stable and rigid to hold the
where ball bar in its positions and to not deflect while probing
E  modulus of elasticity the balls. It must not obstruct access to the balls when
F  force the bar is oriented in the various positions. For low and
I  second moment of area of the bar moderate accuracy AACMMs, most fixtures will be ad-
L  distance from the support equate for testing volumetric performance. For more ac-
curate AACMMs, the stiffness of the ball bar system
The weight of the bar will cause an additional de- must follow proper design guidelines. In general, the
flection of objective is to keep the deflection as low as possible,
qL4 which again illustrates why the bar should be mounted
2, mm   as close to the end as possible.
8EI
It should be noted that a major attribute of the artifact
where mounting is I (the area moment of inertia, more correctly
q  weight per unit length of the overhanging bar termed the second moment of area). For a circular or tu-

30
NONMANDATORY APPENDIX F ASME B89.4.22-2004

0.50

0.45

0.40

0.35
Deflection, mm

0.30

0.25

0.20

0.15

0.10

0.05

0.00
0 100 200 300 400 500 600 700
Overhang, mm

 0.5 N
 1.5 N
 2.5 N

Fig. F-4 Static Deflection as a Function of Force and Overhang Length for a Steel Bar

0.30

0.25

0.20
Deflection, mm

0.15

0.10

0.05

0.00
0 200 400 600 800
Overhang, mm

 0.5 N
 1.5 N
 2.5 N

Fig. F-5 Static Deflection as a Function of Force and Overhang Length for an Invar Tube

31
ASME B89.4.22-2004 NONMANDATORY APPENDIX F

Fig. F-6 Support Bar Mounted Horizontally

bular section, this is obviously a nonissue because of sym-

metry. For a support bar made from a rectangular sec- Fig. F-7 Support Bar Mounted Vertically
tion, the second moment of area is found from
b3
d
I, mm4   This is particularly true when bringing the ball bar
12
from the outside into a laboratory. Temperature
For a rectangular bar 25 mm  10 mm, the second mo- changes of 10°C–15°C are not uncommon. When this
ment of area can be either 2 083 or 13 021 mm4, de- situation occurs, it is desirable to have some estimate
pending on the orientation. This is illustrated in Figs. F-6 of the thermal soak time.
and F-7. The thermal soak time is based on a thermal time con-
When the ball bar is being used for the volumetric stant, , for an item and its surrounding environment.
testing and is mounted in a vertical orientation, a fore- The thermal time constant depends on the thermal ca-
shortening error will occur due to gravitational effects. pacitance, volume, and surface area of an item, and the
The end compression in millimeters can be found properties and movement of the surrounding environ-
(with a consistent set of units) from L2/2E, where  is ment.
the specific weight. The foreshortening, in mm, of 20 The temperature of an item over time, after experi-
mm diameter steel bars and 25 mm OD invar tubes of encing a sudden change in the temperature of its sur-
various lengths is shown in Table F-2. rounding environment, may be expressed as
f(t)  (Tf  Ti)(1  et/)
F-6 THERMAL ERRORS
where
Temperature changes have a detrimental effect on f(t)  change in temperature of the item for a given
accurately measuring a ball bar. One event that causes time interval t
temperature errors is moving a ball bar system from t  time interval from the start of the change in en-
one environment to another at a different temperature. vironmental temperature to any later time

Table F-2 Foreshortening of Bars as a Function of Length

Length, mm
Bar
Dimensions 500 1 000 1 500 2 000 2 500

Solid steel, 20
mm OD 0.0001 0.0002 0.0004 0.0008 0.0012
Tube invar, 25
mm OD, 20
mm ID 0.0001 0.0003 0.0006 0.0011 0.0017

32
NONMANDATORY APPENDIX F ASME B89.4.22-2004

Tf  final temperature of the item, which is the tem- Table F-4 Ball Bar Area and Volume
perature of the surrounding environment after as a Function of Length
the sudden change
Ti  initial temperature of the item Length, Length of Surface Volume,
mm Cylinder, mm Area, mm2 mm3
  thermal time constant for the item
300 275 17 281 86 405
When t  , f(t)  (Tf  Ti)(1  e1)  0.632(Tf  Ti).
500 475 29 849 149 245
This means that the item has experienced a change in 700 675 42 417 212 085
temperature of 63.2% of the difference between its ini- 900 875 54 985 274 925
tial and final temperatures. 1 000 975 61 269 306 345
When t  3, the change in temperature has attained 1 200 1 175 73 837 369 185
95%. Similarly, t  5 represents a change in temperature
of 99.3%. As time (t) increases beyond 5, the change in
temperature approaches 100% of the original difference
Ball bars are usually manufactured in millimeters.
in temperature. For many practical applications, 4 is
Table F-4 shows the surface area and volume for a num-
used for the thermal soaking as the item will then be at
ber of different-length ball bars.
greater than 98% of its steady-state temperature.
The total surface area and volume can now be found
 may be calculated from the following equation:
by simply adding the three elements together as shown
in Table F-5.
  CV/hA (1)
As can be seen, the ratio is almost constant, as one
would expect because the volume/area for the cylinder
where is constant at r/2 and the contribution of the spheres is
A  surface area, m2 very small.
C  thermal capacitance, J/°C m3 From Eq. (1), using the values of volume/area from
h  convective film coefficient, W/m2°C Table F-5, and taking the value of thermal capacitance,
V  volume, m3 C, for steel as 990 J/°C m3 and the convective film co-
In most cases, the fluid medium will be slowly mov- efficient, h, as 11.5 W/m2°C, the thermal time constant 
ing air. For such a situation, assuming a value of 11.5 for (990  4.93)/11.5  0.38 hr or 22 min.
the convective film coefficient is usually adequate. For The ball bar should be allowed to soak in a constant
additional information on estimating thermal time con- temperature environment for 4 or about 1.6 hr prior to
stants, thermal soak times, or fluid heat transfer proper- any measurements being taken.
ties, see ASME B89.6.2 or any textbook on heat transfer. Handling the ball bar also causes thermal errors. This
The following illustrates the calculation of a thermal can occur during assembly of the ball bar to add or re-
soak time for the ball bar. The ball bar can be consid- move an extension or by grabbing hold of the bar to re-
ered as being made up of three elements, two spheres orient it. Using gloves when handling the bar will re-
and a cylinder. The spheres are 25 mm in diameter and duce the heat transfer from one‘s hand to the bar. Also,
the bar is 20 mm in diameter. The surface area and vol- a plastic sleeve on the bar will reduce the thermal
ume for the elements are shown in Table F-3. growth when handling the bar. Experiments have de-
The properties of the spheres will be constant regard- termined that a typical steel ball bar will return to its
less of the length of the bar. Substituting the value of r  original length to within 1
m in 30 min after minor
12.5 mm into the equations for the sphere gives the sur- handling. Invar ball bars did not exceed 1
m change
face area as 1 964 mm2 and a volume of 8 181 mm3. in length during a similar test.
The length of the ball bar is from the center of one The ball sphericity as defined in the Standard should
sphere to the center of the other. The length of the cylin- be less than 20% of the AACMM repeatability.
der can then be found by subtracting 25 mm, i.e., two
radii.
Table F-5 Ratio of V/A for Ball Bars of
Different Lengths
Table F-3 Sphere and Cylinder Surface Area Length, Ratio
and Volume mm Surface Area Volume of V/A

