How does Vrms of raw noise signal compare to power spectral density?

I am using the method described in the link below to take my real world noise data, perform an fft on it, and change it into a power spectrum.

http://www.mathworks.com/help/signal/ug/psd-estimate-using-fft.html

To verify that it's working properly, I'm trying to sum all points in the power spectrum and compare that number to the average power of the real, time based signal.

Is this the correct way of thinking about this problem? What do I need to do to get the results I need?

Thanks in advance.

Answer 1

[Jeremy]

The square root of the integral of the PSD will equal the rms amplitude of the original time history. The integral of the PSD is the sum of the PSD multiplied by the bin width, or just the sum if you are using 1 Hz bins.