How to take the engineering FFT (magnitude spectrum scaled correctly) of Signal in Matlab?

Hi All

the fft function in matlab gives a theoretical mathematical fft of the signal, and is not scaled to units and is real immaginary and not amplitude-frequency

I think in better terms, I want to plot the amplitude spectrum of the signal,, with the X axis being the frequencies of the input signal.

is there an alternative way of fft to do this ? and is the following link correct?

for example if this is my signal

using the code in the link I get

which its extremums do not meet the extremums of the time signal

[farzad]

Anyone ?

[Walter Roberson]

The link looks ok under the important restriction that the inputs are real-valued

[farzad]

well I updated my question, For the time signal I have I expected, like fft, it would give the maximums of the time signal, but it does not, and is much smaller

[Walter Roberson]

The code from that link can only emit positive frequency numbers.

[farzad]

ok but I expect to see peaks of 40 mm in fft or amplitude spectrum, why is it even smaller than 8 ?

[farzad]

any ideas from anyone

[Walter Roberson]

Data would help, along with the code you are using.

[farzad]

It was not possible to attach the signal file here I dont know why. It is an xlsx file of 2.8 MB. but does not get attached. I will add the code to the question as well. I have tried both of these. you can generate a random signal and check with it. my problem is not the signal

function [frq, amp] = simpleFFT(signal)
            
                n = length(signal); 
                 %generate the vector of frequencies
                
                halfn = floor(n / 2)+1;
                
                deltaf = 1 / ( n / ScanRate);
                
                frq = (0:(halfn-1)) * deltaf;
                z = fft(signal, n); %do the actual work
                x=signal(:,1);
                xFFT = abs(fft(x))/length(x);
                amp = abs(fft(x).^2);
                
end

function [frq, amp] = simpleFFT(signal,ScanRate)
            
                N=length(signal);
                N1=2^nextpow2(N);
                X=fft(signal,N1);
                X=X(1:N1/2);%Discard Half of Points
                X_mag=abs(X); %Magnitude Spectrum
                X_phase=phase(X); %Phase Spectrum
                frq=ScanRate*(0:N1/2-1)/N1; %Frequency axis
                amp= X_mag/N1;
end

Answer 1

[Walter Roberson]

The unit for the y axis of the discrete fft is the original unit per hertz. You need to integrate it over frequency to get the original unit.

[farzad]

Thank you , so if I integrate, what will I have ? still a magnitude spectrum ?