Sub-MicroWatt Power Management Circuits and Systems for Self-Powered Internet-of-Things Applications

Emerging trillions of wireless sensor nodes for Internet-of-Things (IoT) applications, such as wearable healthcare, structure health monitoring, and smart home and cities, are dramatically improving our life quality and productivity. To truly enable the IoT era, those sensor nodes need to be fully autonomous and deployable, which requires them to have ultra-low power consumption to increase life time, self-powered and batteryless capability to avoid frequent and a large number of battery change, and highly efficient power delivery train to enable deployment under a variety of environmental conditions. To meet those requirements, the design of power management units (PMUs) including energy harvesting interface circuits and voltage regulators for self-powered system-on-chips (SoCs) is becoming critical and challenging. Especially, in recent years, with power consumption of different loading components gradually reducing from µW down to nW or even pW, and reduced energy from energy harvesters due to limited form factor and energy availability in the environment, PMUs need to be power efficient enough to deliver pW to nW output power from energy sources to loads, which requires them to have ultra-low quiescent power within sub-µW range and meanwhile maintain a high performance.

This dissertation aims to explore sub-µW, high-performance, and highly power-efficient architectures for power management circuits and systems, which includes two main categories and covers the entire power delivery train for self-powered IoT systems. The first category is from energy harvesting perspective looking into how to extract maximum energy from the environment. By extracting more energy, it opens up more applications where energy limits the deployment of IoT nodes. The second category explores how to design voltage regulators to efficiently power nW loads, which requires voltage regulators themselves to consume nW or sub-nW quiescent power. In this category, we provide a complete sub-nW power management solution including low-dropout regulators and a bandgap reference. To achieve the goals for these two categories, four research work has been conducted in this dissertation.

The first research explores a highly efficient piezoelectric energy harvesting system with maximum power-point tracking (MPPT). In this work, a high-performance parallel synchronized-switch harvesting-on-inductor (SSHI) rectifier with >400% figure-of-merit (FOM) has been implemented together with a highly efficient MPPT scheme with >95% tracking efficiency. The second work in the energy harvesting category is to design a multi-input single-inductor multi-output (MISIMO) energy harvesting and power management unit (EHPMU) with nW quiescent power. This MISIMO EHPMU can extract energy from thermal, solar and vibration energy simultaneously and provide four voltage outputs for loads, which greatly extends the energy extraction and power delivery capability. It also integrates a multi-modal cold start-up block and combines the energy harvesting interfaces and voltage regulators in one power stage to minimize the form factor and eliminate the cascaded power loss.

The next two research provides a sub-nW power management solution for nW IoT systems. The first work is to explore the design space of sub-nA low dropout regulators (LDOs), which includes two designs, a digital LDO (DLDO) and an analog LDO (ALDO). The DLDO uses a hybrid synchronous and asynchronous control scheme and keeps sub-nW quiescent power. As a comparison, a traditional fully integrated sub-nA ALDO using an analog feedback loop is also presented, which is well suited for powering analog and RF blocks in fully integrated nW IoT systems. The final work is to design a sub-nW bandgap reference (BGR) with a wide input voltage range. By directly biasing the bipolar junction transistors (BJTs) with pA current and using an input charge pump to increase input voltage range, the proposed BGR achieves sub-nW power and an input voltage range from 0.45V up to 3.3V in simulations. The designed sub-nW BGR can be used together with sub-nW voltage regulators to generate voltage supplies with a good stability against process, voltage and temperature (PVT) variation.