Professional Documents
Culture Documents
SAE TECHNICAL
PAPER SERIES 2000-01-2565
400 Commonwealth Drive, Warrendale, PA 15096-0001 U.S.A. Tel: (724) 776-4841 Fax: (724) 776-5760
THIS DOCUMENT IS PROTECTED BY U.S. AND INTERNATIONAL COPYRIGHT
It may not be reproduced, stored in a retrieval system, distributed or transmitted, in whole or in part, in any form or by any means.
Downloaded from SAE International by General Motors LLC, Monday, December 10, 2012 10:19:53 PM
The appearance of this ISSN code at the bottom of this page indicates SAE’s consent that copies of the
paper may be made for personal or internal use of specific clients. This consent is given on the condition,
however, that the copier pay a $7.00 per article copy fee through the Copyright Clearance Center, Inc.
Operations Center, 222 Rosewood Drive, Danvers, MA 01923 for copying beyond that permitted by Sec-
tions 107 or 108 of the U.S. Copyright Law. This consent does not extend to other kinds of copying such as
copying for general distribution, for advertising or promotional purposes, for creating new collective works,
or for resale.
SAE routinely stocks printed papers for a period of three years following date of publication. Direct your
orders to SAE Customer Sales and Satisfaction Department.
Quantity reprint rates can be obtained from the Customer Sales and Satisfaction Department.
To request permission to reprint a technical paper or permission to use copyrighted SAE publications in
other works, contact the SAE Publications Group.
No part of this publication may be reproduced in any form, in an electronic retrieval system or otherwise, without the prior written
permission of the publisher.
ISSN 0148-7191
Copyright © 2000 Society of Automotive Engineers, Inc.
Positions and opinions advanced in this paper are those of the author(s) and not necessarily those of SAE. The author is solely
responsible for the content of the paper. A process is available by which discussions will be printed with the paper if it is published in
SAE Transactions. For permission to publish this paper in full or in part, contact the SAE Publications Group.
Persons wishing to submit papers to be considered for presentation or publication through SAE should send the manuscript or a 300
word abstract of a proposed manuscript to: Secretary, Engineering Meetings Board, SAE.
Printed in USA
THIS DOCUMENT IS PROTECTED BY U.S. AND INTERNATIONAL COPYRIGHT
It may not be reproduced, stored in a retrieval system, distributed or transmitted, in whole or in part, in any form or by any means.
Downloaded from SAE International by General Motors LLC, Monday, December 10, 2012 10:19:53 PM
2000-01-2565
Coaxial splines are also frequently found inn As demonstrated later in the paper, there is a fourth
electromagnetic clutches. It is these types of application technology required in this evaluation, that of mechanics.
that serve as the focal point for this discussion. A clutch Failure models that consider the interactions between
typically transmits torque through frictional contact of two various loading conditions have not been found in the
rotating surfaces. One of these surfaces must be free to literature. Therefore, a model was developed and
move axially so that the clutch can be engaged or presented which determines the loads, sliding speeds,
disengaged. Once engaged, the clutch torque is driven and other parameters involved in the spline teeth
through the spline. One design, common among interface.
industrial clutches, is to have teeth of the splined hub
made of plastic, while the teeth of the mating part, the
clutch armature, are of low carbon steel. Whereas in
THIS DOCUMENT IS PROTECTED BY U.S. AND INTERNATIONAL COPYRIGHT
It may not be reproduced, stored in a retrieval system, distributed or transmitted, in whole or in part, in any form or by any means.
Downloaded from SAE International by General Motors LLC, Monday, December 10, 2012 10:19:53 PM
The literature does not identify a direct relationship Debris plays a complex role in fretting wear and
between the slip amplitude of gears and wear. Figure 1 corrosion, along with the material and the combinations
shows the typical wear mechanisms found in gear of materials. Fretting wear and corrosion change with
designs [2]. In evaluating wear characteristics of gear time, and are situation dependent [5]. While true models
teeth, the scoring criterion is commonly used. Should have not yet been developed that predict fretting wear,
the product of load and velocity fall below a limit, scoring there are several observations that are useful. For
will not occur. example, it is usually found that the fretting wear rate is
nearly proportional to the load [6].
