UNCLASSIFIED/UNLIMITED

Understanding Scuffing and Micropitting of Gears

R W Snidle, H P Evans, M P Alanou, M J A Holmes

School of Engineering
Cardiff University
Cardiff CF24 0YF, UK
[email protected]

ABSTRACT
The paper describes the results of basic research, both theoretical and experimental, on scuffing and
micropitting of gears used in naval and aerospace applications. Scuffing experiments have been carried
out under severe conditions of load, sliding speed and temperature using typical gear steels with different
surface treatments including case-carburising, nitriding, superfinishing, and super-hard coatings. The
results of this work demonstrate the importance of surface finish not only in improving scuffing resistance,
but also in reducing contact friction and energy losses. Theoretical work on microelastohydrodynamic
lubrication shows the importance of surface roughness in terms of high pressures generated at surface
asperities on gear teeth and the effect this has upon the distribution of stress very close to the tooth
surface. Some new microelastohydrodynamic parameters, derived from this work, are of potential use in
discriminating between different types of gear finishes in terms of their tendency to generate extremes of
pressure and film thinning which are of direct relevance to micropitting.

1.0 INTRODUCTION
Gears transmit power through rolling/sliding contacts between their teeth, which must operate reliably for
many thousands of hours with negligible wear. Lubrication of tooth contacts, both as a means of
minimising surface damage and as a way of reducing friction due to sliding, is therefore of crucial
importance. Gears have been made from various materials, including polymers, but in this paper we are
concerned with steel gears which are the most important type in use in vehicles, aerospace, marine and
industrial applications. It is now known that elastohydrodynamic lubrication (EHL) is the mechanism
responsible for the protection of steel gear teeth. The basic kinematics of involute gears (by far the most
important type) necessitate a relatively low geometrical conformity between the mating teeth (Figure 1)
and this gives rise to contact pressures, which can be as high as 1.5 GPa if high strength surface-hardened
steels are used. Under these conditions the viscosity of typical lubricating oils is increased by several
orders of magnitude and significant “Hertzian” elastic deformation of the surfaces occurs. Sliding also
takes place between the teeth given by
u s = (ω 1 + ω 2 ) s

where s is the distance of the instantaneous contact between the teeth from the pitch point Z shown in
Figure 1 and ω1 and ω2 are the angular velocities of the two gears. The entraining velocity, which is
responsible for the formation of any oil film by hydrodynamic action, is given by

u = (rb1 tan ψ + s )ω 1 / 2 + (rb 2 tan ψ − s )ω 2 / 2

Both sliding and entraining velocities therefore vary throughout the meshing cycle as does the relative
contact geometry. A typical ratio of the sliding to entraining velocities at the extremes of tooth contact is

Paper presented at the RTO AVT Specialists’ Meeting on “The Control and Reduction of Wear in Military Platforms”,
held in Williamsburg, USA, 7-9 June 2003, and published in RTO-MP-AVT-109.

RTO-MP-AVT-109 14 - 1

UNCLASSIFIED/UNLIMITED
 UNCLASSIFIED/UNLIMITED

Understanding Scuffing and Micropitting of Gears

about 0.5, but can be as high as unity in very coarse gears. Contact loads are uncertain due to load-sharing
between simultaneous, overlapping tooth contacts, and dynamic effects at high speeds. A necessary
compromise in high speed gear design is the need to minimise sliding by using rather small teeth whilst
choosing teeth of sufficient size to resist the bending stresses caused by tooth loading. This explains the
use of rather fine pitch teeth combined with relatively large face-width in high speed naval reduction units.

ω2

ψ rb2
ω2
base circle 2

ω1 R2
P
R1 Z

ψ
rb1
base circle 1

ω1

Figure 1: Illustration of gear teeth in contact showing equivalent rolling/sliding circles. Sliding is
proportional to s = ZP, where Z is the pitch point and P is the instantaneous contact between the teeth.

Lubricant film thickness in involute gear teeth is usually assessed using EHL line contact theory [1] in the
form of Dowson and Higginson’s formula

hmin R = 1.6(α ) (η 0 u )0.7 (E ′)0.03 R −0.57 w′ −0.13

0.6

which has been experimentally validated for smooth surfaces, most notably by Dyson et al [2].
Modifications are available for situations where the underlying assumptions do not apply, e.g. to correct
for temperature rise in the inlet to the contact [3]. This theory is for infinitely long “line” contacts which is
the situation in the majority of both spur and helical involute gears. Gears that have “point” contacts such
as those in “Wildhaber-Novikov” [4] conformal gears, and angle-drive gears such as hypoids and spiral
bevels, require formulas for predicting film thickness under side-leakage conditions [5,6].

Under ideal conditions gear teeth operate with films which effectively separate the surfaces, and the above
EHL film formula may be used. However, under demanding conditions of low speed or high temperature,
the film predicted by classical EHL theory may be small compared with the roughness present on the
teeth. The gears then operate under conditions described as "partial", "mixed" or "micro" EHL. A
conventional measure of the severity of lubrication is provided by the lambda-ratio, which is defined as
the mean oil film thickness divided by the combined RMS roughness of the two surfaces, i.e.

Λ=h σ 12 + σ 12

Parameter Λ has been used as an indicator of relative life in rolling element bearings [7]. Most gears,
however, are much rougher than bearings and rarely have a finish better than about 0.4 µm Ra with

14 - 2 RTO-MP-AVT-109

UNCLASSIFIED/UNLIMITED
 UNCLASSIFIED/UNLIMITED

Understanding Scuffing and Micropitting of Gears

nominal oil films typically 1 µm at most. Clearly, many practical gears operate with Λ significantly less
than unity. In theoretical solutions of both the dry contact and micro-EHL problems significant roughness
leads to a severe rippling in the contact pressure distribution with maximum values far in excess of the
Hertzian values for smooth surfaces. Surface roughness effects are important in the gearing problems of
"micropitting" and "scuffing" [8] which are discussed in Sections 2 and 3.

