Last fire.. :-(

[Frank in Thailand]

And that is the third possible reason for the pack fault. No proper top balance.

Cells have a charging V, a discharging V and a resting/relaxed V in any SoC.
With proper top balance you can have they charging and resting V equalized so they reach 100% SoC nearly the same time.
Even better if you do a 2-3 step top balance. Put all cells parallel and charge to
- 3,4-3,45V ... it will be a long time. If charging stops (low A) then
- 3,5-3,55V ... it will be much shorter. If charging stops (low A) then
- 3,6-3,65V ... it will be fast.
The reason is that with the 3,4-3,45V charge you do not float the full cells. It is close to the full cells relax V.

All the packs I saw that had a problem, always, every time ... the top balance was not properly done.

I guess you didn't read how I top balanced my cells.
low C rates, yes, it took weeks.
go to 3.5v, what does take a long long time.
then top to 3.65
when all the cells are finished after weeks, you will see that the first cells aren't 3.65v.
they do not stay at 3.65 but reduce till 3.35-3.45 not all the same, slight difference.
charging them to 3.65 with less then 1 A and slowly reducing takes about 30 minutes per cell.
while the SOC might be about the same, the voltage is not.

as you remember, i have cells in use that where supposed to be joined, as result, voltage and SOC needs to be as equal as possible.

while some say you can only top balance before installation, they are singling out the word for just that purpose.
Most balancing done by any BMS is done at the top charge, burning off excess charge so the lower charged calls can catch up.
Some BMS can balance also at low state, equalizing the voltage of the cells. Yes, active balancers.

i like to use the term top balancing when it is balancing at the high range of the voltage, and bottom balancing at the lowest range.
this doesn't matter if it is done by hand or BMS!
what happens is exactly the same, charge or discharge the sells to a known SOC voltage. 2.5v for low, 3.6v for high.
"0"% SOC and "100%"SOC not exactly but close enough as reference point


Do not really understand this part. Not connected cells caused the fire ?
- neither do I.
the Cells on the top layer and the 8 from bottom where not connected. Last 8 had busbars, top 40 zero busbars, = no connection.



Not a good idea. If you store a cell it must be done at ca 60% Soc (it is on the cells manual).
Like on 3,3V (as you received it from the seller as we all did)
The cells don't like to be 100% charged. Even more I do not recommend to charge them over 95% ... ever.
- I don't agree.
as written above, totally good idea. I can not bent space and time.
charging 40 cells from their "storage" 70%, (= temporary not installed, not storage as for a winter or several months / years / transport or sales!!)
takes weeks.
i can not store them at 70% and have them at 100% in a few days.
the time in between, you just referred to as storage ????

you recommend to do top balance at 95%... yet the only real given value is 100% at 3.65
once that is reached, you can use that as reference point.

Not ever using this voltage is absolutely wrong advice.

I'm sure we can all agree lifeop4 is most happy between 10 and 90% or on the edges, 5 and 93%
lead acid is happy at 100% , lifepo4 is not.
I'm around a few days..... and promoted this in many. many posts.
That i show my screenshot where I'm trying to reach 100% SOC as preparation to join the cells who are waiting on a few sunny days, to have the same SOC... is absolutely not a sign that i top charge each day.

would be great if i would.... enough solar for rain season to always have the cells top charged..
i doubt 100 extra will be enough to have tis, as during rainy days, we get 10%.
If i would have them, yes, my overall stop will be at 3.45-3.5 v, 90-95%

So high humidity that can condensate on cold surfaces. A physical container would be a very good idea for the packs Like in marine applications.
- yes, high humidity can do this, henceforth the air-conditioning

So the cell causing the fire was connected.
- Yes it was, for about 8 weeks

You did a lot of unnecessary DIY I would not do
Also I think thick aluminium busbars are better. No worry with the corrosion and the copper-alu reaction.

Yes, aluminum will be better. In the land where stainless steel grub screws need to be imported, aluminum strips where really unobtainable.
it tool me months to find copper bars.
in between time, i tried to make my own..
nice thread about me building a furnace, and melting copper, making sand molds. Mainly because buying wasnt an option.
Aluminum would have been a lot more easy, lower temperature. sadly, not enough scrap alu, but enough copper.

Good


14 kWp solar, nice
3x 156Ah = 486 Ah ... lets be max 0,5C charging (for longer life) that is 243A. With 48V that 11 kW.
OK I do not think all 14kWp solar panels could produce on max so not too fast charging for LFP.
(not exactly sure how many battery strings was already in use)

432Ah was in use, 2 strings, one of 280, one of 152Ah.
while i have +14kw solar, 18 x 325w on the front, not optimal tilted, 9 x 325w on de side, facing evening sun
16x 345w on the greenhouse.
My Sorotec REVO II hybrid inverters are officially rated 5000W.
With my AM Meter i can see 4500W, where the Revo measures 5000, what also better fits with the rest of the numbers, otherwise 500W should go to heat... vanish into air
that brings it down to 9KW on the 2 main sections. the 9x325 start to work around 10.30, before that shade, 3/400 watt.
that max 2925 is never met, but I do get 2750 on a good day.

Thailand, our location have so much dust, cleaning once a week increase sometimes 20-25%
next is the heat what reduces the total production.
while im working on getting the 5000 from the Revo II, i dont loose a lot.
production is 20-30% below t=rated capacity due heat and dust.

