Idea: Intelligent Battery Charging

[Izomiac]

Most people here probably know about how lithium ion batteries lose capacity with age. The rate at which this happens depends on charge level and temperature. If a battery is kept charged at 100% capacity, then at room temperature it'll lose about 20% of it's total capacity after a year. However, if kept at 40% capacity only 4% of the total capacity is lost. While temperature can't really be controlled, charge level can. Therefore, it seems like one could exploit the fact that most people won't use their Pandora for 10 hours on a single charge everyday to prolong the life of the battery.

Here's a quick rundown of what I'm thinking. Say a person generally uses their Pandora for 3 hours per charge (determined by logs, i.e. charge rarely drops below 70%). It'd make more sense to charge the Pandora to only 65% capacity. This provides an average charge range of 35% - 65% rather than 70% - 100%. Keeping the average charge level at 50% - 65% rather than 85% - 100% should, in theory, prolong the life of the battery. (Of course, one should always have the option of charging to 100% for a long trip or whatever.)

Personally, I don't know how to write such a program, or if the Pandora will even allow software to control the charging circuit. But, it seems like something most people would find useful since it'd mean having nearly double the battery capacity after four years. Thoughts?

 

[hobbyman II]

battery's are cheap, they last ages.

IMO you add some app like this, then forget about it, two months later you go away on a trip for a few days, don`t take the charger cos the Pandora lasts over ten hours, plenty of runtime, right?, you see the problem?, I would rather have the battery at 100% after every charge, you never know when you will need it, besides, at the rate current battery`s are improving, in three years time I am gonna want to fit that new 50000mah pack anyway (its gonna take two years for the battery to drop to the capacity you are charging it too anyway, why deprive yourself of that extra capability for two years?)

hobbyman

 

[Squidge]

I know there are options to do what you suggest (Monitor battery charge and enable/disable battery charging) but I don't know if they are built into the Pandora (ie, I've not checked the hardware side).

However, I've had my current phone for ~4 years(!), and the battery still lasts for a week between charges.

I wouldn't worry about it personally, in 4 - 5 years, there will probably be the Pandora II anyway

(You'll only end up biting yourself in the ass, as hobbyman II says)

 

[sehs33]

However, I've had my current phone for ~4 years(!), and the battery still lasts for a week between charges.

Squidge is your phone a Nokia 6310i or Siemens SL45i? These are the only two phone models I can think of from 4 years ago that would last this long :huh:

 

[centus]

Hello,

New to writing in the forums (been mostly reading till I finally got permission from my wife to preorder one), I think this is a great idea, as for the implementation, it should be doable, my A2 (which was stolen about a year ago) was only half open source but I could easily find the part in code which controls the charging, I think it should be simple enough to implement and if it's not implemented as part of the original software, I think we can make this happen.

I'll keep your idea till I get my Pandora and the development environment set-up, when I get the chance to try it out I'll let you know.

As for forgetting about this feature, I can hardly see how you would forget about it, since the Pandora is probably going to have some sort of software indication of the battery charge and you can't miss the fact it's not fully charged before that long trip.

Thanks for the great idea!

 

[atomicthumbs]

My solution:

I plan to put together portable Pandora Travel Battery Packs with high-capacity lead-acid batteries and sell them. You can take them with you in a backpack or something and have triple capacity!

 

[Squidge]

Squidge is your phone a Nokia 6310i or Siemens SL45i? These are the only two phone models I can think of from 4 years ago that would last this long :huh:

No, it's a

 

[conso]

I'd implement it on demand.
Say you know, you won't have the opportunity to use the pandora for a serious amount of time. You click the "shutdown"-button in the GUI and then "discharge to 40% and then shutdown".
That way, you won't have to control the battery-level for storage.

 

[trooper]

Squidge is your phone a Nokia 6310i or Siemens SL45i? These are the only two phone models I can think of from 4 years ago that would last this long :huh:

No, it's a

I still have mine with the original battery, Mmmm, Must be easily pushing 4 years now, And still holds a charge for around 4 days.

Trooper

 

[aaron11193]

A lot of phones will last a week on one charge if you put them in flight mode

Sadly, My pink O2 XDA mini/i-mate JAM's battery seems to only last around 9 hours while i dont have flight mode on

Drops down to about 2-3 hours while using wifi

 

[Tor]

I wouldn't bother with the 65% charge option on the Pandora, unless it'll be much hotter inside than I expect. The lose-lifetime-of-battery-if-on-100% is very true on laptops, simply because they are hotter inside. It's much less of a problem as soon as the device is just slightly cooler. (In any case the only laptop I know about which tries to handle this has an 80% charging option, they don't go as low as 65%).

It's also really bad for the battery to be run completely down, if you charge to 65% you are much more likely to accidentally end up in that situation. And, according to lots of reports, that will be worse for the battery than if you charge it to 100% routinely.

 

[WizardStan]

It's also really bad for the battery to be run completely down, if you charge to 65% you are much more likely to accidentally end up in that situation. And, according to lots of reports, that will be worse for the battery than if you charge it to 100% routinely.

