AD-A250 770

NASA AVSCOMContractor Report 189125 Technical Report 91-C-048

Enhanced Automated SpiralBevel Gear Inspection

DTICHarold K. Frint and Warren Glasow D TIIUnited Technologies Corporation IMAY 14 1992 1Stratford, Connecticut

March 1992 O 0

Prepared forLewis Research CenterUnder Contract NAS3-25961* 92-12728*~IIuIIIEEIfnII

NASA USAMNational Aeronautics and AVIATIONSpace Administration SYSTEMS COMMAND

92 1 ii $4

ENHANCED AUTOMATED SPIRAL BEVEL GEAR INSPECTION

Harold K. Frint and Warren GlasowUnited Technologies Corporation

Stratford, Connecticut 06602

SUMMARY

An enhanced manufacturing technique for the design and in-processinspection of spiral bevel gears, utilizing a computer-controlledmulti-axis coordinate measuring machine, has been demonstrated atSikorsky Aircraft in a Manufacturing Methods and Technology programsponsored by the U.S. Army AVSCOM PF.opulsion Laboratory, ClevelandOhio.

The technique uses a Zeiss universal measuring machine in conjunctionwith an enhanced Gleason Works software package that permits rapidoptimization of spiral bevel gear tooth geometry during initial toothform development, and more precise control of the tooth profile inproduction. The process involves three-dimensional mapping of spiralbevel gear teeth over virtually their entire working surface, usingthe Zeiss machine, and quantitative comparison of surface coordinateswith nominal master gear values at some 45 grid points on the toothsurface. In addition this technique features a means forautomatically calculating corrective cutting and grinding machinesettings, involving both first and second-order changes, forcontrolling the tooth profile to within specified tolerance limits.

This enhanced positive control method eliminates all of thesubjective decision making involved in the tooth patterning method,which compares contact patterns obtained when the gear set is rununder light brake load in a rolling test machine. The inclusion ofthe second-order change calculation in the automatic correctionprocess, demonstrated in this program, reduces themanufacturing/inspection time by 1.72 hours per gear compared to thebaseline process which included first-order changes only.

&OO@89on Yoru.KTIS GRA&IDTIC TAB 0U.•aruoouzced 0JuSt!fIcat lo_

ByDistrlbuLcmAvallobilltr Co4de

A TAv sif|ii

PREFACE

This report presents the results of a follow-on program to develop anenhanced inspection method for spiral bevel gears. The initialprogram covered the definition and development of a final inspectionmethod utilizing a multi-axis coordinate measuring machine andfeatured automatic calculations of corrections for first-ordergrinding machine settings. The program reported on herein involvesthe extension of the method to include second-order machinecorrections.

The work outlined herein was performed under NASA contract NAS3-25961with funding provided by the U.S. Army Aviation Systems Command(AVSCOM). The technical monitor for the project was Timothy Krantzof AVSCOM's Propulsion Directorate at the NASA Lewis Research Center,Cleveland, Ohio.

This program was conducted by Sikorsky Aircraft, Division of UnitedTechnologies, under the technical direction of Harold Frint, ProgramManager,and Charles Isabelle, Chief of Design and Development ofTransmis ons. Principal investigator was Warren Glasow, SeniorManufacti.-ing Research Engineer.

Acknowledgement is gratefully made to Theodore Krenzer, RobertHotchkiss and John Thomas of the Gleason Works, Rochester, New York,for their support and assistance.

ii

TABLE OF CONTENTS

Page

SUMMARY ........................................................

PREFACE ........................................................ ii

INTRODUCTION ................................................... 1

THE AUTOMATED INSPECTION PROCESS ............................... 6

Universal Multi-Axis Coordinate Measuring Machine ......... 6

Determination of Nominal Values ............................. 10

The Measurement Process ................................... 12

Current G-AGE Correction Process ............................ 21

ENHANCEMENT OF THE AUTOMATED INSPECTION PROCESS ................. 22

Task 1- Selection of Components ............................. 22

Task 2- Establishment of Baseline Values ................... 24

Task 3- Establishment of Second-Order Corrections ......... 24

Selected Gear Set No. 1 .............................. 24

Selected Gear Set No. 2 .............................. 34

Alternate Gear Set ................................... 34

Task 4- Verification of the Enhanced Correction Process ... 39

Task 5- Economic Analysis ................................. 39

Basis for Economic Analysis ............................ 39

Income/Expense Statement ............................... 44

Discusion of Results ...................................... 44

CONCLUSIONS .................................................... 48

iii

LIST OF FIGURES

Figure Page

1. Gleason Test Machine ...................................... 2

2. Typical Gear Contact Patterns ............................. 3

3. Typical 3-D Plot from Zeiss ............................... 7

4. Digital Printout from Zeiss ............................... 8

5. Corrective First-Order Changes ............................ 9

6. Zeiss ZMC-550 Measuring Machine ............................. 10

7. Gear Set-Up on ZMC 550 .................................... 11

8. Pinion Set-Up on ZMC-550 .................................. 11

9. Generation of Network Points .............................. 12

10. Gleason Gear Grinding Summary ............................. 13

11. BLACK HAWK Utility Helicopter ............................. 23

12. BLACK HAWK Drive Train .................................... 23

13. Cross Section of BLACK HAWK Main Gearbox ................... 25

14. Selected Gear Sets ........................................ 26

15. Baseline Plot OF Gear Member-Set No. 1 ..................... 27

16. Baseline First Order Setting Changes ....................... 28

17. Zeiss/Gleason Measurement Results- Gear Set No. 1 ......... 30

18. Zeiss/Gleason Measurement Results- Pinion Set No. 1 ....... 31

19. Measurement Results-Gear Set NO. 1 Corrected .............. 32

20. The Phoenix 400PG CNC Bevel Gear Grinder ................... 33

21. Zeiss Plot of Gear Set No. 2 .............................. 35

22. Zeiss/Gleason Measurement Results-Alt. Pinion Corrected ... 37

23. Zeiss/Gleason Measurement Results-Alt. Gear Corrected ..... 38

24. Verification Measurement Results-Gear, Set No. 1 .......... 40

iv

LIST OF FIGURES (Con't)

Figure Page

25. Verification Measurement Results-Pinion, Set No. 1 ........ 41

26. Verification Measurement Results-Gear, Alternate Set ...... 42

27. Verification Measurement Results-Pinion, Alternate Set .... 43

28. UH-60 Test Main Gearbox ................................... 44

v

LIST OF TABLES

Table Page

1. Time Study Results ........................................ 28

2. Basis for Economic Analysis ............................... 44

3. Income/Expense Statement .................................. 45

4. Cash Flow and Payback Analysis .............................. 46

vi

INTRODUCTION

Proper and reliable service from a pair of spiral bevel gears can beobtained only when they are manufactured accurately and mounted intoprecision-machined gearbox housings that position and maintain thedriving and driven gear members in a specified three-dimensionalrelationship throughout their useful life. Gears produced onexisting gear-generating and grinding equipment will run smoothly andcarry the design load without distress if tooth spacing ismaintained, the teeth are machined concentric with the rotating axis,and the tooth profile contour is controlled so that maximum toothpair conjugation is achieved when operating under full loadconditions.

