Compact Design for Non-Standard Spur Gears 1 - Scientific ...
Compact Design for Non-Standard Spur Gears 1 - Scientific ...
Compact Design for Non-Standard Spur Gears 1 - Scientific ...
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
Volume 2, Issue 1, 2011<br />
<strong>Compact</strong> <strong>Design</strong> <strong>for</strong> <strong>Non</strong>-<strong>Standard</strong> <strong>Spur</strong> <strong>Gears</strong><br />
Tuan Nguyen, Graduate Student, Department of Mechanical Engineering, The University of Memphis,<br />
tnguyen1@memphis.edu<br />
Hsiang H. Lin, Professor, Department of Mechanical Engineering, The University of Memphis,<br />
hlin1@memphis.edu<br />
Abstract<br />
This paper presents procedures <strong>for</strong> compact design of non-standard spur gear sets with the objective of<br />
minimizing the gear size. The procedures are <strong>for</strong> long and short addendum system gear pairs that have an<br />
equal but opposite amount of hob cutter offset applied to the pinion and gear in order to maintain the<br />
standard center distance. Hob offset has shown to be an effective way to balance the dynamic tooth stress<br />
of the pinion and the gear to increase load capacity. The allowable tooth stress and dynamic response are<br />
incorporated in the process to obtain a feasible design region. Various dynamic rating factors were<br />
investigated and evaluated. The constraints of contact stress limits and involute interference combined with<br />
the tooth bending strength provide the main criteria <strong>for</strong> this investigation. A three-dimensional design space<br />
involving the gear size, diametral pitch, and operating speed was developed to illustrate the optimal design<br />
of the non-standard spur gear pairs.<br />
Operating gears over a range of speeds creates variations in the dynamic response and changes the<br />
required gear size in a trend that parallels the dynamic factor. The dynamic factors are strongly affected by<br />
the system natural frequencies. The peak values of the dynamic factor within the operating speed range<br />
significantly influence the optimal gear designs. The refined dynamic factor introduced in this study yields<br />
more compact designs than those produced by AGMA dynamic factors.<br />
Introduction<br />
<strong>Design</strong>ing gear transmission systems involves selecting combinations of gears to produce a desired speed<br />
ratio. If the ratio is not unity, the gears will have different diameters. The tooth strength of the smaller gear<br />
(the pinion) is generally weaker than that of the larger one (the gear) if both are made of the same material.<br />
In some designs, pinions with very small tooth numbers must be used in order to satisfy the required speed<br />
ratio and fit in the available space. This can lead to undercut of the teeth in the pinion which further reduces<br />
strength.<br />
One solution to the pinion design problem is to specify nonstandard gears in which the tooth profile of the<br />
pinion is shifted outward slightly, thus increasing its strength, while the gear tooth profile may be decreased<br />
by an equal amount. These changes in tooth proportions may be accomplished without changing operating<br />
center distance and with standard cutting tools by withdrawing the cutting tool slightly as the pinion blank is<br />
cut and advancing the cutter the same distance into the gear blank. This practice is called the long and<br />
short addendum system. When gears are cut by hobs, this adjustment to the tooth proportions is called hob<br />
cutter offset.<br />
Several studies of non-standard gears have been per<strong>for</strong>med to determine the proper hob cutter offset to<br />
produce gear pairs of different sizes and contact ratios with similar tooth strength. Walsh and Mabie [1]<br />
investigated the method to adjust the hob offset values <strong>for</strong> pinion and gear so that the stress in the pinion<br />
teeth was approximately equal to that in the gear teeth. They developed hob offset charts <strong>for</strong> various velocity<br />
ratios and <strong>for</strong> several changes in center distance. Mabie, Walsh, and Bateman [2] per<strong>for</strong>med similar but<br />
more detailed work about tooth stress equalization and compared the results with those found using either<br />
the Lewis <strong>for</strong>m factor Y or the AGMA <strong>for</strong>mulation. Mabie, Rogers, and Reinholtz [3] developed a numerical<br />
procedure to determine the individual hob offsets <strong>for</strong> a pair of gears to obtain maximum ratio of recess to<br />
approach action, to balance tooth strength of the pinion and gear, to maintain the desired contact ratio, and<br />
to avoid undercutting. Liou [4] tried to balance dynamic tooth stresses on both pinion and gear by adjusting<br />
hob offset values when cutting non-standard gears of long and short addendum design.<br />
<strong>Design</strong>ing compact (minimum size) gear sets provides benefits such as minimal weight, lower material cost,<br />
1
smaller housings, and smaller inertial loads. Gear designs must satisfy constraints, such as bending strength<br />
limits, pitting resistance, and scoring. Many approaches <strong>for</strong> improved gear design have been proposed in<br />
