peak search problem

[tchendu]

I am using peak search function with power spectrum function. But the found peak is not right at the point on the dataview (see the figure). Can someboby tell me is there any problem with the results?

Thanks,

Du

[dsb@NI]

The energy of the tone spreads out (leaks) into neighboring bins when the tone frequency is not periodic within the analysis block. The spectrum peak search identified the true frequency and amplitude of the tone in the input signal. It does not do simple bin picking of the spectrum which is limited by the frequency resolution of the spectrum. If you are in control of the input signal, you can verify that the computed result is correct. If you are processing an acquired signal, try changing the input block size so that 12 Hz is periodic (on bin). A tone is 'on bin' if the frequency is a multiple of the spectrum resolution (spectrum resolution = sample rate / block size).

The maximum difference between the spectrum magnitude and the true tone amplitude is a property of the window used in the frequency analysis. This property is known as the scallop loss.

refer to http://zone.ni.com/reference/en-XX/help/372416C-01/TOC11.htm

In the end, use the Peak Search Step in order to correctly identify spectrum peaks with unknown frequency without having to change the input block size to the Power Spectrum.