I have set up this scene watching tutorials..
and I followed exactly how the instructor did, but my particles are passing through the collision (the cup).

I tried increasing the division numbers, but it didn't work…

Do you know how I can fix this?

Thank you.

Thicker collision geometry or more substeps for the solver.

Thank you
I just tried thicken the collision, also the substep number…

thickening the collision seems helpful, but still not perfect…
and the instructor used the same model as I am using, so shouldn't it be a problem?
also the substep, I don't know which parameter you meant, so I tried
increasing max substep parameter under Particle Fluid Solver.
Also, tried increasing Substep Repetition Tolerance number under Advanced tap, but didn't work…

If you are using SPH you might need to also change the maximum substeps on the AutoDopNetwork.

Are you sure the instructor did not change anything? The model seems awfully small and thin for it to work with default settings.

From my recent “floundering about with wires”, I found that creating a hidden thicker version of the geometry and using that for collisions worked best, as the wires were penetrating on thin geo despite turning the numbers (including substeps) up to “super”.

I probably could have found a value that worked up in the “super-duper” range, but feedback was getting pretty boggy…

Post the hip file if you like. I'm not good at this either yet and would be glad to have a go…

Yea……. I guess I can use thicker version for collision
altho, I wish to know why mine didnt work, when my instructor had no problem. I used his beaker.bgeo, so it shouldnt have been a problem though.
Thank you !

Hey, bluciensky, don't sweat about it!
I think the important point to understand here (and anyone please correct me if I'm wrong) is that when particles or objects are moving too fast the simulation suffers from loss of precision. The substeps parameters create more calculations between frames to estimate values of forces, velocities, positions, etc, thus keeping the simulation more accurate.
Once you kill the collision geometry issues and your nodes are wired correctly (since you were following a tutorial that should be right), I would say substeps is the next thing to look at.
Why the example worked for the instructor and not for you?! No idea, but if it makes you feel better, it didn't work at first in my machine either!

Cheers

PS: you might want to check also if the laser scan isn't the issue, cause in my test geometry it did create a wrong representation for the interior of the cup.