What are some practical techniques for sorting cells in bulk quickly?

dc443

New member
Joined
Dec 10, 2020
Messages
19
I have around (maybe slightly under) 3000 (three thousand) ICR 18650 cells that are almost all in 4-cell series packs. They are labeled:

CORUN
ICR 18650 2600mah/37.44Wh 14.4V

And these have dates mostly from 2015 on them.

I expect all of them to contain Samsung ICR18650 cells in them, though I have samples here that are 26H and 26F already but I think most of mine are 26F. Although these seem fine in terms of quality and would be convenient to keep inside these 4S1P packs, I think they clearly need to be disassembled in order to be used in larger packs. I’ve not totally thought through the dynamics of putting many 4S cells together in parallel, or abominable sounding configurations like 3S4P4S, but it seems to me like even though these are 4S packs with a functioning built-in BMS circuit in each, I’ll probably want to tear them down for a safe approach in assembling larger packs.

I have done a bit of prep, I've got a 24V 14A DC power supply and 4 of these DROK CC/CV driver boards (2 each in a small and large size, where the small ones can pump out 3A, the large ones I've got to look it up but can output at least twice as much). I got these ahead of time in anticipation of the work involved in conditioning, charging, measuring all of these cells, and I haven't got around to seriously planning to start the job till now.

I'm realizing as I learn a bit more, that it'll be somewhat difficult to establish cell capacities using just this equipment. I have an OPUS BT-C3100 charger here which is capable of doing automated capacity testing, but it's only able to do 4 cells at a time which will be somewhat limiting. I also plan on doing some tests on that unit to make sure it measures capacity reliably across its own 4 lanes (i suppose i'll run the same cell through it once in each lane and do analysis on the numbers I get). I suppose if it comes out well I could get 3 more of those and it may be productive enough for me at that point.

I knew that it is better to store cells in cool temperatures with a 40% or so charge level for best longevity, and unfortauntely it's been almost 3 years i've been in possession of these and i haven't had the time to do anything remotely close to discharging all of these packs half way. I'm glad enough already that I haven't been forced to dispose of them or had them burn up any houses.

I rambled too much. Here are my questions:

1. Alternatives to measuring capacity: as a first stage of filtering, bad cells can be filtered out by monitoring if they heat up more compared to others while they are charged in parallel with other cells. This can also be done while discharging some cells. I think IR should account for most of this heating phenomenon. And it is known that higher IR is correlated with lower capacity/worse condition cells.

At first I was planning to set up clamps with resistive load and voltmeter integrated, so I could pop a cell in and quickly measure its IR via voltage drop. But using temperature to compare cells to each other seems like a clever way to skirt around this as a separate step. It won’t be as high a resolution of measurement, and also only a relative measurement. But I think it will work great to reject bad cells quickly as a first pass. Capacity will probably still be painstakingly measured for cells that are staged for building packs. And for large packs, IR is not an impactful property. I am trying to use IR as a way to estimate capacity/quality.

2. Discharging procedure/equipment. I think I probably have charging hardware sorted, since I could easily configure many cells to charge together in parallel using my CC/CV supply units I already have, especially if I 3D print a bunch of cell holders. But let’s say I want to bring my cells down to storage level, I would like to make loads for doing that. I think incandescent light bulbs would work pretty great, since I have them on hand, and the analog voltage/current feedback is great, but I’d need to make a circuit to allow me to dial in a specific voltage to stop discharging at, since I’m not interested in over-discharging any cells. I know I can get many off the shelf dummy/tester loads and don’t really need to save cash that much, but doing so cleverly is its own reward!
 

daromer

Moderator
Staff member
Joined
Oct 8, 2016
Messages
5,468
I have 10 Opus testers. Thats 80 cells per Day tested. Thats alot more cells per week tested than i can build.

You can build complex testing scenarios but how much is your time worth?
 

OffGridInTheCity

Well-known member
Joined
Dec 15, 2018
Messages
1,366
Agree with @daromer

As you process 100s or 1000s of cells - they will have common characteristics by cell batch. For example, 1 batch I did had heaters - so I knew to put extra focus on that. All my other batches did not - so I just lightly check. It will become clearer to you as you process a lot of cells....

