“Smell-O-Vision” was right there, waiting for you!
1 Like
That is a sign of a device that was slapped together (I won’t even say designed here) by someone who has no idea on how to work with batteries. You can’t rely on humans to not overcharge a battery, especially with the fact that they can fail so catastrophically when overcharged. What makes it even worse is that there are cheap and easy ICs that do all this for you, but they cost about a dollar or two more than what I call the “asshole charger” which is typical in products like that, but they’re not used because of ignorance, not caring and the slightly higher cost. These fly-by-night generic trend followers try to pump as much money out of a trend as possible by making things dirt cheap and not giving an ass about design or safety. It just has to work for a short enough of a time that they can just disappear and not be called out on totally scamming people.
Back to the asshole charger. So this image, courtesy of Digi-Key, shows the voltages and currents for charging lithium-ion batteries from dead flat to fully charged: http://www.digikey.com/en/articles/techzone/2012/sep/~/media/Images/Article%20Library/TechZone%20Articles/2012/September/A%20Designers%20Guide%20to%20Lithium%20Battery%20Charging/article-2012september-a-designers-guide-to-lithium-fig5.jpg
(Note, the C in this graph for those interested is the battery capacity in Amp-hours divided by 1 hour.)
There’s four phases in a lithium ion battery charge cycle, especially from flat. The first phase is a conditioning/“bootstrap” phase where you provide a low constant-current charge to get the battery up to 3V, then the constant-current phase in which you apply no more than C (the 1-hour battery capacity) in current to charge the battery. At 4.1V there’s a transition to a constant-voltage charge in which the charging is completed by providing a constant voltage and the current will taper off, and when the current hits a cutoff value the charging is stopped and then occasionally topped up by a few more constant voltage pulses to make sure it’s kept full. This final phase doesn’t strictly need to happen, one can just turn off all charging once it’s done, since Li-Ion and LiPo batteries have good shelf lives.
The asshole charger is a current follower to try and mimic phase 2 and 3 of a proper charge and typically never disconnecting the battery. Basically they have a very simple constant-current circuit that has a max voltage of 4.1V, so it’ll provide a constant current until the voltage hits that 4.1V mark. The circuit at this point is using it’s provided voltage (the voltage coming out of whatever crap regulator they slapped together for the charger) and that provides the constant voltage portion of the charge. It would be a nice, simple circuit if that was all that needed to be done, but the other phases are forgotten. So this will keep charging past the 0.1C point, slowly increasing the charge on the batteries and wearing them down, and if left too long they’ll overcharge and burst (sometimes into flames). Not something I want to have in my house, that’s for sure.
Side note on how this all relates and constant-current sources work:
One of many things you can never get away from in electronics is Ohm’s law; it’s probably the most important thing ever. I = V/Z, the current through an element is equal to the voltage across the element divided by that elements electrical impedance (Z should be preferred over R because of multiple variable conditions of an element). This is important to know, that as you increase voltage on a load the current will increase, if you increase the resistance either the current or the voltage will go down, depending on what’s held constant.
It’s easy, for the most part, to hold voltage constant (at least mostly constant). This is what most power supplies do, it will convert a range of voltages to one specific voltage and keep that voltage constant. From there a load draws whatever current it needs; it can be a dynamic load or static, having either a constant impedance or a varying one with time or other variables. On the other hand, you can also have a power supply that will hold the current constant. This is done by varying the voltage to match the target current. So, as the load’s dynamic impedance changes the voltage will change in response to that impedance change to keep the current at the same value. An ideal constant current supply has one interesting and impossible side effect, if you have an open circuit you’ll have an infinite voltage; not a thing that can happen. But the real-world corrects for this singularity by the fact that every constant-current power supply will, in effect, have a maximum voltage (in most manufactured ones, that’s the supply voltage). So, in reality, a constant current supply is one where it’s constant current to a certain maximum impedance, then it starts to lose regulation (while most constant voltages are constant voltage to a minimum impedance, then they start to lose regulation).
10 Likes
Hoverboard and/or private dragon.
1 Like
smells like hoax to me
Had a powerbook battery do the puffy thing…like it just rather swelled up and came part. Didn’t catch fire tho…but quite a few of them did.
1 Like
A device which caught fire under load would be more believable.
No, this is constant with an H-bridge failure. In any 2-way motor device you need a way to switch the polarity of the motor, and the standard way to do that is an H-bridge. An H-bridge is also something that can go boom very easily if made or controlled improperly, and this is because on these cheap asshole-made hoverboards the H-bridge is directly across the battery with no safety devices (hey, those cost money and the generic makers can’t be bothered with anything that’s “optional” that costs money!).
