Switched-inductor power supplies are valued for their high efficiency despite the bulkiness of off-chip inductors. When designing compact systems like portable consumer electronics and wireless microsensors, single-inductor topologies are therefore preferred. Specifically, single-inductor multi-input and multi-output (SL–MI/O) power supply designs pose unique challenges that have yet to be fully addressed. This paper aims to provide design guidelines for maximizing efficiency in the design of SL–MI/O systems, especially in the sub-5W domain. To simplify the choice between NFETs and PFETs for the multitude of power switches in SL–MI/Os, which is not straightforward, an intuitive metric called the Favorability Index (FNP) is proposed. A new, optimal supply voltage theory is also presented, suggesting that the most efficient voltage to supply power switches’ gates is around twice the threshold voltage (vT). The paper also proposes using dynamic selectors in gate drivers. This allows for blocking cross conduction without increasing vSUPdrastically, ensuring efficiency. A two-transistor selector is recommended as a simple implementation, and the tradeoffs are discussed. An example topology is designed using guidelines proposed by the paper to demonstrate the design flow and efficiency improvements.

Integrating and conforming emerging wireless IoT microsensors into tiny form factors is challenging in many ways. To conserve energy, for example, the system should idle whenever possible, activating functional blocks on demand only. Internal power circuits must therefore supply and cut-off power quickly. In the interim, as these react, capacitors supply and sink the mismatch in power. The low capacitance that small capacitors afford, however, cannot supply or sink much power for long. Hysteretic power supplies are appealing in this respect because they respond quickly. But how fast and reliably they respond depends on design, which hinges on understanding. This paper uses and develops insight to explain and analyze the feedback dynamics and stability requirements of hysteretic current-mode dc-dc switched-inductor converters, which are largely abstract and algebraic in literature today. Moreover, this paper also outlines and analyzes possible practical design issues related to IC implementations. To this end, the paper derives accurate and insightful expressions, uses and applies them to a design, and validates them with SPICE simulations.

Light Emitting Diodes (LEDs) have become pervasive in modern lighting and automotive applications. LED drivers regulate LED current which sets their luminous output, where dimming is an important attribute. Dimming techniques fall in one of two categories: "analog" or "duty-cycled" (pulse-width-modulated), and duty-cycled (PWM) dimming decomposes into three further classes: shutdown, shunt- and series-switched. However, a comprehensive analysis of dimming techniques, corresponding power losses, and their dimming capabilities is lacking in the literature. This paper explains and quantifies those in the context of a switched inductor (SL) DC–DC converter. Presented analysis incorporates SL conversion efficiency and models luminous flux, dimming range, and luminous efficiency. This paper reveals and verifies that analog dimming is up to 57% more efficient with the widest dimming range.

Power-conversion efficiency is critical in power supplies. Switched inductors are popular in this space because they can deliver a large fraction of the power they draw. This fraction hinges on the power that switches, diodes, resistances, and capacitances need to conduct and transfer power to the output. So, understanding how these loss mechanisms set and dictate efficiency across power levels is important, especially when designing and targeting particular load levels. This article details how these losses scale, when they dominate, and how and when they balance. Gate drive and controller losses are the ones that become a smaller fraction of input power as output power increases, leading to the increase of the power efficiency at the low-end of discontinuous conduction scale, while only gate charge loss plays this role at the low-end of the continuous conduction scale. Ohmic loss is the one that reduces power efficiency at the high-end of discontinuous and continuous conduction scale as ohmic loss becomes a larger fraction of input power as output power increases. Power efficiency peaks in continuous conduction when ohmic loss and gate charge losses balance. Overlap and dead time losses, although still important, do not shape the power efficiency in continuous conduction mode. In discontinuous conduction mode, all losses play a role and efficiency peaks when they all trickily balance. With this insight, predicting and controlling when efficiency rises, peaks, and falls across loads are possible. The fractional loss analysis and the design insight that make this possible are new contributions to the state of the art.

