Sei sulla pagina 1di 24

The Engineering

Resouff(;e For
AdvancingMobility 400 COMMONWEALTH DRIVE WARRENDALE. PA 15096
851512
Production Measurement Technique for
In=Process Control of Spiral Bevel Gearing
A. J. Lemanski
Sikorsky Aircraft Div.
United Technology Corp.
1985 International Off-Highway & Powerplant
Congress & Exposition
MECCA, Milwaul<ee, Wisconsin
September 9-12,1985
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
The appearance of the code at the bottom of the first page of this paper indicates
SAE's consent that copies of the paper may be made for persona! or i n t r n ~ use, or
for the personal or internal use of specific clients. This consent is given on !he con-
dition, however, Ihat the copier pay the stated per article copy fee through the
Copyright Clearance Center, Inc., Operations Center, P.O. Box 765, Schenectady,
N.Y. 12301, for copying beyond thai permitted by Sections 107 or 108 at the U.S.
Copyright Law. This consent does not extend to other kinds of copying such as
copying for general distribution, for advertising or promotional purposes, tor
creating new collective works, or tor resale.
Papers published prior to 1978 may also be copied at a per paper fee of $2.50
under the above stated conditions.
SAE routinely stocks printed papers for a period of three years following date of
publication. Direct your orders to SAE Ordp.r Department.
To obtain quantity reprint rates, permission to reprint a technical paper or per-
mission to use copyrighted SAE publications in other works, contact the SAE
Publications Division.
No part of this publication may be reproduced in any form, in
an electronic retrieval system or otherWise, without the prior
written permission of the publisher.
ISSNOI48 7191
Copyright 1985Societyof Automotive Engineers,lnc.
This paper is subject to revision. Statements and opinions ad"
vam'cd in papers or discussion arc the aethor's and arc his
responsibility, not SAl's; however, the paper has been edited
by SAt for uniform styling and format. Discussion will be
printed \vith thc paper if it is published in SAl Transactions.
ror permission to publish this paper in full or in part, contact
the SAl Publications Division.
Persons wishinl;! to submit papers to be considered for pre-
sent;Jtion or publication through SAE should send the manu
script or a 300 word abstract of a proposed Ol;Jnuscript to:
Secretary, Engineering Activity Board, SAL
Printed in U.S.A.
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
851572
Production Measurement Technique for
In-Process Control of Spiral Bevel Gearing
A. J. Lemanski
Sikorsky Aircraft Div.
United Technology Corp.
Abstract
An advanced measuring technique for in-process inspection of spiral bevel
gearing utilizing a computer controlled multi-axis precision coordinate mea-
suring machine has been developed at Sikorsky Aircraft under sponsorship of the
U.S. Army Aviation Systems Command, St. Louis, ~ l out of AVSCON Propulsion
Laboratory, Cleveland, Ohio.
This paper describes both the current state-of-the-art inspection tech-
nique and an advanced production technique for the in-process inspection and
manufacture of high quality spiral bevel gearing. The inspection method
currently used for the control of spur and helical tooth profiles is not
feasible for spiral bevel gears because the tooth shape and thickness varies
over its face width instead of being constant as in the case of spur and
helical gears. Spiral bevel gears are currently inspected on Gleason test
machines which provide rotating check simulating no-load operation under
specified gearbox mounting conditions. Tooth contact patterns are observed by
painting the teeth with a marking compound and running the gear with its mating
master control gear in a Gleason tester with a light brake load. The gear
typically exhibits a localized composite tooth bearing contact pattern \",hich,
ideally, should spread out under the operating load, filling the working area
of the tooth. The inspection task is a subjective one to ascertain an accept-
able full load pattern from a localized composite pattern. The machine oper-
ators task is even more difficult because he must make a judgment to change the
machine settings in order to correct an undesirable feature in the test pat-
tern.
The advance production measuring technique involves mapping of spiral
teeth over their entire working surfaces using a precision coordinate measuring
machine and quantitatively comparing surface normals with nominal master gear
values or \",ith theoretical values at some 45 grid points. In addition, this
technique features a means for rapidly calculating corrective grinding or
cutting machine setting changes for controlling the tooth profile geometry
within specified tolerance requirements.
Both the gear designer and the gear manufacturing Engineer will have more
options in the analysis and manufacture of bevel gearing due to the positive
control of the tooth profile geometry and related gear dimensions as made
possible by the advanced technique.
Introduction
The inspection of tooth profiles that is commonly performed on spur and
helical gears is not feasible for spiral bevel gears because the tooth shape
and thickness varies over its face width instead of being constant as in the
case of spur and helical gears. Spiral Bevel Gears are currently inspected on
Gleason test machines which provide a rotating check simulating no-load opera-
tion under actual gearbox mounting conditions. Tooth contact patterns under
these conditions can be observed by painting the teeth with a marking compound
and running the gear with its mating master control gear for a few seconds in
the Gleason tester with a light brake load. The gears typically exhibit a
localized composite tooth bearing contact pattern which, ideally, should spread
out under the operating load) fill ing the working area of the tooth. The
inspection task is a subjective one to ascertain an acceptable full load
pattern from a localized composite pattern. The machine operators task is even
01487191/85/09091572$02.50
Copyright 1985 Society of Automotive Engineers, Inc.
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
2
more difficult because he must. make a judgment to change the machine settings
in order to correct an undesirable feat lire in the test pattern. This
faeturing process for spiral bevel gears is more time consuming and costly ,,,,hen
compared to the spur and helical process of control.
Experience has shown that highly londed aircraft spiral bevel gears can
develop a tooth scoring failure during production acceptance tests of trans-
missions even though the tooth contact pattern loJas within specification limits
as determined by the present method of control.
The current bevel gear met.hod requires the use of master control gears
which are manufactured by the same process used for the production gears.
Also, duplicate masters are required for second and third source production and
additional levels of master gear control. From a metrology point of view, it
is most desirable to manufacture the control master gears to a higher level of
accuracy which is currently not possible by the present method.
State Of The Art Of Spiral Bevel
Gear
Spiral bevel gears are currently inspected on a Gleason
test machine, shown in Figure 1, \l"hich provides a rotating test of the gear
its control master gear simulating no-load operation under actual gearbox
mounting conditions. Tooth contact patterns under these rotating conditions
are observed by painting the teeth a marking compound and running the
gears with their mating master control gcars for a few seconds in the gear
tester under a light brake load. Because of the compound curvatures inherent
in the spiral bevel gear tooth form and the profile modifications designed into
the tooth geometry, these gears typically exhibit a localized composite tooth
bearing pattern (Figure 2) v:hich; ideally, should spread out under full load,
filling the twrking area of the tooth with some easing off at the end areas of
contact. The size, shape, and pOSition of this tooth bearing pattern is a
gross indication of the tooth topology both up and dotm the tooth profile and
lengthwise along the tooth face.
FIGURE 1 GLEASON ROLLING TEST MACHINE
851572
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
851572
PATTERN TAPE RECORD
!.STER GEAR PH/l011
WnRAl BEARlNG - DRIVE