OK too high A charge it could not be

I do not know this inverter so I do not know how good it is. Some seems like Voltronic/MPP Solar or clone (donno)
The only thing it is important because how good it keeps the CC-CV charging. Too much charging or to big V spikes hurts battery.
Also no need to charge all up to 3,65. Enough 3,45-3,5V per cell. Already 95% SoC.
- LOL. yeah we talked about this.

REVO II from Sorotec is not a voltronic mass production crap thing.
the unit from EASun was, heck... EASun now OEM the Sorotec Revo II
Growatt did not have a unit that could do parallel at that time, MPP solar seems to sell only Voltronic crap (at that time) ( October/ November 2019)
The support from Sorotec has been really outstanding.
I guess im bad luck magnet for too soon launched products, and doing stress test in the field..
While not good it was needed, they addressed and fixed each issue.
except the 4500-5000watt MPPT


Fourth possible reason for the pack's fault.
You can not thrust a thin blue wrap. It can be damaged, torn, stretched, ... by anything.
Also the bottom of the cells is not covered with it. And the cells body is not neutral (maybe +?).
- agreed. I know this now.
many knowledge gained in the 18 months.
blue wrap better should NOT be done at all, as well as supply screws and rigid busbars.
yet.. they arrive this way, and if you look at Wills videos 1.5 years ago.... He didn't know either.
no one to blame or fault, except perhaps the sellers, providing not enough information, and parts that absolutely should not be used with the cells they sell.


That is good Lower humidity, lower Celsius.


Animals can do a lot of harm. Like little bugs too.
A bug dying between two cells. Charge comes cells bloat a bit smashing the bug. Its shield penetrates the two cells thin blue wrappings ...
Or on top of the cells causing short between terminals.
I would do a complete closed battery box with tipple mosquito nets on all openings ... and maybe ventilators. So nothing could get it

- I prefer my way, buffer zones, each with reduced change that unwanted guests enter.
not really possible to have enough ventilation and no openings for the animals and bugs to enter.
totally seal a room...... thats almost impossible.
especially with Thai(rice farmer) builders.
Real builders work in BKK, and absolutely wont take a job 30-60 minute from the main road..
I wasn't able to find them, after 6 months, we decided to build with local manpower and construction "knowledge".
in short, anything that needed to be strong and was advised to use 15x15 concrete pillar, I choose 20x20.
iron 2 inch, I choose 3 inch better safe then sorry.


35 mm2 from every battery string to the busbar I suppose ?
yes real 35mm2.
not the orange welding cable that is sold as 35mm2m, that's little over 4 mm thick..
Thai have many different ratings, each more flexible then the other..



OK that would be helpful

Thanks I and wish good luck to your new battery pack

Thank you!!

 

[rhino]

I think my key takeaways from this since we'll likely never know the true cause is:

1. Keep your wiring clean and easy to follow and not touching live components. Even without a fire, with this system if there was ever a problem with one of the cells it seems it would be extremely tedious to pull it out and more likely to damage wiring doing it.
2. Have enough physical separation between large sets of cells. There was a second layer of cells on top of the first layer with only 4" or clearance and the fire started after adding that 2nd layer. If #1 is followed this may be less of an issue as long as they are easily accessible.
3. Be extremely cautious in using metal to enclose your cells. There looks like the 2nd layer of cells are sitting on metal bars right above the first layer.
4. With such a large bank using a reliable BMS is even more important. Unless you really know what you are doing, making sure you solder and assemble all of the boards for the diyBMS correctly is a big undertaking. On top of that you have to make sure that JLCpcb where most people have those diyBMS fabricated has assembled the board correctly (they had reversed a surface mount component 180 degrees for some people). On top of that the diyBMS is not guaranteed by anyone, it is use at your own risk. At least with the Batrium that was mentioned there is an entity responsible for it who could technically get in trouble.
5. If you don't feel confident in being able to adhere to the above 4 you are better off purchasing a pre-made battery bank that more likely addresses the above 4 for you.
6. Take steps to assume the battery bank may catch on fire. Don't use flammable material to enclose your battery bank. Separate structure ideal but even if in your living area it can be made fairly safe by using all non-combustible materials anywhere near the bank (especially above it). Have fire alarm and possibly automated extinguisher (extinguishing ball above bank) and hand held fire extinguisher nearby.

 

[bdserv]

Maybe this was mentioned already... there is threaded rod that passes between the cells for clamping. I can't tell from the pictures, but if the threaded rod is unprotected where it passes between the cells the sharp edges of the threads may have cut through the thin insulating covering on a couple of cells, effectively causing a short between cell terminals [case is common to either pos or neg terminal depending on mfgr].

I appreciate the willingness of the OP to post information about this experience which we all can learn from. It has certainly caused me to reflect on my plans for a DIY pack for my golf cart.

 

[Frank in Thailand]

The threaded rods have "sharp" edges, like any threaded stud has.

The BMS wires where protected with tape on the hight if the rods.

I did see however some blue plastic slightly damaged and show aluminium.
None of it what caused any damages in its own.

People longer on this forum might remember the "leaking current".
Or more technical term, Electro magnetic field.

That's a dangerous word I noticed for the states, and related to people wearing tin hats...

For those who aren't affected by that fear, you should be aware how closely electricity and magnetism are related.
Your inverter does it all the time, convert to magnetism and from that to electricity!

Anyway..
High current cables do "leak" as bare Bus-bars, charging things around them, up to high voltages!
Not a lot of capacity, a spark (or bite) and it's gone.

