LTO battery fire

[MisterSandals]

And more to the point why do these batteries hurt you so much when you drop them on your foot LOL

I recall someone once saying that "stupid is supposed to hurt".

 

[Checkthisout]

I recall someone once saying that "stupid is supposed to hurt".

This best explains why I'm in constant pain. ?

 

[Kenny_]

I recall someone once saying that "stupid is supposed to hurt".

If it did we wouldn't have the problems in the world we currently have.

 

[summit]

30-75% based on your capacity tests?

no, rough bulk voltage between 51.5v - 53.1v

 

[Dagoth Ur Does Solar]

Wow this thread was a doozy.

Can someone smarter than a smoothbrain like me tell me: those cell terminals and conductors look really damn close to that metal rack he has it sitting on... Would you guys build your packs like that? Serious question. I feel like if I was doing this I'd want some serious insulation. I feel like if I leaned into inspect that thing my jacket zipper might short out and kill me.

Also I'm not going to even comment on using alligator clips, what the hell, man.

 

[Tom5051]

there are so many things wrong with this battery setup that it is not surprising that it started a fire.
it is a lethal hazard with exposed bus bars , which are also close to the non insulated metal rack, relying on alligator clips for balancing.
And the OP claims the be an expert.
The most dangerous person of all is someone who does this sort of dodgy work and then tries to help others with their "knowledge"

 

[Made in North Pole]

If you have a very large bank, just run 2 CORE's and 2 shunt trip breakers. Only lose one half the bank. Takes the same number of K9's, just an additional CORE and shunt trip. I almost set up my bank that way but decided I would add another CORE, K9's and shunt trip if I add to the bank.

I've been extremely pleased with my decision to run Batrium. It isn't cheap but is certainly flexible.

You could buy a Batrium Expansion Board instead. You can run multiple shunt trip breaker on one Watchmon Core with expansion board.

 

[sunshine_eggo]

I've discovered that my active balancers make my battery worse than when it's left on its own with an occasional passive balance via Batrium.

They are individual units, and I'm pretty sure their voltage measurement error forces a 0.04V drift in my cell voltages. They'll seek that dV after about a week following a Batrium 0mV balance. When I leave them off, it takes about a month before my Batrium needs to be fired up for a few Ah of burning. Batrium cell voltages are within ±.005V of my Fluke. I have no idea what the balancers are measuring.

@ConnerLabs Is there something I can do to help you be less unsure? It's pretty simple. With the active balancers installed, my cell voltages diverge more quickly than with them on. The only reason I can see for this is the voltages the units are measuring. With the Batrium, I can confirm BMS readings against my Fluke. With the balancers, I have no idea what they're measuring.

It's nearly 9 months later, and my balance continues to be improved over what it was with the active balancers on.

 

[ConnerLabs]

@ConnerLabs Is there something I can do to help you be less unsure?

The unsure face was more about whether balancers do more harm than good in general, than about doubting the quality of your experimental results. A balancer can save a cell, but also destroy one by malfunctioning. As the old saying goes, if it isn't there, it can't go wrong.

 

[sunshine_eggo]

The unsure face was more about whether balancers do more harm than good in general, than about doubting the quality of your experimental results. A balancer can save a cell, but also destroy one by malfunctioning. As the old saying goes, if it isn't there, it can't go wrong.

Gotcha. Sadly, that's how it is with all this stuff... Even a BMS can malfunction (welded FETs) damaging cells. DIY batteries can be a tightrope walk.

 

[Kenny_]

DIY batteries can be a tightrope walk.

My thought exactly....

 

[Lt.Dan]

Even a BMS can malfunction (welded FETs) damaging cells. DIY batteries can be a tightrope walk.

This is also the case with pre-made batteries too though. All FET based BMS/batteries are subject to the same flaws.

On a slight tangent, this is a big reason why I (and I'm sure others) think it's important to have multiple batteries in a bank, instead of one large battery. Not only do you have redundancy with multiple BMS's, but you are also splitting the load to all of them, ensuring a longer life for the batteries/FETs

 

[SeaGal]

All FET based BMS/batteries are subject to the same flaws.

