Efficient micro-Power Management for Solar Cells with Time Domain Array Reconfiguration

Abstract

The goal of this dissertation is to demonstrate improvements to micro-power solar energy harvesting with a scalable approach. The improvements are demonstrated through a microchip designed to harvest energy optimally from a miniature, partially shaded Photovoltaic (PV) array with an adaptive, light weight, high efficiency power converter. The power converter dynamically reconfigures the array to operate each cell at its individual maximum power point using a new technique termed 'time-domain array-reconfiguration' (TDAR) to extract maximum energy from the miniature PV array. Directly harnessed solar power is an attractive primary energy source for integration into fully wire-free, self-sufficient, portable electronic devices. Photovoltaic cells are most efficient at harvesting when operated at a particular voltage and current, termed ‘maximum power point’. Density of harvested energy increases and cost per watt decreases when PV cells are operated efficiently, requiring dedicated PV management electronics. For both high power and portable applications, arraying PV cells increases voltage and current output. However, simply connecting PV cells in parallel or series leads to inefficient topologies where the weakest cell limits the output capacity of the array. The research challenge in PV power systems is thus to develop better power management techniques to boost PV efficiency. Portable PV arrays are small, low voltage, and subject to varying illumination and mechanical stress. Due to the additional constraints on portable systems, efficient PV management for portable and low-power systems presents a significant design challenge. This work addresses technical challenges of micro-power array management in the context of a portable sensor aimed at 1 gram system weight and 100 micro-watt average power. Energy is stored in a Lithium-polymer battery. Only 2-3 PV cells are available, and illumination is expected to vary quickly. The system is thus designed to efficiently boost PV voltage to charge the 3.7V-4.1V battery while tracking individual PV maximum power points, and keeping power components under 0.5 grams. Efficient analog power tracking and TDAR are leveraged to meet these significant design challenges, and constitute the contribution to the state of art.