House burned down

seneysolar said:

I worry about compressed cells on a shelf that sags in the middle, it squeezes the cells at the top like an accordion.

Click to expand...

Because this^

Composite shelves simply do not hold that much weight for extended periods of time. They can/will/do sag over time.

That style shelving does not support the center and whatever material used is too thin for span/weight.

Periodic maintenance is recommended for other items such as automobiles.

Currently planning a tear down, detailed inspection and reassembly in a battery case with dividers and a redundant fuse on each case. This was already planned. This thread reinforces need for me to follow through with this plan. Batteries already have dividers.

seneysolar said:

Because this^

Composite shelves simply do not hold that much weight for extended periods of time. They can/will/do sag over time.

That style shelving does not support the center and whatever material used is too thin for span/weight.

Click to expand...

The rack itself was steel and rated for 300 kg/shelf if i remember correct. There was little to no sagging. The hpl i just added as divider between the system and metal rack.

I read the entire thread and have an uneducated question:
I currently have 5 sok lifepo4 51.2 100ah units paralleled to a busbar, then a 125a class T fuse before the inverter. It seems a better practice is to have 5 class T fuses between each battery and the busbar.

Assuming I'm correct so far, if I were to perform this upgrade, would I leave the current fuse in place and add the other 5, or would the current fuse become unnecessary or even counterproductive?

Thanks everyone for what you do and for OP sharing his nightmare so that it may help others.

HalfBaked said:

I read the entire thread and have an uneducated question:
I currently have 5 sok lifepo4 51.2 100ah units paralleled to a busbar, then a 125a class T fuse before the inverter. It seems a better practice is to have 5 class T fuses between each battery and the busbar.

Assuming I'm correct so far, if I were to perform this upgrade, would I leave the current fuse in place and add the other 5, or would the current fuse become unnecessary or even counterproductive?

Thanks everyone for what you do and for OP sharing his nightmare so that it may help others.

Click to expand...

Ideally you would fuse everything, but the SOK does have a circuit breaker that should be able to isolate the unit from the system automatically so the string-level fuse becomes supplemental protection.

I like adding the fuse because I don't trust the circuit breakers to reliably interrupt a dead short.

Jejochen said:

The rack itself was steel and rated for 300 kg/shelf if i remember correct. There was little to no sagging. The hpl i just added as divider between the system and metal rack.

Click to expand...

Ok, so it was metal beams with a metal sheet for the shelf (not particle board or any other material) with the HPL on top of/in addition to the original metal shelf?

My apologies if I misunderstood what you have. I see a ton of people using shelves where the beams only go around the outside and a board lay in the middle, nothing supporting the center of the load. There is another post on the forum where a shelf like that likely collapsed.

300kg is not a lot. Even some of the particle board ones from U-Line are rated at 800lbs. They are also rating it in "Uniform Distributed Load" assuming the max weight is the exact footprint of the shelf. It's important that people don't make this mistake, this is the reason for my post.

I'm definitely not trying to be critical of your build, I'd say it's likely in the top 10 for cleanest install, but... sh*t happens. You seemingly had a cascade of failures, and all these bits of information are critical to system design and safety. Let's get it out there for others to learn.

HalfBaked said:

I read the entire thread and have an uneducated question:
I currently have 5 sok lifepo4 51.2 100ah units paralleled to a busbar, then a 125a class T fuse before the inverter. It seems a better practice is to have 5 class T fuses between each battery and the busbar.

Assuming I'm correct so far, if I were to perform this upgrade, would I leave the current fuse in place and add the other 5, or would the current fuse become unnecessary or even counterproductive?

Thanks everyone for what you do and for OP sharing his nightmare so that it may help others.

Click to expand...

So I would leave the class t in place (on the positive between the bus bar and the inverter?) that will protect that wire in case something fails badly in the inverter.

seneysolar said:

Ok, so it was metal beams with a metal sheet for the shelf (not particle board or any other material) with the HPL on top of/in addition to the original metal shelf?

My apologies if I misunderstood what you have. I see a ton of people using shelves where the beams only go around the outside and a board lay in the middle, nothing supporting the center of the load. There is another post on the forum where a shelf like that likely collapsed.

