Li-ion cells cycle ageing

docware

1 kW
Joined
Sep 16, 2018
Messages
339
This topic arise from the Pajda´s idea in another thread :
https://endless-sphere.com/forums/viewtopic.php?f=14&t=102387&p=1507776#p1507776

Frustrated and tired by never ending unfruitful debate about cycle life of various Li-ion cells, I finaly decided to start some long term cycling to see the real results. Parameters was chosen in effort to imitate real utilization, but on the other hand to get some results in acceptable time. Important was to not use parameters like 100 % DOD, 0,5 C charging, 1 C discharging, as these are in my opinion too severe and unrealistic in the real life.

As representants of the typical cells was chosen LG M36, Sanyo GA, Panasonic PF and Samsung 29E7. Chosen parameters for high energy cells M36 and GA : charge 1A to 4,1 V, 5 minutes wait, discharge 2,5A to 3,4 V, 5 minutes wait. For „middle energy“ cells PF and 29E was discharge modified to 3,3 V to keep the same DOD. This is working for M36 and PF, to my surprise GA and 29E are quite out of the DOD range.
In the hindsight discharge could be lower, maybe 2,2 A .

Regarding using terms 1C, 0,5C, …. This terminology is on one hand very useful, on the other hand it is means that for example discharging SONY VTC5A at 2,5A and Samsung 35E at 3,4 A is considered as the equivalent load to both cells. Which is not true.

Equipment : cycle tester ZKETECH EBC-X0510 (8 channels), true 4-wire fixture BF-2A, electronic load Maynuo M9712, DMM GW Instek GDM-8351, thermometer Omega HH 520, K thermocouples.


ZKETECH EBC-X0510.png
BF-2A a.jpg
BF-2A b.jpg
Maynuo.jpg
GDM-8351.jpg
Omega HH 520.jpg

ZKETECH EBC-X0510 and electronic load Maynuo M9712 calibration was checked with the help of DMM GW Instek GDM-8351, DMM Brymen 869s and 10 miliohm 0,1% etalon Riedon RSN-10-100B concentrating especially to long term current stability. ZKETECH EBC-X0510 accuracy is better than 0,3 %, Maynuo M9712 accuracy is better than 0,05 %.

True 4 wire fixtures BF-2A or its similar predecessors are used for all measurement including cycling. Ambient temperature is in the range 25 ± 1°C for DCIR and 25 ± 2°C for cycling.
DCIR is measured according IEC 61960-3:2017, paragraph 7.7.3. Measuring parameters are unified to 0,6/3 A at cca 3,74 V for all cells.

Here are first data in the graphic form :

Panasonic  PF cycling.jpg

LG M36 cycling.jpg

Sanyo GA cycling.jpg

Samsung 29E.jpg

It is always useful to state as much relevant data as possible. All these data are useful for those who are in bussines and also for author, to remind in future immediatelly the conditions during measurement. Without them is measurement not very credible. Authors of the reports should also be able to answer additional relevant questions.
 
The Luna Wolf offers two options in their battery pack ... https://lunacycle.com/luna-wolf-v2-52v-battery-pack/

  • Liquid - Samsung 30q 12Ah pack (15A cell rating)
  • Liquid - LG MJ1 13.5Ah pack (10A cell rating)

If you have time could you run a comparison test on these two 18650 cells.
Are they both NCA/NCR chemistry or is the Samsung NMC chemistry?
 
brone said:
Why didn't LG MJ1 get chosen?

You could also ask : why not Samsung 35E, Sony VTC6, Sanyo 20700B, Samsung 30Q , Sony VC7, LG MH1, Samsung 50E, LG M50T, …….. ?
LG M36 got priority because were discussed in the thread https://endless-sphere.com/forums/viewtopic.php?f=14&t=102387
and finally also purchased by author of the topic (Jan-Erik-86).
 
eMark said:
  • Liquid - Samsung 30q 12Ah pack (15A cell rating)
  • Liquid - LG MJ1 13.5Ah pack (10A cell rating)

If you have time could you run a comparison test on these two 18650 cells.
Are they both NCA/NCR chemistry or is the Samsung NMC chemistry?

Hopefully Pajda give you comprehensive answer.
 
doc you really didn't answer our posts as whether or not you will run a test on both the Samsung 30q and the LG MJ1. At least you could have provided a link, but apparently my ? isn't relevant to your thread :wink:
 
docware said:
brone said:
Why didn't LG MJ1 get chosen?

You could also ask : why not Samsung 35E, Sony VTC6, Sanyo 20700B, Samsung 30Q , Sony VC7, LG MH1, Samsung 50E, LG M50T, …….. ?
LG M36 got priority because were discussed in the thread and finally also purchased by author of the topic.

