Gear Design

1
lY
1
1 Technical Memorandum No. 33-139
1
I
I
I
1
J a m e s Clyde Randall r-, /763 /+
/25
1
1
I Spacec
tp f t Development Seckon
I REPRODUCED BY
NATIONAL TECHNICAL ,
INFORMATION SERVICE
U. S. DEPARTMENT OF COMMERCE
I SPRINGFIELD, VA 22161
I
I
I
Copyright 0 1963
Jet Propulsion Laboratory
California Institute of Technology
Prepared Under Contract No. NAS 7-100

National Aeronautics & Space Administration
~
I JPL Technical Memorandum No. 33-139
I CONTENTS
I I. Introduction .................. .. . . . . . 1
11. Friction, Surface Damage, and Wear . . . . . . . . . . . 4

I 111. Short Discussion on Gear-Load Calculations. . . .... . 19
I IV. ....
Fundamentals of Involute Spur G e a r s . ...... 23
V. Gear Wear T e s t s and Results . . . . . . . . ... . . . . 46
I IV. Summary . . . . . . . . . . . . . . . . . . . ... . . .. 76
References . . . . . . . . . . . . . . . . . . . . . ... . . ..
I Appendix A: Sliding Velocity of Gears . . . . . . . . . . . . ..
80
82
I Appendix B: Calculation of Test-Gear Tooth Loads and

S t r e s s e s Using Buckingham' s Formulae and
American Standards As sociation Specification
B6.11-1951 . . . . .
... ......
... ....
I Appendix C: Calculation of Test-Gear Tooth Loads and
84
S t r e s s e s Using Buckingham' s Formulae and

I Tuplin' s Method for Effective E r r o r . ....... 96
Appendix D: Wear Rate Data . . . . . . . . . . . . . . . . . . . 107
1 Appendix E: Calculation of Wear Rates . . . . . . . . . . . . . 110
Calculation of Fatigue Lives . . . . . . . . . . . .

I Appendix F: 116
TABLES
1
1. Wear r a t e s and calculated s t r e s s e s f o r m a t e r i a l
. ... .. . .....
1 2.
combinations and loads . . . . . ..
Wear r a t e s and s t r e s s e s f o r t e s t m a t e r i a l s . . . . . . . .
64
66
I 3. S t r e s s e s and calculated fatigue lives f o r t e s t g e a r s . . . . 75
1
1
I
I
JPL Technical Memorandum No. 33- 139
I
FIGURES
I
1. Comparison of friction and adhesion f o r s t e e l on I
indium . . . . . . . . . . . . . . . . . . . . . . . . ... 7
2. Comparison of calculated p a r t i c l e s i z e and m e a s u r e d

average p a r t i c l e s i z e . . . . . . . . . . . . . . . . . .. 10
I
3. Comparison of dynamic loads ..... . ......... 22
I
4. Two mating involute profiles . . . . . . ......... 25
5. Dimensions f o r gear teeth and rim . . . ......... 36 I

Factors pertaining to s t r e t c h of rim . . .........
6.
7. Relationship of measured tooth-to-tooth e r r o r by center
38
I
distance deviation to actual pitch e r r o r of g e a r teeth. . . 38
8. Comparison of ratio of insertion time to n a t u r a l period

I
( t l / T I ) and ratio of effective e r r o r in action to actual
e r r o r in action . . . . . . . . . . . . . . . . . . . . . .
9. Surface fatigue curves for 303 s t a i n l e s s s t e e l and

44
I
2024-T4 aluminum. .................... 45
10. Gear t e s t f i x t u r e . ..................... 48

I
11. Wear r a t e for 303 stainless s t e e l on 303 s t a i n l e s s
s t e e l (no load, 3800 r p m ) . . . . . . . . . . . . . .... 54 I
12. Wear r a t e f o r 303 stainless s t e e l on 2024-T4
aluminum (no load, 3800 r p m ) .............. 54 I
13. Wear r a t e f o r 303 s t a i n l e s s s t e e l on anodized 2024-T4
aluminum (no load, 3800 r p m ) .............. 55 I
14. Wear r a t e f o r 303 s t a i n l e s s s t e e l on anodized 2024-T4
aluminum t r e a t e d with molybdenum disulphide (no
load, 3800 r p m ) ...................... 55 I
15. Wear r a t e for anodized 2024-T4 aluminum on
anodized 2024-T4 aluminum (no load, 3800 r p m ) . .... 56 I
16. Wear r a t e for 303 s t a i n l e s s s t e e l on d e l r i n (no
load, 3800 r p m ) ...................... 56 I
17. Wear r a t e f o r 303 stainless s t e e l on 303 s t a i n l e s s
.................
18.
s t e e l ( 4 in. -oz, 380 r p m )
Wear r a t e f o r 303 s t a i n l e s s s t e e l on 2024-T4

57
I
..............
aluminum (4in. -02, 380 r p m )
- iv -
57
I
I
FIGURES (Cont' d)
19. Wear r a t e f o r 303 stainless s t e e l on anodized 2024-T4

aluminum ( 4 in. -oz, 380 rprn) . . .
..... ...... 58

aluminum t r e a t e d with molybdenum disulphide
( 4 in. -oz, 380 r p m ) . . . ............ .. ... 58
21. Wear r a t e f o r anodized 2024-T4 aluminum on anodized

2024-T4 aluminum ( 4 in. -oz, 380 r p m ) . . ....... 59
22. Wear r a t e f o r 303 stainless s t e e l on d e l r i n ( 4 in. -oz,

380 r p m ) ........ .... .............. 59
23. Wear r a t e f o r 303 stainless s t e e l on 303 s t a i n l e s s

s t e e l ( 6 in. -oz, 76 r p m ) ....... . ... ....
. . . 60
24. Wear r a t e f o r 303 stainless s t e e l on 2024-T4
aluminum ( 3 in. -oz, 76 r p m ) . . . . ...
.. ...... 60

aluminum ( 3 in. -02, 76 rpm) . . ... . . .....
.. . . 61
( 3 in. -oz, 76 r p r n ) . ..
... .. .......... .
. . 61
27. Wear r a t e f o r anodized 2024-T4 aluminum on

anodized 2024-T4 aluminum ( 3 in. -oz, 76 r p m ) ..... 62
28. Wear r a t e f o r 303 stainless s t e e l on d e l r i n ( 3 in. -oz,

76 r p m ) . . . . . . .................... 62
29. Design curve f o r 303 stainless s t e e l indicating

calculated s t r e s s and corresponding depth-of -
wearrate.. . . . . . . . .... .. ........... 67
30. Design curve f o r 2024-T4 aluminum indicating

calculated s t r e s s and corresponding depth-of -
wear r a t e . . . . .... .. .. ..... .. ....
. . 68
3 1. Design curve f o r anodized 2024-T4 aluminum indicating

calculated s t r e s s and corresponding depth-of -wear
rate . . .. ....... .... . .. .......
. . . 69
32. Design curve for anodized 2024-T4 aluminum t r e a t e d with
molybdenum disulphide indicating calculated s t r e s s and
corresponding depth-of -wear r a t e . .... ....... 70
33. Design curve f o r d e l r i n indicating calculated s t r e s s and

corresponding depth-of -wear rate . . ........ . . 71
- v -
JPL Technical Memorandum No. 3 3 - 139
I
ABSTRACT I
I /
This paper establishes a method of determining w e a r r a t e s f o r
/i ;- I
non-lubricated, fine -pitch, precision instrument spur g e a r s . The
concepts of wear and the problems associated with applying t h e s e con-

I
cepts to the unique action of s p u r - g e a r s u r f a c e s a r e discussed. The
I
p r o p e r t i e s of the involute curve a r e included only to the extent that is
deemed necessary to analyze thoroughly the g e a r -wear problem. 1

Wear data for t e s t g e a r s run at various loads and speeds a r e collected
to determine the w e a r r a t e s f o r the m o s t popular m a t e r i a l s in u s e

I
today. Design curves a r e made for five m a t e r i a l s ( o r s u r f a c e s )
I
relating wear r a t e s to calculated H e r t z ' s t r e s s e s :
1. 303 stainless steel 1

2. 2024-T4 aluminum
3. Anodized 2024-T4 aluminum

I
4. Anodized 2024-T4 aluminum t r e a t e d with molybdenum
di sulphide
I
5. Delrin I
Design curves consisting of any one o r combination of the above m a t e -
rials can be used to analyze w e a r r a t e s of a g e a r train. In addition,

I
the w e a r r a t e s a r e established in a depth-of-wear p e r revolution s o that
the expected life of a s y s t e m can be determined, realizing that s o m e

8
s y s t e m s can t o l e r a t e m o r e w e a r than o t h e r s y s t e m s before they can be I
said to have failed.
In addition to the data presented i n the w e a r c h a r t s , this paper I

proposes a method f o r using the w e a r data to s e l e c t between two
popular methods of computing dynamic load; namely, the American

I
I
-vi -
I
J P L Technical Memorandum No. 3 3 - 1 3 9
I. INTRODUCTION
Wear is the l o s s of m a t e r i a l f r o m s u r f a c e s that slide and b e a r on
each other. Almost all mechanisms m u s t have v e r y p r e c i s e sliding
s u r f a c e s in o r d e r to function properly.k The deterioration of these
s u r f a c e s because of w e a r usually h a s a detrimental effect on the p e r -
f o r m a n c e of such mechanisms. Although wear o c c u r s f r o m the t i m e
that these s u r f a c e s come into contact, the w e a r rate determines the
life of the mechanism until i t is deemed useless.
Studies on w e a r have been somewhat neglected historically because
of the inability to accurately m e a s u r e wear on surfaces. This w a s not
a problem in the past, however, because of the inability to hold close
tolerances. F o r example, tolerances as l a r g e as a couple of thou-
sandths of a n inch w e r e as close a s could be held, and a w e a r of a few
tenths of thousandths of an inch w a s not important. The "state of the
art" has p r o g r e s s e d , however, and tolerances of a few tenths of thou-
sandths of an inch are held quite easily. A wear of a few tenths of
thousandths of an inch becomes relatively important in this case.
Recent introductions of radioactive-isotope measuring techniques
make a c c u r a t e and reproducible w e a r r e s u l t s possible, which w i l l
undoubtedly stimulate i n t e r e s t in the subject of w e a r in the future.
In the discussion of w e a r i t should be pointed out that t h e r e are
specific methods to combat w e a r . One of the m o s t significant ways to
reduce the w e a r r a t e i s with a lubricant. A lubricant i s a low s h e a r -
strength m a t e r i a l placed between the sliding surfaces. A lubricant
does not prevent wear, but, reduces it by reducing the number of
- 1 -
JPL Technical Memorandum No. 33-139
contact points between the sliding s u r f a c e s . Although a lubricant is
generally thought of a s an oil, g r e a s e , o r s o m e s o r t of dry-film, some
precious metal platings have been used a s lubricants b e c a u s e of their
low s h e a r -strengths.
T h e r e a r e reasons, however, why a lubricant should not be used
a t all times. A t high altitudes, for example, low v a p o r - p r e s s u r e oils
o r g r e a s e s tend to evaporate, leaving harmful residues on the surfaces.
Excessive amounts of lubricating fluids could cause e x t r e m e power
l o s s e s and ultimate failures because of their heating and braking
effects. Another serious problem associated with a lubricant i s that i t
often c a r r i e s contaminants to the sliding p a r t s . As the surfaces b e a r
on each other, particles a r e abraded away and tend to be c a r r i e d by
the lubricant, which adds to the w e a r problem. Filtering of the lubri-
cant is possible, of course, but a m o r e complex s y s t e m results.
It i s apparent that a lubricating s y s t e m is quite n e c e s s a r y when
near infinite life is required. Quite frequently, however, s y s t e m s a r e
designed in which the expected life need b e only a few hours. Many
instrumentation devices, for example, a r e designed with a n operating
life of only 1000 hr. With this in mind, many designers have aban-
doned the lubrication s y s t e m to lower c o s t s and complexity. Unfor-
tunately, very little information exists f o r predicting w e a r l i f e f o r
sliding and rolling surfaces in the absence of a lubricant.
This paper has been p r e p a r e d to establish a m e a n s of predicting
wear l i f e f o r a special type of sliding and rolling s u r f a c e ; namely,
those surfaces of nonlubricated, fine-pitch, precision, instrument
spur g e a r s . Many nonlubricated g e a r t r a i n s have been built on an
intuitive basis, a s to the size of g e a r s and m a t e r i a l s to use to
- 2 -
minimize wear. That is, in the absence of specific g e a r formulae to
accurately predict life, m o s t design work has been done by trial and
error. Often a change in m a t e r i a l s would improve the life of a g e a r
t r a i n by many o r d e r s of magnitude.
Many m a t e r i a l combinations a r e conceivable in a g e a r train, but
only the most popular combinations of steel, aluminum, and delrin w i l l
be considered in this paper. Delrin is a stabilized f o r m of nylon which
has good dimensional stability. Gears made of various other plastics
have been used in nonlubricated gear t r a i n s and appear to w e a r quite
well. The chief disadvantage to plastics, however, is their poor di-
mensional stability, low strength, and low elastic modulus in
comparison with metals.
- 3 -
J P L Technical Memorandum No. 33-139
11. FRICTION, SURFACE DAMAGE, AND WEAR
In the sliding of two surfaces over each other, there appears to be
two major experimental observations. One observation i s that the a r e a
of contact between two surfaces is v e r y small. Although the techniques
of grinding and polishing have advanced to the point w h e r e s u r f a c e

0
finishes within 100- to 1000-A units a r e not unusual, intimate contact
is still anticipated since the range of molecular attraction is only a few
Angstroms. The a r e a over which the surfaces a r e within molecular
range will, even for carefully p r e p a r e d s u r f a c e s , be quite small. The
contact p r e s s u r e s a t these s m a l l contact a r e a s a r e high enough to
cause plastic deformation of the surfaces.
The second observation i s that the sliding speeds a t which typical
sliding m e m b e r s a r e used cause the surface t e m p e r a t u r e s to r i s e to
v e r y high values. When one body slides over another, some of the
work done against the frictional f o r c e i s liberated as heat between the
surfaces. This heat i s then c a r r i e d away f r o m the surfaces by con-
duction, convection, and radiation. Quite primitive calculations,
however, indicate that v e r y high surface t e m p e r a t u r e s a r e attained
even with moderate loads and speeds. This high t e m p e r a t u r e tends to
promote plastic flow and w e a r , caused by the softened surface.
A third item which needs investigation i s the type of interaction
between the moving surfaces and the physical changes which occur in
them during sliding. One of the m o s t striking conclusions drawn f r o m
extensive studies of the friction of metals is that the magnitude of the
frictional f o r c e and the extent and type of surface damage caused by
sliding a r e determined p r i m a r i l y by the relative physical properties
-4-
1 JPL Technical Memorandum No. 33- 139
I of the two sliding surfaces. Specifically, the behavior is quite de-
pendent upon the relative hardness of the two s u r f a c e s and, if the
sliding speeds a r e high, upon their relative softening o r melting points.
1 Sliding friction can be broken down into two main types:
1. Hard s u r f a c e sliding on a soft one

I 2. Surfaces of s i m i l a r hardness sliding on e a c h other
I Many t e s t s have been conducted, but j u s t the r e s u l t s of one t e s t w i l l b e
indicated here. Bowden and Tabor

1
performed experiments with the
1 following results: In Case 1, the h a r d s u r f a c e sliding on a softer one
r e s u l t e d in high w e a r in the soft surface, as expected, and a coefficient

I of friction of about 0. 9. A groove w a s dug out of the softer m a t e r i a l
I and v e r y little w e a r was perceptible on the h a r d e r material. In Case 2 ,
with s i m i l a r s u r f a c e s sliding together, the damage w a s profound f o r
both p a r t s , and the coefficient of friction was considerably higher at
1. 2. Although no actual d e g r e e of w e a r w a s given f o r these t e s t s , the
r e s u l t s a r e conclusive in that the coefficient of friction w a s consid-
e r a b l y higher for s i m i l a r surfaces than f o r d i s s i m i l a r surfaces.
The friction force, however, is unique among the f a c t o r s involved
in rubbing. I% u s e d ta be generally acknowledged that w e a r and f r i c t i o n

w e r e directly related. However, t h e r e a p p e a r s to be a mounting
number of careful experiments on the subject that demonstrate that
this cannot be generally true. F o r example, t h e r e can be a high f r i c -

2
tion f o r c e with a v e r y low amount of wear, and vice v e r s a . Whittaker
1
F. P. Bowden, and D. Tabor, The Friction and Lubrication of Solids.
(Oxford, Clarendon Press, 1954) pp. 78-79.
'E. J. W. Whittaker, "Friction and Wear, Nature, 159 (1947) p. 54.
- 5 -
and Savage3 have shown, generally f r o m typical w e a r data, that not
m o r e than 1% of the frictional work could have been absorbed by r e -
moving the worn-off m a t e r i a l directly, and, in general, the actual
proportion was v e r y much l e s s than 1%. Thus, a correlation between
friction and wear i s not necessarily to be expected.
Friction does affect w e a r indirectly through the intermediate
factor of temperature. If high friction i s present, the m a t e r i a l s may
tend to adhere to one another m o r e easily because of elevated t e m p e r a -
ture, thus adding to the w e a r problem. In addition, the elevated t e m -
p e r a t u r e s may change the h a r d n e s s of the m a t e r i a l s , which would have
a m a r k e d effect on the wear c h a r a c t e r i s t i c s exhibited by two m a t e r i a l s
in operation.
A factor which should be called to mind when attempting to r e a c h
some conclusions about w e a r f r o m friction i s the relationship between
the coefficient of friction and the coefficient of adhesion. It has been
shown that a near l i n e a r relationship holds between friction and ad-
hesion. This point is mentioned since the amount of cold welding, and
thereby wear, i s directly related to the adhesion coefficient of the
materials. These observations support von M i s e s ' relationship that
u2 = 0. 3v2 - 0. 3 for plastic flow under combined n o r m a l and tangential
stresses. A plot of the coefficient of friction and the coefficient of a d -

4
hesion f o r relative motion between s t e e l and indium i s shown in Fig. 1.
The curve i s essentially the s a m e f o r m o s t metals.
3R.H. Savage, "Graphite Lubrication, Journal of Applied Physics,

19 (1948) p. 1
4F. P. Bowden and D. Tabor, The Friction and Lubrication of Solids.