This paper briefly reviews some current research on gear tooth contact lubrication that will lead to a better
understanding of the problems of micropitting, scuffing and gear and transmission lubrication in general.

2.0 MICROPITTING
Micropitting is pitting (rolling contact fatigue) on the scale of the roughness, as opposed to classical
pitting which occurs on the scale of the nominal Hertzian, or macro, contact. With increased use of
hardened gears it has become of widespread concern to the gearing industry. Although it can affect all
types of gears, it has become particularly troublesome in heavily loaded gears with hardened teeth. The
problem has been extensively studied by researchers [9,10,11,12] and is well known to gear designers
[13]. Micropitting is characterised by the presence of fine surface pits and the occurrence of local plastic
deformation and shallow surface cracks. It produces significant wear of the surfaces causing loss of
profile of the teeth, leading to noise. Serious cases can precipitate scuffing and even complete fracture of
the tooth. The problem is widespread and its solution has a high industrial priority. Figure 2 shows a gear
tooth with characteristic micropitting damage.

Figure 2: Gear tooth showing micropitting damage in

the root region. (Design Unit, Newcastle University)

The causes of micropitting are not fully understood, but surface roughness and high contact stresses at
asperity or micro contacts within the overall Hertzian contact are key features, so that a thorough
understanding of the stresses at asperity contacts is clearly needed. Useful insights into the behaviour of
the contact of real surfaces can be obtained using dry elastic contact simulations [14,15] or elastic/plastic
contact simulations [16]. However, it is necessary to consider the micro-EHL effects of the oil film since
these influence the contact pressures and, through non-Newtonian and time-dependent behaviour,
determine the traction forces at the micro contacts. The important sub-surface stress field is therefore
affected by the presence of a micro EHL film. In general when roughness is present the maximum
sub-surface shear stress field occurs much closer to the surface as seen in the micro EHL simulations in
section 4.

The mechanism of plastic ratchetting brought about by asperity pummelling or cyclic stressing at the
edges of micro contacts described by Kapoor and Johnson [17] may also be a factor in micropitting, and
work directed towards an understanding of failure of railway wheel contacts undergoing plastic

RTO-MP-AVT-109 14 - 3

UNCLASSIFIED/UNLIMITED
 UNCLASSIFIED/UNLIMITED

Understanding Scuffing and Micropitting of Gears

deformation may also be relevant [18]. Cavitation occurring within the nominal EHL contact due to
asperity action discussed in Section 4 may also be a factor contributing to micropitting.

Recent fatigue testing of gears in the ground and superfinished state [19] shows the benefit to be obtained
in terms of pitting life by superfinishing. Figure 3 is a Weibull plot comparing performance of hardened,
ground steel gears (mean Ra = 0.38 µm) with superfinished gears (mean Ra = 0.07 µm) of the same
material. The lives of the gears with superfinished teeth in this programme were found to be about four
times greater than those of the ground gears.

3 tests
99 ground
90 superfinished 7 tests
% failures

50

20
10
5

2
1
2 5 10 20 50 100 300
Million Stress Cycles

Figure 3: Results of micropitting fatigue tests on ground and superfinished gears.

3.0 SCUFFING
Scuffing is a serious form of surface failure that can affect gears running at high speeds or at high
temperatures. Although there is no accepted explanation for the mechanism of scuffing, in many
situations it is linked to physical failure of the EHL mechanism. To avoid scuffing problems gear
designers often specify chemically active extreme pressure (EP) additives in the lubricating oil, some form
of surface treatment for the gears such as case-carburising/hardening, nitriding or coatings, or a
combination of these. Two particularly promising methods for upgrading the scuffing resistance of steel
gears are: improvement of surface finish; and use of super hard coatings, e.g. diamond-like carbon (DLC)
coatings. Polishing hardened and ground steel discs can produce a significant improvement in scuffing
performance [20]. Superfinished discs used in scuffing tests showed a two fold load limit increase in
comparison with ground disks, with the additional benefit of significantly lower friction and bulk metal
temperatures. It should be emphasised that the ground and superfinished discs were of the same steel
which was case-carburised and hardened to the same specification for both types of finish. The quality of
the polishing process used to achieve these results is illustrated in Figure 4, and Figure 5 shows collected
results from the series of tests. The lubricant in all tests reported in this paper was Mobiljet 2 at a feed
temperature of 100oC.

Thin hard coatings have been found to offer benefits in many non-gearing applications in terms of wear
resistance, chemical inertness and low coefficient of sliding friction. They are deposited using different
variations of the physical vapour deposition technique using magnetron sputtering [21]. The DLC types
have been selected for a number of industrial applications both for tooling and machine elements. Part of
this same family are the metal doped DLC coatings (Me-DLC), which appear to offer a better adherence
and greater toughness than bare DLC coatings and show particular promise as coatings for gears. A second

14 - 4 RTO-MP-AVT-109

UNCLASSIFIED/UNLIMITED
 UNCLASSIFIED/UNLIMITED

Understanding Scuffing and Micropitting of Gears

type of coating that has been considered is boron carbide (B4C) which is an attractive alternative to Me-
DLC, as outlined by Eckardt et al. [22], giving enhanced chemical and thermal stability combined with
extreme hardness. Moreover, the coating process requires a less reactive gas, which is advantageous. The
chemical structure of the coating is amorphous with a majority of B4C carbides. Some boron oxides as
well as boron oxy-carbides are also normally present with about 15-18 % oxygen located mostly at the
surface [23].

