Hacker News new | past | comments | ask | show | jobs | submit login
The life and death of a laptop battery (skolelinux.org)
81 points by wielebny on Sept 25, 2015 | hide | past | favorite | 58 comments



For some reason, despite the fact that keeping a Li-I battery near full charge and in high temperatures drastically decreases its capacity over time, most laptop (and phone) manufacturers don't bother to have battery management software (to limit state of charge). I only know of Windows drivers for Sony and Lenovo.

http://superuser.com/a/664583/32370

This is probably because of some sort of planned obsolescence - they want you to replace your device.

On the other hand, in electric cars (using the same battery chemistry), where it is in the company's best interest not to replace the battery, it is charged to a maximum of 70-80%, except when soaking up extra energy going downhill, and basically maintains its full capacity for much longer.


Well, people also complain (extremely loudly) that they paid for a 50Whr battery, and by golly, they mean to get 50Whr of power out of it! They're not going to let some slick talking corporate drone trick them into believing they should only use half of what they paid for.

Practically speaking, even assuming my battery did idle at 50% sitting on my desk, what happens when I decide to take it out to the coffee shop (or the airport!) and want it at 100%? I don't want to micromanage charging my battery and filling it up before I go out. I have enough fun discovering my car is only ever out of gas when I'm about to depart on a long drive.


So why not have it as an option so people like you who don't want to have to manage it can have it go to 100% and people like OP can stop st 80%?


> I only know of Windows drivers for Sony and Lenovo.

My Samsung has a BIOS setting to limit charging to 80%.


My Toshiba Satellite S55t-B has the Toshiba-eco utility that allows me to put it in eco-charge mode which also keeps the battery at a max-charge of 80%.


For phones, a smart charger that connects via Bluetooth to a custom app might be a realistic workaround. The charger would just need to be able to pair and to understand a simple on/off command, while the app would only need to be able to pair, configure desired maximum charge, monitor the battery level and provide some form of override GUI while the phone is charging. This feature set seems quite achievable by moderately advanced hobbyists and should even be somewhat resistant to dangerous feature-creep in case someone did a kickstarter.


This isn't so bad when the battery is removable. The idea that they are pushing for planned obsolescence is more believable when they deliberately make it difficult to replace the battery.


On the other hand, using the battery also wears it down. So it's better to stay plugged in all day at 100% and pull off the mains rather than to run off the battery and then recharge it.


My ThinkPad (T410) has the functionality built-in to the firmware (in the batteries themselves, I believe). While only Windows has GUI facilities to set them, I've been able to access the settings from the CLI in Linux as well.


Dell's battery utility allows you to set percentages when charging stops and starts.


I end up punishing the batteries in my stuff with 100% -> 0% -> 100% cycles frequently and don't feel bad about it. The battery exists to conform to my needs, not the other way round! I'm not going to use my stuff in patterns I find inconvenient just to baby an inanimate object.


I do this too, but for me it works really well.

my 4 year old iphone 4 got 3 days on a single charge when I got rid of it.

My year old 6 still gets 2-2.5 days per charge.


Fantastic! I will be giving this a trial on my laptop. My laptop is running Debian and it usually sits connected to the power all the time.

When I disconnect the charger I have seen estimated battery remaining times of anything from 2 hours to 87 hours! However in reality it's more like 45 minutes :(


You do know that keeping your li-ion battery plugged 24/7 is best way to kill it, right?

Do a full battery cycle at least once a month and you won't prevent the catastrophe.

Source: http://batteryuniversity.com/learn/article/how_to_prolong_li...

Storing a battery at 40 degrees C at 100% charge will degrade it by 35% after a year. Drop that to 25 degrees C and it'll be degraded by only 20% after a year. If it was stored at 40% charge instead of 100% charge that drops to only 4% degradation after a year.


