Li-ion cells cycle ageing

[Pajda]

My unwanted remarks to calendar aging:

1) If there are few people who professionally measure the cycle life, there is almost no people who measuring calendar aging and give his results for free. I am also not measuring this particular parameter, at least for now.

2) My personal opinion is that almost all information about calendar aging available on internet are belonging to my favourite category "old wives' advice". It is no offence to other experts, but I think that most of the available results are either outdated or measured at overal shitty cells (like the cycle life tests results) and so did not reflect parameters of newer generations of cells.

3) I am expecting huge difference in this parameter, even for cells from the same category, like you can see in docware cycle life tests, so generalization won't work here either.

 

[docware]

Wow, Pajda that is very brave statement with regard to the fact you never tested calendar ageing. What is your opinion based on ?
Do you consider both studies I mentioned also worthless ?

 

[Tolkano]

Good thread so far, been said many times though can you test 30q? Not everyone wants to build a 30kg battery out of 29e's

 

[Pajda]

Wow, Pajda that is very brave statement with regard to the fact you never tested calendar ageing. What is your opinion based on ?
Do you consider both studies I mentioned also worthless ?

Brave statement :thumb: ..good one

There must be always some disruptive element, or naysayer. Yes, I know both works and in one case is used as test sample: Panasonic PD, which is without hesitation a piece of crap if we talking about cell life parameters. In the second case is used sample described as: "A 26 Ah commercial pouch battery cell" which convinces me, that they used the best available cell on the market. :wink:

It is all about generalization. If you do not have all data/informations it is absolutely ok, that you expect if Panasonic PD was chosen by the expert (by the way he thinks that it is a Tesla Model S cell, which is obviously not) that other cells from the same size of "2900mAh" would show the similar results and then the "3500mAh" cells must be a total disaster.

 

[john61ct]

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%

No, any SoC point is completely objective, and not hard to measure even very accurately.

But there are a thousand different objective "recipes" for getting to that specific point, in fact an infinite number.

And the exact resting voltage corresponding to that SoC varies for each battery, and in fact varies by temperature and changes as cells age.

So just as with the infinite varieties of gender, sexual preferences and relationship status, "it's complicated"

Good news though, the whole issue is moot, for healthy storage parameters, precision is 100% unnecessary. 40% is better for long term storage than 80%, but for a few hours 80% is better than 90%.

The relative difference in "damage", i.e. cycles lost off the back end, between these different points will never be quantified, too many variable and other intertwined factors.

And extreme longevity is not even important to many owners, so just inform yourself and use your own judgment, in line with your needs and preferences.

 

[Pajda]

Just the last comment about calendar aging topic before I went banned from ths forum as heretic.

Common sense told me that if ca 10 years ago was problem to keep lithium cell charging to 4.2V unattended without fire extinguisher and then if the cell survived the charging process, it was dead in 3 month after you left it rest on this voltage. Today we have cells with 4.4V charging voltage (3.9Vnom) used in almost all smartphones and so one might think, that manufacturers could be inspired by this HV materials and so materials used in modern cells with 4.2V working voltage could last a little bit longer.

But you are right, it probably doesn't work that way and so 4.2V is bad and ugly voltage.

 

[john61ct]

I think the general factors involved in reducing the impact from calendar aging are clear and well proven.

Disconnect cells from each other, and any potential parasitic loads.

Temperature, lower the better, up to a point

SoC, lower the better, up to a point

That's it!

The individual chemistry and build quality differences will impact the actual numbers involved, those of course are all over the place.

My experience with LFP has shown me that unused NOS cells stored properly are perfectly usable even after 5 years.

If the use case involved daily heavy DoD usage, so not trying for 20-year lifespan, I'd pay 70-80% compared to new (fresh mfg date) price for such cells.

Otherwise maybe 50-60%.

With the other LI chemistries my bid would likely be much lower, just from lack of knowledge.

 

[eMark]

Good thread so far, been said many times though can you test 30q? Not everyone wants to build a 30kg battery out of 29e's

Thanks doc for posting that chart ... https://endless-sphere.com/forums/download/file.php?id=261819 ... including both the Samsung 30Q and LG MJ1. Just in case you didn't notice the 30% SOC for the MJ1 should be 3.57% instead of the typo error of 3.87%.

Also interesting to compare both the Samsung 35E 3500mAh 8A and 30Q 3000mAh 15A against the LG MJ1 3500mAh 10A below 50% to 0% SOC. At first sight this would 'suggest' that the LG MJ1 is a better chemistry and a better buy being that all three are the same price at IMR. When taking into account that considerate ebikers charge/discharge from 80% to 50% (increase cycle life of their pack) the chart (50% to 0%) does seem to 'suggest' that the LG MJ1 is a better buy.

