Electronics Guide

Matching Supply to Demand

The core difficulty is a mismatch of scale, not merely of timing. An indoor photovoltaic cell of a few square centimeters under office lighting yields on the order of microwatts, while a single Bluetooth Low Energy advertisement draws several milliamperes for a few milliseconds. Peak load current can exceed average source current by three orders of magnitude or more, and no practical harvester can supply that peak directly. Storage closes the gap by accumulating energy slowly and releasing it quickly.

Design therefore begins with an energy budget rather than a power budget. The designer estimates the energy harvested per day under worst-case conditions, sums the energy consumed per operating cycle—sensing, computation, radio transmission, and sleep leakage—and derives the achievable duty cycle from the ratio. A node that harvests 50 millijoules per day and spends 5 millijoules per measurement-and-report cycle can afford roughly ten cycles per day, minus conversion and leakage losses. The storage element must then hold enough energy to cover the longest expected drought and to supply the largest single burst without excessive voltage sag.

Two further quantities shape the choice. The first is source impedance: harvesters behave as current-limited or high-impedance sources, so connecting them directly to a discharged cell collapses the source voltage and stalls the transfer. The second is the system's cold-start condition, in which storage is empty, the microcontroller cannot run, and the power path must bring itself up from nothing using the harvester alone.

Storage Technology Selection

Choosing a storage technology means balancing energy density, power density, cycle life, self-discharge, temperature range, and cost. Lithium-ion cells based on cobalt or nickel oxides dominate where energy density matters, typically delivering 150 to 250 watt-hours per kilogram but lasting only several hundred to roughly 1,500 full cycles. Lithium iron phosphate trades energy density for endurance, offering roughly 100 to 140 watt-hours per kilogram with cycle lives in the thousands. Lithium titanate goes further still, sustaining well over 10,000 cycles at 50 to 80 watt-hours per kilogram, which suits deployments where a decade of daily cycling matters more than mass.

Temperature limits often decide the outcome for outdoor nodes. Lithium-ion cells must not be charged below approximately 0 degrees Celsius, because metallic lithium plates onto the anode instead of intercalating, permanently reducing capacity and creating a safety hazard. A harvesting charger deployed in a cold climate therefore needs a temperature sensor and the logic to suspend charging, which means the node may collect no energy at all during the coldest part of the year unless the design accounts for it.

Supercapacitors invert most of these trade-offs. Electrical double-layer capacitors store far less energy, commonly 5 to 10 watt-hours per kilogram, yet sustain 500,000 to more than a million cycles, accept power on the order of 1,000 to 10,000 watts per kilogram, and operate at temperatures well below freezing without the charging restrictions of a lithium cell. Their weakness is self-discharge: leakage currents in the microampere range drain a charged cell over days to weeks, whereas a lithium-ion cell typically loses only a few percent of its charge per month. A supercapacitor is an excellent buffer for hours and a poor reservoir for seasons.

A capacitor also stores energy as one-half times capacitance times the square of voltage, so its terminal voltage falls continuously as it discharges rather than holding a plateau. Extracting a useful fraction of the stored energy demands a converter with a wide input range: discharging from the rated voltage down to half of it recovers three-quarters of the stored energy, while a converter that gives up at 80 percent of rated voltage recovers barely a third. The compensation is that voltage maps directly to state of charge, which makes fuel gauging almost trivial.

Several intermediate and specialized options fill the remaining gaps. Lithium-ion capacitors combine a capacitive electrode with a lithium-intercalating one, landing between the two families in both energy and cycle life. Thin-film and micro-batteries offer small capacities in solderable, reflow-tolerant packages with long cycle lives and low leakage, which suits chip-scale and implantable nodes. Solid-state cells promise wider temperature tolerance and reduced flammability. Where the duty cycle is extremely low, a primary lithium cell sometimes serves as a non-rechargeable backstop that the harvester merely spares, extending service life without the complexity of a charge controller.

Power Management Architectures

Power management circuits regulate energy flow from harvester to storage and from storage to load. Core functions include maximum power point tracking to extract optimal power from a variable source, charge regulation to protect the storage device, voltage conversion to match load rails, and power-path control that decides whether the load draws from the source, from storage, or from both.

