Li-ion cells cycle ageing

[docware]

what is your indication why the 30Q suck so much?
spec sheet wise it should behave better.

If we read 30Q specification, it´s obvious that this cell was designed to provide high power output. Low internal resistance and high charge and mainly discharge currents is characteristic for high power cells, typical use is in accu tools.
From the 30Q spec it´s obvious that high power is preffered, not cycle life.

For high power cells in our cycling test is characteristic fast capacity loss in the first 20 cycles. 30Q is the most pronounced example, VTC6 is similar, strangely enough HG2 is evidently better. Obviously LG is leading both in high energy (MJ1, M36) as well as high power category.
Pajda is probably chuckling how much I am praising LG products now.

What is reason for such quick initial capacity loss ? Frankly, I don´t know. Maybe it´s connected to higher surface of electrodes, thinner electrodes and lower electrolyte amount to electrodes surface ratio. This lower ratio could lead to quicker loss of active lithium (loss of lithium inventory) during initial SEI build up. Silicon presence in the anode is aggravating factor. SEI growth leads to capacity fade due to the consumption of lithium that would otherwise be available for cycling.

NCA chemistry of 30Q and VTC6 may also be reason for generally worse capacity decline comparing to probably NMC chemistry of HG2, even probably NMC 811, according to this study :

https://www.mdpi.com/2313-0105/5/4/64/htm

View attachment Design Strategies for High Power vs. High Energy Lithium Ion Cells.pdf



It may be myriads of another reasons and their combinations. Frankly, I am not expert in this area and all above written is my pure speculation. I prefer testing of real samples instead of speculations.
I take the liberty to modify well known quote : One good test is worth more than a thousand opinions.

 

[docware]

From the chart it looks like 35E could be the new GA in terms of capacity decay.
I also wonder how a 18650 cell of LifePO4 would come out in tests like these.
It might reveal something we never thought of, or just imagined how it should behave by myth.

Frankly, I expected 35E to be worse than GA, but now after 300 cycles 35E has better trend.
Surely it would be interesting to compare LifePO4, just not sure what specific cell. Moreover, in April or May I´ll suspend testing because I will not be able to keep temperature tolerance.

New data : Samsung 30Q 700 cycles, LG MJ1 658 cycles, SONY VTC6 416 cycles, Samsung 50E 350 cycles, Samsung 35E 324 cycles, LG HG2 281 cycles.

 

[john61ct]

it would be interesting to compare LifePO4

Yes that would be fantastic.

3.0 - 3.45V range, for longevity 1C max discharge 0.4C charging

Anyone know what Mr Ping uses?

A123, now LithiumWerks is the gold standard, would I think blow the rest away. APR18650M1-B

ANR26650M1A are old model, ANR26650M1B newer

______
Full River (fullriver) makes 32700FE

Shandong Goldencell Electronics Technology Co
(is also Full River? same as Heter/ENERpower?)
HTCFR26650

______
VariCore apparently is Liitokala, as well as Vbatty? 32700

______
K2 LFP26650EV

Hixon IFR26650-35A

 

[docware]

New data : LG MJ1 700 cycles, SONY VTC6 482 cycles, Samsung 50E 407 cycles, Samsung 35E 400 cycles, LG HG2 362 cycles.










Samsung 30Q and LG MJ1 finished testing. At 4,2 – 2,5 V / 1 A discharge has 30Q after 700 cycles 93 % of initial capacity and 92,8 % of initial energy and 84,6 % of initial capacity in the cycling interval 4,1 – 3,3 V / 2,5 A.



MJ1 after 700 cycles has 95,1 % of initial capacity and 95,4 % of initial energy at 4,2 – 2,5 V / 1 A discharge and 94 % of initial capacity in the cycling interval 4,1 – 3,4 V / 2,5 A.

 

[john61ct]

Seems to be there is a tendency for the lower resistance (higher power) cells to lose capacity more quickly?

 

[docware]

30Q is the worst from high power cells. During testing after 700 cycles i have learned that 30Q No3 has some inner instabilities. See the first discharge 1 A 4,2 – 2,5 V :



Here see comparison of 30Q No4 (non cycled) with 30Q No3 initial test, again some strange anomaly :



Moreover, voltage of fully charged 30Q No3 is falling quite fast, also had some strange discrepancy during after 700 cycles SOC measurement. It seems to me that Cephalotus´s (and SamRich´s) remarks on 30Q bad reputation are based on true facts:
https://endless-sphere.com/forums/viewtopic.php?f=14&t=103092&start=100

 

[Folkert]

Hi docware,
Following this topic with great interest, especially since these sort of tests are not done a lot. Thanks, keep up the good work!

