Does anyone have information on a possible 4dB error with the autopowerspectrum routine?

[Budo]

I'm trying to develop a PSD of some sampled data. However, I've found that the results are off by approximately 4.2dB from the results produced by other tools, e.g. LMS and other 3rd party PSD routines. After investigating for several days I've found that the autopowerspectrum routine in measurement studio produces results that are off from the expected results for a pure sine wave 1.0V peak without noise by rms (.707). If I consider 6dB to be 100% (which we do) then I find that .707 of 6dB is 4.2dB which is my original error in PSD calculations. Does anyone know why the autopowerspectrum routine would be producing results that are lower by .707 from other 3rd party power rout
ines.

Thanks,

[bilalD]

Hello

I made a small example that calculates the power spectrum, but I did not get any unexpected results. I tried it for a Sine function with 1V peak to peak because I would expect an output of 0.5 for a 1Vpp since wave since its the square of the FFT.

There is a good document that explains the process of the fft analysis ( it talks about LabVIEW and LabWINDOWS, but the API is the same for Measurement Studio))

http://zone.ni.com/devzone/conceptd.nsf/2d17d611efb58b22862567a9006ffe76/c045a890751303a6862568650061ea98?OpenDocument

I hope this helps

Bilal Durrani
NI

[Budo]

thanks