Journal of Mechanical Science and Technology 25 (2) (211) 273~277 www.springerlink.com/content/1738-494x DOI 1.17/s1226-1-1216-4 Effect of the number of poles on the acoustic noise from BLDC motors Kwang-Suk Kim 1, Chang-Min Lee 1, Gun-Yong Hwang 2 and Sang-Moon Hwang 1,* 1 School of Mechanical Engineering, Pusan National University, Busan 69-735, Korea 2 Department of Mechanical Engineering, Youngsan University, Yangsan-city, Kyungnam-do 626-847, Korea (Manuscript Received September 22, 29; Revised August 14, 21; Accepted November 13, 21) ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Abstract The goal of this study is to examine the effect of the number of poles on the noise from BLDC motors. To this end, the number of slots was fixed to six and the number of the poles was set to four or eight before noise was measured. Motors having different numbers of poles showed clear differences in noise; cogging torque, torque ripple and normal local force were interpreted, analyzed and compared to determine the reason for the differences. To conduct precise comparisons of noise, efforts were made to prevent changes in other variables when constructing the four-pole and eight-pole motors. Noise measurements were conducted after going through the above preparation. The results show that the eight-pole motor produced a lower noise level than the four-pole motor. To determine the cause of the noise, the cogging torque, torque ripple and radial local force were calculated, which are representative noise sources in BLDC motors, and FFT was performed to analyze their frequency components (harmonics). The calculations were made using a finite element method with the Maxwell stress tensor. Keywords: Number of poles; BLDC motors; Acoustic noise; Cogging torque; Torque ripple; Local force ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- 1. Introduction The motors used in automobiles must be improved in order to keep up with the trend that continuously requires higher quality and durability in cars. More comfortable driving environments should be provided to meet the need for highergrade cars, and the automobile parts should be improved and replaced to satisfy their required life span so that the demand for higher durability can be met. The DC motors currently used in cars fall short of resolving these problems and meeting the requirements. Of these problems, the problem of higher durability can be solved by using BLDC (brushless direct current) motors. BLDC motors electrically adopt a noncontact rectification method, and they can be used semipermanently as long as there are no problems in the other parts. However the issues related to the achievement of higher quality still remain unresolved. Of these issues, the problem of the acoustic noise is more serious than anything else. This problem must be addressed more actively because it is unpleasant for the driver. The problem of durability can be resolved by adopting BLDC motors but various approaches are needed to address This paper was recommended for publication in revised form by Editor Yeon June Kang * Corresponding author. Tel.: +82 51 51 324, Fax.: +82 51 582 314 E-mail address: shwang@pusan.ac.kr KSME & Springer 211 the issue of acoustic noise and to ensure further advanced grades. To this end, diverse methods have been suggested and demonstrated, and such methods are frequently applied to actual designs and manufacturing. Previous research focused on reduction of torque ripple to reduce noise of BLDC motors by current control or reduction of cogging torque by alterations of teeth shape of the slots [1, 2]. This paper focused on the number of poles in BLDC motors. This is because the basic driving cycles vary as the number of poles changes in the motor. This result in changes in the electromagnetic vibration forces which cause acoustic noise. In particular, the vibration forces in the motor are all cyclic functions, and the ultimate purpose is to analyze those changes [3-5]. To investigate these effects, two specimens were made with the same number of slots but different numbers of poles. To see the effects for the number of poles, the number of slots is not important. In this paper selected six slots motor most used in blower of the automobile and four and eight poles motor. After making the specimens, noise measurements were conducted. According to the test results, the two motors showed clear differences in the acoustic noise level. To determine the cause of the differences, the electromagnetic vibration forces were interpreted and analyzed. The electromagnetic vibration forces interpreted included the cogging torque, torque ripple and radial local force, and they were calculated using the finite element method. In addition, frequency analysis was carried
