2005-01-2302

Identification and Reduction of Gear Whine Noise of the Axle System in a Passenger Van
Sang-Kwon Lee, Sung-Kyu Go, Dongjun Yu, Joo-Young Lee and Sung-Jong Kim
Department of Mechanical Engineering, Inha University

Yun-Kyung Jo and Byung-Jae Choi

NVH Research, DYMOS Co.
Copyright 2005 SAE International

ABSTRACT
This paper presents practical work on the reduction of gear whine noise. In order to identify the source of the gear whine noise, transfer paths are searched and analyzed by operational deflection shape analysis and experimental modal analysis. It was found that gear whine noise has an air-borne noise path instead of structure-borne noise path. The main sources of airborne noise were the two global modes caused by the resonance of an axle system. These modes created a vibro-acoustic noise problem. Vibro-acoustic noise can be reduced by controlling the vibration of the noise source. The vibration of noise source is controlled by the modification of structure to avoid the resonance or to reduce the excitation force. In the study, the excitation force of the axle system is attenuated by changing the tooth profile of the hypoid gear. The modification of the tooth profile yields a reduction of transmission error, which is correlated to the gear whine noise. Finally, whine noise is reduced up to 10dBA.

structure-borne noise is not significantly important for whine noise. As the next process, operational shape analysis and experimental modal analysis are used to identify the source of airborne noise due to the structural behavior of the axle system during acceleration of the passenger van. It is found that the major reason for airborne noise is the vibro-acoustic noise of the axle system. This vibro-acoustic noise is controlled by reducing the excitation force due to gear meshing. The excitation force is reduced by modifying the tooth profile, which is related to the transmission error. Transmission error is very much correlated with gear whine noise [913]. In the paper, the transmission error of the ring gear has been reduced by modification of the tooth profile of the ring gear under test. Finally, whine noise in the compartment of the passenger van is reduced to 10dBA. The sound quality due to whine noise is also improved.

MEASUREMENT OF INTERIOR NOISE

An axle system with a gear ratio of 3.616 is developed. The number of teeth of a pinion gear is 13 and the number of teeth of a ring gear is 47 in an axle system. A passenger van loaded with this axle system is accelerated in order to have a subjective evaluation of gear whine noise. During acceleration, the gear whine noise in the compartment of the passenger van is especially heard at 60km/h and 120km/h. To identify the source of this whine noise objectively, first, the interior noise is measured at the front seat. Fig. 1 shows the waterfall map for the noise signal. On this waterfall map, two peaks are clearly identified at 1800 rpm and 2300rpm as shown in Fig. 2. These two peaks occur at the frequency band, which is the meshing frequency of the hypoid gear of the axle system. These peaks well correspond with the previous subjective evaluation. Meshing frequency of the hypoid gear is 13th times the rotation speed of driveshaft since the number of teeth of a pinion gear is 13. In Fig. 2, it infers that gear whine noise is related to the harmonics of the meshing frequency of hypoid gear in an axle system.

INTRODUCTION
In todays highly competitive automotive industry, the vehicle sound quality becomes more and more important [1-6]. There is a constant pressure to achieve a low vehicle interior noise level in system design, especially for a passenger van. For an axle system, gear whine is one of the major factors in achieving the vehicle interior sound quality. This paper presents the results of the practical research to reduce the whine noise of an axle system of a passenger van. The problem occurs in the compartment of a passenger van. In general, gear whine noise has two paths: airborne and structure-borne [7-8]. In order to identify the transfer path of whine noise, the accelerations on all transfer paths from the axle to the car body are measured together with interior noise of the passenger van simultaneously. From vibration tests, it is found that the isolator between the body and axle system can reduce enough the vibration transfer energy from the axle system to the car body. Therefore, the

13 order

Frequency (Hz)
Figure 1 Waterfall map for the interior noise data

IDENTIFICATION OF TRANSFER PATH

Gear whine noise originates from the excitation force by the meshing of a hypoid gear as shown in Fig. 3. There are two paths for whine noise: structure-borne noise path and airborne path. The transfer of the vibration energy from the axle system to car body induces structure-borne noise. The direct transfer of vibroacoustic sound due to the shell vibration of an axle system from axle system to compartment of a car induces air-borne noise. STRUCTURE-BORNE PATH The structural vibration of an axle system is measured at several points as shown in Fig. 4. Three accelerations are measured on the housing of axle and eight accelerations are measured on the bracket of isolation system during acceleration from 1800 rpm to 3500rpm of driveshaft speed. Fig. 5 shows the photo information about the attachment of three accelerometers attached on the axle housing.

