Dear Nima

I suggest you use the following program steps, but first you must have a data vector x(n) sampled at Ts, time period from n to (n-1);

%(( Example of plotting sampled sin wave at Fs=20Hz (i.e. Ts=0.05sec) ))%

%% [Note: the length of x must equal to the length of t] %%

Fs=20; % sampling freq. Fs=20Hz

t=0:0.05:10; % Ts=1/Fs each step=50msec (the no. of t vector must be

% equal to the imported data length)

time=t'; % time= transpose(t);

y=5/(max).*x; % max=the maximum value of the imported data (x)

subplot(4,1,1);plot(time,y)

title('Sin wave 1Hz');

ylabel('Amplitude');

xlabel('time index----->N');

L=length(y); % L is the no. of points or the no. of reading

NFFT=2^nextpow2(L); % make the sequence length equal to 2n

% (i.e. 256 , 512 ,1024 , 2048 ……)

b=fft(y,NFFT); % determine the fft of y

z=abs(b);

freq=Fs/2*linspace(0,1,NFFT/2+1); % freq=0…..Fs/2

n=length(freq);

subplot(4,1,2);plot(freq,z(1:n))

Good luck

You have an FFT order in mat lab .

You have to keep in mind that for N points of x(n) you have N points FFT.

Marwan Abdulkhaleq Al-Yoonus

In this expression, y=5/(max).*x

what is 5 exactly?

And in this "freq=Fs/2*linspace(0,1,NFFT/2+1)" line, why Fs has been divided by 2?

I think it should be like this, "freq=Fs/L*linspace(0,Fs,Fs)"

I would be grateful if you answer my questions.

Thanks.

Dear Nima

1) y=5/(max).*x ; It is used to normalize the reading data to a specific rang (i.e. from 0 to 5)

you can skip this expression.

2) "freq=Fs/2*linspace(0,1,NFFT/2+1)" line, why Fs has been divided by 2?

It is familiar to plot the frequency domain from 0 to Fs/2.

see the attached file.

Marwan Abdulkhaleq Al-Yoonus

Thank you for your explanation. Now to return to time domain after applying some filtering methods to remove noise, is same code applicable to return from frequency domain to time domain?

Fs= 0:125/16383:125; %Fs=125/16383

t=131.064; %sampling time, t=131.064 s

freq=Fs'; %freq=transpose(Fs)

max(Inputa); % max=the maximum value of the imported data

subplot(2,1,1);plot(freq,Inputa)

title('acceleration-frequency')

L=length(Inputa) % L is the no. of points or the no. of reading

NFFT=2^nextpow2(L) %make the sequence length equal to 2n

b=ifft(Inputa,NFFT) %determine the fft of Inputa

z=abs(b)

time=t/L*linspace(0,L,L)

n=length(time)

subplot(2,1,2),plot(time,z(1:n))

title('acceleration-time')

In this code I have changed some expressions, for instance I have used ifft function to return to time domain. Do you think it is true?

Kind Regards.