AnsweredAssumed Answered

spectrum analysis by VNA - averaging?

Question asked by meta5718 on Aug 25, 2014
Latest reply on Aug 26, 2014 by meta5718

I am trying to do some spectrum analysis with an VNA (HP 8720). To start, I attempted to capture a CW signal from a second VNA (set to 10 GHz). Although there are harmonics/spurious, I were able to identify the signal by a sweep with the receiving VNA at just about 10 GHz (some drift < 100Hz). I think the measurement mode is ratio as I am acquiring S21, although the port 1 of the receiving VNA is open.
I noticed that the SNR was better without averaging, which I hope to understand.

I found another post here:
8753C Average out white noise for each frequency point 

+"When it does averaing, the averaing is complex (re and im) and if it is measuring a response such as gain (B/R) then the averaging works fine as the phase of the signal is constant except for the noise.+

+But, if you are measuring a non-ratioed signal (like B only) then the phase changes from sweep to sweep on the signal, and it will average down to zero just like noise."+

My questions are: Am I measuring non-ratioed signal here? Is that why averaging in my case does not help? 
how to distinguish between ratioed and non-ratioed signal, by whether stimulus is applied to and collected from DUT?
For non-ratioed signal, should I change any setting on VNA? Or can I still acquire S21 as what I have been doing?

thank you!