I'd recommend a straight-forward test process which involves
1) Charging them up to max... looking for heaters (touch them or measure them as they charge / near max charge)
2) Letting them sit for a few weeks and weed out self-discharging cells. This could be done before capacity test if you want
3) Test IR.
4) Capacity test (full charge to full empty) - sort the cells by 100mah groups for even distribution between packs later on.

My process is to
1) Charge it to full
2) Test (discharge - write capacity on the cell - re-charge to full) *look for heaters
3) Test IR
4) Set aside for a few weeks in groups of 100mah test results
5) Assemble in to packs
6) Test individual cells for self-discharge - replace any bad ones from the 'extras'
7) Solder up packs (or spot-weld)
8) Discharge the pack to Storage voltage (3.6v)
9) Put in series for a battery (they are all at 3.6v due to step #8)
 

Bubba

Active member
Joined
May 9, 2018
Messages
289
Agree with @daromer

As you process 100s or 1000s of cells - they will have common characteristics by cell batch. For example, 1 batch I did had heaters - so I knew to put extra focus on that. All my other batches did not - so I just lightly check. It will become clearer to you as you process a lot of cells....

I'd recommend a straight-forward test process which involves
1) Charging them up to max... looking for heaters (touch them or measure them as they charge / near max charge)
2) Letting them sit for a few weeks and weed out self-discharging cells. This could be done before capacity test if you want
3) Test IR.
4) Capacity test (full charge to full empty) - sort the cells by 100mah groups for even distribution between packs later on.

My process is to
1) Charge it to full
2) Test (discharge - write capacity on the cell - re-charge to full) *look for heaters
3) Test IR
4) Set aside for a few weeks in groups of 100mah test results
5) Assemble in to packs
6) Test individual cells for self-discharge - replace any bad ones from the 'extras'
7) Solder up packs (or spot-weld)
8) Discharge the pack to Storage voltage (3.6v)
9) Put in series for a battery (they are all at 3.6v due to step #8)
I would suggest that TEST IR is too low on the list. a proper 4 wire tester works reliably regardless of the cell voltage. I have been checking IR using RC3563 which tells me the voltage and IR at the same time. This will help eliminate dangerous cells before you put them in the charger.
 

dc443

New member
Joined
Dec 10, 2020
Messages
19
Thanks. I have ordered 2 (and will probably get a third) Chinese battery discharger devices that use a 4 wire method. it should prove very useful for accurate cutoff voltage shutoff. I don't think I will have any need for accurate IR measurement. It's my understanding that those fancy $90+ 4-wire IR meters will use a 1000hz method to measure IR? This is different from measuring IR using voltage drop across the battery when hooked up to a resistive load, but that could be done for $2 so I'd probably start there.
 
Last edited:

dc443

New member
Joined
Dec 10, 2020
Messages
19
In related news, I'm finding out that these packs that i'm disassembling are a real pain and take multiple minutes to take apart each pack. Sadly the heatshrink is the only thing separating the cells from each other (mega problem), and often these heatshrinks are so tightly stuck to each other that to separate the cells I'll be ripping the cell wholesale out of its heatshrink just to physically separate it from its neighboring cell. As I've realized (since i first got these) that the voltage is obviously different on the negative cell casings if they were to touch, now it has become an imperative merely for the safe storage of these that I need to shuck them all. This will take a long time.
 
Last edited:

OffGridInTheCity

Well-known member
Joined
Dec 15, 2018
Messages
1,366
In related news, I'm finding out that these packs that i'm disassembling are a real pain and take multiple minutes to take apart each pack. Sadly the heatshrink is the only thing separating the cells from each other (mega problem), and often these heatshrinks are so tightly stuck to each other that to separate the cells I'll be ripping the cell wholesale out of its heatshrink just to physically separate it from its neighboring cell. As I've realized (since i first got these) that the voltage is obviously different on the negative cell casings if they were to touch, now it has become an imperative merely for the safe storage of these that I need to shuck them all. This will take a long time.
Yes sir - its classic DIY. Takes significant effort but you can create batteries for a lot less. I've been thru - ebike packs (very difficult, sore hands), modem packs (not so bad), medical packs (had to use a table saw - horrible), and most recently the RING batteries (not bad - just crack open in the vice). After each battery created I say "I'm done!". Then 4 months go by and I see a great deal.....
 

dc443

New member
Joined
Dec 10, 2020
Messages
19
Haha! Yeah I was surprised to see great deals readily available on battery hookup, looks like we can score cells for well under a dollar a pop in this kind of market, all it requires is rolling up your sleeves a bit. Yeah the modem battery packs look really easy to deal with. These packs I've got are a pain since they have so many things to peel (2 layers of shrink, bunch of stickers all over, though at least they don't have an extra nasty adhesive) and many tabs to carefully cut out first (ripping stuff apart quickly ends up invariably shorting something and i definitely prefer to avoid that).

