AutoPowerSpectrum

[Rod Griffin]

I am wondering if anyone could tell me what the accuracy of the FreqPeak in
the PowerFrequencyEstimate function is?

We are developing a vibration analysis system that uses a pcmica card with
it own microprocessor.


Problem example:
Collect 1024 points of a 20 Hz sinusoid ( produced by a function generator)
and give results to AutoPowerSpectrum function , calculate
PowerFrequencyEstimate reset the sample rate based on FreqPeak and collect
data again. The results are correct and the maximum power is in the
appropriate bin on the power spectrum graph.

Set the frequency to 20.8 Hz and the result is a high percentage of power on
one bin up and down. The frequency result can vary form .02 Hz to .75 Hz of
the actual frequency set by the generator.

Each time I
run the program without changing the generator I get variations
as stated above. Any comments would be appreciated.


Rod

[Jeremiah Cox]

Hello Rod,
The accuracy of the PowerFrequencyEstimate() function is limited by the data and parameters we provide to it. It really just interpolates the power spectrum data supplied based upon its other parameters (span, df). Inspection of your raw data and power spectrum along with information on your data acquisition configuration (sample rate and so forth) should help us to isolate the source of variations in the peak returned by PowerFrequencyEstimate(). A change in the sample size, sample rate, or the use of a filter may be necessary. You may want to repost your question with this additional information in the Multifunction IO (DAQ) forum, as they likely could provide increased feedback.

Jeremiah Cox
Applications Engineer
National Instruments
http://ww
w.ni.com/ask