Charge vs Discharge capacity

So I'm testing~200 cells that I harvested from laptop packs, and I came across some data that I think can be useful to sort bad/good cells out, and I'm planning on getting some feedback from you guys.

So this is it:
step 0) charge to 4.2v, wait for two weeks and take out <4.15v cells.
step 1) Discharge some cells down to 2.9v with a low current, just so that they only bump up to 3v after I stop discharging. This is to "ensure" that all cells were discharged to the same voltage/level, because some cells kept bumping back to 3.4, 3.2, 3.7v after discharging at like 0.5A.
step 2) Charge them up to 4.2v, write down the mAh. For this I am using a Liitokala Lii-S6 (bought by mistake, it's only a charger and will not do discharge tests).
step 3) Do a NOR test on a Liitokala Lii-500, discharging at 0.5A (choosing 1A charge rate at the display)
step 4) Compare the charge and discharge capacity of the cells.

Sample values from some of the cells I have:

Kind of cell discharge |charge | resistance | percentage of charge that went out vs in
Red Sanyo 2236| 2393 | 15 | 93,44%
Pink Samsung 2323 | 2493 | 20[size=small][size=small] [/size][/size] | 93,18%
Red Sanyo 2254 | 2369 |33[size=small][size=small][/size][/size][size=small][size=small][size=small][size=small][size=small] [/size][/size][/size][/size][/size] | 95,15%
Red Sanyo 2205 | 2374 | 29 [size=small][size=small][/size][/size][size=small][size=small][size=small][size=small][size=small] [/size][/size][/size][/size][/size]| 92,88%
Blue LG 2207 | 2337 | 38 [size=small][size=small][/size][/size][size=small][size=small][size=small][size=small][size=small] [/size][/size][/size][/size][/size]| 94,44%
Blue LG 2133 | 2442 | 125 [size=small][size=small][/size][/size][size=small][size=small][size=small][size=small][size=small] [/size][/size][/size][/size][/size]| 87,35%
Grey LG 1709 | 2122 |90 [size=small][size=small][/size][/size][size=small][size=small][size=small][size=small][size=small] [/size][/size][/size][/size][/size]| 80,54%
Red Sanyo 1665 | 2293 | 63 [size=small][size=small][/size][/size][size=small][size=small][size=small][size=small][size=small] [/size][/size][/size][/size][/size]| 72,61%

Also, the Lii-500 measures resistance, but in a very unprecise way (same battery gives a lot of random readings if you just keep taking it out and putting it back in, same slot). I also know that resistance is exactly what I'm talking about, the cell needing more charge than it's able to discharge. I see that the higher spectrum of the resistance values tend to be related to these "bad" cells, just not very consistently.

I'm assuming these three last cells will be harmless to a pack, in the sense that they take a lot more charge to reach the same capacity. I am thinking that this can throw down all the "math" on tools like repackr and lead to unbalanced/unmatched packs.

I am planning on using a cell balancer on this 7s pack, with a cheap 7s bms also, but I am not planning on using a balance charger with it, kinda like what a solar charger would do.

Is this a problem, or am I overthinking?

This is something you should take into account. I would say the last 3 cells have issues with either high internal resistance, self discharge, or both.

Most likely internal resistance is the biggest player here. You need a decent 4 wire tester to achieve sufficient accuracy.

Also, you need to measure charge and discharge in watt hours. Measuring amp hours alone does not count for voltage drop under load.

Before you start running tests with 1% variance resultat change tester..... Get a proper RC tester like icharger. Lii tester is not Good enough

A lithium that is ok is closer to 98+ efficiency or above. Ur resultat are Way off or the cells are dead meat

daromer said:

Before you start running tests with 1% variance resultat change tester..... Get a proper RC tester like icharger. Lii tester is not Good enough

A lithium that is ok is closer to 98+ efficiency or above. Ur resultat are Way off or the cells are dead meat

Click to expand...

Funny, everyone uses these testers like the OPUS, gpower 4s and Lii-500. My cells are probably dead meat, and I have a lot of cells that are like 95% good according to the datasheet.

I have an imax b6 that can test cells 1 by 1, and the discharge values when I test cells in it are pretty close to the Lii-500.

1% variance is accepted in my case. But thanks.

---

Geek said:

This is something you should take into account. I would say the last 3 cells have issues with either high internal resistance, self discharge, or both.

Most likely internal resistance is the biggest player here. You need a decent 4 wire tester to achieve sufficient accuracy.

Also, you need to measure charge and discharge in watt hours. Measuring amp hours alone does not count for voltage drop under load.

Click to expand...

Yeah, 4 wire resistance tester is kinda like out of the equation as of now.

How do you suggest we measure watt hours when testing cells individually?

Many uses them because they give you a quick indication of whats going on with the cells. Example: Opus is generally 5-10% above the true value.

