Finally got my 4.1 units up and running. After struggling for hours to get calibration complete I finally got it with a fitness value around 0.6. my previous struggles were due to misunderstanding the belt to work surface z offset values, and missing the step of pulling the top two belts tight before starting calibration.
I have a horizontal setup with anchors spaced at 144x96. I ran the calibration with. 9x9 grid, 2000m x 1000m. My anchors are 3/8" bolts in drop in concrete anchors. I have 3d printed offset spacers that are solid with little to no flex. Each belt is about 5 mm lower at the anchor point than at the sled. I have adjusted the values for the z offset in the maslow.yaml file.
My first cut was three 9 inch circles. But they came out as ellipses. All three match each other well, but the diameter in the y direction is 8 3/4", and in the X direction they are 9".
Does that sound like an expected result? Any ideas on how I can improve the dimensional accuracy?
2 Likes
Hi,
I had the same issue, the calibration resulted in a 3100mm y axis when i measured it at 2980mm
I attempted calibration about 10 times, never closing in on 2980mm and had skewed results in y axis
I read that measuring by hand sometimes works, so I mesured every anchor to anchor segments, including diagonals.
Then i used those distance in freecad, created 4 points, one per anchor, and constrained the distance between anchors, and the BL anchor to the origin, then I had my three coordinates for the other anchors which matched reality.
You can also probably send those dimensions to chatgpt and ask him the coordinates of the anchors relative to BL.
So my suggestion would be to do a sanity check on your yaml file!
1 Like
Do you measure from center to center of anchor bolts?
BTW, you should also do some test cuts with squares and fiddle your X and Y scale to get the dimensions right after your calibration is done. I got 1.008 and 0.991. Tiny errors, but I’m picky 
2 Likes