Cause of Battery Fire

[ndrober]

First, everyone was fine and I was able to stop the fire before it spread to other parts of the RV.



With that out the way I had an unexpected fire in one of my two battery packs(photos attached). Each one of them is a “48” volt system with 16 of the Basen 280 AH LiFePo4 cells. Each pack has a Chargery BMS and DCC relay, in addition to a 300 amp class T fuse and 250 amp DC rated breaker with 265 amp conductors. The negative is bonded to the chassis of the RV. Each pack is installed in a metal cage and then covered with a metal box. This setup worked for about 3-4 weeks before one battery pack developed a hard short between the positive cabling and the vehicle chassis, thus destroying the fuse as intended. I tried to switch to the secondary battery pack and screwed up on turned on both at one and destroyed both fuses. Yes I forgot to check the time constants on the breakers, so I just have expensive switches. At this point I am about a week from home and both battery packs done. Both are mostly charge ~54 volts.

Since the fuses are hard to replace, I decided to limp home and stop at RV parks one the way home vs a field replacement of the fuses. I plug this RV every night for 6 or 7 days. I have not issue with the 50 amp connections but have some trouble with ground faults when I had 15 amp connections. Then on my last stop I plug in the 50 amp connection, start up the AC (1500 watts) and water heater (2300 watts) and walk back out to see smoke coming from the battery box. I was able to get the fire extinguished with the fire extinguisher and water hose. The power at the RV park looked good, roughly 240 L1-L2 and 120 L1-N and L2-N, but I did not write it down.

So not I am trying to figure out what happened so I can fix it. The fires destroyed the most positive cell of the 16, while the other 15 are still holding a charge. I bought the LiFeP04 chemistry because it was not supposed to be fire prone. A little reading and what I could find says they will catch fire because of: 1) manufacturing defects, 2) excess charge rate, 3) excess discharge rate. Since the fuse was blown there was no intentional charge or discharge.

After I removed the battery pack I found two cells had continuity to the metal case (which was not the case when it was installed. The cell that burned (Most positive, #16) had continuity from positive and negative to the case. Cell #14 also had continuity from the negative terminal only. I do not know if this was caused by the fire. This battery pack had three previous shorting incidents. Once during assembly the main fuse went. I never noticed this but suspect I bummed one of the fuse terminals with the RV chassis. I went to replace this fuse and dropped the hex nut from the positive terminal creating an arc and welding the nut. Then trying to solder the BMS back on the solder shorted out and melted. While discharging during these shorts, cell 16 is survived all these insults without catching fire before my 3 week trip.

I inspect my DC wiring, in particular the bonding between the 48 volt negative terminal and the RV chassis and found no damaged wiring. I am not sure quite what happened but am looking for ideas. The suspected cause of battery 16 making contact with the chassis ( before the fuse allowing for uncontrolled discharge) doesn’t seem to make sense since it happened while stopped and just after I plugged it into an 50 amp RV plug. Lack of a clear cause also makes it hard to decide to rebuild with metal for fire protection or wood for isolation.

 

[circus]

Hard to visualize using a few snapshots but damn, those metal mounts are close to those terminals. Specially considering the paint shaker ride of a bus. I'm no expert but I'd use a lot of 1/16"(2mm) rubber mat sheeting while assembly. The good stuff you can't tear.
Or even better, eliminate the possible ground somehow.
I just looked how my starter battery is tied down. The brace extends down the side of the battery making movement of the battery or brace impossible.

 

[ndrober]

I wanted to put the top support across the middle before I found out the case of the battery is connected to the negative post, and the small amount of the case exposed near the vent was providing continuity to the bar across there, so I only placed that one in the middle like your starter battery. After trying to change out the fuse while attached to the bus(RV) I had decided to replace the case with wood this winter. Torn between fire load and shorting hazard and went with fire hazard on this first iteration.

 

[Steve_S]

First of all, let me say I'm glad you are OK and that your custom bus survived the experience.

