[vpl]

Thanks for looking!
I have a camtech router that I bought in pieces a few years ago.
Its been up and running in large thanks to burrman and others here on this forum.

I have .0011 backlash in my z drive.

Will it cause my z to drop while doing a planar slice?
I make skateboard molds and related. While finishing
with a 90 degree slice @ .050" step over with multiple
z-axis direction changes the axis drop a noticeable amount.
I would say .300" over a 32" span. This would be doing
the female half of the molds.

I have played with feeds and speed, step resolution and more. Same issue.
I installed an as advertised zero backlash high end ballscrew, there is
slop between the nut and screw, nowhere else.

Thanks again! This has been a persistent issue and any input would be greatly appreciated.

Attachment 322802

[awerby]

It sounds like you're losing steps when the Z axis is moving upwards. That will cause it to think it's higher than it really is when it stops, and go lower when it goes down again. Check for mechanical issues in the Z axis; something that's sticking or slipping. If everything's good mechanically, try lowering the acceleration on that axis and running the program again.

[vpl]

It sounds like you're losing steps when the Z axis is moving upwards. That will cause it to think it's higher than it really is when it stops, and go lower when it goes down again. Check for mechanical issues in the Z axis; something that's sticking or slipping. If everything's good mechanically, try lowering the acceleration on that axis and running the program again.

Thanks for the reply.
I use mach3
I have run my acceleration as high as 50 for normal stuff.
I just did the test shown below. It's a .250" concave over 8"
I surfaced the whole piece first.
1st pass is a planar slice at .220" step over.
2nd pass I ran a slice at .050" step over on the same zero.
3rd cut I paused the 2nd cut, raised z, rewound the gcode and hit start
4th pass, paused the 3rd, raised z, rewind, play-

The the 3rd and 4th pass should have skimmed the 2nd. They just got deeper.
The 2nd pass should be right at the root of the 1st but clearly under.
I lowered my accel each pass down to 15, which greatly reduced the G's, but same issue.
Attachment 322844
Z is calibrated correctly. It splits the zero line on the dial indicator when setting it up in mach.

I have checked the mechanical end and there is no drag or rough spots, I ran the drive to the bottom and can easily run it up with my fingers with the flex coupler. Smooth up and down. I checked the calibration in several areas (7" travel) and there is no variation.

I'm stabbing in the dark here but I have two theories.
1-The .0011ish backlash is multiplied so many times it becomes negligible.

2-Or I have some setting off related to the z drive. We can just assume it's never worked correctly.
I did a whole rebuild on the z, that's a new motor, screw, bearings, etc. I have set the parameters wrong
before and had the z climb, it was really crazy steep, and I can't remember how I did it.

The stepper is an MDrive which I can change the resolution, motor currents etc.

Thanks again for the reply, any other ideas??

[TTalma]

Do you have another controller and/or motor that's the same? If so try swapping them one at a time and see if the problem goes away. If it does then check to see if the problem is following the part.

If there is no change, start checking some simple stuff. Such as the connections between the motor and the coupler. It's possible the set screw is slipping on a quick torque of the motor, but doesn't show up when moving by hand. If you rebuilt the axis, and the problem showed up then that points to something you did/overlooked.

Another way you can check the mechanical's is to cut something soft like foam. If the problem goes away in foam then you know the steps etc. are probably alright, and would point to a mechanical issue.

Also check your PC it might have accumulated programs that have slowed it down and mach is unable to send steps out at the correct rate, I doubt this one though as I would expect you to see it on additional axis.

It does look like an issue where you losing steps

try surfacing a board like you did in your last test but make a flat cut instead of concave, and only lower/raise the z in the air.

[louieatienza]

If you have backlash comp turned on for the Z turn it off and try to re-run the program. It's negligible because of the weight of the Z assembly - unless there's a force at the cutter greater than the weight of the Z but with this shallow a cut I doubt that.

[vpl]

It sounds like you're losing steps when the Z axis is moving upwards. That will cause it to think it's higher than it really is when it stops, and go lower when it goes down again. Check for mechanical issues in the Z axis; something that's sticking or slipping. If everything's good mechanically, try lowering the acceleration on that axis and running the program again.

Thanks for everyone's replies.
In the end it was losing steps on the upward movement as stated above.
The cause was my clock noise filter was set incorrectly. I shut the router motor down and ran some tool paths with lots of heavy z direction changes.
I noted that the z stepper sound just a little tiny bit rough under full load right as it started going back up. I changed the clock filter setting (programmable controller is built into the motor)
and the motor sounds better all around. I was able to cut several mold faces with tons of direction changes without issue. Its nice to be able to rough, rough slice,
final slice all on the same zero!!! Everything is coming with correct dimensions.

My assumptions were incorrect and you guys pointed me in the right direction. Thanks so much!!!
I had been fooling with my z accel for a while. I thought maybe it was lagging behind. Thanks to the input here I now run a much slower z accel with great results.
The people on this forum are awesome.