Inductively coupled power receivers for embedded microsensors are often tiny and distant from their transmitting sources. To sustain microsensors as long as possible, the power receiver should draw power whenever possible and output the highest power possible. Of reported state-of-the-art technologies, switched resonant half-bridges require fewer components, are less breakdown-limited, and output as much or more power than the others. This paper derives and shows with measurements the highest possible maximum power point (MPP) for switched resonant half-bridges. The theory predicts the optimal time, duration, and frequency of the energy transfers that charge the battery. Measurements of a 0.18-μm CMOS power receiver demonstrate that the receiver outputs more than 98.7% of the actual MPP at the predicted settings when the coupling factor between the transmitting and the receiving coils is 0.15%-1.14%.

Emerging applications demand compact, battery-powered, and highly functional microsystems that require on-chip integration, low average power, and high peak-to-average power ratios. Switched inductors are popular power supplies because they are power efficient. Switching a power inductor, however, generates a nonlinear current ripple that is often difficult to tolerate and manage. The problem is more severe with on-chip nH inductors and pF capacitors. This paper explores how switching power supplies can manage and reduce this ripple. Although multiphase and filter suppressors help, analog cancellation can be 38× to 77× more effective, but also less power efficient and 23× to 55× more sensitive to mismatch from drift.

Wireless microsensors can sense and share data that can save lives, energy, and money. Recharging or replacing thousands of tiny, easily exhaustible batteries, however, is too costly. Fortunately, photovoltaic (PV) cells can generate 100x more power from sunlight than other transducers can from motion, heat, or radiation. But since PV cells cannot supply the milliwatts that microsystems can at times require, this paper shows how to design battery-assisted PV-sourced CMOS charger-supplies that supply PV power to the system, excess PV power to the battery, and battery power to the system when PV power is insufficient. The design process proposed accounts for power losses and silicon area. This way, simulations show that 10% of losses are from switches when inductor resistance Resr is 2.2 Ω and silicon area can be 80% smaller when Resr is 5.5 Ω.

Summary form only given. A fundamental challenge wireless microsystems face is size, and in consequence, lifetime because tiny batteries exhaust quickly. Although small fuel cells and atomic sources store more energy than lithium-ion batteries and super capacitors, they source less power, so they cannot power as many functions. Small batteries and capacitors, however, cannot sustain life for long. Thankfully, the environment holds vast amounts of energy. And of typical sources like light, motion, temperature, and radiation, sunlight produces the highest power density, but only when available. Combining photovoltaic (PV) cells with tiny batteries or capacitors can therefore be more compact, reliable, and longer lasting than any one of these technologies alone. Managing a hybrid system of this sort to supply a milliwatt application, however, requires an intelligent, low-loss charger-supply system. This talk surveys and describes how smart PV-sourced microsystems can draw power from tiny PV cells and supplementary power from small batteries to supply a load and replenish the battery with excess PV power. To that end, the material reviews and discusses miniaturized PV cells, power-efficient charger-supply circuits, and reliable feedback controllers. The presentation ends with measurement results from a prototyped example.

Because wireless microsystems can only incorporate tiny batteries, they typically exhaust stored on-board energy quickly. Fortunately, harvesting ambient energy is a viable means of extending their operational lifetimes, except starting and re-starting miniaturized microwatt harvesters from nocharge conditions is difficult. The challenge is drawing usable energy from millivolt signals under micro-scale constraints. This paper proposes a nonlinear on-chip starter that borrows the harvester's steady-state inductor to start the system from nocharge conditions. Simulations show that the starter draws power from 250-500 mV to charge 100 pF to 3 V in 48 μs. The 100-pF temporary supply then powers the harvester's 1-V, 4-μA controller to charge 100 nF by 100 mV in 65-μs cycles until the 100-nF battery charges enough to supply the system.