Since it is impractical to design and fabricate gear teeth and gearmounts that are free from deflections when operating under load, mosthigh-power gears are designed with tooth profile modifications alongthe tooth face and in the profile direction to compensate for load-induced deformations and tooth errors, and to prevent loadconcentration at the ends or tips of the teeth which could result inexcessive wear, scoring, or even tooth breakage. This is as true forspiral bevel gears as it is for spur and helical gears.

The elemental conformity inspection of tooth profiles that iscommonly performed on spur and helical gears, however, is notpractical for spiral bevel gears because the size and shape of abevel gear tooth varies along its face width instead of beingconstant as in the case of a spur gear. Prior to the development andimplementation of the automated inspection process, spiral bevelgears were inspected on a specifically designed Gleason test machine,shown in Figure 1, which provided a rotating test of the gear pairsimulating no-load operation under simulated gearbox mountingconditions. Tooth contact patterns under these rotating conditionscould be observed by painting the teeth with a marking compound,similar to jeweler's rouge, and running the gears with their matingmaster control gears for a few seconds in the gear tester with alight brake load. Because of the compound curvatures inherent in thespiral bevel gear tooth form and the profile modifications designedinto the teeth, these gears typically exhibited a localized compositetooth contact pattern, which, ideally, should spread out under fullload, filling the working area of the tooth with some easing off atthe end areas of contact. The size, shape, and position of thistooth bearing were a gross indication of the tooth topology both upand down the tooth profile and lengthwise along the tooth face.Typical tooth contact patterns are shown in Figure 2. If theresultant tooth contact pattern did not duplicate, within limits, theshape, location, and percentage of contact of the master gear set,run under identical conditions, the gears were disassembled forregrinding. A gear engineer analyzed the pattern and made a judgmentas to what machine setup changes were required to improve thepattern.

1J

Figure 1 - Gleason Test Machine

When necessary, the pinion was reground to the new adjusted settingsand the testing process repeated. The number of iterations necessaryto obtain a satisfactory gear profile depended upon the skill andexperience of Lhe. test machine operator or the gear engineer. Thisjudgment process was probably the weakest link in gear tooth patterndevelopment, even with experienced machine operators.

This method of manufacturing primary drive spiral bevel gearsrequired an experienced and qualified organization. It has been saidthat the development of a spirai bevel gear is more of an art than ascience. This expression is based on the requirement for skilledbevel gear machine operators who use their background experience toevaluate the position, shape, and contour or the gear tooth contactpattern produced by the rolling test in the test machine. Themachine operator's judgment is relied upon to determine what grindingmachine setting or combination of settings is best used to correct anundesirable feature in the test pattern.

Gleason gear grinding machine settings involve first, second, andthird order changes. First-order changes affect heel and toeposition of the contact pattern as well as top and flank position.These changes are used in the final positioning of the tooth contactpattern. Second-order changes include bias (diagonal movement)changes, profile changes, and wheel diameter changes. Third-orderchanges include wheel dresser offset changes and heel and toe lengthchanges. There are approximately 14 machine settings that are usedby the machine operator in changes that affect t.,e shape and

2

position of the gear tooth pattern. Second and thirn-order changesrequire a calculation of values, using formulas provided by the

6rVi ~Al 54f /011•53.5"/-I1/5O'/-/O/ FolK" MASTER PATTERN TAPE RECORD ',.S/-//O2•'-/o'/7-O/5

ACTUAL TOTAL B/L - I

MASTER GEAR T - /,5 . S/N //7/ MASTER PINION T - /..".L I. N ,/ L.

MASTER GEAR B/L MASTER PINION B/L

CENTRAL BEARING - DRIVE

CENTRAL BEARING - COAST

AC____

V ONLY HEEL

V •t4-2Y- H" - ACTUAL V - a ACTUAL H

V M H TOE CHECK

V • H • - ACTUAL V - -- ACTUAL H • -

V I H HEEL CHECK

V t H , ACTUAL V . / ACTUAL H "

PROFILE PROFILE

v 2- 10 H. v- .- 0o. H -7,0--920•

ACTUAL V - j ACTUAL H * ACTUAL V " ACTUAL H "

Figure 2 - Typical Gear Contact Patterns

Gleason Works, by a gear engineer who is consulted prior to makingthese changes.

3

The quality control process described above had certain inherentdisadvantages. First, the acceptance or rejection of a productiongear was based upon a visual comparison of tooth contact patterns.Not only the size of the pattern, but its shape and location, wassignificant. Acceptance limits for these features were difficult todefine quantitatively, therefore the accept/reject decision became asubjective one and was subject to the human frailties of theoperator. Second, the size, shape, and location requirements of thetooth contact pattern were peculiar to each gear mesh and gearboxmounting and no particular area, shape, or position could beconsidered universally ideal. Third, since the tooth contact islocalized, and tested under a very light brake load, it was necessaryto determine not only that satisfactory contact patterns wereobtained when the gears were mounted in their equivalent runningposition in the gear tester but to what extent this pattein could bechanged by axial and radial movements of the pinion axis with respectto the gear axis, that would move the pattern to the limits of thetooth contact zone. This is known throughout the industry as the Vand H check. By comparing patterns at these extreme V and Hsettings, a cursory check on lengthwise and profile curvatures wasmaintained. It should be noted, however, that, in some cases,particularly with small cutter geometry, it was impossible to extendthe contact to the extreme corners of the tooth by this method.

It is apparent from the above discussion that a definite need existedfor a more definitive and objective way of determining whether abevel gear profile is acceptable, and what specific changes arenecessary in the grinding machine settings to most efficiently bringan errant pattern situation under control before it gets too far outof hand. It is well known how important it is to control the toothprofile on highly loaded gears to within rather narrow limits. Atooth profile with excessive profile or spiral angle error couldresult in concentrations of load that could cause scuffing, pitting,or even tooth breakage.

In June, 1982, Sikorsky Aircraft was awarded a contract (NAS3-23465)under the sponsorship of U.S. Army AVSCOM to define and develop anautomated inspection and precision grinding procedure for spiralbevel gears utilizing a three-coordinate measuring machine. Thiseffort was completed in August, 1985 and was considered highlysuccessful. This improved inspection system is now in place atSikorsky Aircraft and is gaining wide acceptance throughout theindustry. A unique feature of this technique is that correctivefirst-order grinding machine settings are rapidly and automaticallycalculated for controlling the tooth profile within specifiedtolerance limits.

The objective of the present program, reported on herein, is todemonstrate and validate the theoretical corrective matrix for bothfirst-order and second-order machine change capability. Thisenhanced correction process will result in a minimum of machineadjustments in a production mode producing a higher quality gear witha further reduction in inspection time.

4

This is the second step toward the ultimate closed-loop automatedinterface system linking the automated coordinate measuring machinewith the new generation of CNC spiral bevel gear grinders such as the#463 CNC Gleason Grinder and the Phoenix 400PG Gleason gear grinder.

5

THE AUTOMATED INSPECTION PROCESS

The objective of this improved gear measurement system is thequantitative comparison of the actual manufactured spiral bevel gearsurface topology with an idealized surface, in this case representedby a "hard" master control gear. The computer-controlled measuringmachine uses the XYZ coordinates of this nominal or reference surfaceas a guide for probing and comparing the actual production gear toothprofile.

