Greetings all,

First, let me say that this is a great forum with lots of information and knowledgeable people. I am glad that I have found it as I am really stuck on something now that I don't quite understand... so let me get straight to the topic.

Basically I am trying to do a full 2-Port 16 error term calibration. I have read a lot of Doug Rytting's papers and I am following the steps he described to do my calibration. My setup is such that I have a RF stimulus unit, and I have two couplers and connectors and such. I have attached an image from one of Agilent's service manuals, so I basically followed that.

I started out by trying to calibrate 1-port first by calculating the 3 error terms associate with it. I did this by connecting a load, short, open. so my RF out port generates the frequency range desired, it goes into a splitter (resistive), one side of the splitter goes into a 10dB attenuator then to my REF receiver port. The other end of the splitter goes through a 10dB attenuator and goes into a directional coupler. The output of the coupler would be my PORT 1 for S-Parameter measurement, the coupled port goes into port A (2nd reference port) of my receiver. REF port is one of the receiver ports, and Port A is the second receiver port that measures the reflected signal. However, I am stuck on how I can calculate the Gamma from my data. Basically my receiver spits out I (inphase) and Q (quadrature) data of however many points was put in. Supposed the magnitude is (square root (I^2 + Q^2)) and the phase is arctan (Q/I) Do I calculate the Gamma by dividing the reflected signal (magnitude) over the incident signal?

I have calculated the 3 error terms for 1 port, I used a 3dB attenuator for my test case, so I am supposed to see a 6 dB return loss (S11). But I am unable to get that result. Am I missing something or doing something something wrong? I am really stuck here...

Thanks in advance for all your help!

Kevin

First, let me say that this is a great forum with lots of information and knowledgeable people. I am glad that I have found it as I am really stuck on something now that I don't quite understand... so let me get straight to the topic.

Basically I am trying to do a full 2-Port 16 error term calibration. I have read a lot of Doug Rytting's papers and I am following the steps he described to do my calibration. My setup is such that I have a RF stimulus unit, and I have two couplers and connectors and such. I have attached an image from one of Agilent's service manuals, so I basically followed that.

I started out by trying to calibrate 1-port first by calculating the 3 error terms associate with it. I did this by connecting a load, short, open. so my RF out port generates the frequency range desired, it goes into a splitter (resistive), one side of the splitter goes into a 10dB attenuator then to my REF receiver port. The other end of the splitter goes through a 10dB attenuator and goes into a directional coupler. The output of the coupler would be my PORT 1 for S-Parameter measurement, the coupled port goes into port A (2nd reference port) of my receiver. REF port is one of the receiver ports, and Port A is the second receiver port that measures the reflected signal. However, I am stuck on how I can calculate the Gamma from my data. Basically my receiver spits out I (inphase) and Q (quadrature) data of however many points was put in. Supposed the magnitude is (square root (I^2 + Q^2)) and the phase is arctan (Q/I) Do I calculate the Gamma by dividing the reflected signal (magnitude) over the incident signal?

I have calculated the 3 error terms for 1 port, I used a 3dB attenuator for my test case, so I am supposed to see a 6 dB return loss (S11). But I am unable to get that result. Am I missing something or doing something something wrong? I am really stuck here...

Thanks in advance for all your help!

Kevin

Normally, you want the loss from the splitter to the Ref channel to be the same as the coupler's coupling factor, roughly. So, if you have a 20 dB coupler, you would want 20 dB attenuation to the Ref channel, and no attenuation between the splitter and the coupler, where the coupled arm goes to the A receiver.

Now |A| should approx equal |R|, where |A|=sqrt(I^2+Q^2). Phase of A is atan2(Q,I). You need to use ATAN2 or equivalent to get 360 degrees. Arctan functions of the ratio always return -90 to +90.

Now, before we do any correction, measure the open, the|A|/|R| should be 1. If it is no 1, then 20 Log (open) is approximately your reflection tracking error term.

Measure the short, it should also be nearly 1, or rather, nearly the same as the open. The vector difference sum of the open and short is approx 2 times your source match.

Measure the load, it should be less than .1, this is your directivity. The numbers I give are in linear, do 20*log10(number) to get the dB value.

Measure the pad, with it left open. It should have a value of .5, or rather, 0.5*open value.

Post some results and we can sort out calibration later.