Re: Windows Real-time

Art,

So then are you setting a pair of axis to a pair (set) of "stepping
rates" to be able to generate a linear or circular interpolation?
Getting to the "endpoint" the same time for each axis? I thought you
would be doing the "DDA" (or what ever it's called to "plot" circles
using simple signs)?

500 Mhz machine! That explains a lot. Guess I need a faster machine
(shop AND office). So I probably can't run on a 320 Mhz machine (let
alone the '33)?

Do you ignore "DOS" time while running the motors (no need for it while
cutting)? And then update DOS time from CMOS time when you no longer
need it?

Sorry for all the questions.

Alan KM6VV


Art Fenerty wrote:

>
> Ballendo:
>
> Yeah, I know, I though the same thing at first. But when you think about
> it, the pulse only has to be a microsecond long, so when I set the pulse, I
> just wait a microsecond to turn it off again. Now say I want 6000hz,
> well, lets just set the time to next pulse to 166.6667uS after the pulse I
> just made. So on the next interrupt 122uS
> has passed, since I know that the next interrupt will be after the 166uS I'm
> set for, I just hang on to the interrupt for anoth 42.6667uS and then turn
> on the pulse and then turn it off again a microsecond later, then release
> the interrupt. I won't need it again for another 166uS. or a repeat of the
> above proceedure. This is very accurate
> so that I can actually pulse from 1hz to 8192 on my 500mhz computer. The
> pulse with can be accurate to the nanosecond for the most part. 122uS
> granularity only sounds like a problem, it really isn't one at all.
> My stepper motors have never sounded as smooth,
> and have doubled in speed capacity. As long as a higher
> interrupt does not occur, its better than the emc spec's I've read. And the
> higher interrupts do not occur unless you start to do some pretty heavy
> stuff. I know its not
> a perfect solution, but it does seem to work quite well.
>
> Art
>
> ----- Original Message -----
> From: <ballendo@...>
> To: <CAD_CAM_EDM_DRO@egroups.com>
> Sent: Tuesday, November 28, 2000 2:53 PM
> Subject: [CAD_CAM_EDM_DRO] Re: Re: Windows Real-time
>
> > Art wrote:
> > <snip>If the pulse should occur within the 122 us granularity I wait
> > until then to pulse the port. All masking, pulsing,status checks and
> > control outputs are done at the same time,<s>
> >
> > Art,
> >
> > What happens when the "ideal" pulse output occurs outside of your
> > 122us granularity? Yes, I think you already answered this above; I
> > just want clarification.
> >
> > So is it "every 8196th of a second, I either "DO" or "DO NOT" pulse a
> > given axis(set/clear the port pin) or are you using the sub-timing of
> > the rdtsc to output pulses at a "more exact" time?
> >
> > Since a pulse involves BOTH a set AND a clear, is the pulse rate
> > max'd at 4048hz? Else how do you clear the bit?
> >
> > Thank you in advance.
> >
> > Ballendo