This article was downloaded by: [The University of Manchester Library]

On: 15 October 2014, At: 23:01

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

A S L E Transactions
Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/utrb19

Gear Failure Modes—Importance of Lubrication and

Mechanics
a
P. M. Ku
a
Southwest Research Institute , San Antonio, Texas, 78284
Published online: 25 Mar 2008.

To cite this article: P. M. Ku (1976) Gear Failure Modes—Importance of Lubrication and Mechanics, A S L E Transactions, 19:3,
239-249, DOI: 10.1080/05698197608982799

To link to this article: http://dx.doi.org/10.1080/05698197608982799

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
 Gear Failure Modes-Importance of Lubrication
and Mechanics
P. M. Ku, ASLE
Southwest Research Institute, San Antonio, Texas 78284
Downloaded by [The University of Manchester Library] at 23:01 15 October 2014

Any rational approach to gear design must consider, as a be placed on these requirements depends to a substantial
minimum, the impact of the various modes of gear-tooth failure degree on the specific application. Since these require-
on gear performance; operating life and reliability; andsize, weight, ments are in many respects mutually conflicting, the
and cost. Gear teeth may fail basically by either strength-related optimization of the design, or the achievement of an
or lubrication-related causes. However, many strength-related overall compromise, is not easy. Nevertheless, in any
failures are intuenced by lubrication, or may be induced by lubrica- rational approach to gear design, a knowledge of how
the gear performance is influenced by the various modes
tion-relatedfailures. This paper emphasizes the principal lubrica-
of gear-tooth failure is essential.
tion-related failure modes of gear teeth, and examines the complex
A complete gear system which includes the mating
interactions between gear mechanics andgear lubrication in affecting gears, the shafts and support bearings, the seals, and the
such failure modes. housing, may fail in many different ways. However, most
such failures manifest themselves in failure of the gear
-
INTROD!JCTION teeth-which may be the direct result of improper gear-
tooth lubrication, or the result of failure or improper
he subject of gear lubrication, and 'particularly the condition of the other components which lead to some
effect of lubrication on gear-tooth scoring (scuffing), was form of gear-tooth failure. The impact of gear-tooth
discussed at the First ASLE Gear Lubrication Symposium failure on gear performance is thus obvious.
held in 1959 (1). Since that time, many other conferences
have dealt with various facets of gear lubrication, design,
and manufacture (2-12). The objective of this Second MAJOR MODES OF GEAR-TOOTHFAILURE
ASLE Gear Lubrication Symposium is to take stock of
the progress that has been made in recent years, with AGMA cites 21 modes of gear-tooth failure (13), di- '

emphasis on the practice of gear lubrication. In this vided into four broad headings of wear, surface fatigue
opening paper, it is the author's assignment to provide (pitting), plastic flow, and breakage. It is believed more
a broad perspective to gear lubrication and gear-tooth logical to classify the gear-tooth failure modes under two
failure modes, hopefully laying some groundwork for the basic categories, namely, lubrication-related modes and
more specialized papers to follow. It must be emphasized strength-related nmdes. Major modes of lubrication-
that gear lubrication is a highly complex subject, which related failure are rubbing wear, scoring, and pitting.
involves not merely filubricationv as the term is commonly Major modes of strength-related failure are plastic flow
applied to a steady state of affairs; but must take into and breakage.
account the consequences of the transient nature of Rubbing wear is a loss of metal by the rubbing action
gear-tooth action, and this entails a consideration of gear between two relatively moving surfaces, when there is a
mechanics. lack of an intact oil film of sufficient thickness to separate
The principal requirements of gear design are gear the surfaces (14, 15). One form of rubbing wear is adhesive
performance (such as load-carrying capacity, vibrations, weaG or that caused by metal tr~~nsfer due to localized
noise, etc.); operating life and reliability; and size, weight, adhesionorasolid-~haseweldingProcess, and subsequent
and initial and operating The relative emphasis to detachment of particles from one or both surfaces. The
other form of rubbing wear is abrasive wear, caused by
Presented at the 30th ~,,,,,,~l ti ti^^ abrasive action between the relatively moving surfaces,
in Atlanta, Georgia, M a y 5-8, 1975 or by the presence of abrasive particles between them.
239
 These particles may be dirt or other solid contaminants, of tooth breakage is that due to bending fatigue (20).
or particles detached from the surfaces themselves due Breakage is basically a strength-related failure. However,
to severe pitting or wear. severe pitting or wear may so weaken the tooth (otherwise
Rubbing wear is of course harmful if severe, or if it adequate for the service) as to cause breakage. In this
continues at an undiminishing rate. However, rubbing connection, a lubrication-related failure may lead to
wear which diminishes with time, such as that associated breakage failure.
with a break-in process, is not damaging but in fact It is seen that gear-tooth failures may be due to either
beneficial. strength-related causes or lubrication-related causes.
Scoring is a severe form of adhesive wear, which results However, many strength-related failures are directly or
in rapid damage to one or both surfaces in relative motion indirectly influenced by lubrication.
(16, 17). In contrast to the other modes of lubrication-
related tooth failure, which generally take time to develop
IMPACT OF GEAR MECHANICS
or reach destructive magnitude, scoring occurs quite
precipitously and is, therefore, the most urgent one con- Lubrication is concerned with the behavior of inter-
fronting the designer. Obviously, as long as scoring cannot acting surfaces in relative motion. Accordingly, the study
Downloaded by [The University of Manchester Library] at 23:01 15 October 2014