Enough to intervene with my computer (mining rig, placed above inverters) and charge the threaded rods.

None of it is a possible source for fire on its own.
Tiny sparks are nonrmal with electricity.
Look at the old doorbell relay...
Sparks nicely in the dark.

It does add some instability in any setup.

Bus-bars best are flexible, and coated.

Threaded rods are best coated, insulated to prevent absorbing the electricity around them.

Cells should have separator sheets (non conductive)
And between the rods, protection sheets are a really good idea.
(So the blue wrap is protected from 4 sides.)
Actually...
Not all cells have the bottom sealed...
Best there also.

Yes DIY BMS takes months to build.
Especially if you forget a 0402 (0204?) thermistor to be soldered.

With many hours of patience soldering iron, electronic microscope and toothpick I was able to do 2 or 3 a day...
(And had 30 to do)
Not every day, so that was 2.5 months...

It did improve my soldering skills, and for sure, I whish I had known about soldering in the oven on 140c...
That would have made it so much more easy.
Solder paste, ...

They are finished now, of the 60 boards of 4.2, 45 are functional.
My soldering skills took 15 as casualties..
The new 4.4, all 50 are Ok, thanks to the help of my family and oven baked at 140c

Would be ashame not to use them.

Besides...
Jikong isn't known yet for years stability.
DIYBMS is!

For Batrium..
Nope, $1000 per S16, is beyond my budget and wishes.
Some one who can afford Victron or outback, probably would not think twice about that cost.

For me paying about the same price for BMS as the cells... Not really an option.

Insurance..
Thailand?
Lol.
Well... Atleast here you know for sure they only collect and never pay out.
No insurance possible, what is totally unlike me, coming from Netherlands.
We have normally almost always everything insured...

Only mandatory insurance for vehicles and hospitalisation.
That's it.
The rest is here only a piece of paper not worth th time and effort placed in it to make it as it's totally worthless.

My fire protection worked, just as well as I could expect.

The fire was too small to start the automatic fire distinguisher, not yet hot enough.

All parts to stop all contact (from solar and to inverter) worked at they should and fire delay walls did not burn at all, just blackened.

Now the hand held distinguisher is there, I suppose better then the sand for most fires.

For this one, self discharge of (432Ah, 15% charged,) about 3.5 kWh of energy from 32 cells..
+8 fully charged cells, +/- 4 kWh,
Sand was probably the only thing what would keep it safe during its self discharging.

Fire distinguisher would been needed for hours and hours as I doubt it would be possible to cool enough to stop the cascade effect from the heat.

Anyway...
Done is done

No real harm.

Mostly due smoke detector and fire delay measures.

 

[Sanwizard]

Last week we had a nasty fire.

I worked for months, used only high quality materials, tinned silicone wire that self fuses at 10A, and weeks between each step of construction to be absolutely sure things are safe

The lower layer was good.
Might look a mess with the wires, but absolutely safe, no terminal got more warm than any other, even with +150A charge.
View attachment 57365

How I know for sure it's safe?
Worked like that for over 6 weeks.

Week before started to build the next layer, placing the metal rack, plywood sheets and the repaired tread cells.

All cells where placed and all BMS wires (the pile spaghetti on the picture) was nicely between the cells to the top.
No crushed wires, no damage, all good.
View attachment 57366View attachment 57367View attachment 57368


Top layer received it's compression and that's it.
For days and days.

What happened?
I don't know.
Few days earlier one of my MPPT did not charge, so just one to charge 16* 280 and 16x 152Ah.
That's all that was connected on the lower base.

It went empty that night, my inverter stops at 48v.
Around 07.30 I looked and all was OK, not yet charge, rain outside.
We snugged back to bed
Half hour later lights went on again, 49v, restart voltage of the inverter.
08.30-09.00... beep beep beep, smoke detector allarm!

What happened... ??
Top layer of cells where placed like 60 hours before, and no bus-bar yet.

Previous day's we were in the town to visit lung doctors for my COPD, to time or energy to do anything new with the setup.

No high capacity charge, few hundred watts.
This is something that didn't move or got changed in days.

Perhaps a mouse?
Almost can't be.
Even if it did chew the BMS wires, it should self fuse. AWG22

After investigation of the scorching and location most likely is a instant selfdischarge of one of my 152Ah EVE cells.

Those where a trouble from the start. Who is interested.. enough old posts about them.
Use seach..
FIY, the same pouch case is used for 120 and 135Ah cells.
There is a limit hou many sheets you can squeeze in, without getting into trouble.
152Ah was that limit.

Back then, start 2020, the 152's where the best bang for the buck, 280 came just looking and where high priced, compression as we now know is mandatory, wasn't even discussed.

Sure, taping together for ease of handling.. that ain't compression

And they bloated, quickly.
Little over a mm.
It did not seem to hurt the functionality. After a year of useage I tested it's capacity, overall 147, top 151, low 144Ah (that's 52 cells)

I did compress them afterwards, slightly.
Strong enough to have 6 cells together and not being able to lift one out from the middle, not so strong that the bloating disappeared.

I did notice that after a month or so, I could easily compress a little bit more, 1/4 threaded rods, tighten with pipe wrench by hand, just the round wrench, nothing extra needed.

That was 8 weeks ago or so since I tightened it for the last time.
And no need, busbars where good, with +4nm torque each on M6, +5Nm on M8
All worked tip top since.
Yes, only 432 of my total 1016Ah, but stable.

View attachment 57372

View attachment 57373
Bloated??
View attachment 57374
Casing fused together.
View attachment 57375
Now they are bloated..