A challenge wireless microsensors and other microsystems face is short lifetime, because tiny batteries store little energy. Fortunately, the environment holds vast amounts of energy, and of available sources, like light, motion, temperature, and radiation, solar light produces the highest power density. Still, micro-scale photovoltaic (PV) cells harness a diminutive fraction of light and artificial lighting avails a small percentage of what solar light can, which means the PV cell needs assistance from a battery. Mixing PV and battery power to supply a microwatt system, however, requires a smart and low-loss circuit. For that, the battery-assisted pulse-width-modulated (PWM) buck-boost single-inductor 0.18-μm CMOS harvester-supply proposed and simulated here uses up to 100 μW from a PV cell to supply up to 1 mA and regulate 1 V within 25 mV at 10–80 kHz and with 77%–89% efficiency.

Battery-powered electronics rely on integration and power efficiency for size and operational life. Switched-inductor converters play a critical role in this because most portable systems depend on dc-dc converters to supply power efficiently. Understanding the collective impact of shrinking dimensions on total power losses in a switching converter is therefore important when selecting a process technology for the power-supply chip, because the optimal choice results in longer battery life. This paper analyzes and validates the effects of finer CMOS technologies (which feature shorter minimum channel lengths LMIN, higher oxide capacitance, and lower breakdown voltages) on the efficiency performance of switched-inductor dc-dc converters in continuous- and discontinuous-conduction modes (CCM and DCM). Simulation results show that conduction and gate-drive losses in switches rise with LMIN1.5and bias and bandwidth-critical quiescent losses with LMINand LMIN3, respectively. In other words, because parasitic components and gate-drive voltages rise with LMIN, efficiency drops with coarser technologies: E.g., the efficiencies of optimized 0.18-, 0.35-, and 0.5-μm buck converters peaked at 93%, 89%, and 79%.

Because small batteries store little energy, micro-scale systems often trade functionality or lifetime, or both, for integration. Harnessing ambient energy can abate the sacrifice, but only to the extent transducer and circuit efficiencies allow. Optimally adjusting the electrical damping force in the transducer is therefore as important as lowering power losses in the circuit. In kinetic electrostatic harvesters, raising the voltage across the moving parallel plates increases this force, which is what the energy-harvesting 0.35-μm CMOS charger proposed achieves with a 10-nF capacitor CCLAMP. The system presented harnesses fifteen times (15-×) more energy at 16 V (with 15 nJ/Cycle) than at 4 V (with 1 nJ/Cycle) from 50 -- 250-pF, 60-Hz variations to generate (after discounting loses in the system) a net gain of 8.8 nJ/Cycle at 16 V.

Miniaturized systems like wireless microsensors suffer from short operational lifetimes because they lack space to store the energy that wireless transmission, signal conditioning, and monitoring require to operate across time. Harvesting ambient energy circumvents this limitation because the environment is a virtually boundless reservoir of energy. Of available sources, solar light produces the highest power density, and although artificial lighting is not as rich, thermal and magnetic sources produce even lower power densities and mechanical and chemical transducers are difficult to integrate. The problem is microscale photovoltaic (PV) cells only produce 1 and 100 µW/mm2for artificial and solar lighting, so the act of conditioning and transferring power can dissipate most, if not all, of the power available. The focus of this paper is to introduce and discuss the design challenges associated with harvesting circuits when harnessing, conditioning, and transferring power from tiny PV cells that only generate 1 – 100 µW.

The potential application space for miniaturized systems like wireless microsensors is expansive, from reconnaissance mission work and remote sensors to biomedical implants and disposable consumer products. Conforming to microscale dimensions, however, constrains energy and power to such an extent that sustaining critical power-hungry functions like wireless communication is next to impossible. Harvesting ambient energy offers an appealing alternative, except the act of transferring energy requires power that could easily exceed what the transducer generates in the first place. This paper presents how to design low-power switched-inductor converters capable of producing net energy gains when supplied from low-power piezoelectric and electrostatic kinetic-harvesting sources.

