,24 GEAR TECHNOlOGV

Multi- MetalComposite Gear/Shaft

TechnologyRaYll10nd .J...Clax.ton

Mater,iall!s Analysis, Inc...ID.allas,TXAlbert SOl Wadlleigh

Inte.rfac·e Welding', ICarson, leA

Ab tractA research program. conducted in conjunc-

tion with a U.S ..Army eontraet, ha resulted inthe development of manufacturing technologyto produce a multi-metal composite gear/shaftrepresenting a substantial weight savings com-pared to a solid steel component. Inertia weld-ing is used to join a steel outer ring to a light-weight titanium alloy web and/or shaftthrough the use of a suitable interlayer materi-al such as aluminum.

Fabrication is accompli hed in a two-stepinertia welding operation first joining steel 10

aluminum and then joining titanium [0 the alu-minum side of the steel/aluminum weldment,

surface-hardened by techniques uch c carbur-izing .. nitriding or carbo-nirriding. This proce-dure impart. 1.0 the tooth contact urfacestheu eful properrie of high hardne s, strength andwear re istance, The remainder of the gearcomponent very often only transmits torsionalloads and does not require the hardness andstrength properties characteristic of the alloysteel contact surfaces. In a program begun in1987 for the U.S. Army, a technology for PI'O-dueing lightweight gears and gear hafts hasbeen developed tharallow the use of hardenedsteel in the power tran mission contact sur-face and lighter-weight titanium alloy in thenon-contact area of the web, bub and/or shaf].

The development program, carried out over a Since the density of titanium is 43% le thanfive-year period, included testing and metal-lurgical evaluation of torsion components,small fatigue test gears and prototype demon-stration sun gears for the Allison 250-B17Fgearbox. [11 addition 10 development of fabri-cation technique, criteria were established toaUow design of a new or retrofit componentusing the multi-metal welding technology,

A prototype multi-metal compo ire lJJngearperformed withou: incident when ubjected toa [6O-hollr irnulated flight endurance test inagearbox. te trig.

IntroductionGears used in high-performance aerospace

application are typically forged from a singlepiece of alloy steel. Gear component for cer-tain applications are then selectively or totally

that of steel, slgnlflcam weight savings areoffered by the concept, provided the gear com-ponent geometry and contact surface configu-rations allow a igrnficant titanium-fer-steelsubstitution.

The use of lightweight, high-strength alloysfor engine component. is nalural.ly an attrac-tive proposition. Case-hardened steelalloysenjoy a well-established position of provenreliability and repre ent the very core of gear-box design technology. The objective of theprogram begun in 1.987 was to develop amethod for using surface-hardened steel whereit was required and lighter weight titanium inother locations within the gear component.Merallurglcalirrcornpatibilities, however,.l1aveto dale prevented titanium from being welded

------~---------------------------------------------------

or otherwi e metallurgically bonded directly [10

steel. This program developed a method forjoining steel and titanium alloys by the solid-state inertia welding proce suing aluminumas a mutually compatible mterlayer materialbetween Lite two, The ini.tial project was COIl-

ducted over a period of flve years by MaterialsAnalysis, Inc., with its primary subcontractor,Interface Welding, under the allspices of U.S.Army AVSCOM, Ft. Eusti , VA. Additionalconrractessistance in the design, manufactureand test of flight-quality hardware was provid-ed by Alii 0.11 Engine Company (formerlyAllison Gas Turbine Div. GMC).

Manufacturing Proeess Development.The five-year program to develop the multi-

metal gear technology involved the de ign andfabrication by inertia welding of test pecirnensfor the followlng purpo es:

I. To develop the feasibility of the multi-metal approach.

2. To evaluate the metallurgical integrity ofthe multi-metal joint,

3. To generate the mechanical properties to'be used in the de ign of test and prototypecomponent.

The inertia welding machine which wasemployed by Interface Welding for fabricationof the majority of the te J coupons, mechanicaltorsion test specimen and fatigue test gears ipictured in Fig. I. A typical alloy teel te I

coupon ring is pictured in Fig. 2. In generalthese rings were 3,75" O.D,. and 2.2" I.D. withiii conical bevel machined in the LD. all oneside of the ring. The entrance diameter of theconical bevel on the face of the ring was main-tained at 2.90". The conical bevel machined inthe ring pictured in Fig, 2 was configured suchthat the beveled urface was inclined at anangle of 45° from the rotational axi of thering. For the evaluation of joint geometryeffects, which axe discussed sub equently, theconical bevel on the weld surface was inclinedat angles of both 45° and 30° from the axis ofrotational symmetry. Although weld parame-ter and machine settings differed somewhatfor the two different weld configurations. thetechnique employed for development of theweld procedures was imilar.