 

[docware]

New data : SONY VTC6 536 cycles, Samsung 50E 450 cycles, Samsung 35E 450 cycles, LG HG2 412 cycles.

 

[Dak77]

The samsung "E" cells all seem to mirror each other in the DCIR track. I know, duh, they're probably all based upon the same manufacturing process and cathode make-up . So, you could say, the trade offs are clear. As far as samsung is concerned, if you want the highest capacity and cycle life, you go with 50E, but sacrifice bulk and cost. If you want tolerance to some abuse with a little more balance of the tradeoffs , you go 29E . The 35E for some reason seems to be the choice for prebuilt packs, maybe because of availability and bulk discounts to vendors, but it doesn't seem to offer more usable capacity or cycle life than 29E. It does have very similar DCIR throughout it's life, so it will have similar voltage sag as 29E, 50E, and LG's M36 . I guess the point is, there doesn't seem to be a clear reason why they would not use 29E over 35E , but it isn't a horrible option ,since that's the better choice over the other common pre built offerings, which usually include GA, 35E , or PF . The rising internal resistance of the Panasonic/Sanyo offerings make them a no-go for me. Sag just sucks. It seems needless to say, if MJ1 for low power needs , M36, or M29 are available at similar cost you take those above all else. /rambling observation over

 

[docware]

New SOC comparative chart after 700 cycles. Samsung 30Q and LG MJ1 added.

 

[Cephalotus]

Moreover, voltage of fully charged 30Q No3 is falling quite fast, also had some strange discrepancy during after 700 cycles SOC measurement. It seems to me that Cephalotus´s (and SamRich´s) remarks on 30Q bad reputation are based on true facts:
https://endless-sphere.com/forums/viewtopic.php?f=14&t=103092&start=100

Yes. Some 30Q cells will develop very high self discharge rates. This can happen after just 20-30 cycles and makes the cell packs more or less unuseable, because most BMS are not able to balance those packs, especially if you do not charge the packs very frequently.

Thank you for testing the HG2. This is the cell I did chose over the 30Q.

 

[Cephalotus]

is that right, do you think?
is it the same with the usual e-bike batteries?

(car) Lithium-ion battery degradation over time
Temperature 40% charge 100% charge

0 °C (32 °F)
40% charge: 2% loss after 1 year
100% charge: 6% loss after 1 year

25 °C (77 °F)
40% charge: 4% loss after 1 year
100% charge: 20% loss after 1 year
...

I did store Panasonic PF at +5°C and ca. 40% SOC for 5 years and they lost maybe 3% of capacity.

My oldest cell is a LCO cell from sharp. Had more than 90% capacity after 18 years of storage at room temperature. Now after 22 years it degraded more rapidly

 

[Pajda]

My oldest cell is a LCO cell from sharp. Had more than 90% capacity after 18 years of storage at room temperature. Now after 22 years it degraded more rapidly

Yes this was definitelly a bad one, next time they should offer a better cells or I never again buy anything from Sharp!!!

But back to the results, after some arguing with docware i decided to perform a simple calendar (shelf) life test for modern cells. It will be LG Chem vs Samsung SDI duel. The cells are: M36, M29, 50E and 29E7 (aka best cycle life cells from docware tests). Each model have two contenders. First is stored at 100% SoC and the second at 50% SoC. Each month there will be performed a nominal test (standard battery testing porn stuff like: discharge retention capacity, nominal capacity test, DCIR 10s and IEC test at 50% SoC). And those test will be performed with professional $$$ tester (this time no ZKE, but it will handle it too).

 

[john61ct]

Once a year would be enough.

I've seen LFP prismatics isolated at 50% for ten years, lost less than 2%, still much higher than mfg capacity rating.

Hot in the summers too. . .

 

[john61ct]

I think actual literal self-discharge

meaning no cycling factors at all, just shelf storage at midpoint / calendar aging only

is actually just not a thing with LI.

Holdover concept from lead thinking.

Testing every month could actually introduce interference from the cycling wear.

Since storing at Full is known to be bad practice, not sure worth actually testing that?

Midpoint isolated is the way to go.

 

[zeccato]

is that right, do you think?
is it the same with the usual e-bike batteries?

(car) Lithium-ion battery degradation over time
Temperature 40% charge 100% charge

0 °C (32 °F)
40% charge: 2% loss after 1 year
100% charge: 6% loss after 1 year

25 °C (77 °F)
40% charge: 4% loss after 1 year
100% charge: 20% loss after 1 year
...

I did store Panasonic PF at +5°C and ca. 40% SOC for 5 years and they lost maybe 3% of capacity.

