Hacker News new | past | comments | ask | show | jobs | submit login

> Now Steve wants to warn the public about the dangers of LiPo batteries if left on charge for a substantial amount of time.

I've also occasionally heard recommendations from people to not leave devices charging because it may potentially wear out the battery prematurely. This may be a myth. But I always wonder -- why would the charging circuit not moderate the charging and mitigate such risks? Why should I as a user be responsible for providing just the right amount of charging?




> the dangers of LiPo batteries

I think it's Li-Ion batteries in laptops, not LiPo.

LiPos have a tendency to degas/explode unexpectedly even when the charger is good.

Multicopter/drones use LiPo batteries and there are multiple horror stories of pilots having their house destroyed. We have now learned to store them in special fireproof bags/containers even when not charging, never charge them unattended, always have a fire extinguisher near the charging station, etc.

BTW I see a phantom hanging on the wall... Maybe he tried to charge the laptop battery with the LiPo charger?


> I think it's Li-Ion batteries in laptops, not LiPo.

The important difference between the two is physical construction more than chemistry, though the electrolytes differ slightly. Cylindrical Li-ion cells have a rigid metal case and a breakable seal that will release gas and permanently disable the cell in the event of excessive internal pressure. Pouch and prismatic cells do not have a rigid metal case and are at higher risk of mechanical damage. Since they're made of softer materials, they often don't have a valve to release pressure, and bulge instead.

Pouch cells may be a little less safe than cylindrical cells for these reasons, but are present inside literally billions of consumer devices such as phones, laptops and tablets. Catastrophic failures are quite rare, hence Samsung's incendiary phones being a huge news story.

I suspect the high failure rate of drone batteries is due to the extreme loads drones put on batteries. Very few other devices are continuously operated at loads that will drain their batteries in a few minutes. There may also be an issue of sellers making claims about the maximum safe discharge rate of their batteries that are not based in reality.


> I suspect the high failure rate of drone batteries is due to the extreme loads drones put on batteries. Very few other devices are continuously operated at loads that will drain their batteries in a few minutes.

Indeed. The charge / discharge cycle causes physical strain on the carbon anode. Microscopic pieces of it break off with use, and are part of the cause of capacity degradation over time.


I absolutely loathe the degree to which manufacturers will embed LiPo's into just about everything including wearables without disclosing the fact. If you're clever about this sort of stuff and see a certain form factor coupled with larger than expected operating times you can infer the presence of LiPo's with some accuracy but your average consumer doesn't stand a chance and has no idea what the difference between all these different battery chemistries is in the first place and what the implications are.


But what good would a wearable or phone be with a nicad battery? Sure, there are dangers with lithium chemistry batteries, but sadly there are no viable alternatives. And as far as manufacturers not disclosing, what do you want for disclosure? Should we add yet another useless warning label to ignore, like the prop 65 warnings?

These incidents are relatively rare in consumer devices (so exempting RC lipos and cheap knockoff products with no protection).


https://en.wikipedia.org/wiki/Lithium_iron_phosphate_battery or LiFePO4 has much better safety properties. Any device sold to consumers should be required to have temperature monitoring, over voltage and over current protection. If it fails any onboard safety check it should render itself inoperable.


That sounds like a good development. I'm especially concerned about stuff worn against the body (earbuds, sports wearables, watches, cellphones). Laptops and tablets are also potentially dangerous but it's not as if every fuck up will immediately lead to injury.


I review flashlights and headlamps as a hobby. Many modern high-powered ones use removable Li-ion cells, most commonly 18650 size.

Temperature monitoring has become very common. The primary heat source is almost always the LED, not the battery, but it would be sufficient to detect a battery nearing thermal runaway. Low-voltage protection is common, but sometimes only a warning blink rather than a hard cutoff based on the idea that leaving the user in the dark might be worse than damaging a battery.

