See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/275382387

Helical Gears, Effects of Tooth Deviations and Tooth Modifications on Load

Sharing and Fillet Stresses

Article  in  Journal of Mechanical Design · March 2006

DOI: 10.1115/1.2167650

CITATIONS READS

26 419

3 authors:

Raynald Guilbault Claude Gosselin

École de Technologie Supérieure Involute Simulation Softwares
42 PUBLICATIONS   275 CITATIONS    37 PUBLICATIONS   448 CITATIONS   

SEE PROFILE SEE PROFILE

L. Cloutier
Laval University
39 PUBLICATIONS   365 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

CnC Gear Manufacturing View project

Spiral bevel gears View project

All content following this page was uploaded by Claude Gosselin on 11 August 2016.

The user has requested enhancement of the downloaded file.

 Helical Gears, Effects of Tooth
Deviations and Tooth
Raynald Guilbault
Professor
Modifications on Load Sharing
e-mail: raynald.guilbault@etsmtl.ca
Department of Mechanical Engineering,
Ecole de technologie superieure,
and Fillet Stresses
Montreal, H3C 1K3, Canada
Based on a few specific cases, this paper presents a comparative investigation of the
effect of helix slope and form deviation tolerances as specified by grades 5 and 7 of the
ANSI/AGMA ISO 1328-1 Standard for Cylindrical Gears. In addition, the consequences
Claude Gosselin of longitudinal flank crowning and radial tip relief modifications are investigated, as
applied on a misaligned helical gear set. For all simulations, the express model (Guil-
Louis Cloutier bault et al., 2005, ASME J. Mech. Des., 127(6), pp. 1161–1172) is employed. The bending
Professor
deflection and fillet stresses are obtained from a combination of finite strip and finite
difference meshes. The rolling-sliding motion of mating gear teeth is modeled with a cell
Department of Mechanical Engineering,
discretization of the contact area, which offers fast and accurate results. Similar contact
Laval University,
conditions arise from a helix slope deviation or a misalignment of the gear set: the first
Québec, G1K 7P4, Canada
contact point is driven to a theoretical contact line endpoint. Such a condition produces
a localized, and clearly impaired, contact area subject to overloading. Consequently,
flank crowning and tip relief corrections must be carefully regarded in the design process.
The presented results highlight that, if improperly combined, profile modifications can
amplify the overloading condition. 关DOI: 10.1115/1.2167650兴

1 Introduction the reference section 关2–9兴. One or more of these aspects, under
specific meshing conditions, are explored in these studies. For
With the industrial revolution, early gears made of wood were
instance, Wagaj et al. 关5兴 present a parametric study of the effects
transformed into modern gears, offering remarkable tooth shapes. of profile modifications. The authors illustrate the influence of a
Nowadays, gear design is largely standardized. Organizations perfectly assembled gear set over one mesh cycle.
such as AGMA or ISO supply standards that establish accuracy The main objective of this paper is to illustrate the need for the
classification, rating formulas, etc. Common gear practice also careful study of variation in design parameters, especially the po-
offers common design techniques that are largely employed, but tential consequences, using specific case studies. The intention
not always fully or explicitly normalized, and which refer to past here is not to portray the effects over the complete meshing period
experience as a guide. Tooth tip relief is a good example. In this of a tooth pair, which could be cumbersome or could average
context, the gear design process could appear to be a straightfor- some local upshots. Instead, it is preferred to select a particular
ward and easy task, with no deficient result likely to occur. In position that includes all possible contact types at the same time.
reality, however, the guidelines provided by the standards should From our point of view, this will offer a clearer understanding of
be combined with complementary experience and analysis, other- the phenomenon and be more condusive to general conclusions.
wise poor gear sets may result. The gear designer should be able The express model 关10兴 is used for all simulations. This model
to predict the performances of a gear pair from its accuracy grade combines the finite strip method with a pseudo 3-D model of the
and assembly quality. tooth base solved with the finite differences method to establish
The ANSI/AGMA ISO 1328-1 Standard for Cylindrical Gears tooth displacement and fillet stresses. This representation is
关1兴 sets 13 gear accuracy grades with allowable helix slope and coupled to a cell discretization of the contact area in order to
form deviations. While these inspection items are not individually determine the contact pressure distribution.
subject to mandatory tolerances, they are recognized as having a
significant influence on the performance of a gear pair. A com-
parative investigation is conducted on the effects of helix slope 2 The Express Model Brief Description
and form deviation tolerances, as specified by grades 5 and 7 for The load distribution between and along meshing teeth is ruled
this standard. by the stiffness of the contacting tooth pairs. Equations 共1兲 and 共2兲
The assembly precision also has a strong influence on gear express the displacement compatibility condition between mesh-
behavior. To reduce the effects of assembly deviation such as ing tooth pairs, and the torque equilibrium between the loaded
misalignment, flank crowning and tip relief are often recom- surfaces and the external torque, respectively.
mended 关2,3兴. The resulting consequences of such tooth modifi- ជi + ⌺
ជc + ⌺
ជe + ⌺
ជg = ⌺
ជa
cations are studied when applied on a misaligned gear set. The ⌺ kc kc kc kc 共1兲

