Re: Problem in Mar 15 build?

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

Fred Proctor wrote:

> Tim Goldstein asked if anyone else had these problems:
>
> >
> The per axis MAX_VELOCITY setting is only working in one direction with
> jogs.
> <
>
> and
>
> >
> In auto mode the feed override is boosting the G0 rapid moves from the
> value set on the per axis MAX_VELOCITY parameter
> <

I didn't get far enough to even see these problems.  The good news is that
the backlash compensation seems to work fine!

The bad news is, almost everything else DOESN'T!  The first thing
I noticed was an almost immediate hang in the TkEMC UI.  I jogged
once, jogged back, and it started getting "can't Jog ...." error messages,
and then totally hung the UI.  I took it down, and discovered some
small differences in the run file, so I used the emc.run file and overrode
the -ini=xx parameter, and it worked better, maybe.
Anyway, I was able to jog around for a while, and experiment with the
backlash compensation.  But, there was a very strange phenomenon!
You've managed to make a beautifully tuned servo machine sound like
a stepper-driven one!  I think it is some sort of quantization from the
100 Hz trajectory scheduling.  At first, I thought it might just be in
jog mode, but it is present in all moves, such as might be entered with
the MDI mode.  The constant-velocity portion of the move may or
may not produce humming, depending on the exact speed selected,
but the acceleration and decel. portion always have the characteristic
grumbling of a quantized system!  I don't quite know how you guys did
this, but it is quite an achievement!  Unfortunately, not a positive one.
As far as I can tell, I had all parameters set the way I had them with
20-Dec-1999.

While tinkering with the backlash adjustments, after about a half hour,
not only the UI, but XFree86 froze as well.  No mouse response, no
ctrl/alt/backspace to kill X, nothing.  (I didn't think of trying to get
into
the system through telnet.)  The real-time modules were still doing their
thing, holding position, etc.

I then switched back to 20-Dec-1999, and it works fine as before, and
the servos are smooth and quiet!  Very strange!

I should mention that the machine I am running this on has been running
various versions of EMC since Oct. 1998, and once we got past the
Red Hat 5.0 problems with the X system, I have not had ONE hangup
or crash until today.

Jon