1、06FTM08An Evaluation of FZG Micropitting TestProcedures and Results for the CrownedAGMA Test Gearsby: Dr. D.R. Houser, S. Shon and J. Harianto,The Ohio State UniversityTECHNICAL PAPERAmerican Gear Manufacturers AssociationAn Evaluation of FZG Micropitting Test Proceduresand Results for the Crowned A
2、GMA Test GearsDr. Donald R. Houser, Sam Shon, and Jonny Harianto, The Ohio StateUniversityThe statements and opinions contained herein are those of the author and should not be construed as anofficial action or opinion of the American Gear Manufacturers Association.AbstractThis paper discusses the p
3、rocess of developingan understandingof thestresses andwear predictorsusingdurability tests, with a focus on micropitting. The FVA Information Sheet No. 54/I-IV was used as an initialprocedure, and later modified due to the higher contact stress levels predicted for the crowned AGMA testgears. A new
4、procedure involving strain-gaging shafts on the FZG test machine was developed to producebetter repeatability as well as more precise setting of load levels. Additionally, two oils were used and themicropittingwearamountsarecompared. Topographymeasurementsoftoothprofiles,predictionsofcontactstresses
5、 and contact patterns, and comparisons of micropitting wear to these parameters as well as to flashtemperatures and local film thickness are made.Copyright 2006American Gear Manufacturers Association500 Montgomery Street, Suite 350Alexandria, Virginia, 22314October, 2006ISBN: 1-55589-890-41An Evalua
6、tion of FZG Micropitting Test Procedures and Results for the CrownedAGMA Test GearsDr. Donald R. Houser, Sam Shon, and Jonny Harianto, The Ohio State UniversityIntroductionThis paper reports on the surface fatigue testing ofgears that were manufactured as a part of theAGMA macro-pitting and tribolog
7、y testing program1. The overall goal of this work is to develop mod-els for predicting micropitting wear and to ascertainthe appropriateness of the AGMA test gears for mi-cropitting wear evaluation. As part of this goal, thestudy reported upon in this paper focuses on devel-oping an understanding of
8、 the stresses and wearpredictors using FZG durability tests. Since the fo-cus is on micropitting, the first tests used the meth-od described in FVA Information Sheet No. 54/I-IV2. Later, the procedure was modified to accountfor the much higher contact stress levels that arepredicted for the heavily
9、crowned and tip relievedAGMA test gears. This paper provides extensiveanalysis that includes detailed topography mea-surements of the tooth profiles, predictions of con-tact stressesand contactpatterns, anddiscussionsregarding factors that affect contact stresses, flashtemperatures and local film th
10、ickness of the testedgears.Testing SummaryThegearstestedinthisstudyarepartofthesecondset ofgears manufacturedby AGMAand haveseri-al numbers, 150 to 159. Since this study was a“learning experience” using the heavily modifiedAGMA test spur gears,the testingprogram wases-sentiallyaworkinprogress, withi
11、nitial resultscaus-ing methodologies to change as we progressed.Listed below is a summary of the tests that wereperformed with their results:1. Staged loading at the FVA loading stages:Loads were at the low end of the range used byOConnor 1, Hoeprich 3, and Buzydon andCardis 4 so they were felt to b
12、e a reasonablestarting point. Results showed micropittingwear and topographychange evenat thelowestloads, while at higher loads, macropitting oc-curred. Also, a peculiar “T-shaped” contactpat-tern was observed.2. Based on the above results, more detailedstress predictions were made, a more detaileds
13、tudy of the FZG loading procedure was per-formed, and the procedure for measuring toothwear at each loading stage was refined.3. Further tests were performed with staged load-ing,butatloadlevelsthatcompensateforthein-creased contact stresses of the crowned gears.Micropitting still occurred, causing
14、further con-cern about the predicted contact stresses.4. Using revisedloading stages,a higherviscosity,EPlubricant(ISOviscositygrade220)wasusedas a replacement for the MIL L 23699 lubricantthat was used inthe previoustests. Micropittingwear was much less, but still occurred at thehigher loading stag
15、es.Tooth TopographiesThe tooth geometry of the test gears is shown inTable 1 and the print specifications for leads andprofilesareshowninFigs.1and2. Theleadspecifi-cation of the test gear is straight with a tolerance of0.005 mm.It should be noted that the major “break” of the tiprelief is about half
16、way between the highest point ofsingletoothcontactandthetipofeachtooth.Figs.3and 4 show lead and profile measurements on fourteeth of each gear. Comparing the charts, the pro-filesandleadsare madevery closeto thespecifica-tion, however, the lead crown appears to be slightlygreater than the specifica
17、tion. As will be shown lat-er,themeasuredvalueofleadcrowndependsuponwhere on the tooth height the measurement is tak-en. During initialtesting ofthe gears,a ratherpecu-liar, T shaped contact pattern that is shown in Fig.5appeared, thus implying that the tooth topographywas not as simple as expected
18、from the original in-spection charts. Therefore, more extensivemeasurements of the tooth topographies wereperformed.2Figure 1: Profile and lead specifications forAGMA test pinion 5Figure 2: Profile specification for AGMA testgear 5Table 1: Geometry of AGMA Test Gear Sets 5Dimension Unit Pinion(Gear1
19、)Gear(Gear2)Number of Teeth - 20 30Module mm 3.62857Center Distance mm 91.5Pressure Angle deg 20Contact Ratio - 1.58Face Width mm 14 20Outside Diameter mm 81.64 116.36Root Diameter mm 64.338 99.162Base Diameter mm 68.194 102.293Operating PitchDiametermm 73.57 108.86Standard PitchDiameter (SPD)mm 73.
