Why do battery fires happen?

[Mono]

A talk by the owner of Grin technology who have also designed their own modular battery pack. Time stamped at 20:43 where he starts to talk about the modern era of 18650 Li-Ion batteries:

The main takeaway: a modern nondefective battery cell is apparently quite hard to set on fire by abuse or by a failing BMS. Most fires as of today, he suspects, are caused by cheap cells and/or unlucky manufacturing issues which are much less likely to happen with reputable brands.

I am not happy that my newish 584Wh V8F has now Lishen battery cells https://ecodrift.ru/2023/05/29/inmotion-v8f-razbiraem/

Edited March 18 by Mono

[xiiijojjo]

Thanks for posting, that was a good watch. 

This has me wondering if 18650 are a lot safer than 21700 cells, but I wouldn't know.

So far for this forum, as far as I can remember most fires were the result of the M50T cell, physical damage to the cell from crashing or custom battery packs. 

Considering their QC track record, do y'all think Gotway is the type of company that would potentially use counterfeit cells? I'm not claiming they or any other EUC company are, but it sure would make for some juicy headlines on this forum if someone ever found that to be the case... I know Ewheels have done testing to a lot of stuff they buy so they probably know better than anyone.
 

I wonder what the fire rates on the V13 will be a few years from now. I was always intrigued by how they built their battery. Sorry this is the best picture I could find.



It just came across as safer because of how each cell is fixated separately, causing no friction between each cell in case of a violent crash, I believe each cell is incased in a foam/silicone-like material for further shock proving.
It's really no wonder so many Gotways kick the bucket and erupt into a great fire in case of a crash, considering how they make their packs, with all the cells fixated right up against each other, as with traditional e-bike batteries. 

No matter the future of Inmotion, you gotta give it 'em: That idea with separating and encasing each cell makes for a much safer battery in case of a crash, at least in theory. 

Edited March 18 by xiiijojjo

[alcatraz]

Heat, corrosion, overcharging/discharging, lack of cell separation and missing spacers between positive and nickel.

[BKW]

I have an immense amount of respect for battery/electrical engineers. It terrifies me. You really have to now your stuff I feel like or it's a dangerous game

[Planemo]

17 hours ago, xiiijojjo said:

It's really no wonder so many Gotways kick the bucket and erupt into a great fire in case of a crash, considering how they make their packs, with all the cells fixated right up against each other

Gotways had plastic cell separators as far back as the MSX, maybe even earlier.

Granted, the current Inmotion packs do look well done.

Apart from the LG debacle (which even then isnt the cells fault IMO), its not the cells themselves that have ever really been a problem. Poor construction/badly welded interconnects/shitty bms' and water ingress are the issues.

I've deliberately treated some 18650's incredibly badly for testing purposes, even repeatedly charged one that had gone down to 0.3v. Believe it or not that cell is still working fine 4 years later, (and yes its always isolated in case it does decide to grenade itself!)

[Mono]

4 hours ago, Planemo said:

Apart from the LG debacle (which even then isnt the cells fault IMO), its not the cells themselves that have ever really been a problem. Poor construction/badly welded interconnects/shitty bms' and water ingress are the issues.

I've deliberately treated some 18650's incredibly badly for testing purposes, even repeatedly charged one that had gone down to 0.3v. Believe it or not that cell is still working fine 4 years later, (and yes its always isolated in case it does decide to grenade itself!)

Don't these two paragraphs directly contradict each other?

How do you actually know that "Poor construction/badly welded interconnects/shitty bms' and water ingress are the issues"?

Wouldn't you agree that a single quality 18650 cell should and does virtually never start burning with the single exception of physical destruction? Does it burn when you put it under water? Does it burn when it is shorted? Does it burn when it is overcharged? I don't think so, am I wrong?

Edited March 19 by Mono

[alcatraz]

A single cell might not fail spectacularly, but in close proximity and connected to other cells, new failure modes are introduced.

I personally don't believe that the infamous LG M50T were bad cells necessarily. I think they were simply wrong for this application. The genius that decided to design a pack with four 7A cells to succeed packs with six 10A cells (NCR18650GA) and then draw more power, that guy is an insane engineer. 

