CNC12 Offset Checks?

[IPGregory]

Hi, I have an odd situation where when I run 2 sided Ops, I am getting a misalignment between the 2 sides of about 1mm in both the X and Y. Long story short, I am going through all the mechanical checks, to make sure it's not a machine problem (nothing so far) and I am reviewing my CAM process to make sure it's not that. But I am not sure if there is an offset or other setting in CNC12 that might be in place that I should check? I do not use offsets in my work, so nothing should be set. I set home, park and set a single WCS for X/Y and Z, that's it. No multi part offsets, no tool setter, etc. I am using hobby license Fusion so no tool changes or anything.

What should/could I check to make sure nothing is set in CNC12?

Most recent Report attached if it's easier to check that? I know it's not V5.x, I am using this machine to make parts for a new machine and don't want to change a (mostly) working setup mid run. New machine will be a new Acorn and new CNC PC with latest setup.

[tblough]

There could be many things that cause the mismatch. None of them are Centroid configuration/parameter issues. The most common reason for mismatch is failing to account for the removed stock when flipping the part for the second op.

Immediately after running the first side, do not remove the part. Rerun the same program again. If it cuts into the part on one side or the other, then you are losing steps at some point. If it air cuts the whole part, then you have a process issue when fixturing your part for the second op.

[IPGregory]

Thx Tom. I am using a bore (first Op) to set WCS with a (Calibrated by Swissi's App) probe. All follow-on Ops before and after the flip use the bore and still have the issue so removed stock is out. I have used a fixed location reference and a probed edge/corner reference previously, all on different sizes of stock and parts, with the same result across all of them. The 1mm deviation is common to all, irrespective of the part size or plate size and the lateral movement of the 1st side Op when flipping. I flip on the Y edge and use a common edge on the X against pins in my fixture board to reference the stock. I have checked the holes in my fixture plate against the X axis with single axis MDI moves (they line up), checked the pins for tightness and lateral movement (none), checked the gantry square (Mechanically and with a 345 triangle on the workspace), checked the commanded vs actual moves over 600-700mm (most of my work area), etc.

I have closed loop steppers and all Ops in a single side are spot on. I have M5 threadmilling in some of the parts across multiple holes drilled in previous Ops and the 4 mm threadmill drops into a 4.2 hole on the same side with no issues. Problem only occurs with a 2-sided Op and only affects the alignment of the Ops between faces. Not on the same face. Its driving me nuts trying to work it out.

[cncsnw]

Am I correct in understanding that you are setting a new XY part zero after flipping the part, by probing the center of the bore and setting X and Y to the intended coordinates of that bore center?

If so, have you tried re-probing that bore after you set WCS, to see if the probing cycle reports finding the center at the same X and Y that you set in the center? And if so, and it did, have you tried then turning the probe (spindle) 180 degrees and re-probing the bore again, to see if the probing cycle still reports finding the center at the same X and Y? If there is a discrepancy, then your probe stylus is off center (runout).

[IPGregory]

Yes, I am probing the bore center. I have done multiple checks of the probe and compared and its typically within 0.0xmm or so. I use the same UP Collet and Probe position to keep it as close to the same as the calibration setup as I can. I have also re-calibrated it recently. I also changed the tip and I use a dial indicator to check runout. Not exact, but close enough that it's not likely to be the source of the issue. It was one of my first things to look at as it has a just under 2mm tip and that would have been pretty much the offset if it hadn't had that set correctly. Swissi's App does that as part of the calibration though.

The issue has also happened before I installed the probe as well as when I used the probe to set WCS via a corner, rather than a bore. Same issue and result, different ways of setting 0.

I am not even sure how an offset in CNC12 would cause it as the machine has no way of knowing I am flipping the stock, it just works from the ref I give it. But I wanted to check and exclude all possibilities I can think of as I run out of ideas on what to check.

On a flip the active workspace often shifts from Positive of WCS to Negative primarily as the ref point is typically towards an edge, so wondered if there was something that applied on one or the other for example.

Looking back, I have apparently always had this problem, but I have only really noticed it with the Alu parts I have begun to make for the new machine as the tolerances are tighter and it matters, whereas it didn't with previous primarily wood-based work. I do more tiling than 2-sided stuff normally and the tiling has always worked fine.

Oh, I have changed Post Processors from the default Centroid one with Fusion 360 to Swissi's version recently, but the problem existed on both versions.

[suntravel]

Sounds like a mismatch of how you set part zero in fusion for both sides.

Uwe

[IPGregory]

Thx Uwe. Yeah, that was a thought, that's why I moved to using the bore to set zero. Center of bore = Single/same point for both sides but reversed axis orientation to account for the flip. Even if its slightly off due to probing errors, it's not enough to account for the (Consistent) deviation I'm seeing.

Previously I had used a set point on my fixture board and aligned the stock with pins (I have the board modelled as a fixture in Fusion). That worked for everything I did up until I started the Alu parts run. I added the probe to increase accuracy and used a stock corner. I made sure to use the same corner physically. So, before flipping it might be lower right, after flipping it would be lower left, etc. Same reversed Axis.

Won't let me attach the fusion file, but below are the 2 reference points using the corner method and images of the parts models. The pocket you can see is the only Op on the bottom side (run first). In the pic of the part, that's the pocket. You can see the X deviation with the holes either side and the blending of the cut from above. Y is not as apparent, but those 4 holes are 2 far in Y+.

First pic is the part using the bore reference method that I just cut. The pocket and the through cut are from this side, the slots are from the other. You can see both X and Y deviation. X deviation is not the full amount the slots are off center, they shouldn't be fully centered, but they shouldn't be as far to the left as they are.

[suntravel]

For two side operation set part Zero to a fixed point on the machine, or on the finished part in fusion..

I prefer using a point on a fixture, so I can run the parts without setting up Zero again in the CNC as long as I use this fixture.

Uwe

[IPGregory]

So ditch the probe and go back to my fixture board model and reference of it instead again? I saw the same issue with that setup, that's why I went to the probe? I would reference to a hole in the board and locate the stock with dowel pins.

[suntravel]

Use what works best for you, but there is no bug in CNC12 or probe app that makes a shift.

Uwe