MoI discussion forum
MoI discussion forum

Full Version: XNurbs for MoI

Show messages: All  1-2  3-16

From: Michael Gibson
1 Aug 2024   [#3] In reply to [#1]
Probably the way with the least amount of work overall would be that I'd do all the XNurbs/MOI integration work myself (because it's a lot of work to make a full C++ SDK, I can make a specialized one with just what is needed for XNurbs a lot more easily) but the resulting plug-in would be sold and distributed only by the XNurbs team as they do with Rhino and SolidWorks plug-ins currently.

That way there doesn't have to be any IP licensing set up which greatly simplifies things.

For timing probably near the start of the MOI v6 cycle would work because I'm heavily occupied with ACIS integration work for trying to wrap up v5.

Thanks, - Michael
From: MajorGrubert (CARLOSFERREIRAPINTO)
2 Aug 2024   [#4]
Hello,
Excellent idea! Thanks PaQ!
I'll buy the plugin.
Thanks Michael for the interest.
From: Billabong
2 Aug 2024   [#5]
@Michael,

I'm sure this is a really dumb question, but is it not viable for you to create your own xnurbs instead of relying on a third party?
-B
From: Michael Gibson
2 Aug 2024   [#6] In reply to [#5]
Hi Billabong,

re:
> I'm sure this is a really dumb question, but is it not viable for you to create
> your own xnurbs instead of relying on a third party?

It could be possible but that type of surface fitting is a finicky area of work. It tends to be difficult to avoid oscillations.

It's hard to accurately estimate how much work would be involved, my initial guess is maybe 5 to 10 years?

- Michael
From: Billabong
2 Aug 2024   [#7] In reply to [#6]
well that pretty much cancels that idea, LOL
From: Psygorn (DRILLBIT)
2 Aug 2024   [#8] In reply to [#3]
Hi Michael,

I would like to know your anticipation on how far would it be? - in terms of years-

-Psygorn
From: Michael Gibson
2 Aug 2024   [#9] In reply to [#8]
Hi Psygorn, probably sometime next year. But that depends on whether that approach is accepted by XNurbs, I have not discussed it directly with them as of yet.

- Michael
From: Psygorn (DRILLBIT)
2 Aug 2024   [#10] In reply to [#9]
got it :-)
From: PaQ
2 Aug 2024   [#11]
That's really encouraging, thanks for your input Michael !
From: MO (MO_TE)
6 Aug 2024   [#12]
It's really nice to see some new surfacing tools, either x-nurbs plugin or implementing ACIS kernel into network and n-sided commands to add more features to them.
While it is not completely related to the x-nurbs discussion, I was thinking, Is it possible to make a plugin to connect moi to freecad library to use its surface filling tool inside moi?

I know it's not as powerful as x-nurbs, But, It has these features:
Making a surface with boundary curves/edges
Accepting guide curves/points
G1/G2 continuity

https://wiki.freecad.org/Surface_Filling
From: Michael Gibson
6 Aug 2024   [#13] In reply to [#12]
Hi Mo, that could be a possibility but I haven't heard much about how well it behaves.

Have you used this function in FreeCAD very much?

- Michael
From: MO (MO_TE)
7 Aug 2024   [#14] In reply to [#13]
Hi Michael
Actually, I haven't used it that much except for a couple simple surfaces. But here is a test for more complex situations.
To be fair, I tried to use these videos as reference to compare freecad and x-nurbs results. ( I just quickly glanced at the curves in the videos, so it's not exactly the same. )
https://www.youtube.com/watch?v=mGbYDIqt_MM
https://www.youtube.com/watch?v=a1Himxz1C-Q

As already known, using freecad's surface filling tool was challenging. You can't just select all the boundaries at once and get the x-nurbs magical results.
I tried to minimize the patch area to make it simpler for freecad. You can see the results in the attached file.

Results are not perfect for sure and I totally understand if you don't want to put your time on it. :)

Attachments:
Freecad_surfaceFillingTool.zip


From: Michael Gibson
7 Aug 2024   [#15] In reply to [#14]
Hi Mo, thanks for the tests of the Freecad surface filling module.

But yeah it does not look promising, it looks like it is afflicted by oscillations which is often a complication with this type of surface fitting calculation.

You can see them particularly in this surface:







The control points kind of wildly spike back and forth to opposite sides of the surface in a small area. This makes the surface have a kind of subtle bumpy shaping to it in those spots.

Some of the other surfaces in there are not so bad as this one but it does not really bode well for being very reliable.

There are kind of opposing forces between maintaining surface smoothness versus accuracy of forcing the surface to pass through curves. It is quite tricky to balance between these.

There will often be either oscillations or inaccurate results that aren't close enough to join up with the existing edges. Or both.

- Michael

Image Attachments:
freecad_filling1.png  freecad_filling2.png  freecad_filling3.png 


From: Mauro (M-DYNAMICS)
12 Aug 2024   [#16]
If so, make cars will be more easy here.
You could recover all users that migrate to Plasticity because of Xnurbs.

M

Show messages: All  1-2  3-16