Phase shift in in FFT Harmonics results

Hi Guys,

I am using the fft_powerscope function to evalute a simulink output signal. My problem is, that the recreated signal does not ressemble the original signal. There seems to be a phase shift somewhere. Can you help me find the error? I shared all the data that I used.

order= z_46.fft_iabc.freq./z_46.fft_iabc.fundamental;

phase= z_46.fft_iabc.phase;

for y=1:length(phase)

phase2(y,1)=phase(y)-phase(2)*(y-1);

end

fs = 1000;

Ts = 1/fs;

duration = 10;

t = 0 : Ts : duration-Ts;

mag= z_46.fft_iabc.mag./ z_46.fft_iabc.magFundamental;

x=mag(1);

M=[order,mag,phase2];

phase=phase*pi/180;

for q=2:length(M)

x=x+mag(q)*sin(2*pi* (q-1)*t+phase(q));

end

Would gladly appreciate your help. The fft start time from the orginal signal is in the data set.

Answer 1

[Pratyush Swain]

Hi Dom,

You can try with following workarounds to correct your signal:

1.Here you are adjusting the phase linearly based on phase(2), which could be distorting the phase across harmonics. This may not be necessary and could be introducing a cumulative shift. Consider using the original phase values directly.

for y=1:length(phase)
    phase2(y, 1)=phase(y)-phase(2)*(y-1);
end

2. Consider this revised approach to reconstruct the signal. Use 'order(q)' instead of 'q-1' , the harmonic number might not simply be 'q-1', so try:

x = mag(1);
for q=2:length(M)
    x = x + mag(q)*sin(2*pi*order(q)*t + phase(q));
end