EVALUATION OF A SURFACE ACOUSTIC WAVE
MOTOR OUTPUT FORCE
Makoto Chiba, Masakazu Takahashi, Minoru Kurosawa, and Toshiro Higuchi
Dept. of Precision Machinery Engineering, Graduate School of Engineering,
The University of Tokyo
7-3-1 Hongo, Bunkyo-ku, Tokyo 113, Japan
Phone;+81 3 3812 21 11 ext.6466, Fax;+81 3 5800 6968
ABSTRACT
Evaluation of optimum pressing force, namely pre-load
for a slider to obtain superior operation condition of a
surface acoustic wave(SAW) motor is described. As a first
experiment, we used steel balls for sliders. The steel balls
that we used were 0.5, 1 and 2mm diameter to change
contact conditions such as contact pressure, contact area
and deformation of the stator and the slider. As a result,
the deformation of the stator and the slider by the pre-load
should be half of the normal vibration amplitude. This
condition was independent to the diameter of the balls.
The maximum driving force was 0.95"
and the driving
force density was 50NImm2. Secondly, in order to enlarge
the contact area, multi contact points slider was made on
trial. As a result, the driving force became 25mN, 25
times as high as the first result, 0.95".
The SAW
motor has high potential to be a large output force, quick
response, long traveling distance and low in profile micro
linear actuator.
wave, a kind of surface acoustic wave is applicable to the
ultrasonic motor. For small size linear actuator, the
surface acoustic wave motor has a lot of merit such as
high output force, high speed (about lmlsec), long stroke
up to centimeter order, high energy density, easy holding,
high resolution positioning.
In this study, we intended to make'the output force of
SAW motor higher by introducing an optimum high
pressing force, namely pre-load for the slider and the wide
contact area.
EVALUATION OF THE OPTIMUM
PRE-LOAD
Experimental set up
To improve the friction drive condition, namely to take
out high output force, the higher pre-load condition was
investigated using the set up as shown in Fig.1. A stator
transducer for the experiments is same as previous
paper[ I]. A piezoelectric wafer was 128 degrees y-rotated
x-propagation LiNb03. Two lDTs (interdigital
INTRODUCTION
Surface acoustic wave (SAW) devices are widely utilized
in the field of communication instruments and so on.
These devices are operated at high energy density
condition to save materials and space. As well known,
the energy density of piezoelectric device is roughly in
proportion to the operating frequency. Therefore, usage of
surface acoustic wave devices for actuations at high
frequency operation condition produce good results.
Function
Generator
II
To realize a surface acoustic wave motor, friction drive at
high frequency such as IOMHz is a significant problem.
But the difficulty of the friction drive has been overcome
by control of the contact pressure between the slider and
the stator. We have demonstrated that the HF band (3 to
30 MHz) ultrasonic motor is available[ 11. The Rayleigh
0-7803-3744-1/97/$5.000 1997 IEEE
03
iezoelectric wafer
Fig. 1. Exper.imenta1 set up of a surface acoustic wave
motor with a steel ball slider.
250
force, the contract pressure distribution between the stator
(LiNb03 wafer) and the steel balls are calculated using the
Hertz contact theorem. At the center of the contact
region, the pressure become maximum. The maximum
contact pressures are shown in Fig.2. The contact
pressure was able to change up to 400MPa for 0.5mm
diameter ball and up to 600MPa for 2.0"
diameter ball.
In the previous paper, the contact pressure was around
100MPa[l].
Because of the elasticity of the materials, contact area of
the slider ball is finite. The slider and the stator deform as
illustrated in Fig.3. The contact radius is able to evaluate
with the Hertz contact theorem approximately. The actual
contact radius was several micro meter at most when the
steel ball sliders were put on the stator with attractive
force by the permanent magnet. For example, the contact
radius, the depression and the maximum contact pressure
were about 3pm, 9nm and 390MPa when the ball
diameter was lmm and the pre-load was 7.2".