be avoided, all time-dependent modes of failure are of the lubrication of any machine element must, by
essentially academic. In other words, with gears that are necessity, include a consideration of the total effect of the
adequately designed and manufactured strengthwise, following participating factors:
scoring is the first performance barrier that must be
crossed. It is only after scoring can assuredly be controlled 1. Motions without regard to the forces acting, or a
that the other failure modes become truly relevant. This study of kinematics.
accounts for the enormous emphasis to date on gear 2. Forces, displacements, and motions, which are the
scoring by researchers and designers alike. concern of what may be termedpseudo-statics(where a state
Since scoring is a form of adhesive wear, it cannot occur of rest may be assumed without introducing serious errors)
if an oil film of sufficient thickness separates the surfaces. and dynamics (the general case).
However, mere lack of an oil film of sufficient thickness, 3. Material and surface characteristics, both physical
while it inevitably leads to adhesive wear, may not cause and chemical.
scoring. In order for adhesive wear to advance to scoring, 4. Lubricant characteristics, both physical and chem-
another necessary condition must be satisfied. Although ical.
the precise mechanism of scoring is, at present, not yet 5. Characteristics of the surrounding atmosphere, both
understood, the concensus is that it is the result of intense, physical and chemical.
localized frictional heating at the rubbing contact and
is thus thermal in character. In other words, lubrication deals with the total inter-
Pitting is the consequence of repeated stress cycling of action including all physical and chemical causes and
the surfaces beyond the metal's endurance limit, which effects, which are in most respects time-dependent in
leads to surface or subsurface cracks and eventually the character. In an effort to gain an insight into some specific
detachment of fragments from and the formation of pits aspects of the total problem, one customarily begins by
on one or both surfaces (17, 18). Being a surface fatigue isolating the problem into neat, individual packages that
phenomenon, pitting takes time to develop. However, can be more readily attacked. This is a logical and neces-
while rubbing wear and scoring cannot take place if an sary learning process; but one must not lose sight of the
intact oil film of adequate thickness separates the two fact that his prescribed packages may or may not neatly
surfaces, pitting can occur even though it takes more time. simulate gear operation. It is perhaps not an over-
This is because the presence of such an oil film merely statement that although much has been learned about
modulates the intensity of the repeated surface stressing; the lubrication of idealized sliding-rolling systems in
but does not eliminate it altogether. steady-state operation, certain basic issues still remain.
Plastic pow is the surface deformation resulting from Additionally, in order to translate such idealized knowl-
plastic yielding of one or both surfaces in relative motion, edge into practice, the mechanical behavior of gears must
usually associated with heavy loads or high temperatures. be brought into focus; but this is an area that has not
Although basically a strength-related phenomenon, it can engaged the needed attention of lubrication engineers.
nevertheless be influenced by lubrication. For example, Gears employ counterformal surfaces and are thus
high temperature which results in a reduction of the subject to high contact stresses. They experience relatively
metal's yield strength may be due to inadequate lubrica- high sum velocities combined with relatively high sliding
tion. Moreover, rippling (a form of plastic flow) is appar- velocities, which may be cyclic or sustPined depending
ently related to a complex interaction between the oil film on the gear type.
and surfaces (19). Gear kinematics can be precisely defined by assuming
Breakage of a gear tooth is caused by the bending stress completely rigid gears (21, 22). Even so, the matter ac-
imposed on it by the transmitted torque. Outright break- quires much complexity with such gear types as the
age due to excessive bending stress beyond the fracture hypoids, spiral bevels, and worms. In reality, gears are
strength of the tooth is rather rare. A more common form never completely rigid, hence one must deal with the
 Gear Failure Modes-Importance of Lubrication and Mechanics 24 1

interactions between forces and displacements or motions. of gear teeth of simple geometry, under idealized condi-
One then encounters the problems of pseudo-statics and tions, has been a subject of much study (28-31), mainly
dynamics of gears, which will now be highlighted. with regard to strength-related failures. Even so, the
Surface Deformation-Since gears are not completely dynamics of a complete gear system, and also the dy-
rigid, one must consider the consequences of this fact. namics of lubricant flow to and over the gear teeth, are
One important consequence is the local elastic deforma- quite different matters. The effects of gear and lubricant
tion of the counterformal surfaces under load, which flow dynamics on lubrication-related failures, as well as
gives rise to elastohydrodynamic lubrication (2, 8). Fur- the-time-dependent chemical interactions involved in the
ther comments about EHD lubrication as applied to failure processes, are by and large not well understood
gears will be deferred until later. at present.
Tooth Dejection-The elastic deflection of the gear teeth
necessitates tooth profile modification, which affects load
ELASTOHYDRODYNAMIC LUBRICATION
sharing (23-25). Consider, for example, a set of involute
spur gears (assuming no manufacturing errors) with a In considering gear-tooth failures and their effect on
contact ratio of less than two, for which the load is gear performance, mention should first be made of elasto-
Downloaded by [The University of Manchester Library] at 23:01 15 October 2014

carried by two pairs of teeth at the beginning and end hydrodynamic lubrication. It is clear from the foregoing
of the mesh cycle, and by only one pair of teeth during that the attainment of an intact EHD film between the
the remaining portion of the mesh cycle. In this simple surfaces of mating gear teeth is a desirable design goal,
case, the relation between the load sharing pattern and as it precludes rubbing wear and scuffing, and minimizes
tooth profile modification for a particular design load pitting danger. In order to ensure full EHD lubrication,
can be established by pseudostatics with relative ease, a ratio of oil film thickness to composite surface rough-
but still with some measure of empiricism. If the contact ness of the order of two to three is believed necessary
ratio is, say, between two and three, the load is carried (17, 32). This condition must, in general, be approached
by three pairs of teeth at the beginning, middle, and end much of the time in the operation of gears lubricated
of the mesh cycle, and by two pairs of teeth during the with straight mineral oils, as otherwise excessive wear,
remaining portions of the mesh cycle. The load sharing if not scoring, is likely to occur. However, unless the
and profile modification problem of high-contact-ratio choice of gear steel is very unfortunate, it is no longer
gears is considerably more difficult to solve. Design opti- such a serious concern even when operating with straight
mization is far more complex, because the propensity of mineral oils, and certainly not with oils which provide
both strength-related and lubrication-related failures significant or substantial scoring and wear protection in
depends markedly on how the high contact ratio is boundary lubrication. Accordingly, a design based on
achieved (26). Nevertheless, high contact ratio normally the assumption of full EHD lubrication of gear teeth is
exists in such gears as helicals, hypoids, and worms; and not only unnecessary in the general context; but is, in
it is gaining in popularity for aircraft spur gears. fact, too conservative from the standpoint of size and
Other DeJlections-The gear bodies, shafts, support weight.
bearings, and housing also deform under load. These It should be emphasized that whether one desires to
deflections may modify load sharing among the teeth, design gears to operate at some high or low ratio of oil
or cause tooth misalignment. Analysis of these deflections film thickness to composite surface roughness, one has
is even more difficult than that of tooth deflection; a no real assurance that such a design condition will indeed
rational approach is currently lacking. be achieved in practice. Strictly speaking, the classical
Tooth Misalignment-Tooth misalignment may be due minimum EHD film thickness equation, such as Dow-
to the numerous bulk deflections mentioned above, man- son's earlier equation (33) or his more recent one (34),
ufacturing errors, stackup of tolerances in the assembly applies to a steady-state, flooded, isothermal flow
process, or differential thermal expansion. Whatever the through a rectangular conjunction of perfectly smooth
causes, misalignment can substantially affect both surfaces. Use of such an equation for other than these
strength-related and lubrication-related failures (20, 27). stated conditions requires extreme caution and can often
Misalignment is one of the most nasty problems to han- be misleading. The reasons are many:
dle, because it is difficult to measure and control in 1. The gear-tooth action is certainly not steady-state,
practice, and reliable prediction of its effects is still not and its consequences on EHD film development needs
available. to be examined. There are, for example, the questions
Dynamics-The dynamics of gear-tooth behavior, due of dynamic tooth load and squeeze-film effect. Dynamic
to the transient nature of tooth engagement, operation load as such is probably not a serious obstacle to the
away from the profile-modified design point, manufac- use of the classical EHD film thickness equation, because
turing errors, and externally imposed dynamic condi- the equation states that the film thickness is quite insen-
tions, is a n exceedingly complex subject. Clearly, if the sitive to load. Even the squeeze-film effect due to normal
actual tooth load is much higher than that derived for approach of the surfaces does not appear to have a sig-
the pseudo-static case, then estimates for both strength- nificant impact on EHD film thickness (35). However,
related and lubrication-related failures based on the regardless of these, the flow through the gear mesh neces-
quasi-static load can be overly optimistic. The dynamics sarily involves cyclic fluid acceleration and deceleration,
 and their effect on EHD film thickness is by and large even with a conjunction-geometry correction, may be
not well understood. questionable.
2. Due to the action of the gear teeth and the conven-
tional manner of oil supply, the state of gear-tooth lubri-
BOUNDARY LUBRICATION
cation is probably always starved, or far from the flooded
assumption. Although the effect of starvation on film The above remarks are not intended to minimize the
thickness behavior is quite well understood by assuming important contributions of the EHD theory to the un-
a n arbitrary inlet boundary location and shape with a derstanding of the lubrication of counterformal surfaces.
uniform temperature distribution across the film (36), It is only that, as knowledge on the details of EHD
these assumptions are not realistic for gears. Moreover, lubrication expands, complications begin to emerge and
there is presently no reliable way to relate the extent of further refinements appear necessary. In particular, once
starvation (i.e., the inlet boundary location) to lubricant, the operation leaves the full EHD lubrication regime,
design, and operating parameters, even under these a continuous and undisturbed oil film no longer exists
idealized conditions. between the mating surfaces. The operation then enters
3. The assumption of an isothermal flow process does. the rather ill-defined micro and partial EHD lubrication,
Downloaded by [The University of Manchester Library] at 23:01 15 October 2014

