Off-line identification of digitally controlled power converters using an analog frequency response analyzer

Abstract

The use of a digital architecture in PWM controllers for point-of-load (POL) applications, together with system identification techniques, allows the development of fully automated routines for in-situ system performance optimization where controller parameters are specifically tailored to the application. In this context, this paper proposes a method for performing parametric system identification of digitally controlled power converters using a conventional analog frequency response analyzer (FRA). Nonlinearities intrinsic to the digital loop are taken into account, thus leading to accurate estimation of converter parameters. The proposed method has been verified on a digitally controlled POL with Vin=12V, Vout=1.2V, Iout=10A and fsw=400kHz for various bulk capacitor scenarios.