How to measure noise amplitude on the required frequency

[BPS]

Hello,

I want to measure noise amplitude(RMS) of the resistor. Noise voltage measures with DAQ. Then noise specrum calculates with Spectral Measurements (FFT without avereging) in express VIs.  I don't understand why noise amplitude decreases with samples number increase on the same rate. And how to choose rate and number of samles (for DAQ) to obtain noise amplitude in agreement with Nyquist formula (U^2=4kTRdf).

[Asad_Tirmizi]

What window have u selected Hanning ?

[Asad_Tirmizi]

As far as the noise amp question is concerned. sampling rate should be atleast double the frequency although i personally prefer a higher multiplier but number of sample is purely a matter of convenience. Are u using Elvis for acquiring signal ? Kindly fill me in with a little more detail about the experiment and the appratus as ive played with such VI's for some time now and maybe i can help u with whatever little i know.

[HeatherBaker]

Here is a link to a good overview of analog sampling basics.  We recommend a sampling frequency that is 10x the number of samples you acquire.  Asad is correct that we need a lot more information about your hardware and overall set up to provide more help.

[BPS]

I measure thermal noise voltage of different resistors. I use 2 channels of NI USB-6221 in differential mode. I acquire signal with DAQ Assistant vi. Then noise specrum (RMS) calculates with Spectral Measurements vi (FFT without avereging and without windows). I observ that noise amplitude decreases with samples number increase on the same rate. The same thing I see when I change DAQ Assistant vi with Simulate signal vi(DC=0+noise). So I don't understand how to choose correct number of samples for the given rate. In my application (thermal noise voltage measurments) I set sampling frequency that is 10x the number of samples as you adviced. It works correctly (I checked the rezults with Nyquist formula: U^2=4kTRdf) but I don't understand why koefficient between rate and number of samples is 10.