My Community

I have a few final suggestions. You can try the following:

• Use PBEsol (GGA=PS)
• Run with ALGO=All, which usually improves convergence for magnetic systems
• Vary POTIM and check if the results are stable
• Make sure that you have a sensible MAGMOM. There is a change you don't reach the correct magnetic groundstate and are stuck in a local minimum. Changing MAGMOM or preconverging the calculation with a nonmagnetic calculation can help with convergence.

I'm afraid this is all the advice we can give. Make sure that you systematically double-check all your convergence criteria (KPOINTS, ENCUT, EDIFF, POTIM). This might still be due to an underconverged calculation, though I doubt that it's related to ENCUT.