My oldest cell is a LCO cell from sharp. Had more than 90% capacity after 18 years of storage at room temperature. Now after 22 years it degraded more rapidly

Good to know.
But I don't think you used them on e-bikes,
it's not like you put a fridge on your bike

 

[docware]

... i decided to perform a simple calendar (shelf) life test for modern cells.

This type of test has great advantage – you don´t need a testing cycler nor fixtures.
What temperature do you consider for this test ? By the way, for acurate DCIR measurement the cell´s body temperature ( not ambient air) should be within 0,1 -0,2 °C.

Regarding frequency of capacity and DCIR testing, every month is maybe unnecessarily often, but once a year is definitely too big interval.

 

[john61ct]

In hot temperatures biannually maybe quarterly

In cold storage that frequently the test inaccuracy will be higher than actual changes

 

[Pajda]

Thanks for a valuable remarks, yes I can imagine other test setup and maybe in the near future me or someone else will perform them as well. To my test I have to say that I am not fan of using absolute values and I prefer relative comparison. This is also the answer on what temperature did I chose. It is funny to look in datasheets of different cells from one manufacturer, where they did not follow one settings. You can see that for example LG Chem cells have RT declared as 23, 24 and even 25°C also with different range, mostly +/-2°C but you can find even +/-5°C. Also value of nominal DCIR have significant range, for example LG M36 should have DCIR 30s of 36 mOhm +/- 6 mOhm so anything in the nominal range of 30-42 mOhm is OK.

So yes, I can also use $$$ temperature chamber with better than 0.1°C precision, but there two scientific reasons why I decided not to do it. The first is my laziness, the second is that I prefer to test all samples under the same conditions all the time, no matter what those ambient conditions are. But I'll consider the chamber.

My interest is focused on verifiing of these two phenomena:
1/ Check the monthly discharge rate aka "Storage Characteristics". In datasheets most producers declaring from my point of view very conservative "Energy remaining rate ≥ 90%". This datasheet value is declared for 100% SoC.
2/ The wear level of cell, particularly to verify if storage at SoC 100% is significantly worse than 50% SoC for modern High Energy cells. This is aiming to the presented value(I find it on more sources) of 20% wear after 1 year storage at 100% SoC @ RT. This is so huge wear level where the dozen of nominal cycles cannot play significant role.

 

[ginekolog]

Thanx for lab tests, its good work.

Now real life test.
I have 900Wh Samsung 35E 52V 14S 5P Jumbo Shark battery ( https://fasterbikes.eu/de/akkus/225/900wh-flaschenhalter-akku-52v-14s-5p-jumbo-shark?c=47 ) on Giant Trance X2 + BBSHD. I ride single mountain trails.
I take good care of battery:
- charge to 90% with 5A, discharge to 10% (~45V) with average of 20A (1000W). Peaks 1600W for few minutes.
- storage and charge at 22C
- dont ride with cold battery
- average batt temp is ~40C, max 57C (very rare).

After about 1 year, 4000 km and 100 cycles I measuerd my battery with 200W load. Capacity went down from 17Ah to 15Ah (850Wh to 750Wh). So now I am at 88%. :? This is confirmed with my 200W load calculation and BMS display (BT app).
Internal resistance is ok for now, at 30A it sags for 5.5V. This means 65 mOhm per cell.

I am still satisfied with performance if it stays this way for some years. From your graphs it seems it hopfuly will :wink:

 

[serious_sam]

Internal resistance is ok for now, at 30A it sags for 5.5V. This means 65 mOhm.

R = V/I = 5.5/30 = 183 mOhm

 

[Honk]

I think he means 65mOhm per cell.
And that makes sense when calculating.

 

[Dak77]

I think I remember reading a while back that 35E is very sensitive to charge current , so they may be very decent when charged at low rates around 500-700ma . Just a hunch.

 

[serious_sam]

I think he means 65mOhm per cell.
And that makes sense when calculating.

Yes, that makes more sense.

 

[docware]

..... but there two scientific reasons why I decided not to do it. The first is my laziness, the second ......

Yes, those scientific principles are important, thanks for inspiration ……

Well, the absolute storage temperature is not so important, as all samples will have the same conditions and this is comparative test. You probably store them at room temperature, which is Ok as it simulate normal conditions.
However, if you want to be able to see subtle DCIR changes, temperature and soaking time during measurement should be always the same as much as possible.

I was rethinking of the testing frequency. I think that testing each month is OK, because it simulate in same way not frequent using of the cells. Thus the case when cycle wearing is not dominant over calendar aging. Which is closer to the real life than mere storage.