RESEARCH LABORATORY - DTIC · parallel shafts. In helicopter applications, spiral-bevel gears are used in main-rotor and tail-rotor gearboxes to Various investigators have studied

AD-A263 116

NASA Army Research LabioratoryTechnical Memorandum 106080 Memorandum Report ARL-MR-71AIAA-93-2149

Low-Noise, High-Strength, Spiral-BevelGears for Helicopter Transmissions

DTICELECTEE

David G. Lewicki and Robert F Handschuh APR2 09

Vehicle Propulsion DirectorateU.S. Army Research LaboratoryLewis Research CenterCleveland, Ohio

Zachary S. HenryBell Helicopter Textron, Inc.Fort Worth, Texas

and

Faydor L. Litvin A2PpreY4 ~IPl= '~"

University of Illinois at Chicago wi.. u.

Chicago, Illinois

Prepared for the29th Joint Propulsion Conference and Exhibitcosponsored by the AIAA, SAE, ASME, and ASEEMonterey, California, June 28-30, 1993

U.S. ARMY

93-08197

I•ASA U!I i\ Ii RESEARCH LABORATORY

OR 4 19 091

VI Q I[ I I IM I I

LOW-NOISE, HIGH-STRENGTH, SPIRAL-BEVEL GEARS FOR HELICOPTER TRANSMISSIONS

David G. Lewicki and Robert F. Handschuh ACIA$ F

Vehicle Propulsion Directorate U | CRA&IU.S. Army Research Laboratory D T "

Lewis Research Center D U obCleveland, Ohio 44135

Zachary S. HenryBell Helicopter Textron, Inc. B __y _,

Fort Worth, Texas 76101 LDbr btXO' I

and .1 Awvmt doir•€lo

Faydor L. Litvin Di~t S tc•!

University of Illinois at ChicagoChicago, Illinois 60680 / I

Abstract future aircraft (Vialle, 1991). An effort to improve thetechnology of components such as spiral-bevel gears has

Improvements in spiral-bevel gear design were been the Advanced Rotorcraft Transmission (ART)investigated to support the Army/NASA Advanced program.Rotorcraft Transmission program. Program objectiveswere to reduce weight by 25 percent, reduce noise by The ART program was an Army-funded, Army/10 dB, and increase life to 5000 hr mean-time-between- NASA program to develop and demonstrate lightweight,removal. To help meet these goals, advanced-design quiet, durable drive systems for next generation rotor-spiral-bevel gears were tested in an OH-58D helicopter craft (Bill, 1990). The ART program goals were totransmission using the NASA 500-hp Helicopter Trans- reduce drive system weight by 25 percent, reduce noisemission Test Stand. Three different gear designs tested by 10 dB, and increase life to 5000 hr mean-time-included: (1) the current design of the OH-58D trans- between-removal by using new ideas in gear configura-mission except gear material X-53 instead of AISI 9310, tion, transmission concepts, and airframe-drive train(2) a higher-strength design the same as the current but integration. The success of the ART design configura-with a full fillet radius to reduce gear tooth bending tions in meeting the program goals depended on thestress (and thus, weight), and (3) a lower-noise design successful incorporation of certain critical, advancedthe same as the high-strength but with modified tooth technologies into the preliminary designs. The U.S.geometry to reduce transmission error and noise. Noise, Army Vehicle Propulsion Directorate, NASA Lewisvibration, and tooth strain tests were performed and Research Center, Bell Helicopter Textron (one ARTsignificant gear stress and noise reductions were contractor participant), the University of Illinois atachieved. Chicago (subcontractor to Bell Helicopter Textron), and

the Gleason Works (subcontractor to Bell HelicopterTextron) were involved in a joint project to improve

Introduction spiral-bevel gears. The project goals were to reduce bevelgear noise and increase strength through changes in gear

Spiral-bevel gears are used extensively in rotorcraft tooth surface geometry, and tooth fillet and root designsapplications to transfer power and motion through non- (Henry, 1991; Henry, 1992).parallel shafts. In helicopter applications, spiral-bevelgears are used in main-rotor and tail-rotor gearboxes to Various investigators have studied spiral-beveldrive the rotors. In tilt-rotor applications, they are used gears and their influence on vibration and noise (Litvinin interconnecting drive systems to provide mechanical and Zhang, 1991a; Gosselin, 1991; Fong and Tsay,connection between two prop-rotors in case one engine 1992). Most agree that transmission error, defined as thebecomes inoperable. Even though spiral-bevel gears have difference in relative motion of an output gear withhad considerable success in these applications, they are respect to the input pinion, is the major contributor toa main source of vibration in gearboxes, and thus, a undesirable vibration and noise. A common practice ismain source of noise in cabin interiors (Lewicki and Coy, to modify spiral-bevel gear surface topology to permit1987; Mitchell, et al., 1986). In addition, higher strength operation in a misaligned mode. Over compensation forand lower weight are required to meet the needs of this type of operation, however, leads to large

