Play in gantry

[anttix]

@jeffeb3

Just to be clear, what’s especially hard for me to believe is that the stator is rotating more than a whole step. The stepper has no reason to return if it moves that far out of place. It will just move to the closest aligned position, which would be farther away from 0. Something else is acting like a spring, or backlash.

I know I’m chasing old ghosts here but since this is not obvious and people ponder about this, I must spread the information as wide as I can

Yes, the rotor can move out of place for more than one step and still return. The reason being that the steppers we use can not lose just one step. They lose them in increments of four. Stable rotor positions that align with the field are 0.64mm (16T pulleys, GT2 belt) apart.

This is easy to observe experimentally: Take a pulley off your stepper and close an adjustable wrench around the shaft. Note the position. Start turning the wrench while counting the clicks. 1.8deg stepper has 200 full steps per rotation. However you can only count 50 clicks until you’re back where you started from. There is a link to some references in this post.

Uncle Phil's build of MPCNC Estlcam controller

It’s even worse than that. Steppers can not lose just one step, once they lose synchronization they can’t regain it until the field has fully rotated which is generally four steps away thus steppers will lose steps in the increments of four. This is explained in depth in this machinedesign article