[Plugin][$] Curviloft 2.0a - 31 Mar 24 (Loft & Skinning)

AcesHigh · 13 Jun 2016, 14:37

well, ok, so I deleted LibFredo from the AppData folder, leaving the only version installed at ProgramData.

Now all Fredo6 plugins bugged. Except Curviloft, which now is working...

It seems I need to do the inverse: install all Fredo plugins, including Curviloft, at AppData.

EDIT

Ok, it turns out I can´t install at the default plugins folder. no permissions. Following all the instructions available here at the site did not help. I just can´t change the permissions for the folder. Options greyed-out. Tried to take ownership. No luck.

So I installed two versions of LibFredo6. One using the automatic Sketchucation Store installer at Program Data folder.

The other I installed MANUALLY at the main plugins folder, using RBZ extension, so previous Fredo plugins installed there would work.

Now everything is working again and I do not get the Error Message.

Solved? NO!!!

JUst can´t find Curviloft anywhere. No toolbar (nor toolbar available at "view/toolbars"). Did not find it either at Plugins or the Tools drop down window...

EDIT:

solved, installing Curviloft MANUALLY (RBZ) worked.

robertWan · 16 Jun 2016, 21:44

What am I doing wrong?
At the beginning of the film is not formed skin.
http://www.screencast.com/t/aZEj6Qsq
SU8, Win7.

Robert

pilou · 16 Jun 2016, 23:33

Just add a 4 th curve somewhere!

jaaphu · 13 Oct 2016, 11:43

Great Thanks!

surfmike · 12 Dec 2016, 09:34

I'm a big fan of Curviloft.

However I ran into an issue...

These contours seem impossible to skin. Maybe it's to many segments (will use it for 3d printing)...
The left one just seems to stop.
The right one (with a small added triangle) starts, but doesn't go all the way.
File added.
//Mike

test skinning

pilou · 12 Dec 2016, 11:37

My advice
Your curves are not clean! (segments added to extremities, overwrited itselves etc...)
I redraw above with 4 FSPlines From Bezier Curves by Fredo
(6 general Points each if you want many patches)
= no Problem!
Just be carreful to the snap points!
Surfaces are made in half second with CurviLoft by Fredo!

Free Visualizer as render (till SU 2016)

surfmike · 12 Dec 2016, 13:06

I believe I was really careful with those mentined issues...
However with my file it takes a long time, goes to 90%, then it seems to stop.
With lightweight curves it works.
Did you use high resolution aswell (300s)?
Mike

pilou · 12 Dec 2016, 13:32

Resolution has quasi nothing to do with curviloft!
it's only matrix of vertices and surface resulting!

Problem are more with no natural form curves!
Corners, interruptions, no fluidity etc...

If your time result of Curviloft is too long and not reasonable,
that is that your curves are false for any reasons!

Redraw above it! (Group the old with a little line beside for an easy selection)
As like you can redraw above the old curves without effort of selection at the end!

surfmike · 13 Dec 2016, 21:55

@pilou said:

Resolution has quasi nothing to do with curviloft!
it's only matrix of vertices and surface resulting!

[attachment=0:syi8gtsw]GIF_line.gif[/attachment:syi8gtsw]

Pilou, thanks for your assistance and informative help.
In a hurry "resolution" was the word I found for expressing the amount of segments/verticies.

However - the amount of segments in a curve is limited to 300(s). While modelling for 3D printing, is there a possibility to override that limitaton?
//Mike

surfmike · 16 Dec 2016, 19:57

Hi.
The model consists of two components.
I've done all curves with "Classic bezier curve", and all are "infinite curves" following the green/red/blue axis when placing control points.
Still I got some kind of "V shape" in the front curved panel, see picture and file (concerning file size I had to remove the mesh).

How can I avoid this? I tried to elaborate by changing settings but with no luck.
The person solving this will be a Nobel prize prospect...

Thx!
//Mikael

Front 1.5 fix SUC.skp

pilou · 17 Dec 2016, 00:07

If you want make 3D Print maybe export in OBJ or STL and use MeshMixer! (it's free)
You will obtain objects more smoothed!

surfmike · 17 Dec 2016, 08:39

Yes, I will export to STL and use Meshmixer for scaling and hollow purposes, but I think Meshmixer can't handle this kind of artifacts. Also this surface will be a part of a bigger solid.
All three curves are perfect beziers ending 90° to each other, which preferably wouldn't give this minor V shape.
Is there any settings or any way to avoid this "V"?
Thx
//Mikael

jgb · 17 Dec 2016, 19:16

From the image posted, it looks like you have another mesh inside this outer mesh that is poking through, and you are trying to fix the outer mesh instead.

I can't say for sure as your SKP file is SU-2016 and I am using SU-2015 which cannot load your file. If you can post the file in 2015 format I will look at it closely.

gilles · 17 Dec 2016, 20:40

[a
Hi.
The model consists of two components.
I've done all curves with "Classic bezier curve", and all are "infinite curves" following the green/red/blue axis when placing control points.
Still I got some kind of "V shape" in the front curved panel, see picture and file (concerning file size I had to remove the mesh).