Fig. 3 pictures a typical alloy ,steel ringchucked in the spindle collet of the inertiawelding machine at location l. The steel mass

Fig. Jl- Inertia welding machine used in majori-ty of welding trials.

Fig. 2- Allo}'steel ring prepared for inerUa weld-ing wilh ..So cnnlcal bevel.

Fig..3 - Iner.in ~\eldlillg machlne setup for we'ld-ingaluminum. bar 10 alJoy teel ring.

surroundingthe steel ring at Location 1 intheinertia welding machine is commonly referredto as the flywheel mas . Pictured at location 2in Fig, 3 is a section of 3.5", diameter 606l-T6aluminum alloy round bar with a conical malebevel machined on one end. For the particularweld procedure pictured in Fig. 3, the matingsurfaces of both the aluminum conical malebevel and the steel conical female bevel wereinclined 45° from the axi of rotational. sym-metry of the pan .

After the two component to be inertia.welded are properly machined, cleaned andchucked in the inertia welding machine. thespindle, containing the collet and attached tothe flywheel pictured 'toward the len side ofFig. 3, begins to rotate. The rotational speed of

IRaymond J.ICllaX10nis the founder ofMaleri(l/s Analysis, ln«.The activities ofMaterial.I' Analysisinclude metotlurgicaland mechanical engi-neerins; design, II'.SI;'lll,

failure analysis andresearch.

A~beIl1S.Wad!I'eighIS Presiden! of tnterfac«We/diIiX. whicll isinvolved with design,metallurgy and produc-tion of parts using theinertiaifriction welding

JANUARYJF'EBRUARY 19·&5 25process.

26 GEAR TECHNOlOGY

tile spindle assembly is a variable parameterthai is selected for the individual weld. Afterthe pindle and flywheel 8S. embly reaches 8

preset rotational speed (rpm),lhe spindlea sembly is disconnected from the drive rnech-ani m in the inertia welder and allowed to free-wheel with only the preci 'e kinetic energy(inertia) provided by rhe rotating mass of theflywheel assembly. The flywheel mass itself isvariable 8I\d is elected during machine etupba ed on the material 8I\d configuration of thecomponents to be welded. While the pindleand flywheel as embly are rotating freely. thetail stock, containing the prernachined alu-minum bar picturedtowardthe fight of Fig. 3,is stationary. The rotating component i thrustagain I the stationary component by hydraulicpre sure provided by the inertia weldingmachine. The freely rotating alloy steel ring,pictured at Location 1 in Fig, 3, surrounded bythe spindle and flywheel mas. is thru t againstthe non-rotaring component. pictured atLocation 2 of Fig. 3. The resulting frictionbetween tne mating surface material of thetwo components, D6ac steel and 6061- T6 alu-minum.generates enough heat to raise tiletemperature of the aluminum component to

Fig. 4 - Electron micrograph of multi-meta]weld cross section showi,ng Ti 6AI-4V. 6061Al,D6ac steel components, lower left to up.per right.

F,ig. S - Elcdl"on mi rograpboll weld inlerfacebetween Til 6·AI-4V, lower I' fl, and 60MAI" upp rright. (1S0X)

the forging temperature range, but not into themelting temperature range. In addition to cre-ating heat and providing deformation, or nash.of the two components at their interface, thefriction between the two components alsobrings the flywheel and, pindle assembly to 3.

stop, Thus the rotating (kinetic) energy of thespindle and flywheel as embly is trnn formedinto thermal and strain energy, which cau ethe two components to be weldedtogetherwithout either component changing to 'the liq-uid phase duringthe proce s. The resultingweldment is then removed from the inertiawelding machine. and the aluminum bar is cutoff adjacent to the completed weld,

