The Passion of the Li-Ion 18650

[Wolf]

As we all are aware there are hundreds of millions if not billions of 18650 cells in all kinds of devices and there will continue to be a demand and supply of this form factor. Too many designs for this form factor would have to be changed, molds redesigned, BMS redeveloped, loads recalculated which all costs money.
I think the cylindrical Li-Ion cells are here to stay for quite some time. Just the scooter and power tool industries alone will continue to use them for the foreseeable future as they are compact and deliver substantial power for a reasonable cost.

As far as I am concerned till personal fusion generators become a reality we will be using some form of energy storage for quite some time.

Since I have finished my conservatively speaking ≈30kWh powerwall, and just waiting for the last lexan sides for the packs to finish hooking my 4th and last 14s80p battery to the build. Also this is not a high solar time of the year in the north east of the americas so I have delved back into cell study and analysis of the Li-Ion 18650 cell. I have gobs of leftover tested and untested cells, battery packs of all kinds medical, power tools and yea laptop packs.
I have stepped up my game as far as cell testing equipment with 2 Megacell chargers and 5 SKYRC MC3000 and one lonely OPUS just for sentimental reasons.
By the way the fan Mod on the OPUS is highly recommended if you are going to C/D/C cells at 1A. The cells will stay cool and the tester will not overheat causing it to pause during the cycle extending the test time.


The forum title is Li-ion cell recycling and testing procedures and I am posting this thread in this Forum to share some of my findings with you in hopes to help some of you with your harvesting endeavors.
As of this writing I have retested 544 of the original 6225 cells that I started with in Dec. 2018 and ran through Aug. 2019.
The first of hopefully many hints, tips and discoveries that I will share with you is cell markings. Now Samsung, LG, Panasonic, Sony, and Sanyo all have their product markings and date codes as they are the major players and those are the ones I will concentrate on. Today I will concentrate on the Sanyo as it has baffled me and many others. Most all Sanyo cells have a red wrapper and have an identifying insulator ring at the positive terminal. There is an Orange w/purple insulator ring and Purple w/yellow insulator ring style wrapper also but it is only for the UR18650ZT (2650 mAh) and UR18650ZTA (2620mAh) respectively.
We have all complained that there are no markings on these batteries yet I am here to tell you that every one of them has identification marks on them. At least the real ones. You basically need a hubble telescope and a very good flashlight to find the numbers but believe me they are there.
All Sanyo cells start out with UR18650 and then the identifier. One technique I used to use to find the cell number is to use a fat sharpie and quickly run it over the area which looked like it had numbers / letters on it and then quickly use a paper towel and rub the area hoping the recessed numbers would pop up.
Sometimes it was very successful other times it was not. I now use the magnifier on my cell phone and boy do they show up nicely. Here are some of the pictures. Also a description of the way the date code works. If there is an R with 4 numbers on the right of the cell part number that is not the date code. I have no idea what it is. If there is an R with 4 numbers (most likely will be R1122) then usually below the part number on the left side you will find the date code. It will always be a letter 2 numbers and a letter. If you have a letter 2 numbers and a letter to the right of the part number than that is your date code.

These particular cells are UR18650FM 2500mAh cells they are defiantly old but the ones I have are from 2008 to 2013 with a few of them actually still testing well.There are basically 3 different reds (shades) of the UR18650FM that I have found. One is a bit more pinkish, one a nice deep red, and one more of a translucent clear tinted red wrapper. You can definitely see the labeling on the can of the cell very well with this wrapper. Unfortunately the picture does not due the translucent wrapper justice as it is the very right one in the picture.

I have yet to encounter a heater but as these cells have already gone through a previous testing cycle I do not anticipate any.
One note as Panasonic has bought Sanyo some of their cells are still using their heatshrink and marking machinery here is an example of their NCR18650GA
Barely readable but there.

And here is the updated Lab
Wolf

 

[OffGridInTheCity]

Yes sir. While LifePo4 is the current rage and I can see benefits such as bigger units/a lot less work and potentially more cycles - the shallower voltage range, procurement of class A cells, and 'pack' maintenance offer different challenges.