It should
be pointed out that the deformation is almost same order
with the stator vibration deformation.
Spacer [mm]
Fig. 2. Maximum contact pressure of the ball sliders.
transducers) were arranged on 3 inches wafer. The IDT
pitch was 400pm and the electrodes strip width was
100pm. The electrode strip was 10 pairs for each
transducer. The driving frequency was 9.6MHz. The
normal direction vibration amplitude of the surface
particle was 16 nano meter at driving voltage of 150V.
The tangential direction vibration velocity was 0.95
m/sec.
Transient response
A slider of this motor was small steal ball. In order to
increase a pre-load, namely pressing force of the slider to
the stator, a neodymium permanent magnet was put
beneath the wafer. The pre-load was changed by spacer
thickness between the magnet and the wafer. Owing to
the effect of the magnetization of the steal ball, the slider
moved without rotation.
For optimum pre-load condition, the high speed and the
large driving force should compromise. In the case of the
light pre-load, the stationary speed is high but the driving
force is tiny. On the contrary, the high pre-load
condition, they are vise verse. At the excessive pre-load, a
slider never moves any more. The high speed and the
large driving force is trade off. Moderate pre-load
condition should be investigated to pull out the
performance in itself.
The attraction force between the permanent magnet and
steel balls of 0.5, 1.O and 2.0"
diameters was measured
with electric balance. The gap between the magnet and
the ball was changed from 1 mm (thickness of the wafer)
to 8.5"
using several spacers. From the attractive
The transient responses of the slider measured with a high
12
I
I
I
I
I
I
31 OMPa
0
5
10
15
20
Time [msec]
25
30
Fig.4. Transient displacement of lmm ball slider.
Fig. 3. Deformation of the slider and the stator.
25 1
I
j
UIU
0 ;
I
I
_____________~__.__..........
b..n .......i...............j .............
0
Time [msec]
Fig.5. Transient response of the lmm ball slider.
speed video camera and an image processing equipment
are shown in Fig.4. From the differential of the
displacement data, the transient speed curves were
obtained as indicated in Fig.5. The transient curves
indicated that the transition region was around 3msec. We
decided, therefore, to measure the displacement in 3msec
drive at various operation conditions. At the driving
voltage of 18OV, the displacement of the slider in 3msec
drive were measured. The driving performance has been
improved significantly comparison with the previous
result as shown in Fig.6. The sliders were 0.5, 1 and
2mm diameters. The maximum displacement conditions
depended on the diameter as well as the pressure. For the
2mm diameter ball, the optimum pre-load was about
270MPa. However, it was 350MPa for the lmm diameter
ball and seemed to be much higher for the 0.5"
diameter ball. The optimum pressure depended on the
5
10
15
20
Depression [nm]
25
Fig.7. Displacement of the slider with 3 msec drive
against a depression.
geometry of the sliders.
On the other hand, the depression of the sliders gave a
good result. Fig. 7 indicates the displacement in 3msec
drive as a function of the depression due to the pre-load.
The optimum depression of the lmm and 2mm ball
sliders were the same around 9nm. The normal direction
vibration displacement of the stator at 180V driving
voltage was 19nm. Hence the optimum depression of the
slider was almost half of the particle vibration
displacement of the surface wave.
To confirm the optimum depression condition at the
different vibration amplitude, we estimated in the same
way at the different driving voltage of 120V and 60V
using 0.5"
and lmm diameter balls. The result of
-
2000
0.5-180V
1-18OV
v 0.5-120V
1-120v
I
_
-E
1500
=L
Y
Y
3
8
1000
cd
U
*
.-5?
a
500
n
-
0
100 200 300 400- 500 600
Pressure [MPa]
0
5
10
15
20
Depression [nm]
25
Fig.8. Displacement of the sliders by 3 msec drive with
variation of voltage.