Observations: from what I can see in Photo's.
- Cells were not Bound Together to prevent friction/vibration and reduce expansion/contraction issues.
- Cells should Never Ever make contact with metal. Metal Battery Boxes/Cases etc "MUST" be lined with a nonmetallic, nonconductive liner.
- It Appears to be that the connection to the Cell was either loose or failed somehow. This is not going to be diagnosable due to the damage levels.
- Cells that have been shorted CAN present problems internally resulting in failures. While we think in terms of a Lead Battery taking it (they are brute force tech) because they can take some abuse and a momentary short is not a big deal, BUT Lithium Based is not very tolerant (Any Flavour).

* IF you want to keep batteries under there in that manner, I VERY STRONGLY Urge you to install a Skid Plate to protect that battery assembly from road debris, bumps etc...

 

[HRTKD]

Ratchets straps are not an acceptable (long term) compression device. The straps stretch, requiring frequent adjustment. Probably not the cause, but cells that are allowed to wiggle is not a good thing.

The balance leads should be under the main nut that goes against the bus bar. Placing the balance lead on top of the main nut is generally discouraged.

 

[Sanwizard]

First, everyone was fine and I was able to stop the fire before it spread to other parts of the RV.



With that out the way I had an unexpected fire in one of my two battery packs(photos attached). Each one of them is a “48” volt system with 16 of the Basen 280 AH LiFePo4 cells. Each pack has a Chargery BMS and DCC relay, in addition to a 300 amp class T fuse and 250 amp DC rated breaker with 265 amp conductors. The negative is bonded to the chassis of the RV. Each pack is installed in a metal cage and then covered with a metal box. This setup worked for about 3-4 weeks before one battery pack developed a hard short between the positive cabling and the vehicle chassis, thus destroying the fuse as intended. I tried to switch to the secondary battery pack and screwed up on turned on both at one and destroyed both fuses. Yes I forgot to check the time constants on the breakers, so I just have expensive switches. At this point I am about a week from home and both battery packs done. Both are mostly charge ~54 volts.

Since the fuses are hard to replace, I decided to limp home and stop at RV parks one the way home vs a field replacement of the fuses. I plug this RV every night for 6 or 7 days. I have not issue with the 50 amp connections but have some trouble with ground faults when I had 15 amp connections. Then on my last stop I plug in the 50 amp connection, start up the AC (1500 watts) and water heater (2300 watts) and walk back out to see smoke coming from the battery box. I was able to get the fire extinguished with the fire extinguisher and water hose. The power at the RV park looked good, roughly 240 L1-L2 and 120 L1-N and L2-N, but I did not write it down.

So not I am trying to figure out what happened so I can fix it. The fires destroyed the most positive cell of the 16, while the other 15 are still holding a charge. I bought the LiFeP04 chemistry because it was not supposed to be fire prone. A little reading and what I could find says they will catch fire because of: 1) manufacturing defects, 2) excess charge rate, 3) excess discharge rate. Since the fuse was blown there was no intentional charge or discharge.

After I removed the battery pack I found two cells had continuity to the metal case (which was not the case when it was installed. The cell that burned (Most positive, #16) had continuity from positive and negative to the case. Cell #14 also had continuity from the negative terminal only. I do not know if this was caused by the fire. This battery pack had three previous shorting incidents. Once during assembly the main fuse went. I never noticed this but suspect I bummed one of the fuse terminals with the RV chassis. I went to replace this fuse and dropped the hex nut from the positive terminal creating an arc and welding the nut. Then trying to solder the BMS back on the solder shorted out and melted. While discharging during these shorts, cell 16 is survived all these insults without catching fire before my 3 week trip.

I inspect my DC wiring, in particular the bonding between the 48 volt negative terminal and the RV chassis and found no damaged wiring. I am not sure quite what happened but am looking for ideas. The suspected cause of battery 16 making contact with the chassis ( before the fuse allowing for uncontrolled discharge) doesn’t seem to make sense since it happened while stopped and just after I plugged it into an 50 amp RV plug. Lack of a clear cause also makes it hard to decide to rebuild with metal for fire protection or wood for isolation.View attachment 58887

Glad everyone is safe and OK. The blue plastic on our battery cells is not thick. You need to insulate the cells from each other and from the frame. The strapping also needs to go, as the metal part looks like it will chafe with vibration. For mobile applications, build overkill is better. Use heat resistant plastic and kapton tape, not cardboard.
Open up an SOK or Marine grade battery, and look inside. Thats how you need to secure everything.

