Li-ion cells cycle ageing

Batteries, Chargers, and Battery Management Systems.
Post Reply
Dak77   100 W

100 W
Posts: 102
Joined: May 24 2019 11:57pm

Re: Li-ion cells cycle ageing

Post by Dak77 » Nov 22 2019 4:55pm

eMark wrote:
Nov 22 2019 3:41pm
docware wrote:
Nov 22 2019 4:14am
Hopefully I am not annoying too much, if I again recommend to consider Pajda´s posts as a required reading. Just open some text editor and with ctrl C / ctrl V do your own Reader´s Digest.

The quantity of needed separated cycle-life tests starts in my opinion at number 3 or 4 at least, ending somewhere at ….. ?
Docware's reply provided the means to answering john's two questions.

For the sake of discussion let's assume a pack with 3000mAh 10A cells can be charged and discharged at least 100 times at it's 10A rating from 4.1V to 3.3V. Furthermore we'll assume it can be charged 500 times to 4.1V with a 2.5 amp charger (.83C) with worthwhile useable energy for 500 cycles.

So for the sake of discussion we'll assume you can charge the pack 500 times at .83C (2.5A) and discharge it 500 times at .83C (2.5A) over three years of occasional easy-going ebiking without any noticeable voltage sag. So, following that ultra simplistic TLC logic if you want at least 500 charge/discharge cycles over three years don't discharge the cells at more than .83C. Using that KIS logic it's possible to get at least 500 charge/discharge cycles from either a: VTC6, HG2, 30Q, PanasonicPF, SanyoGA and almost any 8P14S 18650 pack and possibly a hundred more cycles :thumb:

Now for those that contend this post is not practical and is derailing the thread then is this thread purely hypothetical bench testing that may not even apply to a real-life daily ebiking commuting routine?

The problem with my scenario is that you can't go 40 or 30 mph on your ebike if you want to get at least 500 cycles. According to Dak77 it's not possible to get 500 cycles with a daily commuting ebike routine using say, 3450mAh 5A LG Chem M36 cells or any other 18650 cell pack (eg. 8P14S) ... unless you're only going 25mph with a fair amount of pedaling and even then 500 charge/discharge cycles may be pushing it according to Dak77 ??
I think you misunderstood what I was saying. I have no idea if it's possible to get 500 cycles using an M36 at high loads. In your 8p configuration , I would bet that it will based on testing posted here, if your maximum current drain is less than 30 amps total . It seems more likely the M29 will fit that role better though. That's more for someone like Doc/Flippy/Padja to say. What I meant was, by the time you wear down a high discharge cell , "most" people would be ready to upgrade to a newer cell or chemistry , whichever. If you can get 3 years use out of a cell that can handle full throttle smashing with little sag, it makes more sense to go that route than to build a monster sized pack so you can keep it for 1000 charges, in which it will die of old age long before it dies from charge/discharge cycles. This point of view is only from someone that rides recreationally rather than for transportation every day.

docware   1 kW

1 kW
Posts: 313
Joined: Sep 16 2018 5:36am

Re: Li-ion cells cycle ageing

Post by docware » Nov 23 2019 3:05am

SOC % versus voltage chart updated with LG HG2 and SONY VTC5A, 0 % and 5 % columns modified to gray.
Take this chart with some restraint, only as informative, not absolutely true values, especially at 0 % and 5 % also. Chart was initially created to specify rough 50 % and 30 % SOC values for storage. Moreover, measurement was done on relatively new, not very used cells. Not sure if these values will be the same after few hundres cycles.

