How do i use a FFT to acquire a nice spectrum?

[Dennis_Knutson]

Attach the converted file. All you have to do is zip it up and click the Browse button by the Attachments. Few people have this toolkit and it is not reasonable to expect someone to install it and do the conversion when you have already done that.

[GG_Jialat]

I tried to zip the file but it still says out of file extension. I used winrar to zip the file. Is there any other way to make it into the available file extension?

[GG_Jialat]

Heres the attachment

[GG_Jialat]

Sorry for the inconvenient 

[ejkaiser]

Hi,

looking at the waveform you finally managed to provide, I think the FFT is providing pretty much what you can expect. The spectrum shows strong low frequency components and a lot of higher frequency noise.

You won't get a 'nice' clean spectrum with isolated frequency lines because it is not a 'nice' harmonic waveform. If you want to see the low frequency periodicity of the heart beat, you must at least apply a low pass filter to the waveform.

Cheers

[GG_Jialat]

Thanx for the reply but there are many different kind of FFT in labview. So which one should i use? 🙂

[ejkaiser]

The different FFT options refer mainly to the input and output format (waveform vs. array and real vs. complex). You must adapt it to availbale data format and intended tasks. Also the averaging and window options must be chosen according to your needs. One window provides less spectral leakage, the other provides better amplitude accracy. Difficult to give an advice.

You have the software and can try it all and decide which results fit best your needs. That's the beauty of software. All options are readily available, no need to make hardware and acquire components. And I assume there must be a wealth of scientific literature on how to process this kind of signals.

Cheers

Edgar

[GG_Jialat]

Here is my vi. In the vi i tried to use the FFT but it only display 2 line on the graph. Is it i miss understand the FFT or is it i use it in the wrong situation? need some guidance plz >.<

[ejkaiser]

You should take a closer look into the FFT. Scale the y-axis to 100 and the x-axis to 0.1 and you will see that there is a spectrum. It is covered by a high DC offset that shouild be removed from the waveform before FFT. You can do that by subtracting the average vailu of the waveform. There may even be a vi for that also.

It seems you should get yourself a little more familiar with FFT and signal processing in general. And: play with the option the software gives you.

Cheers

Edgar

[GG_Jialat]

Thanx for the info! but the x-axis is the frequecy range right?