(Oxford, Clarendon, P r e s s , 1954) p. 313.
- 6 -
JPL Technical Memorandum No. 3 3 - 1 3 9
I
- THEORETICAL CURVE p e r 0.3 u Z - 0.3
I I
DATA
a
i
-
0
I-
;
LL
LL
0
I-
/
2
w
0
LL
LL
w
0
0
COEFFICIENT OF ADHESION, u
8 IO
.
Fig. 1. Comparison of friction and adhesion f o r steel on indium
- 7 -
A w e a r equation has been proposed by Archard5 and a few others.
The equation is a s follows:
v = KW/P
where:
v = volume l o s s p e r unit distance of sliding
K = wear constant
W = normal load
P = flow p r e s s u r e of the m a t e r i a l s
The W / P t e r m i s generally considered to determine the r e a l a r e a of
contact between the sliding s u r f a c e s . The constant K is related to the
nature of the w e a r p r o c e s s itself. Experiments by Spurr' in 1955 have
shown that this equation gives the r a t e of w e a r of small flat s a m p l e s of
w a x loaded against a rotating disk, provided the variation of P with
s u r f a c e t e m p e r a t u r e is considered. Additional experiments indicated,
a s one would expect, that the s u r f a c e finish of the disk had a very
m a r k e d effect on the value of K. A few papers have been published on
the effect of surface roughness on w e a r in the absence of lubrication.

7
Brownsdon showed in 1936 that w e a r did i n c r e a s e with increasing
5J. F. Archard, !'Contact and Rubbing of Flat Surfaces, 'I Journal of

Applied Physics, 24 (1953) p. 981.
6
R. T. Spurr, "Creep and Static Friction, B r i t i s h Journal of Applied
Physics, 6 (1955) p. 402.
7H. W. Brownsdon. "Metallic W e a r . I -

Journal of the Institute of
Metals, 18 (1936) p. 15.
surface roughness. L a t e r Taylor and Holt 8 found that w e a r w a s ap-
proximately proportional to the surface roughness a s determined by a
profilometer. Thus, it appears that surface finish may be an important
variable in wear.
Mr. Rabinowicz', of Massachusetts Institute of Technology (MIT),
has done a considerable amount of work i n the specific a r e a of w e a r ,
and has expounded s e v e r a l theories that laboratory investigations tend
to confirm. One theory is that the same materials always w e a r i n
f r a g m e n t s of the s a m e characteristic size. The particle size is
r e l a t e d to the amount of elastic energy the m a t e r i a l can a b s o r b before
it yields. Calculations show that the fragment s i z e has a lower limit.
This is because the shattering of the m a t e r i a l t r a n s f o r m s the elastic
energy into surface energy. Since t h e r e is a l i m i t to the amount of
energy a given m a t e r i a l can absorb, t h e r e i s also a l i m i t to the f r a g -
m e n t size. The energy required to f o r m a w e a r particle at the surface
m u s t be provided by the elastic energy of the material in the immediate
vicinity. The minimum fragment s i z e calculated on the b a s i s of elastic
energy should be, f o r a specific material, closely related to the
c h a r a c t e r i s t i c s i z e of i t s w e a r particles.
Figure 2 i l l u s t r a t e s the relationship between calcuiated minimum
particle s i z e and actual a v e r a g e particle size?' A s expected, the actual
8
R. H. Taylor and W. Holt, "Effect of Roughness of C a s t Iron Brake
D r u m s in Wear Tests of B r a k e Linings, Journal of R e s e a r c h ,
National Bureau of Standards, 27 (1941)p. 395.
'Ernest Rabinowicz, "Wear, It Scientific American, 206 (January, 1962)

pp. 127-136.
- , p.
"bid. 132.
- 9 -
1000
0
D 0
0
0
100 ---T---
v)
c 0
2
.-0
E
w
N
cn
W
W IO
a
(r
W
2
0
W
LL
3
cn
a
W
z
I
- WHERE POINTS WOULD FALL IF ACTUAL

SIZE EQUALLED CALCULATED MINIMUM SIZE
0 DATA FOR VARIOUS MATERIALS
0.1 ~~
~
0.I I 10 100
CALCULATED MINIMUM S I Z E , microns
Fig. 2. Comparison of calculated particle size and m e a s u r e d .

average particle size
- 10 -
I JPL Technical Memorandum No. 3 3 - 139
1 a v e r a g e particle size is slightly l a r g e r than the calculated minimum
I particle size. In a s y s t e m i n which w e a r particles of a certain s i z e
are being generated, the surfaces take on a corresponding roughness.
I The height of the hills and valleys on the surface will b e roughly the
s a m e as the d i a m e t e r of the average w e a r particle. A s particles a r e

I worn away, the surface finish tends to r e m a i n the s a m e .
1 If the theory proposed by Rabinowicz is c o r r e c t , which preliminary
investigations s e e m to b e a r out, it can account for the w e a r phenom-
m enon as we o b s e r v e it. At first, there s e e m s to b e a period of v e r y
rapid w e a r followed by a longer period of time a t a much reduced
i w e a r rate. The period of rapid wear could be caused by the rough
I s u r f a c e s working together during the first hours of operation.
particles a r e generated, however, the surfaces become polished to a

As
M finish roughly equivalent to the particle size, and the w e a r r a t e de-
clines to a f a i r l y constant rate. W e a r would then continue a t this r a t e

I until the p a r t is "worn-out" because of excessive clearance o r play.
I This w e a r would b e the abrasive type in that particles would be worn
off by the sliding action.
I In addition, Mr. Rabinowicz

11
h a s done some r e s e a r c h on a l e s s e r
known type of w e a r called adhesive wear. When two smooth surfaces
1 slide over each other, patches of one surface a d h e r e to the other and
I a r e pulled away. Adhesive w e a r results f r o m the strong f o r c e s estab-
lished between a t o m s that come into intimate contact with one another.
I When a bond is made between two atoms, t h e r e is a c e r t a i n likelihood
I -
"Ibid., p. 129.
1
I -11 -
JPL Technical Memorandu
I
I
' I
%-
that, when the contact i s broken, the b r e a k w i l l not occur a t the I

original boundary. Instead the b r e a k will o c c u r within the surface
l a y e r s of one of the m a t e r i a l s and an adhesive w e a r fragment w i l l b e

I
produced. Moreover, this adhesive w e a r o c c u r s in two m a t e r i a l s I
m e r e l y contacting a s well as sliding.
In an experiment p e r f o r m e d by Rabinowicz at the University of I

Cambridge, a copper rod with a hemispherical end w a s p r e s s e d
against a steel surface. The rod was p r e s s e d with a 2-lb f o r c e p e r -

I
pendicular to the s t e e l surface with no tangential motion allowed. By I
radiation-tracer techniques, i t was found that g of copper had
-1 0
g of s t e e l had t r a n s f e r r e d to the
t r a n s f e r r e d to the steel and that 1 0 1
-6
copper. It w a s also found that a s much a s 1 0 g could be t r a n s f e r r e d
if the rod were p r e s s e d against the plate a t a n angle, but not allowed
I
to slide. Although a l l the laboratory conditions a r e not available s o
1
that one could determine these unit s t r e s s e s , the above information
gives a useful description of adhesive wear in a qualitative manner. I

Adhesive w e a r can b e reduced, but it cannot be eliminated.
surfaces operate, f o r example, m a t e r i a l is worn away, leaving clean

As
I
surfaces to contact the mating parts. If these surfaces a r e running in
I
air, they usually have a chance to oxidize slightly before coming into
contact again. T h e s e impurities impede the adhesion of the s u r f a c e s I

and the amount of adhesive w e a r i s reduced.
however, these s a m e surfaces a r e m o r e susceptible to adhesive w e a r

In a vacuum environment,
I
in that they do not oxidize as they do in air. A clean b a r e s u r f a c e is
I
exposed during operation because of abrasion and these s u r f a c e s show
excellent tendencies to w e a r adhesively. I

1
- 12 -
I
I A simple formula has been developed by Archard" f o r adhesive
I wear. This formula a s s u m e s a certain probability k that intimate con-
t a c t w i l l r e s u l t between two contacting surfaces:

I v = kW1/3p (2)
I where:
v = volume of w e a r , m m
3
I W = load on sliding surfaces, kg

1 = sliding distance, m m
I p = penetration h a r d n e s s of the softer contacting surface,
2
I kg / m m
k = probability of intimate contact
I Thus, the w e a r is directly proportional to the load and sliding distance,
and inversely proportional to the penetration hardness, according to

I Archard.
I The above equation, however, makes no r e f e r e n c e to s u r f a c e
finish o r sliding velocity. A s Mr. Rabinowicz13 has pointed out, the
I s u r f a c e finish s e e m s to affect only the e a r l y w e a r , but, after this
period, w e a r r a t e s tend to stabilize because of the generation of

I constant-size w e a r particles. Rabinowicz has a s s u m e d cubic particles
2 3
I and equated the volume energy of the particle, u d /2E, w i t h the
s u r f a c e energy of the particle, 6yd

2
.
r
This resulted i n a d i a m e t e r f o r
I the particle, d = 12Ey/crr2. Experimental r e s u l t s ( s e e Fig. 2) confirm
I
I 13
E r n e s t Rabinowicz, "Wear, Scientific American, 206 (January, 1962)
p. 135.
I
I - 13 -
I
this assumption, and it s e e m s logical, therefore, to a s s u m e that s u r -
I
face finish has v e r y little effect on the w e a r r a t e after the initial run-in
I
period. Research by Archard also tends to indicate that sliding velocity
does have an effect on w e a r , but this is ordinarily slight, and thus not I
accounted f o r in the expression he proposed. He maintains that the
effects of sliding velocity on w e a r a r e negligible if this sliding velocity

1
i s l e s s than 500 ft/min.
I
In a wear problem, it is generally the depth-of-wear r a t h e r than
the volume that i s important. In addition, the penetration h a r d n e s s of 1

a given m a t e r i a l i s not usually a s well-known a s the yield strength,
although a correlation does exist between the two. Rewriting Eq. (2)
I
in English units, considering penetration depth in t e r m s of yield
I
strength, and finding the depth of w e a r r a t h e r than volume of wear,
Archard obtained the following equation: I

h = kW1/9As
where:
(3)
I
h = depth of w e a r , in.
W = load, l b
I
1 = sliding distance, in.
2
I
A = surface area, in.
s = yield strength of softer material, lb/in.

2 I
k = probability of intimate contact
The w e a r rate of the h a r d e r m a t e r i a l is less than the s o f t e r m a t e r i a l

I
by the following expression: I
I
I
- 14 - I
~~
I JPL Technical Memorandum No. 33- 139
I (4)
I where:
= depth of w e a r of the softer material, in.

1 hl
h2 = depth of w e a r of the harder material, in.
2
=
I s1
sz =
yield strength of the softer material, lb/in.
yield strength of the harder material, lb/in.

2
I The equations show the effect that changing the yield strength, load,
a r e a , o r sliding distance has on the depth of wear.

I In determining the amount of a b r a s i v e wear of a sliding system,
1 Eq. (2-4) a r e applicable provided the probability coefficient k is r e -
placed by an abrasive constant K. The abrasive constant is on the
I o r d e r of to 10
-3
, while the probability coefficient for
-4
adhesive
w e a r is on the o r d e r of 10 to in a normal E a r t h atmosphere.

I Thus, in the atmosphere, one would expect the wear of a nonlubricated
1 s y s t e m to be largely of the abrasive type, with a v e r y minute portion
being worn away by adhesion. In the h a r d vacuum of space, however,
I where p r e s s u r e s a r e on the o r d e r of 1 0 - l 2 to 10
-1 6
m m of r m rcury,
the likelihood of adhesion w i l l increase, and one would expect m o s t of

I the wear to occur because of adhesion and a s m a l l amount because of
I abrasion. It s e e m s unlikely that the magnitude of abrasive wear would
change significantly f r o m the atmosphere to v e r y low p r e s s u r e s since
I the s a m e general w e a r - p a r t i c l e size would be generated. It is only the
relative amount of a b r a s i v e wear with r e s p e c t to adhesive wear that
I would change f r o m the atmosphere to a h a r d vacuum.
Although the rolling and sliding action of s u r f a c e s gives both

I abrasive and adhesive wear, the quantities would s e e m to be additive
I
1 - 15 -
I
if considered separately. In predicting the life of a nonlubricated
I
system, the designer is not usually interested in the mode of failure,
but simply the length of time to failure. Since the w e a r quantities a r e

I
additive, the adhesive constant k and the a b r a s i v e constant K could be I
added together to f o r m one constant which would yield the w e a r r a t e
caused by combined abrasive and adhesive action. Equation (3) could

I
be rewritten as:
h = K1W1/9As (5)
I
where: I
K1 = constant for adhesive and a b r a s i v e w e a r , K t k
Since W is the applied load and A i s the surface a r e a , the quantity W/A
I
is a s t r e s s value, and the quantity Sc can be substituted f o r W / A to
find the depth of w e a r i n t e r m s of s t r e s s :

I
h = K1Scl/9s (6) I
where:
S = compressive s t r e s s , lb/in.
2 I
C
Deviating to some extent f r o m Rabinowicz' theories, Mr. Maschmeyer

14
I
claims that wear f a i l u r e s a r e caused by exceeding the endurance l i m i t s
of the material a t which time particles begin to flake f r o m the surface.

I
Maschmeyer expounded this theory b a s e d upon w e a r patterns he had
I
observed in the operation of gears. F i r s t , t h e r e s e e m e d to b e a run-in
phase in which slight w e a r w a s observed. Sharp edges and s u r f a c e I

imperfections w e r e burnished smooth and residual c o m p r e s s i v e
s t r e e s e s were induced in the microscopic surface of the contact area.

I
14A. H. Maschmeyer, "Wear Life of Aluminum G e a r s , 'I Product
I
Engineering, 27 (September, 1956) p. 162.
I
- 16 - 1
I JPL Technical Memorandum No. 3 3 - 1 3 9
I This run-in period w a s followed, Maschmeyer continued, by a phase i n
I which the contact a r e a w a s mechanically stabilized and no appreciable
w e a r took place. Finally, the endurance l i m i t w a s reached, and the
I contact surface fatigued, increasing the w e a r rate sharply.
Several authors, notably Buckingham,15 a g r e e that w e a r is a fatigue

1 phenomenon. While this theory may be t r u e f o r a well-lubricated g e a r
I s y s t e m in which the effects of abrasion and adhesion a r e greatly r e -
duced, this w r i t e r believes that surface fatigue is a definite life
I problem, but that abrasion and adhesion w i l l play an increasingly i m -
portant r o l e since precision gear t r a i n s a r e continually being designed

I without lubrication. F a c t o r s such as adhesion w i l l become c r i t i c a l
B w i t h no lubrication and at reduced p r e s s u r e s .
sonable to attribute all w e a r to surface fatigue.

It does not s e e m r e a -
I Experience tells us that lubrication definitely reduces wear, but
m o s t people will be willing to admit that surface fatigue i s not depend-

I ent upon intimate contact between sliding o r rolling surfaces; yet the
I m a i n function of a lubricant i s to prevent intimate contact.
into the problem a bit deeper, i t may be logical to a s s u m e that rolling

Thinking
I s u r f a c e s would be m o r e likely to f a i l because of surface fatigue than
because of sliding surfaces. This conclusion can b e reached f a i r l y
i easily when one thinks of an abrasive particle being r e l e a s e d i n both
systems. In the sliding system, the particle would be c a r r i e d along

I and gouge m a t e r i a l out a s i t proceeded. In the rolling surfaces, how-
1 ever, the particle would m e r e l y be embedded in the surface with v e r y
1 1c
A 2
E a r l e Buckingham, Analytical Mechanics of G e a r s . (New York,
McGraw-Hill, 1949) pp. 527-529.
I
I - 17 -
I
little w e a r occurring. The sliding surfaces would s e e m m o s t likely to I
wear out, whereas the rolling surfaces would tend to fatigue.
In this Chapter, some general theories on w e a r have been outlined.

I
Wear has been with us since the beginning of time, but the technology of 1
wear i s still in i t s infancy. Many theories a r e offered with little o r no
proof o r experimentation to substantiate them. At the onset, i t would I

appear that the work being done by Rabinowicz a t MIT is probably the
m o s t advanced in the country. A considerable amount of work in the

I
field of wear is going on a t MIT and m o s t of the theories, although
I
quite recent, s e e m to support physical observations far better than
m o s t theories presented heretofore. If these theories a r e c o r r e c t , I

wear w i l l occur during the entire operation because m a t e r i a l i s being
abraded and pulled away f r o m the surface. As Maschmeyer and o t h e r s

1
point out, however, the number of s t r e s s cycles reaches the endurance
I
limits of the m a t e r i a l s and the w e a r r a t e r i s e s sharply because of g r o s s
removals of the tooth profiles. Thus, if the c r i t e r i o n that d e t e r m i n e s I

depth-of-wear i s v e r y low, the surfaces will fail by pure w e a r , a c c o r d -
ing to Rabinowicz. If the c r i t e r i o n is selected higher, the surfaces

I
w i l l m o s t likely fail because they exceed the surface fatigue l i m i t s
r a t h e r than because of adhesive and a b r a s i v e wear.

I
I
I
1
I
I
- 18 -
I
L -
I III. SHORT DISCUSSION ON GEAR-LOAD CALCULATIONS~6
I In the analysis of w e a r on specific surfaces, namely nonlubricated
I spur gears, it would s e e m imperative that an accurate estimate of g e a r
loads and s t r e s s e s b e determined. This chapter gives a brief h i s t o r -
I i c a l background of g e a r calculations and points out the lack of precision
that exists in the field.

I In 1879, John Cooper made an investigation of the strength of g e a r
I teeth and found that t h e r e w e r e then in use about 48 well-established
r u l e s f o r working strengths of gear teeth. These rules differed, in
I the e x t r e m e c a s e s , by about 5000/. In a l a t e r study by W i l l i a m Harkness
in 1886, an examination of the l i t e r a t u r e dating back to 1796 indicated

I that, according to the constants and formulae used by various authors,
I t h e r e w e r e differences of 1500% in the calculated power capacity of a
given gear set. In 1892, Lewis presented a paper to the Engineer's
I Club of Philadelphia entitled "Investigation of the Strength of G e a r
Teeth", which introduced a formula for the load capacity of gears:

I
i where:
W T = total load, l b
I S = safe working s t r e s s , lblin.

2
F = face width, in.

1 Pn
= c i r c u l a r pitch, in.
1 y = tooth f o r m factor
I 16
E a r l e Buckingham, Analytical Mechanics of Gears. (New York,
McGraw-Hill, 1949) pp. 385-389.
I
-
I - 19
I
Values of y and S w e r e tabulated by Lewis f o r various m a t e r i a l s and I
speeds. Although the Lewis f o r m u l a is generally accepted in industry
today, some arguments have a r i s e n as to the meaning of the total load.

