2

Part of the research we are conducting requires us to estimate how detrimental our proposed algorithm to the battery life of the wearables. To do so, I opted to simply charge the wearable to 100% and have it operate for an hour while sampling the battery percentage on a per-minute basis. After that, for example, if the battery level has dropped to 95%, I'll reason that it consumes 5% battery per hour, hence it has a battery life of 20 hours.

What makes me consider this thought is my previous experience with battery-operated devices, specifically smartphones. The time required to have my old phone go from 100% to 95% was significantly longer than it took it to go from, say, 50% to 45%. Furthermore, the former timing was also influenced by how long the phone stood in charge after it was fully charged.

With this observation, I'm not sure if following an experiment procedure as described would yield reliable results. What would be the correct way to proceed with this? The device in question is a series 3 38mm iWatch.

Lastly, I'd like to apologize in advance if I'm not using the correct battery terminology.

cngkaygusuz
  • 129
  • 1
  • You can easily find the answer to your question if you would search for "battery discharge graph". – Oldfart Feb 20 '19 at 16:53
  • It seems battery discharge graph shows the change in voltage with respect to time. I do not have access to the device's battery's voltage. I am able to read the battery percentage, but even in case the voltage level is used to determine the percentage, it may or may not be normalized to show a linear degradation. I'm not sure if I can use the graph in question. – cngkaygusuz Feb 20 '19 at 16:59
  • How do you measure the "percentage" ? Power-wise the drain should be somewhat constant over time. The fact you have seen first 5% drained slower than the others is only meaning that *that* percentage was not calculated properly. – Eugene Sh. Feb 20 '19 at 16:59
  • 1
    @EugeneSh. The API provides a method, which gives the battery percentage. It is exactly the same thing you'd see if you check your phone's battery level on its screen. – cngkaygusuz Feb 20 '19 at 17:00
  • 1
    See update on my comment. Anyway, why can't you simply give it more time and take more data points? – Eugene Sh. Feb 20 '19 at 17:02
  • Well, to get a completely reliable measurement, I need to fully charge the device and measure the time it needs for a complete discharge, twice: both for when the device is idle, and running the application. For an experiment that is to be repeated, this simply takes too much time. Adding more time can help, but operating within reasonable time frames, the question of the drop in percentage being linear still lingers. – cngkaygusuz Feb 20 '19 at 17:08
  • 2
    Again, given a constant power usage, the percentage drop should be linear. In case the percentage is actually showing the remaining usable *energy* and not something else. But this is not that easy to calculate as for the modern batteries there are too many variables to factor in, and some of these are not always known. Sometimes it requires continuous operation to let the algorithm to collect some data and self-adjust, so your readings will get more accurate over time. – Eugene Sh. Feb 20 '19 at 17:17
  • Your answer was helpful, thank you. It seems after this point it boils down to how the percentage is calculated in the device in question. – cngkaygusuz Feb 20 '19 at 17:25

0 Answers0