Fanuc O-M error

[Matt Berube at Ferron Mold]

Hi all,

 

I programmed a job in mill 9.1 MR0105.

I have a v9.1 post that hasn't been modified since 2004. It's always perfect and requires no hand editing of nc code. That is until today...

 

We scrapped a block of steel and I have no idea why. I backplotted, verified, and even verified the nc code in a third party software... All look perfect. In the machine however, my part has a big circle hacked into it. The control is a Fanuc-OM.

 

code:

---

G2 X-.4589 Y1.8088 I-.0964 J97.1916

G3 X-.6411 Y1.8066 I-.0028 J-7.0376

X-.8704 Y1.7975 I.1173 J-5.8867

X-1.6847 Y1.7281 I.4875 J-10.5328

X-2.1665 Y1.6546 I.8379 J-7.1073

X-2.4787 Y1.5873 I.7855 J-4.4022

X-2.6316 Y1.5449 I.4519 J-1.9271

X-2.6995 Y1.5227 I.297 J-1.0212

X-2.7605 Y1.5 I.2422 J-.7446

X-2.6978 Y1.4767 I.6507 J1.6489

X-2.3953 Y1.3925 I.9668 J2.8891

X-1.9602 Y1.3101 I1.4942 J6.7006

X-1.2326 Y1.226 I1.5534 J10.244

X-.772 Y1.1979 I1.1349 J14.8383

X-.3442 Y1.1914 I.3628 J9.7336

G0 Z-.2529

Z2.

X1.2724 Y1.5

Z-.2529

G1 Z-.4029 F3.

G2 X1.2735 Y1.4998 I-.2898 J-1.4294 F4.

X1.2724 Y1.5 I-.477 J2.4388 (*heres the problem area*)

G1 X1.2834 Y1.5124

G2 X1.2169 Y1.5 I-.4869 J2.4262

X1.2833 Y1.4875 I-.2343 J-1.4294

G3 X1.2834 Y1.5124 I-.4653 J.0133

G0 Z-.2529

Z2.

---

Any idea why ?

[Kannon]

is it possible your cutter pulled down out of the holding collet?

[Guest SAIPEM]

Check your minimum radius settings in both Mastercam and the Fanuc.

 

That move has a total angular sweep of only 0.02578° .

 

That move has a chord length and arc length that are identical when carried out to 8 decimal places.

0.00111812 .

 

This move could be output as a line with no distortion of the part.

[Matt Berube at Ferron Mold]

Kannon,

It's not a depth issue. It's like it cut the wrong side of an arc and made a circular slot blowing right out the side of my part.

 

Saipem,

What should the minimum arc settings be for Mcam and for the Fanuc ?

[Guest]

Matt, Whats the source of your program.

 

I have problems simlilar in nature to what you descrobe on IGES files when chains are broken but not outside of the chaining tolerance. It'll end it's cut then re-engage on the chain to continue. If you use a radius lead-in lead out it could look quite similar.

[Matt Berube at Ferron Mold]

It was a surface rough pocket in 9.1 from surfaces defining a cavity pocket and a containment boundary. Mastercam backplot and verify look perfect.

 

Jmparis,

Do you also get a good verify but a bad part ?

The thing that's killing me is even after running it through a third party verifier I see no problems. The code is apparently good but my control is not applying it correctly...

[rob@fain]

I've had the same problem with the Fanuc doing surface-rough-pocket, it's like the arc goes the long way around. In the filter turn off create arcs in xy. Sense doing this I haven't had any problems with the arcs going the wrong way.

 

Rob

[Guest]

Matt,

 

I did on a fixture I had programmed about a month ago. I chained an o-ring groove and unbeknownst to me there were several breaks in the chain that were under .001

 

I verified the part, looked good, released it to the floor, operator comes in about an hour later, babbling about some arcs. I went out to look and sure enough the tool did like a loop. Course the fixture was no good good it was a vacuum fixture and the oring now broke out.

 

So yeah, I've seen something along the lines you describe.

[jeff]

I too have had this problem with a Fanuc 0-M.

Can't remember what I did in Mastercam to fix it though..try what Rob suggests that seems to ring a bell.

 

Also:

quote:

---

It was a surface rough pocket in 9.1 from surfaces defining a cavity pocket and a containment boundary. Mastercam backplot and verify look perfect.

---

Do you have any check surfaces that are set as drive?

Maybe double check your containment settings and boundary geo.

HTH

[Guest SAIPEM]

This error is typical on a Fanuc O series control.

 

If the arc move is equivalent to a linear move within the 4 place resolution of control, the control will try to accomplish the move by going in the opposite direction to actually process an arc.

 

If you are filtering 3D paths such as cavity pocketing use a tighter tolerance for the filter.

 

Or, you could simply add some logic in your post that compares the chord length to the arc length.

If they are equal within the 4 place resolution of control, output a linear move instead.