0
\$\begingroup\$

I have a broadband noise source based on a 24V Zener diode that looks like:-

waveform

You'll notice that it's asymmetric about 0V, but let's ignore that for ease of calculation. The noise is buffered prior to sampling with a TL081 op amp that has a bandwidth product of 3MHz. I think that makes 3MHz the frequency choke point within my little circuit. The Ampl value of 1.02V is actually a peak to peak measurement.

How can I determine the RMS noise level for only the audio range?

\$\endgroup\$
  • 1
    \$\begingroup\$ If you band limit your buffer to the audio range, then the builtin RMS function of the scope is all you need. Otherwise you need to FFT and integrate the PSD for your band of interest. \$\endgroup\$ – sstobbe Jun 25 '17 at 3:30
  • \$\begingroup\$ Is zener noise actually flat? \$\endgroup\$ – Andy aka Jun 25 '17 at 9:43
  • \$\begingroup\$ @Andyaka I think we have to assume so to proceed. The FFT function on a kwality Rigol is fairly naff so can't categorically say otherwise. \$\endgroup\$ – Paul Uszak Jun 25 '17 at 11:28
1
\$\begingroup\$

You can also use Audacity to record audio signals and measure Spectral Analysis.

Using a sine wave on Line out and record on Line In or Aux you can use a sine to see \$V_{RMS}=V_{PK}/\sqrt2 ~~\$ or Vrms=Vpp/2.828 as @Barry said, then repeat with your white noise source ( asymmetric from poor input bias perhaps for CM range) It also has a white noise PRSG source and FM sine log sweep.

Depending on duration of the noise sweep or sequence of a PRSG 98% of the time the RMS value of white noise is about 25% of the peak value. while @Barry suggested Vpp=Vrms*3.5 or Vrms=28.5%

  • which implies a longer capture time (1sec) with bigger peaks), so I still agree with him., but you only show a fastsweep
\$\endgroup\$
0
\$\begingroup\$

You can get a reasonable estimate of the RMS value by dividing the peak-to-peak value by 3.5 (the factor is 2.828 for a sine wave but higher for white noise). However if you want the value only for the audio range (say 20-20000 Hz), you will either have to filter the noise or perform an FFT and estimate from that. If you think the noise is white (flat with frequency), then you could reduce the RMS value by 20log (20 kHz/3000 kHz) = 43 dB to get a value.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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