Low current charging hack - what am I missing?

[harrisonpatm]

First, really enjoy reading about your test results!

But I had a thought - e.g. I didn't read your discharge rate (forgive me if I missed it). To reiterate consistent advice from @daromer (one of key members of this forum) it's good to discharge test your cells at >= ma that you plan for the overall battery / use you're planning. This will reduce surprises in use behavior under load. 250ma / cell might be a bit light - but maybe that's your battery design.

A lot folks do 1000ma for discharge/capacity tests because it's standard OPUS and faster. I do 500ma because my OPUSs overheat and my battery use averages <200ma/cell.

I didn't note my discharge rate for capacity tests (it's 750, goldilocks zone apparently), because that's not the interesting result i was sharing, ie, irrelevant to the high success rate for these low voltage recovery. But you make a good point that I'll clarify: after the 25ma charge rate for 5-8 days, and the 30 day rest to weed out self dischargers, I used the same discharge rate for capacity tests that I use for all cells (750ma). Which is to say, the result of this low volt batch of cells being a better yield and better capacity than other categories is not a result of different method of capacity testing. If that makes sense?

 

[Oz18650]

Regatding your finding that low voltage cells recover and test well:
My guess would be that the cause of these cells being low voltage has not been that the cells being bad. Maybe something else, such as a bad bms.
I have heard of BMS which constantly.take power and can draw cells down.
Then, the cells have gone low voltage and the battery does not work any more - before the cells have done many cycles and have lost capacity through getting worn out.
Then, when you slowly charged the cells, they work again and have decent capacity.
This is just my own ideas - completely theoretical and backed by almost no facts beyond your description of what you are finding.
Personally, I have a different experience (from a pretty small sample space).
My cells are almost all one model, from one source.
With these cells, i have tried reviving low voltage cells a few times with a tp4056 connected to 8 cell holders in parallel.
I have only tried filling the 8 places 4 or 5 times, but each time, charging seemed to stall. I think this was due to a heater wasting whatever energy goes into the cells. On removing the wam cell charging seemed to continue, until another cell got warm.
I have put the low voltage cells aside and may try further at a later stage.
This will involve IR testing to see if that helps narrow down which cells may still be useful for something other than practicing spot welding (eg torches).

 

[paddy72]

Imho you cannot use just one TP4086 for more than one 18650 cell at a time. In the IC there is a built-in charge process with low current until the cell reaches 3.0 V. After that the charge current is increased up to 1 A depending on the little SMD resistor controling the charge current. The TP4086 is already at its limit at 1 Amp - you should better go for less. The chip gets quite hot at 1 Amp continous - so its better to use a little heatsink. I mounted the plates on an aluminiumstrip.

 

[Redpacket]

Imho you cannot use just one TP4086 for more than one 18650 cell at a time. In the IC there is a built-in charge process with low current until the cell reaches 3.0 V. After that the charge current is increased up to 1 A depending on the little SMD resistor controling the charge current. The TP4086 is already at its limit at 1 Amp - you should better go for less. The chip gets quite hot at 1 Amp continous - so its better to use a little heatsink. I mounted the plates on an aluminiumstrip.

I think it is actually OK to do this with a TP4056 - the TP4056 limits the total current to 1A & this is shared between cells.
Yes putting a heatsink on the chip is a good idea (he did that already)+ maybe instead using the more efficient TP5100 chip instead.

 

[paddy72]

If you charge multiple cells with one TP4056 you don't have any control of the individuell current per cell! I think this is the wrong way to save money as the IC is quite cheap. You can provide the same power source to multiple TP4056 tough without problem, just make sure the input voltage is in the specified range. They don't like overvoltage on their input and are easily destroyed. Happened to my with a badly controlled PSU, due to a bad potentiometers (cheap 10-turn pot from china...).

But i destroyed probably 10 of the Chips by wrongly inserted cells

 

[harrisonpatm]

If you charge multiple cells with one TP4056 you don't have any control of the individuell current per cell!

1000ma/40cells=25ma per cell. Give or take. For this application, I don't need any more control than that.

 

[Korishan]

If you charge multiple cells with one TP4056 you don't have any control of the individuell current per cell!

This would really matter at higher current charging rate. When talking about 10's of mA, then parallel grouping isn't going to make much of a difference. Unless the IR is sooo high that it burns the energy as heat, then it won't make a difference. But as Oz18650 mentions, these are heaters anyways and can be removed from the chargers by noticing the heat increase.
A good way to weed these out is to do an IR test first, and you'd know before charging if that cell was going to be a dud, even at low current.

If you are really wary of this, then test all your cells IR values, and then batch charge the ones that very close values, or even the same values. Then those cells will charge at the same rate, regardless.
In any case, the cells will self-regulate the current between them as a cell will only accept a certain amount of current at a certain voltage, and a cell in parallel with others cells can not have a different voltage than the others while connected to the others.

 

[paddy72]

Yes, thats right, you won't hurt the cells with that little current and the TP4056 doesn't care or know if it charges just one cell or 4 or even 10 of them. Anyway i use these ICs to fully charge my cells and it is programmed to start with low current (max. 100 mA) below 3.0V and increases the current up to 1 A (i prefer 0.7 to 0.8 A max.) if the voltage climbs. It just takes me too long to charge 10 cells in parallel, although it might be even less stressy for each cell.