Re: Re:Stepper program w/Encoders?

Tim Goldstein wrote:

> From: "Tim Goldstein" <timg@...>
>
> Thanks Peter,
>
> The reason I was asking is Fred Proctor is has been looking into it for EMC.
> If I remember what he said in his e-mail correctly (I am out of town and
> don't have his previous e-mail on this system), He is concerned that if he
> implements it in the same manner as servo feedback the steppers will
> continually hunt between step. He does have an idea of how to get around it,
> but is not sure if it would really be suitable. He was inquiring if I knew
> of any programs that implement closed loop on steppers.

I've heard of systems that do this. What you do is put a non-linear
function on the positioning error which creates a deadband. A deadband
of just one step might be enough, but there could be some positions where
a combination of step accuracy and encoder 'flicker' could make more
than one step of deadband necessary. A parameter in the .ini file for
this would probably be the best way. I really think that much of the
servo code could be used, where the servo's velocity command out
would be replaced by step rate. Actually, a deadband might not be such
a bad idea for the servo system, too. One would have to be careful
not to create too much nonlinearity, of course, or it would destablilize
the servo loops.

Jon