冉冕 冊
investigation is conducted from the perspective of static load shar-
ing and its influence on surface contact and fillet bending stress.
Various papers dealing with assembly errors, profile modifica- 兺D
p共␰兲储rជ储共nជ · ជt 兲d␰ =T 共2兲
tions, and errors have been published, many of which are listed in Lj D
Equation 共1兲 is the contact condition for a point with coordinates
Contributed by the Power Transmission and Gearing Committee of ASME for
kc in the plane of contact—when Eq. 共1兲 is satisfied, point kc
publication in the JOURNAL OF MECHANICAL DESIGN. Manuscript received November 8, belongs to the loaded surfaces. This equation regroups five angu-
2004; final manuscript received May 26, 2005. Review conducted by Teik C. Lim. lar terms:

444 / Vol. 128, MARCH 2006 Copyright © 2006 by ASME Transactions of the ASME
 ជ a:
⌺ ⫽ the angular approach of the mating teeth;
ជ c:
⌺ ⫽ the contact deformation;
ជ e:
⌺ ⫽ the tooth bending and shearing deflection;
ជ i:
⌺ ⫽ the initial profile separation;
ជ
⌺ g: ⫽ the surface deformation due to frictional traction;

Fig. 1 Displacement-stress model of a helical gear tooth Fig. 2 Pressure cells in contact plane

Equation 共2兲 joins the individual torque produced by the load Over the segments the 2-D governing equations are solved with
distribution p共␰兲 along each of the D tooth pairs in mesh and the the finite difference method 共FDM兲. The main advantage of the
external torque 共T兲. Terms nជ , 储rជ储 and ជt are the unit normal to the FDM is the mesh where irregular grids are very easy to create.
contact plane, the radius of the meshing position, and a unit vector Since in the FDM, nodes are not attached to elements, there is no
in a plane perpendicular to the axis of rotation, respectively. need to impose common node sharing as with FEM meshing.
The simultaneous solution of the preceding equations leads to Thus, the mesh density can be refined rapidly in regions of stress
the load distribution along each meshing surface. Over the years, concentration.
numerous models have been presented 关2,3,6–9,11–21兴. Due to Contact analysis is a major concern in this kind of simulation.
the 3D integration implied and the nonlinear behavior of the con- The rolling-sliding load on meshing tooth surfaces is composed of
tact zones, most of the calculations were based on analytical rep- normal pressure and shear traction. An accurate and economic
resentations, often including important approximations 关2,6–12兴 way to handle such a problem is the numerical algorithm based on
or on different numerical approaches such as the finite element a discretization of the contact plane in pressure shear cells—
method 共FEM兲 关3,9,13–21兴. The express model was developed in influence functions for uniform pressure and shear tractions acting
order to eliminate the loss of precision associated with common on a rectangular zone are introduced in an iterative calculation
simplifications, while reducing the computer time usually attrib- scheme. The contact cells easily cope with any surface irregular-
uted to numerical simulation of contact gear problems. Equations ity, without increasing time consumption. The basis of the method
共1兲 and 共2兲 can be solved concurrently for up to five tooth pairs. is given in Hartnett’s work 关23兴. The approach has been applied to
The finite strip method 共FSM兲 introduced earlier in the gear the gear contact problem by a few authors 关15,17,18,24兴. Figure 2
domain 关22兴 establishes the tooth displacement from its neutral schematically illustrates the procedure. Directions ␰ and ␩ run
surface represented by a clamped 2-D strip. This approach con- along and across the contact line, respectively. This contact model
siderably reduces the node number while ensuring a precision in was also validated in a previous paper 关10兴 by comparison to
bending and shearing displacement within 5% compared to a 3-D Hertz’s prediction: for a spur gear set, the agreement on the maxi-
FEM analysis. mum contact pressure was higher than 95%.
The FSM clamps the tooth at the root radius. Therefore, to work Figure 3 graphically presents the express model.
out the complete tooth displacement, the strips are combined to a
tooth base discretization. This base model calculates the displace- 3 Helix Slope and Form Deviations: Effects of Toler-
ments of the tooth support as well as the fillet stresses. Figure 1
ance Margins for ANSI/AGMA ISO 1328-1 Grades 5
illustrates the resulting representation of a helical tooth used to
ជ e兲. and 7 on Contact and Fillet Stresses
calculate the complete tooth deflection 共⌺
Figure 1 shows that the tooth base is divided into a series of 3.1 Helix Slope Deviation „fh␤…. Real-life gear teeth must
segments of equal width. This strategy is comparable to the one comply with standardized tolerance or deviation margins that
employed in other approaches such as the models of Elkholy 关11兴 specify the grade to which the gears belong. The deviations are
or Smith 关9兴, and thus offers similar speed gain. Nevertheless, the normally measured in the direction of the transverse base tangent
loss of physical continuity usually associated with such a tech- and apply to an evaluation length that can be shorter than the tooth
nique is eliminated in the model—the junction between the base face width. In this study, the evaluation length is equal to the face
and the strips that are lengthwise continuous links the response of width.
the base segments. Since at any position along the tooth, there is a A 200 N · m torque is applied to the helical pinion of Table 1.
direct relation between load distribution and displacement, the The studied meshing position is shown schematically on the pin-
discrete load acting on each tooth base segment is calculated from ion member in Fig. 4, where two tooth pairs mesh simultaneously.
the FSM displacement at the junction line. Reference 关10兴 com- Table 2 gives the contact line endpoint coordinates. Figure 4 also
pares and validates the model with 3-D FEM analysis. presents the tooth meshing sequence: tooth pair 0 is the main

