Charge/discharge voltage conundrum

In an effort to eliminate bad cells, we usually charge and let it rest for a few days or more.
Then check to see if the voltage has dropped.
We reject cells that have dropped below 4.1 (for example, you can reject at whatever voltage you want)

The problem lies in the TP4056 chargers.
Sometimes the 4056 stops charging,and I pull the cell off the charger, and it will be at 4.1v
Other times, it will stop charging at 4.18.
Sometimes if you put in a cell that is at say 4.05, the TP4056 will not even start to charge.
I assume this is due to varying internal resistance of the cells.

So lets say I pull 2 cells off the charger and put in the "charged" box.

I come back a week later to discharge test.
I put the cells in the tester, both say 4.09v

BUT, one came off the charger at 4.10v and the other at 4.19v.
So one has dropped by 0.01v and the other 0.10v, ten times as much voltage drop.

Which is the healthier cell?

So if I didn't write the fully charged voltage on the cell (I dont), I have no way of knowing how much of a drop really occurred.

Also, I can often take cells that only charge to 4.1v and put them on my nitecore charger, and take them up to 4.19.

Now I know that the 1/10th of a volt means little in terms of cell capacity, but it does mean a potential voltage draw when assembled in a pack.

How do you guys handle this?

When assembled into my powerwall, I wont ever charge above 4.1 (for longevity), so does it even matter?

Also, what is your "reject voltage"?

Use a charger that give consistent results instead. Im not using TP charger for testing at all and therefore i have not seen this issue. I pree-charge and then i let opus to charge-discharge-charge. Then i save the first charge cycle.

Above problem you have is huge if you want to check the self-discharge over time since you cant for sure tell if the charge was done fully or not.
Thing is that it need to stop charging at same voltage and at same current every time. Otherwise the voltage drop after 1day or 1 week is not that useful. A cell that is stopped directly or close when it hit 4.2V will drop alot more than a cell trickled untill current/10 or whatever is reached.

My opus chargers arent that accurate either but i always check voltage after an hour or a day and if they are just under 4.15 after 1 day i retest. And thats not often i get that i can say so they do charge the cells up pretty decent.

So that's how i solve or actually go around above issue you have seen.

One potential low tech solution, rack your charged cells in fixture that will put them in parallel for a day to equalize the voltages, next day check the voltage on the paralleled cells and then remove them for leak testing.

I suspect a lot of these ultra cheap chargers and such are made with counterfeit chips in the first place, that shows up even on more expensive equipment on Amazon, more astute customers are pointing out how one unit is supposed to work at 150kHz and is actually running at 50kHz and jumping around on frequency.

From what I had learned so far:
- 4056 modules from ebay aren't the precision tool. I use 4 of them, and the "charged" voltages are in the range from 4.14 - 4.18 V. The current at the end of charge is about 0.05 A and rather consistent.
- I relay on two measurements - first is done about an hour or two after the cell is charged and picked up from the charger (it must "settle" after the charge). I write it down. Second measurement is made after 2-3 days. Really bad cells show about 4.08 or less at the first check. They are discarded. From the remaining very few drop under 4.05 at second test - they also go to bin together with those which lost more than 0.05 between the tests.
The ones I am going to use in banks are about 4.12 - 4.15 at first test and loosing 0.01 - 0.02 till second. Cells with lower specs will wait for flashlights and few small projects (WiFi sensor etc.).
- Each cell before capacity test is topped by imax B6 clone to 4.20.
- And I think it matters because, if a cell is already discharging at about 4.2 or 4.1, than the deterioration has started and it's likely the evil process will appear at 3.9 or even less voltage sooner.

Some automation will follow - this process takes some time, but I liked it, because I learned a lot about the behavior of cells. As my first goal is to have 100 "good" ones (one 4s5p and one 4s20p pack), it's not eternity. So far have 56 from about 320 checked.

One potential low tech solution, rack your charged cells in fixture that will put them in parallel for a day to equalize the voltages, next day check the voltage on the paralleled cells and then remove them for leak testing.

Click to expand...

All this will do is bring them down to the same lowest voltage. It wont tell me which one causes the voltage drop. This would make the problem worse!

APD said:

One potential low tech solution, rack your charged cells in fixture that will put them in parallel for a day to equalize the voltages, next day check the voltage on the paralleled cells and then remove them for leak testing.

Click to expand...

All this will do is bring them down to the same lowest voltage. It wont tell me which one causes the voltage drop. This would make the problem worse!

Click to expand...

The leak testing proceeds after the voltages are all equalized, then all you need is one go/no go voltage reading rather than a different one for each cell.

Elmo said:

APD said:

One potential low tech solution, rack your charged cells in fixture that will put them in parallel for a day to equalize the voltages, next day check the voltage on the paralleled cells and then remove them for leak testing.

Click to expand...

All this will do is bring them down to the same lowest voltage. It wont tell me which one causes the voltage drop. This would make the problem worse!

Click to expand...

The leak testing proceeds after the voltages are all equalized, then all you need is one go/no go voltage reading rather than a different one for each cell.

Click to expand...

Im not following your method?
By connecting them together, it brings them all down. I have thus lost the ability to distinguish which one brought them down. And if I test for that afterwards, there was no reason to connect them in the first place.

Please explain your "Leak test"

Elmo said:

One potential low tech solution, rack your charged cells in fixture that will put them in parallel for a day to equalize the voltages, next day check the voltage on the paralleled cells and then remove them for leak testing.

Click to expand...

That's the best way IMO ... the small chargers don't all charge to the same voltage ... I'm working ona system that charges cells up individually , then putting them in parallel by throwing a switch (one for each cell) , they don't need to be taken from cell holder ... leave overnightforthem to balance , and get over the natural fall in voltage that happens after charging .... then measure the voltage of the batch still in parallel if over 12hrs it drops by more than 0.02Vthen it might be worth finding the culprit in the batch.

No point in doing the above at 4.2 or 4.1 ... we need to know the self discharge at 3.9 , our maximum operating voltage...

ozz93666 said:

Elmo said:

One potential low tech solution, rack your charged cells in fixture that will put them in parallel for a day to equalize the voltages, next day check the voltage on the paralleled cells and then remove them for leak testing.

Click to expand...

That's the best way IMO ... the small chargers don't all charge to the same voltage ... I'm working ona system that charges cells up individually , then putting them in parallel by throwing a switch (one for each cell) , they don't need to be taken from cell holder ... leave overnightforthem to balance , and get over the natural fall in voltage that happens after charging .... then measure the voltage of the batch still in parallel if over 12hrs it drops by more than 0.02Vthen it might be worth finding the culprit in the batch.

No point in doing the above at 4.2 or 4.1 ... we need to know the self discharge at 3.9 , our maximum operating voltage...

Click to expand...

So it only saves you time if you DONT find a voltage drop.
If you do have a drop, you still have to individually examine each cell in the parallel group.
And you will have a drop in voltage.
All cells drop a bit after full charge.

Putting them together in parallel only makes it harder to determine which one is the problem cell.
Seems best to check each individually, as you will have to do that anytime you find a drop in voltage.

You only put the cells in parallel temporarily for 24 hours, after that they will all be at the same voltage. Let them sit for a week or longer disconnected and any that drop significantly from the value they all had when disconnected are leakers.

The point is to make the later voltage measurement easy by having a single go/no go voltage rather than having to do each cell individually according to what voltage it happened to terminate charge at.

It takes longer in hours but it saves actual working time.

Elmo said:

Let them sit for a week or longer disconnected and any that drop significantly from the value they all had when disconnected are leakers.

Click to expand...

You left this part out, now I understand your method. thanks