S-parameters; reflexion factor and transmission factor

[zeroname]

Hello!

I hawe a question regarding this to things.

well i know that s11,s22,..,sxx is representing the reflexion factor.
so we sayed that the reflection factor should be better than -10dB

and we sayed that the s12,s21 are representing the transmission factor which should be near 0dB to be the best.

well but i dont hawe a clue, if iam plotting in HFSS a modal solution Rectangular Plot of the s paramaters, how or what should i look at the plots.

So if iam plotting a s11 Plot the line should at the desired freq drop under -10dB. Or how?!

But if i plot the s12 or s13 parameters the line should be as near to 0dB as possible??! what if the line drops to lots say -11dB i think this is acceteble or??!

Maybe this question sounds stupid but iam confused... or can it maybe that i can plot with the s12 reflexion too???

 

[jgruszynski]

What are the axes you are choosing? That should tell you want you are looking at. Usually a rectangular plot is complex sij vs. frequency. For sii or sjj you could use a rectangular plot but the more conventional and useful thing is to use a Smith chart.

In terms of interpretation, remember s-parameters are just N-ports akin to y-parameters, z-parameters or h-parameters. sii terms are basically like input/output impedances of the port i (with a Moebius transform and normalization to characteristic impedance). sij are the transfer functions between ports i and j.

A "dB drop" is for sii tells you something about impedance mismatch, up or down stream. For a filter, mismatch reflection is another way of visualizing how a filter actually works. It's also the same as mismatching lumped impedances.

A "dB drop" for a transfer function (sij) is a measure of insertion loss for a passive component (0 dB or less) or gain roll-off for an active device. Translate the -11 dB to magnitude and ask your own question again. That's like a 0.23 x power transfer in magnitude.

It might be OK at a given frequency or if you are expecting such an insertion loss. If you device is a 10 dB pad/attenuator, then that would be about right.

You may want to look at this PDF (I'm one of the co-authors):

**broken link removed**

**broken link removed**