Because harvested power is small, converter quiescent current and cold-start behavior are decisive. A regulator that consumes tens of microamperes idle would waste more than an indoor photovoltaic cell collects. Dedicated harvesting power-management integrated circuits are built around this constraint: the Texas Instruments BQ25504, for example, draws roughly 330 nanoamperes of operating current, cold-starts from an input of about 330 millivolts, and continues to harvest down to 80 millivolts once running. For sources that never reach even that level, such as small thermoelectric generators across a few degrees of temperature difference, the Analog Devices LTC3108 uses an external step-up transformer in a resonant oscillator to operate from inputs as low as 20 millivolts.

Tracking strategies scale with the power available. Full perturb-and-observe or incremental-conductance algorithms cost computation and sampling energy that a microwatt-class node cannot spare, so harvesting ICs commonly use fractional open-circuit voltage tracking instead: the controller periodically disconnects the source, samples its open-circuit voltage, and regulates the input at a fixed fraction of that value, typically near 80 percent for photovoltaic cells and 50 percent for thermoelectric generators, where source impedance matching governs. The approximation gives up a few percent of the theoretical maximum in exchange for a control loop that costs nanowatts.

Converter topologies range from simple low-dropout regulators, which are acceptable only when input and output voltages are close, through switched-capacitor charge pumps that need no inductor and suit fully integrated designs, to inductive boost, buck, and buck-boost stages that handle wide and shifting input ranges. Many harvesting front ends are multi-stage: a low-voltage boost charges an intermediate storage node, and a second, higher-efficiency converter serves the load rail from that node so each stage operates near its own efficiency peak.

Charge Management and Protection

Charge-management circuits, implemented as integrated chargers or discrete designs, control how energy enters the storage device. They provide linear or switching charge paths, chemistry-appropriate profiles such as constant-current followed by constant-voltage for lithium-ion, temperature-qualified charging windows, and trickle and fast-charge modes, together with protection against overcharge, over-discharge, and overcurrent.

Harvesting imposes an additional requirement that conventional chargers do not meet. A mains-powered charger demands a fixed charging current from an effectively infinite source; a harvesting charger must instead draw only what the source can give at that instant, backing off as light dims or a thermal gradient collapses, and resuming without supervision when energy returns. Chargers designed for this role regulate the input side to the harvester's maximum power point and let the resulting charge current float.

Protection matters more than usual because these systems are unattended for years. Over-discharge lockout is critical: taking a lithium-ion cell below the manufacturer's cutoff, commonly 2.5 to 3.0 volts, risks dissolution of the copper current collector and permanent capacity loss, so the protection circuit must disconnect the load and then hold that state at negligible current until the harvester recovers. Supercapacitor stacks need cell balancing, since series cells with unequal leakage drift apart in voltage and the weakest cell is driven past its rating. Passive resistive balancing is simple but adds continuous leakage, which may be unacceptable in a microwatt budget; active balancing costs complexity but preserves the energy budget.

State Estimation and Monitoring

Accurate knowledge of storage state underpins reliable operation. State-of-charge estimation reports the remaining usable energy, while state-of-health tracking follows capacity fade and impedance growth as the cell ages. Practical estimators combine coulomb counting with open-circuit-voltage correlation and, in more capable systems, impedance measurement or model-based filtering such as an extended Kalman filter, which remains accurate across temperature and aging.

Each method has a characteristic failure mode. Coulomb counting integrates current and therefore accumulates offset error without bound, requiring periodic recalibration at a known reference point. Open-circuit voltage is accurate but requires the cell to rest, and lithium iron phosphate has such a flat discharge curve that voltage barely distinguishes 30 percent from 70 percent state of charge. Capacitive storage escapes both problems, since terminal voltage gives state of charge directly. In every case the measurement circuit must itself be frugal, because a fuel gauge that draws more current than the sleeping load defeats its own purpose.

In harvesting systems this data feeds energy-aware scheduling directly. The controller uses reserve level, recent harvest rate, and predicted availability to decide whether to take a measurement now, defer it, or drop to a minimal survival mode, so state estimation becomes an input to system behavior rather than a diagnostic readout.

Hybrid Storage Architectures

Combining technologies lets a system approach the energy density of a battery and the power density of a capacitor at the same time. The simplest arrangement connects a supercapacitor in parallel with the cell, letting the capacitor's low equivalent series resistance absorb transmit-current transients that would otherwise sag the battery terminal voltage and trip a brownout detector. This passive approach costs almost nothing but fixes the two elements at a common voltage.

