V0.20.2 Curves WB issues, Macros & Silk WB

Need help, or want to share a macro? Post here!
Forum rules
Be nice to others! Respect the FreeCAD code of conduct!
LQ_6
Posts: 6
Joined: Tue Mar 28, 2023 4:04 pm

V0.20.2 Curves WB issues, Macros & Silk WB

Post by LQ_6 »

I was working on a model with Curves WB, days prior I´d updated to V0.20.2 after updating to newest version all macros I have started to stop working, the Trim Tool and Geominfo mesh from Curves WB stopped being parametric and the worst of them all I started having issues when mirroring surfaces, even though my curves are tangent.

I was chatting with a fellow redditor named gnosys about these issues and he both suggested me use Silk also he used it to finish the pipe I was attempting to model and kindly sent it back to me as a demo

My main porpuse is to model a continous F1 like car (G3 continuity if possible in the nose, side pods, engine cover and halo if possible) for my bachelor´s thesis with freecad due to the support of Open Source sowtware is included in thesis which is about aerodynamics.

I want G3 Continuity because I belive is the only way the engineering school will take a look to freecad since neither product design nor NURBS is taught in the carrer´s program, so if freecad can produce g3 surfaces then is more than enough for mechanical design

The F1 car was modeled around 2021 when I didnt know how to properly use freecad and wasn´t concer about topo issue for e.g, last year due to personal issues I was not able to continue the model nor the thesis but know that I know how freecad behaves and how to work with it got these issues
Attachments
F1 no continuity zebra.JPG
F1 no continuity zebra.JPG (65.38 KiB) Viewed 785 times
Mirror no tangent.JPG
Mirror no tangent.JPG (67.44 KiB) Viewed 785 times
Pipa_silk_demo.FCStd
(116.59 KiB) Downloaded 8 times
Pipa.FCStd
(317.84 KiB) Downloaded 9 times
arc tangency.JPG
arc tangency.JPG (63.45 KiB) Viewed 785 times
emills2
Posts: 868
Joined: Tue Apr 28, 2015 11:23 pm

Re: V0.20.2 Curves WB issues, Macros & Silk WB

Post by emills2 »

if the entire model is too big to load, could you isolate the mirrored shape + the curves used to generate it, and upload a file containing those elements? please also include any sketches used to generate the curves (if they exist).

I believe i see a way to do this part with Silk, and i think the fill surface of the Surface workbench could be used as well. of course Curves may have a valid alternative construction as well. For silk, the big question will be whether your base curves can be replicated with the limited number of control points/degree currently allowed by silk curves.

The problem may also be simply within the mirror tool. although more stable than offset, mirror can choke on complex surfaces. your topology is fairly simple, but some tools create surfaces that are very complex internally, even is the result looks very simple (surface filling does that for example). this can be very nice from a user perspective, because you don't need to worry about what the tool is doing, and you get great results, until things break.

[edit: i see the files now, not sure if i missed them the first time reading]
Last edited by emills2 on Tue Mar 28, 2023 7:14 pm, edited 1 time in total.
emills2
Posts: 868
Joined: Tue Apr 28, 2015 11:23 pm

Re: V0.20.2 Curves WB issues, Macros & Silk WB

Post by emills2 »

something i see right off the bat is that the edges of your combs do not line up (not parallel to each other, and not on the same line). this would indicate that the curves are not parallel to each other at the end points. if using a mirror plane, you likely want your curves to be perpendicular to the mirror plane.

maybe the comb is using diffrent endpoints than what i interpret from the picture though.
Last edited by emills2 on Tue Mar 28, 2023 7:16 pm, edited 1 time in total.
LQ_6
Posts: 6
Joined: Tue Mar 28, 2023 4:04 pm

Re: V0.20.2 Curves WB issues, Macros & Silk WB

Post by LQ_6 »

emills2 wrote: Tue Mar 28, 2023 6:56 pm if the entire model is too big to load, could you isolate the mirrored shape and the curves used to generate it and upload a file containing those elements? please also any sketches used to generate the curves (if possible).

I believe i see a way to do this part with Silk, and i think the fill surface of the Surface workbench could be used as well. of course Curves may have a valid alternative construction as well. For silk, the big question will be whether your base curves can be replicated with the limited number of control points/degree currently allowed by silk curves.