I agree with @Wolf thinking that the shear volume of 18650s will continue for a loooooooong time. The effort it's taken to create billions of this standard unit took many years - and anything that overtakes it will also take a long time. As Elon said, it's not creating the car that's the big challenge, its the mass production.

For me, its just too late. I'm so far down the lithium-ion path that 18650 (or at least lithium-ion) is likely my future. Eventually I'll need to repair packs etc - but I can see this as being the same 'regular process' as creating the packs in the first place. So I have a firm, long term maint vision - which is good for the 10-20yr plan. Also, these cells are working GREAT and at <40% DOD I'm not anticipating the need to replace them for a very long time.

Meanwhile, working to expand from 81kwh (6 x Batteries) to 121kwh (9 x Batteries) one cell at a time. Battery 7 (green NCR18650As) is ready for longmons and a shelf to hold it. Battery 8 (pink ICR18650-26Fs) is underway as the addiction continues...



As I read about LifePo4 and practical issues such as balancing - I'm struck by how flat the charge/discharge curve is. For example, with my lithium-ion I'm doing a touch-up Batrium Auto-Balance right now in the lower 1/3 of my operating voltage range and I'm not sure that's even possible at a practical level with LifePo4 because of the shallow curve.


It strikes me that 18650 (lithium-ion), LifePo4, LTO, 'Solid State', x, y, z have key differences - more than you might think at the surface thought - as you operate and maintain a battery bank over the long run.

 

[Wolf]

The next cell in my discussion is the LG S3 2200mAh, or more commonly known as the ICR18650S3 which includes the part numbers :
LGAAS31865
LGCAS31865
LGCS318650
LGDAS31865
LGDS318650
LGEAS31865
LGES318650
From what I gather these are the workhorse of the LG 2200mAh ICR battery lineup. They are generally always blue with a white insulator.
Here is an example of the variation of labeling.

The Date key is always on the second line starting with a letter A=2001 B=2002 C=2003 and so on. The next 3 numbers are the day of the year so in the example above G330 would be G =2007 and the 330th day is November 26th so the battery was built on Nov. 26th, 2007.
This date key is the same on all LG cells as far as I know.
I don't get quite that picky, for me, month and year is good enough in my spreadsheet.
At first I dismissed this little gem of a cell as the AC IR was always high >60mΩ up to 120mΩ AC IR.
Manufacturer recommends ≤80mΩ. I recommend ≤75mΩ although I have had a cell with an initial AC IR of 125.8mΩ test out at 2103mAh on the MCC and a 2129mAh return on the SKYRC with a final AC IR of 103.5mΩ. Go figure. \_(ツ)_/¯
Would I use that high IR cell in a parallel pack? Not on your life. If I had enough cells and all were between 60mΩ and 70mΩ I would not hesitate to build a battery with them. I would use all of the same style S3 cells though and more than likely not mix and match with other ICR chemistry. Considering I have some cells that were manufactured July 2007 that have an IR between 63mΩ and 68mΩ and have a 100% SOH ie 2200mAh return on testing makes me wonder who came up with the 2 to 3 year life span on a Li-Ion battery. Oh BTW these were also previously tested and left at 4.1+V sitting in a box in my shed through various temperature changes including a cold winter.
Nevertheless from theses cells I never had a heater and the cells performed well even at the upper IR range. The cells come in laptop packs, medical packs, various multipurpose power packs that have relatively low amp draw requirements.
The actual cells specs are Standard charge 1075mA, standard discharge 430mA, Max charge 2150mA and Max discharge 3225mA. So all in all quite a good cell if you can find enough.
Here is a graph of the 231 cells I have tested with AC IR compared to % capacity with trendlines.
It clearly shows as AC IR increases SOH decreases and the sweet spot is anything below 75mΩ. AC IR
All cells were tested at 1A C/D/C (4.2V--3.0V) on both the MCCs and the SKYRCs.
The first chart below is the AC IR reading from the RC3563 an the SOH results from the MCC. The second chart are the same cells with the SKYRC DC IR and SOH correlation. Amazing similarities don't you think.