274 K.-S. Kim et al. / Journal of Mechanical Science and Technology 25 (2) (211) 273~277 out by performing FFT. 2. Preparation of the specimen Some preparations were needed for an exact comparison of the two specimens with different numbers of poles. As mentioned above, as the number of poles varies in the motors, the cycles of the driving mechanisms also change, and thus, the motor driving circuits and winding methods can show considerable deviations. As long as these deviations exist, the noise comparison may be meaningless. Accordingly, this study included three types of preparatory work to prevent these deviations. First, this study prevented deviations in the magnetic circuit. When the number of poles changes, there can be a change in the magnetic circuit of the motor, and consequently, the degree of magnetic saturation appearing locally can vary. Therefore, in this study, the specimens were designed and manufactured in such a way that the magnet volumes of the two motors were equal. This is illustrated in Fig. 1. Second, balancing work was conducted to eliminate the unbalanced magnetic forces that occur when the rotor rotates. Unless the unbalanced magnetic forces of the two motors are equal above a certain level, the effects are included in the noise measurements, and they might then influence the determination of the effects of the pole number. This is why the balancing work was conducted. In the case of motors, double correction balancing is frequently used. According to the Korean Industrial Standard (KS B 612), the allowance is.1 g or less. Table 1 shows the final results of balancing. As shown in Fig. 2, Rear-L indicates the left-hand correction, and Front-R shows the right-hand correction. As confirmed in Table 1, the balancing corrections of the two rotors are similar to each other, and therefore, it appears that there is no problem with the noise comparison. The third preparation, the most important part, was to minimize the deviations caused by the output size. It is meaningless to compare the noises of motors that have different output sizes. Naturally, a motor with a larger capacity will produce a higher level of noise. Fig. 3 shows the TS curves of the two specimens in terms of the measured values. It can be confirmed that the outputs of the two motors are identical. Table 1. Real unbalanced weight (after balancing process). Number of pole 3. Comparison of the acoustic noises Noise measurements were conducted after going through the above preparation process. As Fig. 4 shows, a loader was installed to measure the motor noise at the loading time, and the loader and the test-piece motor were coupled together. A rate voltage of 13.5 V DC and a load of.6 Nm were set before taking the noise measurement. The distance between the specimen and the microphone was.1 m. The test duration was 2 minutes. The cumulative noise level in the audible frequency range of 2 to 2 khz was measured. As Table 2 shows, the noise of the eight-pole motor showed an improvement of as much as 4.8 db, as compared with that of the fourpole motor. 4. Analysis of the noise difference Real Unbalanced Weight (g) Rear L Front -R 4 Poles.89.98 8 Poles.87.95 4 Pole 8 Pole Volume= 2634 mm Fig. 1. The volume of the magnets. Rear - L Fig. 2. Balancing schematic. Rotating Speed [RPM] 5 4 3 2 1 6 Slot 3 Volume=26352 mm Front - R 1 2 3 4 5 6 7 8 9 Torque [kgf.cm] Fig. 3. TS curve of the four- and eight-pole motors (experimental results). The above results indicate that the eight-pole motor showed 3
K.-S. Kim et al. / Journal of Mechanical Science and Technology 25 (2) (211) 273~277 275 Table 2. Results of the acoustic noise test. The Number of pole Total Noise 4 Poles 74.8 db 8 Poles 7. db SPL [dba] 8 7 6 5 4 3 2 4 Pole 8 Pole 1 2 2531.5 4 563 8 1125 16 225 315 45 63 81k 1.25k 1.6k2k Frequency [HZ] 2.5k 3.15k 4k5k 6.3k 8k1k 12.5k 16k Fig. 5. 