Gotway did it first, but then Inmotion and Kingsong fell in the trap too. A batch of the Kingsong S18 had the craziest configuration of them all: three 7A cells (LG M50T). That's 21A power draw budget at room remperature.  

[alcatraz]

It possibly displays one problem with people and it's brand loyalty. Just because it says LG, Sanyo, Samsung or whatever on a cell, you can't just use it for any application.

A noname cell used properly should outlast an LG used improperly etc. 

Edited March 20 by alcatraz

[Planemo]

7 hours ago, Mono said:

Don't these two paragraphs directly contradict each other?

No?

I was supporting the fact that virtually all cells themselves are very tough and durable, even the LG 50T. The problems with the LG debacle was drawing too much current out of them combined with shitty interconnects - I've seen several stripped packs which show failure points.

Not sure what point you are debating.

[Cam]

I would not trust shoddy cell brands either (they literally lie about capacity) but the concern that Grin is raising does not seem to concern us: no EUC manufacturer uses no-name / dodgy brand cells.

Regarding EUC battery fires: AFAIK thermal runaway in a li-ion battery pack can happen from (but not exclusively):

- Overcharge: this can happen when the pack gets misbalanced: a weak cell becomes a short and the charge voltage is spread over the remaining cells.

- Damaged conducting material heating up from current: this can happen from corrosion or cracks due to poor construction or violent shocks.

Sidenote:

- According to the "fire history" thread, the EUC manufacturer with most fires is the one famous for poor build quality and lacking weather protection.

Edited March 21 by Cam

[Mono]

11 hours ago, Planemo said:

Not sure what point you are debating.

I am not debating anything. I am trying to figure out conceptions that most likely agree with the real world. I was asking, among others, the following questions.

Does a single quality 18650 cell burn when you put it under water? Does it burn when it is shorted? Does it burn when it is overcharged? To all I can see, the answer is "very very rarely" to all of these questions. What is the evidence we have to the contrary?

7 hours ago, Cam said:

Regarding EUC battery fires: AFAIK thermal runaway in a li-ion battery pack can happen from (but not exclusively):

- Overcharge: this can happen when the pack gets misbalanced: a weak cell becomes a short and the charge voltage is spread over the remaining cells.
- Damaged conducting material heating up from current: this can happen from corrosion or cracks due to poor construction or violent shocks.

How do you actually know that these are a relevantly common events? I am asking, because it seems that a pretty smart engineer running a company for a decade and counting which designed and builds their own battery pack did not come to the same conclusion.

17 hours ago, alcatraz said:

A single cell might not fail spectacularly, but in close proximity and connected to other cells, new failure modes are introduced.

How so, assuming a single cell doesn't start burning on its own under any discharge/charge/depletion situation? Obviously we can have an electrical fire e.g. of wires or the control board which could then set the batteries on fires. However we should not call this a battery fire, like we don't want to call a battery fire that sets the tire on fire a tire fire.

7 hours ago, Cam said:

no EUC manufacturer uses no-name / dodgy brand cells

Implying that you do know Lishen? Could you tell me more about their reputation?

Edited March 20 by Mono

[Planemo]

15 hours ago, Mono said:

I am not debating anything.

? I'm even more lost now. You're clearly debating/discussing/questioning something but I'm not sure what you are trying to prove/disprove.

15 hours ago, Mono said:

Does a single quality 18650 cell burn when you put it under water? Does it burn when it is shorted? Does it burn when it is overcharged? To all I can see, the answer is "very very rarely" to all of these questions. What is the evidence we have to the contrary?

Single cells are indeed very robust IMO as I have already said. I'm not about to scour the internet for videos on submersion/shorting/overcharging to prove/disprove my thoughts.

All I will say is that a single cell in isolation is a far cry from what we have in our wheels. Which means that control systems and construction of the packs become paramount, not to mention making sure the pack output is sufficient for the job in hand (LG50T scenario). Crucially, shorting a single cell is certainly very different to shorting 24 of them in series at 4P.