20、20 109.80Tooth Thickness atSPDmm 6.415 5.072SAP Roll Angle deg 9.280 12.111EAP Roll Angle deg 37.710 31.065Figure 3: Standard Lead and ProfileMeasurements of the Pinion3Figure 4: Standard Lead and ProfileMeasurements of the GearFig. 6 shows such a set of measurements on onetooth,where16leadchartsare
21、shownononegraphand 16 profiles are shown on the other graph. Thelead crown varies from 0.02 mm at the root to 0.06mm at the tooth tip. This would imply that tool shiftwas used to obtain the crown and the amount oflead crown that occurs due to tool shift will beproportionaltothesineofthepressureangle
22、. Inthiscase,thepressure anglevaries from9.1 degreesattheSAPto33.3degreesatthepiniontipsotheratioof the sines of these angles is almost exactly theratio of the lead crowns observed in themeasurements.The leads and profiles of each set of the nine testpinions and gears are virtually the same within0.
23、0025 mm so it is fair to say that separate stresscalculations for each gearset wouldnot beneeded.In the early testing, various versions of single andmultiple profile and lead inspections were per-formed after each load stage. However, the finalprocedure that was chosen requires measurementof both th
24、e conventional single lead and profile onfour teeth spaced at roughly 90 degrees apart andthemeasurementof16leadsand16profilesononetooth. Teeth are numbered and marked so that thesame teeth are measured after each stage.Figure 5: T-Shaped Contact Pattern on Pinionand Gear after Load Stage 5 (origina
25、l loads)4Figure 6: Detailed lead and profile measurements for a single tooth on the pinion and gear,respectivelyFigs. 7 and 8 show the respective profile measure-ments of the pinion and gear following the finalstage of the revised loading procedure (to be dis-cussedlaterinthispaper). Here,itcanbesee
26、nthatthe micropitted wear region shows up as “hollows”intherespectiveprofilechartswiththehollowoccur-ring in the root region of the driving pinion and nearthe tip relief “break” in the driven member. Thedashed lines ofFig. 7show theprogression oftoothwear as the load stages were increased, with the
27、fi-nal dashedline correspondingwith thewear areaatthe end of the final loading stage. In all cases, thewear seemed to start near the extreme end of thecontact patch (root region of pinion and tip of gear)and progress towards the pitch point.Contact PatternsInitially, Dykem bluing was used to detect
28、contactpatterns. Since contact was extremely repeatablefrom one tooth to the next and because the patternwas detectable by simply observing the “polishing”of the tooth, bluing was discontinued andphotographs of the polished pattern after eachstageofloadingweretaken. Also,magnifiedphotoswere taken of
29、 any regions that had any indicationofwear on the measured profile charts.The contact patterns of the worn pinion and gearteeth are shown in Fig. 9. The wear occurs in theellipticalpatternatthetipoftheTofeachofthepat-terns, pointed out by the arrows. The contact pat-tern width in the base of the T w
30、idened as load wasincreased,muchastheorywouldpredict. Thewearat the tip of the T, however was unexpected and re-sultedinmuchanalysisthatisreportedlater. Fig.10shows a magnified view of the worn region.5Figure 7: 16 Profile Traces for Tooth 1 of the Pinion after the Final Loading Stage (revised testi
31、ngprocedure), with outline of wear pattern progressionFigure 8: 16 Profile Traces for Tooth 1 of the Gear after the Final Loading Stage (revised testingprocedure), with outline of wear pattern progression6Figure 9: Final Contact Pattern of Pinion andGear Under the Revised Procedure (afterendurance s
32、tage)Figure 10: Magnified Worn Region Pinion(20x magnification)Durability Testing MethodologyPrevioustestingoftheAGMAgearswasdoneusingaconstantloadandtrackingthewearandpittingviaeither weight reduction or visual examination of thetooth degradation 1. Since there has been grow-ing interest in micropi
33、tting, the FVA 54/I-IV proce-dureformicropittingwaschosenforthistesting. Ini-tially, the original FVA load levels that are specifiedinTable2wereused. MilL23699lubricant,aratherlowviscosityturbineoilthatiscommonlyusedintheaerospaceindustryasatransmissionlubricant,waschosenasthelubricant. FVAprocedure
34、swereusedin the testing with the exception that dip lubricationwas used.SincetheOSUGearLabhasseveralFZGmachinesandisinterestedinrunningidenticaltestsoneachofthe machines, we became very interested in under-standing machine-to-machine variability. There-fore, for the testing machine used in this stud
