Particles Collision Inside Deforming Geo.. Problem

[3dFrosty]

Hello,, I'm hoping someone can really help me out here..

I've been doing alot of tests with particles birthing inside an object which are when bound within it bouncing around like particles do.. My initial tests worked, but for some reason the collision based on deforming geometry seem to be problematic to say the least..

The static object and the translating object work fine, but I need the geometry to be deforming.. As soon as I roll on a few frames and some of the particles collide they incur a huge amount of velocity and fly off, out of the collision object.. I can use a speed limit but according to how fast my geometry is moving the particle get left behind and are not wholly reacting to the deformations..

I've attached my .hip file, so you can see what the flip is happening.. at a major problem for me in my schedule at the moment..

I'm looking for some help on this one, I've been trying on my own for a long time, but now it's time to see what you guys/girls think, because time is running out!!

Thanks

Ben..

inMo_deforming_problem.hip

[edward]

Sorry, no time to look at your file. One thought is that make sure you triangulate your deforming geometry for the Collision POP and turn on the hint in it that you're using deforming geometry (I forget the exact parameter name right now). If that doesn't work, then starting increasing the number of substeps that you do in your particle simulation.

[3dFrosty]

Sorry, no time to look at your file. One thought is that make sure you triangulate your deforming geometry for the Collision POP and turn on the hint in it that you're using deforming geometry (I forget the exact parameter name right now). If that doesn't work, then starting increasing the number of substeps that you do in your particle simulation.

Thanks for the response ed.

I worked it out before I read your reply, but fundamentally it was a triangulation problem..

I made the test .hip in the same vain I intended to use the simulation... in which i use a animated human with particles bouncing around inside.. But (here's the error) in the file I was sent the model wasn't triangulated, meaning when I divided it in houdini it updated the triangulation per frame. And I guess thats why the particles went crazy..

So I hacked into the original file and converted it to triangles before bringing it into houdini and like magic the simulation started to work as I previously tested a few weeks ago..

I'll post my result when I get a chance..

Thanks again for the response, it's greatly appreciated

Ben...

p.s. if your interested you can check my blog to see what this whole thing is about.... http://inmotionfx.blogspot.com

[Jason]

I've found that for best results, yes; triangulate your geometry and compute velocities on the geometry yourself (Trail SOP), and use a bit of Collision Tolerance.