I think you need to step away from what I think you're trying to say in that a single cell is robust (which I agree with) and look more at the application we are using them in. I think I get where you're coming from in that if one cell can't flare up then why should a larger quantity do so. But ask yourself...why did the Begode LG50T fires happen? The batteries were clearly going into meltdown...are you suggesting that this stemmed from a wire/some other ignition source and not the cells themselves?

 

[alcatraz]

17 hours ago, Mono said:

How so, assuming a single cell doesn't start burning on its own under any discharge/charge/depletion situation? Obviously we can have an electrical fire e.g. of wires or the control board which could then set the batteries on fires. However we should not call this a battery fire, like we don't want to call a battery fire that sets the tire on fire a tire fire.

The heatshrink gets soft and you short out the cell casing against nickel or another cell.

Then there's the problem of mass. If you impact a battery pack it'll damage cell casing sooner than if you impact a single cell.

Then there's balancing requirement causing lots of failure modes. That doesn't exist with single cell.

Then there's the problem of cells internally shorting out for whatever reason. When it happens for a single cell the charge of one cell gets rapidly discharged into heat. When it's connected to other cells in parallel you dump all their charge into the shorted cell causing a more spectacular failure. Much more heat that can start a chain reaction.

There are more reasons why packs are more dangerous but these are the ones I can think of at the time. It's dinner time. 

[alcatraz]

Heat. If you had cells spread out more they wouldn't concentrate heat in one area so they could tolerate more amps before failing.

[sbb]

On 3/19/2024 at 6:24 PM, Mono said:

How do you actually know that "Poor construction/badly welded interconnects/shitty bms' and water ingress are the issues"?

I heard it from some guy named "Jason". There's a rumor that he knows a thing or two about it.

 

22 hours ago, Mono said:

Does a single quality 18650 cell burn when you put it under water? Does it burn when it is shorted? Does it burn when it is overcharged? To all I can see, the answer is "very very rarely" to all of these questions. What is the evidence we have to the contrary?

Single cells aren't the issue with water.

The cross connects are the issue with water. The current drawn by these motors travels along those little metal strips on the surface as a skin effect.

Add some non-conductive corrosion or a non-conductive hard water spot to one of those strips, and the current flows around that spot - increasing current density in the remaining skin at that location. That's the issue. Lots of batteries? Lots of current, and a bigger problem when the surface area is reduced. It's no different than having a cracked spot-weld of one of those tabs to the battery; the density over the remaining weld increases.

 

Edited March 21 by sbb

[Cam]

On 3/20/2024 at 5:30 PM, Mono said:

How do you actually know that these are a relevantly common events? I am asking, because it seems that a pretty smart engineer running a company for a decade and counting which designed and builds their own battery pack did not come to the same conclusion

I can't understand why you are so hellbent on "challenging" established understanding about li-ion battery packs based on a video that does not do so, and mostly does not concern us. I'm not going to take an hour of my time to explain and source to you how a battery pack is build, and why single cells do not have the same challenges as packs of 60 or more with the voltages and currents that involve. Example: A Sherman OG twin pack has 24x higher voltage, 5x higher current giving 120x more Watts than a single cell.

Even the guy from Grin, at 23:25, mentions "construction improving with mechanical holders for the cells" (this to mitigate risk of rupture of connection between cells from mechanical stress), "Better enclosures so there is less water ingress" (water will corrode cell interconnections (and BMS)).

These are the exact two points I made about the conductive strips, mentioned by Grin, in your video. He does not dwell on them because he wants to talk about bad quality cells.

If you do not believe or understand my explanations on how thermal runaway can happen in a li-ion battery pack, how do you explain EUC fires? Because almost no EUCs use the crappy brand cells Grin talks about, and but they still burn.

 

On 3/20/2024 at 5:30 PM, Mono said:

Implying that you do know Lishen? Could you tell me more about their reputation?

Their presence in a few IM EUCs is statistically insignificant compared to all the Panasonic Sanyo / LG / Samsung that are in virtually all other EUCs, and who still catches fire. Maybe they are yet more risky than quality brands, that do not change the fact that quality cells end up burning just the same, due to the reasons several people have been describing, and that Grin mentioned as concerns being worked on.