1l

trznsmission error and higher noise and vibration levels, support the bevel-gear shaft. Both pinion and gear areIn the Army/NASA/Bell joint project, gears with tooth straddle mounted.surfaces designed for reduced transmission errors usingmethods of Litvin and Zhang (1991a) were manufac- A planetary mesh provides the second reductiontured and tested. The teeth were designed to exhibit a stage. The bevel-gear shaft is splined to a san gear shaft.parabolic function of transmission error at a controlled The 27-tooth sun gear meshes with four 35-tooth planetlow level (8 to 10 arc set). The low level of transmission gears, each supported with cylindrical roller bearings.error reduces the vibration and noise caused by the The planet gears mesh wvith a 99-tooth fixed ring gearmesh. The new tooth geometries for this design were splined to the transmission housing. Power is taken outachieved through slight modification of the machine tool through the planet carrier splined to the output mastsettings used in the manufacturing process. The design shaft. The output shaft is supported on top by a split-analyses addressed tooth generation, tooth contact inner-race ball bearing and on bottom by a roller bear-analysis, transmission error prediction, and effects of ing. The 62-tooth bevel gear also drives a 27-toothmisalignment (Litvin and Zhang, 1991a; Litvin et al., accessory gear. The accessory gear runs an oil pump,1991b; Litvin et al., 1991r). which supplies lubrication through jets and passageways

located in the transmission housing.Also as part of the Army/NASA/Bell project, gears

with tooth fillet and root modifications to increasestrength were manufactured and tested. By increasing Spiral-Bevel Test Gearsthese radii, reduced stresses were achieved, and thus,increased strength. Tooth fillet radii larger than those on Three different spiral-bevel gear designs werecurrent gears were made Fossible by recent advances in tested. The first was the baseline OH-58D design.spiralbevel gear grinding technology (Scott, 1991). Table I lists a variety of parameters for this baseline set.Advanced gear grinding was achieved through redesign The reduction ratio of the bevel set is 3.26:1. The gearsof a current gear grinder ard the addition of computer were made using standard aerospace practices where thenumerical control. surfaces were earburised and ground. The material used

for all test gears was X-53 (AMS 6308) rather than theThe objective of this report is to describe the conventional AISI 9310 (AMS 6265).

results of the experiments to evaluate advanced spiral-bevel gear designs. The work was part of a joint Army/ The second spiral-bevel design tested was anNASA/Bell project in support of the ART program. increased strength design. The configuration was identi-Experimental tests were performed on the OH-58D cal to the baseline except that the tooth fillet radius ofhelicopter main-rotor transmission in the NASA 500-hp the pinion was increased from 0.51 to 1.02 mm (0.020 toHelicopter Transmission Test Stand. The baseline 0.040 in.), and the gear was made full fillet (Fig. 2). TheOH-68D spiral-bevel gear design, a low-noise design, high-strength design was made possible by recentand a high-strength design were tested. Results of noise, advances in gear grinding technology (Scott, 1991).vibration, and tooth strain tests are presented.

The third spiral-bevel design tested was a low-noisedesign. The low-noise design was identical to the

Apparatus increased-strength design except the pinion was slightlyaltered to reduce transmission error. The gear member

OH-58D Main-Rotor Transmission was unchanged. The low-noise design was based on theidea of local synthesis that provided at the mean contact

The OH-58 Kiowa is an Army single-engine, light, point the following conditions of meshing and contactobservation helicopter. The OH-58D is an advanced ver- (Litvin and Zhang, 1991a): (1) the required gear ratiosion developed under the Army Helicopter Improvement and its derivative, (2) the desired direction of theProgram (AHIP). The OH-58D main-rotor transmission tangent to the contact path, and (3) the desired majoris shown in Fig. 1. It is currently rated at maximum axis of the instantaneous contact ellipse. The localcontinuous power of 348 kW (464 hp) at 6016 rpm input synthesis was complemented with a tooth contactspeed, with the capability of 10 sec torque transients to analysis. Using this approach, the machine tool settings485 kW (650 hi), occurring once per hour, maximum. for reduced noise were determined. As with the high-The main-rotor transmission is a two-stage reduction strength design, precise control of the manufacturedgearbox with an overall reduction ratio of 15.23:1. The tooth surfaces were made possible by advances in thefirst stage is a spiral-bevel gear set with a 19-tooth final grinding operation machine tool (Scott, 1991).pinion that meshes with a 62-tooth gear. Triplex ball Figure 3 gives a topological comparison between a low-bearings and one roller bearing support the bevel-pinion noise and baseline spiral-bevel pinion tooth. The dottedshaft. Duplex ball bearings and one roller bearing lines are the baseline tooth datum and the solid lines are

2

the measured difference in topology of a low-noise gear input shaft. At 18 310 N (4120 lb) mast lift load, thecompared to the baseline. Solid lines above the dotted elastomeric corner mounts of the OH-58D trantsmissionplain indicate an addition of material and lines below housing deflect such that the transmission is properlythe plain indicate a removal. The effect of the topologi- aligned with the input shaft. (In the actual helicopter,cal change in the low-noise design was a reduction in this design serves to isolate the airframe from the rotoroverall crowning of the tooth, leading to an increase in vibration.)contact ratio and reduced transmission error.