docware   1 kW

1 kW
Posts: 313
Joined: Sep 16 2018 5:36am

Re: Li-ion cells cycle ageing

Post by docware » Nov 24 2019 2:57am

New data, PF, M36 and GA 350 cycles, 29E 300 cycles.
We are slowly approaching to 500 cycles, which rises question to Pajda : should we continue beyond this target or start to cycle new cells instead ? Does cycling beyond 500 cycles provide information interesting enough to give reasons for such long term locking of testing rig ?
Panasonic  PF 350 cycles.jpg
Panasonic PF 350 cycles.jpg (162.63 KiB) Viewed 2993 times
LG M36 350 cycles.jpg
LG M36 350 cycles.jpg (160.09 KiB) Viewed 2993 times
Sanyo GA  350 cycles.jpg
Sanyo GA 350 cycles.jpg (164.03 KiB) Viewed 2993 times
Samsung 29E7 300 cycles.jpg
Samsung 29E7 300 cycles.jpg (161.25 KiB) Viewed 2993 times
Comparison 300 cycles.jpg
Comparison 300 cycles.jpg (45.34 KiB) Viewed 2993 times

eMark   100 W

100 W
Posts: 210
Joined: Nov 02 2019 11:53am

Re: Li-ion cells cycle ageing

Post by eMark » Nov 24 2019 11:02am

docware wrote:
Nov 24 2019 2:57am
We are slowly approaching to 500 cycles, which rises question to Pajda : should we continue beyond this target or start to cycle new cells instead ? Does cycling beyond 500 cycles provide information interesting enough to give reasons for such long term locking of testing rig ?
From Dak77's ebiking real-life knowledge acquired over years about ebikers and their packs it's his belief that 100-200 cycles is more realistic to the actual cycle life of most ebikers ageing packs. So, even 350 cycles may be quite sufficient for docware's life cycle ageing tests as most ebikers may never rack up 350 charge/discharge life cycles whether it be a 36V, 48V or substantial 52V pack.
docware wrote:
Nov 19 2019 12:27pm
Some people on the Pedelec forum mentioned soldering as a root cause of the issue (30Q self-discharge).
Cephalotus wrote:
Nov 19 2019 5:34pm
This was the guess at the beginning that they (30Q) don't like soldering. But later spot welded packs had shown the exact similar behavior.

Some people have been very happy with the 30Q at least for some weeks until their packs got faulty, too and it does not affect all cells, maybe "just" 5-10 cells out of 100 (this is a wild guess), but sadly if you build packs with 50 cells and more your chances are very high that you will have faulty cells.

Some people are desperate to identify "winners" and some people develop blind spots after identifying their chosen winner.
Something to consider is what was once a fault of 30Q "136" cells may no longer be a fault with "141" cells. Wouldn't Samsung make every effort if the fault could be corrected with re-engineering and/or Quality Control. Japan isn't the only country with "zero tolerance" for defects when it comes to QC (something Chevy and Ford had to learn from Honda and Toyota). At least since 9/19 current 30Q cells are now labeled as "141" instead of "136". Whether or not this ("141") is of any significance when it comes to resolving a "heavy self-discharge" of say 5-10% of previous 30Q 136 cells is unknown.

JMO, that Samsung's priority for the importance of "zero tolerance" for defects is just as critical as is Japan's ... so it's probable that Samsung has already resolved the 5-10% defective issue with the "self-discharge" [previous] fault of the 30Q 136 cell.
Pajda wrote:
Nov 18 2019 4:59pm
3. "3000 mAh" category of high power cells with discharge currents more than 3C continuous and up to 10C peak. "Big four" offers: Samsung 30Q, LG HG2 and SONY VTC6
Is there a single cell -- 30Q, HG2 & VCT6 ... ageing test that docware or Padja could conceive that docware would be willing to conduct that might possibly confirm that say 5-10% of "30Q 141" cells are still more prone to suffer from [heavy] self-discharge moreso than say a HG2 or VCT6 cell ???

