Problem with SOL Calibration in AWR

[ffrank]

Hi everyone, I got a problem with SOL (short-open-load) calibration testbench for matching network for power detector in AWR Design. I built 3 testbenches (short open and load to measure S11 parameter of cable TLine) and came up with output equations to exclude influence of cable to measurements.

But it didn't work properly, because calculated parameter S11 (GammaDetector) doesn't match original S11 parameter for power detector LTC5587. Can you help me find the problem? Thank you 

[Joel_Dunsmore]

OK, looking again looks like you missed a minus sign in the last term; from my book, second edition page 136: S11A=(S11M-EDF)/(ERF+(S11M-EDF)*ESF)

So in your terms I think it is :

GammaDetector=(GammaIn-S11ml)/((pow(S21k,2)+gammaIn*S22k - S11ml*S22k)

[Joel_Dunsmore]

And, whenever I get stuck like this, I just plot all the bits and pieces of the equation.  So plot GammaIn, and also GammaIn-s11mL, should be almost the same.
And plot S21K, should be almost 1 (0 dB).
And plot S22K, should be almost zero (-200 dB or something very small).
And plot S11 mL (should be almost zero). 
Then GammaDetector should be almost GammaIn/(almost 1 +almost zero - almost zero).

Good to do this for intuition as to how all the parts fit together. 

[ffrank]

Thank you very much! I think that the problem is S21k on +77 dB (Oh God )) and S22k on +154 dB.. S11ml is on -316 dB on 5GHz, but sometimes goes down to 800.
GammaIn - S11ml is right at the position of GammaIn.

 Perhaps the schematics are wrong, because equations are well known :/ 

[Joel_Dunsmore]

GammaIn and GammaIn=S11ml should be the same in amplitude, as the transmission line doesn't affect amplitude unless it is non-ideal (lossy or not Z0).  but you will see GammaIn - Corrrect(GammaIn) has a value related to the (Sine(of the phase of the line).    As you have it, everything is ideal so only the phase of the detector impedance is changed.  That's why S11ml is -infinite dB : 20log10(0) = big negative number and except for numerical rounding's, S11ml should be 0.

[ffrank]

Thank you for explaining me this 
Unfortunately, I still got these values as You can see on image.
S11mo - S11 ms = 1-1, so everything goes up to infinity...

[Joel_Dunsmore]

With apologies to The Graduate: "One word, phase"
You have to retain the complex values for everything you do. so GammaOpen is 1 angle 0, and Gammashort is 1 angle 180 so gammaOpen-Gammashort=2.  It will make things better.

[ffrank]

Thank you! Phase is King, I changed signs in equations and I got the result