Differences between the production gear tooth surface coordinates andthe nominal values, stored in the measuring machines computer, aredisplayed either as topographical plots or digital printouts. SeeFigures 3 and 4. The corrective first-order machine setting changesare automatically calculated and printed out as shown in Figure 5.

Universal Multi-Axis Coordinate Measuring Machine

When checking the topology of a three-dimensional curved surface,such as a spiral bevel gear tooth flank, using computer-controlledmulti-axis measuring machines, the following requirements must bemet:

The nominal or reference surface must be expressible eitheras a mathematical model or as a matrix of discretecoordinate values representing the desired surface.

The actual surface must be measurable with precisionaccuracy in a reasonable period of time.

Quantitative comparison of the actual and nominal toothsurfaces should be possible.

* The causes of any deviations from nominal values must beinterpretable to permit corrective grinding machine setupwhen the deviations exceed specified tolerance limits.

The ZeissrM Universal Measuring Machines, either Model UMM 500 orModel ZMC 550, satisfied the above requirements and offered aneffective solution to the problem of spiral bevel gear toothmeasurement. The ZMC 550, recently purchased by Sikorsky to satisfyBLACK HAWK/SEAHAWK production requirements, is an accu'rate multi-axiscoordinate measuring machine with an integrated Hewlett-Packardcomputer system that permits unlimited spatial probing in any of thethree orthogonal directions. This machine, in conjunction with asophisticated Gleason/Zeiss three dimensional software package,provides a distinct and quantitative means of measuring and mappingthree dimensional surface contours. In order to accommodate thecomplex surface of the spiral bevel gear tooth, a precision indexingtable is used as the fourth axis in the gear measuring programs. Thecomputer program package for gear measurement permits thedetermination of the face profile coordinates of spiral bevel teethusing as many as 243 (9X27) probe points on the tooth surface and a

6

F- 0 0ru

Z L -

NZI

300

4-4)

C-D0N'4I

001-4

z zf

Z C

0 0z It

o54)

-uiCC< tnz-

LUi

a: 0

0 *0

LL z W

- J Im 0 z

7

C L RVERAGE S MIN MAX

1 1 0.00005 0.00011 -0.00009 0.000101 2 -0.00016 0.00010 -0.00029 -9.000101 3 -0.00009 0.00009 40.00019 -0.000021 4 -0.00019 0.00007 -0.00028 -0.000141 5 -0.00028 0.00007 -0.00037 -0.00023

2 1 -0.00000 0.00010 -0.00013 0.000052 2 -0.00021 0.00008 -0.00032 -0.000172 3ý -0.00008 0.00008 -0.00018 -0.000032 4 -0.00011 0.00009 -0.00021 -0.000052 5 -0.00020 0.00005 -0.00026 -0.00017

3 1 0.00006 0,00008 -0.00005 0.000103 2 -0.00010 0.00006 -0.00018 -0.000083 3 -0.00008 0.00006 -0.00016 -0.000033 4 -0.00002 0.00004 -0.00007 0.000013 5 -0.00003 0.00005 -0.00009 0,00001

4 1 0.00010 0.00006 0,00001 0.000134 2 -0.00004 0.00005 -0.00010 -0,000024 3 -0.00006 0.00003 -0.00011 -0.00.0054 4 0.00004 0.00004 -0.00001 0.000074 5 0.00003 0.00004 -0.00001 0.00006

5 1 0.00011 0.00005 0.00005 0.000155 2 -0.00001 0.00001 -0,00003 -0.000015 3 0.00000 0.00001 -0.00003 -0.000015 4 0.00006 0.00002 "0.00002 0.000065 5 0.00008 0.00003 0.00004 0.00010

6 1 0.00010 0.00002 0.00007 0.000126 2 0.00008 0.00000 0.00006 0.000076 3 0.00005 '0.00000 0.00003 0.000036 4 0.00010 0.00001 0.00007 0,000096 5 0.00004 0.00003 0.00000 0.00006

7 1 0.00010 0.00003 0.00005 0.000117 2 0.00017 0.00003 0.00012 0.000177 3 0.00013 0.00003 0.00010 0.000157 4 0.00013 0.00002 0.00009 0.000147 5 0.00000 0,00003 -0.00005 0.00002

8 1 0.00011 0.00003 0.00007 0.000128 2 0.00016 0.00004 0.00011 0.000198 3 0.00027 0.00003 0.00023 0.000280 , 0,00013 0.00005 0,00006 0.000160 5 0.00002 0.00006 -0.00004 0.0000'7

9 1 -0.00006 0.00004 -0.00011 -0. 00u(-49 2 C.00020 0.00004 0.00015 0.00072

v 3 0,00026 0.00004 0.00022 0.000309 4 0 .00026 0.00006 0.00!19 0.00031S5 -0. 00002 0.00005 -0, 10. u06 0.00003

Figure 4 - Digital Printout from Zeiss

, m l I I I8

U17

*a c o£ -o

* w.- - .- 0 0

<Z N - -ý -0"

S-# a I -N 0 M. r4 0 0CO 0 N~ UN N C> ; b00 (D~O. 00''O0 0C o00N 0 C>-0 00CC>0 0000 C

I. C3~ ITI U% N - -0 00 40 0 ti, . . .0 . .00D - - -

1 0 C3-Z t~ A-1't ,

I CL Z O E n a- . . . . .

La a E.: . ..Z . . .a . -LA LO => . . . .

a 1 cc- j LLJ z- IL.- - - - - - - -

In- " z.J w "i. -j a--

W- IIciaLL- all= LUZ Li ;E -. 0.......................... - -c 4)ci-% jRr Z Ci- CL 4 . . ~. .......... ...... ...... ............... La n

II a LL 0 C3 Z J - LN LL 11 0 -4 3 0 4-a y a L.)LL.I-Z 1- Lij '4 -4 -. . .. . .. . -. 0- . W

a = cr a- C1 Z Li ,- (N M d U, L ........ .................. " -4 CD 0 Ci z

*-7 af M-o Z L MLL - - : -~. . . .. . . LL.--4WýfU

a. a ý . . ... . . . LL s-- W~i.~ f (n

ita Z-' Z cc -W P........ ...... .................. Li >

I3 " .. J.. L.-I. 0- _j -C 3f.1 . - - . 9= = Z - Li i

IX C U E0C. fl- Z X Z M -. . - : -~~L C&C.X)

M u ~ rX 0NE-' LO -. . .J .*... UN '-4 ILA- C.3 Li w a) -3 u m . . . Or Or Or-

-a -a C-4 N4 LA LL.. LLJCCLL Z LO LfL Q: r4 'O - u) cflu) 0(N No LL 00 u m0 zi M- ccD CD Ouli um . - - Ul

II a l 0' - 0 1-- z ..j Z En " Lu -. 11Ca.L IL- -J!. CC . -cfmUw fin--II f. cm ~ LiJ LD (n C ~LOILL.I- I- a-L .1J L-C3-.1 -1 0CD U u ccc j

cc z I- -0 Cl W'- M LO C) -a)F 2 c z -. C mCDDC3 ifw LO CC~ cr 00C =i CCC i C1c ýaL )0:xa

M ' L) Li C= 0m '-a A L ln icU <=o W L.)w LiLI IC

L* Ii C3 ** n WW WL ' WZc WWuu Li - . L- ILL.

a a - jM*4"_ 0..J0 ZZ. LL . cc-4¶-4 C C) C3 =LLC

II :. z ir 3 Ln 0- 0 j 5 L i w - i- = -4 Li 0 ccL Li)C a-4 En

9 a wo =9o. L0 3rLi3 r C3CE,)..U)m m L 1-1-IxCf a:,C4I >-a- < jc a C 0 cE0

U ft

U It

point-by-point comparison with the stored nominal reference values.Generally a grid network of 5 lines and 9 columns, however, aresufficient to map the tooth surface.