I said pretty much the same several months ago when I proposed a 40% discharge tool, and it was pointed out that commercial battery packs (most likely including the Pandora) actually have shutdown circuits built in which kill the connection if the charge drops below a certain point to prevent draining it out completely and ruining the battery.

 

[Tor]

That is true, but you don't want to get down to even that level - that security circuit is the last guard against destroying the battery. If you hit that level you will need special equipment to get the battery functional again.

No, just normal 'full discharge' (which is way above the 2.5V typical "final shutdown" level) is enough to reduce both capacity and lifetime of your battery.

 

[Izomiac]

The lose-lifetime-of-battery-if-on-100% is very true on laptops, simply because they are hotter inside. It's much less of a problem as soon as the device is just slightly cooler. (In any case the only laptop I know about which tries to handle this has an 80% charging option, they don't go as low as 65%).

It is less of a problem, but the effect due to charge level is actually greater (relative) at lower temperatures. In a 40° C laptop the battery loses 35% VS 15%, in a (mostly) room temperature pandora the loss should be about 20% VS 4%. If you wanna play with logs, keeping the charge at around 40% entails a 5.5X increase in battery lifespan at room temperature VS a 2.4X increase at 40° C. I couldn't find any graphs of charge level VS capacity lost, so it could be that you get most of the benefit at not charging past 80%, but with the pandora's battery life it should be safe to go a bit below that and not risk hitting 0%. (Especially since it should be logging what the normal usage is before it tries this technique, hence "intelligent" battery charging.)

Also, I would assume that the battery would cut-out when it hits the lowest "safe" charge level... I mean, how many people would want a few extra minutes of playtime at the cost of battery capacity? If that's not the case I'd consider it a design flaw. Of course, given some of the boneheaded design flaws I've seen in other devices I wouldn't be too surprised if lithium ion batteries behaved like this.

Gee, I wish my PDA and laptop were like your guy's phones. <_< (Other end of the spectrum... Need I say more?)

 

[senquack]

I love this idea of intelligent charging. Seems like a no-brainer to eventually include it as a firmware feature if possible.

 

[BookTracker]

i'd like to expand on this, one of the features id love to see on teh pandora is a power-charge button.
my laptop has one, it looks like a small bar on the left side of teh keyboard, when you press it a led goes up and it starts charging, so basically i have charge on-demand function with it, it solves pretty much every 'intelligent battery charging' issue, with a simple solution: your brain. ^^;



cheers lads, thats my 2 ¹/² cents.

 

[lardman]

Is the battery charging performed by the kernel rather than some hidden hardware + binary contraption (like the Nokia tablets)? In which case people can try out all of these ideas if they so desire by tweaking then recompiling it (and probably adding some sysfs entry to communicate with the outside world)

 

[Tor]

On a Palm Tungsten there are several charge levels which change the behavior of the PDA. Some of them are adjustable.

It stops charging at 4.24 volts (the batteries themselves typically have a built-in charge stop at 4.30 volts, while the dangerous level is 4.40 volts - burn time).

There's a warning at 3.78V, and a cut-off (it goes to sleep mode) at 3.76V. The latter two can be changed by software, so that you could e.g. change cut-off to 3.65 volts, giving you quite a lot of extra capacity in fact.

These cut-off levels are quite higher than for many phones out there. The lower you set the cut-off level the more you affect the lifetime/capacity of the battery. The built-in cut-off of the internal battery circuit is around 2.5V (that's when you will need special equipment to turn the battery on again).

I still wouldn't bother much with the less-than-100% charging though. I have many gadgets (PDAs and others) that I have owned for several years, some of them have a replaced battery, others have the original battery. They have been exposed to different charging schemes, but in all cases it looks like the capacity of the battery starts to get noticably down after 3 years or so and then it goes rapidly downhill. Much more so for phone batteries as far as I can tell, possibly because they are more exposed to lower discharges.

PDAs that have been routinely (that is, as often as possible) charged to 100% overnight seem to keep their battery as well as can be expected - I see no effects that couldn't equally well be attributed to age. I'm not saying that there's no bad effects at all of charging to 100%, I'm only saying that it may not be noticable relative to the general loss due to the age of the battery. For example, a 3.5 year old PDA I have, with its original battery, is in very good shape. It's been charged to 100% as often as possible.

Laptops, on the other hand.. I have lost count of the laptops which drops down to a battery capacity of 15 minutes or worse, after a year of being on the mains most of the time. I replaced the battery in one of them though and carefully avoided staying on the mains, I ran it on battery until it was down to 60%, then charged it, then unplugged it again (this was feasible because this particular laptop lasted 5 hours on a charge). After 3 years it has finally started to drop capacity a bit.

 

[BookTracker]

Is the battery charging performed by the kernel rather than some hidden hardware + binary contraption (like the Nokia tablets)? In which case people can try out all of these ideas if they so desire by tweaking then recompiling it (and probably adding some sysfs entry to communicate with the outside world)

im assuming this is a question to me, and so, yes its not built-in.
actually it didnt even work before kpowermanager got its update, now it works in windows(where there is a software for that), and in kde. maybe in gnome it will work too, but i've never tested.
and yeah i believe its kernel dependent.