Figure 4: A superfinished steel test disc showing high quality finish achieved.

6 ground superfinished
Scuffing Load (kN)

0
0 10 20 30
Sliding Speed (m/s)

Figure 5: Results of scuffing tests on ground and superfinished discs.

Me-DLC and B4C coatings, whose details are given in Table 1, were used in scuffing tests carried out at
sliding speeds ranging from 7 m/s to 20 m/s. The tests demonstrate the potential benefits of these coatings
on both case-carburised and nitrided steels, although superfinishing achieved a similar performance in
terms of scuffing load capacity as may be seen in Figure 6 which ranks a wide range of different material
and finish combinations (see Table 2) which have been tested as part of an ongoing programme. The
triple combination of a nitrided steel substrate plus superfinishing and a diamond-like coating
(configuration 9) shows particularly impressive scuffing resistance. The figure also shows the relative
performance of the combinations in terms of average mean bulk temperature of the discs just prior to
scuffing and the average traction coefficient under the same conditions. Again, superfinishing appears to
confer similar benefits in terms of reduced temperature and traction as that obtained from super hard
coatings.

RTO-MP-AVT-109 14 - 5

UNCLASSIFIED/UNLIMITED
 UNCLASSIFIED/UNLIMITED

Understanding Scuffing and Micropitting of Gears

Table 1: Details of scuffing test coatings

Coating Thickness Elastic Modulus Hardness

Me-DLC 3.2 µm 121 GPa 11.4 GPa

B4C 2.7 µm 180 GPa 17 GPa

Table 2: Different substrate/finish/coating configurations tested

Configuration Substrate Finish Coating No of tests at each

sliding speed
7 16 20
m/s m/s m/s
1* case-carburised ground none (reference) 3 1 1
2** case-carburised superfinished none - 2 2
3 case-carburised ground Me-DLC 3 3 3
4 case-carburised ground B4C 2 4 3
5 nitrided ground, compound none 4 3 3
layer off
6 nitrided ground, compound none 4 3 3
layer on
7 nitrided superfinished none - 3 3
8 nitrided ground Me-DLC - 1 -
9 nitrided superfinished Me-DLC - 3 -
* one other test was performed at each of the sliding speeds of 10, 13, 18.5 m/s, plus two other
tests at 26 m/s.
** two other tests were performed at each of the sliding speeds of 10, 13, 23 m/s, plus three other
tests at 18.5 m/s.

Analysis of EHL between typical gear tooth surfaces in contact strongly suggests that eventual scuffing of
ground surfaces can be explained by the physical failure of the lubrication mechanism due to sideways
leakage of lubricant near the edges of nominal point contacts, and within nominal line contacts due to
waviness in the direction transverse to that of rolling/sliding. Figure 7 shows a scuffing scar on an
elliptical contact in a disc machine test. It is possible to envisage the leakage of oil from the loaded area in
a transverse direction along the relatively deep valleys between asperities produced by the grinding
process as a lubrication failure mechanism. It is interesting to note that for rough ground disc tests,
scuffing was initiated at the edges of the contact without exception, and not near the contact centreline
where pressures and “flash temperature” are highest. Roughness valleys are therefore seen as vulnerable
features of the surface when considering lubrication effects. This concept has been pursued theoretically
[24] in a simple model of EHL film failure in which a dry contact analysis was used to provide the
geometry of gaps (due to valley features of roughness) between lubricated surfaces under thin film
conditions. The magnitude of the transverse pressure gradients needed to cause significant sideways

14 - 6 RTO-MP-AVT-109

UNCLASSIFIED/UNLIMITED
 UNCLASSIFIED/UNLIMITED

Understanding Scuffing and Micropitting of Gears

leakage in these gaps, and hence collapse of the film, is consistent with that present at the edges of real
EHL contacts under similar operating conditions. These ideas are being explored more fully with the aid
of detailed micro EHL solutions that include these side leakage effects, as discussed in section 4.

(a) (b) 250

5000
4500
Average Scuffing Load (N)

4000 200

Average Mean Bulk

Temperature (°C)
3500
3000 150
2500
2000 100
1500
1000 50
500
0 0
6 3 1 5 4 7 2 8 9 6 3 1 5 4 7 2 8 9

Configuration Configuration

(c) 0.035
Average Traction Coefficient

0.030

0.025

0.020

0.015

0.010

0.005

0.000
6 3 1 5 4 7 2 8 9
Configuration

Figure 6 Comparison of performance between the different configurations (defined in Table 2)

for tests performed at 16 m/s sliding speed. Uncoated configurations added for comparison.
(a) scuffing load; (b) average mean bulk temperature at scuffing load;
(c) average traction coefficient at scuffing load.

4.0 MICRO EHL THEORY AND APPLICATION TO MICROPITTING

Considerable progress has been made with modelling of roughness in EHL contacts starting with idealised
single features or sinusoidal waves and progressing to measured (real) roughness on both line and point
contacts. Most analyses are based on a two-dimensional or nominal line contact model. This is a
reasonable assumption in the case of conventional involute, parallel axis gears where rolling/sliding takes
place across the finishing marks which are close to being two-dimensional, as may be judged from
Figure 2.

The early attempts to model micro EHL were based upon a sinusoidal representation of roughness under
steady state conditions (Goglia et al [25]; Lubrecht et al [26]; Kweh et al [27]; Elsharkawy and Hamrock
[28]) More recently real surfaces as represented by profilometer traces of actual surfaces have been used
in steady-state solutions (Sadeghi [29]; Kweh et al [30]; Venner, et al [31]). The problem of non-steady
(i.e. moving roughness) conditions has been addressed by Chang and Webster [32], Chang [33], and
Chang et al [34] using sinusoidal roughness, and by Ai and Cheng [35] with random roughness. The
important influence of non-Newtonian (shear thinning) effects in micro EHL solutions has been reviewed
and studied by Chang and Zhao [36]. The inclusion of non-Newtonian lubricant behaviour and
consideration of the roughness moving relative to the contact tend to reduce the pressure ripples but also
tend to give thinner oil films.

