lorentzian curve fitting on power spectrum

[rom9]

Hi

I am having trouble analysing Power Spectrum and Power Spectrum Density. I am measuring signals from a channel from which I need to get the squaredVrms values which is obtained by intergrating the area under the curve of a Lorentzian Peak from the Power Spectrum. I tried using both Power Spectrum and Power Spectrum density from the Express VIs. However as I increase the number of samples from say 5000 samples to 20000 samples at 50KHz, the area under the resonance peak at roughly 15kHz keeps decreasing leading to a different value everytime I change the sampling rate. Is there an explanation to this ? I kept all the other factors like number of averages etc constant.

Would really appreciate the help

Thanks

[Anders Björk]

Could you give an example how much less i precent it gets? I think it has to do with the scaling the y-axis of the powerspectra, you could look it up in DSP book.

[rom9]

QUOTE (Anders Björk @ Nov 10 2008, 10:52 PM)

Could you give an example how much less i precent it gets? I think it has to do with the scaling the y-axis of the powerspectra, you could look it up in DSP book.

Hi Anders

Thanks for taking your time.

The value changes from 2E-4 at 5000 samples to 8E-5 at 10000 samples, thats an order of magnitude in change. This is what bother me, since if it was only a few percent, i could attribute it to noise or the curve fitting. I used Igor and Origin both to do the curve fitting. I also tried to use power spectral density VI instead of the power spectrum, but got similar results.