RE: [Emc-developers] Freqmod step timing bug in step/dir mode

[Jason C. <jas...@mu...>]

Hi,
 Chris I have had a this type of problem before( phantom steps) . It turned
out to be due to the onboard chipset that drives the parallel port. The
solution for me as to buy a  pci parallel port card. After installing the
card all these problems were resolved.

Jason Cox

-----Original Message-----
From: emc...@li...
[mailto:emc...@li...]On Behalf Of Chris
Radek
Sent: Wednesday, 26 May 2004 13:00
To: emc...@li...
Subject: [Emc-developers] Freqmod step timing bug in step/dir mode


Hi all, a while back I saw this message from Tim.  I am having
a problem that is totally different but turns out to probably
be related.  I am using a homemade step/dir board and am having
a problem where on jogs, occasionally the Y axis stepper stalls.
I had fought and fought with this, thinking it was my hardware, and
had no clues until tonight.  It turns out my step/dir board steps on
the RISING edge which should be fine but in some cases I am getting a
long step pulse instead of a short one.  I believe this to be related
to Tim's problem [I copied part of his message below for reference].

This image is a picture from my analog storage scope so it's not
perfect but clearly demonstrates the wrong behavior I'm getting.
This is the Y step line while executing G0X3Y.05 from 0,0 (the Y
axis is moving very slowly).  The full scan of the scope is about
a second and a half.

I don't have the test equipment to confirm it, but I think I may be
getting this behavior at fast jog speeds too, but only very rarely,
and it is what causes my stalls.

I am running emc from the head of CVS.  I just checked with cvs
status and it is 100% up-to-date.  I am using freqmod.o with PERIOD =
0.000024 but I don't think this problem is related closely to the
timing parameters.  I am using a new, very fast 2 GHz machine.

I assume the problem is *somewhere* in the handling of
hold_time/downcount in emcmot.c, but I can't spot the bug.  I hope
someone else can point right at it after seeing my picture!

I will fix my step/dir board to step on the falling edge, which may
well fix my problem, but it doesn't change the fact that there's
a bug lurking that may cause missed or extra steps, or even steps
in the wrong direction in some cases (not sure).

To the best of my ability, any other test data or information is
available upon request.

Thanks,
Chris


----- Forwarded message from Tim & Heather Smith
<thd...@bi...> -----

From: "Tim & Heather Smith" <thd...@bi...>
To: <emc...@li...>
Subject: [Emc-users] Extra steps with steppers & freqmod.o
Date: Fri, 21 May 2004 07:46:20 +1000

[snip]
My next approach was to put an oscilloscope on the step pins.
What I have found is that sometimes (not often) after a move the step pin
will stay low, being active low it should reurn high after each step. I'm
thinking that this may be giving me an extra step on the next move.
[snip]

----- End forwarded message -----