One parallel group of cells in a Lithium batterh pack discharged quicker after fully charged.

I built a few large 18650 Lithium battery packs recently, one using smart BMS and the rest using no intelligence BMS. for one battery packs using no intelligence BMS, one parallel group of cells discharge quicker than the other parallel groups of cells (when there is no load applied). for example, when fully charged each parallel groups of cells to 4.19v, one day later, the voltage of the quicker discharged parallel group of cells decreased to 4.13v (decreased about 60mv in one day), while the voltage of the other parallel groups of cells decreased to 4.18v (decreased about 10mv in one day).

when assembling, there wasn't short circuit occurred for this quicker discharge parallel group of cells, while there was loose soldered wires (balance connector cable) discovered for this parallel group of cells when the whole battery pack assembly was finished and was being charged, then I re-soldered the wire for this parallel group of cells, then found out that this parallel group of cells discharge quicker (when no load applied) than the other parallel groups of cells.

don't exactly know why this parallel group of cells discharge so quick. could this parallel group of cells make impact to the whole battery pack when used in ebike application? is there a way to repair it?

all the other battery packs assembled are in good condition.

What cells are we talking about ? Where'd you get them. Your battery is only as strong as your weakest parallel group of cells.

You have one or more cells in that group damaged and that's eating current slowly. To fix you need to find out which ones and replace, or replace the entire group.

when I had a similar problem, I had to take the pack apart. The problem is usually one cell that self discharges too quickly. You do need to fix it otherwise the pack will always have balancing problems.
Problem is usually easy to find, charge all cells to 4.20 volts, after 1 hour test the voltage on all cells and you will find the problem cell.

A.W. did a thread where you isolate that Parallel Group. Cut that Parallel Group in half and see what the weak half is. Then cut that parallel group in half and find thw weakhalf and so on. Could be one or ? Cells

This will always happen to some extent.

Starting with good quality brand-new cells, maybe only after SoH has declined past the EoL point anyway, had a good life, just toss it out and replace with a new pack.

Starting out with scrapped ones to start, obviously will start much much sooner,

if you haven't done thorough capacity load testing and matching before then, likely right from the beginning like here.

The ability to monitor and test per-cell after bank assembly while in use will help.

Non-weld no-solder pack construction makes replacing cells much easier.

But so many people just want to do things the "normal", as-cheap-as-possible way, to me it just makes no sense.

999zip999 said:

What cells are we talking about ? Where'd you get them. Your battery is only as strong as your weakest parallel group of cells.

Click to expand...

these are all brand new cells, i just finished assembling them with BMS and balancing, charging them fully and leaving them there for a few days or a couple of weeks to see their self discharge. haven't installed them in the ebikes for load test yet.

for this particular group of cells in one of the packs, after fully charged the pack (with BMS plus balance), i measured the voltage of each group of cells with multi-meter (i installed an external connector, similar as the balance connector, to each pack, for monitoring the pack status in the future), and noticed this particular group of cells only has 4.03v (and the other groups of cells all have 4.19v/4.20v), then i examined the pack and noticed the wire soldering for this group of cells was loose, then i re-soldered the wire, then used a very simple charger (4.2+v) i modified to particularly charge this group of cells to bring its group voltage to 4.19v (as when all the other groups of cells fully charged, the installed BMS couldn't bring the voltage of this particular group of cells to the fully charged voltage, say 4.20v, or it may take very long time to do so, so i used a simple charger i modified to quickly charge this group of cells to about 4.19v), however one day later its group voltage dropped to 4.13v, then the 2nd day dropped to 4.08v (the self discharge of the other parallel group of cells are all normal).

could the loose wiring of this parallel group of cells cause this problem when charging this group of cells with BMS plus balancing? if its wiring hadn't been loosed, then it wouldn't have caused this problem when charging with BMS (or with the simple charger i modified)?

i think it might be the loose wiring initially cause the problem for this parallel group of cells when charging (the battery cells might not have problem initially), just don't how this problem could be introduced.

A lesson to learn is to make sure the wire tightly connected when soldering.

jonyjoe303 said:

when I had a similar problem, I had to take the pack apart. The problem is usually one cell that self discharges too quickly. You do need to fix it otherwise the pack will always have balancing problems.
Problem is usually easy to find, charge all cells to 4.20 volts, after 1 hour test the voltage on all cells and you will find the problem cell.

