Equiv. impedance of a loaded TEM cell: series or parallel?

[Decesicum]

Hello again!
I have computed the S-parameters of a TEM cell and determined the absorbed power as P_abs=P_incident(1-|s11|^2-|s21|^2)
The absorbed power determined as such is actually the active power (not the reactive, nor the apparent), isn't it?
Then I loaded the cell with some lossy probes and re-computed the S-parameters and again the absorbed power (in cell and probes) What I got was then, wheb the probe is inserted, the absorbed power is sometimes higher than in the case of the empty cell, sometimes is lower, varying with frequency.
Why is it so?? I expected to have a higher power absorption all the time (regardless the frequency) What am I missing?
C.

 

[RCinFLA]

First you have to make sure your frequency is within the range of the TEM cell's dimensions. It will jump modes (no longer TEM) if operated too high in frequency.

Second anything you put in the TEM will have an influence on the impedance versus distance creating reflections back to the source. The general requirement is you should not put anything in the gap that is more then half the gap. This is overly simplified as a solid metal brick of half the space will significantly change the impedance at that point in the TEM cell.