Particle bounce

[wiked lo]

Hello,

I am working on project wherein there are two seperate popnets;

1) controls the breaking of an object.

2) controls the emitting of particles from a curve and trail sop to create a smoke effect.

Both popnets use the same collision object, a roughly modeled nurbs surface, and should use a bounce on collision. However, the particles directly pass through the collision object. This may be due to the distance between the object and collsion object, which is very small. But if a "stick or stop on collision" is used there is no problem. Utilizing anything other than bounce does not allow for further manipulation later in the timeline. Also if the objects are given more distance between them then some particles bounce, but the great majority still pass thru the collision object.

On a final note, all theses objects were tested earlier in the project and worked successfully, prior to the nurbs surface being implemented which is to allow for correct camera matching and compositing in post-production.

[michael]

look into Oversampling and the Collision Tollerance (on the Collision Tab)

[wiked lo]

I tried initally to adjust the collision tolerance, however this had no effect other then greatly slowing the calculation time. Thanks thou...any other ideas?

[Lukich]

I think I remember having something like that. Do your particles pass through the geometry perpendicularly to it? If so, add a bit of variance to them, it just might fix it.

[wiked lo]

Thanks Lukich..that helped a bunch...I'm alot closer that I was...but it still needs some tweeking, it isn't quite working correctly yet.

[will]

halo wiked,

if you could attach a zipped.hip, we could more easily see what the problem is and likely find a solution. There are many intrepid houdini explorers here who seem to consider a technical challenge an everest that must be summited! ahh the view

will

[wiked lo]

I think I may have figured it out...seems to be some complication with the force POP...when it's used it continues though the collision plane, but when I simply set the inital attributes in the source POP it has not problems with the collision. Kinda wierd but it's working.....for the moment...haha

I will post the file with the next question...there's bound to be more.

[old school]

If the particles are moving very slowly as they approach the collision plane, then the force POP will push them right through on the next timestep.

Removing the force POP from the particles before they collide will solve that problem. You can always use a Group POP and set the group condition to "Has Collided", create an inverse group in the Combine folder and use the "!" Not button like this:

my_inverse_group = ! my_collision_group

Now you can apply nodes that affect the particles before and after the collision by using the groups. This allows you to use force, but after the collision.

Did you use Drag? Always use Drag with Force.

Did you try the oversampling? This does help in some cases with collision. The oversampling parm is located in the Pop network node. Tap U while in the POP network and you will be at the top POP directory. Select the Network Node and make sure parms are on. Make sure to set the same values to the Pop SOP when you move on over to SOPs.