Journal of Mechanical Design MARCH 2006, Vol. 128 / 445

 Table 1 Helical gear set parameters

Helical gear
pinion/gear

Pressure angle 20°

Helix angle 15°
Module, normal 6.0 mm
Tooth number 20/31
Addendum factor 1
Dedendum factor 1.25
X factor 0.163/⫺0.163
Face width 70.0 mm
E; ␯ 200 GPa; 0.3

are assumed to vary proportionally. The simulation domain de-

fined for meshing tooth pairs can be represented by a two-factor
plane, i.e., the combined deviations of pair ⫺1 and pair 0. Figure
5 shows the simulation domain.
Using a two-level experimental strategy 关25兴, the simulation of
points I–IV covers all possible grade variations. However, points
II and IV represent the worst operating conditions that helix slope
deviations can produce and have been selected in order to simplify
the investigation. Points IV and II are referred to as case A and
case B, respectively. Even though these cases have a low prob-
ability of occurrence in a real gear set, they are possible and will

Fig. 4 Contact lines for the meshing position considered

Table 2 Coordinates of contact lines endpoints

共R , X3兲
共mm兲 Endpoint I Endpoint II

Pair ⫺1 共59.160; 16.800兲 共63.020; 70.000兲

Pair 0 共63.297; 0.000兲 共68.899; 47.600兲

Fig. 3 Flowchart of the express model Table 3 Displacement-stress and contact cell model
parameters

meshing tooth pair, while tooth pair ⫺1 follows. Value Pinion Wheel
The theoretical contact lines observed on the pinion teeth ex-
tend, for tooth pair ⫺1, from the fillet radius to a radius just above D⬘ 30.00 mm 30.00 mm
M 15.00 mm 15.00 mm
the pitch circle and, for tooth pair 0, from a radius above the pitch
l⬘ 14.09 mm 13.88 mm
circle to the tip circle 共Fig. 4兲. Table 3 presents the parameter
␤ 11.70 deg 15.11 deg
values introduced in the express model 关10兴. Finite 0.23 mm 0.23 mm
The measurement of helix slope deviation is obtained between differences
two helix traces that intersect the mean helix trace at the endpoints increment size
of the evaluation length 关1兴. Table 4 lists the tolerance values for No. of finite strips 14 14
grades 5 and 7 from their tooth numbers. 共5 nodes per
strip-radial
For a given grade, the combined deviation of a tooth pair is a direction兲
function of pinion and gear tooth accuracy. In order to limit the Contact plane 18 cells along, 11 cells across contact line
combined pinion and gear deviation range, individual deviations

446 / Vol. 128, MARCH 2006 Transactions of the ASME

Table 4 Helix slope deviation for grades 5 and 7 „ANSI/AGMA
ISO 1328-1…

fh␤共␮m兲 Pinion Gear

Grade 5 7.00 7.50

Grade 7 14.00 15.00

serve here for the purpose of analysis. Table 5 lists all the condi-
tions modeled with helix slope deviation. For comparison pur-
poses, the results of the unmodified theoretical gear set are also
presented.
The results obtained for the operating conditions of Table 5 and
the perfect gear pair are shown in Figs. 6–16.
The helix slope deviation of a tooth, if not compensated by an
inverse deviation in the mating tooth, results in a shift of the
contact zone along the theoretical contact line. In case A, the
contact zone on tooth pair ⫺1 shifts toward endpoint I, Fig. 1, and
toward point II for tooth pair 0. For case B, the contact zone
moves toward point II for tooth pair ⫺1 and toward point I for
tooth pair 0. In short, for both cases, the deviations will result in a
load reduction of the tooth pair where the deviation is negative,
and in a load increase for the tooth pair with a positive deviation.
The first series of results, Figs. 6–10, shows the pressure dis-
tribution in the contact plane of the active tooth pairs.
For the theoretical gear pair 共Fig. 6, no deviation兲, pressure Fig. 6 Pressure distributions on contact plane—no helix slope
peaks related to tip edge contact can be observed. For tooth pair deviation
⫺1, the pressure elevation is clearly visible; for pair 0, the phe-
nomenon is of lesser importance. The real contact pressure value
at tooth edge cannot be predicted exactly, since the model is based The assumption made in this study is that a similar convergence
on a contact cell discretization 关10兴 where the solution gives an could be reached using the cells in edge contact simulation. In
average pressure for each contact cell, and thus depends on cell fact, even if a convergence demonstration is not presented, it is
size and the position of the pressure center relative to the tooth assumed that the 80 ␮m width cells 共ten divisions along ␩ or the
edge. On the other hand, Bhushan 关26兴 suggested that, when the 0.8 mm width of meshed contact plane兲 employed here produce a
contact cells are small enough, the predicted pressure on contact- sufficient precision level in the flank limit vicinity. To the authors’
ing asperities of rough surfaces converge to the real amplitude. knowledge, the current literature does not offer experimental
works dealing with such a particular contact type that are suitable
to validate this constant pressure cells approach.
The effects of helix slope deviation on contact pressure are
observed in Figs. 7–10. For the studied meshing position, case A

