Ultra-low-power IoT design is shaped by a practical challenge: sensors often need to operate where convenient access to electrical power is unavailable. For industrial automation engineers and system integrators, this matters when considering distributed monitoring points, embedded sensing and other applications where devices must fit close to the physical condition being measured. The design challenge is not simply to reduce the power of a processor. It is to balance sensing, timing, communications, energy supply, device size and service life as parts of one system.
IoT devices differ from conventional computers because they interact directly with the physical environment. Sensors provide inputs, while wired or wireless links convey information to other computing resources. This permits devices to be placed in locations and applications that are poorly suited to systems built around large human-facing interfaces. But the breadth of possible uses means there is no single set of hardware requirements that fits every sensor node.
Why IoT requirements vary so widely
A useful way to understand an IoT application is to consider four properties: sensing modality, form factor, energy source and required lifetime, and connectivity. Each can take many forms, and the combination determines the system’s design priorities.
- Sensing modality: A device may measure pressure, chemical properties, strain or temperature. The measurement schedule may be continuous, periodic, triggered by a change in the measured value, or initiated by an external event.
- Form factor: Devices range from millimeter-scale sensors to larger units used for infrastructure monitoring. Size influences what components and energy sources can be accommodated.
- Energy and lifetime: A node may rely on a battery, harvested energy or a combination. Required operating life affects how much average power the system can consume.
- Connectivity: Communication needs include how often a device sends information, over what distance, and with what data rate or data size.
These factors lead to very different operating profiles. A sensor that takes an occasional temperature reading and communicates infrequently has different needs from a device that monitors a signal continuously and transmits when an event occurs. Infrastructure monitoring with a long intended lifetime also places distinct demands on the available energy budget.
Ultra-low-power IoT design starts with the power budget
When a device is untethered, its energy supply limits how much power it can use over time. Smaller batteries and longer operating lifetimes impose tighter average-power constraints. At millimeter scale, the available budget can reach the nanowatt range, while larger systems may support higher average consumption, depending on their battery and lifetime requirements.
Average power is not determined by a single component. It depends on how the device operates. In a low-duty-cycle system, a sensor may spend much of its time inactive, making standby consumption a significant concern. In a system with higher data-rate or bandwidth needs, the radio can become a dominant part of the power budget. Designers therefore need to assess the complete use pattern, including sensing intervals, processing, sleep periods and communication.
This system-level view helps identify which design improvements are relevant. Reducing active power may have limited effect if standby consumption dominates. Conversely, optimizing an always-on timer or radio can matter when it is active often or contributes substantially to total energy use.
Circuit techniques for reducing consumption
Several circuit innovations address functions that IoT devices may need to perform with very little energy. One example is a low-power voltage reference based on the difference between the threshold voltages of high- and low-threshold devices. By sizing the devices to offset temperature dependence, the design provides a stable reference while drawing current in the tens of picoamps. A low-power reference can be important when a system needs a continuously available voltage point without a large standby cost.
Timers are another important consideration because they may need to remain active while much of the device is asleep. A reported crystal oscillator design consumes 5.5 nanowatts. It reduces power by lowering the oscillation amplitude and activating the driver only briefly when the oscillator input reaches its minimum or maximum values. The approach illustrates how both signal amplitude and circuit activity can influence the energy required for an always-on function.
Energy harvesting also creates a conversion challenge. Input power can vary, and small devices may benefit from having conversion circuitry integrated on-chip. A reported voltage doubler combines an oscillator with its switched-capacitor structure, avoiding a separate clock generator and level shifters. The design operates across a broad current range and can start from a low input voltage. These examples show why the energy source, conversion circuit and load should be considered together rather than as independent blocks.
Modular systems and inter-die communication
The diversity of IoT applications makes it impractical to design a unique chip for every use. A modular approach instead combines compatible dies—individual semiconductor chips—according to the needs of a particular device. This can allow system builders to select relevant functions while reusing components across different configurations. The challenge is to connect those components without imposing excessive power, wiring or size costs.
Conventional bus choices can involve trade-offs. SPI uses a relatively large number of wires, while I2C uses pull-up resistors that can contribute to power consumption. The MBus protocol described in the source material was designed for inter-die communication using two unidirectional wires in a daisy-chain topology. It reduces active power by eliminating the need for a clock generator on each slave layer and uses power gating to reduce sleep power.
Using this approach, three millimeter-scale sensor systems were assembled from subsets of an eight-die set. The available dies included sensing, processing and memory, power management, energy harvesting, and near- and far-field radio functions. Stacking selected dies and connecting them through MBus demonstrates how modular components can be combined into different sensor configurations.
Practical implications for automation teams
For teams evaluating distributed sensing, the central lesson is to define the operating requirements before selecting components. Specify what is measured, how often measurement occurs, what events trigger activity, how much information must be sent and how long the device must operate. Then use those requirements to identify whether sensing, standby, timing, conversion or communication is likely to dominate power consumption.
Modularity can support reuse across varied applications, but only when the interconnect and component choices suit the system’s power and size constraints. Ultra-low-power IoT design is therefore a coordinated systems problem: energy-aware circuits, a suitable energy source, deliberate duty cycling and efficient communication must work together. A clear view of these trade-offs gives engineers and technical decision-makers a stronger basis for designing sensor nodes that fit their intended physical and operational environment.