9f'bn
wnRAL BEARHlG COAST
--
AM
..,..::
--
FIGURE 2 SPIRAL BEVEL GEAR TOOTH CONTACT PATTERN
The goal of the design and dC\'c]oplIIent phase of spiral bevel gear
manufacture is to obtain a localized test machine pattern of a size, shape, and
location that will produce the desired full load pattern in the gearbox. The
goal of the gear production phase consists of dupl iC<1ting the desired tooth
shape during a production run and from Dill' prodllCl iOIl run to another.
This current method of manufacturin
h
prim,lry drive spiral bevel gears
requires an experienced and qual ificd orgiln iZ:Jl ion. I t is often expressed that
the development of a spiral bevel gCill- of all art. than a science. This
expression is based on the requirement for ski bevel gear machine operators
who must use their background expcricllcp to pV.lluate the position, shape and
contour of the gear tooth pattern prouucl'd by the rolling test in the test
machine. The machine operator's is relied upon to determine what
machine setting or combination of is best used to correct an un-
desirable featured in the test pattern.
The Gleason gear grinding is ;j culmin3tion of motions and tool
paths that generate the bevel gear tooth !orm inte i1 continually varying
non-involute curve. Basically, thr, G11\<'15011 gl'ilr grinder, shown in Figure 3,
has a cradle that supports the formed grillding l.\'hecl shm.;n and has a radial
oscillating motion while the ill ilnd out of the gear tooth space.
FIGURE 3 GLEASON HYPOID GEAR GRINDER
3
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
4
This cradle motion is controlled by <l genl'raL illg cam that can be adjusted
through the cradle angle setting to modify Lill' r,\tio of motion at one end of
the oscillating arc in relation to til(' othl'!" r-nd. 1'11{' gear to be ground is
mounted on a work holding fixLUrf' Cl'lll,:red to the work spindle that
is in constant rotational motiOIl ill iJ cOllt.l"ollld Tatio La the cradle. The
grinding wheel is mounted concentric l(, till' crildll' axis (Figure 4) in a fixed
relative position to the cradle center depl'lldpllt upon the wheel radius, the
spiral angle, and hand of spiral. Th,< grindillg 1,..heel. in effect, acts as a
single tooth of an imaginary general illg gPilr TIll' t.;hcel is dressed automati-
cally at prescribed stages in thf' griudillg :-'1'lj1H'IICI' to muinr.ain surface finish
and profile accuracy.
_Cradle 1 - I,;radle
Center I

I ,jr,.aglnary Generatlng Gear


\ I ;.. : rlean Cone iJ15tanCe

Ii ..; 11ean SOlral Antie


,Gear Slan' "
!
/ \ ! "'. \ r t : Cu:;:er RadlUS
'\ I \ \ et : Sa51e Cradie ';nole

-............::_\'\.,f.' I, 11: M091e"


