IFFT usage from a frequency signal to a time domain

Hi,

I know that there are a lot of questions about this function, but I have not found an answer that can fit to me.

I have a signal in frequency domain. I have N=2001 samples, the maximum frequency of interest is f_max=10 GHz. I have also the corresponding signal in time domain.

What I want is

• to check if, performing the ifft of my frequency signal, I get a signal equal to my original time domain one.
• to perform the fft transform again, in order to check if I get back my original frequency signal.

Nor ifft neither fft show me the results I expect.

My signal is stored in a 3D matrix.

I am doing like :

 for i=1:row
        for j=1:row

            Fs=2*max(freq); %sampling frequency
            Ts=1/Fs; %time sampling
            dF=freq(2)-freq(1);

            x_aux(i,j,:)=[squeeze(X(i,j,:))' fliplr(conj(squeeze(X(i,j,2:end-1))))'];
            nfft=2^nextpow2(Laux);

            x_ifft1(i,j,:)=real(ifft(x_aux(i,j,:),nfft))*Laux*dF;

            x_ifft(i,j,:)=x_ifft1(i,j,1:Laux);

            t_ifft=(0:Laux-1)*Ts;%/Laux;

            Lifft=length(x_ifft1(i,j,:));
            NFFT=2^nextpow2(Lifft);
            dt=t_ifft(2)-t_ifft(1);
            fs=1/dt;
            time_window=t_ifft(end)-t_ifft(1);
            ncampt=size(t_ifft,2);

            X_FFT1(i,j,:)=fft(x_ifft1(i,j,:),NFFT)*(dt/Lifft);
            X_FFT(i,j,:)=(X_FFT1(i,j,1:1+ncampt/2));
            f_fft=(0:ncampt/2)*(1/time_window); 
end
end

Can somebody explain what is wrong in this code?

Best