The automatic measuring and data processing system presentlyinstalled at Sikorsky consists of several instruments (Figure 6),which are controlled by a central computer. The system shownincludes a Hewlett-Packard 300- series Desk Top Computer, a ZeissTMUniversal Measuring Machine ZMC 550, a Hewlett-Packard WinchesterDrive, an X-Y Plotter and a Impact Line Printer. The pinion and gearsetup on the ZeissTM ZMC 550 is shown in Figures 7 and 8.

Figure 6 - ZeissTM ZMC-550 Measuring Machine

Determination of Nominal Values

The simplest method for determining the nominal reference points on aspiral bevel gear tooth flank is by digitization of the ReferenceMaster Control Gear. The measuring machine is made to probe actualpoints on the flank of the master gear tooth, as described below,for storage on a data disc. This disc, in effect, becomes theunvarying "soft" master in this improved inspection method.Gleason/Zeiss software permits rapid generation of an evenlydistributed point network over the tooth profile after calculatingthe corner points and defining the network density. Care is taken toexclude the edge breaks or corner rounding when establishing thecorner points. The vector of the surface normal at each networkpoint is determined mathematically from several automatically probedpoints in the near vicinity of the specified point (see Figure 9).These normalized values are stored on the disc along with thecoordinate values. A network of 45 points (a 9 by 5 matrix) was

10

Figure 7 - Gear Set-Up on ZMC-550

Figure 8 - Pinion Set Up on ZMC-550

11

Cz

X D

Y .0

IV OUTSIDE% IINID"* . .

INID

POINTS ADJACENT POINTS

WITH RADIUS R.

Figure 9 - Generation of Network Points

chosen because it was felt that this size grid would provide anadequate map of the tooth surface without resorting to time-consuminglinear interpolation. Finer or coarser grids are, of course,possible.

Even though spiral bevel gears possess a high degree of geometriccomplexity, it was reasonable to expect that the nominal surfacecould also be generated numerically by computer simulation of themanufacturing process. This, in fact, was accomplished by theGleason Works. Gleason provides the software that converts finalgrinding machine settings, as reflected on a Gleason GrindingSummary, (See Figure 10), into theoretical profile coordinate pointswhich can also be stored in the ZMC 550 computer as nominal values.This method provides a more theoretical baseline for the measuredmaster gear values, which themselves are subject to manufacturingerrors. These theoretical points are used in the Gleason G-AGETMProgram to calculate the corrective matrix.

The Measurement Process

The inspection process consists of setting up the gear in the ZeissT'machine and automatically probing the surface at the previously- -determined 45 network point locations. To accomplish this, the gearis mounted on the coordinate measuring machine rotary table with itsaxis parallel to the Z axis of the machine (see Figures 6, 7, and 8),care being taken not to deform it while clamping. Part alignment isachieved by bringing the probe into contact at a series of points

12

NE ICI4 - to

li - -*i "" SI

ld~ !g ct

CD ;

;s polar

454

IL f

CL LLP -w.-g . t-2

0aWWW~Lcv

130

cc(a

0c

z Es-

0W4

01

ILl

gag - MU

CV) )

I cn

-_ 0- W x-

It) (a 01 U; -

An 00

tlV-

0n

m, I- s NWM 90 ft

t!2 .-. 55 .

I-IC-

--w

14

Iiawesom-

is M. EMa

4I~ ID

U .L3~K-L~ss-i ''

tj _.- :001U~21ak ei usX

CJ 2

Ly CD CD

e- N 0 x -

. . .LUZ 4

AwL.aZ* I.-0*.

L cr

xx CL

.20- "'.

> CLl

15

4;

Ec 4--

a = o

b0 z

w * ~Id='Misma M-0oh- W

. R o Cnm'o VlW rMCI ~ ~ ~ n~"w'wapWIDD .

WeO 02--.4 CO

ID x e01Mwcl 'oW*- -IX - I

31M V.. . .... ft mf ~ W~ . ...... a

5-0 * 0 "~ OW W~WWg4Q

mN'QU.N Q WLU

0 me

2-. -G -,a e '. 1. U 4

6wi 11A oM -.F

ru'.§ c~x U- -Cfl(.i o oo

i~iZx _ WU, o-

16

~e8 a8B a8d S aa a, US

Quu S;.- ISm a-

4 I9

1b -H-

H a a9us :231

qOO~q~ m agJma 0mm

Issme Nirn - n w-dw 040 e01 E wBo t~ov fa rN~~.3 al o

ca

Las

LAS A rlQU

CL~: Cc ... 19 _nL

I17

ITi3i

p-,, - u • • 8 i.A •8.. . .- .. . . ..s

•.~ ~ ~ 1 #-. .. . ._.- o .V!+ .

44

I .. .. ... .... ... . . . . . .. oI !Cd. S

I • • .. ~~.. .- .......V. .. . . . . . . . •0 A ..... + ......... t:'

b S49 0

VC1 . Ua-lo g8o

mm~p . -.++:m +m . --+ s•+. Ow +++, Wta 5b

us cz L LLA-9 . L3-& 5 - 0'

S• • w-• • " • u-• w ,- o • "-,-

sex cc1M -f

-si

g 04

. ) C..... ON-iR X C

ada

EEr~- CL ur~~n

~ft-o tsSS54 ww *oenSQ185

cc.

-SO).

zh

Lii

It IT amm C)

IW - L.en *s.A"

41: ***J

II

19

I>CO)

4-

0 44

- zL a!9MqNib-II-

-- -C a -aH o vu A O ý2 2 E v

, d - , aa

200

on a reference diameter to establish the location of the Z axis ofthe gear in relation to the machine axis.

The reference coordinate system for the nominal data for the gear isthen located along the gear axis, generally at the pitch cone apex ormounting shoulder. In order to determine the angle of rotation ofthe gear's polar coordinate system relative to the machine'scoordinate system, a known point on the tooth flank is contacted andthe deviation of this point from nominal set to zero.

The tooth flanks are measured in CNC mode. Nominal points on thenetwork are loaded from the disc into core memory and transformedinto machine coordinates. The computer keeps track of the momentaryposition of the probe and determines the path to the next point. Themeasured deviations from the nominal surface are determined along theprojected surface normwls.

