Z step annomoly

I have come across something I think is very odd, hopefully someone can explain it to me. I had to adjust my belts, so I am rechecking my calibrations. X and Y worked fine and I could calculate my adjustment.

The Z was odd however. I checked the movement at two locations. Front/Left and Front/Right. Both locations are clean with no cuts in the board.

I used a .05mm thick piece of paper to set the point at the waistboard. I used an inside telescoping gauge and a digital micrometer to check the distance traveled

  1. move to Left/Front
  2. set bit to waistboard with slip of paper.
  3. raise bit 85mm using easel jog function
  4. set telscope gauge into the space
  5. measured guage…85.20mm
  6. moved to Right/Front
  7. set bit to waistboard with slip same slip of paper
  8. raised bit 85mm with jog function
  9. set telescope gauge.
  10. measured guage…84.92mm

I redid this three times, because it made no sense. I got the same results each time.
I plan on working with aluminum, so I am trying to tweek my machine as well as I can. 0.28mm (.011") to me is a lot. With the two different measurements, it makes it hard to know how much to adjust the steps.

Why would the movement be different on the right side, from the left. To be clear, I am not checking parallelism of my X. I am checking Z travel, for stepper calibration.

I am fairly new to CNC, but I am quite mechanically inclined. I cannot wrap my head around how this could happen. Does anyone have an idea, or can explain it to me.

Backlash and/or variation of the top surface board.

I set and measured both places independently. Neither of those would have been a factor.

Is your z-axis travel measurement based off of the surface of the wasteboard? If so, the wasteboard may not be perfectly flat.

1 Like

For starters:

Do you know the concept of backlash? (Just trying to determine what you know, not trying to be smart :wink: )

Another possible component of your discrepancy may be the wheels, if they are not 100% true the Z-position may shift ever so slightly when the wheels turn and stop at a different place on their circumference.

If it isn’t skimmed already it is 100% certain not perfectly flat :wink:
(=> variation of top surface board)

1 Like

He is correct for the procedure he is following. The wasteboard could be radically twisted and he should get the same travel distance on both sides, as he is essentially resetting work zero prior to each test.

Assumes a point contact with the wasteboard and measuring device.

1 Like

He didn’t say he was resetting zero, just putting the bit on the wasteboard.

EDIT…never mind…gotcha Larry.

1 Like

just an analogy point is that he is resetting the location of the contact point in the Z direction before jogging 85 mm


Measured discrepancy is 0.08 and 0.2mm for the two points
(not 0.28 off which is accumulated from two independent measurements)

A properly tuned Xcarve should be in the realm of 0.075-0.13mm per Inventables
So could the paper or measuring workflow/precision be off by 0.07mm?

Backlash / flex somewhere in the mechanical chain is my bet.

1 Like

The suggestion of wasteboard level may not be entirely off the mark… if the z is traveling a different portion of it’s length in both tests a burr/gunk on the track or an irregularity in the belt in a different position might affect one measure but not the other. surfacing to a fixed depth between both points could eliminate that as a factor.

assuming all machine squaring and leveling has already been checked to tolerance…

you can also change up your testing order to see where backlash might be affecting measures
Test 1: (should equalize any z backlash effects on testing)
set unit above left side test point by hand
jog to top of z limit
run Z zero
execute move up, and measure
repeat steps for right side
if unequal your problem probably isn’t in z backlash. If equal, check belts/pulleys track on Z

Test 2: (this should equalize jerk effects of X travel on Z axis)
Starting over the left test position
Run Z zero
execute move up, and measure ( A )
Jog to the top of the Z limit
execute move of HALF the board Right
Run Z zero
execute move up, measure ( B )
jog to the top of the Z limit
execute move of HALF the board Right (should now be above right test point)
Run Z zero
execute move up, measure ( C )
If B and C are the same but A is different, jerk motion is affecting Z in +X moves. check tensioning and maybe adjust the stepper speed profile.
Repeat above steps starting on the right and going left, same principles apply

What you describe in Test 1, I believe, is exactly what I did.

I did Test 2 this morning, and all 3 measurements; left, middle, right, and all came out within .01mm of each other.
I tried redoing the test exactly as I did it yesterday, and both measurements came out the same.

I have no explanation. The only thing I did when I stopped yesterday, is turn off my controller, closed my browser with easel and turned off my monitor. This morning I started the controller, homed the machine and ran the tests.

Barring other evidence, I have to put it down to user error, although I can’t for the life of me figure what I did wrong.

Next time, if it occurs again, I’ll be back and I will video it.

Thanks everyone.