not hold in practice, due to heating caused by the viscous mixed lubrication, and classical boundary lubrication
shear of the oil in the conjunction inlet. Although a regimes-herein collectively called the boundary lubri-
correction may be applied for this effect, such as that cation regime for the sake of brevity.
due to Cheng (37), or the problem may be solved by What is even more important is the fact that the EHD
direct numerical analysis, these solutions can still be theory is basically a physical proposition, which does not
misleading unless the nonuniform temperature distribu- consider the chemical interaction that takes place. It has
tion across the inlet film is taken into account (38). Reli- been argued previously that of the three major lubrica-
able assessment of the temperature gradient across the tion-related gear-tooth failure modes, EHD lubrication
inlet film, particularly considering sliding (39) and the is not a necessary condition for pitting and not a suffi-
complex participating flow and heat transfer involved, cient condition for scoring. Therefore, in assessing the
is, to say the least, no easy task. effect of lubrication-related failure modes on gear per-
4. Actual gear-mesh conjunctions are generally not formance, the crucial question is not when and how full
strictly rectangular, but elliptic in shape. If the aspect EHD film ceases to prevail; but rather when and how
ratio of the ellipse normal to the flow or motion is large, the boundary film formed by the oil-metal-atmosphere
little error results from the rectangular assumption. interaction ceases to inhibit or minimize surface failures.
However, there are instances in gear design where this Boundary lubrication is, of course, the most investi-
condition is not satisfied. In that event, a correction for gated but perplexing subject in lubrication. Literally
the hnjunction-geometry effect, such as that due to thousands of references exist that pertain to various as-
Cheng (40), may be applied, or the problem may pects of boundary lubrication; b u t a few broad treatises
be solved by direct numerical analysis. In either case, should suffice to illustrate its scope and tremendous com-
the result is only approximate unless the complex flow plexity (6, 7, 46-50). It is not possible to deal with the
and heat transfer in the inlet region are taken into subject of boundary lubrication in simple terms. What
account. appears particularly important as pertains to the lubri-
5. Actual gear-tooth surfaces are never perfectly cation-related failure modes of gear teeth will be com-
smooth. Surface roughness and surface texture affect the mented on later, in the light of current understanding.
EHD film development in a complex manner (41-43); In any case, as the operation moves into the boundary
but there is as yet no confident way to assess their effect. lubrication regime; i.e., when contact between the gear-
Indeed, if the composite surface roughness involved is tooth surfaces takes place, rubbing wear becomes inevi-
about the same order of magnitude as the "nominal" table, scoring becomes a possibility, and pitting becomes
film thickness, EHD flow in the classical sense no longer more severe. The manifestation of rubbing wear and
prevails and the meaning of the film thickness calculated pitting damages is time-dependent, and their rates of
by the classical equation becomes quite obscure. damage depend upon the physical and chemical oil-
6. In the development of the classical EHD film thick- metal-atmosphere interactions, which are, as emphasized
ness equation, the assumption of Newtonian flow is im- previously, strongly influenced by gear mechanics. The
plicit in the use of some suitable value of pressure- occurrence of scoring is quite precipitous, and is appar-
viscosity coefficient for the oil. Although the effect of ently also controlled by boundary lubrication consid-
lubricant rheology on EHD film thickness is not very erations in some way.
large, there is considerable evidence that it is not negligi- It is very difficult to describe the lubrication-related
ble (44, 45). modes of gear-tooth failure, or to evaluate their impact
7. As mentioned previously, gears are extremely sus- on gear performance with much confidence, mainly be-
ceptible to misalignment. The presence of misalignment cause the mechanisms of the failure modes are still not
tends to distort the tooth conjunction shape so that it well understood and the effect of gear mechanics on these
is no longer elliptic with a large aspect ratio, in which failure modes-while known to be large-cannot be
case the use of the classical E H D film thickness equation, accurately assessed. However, regardless of the approxi-
 Gear Failure Modes-Importance of Lubrication and Mechanics 243

mations involved, an analysis made in 1958 by Blok (51) sliding-rolling disks, under substantially boundary lubri-
is instructive. cation conditions, is affected not only by surface rough-
In his analysis, Blok derived general expressions for ness, but also by surface texture (62). Their unpublished
the maximum power transmittable by a set of homolo- work also shows that lubricant, metal, and surface treat-
gous gears, or gears having similar design and materials, ment affect the friction coefficient. They have further
using a straight mineral oil. He assumed that rubbing observed that the friction coefficient is dependent upon
wear would not take place if the operation is in the full whether the rubbing sites are precisely synchronized, or
EHD regime. He considered that scoring was governed not precisely synchronized, in the operation of the disks
by his critical temperature hypothesis (16), with the criti- (63). Their experiments show that cross-ground disks; i.e.,
cal temperature and instantaneous coefficient of tooth disks with grinding marks running perpendicular to the
friction assumed constant. He regarded pitting as strictly direction of sliding, give lower friction than circumferen-
a mechanical consideration by ignoring the effect of tially-ground disks; and this is important because spur
lubrication. As will be seen later, these assumptions are gears, for example, are usually ground crosswise to the
rather drastic oversimplifications of very complex pheno- sliding direction. As to the effect of rubbing-site synchro-
mena. Quite apart from the difficulty of defining the nization, the unsynchronized mode of operation gives
Downloaded by [The University of Manchester Library] at 23:01 15 October 2014

