0

I am wondering if anyone knows of situations where one would need to accurately estimate the power spectrum of a signal at a frequency where the power is known to be very low, much lower than at other regions of the power spectrum? Suppose the spectral dynamic range of the power spectrum was greater that 100 dB or so, is there an application that may require the accurate estimation of the low powers?

Thanks!

JMcB
  • 101
  • 1
  • You need to have a very sharp bandpass filter that has gain close to 0 dB gain for the frequency (or frequencies) of interest and then sharply transitions to -100 dB gain (or even more negative dB) for all other frequencies than those of interest. – robert bristow-johnson Aug 12 '23 at 18:30

1 Answers1

1

Yes, there are many such applications. One specific one I am intimately familiar with is in the design of atomic clocks where we require noise floors to be in far excess of the stated range by the OP.

Another one that comes up frequently is power spectral density plots (power per unit bandwidth in Hz) relative to a carrier power level such as phase noise. See the example plot below that I copied from Wikipedia. This one shows a noise floor above -100 dBc, but I have worked with oscillators and sources with phase noise profiles going significantly lower than this.

Phase Noise

Another plot from the same link:

phase noise 2

Dan Boschen
  • 50,942
  • 2
  • 57
  • 135