diy solar

diy solar

Fire!! Never cover LiFePO4 with wood!!!

From the 8 cells in my van four were completely melted, and four had medium fire damage, melted top vent etc. I've cycled the four cells lots of times now and they still perform like before the fire. They smell very badly so I did nt fit them to the new camper.
 
Second fire (or at least, near fire) which I've seen on the forum here.

However, the cause here looks more obvious than the one from @stobbie , since it does make sense if the terminals are pushed down by the cover/weight and are in direct contact with the top cover.

Which cells do you have? Looks like CATL? Wondering if it's the same cells/vendor as stobbies fire, he had Lishens afaik, but your cell terminals do look different (I'm missing the raised square bottom on the terminals as found on Lishen, and they don't look as EVE either (EVE has the notches facing sideways on the higher parts, and not at a 45 degree angle as on your pictures)

Even tho its 'only' the second I've seen here, I would be really pissed if it was my setup.. My van didn't burn down yet, and I hope it won't happen tho (Always a pita to find out if you're insured properly once something happens, or the insurance company will make an issue of even the slightest they can find)

ATM I do have a risk, since my van build isn't fully finished yet (well, 99%) they might base the value on the original (construction) van... And that's not going to be much, based on a 2007 van with 250k km driven. Need to re-evaluate the final value after conversion once I'm done to make sure I'm properly insured.
 
Thank you very much for sharing. I'm sorry this difficulty, but very glad the results were not worse.

Even though we don't know the cause for certain, your experience impresses on me the potential hazard represented by the massive kWh stored in my battery bank. It speaks to the need for all reasonable precautions;
Proper fusing
Care in mechanical assembly
Assuring that stray debris does not accumulate on cells
Care with respect to combustible materials in the vicinity
Consideration of mechanical stresses imparted through rigid bus bars
A dedicated smoke/temperature detector/alarm (which I have not had, but will install immediately)
General good workmanship-like practices.

I will look at my setup even more critically now. The un-fused BMS leads have always nagged at me as a potential hazard. Normally, they carry no current, but are there hazardous modes of BMS internal failure that happen? But there are so many of these leads that I haven't thought of a workable solution. I don't want to unreasonably over-react, but maybe a fuse in the at least the BMS positive and/or negative terminal leads would be a worthwhile precaution? They are all such small gauge wire they might serve will enough as their own fuse?
 
Thank you very much for sharing. I'm sorry this difficulty, but very glad the results were not worse.

Even though we don't know the cause for certain, your experience impresses on me the potential hazard represented by the massive kWh stored in my battery bank. It speaks to the need for all reasonable precautions;
Proper fusing
Care in mechanical assembly
Assuring that stray debris does not accumulate on cells
Care with respect to combustible materials in the vicinity
Consideration of mechanical stresses imparted through rigid bus bars
A dedicated smoke/temperature detector/alarm (which I have not had, but will install immediately)
General good workmanship-like practices.

I will look at my setup even more critically now. The un-fused BMS leads have always nagged at me as a potential hazard. Normally, they carry no current, but are there hazardous modes of BMS internal failure that happen? But there are so many of these leads that I haven't thought of a workable solution. I don't want to unreasonably over-react, but maybe a fuse in the at least the BMS positive and/or negative terminal leads would be a worthwhile precaution? They are all such small gauge wire they might serve will enough as their own fuse?
Cover the large BMS connections and position the BMS so you can't easily short the BMS body to a terminal.
 
During the installation of the AMmeters I placed a piece of MDF over the cells, and placed my knee on there to keep balance.
It wasn't my full weight, but apparently too much.

The threaded rod of the affected cell was slightly higher then the rest (5mm)
Here I'm happy to even find threaded rods with "allen wrench" (hex) , I can't be choosy.

Pressure on the thread/terminal apparently pressed out of place, and created a short inside the cell.

That got too hot and set the MDF sheet on fire, smoldering as MDF doesn't really burn well.

My walls and ceiling are covered with concrete fiber sheet, to contain accidents, but I doubt they can hold back fire long...

I guess I can say I got lucky, that it didn't burn our house down.

That would be a real shame.

If you are correct that the longer stud was pushed down into the cell causing a short because of your kneeling on the top cover that stud should be able to be pulled right out - for it to go down the threads can not be engaged.
 
OP:

To help us all on here can you tell us:

1) Did you use any isolation between cells?
2) What was under the cells?
3) What type of compression setup and how tight?
4) How did you prep your cells fro use?
5) What type of BMS?
6) Any history on the cells - when ordered, from who, date-codes, QR codes, etc, etc
7) The studs were fixed into the cells how and to what torque? I see you used lock-tite but any other details?

