Letian Wen

Nanjing Institute of Electronic Technology

When perform a inverse Chirp-Z transform of a frequency domain signal (like S parameter measured in VNA).It's possible that we can observe the signal that is before zero second.Is this signal real?

As known,we can do measurements on the VNA(vector network analyzer )and get both real and image parts of the frequency-domain Scattering parameter,the signal(S_(f)).Then we transform it to the time domain (S_(t)).Usually we just keep the real part of the time-domain signal as we do in the DSP.But since Scattering parameter in frequency domain is a ratio of microwave signal that goes in and reflects out,and when it comes to time domain.Can the TD waveform still represent the ratio of signals in and out?Where does the phase part go?The actual waveform is absolutely not the TD signal from the DFT,that's sure,right?

To transform a band limited (f1~f2) frequency domain signal to time domain (0~Tmax).Chirp-z can do the transformation in the direction from time domain to a partial frequency domain,e.g. f1~f2 . So I wonder if I can perform an inverse CZT? Any help will be appriciated.

Hi everyone,

I've simulated a microstrip device using HFSS with RO4350.The ROGERS data sheet says the Er(relative dielectric constant)is 3.48.But the suggest Er in the software is 3.66.CST also says 3.66 is recommended but 3.48 is for specification.

After fabrication of my design,it turns out that the simulating with Er=3.48 results in a lower centre frequency(40MHz below 2.4GHz).Back to simulation,I change the substrate Er to 3.66,the result agrees well with the measurement.

One of the devices ,produced when I operate the a engraving machine incorrectly,suffered a greater erase depth.So the substrate is thinner. I assume that the effective Er is lower than 3.48.This one is right on my required centre frequency.I begin to find out that the value 3.66 is correct. I made a mistake.

Usually,this difference can be ignored.However in my work here a different Er comes to a different centre frequency.

So I get confused.I know that I should simulate using the recommended value 3.66.But I just want to know why.Why I should use 3.66 rather than 3.48.Does anybody know?please help me.

Hi,everyone,

Is there any good methods or tools for me to draw a geometry of a microstrip design in my dissertation?

Thanks a lot!

Hi,

How can I remove the phase loading effect of a simulated microstrip filter so that I can extract the poles of the admittance matrix of it？

Which paper or book should I refer to?

Thank you.

I use lumped port in in transient network analysis to replace the ones in frequency domain.

Error: Port refinement, process hf3d error: Port '2' is not a TEM port as required for transient analysis. (9:45:34 AM Mar 29, 2016)

"

I am working on extracting the coupling factor of microstrip open-loop resonator with a center frequency of 2.5 GHz.

Somehow I cannot find too peaks in S21 with which I can figure out the coupling coefficient,only in the ELECTRIC coupling case,the peaks appear around  5 Ghz instead.Is it common in some occasions?

I will later attach the HFSS project files (version 15) and a photograph of the layout if possible.Do I miss something important?

However,with the help of the HFSS Eignemode,too resonant frequencies can be found around 2.5 GHz.Is it accurate to extract coupling coefficients this way?