How can I avoid this? I tried to elaborate by changing settings but with no luck.
The person solving this will be a Nobel prize prospect...

Thx!
//Mikael

[attachment=1:ni34c5df]Front 1.5 fix SUC.skp[/attachment:ni34c5df][/quote]

[attachment=0:ni34c5df]tangent.gif[/attachment:ni34c5df][quote="surfmike"][attachment=2:ni34c5df]Screen Shot 2016-12-16 at 20.44.08.png[/attachment:ni34c5df]

Give me the Nobel (don't forget the$ ).

pilou · 17 Dec 2016, 23:18

Sorry i have forgotten the $$$ !

surfmike · 18 Dec 2016, 09:46

@jgb said:

From the image posted, it looks like you have another mesh inside this outer mesh that is poking through, and you are trying to fix the outer mesh instead.

I can't say for sure as your SKP file is SU-2016 and I am using SU-2015 which cannot load your file. If you can post the file in 2015 format I will look at it closely.

Thanks, a SU2015 file is attached.

@gilles said:

[a
Give me the Nobel (don't forget the$ ).

I had exactly the same idea in mind, but it ended up with two "V's" with a flat surface inbetween...
If you look real close in your animation/file you will see the shadow in the upper area telling there's still a "V", right?
In SU when using "soft" and "smooth" it all looks good, but when aiming towards 3D printing soft/smooth just hiding artefacts.
I placed a surface in the middle of the "problem area" to show it more clearly, see attached pics.

Thanks all!
//Mikael

Front 1.5 fix SUC (SU2015).skp

jgb · 18 Dec 2016, 19:12

I looked at your model; that's all I could do, is look at it.
One half of the model is composed of almost 2,000 UNCONNECTED very short edges. I could not even weld them into a curve. How you got a surface from them is a bit of a mystery.

However, the "V" you got is no mystery, as this happens to me where I flip a curved surface to form a mirrored other half. It also happens every time I create a large complex surface in several parts. I have to then massage the joins smooth.

The finite lines that form the curve at the "join" are not perfectly aligned in parallel to each other, and when you do the flip there is a small angle to the curve at the join, that appears as a ridge or "V" to the visual rendering. That angle at the "V" is a bit larger than the angles between the lines elsewhere on the curve, and as a result look larger.

To correct that, there are a few ways to do it.

If it was only for a visualization, then making the join lines between the 2 halves smooth and soft should make the join look smooth.

But you want to 3D Print this, so that probably will do nothing for the physical printed part. The "V" will form a bit of a ridge.

What you need to do is delete the surface triangles that form the join line, leaving a gap. But delete the end lines last. Why? Later down.
Explode both halves, then draw a line at the top and bottom corners, joining the 2 halves. You now have a valid curved perimeter for Curviloft to fill.
Explode the Curviloft group, and the "V" should disappear.

If the join is now too flat, then undo the Curviloft surface and the 2 lines you drew. Then undo the deletion of the 2 end lines that formed the "V". (That's why!)

Draw a line mid angled to the original lines and a line joining the 2 halves, then Curviloft that perimeter. This flattens the "V" but not quite flat, so the 3D printed part should not have a ridge.

Sorry, but I do not have the time to draw and annotate the fix.

surfmike · 19 Dec 2016, 11:42

JGB,
Thanks for helping out, I fully understand your instructions.
However, when I color the line by axis it shows beziers with parallell red end lines (7,2mm on above, 0,045mm in the below line) in each half.
Maybe 0,045mm is too short to be recognized correctly, but in the final skinning group there's parallell lines alright even though 0,045mm became 1,06mm (see attached pic).
But the issue is that inbetween those perfectly parallell lines the skin gets a "V" shape... (???)
//Mikael

P.S The unconnected edges was a unpleasant result of copy/paste 300s beziers from SU2016 to 2015. D.S

gilles · 19 Dec 2016, 12:43

@surfmike said:

JGB,
Thanks for helping out, I fully understand your instructions.
However, when I color the line by axis it shows beziers with parallell red end lines (7,2mm on above, 0,045mm in the below line) in each half.
Maybe 0,045mm is too short to be recognized correctly, but in the final skinning group there's parallell lines alright even though 0,045mm became 1,06mm (see attached pic).
But the issue is that inbetween those perfectly parallell lines the skin gets a "V" shape... (???)
//Mikael

[attachment=0:vpj5k8ov]red_line.png[/attachment:vpj5k8ov]

P.S The unconnected edges was a unpleasant result of copy/paste 300s beziers from SU2016 to 2015. D.S

You don't need to copy/paste from SU2016 to 2015, just Save as and choose the SU version.

Also when you work at this scale make a component, copy/paste this component, scale the copy X100 or X1000 and work with the copy, 0,045mm is far from the tolerance of SU to create faces.

jgb · 19 Dec 2016, 16:29

Maybe 0,045mm is too short to be recognized correctly,

Yes, that is a problem. I noticed it but forgot to say so in my reply. SU really does not like very short lines.

To get around this scale up the whole object 10 times and do the required functions on that, then scale back to 10% to return to original scale.

SU can display short lines, but has some difficulty working with them.

As Gilles said, save the SU-16 version as SU-15 and re-post it so I can look at it again.