A quick guide to accurate battery lifetime estimation of LoRaWAN devices

Qoitech

The Things Conference Partner

Posted on 10-12-2020

Every new development project starts with an educated guess. How should the product’s form factor look like? What type of battery will it support? For how long will this battery power the device? And so on. With a specific battery capacity in mind and an average current estimated it’s really just a simple calculation: capacity divided by the average current and voilá: the estimated battery life.
file

As easy as this calculation is, as seldom is that really an accurate prediction. Unfortunately. For a good reason, most battery manufacturers provide a comprehensive data sheet with information reaching from nominal voltage and capacity, operating temperature range, maximum continuous current, and much more. And yet it may not be enough to give you the insights for your particular use case and the scenarios that your device will work in.

This article will show you a few methods that you can apply daily throughout your development project to get more accuracy on your battery lifetime estimations.

#1 Measure your device and use the Otii Battery Life Estimator

For a quick, yet more elaborate and accurate estimation than the simple math above, you can use a feature included in the Otii desktop application. We call it the Battery Life Estimator. No need for HW or subscription, all you need are your device’s – preferably measured – characteristics.

file
Fig. 1 Battery life estimation for a LoRaWAN device with spreading factor SF12.

Three variables are important here to insert into the estimator: The battery’s capacity which you got from the data sheet, your estimated current consumption and the time spent in active respectively sleep mode during the duty cycles.
You can also collect the information directly from your measured graph. In just a few clicks you get the battery lifetime as well as the number of duty cycle iterations estimated for the chosen battery.

Assume a LoRaWAN prototype with a 620 mAh battery capacity and measured current consumption of 17 mA in active mode for 3.8 s and 1.3 uA for a bit less than 15 min in sleep mode, with spreading factor SF12. The battery life is estimated at 11.4 months, as shown in Fig. 1.

The real advantage of using the Battery Life Estimator is that you can check for various different states your device may end up in - for example changing proximity to the gateway and therefore, a changing spreading factor. In other words, it allows you to do quick sanity checks for simply every change in hardware or software, for every use case and scenario.

#2 Talk to the battery manufacturer

When you decide on a battery type for your project there is probably no-one else that knows more about it than the battery manufacturers. For example, French battery maker Saft provides comprehensive data sheets for each of their batteries handing you data for all kinds of scenarios and explaining what happens to batteries in various conditions. Saft has also taken this further by offering their own battery life estimation tool for their portfolio, worth checking out.

What a simple battery estimation completely disregards is the temperature effect on electrochemical systems. Low temperatures slow down electrochemical reactions significantly and increase the internal resistance. On the other hand, high temperatures increase the self-discharge and in some batteries generate passivation on the liquid cathode system. Consequently, in order to set a sufficient capacity level for the device, its characteristics needs to be checked against the battery. However, the easiest way forward might be to actually talk to Saft and ask for their expertise.

#3 No more short-cuts, do a proper estimation

Now you have talked to the battery manufacturer and perhaps even requested samples of the batteries you are considering. You also did your sanity checks with the Battery Life Estimator, but there is still room for improvement. The battery self-discharge isn’t accounted for and neither is the device cut-off voltage that may limit the usable battery capacity, for instance.

To account for this, you want to introduce the Otii Battery Toolbox to your development work. With this feature, batteries can be profiled for specifics loads representing the device’s behavior. Connected to Otii Arc, you can test how the battery performs with a reasonable profiling time - something we call an accelerated device activity profile. Obviously, the more time you spend, i.e. less accelerated your activity profile is, the more accurate the results will be. However, unless you have a couple of years to wait until a battery is drained, you do need the acceleration to have the test results.

The profiling will not only provide you a discharge curve that you can later apply to your device, but also the battery capacity that is more realistic to expect. Using the Battery Life Estimator, you can then apply the new, more realistic battery capacity and the device’s behavior to receive a more accurate battery life.

The battery for the LoRaWAN prototype in previous example was profiled with an accelerated activity, resulting in a capacity of 492 mAh. With a measured current consumption of 17 mA in active mode for 3.8 s and 1.3 uA for a bit less than 15 min in sleep mode, the battery life was originally estimated to 11.4 months, however now resulting in 9 months.

Ok, let’s go one step further now. You have your device profile and also your battery discharge profile and now you want to find out the actual usable capacity. This can be done manually with the Otii Battery Toolbox, by changing the used capacity and observing when the device shuts down or reboots or you can automate this with the Otii Automation Toolbox where scripting is used to get the usable capacity faster and with more accurate steps.

…but this is a topic for another time! Book a demo with the Qoitech experts and we’ll tell you all about it.

#4 To sum up

Battery life estimation is an important part of the development work and you want to check it regularly throughout the project. Use different methods in different stages of the development and talk to the battery manufacturers to save yourself a lot of time experimenting with unsuitable batteries. Make use of the Otii Battery Life Estimator that you can download for free as part of the Otii application and utilize the tool provided by Saft to account for different temperatures and conditions (there is a really good video from Saft during last The Things Conference about this).

And like always, these are the best practices that you can start with today, but be assured that we are working hard to bring you even better and more realistic ways to understand your current consumption and calculate your battery lifetime.

So, stay tuned and join us at The Things Conference in January 2021 for more discussions, how-tos and examples on the topic!

Register today and use our discount code: TTC21-FRIEND-OF-QOITECH

Join The Things Conference

file