Well it was a snow day in the NE and I used the opportunity to sort and organize the cells I will be testing. More to come soon.
Wolf

 

[OffGridInTheCity]

One of my 14s120p'ish batteries in my powerwall is made of all LGDAS31865(s) from BatteryHookup. Have about 200 good ones and 200 slight-to-moderately dented left overs from that effort. Here's 3 of the good ones at random....

Which would make M = 2013 and N = 2014 dates - cool!

Couldn't bear to throw away the dented ones yet - thinking about using good (and dented) in a large Frankenstein 14sXXp shared battery for my APC UPS 3000s as the cells would just sit at 3.95v all the time - not the same kind of risk as actively charging/discharging.

Here's what the dents look like - they were that way as they came out of the packs....

I know - not recommended - battling with myself .

 

[Wolf]

The next cells are the Panasonic cells
They are the CGR (Cobalt-Manganese-Nickel) and NCR (Nickel-Cobalt-Aluminum Oxide) which makes them Hybrids.


Panasonic has been and still is (Tesla chose to partner with Panasonic early on) a pioneer in the Li-Ion battery manufacturing. These are some of my favorite cells as I have some NCR18650A (2900mAh) that are dated June 2010 have an AC IR of 41mΩ (Manufacturer spec is ≤48mΩ) and have a SoH of 95%+ and this is after sitting with at least a 4.1V charge for over a year. For a 12 year old battery to perform this well is in my book impressive.
I have both NCA cells of all flavors and CGR cells of the same with those impressive longevity numbers. Not to dismiss other manufactures as they also have some very good performers.
But we are talking about Panasonic
To the date code.
The 4 digit date code is actually quite simple 1 is the year 2 is the month and 3+4 is the day.


In this example 0618 0 = 2010 6 =June 18th day
9X24 9 = 2019 X =November 24th day
2Y02 2 = 2012 Y = September 02nd day
Months go from 1 to 9 and then XYZ.
Got it?
Good.
I won't bore you with the results graphs unless you indicate to me you want them but they are basically always the same. The higher the IR the lower the SoH.
The CGR chemistry has a slightly higher AC IR results usually between 45mΩ to 60mΩ depending on the suffix of the cell A, C, CG, D, E, etc. and is the lower mAh representation of the Panasonic lineup. 1400mAh to 2550mAh
The NCR chemistry is usually in the 35mΩ to 50mΩ AC IR range, also depending on the suffix of the cell A, B, BD, BE, etc. and represents the higher mAh offering by Panasonic with a 2700mAh to 3350mAh range.
In conclusion I really like these cells and its another one of those conundrums whether or not I would I use these in another battery build with ICR cells. I did put some in my Frankenstein battery. To be exact 144 of both NCR and CGR cells.
I said I wasn't going to bore you with graphs but I changed my mind. Here is an Analysis of my Frankenstein battery cell information just for kicks.
Wolf

 

[OffGridInTheCity]

Yes sir - as you can see from the pic above, battery #7 for me is NCR18650A(s) from RING packs.

I have 2 other batteries online with these NCR18650A cells. Power2Spare ( https://power2spare.net/collections/all ) specializes in medical packs that have these which is were I first started acquiring them. Too bad the RING packs are gone. Power2Spare is a bit pricey but if you need a couple of hundred of this specific type to round out a build...

One thing I notice about NCR18650A(s) is they dent really easily compared to other cells... thin container? so extra care is needed when prying them out.

 

[OffGridInTheCity]

@hbpowerwall recently did a live chat - and one of the topics is the death? of 18650 powerwalls. One comment by @hbpowerwall is that if you have time and no $ then 18650 may still have a place.... which is my case to a T. I'm retired and process 18650 as a background task while listening to content.

BUT ALSO... I'm 4 years in and have a working 9,000 cell (81kwh) powerwall with 1287 cycles and going strong. I'm in the process of adding another 3000 cells to reach 108kwh and have room for another 67kwh for a grand total of 175kwh battery bank a few years from now.

