[time-nuts] Ultra low phase noise floor measurement system for RF devices.

[Dr Bruce Griffiths]

SAIDJACK at aol.com wrote:
>  
> In a message dated 3/30/2007 19:14:31 Pacific Daylight Time,  
> bruce.griffiths at xtra.co.nz writes:
>
> With a  +13dBm carrier this corresponds to a noise voltage of 100pV/rtHz 
> across a  50 ohm resistor or a noise current of 2pA/rtHz  flowing through  it.
>
>
>
> Hi Bruce,
>  
> but isn't the thermal noise at room temp about 130pV/rtHz or so for a 1 ohm  
> source resistor by itself?
>  
> Wouldn't the source resistance's thermal noise be the limiting  factor?
>  
> bye,
> Said
>
>
>
> ************************************** See what's free at http://www.aol.com.
> _______________________________________________
> time-nuts mailing list
> time-nuts at febo.com
> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>
>   
Said

No, with a correlation system only signals common to both channels 
appear at the output.
Thermal noise and other noise in each channel are statistically 
independent and their product averages to zero.
The actual residual decreases as the number of spectra averaged increases.

In Johnson Noise thermometry 2 amplifiers amplify the thermal noise of a 
single resistor.
When their outputs are cross correlated only the input resistor noise 
which is common to both channels remains.
Thus amplifiers with input noise voltages greater than the thermal noise 
of the resistor can be used to make accurate temperature measurements 
using this method.

It should be possible to illustrate this by using a pair of high gain 
amplifiers, one for each channel of a PC sound card.
If each amplifier (gain = 80dB or so) looks is connected to a 50 ohm 
resistor and the amplifier input noise is 3 nV/rtHz, then it is possible 
to accurately determine the noise spectrum of the 50 ohm resistor if 
enough averages of the cross spectrum are taken.


Bruce