how can I measure electrical noise if the average comes out at ZERO?

Question

Can an oscilloscope or a spectrum analyser measure random noise that averages out at zero?

Surely, an oscilloscope or spectrum analyser has to do its measurement over at certain interval of time during which the error will average out to zero and become impossible to measure?!

So I have +ve and -ve errors that average out at zero but there's still gonna be a standard deviation for all those errors.

am I missing something?

Answer 1

An oscillosope does not measure anything (assuming not a digital scope)



It presents a realtime display of voltage against time



You do the measuring



If you want to measure noise on a dc signal - then set the coupling ot DCm set the display to the centre of the screen and crank up the gain (you may have to recentre)

This will show the noise on the signal

You can then measure the noise as Peak to Peak using the graticule and the gain and then converrt to rms

Answer 2

The energy in the noise is proportional to the square of the voltage. If you plot that, it's always positive.

RMS voltage is the square root of the mean (averaged) square of the voltage, and that's what engineers use most of the time because mostly people are interested in power levels, e.g. how hot an overn will get hot how much will their electric bill be..



Digital oscilloscopes bandwidth-limit the input to less than the sampling rate to avoid aliasing and sampling errors. That means that there are multiple measurements in each phase of the waveform. A digital scope can then compute RMS or any other function (maximum, minimum voltage etc.)

Answer 3

its an energy measurement and if the charge involved is positive or negative is irrelevant. You just take the area under the curve on both sides of the datum and add them together. And Yes you are missing something and its not your appendix. Of course the datum is not necessartily a straight line axis at zero. If you are measuring the noise in a sinewave then the sinewave itself becomes the datum. To calculate the noise as a percentage you would calculate the area under the curves on either side of the datum and divide that by the area under the curve of the sinewave. Or did I just dream that?





OK heres an idea of how you might measure noise with an oscilloscope.



You take the original un corrupted signal and set that as one trace

Set the amplified signal with noise as the other trace and invert it

set the scales so that the signals are equal amplitude on the oscilloscope

Combine the traces and whats left is the noise arrayed along a straight line axis



I don't know if you can do that with an old school scope but with my mark1 digital dreameduposcope its easy

Answer 4

The average is zero. But the power or square of voltage is not zero. Noise is squared, and then averaged to measure noise.

For that matter average of ac is also zero. It is measured also by squaring it and finding the average.

Answer 5

All noise averages out as zero. All AC waveforms, for that matter, average out as zero.



Use RMS voltage readings for a more meaningful result.



.

Answer 6

You need something called an Ostrichilloscope I hope this helps as you need help :)