Converter-mediated architectures separate the voltages and give the controller a power-splitting decision: steady, low-current draw comes from the battery, while transient and high-current demands come from the capacitor, which the battery or harvester then refills slowly. Reducing peak current through the cell lowers internal heating and slows capacity fade, and the arrangement is especially valuable in cold weather, when battery internal resistance rises sharply and the capacitor carries the burst the cell no longer can. A common variant reverses the roles at the input, charging a supercapacitor first so the node can start working immediately, then diverting surplus energy to the battery for long-term reserve.

The cost is real. Every additional conversion stage subtracts efficiency, and every additional component adds quiescent current, volume, and failure modes. Hybrid storage earns its place when peak-to-average current ratios are high, when temperature extremes are wide, or when battery cycle life is the limiting factor in service life. When the load is nearly constant and the environment is mild, a single well-chosen element is the better engineering answer.

Energy-Aware System Design

Storage integration succeeds only when the load adapts to what the source provides. Duty cycling is the primary lever: a node that sleeps at hundreds of nanoamperes and wakes briefly to work can average microwatts while performing milliwatt-class tasks. Dynamic voltage and frequency scaling extends the same idea into the active period, trading execution speed for energy per operation. Predictive management uses the harvest history—diurnal solar cycles, machine duty schedules—to anticipate availability and schedule expensive operations when energy will be plentiful.

Energy-neutral operation is the governing principle: over a chosen horizon, average consumption must stay below average harvest, with storage absorbing the difference within the horizon. Once that condition holds, service life is limited by component aging rather than by stored energy, and the node can run indefinitely. Where storage is minimal or absent, intermittent computing techniques let execution stop and resume across power failures by checkpointing state to nonvolatile memory, so a task completes across many brief power-on intervals rather than requiring one continuous supply.

Sizing and Lifetime Trade-offs

Sizing the storage element is a worst-case exercise. The capacity must cover the longest expected period without harvest—overnight for a solar node, a plant shutdown for a vibration-powered sensor, the darkest weeks of winter for an outdoor deployment at high latitude—plus a margin for the capacity the cell will have lost by end of life. Undersizing produces a node that dies in its first bad week; oversizing wastes cost and volume, and, in the case of a battery, adds self-discharge and calendar aging that the node must keep paying for.

Aging follows two independent clocks. Cycle aging depends on how much charge passes through the cell and how deeply it is cycled; calendar aging proceeds regardless of use and accelerates at high temperature and high state of charge. A harvesting node that keeps a lithium-ion cell at full charge and elevated temperature can lose more capacity to calendar aging than to its own gentle cycling, which argues for holding a lower float voltage and accepting shallower depth of discharge. Because most such deployments are meant to outlive any practical maintenance visit, storage frequently determines the system's service life, and the honest measure of a design is the energy it can still deliver after ten years rather than the energy it delivers when new.

Applications

Energy storage integration enables practical deployment of harvesting-powered systems across many domains. Wireless sensor networks run for years without maintenance on small cells or supercapacitors charged by light or vibration, reporting infrequently enough to stay within a microwatt budget. Industrial condition-monitoring nodes capture machinery vibration to power their own accelerometers and radios, which removes the wiring and battery-replacement costs that otherwise make dense monitoring uneconomical.

Wearable and implantable devices draw on body heat and motion to supplement or replace battery charging, where thin-film and flexible storage matters as much as the harvester itself. Building systems buffer photovoltaic and thermal energy for wireless switches, occupancy sensors, and environmental controls that install without wiring. Electronic shelf labels, asset tags, and structural monitors in bridges and pipelines rely on the same pattern: a modest harvester, a carefully sized buffer, and a load disciplined enough to live within the budget. In each case the storage and management layer is what turns an intermittent source into a dependable supply.

Articles in This Category

About This Category

This category covers the technologies and techniques that connect harvesting sources to electronic loads through appropriate storage and power management. From storage device characteristics to charge control and energy-aware system design, these topics provide the knowledge needed to build reliable energy-autonomous systems—devices that operate indefinitely from ambient environmental energy.

The material sits between the source-side and system-side topics treated elsewhere in this guide. The converters and front ends that feed the storage element are developed in Circuit Design and Power Management, and in particular in Energy Buffer and Storage Interface. The behavior of the sources themselves appears under Solar Energy Systems, Thermal Energy Harvesting, and Mechanical Energy Harvesting, while combining several of them is the subject of Hybrid Energy Harvesting. Methods for sizing and validating a complete node are covered in Design and Optimization, deployed examples in Applications and Systems, and the hazards specific to stored energy in Safety and Reliability.