It stands to reason that docware and Padja could conceive of an [abusive(?) 100-200 cycle] "ageing" test on a 30Q, HG2 & VTC6 that might confirm whether the more recent "30Q 141" cell still suffers from [heavy] self-discharge. Maybe, three comparative [rather abusive?] 100-200 cycle ageing tests of: 30Q, HG2 & VTC6. So, how would docware or Padja construct a 100-200 cycle ageing test that possibly might exacerbate (make evident) what is believed to still be a self-discharge failure of 30Q "141" that has never been observed/mentioned as a fault with a HG2 or VTC6 cell ?

The PROBLEM in devising such a test is that in order to determine if 5-10% of 30Q 141 cells are still inferior one would have to test as many as one hundred 30Q 141 cells :( and that in itself makes such a test most unlikely.

Pajda   1 kW

1 kW
Posts: 339
Joined: Jan 24 2016 3:45am
Location: Czech Republic

Re: Li-ion cells cycle ageing

Post by Pajda » Nov 24 2019 2:03pm

docware wrote:
Nov 24 2019 2:57am
New data, PF, M36 and GA 350 cycles, 29E 300 cycles.
We are slowly approaching to 500 cycles, which rises question to Pajda : should we continue beyond this target or start to cycle new cells instead ? Does cycling beyond 500 cycles provide information interesting enough to give reasons for such long term locking of testing rig ?
Nice question :mrgreen: I will start with the answer that there is no good solution. One way or another you will be blamed that you chose the other way. Not only by the community but even yourself. At this settings (70% SoC window with ca 0.3-0.4C charge and 0.7-0.9 discharge rate) you should expect at least 1500+ cycles for good quality cells (by my standards) before they reach the EoL. Now it is clear that PF does not make it and GA will be on the border line at best (particularly because of DCIR rise). There is always a dilemma if you prematurely break the test and then you decided to continue some months or even years later, if this is good for methodology. I am ending all of my tests at 1000 cycles. I only tested LG HG2 up to 3500 cycles for about two years at ca 80% SoC window (4.1V-2.5V, 0.5-1C) and "this beast just didn't want to die".

If you insist on my clear answer :D I think that for the community will be right now more usefull to end your cycle life tests at 500 cycles and start with four new cells. But I am aware that this does not help much, because it only create bunch of another questions, particularly which cells and which settings? For my point of view there are two basic approaches:

1) continue with the same settings on the other HE cells or best price per kWh cells: the clear candidates are for example M29, 35E and MJ1
2) switch to HP cells. But there is another dilemma if it makes sense to test them under this (low power app) settings or to choose another, more aggresive setting. I will suggest 5C discharge rate, but you are limited by your tester up to 3C rate for 3000mAh cells. The problem with 3C discharge rate is that cells like M29 and 29E can still give to HP cells "a good fight" at 3C continuous. Yes, M29 and 29E will show significantly higher DCIR for the whole test, but they still can do at least 500 cycles (M29 will do 1000+ cycles) at this torture before reaching their EoL.

So you have my full confidence, that you choose the best option. I will only suggest to add as soon as possible as a "special guest" Samsung 30Q no matter on which settings it will run.

docware   1 kW

1 kW
Posts: 313
Joined: Sep 16 2018 5:36am

Re: Li-ion cells cycle ageing

Post by docware » Nov 24 2019 4:56pm

Thank you for your thoughts, they are helping me to assort my own. I don´t want to go to 1 000 cycles, my idea was to finish somewhere at 600 – 700 cycles.
Another testing fixture arrived this week, so from Friday is 30Q already running on fifth channel at the same parameters. MJ1 will start next week completing my max testing capacity 6 channels.
Other candidates are 35E, VTC6, M29, HG2, 50E, …..
My testing parameters are different than your, but from your remarks and answers I comprehend that up to now are cells doing according your expectation, isn´t it ?
What is your criteria for EoL ?

takyka   10 mW

10 mW
Posts: 30
Joined: Apr 01 2015 2:25am
Location: Hungary

Re: Li-ion cells cycle ageing

Post by takyka » Nov 25 2019 3:16am

Docware,

First, I would like to thank you for your efforts, I follow it with great interest!
I would vote for longer test run, as I'm not the one who willing to spend on new batteries as long the present one works, just because in a few years a better, fancier is out...
For me, would be good to see if PF capacity fade (and DCIR rise) will stop at a certain point or will continue the current trend until death. I'm disappointed on the performance until now. Somehow it was common opinion here, by using it gently, it will live very long. But the trend and capacity loss looks exactly the same like the graph in the datasheet, for 0.5/1C full cycle tests. (~80% capacity loss by 500 cycles)

I'm wondering if the datasheet lies, or the PF does not really care about discharge depth?