Current G-AGET' Corrective Process

One of the prime requirements identified at the outset for animproved spiral bevel gear inspection system was that if the profiledeviations of a production gear, as measured on the coordinatemeasuring machine, are beyond acceptable limits; these deviationsmust be interpretable in terms of specific delta changes to thegrinding machine settings used to produce that gear. The procedureis essentially the inverse of the mathematical simulation processdescribed earlier and is accomplished by the Gleason Works G-AGETsoftware package described below.

After a spiral bevel gear set has been approved for operation in aparticular gearbox the final grinding machine settings are used tocalculate the theoretical surface coordinates and nominal values.This information is down-loaded, through a modem, and stored on adata disc . Along with this theoretical surface data, a correctivematrix is also generated and stored on the same data disc. Thecorrective matrix can be considered as a surface sensitivity matrix.For example, changes that affect the pressure angle and spiral angleof the tooth surface are defined. The sensitivity of the surface tothese changes is calculated and stored in the corrective matrix.Changes are defined for each Gleason cutting or grinding machine.When the tooth surfaces of the individual gears are measured andcompared to the nominal value matrix (either calculated theoreticalsurface points or measured surface points from a master gear), amatrix of error data is computed and stored. The error data is thenmultiplied by the corrective matrix and the first-order correctivesettings for the grinding machine are calculated and printed out.

21

ENHANCEMENT OF THE AUTOMATED INSPECTION PROCESS

To further improve the spiral bevel gear inspection process describedabove, Sikorsky performed the following tasks to demonstrate andvalidate the theoretical corrective matrix for both first and second-order machine change capability. Once validated, the improvedcorrection matrix can then be used by the coordinate measuringmachine to automatically compute first and second-order machinesetting changes in a production mode. The benefits of thisenhancement in terms of a further reduction in fabrication time andcost are evaluated by the economic analysis of Task 5.

Task 1 - Selection of Components

The specific components or gear sets that were used to verify theimprovements i-n the gear inspection process, realized by theautomation of the second-order change calculations, were selectedfrom Sikorsky's BLACK HAWK/SEAHAWK power transmissions.

The BLACK HAWK shown in Figure 11 is the Army's advanced twin enginetactical transport helicopter manufactured by Sikorsky Aircraft toperform the missions of assault, resupply, medical evacuation,command and control, and tactical positioning of reserves. Two GE-T701C turboshaft engines deliver 1,700 horsepower each to the BLACKHAWK drive system. The main transmission, shown in Figure 12,consists of a main module, two interchangeable input modules, and twointerchangeable accessory modules. The main transmission transmits3,400 maximum continuous horsepower with an input speed of 20,900RPM.

The Navy derivative of the BLACK HAWK is the SH-60B SEAHAWK. Thedrive trains are identical except for the fact that the Navy aircrafthas rotor- braking and tail-folding capability.

The two bevel gear meshes originally selected to verify theenhancement of the spiral bevel gear inspection process were:

(1) Main Module Pinion and Gear SetP/N 70351-38104 and 70351-38114

(2) Input Module Pinion and Gear SetP/N 70351-08205 and 70351-08221

The input module bevel gear mesh has a speed reduction ratio of 3.64and rotates at the engine input speed of 20900 RPM. It transmits1700 horsepower each on a continuous basis and has a single-enginecapability of 1900 horsepower.

The main module bevel gear mesh has a reduction ratio of 4.76 with aninput speed of 5748 RPM. It is the second stage mesh and deliversthe same horsepower as the input mesh.

These two selected spiral bevel gear sets are part of the ImprovedDurability Gearbox (IDGB) design used in both the BLACK HAWK and

22

TAIL RoTO~kDSIVE SH4AFT

t. AloG H-TT ryOC L

TPAHENGINE OUTPUT~

,jgure 12

RA,%e yr2-

23

SEAHAWK transmissions and are shown highlighted in Figure 13 and

close up in Figure 14.

Task 2 - Establishment of Baseline Values

In order to establish a baseline process and quantify the advantagesof an enhanced spiral bevel gear inspection system, which includesthe automatic calculation of second-order corrections, the selectedgear sets were measured on the Zeiss three-coordinate measuringmachine, using the Zeiss/Gleason software package featuring theautomatic calculation of corrections for first-order changes only.These measurements were made during production runs on these selectedgears and, therefore, did not require the manufacture of specialtest-gear sets. Typical outputs from these measurements are shown inFigures 15 and 16. Figure 15 shows the bevel gear, P/N 70351-38114,with a second-order bias-out condition. Figure 16 shows theindicated first-order correction for this gear which, it should benoted, does not correct this second-order variance. The biascorrection, which involves a cam guide angle change, was handcalculated.

The man-hours expended for each of the required machining,inspection, remachining steps were documented as well as theadditional hours required for the manual calculations and iterationsof the second order changes, where necessary. The results of thesetime studies are shown in Table 1.

Task 3 - Establishment of Second-Order Corrections

Selected Gear Set No. 1.

The first step in this task was to establish the basic dimensionsheets and summary files for each of the selected components. Thiswas accomplished using the Sikorsky Tektronix computer terminal whichis on-line with the Gleason Works mainframe computer. The followingGleason programs were run for each gear set.

T2000 The Gear Dimension Sheet

T2000 The TCA and Summary Sheet which includes all of themachine settings and related gear blank dimensions.

A 622 The Gear Grinding Sequence Program

T606 This Program Converts Grinder Settings to BasicSettings

T801ZO The Tooth Form Generator and Correction Package. Thisprogram generates the theoretical XYZ surfacecoordinate points and the corrective matrix includingnow first and second order changes.

A Special Analysis File (SAF) was thereby established, for eachselected gear set, which included the final machine settings and the

24

00

oD

4-40

0

-'4

rXJ25

252

'4.'

f4

I4

4�i4)U)

o S-iZ

4)4J (D4)

U) 4)4-)

4)4) .1-i

H H4)

U)4�)4)

U)

Co4)CD

4)A 4-)- C.)4)

HFU)

'-4

4)

H -�

0

4-,4)

U)

4) 4 � LII�� (4-,C-)

H4)U)

26

cu H (

00

z

z

.L< '-I

w -0CD 0

0

x0 u

(P ~ 14

0 U0

o~~~ '-L o 0I-CCfl~U I Cl

cu27

.H I IIP.I -•I 'I . 1I?

~40 ý rlctI:N I NV 'nI- rII' NI' I'HAW-JI*"-ý Ist flRl)I R r.FAR 'f INUFXF IIF" '.I. fl

I•f':F'NI W I f' ANI41 I- ............ fl m in'A0I)I F ANNII F ............... II min

'•-I I'INr4 HA.I-. ... hAI TH nl. nllulil inhr..hMA",H I NF RIlIIT ANII F" .......... n mi nI)1 I-,4,•- R fl.H.ANI41 F .......... 19 minI'AM ITIOIF ANI-41 F .......... r.IAI 1) minf:"AMrl s T I NI4 ................ fl. MITII i n hr.h .s

fAUFRAU4F FRRfIRS In Ii OF 1 9PII- •Ik? A I• F ............. 13,7R n. nln min

.;P I R•AI ANIPI F ............... 11. 92 11.11I11 m in

IAlf. & F A, 'fnI1 .. 1 ..' . .I. .l. .. I1M MFAz; FIWRIINS' I,1•, nri;n I1. I1111i1119ij7,"3

"l.wt nrA Ifl. 1 I 1 lITIf, 3,5•R

Figure 16 - Baseline First Order Setting Changes

TABLE - 1. SAVINGS ANALYSIS

Savings in Hours per Lot (Set-Up)

Hand Calculated #463 Grinder Total1st & 2nd Order ist & 2nd Order Savings

Gear Pinion Gear Pinion in Hours

70351-38151 Pinion 54.6 19.8 34.870351-38167 Gear 50.7 24.2 26.570351-38104 Pinion 65.2 19.7 45.570351-38114 Gear 64.6 34.1 30.5Totals 115.3 119.8 58.3 39.5 137.3

28

theoretical XYZ coordinates of the tooth profile. The correspondingcorrective matrix was also established for each mesh.

