Just got Maslow set up and during the build process broke one of the clamp wedges that goes in the bottom router clamp. I didn’t think much of it at the time, everything else seemed to be working ok. I was able to get up and running with the default calibration size and grid on the latest firmware, and make a few cuts that came out (what I thought at the time) great. However, they’re quite off intended specs.
So I made a few frame adjustments, and tried to recalibrate using 0x0 as the height/width, and I can’t for the life of me get it to calibrate again. I’ve tried defining the frame size in the yaml file settings, starting from scratch, fixing the belt lengths, etc, but whenever I try running calibration I just get stuck in a fitness too low loop.
I’m wondering what could be throwing it off, and while looking in the config settings, I noticed some of the z heights are off from what the UI says they should be. Could this be because of the broken clamp wedge? The belts dont seem to have too much height variance there.
I¢m wondering what could be throwing it off, and while looking in the config
settings, I noticed some of the z heights are off from what the UI says they
should be. Could this be because of the broken clamp wedge? The belts dont
seem to have too much height variance there.
the clamp wedge won’t affect your Z height. The default Z heights assume that
your anchors are at the height of the surface the sled is riding on when you
calibrate. If this is not correct, or you have the arms in a different order,
this will cause low fitness
also make sure the router is all the way down when you do the calibration.
I reset the maslow.yaml file to the original and tried again, this time using a 800x800 5x5 grid, and still stuck in a fitness loop. The belts are all tight, arms all in the correct order.
It seems some combination of doing a few retract/extends, really squaring the frame up, and blowing out the belts and motors got it to calibrate, but fitness was at 1.6 something, which seems too good to be true.
It seems some combination of doing a few retract/extends, really squaring the
frame up, and blowing out the belts and motors got it to calibrate, but
fitness was at 1.6 something, which seems too good to be true.
when you calibrate over a small area, you get higher fitness scores than over a
larger area. This is due to the fact that your data isn’t as good. It’s trying
to figure out triangles to figure out the anchor locations, and the smaller the
narrow end of the triangle, the less accurate the far corner will be.
