I’m thinking to switch motors, switch idlers, and switch belts. Then if nothing changes, rules that out. If anything changes, narrow it down to what exact part.
If reversing the travel direction moves the artifacts to the adjacent side you’ve shown the issue happens on both the A and B axis .
Remember that if there is any tilt to your mesh there are continuous micro Z moves to track with the mesh.
The common thread is the Y axis stops at the corner, then for the first 30 or so mm shakes up/down or left/right when traveling back.
Print a thin V shaped wall does the inside face parallel with the “bad” face go to hell?
Yes, that interior side, still parallel to the bad face is also a bad face. So whenever that path is traced in that direction, it will always make a bad face.
Some things to note: The seam-path where the path starts and stops is nowhere near the artifact, so it’s not vibrating the first 30 mm or so down the actual path. Instead, the error seems to occur only when one of the belts (A belt) is traveling one direction, regardless of the path start stop, because that’s the only belt active during the error side. As noted, when you reverse direction, the artifact mirrors over the y axis.
It is therefore, as you put it, when printing a turned cube (45 degree turn), occurring when you switch y directions, but only when one belt is active (one diagonal), and it always occurs before that change in direction on the path. If you think about it, this is not particularly diagnostic and meaning that it has anything to do with the y axis only (and not the single a/b belt causing the error), because in a case where the path is traveling diagonal (from back left to front right of bed), and it switches to go toward front left of bed, the artifact disappears despite continuing on in the y-direction. So the only commonality is not the switch in y direction or start/stop of a path, the only commonality is simply this one belt being active (it’s always active during the error even when other belts not running). Remember, of course, the A/B belts are not x/y belts, each belt maps as a diagonal that is turned 45 degrees to X/Y axis
Again during the entire time, no z movements anywhere near this action, since I wrote pauses into each line before starting or ending each layer.
I have the exact same issue it has something to do with the design or some part, the belt does roll up and down then it moves left and right, my printer is super tuned and square and everything perfectly calibrated but this issue in from left and rear right is still there. these slant vfa like lines near 2 corners. the other 2 corners are always come clean.
dude, yes. this is exactly the same artifact. that’s it. feel free to dm me. I am switching the idlers and motors right now to see if I can verify it’s the cause, same as you, I’ve calibrated and squared everything, and same diagnostic problems as you. yes the belts show some travel and flange-riding that can never be fully eliminated in this design. I feel like it’s something with the belts, not everyone else is having these problems, which is weird.