The T015 grand master gear and pinion for the first selected gearset were then set up in the ZeissTM machine and measured. The flankform on both the concave and convex sides of both master pinion andgear were mapped at 45 grid points covering the active surfaces ofthe tooth. The deviations from the theoretical nominal surface wereobtained and the corrective settings, for both first and second-orderchanges, generated. The results demonstrated a measurable variationbetween the measured tooth surface and the theoretical nominal XYZvalues derived from the final machine settings. Figures 17 and 18show the Zeiss/Gleason measurement results for the TO15 Grand mastergears, P/N 70351-38114, and pinion, P/N 70351-38104, based on theSAF, adjusted to the same set-up values used to grind the samemasters on the Gleason #463 grinder. This variance indicates thatthe theoretical model in the Gleason mainframe computer did notduplicate the form ground on the #463 grinder and an adjustment ofthe theoretical points in the SAF would be neiessary.

Using the Tektronik terminal and the Gleason T606 program, new XYZtheoretical points and a new corrective matrix were generated anddown-loaded to the HP computer on the ZeissTM machine as before. Thisprocess was repeated until it was confirmed that the correctedtheoretical data adequately duplicated the measured tooth profile. Asample gear, ground on the #463 Gleason Grinder, was then measured onthe Zeiss using the new theoretical data. The final result is shownin Figure 19. The correction data indicated some machine changeswhich would be difficult to make on the grinder due to the inherentsensitivity of the manual settings on this machine. Two or moreiterations generally would be required to make the gear acceptable.The Gleason #463 and 137 grinders consist of machine settings whichuse verniers, dials, and slides which are all manually adjusted.Backlash in the screws must be considered as a source of error. Thisinability to make accurate adjustments make it difficult to obtaincorrect settings the first time. For these reasons, it was expected,and later verified, that the correction program would work muchbetter on the new CNC Phoenix Grinder.

The Phoenix 400PG, recently installed at Sikorsky and shown in Figure20, is a full 6-axis CNC machine tool specifically designed to moreefficiently grind generated spiral bevel and hypoid gears. Themachine uses a new concept whereby all necessary relative motions areprovided by six CNC axes. Three axes of motion are rotationalincluding the cutter spindle, work spindle, and the swinging base.The X-horizontal cutter axis, Y-vertical cutter axis, and Z-slidingbase are linear axes of motion. Each axis is controlled byindependent AC servo drives and precision ballscrews. Incrementalrotary encoders indicate the position of the rotary axes andincremental linear encoders mounted directly to the moving slidesprovide position feedback for the linear axes of motion.

29

0 0

Wi ED Cr.l

LU 1LUd- < z

En 0 C3

0

LUJ

o ~ to

zZ

IL <C I -

wI 0 4.'CD0 C:0 - (D

I- C

x cu

z C) CC ,

EL < -4 o~

xL CD)U

0 LrLL m

z

aL zL'4J .4 Z

0 *r .w0 u x cc

wo .4I-Cx 0 I-4U ...J ý

In- I- U)j <4 ) lF

> tn %-# CC30

+cu 0 (10 00 I

0

www a;o

cc I- 0

N0 0j

q r, D U)

co ,m I

6 U)

z -

020

Z Z)CC 0

LL z

0 L

Z

0~~~~ 2:C 0 -cu0..JI~~C cnW 0 -U~ I_ _

31C C

00 0m

N 0 01

P'-

I ~0I-

Hz

0- 0)

00

Li 4

'-U0 0 '-CU a;zR H: CCJ1

32)

00

Figure 20 - The Phoenix 400PG CNC Bevel Gear Grinder

All six axes are simultaneously controlled by CNC control andsoftware. This approach translates the basic motions required togenerate a tooth form into the simplest linear and rotationalelements. Programming is accomplished by user-friendly menus so thatgear grinding summary data can be easily input from a keyboard orsimply called up from the control memory. The mechanical setupadjustments, required on conventional gear grinding machines, havebeen eliminated. Set up time and machine changeovers aresignificantly reduced. As a result, operator efficiency isdramatically improved.

Set up time is accomplished virtually in minutes. Random batch andsmall lot size processing are as simple as calling up a program andmounting the appropriate tooling. Grinding wheel dimensions aresimply entered into the control and the CNC then automaticallycalculates the movements and positions necessary to grind the gear.

Advanced communication links to a mainframe or personal computer arepossible to allow setup, proportional changes, or corrective settingsfrom the Gleason G-AGE software, automatically.

7o further test the validity of the enhanced correction program,first and second order changes were made to the flank form by

33

changing the theoretical setting data instead of grinding the actualgear. This method offers more flexibility in assessing thecorrection data since the physical problems associated with resettingthe Gleason grinder are eliminated. The master gear and pinion werethen measured on the Zeiss with the G-Age correction program. Thecorrection program dictated machine setting changes to correct theerrant flank form. This result, which was also noted in the earlierprogram, highlights the fact that the Gleason correction program candictate two or more alternate setting-changes to correct onedisturbed setting. The changes, dictated by the correction program,were made, precisely as indicated, and the gears remeasured. Theresults demonstrated that the indicated G-Age correction valuesdid, in fact, correct the variance in the flank form.

Selected Gear Set No. 2.

The Gleason Summary, Dimension Sheet, TCA, and Special Analysis File(SAF) was also established for the second Selected Gear Set P/Ns70351-08205 and 70351-08221. Attempts to adjust the SAF to duplicatethe flank form of the master gear and pinion, as was done for gearset No. 1, however, were unsuccessful. The culprit was the gearmember, P/N 70351-08221. In 1980, before the Zeiss machine wasintroduced as a vehicle to inspect spiral bevel gears, and before theon-line computer link-up with the Gleason Works, a change was made inthe index interval and cam number for grinding this gear. Thischange produced a generating error in the flank of the tooth on theconvex (drive) side which resulted in a severe undercut condition inthe dedendum at the toe end of the tooth. This undercut can be seenin Figure 21. This condition was not discernable by the conventionaltesting methods in place at that time, and the gear was put intoproduction. The SAF would not accept this undercut condition andattempts to overcome the difficulty, and develop correction data forthis gear, were futile. The Gleason computer system doesn't accepta tooth form which has a generated undercut or has indicated machinesettings which are outside their predetermined limits. Attempts weremade to change the cam number and machine settings to values whichthe Gleason computer system would accept. These changes, however,caused a large discrepancy between the theoretical and actual part.It was finally determined that it would not be possible to establishcorrection data for this gear member. At this point it was proposedthat an alternate gear mesh be used as the second selected gear setfor this program.