Fig.6. Comparison with different diameter of the slider
traveling displacement with 3msec driving time.
252
Fig.8 indicated that the half of the vibration ,displacement
was optimum. The optimum depression was about 6nm
for 120V and was about 3nm for 60V. They were also the
half of the vibration displacement.
It is concluded that to obtain large driving force and high
speed, namely quick response, the pre-load should be
adjusted that the depression is the half of the vibration
displacement.
?
I
0
0
Output force of the friction drive
MULTI CONTACT POINTS SLIDER
Slider design and evaluation
Owing to the large thrust density such as SON/"*,
it
seemed to be effective to widen the contact area between
the slider and the stator. In order to obtain the large
contact area, putting a flat contact surface slider on the
flat surface substrate is quite common. From a
tribological point of view, however, it is almost
impossible to make contact the flat surface of the slider
uniformly to the stator. Therefore, the multi contact
points slider was made on trial.
The multi contact points slider was made of an aluminum
plate and several hundreds steel balls glued on the plate as
illustrated in Fig.11. The aluminum plate size was
6 ~ 6 m m 2and the diameter of the steel balls was 0.2".
As mentioned above, to obtain the high output force, the
pre-load should be adjusted that the depression is half of
the vibration displacement. When the driving voltage is
1
e
>
0.8
contact pressure for the same deformation. If a slider size
is 6 ~ 6 m m 2and the true contact area is 1/1000, the
resultant driving force will become 1N.
The output force of the several driving conditions are
plotted in Fig.9. as a function of the depression ratio to
the normal vibration displacement. At the depression
ratio of around 0.5, the output force was maximum. But
the value of the force depended on the ball diameter
because the contact area and the pressure were different.
The output force divided by the contact area is shown in
Fig.10. In the case of the 0.5"
diameter steel ball, the
thrust density was about SON/";?.
This was larger than
the lmm diameter ball because of the difference of the
E
0.2
0.4
0.6
Depressioddisplacement
.I
I
Fig.10. Output force density of the ball sliders as a
function of the depression/displacementratio.
To estimate the output force of the surface acoustic wave
motor, available thrust and optimum condition of
operation were investigated. For the quantitative
evaluation, the dynamic response should be measured.
For this purpose, the transient motion of the sliders were
measured using the high speed video camera and the
image processor unit. This method was effective to
measure the dynamic response of the motor, but it was
not so efficient. It took several hours for the image
processing so that the measured condition were limited.
e,
0
.
0.6
0.4
.e
z 0.2
n
01
0
.
.
.
i
i
.
I
..
i
.
..
i
.
..
i
,
..
i
..
.
i
0.4
0.6
DepressiodAmpli tude
0.2
I
0.8
0.2"
diameter steel balls
Fig. 11. The multi contact points slider.
Fig.9. Output force of the SAW motor when changing
the diameter and the driving voltage.
253
165V, the vibration displacement of the SAW is 16nm,
so that the deformation of the slider and the substrate
should be 8nm.
Assuming that the plate size and the depression was fixed
in 6 ~ 6 m m 2and 8nm respectively, the optimum pre-load
and the contact area were calculated as a function of the
diameter of the steel balls. Fig. 12 illustrates the relation
between the optimum pre-load for the slider and the
diameter of the steel balls on condition
that the steel balls are glued maximum amount on the
6 ~ 6 m m 2aluminum plate. Fig. 13 illustrates the relation
between the contact area and the diameter. From Figs.12
and 13, as the diameter of the steel balls increases, the
optimum pre-load and the contact area decrease. For the
20
15
70
10
E
85
.e
higher pre-load operation, the diameter of the steel balls
should be small.
A diameter of steel balls glued on the aluminum plate
was 0.2"
because they were obtained easily. The
maximum amount of the steel balls was calculated to be
about 1000. The contact area of the slider was also
calculated to be 5.3~10-3mm2.This contact area was
about 200 times as large as that of the former experiment
using the 1.0"
diameter steel ball slider. The number
of the steel balls glued on the trial slider was about 500.