13 order

Speed of driveshaft (rpm)

Figure 2 Order analysis for interior noise

Excitation due to Tooth Meshing

Structural Vibration Of Axle Housing Transfer Path Resonance

ACC 3

ACC 1

ACC 2
Isolation of Mount Body vibration Attenuation of noise by body
Figure 5 Photo explanation for measurement point of the structure vibration of an axle system

Inertial noise in the compartment of a passenger van

Figure 3 Transfer paths of air-borne noise and structure-borne for gear whine noise in an axle system; : Structure-borne path, : air-borne path

Figure 4 Measurement point of the structure vibration of an axle system

Figs. 6, 7 and 8 show the waterfall map for the vibration data measured at the point 1, 2, and 3 of axle system respectively. From these results, there is one structure resonance around 3100rpm. Resonance frequency is about 670Hz (3100rpm x 13 / 60). The other resonance of house structure occurs at 2600 rpm. Resonance frequency is about 1100Hz (2600rpm x 26/ 60). Two peaks caused by resonance are clear in Fig. 6 and Fig.7 , but in Fig.8 only one peak is clear caused by 13th order of driveshaft speed. Points 1 and 2 are the places on the hypoid gear housing. Pont 3 is the place of the axle system. Therefore, at 670Hz, the resonance caused by 13th order of driveshaft speed is global mode of axle system. At 1100Hz, the resonance caused by 26th order of driveshaft speed is the local mode of axle housing itself. This local mode does not contribute to the interior since there is no gear whine noise at 1100Hz as shown in Fig. 1. The global mode excites the whole axle system. Therefore, the axle system has high vibration level at all other four points (point 4,6,8 and 10 in Fig. 4) before the isolator as shown in Fig. 4. It is confirmed with vibration result measured at point 4 as shown in Fig. 9. The isolator, which is installed between the axle system and the car body, attenuates this horrible vibration due to global mode of the axle system. Fig. 10 shows the waterfall map for the vibration data measured at point 5. This point is the place after the isolator rubber; there is no resonance peak. Although there are a few resonances in the axle system, all vibration energy is dissipated through damping of the rubber isolator, which is installed between the axle system and the car body. Therefore, the gear whine noise as shown in Fig.1 is not caused by structure-borne noise.

13 order

26 order

Frequency (Hz)
Figure 6 Waterfall map for the vibration data measured at the point 1 of axle housing.

26 order

Frequency (Hz)
Figure 7 Waterfall map for the vibration data measured at the point 2 of axle housing.

13 order

Frequency (Hz) Figure 8 Waterfall map for the vibration data measured at the point 3 of axle housing.

13 order 26 order

Frequency (Hz) Figure 9 Waterfall map for the vibration data measured at the point 4 before isolation rubber

Frequency (Hz) Figure 10 Waterfall map for the vibration data measured at the point 5 after isolation rubber.

Finally, Fig.11 shows the set-up of the measurement equipment for noise and vibration testing for a passenger van.

MODAL ANALYSIS AND OPERATIONAL SHAPE ANALYSIS Although the two resonances existing in the axle system do not affect the gear whine noise inside of the car throughout structural vibration-transmission, they can affect the gear whine noise as the air-borne sound. In order to identify the structure mode, experimental modal analysis and operational shape analysis are used. Operational shape analysis (OSA) can be obtained by using phase difference between reference point X and moving point Y. The governing equation is given by [14],

i = Yi
where

X* = B e j( b a ) X

(1)

X = A e ja Y = B e ja

(2)

Figure 11 Set-up of measurement equipment for noise and vibration test for a passenger van

Fig. 12 shows the measurement point for operational shape analysis and experimental modal analysis. The number of points is 125. At these 125 points, vibration is measured during acceleration from 1800 rpm to 3200rpm of driveshaft. The phase difference for each point as shown in Fig. 12 is calculated by using Eq. (1).

resonance frequency obtained operational shape analysis and that obtained experimental modal analysis.

Figure 12 Measurement point for operational shape analysis and experimental modal analysis.

(a)

Operation shapes are obtained and plotted in Fig. 13. There are three major operational shapes. Two are global modes and the other one is a local mode. One global mode is a tensional mode as shown in Fig. 13(a), the other a global mode is the bending mode as shown in Fig. 13(b). The local mode is the bulge mode of the axle housing as shown in Fig. 3(c). The tensional mode of the axle system occurred at 1800 rpm with 13th order of drive shaft and its frequency is 433Hz. It affects gear whine noise and the air-borne noise. The bending mode of the axle system occurred at 3100 rpm with 13th order of drive shaft and its frequency is 670Hz. It affects gear whine noise and the air-borne noise. The bulge mode of the axle housing occurred at 2600 rpm with 26th order of drive shaft and its frequency is 1100Hz. It does not affect the gear whine noise as much as the two global modes as shown In Fig.1, since it is a local mode and vibration energy is not enough to radiate the sound. Major three modes appearing in the operational shape analysis in Fig. 13 are confirmed with experimental modal analysis. Experimental set-up for modal analysis is arranged as shown in Fig. 14. The number of excitation points is the same point as points as shown in Fig. 12. In order to identify the major modes of the axle system, the sum of transfer functions measured at the 125 points as shown in Fig. 12 is calculated and plotted in Fig. 15. In this figure, two peaks at low frequency around 115Hz and 300Hz are found and one peak is identified around 900Hz. The first two peaks are global modes: tensional mode and bending mode. The last one is the bulge mode of the axle housing, which is a local mode as shown in Fig. 16. The natural frequency of each mode is shifted down to low frequency compared with the frequency obtained from the operation shape analysis as shown in Fig. 13, because the boundary condition of the structures between operational shape analysis and experimental modal analysis is different. For operational shape analysis, the axle system is installed on the car. Therefore the structure is constrained by the suspension system of a real car. For experimental modal analysis, the axle system under test is installed freely on the frame as shown in Fig. 14. Therefore, there is a frequency difference between the