View attachment 57376
And the cell that most likely started the fire.

Instant selfdischarge, at +/- 15% charge
Creating enough heat to infect it's neighbours, who also selfdischarge, chain reaction.

The flames most likely are the Ox-guard, it's grease, and that self combust at 350-450c
Perhaps a light glow of self fusing wire was enough to ignite the grease.
Who will be able to tell.

Flames where calm, heat immense.
I worked at a bakery for several years and know the heat of 275c oven.
This was a lot more.

All lower level cells selfdischarge, bloating like a real pillow.
Top level, all precharged to 100% ...
Some bloated more, most a little or none at all.
They did get really hot feet.
I no longer thrust them for Safety reasons.

I reached out for a loan to my family in Netherlands, and now am the proud owner of 6 X 24v BYD packs with 3* Jikong 2A active balancer BMS.
In other words 3 X S16, 51.2v 260Ah, total 780Ah about 40 kWh.

Funny how life works.
That was my first BMS when all the troubles started.
I was a happy puppy the whole 3 days the Jikong worked.
Total pre-mature launch, where almost all units of first shipment got returned to the seller in a few weeks....
Now it's a grown up product.

View attachment 57378
I'll post some pictures later

What a development there has been in the last 2 years.

This wasn't available, if it was I absolutely would have started with this
Price?
41.500 THB, 1000 THB transport ($ 1.260 + 30)
Under 1300 for 260Ah 51.2v)
Including BMS.

All tax paid, delivery at my doorstep.

Roughly 60kg per pack.
That's 15-20 kg on structural integrity, safety, quality.
280Ah is 5.5kg, *8 is 44kg.
And it shows.

People... What the F are we still doing buying EVE or Lishen and try to do what robotics laser welding does?

Not anymore for the price!!
It's even cheaper then the. 200Ah deep cycle lead Acid we started with!!
(11.000thb / $335 for 12v, 48v 44.000 or $1360)

I'll post a separate thread about the BYD soon.

Take care and stay safe.
My experience is now that any compromised cell is a potential bomb.
Even with 15% charge, results are catastrophic.
In this case only for the cells, due my smoke detector alarm, concrete sheets around the battery and the help of my family.

Releasing the compression, last center nut did not twist smoothly.
A tiny bit of force and it snapped...
The 1/4 threaded rod.
That's how much force there can be on the compression, and even with this, just a few cells the vent bursted...

Know what you are playing with!!

Special thanks to @Just John , sorry the epoxies are used for demonstration, I didn't get the chance to test it with Bus-bars connected...
And won't be using the cells now..
Even with almost perfect repair ?

I noticed you did not have an insulator between the cells. I believe there is a potential for a short to the alumimum casing without protection. Just posing a possible reason.

 

[Mcgivor]

Getting a headache reading this, I'm less than 100Km. from the OP with 2 systems in the same conditions except my banks are not climate controlled, temperature is not a major concern, the cell temperatures are generally 10°C below maximum, unless >40°C persists for a number of weeks, which is rare and usually occur during April and May. Same incects, rodents, reptiles, spiders and snakes

There are many things I could disagree with, along with many comments I could make that may be perceived as offensive, so I restrain myself, however I have the same challenges without the drama.

My background is in electrical/electronic control, my training was the best one could get, my focus is on neatness in wiring practice, something I'm not seeing here, I've passed this philosophy on to to others throughout my career, training others who are grateful for the guidance provided.

There is a simple solution, but sometimes stubbornness and self preservation get in the way of persuasion, I'm willing to offer my services for free, if you are willing to accept @fhorst your move.

 

[rhino]

For Batrium..
Nope, $1000 per S16, is beyond my budget and wishes.
Some one who can afford Victron or outback, probably would not think twice about that cost.

Just to clarify this, a WatchMon CORE with 4 CellMate K9s which can monitor 64 cells plus a 1000A "ShuntMon" runs $1221 USD which is about $19 USD/cell.

 

[bdserv]

The threaded rods have "sharp" edges, like any threaded stud has.

The BMS wires where protected with tape on the hight if the rods.

I did see however some blue plastic slightly damaged and show aluminium.
None of it what caused any damages in its own.

Let's say your cells have case common with positive, as mine do.

Now consider this example: cell #1 has negative terminal connected to cell #2 positive terminal. This means that an electrical connection between the case of cell #1 and the case of cell #2 is the same as a short across the terminals of cell #1. Something like this is certainly possible if you have the cell cases touching the sharp edges of the threaded rod. That is why it is a good idea to protect the threaded rod with tubing such as thin wall aluminum or better yet, a non-conductor such as hard plastic.

 

[Supervstech]

Getting a headache reading this, I'm less than 100Km. from the OP with 2 systems in the same conditions except my banks are not climate controlled, temperature is not a major concern, the cell temperatures are generally 10°C below maximum, unless >40°C persists for a number of weeks, which is rare and usually occur during April and May. Same incects, rodents, reptiles, spiders and snakes

There are many things I could disagree with, along with many comments I could make that may be perceived as offensive, so I restrain myself, however I have the same challenges without the drama.

My background is in electrical/electronic control, my training was the best one could get, my focus is on neatness in wiring practice, something I'm not seeing here, I've passed this philosophy on to to others throughout my career, training others who are grateful for the guidance provided.

There is a simple solution, but sometimes stubbornness and self preservation get in the way of persuasion, I'm willing to offer my services for free, if you are willing to accept @fhorst your move.