I
E r r o r s on gear tooth profiles, caused by elastic deformation under I
load o r by inaccuracies of production, a c t to change the relative
velocities of the rotating m e m b e r s . This varying velocity of the I

rotating m e m b e r s r e s u l t s in a varying load cycle on the gear teeth.
The amount of load variation depends largely upon the effective m a s s e s

I
of the rotating gears, amount of effective errors-in-action, and the I
speed of the g e a r s . It i s not unusual for the dynamic load on a gear
tooth to be several times the static o r transmitted load. The problem, I

however, i s in accurately determining the amount of this dynamic load.
In the early 19001s, Buckingham generated a s e t of formulae which

I
tended to be a v e r y popular method in calculating dynamic loads and I
which, in turn, could be used in conjunction with the Lewis formula.
Buckingham did a considerable amount of gear analysis and testing at I

MIT, and his dynamic load formulae a r e the m o s t widely used in in-
dustry today. His formulae a r e generally acknowledged to give loads
I
considerably higher than the actual loads, thus incorporating a safety
I
f a c t o r in a rather inexact science.
In the past 20 y e a r s , however, numerous formulae based upon 1

empirical and analytical considerations have been generated in an
attempt to predict m o r e closely the actual dynamic loads. These

1
methods vary in magnitude by approximately 1000% f o r the amount of
I
I
I
- 20 -
I
power that can b e transmitted by a given g e a r set. Figure 3 illustrates

17
this variance f o r s o m e of the m o r e popular methods.
Although g e a r s have been used f o r many centuries, this chapter
shows that the state of the art in gearing has not p r o g r e s s e d to the
point at which a c c u r a t e g e a r calculations can be made. It appears,

however, that a valid w e a r analysis can b e performed if a good c o r r e -
lation exists between the actual loads and the calculated values of load.
7Darle W. Dudley, G e a r Handbook. (New York, McGraw-Hill, 1962),

pp. 14-31.
- 21 -
I
I
I
I
I
I
1
I
1
I
I
I
I
I
I
PHERIPHERAL VELOCITY, cm/sec 1
Fig. 3. Comparison of dynamic loads
I
I
- 22 - I
-
~ ~~~
IV. FUNDAMENTALS OF INVOLUTE SPUR GEARS
The wear analysis of g e a r s first necessitates a c l e a r under-
standing of the involute spur gear. A brief description of the involute
c u r v e and g e a r formulae a r e presented i n o r d e r to make a thorough
analysis of the wear of spur gears.
At the p r e s e n t time, the involute curve is used a l m o s t exclusively
for spur g e a r tooth profiles. The involute c u r v e is the curve that is
d e s c r i b e d by the end of a line that is rolled without slip f r o m the c i r -
cumference of a c i r c l e called the base circle. The length of the
generating line that i s rolled f r o m the b a s e c i r c l e is the radius of
c u r v a t u r e of the involute curve a t any particular point on the curve.
The involute curve m e e t s all the requirements for a gear tooth profile,
the m o s t important being the transmission of uniform r o t a r y motion.
In o r d e r to t r a n s m i t uniform rotary motion, the values of momentary
pitch r a d i i a s defined by the instant center m u s t r e m a i n in the s a m e
proportion to each other f o r a l l operating positions of the contacting
profiles .
An advantage of the involute curve is its ability to t r a n s m i t uniform
motion even though the center distance be varied. If one involute, for
example, rotating at a uniform rate, acts against another involute p r o -
f i l e of the s a m e pitch and p r e s s u r e angle, it w i l l t r a n s m i t a uniform
angular motion to the second profile r e g a r d l e s s of the distance between
the c e n t e r s of the two b a s e circles.
- 23 -
J P L Technical Memorandum No. 33- 139
I
Figure 4 shows two involute curves with the generating l i n e s at I
equal angular intervals. l 8 The p a r t bc on one involute c u r v e comes
into contact with hi on the second involute curve. Since bc is much

D
n e a r e r to its b a s e c i r c l e than is hi to i t s b a s e circle, the a r c bc is I
much shorter than the a r c hi. The two profiles m u s t slide against
each other a distance equal to their difference i n length in o r d e r to I

have uniform r o t a r y motion. The length cd is s t i l l much s h o r t e r than
i t s mating section ij, but the amount of sliding will not b e as l a r g e as

1
I
with the previous sections of the mating profiles because the difference I
in lengths i s not as large. The sections de and jk a r e a l m o s t equal in
length, so very little sliding occurs in this portion of the gear profiles. m'
It should be noted that the profile ef on the f i r s t involute curve is
slightly longer than its mating section kl on the second involute, and
i
the small amount of sliding will now a c t i n the opposite direction. The I
r a t e of sliding between two involute curves acting against each other i s
proportional to the distance f r o m the point of contact to the instant 1

center o r pitch point. The sliding velocity s t a r t s quite high, reduces to
z e r o at the pitch point, then i n c r e a s e s again to a maximum value i n the

1
opposite direction. The only point at which t h e r e is p u r e rolling action
I
is the pitch point; a t all other portions of the profile s o m e sliding action
occurs. I
The sliding velocity at any point on a pair of involute gear teeth
can b e derived. The magnitude of the sliding velocity will be the

1
difference i n velocities of the ends of the generating lines of the
1
8 E a r l e Buckingham, Analytical Mechanics of Gears. (New York,
McGraw-Hill, 1949) p. 6.i.
I
I
- 24 -
I
1 Fig. 4. Two mating involute profiles
I
1 - 25 -
involute curves as they pass through the line of action. The angular
I
velocity of these generating lines w i l l be the s a m e a s the angular
velocity of the g e a r s themselves since the profiles and the g e a r blank

1
a r e in one rigid piece. The actual sliding velocities w i l l be the I
products of these relative angular velocities and the lengths of the gen-
erating lines o r radii of curvatures of the involute curves. If we let:

I
-
- angular velocity of driving gear, r a d / m i n
angular velocity of driven g e a r , r a d / m i n

1
*2 =
n = speed of driving gear, r e v / m i n I
v = pitch line velocity of gears, f t / m i n
vs
-
- sliding velocity, f t / m i n I
- pitch radius of driving gear, in.
R1 -
- pitch radius of driven gear, in.
R2 -
c = center distance, in.
Rbl = base c i r c l e radius of driving gear, in.

base c i r c l e radius of driven gear, in.
Rb2 =
b = p r e s s u r e angle, deg
- radius of curvature of driving gear at r l , in.
Rcl -
Rc2 -- radius of curvature of driven gear at r
2'
in.
r = radius of driving g e a r tooth at point in question, in.

1
r = radius of driven gear tooth at point in question, in.
2
The pitch line velocity can b e written
V = 2aR1n/12 = R 1 w 1 /12
f r o m which the angular velocity f o r the driving g e a r is found
w1 = 12V/R1
- 26 -
L
and s i m i l a r l y for the driven gear
O2 - 12V/R2
The angular velocity of the driven g e a r is r e l a t e d to the driving g e a r by
The sliding velocity is
By geometry
Rcl t Rc2 = C sin0
These expressions are combined and simplified to yield the sliding
velocity at a given radius
As stated previously, the sliding velocity is continually changing as

the g e a r teeth operate. F o r the w e a r analysis, some s o r t of a n a v e r -
age velocity would be appropriate. Since the sliding velocity is
nonlinear with r e s p e c t to the a r c length traveled, an a v e r a g e sliding
velocity can b e obtained by integrating under the sliding velocity v e r s u s
angular position c u r v e and dividing by the angular a r c through which the
sliding velocity takes place. Before this can b e done, however, the
length of the contact a r c m u s t be established.
The a r c of action i s the a r c through which one tooth t r a v e l s f r o m the
time it first makes contact with its mating tooth until it c e a s e s to b e in
- 27 -
contact. The arc of action is often separated into the arc of approach
and the a r c of r e c e s s . The a r c of approach is the a r c through which
the tooth moves f r o m the point at which it f i r s t comes into contact with
i t s mating tooth until contact is made at the pitch point. The a r c of
r e c e s s is the a r c through which the tooth moves f r o m contact at the
pitch point until it c e a s e s to be in contact with its mating tooth. The
following equations can b e written:
Pa = (4-
where:
Pa = a r c of approach
P, = a r c of r e c e s s
Rol
= outside radius of driving gear, i n
RoZ = outside radius of driven gear, i n

The average sliding velocity can b e found by dividing the a r e a
under the velocity curve by the a r c length. In Eq. (8) for sliding
velocity, i t can b e seen that the conditions of sliding velocity a r e m o s t
s e v e r e when rl = Rbl and r 2 -- Ro2 on the approach eide and when

rl - Rol and r
2 = Rb2 on the r e c e s s side. The a v e r a g e sliding
velocity then becomes:
- 28 -
S = Rbl / (Jm ) SrliR1 - R2 sin 0

Il=Rbl
V(l/R1 t l / R 2 )
s rl =R
rl =R1
01
V(l/R1 t l / R 2 ) ( , / m - R1 sin O))drl
which reduces to:
S
= Rbl V(l/R1 t '/R2)/2(J~02 2 - Rb22 - R2 s i n s>
- 2R12 sin b t 2 R 1 s l sin fJ

1
- 2R1Rol sin 0 t 2R12 sin fl]
With Eq. (1 1) the average sliding velocity can b e calculated f o r a
given g e a r set.
In addition to the sliding velocity, the load on a g e a r tooth is an
important w e a r consideration. In high-speed applications, high pitch
line velocities develop v e r y high dynamic loads while the g e a r s a r e
- 29 -
transmitting extremely low values of torque. The load on the g e a r
teeth consists of a transmitted load plus a dynamic load and can be
e x p r e s s e d as follows:
where:
total load, lb/in.
transmitted load, lb
dynamic load, l b
F = face width, in.
Since the transmitted load is equal to the torque divided by the pitch
radius, Eq. (12) c a n be rewritten:
WT = T/RFtWd/F
where:
T = transmitted torque, in-lb
R = pitch radius, in.
The magnitude of the dynamic load Wd is a function of the pitch
line velocity, effective errors-in-action, m a s s effects, and p r e s s u r e
angle. Errors-in-action a r e any e r r o r s in the tooth profile that cause
the contact action of the g e a r s to differ slightly f r o m the pure conjugate
action. This e r r o r -in-action causes changes in the relative angular
velocity of the mating g e a r s and excites a varying load cycle on the
gear teeth. The amount of error-in-action has been defined by the
American Standards As sociation: 1 9
19American Standards As sociation, “Inspection of Fine Pitch G e a r s ,

American Standards Association, Specification B6. 11 -1 951
‘(1951)pp. 5-9-
- 30 -
~
J P L Technical Memorandum No. 3 3 - 139
= total composite e r r o r t 1 / 2 tooth-to-tooth e r r o r (14)

El
According to Buckingham, 2o there are two distinct load s u r g e s
during each tooth engagement. The first s u r g e is caused by s m a l l
changes in the relative velocity of either gear and usually o c c u r s
during the f i r s t portion of engagement. This s u r g e is a n acceleration
load, since any deviation f r o m pure involute action results in a c a m -
ming action of the g e a r teeth. As the succeeding p a i r of teeth c o m e s
into contact, the acceleration load becomes z e r o on the f i r s t set of
teeth. However, as the g e a r teeth come together for a second time
during one mesh, a second s u r g e o c c u r s called the impact load. This
impact load and the acceleration load constitute the dynamic load. If
the i n e r t i a load is small, as is usually the c a s e f o r instrument gearing,
the effective m a s s of a g e a r set may b e e x p r e s s e d as:
m = m m /(m + m )
e P g P g
where:
m = effective m a s s of pinion
P
m = effective m a s s of gear
g
The effective m a s s of a n individual s p u r g e a r is the polar moment of
i n e r t i a divided by the radius squared. Since the polar moment of
i n e r t i a equals wRL/2g, Eq. (15) can b e written a s follows:
m e = w w /2g(w t w )
P g P g
where:
w = weight of pinion, lb
P
w = weight of gear, lb
g
2
g = acceleration due to gravity, i n / s e c
2 0 E a r l e Buckingham, Analytical Mechanics of Gears. (New York,

McGraw-Hill, 1949) pp. 427-452.
- 31 -
The m e a n ac.celeration f o r c e f resulting f r o m the acceleration and

m
impact load is e x p r e s s e d by Buckingham as:
m = f a f d / ( f a t fd)
and f o r a 20-deg, p r e s s u r e angle g e a r system:
2
f = O.O012(1/R t 1 / R ) m e V (18)
a P g
fd
and the magnitude of the dynamic load is:
/(1/2E t 1 / E )
P g 3 (19 )
Now Eq. (20) could b e substituted into Eq. (13) to yield a n expression
f o r the total load p e r inch of face width.
The compressive s t r e s s developed on the g e a r teeth can b e de-
rived using the Hertz equation f o r s t r e s s e s on friction d i s k s 2 l The
maximum compressive s t r e s s f o r two disks in contact is:
SL = 0. 35WT(1/R1 t l / R 2 ) / ( l /E1 t 1 / E 2 ) (21)

C
where:
2
sC = maximum compressive s t r e s s , lb/in.
- radius of one disk, in.
R1 -
-
R2 - radius of other disk, in.
-- modulus of elasticity of one disk, lb/in.
2
--
2
modulus of elasticity of o t h e r disk, lb/in.
E2
- total load, l b / i n .
wT -
-
"Ibid., p. 528.
- 32 -
I F o r spur gears:
I R1 = D sin 0 / 2
P
R2 = D sin 0 / 2
1 g
l / R 1 t 1 / R 2 = 2(1/D t 1 / D ) / s i n @
P g
I Substituting these equations into Eq. (21), one obtains the maximum
c o m p r e s s i v e s t r e s s f o r a s e t of spur g e a r teeth
I S = 0. 84J[(Dp t D g ) / D D ][E E /(E t E ) WT/sin 01 (22)
P g P g P g
I
C
where:
I D = pitch diameter of pinion, in.

P
D = pitch diameter of gear, in.
g
I Thus, Eq. (22) gives the maximum c o m p r e s s i v e s t r e s s f o r a given s e t
of g e a r s operating under a certain total load p e r unit face width.
I A s mentioned previously, the error-in-action i s taken a s the
maximum total composite e r r o r plus one-half the maximum tooth-to-

I tooth e r r o r , according to the American Standards Association Specifi-
I cation B6. 11 -1951. There seems to be some doubt, however, that this
is the appropriate value to use. Experience indicates that the
I Buckingham f o r m u l a e yield much l a r g e r values f o r dynamic loads than

22
actually occur when this value for error-in-action is used. Tuplin,
1 of the University of Sheffield in England, has conducted numerous
B t e s t s on actual g e a r s in operation, and has derived his own values for
e r r o r s -in-action, based largely on empirical considerations.
I
I 22W. Tuplin, "Dynamic Loads on G e a r Teeth, Machine Design, 25
(October, 1953) pp. 203-211.
I
I - 33 -
I
Tuplin indicates that an effective e r r o r - i n - a c t i o n should be used,
I
which i s i n some proportion to the c i r c u l a r pitch e r r o r .
effective error-in-action is based on the speed of operation of the

Also, the
I
gears. The speed of operation of the g e a r s is not determined by the I
pitch line velocity alone, but by the natural period of the dynamic s y s -
tem as well, The dynamic system consists of two g e a r s , each rotatable I

about i t s axis and each coupled with the engaged teeth. Thus, the s y s t e m
i s a n angularly vibrating one with two m a s s e s elastically connected. If

I
the gear blanks a r e very s t i f f , the elasticity i s that of the g e a r teeth I
alone. The remaining p a r t s suffer no comparable distortion and may be
regarded a s rigid m a s s e s . If the gear blank i s highly elastic, however, I

the elasticity of the blank m u s t a l s o be taken into account.
The compliance, the i n v e r s e of stiffness, i s used because the

I
compliance of a n assembly of loaded m e m b e r s i s simply the sum of the I
compliances of the components. Tuplin has found the compliance of
two mating gear teeth rigidly "built in" to the m a s s of the g e a r blank. I
Elastic flattening of the tooth surfaces is neglected since it i s very
s m a l l in comparison with other compliances. The expression for the

I
compliance of two mating g e a r teeth a t the pitch c i r c l e is: I
l / k t = 3(1/Ep + 1 / Eg ) (23)
This expression i s nearly independent of tooth thickness because a n

I
i n c r e a s e in thickness i s compensated for by a n i n c r e a s e i n length of
I
the tooth. This expression was derived by e m p i r i c a l investigations
over a wide range of d i a m e t r a l pitches. I

If a tooth i s cut in a very thin r i m , however, the tooth is not
"built in, 'I and extra compliances need consideration. The bending
I
I
- 34 - I
I moment at the root of the tooth causes the p a r t of the r i m below the
I tooth to undergo angular displacement and the tooth h a s m o r e com-
pliance because of this displacement. The compliance of this displace-
1 ment h a s been estimated by a r b i t r a r i l y assuming that the tilt of the
rim under the tooth is caused by radial s h e a r deformation in the r i m

I sections?3 ABCD and EFGH i n Fig. 5. If a s p u r g e a r is taken with a
I load f p e r unit face width applied at the midpoint of the working depth,
a moment is produced about the axis 0 of f ( O . 4 n

~ + 0. 5H). This
I moment is balanced by a n equal shear f o r c e i n the planes BC and EH.
The magnitude of each s h e a r force is:

I f(0. 4pn t 0. 5H)/0. 8pn
I which simplifies to:
f(0. 5 t 0. 625H/p n )
I The s h e a r s t r e s s i n the vertical sections between AD and BC, and
I between EH and F G , is the shear f o r c e divided by the depth of the r i m
(unit width was a s s u m e d when f w a s selected):
I f(0. 5/H + 0. 625/pn )
I Tlii.is, the upw-ard motion of BC reiative to AD and tne downward
motion of EH relative to F G is:
1 (f/G) (0. 5/H t 0. 625/pn) 0. 2pn
where:
1 G = s h e a r modulus of g e a r material, lb/in.
2
I The tilt of the p r i s m BEHC i s this movement divided by one-half the
b a s e CH. The horizontal movement of the tooth at the point of
1 -
231bid., p. 208.
I
I - 35 -
I
I
I
I
I
1
I
--I----
-f
I
0.4 Pn
I
_ - \--
A
I
II I I
1
I I \ /
I I
\
\
H
I /
/ I
I
I
I I
I I
I
\
\
I
\
I I I
I
Fig. 5. Dimensions f o r g e a r teeth and rim
I
I
I
- 36 - I
I JPL Technical Memorandum No. 3 3 - 1 3 9
I application of the load i s the tilt multiplied by 0. 4pn, since B and E
I have no horizontal movement due to this compliance.
pliance for the tooth in the r i m is:

