I accidently deep discharged my LiFePo4 prismatic cells

[merlin077]

Hi folks,

i did something rather dumb... My 16S LiFePo4 pack consisting of 16 105Ah Liitokala cells was discharged to 1,4V per cell.
The charge controller went dead two weeks ago and i sent it in for repair. Unfortunately i forgot to disconnect the BMS and now the cells are down to approx. 1,4V each and swollen (see attached photos).


They were very lightly compressed, which might had benefit the swelling.

Under close inspection, i took one cell and applied charging voltage with very little current, this cell is now up to 3,2V, the bulging did not got worse.

What do you think? Might they be still usable or a potential fire hazard? I don't want to waste 16 quite new batteries but otherwise i like my house unburnt

Thanks a lot!
merlin

 

[upnorthandpersonal]

If they keep a charge over time, and stay at a pretty constant voltage after settling, they could be fine. LiFePO4 is, for a lithium battery, quite forgiving. I would still go through the trouble of testing each and every one of them for capacity. Also, the cells that are bloated: you can try to slowly charge them to 30% or thereabout, then disconnect them and let them sit for a few months. I've noticed they can de-bloat substantially over time.

That said, getting a cell below 2.5V is not good at all and you should probably consider them very damaged. Don't keep them in the house; put them in a shed in a metal box or some other place where if something catastrophic were to happen, they can't cause more damage.

i forgot to disconnect the BMS

That's why I'm a fan of the JK BMS: it can shut itself off when cells get below a certain value preventing draining the cells to power the BMS.

 

[hankcurt]

@merlin077 , That is just painful to look at.
Sorry to hear about the damage, but I am curious which BMS you were using.

 

[Bud Martin]

Wow, I did not know that the cell can bloat when it is discharged to really low Voltage.

 

[merlin077]

Thanks for your condolences...
I am still a bit concerned putting those batteries into service again, even if they hold charge. If there is a internal short, my solar powered garden shed gets local news coverage Or do i overreact? I know, nobody can see inside the cell chemistry, but it's a bit disappointing, that they went bad so quick...

Another theory of mine: one cell went bad and because it was a 16S configuration, the higher resistance cell drained the others, but that's pure speculation...
What a mess

I used a DALY BMS. It's supposed to cut off at 2,2V (which it did i think). Connected to the BMS is a ESP32 which reads the operating values and reports it to my Home Assistant instance. But i doubt that those few mA drains 16 beefy LiFePo4 to their death.

After two days, the cells read 2,0V each - directly after disassembling the pack it was 1,4V.

 

[littleharbor2]

Maybe this will ease your pain a little. My first attempt at LiFePo4. Bought some questionable cells (as I realized when unpacking the very poorly packaged container, still paid full price). I was top balancing them and accidently bumped up the max voltage setting on the power supply. After about 7 long, slow days of charging I found they had jumped up to over 5 volts overnight. What a sick feeling seeing these cells all bloated like this. There were 8 cells BTW.

 

[JoeHam]

I would not use those bloated cells in my house.

Google LiFePO4 dendrite formation due to severe over or under charging.

It is irreversible and in extreme cases the dendrites could puncture the cell casing.

Sorry for your loss but it is tuition paid to the school of hard knocks. That tuition is non refundable and I have paid it myself more than once. In many different ways over the years.

I did ruin some cells with a Chargery BMS the exact same way

Eight of my 16 cells were unbloated and able to hold a charge so I built two 12V nominal batteries from them.

 

[nt40lanman]

You people are scaring he hell out of me!!!

 

[Charleswgibbs]

Was this a failure of the bms? I thought one of the main purposes of a bms was to disconnect if any cell voltage gets too low?

 

[time2roll]

For me... those cells are done. Although I am not very clear on how the BMS by itself discharged the cells that low. Done is done, time to move forward with something else.

 

[dweick]

Sell them on Alibaba as Grade A cells.

 

[nt40lanman]

I would try to cycle them or work with them a little, even if only to learn something

 

[RCinFLA]

You took them lower than 1.4v. Maybe 1.4v was after you took the load off of them. They will not bloat on solely discharge until significantly below 1 volt.

They will start to grow metal dendrites if not recharged soon. Recharge current has to initially be kept low until cell gets above 2.5v.

 

[merlin077]

You took them lower than 1.4v. Maybe 1.4v was after you took the load off of them. They will not bloat on solely discharge until significantly below 1 volt.

They will start to grow metal dendrites if not recharged soon. Recharge current has to initially be kept low until cell gets above 2.5v.

Do i have to charge them full to stop the forming of dendrites? My plan was to bring them up to 3.2V but not fully charged in case they blow up… Currently, they are charged with 0.5A at 3.2V

 

[RCinFLA]

Do i have to charge them full to stop the forming of dendrites? My plan was to bring them up to 3.2V but not fully charged in case they blow up… Currently, they are charged with 0.5A at 3.2V

No, just get them above 3v. Going to at least 3.2v gives you several months of self discharge before you drop below 2.5v again.

For most electrolytes used for lithium ion cells, they will not experience much electrolyte breakdown, which is the cause of bloating, until cell gets above 4.3vdc or below about 0.1 vdc cell voltage. These voltages are not an abrupt line in the sand. Don't leave LFP cells at greater than 3.7v for weeks or you will see some bloating. Below 1.0v is another issue as it allows lithium plating to form. The manufacturers spec of 2.5v minimum discharge is just including some margin to the 1v since it does not take much additional discharge to quickly drop from 2.5v to 1.0v.

