cst expert system mesh
Sorry I have no specific advice regarding your particular problem.
I've seen this happen alot with this software. Sometimes it's understandable; for example, maybe you have some small feature that would be expected to have a relatively large effect, e.g., a small air gap in a high permeability magnetic core, and it just doesn't get meshed properly until you intervene and modify the mesh.
Other times, you just get significantly different results with no obvious reason for it.
What irritates me most is the difference in results I often see between hex and tet mesh, or between the "expert system" and "energy based" hex adaptive meshing.
Another annoyance is the poor documentation.
I've been putzing around with CST EM Studio for quite a while, and I now have very little faith in it...it's hard to trust the software when the software doesn't even agree with itself.