Same 18650 battery size, different voltages, how to avoid a catastrophe?

[amyk]

Playing "spot the industry shills" in this thread is far too easy...

But seriously, all you need to do is specify the required battery type clearly. It's not your problem anymore after that if some idiot decides to put the wrong one in. I'd not choose phosphate unless there's a special need for that type, as they are more rare. Regular 4.2V lion is far more widely available.

[westfw]

There is no way to detect which battery is being used, as the voltage depends on the charge level

Can't you check the charge current at the end of the of the LiFePO4 charge cycle?  Ie, if you you set the charge voltage to 3.7V and have an LiFePO4 cell, it'll be fully charged and in "trickle" mode by the time cell voltage reaches 3.7V, but a normal LiIon cell will still be drawing significant current...

[westfw]

If you apply 4.2V with a 500 mA current limit to a 3.2V LFP cell

But if you apply 3.7V with a 500mA current limit to a 3.2V LFP, it will draw 500mA *until* it approaches full charge, at which point current would die off, but the LiIon cell would still be drawing close to 500mA.

[tooki]

The only rechargeable products I've ever seen that use user-accessible, standard-size, loose rechargeable cells (not battery packs) are cordless phones using NiMH cells. Every product with NiCd that I have ever seen used either custom cells (like the early gumstick batteries in some old Walkmans, before they switched to NiMH) or battery packs. Is there such a product out there somewhere? Probably. Doesn't mean it was a good idea then, either.

The products were anything where an end user might use rechargable NiCd cells in place of the intended alkaline cells, like my calculator, my flashlight, my remote control toy car, my portable cassette player, etc.  My retired cordless phone is the only thing I have seen that came with loose AAA NiMH cells, but that was only a few years ago.

I have a whole collection of AA and AAA Eneloop cells for various things.  If there were Eneloop 9 volt batteries, then I would have some of those also for my multimeters.  Nothing I use them in was intended to use rechargable cells.

But I did not say “devices that NiCd batteries can be used in”, I explicitly said rechargeable products, that is, products where rechargeable batteries are the intended power source. My point being that manufacturers of products intended to be used primarily or exclusively with rechargeable batteries rarely, if ever, designed them to use loose NiCd cells.

I have a whole collection of AA and AAA Eneloop cells for various things.  If there were Eneloop 9 volt batteries, then I would have some of those also for my multimeters.  Nothing I use them in was intended to use rechargable cells.

Putting loose cells into something designed for alkaline cells is expressly outside the scope of my claim. 

So... use insulated wire of adequate gauge for all of the wire before the fuse?

I think the rest of us are struggling to envision the problem you have in mind. Can you give an example?

Sometimes a wire inadvertently gets pinched, but like modern lithium cells, the insulating sleeve around the battery could be abraded, or I saw one time where a left over cut lead fell into the battery area and shorted across the terminals.  Does anybody remember metal battery holders from companies like Keystone?
[/quote]So… poor battery pack design or abuse.

Not the

The fuse has to be located somewhere leaving some of the wiring unprotected.

to which I was responding. That statement implies that it is unavoidable to have exposed wire.

Meanwhile, risk of cell sleeve damage is likely one of the exact reasons that bare LiIon cells aren’t intended for consumer use, and the reason that product engineers of LiIon and NiCd rechargeable devices either build the cells into the device itself, or into a battery pack that has a lot more insulation or even a hard shell.

The point is that NiCd cells can also be a fire hazard under short conditions, so this is not a new hazard and it was something to be aware of before lithium batteries became available. At least NiCd cells were not combustible in of themselves.

Nobody said it was a NEW hazard!! We are just saying that it IS a hazard. And LiIon’s own flammability is precisely why they’re so much more hazardous.

You keep arguing based on things I did not say, and then ignoring things I did say. Cut it out, man.

[tooki]

Playing "spot the industry shills" in this thread is far too easy...

I don’t think anyone here is an industry insider. (And referring to people as “shills” is just cheap pot shots, typically used by ignoramuses.)