Wow, how kind an offer! I would LOVE to see some of your work on your build.

 

[upnorthandpersonal]

@Supervstech He has shown his set-up here:

Last fire.. :-(

Those BYD modules look nice. I've just seen them being advertised by FPV Power in australia via a facebook page. Price is about double what it would cost to build DIY from bare cells, which I think isn't bad. The bloke selling them says they are taken from cars which have been test driven or...

diysolarforum.com

 

[Mcgivor]

Wow, how kind an offer! I would LOVE to see some of your work on your build.

There is a picture on page 5 of this thread, taken after the switch over from LA to LFP, so basically added some extra components to my existing install, BMS etcetera. Will at some stage redo the whole installed with a fresh layout to address things Im not happy with.

The years have taken their toll, a constant battle with spider webs, birds, wasps rats and so forth, the redo will address these issues by installing in an enclosure with stainless steel mosquito mesh ventilation. In retrospect I should have gutted the whole installation because starting from scratch allows planning as opposed to adaptation.

 

[robby]

This is why this stuff will never be in my house and has to always be contained against fire.
If it can happen on a multi million dollar project it most certainly can happen with any home system.

This just happened 36 hours ago!

 

[robby]

Robbie,

I could not agree with you more.

Few things to consider in my situation:
Theft.
On our land we have like 30-35 dogs.
Slamdog breed.. so only bark, no bite.
Like all average Thai dogs.
We had a guy stealing a lead acid battery, 70kg was too much to handle for him.
That went without any real damage.
Cut the wires with an ax.
That was damage.

Theft is a real thing here.

Besides this, I need to be able to keep an eye on things, as it's not always working as one would like.

My first hybrid inverter (EASun) charging 78v went unnoticed for about 14 days, then the 16* 200Ah where about dead.

No equipment in Thailand is thrusworthy.
People living here will agree on this.
People who don't live here might be puzzled.
We can only buy the junk that doesn't pass QA testing for Europe or USA.
And I wish I was joking.

So stuff fails, a lot.
Everything.
Low quality, high quality needs to be imported from USA or Europe.

I for sure wish I could have it outside, where it's easy accessible, naturally cooled to under 30c
Or special airconditioning just for that room...
(Would require 6-8 additional solar panels, more AH and addition point of failure what isn't seen 24/7)

I wish, but can't.
It would not be safe.

Other probelm..
If one does break in, and get fried because the guy/girl have only a screwdriver and a hamer...
I will pay for his/her family loss.
In Thailand no forigner get justice towards a Thai national.
They always are right.

I would need to build a fortress, 24/7 Thai guard outside, alarms all over and a warming system on each and every item that might fail, and become dangerous...

You don't want to know how official Thai airconditioning installers work..
No wonder many fail.
Brazing without CO2, and then wondering how the compressor stop in a few weeks...
(Perhaps black powder from brazing inside the AC line ???)

That's a professional installation company who does it for a large shop (HomeHub) in Udon Thani..
HomeHub is all over Thailand, like Walmart in USA.

I'm not complaining!!
Just explain how thinks might seem easy to do, and probably in Europe/USA it is.

$5500 LiFePO4, $2500 on inverter, 43 solar panels..
All adds up.
(Actually that's just bare purchase price, about $5500, add an other grand for tools and equipment, wires, BMS, etc etc etc )

For people who get $500-600 for working 60 hours a week...

That's a tempting jackpot out in the middle of nowhere...

Theft is a issue for me also but not nearly to the level that you describe. I would imagine that labor is cheap over there and building a small room out of reinforced concrete should not be all that expensive??
I would certainly do that and then have proper burglar bars on the door, small vent opening for the exhaust fan and intake air. It's may cost a lot more money than you want to spend but I just cannot imagine sleeping soundly at night knowing a potential fire hazard is in the house with me and my family.

 

[Frank in Thailand]

Getting a headache reading this, I'm less than 100Km. from the OP with 2 systems in the same conditions except my banks are not climate controlled, temperature is not a major concern, the cell temperatures are generally 10°C below maximum, unless >40°C persists for a number of weeks, which is rare and usually occur during April and May. Same incects, rodents, reptiles, spiders and snakes

There are many things I could disagree with, along with many comments I could make that may be perceived as offensive, so I restrain myself, however I have the same challenges without the drama.

My background is in electrical/electronic control, my training was the best one could get, my focus is on neatness in wiring practice, something I'm not seeing here, I've passed this philosophy on to to others throughout my career, training others who are grateful for the guidance provided.

There is a simple solution, but sometimes stubbornness and self preservation get in the way of persuasion, I'm willing to offer my services for free, if you are willing to accept @fhorst your move.

I gratefully accepted.

Where +40 is indeed just for a few months, and during those months just 4 or 5 hours a day between 44 and 47 in de shade, with slight windy breese.

Perhaps I misunderstand the Temperature issues, that charging at -10 isn't a problem if it's just a few hours a day, a few months a year
Other way around, the +40..

What I understood is that it is a probelm, especially as during charge, temperature will rise and if not controlled, it goes up to dangerous temperatures

I'm not an electrician, and not claim to be, just trying to do my best here, with limited knowledge and materials.

@Mcgivor , you are absolutely welcome to our place.
Always willing to learn.

I can't bend space and time, my resources and skills are limited.
Resources not that much financial, but obtainable is the big word in Thailand

I'll send you PM.