Zhu and co-workers [37,38] were among the first to develop a “mixed” lubrication EHL model for
elliptical contacts in which films below an arbitrary low value are regarded as creating contacts with a
corresponding simplifying modification of the Reynolds equation presumed at the contacting areas. Their

RTO-MP-AVT-109 14 - 7

UNCLASSIFIED/UNLIMITED
 UNCLASSIFIED/UNLIMITED

Understanding Scuffing and Micropitting of Gears

analyses have predicted a considerable degree of direct surface contact [39], which is stated to be
insensitive to the (small) film level at which contact is taken to occur. The treatment of contact in this
model is somewhat arbitrary however, and the assumptions made in dealing with contact remain to be
verified. In contrast our contribution to rough surface EHL under severe thin film conditions is based on a
theoretical model for line contacts (the appropriate configuration for spur gears) that solves the elastic and
hydrodynamic film thickness equations simultaneously using a consistent, mass-conserving, fully coupled
method [40, 41]. This new robust numerical technique has enabled results to be obtained for extremely
low Λ values with contact, if it occurs, established directly from solution of the time-dependent equations.
A second important feature that emerges using this approach, dependent upon the oil rheology model, is
that of lubricant cavitation under severe conditions of thin films/high roughness [42].

10 mm

Figure 7: Scuffing scar on an axially-finished test disc. Superimposed

ellipse indicates approximate equivalent Hertzian contact.

In some circumstances contact occurs between the surfaces as part of the solution process. If, in a
particular timestep a converged result is obtained with a negative calculated film thickness, then the most
negative film thickness is set to zero. The hydrodynamic equation at this nodal point is deleted from the
problem matrix, but the elastic deflection equation is retained. (In this way the film thickness value of
zero is a boundary condition for the elastic deflection equation which ensures that the pressure distribution
obtained remains entirely consistent with the film shape. The pressure developed in the coupled solution
at the contacting node is then an automatic boundary condition for the hydrodynamic equation on both
sides of the contacting node.) The timestep is then re-calculated. This procedure is repeated for the
timestep, adding no more than one contact point per timestep re-calculation, until a converged result with
no negative calculated film thicknesses is obtained. Contact, as calculated in this way, is found to be a
relatively infrequent event in the line contact results obtained as discussed below.

A further instance where a full film solution is not obtained at each node point is when what may be
termed “in-contact cavitation” occurs. Cavitation at the exit of a contact, where the calculated pressure
falls to sub-ambient values, is a natural part of any EHL solution. Situations occur in rough EHL contacts
where composite valley features are increasing in volume with time. If flow into such a feature from the
micro contacts at either side is insufficient, then the pressure in the composite valley feature will fall. If
this effect is sufficiently strong then the oil ‘trapped’ in the composite valley can become entirely de-
compressed, returning progressively to ambient pressure. When the calculated pressure becomes negative
the lubricant is regarded as cavitated. In these solutions the occurrence of in-contact cavitation is found to

14 - 8 RTO-MP-AVT-109

UNCLASSIFIED/UNLIMITED
 UNCLASSIFIED/UNLIMITED

Understanding Scuffing and Micropitting of Gears

increase with the degree of sliding and this suggests, perhaps, a possible erosion wear mechanism that may
be related to micropitting damage.

Both localised contact and in-contact cavitation are effects determined from the numerical analysis. The
sensitivity of these calculated events to the details of the numerical scheme is the focus of a current
investigation by the authors that will be reported in due course. There seems to be little doubt that micro
contact does indeed take place between rough surfaces in EHL, as electrical contact resistance
measurements provide strong supportive evidence of this phenomenon. The occurrence of calculated in-
contact cavitation is a new feature of this kind of analysis, which may be potentially damaging to the
surfaces as it takes place in isolation from the ambient atmosphere.
The line contact micro-EHL analysis technique has been applied to a set of profiles corresponding to the
later (most heavily loaded) stages in an FZG gear test. The results form part of an investigation
programme comparing micropitting performance of test gears using different oils. Use of experimental
gear tooth profiles under the real operating conditions of the gears is judged to be an important
requirement as surface profile topography is significantly modified by running-in, and engineering interest
must be primarily focussed on the stress and film thickness response actually encountered by the
components.

8
6

6
4

4
p / GPa

2
h / µm

2
0

-2 0

-4 -2
-1 -0.5 0 0.5 1

0.1 0.3 0.5 0. 1

0. 1

0.2
0. 3

0.3
1
0.

0.4
z/a

0.4
0.6
0.3

0 .4

0.8
0.45
1
-1 -0.5 0 0.5 1
x/a

Figure 8: Pressure and film thickness at a particular timestep in a micro EHL solution
for gear tooth contact in an FZG test. Corresponding maximum Hertzian pressure = 1.4
GPa. Also shown are contours of τ/GPa, the sub-surface maximum shear stress.

