Reference impedance not equal to characteristic impedance

Hi all.

Regarding S parameters and their measurement with modern VNAs, most of the times reference impedance is 50ohm, as is the characteristic impedance of the waveguides that connect the VNA and the DUT. This way, TRL calibrations (that set the value of Zref) are consistent and the VNA measures the traveling waves with generator and load, no reflected wave in the load port of the DUT.

Is there any real-life situation where one wanted to set a reference impedance different from the characteristic impedance of the VNA waveguides, and different from 50 ohms? I can only think of CAD EM simulations, where modal waveguide ports measure generalized S parameters that can then be"normalized" to 50 ohms.

Regards.

Iban

For waveguides : you use coax to waveguide transitions which are low-loss and cover a full waveguide bandwidth. For other line impedances like 75 or 60 Ohms, and 600 Ohms symmetrical lines, use appropriate transformers and read VNA instructions for such cases. Calibration kits are on the market.

Hi jiripolivka

If I got it right, it is necessary in any situation in which I have to insert adaptors from my input and output coaxes to any other waveguide structures in which I embed my DUT for measure, and that are not 50ohm.

Because then I'm setting my calibration plane in the midle of the inserted transitions, which are not 50 ohm but any other impedance given by their cross section, and this impedance value is now used for S param definition for my DUT.

Is this correct?

The question may be silly but I have always worked with 50 ohm microstrip only, so I never stopped to think about it.

Thanks for your answer

Iban

I think you are right. Concerning the reference plane located "between the transitions" I am not sure. That point is in a waveguide having a different impedance and delay. I would assume better to locate the reference plane at some point along the 50 Ohm line.
If you need to locate the reference plane in the waveguide, at a particular frequency you should calibrate that point by using a shorting shim, and a waveguide termination, to calibrate S11 for a maximum and minimum reflection. The same can be done for S22 from the load side.
I did it that way, and the only difference was the delay from VNA output to reference plane for a coaxial line or the waveguide section.