Although the benefits of incorporating noninvasive intelligence (e.g. wireless micro-sensors) to state-of-the-art and difficult-to-replace technologies are undeniable, micro-scale integration constrains energy and power to the point lifetime and functionality fall below practical expectations, forcing technologists to seek energy and power from the surrounding environment. To this end, a piezoelectric energy harvester circuit is proposed. The 2 mum CMOS design circumvents the need for (and losses and low-voltage restrictions associated with) a rectifier by extracting and transferring energy directly from the piezoelectric transducer to the battery via a switched inductor. Simulation results show that the proposed system can harvest 45 nJ and 10 nJ per period at 71% and 69% efficiency from 3 V and 1.5 V peak piezoelectric voltages, respectively.

Wireless micro-sensors and similar technologies must derive their energy from micro-scale sources (e.g., thin-film Li Ions, etc.) to function in volume-constrained environments like the human body. Unfortunately, confining the source to small spaces limits the total energy available to such an extent that operational life is often impractically short. Ambient energy offers an alternate and virtually boundless source, except small volumes restrain harvesting power. Voltage-constrained electrostatic CMOS harvesters, for example, draw energy from the work done against the mechanical plates of a MEMS variable capacitor at relatively slow rates, producing low output power. This paper discusses how much energy is available in such a system before and after harvesting and offers energy-conservation schemes for increasing its net energy gain (i.e., power output) during all operational phases.

Portable electronic devices not only require switching DC-DC converters to be compact and integrated but also compliant to wide off-chip LC filter variations, which are subject to manufacturing tolerances, temporal and thermal parameter drifts, and more often than not, application-driven constraints. While optimal LC compliance has been demonstrated in ΣΔ buck converters, little has been done in boosting applications. This paper presents an asynchronous ΣΔ boost converter and describes how LC variations affect stability, steady-state error, and switching frequency, and how a frequency-dependent gain mitigates these effects. Simulations show the circuit is stable for 1-30μH inductances and 15-350μF output capacitances, its steady-state error is less than 1%, and its switching frequency varies 15% less (over load and line variations) than in conventional ΣΔ converters.

A hybrid energy source has become a necessary solution for micro-scale applications, like wireless sensors, because of small form factor and extended lifetime requirements. Conventional power mixer circuits use two or more inductors, which become prohibitively area intensive in micro-scale applications where area for discrete inductors is scarce and on-chip inductors are poor. Although single inductor multiple input or multiple output converters appear in recent literature, they cannot be applied to the hybrid energy source directly. A single inductor, multiple input, multiple output (SIMIMO) power mixer-charger-supply system is therefore proposed for a hybrid fuel cell-lithium ion source. The SIMIMO system adopts a novel nested hysteretic mode dual-loop control architecture, regulating both the fuel cell current and the output voltage within predetermined hysteres is windows. A SIMIMO with 67% average efficiency was designed and simulated, regulating the fuel cell current to 10mA and output voltage to 1.8V within ±20mV.

Power management is an essential component of any electrical system, and nowadays a limiting factor in the miniaturization of portable electronic devices. Not only are the battery and power components difficult to integrate but their performance requirements in mobile environments are more stringent. And although point-of-load (PoL) regulation techniques and monolithic controllers are industry standards today, more integration is indispensable. To address these issues, system-in-package (SiP) self-renewable energy source and storage devices are proposed alongside an array of circuit techniques designed to circumvent the shortcomings of such a miniaturized environment, like smart load-sharing schemes, customizable and self-adaptive PoL regulators, active inductor and capacitor multipliers, and robust self-calibrating and self-stabilizing dc-dc converters. On their own, each seeks to push the limits of integration while maintaining and many times improving performance. As a whole, they promise the birth of a new generation of ICs.

While the large passive elements of power converters are in the way of converging walls of shrinking cell phones and cameras, the new capabilities these devices flaunt are creating additional burdens and making it difficult to meet specifications without even bigger elements. Active circuits that enhance the effects of passive elements will allow power converters to handle larger loads and get smaller at the same time. This paper presents a predictive inductor multiplier circuit that amplifies the effective inductance in a Buck converter. The output ripple of the simulated converter is so small that the converter appears to have an inductance thirty-eight times the value actually used. Compensating for small inductors introduces new power losses, but it is discovered that linear regulators and faster switching converters can be even less efficient