3D Surface - Face Selection (New feature) [Merged]

[-alex-]

5:03PM:

I think we would only want/expect a path generated in the front left rectangular region of the selected yellow face. Correct?

5:24PM:

Fixed.

Well done
And thanks for such improvements and great features

[freman]

Nice work, that looks more sensible.

The diamond-circle-square test should be interesting.

[Russ4262]

Here is a stress test model I submitted to sliptonic's CNC-torture collection.

Russ, you may like to throw a 3D over that and look at accuracy where the squares and circles touch. This brought out a bug in one of the 2.5D tools which did not handle a zero dimension correctly.

There should be an unbroken vertical line from the corners of the top square down to the base, touching both cuboids and the cylinder.

Attached image is at 0.015mm sample interval and 10% step over on a 3.175mm endmill - a fairly moderate resolution.

Russ

[freman]

Thanks, that looks nice. What's happening on the LHS of the base there?

Is it cutting the whole height of that face in one cut or not machining it at all?

This obviously not the ideal tool to cut this part, it would be there all day to get a good resolution finish. But it's an interesting exercise to see how it handles it.

[RatonLaveur]

Thanks, that looks nice. What's happening on the LHS of the base there?

Is it cutting the whole height of that face in one cut or not machining it at all?

This obviously not the ideal tool to cut this part, it would be there all day to get a good resolution finish. But it's an interesting exercise to see how it handles it.

LHS: the cutting direction being fully parallel to the edge there, the drop-cutter algorithm cannot distinguish it. Similar to waterline being unable to detect a horizontal plane. So technically yes it's "cutting the whole face at once" because the pattern encompasses the edge.
It's indeed not the right tool for that part. But good way to test the algo. If you want a nice finish on this part waterline it.

[freman]

Thanks, this is what is thought was going on but I wanted to confirm it.

So that path would simply break the tool. There is also the question of tool length and machine clearance.
Maybe there needs to be a control on cutting depth.

Similar to waterline being unable to detect a horizontal plane.

But waterline will correctly cut across a horizontal plane at a suitable speed and will clear the surface won't it ?

This case is pretty obvious to the trained eye since it is a large area with no paths and it is visible. I wonder about what happens on a less expansive face maybe not in plain sight. Perhaps the path tool should be throwing a warning about this kind of condition.

Even better, I think Russ showed that 3Dsurface can now do vertical planes and non planar faces, maybe this can be integrated if this condition is detected.

BTW this is not a "real" part, it is an ISO standard test piece to detect machine accuracy and repeatability. How about trying one on your laser mill

It turns out to be quite good for testing CNC path generating software too.

If you want a nice finish on this part waterline it.

Yes, that would be another good test of the algo.

[freman]

Russ, what do I need on top of current master to test this?

[Russ4262]

Evening Freman,

Russ, what do I need on top of current master to test this?

I just updated the script in the OP of this thread; I also added installation instructions/options for Windows users and those building from source.

Cheers,
Russ

[freman]

Thanks Russ, I'll try to find time to build it this w/e. It's come a long way since last time I used it.

[Gauthier]

Hi Russ,

Many thanks for your great job.
It seems to work fine for simple surfaces, but for my complex geometry that I need to mill, it don't work at all

Surface3D-v0.19-Exception.png (235.38 KiB) Viewed 892 times

This is the output in report view :

Code: Select all

<Exception> shapes are not planar
Traceback (most recent call last):
  File "/home/gauthier/src/FreeCAD/build/Mod/Path/PathScripts/PathUtils.py", line 59, in new_function
    res = function(*args, **kwargs)
  File "/home/gauthier/src/FreeCAD/build/Mod/Path/PathScripts/PathOp.py", line 513, in execute
    result = self.opExecute(obj) # pylint: disable=assignment-from-no-return
  File "/home/gauthier/src/FreeCAD/build/Mod/Path/PathScripts/PathSurface.py", line 324, in opExecute
    (faceShapesList, voidShapesList) = self._preProcessModel(JOB, obj)
  File "/home/gauthier/src/FreeCAD/build/Mod/Path/PathScripts/PathSurface.py", line 540, in _preProcessModel
    clsdAreas = self._extractFaceLoopFeatures(base, shape)  # returns all closed loops
  File "/home/gauthier/src/FreeCAD/build/Mod/Path/PathScripts/PathSurface.py", line 860, in _extractFaceLoopFeatures
    fcShp = self._wireToFlatFaceShape(fc, wire)  # complete object
  File "/home/gauthier/src/FreeCAD/build/Mod/Path/PathScripts/PathSurface.py", line 899, in _wireToFlatFaceShape
    env = PathUtils.getEnvelope(NW.Shape, depthparams=self.depthParams)
  File "/home/gauthier/src/FreeCAD/build/Mod/Path/PathScripts/PathUtils.py", line 313, in getEnvelope
    sec = area.makeSections(heights=[0.0], project=True)[0].getShape()
<class 'ValueError'>: {'sclassname': 'N4Base10ValueErrorE', 'sErrMsg': 'shapes are not planar', 'sfile': '', 'iline': 0, 'sfunction': '', 'swhat': 'shapes are not planar', 'btranslatable': False, 'breported': True}
6.9e-08 <App> Document.cpp(3462): Failed to recompute MouleCabane3#Surface: {'sclassname': 'N4Base10ValueErrorE', 'sErrMsg': 'shapes are not planar', 'sfile': '', 'iline': 0, 'sfunction': '', 'swhat': 'shapes are not planar', 'btranslatable': False, 'breported': True}
Recompute failed! Please check report view.

The old full solid selection worked in both v0.18 & v0.19 with little bugshttps://forum.freecadweb.org/viewtopic.php?f=15&t=43310, but worked :

Capture d’écran_2020-02-15_16-27-18.png (312.29 KiB) Viewed 892 times

Perhaps it will be possible to keep both 3D_Solid (the old 3D surface algorithme) and your new surface algo ?


OS: Debian GNU/Linux bullseye/sid (XFCE/xfce)
Word size of OS: 64-bit
Word size of FreeCAD: 64-bit
Version: 0.19.19535 (Git)
Build type: Release
Branch: master
Hash: 8deed98acec8f31ec946c95d0aa0c2767223baa3
Python version: 3.7.5
Qt version: 5.12.5
Coin version: 4.0.0a
OCC version: 7.3.0
Locale: French/France (fr_FR)

- Path/PathScripts/3DSurface : Version 9d

The model is too big to be uploaded in this forum, you can download it on my Google drive : https://drive.google.com/file/d/1V_2wbv ... sp=sharing

@++;
Gauthier.