Some results from this work are shown as follows. Figure 8 shows the pressure and film thickness
variation at a particular timestep corresponding to load stage 9 in a standard FZG test in which the
equivalent Hertzian (smooth surface) maximum contact pressure is 1.4 GPa. The profiles used in the
simulation are taken from the gears after running at this stage, and the properties of the oil are those for the
lubricant used in the tests. The FZG pinion speed was 2250 rpm and oil viscosity (0.028 Pas) was taken to
be that at the feed temperature of 60oC. Non-Newtonian lubricant behaviour was modelled according to
the Eyring equation. For the result shown in Figure 8 a pressure-dependent Eyring stress given by
τ0 = 0.03p was used subject to a minimum value of 10 MPa. Also shown beneath the film thickness

RTO-MP-AVT-109 14 - 9

UNCLASSIFIED/UNLIMITED
 UNCLASSIFIED/UNLIMITED

Understanding Scuffing and Micropitting of Gears

distribution is a contour plot of the subsurface maximum shear stress corresponding to the pressure
distribution shown. It is clear that pressures well in excess of the Hertzian values are generated on
individual micro asperity contacts. Pressure values in the composite valley areas are much lower, but
clearly contribute significantly to the load carrying of the overall EHL film. The elevated values and rapid
spatial variation of pressure leads to concentrations of shear stress levels close to the surface. Maximum
shear stress values as high as 0.7 GPa occur compared with the smooth surface value of 0.42 GPa. Video
sequences made up of such figures enable the variation of pressure and film thickness to be assessed in a
qualitative way and are invaluable for developing an understanding of the mechanisms at work as
individual asperity features enter the Hertzian contact area and subsequently pass through it, as discussed
in [41]. Videos for results such as those presented in this paper are available from the authors on CD by
request. Contact between individual asperities and internal cavitation have been observed in such
numerical solutions, as discussed above and in [42] where a thinner oil (η0 = 0.0048 Pas, α = 11.1 GPa-1)
was used for the analysis.

The EHL response of these contacts requires a statistical evaluation of loading severity with parameters
obtained by averaging over a sufficiently long transient analysis. Figure 9 shows two such parameters
(defined in the Notation section) used to compare the EHL response of profiles taken from FZG gear teeth
at four different load stages. Parameter Φ(p) is the fraction of the Hertzian contact area that has a pressure
that exceeds p. Load stage 6 has the least amount of its contact area subject to extremes of pressure, and
load stage 9 has the greatest amount. However all cases tend to the same level of extreme pressure with
about 0.25% of the area subject to pressures of 4.5 GPa. The maximum pressures are thus seen to be
significant in relation to the hardness of the gear teeth (700 Vickers hardness number) suggesting that
surface modification is achieved by plastic deformation under the action of the EHL asperity pressures.
The fraction of the contact area subject to pressures between 1 and 2 GPa is seen to rank the profiles
according to the applied load, whereas true extremes of pressure may well be governed by surface
hardness considerations. Figure 9 also gives the cumulative film thickness distribution for small film
thickness values for the four profiles. Parameter Φ(h) is the fraction of the Hertzian contact area that has a
film thickness that is less than h. The lowest film values observed are of the order 0.075 µm, and the
earlier load stages have a greater tendency towards the lower film values, for example 2% of the contact
area for load stage 6 is below 0.15 µm, a figure that reduces to 1% for load stage 9. These results thus
suggest that the surface modification (running in) induced at each load stage causes the rough surface to
adopt a shape that is more able to generate lubricant films on the micro asperities.

20% 6%
stage 6
18% stage 6 stage 7
16% stage 7 stage 8
stage 8
14% stage 9
stage 9 4%
12%
(p)

(h)
φφ(h)
φφ(p)

10%
8%
2%
6%
4%
2%
0% 0%
1.0 2.0 3.0 4.0 5.0 0.0 0.1 0.2 0.3

pp//GPa
GPa hh/ /µµm
m
Figure 9: Maximum pressure Φ(p) parameter and low film parameter Φ(h) obtained by averaging
results of micro EHL solutions corresponding to four different load stages in an FZG gear test.

14 - 10 RTO-MP-AVT-109

UNCLASSIFIED/UNLIMITED
 UNCLASSIFIED/UNLIMITED

Understanding Scuffing and Micropitting of Gears

The above micro-EHL analyses are based on a line contact model, but in all real gear contacts side-
leakage effects will occur even in nominal line contact gears due to edges and waviness of the finish in the
axial direction. In gears which are deliberately arranged to have point contacts (such as crowned spur
gears, crossed axis gears and special types such as Wildhaber-Novikov) a full side-leakage analysis
becomes a necessity. Progress in modelling real roughness on an elliptical contact corresponding to the
Cardiff crowned disc machine has been made by Holmes [43]. Figure 10 shows predicted contours of the
film under rolling/sliding conditions with roughness of Ra = 0.26 µm (0.25 mm cut off) on both surfaces.
Entrainment is in the direction of the minor axis of the contact. Over most of the contact the behaviour is
effectively one-dimensional, but thinning of the film occurs at the transverse edges due to sideways flow
of the lubricant in the deep valley features. This thinning (compared to that on the centre line, y = 0) is
much more severe than that occurring in the corresponding smooth-surface case This predicted behaviour
correlates, qualitatively, with the observation that scuffing in the disc tests initiated at the transverse edges
of the contact.

-1

2.00
0.70
0.50
0.40
x/a

0 0.30
0.20
0.10
0.05
0.00

0 1 2 3 4 5

y/a

Figure 10: Contours of film thickness/µm in micro EHL solution corresponding to contact in a crowned
disc test as shown in Figure 7. Arrow indicates position of minimum film thickness (0.011 µm).