Some models have onboard charging, most often through MicroUSB. Correctly charging a single Li-ion cell is not difficult, and I haven't seen one get it seriously wrong yet.

LiFePO4 has seen a little use, but is not popular due to its much lower energy density. I suspect this has kept it out of other portable devices as well since users always seem to want longer runtime in ever smaller devices.


It was popular for pro power tools for a while. Those get a lot of abuse, and people are constantly changing battery packs so they max out on charging cycles. DeWalt used LiFePO4 from A123 Systems. But they seem to have given that up.


I honestly wonder why. I developed a portable small-quantity project a few years ago and designed in a 20AH LiFePO4 battery pack; it seemed like the only sane choice. The cost was slightly higher and density was slightly lower, but the difference was less than 10%. Not having to be concerned with battery failure modes was a huge advantage. But those factors must be different at larger scales; the largest scale I can think of where density matters would be for a commercial electric 18-wheeler tractor; what chemistry do those use?


Price. Ordinary lithium-ion batteries have dropped in price due to volume production and competition. LiFePO4, not so much.


Unfortunately we're getting to the point where there's rarely any choice.


30 years ago we didn't walk around with half a dozen bombs on our body. As a society we have decided that the benefits outweigh the risk, but it's still easy to not walk around with battery devices in our pockets/bags, or even charging at home. It's not as hard as opting out from cars and the roads, which kill dozens every week in the UK alone.


LiPos or Li-Ion?


LiPo, Li-Ion (not the LiPo variety, LiPo is also Li-Ion) is also dangerous but much less so from my own personal experience.


The second last sentence was a big red flag to me:

"Steve bought his HP Envy laptop in 2014, and had never had an issue with the charger. He only decided to leave it on charge when he unplugged it and it turned off immediately."

He knew there was a problem with the battery. He just didn't know the possible consequences.


I think it’s reasonable to expect users of self-assembled drone batteries to be aware of and careful about those batteries. These batteries are intended for RC hobbyists, and are likely to be subjected to physical abuse. I’m talking about the battery packs covered in thin plastic wrap with no integrated protection circuitry, like this: http://www.rapidrcmodels.com/turnigy-5000mah-2s-20c-lipo-pac...

But batteries in consumer products, like laptops and ready-to-fly drones, should be protected both physically (with a harder plastic case) and with integrated charging circuitry. That seems to be the case with consumer drone batteries, like the DJI Phantom: https://m.dji.com/product/phantom-3-intelligent-flight-batte...


A 20C discharge rate means that battery claims it can supply 100A @ 7.4V (740W) and be discharged in 3 minutes of operation. I don't have experience with RC batteries, but I'm a little skeptical. That's an awfully fast discharge rate, and a lot of power from a 282g battery.


In the RC hobbyist community it's fairly widely recognized that manufacturers inflate their C ratings. That said, these things can definitely put out a scary amount of power, and discharging batteries in 3-4 minutes of flight isn't unusual for some high performance RC aircraft.


That is absolutely normal in the high performance RC world. You even get 40C and higher rates cells. No point carrying around a heavier battery than necessary for the power required.


This is from a respected manufacturer and they've got 2-cell 7.4V LiPo batteries with a 120C discharge rate. There must surely be a way to verify this claim? https://teamorion.com/en/news-and-racing/team-orion-ultimate...


I have a bunch of 4s Quadcopter batteries that are rated to 95c. If I fly hard I can run it flat (to 3.6v per cell) in 3 minutes


I'd bet on that 95C being fraudulent (at least for more than a couple seconds at a time), but discharge in 3 minutes is 20C, and that's a lot.


Configuration: 4S1P / 14.8V / 4Cells Discharge Rate: 95C Max Burst discharge Rate: 190C


The short happens inside the battery itself because of the dendrite growth (https://www.youtube.com/watch?v=zwLUD41f15U)

By the time it happens, it is generally too late to shut off the charging circuit -- it helps and it happens but the battery destruction is already underway. The question next is whether it will catch fire or simply 'swell up'. At this point it's mostly up to luck, the available fire-prevention mechanisms are stronger battery enclosures, which would be heavy, bulky and expensive -- exactly the things most consumers do not want and would vote against with their money, reducing the profits of the company innovating in this direction.