But seriously, all you need to do is specify the required battery type clearly. It's not your problem anymore after that if some idiot decides to put the wrong one in. I'd not choose phosphate unless there's a special need for that type, as they are more rare. Regular 4.2V lion is far more widely available.

Hah, how naive are you? Product liability is a serious concern. You have to design products around what people are likely to do, and fine print cannot definitively absolve you of that. (Various laws cover such things, and you cannot EULA yourself away from that.) And a key consideration is, as a manufacturer do you really want to test that theory in court? Do you want to spend hundreds of thousands, maybe millions of dollars defending yourself in court if someone dies? What will that do to your company’s reputation?

Heck, even the mere accusation can ruin your reputation, even if you win. Major companies (like Samsung) have had to perform major damage control in the past when there were reports of their devices, like phones or cameras, catching fire in people’s pockets, even after it was determined that cheap third-party batteries were to blame!

[tooki]

Primarily, you can and should use a thermocouple or NTC resistor to make sure the cell isn't overheating during charge or discharge. Every commercial device I've owned that uses 18650s contains one of those, as does every reputable eBike BMS. It might just be the vape pens that skip that one.

Thermal monitoring is considered a baseline safety feature for practically any LiIon/LiPo system. Only very small packs do without.

But like the battery protection (over-/undervoltage cutoff, overcurrent cutoff) chips used on rechargeable lithium, that’s considered a “backup”, it’s not intended to be a primary charging limiter like you’d be using it as. The primary cutoff is supposed to be a properly configured charging algorithm.

Note also that safety isn’t the only reason for wanting proper, accurate charging parameters: even if you are within what’s safe for the non-identical replacement cell, that might not be what’s good for it, resulting in reduced cell lifespan.

[westfw]

The only reason a li-ion cell "still draws close to 500 mA" when it's at 3.7V is that the charger's voltage limit is set to 4.2V. If the voltage limit is set to 3.7V, no current will flow when the cell reaches that level.

Fair enough.  I was assuming that LiFePO4 cells had a charging profile similar to LiIon, but with a lower termination voltage.  Apparently that's not the case, and the LFP profile is closer to "constant current" throughout.  :-(


With my original (incorrect) assumption, if you put 3.0V (discharged) cells into a charger at constant current (voltage setting doesn't actually matter during the constant-current phase) then by the time the cell voltage reached 3.7V the LFP cells would be fully charged and drawing minimal current, while the LiIon cells would still be in their constant current phase and drawing significant current.  It's a shame that's not the case :-(

When the charge voltage is removed, the LFP cell will gradually settle back to 3.2V.

Does that happen after charging, or only when you start to discharge the cell?  And how gradually is "gradually"?
Either way, this seems like a "detectable behavior" for a sufficiently intelligent charger.  A LiIon cell charged to 3.7V will remain at approximately 3.7V for quite a while; an LFP cell won't.

OTOH, having done a bit more research, it seems that many sites claim that it is safe to charge LFP cells to 4.2V, in essentially a "constant-current only" cycle.  They don't say what effect this might have on cell life, if any.

OTTH, the web info on LiFePO4 seems to be relatively polluted with people selling 12V and 24V "lead acid replacement" batteries, happy to claim that you can just throw one into your car or into your solar system, and have it work with the standard lead-acid charging electronics.

[KE5FX]

So, after many years i'm finally starting a project which involve the terrifying (according to media and shipping services) lithium battery. I identified a balancing IC and a charger IC that i like but when looking for 18650 batteries i noticed something that made me worrying: both LiIon and LiFePo are sold as 18650 cells.
They have vastly different nominal and charging voltage,  and now I have many doubts. What happens if I design a product to use 18650 LiIon cells and someone uses LiFePo cells instead? LiFePo cells have a much lower charging voltage (3.7 max) and it can get dangerous to go above that, especially at LiIon levels (4.2V). There is no way to detect which battery is being used, as the voltage depends on the charge level, so can't configure the charger and balancing circuit on the fly with some microcontroller, so what to do?
One solution i could find is to design everything to use LifePo cells to maintain a safe voltage for both types, but this way i'm stuck with lower capacity and if someone uses LiPo cells i'm basically throwing away half of the capacity.
Another solution would be to consider everything a LiFePo cell untill a full charge cycle where i can accurately measure the voltage curve, and with that detect the type of cells, but sounds quite complicate and prone to error...