I really like the wider voltage range of lithium-ion compared to lifepo4 - where I can choose 80% DOD or 50% DOD or 20% DOD (if I reach 175kwh) type of control for long life.

I believe 18650 can be still be significantly cheaper 1) if you have the time and 2) you can get good priced cells.

Long term is something I think about these days.....
18650 - not sure what's going to happen when 18650 packs start to fail. Will it be gradual? or hit/miss pack here or there? Some of my 84 packs have 1287 cycles, and some are only 800 cycles and some 400. The new packs to go online later this year are at 0 cycles. Will the process of retiring bad packs be 14 at a time? or packs here and there? - will be interesting to see. If packs fail here and there - it will be interesting to disassemble and retest the cells.

Lifepo4 - for a 167kwh battery bank that's still a large number of bricks! and requires a larger space by 20% or so. If one did 200ah units, it would be ~300 units for 167kwh. That's a lot of LifePo4 cells as well. And how does it fail? one cell at a time? or 16s at a time?. Same questions as 18650. As far as I can tell, no one really has a handle on long-term yet.

Batrium's visibility into each pack (or cell for LifePo4) is a key component in my mind of being able to diagnose failure trends. Partly why I don't resent the $ so much compared to 'regular' BMSs.

The fact that the 18650 powerwall is working so well with minimal attention for over 3 years now + I can power my entire house including AC, Dryer, Cooktop, HotWater, and EV (if I had one) is just a fantastic result/experience so far.

 

[harrisonpatm]

One comment by @hbpowerwall is that if you have time and no $ then 18650 may still have a place.... which is my case to a T

Me too! I would also like to add the recycling aspect. I get peeved when pundits or ignorant friends/relatives say that the solution to environmental concerns is to "just recycle." Ok...who's gonna do it? In addition to 18650's, I process all kinds of e-waste, and there's no way around it, recycling takes work and there's not much profit in it. So even if powerwalls have "died," building the skill set to process old batteries and make them new again is not going to go out of fashion anytime soon.

 

[Wolf]

Indeed @harrisonpatm and @OffGridInTheCity I wholeheartedly agree that the future is still bright for 18650 recycling. I personally find that the solution is for a decentralized energy production environment to be the future. As I say till the personal fusion device becomes readily available and affordable we are unfortunately somewhat enslaved to cheap energy from hydrocarbons. As 18650 and for that matter of fact any battery recycling that can be done by the savvy homeowner, who can add some solar panels to their property and store the excess energy produced in those salvaged cells for later use, should be heralded as a pioneer of independent energy production.
I am also talking about the actual harvester of cells not one who purchases new cells whose carbon footprint let alone environmental impact during manufacturing is massive and is not keeping good used cells out of the waste/recycle stream.
Besides keeping these still useful batteries out of the waste/recycling stream the one thing that doesn't seem to be mentioned though, unless I have completely missed it, is the reduction of CO2 that we participate in.
According to the latest info I was able to find the national average carbon dioxide output rate for generated electricity in 2019 was 884.2 lbs CO2 per megawatt-hour (EPA 2021), which translates to about 953.7 lbs CO2 per megawatt-hour for delivered electricity (assuming transmission and distribution losses of 7.3%) (EPA 2021; EIA 2020b).
Naturally that is excluding renewables of about 20% consisting of Wind, Solar, and Hydro. Nuclear contributing about 19% for a total of "CO2" free production of about 40%.
Photovoltaic is only 2.8% in the US but at my house it is closer to 98% especially as we approach the summer months.
So taking the average of 900 lbs CO2 per megawatt-hour, in my case I have just crossed the 5000 lbs of CO2 reduced
or 2.268 mt (Metric Tonnes) not a great deal, and I am sure @OffGridInTheCity has me beat by many mt's but every bit helps. Especially if more homeowners would consider their own personal energy production systems.
Go 18650

Wolf

 

[100kwh-hunter]

As always your post are a welcome tread to read, and Spock got some friends!
Yes i was looking for him, sorry, i am a fan

And yes i totally agree that the 18650 form factor will stay for a long time.
If i look at my power tools alone........They will not change easy, look at Hilty(my brand) Makita de walt and what ever, a lot of laptops....
I agree it will be a standard for a loooooooooong time, Elon started something.

