Headway battery pack FIRE! <lot pics>

[Jeremy Harris]

sounds like the charging mosfet overheated and caught the pcb on fire.

i am sure they will use this as proof that lifepo4 catches on fire when overcharged like the bogus story about gregg's garage catching fire because of lifepo4 overcharging.

Believe me, Greg's story wasn't in any way bogus - I do wish you'd stop spouting crap on here, especially when you're technically wrong more often than not.

Greg's car DID burn out whilst on charge, but as I've said here before, and as Greg himself said after trying hard to find it, the cause remains unknown. All we do know for sure is that when a fire starts LiFePO4 cells will burn fiercely just like pretty much any other battery.

Any bike/car battery has enough stored energy to initiate a fire pretty easily if a fault develops, irrespective of chemistry.

 

[Jeremy Harris]

My money is on there being a short between two cell cans, probably for the reason Ed Lyen has given, the metal connecting straps being slightly too tight and allowing two cell cans to be in contact.

I hate the plastic film on cells like this, as it isn't very abrasion resistant and it causes a cascade failure when there is a problem. If a small short developed between two cell cases, caused by a bit of wear from vibration, then the resulting heat would cause the insulating film to shrink and expose a greater area to short, which in turn causes more heat and starts to shrink back the film on adjacent cells, creating a cascade failure.

I had exactly the same thing happen to me with some film insulated NiMh cells in my very first ebike back. A single cell cooked (my fault - I over charged it) and that then damaged the insulation on several surrounding cells, creating a similar fire to this one.

Perhaps one thing to learn from this is that even an outwardly very well built battery pack has the potential to catch fire, no matter what the chemistry. Personally I prefer cells that have cardboard sleeves (like the older Headway cells) as these are less likely to fail in this way.

 

[circuit]

I usually use some heat resistant two sided tape, to stick these together. The tape is ~1mm thick, giving it some room. But still, after seeing this, will do some things a bit differently.

 

[Punx0r]

Personally I prefer cells that have cardboard sleeves (like the older Headway cells) as these are less likely to fail in this way.

Interesting point. I'd also assumed card covered cells to be of cheap quality, but it seems that's an incorrect assumption. Judging a cell by it's cover

 

[dogman dan]

I'm still pretty curious about how the battery was mounted too. Tossed in a bag on the rear deck? Doesn't take much movment of a heavy object to start things chafing, leading to the short that likely caused the fire. Could have been the cells that chafed, could have been wires to the bms, whatever.

Like the duct tape pack fire I saw posted on the V years and years ago, once you get em going with a good short, it still burns good regardless of the chemistry.

Maybe a loose mounting method wasn't the real cause, but it will be causing other packs to burn.

 

[Alan B]

Great photos Edward.

So I see a main fuse on the plus lead. Looks like the Signalab is switching the negative main lead. I don't see any fusing between the battery and the BMS.

I also don't see any fusing on the charging lead which probably also goes through the Signalab.

The balancing leads are fairly small gauge. One question is whether they are small enough gauge to act as fuses. But the charging leads and the negative main lead are heavy enough to carry enough current to a short to heat up the Headways.

The great looking Headway plastic supports do hold the batteries apart so they don't chafe, but at the first sign of heat they fail, as does the plastic coating on the cells, and then the cells develop local shorts and starts a chain reaction of heating, melting, and shorting.

Perhaps a couple more fuses here would have helped to separate the pack from a failing BMS? Perhaps the one fuse that is there is not located optimally? It is close to the battery which is excellent, but it is not between the battery and BMS?

 

[kfong]

Sorry to hear about the fire chroot. Sounds like a short since no charging was taking place at the time. Seems like the space between batteries could of used some extra insulation or even injections of silicone caulk to secure things.

I've gone as far as wrapping my lipos with thicker tape, separating them with a wide double stick layer. Injecting silicone caulk between any spaces. Packs are then secured to a custom tray that is hard mounted to the bike with kevlar straps. Then the whole pack is further wrapped in duct tape to secure any loose wiring with separarating layers. On top of all this is a carbon fiber shell that should be able to contain the fire. I trail ride all the time with plenty of crashes. It's held up well for me. Shorts are something you can't afford to let happen bettween adjacent packs due to mechanical stresses. Something that seems easily overlooked by many builds. From what I've read most BMS failures tend to drain out batteries. So my bet is on a mechanical short.

I don't think a fuse could help in this situation unless you have fuses between every cell. You can't easily predict where mechanical shorts can happen.

 

[circuit]

Alan,
and why do you assume that BMS failed? There are no signs of failure on it. The PCB is intact, all components in place. Ant the plexiglas sheets are in good shape around the power switches.

 

[Alan B]

Alan,
and why do you assume that BMS failed? There are no signs of failure on it. The PCB is intact, all components in place. Ant the plexiglas sheets are in good shape around the power switches.