Surface Area, 300 (2)(1 964)  17 281 (2)(8 181)  86 405 4.85
Element mm2 Volume, mm3 500 (2)(1 964)  29 849 (2)(8 181)  149 245 4.90
700 (2)(1 964)  42 417 (2)(8 181)  212 085 4.93
Sphere 4r 2
4
 r 3 900 (2)(1 964)  54 985 (2)(8 181)  274 925 4.94
3 1 000 (2)(1 964)  61 269 (2)(8 181)  306 345 4.95
Cylinder 2rL r 2L 1 200 (2)(1 964)  73 837 (2)(8 181)  369 185 4.96

33
ASME B89.4.22-2004 NONMANDATORY APPENDIX F

Table F-6 Combinations of Length from a

Commercially Available Invar Ball Bar Kit
Total Total Total Total
Total Number Length of Length of Length of
Invar Bar 125 mm 250 mm 500 mm Number of End Invar, Steel, Ball Bar,
Length (Two Bars) (Two Bars) (Four Bars) of Bars Plugs mm mm mm

125 1 0 0 1 2 100 25 125

250 2 0 0 2 4 200 50 250
250 0 1 0 1 2 225 25 250
375 1 1 0 2 4 325 50 375
500 0 0 1 1 2 475 25 500
500 0 2 0 2 4 450 50 500
500 2 1 0 3 6 425 75 500
625 1 0 1 2 4 575 50 625
625 1 2 0 3 6 550 75 625
750 0 1 1 2 4 700 50 750
750 2 2 0 4 8 650 100 750
750 2 0 1 3 6 675 75 750
875 1 1 1 3 6 800 75 875
1 000 0 0 2 2 4 950 50 1 000
1 000 0 2 1 3 6 925 75 1 000
1 000 2 1 1 4 8 900 100 1 000
1 125 1 0 2 3 6 1 050 75 1 125
1 125 1 2 1 4 8 1 025 100 1 125
1 250 0 1 2 3 6 1 175 75 1 250
1 250 2 0 2 4 8 1 150 100 1 250

Ignoring the thermal correction may not be the best 125, 250, and 500 mm, although other lengths up to 1 500
strategy for ball bar kits that are screwed together to form mm are readily available. In a commercially available kit,
longer lengths. Each bar typically has a threaded stainless all lengths are certified to
0.003 mm or
0.008 mm.
steel plug in each end, usually approximately 12.5 mm in Table F-6 shows how different ball bar lengths can be
length. Typical lengths in commercially available kits are obtained from a commercially available kit.

Table F-7 Natural Frequency of Ball Bars as a Function of Length

Bar
Case Dimensions 250 500 750 1 000 1 250

1 20 mm dia. 186 46 21 12 7
2 25 mm dia. 248 62 28 15 10
3 40 mm dia. 371 93 41 23 15
4 20 mm OD, 224 56 25 14 9
10 mm ID
5 25 mm OD, 310 77 34 19 12
20 mm ID
6 40 mm OD, 447 112 50 28 18
25 mm ID
7 20 mm dia. 154 38 17 10 6
8 25 mm dia. 205 51 23 13 8
9 40 mm dia. 307 77 34 19 12
10 20 mm OD, 185 46 21 12 7
10 mm ID
11 25 mm OD, 256 64 28 16 10
20 mm ID
12 40 mm OD, 370 92 41 23 15
25 mm ID