35、y, thevariability of tooth loading was also studied.Sincetherevisedloadingstagesrequirelowerloadsthan the initial FVA load stages, a new, lightweightmoment arm was designed. With the lower loads,the accuracy of the static loading methodology be-came increasingly important and it was decided toapply
36、a torque measurement strain gage bridge tothe torsion shaft. This allows study of the repeat-ability of thedead weighttorque applicationmethodandthegeartorquevariationas onerotates thetestgears. A photo of the setup is shown in Fig. 11.Figure 11: FZG Machine Setup7Table 2: Load Levels of 54/I-IV Pro
37、cedure (note that the stage torques are not the same as thoseof the conventional FVA durability test)Load Stage No. Torque on Pinion (Nm) Hertzian Contact Pressure at PitchPoint (N/mm2) for the FVA Test GearRunning In (16 hours) 28.8 510.05 (16 hours) 70.0 795.16 (16 hours) 98.9 945.17 (16 hours) 13
38、2.5 1093.98 (16 hours) 171.6 1244.99 (16 hours) 215.6 1395.410(16 hours) 265.1 1547.3Endurance (80 hours) 171.6 1244.9The dead weight moment arm was used to obtainthe calibration curve shown in Fig. 12. Since theangleofthemomentarmvariesasloadisapplied,acorrection was first made for this angle when
39、com-puting the applied torque. Likewise, the inherenttorque of the moment arm itself was added to thecalibrationcurvedata. Loadswereappliedinsuchamannerthatthehysteresisduetoimpendingfrictioncould be ascertained at each load position. Thebands of the calibration curve indicate that this hys-teresis
40、is significant and if one simply applied theweight without knowledge of this hysteresis, therecouldbetorqueapplicationerrorsashighas10%offull scale. For this reason, in all later testing, theload was set using the torque bridge output asopposed to simply using the dead weights.Back-to-back testers s
41、uch as the FZG machinealso have tooth to tooth torque changes due to thedirection of friction forces, spacing errors and run-outofeachpartthateffectivelychangestheamountoftorque thatis lockedin thetorque loop. Byslowlyrolling the loaded gear train over three shaft revolu-tions, the torque trace of F
42、ig. 13 was obtained.Here, the torque variation from tooth to tooth addsanother 15% to the uncertainty of the applied load.Themajorperiodicityisat thepinion rotationspeed,indicating that the index errors of the pinion are themajor factor in creating the torque change that isshown. This variation cann
43、ot be improved bychangingthemeasurementprocedureandcanonlybe improved by reducing manufacturing errors thatresult in index errors of both the test and slavegears. It should also be noted thatthe meantorquethatthetestgearsseeisabout2Nmhigherthanthetorque setting of 46 Nm.Figure 12: Shaft Torque Calib
44、ration Curvewith Increasing and Decreasing LoadHysteresisFigure 13: Torque Fluctuations through 3Pinion Revolutions8Load Distribution and Contact StressCalculationsLoad distribution simulations using the Load Dis-tribution Program (LDP) 6 were performed to de-termine the local contact stresses in th
45、e contactzone for both the profile and leads modificationsspecified on the print and for the as measured pro-file modifications. In addition, local flash tempera-tures and lubricant film thickness were predicted tosee if the location of peak flash temperature andminimum lubricant film thickness corr
46、elate with thewear location that was observed.Fig. 14 shows a series of contact stress plots forthree different loads that were applied and showsvia the outer boundary how the predicted contactpattern changes with load for both the print profiles(constant lead crown) and the pinion with the ta-pered
47、leadcrown. Asexpected,thecontactareain-creaseswithloadforeachcase,withthehigherloadcontact pattern covering nearly 100% of the tooth.The contact pattern of the tapered lead crown toothis wider at the pinion (GEAR1) root since the leadcrownislessinthis regionrelative tothe leadcrownnear the tip. This
48、 could be one reason for the mea-suredTshapecontactpatterns. However,themea-surements show a much more exaggerated Tshape.One other factor that has not been taken into ac-count in the above contact stress analysis is the ef-fect of the tip relief on the local radius of curvature7. If the break is sh
49、arp, the radius of curvaturewould approach zero, and extremely high contactstresses could be observed. For the test pinion(GEAR1),thisregionoccursbetween30and32de-grees of roll and the radius change is significantenough to increase the predicted contact stressesas is shown in the Fig. 15. For the test gear(GEAR2),thisregionoccursbetween26and28de-grees of roll (at mesh, the corresponding GEAR1roll angles are between 14 and 17 degrees). Thefirst set of contact stress patterns use the taperedlead crown and consider the profile modification ra-dius,