The test transmission input and output shafts havespeed sensors, torquemeters, and slip rings. Both load

NASA 500-HP Hel'copter Transmission Test Stand cylinders on the mast yoke are mounted to load cells.The 149-kW (200-hp) motor has a speed sensor and a

The OH-58D transmission was tested in the NASA torquemeter. The magnetic particle clutch has speedLewis 500-hp helicopter transmission test stand (Fig. 4). sensors on the input and output shafts and thermo-The test stand operates on the closed-loop or torque- couples. An external oil-water heo~t exchanger cools theregenerative principle. Mechanical power recirculates test transmission oil. A facility oil-pumping and coolingthrough a closed loop of gears and shafting, part of system lubricates the differential, closing-end, speedwhich is the test transmission. The output of the test increaser, and bevel gearboxes. The facility gearboxestransmission attaches to the bevel gearbox. The output have accelerometers, thermocouples, and chip detectorsshaft of the bevel gearbox passes through a hollow shaft for health and condition monitoring.in the closing-end gearbox and connects to the differen-tial gearbox. The output of the differential attaches tothe hollow shaft in the closing-end gearbox. The output Test Procedureof the closing-end gearbnx connects to the speed de-creaser gearbox. The output of the speed decreaser gear- Two sets of the basedine design, two sets of thebox attaches to the input of the test transmission, high-strength design, and one set of the low-noise designthereby closing the loop. were manufactured and tested. Noise and vibration tests

were performed on all sets of each design. One set of theA 149-kW (200-hp) variable-speed direct-current baseline design and one set of the high-strength design

(dc) motor powers the test stand and controls the speed. was instrumented with strain gages and strain tests wereThe motor output attaches to the closing-end gearbox. performed on these gears. A description of the instru-The motor replenishes losses due to friction since power mentation, test procedure, and data reduction procedurerecirculates around the loop. An 11-kW (15-hp) dc is as follows:motor provides the torque in the closed loop. The motordrives a magnetic particle clutch. The clutch outputdoes not turn but exerts a torque. This torque is trans- Noise Testsferred through a speed reducer gearbox and a chaindrive to a large sprocket on the differential gearbox. The Acoustic intensity measurements wft,, -,rformedtorque on the sprocket applies torque in the closed loop using the two-microphone technique. The microphonesby displacing the gear attached to the output shaft of used had a flat response (±2 dB) up to 5000 Hz and athe bevel gearbox with respect to the gear connected to nominal sensitivity of 50 mV/Pa. The microphones werethe input shaft of the closing-end gearbox. This is done connected to a spectrum analyzer which computed thewithin the differential gearbox through use of a comn- acoustic intensity from the imaginary part of the cross-pound planetary system where the planet carrier power spectrum. Near the input region of the OH-58Dattaches to the sprocket housing. The magnitude of transmission, a grid was installed which divided thetorque in the loop is adjusted by changing the electric region into 16 areas (Fig. 5). For each test, the acousticfield strength of the magnetic particle clutch, intensity was measured at the center of each of the 16

areas. Only positive acoustic intensities (noise flowingA mast shaft loading system in the test stand sim- out of the areas) were considered. The acoustic intensi-

ulates rotor loads imposed on the OH-58D transmission ties were then added together and multiplied by theoutput mast shaft. The OH-58D transmission output total area of the grids to obtain sound power of themast shaft connects to a loading yoke. Two vertical load transmission input region.cylinders connected to the yoke produce lift loads. A14 000-kPa (2000-psig) nitrogen gas system powers the At the start of each test, the test transmission oilcylinders. Pressure regulators connected to the nitrogen was heated using an external heater and pumping sys-supply of each of the load cylinders adjust the magni- tem. For all the tests. #he oil used conformed to atude of lift. Note that in the OH-58D design, the trans- DOD-L-85734 spec;f,'cation. Once the oil was heated,mission at no-load is misaligned with respect to the the transmission input speed was increased to 3000 rpm,