FRETTING FATIGUE
The life prediction models used for gears are of little use
in evaluating coaxial splines for two reasons. First, they
rely on the existence of lubrication. Secondly, although Figure 2: Regions of crack growth.
they are the result of decades of experience, they are
also focused on gear reducers with external teeth. Nishioka and Hirakawa also related the reduction in the
fatigue strength to the slip amplitude. It is claimed that
FRETTING WEAR AND CORROSION slip amplitudes of 7 to 26 micrometers produce
significant reductions in fatigue strength [11]. There is a
Outside of gear design, a major characteristic that damage threshold theory which is related to a cycle and
determines the type of wear which will occur is the slip pressure threshold [12]. Fretting fatigue theories do not
amplitude. With slip amplitudes less than about 300 address the issue of wear, as the majority of the models
micrometers, the wear mechanism will be fretting, higher predict crack behavior and effects on the fatigue life
amplitudes are associated with reciprocating wear [4]. diagram.
The difference between fretting wear and corrosion is the
effect of the environment. Both can occur at low WEAR
amplitude sliding. With fretting wear there is adhesion
and abrasion of the surfaces. Fretting corrosion involves Reciprocating wear is observed at amplitudes over 300
the oxidation of the surface as well. At sliding micrometers (.012 inches). There are many theories to
amplitudes of less than about 70 micrometers, fretting predict wear, the most popular being Archard’s wear
oxidation occurs. Above 70 micrometers, both fretting theory, relating wear to the load [13]. In 1956, Archard
wear and fretting corrosion can occur [4].
THIS DOCUMENT IS PROTECTED BY U.S. AND INTERNATIONAL COPYRIGHT
It may not be reproduced, stored in a retrieval system, distributed or transmitted, in whole or in part, in any form or by any means.
Downloaded from SAE International by General Motors LLC, Monday, December 10, 2012 10:19:53 PM
[14] revised his wear model to incorporate the material The method using the height of wear, h=KPV, is related
hardness. He proposed a nondimensional wear to plastics and the PV limit. The PV limit is the transition
coefficient, k, as given in Equation (1). from mild to severe wear [20]. It assumes that the area
of contact remains constant. For non-lubricated
systems, particularly plastics, a sliding temperature
3PV criteria usually dictates the PV limit [20]. The PV limit
k= (1) model is quite similar to the scoring criteria for gears.
LX Both of them are temperature related, one of lubricant,
Where the other of plastic.
P = Indentation hardness of the softer material,
ADHESION
V = Volume of material lost,
L = Load, and
X = Total sliding distance. Adhesion is not a term typically used in gear design. It
does, however, relate to fretting and wear. Adhesion
Stowers and Rabinowicz [15] then proposed the model in involves material transfer or the production of debris.
Equation (2) relating the coefficient of friction, f, to Sikorskii [21] predicts it to be a function of the surface
Archard’s wear coefficient, k. energies. Crack nucleation is thought to occur at the time
of maximum adhesion. The adhesion of asperities was
found to be a key element in many of the wear theories.
k = Kf n (2)
Stowers and Rabinowicz found the constant K to be ABRASION
based on material combinations and to be typically about
.0005 [15]. Additionally, the value for n has been found Abrasion, along with adhesion, plays an important role in
to be about 4. The wear coefficient, k, versus friction the theories regarding wear. Abrasion is the plastic
coefficient, f, would then look like Figure 3 [16, 17]. deformation of asperities and the production of particles.
These particles can then abrade the surface. There
were no models for abrasion found, although Midlin [22]
did develop a model of elastic asperity deformation. In
gear applications, abrasion is observed at low speeds
and high loads.
of the simpler wear models use the load to a power the coefficient of friction. Their prediction of the
greater than one. Later in the paper, fretting maps will coefficient of friction is shown in Figure 5 where the
be used to further demonstrate the impact of load on variable f is the interfacial film strength or strength of
wear. adhesion.