T.
T.

brone   10 mW

10 mW
Posts: 25
Joined: Jun 26 2018 2:45pm

Re: Li-ion cells cycle ageing

Post by brone » Nov 25 2019 3:41am

takyka wrote:
Nov 25 2019 3:16am
Docware,

First, I would like to thank you for your efforts, I follow it with great interest!
I would vote for longer test run, as I'm not the one who willing to spend on new batteries as long the present one works, just because in a few years a better, fancier is out...
For me, would be good to see if PF capacity fade (and DCIR rise) will stop at a certain point or will continue the current trend until death. I'm disappointed on the performance until now. Somehow it was common opinion here, by using it gently, it will live very long. But the trend and capacity loss looks exactly the same like the graph in the datasheet, for 0.5/1C full cycle tests. (~80% capacity loss by 500 cycles)

I'm wondering if the datasheet lies, or the PF does not really care about discharge depth?

T.
My previous encounter with PF is that they were very saggy at 2C. Later I got the IMR chemistry 30Q's that I'm super satisfied with the sag. Now I know that they aren't comparable because 30Q is a HP cell, but I assume another IMR cell would perform identically, like Pajda mentions that the M29 holds good for 1k cycles even at 3C

docware   1 kW

1 kW
Posts: 313
Joined: Sep 16 2018 5:36am

Re: Li-ion cells cycle ageing

Post by docware » Nov 25 2019 4:48am

takyka, those parameters are behind gentle utilization, so no such big surprise. PF is obviously not the best choice, don´t understand your nostalgia for it. You have better options like 29E or newly M29. Or M36.

User avatar
flippy   1 MW

1 MW
Posts: 1668
Joined: Aug 12 2015 3:07pm

Re: Li-ion cells cycle ageing

Post by flippy » Nov 25 2019 8:05am

docware wrote:
Nov 25 2019 4:48am
takyka, those parameters are behind gentle utilization, so no such big surprise. PF is obviously not the best choice, don´t understand your nostalgia for it. You have better options like 29E or newly M29. Or M36.
the PF's used to be the cheapest kid on the block. now tis overtaken by the superior 29E6 and 7.

i got a bunch of 500+ cycle 29E's. once my new shop is up and running i will put those on the tester and run them until they are dead.
Lithium beats liquid dinosaurs.

Pajda   1 kW

1 kW
Posts: 339
Joined: Jan 24 2016 3:45am
Location: Czech Republic

Re: Li-ion cells cycle ageing

Post by Pajda » Nov 25 2019 3:16pm

docware wrote:
Nov 24 2019 4:56pm
My testing parameters are different than your, but from your remarks and answers I comprehend that up to now are cells doing according your expectation, isn´t it ?
What is your criteria for EoL ?
Yes, they are. Your combination of narrow SoC window and particularly rather slow charge rate helps a lot Panasonic/Sanyo cells to show at least satisfactorilly results. This will be also the case for Samsung 30Q.

EoL assessment depends on application, but for general traction it is the 100% DCIR rise from initial value. Only majority of LG production in 18650 shows similar trend for both DCIR rise and capacity fade and so you can use as estimate only capacity criterion. For other production, particularly from Panasonic/Sanyo in 18650 it is necessary to track DCIR rise, because it can reach 100% rise in just 500 cycles, where the remaining capacity should be still rated as ok.

