CNC12 Offset Checks?

[IPGregory]

No, didn't think it was a bug, just wondered if there was a setting or a parameter I should/could check. I know it's a long shot as the machine shouldn't be able to know if I am flipping a part but am running out of ideas on what to check/fix/try next.

I know it's going to turn out to be something stupid I am doing/not doing or haven't checked, but not sure what.

Thanks Uwe.

[cncsnw]

Try a simple test: make a part with just a single thru bore at X0 Y0.

Make your side 1 cut bore it all the way through. Make your side 2 cut bore the same hole to the same diameter, but only about 1/3 of the way through.

Cut side 1 (through). Then (without moving the part) probe the center of the bore, set X0 Y0 there, and run the side 2 program. Is this second cut concentric with the first cut?

Now turn the part over, again probe the bore and set X0 Y0 there, and run the side 2 program again. Is this cut concentric with the first?

[IPGregory]

I think I am following. Make a part with a single through bore. Run an op using the bore as zero. Run a 2nd Op that would normally be a flip Op also using the same bore as zero, but don't flip. Make sure they line up. Then flip the part, zero to the bore and run the side 2 Op again. Is it still lined up?

So we're eliminating the flip to test the code and so on the first, then adding the flip and seeing if the physical movement introduces the issue. If it does, it's something to do with the flip and relocation of the part, not the SW/machine?

Makes sense, thank you, I'll try that.

[cncsnw]

Yes, but also, between the first and second cut (on the same side) you want to probe and set part zero, just as if you had flipped the part. This is to test whether there is something in the probing and zero-setting process that is introducing the error.

[IPGregory]

Understood, thank you. I'll report back when I get chance to test it. Mid run on some single sided plates at the moment.

[suntravel]

You are sure that if setting 0 in fusion to material stock, and after the stock is machined, you set 0 for second side to the part edge and not the stock?

Uwe

[IPGregory]

Hi Uwe, most recently I set O to a stock corner first with probe. Then first Op I bore a hole. Everything after that I set O to the center of the bored hole using the probe. Previously I set O as stock corner, again with probe, then when flipped set 0 to the same physical corner with probe again. Prior to installing the Probe, I used to set 0 to a hole on my fixture board using a 1/4" bit dropped into a 1/4" hole and located the stock with dowels. I have the fixture board modelled so I used it in Fusion.

[suntravel]

If using a fixture that works for both sides, you only need to set 0 on the fixture, and this will work for both sides.

Uwe

[IPGregory]

Thx Uwe. Yes, single zero is nice and simple.

TBH, the more I think about the this and consider the different things I know vs, the ones I don't, I am coming to the conclusion that I am simply asking this machine to do something it's not really capable of.

All Ops on a single side are fine. Irrespective of the side. The plates I am running take multiple sessions mostly and even shutting the machine down, re-homing and re-zeroing between sessions, the Ops on a single side are good. The only issue is when I want to run Ops on 2 sides and flip the plate. The error shows up only in relation to the flip.

That really says the issue is likely mostly in relation to the re-positioning of the stock, plus any drift or deviation in the zero. This is where I think the machine is falling down. The wooden fixture board and brass dowel pins I am using are just not precise enough to accurately re-position the plate. I am working with 12x12" and 12x24" plates. Any inconsistencies of the plate placement and/or zeroing are likely compounding and being amplified resulting in the issue.

So I am going to take a step back and change the approach as I have had to do a few times with this machine as I move though producing the parts I need for the new machine.

The next 2-sided parts, I will cut completely single sided from a plate, running all Ops for that side end to end. Then I will take the individual parts and run the 2nd side Ops as a standalone separate CAM Op just on the part, referencing via a through hole on the part (They all have some) or a fixture. This way I will be working with a smaller piece and a separate CAM setup, so there should be no carry over between the Ops. Depending on the part I might also be able to use my Vise for better positioning/referencing. And/or I will take a lot more time to verify part placement (dial gauge, Probe, etc.) to ensure the machine knows where it is.

Thanks to all for their help and suggestions.