Click to expand...

these are all brand new cells, i just finished assembling them with BMS and balancing, charging them fully and leaving them there for a few days or a couple of weeks to see their self discharge. haven't installed them in the ebikes for load test yet.

for this particular group of cells in one of the packs, after fully charged the pack (with BMS plus balance), i measured the voltage of each group of cells with multi-meter (i installed an external connector, similar as the balance connector, to each pack, for monitoring the pack status in the future), and noticed this particular group of cells only has 4.03v (and the other groups of cells all have 4.19v/4.20v), then i examined the pack and noticed the wire soldering for this group of cells was loose, then i re-soldered the wire, then used a very simple charger (4.2+v) i modified to particularly charge this group of cells to bring its group voltage to 4.19v (as when all the other groups of cells fully charged, the installed BMS couldn't bring the voltage of this particular group of cells to the fully charged voltage, say 4.20v, or it may take very long time to do so, so i used a simple charger i modified to quickly charge this group of cells to about 4.19v), however one day later its group voltage dropped to 4.13v, then the 2nd day dropped to 4.08v (the self discharge of the other parallel group of cells are all normal).

could the loose wiring of this parallel group of cells cause this problem when charging this group of cells with BMS plus balancing? if its wiring hadn't been loosed, then it wouldn't have caused this problem when charging with BMS (or with the simple charger i modified)?

i think it might be the loose wiring initially cause the problem for this parallel group of cells when charging (the battery cells might not have problem initially), just don't how this problem could be introduced.

A lesson to learn is to make sure the wire tightly connected when soldering.

john61ct said:

This will always happen to some extent.

Starting with good quality brand-new cells, maybe only after SoH has declined past the EoL point anyway, had a good life, just toss it out and replace with a new pack.

Starting out with scrapped ones to start, obviously will start much much sooner,

if you haven't done thorough capacity load testing and matching before then, likely right from the beginning like here.

The ability to monitor and test per-cell after bank assembly while in use will help.

Non-weld no-solder pack construction makes replacing cells much easier.

But so many people just want to do things the "normal", as-cheap-as-possible way, to me it just makes no sense.

Click to expand...

all the cells are brand new, please refer to my latest reply.

for your view "The ability to monitor and test per-cell after bank assembly while in use will help", when the pack is assembled, only can monitor and test each parallel group of cells (i installed an external connector, similar as the balance connector, to each pack, for monitoring the voltage status of each parallel group of cells in the future), how could it be able to monitor and test per-cell in each parallel group? please let me know.

You still have not stated the source, maker, model of cells.

There are tons of ripoff scammers and counterfeiters out there.

As to how to design and build so each cell is accessible to be independently monitored,

that would require bringing per-cell wiring out to a buss-connector setup so the parallel connections can be broken. Or serialing your cells into high-voltage sub-packs with balance connectors as is common in the RC world, then paralleling those packs together to build capacity.

Obviously that goal is impossible using the build methods that do not allow that, like hard-wiring connecting the cells together with welding or solder.

As is easily breaking the pack apart for replacing bad cells.

Individually capacity load testing each cell before assembly is basic in any case. How else will you identify the faulty ones to leave out?

john61ct said:

You still have not stated the source, maker, model of cells.

There are tons of ripoff scammers and counterfeiters out there.

As to how to design and build so each cell is accessible to be independently monitored,

that would require bringing per-cell wiring out to a buss-connector setup so the parallel connections can be broken. Or serialing your cells into high-voltage sub-packs with balance connectors as is common in the RC world, then paralleling those packs together to build capacity.

Obviously that goal is impossible using the build methods that do not allow that, like hard-wiring connecting the cells together with welding or solder.

As is easily breaking the pack apart for replacing bad cells.

Individually capacity load testing each cell before assembly is basic in any case. How else will you identify the faulty ones to leave out?

Click to expand...

there are plenty of Lithium cells out there for different type of applications. for RC application, it use different type of cells.

for my ebike application, the current assembly design and packaging of mine is quite appropriate, and the process is slightly different from the other type of applications. examine each cell before assembly is the very important 1st step. for ebike application, there is no need to independently monitor and test each cell after assembly, but for EV, RC application, it is good to use buses-connector setup.

last night, disassembled the battery pack, cut off the nickel strips for the parallel group of cells with bad cell. let this parallel group of cells self-discharge to identify the bad cell, removed the bad cell, prepared to replace it with a good cell, and tonight just finished re-welding this parallel group of cells with new nickel strips.

repairing battery pack to remove the bad/degraded cells is the have to do task at some stage of the battery pack life cycle, however careful examining the cells before assembling can prevent it from occurring too early.

Yes, but "examining" should mean proper capacity load testing.

Oh I get it Tesla like cells.

Those must be some embarrassingly bad cells if he's not willing to name them. Or just doesn't know.

doesn't want to confirm he got ripped off

I got ripped off when I bought some Samsung 25r cells. In recycle now. It's like in the 8s - 90s chip saling Heise. If you can tell us your vendor your cell and your model number maybe you can save other people from having problems or maybe we can help you match your cell to your amp needs for a battery pack that will work for your amp to needs and motor
Or more info for fine tunning. Help me help you


Yes a lot of talks text so figure it out.