Professional Documents
Culture Documents
DISSERTATION
By
*****
Dissertation Committee:
Approved by
Professor Ahmet Kahraman, Advisor
Mohsen Kolivand
2009
ABSTRACT
milled and face-hobbed hypoid gears produced by Formate and generate processes.
Tooth surfaces are defined directly from the cutter parameters and machine settings. A
novel methodology based on the ease-off topography is used to determine the unloaded
contact patterns. The proposed ease-off methodology finds the instantaneous contact
curves through a surface of roll angles, allowing an accurate unloaded tooth contact
analysis in a robust and accurate manner. Rayleigh-Ritz based shell models of teeth of
the gear and pinion are developed to define the tooth compliances due to bending and
deformation effects are also included in the compliance formulations. With this, loaded
contact patterns and transmission error of both face-milled and face-hobbed spiral bevel
and hypoid gears are computed by enforcing the compatibility and equilibrium conditions
of the gear mesh. The proposed model requires significantly less computational effort
than finite elements (FE) based models, making its use possible for extensive parameter
hypoid gear contact model are also provided to demonstrate the accuracy of the model
ii
The proposed ease-off formulation is generalized next to include various types of
tooth surface deviations in the tooth contact analysis. These deviations are grouped in
two categories. The proposed ease-off based method is shown to be capable of modeling
both global deviations due to common manufacturing errors and heat treat distortions and
The proposed loaded contact model is combined at the end with a friction model
(mechanical) power losses and efficiency of the hypoid gear pairs. The velocity, radius
of curvature and load information predicted by the contact model is input to the friction
model to determine the distribution of the friction coefficient along the contact surfaces.
At the end, the variations of predicted mechanical efficiency with geometry, surface and
iii
Dedicated to my mother
iv
ACKNOWLEDGMENTS
Kahraman, for this great research opportunity, his guidance throughout my research and
his effort in reviewing this dissertation. I would also like to express my appreciation to
Prof. Donald R. Houser, Prof. Gary L. Kinzel and Prof. Henry H. Busby for their patience
and effort in being a part of my dissertation committee. Also, thanks to Dr. Sandeep
I would like to thank the sponsors of the Gear Power Transmission Research
My sincere thanks go to Prof. Hermann J. Stadtfeld from The Gleason Works who
spent valuable time teaching me fundamental concepts of bevel gear design and
I would also like to thank Jonny Harianto, Samuel Shon and all my lab mates for
Finally, I deeply appreciate the love and trust shown toward me by my parents,
my grandparents, my uncles and aunts, my sisters and brother and my fiancée for all of
v
VITA
Nov. 1999 –Apr. 2002 ……………. Design Engineer, Tarh Negasht Co., Tehran,
Iran
Apr. 2002 –Aug. 2005 ……………. Design Engineer, TAM Co., Tehran, Iran
PUBLICATIONS
1. Kolivand M. and Kahraman A., “A Load Distribution Model for Hypoid Gears Using
Ease-off Topography and Shell Theory,” Journal of Mechanism and Machine Theory,
2009.
FIELDS OF STUDY
vi
TABLE OF CONTENTS
Page
Abstract………………………………………………………………………………. ii
Dedication……………………………………………………………………………. iv
Acknowledgments………………………………………………………………….... v
Vita……………………………………………………………………………….….. vi
List of Tables..………………………………………………………………...…….. x
List of Figures...……………………………………………………………...……… xi
Nomenclature………………………………………………………………………... xvi
Chapters:
1 Introduction ……………………………………………………………………... 1
vii
2.2.1 Kinematics ………………………………………………………...… 32
2.2.2 Cutting Tool Geometry and the Relative Motion……………………. 36
2.2.3 Equation of Meshing…………………………………………………. 40
2.2.4 Principal Curvatures and Principal Directions…………………......... 42
2.3 Unloaded Tooth Contact Analysis ………………...………………………… 44
2.3.1 The Conventional Method of UTCA.……………….……………….. 47
2.3.2 Ease-off Based Method of UTCA …………………………………… 52
2.3.2.1 Construction of Ease-off and the Surface of Roll Angle………… 53
2.3.2.2 Contact Pattern and Transmission Error…………………………. 57
2.4 An Example Hypoid Unloaded Tooth Contact Analysis ……………………. 59
References for Chapter 2………………………………………………………. 65
3 Shell Based Hypoid Tooth Compliance Model and Loaded Tooth Contact
Analysis …............................................................................................................. 68
4 Loaded Tooth Contact Analysis of Hypoid Gears with Local and Global Surface
Deviations…………………………………..….………………………………… 94
ix
LIST OF TABLES
Table Page
2.1 Basic drive side geometry and working parameters of the example hypoid
gear pair…………………………………………………………………..….. 60
4.1 Basic drive side geometry and working parameters of the example hypoid
gear pair.…………………………………………………….……………….. 112
4.2 The transmission error amplitudes of theoretical and deviated surfaces…….. 118
4.3 Basic drive side geometry and working parameters of the example hypoid
gear pair.……………………………………………………………………... 121
4.4 The transmission error amplitudes of theoretical and deviated surfaces…..… 128
5.1 Parametric design for the development of the friction coefficient formula….. 147
5.2 Basic parameters of the 75W90 gear oil used in this study.…………….….... 148
5.4 Basic drive side geometry and working parameters of the examples hypoid
gear pairs……………………………………………………………………... 153
x
LIST OF FIGURES
Figure Page
1.2 A sample hypoid gear pair with a shaft angle and a shaft off-set d a .….... 4
1.4 Flowchart of overall hypoid gear loaded tooth contact analysis methodology 18
2.3 (a) Cutter head, (b) blade and (c) cutting edge geometry……………………. 37
2.8 Unloaded TCA computation procedure: (a) gear projection plane, ease-off
and Q surfaces, and (b) instantaneous contact curve, contact line and
unloaded transmission error…………………………………………………. 58
2.9 Unloaded transmission error of the example gear pair with misalignments
E 0.15 mm, P 0.12 mm, G 0, 0 ………………………….. 61
xi
2.10 Unloaded contact pattern of the example gear pair for three adjacent tooth
pairs i 1 , i and i 1 (i-1), (i) and (i+1) with
E 0.15 mm, P 0.12 mm, G 0 and 0 ……………………… 63
2.11 Unloaded contact pattern of the example gear pair (a) at nominal position
with E P G 0 , (b) at toe with E 0.08 mm, P 0.10
mm and G 0 , (c) at heel with E 0.15 mm, P 0.10 mm
and G 0 and (d) at toe with E 0.05 mm, P G 0 , and
4 min…………………………………………………………………. 64
3.5 Static equilibrium between torque applied on gear axis and torque produced
by the force of all contacting segments.………...………………………….... 85
3.6 Loaded transmission error of the example gear pair with E 0.15,
P 0.12, G 0 and 0 at (a) Tp 50 Nm, (b) Tp 250 Nm, and (c)
Tp 500 Nm.…………………………………………………………............ 87
xii
4.2 Graphical demonstration of the procedure to update ease-off surface for
surface deviations.………………………...…………………………………. 105
4.5 Example local deviation surfaces for the gear and pinion tooth surfaces..….. 113
4.6 Ease-off update for the example deviation of Fig. 5. (a) Three-dimensional
view of the projection plane, and , , Q and Q surfaces, and contour
plots of (b) , (c) , and (d) the change of ease-off topography..………… 114
4.7 Predicted unloaded tooth contact pattern for separation value of 6 μm … 116
4.8 Transmission error (UTE) curves for theoretical and deviated surfaces at (a)
unloaded conditions and (b) loaded conditions at a pinion torque of 200 Nm. 117
4.9 Predicted contact pressure distribution for a pinion toque of 200 Nm for (a)
theoretical and (b) deviated surfaces………………………………………… 119
4.10 Example global deviation surfaces measured by CMM for the gear and
pinion tooth surfaces, (a) pinion measured deviation, (b) gear measured
deviation, (c) pinion deviation distribution in tooth active region and (d)
gear deviation distribution in tooth active region……………………………. 122
4.11 Ease-off update for the example deviation of Fig. 4.10. (a) Theoretical ease-
off topography, (b) updated ease-off topography only with pinion deviation,
(c) updated ease-off topography only with gear deviation, and (d) updated
ease-off topography with both pinion and gear deviations…………………... 124
4.12 Predicted unloaded tooth contact pattern for separation value of 6 μm … 125
xiii
4.13 Transmission error curves for theoretical and deviated surfaces; (a)
unloaded conditions and (b) loaded conditions at a pinion torque of 200 Nm 127
4.14 Predicted contact pressure distribution for a pinion toque of 200 Nm for (a)
theoretical and (b) deviated surfaces………………………………………… 129
5.2 Sliding and rolling velocities and their projection in tangential plane along
and normal to the contact line direction.……………………………………... 141
5.3 Ease-off topography of (a) Design A with d a / Da 0.07 and (b) Design B
with d a / Da 0.14 …………………………………………………………... 154
5.5 Rolling velocity distribution of (a) Design A with d a / Da 0.07 and (b)
Design B with d a / Da 0.14 at p 1500 rpm …………………………… 157
5.6 Sliding velocity distribution of (a) Design A with d a / Da 0.07 and (b)
Design B with d a / Da 0.14 at p 1500 rpm …………………………… 158
5.7 Slide-to-roll ratio distribution of (a) Design A with d a / Da 0.07 and (b)
Design B with d a / Da 0.14 at p 1500 rpm …………………………... 159
5.9 distribution of (a) Design A with d a / Da 0.07 and (b) Design B with
d a / Da 0.14 at p 1500 rpm , T p 500 Nm , Toil 90 C
and S1 S 2 0.8 m ……………………………………………………….... 161
xiv
5.10 Friction coefficient distribution of (a) Design A with d a / Da 0.07 and
(b) Design B with d a / Da 0.14 at p 1500 rpm , T p 500 Nm ,
Toil 90 C and S1 S2 0.8 m …………………………………………... 162
5.11 Power loss and efficiency of Design A (a1, b1) and Design B (a2,b2) at
Toil 90 C and S1 S2 0.8 m ……………………………………….…. 163
5.12 Efficiency of (a) Design A with d a / Da 0.07 and (b) design B with
d a / Da 0.14 for different surface finish and oil temperatures at
p
1500 rpm and T p 500 Nm …………………………..……………… 166
xv
NOMENCLATURE
e1 , e2 Principal directions
E Overall efficiency
Eb Blank offset
iT Tilt angle
js Swivel angle
k1 , k2 Normal curvatures
K1 , K 2 Principal curvatures
Segment index
mg Number of surface grid in lengthwise direction
xvi
M ctb Machine center to back
n Normal to the family of cutter surface
ncl Number of potential contact lines at each time step
Ease-off surface
s Distance of an arbitrary point to reference point on the blade edge
Sr Radial setting
xvii
SE Strain energy
SR Sliding to rolling velocity ratio
Seq Equivalent surface roughness
xviii
Y Vector of slack variable
Variable of curvilinear cylindrical coordinate system (in tooth lengthwise
direction)
b Blade angle
mn Shear strain
Effective viscosity
0 Ambient viscosity
xix
c Cradle angle
p Pinion speed
Tangential plane
Superscript:
( ) Real or updated
( ) Theoretical
(ˆ) Conjugate
( ) Interpolated / Extrapolated
p Pinion
g Gear
a Action
xx
CHAPTER 1
INTRODUCTION
Hypoid gears are widely used in many power trains to transfer power between
two non-intersecting crossed axes. Their most common and highest-volume applications
can be found in front and rear axles of rear-wheel-drive or all-wheel-drive vehicles [1.2].
Figure 1.1 shows a sample of hypoid gear application for the rear axle. A rear axle has
three primary functions: (i) transmit power from the drive train axis to the wheel axle,
that is usually perpendicular to the drive train axis with an offset, (ii) provide the
capability to the vehicle to turn corners without any slippage at its wheels through its
differential, and (ii) provide the final stage of speed reduction (torque increase) that is
1
Hypoid
gear
Input
Output Hypoid
pinion
Figure 1.1: A cut-away of an ‘auxiliary’ axle (Rear Drive Module) used in midsize
passenger cars and SUV’s (Courtesy of American Axle & Manufacturing Inc.).
2
A pair of hypoid gears is commonly used to deliver this third final drive function.
In the arrangement shown in Figure 1.1, the smaller of the hypoid gears, called the
pinion, is at the end of the drive shaft and is in mesh with the larger hypoid gear (called
the gear).
Hypoid gears can be considered as one of the most general cases of gearing based
on their geometry, such that other gear types can be obtained from it by assigning certain
values to some of the geometric parameters [1.1,1.3-1.5]. The main function of the
hypoid gear pair in a rear axle is to transmit power between two axes that are at a shaft
angle (usually 90 ) [1.2,1.6] and at a certain amount of shaft off-set d a as shown in
Figure 1.2. A higher level of power transmission through such a kinematic configuration
is possible through use of a hypoid gear pair, which can provide a better balance amongst
all primary design requirements such as strength, noise and power density. The trade-off
arrangements. The shaft offset, being the main difference between spiral bevel and
hypoid gears, provides several advantages to hypoid gears including larger pinion size,
smaller pinion tooth counts, higher contact ratio, and higher contact fatigue strength. On
the negative side, hypoid gears experience higher sliding velocities, resulting in higher
power losses due to excessive sliding friction. Increasing the pinion size without
3
da
Figure 1.2: A sample hypoid gear pair with a shaft angle and a shaft off-set d a .
4
shaft offset (spiral bevel) increases the size of the final drive significantly, while the
hypoid pinion can be made larger due to shaft off-set to increase the strength of the gear
Any attempt to improve the functional attributes of a hypoid gear pair in terms of
its strength, quality, noise and power efficiency requires an optimization of its design
either by fine-tuning its key parameters that have traditionally been chosen based on
by new-generation hypoid gear cutting machines that allow application of many kinds of
surface modifications [1.7, 1.8]. Hypoid gear design procedures were developed within a
small number of hypoid gear cutting machine tool and cutting tool manufacturers and
practical and theoretical details of hypoid development are still propriety to these
companies [1.9].
In general, two different basic cutting methods are used to generate hypoid gears,
indexing, which have their own advantages over each other. The FM process that was
the primary hypoid cutting method for decades has been taken over by the FH process in
continuous indexing [1,2, 1.10-1.12]. However, it is safe to state that the technology
level and design understanding of the FH process is almost a decade behind the face
milling process [1.13]. One reason for this is that newer machining methods
5
Worm gears
Hypoid gears,
Face gear
Efficiency
Pinion size
Bevel gears (Straight,
Sliding
6
such as grinding are applicable to FM process, while there is still no such alternative
Having high gear ratios in hypoid gears in automotive applications causes the gear
to have usually 3 to 4 times the number of teeth of the pinion, which justifies designing
the gear surface as simple as possible to increase production efficiency and minimize
manufacturing time. One typical cost-effective cutting method, called Formate®, is much
faster than the Generating methods. In Formate®, only a few degrees of freedom of
motions are allowed between the cutter and the gear blank (compared to the Generate
cutting method). Therefore, most of the surface modifications are applied to pinion tooth
characteristics including its contact pattern, the motion transmission error (TE),
efficiency and sensitivity to misalignments. The geometric accuracy of a single gear has
limited significance here as the geometry of the mating gear and the assembly errors can
have been quantified either by using FE-based hypoid gear load distribution models or by
experimental means, both of which are very time-consuming and expensive. Due to their
significant computational burden, FE-based hypoid gear contact models are not suitable
for design and parameter and misalignment sensitivity studies. The aim of this study is to
develop computationally efficient, semi-analytical loaded tooth contact models for both
7
FM and FH hypoid gears with or without misalignments. The main motivation for this
surfaces and employ them in a gear contact mechanics formulation to predict unloaded
and loaded contact characteristics as well as functional metrics such as the transmission
In his writings, Aristotle (about 330 BC), made mention of gears and their
commonality. The earliest recognized relic of ancient time gearing is the south pointing
chariot with pinned gears used by the Chinese in about 2600 BC [1.14]. According to the
know today starts with Euler (1781) who proposed the concept of an involute curve
(1781), followed by others such as Willis (1841), Olivier ( 1842) and Gochman (1886)
who developed basic ideas of conjugacy and the foundations of modern gear geometry.
As for spiral bevel gears, Monneret (1899) filed the first patent for spiral bevel
generating method. About two decades later in 1910, Böttcher was issued a series of
patents that addressed both face hobbing and face milling methods [1.2]. Wildhaber’s
earlier papers and patents formed the basis for many of today’s hypoid gear geometry and
generation approaches [1.1, 1.3]. Wildhaber pointed out the significance of using
principal curvatures and directions in establishing hypoid gear geometry [1.16, 1.17].
8
Baxter’s later formulas, based on vector notation, helped condense the formulations to
facilitate the use of computers for definition of surfaces and tooth contact analysis (TCA)
[1.18]. He also developed one of the first unloaded tooth contact model for simulation of
mismatched surfaces of gears generated by Gleason type machines and studied the effects
Coleman [1.20, 1.21]. Krenzer published a series of formulations for unloaded tooth
contact analysis (UTCA) of spiral bevel and hypoid gears [1.22]. These formulations
were useful for the gears manufactured by a class of machinery but were quite difficult to
adapt since the logic behind his formulae were not given. Nearly two decades later, he
proposed a loaded tooth contact analysis model without providing details of the geometry
and the contact analysis. This model used the simplified cantilever beam formula of
Westinghouse to estimate the compliance of the tooth [1.23]. Within the same time
frame, Litvin and Gutman [1.24-1.27] published a series of papers on synthesis and
contact characteristics at a mean point. They determined the contact points, the
instantaneous contact length and direction by conventionally using the surface principal
curvatures and directions. They used a conventional approach to find the contact points,
the instantaneous contact directions and the instantaneous contact lengths utilizing
optimized machine settings for limited cutting methods such as spiral bevel gears cut by
face-milling method was later continued by Litvin and Fuentes [1.28]. As these studies
9
focused on calculating surface coordinate of FM spiral bevel and hypoid gears, there are
Among the few published studies on calculating and “optimizing” TCA, Stadtfeld
[1.2] appears to be the only investigator who used the ease-off approach. He provided a
consistent definition for ease-off as well as a procedure to calculate TCA from ease-off,
and calculated instantaneous contact between two surfaces as a line that maintains its
orientation over the tooth area [1.2]. Moreover, he utilized ease-off to optimize UTCA of
settings and cutter geometry [1.34]. However, he did not provide a detailed procedure on
how to determine orientation of the instantaneous contact curves. Meanwhile, Fan [1.35]
focused on how to calculate surface coordinates and normal vectors for FH spiral bevel
and hypoid gears cut by using the generation method. The solution to the set of equations
that determines contact points where the collinearity condition for the normal vectors of
two mating surfaces is satisfied is typically subject to various numerical instabilities. Fan
[1.35] used the conventional approach for UTCA in conjunction with the Euler-
Rodrigues’ formula to avoid these stability issues. He also used minimization of the
separation between the tooth surfaces to determine the direction and length of the
instantaneous contact lines [1.8]. Later Vogel et al [1.36, 1.37] proposed an alternate
approach to compute both tooth surfaces and the UTCA by using Singularity Theory.