1% variance is what you think you have on above values but you dont. Its closer to 10% or even above. So if you want down to 1% or lower you need a better tester. Above numbers going into 1% does not work in the terms you had them. It works great comparing between cells but you cannot go into 1mAh with that tester since it easy varies 100-300mAh per test and tester

If you want to test what you do ie the difference between charge and discharge energy you need to have better equipment like said a 4 wire tester for couple of quids and a capacity tester that can do alot better results.

If you look at some of the other threads where guys have done some more extensive testing of cells comparing to IR and other you will see that they didnt use very expensive equipment but still alot better than the avera tester.

For instanve i sort my cells based on a very simple method where i group them in 100mAh buckets. No use grouping more distinct with that since the testers can be 10% off. I then randomize them 100% within the buckets. This actually leaves me with a theoretical max difference of 10% in the end.

If you dont take this in account when testing you can get very weird results especially if you use 1mAh refferences when thats very inaccurate.

I hope i made it understandable


With this said I think you test is interesting to see but we need to see it with better accurazy for being able to give a proper result. It might be that the % off will be same with a better tester statistically it should but... There are many elements involved.

daromer said:

Many uses them because they give you a quick indication of whats going on with the cells. Example: Opus is generally 5-10% above the true value.

1% variance is what you think you have on above values but you dont. Its closer to 10% or even above. So if you want down to 1% or lower you need a better tester. Above numbers going into 1% does not work in the terms you had them. It works great comparing between cells but you cannot go into 1mAh with that tester since it easy varies 100-300mAh per test and tester

If you want to test what you do ie the difference between charge and discharge energy you need to have better equipment like said a 4 wire tester for couple of quids and a capacity tester that can do alot better results.

If you look at some of the other threads where guys have done some more extensive testing of cells comparing to IR and other you will see that they didnt use very expensive equipment but still alot better than the avera tester.

For instanve i sort my cells based on a very simple method where i group them in 100mAh buckets. No use grouping more distinct with that since the testers can be 10% off. I then randomize them 100% within the buckets. This actually leaves me with a theoretical max difference of 10% in the end.

If you dont take this in account when testing you can get very weird results especially if you use 1mAh refferences when thats very inaccurate.

I hope i made it understandable


With this said I think you test is interesting to see but we need to see it with better accurazy for being able to give a proper result. It might be that the % off will be same with a better tester statistically it should but... There are many elements involved.

Click to expand...

I understood what you meant now. Even so, I will not be able to invest some more in equipment as what I am building is just a test run battery, as I have no means of getting solar energy in my current home.

That being said, I am "confortable" with some results that I am getting from the tests in the sense that I'll be able to sort out the "really bad" cells, always considering the universe of cells that I have, not comparing it to others, much less to brand new cells.

At the end, I'll just retest the bad cells to make sure they are really bad, then move forward.

Thanks for the feedback!

You should only really look discharge capacity when determining cells for packs. Its what you can get back out that really matters. That said, metering charge in is pretty interesting as well, greater the delta the more inefficient the charging of that cells is. You should monitor temperate as well because Im sure most of that is getting lost as heat.

Capacity must be tested on DISCHARGE.

Charged capacity is not relevant because used-up cells can only store a part of the drained power.

However, the discharge / charge rate may be a good indicator of the quality /state of the tested cell.

@arfurtado would be nice to have the exact cell type.
Pinks are 2600, blue LGs are 2400mAh, grey LGs should be 2200mAh.
Red sayons will vary (you can usually tell'em by the cap color).

At an overall operating level, I compare the kwh IN as reported by my Midnite Classic 150s and kwh OUT as reported by some cheap meters on 240v inverter output and for the year of 2019 (to date) I've gotten86.2% efficiency.

AIMS documents theinverter as 88% peak efficiency. The battery charge/discharge was 46% of the overall output (the other 54% was direct PV -> inverter).

First, as noted above, its difficult to rely (scientifically) on measurements from the Charge Controller and especially cheap meters from Amazon. However, its what I have.

So I account (as my operating thought process) for my 14% overall system loss as mostly from the Inverter - I don't even give 18650 cell inefficiency significantthought. Its like the wiring - I don't operate as if its asignificant factor even though I have hundreds of feet of 10 and 6 AWG wiring between the panels and the charge controllers.

However, I agree, it would be interesting to actually know the'real details'

Lithium as i sida should be above 98% efficienct. IF they Arent you got issues or just wrong Numbers.

I have tested this mnay many many Times before in My Early dys but you need a proper tester

As a quick test of IR, has it been proven that a 4 wire result is significantly different the recording the initial voltage drop under a standard load

For ex: From full charge, but cell under a 1A discharge.
Record the initial voltage drop around 3-5 seconds after start

Most power tool cells will not drop below 4v

Most laptop cell will not drop below 3:9

If a cell drops below 3.8 scrap it

Power tool cells can drop more than standard laptop cells.
My premiums keep at 4.15 anyway, but in time, you will notice more drop on power tool cells like Samsung 15Q/L compared to standard ones. The explanation is that the power tool cells have lower nominal.