Convergence Problem

[Anders Blom]

We have discovered a possible bug in the Neumann conditions. I suggest you use multigrid, but periodic boundary conditions in all directions. This will work fine since you have a double-gate.

[Anders Blom]

Oh, you have a single gate only. In that case, try to remove the gate from the absolute edge (this seems to trigger the bug); place it 0.1 Å or something inside the cell (as you had originally, in fact..., but a bit wider than before).