Thus, the com-
1 l / k r = ( l / G ) ( O . l pn / H + 0. 125)
I Taking a as the ratio of r i m depth to c i r c u l a r pitch, H/p,,
expression f o r compliance becomes:

the
.I l/kr = (l/G)(O. l / a + 0. 125) (24)
I In addition to this tilting of the p r i s m under the tooth, the r i m
undergoes some circumferential s t r a i n because of the t r a n s m i s s i o n of
1 the tooth load to the m a s s of the rim. F i g u r e 6 r e p r e s e n t s the c i r -
cumference of a toothed r i m subject to a tangential f o r c e f p e r unit
I width a t AF4 Rotational acceleration is prevented by the tangential
f o r c e f p e r unit width a t B. (The effective depth of the r i m may be

I somewhat l a r g e r than H because of the stiffening effect of the teeth.
+ 0. 2pn o r
I The effective depth i s taken a s H (a f 0. 2)pn. ) If x is the
circumferential movement of A relative to B because of the lengthening
I of a r c AB and the shortening of a r c ACB, the tension force in AB p e r
unit width is the product of strain, effective depth, and the modulus
I of elas ticity:
+ 0. 2)pn E + 0. 2 ) p n E / r
1 (x/AB) (a = (x/€I)(a
The compressive force in a r c ACB i s given by a s i m i l a r expression,

1 and the sum of the two f o r c e s is equal to f:
-6)
I f = ( x / r ) ( a t 0. 2)p E(1/ 6 + 1/2n
n
I
- , p.
241bid. 210.
I
I - 37 -
I
f
Fig. 6. F a c t o r s pertaining to s t r e t c h of r i m
Fig. 7. Relationship of m e a s u r e d tooth-to-tooth e r r o r by center

distance deviation to actual pitch e r r o r of g e a r teeth
- 38 -
1 JPL Technical Memorandum No. 3 3 - 139
1 The compliance of the r i m f o r unit loading a t points separated by an
1 angle 0 is:
1 The average value of compliance for all values of 0 f r o m 0 t o is:
I l / k c = [ r / ( 2 n 2 ) (a t 0. 2)pnE] 0(2n - 8)d0

I = m/(a + 0. 2)pn3E
1 Since 2 n r / p
n
equals the number of teeth, N, i n a spur gear, the c i r -
cumferential compliance f o r the r i m can be written as:
1 l/kc = N/6(a + 0. 2)E (25)
The total compliance of the elastic connection between the two

1 g e a r m a s s e s is simply the total of a l l the individual compliances:
I l / k T = 3(1/E
P
+ 1/Eg ) + (1/GP )(O. 125 + 0. l/a )
P
+ (1/Gg ) (0. 125 + 0. l/a )
I r
g
1
I Tuplin continues by finding moments of i n e r t i a for the two mating
-gears. Since, i n instrument gearing, the connected m a s s e s of the

1 shafts and other p a r t s a r e generally quite s m a l l i n comparison with the
1 g e a r s themselves, only the g e a r i n e r t i a need be considered.
compliances discussed heretofore a r e f o r unit face widths, the i n e r t i a s

Since all
I f o r the g e a r s m u s t a l s o be p e r unit face width. The equivalent m a s s of
each g e a r is the moment of inertia for a unit face width of the g e a r

I divided by the square of the radius of the pitch circle. This is i n
1 a g r e e m e n t with Buckingham except that Buckingham c a r r i e s h i s work
through on a n actual g e a r thickness r a t h e r than working with unit
1
- 39 -
I
face-width values. Although the effects a r e not exactly l i n e a r because
I
of end effects, etc., no problem is anticipated using the two c r i t e r i a s
since Tuplin's work i s used only to find the effective error-in-action.

I
The effective m a s s f o r Tuplin's formulation can be found by using I
values as calculated f r o m Buckingham's approach and dividing by the
f a c e width. The r e s u l t is a n effective mass p e r unit face-width.

I
If the effective m a s s and the compliance of the dynamic s y s t e m a r e
known, the n a t u r a l period of vibration c a n be calculated. The n a t u r a l

I
period f o r a s y s t e m with two m a s s e s is: I
T1
= 2.rr J 1/M
l'k~
+
l/Mg I
P
The time of i n s e r t i o n of a pitch e r r o r is simply the time f o r the I
g e a r tooth to advance one position. This can be found by dividing the
c i r c u l a r pitch by the pitch line velocity: I

where:
tl = 5pn/v
I
pn = c i r c u l a r pitch, in.
V = pitch line velocity, f t / m i n

I
tl = insertion time, s e c
If the time of insertion of a pitch e r r o r is l a r g e i n comparison with the

I
n a t u r a l period T1,the effective e r r o r - i n - a c t i o n i s low and the dynamic I
loads a r e low. If, on the other hand, the time of i n s e r t i o n is v e r y
s m a l l i n comparison with the n a t u r a l period, the effective e r r o r - i n - I

action is large, and a high dynamic load results.
Tuplin's r e s e a r c h indicates that the adjacent pitch e r r o r s a r e much

I
m o r e important than the composite pitch e r r o r s i n the determination of I
I
- 40 - I
dynamic loads,25 w h e r e a s the American Standards Association came
to the opposite conclusion. Tuplin's line of thinking would s e e m to be
c o r r e c t in that the pitch e r r o r of adjacent teeth would cause m o s t of
the accelerations of the gear masses. Since the tooth-to-tooth e r r o r
is usually m e a s u r e d by mating the g e a r with a m a s t e r gear and ob-
serving the change i n center distance as the g e a r s rotate, the pitch
e r r o r can be found by multiplying this tooth-to-tooth e r r o r by the
tangent of the p r e s s u r e angle. This fact is illustrated i n Fig. 7.
Since t h e r e a r e two g e a r s i n contact, this e r r o r i s doubled to find the
actual pitch e r r o r s for two mating gears:
E l = 2(tooth-to-tooth e r r o r ) tan 6 (29)
The significant factor i n this analysis is the summing of maximum
pitch e r r o r s of any two adjacent teeth i n one gear with the maximum
pitch e r r o r s of any two adjacent teeth i n the mating gear. The u s e of
this sum i s pessimistic t o a certain d e g r e e since the maximum e r r o r s
do not usually come together i n every revolution of the pinion. If the
number of teeth i n the g e a r i s exactly divisible by the number of teeth
i n the pinion, the maximum e r r o r s may never come together. Since
t h e r e is some s m a l l probability that maximum e r r o r s will come
together each revolution, it seems wise to u s e this value f o r calcula-
tions and realize that actual dynamic loads will never exceed the
calculated values, thus incorporating conservatism i n the analysis.
With the actual pitch e r r o r now defined in Eq. (29), an effective
pitch e r r o r can be ascertained from the ratio of insertion time t to

1
-
251bid. , p. 210.
- 41 -
I
the natural period of vibration T. Tuplin h a s a s s u m e d that the pitch I
e r r o r can be equivalent to one of t h r e e types of wedges acting on the
g e a r teeth.26 The t h r e e types a r e :

I
1. Uniform t a p e r I
2. Concave c i r c u l a r a r c
3. Convex c i r c u l a r a r c
I
RigGrous calculations f o r each of the above types of wedges have been
made for estimating effective pitch e r r o r s . One thing that became

I
apparent, however, was that the type of wedge a s s u m e d seemed to I
have very little effect on the effective errors-in-action. The ratio of
effective e r r o r s - i n - a c t i o n to the actual e r r o r s - i n - a c t i o n can be related I

to the ratio of insertion t i m e to n a t u r a l period of the dynamic system.
Such a relationship i s indicated i n Fig. 8.27 Thus a n effective e r r o r -

I
in-action can be calculated by knowing the actual pitch e r r o r f r o m 1
Eq. ( 2 9 ) a n d the r a t i o t l / T l e This effective e r r o r - i n - a c t i o n i s then
used with Buckingham' s formulae to establish dynamic loads, which, I

i n turn, a r e used with H e r t z ' equation to find s u r f a c e c o m p r e s s i v e
stresses. In addition, a n important conclusion can be drawn f r o m

I
F i g . 8. If the ratio of t / T 1 is g r e a t e r than 3. 0, the effective e r r o r -
1 I
in-action becomes essentially zero, and the dynamic load can be
neglected. Thus, a c r i t e r i o n i s established to determine when incor- 1

poration of dynamic loads i n the g e a r analysis i s necessary.
I
1
I
I
- 42 -
I
~~ ~
J p L Technical Memorandum No. 33- 139
In the general discussion on wear i n Chapter 11, it was mentioned
that s e v e r a l authorities were of the opinion that w e a r i s a surface
fatigue phenomenon. The surface fatigue life for a particular gear
s e t can be found by f i r s t calculating the c o m p r e s s i v e s t r e s s f r o m
Hertz' equations. The average number of load cycles to failure for
this s t r e s s c a n be determined from a fatigue life chart. Such a c h a r t
is illustrated in Fig. 9 for 303 stainless steel and 2024-T4 aluminum
alloyF8 The expected hours of operation can be derived a s follows:
Let
E = expected life, h r
c = number of cycles to failure
n = angular velocity of gear being investigated, rpm
F o r every revolution of the investigated g e a r , any one surface under-
goes one s t r e s s cycle. The expected life i n minutes f o r the g e a r would
simply be c/n. The expected life in hours would then become:
E = c/60n (30)
28A. H. Maschmeyer, "Wear Life of Aluminum G e a r s , Product

Engineering, 27 (September, 1956) p. 163.
- 43 -
JPL Technical Memorandum NO. 33-139
I
I
0.5
'I / r,
Fig. 8. Comparison of ratio of insertion t i m e to natural period
( t l / T I and ratio of effective e r r o r i n action to actual
e r r o r i n action
- 44 -
I
303 STAINLESS
W
n
3
-
J
a
LL
0
I-
v)
w
J
0
>
0
v)
v)
W
n
I-
v)
LL
0
t
LL
w
m
z
I
3 2024- 1 4
z ALUMINUM
BO 100
COMPRESSIVE STRESS, lb/in? x IO3
Fig. 9. Surface fatigue curves for 3 0 3 stainless steel and

2024- T4 aluminum
I
I - 45 -
V. GEAR WEAR TESTS AND RESULTS
The analysis of w e a r on s p u r g e a r s r e q u i r e s actual w e a r testing.
S e v e r a l w e a r t e s t machines have been developed and used i n the p a s t
i n which two cylinders r u n against each other. With gearing between
the two cylinders, the proportion of sliding to rolling action i s variable,
and the contact p r e s s u r e can be varied t o give the surface s t r e s s r e -
quired. Since g e a r s behave by sliding and rolling i n varying propor-
tions, it was decided that the b e s t w e a r t e s t for g e a r s would be to
m e a s u r e wear of actual g e a r s i n operation r a t h e r than cylinders, o r
some other form of friction surfaces. Another factor which would be
difficult to produce on rubbing s u r f a c e s other than g e a r s is the surface
finish left by a g e a r hob o r shaper. Although i t h a s been f a i r l y well
established that s u r f a c e finish h a s little o r no effect on wear, it is
believed that wear data obtained directly f r o m the g e a r s would be
m o r e representative than would that obtained by duplicating all the
variables on other rolling and sliding parts.
A g e a r testing fixture was built by simply using two 1/4-in. plates
t o hold the bearings, and by spacing t h e s e about a n inch apart. These
plates w e r e precision jig-bored to reduce the possibility of misalign-
m e n t of the gear s e t s and to i n s u r e p r o p e r operating center distances.
A d i r e c t c u r r e n t g e a r motor w a s placed at the c e n t e r and g e a r t r a i n s
w e r e originated at this point, extending radially i n s i x directions.
P r e c i s i o n 1 gears w e r e used, and all had a pitch d i a m e t e r of 1 in. and
96 d i a m e t r a l pitch. P r e c i s i o n 1 is a n A m e r i c a n G e a r Manufacturers
Association designation and indicates a total composite e r r o r of
- 46 -
0. 0010 in. o r l e s s and a tooth-to-tooth e r r o r of 0. 0004 in. o r less.
A stainless steel gear was placed on the motor to drive six aluminum
gears. This procedure tended to equalize w e a r among the driving
pinion and the driven gears. F r o m the aluminum g e a r s , the t e s t g e a r
trains w e r e run. Figure 1 0 s h o w s the t e s t fixture with the top plate
removed and one t e s t g e a r train. T h r e e m e s h e s w e r e used f o r each
m a t e r i a l combination i n o r d e r to obtain statistical information about
the w e a r patterns.
The m a t e r i a l combinations used for the t e s t s were:
1. 3 0 3 stainless steel on 303 stainless steel
2. 303 stainless steel on 2024-T4 aluminum
3. 303 stainless steel on 2024-T4 anodized aluminum
4. 303 stainless s t e e l on 2024-T4 anodized aluminum treated
with molybdenum disulphide
5. 3 0 3 stainless steel on delrin
6. 2024-T4 anodized aluminum on 2024-T4 anodized aluminum
Although many other material selections and combinations a r e
available and i n use, the above combinations a r e the m o s t widely used
i n precision, nonlubricated gear trains. A s e t of useful tables o r
f o r m u l a developed for these m a t e r i a l s would be quite useful. The
molybdenum disulphide w a s tested m e r e l y t o a s c e r t a i n the effects of a
dry-film lubricant. Other dry-film lubricants a r e available, but the
molybdenum disulphide s e e m s to be the m o s t popular because of i t s
versatility. Graphite, for example, is a good dry-film lubricant under
c e r t a i n conditions. In a high vacuum, however, graphite tends to lose
i t s absorbed m o i s t u r e and become a s e v e r e abrasive. Molybdenum
- 47 -
J P L T e c h n i c a l M e m o r a n d u m No. 33-139
Q
@
- 48 -
disulphide, on the other hand, retains its lubricity quite well i n a
vacuum and is a good dry-film lubricant f o r space applications.
F i r s t , it was decided to use a friction clutch a t the end of each
t r a i n to provide the load on the gear teeth. However, with the possi-
bility of placing the package on board a space probe f o r gear w e a r
checks at extremely low p r e s s u r e s , a method was looked for which
would consume a low amount of power. It was finally decided to load
the g e a r s i n a manner s o that the motor need only be l a r g e enough to
supply the l o s s e s in the gear train.
W e s t e r n Gear Company tests two s e t s of g e a r s at 0 n c e . 2 ~ The
g e a r s a r e clamped on two parallel shafts with a small motor connected
to one of the shafts. One gear is then rotated on its shaft until the
torsional deflection in the opposite shaft gives r i s e to a tooth load
equivalent to the actual working conditions. The s m a l l motor then
brings the g e a r s up to the desired speed. The g e a r s a r e transmitting
the design torques, but the motor i s simply supplying the l o s s e s i n
o r d e r to maintain the speed.
The loading f o r the gear testing i n this p a p e r was accomplished
i n a s i m i l a r manner. Antibacklash g e a r s w e r e used to establish the
tooth loads. The tooth loads were established by using a torque wrench
s e t at the d e s i r e d load. One face of the antibacklash gear was held
while the other face was turned through a n angle with the torque
wrench. When the load was reached, the mating pinion was brought
into contact s o that the antibacklash faces could not rotate, thus giving
the d e s i r e d preload on the gear teeth.
29Joseph Beggs, Mechanism, (New York, McGraw-Hill, 1955)pp. 94-95.
- 49 -
The motor used was a d i r e c t - c u r r e n t type. The t e s t s w e r e run at
various speeds i n o r d e r to obtain various data points that incorporated
both dynamic and t r a n s m i t t e d loads. The speed of the g e a r t r a i n s was
varied by adjusting the motor voltage. Rotational speeds of 76, 380,
and 3800 rpm w e r e selected that resulted i n pitch line velocities of
20, 100, and 1000, ft/min, respectively.
C l a s s 7 bearings w e r e used on the g e a r shafts to reduce the
amount of play that might feed back to the w e a r measurements. All
g e a r shafts were shimmed to reduce the r a d i a l play of the bearings to
a minimum. The g e a r s w e r e checked f o r surface finish and found to
have surfaces on the o r d e r of 1 6 p i n The g e a r s w e r e a s s e m b l e d on
the g e a r shafts and the entire a s s e m b l i e s w e r e ultrasonically cleaned
to remove t r a c e s of lubricants o r foreign matter. This cleaning w a s
deemed n e c e s s a r y since the gear manufacturing p r o c e s s usually in-
volves cutting o r machining oils and a n oil coating to reduce the possi-
bility of r u s t o r corrosion. A s m a l l amount of this oil would m o s t
likely have large effects on the observed life of a g e a r set. Although
g e a r s generate p a r t i c l e s a s they operate, it was a l s o felt n e c e s s a r y to
p e r f o r m these t e s t s i n a dust f r e e environment. This procedure would
tend to reduce the h a z a r d s of foreign p a r t i c l e s starting the chain
reaction of the wearing process.
The wear was a s c e r t a i n e d by m e a s u r i n g the rotation of one g e a r
while the mating g e a r was held. A dial indicator and l e v e r a r m s w e r e
used to record this angular rotation i n inches. Although the g e a r
m e s h e s were antibacklashed, the amount a g e a r space exceeded the
g e a r tooth thickness could be m e a s u r e d the s a m e manner as r e g u l a r
- 50 -
~
backlash. Instead of measuring backlash as the f r e e play f r o m one
direction to the other, the space i s m e a s u r e d by pulling against the
antibacklash spring f r o m one side to the other side. More c a r e m u s t
be used, however, since the limits a r e not as easily discernible. In
o r d e r to eliminate eccentricities, total composite e r r o r s , and tooth-
to-tooth e r r o r s , s c r i b e lines were put on the gear faces s o that
"backlash" readings would always be taken a t the s a m e relative posi-
tion. The i n c r e a s e in "backlash" is a d i r e c t m e a s u r e of the depth of
wear on the tooth surface. The amount of w e a r p e r revolution o r
cycle can be determined f r o m backlash a s follows:
h = AB/60nt (31)
where:
h = depth of wear r a t e , in. /cycle
AB = change i n backlash (finial - initial backlash), in.
n = speed of gear, rpm
t = time over which A B is taken, h r
The "backlash" w a s closely monitored during the early periods of
operation when the wear r a t e s were m o s t e r r a t i c . As the t e s t s pro-
g r e s s e d , wear data was taken a t increased intervals. The t e s t s w e r e
terminated when the "backlash" had increased by a value of 0. 004 in.
Although somewhat a r b i t r a r y , this value w a s selected since it r e p r e -
sented roughly 25% of the tooth thickness on the pitch c i r c l e of a
96 pitch gear. The tooth thickness f o r a 96 pitch gear is approximately
0. 016 in. at the pitch line. W i t h a removal of 0. 004 in. f r o m the tooth
thickness, the gear could no longer be classified a s a precision gear.
Since different s y s t e m s can tolerate a different amount of profile
- 51 -
removal before accuracy is impaired, the w e a r was determined i n a
r a t e of depth of wear p e r revolution s o that the data could readily be
used by many people designing systems.
In the discussion of g e a r testing, it should be pointed out that this
testing was performed on g e a r s of equal n u m b e r s of teeth on both the
pinion and gear. Most authorities a g r e e that g e a r t r a i n s should be
designed with a "hunting tooth" a r r a n g e m e n t to reduce wear. The
"hunting tooth" a r r a n g e m e n t is simply the selection of the g e a r s of the
proper number of teeth s o that s e v e r a l revolutions a r e made before
any two mating teeth come into contact again. The testing was pur-
posely not conducted i n this manner f o r s e v e r a l reasons. First, it i s
obviously not possible to design all g e a r t r a i n s with the "hunting tooth"
scheme. F u r t h e r , it was important, if possible, to achieve a reason-
ably accurate s e t of w e a r r a t e s that could be used f o r all c a s e s of

