FFT of discrete time domain data

Dear all,

I have vibration data from a setup in time domain. I need to convert into frequency domain,for which i have coded as below

n=length(data);
ts=t(2)-t(1);
p=abs(fft(data));
p=p(1:n/2).^2;
freq=[1:n/2]/ts;
figure(2)
plot(freq,p)

But the frequency components obtained in the plot, are not matching with actual ones,which i think the error in frequency axis of fft plot.Is any normalisation of axis needed . how to rectify this error .

Regards.

Answer 1

[Arnaud Miege]

I would suggest following the example in the fft documentation:

n = length(data); 
Ts = t(2)-t(1); % sample time
Fs = 1/Ts; % sampling frequency
NFFT = 2^nextpow2(n); % Next power of 2 from length of data
Y = fft(data,NFFT)/n;
f = Fs/2*linspace(0,1,NFFT/2+1);

% Plot single-sided amplitude spectrum.
plot(f,2*abs(Y(1:NFFT/2+1))) 
title('Single-Sided Amplitude Spectrum of y(t)')
xlabel('Frequency (Hz)')
ylabel('|Y(f)|')

HTH,

Arnaud

[shravankumar]

Thank you Arnaud. that helped.

Regards.

[Ella]

Hi Shravankumar do you get the same amplitude of vibration as your time domain ? I tried using the same code to get FFT of my signal but the amplitude drops. How can I solve this ? your suggestion is appreciated

[Imantha Ahangama]

@Ella, Try

data = data - mean(data);

before typing the above code.