1、90 FTM 8Investigations on the Scuffing Resistanceof High-Speed Gearsby: H. Winter, K. Michaelis and H. F. Collenberg,Technical University of MunichiJ,American Gear Manufacturers AssociationI IlllTECHNICAL PAPERInvestigations on the Scuffing Resistance of High-Speed GearsH. Winter, K. Michaelis and H
2、. F. Collenberg,Technical University of MunichThe 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.ABSTRACT:Test results with a high speed four square gear test rig (nmax = 26 0
3、00 rpm) show the influence of speed,viscosity, base oil, type and amount of additive on the scuffing load. In some cases at high speed thescuffing load can be more than twice the value calculated according to ISO/DIS 6336/4. The reason forthe speed-dependency of the scuffing load is explained by the
4、 kinetics of the chemical reaction betweenthe metal and the additive. It is described, how a calculation method could take account of the highscuffing load at high speed.Copyright 1990American Gear Manufacturers Association1500 King Street, Suite 201Alexandria, Virginia, 22314October, 1990ISBN: 1-55
5、589-560-3Investigations on the Scuffing Resistanceof High-Speed GearsH. Winter, K. Michaelis and H.F. Collenberg *)Test results with a high speed four square gear test rig (nmax=26 000 rpm) show the influence of speed,viscosity, base oil, type and amount of additive on the scuffing load. In some cas
6、es at high speed the scuffingload can be more than twice the value calculated according to ISO/DIS 6336/4. The reason for the speed-dependency of the scuffing load is explained by the kinetics of the chemical reaction between the metal andthe additive. It is described, how a calculation method could
7、 take account of the high scuffing load at highspeed.1 Introduction It is known, that the scuffing load of mediumspeed gears is lower than that of low speed gears.Besides tooth breakage and pitting, scuffing is one For high speed applications, however, increasingof the main factors, limiting the pow
8、er transmitting scuffing load was first reported in the 1950s bycapacity of gears. Of course, scuffing damage can Borsoff 1. Ku and Baber 2, Lechner 3, Droz-nearly always be avoided by an adequate amount dov and Gavrikov 5 and Seitzinger 4 also foundof Extreme Pressure (EP)additives. But especially
9、a minimum in scuffing load at a certain speed,in cases, when the lubrication of gears is not the while other investigators, e.g. Schauerhammer 6,only task of the lubricant, it is often not possible to did not. None of these authors could give a corn-add enough EP-additive to give sufficient scuffing
10、 plete explanation for this effect, and no commonlyprotection: used calculation method considers the increasingscuffing load at high speed.In applications where the lubricant is also used asa hydraulic fluid (e.g. for the control of steam- or It was the objective of the research work, the re-gas tur
11、bines or in hydraulic torque converters in sults of which are presented here, to investigatelocomotive drives) the air release property of the the main factors which influence the increasingoil may be reduced because of EP-additives. scuffing load with increasing speed, to contributeto possible expl
12、anations and calculation methods.- Marine gears are often lubricated with diesel engi-ne oils. 2 Test Rig, Test GearsThe oxidation properties of the oil as well as seal A four-square gear test rig with a mechanicallymaterials are affected by EP-additives. closed power circuit having 91.5 mm center d
13、istan-ce was used (Fig. 1). The pinion speed is conti-*) Prof.Dr.-Ing. Hans Winter is Emeritus of the Gear Re- nously variable between 500 and 26 000 rpm, thesearch Center (FZG) at the Technical University Munich maximum pitch line velocity is 100 m/s (= 20 000(head:Prof. Dr.-Ing. B.-R. HOhn).Dr.-In
