AllCell Battery Fire

I've been using an AllCell 36V battery for the last 3 weeks and today it caught fire !! Towards the end of my ride.

Using the CA3, 9C 1000W motor. Previously the BMS on the AllCell functioned properly and after discharging to about 33V would cut off power.

Towards the end of this ride, it cut off as usual. I was braking going downhill and sending some juice back into the battery via regenerative braking. Started to smell something, turned around to see smoke coming out of my bike bag where it was sitting next to my 35A controller from Grin.

See picture, fire occurred around top of pack around the BMS.

This isn't supposed to happen with an AllCell ! Anyone heard of this happening before...advice ?

Thanks

The rule is not to run 2 lvc cutoff but leave something in there ( the battery) and put it back in the charger. Running to lvc will unbalance the pack.
Do you drive your car to the gas tank is empty sucking all the water and dirt in the fuel injection system to have it Clog ? Not a good rule.

Talk to Allcell ?..im sure they would like to know of the problem and help diagnose the cause at least.
Hard to tell from one photo and your description, but initial thought is it sounds like a wireing fault or BMS defect.

You're lucky it didn't burn more. There's enough of it left to maybe actually find out exactly what caused the fire.

If the BMS tripped on low voltage, then you pound it with high regen from the controller, all the regen current has to pass through the FETs, which are turned off. In the regen direction, the FETs behave like a diode and pass the current. Diodes heat up a lot more than a FET that's turned on.

You want to be real careful with that pack, as the insulation between cells might be damaged. If cells short, the whole pack could burn. Keep outside where it won't set your house on fire.

It may be repairable if the cells are intact.

Call allcell. Does look like something shorted on the bms or a cell reverse polarity and shorted or the positive shorted against the negative case may be rough handling. Just guessing as I haven't peeled back the onion.

fechter said:

If the BMS tripped on low voltage, then you pound it with high regen from the controller, all the regen current has to pass through the FETs, which are turned off. In the regen direction, the FETs behave like a diode and pass the current. Diodes heat up a lot more than a FET that's turned on.
.

Click to expand...

If that is true....and i bow to your experience on this... Then that sounds like a really flawed system design where someone has overlooked a common situation and failed to build in a failsafe.
It will be very interesting to see the diagnosis of this !....

You said the controller was in the bag with the battery?

If the BMS was next to the controller (especially next to the FET side), the heat buildup inside the bag could exacerbate an existing issue, or even cause one if it got the FETs (or some other part) in the BMS hot enough to fail in some way.

Since Fechter's analysis is probably correct, then the heat buildup from that would probably be the main failure cause, but the extra heat from teh controller might be what pushed it over the edge, if it has gone thru this before but did not fail and was differently packed in the bag where controller heat didn't go into the BMS.


As an aside, generally it's not recommended to run controllers inside bags as it can damage them (over time, not usually immediately). They typically need airflow, or else some other path for the heat to escape.

999zip999 said:

The rule is not to run 2 lvc cutoff but leave something in there ( the battery) and put it back in the charger. Running to lvc will unbalance the pack.
Do you drive your car to the gas tank is empty sucking all the water and dirt in the fuel injection system to have it Clog ? Not a good rule.

Click to expand...

Exactly. Seems a ridiculous way to manage a pack.i bow to Fechter and the more knowlegable, but common sense would dictate not pushing my packs to lvc on a regular basis. As a matter of fact i always avoid. I also avoid contact between controller and pack. Again seems like common sense for even a slow learner like me. Over heat a battery next to a potentially hot controller?

33v lvc is not exactly into unbalancing levels for any half decent cell.
Its not even fully discharge for many which have a 2.5v recomended min voltage.
And isnt the point of a lvc to prevent that type of damage being possible ?

What model of Allcell battery are you using, what is the capacity?

Some observations:
1. I don't see a battery fire, as in, I don't see cells that have gone into thermal runaway and burnt. I see a burnt out BMS, maybe some burnt cables, not more. Happy to be proven wrong.
2. Your practice of running the pack to trip on every use, is bad practice.
3. It sounds like somehow/possibly the pack/bms had tripped and then the controller was trying to charge the pack during regen, but with a tripped BMS, so that energy went into the BMS (instead of charging the cells) and fried the BMS. I don't really understand how that could happen and I suspect that the battery/bms couldn't prevent regen into a tripped pack under certain circumstances.
4. You shouldn't have the controller in the bag, it will produce lots of heat.

You really should be speaking to Allcell first before coming on here. I also suggest you do some more research, read the specs etc. to better understand good/bad practice. I'm not saying that Allcell are without fault or responsibility, but at the same time nobody knows enough detail to start laying all the blame at their door either.

(edited as i came in with a wrong assumption)

cell_man said:

What model of Allcell battery are you using, what is the capacity? That pack looks pretty small and from my basic knowledge of Allcell packs, there is no way that pack was designed to power a 35A controller.

Some observations:
1. I don't see a battery fire, as in, I don't see cells that have gone into thermal runaway and burnt. I see a burnt out BMS, maybe some burnt cables, not more. Happy to be proven wrong.
2. Indications are that you were using the pack at well beyond the power levels it was built to provide (happy to be proven wrong on this point). The fact it was tripping at 33V (which I suspect is the voltage it bounces back to after tripping) suggests the pack is being discharged at far higher current levels than it was intended for.
3. Your practice of running the pack to trip on every use, is bad practice. Running a pack at well above it's intended discharge rate is also bad practice and will at the least shorten the battery life and quute likely result in a failure of some part or other.
4. It sounds like somehow/possibly the pack/bms had tripped and then the controller was trying to charge the pack during regen, but with a tripped BMS, so that energy went into the BMS (instead of charging the cells) and fried the BMS. I don't really understand how that could happen and I suspect that the battery/bms couldn't prevent regen into a tripped pack under certain circumstances.
5. You shouldn't have the controller in the bag, it will produce lots of hear and only make a bad situation worse.

Respectfully, it is a little off to come onto a forum talking about a battery fire when it doesn't appear to be an actual fire and also, it really does appear to me, that the extreme way (for this model) the pack was used and an unusual set of circumstances is what caused (or at least very significantly contributed to) this issue.

You really should be speaking to Allcell first before coming on here. I also suggest you do some more research, read the specs etc. to better understand good/bad practice. I'm not saying that Allcell are without fault or responsibility, but at the same time nobody knows enough detail to start laying all the blame at their door either.

Click to expand...

No way that pack was designed for 35A discharge why doesnt the bms cut-off to protect the cells?????????if this is the case then they put the wrong bms in this pack :lol:

Sure, not an invalid point and that is only 1 of several points made. I also stated that i don't know the specs of the battery and I'd be happy to be proven wrong (I've since had a look and it seems like their typical 36V packs are bigger than i thought, so my point is likely invalid).

The battery maker doesn't control what is connected to their battery and it needs to be able to deliver peak currents without tripping, so it isn't as simple as it sounds, to have the BMS to be reliable, not trip when it shouldn't, yet trip when the pack is worked harder than it should be.

The point i was trying to make, is that it is quite likely the user brought on this issue to some extent and that this doesn't look like a battery fire as such. I personally think that somehow the regen current went into a tripped battery and burned up the BMS. Not quite sure how that could happen and it is something that would be difficult to protect against.

The battery needs to be properly inspected by someone that knows what they are doing, to determine what failed and how that might have happened.

I have no links to Allcell, in fact they could be seen as a competitor of mine. However, from what i have seen, they seem to be a professional company. I just think, the battery needs to be properly inspected by an expert to determine what happened, before potentially very damaging statements are made.

Thanks for your responses everyone.

Yes I'm a beginner and this is my first e bike. I had been trying unsuccessfully to build a battery for the last few months and was relieved when a friend sold me this unused brand name pack.
I have not normally been running the pack down to LVC. This was only the second time it discharged that far.

