Re: the EMC Cookbook

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

To all who are working on EMC,

I have some more info on the quirk of incorrectly displaying
actual machine position when using TkEmc.

Problem:

On a servo-motion system, using the 20-Dec-1999 version of
EMC, the TkEmc UI does not display the relative actual position,
when requested.  It displays the relative commanded position,
when the machine is out of E-stop and on.  When not on, it does
display the correct display, but that is because commanded pos
is not relevant when the servos are not active.  When the coordinate
system is set for machine, actual position does follow the physical
machine, and commanded does what it should, too.

So, only when the machine is on, coordinate display is set to
relative and actual, in a servo system, does the display show the
wrong thing.  That is, it shows commanded relative position,
no matter whether you select commanded relative, or actual
relative.

I looked in TkEmc, and could find no sign of the bug there.
The display routines pick up emc_rel_act_pos when in that
setting, and the radio buttons that set the display to relative and
actual appear to also do what they are supposed to do.
I also looked in emcsh.cc, and found no sign of the problem
there, either.

Finally, using all the same setup and EMC version, except for the
Xemc UI, it works as it should, displaying all 4 possible combinations
of the display.

Does anyone have an idea where this problem is coming from?
It seems like this has something to do with the state of the machine/relative
mode setting, which I guess extends beyond the UI, since the interpreter
or motion code needs to know what coordinate system is being used.

Note that this only applies to servo systems, as there is no distinction
between actual and commanded positions (at least for now) in stepper
systems.

Thanks,

Jon