ifft of spectrum from VNA

Hello, I'm doing some project to school and I've got a problem..spent a lot of time on it, but without any solution.

I've got a output from vector network analyzator (s-parameters) in frequency domain. I need to transform it into the time domain. But the output of my matlab calculation isn't the same, as output in time from the VNA (using the built-in features of the VNA). In the following picture you can see the differences. It looks like that there is some strange shape "added" or what:

http://s13.postimage.org/qydriepxz/matlab2.jpg

The measurement was from 5 GHz to 50 GHz; 801 points measured (is it enough points considering the huge freq range?)

Here is my code:

 %s11_komplex - measured s11 parameters in complex numbers (contains magnitude and phase)

 %s11_komplex = [zeros(89,1);s11_komplex]; % I was trying to put zeros into the missing freq range 0-5 GHz, but it makes no difference

 Y1=[s11_komplex(:,1);flipud(conj(s11_komplex(:,1)))];
 time = ifft(Y1);
 plot (20*log10(abs(time))) % I've also tried "real" instead of "abs" - no significant difference 
 plot(import(:,2),'green') % plot data from VNA to compare

Can you help me, where is the problem, that the results are in some areas different? Thank you a lot, I'm quite frustrated..

Answer 1

[signal processer]

dear friend did you find a way to fix your problem？I have the same problem with you，and i think the key of the problem is the raw data from VNA.if you use the mag and phase，the signal should be magX.*exp(li*angX).or if you use real and imag,the signal can be real+imag*li.