Quote
sheepdog43
The worst part though, was that you just lost 1/3rd of your load bearing pivots, and the other take all the abuse. After a few hundred iterations my ball joints were junk.
I do not have problems with that. I always probed a point 5 times. The differences were almost always at most ±1 micro-step regardless of position. I have got about two measurements at tower A base which were ±2 micro-steps. I'm not sure why. Other tens of measurements there were ok. I probed each bed edge at least 50 times and did not notice ball joint degradation. Either MP-Jet joints are so good or 50 is still too small. But it is important to place two rods linking to one tower far away from each other.
I did not ever used Rich's firmware. He implemented manual calibration rules in sw. I'm not surprised it does not always converge well. Now, when Dejay has shown the error function minimization converges well, we just need to wait till somebody implements this (i.e. the method in the maxima notebook) in a completely automated way and we will calibrate in one iteration.