Fast Fourier Transform vs Linearize

Hi,

I have a simple 5 degree of freedom Simulink model, and I want to extract the frequencies of my model. I use two ways to do this:

1- Run the Simulink model, get the time domain responses, and get the frequency response by using FFT.

2- use linearize(mdl) in MATLAB, plot Bode diagram and get the frequencies

Both of them have almost similar results, but when I use first method(FFT), my frequency response is missing some of the frequencies. While, using linearize(mdl) shows me all the frequencies.

Any information on why they are different and can someone help me with the basics here?

[Mathieu NOE]

hello

regarding point 1/ to have a fair comprasion with 2/ , it has to be a transfer function estimate based on time domain output and input signals

see fonctions tfe.m / tfestimate .m