14、g Klaus Michaelisis ft/min). The torque is applied while the gears arechief engineer, Dipl.-Ing. Hans F. Collenberg is research running using a hydrostatic load device, measuredengineerat the FZG. by strain gages and controlled by an electrical con-trol circuit. The maximum pinion torque is The dist
15、ribution of load, speed and contact time is350 Nm, which corresponds to a Hertzian stress of shown in Fig. 2 and Fig. 3, the main data of thepc = 1500 N/ram 2 (=217 500 psi) at the pitch point standard test gears type A and C are shown inof the test gears (type A and C). Fig. 4. The material of the
16、test gears is20 Mn Cr 5, case carburized (case depth 0.6 - 0.8, “; ram), surface hardness 60 - 62 HRC. Retained_ !_. austeniteat the surfaceis less than 20%._ normal load FbL _2o0oP,_, ?, _, _, t,h“,114,o_Nlmm_ /J _ooJ_- _I _m_Hx 1600_ k oo-“-_ 120Q- Hertzian / / I_“ fl 500“_“- / / _ 4-400_oo_ / / _
17、2ooF._oo/ / _t.orque _ /0 _ 400_“ / /_ 100_-200_ 0 j,it,i, , i _ _ O I n -i-0 468 10 12 14o_ load stageFie. 2 : Relation between load.stages, torque, normal loadFbt and Hertzian pressure Pc at the pitch line of the test.u u_ gears_O._o 0 1 2 3 4- 5 6 kHz 8i 1,1,1 ,111 ,1,1,m/s- _ /AC-type_cgours/so
18、vs_60 A-type_Qe o,r“sC-typeX 20 _gectrsVSDFour square high speed gear test rig with hydro- ostatic load device. Center distance 9t.5 ram. gear speed n2Maximum transmittable power I000 kW _ _ _ ?,_, _,llO,112,1,4,1,6rl_m210 x 10 _0 _ I 112 16 210 IrpLml310 5The temperature at the center of one pinion
19、 tooth, pinion speed n 1(1“ module below the addendum circle) is mea- Fig.3 Pitch line velocity vt, sum velocity v_x:at pitchsured by an K-type-thermocouple (chromel-alu- point C, sliding velocity vsE, VsD at the engagement pointsreel). One rotating amplifier each modulates the D and E of the test g
20、earstemperature and the torque signal and transducesthem by an infrared LED / phototransistor (PCM- All test gears are 15-MAAG ground. The devia-system) to the stationary display and registration tions of pitch, profile and involute were measuredunit.2typeA typeCcenterdistance a (mm) 91.5number of t
21、eeth (pinion / gear) z1 / z2 16 / 24module m(mm) 4.5toothwidth b (mm) 20working pitch diameter (pinion/gear) clwt/ dw2 (mm) 73.2 / 109.8addendum diameter (pinion / gear) dal da2 (mm) 88.77 / 112.50 82.64 / 118.64pinionx1 0.8635 0.1817addendum modification factorgearx2 -0.5000 0.1715pressure angle a
22、n (deg) 20.0helixangle 1“5(deg) 010.00 100.00 10.00 100.00Hertztan8.00 80.00 width bH 8.00 80.00,0 Y t,ftm | ftmcontact2.00 20.00 “ hmetl,t 2 2.00 20.000.00 0.00 0.00 0.00Hertziant t pressurePH,. ,. t IfL fHsum veloclty vT ll slid,ng vel. vs l vzlvtpitch tine; velcityvtvslvf !VslVt_ A C8 3 E A B C O
23、 E,_- “ geometry .,dwz _._2 :=“:-, -:.;.:.!/-:.,.z 50 m/s), on the addendum ofrate of 5.5 l/min, the oil temperature is 90 _ 3 C. pinion and gear there is nearly no material loss.The flanks look grey. In some cases, the grindingThe test starts at the lowest load stage No.1 (Fbt/b marks are still vis
24、ible in the scuffed areas. The= 50 N/mm). After 15 minutes running time the corresponding dedendum shows a large amount ofgears are inspected visually with a stereo micro- material loss, the flanks are very smooth. Thisscope. If there are no scuffing marks visible, the effect is independent of wheth
25、er the pinion or theload is increased by 20 % (load stage No.2) and gear are driving (Fig. 5, 6a,b; 7a,b).the same process starts again, etc In the medium speed zone (vt = 20.50 m/s)smooth parts with and rough parts without materi-If there are scuffing marks visible, the patterns of al loss are visi
