LiFEPO4 sudden voltage drop on one cell

[bitrate001]

Hi and thank you for an excellent forum!

I have a 200 Ah LiFEPO4 battery with a built in jdb BMS (JBD-SP04S034-L4S-200A). The battery powers from a 170W solar cell through a Victron BlueSolar MPPT 75/15.
The loads on the battery are usually low 2-5A. Occasionally 10-15A when using 230V AC equipment through a Victron Phoenix Inverter 12/500.

Now to the problem. I have been using this setup for around a year now and at 3 times the battery has gone from 90-100% capacity to zero in a few hours without any load at all. At first I thought it was temperature related (first time it was around -15 C) but it have happened two times after that in 15-20 C.

I have opened the battery and checked for loose or corroded wiring but did'nt find any (as far as I can tell that is).

Attached pics shows current load vs individual cell voltages at one of the outages.

Any thaughts on why this is happening?

Best Regards Håkan

 

[Sverige]

Hej!

I don’t think it’s going from 90-100% charged to zero. It seems to be not very charged at all and simply is running down to zero state of charge and you see the weakest cell drops in voltage first. The battery is either out of balance, or that cell which drops lowest has lower capacity than the rest.

Your bms should cut the power off before a cell gets down below 2.5V, which is what the graph shows. So I guess your BMS is not set up right.

 

[mikefitz]

Cells are out of balance and not being charged to a high enough voltage. Cell balance won't take place unless the cell volts are above 3.4 volts.

 

[Q-Dog]

That battery is nearly dead.

 

[TomC4306]

I agree with qdog. Battery is dead.
Focus your attention on a full charge.
You need more solar if that is how you want to charge.
170watt panel x.6 (age, degradation, improper aim, clouds) = 100 watts. X 5 hours= 500 watt hours of charging (loads off, inverter disconnected) per day if you're lucky.
Battery is 200ah x 12.8v = 2560 watt hours.

So, 6 days of uninterrupted charging.

 

[jjww]

since you can open the battery, measure the individual cells when at 0% will tell you your bad/weak cells

 

[Suijkerbuijk]

I agree with qdog. Battery is dead.
Focus your attention on a full charge.
You need more solar if that is how you want to charge.
170watt panel x.6 (age, degradation, improper aim, clouds) = 100 watts. X 5 hours= 500 watt hours of charging (loads off, inverter disconnected) per day if you're lucky.
Battery is 200ah x 12.8v = 2560 watt hours.

So, 6 days of uninterrupted charging.

He do not need it.
He do not use the battery a 100% .
I think one cell its low or to high and the bms stop charging wel he still use the battery output .
And the bms dit not reconnect.
So it looks like that his cell came out off balance.


I use a 150ah and 250watt solar . (Panels are flat )
If i pull in the night 40% out of it.
Next day its back up a 100% and the fridge still running on the day time by the charger

Than the damage is done by try to charger on -15
You can not charge lfp under 0 .
So its solar Control normal not charge if the bms is set correct..

So his real problem looks like 1 dead cell and other cells are over volt to compensation.
The charge will try to Pump up to that voltage.

 

[TomC4306]

No. Look at the graph again.
Three of his cells went below 3.0v.
This pack is empty.
Operating range for LFP is generally accepted as 3.0 to 3.4 volts.
Recharge to 13.9 or 14 volts and observe cell behavior there.
Not enough data to support your dead cell theory.
Below 3v, a dive to cutoff is possible, particularly in an unbalanced battery.
Collect data.

 

[Suijkerbuijk]

No. Look at the graph again.
Three of his cells went below 3.0v.
This pack is empty.
Operating range for LFP is generally accepted as 3.0 to 3.4 volts.
Recharge to 13.9 or 14 volts and observe cell behavior there.
Not enough data to support your dead cell theory.
Below 3v, a dive to cutoff is possible, particularly in an unbalanced battery.
Collect data.

The start all cell are almost the same.
Than you see 2 cells really go fast discharge.
The purple cell is the one that trikker the bms .
You see that one drop really fast
That cell is damage.

Agree with you that the battery is not full charge.
He can try to charge them singel one by one.
Than top balance them in parallel.

Still i think one cell is damage .
Not dead like dead under the ground.

 

[bitrate001]

Thanks for your answers! Yes the battery probably defect. When it went dead this time soc reported still around 50% and capacity left 100aH.
The first two times it went dead soc reported around 80-90%.
It's now fully charged but one cell is 0.1V lower than the others that is completely balanced.
The cells are fixed to the box with a thick layer epoxy resin so I doubt that I can replace any cells.

I guess I have to replace the battery.

Best Regards Håkan

 

[rilonau]

The BMS typically has problems with accuracy at low charge / discharge rates. When you just charge by a few amps and discharge nearly the same my JK-BMS switches between -1.5A and +1.5A. The calculation of actual SoC gets very inaccurate over time! You have to check voltage at the cells from time to time and to keep the BMS up and accurate you need full cycles every few months. Only at your set 100% SoC voltage the BMS 'knows' the cells are full and only at 0% and the corresponding voltage it 'knows' cells are empy. Everything in between is math and best guess, esp. with low and pulsing currents. Be aware: most inverters draw current in pulses of 100Hz and most AD-converters can't handle this correctly and with needed accuracy in low load area.

 

[Q-Dog]

I don't think there is anything wrong with the battery. It just needs to be fully charged, slowly, and given some time for the BMS to balance the cells.

 

[Sverige]

I agree, no evidence in the thread the battery is damaged, just simply that it was fully discharged and out of balance.

Some cells *may* be damaged due to the depth of discharge the BMS seems to be allowing, but that’s yet to be confirmed.

 

[Skypower]

Until the battery has reached a full state of charge (3.45 vpc) and is balanced at that point (delta less than .015V and balance only above 3.410 vpc) also sure that ALL terminal connections are good, I wouldn’t call that a bad cell. Then discharge again and if it has the same issue, only then I’d consider it a bad one. How a battery is treated at the top greatly affects the bottom and the whole performance. Balance could take time.