I knew that the battery would not fully power my 35A controller or 1000W motor. At full throttle the CA3 was showing that the battery was delivering betw 15 and 20 amps.
My understanding was that the BMS would limit the demands of the controller, is that not correct ? And I'm confounded by the Regen frying the BMS, as I thought that the whole point of the BMS was to limit power flow in and out to safe levels. It can't be that unusual to trip the LVC and on the same ride still use your brakes going downhill.
I was well aware of the damaging effects of heat, but I had not heard specifically that you should never put your controller in a bag with the battery.
For the last few weeks after every ride I touched the battery and controller, and the controller would be slightly warm, not hot, and the battery was not warm at all.
However, on this last ride I put the battery and controller in a smaller bag, in an attempt to keep everything from flopping around less, which most have further trapped any heat produced. Although in the smaller bag, the pack and controller were side by side, just as they were in the larger bag, and the heating effects of the smaller bag didn't seem significant.
Cell_Man the reason I named this post Battery Fire is because when I opened the bike bag and then threw the battery on the ground, there were flames shooting out of the top of the pack, which I able to smolder with my sneaker.
As far as the picture above, at first I thought that the controller had burned through the shrink wrap and caused that rectangular line, but there's an equivalent line on the other side of the pack, must be a wiring pathway that heated up along with the BMS.
It seems that I relied too heavily on the BMS, and in the future I'll be sure to separate the controller from the pack.
Thanks again.

Your 20Ah pack should have been well abLe to power the 35 A peak controller.

dougcullen said:

Thanks for your responses everyone.

Yes I'm a beginner and this is my first e bike. I had been trying unsuccessfully to build a battery for the last few months and was relieved when a friend sold me this unused brand name pack.
I have not normally been running the pack down to LVC. This was only the second time it discharged that far.

I knew that the battery would not fully power my 35A controller or 1000W motor. At full throttle the CA3 was showing that the battery was delivering betw 15 and 20 amps.
My understanding was that the BMS would limit the demands of the controller, is that not correct ?

no, the bms can only cut the power, it can't actually limit the current. It's the motor controller that primarily controls the current. The current you are seeing, is likely just the result of the speed you are travelling and therefore power user, rather than any sort of limiting. That pack should be ok with a 35A controller, especially if you are not using it particularly hard, so that shouldn't be a concern

Click to expand...

And I'm confounded by the Regen frying the BMS, as I thought that the whole point of the BMS was to limit power flow in and out to safe levels. It can't be that unusual to trip the LVC and on the same ride still use your brakes going downhill.
I was well aware of the damaging effects of heat, but I had not heard specifically that you should never put your controller in a bag with the battery.

most bms are not really intended to receive current through the discharge leads, as you get with regen. You can get away with it, but the bms wouldn't normally protect the pack from regen current. I and some others suspect the bms could have tripped, then the controller applied regen current to the tripped pack. That is not a normal situation and I'm a little surprised the controller could do that.

BMS have several functions. Preventing over discharge and excessive discharge current the discharge lead. Balancing out any high voltage cells when charged. Preventing any over charged cells when charged through the charge lead.

Click to expand...

For the last few weeks after every ride I touched the battery and controller, and the controller would be slightly warm, not hot, and the battery was not warm at all.
However, on this last ride I put the battery and controller in a smaller bag, in an attempt to keep everything from flopping around less, which most have further trapped any heat produced. Although in the smaller bag, the pack and controller were side by side, just as they were in the larger bag, and the heating effects of the smaller bag didn't seem significant.
Cell_Man the reason I named this post Battery Fire is because when I opened the bike bag and then threw the battery on the ground, there were flames shooting out of the top of the pack, which I able to smolder with my sneaker.
As far as the picture above, at first I thought that the controller had burned through the shrink wrap and caused that rectangular line, but there's an equivalent line on the other side of the pack, must be a wiring pathway that heated up along with the BMS.
It seems that I relied too heavily on the BMS, and in the future I'll be sure to separate the controller from the pack.
Thanks again.

