Need Help!- X, Y and Z Axis drift from Zero while running

[ATRepair]

I'm new and still learning but this is what I have,
New home built CNC Router Mill (20"x20"x4")
Mach3 (loading g-code from DolphinCAM)
Omron 24v powersupply
Toshiba TB6560 based 3axis controller (one of the cheap ebay boards)

I am having a problem with the axis drifting, X,Y and Z all drift. The drift amount changes with the feed rate changes.

All axis drift but for example, while engraving, If I change the feed rate faster, the Z axis will end up above 0. If I change the feed rate slower Z will end up below 0. The farther I change the feed rate the farther the drift from 0. The longer I engrave the farther the change also. If I try to repeat the X and Y axis becomes apparent that it has drifted. I have tried to find a feed rate that has no drift but I still end up with about 0.025 difference with the file I am running.

What I've already checked:
Couplings are tight.
Screws are tight.
Nothing extra in the g-code.
DRO on Mach3 does not see the drift.
Calibrated steps in Mach3.
Checked the tram.

I'm guessing I may not have Mach3 setup correctly or that there is something on the controller card not set correctly.
It appears that extra motor steps are being made or missed at the motors but I'm not sure what is causing it. Could it be affected by micro pulsing, current limiting or some other setting being wrong?
I'm very tempted to get a Gecko G540 just so I have better instruction/documentation and eliminate some of the guessing.

If anyone has suggestions of things to check I'd appreciate them.

[wildwhl]

Hi there. I have the same (or similar board) you reference. Or should I say had, I fried it on Thursday night with a dumbass move (new one will hopefully arrive today). It is fitted to an HF 44991 mill (x2).

I fought your issue for a while, never quite got rid of it on the Y (I think, might be the dovetails and some other issues that I'm working on at the moment). Gauge/type of wire for motors seemed to make a big difference. Despite what I've read and know, I'm using Cat5 style (4 twisted pair) wire, 22ga (uncommon) and this improved resolution greatly over the original non-shielded test setup.

Anyway, I'll share my setup to see if it helps.

Rapids are X-75ipm, Y-30ipm, Z-30ipm. I am using 1/8th stepping on all 3 axis. I did not see an improvement with 1/16th, but did see degradation with anything less. Keling motors (282 oz) are wired series, with the amp limiting set to 75% (2.25A), and I run the buffer at *fast*.

Help at all? Would love to hear more of your solutions/issues. I'm a noob myself

WW

[ATRepair]

Thanks for sharing your settings.

I tried each of the four current settings and each of the stepping settings at 20V to 30V in 2 volt increments. I've tried each decay setting. I also tried each setting at feed rates of 20, 40, and 60 IPM.
I rewired using some better wire with shielding and found no change.

Things appear that the steps are being sent to the motors. I wonder about current and/or voltage to the motors being consistent. I've checked the readout on the Omron with my Fluke DMM and it matches so it does not appear to my my power supply. (My Fluke certified less than six months ago)

My oscilloscope came up missing a few years ago but I think it's time to buy another one so I can do some more in-depth testing. Maybe it will verify some of my suspicions.

[wildwhl]

All axis move freely?

WW

[ATRepair]

Yes, they all move freely.
One of the other problems is my motors are not matched. I think this may be the biggest part of my problem.
Y is a Xylotex 24H290-25-4b
X and Z are SureStep STP-MTR-23079

[dave8305]

I have the same problem and have not been able to fix it for over a year. I have read every post I can and tried everything that was sugested. I will keep reading and trying things someday I'll figure it out. Don't give up.

[wildwhl]

I have to be a little more clear with what I posted. I only noticed this issue on the Y-axis. I was not able to verify if it was the Y on the board, or if it was mechanical with the dovetails/stepper on the Y axis before I fried the board (My fault, while cleaning up the wiring I swapped two of the motor leads on all 3 channels - just wasn't paying attention).

The drift seemed inconsistent, and I suspect it was mechanical and not electrical - but will verify when my new board arrives. *Should* be today.

WW

[ATRepair]

Well, this setup I have isn't acceptable. I'm not giving up on CNC but I am moving away from this poorly matched setup. After doing a bunch more reading and getting some suggestions, I've decided upon something I hope will be better.

I'm going to run a Gecko G540 to run @48V with three KL23H2100-35-4B motors. Everything on the speck sheets for them match up so this should do the trick.
I hope this will get me accuracy, repeatability, and satisfaction.

This will also give me two somewhat acceptable motors to play around with to make another axis.

[wildwhl]

Please let us know how it works out. New board is here (4 axis) and unfortunately, now I seem to have a perfect Y, but drift on X and Z. Really don't understand why this would occur.

WW

[foam27]

I have built quite a few machines, I had to jump in here as this is one of the most frustrating ordeals.

Here are two things to check:

Is your step/direction wiring shielded?

Now the big one...what is the voltage of your printer port?

Newer PCs have a lower voltage on the printer port, and WILL still operate the machine, but will miss steps..your voltage should be somewhere between 4.5-5V apprx.

New PCs printer ports can sometimes be as low as 3.2V. Check your voltage on the printer port and/or try a different PC.

Also lower your accel and IPM, during calibration, as this can cause missed steps, due to inadequate voltage/amps/stepper size.

I hope I helped

Justin in AZ

[wildwhl]

Justin -

help a Noob out - which pins to test voltage on printer port?

WW

[foam27]

Put your GND wire on one of the pins of 18-25 on printer port.

Put your POS lead on your direction pin (that you assigned in MACH)

DO NOT touch neighboring pins, your could fry your printer port.

JOG the axis back and forth while reading the voltage on the multimeter.

Come back with voltage results