US18650V3 cells

Announcement - Help us fight the BOTS! Please report all spam including stuff in your inbox!

ziporah

Member
Joined
Apr 26, 2021
Messages
45
I'm inventorying my US18650V3 cells, they all rested for at least 3 weeks if not more after being fully charged. I'm measuring the IR off all cells with my YR1030+ None of the 200 cells I've tested are below the IR suggested in the spec (31.5m). I randomly picked 10 cells to test, they vary from 1770 - 1830 mah (after rest period, no top up, discharge to 3V). I`m wondering if I should continue testing them all or consider selling them and find some other cells to replace these.
 

Korishan

Moderator
Joined
Jan 7, 2017
Messages
6,552
Test all cells when doing an IR test. You never which cells are good if you only randomly test them.
The only time that random testing is good idea is if the cells are brand new and you received them from the same supplier.
With used cells, always test every cell. Never assume. It only takes 1 cell going bad to make for a very bad situation.

With IR being that high, I wouldn't use any of those cells in a production powerpack application. Perhaps they'd be good for single use applications like torches/flashlights, solar garden lights, flood lights, etc. Things that have extremely low current load/recharge
 

floydR

Member
Joined
Aug 23, 2017
Messages
1,312
What are the IR readings?
31.5mΩ is a very good reading for most cells most of mine are 35-40mΩ
Do you have a datasheet on the cells?
What are your cell's chemistries?

The cells will be higher in IR than the IR listed on the datasheet.
When testing for capacity you should top off the cells.
Later floyd
I never had luck with sony cells
 
Last edited:

ziporah

Member
Joined
Apr 26, 2021
Messages
45
I am testing the IR of all cells, as stated, none of them are below 31.5 mΩ.
The IR spread is like this graph. A top level IR indicator of other members would be nice, for those who are also using these cells
1627845633937.png

I collected a random sample of 10 cells to measure the capacity of these 10 cells.
@Korishan : I am always testing all cells I will be using, this just was my first triage to find out if I would bother top charging them all and pushing them trough my testers to find out their left capacity.
The Y axis is just the number I assigned to the cells in my database.
All together, these are 280 cells, so I'm not quite sure I want to toss them out, I might have to rethink what I am going to use them for.
I've also got a bunch of US18650V cells (numbers not counted yet, guessing around 200'ish as well) , I might add them even tough they are lower spec and put them to use for my UPS if they can't make it in my powerwall.
 

floydR

Member
Joined
Aug 23, 2017
Messages
1,312
31.5mΩ is a typical cell when new, some lower some higher. Looks like you have many in the 31.5mΩ- 35mΩ range, even more if you included up to 40mΩ. Are these new cells or NOS used cells or used cells?
With new cells they should be close to typical cell IR.
With NOS cells close to typical some higher
With Used cells the IR will be higher than typical.
A top level IR indicator of other members would be nice, for those who are also using these cells
1627845633937.png
I also use a YR1030+.
I have some Ls LR1865LA cells That are at least 5 years old with IR readings of 18-20mΩ from hoverboard packs. the packs were NOS.
I have 2000+ LGGBM261865 cells with IR between 35mΩ-40mΩ.
since I haven't tested any sony cells lately I have no info to post. Just a general observation that I would have used cells between 35mΩ and 45Ω if I didn't have enough. so 31.5mΩ to 41.5mΩ should work.
Later floyd
 

Korishan

Moderator
Joined
Jan 7, 2017
Messages
6,552
I misread your original post as the 1780 - 1830 mAh as the IR values. I over looked the "mAh" on the next line and thought those numbers represented the IR values you were getting. That's why my statement "...that high...." was included.
As floyd mentions, those are pretty good cells overall. The 50's might be getting a little high for mixing in with the 35's, though. So perhaps you could make two different strings. One string uses the <40-Ohm cells, and the other string uses the 40 - 50-Ohm cells. This will help keep resistances between neighboring cells pretty balanced during high charge/discharge currents
 

Wolf

Active member
Joined
Sep 25, 2018
Messages
1,581
The US18650V3 is a 2250mAh hybrid and an early attempt by Sony to harness the power of nickel and magnesium to build a high drain battery
(Lithium Nickel Cobalt Manganese Oxides) LiNiMnCoO2 which is Lithium manganese nickel also known as INR or NMC (high amp draw)