300kg is not a lot. Even some of the particle board ones from U-Line are rated at 800lbs. They are also rating it in "Uniform Distributed Load" assuming the max weight is the exact footprint of the shelf. It's important that people don't make this mistake, this is the reason for my post.

I'm definitely not trying to be critical of your build, I'd say it's likely in the top 10 for cleanest install, but... sh*t happens. You seemingly had a cascade of failures, and all these bits of information are critical to system design and safety. Let's get it out there for others to learn.

Click to expand...

Unfortunately the OP does not seem to have photos of his final configuration before the fire. Photos of several years ago from his old threads are not informative if major changes were made. The 1 photo posted of after the fire is insufficient to establish much. His description of things can be a bit easy to misinterpret. He never stated there was metal shelves only that it was a steel rack. Rack is the frame and may not be a shelf.

I gotta say, this thread has been the exception out of all the fire threads on here I've read. OP's installation looked very professional, to my untrained eye, nothing immediately stuck out as bad. The other threads have had stuff like alligator clips everywhere, cells sitting on particleboard, etc.

Glad OP had insurance and that no one was hurt.

I remember one of Will's recent videos was showing some of the modern powerwall-style batteries seemed to have some kind of fire suppression system built in? If those work at all, I could see the value in paying extra for that.

I am not in any way implying OP did anything wrong here, because I'm too uneducated to be able to tell, but I do notice that these fire threads often have the in-common factor of a DIY battery build as opposed to a prebuilt. Just makes me further shy away from trying it.

Dagoth Ur Does Solar said:

I gotta say, this thread has been the exception out of all the fire threads on here I've read. OP's installation looked very professional, to my untrained eye, nothing immediately stuck out as bad. The other threads have had stuff like alligator clips everywhere, cells sitting on particleboard, etc.

Glad OP had insurance and that no one was hurt.

I remember one of Will's recent videos was showing some of the modern powerwall-style batteries seemed to have some kind of fire suppression system built in? If those work at all, I could see the value in paying extra for that.

I am not in any way implying OP did anything wrong here, because I'm too uneducated to be able to tell, but I do notice that these fire threads often have the in-common factor of a DIY battery build as opposed to a prebuilt. Just makes me further shy away from trying it.

Click to expand...

I think the last big fire that was discussed (Germany, Poland?) ended up being a basen green commercial rack style battery

Dagoth Ur Does Solar said:

I gotta say, this thread has been the exception out of all the fire threads on here I've read. OP's installation looked very professional, to my untrained eye, nothing immediately stuck out as bad. The other threads have had stuff like alligator clips everywhere, cells sitting on particleboard, etc.

Glad OP had insurance and that no one was hurt.

I remember one of Will's recent videos was showing some of the modern powerwall-style batteries seemed to have some kind of fire suppression system built in? If those work at all, I could see the value in paying extra for that.

I am not in any way implying OP did anything wrong here, because I'm too uneducated to be able to tell, but I do notice that these fire threads often have the in-common factor of a DIY battery build as opposed to a prebuilt. Just makes me further shy away from trying it.

Click to expand...

Looks to most of us that the OP did everything correctly at the time he did them. Several things in the realm of DIY battery build have changed over the years. Namely separators between cells and compression methods.

And I agree, the attention to detail in this was amazing. Very professional looking and in all cases. He followed the vendor recommendations on the design and fusing.

The type fuse used probably shouldn't have been even though the vendor recommends it - i.e. the AIC was low and the plastic it was made from could catch fire along with the flame resistant fuse holder. Flame resistant doesn't mean it can't catch, just that it won't in expected operating conditions for running and failure.

What IMO caused the fire is a combination of a cell in one bank shorting and the fuse blowing and igniting into flames because of the resistance of it when faced with full amps from all other banks. i.e. 125a x 6 = 750amps. Math shows the fuse/wire was hot enough to at the minimum melt the fuse and wire sheath to the max of instant flames. That high temp would have happened between 0.01s and 0.1s. High enough to spontaniously ignite hydrogen from a vent or catch anything flamable in contact with the wire between the bus bar and battery. That combined or instead of a vent could have started the fire.

