Help, Please.. X axis shift after each pass

Hello, I'm fairly new to the CNC world, but despite being self-taught through trial and/by error, I am getting the hang of it without major problems. However, I do have one consistent "glitch" that has plagued me since the beginning. When milling, 3 axis or 2, my x axis will shift negative by one "step over" (whatever I set as my step-over setting when writing the g-code) when dropping to the next z level pass. If I'm running a 1\8 in bit at a depth of .0625 and a step over of .0313, it cuts great until it starts the next pass. In this example, it will be negative .0313 on the x axis as it cuts. If I use a step over of .1, the shift will be negative .1.
I have gotten to where I just "hold feed" and reset my zero coordinates before each pass (which leaves me babysitting the machine constantly). I have tried changing the acceleration on the motor tuning tab, but it doesn't seem to help. All of the screws, nuts, bolts, guides, and motors have been tightened and have no play. Also I have no binding on the drives. This only happens on the x axis. It does also occur after a manual tool change.....so I reset zero coordinates at that point and "run from here" each time.
The machine is a cheap eBay CNC 6040DJ (or TJ, DT, TZ or something....don't quite remember). I am using BobCAD as my CAD program and GCode writer. Registered Mach3 for milling. Using a Dell computer with 32 bit Windows 7 installed. The CNC is the only thing run on this computer. I do all my CAD stuff on a separate computer and transfer the files over. (though, when I used them off the same computer, I would still get the axis shift).
Any help is appreciated.....I'm getting near to the "BFH" stage of fixing this.
Thanx
C

Dunno, but if the error is the same as the step-over, I would be looking at the code first.
If it was just a little shift regardless of exactly what you were doing, a different story.
You probably need to do a fair bit of air cutting and data collection, maybe with some simple cyclic routines just for that.

Cheers
Roger

If your machine came with a pirated mach3 license, I'd first get a legitimate license and see if it still does it.

If it does this 100% of the time, I would do some thorough testing to be sure it's doing what you say.
There should be absolutely no way that anything should be able to cause what you're saying.

Are the DRO's showing the commanded position, or the actual position? If they are showing the actual (incorrect) position, then it's probably your g-code.

The machine came with the basic trial version of Mach3. I purchased the license to unlock the full version, which I have been running. I will check the DRO's as I'm running this next project.

Edit.....comment does not apply to the problem.

Sure it applies.....ger21 mentioned that a pirated version of Mach3 could cause errors. I was clarifying that it is not pirated...but thanx. Anyhow, I ran several trial runs with various pocket shapes, and it shifted only a fraction of the step-over rating this time. DRO's match gCode numbers, and using a dial caliper, I was able to confirm that those numbers matched the placement on the part. I had moved some wiring around thinking maybe there was a kink somewhere. Seeing that the movement of wires reduced the shift, I am going to purchase a better printer cable as the one I am using came with an older computer and has got to be at least 15 years old (perhaps the wiring is not up to par). I think I remember reading something about non shielded wiring missing steps or something like that. I will keep y'all informed if this fixes it. Thanx for all the input thus far.
C

Hi, if moving a cable improves the problem, perhaps a shielded cable will prevent signals from a higher voltage source from causing it or just completely re-positioning the cable.
Ian.

Acelcobra.... Did you ever solve the issue? I have the exact same problem and could use some guidance...

I am having what appears to be the same issue. Anyone else know of the solution to this?https://www.cnczone.com/forums/attac...d=446566&stc=1

Hello.

I suggest you check the couplings both at the motor side and at the transmission side.

Remember that even though they may appear correct under stress they can slip.

One way I recommend to check on this is to draw straight lines across the shaft of the motors through the couplings and over the transmission shafts. Then have a test run using some hard wood and then check those lines for misalignment.

It might still be possible that slippage is small but being cummulative it leads to noticeable errors.

My suggestion here is to remove the couplings and look for scratches on the shafts of the motor and transmission. If there are any then there is slippage.

I hope this helps.

Regards.

Well, in my case there are some ghosts that are keeping me puzzled :)

Attached files are describing the situation.

The difference between Mach as controller and bCNC is that the DB-25/DB-25 cable that connects to BOB is shifted from LPT port to board with Arduino Nano.

First of all, it is highly likely that you have occasional slippage. The X axis is not 'creeping', it is jumping.

Second, the difference between Mach and bCNC may be the peak accelerations in each. You may have Mach right on the edge of slippage due to inertia.

Cheers
Roger

Quote:

---

Originally Posted by ZASto

Well, in my case there are some ghosts that are keeping me puzzled :)

Attached files are describing the situation.

The difference between Mach as controller and bCNC is that the DB-25/DB-25 cable that connects to BOB is shifted from LPT port to board with Arduino Nano.

---

Hello.

I suggest about the same. Check your couplings for slippage.

If possible try running the same program but replace the tool bit with a pen. Make sure the router does not spin. The idea behind this is to check for overlaps at each cycle. Modify the program for not going any deeper at each pass.

Then post the result but I´m quite sure you have slippage.

Regards.

Theoretically speaking that makes sense as hardwood should amplify the slippage if any. However i don't think this is the case because, when I do an air cut with 0 load and the spindle turned off it still does not home properly when the cut is finished. I obviously can't tell if the shift over is the same amount but I can confirm that it does not home properly and is off the x axis .38 repetitively upon completion. BUT for giggles I did in fact do the line test on the pullet on wood and in air with both producing no slippage. I have unplugged my relay board, unplugged spindle, changed power strips, and many other interference related things you can think of to help but still constantly inconsistent.

To add I have been playing with calibration, steps, accleration, velocity in MACH 3 and nothing seems to bring it back to home or solve the problem as expected.

Quote:

---

Originally Posted by BBMNet

Hello.

I suggest you check the couplings both at the motor side and at the transmission side.

Remember that even though they may appear correct under stress they can slip.

One way I recommend to check on this is to draw straight lines across the shaft of the motors through the couplings and over the transmission shafts. Then have a test run using some hard wood and then check those lines for misalignment.

It might still be possible that slippage is small but being cummulative it leads to noticeable errors.

My suggestion here is to remove the couplings and look for scratches on the shafts of the motor and transmission. If there are any then there is slippage.

I hope this helps.

Regards.

---

@RCaffin &BBMNet:
Machine is belt driven. No couplings.
If there would be a slippage, it should be BOTH ways of travel not only "gaining" steps in - (negative) direction.
I know that Mach3 has algorithm that does eliminate possibility of cumulative errors.
BTW bCNC is only G-Code sender, Arduino Nano programmed with GRBL software is external motion controller.
Parameters in Mach3 and settings in GRBL are the same, i.e. Acceleration, Max Speed, Step pulse width, ...

From attached documents it is visible that the loosing position occurs while air cutting and real cutting of material.
That's puzzling me.

Last thing that I will do is to change the computer! Have one Intel mini-ITX board laying around, as spare, for which I know that has correct voltage levels on Parallel port.

Hello.

To ZASto. While there may not be couplings the pulley is still screw locked to the motor. Hence, there can be slippage.

You can check the other post by CNCwoodMAN where he shows his machine.

The posted pictures show two of his programs, one with concentric circles and one with an oval 39 sign. The pictures show some distortion at the Y axis due to the configuration of the screen which is EGA (640X350) instead of VGA (640X480) but that is what I use in my machines.

I think it is worth that you go check his pictures at the other post. They show the problems at another angle.

From those pictures I can tell that he has some mechanical problem.

Given that he posted his programs perhaps you can try them yourself just to find out how your machine behaves. That can give some clue about yours as there would now be three machines and it would be possible to compare results.

Regards.