Averaged Behavior Model of Current-Mode Buck Converters for Transient Power Noise Analysis

Accurate evaluation and simulation of power noise is critical in the development of modern electronic devices. However, the widely used target impedance fails to predict the low-frequency noise generated in a device due to the existence of the dc–dc converter, whose output impedance can change under different loading conditions. A physical circuit model is then desired to replicate the behavior of a voltage regulator module, and the average technique is an efficient method to estimate the noise of a pulsewidth-modulated (PWM) converter. With the emergence of converters with adaptive on-time (AOT) controllers, more complex averaging methods are required, but none of them supports transient simulation. A general, efficient, and accurate modeling technique is presented in this article, whose framework supports both current-mode PWM and AOT controllers. In addition, a novel two-step parameter extraction method is proposed, which can be used to evaluate the equivalent values of internal feedback parameters of an encrypted simulation model or from measurement. The modeling method is validated by both simulation and measurement.

I. Introduction

Design and optimization for power distribution networks (PDNs) are critical for the state-of-the-art applications, such as laptops and smartphones. The PDN is designed to maintain a constant supply voltage for the chips and keep it within a narrow tolerance band [1], [2], [3]. The demand for low-voltage operation of a high-speed digital interface is increasing due to the faster logic transition [4], [5]; however, the noise margin is also compromised. Evaluating the fluctuation of the power rail voltage under different loading conditions is increasingly important.