Hello everybody,

I am working on a relatively simple soft body-project to deepen my understanding of vellum. The setup is easy: take a mesh, cut it in pieces and then each frame turn one of those pieces to a soft body and delete it from the static collider (so eventually the whole mesh turns to soft bodies).
In the DOP-net the collider is set to animate and the vellum source is set to emit continuously. I even scaled down the vellum-pieces to prevent them from clipping into the collider. But even with this and lots of substeps the solver fails to produce correct results. I tried tweaking every parameter but cant get it to work… has anybody an idea how I can fix this? The collision geo should be correct, it is concave and I double checked collision geo, guide geo etc.

Statue from: http://threedscans.com/lincoln/loewental/ [threedscans.com]

Thank you for your help!

So far I spent a lot of time trying out different configurations of the vellum-solver and tried playing with the substeps of the dop-net but the problem is still there. I really don't know that else I can try and would appreciate a small pointer towards the right direction.

This happens if I increase my dop-supsteps, some pieces get launched into the air. Usually I would expect this behaviour if the pieces collide with the inside of the static geo. But this can't be the case because I shrank them down a little bit to prevent exactly this from happening. The subdivs of my static geo and the proxy geo should also be high enough to prevent any clipping from happening… :/

Hello
I opened your scene and what i think is :as you set the emission type to continuous on the Vellum source
you have one emission not every frame but every substep ! so you get 3 time each geo in the same place
that's why it get messy !
setting the substep to 1 on the Vellum Solver looks allready better :

My advice ; find a way to emit only once for each frame !

I hope it helped

Benjamin