Links

Tools

Export citation

Search in Google Scholar

Step-down switched-capacitor DC-DC converter for system-on-chip applications

Published in 2016 by Giovanni Perini
This paper was not found in any repository; the policy of its publisher is unknown or unclear.
This paper was not found in any repository; the policy of its publisher is unknown or unclear.

Full text: Unavailable

Question mark in circle
Preprint: policy unknown
Question mark in circle
Postprint: policy unknown
Question mark in circle
Published version: policy unknown

Abstract

This thesis presents a fully integrated switched capacitor DC/DC converter for System on Chip applications, designed in 40nm CMOS process. The converter is designed for a 2.5 ÷ 4.5V input voltage, a 0.5 ÷ 1.5V output voltage and a 500n÷1.5mA output current. A system efficiency grater than linear regulator’s is achieved for every loading condition. A review of losses mechanism in basic a charge pump circuit is provided and a generic equivalent model is introduced. With the simplified model the charge-pump is represented as an ideal transformer followed by an equivalent output resistance. In particular, charge vector analysis is reviewed and used to evaluate the output resistance of different charge pumps circuits and hence their performances. A specific charge-pump topology is chosen that meets our technological requirements in terms of switches maximum voltage rating. The charge-pump’s driving circuits are designed to properly drive the switches using a 2V externally provided supply voltage. Capacitors layout is chosen to mini- mize parasitic losses and to achieve the highest capacitive density at the same time. Later, the performances of the charge pump are investigated and the degrees of freedom (DoF) - including switch size and operative frequency - are chosen to improve system efficiency. A double control loop is designed, providing a coarse and a fine regulation: the inner loop adjusts the effective frequency of the charge-pump’s driving circuit so that output regulation is achieved, while the coarse feedback loop changes the charge-pump’s configuration to track peak efficiency depending on the loading conditions. For this purpose, a digital block is designed for loading condition detection, making possible to chose the best system configuration with no trial and error approach. Both control loops are first simulated using Verilog-A blocks and the specifications of each block are defined to meet the required performances of the converter. In particular, the bandwidth of the inner regulation-loop’s comparator is designed to meet output ripple specification and a total biasing current budget is set to achieve higher than LDO efficiency in sub − 1μW output power. Later, all analog blocks of the control loop are designed and simulated in Cadence suite, over a variety of temperatures, process corners and mismatches. Finally, system performances - such as efficiency, output ripple, system step-response, battery charge/discharge and reference voltage variations - are evaluated. Our converter is designed to work over a very wide range of loading conditions, from sub-μW to few mW of output power, achieving a 70% peak efficiency. Remarkable performances are achieved under very-light load condition thanks to a particularly optimized static biasing of the analog blocks.