1
nonlubricated g e a r de signs, whether f o r the "hunting tooth'' o r the
ordinary multiple arrangement. In short, a conservative but plausible I

method f o r predicting gear life was sought.
A s mentioned previously, each m a t e r i a l combination w a s tested at

I
t h r e e values of load and speed. One s e r i e s of t e s t s run at 20 f t / m i n
D
and a load of 3 in. -02 (6 in. -oz, for 303 stainless steel on 3 0 3 s t a i n l e s s
steel) was intended to be p r i m a r i l y a transmitted-load t e s t , with little I

o r no dynamic load. At the other e x t r e m e , a second s e r i e s of t e s t s ,
run a t 1000 f t / m i n and with no load, was intended to a pure dynamic-

I "
I
loading test. The l a s t s e r i e s of r u n t e s t s a t 100 f t / m i n and 4 in.
was intended to be a combination of both dynamic loading and

-02
I
I
I
1
I t r a n s m i t t e d load. The t h r e e tests w e r e performed to establish w e a r
I r a t e s at various s t r e s s levels and a l s o to u s e the wear r a t e s to check
Although loads cannot be found directly

on dynamic load calculations.
I by knowing the wear r a t e s , one can determine r a t h e r easily whether
the calculated s t r e s s e s a r e of the right o r d e r of magnitude. The

I higher the s t r e s s e s f o r a given g e a r set, for example, the higher the
I w e a r rate w i l l be. With this information, one can a s c e r t a i n , to a
limited degree, which methods a r e m o s t nearly c o r r e c t i n calculating
I dynamic loads.
S t r e s s e s have been calculated for the above loads and speeds

I using Buckingham's formulae a s outlined in Chapter IV. Two calcula-
I tions of s t r e s s w e r e made for each value of load and speed, using the
two methods of determining effective errors-in-action. One s t r e s s
I calculation was based upon Eq. (14), while the other calculation was
based on Tuplin's work. The calculations a r e performed in

I Appendices B and C.
I The data collected f o r the gear t e s t is given in Appendix D.
F i g u r e s 11 through 28 illustrate this data pictorially. F r o m the data,
I w e a r r a t e s can be established. With the wear rate for the total gear
set known, the amount of wear for the pinion o r the g e a r can be d e t e r -
I mined f r o m Eq. (4). F o r t e s t gears of the s a m e m a t e r i a l , the w e a r
I r a t e on each gear i s one half the total wear of the g e a r set.
of different m a t e r i a l s , however, the softer gear is worn a g r e a t e r

F o r gears
I percentage of the total than the harder gear. In s e v e r a l of the g e a r
t e s t s conducted, 303 stainless steel was run on itself, o r on aluminum

I o r delrin. Since the ratio of the yield strength of either aluminum o r
I
I - 53 -
0 50 too I50 200
TIME, hr
Fig. 11. W e a r r a t e f o r 3 0 3 stainless s t e e l on 3 0 3 stainless steel

(no load, 3800 rpm)
II
IO
3
0 so 100 I so 20
TIME, hr
Fig. 12. Wear r a t e for 303 stainless steel on 2024-T4 aluminum

(no load, 3800 r p m )
- 54 -
Fig. 13. Wear r a t e f o r 303 s t a i n l e s s s t e e l on anodized 2024-T4

aluminum (no load, 3800 r p m )
I1
IO
8 .
Fig. 14. Wear r a t e f o r 3 0 3 s t a i n l e s s s t e e l on anodized 2024-T4

aluminum treated with molybdenum disulphide
- 55 -
I
I
I
I
I
I
I
I
0 50 100 150 200
TIME, hr
Fig. 15. Wear r a t e f o r anodized 2024-T4 aluminum on anodized

2024-T4 aluminum (no load, 3800 r p m )
I
I
II
IO
I
9
TIME, hr
F i g , 16. Wear r a t e for 303 stainless s t e e l on d e l r i n

II
10
ln
0
.c 9
0
c
.-
I
-
.-
-
.- 8
E
r 7
v)
a
-I
Y 6
0
a
m
z 5
Q
W
s
4
3
0 100 200 300 400 500 600 700 8 00
TIME, hr
Fig. 17. Wear r a t e f o r 303 stainless s t e e l on 303 stainless s t e e l

(4 in, - 0 2 , 380 r p m )
m
al
S
V
c
.-
I
._
-
-
.-
E
I
v)
a
J
Y
V
a
m
z
a
w
I
0 50 100 150 200
TIME, hr
Fig. 18. Wear r a t e f o r 303 stainless s t e e l on 2024-T4 aluminum

(4 in. -oz, 380 r p m )
- 57 -
I
I
I
I
I
I
I
T I M E , hr
Fig. 19. Wear r a t e for 3 0 3 s t a i n l e s s s t e e l on anodized 2024-T4
I
aluminum (4 in. -oz, 3 8 0 r p m )
I
1
I
10
al
v)
c
0
.-
C
I
-I
.-
.-
E
x
v)
I
a
J
Y
V
a
m
I
z
I
a
W
I
100 200 3G0 4 00
TIM E,hr
5 00 600 700
1
Fig. 20. Wear r a t e f o r 3 0 3 s t a i n l e s s s t e e l on anodized 2024-T4
(4 in. -oz, 380 r p m )
I
- 58 -
I
n
u
r
0
.-
C
I
-
-
.-
.-
E
I
v)
a
-I
Y
V
a
m
z
a
W
5
TIME, hr
Fig. 21. Wear r a t e for anodized 2024-T4 aluminum on anodized

2024-T4 aluminum (4in. -02, 380 r p m )
I
-
cn
a
-I
Y
V
a
m
z
a
W
I
TIME, hr
Fig. 22. Wear r a t e for 303 s t a i n l e s s s t e e l on d e l r i n

(4 in. - o z , 380 r p m )
I
I - 59 -
JPL Technical Memorandum No. 3 3 - 139 I
I
Fig. 23. Wear r a t e for 3 0 3 stainless steel on 3 0 3 stainless

steel (6 in. - o z , 76 rpm)
I
I
TIME, hr
F i g . 24. Wear rate f o r 303 stainless s t e e l on 2 0 2 4 - T 4 aluminum

( 3 in. - o z , 76 rpm)
- 60 -
I
v)
a
T I M E , hr
Fig. 25. Wear rate f o r 303 stainless steel on anodized 2024-T4

aluminum ( 3 in. -02, 76 r p m )
Fig. 26. Wear r a t e f o r 303 stainless steel on anodized 2024-T4

aluminum treated with molybdenum disulphide
( 3 in. -oz, 76 r p m )
- 61 -
JPL Technical Memorandum No, 3 3 - 139
In
0)
c
0
.-c
I
.-
-
.--
E
"
I
v)
a
-I
Y
V
a
m
z
a
W
H
0 IO0 200 300 400 500 600 7 00 80 0
TIME, h r
F i g . 27. Wear r a t e f o r anodized 2 0 2 4 - T 4 aluminum on anodized
2 0 2 4 - T 4 aluminum ( 3 in. -oz, 76 r p m )
5 -
43
30 IO0 200 300 400
TI ME, h r
!
Fig. 28. Wear r a t e f o r 303 stainless s t e e l on d e l r i n

( 3 in.-oz, 7 6 r p m )
- 62 -
delrin to 303 stainless steel is small, the square of this value i s even
s m a l l e r , and m o s t of the wear occurs i n the softer aluminum o r delrin.
F o r the purposes of this paper, the total amount of wear will b e
a s s u m e d to be totally on either the delrin o r aluminum when these
m a t e r i a l s a r e mated with stainless steel. F o r stainless steel o r
aluminum mated against itself, one half the w e a r is assumed to occur
on e a c h surface. Observations of the t e s t g e a r s under a microscope
tend to support these assumptions.
Wear r a t e s a r e calculated in Appendix E and the r e s u l t s a r e shown
i n Table 1, with calculated stresses. Investigation of this Table gives
some indication of which s t r e s s calculations a r e m o r e valid. The
w e a r r a t e s c o r r e l a t e quite well with the s t r e s s e s calculated f r o m
Tuplin's method. On the other hand, the s t r e s s e s as calculated using
the A m e r i c a n Standards Association Specification B6. 11-1951 s e e m to
have a negative correlation with the wear rate. With 303 stainless
s t e e l mating against 303 stainless steel, for example, the higher s t r e s s
values s e e m to r e s u l t in lower wear rates. Since this i s contrary to
the general theories on wear, all s t r e s s e s used f o r purposes of estab-
lishing wear r a t e s will be calculated using Tuplin's analysis f o r
effective error-in-action, Buckingham' s formulae for dynamic loads,
and Hertz' equations for compressive s t r e s s e s .
A second table can be derived f r o m Table 1 giving the w e a r r a t e
on a specific surface f o r a certain value of s t r e s s . These r e s u l t s a r e
indicated i n Table 2. F r o m the information i n Table 2, design curves
can be drawn which relate calculated s t r e s s e s to wear depth r a t e s for
the five m a t e r i a l s and surfaces. These curves a r e illustrated i n
Fig. 29 through 33. Although Eq. ( 6 ) indicates that the relationship
- 63 -
J P L Technical Memorandum No. 33-139 I
Table 1. Wear r a t e s and calculated s t r e s s e s for m a t e r i a l
I
combinations and loads
4 *#: I
St r e s s, Stress, Wear r a t e s
Material and Load
303 Stainless Steel

l b / i n .2 l b /in? pico-inches/ cycle
I
on
303 Stainless Steel: I
a. 1000 ft/min, no load 66,000 19,500 11. 6
b. 100 ft/min, 4 in. -oz 32,300 22,400 111 1

c. 20 ft/min, 6 in, -oz
303 Stainless Steel

29, 500 27,400 605
I
on
2024- T4 Aluminum: I
a. 1000 ft/min, no load 38,200 9,750 202
b. 100 ft/min, 4 in. -oz 21,300 16, 900 1170 Il

c. 20 ft/min, 3 in. -oz
303 Stainless Steel

16, 100 15, 100 57 5
I
on
Anodized 2024-T4 Aluminum: I
a. 1000 ft/min, no load 38,200 9,750 23 7
b. 100 ft/min, 4 in. -02 21,300 16,900 1070 I

c. 20 ft/min, 3 in. -oz
303 Stainless Steel

16, 100 15, 100 58 5
I
on
Anodized 2024-T4 Aluminum
Treated With MoS2:
D
a.
b.
1000 ft/min, no load
100 ft/min, 4 in. -oz

38,200
21,300
9,750
16,900
14. 6
110
1
c. 20 ft/min, 3 in. -oz 16, 100 15, 100 59. 4 1
:;:
.,. J,
T. .<,
F r o m ASA Spec. B6. 11-1951
1
F r o m Tuplin's method
I
- 64 -
I
-
Table 1 (Cont'd)
Stress,
.
I
I
.-
Stress,
** Wear r a t e s
Material and Load lb/in2 1b/in2 pico-inches/cycle
303 Stainless Steel

on
Delrin:
a. 1000 ft/min, no load 3,720 1,960 29. 4

b. 100 f t / m i n , 4 in. -oz 3,940 3,720 23 2
c, 20 ft/min, 3 in. -oz 3,270 3,220 118
Anodized 2024-T4 Aluminum

on
Anodized 2024-T4 Aluminum:
a. 1000 ft/min, no load 27,200 5,740 82
b. 100 f t l m i n , 4 in. -oz 17,400 14,700 560
c. 20 ft/min, 3 in. -oz 13,500 12,800 3 63

__
J
.
-8-
F r o m ASA Spec. B6. 11-1951

**F r o m Tuplin' s method
- 65 -
Table 2. W e a r rates and s t r e s s e s for test materials
Stress , W e a r rate,
Mat e rial l b / in2 pico- inc he s / cycle
303 Stainless Steel 19,500 11. 6
22,400 111
27,400 605
2024-T4 Aluminum 9,750 202
15, 100 575
16,900 1170
Anodized 2024- T 4 Aluminum 5, 740 82. 0
9,750 237
12,800 3 63
14, 700 560
15, 100 585
16,900 1070
Delrin 1,960 29. 4
3,220 118
3,720 232
Anodized 2024- T 4 Aluminum 9,750 14. 6

W i t h MoS2
15, 100 59. 4
16,900 110
- 66 -
1000
I
’
500
<
-
al
0
a
rn
0)
r
0
.-
0
K
I
.-n
V
W”
F
a
a
a
U
W
3
100
50 E
f c
7
I
I Fig. 29.
I 20,000 22,000
1 24,000
CALCULATED STRESS, I b/in?

26,000 28,000
Design curve for 303 stainless s t e e l indicating calculated

s t r e s s and corremondine deDth-of-wear r a t e
- 67 -
/
2
1000 o o ~
f
-W
U
f
f
5r
<ul
W
r
U
.-c
I
0
U
.-a
W
l-
a
[L
[L
a
W
3
t
--
IO, u 00 12 00 14,000
CALCULATED STRESS, lb/in2

16,000 18,OO
I
Fig. 30. Design curve f o r 2024-T4 aluminum indicating calculated
s t r e s s and corresponding depth-of-wear r a t e
2000
1000
f
-
0)
0
>r
5-
m
.c 500
0
.-c
I
0
.-n
0
J
Ga
LL
a
w
3
IO0
I
6,000 8,000 IO.000 IP.000 14,000 16 30 ie,ooo
CALCULATED STRESS, Ib/in?
Fig. 31. Design curve f o r anodized 2024-T4 aluminum indicating

calculated s t r e s s and corresponding depth-of-wear r a t e
- 69 -
8,000 l0,OOO 12,000 14,000 16,000 18,000
CALCULATED STRESS, Ib/in.2
Fig. 32. Design curve f o r anodized 2024-T4 aluminum t r e a t e d with

molybdenum disulphide indicating calculated s t r e s s and
c o r responding depth- of-wear r a t e
- 70 -
1000
500
Y)
u
c
c
U
.-
I
0
.-U
n
W 00
c-
a
[I
a
a
w
3
so /
Fig. 33.
L
I O 1,000
I
1 %OOO
CALCULATED STRESS, l b / i w 2
4.001
Design curve f o r delrin indicating calculated s t r e s s and

c o r responding depth- of -wear r a t e
- 71 -
between s t r e s s and w e a r should be l i n e a r , the data points fall m o r e
nearly i n a straight line if the s t r e s s and w e a r depth r a t e s a r e plotted
on logrithmic scales. In any case, the r a t e of w e a r f o r each g e a r of
a s e t can be determined by knowing the s t r e s s levels and using the
appropriate wear curves. F o r a typical g e a r s e t i n which the pinion i s
usually much s m a l l e r than the g e a r , the depth-of-wear p e r revolution
can be found for both the pinion and the gear. Then, realizing that the
pinion t r a v e l s through m o r e revolutions than the g e a r i n proportion to
the number of teeth on each gear, the total w e a r r a t e f o r the g e a r s e t
can be determined.
As a n example, a s s u m e that it is d e s i r a b l e to determine the w e a r
r a t e of a pinion and g e a r of not n e c e s s a r i l y the s a m e m a t e r i a l , and
that the g e a r is five t i m e s the d i a m e t e r of the pinion. The s t r e s s can
be calculated as previously outlined, and, with the design curves,
wear-depth rates can be established for this s t r e s s for both the pinion
and the gear. The wear-depth r a t e s will b e i n t e r m s of depth-of-wear
p e r revolution of the g e a r i n question. F o r every revolution of the
g e a r , the pinion will have traveled through five revolutions. Thus,
the total wear p e r revolution of the g e a r , o r five revolutions of the
pinion, would simply be the sum of the wear-depth r a t e s p e r revolu-
tion of the gear and five t i m e s the w e a r p e r revolution of the pinion.
This can be generalized into a n equation f o r establishing total w e a r
rates:
(32)
or
(33)
- 72 -
where:
W = total wear of gear set p e r revolution of the pinion,

TP
in /cycle
W = total wear of gear set p e r revolution of the g e a r ,

Tg
in /cycle
W = wear of pinion p e r revolution, in. / c y c l e

P
W = wear of g e a r p e r revolution, in. / c y c l e
g
N = number of teeth i n pinion
P
N = number of teeth i n g e a r
g
If the pinion is of one m a t e r i a l and the gear of another, the appropriate
c h a r t is consulted and W i s found f o r the pinion. In a s i m i l a r manner,

