Fire!! Never cover LiFePO4 with wood!!!

[Luthj]

As far as how a terminal connection could go south, with that many cells in parallel, and non matched bus bars, and triple stacked in some places... Well, one cell lags behind the others due to higher contact resistance, once you hit the knee, or close to it, that slow charging cell can see 80% of the total current instead of its share.

With a 100mv drop at that terminal, to get 25w dissipated power, you would need P= V*I 25/0.1=250A. Of course this is all guesswork.

What is the voltage of this battery pack?

 

[Just John]

That doesn't look like what I was expecting to see under there.
I expected the terminal to be gone.

I don't understand what you are saying caused the fire.

My understanding is that one terminal stood up farther than the rest, and was physically pressed into the cell when he stood or sat on them the previous day. Seems likely, I doubt they have a lot of structural integrity to support significant weight placed on a single terminal.

 

[Luthj]

I doubt they have a lot of structural integrity to support significant weight placed on a single terminal.

I must have missed the sitting issue. Sounds like user damage. If a terminal is showing pressed inward, that could have easily shorted the cell internally, or caused a high resistance busbar. These cells only have 5-10mm of clearance between the terminal bottom and the top of the jelly roll. So any significant movement causes issues.

Anyways, I think the OP is damn lucky to still have a home and battery pack. There are a lot of issues with this pack, and some obvious carelessness.


Only fool would consider sitting on a battery pack. Heck, those terminals can be damaged by overtorquing them.

 

[Frank in Thailand]

I would try to use all of them except the one that shorted. Unless there is something I'm not seeing.

A lot of them hot hot, so hot the black plastic cover melted.
Some cells the plastic around the terminal got funny an some spots.

I will start to fully charge them, and then do capacity test.

Sadly only one meter, and 8 cells optical damage.
4 others I don't thrust atm, as they where close to the heat as well.

First one looks bad.:

Like a blister on the side of the terminal.

The others have mostly retracted top plastic.

.

This last one is clear...
Not to be used!!



All the 24 cells got warm/hot.
Obviously, the location where the fire / smoldering was, got the worse beating.

I was surprised that the cells on the other end where not that much cooler!

Warm/hot around/over 45 degrees Celsius.
1.5 hours after the fire, but they are massive and compacted, the temperature would not have dropped a lot.
I was mostly surprised how little difference the temperature of the other cells where.

That does show how thermal efficient their build is.
One other thing that this accident showed, is that the extraction under high heat of my thick copper Bus-bars didn't affect the terminals / threaded rods.

None of them where bend.
I'm going to test the Bus-bars on a mirror to see if they bend.

If not, the whole discussion about this can end.

If this fire have no influence, for sure daily usage won't have!!

 

[Frank in Thailand]

I must have missed the sitting issue. Sounds like user damage. If a terminal is showing pressed inward, that could have easily shorted the cell internally, or caused a high resistance busbar. These cells only have 5-10mm of clearance between the terminal bottom and the top of the jelly roll. So any significant movement causes issues.

Anyways, I think the OP is damn lucky to still have a home and battery pack. There are a lot of issues with this pack, and some obvious carelessness.


Only fool would consider sitting on a battery pack. Heck, those terminals can be damaged by overtorquing them.

Over torque is totally different then pressure from above.

I guess it takes a fool to know one.

 

[fafrd]

Over torque is totally different then pressure from above.

I guess it takes a fool to know one.

This is a very serious matter and insults in any direction are really uncalled for (and I know it’s not you that threw the first insult).

I appreciate your bringing this failure to the Forum and would like to see us all focused on what could have caused this fire.

Seeing your photos brings into focus exactly how dangerous these LiFePO4 Calle can be if not treated correctly and protected with the proper safeguards.

I’m very happy your accident was not more catastrophic and would like to see the community help you understand what might have caused this fire.

One terminal clearly reached the highest temps and likely caused the wood to catch fire, but it seems to me that many terminals were reaching dangerously high temps that resulted in heat damage.

Can you tell us again what currents were flowing between the last time you saw that the battery was fine and the discovery of the burning fire?

I had not thought about installing a smoke alarm in my cellar, but I will now.

 

[fafrd]

Well....
The contact between the cell terminals and bus-bar is absolutely tight as possible with aluminium threads, that where reinforced with Loctite, and firmly connect.
Internal resistance meter showed practically no increase, making that the contact as strong as possible.

It wasn't under high load.
500-750 watt on +1000Ah
And at 09.00 the sun wasn't providing a lot of energy.

when you say this, are you saying that after you last saw the battery it wasn’t charging and was discharging at an average rate of 0.5-0.75kW (meaning 10 to 15 Amps @ 48V)?

