0
\$\begingroup\$

I guess this is more of a conceptual question but for the problem below Im having trouble getting started on how to derive the S parameters for the network.

Blockquote

I noticed that the network at 28.025 GHz is resonant, so the shunt capacitor and inductor are open, leaving the equivalent circuit at that frequency to be just the resistors

To derive the s parameters I understand that I have to terminate a port with a matched load, at either 1 or 2 to so that the Reflection Coefficient = 0

Now here is a my confusion, it seems to me that going from port 1 to 2 the characteristic impedance of the circuit (Z_0) would be different than going from port 2 to port 1 leading to different matched loads.

I think that they are different because if you inject a signal into port 1 and short port 2 to ground you encounter an impedance of:

enter image description here

But if you inject a signal starting at port 2 you get:

enter image description here

I believe there's a high chance of doing those resistance calculations wrong, but Im just not sure where to start or if im headed in the right path.

Which boils down my question to

Is it true that the Characteristic Impedance of this network changes depending on where you inject the signal, leading to different matching loads? If not how do you calculate the system impedance this network in particular?

My professor in class has always shown examples and given us homework of simple two port networks that look symmetrical from either port where the system impedance is the same whether looking into port one or port 2 matched loads will be equal so im pretty confused right now.

\$\endgroup\$

3 Answers 3

1
\$\begingroup\$

It is possible to define S-parameters for networks intended to be used between systems of different impedance, for instance transformers or pads intended to work between 50 Ω and 75 Ω systems.

However, when you are asked without further clarification to compute the S-parameters of a network, you should assume it's being used in a 50 Ω system.

There may be many practical reasons a network does not present a good match to a 50 Ω system. Perhaps it's actually intended as an antenna match.

Regardless of what actual impedance a 2-port is intended to present, and regardless of what terminates the other port, the S-parameters can be combined so that the measurement in the 50 Ω system can be used to design for other impedances.

You are not being asked to compute the system impedance of this particular network. What you are being asked to do is

a) Terminate port 2 in 50 Ω
b) Drive port 1 with a 50 Ω source
c) Compute S21 as the ratio output/input
d) Compute S11 as the ratio reflected/input
e) Swap ports 1 and 2, and do the same for S22 and S12

\$\endgroup\$
2
  • \$\begingroup\$ hmmm I see! So regardless of what the actual network looks like the system its being used in should be assumed is Z_0 = 50ohms. and the network above is just part of the input impedance and the input impedance changes depending on which port you're looking into. For example S11 = V1-/V1+ = (Z_in- 50)/(Z_in + 50) \$\endgroup\$
    – brian736
    Oct 21, 2020 at 17:48
  • \$\begingroup\$ Input impedance being different on the two ports is just saying the network is not symmetrical. You don't have to assume it's going to be used in a 50 ohm system, though it usually is. You are measuring it and defining it in a 50 ohm system, and those measurements are easily definable, and stable, and can be transformed mathematically into any other impedance system. Hybrid parameters use shorts and opens on the 'other'port, but those are often unstable for RF and microwave amplifiers, hence the invention of S parameters. \$\endgroup\$
    – Neil_UK
    Oct 22, 2020 at 7:00
0
\$\begingroup\$

I think that they are different because if you inject a signal into port 1 and short port 2 to ground you encounter an impedance of:

enter image description here

First, you are not calculating \$Z_0\$ here. \$Z_0\$ is a given of the problem, typically 50 or 75 ohms, but in principle it could be chosen arbitrarily for your system. What you're calculating here is \$Z_{in}\$ the input impedance.

Second, if you short port 2, then the 200-ohm resistor is shorted, and shouldn't appear in the equation for \$Z_{in}\$. The actual equation with port 2 shorted would just be

$$Z_{in}=15+ (36||500)$$

Third, if you want to calculate the S-parameters, you should be terminating port 2 with the system characteristic impedance (again, usually 50 or 75 ohms, but possibly arbitrarily chosen by your instructor or the system designer), not with a short.

Is it true that the Characteristic Impedance of this network changes depending on where you inject the signal, leading to different matching loads?

No, the characteristic impedance is a system-level parameter that needs to be chosen before you design or analyze the elements in the system.

If the 2-port isn't well-matched to the system \$Z_0\$ then that just means the 2-port isn't well matched. It doesn't mean the system \$Z_0\$ has changed.

\$\endgroup\$
2
  • \$\begingroup\$ I see thank you for pointing out what I did wrong clearly, so say I wanted to calculate S11, now that Z_0 is assumed to be 50 ohms the goal is to terminate Port 2 with a resistor that creates an input impedance of 50ohms as well correct? Such that the reflection coefficient = 0? Or is port 2 also terminated with 50 ohms? \$\endgroup\$
    – brian736
    Oct 21, 2020 at 19:31
  • \$\begingroup\$ @brian736, no, you terminate port 2 with 50 ohms and measure/calculate the impedance looking into port 1. Then you can calculate \$S_{11} = \frac{Z_{in}-Z_0}{Z_{in}+Z_0}\$. You're trying to measure the properties of your 2-port, not trying to optimize the system to work with your 2-port. \$\endgroup\$
    – The Photon
    Oct 21, 2020 at 19:32
0
\$\begingroup\$

S11=0.236+j2.75u (50 ohm)

S12=0.53+j5.81u (50 ohm)

S21=0.53+j5.81u (50 ohm)

S22=-0.081+j12.4u (75 ohm)

Yes, it's reciprocal.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.