User avatar
multifrag   100 W

100 W
Posts: 120
Joined: Feb 19 2016 5:44am
Location: London, UK

Re: Li-ion cells cycle ageing

Post by multifrag » Nov 25 2019 4:28pm

docware wrote:
Nov 24 2019 4:56pm
Thank you for your thoughts, they are helping me to assort my own. I don´t want to go to 1 000 cycles, my idea was to finish somewhere at 600 – 700 cycles.
Another testing fixture arrived this week, so from Friday is 30Q already running on fifth channel at the same parameters. MJ1 will start next week completing my max testing capacity 6 channels.
Other candidates are 35E, VTC6, M29, HG2, 50E, …..
My testing parameters are different than your, but from your remarks and answers I comprehend that up to now are cells doing according your expectation, isn´t it ?
What is your criteria for EoL ?
I'd be really interested in M29 cell testing as it has similar price to performance ratio to 29E6. It's been reported that it might have similar cycling life to 29E6 too. Planing on order 250 cells, would be nice to know if they will last

takyka   10 mW

10 mW
Posts: 30
Joined: Apr 01 2015 2:25am
Location: Hungary

Re: Li-ion cells cycle ageing

Post by takyka » Nov 25 2019 4:51pm

docware wrote:
Nov 25 2019 4:48am
takyka, those parameters are behind gentle utilization, so no such big surprise. PF is obviously not the best choice, don´t understand your nostalgia for it. You have better options like 29E or newly M29. Or M36.
I don’t have nostalgia, just arround 170 cells in two packs. With (the original) life expectancy arround 10years/~1000cycles in similar usage scenario to your tests! My purchase decision was made based on datsheet figures, the general consensus about much longer cycle life with shallow discharge. It looks, this isn’t hold with PF.

T.
T.

docware   1 kW

1 kW
Posts: 313
Joined: Sep 16 2018 5:36am

Re: Li-ion cells cycle ageing

Post by docware » Nov 26 2019 1:49am

takyka wrote:
Nov 25 2019 4:51pm
I don’t have nostalgia, just arround 170 cells in two packs.
170 cells, that´s different story. If possible try to have lower DOD and charging current. DOD for PF in this testing is cca 73 %, which is quite high.
Both PF and GA are now slowing down capacity decline. We should also realize the we are comparing PF and GA to 29E and M36, members „Best of“ team.
On the other hand, this is only „laboratory“ testing, you will see what lifetime your packs will have in real life. Maybe not 10 years, but your battery packs may serve you long time.
I have 50 GA cells in my ebike battery. This is my third year, I am still satisfied. But my DOD is shallow, charging current formerly 0,8 A / cell, now cca 0,5 A.

Dak77   100 W

100 W
Posts: 102
Joined: May 24 2019 11:57pm

Re: Li-ion cells cycle ageing

Post by Dak77 » Nov 26 2019 2:01am

Just know that what you're doing is extremely helpful and valuable to the community. Anybody that implies otherwise is an idiot. If this group of tests had been posted 6 months ago, I would have never used PF over 29E. The info was out there, just not as clearly and nothing as blatant and matter of fact is it is now. Thank you for your efforts.

mighty82   10 mW

10 mW
Posts: 34
Joined: Jan 29 2017 5:45am

Re: Li-ion cells cycle ageing

Post by mighty82 » Nov 26 2019 6:36am

docware wrote:
Nov 23 2019 3:05am
SOC % versus voltage chart updated with LG HG2 and SONY VTC5A, 0 % and 5 % columns modified to gray.
What happened to the chart? I get this error:

"The selected attachment does not exist anymore"
Last edited by mighty82 on Nov 26 2019 6:37am, edited 1 time in total.