.1

a nominal amount of torque was applied, and mast lift Strwii TeStsload was applied to align the input shaft (18 310 N,4120 Ib). The transmission input speed and torque were Twenty strain gages were mr~ounted on the spiral-then increased to the desired conditions. The tests were bevel pinions and :6 gages were Tuounted Ols tile spti-performed at 100-percent transmission input speed bevel gears- of one set each of the ba.hlti:e ad high-(6016 rpm) and torques of 50, 75, 100, and 125-percent strength design- (Figs. 8 &and 9), Gage- were po:,tinrdof maximum design. The tra.asmission oil inlet tempera- evenly across the tooth face width: with s•,ue in threture was set at 99 'C (210 *F). After the transmission fillet area and some in the root area of the teeth. Theoil outlet stabilized (which usually required about fillet gages were placed on the drive side of the teeth20 min), the acoustic intensity measurements were The fillet gages were also positioned at a point ,z thetaken. The time to obtain the acoustic intensity mess- tooth cross-section where a line at a 45' angle withurements of the 16 grid points was about 30 min. F-r resptc:t to the tooth centerline intersectE the tooth profileeach acoustic intensity spectrum at a grid point, 100 (Fig 8(b)) The fillet gages were placed there to measurefrequency-domain averages were taken. This data was maximum tooth bending stress Previous studies on spurcollected by a computer. The computer also computed gears showed that the maximurn' stresses were at a linethe sound power spectrum of the giids after all the 30° to the tooth centerline (Hirt, 1976) 4W was chosenmeasurement2 were taken. for the current tests to minimie the pvssiibilty of the

gages being destroyed due to tooth contact In additionto maximum tensile stress", root stressft can 1:,ecome

Vibration Tests significant in lightweight, thin-rimmed arrospaxe gearapplications (Drago, 1990). Thus. root gages were

Ten piezoelectric accelerometers were mounted at centered between teeth in the root to measure gear rimvarious locations on the OH-58D transmission housing stress. Tooth fillet and root gages were plaued on(Fig. 6). The accelerometers were located near the input successive teeth to determine loading consistency, Thespiral-bevel area (accelerometers 1, 2, and 10, measuring grid length of the gages was 0,3$1 mm (0.01S in,) andradially to the input shaft), the ring gear area (3, 4, the nominal resistance was 120 1l The gages wereand 9, measuring radially to the planetary), and on the connected to conditioners using a Wheatstone bridgetop cover (5 to 8, measuring vertically), Accelerometers circuitry and using a quarter-bridge arrangement.1 to 8 had a 1 to 25 000-Hz (±3 dB) response, 4 mV/gsensitivity, and integral electronics. Accelerometers 9 Static strain tests were performed on both theand 10 had a 2 to 6000-Hz (±5 percent) response and spiral-bevel pinions and gears. A crank was installed onrequired charge amplifiers, the transmission input shaft to manually rotate the

shaft to the desired position (Fig. 10). A sensor wasThe vibration tests were performed in conjunction installed on the transmission output shaft to measure

with the noise tests. After collecting the acoustic intens- shaft position. At the start of a test, the transmissionity data for a given test, the vibration data was recorded was completely unloaded and the strain gage condition-on tape and processed off-line. The vibration data was ers were zeroed. Conditioner spans were then determinedlater analysed using time averaging (Fig. 7). Here, the using shunt calibrations. The transmission was loadedvibration data recorded on tape was input to a signal (using the facility closed-loop system) to the desiredanalyzer along with a tach pulse from the transmission torque, the shaft was positioned, and the strain readingsinput shaft. The signal analyzer was triggered from the were obtained using a computer. This was done for atach pulse to read the vibration data when the transmis- variety of positions to get strain as a function of shaftsion input shaft was at the same position. The vibration position for the different gages. At the end of a test, thesignal was then averaged in the time domain using 100 transmission was again completely unloaded and theaverages. This technique removed all the vibration conditioner zeroes were checked for drift. All static testswhich was not synchronous to the input shaft. Before were performed at room temperature.averaging, the major tones in the vibration spectrum ofthe OH-58D baseline design were the spiral-bevel and Dynamic strain tests were performed only on theplanetary gear fundamental frequencies and harmonics. spiral-bevel pinions. The pinion gages were connected toTime averaging removed the planetary contribution, slip rings mounted on the input shaft. A slip ring assern-leaving the spiral-bevel contribution for comparing the bly for the spiral-bevel gear was unavailable, and thus,different design configurations. dynamic strain tests of the gear were not performed. The

4

test procedure was basically the same as the noise and shows the dominant spikes for the bseline design at thevibration tests, except that the tra~nsmission was not run spiral-bevel rmeshing frequencies, and the significantas long in order to maximize strain gage life. reduction in spiral-bevel gear vibration for the !ow-noise

design. The results of the other nine accelerometers westsimilar.

