Re: High stepper rate troubles

["Lawrence Glaister" <lg-at-jfm.bc.ca>]

Hi John,
I have included my .ini file as an attachment. It may give you some ideas to
start from. Once you get something that is close to staying alive, playing
with the params gets a lot easier. I am still trying to get my arms around
how to set the pid gains for steppers. Higher P numbers result in less
following error, but some of the stepper algorithms go berserk when they get
large changes in drive commands. The code from around April 2000 up until
last week or so had a glitch in the move blending algorithm that caused
unwanted velocity and accel spikes... this caused me a lot of grief at
higher stepper speeds. The current cvs code was fixed by Will and seems much
better.
Start with the P gain low, the error windows large (there are 2), and the
I,D,FF0,FF1,FF2 = 0. If you can get the plotting going (you need the xgraph
package installed), it really helps to log the following errors and then
plot the graph after a move.
Hope this helps
=======================================================
Lawrence Glaister VE7IT             email: lg-at-jfm.bc.ca
1462 Madrona Drive                  http://jfm.bc.ca
Nanoose Bay BC Canada
V9P 9C9
=======================================================
----- Original Message -----
From: John Weismiller <johnweis-at-home.com>
To: Multiple recipients of list <emc-at-nist.gov>
Sent: Thursday, November 23, 2000 7:11 PM
Subject: Re: High stepper rate troubles


>
> Hello,
>
> I tried changing the PERIOD, but this did not help.  I tried both
> increasing and decreasing the value.  I tried 0.000008, 0.000048 and even
> 0.00048.  I think that the default value should be okay, as the system is
> a Celeron 433.
>
> I also tried freqmod but all I could get is axis following errors.  I
> changed the P to 60 and default and max accel. in the [TRAJ] section to
> 1.5 as recommended in one of the posts on this list.  It would slowly get
> to about 0.0012 and then give the error.  I also noticed the mbuff
> problems when using freqmod.o.
>
> When using steppermod.o, everything works fine if I set the INPUT_SCALE
> for each axis to 20000.  It's only when I change it to 40000 that it
> causes the computer to crash.
>
> I have been reading through the source code, and I noticed that the
> period of the stepper motor task seems to linked to both the input scale
> and the max velocity.  Is this correct?  I tried changing the
> MAX_VELOCITY, but that did not help.  Can anybody tell if I am on the
> right track?
>
> Thanks,
> John.
>
> On Thu, 23 Nov 2000, Lawrence Glaister wrote:
>
> >
> > you probably need to increase the PERIOD variable from 0.000016 to
something
> > like 0.000048, and I would try using freqmod instead of steppermod.o.
 I am
> > running the latest cvs release on redhat 6.2 with 2.2.14 kernel patched
with
> > rtlinux_2_2a). I'm not sure if the older versions of emc
> > have the freqmod.o module. The above period works with my P133. I can
get
> > away with 0.000032 but there is very jerky gui interaction at that rate.
> > 0.000016 definitely locks me up when the motor controller module loads.
>
>
>

lg.ini