Do you think LG M36T is the same as the M36, because at nkon there's only a 5A rated M36T, that's why I'm asking about the MJ1 - because it's a 10A rated cell
 
docware I forgot to warn you when I was writing my post to create a new topic, that you should expect a flame of posts with question why did you do not test another cells. And here it is... :mrgreen: :mrgreen: :mrgreen:

I have some unwanted advice for readers of this topic.

I am in the "battery expertise business" for more than ten years and I can say, that docware`s cell aging (cycle life) tests are valid. It means that in the first case he does not make errors in measure chain (like use of crappy charger/tester, ignoring the 4-wire measure, ignoring the ambient temperature...) and so this is one of the few trustable results which you can found on internet for free!!! We all should thanks to docware that he makes his results public. I and my donor decided to keep our test results private for many reasons. And so I am only commenting them.

From my point of view there are two working approaches if you would like to see results for another cells:

1) Inspire with the test settings and fixtures and do your own tests. I use a proverb that good charger is usually a very bad cell tester. So If you think about your own test equipment, forgot about all "modeler chargers" (even expensive/profi ones) and buy ZKE Tech equipment. ZKE testers have also many flaws but it was proven than they can reliable do the job with satisfactory results.

2) Make a donation project for docware, that he can buy another 8-channel tester, cell fixtures and cell samples. I think he does not want any income for the work (which is of course the most expensive part of this fun). I would be one of the first donor.
 
Padja,

Take Xception to your "flaming" remark as flaming in itself. Your reply isn't helpful!

My previous reply including the chemistry NCA/NCR or NWC of the cells doc is testing and planning on any future testing is for sure more valid than is your remark ... "here it is... :mrgreen: :mrgreen: :mrgreen:" ... which does nothing to further the intent/clarification by the Op of this thread's purpose.
.
Possisbly because brone and myself are recent ES members U may mistakenly assume our posts are not relevant to ebiking 18650 cell longevity bench testing interests of ebikers and possibly two of the most produced and popular cells in use by DIY ebikers.

STOP posting if your only purpose is to make a false flamming accusation when my posts are certainly relevant to this thread.

What you could/should have posted is the link that doc was referring to in his reply to me OR was he just making jest?
 
eMark, first read all Pajda´s posts here on ES and you will know answers to your questions.
 
The bottom line is, those publishing objective real-life test results are performing a hugely valuable public service **for free**.

Expressing **anything** but gratitude, much less an entitlement to more than what they are volunteering for, is just pure scumbaggery.

Sure OK, politely suggest or request other cells, but expressing an opinion implying "they should" do what you think, really, just beyond the pale.
 
That 29E is looking like a really good deal for daily 80% DOD use under moderate load. I'm curious if it ever starts degrading rapidly or just trends slightly down over time.

From a purely selfish point of view I'd love to see some of the cheap jumbo QB cells tested since they are excellent value per wh but little is known about lifespan.
 
M36 probably has never been 10A cell, always 5A only. At least it´s not possible to find such 10 A specification. However, there is a contradiction between specification description (M36T) and cell label (M36). By the way, there is no such thing as M36 3600 mAh cell, real capacity is always bellow 3 400 mAh (fresh new cell). It is only nkon bussines label as probably 10 A was some time ago.
 
Very nice and highly appreciated.

But may I suggest something.

For comparison I would prefer if you would keep the axis fixed, lets say capacity from 1000mAh to 3000mAh on each diagram and Ri from 0 mOhm to lets say 150mOhm on the other axis.

Maybe x-axis fixed to 0-1000 cycles, too. (I don't know how many cycles you plan to do)

On your charts the axis differ and so comparison is much more difficult than it needs to be. (yes, I notcied that your 2nd axis is alway a delta of 5mOhm, but I still would prefer it to be fixed)
 
Pajda said:
2) Make a donation project for docware, that he can buy another 8-channel tester, cell fixtures and cell samples. I think he does not want any income for the work (which is of course the most expensive part of this fun). I would be one of the first donor.

If this is an option I could imagine to donate the costs for a cell test via Paypal depending on how much it costs. An estimation would be nice.
 
Thank you very much for publishing your results, it's for sure been very helpful and interesting for me, and I'm sure it is (and will be) for many others as well for years to come! :)
I'm really looking forward to seeing how they compare after more cycles, both in terms of DCIR and capacity loss.

