Steps to safely recover low voltage NCM cells?

[lord_gammon]

Hi,
I have a prismatic NCM pack of 20S 92Ah 3.65V nominal cells in an electric motorbike. They have a Daly BMS set to protect them from over discharge, but somehow it didn't do that and whilst stood through winter they have run down to a range 1.8V to 2.2V. Since it's like $900 worth of cells, I would like to recover them, but not burn the house down trying. I hope that since they had no significant load whilst under-voltage this will help.
Anyone got some practical experience of this, or handy tips?
Should I connect all the cells in parallel and balance them first?
What is a safe current and voltage to use for charging?
Are they likely to be permanently damaged?
P.S. Daly BMS are useless, in case anyone didn't know already.
Thanks, I appreciate you helping.
- N

 

[amberwolf]

Below the cells' spec sheet's minimum voltage limit, there is always the chance of cell damage. The lower the voltage, the higher the chance of damage and/or the worse the damage could be.

If they are good quality EV-grade cells (such as those recovered from used EV packs), they may be safe enough even if degraded by the damage, but they may not have the same capacity or current capability that they did before.

I had a 4s1p lighting pack made of used EIG NMC C020 cells drained to virtually nothing that I recovered as an experiment, documented in my CrazyBike2 thread around a decade or so ago if you want to read up on the procedure I used. It did still work, and did not swell or otherwise have external indications of problems, but it didn't perform as designed anymore; if it had been a traction pack it would have had to be relegated to some other lighter duty with less critical function. Some years later this pack was again drained severely on the SB Cruiser trike, and at least one cell did not survive this, and actually reversed voltage and swelled up, requiring replacement.


I would not recommend paralleling the cells to rebalance them; it may cause higher current flow than you want them to see in this state. (you can calculate the potential current that could happen by dividing the worst-case difference in voltage between cells by the internal resistance of each cell that is parallelled (ohms law)). Keeping recharge current in overdischarge cases to a few mA or maybe few dozen of mA (lower is better) until it is above normal minimum safe recharge voltage (typically 2.8-3v or more) would be easiest on the cells.

Regarding Daly BMS, Methods has a few threads from the last several months with good info on the issues. There are a few other threads around the forum reporting problmes with them too, and speculation on why some and not others.

 

[harrisonpatm]

I have had success recovering low voltage cells with very low CC/CV, like in the range of 25ma per cell. That'll take quite long, of course. If you want, you could parallel all the low voltage cells and connect to a power supply set to the correct current. I have a rig of 40 18650 slots, all connected in parallel to 1ea TP4056, which can deliver 1A of charging current. Divided by 40, you get 25ma per cell.

Can take 10-20 days. Be exceedingly careful if you choose to do this, it's at your own risk. Check frequently for warm cells, maybe do it outside. And even if you manage to recover any cells, they are likely to have reduced capacity.

The cells I recover are low-drain laptop cells, destined for further low-drain use, if they pass the rest of my test hurdles, which they don't always.

 

[lord_gammon]

Thanks for the feedback.
I'm charging each cell up from 2V to 3V @1A (they are 92Ah cells). The first one got there in 30 mins so 6 hours for 20S.
I'm interested to see how they perform afterwards. They haven't been used much.

 

[harrisonpatm]

I'm charging each cell up from 2V to 3V @1A (they are 92Ah cells). The first one got there in 30 mins so 6 hours for 20S.

Ok, but my recommendation is to continue the full charge to 4.2v (they are Li-Ion, right?) at that continued lower amperage. Not just up to 3v, then normal charge current after that. This will take some time if you do it my way. Again, just a recommendation, you don't have to follow it, but it's what I do. If it were me, and I was trying to revive hundreds of dollars worth of cells, I would be fine with it taking months, rather than risk damage by doing it faster.

To figure out a safe amperage for you, compared to what I do: I charge low voltage 18650s, roughly 2500mah cells, at 25ma. Meaning 1/100C. I suppose it'd be reasonable to take that to 1/50C. For your 95ah cells, 1/50C would be 1.9 amps per cell, meaning your 1A per cell was quite safe, in my opinion. If you parallel all 20, you could give it 38 amps. In any case, charging all 20 cells from below 3V at 1.9 amps per cell would take 50 hours. So maybe give them just 1A per cell, 20amps if you parallel all 20, and let it take 100 hours. That's only 4-5 days. 10 days, if you stop charging them while you sleep, which I would recommend.

Also:

2V to 3V @1A (they are 92Ah cells). The first one got there in 30 mins

This only happens because the discharge curve (for both Li-ion and LFP) drops off hard after 3v. So very little ampacity is required to bring them back over 3v. This is why I recommend doing the full charge at 1/100C if the cell is questionably damaged, and if you're attempting recovery.

 

[amberwolf]

Assuming that's a 5S section of cells in the image (based on the arrangement of posts, and shape of the blue casings), those are pretty physically small for 92Ah. Based on the probes (5"?) and alligator clips (2"?) these cells look like they are about 4" tall (not counting terminals), 2" thick, and 8" long.

