Removing DC component at 0 HZ from acceleration data after fft

Hello,

After i compute the FFT of my acceleration data (Fs = 50 Hz) i am getting a large spike at 0 Hz (the leftover DC component). I have tried subtarcting the mean and detrending my data (detrend(X,2)).

Here is my code:

n = length(acc(:,1));

accX_fft = fft(acc(:,1),n);

PSD = accX_fft.*conj(accX_fft)/n;

freq = linspace(0,50,length(accX_fft));

L = 1:floor(n/2);

plot(freq(L),PSD(L));

The amplitudes of the frequecieis i want to visualize are already very very small, so the spike at 0 Hz is really distorting the view.

Id appreciate any reccomendation on how to get rid of this spike.

[Ameer Hamza]

If subtracting mean does not work, then it might be some other frequency?

[Nidhi Singh]

I am having the same doubt. If subtracting the mean doesn't work, than can we use high pass filter here with fft graph?

Answer 1

[Ameer Hamza]

As described here: http://blog.originlab.com/how-to-remove-dc-offset-before-performing-fft try using a high-pass filter to remove very low frequencies from your signal. See highpass(): https://www.mathworks.com/help/releases/R2020a/signal/ref/highpass.html

[Adam Levschuk]

high pass filter worked great. Thanks for the help, Ameer.

Adam

[Ameer Hamza]

I am glad to be of help!