How can I do inverse FFT? I have only frequency domain signal.

%data load

freq_range=[50 100 150 200 250 300 350 400 450 500 ...

    550  600  650  700  750  800  850  900  950  1000  ...

    1050  1100  1150  1200  1250  1300  ...

    1350  1400  1450  1500 ...
]';

input_V=[0.044668359 0.044668359 0.044668359 0.044668359 ...

    0.044668359  0.044668359  0.044668359  0.044668359  0.044668359  ...

    0.044668359  0.044668359  0.044668359  0.044668359  0.044668359  ...

    0.044668359  0.044668359  0.044668359  0.044668359  0.044668359  ...

    0.044668359  0.044668359  0.044668359  0.044668359  0.044668359  ...

    0.044668359  0.044668359  0.044668359  0.044668359  0.044668359  0.044668359 ...

    ];

%% time setting

dt = 800e-9;

% sampling frequency

fs = 1/dt;

%Input signal-frequency

input_freq = freq_range;

input_voltage = input_V;

figure(1)

stem(input_freq,input_voltage)

ylim([0 0.1])

xlabel('Frequency [MHz]')

ylabel('Amplitude')

title('Frequency Domain')

grid minor

% zeros pad

s = length(input_voltage);

n = 2^nextpow2(s);

%Input signal-Time

dim = 1;

initial=input_voltage(1);

last=input_voltage(end);

input_voltage=input_voltage(2:end-1);

input_voltage(2:end-1)=input_voltage(2:end-1)/2;

len=length(input_voltage);

for i=1:len

    ifft_voltage(1,len-i+1)=input_voltage(1,i);

end

Y=[initial input_voltage last ifft_voltage initial]*abs(n);

y2=real(ifft(Y,n,dim))'; %denormalization

figure(100)

plot(y2);

title('time domain signal')

xlabel('time[ns]')

ylabel('|P1(t)|')

---------------------------------------------------------------------------------------------------------------------

I have only a frequency signal, and I want to convert it to a time domain signal. But I think this code is not working. Can you tell me why this code is wrong? And Can you tell me correct answer? Please, I need your help.

I want to get the exact result waveform of the IFFT of the input voltage.

[KALYAN ACHARJYA]

How do you know this code is wrong? What about y2 plot, its generates time domain signal. By the way, code is perfectly working, the result is correct or not, this is different issue.

[HYUNWOONG KIM]

The resulting waveform of the IFFT for the frequency signal is odd. y2 is the resultant waveform of the IFFT. But, I think y2 is not correct about input voltage time domain signal

[madhan ravi]

Plot of y2 argument is time right?

[HYUNWOONG KIM]

yes, I know. y2 plot is time domain signal, But it is not exactly IFFT result of input voltage signal. Do you think that y2 is correctly converted input voltage?

Answer 1

[Naman Chaturvedi]

I understand that the voltage signal you want to take IFFT of is a constant signal. If you just do

ifft(input_V);

you should get an impulse as the output. Which is fundamentally correct. But, if you want summation of different frequencies mentioned in 'freq_range' with different strengths mentioned in 'input_V', you have to perform zero padding according to the sampling period or sampling frequency.

To understand more about how the IFFT works, refer to the examples given in the link.