5.0 WORM GEARS

The lubricated contacts in worm gears are similar to those occurring in high conformity Wildhaber-
Novikov gears [44] giving rise to particularly severe conditions because of the elongated shape of the dry
contact region. The contact geometry is non-symmetrical/non-Hertzian and although entrainment of the
oil is predominantly along the major axis of the contact its direction varies throughout the area of the
conjunction (i.e. a significant spin component of velocity is present). High sliding and heat generation in
the contact necessitate a thermal/non-Newtonian EHL treatment. A typical solution obtained for a worm
gear contact [45] is shown in Figure 11(a). An interesting feature of the contours shown in this figure is
the appearance of a narrow region of almost zero film thickness (within the 0.47 µm contour towards the
exit) which effectively separates the overall contact into two regions. This behaviour is explained as
follows. A feature of EHL is that within the thin-film main load-bearing part of the contact
(corresponding roughly with the equivalent Hertzian region) flow of the lubricant is dominated by the
entraining motion of the surfaces (Couette flow). The pressure-induced (Poiseuille) flow is negligible
because of the tremendous increase in viscosity due to pressure. Oil entrained into this region will

RTO-MP-AVT-109 14 - 11

UNCLASSIFIED/UNLIMITED
 UNCLASSIFIED/UNLIMITED

Understanding Scuffing and Micropitting of Gears

therefore follow the entrainment streamlines, and these may be determined from purely kinematic
considerations of the gears. It also follows that at points where entrainment lines are tangential to the edge
of the high pressure region there will be zero net entrainment of oil into the contact at the tangency point,
and this indicates a "line" of poor film formation. This hypothesis is reinforced by consideration of Figure
11(b) which shows the entrainment streamlines corresponding to Figure 11(a) together with the dry
contact region. The striking similarity between the region of severe film thinning and the path of the
tangential entrainment streamline (tangential at point A) within the contact is clear.

(a) (b)
Figure 11: (a) contours of film thickness/µm obtained from EHL solution for a worm
gear contact (inlet is on the left); (b) corresponding dry elastic contact
with key entrainment velocity flowlines superimposed.

6.0 FRICTION
The efficiency of work transfer in a gear mesh is generally in excess of ninety percent, so in many
industrial applications overall efficiency of a geared transmission system is not of major concern.
However, in high technology applications such as aerospace the losses arising from friction can become of
critical importance. For example, the need for ever greater efficiency of propulsion systems for large
transport aircraft has led to serious consideration of geared propeller engines or "geared fans" where a
reduction gear drive is used to match the speed of an optimum-efficiency gas turbine to that of an efficient
propeller or fan. The powers involved are high, typically 30 MW or more. In these applications reduced
gear tooth friction can potentially lead to significantly lower amounts of heat being generated at the tooth
contacts and hence less bulky cooling arrangements. High gear tooth temperatures are also the
predominant factor associated with scuffing, which is a further important reason for reducing friction

Measurements of losses in gearing have been made by Britton et al [46] using a special gearbox mounted
on trunnion bearings as illustrated in Figure 12. Power is recirculated between the two pairs of meshing
gears which are loaded by means of a torque shaft. After calibrating and deducting the churning and
bearing losses, the torque required to drive the rig gives the frictional power at the tooth contacts. Figure
13 shows a comparison of tooth friction losses for both conventionally ground and superfinished gears.
This work demonstrates the potential energy-saving benefits of improved surface finish in gearing. Work
is now underway to develop theoretical models of gear tooth traction based upon both Eyring and limiting
shear stress formulations for the lubricant.

7.0 CONCLUSIONS
This brief review of surface distress and elastohydrodynamic lubrication effects in gears demonstrates the
importance of the real effects of surface roughness including: non-steady behaviour due to the motion of
roughness features relative to the nominal contact; the presence of relatively high degrees of sliding and
the consequent frictional heating; non-Hertzian and non-symmetrical geometries. Encouraging progress is
being made in the development of new, robust solvers that can deal with the severe conditions of thin
films/rough surfaces that characterise real gear tooth contacts. The reality of what has long been called
“mixed” lubrication in an otherwise hydrodynamically-lubricated conjunction is beginning to be modelled

14 - 12 RTO-MP-AVT-109

UNCLASSIFIED/UNLIMITED
 UNCLASSIFIED/UNLIMITED

Understanding Scuffing and Micropitting of Gears

and seen in these solutions. A second real feature is that of what we have called “in-contact cavitation”
due to the presence of deep valley features within the contact, particularly under sliding conditions. The
tendency for micropitting to occur in practice at the beginning and end of the meshing cycle, where sliding
is greatest, leads us to suggest cavitation as a possible mechanism for erosive wear that may be part of the
micropitting explanation. Edge effects in rough contacts are seen to be of crucial importance in allowing
the otherwise stiff and robust EHL mechanism to weaken due to sideways leakage of lubricant. This
effect is present in rough contacts to a far greater extent than occurs with perfectly smooth surfaces and
reinforces the view that it is strongly related to the real phenomenon of scuffing.

Figure 12: Gear friction rig.

350

300 ground
superfinished
250
Power Loss (W)

200

150

100

50

0
0 1000 2000 3000 4000 5000 6000
Speed (rpm)

Figure 13: Measured gear mesh friction loss versus

sped for ground and superfinished steel gears.

RTO-MP-AVT-109 14 - 13

UNCLASSIFIED/UNLIMITED
 UNCLASSIFIED/UNLIMITED

Understanding Scuffing and Micropitting of Gears

8.0 ACKNOWLEDGEMENTS
The authors acknowledge the contribution of Mr C Aylott (Newcastle University) in providing a figure for
the paper. Financial support for the research was provided by the British Gear Association, EPSRC and
the Ministry of Defence.

