At 4/24/2007 09:50 AM, you wrote: >Could anybody clarify how one does lower the noise floor of the >analyzer? It was my understanding that the noise floor is a intrinsic >characteristic of the instrument itself, and is so to say a >measurement limit that cannot be varied without external aids (or >maybe with a low-noise LNA?). But if one uses an external amplifier, >wouldn't this also raise the site noise floor on the analyzer screen?
Not if the SA is self-noise-limited. For example, for a 10 kHz resolution bandwidth, the noise floor of a typical environment @ UHF is going to be somewhere around k*290*10000, or ~134 dBm. If your spectrum analyzer's noise floor at that resolution BW is -110 dBm, then you need at least 24 dB of gain ahead of its input to see down to that level. Fortunately, you don't need weak-signal DXer-grade sensitivity, so the 16 dB or so you get from an Angle Linear PHEMT LNA should be sufficient. Just subtract the gain of the added LNA from the display reading & you've got your amplitude. As others have pointed out, you can also reduce your resolution bandwidth, but once you reduce it to less than the bandwidth of the signals you're looking for (10 kHz for NBFM, already pushing the limit a bit), the signals of interest are reduced as well as the noise. You'll still see unmodulated or weakly modulated carriers at resolution BWs down to Hz, but when modulated they may disappear into the noise. >Or if I am wrong, how could one lower the noise floor of the >measurement in order to be able to take measurements at lower levels? >Or is the noise floor also a function of the SITE noise level per se? I'm familiar with quite a few comm. sites in SoCal & know of only one that seems to have a minor broadband noise problem @ UHF. If there's a lot of RF at the site you're investigating, you might need a 2 MHz window filter in front of the LNA. A cavity filter with low-loss loops may also just cover the 2 MHz span you need, though there will be a little attenuation at the edges. Bob NO6B
