FFT of discrete time domain data

Hello,

I would appreciate if anyone could help me to figure out the following problem.

I have 512 samples, the samples are taken within 1 microseconds of time interval. I would like to see the signal from 6 microseconds to 5.17e-4 seconds in frequency domain.

I have already tried the following codes in MATLAB, but it only gives me the time-domain representation of signal.

T = 5.17e-4;
N=512;
t=linspace(6e-6,T,N);
X=fft(data);
X = X/N;
dt = t(2)-t(1);
fs=1/dt;
fn=fs/2;
f=linspace(0,fn,length(t)/2);
X1=X(1:length(t)/2);
X1=abs(X1);
plot(f,X1);

I look forward for your reply. Thanks.

[Rick Rosson]

1. Can you tell me the size of |data| and what type of signal it represents?

2. Why are you calling the |fft| function twice? Why not just once?

[PK]

1. the size of the data is 512, and it represents discrete time domain continuous signal. That means the samples are taken after 6e-6 sec at each 1e-6 sec time interval.

2. the fft function is called only once at line 4

I have tried plot(abs(x)); it still does not work

Answer 1

[Rick Rosson]

   figure;
   plot(abs(X));

[PK]

it does not work.

Answer 2

[Rick Rosson]

In the following code, I am assuming that data is predefined as a column vector of size N x 1. If it is a row vector of size 1 x N, then either transpose data or change the first line of the code to take the number of columns instead of the number of rows.

   %%Number of samples:
   N = size(data,1);        % assumes data is an N x 1 column vector

   %%Compute the time domain:
   Fs = 1e6;                % samples per second
   dt = 1/Fs;               % seconds
   t = dt*(0:N-1)';
   T = N*dt;

   %%two-sided spectrum, centered on DC
   X = fftshift(fft(data))/N;     

   %%Compute the frequency domain:
   dF = Fs/N;
   f = (-Fs/2:dF:Fs/2-dF)';

   %%Plot the time domain signal:
   figure;
   plot(t,data);

   %%Plot the magnitude response:
   figure;
   plot(f,abs(X));

HTH.

Best,

Rick