I would have thought there'd be more chance of a failure with an active vs. passive balancing BMS in the case failure, but haven't got my head round to working out why I have that gut feeling.

 

[ConnerLabs]

I would have thought there'd be more chance of a failure with an active vs. passive balancing BMS in the case failure, but haven't got my head round to working out why I have that gut feeling.

The more I think about this the more I like the contactor or shunt trip breaker style of BMS. I wish they sold the Midnite Solar 250A shunt trip DC breaker in the UK, it seems like about the best safety shutoff you could imagine for a lithium battery.

Both styles of balancer can stick in the on position and flatten a cell. I guess the active ones have more current at their disposal to do it quicker. I have to wonder how much balancing current a pack of good cells that were balanced to start with actually needs?

 

[OffGridInTheCity]

I use Batrium + longmons (6+ years and counting) to monitor the 126 x 18650 packs @ 260ah each. Over the years I've discovered that balancing isn't needed day to day for healthy packs (or cells).


I do touch up balancing every 9 months (or so) as a 40-60mv max difference between packs will drift to 50-100mv max difference over those months. I can't effectively get closer than 30-50mv max difference because these packs are made of cells from several different manufacturers and the max difference varies as the voltage goes thru the charge / discharge cycle.

I set 100mv max difference as my trigger for a touch-up. What I find is a combination of 20mv'ish drift in general (from 30-50mv to 50-70mv) over those months plus 1 or 2 packs will sag an additional 20-30mv triggering the 100mv max difference. However, it's not the same 1 or 2 packs each time - mystery to me.

The touch up is 1) put an individual charger on the 1 or 2 packs and pump in 20ah'ish (~8% capacity boost) to bring them back in the mix and then auto-balance (balance the live system as the voltage goes up/down each day) for several weeks and wa-la - good to go for another 9 months.

These packs are healthy as evidence by multiple years of steady operation... BUT there seems to be some amount of disturbance in things as the pack voltages go up and down the charge/discharge cycle. I don't think the longmon voltages are changing over time since I've had the some of these longmons for 6+ years and they appear to consistently report voltages.

I do agree that it's all a bit of smoke and mirrors as I balance to what the longmons report - haven't actually measured the actual voltages since building/adding the packs to the powerwall.

 

[ConnerLabs]

I use Batrium + longmons (6+ years and counting) to monitor the 126 x 18650 packs @ 260ah each. Over the years I've discovered that balancing isn't needed day to day for healthy packs (or cells).

The touch up is 1) put an individual charger on the 1 or 2 packs and pump in 20ah'ish (~8% capacity boost) to bring them back in the mix and then auto-balance (balance the live system as the voltage goes up/down each day) for several weeks and wa-la - good to go for another 9 months.

These packs are healthy as evidence by multiple years of steady operation... BUT there seems to be some amount of disturbance in things as the pack voltages go up and down

So you pump in 20Ah per 9 months of balancing current from an extra charger, that works out as 74mAh per day? The Batrium wiki says 50-150mAh per day for cells "at the end of life", so I guess this is in the ballpark. The question then would be why the Batrium BMS needs help from an external charger and doesn't take care of this by itself. Their balancers seem to all be good for a lot more than 74mAh per day.

 

[Zwy]

The more I think about this the more I like the contactor or shunt trip breaker style of BMS. I wish they sold the Midnite Solar 250A shunt trip DC breaker in the UK, it seems like about the best safety shutoff you could imagine for a lithium battery.

The ABB commonly used is better than the Midnite. Rated for much higher DC voltage and amps. It is a beast and takes more room. It will accept 4/0 battery cable easily.

Both styles of balancer can stick in the on position and flatten a cell. I guess the active ones have more current at their disposal to do it quicker. I have to wonder how much balancing current a pack of good cells that were balanced to start with actually needs?

That all depends on how close cells were matched according to IR.

 

[ConnerLabs]

The ABB commonly used is better than the Midnite. Rated for much higher DC voltage and amps. It is a beast and takes more room. It will accept 4/0 battery cable easily.

