Reviving a Dead Battery

[Deleted member 4366]

I was given one of those very common 36v behind the seatpost silver fish batteries to have a look at. It wouldn't take a charge, so was effectively dead. I measured the voltage at only 17.8v at the charger socket. No wonder it wouldn't charge.

To strip it down is pretty easy: 4 screws top and 4 screws bottom, and then there's two screws that hold the connector block down at the bottom. Once removed, you can slide the entire contents out of the case leaving the top attached by its wires. You need to get to the balance wires, where they connect to the BMS to measure the individual cell voltages. This will give a good idea what's wrong. The BMS won't allow charging if any cell is below a certain level - probably about 2.5v.

This one was shrink-wrapped, so that had to be cut off, and then it had thin fibreglass panels stuck to each side with double-sided tape, so these had to be prised off to get to the cells. It always looks a bit of a mess, but you can normally put it back fairly easily and use gaffa tape instead of shrink-wrap. This one had a multi-pin connector into the BMS. To measure cell voltages, you start at one end and check between adjacent pairs of holes. Some batteries use the main wires as the last one, so check that the voltage from first to last is the same as the voltage at the main socket. or check between the lowest and the main black and the highest and the main red. for a 36v Li-ion battery, there's normally 10 cells. and 13 for a LiFePO4.

I measured the cell voltages:
0.2, 0.6, 2.7, 1.2, 1.2
1.3, 2.5, 2.5, 2.7, 2.0

Most people would at this stage consider the battery scrap, but all is not yet lost. Most chargers won't do anything with cells that low, but there'es a trick you can use. You need a charged up single spare cell. If you connect it in parallel to a low group by poking wires directly into adjacent holes in the multi-pin connector for just a few seconds, it'll transfer a minimal amount of charge, which is all it takes to get the cell up to about 2.5v; however, it rarely works if the cell is down to zero volts. It worked on these, so I was able to get them all high enough for my lipo charger to take over. Mine can charge up to 6 cells individually.

My lipo charger is very versatile. It can charge single cells in groups directly through the balance leads. Here's a photo of the battery innards showing the BMS and the multi-pin connector. I made an adaptor that can charge 6 cells at a time. The single cell is in the foreground:



A close-up of the BMS and my adaptor in the multi-pin balance lead connector:

 

[Deleted member 4366]

I can only charge 6 cells at a time, so two half hours later, all the cells were balanced and at 3.34v, an arbitrary low figure I chose. I could've charged them all the way to 4.2v, but I wanted to see that a normal charger would work.



Here's the charger charging nicely now. Don't forget to plug the balance lead connector in before using the charger. It's been going for about an hour and it's up to 36.7v, so I've every confidence that this battery will be restored to usefulness. i'll test the final cell voltages before re-assembly, and I'll leave it over night just to check that they don't self discharge because I still don't know the reason why it was down, but at a guess I'd say that it was left in a prolonged discharged state. This is why I recommend that you don't fully discharge your battery.

 

[Deleted member 4366]

I forgot to mention that this one has 13 balance wires to the BMS, but the last three were joined together. Presumably this BMS is also used for LiFePO4 batteries. I wonder what the cut-off voltage is because LiFePPO4 is normally 2.0v - much lower than these should have. I bet the bike's relying on the controller for low voltage cut-off. I might test it later to see. My Turbo Trainer makes a great test rig because you can adjust the load on the wheel.

 

[Alan Quay]

Great writeup D8veh, very useful indeed.

 

[Deleted member 4366]

At 21:00 it was still charging at 41.5v, and now it's just finished. Cell voltages are:
4.15, 4.15, 4.17, 4.16, 4.16,
4.16, 4.16, 4.16, 4.17, 4.17.
The charger was set to 41.7v, which is 4.17v per cell, so that looks pretty good. I'll check again in the morning before doing the discharge test.

I'm not sure 4 1/2 hrs is enough for a 15ah battery.

 

[Oxygen Bicycles]

Genious post

 

[manalog]

Great Post, thanks for sharing.

 

[Deleted member 4366]

After, my success with this battery, somebody sent me a couple more to look at. These ones were both not accepting charge and showing 0v on the main connector, but some volts on the charge socket.

I opened them both up and measured the cell voltages on the multi-pin BMS connector. One showed all cells at a healthy 3.9v except one at 0v. The other showed 7 cells at 3.8v, two at about 2,5v and one at 0v volts, so further investigation was required.

This is what I found:
Battery 1 corner cell squashed during manufacture and now leaking.


Battery two had 5 leaked cells that corresponded with the low strings that i measured.


So two scrap batteries - or are they?

 

[Deleted member 4366]

As battery 1 had a faulty string right at the edge, I figured out that I could cut it off and replace it with one of the good strings from the other pack,
Here's cutting it out leaving the remaining tags as long as possible. I did the same with the other strip but leaving the tags as long as possible on the string I cut off:


Soldering the overlaps:

 

[Deleted member 4366]

I plugged in the BMS multi-pin connector. Now there's a nicely balanced 39v on the main connector. I plugged in the charger, and now it's charging happily.

Here it is charging and looking like nothing was ever wrong with it.


Not everybody will have a spare battery lying around to do this repair,but as long as it's a faulty string in the edge, you could replace the cells in the same way with bought cells.

 

[Zebb]

Great thread thanks for writing it up. Just goes to show what can be done.

 

[morphix]

Imagine how many good batteries end up in landfill or wherever they end up?! You could start a mini-industry here recycling them!

 

[trex]

problem is often the BMS gets damaged when you go over bumps. There is a need for someone (d8veh?) to design replacement protection switches. They don't seem to be available on aliexpress/ebay.