Suggestion: (Imprecise) Time Remaining/Time Till Battery Dies

[Patrickode]

I know these things are going to vary wildly depending on what headset you're using, how old it is, or whether you've replaced the battery entirely like I have for my headset, but I think it'd be nice if the tool could display how much time the battery's got left before it dies, even if it was definitely inaccurate.

For example, let's say Headset A has a marketed/rated battery life of 40 hours. If we assume that it degrades linearly (it almost certainly doesn't), that'd suggest it has about 7 hours and 12 minutes left at 18% battery. Maybe a site like RTINGS or similar tech review site out there has some kind of data about what percentage levels mean on different headsets.

If this were a thing, it should probably be marked (ROUGH/NAIVE ESTIMATE) or something. It'd also probably be better to have the estimate be pessimistic, since the reason I would like and others might want a feature like this is to determine if the battery's low but has several more hours in it, or if it's in imminent need of a charge. An estimate that's too optimistic (like the above example?) could give false confidence and get in the way of that. You know, set expectations to exceed them rather than miss them and all that, à la the WoW resting rework. This would also mean, if the headset can't meet pessimistic estimates, you've got an early warning sign that it might need a repair or replacement in the not-too-distant future.

_{I know issues aren't exactly the place to put thank yous, but since I'm writing one anyway, I'd like to go ahead and say I really appreciate this little tool! I've been using it for a few months, it looks nice in my system tray, and works great. Kudos and well wishes.}

[auto94]

Oof, that sounds like it'd be a fun thing to add. Off the top of my head we could:

1. Have the user run a "benchmark". At 100% charge you press a benchmark button that tells you not to recharge until at least 5 or 10% battery remaining. During this time the app polls the headset every X mins (lets say 3 so its low but not too spammy), and logs the time of use + battery percentage. Then we can just put the time-until-5/10% into the taskbar pop-up.

• Pros: Accurate for each individual user regardless of battery health and even if they replace the battery. Could also ask the users to post their findings on Github (headset type, headset age, battery life in hours) and eventually have a decent estimate model.
• Cons: Requires user input.
• This could technically be done without user input, but it wouldn't be accurate at all if the user would charge it at eg. 40% already.

2a. Could just add the manufacturer/rtings battery life estimates, and hope the battery level is roughly linear.

• Pros: Easiest to add.
• Cons: With the Alpha (300hrs battery life), any battery degradation could probably cause the estimate to be off, by a lot. Probably also inaccurate relative to microphone usage/headset volume, but that's unavoidable i think.

2b. Have the user input their own estimate and linearize the expected battery life.

• Pros: Can't blame me for inaccuracies 😄
• Cons: Same as 2a probably.

I don't really think the battery percentage is linear to the time left. Personally I feel like it goes down faster at lower percentages but that might just be some psychological thing where I'm checking the battery more the lower is gets so I notice the change quicker. What have you noticed?

There's probably plenty of other ways, I'm open to any suggestions, though I don't feel like over-engineering this,

[Patrickode]

You're right, in my experience, it definitely isn't linear, but all I really knew was that it's complex to figure out how much life a battery has left. After reading this, I decided to look up how phones do it, since a number of them, including mine, give you estimates on how much time you have left. It's more complex than I thought, even, which I should have expected, as a programmer myself. I've seen mention of "coulomb counters" and such, which is out of my software wheelhouse. Measuring voltage + speed of energy drain + maybe a temperature sensor + maybe keeping track of how long the battery's lasted up till now, etc.

I also didn't think of volume or microphone use impacting battery life. So, it might not be feasible to provide a helpful estimate without serious shenanigans. The best bet I can think of without really sitting down and trying to plan something out is taking the manufacturer estimate, maybe plotting that against one of the charts I've found looking this up (linked below), then just rounding down or chopping off 30 minutes to make the prediction more pessimistic.

Maybe, like you say, there could be a setting you could turn on that'll keep track of your battery level over time, and use that with some or all of the above to adjust the estimate. If all that sounds like it'd be over-engineered for this tool, I wouldn't worry about it. It's not like this suggestion is critical in the slightest!

Li-ion Battery and Gauge Introduction [Voltage/Capacity in mAh]: https://www.richtek.com/Design%20Support/Technical%20Document/AN024

https://batteryuniversity.com/article/batteries-will-eventually-die-but-when-and-how

Compare voltage to manufacturer SOC (state of charge) curve: https://www.reddit.com/r/ElectricalEngineering/comments/nabl4b/comment/gxsn4hd/

How can software give an accurate estimation of remaining usage time for a worn battery?: https://electronics.stackexchange.com/a/368881

Cheers, and sorry for nerd-sniping you/myself. 😁

[auto94]

Yeah, it seems pretty complicated to do accurately 😄. I appreciate the research tho!

I think I'll keep this issue open for a bit and when I have the time, I'll try to implement my "user benchmark" method and see how accurate it ends up being. If you end up thinking of any other method, feel free to suggest it, it's a fun challenge.