Fig. 5 Simulation domain and studied points

Table 5 Helix slope deviation, studied combinations

Pair ⫺1 Pair 0

Case A Grade 5 − ⫹
Grade 7 − ⫹
Case B Grade 5 ⫹ −
Grade 7 ⫹ − Fig. 7 Pressure distributions on contact plane—case A, grade
5

Journal of Mechanical Design MARCH 2006, Vol. 128 / 447

Fig. 8 Pressure distributions on contact plane—case A,
grade 7
Fig. 10 Pressure distributions on contact plane—case B,
grade 7
is more damaging since it increases edge contact pressure on tooth
pair 0. The increase appears to be proportional to slope error: in
case A, the maximum pressure increase is 1162 MPa for grade 5
and 2017 MPa for grade 7 共Figs. 7 and 8兲, while the deviation of
grade 7 is twice that of grade 5.
Figure 11 displays the longitudinal load distribution obtained
from the contact pressure integration along ␩. This averts the
uncertainty on the contact pressure near the tooth’s edges. The
relationship between maximum load distribution elevation and
gear grade is clearly shown.
Table 6 displays the peak values of the integrated contact pres-
sure distribution of Fig. 11. In case A, the load transfer is from
tooth pair ⫺1 to pair 0, whereas in case B it is from pair 0 to pair

Fig. 11 Load distributions on contact lines, theoretical teeth,

Fig. 9 Pressure distributions on contact plane—case B, grades 5 and 7 with positive and negative helix slope
grade 5 deviations

448 / Vol. 128, MARCH 2006 Transactions of the ASME

 Fig. 14 Maximum principal pinion tooth fillet stress
Fig. 12 Maximum principal pinion tooth fillet stress distribu- distributions—case A, grade 7
tions, no deviation

entire contact line. On the other hand, when the first contact point
⫺1. is at the tip radius, such as tooth pair 0 共Case A兲, the perturbation
The analysis of Table 6 and Fig. 11 reveals the importance of is localized. In others words, the effect of deviations depends on
the load increment and the influence of the helix slope deviation, the nature of the contacting surfaces.
depending on the position of the contact line. With a first contact As Table 6 clearly shows, load elevation for grade 7 is almost
point moved to a position close to the pitch radius, for tooth pair twice that of grade 5, and the deviations double from grade 5 to 7.
⫺1 共case A兲, the helix slope deviation influence spreads over the The results in Figs. 12–16 present the maximum principal tooth

Fig. 13 Maximum principal pinion fillet stress distributions— Fig. 15 Maximum principal pinion tooth fillet stress
case A, grade 5 distributions—case B, grade 5

Journal of Mechanical Design MARCH 2006, Vol. 128 / 449

 Fig. 17 Helix form deviation

3.2 Helix Form Deviation „ff␤…. Helix form deviation mea-

surement is made between two parallels to the mean helix trace.
The two parallels are separated by a given distance, so as to en-
close the trace of the real profile over the evaluation length. Tol-
erance values admitted with grades 5 and 7 are the same as those
for helix slope deviation, as presented in Table 4.
In this paper, the studied helix form deviation is given by a sine
wave presenting three protuberances over the face width, such that
the pinion and gear deviations are in phase 共Fig. 17兲.
It is assumed that the effects of helix form deviation are limited
to the deviated tooth pair itself. Therefore, the pinion and gear of
tooth pair 0 will be erroneous and pair ⫺1 will maintain its true
form. If we refer to Fig. 5 with these new factors, points I and III
become the worst gear set conditions. However, since point III
represents an inversion of the deviation shown in Fig. 17, only
point I will be considered. Table 8 lists the considered helix form
deviations.
Fig. 16 Maximum principal pinion fillet stress distributions— Figures 18 and 19 show the contact pressure graphs, while Fig.
case B, grade 7 20 presents the resulting load distribution along the contact lines.
Figure 20 also includes, from Fig. 11, the load distribution ob-
tained with the theoretical, unmodified, gear set.
fillet stress on the tensile side of the pinion member 共see Fig. 2兲. Pressure and load distribution diagrams for tooth pair 0, Figs.
The stress charts display values calculated in the tooth fillet but do 18–20, clearly show the practical effect of helix form deviations
not include the effects of contact proximity to the fillet area.
Load fluctuations on the tooth flank influence tooth fillet stress.
The fillet stress variation is in line with the maximum or minimum Table 8 Helix form deviations „tooth pair 0…
value of the load distribution, as shown in Figs. 12–16. Even if the Deviation
contact pressure attained very high values in Figs. 6–10, the cor- 共␮m兲 Pinion Gear
responding bending stress stays under tolerable amplitudes.
Table 7 presents peak values of the principal fillet tensile stress Grade 5 7.0 7.5
in relation to the corresponding theoretical value 共i.e., without Grade 7 14.0 15.0
helix slope deviation, same position as the maximum stress point
on deviated teeth兲. As expected, the results of Table 7 indicate that
an increase in tooth load 共Table 5兲 has an almost direct correspon-
dence with fillet stress increase. The stress increase also appears
proportional to the error amplitude of the tooth pair.

