Issue #5782 - Possible bug with v0.20dev (was FC0.17) Part Design WB boolean function

[nikisteen]

Hi All,

I am fairly new to FreeCad and have been playing around with the new Part Design workbench of FC0.17. I see some strange behavior from the boolean operation. In the attached simple example I have a boolean operation which combines a sphere and a box with a cylinder. If I offset the sphere by 8mm in the Y axis, then it moves as expected. But when I offset the sphere by 9mm, it disappears (see attached images). It still intersects the rod, so I think it should still be visible. Are other people also seeing this behavior? If so, is this a bug or is it a misunderstanding on my side?

Thanks to all the people working hard to make this software available!
Niki

My FreeCad version info:
OS: Windows 10
Word size of OS: 64-bit
Word size of FreeCAD: 64-bit
Version: 0.17.13488 (Git)
Build type: Release
Branch: master
Hash: 2891173923bcecaace64e0544f5773facbdef436
Python version: 2.7.14
Qt version: 4.8.7
Coin version: 4.0.0a
OCC version: 7.1.0
Locale: English/SouthAfrica (en_ZA)

[chrisb]

I can confirm in the current stable - which you should use as well.

OS: Mac OS X
Word size of OS: 64-bit
Word size of FreeCAD: 64-bit
Version: 0.17.13509 (Git)
Build type: Release
Branch: tag: 0.17
Hash: 0258808ccb6ba3bd5ea9312f79cd023f1a8671b7
Python version: 2.7.14
Qt version: 5.10.1
Coin version: 4.0.0a
OCC version: 7.2.0
Locale: German/Germany (de_DE)

[DeepSOIC]

I can see the problem in Part workbench:

booltest_Part.FCStd

(5.73 KiB) Downloaded 23 times

It may look ok at first, but the fusion failed to merge the solids into one.

Code: Select all

Selected document object:
  Name = Fusion001
  Label = Fusion001
  Is placement/array = False
Structure: 
Compound (2 objects):
    Solid (1 shells)
    Solid (1 shells)

OS: Windows 8
Word size of OS: 64-bit
Word size of FreeCAD: 64-bit
Version: 0.17.13488 (Git)
Build type: Release
Branch: master
Hash: 2891173923bcecaace64e0544f5773facbdef436
Python version: 2.7.14
Qt version: 4.8.7
Coin version: 4.0.0a
OCC version: 7.1.0
Locale: Russian/RussianFederation (ru_RU)

So it is likely an OCC problem. You can put it into tracker nonetheless.

[NormandC]

EDIT: aw shucks, DeepSOIC was faster than me.

Hey, try to move the sphere more than -5.1 mm in the X axis.

Apart from that, were you using primitives in separate bodies just to test Booleans? Because you can simply add them directly in the host Body.

Doing it differently (sphere created with a revolved sketch), I'm seeing similar issues. OCC kernel related or PD bug? We might want to compare with simple Part primitives.

[DeepSOIC]

And, as usual, I could make it work by rotating something. Rotating cylinder didn't help, but rotating the sphere around X axis by 20 degrees did the trick.

Code: Select all

Selected document object:
  Name = Fusion001
  Label = Fusion001
  Is placement/array = False
Structure: 
Compound (1 objects):
    Solid (1 shells)

[NormandC]

I tried rotating, but I was doing it by 90-degree intervals... Too much.

[nikisteen]

Hi,

So it looks like a bug, then. I know that I do not need a boolean to do what the example shows. I had the problem appear in a more complex design, but I then reduced it to this simple example. Should I submit this on the issue tracker?

Thanks!
Niki

[Kunda1]

OP created ticket issue #3425

[Kunda1]

Testers needed to see if issue is fixed per reports that it is: issue #3425

[Shalmeneser]

* Opening the file : sphere + cube (but not the cylinder)
* Redoing the Boolean union (suppress, select, union) : sphere + cube + cylinder

Code: Select all

OS: Linux Mint 20.3 (MATE/mate)
Word size of FreeCAD: 64-bit
Version: 0.20.27008 (Git) AppImage
Build type: Release
Branch: (HEAD detached at 9162733)
Hash: 91627338a44371fe9a8e815b4900f62558af98a1
Python version: 3.9.9
Qt version: 5.12.9
Coin version: 4.0.0
OCC version: 7.5.3
Locale: French/France (fr_FR)