Battery Changing issue

So you bought old dead batteries off ebay and are running into issues? Strange.

The battery is fine, but the BMS is strange. I just don't understand how the BMS could detect that i am charging the battery behind it's back. The battery is only going up a few mV per hour.

The ebay posting makes it clear these are end of life, thrown away, then pulled from a dumpster and being resold type deals. If the batteries are perfect like you seem to think, why do you care about the protection circuity thrown in?

I am just trying to understand how the BMS works. For output, the 20A short protection and 30A fuse is good to have. I can charge the raw battery directly if necessary. More convenience if charge through the BMS, with balancing.

The battery should have around 70% to 80% of original capabilities.

SOC logic is dependent on preconditions and alarms in the BMS.

The state machine may require a higher than 11V input to permit charging after a low voltage cutoff. Post-high-current discharge cut-off float is >11V for healthy cells. 10V self-float suggests issues, if measured across actual cells.

Blinky lights can mean anything if the logic is bypassed, as you've done - may simply indicate that charging switch is set.

A 12V LiFePO4 pack has 4 series sections. Undervoltage cutout between 10 and 11.2V - earlier if one cell goes lower than

2.8V and the BMS is watching (not all do).

End of charge will be <16V. Full charge self-float ~13.8.

Overcharge will severely reduce capacity, permanently. Overdischarge will form gas - forming tell-tale bloat and degrade capacity.

RL

It's 3S9P of Regular Li cells. Monomial voltage of 10.8V. So, 11V (that's what i have for the moment) should be able to start charging. I will try to get a higher voltage supply.

Some packs indicate that they cannot be charged, perhaps with some error codes internal to the BMS. However, bad BMS might mean better cells, since they were cycled less.

Bullshit. You bought garbage somebody threw away, which was likely diverted from proper recycling. Buying used batteries is always a fool's errand.

The cell pack itself is good. Bypassing the BMS, I have it charged from 9.6V (20%) to 10.7V (50%) in three days. Someone mentions that the BMS config needs to be reset via I2C. Plan B is to use external BMS, or not at all.

Markings on the BMS did not turn out any result. I guess i have to try every known BMS I2C possible.

It's better to reuse the battery first than to recycle.

Says Lithium Iron in the ad. That's one creepy ad if what you say is true.

RL

It's 3.7V Panasonic 18650 cells. One of the best cell as in the original Tesla.

<snip>

In that case, I suggest you ignore anything in the ad, inspect for rewelded tabs, then charge-discharge every individual cell to establish their capacities.

You can then try to rebuild a pack using known-good cells with some chance of predetermining equal capacity balance between the '9P' blocks.

RL

Yes, it's very popular for people to take out the cells to rebuild balanced pack. But why? I can just use the pack as is. Even if it's unbalanced, same amount of energy goes in and out. I just have to pump less energy into the less efficient cells, but no loss of energy overall.

If i re-charge each 3.7V cell group separately, I don't really care if they match or not.

If your BMS is smart, you'd get away with it at the capacity of the most degraded group.

If the BMS is not smart, you just agravate the imbalance through series overcharge/overdischarge of the lowest cap group.

A few things the ad said that made sense - 'as-is for parts' and 'usable for flashlight'.

I'd look up the panasonic cell type, stamped on the case;

18650 is only the package, not the chemistry.

might help in prolonging life of what's there.

RL

It doesn't take 3 days to charge a good battery.

I am slow charging 8 packs in series. Around 100V, 24S9P at 0.05C.

Worst possible way to assess or to nurture remaining capacity.

RL

The target will be in series, but hooking them up all in series run the risk of unbalanced cells or modules. The BMSs are unpredictable. Using a 13v power source, half of them work correctly. Namely, voltage went up slowly until cut-off voltage of 12v or so. At cutoff, it can still source at least a 20mA LED meter, but charging immediately bring it up to 13v. Unfortunately, the cutoff voltage can be anything: 10.8v, 11.4v, 11.8v or 12v.

The BMS might need to be reprogrammed via I2C. Until them, I just have to by-pass them and take a chance.