How Do S-Parameters Affect Attenuation in WG Transitions and Circulators?

[lennybogzy]

I was hoping someone might have some insight as to how I can correlate some attenuation measurements.

Here is my setup:

I have a KU-band CW source going into a 3-port switch, into 3 semi-rigid SMA's into 3 waveguide transitions, into 3 circulators, into 3 long pieces of waveguide. On the back end the three go into 3 more sets of circulators, transitions, SMA's and then back into a switch on which I have a power meter.

I can then switch between the three different paths and measure 3 DUT's, one at a time.

The issue that I have is that I have different attenuation measurements from port to port. Some can differ by as much as 1dB.

Here is my question:

I'm pretty sure it is the transitions and/or the circulators that are causing this port to port deviation. What should I look at in terms of the s-parameters of the transitions and/or circulators to see which ones are giving me this headache, which ones to pair together, etc. What would give me the greatest insight?

 

[Corneo]

You should look at the S21. If you want to know the loss between transitions, just measure it on a network analyzer, however this could get tricky in terms of calibrating on a VNA though.

 

[lennybogzy]

I understand the importance of looking at the loss. However, when I take my attenuation readings, I zero-out the insertion loss of each port every time. Therefore, I would expect that the loss in the transitions shouldn't affect my attenuation readings, even if they're all different because it just gets offset. My feeling is that maybe this has to do with s11. Anyone have any input as to how VSWR interactions would affect attenuation readings (after insertion loss is zeroed)?

Hypothetical example: I connect waveguide to waveguide, REL my power meter so it reads 0.00dB in each port.

Then I put in some sort of pad or coupler or whatever. 1 port reads 5.2dB one reads 4.8dB and another reads 4.43dB. What could cause this.

 

[Corneo]

When you say you get 5.2 dB and another reads 4.8 dB, are you saying you have 2 identical pad/coupler/whatever? It could be just variations along between one device to another.

There is something call mismatch loss and it related to S11, $$ML = 10 \log_{10} (1-|\rho|^2)$$.

 

[lennybogzy]

Right, what I mean is if I take the same one and move it from port to port, I'll have these different readings.

I'll look into mismatch loss, thanks.