Hi,
I get confused about the CST calculations for reflection and transmission values (S- S-Parameters) for different kinds of transmission lines.
1)First could you please let me know where is the evanescent port mode selection available in CST for waveguide port?
2)If I want to see the S11 parameters of a rectangular waveguide, how does the CST calculate it?
Explain more:
The input impedance of a rectangular waveguide is its wave impedance of dominate mode. So the Zte impedance is not fixed and it varies with frequency. (attachment formula)
So, what type of and what value of waveguide port impedance is considered, for calculating the S11 parameters during the total bandwidth of a dominant mode and also more than dominant mode bandwidth?
3) If the type of TLs are not common or some changes are made, how does CST calculate the S11 for these structures? For popular TLs such as waveguides, we have a common formula.
I appreciate your time and explanation in advance.
CST excites the port to generate the port field pattern, and calculates the impedance from this. I expect it only does this at the centre of the band you choose, because this is what it displays in the 2D results.
I expect it looks for this pattern in the fields exiting the port. If you interrogate the total fields in the port by integrating the mode pattern, electric and magnetic, with the mode pattern that you want, you get the relative excitation, which is the S-parameter. Reversing the sign of the magnetic field switches between ingoing and outgoing power. The modes are orthogonal, so this separates the power in each mode, including evanescent modes.
Dear Malcolm White, I am appreciative of your prompt response. Your explanation helps me deepen my understanding. However, I test a simple waveguide, I design a WR90 in CST with standard dimensions. I set frequency Fmin=1Ghz and Fmax=15Ghz. I see the port impedance (at 2D result), and it shows me the wave impedance of 657 ohms at 8Ghz( center frequency). So I consider it at the impedance @8Ghz. I changed the Fmin to 7 GHz and started the simulation. I see the wave impedance of 469ohm @11Ghz. I calculated S21 (by considering it as a lossless medium) with 469 and 657 ohm and I obtained S21 about -.12dB but CST shows me -.005dB at 8Ghz. It means the impedance at 8Ghz was not the same in the calculation.
I wouldn't expect CST to be that accurate for S21. If you don't run it for very long you have power still bouncing around inside the problem, and if you run it for a long time then numerical errors start to build up. How much ripple was there left on the S11? If you watch the S11 as the solution progresses you can see the ripple reduce every time a lump of energy arrives at the port (see the energy monitor). If the finishing point is -40 dB energy left there is still 0.01% of the energy left in the problem which has the potential to make +/-1% difference in the voltage at the output port which is */-2% of the power. The absence of this from the answer leaves ripple in the S parameter. This ripple is bigger than the differences you are seeing. I'm not sure that the impedance is calculated as you would hope, but I think you can't tell from these results if it's right or not.
Dear Malcolm White, Thank you for your explanation and your time. You are right.
I tried to investigate the value of S11 in different phases.
Phase 1) I calculated the wave impedance at 8 GHz (center frequency) and I found its value: 657ohm and S11 was -144dB.
Phase 2) I enabled “Normalize to fix impedance” in simulation settings and set the value to 50 ohms.
Here is my results at 8Ghz: Impedance: 657 , S11 was:-14dB
My calculation is (Gama=(657-50)/(657+50)=.85) and S11:-1.32dB
This is a big difference.
Phase 3) I changed the Fmin to 7 GHz and started the simulation. (Center frequency is:11Ghz), the impedance was: 465@11Ghz , S11:-115dB @ 8Ghz
My calculation is (Gama=(657-465)/(657+465)=.1711) and S11:-15 dB
This is a big difference again.
Did I make a mistake in each step?
I appreciate your help again.
- Badges
- Science topic