I think if these batteries didn't have this sentimental value attached to them i might consider offloading them. They're pretty close to break even in terms of feeling like it's worth my time.
 

Hyraelle

New member
Joined
Dec 14, 2020
Messages
10
Yes sir - its classic DIY. Takes significant effort but you can create batteries for a lot less. I've been thru - ebike packs (very difficult, sore hands), modem packs (not so bad), medical packs (had to use a table saw - horrible), and most recently the RING batteries (not bad - just crack open in the vice). After each battery created I say "I'm done!". Then 4 months go by and I see a great deal.....
I agree, i bought 86 of those modem packs batteryhookup sells. It takes 15min each approx to open and most of it is managing to unglue the casing without damaging the cells. I open like 5 a day because that's what i can process ( 1 charger lol ). I could be faster or find more efficient techniques but i'm fine with a hacksaw and a screwdriver :)
 

hbpowerwall

Administrator
Staff member
Joined
Oct 7, 2016
Messages
1,908
This is how I've done it.. not 'fast' but works

 

Overmind

Active member
Joined
Jan 16, 2019
Messages
412
I color-code the capacity using colored dots. It's too ugly to write it with a marker on them.
 

dc443

New member
Joined
Dec 10, 2020
Messages
19
I color-code the capacity using colored dots. It's too ugly to write it with a marker on them.

Cool, this is interesting. What granularity do you group them into (how many different colors do you use)?

I'm still on the lookout for decent setups for evaluating cell capacity. I have already taken apart my OPUS BT-C3100 since its 20mm fan was driving me absolutely nuts with its vibration, I was planning on 3d printing a ducted mount attachment to put a 40mm fan onto it, but luckily after spraying the bearing on the original fan with silicone oil the unit has been running completely silently for a week now. I have two other 35W or so discharging, capacity-reporting devices coming in the mail from China, almost here, those will let me even log a discharge voltage profile for a single cell. Which would be nice.

But what I really want in order to be able to move the needle on the volumes of packs I've remaining to disassemble is stuff that'll let me run a discharge from full on many cells and measure capacity.

I want:

- To be able to discharge at say 2A or 3A (the things coming in the mail will let me do that, but the Opus won't)
- Let me discharge *and* measure the capacity of ~30 cells at a time concurrently and independently (I'm not willing to spend $20 * 30 = $600 on this)

I already have:

- a few garbage-tier li-ion cell chargers that make awful audible EMI noise
- the opus
- a number of very capable CC/CV Buck/Buck-Boost drivers and a 24V (as well as plenty of 12V old PC) power supplies, that'll let me charge many cells without balancing
- a dual 6s lipo charger unit, that'll let me charge 12 cells at a time (with balancing!)

for handling charging, so it's easy and affordable to accumulate equipment for charging and topping up the cells.

Once I get a plan for this sorted i'll have a plan in motion to be able to crank through my battery packs in a few short days and can move onto next steps such as ordering a kWeld.
 

OffGridInTheCity

Well-known member
Joined
Dec 15, 2018
Messages
1,366
>- To be able to discharge at say 2A or 3A (the things coming in the mail will let me do that, but the Opus won't)
Most DIY powerwall / second hand cells / laptop-modem cells (that I've processed) aren't ment to perform at 2a or 3a... My own powerwall is targeted for 0.5a/cell max. That's why a lot of folks are OK with 1a max of OPUS. :)
 

dc443

New member
Joined
Dec 10, 2020
Messages
19
Yeah, I know. I just sorta figure it may be another capability that could be used to discriminate cells from one another, produce a better signal for a temperature sensor to pick up, etc.