 

[Frank in Thailand]

Just to clarify this, a WatchMon CORE with 4 CellMate K9s which can monitor 64 cells plus a 1000A "ShuntMon" runs $1221 USD which is about $19 USD/cell.

I'm not sure where you are getting your information.

One cell module is/was $45
Then the controler unit is needed, and the contactors who have a price tag of $200 for 400A.
Shunt $ 100,
Additional WiFi, and toys, made a 16 cell setup about 900-1250, depending on modules chosen.
$900 USA, bare setup, no shunt or toys at all..

Only cell modules, 45 each,
$2900 + the $1250 starter set..
And..
Oeps..
More then 64..
So 2 start sets, 1250-900 + 48x $45
$4300,- for a complete set for 80 cells.
Pricing of April last year

Now I'm sure they follow the development and improve their products.
Perhaps no longer one module per cell needed, it they greatly reduced their module prices.

Bit that was the price (by head) in April last year.
I did not yet look what one cell module cost today

Edit:
Just did:
Price 400 Australia Dollar about $300 USA

Less then $20 per cell module.
That's a nice reduction in price, and more "realistic" to spend.

A year ago..
Not so much.
$45 per unit when you bought 16..

 

[curiouscarbon]

This is why this stuff will never be in my house and has to always be contained against fire.
If it can happen on a multi million dollar project it most certainly can happen with any home system.

This just happened 36 hours ago!

450MWh capacity!! (before fire)

UPDATE: They just let the fire burn itself out and cooled down the surrounding units to help stop it spreading. Nearly 24hrs later the same megapack is still alight, but it appears as though the half that ignited the fire is now burnt out, but the other half of the megapack is now alight. Doesn't appear to have spread to any other megapacks other than the outer casing heat damage shown in the video. The fire department have doen a great job containing it.

The batteries are exothermic. I doubt a fire surpression system is going to contain it. This is why the fire department simply let it burn out.

Some gems from the comments on that video :

How about LN2 deluge systems for each bank? Cryo temps and O2 displacement to address exothermic head generation and fire suppression in one. Match the LN2 tanks to hold an amount to negate all of the heat from the exothermic reaction.

Surely a liquid N2 (similar to the old halon) dump would be viable? Thinking if you cool and starve the fire it could be extinguished? Dunno, just thinking.

Lithium react with Nitrogen, Carbon Dioxide and Water so it doesn't leave many common materials available for fire suppression ??‍

Seal them completely, no air = no fire, get rid of the external to internal fans, use internal fans only and use water cooling

Problem here is that they need to make money. Deluge systems for each of these packs might set them back a good deal. I think the cheapest way would be to place them further apart. It'll still burn, but it won't spread as much.

Shame on Dave for the photo of him pointing and grinning at the fire.

Did anyone get hurt? I thought it was just a huge company losing money and time.

Gee, I just thought it was common decency to not laugh at a fire.

Down a rabbit hole I went. There seems to be a halon substitute that is non ozone depleting tested to just as effective as halon. The only problem with such a system is the limited quantity and high price. For something large like this I'm guessing water is still going to be the main direction.

people are capable of "go slow and safe" (which some disasters still happen), but dangle huge money in front of them and they’ll stare at it and trip over every rock on the way up the hill.. hope no injuries resulted from that fire. i know i'm very eager to operate DIY LFP ESS, but also i must go slow enough to understand many things! what a worthy challenge.

i still fear and respect LFP chemistry, but recognize that it’s a fundamentally less volatile than eg NCA chemistry.

my general plan still involves a ≥2mm wall thickness airtight stainless steel box, walls thermally insulated, *no air exchange with outside*, *only thermal transfer*. moisture capture and removal system internal. active thermal regulation using peltier to LFP's optimal lowest degradation charge temperature. exchange heat with water/glycol coolant going through silicone tubing and in/out of aluminum heat blocks and aluminum radiators. it's more cost added, but i am trying to amortize it over multiple packs. pump heat. overpressure relief valve too i guess.....

still working on the pack thermal interface design. don't need fast heat conduction. trying to minimize materials for cost.

maybe i’ll do what another user did and what @robby suggests and hire a mason to build a specific concrete outdoor shed.. a worthy single time construction cost... that beats garage. i prefer to sleep far enough away that if the alarm system fails, i wake up alive... and with no lung damage from sleeping and inhaling fumes >_<

 

[A.Justice]

3.2v, maximal 1.5A.. 5 watt.
Discharge from 3.65 to 3.55 of 280 or 152ah battery.
That's 5 watt for a minute?
On a heat sink...

You keep mentioning "only 5 watts" like that's a small amount of power.

I have a vape that heats up a coil red-hot with 5-10 watts. ANY watts can generate a tremendous amount of heat.

So being "realistic" as you say, 5 watts is more than capable of starting a fire. Not saying that that was the cause, but you can't rule it out.

 

[Frank in Thailand]

I can rule it out as the device that would be possible to draw this 5 watt was not connected.

Like your vape, or any other Device won't draw power when not connected, it can't be a source of fire.

Sure Wolfram wire in a small light bulb glow hot! With 5 watt.

Awg20 wire...
Rated for 10A...
Not so.
Or better, not at all.

If one would take one strand of the tinned copper wire, then yes, that might get hot.

But.. it wasn't one strand, and it wasn't connected to the DIYBMS yet.

That was theoretical story where some forum members find it dangerous to clamp the BMS lug between 2 stainless steel nuts, as it's high resistance would give Total unable readings and using the BMS at maximum (5 watt) would absolutely make a fire.