P
W i s found for the gear. These t e r m s a r e then combined by either
g
Eq. (32) o r (33) to a r r i v e at the total wear r a t e f o r the g e a r set.
It is interesting to note the effects of the molybdenum disulphide
on w e a r rates. The wear r a t e decreased by a n o r d e r of magnitude
f r o m plain anodized aluminum to anodized aluminum t r e a t e d with
molybdenum disulphide. It cannot be a s c e r t a i n e d whether the same
effect could be observed on other t e s t s u r f a c e s because of the limited
testing that was performed. It can definitely be concluded that the
molybdenum disulphide dry-film lubricant does help, but its u s e i n
p r e c i s i o n instrument gearing i s not recommended where high accuracy
i s required. The molybdenum disulphide coating ranges f r o m
0. 0002 to 0. 0005 in. thick and this amount would radically change the
c h a r a c t e r i s t i c s of the involute gear surface.
In the discussion of wear, i t was pointed out that s e v e r a l authors
believed that wear is a surface-fatigue phenomenon. Fatigue l i v e s have
- 73 -
I
been calculated in Appendix F for the t e s t g e a r s using the fatigue I
c u r v e s i n F i g . 9.
the t e s t gears.
Table 3 s u m m a r i z e s s t r e s s e s and fatigue l i v e s f o r
6
Since 10 h r is roughly equivalent to 100 y r , it can be
I
r a t h e r emphatically said that the wear observed on the t e s t g e a r s was I
of some other nature than fatigue. This statement does not imply that
fatigue is not an important problem. It does imply, however, that I

wear and surface fatigue a r e vastly different when applied to nonlubri-
cated surfaces. Wear, i n addition to fatigue, m u s t be considered

I
whenever nonlubricated g e a r s y s t e m s a r e proposed. Fig. 29 through I
3 3 provide a basis for predicting lives of nonlubricated g e a r t r a i n s
caused by wear alone. I

1
I
I
D
M
I
1
I
I
I
I - 74 - I
I Table 3. S t r e s s e s and calculated fatigue l i v e s for the t e s t g e a r s
I Material and Load

* Stress,
lb /in2
Fatigue life,
hr
I 303 Stainless Steel on

303 Stainless Steel:
a. 1000 ft/min, no load 19,500

I b. 100 ft/min, 4 in. -oz 22,400
03
03
I c. 20 ft/min, 6 in. -oz 27,400 03
303 Stainless Steel on
I 2024-T4 Aluminum:
a. 1000 ft/min, no load 9,750

03
I b. 100 ft/min, 4 in. - o z 16, 900

2 . 6 ~
106
6
c. 20 ft/min, 3 in. -oz 15, 100 22 x 10
I 2024-T4 Aluminum on
2024-T4 Aluminum:
I a. 1000 f t / m i n , no load 5, 740 03
4.3 x 10
6
b. 100 ft/min, 4 in. -oz 14, 700
I c. 20 ft/min, 3 in. -oz 12, 800 66 x 106
I *Fatigue data not available f o r d e l r i n
I
I
I
I
I
I
I - 75 -
I
VI. SUMMARY I
The application of the wear data t o qualitatively analyze dynamic I
loads and s t r e s s e s i s one of the m o s t significant discoveries to r e s u l t
f r o m this study. Since the object of this paper was to find wear r a t e s
I
f o r calculated s t r e s s e s , this might sound a s though the "cart is being
placed before the horse. 'I But the fact r e m a i n s that the observed wear
I
gave definite clues a s to which s t r e s s calculation seemed correct. I
Using this a s a guide, it was then decided to u s e Tuplin's method to
establish an effective error-in-action r a t h e r than using the American

I
Standards Association Specification B6. 11- 1951. This effective e r r o r -
in-action was then used with Buckingham's equations to establish

I
dynamic loads, and H e r t z ' s equation was used to find the corresponding I
stress, It should be noted that the u s e of B6. 11-1951 resulted in
dynamic loads much higher than actual loads, and, a s such, gave I
r e s u l t s which tended to be safe.
Wear r a t e s w e r e established for five m a t e r i a l s o r surface con-

I
ditions with the p a r t s running i n a n o r m a l atmosphere and I
nonlubricated, They were:
1. 303 stainless steel I

2.
3.
2024-T4 aluminum
Anodized 2024-T4 aluminum

I
4. Anodized 2024-T4 aluminum treated with molybdenum I
disulphide
5. Delrin I
The wear rate for gear s e t s can be established f o r any combination of
the above materials. Although all the data w e r e taken on t e s t g e a r s of

I
I
- 76 -
I
I ~~~ ___
I 9 6 diametral pitch, the wear-rate curves a r e applicable to all fine-
I pitch gears. A change in pitch effectively changes the sliding distance
by changing the length of the tooth. This is compensated for, however,
1 by the volume of wear, and, thus, the depth of wear would r e m a i n the
same.
I To s u m m a r i z e the method used to establish wear r a t e s for nonlu-
i bricated spur gears:
Find transmitted load from: Wt = T / R , lb

1.
I 2. Find the pitch line velocity from: V = ( r p m )IT R/ 6, f t / m i n
3. Find the time of insertion from: 5pn/V, s e c

1 4. Find the effective m a s s of e a c h g e a r from:
2 2
I 5.
m = Weight/ (386 x F), lb- s e c /in.
Find the r i m depth from:
1 H = outside radius - tooth thickness - bore, in.
6. Find a from: a = H/pn

I 7. Find the compliance of the gear s e t from:
+ 1 / Eg ) + (1/GP) (0. 125 + 0.

I l / k T = 3(1/E
P
l/a)
+ (1/Gg) (0. 125 + 0. l/a)

I + ( 1 / 6 ) [ N p / ( a + 0. 2 ) EP + N g / ( a + 0. 2 ) E
I 8. Find the natural period of the g e a r s e t from:
I T1 = ‘ Y 1 / M p
J.
+ 1 / Mg , sec
I 9. Find the insertion time to natural period ratio from:
V T 1
I
I
,I - 77 -
10. Find the actual error-in-action f r o m :
E l = 2(tooth-to-tooth e r r o r ) tan b
11. Find the ratio E e / E 1 f r o m Fig. 8 and the r a t i o t l / T

1
12. Find the effective error-in-action Ee f r o m 10 and 11 above:
13. Find fa from:
f = 0. 0012(1/R t 1 / R ) [M M / ( M t M )] V2
a P g P g P g
14. Find f from:
d
fd
(0. 111)E / ( 1 / 2 E t l / E g )
e P 1
15. Find fm from:
m = fafd/(fa fd)
16. Find the dynamic load from:
- Jfm(2fd - fm) t 1b
wd
17. Find the total load from:
W T = (Wt t W d ) / F , lb/in.
18. Find the s t r e s s from:
S (EpEg)/(Ept E g ) (DP t D ) / ( D D ) WT/sin 0

C g P g
19. Find the w e a r depth p e r revolution f o r the pinion and the g e a r
from Fig. 29 through 3 3
20. Find the total w e a r r a t e f o r the g e a r set from:
w = w t ( N ~ / N ~ ) Wo r~
TP P
W = W t (N / N )W in. /revolution
Tg g g P P
- 78 -
u Although only five m a t e r i a l s o r surfaces a r e covered in this paper,
I they a r e the m o s t widely u s e d i n industry today. The techniques have
been established here, and an extension of this work could be p e r -
I formed. T h e r e a r e undoubtedly other m a t e r i a l s and surfaces of
interest, as well as different conditions. Testing i n a vacuum, f o r

I example, would m o s t likely l e a d to a different s e t of w e a r curves. At
I any rate, the work in this paper does provide a basis f o r establishing
w e a r r a t e s f o r precision, instrument, nonlubricated, fine pitch, spur
I g e a r s for a very limited number of m a t e r i a l s and/or surfaces. The
r e a d e r is cautioned that the design curves relate actual wear to cal-

I culated s t r e s s e s . If a safety factor is desired, the w e a r r a t e s should
I be increased proportionately.
I
I
I
I
I
1
I
I
I
I - 79 -
I
REFERENCES
I
1. American Gear Manufacturers As sociation. "Standard
Nomenclature of Gear Tooth Wear and Failure. American
I
G e a r Manufacturers Association, Specification 1l b . 02, 1954.
2. American Standards Association. "Inspection of Fine Pitch

I
Gears, American Standards As sociation, Specification
3.
B6, 11-1951, 1951.
Archard, J. F. "Contact and Rubbing of Flat Surfaces. 'I

I
4.
Journal of Applied Physics, 24:981-988, 1953.
Beggs, Joseph S. Mechanism. New York, McGraw-Hill, 1955.

I
5. Bowden, F. P. and D. Tabor. The F r i c t i o n and Lubrication of
Solids. Oxford, Clarendon P r e s s , 1950.
I
6. Brownsdon, H. "Metallic Wear.
Metals, 18:15-27, 1936.
Journal of the Institute of
I
7. Buckingham, Earle. Analytical Mechanics of Gears.
New York, McGraw-Hill, 1949. I
8. Buckingham, Ea rle. Spur Gears. New York, McGraw-Hill,
1928.
I
9. Burwell, John T. Mechanical Wear. New York, American
10.
Society f o r Metals, 1950.
Burwell, J. and C. Strang. "On the E m p i r i c a l Law of Adhesive

I
Wear. I t Journal of Applied Physics, 23:18-28, 1952.
11. Carmichael, Colin. Kent's Mechanical Engineers' Handbook.

I
New York, John Wiley, 1955.
12. Dudley, Darle W. Gear Handbook.

1962.
New York, McGraw-Hill, 1
13. Dudley, Darle W. P r a c t i c a l Gear Design.
Mc G raw - Hill, 19 54.
New York, I
14. Maschmeyer, A. H. "Wear Life of Aluminum Gears.
Engineering, 27:160- 166, September, 1956.
Product
I
15. Merritt, H. E. Gears.
1955.
New York, P i t m a n Publishing Company,
I
16. P r i c e , R. "Gears Have Special Lube Problems.
Engineering, 16:119-123, May, 1962.
If Plant
I
I
- 80 -
I
I J P L Technical Memorandum No. 33-139
I 17. Rabinowicz, Ernest. "New Coefficients P r e d i c t W e a r of Metal

Parts. I f Product Engineering, 29:71-73, June, 1958.
I 18. Rabinowicz, Ernest. "Stick and Slip.

194: 109- 118, May, 1956.
It Scientific American,
I 19. Rabinowicz, Ernest. "Surface Energy and F r i c t i o n and Wear

Phenomena. ' I Journal of Applied Physics, 32:1440-1444, 1961.
I 2 0. Rabinowicz, Ernest. "Wear.

206:127- 136, January, 1962.
Scientific American,
I 2 1. Read, Robert H. "Metal Wear.

August, 1962.
F r o n t i e r , 25:8- 13,
22. Roark, R. J. F o r m u l a s for S t r e s s and Strain. New York,

I McGraw-Hill, 1943.
23. Ryder, E. A. "A G e a r and Lubricant T e s t e r . I t American

I Society f o r Testing Materials Bulletin, 148:69-73,
October, 1947.
I 2 4. Savage, R. "Graphite Lubrication.

19: 1- 10, 1948.
" Journal of Applied Physics,
2 5. S p u r r , R.. T. "Creep and Static Friction. B r i t i s h Journal of

I
If
Applied Physics, 6:402-403, 1955.
2 6. Spurr, R.. and T. Newcomb. "The F r i c t i o n and Wear of Various

I Materials Sliding
- Against
- Unlubricated Surfaces of Different
Types and Degrees of Roughness. I t Proceedings of the
Conference on Lubrication and W e a r . October 1-3, 1957,
1 London. The Institution of Mechanical Engineers,

pp. 269-275.
- 1957.
2 7. Taylor, R. and W. Holt. "Effect of Roughness of Cast I r o n

I Brake D r u m s in Wear Tests of B r a k e Linings. Journal of
Research, National Bureau of Standards, 27:395, 1941.
I 2 8.
29.
Tuplin, W. G e a r Load Capacity. New York, John Wiley, 1962.
Tuplin, W. "Dynamic Loads on G e a r Teeth. 'I Machine Design,

25:203-211, October, 1953.
I 30. Underwood, 0. A. "Operating Life of Gears. Product
Engineering, 32:65-74, May, 1962.
I 3 1. Way, S. "Pitting Due to Rolling Contact.
Mechanics, 2:49-58, 1935.
J o u r n a l of Applied
I 3 2. Whittaker, E. J. W.
1947.
"Friction and Wear. Nature, 159:541,
I
I - 81 -
I
APPENDIX A I
Sliding Velocity of G e a r s
I
The maximum sliding velocity of a g e a r s e t i s produced at the
point where the g e a r s f i r s t come into contact and a t the point where
I
the g e a r s l a s t contact during a given mesh.
velocity for a given radius f r o m Eq. 8 is:

The maximum sliding
I
vS = V (l / R 1 t l / R 2 ) ( , / r 1 2 - %f- R1 sin+)) I
The maximum sliding velocity achievable with a s e t of precision f i n e - I
pitch g e a r s can b e found a s follows:
1. Maximum p r a c t i c a l pitch line velocity is 1000 f t / m i n I

2.
3.
Coarsest pitch is about 48 diametral pitch
P r e s s u r e angle + of 20" is a l m o s t universally used

I
4. The s m a l l e s t r a d i i (R1 and R2) f o r 48 pitch g e a r s is about
0. 300 in.
m
F r o m the above assumptions and the standard configurations, the I
following information can b e listed:
1. R1 = R2 = 0. 300 in.
D
2.
3.
Rbl = Rb2 = (0. 300)
Rol = R02 = 0. 321 in.

COS 20" = 0. 282 in.
I
4. V = 1000 f t / m i n I
Thus, the maximum sliding velocity to b e expected with a s e t of p r e c i -
sion instrument spur g e a r s is:

I
Vs = 1000 ( l / O . 3 t 1 / 0 . 3 ) ( J ( . 321)2 - (. 282)2 - 0. 3 s i n 20") I
= 6667 (do.103 - 0. 079 - 0. 102)
V,
Vs = 6667 ( 0 . 155 - 0.102) = 353 f t / m i n

I
I
- 82 -
I
~
The maximum sliding velocity for the t e s t g e a r s can be calculated
in a s i m i l a r manner. F o r t e s t gears:
1. R1 = R2 = 0. 500 in.
2. Rbl = Rb2 = ( 0 . 500) C O S 20" = 0.470 in.
3. Rol = R02 = 0. 5104

4. V = 20, 100, and 1000 f t / m i n
F o r 20 f t / m i n
vs
J (0.500)2 - (0.470)2 - 0. 5 s i n 20")
= 20 (l/O. 5 + 1/0. 5)(
= 20 (4)(G-
0.171)
vs
= 80 (0.1711 - 0. 171) = 0. 008 ft/min
F o r 100 f t / m i n
V S = 100 (I/O. 5 4- 1/0.5) ( J ( 0 . 500)2 - (0.470)2 - 0. 5 sin 20")

= 100 (4)(,/. 0293 - 0.171)
Vs = 400 (0.1711 - 0. 171) = 0. 04 ft/min
F o r 1000 f t / m i n
V S = 1000 (l/O.5 t 1/0. 5 ) ( , / ( 0 . 500)2 - (0.470)2 - 0. 5 sin 20")
= 1000 (4)!\=-
0.171)
Vs = 4000 ( 0 . 1711 - 0.171) = 0. 4 ft/min
- 83 -
APPENDIX B I
Calculation of T e s t Gear-Tooth Loads and S t r e s s e s Using
Buckingham's Formulae and American Standards
Association Specification B6. 11-1 951
I
F o r purposes of the calculations, the following numbers will b e
I
used:
1. 303 Stainless Steel:

I
(a). p = 0. 283 p s i I
(b). E = 30 x l o 6 p s i
( c ) . G = 1 2 x 106 p s i
2. 2024-T4 Aluminum:
(a). p = 0. 095 p s i
(b). E = 1 2 x l o 6 p s i
( c ) . G = 4 x 106 p s i
3. Delrin:
(a). p = 0. 052 p s i
(b). E = 0. 41 x l o 6 p s i
(c). G = 0. 16 x l o 6 p s i
4. Error-in-action:
E1 = total composite e r r o r t 1 / 2 (tooth-to-tooth e r r o r )

= 0. 0010 t 1 / 2 (0. 0004) = 0. 0012 in.
The calculations a r e divided into t h r e e sections, one f o r each
value of load and speed.
1. V = 1000 f t / m i n , no load
.'. Wt = 0 lb
- 84 -
I (a). 303 Stainless Steel on 303 Stainless Steel
I m
e
= w w / 2 (w t w )
P g g P g
w = I T R ~ F ~
I w
P
= (3. 14) (0. 5)2 (0. 0625) (. 283)
P
I = 0. 014 lb
w = (3. 14) (0. 5)2 (0. 125) (. 283)
I g
= 0. 028 l b
I m [
e = (0. 014) (0. 028)/ 2 (386) (0. 014 t 0. OZ8)]
I = 1.21 x lb-secz/in.
a = 0.012 (1/R1 t 1 / R 2 ) me V2
f
I = 0.0012 (110.5 + 110. 5) (1.21 10-5) (~ooo)~

= 0. 058
I = Wt t F [(O. 111) E 1 / ( 1 / 2 E t l/Eg)]
fd P
1 [
= 0 t . 062 (0.111) (0. 0012)/(10-6/60 t 1 0 - ~ / 3 0 ) ]
= 165
I m = fa f d / ( f a t fd)
I = (165j (0.0 5 8 ) / ( 0 . 058 t 165)

= 0. 058
I
1 = d (0. 058) (330 - 0. 058)
=
I Total Load
4. 37 lb
1
It
I - 85 -
J P L Technical Memorandum NO. 33-139
I
wT = (4. 37 t 0. O ) / O . 062
I
Stress
= 7 0 lb/in.
I
S
C
= 0.84 E E / ( E p t Eg)
P g
(DP t D g ) / D D
P g
WT/sin 20" 1
= 0.84 (15 x 106) (1) (70/0.34) I
= 66,OOOpsi
(b). 303 Stainless Steel on 2024-T4 Aluminum
I
w
P
= 0. 014 l b f r o m (a)
1
w = (3. 14) ( 0 . 5)2 (0. 125) (0. 095)
g
= 0.0093 l b
1
m
e
= ( 0 . 0093) ( 0 . 014)/ [2(386) ( 0 . 0093 t 0. 014)] I
f
a
= 7. 25 x lb-sec2/in.
= 0. 0012 ( l / O . 5 t 1 / 0 . 5) (7. 25 x (1000)2

I
= 0. 0348 I
f d = 0. 062 [(O. 111) (0. 0012)/(10-6/60 t 10-6/13)]
= 83
I
f
m
= (83) (0. 0348)/(83 t 0. 0348) I
= 0. 0348
Wd = J (0. 0348) (166 - 0. 0348)

I
= 2. 40 l b I
Total Load
wT = (0. 0 t 2. 40)/0. 062

I
= 38. 8 lb/in. I
I
- 86 - I
Stress
sC = 0. 84,/[(30) (12) (106)/42] (1) (38. 8/0. 34)
= 38, 200 psi
( c ) . 2024-T4 Aluminum on 2024-T4 Aluminum
w = 0. 093 lb f r o m (b)
i3
w = (3. 14) (0. 5)2 (0. 062) (0. 095)

