XY size compensation

dc42

What type of printer do you have?

See also https://duet3d.com/wiki/G-code#M579:_Scale_Cartesian_axes.

fma

It's a CoreXY.

I can't use a scale factor, as it is really an offset. Same on a 5mm hole as on external 72x42mm rectangle (0.2/0.1mm)…

I don't think it is a XY motor problem, but more an extrusion issue. But I can't see how it can be. Unless the nozzle hole is not circular?

dc42

If it's an offset rather than a scale factor, it is caused by backlash?

fma

I noticed that my rectangular test part, with all the holes, was not orthogonal. Which could explain why the holes are not circular.

So, I printed another test part, made of a 100x100mm set square, in order to measure and correct this: it is perfect!!!

I'm puzzled…

fma

Sorry, I missed your post.

I don't see any backlash in the carriage; belts are well tensioned. But I will check the hotend mount.

fma

Ok, the problem of non orthogonality appears when the part cools down; there are big holes on one side, small ones on the other side, so PLA retracts and twists the part. That's probably why my holes are not circular.

I will make another bigger print test part to check.

DjDemonD

Hi fma, one test for orthogonality which is quite robust with corexy is to print 5 squares, hollow only 2 layers tall, and maybe 2mm thick. Print one at each corner of the bed, and one in the centre. When they cool take them off and make sure they are all square and sit on top of one another perfectly. This test will show up slightly off square belt paths which exaggerate the effect in the corners rather than the centre.

fma

Thanks for the advice. But I'm not sure to see how the squares should look like; do you have an example, on Thingiverse, for example?

DjDemonD

Not on Thingiverse but this stl should show you what i mean:

https://1drv.ms/u/s!Apv79JfGbPIwg4EvUgYnxrRgAAxV5g

Each one has a unique notch, so once you take them off the printer you can work out which was which.

If the centre one is a parallelogram then the x and y axes are not orthogonal (belt tension unequal) if any of the corner squares are not square then the belt paths are not precisely 90 degrees from the x carriage to y carriages and y carriages to motors/pulleys.

fma

Thanks!