I have a DUT to measure on a spectrum analyzer. The noise floor is -100 dbm. My signal has a bandwidth of 10mhz.

I want to measure the amplitude of my signal and I measure the signal amplitude to be -97dbm over the entire 10mhz bandwidth.

The noise floor is still -100dbm.

Is the amplitide of my signal then 3dbm?

  • \$\begingroup\$ That must have been a noisy measurement so close to the noise floor. Could that -97 dBm measurement be unsteady? Averaging measurements on a logarithmic scale is prone to error. \$\endgroup\$
    – glen_geek
    May 15 at 17:07
  • 2
    \$\begingroup\$ 10 millihertz? Are you sure? (I think you mean megahertz, so MHz, not millihertz, mHz. Capitalization matters, a lot! It's also dB, not db.) \$\endgroup\$ May 15 at 17:41
  • \$\begingroup\$ 10 mhz bandwidth? Which spectrum analyzer goes that far down? \$\endgroup\$
    – MrGerber
    May 15 at 17:41

1 Answer 1


Is the amplitide of my signal then 3dbm?

No, dBm is a power, so, it's not an amplitude. It's "decibel of milliwatt"!

The noise floor is -100 dbm

Are you sure? Noise floors are typically, in continuous-time systems, given as a spectral power density, which means as power per bandwidth; one typical unit is dBm/Hz. "dBm" is a power; it's very unusual to know the total noise power over an infinite bandwidth.

I measure the signal amplitude to be -97dbm over the entire 10mhz bandwidth.

That makes no sense; measuring something over milli- or megahertz (I guess you mean megahertz, MHz, not millihertz, mHz) only makes sense when it's a "per bandwidth" size.

Note that your spectrum analyzer might actually show you dBm, not dBm/Hz, but you need to take into account that this would be the power density integrated over the resolution bandwidth filter passband bandwidth. So, if you, for example, have set your spectrum analyzer to a resolution bandwidth of 100 kHz, then it showing your -97 dBm for any point on the graph means that it sees -97 dBm in the 100 kHz bandwidth applied around that frequency, so -97 dBm / (100 kHz) = -97 dBm / (50 dBHz) = -147 dBm/Hz; multiplying that with a 10 MHz width over which you see that value, you get -77 dBm as signal power.

But it really boils down to knowing what your units actually are, and knowing to what you've configured your spectrum analyzer! There's good introductions to spectrum analyzer practice and theory online, you should definitely read or watch some!


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.