P
= 0. 0047 l b
m
e
= ( 0 . 0047) (0. 0093)/ [2 (386) (0. 0047 t 0. O093)]
= 4. 1 x lb-sec2/in.
f
a
= 0.0012 ( l / O . 5 t 1 / 0 . 5) (4.1 x 10-6) (1000)2
= 0.0197
fd = 0. 062 [(O. 111) (0. 0012)/(10-6/24 + 10-6/13)]

= 72 lb
f = (72) (0. 0197)/(72 t 0. 0197)

m
= 0.0197
Wd = J(O.0197) (144 - 0.0197)
Total Load
W T = (0. 0 t 1.69)/0. 062
= 27. 2 lb/in.
Stress
SC = 0. 84/[(12) (12) (106)/24] (1) (27. 2/O. 34)
= 27,200 psi
- 87 -
I
(d). 303 Stainless Steel on Delrin
I
w
P
= 0. 014 l b from (a)
I
w = (3. 14) ( 0 . 5)2 (0. 125) (0. 052)
g
= 0. 0051 l b
I
m e = ( 0 . 0051) ( 0 . 014)/[2(386) ( 0 . 0051 t O.014)]
I
= 4. 83 x 10-6 lb-sec2/in.
f
a
= 0. 0012 ( l / O . 5 t 1/0. 5) (4.83 x (1000)2
D
= 0. 0232 D
*d = 0. 062 [(o. 111) (0. 0012)/(10-6/60 t 4111
= 5. 5
I
f
m
= (5. 5) (0. 0232)/(5. 5 t 0. 0232) i
= 0. 0232
wd = J (0. 0232) (11 - 0 . 0232) R
= 0. 505 l b N
Total Load
WT = (0. 0 t 0. 505)/0. 062 I

= 8. 1 5 lb/in.
1
Stress
SC = 0. 8 4 4 [(30) (0.41) (106)/30] (1) (8. 15/0. 34)

1
= 3,720 p s i I
I
I
1
I
2. V = 100 f t / m i n , 4 in. -02 load
Wt = ( 4 in. -oz)/O. 5 in. (1 lb/16 oz)
= 0. 500 l b
(a). 303 Stainless Steel on 303 Stainless Steel
w = 0. 028 l b f r o m ( l a )
g
w = 0. 028 l b a l s o (face width increased)

P
m
e [
= (0. 028)2/ 2 (386) (0. 028 t 0. O28)]
= 1.81 x lb-sec2/in.
f
a
= 0.0012 ( u o . 5 t 1/0. 5) (1.81 10-5) ( i 0 0 ) 2
= 0. 00087
fd [
= 0.500+0.062 (0.111) ( 0 . 0 0 1 2 ) / ( 1 0 - ~ / 6 0 + 1 0 ~ 6 / 3 0 ) ]
= 166
f = (166) (0. 00087)/(166 t 0. 00087)

m
= 0. 00087
(0. 00087) (332-0. 00087)
Wd =
= 0. 539 l b
Total Load
= (0. 500 t 0. 539)/. 062

wT
= 16. 8 l b / i n .
Stress
SC = 0.84J [(30) (30) (106)/30] (1) (16. 8/0. 34)
= 32,300 psi
- 89 -
w = 0. 028 l b f r o m (a)
P
w = 0. 0093 l b f r o m ( l b )
g
m = ( 0 . 028) ( 0 . 0093)/ [2(386) (0. 028 t 0. 0093)]

e
= 9. 05 x lb-secz/in.
f = 0. 0012 ( l / O . 5 t 1 / 0 . 5) (9. 05 x
a
= 0. 00043
f d = 0. 500tO. 062 ( 0 . 111)(0. OO12)/(10-6/60t10-6/12)]
= 83
f
m
= (83) (0. 00043)/(83 t 0. 00043)
= 0.00043
Wd =J (0. 00043) (166 - 0. 00043)
= 0. 267 lb/in.
Total Load
WT = (0. 500 t 0. 267)/0.062
= 12. 4 lb/in.
Stress
sC = 0. 8 4 1 [(30) (12) (io6)/4z] (1) (12.4/0. 34)
= 21,000 p s i
(c). 2024-T4 Aluminum on 2024-T4 Aluminum
w = 0. 0093 l b f r o m ( l b )
g
w = 0. 0093 l b a l s o (face width i n c r e a s e d )
P
- 90 -
I m e = (0. 0093)2/ [2(386) (0. 0093 t 0.0093)]
I = 6. 03 x lb-sec2/in.
f = 0. 0012 (1/0.5 + 1/0. 5) (6. 03 x

1 a
1
= 72
f
m
= (72) (0. 0 0 0 2 9 ) / ( 72 + 0. 00029)
= 0.00029
Wd = ,/ (0.00029) (144 - 0. 00029)
= 0. 204 l b
Total Load
W T = (0. 500 t 0. 204)/0. 062
= 11. 3 lb/in.
Stress
SC = 0. 8 4 J [(12) (12) (106)/24] (1) (11. 3/0. 34)
= 17,400 psi
w = 0. 028 l b from (2a)

P
w = 0. 0051 lb f r o m (Id)
g
m
e
= (0. 028) (0. 0051)/ [2 (386) (0. 028 + 0.0051)]
= 5.60 x lb-sec2/in.
f
a
= 0.0012 ( l / O . 5 t l / O . 5) (5.60 x
= 0.000269
- 91 -
fd
= 0. 500 t 0. 062 (0. 111)(0. 0012)/(10-6/60 t 10-6/0. 41)
= 6. 0
f
m
= (6. 0) ( 0 . 000269)/(6. 0 + 0. 000269)
= 0.000269
wd
= 4(0. 000269) (12 - 0.000269)
= 0. 057 lb
Total Load
WT = (0. 500 t 0. 057)/0. 042
= 9. 1 5 lb/in.
Stress
sC = 0. 84J [(30) (0. 41) ( 1 0 ~ ) / 3 0 ](1) (9. )01.5/
= 3,940 p s i
3. V = 20 ft/min, 3 in. -oz load
(6 in. -oz for s t e e l on steel)

Wt = ( 3 in. -oz)/O. 5 in. (1 l b / 1 6 oz) = 0. 375 l b
(0. 750 l b for s t e e l on steel)

m
e
= 1. 81 x lb-secz/in. f r o m (2a)
f
a
= 0.0012 (110.5 t 110. 5) (1. 81 10-5) (2012
= 0.000035
fd
= 0. 750 t 0. 062 (0.111) (0. 0012)/(10-6/60 t 10-6/30)]
= 166
f = (166) (0. 000035)/(166 t 0. 000035)

m
= 0.000035
- 92 -
I Wd = J (0.000035) (332 - 0. 000035)
I = 0.108 lb
I Total Load
W T = (0. 750 t 0.108)/0.062
1 = 13. 8 lb/in.
Stress
I Sc = 0. 8 4 J [(30) (30) ( 1 0 ~ ) / 6 0 ](1) (13. 8/0. 34)
I = 29,500 psi
I (b). 303 Stainless Steel on 2024-T4 Aluminum
m = 9. 05 x lb-secz/in. f r o m (2b)
e
I f
a = 0. 0012 ( l / O . 5 t l / O . 5) (9. 05 x
(20)2
1 = 0.0000174
f d = 0. 375 t 0. 062 (0.111) (0. 0012)/(10-6/60 t 10-6/12)]
1 = 83
I f
m
= (83) (0. 0000174)/(83 t 0. 0000174)
= 0. 0000174
I Wd =
(0.0000174) (166 - 0. 0000174)
I = 0. 054 lb
Total Load
I WT = (0. 375 t 0. 054)/0. 062
I = 6. 9 lb/in.
I
I
I - 93 -
I
Stress I
Sc = 0. 84J [(30) (12) (1o6)/42] (1) (6. 9 / 0 . 34)
I
= 16,100 p s i
(c). 2024-T4 Aluminum on 2024-T4 Aluminum I
m
e
= 6. 03 x
=
lb-sec2/in. f r o m (2c)
0. 0012 (l/O. 5 t 1/0. 5) (6. 03 x 10-6) (20)2

I
f
a
= 0.0000116
I
f d = 0. 375 t 0. 062 [(O. 111) (0. 0012)/(10-6/24 t 10-6/12)]
I
= 72
f
m
= (0. 0000116) (72)/(72 + 0. 0000116) I
= 0. 0000116 I
wd
= 4 (0. 0000116) (144 - 0.0000116)
= 0. 041 l b
I
Total Load
WT = (0. 375 t 0. 041)/0. 062

I
= 6. 7 lb/in. I
Stress
1
SC = 0. 8 4 4 [(12) (12) (1o6)/24] (1) (6. 7/0.34)
= 13,500 p s i
I
m = 5.60 x lb-sec2/in. f r o m (2d)

I
e
f
a
= 0. 0012 (l/O. 5 t 1/0. 5) (5.60 x (20)2
I
= 0. 0000108 I
I
- 94 -
I
~
I fd
= 0. 375 to. 062 [(O. 111) (0. 0012)/(10-6/60 t 10'6/0. 4111
I = 6. 0
f = (6. 0) (0. 0000108)/(6. 0 t 0. 0000108)
1 m
= 0. 0000108
I Wd = 4 (0.0000108) (12 - 0. 0000108)

= 0. 0114
I Total Load
I WT = (0. 375 t 0. 0114)/0. 062
= 6. 25 lb/in.
I Stress
,/
~
I sc
= 0. 8 4 [(30) (0.41) (lO6)/3O
I (1) (6. 25/0. 34)
= 3,270 psi
1
I
I
1
I
I
1
I
I
1 - 95 -
I
APPENDIX C I
Calculation of T e s t G e a r Tooth Loads and S t r e s s e s Using
Buckingham’s Formulae and Tuplin’s Method f o r
Effective E r r o r s
I
F o r purposes of the calculations, the s a m e numbers used i n the
I
previous calculations will now b e used, with the following exceptions:
1. Actual E r r o r :
I
El = 2 (tooth-to-tooth e r r o r ) tan + I
2. R i m depth:
= 2 (0. 0004) (0. 364) = 0. 00029 in.
I
H = radius of g e a r - depth of tooth I
= 0. 500 - 0. 021 = 0.479 in.
3. Ratio of rim depth to c i r c u l a r pitch: I

a = H/pn = 0.479/0.032 = 1 5
The calculations a r e divided into t h r e e sections, one f o r e a c h value of

I
load and speed: I
1. V = 1000 ft/min, no load
.*. W t = 0 lb
I
(a). 303 Stainless Steel on 303 Stainless Steel I
l / k T = 3 ( 1 / E t 1 / E ) t l / G (0.125 t O.l/a)
t l/G
P g P
(0.125 t 0. l / a )
I
g
t 1 / 6 [ N p / ( a t 0. 2) E
P
t N / ( a t 0. 2) E
g g 1 I
= 3 ( 1 / 1 5 x l o 6 ) t ( 2 / 1 2 x l o 6 ) (0.125 t 0.1/15)
I
t 1 / 3 [(96)/(30 x l o 6 ) (15 t 0. 2) 1 1
I
- 96 -
I
l / k T = 1 2 x 10-6/60 + 1. 3 x 10'6/60 + 4 . 2 x 10-6/60

= ~ 9 10-72 ~
m = w/gF
m = m = (0. 014)/(386) (0. 062)

P i3
= 5.8 x
= 27r d 2 . 9 2 x 10'7)/(2/5. 8 x
= 5. 8 x sec
tl = p,/v
= (0. 032) (60)/(1000) (12) = 1. 6 x sec
tl/Tl = (1. 6 x 10-4)/(5. 8 x
= 2. 76
Ee/E1 = 0. 035 at t l / T 1 = 2. 76 f r o m Fig. 8
E
e
= 0.035 El
= (0. 035) (0. 00029) = 1. 01 x in.
f
a
= 0. 058 from page 87
fd
= 0. 062 [(o.
r 111)(1. 01 ~ i o - ~ ) / ( l 0 - ~ / 6 o + l o - ~ / 3 o ) ]
= 1.39
f
m
= (0. 058) (1. 39)/(1. 39 + 0. 058)
= 0.056
wd
= J(0.056) (2. 78 - . 056)
= 0. 390 l b
- 97 -
Total Load
W T = (0. 0 t 0. 390)/0. 062
= 6. 3 lb/in.
Stress
SC = 0. 8 4 4 [(30) (30) (106)/60] (1) (6. 3/0. 34)
= 19,500 p s i
m = 5 . 8 x l o m 5f r o m ( l a )
P
m = ( 0 . 0093)/(0. 125) (386)
g
= l . 9 4 x 10- 5
t (10-6/12) (0. 125 t 0 . 1 / 1 5 )
t (10-6/4) (0. 125 t 0. 1/15)
t 1/6
[96/(15 t 0. 2) (30 x l o 6 )
t 96/(15 t 0. 2) (12 x lob)]
= 21 x 10-6/60 t 0.66 x 10'6/60
t 1.98 x 10-6/60 t 2.10 x 10-6/60
t 5. 25 x 10-6/60
= 5 . 1 7 ~10-7
T1
= 2n J ( 5 . 1 7 x 10-7/(104/5. 8 t i o 4 / i . 94)
= 5.45 x 10-5 s e c
tl = 1.6 x sec f r o m (la)
- 98 -
JPL Technical M e m o r a n d u m No. 33- 139
= (1. 6 x 10-4)/(5. 45 x
tl/T1
= 2.94
Ee/El
= 0. 02 at t l / T 1 = 2. 94 f r o m Fig. 8
E = (0. 02) (0. 00029) = 0. 58 x

e
f
a
= 0.0348 f r o m page 88
= 0. 062 [(O. 111)(0. 5 8 ~ 1 0 - ~ ) / ( 1 0 - ~ / 6 0 + 1 0 - ~ / 1 2 ) ]

fd
= 0.40
f
m
= (0. 40) (0. 0348)/(0. 40 t 0. 0348)
= 0.032
Wd = J(0.032) (0.80 - 0. 032)

= 0. 157 l b
Total Load
W T = (0. 0 t 0. 157)/0. 062
= 2. 54 lb/in.
Stress
sC = 0. 84,/[(30) (12) (io6)/42] (1) (2. 54/0. 34)
= 9,750 psi
l / k T = 6 (10-6/12) t (10-6/2) (0.125 t 0.1/15)
t ( 1 / 3 ) [96/(15 t 0.2) (12 x lo6)]

= 6 x 10-6/12 t 0. 79 x 10-6/12 t 2.1 x 10-6/12
= 7.4 10-7
- 99 -
I
2.rrJ7.4 10-7)/(2/1.94 10-4)
I
5.32 x sec 1
1 . 6 x 10-4 sec f r o m ( l a )
(1.6 x 10'4)/(7.4 x 10-7)

I
3. 00 I
0. 015 at t l / T 1 = 3. 00 f r o m Fig. 8
I
(0. 015) (0. 00029) = 0 . 4 4 ~
10-5
0. 0197 f r o m page 89
I
r
0. 062 [(O. 111) (0. 4 4 ~ l O ~ ~ ) / ( l O - ~ / 2 4 + l 0 - ~ / 1 2 ) ] P
0. 0183
J (0.0183) ( 0 . 328 - 0. o m j
0. 075 l b
Total Load 1
W T = ( 0 . 0 t 0. 075)/0. 062
= 1. 21 lb/in.
I
Stress
1
SC = 0. 84,/[(12) (12) (106)/24] (1) (1. 21/0. 34)
= 5,740 psi
I
I
I
I
I
- 100 -
I

m = 5.8 x 1 0 - 4 f r o m ( l a )
P
m = (0. 0051)/(386) (0. 125)
g
= 1.06~10-4
l / k T = 3 (1OT6/3O t 10'6/0.41)
t (10-6/12) (0.125 t 0.1/15)
t (10-6/0. 16) (0.125 t 0. 1/15)
t 1 / 6 [96/(15 t 0.2) (30 x 106)
t 96/(15 + 0. 2) (0. 41 x 1 06)]

= 6 x 10'6/60 t 439 x 10-6/60 t 0.66 x 10-6/60
t 49. 5 x 10'6/60 t 2.1 x 10-6/60
t 154 x 10-6/60
= l O . 9 x 1 0 -6
T1 = 2a,/(lO. 9 x 10-4)/(104/5. 8 t 1 0 4 / l . 06)
= 19.6 x 10-5 s e c
= (1.6 x sec f r o m ( l a )
tl/T1 = (1.6 x 10'4)/(19. 6 x
= 0.818
Ee/E1 = 0. 36 a t t l / T 1 = 0. 818 f r o m Fig. 8
Ee = (0. 36) (0. 00029)
= 1 . 0 4 x 10-4 in.
f
a
= 0. 0232 f r o m page 90
- 101 -
f d = 0.062 [ ( O . l l l ) ( 1 . 0 4 ~ 1 0 - ~ ) / ( 1 0 - ~ /t10-6/0.41)]
60
= 0.458
f = (0. 458) (0. 0232)/(0.458 t 0. 0232)
m
= 0.0221
wd
(0.0221) (0. 916 - 0. 0221)
= 0. 140 lb
Total Load
W T = (0. 140 t 0. O ) / O . 062
= 2. 27 lb/in.
Stress
sc = 0. 84J[(30) (0.41) (106)/30] (1) (2. 27/0. 34)
= 1,960 p s i
2. V = 100 f t / m i n , 4 in. -oz load
W t = (4 in. -oz/O.5 in. ) (1 lb/16 oz)
= 0. 500 lb
tl’T1
= ( l O O O / l O O ) (2. 76) = 27. 6 f r o m ( l a )
:. E
e
= 0, Wd = 0
Total Load
W T = (0. 500 t 0. O ) / O . 062
= 8. 07 lb/in.
- 102 -
Stress
sC = 0. 8 4 J [(30) (30) (106)/60] (1) (8. 0 7 / 0 . 34)
= 22,400psi
T1
= (1000/100) (2. 94) = 29. 4 f r o m ( l b )
:. E, = 0, Wd = 0
Total Load
W T = (0.0 t 0. 500)/0. 062
= 8. 07 Ib/in.
Stress
sC = 0.84,/ [(30) (12) (io6)/42] (1) (8. o7/0. 34)
= 16,900 p s i
= (1000/100) (3. 00) = 30. 0 f r o m ( I C )

