For single frequency only. Very high Q at 2,4 GHz? For example, if VCO have -105 dBc@10 kHz, your resonator must have 85dB attenuation at 10 kHz offset. So, it's practicaly impossible for single frequency too.biff44 said:A fair bit of misinformation here. -174 dBc/Hz thermal noise floor? Boy is THAT wrong! Check your units.
As far as measuring this, you will need to build two separate oscillators, have some means to phase lock (in a very narrow loop bandwidth) one to the other, put them in quadrature, and beat the two outputs in a mixer (acting as a phase detector). You then video amplify the mixer output and can make your measurement.
As far as actually making the oscillator, you have two choices to TRY:
1) get a very high Q cavity (maybe whispering gallery mode in saphire) and run the oscillator at a very high power level (maybe 1 watt). Look up "Lesson's equation" to see what loaded Q and oscillation power you would need.
2) have a pretty good oscillator, and use a very high Q resonator as a frequency discriminator, and use the frequency discriminator output in a feedback loop to degenerate the original oscillator's phase noise.
biff44 said:Listen, I am not going to hold a school on phase noise measurement techniques or oscillator design. Go do some reading, and stop making such outlandish assumptions about needed dynamic range, or thermal noise floors, etc.