For optimal SOH IR should be ≤ 31.5mΩ anything above that and the cells SOH has been compromised. These are ≈10 year old cells and are more than likely used up. With the high IR results that are >31.5mΩ especially for this hybrid chemistry you will see a degradation of SOH very quickly.
As you wrote
they vary from 1770 - 1830
That's ≈78% to 81% SOH really skimming the low end for a successful battery build.
Sony has done some strange things with cell chemistry that in some cases makes a very good cell. Unfortunately they don't play well with others in the sandbox,. If you only use Sony cells and they are all the same great but I wouldn't mix and match them with anything other than themselves.
Remember INR chemistry = low IR, usually ≤ 30mΩ and more than likely with a good cell <25mΩ. Most newer high drain cells have an IR in the low 15mΩ.
LGGBM261865 cells with IR between 35mΩ-40mΩ.
ICR chemistry which the LG M26 cells are, 35mΩ to 40mΩ is just about right for a close to 100% SOH.
You cannot compare ICR to INR "IR" as a 35mΩ IR value for an ICR cell is very good whereas a 35mΩ value for an INR cell is horrid.
Wolf
 

floydR

Member
Joined
Aug 23, 2017
Messages
1,312
I see. Knew Wolf could correct me if I was wrong. So my Lishen Ls LR1865LA cells are be INR cells?
Later
floyd
 

floydR

Member
Joined
Aug 23, 2017
Messages
1,312
I am confused this is the datasheet I have been using for the LG m26 cells https://www.batteryspace.com/prod-specs/11609.pdf
Screen Shot 2021-08-01 at 10.40.01 PM.png

ultra low IR are IMR ≤ 20mΩ lowIR are hybrid INMR ≤ 30mΩ mediumIR are INR ≤60mΩ ?
and I thought I had thechemistries figured out. more I know the less I understand.
Later floyd
 

ziporah

Member
Joined
Apr 26, 2021
Messages
45
....

That's ≈78% to 81% SOH really skimming the low end for a successful battery build.
Sony has done some strange things with cell chemistry that in some cases makes a very good cell. Unfortunately they don't play well with others in the sandbox,. If you only use Sony cells and they are all the same great but I wouldn't mix and match them with anything other than themselves.
Remember INR chemistry = low IR, usually ≤ 30mΩ and more than likely with a good cell <25mΩ. Most newer high drain cells have an IR in the low 15mΩ.
So if I would make a seperate string out of the sony cells, say 14s20p or 14s40p, depending on what can be used from my other US18650V cells and use that string together with another string of mixed samsung INR-18650-22P and/or Sanyo UR18650, the strings should play nice together.

These are the number of cells I have gathered up to now:
# CellsCapacityPart Number and ColorManufacture
4443141,3492ICR18650-22PSamsung
49342,2916INR18650-20RSamsung
27179,9604UR18650RSanyo
142948,3822Blue cellsunknown
280to be measuredUS18650V3Sony
300+to be measuredUR18650AASanyo
200-300to be measured and opened form packsUS18650VSony

The ones with capacity meantioned are all tested with good IR (well within specified range by manufactorer)
It's because I've read a lot off confusing messages about these Sony cells, I'm trying to decide clearly what to do with them and if they are worth my time to clear them from the packs and test them all.

Is there a way to detect the chemistry of a cell ? I've got the so called Blue cells, they are cells from hooverboards and e-steps by the looks of the design from the packs I've dissassembled, but there is no serial or whatsoever on them. They all come out in this IR trend:
1627884713308.png

This is the capacity graph of the cells:
1627885002799.png

The ones at the end are the "bad ones" they were 0V or high IR before putting them in the charger, but just for statistics I added them in the sheet so I could detect and the ones with 0 are offcourse the ones that didn't even measure anything at start and were binned. This batch I acted differently. Pre-measure each parallel pack (0V => keep to end), open up, numbered all cells, measure IR, good IR => into charger, bad IR => bin. End processing : try reviving the 0V cells with low charge for 10 minutes. Rest, and repeat until they are recognized as Li-ion. some were recovered, some will be binned. Strangely enough, some high IR cells are also the ones with the highest capacity. They all "rested" the same amount of time (3 weeks) before being tested
 

Korishan

Moderator
Joined
Jan 7, 2017
Messages
6,552
So if I would make a seperate string out of the sony cells, say 14s20p or 14s40p, depending on what can be used from my other US18650V cells and use that string together with another string of mixed samsung INR-18650-22P and/or Sanyo UR18650, the strings should play nice together.
Yeah. The full strings are easier on the individual cells when they are batched together on likeness than if they are batched together in bulk.
With the full string batteries in parallel (absolute ends are paralleled) the current flowing is more balanced across the cells.
Is there a way to detect the chemistry of a cell ?
1) go by the label on the shrink wrap or stamped on the cell
2) take it apart and have the guts sent to a lab for analysis.