What is different about the MJ1 data sheet compared to the 30Q data sheet is that the MJ1 lists its minimum rated capacity at 3400mAh suggesting the chemistry of the LG MJ1 is different than the chemistry of the Samsung 35E and 30Q. The minimum rated capacity of the 35E is 3350mAh; whereas the 30Q doesn't list a minimum rated capacity. This suggests that possibly the higher in mAh a cell is rated (i.e. 3500mAh) with the possiblity of more disparity between cells (with increasing cycle aging) with a 35E and MJ1 battery pack than with a 30Q battery pack should doc ever conduct a comparable bench testing comparison.

Thanks again doc for posting that chart

PS: Aging is the preferred spelling in North America. The British prefer "ageing", while Americans, Canadians and Australians prefer "aging"

 

[docware]

Just the last comment about calendar aging topic before I went banned from ths forum as heretic.

Common sense told me that if ca 10 years ago was problem to keep lithium cell charging to 4.2V unattended without fire extinguisher and then if the cell survived the charging process, it was dead in 3 month after you left it rest on this voltage. Today we have cells with 4.4V charging voltage (3.9Vnom) used in almost all smartphones and so one might think, that manufacturers could be inspired by this HV materials and so materials used in modern cells with 4.2V working voltage could last a little bit longer.

But you are right, it probably doesn't work that way and so 4.2V is bad and ugly voltage.

Heretic, disruptive elements are sometimes needed to help break and change the old rigid way of thinking.

Pajda, I know that you are very experienced in this area, therefore I assumed that maybe you know something more, than you want or can share. I know you as a realistic guy who after all these experiences trust only verified facts. Also understand your Panasonic PD argument, however, exist any study on any cell which has different conclusions on calendar ageing ?

I agree that each cell is different and also probably can have different calendar ageing characteristics, just wanted to remind that there is another aspect of ageing except cycle ageing. We just miss more reports on various cells. Otherwise, we are discussing cells commercially obtainable 18650 and 21700.

 

[docware]

Good thread so far, been said many times though can you test 30q? Not everyone wants to build a 30kg battery out of 29e's

OK, as 30Q is also very popular at the opposite side of the planet, it is next cell to test.
Initial 30Q capacity is slightly above 3000 mAh, initial 29E capacity is slightly bellow 2 800 mAh.

 

[eMark]

Disconnect cells from each other, and any potential parasitic loads.

"Plz Xpln" (Ricky to Lucy)) how you would "Disconnect cells from each other" (a DIY battery pack) for a Rip Van Winkle cool/dry storage scenario ?

The individual chemistry and build quality differences will impact the actual numbers involved, those of course are all over the place.

Xactly so doc's comparative testing is interesting, but in real-life ????? For example it is a waste of time to even think about doing comparative bench tests between a specific 2017 SDI cell serial run from South Korea, another SDI 2018 cell serial run from Malaysia and a SDI 2019 cell serial run from China.

Question: If IMR can't fill a specific 18650 battery order from the same serial run, but can supply half of the order with one serial run and the other half from another serial run how would you distribute these two different serial runs in a pack assuming IMR would be so considerate to comply ?

 

[Pajda]

Good thread so far, been said many times though can you test 30q? Not everyone wants to build a 30kg battery out of 29e's

OK, as 30Q is also very popular at the opposite side of the planet, it is next cell to test.
Initial 30Q capacity is slightly above 3000 mAh, initial 29E capacity is slightly bellow 2 800 mAh.

I also think that 30Q is the next candidate no.1 for docware test procedure. Lets see if at the opposite side of the planet this cell performs better than a total crap (even worse than PF) in cycle life and particularly DCIR rise parameter than on our side of the planet.

 

[docware]

Thanks doc for posting that chart ... https://endless-sphere.com/forums/download/file.php?id=261819 ... including both the Samsung 30Q and LG MJ1. Just in case you didn't notice the 30% SOC for the MJ1 should be 3.57% instead of the typo error of 3.87%.

eMark, thank you for notice on the typo error, MJ1 at 30% is on 3,57 V, not 3,87 V. Table is updated including new cell SONY VTC6.

 

[eMark]

OK, as 30Q is also very popular at the opposite side of the planet, it is next cell to test.

Apparently it helps if an ES member with at least 100 posts (tolkaNo with 116) is able to bend Pajda's ear and subsequently get doc's endorsement ...

Good thread so far, been said many times though can you test 30q? Not everyone wants to build a 30kg battery out of 29e's

I also think that 30Q is the next candidate no.1 for docware test procedure. Lets see if at the opposite side of the planet this cell performs better than a total crap (even worse than PF) in cycle life and particularly DCIR rise parameter than on our side of the planet.

