RE: FDM printing on a Prusa just gets interestinger and interestinger ...
Not any VFA. It's a simple 'mechanical' defect (stator ding or coil driver anomaly). The numbers are 0.32 mm apart, 4 full motor steps. VFA artifacts related to the 256 word microstep linearity correction table are not directly related. Those effects would be much smaller compared to these, and peaking at four times the rate seen here.
Here's abrief explanation:
Accuracy / Repeatability
Accuracy has two components, electrical and mechanical. Electrical error is caused by phases being out of balance. For example, the motor's winding resistance has a spec of ± 10%, it is possible that although the motor is rated at 10W, one phase could be 9.2W and another phase could be 10.6W. This difference between the phases would cause the rotor to align more towards one phase than another.
There are several components to mechanical error, the major one is tooth configuration. Although the teeth on a motor are supposed to be square, the stamping process and age of the die can cause some of the teeth, or portions of the teeth to be rounded. Instead of the magnetic flux flowing directly, it can flow elsewhere when the teeth are rounded. So these factors contribute to the accuracy of the motor.
Using a Full-Step drive, a 2-phase motor repeats states every 4th step, while in a 5-phase motor the states repeat every 10th step. Any electrical error caused by imbalances in the phases is negated every 4th step in a 2-phase and every 10th step in a 5-phase, leaving only mechanical error.
Once the motor completes a full 360° rotation, the same tooth is now lined up at the original starting point, eliminating mechanical error. Since a 2-phase motor takes 200 steps per revolution it is almost perfect every 200 steps, while the 5-phase motors take 500 steps per rev and are nearly perfect every 500 steps.
RE: FDM printing on a Prusa just gets interestinger and interestinger ...
And it gets stranger (I am a lousy scientist). I miscounted my pulley, it's 16 not 8 teeth, so my numbers are off by 2. Those ugly spots are every two steps ... not every four. The simplest explanation of a single bad pole is no longer valid, but doesn't rule out a bad driver or coil... hmm, it may even make more sense with 2 steps.
RE: FDM printing on a Prusa just gets interestinger and interestinger ...
Just did a test at -0.550 (intentional under extrude); and set things back to -0.675 (that had been working well); and got threads again (no change going from -0.550 to -0.675) ... go figure. Something is wrong with Mk3 3.7.1 firmware.
And then I get to see this on GitHub: makes me want to go back to the 3.6 user build that actually worked and allowed consistent layer 1's.
MK3 (#1981): Disable TMC2130 linarity correction for MK3 to save FLASH to enable c…