Issue
153 April 2003
Muscle
for High-Torque Robotics
ANALOG
SWITCH
The
heart of this little circuit is the MC74VHC1G66 analog
switch. After receiving the pulse from Timer1, the analog
switch transfers the voltage on C3 (i.e., the PW-V converter)
to C8. Recall that Timer1’s pulse occurs at the falling edge of the input
pulse; thus the voltage on C3 will be ramped up to a
peak corresponding to the width of the input pulse.
Recall
also that Timer2, which generates its pulse after the
completion of the sample and hold activation pulse,
discharges C3. Photo 4 shows the output of the sample
and hold circuit versus the input pulse. (Note that
the input pulse repeats in 15-ms intervals. The pulse
repetition interval isn’t critical and doesn’t contain
any information. It varies from 10 to 30 ms, depending
on the brand of RC transmitter.) The voltage stored,
or held, on C8 is buffered by another MC33078 op-amp—the
other half of this dual op-amp chip—and applied to the
position reference input of the MC33030.
|
(Click
here to enlarge)
|
Photo
4—The radio control servo control pulse has an ~15-ms
repetition interval. Note that it’s the width of
the pulse (0.5 to 1.5 ms) that’s critical, not the
pulse repetition interval. |
TEST-PULSE
GENERATOR
This
servo test-pulse generator is simply a test-pulse generator
that’s added to the board so that it can be tested without
an RC transmitter and receiver. Jumper JP2 is provided
so that the PW-V converter can receive a pulse from
an external receiver or the on-board test-pulse generator.
WATCH
YOUR FINGERS
I
used a simple, single-reduction motor and gear assembly
for testing the circuitry. A worm gear was mounted to
the shaft of a 5-A motor, and the position feedback
potentiometer was mounted to the mating worm wheel.
The potentiometer provided 320° of rotation.
Testing
showed that the system responded to position commands
with the high speed associated with a single-reduction
gear and the high torque associated with the 5-A motor.
In fact, the speed and torque were of a magnitude that
merited physical precaution for my fingers!
Several
alternatives would be practical for the gearing and
potentiometer assembly. For instance, if a multi-turn
potentiometer (e.g., 10 turn) were used in place of
the single-turn device, a winch servo would be created
without sacrificing positional accuracy. Likewise, a
pure linear-force mechanical arrangement could be accomplished
via an ACME lead screw and follower nut, which would
be attached to a linear potentiometer.
With
deference to my MCU-focused colleagues, I acknowledge
that this entire exercise could have been done with
one HC08 microcontroller, one MC33887DH, and about 2
KB of code. But wasn’t sticking strictly to a hardware
solution more fun?
