Energy processing circuits for low-power applications

• Main Author: Ramadass, Yogesh Kumar
• Other Authors: Anantha P. Chandrakasan.
• Format: Others
• Language: English
• Published: Massachusetts Institute of Technology 2011
• Subjects: Electrical Engineering and Computer Science.
• Online Access: http://hdl.handle.net/1721.1/63026

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2009. === Cataloged from PDF version of thesis. === Includes bibliographical references (p. 199-205). === Portable electronics have fueled the rich emergence of new applications including multi-media handsets, ubiquitous smart sensors and actuators, and wearable or implantable biomedical devices. New ultra-low power circuit techniques are constantly being proposed to further improve the energy efficiency of electronic circuits. A critical part of these energy conscious systems are the energy processing and power delivery circuits that interface with the energy sources and provide conditioned voltage and current levels to the load circuits. These energy processing circuits must maintain high efficiency and reduce component count for the final solution to be attractive from an energy, size and cost perspective. The first part of this work focuses on the development of on-chip voltage scalable switched capacitor DC-DC converters in digital CMOS processes. The converters are designed to deliver regulated scalable load voltages from 0.3V up to the battery voltage of 1.2V for ultra-dynamic voltage scaled systems. The efficiency limiting mechanisms of these on-chip DC-DC converters are analyzed and digital circuit techniques are proposed to tackle these losses. Measurement results from 3 test-chips implemented in 0.18pm and 65nm CMOS processes will be provided. The converters are able to maintain >75% efficiency over a wide range of load voltage and power levels while delivering load currents up to 8mA. An embedded switched capacitor DC-DC converter that acts as the power delivery unit in a 65nm subthreshold microcontroller system will be described. The remainder of the thesis deals with energy management circuits for battery-less systems. Harvesting ambient vibrational, light or thermal energy holds much promise in realizing the goal of a self-powered system. The second part of the thesis identifies problems with commonly used interface circuits for piezoelectric vibration energy harvesters and proposes a rectifier design that gives more than 4X improvement in output power extracted from the piezoelectric energy harvester. The rectifier designs are demonstrated with the help of a test-chip built in a 0.35pm CMOS process. The inductor used within the rectifier is shared efficiently with a multitude of DC-DC converters in the energy harvesting chip leading to a compact, cost-efficient solution. The DC-DC converters designed as part of a complete power management solution achieve efficiencies of greater than 85% even in the micro-watt power levels output by the harvester. The final part of the thesis deals with thermal energy harvesters to extract electrical power from body heat. Thermal harvesters in body-worn applications output ultra-low voltages of the order of 10's of milli-volts. This presents extreme challenges to CMOS circuits that are powered by the harvester. The final part of the thesis presents a new startup technique that allows CMOS circuits to interface directly with and extract power out of thermoelectric generators without the need for an external battery, clock or reference generators. The mechanically assisted startup circuit is demonstrated with the help of a test-chip built in a 0.35pm CMOS process and can work from as low as 35mV. This enables load circuits like processors and radios to operate directly of the thermoelectric generator without the aid of a battery. A complete power management solution is provided that can extract electrical power efficiently from the harvester independent of the input voltage conditions. With the help of closed-loop control techniques, the energy processing circuit is able to maintain efficiency over a wide range of load voltage and process variations. === by Yogesh Kumar Ramadass. === Ph.D.