1/3 octave band. Fig. 4. Schematics of the acoustic noise test. an improvement of as much as 4.8 db as compared with the four-pole motor. To analyze the difference more closely, the 1/3 octave and narrow band are examined in Fig. 5 and Fig. 6. As shown in Fig. 5, the dominant noise differences appear near 6 Hz and 2 khz. More precise differences are shown in Fig. 6. It can be confirmed that the 12 th, 32 th, 36 th, 44 th, and 48 th higher order noise components are the major components of the differences. These noise results can be attributed to the electromagnetic exciting forces in the motors, and each higher order component may be caused by the electromagnetic exciting force components. Accordingly, it is necessary to determine the causes of the differences by analyzing and comparing the electromagnetic exciting forces of the two motors. To determine the cause of the noise, cogging torque, torque ripple and radial local force were calculated, which are representative noise sources in BLDC motors, and FFT was performed to analyze their frequency components (harmonics). The calculations were made using a finite element method with the Maxwell stress tensor [6, 7]. The cogging torque is caused by the interaction of the permanent magnet and the tooth blade depending on the rotor position even when there are no input currents and it can be obtained as shown in Eq. (1): T cog R = BRBθ ds µ (1) where R represents the radius of the air gap, µ indicates the permeability of air, and B R and B θ show the flux density in the radial and tangential directions in the air gap, respectively. When the torque occurring during the motor operation is not constant but pulsates within certain minimum and maximum values, the size of the pulsation is called the torque ripple. The torque ripple can be expressed as the difference between the minimum and maximum values of the waveform following the calculation of the torque. Eq. (2) shows how to calculate the torque: SPL [dba] 8 7 12X 6 32X36X 44X48X 5 4 3 2 1 4 Pole 8 Pole 1k 2k 3k 4k Fig. 6. Narrow band. Cogging Torque ( N.m ).15.1.5. -.5 -.1 Frequency [Hz] -.15 15 3 45 6 Rotor Position (Mdeg.) Fig. 7. The waveform of the cogging torque. Torque ( N.m ) 1..8.6.4.2. 15 3 45 6 Rotor Position (Mdeg.) Fig. 8. The waveform of the torque (ripple).
276 K.-S. Kim et al. / Journal of Mechanical Science and Technology 25 (2) (211) 273~277 Cogging Torque ( N.m ).2.15.1.5. 12 24 36 48 Fig. 9. The waveform of the normal local force. Fig. 1. The FFT of the cogging torque. e i T =. (2) ω Here, e and I represent the back-emf and currents of each phase, respectively. ω is the rotating speed. That is, the size of the pulsation can be determined after the torque is calculated according to Eq. (2). Normal local force refers to the magnetic traction force acting on the surface of the stator tooth of BLDC motors. This directly increases the acoustic noise when the motor rotates. This can be expressed as in Eq. (3); Torque ( N.m ).2.15.1.5. 12 24 36 48 F 1 2 2 = ( BN BT )ds 2µ. (3) N - B N represents the normal flux density, and B T indicates the tangential flux density of the tooth surface, respectively. According to the results of the analysis of each vibration force shown in Figs. 7-9, the cogging torque shows differences in both size and cycle. In the case of the eight-pole motor, the size is much larger, but the cycle is two times shorter than that of the four-pole motor. Therefore, the number of additional harmonics is smaller. In the case of torque ripple, the ripple of the eight-pole motor is definitely smaller than that of the four-pole motor, and the radial local force also shows the same tendency. Accordingly, the difference in the noise caused by the change in the number of poles can be ascribed to the torque ripple and the radial local force among the vibration forces tested. For a clearer comparison, FFT was conducted on these vibration forces, as shown in Figs. 1-12. In the case of the cogging torque, it was found that the eight-pole motor had a greater amplitude, but the period was reduced by half, when compared with four-pole motor. Therefore the 12 th and 36 th higher order components did not appear. This means that the noise measurement results involve the 12 th and 36 th components, and this can actually be confirmed in the noise measurement results of Fig. 5 and Fig. 6. In the case of the torque ripple, the eight-pole motor has a remarkably small Fig. 11. The FFT of the torque ripple. Normal Force ( N ).2.15.1.5. 4 8 12 16 2 24 28 32 36 4 44 48 Fig. 12. The FFT of the normal local force. torque ripple and this can be found in terms of periodicity. It can be found that the overall higher order component amplitude is smaller than that of four-pole motor, and the 12 th and 36 th components do not appear due to periodic gains. In the case of the normal force, it can be confirmed that the differences in the 32 nd, 36 th, 44 th and 48 th components were especially notable, as shown in Fig. 12. It can also be found that, in addition to the periodic gains, the fact that the 32 nd, 36 th, 44 th and 48 th components of the eight-pole motor are much smaller than those of the four-pole motor also affects the noise results given in Fig. 5 and Fig. 6.
