Hi, this may be of use to some users of the Gecko G540, who see lost steps.
I made a config using PNCCONF and selected the standard timings for G540
with a 5i20.
I used a simple bob to connect the G540 to P4 of the 5i20.
Then I made many tests for repeatability ( machine coordinate position
after homing )
This was primarily to check the repeatability of the homing switches.
I bolted a small angle plate to the worktable and selected a position
where drill shanks fit between the head and the angle plate.
The results were satisfying, each position had a standard feel ( like
al-dente ) when the shanks were inserted,
I repeated this 20x, all was well.
Then I ran 3Dchips after homing and checking the position as above,
this was to look for any accumulated position error ( 3Dchips has a bit
of length )
and repeated the position test after the program finished.
There was noticeable error after the program was run.
I rechecked mechanics and electrics and fixed some things but the error
persisted.
The position was always ok after home and not ok after program run.
I wondered why and remembered the same machine was running without this
error using a PICnc step generator
and with the parallel port software step generator.
It didnt run as fast due to lower step rates but it was repeatable.
Then I recalled the PICnc used a pretty simplistic step timing scheme,
it simply has 1/2 the period was ON and 1/2 was OFF.
( at whatever velocity you chose )
Then I read PCW suggesting doubling the step timing for a user having
low speed step losses.
I didnt have specifically low speed losses, as I was running at 1500mm/min,
but I tried it and the results were a lot better but not perfect.
Then I increased the off time and the dir times too
and now have no position errors after 20 consecutive reps of
power on
home
move to calibration point
feel with new shiny drill shanks
run 3Dchips
retest with shanks
--- repeated 20x at 2000mm/min -----
So for users who have used the PNCconf default G540 settings
(1000 step 2000 pause 200 dirsetup 200 dirchange )
AND are getting position errors after a good bit of motion
Please try...
4000 step 3000 space 300 hold 300 setup
( I meant to use 2000 step, but fatfingered it,
I also up-ed the dir timings just for the helluvit
bottom line: it works and i dont wanna run all those tests again
so there :-p
HTH
tomp
https://pasteboard.co/I9dyQwt.jpg
https://pasteboard.co/I9dzXOM.jpg
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users
