Freuency analysis of non-periodic signals

[Ken Kearney]

Hi,

What I was wondering is if there is a way in Labview to get the frequency response i.e. power spectrum of a non-periodic signal i.e. a transient of some decription?

What I basically have is a 'spike' in a waveform happening every few seconds and I want to analyse it for high-frequency components.

Best Regards,

Ken

[SachaE]

Hi,

  you could try putting the waveform through a high pass filter, and then FFT / Power spectrum it if you have the Full or Professional version.

That way you'll remove the normal signal in theory and be left with the main signal. Another option would be to threshold away from the voltage spike (so vertically).

The frequency resolution of both of these isn't going to be great if you're acquisition rate isn't in excess of the 2x the highest frequency component however. Do you know the frequency with which the spikes are appearing? (i.e. you've said every few seconds, but do you know that more accurately so you can go looking for environment based sources).

Do you have the signal you could post back here in some suitable form?

Also which version of LabVIEW are you using?

Thanks

Sacha Emery
National Instruments (UK)

[Ken Kearney]

The transient could be anything i.e. a square pulse 5 V in amplitude and 10 ms in amplitude. These pulses could then be 1 second apart. What I would expect to see in a transient frequency response would content at 100Hz (i.e. the 10ms pulse width) and also 300, 500 Hz etc. i.e. the harmonics of the square wave. If I put it straight into a power spectrum I simply get a lot of low-frequency content and not at all what I was expecting.

Ken

[SachaE]

Hi Ken,

   what rate are you acquiring at?

Can you post the captured waveform back here please?

Which version of LabVIEW are you using?

Thanks

Sacha Emery
National Instruments (UK)