,/!" i "\ ,! I:Vertleal OlStance of Cutter
// '''',}''/ I Center frem Cradle Center
V 1('0 -----,- H : HOr);:ontal Ogtance of Cutter
<', \ /i \ / Center fre.- Cradle Center
5 I ' ' S $ Radial DiStance of Cutter
. /1 , 1 Center fro", Cradle Center
.... _ IOhee
/ r
N
-i Center 1:2 : riachine Ettentr;t Constant
Ettentric Pivot
FIGURE 4 GRINDING MACHINE GEOMETRY
The geometry and nomenclature of il :-.pirill !w\'cl gear set is shown in
Figure 5.
851572
Dedendum
Angle
Pitch Diameter.
A
OUISi\de Diameler v.
:j
",f
Q'<J
Back
Axial Section ,,0
iT
1','"
FIGURE 5 BEVEL GEAR NOMENCLATURE
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
851572
Gleason gear grinding miJchine settillg in'.-alve first, second, and
third order changes. First. order dfltct ilf'cl and toe pattern position
as well as top and flank position. Tllesl' are u5ed in the final posi-
tioning of the tooth p<lttern. St'CDlld orlkr changes include bias (diagonal
movement) changes. prof i 1c. changp alld ".;hpi': d i I1mlltc-r changes. Third order
changes include wheel dresser illlt! IIp\l ilnd t.oe length changes. There
are approximately fourteen ffinchitlP SI'tt Illg.... th,lL arc used by the machine
operator in first order changes thill afftct till' shap' and position of the gear
tooth pattern. Second and tit i rd ord, r rNjll i rc Q Cil leu lation of values I
using formulas provided by the Glt'ilsOll Ir,"orb., by il gear engineer who is con-
sulted prior to making second Of t!lirt! ordef changes.
When a new bevel gear set is to be produced in quantity, it is first
necessary to "developll the pair - that is, to determine the desired location
and shape of the tooth bearing in the Gleason test machine that will prOVide a
satisfactory full and uniform load bearing pattern when run in the production
gearbox. This is currently accomplished by a trial and error process. The
gear teeth are first semi-finish cut to size on a Gleason bevel gear generator
(Figure 6). The gear member of the pair is then set up in a Gleason bevel gear
grinder (Figure 3) to the calculated but unconfirmed machine settings provided
by a Gleason gear summary. This summary consists of approximately thirty
machine settings for each side of the tooth. The gear member is ground lIspread
blade" (both concave and convex sides ground at the same time). The pinion
member is set up and ground in the grinding machine to the unconfirmed pinion
settings indicated on the summary. The pinion is ground "single side" re-
quiring a separate set up for both the concave and the convex side.
FIGURE 6 GLEASON HYPOID GEAR GENERATOR
After the gears are ground, they are installed in a Gleason universal test
machine (Figure 1) that is set up using precision gage blocks or set up gages
to the theoretical gear mounting distance. Using precision work holding
equipment) the gear and pinion are mounted in the same relative position to
each other as they will be when run in the actual transmission gearbox. The
test machine also allows calibrated adjustments along the gear cone axis) along
the pinion cone axis) and in the vertical offset direction.
The gear and pinion are rolled together in the test machine at a predeter-
mined light brake load (approximately 100 in-Ibs of torque) applied through the
pinion spindle. Prior to running, the gear and pinion teeth are painted with a
gear marking compound (similar to jev,'eler's rouge) that produces a rolling
contact pattern on the gear and pinion flanks due to the surface contact
between the mating teeth and wearing away of the marking compound (Figure 2).
5
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
6
The gears ground to the undeveloped summary settings are then installed in
a test gearbox and run under load. The observed composite gear contact pat-
terns are a final indication of the acceptability of the manufactured tooth
profile shape.
If the tooth profile contact does not meet the desired shape location and
percentage of contact required by the application l the gears are disassembled
for regrinding. The usual practice is to regrind, or develop, only the pinion
member because it takes less machining time (due to fewer teeth) J and because
of the Gleason system convention for single side grinding of the pinion. At
this point a gear engineer conducts an ana lyses of the dynamic load pattern,
evaluates the Gleason test machine no-load pattern) and makes a judgment as to
what changes are required on the pinion tooth to improve the dynamic load
pattern. To assist the gear engineer in determining what move or correction to
the Gleason grinding machine set up is most appropriate) the pinion cone axis
and the vertical offset in the test machine can be adjusted to change the
pattern size and location. These adjustments provide an indication to the gear
engineer as to what grinding machine setting will be most effective. In most
cases it takes a combination of two or more moves to correct a pattern and more
than one combination can produce similar results; however) only one combination
is optimum.
The pinion is reground to the new adjusted setting and the testing process
repeated. The number of iterations necessary to obtain a satisfactory gear
profile depends upon the skill and experience of the test machine operator
and/or the gear engineers. This judgment process is probably the weakest link
in gear tooth pattern development even with experienced machine operators.
Once the development is complete) several sets of master control gears are
made that duplicate the newly developed pair as precisely as possible. These
control gears are used to inspect the production gears. They are run in the
Gleason test machine against each mating gear subsequently produced by the
final machine settings to visually inspect the contact patterns against those
obtained from the developed master gear pair in order to maintain uniform
quality.
The production process control for spiral bevel gears is) in effect) a
miniature development process except that the changes required to keep a
drifting pattern situation under control are more subtle and involve the visual
comparison of a production gear pattern with the established master gear
pattern and the necessary corrective changes to keep the two in agreement.
Improved Gear Nanufacturing and
Inspection Process
The quality control process described above has certain inherent dis-
advantages. First) the acceptance or rejection of a production gear is based
upon a visual comparison of tooth contact patterns. Not only the size of the
pattern) but its shape and location, are significant. Acceptance limits for
these features are difficult to define qualitatively; therefore, the accept/
reject decision becomes a subjective one and is subject to the human frailties
of the operator. Second, the size) shape and location requirements of the
tooth bearing pattern are peculiar to each gear mesh and gearbox mounting and
