Pairing mismatched cells

[Peterbylt]

Looking for suggestions, Through an apparently bad choice, on my first purchase of 8 Catl 310ah Cells, the cells arrived in bad shape, some with dents in the case, some bent posts, some of them bloated, one of them leaking and unusable.

These were supposed to be Grade A cells, as things turned out it will be impossible to return or get my money back.

I have tested all the usable cells to be 3.2 volts and top balanced them to 3.5 volts successfully.

I have ordered 9 more Catl 302ah cells from a hopefully better source, 8 cells to make up the 24 volts needed and one cell to maybe fill out the questionable original cells and build a 2P8S 24 volt battery.

I am using an Electrodacus SBMS0 so they need to be 2 parallel and then 8 pairs in series.

My Question what do you think would be the best pairing of the cells?

My initial thought is to pair the original cells, then pair the new cells, leaving the one pair of an original and new cell, these would be the individual cells from both sets that matched voltage wise the closest after top balancing.

I understand the Battery pack will only be as good as the weakest pair in the pack.

Just trying to make the best of a bad situation.

Peter

 

[MisterSandals]

I understand the Battery pack will only be as good as the weakest pair in the pack.

Exactly. You need to rank them from strongest to weakest. Then pair strongest with weakest. Second strongest with second weakest...

I did not have a capacity tester nor the desire to do a cell by cell capacity test.

I created a 2P4S battery this way (i will change steps to 8S for your use case):

Label the cells, i like using letters (A, B, C...). Create 2 separate 8S batteries.

Run each battery thru a couple charge/discharge cycles taking notes on which cells charge fastest and slowest.
This will let you rank the 8 cells in each battery.

Take the 4 weakest of battery1 and swap with strongest of battery2.

Repeat charge/discharge cycles and rank cells in each battery. Swap 1 or 2 cells moving weakest 1 or 2 cells from battery1 and strongest of battery2.

Run a couple charge/discharge cycles and rank cells.

Pair strongest of battery1 with weakest of battery2. Second strongest of battery1 with 2nd weakest of battery2. Do this for all 8 pairs.

This is fairly rough, if you see a chance to improve getting all strongest in one 8S battery and weakest in the other battery, do it. And you might be able to rank cells with a single charge/discharge cycle.
Overall, this should make your pairs pretty equal.

 

[BigGun]

How many cells will you have from each batch --- I would think you would need one from each batch per parallel pair -- then strongest with weakest and so forth --- Seems like capacity testing each cell will get you a lot closer

 

[Peterbylt]

How many cells will you have from each batch --- I would think you would need one from each batch per parallel pair -- then strongest with weakest and so forth --- Seems like capacity testing each cell will get you a lot closer

I will have 7 cells from Batch 1 and 9 cells from Batch 2.

Peter

 

[Peterbylt]

Exactly. You need to rank them from strongest to weakest. Then pair strongest with weakest. Second strongest with second weakest...

I did not have a capacity tester nor the desire to do a cell by cell capacity test.

I created a 2P4S battery this way (i will change steps to 8S for your use case):

Label the cells, i like using letters (A, B, C...). Create 2 separate 8S batteries.

Run each battery thru a couple charge/discharge cycles taking notes on which cells charge fastest and slowest.
This will let you rank the 8 cells in each battery.

Take the 4 weakest of battery1 and swap with strongest of battery2.

Repeat charge/discharge cycles and rank cells in each battery. Swap 1 or 2 cells moving weakest 1 or 2 cells from battery1 and strongest of battery2.

Run a couple charge/discharge cycles and rank cells.

Pair strongest of battery1 with weakest of battery2. Second strongest of battery1 with 2nd weakest of battery2. Do this for all 8 pairs.

This is fairly rough, if you see a chance to improve getting all strongest in one 8S battery and weakest in the other battery, do it. And you might be able to rank cells with a single charge/discharge cycle.
Overall, this should make your pairs pretty equal.

Thank You for this, I had not thought to pair the Weakest and Strongest cells to attempt to gain a middle ground.

Peter

 

[coplate]

I'm actually considering doing this right now as well, so I thought I would add onto this thread instead of making a new one.

I ordered 16 "grade A 90ah" cells, but they are all testing between 50 and 70, in a fairly even distribution.
Instead of making a 16s that is limited to 50ah ( 50ah @ 48v =2400Wh), I am looking at making 2p8s with the pair being 120ah( 50+70, 60+60 etc ) for ( 120ah@24v =2880Wh).

I was hoping to use 48v for smaller a wires, and I already had a 48v inverter, so I am trying to decide if the extra 400Wh is something I need.

I don't know if having the strongest cell with the weakest will keep adding stress to the weakest, and it will degrade faster this way, or the other way, and wondered if anyone had thoughts on that.

 

[MisterSandals]

and it will degrade faster this way, or the other way, and wondered if anyone had thoughts on that.

I would think the other way around from what i've seen. Having a stronger cell paired with a weaker cell keeps in from running away on the high or low ends, which would be the most stressful and most degrading.

 

[MisterSandals]

Instead of making a 16s that is limited to 50ah ( 50ah @ 48v =2400Wh), I am looking at making 2p8s with the pair being 120ah( 50+70, 60+60 etc ) for ( 120ah@24v =2880Wh).

Wow this is a big adjustment but I get what you are thinking and cannot say i disagree with the numbers.
Since you have a 48V inverter and maybe more 48V stuff, I'd strongly consider a 5A active balancer to see if that's enough to keep your 16S inline with your use case. A gentler charge and discharge would make a balancer more effective. Not sure if this works with your use case.

I'd bet if you bought a 16S 5A balancer and decided it wasn't for you, you'd be able to sell it here without much trouble or loss. Just a thought.

 

[Peterbylt]

I took all 16 cells (302AH CATL) and charged them in series, 8 at a time as a 24 volt battery, Batch one and then batch two.

I used the Electrodacus SBMS0 to monitor the Batteries.

Surprisingly when I brought them up to 3.56V they were all very close to each other with the highest at 3.56 and the lowest 3.52, I ran them through a couple charge cycles.

Even though Batch one were questionable batteries dented and bloated they performed as well as Batch two the Grade A cells.

I did a capacity test on both sets, I ran a window air conditioner off the AIMS 3K inverter that pulled approx. 23 amps, It took slightly over 13 hours to get to 2.8 volts where the SBMS0 cut it off, Both sets ran almost exactly the same, 299AH can’t complain for Advertised 302AH cells.

The Electrodacus cuts off the Charging or Load when one cell reaches the target value for 5 seconds, currently set to 3.56 on the high end and 2.8 on the low end.

The object was to find the highest and lowest cells from each batch at the top and bottom of the charge and match the highest and lowest with each other for the parallel pairs, the differences were so insignificant that it’s not really worth the effort.

My current thought is to match one of the Grade A cells with one of the Questionable cells in parallel pairs and run a capacity test on the entire 2P8S pack.

Peter

 

[coplate]

... I'd strongly consider a 5A active balancer...

That's an option I had completely forgotten about. The benefit of 48v is that to get 1500W AC, I need a much lower amp draw from the battery pack, and a balancer can probably keep up with that without too much trouble.
I'll look into this, thanks again!