Click to expand...

hmmm. I think that Allcell would be interested to see what happened, even if you bought it used. It really needs to be properly inspected in order to know what happened. After you've explained more, I'm tending towards you weren't using the pack overly hard.

Click to expand...

Apologies if I was a little premature and critical with some of my comments. Battery packs going on fire is every pack assemblers worst nightmare. I have no skin in it, but at the same time, I wouldn't like to see a reputable company getting bad press, when the details are unknown. At the same time, if there is an inherit issue with a battery or bms, and there is any chance of a dangerous situation, then it needs to be put right.

Yikes! A second hand battery will not draw much attention from any maker. I'm sorry but it reeks of bad management, not manufacture. Running to LVC is one admission as well as poor regen practice. To often we don't get or see the whole picture.
I'm surprised there aren't more of these events and suspect there are, just fewer reveals. Dogman's post today was an event that made big changes in practice and management around here.

Ok guess number 3A
The BMS is cutting off at 33 volts because one or more cells is cutting off early or has less capacity. Plus I second regen on low battery. Regen can have huge instant Spike of voltage.

My suspicion of BMS systens has dropped to an all time low if this idea that the system cannot protect a battery at lvc.. (or fully charged ?)..from a regen charge.!
I can well understand why a fully charged pack needs the BMS to prevent overcharge, but fail to see why a pack at lvc should need to be protected from a charge.
Voltage spikes, excess current , high temperature, etc , yes...that is the job i expect a BMS to do, and if it needs to block the cells from any further charge input , regen or other, then it should be able to do that safely, not just churn the excess into heat !. If regen charge becomes too much, couldnt a simple thermal cut out protect the system ?
Im sure there must be "smart" BMS systems that safely do something similar, whilst others may not be so well thought out, but how do we know which to trust ?

Yes isn't the point of the BMS to cutoff power at a Safe level before voltage gets too low ? Tomjasz - What is the proper use of regen braking ? Never use it when your battery is getting low ? Isn't the point of regen to add power to your battery and increase range ?

I've sent this story and pics to AllCell, and I'll let everyone know how things progress.

When any cell gets down to the cutoff voltage, the BMS is supposed to turn off the power going to the controller. Once the motor stops, the voltage will usually bounce back up, maybe enough for the BMS to turn back on.

If the BMS is tripped, then the controller won't have any power and won't be able to activate regen. If the controller is on, then the BMS isn't tripped and can safely pass regen current to the pack.

If you have some way to create regen even when the BMS is tripped, you could have a problem.

Even if the BMS is on, if the regen current is too high, there is no circuit to detect that, nor is there a way for it to turn off in that direction. The controller should be taking care of this.

Most BMS should have no problem handling some regen current when the battery is at a low state of charge. Most bms are a 3 wire type, that have a seperate charge and discharge cable. 2 wire type shares the charge and discharge on the same cable. 2 wire type requires more mosfets to handle the same current levels, they need back to back mosfets.

A 3 wire bms cannot protect against over charge through the discharge lead. This means you shouldn't run a fully charged battery down a big hill and apply regen current. The way to protect against this, is to have a regen voltage limit in the controller.

cell_man said:

This means you shouldn't run a fully charged battery down a big hill and apply regen current. The way to protect against this, is to have a regen voltage limit in the controller.

Click to expand...

That's interesting. When both my bike and my car battery are full, regenerative braking virtually doesn't work. It's that "virtual" part that I'm interested in. If it simply opened the circuit when the voltage hit a certain point, that would be easy.

The first time it happened to me, I got a bit of a shock - going down a steep hill and a truck turns in front of me, twist the throttle backwards, and the Cycle Analyst shows that I'm regenning about 2amp, instead of the normal 20-40a, and I'm not slowing down. The car exhibits very similar behaviour. For the first 2km or so after coming off the charger, the regen graph never shows more than 2-3kw, and there's a big "dead spot" until you push hard enough to engage the mechanical brakes.