Severe Wear
SLIP SPEED
CYCLES white cast irons can also increase wear resistance, along
with higher levels of austenite. The presence of
With respect to gears, scoring occurs first, and then chromium and nickel also serve to reduce scuffing.
pitting is said to occur after 20,000 to 500,000 cycles Small concentrations of carbon in iron are quite effective
[27,28]. There are mixed opinions regarding the type of in reducing adhesion. The use of identical or similar
wear that occurs in the early stages. Debates involve materials has mixed results. Their adhesion is usually
whether severe wear occurs or not [5]. In fretting fatigue, higher, but if their hardness and surface finish are
cycles do not have a significant effect on the life. It is different, wear reductions can be achieved [26].
observed, however, that the growth rate of the crack
decreases as it moves further from the surface, due to SURFACE FINISH
the decreasing effect of the surface stresses [8]. In
reciprocating wear theories, the number of cycles has Surface finish plays an important role in gears, as it
not been identified as an influencing parameter either. usually effects the function of the lubricants. If one
surface is rougher than the other, they will eventually
CYCLE FREQUENCY wear to the same level of roughness [34]. In fretting
wear, however, a hard rough surface is beneficial, as it
The cycle rate of sliding only has an effect on fretting allows a place for debris to collect without causing future
wear and corrosion. With mild steel, the wear decreases damage. If the debris is allowed to oxidize, however, the
as the frequency increases from 0 to 30 hertz. At benefit is lost. To reduce fretting fatigue of dry surfaces,
frequencies over 30 hertz, oxides no longer affect the Hoeppner [12] recommends a finish of less than 16 or
wear rate [29]. In the case of coaxial splines, the greater than 63 microinches Ra. With respect to
minimum frequency at which the oxides are effective polymers, the asperity size was found to impact the wear
may increase due to the extended time that the teeth rate as well [33].
surfaces are not in contact. Thirty hertz, or 1800 rpm, is
the speed of the most motors. This suggests that the CONTACT AREA
frequency produced in the majority of the direct motor
drive applications, as well as those with a speed reducer, While a larger contact area reduces the contact stresses,
is significant. it also creates more of an opportunity for wear particles
to become trapped, producing plowing wear [6]. In the
COMPONENTS PARAMETERS case of gear teeth, this is less of an issue because the
teeth are not always in contact, allowing particles to
SLIDING AMPLITUDE freely escape.
maps to establish the effectiveness of changes to the three-dimensional map. This map could then be used to
design. evaluate any displacement and force combination. An
examples of such a 3-D map is shown in Figure 9.
While the fretting maps proposed by Vingsbo and One rule of thumb which has been used by gear
Soderberg help to set a standard for evaluation, they engineers is to design the spline with the assumption that
might be improved to enhance usefulness. In the case half of the teeth make contact. While this may produce a
of Figures 7 and 8, the ordinates remain identical while functional design, it is not based on an understanding of
the abscissas change. While the displacement in Figure the limits of the gear set and therefore, it is not a good
7 changed, the load must have been fixed, and visa method of predicting their performance. Clearly,
versa for Figure 8. A suggestion would be to create depending on the loading and geometry, more teeth may
each map at various displacements and loads. Then
findings could be interpolated between them to form a
THIS DOCUMENT IS PROTECTED BY U.S. AND INTERNATIONAL COPYRIGHT
It may not be reproduced, stored in a retrieval system, distributed or transmitted, in whole or in part, in any form or by any means.
Downloaded from SAE International by General Motors LLC, Monday, December 10, 2012 10:19:53 PM
TOOTH GEOMETRY 2
− x 4 + 4x 3 rb + 5x 2 rb2 + 2xrb3 (7)
rb
In order to determine the deflection of the tooth under where
load, the thickness of the tooth along its entire length
must be known. The geometry of a spline tooth is found Pd = Pitch diameter.
in the study of involutometry. Equations (3) through (6)
calculate tooth thickness for an external spline at any Unfortunately, this equation yields the arc thickness of
radius, as depicted in Figure 11. the tooth. In order to determine the area moment of
inertia, which is required to calculate the tooth deflection,
the cross sectional thickness, t’, of the tooth is required,
æ tp ö not the arc thickness. The tooth thickness, t’, can be
t(r) = 2rç + invφ − invϕ ÷ (3) determined by taking the sine of the arc angle at r times
ç 2r ÷
è p ø the radius at r as shown in Equation (8) and Figure 12.
