DIY Electric Car Forums banner

1 - 20 of 31 Posts

·
Registered
Joined
·
1,150 Posts
Discussion Starter #1
Sort of newsy: https://www.youtube.com/watch?v=WdDi1haA71Q

Many of you I'm sure know of Rich Rebuilds, the guy who got famous for buying a flooded Tesla and restoring it. This isn't his shop, but it's a shop he was at.

So there is this little princess car thing, and what seems to be an EV shop (they had partially disassembled Teslas sitting in the lot), but might not, hooks up a pair of 400-cell Tesla modules to them to goof around a bit. Seems to work fine while driving and such.

Later the fire starts. A guy manhandles the car outside the shop in time, and the batteries go off like fireworks, raining ash, individual cells so hot they half-submerged themselves into the asphalt. Melted through the box, etc.

Happened yesterday so not much of a post-mortem yet (or maybe ever, if they don't want to disclose it).

Rich says in the video that the cells had no BMS and no cooling... as if that is the lesson to be learned. But that's all he says about the cause.

Best guess from the comments is that they were charging a 12S pack (4.2v/cell = 50.4v) with a 48v forklift charger. A 48v lead acid charger will of course actually charge up around 60v, which is 5v per cell so, naturally the whole pack cooked off.

I don't think that the issue was a lack of cooling. Yeah cooling might have helped, but the batteries were still being massively overcharged.

I don't think that the issue was a lack of BMS. I suppose cells might have been out of balance, and, well, of course if you were monitoring the voltage of every cell it would have shut off, but, that's a bit misleading. You could charge a pack just fine in bulk.

I think they just got overcharged.

One of the commenters pointed out my first thought... that they didn't have the right charger so they hooked up the forklift charger and meant to go check on the voltage every 5 or 10 minutes and then forgot. At least, I'd hope that people doing EV work don't think that a "48v" charger would have been fine to leave on there.

Anyway, thought it interesting, food for thought.
 

·
Registered
Joined
·
565 Posts
I don't think that the issue was a lack of BMS. I suppose cells might have been out of balance, and, well, of course if you were monitoring the voltage of every cell it would have shut off, but, that's a bit misleading. You could charge a pack just fine in bulk.
Think again... a bms would have stopped the charger before overcharging would occur.

only other possibility is debris on the module, shorting a cell's positive and negative terminal. But seeing how fast the other cells started popping, I'm 99% sure either the whole battery, or one cell group was way overcharged.
 

·
Registered
Joined
·
5,041 Posts
I don't think that the issue was a lack of BMS. I suppose cells might have been out of balance, and, well, of course if you were monitoring the voltage of every cell it would have shut off, but, that's a bit misleading. You could charge a pack just fine in bulk.

I think they just got overcharged.
If they were overcharged, a BMS would have solved the problem:
Think again... a bms would have stopped the charger before overcharging would occur.
"BMS" means "battery management system", not "battery balancer". Whatever a BMS might do for balancing and state-of-charge calculations, the most basic functionality is to cut off charging and discharging at safe limits.
 

·
Registered
Joined
·
1,150 Posts
Discussion Starter #4
Think again... a bms would have stopped the charger before overcharging would occur.
Obviously. I said that. You quoted it. "and, well, of course if you were monitoring the voltage of every cell it would have shut off".

To be more clear, what I was saying is that, the problem wasn't that there was no BMS. A BMS would have prevented the problem from having a disastrous outcome, but one being absent is not an issue itself. The problem is that the batteries were being given 4.8-5v per cell.

Similarly, a cooling system might have prevented catastrophe... maybe. I suppose if a cooling system is powerful enough you can leach heat out of the batteries faster than they can make it, even when being given 15% overvoltage. I'm not sure what's happening, chemically, inside the battery at that point. They might still pop, cold.

It's like saying, in a runaway diesel, the issue was that the vehicle didn't have airbags or strong enough brakes. True, but, the actual problem was that there was a runaway diesel.

The problem here (presumably, if they hooked up the old 48v lead acid charger) isn't insufficient ancillary systems, the problem is giving the batteries too high a voltage.
 

·
Registered
Joined
·
5,041 Posts
To be more clear, what I was saying is that, the problem wasn't that there was no BMS. A BMS would have prevented the problem from having a disastrous outcome, but one being absent is not an issue itself. The problem is that the batteries were being given 4.8-5v per cell.
But with a properly operating BMS, the potentially high supplied voltage is not a problem. The charger output would never reach its maximum level, no cell would ever see 4.8-5 volts, and no cell would be overcharged with a BMS; it is real solution, not just a disaster-mitigation system.
 

·
Registered
Joined
·
1,150 Posts
Discussion Starter #6
The charger output would never reach its maximum level
Yeah, I guess I see your point. Seems like playing with fire though. I mean, a bench supply would have also (presumably, we don't actually know that the charger in use was the original one) prevented this and also charged them to the correct voltage.

In that sense, you could hook up a 240v charger to the pack, and expect the BMS to terminate at 50.4v, but, seems a little bit like walking on the knife edge.
 

·
Registered
Joined
·
5,041 Posts
I mean, a bench supply would have also (presumably, we don't actually know that the charger in use was the original one) prevented this and also charged them to the correct voltage.
Not quite, because a BMS monitors the voltage across each cell level, while an external regulated supply can only see the overall pack voltage.

In that sense, you could hook up a 240v charger to the pack, and expect the BMS to terminate at 50.4v, but, seems a little bit like walking on the knife edge.
Any charger/power supply should be current-limited as well, and something needs to ensure that charging current is reasonable... although acceptable charge current is high for Tesla modules (hundreds of amps?) compared to what a random power supply that cheapskate amateurs might have lying around.
 

·
Registered
Joined
·
1,150 Posts
Discussion Starter #8
Not quite, because a BMS monitors the voltage across each cell level, while an external regulated supply can only see the overall pack voltage.
Well obviously, I get what a BMS is and does, but, there are guys who manually balance their cells once a season and otherwise have no BMS at all, and no issue for years. Bulk charging can be just fine.

I mean, I suppose the problem here could have been a balancing issue, but, more likely was just straight overcharging the whole pack.

It's I suppose a silly hypothetical, obviously you're best off with both (I suppose not obviously, some are strong proponents that BMSs are a dead pack waiting to happen), but, if you had to pick one or the other, or blame the absence of one versus another, I'd pick a charger with the correct max voltage and rely on it versus relying on a BMS to keep a higher voltage from murdering a pack. One's only an issue if there's significant balancing issues, the other is certainly a significant issue every single time.
 

·
Registered
Joined
·
565 Posts
Well obviously, I get what a BMS is and does, but, there are guys who manually balance their cells once a season and otherwise have no BMS at all, and no issue for years. Bulk charging can be just fine.
No it's not. yes it can be fine for seasons...but no guarantee, and most people end up killing cells, or packs.

But...this practice usually was with LFP cells, they are a bit more difficult to ignite.

Also this was when some available bms systems did more wrong then good.

I mean, I suppose the problem here could have been a balancing issue, but, more likely was just straight overcharging the whole pack.
That's basically has the same effect, though it is likely that the whole pack was overcharged.
Another possible issue is the way Rich handles the modules...lots of physical abuse, and coolant everywhere... Leak coolant into the modules and they will get out of balance...seen it happen... I won't buy separate modules from car wreckers...

It's I suppose a silly hypothetical, obviously you're best off with both (I suppose not obviously, some are strong proponents that BMSs are a dead pack waiting to happen), but, if you had to pick one or the other, or blame the absence of one versus another, I'd pick a charger with the correct max voltage and rely on it versus relying on a BMS to keep a higher voltage from murdering a pack. One's only an issue if there's significant balancing issues, the other is certainly a significant issue every single time.
Better think again...when using LFP cells (and some others might also be more forgiving / harder to get to burn) this might hold -some- truth (5 years ago...)
But when using Tesla modules, or other OEM modules that are highly flamable...

