An optimization algorithm to design fast and robust analog controls for Buck converters

Abstract

Ripple-based controls are popular to achieve a fast dynamic response, but the design of these controls assuring robustness is not easy due to its intrinsic nonlinear nature. These techniques often rely on sensing networks heavily dependent on parasitic elements to estimate the inductor current or the capacitor current. Consequently, a modeling technique that takes into account these sensors and parasitic elements is needed. This paper proposes an optimization algorithm to design a wide variety of controls that can take into account the parasitic elements and tolerances of the converter. The proposed algorithm can be used to design very fast controls that are also robust in a real-world implementation. This algorithm is verified on a 300kHz Buck converter with Voltage mode and a 1.3MHz Buck converter with V2Ic control.