34
NONMANDATORY APPENDIX F ASME B89.4.22-2004

For example, for a change in temperature of 5°C, a Distortions due to gravity cause two effects, which
nominal 1 250 mm long bar, made up of (2  125)  can be combined for tilted bars: an overall shorten-
(2  500) bars and treated purely as an invar bar, will ing of the ball bar from compression along the verti-
expand by (1 250) (1.26  106) (5)  0.0079 mm. cal direction, and a shortening of the ball bar due to
However, including the steel plugs will give an ex- the sag of the bar when it is held horizontally. For
pansion of (1 150) (1.26  106) (5)  (100) (11.7  106) ball bars as commonly built, the first effect is very
(5)  0.0131 mm, or a difference of about 60%. small and the second is insignificant because it is a
cosine error.
F-7 HYSTERESIS ERRORS
Applying a force parallel to the ball bar in one direc- F-9 NATURAL FREQUENCY OF VIBRATION
tion and then applying it in the opposite direction can
Free-standing ball bar systems can be very suscepti-
check these. The applied force should be twice the prob-
ble to vibrations. For many designs, both the post and
ing force. The test should then be repeated, but with the
the ball bar act as cantilever beams with no restraints to
direction of the force perpendicular to the ball bar axis.
dampen the motion. The vibration can be caused by ex-
The hysteresis in each direction should be less than 20%
ternal sources or can be induced by the motion of the
of the AACMM repeatability.
AACMM operator.
The natural frequency of vibration of the overhang-
F-8 PROBING ERRORS ing section can be found from
Probing forces may cause deflections in the ball bar
1 g
and the fixture. Lateral deflections have a direct influ-
ence on determining the distance between ball centers,
n  
2
 Hz
because symmetrical deflections cause the ball diame- The resultant values are shown in Table F-7. Cases 1
ters, as fit to the data, to appear smaller and thus a longer through 6 are for steel, while Cases 7 through 12 are for
ball bar length (center-to-center distance) is calculated. invar bars of the same cross section.

35
ASME B89.4.22-2004

NONMANDATORY APPENDIX G
KINEMATIC, CONICAL SEAT, AND
CHAMFERED HOLE DESIGN RECOMMENDATIONS

G-1 PURPOSE G-3 TRIHEDRAL SOCKET CONSIDERATIONS

Simply stated, the principle behind kinematic mount- There are several ways to construct the trihedral socket
ing is that a rigid body should be constrained by no more or “kinematic seat.” Probably the easiest is to place three
than the number of single points equal to the number of spheres (could be ball bearings) in a flat bottom hole or
unwanted degrees of freedom of that body. Since a rigid “cup” as shown schematically in Fig. G-2. The sizes of the
body has six degrees of freedom (three translations and spheres and the hole will depend on the size of the sphere
three rotations), this dictates that the number of point con- that is to be constrained by the seat. Similarly, the depth
tacts be six or less. Constraining a rigid body in such a of the hole will depend upon the degree of access required
manner ensures that no stress is induced in the body as a for the seated sphere. In order to maintain stable contact,
result of the mounting. Under the right applications, this the angle of contact as measured from the horizontal
provides for a very stable and highly repeatable (de- should be 45 deg. In this case, the relationship between
pending on the choice of materials) mount. However, the diameter of the three spheres forming the seat (d) and
kinematic mounts are not ideal for all situations, due to the diameter of the sphere that is to rest in the seat (D)
the high local stresses that may result from point loading. will be
d  1.58D
G-2 GENERAL
and the diameter of the cup (CD) can be found from
For the purposes of this Standard, kinematic mount-
ing is used for locating precision spheres, either a ref- CD  2.16d
erence ball or an AACMM probe tip. This is most ef- Variations on this particular version of the trihedral
fectively accomplished through the use of the trihedral socket use surface-mounted hemispheres that are spread
socket. In this case, it will only be necessary to constrain out, i.e., not touching one another, which allows the use of
the three translations of the sphere. Therefore, three smaller spheres to form the seat and allows the addition
point contacts are all that are required (the three rota- of a magnet, if so desired, to help keep the seated sphere
tions have no effect on the sphere, since it is symmetric in place. If possible, the hemispheres should be placed on
about all three of its axes). Forms of the trihedral and a flat plane and spaced 120 deg apart (see Fig. G-3), as is
guidance on how to make them are shown in Fig. G-1. inherent in the three spheres in a cup case. When design-
ing this variation of mount, be sure that the sphere to be
seated does not contact the hemisphere support surface be-
fore resting on the hemispheres. Again, assuming that the
spheres contact at 45 deg, this leads to the relationship
d  0.414D

Fig. G-1 Degrees of Freedom for a Rigid Body Fig. G-2 Three-Sphere Kinematic Seat