There is one more thing I'm curious about, and maybe someone can shine some light on it. It might be a bit off topic as far as "cycle life ageing" goes, but it for sure does have to do with ageing still.
It's said a cell should not be stores at high SoC for any period, but I'm sure many of us still do. Personally I charge my bike up every evening before bed so it's ready for use next morning, and while I know this is not ideal, I charge indoors and would rather wear more on the battery then not wake up one day.

So here's my question; Does all these cases wear equally much on the battery, as they have the same total time (ignore the wear from cycles)?:
1) Battery charged to 4.2v and kept at that over night (say ~7 hours) for 100 cycles. Total "high SoC storage time" = ~30 days.
2) Battery charged to 4.2v and kept at that for 3 days for 30 cycles. Total "high SoC storage time" = ~30 days.
3) Battery charged to 4.2v and kept like that for a month. Total "high SoC storage time" = ~30 days.

If the answer to this is "yes", I would assume that storing a battery at full charge for about 3-4 months would equal to the same "high SoC storage damage" as charging up the battery before going to bed for about a year. I know this would be a time consuming test, but it would be very interesting to see how this actually affects the different cells in terms of capacity loss and DCIR. Maybe also a comparison of storing at 4V, 4.1V and 4.2V, to see what difference it actually make, and how it affects capacity and DCIR.
If total "high SoC storage time" affects the battery a lot, I think a test/comparison of this for different cells could be almost as interesting and important as cycles alone.

Of course, this was just a thought crossing my mind... :)
 
Jan-Erik, you obviously know that you are unnecessarily wearing the cells by your charging scheme.
Why don´t you buy timer and start charging in the night to finish just in time cca 30 – 40 minutes before your departure ?
 
Jan-Erik-86 said:
1) Battery charged to 4.2v and kept at that over night (say ~7 hours) for 100 cycles. Total "high SoC storage time" = ~30 days.
2) Battery charged to 4.2v and kept at that for 3 days for 30 cycles. Total "high SoC storage time" = ~30 days.
3) Battery charged to 4.2v and kept like that for a month. Total "high SoC storage time" = ~30 days.

For daily use I believe it is wise to use a battery that does not need 1005 of ist rated capacity in everyday application. It#s wise to use just 50-60%.

if you only use 50-60% of the capacity you do not need to charge to 4,2V/cell, but 4,0V~4,1V/cell should be fine, too.

If you use the byttery that way for maybe 200-250 times a year calendar aging should not be to dominant, you should still be able to get 1000 cycles (=4-5 years) from good quality cells used that way.

if the battery ages over time and loses useable capacity you may charge a bit higher and compensate for some capacity loss and higher Ri that way.
 
It's not really about having a timer on it or not for me, it's more because i don't completely trust the safety of charging li-ion cells unattended. Surely it goes well in 99.9% of the cases, but after reading what happened to dogman dan for example, I simply don't want to take that chance as long as I'm charging indoors two floor below me in a wood house.

Like you say, I know my charging scheme causes unnecessarily wearing on the cells, but my question is simply how much wear are we actually talking about? And is it just capacity loss, just increased DCIR, or both? Have there every been made any tests on this at different SoC (and maybe temps)? And if so, is the capacity loss/increased DCIR "identical" for "all" cells, or does it vary a lot from cell to cell, just like cycle life?

I've seen a wide range of rated "capacity loss after storage" in spec sheets, but they are with different temperatures, storage time, and mostly at 100% SoC - but this only tells you how much capacity the cell will have left, and nothing about the permanent capacity loss caused by the storage. For the M36 for example, it say: "30 days storage at 4.2V at 24c +/- 2c = 90% capacity remaining. "
Sure, but if you recharge it again, will it still only have 90% of original capacity, or is it back to normal? What about the DCIR?

Cephalotus said:
For daily use I believe it is wise to use a battery that does not need 1005 of ist rated capacity in everyday application.
Very true. I only wrote 4.2V as an example. I personally charge to around 4.07V at 0.2C, and with a room temp of around 15c i hope overnight storage won't affect them much, but I'm still curious to know, that's all.. :)
 
eMark, first read all Pajda´s posts here on ES and you will know answers to your questions.
Your WG comment is condescending and not helpful. If it isn't too much bother U could've given a link to a post where Pajda mentions the chemistry of the Samsung 30q and LG MJ1 as both being NCA/NCR chemistry or if the Samsung 30q is NMC chemistry and the LG MJ1 NCA or NCR chemistry. Is it possible that even Pajda doesn't know?
Knowing the Li-ion chemistry of the cells doc is testing is relevant to this thread.