Results and DiscussionThe effect of torque on vibration for accelerometers

Noise Tests I and 5 is given in Fig. 14. Shown in the figure i's time-averaged accelera.ion processed up to 10 000 fi. The

The noise spectrum (sound power versus frequency) results are root-mean-square (rms) calculations of theat 100-percent torque is given in Fig. 11. The results time-domain signals. Since the time-averaging removedshown are for set I of the baseline configuration and vibration nonsynchronous to the input shaft, the resultsset I of the low-noise configuration. Among the domi- in Fig. 14 were basically the cumulation of the spiral-nant spikes in the spectrum for the baseline design are bevel meshing frequency (1905 flzi and second throughthe spiral-bevel meshing frequency (1905 Hz) and second fifth harmonics.harmonic (3810 Hz). Note that these tones are signifi-cantly reduced for the low-noise design. Other dominant As with the noise measurements, the vibration fortones in the spectrum are at the planetary meshing fre- the baseline and high-strength designs were similar butquencies (fundamental at 852 Hz). The planetary tones with scatter. Again, the figure clearly shows a significantwere not affected by the low-noise design. Tones from reduction in spiral-bevel gear vibration for the low-noisethe facility closing-end gearbox (Fig. 4) were also design compared to the baseline and high-strengthdominant in the spectrum (fundamental at 790 It) and designs. Like the noise results, the reduction in vibrationas expected were not affected by the low-noise design. for the low-noise design was greater at the higher

torques (100 and 125 percent). The results of the other

The effect of torque on sound power at the spiral- eight accelerometers were similar. From the results of allbevel frequencies is given in Fig. 12. Both sets of the 10 accelerometers and at 100-percent torque, the vibra-baseline and high-strength designs and one set of the tion for the low-noise design due to the spiral-bevellow-noise design are included. The sound power is the mesh was on the average 5 to 10 g's lower than that ofcumulation of the spiral-bevel meshing frequency the baseline and high-strength designs.(1905 Hz) and second harmonic (3810 Hs). The baselineand high-strength designs produced basically the samenoise since the difference between them was in the tooth Strain Testsfillet geometry. There was some scatter in the baseline

and high-strength results due to manufacturing toler- Figure 15 shows the results of a typical staticances of the different sets and assembly tolerances. To strain test of the spiral-bevel pinion. A uniaxial stresscheck assembly tolerances, the low-noise tests were field was assumed to exist at the strain gage and therepeated two times. Here, the gears were completed stress was determined by multiplying the measureddisassembled and reassembled in the transmission, and strain by Young's modulus for steel. For a pinion fletthe tests were repeated. The results showed the same gage, the stress was first compressive, then tensile. Sincetrend and were repeatable to within about 2 dB. The the pinion drove the gear, the compression occurredgeneral trend was a significant decrease in spiral-bevel when the tooth in mesh prior to the strain-gaged toothgear noise for the low-noise design compared to the was loaded, causing compression in the gage. As thebaseline and high-strength design. At 100-percent pinion rotated, the strain-gaged tooth was loaded intorque, the noise due to the spiral-bevel mesh was 12 to single-tooth contact and the gage measured the maxi-19 dB lower than that of the baseline and high-strength mum tensile stress. Similar conditions existed for thedesigns. Also, a decrease in noise was most prevalent at pinion root gage except the gage measured the stress of100 and 125-percent torque and less prevalent at 50 and the pinion rim rather than tooth bending. The results75-percent torque. for the spiral-bevel gear were similar to the pinion

except the tensile stress occurred before the compressionsince the pinion drove the gear.

Vibration ResultsFigure 16 shows the distribution of maximum ten-

The vibration spectrum (time-averaged acceleration sile and compressive stress during contact along theversus frequency) for accelerometer I (input spiral-bevel tooth face width for the baseline and high-strength de-housing) at 100-percent torque is given in Fig. 13. As signs. The most important item to note is the reduction

with Fig. 11, the results compare set 1 of the baseline to in maximum tensile bending stress of the high-strengthset 1 of the low-noise configuration. The figure clearly design compared to the baseline design. The maximum

5

tensile stress of the high-strength design was reduced on notse (sound power) due to the spiral-bevel meshing fre-the average 27 percent compared to the baseline for the quencies of the low-noise design was 12 to 19 d13 lowerspiral-bevel pinion (Fig. 16(a)). There was, however, an than that of the baseline and high-strength designs.increase in the ma.dimum compressive fillet stress for the Using a time-average processing scheme, the spiral-bevelspiral-bevel pinion. Thus, the alternating stress of the gear vibration of the low-noile design was 5 to 10 g'shigh-strength design was reduced on the average 14 per- lower than that of the baseline and high-strengthcent compared to the baseline (the alternating stress is designs.defimed as the maximum tensile stress plus the absolutevalue of the maximum compressive stress). For the 2. The increased fillet radius of the high-strengthspiral-bevel gear, the maximum tensile stress of the design had a significant benefit in decreasing toothhigh-strength design was reduced on the average 10 per- bending stress. For tests at 100-percent torque, thecent compared to the baseline and the alternate was spiral-bevel pinion maximum tooth bending stress of thereduced on the average 12 percent (Fig. 16(c)). Thus, high-strength design was on the average 27-percentthe increase in fillet radii of the high-strength design has lower than that of the baseline design. There was,a significant benefit in increasing the tooth bending however, an increase in the maximum compressive stresscapacity of the gear tooth. at the center of the tooth root.