Another way to prevent fires would be attempting to detect dendrite growth by the changes in observable battery state (capacity, amperage, etc). This is hard, expensive and will likely give a lot of false alarms requiring warranty replacement, once again hitting the would-be innovator in their pockets.


Literally last week I had a conversation with Apple about a swollen laptop battery.

They said it’s not a fire hazard but I should replace it and not use the laptop.

And, refused to replace it because it was in a “vintage” laptop.

Carefully crafted language.


That is why I run my "vintage" MacBook pro without a battery (first unibody, with swappable battery), always plugged in. It works for me as it stays stationary at home. If I need to unplug for any reason, I close the lid and macOS saves everything, then unplug, replug later, and nothing is lost.


Wow, I love that. Still, there must be some small coin-sized battery on the motherboard that's saving state, which will have to be replaced at some point?


No, it's hibernating; state is saved to the hard drive.


Problematic if you have a magsafe connector in combination with kids or pets running around though. Or just if you trip over the cable yourself. In my unfortunate experience :P


OK, machine turns off, rather than smashing on the floor (dog knocks it off, kid chasing dog puts foot through the screen)

I still haven't grokked the decision to abandon magsafe.


You should have sat down at one of their tables and plugged it in. :)


Second time my 2012 13" MBP battery swelled up. Glad on the older model you can at least replace parts yourself.

I think my problem might be that I treat my laptops more like desktops and keep them plugged in 24/7 and hardly ever run them off batteries. Last battery only had 2 cycles on it.

I want to get into vlogging at some point, so might at some point upgrade to a 15 inch with a GPU. I love Apple's software but disappointed the newer hardware isn't user serviceable anymore.


6 months ago is 'vintage' to them. How old was the laptop?


Actually "vintage" has a specific meaning and is at least 5 years after the product is no longer manufactured. See https://support.apple.com/en-us/HT201624.


It’s old. I think 2011.

Just replaced the battery with one from ifixit


That's exactly what charging circuits are supposed to do. I suspect the circuit in this case was damaged or faulty.


Probably faulty, since HP has a laptop battery recall for HP ENVY laptops from 2014. https://www.consumerreports.org/computers-internet/hp-recall...

This is just a good reminder people should keep aware of recalls for products they own, especially ones with batteries.


This probably takes HP off the hook. Still, as a consumer, how would one know? Do you have to periodically search yourself for the case one of your products is faulty? This seems hard.


Those product registration cards/forms/preinstalled apps that you just throw away/uninstall when you get a new computer or other product are there for exactly that reason. If you register your hardware, and it gets recalled, you should be notified.


What about those of use who prefer not to be tracked by every piece of technology they own. I don't want Samsung to know I have an S4 and Note 7, but I would like to know if there are any recalls for those devices.

It would be nice if one could run some service _locally_ that would parse some standards-compliant recall notice format from some standards-compliant location (like robots.txt, how about recalls.txt). I know, not going to happen.


It seems like it would be better if HP just (or also) notified their resellers, who could in turn notify their customers.


That works okay at the "notify CDW and Best Buy" level (and to some degree, they probably do), but not at the "notify every non-chain store that is reselling computers" level. And then you're also placing the burden on every reseller to keep track of which PC models (and serials) they sell their customers indefinitely. Which, again, not a big deal at the CDW and Best Buy level, but much less practical for many others.


Stick it in a spreadsheet.

Useful reason to have for why you're storing marketing data under the GDPR.


Lenovo does this. I was recently notified about the X1 Carbon (2017) recall/inspection (potentially loose screws inside the case) by the reseller.


How is the reseller supposed to track me after I move to the other end of the country and my mail-forwarding order at the Post Office has expired?