9.0 NOTATION
b semi-width of Hertzian line contact m
E elastic modulus Pa
E' {( ) ( ) }
= 0.5 1 − ν 12 / E1 + 1 − ν 12 / E1
−1
Pa
h film thickness m
hmin minimum film thickness Pa
p pressure Pa
rb base circle radius m
R radius of relative curvature of contact m
R1,R2 profile radii of teeth at the contact point m
s distance between contact and pitch point m
u mean entraining speed m/s
us sliding speed m/s
w′ load per unit length of line contact N/m
α pressure coefficient of viscosity Pa-1
η viscosity Pas
η0 viscosity at zero pressure Pas
Λ ratio of film thickness to composite roughness
ν Poisson’s ratio
σ rms surface roughness m
τ0 Eyring shear stress Pa
Φ(h) fraction of dry contact area whose
film thickness is less than h
Φ(p) fraction of dry contact area whose
pressure exceeds p
ψ pressure angle rad
ω1,2 Angular velocity of wheel and pinion rad/s

10.0 REFERENCES

[1] Dowson, D and Higginson, G R, 1966, Elasto-hydrodynamic Lubrication, Pergamon Press, Oxford.
[2] Dyson, A, Naylor, H and Wilson, A R, “The measurement of oil film thickness in
elastohydrodynamic contacts”, Proc. Instn Mech. Engrs Vol 180(3B), 1965, pp 119-134.
[3] Greenwood, J A and Kauzlarich, J J, “Inlet shear heating in elastohydrodynamic lubrication”, Trans
ASME Jn. Lub. Tech. Vol 95, 1973, pp 417-426.
[4] Dyson, A, Evans, H P and Snidle, R W 'Wildhaber Novikov circular arc gears: some properties of
relevance to their design', Proc. R. Soc., London, A425 (1989), pp. 341-363.
[5] Chittenden, R J, Dowson, D, Dunn, J F and Taylor, C M, "A theoretical analysis of the isothermal
elastohydrodynamic lubrication of concentrated contacts", Proc. R. Soc. Lond., Vol A397, 1985,
pp 245-269, 271-294.

14 - 14 RTO-MP-AVT-109

UNCLASSIFIED/UNLIMITED
 UNCLASSIFIED/UNLIMITED

Understanding Scuffing and Micropitting of Gears

[6] ESDU, 1989, “Film thickness in lubricated Hertzian contacts. Part 2: point contacts”
[7] Harris, T A, Rolling Bearing Analysis, Second Edition, Wiley, New York, 1984, p 508.
[8] O.E.C.D., Glossary of Terms and Definitions in the Field of Friction, Wear and Lubrication, 1969,
pp 52-53.
[9] Berthe, D, Flamand, L, Foucher, D and Godet, M, 1980, "Micro-pitting in Hertzian contacts", Trans
ASME, J. Lubr. Tech., Vol 102, pp 478-489.
[10] Olver, A V, Spikes, H A and MacPherson, P B, 1986, “Wear in rolling contacts”, Wear Vol 112
pp 121-144.
[11] Zhou, R S, Cheng, H S and Mura, T, 1989, "Micropitting in rolling and sliding contact under mixed
lubrication", Trans ASME, J. of Tribology, Vol 111, pp 605-613.
[12] Webster, M N and Norbart, C C J, 1995, "An experimental investigation of micropitting using a
roller disk machine", Tribol. Trans, Vol 38, pp 883-895.
[13] Shotter, B A, 1981, "Micropitting, its characteristics and implications, in performance and testing of
gear oils and transmission fluids", Institute of Petroleum.
[14] Webster, M N and Sayles, R S, 1986, “A numerical model for the elastic frictionless contact of real
rough surfaces”, Trans ASME, Jn Tribology, Vol 108, pp 314-320.
[15] Snidle, R W and Evans, H P, 1994, "A simple method of elastic contact simulation", Proc. Instn
Mech. Engrs Part J Jn of Engng Tribology Vol 208, pp 291-293.
[16] Mayeur, C, Sainsot, P and Flamand, L, 1995, "A numerical elastoplastic model for rough contact",
Trans ASME, J. of Tribology, Vol 117, pp 422-429.
[17] Kapoor, A and Johnson, K L, 1995, "Plastic ratchetting as a mechanism of erosive wear", Wear Vol
190, pp 86-91.
[18] Tyfour, W R and Beynon, J H, 1994, "The effect of rolling direction reversal on fatigue crack
morphology and propagation", Tribology International, Vol 27, pp 273-282.
[19] Krantz , T L, Alanou, M P, Evans, H P and Snidle, R W, “Surface fatigue lives of case-carburized
gears with an improved surface finish”, Trans ASME, Journal of Tribology Vol 123, pp 709-716,
2001.
[20] Patching, M J, Kweh, C C, Evans, H P and Snidle, R W, 1994, "Conditions for scuffing failure of
ground and superfinished steel discs at high sliding speeds using a gas turbine engine oil", Trans
ASME Journal of Tribology Vol 117, pp 482-489.
[21] Sproul, W. D., Physical vapor deposition tool coatings, Surface & Coatings Technology, Vol 81
(1996), pp. 1-7
[22] Eckardt, T., Bewilogua, K., van der Kolk, G., Hurkmans, T., Trinh, T., Fleischer, W., Improving
tribological properties of sputtered boron carbide coatings by process modifications, Surface and
Coatings Technology, Vol 126 (2000), pp 69-75
[23] Hu, T., Steihl, L., Rafaniello, W., Fawcett, T., Hawn, D. D., Mashall, J. G., Rozeveld, S. J., Putzig,
C. L., Blackson, J. H., Cermignani, W., Robinson, M. G., Structures and properties of disordered
boron carbide coatings generated by magnetron sputtering, Thin Solid Films, Vol 332 (1998), pp. 80-
86
[24] Evans, H P and Snidle, R W, 1996, "A model for elastohydrodynamic film failure in contacts
between rough surfaces having transverse finish", Trans ASME Journal of Tribology, Vol 118,
pp 847-857.