Am I missing something? I find very dangerous that 18650 cells of the same exact size can be so different. Granted they have integrated protection circuit to prevent overvoltage, but still...

Would it make sense to consider 14500 LiFePO4 cells, like the ones that Tenergy used to sell as the 30225?  Looks like similar parts are sold by DigiKey from a company called Ultralast.  Unlike 18650s these are intended for consumer use, specifically in outdoor path lighting.

The scary thing about these batteries, and probably the reason why they haven't really caught on, is that they are virtually identical in size to AA primary cells.  They have both the voltage and the current-delivery capacity to utterly destroy almost any AA-powered device that an unwitting consumer plugs them into. 

There are also AA and AAA NiMH cells, which are very widely sold over the counter here in the US.  I've used quite a few of them in isolation amplifiers and other gadgets built into Hammond boxes that I keep around the workbench.   They offer a good combination of low self-discharge (similar to Eneloops, which have always been famous for that) and capacity, but aren't competitive with lithium of course.

Failing all those ideas, I'd vote for tooki's suggestion of using a pack form factor that's already been semi-standardized in other industries. 

[james_s]

But if you apply 3.7V with a 500mA current limit to a 3.2V LFP, it will draw 500mA *until* it approaches full charge, at which point current would die off, but the LiIon cell would still be drawing close to 500mA.

The current dies off because the output voltage of the charger is limited according to the cell chemistry. If you connect a 3.7V LiFE cell to a 4.2V Li-ion charger the cell will happily charge right up to 4.2V, if it doesn't catch fire or trip the internal safety first. With most battery types the cell voltage is inherently limited but that isn't the case with lithium ion chemistries, or if it is, it's above the voltage at which the cell will destroy itself. You can't trickle charge a lithium ion battery with a constant current source, it will keep right on charging beyond the maximum cell voltage.

[tooki]

Yep, this fooled me too for a long time. The "constant current" mode is actually just a constant voltage mode with a maximum current limit applied after the constant voltage.

That’s literally what a constant-current supply is. There can always be only one limit acting at a given moment, whether you’ve configured both or just one of the two. Only a conceptual ideal current source has infinite voltage available. In reality, the supply voltage always has a maximum. The maximum voltage  of a constant-current power supply is know as the “compliance voltage”. Your constant-current output voltage will never exceed your compliance voltage because that’s all you have available.

[David Hess]

Speaking of catastrophe:

https://youtu.be/qJ8BzvbeRXQ?t=155

[amyk]

Playing "spot the industry shills" in this thread is far too easy...

I don’t think anyone here is an industry insider. (And referring to people as “shills” is just cheap pot shots, typically used by ignoramuses.)

You sure do sound like one. The same as the anti-right-to-repair authoritarians who spread fear and paranoia because it's against their greedy $$$ interests.

Hah, how naive are you? Product liability is a serious concern. You have to design products around what people are likely to do, and fine print cannot definitively absolve you of that. (Various laws cover such things, and you cannot EULA yourself away from that.) And a key consideration is, as a manufacturer do you really want to test that theory in court? Do you want to spend hundreds of thousands, maybe millions of dollars defending yourself in court if someone dies? What will that do to your company’s reputation?

There's a lot of lawyer-trolling, but we also have something called personal responsibility here. It's not "fine print" if it's clearly stated.

Heck, even the mere accusation can ruin your reputation, even if you win. Major companies (like Samsung) have had to perform major damage control in the past when there were reports of their devices, like phones or cameras, catching fire in people’s pockets, even after it was determined that cheap third-party batteries were to blame!

