Negative R and G Values in High-Frequency Range Using s2rlgc in MATLAB

Hello,

I have been analyzing a transmission line using the s2rlgc tool provided by MATLAB.

While I was able to obtain the desired results for L and C values, I encountered an issue where the R and G values became negative in the high-frequency range.

The s2p file I used was generated from HFSS, and even when I varied the length of the transmission line, the sign inversion issue occurred in roughly the same frequency range.

I have searched extensively for potential solutions and explanations for this problem, but I haven’t been able to find anything conclusive.

I would be extremely grateful if anyone could help me with this issue.

Thank you in advance.

https://kr.mathworks.com/help/rf/ref/s2rlgc.html

Answer 1

[akshatsood]

Hi @Donggil,

I understand that you are encountering negative R and G values in the high-frequency range when using the "s2rlgc" function, which is not the intended behavior.

Based on my understanding, I would like to offer an explanation for the discrepancies you are observing.

1. An initial consideration could be that the .S2P file may contain noise or measurement errors.
2. The "s2rlgc" function is designed to model structures with a shared ground between both sides. If your structure does not meet this criterion, the "s2rlgc" function may not operate as intended, resulting in inaccurate outputs.
3. Behind the scenes, the "s2rlgc" function employs the "Eisenstadt" algorithm, which is effective for high-quality S-parameter data. However, it struggles when S-parameters exhibit significant non-transmission-line effects, which could contribute to the negative values observed.
4. As referenced in the documentation for the "s2rlgc" function, that directly calculating RLCG parameters from measured S-parameters is not considered viable for GHz frequencies.

https://in.mathworks.com/help/rf/ref/s2rlgc.html?searchHighlight=Eisenstadt&s_tid=srchtitle_support_results_1_Eisenstadt#:~:text=the%20s2rlgc%20function.-,References,%5B1%5D,-Degerstrom%2C%20M.J

I hope this helps.