Tuning a filter using the time-domain response of S11 2

[IW6OWB]

It is possible to tuning a filter using the time-domain response of its return loss: S11.
To do this I need a mask for the filter: S11 vs. Time.
With this method it is possible to tune each resonator individually,
since time-domain measurements can distinguish
the individual responses of each resonator and
coupling aperture. Such clear identification of
responses is extremely difficult in the frequency
domain.
Normally a RF simulator gives a plot S11 vS frequency, how I can obtain the filter time-domain response with ADS2003?

For more details you can see Agilent AN 1287-8.

Thanks.

 

[biff44]

Well, it looks like you design a filter, measure its reflection coeeficient magnitude OVER ONLY THE PASSBAND, do a fourier transform of the Refelection coeficient vs frequency to get it into the time domain in some program like Mathcad, and print it out.

Sounds like a gimmick to sell more grossly overpriced Agilent Network Analyzers to me! You can do the same thing in the frequency domain using a polar plot of S11. Crystal wrote a paper on it once.

 

[biff44]

You tune the filter's S11 to look like "petals of a rose".

 

[Higgler]

I heard it's useful to tune in time domain as a preliminary setup for the filter. After you've got the time domain plot looking fairly good, switching back to frequency domain is required for final tuning.

I suppose the need for time domain tuning would depend on the range of adjustment of your tunable components. If a technician is not too bright, and the tunable components have large tuning range, then it could be helpful.

kch