Curves workbench

[fc_tofu]

… sorry fc_tofu, no retrieval. Too long ago.
Today I would do it in another way - and I have "only" FC.18.4.

Thank you.
I find @Chris_G's latest example of blend surface. "blendsurface.FCStd"
https://forum.freecadweb.org/download/f ... ?id=104829
My understanding is that,
The right tool to convert blender suface network into surface is Gordon surface, not Aproximation.

[fc_tofu]

No, unfortunately I have not found a possibility but I have asked several times that we can make Gordon surfaces tangential to neighbouring surfaces.

You can call me Tofu, and many thanks for your explanation.
So basically I know, G1 Tide bottle is impossilbe with CurvesWB for now.
But I am still not very clear about the clause in your translated sentence.
"we can make Gordon surfaces tangential to neighbouring surfaces", is this a fact or a question?

[freecad-heini-1]

"we can make Gordon surfaces tangential to neighbouring surfaces", is this a fact or a question?

It's a translation with:
https://www.deepl.com/en/translator

[microelly2]

Hi, heini,
have you found a workaround to get a G1 surface Tide bottole?

Where I can get this model?
I want to see the reason why there is this trench.
clamped curves and surfaces must be smooth. https://pages.mtu.edu/~shene/COURSES/cs ... -type.html
I think there is a hard border which can be eliminated.

[fc_tofu]

Where I can get this model?

Discussion:
https://forum.freecadweb.org/viewtopic ... &start=100

https://forum.freecadweb.org/viewtopic. ... 62#p204162

Sample file:
https://www.dropbox.com/s/5d4tbbjqalvwy ... fcstd?dl=0

https://forum.freecadweb.org/download/file.php?id=49055

This topic was posted in 2017, when @Chris_G told this is a limiation.
I am just curious, have some workaround been found during the last 3 years.

[microelly2]

https://forum.freecadweb.org/download/file.php?id=49055

This topic was posted in 2017, when @Chris_G told this is a limiation.
I am just curious, have some workaround been found during the last 3 years.

To get a smooth border the tangents must be perpendicular to the mirror plane (yz).
I do this by given the 2nd and last but ones poles row the same y values as the first and last rows have.
The change is minimal but surfaceand mirror are now tangent.

dp_04_23_001.png (208.16 KiB) Viewed 1762 times

here the model and the script

tide_bottle3_mod.fcstd

(106.34 KiB) Downloaded 66 times

Code: Select all

f=App.ActiveDocument.Approximastion_Curve.Shape.Face1
c=f.Surface

poles=c.getPoles()
uk = c.getUKnots()
mu = c.getUMultiplicities()
ud = c.UDegree
uperiodic = c.isUPeriodic
vk = c.getVKnots()
mv = c.getVMultiplicities()
vd = c.VDegree
vperiodic = c.isVPeriodic
poles=c.getPoles()

polesa=np.array(poles)
polesb=polesa.swapaxes(0,1)

# polesb=np.round(polesb,1)

polesb[1,:,1]=polesb[0,:,1]
polesb[-2,:,1]=polesb[-1,:,1]


polesc=polesb.swapaxes(0,1)
 
sf=Part.BSplineSurface()
sf.buildFromPolesMultsKnots(polesc,mu,mv,uk,vk,False,False,ud,vd)

a=FreeCAD.ActiveDocument.getObject("ShapeA")
if a is None:
	a=FreeCAD.ActiveDocument.addObject("Part::Feature","ShapeA")

a.Shape=sf.toShape()

[fc_tofu]

To get a smooth border the tangents must be perpendicular to the mirror plane (yz).
I do this by given the 2nd and last but ones poles row the same y values as the first and last rows have.
The change is minimal but surfaceand mirror are now tangent.

I donn't know coding, but I know this is a cheerful workaround and deserves a celebration.
Thank you.

[microelly2]

To get a smooth border the tangents must be perpendicular to the mirror plane (yz).
I do this by given the 2nd and last but ones poles row the same y values as the first and last rows have.
The change is minimal but surfaceand mirror are now tangent.

I donn't know coding, but I know this is a cheerful workaround and deserves a celebration.
Thank you.

You can do it by design of the rib sketches without coding: the endings of the arcs must be perpendicular to the yz plane.

[fc_tofu]

You can do it by design of the rib sketches without coding: the endings of the arcs must be perpendicular to the yz plane.

In your way, I can got a G1 mirrored surface now. Thank you!

Another question. If you check mirrored surface in "Object Property >View > Lighting = One Side", you will find Mirror command (both PartWB and DraftWB) donnot correct surface normal.
So, after Boolean Union, we finally get a polysurface with adjacent reversed normal. But it looks weird.

fsc_2020-04-24_183014.jpg (20.63 KiB) Viewed 1661 times

In this case, PartWB >Reverse shape can be used before Union command to get consistent normals between adjacent faces.

But "Reverse shape" command is crappy for now, it can only handle objects which LCS loates at world origin.
https://forum.freecadweb.org/viewtopic. ... 78#p384478

So, just be careful.

[microelly2]

The better way to create such volumes (tubes) is to see them as variant of a torus.
ribs are closed and the surface is cyclic in both directions.See an example here.
https://youtu.be/sWlaE4i2MAQ
There is no need of mirror and normalizing faces.