Table 6 Maximum load distribution increase

Theoretical Grade 7
value Grade 5 Grade 7
共N/mm兲 共N/mm兲 共N/mm兲 Grade 5

Case A 409 ⫹344 ⫹638 1.85

pair 0 共⫹84.1%兲 共⫹156.0%兲
Case B 361 ⫹262 ⫹517 1.97
pair ⫺1 共⫹72.6%兲 共⫹143.2%兲

Table 7 Increase in maximum principal stress compared to

the theoretical gear set

Theoretical Grade 7
value Grade 5 Grade 7
共MPa兲 共MPa兲 共MPa兲 Grade 5

Case A 120 ⫹48 ⫹101 2.10

pair 0 共⫹40.0%兲 共⫹84.2%兲
Case B 84 ⫹50 ⫹102 2.04
pair ⫺1 共⫹59.5%兲 共⫹121.4%兲
Fig. 18 Pressure distributions in the contact plane—grade 5

450 / Vol. 128, MARCH 2006 Transactions of the ASME

 Table 9 Increase in the maximum load distribution value in
relation to the unmodified gear set

Theoretical Grade 7
value Grade 5 Grade 7
共N/mm兲 共N/mm兲 共N/mm兲 Grade 5

Mean 346 594 831 ...

value
Increase ... 248 485 1.96
共%兲 关71.7%兴 共140.2%兲

clearly visible. In addition, in Figs. 19共b兲 and 20共b兲 共grade 7兲, the
elastic deformation fails to produce contact in some areas of the
tooth where load distribution falls to zero.
In the case of grade 5, the contact areas where the contact
pressure is reduced sustain an average load of 128 N / mm, equiva-
lent to 21.5% of the mean peak value of 594 N / mm and 37.0% of
the theoretical 共unmodified teeth兲 average value. Moreover, from
Table 9, one can see that the maximum load value nearly doubles
when the grade changes from 5 to 7, where the deviations are
doubled.
Helix form deviations also have a visible effect on fillet stress,
as shown in Figs. 21 and 22 where the maximum fillet stress
Fig. 19 Pressure distributions in the contact plane—grade 7 follows the evolution of helix form deviations. However, the high-
est values remain under acceptable levels, while the intensity
on tooth pair 0, while tooth pair ⫺1 remains unaffected. reached by the contact pressures 共Figs. 18 and 19兲 was significant.
Table 9 lists, for tooth pair 0, the maximum load increase for Table 10 lists the increases in maximum fillet stress. These
the modified teeth; the mean value is calculated along the contact results are averaged values calculated at the peak location on the
line of the unmodified tooth. deviated tooth flanks and at the corresponding position on the
The importance of load increase under helix form deviation is theoretical unmodified pinion tooth 关Fig. 11共b兲兴. The values in
Table 10 show a much weaker influence of the helix form devia-
tions on fillet stress than on contact pressure, since fillet stress
distribution is a response to flank loading, which depends on the
shape, amplitude, and position of the helix form deviation. Con-
sequently, helix form errors will affect contact surface pressure

Fig. 20 Load distributions on simultaneous contact lines „un- Fig. 21 Maximum principal pinion fillet stress distributions—
modified teeth and grades 5 and 7 with undulated profiles… grade 5, deviated tooth

Journal of Mechanical Design MARCH 2006, Vol. 128 / 451

 Fig. 23 Crowning modification

关2兴, but no consequence on tooth pair 0. A tip relief modification,

conversely, would have a result on tooth pair 0 关11兴, but not on
pair ⫺1.
In addition to the original gear set, the following three modified
sets are studied:
1. Crowned tooth flank
2. Tip-relieved tooth
3. Crowned and tip-relieved tooth

The crowning modification applied to both the pinion and gear

teeth is defined in Fig. 23. The tooth ends are corrected using a
sine shape.
Tip relief correction is defined on the tooth of the rack cutter
using a quadratic form. Figure 24 illustrates the modified profile
of the tooth used to generate the pinion and the gear. Table 11
Fig. 22 Maximum principal pinion fillet stress distributions— gives the parameter values.
grade 7, deviated tooth The results in Figs. 25–28 show the pressure distribution in the
contact zone for the four misaligned gear sets. Overloading at the
ends of the contact lines is obvious.
more than bending stress, and thus gear life. Fillet stress increases From the results shown in Figs. 25–28, it is clear that misalign-
between grades 5 and 7 also reveal a linear correspondence with ment on a helical gear overloads the ends of the potential lines of
the amplitude of the deviation. contact rather than the tooth ends.
Consequently, during mesh, overloading will follow a path
4 Flank Crowning and Tip Relief: Effects on a Mis- along the tooth edges. Figure 29 illustrates the progression of the
aligned Gear Set contact lines on the pinion tooth. For a positive misalignment
共current case兲, overloading starts at point D and reaches point F
The misalignment of a gear may be of limited consequence for through point E. It goes without saying that parameters such as
the position of the potential lines of contact. However, misalign-
ment will significantly affect load distribution along the lines of
contact. As for helix slope deviations, an alignment error shifts the
first contact point towards an endpoint of the potential line of
contact 关10兴. The gear described in Table 1 and in the contact
position of Table 2 is rotated by 0.05 deg 共about the X1 axis兲 from
its perfect assembly position. Since the location of the lines of
contact is unmodified from the perfect assembly, the coordinates
of Table 2 remain valid. Thus, the first contact points reach end-
points II of the contact lines. Consequently, the loaded zones are
initially localized at the pinion tooth end for pair ⫺1 and at the
pinion tip edge for pair 0 共Fig. 4兲, and load distribution in the
contact zones depends on tooth deformations.
To improve the adjustability of an assembly and reduce mis-
alignment effects, the tooth flanks are crowned. Such a modifica-
tion of the flank would have a beneficial effect on tooth pair -1

Table 10 Increase in maximum principal stress compared to

the theoretical gear set Fig. 24 Modified rack cutter profile

Theoretical Grade 7 Table 11 Modified rack cutter parameter values

value Grade 5 Grade 7
共MPa兲 共MPa兲 共MPa兲 Grade 5 Parameter Value

Mean 120 132 144 ... Module, normal 共mn兲 6.0 mm

value Addendum factor 共a兲 1
Increase ... ⫹12 ⫹24 2.00 Dedendum factor 共b兲 1.25
共%兲 共⫹10.0%兲 共⫹20.0%兲 Tip relief factor 共d兲 0.2
Tip relief depth 共A⬘兲 0.036 mm

452 / Vol. 128, MARCH 2006 Transactions of the ASME

Fig. 25 Pressure distributions on the contact plane—
unmodified, misaligned gear set Fig. 27 Pressure distributions on the contact plane—tip re-
lieved, misaligned gear set
the helix angle, the face width and the contact ratio will control
the exact position of the critical point as well as the local load
overloading are directly related to meshing position. Thus, crown-
amplitude.
ing and tip relief corrections should always be used together. For
Figure 30 presents the load distribution on each contact line,
instance, the crowned and tip relieved gear set shows load reduc-
while Table 12 compares load change in reference to the original,
tions of about 17% and 34% on the contact lines of tooth pairs ⫺1
unmodified, misaligned gear set.
and 0, respectively.
Analysis of Fig. 30 and Table 12 indicates that for the studied
Figures 31–34 show the maximum principal stress distribution
contact position, flank crowning correction reduces contact stress
calculated on the tensile side of the pinion teeth, as affected by
at the tooth ends and increases overloading at the tip edge in the
crowning and tip relief.
center of the tooth.
Table 13 compares the resulting peak stress values to those of
In contrast, tip relief has a reducing effect in the middle portion
the original, unmodified, misaligned gear set.
and overloads tooth ends. Of course, increase and reduction in

Fig. 26 Pressure distributions on the contact plane—crowned, Fig. 28 Pressure distributions on the contact plane—crowned
misaligned gear set and tip relieved, misaligned gear set

Journal of Mechanical Design MARCH 2006, Vol. 128 / 453

 Fig. 29 Path of overloading along tooth edges

The stress diagrams show a trend in bending stress that is

analogous to that of the contact pressure. However, a partial cor-
rection, i.e., either crowning or tip relief, seems to be more dam-
aging to fillet stress than to contact pressure.

Fig. 31 Maximum principal pinion fillet stress distributions—

original, misaligned gear set

For example, flank crowning increases fillet stress of tooth pair

0 by about 14%, whereas contact load increase is less than 5%.
For tooth pair ⫺1, tip relief increases fillet stress by 31%, while
contact load is increased by about 28%.
On the other hand, the useful effects obtained with the complete
correction are less significant on bending stress than on contact

Fig. 30 Load distributions on simultaneous contact lines of

misaligned sets

Table 12 Maximum load distribution variation in relation to the

original set

Crowned
Theoretical Tip and tip
value Crowned relieved relieved
共N/mm兲 共N/mm兲 共N/mm兲 共N/mm兲

Pair -1 858 −233 239 −143

共−27.2% 兲 共27.9%兲 共−16.7% 兲
Pair 0 912 41 −357 −309
共4.5%兲 共−39.1% 兲 共−33.9% 兲 Fig. 32 Maximum principal pinion fillet stress distributions—
crowned, misaligned gear set

454 / Vol. 128, MARCH 2006 Transactions of the ASME

 Table 13 Increase in maximum principal stress compared to
the original gear set

Crowned
Theoretical Tip and tip
Value Crowned relived relived
共MPa兲 共MPa兲 共MPa兲 共MPa兲

Pair ⫺1 184 −38 57 −17

共−20.7% 兲 共31.0%兲 共-9.2%兲
Pair 0 154 21 −27 −14
共13.6%兲 共-17.5%兲 共-9.1%兲

Even though the comparative numbers calculated here are spe-

cific to the studied cases, the highlighted phenomena remain gen-
eral. Consequently the observations clearly show the need to prop-
erly combine the two types of tooth modification.