After completion of the first inertia weld,the aluminum center (interlayer) in the inter-mediate weld coupon was machined to create aconical female bevel on the inside surface ofthe steel/aluminum subassembly. At the con':elusion of machining. the intermediate weldcoupon with a newly machined aluminum sur-face was chucked in the inertia weldingmachine in a manner similar to that previoaslydescribed for the [eel ring component. Inertiawelding of the titanium alloy bar to the alu-minum interlayer was accomplished in a man-ner similar to thai described for the aluminum-interfayer-to- teel weld, although differentweld parameters were employed. This two-stepinertia welding technique was employed withseveral variations to produce test coupon andcomponents required iothe R&D program.

For the coupon configurations describedabove, the inertia. welding proce s generallyproduces welds of rotational symmetry.Machine settings, or weld parameter . whichcan be varied in the process. are flywheel.mass, spindle rpm (weld speed) and weld pre -ure (the force with which tile rotating weld

component is 'thrust against the non-rotatingweld component during the welding process),The length of total upset, or axial length loss,lis a IU eful parameter to monitor; however, it ia variable resulting from the three machine set-ting parameters ..the weld configuration and theproperties of the materials being welded. Allof the e parameter are typically determinedshrough a weld development program.

After development. of the welding processparameters suitable for multi-metal weldmentfabrication, metallurgical micro sections were

taken through the composite joints of each ofthe alloy combination evaluated. Fig. 4 pic-tures at 35X magnification an electron micro-graph showing the overall appearance of ajoint profile of weldment materials Ti 6AI-4V,6061AI and D6ac steel; lower left, center andupper right, respectively .. The titanium/alu-minum inertia weld profile is pictured at 750Xmagnification in Fig. 5 with the Ti 6AI-4Vand 6061AI materials pictured lower left andupper right, respectively. Simi Iarl y, theteet/aluminum profile is pictured at 750X

magnification in Fig. 6 with the 6061AI andD6ac reel alloys pictured lower left and upperright, respectively.

Mechanical CharactierizationMechanical characrerizarionconducted OIl

the multi-metal composite weldments consistedof microhardness surveys and tatictorsion test-lng, Microhardness surveys conducted on pol-ished cross sections, such as the one pictured inFig. 4, revealed that the heat treated propertieof the titanium and steel alloy components werevirtually unaffected by the inertia weldingprocess. The aluminum alloy interlayer, howev-er, being the lowest melting and the mo t sensi-tive to heal of the three materials in the stackup,was briefly ubjected, during the inertia weldingproce s, to temperature which were at or nearthe solution treatment temperature for the all oy .Consequently. the 606 [-T6 aluminum barstock, which was the primary interlayer materialused, experienced over-aging and thermal soft-ening at locations immediately adjacent to thebond lines. Metallographic examination of the ethermally oftened region did not reveal thepresence of any defective condition, such aseutectic melting.

Specimens for mechanical torsion testingwere de igned and fabricated using the multi-metal inertia welding proce s. An example ofone such torsion specimen is pictured in Fig ..7.Evident in the picture are the two Inch-squaredrives machined into the titanium component tothe right and the steel. component. to the left sideof the photo. Specimens were made with inter-layer thicknesse in the range of 0.020-0.050".A further joint geometry variable was evaluatedby fabricating torsion specimens with the coni-cal, aluminum interlayer oriented at 30° and45° from the rotational axis of the specimen,

The torsion testing machine itself consisted

Fig. ,6- Electron micrograph of weld interfacebetween ,6061AI, lower left, and D6ac steel, upperright. (7S0X)

Fig. 7 - Typlcal compo ite tor ion specimenprior to testing.

of a drive shaft mounted in a turdy set of bear-ings and operated. through a 16-in. crank armby a servohydraulie cylinder fitted with anLVDT and a 50,000 lb. load cell. The digitalcontrol system and hydraulic power supplywere comparable to those that would beemployed on a servohydraulic mechanical test-ing machine. Nine each of the 30° and 45°specimens were successfully torsion tested tofailure in this program. The average maximumtorque for the 45° weld specimens was 9,409ft-lbs., while the corresponding value for the30° specimens was 13,302 ft-lbs. Geometricalanalysis of the surface area of the conicalweldment in each of the two weld angle con-figurations revealed total weld areas of 3.96square inches and 5.6 square inchs for the 45°and 30° configurations respectively. When themaximum torque values were calculated on aunit area basis, it was discovered that for boththe 300 and 45° conical weldment configura-tions, the maximum torque values reduced to2,375 ft-lbs, per square inch of interlayer mate-rial Tins torsional trength consistency wouldprove useful during subsequent stres analysiand design activities associated with the manu-facture M flight-quality demonstration gears.