Cephalotus   10 kW

10 kW
Posts: 672
Joined: Jun 18 2012 12:27pm

Re: Li-ion cells cycle ageing

Post by Cephalotus » Nov 26 2019 6:36am

docware wrote:
Nov 24 2019 2:57am
New data, PF, M36 and GA 350 cycles, 29E 300 cycles.
We are slowly approaching to 500 cycles, which rises question to Pajda : should we continue beyond this target or start to cycle new cells instead ? Does cycling beyond 500 cycles provide information interesting enough to give reasons for such long term locking of testing rig ?
If you cycle your cells within a limted SOC window as you do it, you do that with the intention to get 2000+ cycles.

I see no point in stopping the tests at 500 cycles.

If you use the full SOC window you optimce for Maximum capacity and than 300-500 cycles could be fine.

Now many people blame the Panasonic PF as an inferior cell.

I do not believe that. PF always lost 10-15% capacity early and stabilised after that. It still can hold very well for 2000cycles+ and end up better than the other options, we just don't know yet.
Last edited by Cephalotus on Nov 26 2019 6:53am, edited 1 time in total.

Cephalotus   10 kW

10 kW
Posts: 672
Joined: Jun 18 2012 12:27pm

Re: Li-ion cells cycle ageing

Post by Cephalotus » Nov 26 2019 6:47am

docware wrote:
Nov 24 2019 4:56pm

Another testing fixture arrived this week, so from Friday is 30Q already running on fifth channel at the same parameters. MJ1 will start next week completing my max testing capacity 6 channels.
Other candidates are 35E, VTC6, M29, HG2, 50E, …..
You should do what you prefer. Ist is your setup and your work and your money. There is also no need to hurry...

My suggestion would be HG2 and 35E (and others if you like) and test them the the exakt same way as you did the other cell tests so far.

Only so the tests are compareable

If you finihed your first round of cycle tests and your 4.1V~3,4V Setting use the same cells and test them on more demanding settings or however you would like it and you will be able to compare them again.

I would not suggest testing diferent cells only at different parameters. If you change more than one parameter (different cells already is one parameter) test results are of low value imho.

john61ct   100 GW

100 GW
Posts: 5002
Joined: Dec 18 2018 2:06pm

Re: Li-ion cells cycle ageing

Post by john61ct » Nov 26 2019 7:49am

Cephalotus wrote:Now many people blame the Panasonic PF as an inferior cell.

I do not believe that. PF always lost 10-15% capacity early and stabilised after that.
I see that as a major flaw, for me a dealbreaker since they aren't much cheaper.

If they were accurately sold labeled as 2500mAh capacity. . .

takyka   10 mW

10 mW
Posts: 30
Joined: Apr 01 2015 2:25am
Location: Hungary

Re: Li-ion cells cycle ageing

Post by takyka » Nov 26 2019 9:27am

docware wrote:
Nov 26 2019 1:49am
170 cells, that´s different story. If possible try to have lower DOD and charging current. DOD for PF in this testing is cca 73 %, which is quite high.
Both PF and GA are now slowing down capacity decline. We should also realize the we are comparing PF and GA to 29E and M36, members „Best of“ team.
On the other hand, this is only „laboratory“ testing, you will see what lifetime your packs will have in real life. Maybe not 10 years, but your battery packs may serve you long time.
I have 50 GA cells in my ebike battery. This is my third year, I am still satisfied. But my DOD is shallow, charging current formerly 0,8 A / cell, now cca 0,5 A.

I don't think calendar life alone will be an issue for me.
I have a 6S4P PF pack in stand up scooter for four years now. Less than 100 cycles, but pulling 3C for acceleration and up to 1,5C continously. A bit more saggy now, but no drama on range, luckily. If trend will be the same for the future, I can believe in another four years.

T.
T.

takyka   10 mW

10 mW
Posts: 30
Joined: Apr 01 2015 2:25am
Location: Hungary

Re: Li-ion cells cycle ageing

Post by takyka » Nov 26 2019 9:29am

Cephalotus wrote:
Nov 26 2019 6:36am
I do not believe that. PF always lost 10-15% capacity early and stabilised after that.
This is what I'm hoping, but yet to see...

