I measured the VCO noise again.  2 methods that I tried

1) Measuring noise from a) suppressed signal (signal from pre-amp which is fed back to IFR2023) and b) error signal ((beat signal coming out of the mixer)) and match them together

Setup for 1)

ifr2023b,  carrier f = 79.994620 Hz, power =    7.0 dBm,     FMdevn = 500kHz

mixer     Mini Circuit ZFM-3-S+   (Lo input = 7dBm, RF input = -1.05 dBm)

pre amp, SR560, Gain inv10, low pass at 1Hz

2) Lowering the unity gain frequency and measuring only the error signal (beat signal coming out of the mixer)

Setup for 2)

ifr2023b,  carrier f = 79.995 316 Hz, power =    7.0 dBm,     FMdevn = 1kHz

mixer     Mini Circuit ZFM-3-S+   (Lo input = 7dBm, RF input = -1.05 dBm)

pre amp, SR560, Gain inv20, low pass at 0.03Hz

Converting V/rHz to f/ rHz

1)The feed back signal can be convert to freq noise by using the calibration from last week, which is 0.35 MHz/V, since two setups on IFR2023 are similar.

We can obtain the calibration by giving input voltage to the LO and see how the carrier freq changes.

2) For the  error signal out of the mixer, disconnect the feedback signal and measure the slope of the signal output. That will be [V/rad] calibration.

Then we can convert V/rHz -> rad / rHz , multiply by the corresponding f to get f/rHz

 Red and Green plot are from the 1st method,

Blue plot if from the 2nd method.

The results between two methods are not even close to each other, I'll check tomorrow to see if I did something wrong,