CST MWS stability problem

Hallo,
i am working on a microstrip via simulation.
The microstrip configuration is inhomogeneously; means the room between
the strip and the ground plane is filled with dielectricum AND air.
For my understanding, the energy balance curve gives information
about the power transmitted and received on a port (e.g. waveguide port).
Therefore it must be below 1.
I got two phenomena.
1) The whole calculation has balance values below 1, but after
deembedding the structure i get transmission coefficients (S21) with
gain (>0dB) in a passive structure. Suspiciously !
2) The whole calculation has balance values greater than 1.
Any statement would be helpfull. Thanks.

Hello,
Make sure you don't mesh to dense, CST sometimes gets into trouble if you have smaal mesh cells. If you have a coarse mesh, consider mesh refinement. The energy balance curve should never go upwards only downwards or horizontal for a few time steps.
Bye.

Now I know the reason for the active behaviour of my passive structure.
Its just a deembedding problem. I calculated Zline, beta for only one frequency, although the values are frequency dependent. Therefore the algorithms DEEMBEDDING and RENORMALIZATION dont work.