Re: [CAD_CAM_EDM_DRO] Re: Re: Re: GeckoDrives.. Can anyone Help?

Tom Eldredge wrote:

> Gentlemen,
>
> Could we step/direction guys bridge this gap with Marris' help? Let me make
> a suggestion:
>
> Could we not make a hardware (IC based) adapter which included an enocder
> driven counter, a hardware comparator, and another hardware counter driven
> by our step /direction signals, and then a D/A converter driven by the
> hardware comparator,

As far as I know, this is exactly what the Gecko Drive G320 and G340 do.

> which would then drive one of Marris' upcoming +_
> servo amps? Would we not be thus able to provide the P, and he provide the
> I, D, and the actual amplifier to drive the servo motor?
>
> I have already emailed Marris with this idea. I suppose I should have put
> the whole thing on this forum rather than bother him, but after I emailed
> him, I thought others might have some good ideas related to this.
>
> It does not seem as though the hardware for the P of the PID would be much
> to come up with, and maybe it would provide another compromise helping us to
> get closer to the true velocity control desired in some applications.

I can't see why you would not use the G320 or G340 for your application.

> I probably do not see the big picture, but I thought I'd ask.

OK, I guess the big picture is that as long as step pulses are coming out of
the computer, then the motion is in the form of discrete positions. If you make
the resolution of these discrete positions extremely small, then the quantization
should be small enough for any particular application, but then the computer
has to output step pulses at an incredible rate to get a decent rapid feed rate.
The advantage to a true velocity servo, straight from the computer, is that
the VELOCITY is not quantized by the resolution. So, at very slow speeds,
the motion remains smooth. My velocity servo remains smooth down to .01
IPM, at which point the stick slip friction takes over. This is at a rate of
3.3 encoder counts/second on my system (20,000 counts/inch). With a
pseudo-servo system, step pulses would come out at a rate of 3.3 per second,
and the machine would most likely move in discrete steps.

This becomes most important in things like cutting very slight tapers on
a lathe, or cutting angles that are very nearly parallel to a machine axis on
the milling machine. It also shows up in interpolating circles, especially of
larger diameter. 4 times in the full circle, one axis slows down to zero speed
and then reverses, moving very slowly at first.

Jon