no particular area, shape, or position can be considered universally ideal.
Third, since the tooth contact is localized and tested under a very light load)
the edges and corners of the tooth cannot be measured directly. It is neces-
sary to determine) not only that satisfactory bearing patterns are obtained
when the gears are mounted in their eqUivalent running position in the gear
tester) but to what extent this pattern can be changed by axial and radial
movements of the pinion axis, with respect to the gear axis) that would move
the pattern to the limits of the tooth contact zone. This is known throughout
the industry as the V and H check. By comparing patterns at these extreme V
and H settings) a cursory check on lengtln"'ise and profile curvatures is main-
tained.
851572
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
851572
It is apparent from the above d isCllSS iOll that a need exists for a more
definitive and quantitative means of dingoos illg ":hether specific in-process
changes are necessary in the grinding molehin!' settings to control the tooth
profile geometry within rather narrm.; 1imi Ls. This control is important for
highly loaded gearing to prevent conc:enLr"l ions of load that could cause
scoring, pitting or tooth breakage.
A Precision Coordinate Neasuring such as the Zeiss Hodel UHH500
shown in Figure 7 offers an effective "lterll.1Live solution to the problem of
spiral bevel gear tooth measurement ilnd cOlltfol. The UNN500 is a highly
accurate multi-axis-coordinate measuring machine t<,'ith an integrated Hewlett
Packard (HP) 9836 computer system that fwrmits spatial probing in any
of the three orthogonal directions. This IJlilC.:hillP.. in conjunction with a
sophisticated 3D software package, pro\" ides il d ist inc!. und quantitative means
of measuring and mapping three dimcnsioHi11 sllrfilcP contours, In order to
accommodate the complex surfacc of the sl,ir;ll ben1 gear tooth, the Zeiss
system also uses a precision indnxillg t,lldl', shm;lI in Figure 8, which is
integrated as the 4th axis in gf'ilr lllg prograllls, The computer program
packet for gear measurement permits Ul(' dpLl'rminn'tion of the face profile
coordinates of spiral bevel gears nt an IllJlimiu'd Humber of probe points on the
tooth surface and a point by poinL comp.1riso:1 ,..;i!.11 sLored nominal reference
values representing the mastcr gcar tooth This reference data can be
obtained either by measuring iJ mastcl' gedr ...ith tlIP developed profile on the
UNN or by computing the theoretical coordilutf' \',llues at specific probe loca-
tions using the final grinding machinr: sett ings. Hy considering the deviations
from the nominal values in the axes i1:1d thl' Hormal vector. the computer
determines the deviation of the surface normal ilL each probe point and prints
it out. These values can be drat.:!l by till' III' plotLcr at a preselected scale)
and interconnected to make up u thrcp diml!IlSioll<11 C'!rror diagram (Figure 9).
FIGURE 7 PROTOTYPE GEAR MEASUREMENT SYSTEM
7
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
8 851572
FIGURE 8 ZEISS UMM 500 ROTARY INDEXING TABLE
II
,
...LIGWlEIIf HOOf! - 0 PART 110.
PROFILE MEASURING RH1G GEAR ! Z OPERAT.:
o NO: Tooru-n(l: 20
DAlE
FIGURE 9 TOPOGRAPHICAL MAP OF GEAR TOOTH PROFILE
This improved method of tooth measuremellt and numerical representation of
tooth errors at each specified probp. point provides a means for quantitative
evaluation of the spiral bevel gear tooth profi Ie ill physical and measurable
geometric terms without resorting to \"isual comparisons of tooth
contact patterns. This method Clln eliminatp the need for a Gleason running
test machine, Gleason gear blank checker. looth spacing checker, tooth index
machine, master control gears and associaud gages.
The overall objective of this program t.o develop a manufacturing
method and the technology reqUired for an and final inspection of
spiral bevel gears using a prototype, automated. mechanical contact coordinate
measuring machine. The program comprisC'd four phases under Contract NAS 3
23465.
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
851572
9
Phase I involved the definition and development of a final inspection
technique for spiral bevel gears utilizing a rnulti-axis-coordinate measuring
machine. The UH-60A BLACK HA\;rK main module bevel gear set was selected for
evaluation and study. In this initial phase, a Zeiss UHH-500 universal mea-
suring machine was procured, installed in the Sikorsky gear inspection depart-
ment and converted for spiral bevel gear measurement by the addition of a
rotary table and a He\.;lett Packard (lIP) computer package. Tooth profile
measurements and pattern checks wer.e taken on the master gears for the selected
gear set, and on a number of production gear sets, using both the UHH-500 and a
standard Gleason test machine. In Phase I, a final inspection process for
spiral bevel gears, utilizing a computer controlled multi-axis-coordinate
measuring machine, was demonstrated pOSitive economic and technical
results.
Phase II involved the development of an in-process inspection technique
for spiral bevel gears where the measured deviation from the nominal profile
are converted into corrective grinding machine settings during the production
cycle. A sensitivity study made in \.'hich gear test specimens) ground
machine settings which purposely deviat.e from the normal settings, were mea-
sured on the UHH-SOO and the corrective settings determined by the Gleason
G-Age software package. A comparison of the calculated setting changes with
the actual settings used to grind the gear specimens validated the computer
program. An alternate empirical approach, using the matrix of deviations
caused by each independent setting change, \<,'as also evaluated.
Phase III included a pilot product.ion monitoring program on the selected
production gear set and an acceptance test procedure.
Phase IV included the documentation of a new manufacturing and inspection
specification for spiral bevel gears and a manufacturing methods and technology
plan.
Advanced Bevel Gear Neasuring Technigue
The initial task was to demonstrate that the topology of a spiral bevel
gear can be quantitatively measured and compared with the desired nominal
values stored in the computer.
The simplest method for determining the nominal points on a gear tooth
flank is by digitization of a master gear. The master gear is a carefully
manufactured comparison gear \<,ith a developed profile that has been proven
under load in an actual gearbox installation. The measuring machine is pro-
grammed to probe points on the flank of a master gear tooth for storage on a
magnetic cassette tape. Specialized software permits rapid generation of an
evenly distributed point ... ork along either radial or axial sections after
manual probing of the corner points (Figure 10). The vector of the surface
COORDINATE MEASURING
SET-UP TECHNIQUE
.v
R