previous literature of references [5] to [16]. Previous research presented different approaches <strong>for</strong> optimal<br />
gear design. Savage et al. [12] considered involute interference, contact stresses, and bending fatigue. They<br />
concluded that the optimal design usually occurs at the intersection point of curves relating the tooth<br />
numbers and diametral pitch required to avoid pitting and scoring. Wang et al. [13] expanded the model to<br />
include the AGMA geometry factor and AGMA dynamic factor in the tooth strength <strong>for</strong>mulas. Their analysis<br />
found that the theoretical optimal gear set occurred at the intersection of the bending stress and contact<br />
stress constraints at the initial point of contact.<br />
More recently, the optimal design of gear sets has been expanded to include a wider range of<br />
considerations. Andrews et al. [14] approached the optimal strength design <strong>for</strong> nonstandard gears by<br />
calculating the hob offsets to equalize the maximum bending stress and contact stress between the pinion<br />
and gear. Savage et al. [15] treated the entire transmission as a complete system. In addition to the gear<br />
mesh parameters, the selection of bearing and shaft proportions were included in the design configuration.<br />
The mathematical <strong>for</strong>mulation and an algorithm are introduced by Wang et al. [16] to solve the multiobjective<br />
gear design problem, where feasible solutions can be found in a three-dimensional solution space.<br />
However, none of them was <strong>for</strong> compact design study.<br />
Most of the <strong>for</strong>egoing literature dealt primarily with static tooth strength of standard gear geometries. These<br />
studies use the Lewis <strong>for</strong>mula assuming that the static load is applied at the tip of the tooth. Some<br />
considered stress concentration and the AGMA geometry and dynamic factors. However, the operating<br />
speed must be considered <strong>for</strong> dynamic effects. Rather than using the AGMA dynamic factor, which<br />
increases as a simple function of pitch line velocity, a gear dynamics code by Lin et al. [5, 6, and 17] was<br />
used here to calculate a dynamic load factor.<br />
The purpose of the present work is to develop a design procedure <strong>for</strong> a compact size of non-standard spur<br />
gear sets of the long and short addendum system incorporating dynamic considerations. Constraint criteria<br />
employed <strong>for</strong> this investigation include the involute interference limits combined with the tooth bending<br />
strength and contact stress limits.<br />
Model Formulation<br />
<strong>Spur</strong> <strong>Gears</strong> Cut by a Hob Cutter<br />
The following analysis is based on the study of Mabie et al. [3]. Figure 1 shows a hob cutting a pinion where<br />
the solid line indicates a pinion with fewer than the minimum number of teeth required to prevent<br />
interference.<br />
Figure 1 <strong>Spur</strong> gear tooth cut by a standard hob or a withdrawn hob.<br />
The addendum line of the hob falls above the interference point E of the pinion, so that the flanks of the<br />
pinion teeth are undercut. To avoid undercutting, the hob can be withdrawn a distance e, so that the<br />
2
addendum line of the hob passes through the interference point E. This condition is shown as dotted in<br />
Figure 1 and results in the hob cutting a pinion with a wider tooth. As the hob is withdrawn, the outside<br />
radius of the pinion must also be increased (by starting with a larger blank) to maintain the same clearance<br />
between the tip of the pinion tooth and the root of the hob tooth. To show the change in the pinion tooth<br />
more clearly, the withdrawn hob in Figure 1 was moved to the right to keep the left side of the tooth profile<br />
the same in both cases.<br />
The width of the enlarged pinion tooth on its cutting pitch circle can be determined from the tooth space of<br />
the hob on its cutting pitch line. From Figure 2, this thickness can be expressed by the following equation:<br />
p<br />
t = 2etanφ<br />
+<br />
( 1 )<br />
2<br />
Figure 2 A hob cutter with offset e <strong>for</strong> enlarged pinion tooth thickness<br />
Equation (1) can be used to calculate the tooth thickness on the cutting pitch circle of a gear generated by a<br />
hob offset an amount e; e will be negative if the hob is advanced into the gear blank. In Figure 1, the hob<br />
was withdrawn just enough so that the addendum line passed through the interference point of the pinion.<br />
The withdrawn amount can be increased or decreased as desired as long as the pinion tooth does not<br />
become undercut or pointed. The equation that describes the relationship between e and other tooth<br />
geometry is<br />
There<strong>for</strong>e,<br />
e = AB + OA - OP<br />
k<br />
= + Rb<br />
cosφ<br />
- R p<br />
Pd<br />
e= k<br />
P - 2<br />
R p(1-<br />
)<br />
d<br />
e= 1<br />
P<br />
d<br />
(k- N<br />
2<br />
( 2 )<br />
cos φ ( 3 )<br />
)<br />
2<br />
sin φ ( 4 )<br />
Two equations that were developed from involutometry find particular application in this study:<br />
3
cos = cos<br />
R A<br />
φBφ A<br />
( 5 )<br />
RB<br />
t B=2R B(<br />
t A<br />
2 R<br />
A<br />
+inv -inv )<br />
φA φ B<br />
( 6 )<br />
By means of these equations, the pressure angle and tooth thickness at any radius RB can be found if the<br />
