Li-Ion battery failure.

[Nigel Goodwin]

Just wondered if anyone has any thoughts?.

Just had a failed battery pack I built in November 2020, they are 4S (so 16.8V) and have a BMS board with charge balancing built-in.

Checked the cells - three are perfect, one only read 0.7V - the batteries are UK sourced from a major battery supplier, and are Samsung branded.

Anyway - I built a new pack up - so the issue is sorted, but I was interested in the reason?. So I removed the faulty cell, and stuck it on my bench PSU - set to 4.2V and 130mA - as expected the cell started charging, and the voltage slowly crept up. But it's been sat at 3.58V for a couple of hours now, so I presume that means the battery is no good?.

So was it just a faulty cell?, will it not charge properly because it was discharged too far?, and if so why didn't the protection board/balancer protect it?.

 

[Diver300]

It looks to me as though the cell has developed internal leakage. It might just be a purely resistive short of around 27 Ohms, but it could be more complicated.

Anyhow, it's leaking at 130 mA. The balancing won't be powerful enough to keep up. That will be dissipating just under half a Watt, and it's unlikely that a BMS will be rated to discharge 3 cells at 0.5 W each. Also the balancing will probably cut off at 3.4 V or so, in order not to destroy cells just because one cell in the pack is intent on destroying itself.

In some systems, balancing only happens when charging. It's only meant to deal with really small leakage that would unbalance cells over months, not hours.

 

[sagor1]

When dealing with Li-ion, if in doubt, throw it out (recycle it that is...). It is likely that battery is bad now, and you don't want to mess with it. Those batteries are famous for blowing up in flames when something goes wrong with them.

 

[rjenkinsgb]

I've had a similar experience; it turned out in that case that the way the BMS was built with separate protection and balance sections, that the voltage protection was kicking in and cutting the charge off before the balance circuit switched on...

The balance just never operated correctly and only one cell was ever fully charging.

 

[Externet]

Hi.
There is always the possibility that the charging balancing circuitry failed, not the cell itself.
The way I determine is by applying some limited current voltage only to the discharged or 'suspect bad' cell until rises over 3 Volts. That makes it 'acceptable' to the circuitry.
Then charged to specs in the pack. Next day, confirm it held charge.
May mean the charge balancing circuitry failed or has a drain.

If next day is again in trouble, the cell is kaput and gets discarded.

If the 're-accepted' cell fails again in the pack; the circuitry is bad.

Yesterday I picked three 36V LiFePO4 DeWalt battery packs. Each has one cell at 0V, 9 good at 3.3V !

 

[gophert]

Lithium ion batteries also suffer from any breach in the gas tight seal of the cap to the can. Or, a poorly formed seal in the first place. The oxygen and/or humidity convert the lithium carbonate to lithium oxides and hydroxides which do not charge as easily (higher over potential) or emit gas on charging and further damage the seal. Unfortunately, there is no perfect way to determine the quality of a seal in a manufacturing process.

 

[Tony Stewart]

If a balancer can bypass 50mA out of 2A that's only 2.5% mismatch limit and ~ 1/4W at CV charge and they need to be Ah matched << 1% to begin with for best performance. But if in operation they are drained in 30 minutes the BMS won't help to accelerate the degradation of the weakest cell at cutoff well below the threshold. Best to use more conservative cutoffs like 3.2.V to extend life or 10~20% SoC min.

 

[Glatensi]

Or, a poorly formed seal in the first place.

But it is always possible to use an extra seal if there is any doubt about the original one. It is true that this will not be very easy, but it is possible if necessary.