Alternate selected Grir Set

Formal permission ,as received from the Contract Officer tosubstitute the BLACK HAWK/SEAHAWK Tail Take Off bevel gear mesh forthe selected gear set #2. This mesh is also shown highlighted inFigure 12.

The Tail Takeoff Pinion and Gear Set, (P/Ns 70351-38167 and 70351-38151) selected as an alternate mesh has an increasing speed ratio of3.409, rotates at an output speed of 4115 RPM, and delivers 524horsepower to the tail rotor. The Special Analysis File (SAF) was

34

in '0a)

0Z 001

uJ z CD

m KIL(D 0 00Y)01

rNN

UD Cl) 0

uj 0 (1)HCCLD IZ

tn < L! : 4

10

H- 0

w 0O w

0r-~

() 05C)< CU~W cuF- x C

00

0 CC~0 j (N

0yL r-z

00PU- z

00I--~ wU.)~~U L)U0 0 -L

U)cr I- J____ ___ ___ ___ ___ ___ ___

R > w35

successfully developed for this gear set and the correction programsatisfactorily demonstrated. The measurement results for this gearset are shown in Figures 22 and 23.

Concurrently with the performance of this development procedure, TheGleason Works, of Rochester, New York, as the originator and soleproprietor for the machine correction part of the RAM 300 program,made a number of changes to the software to improve its performance.Some of the more significant changes made are as follows:

1. A more convenient method for adjusting and correcting thebasic settings in the Special Analysis File was provided.

2. More options in the correction data were included; such as:

a). Zero First Orderb). First Orderc). First Order and Second Orderd). First Order and Second Order with RC

3. Gears can be measured as they are ground; eitherSingle-Side or Spread-Blade

4. Changes to the Eccentric Angle were previously indicated indegrees and minutes. Now the angle is given in hundreds ofa minute, an accuracy which is required to properly controlthe flank form.

The final step in this task was to establish and store the nominalvalues for each component in the ZeissTM HP computer on a floppydisc. This was done by digitizing each master gear and pinion at thesame 45 grid points covering the gear tooth surface. These becamethe coordinate points representing the nominal gear tooth surface towhich the production parts would be compared.

It should be noted that very small differences may exist between thenominal surface coordinates, represented by the master gears, and thetheoretical coordinates, generated from the final machine settings.Since the master gear represents the desired profile, determined fromdevelopmental testing, the production parts are compared to thenominal values digitized from the master gears and deviations fromthese nominal values are calculated. The theoretical coordinatevalues are derived directly from the theoretical model located in theGleason mainframe computer and are used in the corrective program tocalculate the required machine-setting changes.

These theoretical values are developed by adjusting the basicsettings in the SAF. This is accomplished by trial and error bymeasuring the master gear and pinion during each iteration and usingthe correction program to provide the necessary changes.

36

W I-i týw M ci)

CO w

-4

-4

ID ILcc:,,

wi

04

LI 0 ':JLfli.I Ui)(E

cn n.

z

LLJ 0

____________ ________________________________________ _________ _______________________Z_

10 + <--> CO

00 c

as CflU-Lu C!

ui -cI 4-J

Z o CEw 0'Z W I-- 0rI 10.49

0

0u 4)C

Cu4J

0

S4.)

z U)0 0)

Z 4)

1r)4)

0 a) 0Z ~LL u)

04)

OZ zw w

00

U) a

ow 1 Hr_~ Q9

4-) En -

44 mu ooz

-0 H

W L) 00 U) L

38

Task 4 - Verification of the Enhanced Corrective Process

To verify the enhanced corrective process, which now includes theautomated second-order change capability, a pinion and gear for eachselected gear set were followed through the production process.

The machined gears were set up in the Gleason grinder and ground tofinish dimensions. The gears were then measured on the ZeissT'machine and reground as indicated by the grinding machine changescalculated by the enhanced correction program. The pinion and gearwere then remeasured to verify that the indicated changes wereeffective. The man-hours expended for each of the required stepsduring the enhanced spiral bevel gear manufacturing/inspectionprocess were recorded and are shown in Table 1. Figures 24 and 25demonstrate the Zeiss/Gleason measurements of P/N 70351-38114 andpinion P/N 70351-38104 after developing the flank form using theenhanced Gleason correction program. Similar plots are shown for thealternate gear set in Figures 26 and 27.

After final grinding, the selected gear sets were processed andassembled into a production main gearbox and a production acceptancetest (ATP) conducted.

The ATP is an integrated gearbox system back-to-back test run on theUH-60 main gearbox (see Figure 28) in the UH-60A Test Facility beforeit is installed on the aircraft. Since this test is part of theproduction qualification process, the test gear box is notdisassembled for detail inspection unless there are signs of surfacedistress, or excessive concentrations of load, such as scoring,surface pitting, or chipping.

All gears and pinions which were ground and measured using theZeiss/Gleason enhanced correction program have demonstrated goodperformance in the production gearbox, and none were the cause ofgearbox rejection.

Task 5 - Economic Analysis

Based on a manufacturing gear lot size of 20 gears, the projectedsavings in inspection and manufacturing time realized from theinstallation of the enhanced measurement process described herein wasestimated to be 1.72 hours per gear. The following analysis showsthe equivalent dollar savings and resulting cash flow over a fiveyear period.

Basis for Economic analysis

The data upon which the economic impact of the enhanced spiral bevelgear inspection process is based is shown in Table 2. It assumesthat 50 percent of the BLACK HAWK and SEA HAWK spiral bevel gears areproduced at Sikorsky Aircraft, and estimates the benefits derivedsolely from that production for each year.

39

CI'

a, PC)4I- C .)o C

F- LU flC

H; cv_ wi

z l~i I--

ED~

4-)

C2l)

z I U)a,

U-4-)

0 4J

I-I 0 -U

In 0 4-4om j 04q

I-4~ I cu itLn w >

z 3,I

DD

En<C

w (n 44

LL z Cý N<I

Lz0 z 0

40

+ +

0 0'-AwL z a)

U)0 0 a)

LtI2

0z *

o- ZiccI m

wL 00D 0 41

z

0 Lo W 0 U

('1 . 4 En -q

F- m: ) 0H

LL (1 44cu0

-J 0LL- Z

0 L

0- W

0 ccw<-M

1-lU 0 tnU0J H 4 rE

41-

CU->-'cn-4

U)

00

M 3m o C30~ 4-)

00

oc

0 4,

CC000

I~ z

z cl () Icr)

U ') 4-)

(n.1 (0-

Z 3 ujju cu 1-

OZZ w -taa

LI~Q L) >

z4

01 C;

F- 00 = 4-

w -ca

F- cowL U) C3 4

I I a

ED

/I,

ILLCD

004T- 0 -

<CU.n C H

z >w I~

I Z W' C L4

mr U ) P

xzo0 0) Dz

-i CC, 0 0 -

43

Figure 28 - UH-60 Test Main Gearbox

TABLE 2. BASIS FOR ECONOMIC ANALYSIS

No. of aircraft - BLACK HAWK, SEAHAWK, spares

year 1 2 3 4 5

A/C 158 163 163 163 i58

17 bevel gears per aircraft50% of gears produced at Sikorsky1.72 hrs saved per gearLabor rate - $14.48 per hour (1991 dollars)Overhead rate - 223%Tax bracket - 40%

Income/Expense Statement

Table 3 lists the annual dollar savings and costs associated with theenhanced insrpction method in each of the five years. Table 4presents the annual and cumulative cash flow situation.