RTO-MP-AVT-109 14 - 15

UNCLASSIFIED/UNLIMITED
 UNCLASSIFIED/UNLIMITED

Understanding Scuffing and Micropitting of Gears

[25] Goglia, P R, Cusano, C and Conry, T F, 1984, "The effects of surface irregularities on the
elastohydrodynamic lubrication of sliding line contacts. Part I - single irregularities; Part II - wavy
surfaces", Trans ASME, Jn. of Tribology, Vol. 106, pp 104-112, 113-119.
[26] Lubrecht, A A, Ten Napel, W E and Bosma, R, 1988, "The influence of longitudinal and transverse
roughness on the elastohydrodynamic lubrication of circular contacts", Trans ASME Jn of Tribology,
Vol 110, pp 421-426.
[27] Kweh, C C, Evans, H P, and Snidle, R W, 1989, "Microelastohydrodynamic lubrication of an
elliptical contact with transverse and three-dimensional roughness", J. of Tribology, Trans ASME,
Vol 111, pp 577-584.
[28] Elsharkawy, A A and Hamrock, B J, 1991, "Subsurface stresses in micro-EHL line contacts", ASME
Journal of Tribology, Vol 113, pp 645-655.
[29] Sadeghi, F, 1991, "A comparison of the fluid model's effect on the internal stresses of rough
surfaces", ASME Journal of Tribology Vol 113, pp 142-149.
[30] Kweh, C C, Patching, M J, Evans, H P, and Snidle, R W, 1991, "Simulation of elastohydrodynamic
contacts between rough surfaces", J. of Tribology, Trans ASME, Paper 91-Trib-36.
[31] Venner, C H and ten Napel, W E, 1992, "Surface roughness effects in an EHL line contact", ASME
Journal of Tribology, Vol 114, pp 616-622.
[32] Chang, L and Webster, M N, 1991, "A study of elastohydrodynamic lubrication of rough surfaces",
ASME Journal of Tribology, Vol 113, pp 110-115.
[33] Chang, L, 1992, "Traction in thermal elastohydrodynamic lubrication of rough surfaces", ASME
journal of Tribology, Vol 114, pp 186-191.
[34] Chang, L, Webster, M N and Jackson, A, 1993, "On the pressure rippling and roughness deformation
in elastohydrodynamic lubrication of rough surfaces", ASME Journal of Tribology, Vol 115,
pp 439-444.
[35] Ai, X and Cheng, H S, 1993, "A transient EHL analysis for line contacts with measured surface
roughness using multigrid technique", ASME Journal of Tribology, Paper No 93-Trib-56.
[36] Chang, L and Zhao, W, 1995, "Fundamental differences between Newtonian and non-Newtonian
micro-EHL results", ASME Journal of Tribology, Vol 117, pp 29-35.
[37] Zhu, D. and Hu, Y Z., 1999, “The study of transition from full film elastohydrodynamic to mixed
and boundary lubrication” Proc STLE/ASME H.S. Cheng Tribology Surveillance, 150-156.
[38] Hu, Y Z, and Zhu, D., 2000, “A full numerical solution to the mixed lubrication in point contacts,”
ASME Journal of Tribology, Vol 122, 1-9.
[39] Zhu, D. and Hu, Y Z., 2000, “Effects of rough surface topography and orientation on the EHD and
mixed lubrication characteristics” Proceedings International Tribology Conference, Nagasaki,
625-630.
[40] Hughes, T.G., Elcoate, C.D. & Evans, H.P. Coupled solution of the Elastohydrodynamic line contact
problem using a differential deflection method. Proc. Instn Mech. Engrs Part C Jn of Mechanical
Engineering Science, Vol 214, pp 585 598, 2000.
[41] Elcoate, C D, Hughes, T G, Evans, H P., and Snidle, R W., 2001, “Transient elastohydrodynamic
analysis using a novel coupled differential deflection method” Proc. Instn. Mech. Engrs Part J, Jn of
Engng Tribology , Vol 215, pp 319-337.
[42] Tao, J, Hughes, T G, Evans, H P, & Snidle, R W., 2002, "Elastohydrodynamic response of transverse
ground gear teeth". To appear Proc. 28th Leeds-Lyon Symp. on tribology, Elsevier, Amsterdam.

14 - 16 RTO-MP-AVT-109

UNCLASSIFIED/UNLIMITED
 UNCLASSIFIED/UNLIMITED

Understanding Scuffing and Micropitting of Gears

[43] Holmes, M.J. Transient analysis of the point contact elastohydrodynamic lubrication problem using
coupled solution methods. PhD thesis, University of Wales, 2002.
[44] Evans, H P and Snidle, R W, 'Wildhaber-Novikov circular arc gears: Elastohydrodynamics', Trans
ASME Jn of Tribology, Vol 115, pp 487-492, 1993.
[45] Sharif, K, Kong, S, Evans, H P and Snidle, R W, “Contact and elastohydrodynamic analysis of worm
gears: Part 2, results”, Proc. Instn Mech. Engrs, Part C: Journal of Mechanical Engineering Science
Vol 215 (2001), pp 831-846.
[46] Britton, R D, Elcoate, C D, Alanou, M P, Evans, H P and Snidle, R W, “Effect of surface finish on
gear tooth friction”, Trans ASME Jn. of Tribology, Vol 122, pp 354-360, 2000.

RTO-MP-AVT-109 14 - 17

UNCLASSIFIED/UNLIMITED
 UNCLASSIFIED/UNLIMITED

Understanding Scuffing and Micropitting of Gears

14 - 18 RTO-MP-AVT-109

UNCLASSIFIED/UNLIMITED