Chirp-Z Transform with VNA data

[miguelc]

Has anyone worked with converting VNA frequency data to time data using the Inverse Chirp-Z Transform? I understand the theory alright. I have a bandlimited frequency signal (due to the finite range of the VNA), and I want to calculate the ICZT using that range of frequencies (i.e. compute the ICZT on a portion of the unit circle in the z-plane). My main questions are the input to the ICZT for start and increment. If I had done a CZT in Labview, this would be apparent; however the data came directly in the frequency domain from the VNA. Usually the start would be defined by exp(j*2*pi*fmin/fsample), but I don't know if we can necessarily define fsample for a VNA.

Any insights/experience anyone ahs ahd would be welcome.

Michael

[TsTan]

Hi,
   I am doing a project involving scattering matrix (S parameter) using frequency and time domain analysis. In order to make use of S parameter in frequency domain from Vector Network Analyzer, I need to convert S parameters from frequency domain (sweep from 8 GHz to 12 GHz) to time domain using inverse chirp z transform. I have no idea how to implement it. Could anyone please help or guide me how to implement them.
Thank you so much for your help.
Regards,
Tantong

[xseadog]

Hi both of you

A quick search on Ni site
reveals that the function not in the base package

http://zone.ni.com/reference/en-XX/help/371361B-01/lvanls/inv_chirpz_trans/#details

remember help files are there to assist you

Regards
xseadog

[miguelc]

>>A quick search on Ni site reveals that the function not in the base package

Don't know if you meant this for other users; I have the ICZT function.

>>remember help files are there to assist you

I read through the help files before posting the first time, I'm still a little lost. My problem stems from not knowing where on the unit circle to start. Using VNA data I know my beginning and ending frequency, but because I don't know the sample frequency I can't determine where on the unit circle to start, or how much to increment by. I tried using the Nyquist (2 x f_end) for the sampling frequency, but that seems to give me a Gaussian bell curve.

I was hoping someone had experience with transforming bandlimited VNA data to the time domain...

Tantong, let me know if you figure anything out, I'll do the same for you.

Michael

Message Edited by miguelc on 03-12-2007 10:25 AM

[TsTan]

Hi Michael,
       There was some people done this transformation using inverse chirp Z transform with Mathcad for long time ago. He said it works well. You can check it out here   http://www.rfglobalnet.com/forums/Default.aspx?g=posts&t=16806&VNETCOOKIE=NO

Tan

[xseadog]

hi Michael,

if you know your start frequency and your stop frequency  surely you also know how many samples you have between the two extreme points?

Agilent Network analysers usely have a range sequence 51, 101,201,401,801,1601.

The more points the better the resolution but the slower the time

Xseadog

[Viper]

Fstep = (Fstop - Fstart)/ (N-1) where N = number of points on the trace.

Another question for this topic: TDR usually deals with 1-port measurements. Can this technique be expanded to 2-ports? I am trying to identify reflections in an anechoic chamber that causes multipath from point A to B. Is it as simple as applying the ICZT to S21 instead of S11?

[MehakD]

Is it possible for you to supply us with datasample of the data that you are talking about?