How many hours passed between when you left the battery and next checked it? Overnight? 12 hours?

Whatever was consuming the 0.5-0.75kW, was it relatively constant over that full period? Were there any spikes in consumption and if so, to what power?

The battery was is "resting phase" not actively charger or discharged.

So I don't know.
I do know that it wasn't a problem for weeks and weeks.

I do know that the one day I did cover it with MDF sheet, that sheet catched fire.

As that was the only thing changed in that area, it's a solid assumption that this was the cause of the fire.

All other suggestions are more speculation.

Even high resolution pictures won't give you the answer you look for.
As it's after the fire, where you can see the terminals where hot enough to melt the plastic around them.

I'm quite sure they have no longer optimal contact now, after the fire.

Throughout the ‘weeks and weeks’ was the surface of the battery exposed to air? Leaving the MDF on top of the battery is obviously going to affect the passive cooling of the battery surface - had you ever measured the temperature of your cells after a similar night of use?

 

[Frank in Thailand]

It will stay a mistery as I'm not planning to make the same fire again.

For fires caused by wires:
Yes, possible.
Not the BMS wires. 1.5mm2 with thick insulation, no way.
The screws used had about 1mm spacing between the plastic cover and the bolt/screw.

There was a thin wire from temperature sensor, with an iron top.
I noticed in the past that extending the BMS spaghetti wires with cat6 leads isn't safe.
A tip from @ghostwriter66 , I thought was a good one.
Turns oute they hve so thin insulation that it's not enough to prevent a short in high amperage environment.

Using thin wires as suggested isn't a solution here, it doesn't melt if it touches the MDF, it will start glowing, and might have been the start.

BMS wasn't connected yet.
For the people who want to start a discussion about this, yes, absolutely right it should not.

After 6 days of building the battery pack (total 80 cells) that is placing and connecting the Bus-bars to be able to get the 48(51.2) volts, dark at night, no airconditioning, you'll be happy to just have electricity.

It's hot season. +35c
Those high temperatures at night aren't good for sleeping.
The high temperatures during day time not the best also for working

6 days?
Yes.
Check, double check, tripple check.
Ask second person to look for errors after explaining what they might be.
(In this situation just not tightened nut and bare wires or end of wires)

I have a medical condition, chronic neuropathic pain after surgery, now 17 years ago.
When a doctor see my pile of pain meds the ask "how can you still be walking?". And they are genuinely amazed. Normally they see their patients in half coma with it.

Needless to say I haven't been working since.

The first few years I could not read Donald Duck comic, I forgot the first picture halfway.
A walk to the store (750m) and bring back light groceries was enough to nock me out for 3 days.

After 17 years I've learned to live/work with the energy limitations the pain give me, and the mental limitations the medicine give.
The meds mostly take away the pain, they don't give back the energy pain eats.

Ever had real tooth pain? The type that requires dentist to fix (or pull out yourself)
That's neuropathic pain.

And it's not always , but constant waves.
Like when you hit your finger with a hammer. Not the first sharp pain, but the pounding waves with sharp princes.
My whole lower right rib feels like that, like some one hit it with a baseball bat.
Always. Day and night.

Good news is that I can't damage it, only irritate, what gives more pain, that takes a week or so to reduce.

Sure.
I'm making mistakes.
Some might call them foolish, but they probably are fools themselves.
My wife isn't foolish, yet, she would sit on there without thought.

My meds make that I'm less sharp in seeing all the connections, all the variables. That makes me just human, no longer super human.

Silly, Yes. I was "Super human" with a large pile of certification from Microsoft (mcse), Novell (master cne) Oracle, Cisco, SAP, etc.
My job was to fix the problems the rest could not and flew all over Europe, ME and Africa.
Third line software support for really big Japanese company
Thrust me, you must have a lot of certification before Vodafone UK HQ let's you play with their servers.

That was difficult to accept.
Being almost brain dead
Also no more gym, u used to be there 5 days a week 90 minutes.
Again, hard to accept. The body build in 15 years see and feel getting weaker every day/week/month/year.

I'm still here!
And sure, I'll make mistakes.
Not mistakes others would not have made with less in-depth knowledge of the materials.
I do have it, sadly can't always combine to notice all variables.

Was it a dumb action to cover 24/36 threaded rods with a sheet of wood? No, it wasn't

Was it a dumb action to rest on that wood, apply like 40 kg or less?
No it wasn't. 1.5 kg on a terminal isn't a lot.