So there are two ways the H-bridge can fail, leading to a direct short of the battery and that’s by control failure or device failure. An H-bridge is called that because it’s in the shape of an H, H-bridge - Wikipedia it’s 4 switches that control the direction of the motor. Since the H-bridge is connected directly to the power supply (on good devices also through safety mechanisms, but ain’t nobody got time for that these days!) on the top and bottom. Besides the power each of the switches (in modern H-bridges they are usually MOSFETS) has a control signal to turn on or off. If you turn s1 and s3 on the motor will go one way and s2 and s4 it’ll go on the other way. If there’s a failure in software or other hardware that causes s1 and s2 or s3 and s4 at the same time, you get a direct short of the battery which will cause wires to heat up and since it’s on a high-discharge LiIon it’ll start whatever’s nearby on fire (and not the cool flame decal type, real hot fire).
The other failure mode is where you have a silicon device actually fail. The problem with good design with power FETs, IGBTs, BJTs and other power silicon devices is that when they fail they fail closed. This means you have something that is on when it’s in a failure state; not failsafe. So, when that fails it’s another dead short condition (and again, with no safety devices so stand in our way of getting awesome flame effects!)
To me it looks like it’s more of the first failure mode I described, I think there was either a manufacturing short or other defect that caused the H-bridge to have both parts of the leg on on the right side of the device (it very well could have been that upstream components failed in the second fail-on way, signaling the H-bridge to be on-short). This wouldn’t have presented itself until the device was switched onto the drive mode, and the fact that none of the lights worked and the other motor didn’t work is also consistent with this; the short would have lowered the available voltage to the rest of the device lower than the on threshold.
He still could be doing this for views and taking into account the current sub-craze of “Hey, look at all them hoverboards exploding”, but I doubt it. Most hoaxers go for the “movie approach” where they try to play it up as much as possible, and go for the improbable or impossible failure modes that people have in their heads, not ones that are easily picked apart by engineers 
7 Likes
Same here, note how far back he is after he gets off and inspects it. He also doesn’t seem to really be trying to move it forward so much as twist it, and when he’s doing the one foot movement right before it smokes he again doesn’t seem to be attempting motion per se, (and the foot is opposite the point of combustion).
My guess is he took it apart, saw it was badly made, tweaked it a little to force the issue, et voila’, hover-brulee’!
2 Likes
eeeeeewwww thunderfoot
4 Likes
No, the battery luckily didn’t ignite here, I suspect capacitors are what actually caused the flames with the H-bridge shorting to generate the heat. Probably also some added low flashpoint plastics and some of the silicon devices. The smoke is consistent with silicon and packaging burning, along with the flareup caused by a shit electrolytic cap going.
3 Likes
I’m so glad here in the US we have a governmental Consumer Product Safety Commission dedicated to keeping us safe from things like this device, which I assume must be a dangerous UK- or EU-only product that wouldn’t be allowed near the US thanks to our strong regulations.
1 Like
photoshopped!!
2 Likes
You forgot the sarcasm marks.
This is the reason airlines won’t allow them on board.
Over my vacation, I saw two people riding these through stores in the goddamn mall. Why couldn’t THOSE have burst into flame?
5 Likes
Don’t some Li batteries release HF gas when burning? Don’t breathe that smoke!
1 Like
That’s actually exactly what the FAA are advising cabin crew to do with Li-ion batteries - water, cola, etc. There’s really not enough lithium in them for it to pose a bigger risk.
Lithium-metal batteries are a different matter, but you won’t find them in so many gadgets because they’re not rechargeable. The batteries you find in phones, laptops, etc - feel free! Now that most of us drag at least one of these batteries around with us all day every day, I think more people should be aware of this.
(Personally I would have let this one burn out because it’s outdoors, it’s a non-flammable surface, etc. You can just go inside, away from the fumes, and make a cup of tea. But if you have a rechargeable battery fire indoors, put it out, and feel free to use the tea kettle if it’s all you have to hand!)
3 Likes
I thought I was supposed to make piddles on it, or something.
Do you have a Nest thermostat? Google does own both YouTube and Nest…
(If you do have a Nest I recommend staying away from skiing/snowboarding videos until summer)
1 Like
Looks to me like a sudden release of a fair amount of flammable vapors, which then gets ignited.
Would be consistent with outgassing of propylene carbonate (which the flame color also suggested, together with the rapid die-off of the flame), which is a common lithium battery electrolyte.
Not in this kind of battery. Worst case you get some minor amount of lithium hydroxide.
The propylene carbonate electrolyte, on the other hand, is very water-soluble. So you get rid of the flammable volatile part pretty fast.
In this case, the short was already there. Once the battery gets through the “rapid spontaneous disassembly” process, all bets are off. Enough water can overwhelm the heat generation capability of what remained, and sink the heat effectively.
Pure, non-salty, water also won’t do much of a short, its conductivity is rather poor.
Salt water, however, may generate chlorine by electrolysis. This is the bane (one of many) of submarine operators; many succumbed to this effect. But I wouldn’t consider it a major concern for small batteries (this one counts as small in this context) in open area.
A dodgy cell can be rendered safe by storing it overnight in slightly salty water. Any thermal runaway will be nipped in the bud with the water’s heatsinking capability, and the cell will discharge. You can then cut it open and salvage the thin copper foil and the microporous polyolefin separator foil to have fun with later.
5 Likes
Ah…Lithium. The angry mood stabilizer.
6 Likes