I've seen a wide range of rated "capacity loss after storage" in spec sheets, but they are with different temperatures, storage time, and mostly at 100% SoC - but this only tells you how much capacity the cell will have left, and nothing about the permanent capacity loss caused by the storage.

The above quote by Jan-Erik-86 is relevant to this thread and would be nice to know when buying cells (e.g. IMR or BQ) for a DIY 18650 project. For example the data sheet for LG MJ1 lists the capacity at 3500mAh with a minimum capacity of 3400mAh ... suggesting it may be a NCA or NCR chemistry?? Thus, one reason for doing an extended bench test between the "Liquid" - Samsung 30q 15A cell and so-called "Liquid" - LG MJ1 10A cell would be nice to know for DIY builder being that they're two popular cells mass produced in large numbers. A DIY builder would like to know not only the chemistry of certain IMR and BQ cells they plan to purchase, but also the same run number for the cells they are considering purchasing (before purchase) for their DIY build project.
 
Cephalotus said:
I personally charge to around 4.07V at 0.2C, and with a room temp of around 15c i hope overnight storage won't affect them much, but I'm still curious to know, that's all.. :)
Have had my 10S4P easygoing recreational pack at 4.00V for the past 4 days as its been too chilly outside. As you know an 80% charge (3.96V to 4.05V Li-ion cell voltage) depends on who you want to believe establishes what is considered 80% ... i just settle at 4.00V

My easygoing recreational pack will be drained somewhere between 3.6 to 3.7 volts this evening for longtime winter storage until next spring. Thought i'd get in one more ride this year, but doubtful now. JMO that it does no harm to have a pack at 80% charge for 2-3 days. doc should have an opinion if even a week is OK at an 80% charge. The problem with any thread is having doc set some parameters for discussion. For example how long one can store a pack at an 80% charge may not be within the scope and breadth of this thread by doc.
 
Jan-Erik-86 said:
It's not really about having a timer on it or not for me, it's more because i don't completely trust the safety of charging li-ion cells unattended. Surely it goes well in 99.9% of the cases, but after reading what happened to dogman dan for example, I simply don't want to take that chance as long as I'm charging indoors two floor below me in a wood house.

OK, I understand. Then you have to accept some little bit higher ageing. If you have enough capacity reserve, you can charge to lower SOC, see the table SOC % versus unloaded, resting voltage. If you have possibility to charge at your job, even lower SOC is possible. Take the chart as informative only, especially SOC low end.

SOC% versus voltage 22.11.2019.jpg

I doubt you find exact answers to your highly theoretic speculative questions. Here is again some reading, but don´t overtax your head needlessly. :)

https://www.researchgate.net/publication/328086745_Extending_Battery_Lifetime_by_Avoiding_High_SOC
https://mediatum.ub.tum.de/doc/1355829/file.pdf
 
Thank you very much for the links and voltage SoC chart for different cells, i found it very useful. :)
I'm unable to charge at work, but I have a lot of reserve capacity and will lower my charge voltage a bit more.

I understand my question may be speculative and theoretical, but I can't be the only one curious if there's a SoC where you are able to get a decent amount of enough energy out of the pack and still do minimal damage by keeping the cell at, and if it is different for each cell or not. There's a lot of tests and discussion about DoD and it's impact on DCIR and capacity over time, but very little about what impact different SoC actually haver over time (other then any time at 100% SoC is bad, and 30-50% SoC is good for storage).
But I'll start reading your links and see if i can learn some more about it. :)
 
Citation from „Extending Battery Lifetime by Avoiding High SOC“ :

The result showed that, when only considering ageing from different types of driving in small Depth of Discharges (DODs), using a reduced charge level of 50% SOC increased the lifetime expectancy of the vehicle battery by 44–130%. When accounting for the calendar ageing as well, this proved to be a large part of the total ageing. By keeping the battery at 15% SOC during parking and limiting the time at high SOC, the contribution from the calendar ageing could be substantially reduced.

Citation from „Aging of Lithium-Ion Batteries in Electric Vehicles“ :

Calendar aging strongly depends on time, SOC and temperature.

A picture is worth a thousand words, more to five pictures :

1.jpg
2.jpg
3.jpg
4.jpg



Your 15 °C is helpful for inactive period, but is adverse for charging and discharging.
Of course, these studies were done on cell with NCA chemistry. Different chemistry may have different results. But generally accepted agreement higher SOC is more adverse to cells is valid.
We only have to get into the fact, that 30 – 50 % is not only good for long term storage but also for storage generally. An we have to get into the fact that not charging or driving (discharging) means inactivity period, thus storage.
I personally would not go lower than 20 %, to be away enough from critical 10 % range.
 
Back
Top