When setting your tet size, we recommend following these two key rules:
Use the largest tetrahedra that will capture the deformations of interest.
The tetrahedra can often be much larger than the triangles in the surface mesh. Smaller tetrahedra invert more easily, are less stable, require smaller timesteps, take more memory, and more time to solve. Remember that you can use adaptive refinement to only make tets smaller in the specific areas where you need them to be. See the section on Tetrahedal Meshes for more details.
Above: The resolution of the above tet mesh captures the desired deformation without unnecessary resolution.
Use tetrahedra no smaller than the triangles in the tissue’s surface mesh.
The simulation operates at the resolution of the tetrahedral mesh. If the tetrahedra are smaller than the triangles, then it is doing more work than will be visible. This is wasteful and the solver is not designed to work in this regime. Attachments will behave oddly in this scenario and the surface dynamics will be a poor representation of the solver’s internal state.
Above: The resolution of the tet mesh for the sphere on the right is too high, given the resolution of the embedded surface.