mechanisms of the failure phenomena, which will hope- lower friction than the synchronized mode; and this is
fully come about in due time, it is believed that the relevant to the performance of hunting-tooth vs non-
time-dependent nature of the damages due to rubbing hunting-tooth gears.
wear and pitting can be introduced in the scheme of The generalization of the friction behavior in bound-
analysis. While such refinements will alter Blok's pre- ary lubrication is clearly very difficult; and since the
dicted trends substantially in some regimes of gear oper- participating chemical effects cannot be ignored, it ap-
ation, it is not believed that his major conclusions will pears likely that specific situations can only be dealt with
be greatly altered: In the absence of strength-related separately and, at least for the time being, empirically
failures, the maximum power transmittable through a by experimental determinations. To extend the treatment
set of homologous gears is primarily limited at low speeds to the whole range of EHD and boundary lubrication
by rubbing wear, at intermediate speeds by pitting or regimes only makes a difficult problem even more so.
scoring, and at high speeds by scoring. References (60) and (64) are examples of empirical corre-
lations of friction coefficient for some specific situations;
but great caution should be exercised in using such data.
FRICTION
Gear-tooth friction behavior is of particular interest
SCORING
not only because it affects gear efficiency and the attend-
ant friction heat removal problem; but also for its influ- Although the basic mechanism of the scoring phenome-
ence on pitting'and especially on scoring. The problem non is still largely not understood, there is good agree-
is complex because the instantaneous loads and velocities ment that the breakdown of the EHD film is only a
in the gear mesh do not remain constant. The matter necessary but insufficient condition for scoring
is further complicated by the almost inevitable occur- (16, 17, 27, 62-68). In other words, in order for scoring
rence of misalignment and dynamic loads. to occur, the operation must move not only into the
Gears operate, at least some of the time and invariably boundary lubrication regime, but must also meet an
under rubbing wear- and scoring-limited conditions, in additional requirement. However, largely because the
the boundary lubrication regime. Pitting can take place mechanism of scoring is basically unsettled, what form
in the full EHD regime; but is aggravated when the this additional scoring criterion must take is still very
operation moves into the boundary lubrication regime. much of an open question. All available evidence ap-
Friction behavior through the whole range of EHD and pears to suggest that how deeply the operation may
boundary lubrication regimes is, therefore, of interest. safely extend into the boundary lubrication regime with-
The friction behavior of counterformal conjunctions out resulting in scoring depends upon the physical and
under full EHD conditions is pronouncedly influenced, chemical nature of the oil, the metal and surface, the
among other factors, by rheological and thermal consid- surrounding atmosphere, as well as the operating condi-
erations (52-56). Probably the most generalized treat- tions. And if there is a generalized scoring criterion, the
ment of EHD friction behavior is that given by Archard concensus is that it is thermal in character; i.e., it is the
and Baglin (57, 58); but the work suffers from the same consequence of the intense frictional heat generation at
objections as those discussed previously under EHD film the potential failure site.
thickness behavior. Moreover, in dealing with EHD fric- Of the various thermal scoring models that have been
tion behavior, the effects of lubricant rheology and starva- examined, the most famous is no doubt the critical tem-
tion are expected to be even more marked. perature model (16). Other principal models include the
The friction behavior in boundary lubrication is even critical power intensity model (66) and the critical power
more complex, because it is so strongly influenced by the model (64). These models have been compared (64, 68);
oil-metal-atmosphere interaction (46-50, 59-64). Ku but the comparisons are basically inconclusive. In short,
and associates have found that the friction coefficient of there are no lack of data or arguments either to support
 or to refute any one of these models in some way. How- film thickness to composite surface roughness ratio and
ever, from the point of view of generality; i.e., the poten- oil type have been reported also (17); it is shown that
tial of the model to account for the greatest number of the pitting life is not merely a function of the ratio but
design, material, lubricant, and operating variables, it also of oil composition.
appears, at this time, that the critical temperature model Published work on pitting in gears has apparently
is the most promising, provided suitable refinements are been limited to a study of the effect of lubricants (78)
added to Blok's original concept. Some of these refine- and another on the effect of materials (79). The former
ments include the recognition that both the critical tem- shows that with oils of different chemical types, oil vis-
perature and the friction coefficient are not merely func- cosity (or oil film thickness to composite surface rough-
tions of the oil-metal-atmosphere combination; but ness ratio) is not a unique measure of pitting life. The
depend also upon the operating conditions and surface latter shows, among other things, that the pitting life is
details. influenced by load sharing and tooth profile modifica-
Clearly, the usefulness of any predictive model, no tion. These two papers are concerned with gears with
matter how simple or elaborate, can only be established a contact ratio of less than two. For gears with higher
by applying it to actual performance results. Without contact ratios, the load sharing pattern is quite different
Downloaded by [The University of Manchester Library] at 23:01 15 October 2014

going into details, it can be said that the correlation of and the maximum Hertz stress may very well occur in
predicted scoring-limited performance of sliding-rolling the mesh cycle where the sliding velocity may be high
disks with disk test data has been reasonably satisfactory (26). Accordingly, the effect of sliding friction on pitting,
to date (64). However, application of the same technique under EHD and particularly boundary lubrication con-
to correlate with the scoring-limited performance of gears ditions, cannot be overlooked.
has not met with similar success. It is the author's con- Possibly the most interesting work, now in process of
clusion that the chief difference between disk tests and publication, is a comparative experimental study per-
gear tests is that the disk operation is steady-state while formed by Onion and Archard on disk and gear pitting
the gear operation is not. This problem has been dis- (80). They have found, among other things, a surface
cussed at some length earlier. Analysis of several sets of texture effect on disk pitting rather similar to that on
aircraft-quality gear test data available appears to sug- disk scoring mentioned earlier. Just as significantly, they
gest that the major problems are gear misalignment and have found that the gear pitting life is only about I/l,,,th
dynamic tooth load, with the result that the actual scor- of the disk pitting life under equivalent operating condi-
ing-limited load-carrying capacity of gears may be very tions. These authors believe that misalignment was not
much lower than the equivalent disk-test performance. a significant problem in their tests, and, accordingly,
The "correction" for each effect that is required to obtain conclude that the difference was due principally to the
reasonable disk-vs-gear scoring performance correlation dynamic tooth load experienced in gear operation. If
may, under unfavorable circumstances, amount to as their conclusion is accepted, it would mean that the
much as twenty-five percent or more of the disk-test dynamic tooth load would be about 4.5 times the quasi-
load-carrying capacity. static tooth load! This value appears to be inordinately
high unless the gear-tooth error is excessive. O n the other
hand, gear misalignment is an elusive matter as said
before; even a slight misalignment can conceivably make
PITTING
a great deal of difference.
Although pitting in nearly pure-rolling systems, such
as in rolling-element bearings, has received a great deal
of attention, pitting in rolling-sliding systems, such as in
RUBBING WEAR
gears, has so far been largely overlooked. This latter
oversight may be due to two reasons. Firstly, as men- Gear wear has not commanded much attention to
tioned before, scoring has an overriding influence on date, mainly because it takes time to become damaging
maximum gear performance. Secondly, with gears of low and thus presents a less urgent problem than scoring.
contact ratios, pitting usually occurs near the pitchline, On the other hand, in applications where scoring has
where the Hertz stress is maximum and the motion is been avoided, the gear life becomes a matter of competi-
nearly pure rolling. These are, of course, gross simplifica- tion between pitting and wear. In other words, for long-
tions of the problem. life applications, wear is certainly a failure mode to
Consider first the case of nearly pure rolling. A popular contend with, just as is pitting.
view appears to be that the pitting life is largely a func- Wear may take place under rubbing conditions or
tion of the oil film thickness to composite surface rough- nonrubbing conditions. The present discussion will be
ness ratio (69-71). However, even in pure rolling, there concerned with rubbing wear, the more common type
is enough evidence to show that, with oils of different in gear operation. As said before, rubbing wear cannot
chemical compositions, the chemical effect on pitting life occur without entering into the boundary lubrication
cannot be ignored (17, 72-74). In addition, the effects regime. Hence, a safe but conservative design is to aim
of steel composition, metallurgical structure, etc., have for full EHD lubrication a t the mating gear-tooth sur-
also been observed (75-77). The combined effects of oil faces, or close to this condition. In practical gearing, such
 Gear Failure Modes-Importance of Lubrication and Mechanics

