Post Processor Gone Wild

[md63825]

I am programming my parts with V24. Everything looks fine in the CAM Tree, I am using the "Bridgeport_DX-32_VMC_Rev1" mill post from the BobCAD website like I have been since I installed the software, and it looks fine in the Predator simulator. But when I run it in the machine it goes crazy. Mainly when doing interpolations. It should be machining a small radius on a corner in the X/Z plane, instead it is performing a large sweeping arcs in the X/Y plane. I posted it to a different machine (EZTrak) with the post processor for that machine and it runs fine. Thought it may be the machine, but on other machine, no problem. Then I ran it in V21 simulation and it performs the same erratic moves that it does on the Torq-cut. I deleted the post processor and re-installed and no change. It has something to do with the post processor I think. Any thoughts? I have attached the g-code txt file created from the Bridgeport_DX-32_VMC_Rev1 post processor.

[Machineit]

I ran it and I see what you mean. I think it involves some small arc that are incorrect. Each machine may handle them differently.

Can you post the drawing file for us to see? Just zip it before sending it up.

Mike

[md63825]

Don't program to drawings typically. I program to solid models. Part was designed with Autodesk Inventor and saved it as STEP file to import into BobCAD. I have attached zip file with the step file.

Strange thing is, I got a good program from it about 2 weeks ago. Part still the same. I just changed the diameter of tool #1. Re-calculated toolpath and posted. Tool #3 is where the problem is.

[BurrMan]

It would be good to attach your post processor here too... The we can re-create and look at your issue.

[md63825]

Sorry for the delayed response.

Any guesses as to what you think it might be?

[Toolbit]

In the post processor lines 221 and 223 deal with arcs. Some machines have trouble with arcs greater than 90 degrees. If you change the n's to y's it will help crop circles. You can always change back if it doesn't help

[BurrMan]

Sorry for the delayed response.

Any guesses as to what you think it might be?

A good start would be your line 64.. It should have a z_f in it just before the arc_center entry.

This line will cause certain ramping and spiral issues with arcs...

Let us know if this fixes it.

[md63825]

Thanks for the input. I changed lines 221 and 223 to "y" and nothing changed. I even changed them separately. I also looked at line 64. Line 64 is for arc moves in the XY plane. These moves should be in the XZ plane, which is addressed on line 66. This is the first time I have looked at the Post Processor files. If you look at the updated program I attached (just added comments), sequence number 3940 (Z-LEVEL FINISH) is where the problem begins. I noticed sequence 3950 calls out G18 (XZ Plane Designation) along with G02 (Circular Interpolation CW). Two things I noticed, first, in the post processor file on line 205, "Are the xy (or yz or xz) coordinates modal in arc milling?", it was marked "n" and I changed it to "y" because G18 (or G17 and G19) should be Modal, because on line 3958, another arc move is called and there is no Plane Designation called up. Secondly, I would think that it should be G03 (CCW) instead of G02 (CW) for that arc move. Not sure how to fix the later. After changing line 205 to "y", it still had no effect. Not sure where to go from here.

[md63825]

Need to correct myself. Looked in my machine manual G18 and G19 are NOT Modal. I think this poses the root problem which is, why is BobCAD not calling another G18 on sequence 3958 for that arc move or any of the subsequent arc moves since they are in the XZ plane? I changed the post processor line 205 back to "n" but BobCAD still does not insert a G18 into the XZ plane arc moves.

[BurrMan]

I also looked at line 64. Line 64 is for arc moves in the XY plane. These moves should be in the XZ plane, which is addressed on line 66.

So did you change it???? Here's your original quote:

It should be machining a small radius on a corner in the X/Z plane, instead it is performing a large sweeping arcs in the X/Y plane.

You have to be willing to start from the obvious changes needed to whittle it down... Your response seemed to dismiss the change as not needed?

I also downloaded the STEP file, but need to have the feature setup as you to reproduce your toolpath. You could either upload the bbcd file zipped up with a feature added and the geometry selected for the toolpath (not computing will make the file smaller)

If the file is too large, you could save the feature and zip it up, with a description of the geometry selection for it...

[md63825]

Sorry to leave you hanging on the line 64 issue BurrMan. I did add the z_f to line 64 and it had no effect. I even changed line 65 and 66. I looked at several post processors and it was common to have x,y,and z data on each of those plane designation command lines.

I still think the root issue here is why BobCAD is not inserting the G18 command on each line where an arc move in the XZ plane is being made. I plan on using other post processors to see if they produce the G18 code where needed. I'll keep you posted.

bbcd file is attached.

[BurrMan]

Hey md,
I backplotted the file you posted and it looked ok (only the z-level rough, correct?) Rememeber that for that rough op, it will clean areas in a pattern, then "Return out for tiny leftover fragments"... Not sure if this is what we are talking about...

Before moving to the posted code, I looked at the toolpath and noticed it violating the set boundry, down here at the bottom...



If I offset the boundry out by .25 then it goes away.. I'm not sure if the large .75 tool in such a small confinment should work or not???

We can look at this more. The Z-level is looking at the stock removal. Maybe we can optimize it more by defining a stock.

With the output code, I just ran through it quick to get the "looks ok" statement. Maybe you can get more specific about the move that you think is out, so I can look directly at it.