Please reconsider these thoughts...and don't promote people to use these ideas...
 

·
Registered
Joined
·
248 Posts
Just like that guy with the Jeep that burnt down his house. The site has been wiped of the story but he wasn’t using a BMS on these modules either. Proof that you NEED TO USE A BMS! Don’t listen to Jehu. Lol
 

·
Administrator
Joined
·
6,157 Posts
Just like that guy with the Jeep that burnt down his house. The site has been wiped of the story but he wasn’t using a BMS on these modules either. Proof that you NEED TO USE A BMS! Don’t listen to Jehu. Lol
Nope
You don't need a BMS if you are sensible
If you are not sensible then a BMS will probably not help anyway
 

·
Registered
Joined
·
735 Posts
Sorry Duncan, sensiblity is just not in the DNA of a lot of people, Also, think about it with these Tesla modules. They get old and start to drift(as eventually we all do) on the voltages. You get a little lazy about manually monitoring voltages and temperatures. You forget just one time and leave the charger on just a little too long: KABOOM!

Judging by the Tesla module failure incidents, it IS most likely going to be a KABOOM. Not a swell, stink and sizzle failure like with your Volt cells, and other safer cells. I really can't believe if you had Tesla modules in your EV, that you wouldn't be using a good BMS.
 

·
Administrator
Joined
·
6,157 Posts
Hi electro wrks

Disagree
Every failure I have heard of (other than total loonies like the video) has been a two stage failure
(1) cells die
(2) charge up as usual - leading to an overcharge

What is needed is a warning that a cell has died - I use the Lee Hart Batt Bridge

I also charge and discharge conservatively

if I could use the Tesla (or Chevy) BMS I would but the aftermarket BMS's appear to simply CAUSE problems
 

·
Registered
Joined
·
4,056 Posts
"Tesla Battery Fire, What Went Wrong?"


Good second video Kevin.

It was a single cell that appeared to overheat FIRST. This could be because of it being unbalanced more than another and hit a higher peak voltage first, or that it has a higher IR than another causing overheating. But it was the one that started the chain reaction. Not only that, but at the top of the charge, the cell has the most energy in it, so the most potential to be volatile.

That cell was the instigator. The rest was all thermal runaway.

IMHO They should have immediately sprayed water, and kept spraying the pack with water to cool it down. From the video, it doesn't look like they did a lot during the fire.

We had to do this with a Motoczysz bike with Kokam cells when they overheated during a race at Portland International Raceway for the TTXGP. It saved the bike, and kept the other cells from igniting.
 

·
Registered
Joined
·
735 Posts
Hi electro wrks

Disagree
Every failure I have heard of (other than total loonies like the video) has been a two stage failure
(1) cells die
(2) charge up as usual - leading to an overcharge

What is needed is a warning that a cell has died - I use the Lee Hart Batt Bridge

I also charge and discharge conservatively

if I could use the Tesla (or Chevy) BMS I would but the aftermarket BMS's appear to simply CAUSE problems

Could you please show us how you would use Lee's Batt Bridge with Tesla modules? Maybe a schematic?
 

·
Registered
Joined
·
1,150 Posts
Discussion Starter #17
It was a single cell that appeared to overheat FIRST. This could be because of it being unbalanced more than another and hit a higher peak voltage first, or that it has a higher IR than another causing overheating. But it was the one that started the chain reaction.
I disagree with what I think you're concluding with that assessment.

If you put 800 of anything flammable in a box, and then apply heat, it will always be 1 of them that ignites first. They won't synchronize and all burst into flame at the same moment.

If you take 800 strands of spaghetti and bend them, 1 of them will snap first. If you keep bending, soon after, more and more strands will break. This does not mean that the first strand of spaghetti "set off a chain reaction", as if all the other strands were still okay at that point. They were all already on verge of snapping.