J ...N U A R Y f FeB R U A R V 1 H 5 2J

28 GEAR TECHNOLOGY

The 30° interlayer torsion test specimen.pictured in Fig. 7. is shown after torsion testingto failure in Fig. 8. The interiayer fracture : ur-face was carefully examined on all torsionspecimens to veriiylhal fracture had occurredentirely through the aluminum interlayer, andthat no de-bonding of the various componentshad occurred during the testing sequence. Fig.9 is an electron Iractegraph, taken at IOOOXmagnification, showing the typical shear dim-ple microfeature as ociated with the shearoverload fracture mode in the aluminum inter-layer material.

The fracture path through the inter layermaterial on pecimen subjected to static tor-sional overload bore certain characteristicsworthy of note. The shear fracture path in athick-walled lube composed of a homogeneousand isotropic material would be predicted tooccur on atransverse radial plane; however, in

Fig. 8 - Fracture surfaces of 300 int,erlayn tor-sion specimen aft.er testing ..

Fig. 9 - Electron micregraeh 01 sheared alu-minum :interl'ayer from tor ion specimen fracturesurface. (1000X)

Fig. 10 - Pair of multi-metal fatigue 'lest gears.

the case of the composite weldment, fracturethrough the aluminum is forced to occur alonga line defined by Ihe configuration ofthe coni-cal interlayer. The longer the path of the frae-ture, the greaterthe strain energy required tocreate new surface during the fracture process.[0 appreciably thick: interlayers, the fracturepath occurred from one side of the interlayeron the 0.0... (0 the 01'1'0. ite ide of the interlay-er on the 10. in uch a manner as to minimizethe cross sectional area of the fracture. Thefracture was simply passing through the path oflea I re istance, minimum area. In the eighteentorsion te t weldrnents 1I ed for mechanicalbehavior characterization, an effort was madeto maintain the interlayer below .0..050" inthickness . Once the imerlayer wa reduced to. athicknes on the order of 0.050.". and below,the tendency to. fracture from one side of theinterlayer at the 0.0. to the opposite side of theinterlayer at the LD. was not as predictablyobserved. At the mall imerlayer thicknesses,other factor , such as local thermal softening ofthe aluminum alloy. probably exerted an influ-ence on fracture path. This ameliorating effectof a thin interlayer configuration is consistentwith the fact that the thinner the Interlayer, theless difference ill strain energy requirementsfor the v,e.ry small. angle of optional fracturepaths through the inrerlayer, For this thin inter-layer condition, other factors influencing frac-ture stress are probably of the same order ofmagnitude as the variations possible fromavailabfe fracture paths. As the interlayerbecomes thicker, the optional fracture pathband increases, as does the percentage differ-ence in strain energy required for separationalong thevarious possible fracture paths.

In summary. joint geometry exerted a sub-stantial influence on the static torsion strengthof a composite weldment This influenceappears to be directJy related to the ere ec-tional area of surface that the fracrure separa-tion is forced to create. For thicker interlayerthere is 3 distinct. tendency for the fracture topass through the path of least resistance fromone side of the interlayer at the O.D. of theweld to the opposite side ef the interlayer al

the 1.0. of the weld. The tendency observedwa for the fracture path topa s through theweaker interlayer material, but at the sametime to cut comers and pass through the short-

These test. conditions were selected to be inertia welding process. Selection of the gearsimilar tothose employed routinely by Allison to be manufactured included fabrication andto test steel gears. Employing this standard testgear design and test protocol did not. subjectthe tri-meral weld joint to either temperature ortorsional stress that would be consideredsevere. The test did, however, subject the tri-metal weld to vibratory loading similar to thatexperienced in an aircraft gearbox.