Do you happen to know the manufacturer and model? I'd be interested in something that much smaller (and presumably lighter) than the ancient EIG C020 cells I'm using, which are about a pound per 20Ah cell, not including the busbars and housings. https://www.researchgate.net/figure/Specifications-of-EiG-ePLB-C020-Battery-Cell_tbl1_356718687 My 14s2p 40Ah pack is around 35lbs+ and something like 6-7"x9-10"x9-10" at a guess.

If my size guesstimate of your cells is correct, then a pack about 28" long (14s x 2") x 4" x 8" would give me more than twice the capacity I have now for not much more volume (not including the compression setup). (900 cubic inches vs 700 cubic inches). If I make two 7s packs to "stack" it's more like 14" x 10" x 8" (giving space for posts and insulation between the two stacked packs).

Since I use 14s instead of 20s, then assuming the same per-cell cost, it ought to work out to around $900 / 20s = $45 x 14s = $630. (plus whatever busbars and compression stuff I make or buy).



Also, what constraining / compression setup are you using on the cells to keep them from swelling during high-rate charge/discharge? (all of the non-cylindrical cells I've seen are designed around this being provided by the pack construction method; cylindrical cells do this with their casings by nature).

 

[lord_gammon]

Assuming that's a 5S section of cells in the image (based on the arrangement of posts, and shape of the blue casings), those are pretty physically small for 92Ah. Based on the probes (5"?) and alligator clips (2"?) these cells look like they are about 4" tall (not counting terminals), 2" thick, and 8" long.

Do you happen to know the manufacturer and model? I'd be interested in something that much smaller (and presumably lighter) than the ancient EIG C020 cells I'm using, which are about a pound per 20Ah cell, not including the busbars and housings. https://www.researchgate.net/figure/Specifications-of-EiG-ePLB-C020-Battery-Cell_tbl1_356718687 My 14s2p 40Ah pack is around 35lbs+ and something like 6-7"x9-10"x9-10" at a guess.

If my size guesstimate of your cells is correct, then a pack about 28" long (14s x 2") x 4" x 8" would give me more than twice the capacity I have now for not much more volume (not including the compression setup). (900 cubic inches vs 700 cubic inches). If I make two 7s packs to "stack" it's more like 14" x 10" x 8" (giving space for posts and insulation between the two stacked packs).

Since I use 14s instead of 20s, then assuming the same per-cell cost, it ought to work out to around $900 / 20s = $45 x 14s = $630. (plus whatever busbars and compression stuff I make or buy).



Also, what constraining / compression setup are you using on the cells to keep them from swelling during high-rate charge/discharge? (all of the non-cylindrical cells I've seen are designed around this being provided by the pack construction method; cylindrical cells do this with their casings by nature).

Hi, they are 3.65V Li-ion NMC cells, commonly available on Aliexpress or Alibaba. I bought from Aliexpress store Deligreen and they offer many sizes depending on your needs. I stick them together with industrial double sided tape, which holds them fast together but is spongy to allow for some sideways movement.

 

[nicobie]

I would not trust any abused Alibaba cell that hasn't been compressed to be inside my house or shop, much less charge them there.

Could you replace your home for $900?

 

[amberwolf]

I stick them together with industrial double sided tape, which holds them fast together but is spongy to allow for some sideways movement.

The pack framework itself should secure the cells from being able to move relative to each other, and perform the compression these cells are designed for.

FWIW, you shouldn't have (allow) any movement of any cell relative to any other. If there is movement, there is then possibility of stress on the terminals or the interconnects (whichever is stiffer will then flex the other). Tiny flexings over time add up and can cause stress fractures that increase resistance of the interconnects, or even disconnect a cell or interconnect entirely.

You should also not have any spacing between the cells (like with foam tape) because these kinds of cells are designed to be compressed evenly across their surfaces, and if they are not then the layers inside can form gas bubbles during charge and/or discharge, which will affect the cell's resistance, capacity, etc. Normally the pack is installed in a compression framework that may look like these:

 

[lord_gammon]

The pack framework itself should secure the cells from being able to move relative to each other, and perform the compression these cells are designed for.

FWIW, you shouldn't have (allow) any movement of any cell relative to any other. If there is movement, there is then possibility of stress on the terminals or the interconnects (whichever is stiffer will then flex the other). Tiny flexings over time add up and can cause stress fractures that increase resistance of the interconnects, or even disconnect a cell or interconnect entirely.

You should also not have any spacing between the cells (like with foam tape) because these kinds of cells are designed to be compressed evenly across their surfaces, and if they are not then the layers inside can form gas bubbles during charge and/or discharge, which will affect the cell's resistance, capacity, etc. Normally the pack is installed in a compression framework that may look like these:
View attachment 334166View attachment 334167View attachment 334168

Thanks for the time to advise.