FFT, IFFT and phase shifting

Hi everyone, please i need some help !

i have a rectangular pulse i did the FFT it gaves me a sinc forme and i try to make a phase shifting of this pulse, but when i applicate the ifft on my code i get an horrible forme !!! here is my code :

%XXXXXXXXXXXXXXXXXXXXXXXXXX FFT with bilt in function XXXXXXXXXXXXXXXXXXXX
fs=1e+10;
ts=1/fs;
t=[0:ts:(1.6384e-6)-ts]; 
A=2;   
y=A*trapmf(t,[100e-9 105e-9 205e-9 210e-9]); 
subplot(2,1,1)
plot(t,y);
grid;
N=length(t);   %which is equal to 16384 so pow of 2
%XXXXXXXXXXXXX FFT XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
Xf=fftshift(fft(y));
f=fs.*(-N/2:N/2-1)/N;
subplot(2,1,2)
stem(f,abs(Xf)/length(y));
title('FFT');
%XXXXXXXXXXXXXXXXXXXXXXXXX PHASE SHIFT XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
 phi= -100; 
 ind = 1:numel(Xf)/2; %Postive frequencys
 Xf(ind) =abs(Xf(ind)).*exp(1i*angle(Xf(ind))-1i*phi); %// add 1i*phi
 ind = numel(Xf)/2+1:numel(Xf); %// upper half (negative frequencys)
 Xf(ind) =abs(Xf(ind)).*exp(1i*angle(Xf(ind))+1i*phi); %// subtract 1i*phi
 %XXXXXXXXXXXXXXXXXXXXXXXXXX IFFT XXXXXXXXXXXXXXXXXXX
 y1=ifft(ifftshift(Xf));
 figure(2)
 plot(t,real(y1));
 grid on;
 title('IFFT');

Answer 1

[Catalytic]

Is this what you want

Xf=fftshift(fft(y));
f=fs.*((0:N-1) -ceil((N-1)/2))/N;
subplot(2,1,2)
stem(f,abs(Xf)/length(y));
title('FFT');
%XXXXXXXXXXXXXXXXXXXXXXXXX PHASE SHIFT XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
 phi= -100; 
 Xf=Xf.*exp(+1i*phi*2*pi*f);
 %XXXXXXXXXXXXXXXXXXXXXXXXXX IFFT XXXXXXXXXXXXXXXXXXX
 y1=ifft(ifftshift(Xf));
 figure(2)
 plot(t,real(y1));
 grid on;
 title('IFFT');

[RH ZA]

Yess !!! thank you soo much !!!

a little question please ! if the phi ( phase shift) is a vector i can do the same thing ?

cause my problem is that i want to modelisate a coax cable and the phi is giving by an equation which is depend on the frequency

Thank you again

[Catalytic]

Yes, you can do that.

[RH ZA]

Thank you, but how can i get the write frequency vector make it in my equation !

[Catalytic]

You have f.

[RH ZA]

But if i applicate the whole f look at what i get ! and also my equation gives complex numbers

[Catalytic]

Who besides you would know what phi(f) is supposed to be...?

[RH ZA]

it supposed to do the the same things, just shiftting the pulse, i'm sorry for asking too musch questions !

[Catalytic]

No change to the pulse shape? But you can already do that!

[RH ZA]

No, just shiftting it, i tried that with your method but no results, should i give all the frequencys to the equations or just the half ? because with all the frequencys the equation gives complex numbers