How to test if a cell is "weak/bad" or simply out of balance in SOC?

[apctjb]

I recently received 32 EVE 280 cells (from Xuba-Amy) and have assembled them in a 2P16S configuration with Chargery 16T BMS. I opted to skip top balancing as resting cell voltages were within 25mv (lowest to highest), I did not have a power supply available and impatient to build after waiting 60 days for my cells

I began charging at a constant 25A (utility power to the inverter/charger). All good until individual cell voltages approached 3.4V when one cell voltage (#16) began to rise faster than other cells. Within a short period of time the highest voltage cell had reached 3.65V while all the others were still in the 3.4V range at which point I stopped charging. Next morning I checked and all cell voltages had dropped to just under 3.4V and the spread between highest/lowest cell was 20mv. Started charging again at 25A and within a short time the "weak?" cell started rising rapidly to 3.6V while the others stayed in the 3.4V range.

Question; What testing can I do to determine if the cell that voltage rises rapidly above 3.4V weak/bad or simply at a higher SOC than the other cells ?

 

[fafrd]

You are ahead of me because I have not yet hooked up my charger for the first time (but am planning to).

Do you have a balancer (other than your BMS)?

From what I’ve read:

1/ voltage is a very, very poor inficator of SOC, especially with LiFePO4 which has a very flat curve.

2/ Because of the flat curve, passive balancing is close to nonexistent (since there is next to no voltage pushing charge). Your 3.65V cell quickly pushes a tiny bit of charge into the adjoining cells until it’s also in the same 3.4V range as it’s neighbors and then charge flows essentially stops.

3/ Your BMS May provide some degree of active balancing but since it is only active while charging and only kicks in near the end and based on voltage levels, it actually can’t do much to compensate for wildly-off cells (wildly different SOC) [and this is why top balancing is important - to equalize the SOC of all cells so no cells are wildly off from the others].

I built my first 90Ah battery just like you - checked that voltages were close and hooked it all up without balancing (this was before I discovered this site).

I picked up an ISDT BattGO which is a monitor with active balancing capability.

I plan to charge until any high cell gets to 3.65, then switch to the BattGO and put it in balance mode then iterate until all cells are above the ‘knee’

If that takes too long (too many iterations) it probably means that balancing cells based on voltage is just a losing proposition if any cells are at wildly different SOC than the others and going to the trouble of reconfiguring the battery with parallel busbars and charging with a 3.65V charger is the only realistic to get starting SOC balanced.

Your ‘weak’ cell is likely not weak, it is likely just much more full than the others - for example, 7 cells at ~50% SOC and one at 99%...

 

[nosys70]

as long as the cell stay in the safe range, i think you should not worry too much for the cell.
it will go 3.7 when others stay at 3.4, but it does not hurt the cell.
for the SOC of the battery it is more problematic since the SOC is the sum of all cells.
so if your read 27.4, but in reality you got less, except one cell is pushing the level up, your charger could stop the charge too early.
I think there is no solution than replacing the ""bad cell .

 

[newbostonconst]

You need to put like a .5 ohm resister (or higher ohm - higher resister just takes longer - 10 ohm is on the big size) across the high cell to bring it down or do what you are doing and let the BMS balance it (this will likely take a week). I used the resistor method and only charged when I could check it every 20 minutes and bring it down as needed. It still took almost 3 days of fiddling.

LifePo4 cells have a very flat charging curve, so you can have cells at nearly the same voltage and they are like 10% different in charge state even though the voltage is almost the same. I learned the same way you did but had to buy a new cell After I ruined one cell I did what I explained above. The best way is to link them all together and charge to 3.5

 

[Bob B]

I would set the Chargery BMS to balance in all modes temporarily and charge at a very low rate if you have the option to do that. See if the Chargery will balance them out. The Chargery does balance at 1.5A.
If you don't have the option to charge at a low rate, charge until you see that cell going higher again let it sit and keep repeating that.

Also triple check to make sure you don't have any poor connections.

 

[fafrd]

You need to put like a .5 ohm resister (or higher ohm - higher resister just takes longer - 10 ohm is on the big size) across the high cell to bring it down or do what you are doing and let the BMS balance it (this will likely take a week). I used the resistor method and only charged when I could check it every 20 minutes and bring it down as needed. It still took almost 3 days of fiddling.

LifePo4 cells have a very flat charging curve, so you can have cells at nearly the same voltage and they are like 10% different in charge state even though the voltage is almost the same. I learned the same way you did but had to buy a new cell After I ruined one cell I did what I explained above. The best way is to link them all together and charge to 3.5

How did you ruin 1 cell?

And how did you charge to 3.5V?