In the Navy they would call this and all fires a failure of imagination. Failure to imagine all ways the system could possibly catch fire.

I've done a back of a napkin calculation with a colleague today while discussing this incident. We assumed an internal resistance per cell of 0.3mR, a 16s bank and 7 banks in parallel. We also assumed one cell failed in a pack with a 0 Ohm/0V resulting short link of that one cell (not sure if this is realistic to assume, but I guess it's worst case). In this case, there would be a potential 5kA current flowing into that battery from the others, or 830A per battery.

While I'm pretty sure the aforementioned mega 48v 300A Mega fuse can break that 830A, that 5kA is likely above the AIC rating (didn't look it up). That said, even if the fuse can't handle the 5kA, the other packs should have blown their fuses and the current dumping should have stopped. With the other packs disconnected, there should now be no more issue in the faulty pack, except that once cell if e.g. internally shorted.

But then, internally shorted cells due to dendrites are not low resistance shorts and they can't sustain large currents and tend to be fragile and not sustained...

robbob2112 said:

buy your own... I paid $38 for these, the largest part was for the ceramic holders...... and I don't have the setup to safely do a dead short.

Click to expand...

This is what I ordered, it has shipped.

Do we know 100% that the fire was started by the batteries / associated wiring / fuses, apart from what the fire department person said?

SeaGal said:

Do we know 100% that the fire was started by the batteries / associated wiring / fuses, apart from what the fire department person said?

Click to expand...

No, we only know that the fuse blew and arced, and that the fire inspector said that was the cause.

Tulex said:

No, we only know that the fuse blew and arced, and that the fire inspector said that was the cause.

Click to expand...

And given the amount of destruction we will never know for sure. The records from the batrium about cell voltage burned up with it. Had that survived it could prove/disprove the theory of a venting/low voltage/shorted cell.

Jejochen said:

A few data screenshots from the system.
The cells were bought as grade A cells from luyuan. Imo the leaking cell was an exception. Sometimes things just break. Thats why i didnt gave it any futher attention. All other 15 cells in that string were fine.
The cells are in a room with steady room temperature, were never discharged or overcharged. Charging specs per cell 2.8 - 3.55 v.
Batrium bms i thought and still think is a great bms. I have no idea why 1 central bms for 112 cells is "bad". If something is wrong with the system design, thats possible, but thats not the bms to blame, it did always a great job.
About being the fuse the source as the fire according to the fire inspector: i never said he said the fuse is the reason the system started to fail. Only it is there the fire started, and as he talked me through the entire research process, i strongly tend to agree with him.
View attachment 212765View attachment 212766

Click to expand...

Looks like only 1 of the Phases (L2) had a large draw? Am I reading that correctly?

So when people are running multiple batteries, there is potential for the others to dump into the faulty one? I feel this is ignorance on my part, hence my clarity seeking question.
Is there potential for this if batteries are paralleled with eachother and then to bus only? Or can this also happen if each battery is wired directly to the bus? If each battery has a correct fuse and breaker, does that eliminate the potential?

To the OP, I realize now my contributions in this thread have only been in regards to discussion of future system safety.
Please forgive my insensitivity. I wish you and your family well, hope the path to rebuild is painless, and thank you very much for sharing on this forum as it has been a tremendous learning opportunity and discussion for others that we may also attempt to avoid a similar fate

A class T should prevent this from happening.
In theory, IF that's what happened to the OP, his mega fuse setup was supposed to, but the mega fuse got overwhelmed by the available short circuit current and simply arced. I guess it maybe made it worse. Class T won't do that though.

SparkyJJO said:

A class T should prevent this from happening.
In theory, IF that's what happened to the OP, his mega fuse setup was supposed to, but the mega fuse got overwhelmed by the available short circuit current and simply arced. I guess it maybe made it worse. Class T won't do that though.

Click to expand...

My wonder is mainly if this can happen due to the way the batteries are paralleled or if it could happen in any configuration (all other batteries in system dumping into a single battery)