Estimation of Voltage Regulator Stable Region Using Radial Basis Function Neural Network

Abstract

Disturbance to the voltage regulator (VR) output caused by the abrupt change in load current can be compensated using an output capacitor with an internal parasitic element called the equivalent series resistance (ESR). However, the ESR value changes due to aging and temperature change factors, thereby creating a VR stable region in terms of ESR. In practice, time-consuming and high-expertise manual characterization is required to characterize the VR stable region during the design and manufacturing phases. Therefore, this research aims to develop an efficient and effective VR characterization method. In this work, the radial basis function neural network (RBFNN) approach was implemented to estimate the stable region. Results show that the RBFNN approach yields a stable region with higher estimation accuracy and faster characterization time than those of manual characterization. VR characterization using the RBFNN approach can efficiently and effectively estimate the VR stable region.