Not assuming. Just observing. Looks like a lot of heat from the BMS balancing circuits area. Looks like local heating?

If it was a wiring short why did the main fuse not stop it? Is the main fuse blown? Is the main fuse properly voltage rated?

Whether it is the BMS or a wiring short the fusing was not adequate to isolate the battery from the fault.

 

[circuit]

Alan,
and why do you assume that BMS failed? There are no signs of failure on it. The PCB is intact, all components in place. Ant the plexiglas sheets are in good shape around the power switches.

Not assuming. Just observing. Looks like a lot of heat from the BMS balancing circuits area. Looks like local heating?

That side of BMS is burnt, I agree. However, this could be simply from the pack burning. Also, since the balancing wires are non-heat-tolerant, their insulation could have simply melted due to external heat source and then short together. The circuit board in that corner looks burnt from outside, but I bet its all good under the ash.

If it was a wiring short why did the main fuse not stop it?

Because main fuse is intended only to protect the internal wiring from external short. This is clearly not the case.

Is the main fuse properly voltage rated?

What? You mean current, I guess?

 

[Jeremy Harris]

Even with those plastic end caps the cell cases are pretty darn close together. If the link plates are a little bit tight (in as much as the holes are slightly too close together, as some are known to be) then it's pretty easy to pull two cells close enough together to chafe, notwithstanding the plastic end caps.

There's no apparent BMS failure here, looking at the photos, and no amount of fusing can protect from a case-to-case short.

 

[Alan B]

The Headway plastic support blocks I have won't allow cells to touch, at least until they melt. The plastic of the two blocks is physically between each cell pair at the ends, so unless that plastic is compressed to nothing (or melted) or the cell flexes they won't touch. So I'm not expecting the first fault to be in the battery itself, but somewhere else.

But when the cells get hot they can short and no stopping that. So I expect the cell supports were victims and the first fault is elsewhere. But that remains to be determined.

But can we discern the first fault?

One question for chroot, is the fuse blown? Did it try to blow and re-short? Fuses have voltage ratings, and using a fuse beyond its voltage ratings can cause problems, like a failure to remain open during a fault. This is not likely. More likely is the housing can melt and re-short during the current surge. Fuse housings have a maximum current interruption capabilty and the surge currents on these lithium batteries are pretty high, so the time it takes for the fuse to blow can allow enough heat in a low cost fuse housing to destroy it. But it doesn't look that distorted so that is unlikely.

What was the fuse rating? Is it appropriate for the wire gauge?

Looks like the main power was through a powerpole. So main current wiring is 12 or 10 gauge most likely.

Is the damage on the frame indicative of a first-fault wiring short?

I see about three scenarios here, related in no particular order:

1) first fault was in the battery. game over. no help from fusing.

2) first fault was in the BMS. then battery overheats and local shorts. fuses could help but are not properly placed between battery and BMS.

3) first fault was in wiring. then battery overheats... fuses could help, why did they not?

 

[Jeremy Harris]

The Headway plastic support blocks I have won't allow cells to touch, at least until they melt. The plastic of the two blocks is physically between each cell pair at the ends, so unless that plastic is compressed to nothing (or melted) or the cell flexes they won't touch. So I'm not expecting the first fault to be in the battery itself, but somewhere else.

But that's not taking account of the known problem. There were some connecting strips around that had the holes punched too close together, as Ed Lyen and I have both mentioned already. What then happens is that the constant pressure and stress causes the plastic end caps to crack right at the squeeze point and allow the cell case insulation to touch. It's not a new issue, if you look back here to the early days of these plastic blocks you'll find that there were a few issues with this.

It's not obvious, looking at the pack, that the blocks have split at this very narrow region between the cells, either. I believe that Headway have fixed the issue now by making the blocks a bit thicker here and getting the hole spacing right on the connecting strips. My guess is that this pack may have had the older style blocks and strips whilst yours has the newer ones.

 

[torqueon]

Is it possible that the cells have swelled up over time, thus making contact

 

[dnmun]

the positive charger wire appears to have melted all of the insulation off all the way back behind that first wire tie. it shows signs of molten metal attached to the strands where it is pulled up above the BMS in that picture, and the strands have burned open and some are twisted out away from the wire itself.

is there some way to guess about where that wire was located before this incident started? it seems to have some arch, or curve similar to the tubing on the frame, so i wondered if the charging wire came out of the bag right above the frame there where the metal spittle is on the frame. there is no fuse in that line back to the connection to the B+ lead where it comes off the battery, below the main fuse.

 

[Alan B]

Good observation.

The positive charger wire is a good candidate.

If it was this wire to the frame that would imply the frame was grounded to the pack. Where might the system have been grounded? There might be evidence of high currents in other parts of the grounding path as well.