As many pictures during assembly as you have would also be appreciated.
 
If you are correct that the longer stud was pushed down into the cell causing a short because of your kneeling on the top cover that stud should be able to be pulled right out - for it to go down the threads can not be engaged.
It would be tough to push the stud through the nut. And he said the whole terminal pushed down.
But that seems to have caused a short inside the cell.
 
I will look at my setup even more critically now. The un-fused BMS leads have always nagged at me as a potential hazard. Normally, they carry no current, but are there hazardous modes of BMS internal failure that happen? But there are so many of these leads that I haven't thought of a workable solution. I don't want to unreasonably over-react, but maybe a fuse in the at least the BMS positive and/or negative terminal leads would be a worthwhile precaution? They are all such small gauge wire they might serve will enough as their own fuse?

Consider the use of silicone tubes to insulate the wires as shown on page 21 of these instructions.
That is extra protection against shorting, and if the wire does fuse it will contain the molten metal.

 
Definitely important to think through all the possibilities....

I have to say the idea that putting pressure on a terminal can cause an internal short is scary - especially since these batteries expand and contract when charging/discharging - not sure if that argues for more compression, getting rid of busbars, or no compression and leaving space for cells to expand/contract without stressing terminals...
 
Wow! Glad everyone is okay! That could have been very bad.

Conductive Wood? Even if it was from the humidity, the heat would drive the moisture out.

Really doubt it was the wood ... but, I guess if the drywall shipped here can be radioactive why not have something in the glue in the MDF that's got a low electrical resistance? MDF is usually treated to have some level of fire classification, might have been those chemicals. Do you have any specs on the MDF?

Do you have an ohm meter and any non-carbonized MDF left to measure the resistance? That would help confirm/deny the possibility. Perhaps a little dielectric grease at the point of probe contact to ensure good contact?

Update:
During the installation of the AMmeters I placed a piece of MDF over the cells, and placed my knee on there to keep balance.
It wasn't my full weight, but apparently too much. Pressure on the thread/terminal apparently pressed out of place, and created a short inside the cell.
That got too hot and set the MDF sheet on fire, smoldering as MDF doesn't really burn well.
Guess I should have read through the whole thing first ... thanks for sharing your experiences and the updated analysis as to the cause of failure
 
Last edited:
I think you are implying that the short was to the wires. I guess that could make sense. Because it looks like all of the heat was above the busbar.

No, just for Johncfii's concern about balancing wires being something that could short. I think they would blow out and splatter, probably nothing more after that.

As for fhorst's battery pack, I saw how there was thermal damage all around the terminal, and it extended for a ways onto busbar.
I didn't think that mild heating of stud by MDF would do so much to the battery case. Stud is poor thermal conductor and busbar is good thermal conductor with good contact to terminal.

So I figured terminal is what got hot. If not external poor contact, then internal poor contact (either requires current flow). I suppose internal short could do it too. Shoving terminal against layers of other electrode? Not sure of exact construction.
 
Last edited:
First of: ‘happy’ that it didn’t totally catch fire, could have been worse.

@fhorst Could it also have been one of the BMS wires being pinched because of the weight on the MDF? It only requires one of the bms wires to break its insulation, the other side is always at a different potential so energy would flow.
 
First of: ‘happy’ that it didn’t totally catch fire, could have been worse.

@fhorst Could it also have been one of the BMS wires being pinched because of the weight on the MDF? It only requires one of the bms wires to break its insulation, the other side is always at a different potential so energy would flow.
As a newbie I’m now also asking myself if a tiny/smallest bms wire would actually be safer since it would melt away quickly in case of a short.
 
No, just for Johncfii's concern about balancing wires being something that could short. I think they would blow out and splatter, probably nothing more after that.

I think you are correct. Fuses are probably not necessary given the tiny wire size. But the idea of adding some protective flexible tubing around these small wires, for further abrasion resistance, is something I will consider.
 
It would be tough to push the stud through the nut. And he said the whole terminal pushed down.
But that seems to have caused a short inside the cell.

Yes, that was my point. If it pushed down it had to be loose and the nut did not act as a stop so the threads were pushed out and that means it should also "pull" up to prove the culprit.
 
Any recommendations for an insulated battery box, fireproof/electrically insulated materials?
 
Any recommendations for an insulated battery box, fireproof/electrically insulated materials?

Metal, concrete (perhaps tile backer board), sheetrock.
Sheetrock/plaster has the particular characteristic of absorbing heat as it breaks down and releases water. It is used in fire resistant safes.
 
Back
Top