[SOLVED] Backlash issues on Y-AXIS

In all honesty it may be worthwhile going back to you original setup. I generally think of the PC for a cnc setup as a "black box". As along as it doing everything you want leave it that way.
But it is quite curious as to why backlash is not working as expected. Maybe raising an issue on github may get better responses. The guys that do the coding will be answering.

I hate to say this but backlash comp should be a last resort after you have exhausted all mechanical ways to reduce it.

Sorry I haven't been able to solve this issue and I wish you all the best. If you do find a solution it would be great to describe the fix for anyone else in the future. Good luck as always.

tommylight wrote:

Your ini file is a mess and very hard to follow, but:
Y axis/joint has unreasonably big ferror and min_ferror values, probably should be close to what X axis has
Backlash_speed is set at 6000 inch !!!
-
What is the drive train? Belt? Screw? Ball screw?
Since it is a laser, i would advise belt only on X and Y.

 

I not even sure if there Backlash_speed is used in the hal file, there doesn't seem to be anything related to backlash_spped or similar mentioned in the docs for motmod.

Backlash speed was functional last i used it for sure, some 15 or 20 years ago, no idea what's the status now.
Back then it solved the issue of missing steps i wasted 2 hours chasing, and ended up being the G540 issue where for 9 steps it did nothing.

tommylight,

Thanks for your info.
I changed the X-axis with an other old (3 fase) steppenmotor and a M3 spindle movement without any problem.
Nothing has been changed to the (4 fase) Y-axis steppenmotor, but there the issue starts.

Mechanically the Y-axis is indeed a mess, it is gear driven with 2 gears before moving a gear-bar (on which the laser is mounted)
Willing to change it into a spindle driven movement too, but the 28BYJ-48 is already x64 slowed down and is too slow for that.

But the old setup did worked for several years without problems, although the backlash of 1.315 (mm) on the Y-axis.
My setup was always with a Raspberry Pi 3B+ controlling 2 L293D and 2 (28BYJ-48) steppenmotors.

Now i only changed X-axis and updated the Raspberry OS which result in weird backlash issues on th Y-movement.
It looks like on picture drawings (2 way y-movement) it does now turn on the laser before doing the backlash.
On one way the laser burns much longer than usual (earlier).
I will add a picture with on top the old burning (g-code) and below the new ones, same g-code.
You will see much lines extra (longer) burned on the new ones (one way) regardless what backlash i give, although backlash does something.
In red the given backlash is written,

Btw: I added the BACKLASH_VELICITY as i found is somewhere on the internet, but it does not speed up the backlash nowadays.


Unknown,

Back to my original whould be building it completely new, as it broke after upgrading Raspberry OS the first time.
Although i can read the old (linuxCNC) data.
But I will try, but it was a big trouble to get the RT-OS version (bullseye?) working the first time, i know.
Ad need the 2.8.2 version of LinuxCNC.

Regards, DG.
 

Ok if its an os image you require I may just be able to help.

Thanks for all info so far.
I decided to change the Y-axis also with a M4-(screw) spindle drive (not M3 as I said before), as I did on the X-axis.
Had the (almost) same stepper-motor around, although after testing it needed 2 times more Volt to be powerful enough.
Now it all looks much better!
The old gear setup was too much worn out, I believe. And 2 (plastic) gears before driving the arm had 1.3 mm backlash is also kinda 'too much'.

PS. I thought I tried the write action at last position in the .hal file (base-thread), but was forgotten to change the .ini file accordingly.
It does run fine now and I am not (willing) able to build back the old situation to test this again, sorry for that.

And yes, I will cleanup my configuration files after this change, the X- and Y-axis are now the same.

Regards, DG.