pressure angle and tooth thickness are known at a reference radius R A. This reference radius is the cutting<br />
pitch radius, and the tooth thickness on this cutting pitch circle can be easily calculated <strong>for</strong> any cutter offset.<br />
The reference pressure angle is the pressure angle of the hob cutter.<br />
When two gears, gear 1 and gear 2, which have been cut with a hob offset e 1 and e 2, respectively, are<br />
meshed together, they operate on pitch circles of radii R ' 1 and R ' 2 and at pressure angle φ. The thickness of<br />
the teeth on the operating pitch circles can be expressed as t ' 1 and t ' 2, which can be calculated from<br />
Equation (6). These dimensions are shown in Figure 3 together with the thickness of the teeth t 1 and t 2 on<br />
the cutting pitch circles of radii R 1 and R 2.<br />
Figure 3 <strong>Non</strong>-standard gears cut by offset hob cutters and operated at new pressure angle φ’.<br />
To determine the pressure angle φ at which these gears will operate, we found<br />
and<br />
ω2<br />
=<br />
ω<br />
t<br />
'<br />
1<br />
1<br />
t<br />
'<br />
2<br />
N 1 =<br />
N<br />
2<br />
2πR<br />
=<br />
N<br />
R<br />
R<br />
1<br />
'<br />
1<br />
'<br />
1<br />
'<br />
2<br />
2πR<br />
=<br />
N<br />
2<br />
'<br />
2<br />
+ ( 8 )<br />
Substituting Eq. (6) into Eq. (8) and dividing by 2R ' 1,<br />
4<br />
( 7 )
1<br />
[<br />
t<br />
+ ( inv<br />
2 R1<br />
Rearranging this equation gives,<br />
t<br />
2R<br />
1<br />
R<br />
'<br />
' 2 2<br />
'<br />
φ - invφ<br />
) ] [<br />
t<br />
+ ( inv φ - invφ<br />
) =<br />
'<br />
R1<br />
2 R2<br />
π<br />
N<br />
+ ( 9 )<br />
'<br />
'<br />
R2<br />
t2<br />
π R2<br />
'<br />
+ = + ( 1 + )( invφ<br />
− inv )<br />
( 10 )<br />
'<br />
'<br />
R 2R<br />
N R<br />
1 φ<br />
1<br />
2<br />
1<br />
By substituting Eq. (7) and 2R = N/Pd into the above equation and multiplying by N 1/P d ,<br />
π<br />
1 2<br />
t 1+ t 2=<br />
+<br />
P d d<br />
N + N<br />
By substituting Eq. (1) <strong>for</strong> t1 and t 2,<br />
2 e<br />
P<br />
1<br />
(inv -inv )<br />
φ′ φ (11)<br />
p<br />
p π N 1+<br />
N 2<br />
φ + + 2 e2<br />
tan φ + = + (invφ<br />
′ - inv φ )<br />
2<br />
2 Pd<br />
Pd<br />
1 tan ( 12 )<br />
Simplifying the equation above gives,<br />
2 (e + e )+ p= +<br />
P N π + N<br />
tan φ 1 2<br />
d Pd<br />
By substituting p = π /Pd and solving <strong>for</strong> inv φ ' ,<br />
or<br />
1 2<br />
inv = inv + 2 P d( e 1+ e 2)<br />
tan φ<br />
φ′ φ<br />
N 1+ N 2<br />
( N1<br />
+ N 2 )(invφ<br />
′ - inv φ )<br />
e1+<br />
e2<br />
=<br />
2 Pd<br />
tan φ<br />
(inv -inv )<br />
Expressing the above equation in metric unit with m as module,<br />
1 2<br />
e 1+ e 2=<br />
m ( N + N )(inv φ′ -inv φ )<br />
2<br />
tan φ<br />
1<br />
φ′ φ ( 13 )<br />
In this study, we limit our investigation to the case that the hob cutter is advanced into the gear blank the<br />
same amount that it is withdrawn from the pinion, there<strong>for</strong>e, e2 = -e 1 and, from Eq. (15) or (16), φ ‘ = φ.<br />
Because there is no change in the pressure angle, R ' 1 = R 1 and R ' 2 = R 2, and the gears operate at the<br />
standard center distance. If the offsets are unequal (e2 ≠ -e 1), then the center distance must change. This<br />
problem is not considered in the present investigation.<br />
Objective Function<br />
The design objective of this study is to obtain the most compact gear set of the long and short addendum<br />
system satisfying design requirements that include loads and power level, gear ratio and material<br />
parameters. The gears designed must satisfy operational constraints such as avoiding interference, pitting,<br />
scoring distress and tooth breakage. The required gear center distance C is the chosen parameter to be<br />
optimized.<br />
where<br />
C = R + R<br />
p1<br />
p2<br />
5<br />
( 14 )<br />
( 15 )<br />
( 16 )<br />
(17)
R p1 pitch radius of gear 1<br />
R p2 pitch radius of gear 2<br />
<strong>Design</strong> Parameters and Variables<br />
The following table lists the parameters and variables used in this study:<br />
<strong>Design</strong> Constraints<br />
Table 1 Basic gear design parameters and variables<br />
Gear Parameters <strong>Design</strong> Variables<br />
Bending strength and contact strength Number of pinion teeth<br />
limits<br />
Diametral pitch<br />
Operating torque<br />
Operating speed<br />
Gear ratio<br />
Face width<br />
Pressure angle<br />
Hob cutter offset<br />
Involute Interference Involute interference is defined as a condition in which there is an obstruction on the<br />
tooth surface that prevents proper tooth contact (Townsend et al. [18]); or contact between portions of tooth<br />
profiles that are not conjugate (South et al. [19]). Interference occurs when the driven gear contacts a noninvolute<br />
portion (below the base circle) of the driving gear. Undercutting occurs during tooth generation if the<br />
cutting tool removes the interference portion of the gear being cut. An undercut tooth is weaker, less<br />
resistant to bending stress, and prone to premature tooth failure. DANST has a built-in routine to check <strong>for</strong><br />
interference.<br />
Bending Stress Tooth bending failure at the root is a major concern in gear design. If the bending stress<br />