It is well-established that in most. aircraftgearbox applications, oil supply temperaturesof I75-250°F are normal. In this test, the oilsupplytemperature was maitnained at 119°F toprovide a thick lubricant film and prevent

est, most nearly radial path possible. As theinterlayer material becomes thinner, the strainenergy difference between the optional fracturepaths becomes less, as does the tendency of thefracture to predictably follow a given path.

Small Fatigue Test GearFabrication and Evaluation

One of the major goals of this project was toevaluate the performance of the multi-metalgear manufacturing concept in a gearbox. envi-ronment. Simulation of such a gearboxenvi-ronment had been uccessfullyachieved byAllison using a four-square gear fatigue testmachine. TIlls machine, reportedly developedby Allison, was similar in design to rigs used atNASA Lewis Research Center. A standardgear configuration has beenestablished forevaluation ill the four-square test rig. Themulti-metal inertia welding concept was incor-porated into this standard gear configurationwhile maintaining all exterior dimensions of[he gear. The multi-metal inertia weldingprocess was employed to fabricate gear blankssuitable for subsequent machining of the testgear configuration. Several pairs of smallfatigue test gears were manufactured with apitch diameter of approximately 3.3". Fig, 10pictures a typical set of these test gears with atitanium aHoy hub, aluminum alloy inter layerand hardened alloy steel teeth. The smallfatigue test gears were tested :in the four-squarer,ig under the following conditions:

1..Speed 10,000 rpm2. Torque 400 in-lb.3. Oil Inlet Temperature 119°F4. 011 Type M1L-L-23699C5. Duration 50 hrs ..6. Max. Contact Stress 261 ksi

scuffing or scoring ofthe gear teeth. The actualsurface temperature of the gears is usuallysomewhat higher than that of the oil supply.The web of the test gear is over-designed toprovide rigidity and ease of manufacturing.Because of this, the maximum stress in theweld interlayer during the test wa. calculatedto be 0.168 ksi, For comparison, the torsionaldesign allowable stress utilized in the demon-stration gear was 27.2 ksi,

Post-test inspection of the fatigue test gearsafter the scheduled teardown indicated ateardrop-shaped contact patch on the gearfaces. This contact pattern, pictured on the vis-ible faces of the right hand gear in Fig. II.wasuniform in size, shape and I.ocation. The gearsshowed no sign of pitting or other distress thatwould have prohibited continued testing. Thetested gears passed magnetic particle inspec-tion according to EIS Il69, Method D, andfluororescent particle inspection according toMethod AMS 2640.

Aircraft gears are generally designed withcontact stresses kept below 160 ksi, Using theAmerican Gear Manufacturers AssociationMethod 2001, the successful operation of thisgear for 50 hours at 261 ksi is equivalent to

300,000 hours operation at 160 ksi, Althoughnot statistically significant, the success of thistest verified the soundness of the approach loamulti-metal composite gear and paved the wayfor design and manufacture of a. full-scale,f1ighr.-qual.i.ty,demonstration gear.

Demonstration Gear Design,Fabricatien and Test

One of the primary goals of the project wasW produce a sample quantity of flight-qualitydemonstration gears usingthe multi-metal

Fig.U- Contact surfaces of mum-metal. faCiiguetest gears after testing.

JANIJARY/FEBRUARY 1995 29

30 ClEAR TECKNOLOClV

application considerations. The completed partwould be subjected to high frequency dynamictesting and simulated flight endurance testing.

It is necessary to consider several importantcriteria when designing a multi-metal gear orgearshaft, some of which can only be evaluat-ed through testing of a prototype part. A fewof these criteria follow.

1. Weight savings must be realized sum-cient to justify increased manufacturing costassociated with the multi-metal, process.

2. Weld geometry and strength must beconsidered. The [ower-strength aluminuminterlayer material may necessitate a largershafting cross section at the Iocation of theinertia welds.

3 .. The difference in modulus of elasticitybetween steel and titanium introduces inherentstiffness differences between the multi-metalgear and an. an-sled gear. Design. modifica-tions may be required to maintain satisfactorygear web or shaft dynamics when designing acomposite component

4, In some cases special tooling is requiredto control weld component concentricity dur-ing the fabrication process. Excess runoutfrom eccentricities introduced during weldingcan lead to poor spline and gear loading anddetrimental vibration.

