Magnetostatic CST boundary condition problems

Hi All,

I'm simulating a magnetostatic problem involving two magnets separated by a thin copper spacer. The total magnetic assembly is on the order of 10 mm cubed. The magnetization is in the same direction and I use open boundary conditions on all sides.

The problem is that I find that the height of the magnetic field zero point moves around 200 um each way depending on my selection of free space that I select for the background. I have done simulations for 50-300 mm of space on all sides and there is quite a spread of values of the nil in a non linear way. My mesh size is 200 at the moment with a minimum measurement value of 0.05 of a mesh. Please, does anyone know why this would be happening and what I can do to get a consistent magnetic field zero zero height result from my simulations.

Many thanks,

mimeo