For Li-Ion battery cathode chemistries that have a full charge voltage spec of 4.2v +/- 0.05vdc this does not give a lot of margin on overcharging voltage.

Manufacturers often have some of their own special sauce additives to electrolyte to get better attribute out of electrolyte. These might be better low temp performance or some other attribute. Usually there is always a downside consequence trade off. Samsung for example has electrolytic additives to allow exceeding 4.3v maximum charge voltage so they can get the 'quick charge' feature in their mobile phones. What they don't tell you is the quick charge function will not result in a fully charged battery, but they put a larger AH cell in the phone to make up for it.

 

[merlin077]

Did i mention, that i have a monitoring system set up? Your guess might be right, the cells were down to about 1.1V

And i have an idea what caused the pack to over discharge: Load was disconnected and besides the Daly BMS an active balancer (the Heltec type units with an array of SMD caps) was connected to the cells. I‘m not entirely sure if this thing has some sort of cutoff voltage - looks like it doesn‘t…

I‘m now veeery gently charging the cells to over 3V. I hope i can charge them quick enough before they look like footballs…

 

[RCinFLA]

At some point there may not be enough total stack voltage to run the BMS so you could have lost info on the lowest voltages.

 

[JoeHam]

Although I am not very clear on how the BMS by itself discharged the cells that low.

Most BMS that I am aware of have a standby current draw.

Given enough time it drains the battery.

They typically can turn off other current draws but still need power to keep themselves “awake” and monitoring.

 

[time2roll]

Most BMS that I am aware of have a standby current draw.

Given enough time it drains the battery.

They typically can turn off other current draws but still need power to keep themselves “awake” and monitoring.

From the Overkill manual FAQ:

Q: What is the quiescent current? What can I do for long-term storage?
A: The quiescent current is as follows: (this was measured on a 4 cell BMS)

• ● 5.5 milliamps with everything off, when the BMS is active, but no bluetooth.
• ● 15 milliamps with the bluetooth active (after about 10 seconds it drops to 0.8 milliamps.
  Reconnecting it wakes it up again).
• ● 0.8 milliamps when the BMS is inactive.
  So, assuming your battery setup is 100 amp hours, the BMS would run for 17 years. This proves that the BMS can be connected for long periods of time without any fear of it draining the battery.
  Unplugging the balance connector would ensure complete shutdown
  NOTE: The cell’s self-discharge rate will always cause the battery to drain over time, which may be several percent per month. This is simple chemistry and physics; there’s nothing that you or the BMS can do to avoid the battery cells from self-discharging over time, other than to occasionally top up the batteries.

 

[merlin077]

„Pregnant and yet full of energy - Birth preparation course with merlin077“.



I decided to make the best out of this mess and make it an experiment… Charging each cell veeeery carefully to 3.2V and measuring the swelling with a caliper over time.
An internal resistance meter is on the way, also some sort of load with a shunt to measure capacity will be built/bought, so that we know how much (if any) capacity is left.

Of course they will be thorough testing going on before they even remotely see a charge controller again… Even after that, precautions will be taken to sort out any possible danger.

Unfortunately i didn‘t measure the capacity when the cells were brand new, but from experience i guess they were about 85-90Ah each (rated 105 chinese Ah from the factory)

If something feels fishy and/or unsafe, of course they go into the dumpster immediately - but i think it‘s worth at least trying. Gaining knowledge from other peoples stupidity, here we go!

Last but not least: Thank you all for sharing your opinions and knowledge. Really appreciate it!

 

[SteveCA]

From the Overkill manual FAQ:

Q: What is the quiescent current? What can I do for long-term storage?
A: The quiescent current is as follows: (this was measured on a 4 cell BMS)

• ● 5.5 milliamps with everything off, when the BMS is active, but no bluetooth.
• ● 15 milliamps with the bluetooth active (after about 10 seconds it drops to 0.8 milliamps.
  Reconnecting it wakes it up again).
• ● 0.8 milliamps when the BMS is inactive.
  So, assuming your battery setup is 100 amp hours, the BMS would run for 17 years. This proves that the BMS can be connected for long periods of time without any fear of it draining the battery.
  Unplugging the balance connector would ensure complete shutdown
  NOTE: The cell’s self-discharge rate will always cause the battery to drain over time, which may be several percent per month. This is simple chemistry and physics; there’s nothing that you or the BMS can do to avoid the battery cells from self-discharging over time, other than to occasionally top up the batteries.

100% agree. I good BMS will not drain your 100AH batteries in the slightest. Something else is in play here, or the Daly is defective.

 

[merlin077]

100% agree. I good BMS will not drain your 100AH batteries in the slightest. Something else is in play here, or the Daly is defective.

My guess is the Heltec Baöan

 

[merlin077]

I‘m also pretty confident the Heltec thingy led to the discharge. Will investigate that further when the pack is reassembled

 

[littleharbor2]

A dc clamp meter is real handy for troubleshooting this kind of stuff.

 

[merlin077]

A dc clamp meter is real handy for troubleshooting this kind of stuff.

I have one - didn‘t had to use it cause the Heltec balancer did a good job until it effed up But just for curiosity, i will investigate, what led to the deep discharge situation...