I simulated a smooth waveguide in CST Studio and calculated the S parameters. However, even if the waveguide is matched (using ports), there is a rippling behavior of those. What is the cause of this behavior and how is connected with the length of the waveguide and maybe the frequency spectrum of the excitation pulse?
Probably due to multiple reflections caused by mismatch between the connectors, adapters, cables etc that you did not take into account in the modelling. Usually calibration will remove these.
I expect the ripple period will be close to the frequency which would have half a wavelength along the guide, because it is due to reflections between the ends. I know that this frequency doesn't propagate, but it is a difference between two frequencies that do, and differ by one cycle over the distance to the other end and back.
I'm not sure why CST isn't seeing a perfect match at each end. Perhaps your bandwidth is wider than the bandwidth of the guide?
I don't know if CST uses one impedance for the port, at all frequencies, or adjusts the impedance for each frequency. I think it has only one impedance. This would matter because each frequency has a different impedance in waveguide, so if the port has only one impedance it is only matched at one frequency. Does the ripple get smaller in the centre of the band?
Thank you for the answers. @Malcolm White, yes the ripple is getting smaller towards the center of the band. I don't understand what you mean by "bandwidth of the guide".
Waveguide does not propagate DC. It has a cut-off frequency. Even above this frequency, different frequencies travel at different speeds and with different impedances.
@Malcolm White, does the Gaussian pulse disperse in the medium? I was wondering if not only changes it's shape in time domain but also in frequency domain. This can be a reason of the rippling I guess.
The pulse will disperse. It also isn't a simple Gaussian pulse because this has a DC and low frequency components, which won't propagate, so in the frequency domain they are missing in the waveguide.
Ripples in the S parameters during simulation means the energy has not decayed, the simulation ended with appreciable energy left on the simulation volume. Solutions to that involve increasing the accuracy and/or increasing the maximum number of cycles.
Check on the enegy folder and also the output pulses (time domain), they should ideally reach zero amplitude @ the end of the simulation.
Marcelo B. Perotoni , the energy has been decayed. The energy is smaller than -120dB. Therefore, what other reasons could introduce rippling in the S Parameters? I suspect the fourier transform itself. The pulse of course will disperse but this doesn't change the frequency spectrum of it.
Nadeem Ashraf , simulating the structure in frequency domain will not help me understand the cause of the rippling behavior.
I hope you took wave-guide ports, so must be perfectly matched (by default). if you are talking about ripples near cut off frequency, those appear practically also in a real wave-guide (due to resonant behavior of Wave-guide as HPF). If ripples are in full frequency range, try increasing frequency range, or increase the wave-guide length. Hope this help
Manjeet Ahlawat Thank you for the suggestion. However when the waveguide is not cylindrical but also has some discontinuities, I expect that some frequencies will resonate and not absorbed by the waveguide ports. The waveguide ports to what modes are matched ?
I think all modes are matched. This can be checked by exciting individual mode only, and measuring s parameters. Please check and let us know
pls send the images, some time it is due to the large step size chosen during simulation. and any kind of mismatching occurring at ports
i am also facing the same issue; ripples in the S21 plot.
when i sweep the frequency from 60 to 300 GHz, S21 plot has ripples, above and below 0 in the passband. but when I reduce the sweep frequency (60-80 GHz), then I am getting a better plot.
what should I do now to analyze the complete frequency band of 100 to 300 GHz
- Badges
- Science topic