Was it a dumb action to have (slightly) different size of rods?
Again, no it wasn't.
Thailand is not USA.
Especially rural areas.
You are happy to find studs, even that they are 5mm different length and not stainless but carbon steel.
Stainless long threaded rods (e.g.50cm) can't be found to make yourself.
Galvanized iron, only 1/4 inch.
Not 6mm.

Is working with the materials that are available foolish? It isn't.

It does increase the chances of things going wrong.
That's for certain.

One of our forum members has a nice tag "some people ask "why" others ask "why not?"

I guess I'm part of the last group.
This got me from not being able to read the Donald Duck to building +200 1060 GPU mining farm, and now our house in Thailand.

 

[fafrd]

As far as how a terminal connection could go south, with that many cells in parallel, and non matched bus bars, and triple stacked in some places... Well, one cell lags behind the others due to higher contact resistance, once you hit the knee, or close to it, that slow charging cell can see 80% of the total current instead of its share.

With a 100mv drop at that terminal, to get 25w dissipated power, you would need P= V*I 25/0.1=250A. Of course this is all guesswork.

What is the voltage of this battery pack?

Sounds like he has 48 280Ah cells for 1000Ah, so 36V (6S8P) battery?

The discharge rate was apparently 500-750W which would translate to 15 to 21A or 0.007C to 0.01C.

Also, sounds like the battery was nowhere near the knees.

The impact of insulating the top of the battery pack during this (probably overnight) modest discharge rate would be useful to understand.

Though the two cells with the highest grubscrews certainly seem to have been the most severely impacted, so the possibility of mechanical damage associated with pressing a terminal down into the cell should not be discounted.

There are a large number of fragile parallel connections to the terminals internally, so if mechanical stress downward can cause some of many of those internal connections break or in some other way disconnect or become stressed, that could add a lot of internal resistance just under the terminal.

I think you may be in the right track thinking about differences in IR at those two cells, but the cause is likely to be internal mechanical damage rather that poor busbar connections, the magnitude of the IR is possibly much higher than you are considering, and the generated heat is likely to have been generated far from the knees and under very modest discharge rates.

 

[Frank in Thailand]

The sheet of wood was there for a few hours, I just forgot to remove it.

That wasn't the problem.
What might have been, the thin thermal sensor wire, could have been on top of a stud.

They are from the temperature based fan controllers and where not yet tugged nicely.

Having a metal head...
That could have made contact with an bus-bar.
Being squashed between wood and stud might have given enough contact to start glowing there.
And start the fire.

Could have, might.

I don't know.
I got lucky and have installed a battery based smoke detector now.

It just shows the dangers there are, working with our high capacity batteries.

There is one thing funny about my medical condition..
I thought my father at 85 how to cope with less energy after he got leaking hard valve.
And though my mother how to deal with the "fuzzyness" that happens when you get too tired, how to spot and how to act.
Fuzziness... Compare it to severe sleep deprivation, then run a marathon and after that, directly go to birthday party and try to have a "normal" conversion.
That's about the level of brain activity I have at the end of a day, or when using too much energy.
I can dig a hole for 5 minutes, and need 2 days to recover.
Or use my excavator for a few hours.
Choises

 

[Frank in Thailand]

The setup now.

32 X 280Ah S16, P2
48 X 152Ah S16 P2
No longer P3 as cells are lost.

60 cm high, 70 cm deep, 75 cm wide.

As you can understand, hanging over 70cm is not that easy for longer periods of time.
Even with a young healthy body.

 

[Frank in Thailand]

The wire mess.
I was in process of installing the units in parallel of eachother.
Together 3*3.2kw, 9.6.
The cables needed are long, and not able to make shorter without soldering and buying new plugs.

After that, I started to install the AMMeter, that gave a bang!

Rated for 600v, the 350v from solar was clearly too much.

If I installed correctly?
Yes, absolutely. I first tested the installation with my lab power supply at 30v and light loads.
If that works..
It air rocket science.
Even with a fuzzy brain I know where to place + and -.
And it didn't say bang till I gave it 12v external power, that is rated between 4.5 and 30v.

After that...
My day stopped.
Next day, 12-14 hours later, back in the room and that story you know.

Before someone thinks a part of the AMmeter might have set the fire...

