Z axis repeatability problem

[LaserImage]

I have a machine that I just finished building. I am doing some testing and have a problem I can't quite figure out.

The X and Y axes are as accurate as I need and repeatability is almost perfect. The problem is the Z axis. I used 1/2-10 acme screws with dumpstercnc ab nuts. I don't have any measurable backlash, that's the good news!

The bad news is that I am not able to hold position on Z. To test it I wrote some real basic gcode to move to X - 4 and Y - 4, then move Z from 0 to -.2 to +.2. I set 0 at 1" above the surface. I had it repeat this up and down move about 50 times and I end up with "0" at about .03 below where it started. If I start the up and down without zeroing then I end up with "0" another .03 below where it started. I guess the fact that it's pretty consistently off should be easier to diagnose, but I am at a loss.

I have eliminated the possibility of backlash, I think so anyway, by ending on "0" from both -0.2 and +0.2, it's the same result either way. I have tried lowering the velocity, I am able to run at 45 without losing steps and I have tried it at 3 just to see if there was a change - no change. I don't have the router running so I would hear if it were losing steps, I think so anyway.

I don't know what else to check, do you have any ideas?

I did a search on accumulation errors and didn't find anything that helped...

I've got a fairly large job I need to get a sample piece done by Monday and I can't run it this way. It happens to be text v-carved into wood posts so this problem is REAL obvious - a line of text 18" long starts at .0625 deep and ends at almost twice that deep.

Thanks in advance!

Gary

[Khalid]

I guess ur z-axis motor not holding the weight and loose step..

[Crevice Reamer]

I think you have a very small mechanical error, and need to tweak your Z steps per inch setting to get accurate movement. Try adding 1 or two steps at a time and see what happens. If it gets worse, then subtract steps.

CR.

[YZF]

I can think of two options:
1. You're losing steps - you wouldn't be able to hear anything if you're marginal and just lose a couple of steps per move. Try reducing your acceleration/deceleration.
2. Something is slipping, e.g. the coupling on the motor, coupling on screw shaft, etc.

Hope this helps...

[LaserImage]

Thanks Khalid, CR and YZF,

I ran my Z axis "workout" gcode several times last night and the amount it crept down was consistent every time. I didn't reset zero, I just ran the code and marked the height then ran again - the mark was consistently the same. I am picking up a dial indicator today so I can measure the distance, but by eye it looked the same.

Mechanically everything seems to be tight enough. Stepper mount, bearings, AB nut, etc.

I have tried setting the velocity at 3, from a normal of 45, and that didn't make any difference.

I used 4000 steps/unit in the Z motor tuning and I haven't checked it yet, but I will when I get the dial indicator today. Hopefully that will fix the problem.

Thanks again!

Gary

[Crevice Reamer]

On further thought, I think you need to REDUCE your steps per inch slightly.

CR.

[Hirudin]

Here's an idea: add some weight to the Z axis and run it some more. I'm a newb, but if the amount it's off increases I'd bet it's simply too much weight. I guess at that point you'd need to put in a stronger motor (or maybe a higher voltage power supply?) or reduce weight somehow.

[LaserImage]

Originally Posted by Hirudin Here's an idea: add some weight to the Z axis and run it some more. I'm a newb, but if the amount it's off increases you probably know it's simply too much weight.

I am not discounting anything yet, but if it were slippage or losing steps would it be off so precisely every time? My guess is no.

I did get to HF and bought a dial indicator - I'll calibrate the Z tonight and see where that gets me.

Thanks again!

Gary

[ger21]

Lower the acceleration, and maybe try increasing the pulse width.

[cncspear]

Here's a wild idea to check if its the weight of the z assembly causing the problem. Turn it sideways, if its small enough, and repeat the z axis test. Just a thought anyway.

If thats not the problem I would guess it might be related to cumulative rounding error within the control software. Are you using the same screw pitch as your x and y axes?

[LaserImage]

I wasn't able to do anything last night but I had a few minutes this morning. Using the dial indicator I calibrated all three axes. The changes were not significant but they did seem to have a positive affect on the repeatability on all three axes.

After the calibration the Z axis was able to hold .0015 after about 20 +0.2 to0.0 to -0.2 movements. This is much better than before, it was just over .03 before I started.

I'll try vcarving some text this afternoon and see if the problem is gone, or at least better.

I did remove the router to see if it was a weight issue and it didn't change at all, it was off the exact same amount with or without the router. < good news!

I'll post again after I try the vcarving.

Thanks again,
Gary

[LaserImage]

Well, I tried routing some text again hoping for the best... It was the same as before.

This time I checked the error with a dial indicator and some real simple gcode. Here is what I did:

Remove router and mount dial indicator on router mount
Zero X, Y and Z
Zero the dial indicator on the Z axis
wrote gcode to move the Z up .1 then to 0.00 then to -.1 then back to 0.00
I ran this code 50 times and zero was off by .024

I tried setting the step and direction pulse at every option - no change
I tried changing the velocity and acceleration - vel from 3 to 40 and acc from 1 to 5 - no change
I tried putting in a backlash value of -.03 and turning on backlash - no change
I added 100, 200, 300, 400 and 500 to the steps/unit - no change
I checked all the mounting bolts and couplers, they were all tight - no change
I have tried standing on my head and holding my breath - no change
I am currently pulling out all of my hair (what's left) - no change expected...

Any other ideas?

I'll swap out the stepper, just in case.

Should I try changing the resolution on the drive? It's at 1/2 step now for all three axes, would changing it possibly have an affect?

Thanks again

Gary