The problem may also be simply within the mirror tool. although more stable than offset, mirror can choke on complex surfaces. your topology is fairly simple, but some tools create surfaces that are very complex internally, even is the result looks very simple (surface filling does that for example). this can be very nice from a user perspective, because you don't need to worry about what the tool is doing, and you get great results, until things break.
Don´t know if is the correct way to answer but I didnt know how how posting would work in the original comment there are 2 files "pipa " is the one that failed and the other one was done by gnosys_ with silk

As far as I know we need 8 control points for a G3 but if it only can do G2 it still good enough, I could say in the thesis that with more support it could probably do G3

Didnt know trim could fail that easily
LQ_6
Posts: 6
Joined: Tue Mar 28, 2023 4:04 pm

Re: V0.20.2 Curves WB issues, Macros & Silk WB

Post by LQ_6 »

emills2 wrote: Tue Mar 28, 2023 7:13 pm something i see right off the bat is that the edges of your combs do not line up (not parallel to each other, and not on the same line). this would indicate that the curves are not parallel to each other at the end points. if using a mirror plane, you likely want your curves to be perpendicular to the mirror plane.

maybe the comb is using diffrent endpoints than what i interpret from the picture though.
You are right, I tought it was the view but weird thing is that they are circles arcs but the middle one
emills2
Posts: 868
Joined: Tue Apr 28, 2015 11:23 pm

Re: V0.20.2 Curves WB issues, Macros & Silk WB

Post by emills2 »

LQ_6 wrote: Tue Mar 28, 2023 7:16 pm As far as I know we need 8 control points for a G3 but if it only can do G2 it still good enough, I could say in the thesis that with more support it could probably do G3
you might find the conversation starting here viewtopic.php?p=223101#p223101 interesting on topic of Gs and CPs
emills2
Posts: 868
Joined: Tue Apr 28, 2015 11:23 pm

Re: V0.20.2 Curves WB issues, Macros & Silk WB

Post by emills2 »

see also here viewtopic.php?p=234844#p234844 to see "current best" for Silk and when i switch over to Surface workbanch for final touches. talks about trade offs for continuity along patch edges versus within patches, shows zebras, etc.
LQ_6
Posts: 6
Joined: Tue Mar 28, 2023 4:04 pm

Re: V0.20.2 Curves WB issues, Macros & Silk WB

Post by LQ_6 »

LQ_6 wrote: Tue Mar 28, 2023 7:23 pm
emills2 wrote: Tue Mar 28, 2023 7:13 pm something i see right off the bat is that the edges of your combs do not line up (not parallel to each other, and not on the same line). this would indicate that the curves are not parallel to each other at the end points. if using a mirror plane, you likely want your curves to be perpendicular to the mirror plane.

maybe the comb is using diffrent endpoints than what i interpret from the picture though.
You are right, I tought it was the view but weird thing is that they are circles arcs but the middle one
Nop T_T I corrected the arc tangency, zebra analysis still breaks
emills2
Posts: 868
Joined: Tue Apr 28, 2015 11:23 pm

Re: V0.20.2 Curves WB issues, Macros & Silk WB

Post by emills2 »

LQ_6 wrote: Tue Mar 28, 2023 7:16 pm Didnt know trim could fail that easily
unless a trim curve falls exactly on an isoline (u or v), it is an approximation composed of many many segments. the workload increases in difficulty the more you trim arbitrarily. at some point you have to remap uv and re-approximate the surface to clean it up after too many operations.

right now Silk only 'trims' along isos, and i call them segments, to differentiate. a segment of a surface is at most as complex as the source surface, and can become simpler. this simplification rule is the driver for the current limitation. it actually has a lot in common with Class A design strategies...but obviously is not yet a full Class A system. you can trim Silk surfaces with FreeCAD, but the trimmed edges may not be directly usable in Silk for futher work. Although Silk cannot use any arbitrary curves to work with, it is designed to be attached to "standard" CAD surfaces like planes, cylinders, spheres. so it's not completely isolated, but it's very specific in what it interacts with.
emills2
Posts: 868
Joined: Tue Apr 28, 2015 11:23 pm

Re: V0.20.2 Curves WB issues, Macros & Silk WB

Post by emills2 »

i could do some silk stuff here for the pipe, but i don't think it fits best here. this is a great candidate for the Surface Filling tool.

however, for other things like your vehicle fairing, Silk might be very appropriate.

i'm at work right now so i probably can't get back to you until tomorrow.
Post Reply