36
NONMANDATORY APPENDIX G ASME B89.4.22-2004

cd
120 deg

Fig. G-3 Three-Hemisphere Kinematic Seat

or requires that the hemisphere support surface be re- materials can be used, the following guidelines address
lieved to accommodate the seating sphere. A second the use of a steel seat or an anodized aluminum seat.
useful relationship for this configuration gives the di- Hardness, or the resistance to permanent indentation,
ameter of the circle (cd) that circumscribes the triangle is denoted by a number of different scales. Commonly
formed by the center points of the three hemispheres: used scales for engineering metals are Brinell (HB),
Rockwell (HR), and Vickers (HV), depending to a cer-
cd  0.707(D  d)
tain extent on the material being measured. The differ-
ent tests involve pushing a ball, cone, or pyramid into
G-4 CONICAL SEAT CONSIDERATIONS the surface and expressing the hardness as a ratio of the
load and the depth or surface area of the indentation.
In less-demanding applications, an alternative to the As such, for some of the scales, empirical relationships
kinematic seat is the conical socket. These sockets or seats have been developed with other material properties, but
can be easily constructed on a drill press, grinding ma- careful attention should be paid to the units.
chine, or milling machine using a tool with the proper in- While wrought aluminum alloys are typically rela-
cluded angle. It is recommended that the cone angle be tively soft, their use as a chamfered hole seat can be con-
within the range of 60 to 120 deg and that the depth of sidered if they are anodized or, more specifically, hard
the conical socket be such that a minimum of one-third anodized. Anodizing is an electrochemical conversion
of the ball diameter lies inside the cone. If two or more process that changes the outer structure of the metal.
conical sockets are used to define distance points on an Aluminum, on exposure to air, naturally develops a thin
artifact of length, reasonable care must be exercised to oxide film that seals the surface from further oxidation.
ensure that the axes of the cones are parallel. Burnishing For most purposes, this thin oxide layer doesn’t enhance
an annulus of contact at the proper contact diameter can surface hardness, as the depth is quite shallow. Hard
further refine the conical socket. This can be accom- anodizing makes a much thicker oxide coating, up to
plished by placing a “sacrificial” hardened steel ball, of 0.08
m in thickness. The hardness of a correctly treated
the same nominal diameter as the sphere that the coni- anodized aluminum oxide coating can approach that of
cal socket is designed to locate, and hitting it with a ham- diamond, making it very suitable for consideration in
mer. Although this will destroy the ball, it will provide artifact design.
a more repeatable seat for locating spheres. It is easy to recognize that a hole with a poor surface
finish will change size and shape due to the continual
insertion of the probe tip. A poorly machined surface
G-5 CHAMFERED HOLE SEAT CONSIDERATIONS
with lots of chatter will change over a period of time.
A number of material parameters and design consid- To minimize the effect, the surface finish of the cham-
erations for a chamfered hole seat are discussed. Some fer should be 0.8
m or better, paying attention to the
of the important design requirements are hardness, sur- material selection and manufacturing process. As a com-
face roughness, wear, elastic modulus, yield strength, parison, this is about the same finish as found on com-
and mounting configuration, although each may have mercially available gear teeth.
significantly different contributions depending on Over the duration of a performance evaluation test,
whether the chamfered hole is being used for perform- wear should not be a problem for a correctly selected
ance acceptance or periodic recalibration. While many combination of suitable materials. Although wear gen-

37
ASME B89.4.22-2004 NONMANDATORY APPENDIX G

erally alters the surface topography and may result in point. A question, of course, is if it was a global plastic
surface damage, it also has a beneficial effect of reduc- deformation or a local plastic deformation due to a poor
ing surface roughness by removing the peaks from as- surface finish.
perities. Of the most common wear mechanisms, only Many times in artifact design it is desirable to con-
adhesive and abrasive wear are discussed. Both are usu- strain the artifact with no more points of contact than
ally associated with sliding, which for this application one per degree of freedom that is to be constrained. Un-
may be considered minimal. For adhesion wear to oc- fortunately, these point contacts may give rise to con-
cur, fragments of one material are detached from it and siderable stress levels in a very localized region. The re-
adhere to the other material. Abrasive wear will occur sulting contact deformation can be a significant error in
when the harder of the two materials slides into the any subsequent measurements of the artifact. The term
softer material. This second process is highly depend- Hertzian stress is sometimes used to describe this con-
ent on the force applied to the material interface. tact stress/deformation phenomenon.
This is actually a measure of the interatomic bonding The formula shown below is a typical example for a
force, but in general engineering and metrology is more spherical indentation on a flat surface, taken from a ba-
commonly associated with the linear relationship be- sic strength-of-materials textbook. While other factors
tween stress and strain in materials. For many engi- are present, it gives a good approximation of the effects
neering alloys, the relationship that the yield strength for many cases. Friction effects and form deviations will
equals 0.003 times the modulus of elasticity holds true. cause deviations from the Hertz equations.
However, it should be noted that for alloy steels this is The deformation can be approximated from Hertz’s
not always the case. Reference should be made to rele- formula
vant handbooks for precise material specifications being
used. One important use of the value of the modulus of

9P2 1 
12 1 
22
3 2
0     
elasticity is in the calculation of the elastic deformation. 8D E1 E2
For many practical applications, for common engineer-
ing materials, this number is usually negligible. where
This property is the ability of a material to resist plas- D  sphere diameter
tic deformation and is usually considered the point that Ei  modulus of elasticity
plastic deformation first takes place. For ferrous alloys, P  perpendicular measuring force
the yield point is very noticeable, but for other materi- 0  depth of elastic compression
als, including aluminum, there is no pronounced yield
i  Poisson’s ratio

38
 ASME B89.4.22-2004

NONMANDATORY APPENDIX H
DETERMINATION OF THERMAL ERROR INDEX

H-1 PURPOSE change in length of a body due to change in tempera-

ture is described by the equation
The purpose of this Appendix is to introduce the user
to the concept of the thermal effect index (TEI), a means L  L T
of assessing thermal influences on the AACMM per-
where
formance tests and/or workpiece measurements. It is
L  nominal length of the part
approximately the percent degradation of the perform-
T  ambient temperature during the test or
ance test or measurement under consideration due to
measurement
thermal effects. The thermal error index is composed of
  nominal material coefficient of thermal
four parts: the nominal differential expansion (NDE) er-
expansion
ror, the uncertainty of nominal differential expansion
L  change in length
(UNDE), the uncertainty due to temperature measure-
T  change in temperature (T  20), °C
ment (UTM), and the thermal variation error (ETVE).
The methods used in this Appendix closely parallel For measurement instruments, the thermal expansion
those of the ISO Technical Report 16015, Geometric manifests itself in two ways: the expansion of the work-
product specification (GPS) — Bias and uncertainty of piece and the expansion of the instruments‘ scale. The
dimensional measurements due to thermal influences, resultant of these two are known as differential thermal
which is a modernization of ASME B89.6.2, Tempera- expansion or, when using a nominal value for the COTE,
ture and Humidity for Dimensional Measurement. The as the nominal differential expansion (NDE). This is cal-
user is referred to these documents, along with the ISO culated from the equation
Guide to the Expression of Uncertainty in Measurement
(GUM), for a more thorough treatment of this and re- NDE  Lw  Ls  wLw Tw  sLs Ts
lated subjects. where
s  subscript denoting scale
H-2 NDE, UNDE, UTM, AND ETVE w  subscript denoting workpiece
H-2.1 Nominal Differential Expansion Error (NDE) For the AACMM, the nominal length of the scale can
be considered to be the length of the arm. If the
All materials exhibit the property that they expand or
AACMM has provisions for correcting for workpiece,
contract when they undergo a change in temperature.
scale, or both, then the corresponding values can be set
This behavior is quantified by a parameter known as the
to zero in the above equation. If the supplier does not
coefficient of thermal expansion (COTE), which may be
provide an automated means to correct for the differ-
positive, negative, or very near zero. Typical values of
ential expansion between the AACMM and the work-
COTE for some common engineering materials are
piece, then this error must be calculated and included
listed in Table H-1. In a simple first-order model, the
in the TEI.

