[time-nuts] 3 corner hat for phase noise

Martyn Smith martyn at ptsyst.com
Wed Aug 26 05:25:49 EDT 2015


I asked for a 3 corner hat measurement for phase noise.

Managed to work it out myself.

In case anyone else wants to know:

First of all Timepod does do it but only for Allan Deviation.

But I wanted it for phase noise.

So if you make three phase noise measurements A vs B (call result X), A vs C 
(call result Y), B vs C (call the result Z)

So you have three results X, Y, Z in dB.

Then you do the following.

Divide the dB by 10
Antilog them
Then Osc A = (X+Y-Z)/2
Osc B = (X+Z-Y)/2
Osc C = (Y+Z-X)/2
Then log10 the above three results
Then multiply by 10 to get back to phase noise in dB.

So I made three measurements and got -115 dB, -114 dB and -113 dB

The individual phase noise of each oscillator worked out to 
be -119.72, -116.78 and -115.35

The results are only accurate when the 3 oscillators are within 3 to 5 dB of 
each other, so its just an approximation.



More information about the time-nuts mailing list