Help Obi fluid tunneling through angled objects

[felste]

I'm new to Obi fluid and encoutered and issue when creating containers for fluids. Collision detection works well to the point where I put the Object on an angle.
In my attached example I build a simple box consisting of overlapping box colliders. At an angle the fluid leaks through the point where the box colliders overlap.
What am I missing here or how can this be fixed?

[josemendez]

I'm new to Obi fluid and encoutered and issue when creating containers for fluids. Collision detection works well to the point where I put the Object on an angle.
In my attached example I build a simple box consisting of overlapping box colliders. At an angle the fluid leaks through the point where the box colliders overlap.
What am I missing here or how can this be fixed?

Using more collision constraint iterations may help. This can be found in the solver component, under the "constraints" foldout. See:
http://obi.virtualmethodstudio.com/manua...olver.html

So can slightly increasing the collision margin parameter in the solver.

Overlapping colliders don't constitute a fail-proof barrier in any engine. They're still individual, separate colliders, so any iterative solver (like Obi, Unity, Havok, etc) will solve a collisions with the first collider, then with the second, then with the first, etc.. up to the amount of iterations you specify. This means particles or other objects can still slide trough the overlap between colliders, since each collider is considered individually. The smaller the objects are (and in Obi, particles are pretty small), the higher the chance for collisions to be missed.

Using a distance field is a better approach for complex collision surfaces. See: http://obi.virtualmethodstudio.com/manua...ields.html

[felste]

Using more collision constraint iterations may help. This can be found in the solver component, under the "constraints" foldout. See:
http://obi.virtualmethodstudio.com/manua...olver.html

So can slightly increasing the collision margin parameter in the solver.

Overlapping colliders don't constitute a fail-proof barrier in any engine. They're still individual, separate colliders, so any iterative solver (like Obi, Unity, Havok, etc) will solve a collisions with the first collider, then with the second, then with the first, etc.. up to the amount of iterations you specify. This means particles or other objects can still slide trough the overlap between colliders, since each collider is considered individually. The smaller the objects are (and in Obi, particles are pretty small), the higher the chance for collisions to be missed.

Using a distance field is a better approach for complex collision surfaces. See: http://obi.virtualmethodstudio.com/manua...ields.html

Thanks for your advise. I tried distance fields before and just got it to work. Awesome