adjacent points
with radius R
Network
points
z
-- -----
C

Outside
---
y.,,,'-/<.=-----
I I _---
, ,
, ' ,
I (
1\ '...\ 'fI
I '... ....
\ ....
..................... _-------
Inside
---------
FIGURE 10 GENERATION OF NETWORK POINTS
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
10
normal is determined mathematically from :-ievel-al lluLoffiatically probed points in
the vicinity of the specified netv.'ork point. \,;IWll llIilking measurements l actual
points on the test gear flank are numl'ricillly compared with stored point
locations on the master gear flank. TIll' milS tp r gP;J r, s i Ilee it is produced on
the same equipment as the test gear, ....ill gt'lll'rally contain some random form
deviations that vary in magnitude from tootll to tooth on a given master gear
and from one master gear to anothr.r. III it may become necessary, in
this method, to create an average nom i Hill mns LC'r gcn r profi Ie by integration
over a large number of actual master gCell flanks.
Even though spiral bevel gears ho\'P a high of geometric complexity,
the theoretical surface can be gcncrnled I\umpr ica I ly by computer simulation of
the manufacturing process. This, in faLL, hn!-> ilccomplished by the Gleason
G-Age software package. G-Age is designed to provide the computer on the
UHM-500 with the ability to:
851572
1. Download theoretical lOOlh
matrix from the Gleason maill
sllrfncp Loordjl\i1tp
fr:JnII' CO:;lplltllr.
data and corrective
2. Direct the machine to me;lsun' til!' tooth sllrface.
3. Compare measured data to theoretical or mastcr gear data and cal-
culate differences bcn.:ccn til, tl,.,O surfaces,
4. Calculate corrective cutting/grinding lnilchillC settings based on the
measured differences.
5 Make tooth spacing checks,
G-Age permits more freedom in tilt choicE' of till' form and density of the
point network and provides a more theorf't i cdl J j liP thilll the measured master
gear values) which themselves are SUb,il'ct to mnHufacturing errors.
The inspection process for n production geilr consists of setting up the
gear in the UMM-500 and automatically prohing the surface at the specified
network point locations. To accompl ish thl' gear is mounted on the
coordinate measuring machine tabJ!, or (lll lht' indexing table with its axis
parallel to the Z axis of the m,1chil10 (Figllrl' 11) \t,'ith care being taken not to
deform it while clamping. Part illignml'ilt by bringing the probe
FIGURE 11 PINION SET UP IN THE UMM 500
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
851572
into contact at a series of Oll ,I diameter to establish the
location of the Z axis of th(' gf'iJ:' ill t"I'lillion to the machine axis. The
reference coordinate syst.em for LllI' nor.lilll] datil for Lhe bevel gear is then
located along the gear Any i n,d ZI't"O POillt. can be selected along this
axis. In order to determine Ull' ,lflgil' of rOlillioll of the gears polar
ordinate system relative to til<' L00rdinillp s\'sr.em, a known point on
the tooth flank is contacted and tit, df".. iat iOlL of point from nominal is
set to zero.
measuring the compound curved surfaces of spiral bevel gears, the
"continuous probing" mode of the UNN-SOO system is particularly beneficial.
The machine can follow the contour of the part in a predetermined direction in
the same manner as the follower head on a 3-D copy mill. The automatic posi-
tioning control that is actuated at probe contact scans the free axis of the
machine until the inductive measuring system in the probe head is brought to
its nu 11 point. The moment this condition is achieved all three machine
coordinates are automatically transmitted to the computer; therefore, the probe
may be locked in the X axis and be made to traverse to predetermined locations
in the Y axis, while automatically following contour changes of the part in the
Z axis, and the machine will remain at a preselected X-Y' location until the
probe has been nulled in the Z direction and the position information trans-
mitted to the computer. It will then proceed to the next X-Y location.
Tooth flanks are measured in eNC mode. Nominal points on the network are
loaded from the magnetic tape cassette into core memory and transformed into
machine coordinates. The computer keeps track of the momentary positi':)fi of the
probe and determines the path to the next point. The measured deviations from
the nominal surface are determined along the projected surface normals.
After measurement of a complete tooth flank, requiring approximately 3
minutes for 45 measuring points (Figure 12), the probe automatically returns to
its starting position and awaits further commands.
COORDINATE MEASUREMENT
TOOTH PROFILE PLOT
11
I
I"