With best regards Igor K

 

[OffGridInTheCity]

Heavy into 18650 but... @LithiumSolar did a youtube today showing a US distributor of EVE cells. 280ah @ $599 for 4. For a 48v@280ah battery that would be 4 x $599 = $2396 + $200? shipping = $2600.

View: https://youtu.be/kVpFS8_PcjQ

For a comparable 18650 using 2600mah cells at avg 95%, that would be ~ 14s114p. If we divide $2600 / 1,596cells = $1.62/cell for the same capacity but shorter life. I paid (this year) to Jag35.com, BatteryHookup, and BatteryClearing house about $1900 to get 1,500 cells at 95% avg original capacity. $1900 / 1,596 = $1.19 / cell.

The $1.62/cell (LifePo4 cost) vs $1.19/cell (18650 cost) for 280ah@48v is getting slim these days. I wish 1 Batrium controller could do BOTH 18650 packs and LifePo4 cells

 

[vicriz]

Am new here and find you @Wolf
Yo're amazing!!!

 

[OffGridInTheCity]

Just finished processing 3,000 cells looking for pink ones

The journey started with jag35.com - they advertised 2600mah modem packs with no details (buyer beware). Ordered all they had - 2,000 cells for $1111.88 Turned out to be about 50% keep rate with 90% or better criteria. This gave me about 1,000 at 90% capacity.

Moved on to 17 by 24cells each BatteryHookup medical packs advertised as 'fully tested' for $543.96 but unfortunately there were many with corrosion and cells were not so hot. They gave a $246.42 refund based on stats below

This gave me about 138 at 90% capacity.


Finally - got 800 ( https://www.batteryclearinghouse.co...0-cells-in-large-modem-packs-4-cells-per-pack ) more from Battery Clearing House with 75/75 guidance in the description. This was $618.51. Just FYI - here's what I found after processing them - e.g. pretty accurate description.

This gave me about 419 at 90% capacity.

RECAP:
Was able to get a grand total of 1,557 cells of same type at >= 90% capacity for a 14s108p battery - e.g. 260ah. **Met my goal**
The total cost was $2027.93 / 1,557 = $1.30/cell

$1.30/cell is pretty normal in my experience with purchasing cells BUT the 50% keep rate was low... usually get more like 70% keep rate. This was a lot of extra work.

However, I have the following left over....
250 @ 2200mah
150 @ 2100mah
200 @ 1900-2000mah

Thinking of a large APC UPS battery made up of my leftovers.

 

[vicriz]

Just finished processing 3,000 cells looking for pink ones

Ohh man, yo're trully entusiast with 18650
Salute!!!

 

[OffGridInTheCity]

I put an new DIY 18650, 14s108p Battery online today - and it reminds me that my original battery (as of today) has 1,375 cycles @ an average of 37.4% DOD / cycle. I have proof that low stressed, second-life, 18650 batteries can easily exceed 1,000 cycles.

Will these cells do 2,000 or 4,000 or ... ? Based on performance so far, I'm hopeful but it will take a lot more time to find the limits.

 

[Wolf]

Well according to @Generic he got reasonable results way into the 1500+cycles(on well manufactured cells) and that was I believe an OPUS charging to 4.2 and discharging to 2.8. Don't recall the amps but nevertheless FULL cycles. So if our PWs are not even scratching full cycles and at very low amps as in milliamps
we are in uncharted territory. Just think we are pioneers of the 18650 recycle revolution.

Results of Capacity Testing 18650s for 1,600+ Cycles

@Diegobg - Nice work! The 2 Amp discharge is putting more stress on the cells, so it looks like they are degrading faster, about 9% loss in capacity, right? Are the cells heating up at all? Thanks!It is right. For now, the Liitokala NCR18650B cell has lost 10% or so degradation in 300...

secondlifestorage.com

Wolf