Table H-1 Nominal Coefficient of Thermal

H-2.2 Uncertainty of Nominal Differential
Expansion for Various Materials
Expansion (UNDE)
COTE COTE It is impossible to exactly know the coefficient of ther-
Material  106/°C  106/°F
mal expansion of either the AACMM or the artifact/
Plain carbon steel 11.5 6.4 workpiece under test. Consequently, any calculation of
Stainless steel 17.0 9.4 the differential thermal expansion will be in error. The
Aluminum alloys 23.0 12.8 UNDE attempts to account for this error by assigning
Carbon graphite 0.9 0.5 an uncertainty to the nominal differential expansion cal-
Plastics 80–200 44–110
culation. For those cases where the COTE is not explic-
Invar 1.0 0.6
itly calibrated, the error in the COTE value is typically

39
ASME B89.4.22-2004 NONMANDATORY APPENDIX H

estimated to be
1.0  106/°C for both the scale and plish this, a low thermal expansion artifact (e.g., invar,
the workpiece. As this is assumed to be uniformly dis- super invar, or steel with a known coefficient of ther-
tributed, the standard uncertainty in the COTE then be- mal expansion) is used to generate a thermally stable
comes point within the working volume of the machine at a
distance of between 75% and 90% of the length of the
1  106/°C
u()    0.577/°C arm. The thermal expansion coefficient of the artifact

3 should be chosen such that the change in length of the
Using this, the length uncertainty due to uncertainty in artifact during the test, or the uncertainty in the change
the coefficient of thermal expansion can be calculated as in the length, should be less than the single-point artic-
ulation test (SPAT) specification for the machine under
uNDE(L) 
2 
Lw Tw2 u
2( ) 
w L2s 
T 2s u2
(s) test. This will be accomplished if
For the case where Tw  Ts  T, this reduces to L  T  SPAT/2


2  L2) T2
(Lw
uNDE(L)   s where
3 L  length of the arm
  thermal expansion coefficient of the artifact bar
T  temperature deviation from 20°C that is
H-2.3 Uncertainty Due to Temperature expected during the test/measurement, °C
Measurement (UTM)
H-2.4.2 ETVE Test Methodology. The artifact is
Similar to the UNDE, this uncertainty source arises mounted between the AACMM base and the probe axis
from the inability to measure the temperature of the using a ball and magnetic trihedral socket arrangement.
AACMM and the workpiece without error. This again For more information on the construction of kinematic
results in an error in the nominal differential expansion seats, see Appendix G. It does not matter whether the
correction that depends on the accuracy of the ther- socket or the ball is substituted for the probe; however,
mometer. For calibrated thermometers, the standard un- if the socket is used to replace the probe, the artifact then
certainty is calculated from the information given by the becomes the familiar ball bar. If it is necessary to sup-
uncertainty of that calibration. For typical uncalibrated port the arm and/or artifact in order to perform this test,
thermometers (e.g., thermistor, thermocouple),1 an esti- the artifact should be located in a horizontal plane at the
mate of
0.2°C (
0.4°F) can be used, resulting in a stan- height of the first machine axis. Additionally, the arm
dard uncertainty for temperature measurement of should be constrained only by its own weight and us-
0.2 ing the minimum number of contact points necessary in
u(T)    0.12°C order to minimize external forces, both thermally in-

3
duced and otherwise. The goal is to ensure that the
The uncertainty in temperature measurement can metrology loop (i.e., the path through the structural
then be used to calculate the length uncertainty due to components that can influence the measurement) is
temperature measurement: comprised entirely of the AACMM and the test artifact.
While, for example, clamping the arm or test artifact in
uTM(L) 

 2 u2(

w Lw
2 Tw) 
2s L
2 u2(T)
s s
a fixture that is rigidly mounted to a surface plate might
appear to be a good idea from a structural stability
H-2.4 Environmental Temperature Variation Error (ETVE) standpoint, what will actually be measured during the
Environmental temperature variation error (ETVE) is course of the test is the thermal drift of the arm, the ar-
determined by a drift test and is sometimes referred to tifact, the surface plate, and possibly the fixture.
by that name. Any drift test should be conducted for a Prior to conducting the ETVE test, the AACMM shall
period equal to the duration of the longest performance be maintained in the thermal environment under eval-
test/measurement task. This short period is a compro- uation for a period of at least 8 hr (overnight is recom-
mise, and users are strongly advised to run this test for mended) with the machine in its “resting” position.
a time period of at least 24 hr, as many temperature ef- With the artifact securely positioned between the
fects exhibit daily periodicities. AACMM base and probe axis, measurements of the
length of the bar are then conducted for a time period
H-2.4.1 ETVE Test: General. The procedure for mea- at least as long as the time required to conduct the
suring the ETVE of the AACMM requires monitoring longest performance test specified in this Standard. The
the apparent drift of a fixed point in space. To accom- test period shall be divided into intervals of approxi-
mately 1 min, during which three measurements of the
1 This is a very conservative estimate of uncertainty for these
bar length are made. In order to minimize the effect of
types of temperature systems and does not reflect the level of un-
nonrepeatability, the mean value of the three lengths for
certainty that could be achieved when these systems are properly each interval shall be calculated and reported as a sin-
calibrated. gle value for that interval.