26、ble in addendum and dedendum ofthe scuffed areas of pinion and gear are noted. If pinion and gear. The higher the speed, the lower isthe damaged area is more than 20 % of the active the material loss at the addendum.flank of pinion and gear the failure limit is reachedand the test is terminated. If
27、the failure limit is not The described effect of speed on the profile of theyet reached, the load is increased by only 10 %. scuffed area is nearly independent of the load orThe load corresponding to 20% of scuffed area of of the oil used. The scuffing appearance changes atpinion and gears is interp
28、olated and reported as that speed, where the scuffing load has its mini-scuffingload. mum.Because scuffing as an instantaneous failure is not The grey areas look macroscopically very similar toinfluenced by the load cycles, but mainly by the areas of micropitting, but the damage is different.tempera
29、ture of the gear, the same running time is Scanning electron micrographs (Fig. 6) show thatapplied for the whole speed range. The bulk-tern- the damage mechanism at low speed and at highperature of the gears has to reach steady state speed is the same: parts of material of the gearconditions. This i
30、s the case after 8-12 minutes run- flanks weld with the corresponding pinion flanks.ning time, depending on the load and only very Material is torn out of the dedendum flank andlittle on the speed. Therefore a running time of 15 adhered to the addendum surface. At high speed,minutes in each load sta
31、ge was chosen, these parts are very small and thin, so the grindingmarks are sometimes still visible under the weldedIt is known, that the scuffing load of gears is parts. At low speed the parts are larger andstrongly affected by smoothing of the surfaces due grinding marks can no longer be detected
32、.to running-in 7. For the test it is required, thatthe surface roughness does not change during the The reason, why at the beginning of the scuffingtest procedure. Therefore all gears are run-in befo- process always parts of the dedendum adhere tore the test at medium load (7 th load stage) and the
33、addendum of the mating gear is, according tolow speed (v t = 8.3 m/s). So the roughness of the Michaelis 7, that the shearing stresses below therun-in flanks before scuffing occurs is always the surface, resulting from frictional force and flashsame (R a = 0.39 _ 0.04 _m (pinion); R a = 0.30 +_0.07
34、_m (gear).A _“_ _“120_m “_ 10pm -A,32 890 _ 32 8/+2la Ibt-30802 , ._ 308002cFi2. 5a: Photographs of scuffed teeth (a: gear, b: pinion); changes of the involute and longitudinal profile related to scuffing1.: vt = 8.6 m/s, Fbt = 5.6 kN, A-type gear, pinion driving 2.: vt = 8.6 m/s, Fbt = 5.6 kN, A-ty
35、pe gear, pinion driving2c: Mating deviations of pinion and gear; BEa“ -AI I / 1 i32922 ,/ _., 32874A !/“_.;0, k2=0 ) according to the simplifiedand can be calculated according to the equation of equationArrhenius : Fe + A kt(ot) FeA: (3)k = A e _ (2) (a-x) xwith _ : temperature in K During the time
36、t1 a number x of additive compo-A,C : constants of the equation nents reacts:x =a “(1 - e -t4) (4)The constants A, C of equation (2) are differentfor k1 and k2. At medium temperature the net rate with k 1 according to equation (2),(k1 - k2) is positive, xe (the equilibrium value of x) _1 = (0z+#0)/2
37、is positive. The number of reaction products FeA t1 = (tg-tO)/2at equilibrium (t -, co) is higher than at the begin-ning of the reaction. If the temperature rises above t)., no further reac-tion layer compositlon is assumed (k20 , kl=0 ).At high temperature the rate of the forward reac- The reaction
