PDA

View Full Version : nparticle collision with plane problem


royter
10-16-2010, 10:20 PM
i am choosing the point option for my nParticle emitter but when they collide with the plane (made passive collider) they always collide before they touch the plane's surface as if there is an offset, or another hidden plane that collides with them, why is that?

another problem i am getting through is that sometimes the particle wouldnt collide at all with my plane, even if they are on the same layer and collide is checked ON in both nodes.any ideas?

mandark1011
10-17-2010, 11:18 PM
for your first problem with the collision looking like it has an offset, check your passive rigid settings for collision. Turn it on and you will see a yellow approximation of your collision tolerance. For some reason on ncloth objects by default this is very hi lower it and that will correct your collision issues.


as for the second one if your particles are moving too fast they may be missing the collision calculation all together try uping your subsamples and trying again.

good luck

royter
10-18-2010, 11:05 PM
for your first problem with the collision looking like it has an offset, check your passive rigid settings for collision. Turn it on and you will see a yellow approximation of your collision tolerance. For some reason on ncloth objects by default this is very hi lower it and that will correct your collision issues.


as for the second one if your particles are moving too fast they may be missing the collision calculation all together try uping your subsamples and trying again.

good luck

thanks for the info.
i tried tweaking the thikness value but this didnt change anything: image (http://www.youfile.net/download.php?file=77b4665b6598b95f6b85815c42342a7d)

Duncan
10-19-2010, 12:26 AM
The collision distance is based on both the particle radius and the nRigid thickness. Set both values to zero.

Duncan

royter
10-19-2010, 01:35 AM
The collision distance is based on both the particle radius and the nRigid thickness. Set both values to zero.

Duncan

a zero thikness value to the plane makes the particles bounce once and get through it.
a 0.01 value fixed the problem.
thanks anyways.

Duncan
10-19-2010, 04:01 AM
Yes, zero width collisions can fail... basically the particle does not know which side of the surface it is on. One can use pushOut on the nRigid to keep particles pushed out on the normal size of the surface. Thus if you make pushOut 1 and the pushOut radius at least the distance the particle falls (from a standing start) in one step then zero collision width will work OK.


Duncan

royter
10-19-2010, 04:55 AM
Yes, zero width collisions can fail... basically the particle does not know which side of the surface it is on. One can use pushOut on the nRigid to keep particles pushed out on the normal size of the surface. Thus if you make pushOut 1 and the pushOut radius at least the distance the particle falls (from a standing start) in one step then zero collision width will work OK.


Duncan

thanks for the info

Robbert
05-18-2011, 09:24 AM
Is there a trick to do the thickness in a negative value?
So that the particles bounce under the plane instead of on the plane itself.
I needed that for a couple of jobs but I always need to duplicate the object and do a transform component so that the particles collide more inside the object. I mostly need it for colliding nParticles raindrops as output mesh. It's nicer to have the raindrops intersecting the object instead of colliding on the object. Is there a trick for that?

mandark1011
05-18-2011, 05:24 PM
i think that IS the trick duplicate do a transform component and voila.

Robbert
05-24-2011, 10:36 AM
Yes i know, but you are not really flexible if you have 50 objects.
Right now I have to script my way...

Calibrix
05-24-2011, 05:53 PM
Maybe I'm not understanding what you're trying to achive but couldn't you just move your collision plane instead?

mandark1011
05-24-2011, 05:58 PM
he is using a plane in its most simplistic sense..really he means objects that the nParticles collide with. ....i think haha.

I also have had to script this workaround, doesn't seem like a huge hassle for me tho, i would much rather see the ability to have an event recorded on a collision that has no collision like with normal particles. that would be very very useful. ie -1 resilience on normal particles.

CGTalk Moderation
05-24-2011, 05:58 PM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.