5. The complexity of the component to beproduced may simply require so manyrwo-srepinertia welds or so much specialized toolingthat it may be impractical to replace in its cur-rent configuration with a multi-metelgear.

The demonstration gear was selected by ateam from Materia.ls Analysis, InterfaceWelding and Alii on. The major considerationof the team in its final selection process was toselect a gearthat, when manufactured using themulti-metal inertia welding process, wouldsubject the weld locations to relatively highloads and at the same time would be a relative-ly simple gear shape 10 fabricate. The intentwas to select a gear shaft that was not everlycomplicated to fabricate, but that would chal-lenge the ultimate performance of the welds byproviding a high service load application. Aftercareful consideration the un gearshaft, PIN23033882 from tile Allison 2S0-B 17F gearbox,was selected as the demonstration gear to befabricated Ilsi.ng the multi-metal technology.Prototypes would be subjected to a perfor-mance evaluation and comparison to theincumbentall-steel sun gear shaft.

In order to evaluate the applicability of sev-era] manufacturing techniques, it was decidedto produce a small quantity of flight-quality,demonstration sun gears using two parallelmanufacturing approaches to produce thehardened steel portions of the gear shafts, Onone set of demonstration gears, dual pulseinduction hardening (DPIH), an. Allison pro-prietary process, would be employed to indue-lion surface harden the gear and spline teeth asthe final step in the multi-metal sun gear man-ufacturing process. A second set of demon-stration sun gears would be manufacturedemploying conventional gas carburizing to

,0003 A·a

If------------. 12.850 --------------l0003 A-B ,

St!*

WeldS

• Ex. 148797 - 9310 (ASM 6265) Gas Carburized Steel orEx 148798 - D6ac (AMS 6431) induction Hardened Steel

Fig. 12 - Allison 250·817F mull i-rnetal slIn gear shaft weld locations.

prefabricated 9310 aUoy steel gear and splinecomponents prior to the inertia welding fabri-cation process and finish grinding of the sungear component.

Weld locations were selected within thegear shaft so as to maximize the amount ofsteel replaced with titanium and thus maxi-mize weight savings. These weld locations arepresented in the schematic gear shaft crosssection of Fig. 12. The welds involving alu-minum interlayers (welds 1 and 2 and weld 4and 5), were designed as conical welds withthe interlayer oriented at an angle of 30° fromthe rotational axis of the gear shaft. Weld 3,identified in Fig. 12, was simply a Ti 6AI-4Vbutt weld employed in the fabrication processto reduce machining costs and material waste.

The torsion test information previouslydescribed was used to determine the allowabletorsional shear stress for the weld joints in themulti-metal. sun gear, The failure torques forthe 30° torsion specimens were compiled, andre ulting torsional shear stresses were calcu-lated. The resulting torsion design allowablestress within the interlayer was found to be27.2 ksi. Analysis of the sun gear shaft pre-dicted torsional stress in the shaft at the splineend to be 26.3 hi. Since this was below the27.2 ksi design allowable, no modification tothe shaft section thickness was required. Sincethe service stresses at the location of the shaftspline were considerably greater than those inthe region 0:1' the spur gear, the tri-rnetal weldin the spur gear web was also expected to beof sld'fici:ent. strength without increasing theweb thiclrnes at the weld location.

A lateral and torsional dynamic analysiswas performed at Allison. These dynamicstudies were conducted to verify that the tor-sional and lateral natural frequencies werelocated weUaway from the frequencies associ-ated with normal operation. Sufficient marginbetween the operating speed and both torsion-al and lateral natural. frequencies assured noexcessive engine vibration or dynamic loadingof engine components.

Once the design analysis for the 250 sungear was complete, manufacturing procedureswere conducted in parallel for the D6ac and9310 multi-metal gear shafts. The inertia

welding process parameters were developed Fig. ~5_ Spline end 0('9310 multl-metal demon-for each of the five welds cited in the gear stratioa sun gear.

