Noise is of stochastic nature ...
There is no unique truth.
common ways:
- capture a declaired number of samples and make a histogram
- measure Peak to Peak Noise (IIRR about 3x RMS noise)
I had a look at the NI5922 calibration procedure, an adaption:
Connect a imput impedance matching your task, (1M or 50Ohm resistor ?)
configure your Task (for every SR used)
capture 65536 samples
FFT
and start with the freq. bin depending of freq range and SR choosen. FFT bin 0 is the DC and should always be ignored. (That's offset 😉 )
Document the way you have choosen together with the result.
Greetings from Germany
Henrik
LV since v3.1
“ground” is a convenient fantasy
'˙˙˙˙uıɐƃɐ lɐıp puɐ °06 ǝuoɥd ɹnoʎ uɹnʇ ǝsɐǝld 'ʎɹɐuıƃɐɯı sı pǝlɐıp ǝʌɐɥ noʎ ɹǝqɯnu ǝɥʇ'