5 Conclusion
The first part of this paper illustrates the influence of helix slope
and form deviations. Longitudinal flank crowning and radial tip
relief modifications were studied.
Helix slope deviation and misalignment of a gear set have com-
parable consequences: the first contact point is driven to an end-
point of the theoretical line of contact, which results in localized
overloading. As a consequence, contact pressure is increased lo-
cally, which generates a corresponding increase in bending stress
Fig. 33 Maximum principal pinion fillet stress distributions— in the fillet zone.
tip relieved, misaligned gear set Flank crowning and tip relief modifications are then introduced;
results show that they should always be considered in the design
process and should be used in concert, since partial correction
pressure. For example, the fillet stress reduction obtained is on the may cause load increase rather than decrease: when used sepa-
order of 9%, which is an improvement. It is also worth emphasiz- rately on a misaligned cylindrical gear set, flank crowning, while
ing that the stress levels always stayed within a normal range, reducing contact stress at tooth ends, will amplify the overload at
whereas the contact pressure amplitudes attained dangerous inten- tip edge. Conversely, tip relief will overload tooth ends and dis-
sities. charge the tip edge.
While helix slope deviations and misalignment errors have
similar consequences, helix slope deviations are potentially more
dangerous since opposite errors on meshing teeth may occur. This
condition results in load transfer from one tooth pair to another.
For example, for a grade 5 gear pair with helix slope deviation,
our results show that load increase is more than 70% with fillet
stress increases of more than 40%. A comparative analysis of
grades 5 and 7 reveals values virtually proportional to the ampli-
tude of the error.
The helix form deviation, modeled as a sine wave superim-
posed on a theoretical tooth surface, presents the same direct re-
lationship between the amplitude of the error and the load state of
the teeth.
Therefore, it can be concluded that reducing or increasing the
quality class of a gear pair will proportionally influence the gear
performance as well as its life.
The analysis presented here also indicates, for all cases, that
contact perturbations led to very high or even excessive contact
pressure, while the corresponding fillet stress remained moderate.
Comparable observations have also been made by other investi-
gators 关5兴. Thus, we can assume that any contact disturbance will
primarily damage the contacting surfaces and might have no sig-
nificant effect on the bending life.

Nomenclature
D ⫽ number of tooth pairs in mesh
D⬘ ⫽ width of flexible support under the tooth
Lj ⫽ length of contact line j
M ⫽ height of flexible support under the tooth
T ⫽ applied torque
Fig. 34 Maximum principal pinion fillet stress distributions— X ⫽ addendum modification coefficient
crowned and tip-relieved, misaligned gear set l⬘ ⫽ height of the tooth base model

