A radio and a coin cell — the simple part of the system, until the first cold morning.
A BLE asset tag is the simplest thing in the whole system: a radio chip and a coin cell in a sealed shell. We’ve said before that when telemetry goes missing the gateway is almost always the culprit, not the tag. That’s true. But the tag has one failure mode that’s entirely its own, and it’s the one customers ask about most: how long does the battery last?
The honest answer is “it depends,” and the things it depends on are not the things the datasheet emphasizes.
The datasheet number is a lab number
“Up to 5 years” is a real measurement. It’s taken at a fixed advertising interval, at room temperature, with a fresh cell, broadcasting into open air. Every one of those conditions is generous, and your machine yard violates all four.
Four things set actual battery life. Three of them are tuning decisions; the fourth is physics you don’t get to argue with.
1. Advertising interval — the biggest lever
A beacon doesn’t transmit continuously. It wakes up, broadcasts, and sleeps. How often it does that — the advertising interval — is the single largest factor in battery life, and it’s a direct trade against how quickly the gateway finds the asset.
Roughly: a tag advertising every 100 ms is found almost instantly and might last a year. The same tag at one second is found within seconds and lasts several years. At ten seconds it sips power and lasts most of a decade, but now an asset that’s only in range for a moment — a machine passing through a gate — can be missed entirely.
There is no correct interval on the datasheet. There’s a correct interval for your deployment, and it depends on how fast assets move through coverage and how soon you need to know they arrived. Shipping every customer the factory default is how you get a tag that either dies early or misses detections.
2. Transmit power
Higher TX power buys range on a yard full of steel, which absorbs and reflects 2.4 GHz badly. It also costs battery on every single broadcast. Cranking power to punch a signal off a machine that’s acting like a giant ground plane is sometimes the right call — but it’s a withdrawal from the same account as the advertising interval, and the two compound.
3. Temperature — the physics you don’t get to argue with
Coin cells hate the cold. At -20 °C the lithium chemistry sags: internal resistance climbs and the effective capacity you can actually pull drops well below the rated milliamp-hours. Heat is the other tail — sustained high temperatures raise self-discharge, so the cell quietly drains even when the tag is barely transmitting.
Off-highway equipment lives in both extremes, often in the same week. A tag rated for “5 years” at 20 °C on a bench is a different tag on an excavator that spends January at -25 °C and August baking on black steel in the sun. The duty cycle that the cab is built to survive is the same one quietly eating your coin cells.
4. The cell itself
Not all coin cells are the same battery in a different label:
- CR2032 — the default. Cheap, ~225 mAh, fine for indoor and mild outdoor.
- CR2477 — physically bigger, ~1,000 mAh. The obvious upgrade when the duty cycle is harsh and there’s room in the enclosure.
- Lithium thionyl chloride (e.g. ER14250) — wider temperature range and far lower self-discharge. The right choice for genuinely brutal cold or long-life sealed tags, at a higher cost.
Picking the cell for the brochure conditions instead of the worst month is one of the quiet ways a deployment ages badly.
Why tags die suddenly, not gradually
Here’s the part that surprises people. A coin cell has meaningful internal resistance, and that resistance rises as the cell ages and as it gets cold. Every time the radio transmits, it pulls a brief current spike, and that spike causes a momentary voltage droop across the internal resistance. Cold cell, plus aged cell, plus transmit pulse, and the supply voltage dips below the chip’s brownout threshold — the tag resets or simply fails to broadcast, even though a multimeter on the cell at rest still reads a respectable voltage.
So a tag with “charge left” can stop working on the first cold morning after it crosses some age. It doesn’t fade. It falls off a cliff. That’s why battery life in the field is lumpier and earlier than the linear-drain math on the datasheet predicts.
The problem you can’t see from the cloud
A dead tag and a tag that’s simply out of range look identical from the gateway’s perspective: silence. You cannot tell “battery died” from “asset left coverage” from “asset got buried behind a container” by absence alone.
The only thing that distinguishes them is history. A tag that reported battery telemetry and a steady detection pattern, then trended its voltage down over weeks, then went quiet — that’s a battery. A tag that was strong and then vanished mid-shift — that’s range, or the asset moved. Fleet watches the voltage trend and the last-seen pattern precisely so a tag can warn you before it goes dark, instead of you discovering it the day you needed the asset’s location and got silence.
What we recommend
- Tune the advertising interval per deployment, not per datasheet. Match it to how fast assets cross coverage and how soon you need the detection. This is the biggest dial; set it deliberately.
- Spec the cell for the worst month. If the asset sees -25 °C, a CR2032 sized for a warehouse is the wrong part. Bigger cell or lithium thionyl, every time.
- Replace critical-asset tags on a schedule, not on failure. For an asset where an hour of “where is it?” is expensive, a $3 coin cell on a calendar beats waiting for the cliff. Batch the swaps with a maintenance visit you were making anyway.
- Trust the trend, not the resting voltage. A cell that reads fine at rest can still brown out under a cold transmit pulse. Watch the telemetry curve, and act on the warning before the gap.
The tag really is the simple part of the system. But “simple” and “maintenance-free” aren’t the same word, and the gap between them is a coin cell that the datasheet measured under conditions your yard will never offer.