Yeah, could be.
After i stopped I was busy cleaning th tools and stuff for about 30 minutes.
I didn't notice anything.
Nothing was laying on top of the wood.
Seeing the damage as it is, the source came from underneath the wood.,

 

[Frank in Thailand]

Discharge rate of 500-750 is already high.
Just airconditioning, lights, mobile chargers and refrigerator use about 130-150 watts in the early morning (outside +/- 24, airconditioning at 26)
Sure, at higher temperatures, the Aircon works harder, but max 500w at turbo. 350 normal max speed.
(Daikin high efficient units, 8500BTU, at turbo 11500)

Total capacity 1016Ah, S16.

The discharge was practically "non existent"

For those who wonder why we want that high capacity of LiFePO4..
They don't live off grid.
And need to use their cells for 10 days or longer with minimal charge from the solar due rain.

The rest of the year the extra energy can be used for Crypto Mining.
No +200 GPU 1060 anymore..
Just 6 and one 3070.
Still a nice extra income

 

[Frank in Thailand]

Yesterday I did too much work in one time.
I walk/drift around like a person with high fever (or drunk) with probably the same level of brain activity...
If I talk too fuzzy, sorry for that.

The stuff works for now.
I will keep it in the 30-70% range as I'm sure not all contacts are optimal and no BMS installed yet.

During installation of the lower set of 152Ah sets I stripped 4 cells from thread.
(Working too fast, not careful enough. 5 or 6 others I felt "slide" and stopped twisting.. the 280's are OK, I was more careful there)

Tap set (3 version M8) with bottom tap is ordered, as 8M stainless steel rods (32cm) washers and nuts.
They should arrive In the next 2 weeks.

Time enough to slowly charge the cells to full, and do capacity test.

If my calculations are correct it should take 5 to 6 hours to get empty with my tester.
Enough time to prepare

And time to study how to do "blind taps" in cells (and first practice on something like heatsink)

Thanks for the positive feedback, and thinking with me!
(Most of you)

I share my mistakes for educational purposes. Not to be bashed around.
Mistakes that are not unique to me and my installation.

While it might not be what you would do, you aren't here and you don't have the limitations in materials and the rest as I have.

Be kind, be silly, be happy

Time for recharge my internal battery. That will take a day or 3 before I can do anything useful.
Compare it like a flue, with fever you won't do much activities also

 

[diyernh]

I was using loctite to permanently install my studs, not to attach the nuts. It's really about avoiding the stripping problem and making the stud install solid. Red loctite is used for this and you would never want to use it on the nuts because it's basically permanent. Blue loctite is used for assemblies which might need to be taken apart again.

I did a write up of the red loctite stud install process (link below) that explains how to do it without making a mess of your terminals. After making a mess of my first pack, I figured out the process and the rest of the cells were fine.

Loctite install process for studs

I've seen much debate on Loctite vs. helicoil, etc. to strengthen studs. I'd say helicoils are the way to go if you want the strongest option, but I decided to go with red loctite since my threads seemed pretty good. I did my first 4 cells and ended up making a bit of a mess of the terminals...

diysolarforum.com

In my job (military defense contractor), the application of Loctite is only a drop. Apply it to the screw/stud. You will see the drop wick a little around the screw, continue until it covers a 2-4 threads. As the screw goes into the nut/terminal, it will coat the thread surfaces. Once it's tightened, it will cure the Loctite when the two metal surfaces of the thread compress.

Loctite is anaerobic. It will only cure in the thread areas that are in close contact. Having liquid Loctite in the bottom of the hole won't cure, but will be leaking out for a while. Just a drop is all you need

 

[diyernh]

It will stay a mistery as I'm not planning to make the same fire again.

For fires caused by wires:
Yes, possible.
Not the BMS wires. 1.5mm2 with thick insulation, no way.
The screws used had about 1mm spacing between the plastic cover and the bolt/screw.

There was a thin wire from temperature sensor, with an iron top.
I noticed in the past that extending the BMS spaghetti wires with cat6 leads isn't safe.
A tip from @ghostwriter66 , I thought was a good one.
Turns oute they hve so thin insulation that it's not enough to prevent a short in high amperage environment.

Using thin wires as suggested isn't a solution here, it doesn't melt if it touches the MDF, it will start glowing, and might have been the start.

BMS wasn't connected yet.
For the people who want to start a discussion about this, yes, absolutely right it should not.

After 6 days of building the battery pack (total 80 cells) that is placing and connecting the Bus-bars to be able to get the 48(51.2) volts, dark at night, no airconditioning, you'll be happy to just have electricity.