Discussion of results

It has been demonstrated, in this program, that the automaticcalculation of both first and second-order grinding machine changesworks very well, especially in conjunction with the new class of CNCbevel gear grinders represented by the Phoenix 400PG. Verification

44

ý4 Lf 00l C)0 D 1 )>1 w l C) cy' D c cq

v) c l > C 0(Y) 0w OD00l

U)(Y - C) C 0 Cýk

C)) r- r.-4LI) 0')(. v 0

zý4() r- C r-

CI >1 aU LO If) U' 0 Or

a- ) HY -r4- H HkDU) (d[-H ~

z

S >1 C> I) If) C - ) U)

c')m HT H W0

H-H H

(aC~

* >1

a)Q )

C; 0.ir to C)0) )

U)> -

U) UrI4)9 ) A) 0)0U~a p U)H -))

04 0 -1 04 U) C0C:-' a)l 0HlV)H

> ) a)- 0 (0U' >' U' 0p 4 - 44-

'UE- 0~Q~4 0V)

"I ) a

04

0*JU

Q ý4X

0 ~ a)~-45U. I (

$.IU- co (Y) (V Y) O>1 (Y C\J4 ( k

>1 LA) N-r r

U) 000 co -

CN

>1 LAN N N

CO 0-0. OD 00 L

>1 U-; r- r- GVz OD OD CID t

.r-4 OD(y) )

* >1 m )N NI4) ODz %W- - IU) 0k

>1 (UY)

U 00CO LA

rI-

0)0)

0 0 $4

04 04 >

4-) 4J `4 -4 4)

4) 41 4.)m 44- 0 ý

46

measurements on two selected production gear sets have shown that thebevel gear tooth profile can be held to within acceptable limits withonly one or, at the most, two iterations.

The inclusion of the automatic second order change capability hasresulted in an additional savings of 1.72 labor hours per gear.

Based upon the cash flow picture presented in Table 4, the calculatedpresent worth, with an assumed acceptable rate of return of 23percent, is $185,083 for this second order change enhancement.

Based upon the success of this program, the final step in theautomated inspection process for spiral bevel gears is now possible.This involves a closed-loop or hard-wire interface system linking theZeiss coordinate measuring machine with the Gleason CNC Pheonix geargrinder. This completely automated system is expected to be in placeat Sikorsky Aircraft within the next three years.

47

CONCLUSIONS

1. An enhanced inspection method for spiral bevel gears involvingautomatic first and second-order change capability wasdemonstrated and verified.

2. The validated process automatically calculates first and second-order grinding machine setting changes necessary to correct anout-of-tolerance spiral bevel gear tooth profile in only twogrinding cycles.

3. Manufacturing and inspection time for spiral bevel gears isreduced by 1.72 hours per gear, resulting in significant costsavings.

4. The enhancement was demonstrated on two selected BLACKHAWK/SEAHAWK gear sets on both the Gleason #463 and thePhoenix 400PG grinders. The process worked much better on theCNC Phoenix with fewer grinding iterations.

5. All gears inspected with the enhanced process were subjected toa final ATP test in a production gearbox without any signs ofsurface distress or abnormal distribution of load.

6. The technology developed in this program can be applied to allbevel gears manufactured by Sikorsky Aircraft and Supplierswhich use the Zeiss/Gleason system.

7. This technology was required to permit the successful operationof the Phoenix Grinder.

8. The enhanced inspection system will produce higher-quality gearswith fewer anomalies in acceptance test results.

48

REPOR DOCMENTTIONPAGEForm Approvedl

REPOT DCUMNTAIONPAG OMB No. 0704-0 r88Public reporting burden for this collection of information is estimated to average I hour per response, including the time for reviewing instructions searching existing data sources.gathering and maintaining the data needed, and completing and reviewing the collection of information Send comments regarding this burden estimate or any other aspect ot thiscollection of information, including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for information Operations and Reports. 1215 JetfersoriDavis Highway. Suite 1204. Arlington. VA 22202-4302. and to the Office of Management and Budget, Paperwork Reduction Project (0704 0188) Washington. DC 20503

1. AGENCY USE ONLY (Leave blank) 2. REPORT DATE 3. REPORT TYPE AND DATES COVERED

March 1992 Technical Paper4. TITLE AND SUBTITLE 5. FUNDING NUMBERS

Enhanced Automated Spiral Bevel Gear Inspection

WU-505-63-36

C-NAS3-25961SAUTHOR(S)

Harold K. Frint and Warren Glasow

7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) 8. PERFORMING ORGANIZATION

United Technologies Corporation REPORT NUMBER

Sikorsky Aircraft DivisionStratford, Connecticut (16602 None

9. SPONSORINGIMONITORING AGENCY NAMES(S) AND ADDRESS(ES) 10. SPONSORINGIMONITORINGNASA Lewis Research Center AGENCY REPORT NUMBERCleveland, Ohio 44135-3191and NASA CR-189125Propulsion Directorate AVSCOM-TR-91--C--048U.S. Army Aviation Systems CommandCleveland, Ohio 44135-3191

11. SUPPLEMENTARY NOTES

Project Manager, Timothy L. Krantz, Propulsion Systems Division, NASA Lewis Research Center, (216) 433-3580.

12a. DISTRIBUTIONJAVAILABILITY STATEMENT 12b. DISTRIBUTION CODE

Unclassified -UnlimitedSubject Category 37

13. ABSTRACT (Maximum 200 words)

This report presents the results of a Manufacturing Methods and Technology program to define, develop, and evaluatean enhanced inspection system for spiral bevel gears. The method utilizes a multi-axis coordinate measuring machinewhich maps the working surface of the tooth and compares it to nominal reference values stored in the machine'scomputer. The enhanced technique features a means for automatically calculating corrective grinding machinesettings, involving both first and second order changes, to control the tooth profile to within specified tolerance limits.This enhanced method eliminates the subjective decision making involved in the tooth patterning method, still in usetoday, which compares contact patterns obtained when the gear set is run under light load in a rolling test machine. Itproduces a higher quality gear with significant inspection time and cost savings.

14. SUBJECT TERMS 15. NUMBER OF PAGES

Spiral bevel gears; Gear inspection; Gear manufacturing; Zeiss I !l,,; Gcars 5616. PRICE CODE

A04

17. SECURITY CLASSIFICATION 18. SECURITY CLASSIFICATION 19. SECURITY CLASSIFICATION 20. LIMITATION OF ABSTRACTOF REPORT OF THIS PAGE OF ABSTRACT

Unclassified Unclassified Unclassified

NSN 7540-01-280-5500 Standard Form 298 (Rev 2-89)Prescribed by ANSI Std Z39 16298-102