Why my battery dies in cold? (Part 2)

Finally, part 2 has arrived to shed more light to the question, why my battery dies in cold.

In Part 1 we talked about the unavoidable performance loss of lithium batteries in cold, and how the issue can be mitigated. Now we take a look at how Apple and Samsung, the biggest players in the smartphone industry actually handle the cold.

“Nothing burns like the cold.”
George R.R. Martin, A Game of Thrones

As Mr. Martin quoted, the cold really does seem to burn out your battery faster. We tested Apple iPhone 6s and Samsung Galaxy S7 in temperature controlled chamber. Our first test was to see how the battery voltage behaves under low load (video playback) in different temperatures. The devices were kept in each test temperature with full battery for one hour before turning on the use case.



As expected, the average battery level gets lower and lower as the temperature drops. Iphone is notorious for refusing to work in minus degrees, but keeping the device in charger helps in getting around this hindrance. Still, iPhone turns itself off after 5 minutes in -25 degC.

The next graph shows how the battery charge level behaves as the ambient temperature starts dropping quickly to -30degC. The use case was video recording, and in normal temperature our both test devices can sustain the use case for over 3 hours.



Apple’s approach to cold is very straight-forward : The device just shuts down when the battery temperature reaches zero degrees Celsius. This is not the most user-friendly way, but it makes things really simple for the battery monitoring algorithm and the cold temperature does never cause any surprises. Samsung, on the other hand, seems to have a great battery which works really well even in ultra-low temperature.  However, both devices are using lithium batteries, so the difference between their behavior shouldn’t be this drastic.

To see the realistic cold behavior of the devices we need to look past the battery monitoring algorithm. We did a test where we checked how the battery voltage behaves in different temperatures. The battery was at 100% level all the time. This was done by keeping the device in charger between test steps, and this method also kept iPhone running in colder temperatures. In the next graph you’ll see the minimum voltage values during the tests. The use case was to take 5 still images in a row. The rest time (device in charger) between test temperatures was 1h.



The maximum voltage during the test was 4,2V – 4,3V (the target voltage for charging is reduced in cold). The lowest measured voltages are pretty much at the same level for iPhone and Samsung, but at -20C temperature iPhone initiates a shutdown. This is most likely due to battery voltage level going below 3V which practically means an empty battery. Samsung just keeps on going to -30C, and there it shuts down without any warning after taking a picture. The reason for this is the extremely low battery voltage; anything below 2,8V can potentially initiate a HW level shutdown immediately.

It seems, that Samsung is taking quite a ballsy approach to cold temperature. They don’t really care how low the battery voltage goes, and the approach actually seems to work quite well, at least in these tests. However, you have to remember that the battery was full, giving a lot of headroom for voltage to drop. If the battery was at more typical level of 3,8V (50-60%), the shutdown could have happened already at -15C. Apple’s method in cold is predictable but their battery together with their low-consumption engine would also enable a bolder approach.

So while you are outside and it’s cold, remember that your device battery is losing performance. It doesn’t matter whether it’s a phone, wearable smartwatch or any other battery operated device.  After spending 2 hours building a snow castle, taking a picture of it may be more difficult than you thought.  Just like you like to keep yourself warm, remember to keep your batteries warm too !




Leave a comment

Your email address will not be published. Required fields are marked *