This immediately made me think of haveibeenpwned.

Reading the replies in this subthread, it becomes (somewhat) blindingly obvious that this is an excellent space for a viable startup.

- Mobile app (all platforms) that you can carry around with you at all times so when you go "oh! I haven't registered my blah" you can do it in 30 seconds on the spot (or at least start the process) - bam, inertia gone

- App works by having you take photos of device(s); it figures out what kind of thing it is, clarifies specs it can't work out from the photo, and tells you where to find the serial number for the given device (and I can't see why you can't take a photo of that too and the app can OCR it)

- It partners with hundreds of manufacturers to communicate and electronically sign your registration information

- When you move it handles the info update process for you

- (Manufacturers that require unambiguity could send you verification emails, and while the app would (obviously) not have access to these emails or the verification clicks, it would know what files had complete/correct/wrong info and be able to call out the action items in the flood of emails you would periodically receive)

And obviously, last of all,

- When a device is recalled or anything goes wrong with it, you receive recurring app notifications and periodic emails to let you know what's going on

- The app has a one-tap flow to allow you to say "I want to return my device" which would kickstart the process of for example the manufacturer mailing you a return box

---

This would be hard. Really hard. Stupidly ridiculously hard; the key word here is "vendor compliance".

But if you could pull this off, you'd wind up "baked-in" as a part of worldwide industry. You'd be the app everyone used to handle recalls if you got everything right.

I don't think this exists yet, at least not in the form I've just described.

I constantly see little recall dockets (A4 papers) at my local supermarkets (in Australia). Sometimes there's a photo, sometimes there's only a text description of the product.

Everyone's handling this themselves, and duplicating work. This system could provide centralized returns handling both for vendors and consumers. Hmm.

This seems like a fun one to take a crack at, and I'm interested enough to have a go. It'd probably take 5 years to get to minimum viability though, FWIW.


No company truly wants its customers to be made aware of recalls for its products. A business has very little to gain (some goodwill), but much to lose (a lot of money). They put in the minimum effort mandated by law, and have no incentive to open themselves up to increased recall numbers. Companies are not going to willingly provide APIs that allow third party companies to simplify or centralize registrations.

Thus, a company wanting to do this would be left performing manual data entry or automated scraping of existing sources. It would also be nice to notify users of products which have multiple reports of a defect, before recalls are put out. However I suspect companies would be quick to sue over unsubstantiated claims.

Monetization of such a product would be limited. Very few people would pay for this service. You're essentially left with building a profile based on a user's purchased products, used to sell ad space and/or push affiliate links. With how infrequently a user would be opening the website or mobile app, I can't imagine the income being enough to sustain even a small startup. Assuming there are existing public catalogues of recall information that can be scraped, I could see a solo developer doing this as a quick project for some side income. I don't see this as being a full startup due to, as you put it, "vendor compliance".


Thanks for the insight.

I just thought about some of the recall info sheets I've seen at the local supermarket: the cable entry on an electric blanket was misdesigned and had the potential to catch fire (:D), certain food items unknowingly contained allergens (presumably due to insufficient communication with manufacturers), the packaging (removable twist top) of a package of baby food could disintegrate (small pieces of the safety seal could break off) and present a choking hazard, etc.

These examples of recalls all share the element of preventing further injury and minimizing the chances of some disaster occurring. Hmm, I see what you mean by the goodwill aspect now.

As for sourcing data and payment, I was actually envisaging a two-sided product - creating the app described before for consumers, but also providing retailers and manufacturers with business-specific portals that allowed them to manage all aspects of a recall internally. I was thinking along the lines of "you store your data in the cloud so you don't have to maintain the associated infrastructure; this is the same, let us understand how you manage your recall processes so we can add this knowledge to a mega-system we build for you that will be ten times as good as what you've already got, and then just pay us $lessthanexpected to maintain it". Essentially solicited outsourcing.