40
NONMANDATORY APPENDIX H ASME B89.4.22-2004

H-2.4.3 ETVE Test Data Analysis. The range of varia- error index. This represents the maximum thermal er-
tion of these mean length values shall be the ETVE. Un- ror contribution to a length measurement as a percent-
less otherwise stated, the ETVE will be assumed to be an age of the desired artifact or workpiece tolerance (TOL).
uncertainty source that is uniformly distributed and of full The TEI is given by the equation
width given by the range of value of the drift test, i.e., the
ETVE. The resulting uncertainty in length measurement
due to environmental temperature variation error is then:
TEI  2  
NDE  2u (L)
TOL
T
  100%
ETVE or in its complete form
uETVE(L)  

2
3


ETVE2
  L 2w Tw 2 u2( )
w
12
H-2.5 Calculation of TEI NDE  2  Ls Ts u (s)  w L2w
2 2 2 2

u2(Tw)  2s L2s u2(Ts)

In accordance with the Guide to the Expression of TEI  2
Uncertainty in Measurement, the above uncertainty TOL
sources (UNDE, UTM, and ETVE) can be combined in  100%
a root-sum-squares fashion to provide an estimate of the
standard uncertainty of a length measurement due to The above formulation(s) of the thermal error index as-
thermal effects: sumes a coverage factor of k  2 for the thermal uncer-
tainty term. Substituting the combined standard un-
uT(L) 

u2ETVE(
L)  u2
TM(L) 
 u2ND
E(L)
certainty of the length measurement for the artifact/
This standard uncertainty can then be combined with workpiece tolerance gives an estimate of the uncertainty
the bias term (NDE), if applicable, to give the thermal attributable to thermal effects.

41
ASME B89.4.22-2004

NONMANDATORY APPENDIX I
STATISTICS USED IN SPECIFYING
AACMM PERFORMANCE EVALUATION

I-1 PURPOSE tances from the reference location. The maximum devia-
tion represents the largest distance away from mean point
This Standard employs a variety of statistics to describe coordinates that should occur with a well-functioning
the performance of AACMMs. Typically these are the AACMM. This statistic has a simple physical interpreta-
maximum deviation from a reference value (which is the tion, a simple calculation, and places an upper bound on
calibrated value of an artifact if one is available), the range the distance deviation one might expect from an AACMM.
of deviations from a reference value, twice the root-mean- The second statistic, 2sSPAT, is a measure of the vari-
sum-of-squares (2 RMS) of the deviations, and, in the case ation of the ten point locations measured in accordance
of the single-point articulation test, twice the standard de- with para. 5.3.
viation of the point location. The rationale for the use of This statistic is derived as follows. Compute the stan-
these statistics is described below. dard deviation of each point coordinate; for example,
the X coordinate standard deviation, sx, is given by


I-2 EFFECTIVE DIAMETER TEST n

This test is primarily concerned with detecting errors in  (Xi  Xa)2

sx  
i1
the AACMM probe tip size. This is an important param- n1
eter, as an incorrect probe tip size will be transferred into
all measured features of size, e.g., the diameter of a bore. where Xa is the average X coordinate, and similarly for
The test uses a calibrated test sphere that is measured three sy and sz.
times in accordance with para. 5.2 and the largest devia- Create the statistic sSPAT from
tion (error) from the calibrated diameter is reported. The
maximum deviation statistic is employed because sSPAT 

s 2x  sy2
 sz2
(a) it is the diameter error (not the range of errors)
which represents the radius of a spherical region, cen-
that concerns users, since it is this error that transfers
tered at the average point coordinate, in which the meas-
additively into all features of size
ured point coordinates are likely to be found (sSPAT is
(b) the probe tip size error is typically due to a sys-
the standard deviation of the point coordinate vector).
tematic effect, e.g., a poorly calibrated artifact supplied
Using the definition of sx (from above) and i [from Eq.
with the AACMM, hence a systematic effect is best (and
(1)], we get
most clearly) reported as a diameter error and
(c) given the significance of this error, reporting the


n
maximum error is prudent  2i
sSPAT  
i1
n1
I-3 SINGLE-POINT ARTICULATION
which is equivalent to Eq. (2).
PERFORMANCE TEST
Twice the standard deviation of the point locations
The SPAT seeks to measure the variability in meas- is reported (2sSPAT), as is customary with the ISO Guide
uring a single physical point in space. Since a point in to the Expression of Uncertainty in Measurement
space does not have an associated calibrated value, the (GUM), to provide a value that contains most (typi-
average point coordinates are taken to provide a refer- cally 95%) of the distance variation of a point location
ence location. The physical quantity of interest to the from its average value. The 2sSPAT statistic is more ro-
user is the distance [given by Eq. (1)] of each data point bust than the maximum deviation, since it includes all
from the reference location. This represents how far the data and is not solely determined by one extreme
away the measured point coordinate might be from the value. Furthermore, the 2sSPAT statistic is useful when
mean point coordinate. performing an uncertainty evaluation, as it represents
Paragraph 5.3 reports two statistics about the magni- a Type A uncertainty contributor that is in a form that
tude of the point coordinate errors as measured by dis- is utilized by the GUM.

