How to Choose a Battery for IoT Sensors

Choosing a battery for an IoT sensor looks simple until the sensor is deployed. A sensor that works on the bench for a week may fail in the field after a year, because the battery was sized for the bench, not for the deployment. This guide works through the selection steps in order, so the battery chosen is the battery the deployment actually needs.

Step 1: Voltage

Start with the voltage the sensor's electronics require. A sensor designed for a 3.6V rail points to LiSOCl2 cells. A sensor designed for a 3V rail points to LiMnO2 cells. A sensor designed for a 1.5V supply points to LiFeS2 cells in AA or AAA sizes. Voltage is a constraint set by the electronics, not a preference to trade against capacity or price. If the voltage is wrong, nothing else about the cell matters.

Step 2: Energy budget

Next, estimate the energy the sensor needs over its required service life. That means adding up the sleep current over time, the measurement current during each reading, and the transmission current during each report, multiplied by how often each happens. A sensor that measures once an hour and reports once a day needs far less energy than a sensor that measures every minute and reports every hour. The energy budget, divided by the cell voltage, gives the capacity the battery must deliver over its life — and that capacity, with margin, points to a cell size.

Capacity figures in catalogues are measured under specific test conditions: a stated discharge current, a cut-off voltage and a test temperature. A cell rated at a given capacity under light load will deliver less under heavy load. Read capacity figures with their conditions, and leave margin for temperature extremes and for the end of the cell's life.

Step 3: Pulse loads

Wireless transmission is the step most often underestimated. A sensor that sleeps at microamps may draw hundreds of milliamps, or more, during a transmission burst lasting milliseconds or seconds. A lithium thionyl chloride cell handles steady low current well but handles high pulses poorly, especially after long storage, when passivation raises its internal resistance. If the sensor transmits frequently or at high power, consider a hybrid pulse capacitor pack: the capacitor supplies the pulse, the cell supplies the long-term energy, and the cell's service life is protected.

For sensors with rare, low-power transmissions, a single ER cell is usually enough. The dividing line is the pulse current and frequency, and it should be established from the radio module's datasheet, not guessed.

Step 4: Temperature

Outdoor sensors face the full local climate. The ER series detail page states -55 to +85 degrees Celsius, the LiSOCl2 category states -60 to 85 degrees Celsius, and the LiMnO2 category states -40 to 85 degrees Celsius. A sensor in a cold climate needs a cell that works at the lowest expected temperature, not just at room temperature, because capacity and voltage both fall in the cold. A sensor in a hot enclosure needs a cell rated for the highest expected temperature, because heat accelerates self-discharge.

Step 5: Size, service life and ordering

Finally, check that the chosen cell fits the housing, including any holder, tabs or leads required for assembly. Then confirm the expected service life against the load profile, not against a category shelf-life figure alone. Shelf life of more than 10 years for LiSOCl2 is a storage figure; service life in a transmitting sensor will be shorter, and the gap depends on the transmission pattern.

Before ordering, confirm the exact model, the capacity variant, the pulse requirements, the temperature range and the final datasheet. Samples are available for qualification. Minimum order value is USD $500, according to the company FAQ. Lead time is confirmed per order. Certificates are available upon request.

Related pages: Batteries for IoT Devices, LiSOCl2 Batteries, Hybrid Pulse Capacitor Packs, ER14505.

Worked example: sizing a sensor battery

Consider a soil moisture sensor that wakes once an hour, measures for one second at 5 mA, and transmits once a day for two seconds at 100 mA, sleeping the rest of the time at 5 microamps. Over a day, sleep consumes about 0.12 mAh, measurement consumes about 0.033 mAh, and transmission consumes about 0.056 mAh, for a daily total near 0.21 mAh. Over five years, that is roughly 380 mAh, before margin. A compact ER14250 cell at 1,200 mAh would cover that budget several times over, even allowing for cold-temperature losses and end-of-life fade. The same sensor transmitting every hour instead of once a day would consume roughly 1.3 mAh per day from transmission alone, or about 2,400 mAh over five years — a budget that points to a larger cell such as ER14505 at 2,400 mAh at the edge, or ER26500 at 8,500 mAh with comfortable margin, or a hybrid pulse capacitor pack if the transmission pulses strain a bare cell. The arithmetic is approximate, and the datasheet for the exact cell and the radio module should replace it, but it shows why transmission frequency dominates the sizing decision for IoT sensors.

The same arithmetic explains why two sensors with identical hardware can need different batteries: change the reporting interval, and the energy budget changes by an order of magnitude. That is why the first question for any IoT battery selection is not which cell, but how often the device transmits, at what power, and for how long it must run. Answer those three questions, and the cell choice usually follows.

Qualification before bulk order

Before committing to a bulk order, qualify the chosen cell in the actual device, under conditions close to deployment. Test at the lowest expected temperature, not just at room temperature, because capacity and voltage both fall in the cold. Test the transmission pattern the device will actually use, including the worst case of repeated transmissions, because a cell that handles one report may sag on the tenth. Test after storage, not just on fresh cells, because passivation after months on a shelf changes the voltage behaviour under load. Samples are available for exactly this qualification, minimum order value is USD $500 according to the company FAQ, and lead time is confirmed per order. A week of qualification testing costs far less than a field recall.

For a sensor developer starting a new design, the practical next step is to prototype with samples: the cell size the energy budget points to, tested in the sensor itself under the real reporting pattern and at the lowest expected temperature. Samples are available, minimum order value is USD $500 according to the company FAQ, and lead time is confirmed per order. A prototype test answers the questions no spreadsheet can: the voltage during transmission, the capacity over the required service life, and the behaviour after storage. Design the battery compartment around the tested cell, not the other way around.