You can’t bring a product to market until you know how it performs.
For IoT developers, that requires a strong understanding of two things: your device’s power usage and the battery’s ability to supply that power.
IoT power usage is a complex subject. We won’t get into the topic here, except to recommend that you test (and retest) devices in every conceivable use case.
To get an idea of how your battery will perform you have a go-to resource in the battery datasheet. Battery manufacturers ship their products with this official documentation. These datasheets provide a basic set of battery specifications, such as the following:
- Operating temperature ranges
- Discharging/charging current
- Capacity
- Nominal voltage
- Internal impedance
- Performance examples based on standard use cases
Battery datasheets may also include safety information, recommended testing procedures, and other technical details, depending on the manufacturer. They’re a great source of baseline expectations for how a battery will perform, so you can begin to determine how long your device will last in the field.
The key word here is begin. Battery datasheets can’t predict the actual performance of your IoT device. We’ve seen IoT developers design products based on datasheets and hope for the best—only to end up with device failure in as little as one-fifth of the predicted lifespan.
Of course, that doesn’t mean we don’t need datasheets. It’s just that IoT developers must understand the limits of this resource—and learn how to get the performance data that’s missing from the datasheet.
Where Battery Datasheets Fall Short
Battery datasheets are vital sources of information for the IoT product developer. They give you the basic information needed to work with the battery—from the charging voltage to a general idea of available energy.
Battery manufacturers know their batteries. What they don’t know is your device.
Regardless of the chemistry, battery life depends on a range of factors that are entirely use-case-dependent. Change the temperature and performance changes. The same is true with elevation, vibration, and even gravitational force.
Here are just a few more examples of the factors that affect battery life, and may be unreliable or absent on datasheets:
- Pulse vs. steady discharge. Some battery technologies handle pulses—quick, high-current discharge—without losing capacity. Coin cell batteries famously do not. The information on the datasheet is probably based on continuous discharge, not pulses, so it’s hard to determine the actual life of your battery.
- True ambient temperature effects. Every datasheet gives an effective temperature range, but these aren’t always reliable. The datasheet for one lithium-ion pouch cell battery gave an operating temperature range of -20 to 60 degrees Celsius. We froze it to -10 degrees—well within the reported range—and found that our device wouldn’t even start.
- Battery recovery time. Battery engineers disagree about which chemistries require recovery between discharge cycles. Our experience suggests that more batteries benefit from recovery than we might think, however—and that skipping recovery intervals reduces battery life in some instances.
What do all these examples have in common? They depend on the real-world use case: where you deploy your device, how it uses power, how your sleep cycles are set up, and more.
In short, datasheets can’t tell the future. They can only help you make an educated guess as to which battery is best for your device. So where can you get the data you need to accurately estimate device performance?
Beyond the Datasheet: How to Accurately Predict Battery Life
There’s only one way to see how a battery works for your device, and that’s to test it in your specific use case.
Test battery performance as you continue to optimize your device for power consumption. Test the device in different temperatures, at different elevations, and in the very real-world scenarios your device is built to function. Continue testing before you send over-the-air updates. Build battery testing into your continuous integration cycle.
Don’t be intimidated by all the unexpected ways your customers may use the device. You probably won’t be able to test your battery in every possible real-world scenario. The more tests you run, however, the more you’ll learn. That can give you the data you need to make promises your device can keep.
And for IoT technology to become truly transformative, we need to make better promises. Battery-life anxiety leads many manufacturers to frame lifespans as maximums: up to a year, for instance.
That’s not helpful. A single day is less than a year. So is an hour.
With robust battery testing, you can start advertising device lifespans in terms of minimums, instead of “at least” a year. That’s information consumers can use to plan their deployments.
Just don’t rely on the datasheet alone to create your battery-life predictions for IoT. Energy optimization tools are available. So are battery-profiling software and battery-emulation hardware. With these tools in hand, you can go far beyond the datasheet—and start predicting battery life with data-backed confidence.
