noise floor in a power spectrum analyzer

Good God man! You've found a way to decrease spam bandwidth? Geez, package it and sell it, you'll make a million bucks!

;-) Rich

I hope you mean "span". I haven't heard of spectrum analyzers with email functions.

10Log(BW) should explain what you are seeing in the dB world. Where BW is the effective noise bandwidth of the analyzer IF filter (some analyzers call this resolution bandwidth). Your analyzer may automatically change the resolution bandwidth when you change the frequency span.

When measuring noise, it is common to measure the spectral density in V/(root Hertz) or dB/Hz, i.e. noise in a one Hertz bandwidth. You can figure out the total noise for a given bandwidth by adding 10Log(BW) to your spectral density number.

If you have access to a university library, try look for "Low-Noise Electronic Design", by Motchenbacher and Fitchen. It will give insight to your observation.

--
Mark

I read in sci.electronics.design that Harry wrote (in ) about 'noise floor in a power spectrum analyzer', on Wed, 21 Sep 2005:

The basic electrical noise power equation is PN = 4kTB, where PN is noise power, k is Boltzmann's constant, T is the absolute temperature and B is the bandwidth in Hz. Your spectrum analyser displays voltage levels in dB, and those voltages are measured across a fixed 50 ohm resistance.

The noise voltage VN = sqrt(4kTBR), where R is the 50 ohm resistance. If you halve the bandwidth, the voltage goes down by sqrt(0.5) = 0.71, which is 3 dB down, just as if you had done the calculation in terms of PN.

--
Regards, John Woodgate, OOO - Own Opinions Only.
If everything has been designed, a god designed evolution by natural selection.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk