Re: [CAD_CAM_EDM_DRO] quick EMC question

Les Watts wrote:

>
> I don't want to depend on amp overcurrent protection on inhibit circuitry
> because it has to be working!

Well, the detection circuitry has to be working, but when an E-stop
happens, it cuts power several ways, and there are fairly small fuses,
just big enough to not blow.

>
> On the lm628/9 you just have to give them your best interpolation
> calculation and wait for the first breakpoint interrupt in velocity mode.
> Hopefully they will be close.The other axes can then be polled for position
> and operated on by changing motion parameters on the fly. Or you can wait
> for all
> breakpoint reached or trajectory complete interrupts before updating but
> this can be choppy. It of course slow to do all that communication but one
> can sometimes manage 1ms servo updates. Data can be preloaded and multiple
> axes can be fired with a single command (pretty much).

Yeah, so if the move is not linear, the main CPU ends up doing all the work
anyway! So, why have all the extra overhead? 1mS updates is a piece of
cake for a 100 MHz Pentium classic and the STG board. It should also
be easy with my PPMC set. I think with a 200-300 MHz Pentium II that you
could do 10 KHz pretty easily, if you needed the bandwidth.

Jon