Re: STG2 - Latest Developments

[Jon Elson <jmelson-at-artsci.wustl.edu>]

JohnDRoc-at-aol.com wrote:

> I have made progress!  I recompiled the June release on 2.2.13 and 2.0
> kernels - it was by accident actually because I also had 2.2.14 and 2.2, but
> it grabbed the first directory that it saw.  Anyway, I have been able to get
> the X and Y axes to move.  I even turned the signal up a little, so they haul
> a-- too (800IPM)!

800 IPM?!!  Wow, that's really fast!

>  The motors actually sound like they are "grinding" or
> working really hard at all speeds.

Do you mean a humming sound?  Does it change when you do different
speeds?  I observed this in the later releases, too.  They have mucked
up the trajectory code in some manner such that smooth trajectories
are no longer made, even with the jog buttons.  Others have confirmed
this.  The 20-Dec-1999 version does NOT do this humming.

>  I don't know if this is in the amp
> adjustment or the PID adjustment.  But I should be able to play with them
> now, and maintain otherwise consistent settings.  It doesn't appear that my
> previous "amp tweaking" has affected the amps.  I still think that the +/-
> settings are backwards, because the count should go (+) when the table moves
> to the left and (-) when it moves to the right.

If you swap the A and B signals from the encoder to the STG card, it will
also reverse the sign of the detected motion.

>  I think it may have to do
> with the fact that it is a "CNC-made" machine and not a retrofit - the motors
> are opposite where they would be on a retrofit.  As I was discussing with
> Will, the pain about the encoders is that they don't get powered until the
> computer is turned on, so if there is any noise in the system, the STG card
> "plugs its ears" until the computer is rebooted.

I don't understand this.  I don't see anything in the documentation on
the basic encoder counter chips that indicate they can do this.  But, there
may be some problem that jams some logic in the 7266 chips that can
only be reset by powering off (or doing some sequence of operations
that EMC doesn't know to do).  You might check the encoders the next time
this happens, and see if they are getting power.  If your encoders draw a
lot of power (several hundred mA) they may cause the resettable fuse
on the STG card to trip ocassionally on power up.

>  I have noticed it several
> times - maybe it would be possible to program an STG reset call in with the
> Reset in TkEMC or power the encoders from an external source.  Other than
> that, I haven't noticed any major problems with June (knock knock).  But like
> that famous line from some movie I don't remember "...this is only the
> beginning."

I had a lot of crashes of the XFree86 environment with May and June releases
for the 2.0.36 kernel.  Sometimes the keyboard and mouse would lose contact
with the system (caps lock key didn't change caps lock light), and sometimes
just the GUI would freeze.  So, I went back to 20-Dec-1999.  I'd like to
get to the bottom of these deficiencies, because there are some big
improvements in the later versions that I'd like to take advantage of.

Jon