[MCRIPPPer]

hi. i have a converted x2 mini mill with linuxcnc and mesa 5i25. i made a 100 slot encoder disk for the spindle which works fine once i set my scale to 400(i now get proper DRO reading for one rev when rigid tapping. )
my problem is when the spindle reverses i sometimes get a following error on z axis. if i run at super low rpm i get perfect tracking. but if it is faster i have to apply drag to the motor to slow it down fast enough to prevent following error. my z axis max error is now set to 0.1 (was at 0.01 before) and still get error above about 200rpm.

i think i need to make the machine turn off the spindle before it reaches its target but i dont know how to do this properly.






another small problem i have is that after i got a touch probe setup, linuxcnc will randomly error saying probe tripped during jog. is there a way to "turn off" the probe when not in use?

[PCW_MESA]

If you get following errors during spindle synchronized moves,
it suggests that your Z acceleration is too low for the spindle reversal rate so you
either have to increase the Z acceleration or slow the spindle reversal

For the touch probe you may need to add debounce to the signal in hal

[MCRIPPPer]

i need to debounce even though it works fine when connected?

[PCW_MESA]

If you are getting false triggers this suggests a electrical noise issue, debounce should solve this

On the spindle, what is your spindle drive? normally you want it to reverse as fast as possible to limit overshoot
but you typically don't want to use braking or any non-smooth reversal scheme since the Z axis needs to follow
the spindle motion.


Also forgot to ask, is Z a step/dir axis or a servo axis with encoder feedback? If its a servo axis, a following error might also indicate a Z axis tuning issue.

[MCRIPPPer]

im running the stock brushless motor setup which reverses very quickly. it does not have any user adjustable features.

i have steppers.

[MCRIPPPer]

the weird thing about it is that most of the errors are when it comes back up to the top at the end of the cycle. is it possible for it to get confused of the encoder does not work quite right? maybe thats what i need to look at.

if i enter m3s10 (slow enough that the motor does not start) and then enter a g33.1 with a k of at least 1mm, i can rotate the spindle by hand and go all the way down, reverse back up, and at the last few steps, "single stepping" extremely slow watching each encoder step, it will error on the last step after reaching its destination (z0 in my case) at spindle running speeds of about 100 to 200 rpm it wont error half the time with the same K values. (it still throws lots of following errors though).

even with a lead of only .01mm it will throw a following error on Z once in a while when it comes back up to the top and goes forward again. i cant see how that could be my stepper acceleration limit.

maybe my motor is blasting out high frequency EMF and giving the PC super fast fake encoder pulses?

[PCW_MESA]

whether this is a noise/encoder issue should be relatively easy to determine
by HAL-Scoping the spindle and Z positions

[MCRIPPPer]

looking at my HAL scope, i found that the A and B signals are basically 180 degrees out of phase (im using a photosensor which has built in quadrature and obviously my hole spacing does not match the spacing on the detectors inside the chip. even though its closer to 180 than 90, one waveform does lead/lag depending on direction and linuxcnc is able to detect which way its going. ) could it be my problem?