a conservative approach is seldom necessary, even when CONCLUSIONS

operating with a straight mineral oil, with reasonable
This paper examines, in broad terms, the current state
choice of gear material. If oils possessing EP quality are
of understanding of gear lubrication, with emphasis on
used, considerable penetration into the boundary lubri-
the impact of the principal lubrication-related modes of
cation regime may be possible w'thout excessive wear.
d
In these latter instances, the boun ary lubrication inter-
action at the rubbing site becomes the dominant consid-
gear-tooth failure on gear performance. The implications
of elastohydrodynamic lubrication and boundary lubri-
cation on these failure modes are considered. The conse-
-~ ~ ~

eration.
quences of the transient nature of gear-tooth action are
The subject of rubbing wear under steady-state sliding
discussed. In particular, the overwhelming influence of
conditions has, of course, been extensively studied
gear mechanics on gear lubrication is emphasized.
(6, 14, 15,46-50). Howiver, the literature contains little
dealing with rubbing wear under steady-state sliding-
rolling conditions, or under the transient conditions of
gear-tooth action. If it can be a sumed that the knowl- REFERENCES
k
edge of wear gained from stea y-state sliding can be
Downloaded by [The University of Manchester Library] at 23:01 15 October 2014

( I ) Gear Lubrication Symposium, ASLE, Chicago, Ill., Jan. 1959;

applied to the transient state of gear-tooth action, then papers published in Lub. Engrg., 15, and A S L E Trans., 2, (1959).
a tentative approach to rubbing wear in gears becomes ( 2 ) Elastohydrodynamic Lubrication, Proc. IMechE, 180, Pt. 3B, (1965-66).
possible. For example, Rowe's model for lubricated wear ( 3 ) Gear Lubrication, Inst. Petroleum, (1966).
(4) Conference on Lubrication and Wear, Proc. IMechE, 182, Pt. 3A,
(81, 82), which is an extension of Archard's model for (1967-68).
unlubricated wear (15), is, though rather involved, quite (5) Proc. International Conference on Gearing, IMechE, (1970).
amenable to such a treatment. To this author's knowl- ( 6 ) Ku, P. M., editor, Interdisciplinary Approach to the Lubrication of
Concentrated Contacts, N A S A Spec. Publ. 237, (1970).
edge, Rowe's wear model has not been tried on gears. ( 7 ) Ku, P. M., editor, Interdisciplinary Approach to Liquid Lubricant
However, a simpler, empirical wear equation has been Technology, N A S A Spec. Publ. 318, (1972).
applied to compute the tooth wear profile of a spur gear (8) Elastohydrodynamic Lubrication, 1972 Symposium, IMechE, (1972).
( 9 ) International Symposium on Gearing and Transmissions,
by taking into account the variations of sliding velocity IFToMM-ASME-AGMA, San Francisco, Calif., Oct. 1972;
and tooth load through the mesh cycle (83). The tooth papers published in Mechanism 6' Machine Theory, 8, (1973) and
A S M E Tram., Jour. Engrg. for Indurlry, 95B, (1973) and 96B, (1974).
wear profile so computed has been found to be qualita- (10) Gear Scuffing-State of Calculation Methods, FZG Colloquium,
tively quite similar to that commonly observed, although Munich, Germany, April 1973.
refinements appear possible. (11) Guichelaar, P. J., Levy, B. S., and Parikh, N. M., editors, Gear
Of course, the real test of any gear wear analysis lies Manufacture and Performance, Am. Soc. of Metals, 1973.
(12) First Power Transmission and Gearing Conference, ASME, New
in quantitative correlation with gear test results. This has York, N. Y., Oct. 1974.
not yet been done, at least to the author's knowledge. (13) "Nomenclature of Gear-Tooth Wear and Failure," A G M A Standard
It is likely that such an effort will show, as in the in- 110.03 1962.
(14) Rabinowicz, E., Friction and Wear ofMaterials, John Wiley & Sons,
stances of scoring and pitting, that the dynamic tooth New York, 1965.
load and misalignment play very important roles. (15) Archard, J . F., "Wear," Interdisciplinary Approach to Friction and
Wear, N A S A Spec. Publ. 181, 267-334, (1968).
(16) Blok, H . , "Les temperature des surface, dans des condition de
graissage sous pression extreme," Congr. mondial pCtrok, 2 me Congr.,
3, (1937).
HEAT TRANSFER (17) Ku, P. M., "Tribology of Gears and Splines," Proc. Tribology
Workshop, NSF, 402-427, (1974).
It must be clear that the ultimate performance of a (18) Littmann, W. E., "The Mechanism of Contact Fatigue," Inter-
lubricated system is limited in many respects by cooling. disciplinary Approach to the Lubrication of Concentrated Contacts, N A S A
T h e problem consists of two parts, frictional heat genera- Spec. Publ. 237, 309-368, (1970).
(19) Landen, E. W., "Slow-Speed Wear of Steel Surfaces Lubricated
tion at the rubbing sites, and heat transfer from these by Thin Oil Films," A S L E Trans., 11, 6-18, (1968).
points ultimately to the surrounding atmosphere. The (20) "Rating the Strength of Spur Gear Teeth," A G M A Standard
difficult problem of handling friction in sliding-rolling 220.02, 1966.
(21) Buckingham, E., Analytical Mechanics of Gears, McGraw-Hill Book
contacts, and particularly in a gear mesh, has already Co., 1949.
been alluded to. As to the heat transfer problem, little (22) Dyson, A,, General Theory of Kinematics and Geome~ryof Gears in Three
has been done except for the recent work on the fling-oE Dimcmiom, Oxford Univ. Press, 1969.
(23) Walker, H . , "Gear Tooth Deflection and Profile Modification, I,"
oil cooling of the gear teeth by Blok and associates The Engineer, 166, 409-412, (1938).
(84, 85). (24) Walker, H . , "Gear Tooth Deflection and Profile Modification, 11,"
The overall problem of heat transfer in a complete The Engineer, 166, 434-436, (1938).
(25) Walker, H., "Gear Tooth Deflection and Profile Modification,
gear transmission has been described by Blok (86), and 111," The Engineer, 170. 102-104, (1940).
also by Bathgate and associates (87), by treating the (26) Staph, H . E., "A Parametric Analysis of High-Contact-Ratio
system as a network of thermal potentials and imped- Spur Gears," ASLE Annual Meeting, Atlanta, Ga.,
May 5-9, 1975.
ances. Unfortunately, rigorous analysis of this character (27) Kelley, B. W. and Lemanski, A. J., "Lubrication of Involute
requires a detailed knowledge of the heat transfer proc- Gearing," Proc. IMechE, 182, Pt. 3A, 173-184, (1967-68).
esses from various rotating machine elements, such as (28) Reswick, J. B., "Dynamic Loads in Spur and Helical Gear
Teeth," Trans. ASME, 77, 635-644, (1955).
bearings and gears, and the associated coefficients. Such (29) Harris, S. L., "Dynamic Loads on the Teeth of Spur Gears," Proc.
information is currently not available. IMechE, 172, 87-100, (1958).
 (30) Tuplin, W. A., "Dynamic Loads on Gear Teeth," F'roc. Int. Con$ (59) Misharin, J. A,, "Influence of the Friction Conditions on the
on Gearing, IMechE, 18-23, (1958). Magnitude of the Friction Coefficient in the Case of Rolling with
(31) Seireg, A. and Houser, D., "Evaluation of Dynamic Factors for Sliding," Proc. Int. ConJ on Gearing, IMechE, 159-164, (1958).
Spur and Helical Gears," Trans. ASME, Jour. Eng. for Industry, 92B, (60) Benedict, G. H. and Kelley, B. W., "Instantaneous Coefficients
504-515, (1970). of Gear Tooth Friction," ASLE Trans., 4, 59-70, (1961).
(32) Akin, L. S., "EHD Film Thickness Formulaes for Power Trans- (61) Bell, J. C. and Dyson, A,, "Mixed Friction in an Elastohydro-
mission Gears," Trans. ASME, Jour. Lub. Tech., 96F, dynamic System," Elastohydrodynamic Lubrication, 1972 Symposium,
426-431, (1974). IMechE, 68-76, (1972).
(33) Dowson, D. and Higginson, G. R., Elastohydrodynamic Lubrication, (62) Staph, H. E., Ku, P. M., and Carper, H. J., "Effect of Surface
Pergamon Press, 1966. Roughness and Surface Texture on Scuffing," Mechanism and
(34) Dowson, D., "Elastohydrodynamic Lubrication," Interdisciplinary Machine Theory, 8, 197-208, (1973).
Approach to the Lubrication of Concentrated Contacts, NASA Spec. (63) Carper, H. J., Ku, P. M., and Anderson, E. L., "Effect of Some
Publ. 237, 27-76, (1970). Material and Operating Variables on Scuffing," Mechanism and
(35) Sanborn, D. M. and Winer, W. O., "Fluid Rheological Effects Machine Theory, 8, 209-232, (1973).
in Sliding Elastohydrodynamic Point Contacts With Transient (64) Carper, H. J. and Ku, P. M., "Thermal and Scuffing Behavior
Loading: I-Film Thickness," Trans. ASME, Jour. Lub. Tech., 93F, of Disks in Sliding-Rolling Contact," ASLE Tram., 18,
262-271, (1971). 39-47, (1975).
(36) Castle, P, and Dowson, D., "A Theoretical Analysis of the Starved (65) Leach, E. F. and Kelley, B. W., "Temperature-The Key to
Elastohydrodynamic Lubrication Problem for Cylinders in line Lubricant Capacity," ASLE Tram., 8, 271-285, (1965).
Downloaded by [The University of Manchester Library] at 23:01 15 October 2014