38、 layer decomposes according to thetion becomes smaller than that of the reverse reac- equationtion (k 1 (y/X)min). The principal depen- time of contactdency of contact time tC on the critical contacttemperature eS (y/x) = (y/X)min) is shown in the Fill. 16 Influenceof contact time t on the critical
39、contactcalculated curve a of Fig. 16. For long reaction temperaturetime, the reaction (according to eq.(1) is balancedand independent of time. For short contact time,the critical temperature which is necessary for 14decompositionof the reaction layerincreases, kN -A .,_,12 -T no damage9 Relation Bet
40、ween Contact Temperature and L_ _-“ZhContact Time at Scuffing _ 10 _t_ / f6 _uoThe time tC one point of a pinion flank is in con-tact with the gear flank is the time that the point t_ 8._needs to cross the Hertzian contact width (2. bH). _ 6 _ a2bH otx_ - (7) co -V1,2 4with Vl, v2: tangential veloci
41、ty of pinion, gear. 2 _ o testA_typeresultSgears toil Z49 /0 , I , I , I I , I ,The contact time increases with load and decreaseswith speed and distance from the addendum of the 0 20 40 60 80 m/s 120tooth (Fig. 4). pitch line velocity vtFig. 17 Dependency of scuffing load on speedFig. 16 shows the
42、relation between the contact a: calculationwith constant criticaltemperatureace. tosurface temperature _C for test results, calculated Fig. 16, curve c (ISO 6336/4, CTC)in point D according to 9 and the contact time t1 b:calculationwithtimedependentcriticaltemperature0sat D (see Fig. 4), compared wi
43、th the calculated ace.to Fig. 16, curvebtheoretical relation according to equations 2, 4 and6 (curve a). The theoretical curve fits the tendency 10 Calculation of Scuffing Loadof the measured points very well.According to ISO/DIS 6336/4 9 the critical value,which must not be exceeded is themaximum l
44、ocal and instantaneous total contacttemperature (Contact Temperature Criterion,CTC), in the original according to Blok 11,13respectively the perature-time relation fits the results of gear testsrather well.- mean weighted surface temperature across the A proposal is made, how the calculation methodc
45、ontact according to Michaelis 7 (Integral according to ISO/DIS 6336/4 9 could takeTemperature Criterion ITC). account of the increasing scuffing load at highIn 9 both limits are assumed to be independent speed.of speed, so the scuffing load decreases with loadup to highest speed. Therefore the calcu
46、lation of 12 Acknowledgementgears lubricated with non-EP oils is very conserva-tive at high speed (Fig. 17,curve a). The authors would like to thank the Deutsche For-To improve the calculation methods at high speed, schungsgemeinschaft (DFG, German Researchit must account for the fact, that the crit
47、ical tempe- Association) for the financial support of this re-rature _S is dependent on the contact time. Fig. 16 search work.shows an example. The calculated curve (a)according to chapter 8 is approximated by two 13 Referencesstraight lines (b). This modification of the CTC-method gives the require
48、d result: the calculated 1 BORSOFF,V.N.:On The Mechanism of Gear Lubrica-tion. ASME Trans., Journal of Basic Engineering,and the measured scuffing load are close together March 1959,pp. 79-93(Fig 16, curve b).At this time further theoretical investigations of the 2 KU,P.M.;BABER,B.B.: The Effect of
49、Lubricants ontest results are performed to see whether this too- Gear Tooth Scuffing.ASLE Trans., 2, (1959)del is of general value to explain some other dif-ficulties in calculating the scuffing load (e.g. dif- 3 LECHNER,G.: Die Frel3-Grenzlastbei StirnraOernausStahl. Diss. TU Miinchen 1966ferent gear geometry, tip relief, influence of re-tained austenite etc.). These results will be pre- 4 SEITZINGER,IC: Die Erwarmung einsatzgehartetersented later. Stirnrlider als Kennwert