----- - - -------~-~~~~~~~~~~~~-~-~~~~

shaft schematic of Fig. [2 ..Two of the completed demonstration gears

are pictured in Fig, .13. In that photo the gearshaft with 9310 alloy steel and gas carburizedteeth is shown at the top of the photo and thegear shaft containing induction hardened,D6ac gear teeth is shown at the bottom of thephoto. The geometry of the two gears is iden-tical. Physical weighing of one of the finishedcomposite sun gear shafts showed u to weigh1.75 Ibs., a weight savings of 28% over the2.45-10. weight of the incumbent all-steelcomponent. The gear shaft with the 93 W alloysteel teeth was inertia welded into the tri-metalgear form using previously manufactured gascarburized steel parts, The gear shaft with the

F.ig. 13 - Mu'lti.-metal demonstration sun gearswith gear and spline portions eomposed of 9310,tnp, and D6ac. bottom.

Fwg. 14 - Spur gear end of 9310 multi-metaldemonstration sun gear.

32 GEAR TeCHNOLOGY

D6acaUoy steel components was fabricatedby inertia welding and completely machinedprior 10 induction hardening of the steel per-tions of the part. Cia e-up views of the spurand spline ends of the 93 W demonstrationgear are pictured in Figs .. 14 and 15, respee-uvely, Fig. 16 and .17picture the internal pro-file of the multi-metal weldments on the SpUTand spline end of a demonstration sun gear.

One of the 9310 multi-metal sun gear shaftswa. subjected to dynamic tre s characteriza-tion at Allison. This work complemented theanalytical. tor ional and lateral dynamic analy-sis previously described. The physical testingassociated with dynamiccharacterization wasdesigned to further verify that the natural fre-quencies of the lin gear were located farenough fromthe operating speed frequenciesso fhu no excessive engine vibration ordynamic loading of engine components wouldlikely occur. Physical test results indicated thecomposite gear shaft exhibited natural frequen-cies very close to those observed ill the all-steel production gear shaft. Since both the tri-metal gear shaft and the incumbent all-steelgear shaft exhibited very similar high frequen-

F'ig..16 - Segm-nc removed from spur gear endof .D6ac raultl-metal sun gear fur m.etallugraphic'evaluation.

Fig,. 17 - Segment removed from spline end ofD6ac mum· metal sun gc.ar fur meta'llurgj.eal eval-uat'iun.

cy vibrational behavior, and s:ince no problemhas been observed in the long history of opera-tion of the all-steel sun gear shaft, it seemedsale to conclude that no vibrational difficultieswould be expected with the dynamically simi-lar compo ite sun gear shaft.

The high frequency vibrational. behaviorimilarities between the multi-metal. composite

sun gear shaft and theall-steel versionof thaIeornpenent wa alii unexpected finding lin light'Of the significant difference in weight, 2.45 lbs.versus 1.75 lbs. for theall-steel and multi-metal sun gears, respectively, One possibleexplanation for this observed v ibrationalbehavior similarity is that, although the densityof titanium is considerably less than that ofsteel (0.160 .lb/in.3 versus 0'.282 Ib/in.3), tiledifference in the elastic modulus of the twomaterials (] 6.5 x [()6 psi for titanium versus 30x 1.06psi for steel) resultsin a ratio of elasticmodulus/density which is virtually identical forthe two material .

Finally, one ofthe 9310' alloy multi-mesalsun gears was ubjected to a 160-hour simulat-ed flight endurance test by Allison in a BI7Fgearbox Lestrig. This test consisted of approxi-mately 500 duty cycles, each consisting of idle,take-off, cruise and landing service segments.The multi-metal composite sun gear shaft. per-formed without. incident and was undamagedduring the test. Successful completion of thesimulated fl:ight endurance test verified that themulti-metal composite gear technology is uf~ficienUy mature for incorporation iatoa fun-scale engine development program and isaviable candidate for retrofit, where feasible,into existing power transmission and accessorygearbox application.O

Acknowledgements: The authors wish to acknowl-edge the technical input and contractual ,~llpport ofMr. Clay Ames. AATD, Ft. Eustis, VA, and the valu-able technical support and assistance of Messrs.DO/lg 'Wagner and Ron Sciloolcr(tji, Allison EngineCo~ptJny, Indianapolis, IN.

Tell Ills What You 1bink..Jf you found this article 'ofinterest and/or useful. please circle 'Reader ServiceINumber A.-49.