Re: Low cost Servo Controller

Bertho

I spent years keeping up a bunch of Hardinge lathes that were controlled by
AB 73xx controllers. They used 15 bits from the digital to drive their D/A
board. That's 1600+ steps plus direction if mscalc works the way I think it
should. They would do an excellent job on shallow tapers where one axis
runs very slow compared to the other and on cutting a radius or half circle.

To me that's the critical question. What are the very shallow tapers going
to look like? What are the four corners (where axis motion slows to a stop
and then reverses) of a milled circle going to look like? 128 steps are
more than adequate for the high end of axis motion. Will they be enough
for the low end of feedrates.

The stg card uses 13 bit for both D/A and A/D. (8000+ levels)


Ray

-----om-----

From: Bertho Boman <boman@...>

>Subject: Re: Low cost Servo Controller
>
>I do not know anything about the AMC amps but is there a chance that the

controller actually sends a velocity command as a

>digital signal. One byte would give sign + 128 speed levels.
>
>Actually, it makes sense to communicate digitally to the servo amp and

then put the D-A internally in the servo amp. Less noise

>and less headaches in general. I have been thinking of using the parallel

port to send a byte per channel for the servo amp. I

>have not checked if 128 levels are enough but I think it is.
>
>Although, sticking to the +/- 10 VDC standard is very desirable as I

mentioned earlier.

>
>One advantage with the digital link is that the servo amp can be either

analog all the way or with PWM or it can be DSP based

>and all digital and read the tach for velocity feedback. The encoders

wires can continue and be connected to the PC as usual,

>the DSP would just monitor the encoders too.