Comment on narrow filters

[Andre' Kesteloot]

Nick (G4WHO) wrote:

> [...] I do make my
> living out of EMC measurement and trouble shooting though so have a fair idea of
> noise and signal performance,
>
> The theoretical improvements are those that would take place against random
> noise, not close spaced strong signals, so if static and atmospheric noise were
> the limiting factor then what you say is true.
>
> If on the other hand strong local signals were the problem then nose bandwidth,
> skirt shape/depth, and stop band are critical in keeping them out to be able to
> hear the ones you want and the improvement  because of a reduction in the
> adjacent signal can be much greater than the figures mentioned.
>
> When you have spectrogram on the end the actual bandwidth you are looking at the
> signal through is that of the FFT process and is MUCH narrower than the receiver
> filter.
>
> In this case the rx bandwidth is only a 'roofing filter' and, provided all the
> chunk of signal arriving is around the same signal level, the rx bandwidth is
> irrelevant. If there are strong signals coming and going within the passband
> however this will cause AC pumping and all sorts of problems associated with
> that. In this case a narrower bandwidth would help a lot if  it kept the
> unwanted strong signals out