unstable peak control loop of boost converter with compensator type 2?

Question

hello,

I have designed a compensator type 2 (according to the k-factor approach)for my peak current mode control of boost converter for chosen crossover frequency of 30KHz and its transfer function is :

Gc=(2.188 s + 1.066e05)/(1.372e-06 s^2 + s) where the power stage gain is -6.8dB. so my circuit is attwached as you can see.

my problem when I run circuit using this conpensator in PLECS I get a huge oscillation of the system eventhough when the bode plot of the complete system gives a good result accorcing to its complete transfer function, I have noticed that whenever I set an initial value to the output capacitor to 9.5V my system works perfectly.So my question is my compensator is wrong which i doubt ot I must include an initial value to the capacitor thanks

Answer 1

You have it backwards: In order to calculate an Impulse Response Analysis PLECS first does a Steady State Analysis. The impulse response is then calculated for that operating point.

With your model, using an initial capacitor voltage of zero, the steady state is not reached or the steady state that is reached is not the one you are aiming for (a system may have more than one steady operating point or the steady operating point may only appear to be stable within the given tolerances). The result of the Impulse Response Analysis may be misleading in this case.

So you should fix up the model such that your controller actually allows the output voltage to be reached (hint: limiting the current to negative values may not be what you want) in a transient simulation. Then see whether the steady state analysis reaches a sensible operating point. Then do the impulse response analysis.

Comments

thank you Mr. Oliver Schwartz for your hint.

my major problem is that designing the controller. for instance, a crossover frequency of 30KHz or higher results in a non-stable closed loop. you have asked me to fix the model such that your controller actually allows the output voltage to be reached in a transient simulation and that is what I want. so, I have used the k factor approach which shows a good frequency response for the controller and the complete closed system. but implementing in  PLECS results in an oscillating system without the reaching the operating point when the initial value of the capacitor is not initialized. theoretically, the controller must stabilize my loop under any crossover frequency fc of less than fsw/15. however in my design, only the fc less than 6Khz that allows the output voltage reaches the operating point in transient simulation