Vellum strut setup

[Jesper Rahlff]

Hi everyone.

can anyone tell me how to make strut constraints in vellum work with a thin object?

 

Attached is a hip file. Try switching between the thick and the thin tube and you will see that the thin tube collapses while the thick one behaves as one would expect.

 

Thanks

odforce_example.hiplc

Edited May 1, 2019 by Jesper Rahlff

[j00ey]

I'm not a vellum expert but if you turn up the bend stiffness on the cloth constraint [as opposed to the struts] to say 50, the behaviour is similar 

[Atom]

Thinner objects need more iterations. On the solver set Substeps to 2 and Constraint Iterations to 4,000. Size is a factor in Vellum. That number of constraint iterations is ridiculous, of course, but it does give a similar response to the thicker tube. So you might want to build everything a little bigger, from the start, then scale it back down to match the scene at the end. Or play around with adding more subdivisions to the tubes, rather than relying on those very long primitives.

Edited May 1, 2019 by Atom

[Jesper Rahlff]

Thanks so much to both of you. I have some stuff to play around with.

As for the number of constraints you are suggesting @Atom, I agree. That value seems totally out of proportion.

[Zybrand]

I think there are a couple of things going on the effects your sim results. Because you thin shape is so much thinner that the thick shape your point thickness changes which will effect the sim, if you visualize thickness in the sim you will see that changing between sims because of overlap. 

Then the bigger difference between the width and height of the primitives of the thin tube will make it bend easier and that has more to do with the cloth settings than the struts. Increasing the stiffness and bend stiffness a lot on the cloth constraints help as much if not more that the struts. I also upped the strut constraints stiffness a bit.

Making those changes gives you good results while only having to set the solver substeps to 2 and the constraint iterations to 200. So you can avoid 4000 iterations

odforce_example_zj.hiplc

[Atom]

That does work for the thinner tube, but when you change the switch to compare the two sizes the results are still very different. Can you think of a way to make that part work? So the bend experienced is approximately the same between two different sizes?

 

[Jesper Rahlff]

using Zybrands example, if you increase the bend stiffness on the cloth from 10 to 100 you get a much closer result between the thick and the think tube

odforce_example_jr.hiplc

Edited May 1, 2019 by Jesper Rahlff

[Zybrand]

I didn't consider having both the thin and thick tube looking the same with the same setting and just switching. I expect them to need different setting to behave in a similar way because of their shape differences and the point thickness changing between the two shape, if they were behaving the same I would be disappointed in the vellum solver. 

If you want to switch between them without adjusting cloth and strut settings your best bet is probably to multiply the stiffness parameters with attributes on the geometry. Maybe calculate the the area of the shapes or something similar and multiply the stiffness by that.