With that done, that solves the industry integration problem (oh, and keeping the lights on); once vendors hit "publish" I'd have my customer-specific data.

Besides B2C this could work for B2B recalls of industrial and commercial equipment too. There could be pricing tiers to manage configurations of fleet or other large volumes of business-related equipment too.

Also, while pondering the idea earlier, I realized I could also expand horizontally to include warranty - it seems relevant, and I'm collecting product info anyway. Although about two seconds after the "that would be really cool" little red alarms started going off as I thought about all the spam (in the form of fraudulent claims) that I'd need to figure out a good way to filter.

I was wondering about ways to benefit (conscientiously, if you'd believe it) from customers' product profiles as well. FB (as one example) loves data, but I don't know what their appetite is like in terms of data volume or depth, and whether I'd be comfortable with what that industry would want.

Lastly, I envisaged the app as "install and forget" - maybe even going so far on Android as to hide the app icon completely (if that's deemed polite?). No bandwidth is used (with background messaging etc) until a recall occurs - only then does it spring into action. (Maybe if warranty stuff is factored in it could alert when devices are about to go out of warranty too.)


> I can't see why you can't take a photo of [the serial number] and the app can OCR it

I recently bought a toaster which came with a "Photoregister" [1] icon, which apparently works exactly like this.

[1] https://photoregister.com/


Thinkpads can be told to limit charging, say, between 50-70%, but that is a very rare feature.


All Fujitsu Lifbooks I had in the last 10 years have another setting. They do not start to charge the battery if the are more then about 95% charged. So there is no need to take the battery out while working "on cable power". I often see somehing like "97% full, not charging"


That's a battery longevity feature and not a safety feature, though. Any consumer electronics device should be safe to charge to 100% and be left plugged in.


You can do this on the Dell XPS series laptops, as well as the (older?) Sony Z-series.

I have my Dell set to start battery charge at 75% and stop at 85%.


The XPS line also tries to detect what to do based on usage patterns. Neat BIOS feature.


>>> ... charging because it may potentially wear out the battery ...

It is true that overcharging a battery will do damage, potentially fire-type damage. The issue is that every "device" contains a charge controller to manage this. A laptop should not / will not overcharge its battery. By leaving it plugged in you are relying upon the charge controller to do its job. But remember that we trust such voltage regulators 24/7 in our cars, on our planes, and with nearly every consumer electronic plugged into a wall outlet.


The user may want the battery charged asap but the intelligent charger might decide that now isn't the time for that and really annoy you. If the battery requires complicated charging and discharging schedule, like old ones that had to be discharged before charging, or periodically, then this has to be controlled by the user because you won't want to wake up one morning and find your charger has intelligently flattened the battery for you, or failed to charge it because it's time to let it drain all the way down.


A toggle "prepare off-grid use", nothing more complicated than that.

You could even try some implicit haptic UI ideas like having two power ports, one for operating (that only charges when the battery is dangerously low) and one for complete charging (that also runs the device, but it's clearly designated as the "charging port"). With USB power delivery that would not even waste case surface, it would be an entirely internal change to make one more USB-C power capable.


The other part being the recommendations not to always fully charge Lithium batteries. How come you can't configure the device or charger to do this? its 2018 already.


Some devices does this, if you are on Linux and are using a Thinkpad, you can fetch tp-smapi-dkms or acpi-call-dkms, and run the following:

    echo 40 | sudo tee /sys/devices/platform/smapi/BAT0/start_charge_thresh
    echo 80 | sudo tee /sys/devices/platform/smapi/BAT0/stop_charge_thresh
Then it will stop charging once it reached 80% and only start charging when it drops below 40%.


I don't know if that it's exclusive on ThinkPads but I love that feature. Now I'm using TLP[0] to re(calibrate) the batteries (main and extra) on a ThinkPad running Linux, it uses "tp-smapi-dkms" or "acpi-call-dkms" behind the scenes.