Aren't lithium ion batteries micromanaged to within an inch of their lives by electronics within the computer and the battery? Doesn't whether your machine is plugged in or not have very little to do with what the charge current to the battery itself actually is? Plus, of course, there's no guarantee that when the battery meter reads 100% the battery is actually full --- mapping 100% in software to 80% on the battery seems like the easiest way to improve lifetime.

e.g. it's very obvious that Macbooks have multiple different charging regimes, and they use different ones depending on what the laptop's doing and the state of the battery.


They are managed, yes, to provide the greatest out-of-box specs that they can then use for advertisement. "No one" cares about long-term effects, especially if those effects will be beyond warranty periods anyways.


Why can't this problem be solved in software? OS knows that AC power is used 100%, right?


I believe a better question is: Why can't this problem be solved in hardware? So, it won't depend on the OS and if I shutdown my laptop and leave it on the charger it will still work.

My 3.5 year old Asus laptop has still similar battery life as it had when it was new. I always keep it on the charger when I can and sometimes I just shut it down and leave it on the charger overnight too (but this is rare). When I bought it, I have noticed that the battery was never above ~95%, but I have never studied how it works exactly (is it done by HW or by the driver). I have used it with Win7, Win8, Win8.1 and now Win10.


Thinkpads (used to, don't know about these days) ship with a utility that allowed to prolong the lifetime of the battery by limiting the amount it will be charged. The article mentions "tlp" software which apparently does similar thing. So yes, it can be solved in software.


My lawn mower's LI battery is connected 24/7 and still works after 10 years, using it every summer. My phone battery recently died after almost 3 years but my previous phone battery still works 5 years later, both 24/7 on charge unless in mobile use. I think/thought the prevailing advice was that car batteries like lead/acid should be cycled (actually I know this) and the same for NICAD and for LI it has always been best to charge 24/7. Sometimes batteries including LI go bad, it just happens.


Doing full battery cycle every month sounds like a very bad idea for lithium batteries. You generally want to avoid deep discharges as much as possible, this is even pointed out in the link you posted:

"The shorter the discharge (low DoD), the longer the battery will last. If at all possible, avoid full discharges and charge the battery more often between uses"

Keeping the battery constantly at stable 60-80% charge is far better than doing full discharges every month.


To a first approximation modern li-ion batteries don't really care if they drain 100->50 once or 100->90 five times: the wear on the battery is more or less equal. Exposing your batteries to extreme temperatures (below freezing, above 40C) or allowing them to self discharge below the safe low voltage limit has a far greater impact on overall longevity.


That is to say: a lithium ion battery under use in a personal electronic device, whether kept charged or cycled, will be pretty well had it after 2 to 3 years, or at least have you thinking "gee, this battery sure isn't what it once was." I don't think it's possible to avoid this.


Wow down to 7% of original capacity? I've always had battery woes but that's exceptionally bad. If you started with 6 hours now you'll get 25 minutes of battery life!

Does anyone know of any good power management software for Linux? I am about to get my first Linux laptop and this is one of my primary concerns (since the laptop is designed for Windows 10).


If you have not purchased already I strongly recommend re-considering and buying a made for Linux laptop. Putting Linux on hardware intended by the OEM of Windows is not always a good idea. I've not tried them yet but my next laptop will come from System76 (unless I see something better before then).


Almost no laptops are "made for Linux". For example, the last time I looked, System76 just sold rebranded Clevo laptops. So while you might be able to make the argument that System76 does more Linux testing than the likes of HP or Dell, trying to make the claim that System76's laptops are "made for Linux" is incorrect. I'm pretty sure that Clevo's are made for Windows.

They do the same thing you or I do before ordering a laptop we intend to throw Linux on: do some research first to find one that is mostly compatible, and then tweak the hell out of it until it mostly works. They just take the extra step of then selling the laptop. However, I can tell you from personal experience that System76 Linux laptops are far from perfect Linux experiences.


Unfortunately, I can't recommend a System76 laptop. I use the Galago UltraPro as a work laptop, and it is not an enjoyable experience at all. The construction is flimsy, and the keyboard is _terrible_. I have never used a worse keyboard on a laptop. I wanted to love it, but I can't recommend it at all.