(b)

(c)
Figure 13 Operational deflection shapes during acceleration from 1800 rpm to 3200rpm of driveshaft

(a)

Figure 15 Sum of transfer functions measured at the 125 points

(b)

Figure 16 Mode shape for 3 mode (bulge of axle housing) of the axle system.

rd

DESIGN MODIFICATION OF AXLE

In the previous section it was found that structure-borne noise does not contribute to gear whine noise because all vibrational energy caused by the resonance of the axle system is damped out through the isolator between the body and the axle system. Therefore, it is inferred that part of the vibration energy caused by resonance of the axle system is radiated as vibro-acoustic energy. This vibro-acoustic energy is air-borne noise and a major source for gear whine noise in the compartment of a passenger van. There are three methods for reducing this air-borne noise. The first is a design modification of the car body with high acoustic transmission loss required. The second is a structure modification of the axle system to avoid resonance. The last one is to attenuate the excitation force by modifying the profile of the teeth in the hypoid gear. For economical efficiency,

(c)
Figure 14 Photo illustration for experimental modal analysis of axle system

in the paper, the last method is adapted for the reduction of whine noise. MODIFICATION OF TOOTH PROFILE According to previous research results [8-13] and the company experience, transmission error is correlated to the gear whine noise. In the paper, the tooth profile is modified from a flat form to bent form listed in Table .1.

TRANSMISSION ERROR Gleason 500Ht is used for the measurement of transmission error of two-ring gears. The total transmission error is reduced from 0.104 to 0.056 as listed in Table.2. According to Table 1, the initial contact area of two gears has the same area. However, the contact area of the ring gear with high transmission error goes out of the boundary of tooth face under full-load condition while the contact area of the ring gear with low transmission error is inside of the boundary of the tooth face. Fig. 17 shows a photo of two ring gears. The upper one shows the ring gear with high transmission error; the bottom one shows the ring gear with low transmission error.

Table 1 The profile of tooth in the ring gear

Cutter Dia. Original 9 Inch Tooth Profile Slow Steep Contact Ares

Modification 7.5 Inch

In order to modify the tooth profile, the original ring gear is processed by using a 9-inch cutter, but the modified ring gear is processed by using a 7.5- inch cutter. In the contact area, the black area is the initial contact area of teeth. The outline is the contact area of teeth under full load condition. The design details of the ring gear under test are listed in Table.2.

Table 2 The design details of the ring gear

Cutting by 9 Inch Cutter Tooth Profile Cutting Method Cutter Diameter Module Number of Teeth Face Width Shaft Angle Pinion Offset Spiral Angle Whole Depth Addendum Dedendum Face Angle Root Angle Pitch Angle Gear Ratio Mating Gear Backlash Transmission Error Helix Form Gleason 9 inch 5.268 47 31.75 90 38.10 26 49 9.83 1.13 8.70 76 12 74 9 79 21 4.89R (44X9) 3009 4580 0.15~0.20 0.104

Cutting by 7,5 Inch Cutter Helix Form Gleason 7.5 inch 4.907 47 31.75 90 38.10 22 5 10.57 1.78 8.79 76 12 69 53 75 4.89R (44X9) 3009 4580 0.13~0.20 0.056

Figure 17 Photo of two ring gears of axle system

GEAR WHINE NOISE REDUCTION The modified axle system with new profile tooth is replaced of noisy axle system and installed on the passenger van under test. Interior noise for a passenger van is measured and its results are plotted in Figs. 18 and 19. Vibration on the axle housing at point 2 is also measured. Fig. 20 shows results of the vibration test. From these results, high vibration around 1800rpm and 3200rpm of driveshaft speed, as shown in Fig. 7, is removed since the excitation force due to tooth meshing is significantly attenuated by reducing transmission error although structure resonance exists. Therefore, Interior noise also reduced up to 10dBA at the same speed of driveshaft speed as shown in Fig. 18 and 19 comparing with Fig. 1 and2. Around 2600rpm of driveshaft speed, the 26th order component of vibration is still high level since it is a local mode of axle housing and it does not significantly contribute to the gear whine noise.

