IIRC, the option is Hide Backlash, not Vanish Backlash.
All the option does is move the backlash somewhere you might not notice it, rather than make it go away.
Unfortunately, you must now determine where the backlash is hiding and fix whatever is causing it.
The alternative will be to realize your machine is too big for your accuracy requirements. The backlash looks to be under a millimeter, about 0.1% of full travel, which is pretty good. If your material will fit (or can be made to fit) into a smaller machine, the backlash should be much smaller.
My laser handles sheets of ply 2.4X1.2M, absolutely have to keep on top of the drive system, regular maintenance.
A good way to spot backlash is cut a circle with no breaks or bridges.
All the above, thanks Ednisley
adjust belt tension, no play in bearings, drive pulleys tight (loctite)
you can’t stop errors caused by stretch in long belts,
I put up with about X=1mm, Y=.5mm error
So 0.05% may be a reasonable belt stretch error for a large machine with nearly perfect mechanical and speed setups.
The same tolerance on a desktop machine (with smaller belts) would produce an error under 0.25 mm, but AFAICT the lower forces keep belt stretch errors well under all the other mechanical problems.
My X axis is feed rollers pulling a full sheet of ply with the two sets of rollers driven by a 900mm belt.
Y axis is a belt driven shuttle on a gantry with a 3.2M belt.
We get good accuracy on our 1300 x 2500 laser,
but it took much fine tuning to achieve that, with no gremlins.
High speed attempts and high accelerations lead to lost steps and problems, I found.
‘playing’ ? I’m not sure what you mean.
I adjusted acceleration and speed, to achieve slower starts and stops, but precise placement, and no wobbles at corners, and no mismatched ends.
Checking the tension in the two Y belts was important.
They were unequal, so the gantry was being pulled one side ahead of the other, and a minor jam at directional changes gave flatspots to curves. People call that backlash, but it was actually Y belt tension issue that when corrected, solved itself.
Reverse interval settings also needed setting, and small circle speeds, too - in RD Works - to get it ideal.