[0] http://linrunner.de/en/tlp/docs/tlp-linux-advanced-power-man...


My Sony ZX1 sometimes pops up a message when I plug it in "Battery saver: your battery will be fully charged in time for your alarm at 07:00" or something like that; I can charge it faster with an option on the message if I want to.

It would be nice if this feature defaulted to charging to 95% or whatever. I usually know in advance if I'll need that extra 5% for a long journey.


I believe products already do this. A good few years ago I was asked by a guy why his plugged-in HP laptop was telling him it “wasn’t charging”. I looked up the docs and discovered that between 95-100% it would turn on/off the charging as required.

I’ve heard separately (no source, sorry [edit: see below]) that Apple products also do this, they just don’t confuse the user by telling them that it’s happening. I wouldn’t be surprised if HP et. al. have since changed their user notification behaviour, too.

Source: forum posts going back almost a decade. https://www.google.com.au/search?q=apple+laptop+stops+chargi...


Wow, that's a pretty neat feature. I think there are many people who leave their phone to charge during the night, only for it to 'fast charge' up in 1 or 2 hours and then sit at 100% for the rest of the night.


Here[1] is the screen.

Here, the time isn't based on an alarm (none is set), but on when the charger was disconnected the previous morning.

[1] https://i.imgur.com/1EywDA0.png


Devices already don't charge all the way, the "100%" is an arbitrary lower amount.

It would be nice if you could set it even lower to prolong life, but there's just not enough demand for it to be a common feature.


I could do that on my Thinkpad.


I liked that feature, especially so for office "luggables" that stays on the desk charging most of the time.

That really seems to spare the battery compared to keeping it at 100%. I've replaced the battery in other laptops much more often.


Many charging circuits in larger systems like vehicles do allow setting a limit below the theoretical maximum charge to preserve the battery's health. Options to do this exist in laptops, but mostly just on enterprisey, higher-end models, are not set by default and not very well known.


There are two parts involved here: 1) prevent the battery from exploding? This it should do. 2) Prolong battery life? This requires not charging to 100%. I think it's fairly clear why most users don't want their things to stop charging at 75%.


Just change the external scale. 75% is now 100%. Users are none the wiser and battery life is prolonged.


Then your competitor with the same battery cells says 85% is the new 100% and now they have longer battery life than you do.


And then maybe your competitor gets an announcement on every plane reminding people not to buy their product :-)


No, because again, this is not about preventing batteries from exploding. They already shouldn't explode. This is purely to give them a longer lifespan – which is not that attractive to producers, who I assume want to sell more batteries.


Ah yep, didn't read the thread carefully enough.


Or maybe they'll buy the competing product that has 25% more runtime, or is 25% thinner, or is 10% cheaper etc.


Most smart battery chargers do this. 0% is not really 0% and 100% is not really 100%.


This one goes to eleven.


The charger should prevent this, but if it's not designed to fail-safe it could be contributing to the issue...


And the battery circuitry should prevent this as well. So even if the charger is faulty there is supposed to be a thermal runaway protection circuit right on top of the battery itself that would cut the power to the battery before something catastrophic would happen.

Also, the battery should be given some room to swell in the design.


> because it may potentially wear out the battery prematurely. This may be a myth.

I thought it was a myth too, so when I got my first cell phone, I left it plugged whenever I was home, even when I had used it so little that it was still showing full charge. After a time, however, its battery started to get "fat".

I bought a new battery, and since then only plugged the charger when the phone was no longer showing full charge. That battery still works, and is still as "thin" as it was when bought.


Which is anecdotal, the first battery could have just been defective or failed earlier.


Sounds like it was a defective battery rather than charging controller. Although depending on how long ago you got your first cell phone, I'm sure a lot has changed.


Batteries are of varying quality and constituent purity, this is not a conclusive observation.




Join us for AI Startup School this June 16-17 in San Francisco!

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

Search: