Before my main question, let me ask you if there is any comprehensive app notes about PNA-X Noise Figure converters? I can not find any thing that talking about that in deep.
Recently I have a task to measure Noise Figure on mixer.
The measurement was done with fix IF at 30 MHz and LO-RF sweep between 11-13 GHz
The first thing what I notice was that my noise figure was higher by 9-10 dB (reading around 15-16 dB).
Knowing that double side band, subtracted 3 dB. I got 12-13 dB Noise figure. Still too high.
Next step LO phase noise leak in to mixer IF. I use 3.8 GHz High pass filter get additional 40-50 dB of LO phase noise power and spurs suppression to IF port.
The reading that I get is between 6.5 to 9dB with measured conversion loss 5.4 dB. It that real performance? I am still missing something.
I am wonder how on PNA-X calculate the noise figure with respect of both side bands.
On Noise Figure Analyzer it was impassible to get correct single side band conversion loss with out filters on RF port as input signal (noise source) is wide band. On PNA-X you can do that as the source is selective. Noise power in both NFA and PNA-X cannot distinguish from which side band it come as DUT thermal noise is wide band.
It look like on PNA-X if you measure both side bands conversion loss (most cases there not even) you might calculate more precise noise figure from each side band. Look like problem with equation with two know and two unknown.
I wonder if that approach have sense