Re: [CAD_CAM_EDM_DRO] Re: G320s, New and Improved!

mariss92705@... wrote:

> Jon,
>
> It dithers in the absence of a tach simply because there is no
> feedback from the encoder except on the edges. A PID servo has no
> other option because of the integrator term.

Of course. between steps, when moving slowly, or not at all, it
goes open loop - a discontinuity. The only fix is to make the
encoder resolution high enough so that the error is too small to
be a problem, or find a way to keep the loop closed all the time.

> My solution is to knock the proportional and differential co-
> efficients (gain and damping) down to 5% of set value when the servo
> is within +/- 2 counts of the command position. The interal co-
> efficient is dominant within this range, so this is very effective.

A deadband. yet another discontinuity! Maybe this isn't so bad,
as long as it rarely gets into the deadband during continuous motion.
If it passes through the deadband during continuous motion (maybe
do to a small discontinuity in the mechanical system) that could cause
an amplification of the little bump into a big bump.

> Most inexpensive motors are not equipped with tachs, so it tells you
> which way technology is trending these days. Though not perfect, a
> motor and encoder sans tach gets you within 99.9% of perfect with a
> modern drive.

Yes. I just added tachs to motors that lacked them. Maxon motors
make great tachs, and there are a lot of them floating around.

Jon