According to some Doom thinkers.
All I did was explained to them that they aren't realistic and searching for problems that don't exist.

Stainless steel nuts will have voltage drop, agreed.
Like 0.01v, less then Mosfet based BMS.
That voltage drop generate heat, with the 1.5A.
About ??? Joules?

My math isn't that great at this point.

Perhaps 0.1 degree on stainless steel stud, that is contacted to the aluminium terminals and the thick copper bus-bars.
Probably less.

So far for fire hazard by using awg20 for BMS wire, drawing 5 watt, with the lug clamped between 2 stainless steel nuts.

If your vape would have the same mass as this setup, it won't get warm at all.

On addition, even it it does increase 1 or 2 degrees (what I seriously doubt) that won't be enough to start a fire as the whole set Average temperature is around 30-35c during usage.

So no.
We can't be absolutely sure about anything.
For decades people laughed about Einstein that there is light at the end of a black hole...
Turns out he was right. Atleast according to the latest scientific study.

While it can't be absolutely ruled out, the realistic chance that it would be a possibility in the future when the DIYBMS would be attached, is really small.
With the fact that it wasn't attached yet, it reduces to almost absolute zero chance.

Mass of the copper busbar, 65 gram.
Stainless steel rod and nuts 17 gram.

Fraction of 5 watt to heat this up.
For a relative short time as it's just burning off the excess charge, something a Daly BMS does with 0.08A.
While 1.5A is a lot, it also needs a lot less time to do it.

And...
With cells that where 15-20% charged..
That it wasn't balancing is probably the last tiny percentage chance that it was the cause of the fire.

Not sure.
But no one would believe the "proof" that it can be responsible.
With the fact that it was not installed and even if, it does not balance at 15-20% SOC..

Nothing is sure.
That it the only thing that is sure.

 

[Supervstech]

I can rule it out as the device that would be possible to draw this 5 watt was not connected.

Like your vape, or any other Device won't draw power when not connected, it can't be a source of fire.

Sure Wolfram wire in a small light bulb glow hot! With 5 watt.

Awg20 wire...
Rated for 10A...
Not so.
Or better, not at all.

If one would take one strand of the tinned copper wire, then yes, that might get hot.

But.. it wasn't one strand, and it wasn't connected to the DIYBMS yet.

That was theoretical story where some forum members find it dangerous to clamp the BMS lug between 2 stainless steel nuts, as it's high resistance would give Total unable readings and using the BMS at maximum (5 watt) would absolutely make a fire.

According to some Doom thinkers.
All I did was explained to them that they aren't realistic and searching for problems that don't exist.

Stainless steel nuts will have voltage drop, agreed.
Like 0.01v, less then Mosfet based BMS.
That voltage drop generate heat, with the 1.5A.
About ??? Joules?

My math isn't that great at this point.

Perhaps 0.1 degree on stainless steel stud, that is contacted to the aluminium terminals and the thick copper bus-bars.
Probably less.

So far for fire hazard by using awg20 for BMS wire, drawing 5 watt, with the lug clamped between 2 stainless steel nuts.

If your vape would have the same mass as this setup, it won't get warm at all.

On addition, even it it does increase 1 or 2 degrees (what I seriously doubt) that won't be enough to start a fire as the whole set Average temperature is around 30-35c during usage.

So no.
We can't be absolutely sure about anything.
For decades people laughed about Einstein that there is light at the end of a black hole...
Turns out he was right. Atleast according to the latest scientific study.

While it can't be absolutely ruled out, the realistic chance that it would be a possibility in the future when the DIYBMS would be attached, is really small.
With the fact that it wasn't attached yet, it reduces to almost absolute zero chance.

Mass of the copper busbar, 65 gram.
Stainless steel rod and nuts 17 gram.

Fraction of 5 watt to heat this up.
For a relative short time as it's just burning off the excess charge, something a Daly BMS does with 0.08A.
While 1.5A is a lot, it also needs a lot less time to do it.

And...
With cells that where 15-20% charged..
That it wasn't balancing is probably the last tiny percentage chance that it was the cause of the fire.

Not sure.
But no one would believe the "proof" that it can be responsible.
With the fact that it was not installed and even if, it does not balance at 15-20% SOC..

Nothing is sure.
That it the only thing that is sure.

ok.
the small wires were unlikely to be the fire.

the batteries themselves somehow got hot.

A single cell, or several… we do not know.

Please stop stating “x” couldn’t not have started the fire. You are GLARINGLY GLOSSING OVER a combination of problems… sure, the lead placement wouldn’t have generated the heat needed for a fire. And the wiring crisscrossed over and over other wires wouldn’t have caused the fire, and the BMS not being hooked up wouldn’t have caused the fire. And multiple more things we keep bringing in from the few pictures we have wouldn’t have caused the fire… by itself.
But a combination of micro faults and small issues could have.
How? We don’t know… because each time anything at all is discussed, you write up huge essays explaining why nothing discussed could cause the fire.

Do you not see that overlapping mistakes, and improper everything can accumulate, and result in a fire? Not ONE thing started the fire. Obviously, the cell or cells ruptured… the fire was from a cell… that is plain and obvious…

WHY did the cell get hot?

 

[DerpsyDoodler]

ok.
the small wires were unlikely to be the fire.

the batteries themselves somehow got hot.

A single cell, or several… we do not know.

