I have a question about the True Mode Stimulus for balanced measurements on a 4-port PNA.
I used the Keysight 4-port PNA (N5227A) to measure the mixed-mode S-parameters of the balanced transmission lines (coupled coplanar waveguide lines) on wafer.
At first, without applying any calibration algorithm, I obtained the raw mixed mode S-parameter measurements with the two scenarios as following:
- Classical S-parameter measurement with the Single-Ended stimulus (Sn_raw_SE) then transformation them to mixed-mode S-parameters according to Bockelman/Eisenstadt (Sm_raw_transformed)
- Mixed-mode S-parameter measurement with the True Mode stimulus (Sm_raw_TrueMode)
As my DUT is a passive device operating in linear-region, I supposed that the both scenarios gave the same mixed-mode S-parameter matrix.
But actually, these two matrices (Sm_raw_transformed and Sm_raw_TrueMode) are totally different.
However, if I applied the calibration algorithm (For example: applying the SOLT calibration using the 85058B calibration kit) before doing the 2 above measurements, I noted that these two scenarios gave the same mixed-mode S matrix (Sm_transformed and Sm_TrueMode).
From this observation, I think that the problem with discrepancy of raw mixed-mode S-parameters comes from the non-linearity of the measurement system (some part of the VNA hardware, Microwave probes, etc).
As I’m not really sure of my interpretation, I would like to consulting ideas of experienced people in this domain.
I would be grateful if any experts here could give me some ideas about this discrepancy between the True Mode and Single Ended Stimulus to obtain mixed-mode S-parameters.
Thank you in advance for your help.