If you rolled 800 dice twice, the few that rolled a 6 both times doesn't mean those dice were out of balance. Some are just going to do that. You could argue that it's not "randomly", and, sure it comes down to microscopic differences in the dice and the way they were thrown, but, it's random.

In fact, the cell that first exploded had many other cells in parallel with it. That whole parallel group, by virtue of being in parallel, were balanced with each other, yet they didn't light like a 21-gun salute.

In terms of the battery, it would be misleading to conclude that the cell exploded because it was imbalanced and then its fire caused its neighbors to overheat. I would find that to be the case if the rate that the fire propagated seemed proportional to the heat a cell gained from its ignited neighbor, but that doesn't appear to be what happened here.

You can hold a blowtorch against an 18650 for a long time before you can get it to ignite. That's what you'd expect if it was just one cell that was imbalanced. That the cells burst so quickly in succession, to me, indicates any one of them could've been the one that popped first.

I suspect that *all* cells were right on the very brink of igniting, *all* were simultaneously overcharged, and that the only thing the first cell popping did was provide a small bit of extra heat to tip the rest over the edge.

That is, the fire was not caused by a chain reaction of the cells being out of balance, and would not have been prevented or even delayed by more than a few seconds had the cells been perfectly balanced. Even 800 cells with perfectly equal voltage, one would have ignited first. The fire was caused by all the cells in the pack being massively and cohesively overcharged until one finally lit.

The takeaway should be "Don't massively overcharge your battery pack", of which this result was guaranteed, every single time, not "This is what happens when you have imbalanced cells", which is something that might be an issue 1/1000 times.
 

·
Registered
Joined
·
4,056 Posts
Well, you're wrong with what you think I'm concluding with my assessment. I don't see where I said the others weren't on the verge of igniting.
It just so happens, we're mostly in agreement!

In summary:

One (or more) hit it's thermal limit before the others. That is fact. . The video supports this. It's hard to tell if one or more simultaneously ignited, but it was definitely within the same area on one of the packs.

You and I both seem to surmise that it did so because it was overcharged. That is reasonable and prudent to assume.

You and I both seem to surmise that if that one didn't hit thermal limit due to an overcharging situation, a different one likely would have. That is also reasonable and prudent to assume.

What I do think, is that it all starts with one (or more) going into thermal limit first. From my experience, one cell going thermal can start snowball effect that heats up the next cell (on the edge or not) and then the next, and the next. That is the basic idea of thermal runaway.

Long and short, it is likely that all cells got overcharged, and this one hit its thermal limits first, which caused a chain reaction.
 

·
Administrator
Joined
·
6,157 Posts
Hi electro wrks

All you need to use a Batt Bridge is a centre tapping - and it does not NEED to be "center"

What you are doing is comparing the two halves - or quarters with each other
When I had four strings of Headways I used four Batt Bridges

The "comparison" is ongoing and if the two halves are not the same than a lovely red LED will light

This means that you are comparing the "halves" under load - not just under no-load conditions
If you "lose" a cell it will show up under load - and then you will need to get in there with your meter to see what has gone wrong
 

·
Registered
Joined
·
106 Posts
You and I both seem to surmise that it did so because it was overcharged.
Based on the second video describing how they just used a charger they had on hand, I also agree that overcharging was the main issue. Technically, all the paralleled cells in a cell group in the module would have been at exactly the same voltage, and perhaps all the cells in the group were getting hot, but slight differences in each cell in the group and their point of pressure release and ignition would result in one individual cell failing first.

A charger with an appropriate lower voltage may have prevented the problem, if you assume all the cell groups in the modules were balanced with each other. But a BMS that shuts down the charger as soon as any cell group reaches the cutoff voltage would also have prevented the problem regardless of the characteristics of the charger and any imbalance between the cell groups.
 
1 - 20 of 31 Posts
Top