Topology generation, optimization and design of switched-capacitor converters

Abstract

A unified and comprehensive method is proposed to generate all possible topologies of switched-capacitor converters (SCCs) with N flying capacitors that work in 2- to (N+1)-topological phases. The topologies include both step-up and step-down converters, with voltage-conversion ratios (VCRs) that are integers and rational numbers. Ladder diagrams are introduced to illustrate the operation of SCCs, and duality can be applied to generate step-down SCCs from step-up SCCs, (m/n)X SCCs from (1–m/n)X SCCs for step-down converters, and (n/m)X from (n/(n–m))X SCCs for step-up converters, and vice versa. Criteria are set to identify and eliminate redundant topologies. The exhaustive search identifies SCCs with VCR_min and VCR_max that are better than 1/2^NX and 2^NX respectively. Switched-capacitor converters are then compared based on their slow switching limit resistance (R_SSL) and fast switching limit resistance (R_FSL) that are computed systematically using matrix analysis, which could serve as efficient design guideline. Based on the analysis, a fully-integrated reconfigurable SCC that achieves very small voltage conversion ratios (VCR = 1/4X, 1/5X, 1/6X) is designed using a 65nm CMOS process. For each of the topologies, the SCC is designed to have the minimum output resistance, and the optimal design achieved 50% improvement in power density compared to a conventional Dickson SCC, which is verified by measurement results. A second 3-interleaving-phase fully-integrated 1/4X SCC is also designed. To fully utilize all capacitors, the problem of idling capacitors is solved by the novel capacitor bridging and phase expansion techniques. The drain-source flipping problem associated with reconfigurable switches is solved by a phase clamping technique. The current density is improved by 25% comparing to the traditional 3-phase 1/4X SCC without sacrificing efficiency; and the improvement is also verified by measurement results.

Date of Award	2022
Original language	English
Awarding Institution	The Hong Kong University of Science and Technology
Supervisor	Wing Hung KI (Supervisor)