This, however, will result in a thickness that is actually at
invφ = tanφ − φ (4) a radius less than r, and it is the tooth thickness at the
radius that is desired.
r æ t(x) ö
r= b (6) t' (x) = 2rsin ç ÷ (8)
cosϕ è 2r ø
A direct solution is not readily available because the Which reduces to Equation (14).
tooth thickness is a complex function of r. The exact
value at r may be found by iterative solutions. For the 12
x x
M(x) M(x)
L
Eb ò0 ò0 (t(x) )3
purpose of this evaluation, however, an approximate y(x) = dx − ò dxdx (14)
0 (t(x) )
3
adjustment to r will be made. The adjustment is equal to
the residual of the cosine component to produce r’ as
calculated in Equation (9) and shown in Figure 12. The Due to the complexity of the tooth thickness equation,
remaining error will then be less than the difference in the deflection is best solved using numerical integration.
tooth thickness at r and r’. In this evaluation, the rectangular rule was used with 100
divisions of the tooth. Now a load can be applied and the
ì æ t öü moments determined for each location on the tooth.
r' = r í2 − cosç ÷ý (9) Because the actual loading is unknown at this time, a
î è 2r øþ
point load on the tooth will be assumed. The resulting
deflection curves are shown in Figure 13, for a 10 pitch,
The previous evaluation is for an external splined tooth. 25 tooth spline. The same evaluation is then performed
The calculations for an internal tooth are the same, for the internal tooth. The only change required is the
except that its circular tooth thickness, tp, would be the tooth width and direction of the force.
circular space width at the pitch line. The circular space
width, tE(x), could then be calculated for the external
spline, and the internal tooth thickness along the length
of the tooth, tI(x), would become Equation (10).
2π
t I (x) = (x + rb ) − t E (x) (10)
N
Where
N = Number of teeth.
TOOTH BENDING
[I]{L1 − L2 } = [B]{t1 − t 2 + ∆1 + ∆ 2 − δ } (15) The remainder of the analysis involves the evaluation of
a set of gears using the Coaxial Spline Model. It is
Where for each pair of teeth assumed that the teeth are of perfect form and that the
gear set has the properties listed in Table 1.
L1 = Arc length between two external teeth,
L2 = Arc length between two internal teeth, Number of Teeth 25
t1 = Arc tooth thickness of the external tooth, Pitch 10 /in
t2 = Arc tooth thickness on the internal tooth, Pressure Angle 30 degrees
∆1 = Deflection of the external tooth from the load, Offset .254 mm (.010 in)
∆2 = Deflection of the internal tooth from the load, Torque 67.8 Nm (50 ftlbs)
δ = Gap between the teeth, Material Mild Steel
Class of fit 6
é1 − 1 0 0 ù 0 mm (0 in)
ê0 1 − 1 0 ú Wear
[ B] = ê ect ú, and
ê0 0 1 −1 ú
ê ú Table 1: Baseline gear teeth properties.
ë0 0 0 1 û
The resulting load distribution on the teeth is that shown
[I] = Identity matrix.
in Figure 17.
The arc distance between the teeth, L, and the tooth
thickness, t, are known from the geometry of the
problem. The arc length between teeth, L, is determined
through the geometric relationship of the arc length
between the teeth to one center. This is where the
offset, ∆, has a significant effect. The deflections due to
the offset and gap between the teeth are unknown.
Therefore, Equation (15) becomes Equation (16).
The value for the right hand side of the above equation
can then be calculated. Now the freedom of rotation
needs to be accommodated. First the minimum result
from the right hand side of the equation is subtracted
from each tooth gap. Then the manufacturing clearance
and wear are added to each tooth gap from the right
Figure 17: Tooth loads.
hand side of the equation. Some teeth now have gaps
between them, whereas others have interference.
The baseline results are shown for the condition where
If there is interference, the gap between the teeth, δ, is the clearances between the teeth are at their maximum,
set to zero and the force from the deflection can be or .2235 mm (.0088 inches) for a class 6 fit. Clearly,
calculated as in Equation (17). there must be a distribution of these clearances. They
are not expected to decrease to zero, however, or axial
Fi = K∆ i (17) movement could not occur. The analysis was repeated
THIS DOCUMENT IS PROTECTED BY U.S. AND INTERNATIONAL COPYRIGHT
It may not be reproduced, stored in a retrieval system, distributed or transmitted, in whole or in part, in any form or by any means.
