Help doing an spectral analysis

[alvarokoko]

Hi, I am doing a research study at university using Arduino Uno, LabVIEW and an accelerometer à352C33.

The idea is to attach the accelerometer to a structure and with a hammer make the structure vibrate and then make that signal analysis with a power spectrum in order to know the natural frequencies. I have some questions:

1. - The accelerometer needs at least 18V but with 5V the signal measuring is good for what I need but the spectral analysis is not. Does this have something to do or even with 5 Volts is should show some values?
2. - I have 2 options for measuring and both don’t show correct values for the spectral analysis (with the Fourier Transform). 1à Use a point by point filter and the Fourier for point by point. 2à Use a for loop to keep some values and create an array and then use a normal filter and Fourier to evaluate the signal.

I upload 2 Vi and some captures to show you that none of those solution works and I am stuck at this point. I am happy with the signal measuring cause is sufficiently precise but I can’t get the natural frequencies.

Any suggestion will help. Thanks in advance

[alvarokoko]

No one can help?

[mikeporter]

What is wrong with the spectral analysis?

Mike...

[alvarokoko]

None of my Vis show a proper espectral análisis and dont know if it is the arduino problem or my program problem

[johnsold]

What results do you see in the Loop Rate (Hz) indicators? 

The last time I looked at the LINX Analog Read VI it was only capable of reading at 20-40 Hz.  That means that your analysis of freqeuncies would only be valid up to half that frequency (and then only if you had no aliasing). Every hammer/accelerometer system I have seen generates frequencies into the kHz range or higher.

The Nyquist criterion cannot be ignored just because it is inconvenient for your hardware.

Lynn