Really, I do not understand what you are getting at. Some cells possibly being worse than others do not exclude other fire risks.

Maybe we misunderstand your point, and you could help us understand by reformulating it.. Currently it comes across as "Battery fires happen because of bad quality cells, and not the reasons commonly accepted". Had it been "Battery fires can happen because of bad quality cells." that would have been perfectly fine, even though not relating to most of us.

Edited March 22 by Cam
clarity

[Mono]

On 3/21/2024 at 8:58 AM, Planemo said:

I think you need to step away from what I think you're trying to say in that a single cell is robust (which I agree with) and look more at the application we are using them in.

Where does the fire start then? If it is in the "robust" single cell, what is the mechanism that ignites the cell?

On 3/21/2024 at 11:04 AM, alcatraz said:

The heatshrink gets soft and you short out the cell casing against nickel or another cell.

Then there's the problem of mass. If you impact a battery pack it'll damage cell casing sooner than if you impact a single cell.

Then there's balancing requirement causing lots of failure modes. That doesn't exist with single cell.

Apparently, undervoltage or overvoltage does not ignite a single cell. They also seem to be protected against overheat. So, where does the fire start?

[Mono]

23 hours ago, sbb said:

Add some non-conductive corrosion or a non-conductive hard water spot to one of those strips, and the current flows around that spot - increasing current density in the remaining skin at that location. That's the issue.

The question remains: where does the fire start? You seem to be saying it does not start within any of the cells?

[Rawnei]

1 hour ago, Mono said:

The question remains: where does the fire start? You seem to be saying it does not start within any of the cells?

Of course the fire starts inside the cell but the root cause is not the inside of the cell.

[Funky]

22 minutes ago, Rawnei said:

Of course the fire starts inside the cell but the root cause is not the inside of the cell.

One could say - it's the outside and outside elements.  

If only cells (batter packs) would be protected better..

[alcatraz]

7 hours ago, Mono said:

Apparently, undervoltage or overvoltage does not ignite a single cell. They also seem to be protected against overheat. So, where does the fire start?

As the cells age their internal resistance increases. With it comes more heat for the same current. 

Check out some diy powerwall videos on youtube. There some serious charge is moved across really crappy noname cells. No matching, just mixing all kinds of cells into packs. Reusing of cells that have been overdischarged. Basically junk that can hold a charge. They prepare for the cells to fail in operation by using fuse wire.

Removing shorted cells and rebuilding packs is a regular thing for them. The packs remain in operation but the shorted cells automatically disconnect.

It seems that given enough time all cells fail this way. 

You can even ask in the comments about failures etc.

Unfortunately no euc packs use fusewire like these powerwall guys. It means the pack can't be ridden until the cells fail.

[Promise]

Are C02 extinguishers enough to stop an euc fire?  I always monitor my wheel while it charges, so I can haul it outside as quickly as possible if I need too, but it would be amazing if all I need to do is hose it down with an extinguisher until the temperature drops.

[The Brahan Seer]

6 hours ago, Promise said:

Are C02 extinguishers enough to stop an euc fire?

Sadly No. Its great you monitor your wheel while it charges. Previous history shows 30% of fire occurred during charging but this may have changed with better batteries, BMS and chargers.

Have a read of this thread to learn more...

 

[Cam]

Here is a simplified illustration showing how a damaged conductive strip can lead to heat and thermal runaway of the cells.
It is the same principle where a bad contact can lead to house fires, except in this case it heats up the neighboring li-ion cells.

This does not exclude other ways / reasons for battery packs to catch fire.

Edited March 23 by Cam

[Panen]

I used to ride the Begode RS with LG50LT (not LG50T) batteries, and initially it always charged to 100.8volts.
After around 6 months at max it started to only charge to 100.5 and then 100.2 and 99.8 volts after that.
I used to keep it in the charger for balancing at least once every 2 weeks.
I wonder why it deteriorated that quickly. Maybe the cells are low quality or the cells were recycled or something?

Now i ride the Sherman S with Samsung 50E cells and even after a whole year of riding it still charges to the exact same voltage as when it was factory new.

Edited March 25 by Panen