Hello

I have the .s1p file [S11 f (F)] of an impedance [z=f(F)]. I connect a quadrupole that I measured as a DUT and of which I saved the .s2p file [S11 f (F); S21 f(F)].

Is it possible to * calculate * the .s1p file of the impedance brought back to the right?
--
F1AMM
François

Yes. You need to make available the 2 port s data of each network.
You might assume the |Z1| is symettric. So S11=S22 and S21=S12.

Then form the casacde see...

https://www.microwaves101.com/encyclopedias/computing-the-s-parameters-of-two-cascaded-circuits

Pick off the S11 term from the cascade.

There is a somewhat simplified method... Look up

Generalized Scattering Parameters

This provides the method for finding a two port S data set in terms of arbitrary source and load Z.
Excellent reference on this topic, see Gonzalez, Microwave Transistor Amplifiers, chapter 1.

It may be no more complicated then applying the bilateral S eqn... try this method.

Where Gamma S is your Z termination. A simple network and a hand calculation show
prove this method.

Hello

Thanks for your answer. I will test.

I understand

https://www.microwaves101.com/encyclopedias/computing-the-s-parameters-of-two-cascaded-circuits

Network A Data is a .s1p file
Network B Data is a .s2p file

output is a .s1P file ?
--
F1AMM
François

In the cascade, the output would be a s2p file.

If your 2 port is NOT symmetric, then you will have to turn that test 2 port you mention end for end to obtain all s data.

Your Z1 will require just a S11 measure if you use the bilateral equation, figure posted, as you just need GAMMA.

I would apply the bilateral formula as it is simplest and straight forward to test with pencil and paper.
Or write a script for your favorite program that handles complex numbers. MATLAB, Mathcad or even a spreadsheet might do.

I would expect a 2 port (.s2p) file that is the overall system response.

For what it’s worth, cascading S matrices is mathematically complex.  Usually what one does is transform them to a ABCD (or T) matrix, which are cascadeable by just matrix multiplication, then transform it back to an S matrix. Similarly, for other combinations (series, parallel, etc.) there are “more convenient” matrix representations.
A book like the one from Mason and Zimmerman (from the 1950s) might be a good start (they’ve got several books, one is more circuit theory, one is more about networks).

If you’re interested, scikit-rf is a python package that does all these things (and plots, etc.).  The learning curve is a bit steep - like all specialized packages you have to learn the way it’s used and how they define things.  But they’ve got lots of examples, and once you have it, you can do a lot of things.

Also, scikit-rf also does standard VNA calibrations.

scikit-rf.readthedocs.io ( https://scikit-rf.readthedocs.io/en/latest/ )
( https://scikit-rf.readthedocs.io/en/latest/ )

( https://scikit-rf.readthedocs.io/en/latest/ )

Here is a nice reference from Rutgers... I believe Jim, he needs to just apply the relationship discussed on page 670... see bottom.

https://eceweb1.rutgers.edu/~orfanidi/ewa/ch14.pdf

Yes, I use Orfanides all the time.
And that would work if all you want is the impedances. But the more general solution is a cascade of 2 ports. And once you’ve built the spreadsheet or software, then you’ve got the full calculation whenever you need it.

On Sun, Sep 22, 2024 at 01:55 PM, Jim Lux wrote:

I would expect a 2 port (.s2p) file that is the overall system response.

If you connect a two-port to a one port, the resultant pair only has one port available to "measure".
Another way to look at it is that the two-port is terminated by the one-port.
Draw a block diagram and it is instantly clear. There is nowhere to connect to "port 2".
73, Don N2VGU

On Sun, Sep 22, 2024 at 11:34 AM, alan victor wrote:

Then form the casacde see...

You can do this again and again, adding an additional two-port to the results of the first calculation, for an arbitrary number of two ports.
73, Don N2VGU

Yes Don. The termination, the one port, must be represented as a two port network and assigned symmetry for proper cascading.

I should add, this would be in its series form, not shunt. In the shunt form, the bilateral s parameter equation would be the most straight forward
method to obtain what the original poster requested.