Non-Symmetric Networks

[ptheofan1990]

Hello,
I have tried this code for non-symmetric/reciprocal networks (S11~=S22 and S21=S12) and it seems that the results that I am getting are incorrect. I have been throught the equations on the paper and it seems that at the end the calculation differs from the analysis on the paper:
R1(n) = (w1-b)/(a-cw1);
% R2(n) = (w1-b)/(a-cw1);

alfa = (d-bf)/(a(1-eca));
beta = beta_alfaalfa;

r22p22 = R11t(n)/(aalfa + bgamma);

IRa = [ 1 -b ; -c a ];
IRb = [ 1 -beta ; -gamma alfa ];

Rm = [ R11m(n) R12m(n) ; R21m(n) R22m(n) ];

Rx = 1/(r22p22*(alfa-gammabeta)(a-bc))IRaRmIRb;

S11x(n) = Rx(1,2)/Rx(2,2);
S12x(n) = Rx(1,1) - Rx(1,2)*Rx(2,1)/Rx(2,2);
S21x(n) = 1/Rx(2,2);
S22x(n) = -Rx(2,1)/Rx(2,2);
Could you please let me know if you have seen something similar?
Thanks,
Panos

[la3pna]

What file is this from?

[la3pna]

If this is TRLstone, then increase your line length and see if that works. See the other issue in this repo.

[ptheofan1990]

It is from this one: matlab/vna/trl.m
thanks,
Panos

[ptheofan1990]

I tried that but it did not work. The issue is that the matrix Rm is assymetric and the projection conducted at the end makes Rx reciprocal. I am trying to grasp why this is happening. When my netwrok is asymmetric I should get asymmetric results.
Thanks,
Panos

[la3pna]

Is the Wave cascading matrixes correct? ie, making sure that you feed in S-parameters on the form of amplitude + phase, not dB. The S-parameters needs to be true vectors, with the signs. If so then somthing must be wrong during the quadratic equations, but I don't understand why. This code has been used with a VNA, and compares within 0.1dB to the output response of a HP8720 VNA doing TRL. If you can make a github repo with what you have, then I can take a look, but my experience with this was mostly 4 years or so back, and I would need some time to look into the finer details of this. Please include files, s-parameters, scans of paper math etc, so I can follow your train of thougths. Thomas. Den lør. 28. jul. 2018 kl. 00:18 skrev ptheofan1990 < [email protected]>:

…

I tried that but it did not work. The issue is that the matrix Rm is assymetric and the projection conducted at the end makes Rx reciprocal. I am trying to grasp why this is happening. When my netwrok is asymmetric I should get asymmetric results. Thanks, Panos — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#2 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AEOGvFcnLBeWfJ31VMY-Ly-vtDT1FfNlks5uK5GjgaJpZM4Vkt2z> .

-- With Best regards, Thomas S. Knutsen. Please avoid sending me Word or PowerPoint attachments.

[ptheofan1990]

Thank you Thomas

once I have completed looking into the algorithm I will send it to you.

thanks,
Panos

[la3pna]

Did you have any progress?