Downloaded from SAE International by General Motors LLC, Monday, December 10, 2012 10:19:53 PM
with an estimate, based on experience in industry, of the As expected, this is just slightly higher than the contact
tightest clearance likely to be produced, .0254 mm (.001 stress calculated if the load were divided by the entire
inches). tooth area. That stress is 396 MPa (57.4 ksi).
With the tighter clearance, as shown in the same figure, SLIDING DISTANCE AND SPEED
the tooth stresses are nearly double, and almost all of
the teeth make contact at the same time. Now consider Sliding distance and speed are both measures used to
the case where the teeth are worn by .457 mm (.018 predict the fretting or wear of materials. In the case of
inches), with the .0254 mm (.001 inch) manufactured coaxial splines, the contact of the teeth is periodic. The
clearance. Not only are there fewer teeth in contact, the time of contact is a function of the maximum number of
maximum load on the teeth is significantly reduced. teeth which can make contact. While the sliding
distance is fairly constant, cycle to cycle, the sliding
Several observations can be made from the results speed changes with angular position during the cycle.
shown in Figure 17. The first is that the tighter the fit, the
more each tooth will have to deflect, increasing its load. If, using the earlier convention, the shafts were offset by
Second, as the teeth wear, fewer teeth carry the load. an amount, ∆, the relative position of two opposing teeth
Also, the load imposed on the teeth by torque is
insignificant compared to the load from the teeth ∆ Sin (ω t )
S(t) = (18)
deflections. Cos(φ )
to each other, S(t), at a speed, ω, would be found with
The last and most important observation is that the tooth
Equation (18).
forces due to the offset shafts are extremely high. If the
surrounding system is truly rigid, the teeth would be
expected to fracture. In applications such as clutches, ∆ (1 − Sin (β ))
which are the focus of this paper, the system dS = (19)
surrounding the gear teeth is not rigid. The friction faces Cos (φ )
of the clutch may slip, placing a ceiling on the load and And the sliding amplitude, dS, would be determined with
the gear tooth requirement to bend. The amount of Equation (19).
slippage is not always a known quantity.
Here, β, is the angle where contact is first made. The
The typical clutch which would be used with this spline sliding distance during one cycle is twice the sliding
would have this radial force, about 8900 N (2000 lbs). amplitude.
Under these conditions, with steel teeth, the tooth loads
should be about 8900 N (2000 lbs), maximum, and The velocity, V(t), can now be found with Equations (20)
acting only on one tooth. and (21) by taking the absolute value of the first
The Coaxial Spline Model can be used to evaluate both slip region. Figure 9 illustrates how load and
clutches and couplings. Each of these power displacement data can be combined to predict slip and
transmission components was evaluated in light of wear characteristics. Changes in the environment,
various geometric, material and application parameters. frequency, contact area, and cycles are also expected to
In both the clutch and coupling, the largest reduction in have an effect on the fretting map.
contact stresses was obtained by reducing the Young’s
modulus. The sliding distance was also found to be a CONCLUSION
direct function of the offset, and the sliding speed was
proportional to the speed of rotation in both mechanisms. There are many factors presented that influence the
Increasing the pitch diameter was the most effective wear of spline teeth, and they are associated with a wide
means of reducing the maximum PV value for the clutch. variety of gear design, fatigue, and wear theories. While
each theory is unique, they do have common elements.
For the coupling, the application parameter with the The use of three-dimensional fretting maps predict wear
greatest impact was the offset of the splines. Once a rates as a function of load and slip amplitude. These
limit was exceeded, the loads increased exponentially. theories used in conjunction with fretting maps and the
For the clutch, the radial slip force influenced the loads results of the Coaxial Spline Model, will provide a better
the most. In both cases, the model predicts the best understanding of, and an enhanced ability to optimize,
method of reducing the contact stress and maximum PV coaxial spline designs.
value is by increasing the number of teeth in contact.