Please stop stating “x” couldn’t not have started the fire. You are GLARINGLY GLOSSING OVER a combination of problems… sure, the lead placement wouldn’t have generated the heat needed for a fire. And the wiring crisscrossed over and over other wires wouldn’t have caused the fire, and the BMS not being hooked up wouldn’t have caused the fire. And multiple more things we keep bringing in from the few pictures we have wouldn’t have caused the fire… by itself.
But a combination of micro faults and small issues could have.
How? We don’t know… because each time anything at all is discussed, you write up huge essays explaining why nothing discussed could cause the fire.

Do you not see that overlapping mistakes, and improper everything can accumulate, and result in a fire? Not ONE thing started the fire. Obviously, the cell or cells ruptured… the fire was from a cell… that is plain and obvious…

WHY did the cell get hot?

This is exactly what I said near the beginning of this thread.

 

[Supervstech]

This is exactly what I said near the beginning of this thread.

Yes. I have said similarly several times...

 

[mrzed001]

Yes. I have said similarly several times...

At least we did the investigation/audit and found at least 6 errors that could be the cause of the fire.
A good lesson for everyone.
Especially since there is the 4th cell fire case ...

 

[Frank in Thailand]

Why the cell got hot?
my best guess is that it got damaged along the way and decided to give up on that moment.
it could have done it days of weeks before, or days or weeks later.

many micro mistakes do add up.
I totally agree to that.

there was not a mistake in 4 wires per cell, or having them laying crisscross during installation. It only shows how safe the wires where, as they could layb like that without making a problem. the picture with spaghetti was weeks before the fire, and not like that when the fire started.
i made a special photo shoot on how the wires per cell where before the fire.

first mistake was buying 152Ah cells that have the same format/size as 120AH cells.
too many sheets in a housing.
second mistake was using the sellers screws and rigid busbars
third mistake was not using torque wrench
forth mistake not clamping/ compression.

while i write the word mistakes, its a way of installing many of us have done, and usually goes fine.
152AH cells started to bloat.... and having trouble staying balanced.

one could call it a mistake to keep using them.
Without reference experience, this was how it was supposed to be. new cells, good brand...
besides.... dumping $3000-3500..... just because its not stable at top voltage and not flat anymore?
I don't have that kind of financial space.

what I do see now as a mistake ( a real one) is trying to prevent the bloating from expanding, and maybe even reduce it.
(with light compression, hand tight a pipe wrench on 1/4 inch, after a few weeks it was getting "loose" so i could tighten a bit further)

a few days before, I've charged the cells to 3,.65 range, and they probably will have increased the expanding force.
off grid, = depending on the sunshine. sadly a few clouded days, so it got discharged to about 15%
that was probably the last microdamage that the cell ( or cells) needed to self discharge.

I don't had any way to prevent threads from failing.
many of our forum members have. and using the screws provided... not good.

If I had known better, i would have done better.

And that is for me the most important reason for this thread: Educational.
making awareness that the threads are crap weak, the screws should not be used, same as the busbars (flexible only) and cells should be clamped/compressed already with the first top charge / top balance action before placing the pack together.

I don't know, maybe I should not have shared the picture of the spaghetti. it made some people confused.
I should not have shared the picture of charging till 3.65, as people think i'm having my cells in storage while I'm building the pack (??)
I should have started with the photo shoot on how 4 wires where used for 1 cell, and while for some that might feel dangerous, as the pictures shown, its not. perhaps could have used a cable sleeve for additional neatness... I didn't have 30 meters.

where I thought for a long time slightly bloated cells are safe to use, my story tells different.
Will was/is right, they have only one direction, trash bin., or used as dangerous goods that will self-discharge at some point. metal box, or outside concrete shed... not a problem.
They are no longer safe.

please do remember, you have more lithium batteries around you that you might think of. and each and everyone of them is potential dangerous.

and with this last video, I will put, just like most people my phone on the charger, at by bedside for the night..
sleep without thinking about the possible danger I just put myself into.

 

[mrzed001]

first mistake was buying 152Ah cells that have the same format/size as 120AH cells.
too many sheets in a housing.
second mistake was using the sellers screws and rigid busbars
third mistake was not using torque wrench
forth mistake not clamping/ compression.

5th: no isolation between the cells
6th: no closed box (bug prevention)
7th: no BMS
8th: no proper top balance

If I had known better, i would have done better.
And that is for me the most important reason for this thread: Educational.

This is why this forum exist. You can ask anytime.
I do not understand people sometime.
What I do first is planning. Plan every little part.
Second I share my plans with more experienced people. Asking they opinion.
I correct my plans and only after that I start to build anything.

please do remember, you have more lithium batteries around you that you might think of. and each and everyone of them is potential dangerous.

and with this last video, I will put, just like most people my phone on the charger, at by bedside for the night..
sleep without thinking about the possible danger I just put myself into.

You are wrong. That is an NMC/Li-ion cell. EV-s have all Li-ion battery packs (except the new Chinese Tesla Model 3 that has CATL LiFePO4 prismatic cells)
This is a real test from LiFePO4 / LiPo / LTO cells:

 

[A.Justice]

This is a real test from LiFePO4 / LiPo / LTO cells:

I love that video, and just wanted to add that if you do put multiple holes in a prismatic LiFePO4 cell, they most definitely can catch fire, although not nearly as seriously as other types of Li cells. They don't explode or jet flames everywhere.

It was either @HighTechLab or a YouTube recommendation from his channel, but somebody put two holes in a cased LiFePO4 cell and it did ignite.