Re: Re: Close loop steppers

beer@... wrote:

> I don't think a "proper" stepper system should lose steps. If it does,
> it's either because it's not powerful enough or the feed rate is too
> high.
>

Agreed. with 3000 in-oz for x, 1500 for y and z I have good torque
with the motors I just got.
Inertia match is good, about 3/1 except for z. Feed rate is very
high though- about 10x that of a typical mill. Cutting forces are
very low, only a few pounds.

Slipped steps are more a concern of bad programming, running
into something, or damaged/worn cutting tools. And I know that
will happen.

>
> When a step pulse came in, a step pulse was immediately issued. The
> box then watched the DRO to see if movement resulted. If there was no
> movement ( as indicated by the DRO ), the CNC software was paused and
> additional steps were issued by the little box until suitable movement
> DID occur. The pause line was released and normal operation continued.
>
> This requires that the CNC software have a "pause" input, as opposed to
> an EStop input. ( Doug Yeager of CNCPro did add such a feature, BTW.)
>

I downloaded CNCPro last night. I would like to talk to Doug about
its capabilities. And whether he sells the source code. I guess
he is doing everything with one parallel port so there may be
no place for encoder inputs. One could use the port interrupt from
external counter data to do his pause perhaps. Is his software
good stuff?

Anyway I just see steppers as low speed brushless motors with some
torque cogging. Enough microstepping and you just have the makings
of a nice smooth servo. Mine are just half step though so
the general case of filtered encoder feedback is way too slow.
But the things You and Arne are talking about is non-linear
processing and could provide some protection and be stable
even though they are not suitable for all motions.(in my case)
That is why I am considering step recovery in non-cutting
parts of the cycle only.

Les


--
Leslie M. Watts
ITW Advanced Technology Group
(847)657-4559
http://www.netcom.com/~leswatts/leswp.html