Power systems and battery basics

Key takeaways

Battery life is often the single biggest constraint in cellular IoT design. A device that needs recharging or a battery swap every few months is expensive to maintain at scale, yet many deployments target five, seven, or ten or more years on a single battery. This page explains the power source options available, how to think about a power budget, and the strategies that make multi-year battery life possible.

What you'll learn

The three power source categories and their trade-offs

IoT devices draw power from one of three broad categories:

• Primary (non-rechargeable) batteries: The most common choice for low-power cellular IoT. Lithium thionyl chloride (Li-SOCl₂) is the gold standard for long-life deployments: high energy density, a wide operating range (−60°C to +85°C), very low self-discharge (under 1% per year), and a flat voltage curve that simplifies power management. AA alkaline and CR2032 coin cells suit lighter-duty uses. When depleted, you replace them or retire the device.
• Rechargeable batteries: Lithium-ion (Li-Ion), lithium-polymer (Li-Po), and lithium iron phosphate (LiFePO₄) are used when the device has external charging (solar, USB, vehicle power) or when power demands exceed a primary cell. They deliver higher peak current but have lower energy density per unit weight, degrade over charge cycles, and need more complex charging circuitry.
• Energy harvesting: Capture ambient energy from solar, vibration, thermal gradients, or RF. It rarely powers a cellular device alone (cellular transmissions need peak currents of 200+ mA that harvesting cannot supply), but works well as a supplement that trickle-charges a rechargeable battery or supercapacitor.

How to build a power budget

A power budget accounts for how much energy the device consumes in each operating state and how much the battery can supply. In principle the math is simple:

Battery life = Battery capacity (mAh) ÷ Average current draw (mA)

The catch is that IoT devices do not draw a constant current. They cycle between power states with very different consumption levels.

The four current-draw modes and why average current wins

A device using PSM/eDRX cycles through four distinct modes:

• PSM Sleep: The deepest sleep, drawing as little as 3 µA. The cellular modem is off and only the MCU's real-time clock runs.
• eDRX idle: A lighter sleep where the device periodically listens for paging messages, averaging around 15 µA.
• Active: The MCU is running, reading sensors and processing data, drawing about 2 mA.
• Transmit: The cellular modem is actively transmitting, drawing 220 mA or more in short bursts.

The key insight is that average current, not peak current, determines battery life. A device that spends 99.99% of its time in PSM sleep at 3 µA and only briefly transmits at 220 mA can average single-digit microamperes, enabling years of battery life.

The power-saving features built for cellular IoT: PSM, eDRX, and RAI

• Power Saving Mode (PSM) lets the device enter deep sleep for hours or days while staying registered with the network. It negotiates a TAU (Tracking Area Update) timer that defines how long it can sleep; during PSM it is unreachable but avoids re-registering on wake, saving significant energy. With daily reporting and aggressive PSM settings, a device can exceed 7 years on a standard Li-SOCl₂ D-cell.
• Extended Discontinuous Reception (eDRX) is a compromise between PSM and always-on: the device sleeps for configurable intervals (seconds to minutes) but periodically wakes to check for downlink messages, giving near-real-time reachability while still saving power.
• Release Assistance Indicator (RAI) is a network-level optimization that tells the network the device has finished transmitting. Setting an RAI flag after the final packet asks the network to release the radio connection immediately, bypassing the power-draining inactivity timer. This can cut modem energy per transmission cycle by up to 40%.

Lesson 3.3 Self-assessment

Frequently asked questions