Frequency (Hz) Figure 18 Waterfall map for interior noise of a passenger van with modified axle system

Gear whine noise occurs around 1800rpm and 3200 rpm, at which vibration level is also high. 2. Experimental modal analysis and operational deflection shape are used to identify the vibration transfer path. It is found that two global modes contribute to the resonance of an axle system around 1800rpm and 3200rpm. The former is the torsion mode of axle system around 433Hz; the latter is bending mode of axle system around 603Hz.Vibrations caused by these modes are attenuated through an isolator at the mount point of the axle system. Therefore, it is inferred that the structure-borne noise is not contributing to gear whine noise. However these modes contribute to the airborne noise for gear whine noise. In order to control the air-borne noise, the structure should be modified to avoid resonance or the excitation force should be reduced. For gear whine noise, the modification of structure mode requires very high stiffness in structure. In the paper, instead of modification of the axle system, the excitation force is attenuated by reducing the transmission error by optimizing the profile of teeth of the axle gear.

3.

Figure 19 Order analysis for interior noise of a passenger van with modified axle system

CONCLUSION
1. Basic noise and vibration tests for a passenger van are assessed for the identification of problems.

4.

By reinstalling the new axle system with optimized tooth profile on the same passenger van with gear

whine noise problem, interior noise is reduced up to 10dBA.

Frequency (Hz)

Figure 20 Waterfall map for structure vibration of a axle housing measured on the positing 2 in modified axle system

ACKNOWLEDGMENTS
This work was supported by BK21 Project in Korea. This work was also partly supported by DYMOS Company in Korea. 5. 6. 7. 1. Lee, S. K., Yeo, S. D., and Choi, B. U., Identification of the Relation Between Crankshaft Bending and Interior Noise of A/T Vehicle in Idle State, SAE930618, 1993 2. Lee, S. K. Weight Reduction and Noise Refinement of Hyundai 1.5 Liter Powertrain, SAE940995, 1995 3. Lee, S. K. Vibrational Power Flow and Its Application to a Passenger Car for Identification of Vibration Transmission Path, Traverse City, Michigan, USA. SAE Noise & Vibration Conference & Exposition, SAE 2001-01-1451, 2001, 2001 4. Lee S. K., Chae, H. C, Park, D. C and Jung, S. G. "Booming Index Development for Sound Quality 8. 9.

REFERENCES

10.

Evaluation of a Passenger Car," SAE 2003-01-1497, 2003 Becker S. B. and Yu. S Objective Noise Rating of Gear Whine SAE1999-01-1720 Becker S. B. and Yu. S Gear Noise Rating Prediction Based on Objective Measurement SAE1999-01-1721 Sun Z., Voight M and Steyer G. Driveshaft Design Guidelines for Optimized Axle Gear Mesh NVH Perfprmance, FISITA2004, F20004V287 Sun Z., Steyer G. Meinhardt G. and Ranek R. NVH Robustness Design of Axle System SAE2003-01-1492 M.G. Donley, T.C. Lim and G.C. Steyer Dynamic Analysis of Automotive Gearing Systems, Journal of Passenger Cars: Mechanical Systems, 101(6), pp. 958968, 1992. Glover R. and Rauen D. Gear Transmission Error for Use with Ger Inspection Machine, SAE2003-011663

11. Tarutanl I. and Maki H. A New Tooth Flank Form to Reduce Transmission Error of Helical Gear, SAE2000-01-1153 12. Athavale S., Krishnaswami R. and Kuo E. Estimation of statistical Distribution of Composite Manufactured Transmission Error, A precursor to Gear Whine for A Helical Planetary Gear System, SAE2001-01-1507 13. Houser R. and Harianto J. Manufacturing Robustness Analysis of Noise Excitation and Design of Alternative Gear Sets, SDAE2001-01-1417 14. Wyckaert, K and Van der Auweraer, H., "Opernational Analysis, Transfer path Analysis, Modal analysis: Tools to Understand Road Noise Problems in Cars," SAE Paper 951752.

CONTACT
Dr. Sang-Kwon Lee joined Hyundai Motor Research Center in Mabook-Ri, Korea, working with the Automotive Noise and Vibration Control Group from 1985 to 1994. Dr. Sang-Kwon Lee (B.E., M.E., and Ph.D.) received the Ph.D. degree from the Institute of Sound and Vibration Research (ISVR) in the University of Southampton, Southampton, England in 1998. He has been an Associate Professor in the Department of Mechanical Engineering, Inha University, Inchon, Korea, since March 1999. His research interests are in the application of digital signal processing to the automotive noise and vibration control e-mail:sangkwon@inha.ac.kr.