Con tact," Elastohydrodynamic Lubrication, 1972 Symposium, IMechE, (66) Matveevsky, R. M., "The Critical Temperature of Oil with Point
131-137, (1972). and Line Contact Machines," Tram. ASME, Jour. Basic Ens., 87D,
(37) Cheng, H. S., "Calculation of Elastohydrodynamic Film Thick- 754-760, (1965).
ness in High-Speed Rolling and Sliding Contacts," M T I Rept. (67) Bell, J. C. and Dyson, A,, "The Effect of Some Operating Factors
67-24, 1967. on the Scuffing of Hardened Steel Discs," Elastohydrodynamic Lubri-
(38) Dyson, A. and Wilson, A. R., "Film Thicknesses in Elastohydro- cation, 1972 Symposium, IMechE, 61-67, (1972).
dynamic Lubrication at High Slide/Roll Ratios," Proc. IMechE, (68) Bell, J. C., Dyson, A,, and Hadley, J. W., "The Effects of Rolling
183, Pt. 3P, 81-97, (1968-69). and Sliding Speeds on the Scuffing of Lubricated Steel Discs,"
(39) Turchina, V., Sanborn, D. M., and Winer, W. O., "Temperature ASLE Trans., 18, 62-73, (1975).
Measurements in Sliding Elastohydrodynamic Point Contacts," (69) Dawson, P. H., "The Pitting of Lubricated Gear Teeth and
Trans. ASME, Jour. Lub. Tech., 96F, 464-471, (1974). Rollers," Power Transmission, 30, 208-27 1, ( 1961).
(40) Cheng, H. S., "A Numerical Solution of the Elastohydrodynamic (70) Dawson, P. H., "Effect of Metallic Contact on the Pitting of
Film Thickness in an Elliptic Contact," Trans. ASME, Jour. Lub. Lubricated Rolling Surfaces," Jour. Mech. Ens. Sci., 4,
Tech., 92F, 155-162, (1970). 26-31, (1962).
(41) Fowles, P. E., "The Application of Elastohydrodynamic Film (71) Dawson, P. H., "Further Experiments on the Effect of Metallic
Thickness to Individual Asperity-Asperity Contacts," Trans. Contact on the Pitting of Lubricated Rolling Surfaces," Proc.
ASME, Jour. Lub. Tech., 91F, 464-476, (1969). IMechE, 180, Pt. 3B, 95-100, (1965-6).
(42) Johnson, K. L., Greenwood, J. A,, and Poon, S. Y., "A Simple (72) Rounds, F. G., "Effects of Base Oil Viscosity and Type on Bearing
Theory of Asperity Contact in Elastohydrodynamic Lubrication," Ball Fatigue," ASLE Trans., 5, 172-182, (1962).
Wear, 19, 91-108, (1972). (73) Rounds, F. G., "Some Effects of Additives on Rolling Contact
(43) Archard, J. F., "Elastohydrodynamic Lubrication of Real Sur- Fatigue," ASLE Trans., 10, 243-255, (1967).
faces," Tribology, 6. 8-14, (1973). (74) Appeldoorn, J. K., "Effect of Lubricants-General Background,"
(44) Naylor, H., "The Rheological ~ e h a v i o rof Lubricants," Inter- Interdisciplinary Approach to the Lubrication of Concentrated Contacts,
disciplinary Approach to the Lubrication of Concentrated Contacts, NASA NASA SP-237, 439-462, (1970).
Spec. Publ. 237, 279-308, (1970). (75) Zaretsky, E. V. and Anderson, W. J., "Effect of Materials-
(45) Cheng, H. S., "Rheology in Lubrication," Interdisciplinary Approach General Background," Interdisciplinary Approach to the Lubrication of
to Liquid Lubricant Technology, NASA Spec. Publ. 318, 263-314, (1973). Concentrated Contacts, NASA SP-237, 379-408, (1970).
(46) Bowden, F. P. and Tabor, D., The Friction and Lubrication of Solids, (76) Bamberger, E. N., "Effect of Materials-Metallurgy Viewpoint,"
Pt. I, Oxford Univ. Press, 1954. Interdisciplinary Approach to the Lubrication of Concentrated Contacts,"
(47) Bowden, F. P. and Tabor, D., The Friction and Lubrication of Solids, NASA SP-237, 409-462, (1970).
Pt. 11, Oxford Univ. Press, 1964. (77) Rounds, F. G., "Influence of Steel Composition on Additive
(48) Ku, P. M., editor, Interdisciplinary Approach to Friction and Wear, Performance," ASLE Tram., 15, 54-66, (1972).
NASA Spec. Publ. 181, 1968. (78) Davidson, T. F. and Ku, P. M., "The Effect of Lubricants on
(49) Ling, F. F., Klaus, E. E., and Fein, R. S., editors, Boundary Lubrica- Gear Tooth Surface Fatigue," ASLE Trans., 1, 40-50, (1958).
tion-An Appraisal of World Literature, ASME, 1969. (79) Townsend, D. P. and Zaretsky, E. V., "A Life Study of AISI M50
(50) Ling, F. F., Surface Mechanics, John Wiley & Sons, 1973. and Super Nitralloy Spur Gears with and without Tip Relief,"
(51) Blok, H., "Lubrication as a Gear Design Factor," Proc. Int. Con$ Trans. ASME, Jour. Lub. Tech., 96F, 583-590, (1974).
on Gearing, IMechE, 144-158, (1958). (80) Onions, R. A,, and Archard, J. F , "The Pitting of Gears and
(52) Jefferis, J. A. and Johnson, K. L., "Sliding Friction between Discs," to be published.
Lubricated Rollers," Proc. IMechE, 182, Pt. 1, 281-291, (1967-68). (81) Rowe, C. N., "Some Aspects of the Heat Adsorption in the
(53) Dowson, D. and Whomes, T. L., "Effect of Surface Quality upon Function of a Boundary Lubricant," ASLE Tram., 9,
the Traction Characteristics of Lubricated Cylindrical Contacts," 101-1 11, (1966).
h c . IMechE, 182, Pt. 1, 292-299, (1967-68). (82) Rowe, C. N., "Role of Additive Adsorption in the Mitigation of
(54) Plint, M. A., "Traction in Elastohydrodynamic Contacts," Proc. Wear," ASLE Trans., 13, 179-188, (1970).
IMechE, 182, Pt. 1, 300-306, (1967-68). (83) Thompson, R. A. and Weichbrodt, B., "Gear Diagnostics and
(55) Johnson, K. L, and Cameron, R., "Shear Behavior of Elasto- Wear Detection," A S M E Preprint 69-VIBR-10, (1969).
hydrodynamic Oil Films at High Rolling Contact Pressures," (84) Dewinter, A,, and Blok, H., "Fling-Off Cooling of Gear Teeth,"
Proc. IMcchE, 182, Pt. 1 , 307-319, (1967-68). Trans. ASME, Jour. Eng. for Industry, 96B, 60-70, (1974).
(56) Dyson, A,, "Frictional Traction and Lubricant Rheology in Elas- (85) Heijnigen, G. J. J., and Blok, H., "Continuous As Against Inter-
tohydrodynamic Lubrication," Phil. Tram., Roy. Soc., 266A, mittent Fling-Off Cooling," Tram. ASME, Jour. Lub. Tech., 96F,
1-33, (1970). 529-538, (1974).
(57) Archard, J. F. and Baglin, K. P., "Non-Dimensional Presentation (86) Blok, H., "The Postulate About the Constancy of Scoring Tem-
of Friction Tractions in Elastohydrodynamic Lubrication: perature," Interdisciplinary Approach to the Lubrication of Concentrated
Part I, Fully Flooded Conditions," A S M E Preprint 75-LUB-31, Contacts, NASA Spec. Publ. 237, 153-248, (1970).
(1974). (87) Bathgate, J., Kendall, R. B., and Moorehouse, P., "Thermal
(58) Archard, J. F., and Baglin, K. P., "Non-Dimensional Presentation Aspects of Gear Lubrication," Wear, 15, 117-129, (1970).
of Friction Tractions in Elastohydrodynamic Lubrication: Part 11,
Starved Conditions," ASME Preprint 74-LUB-32, (1974).
 Gear Failure Modes-Importance of Lubrication and Mechanics