No exactly straight forward and hopefully the wrappers are intact and are correct. Some re-wraps are out there where ppl will re-wrap an old or weak cell with a wrap that says it's newer and/or higher capacity/current cell, unfortunately.

Based on your graphs, you at least have a lot of cells with similar IR. Even the capacities are pretty close.
Perhaps leave those cells that have >30-Ohm IR out of the big packs. Doesn't look like you have many so far.
The US18650V3 is a 2250mAh
And as wolf says, the original capacity is 2250 and you now have 1850 (avg), then that's a reduction of 22%, or about 78% capacity left. This shows they have had a hard previous life, for sure. And, after about 75% (maybe 70%(?), I'm not sure of the actual number) the capacity will start to drop off faster.
I suppose if you baby these cells really nicely, you can get a longer life out of them. For example, only pull about 500mA per cell for the rest of it's life.
 

Wolf

Active member
Joined
Sep 25, 2018
Messages
1,581
correct me if I was wrong. So my Lishen Ls LR1865LA cells are be INR cells
No correction @floydR just information:)
From what I found on the Lishen LS LR1865LA Cell https://prom.ua/p560238334-vysokotokovyj-akkumulyator-lishen.html
Resistance: ≤ 15 mΩ
So yes I would say they are INR Cells.
Just remember there really are only 2 chemistries ICR and INR. The rest are just hybrids trying to accomplish a well choreographed dance between performance, cost, and longevity. Also certain battery buyers want specific performance values and the manufacturer will try to accommodate their requests. If you buy 10 million batteries at a wack I will try to meet your demands.;)
Not to sway too much from the OP topic but yes the ≤60mΩ for the LG M26 is quite high considering we get mostly 35mΩ to 40mΩ in our testing.
That being said the highest IR M26 cell I have in my collection was 55.37mΩ and it still came out at 102% SOH so ≤60mΩ spec seems right. Although I would obviously stick with the lower IR 35mΩ to 40mΩ cells to build a battery.
1627907974439.png


Is there a way to detect the chemistry of a cell ? I've got the so called Blue cells, they are cells from hooverboards and e-steps by the looks of the design from the packs I've dissassembled, but there is no serial or whatsoever on them
Without knowing the manufacturer and part number it is quite difficult to determine the "chemistry" of a cell without as @Korishan said "have it analyzed at a lab".
That being said we can go by a certain rule of thumb.
Again ICR chemistry is LiCoO2 which is Lithium cobalt oxide and you will be sore pressed to find a new ICR cell with < 25mΩ the cobalt just won't let it. Alternatively a new INR cell LiNiMnCoO2 which is Lithium manganese nickel will be in the sub 20mΩ range. The nickel gives it the low resistance.
Let's take 2 of your cells as an example.:
The ICR18650-22P (got to love Samsung tells you right away what it is) has a IR spec of ≤35mΩ
The INR18650-20R you guessed it INR got to be a lot lower its the nickel. Spec is ≤18mΩ.
You will also notice that the INR max charge/ discharge rate is ≈ 2x that of the ICR. (less resistance less heat)
My cell database has this information. It is a work in progress but has a lot of information.
Use at your own risk and always verify but I do spend a lot of time on it to make sure its right.
1627912473359.png
Wolf
 

ziporah

Member
Joined
Apr 26, 2021
Messages
45
....
Use at your own risk and always verify but I do spend a lot of time on it to make sure its right.
View attachment 25771
Wolf
@Wolf you can't notice offcourse as I did a text based copy/paste, but I'm actually using your cell database as a starter for my cell table spreadsheet.
I start by copying over the information from your database and then have it calculate the rest of the numbers based on what I input in the additional sheets for each cell. I don't have an automatic process in place for detecting the part number of the cells and didn't want to type in each cell when I start, so I just sort them by part number, give them a sequence number, measure IR and put them in the sheet. If they have a good IR, I start them on their test trial to get their capacity. (hence my thread here, as I didn't find the "Good IR value for these cells" ) I top-charge them, leave them sitting for 3 weeks, test if they are above 4.1 and then put them in the capacity tester. I don't top charge again before testing. I test from whatever the charge value was, down to 3V (not the recommended value from the spec sheet). This might give me lower capacity results, but I'm feeling more confident it will remove more bad cells from my database. If they are below 4.1, they get charged and go to the bench for another 3 weeks. I start the same procedure again, except this time I drain them if they fall below 4.1 again. I figure there must be something wrong with the cell then. If IR detection didn't find this problem, something else must be off.