About the slider made on trial, in case that the driving
voltage was 165V, optimum pre-load was calculated to be
2.7"
for each steel ball, and 1.4N for the slider, using
the Hertz contact theorem. In this case, the contact radius
of each steel ball was 1.3pm. The whole contact area of
the slider was 2.5xl0-3mm2, namely about 1/100000 of
the aluminum plate area. The maximum contact pressure
was 800MPa at the center of a contact point. The value
of the maximum contact pressure was about 2 times as
high as that of the former experiment. From the pre-load
of 1.4N and the friction coefficient of 0.2, estimated
driving force became about 0.3N. If the driving force was
0.3N, the thrust density would increase up to 110N/mm2.
Experimental output force
0
0.5
1
1.5
2
Dimeter of the sell balls [mm]
Fig. 12. The calculated optimum pre-load versus the
diameter of the steel balls.
The experiment was carried out on condition that the
driving voltage was 165V and the driving frequency was
9.6MHz. To control the pre-load, a weight was put on
the slider as shown in Fig. 14. The pre-load was changed
from 7mN to 143".
0
0.0:
-*g
I
I
The driving force increased in proportion to the pre-load
as shown in Fig.15. The maximum driving force was
about 25".
This maximum driving force was 25 times
as high as that of the former result of 0.95".
However
0.0;
Y
m
Multi contact points slider
8
\
c-)
Weight
0
m
2
6
0.01
C
,-
0.5
1
I .5
Diameter of the steel balls [mm]
Fig.13. The contact area between the slider and the stator
versus the diameter of the steel balls.
Fig. 14. Experimental set up using the multi contact
points slider.
254
30
Piezoelectric substrate
25
E
Y
20
IDT
Spring
\
5
Linear guide
\
Multi contact points slider
/
0
0
50
100
Pre-load [mN]
150
Fig. 16. Practical design of the SAW linear motor.
An example of the practical design of a SAW linear
motor is shown in Fig.16. The dimensions of the stator
are several cm in length, one to two cm in width and
several mm in height. The slider has multi contact points
driving surface and pre-load mechanism with a spring.
Much smaller one will be possible.
Fig.15. Out put force as a function of the pre-load.
it was only 10 percent of the estimated maximum driving
force of 0.3N. From the value of the friction coefficient
was 0.2, the driving force would have become 0.28N
when pre-load was the optimum value of 1.4N.
However, the slider did not move under the condition of
the optimum pre-load 1.4N. It seemed that all of the 500
steel balls did not contact to the substrate. Hence the
concentration of the pre-load in a part of the steel balls
prevented the slider from moving. As a result, the contact
area was less than that of calculated value, so that the
thrust was reduced.
ACKNOWLEDGMENT
This work was supported by the Grant-in-aid for general
scientific research of the Ministry of Education, Science,
Sports and Culture.
REFERENCES
It was concluded that the multi contact points slider was
effective to obtain the large driving force. Since the thrust
increased owing to the large contact area and the large preload. Now the slider is being improved to obtain much
larger thrust.
[l]
CONCLUSION
We have succeeded in improvement for friction drive
conditions of the surface acoustic wave motor. The
important point was the deformation ratio of the slider
and the substrate against the normal vibration
displacement and the contact area between the slider and
the substrate. Taking the deformation and the contact area
into consideration, the multi contact points slider was
made on trial. It was confirmed that the slider was
effective to obtain the higher driving force than ever. The
maximum driving force became 25".
It was 25 times
as high as that of the elemental experiment result of
0.95".
255
M.Kurosawa, M.Takahashi and T.Higuchi,
"Ultrasonic linear motor using surface acoustic
waves," IEEE Transactions on Ultrasonics,
Ferroelectrics, and Frequency control, vol. 43,
no.5, pp901-906, 1996.