exceeds the fatigue strength, the gear tooth has a high probability of failure. The AGMA bending stress<br />
equation can be found in the paper by Savage et al. [13] and in other gear literature. In this study, a modified<br />
Heywood <strong>for</strong>mula <strong>for</strong> tooth root stress was used. This <strong>for</strong>mula correlates well with the experimental data and<br />
finite element analysis results (Cornell et al. [20]):<br />
Wj<br />
cos β j hf<br />
6l f . h<br />
tan β<br />
07 .<br />
072 L<br />
j<br />
σ j = [ 1+ 026 . ( ) ][ + ( 1 − νtan β j ) − ]<br />
2<br />
F<br />
2R<br />
h h l h h<br />
where<br />
σ root bending stress at loading position j.<br />
j<br />
Wj transmitted load at loading position j.<br />
β load angle, degree<br />
j<br />
F face width of gear tooth, inch<br />
ν approximately 1/4, according to Heywood [21].<br />
Rf fillet radius, inch<br />
f<br />
f f f<br />
other nomenclature is defined in Figure 4 and References [20] and [21]. To avoid tooth failure, the bending<br />
stress should be limited to the allowable bending strength of the material as suggested by AGMA [22],<br />
St K L<br />
σ j ≤ σall<br />
= (19)<br />
K K<br />
T R<br />
where<br />
σ all allowable bending stress<br />
S t AGMA bending strength<br />
K L life factor<br />
K T temperature factor<br />
6<br />
f<br />
f<br />
(18)
K R reliability factor<br />
Kv dynamic factor<br />
Figure 4 Tooth geometry <strong>for</strong> stress calculations using modified Heywood <strong>for</strong>mula.<br />
Surface Stress The surface failure of gear teeth is an important concern in gear design. Surface failure<br />
modes include pitting, scoring, and wear. Pitting is a gear tooth failure in the <strong>for</strong>m of tooth surface cavities as<br />
a result of repeated stress applications. Scoring is another surface failure that usually results from high loads<br />
or lubrication problems. It is defined as the rapid removal of metal from the tooth surface caused by metal<br />
contact due to high overload or lubricant failure. The surface is characterized by a ragged appearance with<br />
furrows in the direction of tooth sliding (Shigley et al. [23]). Wear is a fairly uni<strong>for</strong>m removal of material from<br />
the tooth surface.<br />
The stresses on the surface of gear teeth are determined by <strong>for</strong>mulas derived from the work of Hertz [18].<br />
The Hertzian contact stress between meshing teeth can be expressed as<br />
σ<br />
Hj<br />
=<br />
+<br />
Wj<br />
β j ρ ρ<br />
π F φ − ν ν<br />
+<br />
E E<br />
−<br />
⎡ 1 1 ⎤<br />
cos<br />
⎢<br />
⎥<br />
⎢ 1 2 ⎥<br />
2<br />
2<br />
cos ⎢ 1 1 1 2 ⎥<br />
⎣⎢<br />
⎦⎥<br />
where<br />
σ contact stress at loading position j<br />
Hj<br />
Wj transmitted load at loading position j.<br />
βj load angle, degree<br />
F face width of gear tooth, inch<br />
φ pressure angle, degree<br />
ρ radius of curvature of gear 1,2 at the point of contact, inch<br />
1,2<br />
ν Poisson’s ratio of gear 1,2<br />
1,2<br />
E modulus of elasticity of gear 1,2, psi<br />
1,2<br />
1<br />
2<br />
The AGMA recommends that this contact stress should also be considered in a similar manner as the<br />
bending endurance limit (Shigley et al. [22]). The equation is<br />
7<br />
(20)
σ ≤ σ = S<br />
C C<br />
Hj c, all C<br />
L H<br />
CTCR (21)<br />
where<br />
σc,all allowable contact stress<br />
Sc AGMA surface fatigue strength<br />
CL life factor<br />
CH hardness-ratio factor<br />
CT temperature factor<br />
CR reliability factor<br />
According to Savage et al. [12], Hertzian stress is a measure of the tendency of the tooth surface to develop<br />
pits and is evaluated at the lowest point of single tooth contact rather than at the less critical pitch point as<br />
recommended by AGMA. Gear tip scoring failure is highly temperature dependent (Shigley et al. [23]) and<br />
the temperature rise is a direct result of the Hertz contact stress and relative sliding speed at the gear tip.<br />
There<strong>for</strong>e, the possibility of scoring failure can be determined by Equation (20) with the contact stress<br />
evaluated at the initial point of contact. A more rigorous method not used here is the PVT equation or the<br />
Blok scoring equation (South et al. [18]).<br />
Dynamic Load Effect One of the major goals of this work is to study the effect of dynamic load on optimal<br />
gear design. The dynamic load calculation is based on the NASA gear dynamics code DANST developed by<br />
the authors. DANST has been validated with experimental data <strong>for</strong> high-accuracy gears at NASA Lewis<br />
Research Center (Oswald et al. [24]). DANST considers the influence of gear mass, meshing stiffness, tooth<br />
profile modification, and system natural frequencies in its dynamic calculations.<br />
The dynamic tooth load depends on the value of relative dynamic position and backlash of meshing tooth<br />
pairs. After the gear dynamic load is found, the dynamic load factor can be determined by the ratio of the<br />
maximum gear dynamic load during mesh to the applied load. The applied load equals the torque divided by<br />
the base circle radius. This ratio indicates the relative instantaneous gear tooth load. <strong>Compact</strong> gears<br />
designed using the dynamic load calculated by DANST will be compared with gears designed using the<br />
AGMA suggested dynamic factor, which is a simple function of the pitch line velocity.<br />
Gear <strong>Design</strong> Application<br />
<strong>Design</strong> Algorithm<br />
An algorithm was developed to per<strong>for</strong>m the analyses and find the optimum design of non-standard gears of<br />
the long and short addendum system (LASA). The computer algorithm of the dynamic optimal design can<br />
start with the pinion tooth number loop. Within each pinion tooth number loop, all of the basic parameters are<br />
needed in dynamic analysis, including the diametral pitch, addendum ratio, dedendum ratio, gear material<br />
properties, transmitted power (torque), face width, and hob offset value. The DANST code also calculates<br />
additional parameters that are needed in dynamic analysis, such as base circle radius, pitch circle radius,<br />
addendum circle radius, stiffness, and the system natural frequencies, etc. One of the system natural<br />
frequencies (the one that results in the highest dynamic stresses) will be used as pinion operating speed.<br />
For this study, the diametral pitch was varied from two to twenty. Static analysis was first per<strong>for</strong>med to check<br />
<strong>for</strong> involute interference and to calculate the meshing stiffness variations and static transmission errors of the<br />
gear pair. If there was a possibility of interference, the number of pinion teeth was increased by one and the<br />
static process was repeated. The number of gear teeth was also increased according to the specified gear<br />
ratio. Results from the static analyses were incorporated in the equations of motion of the gear set to obtain<br />
the dynamic motions of the system. Instantaneous dynamic load at each contact point along the tooth profile<br />
was determined from these motions. The varying contact stresses and root bending stresses during the gear<br />
mesh were calculated from these dynamic loads.<br />
If all the calculated stresses were less than the design stress limits <strong>for</strong> a possible gear set, the data <strong>for</strong> this<br />
set were added to a candidate group. At each value of the diametral pitch, the most compact gear set in the<br />
candidate group would have the smallest center distance. These different candidate designs can be<br />
compared in a table or graph to show the optimum design from all the sets studied.<br />
The analyses above are <strong>for</strong> gears operating at a single speed. To examine the effect of varying speed, the<br />
analyses can be repeated at different speeds. The program will stop when all the stresses are below the<br />
8
design constraints. Thus, a feasible optimal gear set operating at the system natural frequency can be<br />
obtained. As the diametral pitch varies, the optimal gear sets determined from each diametral pitch value are<br />
collected to <strong>for</strong>m the optimal design space. This design space, when transferred into the center distance<br />
space, which is our merit function of optimal design, will indicate all the feasible optimal gear sets to be<br />
selected from with given operating parameters.<br />
<strong>Design</strong> Applications<br />
Table 2 shows the basic gear parameters <strong>for</strong> the sample gear sets to be studied. They were first used in a<br />
gear design problem by Shigley and Mitchell [19], and later used by Carroll and Johnson [13] as an example<br />
<strong>for</strong> optimal design of compact gear sets. The sample gear set transmits 100 horsepower at an input speed of<br />
1120 rpm. The gear set has standard full depth teeth and a speed reduction ratio of 4. In this study, the face<br />
width of the gear is always chosen to be one-half the pinion pitch diameter. In other words, the length to<br />
diameter ratio (λ) is 0.5.<br />
Table 2 Basic design parameters of sample gear sets<br />
Pressure Angle, φ (degree) 20<br />
Gear Ratio, M g 4.0<br />
Length to Diameter Ratio, λ 0.5<br />
Transmitted Power (hp) 100.0<br />
Applied Torque (lb-in) 5627.264<br />
Input Speed (rpm) 1120.0<br />
Modulus of Elasticity (Mpsi) 30<br />
Poisson’s Ratio, ν 0.3<br />
Scoring and Pitting Stress Limits, S S and S P (Kpsi) 79.23<br />
Bending Stress Limit, S b (Kpsi) 19.81<br />
Hob Cutter Offset (in) 0.03, 0.05, 0.07, 0.09<br />
Cutting gears with offset hobs is an effective way to balance the dynamic tooth strength of the pinion and<br />
gear. It reduces the dynamic stress in the pinion and increases stress in the gear to achieve balance. In<br />
general, increasing the offset improves the balance in dynamic tooth strength. However, the best hob offset<br />
varies with the transmission speed and is limited by the maximum allowable offset that renders the pinion<br />
tooth pointed. The hob offset amount is varied from 0.03 inch to 0.09 inch with an increment of 0.02 inch in<br />
this investigation.<br />
Table 3 displays Carroll’s [13] optimal design results <strong>for</strong> the sample gears of standard system without hob<br />
cutter offset. The optimal design is indicated in bold type and by an arrow. In the table, P d is the diametral<br />
pitch, NT1 and NT2 represent the number of teeth of pinion and gear, respectively, CD is the center<br />
distance, FW is the face width, CR is the contact ratio, and S b, S S, and S P are the stress limit <strong>for</strong> bending,<br />
scoring, and pitting, respectively.<br />
Table 3 Carroll’s optimization results of sample standard gear set [13]<br />
(Using Lewis tooth stress <strong>for</strong>mula)<br />
Pd NT1 NT2 CD FW CR Sb S s S p<br />
2.00 19 76 23.750 4.750 1.681 2.872 72.497 51.396<br />
2.25 20 80 22.222 4.444 1.691 3.553 72.162 55.853<br />
2.50 21 84 21.000 4.200 1.701 4.275 72.532 59.950<br />
3.00 23 92 19.167 3.833 1.717 5.820 74.100 67.213<br />
4.00 27 108 16.875 3.375 1.745 9.202 77.876 78.860<br />
6.00 40 160 16.667 3.333 1.805 12.703 66.972 78.309<br />
8.00 53 212 16.563 3.313 1.840 16.191 63.302 78.247<br />
10.00 66 264 16.500 3.300 1.863 19.670 61.463 78.285 ←<br />
12.00 86 344 17.917 3.583 1.887 19.727 53.219 69.603<br />
16.00 132 528 20.625 4.125 1.917 19.726 42.459 57.137<br />
20.00 185 740 23.125 4.625 1.934 19.742 35.708 48.760<br />
The theoretical optimum <strong>for</strong> this example occurs at the intersection of bending stress and contact stress<br />
constraint curves at the lowest point of single tooth contact. This creates a gear set that has NT1 = 64 and<br />
P d = 9.8 <strong>for</strong> a theoretical center distance of 16.333 in. The minimum practical center distance (16.50 in.) is<br />
9
obtained when NT1 = 66 and P d = 10.0.<br />
For comparison with the above results, we used the same AGMA dynamic factor K v (Equation 6) but with the<br />
modified Heywood tooth bending stress <strong>for</strong>mula (Eq. 2) in the calculations. Table 4 lists the optimization<br />
results obtained.<br />
Table 4 Optimization results of sample standard gears<br />
(Using modified Heywood tooth stress <strong>for</strong>mula)<br />
Pd NT1 NT2 CD FW CR Sb S s S p<br />
2.00 19 76 23.750 4.750 1.681 2.847 66.873 52.390<br />
2.25 19 76 21.111 4.222 1.681 3.935 78.539 61.589<br />
2.50 20 80 20.000 4.000 1.691 4.718 76.864 65.812<br />
3.00 22 88 18.333 3.667 1.709 6.384 76.054 73.234<br />
4.00 28 112 17.500 3.500 1.751 8.466 66.656 76.216<br />
6.00 41 164 17.083 3.417 1.808 12.215 58.956 77.066<br />
8.00 54 216 16.875 3.375 1.842 15.785 56.336 77.689<br />
10.00 67 268 16.750 3.350 1.865 19.369 55.015 78.137 ←<br />
12.00 87 348 18.125 3.625 1.888 19.637 47.915 69.817<br />
16.00 134 536 20.938 4.188 1.918 19.638 38.076 57.045<br />
20.00 188 752 23.500 4.700 1.935 19.718 32.006 48.616<br />
As can be seen from the table, the minimum practical center distance (16.750 in.) is obtained when NT1 =<br />
67 and P d = 10.0. This is very close to Carroll’s design but his optimal gear set will exceed the design limit of<br />
19.81 Kpsi (from Table 2) <strong>for</strong> maximum bending stress on the pinion according to our calculations. The<br />
differences between Carroll’s results and those reported here are likely due to the use of different <strong>for</strong>mulas<br />
<strong>for</strong> bending stress calculations.<br />
Figure 5 shows graphically the design space <strong>for</strong> the results presented in Table 4, depicting the stress<br />
constraint curves of bending, scoring, and pitting. The region above each constraint curve indicates feasible<br />
design space <strong>for</strong> that particular constraint. In the figure, the theoretical optimum is located at the intersection<br />
point of the scoring stress and the bending stress constraint.<br />
Figure 5 <strong>Design</strong> space <strong>for</strong> determination of optimized results of sample standard gears using modified<br />
Heywood tooth stress <strong>for</strong>mula.<br />
10
Table 5 displays the data <strong>for</strong> the optimum compact design of the sample non-standard gears (long and short<br />
addendum system) with a hob offset of 0.03 inch. The practical minimum center distance is equal to 15.000<br />
inches and occurs at two values of P d = 10.00 and 12.00, with NT1 = 60 and 72. The practical minimum<br />
center distance decreases as the diametral pitch P d increases from 2.00 to 10.00; it then increases as the<br />
diametral pitch P d increases from 12.00 to 20.00.<br />
Table 5 Optimization results <strong>for</strong> non-standard gears (LASA) with 0.03 inch hob offset<br />
Pd NT1 NT2 CD FW CR Sb S s S p<br />
2.00 20 80 25.000 5.000 1.691 2.684 71.166 23.861<br />
3.00 25 100 20.833 4.167 1.732 5.325 74.722 27.375<br />
4.00 30 120 18.750 3.750 1.762 8.534 79.217 25.057<br />
6.00 39 156 16.250 3.250 1.801 12.526 78.846 60.440<br />
8.00 50 200 15.625 3.125 1.833 16.074 76.410 63.863<br />
10.00 60 240 15.000 3.000 1.854 21.053 78.169 67.826←<br />
12.00 72 288 15.000 3.000 1.872 24.637 76.021 68.245←<br />
16.00 107 428 16.719 3.344 1.904 25.312 61.583 57.990<br />
20.00 151 604 18.875 3.775 1.924 25.647 51.281 49.269<br />
In Table 6, when the hob offset is increased to 0.05 inch, the optimum compact design occurs at P d = 10.0<br />
with a practical minimum center distance of 15.750 inch and NT1 = 63,. As can be seen from the table, the<br />
minimum center distance decreases as the diametral pitch increases from 2.00 to 10.00. When the diametral<br />
pitch is greater than 10, the minimum center distance increases with the diametral pitch value.<br />
Table 6 Optimization results <strong>for</strong> non-standard gears with 0.05 inch hob offset<br />
P d NT1 NT2 CD FW CR S b S s S p<br />
2.00 19 76 23.750 4.750 1.681 3.005 78.361 26.984<br />
3.00 25 100 20.833 4.167 1.732 5.296 73.095 29.197<br />
4.00 30 120 18.750 3.750 1.762 8.611 77.748 30.469<br />
6.00 42 168 17.500 3.500 1.811 14.076 78.573 30.307<br />
8.00 52 208 16.250 3.250 1.838 17.460 75.663 61.022<br />
10.00 63 252 15.750 3.150 1.859 23.580 78.122 65.023←<br />
12.00 79 316 16.458 3.292 1.880 25.628 71.270 62.258<br />
16.00 127 508 19.844 3.969 1.915 25.293 54.128 49.583<br />
20.00 196 784 24.500 4.900 1.936 25.309 41.570 38.476<br />
The optimization results <strong>for</strong> hob offsets of 0.07 and 0.09 inches are very similar. For both cases the practical<br />
minimum center distance is equal to 17.500 and occurs at P d = 6.00 and 8.00, with NT1 = 42 and 56,<br />
respectively. Minimum center distance decreases as the diametral pitch increases from 2.00 to 6.00; it then<br />
increases as diametral pitch increases from 8.00. At P d = 6.00 and 8.00, the minimum center distance <strong>for</strong><br />
both cases is unchanged. Some higher diametral pitch values do not produce optimization results due to the<br />
pointed teeth.<br />
Comparisons between non-standard gear designs at different hob offset values are shown in Figures 6(a)<br />
and 6(b). The compact design improves significantly as P d value increases from 2.00 to 4.00 <strong>for</strong> nonstandard<br />
gears and from 2.00 to 6.00 <strong>for</strong> standard gears.<br />
11
Pinion Tooth Number<br />
Center Distance (in.)<br />
210<br />
190<br />
170<br />
150<br />
130<br />
110<br />
9<br />
7<br />
5<br />
3<br />
1<br />
29<br />
27<br />
25<br />
23<br />
21<br />
19<br />
17<br />
15<br />
13<br />
11<br />
9<br />
7<br />
5<br />
Feasible<br />
<strong>Design</strong><br />
2.00 3.00 4.00 6.00 8.00 10.00 12.00 16.00 20.00<br />
Hob Offset<br />
Diametral Pitch<br />
0.03 in 0.05 in 0.07 in<br />
(a)<br />
Feasible<br />
<strong>Design</strong><br />
0.09 in<br />
2.00 3.00 4.00 6.00 8.00 10.00 12.00 16.00 20.00<br />
Diametral<br />
Hob Offset 0.03 in 0.05 in 0.07 in 0.09 in<br />
Figure 6 <strong>Design</strong> space and optimal gear sets <strong>for</strong> standard and non-standard gears, Gear ratio M g = 4,�<br />
pressure angle φ = 20 0 .<br />
(b)<br />
12
At diametral pitch values less than or equal to 3.00, non-standard gears with hob offset values of 0.07 inches<br />
and 0.09 inches appear to be more compact than standard gears while non-standard gears with hob offset<br />
values of 0.03 inches and 0.05 inches have the same center distance as standard gears (except nonstandard<br />
gears pair at P d = 2.00 with hob offset 0.05 inch that has slightly smaller center distance than<br />
standard gears pair).<br />
At a diametral pitch value of 4.00, non-standard gears and standard gears have the same compact design.<br />
At diametral pitch that is greater than 6.00, standard gears appear more compact than non-standard gears.<br />
The difference is extremely large at high diametral pitch values and high hob offset values. Comparisons<br />
between non-standard gears indicate that the curves in this study are almost identical to the one in Study 8.<br />
There<strong>for</strong>e, similar conclusions are derived in this study.<br />
From Figure 6(b), hob offset values of 0.07 inches and 0.09 inches seem to result in better compact design<br />
than hob offsets of 0.03 inches and 0.05 inches when diametral pitch is less than or equal to 4.00. With<br />
diametral pitches that are greater than 4.00, non-standard gears with hob offset values of 0.07 inches and<br />
0.09 inches become less compact when compared to non-standard gears at smaller hob offset values.<br />
Generally, at a diametral pitch value of 4.00 or less, increasing the hob offset value will result in a more<br />
compact design. When the diametral pitch is greater than 4.00, increasing the hob offset value will result in a<br />
larger gears center distance or a less compact design. When the diametral pitch value is high (greater than<br />
or equal to 10.00), non-standard gears at higher hob offsets (0.07 inches and 0.09 inches) seem very<br />
sensitive to subject load which result in a much larger gear pair. Thus, the degree of sensitivity increases<br />
with hob offset value at high diametral pitch values. As can be seen in the plots, because of the sensitivity in<br />
geometry, we could not obtain results <strong>for</strong> non-standard gears with hob offset values of 0.07 inches at P d =<br />
20.00 and 0.09 inches at P d = 16.00 and 20.00.<br />
Conclusions<br />
This paper presents a method <strong>for</strong> optimal design of standard spur gears <strong>for</strong> minimum dynamic response. A<br />
study was per<strong>for</strong>med using a sample gear set from the gear literature. Optimal gear sets were compared <strong>for</strong><br />
designs based on the AGMA dynamic factor and a refined dynamic factor calculated using the DANST gear<br />
dynamics code. The operating speed was varied over a broad range to evaluate its effect on the required<br />
gear size. A design space <strong>for</strong> designing optimal compact gear sets can be developed using the current<br />
approach.<br />
The required size of an optimal gear set is significantly influenced by the dynamic factor. The peak dynamic<br />
factor at system natural frequencies can dominate the design of optimal gear sets that operate over a wide<br />
range of speeds. In the current investigation, refined dynamic factors calculated by the dynamic gear code<br />
developed by the authors allow a more compact gear design than that obtained by using the AGMA dynamic<br />
factor values.<br />
<strong>Compact</strong> gears designed using the modified Heywood tooth stress <strong>for</strong>mula are similar to those designed<br />
using the simpler Lewis <strong>for</strong>mula <strong>for</strong> the example cases studied here. <strong>Design</strong> charts such as those shown<br />
here can be used <strong>for</strong> a single speed or over a range of speeds. For the sample gears in the study, a<br />
diametral pitch of 10.0 was found to provide a compact gear set over the speed range considered.<br />
13
References<br />
1. Walsh, E. J. and Mabie, H. H., “A Simplified Method <strong>for</strong> Determining Hob Offset Value in the <strong>Design</strong> of<br />
<strong>Non</strong>standard <strong>Spur</strong> <strong>Gears</strong>”, 2nd OSU Applied Mechanism Conference, Stillwater, Oklahoma, (Oct. 1971).<br />
2. Mabie, H. H., Walsh, E. J., and Bateman, “Determination of Hob Offset Required to Generate<br />
<strong>Non</strong>standard <strong>Spur</strong> <strong>Gears</strong> with Teeth of Equal Strength”, Mechanism and Machine Theory, Vol. 18, No. 3, pp.<br />
181-192, (1983).<br />
3. Mabie, H. H., Roggers, C. A., and Reinholtz, C. F., “<strong>Design</strong> of <strong>Non</strong>standard <strong>Spur</strong> <strong>Gears</strong> Cut by a Hob”,<br />
Mechanism and Machine Theory, Vol. 25, No. 6, pp. 635-644, (1990).<br />
4. Liou, C. H., “<strong>Design</strong> of <strong>Non</strong>standard <strong>Spur</strong> <strong>Gears</strong> <strong>for</strong> Balanced Dynamic Strength,” Ph.D. Dissertation,<br />
University of Memphis, (1994).<br />
5. Lin, H.H., Townsend, D.P., and Oswald, F.B., 1989, “Profile Modification to Minimize <strong>Spur</strong> Gear<br />
Dynamic Loading,” Proc. of ASME 5th Int. Power Trans. and Gearing Conf., Chicago, IL, Vol. 1,<br />
pp. 455–465.<br />
6. Liou, C.H., Lin, H.H., and Oswald, F.B., 1992, “Effect of Contact Ratio on <strong>Spur</strong> Gear Dynamic Load,”<br />
Proc. of ASME 6th Int. Power Trans. and Gearing Conf., Phoenix, AZ, Vol. 1, pp. 29–33.<br />
7. Bowen, C.W., 1978, “The Practical Significance of <strong>Design</strong>ing to Gear Pitting Fatigue Life Criteria,”<br />
ASME Journal of Mechanical <strong>Design</strong>, Vol. 100, pp. 46–53.<br />
8. Gay, C.E., 1970, “How to <strong>Design</strong> to Minimize Wear in <strong>Gears</strong>,” Machine <strong>Design</strong>, Vol. 42, pp. 92–97.<br />
9. Coy, J.J., Townsend, D.P., and Zaretsky, E.V., 1979, “Dynamic Capacity and Surface Fatigue Life <strong>for</strong><br />
<strong>Spur</strong> and Helical <strong>Gears</strong>,” ASME Journal of Lubrication Technology, Vol. 98, No. 2, pp. 267–276.<br />
10. Anon, 1965, “Surface Durability (Pitting) of <strong>Spur</strong> Gear Teeth,” AGMA <strong>Standard</strong> 210.02.<br />
11. Rozeanu, L. and Godet, M., 1977, “Model <strong>for</strong> Gear Scoring,” ASME Paper 77-DET-60.<br />
12. Savage, M., Coy, J.J., and Townsend, D.P., 1982, “Optimal Tooth Numbers <strong>for</strong> <strong>Compact</strong> <strong>Standard</strong><br />
<strong>Spur</strong> Gear Sets,” ASME Journal of Mechanical <strong>Design</strong>, Vol. 104, pp. 749–758.<br />
13. Carroll, R.K. and Johnson, G.E., 1984, “Optimal <strong>Design</strong> of <strong>Compact</strong> <strong>Spur</strong> Gear Sets,” ASME Journal<br />
of Mechanisms, Transmissions, and Automation in <strong>Design</strong>, Vol. 106, pp. 95–101.<br />
14. Andrews, G.C. and Argent, J.D., 1992, “Computer Aided Optimal Gear <strong>Design</strong>,” Proc. of ASME 6th<br />
Int. Power Trans. and Gearing Conf., Phoenix, AZ, Vol. 1, pp. 391–396.<br />
15. Savage, M., Lattime, S.B., Kimmel, J.A., and Coe, H.H., 1992, “Optimal <strong>Design</strong> of <strong>Compact</strong> <strong>Spur</strong><br />
Gear Reductions,” Proc. of ASME 6th Int. Power Trans. and Gearing Conf., Phoenix, AZ, Vol. 1,<br />
pp. 383–390.<br />
16. Wang, H.L. and Wang, H.P., 1994, “Optimal Engineering <strong>Design</strong> of <strong>Spur</strong> Gear Sets,” Mechanism and<br />
Machine Theory, Vol. 29, No. 7, pp. 1071–1080.<br />
17. Lin, H.H., Wang, J., Oswald, F.B., and Coy, J.J., 1993, “Effect of Extended Tooth Contact on the<br />
Modeling of <strong>Spur</strong> Gear Transmissions,” AIAA-93–2148.<br />
18. Townsend, D.P., 1992, Dudley’s Gear Handbook, 2nd edition, McGraw-Hill Inc.<br />
19. South, D.W. and Ewert, R.H., 1992, Encyclopedic Dictionary of <strong>Gears</strong> and Gearing, McGraw-Hill.<br />
20. Cornell, R.W., 1981, “Compliance and Stress Sensitivity of <strong>Spur</strong> Gear Teeth,” ASME Journal of<br />
Mechanical <strong>Design</strong>, Vol. 103, pp. 447–459.<br />
21. Heywood, R.B., 1952, <strong>Design</strong>ing by Photoelasticity, Chapman and Hall, Ltd.<br />
14
22. Shigley, J.E. and Mitchell, L.D., 1983, Mechanical Engineering <strong>Design</strong>, 4th Ed., McGraw-Hill.<br />
23. Shigley, J.E. and Mischke, C.R., 1989, Mechanical Engineering <strong>Design</strong>, 5th Ed., McGraw-Hill.<br />
24. Oswald, F.B., Townsend, D.P., Rebbechi, B., and Lin, H.H., 1996, “Dynamic Forces in <strong>Spur</strong> <strong>Gears</strong> -<br />
Measurement, Prediction, and Code Validation,” Proc. of ASME 7th Int. Power Trans. and Gearing Conf.,<br />
San Diego, CA, pp. 9–15<br />
15