They considered the generated tooth flank as a first-order singularity of the particular
10
function that models the generating process. They also used numerical differentiation to
investigate the sensitivity of tooth contact to machine settings [1.36]. Simon also used
the conventional system of five scalar nonlinear equation and six unknowns to find
contact points on both surfaces. He calculated contact lines orientation by minimizing the
separation function between two contact surfaces and applied his method for a FM
Gleason type gear pair with a generated pinion and a Formate® gear [1.38].
are quite sparse. Simon [1.39-1.41] used a FE model to calculate deflection and
displacement under load from which interpolation functions were obtained to estimate
developed a loaded tooth contact analysis (LTCA) model for spiral bevel gears by using
pinion and gear tooth pair. Wilcox et al [1.43] also developed a FE-based model to
calculate the spiral bevel and hypoid gear tooth compliances by using a three-dimensional
model of a tooth including base deformations, which was later employed by Fan and
Wilcox [1.44] to perform LTCA analyses. Vimercati and Piazza [1.45] also calculated
FH gear pair surfaces and incorporated them with a commercially available finite
elements (FE) package [1.46] to calculate both TCA and LTCA. This particular hypoid
FE package that employs FE away from the contact zone and a semi-analytical contact
formulation at the contact zone [1.47] is perhaps the most advanced hypoid LTCA model
11
available to accurately simulate a hypoid gear contact. The major drawback of these FE-
based models is that they require a considerable amount of computation time, making
them more of an analysis tool. Their use for design tasks such as parameter and assembly
Beside the FE method, the Boundary Element (BE) method was also used in
several studies for performing LTCA. For instance, Sugyarto [1.48] sliced gear and
pinion teeth into a number of sections and considered each section as an independent
plate from its neighboring sections, and applied a two dimensional BEM formulation to
each slice to compute bending and shear deflections. Liu [1.49] applied same compliance
methodology to face gears [1.49] with a correction intended to couple each slice with
adjacent slices using Borner’s coefficient [1.50], which was originally proposed for
parallel axis gear. Vecchiato [1.51] used a three-dimensional BE approach for loaded
tooth contact predictions of FH hypoid gears. As for unloaded contact analysis, he used
deformation solutions. Adding linear thickness variation along the profile to the
originally proposed plate solution [1.52], Yakubek [1.53] used the Rayleigh-Ritz Energy
Method to calculate the approximate deflection of a tapered plate for estimating the
compliance of spur and helical gears. Bending deformations of a tooth were considered
12
as a sum of shape functions that satisfy clamped-free and free-free boundary conditions,
and the unknown coefficients of the shape functions were determined by minimizing the
potential energy. Later, Yau [1.54, 1.55] expanded this compliance model to add shear
deformations to the energy function and found more realistic deformation for spur and
helical gears and Stegemiller [1.56, 1.57] used the FE package ANSYS to propose an
approximate interpolation based formula to compensate for base rotation and base
translation. All of these analytical compliance methods are valid for tooth having
constant height along face width and either constant or linearly varying thickness along
its profile, which is not the case for hypoid gears. Vaidyanathan [1.58-1.60] proposed an
analytical compliance model for a tooth with linearly varying thickness in the profile and
lengthwise directions as well as linearly varying tooth height along the face width. His
Rayleigh-Ritz based formulation used polynomial shape functions and was applied to
both sector and shell geometries. The sector model represents straight bevel gear
geometry closely while the shell model is sufficiently close to a spiral bevel gear tooth in
contact analysis in a practical and computationally efficient way have been the major
road block to the development of other models to study other functional behavior of
hypoid gears. One such behavior is the efficiency of the hypoid gear pair. In addition to
the analytical surface geometries and surface velocities, an accurate description of the
13
contact load distribution is required at many rotational increments (of pinion angle) to
predict the distribution of the friction coefficient and the resultant mechanical power
losses. While this hypoid efficiency methodology has been demonstrated by Xu and
Kahraman [1.61-1.63] by using the FE-based loaded tooth contact model of Vijayakar
[1.64], the amount of computations required was reported to be very significant for it to
be used extensively as a design tool. Likewise, other hypoid gear models for simulation
of surface wear [1.65] and finishing processes such as lapping have also been hampered
velocities forms the basis for predicting the required functional parameters of the hypoid
gear pair, including the transmission error, contact stresses, root bending stresses, fatigue
life and mechanical power losses. It is evident from the review of the literature that a
model to compute the load distribution accurately and efficiently without resorting to
computationally demanding FE methods does not exist. This is mainly due to three
primary reasons:
(i) A detailed description of general and reliable formulation to define the geometry
14
(ii) Conventional methods of matching the tooth surfaces (bringing them to contact)
(iii) There is no published model available for hypoid gear LTCA based on semi-
gear contact models that address these issues. This study performs the following specific
surface geometries of hypoid gears including the coordinates, normal vectors and
radii of curvatures.
(ii) Development of a novel formulation for unloaded tooth contact analysis by using
the ease-off topography, surface of action and roll angle surface to predict
15
(iv) Development of a LTCA model for FH and FM hypoid gears to predict pressure
various types.
tooth surfaces, which affect contact patterns and transmission error significantly.
Moreover, wear or lapping simulations (as accumulated wear) changes surface geometry
usually in a very local fashion that conventional tooth contact analysis approaches are not
capable of capturing. A novel ease-off based approach will also be developed to modify
ease-off topography of the theoretically generated tooth surfaces to account for both
global deviations due to the manufacturing process and local surface deviations due to
dependent (mechanical) power losses of hypoid gear pairs. For this purpose, the
proposed loaded tooth contact model will be combined with a new friction model
mechanical power losses and gear pair efficiency including all relevant contact, surface,
and lubricant parameters as well as the operating conditions. This hypoid gear efficiency
model will be used to investigate the impact of basic design parameters, and surface and
16
lubricant conditions, on mechanical power losses of hypoid gear pairs and to arrive at
The overall methodology used to develop the hypoid load distribution model is
illustrated in the flowchart of Figure 1.4. Gear blank dimensions, cutter geometry,
machine settings, assembly dimensions and misalignments, torque and speed are all
included as input parameters for the load distribution model. These parameters are
commonly put together by hypoid gear manufacturers in a standard form that is called a
special analysis file. The pinion and gear cutter surfaces are first constructed and used to
define the extended pinion and gear surfaces (including surface coordinates, normals and
curvatures) by applying fundamental equation of meshing between a gear blank and its
respective cutter surfaces. These extended tooth surfaces are then trimmed in 3D space
so that they are contained by the blanks, and transformed to a global coordinate system
where any misalignments in the directions of shaft offset (ΔE), pinion axis (ΔP), gear axis
(ΔG) as well as the shaft angle error (ΔΣ) can be applied. Next, ease-off and surface of
roll angle are constructed and an UTCA model is developed by bringing the tooth
17
(b)
Generating gear
(a) Generating gear
Circular
FM Method Reading Design File FH Method CG
Extended
epicycloids trace
CG
OB
(Generate / Formate) .HAP or .SPA file (Generate / Formate) t
t
Blade
IB c
Fixed OB Ct
IB
Cutter
Cutter
Loaded Tooth
Contact Analysis
Required Data for Hypoid Gear Mechanical
(MPa)
758
Efficiency Analysis Power Loss and Efficiency
2
Y (mm) Toe
674
0 590
506
-2 421
-4 337
253
Figure 1.4: Flowchart of overall hypoid gear loaded tooth contact analysis methodology.
Next the tooth compliance matrices comprising bending, shear, Hertzian and base
conditions are defined and solved simultaneously to compute the load distribution and the
loaded transmission error of the hypoid gear pair. Moreover, all required information for
In Chapter 2, the hypoid gear tooth surfaces will be defined through simulation of
the face-milling and face-hobbing processes with all relevant cutter and machine related
parameters included. A new formulation of unloaded tooth contact analysis based on the
principles of ease-off and a newly introduced surface of roll angle will be proposed as
well.
loaded tooth contact model will be described. A novel approach will be introduced in
Chapter 4 to compute loaded tooth contacts of gear surfaces that have deviations from
pairs. This mode combines the developed computationally efficient contact model and
19
Kahraman [1.66] to predict gear mesh power losses and mechanical efficiency. A
References of Chapter 1
[1.2] Stadtfeld, H., J., 1993, Handbook of Bevel and Hypoid Gears, Rochester Institute
of Technology.
[1.3] Stewart, A. A., and Wildhaber, E., 1926, "Design, Production and Application of
the Hypoid Rear-Axle Gear." J. SAE, 18, pp. 575-580.
[1.4] Wang, X. C., and Ghosh, S. K., 1994, Advanced Theories of Hypoid Gears,
Elsevier Science B. V.
[1.5] Dooner, D. B., and Seireg, A., 1995, The Kinematic Geometry of Gearing: A
Concurrent Engineering Approach, John Wiley & Sons Inc.
[1.6] Stadtfeld, H., J., 1995, Gleason Advanced Bevel Gear Technology, The Gleason
Works.
[1.7] Coleman, W., 1963, Design of Bevel Gears, The Gleason Works.
[1.8] Fan, Q., 2007, "Enhanced Algorithms of Contact Simulation for Hypoid Gear
Drives Produced by Face-Milling and Face-Hobbing Processes." ASME J. Mech.
Des., 129(1), pp. 31-37.
20
[1.9] Dooner, D. B., 2002, "On the Three Laws of Gearing." ASME J. Mech. Des., 124,
pp. 733-744.
[1.11] Litvin, F. L., and Fuentes, A., 2004, Gear Geometry and Applied Theory (2nd
ed.), Cambridge University Press, Cambridge.
[1.12] Krenzer, T. J., 1990, "Face Milling or Face Hobbing." AGMA, Technical Paper
No. 90FTM13.
[1.13] Stadtfeld, H. J. (2000). "The Basics of Gleason Face Hobbing." The Gleason
Works.
[1.14] Dudley, D. W., 1969, The Evolution of the Gear Art, American Gear
Manufacturers Association, Washington, D. C.
[1.16] Wildhaber, E., 1956, "Surface Curvature." Product Engineering, pp. 184-191.
[1.17] Dyson, A., 1969, A General Theory of the Kinematics and Geometry of Gears in
Three Dimensions, Clarendon Press, Oxford.
[1.18] Baxter, M. L., 1964, "An Application of Kinematics and Vector Analysis to the
Design of a Bevel Gear Grinder." ASME Mechanism Conference, Lafayette, IN.
21
[1.20] Coleman, W. "Analysis of Mounting Deflections on Bevel and Hypoid Gears."
SAE 750152.
[1.22] Krenzer, T. J., 1965, TCA Formulas and Calculation procedures, The Gleason
Works.
[1.23] Krenzer, T. J., 1981, "Tooth Contact Analysis of Spiral Bevel and Hypoid Gears
under Load." SAE Earthmoving Industry Conference, Peoria, IL.
[1.24] Litvin, F. L., and Gutman, Y., 1981, "Methods of synthesis and analysis for
hypoid gear-drives of formate and helixform; Part I-Calculation for machine
setting for member gear manufacture of the formate and helixform hypoid gears."
ASME J. Mech. Des., 103, pp. 83-88.
[1.25] Litvin, F. L., and Gutman, Y., 1981, "Methods of synthesis and analysis for
hypoid gear-drives of formate and helixform; Part II-Machine setting calculations
for the pinions of formate and helixform gears." ASME J. Mech. Des., 103, pp.
89-101.
[1.26] Litvin, F. L., and Gutman, Y., 1981, "Methods of synthesis and analysis for
hypoid gear-drives of formate and helixform; Part III-Analysis and optimal
synthesis methods for mismatched gearing and its application for hypoid gears of
formate and helixform." ASME J. Mech. Des., 103, pp. 102-113.
[1.27] Litvin, F. L., and Gutman, Y., 1981, "A Method of Local Synthesis of Gears
Grounded on the Connections Between the Principal and Geodetic of Surfaces."
ASME J. Mech. Des., 103, pp. 114-125.
22
[1.28] Litvin, F. L., Fuentes, A., Fan, Q., and Handschuh, R. F., 2002, "Computerized
design, simulation of meshing, and contact and stress analysis of face-milled
formate generated spiral bevel gears." J. Mechanism and Machine Theory, 37(5),
pp. 441-459.
[1.29] Fong, Z. H., and Tsay, C.-B., 1991, "A Mathematical Model for the Tooth
Geometry of Circular-Cut Spiral Bevel Gears." ASME J. Mech. Des., 113, pp.
174-181.
[1.30] Fong, Z. H., 2000, "Mathematical Model of Universal Hypoid Generator with
Supplemental Kinematic Flank Correction Motions." ASME J. Mech. Des.,
122(1), pp. 136-142.
[1.31] Tsai, Y. C., and Chin, P. C., 1987, "Surface Geometry of Straight and Spiral
Bevel Gears." J. Mechanism, Transmission and Automation in Design, 109, pp.
443-449.
[1.32] Fong, Z. H., and Tsay, C.-B., 1991, "A Study on the Tooth Geometry and Cutting
Machine Mechanisms of Spiral Bevel Gears." ASME J. Mech. Des., 113, pp. 346-
351.
[1.33] Litvin, F. L., Zhang, Y., Lundy, M., and Heine, C., 1988, "Determination of
Settings of a Tilted Head Cutter for Generation of Hypoid and Spiral Bevel
Gears." J. Mechanism, Transmission and Automation in Design, 110, pp. 495-
500.
[1.34] Stadtfeld, H. J., and Gaiser, U., 2000, "The Ultimate Motion Graph." ASME J.
Mech. Des., 122(3), pp. 317-322.
23
[1.35] Fan, Q., 2006, "Computerized Modeling and Simulation of Spiral Bevel and
Hypoid Gears Manufactured by Gleason Face Hobbing Process." ASME J. Mech.
Des., 128(6), pp. 1315-1327.
[1.37] Vogel, O., Griewank, A., and Bär, G., 2002, "Direct gear tooth contact analysis
for hypoid bevel gears." Computer Methods in Applied Mechanics and
Engineering, 191(36), pp. 3965-3982.
[1.38] Simon, V., 1996, "Tooth Contact Analysis of Mismatched Hypoid Gears." ASME
International Power Transmission and Gearing Conference ASME, 88, pp. 789-
798.
[1.39] Simon, V., 2000, "Load Distribution in Hypoid Gears." ASME J. Mech. Des.,
122(44), pp. 529-535.
[1.40] Simon, V., 2000, "FEM stress analysis in hypoid gears." J. Mechanism and
Machine Theory, 35(9), pp. 1197-1220.
[1.41] Simon, V., 2001, "Optimal Machine Tool Setting for Hypoid Gears Improving
Load Distribution." ASME J. Mech. Des., 123(4), pp. 577-582.
[1.42] Gosselin, C., Cloutier, L., and Nguyen, Q. D., 1995, "A general formulation for
the calculation of the load sharing and transmission error under load of spiral
bevel and hypoid gears." J. Mechanism and Machine Theory, 30(3), pp. 433-450.
[1.43] Wilcox, L. E., Chimner, T. D., and Nowell, G. C., 1997, "Improved Finite
Element Model for Calculating Stresses in Bevel and Hypoid Gear Teeth."
AGMA, Technical Paper No. 97FTM05.
24
[1.44] Fan, Q., and Wilcox, L., 2005, "New Developments in Tooth Contact Analysis
(TCA) and Loaded TCA for Spiral Bevel and Hypoid Gear Drives." AGMA,
Technical Paper No. 05FTM08.
[1.45] Vimercati, M., and Piazza, A., 2005, "Computerized Design of Face Hobbed
Hypoid Gears: Tooth Surfaces Generation, Contact Analysis and Stress
Calculation." AGMA, Technical Paper No. 05FTM05.
[1.46] Vijayakar, S., 2004, Calyx Hypoid Gear Model, User Manual, Advanced
Numerical Solution Inc., Hilliard, Ohio.
[1.47] Vijayakar, S. M., 1991, "A Combined Surface Integral and Finite Element
Solution for a Three-Dimensional Contact Problem." International J. for
Numerical Methods in Engineering, 31, pp. 525-545.
[1.48] Sugyarto, E., 2002, "The Kinematic Study, Geometry Generation, and Load
Distribution Analysis of Spiral Bevel and Hypoid Gears," M.Sc. Thesis, The Ohio
State University, Columbus, Ohio.
[1.49] Liu, F., 2004, "Face Gear Design and Compliance Analysis," M.Sc. Thesis, The
Ohio State University, Columbus, Ohio.
[1.50] Borner, J., Kurz, N., and Joachim, F. (2002). "Effective Analysis of Gears with
the Program LVR (Stiffness Method)."
[1.51] Vecchiato, D., 2005, "Design and Simulation of Face-Hobbed Gears and Tooth
Contact Analysis by Boundary Element Method," Ph.D. Dissertation, University
of Illinois at Chicago.
[1.52] Timoshenko, S. P., and Woinowsky-Krieger, S., 1959, Theory of Plates and
Shells, McGraw-Hill Book Company Inc.
25
[1.53] Yakubek, D., Busby, H. R., and Houser, D. R., 1985, "Three-Dimensional
Deflection Analysis of Gear Teeth Using Both Finite Element Analysis and a
Tapered Plate Approximation." AGMA, Technical Paper No. 85FTM4.
[1.54] Yau, H., 1987, "Analysis of Shear Effect on Gear Tooth Deflections Using the
Rayleigh-Ritz Energy Method," M.Sc. Thesis, The Ohio State University,
Columbus, Ohio.
[1.55] Yau, H., Busby, H. R., and Houser, D. R., 1994, "A Rayleigh-Ritz Approach to
Modeling Bending and Shear Deflections of Gear Teeth." J. of Computers &
Structures, 50(5), pp. 705-713.
[1.56] Stegmiller, M. E., 1986, "The Effects of Base Flexibility on Thick Beams and
Plates Used in Gear Tooth Deflection Models," M.Sc. Thesis, The Ohio State
University, Columbus, Ohio.
[1.57] Stegmiller, M. E., and Houser, D. R., 1993, "A Three Dimensional Analysis of
the Base Flexibility of Gear Teeth." ASME J. Mech. Des., 115(1), pp. 186-192.
[1.58] Vaidyanathan, S., 1993, "Application of Plate and Shell Models in the Loaded
Tooth Contact Analysis of Bevel and Hypoid Gears," Ph.D. Dissertation, The
Ohio State University, Columbus, Ohio.
[1.59] Vaidyanathan, S., Houser, D. R., and Busby, H. R., 1993, "A Rayleigh-Ritz
Approach to Determine Compliance and Root Stresses in Spiral Bevel Gears
Using Shell Theory." AGMA, Technical Paper No. 93FTM03.
[1.60] Vaidyanathan, S., Houser, D. R., and Busby, H. R., 1994, "A Numerical
Approach to the Static Analysis of an Annular Sector Mindlin Plate with
Applications to Bevel Gear Design." J. of Computers & Structures, 51(3), pp.
255-266.
26
[1.61] Xu, H., 2005, "Development of a Generalized Mechanical Efficiency Prediction
Methodology for Gear Pairs," Ph.D. Dissertation, The Ohio State University,
Columbus, Ohio.
[1.62] Xu, H., and Kahraman, A., 2007, "Prediction of Friction-Related Power Losses of
Hypoid Gear Pairs." Proceedings of the Institution of Mechanical Engineers, Part
K: J. Multi-body Dynamics, 221(3), pp. 387-400.
[1.63] Xu, H., Kahraman, A., and Houser, D. R., 2006, "A Model to Predict Friction
Losses of Hypoid Gears." AGMA, Technical Paper No. 0FTM06.
[1.64] Vijayakar, S. M., 2003, Calyx User Manual, Advanced Numerical Solution Inc.,
Hilliard, Ohio.
[1.65] Park, D., and Kahraman, A., 2008, "A Surface Wear Model for Hypoid Gear
Pairs." In press, Wear.
[1.66] Li, S., and Kahraman, A., 2009, "A Mixed EHL Model with Asymmetric
Integrated Control Volume Discretization." Tribology International, Hiroshima,
Japan.