FIGURE 12 DISTRIBUTION OF MAPPING POINTS


Presentation of the results of the measurements are graphical plots, as
shown in Figure 9) and a digital printout (Table I). The digital printout
locates each grid point by column and row number. For each grid point, the X,
Y, and Z coordinate values are lis ted as we 11 as the x) y. and z deviations
from the stored nominal values. The last column in the printout is the devi-
ation in the surface normal dimension and is the value plotted in Figure 9.
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
12
" ,... ""'=".. i===
. ....:..__ __.. .._-----
.. __ ..
_,......_-'-".'c: ".__."" _
,: ..,." ' ,." ..."" ",
" , .. " ,,,... t"
._-,.'-'- , . .......
" , -.."" '."''' ........,.""', ...,.. -,,,., "
851572
.,," "I"
-,." "
,,,, ,,,,,
I' ",0--
,Tt''' '.".n
,,,,,, ..,,"",. , ...,. . -0,"-' <.B'"
.. ... ".,..,"",.
-: ....
-':-; .."",.. 0"","' ..."". '1.'''' ",to
TABLE 1. DIGITAL PRINTOUT OF MEASURED POINTS
A good correlation was found between the Gleason test machine patterns and
the UHH-500 measurements. Figure 13 shows a plot of test machine values
compared to UHN-500 values. The test machine values plotted are the observed
distance from the end of the test machine pattern to the toe end of the tooth.
The Zeiss values are the measured surface deviation at the same point on the
gear. The correlation between this data and the Gleason pattern is as ex-
pect.ed.
The advanced measurement technique has proven to be useful in the analysis
of variations which exist in master control gears and in production gears
manufactured by different manufacturers even though test machine pattern
comparisons showed no reason for rejection.
If the deviations from the nominal values, as measured on the are
not acceptable, G-Age permits calculation of corrective machine settings based
on the magnitude of the measured differences and their location.
TOOTH MEASUREMENT TO
THE TOE OF THE GEAR
.400
.300
Gleason
test
machine
values .200
.100
r----
,
Gleason pattern I
,
,
,
,
,
,
" 'zeIss dala
,
,
,
,
,
,
200
150
Zeiss values
In millionths
01 an Inch
100
50
10 20 30 40 50 60 70
Toolh no.
FIGURE 13 COMPARISON OF MEASUREMENT SYSTEMS
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
851572
For the first time a spiral bevel gear tooth can be measured and compared
quantitatively to its theoretical form. The primary advantages of this system
are the degree of accuracy attainable, the capability to evaluate tooth geo-
metry over the entire surface, the ability to calculate grinding machine
setting changes. and the reduced time needed to accomplish the inspection
process. A flow chart showing the manufacturing and inspection of a production
spiral bevel gear set using this technique
l
is shown in Figure 14.
GEAR TOOTH PRODUCTION PROCESS
0''''''.'0 co."
FIGURE 14 BEVEL GEAR PROCESSING FLOW CHART
Demonstration of G-Age Corrective
Hachine Settings
In order to evaluate the effectiveness of the conversion of UMH-500
measurements into delta machine settings for the spiral bevel gear grinding
machine) parametric testing was conducted in which bevel gears were ground and
reground with machine settings that purposely deviated from the Gleason grind-
ing machine summary settings in accord<lncc \.:ith a specific matriX. This matrix
involved 10 first order changes for the spiral bevel pinion and gear (Figure
15) and B second and third order ch<lnges for the spiral bevel pinion (Figure
16). Each setting change consisted of 5 variations including 1 baseline) for a
total of 142 grinds (91 for the pinion and 51 grinds for the gear).
Gleason grinder first order changes
I(lTUlt; I I ",., ''''01, \til" .. " "a""" n:l"1'''' '''''''>:In, en""",""'''''''''' 'U""01 , .. <0<0 ..
13
UIlIX!
cwaR 10 '10.';1 ......
AlIGtE nTH
1\ .....10HSO

AlIGU
(IWl\( AIKin
Cl.'t WIll! AIlGL[
rOSlllOK
110 OffSEl
AlKia
t!C;EKO
I lUcMM SoUlo? Ch""IO
T .. Z tl ... s I K'thloo Sotttn,1
I Wtrlfl<>tlo. Grlf\d
.- , -
lIT T I _ ..
II TT
.- .. -
11' ,
II Tl
,- .. -
III II
FIGURE 15 GEAR CRINDING MATRIX
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
14
851572
Gleason grinder second and third order changes
UIWG .' I

SCOIiD
BIAS (!WIC!

lUGltr.IS(
IPlllAt ArGlE ClI;l},Ct
(.<,II (II.\)I;,.
Illl"O ORDER ,11')(;[;
ornH
lHllU
G[HMT/NG (!.Ii 'Ill.
, !il"''''!l\ ' .... "'00. oJc... ... .. '''!le'_''!!!.'.:!l1'
I
r ..
I I T 1 '._ ..-_
It,,1
.. -...
I I 1 Y I _ w_
1111/
I "ll
1 s.nl"1
1 l U,,<I I S.tt1"'l Chin",.
1 TtrllltHlon Gd"<l
FIGURE 16 GEAR GRINDING MATRIX
Four specific grinding machine settings were used for evaluation.
An eccent.ric angle change of zero degrees and five minutes
COoS')
Hachine center to back I.o;ithdrawal of .020 inches
A pressure angle change of zero degrees and thirty minutes
(030')
A root angle change of zero degrees and twenty minutes (020')
The eccentric angle change 05' resulted in a maximum deviation of
inches in the bevel pinion profile geometry as shown in Figure 17. When the
pinion was reground to the corrective delta setting calculated by G-Age this
deviation was reduced to -0014 inches (Figure 18). A second regrind resulted
in a maximum deviation of +.0003 inches as shown in Figure 19.
_COIIII
G"P tOR"lClEo
"'''c,'''" '''''''C
(CC"""
cn"Olf ."C
"'OJ",, OUt
"OQ' ."G
pnt"v"! ."C
o
o
0"
OeVIATlONS CONVERTEO ItHO CORRECTIVe MACHINE SETTINGS
FIGURE 17 INITIAL GRIND WITH A ECCENTRIC ANGLE CHANGE OF 0'5'
.-
_
.-

