The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> RF Design >> convert SA data into time domain?
https://designers-guide.org/forum/YaBB.pl?num=1204939382

Message started by sapphire on Mar 7^th, 2008, 5:23pm

Title: convert SA data into time domain?
Post by sapphire on Mar 7^th, 2008, 5:23pm

Hi All,

I am trying to convert the spectrum analyzer's trace data into time domain waveform using Matlab. Any one has experience in how to do that? Actually the SA's trace data contains only amplitude information, no phase information. Can I convert it to time domain correctly?

The SA I am using is R&S FSU model.

Regards,

Sapphire

Title: Re: convert SA data into time domain?
Post by RFICDUDE on Mar 8^th, 2008, 7:03pm

Unfortunately, I believe the answer to your question is "no" you cannot go back to the time domain from just a measurement of the power spectrum of a signal.
Phase information is extremely important in defining the original time domain signal, and there is no easy way to recover phase information from a power measurement.

A vector signal analyzer could at least give you amplitude and phase information regarding the envelope of the signal.

Title: Re: convert SA data into time domain?
Post by pancho_hideboo on Mar 9^th, 2008, 12:12am

Hi.

I have the following instruments.

(1) Sweep Type Spectrum Analyzer
(2) Real Time Spectrum Analyzer
(3) Vector Signal Analyzer
(4) Vector Network Analyzer
(5) Scalar Network Analyzer

IFFT of Spectrum　Analyzer data gives you an autocorrelation function.

But if you want to get impulse response, you have to rely on vector network analyzer not spectrum analyzer.
And most vector network analyzer have IFFT ability as option. Vector network analyzer have frequency offset ability which make it possible to treat frequency conversion device such as mixer.

Unless you measure output from vco, where you don't need signal generator as signal input to DUT, you can't use vector signal analyzer. If you measure vco, you can see I(t)+j*Q(t) directly using vector signal analyzer.
But if you measure DUT's response drived by signal generator, you can't use vector signal analyzer.
Even though you connect 10MHz reference_out of Signal generator to reference_in of vector signal analyzer,
phase are never constant against frequency sweep.
So it is very difficult to evaluate a phase difference of input and output of DUT.

Also current vector signal analyzer don't have a tracking signal output ability to control signal generator.
If my memory is correct, Agilent old VSA89400 had this ability.

Title: Re: convert SA data into time domain?
Post by sapphire on Mar 19^th, 2008, 4:23pm

Thanks very much for you guy's generous postings

Title: Re: convert SA data into time domain?
Post by loose-electron on Mar 20^th, 2008, 10:15pm

silly question - if you have the signal someplace, why not just do a time domain capture and be done with it? The mathematical transformation exists, (with amplitude and phase, thank you Monsier Fourier) but there are a lot of practical limitations and problems doing this.

- Jerry

Title: Re: convert SA data into time domain?
Post by pancho_hideboo on Mar 20^th, 2008, 11:37pm

loose-electron wrote on Mar 20^th, 2008, 10:15pm:

I think this is possible if we use two probe with same trigger signal in oscilloscope.
For example, observe input with probe-1 and output with probe-2.
We can measure magnitude ratio and phase difference(delay).

Title: Re: convert SA data into time domain?
Post by nanion on Mar 22^nd, 2008, 9:47am

By the way, I am not sure if this going to help. But if you have only one single frequency point to look at time domain then you can utilize the zero span of the spectrum analyzer. This will give you a time domain information of one frequency point of interest.