Hey @ all

I record the S-Parameters with a VNA in the frequency range from lets say 1-11 GHz and I want to get the time domain signal via the inverse FFT. This is what I am doing before appliing the inverse FFT

[list=1:3sft0rlc]

I extrapolate the S-Parameters down to DC (adding zeros at the beginning). So far I have N values corresponding to frequencies in the range [DC ... fmax=11 GHz] .

[list=1:3sft0rlc]

I calculate the complex conjugated values form the N values above (but exclude the DC value) and append the former in reversed order to the latter. All these points correspond to the frequencies in the range [DC ... fmax -fmax ... DC[ where the second half represents the negative frequencies. This means I now have N+(N-1) values for feeding the inverse FFT.

According to the rules of FFT the time step DELTA t between two neighboring points in time domain - which is my capability of resolution - equals to the inverse bandwidth in frequency domain, DELTA t = 1/bandwidth (Fourier pairs).

My question now is:

Who has some ideas/information/papers for me dealing with this topic?

I record the S-Parameters with a VNA in the frequency range from lets say 1-11 GHz and I want to get the time domain signal via the inverse FFT. This is what I am doing before appliing the inverse FFT

[list=1:3sft0rlc]

I extrapolate the S-Parameters down to DC (adding zeros at the beginning). So far I have N values corresponding to frequencies in the range [DC ... fmax=11 GHz] .

[list=1:3sft0rlc]

I calculate the complex conjugated values form the N values above (but exclude the DC value) and append the former in reversed order to the latter. All these points correspond to the frequencies in the range [DC ... fmax -fmax ... DC[ where the second half represents the negative frequencies. This means I now have N+(N-1) values for feeding the inverse FFT.

According to the rules of FFT the time step DELTA t between two neighboring points in time domain - which is my capability of resolution - equals to the inverse bandwidth in frequency domain, DELTA t = 1/bandwidth (Fourier pairs).

My question now is:

**Do I calculate the value of DELTA t using the physical bandwidth (i.e. 10 GHz) or using the "iFFT bandwith" (i.e. 20 GHz) which is doubled because of the negative frequencies?**Who has some ideas/information/papers for me dealing with this topic?

You might read the time domain application note to see http://cp.literature.agilent.com/litweb ... 5723EN.pdf for more details on VNA time domain.