27
CHAPTER 2
2.1. Introduction
Unlike most types of gears that have closed-form equations defining their
geometry, the geometry of hypoid gears can only be computed by solving implicit
equations governed by the manufacturing process, including its machine settings and
cutter specifications. Besides the gear blank dimensions and basic geometry
Among them, the contact pattern (location, size, shape) on the gear tooth surfaces and the
motion transmission error amplitude of the gear pair are two of the most common ones
checked routinely in the design of hypoid gear pairs. The contact pattern and the
transmission error are both determined via a contact analysis of the pinion and gear tooth
28
In general, contact of hypoid gear surfaces is single-mismatched. This is the most
general case of point contact condition between two surfaces [2.1]. The purpose of the
unloaded tooth contact analysis (UTCA) is to determine a contact point path (CPP) on
instantaneous contact point that falls in a specified separation distance (usually 6.3
addition, UTCA results in the function of motion transmission error between two gear
axes that is viewed as a key metric used to estimate the noise/level of the hypoid gear pair
In this chapter, as the first basic step in the analysis of hypoid gears, the geometry
of both face-milled (FM) and face-hobbed (FH) hypoid gear pairs produced by using both
Formate® and Generate cutting methods will be computed. This will be done by
simulating individual cutting processes. Basic machine tool settings, cutter geometry
parameters and gear blank dimensions will form the input for this computation. Next, a
novel method based on the ease-off topography will used to determine the unloaded
contact patterns. The proposed ease-off based methodology finds the instantaneous
contact curve through a surface of roll angles, allowing an unloaded tooth contact
29
2.2. Definition of Tooth Surface Geometry
The concept of the generating gear is a key to the basic understanding of hypoid
gears because this hypothetical gear can be treated as cutting tool for both the pinion and
the gear [2.5]. In a FM cutter head, blades are arranged around the cutter head axis on an
equal radius for inside and outside blades (IB and OB, respectively) to form a conical
shape due to cutter axis rotation. The inside blades cut convex side of a tooth slot while
the outside blades cut concave side of the same slot, as shown in Figure 2.1(a). Face-
hobbing cutter heads (such as PENTAC® or TRI-AC®) like the one shown in Figure
2.1(b) roll while cutting such that each set of IB-OB blades (called blade group) will pass
through a different tooth slot. The cutting process can be considered as rolling of two
gears together, except the teeth of one of the gears are replaced by blade group of the
cutter head. By rolling the cutter head and the gear blank together while advancing the
cutter head into the blank, the gear is cut by the continuous indexing method. While the
axis of the generating gear for FM process is fixed, it is located on the center of a circle
Ct for FH process that rolls on the generating gear circle CG , as shown in Figure 2.1(b).
Therefore, the edges of a blade in FH process traces extended epicycloids since they
usually lie on a radius that is larger than the radius of rolling circle Ct .
30
(a) Generating gear
Circular
CG arc trace
OB
t
IB
Fixed
Cutter
(b)
Generating gear
Extended
CG epicycloids trace
t
c Blade
group
Ct
IB
OB
Cutter
31
2.2.1. Kinematics
degrees of freedom and relative motions through machine settings to accommodate the
cutting process of the gear and pinion blanks by means of the cutter blades. Figure 2.2
shows a typical cradle-based hypoid generator with machine settings and relative motions
defined as the cutter phase angle t , the tilt angle iT , the swivel angle js , the radial
setting Sr , the cradle angle c , the sliding base X B , the machine root angle m ,
machine center to back M ctb , the blank offset Eb , the blank phase angle g , angular
speed of the cutter axis t , angular speed of the cradle axis c , angular speed of the
blank axis g , and the cradle angle change q. Newer-generation hypoid gear machine
tools still use settings in the form of older cradle-based machines. While the new cutting
machines are not set as the old mechanical machines, their working principles are still the
same such that they generate the same gear surface with an added capability of
machine settings are typically fixed (kept constant), a number of them are defined as a
polynomial function of q. These parameters that are dependent on q are called higher-
order motions such as the modified roll (ratio of roll change), the helical motion (sliding
32
c
Sr
js
iT
Eb
t g
M ctb
m
XB
33
In the FM process, the teeth are cut individually by blade edges that rotate fast
about the cutter axis. The cutting edges of the blades form a conical surface and their
envelope seen from a coordinate attached to the blank is the gear or pinion surface. In the
Formate® case for the FM process, the blank is fixed, the cutter advances towards the
gear blank while rotating, in the process replicating its surface on the blank. Then, the
cutter head retreats, the blank is rotated by one tooth spacing, and the same cutting
using Formate® requires the blank and cutter head to be rolled together according to a
Ro
g Nt
Rtg (2.1)
t Ng
indexing) of angular velocity of the gear blank, and Nt and N g are numbers of blade
groups and gear teeth being cut. Typically, edges of the blades in FH process do not
intersect with cutter head axis so that the cutting surface is a hyperboloid of revolution.
relative rotation between the gear blank and the cradle axis in Figure 2.1(b). The ratio of
34
g
RaFM , (2.2a)
c
Ge
g
RaFH (2.2b)
c
where RaFM and RaFH are the ratios of rolls for the FM and FH processes (they are zero
for Formate® process), respectively, c is the angular velocity of the cradle axis, g is
the angular velocity of the blank axis during face-milling and Ge
g is the generating
portion of blank angular velocity for the FH process. As a result, the total blank angular
FM
g g , (2.3a)
FH Ro Ge
g g g (2.3b)
FM FM
g Ra q, (2.4a)
FH FH
g ( Ra q Rtg t ) . (2.4b)
35
Here, the relative motion between the cutter and the gear blank for the FM process can be
treated as a special case of the FH process. Therefore, the formulation for the FH process
will be explained here in detail while its differences from the FM process will also be
specified, as required.
Figure 2.3 shows a typical FH blade with its geometry along its cutting edge is
defined by the blade angle b , the rake angle , the hook angle , the blade offset angle
b , the cutter radius rc , and the distance from the tip of blade to reference point h f .
The cutting edge is divided into four different sections as the edge (or tip radius), toprem,
profile and flankrem that are all shown in Figure 2.3(c). The edge and flankrem are
usually circular arcs while toprem is usually a straight line at a slight angle from the
profile section. Most of the cutting is done by the profile section of the blade that is
relative to the local coordinate system Xb fixed to the cutter head (with its origin at
reference point M) where s is the distance of point A to point M along the blade edge.
With this, the unit tangent vector is t r s , and if the cutting edge is a line, it can be
reduced to
36
(a)
yt Reference Plane
xb
x , x
xt rc b
M
zb , z z
Inside Blade 31
(b) (c) b
Flankrem
A
r
xb Profile t s
M xb , xt
x , x
M
n
yb hf
y b Toprem
y 18
Edge
rc
zt zb
Figure 2.3: (a) Cutter head, (b) blade and (c) cutting edge geometry.
37
t [ sin b 0 cos b 1]T (2.5)
where the superscript T denotes a matrix transpose. Position and unit tangent vectors, rt
and t t , in the coordinate system Xt whose z axis coincides with the cutter axis and xy
plane passes through reference point M as shown in Figure 2.3(a) are given as
rt = rc t i s t t (2.6b)
where
(k [ x, y, z ]) such that
1 0 0 0
0 cos sin 0
M x ( ) , (2.6d)
0 sin cos 0
0 0 0 1
cos 0 sin 0
0 1 0 0
M y ( ) , (2.6e)
sin 0 cos 0
0 0 0 1
38
cos sin 0 0
sin cos 0 0
M z ( ) . (2.6f)
0 0 1 0
0 0 0 1
blank as [2.6]:
rg M x ( g ) M Eb M ctb M y ( m ) M m X B M X B c M z (c )
(2.7a)
M z (q ) M Sr j M z ( js ) M x (iT ) M z ( t ) rt
1 0 0 Sr
0 1 0 0
M Sr j , (2.7b)
0 0 1 0
0 0 0 1
1 0 0 0
0 1 0 0
M X B c , (2.7c)
0 0 1 XB
0 0 0 1
1 0 0 M ctb
0 1 0 0
M m X B , (2.7d)
0 0 1 0
0 0 0 1
39
1 0 0 0
0 1 0 Eb
M Eb M ctb . (2.7e)
0 0 1 0
0 0 0 1
Variables influencing rg , with the exception of s and t , are either fixed or dependent
blank whose envelope is the generated surface on the blank. The envelope of rg with
n (Vc Vw ) 0 , which states for each point to lie on the envelope surface that the
normal vector n to the family of the cutter surfaces should be perpendicular to relative
velocity between the blank (w) and the cutter (c) as shown in Figure 2.4.
40
Vc Vw
Vc Vw
Cradle axis
Vc Vw
Cutter head
c w
Blank
Blank axis
41
Using Eq. (2.8), each point on the generated surface can be found by solving a
system of two implicit nonlinear equations for a pair of unknown parameters that can be
chosen as ( s, t ) , ( s, q) or (q, t ) .
curvilinear variables s and t on the gear surface, the unit normal to the surface is given
as
rg rg
s t
n0 . (2.9)
rg rg
s t
of the surface variables s, the change of unit normal to the surface is defined as [2.10]
dn
kt1 t1 t1 v1 . (2.10)
d ( s, t )
Here
P P
t1 1 0 , (2.11a)
P1 P0
42
n2
n1 n0
C2
P2
C1
t2
P1
t1
P0
v1
43
n0 t1
v1 (2.11b)
n0 t1
and ( s, t ) defines the distance from P0 to P1 along the gear surface (along curve C1 )
as a function of s and t (here, t t 0 constant and only s varies). kt1 is the normal
curvature in the t1 direction and t1 is the geodesic torsion in the direction of v1 . Here
n0 , t1 and v1 form a Frenet trihedron. Following the same procedure, but this time
curvatures kt2 and geodesic torsion t2 in t 2 and v 2 directions are found according to
dn
kt2 t 2 t2 v 2 (2.12)
d ( s, t )
with
P2 P0
t2 , (2.13a)
P2 P0
n0 t 2
v2 (2.13b)
n0 t 2
and ( s, t ) defines the distance from P0 to P2 along the gear surface (along curve C2
in Figure 2.5) as a function of s and t (here s s0 is constant and only t varies). With
44
kt1 , kt2 , t1 and t2 in hand, Euler equation [2.11, 2.12] is applied to compute the
Having the principal curvatures and directions of every possible contact point
(points on the projection plane) on the pinion and gear, principal directions of the
difference surface are defined as directions in which two contacting surfaces have the
and gear surfaces in the direction of maximum relative curvatures are used to find the
(i) the contact point path (CPP) on each of the gear surfaces in addition to the zone on
each surface in the neighborhood of each instantaneous contact point that is as close
Two different approaches were used in the past for performing UTCA of hypoid gears
with mismatched surfaces. In the conventional method, tooth surfaces are treated as two
45
arbitrary surfaces, rotating about the pinion and gear axes. The contact point path (CPP)
on each surface is computed by satisfying two contact conditions. The first condition is
the coincidence of position vector tips of the points on the gear and pinion surfaces in
three dimensional space. The second condition is the collinearity of the normal vectors
The second method of performing UTCA is based on the ease-off procedure. The
current literature lacks a clear and accurate mathematical definition of ease-off as well as
its construction including the instantaneous contact lines/curves [2.17]. Ease-off has
often been defined in the literature as the change in pinion surface with the application of
modifications. While these changes directly reshape ease-off, they do not constitute the
ease-off itself.
Finding the location and orientation of potential instantaneous contact lines is one
major step in the UTCA. In general, the instantaneous contact shape in the projection
plane (plane that includes the gear axes) is slightly curved as opposed to commonly used
approximate straight lines, which is the contact shape in action surface. The
46
In the next two sections, a brief overview of the conventional approach will be
thesis.
The position vector r p (1, 1 ) and normal vector n p (1, 1 ) of any point on the
surface of the pinion can be defined by two independent curvilinear local surface
variables 1 and 1 . Similarly, the position and normal vectors of any point on the
surface of the gear are given as r g ( 2 , 2 ) and n g ( 2 , 2 ) where 2 and 2 are the
independent curvilinear local variables of the gear surface. Pinion surface coordinate
r p (1, 1 ) and normal n p (1, 1 ) are rotated about the pinion axis a p as much as an
rotated about the gear axis a g by an angle g to satisfy the two contact conditions
defined below:
47
Here, OE is the offset vector that connects the origins of the pinion and the gear. Eq.
by Eq. (2.14) can be solved for the remaining five unknowns. However, due to the high
level of conformity of the pinion and gear surfaces in the vicinity near the contact point,
addition, the solution is very sensitive to the initial guesses. Provided these numerical
difficulties can be overcame, the solution of Eq. (2.14) yields the coordinates of contact
point path (CPP) on the pinion and gear surfaces as well as the angular position of the
With the conventional method, at each point of the contact point path (CPP), a
direction in which separation between two surfaces (here pinion and gear surfaces) is
minimum [2.11, 2.12, 2.18, 2.19] is assumed to be potential contact line. For this
purpose, the two contacting surfaces are approximated as two contacting ellipsoids with
both pinion ( K1p and K 2p ) and gear ( K1g and K 2g ) surfaces respectively and the
corresponding principal directions ( e1p , e 2p , e1g and e2g ) are all required to find
direction in which relative normal curvature between ellipsoids is minimal. This direction
48
e 2g u
g
e 2p
n
e1p
(a) v
k1p
T
k2g
e1g
M
k1g k2p
p
v u
(b)
1L20
L1
Figure 2.6: General case of approximating gear surfaces as two contacting ellipsoids to
orient instantaneous contact line.
49
u, as shown in Figure 2.6, having an unknown angle from e1p in the tangential plane
T is the major direction of instantaneous contact ellipse. Utilizing the Euler equation
[2.11, 2.12], the pinion normal curvature along any arbitrary direction u with an angle
kupg k1g cos 2 ( ) k2g sin 2 ( ) k1p cos 2 () k2p sin 2 () . (2.16)
dkupg
0 (2.17)
d
such that
1 sin(2)
tan 1 , (2.18a)
2 pg
K12 cos(2)
50
p p
pg k1 k2
K12 . (2.18b)
k1g k2g
where is the unloaded separation distance [2.12]. In the conventional method, it has
been assumed that instantaneous contact lines spread equally on both sides of
contacting surfaces with local ellipsoids results in the same relative curvatures on both
sides of contact points. In other words, the conventional method fails to include the
variation of the curvature along a given contact line, resulting in contact line length
function that defines separation between contacting surfaces in the direction of the
normal to the gear surface at each contact point. Although his approach does not acquire
principal curvatures and directions information on each contact point, it still bears the
some level of computational complexity and inefficiency since such minimization of the
separation function requires the solution of a system of seven nonlinear equations and
seven unknowns. He later mentioned that the instantaneous contact form is a curve
rather than the generally assumed line [2.21]. Fan [2.8] found instantaneous contact line
direction and length without using second order information by finding minimum
51
separation direction. He constructed a cylinder centered at an instantaneous contact point
with an axis that is collinear with the common normal then searched for a direction on the
tangent plane in which the separation between the contacting surfaces is minimal. With
this, another search was performed to find the distance required to move on the both sides
of the minimum separation direction in order to reach the predetermined separation value
. He eliminated the need for the second order surface information and computed more
realistic contact line length estimating different lengths on both sides of contact point
along contact line. However, this method still assumed an instantaneous contact line, as
This study proposes a novel surface of roll angle and utilizes it to orient
instantaneous contact lines/curves without using principal curvatures and directions. This
method requires significantly less computational effort since (i) it does not result in a set
of nonlinear algebraic equations that must be solved numerically, and (ii) it only requires
the coordinates and the normal vector of one contacting surface and the spatial
orientation of the axes of both gears. The instantaneous contact curves are defined
between two conjugate surfaces, namely a given surface and the conjugate surface to the
reference surface with respect to the given spatial orientation of the axes of the gears.
Since the instantaneous contact line orientation is extremely insensitive to local surface
52
changes as it will be demonstrated later, the validity of using a conjugate surfaces instead
Ease-off will be defined in this study as the deviation of real gear surface from
the conjugate of its real mating pinion surface. It can also be defined as the deviation of
real pinion surface from the conjugate of its real mating gear. These definitions are
respectively called the gear-based ease-off and pinion-based ease-off [2.5]. The
conventional method of UTCA seeks a contact between two arbitrary surfaces, failing to
benefit from the fact that the designed gear and pinion surfaces are indeed close to the
conjugate surfaces enables the use of the ease-off concept. The proposed ease-off
(i) providing an overview of the contact pattern and transmission error as well as
(ii) providing more accurate instantaneous contact curves instead of commonly used
(iii) eliminating the need to compute the curvature in order to estimate length and
53
(iv) avoiding the need for initial guesses required by the conventional method to
locate first contact point (some initial guesses might end up divergent solutions).
In this study, the ease-off surface is constructed directly from the relationships between
the continuous cutter surfaces. Therefore, any surface fits to pinion and gear surfaces is
not needed.
The first step in constructing ease-off is to specify an area in the gear projection
plane with the possibility of contact between pinion and gear, as shown in Figure 2.7.
Such an area is the projection of a volume bounded by the faces and front and back cones
of the pinion and gear into the gear projection plane. The projection plane shown in
Figure 2.7 is a gear-based projection plane since its ease-off is defined on gear
tooth area. Using the gear machine settings and Eq. (2.8), the real gear surface
coordinates (shown in Figure 2.7) are computed for every point of the projection plane as
rijg where i [1, mg ] and j [1, ng ] are the indices in lengthwise and profile directions
of the surface point with mg and n g as number of surface grids respective directions.
Then, each point of the gear projection plane is transformed to the pinion coordinate
system to construct the projection plane for the pinion. Next, the real pinion surface
points and unit normals to the surface are computed from Eq. (2.8) for every point of the
pinion projection plane as rijp and nijp . Both real pinion and gear surface coordinates and
unit normal vectors are transformed into the global coordinate system where
54
P
Real gear
surface, rijg
Gear projection
plane
E
Ease-off
surface, ij G
ap
Q Surface ag
Real pinion
surface, rijp
Action
surface, rija
Conjugate of
pinion, r̂ijg
55
misalignments E , P , G and shown in Figure 2.7 are applied as well. Having
rijp , nijp , the pinion and gear axis vectors a p and a g , and the gear ratio R , the action
surface position vector rija is found from conjugacy equation in 3D space as [2.5]:
The same steps are repeated for all points on the real pinion surface to define the
action surface completely. As seen in Figure 2.7, the action surface for a hypoid gear
pair, while quite flat, is not a plane. Any point and its unit normal vector on the real
pinion surface are rotated around the pinion axis in order to satisfy Eq. (2.19). The angle
of rotation ijp of each point that satisfies Eq. (2.19) is then plotted on the projection
plane to construct the pinion roll angle surface Q . The angle ijg = ijp R corresponds to
the amount of rotation from the surface of action to reach the conjugate of pinion surface.
Therefore,
If this conjugate surface of the pinion were to match perfectly with the real gear
surface at any point, then a perfect meshing condition with zero unloaded transmission
error would exist. The difference between these two surfaces (conjugate of pinion and
56
real gear surfaces) in projection plane domain is defined as ease-off surface where the
Any value , of any point on the Q surface is the pinion roll angle. As shown in
Figure 2.8 for a specific pinion roll angle i , intersection of the plane z i and the
Q surface defines x and y coordinates of all points on the projection plane that have the
same roll angle, stating theoretically that they lie on the same contact line/curve. Since
Q is not a plane, this intersection for hypoid gears is usually a curve rather than a
straight line as assumed by most studies. The instantaneous contact curve C (i ) shown
in Figure 2.8 is determined by projecting the intersection curve first on projection plane
and then projecting this projected curve once more on the ease-off surface. The
minimum distance from C (i ) to the projection plane at point H (i ) is the
from point H (i ) along C (i ) within a preset separation distance , gives the unloaded
contact line length L(i ) . Repeating this procedure for every pinion roll angle
increment, unloaded transmission error curve TE () and the unloaded tooth contact
pattern are computed. Here the contact curves are between real pinion and conjugate
gear.
57
Ease-off
z surface
(a) Instantaneous contact
curve, C (i )
Projection
plane
A
Contact curve
projection
y
z i
x
Q i
Q surface
A
(b) z
C (i )
L(i )
H (i )
x
TE (i )
Figure 2.8: Unloaded TCA computation procedure: (a) gear projection plane, ease-off
and Q surfaces, and (b) instantaneous contact curve, contact line and unloaded
transmission error.
58
Replacing the conjugate gear with the real one practically does not change the contact
line orientation and shape, since the effect of microscopic changes of pinion and gear
An example hypoid gear pair whose basic parameters are listed in Table 2.1 for
the drive-side contact (concave side of pinion and convex side of gear) is considered to
demonstrate the capabilities of the proposed hypoid gear geometry computation and
The predicted unloaded transmission error (UTE) curves computed by the model
for three adjacent tooth pairs i 1 , i and i 1 are shown in Figure 2.9. Here UTE is
plotted against the mesh cycles (pinion roll angle) where each of the individual UTE
curves corresponds to a single tooth pair in mesh. Individual curves for two adjacent
tooth pairs are one mesh cycle apart. At the intersection point of the two adjacent UTE
curves ( M1 or M 2 in Figure 2.9), transition from one tooth pair to adjacent tooth pair
occurs. The transmission error value of the intersection point is the maximum UTE,
59
__________________________________________________________________
Table 2.1: Basic drive side geometry and working parameters of the example hypoid gear
pair.
60
30
TE
rad]
i i 1
15
i 1 M1 M2 P-P UTE
1 2 3
Mesh cycles
Figure 2.9: Unloaded transmission error of the example gear pair with misalignments
E 0.15 mm, P 0.12 mm, G 0, 0 .
61
Figure 2.10. The curve marked as CPP is the locus of all instantaneous contact
points between gear and pinion tooth surfaces. On both sides of CPP are right point path
(RPP) and left point path (LPP), which are as far from CPP as it is required to reach
plane shown in Figure 2.8. The CPP, RPP and LPP curves are computed for a single
tooth pair in contact while in case of multiple teeth in contact these curves are partially
active (usually middle part of the curves are active) since the adjacent pairs will take over
the motion. In Figure 2.10, unloaded contact pattern of the tooth of interest is bounded
by the instantaneous contact lines at M1 and M 2 and the RPP and LPP curves.
Using the same example gear pair, the influence of misalignments effects on
unloaded contact patterns are illustrated next. Figure 11(a) shows unloaded contact
contact pattern to toe as shown in Figure 11(b) while misalignments E 0.15 mm,
11(c). Similarly, Figure 10(d) shows another contact pattern near toe for the gear pair
Figures 11(b) and 11(d) indicates that similar shifts in the contact patterns can be caused
62
Tip
M1 CPP
LPP
RPP
M2 Contact lines
Toe
Root
Heel
Figure 2.10: Unloaded contact pattern of the example gear pair for three adjacent tooth
pairs i 1 , i and i 1 (i-1), (i) and (i+1) with E 0.15 mm, P 0.12 mm, G 0
and 0 .
63
(a)
(b)
(c)
(d)
Figure 2.11: Unloaded contact pattern of the example gear pair (a) at nominal position
with E P G 0 , (b) at toe with E 0.08 mm, P 0.10 mm and
G 0 , (c) at heel with E 0.15 mm, P 0.10 mm and G 0 and
(d) at toe with E 0.05 mm, P G 0 , and 4 min.
64
References for Chapter 2
[2.1] Baxter, M. L., and Spear, G. M., 1961, "Effects of Misalignment on Tooth Action
of Bevel and Hypoid Gears." ASME Design Conference, Detroit, MI.
[2.2] Krenzer, T. J., 1981, Understanding Tooth Contact Analysis, The Gleason Works.
[2.3] Smith, R. E., 1984, "What Single Flank Measurement Can Do For You." AGMA,
Technical Paper No. 84FTM2.
[2.4] Smith, R. E., 1987, "The Relationship of Measured Gear Noise to Measured Gear
Transmission Errors." AGMA, Technical Paper No. 87FTM6.
[2.5] Stadtfeld, H., J., 1993, Handbook of Bevel and Hypoid Gears, Rochester Institute
of Technology.
[2.6] Zhang, Y., and Wu, Z., 2007, "Geometry of Tooth Profile and Fillet of Face-
Hobbed Spiral Bevel Gears." IDETC/CIE 2007, Las Vegas, Nevada, USA.
[2.7] Dooner, D. B., and Seireg, A., 1995, The Kinematic Geometry of Gearing: A
Concurrent Engineering Approach, John Wiley & Sons Inc.
[2.8] Fan, Q., 2007, "Enhanced Algorithms of Contact Simulation for Hypoid Gear
Drives Produced by Face-Milling and Face-Hobbing Processes." ASME J. Mech.
Des., 129(1), pp. 31-37.
[2.9] Vecchiato, D., 2005, "Design and Simulation of Face-Hobbed Gears and Tooth
Contact Analysis by Boundary Element Method," Ph.D. Dissertation, University
of Illinois at Chicago.
65
[2.10] Wu, D., and Luo, J., 1992, A Geometric Theory of Conjugate Tooth Surfaces,
World Scientific, River Edge, NJ.
[2.11] Litvin, F. L., and Fuentes, A., 2004, Gear Geometry and Applied Theory (2nd
ed.), Cambridge University Press, Cambridge.
[2.12] Wang, X. C., and Ghosh, S. K., 1994, Advanced Theories of Hypoid Gears,
Elsevier Science B. V.
[2.13] Krenzer, T. J., 1981, "Tooth Contact Analysis of Spiral Bevel and Hypoid Gears
under Load." Earthmoving Industry Conference, Peoria, IL.
[2.14] Weber, C., 1949, "The Deformation of Loaded Gears and the Effect on Their
Load Carrying Capacity (Part I)." D.S.I.R., London.
[2.15] Krenzer, T. J., 1965, TCA Formulas and Calculation procedures, The Gleason
Works.
[2.16] Shtipelman, B. A., 1979, Design and manufacture of hypoid gears, John Wiley &
Sons, Inc.
[2.19] Fan, Q., and Wilcox, L., 2005, "New Developments in Tooth Contact Analysis
(TCA) and Loaded TCA for Spiral Bevel and Hypoid Gear Drives." AGMA,
Technical Paper No. 05FTM08.
66
[2.20] Simon, V., 1996, "Tooth Contact Analysis of Mismatched Hypoid Gears." ASME
International Power Transmission and Gearing Conference ASME, 88, pp. 789-
798.
[2.21] Simon, V., 2000, "Load Distribution in Hypoid Gears." ASME J. Mech. Des.,
122(44), pp. 529-535.
67
CHAPTER 3
3.1. Introduction
surface was developed and used to determine the unloaded contact characteristics of a
hypoid gear pair, including the unloaded single and multiple tooth contact patterns and
the unloaded transmission error. This chapter builds on this formulation to predict the
Published studies on the modeling of tooth contact of hypoid gears under loaded
conditions are quite sparse. They can basically be divided into two major groups:
Computational models that use Finite Element (FE) or Boundary Element (BE)
formulations, and analytical models. As an example of models from the first group,
68
Wilcox et al [3.1] developed a FE-based model to calculate spiral bevel and hypoid gear
tooth compliance using 3D model of tooth including base deformations, which was later
employed by Fan and Wilcox [3.2] to develop a loaded tooth contact analysis (LTCA).
Vijayakar [3.3] developed another FE based hypoid LTCA package. This model employs
a hybrid approach with FE away from the contact zone and a semi-analytical contact
formulation at the contact zone. This model is perhaps the most advanced hypoid LTCA
model available today to simulate the loaded contacts of a hypoid gear pair accurately.
The major drawback of these computational models however is that they require a
considerable amount of computation time, which makes them more of an analysis tool.
Their use for design tasks such as parameter and assembly variation sensitivity studies is
Chapter 1. All of these analytical compliance models were valid for a tooth having
constant height along its face width and either constant or linearly varying thickness
along its profile, which is not the case for hypoid gears. Vaidyanathan [3.4-3.6] proposed
an analytical compliance model for a tooth with linearly varying thickness in the profile
and lengthwise directions as well as linearly varying tooth height along the face width.
His Rayleigh-Ritz based formulation used polynomial shape functions and was applied to
both sector and shell geometries. The sector model represents straight bevel gear
69
geometry closely while the shell model can be deemed sufficiently close to that of spiral
bevel gear.
In this chapter, a Rayleigh-Ritz based shell model similar to the one proposed by
Vaidyanathan [3.4-3.6] will be applied to face-hobbed and face-milled hypoid pinion and
gear teeth to define the tooth compliances due to bending and shear effects efficiently in a
semi-analytical manner. Base rotation and contact deformation effects will also be
included in this compliance formulation. With this, loaded contact patterns and
transmission error of both face-milled and face-hobbed spiral bevel and hypoid gears will
According to methodology outlined in the flowchart of Figure 1.4, the last step
before a LTCA can be performed is determining the tooth compliances of both contacting
contact point due to a unit load applied at various points on the same tooth surface [3.5].
The compliance of a gear tooth must include tooth bending deflections, shear and
Hertzian deformations as well as the base rotation, since each might contribute to tooth
70
model. The model considers a hypoid gear tooth as part of a shell with a linearly varying
thickness and height, as parameterized in Figure 3.1. The tooth thicknesses ttoe , theel
and ttip at the toe-root, heel-root and tooth tip locations are calculated and a linear, two-
hheel htoe
h htoe [ ] rc . (3.1b)
f
Here, f is the face width, rc is the cutter radius, htoe and heel hheel are tooth height
values at the toe and the heel of the tooth, is the angle between any contact points on
the tooth to toe measured from cutter center, and x is the tooth height at that specific
71
Heel (free)
Tip (free)
ttip hheel
htoe
x
theel
6 Root (clamped)
ttoe
Figure 3.1: Basic dimensions of a hypoid tooth used in the compliance formulation.
72
r r r r
iu , cos 1 . (3.2b,c)
sin r r
[3.5]
Um 1 3 U k g m
m , (3.3a)
m g m 2 g m k 1
g k k
1 U U
mn gm ( m ) gn ( n ) (3.3b)
g m g n m gm n g n
where
g3 =1 (3.3e)
In cylindrical coordinates, normal and shear strain relations are derived as [3.5]:
73
1 1 U V A W
, (3.4a)
z A AB R
(1 )
R
1 1 V U B W
, (3.4b)
z
(1 ) B AB R
R
W
z , (3.4c)
z
z z
A(1 ) B(1 )
R U R V
[ ] [ ], (3.4d)
z z z z
B(1 ) A(1 ) A(1 ) B (1 )
R R R R
1 W z U
z A(1 ) [ ], (3.4e)
z R z z
A(1 ) A(1 )
R R
1W z V
z B(1 ) [ ]. (3.4f)
z R z B(1 z )
B(1 )
R R
as the independent curvilinear coordinates along and perpendicular to tooth root line,
74
theory [3.7] that assumes a constant shear strain throughout the thickness [3.5], Eq. (3.3)
is reduces to:
x 2W
x z 2 , (3.5a)
x x
z 1 2W W
, (3.5a)
(rc z ) (rc z )
rc 2
z
z (1 ) , (3.5d)
rc
zx x . (3.5e)
setting the first variation of the potential energy to zero by using the Rayleigh-Ritz
method. The potential energy (PE) for a conservative system is the difference between
the strain energy SE and the work done by the external force WF, i.e. PE SE WF .
Here, SE and WF due to an external force p of a deformed shell surface are [3.5]
E
(rc z ) ( x
2
SE 2 2 x )
2
2(1 ) x z (3.6a)
1 (1 )( x2 xz 2 z 2 ) dzd dx,
2
75
WF pWrc dxd . (3.6b)
x
where E and are the modulus of elasticity and the Poisson’s ratio for the gear material.
Setting the first variation of the potential energy to zero one obtains
E
PE (rc z ) ( x x ( x x )
(1 2 ) x z
(3.7)
12 (1 )( x x xz xz z z ) dzd dx pWrc dxd 0.
x
The transverse deflection and shear rotations are written as linear combinations of
all of which must satisfy the following boundary conditions of a shell-shaped tooth
W ( x, )
W ( x, ) x 0 0 , , (3.9a,b)
x x 0
76
x ( x, ) x 0 0 , ( x, ) x 0 0 . (3.9c,d)
The polynomial functions m ( x) and n () must satisfy all essential boundary
(3.8) is substituted in Eq. (3.5) that is needed to evaluate normal and shear strains defined
where the sub-matrices K ms ns (ms , ns [1,3]) are determined based on volume integral
of material properties and assumed trial functions, and F ms (ms [1,3]) forms the force
vector that depends on type (point, line, etc) and location of the applied load.
Computation of the tooth compliance is done by solving Eq. (3.10) numerically for
coefficients Amn , Bmn and Cmn . Polynomial functions for pinned-free and clamped-free
respectively, where a is the tooth height. The free-free condition for n () is
77
circular segment of the shell of a length equal to the face width of the gear [3.5], which is
approximately equal to the ratio of the tooth length to the cutter radius.
While the task of computing unknown coefficients of shape functions is time consuming,
it is done only once and is valid for all mesh positions. Suppose that total contact
lines/curves as is shown (for one contact line) in Figure 3.3 are divided into N c segments
(and each segment has its own local load, which is yet to be computed), then the total
w11 . wNc 1
C . . . (3.11)
w . wNc Nc
1N c
where wis js (is , js [1, N c ]) is the deflection at segment is due to the load applied at
segment js. With this, the total deflection at segment is due to all of the applied discrete
Closed-form formula of Weber [3.8] was used here for computing the Hertzian
deformations while the base rotation and base translation effects on total tooth
78
Choose number of mode shapes Blank dimensions and machine
IFF: Free-Free settings
ICF: Clamped-Free
79
F1
F2 F3
.
1
2
3 FN 1
c
FN
c
.
Nc 1
Nc
80
In order to validate the proposed tooth compliance computation procedure, a tooth
finite elements package (ANSYS). As an example, a 500 N load was applied in the
middle of the lengthwise direction and the middle of the profile, and the tooth deflections
predicted by ANSYS and the proposed semi-analytical shell model along the middle of
the top-land of the tooth were compared as shown in Figure 3.4. It is observed in this
figure that increasing the number of mode shapes in Eq. (3.8) (IFF mode shapes for the
free-free boundary conditions and ICF mode shapes for the clamped-free boundary
those from ANSYS. However, it also increases the computational time require as shown
shell model with ANSYS for various loading conditions and number of mode shapes and
The number of tooth pairs in contact depends on the gear contact ratio, roll angle
of the pinion (or gear) and amount of applied torque. Under unloaded conditions, a
hypoid gear pair having a contact ratio greater than one has always at least one tooth pair
in contact. Once the load is applied, this number increases due to the deflection of the
contacting teeth. In the loaded tooth contact model, all the tooth pairs that are likely to
81
3
FEM Shell model
ICF 5, IFF 3, t 8 s
2.5
ICF 5, IFF 5, t 15 s
Deflection of free edge (micron)
2 ICF 5, IFF 9, t 50 s
1
82
0.5
-0.5
0 5 10 15 20 25 30 35 40 45
separation distances. Potential contact lines/curves of all contacting tooth pairs are
computed and discretized into a finite number of segments. The length, separation,
surface curvatures of both members along each line segment are computed and used as
the load distribution model [3.10]. According to the compatibility condition, in order for
the contact to occur along each of the contact lines/curves, the sum of total elastic
deformation of two contacting teeth C F and the initial separation vector S must be
C F S ΘR g (3.12)
where F is force vector, C is the total compliance matrix that is the sum of the pinion
and gear tooth compliance matrices C p and C g (Eq. (3.11)), and the Hertzian
compliance matrix Ch , and R g is the vector that contains the distances of each segment
to the gear axis. Eq. (3.11) can be written in form of an equality constraint by
C F ΘR g Y = S . (3.13)
83
Since two bodies must be at contact for any force on a given segment is to exist, either
Meanwhile, the equilibrium condition assures that the total moment caused by
forces acting on all contacting segments about the gear axis as shown in Figure 3.5 must
FT R g T g (3.14)
where superscript T denotes matrix transpose. The load distribution and loaded
simultaneously.
The same face-milled example hypoid gear pair used for the UTCA whose basic
parameters are listed in Table 2.1 for the drive-side contact (concave side of pinion and
convex side of gear) is considered to demonstrate the capabilities of the proposed hypoid
gear load distribution model. This is a FM gear set representative of an automotive rear
84
Tg
ag
F3
F2
R3g F1
R2g
R1g
Figure 3.5: Static equilibrium between torque applied on gear axis and torque produced
by the force of all contacting segments.
85
Loaded transmission error (LTE) of the example gear pair predicted by the
proposed model are shown in Figure 3.6 at three different pinion torque values of
T p 50, 250 and 500 Nm for a set of fixed misalignment values of E 0.15
mm, P 0.12 mm, G 0 mm and 0 . These LTE time histories indicate that the
shape and the average value of LTE change with T p , as expected. Table 3.1 lists the
peak-to-peak value (p-p), first three Fourier harmonics (1st, 2nd, 3rd) and the root-mean-
square (RMS) value of LTE corresponding to the cases of Figure 3.6. The LTE functions
in each T p level are dictated primarily by the first harmonic order. Modest increases in
p-p, 1st harmonic and the root-mean-square values of LTE are observed with increasing
T p.
Figures 3.7(a-c) show the pressure distributions predicted by the proposed model
for the same cases of Figure 3.6. It is seen in Figure 3.7(a) that the contact is localized at
the center of the tooth when T p is low (50 Nm) with no edge loading. An increase in
T p causes the contact pattern to spread, in the process exhibiting edge loading at the tip
and root regions as it is evident from Figures 3.7(b) and (c). Figures 3.7(a-c) also show
the loaded contact patterns (maximum contact pressure distributions) predicted by a FE-
based hypoid contact model [3.11] for the same cases are in good agreement. It is
worthwhile to mention here that each simulation with the proposed model required 45
seconds of CPU time (about 25 seconds for compliance matrix computations and 1
86
(a) T p = 50 Nm
130
120
110
100
90
(b) T p = 250 Nm
310
300
TE 290
[rad]
280
270
(c) T p = 500 Nm
620
610
600
590
580
0.0 0.5 1.0 1.5 2.0
Mesh cycles
Figure 3.6: Loaded gear transmission error of the example gear pair with
E 0.15 mm, P 0.12 mm, G 0 and 0 at (a) T p 50 Nm, (b) T p 250
Nm, and (c) T p 500 Nm.
87
_________________________________________________________________
Table 3.1: The loaded transmission error predictions of the proposed model; G 0
mm and 0 for all cases.
88
Proposed Model (a) T p = 50 Nm , E = 0.15 mm , P = 0.12 mm FE Model [3.11]
Continued
Figure 3.7: Comparison of loaded contact patterns predicted by the proposed model to an FE model [3.11] for (a)
T 50 Nm , E 0.15 mm, P 0.12 mm, (b) T p 250 Nm , E 0.15 mm, P 0.12 mm, (c) T p 500 Nm ,
p
E 0.15 mm, P 0.12 mm, (d) T p 50 Nm , E 0.08 mm, P 0.05 mm, and (e) T p 50 Nm ,
E 0.26 mm, P 0.13 mm ( all at G 0, 0 ).
Continued
Proposed Model (d) T p = 50 Nm , E = 0.08 mm , P = 0.05 mm FE Model [3.11]
90
highlights the main advantage of this proposed model as a design tool, even if it might
Next, the same gear pair is simulated by using the proposed model and the FE
model [3.11] at T p 50 Nm for two other misalignment conditions. Here, two of the
errors are kept constant at G 0 mm and 0 , and the other two errors E and
cause the predicted loaded contact pattern to move towards toe and root, compared to
Figure 3.7(a). Meanwhile, the loaded contact for E 0.26 mm and P 0.13 mm
moves the contact in the opposite direction towards the heel. In the process, the
maximum contact pressure is reduced since there is larger area in the heel that carries the
same load. In addition, equivalent radii of curvature are larger at heel than toe, which
directly decreases maximum Hertzian pressure from Weber equation [3.8]. The FE
simulations of the same error combinations shown in the same figures are again in good
agreement with the predictions of the proposed model. This suggests that the sensitivity
of the hypoid gear contact to gear errors is captured sufficiently by this model. Finally
the LTE parameters listed in Table 3.1 for these two cases reveal slight reduction in LTE
amplitudes compared to the first case of Figure 3.7(a), suggesting that a good contact
91
References for Chapter 3
[3.1] Wilcox, L. E., Chimner, T. D., and Nowell, G. C., 1997, "Improved Finite
Element Model for Calculating Stresses in Bevel and Hypoid Gear Teeth."
AGMA, Technical Paper No. 97FTM05.
[3.2] Fan, Q., and Wilcox, L., 2005, "New Developments in Tooth Contact Analysis
(TCA) and Loaded TCA for Spiral Bevel and Hypoid Gear Drives." AGMA,
Technical Paper No. 05FTM08.
[3.3] Vijayakar, S. M., 1991, "A Combined Surface Integral and Finite Element
Solution for a Three-Dimensional Contact Problem." International J. for
Numerical Methods in Engineering, 31, pp. 525-545.
[3.4] Vaidyanathan, S., 1993, "Application of Plate and Shell Models in the Loaded
Tooth Contact Analysis of Bevel and Hypoid Gears," Ph.D. Dissertation, The
Ohio State University, Columbus, Ohio.
[3.5] Vaidyanathan, S., Houser, D. R., and Busby, H. R., 1993, "A Rayleigh-Ritz
Approach to Determine Compliance and Root Stresses in Spiral Bevel Gears
Using Shell Theory." AGMA, Technical Paper No. 93FTM03.
[3.6] Vaidyanathan, S., Houser, D. R., and Busby, H. R., 1994, "A Numerical
Approach to the Static Analysis of an Annular Sector Mindlin Plate with
Applications to Bevel Gear Design." J. of Computers & Structures, 51(3), pp.
255-266.
[3.7] Mindlin, R. D., 1951, "Influence of Rotary Inertia and Shear on Flexural Motions
of Isotropic Elastic Plates. ." J. of Applied Mechanics, 18, pp. 31-38.
92
[3.8] Weber, C., 1949, "The Deformation of Loaded Gears and the Effect on Their
Load Carrying Capacity (Part I)." D.S.I.R., London.
[3.9] Stegemiller, M. E., and Houser, D. R., 1993, "A Three Dimensional Analysis of
the Base Flexibility of Gear Teeth." ASME J. Mech. Des., 115(1), pp. 186-192.
[3.10] Conry, T. F., and Seireg, A., 1972, "A Mathematical Programming Technique for
the Evaluation of Load Distribution and Optimal Modification for Gear Systems."
ASME J. of Industrial Engineering.
[3.11] Vijayakar, S., 2004, Calyx Hypoid Gear Model, User Manual, Advanced
Numerical Solution Inc., Hilliard, Ohio.
93
CHAPTER 4
4.1. Introduction
axles are subject to various manufacturing errors and heat treatment distortions that
deviate the actual (real) tooth contact surfaces from the intended (theoretical) ones. Such
errors impact the quality of a hypoid gear pair, defined by a number of performance
indicators including its contact pattern, the motion transmission error (TE), efficiency as
manufacturing errors typically follow patterns that shift, rotate or twist the surfaces
relative to the theoretical ones. Therefore, they can be characterized as global deviations.
Other more local deviations occur during the life span of hypoid gears in the form of
94
surface wear. Since the surface wear depths are proportional to the contact pressure and
the sliding distance, deviations due to surface wear are rather local and cannot be
captured by using the surface fitting methods developed to approximate the global
methodology that allows loaded and unloaded tooth contact analysis of hypoid gears
Tooth contact analysis has usually been performed by considering the theoretical
pinion and gear surfaces defined by simulation of the hypoid cutting processes. The
analysis results presented in Chapters 2 and 3 also considered such theoretical tooth
surface errors with no deviations.. There are only a few published studies on hypoid
gear tooth contact analysis using the real surfaces. In such an analysis, Gosselin [4.1]
proposed an approach to compute tooth contact of real spiral bevel gear surfaces. He
interpolated measured surfaces with rational functions to predict their unloaded contact
pattern and transmission error. Since pinion and gear normal vectors of low-mismatch
(high-conformity) surfaces, a wide span of potential contact line around contact point can
be identified and the condition of collinearity for normal vectors is subjects to numerical
stability issues. In order to simplify the task of locating the contact point, Gosselin [4.1]
computed the difference between pinion and gear normal vectors at several points along
the lengthwise direction and estimated a location where the difference between the
95
Zhang et al [4.2] proposed an approach to analyze unloaded tooth contact of real
hypoid gears based on a generalization of the work of Kin [4.3, 4.4] on spur gears. The
real pinion and gear tooth surfaces were divided into two vectorial surfaces of theoretical
and deviation surfaces. Separating theoretical and deviation surfaces, finding the
theoretical surfaces through cutting simulation, and applying interpolation only to the
deviation surfaces made his approach simpler and more accurate. Zhang [4.2] defined
the deviation surface in the normal direction of the theoretical surface by comparing the
theoretical and the real (measured) surfaces, and fit a bicubic surface to it. The normal to
the real surfaces (sum of the theoretical and interpolated deviation surfaces) were
computed by taking the derivative of the real surfaces. Having continuous functions for
the surfaces and normal of the pinion and the gear, he employed conventional system of
five nonlinear equations and five unknowns used in many other studies [4.5-4.7] to
changes to the machine settings that define the theoretical surfaces so that real gear
surfaces can be computed approximately from the cutting simulation [4.8, 4.9]. This
method found machine settings that generate a theoretical surface close to (but not
identical to) the real surface and the difference of the two surfaces was defined as
“residual error surfaces” for the pinion and the gear. He computed transmission error and
the contact pattern of the generated surfaces with this new set of machine settings. Then,
he used the residual error surface to modify predicted transmission error and contact
pattern without providing the details of this process [4.10]. This method is suitable in
96
capturing the effect of global errors such as the pressure angle error, the spiral angle
error, and lengthwise or profile crowning errors for the cases when residual errors are
rather small, while the same cannot be said for localized deviations such as surface wear
In Chapter 2, it was mentioned that the unloaded tooth contact analysis (UTCA)
of hypoid gears with mismatched surfaces were performed using two fundamentally
different methods. The first method that was used widely defines the tooth surfaces as
two arbitrary surfaces, rotating about pinion and gear axes [4.7, 4.11-4.13]. In this
method, the contact point path (CPP) on each surface was determined by satisfying two
contact conditions: (i) coincidence of position vector tips of the points on the gear and
pinion surfaces and (ii) collinearity of the normals of the both of the surfaces. The
second method that was proposed in Chapter 2 was based on the ease-off topography. A
contact curves from surface of roll angle was provided. In Chapter 3, UTCA results were
combined with a semi-analytical compliance model based on shell theory to predict the
loaded contact patters and loaded TE. Using the ease-off approach for TCA of real gear
efficiency of TCA since surface interpolations for measured pinion and gear surfaces as
well as the solution of the system of five governing nonlinear equations are not needed.
In this chapter, ease-off is defined the same way as Chapter 2 as the deviation of
real gear surface from the conjugate of its real mating pinion surface. The formulation
97
that will be proposed to handle local and global deviations is based on the premise that all
surface deviations, both local and global, can be handled through modifications of the
ease-off topography. Having machine settings and blank dimensions, surface coordinates
and normal vectors of both pinion and gear will be defined using the methodology
proposed in Chapter 2. These theoretical surfaces will be used to establish the theoretical
ease-off topography and a theoretical surface of roll angle. As the main contribution of
this study, a procedure will be proposed to update the ease-off topography by taking into
account pinion and gear surface deviations. The updated ease-off and roll angle surfaces
will be used to determine the unladed tooth contact characteristics of the gear pair. These
UTCA results will be combined with the semi-analytical LTCA methodology of Chapter
3 to predict the loaded contact patterns and the transmission error of hypoid gears having
The first step in defining the theoretical ease-off surface is specifying an area in
the gear projection plane with the possibility of contact between pinion and gear, as
illustrated in Figure 4.1. This area represents the projection of a volume bounded by the
faces and front and back cones of the pinion and gear into the gear projection plane. The
projection plane shown in Figure 4.1 is a gear-based projection plane since its ease-off
is defined on the gear tooth area. Using the gear machine settings and applying the
equation of meshing, the theoretical gear surface coordinates of a point of the projection
98
rijg
Gear projection
plane
ij
Q ap
ag
rija
rijp
rˆijg
99
plane shown in Figure 4.1 are computed as rijg where i [1, m] and j [1, n] are the
indices in lengthwise and profile directions of the surface point with m and n as number
of surface grids in respective directions, and the superscript g denotes gear surface.
Then, rijg is transformed to the pinion coordinate system to construct the projection plane
Next, the theoretical pinion surface point rijp and its unit normal to the surface
nijp are computed again through machine settings and applying the equation of meshing
[4.1]. Both theoretical pinion and gear surface coordinates and unit normal vectors are
transformed into the global coordinate system. Having rijp , nijp , the pinion and gear
axis vectors a p and a g , and the gear ratio R, the position vector rija of the corresponding
point on the action surface is found from the conjugacy equation in 3D space as [4.14]:
The same procedure is repeated for all points on the real pinion surface to define the
action surface completely. As observed from Figure 4.1, the action surface for a hypoid
gear pair, while quite flat, is not a plane. Here, any point and its unit normal vector on
the theoretical pinion surface are rotated around the pinion axis in order to satisfy Eq.
(4.1). The angle of rotation qijp of each point ij required to satisfy Eq. (4.1) is then
100
plotted on the gear projection plane to construct the theoretical pinion roll angle surface
Q.
Here, only the relative values of qijp (i.e. the shape of the Q surface) are of
the Q surface is shifted in the direction normal to the projection plane to bring it to
contact with the projection plane such that at least one grid point has zero qijp value.
Mathematically, this shift is equivalent to rigidly rotating pinion tooth surface around the
pinion axis, which has no effect on the pinion surface. The angle qˆijg = qijp R
corresponds to the amount of rotation required to travel from the surface of action to the
conjugate of theoretical pinion surface. Therefore, the position vector of the conjugate
cos qˆ g g
sin qˆij 0 0
ij
g g
0 0 .
M z (qˆijg ) sin qˆij cos qˆij
(4.2b)
0 0 1 0
0 0
0 1
101
If this conjugate surface of the pinion were to match perfectly with the real gear
surface at any point, then a perfect meshing condition with zero unloaded transmission
error would exist. For simplicity, the conjugate of pinion will be called here the
conjugate gear. The difference between the conjugate gear and the theoretical gear is
defined as theoretical ease-off topography . In general, the conjugate gear and the
theoretical gear are located at different angular positions with respect to the gear axis. In
order to compare these two surfaces, conjugate gear surface is rotated around gear axis
a g by an angle such that a grid point of conjugate gear surface touches the
corresponding grid point on the theoretical gear surface. In this position, the radial
distances ij between the grid points ij on these surfaces define the theoretical ease-off
surface . The and Q surfaces were used in Chapters 2 and 3 for both unloaded and
4.3. Updating Ease-off Topography for Manufacturing Errors and Surface Wear
Deviations of the pinion and gear surfaces a grid point ij from their respective
theoretical surfaces are defined as ijp and ijg ( i [1, m] , j [1, n] ) as shown in Figure
4.2. Normal vectors to both pinion and gear surfaces are considered in inward direction.
Measured and/or worn pinion and gear surfaces are written as ( i [1, m] , j [1, n] ):
Here, it is assumed that the normal vectors of the real and theoretical surfaces are the
same, since both surfaces are practically very close to each other [4.4].
Any point of the theoretical ease-off surface ij can be updated by using
deviations on pinion and gear surfaces. The goal here is to update theoretical ease-off
surface directly from surface deviations rather that updating original pinion and gear
surfaces and conducting whole tooth contact procedure between new surfaces as it has
Assume that an ease-off value ij of a point on the theoretical ease-off surface
T T
rijp xijp , yijp , zijp , nijp and rijg xijg , yijg , zijg . At the same grid point ij, the
pinion roll angle is qijp , and hence, the corresponding roll angle of the gear surface
(conjugate to the theoretical pinion surface) is qijg qijp R and the distance of the same
grid points on the pinion and gear surfaces to their own rotation axes ( a p and a g
respectively) are
103
Lijg ( xijg )2 ( yijg ) 2 (4.4b)
as shown in Figure 4.2. With this, the changes in the pinion and gear roll angles due to
the pinion and gear surface errors ijp and ijg are defined as
Here uijp and uijg are the unit normal vectors in the radial (circular) direction of the
pinion and gear axes a p and a g , respectively, as shown in Figure 4.2. They are
defined as:
a p rijp a g rijg
uijp , uijg (4.5b,c)
a p
rijp a g
rijg
Hence, the updated pinion roll angle taking the pinion deviation into account is
which can be used to find locations and directions of contact curves corresponding to the
104
Pinion
projection plane
p
nijp uij
ijp
Lijp
Gear projection
plane
ijg
nijg
uijg rijp
rijg
Q
ap
Lijg
ag
Figure 4.2: Graphical demonstration of the procedure to update ease-off surface for
surface deviations.
105
Change to the theoretical ease-off surface can be described in two components.
The first component is related to the pinion surface deviation that is formulated as
where qijgp = qijp R is the gear roll angle change due to the pinion surface deviation.
The second component is due to the gear surface deviation that is given as
Therefore, the total change to the ease-off topography is the sum of its two components
This updated ease-off surface and the corresponding updated surface of roll
angle Q as defined by Eq. (4.6) are used for unloaded and loaded tooth contact analyses
surface of roll angle Q is also shifted to touch the projection plane since the absolute
values qijp do not have any effect on unloaded and loaded tooth contact analyses.
106
4.4. Unloaded and Loaded Tooth Contact Analyses
plane to provide an insight into how UTCA is conducted. For any value q on surface Q ,
a corresponding instantaneous contact curve can be defined. As shown in Figure 4.3 for
a specific pinion roll angle qk , the intersection of the plane z qk and the surface Q
defines the x and y coordinates of all points on the projection plane that have same roll
angle, stating theoretically that they lie on the same contact curve. Since Q is not a
plane, this intersection for hypoid gears is usually a curve rather than a straight line as
Figure 4.4 shows the theoretical contact curves of the drive and coast sides of a
sample hypoid gear pair on the projection plane for different pinion angles as shown on
the contact curves. The instantaneous contact curve C (qk ) shown in Figure 4.3 is
obtained by first projecting this intersection curve on projection plane and then projecting
this projected curve on the ease-off surface . The minimum distance from C (qk ) to
Moreover, moving in both directions from point H along the curve C (qk ) within a preset
separation distance yields the unloaded contact line length U (qk ) . Repeating this
procedure for every pinion roll angle increment q qk ( k [1, Nl ] where Nl is total
number of contact curves considered), unloaded transmission error curve TE (qk ) and the
Contact curve
projection
z qk
q qk
U (qk )
H
x
TE (qk )
Figure 4.3: Graphical demonstration of the procedure to compute unloaded TCA; (a) gear
projection plane, ease-off and Q surfaces, and (b) instantaneous contact curve, contact
108
(a) Drive Side
70° 85°
40° 55°
Toe 25°
10°
Root
Root
109
It is noted here that the contact curves are defined between the real pinion
surface and conjugate of the real pinion surface (conjugate gear surface), instead of using
the real gear surface. Replacing the real gear surface with the conjugate one causes a
very little change to the orientation and shape of the contact curve, since the effect of
Number of tooth pairs in contact depends on the gear contact ratio, the roll angle
of the pinion (or gear) and the amount of torque applied. Under unloaded conditions, a
hypoid gear pair having a contact ratio greater than one has always at least one tooth pair
in contact. Once the load is applied, this number increases due to deflection of the
contacting teeth. In the LTCA model of Chapter 3, all the tooth pairs that are likely to
share the torque geometrically are taken into consideration with their respective
separation distances. Potential contact curves of all contacting tooth pairs are computed
and discretized into a finite number of segments ( N c ). The length of the separation at
each segment along each contact curve are computed and used as input for the LTCA
model. With the theoretical ease-off topography replaced by the modified ease-off
and 3 can be applied to predict the unloaded and loaded tooth contact conditions,
respectively.
110
4.5. Example Analyses
A face-milled hypoid gear set with local deviations whose basic parameters are
listed in Table 4.1 for the drive-side contact (concave side of pinion and convex side of
gear) will considered for an example loaded contact analysis of surfaces with local
deviations. This gear set is representative of an automotive rear axle gear set. An
example case of local deviations from theoretical surfaces is shown in Figure 4.5 for both
pinion deviation ijp and gear deviation ijg . This case represents worn tooth surfaces
predicted by a hypoid gear wear model [4.15] from a companion study. In Figure 4.5, the
“Root” line refers to the lower limit of active contact region and it is not actual gear root
line.
Following the proposed ease-off update approach, gear projection plane, the
theoretical and updated ease-off surfaces, and , and the theoretical and updated roll
angle surfaces, Q and Q , are computed. Figure 4.6(a) shows these surfaces in relation
to each other while the theoretical ease-off topography, updated ease-off and the amount
of ease-off change computed from Eq. (4.9) are shown in Fig. 4.6(b) to (d) in contour plot
format, respectively. As shown here, the maximum changes take place in the vicinity of
the locations where pinion and gear deviations are maximum, according to Figure 4.5 and
the rest of the projection plane does not exhibit any considerable ease-off change.
111
_____________________________________________________________
Number of teeth 11 41
Hand of Spiral Left Right
Mean spiral angle (deg) 40.5 28.5
Shaft angle (deg) 90
Shaft offset (mm) 20
Outer cone distance (mm) 115 111
Generation type Generate Formate
Cutting method FM
___________________________________________________________________________
Table 4.1: Basic drive side geometry and working parameters of the example hypoid gear
pair.
112
ijp
Toe
ijg Root
Root
Toe
Pinion deviation μm
10
8
Toe 6
4
2
Root 0
Gear deviation μm
5
4
Toe
3
2
1
Root 0
Figure 4.5: Example local deviation surfaces for the gear and pinion tooth surfaces.
113
Gear projection
(a) plane
Q,Q μm
(b) 60
Toe 40
20
Root 0
μm
(c) 60
Toe 40
20
Root 0
μm
(c) 6
5
Toe 4
3
2
Root 1
Figure 4.6: Ease-off update for the example deviation of Fig. 5. (a) Three-dimensional
view of the projection plane, and , , Q and Q surfaces, and contour plots of (b) ,
(c) , and (d) the change of ease-off topography.
114
The corresponding predicted unloaded tooth contact patterns are shown in Figure
4.7 for a maximum separation value of 6 μm for both cases of (a) theoretical and (b)
deviated tooth surfaces, indicating that the unloaded contact patterns are influenced by
the local deviation as well. It is clear from this figure that the length of the contact lines
are elongated for UTCA of deviated surfaces since ease-off in the contact region is
flattened.
The predicted unloaded transmission error (UTE) curves are shown in Figure
4.8(a) for the theoretical and deviated surfaces. The corresponding peak-to-peak
amplitude (p-p), first three Fourier harmonics (1st, 2nd, 3rd) and the root-mean-square
(RMS) value of these curves are listed in Table 4.2(a) to show that all components of
UTE are influenced by the local deviation introduced. Both the root-mean-square (RMS)
Next, LTCA is performed for the theoretical and the locally deviated surfaces as
before. A pinion torque of 200 Nm was applied in this analysis. The predicted loaded
transmission errors (LTE) at this torque value are shown in Figure 4.8(b) for the
theoretical and deviated surfaces. It is noted here that both curves are identical for certain
mesh positions where areas of the local deviation are not in contact while they differ
significantly in certain mesh positions. Table 4.2(b) lists the same LTE amplitudes for
theoretical and deviated surfaces to show that such local deviations also impact the LTE.
Finally, predicted contact pressure distributions are shown in Figures 4.9(a) and (b) for
the theoretical and the deviated surfaces at 200 Nm pinion torque value. Here, it is
115
(a) Theoretical surfaces
Toe
Root
Toe
Root
Figure 4.7: Predicted unloaded tooth contact pattern for separation value of 6 μm .
116
(a) Unloaded TE [µrad]
25
Theoretical surfaces
20
15
10
5
0
25
Deviated surfaces
20
15
10
5
0
0 0.5 1.0 1.5 2.0
Mesh cycles
Figure 4.8: Transmission error (UTE) curves for theoretical and deviated surfaces at (a)
unloaded conditions and (b) loaded conditions at a pinion torque of 200 Nm.
117
(a) Unloaded Transmission Error in [µrad]
__________________________________________________________
p-p 1st 2nd 3rd RMS
__________________________________________________________
Theoretical surfaces 21.1 8.3 3.7 2.0 9.3
Deviated surfaces 15.0 4.3 4.3 1.4 6.3
__________________________________________________________
Table 4.2: The transmission error amplitudes of theoretical and deviated surfaces.