Wait, did you ever get the replacement keyboard the mailed out for free? I run an ultrapro as well, and they sent me the redesigned keyboard without me saying a thing.


This was after they replaced the keyboards. The keyboard is still terrible.


Already ordered. I'll keep Windows in a partition I think, this is just an experiment to see if I can use a non-Mac for the first time in my life.


My Lenovo's laptop (IdeaPad z410) battery is already dead, and I've had it for a year and a half. When it had Windows on it, I get about 2 hours of battery time. After I fully installed Ubuntu, I got 3 hours out of the same battery. Now it never charges, always stuck at around 1%, and it shows that it has a capacity of only 11.4%. Two of my previous laptops' batteries degraded after the second year. This is the first time I have a laptop whose battery dies in less than two years. We have a trend here (as it seems from skimming the other comments).


Usually when you install Linux on a laptop you can expect any software (non-firmware) battery management to go out the window.


And that's a shame. Battery management should be on the hardware level. I can't find any information whether detection of AC-charger presence is on the hardware level or not, which I also think should be on hardware level to stop charging once the battery is full and fully draw current from the charger. That will not waste the battery's cycle life. Both of these speak for the overall quality of the manufacturers' products.


"should"... doesn't buy you much. It's a shame there's no standard driver for this, however, since these Li-ion batteries are all pretty much the same in terms of charge management. It's a safe bet almost all users would be better off with a batter that defaulted to around 70% charge unless explicitly (and temporarily) told to prepare for a long trip.


How do most people manage their devices(iPhone / MacBook / etc.) if leaving them plugged in all the time is detrimental to the battery? I could kill all the apps on my phone and put it in flight mode when I go to sleep, but I may not get the chance to charge it to full if I have to leave the house immediately upon waking up. Should I let my battery drain all the way?


Leaving laptops plugged in all the time doesn't seem to be particularly bad for the battery, as long as (1) the laptop doesn't get too hot, and (2) you actually keep them plugged in all the time, rather than switching frequently between plugged-in and battery-powered.

My Lenovo laptop has been plugged in for the last 3.5 years, only very occasionally (once a month?) going on battery power for short periods of time. Windows reports that I still have 91.7% of the design capacity. I'm not sure if I should trust that number or not... but the last time I unplugged it, the battery seemed to last at least 70% as long as it did when it was new, so it's not totally off, either.


Windows 8+ has a similar feature built in, to generate your battery report you just need to go enter powercfg /batteryreport in cmd

Windows 7 can also generate power reports (powercfg -energy) but it's a little less human friendly

Unfortunately for me it appears that installing windows 10 reset the stats, as my report only goes as far back as about a week after I installed windows 10.


Very impressive. Concise, thorough, useful. I love seeing posts like this.


7% remaining capacity after 2 years seems insane to me. I replaced my 2012 MacBook Pro battery this year as it passed 1000 cycles (I'm bad at plugging in and end up charge cycling daily) and even that battery was still above 70% capacity...


Interesting.

Did someone follow the link to the Macbook Air? It shows that the Mac battery lasts much longer than the Thinkpads. My impression is similar: my Thinkpads (back when IBM still labelled them) where down after 2 years. All Powerbook or Macbook batteries that I ever had still did well even after 3 years.

Shocking, on the other hand, is the amount of detail of the battery usage records that apple stores. Imagine apple had access to them.


One of the things Apple seems to be very good at is managing battery "overcharge": my 5 years old MBP reports ~90% battery capacity but a relatively low amount of cycles (~200) because it's almost always tethered. Other companies seem to handle that less well and have the battery health fall much faster.


It it possible Apple overprovisions their batteries such that they only ever charge the physical calls up to an ideal level (~80% or whatever, of nominal Wh capacity), and then they show in software a 100% charge? If the battery is larger than it needs to be, the extra headroom would allow long usage on a "full" charge while preserving cell longevity. In time when capacity does diminish, software could also raise the full-charge point from ~80% of physical capacity closer to the nominal physical cell capacity.

