Ultrasonic Motor by Measuring Transient Responses

Transcription

1 IEEE TRANSACTIONS ON ULTRASONICS. FERROELECTRICS, AND FREQUENCY CONTROL. VOL. 38, NO. 5, SEPTEMBER An Estimation of Load Characteristics of an Ultrasonic Motor by Measuring Transient Responses Kentaro Nakamura, Minoru Kurosawa, Hisayuki Kurebayashi, and Sadayuki Ueha Abstract-To measure the characteristics of ultrasonic mo- tests cause the temperature of the motor to rise. Theretors, such as the maximum torque, torque-speed relationship fore, the resonance frequency, at which the motor should and the frictional coefficient at the contact surface, a method be driven, drifts and the characteristics vary. It is necesin which the torque is calculated from the transient responses is proposed. The rise curve that is the transitional change in sary to maintain a fixed temperature or to tune the driving the rotor speed soon after turning on the motor gives the load frequency automatically to characterize the motor accucharacteristics, while the fall curve that is the decay of the ro- rately. To carry out the study of ultrasonic motor effector speed after turning off the motor yields the frictional coef- tively, a simpler measurement method that can be perficient of the contact surface. This method requires only a short formed in a short time is needed. time (the transient time of the motor) to complete the measurement. After analyzing the relations between the transient re- This paper presents a method to estimate the load charsponses, the load characteristics and the frictional force, this acteristics (torque-speed curve) of the ultrasonic motor inmethod is applied to a hybrid transducer type rotary motor and stantly by measuring its step responses. One can obtain a traveling wave type linear motor. The load characteristics of the load characteristics without a torque meter, and ignore the rotary motor is given by the present method and various the temperature effects since only a short time is needed kinds of frictional materials for the linear motor are examined. to complete the measurement. This method gives the torque, the speed of revolution and the efficiency of the I. INTRODUCTION motor, as well as the frictional coefficient of the contact N ULTRASONIC MOTOR is based on the concept surfaces of the rotor and the stator. A of driving a rotor or a slider, through frictional forces, In the first part of this paper, the relation between the by a high-frequency elastic vibration excited on the stator step responses and the load characteristics is analyzed. A via the piezoelectric effect. Several basic ideas, such as a system for the measurement and two examples are then traveling wave type [l], [2,] a vibration conversion type described. One is the load characteristics of a hybrid [3]-[5] and a hybrid transducer type (61, [7], have been transducer-type rotary motor 2 mm in diameter; the reproposed and constructed for trial during the last decade. sults with the present method and that with a traditional Few of them, however, has been put to practical use as way are compared. The other example shows the perforyet, for all the excellent potential properties: high torque mance of various frictional materials such as the maxiat low speed, and quick responses. It can be said that the mum speed and the maximum traction force by using a research on the ultrasonic motors is at the developmental traveling wave-type linear motor. stage, and large number of measurements are required to improve the design. For example, one of the difficulties consists in the choice of the contact interface between the rotor and the stator. If it is overcome to some extent. ultrasonic motors will be used widely. Various kinds of ma- terials must be examined experimentally to find the most suitable one, since no theoretical guide line has been found yet. In the laboratory we often measure the load torque of the motor by making use of a torque meter or by simply pulling up a weight. A great deal of time is required to collect data for a load characteristics curve. Also, these Manuscript received April 19, 199; revised November 1, 199; accepted March 2, The authors are with the Research Laboratory of Precision Machinery and Electronics, Tokyo Institute of Technology, 4259 Nagatsuta, Midoriku, Yokohama 227 Japan. IEEE Log Number PRINCIPLE OF THE METHOD A. Typical Step Responses of Ultrasonic Motors Fig. 1 shows the step responses of a hybrid transducer type rotary motor 2 mm in diameter, where J denotes the moment of inertia of the load including the rotor. This describes a typical transient response of ultrasonic motors; the speed of revolution rises as a first-order lag system (its step response has a form of [ 1 - exp (-at)]) and falls linearly. It takes a certain time for the vibration of the stator that drives the rotor to build up. In the case of the hybrid transducer type, however, the time for building up the vibration is much smaller than that for rising up the speed of the rotor, because the rotor is pressed to the stator with a large force to obtain a sufficient frictional force and the vibration is heavily damped. Thus approx /91/9-481$ IEEE