K.-S. Kim et al. / Journal of Mechanical Science and Technology 25 (2) (211) 273~277 277 5. Conclusion This study focused on the number of poles in order to improve the noise characteristics of BLDC motors. The final conclusion of the study is that, as the number of poles increases, the noise characteristics improved. To determine the cause, the cogging torque, the torque ripple and the local force acting vertically on the stator tooth blade were analyzed and compared. When the number of poles increases, there is a change in the periodicity of the electromagnetic exciting forces of the motors, and it was found that a greater number of poles lead to higher gains in terms of the number of exciting forces. The study also found that the periods were reduced by half in all of the exciting forces examined in this study, and the number of exciting forces could be decreased by half within the frequency range of interest. This means that there is a greater possibility to avoiding structural resonance of the motors. Accordingly, as the number of poles increases, the periodic gains of the electromagnetic exciting forces increase, and the probability of avoiding the structural resonance becomes higher, ultimately facilitating the noise reduction further. In terms of the amplitude of the exciting forces, the cogging torque did not show significant gains, but in the case of the torque ripple, a greater number of poles could generate a more stable output torque. In the case of the normal force, the differences in the higher order components were clear, and these differences may be reflected in the noise measurement results. Accordingly, when considering the structural size and the economic rationality allowed in the design and manufacture of BLDC motors, it would be better in terms of the noise characteristics to adopt as many poles as possible. Acknowledgment This research was financially supported by the Ministry of Education, Science Technology (MEST) and Korea Institute for Advancement of Technology (KIAT) through the Human Resource Training Project for Regional Innovation. References [1] J. Y. Hung and Z. Ding, Design of currents to reduce torque ripple in brushless permanent magnet motors, Electric Power Applications, IEE Proceedings B, 14 (4) (1993). [2] S. M. Hwang, J. B. Eom, G. B. Hwang, W. B. Jeong and Y. H. Jung, Cogging torque and acoustic noise reduction in permanent magnet motors by teeth pairing, IEEE Trans. Magnetics, 36 (5) (2) 3144-3146. [3] M. N. Anwar and Iqbal Husain, Radial force calculation and acoustic noise prediction in switched reluctance machines, IEEE Trans. Ind. Applications, 36 (6) (2) 1589-1597. [4] S. P. Verma and A. Balan, Determination of radial-forces in relation to noise and vibration problems of squirrel-cage induction motors, IEEE Trans. Energy Conv., 9 (2) (1994) 44-412. [5] C. Y. Wu and C. Pollock, Analysis and reduction of acoustic noise and vibration in the switched reluctance drive, IEEE Trans. Ind. Applications, 31 (1) (1995) 91-98. [6] A. N. Wignall and A. J. Gilbert, Calculation of forces in magnetized ferrous cores using Maxwell Stress Method, IEEE Trans. Magnetics, 24 (1) (1988) 459-462. [7] S. M. Hwang, K. T. Kim, W. B. Jeong, Y. H. Jung and B. S. Kang, Comparison of vibration sources between symmetric and asymmetric HDD spindle motors with rotor eccentricity, IEEE Trans. Ind. Applications, 37 (6) (21) 1727-1731. Kim Kwang-Seok received his B.S. degree in Mechanical Engineering from the Pusan National University, KOREA, in 2, and his Ph.D. in 21. Currently he is a team-manager of EM- Tech, a Korea microspeaker company. His primary research interest is newconcept structure of speakers, receivers and vibrators. Hwang Sang-Moon received his B.S. degree in Mechanical Engineering from the University of California at Berkeley, USA, in 199, and his Ph.D. in 1994. Currently he is a professor of Mechanical Engineering at Pusan National University, Korea. He is also CTO of Em- Tech, a Korean microspeaker company. His primary research interest is new magnetic circuits in microspeakers.