On Wed, 12 Aug 1998 10:35:44 -0400, Frank E Gennari
>Excerpts from netnews.comp.robotics.misc: 12-Aug-98 Re: Regarding
>>The servo that was almost centered would have already finished moving
>>before the other finished receiving it's pulse. Wouldn't it?
>>Is there such a long delay from the time the pulse is received until
>>the servo actually moves to make that 1 mS difference negligible?
>>How long does it take from the time a pulse is received (trailing
>>edge) until the servo actually 1) begins to move and 2) finishes
>>moving approx 5 degrees.
> Servos update their positions about every 20ms, so the first servo
>probably would not have started moving when the second one recieved its
>pulse. It probably takes at least 50ms to move to its final position.
>You won't even notice the 1ms difference. 1ms is only 1/1000 of a
>second, which is a very short time for mechanical rotation. I made a
>Stamp I control 3 servos simultanneously and it works fine. Why would
>you need that accurate timing?
Servos update their positions at a rate depending on when the pulse is
repeated. I assume when a pulse is first sent, that the servo begins
moving almost immediately.
I agree that a slew rate of 60 degrees in .22 seconds that the first
servo would still be moving when the second one started. But, my
point is that there might be several mili-seconds when the two servos
are bucking each other. Remember that these are controlled by Basic
instructions (not exactly the fastest); they might be controlled by an
SSC chip which takes mili-seconds just for the commands to be passed.
On an RC airplane this might not be a problem since we are talking
about air between the control surfaces. But what about a mechanical
link that depends on two links moving simultaneously with each other.
This results in increased power drain, motor wear, etc.
Seems to me that it should be possible to design a servo controller
that can have servos move simultaneously with respect to each other.
Obviously there would probably still be some overlap but it seems that
it could be minimized.
1) Load up PWM pulse lengths
2) Start them all such that they all finish (trailing edge of pulse)
at the same time.
3) This would allow all to begin movement, with some known lag, at the
Are there any spec sheets on the Futaba servos? I have looked and
cannot find any specifications on their control and operation.