DISCUSSION therefore, support Dr. Ku in his supposition that the

transient loading of the oil film may be as important as
F. 1. BARWELL
the dynamic loading of the tooth flank.
University College of Swansea As regards misalignment, it is obviously more
Swansea, United Kingdom difficult to ensure the proper meshing of a number of
One is indebted to Dr. Ku for his masterly survey of flanks rather than just two cylinders. Perhaps the
the problems of gear lubrication mechanics and, in contact quality is important here. It emphasizes the
particular, for his conclusion that the system must be effect of surface texture and the mutual interaction of
treated as a whole because of the complex interactions the surfaces of the two interacting pairs over a period ol
between geometry, mechanics, and tribological factors. time, again this may be an important factor dis-
A particularly complex area relates to the definition tinguishing between disks and gear tests.
of the lubrication mode which may be operative in a T h e interaction of asperities, therefore, seems to
given set of circumstances. For many years, gears were assume unexpected importance and there would appear
regarded as examples of boundary lubrication because to be the need for much experimentation using actual
Downloaded by [The University of Manchester Library] at 23:01 15 October 2014

calculations of hydrodynamic film thickness following gears rather than disks and to try and quantify the
the work of Martin indicated values which were orders quality of the surface of the film separating them. T h e
of magnitude below the known asperity heights. writer's view is that the technique used by Campbell, et
However, practical experience indicated that there was a1 and by Furey is the easiest and most applicable
very effective lubrication indeed. For example, in the method of progress.
"Queen Mary" the original machining marks could be The breakdown of the elastohydrodynamic film is
seen on the gear flanks after many years o f intensive not necessarily the cause of scoring failure. However,
transatlantic service. The introduction of the concepts there does appear here to be evidence that the onset of
underlying elastohydrodynamic lubrication appeared pitting is closely related to the D-ratio; i.e., the ratio of
to resolve thisdifficulty a n d a great deal of investigation film thickness to measured surface roughness. It is the
was carried out on over-simplified systems, notably that writer's view that some form of film must be generated
involving the interaction of disks. Some of us have in order to keep frictional heating within bounds, but
always regarded the direct translation of the steady-state that extreme pressure lubricants can, by removal of
situation as between disks to the entirely transient offending asperity contacts, provide for operation at
situation of the gear as being somewhat facile and, as films which are very thin indeed. Regarding boundary
indicated by Dr. Ku, recent work by Onions and lubrication, it is possible that a special surface texture
Archard confirms that disk tests can grossly results from working in the presence of chemically
overestimate the pitting life of gears. reactive substances so that the surface ceases to be clear,
T h e discussant is referring to Ref. (80)which has now bright metal, but becomes a mixture of oxide,
been published by the Institution of Mechanical phosphide, sulphide, oil, and an occasional metallic
Engineers. Although Dr. Ku, no doubt rightly, flake. T h e lubricant may cease to be a simple
considers these results to be rather tentative, as indeed Newtonian fluid and, in immediate proximity to the
do the authors themselves, this does not mean that they surface, may become a product of reaction such as a
are unimportant, and it is hoped that further work will viscous soap with non-Newtonian properties. This
be carried on in this aspect of the subject. Onions and acting o n the spongey weakened asperities may allow
Archard first mention that the differences may arise the surfaces to interact with onlymild damage.
from the dynamic loads which are present in gears but With regard to the incidence of scoring(in the U.K. it
absent from disks. This is an important factor, but it is still persistently called "scuffing"), Dr. Ku is right in
might well be expected that a well-developed emphasizing the dynamic nature of the system. Some
elastohydrodynamic film might have a cushioning years ago, A. A. Milne and the discussant put forward
effect. the dynamic hypothesis, namely that, as long as the
Dr. Ku mentions the importance of lubricant situation was such that heat generated in the bearing
starvation. T h e method whereby lubricant is injected system could be removed therefrom by conduction or
between the two gear flanks seems to this discussant to convection, or even by radiation, a stable operable
be very capricious and simple starvation might very system resulted. But where the energy consumed
well be the explanation of Onions and Archard's exceeded that dissipated, as would occur with a partial
observations. Moreover, a certain amount of time may breakdown in lubrication, leading to an increase in the
be necessary to build a film. Here one would expect no value of the surface friction, the frictional heat and
correlation with disks insofar that a portion of dissipated' energy would increase at a faster rate than
lubricant first entering the Hertzian stress area of adisk they could be conducted away, thereby leading
system would be subject to steady forces, but, with the inevitably to the attainment of temperatures which
gear, there is a continual change, the time of meshing would result in damage.
being exceedingly short. T h e discussant would, What seemed like an instantaneous phenomena
 could easily be the result of an exponential growth; i.e., "the attainment of an intact EHD film between the
as the temperature rose so viscosity fell off, so the film surfaces of mating gear teeth is a desirable design goal,
became weaker, friction loads grew higher with a as it precludes rubbing wear and scuffing, and
higher rate of temperature rise, leading to higher rate of minimizes pitting danger." In applications where very
heat generation, so that, once a certain point had long life is the only requirement, full EHD lubrication
passed, failure became inevitable, although perhaps is certainly a legitimate design objective. But even here,
delayed for a short time. Some support for this is noted the designer cannot escape from the difficulties of
