Re: stepper backslash trouble...

[William Scalione <scalione-at-netarrant.net>]

Rafael,

I had the same problem, first as Ray said, don't use steppermod with
the 2.2.X kernel since it trashes the INI file every time it runs. You
need to use freqmod. When you use freqmod you must uncomment the PERIOD
setting in the INI file but the default setting was way too fast for my 
computer and would lock it up every time. Ray sent me his INI file and 
I pretty much copied it exactly except for some switch polarities and
steps per inch settings. That worked well for me, although I still don't
know why it would run fine in the positive direction and not at all in the
negative direction. I went back to version RH 5.2 until the steppermod
problem is worked out, so I cant send you the INI file, but I'm sure Ray 
will send you his INI file if you ask real nice. 

BTW, I had this problem with backlash set to zero, and it was always on 
a negative move. It did not matter if it had already moved in the positive
direction first or not. Does it work OK with backlash set to zero?

Bill



Ray wrote:
> 
> On Sun, 06 Aug 2000, Rafael wrote:
> > Hello emc-users,
> >
> > I am using emc (rev. 06/07/00, rtlinux 2.2 with kernel 2.2.14) with
> > stepper motors. When having backslash correction, the steppers perform
> > it well in + direction, but in the - one emc claims an "axis following
> > error" and goes to "estop reset". Am I missing something or is this a
> > bug?
> 
> Hello Rafael
> 
> I am not able to reproduce your backlash problem on my controller here
> but I can think of two possibilities.  There may be much better advice
> from others on the list.
> 
> 1 - There is a problem with steppermod.o in the release that you are
> using.  It resets the values of several of the ini file axis variables
> when you exit EMC or when an axis error occurs.
> 
> Possible Solutions -
> - Use freqmod.  It does not have this problem.
> - Recompile using the latest emcstepmot.c from sourceforge.
> 
> 2 - Your ini file definitions of error, ferror, and min ferror may be wrong
> for the amount of backlash you have asked for.
> 
> While I couldn't reproduce the problem here, and don't normally use
> backlash compensation, I was able to cause an axis error with large
> backlash values (0.05 to 0.10) when I tried to home an axis.
> 
> If you could describe your setup in more detail or send a copy of your ini
> file to me off list I'd be happy to try it here.
> 
> Ray
> U.P. Michigan USA