I’m in the same situation as the OP and hoping to find a way to top balance my cells without having to disassemble the 8S battery.

The BattGO has active balancing, so I’m hoping that if the battery is either charged so that all cells are above the knee or discharged below the start of the slope to empty that it can get the cells balanced in relatively short order.

Is that a pipe dream?

 

[Ampster]

Question; What testing can I do to determine if the cell that voltage rises rapidly above 3.4V weak/bad or simply at a higher SOC than the other cells ?

others have already started down this path and i will add my current experience.. About all you can do while it is in the pack is try to pull out some charge. I use two 30 Ohm 10 Watt resistors in parallel glued to a heat sink. I have it wired with aligator clips. Next time, based on foregoing advice, I am going to get a bigger resister with less Ohms to make it go faster. I have one pair of cells that were behaving like that and I think I have them balanced at the top with the resistor. I did lower charge current and Absorb to allow some time. There is no way of knowing how long because after the charge cycle is over the cells come to resting voltage. I did it in iterations checking at each charge cycle and the deltas became smaller.
Eventually things were okay at the top but last night my pack went below 50% and these cells were below the others so I know they, or at least one of them, have/has less capacity. The next step is to remove them and individually test them. I purchased spares so I can do that without taking my pack offline for more than an hour.
Inserting two new matched cells will be tricky. I have to pick a point where the pack is at the top of a charge cycle and charge the replacement cells to that same voltage. I may have to crank up my charge voltage to get them more into the knee at about 3.5 volts per cell. (Lately I have only been charging to 3.4 volts)
Of course the alternative is to diisasemble the whole pack and parallel top balance it with the replacement cells. That is a 3 to 5 hour project not counting the time spent in parallel. My timing is complicated by the rolling blackouts in California so I may have to wait until a more opportune time.
At any rate, I hope that helps.

 

[RCinFLA]

You need to drag the higher voltage parallel pair down by 2% to 4% of their full state of charge. 560 AH x 3% is about 17 AH.

If you can get some 1 ohm, 20 watt power resistors you can bleed it down. Its only about 3.6 amps per power resistor so will take a while.

Be careful, the resistor(s) will get very hot.

Other possibility if you have 100 foot roll of 14-2 or 14-3 romex can tie one end's white and black together and use other end white-black as terminals of resistor across 3.6v battery. 200' of #14 solid copper is about 0.5 ohms. Will have about 7.2 amps current and about 0.25 watts of wire pair heating per foot which is fine.

You can do it with less #14 length for more current, just be sure not to exceed 1 watts per foot-pair of wire (about 50' of romex pair, 100' of #14). You might have to uncoil it out to better dissipate heating if you get close to 15 amps of current through #14 wire.

 

[Ampster]

I’m in the same situation as the OP and hoping to find a way to top balance my cells without having to disassemble the 8S battery.

The BattGO has active balancing, so I’m hoping that if the battery is either charged so that all cells are above the knee or discharged below the start of the slope to empty that it can get the cells balanced in relatively short order.

Is that a pipe dream?

Not necessarily. Unlike me you have active balancing at much higher current than my BMS. I also have 2P16S pack.

 

[newbostonconst]

How did you ruin 1 cell?

And how did you charge to 3.5V?

I’m in the same situation as the OP and hoping to find a way to top balance my cells without having to disassemble the 8S battery.

The BattGO has active balancing, so I’m hoping that if the battery is either charged so that all cells are above the knee or discharged below the start of the slope to empty that it can get the cells balanced in relatively short order.

Is that a pipe dream?

I had 16 cells and my inverter was capable of voltage levels to charge a 17 cell pack. I had bought a spare cell when I got my set so I wanted to put it in for more capacity. I waited for all the cells to be the same voltage to the thousandths of a volt (0.001) and put it in the pack.

Well I made the mistake and put the cell in when the pack was at a mid state of charge. So when it came up to near fully charge the cell I put in was charged more then the others(even though the voltage was the same) and when over voltage to just over 5 volts and expanded.

It still worked but I didn't like it so I ordered 2 new ones. Then when I put in the 17th cell the second time I watched very closely every 20 minutes while charging and turned the charger down to 3 amps and it took a full weekend before I could charge the pack fully. I used a half ohm resistor to dump the load every twenty minutes (I also had a 10 ohm resister with alligator clips on it as Ampster explains above) across the terminals of the high cell.

After that it still took my 5 amp ebay active balancer 4 more days to balance it and 3 charge/discharge cycles. It is perfect now.

I will say that ebay 5 amp active balancer works good. I can see it move the voltage around and in the process above shorted it a couple times and with the resistors across the cells it never got ruined.....I was surprised after thinking about what I was doing to it....putting 1/2 ohm loads across its capacitors and shorted it....