The problem occurs when outflow boundary conditions for hydrodynamics are combined with self-gravity. It seems that it only shows up if the blocks are refined at runtime (that is why you don't see it in the dust collapse setup). We have used the sedov setup with gravity.
Config
flash.par
init_block.F90
We show runs of this setup for periodic (lhs) and outflow (rhs) hydro boundary conditions. Although both gravitational potentials look sensible at the beginning, there are large disturbances in the end.
![]() |
![]() |
![]() |
![]() |
| sedov.log | sedov.log |