It's hot season. +35c
Those high temperatures at night aren't good for sleeping.
The high temperatures during day time not the best also for working

6 days?
Yes.
Check, double check, tripple check.
Ask second person to look for errors after explaining what they might be.
(In this situation just not tightened nut and bare wires or end of wires)

I have a medical condition, chronic neuropathic pain after surgery, now 17 years ago.
When a doctor see my pile of pain meds the ask "how can you still be walking?". And they are genuinely amazed. Normally they see their patients in half coma with it.

Needless to say I haven't been working since.

The first few years I could not read Donald Duck comic, I forgot the first picture halfway.
A walk to the store (750m) and bring back light groceries was enough to nock me out for 3 days.

After 17 years I've learned to live/work with the energy limitations the pain give me, and the mental limitations the medicine give.
The meds mostly take away the pain, they don't give back the energy pain eats.

Ever had real tooth pain? The type that requires dentist to fix (or pull out yourself)
That's neuropathic pain.

And it's not always , but constant waves.
Like when you hit your finger with a hammer. Not the first sharp pain, but the pounding waves with sharp princes.
My whole lower right rib feels like that, like some one hit it with a baseball bat.
Always. Day and night.

Good news is that I can't damage it, only irritate, what gives more pain, that takes a week or so to reduce.

Sure.
I'm making mistakes.
Some might call them foolish, but they probably are fools themselves.
My wife isn't foolish, yet, she would sit on there without thought.

My meds make that I'm less sharp in seeing all the connections, all the variables. That makes me just human, no longer super human.

Silly, Yes. I was "Super human" with a large pile of certification from Microsoft (mcse), Novell (master cne) Oracle, Cisco, SAP, etc.
My job was to fix the problems the rest could not and flew all over Europe, ME and Africa.
Third line software support for really big Japanese company
Thrust me, you must have a lot of certification before Vodafone UK HQ let's you play with their servers.

That was difficult to accept.
Being almost brain dead
Also no more gym, u used to be there 5 days a week 90 minutes.
Again, hard to accept. The body build in 15 years see and feel getting weaker every day/week/month/year.

I'm still here!
And sure, I'll make mistakes.
Not mistakes others would not have made with less in-depth knowledge of the materials.
I do have it, sadly can't always combine to notice all variables.

Was it a dumb action to cover 24/36 threaded rods with a sheet of wood? No, it wasn't

Was it a dumb action to rest on that wood, apply like 40 kg or less?
No it wasn't. 1.5 kg on a terminal isn't a lot.

Was it a dumb action to have (slightly) different size of rods?
Again, no it wasn't.
Thailand is not USA.
Especially rural areas.
You are happy to find studs, even that they are 5mm different length and not stainless but carbon steel.
Stainless long threaded rods (e.g.50cm) can't be found to make yourself.
Galvanized iron, only 1/4 inch.
Not 6mm.

Is working with the materials that are available foolish? It isn't.

It does increase the chances of things going wrong.
That's for certain.

One of our forum members has a nice tag "some people ask "why" others ask "why not?"

I guess I'm part of the last group.
This got me from not being able to read the Donald Duck to building +200 1060 GPU mining farm, and now our house in Thailand.

Don't get bogged down in others comments/opinions. There are few people who ever put problems/failures out there. The social media is filled with bragging, selfies, "look at me", etc. You rarely hear from any of those when something goes wrong. If you do mention something to look at, it is taken as an insult. They are not on social media to get information and comments, they are perfect and want to show you that.

But the real power of groups like this is to increase the common good. Putting issues out there for others to learn from and for them to learn from you experiences.

Thank you for posting. It could save others from making mistakes.

 

[jtvt]

I dunno about using loctite. It's more or less liquid plastic that hardens. It would have to be conductive for use on electrical connections. Google says loctite 3888 is a conductor, but not typical red or blue versions.

 

[stobbie]

Please don't use stainless steel scrub screws in the aluminium cells. They will corrode.
Use normal galvanised fasteners

 

[Cdkipp]

Please don't use stainless steel scrub screws in the aluminium cells. They will corrode.
Use normal galvanised fasteners

These metal on metal rules. I am sure there is a chart somewhere I should find.

 

[blutow]

I dunno about using loctite. It's more or less liquid plastic that hardens. It would have to be conductive for use on electrical connections. Google says loctite 3888 is a conductor, but not typical red or blue versions.

The stud is not your electrical connection, the bus bar to terminal mating surface is. Stainless is a poor conductor whether it's got loctite on it or not. These studs and the limited number of aluminum threads is not ideal in a lot of ways. I'd say helicoils is probably the best/strongest solution. Red loctite is another path that I'd argue is significantly better than raw stainless to poor aluminum threads, but opinions vary.