This 3 weeks rest might also explain the lower capacity of the US18650V3 cells. I will just take a random sample of 12 cells (that's how many I can test in one run), top charge them, rest for 1 hour and then run a capacity test on them. See how the figures compare to the rested values.

I believe these cells are still in good shape, as most af them were still well above 4V when I disassembled the packs. I don't know the self discharge rate. Unfortunatly I can't test at 5A discharge rate... I can do a cutoff at 2.5, but the cap graph seems fairly steep from 3V to 2.5V
1627915023432.png

@Korishan as to the load of the packs. I wasn't planning on allowing more then 250mA current to flow trough one cell in the pack. So I will baby these batteries with the best of my knowledge. I should probably create a post in the project page, explaining what I'm planning to build and my current status of the project. Will be next on my todo list when I have another day off work.
 

ziporah

Member
Joined
Apr 26, 2021
Messages
45
So I randomly picked 10 cells out of the 280 and charged them, rest and test. This is the result:
numbercapacityresistancevoltage
30161943,14,128
501997334,139
69200851,34,167
87188744,14,167
1102027334,177
131192644,34,102
171209633,64,149
187178638,24,076
217194048,54,168
231197247,34,146

rated cap = 2150
above 95% = + 2042 => 1 cell
95-90% = 2042 - 1935 => 5 cells
90-80% = 1935 - 1720 => 3 cells (require the baby sitting and <500ma per cell )
below 1720 => 1 cell
The distribution is not that bad. The IR measurements are the least one could say curious ...
I can't find a good IR measurement to determine if I should continue the testing for a cell or just toss them before even getting started.
I should probably just filter these cells on self discharge rate after 3 weeks and capacity and ignore the IR, as long as I put them all in the same pack and string.
 

Overmind

Member
Joined
Jan 16, 2019
Messages
450
V3s were the tech-up of the 1500mAh V1s and they ended up being excellent cells.
I've been using many of them after 7 years or more and they still had full capacity (over 2Ah).
Sold them slowly as replacements for med-current power tool batteries (only have some Sanyos left now).
 

ziporah

Member
Joined
Apr 26, 2021
Messages
45
I tested all of my V3 cells now, the ones which had good juice in them before I started testing, performed well to excellent. The ones that were on the low side or recovered from 0V are all far below the 80% limit. The IR on all cells was in the same range, if you look at the first IR graph, the recovered cells are all located at the end. I'll create a capacity graph, as soon as I entered them in my sheet. If I just had to select on IR, I would say stick between 30-35 and you have selected the best cells. Unfortunatly I don't have enough cells to just select those, so I'm going to have to raise the IR requirement a bit. I plan on building a 48V powerwall, so I'll need a lot of them if I only want these in a single 14S string. I still have a lot unopened packs with V1 cells, but I didn't test any of them yet, so I don't know if they are still any good and if they can be used to increase the number of cells so I could build a 14s20 or 14s40 pack.
 

ziporah

Member
Joined
Apr 26, 2021
Messages
45
finished testing all cells ... from a total 280 cells, only 188 made it in the selection ...
1629532381728.png

so to build at least one pack, I would need to find me at least another 100 good cells in my US18650V packs or find someone who has at least 100 good US18650V3 cells.
total capacity of the tested cells top charged and tested to 3V is 1375,8516 Wh, 170 of the 188 cells are in the 90-100% range, which is quite nice if you ask me.
 

ziporah

Member
Joined
Apr 26, 2021
Messages
45
@Overmind , @Wolf I can buy NC1 cells, would it be "wise" to put them together with the V3 cells in a pack?
I see the IR is between 18-28 mOhm, so probably another chemistry ... not wise to mix and match them, must continue the search
 

Overmind

Member
Joined
Jan 16, 2019
Messages
450
It would really depend on what you want to use them for.
Generally, should be fine to put in parallel amounts of NC1s and V3s and then these packs in series, but there's a chance the V3s degrade faster than normal and that would affect the pack. Since you have plenty in good shape I don't see any good reason not to use them. Just make sure you do some periodical maintenance.
 
Top