Re: G320 Prototypes

Jon,I agree. It should have a soft decay if it is a classic
integrator clamp. However what I did was to force the integrator
(actually the summing amp for P and D co-efficients as well as being
an integrator) to the zero speed voltage of the PWM (2.5V in this
case).

The integator is uncaged at the end of reset. Simultaneously the
power devices begin switching and the counters are free to update.
Since it is at or very near what will be the steady state level, no
transient behavior is noticable.

To test this I super-glued a mirror sliver to the test motor shaft
and bounced a HeNe beam off of it. The total before/after power-up
excursion was 0.2 deg, and thats counting the 0.1 deg dither angle.
So yeah, I'd say it works okay now.

Mariss

> Yup, this was exactly like the trouble I had with my servo amps. I

knew

> this would
> be a problem, and designed in a circuit to accomplish this. The

trick

> is letting
> this circuit unclamp the final servo loop SLOWLY, so that as the amp
> gain just
> increases to where the motor breaks static friction, the system can
> stabilize
> before the gain comes all the way up. It is going from open loop to
> closed loop operation, and has to be done in a controlled manner.
>
> Jon