T.
T.

eMark   100 W

100 W
Posts: 210
Joined: Nov 02 2019 11:53am

Re: Li-ion cells cycle ageing

Post by eMark » Nov 26 2019 10:40am

docware wrote:
Nov 25 2019 4:48am
takyka, those parameters are behind gentle utilization, so no such big surprise. PF is obviously not the best choice, don´t understand your nostalgia for it. You have better options like 29E or newly M29. Or M36.
Cephalotus wrote:
Nov 26 2019 6:36am
Now many people blame the Panasonic PF as an inferior cell.

I do not believe that. PF always lost 10-15% capacity early and stabilised after that. It still can hold very well for 2000cycles+ and end up better than the other options, we just don't know yet.
john61ct wrote:
Nov 26 2019 7:49am
I see that as a major flaw, for me a dealbreaker since they aren't much cheaper.

If they were accurately sold labeled as 2500mAh capacity. . .
Pajda wrote:
Nov 19 2019 2:55pm
Present number one is LG M29. It can handle 3C continuous (single cell on free air) "with the finger in the nose". At the setting: 0.5C charge - 3C discharge at 100% DoD it have still ca 88% of initial nominal capacity after 900 cycles.
Cephalotus wrote:
Nov 19 2019 12:08pm
Some people are desperate to identify "winners" and some people develop blind spots after identifying their chosen winner.
So, if there is a blind spot is it with those still favoring the NCA chemistry of PanasonicPF and SanyoGA cells or with those favoring NMC chemistry like that of the LG Chem M29 cell ?? Perhaps, Cephalotus best sums it up when he says ...
Cephalotus wrote:
Nov 19 2019 9:46am
Cycle tests are only half of the story.
... and furthermore ...
Cephalotus wrote:
Nov 19 2019 12:08pm
I don't try to fine "the best" cell, because this is very difficult.

For me chosing a "good enough" cell is good enough :-) And good enough always depends on your personal specifica.
So, what conclusion(s) can we affirm so far from docware's cycle tests and Padja's posts that a newbie lurking guest can take to the bank? Is it a general consensus that both docware and Padja are downrating the Japan PanasonicPF and SanyoGA cell (NCA chemistry) in favor of South Korean LG Chem cells (NMC chemistry) like 29E, M29 and M36?
docware wrote:
Nov 26 2019 1:49am
Both PF and GA are now slowing down capacity decline. We should also realize that we are comparing PF and GA to 29E and M36 ... this is only "laboratory“ testing ...
I have 50 GA cells in my ebike battery. This is my third year, I am still satisfied. But my DOD is shallow, charging current formerly 0.8 A / cell, now cca 0.5 A.
Has LG Chem of South Korea surpassed Japan in the production of quality NMC chemistry 18650 cells that are becoming more popular with DIY ebikers than are the PanasonicPF and SanyoGA NCA chemistry cells of Japan ??

Typo Correction: NMC not NCM as first typed, but now corrected
Last edited by eMark on Nov 26 2019 12:28pm, edited 1 time in total.

john61ct   100 GW

100 GW
Posts: 5002
Joined: Dec 18 2018 2:06pm

Re: Li-ion cells cycle ageing

Post by john61ct » Nov 26 2019 11:50am

I think your "industry trend analysis" is a bit of a stretch. Is this research for an article? Generalities aren't as useful as the actual test reports.

Up until recently, longevity wasn't even considered an important factor by most.

Also, better check / cite your chemistries

I had Samsung INR18650-29E down as NCA?

eMark   100 W

100 W
Posts: 210
Joined: Nov 02 2019 11:53am

Re: Li-ion cells cycle ageing

Post by eMark » Nov 26 2019 1:43pm

john61ct wrote:
Nov 26 2019 11:50am
Up until recently, longevity wasn't even considered an important factor by most.