2 482 IEEE TRANSACTIONS ON ULTRASONICS. FERROELECTRICS. AND FREQUENCY CONTROL. VOL 38. NO 5. SEPTEMBER 1991 I I Time (ms ) Fig. I. Typical step responses of an ultrasonic motor J indicates the moment of inertia of the load including the rotor. I) J = 17 g. cm', 2) J = 67 g. cm2, 3) J = 86 g. cm'. imately only the equation of motion of the load dq T= J- dt governs the transient curves. Here, T, Q, and t are the torque that drives the load, the angular velocity of the rotor and the time, respectively. In the case of other type motors whose transient time of vibration can not be ignored, the moment of inertia of the load should be chosen large enough. The rise time and the fall time are plotted against the moment of inertia in Fig. 2, for a hybrid transducer type rotary motor 2 mm in diameter. The transient times are almost proportional to the moment of inertia of the load, and these curves converge near the origin. This shows that our assumption is adequate. B. Relation Between the Step Responses and the Load Characteristics In most cases, the load characteristics (torque-speed curve) of the ultrasonic motor, as illustrated in Fig. 3, show that the speed of revolution decreases linearly as the torque increases. This can be written as Q = Q,(l - T/T,) (2) where Q, and T, indicate the speed of revolution without load and the maximum torque, respectively. Eliminating T from (1) and (2), we have a differential equation: Q - Q,, = -J Q d - (Q - QJ. To dt The solution of (3) gives the rise curve, Q = Q(, 11 - exp (-t/~,)], (4) where the rise time r, is given by rr = JnCJ/C,' (5) We can calculate the maximum torque T, if we know the rise time 7,. In order to draw the load characteristics curve, we have to calculate dq/dt at every point on the rise curve and substitute in (1). We measure the step response of the input electric power to the motor P,, simultaneously to determine the efficiency curve as a function of the load. Here, the efficiency q is calculated as c C U. 5 1 Moment of fnertio 5 (g.crn2) Fig. 2. Rise time 7, and fall time 7, against the moment of inertia of the rotor and the load. Fig. 3. Typical load characteristics of an ultrasonic motor. C. Estimation of Frictional Coejicients Between the Rotor and the Stator After the vibrator is turned off. the retarding frictional torque T,, which acts between the rotor and the stator, brakes the rotation:?j = pfr (7) where p, F, and r represent the frictional coefficient, the vertical static force that presses the rotor to the stator and the radius of the contacting surface, respectively. Then, the fall characteristic is readily calculated as Q = Q,(l - t /Tf) (8) where fall time rfis written in the form: rf = JQ,/?j. (9) Consequently. by measuring the fall time, we can estimate the frictional coefficient of the materials used for the rotor and the stator MEASURMENT SYSTEM Fig. 4 illustrates a schematic diagram of the measurement system. In general, an ultrasonic motor needs a twophase driver system: an oscillator, a phase shifter and two power amplifiers. Two power meters are equipped to measure the input electric power to the motor. We employed an analog multiplier for a power meter to measure the effective power. We note here that the transient response of the power meter should be sufficiently shorter than that of the motor. For the measurement of the revolution speed of the rotor, a high resolution rotary encoder (Canon M-l: 5, pulses per revolution) was used, and a high speed f/v converter was employed to convert the pulses to the analog voltage. The necessary resolution of the encoder, the speed of the f/v converter and the power meters should be determined in view of the transient time of the motor. Consequently, one should choose load a with

