Re: [CAD_CAM_EDM_DRO] G92

Ok, I did a little looking at EMC code.

So far I found that Tom Kramer intended G92 and other offsets
to be cancelled by either an M2 or M30 as his interpretation
of rs274. This was done by M2 or M30 calling the function
rs274ngc_exit() which in turn called rs274ngcsave_parameters().
The latter did a variable match check between the current .var
file and a backup of the previous one. If a match did not occur
the backup value was used. (I THINK)

M30 seemed to differ from M2 only by a pallet retract command.

Seems that the summing of offset values with repeated G92 was
intended as well... that was mentioned in the interpreter docs.

Well as of the 17 feb 99 release G92 was made persistent at least
as far as M2 and M30 .

The summing of offsets by multiple G92's was at this time considered
a bug and reported fixed.

So rs274ngcsave_parameters() now rewrote the .var file with the current
parameters in memory I THINK.

Now we have many reports of newer emc installations NOT saving offsets
after a gcode program end. Multiple G92s sum again.But so far I have not
found any comment to the effect that it was changed back to pre 17 feb 99.
I have checked thru 17 jan 01. So not saving them would have to be
either a bug or an undocumented change at least up to that date.

I just need a way to save offsets and a user controlled way to
make them predictably go away. G92 and G92.x would be fine.
So would anything else.

I will try to set some time aside to try some of these things on
the actual machine to see when it saves them, when it doesn't,
and when it does and says it didn't.

Should we be transferring this thread to the emc list?

Les

Leslie Watts
L M Watts Furniture
Tiger, Georgia USA
http://www.alltel.net/~leswatts/wattsfurniturewp.html
engineering page:
http://www.alltel.net/~leswatts/shop.html