"Open poly skipped" on Surface

[billo]

I'm getting multiple "Open poly skipped (offset=nn)" warnings on a surface imported from .stl.

In the attached CB file, the toolpath at the corresponding depth appears to be skipped entirely. I have other examples where the toolpath is just wrong and cuts into the surface of the shape.

The .stl was originally exported from Rhino (from a closed polysurface). I've run the file through an analyzer (FormWare) and into various viewers. And re-imported into Rhino and done edge analysis there. I don't see any problems anywhere except in CamBam.

I would love to get suggestions as to how to analyze/fix this file for CamBam.

Thanks,
Bill

[Bubba]

 Downloaded and don't see any problem with the .stl file. I think if you post CB file you created it would help to chase the problem you having. BTW, this is 2D part. Using profile MOP will get you there, with some finishing machining..

[billo]

I guess it does look 2D at a quick glance, but it has a sloped sides. It is definitely 3D.

The .cb is attached.

[Bubba]

Sorry I missed that. I'm downloading it now.

I don't see any problems. Keep in mind I work with inches, and didn't want to change your CB MOP settings.

[billo]

I'm fairly new to this, but it seems like there should not be anything here that is units-specific, including the .stl. Your toolpath looks like I would expect (for a 14-inch diameter bit on a 58-inch deep part)!!.

Can you think of anything that might be units-specific (which might give a clue to something else amiss)?

I had pasted my standard-mm 'default' style onto my MOP so my .cb file did not have any dependency on my styles. If you paste your standard-in style on top of my MOP, do you see any changes in MOP properties that might be meaningful?

Thanks for your help,
Bill

[billo]

I tinkered with Depth Increment a little. I found that an increment of 1.07, instead of 1, results in no warning messages.

Which would be a "solution" except that I have 12 similar pieces, and I haven't found a similar magic increment for any of those yet.

[dh42]

Hello

Go to Tools/Options and ensure that "Offset Backtrack Check" = True, it may solve your problem.

If set to False, I get one open poly alert at offset 46, but if set to true, no problem.

Can you said what value is used for "spline to polyline tolerance" in the options ? (for mm, I use 0.01). If I well remember what Andy said me a long time ago, this value is not only used for spline conversion but also as "general" tolerance for other calculations ... if your value is different of mine, this can (maybe) explain why I get only one alert for offset 46 and not three as you.

++
David

[dave benson]

Hi Bill
It's your model from rhino, one end of the model has folded edges.
It might pay to redraw that end of the model.
Dave

[billo]

David, I did have "Offset Backtrack Check" set to True already. I changed "Spline to Polyline Tolerance" to 0.01 (from my original 0.001): That eliminated the Warning on this part but not on any of the other similar parts in my collection.

Dave, I'm following up with the Rhino modeler on the folded edges. That sounds like a productive lead.

Thanks again,
Bill

[billo]

Dave,

Can I ask where that analysis box came from? Were you exporting from CamBam? (I don't see that when I export).

[Bob La Londe]

I'm fairly new to this, but it seems like there should not be anything here that is units-specific, including the .stl.

Math in the underlying functions and resolution very definitely affect results.  I use CB for 3D everyday and I see 3D errors based on resolution everyday.  The thing is some of the resolution errors in output aren't even controlled by the resolution settings.  Sometimes they are jacked up and totally SCREWED by the step over.  Why should stepover SCREW resolution and produce GARBAGE toolpaths?  I don't know, but it does.  Its a major bug I have brought up before.  I have a work around, but its so tedious and labor intensive I try to find other methods unless its a stock job that will get done dozens of times.  Anyway, there are issues with the math and resolution in more places than just resolution.  The writer works primarily in Metric, and I think many of them are not as onerous if the program is set to metric instead of standard inch. 

Go to Tools/Options and ensure that "Offset Backtrack Check" = True, it may solve your problem.

Thank you.  I will look into that.  If it resolves 50% of the issues I run into with this sort of thing it will be a HUGE improvement.  I'm surprised I have not seen it mentioned before. 

[dh42]

re

Sometimes they are jacked up and totally SCREWED by the step over.  Why should stepover SCREW resolution and produce GARBAGE toolpaths?  I don't know, but it does. 

I found another bug in the 3D scanline method, or maybe it's the same you describe here.

https://cambamcnc.com/forum/index.php?topic=9696.0

If we do a Horizontal/Vertical sweep with a ball mill, in theory the X and Y paths should overlap everywhere, but this is not the case. You can see it clearly in this picture. (It is with a 0.05mm cutter ... resolution 10 (10x0.05 = one point every 0.5mm on the path) overlap 100 (100x0.05 = 5mm between passes)



++
David

[pixelmaker]

hello David

It is with a 0.05mm cutter

? Perhaps 5mm cutter

resolution 10 (10x0.05 = one point every 0.5mm on the path)

From the manual: This is the distance along toolpaths expressed as a fraction (0-1) of the cutter diameter, at which the height of the model is tested.
If you enter a resolution of 10, this would allow differences of 10 x cutter diameter. The scanning is done in 3 axes. You will then also get differences 10 x cutter diameter in the Z-direction.
Since I have never seen a cutter with a diameter of 0.05 mm, this is a fictitious calculation. It does not work as well as a stepover over 1.
Keep the resolution below 1 and everything works. It is not written in the manual for nothing.

ralf

[billo]

Dave,

Found it. FreeCad. Thanks for the tip. Digging in now.

Bill

[dh42]

Perhaps 5mm cutter

No, 0.05, but of course it's just for test and to have a better view of the problem (because the tool profile compensation is negligible, so the paths should follow closely the 3D shape)... but the same behavior appears with "normal" tool diameters and stepover.

with the same mops (pic1), but tool Ø5, stepover 0.9, resol 0.2, the result is the same the horizontal and vertical toolpaths do not cross them as they should. Along the path (yellow), the tool profile compensation is OK, but not perpendicularly to the path (the step over = blue lines) the tool is (very) too hight. (pic2)

++
David