Whenever this happens, the loads decrease. CONTACT
In order to interpret the results of the Coaxial Spline Frank Bohm is a Principal Engineer and Team Leader in
Model to predict wear, there must be knowledge of the new product design with Warner Linear. He has a BSME
wear mechanisms; fretting wear and corrosion, fretting (1986) and a MSME (1998) from the University of Illinois
fatigue, and reciprocating wear. at Urbana-Champaign. The majority of his 14 years of
professional experience is with electromagnetic clutches
Fretting wear occurs at slip amplitudes less than 300 and brakes and linear and rotary actuators. The Coaxial
micrometers. Fretting oxidation can occur at slip Spline Model was development as part of his master’s
amplitudes less than 70 micrometers. Fretting fatigue is thesis project.
dependent on the stress level and occurs over a wide
range of slip amplitudes. Fretting fatigue produces REFERENCES
cracks that can create debris or fracture the component.
There are many theories involving reciprocating wear, 1. Lee, J. R., "Fretting in Helicopters", AGARG
which is observed at slip amplitudes over 300 Conference Proceedings, No 161, Specialist
micrometers. Several of them see reciprocating wear as Meeting on Fretting in Aircraft Systems, AGARG,
a direct or indirect function of the coefficient of friction, 1975, p. 3-1 to 3-10.
which also has many prediction theories. Some theories 2. Teeter, F. J. & Berger, M., "Wear Protection for
view the coefficient of friction as a function of surface Gears," Gear Technology, March, 1996, p. 27-29.
asperities. Additionally, the coefficient of friction has 3. Calistrat, M. M., "Gear Coupling Lockup: Causes
been shown to change with time, especially at startup and Cures," Power Transmission Design, October,
and when the components are disassembled. 1991, p. 31-32.
4. Ohmae, N., and Tsukizoe, T., "The Effect of Slip
Severe wear can occur when the stresses are within a Amplitude on Fretting," Wear, 27, 1974, p. 281.
certain range which removes the oxide layer from the 5. Bohm, Frank L., "Surface Durability of Dry Coaxial
material and exposes a soft core. Some theorize that Splines", MS Thesis, University of Illinois at Urbana-
severe wear occurs at startup. In fretting corrosion, Champaign, 1998
temperature and humidity play an important role. Cycle 6. Bill, R. C., "Review of Factors That Influence
frequency is important in fretting wear and corrosion, as Fretting Wear," Materials Evaluation Under Fretting
the growths of oxides on the surface are affected. Conditions, ASTM STP 780, American Society for
Testing and Materials, 1982, p. 165-182.
Changes to components can have an impact on the wear 7. Waterhouse, R. B., "The Role of Adhesions and
mechanisms as well. A reduction in slip amplitude Delaminations in the Fretting Wear of Metallic
affects whether there is stick, stick-slip, slip, gross slip or Material," Wear, 45, 1977, p. 355-364.
reciprocating wear. Material can have a significant effect 8. Endo, K., and Goto, H., 'Initiation and Propagation of
on the wear mechanisms. Changing surface roughness Fretting Fatigue Cracks," Wear, 38, 1976, p. 311-
can have several effects, as it changes the coefficient of 324.
friction and the production and behavior of debris. 9. Nowell, D. and Hills, D. A., "Crack Initiation Criteria
Generally, increasing contact area reduces wear. It in Fretting Fatigue," Wear, 136, 1990, p. 329-343.
does, however, increase the likelihood of trapped debris. 10. Szolwinski, M. P., Harish, Farris, T.N., Sakagami,
T., "The Use of Thermal Imiging for In-Situ
Fretting maps are perhaps the most useful tool in Characterization of Frictional Behavior in Mixed-
predicting wear behavior as it relates to the mixed stick- Mode Fretting Contact," World Wide Web (Internet),
THIS DOCUMENT IS PROTECTED BY U.S. AND INTERNATIONAL COPYRIGHT
It may not be reproduced, stored in a retrieval system, distributed or transmitted, in whole or in part, in any form or by any means.
Downloaded from SAE International by General Motors LLC, Monday, December 10, 2012 10:19:53 PM