Cell gets out of balance without any load

[phaberest]

I balanced my cells(200Ah lifepo4 3.2V) connecting them in parallel and charging them at 3.4V 2A for about a week, charging during the day and leaving to rest but still connected in parallel overnight.

When they all raised to 3.4 I connected them in series and to the BMS, then charged them to 100%. After 4 days leaving them at rest I placed them in a box and reduced the length of the BMS cables, plus added a 120A fuse on the main positive and now the voltages result totally out of balance and I lost about 20% capacity without applying any load

Not only that, the problem is that instead of reducing the voltage distance it sounds like it is increasing ?

The more problematic looks to be the 4th cell, but the 3rd has lost a lot as well

What should I do? Do I need to top balance them all over or should I return them?

 

[phaberest]

This is how I assembled it

 

[ArthurEld]

You need to charge to a higher voltage while connected parallel.
Most people charge to 3.65V while connected parallel.
There is information about top balancing in Resources.
You should read it because you can damage your cells if you do it wrong

 

[Steve_S]

Sorry to say, you topped them and even saturated to that voltage but they were never "full".
Initial Topping should take them between 3.600 and 3.650 and then allowed to saturate (sit at low amp for a while as you did.
Then they will all eventually settle to the 3.500 zone +/- a few millivolts. and if long enough they will even go a little lower.
The Main Power Curve for LFP being 3.400-3.000 +/-0.050

Once a BMS is attached, it will draw a "little" power to do its thing, so very slowly that will take out some capacity. IF the BMS has "Passive Balancing" that will cost some too as it burns off hi volt cells to balance with the lowest cell in the paclk. "Active" balancer will pass Hi Voltage to the low voltage cells without burning any off.

Pending on the BMS used, you should just try to charge the entire battery assembly at a reasonable but not big heavy charge rate. 200AH cells @ 0.5C = 100A If you do a 20-50A charge that would be perfect, likely they'll all come up nicely till 3.400 at which time they will start diverging and if you have a runner it will now pop up & show it's ugly head.

Runners can be clobbered a bit by 2 or 3 full Charge & Discharge cycles which can seemingly tame them a little - something to do with getting the chemical interactions inside to further activate (usually for cells that have been stored for a while, or which have actually frozen).

Hope it helps, Good Luck.

BTW, the divergence you see isn't terrible... 10-30mv is not unusual. At the top or bottom of the voltage curve, cells can deviate as much as 1mv per AH of cell capacity, so for a 200AH cell that's 200mv.

 

[phaberest]

All right, thank you @Steve_S for the deep explanation

The gap between the cells keeps increasing even if they are not under charging

To make sure I understand correctly, shall I disconnect the BMS and connect them in parallel and charge them via the PS until they are at 3.6 and then reconnect them in series and let them rest?

BTW the BMS is the one suggested by Will, the Overkill Solar with BT module that has passive balancing

 

[Steve_S]

Ideally, yes, place the cells in Parallel, bring them up 3.600-3.65 and let the amps drawn drop to <2.0A and wait an hour or so at that static level, which saturates the cells. Longer is fine too, the cells can only take so much.
Do Note, the LFP cells will NEVER EVER stay at 3.6xxx volts, they will always settle between 3.450-3.550 (pending on brand and class type). This typically takes between 4-6 hours to occur and that is quite normal.

An Important tip !
Never mess with the BMS wires while they are attached to the BMS.
Disconnect the BMS Harness before connecting / disconnecting the leads. One tiny oopsie & the BMS is cooked. "Magic Smoke Escapes".

I have a JBD BMS in my test rig similar to what Overkill is offering. They are a good BMS for their purpose. The Passive Balancing on such sized cells is so-so. If the cells are very close in Resistance & Impedance passive can work really well.