3 NAKAMUR.4 f'f d.: ESTIMATION OF LOAD CHARACTERISTICS 483 %. Gate Phase Shifter Amp. J Shaft Power Meter 1 AID Converter II Fig. 4. Block diagram of the measuring sy\tem sufficient moment of inertia. The total value of the moment of inertia including the rotor, the encoder and the couplers or joints are calculated or measured before the measurement. The data are taken by a personal computer with an AID converter. 2 Fig. 5. Configuration of a hybrid transducer type rotary motor. l / IV. MEASUREMENTS A. Load Characteristics of a Hybrid Transducer Type Rotary Motor In this section, the load characteristics of a hybrid transducer type rotary motor 2 mm in diameter is measured by the present method as an example. Fig. 5 shows the configuration of the hybrid transducer type ultrasonic motor. The hybrid transducer that acts as a stator of the motor consists of a torsional vibrator and a multilayered piezoelectric actuator. The former drives the rotor to rotate, while the latter vibrates in the axial direction to control the frictional force that acts between the rotor and the stator. The rotor is pressed to the stator by a coil spring to obtain sufficient friction [6]. First, the maximum torque was measured as a function of the voltage applied to the torsional vibrator, where the voltage applied to the actuator and the static force that presses the rotor was constant. As is shown in Fig. 6, the maximum torque measured by the proposed method almost agrees with the one obtained by pulling up weights. The maximum error in this case was about IO%, and it is thought to be caused by the wow and flutter in the revolution. While the conventional method shows the average torque, this method yields the torque obtained at a certain position of the rotor, because the rotor rotates only about IO" to finish the measurement by this method. The prototype motor used in the measurement had about 1% wow due to the irregularity at the contact surface or the error in position of the shaft. Then, the load characteristics curve as well as the eficiency curve were drawn from the rise curve of the speed of revolution and the input power. The transient curves and the resulting load characteristics are shown in Figs. 7 and S, respectively. The speed of revolution and the input power rose gradually. The time delay rb indicated in Fig. 7 corresponds to the transient time of the vibration. In the case of a hybrid-type ultrasonic motor, when the speed of revolution is reduced with the load, the impedance of the torsional vibrator becomes large and,the '5L I 5 IC 15 Applied Voltage (V rms) Fig. 6. Maximum torque as a function of the voltage applied to the torsional vibrator: -*-, measured by the present method; -X-, by pulling up weights. Time - Fig. 7. An example of the measured transient curves of the speed of revolution and the input electric power. input power is also reduced. Consequently, we can attribute the change in the input power after 78 to the transient of the mechanical revolution of the rotor, as the vibration itself had already built up. Next, the efficiency which is the ratio of the output mechanical power to the input electric power (defined by (6)) is calculated. These results showed good agreement with results obtained by pulling up weights. B. Properties of Frictional Materials The performance of several kinds of frictional materials have been examined. Here we employed a traveling wave type linear motor [g] whose set up is illustrated in Fig. 9. Two Langevin type transducers are attached to excite the traveling wave of the flexural vibration on the bar. One is connected to a power source and drives the bar, while the other absorbs the wave motion without reflection because of an electrical matching circuit and an electrical termination. The principle of this motor is summarized as fol-

4 484 5, I1 2 BEAPEE FL-EP i,, ' D,l Q D :'.<a525 Torque (kgfcm) Fig. 8. Measured load characteristics: solid line by pulling up wieght: dotted line by the present method., ' m. m Direction of wave propagation n I n, / meter Fig. 9. Set up of a traveling wave U Termination R tlpe linear motor lows. When an elastic traveling wave propagates, a particle at the surface of the bar vibrates elliptically [9]. In consequence, a slider that is put to the bar is driven to move. The frictional coefficients, the maximum speed of the slider, the maximum force and the maximum efficiency for several kinds of rubber and other plastics were measured by the method described above. The results are summarized in Table I, where an aluminum bar or a stainless steel bar was used, and the force pressing the slider to the bar was 1 N. The maximum speed and the maximum force for the aluminum bar in Table I are plotted in V. CONCLUSION We have proposed a method to estimate the characteristics of the ultrasonic motor by measuring