It's clear that I should definitely simply buy a few more Opuses, for all the time i could save for myself cobbling something together (with temperature monitoring, capacity measuring and more), as much fun and as good of a learning experience as that would be. 1A discharge ability is fine. I was just sorta hoping there'd be some practical ways to use lipo charger/discharger units for li-ions. It's just that capacity measuring is such a special use case that outside of the Opus you just can't get that feature in something cheap.
 

daromer

Moderator
Staff member
Joined
Oct 8, 2016
Messages
5,468
This is a simple way of doing it with Opus:
 

hbpowerwall

Administrator
Staff member
Joined
Oct 7, 2016
Messages
1,908
I color-code the capacity using colored dots. It's too ugly to write it with a marker on them.
ok, that is so obvious! BUT grabbing a different color all the time would be a pain.. but still a neat idea
 

Wolf

Well-known member
Joined
Sep 25, 2018
Messages
1,325
Well of course I am going to be the odd man out. I really don't think putting together a powerwall with used 18650 cells should be a speed test, rather a well thought out project. You are after all repurposing a somewhat dangerous item to store and release energy. Now I don't want to be an alarmist but 18650 batteries or any Li-Ion metal based battery (for that matter of fact anything with an energy storage potential) in the right circumstances can be dangerous and given that knowledge just don't put the cells or yourself into those circumstances. I sleep very soundly with hundreds if not thousands of cells in my house at any given moment, knowing that I am safe. Is there a potential of a problem? Sure but so is every other Li-ion based product we have in our house. Cellphones, Laptops, tablets, etc. I don't see any of us keeping a fire extinguisher next to them at all times. Maybe we should.🙂

My technique is by far not the only technique to assemble a successful powerall. If anything at all it is probably the slowest and most methodical but then again I am not in a speed race but a quality quest. I want my 14s80p battery to last as long as possible without any intervention less the monitoring of it.
These steps taken are all after the cell has been liberated and cleaned for testing. The assumption is also you have a good supply of quality cells.
If you are working on laptop cells or cells with unknown quality I will usually test Voltage and IR before cleaning the cells completely. Kind of a waste of time to clean and deburr a cell if it's absolutely no good.

All data is recorded into an Excel spreadsheet.
So here are my steps.
  1. Excel book is started for this build and Cells are numbered with small number stickers.
  2. Measure IR and Voltage Enter this data and date code if available into a spreadsheet
  3. Cells with higher than acceptable IR are set aside for recycling no matter what voltage
  4. Cells with good IR but low Voltage < 2.5v but >0.5v are slow charged with 50mA to 3.6. Then they are charged with the normal 1A rate to full. (No heaters have been observed using this process so far)
  5. Cells are inserted into testers. Tester make and slot # recorded along with the testers IR reading if available and or desirable to know.
  6. After a 1A (1000mA) C/D/C ( Charge / Discharge / Charge ) cycle resulting mAhs are recorded.
  7. Once the cell has been taken out of the Charger/Tester Voltage and IR are again recorded.
  8. The cell is set aside for >25 days to verify the cell is not a SD (Self Discharger)
  9. After >25 days the cell is again checked for voltage and IR any cell with more than a 0.05 voltage difference from when it came out of the charger is suspect and set aside for recycling.
    Side Note: I have had chargers that claim the end voltage to be 4.2v but yet the finish charge was below 4.15v. So if measuring a cells voltage just by the predetermined assumption it was finish charged at 4.2 you may encounter what appear to be SD cells but are not. Hence I measure SD voltage drop against the actual exit voltage of the tester.
  10. Depending on the volume of the cells I have in my sheet I will filter them down by best IR, best date code, and best mAh results to the amount I need i.e. 1200 cells for a 14s80p. I always add enough for a potential spare pack.
  11. Once the cells have gone through this sorting process I enter the best of the best remaining cell numbers into an Excel based repacker and generate my packs.
  12. This will give me the cell numbers to pull from my tested batch. As I have them in boxes of ~250 cells numerically it is not to difficult to find the cell number and label them with a sharpie of what pack they will go into. (I do like the color sticker idea)
  13. Last but not least you guest it! The cell gets one final IR and voltage check before being committed to the pack. If the cell at this point does not pass the muster it gets replaced with one of ~ equal value from the 80 spare cells.
This is my procedure from what I have learned and it builds very well balanced packs with Ah results deviating ~ 2000 mAh range across 14 packs. Some images for your viewing pleasure.
Wolf
20201221_225253.jpg1609166558643.png
 
Top