Did I kill my LFP cells?

[bluetrepidation]

Hello,

I have a 2p8s pack of 100Ah Fortune cells from Electric Car Parts Company. I've fully cycled the cells maybe 40 times. Recently I've noticed that I can't run my loads as long as in the past. When charging I noticed my BMS seems to be cutting off the pack early. So I decided to investigate with a multimeter and log the min/max voltages on particular cells. When charging at 70 amps (0.35 C) cells 8, 6, and 5 reach 3.6 volts well before the other cells. So I re balanced my pack twice and kept testing over the last two weeks. Then I noticed something that made my heart sink. When running a decent load of 90 amps (0.45 C) cells 8, 6 and 5 also sag in voltage. The cell imbalance on my BMS shoots up to 300 mA. To me this is a sign of high internal resistance. Any ides on what I should do? Or just call it a $2k loss and move on?

I've checked the following....
- Looked for loose connections between cells.
- Verified that my multimeter and the voltages displayed by my BMS matched and are accurate.
- Verified that my BMS cuts off discharge when any of the cells hit 2.5 volts.
- Verified that my BMS cuts off charge when any of the cells hit 3.6 volts.
- Max discharge the pack has been exposed to is 121 amps for 3 seconds averaging 11 amps over the long term.
- Max charge has been 70 amps.

Thanks,
A.J.

 

[sunshine_eggo]

Quadruple check all of your connections - especially cell interconnects and BMS sense leads. Make sure the BMS sense ring terminals or whatever you used are of good quality.

 

[Bud Martin]

If you swap those suspected cells to difference locations, does the problem follow?

 

[John Frum]

Make sure your dvom agrees with your BMS.

 

[Substrate]

Sounds dumb, but have to ask:

How are you connected, ie separate bus bars, or just jumper cables? Are you charging and discharging from the end of a long "ladder" type setup with positive and negative coming off solely from one of the batteries at the end of the chain?

 

[740GLE]

What are your charging settings?

Also cell imbalance is listed as mV not mA.

 

[bluetrepidation]

Hello everyone. I took the pack apart and connected all my cells in parallel to balance them. After this I ran a load test and I'm seeing the same results. I do like the idea of moving the problem cells and seeing if the voltage sag follows. Let me run that test and I'll post pack soon. Thanks for all the ideas. I did purchase all new bolts without the nylon locking ring to be sure everything was torqued correctly. All the connections looked good and all my cell voltages are matching between two Keysight U1282A multimeters and the BMS.

A..J.

 

[bluetrepidation]

What are your charging settings?

Also cell imbalance is listed as mV not mA.

Typo.

 

[sunshine_eggo]

Hello everyone. I took the pack apart and connected all my cells in parallel to balance them.

Did you charge them to 3.65V in this config, or did you just let them sit in parallel?

 

[bluetrepidation]

Did you charge them to 3.65V in this config, or did you just let them sit in parallel?

Yes charged to 3.65 with all cells paralleled. Then I let them stand for 3 days while I got caught up in yard work.

A.J.

 

[sunshine_eggo]

Move suspect cell and triple check all crimps and connections.

Confirm BMS is configured to balance only above 3.40V, 20mV cell deviations and only during charge.

 

[740GLE]

Drop that deviation down to 5mv, with slow balancers you need all the help you can get.

 

[bluetrepidation]

OK good news. I moved this pack to a test bench and dedicated some time to troubleshooting it. Two problems. First, the balance current on my old BMS was not enough to keep up with these large amp hour cells. I installed a JK BMS with a 2 amp active balancer and after manually balancing again I was able to pull 98% of the capacity. Now why couldn't the old BMS keep up? This leads me to the second problem, the terminals on cell 4 where getting warm to the touch. I pulled the bus bars off and cleaned everything with rubbing alcohol then reassembled TWICE. No change. I was thinking of using sand paper next but ordered some MG Chemicals 847 carbon conductive paste. This worked! Zero heat now. I was able to pull 100% capacity finally! I did torque the heck out of the bolts on the other two tests. I guess more torque doesn't always solve a connection issue. Give the paste a try it is amazing.

A.J.