 

[mrzed001]

I inspect my DC wiring, in particular the bonding between the 48 volt negative terminal and the RV chassis and found no damaged wiring. I am not sure quite what happened but am looking for ideas. The suspected cause of battery 16 making contact with the chassis ( before the fuse allowing for uncontrolled discharge) doesn’t seem to make sense since it happened while stopped and just after I plugged it into an 50 amp RV plug. Lack of a clear cause also makes it hard to decide to rebuild with metal for fire protection or wood for isolation.View attachment 58887

In 20210728_190940.jpg there seems to be a liquid metal droplet near you main positive terminal (like when you solder).
Also it seems the top rod was to the right (near the main positive terminal). Maybe the cells (top) moved a bit left.
Could you check if there are metal droplets or arc made holes anywhere in the metal frame ?

In 20210728_130153.jpg the main positive terminals rod seems to be melted. And next to it (and to the middle rod) is a screw head.

Did you had fuse only on the positive cable ?

 

[ndrober]

yes the only a fuse on the positive cable. The negative is connected to the chassis after the main bus after the shunt and DCC.

The nut was welded to the support during my first short while when I tried to remove the main positive before replacing the fuse (photo 20210625_165229.jpg). I was on the road at this point and did not have tools to remove the battery pack, so I soldered the BMS connection back.

Following up on Steve's comments on a loose connection, photo 20210808_201217.jpg shows arcing at the bottom of the photo, along with a burn pattern that is not centered on the stud. When I was dissembling it the main positive (in photo) the stud backed out instead of the nut. So it looks like what happened is the main positive vibrated loose and made contact with frame shorting out the cell.

Overall an expensive lesson. I plan to rebuild my other pack with wood interior and metal shield for road debris. Thankfully, right now I do not have a deadline to get it rebuilt.

 

[GSXR1000]

I would say your batteries didn't "catch" fire.. you had an arc flash with a high current discharge that melted the battery....
you can't have that much bare metal around that semi high voltage and extreme current... I put vhb tape between my cells in my rv in a plastic case and plastic covered all thread in a 12v setup... you have 48v and once a small arc starts it ionizers the air whish drops the resistance and allows a direct short

 

[ereams65]

I would agree with GSXR1000; this was an arc, not a fire in the literal sense. Regardless, it's a painful, i.e. expensive, lesson to learn, but the first warning of a potential issue occurred when the nut arc-welded to the stud. That cell should've been pulled from the pack. Any way to move the next box inside the bus?

 

[DerpsyDoodler]

I would agree with GSXR1000; this was an arc, not a fire in the literal sense. Regardless, it's a painful, i.e. expensive, lesson to learn, but the first warning of a potential issue occurred when the nut arc-welded to the stud. That cell should've been pulled from the pack. Any way to move the next box inside the bus?

Before he moves it inside, I strongly encourage a redesign with some assistance from the knowledgable and helpful folks around here. 1 fire is bad. Let's make sure there's no second.

 

[GSXR1000]

I would agree with GSXR1000; this was an arc, not a fire in the literal sense. Regardless, it's a painful, i.e. expensive, lesson to learn, but the first warning of a potential issue occurred when the nut arc-welded to the stud. That cell should've been pulled from the pack. Any way to move the next box inside the bus?

also double our triple wrap that unistrut with heat shrink

 

[GSXR1000]

good thing it was lifepo. Lipo would have taken that bus down

 

[mrzed001]

yes the only a fuse on the positive cable. The negative is connected to the chassis after the main bus after the shunt and DCC.

OK so the fuse on the positive could not blow because it was a direct short between the main minus (connected to the chase) and the main positive terminal.
If there were a fuse on the minus cable it would have blown (saving the pack).

The nut was welded to the support during my first short while when I tried to remove the main positive before replacing the fuse (photo 20210625_165229.jpg). I was on the road at this point and did not have tools to remove the battery pack, so I soldered the BMS connection back.

