Went from a thin sheet to simming a full tet mesh and still it's impossible. Also the scale dependancy and topology dependancy makes it hard to follow a process, and I always end up value mashing until something looks plausible.
I attached the results I get at 100x scale difference.
This is the starting shape, although not really relevant. Using 3 substeps in these examples.
Please, any tips on how to make sense of all this?