G.,nCOMtCHO
""C"'''! SHl'''G
/
.Ol. tCC .""
CnA"U .""
oUt
0001""0.
00(0.1I01."G
DEVIATIONS CONVERTED INTO CORRECTIVE MACHINE SETTINGS
FIGURE 18 RESULTS OF 1st CORRECTIVE GRIND
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
851572
'"
_,llOOlO
FIGURE 19 RESULTS OF 2nd CORRECTIVE GRIND
G_MET
ConREClO
MACIllNE
SElllNGS
lice. ANG. a"U'
eRA IUHl, 0"0'
SLOIlASE 0"0'
floor ANa. 0"0'
PREss. ANG. _Clor
15
The machine center to back change of .020 inches withdrawal resulted in a
maximum deviation of +.0053 inches in the bevel pinion profile geometry as
shown in Figure 2.0. When the pinion was reground to the corrective delta
setting calculated by G-Met, the deviation was reduced to +.0009 inches (Figure
21). A second regrind resulted in a maximum deviation of +.0004 inches as
shown in Figure 22.
" "[I COR/lleHD
..
ECCA"G .,..'
C/lIIOll "UU ".,"
DAS 0
110<;11 Mill 0
At,,, 0",'-
{)-Hto"J
OEVlAllONS COtiVEnTEO INTO CORRECTtV MACHINE SETTIiGS
FIGURE 20 INITIAL GRIND WITH A MACHINE CENTER TO
BACK CHANGE OF M.C.B. W .020
Gleason Pru"lt ",UlU"
G-ME 1 ""'.,", ""
DEV\ATlONS COtiVEllTED HTO coonecnve MflCHtE SETa4GS
C""COR"leTtO
"00""''''''''<;:';
.,zc ...,c
ellA"' ...... "
'10'''0. ...,' 0
"oOl ...G 0
"",,:ouRl ...n ....
FIGURE 21 RESULTS OF lsI CORRECTIVE GRIND
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
16
(;
""';""'151""""
o
CM<>UAllG
"m,.. 0
o
.".<:UnlA"') If'l"
851572
..i
G-MET
DEVIATIONS COl/VEnTED !lITO CORAECllVE MACHINE SETTINGS
FIGURE 22 RESULTS OF 2nd CORRECTIVE GRIND
A pressure angle change of 0030 ' resulted ill a maximum deviation of -.0009
inches in the bevel pinion profile geometry as shmm in Figure 23. 'olhen the
pinion was reground to the correct i ve de!! La t ing ell I cu lated by G-Age J this
deviation was reduced to -.0004 iTlches (1-' i gun' 2':' I . A second regrind resulted
in a maximum deviation of inches ils sllo",,-u ill Figure 25.
to;lll,

//II'lil/ II
".,JIII/IIIII
- ' ,)'7 If ;; iT
N-4
... ,... """" Po"' ''''._ ,"',,"}
GI C<Json "nOIIU. >A5'mn'OH ''''''',"'
G-HET
DEVIATIONS CONVERTEO INTO conRECnVE MAClme
r; Lll,c"n"t1:1W

Q
no<>l AHG
.nU.:")nEAI<G 0",.
FIGURE 23 INITIAL GRIND WITH A PRESSURE ANGLE
CHANGE OF O' 30'
Gleuson ",,>sur'(><1:111 "'''ll"'
OEYlATlONS CONVERTED ffTO CORRECTED MACHlHE SETTiNGS
" ..t"OMtC''''
>"."',,NI.Ht'Nm
o
o
",''' a
...,..".
FIGURE 24 RESULTS OF 1st CORRECTIVE GRIND
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
851572
17
" '",conn'CllO
.""Cl"..
cc ... " ..
......,.,'" ..
m"'llll...." 0
00' ...."
.=un' ....." ....
"",
,
. .
,/ / / / /
,
j
" ..
/ / / / / / /
-",
;
/ / / / / / /
'>L L J' I .L
,/ /
/ /
---
!--
,. 1

I
.",
......
,
"

.. ,......,..,..-
''''''''''''''''0' '"
IU ..(t" PHlIu" , .
,"0"'" "
G ME 1 ,,-
"0'"
,,,,,,,,...,
,. "
"'"
'"
DEV1ATlONS COHVRTCD liTO COrtrlECTED MACH...." SETnlGS
FIGURE 25 RESULTS OF 2nd CORRECTIVE GRIND
In all four of the above cases. the tH>\'l'l pin ion tooth profile geometry
was restored to within acceptable 1imits ill t\.... o regrinds using only first order
changes. This demonstrated that the procedure was effective
in correcting an out-af-tolerance tooth profill' during the production process.
This technique can eliminate the for il filWl process.
A root angle change of rcslll u'd ill il m;:lximum deviation of -.0041
inches in the bevel pinion profi geom(>try sho...,;n in Figure 26. When the
pinion was reground to the sptting calculated by G-Net. this
deviation was reduced to -.0004 inches (Figllrl' "2.7). A regrind resulted
in a maximum deviation of -. ODD:! i llclH'S <1:-. ShO..... ll ill Fi gl1re 28.
'"

"
,
--If

i
,.
".
., J'-.,
iJ'CI.
,
I VI 1/
-,
1/ 1/
,1
"n

\
!I
..,-""""-
...,"".- .....,.
Gleason
"
.....' ...
G-MEl ...,. ..... ,"',... ..,. , ...
.."
"
,; "OCOMfcll"
MAC!""'UIIINC.
frx"'Nll rf'r
CM"....."11 lI",'
[}
[}
.Ofo:.-Unt ....11 -<:. IT
OEVIATIONS COIIVERTEO INTO CORnE:CnvE MAClfliE SETTflGS
FIGURE 26 INITIAL GRIND WITH A ROOT A!.tGLI'. CHANGE;.OF 0" 20'
_._-
.
, ,

. . .
_.00038
.

