Liquid sim, water moves whit the emitter at certain point, issue

[Muxiku]

The setup is a spinning object with an emitter spinning
In the first 2 frames all goes as normal.


But then at frame 3 is like it cuts and the liquid remains attached to the source moving with it.

I tried a lot of settings and nothing solves it.

Im using phoenix fd 5.0 with 3ds max 2020

[Svetlin.Nikolov]

Ah, could you share your settings in your Dynamics rollout?

[Muxiku]

It's the default values
I tried to change some settings and nothing solves it.

[Svetlin.Nikolov]

Hmm, yes, there is nothing suspicious. Would it be possible to send over the scene file so I can check it?

[Muxiku]

Yes here you have it

[Svetlin.Nikolov]

Thank you!

Looks like it's a matter of tweaking the scene. I changed a few things to get this result:



So first the "emitter" object seemed to be just a flat geometry and this could make the simulation explode or behave weirdly, so I put a Shell modifier on it.

However, the more serious issue in your case was that the emitters were also very small - just a couple of voxels wide, so this made them emit irregularly. Phoenix turns emitters into voxels internally, so when the emitters are very small compared to the size of a voxel, they could start emitting in the wrong direction (due to skewed normals) or could even get completely ignored by the solver if they are too small. So I added a Push modifier on top of the Shell modifier and used Push Value of 0.01m.

This made the emitters larger and they emit all the time now without glitches during the rotation, but now they started to emit in all directions, so I added a Discharge Modifier to the Source to make them emit only downwards.


I used the Normal in the Z direction for the modifier and told it to emit only when the normal is very close to -1, meaning it points almost straight down:



This is pretty much it.

A final change was that I animated the Emit Liquid checkbox on the Source from OFF at frame 0 to ON at frame 1, because otherwise the liquid changed direction during the first frame as Phoenix needs one frame at the start of the simulation to understand if a geometry is moving or rotating.

An additional note is that since you've enabled splash simulation, by default the Splash particles eat away from the liquid and this makes the liquid mesh quite jagged and uneven, so for my screenshot at the top of the post I turned off Splash simulation off.

Hope this helps, cheers!

[Muxiku]

I tried what you did but that give me the wrong result from the start and seeing your image of the Sim your result also isn't the expected.


What I don't understand is why the first 2 to 3 frames work as expected and the Sim starts doing weird things
I tried to increase the voxel size and I gained 1 frame of correct behavior but obviously the Sim slow down quite a bit, and I use flat emitter on more than 60 Sims and I never encounter a problem in fact I have a similar spinning object maybe not going that fast that it works well.

Also, if I slow down the rotation it goes up to frame 5 doing well, it doesn't make any sense.

[Svetlin.Nikolov]

Ah, sorry but you missed the part where I animated the Emit Liquid checkbox - I explained why this matters.

And about the larger shape and behavior of the liquid - in short, it's about the size of the emitter. You could increase the size of the emitter or increase the grid resolution, but if you change the resolution, the scene would slow down.

WHY this is the reason and why exactly it almost looks like it's working - in short - by chance. I am going to have to go into a very long story with a lot of technical details in order to explain it entirely, so in short - if the emitter size or an obstacle size is about 1-2 voxels, then they are too small and the solver could miss it entirely or account of it, but it would create glitches. As the emitter spins, it travels between voxels so sometimes it could intersect 2 voxels, other times 3 voxels, and this changes its behavior - in this specific scene, the normal of the emission changes.

Hope this makes sense, cheers

[Muxiku]

I try increasing the size of the emitter a lot and the result is the same, the strange part is that isn't irregular is on an exact frame that it starts to simulate wrong.
So I don't think that the size of the emitter is what's causing the problem

[Svetlin.Nikolov]

Try following exactly the steps I described and it will work out.

Cheers!

[Muxiku]

No, what you did doesn't solve the problem in fact it gives the wrong result from the start, the image that you shared is not what I want
I already tried what you did and this is not the result that I want, the water should stay in place no movie with the emitter, Like in the first 2-3 frames of the untweaked simulations

[Muxiku]

In the 7 comment I put you an image that that puts bad and good I want the good part for all the simulation.

[Svetlin.Nikolov]

Ah, ok, then turn off the Motion Velocity option in the Source - this way the emitted liquid would not inherit the velocity of the emitter:

[Muxiku]

Okey thanks, finally is what I wanted !!!

[Svetlin.Nikolov]

Hooray! Yep, this way it's really more beautiful to watch, though note that it would not spread to the sides as a side effect.