Also, better check / cite your chemistries. I had Samsung INR18650-29E down as NCA?
You asked for sources so referenced a couple below. With just a little googling you could've easily discovered that 29E is NMC not NCA chemistry. As previously mentioned in another post INR is the same chemistry as NMC ...

"This NMC chemistry adds nickel to the IMR chemistry, making it a "hybrid" chemistry. It combines the safety and low resistance of manganese and the high energy of nickel. There is extensive innovation within this chemistry as well."

Lithium Manganese Oxide LiMn2O4 -- IMR = LMO
Lithium Manganese Nickel LiNiMnCoO2 -- INR = NMC
Lithium Ni Co AL oxide LiNiCoAlO2 -- NCA

"At present, batteries with cobalt (NMC/NCA) are preferred due to their high energy density. However, safety remains an issue as NMC and NCA are prone to thermal runaways — especially as the devices get smaller. NMC chemistries have become the preferred technology of original equipment manufacturers in the automotive industry in recent years.

According to our research, NMC will become the first to dominate the EV sector in the early 2020s, but will be overtaken by LFP batteries when a new rise in cobalt prices coincides with a boost of LFP’s energy density achieved through R&D."

https://www.greenbiz.com/article/who-wi ... ttery-wars

"One of the biggest talking points in the electric vehicle (EV) and lithium ion battery space right now is the emergence of an 811 NCM (Nickel Cobalt Manganese) cathode chemistry that many believe will result in the hammer blow for cobalt and the anointing of nickel. For those newer to the battery industry, 811 quite simply stands for 8 parts nickel, 1 part cobalt, 1 part manganese. It is the natural evolution from the traditional 111 chemistry (equal parts of each), 523 (5 parts nickel, 2 parts cobalt and 3 parts manganese) and 622.

In fact, it is Benchmark‘s expectations that 811 cathode material will reach 25% of NCM’s market share by 2026 – enough to cause disruption in the raw material supply chain, but far from takeover. Much like their Chinese counterparts, Korean and Japanese manufacturers are also developing 811 formulations at the moment and their arm has been forced by two main reasons (continued) ..."

https://www.benchmarkminerals.com/nicke ... n-battery/

When it comes to long term Li-ion storage it's docware's belief that ... "Different chemistry, manufacturer, electrolyte additives, ….. may have little bit different results ... however calendar aging strongly depends on time, SOC and temperature" (see posts #16, #18, #21) ... "This very informative study shows a higher degradation of their NCA cathode material at higher voltages, leading to increased internal cell resistance" (#23) ... http://budgetlightforum.com/node/47730

According to docware Li-ion NCA vs NMC chemistry isn't as important when it comes to the storage of Li-ion cells. Maybe, maybe not (NCA vs NMC) when it comes to self-discharge, but may be relevant to cycle life ageing (e.g. DCIR). It is important for DIYers to know the chemistry of their Li-ion battery pack. For example the Samsung INR18650-29E is not NCA, but rather NMC as relevant info for this ageing thread (e.g. DCIR). Whether or not docware and Padja prefer NMC 18650 cells to NCA cells for DIYers has not yet been clearly stated (so far), but does seem like they favor NMC.

DIYers have their own pros and cons. Often price, availability of useable salvaged cells and personal preference from real-life ebiking battery pack experiences can take precedence over LMO (IMR) vs NCA vs NMC (INR) cell chemistry. Possibly a valid reason for why neither docware or Padja have had much to say about Li-ion chemistry in this cycle ageing thread.

john61ct   100 GW

100 GW
Posts: 5002
Joined: Dec 18 2018 2:06pm

Re: Li-ion cells cycle ageing

Post by john61ct » Nov 26 2019 2:40pm

TL;DR

Just one definitive link please, showing 29E is NCM, no need to quote extraneous kruft

My quick search shows reviews stating several different chemistries, including NCO just now, LiNiCoO2.

Post Reply