118
Mpa
(a) Theoretical surfaces 1200
800
Toe
400
Root 0
Mpa
(b) Deviated surfaces 1200
800
Toe
400
Root 0
Figure 4.9: Predicted contact pressure distribution for a pinion toque of 200 Nm
for (a) theoretical and (b) deviated surfaces.
119
observed that the edge loading condition experienced by the theoretical surfaces on the
gear root (pinion tip) is reduced significantly for the surfaces with the local deviations
A face-hobbed hypoid gear set with basic parameters listed in Table 4.3 for its
drive-side contact (concave side of pinion and convex side of gear) is considered as
example for loaded contact analysis of surfaces with global deviations. This gear pair is
also representative of an automotive rear axle gear sets. The measured pinion and gear
deviation surfaces ( ijp and ijg ) from their theoretical geometry after heat treatment and
the lapping process are shown in Figures 4.10(a) and (b), along with their respective
interpolated surfaces on active surface domain shown as ijp and ijg . Figures 4.10(c) and
(d) are the contour plots of the same, showing a maximum 50 µm of error for pinion in
heel-top region and 25 µm for gear in toe-root. The source of deviation here could be
due to manufacturing errors, heat treatment distortion and surface wear caused by lapping
process. In the most common measurement procedure used by the axle manufacturers,
direction normal to the surface. In Figure 4.10, these deviations are shifted for
Table 4.3: Basic drive side geometry and working parameters of the example hypoid gear
pair.
121
(a)
Toe ijp
Root
ijp
(b)
Toe
ijg
ijg
Root
122
It should be noted here that a simple weighted average is used to interpolate (or
extrapolate) the deviations ijp to ijp and ijg to ijg at a point ij that is not one of the 45
vectors as is the case in previous studies, hence the required accuracy and complexity of
this interpolation is by no means comparable to the case that interpolation is required for
Following the proposed method described in the previous section , the theoretical
and updated ease-off surfaces are computed. Figures 4.11(a-d) respectively show (a)
pinion surface deviations, (c) updated ease-off topography by only considering gear
surface deviations and (d) updated ease-off topography by considering both pinion and
gear surface deviations. Figure 4.11(a) shows a localized well defined ease-off
Although pinion and gear deviation effects on ease-off, when applied alone, are
their adverse effect of alone, resulting in the ease-off topography shown in Figure
4.11(d).
The corresponding predicted unloaded tooth contact patterns of this design are
shown in Figure 4.12(a,b) for a maximum separation value of 6 μm for both cases of
theoretical and deviated tooth surfaces, respectively, indicating that the unloaded contact
patterns are influenced by the deviations given in Figure 4.10 as well. With the deviation
123
(a) µm
120
Toe 80
40
0
Root
(b) µm
120
Toe 80
40
0
Root
(c) µm
120
Toe 80
40
0
Root
(d) µm
120
Toe 8 80
6
4 40
2
0
Root
Figure 4.11: Ease-off update for the example deviation of Fig. 4.10. (a) Theoretical ease-
off topography, (b) updated ease-off topography only with pinion deviation, (c) updated
ease-off topography only with gear deviation, and (d) updated ease-off topography with
both pinion and gear deviations.
124
(a) Theoretical surfaces
Toe
Root
Toe
Root
Figure 4.12: Predicted unloaded tooth contact pattern for separation value of 6 μm .
125
included, unloaded contact pattern slightly shifted toward heel and becomes narrower as
The predicted unloaded transmission error (UTE) curves are shown in Figure
4.13(a) for theoretical and deviated surfaces against mesh cycles. The corresponding
peak-to-peak amplitude, first three Fourier harmonics and the RMS value of these curves
are listed in Table 4.4(a) to show that the all components of UTE are influenced by the
local deviation introduced. The RMS, peak-to-peak and 1st harmonic amplitudes are
Next, LTCA is performed for the theoretical and the globally deviated surfaces as
before. A pinion torque of 200 Nm was applied in this analysis. The predicted loaded
transmission errors (LTE) at this torque value are shown in Figure 4.13(b) for the
theoretical and globally deviated surfaces. Table 4.4(b) lists the same LTE amplitudes
for theoretical and deviated surfaces to show that such local deviations also impact the
LTE. Finally, predicted contact pressure distribution is also shown in Figure 4.14(a) and
(b) for the theoretical and the deviated surfaces at 200 Nm pinion torque value. Here, it
is observed that the contact pressure distributions for the theoretical and deviated surfaces
are rather close since pinion and gear surface deviations tend to compensate each other
126
(a) Unloaded TE [µrad]
Theoretical surfaces
60
40
20
0
Deviated surfaces
60
40
20
0
0 0.5 1.0 1.5 2.0
325
315
305
Deviated surfaces
370
360
350
340
0 0.5 1.0 1.5 2.0
Mesh cycles
Figure 4.13: Transmission error curves for theoretical and deviated surfaces; (a) unloaded
conditions and (b) loaded conditions at a pinion torque of 200 Nm.
127
(a) Unloaded Transmission Error [µrad]
__________________________________________________________
p-p 1st 2nd 3rd RMS
__________________________________________________________
Theoretical surfaces 36.9 15.1 4.0 1.8 15.7
Deviated surfaces 61.5 28.9 2.1 1.5 29.0
__________________________________________________________
_______________________________________________________
Table 4.4: The transmission error amplitudes of theoretical and deviated surfaces.
128
(a) Theoretical surfaces Mpa
800
600
Toe 400
200
0
Root
Figure 4.14: Predicted contact pressure distribution for a pinion toque of 200 Nm for (a)
theoretical and (b) deviated surfaces.
129
4.6. Summary
was proposed to perform loaded and unloaded tooth contact analysis of spiral bevel and
hypoid gears having both types of local and global deviations. Manufacturing errors
causing global errors and localized surface deviations were considered to update the
theoretical ease-off to form a new ease-off surface that was used to perform a loaded
tooth contact analysis. Two numerical examples of (i) face-milled hypoid gear set with
local deviations and (ii) face-hobbed hypoid gear set with global deviations measured by
well as quantifying the effect of such deviations on load distribution and the unloaded
[4.1] Gosselin, C., et al. 1991, "Tooth Contact Analysis of High Conformity Spiral
Bevel Gears." Proceedings of JSME Int. Conf. on Motion and Power
Transmission, Hiroshima, Japan.
[4.2] Zhang, Y., Litvin, F. L., Maryuama, N., Takeda, R., and Sugimoto, M., 1994,
"Computerized Analysis of Meshing and Contact of Gear Real Tooth Surfaces."
116, pp. 677-682.
[4.3] Kin, V., 1992, "Tooth Contact Analysis Based on Inspection." Proceedings of 3rd
World Congress on Gearing, Paris, France.
130
[4.4] Kin, V., 1992, "Computerized Analysis of Gear Meshing Based on Coordinate
Measurement Data." ASME Int. Power Transmission and Gearing Conference,
Scottsdale, AZ.
[4.6] Litvin, F. L., and Fuentes, A., 2004, Gear Geometry and Applied Theory (2nd
ed.), Cambridge University Press, Cambridge.
[4.7] Fan, Q., 2007, "Enhanced Algorithms of Contact Simulation for Hypoid Gear
Drives Produced by Face-Milling and Face-Hobbing Processes." ASME J. Mech.
Des., 129(1), pp. 31-37.
[4.8] Gosselin, C., Nonaka, T., Shiono, Y., Kubo, A., and Tatsuno, T., 1998,
"Identification of the Machine Settings of Real Hypoid Gear Tooth Surfaces."
ASME J. Mech. Des., 120(3), pp. 429-440.
[4.9] Gosselin, C., Jiang, Q., Jenski, K., and Masseth, J., 2005, "Hypoid Gear Lapping
Wear Coefficient and Simulation." AGMA, Technical Paper No. 05FTM09.
[4.10] Gosselin, C., Guertin, T., Remond, D., and Jean, Y., 2000, "Simulation and
Experimental Measurement of the Transmission Error of Real Hypoid Gears
Under Load." ASME J. Mech. Des., 122(1), pp. 109-122.
[4.11] Wang, X. C., and Ghosh, S. K., 1994, Advanced Theories of Hypoid Gears,
Elsevier Science B. V.
[4.12] Litvin, F. L., and Gutman, Y., 1981, "Methods of synthesis and analysis for
hypoid gear-drives of formate and helixform; Part III-Analysis and optimal
synthesis methods for mismatched gearing and its application for hypoid gears of
formate and helixform." ASME J. Mech. Des., 103, pp. 102-113.
131
[4.13] Simon, V., 1996, "Tooth Contact Analysis of Mismatched Hypoid Gears." ASME
International Power Transmission and Gearing Conference ASME, San Diego.
[4.14] Stadtfeld, H., J., 1993, Handbook of Bevel and Hypoid Gears, Rochester Institute
of Technology.
[4.15] Park, D., and Kahraman, A., 2008, "A Surface Wear Model for Hypoid Gear
Pairs." In press, Wear, pp.
132
CHAPTER 5
5.1. Introduction
critical to power train designers. This is mainly because the government regulations in
regards to fuel economy and carbon emissions are becoming more stringent. Forecasted
increases in oil prices also add to the motivation to predict and reduce power losses of
drive trains. In rear-wheel drive vehicles, the rear axle-differential unit is one of the
major sources of power losses. The axle efficiency values can be typically as low as 90
to 95% [5.1]. Considering that rear-wheel drive vehicles comprise a significant share of
the global passenger vehicle market, any sizable improvements to the axle efficiency can
Axle power losses can be divided into two groups. One group of losses is
independent of the torque transmitted. These load independent (spin) power losses are
133
due to viscous bearing losses (including the losses due to pre-load) and gear windage and
oil churning losses [5.2, 5.3]. Such losses are outside the scope of this research. The
other group of losses are induced by friction at bearing and hypoid gear pair locations
under load. These power losses are called load-dependent (mechanical) losses. Focusing
on the hypoid gear pair, mechanical power losses are associated with the relative sliding
and the lubricated contact conditions along the tooth contacts. The shaft off-set, being
the main difference between spiral bevel and hypoid gears, causes increased relative
sliding in hypoid gears and the power losses associated with friction [5.4]. The
geometry parameters and velocities, and the normal load at each contact point from a
tooth contact analysis model as the one proposed in Chapters 2 and 3, (ii) a friction model
to determine the coefficient of friction at each contact points (iii) computing the surface
traction from the distributions of the friction coefficient and normal force, and (iv)
determining the friction torque and the resultant power loss. Published gear efficiency
models differ mostly in the way they determine the friction coefficient. The first group of
models used a constant friction coefficient μ [5.5-5.7] in computing the power losses.
Recognizing the fact that μ is dependent on various contact parameters, including rolling
velocity, sliding to roll ratio, radii of curvature of the contacting surfaces and normal
134
load, all of which vary as gears roll, experiment based µ empirical formulae [5.8-5.11]
were employed by another group of studies [5.8, 5.12-5.16]. However, the applicability
of these models was limited to narrow ranges of the operating temperatures, speed, load,
and surface roughness conditions represented by the empirical formula. The third group
(EHL) theory [5.17-5.22]. This approach, while physics-based and potentially more
are required to predict the mechanical losses of a gear pair. In order to avoid this
formula up-front by using the EHL model of Cioc et al [5.24]. Using this EHL model,
they conducted a large parameter study, covering wide ranges of contact and surface
traction data was reduced into a single formula by using linear regression technique.
All of the models cited above were limited to spur or helical gears. Efficiency
models for hypoid gears are very sparse. Approximating the hypoid gear power loss as
the sum of losses from the corresponding spiral bevel and worm gears, Buckingham
[5.25] recommended a power loss equation. Coleman [5.1] proposed a simple closed-
form formula to estimate bevel and hypoid gears efficiency. This heuristic formula used
a constant friction coefficient of 0.05 at every contact point and was a function of the
135
normal load, pressure angle, and pinion and gear mean spiral angles. Simon [5.26]
The model proposed recently by Xu and Kahraman [5.27] extended their helical
gear efficiency model to hypoid gears. They used a commercially available FE-based
hypoid gear contact model CALYX [5.28] to determine all required contact load and
they computed sliding and rolling velocities at each contact point along and perpendicular
to the contact line. While this model [5.30] was physics-based and included most of the
key surface, lubricant, geometry and operating parameters, its FE load distribution
and parameter sensitivity studies. It relied on the same FE model for its geometry and
curvature information as well. More importantly, the EHL model [5.24] it employed to
derive the friction coefficient formula was not designed for simulating mixed type of
lubrication condition. Therefore, the fidelity of the model Xu and Kahraman [5.30] was
Kahraman developed transient mixed or boundary EHL models for line [5.31] and point
[5.32] contacts that can handle any lubricated gear contact conditions ranging from
136
almost dry to full-film EHL. Li et al [5.33] demonstrated the effectiveness of their line
EHL model by applying them to the methodology of Xu et al. [5.23] to predict helical
gear efficiency.
The hypoid gear efficiency model that will be developed in this chapter improves
the methodology of Xu and Kahraman [5.32] by (i) employing the gear geometry defined
in Chapter 2 to derive contact curvature and surface velocity values at each contact point
instead of relying on any particular commercial FE package, (ii) employing the loaded
tooth contact model of proposed in Chapter 3 for computation of the normal load for
minimizing the computational time required for this task, and (iii) by incorporating a new
µ formula derived by using the mixed EHL model of Li and Kahraman [5.32] such that
Figure 5.1 shows the flowchart of the methodology used to compute the power
losses of hypoid gear pairs in the proposed model. It combines the developed ease-off
based hypoid gear contact model of Chapters 2 and 3 and the EHL-based friction
dimensions, machine settings, cutter geometry, misalignments, load and speed are input
137
Blank dimensions, Machine Lubricant property,
settings, Cutter geometry, Temperature and Surface
Misalignments, Load and Speed
roughness
Req , Vr ,Vs , f
Hypoid Gear Contact Friction Coefficient
, Model
Model
Yes
k ns
No
138
At each time step k with pinion roll angle q q k k q ( k [1, ns ] ), there are
multiple potential contact lines between gear and pinion surfaces of adjacent tooth pairs
(depending on the contact ratio and the pinion roll angle). Here, ns is total number of
time steps per one pinion pitch qa 2 N p , N p is the number of teeth of the pinion and
q qa ns is pinion roll angle increment. At any incremental position k, there are ncl
(integer) number of potential contact lines, each divided into ncp number of contact
segments. With this, a total of n ncl ncp number of potential contact segments are
defined at each increment k . Each line segment k ( k [1, ns ] and [1, n ] ) has a
length Lk and carries a constant load per unit length fk . The contact at the same
segment has an equivalent radius of curvature ( Req )k between contacting surfaces as well
as constant sliding velocity (Vs )k and rolling velocity (Vr )k . Here, (Vr )k and ( Req )k are
computed in a direction perpendicular to the potential contact line they belong to.
At each line segment k , contacting surfaces are approximated as two cylinders in
combined rolling and sliding. Friction at each segment has two components of sliding and
rolling. The friction due to sliding is because of relative sliding between contacting
surfaces and friction coefficient k is defined as the ratio of the tangential force
produced due to sliding to the normal force applied between contacting surfaces.
139
The friction caused by rolling is because of resistance of contacting surfaces
against rolling over each other [5.23] and empirical rolling friction coefficient of
Goksem [5.34] is used to predict friction force ( Fr )k . In addition to the lubricant
properties and the surface roughness amplitude S , other contact parameters at each
contact segment, namely Lk , fk , (Vs )k , (Vr )k and ( Req )k must be defined to determine
the friction coefficient k and the rolling loss k ( Fr )k (Vr )k . Distributions of k
and k are used to computed the instantaneous mechanical power loss P k at increment k
due to sliding and rolling that are averaged over k to compute the average gear mesh
The load fk at each line segment k are computed using the hypoid gear contact
model proposed in Chapter 2 and 3. The computation of the rolling and sliding
velocities, (Vs )k and (Vr )k , require a kinematic analysis beyond what is provided in
Chapter 2. As shown in Figure 5.2(a) for any point on the ease-off surface with position
vector M ij , the surface velocities of the pinion and gear are defined as:
t
uijp
(a)
M ij v ijg
ap
rijp
rijg
da
(b) ag
t
t (uijp )t
u ijp
(uijg )t
u gij
(uijp )t , (uijg )t
Figure 5.2: Sliding and rolling velocities and their projection in tangential plane along
and normal to the contact line direction.
141
where p R g . The surface velocity v ijh of gear h ( h p, g ) at this contact point has
two components: w ijh in the common normal direction nij and uijh in the tangential plane
. Noting
it can be stated that uijh ( h p, g ) lies in the tangential plane and has two components,
one component (uijh )t along the instantaneous potential contact line direction t and
With this, sliding velocities at the same contact point along t and t are given,
respectively, as
2 2
vijs (vijs )t (vijs )t . (5.5)
142
The rolling velocity along t is defined as
Sliding (Vs )k and rolling (Vr )k velocities at the contact segment k ( k [1, ns ]
and [1, n ] ) are found through weighted averaging of the values at four corners of
quadrilateral grid cell on the ease-off surface that contains the middle point of this
segment. These two velocities will be computed for each contact segment and will be
The surface shear traction consists of the viscous shear within the fluid regions
and the asperity traction in the regions of metal-to-metal contact. Considering a one-
dimensional flow, the sliding viscous shear stress acting on the contact segment k is
defined as
(Vs )k
( x, t ) (5.7)
h ( x, t )
143
where h( x, t ) is the instantaneous film thickness distribution and is the effective
viscosity. Parameter x defines the coordinate of a point within the contact in the direction
of sliding. Within the asperity contact regions, the shear stress is defined as
( x , t ) d p ( x , t ) (5.8)
where p( x, t ) is the instantaneous pressure distribution and d is the dry contact friction
coefficient. With these, the average friction coefficient at a contact line segment k
xe
Nt
( x, tn )dx
1 xs
k (5.9)
Nt n 1 f k
here xs and xe are the start and end points of lubricated contact in the direction of
sliding, N t is the number of time steps at which the lubrication analysis performed.
The rolling traction formula of Goksem [5.34] is used here to find the rolling
144
Here W fk Eeq ( Req )k , U 0 (Vr )k Eeq ( Req )k , G pv Eeq , pv is the
pressure-viscosity coefficient for the lubricant used, is thermal correction factor [5.31],
ambient conditions. With this, rolling power loss at each contact segment k is
In Eq. (5.7) to (5.9), the distribution of the surface shear ( ( x, t ) ), normal pressure
are n contact line segments at each rotational increment k and there are a total of ns
distribution of the friction coefficient and the resultant gear mesh power loss. This would
require several hours of CPU time, hampering the usefulness of the model.
detailed parametric study of MEHL conditions of hypoid gear contacts will be performed
here by including all key contact parameters. These parameters and their selected ranges
145
and incremental values are listed in Table 5.1. These ranges of Hertzian pressure ph
ratio SR (0 to 1), and viscosity of a typical axle fluid (75W90) within a temperature
range (25 to 100 C ) cover most of the contact conditions present in automotive hypoid
gear pairs. In addition, a typical measured roughness profile from a gear surface with
amplitudes are obtained from this baseline profile by multiplying it by a constant. The
mechanical properties of 75W90 gear oil are listed in Table 5.2 is used.
parameter values listed in Table 5.1 were analyzed by using the model of Li et al [5.31]
and a regression analysis to the µ values predicted for each or these conditions to obtain
For 1 :
exp a0 H a1 SR a2G a3 2 SR a lnG a ln S
4 5 c
(5.11a)
a G a a ln a9 ln Sc S a10 H ,
P 6 7 8 eq
146
_____________________________________________________________
Lubricant 75W90 gear oil
_____________________________________________________________
Hertzian Pressure ph (GPa) 0.5, 1, 1.5, 2, 2.5
_____________________________________________________________
Table 5.1: Parametric design for the development of the friction coefficient formula.
