Particles collide with PhysX Shape?

[mazi710]

Maybe i am just really dumb, but how do you get particles to collide with PhysX particles? Right now it collides with some mesh through Collision, but i want the particles to also collide with some PhysX objects. Current setup is this, big chunks is PhysX, small ones are particles.
https://imgur.com/uubzM94

[insertmesh]

I don´t think thats currently possible, the particle physics solver and the physx solver are not compatible. I´d just spawn the small particles as physx particles.

Or use cascade sim method:

A) sim physx flow.
B) sim particle physics flow and pick physx flow as collision object.

[tyFlow]

Any particles that you want to collide with PhysX particles must also be PhysX particles themselves....unless you're fine with simple radius-based collisions, at which point the Particle Physics operator will be fine.

However, if using Particle Physics, make sure you uncheck "affect this event only", so it considers particles from other events for its calculations.

[mazi710]

Any particles that you want to collide with PhysX particles must also be PhysX particles themselves....unless you're fine with simple radius-based collisions, at which point the Particle Physics operator will be fine.

However, if using Particle Physics, make sure you uncheck "affect this event only", so it considers particles from other events for its calculations.

I am trying to sim fine sand, so i'm looking at millions of particles. So it's fine they just have a collision radius, they don't need a complex collision shape. So wouldn't the PhysX particles be way harder to simulate if i'm trying to simulate so many particles? But guess that's the only option if there is no way to have PhysX particles collide with normal particles, i just feel like i'm missing something. Like if you have a box of sand and drop chunks of rocks into it. How would you simulate that then? Just all PhysX?

[tyFlow]

There's no 2-way PhysX support at the moment, between PhysX particles and Particle Physics operator (used for grains). And you definitely don't want to try and send millions of particles to the PhysX engine.

You can either convert your large chunks into (very) stiff cloth (to the point where they act rigid) and drop them in, and have everything simulated in the same flow with Particle Physics operators, or simulate your PhysX in one flow, enable mesh and then bring it into the next grain flow to act as a collider.

Or a final option is to simulate your grains at a lower res as PhysX (maybe up to 100k particles before it gets too unruly and slow)...and then upres the sand simulation in a separate flow using particle force operators. I haven't produced example files on how to do that yet but it's planned for the future.

[insertmesh]

Any particles that you want to collide with PhysX particles must also be PhysX particles themselves....unless you're fine with simple radius-based collisions, at which point the Particle Physics operator will be fine.

However, if using Particle Physics, make sure you uncheck "affect this event only", so it considers particles from other events for its calculations.

I am trying to sim fine sand, so i'm looking at millions of particles. So it's fine they just have a collision radius, they don't need a complex collision shape. So wouldn't the PhysX particles be way harder to simulate if i'm trying to simulate so many particles? But guess that's the only option if there is no way to have PhysX particles collide with normal particles, i just feel like i'm missing something. Like if you have a box of sand and drop chunks of rocks into it. How would you simulate that then? Just all PhysX?

Look at the snowman example I uploaded to the Gdrive folder. Its a starting point.

But the easiest way to do would be a cascading sim: First sim the physx particles, then sim the grain and simply pick the physx flow as collision shape.
As long as the physx particles are much heavier than the grain particles and are not being influenced too much by the grain, it should look fine.

But two way interaction between physx particles and particle bind particles is not supported so far (and also might not in the future).

[mazi710]

Thanks for the in depth answers both of you! I will try these methods.

[insertmesh]

Ah, thanks Tyson for the explanation, the uprezzing is an interesting option and thanks for clarifying the particle collision radius thing, that probably explains why it worked so well with the cannonball in my snowman R&D...