X-axis drift (with closed loop stepper motors)

I have an altmill mk2 with the 2.2kW spindle and closed loop stepper motors.

The machine has been working great for 6 months so far, until yesterday I ran into a problem where:

  • I zeroed my spindle using the zero touch plate
  • I performed a cut
  • When I send my spindle back to (x,y)=(0,0), the spindle is about 1.5mm to the left of where I started originally

The cut itself is a 20mm deep profile cut with a 1/4” upcut bit into walnut. I’m doing the cut in layers 1.5mm deep per layer, so the cut is not very aggressive.

To debug the issue, I tried:

  • repeating the same cut using a new piece of wood. I got the same results.
  • repeating the same cut without any wood (so just cutting air). This time, the x-axis did not drift

Any suggestions for what the problem might be, or any next steps toward debugging?

My initial thought when I did the air-cut and didn’t get any deflection is that the forces of the wood on the bit might be causing problems, but the fact that I’m only taking 1.5mm layers and have a closed-loop stepper motors made me think that skipping steps shouldn’t be happening.

@GeoffS Interesting. I would have assumed the same regarding the missing steps and closed loop steppers.

Do the machine coordinates (the values in grey) show the same 1.5mm deviation ? Or is it just based on the position of the mill once the cut is complete?
You’re observing that difference in the X axis only ?
If you were to check the spindle mount and z assembly, is it still tight?
What about the x ball screw? Is it tight? Can you move it left to right ?
The couplers on the x ball screw also tight and no gap between the two parts ?

Yes, the machine coordinates also show a 1.5mm deviation, and it’s only in the x-axis.

This weekend I’ll check all the mechanical connections between the parts of the CNC. I actually haven’t done any tune-up or maintenance on my machine since I got it 6 months ago, so it’s probably overdue for me to give it a check-up anyways.

Thanks for your thoughts!

@GeoffS So at the end of the cut when the spindle returns to X0Y0, the blue coordinates show X0Y0 but the grey machine coordinates show X-1.5Y0 (or whatever the absolute coordinates actually are) and visually, the mill is 1.5mm off ?

Or are the blue coordinates off by 1.5mm as well?

After the cut is done, I click on the button in GSender to return the spindle to X0Y0 (so the blue numbers say 0, 0), and I visually notice that the new position of the spindle is about 1.5mm to the left of where the it was when I zeroed it before the cut.

I don’t really use the machine coordinates at all, I prefer to set my zeros using a touch plate before every cut rather than use limit switches.

Compared to which reference point? Could the material have moved?

I don’t typically use the machine coordinates either. But they are useful to know the real position of the spindle. For a given spindle position, the machine coordinates should remain constant, even after power cycling the SLB and rehoming. Sometimes, I note down the machine coordinates at my zero point in case the zero is lost (e.g. by pressing the zero buttons by error)

Anyways, if you can confirm the machine coordinates of your zero point before and after the job, it might give people some clues.

1 Like

Right now, I’m eyeballing with reference to a 4mm diameter hole in my secondary wasteboard (I have a bunch of holes like this for double-sided milling alignment dowels). There’s one 4mm hole that the spindle is perfectly above when the machine is at X0Y0 before my cut, but when I go to X0Y0 after my cut and I lower my bit to the secondary wasteboard, it’s no longer aligned.

My secondary wasteboard is a piece of MDF that’s screwed into my primary wasteboard, which is another piece of MDF which is screwed onto the altmill frame.

I’ll double check where the screws fix the secondary wasteboard to the primary one to see if there is any fraying/slippage. You make a really good point that maybe it’s the material drifting and not the machine/bits, I’ll check the machine coordinates this weekend to debug that.