the transient L 1 I I I I I I I I I I I.1. I5.2 Frictional Coefficient (b) Fig. 1. (a) Maximum speed and maximum force for an aluminum bar as a function of Young's modulus. The force that presses the slider to the bar is 1 N. The. shows the maximum speed, while indicates the maximum force. (b) Maximum speed and maximum force for a stainless steel bar as a function of frictional coefficient. The force that presses the slider to the bar is 1 N. The shows the maximum speed. while indicates the maxlmum force. responses. The load characteristics curve, the efficiency curve and the frictional coefficient at the contact surface can be readily measured. This method was applied to a hybrid transducer type Fig. lo(a) as a function of the Young's modulus. It can ultrasonic motor, and the maximum torque and the load be said that the materials with larger Young's modulus characteristics were measured. The proposed method can show higher maximum speed and greater maximum force, be accomplished in a time as short as the transient time with thexception of BEAREE FL-EP (teflon) and of the motor. Since the rotor revolves less than about 1" RIC716. The relationship for the stainless steel bar is because of the quick response of the ultrasonic motor, the very complicated as seen from Table I. torque obtained by this method is the value for a certain Fig. 1(b) summarizes the relations between the max- position of the rotor, and any irregularity along the path imum speed, the maximum force and the frictional coef- of revolution is directly reflected in the results. This may ficient for the stainless steel bar. The larger frictional coef- be an advantage for some purposes, but to ascertain the ficient results in the higher speed and the stronger force. average value it is necessary to repeat the measurement Rubber B showed prominently large force. In the case of several times. Since the motor is not operated continuthe aluminum bar, no such simple rule existed. ously in this system, we can avoid effects arising from temperature changes. The properties of many kinds of frictional materials for a traveling wave type linear motor were measured and examined. Though, in the most cases, the relations between

5 NAKAMURA et U/.: ESTIMATION OF LOAD CHARACTERISTICS 485 TABLE I MEASURED PROPERTIES OF FRICTIOVAL MATERIALS Maximum Young's Maximum Frictional Speed Density Modulus Efficiency (X IO8 N/m') (kg/m') Coefficient (cm/s) Force (N) (%) Rubber A' Rubber B* Polyethylene Hi-moler EX1 3' Polyester P9B' P15B' Polyimid K4525' K55 18' RIC5323h RIC534 I h RIC716h Polypropylene SPX MU4K' Polyamide TEIJINCONEX' Teflon BEAREE FL-EPh Carbon fiber I1 I l ". 18gh I l63.15 I IO I 12. I I 8. I l. I I I "Measured values on the top are for an aluminum bar. hmeasured values on the bottom are for a stainless steel 'JlSB345: asbest joint sheet packing. normal type. djis-b345: asbest joint sheet packing, for oil sealing. 'Shiotani-Kasei, Co.. Ltd. bar. 'Toyoba Co.. Ltd. 'Nippon Polymide C. Ltd. hntn Rulon Corp. 'Mitsubishi Petrochemical Co.. Ltd. JTeijin Ltd. ACKNOWLEDGMENT The authors would like to thank the companies who of the ay 1991 supplied us the frictional materials. Authors also wish to thank Mr. Kazuhito Nishita for his experimental contribution. REFERENCES Kentaro Nakamura, for a photograph and biography, please see page 193 issue of this TR.~~S,~C-II~I~S. Minoru Kurosawa, for a photograph and biography, please see page 92 of the March 1991 issue of this TRAYSACTIONS. Hisayuki Kurehayashi was horn in Shizuoka, Ja- [I] Sashida, T. Japanese , no. Pat. Feb. 25, March on pan received He B.Eng. the 121 M. Kurosawa, Nakamura, K. T. Okamoto, and S. Ueha, "An ultra- degree in electrical and electronic engineering and sonic motor using bending vibrations of a short cylinder." leee Trans. the h1.eng. degree from the Tokyo Institute of Ultruson. Ferroelec. Freq. Conrr.. vol. UFFC-36. no. 5, pp Technology. Tokyo. Japan. in 1985 and re- 521, Sept pectively. 131 T. Sashida, "Trial construction and operation of an ultrasonic vibra- $,,, Slnce 1987 he has joined Chubu Electric Power tion driven motor," Oyo Bursuri, vol. 51, no. 6. pp g -Qb *-G*!A Company. Nagoya. Japan. (in Japanese). [4] A. Kumada, "A piezoelectric ultrasonic motor," In froc. 8rh Meeting Ferroelrc , Kobe, Japan, pp [5] J. S. Schoenwald. P. M. Beckman, R. A. Rattner, B. Vanderlip. and B. E. Shi, "Exploiting solid state ultrasonic motors for robotics," Sadayuki Ueha, for a photograph and biography, please sec page 92 ot Proc. leee 1988 Ulrru.son. Symp., vol. I, pp, the March 1991 issue of this TR4hSACTlOhS.