OMG on the pic it seems the nut almost in contact with the main positive and the chassis. How did it not made a huge arc at this position immediately?

Following up on Steve's comments on a loose connection, photo 20210808_201217.jpg shows arcing at the bottom of the photo, along with a burn pattern that is not centered on the stud. When I was dissembling it the main positive (in photo) the stud backed out instead of the nut. So it looks like what happened is the main positive vibrated loose and made contact with frame shorting out the cell.

Overall an expensive lesson. I plan to rebuild my other pack with wood interior and metal shield for road debris. Thankfully, right now I do not have a deadline to get it rebuilt.

Thank you very much that you shared your build and the pics.
This type of "what did I wrong" posts help new users not to repeat the same mistakes.
Saves batteries, money, even life.

 

[Tonigau]

A mobile battery install needs special attention due to the vibration & shock loads (potholes etc..) not to mention moisture & road grime if its not in-vehicle.
. __>> "I plan to rebuild my other pack with wood interior and metal shield for road debris"
1. Plywood is strong, dampens vibration & insulative for low voltage but is flammable so consider carefully, use Fire Retardent Treated(FRT) if you go this route.
Also use SwitchBoard panel insulation sheet between cells & plywood & in key areas where terminals to metal clearance is difficult to maximize. This sheet is very strong, wear & fire resistant & designed to handle arc flash.
Shield the plywood (reduce height) behind insulation sheet to protect wood from direct arc flash heat.

2. For the mechanical bar holding the top of cells, you might find some thick fiberglass C channel or square tube, put steel tube inside if needed. For other steel bars you could use Fiberglass angle insulated or screw on Insulation Panel.
Put some insulating sheet between the 2 cell banks to widen the space between terminals if required & or use a T-shape as it will always be centered between banks (insulate if metal to protect from loose metal objects).
*** Soft plastic (heatshrink, PVC) over metal bars is not appropriate for heavy mechanical insulation due to wear-thru. ***

3. See if you can find someone (competent & experienced) to asses/check your design/work. In my job we all cross check / review each others design & build.

4. Assess how fire can propagate, ie what can catch fire from what & modify design, use Silicone or Tefzel insulated wire in hazard area.
Consider installing smoke/fire sensor alarm for the power system. Include the charge/load controller in the hazard assessment.

5. Recommend improving enclosure to exclude moisture & dust/road grime. (under vehicle is a very harsh environment).

6. Start a build log on this forum & you will get review & advice as you progress.

7. The battery & associated connections need to be easily inspected periodically, (catch that loose connect before it overheats etc...)

With my sense wires I go direct to the bus bar with a tapped screw & spring washer rather than under the terminal nut.

 

[Cheap 4-life]

Seems like flexible busbars should be used for a mobile setup. Chevy volt uses a bunch of thin copper straps for each busbar.

 

[HRTKD]

With my sense wires I go direct to the bus bar with a tapped screw & spring washer rather than under the terminal nut.

When I switch to different bus bars I plan to put the balance leads on the bus bar not the terminal stud. I think that will make the connections a bit easier to manage.

Did you do that for all the balance leads, even the positive and negative power leads? <-- EDIT this is kind of dumb question since the bus bar doesn't go to those terminals. I'm going to blame the lack of caffeine on that.

 

[DerpsyDoodler]

I used 1/0 with appropriate lugs for my "flexible bus bars"



While I could do it, it isn't the most practical to try and put the sense wires in the middle of these. Had I used flexible copper braid and been heat shrinking them myself, different story.

 

[Tonigau]

Did you do that for all the balance leads, even the positive and negative power leads?

Well your right actually, I put short bars on the B+ & B- just for the sense connections. (SBMS100)
Only high current connections are on the main cell bolts. My install is off-grid home so solid Nickle plate CU bars.
You could put a thread in the Lugs, thick copper is a bit sticky to tap so could using tri-lobe screws to form the threads. (I get mine from old electronic products, just make the hole larger than tapping (2.7 for M3 screw) as the thread is rolled & use oil)

 

[HRTKD]

Nice. Thanks for the explanation and follow up. Welcome to the forum!