 

[dnmun]

it may be is shorted to the pack through the frame if the negative battery lead was also shorted to the frame, or it could be that one of the screw heads on one of the cells pinched the charger lead so it shorted out part way through the pack and was not full voltage. or the screw contacted the frame in one place and the charging lead was finally worn through at another place and the current flowed through a short section of the frame.

i am surprised to see both the positive and negative cables wire tied together and actually see the main battery cable runs right through the BMS. not sure why they had to do that except to have the two cables adjacent each other on the output.

of course it coulda shorted later in the fire. if it shorted between casings then you will see the short burned a hole in the casing on the place between adjacent cells where it is supposed to be too tight and you can measure the weight of the cells to see how much electrolyte was boiled out by comparing the current weight to the stock 460 gms.

 

[eva-michael]

Sorry to hear this issue.
No matter what's the cause. The biggest problem is the holder which seems to be plastic one and not the ABS flame retardant one. The cell will not be fuel of this fire as we can see from photos and the major fuel is that plastic holder and some plastic coated wrie.
I also wonder if there is a bag which take the pack directly without isolation of fireproofing. It's a danger with bare battery pack inside.

Actually my first though when I saw the picture is charge wire over heat-> plastic cooted wire get burn as it's too hot->burn all which can be burn in this pack. But Chroot has present it was no charging.
I never hear balance wire which are too heat and then burn. The balance function(max balance current mA) is under the max capability of which the balance wire can load. We should take care and not cut any flexible cords inside the the balance wire.
Also make sure the connection between wire and bms are firm and efficent enough. That's also a critical point.

----Safety Protection is always a very important thing.

 

[deVries]

@deVries - The battery pack protected by the Topeak MTX TrunkBag DXP and it was destroyed.

This a soft nylon bag that could be "bumped" & "pressed" into. It is not a hard shell outer case, so bumping & pressing into the soft battery bag could do damage to the battery inside the bag or apply slanting or angular forces against the battery's exposed skeleton structure.

Children can play, bump into, or knock over things too.

Did you store anything else inside the bag with battery?

Seems like using fuses between such big Ah cells & fuse bms from battery too would not be too expensive? At least it would isolate shorts to bms failure or one cell failure *IF* designed correctly. Also, better cell isolation protection between cells is needed if one shorts, so it remains isolated from "melting into" adjacent cells creating the "cascade effect".

 

[Alan B]

My Topeak trunk bags are pretty substantial, somewhere between hard and soft cases. Definitely not just a soft case.

I think proper fusing is not really done frequently on these battery packs. I would like to see fuses on each hot lead coming out of the battery, but that is a lot of fuses. The balance leads might not need fuses since a short would destroy the thin wire without much danger. But the charging leads and main leads are good candidates for fusing.

 

[deVries]

I think proper fusing is not really done frequently on these battery packs. I would like to see fuses on each hot lead coming out of the battery, but that is a lot of fuses.

This pack only needed 16 fuses to protect each cell from the other. Cheap insurance considering the alternatives. 8) Too bad the assembly design itself doesn't have the fuse built-in as "thin width" power connections, and then support the cells structurally by not using the power connections w/fuse design as support.

 

[dak664]

A single fuse at the mid point gives the least bang for the buck.

 

[bigmoose]

A single fuse at the mid point gives the least bang for the buck.

Are you sure statistically that statement is correct? If you do a formal FMEA and place the fuse in either the main + or main - it only protects against shorts downstream or outside of the battery pack. If one sketches out all the other potential cell to cell shorts within the battery there are other places with a higher probability of the fuse blowing from valid, potential shorts I believe...

 

[chroot]

@Lyen, Thank you for the pictures. That one I bought it from ES member.

@deVries - There is none object or anything in the pannier bag except just only battery pack *isolated* inside. It was been since never change or anything. It never give me any trouble during the charging or ride rough on road daily.

@Alan B - The fuse rated is... hard to read due burnt. I tried look inside the fuse apparently did blow up. *Hard to see inside but can see the black marked from blow up*

Here the picture of fuse.

 

[dak664]

I have a 3 9 amp polyfuses in parallel at the midpoint of two physically side by side 6 cell strings on a 36 volt headway pack in a 2x6 configuration. A short from any cell on one string to any cell on the other string activates the fuse, and that has happened twice so far with just a mild spark. It does not protect against shorts within each half string, but if dropping something metallic has decreasing probability of shorting as distance increases, then since each pair of cells in the 18 volt strings is strapped on one side or the other anyway I figure there is a 4 in 5 chance a that a random short of adjacent cells will go through the fuse. But with only two accidental shorts there is not enough data to support that conclusion.

By "least bang for the buck" I meant the most safety for the money spent on the fuse