Re: [CAD_CAM_EDM_DRO] Re: +/- 10V Gecko

mariss92705@... wrote:

> Jon,
>
> At this point I'm looking a three trimpots; that's all the cans are
> punched out to accomodate.
>
> One trimpot has got to be a current limit. This is a pulse by pulse,
> real time (sub 1uS) hard current limit set.
>
> That leaves two trimpots.
>
> The natural here is scale and offset. The offset trimpot will be
> a "zero" adjust for the input signal. The 3rd pot will be a "gain"
> adjust for the input signal.
>
> My approach to design philosophy is to do the very most with the very
> least. Beauty and perfection lies with sparseness. Like poetry, the
> fewest words are perfect and sublime.

We're a little out of sync on these messages, but I have an important
thing to say. As far as I'm concerned, this is NOT a velocity servo
amp, in fact it is not even a servo amp. It is just a plain voltage amp.
I have tried this, and I can tell you with absolute certainty, that it will
perform extremely poorly in a CNC motion control system. If you
turn up the DC gain enough to get low error when standing still,
it will vibrate badly when moving. If you turn the gain down to
eliminate the instability, it will have HUGE following errors when
moving. The IR compensation will help, but only a little. Since Mariss
says there won't be room enough for a pot to adjust it, I don't know how
you will set the correct IR compensation, anyway. And, the IR comp
needs to be set for each motor type.

I still say that you can't do a velocity servo without velocity feedback.
That either takes a very high resolution encoder or a DC tach.
I also still say that a true velocity servo is a very desirable thing!

Jon