How does a controller limit the voltage/current that goes back without totally cutting connectivity? I'm just curious, as I have been doing more reading into basic electronics, to better understand how a lot of these things work.

Both vehicles are commercially mass produced, so I'm not expecting these features to be available to all DIYers, but I am just curious.

id always assumed its the controller activating PWM on the regen current to keep regen voltage at or below HVC. my current controller (from paul) progressively reduces regen to keep HV at or below limit, a cheaper one I have just cuts it off...which is crappy.
I also have blown a 36v battery pack on a 3wire bms with a 48v controller regening on a full pack (before AW set me straight-ta!)

I have seen a 55amp regen spike on my muxus 3,000 on a 72v system no bms. That's on the c.a. I have big alummiun blocks between the 20ah tabs that the voltage muss get thru. Meaning a dd motor can generate a high voltage spike.

dougcullen said:

My understanding was that the BMS would limit the demands of the controller, is that not correct ?

Click to expand...

No. Not in the way you're thinking.

The only thing the BMS does is shutdown the output completely if the battery voltage drops too low.

Some BMSs are also designed to shutdown the output if the current draw exceeds it's limit, but it doesn't limit the current, it just turns the bike off completely.

Only the controller itself limits the current, to whatever it is programmed to or designed for if not programmable.

Unless you use a Cycle Analyst v3, with it setup to direclty control the throttle input to the controller, via the limiting settings in the CA (see the CA manuals on the ebikes.ca site for how to do that).

And I'm confounded by the Regen frying the BMS, as I thought that the whole point of the BMS was to limit power flow in and out to safe levels.

Click to expand...

Not really, as noted above all it does is cut off output if the voltage drops too far (LVC) (and if designed to it will also cut off if current is too high (OC, overcurrent)), and to cut off input if the voltage gets too high. (HVC).

It can't be that unusual to trip the LVC and on the same ride still use your brakes going downhill.

Click to expand...

It's probably not unusual...and it's also possible that various blown BMS problems have been caused by that in the past, too. But I suspect that most of them aren't blowing becuase the FETs don't heat up enough to fail or to desolder themselves and cause a short, etc.

I was well aware of the damaging effects of heat, but I had not heard specifically that you should never put your controller in a bag with the battery.

Click to expand...

It depends on how hot things get--some people have no trouble with it, and some do. So the safest thing is to simply not put the controller or anything else that generates heat into something that will insulate it and keep the heat inside.

Hillhater said:

I can well understand why a fully charged pack needs the BMS to prevent overcharge, but fail to see why a pack at lvc should need to be protected from a charge.

Click to expand...

It doesn't.

The problem is the heat generated by current flowing thru the FET body diodes. This isnt' even a problem if this heat isnt' trapped, which will depend on:

--how the FETs are heatsinked (if they are)
--how the heat path is setup
--if the FETs are on the outside of the BMS or the inside
--if the BMS is exposed to free air or trapped inside insulation
--how much current is passed thru it for how long.

If regen charge becomes too much, couldnt a simple thermal cut out protect the system ?

Click to expand...

Sure--if the thermal sensor is in the place where the heat is coming from, but if it's not where it's hot it can't know about it, and cant' protect against it.

Im sure there must be "smart" BMS systems that safely do something similar, whilst others may not be so well thought out, but how do we know which to trust ?

Click to expand...

You don't, unless you test all the functions you expect one to perform, under all the situations you need it to protect against (separately *and* in combination).

fechter said:

If the BMS is tripped, then the controller won't have any power and won't be able to activate regen. If the controller is on, then the BMS isn't tripped and can safely pass regen current to the pack.

Click to expand...

Not necessarily true--if the motor is spinning fast enough, the generated voltage from the phases back to the controller could be above the LVC of the controller, and activate it, and then regen could activate when the brakes are engaged. On a downhill run, this could happen.