This would be similar to what SSD manufacturers do with drive sizes. There are more memory cells on the die than the sticker capacity, and as cells fall out of use (due to failure or maximum number of write cycles) extra cells are rotated into use to maintain speed and capacity.


Here's a link to my MBA performance: http://imgur.com/2UVFmIU

Bought Nov 2013, now has over 250 charge cycles and is still at 90% of original capacity.


How would I generate this graph for my MBA?


For each of my laptops, I use two batteries: a high-capacity one that I use when traveling and a regular one for daily use. When unused, I store the high-capacity battery (charged to about 40%) in the fridge, wrapped in a plastic bag. Thanks to this treatment, the 9-cell battery of my old spare laptop (Dell D430) is still good for 4-5 hours of use. It lasted for 6 hours when I bought in 2008.


(I know this post is about linux.) My windows laptop batteries usually die after 2 years. I think Windows laptops have the problem of overcharging. To avoid it, I have to unplug the battery when I need to use a laptop as a desktop. We have to admit Apple does good job in battery management. My 2010 MacBook Pro 13-inch's battery is still in good condition (relatively).


> My 2010 MacBook Pro 13-inch's battery is still in good condition (relatively).

My 2006 MacBook still holds about 30 minutes of charge in the original battery. Most of the time I left it plugged in all the time, what do people do that kills them so quickly?


> We have to admit Apple does good job in battery management. My 2010 MacBook Pro 13-inch's battery is still in good condition (relatively).

To answer with my own anecdote, my three year old thinkpad has had barely any drop in capacity, whereas a friends macbook pro of a similar age has only about 60% of its original capacity.

It's hard to compare these things because it depends so heavily on usage patterns.


My wife's Lenovo has a windows utility that allows you to default the charge max to 60%. My Dell has a similar option in the EFI settings.

After about a year, there has not been a noticable degradation in battery life (and I have noticed it in past laptops).


This should be the norm for all devices with built-in lithium battery charging. Default to a low attrition end voltage, and a simple hardware control to temporarily enable full capacity charging in anticipation of a particularly long off-grid phase.

But dead batteries are what keeps the economy spinning, so it must be in all our best interest that the self-optimizing algorithm of capitalism decidedly skips on this particular efficiency gain, right?


IBM Thinkpad actually has pretty good battery reporting tool that tells you condition, minutes-to-charge, cycle count, voltage, current, capacity Wh, temperature, mfr. name, mfr. date,design voltage, firmware, etc.


Pro-tip: suspend is not zero power state and can hurt your battery. I used it on my first work laptop and ruined the battery (brand new) in 1 year. Use hibernation.


Tradeoff between killing your battery or killing your SSD then I guess (if you have an SSD of course). And wasting time when booting, my experience with hibernation is that it isn't that much faster than a cold boot. Suspend is more or less instant.


I wouldn't worry too much about write endurance with modern SSDs - at least if you're not using them in a server or something. The Samsung 850 EVO (not even the PRO model) is rated for 150TB, and most tests got closer to double or triple that number before the drive failed.

If the full contents of a 16GB memory are written to disk for hibernation daily, with say another 50GB of work, that's over 6 years before you hit the lower limit. I probably wouldn't trust a hard disk that was used that heavily for that long either.


I've written 40TB so far to the 1TB 850 Evo in my workstation and it certainly has problems now that it didn't have before - it just hangs for a few seconds at a time, when copying or deleting large number of files. Not sure what it's doing.


If by chance it's an 840 (rather than the VNAND 850), there's been two firmware updates to improve performance over time - make sure to update the firmware.


How full is the drive? If it's more than 80% full you might actually be running into wear leveling issues.


It's unlikely that your OS is writing enough data to the SSD during boot to noticeably affect the drive life.




Consider applying for YC's W25 batch! Applications are open till Nov 12.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: