SVT 3.1 FFT Spectrum Magnitude VI output is RMS or Peak (db OFF)

[WantToGolf]

Please help with my SVT problem:

1. With db OFF, what magnitude EU value does the SVT 3.1 FFT Spectrum VI output?  More to the point, is the EU an rms value or peak value? I thought it was rms but now I'm not sure.

2. When I double integrate acceleration values to get displacement then feed the displacement waveform to this FFT Spectrum VI, I seem to get high displacement values compared to other portable analyzers in our shop. If the output of the FFT Spectrum VI is for example, mils rms (single sided) I have to convert this to peak-to-peak values. So I multiply the each spectrum value by 1.414 (to get peak) then by 2 for peak-to-peak. Is there a more accurate way to get peak-to-peak displacement values from an acceleration waveform?

thanks,

Glen

[WantToGolf]

I made a little program to answer my 1st question. I created a sine wave with 2 EU peak to peak. I passed this waveform to the FFT in question and it produced a spectrum showing approximately 0.7 EU at the sine wave frequency. So, the output of this FFT is absolutely with out a doubt rms.

Can anybody help with my 2nd question?

Message Edited by WantToGolf on 03-31-2006 10:17 AM

[Tica]

Hi Glen,

There are several FFT-based VIs in LabVIEW and the output units of each function is different.

The following table will give you a better idea of what each function does. In your case you might be better off using the FFT VI that comes with the Signal Processing toolkit since you are looking for Volts output instead of the Vrms given by the FFT Spectrum functions.

If you have doubts about the units returned on a specific function I recommend using the Context Help in LabVIEW. As shown on the image below this gives you more information about the function and the variables used/returned by it. You can open Context Help by going to Help>>Context Help in LabVIEW.

Hope this helps. Regards,

Tica T

Applications Engineer

Message Edited by Tica T on 04-04-2006 10:38 AM

[WantToGolf]

Thanks for the reply Tica T, very good information.

I'm still working on my double integration of acceleration to displacement problem. At low frequencies it is not accurate... the displacement values are much too high. My Kistler accelerometer has a 1 to 6000 Hz frequency range at +/- 5 % signal so I don't think it's an issue with my accelerometer low frequency performance. I've been reading more on the 2 methods of integration. In the SVT manuals it shows:

a) Time domain integration

b) Frequency domain integration

Right now I'm using time domain integration then using the FFT for a spectral display.

According to the SVT manual, I should be using the FFT VI on the accelration signal then use frequency domain integration to convert to displacement instead.

It will take some time to modfiy the program for frequency integration but I think it's the right thing to do.