Almost same estimated plants, but PID Tuner gives total different Kp/Ki/Kd values

Hello,

firstly I followed the example model: Design Controller for Boost Converter Model Using Frequency Response Data - MATLAB & Simulink - MathWorks Deutschland and everything works fine. The example model used a Fixed Sample Time Sinestream.

Then I estimated the plant with a prbs-signal of order 11. After thinning, the result looks quiet good and fit to the result with sinestream.

But why does the PID-tuner give completely different P / I and D (negative) values, when the estimated result with prbs is used as the plant ? With the tuned PID-parameters, the PID-controller could not even follow the reference value (Vref).

Second question: Why isn't there the DC (frequence 0 Hz) component in the thinned result ("estsys_order11_prbs_thinned" - frd object) ?

But in the unthinned result, there is a DC component (as a result of single-side FFT, frequence from 0 to Nyquist frequency)

Many thanks for your help!

Kind regards

Tony