There was a significant increase in the maximumcompressive root stress of the high-strength design corn- Referencespared to the baseline spiral-bevel pinion (Fig. 16(b)) anda slight increase for the gear (Fig. 16(d)). This was Bill, R.C., "Advanced Rotorcraft Transmission Pro-probably due to the removal of material for the in- gram,* NASA TM-103276, 1990.creased fillet, thus lowering the rim thickness. For the Drago, R.J., 'Design Guidelines for High-CapacityOH-58D design, this increase in stress is acceptable, but Bevel Gear Systems," A&-15 Gear Design, Manu-in general, these effects need to be considered in a facturing and Inspection Manual, Society of Auto-design. motive Engineers, Warrendale, PA, 1990,

pp. 105-121.Figure 16 also shows the results of the dynamic Fong, Z.H., and Tsay, C.B., 'Kinematic Optimization

strain tests for the spiral-bevel pinion. The results of the of Spiral Bevel Gears', Journal of Mechanicaldynamic strain tests matched closely to those of the Design, Vol. 114, No. 3, Sept. 1992, pp. 498-506.static. The stress-position plots were similar as well as Gosselin, C., Cloutier, L., and Brousseau, J., "Tooththe maximum and minimum stresses, indicating no Contact Analysis of High Conformity Spiral Beveldetrimental dynamic effects. Gears," Proceedings of the International Confer-

ence on Motion and Power Transmissions,Nov. 23-26, 1991, Hiroshima, Japan.

Summary of Results Henry, Z.S., "Advanced Rotorcraft Transmission (ART)-Component Test Results,' Presented at the 28th

Advanced-design spiral-bevel gears were tested in AIAA/ASME/SAE Joint Propulsion Conference,an OH-58D helicopter transmission using the NASA Nashville, TN, July "-8, 1992.500-hp Helicopter Transmission Test Stand. Three Henry, Z.S., "Preliminary Design and Analysis of andifferent gear designs were tested. The baseline design Advanced Rotorcraft Transmission,' AIAA/was the current design of the OH-58D transmission, ASME/SAE/ASEE 27th Joint Propulsion Confer-except the gear material was X-53 rather than AISI !, Proceedings, AIAA, Washington, DC, 14 p.,9310. The second design was a higher-strength design 1991.which was the same as the baseline but incorporated a Hirt, M.C.O., "Stress in Spur Gear Teeth and Theirfull fillet radius to reduce gear tooth bending stress. The Strength as Influenced by Fillet Radius," Disserta-third design was a lower-noise design which was the tion, American Gear Manufacturers Association,same as the high-strength design except the tooth Ph.D. Thesis, 1976.geometry was modified to reduce transmission error and Scott, H.W., 'Computer Numerical Control Grinding ofnoise. Noise, vibration, and tooth strain tests were Spiral Bevel Gears," NASA CR-187175, 1991.performed. The following results were obtained: Lewicki, D.G., and Coy, J.J., 'Vibration Characteristics

of OH-58A Helicopter Main Rotor Transmission,"1. For the baseline spiral-bevel gear design, domi- NASA TP-2705, 1987.

nant tones in the noise and vibration spectra occurred at Litvin, F.L., and Zhang, Y., 'Local Synthesis and Tooththe spiral-bevel meshing frequencies and harmonics. A Contact Analysis of Face-Milled Spiral Bevelsignificant decrease in the spiral-bevel tones resulted Gears," NASA CR-4342, 1991.from the low-noise design. At 100-percent torque, the

6

Litvin, F.L., Zhang, Y., and Chen, J., 'User's Manual TABLE I.-BASELINE SPIRAL-BEVEL GEAR DESICNfor Tooth Contact Analysis of Face-Milled SpiralBevel Gears With Given Machine-Tool Settings," PARAMETERS OF THE OW-SSD MAIN-ROTORNASA CR-189093, 1991. TRANSMISSION

Litvin, F.L., Kuan, C., and Zhang, Y., "Determinationof Real Machine-Tool Settings and Minimization Number of teeth,

of Real Surface Deviation by Computer Inspec- ger ............................ 62tion," NASA CR-4383, 1991. Diametral pitch ............................ 6,092

Mitchell, A.M., Oswald, F.B., and Coe, H.H., "Testing Pressure axgle, de .............. ......... 20of UH-60A Helicopter Transmission in NASA- Mean spiral angle, deg ......................... 3Lewis 2240-kW (3000-hp) Facility," NASA SW'h `'ge, deg .............................. 9$

TP-2626, 1988. Face width, mm (in.) ................. 36.93 (1.450)Fillet radiua, mm (in.)