Journal of Mechanical Design MARCH 2006, Vol. 128 / 455

 nជ ⫽ unit normal to the contact plane in the pinion Modification of Helical and Double Helical Gears,” Eur. J. Mech. Eng., 36共3兲,
pp. 171–176.
reference frame 关9兴 Smith, J. D., 1995, “Estimation of the Static Load Distribution Factor for
p共␰兲a ⫽ load distribution on a contact line Helical Gears,” Proc. Inst. Mech. Eng., Part C: J. Mech. Eng. Sci., 209, pp.
储rជ储 ⫽ pinion radius of contact 193–199.
关10兴 Guilbault, R., Gosselin, C., and Cloutier, L., 2005, “An Express Model for
ជt ⫽ unit vector in a plane tangent to the pinion Helical Gear Load Sharing and Stress Analysis,” ASME J. Mech. Des.,
axis of rotation 127共6兲, pp, 1161–1172.
␤ ⫽ inclination angle of tooth base model 关11兴 Elkholy, A. H., 1996, “Load and Stress Variation Along Helical Gear Teeth,”
␰ ⫽ position along contact line 关12兴
Trans. CSME, 20共2兲, pp. 159–174.
Börner, J., 1996, “Very Efficient Calculation of the Load Distribution on Ex-
⌺ជ a ⫽ rotation of pinion and gear members ternal Gear Sets – The Method and Applications of the Program LVR,” Proc.
⌺ជ c ⫽ rotation caused by contact deformation of the 7th International Power Transmission and Gearing Conference, ASME,
San Diego, Vol. 88, pp. 219–225.
⌺ជ e ⫽ rotation caused by tooth bending and shearing 关13兴 Li, J., Zhun, Z., Lin, J., and Shouyou, W., 1998, “Finite Element Analysis of
Cylindrical Gears,” Commun. Numer. Methods Eng., 14, pp. 963–975.
共includes tooth base兲 关14兴 Sfakiotakis, V. G., Vaitsis, J. P., and Anifatis, N. K., 2001, “Numerical Simu-
ជ i ⫽ rotation caused by initial tooth profile
⌺ lation of Conjugate Spur Gear Action,” Comput. Struct., 79, pp. 1153–1160.
separation 关15兴 Parker, R. G., Vijayakar, S. M., and Imajo, T., 2000, “Non-linear Dynamic
Response of a Spur Gear Pair: Modeling and Experimental Comparisons,” J.
⌺ជ g ⫽ rotation caused by surface deformation due to Sound Vib., 237共3兲, pp. 435–455.
frictional traction 关16兴 Choi, M., and David, J. W., 1990, “Mesh Stiffness and Transmission Error of
␩ ⫽ position perpendicular to contact line 关17兴
Spur and Helical Gears,” Trans. CSME, 20共2兲, pp. 1599–1607.
Kim, H. C., de Vaujany, J. P., Guingand, M., Bard, C., and Play, D., 1995,
␴ ⫽ normal stress “Stresses of External and Internal Cylindrical Gears. Effects of Rim, Web and
␶ ⫽ shear stress Mechanical Constraint Conditions,” 9th World Congress on the Theory of Ma-
chines and Mechanisms, Milano, Vol. 1, pp. 565–569.
关18兴 Guingand, M., de Vaujany, J. P., and Icard, Y., 2004, “Fast Three-Dimensional
References Quasi-Static Analysis of Helical Gears Using the Finite Prism Method,”
关1兴 American Gear Manufacturers Association, 1999, ANSI/AGMA ISO 1328-1 ASME J. Mech. Des., 126, pp. 1082–1088.
Cylindrical Gears – ISO System of accuracy – Part1: Definitions and Allow- 关19兴 Olakorédé, A. A., and Play, D., 1991, “Load Sharing, Load Distribution and
able Values of Deviations Relevant to Corresponding Flanks of Gear Teeth, Stress Analysis of Cylindrical Gears by Finite Prism Method”, Design Produc-
American Gear Manufacturers Association, Alexandria. tivity International Conference, Honolulu, pp. 921–927.
关2兴 Seol, I. H., and Kim, D. H., 1998, “The Kinematics and Dynamic Analysis of 关20兴 Simon, V., 1988, “Load and Stress Distributions in Spur and Helical Gears,”
Crowned Spur Gear Drive,” Comput. Methods Appl. Mech. Eng., 167, pp. ASME J. Mech., Transm., Autom. Des., 110, pp. 197–202.
109–118. 关21兴 Sundarajan, S., and Young, B. G., 1990, “Finite-Element Analysis of Large
关3兴 Zhang, J. J., Esat, I. I., and Shi, Y. H., 1999, “Load Analysis with Varying Spur and Helical Gear Systems,” J. Propul. Power, 6共4兲, pp. 451–454.
Mesh Stiffness,” Comput. Struct., 70, pp. 273–280. 关22兴 Gagnon, P., Gosselin, C., and Cloutier, L., 1996, “Analysis of Spur, Helical
关4兴 Umeyana, M., Kato, M., and Inoue, K., 1998, “Effects of Gear Dimensions and Straight Level Gear Teeth Deflection by the Finite Strip Method,” ASME
and Tooth Surface Modifications on Loaded Transmission Error of Helical J. Mech. Des., 119共4兲, pp. 421–426.
Gear Pair,” ASME J. Mech. Des., 120, pp. 119–125. 关23兴 Hartnett, M. J., 1980, “A General Numerical Solution for Elastic Body Contact
关5兴 Wagaj, P., and Kahraman, A., 2002, “Influence of Tooth Profile Modification Problems,” Solid Contact Lubr. AMD ASME, 39, pp. 51–66.
on Helical Gear Durability,” ASME J. Mech. Des., 124, pp. 501–510. 关24兴 Vijayakar, S. M., 1996, “Edge Effects in Gear Tooth Contact,” Proc. of the 7th
关6兴 Poritsky, H., Sutton, D., and Pernick, A., 1945, “Distribution of Tooth Load International Power Transmission and Gearing Conference, ASME, San Di-
Along a Pinion,” ASME J. Appl. Mech., 12共2兲, pp. A78–A86. ego, Vol. 88, pp. 205–212.
关7兴 Kubo, A., 1978, “Stress Condition, Vibrational Exciting Force, and Contact 关25兴 Goupy, J., 1988, La méthode des plans d’expériences: optimisation du choix
Pattern of Helical Gears with Manufacturing and Alignment Error,” ASME J. des essais et de l’interprétation des résultats, Dunod, Paris.
Mech. Des., 100, pp. 77–84. 关26兴 Bhushsn, B., 1999, Principles and Applications of Tribology, John Wiley and
关8兴 Winter, H., and Placzek, T., 1991 “Load Distribution and Topological Flank Sons, New York.

456 / Vol. 128, MARCH 2006 Transactions of the ASME

View publication stats