42
NONMANDATORY APPENDIX I ASME B89.4.22-2004

I-4 VOLUMETRIC PERFORMANCE TEST uation of Coordinate Measuring Machines, volumetric

performance specification for Cartesian CMMs.
The volumetric performance test seeks to report in- Twice the RMS of the length measuring errors is also
formation about the AACMM length measuring error reported for the 20 length positions and orientations. This
between two points in space. Since we are detecting test uses a calibrated artifact and hence measurement er-
length errors, a calibrated length is required, typically rors are determined; thus the errors for both the short and
the center-to-center distance of a ball bar. This test re- long ball bar can be combined in this statistic. Since the
ports three statistics: the maximum length error, the artifacts are calibrated, we do not consume one degree of
range of length errors, and twice the RMS of the length freedom in computing the mean error, but rather compute
errors. the RSS value about the true value, taken as the calibrated
The maximum length error is reported, as it has a value of the artifact. Consequently, the denominator of the
simple physical interpretation, a simple calculation, RSS calculation has the value of n (instead of n  1) to re-
and places an upper bound on the magnitude of the flect the fact we are using the calibrated, not mean, value.
length error one might expect from an AACMM. Ad- Twice the RMS is reported, as is customary with the
ditionally, this statistic is similar to the performance GUM (referenced above), to provide a value that con-
value employed by the international standard on tains most (typically 95%) of the length errors of the ar-
CMM performance evaluation (ISO 10360-2, GPS — tifacts as measured using this test procedure. The RMS
Acceptance and reverification tests for coordinate statistic is more robust than the maximum deviation,
measuring machines — Part 2 CMMs used for mea- since it includes all the data and is not solely determined
suring size) and allows some comparison with CMMs by one extreme value. Furthermore, the RMS statistic is
specified to that standard. useful when performing an uncertainty evaluation, as it
The range of length errors is reported for comparison represents a Type A uncertainty contributor that is in a
with ASME B89.4.1b-2001, Methods for Performance Eval- form that is utilized by the GUM.

43
ASME B89.4.22-2004

NONMANDATORY APPENDIX J
APPLICATION OF DECISION RULES

J-1 GENERAL zones, the AACMM is considered acceptable with re-

gard to the performance specifications. If any test result
Most standards define a minimum set of performance lies in a rejection zone, the AACMM is considered to be
tests with associated methods for data analysis and re- out of specification and may be rejected. (A rejected
duction. However, to optimize the benefits of stan- AACMM may be retested after appropriate adjustment
dardization, a decision rule must also be applied to the by the manufacturer; however, the entire set of per-
testing results. For example, the decision rule applied to formance tests must be performed upon retesting.)
an articulated arm coordinate measuring machine may
not be the same as used for a gage block or microme-
ter, but should be the same for all articulated arm co- J-3 FURTHER DISCUSSIONS ON
ordinate measuring machines. REPEATED MEASUREMENTS
ISO 14253-1:1998, Geometrical Product Specifications
(GPS) — Inspection by measurement of workpieces and In general, most performance tests will be conducted
measuring equipment — Part 1: Decision rules for prov- only once. If the output result is considerably outside
ing conformance or non-conformance with specifications, the specified maximum permissible error, the decision
defines the “default rule” as subtracting the expanded un- that the supplier adjust the AACMM is clearly the most
certainty from the specification limits. This is an easy rule viable option, after which the entire set of tests will be
to understand and apply, but can have a significant neg- repeated. Difficulties arise when the output results are
ative economic impact. As a result, ASME B89 standards only marginally outside of the specification limits. If the
have adopted the general statement, “B89 standards that test result lies in the rejection zone by no more than 25%
adopt standards referencing ISO 14253-1 as a normative of the specification value, then the test may be repeated.
standard shall explicitly state a different default decision The arithmetic mean of the two test results shall be taken
rule, where the 4:1 simple acceptance and rejection rule as the test result.
from B89.7.3.1 shall be the default rule unless a different
rule is specified.” Simple acceptance and rejection are de-
fined in para. 4.1 of ASME B89.7.3.1-2001, Guidelines for
J-4 TEST UNCERTAINTY EVALUATION ISSUES
Decision Rules: Considering Measurement Uncertainty in When the simple acceptance decision rule is used, the
Determining Conformance to Specifications. issue of evaluating uncertainty components needs to be
addressed. For the purposes of this Standard, the un-
certainty under consideration includes all sources asso-
J-2 REQUIREMENTS FOR DECISION RULES
ciated with the testing equipment under the conditions
For AACMMs that are tested using the procedures that prevail at the time of testing. It is the test uncer-
described in this Standard, the simple 4:1 acceptance tainty that must be one-fourth the specification. Since a
and rejection rule implies the following: test may involve multiple measurements, the uncer-
The acceptance zone includes all test results up to and tainty of the test may be different from the uncertainty
including the manufacturer’s specification value stated of the individual measurements involved in the test. The
in Fig. 3 and the rejection zone is any value outside the reproducibility of the test value is not one of the uncer-
acceptance zone. tainty sources of test uncertainty; the test value, includ-
The 4:1 ratio of uncertainty in the test result (due to ing its reproducibility, is controlled by the stated test
the testing equipment) relative to the specification must specification value.
conform to para. J-4.
Issues associated with the rejection of measurement
J-4.1 Single-Point Articulation Performance Test
values that comprise the performance tests and issues
associated with repeated performance tests must con- The only test uncertainty associated with this equip-
form to the relevant sections of paras. 5.2.3 and 5.4.3 in ment is the rigidity of the artifact and its geometrical
the Standard and to para. J-3 of this Appendix. perfection. Typically, these uncertainties will be negli-
If all performance test results lie in their acceptance gible relative to the SPAT specification values. However,