View attachment 236469

Thanks, that breaker is a beast, I’ll see if I can find a used one cheaper than importing a MNEDC250RT. I don’t doubt that it’s better than the Midnite, but it’s probably overkill for my application.

 

[Zwy]

Thanks, that breaker is a beast, I’ll see if I can find a used one cheaper than importing a MNEDC250RT. I don’t doubt that it’s better than the Midnite, but it’s probably overkill for my application.

Easily found on ebay, that is where mine came from. 3 or 4 pole is the cheapest, nobody thinks of just using 2 of the poles. Be sure to find one that has the shunt trip. You will need a 48V to 24V converter like the one in my photo for the shunt trip as those are 24V.

 

[Zwy]

I used the SACE S5 model like this one.

The one linked has the same terminals as mine so you can install the battery cable or you can crimp a terminal on the cable with the spade type terminals.

 

[OffGridInTheCity]

So you pump in 20Ah per 9 months of balancing current from an extra charger, that works out as 74mAh per day? The Batrium wiki says 50-150mAh per day for cells "at the end of life", so I guess this is in the ballpark.

Could you provide a link to what you're reading?

The question then would be why the Batrium BMS needs help from an external charger and doesn't take care of this by itself. Their balancers seem to all be good for a lot more than 74mAh per day.

I use auto-level instead of bypass (top) balance - which pulls down the high ones in bands (groups) and cycles thru the 126 packs 10'ish at a time rather than pulling down all hi cells at once the way bypass can. Doesn't do anywhere near 50-150mah per day. Auto-Level takes 3-4 weeks to bring 50-90mv max difference down to 30-50mv max difference on 126 x 260ah packs while the system is 'live' - e.g. it's charging from PV and discharging at night while Auto-Level is operating.

 

[ConnerLabs]

Could you provide a link to what you're reading?

Wiki page for Cellmate K9
(I saw that you’re using the Longmons, but the statement of 50-150mAh/day is about the battery, it doesn’t depend on what cell monitor is used)

I use auto-level instead of bypass (top) balance

I’m interested to know why, is it because you’re using a different lithium chemistry than LFP?

 

[Zwy]

I don't use autolevel with the K9's, I had more trouble keeping cells balanced than just using bypass.

 

[J.H.]

On a slight tangent, this is a big reason why I (and I'm sure others) think it's important to have multiple batteries in a bank, instead of one large battery. Not only do you have redundancy with multiple BMS's, but you are also splitting the load to all of them, ensuring a longer life for the batteries/FETs

Never thought about it like that before….hmmm…so the less % of max load you are putting on the BMS , the less likely it is to fail under load in the open postion…?
So 50% would be better rate than 75
Or 25% better then 50…and so on…
assuming the job your powering gets done…

Is that correct..?
I hope so as I like that line of thought..!

 

[OffGridInTheCity]

Wiki page for Cellmate K9
(I saw that you’re using the Longmons, but the statement of 50-150mAh/day is about the battery, it doesn’t depend on what cell monitor is used)

These comments talk about 8ah/day as a theoretical example to illustrate 300ah would take 38 days.

And in fact, this comment about 'end of life' is weird as it implies it's OK for this situation to exist? If I had a pack or cell sagging 50-150ma/day I hope I'd detect it when it was still 20ma/day and it would be taken offline as it's a hazard just waiting to happen.

A main reason I like Batrium is the monitoring / visibility into what's happening, plus I record the data and can tease out trends if necessary. This gives me confidence the 121kwh powerwall in the basement is safe.


This chart on the link page indicates 5-6ah per day for "Existing". Not sure what "future FW" is.



And yes, I'm longmon which has a higher balance capability but I have mine throttled by current and temperature to increase life span - so no gain.

I’m interested to know why, is it because you’re using a different lithium chemistry than LFP?

Yes, I'm 18650 (e.g. non LifePo4) but this doesn't make any difference to the balance / ah per day. 260ah is 260ah regardless of chemistry - Batrium doesn't care.

And to return the main point - if one needs ongoing, every day, active balancing, then I (personally) would not feel comfortable leaving a battery like that operational for long. Balance capability is not the solution - replacing the bad cells or packs is more appropriate and visibility into what's happening is key.