in connection with some cross cylinder experiments selecting the desired film thickness ratio to use,
where the zone of contact moved on to a fresh material calculating the minimum EHD film thickness in the
continuously; there was evidence sometimes of scoring actual gear mesh, and predicting the gear surface
being initiated but being extinguished after a short roughness in service-all of which involve considerable
path has been traversed. As conditions were worsened, uncertainties but have a strong impact on design. Of
so did the duration of the scores increase until they course, the designer's ,task is much eased if the oil
became continuous. Therefore, an important function selection is not dictated by other considerations. On the
of EP additives appears to be the preservation of a other hand, if the working oil viscosity must be kept
Downloaded by [The University of Manchester Library] at 23:01 15 October 2014

surface texture conducive to the maintenance of within bounds, then the designer's only option is to
elastohydrodynamic conditions. adjust the gear size, which affects the system size,
The theory that pitting type failure originates from weight, and cost. And, in view of the uncertainties
the zone of maximum shear stress within the material is mentioned above, overdesign is an ever-present tempta-
difficult to reconcile with the demonstrable effect of tion.
surface finish in relation to film thickness. Mudd ( A I ) Before leaving the subject of EHD lubrication, it is
indicated that there appeared to be some strengthening perhaps well to emphasize again that these "remarks
factor related more to direct stress than to shear. He also are not intended to minimize the important con-
found in his examples that failure was more likely to tributions of the EHD theory" in general. However, one
occur at the surface than beneath the case of case- does not design gear systems without giving at least
hardened gears. This has important implications some consideration to size, weight, and cost. It is in this
regarding the specification of case depth for gears (A2). contest that, unless the choice of gear material and oil is
The discussion this morning emphasizes the great very unfortunate, "a design based on theassumption of
difficulty in studying a real system, such as a set of full EHD lubrication of gear teeth is not only
gears, with sufficient detail to be able to understand the unnecessary . . . . , but in fact too conservative
tribological reactions at the gear flanks involving ...... 9,

surface finish studies of great detail and precision. The Onions and Archard investigation is, indeed,
Perhaps the role of the disk machine has been highly significant in showing that at the same
overemphasized and the problem now is to find ways of computed film thickness ratio, the gear pitting life was
understanding in basic physical terms what goes on in only about 1/100th of the disk pitting life. As Professor
real gears. The electrical conductivity test appears to be Banvell and the authors themselves have pointed out, a
an ideal way of doing this and some promise is offered multitude of factors could collectively account for this
by new techniques for the collection and study of debris difference. This author's only point was that, while the
within the lubricant. authors may very well be right in singling out dynamic
REFERENCES load as the principal culprit, he wonders if misalign-
ment might not have been an unwitting partner. Since
( A l ) Mudd, G. C., "A Numerical Means of Predicting the Fatique disks and gears are elastic (a fact fundamental to the
Performances of Nitride-Hardened Gears," Proc. I. Mech. E.,
184, Pt. 3.D (1970). EHD theory), a seemingly negligible amount of
(A2) Young, I. T., "Gearing for Modern Marine Propulsion misalignment can change the normal stress distribu-
Systems," G. E. C. Journal of Science6 Technology, 41,143-155 tion across the disk or gear face very substantially, but
( 1974).
still shows a "full wear" across the face. Moreover, the
location of the maximum normal stress (i.e., the
AUTHOR'S CLOSURE location of the scoring or pitting site) need not be at the
Professor Barwell speaks with rare authority from edge of the disk or gear face, because this location
vast experience. Therefore, the fact that he endorses the depends not merely on the misalignment as such, but
author's views in most major respects is both reassuring also on any slight curvature or irregularities that many
and gratifying. In the interests of space, the author shall be present across the face. For the sake of argument, if
not dwell on that on which there is agreement; but will one assumes that the gears were misaligned only to such
only comment on those few points on which there an extent as to cause the local maximum normal stress
appears to be difference in relative emphasis, though to be 1.25 times that calculated for perfect alignment,
perhaps not in substance. then the pitting life could be reduced by a factor of
There is no disagreement with regard to the desirable about (1.25)1°= 9.3. his then leaves the dynamic load to
mode of gear lubrication. It is stated in the paper that account for a factor of 100/9.3 = 10.8, yielding a ratio of
 Gear Failure Modes-Importance of Lubrication and Mechanics 249
1
dynamic load to quasi-static load of about (10.8)3= 2.2, temperature and the surface roughness of the cone; and
which appears to be a fairly plausible figure unlessvery these changes were found to produce essentially the
unfavorable dynamic conditions were encountered. same life vs film thickness ratio relationship. More
Professor Barwell's remarks on the effect of surface recent but unpublished work has also shown substan-
finish on pitting are pertinent and appreciated. The tially the same life (in stress cycles) vs film thickness
author reported on the pitting life behavior of four oils ratio behavior by changing the spindle speed. In-
as determined in ten Barwell-type testers, using 52100 terestingly, different oil types did not yield the same life
steel specimens in 4-ball and 3-ball/cone con- vs film thickness ratio curve; but the curves were
figurations, over a wide range of computed film significantly displaced from one another. A rationale
thickness ratio from 0.1 to 6 (17). For a given oil, the for the observed effect was offered by Beerbower in a
film thickness ratio was varied by changing the test discussion of Ref. (I 7).
Downloaded by [The University of Manchester Library] at 23:01 15 October 2014