44
NONMANDATORY APPENDIX J ASME B89.4.22-2004

if it is desired to evaluate the test uncertainty, proceed J-4.3 Volumetric Performance Test Uncertainty
as follows: Place a precision indicator, having an un-
J-4.3.1 Uncertainty of a Single Length Measurement.
certainty no larger than one-tenth the specification value
The uncertainty of a length measurement will include
against the kinematic seat (or sphere). Insert the
the uncertainty of the calibrated artifact as stated on the
AACMM probe into the seat (or against the sphere). Us-
calibration certificate, the length uncertainty resulting
ing a typical AACMM probing force, probe toward and
from the uncertainty of the coefficient of thermal ex-
away from the indicator and record the deflection, in-
pansion at the temperature that prevails at the testing
cluding its sign, for each of the two probing directions,
time, and the added uncertainty due to the fixturing.
i.e., both positive and negative deflections. Repeat this
This last source can be minimized by using the best prac-
for a total of five different probing directions, yielding
tices as discussed in Appendix F.
a total of ten readings; this should include three direc-
Assuming that all of the input quantities are uncor-
tions that are approximately mutually orthogonal. An-
related with sensitivity coefficients equal to unity, then
alyze the ten readings in a manner identical to the data
the expanded uncertainty for a single length measure-
analysis for the performance test to obtain the uncer-
ment performed with the testing equipment can be
tainty associated with the rigidity of the kinematic seat
found from
(or sphere). That is, the maximum deflection observed
is the test uncertainty associated with max, and twice
the standard deviation of the deflections is the test un- U  2
2
uartifac 
t Cal 
2
uCTE 2
ufixtur 
ing

certainty associated with 2sSPAT.

The preceding analysis assumes that the contribution where the internationally accepted default coverage fac-
of a geometrically imperfect seat (or sphere) is negligi- tor of 2 is used.
ble. This is usually a good assumption, as the form er-
ror of a seat or sphere is typically a factor of 50 less than J-4.3.2 Volumetric Test Uncertainty. For this test, the
the performance specification. If the geometry of the data is gathered using short and long calibrated ball
kinematic seat is not negligible, then this uncertainty bars. In general, although not always, the uncertainty is
contribution needs to be combined (in an RSS manner) length dependent, i.e., typically the uncertainty in the
with that due to the rigidity issue. If F is the (peak to calibrated length of the short ball bar will be less than
valley) form error of the test sphere, or the maximum the uncertainty in the calibrated length of the long ball
range that a perfect sphere can move in the kinematic bar. Since the volumetric performance test involves two
seat, then the test uncertainty associated with max is calibrated lengths, the test uncertainty involves both of
0.5F and with 2sSPAT is 0.58F. these uncertainty sources. However, since this Standard
uses simple acceptance and rejection as the decision
rule, and consequently the magnitude of the test uncer-
J-4.2 Effective Diameter Test Uncertainty
tainty does not affect the size of the acceptance zone (as
The uncertainty sources that contribute to this test un- it does using the default rule of ISO 14253-1), it is usu-
certainty include the uncertainty in the calibrated di- ally sufficient to determine the volumetric test uncer-
ameter of the sphere (as stated on its calibration certifi- tainty based on the uncertainty of the larger of the two
cate), the size uncertainty in the sphere diameter due to lengths, as given below.
the uncertainty in the CTE and the temperature pre- Compute the uncertainty of a single measurement as
vailing at the time of testing (this is usually a small con- described in para. J-4.3.1 for both length standards and
tributor), and the uncertainty associated with the lack select the larger of the two; denote this value as U. (Usu-
of rigidity of the test sphere mount. (It is assumed that ally this will be expanded uncertainty of the longer of
the effect of the sphere’s form error on its diameter has the two length standards.)
already been included in the uncertainty associated with For the volumetric performance test, the following
the sphere diameter calibration.) The standard uncer- uncertainties are associated with the performance
tainty associated with the rigidity of the mounting is values:
identical to the uncertainty associated with max found
in the single-point articulation performance test and can
be evaluated in the same manner.
Performance Parameter Expanded Test Uncertainty
Since these uncertainty sources can be considered un-
correlated, the test uncertainty is Maximum deviation U
Range of deviations 2U
U  2

2

uartifac t Cal
2
uCTE 2
ufixtur ing 2 RMS of deviations 2U

45
Intentionally left blank
 ASME Services

ASME is committed to developing and delivering technical information. At ASME’s Information Central, we make every effort to an-
swer your questions and expedite your orders. Our representatives are ready to assist you in the following areas:

ASME Press Member Services & Ben- Public Information

Codes & Standards efits Self-Study Courses
Credit Card Orders Other ASME Programs Shipping Information
IMechE Publications Payment Inquiries Subscriptions/Journals/
Meetings & Confer- Professional Development Magazines
ences Short Courses Symposia Volumes
Member Dues Status Publications Technical Papers

How can you reach us? It’s easier than ever!

There are four options for making inquiries* or placing orders. Simply mail, phone, fax, or E-mail us and an Information Central
representative will handle your request.

Mail Call Toll Free Fax – 24 hours E-Mail – 24 hours

ASME US & Canada: 800-THE- 973-882-1717 [email protected]
22 Law Drive, Box 2900 ASME 973-882-5155
Fairfield, New Jersey (800-843-2763)
07007-2900 Mexico: 95-800-THE-
ASME
(95-800-843-2763)
Universal: 973-882-1167

* Information Central staff are not permitted to answer inquiries about the technical content of this code or standard. Information as to whether
or not technical inquiries are issued to this code or standard is shown on the copyright page. All technical inquiries must be submitted in writ-
ing to the staff secretary. Additional procedures for inquiries may be listed within.
Intentionally left blank
Intentionally left blank
ASME B89.4.22-2004

L08004