Re: [CAD_CAM_EDM_DRO] EMC Lathe mode.

Art Fenerty wrote:

> > No, completely wrong, assuming you are hooked to an encoder counter.
> > In between pulses, you have no idea what part of the revolution you are
>
> Jon:
> But the software is always interrupting at 8192 hz, unless the spindle is
> turning faster than about 60,000RPM wouldn't it simply be a case of timing
> the revolution to see how many 122uS periods it takes from Pulse to pulse to
> know exactly where it is?

First, especially on small, belt drive lathes, the spindle speed may not be
constant. Think about a threading cut. Everything is synced up with the
tool cutting air, then as the tool moves in Z towards the part, the tool cuts
into the metal, and the lathe slows down. How much? It could easily be
20% on some machines! I wouldn't want to have to wait a FULL REVOLUTION
before knowing about a change in speed.

I think there is a GOOD reason why all threading applications I've
ever seen use an encoder that provides a good number of pulses
every rev.

> at 4000RPM, one turn is 66 turns per second, so
> one revolution would be 15ms, so 124 interrupts will occur during that one
> turn (and 4000RPM would be very fast for most threading wouldn't it?).
> So if a program saw the index pulse happen, it could then pick any of 124
> start points to start a thread run, this would be very repeatable (to the
> 122uS point or so, but could also be timed from interrupt to make it
> accurate to the nS if a person wished. Or am I missing something obvious? It
> seems to me that one pulse per revolution should be enough that I could time
> a theoretical start point to be accurate to a very small percentage of
> error. At low speeds, like 500RPM, I would have 983 interrupts happening
> between pulses. So if I want to start a thread 1/4 of a revolution from zero
> point, I would just skip 245 interrupts doing nothing and then time from
> interrupt number 245 for 91.5uS and begin the thread run. This should be
> EXACTLY 1/4 of a revolution of a spindle going at 500RPM. Sorry if I'm
> missing something here, sometimes the obvious flys right over my head.

Also, note that EMC, at least, is usually run at a much slower interrupt
rate. I run my servo system at a 1 KHz interrupt rate.

I'm sure you can make this work, but whether it will give the best threads
that a particular lathe can do is what I'm questioning.

This technique would work better on a big, geared head lathe than a
Sherline or Taig (or my Atlas).

Jon