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

Magnus Danielson cfmd at bredband.net
Sat Mar 31 07:48:53 EDT 2007


From: SAIDJACK at aol.com
Subject: Re: [time-nuts] Ultra low phase noise floor measurement system for RF devices.
Date: Fri, 30 Mar 2007 22:04:05 EDT
Message-ID: <cd2.e102198.333f1b95 at aol.com>

>  
> In a message dated 3/30/2007 12:26:46 Pacific Daylight Time,  
> cfmd at bredband.net writes:
> 
> Until  I've actually read the paper, ask yourself how our GPS receivers is 
> able
> to  track signals below their internal noise.
> 
> There are a number of tricky  aspects involved in doing that and I beleive the
> paper goes into those. I  have it in print here and will read it as I come  
> back
> home.
> 
> Cheers,
> Magnus
> 
> 
> 
> Weekend reading for me too :)
>  
> But GPS uses highly correlated 'noise' which allows the payload  data to be 
> below the noise floor through simple correlation since the noise  is expected 
> to be random.

This was my point. Correlation. That gets you below the noisefloor since you
can average out the noise.

> Just wondering about the numbers, -200dBc/Hz is 10 Billion times smaller  
> than the carrier, and that's in terms of voltage leves - power is that number  
> squared...

10 Miljard times below the carrier you mean! (I'm nitpicking you, billion to
mean 1E9 is a US feature, here in Europe we have miljard to mean the same
thing, a billion is 1E12 so... this is why you just can't use them across the
pond at all)

I agree, it is indeed a fantastic number, but you have to compare it with
other results. There is comercial tools out there which is only 20-30 dB
behind. Actually, the correlation techniques improves the longer you average,
so then stability of setup and such becomes the hurdles, without looking I
would guess they have put some thought into that. Normal crosscorrelation
techniques looks like crude conceptual setups compared to this more elaborate
setup.

I'm looking forward to read about this.

Cheers,
Magnus



More information about the time-nuts mailing list