V T 1
:. Ee = 0, Wd = 0
Total Load
WT = (0.0 t 0. 500)/0. 062
= 8. 07 lb/in.
Stress
SC = 0. 84J[(12) (12) (lo6)/24] (1) (8. 07/0. 34)
= 14,700 p s i
- 103 -
(a). 303 Stainless Steel on Delrin
tl/T1
= (1000/100) (0.818) = 8. 18
.-.Ee = 0, Wd = 0
Total Load
WT = (0. 500 t 0. O ) / O . 062
= 8. 07 lb/in.
Stress
sc = 0. 84\/[(30) (0. 41) (106)/30] (1) (8. 07/0. 34)
= 3,720 p s i
3. V = 20 ft/min, 3 in. -oz load
(6 in. - o z f o r s t e e l on steel).
Wt = ( 3 in. -oz/O.5 i n . ) (1 l b / 1 6 02)
= 0. 375 l b
(0. 750 lb f o r steel on steel)
tl’T1
= (1000/20) (2. 76) = 138 f r o m ( l a )
.*. E
e
= 0, Wd = 0
Total Load
W T = (0. 0 t 0. 750)/0. 062
= 12. 1 lb/in.
Stress
I
sC = 0. 84/[(30) (30) (106)/60] (1) (12. 1/0. 34) D
= 27,400 p s i
I
- 104 -
= (1000/20) (2.94) = 147

VT1
Total Load
WT = (0.0 t 0. 375)/0. 062
= 6. 0 5 lb/in.
Stress
SC = 0. 84/[(30) (12) (106)/42] (1) (6. 05/0. 34)
= 15,100 psi
= (1OOO/2O) (3.00) = 150

tllT1
.'. Ee = 0, Wd = 0
Total Load
W T = (0. 0 t 0. 375)/0. 062
= 6. 0 5 lblin.
Stress
sC = 0. 84J/i(12) (12) (106)/24] (1) (6. 05/0. 34)
= 12,800 p s i
tl/T1
= (1000/20) (0.818) = 41
.*. E
e
= 0, Wd = 0
- 105 -
J P L T e c h n i c a l Memorandum No. 33- 139
Total Load
wT
(0. 0 + 0. 3 7 5 ) / 0 . 0 6 2
6. 05 lb/in.
SC = 0. €344 [(30) ( 0 . 41) ( 1 0 6 ) / 3 0 ] ( 1 ) ( 6 . 0 5 / 0 . 34)
= 3,220 psi
- 106 -
APPENDIX D
Wear Rate Data
1. 1000 ft/min, no load
Dec. 19, 1962 - Jan. 4, 1963:
Time, Average backlash,

hr milli -in.
l** 2 3 4 5 6
0 4. 71 4. 33 4. 38 4. 71 3. 43 4. 60
5. 0 4. 88 4.43 4. 60 4. 60 3. 50 4. 77
25 5. 71 4. 71 5. 10 5. 17 3. 33 5.10
50 6. 38 4. 67 5. 50 5. 10 3. 33 5. 27
75 8. 43 4. 83 5. 94 5. 21 3.43 7. 05
100 9.33* 5. 05 5. 94 5. 60 3. 76 10.08
150 L _
5. 60 10. l* 5.67 3.83
200 - 5. 67 - 5.77 -
*Test terminated - wear reached 0.004 in.
**Material Combination Code
1. 303 Stainless Steel on 2024-T4 Aluminum
2. 303 Stainless Steel on Delrin
3. Anodized 2024-T4 Aluminum on Anodized 2024-T4

Aluminum
4. 303 Stainless Steel on 303 Stainless Steel
5. 303 Stainless Steel on Anodized 2024-T4 Aluminum

Treated with Molybdenum Disulphide
6. 303 Stainless Steel on Anodized 2024-T4 Aluminum
- 107 -
2. 100 ft/min, 4 in. -02 load

I
Nov. 2, 1962 - Dec. 19, 1962:
I
hr
1 2
milli -in.
3 4 5 6
I
0 4. 12 3. 95 4. 22 4. 95 3. 05 4. 67 I
4.95 4. 62 3. 83 3. 83 3. 66 4. 05
5. 0
25 5. 38 5. 05 5. 45 4. 73 3. 67 5. 50
1
50 6. 1 1
6. 05
5. 05
4. 88
6. 00
6. 22
4. 72
4. 88
3. 78
3. 72
6. 00
6. 22
1
75
100 6. 84 5. 33 6. 78 5. 16 4. 12 7. 00 I
150 8. l l * 5. 67 8. 05* 4. 95 4.22 8. 34*
200 5. 62 - 6. 34 4. 50
1
-
300
400
6. 34
6. 16 -
7. 00
6. 05
4. 05
4.16
B
500 6. 55 - 6. 83 4. 50
600 6. 77 - 7.88 4.00
700 7. 16 - 8.60 4.43
800 8.17* - 9. 0O* 5.05
*Test terminated - wear reached 0. 004 in.
- 108 -
3. 20 f t / m i n , 3 in. -02 load
(6 in. -02 for steel on steel)
Jan. 10, 1963 - Feb. 22, 1963:

hr milli -in.
1 2 3 4 5 6
0 4.83 4. 77 5. 05 5.44 2. 67 4. 67
115 5. 1 7 4. 80 5. 55 7. 17 2. 70 4. 82
237 5. 26 4. 9 3 5. 55 7. 60 2. 76 5. 36
311 5. 50 4. 96 5. 84 7. 93 2. 75 5. 63
571 6. 05 5. 04 6. 84 8. 7 7 2. 81 6. 08
740 6. 75 5. 1 7 7. 50 9. 50 2. 87 6. 64
- 109 -
APPENDIX E
Calculations of W e a r Rates
1. 1000 f t / m i n , no-load, 3800 r p m
Wear Depth Rate = (final - initial backlash)

( r p m ) (60 m i n / h r ) (time)
= (5.77 - 4.71) ( l o m 3 ) / (3800) (60) (200)
= 23. 2 x in. / c y c l e
W e a r r a t e equally divided between pinion and g e a r
:. w = 1 / 2 (23. 2 x
-1 2
= 11.6 x 1 0 in. /cycle
(b). 303 Stainless on 2024-T4 Aluminum
W e a r Depth Rate
= (9. 33 - 4. 71) (10'3)/(3800) (60) (100)
= 202 x 10-1 2 in. / c y c l e
W e a r r a t e totally on aluminum
.*. w = 202 x 1 0 - l ~
in. / c y c l e
(c). 303 Stainless on Anodized 2024-T4 Aluminum
Wear Depth Rate
= (lo. 0 - 4.60) (10-3)/(3800) (60) (100)
= 237 x 1 0-1 2 in. /cycle

Wear rate totally on aluminum
.'. w = 237 x 10-1 2 in. / c y c l e
- 110 -
303 Stainless Steel on Anodized 2024-T4 Alumni-

W e a r Depth Rate
= (3.83 - 3.43) (10-3)/(3800)(60)(150)

-12
= 14.6 x 10 in. / c y c l e
-1 2
:. w = 14. 6 x 10 in. /cycle
303 Stainless Steel on Delrin

W e a r Depth Rate
= (5. 67 - 4. 33) (10-3)/(3800)(60)(200)

= 29.4 x in. / c y c l e
W e a r r a t e totally on d e l r i n
-1 2
... w = 29.4 x 10 in. / c y c l e
Anodized 2024-T4 Aluminum on Anodized 2024-T4

Aluminum
W e a r Depth Rate
= (lo. 1 - 4.38) (10’3)/(3800) (60)(150)

= 164 x in. / c y c l e
W e a r r a t e equally divided between pinion and g e a r
.*. w = 1/2 (164x
= 82 x in. / c y c l e
- 111 -
2. 100 ft/min, 4 in. -02 load, 380 r p m
( a ) . 303 Stainless Steel on 303 Stainless Steel
Wear Depth Rate
= (9. 00 - 4. 95) (1Om3)/(38O)(60) (800)

= 223 x 10-1 in. /cycle
Wear r a t e equally divided between pinion and g e a r
;. w E 1 / 2 (223 x 10-l')
= 111 x 1 0 - l in.
~ /cycle
303 Stainless Steel on 2024-T4 Aluminum
W e a r Depth Rate
= (8.11 - 4. 12) (lC1-~)/(380)(60) (150)
= 1170 x in. / c y c l e
:. w = 1170 x 1 0-1 2 in. / c y c l e

(c). 303 Stainless Steel on Anodized 2024-T4 Aluminum
Wear Depth Rate

1
= (8. 34 - 4. 67) (1Oe3)/(38o) (60) (150)
= 1070 x 10-1 in. / c y c l e
I
Wear rate totally on aluminum I
:. w = 1070 x in. /Cycle
I
II
I
1
I
(d). 303 Stainless Steel on Anodized 2024-T4 Aluminum

Wear Depth Rate
= (5. 05 - 3. 05) (lC1-~)/(380)

(60) (800)
= 110 x 1 0 - l in.
~ /cycle
*. w = 110 x 1 0 - l in.
~ /cycle
(e). 303 Stainless Steel on Delrin
Wear Depth Rate
= (8. 1 7 - 3. 95) (10-3)/(380) (60) (800)

= 232 x 10 - l 2 in. /cycle
Wear r a t e totally on d e l r i n
:. w = 232 x in. /cycle
(f). Anodized 2024-T4 Aluminum on Anodized 2024-T4

Aluminum
Wear Depth Rate
= (8. 05 - 4.22) (10-3)/(380) (60) (150)
= 1120 x 10-l’ in. j c y c i e
Wear r a t e divided equally between pinion and g e a r
:. w = 1 / 2 (1120x
= 560 x in. /cycle
- 113 -
3. 20 ft/min, 3 in. -oz load (6 in. -oz load f o r s t e e l on steel),

76 rpm
(a). 3 0 3 Stainless Steel on 3 0 3 Stainless Steel
Wear Depth Rate
= ( 9 . 50 - 5 . 4 4 ) ( 1 0 m 3 ) / ( 7 6 ) ( 6 0 ) (740)
= 1 2 1 0 x 1 0 - l ~in. / c y c l e
Wear rate equally divided between pinion and g e a r
:. w = 1 / 2 (1210 x l o - 1 2 )
= 605 x 1 0-l' in. / c y c l e
(b). 303 Stainless Steel on 2 0 2 4 - T 4 Aluminum
Wear Depth Rate
= (6. 75 - 4. 83) ( 1 0 m 3 ) / ( 7 6 ) (60) ( 7 4 0 )
= 5 7 5 x 1 0 - l ~in. / c y c l e
Wear r a t e totally on aluminum
-1 2
.: w = 5 7 5 x 10 in. / c y c l e
(c). 303 Stainless on Anodized 2 0 2 4 - T 4 Aluminum
Wear Depth Rate
= ( 6 . 64 - 4. 67) ( 1 0 - 3 ) / ( 7 6 ) ( 6 0 ) (740)
= 585 x in. / c y c l e
Wear rate totally on aluminum
:. w = 5 8 5 x 1 0-l2 in. / c y c l e
- 114 -
303 Stainless Steel on Anodized 2 0 2 4 - T 4 Aluminum

W e a r Depth Rate
= (2. 87 - 2. 67) ( 1 0 - 3 ) / ( 7 6 ) (60) (740)
= 5 9 . 4 x 1 0 - l ~in. /cycle
.e. w = 5 9 . 4 x 1 0 - l ~in. /cycle
303 Stainless Steel on Delrin
W e a r Depth Rate
= (5. 1 7 - 4. 77) ( 1 0 m 3 ) / ( 7 6 )(60) ( 7 4 0 )
= 118 x in. / c y c l e
W e a r r a t e totally on d e l r i n
.'. w = 118 x 10
-1 2 in. / c y c l e
Anodized 2 0 2 4 - T 4 Aluminum on Anodized 2 0 2 4 - T 4

Aluminum
W e a r Depth Rate
= ( 7 . 50 - 5.05) ( 1 0 - 3 ) / ( 7 6 ) ( 6 0 ) ( 7 4 0 )
= 726 x in. / c y c l e
Wear r a t e equally divided between pinion and g e a r
-12
... w = 1/2 ( 7 2 6 x 1 0 )
= 363 x in. /cycle
- 115 -
APPENDIX F
Calculation of Fatigue Lives
1. V = 1000 ft/min, no-load, 3800 r p m
F r o m Fig. 9
c = 03 at s t r e s s = 19,500 p s i
E = c/60xn
= 00/(60) (3800)
= 03 hr
From Fig. 9
c = 03 at s t r e s s = 9,750 p s i
E = m / ( 6 0 ) (3800)
= 00 hr
F r o m Fig. 9
c = 00 a t s t r e s s = 5,740 p s i
E = 03/(60) (3800)
= 00 hr
2. V = 100 f t / m i n , 4 in. - o z load, 380 r p m
From Fig. 9
c = CXI a t s t r e s s = 22,400 p s i
E = 00/(60) (380)
hr
='a
- 116 -
(b). 3 0 3 Stainless Steel on 2024-T4 Aluminum
F r o m Fig. 9
c = 6x lo1* at s t r e s s = 16,900 p s i
E = (6 x 1O1O)/(6O) ( 3 8 0 )
= 2.6 x 106 h r
F r o m Fig. 9
c = 1 x 1 0 l 1 a t s t r e s s = 14, 700 psi
E = (1 x 1011)/(60) ( 3 8 0 )
= 4 . 3 l~o 6 hr
3. V = 2 0 ft/min, 3 in. -oz load (6 in. -oz f o r s t e e l on steel),

76 r p m
(a). 3 0 3 Stainless Steel on 3 0 3 Stainless Steel
F r o m Fig. 9
c = 00 a t s t r e s s = 27,400 psi
E = 00/(60) ( 7 6 )
= 00 hr
(b). 3 0 3 Stainless Steel on 2024-T4 Aluminum
F r o m Fig. 9
c = 1 x 1 0 l 1 at s t r e s s = 15,100 psi
E = (1 x 1011)/(60) (76)
= 22 x l o 6 h r
- 117 -
From Fig. 9
c = 3 x 1011 a t s t r e s s = 12,800 psi
E = ( 3 x 1011)/(60) (76)
= 66 x l o 6 hr
- 118 -

Gear Design

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Gear Design

Uploaded by

Copyright:

Available Formats

1

Prepared Under Contract No. NAS 7-100

I JPL Technical Memorandum No. 33-139

11. Friction, Surface Damage, and Wear . . . . . . . . . . . 4

V. Gear Wear T e s t s and Results . . . . . . . . ... . . . . 46

I IV. Summary . . . . . . . . . . . . . . . . . . . ... . . .. 76

I Appendix B: Calculation of Test-Gear Tooth Loads and

S t r e s s e s Using Buckingham' s Formulae and

Calculation of Fatigue Lives . . . . . . . . . . . .

I 3. S t r e s s e s and calculated fatigue lives f o r t e s t g e a r s . . . . 75

2. Comparison of calculated p a r t i c l e s i z e and m e a s u r e d

5. Dimensions f o r gear teeth and rim . . . ......... 36 I

8. Comparison of ratio of insertion time to n a t u r a l period

9. Surface fatigue curves for 303 s t a i n l e s s s t e e l and

10. Gear t e s t f i x t u r e . ..................... 48

Wear r a t e f o r 303 s t a i n l e s s s t e e l on 2024-T4

19. Wear r a t e f o r 303 stainless s t e e l on anodized 2024-T4

20. Wear r a t e f o r 303 stainless s t e e l on anodized 2024-T4

21. Wear r a t e f o r anodized 2024-T4 aluminum on anodized

22. Wear r a t e f o r 303 stainless s t e e l on d e l r i n ( 4 in. -oz,

23. Wear r a t e f o r 303 stainless s t e e l on 303 s t a i n l e s s

25. Wear r a t e f o r 303 stainless s t e e l on anodized 2024-T4

27. Wear r a t e f o r anodized 2024-T4 aluminum on

28. Wear r a t e f o r 303 stainless s t e e l on d e l r i n ( 3 in. -oz,

29. Design curve f o r 303 stainless s t e e l indicating

30. Design curve f o r 2024-T4 aluminum indicating

3 1. Design curve f o r anodized 2024-T4 aluminum indicating

33. Design curve f o r d e l r i n indicating calculated s t r e s s and

concepts of wear and the problems associated with applying t h e s e con-

deemed necessary to analyze thoroughly the g e a r -wear problem. 1

to determine the w e a r r a t e s f o r the m o s t popular m a t e r i a l s in u s e

1. 303 stainless steel 1

3. Anodized 2024-T4 aluminum

rials can be used to analyze w e a r r a t e s of a g e a r train. In addition,

the expected life of a s y s t e m can be determined, realizing that s o m e

In addition to the data presented i n the w e a r c h a r t s , this paper I

popular methods of computing dynamic load; namely, the American

Wear is the l o s s of m a t e r i a l f r o m s u r f a c e s that slide and b e a r on

each other. Almost all mechanisms m u s t have v e r y p r e c i s e sliding

s u r f a c e s in o r d e r to function properly.k The deterioration of these

s u r f a c e s because of w e a r usually h a s a detrimental effect on the p e r -

f o r m a n c e of such mechanisms. Although wear o c c u r s f r o m the t i m e

that these s u r f a c e s come into contact, the w e a r rate determines the

life of the mechanism until i t is deemed useless.

Studies on w e a r have been somewhat neglected historically because

of the inability to accurately m e a s u r e wear on surfaces. This w a s not

a problem in the past, however, because of the inability to hold close

tolerances. F o r example, tolerances as l a r g e as a couple of thou-

sandths of a n inch w e r e as close a s could be held, and a w e a r of a few

tenths of thousandths of an inch w a s not important. The "state of the

art" has p r o g r e s s e d , however, and tolerances of a few tenths of thou-

sandths of an inch are held quite easily. A wear of a few tenths of

thousandths of an inch becomes relatively important in this case.

Recent introductions of radioactive-isotope measuring techniques

make a c c u r a t e and reproducible w e a r r e s u l t s possible, which w i l l

undoubtedly stimulate i n t e r e s t in the subject of w e a r in the future.

In the discussion of w e a r i t should be pointed out that t h e r e are

specific methods to combat w e a r . One of the m o s t significant ways to

reduce the w e a r r a t e i s with a lubricant. A lubricant i s a low s h e a r -

strength m a t e r i a l placed between the sliding surfaces. A lubricant

does not prevent wear, but, reduces it by reducing the number of

contact points between the sliding s u r f a c e s . Although a lubricant is

generally thought of a s an oil, g r e a s e , o r s o m e s o r t of dry-film, some