FFT of signal

[esra]

Hi

Im trying to figure our the noise in a laser. My set up is Laser- Detector-Oscilloscope. I read the signal using a GPIB and Labview program. I think I need to take the FFT of the signal produced but it does not give me anything. I wired the output signal to the FFT.vi and the result is one peak at 0 (zero). I should al least see a peak around 60 Hz(room light) but I dont see it. My sampling rate is 100Ksamples/second. Also, how can I see frequency on x-axis of the FFT plot and what is the Y plot on FFT chart?

Thank you

[Somil]

use the spectral annalysis VI

[Somil]

also, on x axis is the frequency (since you are dealing fith fourier transform) and on y axis is the signal amplitude

[esra]

Thank you for your reply

Ive used the spectral analysis, Im attaching thefront panel and the block diagram. I still cannot see frequency of the rom light.(frequency numbers are changing from 0.1 to 0.4) Is that the Hz unit?? why? Again my set up is Laser- Detector and Oscilloscope and the scope is connected to the computer using a GPIB interface. 

Thank you so much

[Dennis_Knutson]

The spectral analysis function is not getting any sample rate information. All that you are passing it is a dbl array. Use the Build Waveform function to add the dt and pass that. The dt is probably something that is part of the cluster returned by the scope.

[esra]

so Im gonna build a waveform using delta T (which is the time division in the scope) and the signal (voltage reading from laser) and then Im gonna feed this into spectral analysis VI. Is that correct?

[Dennis_Knutson · Accepted solution]

Yes. Without dt information, the spectral analsyis function will assume a sample rate of 1.

[esra]

so I builded a waveform and put dt and Y (signal) and fed that into spectral analysis. Also I put a chopper on the way 100 Hz frequency to be able to see the signal. Now I should see a peak at 60 Hz, 120 Hz(room light) and the signal around 100 Hz. but I see nothing. The sample rate is 250 Ksamples/second because oscilloscope decides the sample rate. If I let the time/division to be 20 ms then to be able to collect 50,000 data points the sample rate should be 250 Ksamples/second. whatever the sample rate is the oscilloscope wants to collect 50000 data points. Could you please help?

Thank you

[esra]

Hi again

I think I figured it out, I can see the chopper frequency now, but I cannot see the room light still. But part of the problem is solved. Thank you very much

[drummerguy]

could you share what your solution was?  I'm having a similar problem.  Or does anyone have a better idea for how to take an FFT of an array of data (I have one 1D array of time data and one 1D array of the corresponding F(t) data).  I interpolated the dataset so that everything is spaced equally and then built a waveform and fed it into the spectral analysis express vi but get the same noise-like signal shown in the jpeg in the previous post.  I've also tried the regular FFT vi, but I'm not sure what the 'bin' number corresponds to - it seems really strange to me that you only input the dependent variable and only get the dependent variable out - to me it would make more sense to input the dependent variable and the independent variable and then get the transformed dependent variable plus the 'frequency' out.  Any suggestions?