Let's not forget that one of the biggest widespread reports of fires with phones was with Samsung's own batteries.

[bdunham7]

The point is that NiCd cells can also be a fire hazard under short conditions, so this is not a new hazard and it was something to be aware of before lithium batteries became available.  At least NiCd cells were not combustible in of themselves.

I've had NiCd power tool batteries catch fire while in use for no particular reason.  Someone I know in the aviation business went to work one morning to fly a plane that used a NiCd battery pack instead of LA and found the battery pack on the ramp under the plane, smoking and melting the asphalt.  For some reason it had warmed itself up enough to melt out of it's case and continued all the way through the body of the airplane, leaving a set of smoking holes in the fuselage.

[tooki]

Playing "spot the industry shills" in this thread is far too easy...

I don’t think anyone here is an industry insider. (And referring to people as “shills” is just cheap pot shots, typically used by ignoramuses.)

You sure do sound like one. The same as the anti-right-to-repair authoritarians who spread fear and paranoia because it's against their greedy $$$ interests.

I have never worked for a battery company or anything related. In fact I’ve only just entered the electronics industry professionally very recently.

I just take safety seriously, that’s all.

Let’s swap the order of the next two points.

Heck, even the mere accusation can ruin your reputation, even if you win. Major companies (like Samsung) have had to perform major damage control in the past when there were reports of their devices, like phones or cameras, catching fire in people’s pockets, even after it was determined that cheap third-party batteries were to blame!

Let's not forget that one of the biggest widespread reports of fires with phones was with Samsung's own batteries.

I know that, and it cost them billions of dollars. But there were also issues with older Samsung phones with removable batteries, where phones caught fire with third party batteries — many of which were unknowingly bought by users who thought they were buying genuine ones, but got counterfeits instead.

Hah, how naive are you? Product liability is a serious concern. You have to design products around what people are likely to do, and fine print cannot definitively absolve you of that. (Various laws cover such things, and you cannot EULA yourself away from that.) And a key consideration is, as a manufacturer do you really want to test that theory in court? Do you want to spend hundreds of thousands, maybe millions of dollars defending yourself in court if someone dies? What will that do to your company’s reputation?

There's a lot of lawyer-trolling, but we also have something called personal responsibility here. It's not "fine print" if it's clearly stated.

But I’m not a lawyer (though I did take a basic business law class in college). And it’s not “trolling” just because it’s something you don’t want to hear!

You literally just provided an example of product liability causing a multibillion-dollar product recall.

You can state all kinds of things, but that doesn’t necessarily make them legally binding. There are safety regulations for consumer products, and you cannot evade those simply by stating exceptions. A core tenet of contract law is that any clauses that have to do with illegal things are not binding. That’s why you can’t escape a murder conviction just because the other person stated in writing that they wanted you to kill them.

The salient point here is this: You can have a signed liability waiver that says you’re not liable even in cases of gross negligence, but since gross negligence is a crime, that clause is unenforceable, so the other party can still take you to court for gross negligence. And since what constitutes “gross” negligence is quite fuzzy, allowing the use of incompatible batteries could be found to be grossly negligent by a court or jury. As a company, do you want to take that risk?

And what about your insurance company? The fine print in your liability insurance policy could very well require a duty of care, and if something happened and it was shown that you knew about a potential risk, but downplayed it or thought you could just weasel your way out of it with a waiver, the insurer may decide to deny your claim, leaving you on the hook for damages.

[james_s]

There's a lot of lawyer-trolling, but we also have something called personal responsibility here. It's not "fine print" if it's clearly stated.

Wouldn't that be nice? Unfortunately the fact that you (and I) think that personal responsibility should be prevalent doesn't make it so. There are countless cases of people doing stupid things, somebody getting hurt, and then somebody else getting sued for not doing more to prevent that from happening.

[Doctorandus_P]

And of course, stupid is the brother of ignorant, and often combined in the same person, and mixed with just enough slyness to deny any form of responsibility.