I-
..0 .
,(
),
--
,;
,
.-
.01
-

\

.
.........--
Gleason PnoFtlE IEASl$EIEIH f'lfllOtl .,
.....:"''''"
G MET "... ",,,..
'0-
f. t. ...
conneCteD
MACHlIIE SETTINGS
EtC MIO. 0
CR/l.DLE /l.NG. 0
SllDlIIG DJISE 0
I\OOT /l.IIG. 0
pnESSVRE /l.NO. 0<> II'
DEVIATIONS CONVERTED INTO CORRECTNE MACHIlE SETTINGS
FIGURE 27 RESULTS OF 1st CORRECTIVE GRIND
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
18
851572
(lM(TCOIlArCTED
MflClmlE Sntl'lGS
ECCMIG 0
Cl1l1DU ANa 0
StlDmG !IASE 0
floor MiG 0
rnrswl!E IINI; fI' 2'
,"""..... .,.. ...'l
PIIOFILE "USU1JEIIT PlIuml
G-MET
....1_"'''''''''
Gleason
mrr-="
/ / ,/ ,/ .v
""",/f/ ///
--,J;; //7/
-.".m - - ,/ /
,I\c- -="
L\ \
DEVIATIONS CONVERTED INTO CORRECTIVE MACHINE SETTlNGS
FIGURE 28 RESULTS OF 2nd CORRECTIVE GRIND
In all four of the above CilSP:-.. thl' be\'vl pinioll tooth profile geometry
wc:s restored to within acceptabl(' limits ill l\o.'o rlgrinds using only first order
changes. This demonstrated thul the corn'clivp procedure was effective
in correcting an out -of -tolcrancp. looth pro f i 1(' tiu ring thl' production process.
this technique can eliminate the for ;1 fill;!1
ACCEPTA\CE n:ST l'ROCElll"RE I ATI')
Three sets of UH-60 main gcarho::-: spirill hewl "'ere manufactured by
the improved technique for dynnmic load test ing to production acceptance
test procedure (ATP). For the ATP po\.>'c.r SPl' Figur 29.
MIN JSS I 0\ ACCI:I'TA\CI: TEST
PROCI:IIl"RES (ATPI
POIIER TEST
PO\,ER (SIlI'I Dl'R;\TION
L/Il INPUT

TIO
400 400
.,
5
-,
700 700 0) 15
1000 1000 170 10
1250 1250 2:10 5
700 700 6; 5
1400 1400 200 5
FIGURE 29
Results of the three tests conducted lo llw ATP power spectrum are as
follows: (See Figure 30):
Test 1 -
Test 2
Test 3
Visual inspection indiC<1tcd flO di::>lress.
Visual inspection indici1tcd n modcnltt profile scoring condition
on both pinion and genr.
Visual inspection indicated n moderate profile scoring and wear
condition on both pinion and gnar.
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
851572
TEST 1
ATP - DYt\II:'11 C LOA!) TI:ST KESlI.TS
Gear and Pinion Ground lo Produclioll
19
Pinion Condition After Test Gear CO!l(Ii l i 011- ,\ It {'l' Tes l
TEST 2
---
Pinion Ground to il Prr's,.,lJrp Aug!" Error of 07', Resulting in a
+.0004" Profi!e IlI'\'jilLiOJi.
Pinion Condition After Test Gear Condilioll Atllr Tpsl
TEST 3 Pinion Ground Lo d I.PII;.;lh l\is{' (;ur\'aturl' Error of
Wheel Diameter and ,02(j 111::11"'" S id"
Resulting in il +.0007" Profi!,' iJp\'jaLioll.
.040 Inches on
Radial Change
Pinion Condition After Test
Geilr Cond i l i Oil ,\[ l" r Tps l
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
20
Conclusions
Application of the improved inspection method for spiral bevel gears)
described herein, can provide the following benefits:
1. Permits quantitative evaluation of tooth profiles.
2. A substantial reduction in Manufacturing and inspection manhours and
associated cost.
3. Measurement of all blank dimensions and tooth elements, including
tooth spacing and index variation; in one set up.
4. Permits measuring entire tooth surface rather than the localized
contact area measured by the current technique.
5. Eliminates the need for, and costs associated with, metallic master
control gears and working master gears used for production quality
control. Also eliminates the inherent variability of master gears.
6. Produces both digital and graphical records of the tooth geometry.
7. Automatically calculates the most optimum grinding machine corrective
settings necessary to correct profiles tdth deviations outside the
established tolerance range.
References
1. Lemanski, A.J., and Frint. H.K., Automatic Inspection and Precision
Grinding of Spiral Bevel Gears, AHS Paper No. A-84-40-34-5000, presented
at the 40th Annual Forum of the American Helicopter Society, Arlington,
Virginia, May 16-18, 1984.
2. Lemanski, A.J., Frint, H.K., and Glasow. ~ Manufacturing Perspective in
The Design of Bevel Gearing, AGARD-CP-369. Paper No. 15, Gears and Power
Transmission Systems for Helicopters and Turboprops. Presented At The
Propulsion and Energetics Panel 64th Symposium. Helo in Lisbon. Portugal,
8-12 October 1984.
851572
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
This paper subject to revision. Statements and opinions ad-
vanced in papers or discussion arc the author's and arc his
responsibility, not SAE's; however, the paper has been edited
by SAl': for uniform styling and formal. Discussion will be
printed with the paper if it is published in SAE Transactions.
For permission to publish Ihis paper in ful.l or in part, contact
the SAE Publ.ications Division.
Persons wishing to submit papers to be considered for pre-
sentation or publ.ication through S.-\E should send the manu-
script or a 300 word abstract of a proposed manuscript to:
Secretary. Enbtineering Activity Board, SA.
Printed in U.S.A.
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014

Potrebbero piacerti anche