147
________________________________________________________________________
Temperature Pressure-Viscosity Coefficient Dynamic Viscosity Density
Toil (°C) 1 (GPa-1) 0 (Pa.s) 0 (kg/m3)
________________________________________________________________________
25 18.0 0.1626 844.30
50 13.9 0.0499 829.30
75 11.4 0.0208 814.30
100 9.7 0.0106 799.30
________________________________________________________________________
Table 5.2: Basic parameters of the 75W90 gear oil used in this study.
148
For 1 3 :
For 3 :
c3 ln G c4 ln H c5 H c6 ln c7 ln Sc
exp c0 c1G c2 SR U ( SR)
(5.11c)
Here is the lambda ratio (ratio of the smooth condition minimum film thickness to the
Sc Sc Req and Seq Seq Req ( Sc Rq21 Rq22 and Seq Rq1 Rq2 ( Rq1 Rq2 ) ). Here,
lubricant considered. For the 75W90 gear oil, these parameters are listed in Table 5.3.
149
___________________________________________________________________
i ai bi ci
___________________________________________________________________
1 -0.62538 16.1512 5.0435
2 -51.411 -0.60156 -0.00576307
3 -0.0371532 -0.0466305 248228480
4 2.06770 -0.348239 -0.396002
5 -0.031750 -0.358514 -0.405254
6 -0.046276 22.568 35.618
7 -7.754e-5 -11.2295 -0.109382
8 1.18821 -6.49095 -0.112364
9 0.170432 -1.31986 -0.00016056
10 -0.136892 -0.881279 -0.0690238
11 -7.8882 -4398.1 -0.00038810
12 0.052821
___________________________________________________________________
150
5.2.4. Computation of the Mechanical Power Loss of the Hypoid Gear Pair
With the sliding friction coefficient ( k ) and rolling loss ( k ) computed in the
previous section for every contact segment k , the mechanical power loss due to the
With this, the instantaneous gear pair power loss at a given rotational increment k
becomes
nl
P k Pk (5.13a)
1
and the average mechanical power loss of the hypoid gear pair is found as
ns
1
P Pk . (5.13b)
ns k 1
Having input pinion torque T p and speed p and power loss P k loss at each rotational
Pk
Ek 1 , (5.14a)
T p p
151
and overall the average gear mesh efficiency is estimated as
P
E 1 . (5.14b)
T p p
Two sets of face-hobbed hypoid gear designs defined in Table 5.4 are borrowed
from automotive applications to study the influence of shaft offset along with working
conditions that affect on hypoid gear efficiency. Designs A and B have two levels of
offset ratios of d a Da 0.07 and 0.14, respectively, while the other geometry and
operating parameters (such as number of teeth, shaft angle and gear pitch diameter Da )
of the two designs are kept the same or very close to each other to isolate the offset
influences from those of the other parameters ( d a is pinion shaft offset). For this
purpose, similar ease-off topographies as shown in Figure 5.3 are developed for the two
gear sets. Despite all geometrical differences of two designs, they have similar contact
pressure distribution as shown in Figure 5.4 for pinion torque of T p 500 Nm. This is
mainly due to matching ease-off topographies of two design sets through machine setting
changes.
152
_____________________________________________________________________
Set A Set B
Parameter Pinion Gear Pinion Gear
_____________________________________________________________________
Number of teeth 12 41 12 41
Hand of Spiral Left Right Left Right
Mean spiral angle (deg) 40.0 31.3 47.0 29.6
Shaft angle (deg) 90 90
Shaft offset (mm) 15.0 30.0
Gear pitch diameter (mm) 220.0 220.0
Generation type Generate Formate Generate Formate
Cutting Method FH FH
_____________________________________________________________________
Table 5.4: Basic drive side geometry and working parameters of the examples hypoid
gear pairs.
153
(a) µm
100
5
50
0
Toe Root
(b) µm
100
50
0
Toe Root
Figure 5.3: Ease-off topography of (a) Design A with d a / Da 0.07 and (b) Design B
with d a / Da 0.14 .
154
(a) Mpa
1000
800
600
400
200
Toe Root
Mpa
(b)
1000
800
600
400
200
Toe Root
Figure 5.4: Maximum contact pressure distribution of (a) Design A with d a / Da 0.07
155
At p 1500 rpm, Vr and Vs of the two gear sets are compared in Figures 5.5
and 5.6, respectively, for these two designs. It is seen that design B with higher offset
ratio ( d a Da 0.14 ) has higher Vr and Vs compared to design A. Likewise, the slide-
to-roll ratio (SR) values over the tooth face of gear set B is higher than that of gear set A,
as shown in Figure 5.7. Additionally, the Req and distributions shown in Figures 5.8
and 5.9 reveal slight differences in Req values observed between the two gear designs,
the values for the gear set B are larger than those for the design set A. With all the
parameters required by Eq. (5.11) computed, the distributions along the tooth faces can
be determined, as shown in Figure 5.10. Although gear set B has higher ratio
compared to design A, its higher SR levels causes higher values, resulting in higher
speed p , oil temperature Toil and surface roughness S , on mechanical power loss P
and mechanical efficiency E of these gear pairs are quantified. In Figures 5.11(a1, a2)
and (b1, b2), variation of P and E with p and T p are shown. It is observed in Figure
5.11(a2, b2) that the E decreases slightly with increasing p when the speed is low.
This trend is reversed at higher speed ranges as the slope between E and p becomes
156
(a)
Toe Root
(b)
Toe Root
Figure 5.5: Rolling velocity distribution of (a) Design A with d a / Da 0.07 and (b)
157
(a)
1.0
1.5
2.0
Toe Root
(b)
1.5 2.0
2.5
3.0
Toe Root
Figure 5.6: Sliding velocity distribution of (a) Design A with d a / Da 0.07
158
(a) 0.8
0
0.7
0
0.6
0
0
0.5
0
0.4
0
0.3
0
0.2
00.1
Toe Root
(b) 0.8
0.7
0.6
0.5
0.4
0.3
0.2
Toe Root 0.1
Figure 5.7: Slide-to-roll ratio distribution of (a) Design A with d a / Da 0.07 and (b)
159
(a)
35 40 45
50
Toe Root
(b)
35
40 45
50
55
Toe Root
160
(a)
0.30
0.25
0.20
0.15
(b) 0.30
0.25
0
0.20
0
0.15
0
Figure 5.9: distribution of (a) Design A with d a / Da 0.07 and (b) Design B with
161
(a) 0.06
0.05
0.04
0.03
0.02
Toe Root 0.01
(b) 0.06
0.05
0.04
0.03
0.02
Figure 5.10: Friction coefficient distribution of (a) Design A with d a / Da 0.07 and
(b) Design B with d a / Da 0.14 at p 1500 rpm , T p 500 Nm , Toil 90 C and
S1 S 2 0.8 m .
162
(a1)
10,000
1,000
P [w] 100
10
1,000
P [w] 100
10
p Continued
Figure 5.11: Power loss and efficiency of Design A (a1, a2) and Design B (b1, b2) at
163
Continued
(a2)
99.5
99.0
98.5
E[%] 98.0
97.5
97.0
96.5
100 Nm 500 Nm 1000 Nm
96.0
0 500 1000 1500 2000 2500 3000
p
(b2)
99.5
99.0
98.5
E[%] 98.0
97.5
97.0
96.5
100 Nm 500 Nm 1000 Nm
96.0
0 500 1000 1500 2000 2500 3000
p
Figure 5.11 continued
164
p
In the lower speed range, where the power loss is relatively small, higher T
results in higher efficiency since the increase in power loss due to the heavier load
applied is smaller than the corresponding input power increase, while the opposite is true
in the medium and high speed ranges. Between the two designs, gear set A with a
smaller shaft off-set is consistently more efficient (on average, gear set A has about 1.5%
Figure 5.12 illustrates the effects of surface finish as well as the lubricant
temperature on the gear mesh efficiency. It is found that reduction in surface roughness
amplitude effectively increases the mechanical efficiency of both gear pairs. As for the
operating temperature, within the low roughness range, where a substantial amount of
contact area is separated by the hydrodynamic fluid film and the viscous shear dominates,
an increase in temperature will reduces the sliding friction and the power loss P
accordingly through the reduction in lubricant viscosity. In the medium and high
roughness ranges, when the contact zone might experience severe asperity contacts,
lower lubricant viscosity at high temperature results in thinner fluid film and larger
165
(a)
99.5
99.0
98.5
98.0
Ef [%]
97.5
97.0
Toil
96.5
50° C 75° C 100° C
96.0
0.1 0.3 0.5 0.7 0.9 1.1 1.3
S
(b)
99.5
99.0
98.5
98.0
Ef [%]
97.5
97.0
Toil
96.5
50° C 75° C 100° C
96.0
0.1 0.3 0.5 0.7 0.9 1.1 1.3
S
Figure 5.12: Efficiency of (a) Design A with d a / Da 0.07 and (b) design B with
d a / Da 0.14 for different surface finish and oil temperatures at p 1500 rpm and
T p 500 Nm .
166
5.4. Conclusion
In this chapter, a new spiral bevel and hypoid gear mechanical efficiency model is
proposed for both face-milling and face-hobbing cutting methods. The proposed
formulation for rolling loss. The developed model improved the methodology of Xu and
Kahraman [5.30] by (i) employing the gear geometry defined in Chapter 2 to derive
contact curvature and surface velocity values at each contact point instead of relying on
any particular commercial FE package, (ii) employing the loaded tooth contact model
proposed in Chapter 3 for computation of the normal load for minimizing the
computational time required for this task, and (iii) by incorporating a new µ formula
derived by using the mixed EHL model of Li et al [5.33] such that any degree of asperity
Limited numerical results show that the shaft off-set is critical to the efficiency of
the gear set as lower off-sets resulting in significant increases in mechanical efficiency.
Likewise, reduction in surface roughness was also shown to reduce power losses of
167
References for Chapter 5
[5.1] Coleman, W., 1975, "Computing efficiency for bevel and hypoid gears." Machine
Design, 47, pp. 64-65.
[5.2] Seetharaman, S., and Kahraman, A., 2009, "Load-Independent Spin Power Losses
of a Spur Gear Pair: Model Formulation." ASME J. of Tribology, 131(2), pp.
022201.
[5.3] Seetharaman, S., Kahraman, A., Moorhead, M. D., and Petry-Johnson, T. T.,
2009, "Oil Churning Power Losses of a Gear Pair: Experiments and Model
Validation." ASME J. of Tribology, 131(2), pp. 022202.
[5.4] Stadtfeld, H., J., 1993, Handbook of Bevel and Hypoid Gears, Rochester Institute
of Technology.
[5.5] Denny, C. M., 1998, "Mesh Friction in Gearing." AGMA, Technical Paper No.
98FTM2.
[5.6] Pedrero, J. I., 1999, "Determination of The Efficiency of Cylindrical Gear Sets."
4th World Congress on Gearing and Power Transmission, Paris, France.
[5.7] Michlin, Y., and Myunster, V., 2002, "Determination of Power Losses in Gear
Transmissions with Rolling and Sliding Friction Incorporated." J. Mechanism and
Machine Theory, 37, pp. 167.
[5.8] Benedict, G. H., and Kelly, B. W., 1960, "Instantaneous Coefficients of Gear
Tooth Friction." ASLE.
[5.9] O’Donoghue, J. P., and Cameron, A., 1966, "Friction and Temperature in Rolling
Sliding Contacts." ASLE Transactions, 9, pp. 186-194.
168
[5.10] Drozdov, Y. N., and Gavrikov, Y. A., 1967, "Friction and Scoring Under The
Conditions of Simultaneous Rolling and Sliding of Bodies." Wear, pp. 291-302.
[5.11] Misharin, Y. A., 1958, "Influence of The Friction Condition on The Magnitude of
The Friction Coefficient in The Case of Rollers with Sliding." Int. Conference On
Gearing, London, UK.
[5.12] Heingartner, P., and Mba, D., 2003, "Determining Power Losses in The Helical
Gear Mesh; Case Study." Proceeding of DETC3, Chicago, Illinois, USA.
[5.13] Anderson, N. E., and Loewenthal, S. H., 1986, "Efficiency of Nonstandard and
High Contact Ratio Involute Spur Gears." J. Mechanisms, Transmissions and
Automation in Design, 108, pp. 119-126.
[5.14] Anderson, N. E., and Loewenthal, S. H., 1982, "Design of Spur Gears for
Improved Efficiency." J. Mechanical Design, 104, pp. 767-774.
[5.16] Vaishya, M., and Houser, D. R., 1999, "Modeling and Measurement of Sliding
Friction for Gear Analysis." AGMA, Technical Paper No. 99FTMS1.
[5.17] Martin, K. F., 1981, "The Efficiency of Involute Spur Gears." ASME J. Mech.
Des., 103, pp. 160-169.
[5.18] Dowson, D., and Higginson, G. R., 1964, "A Theory of Involute Gear
Lubrication." Institute of Petroleum, Gear Lubrication, Elsevier, London, UK.
169
[5.19] Simon, V., 1981, "Load Capacity and Efficiency of Spur Gears in Regard to
Thermo-End Lubrication." International Symposium on Gearing and Power
Transmissions, Tokyo, Japan.
[5.20] Simon, V., 2009, "Influence of machine tool setting parameters on EHD
lubrication in hypoid gears." J. Mechanism and Machine Theory, 44, pp. 923-
937.
[5.22] Wang, X. C., and Ghosh, S. K., 1994, Advanced Theories of Hypoid Gears,
Elsevier Science B. V.
[5.23] Xu, H., Kahraman, A., Anderson, N. E., and Maddock, D. G., 2007, "Prediction
of Mechanical Efficiency of Parallel-Axis Gear Pairs." ASME J. Mech. Des.,
129(1), pp. 58-68.
[5.24] Cioc, C., Cioc, S., Kahraman, A., and Keith, T., 2002, "A Non-Newtonian,
Thermal EHL Model of Contacts with Rough Surfaces." Tribology Transactions,
45, pp. 556-562
170
[5.28] Vijayakar, S., 2004, Calyx Hypoid Gear Model, User Manual, Advanced
Numerical Solution Inc., Hilliard, Ohio.
[5.29] Litvin, F. L., and Fuentes, A., 2004, Gear Geometry and Applied Theory (2nd
ed.), Cambridge University Press, Cambridge.
[5.30] Xu, H., and Kahraman, A., 2007, "Prediction of Friction-Related Power Losses of
Hypoid Gear Pairs." Proceedings of the Institution of Mechanical Engineers, Part
K: J. Multi-body Dynamics, 221(3), pp. 387-400.
[5.31] Li, S., and Kahraman, A., 2009, "A Mixed EHL Model with Asymmetric
Integrated Control Volume Discretization." Tribology International, Hiroshima,
Japan.
[5.32] Li, S., and Kahraman, A., 2009, "An Asymmetric Integrated Control Volume
Method for Transient Mixed Elastohydrodynamic Lubrication Analysis of
Heavily Loaded Point Contact Problems." Tribology International.
[5.33] Li, S., Vaidyanathan, A., Harianto, J., and Kahraman, A., 2009, "Influence od
Design Parameters and Micro-Geometry on Mechanical Power Losses of Helical
Gear Pairs." JSME International, Hiroshima, Japan.
[5.34] Goksem, P. G., and Hargreaves, R. A., 1978, "The effect of viscous shear heating
on both film thickness and rolling traction in an EHL line contact." J. Lubrication
Technology, 100, pp. 346–352.
[5.35] Wu, S., and Cheng, H., S., 1991, "A Friction Model of Partial-EHL Contacts and
its Application to Power Loss in Spur Gears." Tribology Transactions, 34(3), pp.
398-407.
171
CHAPTER 6
A computationally efficient load distribution model was proposed for both face-
milled and face-hobbed hypoid gears produced by Formate and generate processes.
Tooth surfaces were defined directly from the cutter parameters and machine settings.
This study defined and utilized a new surface of roll angle as an essential tool to simplify
the task of locating instantaneous contact lines of any general type of gearing in the
projection plane. First, the position vector and normal to one of the mating surfaces of
contacting members were computed, and the action surface and the surface of roll angle
were introduced by applying equation of meshing between any general axis arrangement.
Once the surface of roll angle was constructed, the instantaneous contact lines location
and orientation were computed through a novel approach inspired by analogy to parallel
axes gears. Gear surfaces were assumed conjugate only in computation of contact line
172
locations and orientation of the proposed approach. However it was shown that surface
of roll angle is very insensitive to any practical modifications of contacting surfaces, and
hence, real instantaneous contact lines practically remain identical to their conjugate
counterparts. For any other steps of contact analysis, theoretically generated surfaces
Rayleigh-Ritz based shell models of teeth of the gear and pinion were developed
to define the tooth compliances due to bending and shear effects efficiently in a semi-
analytical manner. Base rotation and contact deformation effects were also included in
the compliance formulations. With this, loaded contact patterns and transmission error of
both face-milled and face-hobbed spiral bevel and hypoid gears were computed by
enforcing the compatibility and equilibrium conditions of the gear mesh. The proposed
model requires significantly less computational effort than finite elements (FE) based
models, making its use possible for extensive parameter sensitivity and design
were also provided to demonstrate the accuracy of the model under various load and
misalignment conditions.
Two applications of the proposed model were also introduced. First application
combined the proposed model with a newly introduced approach of modifying the ease-
off topography to investigate the effect of errors occurred in manufacturing and heat
errors typically cause real (measured) spiral bevel and hypoid gear surfaces to deviate
173
from the theoretical ones globally. Tooth surface wear patterns accumulated through the
life span of the gear set are typically local deviations that are aggravated especially in
case of edge contact conditions. An accurate and practical methodology based on the
developed ease-off topography approach was proposed in this study to perform loaded
tooth contact analysis of spiral bevel and hypoid gears having both types of local and
global deviations. Manufacturing errors and localized surface wear deviations were used
to update the theoretical ease-off and surface of roll angle to form a new ease-off surface
that was used to perform a loaded tooth contact analysis. Two numerical examples of
face-milled and face-hobbed hypoid gear sets with local and global deviated surfaces,
methodology as well as quantifying the effect of such deviations on load distribution and
efficiency model was developed next for both face-milling and face-hobbing cutting
methods. The proposed efficiency model combined the computationally efficient contact
model and a mixed EHL based surface traction model to predict friction power losses.
The contact area, pressure distribution and rolling and sliding velocities were determined
employing the developed loaded tooth contact model. The EHL traction model
considered specific ranges of the key contact parameters, including Hertzian pressure,
contact radii, surface speeds, lubricant temperature and surface roughness amplitude of
hypoid type of gears, covering wide range of lubrication conditions from full film to
174
boundary regimes. The efficiency model for hypoid gear was applied to two face-hobbing
examples with similar overall dimensions, but different off-sets, to investigate the effects
of several working and design parameters including offset, load, speed, surface roughness
analysis intended to predict required functional parameters of the hypoid gear pair,
including the transmission error, contact stresses, root bending stresses, fatigue life and
mechanical power losses. The hypoid gear literature lacked a model to compute the load
FE methods. The main potential reasons for that was the absence of a general and
reliable formulation to define the geometry of FH and FM hypoid tooth surfaces from
cutter parameters, machine motions and settings. This void, combined with the
numerical difficulties in matching the tooth surfaces using the conventional methods and
lack of a semi-analytical tooth compliance formulation for hypoid gears, has hampered in
design, analysis and optimization of hypoid gears. This research study fills some of this
void.
The model proposed in this study to simulate the contacts of FM and FH hypoid
gear pairs under both unladed and loaded condition provides major enhancements to the
(i) Methodology that simulates the FM and FH processes to define surface
geometries of hypoid gears including the coordinates, normal vectors and radii of
(ii) The method using ease-off topography to compute the unladed contact conditions
numerical stability and computational efficiency. This method and its surface of
roll angle concept is also general such that it can be applied to the other gear
types.
(iii) Application of mounting errors and inclusion of global and local deviations are
rather straightforward with the proposed model while these have typically been
(iv) The efficiency model that combines the loaded tooth contact model proposed in
superior to any published hypoid gear efficiency model as it includes all key
Its ability to handle variety of lubrication conditions ranging from almost dry
contact to full film EHL makes this model applicable to wide ranges of hypoid
and spiral bevel gear applications from automotive and aerospace systems.
