Phase Noise Measurement using spectrum analyzer (PLL technique)

[fparveen]

Hello,

I am planning to measure the phase noise of a VCO using PLL technique using spectrum analyzer model # R&S FSL (9kHz - 3GHz).

My question is what should i look for in the spectrum analyzer for measuring the phase noise. Do I need to plot the output of the LPF (connected at the output of the mixer) from 9kHz to, say, 100kHz? Then the resulting display will give me the rms value of the mixer output (amplified by a LNA), at different frequencies. Will it give me the phase noise once I divide the measured rms values from the display by the calibration constant, Kd? (Kd is the zero-crossing slope of the low pass Filter's output, which is the proportionality constant between phase noise of the oscillator under test and rms voltage measured at the LPF output)

Or will I have to set a center frequency, say, at 10kHz and span, say 2kHz, then I measure the signal rms power and divide it by the span to get the rms value per Hz. And then repeat it by changing the center frequency. Then calculating the phase noise from the rms value measurements using the zero-crossing slope calibration. And finally plotting the values at each frequency in MATLAB to get the phase noise vs frequency plot?


Thanks in advance,

 

[BigBoss]

You can measure PN with Spectrum Analyzer but it won't be very healthy.Although modern Signal Analyzers have Phase Noise Measurement Features, I'd prefer use a Phase Noise Analyzer for better results..

 

[fparveen]

Thank you very much for your reply. However, I don't have access to any Phase Noise Analyzer. I have the spectrum analyzer R&S FSL in the Lab which I need to use to measure phase noise. I understand the results are not correct with spectrum analyzer (direct method) , so we need to use frequency discriminator method to improve the results by eliminating AM sidebands. What I am not understanding is what should I look for when measuring using frequency discriminator method (with or without PLL). If I want to measure phase noise from 10kHz to 100kHz offsets, then will I have to set the center frequency at 50kHz and set the span 100kHz and get the reading? Or will I have to change the center frequency from 10kHz to 100kHz and set the span say 10kHz every time and then measure the peak voltage and divide it by span to get the value per hertz bandwidth, and repeat it for each of the center frequencies that I set?