Vialle, M., "Tiger MGB-, High Reliability-, Low pinion.............................. 0.1 (0.020)Weight,* Proceedings of the 47th American Heli- pgar ............................ 1.65 (0.065)copter Society Annual Forum Proceedings, Vol. 2,AHS, Alexandria, VA, 1991, pp. 1249-1258.

OutputPlanet

gear---OH-58D baseline design High-strergth design

Rintg '-Sun

gear / gear 0S mL10 n

•' • Spiral

" bevel A_0gear a) Pinion.

ۥInput 1 .91 mm

-Spiral bevel b) Gear.pinion Figure 2.--Compauison of OH-58D and high-strength spiral-

Figure 1.-OH-58D helicopter main-rotor transmission. bevel gear designs.

7

OH-S8O) baseline designLow-noise desiGn

25.4 l~n Drive side

material

Toe Topland Heel removed

Coast sideFigwxe 3.-Topological corrnparison of OM-58D and kiw-noiso

spiral-bevel pinion.

.- Closing-end gearbox

Mast Load cyrtinders

Magneft particle d~utch.. J

Figure 4.-NASA Lewis 500-hp helicopter transmiss ion test stand.

8

prooe noder-

7 8

a) Top view, b) S•Ue wew

Figure 5.---Grid for sound intensity measurements, FiWu•e 6 ---Acceferomeler locons on 0ti.580 transmmsson

m • , . Time- 1

I 17 ý

m•) I =l~tJ]•)[1averaged -, m_________ _ acceleration_

Tape i Computer

recorderTp ie.b)Sd ve

i ~averagesAccelarometerlanoericy ionm

an I

tach pulse _ -- Com--et

Signal analyzer

Figure 7.-Data reduction scheme of vibrasion tests.

9

a)ila l gagbesig

File a e10 Filletgages2

110

10 0 Spiral-bevel frequencies 1 201- " Planetary frequenciess0 h IC Facility closing-end

I eabx r~uncesC Baseline design, set 10.0

80 Baseline design L.100 Baghsteine t design, set 2Low noise design0 Hghsentdsinst1

V 0 0 High-strength, design, set 2 ..70 10 41 Low-noise design, set I

0 9

. 60 -CL

50-c

40 -'0 25 50 75 100 125 1500 1000 2000 3000 4000 5000 Transmission input torque, % of maximum

Frequency, Hz Figure 12I-Noise test results, sound power at spiral-bevel meshFigure 11 .-Ndise spectrum; test at 1 00-percent torque. frequencies (first two harmonics) versus transm'ss&c input torque.

30

CD 25 = a ei edei n e0 Baseline design, set 1

S20 ...---- : High-strength design, set 1.....2 High-strength, design, set 2

is 1 ..... .. , Low-noise design, set 1 -----50

coO

Spra-bvl ............................................................................................. ......

40:1Baseline design % 25 o5 75 100 125 I50- Low-noise design Transmission input torque, % of maximum

b) Accelerometer 1 , iptospia-el housing.F2ur 14-yia-irto etrsls

150 1000

100. 750-

so' 2o500.0 0

.250-50

-40o-•30-20o-1o 0 10 20 3040wniom ",aft pos;son, dog

a) Fiet gage 14.

I50 10 0 0_.... . .. . . . . . .. .. .

100 'SO .

25 W

0~ 0

-50".40-3020-10 0 10 20 30 40

Pirki sha-t position. deg

b) Root gmge 2.Figue 1 5.-Typical strain towt resuts s ra;b.beve! inio.

1 00-w,-nt to•rqu.