176
6.3. Recommendations for Future Work
The following items can be considered as the potential future studies to improve
(ii) The loaded TCA formulation can be improved by enhancing the base rotation
(iii) A shell model for non-circular cylinders can be developed to compute a more
accurate compliance matrix for a generated gear and the face-hobbed generation
(iv) The proposed model lends itself to optimization studies to refine machine settings
for desired transmission error amplitudes and loaded tooth contact patterns.
(v) The way local deviation and mounting errors are included in the proposed model
is general such that this model can be used to predict surface wear as well as to
hypoid gears.
177
BIBLIOGRAPHY
[1] Anderson, N. E., and Loewenthal, S. H., 1982, "Design of Spur Gears for
Improved Efficiency." J. Mechanical Design, 104, pp. 767-774.
[2] Anderson, N. E., and Loewenthal, S. H., 1986, "Efficiency of Nonstandard and
High Contact Ratio Involute Spur Gears." J. Mechanisms, Transmissions and
Automation in Design, 108, pp. 119-126.
[4] Baxter, M. L., 1964, "An Application of Kinematics and Vector Analysis to the
Design of a Bevel Gear Grinder." ASME Mechanism Conference, Lafayette, IN.
[5] Baxter, M. L., and Spear, G. M., 1961, "Effects of Misalignment on Tooth Action
of Bevel and Hypoid Gears." ASME Design Conference, Detroit, MI.
[6] Benedict, G. H., and Kelly, B. W., 1960, "Instantaneous Coefficients of Gear
Tooth Friction." ASLE.
[7] Borner, J., Kurz, N., and Joachim, F. (2002). "Effective Analysis of Gears with
the Program LVR (Stiffness Method)."
[9] Cioc, C., Cioc, S., Kahraman, A., and Keith, T., 2002, "A Non-Newtonian,
Thermal EHL Model of Contacts with Rough Surfaces." Tribology Transactions,
45, pp. 556-562
[10] Coleman, W., 1963, Design of Bevel Gears, The Gleason Works.
178
[11] Coleman, W., 1975, "Analysis of Mounting Deflections on Bevel and Hypoid
Gears." SAE 750152.
[12] Coleman, W., 1975, "Computing efficiency for bevel and hypoid gears." Machine
Design, 47, pp. 64-65.
[13] Coleman, W., 1975, "Effect of Mounting Displacements on Bevel and Hypoid
Gear Tooth Strength." SAE 750151.
[14] Conry, T. F., and Seireg, A., 1973, "A Mathematical Programming Technique for
the Evaluation of Load Distribution and Optimal Modification for Gear Systems."
ASME J. of Industrial Engineering.
[15] Denny, C. M., 1998, "Mesh Friction in Gearing." AGMA, Technical Paper No.
98FTM2.
[16] Dooner, D. B., 2002, "On the Three Laws of Gearing." ASME J. Mech. Des., 124,
pp. 733-744.
[17] Dooner, D. B., and Seireg, A., 1995, The Kinematic Geometry of Gearing: A
Concurrent Engineering Approach, John Wiley & Sons Inc.
[18] Drozdov, Y. N., and Gavrikov, Y. A., 1967, "Friction and Scoring Under The
Conditions of Simultaneous Rolling and Sliding of Bodies." Wear, pp. 291-302.
[19] Dowson, D., and Higginson, G. R., 1964, "A Theory of Involute Gear
Lubrication." Institute of Petroleum, Gear Lubrication, Elsevier, London, UK.
[20] Dudley, D. W., 1969, The Evolution of the Gear Art, American Gear
Manufacturers Association, Washington, D. C.
179
[21] Dyson, A., 1969, A General Theory of the Kinematics and Geometry of Gears in
Three Dimensions, Clarendon Press, Oxford.
[22] Fan, Q., 2006, "Computerized Modeling and Simulation of Spiral Bevel and
Hypoid Gears Manufactured by Gleason Face Hobbing Process." ASME J. Mech.
Des., 128(6), pp. 1315-1327.
[23] Fan, Q., 2007, "Enhanced Algorithms of Contact Simulation for Hypoid Gear
Drives Produced by Face-Milling and Face-Hobbing Processes." ASME J. Mech.
Des., 129(1), pp. 31-37.
[24] Fan, Q., and Wilcox, L., 2005, "New Developments in Tooth Contact Analysis
(TCA) and Loaded TCA for Spiral Bevel and Hypoid Gear Drives." AGMA,
Technical Paper No. 05FTM08.
[25] Fong, Z. H., 2000, "Mathematical Model of Universal Hypoid Generator with
Supplemental Kinematic Flank Correction Motions." ASME J. Mech. Des.,
122(1), pp. 136-142.
[26] Fong, Z. H., and Tsay, C.-B., 1991, "A Mathematical Model for the Tooth
Geometry of Circular-Cut Spiral Bevel Gears." ASME J. Mech. Des., 113, pp.
174-181.
[27] Fong, Z. H., and Tsay, C.-B., 1991, "A Study on the Tooth Geometry and Cutting
Machine Mechanisms of Spiral Bevel Gears." ASME J. Mech. Des., 113, pp. 346-
351.
[28] Goksem, P. G., and Hargreaves, R. A., 1978, "The effect of viscous shear heating
on both film thickness and rolling traction in an EHL line contact." J. Lubrication
Technology, 100, pp. 346–352.
180
[29] Gosselin, C., et al. 1991, "Tooth Contact Analysis of High Conformity Spiral
Bevel Gears." Proceedings of JSME Int. Conf. on Motion and Power
Transmission, Hiroshima, Japan.
[30] Gosselin, C., Cloutier, L., and Nguyen, Q. D., 1995, "A general formulation for
the calculation of the load sharing and transmission error under load of spiral
bevel and hypoid gears." J. Mechanism and Machine Theory, 30(3), pp. 433-450.
[31] Gosselin, C., Guertin, T., Remond, D., and Jean, Y., 2000, "Simulation and
Experimental Measurement of the Transmission Error of Real Hypoid Gears
Under Load." ASME J. Mech. Des., 122(1), pp. 109-122.
[32] Gosselin, C., Jiang, Q., Jenski, K., and Masseth, J., 2005, "Hypoid Gear Lapping
Wear Coefficient and Simulation." AGMA, Technical Paper No. 05FTM09.
[33] Gosselin, C., Nonaka, T., Shiono, Y., Kubo, A., and Tatsuno, T., 1998,
"Identification of the Machine Settings of Real Hypoid Gear Tooth Surfaces."
ASME J. Mech. Des., 120(3), pp. 429-440.
[34] Heingartner, P., and Mba, D., 2003, "Determining Power Losses in The Helical
Gear Mesh; Case Study." Proceeding of DETC3, Chicago, Illinois, USA.
[5] Krenzer, T. J., 1965, TCA Formulas and Calculation procedures, The Gleason
Works.
[36] Krenzer, T. J., 1981, "Tooth Contact Analysis of Spiral Bevel and Hypoid Gears
under Load." SAE Earthmoving Industry Conference, Peoria, IL.
[37] Krenzer, T. J., 1981, Understanding Tooth Contact Analysis, The Gleason Works.
181
[38] Krenzer, T. J., 1990, "Face Milling or Face Hobbing." AGMA, Technical Paper
No. 90FTM13.
[40] Kin, V., 1992, "Computerized Analysis of Gear Meshing Based on Coordinate
Measurement Data." ASME Int. Power Transmission and Gearing Conference,
Scottsdale, AZ.
[41] Kin, V., 1992, "Tooth Contact Analysis Based on Inspection." Proceedings of 3rd
World Congress on Gearing, Paris, France.
[42] Li, S., and Kahraman, A., 2009, "A Mixed EHL Model with Asymmetric
Integrated Control Volume Discretization." Tribology International, Hiroshima,
Japan.
[43] Li, S., and Kahraman, A., 2009, "An Asymmetric Integrated Control Volume
Method for Transient Mixed Elastohydrodynamic Lubrication Analysis of
Heavily Loaded Point Contact Problems." Tribology International.
[44] Li, S., Vaidyanathan, A., Harianto, J., and Kahraman, A., 2009, "Influence od
Design Parameters and Micro-Geometry on Mechanical Power Losses of Helical
Gear Pairs." JSME International, Hiroshima, Japan.
[47] Litvin, F. L., and Fuentes, A., 2004, Gear Geometry and Applied Theory (2nd
ed.), Cambridge University Press, Cambridge.
182
[48] Litvin, F. L., Fuentes, A., Fan, Q., and Handschuh, R. F., 2002, "Computerized
design, simulation of meshing, and contact and stress analysis of face-milled
Formate generated spiral bevel gears." J. Mechanism and Machine Theory, 37(5),
pp. 441-459.
[49] Litvin, F. L., and Gutman, Y., 1981, "A Method of Local Synthesis of Gears
Grounded on the Connections Between the Principal and Geodetic of Surfaces."
ASME J. Mech. Des., 103, pp. 114-125.
[50] Litvin, F. L., and Gutman, Y., 1981, "Methods of synthesis and analysis for
hypoid gear-drives of Formate and helixform; Part I-Calculation for machine
setting for member gear manufacture of the Formate and helixform hypoid gears."
ASME J. Mech. Des., 103, pp. 83-88.
[51] Litvin, F. L., and Gutman, Y., 1981, "Methods of synthesis and analysis for
hypoid gear-drives of Formate and helixform; Part II-Machine setting calculations
for the pinions of Formate and helixform gears." ASME J. Mech. Des., 103, pp.
89-101.
[52] Litvin, F. L., and Gutman, Y., 1981, "Methods of synthesis and analysis for
hypoid gear-drives of Formate and helixform; Part III-Analysis and optimal
synthesis methods for mismatched gearing and its application for hypoid gears of
Formate and helixform." ASME J. Mech. Des., 103, pp. 102-113.
[53] Litvin, F. L., Zhang, Y., Lundy, M., and Heine, C., 1988, "Determination of
Settings of a Tilted Head Cutter for Generation of Hypoid and Spiral Bevel
Gears." J. Mechanism, Transmission and Automation in Design, 110, pp. 495-
500.
183
[54] Liu, F., 2004, "Face Gear Design and Compliance Analysis," M.Sc. Thesis, The
Ohio State University, Columbus, Ohio.
[55] Martin, K. F., 1981, "The Efficiency of Involute Spur Gears." ASME J. Mech.
Des., 103, pp. 160-169.
[56] Michlin, Y., and Myunster, V., 2002, "Determination of Power Losses in Gear
Transmissions with Rolling and Sliding Friction Incorporated." J. Mechanism and
Machine Theory, 37, pp. 167.
[57] Mindlin, R. D., 1951, "Influence of Rotary Inertia and Shear on Flexural Motions
of Isotropic Elastic Plates. ." J. of Applied Mechanics, 18, pp. 31-38.
[58] Misharin, Y. A., 1958, "Influence of The Friction Condition on The Magnitude of
The Friction Coefficient in The Case of Rollers with Sliding." Int. Conference On
Gearing, London, UK.
[59] O’Donoghue, J. P., and Cameron, A., 1966, "Friction and Temperature in Rolling
Sliding Contacts." ASLE Transactions, 9, pp. 186-194.
[60] Park, D., and Kahraman, A., 2008, "A Surface Wear Model for Hypoid Gear
Pairs." In press, Wear.
[61] Pedrero, J. I., 1999, "Determination of The Efficiency of Cylindrical Gear Sets."
4th World Congress on Gearing and Power Transmission, Paris, France.
184
[62] Seetharaman, S., and Kahraman, A., 2009, "Load-Independent Spin Power Losses
of a Spur Gear Pair: Model Formulation." ASME J. of Tribology, 131(2), pp.
022201.
[63] Seetharaman, S., Kahraman, A., Moorhead, M. D., and Petry-Johnson, T. T.,
2009, "Oil Churning Power Losses of a Gear Pair: Experiments and Model
Validation." ASME J. of Tribology, 131(2), pp. 022202.
[64] Shtipelman, B. A., 1979, Design and manufacture of hypoid gears, John Wiley &
Sons, Inc.
[66] Simon, V., 1981, "Load Capacity and Efficiency of Spur Gears in Regard to
Thermo-End Lubrication." International Symposium on Gearing and Power
Transmissions, Tokyo, Japan.
[67] Simon, V., 1996, "Tooth Contact Analysis of Mismatched Hypoid Gears." ASME
International Power Transmission and Gearing Conference ASME, San Diego.
[68] Simon, V., 2000, "FEM stress analysis in hypoid gears." J. Mechanism and
Machine Theory, 35(9), pp. 1197-1220.
[69] Simon, V., 2000, "Load Distribution in Hypoid Gears." ASME J. Mech. Des.,
122(44), pp. 529-535.
[70] Simon, V., 2001, "Optimal Machine Tool Setting for Hypoid Gears Improving
Load Distribution." ASME J. Mech. Des., 123(4), pp. 577-582.
185
[71] Simon, V., 2009, "Influence of machine tool setting parameters on EHD
lubrication in hypoid gears." J. Mechanism and Machine Theory, 44, pp. 923-
937.
[72] Smith, R. E., 1984, "What Single Flank Measurement Can Do For You." AGMA,
Technical Paper No. 84FTM2.
[73] Smith, R. E., 1987, "The Relationship of Measured Gear Noise to Measured Gear
Transmission Errors." AGMA, Technical Paper No. 87FTM6.
[74] Stadtfeld, H., J., 1993, Handbook of Bevel and Hypoid Gears, Rochester Institute
of Technology.
[75] Stadtfeld, H., J., 1995, Gleason Advanced Bevel Gear Technology, The Gleason
Works.
[76] Stadtfeld, H. J. (2000). "The Basics of Gleason Face Hobbing." The Gleason
Works.
[77] Stadtfeld, H. J., and Gaiser, U., 2000, "The Ultimate Motion Graph." ASME J.
Mech. Des., 122(3), pp. 317-322.
[78] Stegemiller, M. E., 1986, "The Effects of Base Flexibility on Thick Beams and
Plates Used in Gear Tooth Deflection Models," M.Sc. Thesis, The Ohio State
University, Columbus, Ohio.
[79] Stegemiller, M. E., and Houser, D. R., 1993, "A Three Dimensional Analysis of
the Base Flexibility of Gear Teeth." ASME J. Mech. Des., 115(1), pp. 186-192.
[80] Stewart, A. A., and Wildhaber, E., 1926, "Design, Production and Application of
the Hypoid Rear-Axle Gear." J. SAE, 18, pp. 575-580.
186
[81] Sugyarto, E., 2002, "The Kinematic Study, Geometry Generation, and Load
Distribution Analysis of Spiral Bevel and Hypoid Gears," M.Sc. Thesis, The Ohio
State University, Columbus, Ohio.
[82] Timoshenko, S. P., and Woinowsky-Krieger, S., 1959, Theory of Plates and
Shells, McGraw-Hill Book Company Inc.
[83] Tsai, Y. C., and Chin, P. C., 1987, "Surface Geometry of Straight and Spiral
Bevel Gears." J. Mechanism, Transmission and Automation in Design, 109, pp.
443-449.
[84] Vaidyanathan, S., 1993, "Application of Plate and Shell Models in the Loaded
Tooth Contact Analysis of Bevel and Hypoid Gears," Ph.D. Dissertation, The
Ohio State University, Columbus, Ohio.
[85] Vaidyanathan, S., Houser, D. R., and Busby, H. R., 1993, "A Rayleigh-Ritz
Approach to Determine Compliance and Root Stresses in Spiral Bevel Gears
Using Shell Theory." AGMA, Technical Paper No. 93FTM03.
[86] Vaidyanathan, S., Houser, D. R., and Busby, H. R., 1994, "A Numerical
Approach to the Static Analysis of an Annular Sector Mindlin Plate with
Applications to Bevel Gear Design." J. of Computers & Structures, 51(3), pp.
255-266.
[87] Vaishya, M., and Houser, D. R., 1999, "Modeling and Measurement of Sliding
Friction for Gear Analysis." AGMA, Technical Paper No. 99FTMS1.
[88] Vecchiato, D., 2005, "Design and Simulation of Face-Hobbed Gears and Tooth
Contact Analysis by Boundary Element Method," Ph.D. Dissertation, University
of Illinois at Chicago.
187
[89] Vijayakar, S. M., 1991, "A Combined Surface Integral and Finite Element
Solution for a Three-Dimensional Contact Problem." International J. for
Numerical Methods in Engineering, 31, pp. 525-545.
[90] Vijayakar, S. M., 2003, Calyx User Manual, Advanced Numerical Solution Inc.,
Hilliard, Ohio.
[91] Vijayakar, S. M., 2004, Calyx Hypoid Gear Model, User Manual, Advanced
Numerical Solution Inc., Hilliard, Ohio.
[92] Vimercati, M., and Piazza, A., 2005, "Computerized Design of Face Hobbed
Hypoid Gears: Tooth Surfaces Generation, Contact Analysis and Stress
Calculation." AGMA, Technical Paper No. 05FTM05.
[94] Vogel, O., Griewank, A., and Bär, G., 2002, "Direct gear tooth contact analysis
for hypoid bevel gears." Computer Methods in Applied Mechanics and
Engineering, 191(36), pp. 3965-3982.
[95] Wang, X. C., and Ghosh, S. K., 1994, Advanced Theories of Hypoid Gears,
Elsevier Science B. V.
[96] Weber, C., 1949, "The Deformation of Loaded Gears and the Effect on Their
Load Carrying Capacity (Part I)." D.S.I.R., London.
[97] Wilcox, L. E., Chimner, T. D., and Nowell, G. C., 1997, "Improved Finite
Element Model for Calculating Stresses in Bevel and Hypoid Gear Teeth."
AGMA, Technical Paper No. 97FTM05.
188
[98] Wildhaber, E., 1946, Basic Relationship of Hypoid Gears, McGraw-Hill.
[99] Wildhaber, E., 1956, "Surface Curvature." Product Engineering, pp. 184-191.
[100] Wu, D., and Luo, J., 1992, A Geometric Theory of Conjugate Tooth Surfaces,
World Scientific, River Edge, NJ.
[101] Wu, S., and Cheng, H., S., 1991, "A Friction Model of Partial-EHL Contacts and
its Application to Power Loss in Spur Gears." Tribology Transactions, 34(3), pp.
398-407.
[103] Xu, H., Kahraman, A., Anderson, N. E., and Maddock, D. G., 2007, "Prediction
of Mechanical Efficiency of Parallel-Axis Gear Pairs." ASME J. Mech. Des.,
129(1), pp. 58-68.
[104] Xu, H., and Kahraman, A., 2007, "Prediction of Friction-Related Power Losses of
Hypoid Gear Pairs." Proceedings of the Institution of Mechanical Engineers, Part
K: J. Multi-body Dynamics, 221(3), pp. 387-400.
[105] Xu, H., Kahraman, A., and Houser, D. R., 2005 "A Model to Predict Friction
Losses of Hypoid Gears." AGMA, Technical Paper No. 0FTM06.
[106] Yakubek, D., Busby, H. R., and Houser, D. R., 1985, "Three-Dimensional
Deflection Analysis of Gear Teeth Using Both Finite Element Analysis and a
Tapered Plate Approximation." AGMA, Technical Paper No. 85FTM4.
189
[107] Yau, H., 1987, "Analysis of Shear Effect on Gear Tooth Deflections Using the
Rayleigh-Ritz Energy Method," M.Sc. Thesis, The Ohio State University,
Columbus, Ohio.
[108] Yau, H., Busby, H. R., and Houser, D. R., 1991, "A Rayleigh-Ritz Approach for
Modeling the Bending and Shear Deflections of Gear Teeth." JSME International
Conference on Motion and Powertransmission.
[109] Zhang, Y., Litvin, F. L., Maryuama, N., Takeda, R., and Sugimoto, M., 1994,
"Computerized Analysis of Meshing and Contact of Gear Real Tooth Surfaces."
116, pp. 677-682.
[110] Zhang, Y., and Wu, Z., 2007, "Geometry of Tooth Profile and Fillet of Face-
Hobbed Spiral Bevel Gears." IDETC/CIE 2007, Las Vegas, Nevada, USA.
190