12

0 Sweline deskgr. static tests

* Baselkne desigr. dyn'arithc stesi

*High-strengtt. de$sigr. dvneam#C Iets,

1S 90 5Eiur S 900 -ý

100 0 -... 100 0 maotr

t o T3 0Q. .. -5 0 -~ 3 0 0 ~ T o-600

-6 0 6 1 18 24 0 3 V0 6 12 1S 24 30 36

position awoig toth gmijy pitat~n *ioflQ tootti, mmfT

al Spira~iavel pn~on fiet gages. b)f'+bvl 01Mt4W

150 900 ________ISO,___9005

1 00 - - 0,o 600 maximw"

j5-------- f50'3o

-50 300-om rsn 0Ccmrtsol

0 6 12 1S 24 30 36 46'0 6 12 18 2' 30 36

Position a"on tooth. mm Posirtion &Jong WONth "W"

C) Spiral-bevel gear fimet gages. d) Speatbeve; gear coot gages-

Figure IS.-Strain test results; I 00"pftcnt t~fqLw

13

REPORT DOCUMENTATION PAGE Fj p

Pu~bic reporting jturo *o in ctaitn ofE aiomaw_ vb tu 4.W'' I e~. ''W *owl'-.), 1416*

qatrwintg &W~ mkwa**aV* aw Odata ý*aq AfwO oomspsng *V fWI'4 I* W d0 ~~~ '"ýWt Vý LIýW O..f C ý' o., urw V 91

Owe~ Hipo", $&As 1204, Aito. VA --2202 43W, &O to~ To orkW O ý'1 Awrov "~ ti"4A k~.-v h*"A- 4A~W' Laý e' .'

1. AGENCY USE ONLY (Leave bWanA) 3.REPORT DATE OT T"PC AND DATES COVERED

4. TITLE AND SUBTITLE S. FUNODIG NUMERS

Low-Noise, High-Strength, Spira1.&BcJ, (.cwi ((-tot Hclicopter -1n~n

4.AUTHOR(S) 11.6221 IA4 7 ADavid G. Ltwicki, Robert F. Harndschuh, 7.jchar,, S ficnrý,and Faydor L. Litvmn

7. PERFORMING ORGANIZATION NAME(S) AMD ADORE SS(flifI PEftfORM*#G O*AIAMZATIONNASA L'ewis Rewatcah Center REPORT NMBJUER

Cleveland, Ohio 441 35-3191andVehicle Propulsion Directorate-US. Army Resea~rch Laboratory

Cleveland, Ohio 44135-3191

9- SPONSORINGMONfTORRIG AGENCY NAMES(S) A340 ADODRESS(ES) I.$OIOWMWT~n

Nation~al Aeronautics and Space AdminsurAtion AGE*"Y REPORT NUNKIK"

Washington, DýC 20546.-4X)Ol NAAT -Ii-and AIAA- 4N 21414U.Sý Army Research LaboratoryA~deiphi. Maryland 2~73l5ARI,- MR..71

11I. SUPPLEMENTARY NOTESPrepared for the 291h Joint Propul-4on Conference and Exhibit ciaortwionsd bv the AlAA, SAP., AVI41. ind AkStt 'ilrtci. ai',

June 28-30. 1993. David G, Ltwicki and Robert F HAndschuh, Vchztel Pwpoulsiort Dotctorvtr. VS Arrn, ltc~artPi laicwil~w%. Lr'.,Research Center- Zachary S. Henry, Bell Helicnopte Textron. Inc. Fort Worth. Trtiai 76101 Favi&or L Liii~. k. nnrvts of Wri & IChicago, Chicago,_Illinois_60680,_Rtspon-ible person,_DavidCt~.(1)4t3t G_________________

I Za. DISTRIBUTIOWAVALABABLIY STATIENEN I 7b. 0IS15 Ta4JToN Coot

Unclassified - UnlimitedSubject Category 37

13. ABSTRACT i(azhuui 200 worao)

Improvements in spiral-bevel gear design were investigated to support the ArmyN*ASA Advanced Rotorcral'iTransmission program. Program objectives were to reduce weight by 25 percent. reduce noise bv 10 dB,. and increaselife to 5000 hr mean-time-between -removal. To help meet these go~al. advanced-design spiral-bevel gears weretested in an OH-S8D helicopter transmission using the NASA 500-hp Helicopter Transmission Test Stand. Threedifferent gear designs tested included: (1) the current design of the OH{- 58D transmnission except gear material X-53instead of AISI 9310. (2) a higher-strength design the same as the current but with a full fillet radius to reduce geartooth bending stress (and thus, weight), and (3) a lower-noise design the same as the high-strength but with modifiedtooth geometry to reduce transmission error and noise. Noise, vibration, and tooth strain tests were performed andsignificant gear stress and noise reductions were achieved.

14. SUBJECT TERMS 15. NUJMBER OF PAGES

Transmissions (machine elements); Gears; Noise, Vibration', Strain measurement ______________

14& PRICE CODE

________________________________________________A03

17. SECURIT CLASSIFICATMO 18. SECURITY CLASSIFICATION it, SECURITY CLASSIFICATION 20. UMIBTATION OF ABSTRACTOF REPORT Of TRIS PAGE I Of ABSTRACT

Unclassified IUnclassified I Unclassified

NSN 7540"1-280-5500 Standard Form, 298 (Rev 2-89)Prescribed by ANSI Std Z39-182WiO12