Do you really need to perform a lengthy bench test to determine that :wink:

Now, if only someone with at least 100 ES posts (brone and myself aren't significant enuf) can twist doc's arm to include the LG MJ1 3500mAh (34oomAh min) 10A popular 18650 cell used by DIY builders. Got to believe that doc really wants to include the MJ1 in the next test, but needs a bug in his ear from an ES member with at least 100 posts and followed up with an endorsement by Pajda to seal the deal :thumb:

 

[Pajda]

eMark it seems to me, that for some reason you did not understand the idea of my message. Because I went through it in my beginning I know that there are no other options than I described in my first post in this topic.

docware has a 1pc of EBC-X0510 tester, this is an 8-channel tester. He is using four channels for the four different cell tests listed in this thread. If he would like to go up to 1000 cycles, it will took another ca 3 months. I remember that at least one channel on his tester is broken and he probably need to keep remaining channels unoccupied to perform immediate short-term tests.

That means that even he would like to perform the tests which you (and many others) asking for, he actually does not have possibility to do that. And even he finds one or two more free channels, there will rise demand for more and more tests (which is natural). Which brings us back to the possibilities I have already mentioned.

So to make it absolutely clear, eMark I am not telling you or any other, not to ask docware to include your preferred cell to his tests. I am only explaining why it won't work and what to do to make this possible.

 

[john61ct]

Post count has **zero** to do with credibility.

"Plz Xpln" (Ricky to Lucy)) how you would "Disconnect cells from each other" (a DIY battery pack) for a Rip Van Winkle cool/dry storage scenario ?

Actually much easier with a DIY pack, but some good builders do follow instructions for custom orders.

And I don't know why you think only useful for long-term storage, same protocol can apply for anytime the cells aren't being cycled, even if just for a day or two.

The answer is, you design your connection methodology with that functionality in mind, not just "possible", but so that it is easy.

 

[john61ct]

these two 18650 cells.
Are they both NCA/NCR chemistry or is the Samsung NMC chemistry?

I am also interested in such chemistry details. Few others here are. It would be a useful contribution for you to take that aspect on, do the required detective work to find out and post that info into the threads where the batteries you know about are being discussed.

 

[docware]

OK, as 30Q is also very popular at the opposite side of the planet, it is next cell to test.

Apparently it helps if an ES member with at least 100 posts (tolkaNo with 116) is able to bend Pajda's ear and subsequently get doc's endorsement ...

Good thread so far, been said many times though can you test 30q? Not everyone wants to build a 30kg battery out of 29e's

I also think that 30Q is the next candidate no.1 for docware test procedure. Lets see if at the opposite side of the planet this cell performs better than a total crap (even worse than PF) in cycle life and particularly DCIR rise parameter than on our side of the planet.

Do you really need to perform a lengthy bench test to determine that :wink:

Now, if only someone with at least 100 ES posts (brone and myself aren't significant enuf) can twist doc's arm to include the LG MJ1 3500mAh (34oomAh min) 10A popular 18650 cell used by DIY builders. Got to believe that doc really wants to include the MJ1 in the next test, but needs a bug in his ear from an ES member with at least 100 posts and followed up with an endorsement by Pajda to seal the deal :thumb:

OK eMark, enough is enough, this post https://endless-sphere.com/forums/viewtopic.php?f=14&t=103092&p=1507985#p1507985 was intended mainly for you.
If you want to continue with your self-centered arrogant and agressive manners, then rather STOP posting here please.

 

[Jan-Erik-86]

Citation from „Extending Battery Lifetime by Avoiding High SOC“ :

Thank you again very much for taking your time to both dig out, post and explain all these stuff. This thread was already good to begin with, but is now pure gold thanks to people like you.
I'm sorry if i dragged it a bit off topic though, from "Cycle Ageing" to "Calendar Ageing".

I find these graphs very interesting, and like you say, a pic is worth a 1000 words. Now i haven't read it all yet, so correct me if I'm wrong, but from what i can see on the graphs it shows 3 important things as far as calendar ageing:
1) store below 25c if you want it to last, period. There is however not much to be gained by storing at less then 25c unless maybe at high SoC.
2) If i understand the 2nd graph correct, the capacity loss due to SoC starts suddenly at around 55% SoC (even though it doesn't clearly appear so based on the other graph due to "medium SoC" is 50% and capacity loss starts at 55%), but there is little difference in the loss from 60% SoC and up as long as temp is at 25c or less, however;
3) Even though the capacity loss does not change much from 60% Soc and up at 25c or less, the DCIR will start to increase once stored above 20% Soc, and have a significant increase if stored above 80% SoC.

So in short, stay below 25c and below 80% SoC for daily use to prevent significant increased DCIR, and discharge to 20-30% SoC for long term storage. If this is correct, and does not vary too much from cell to cell, i think i got my question answered pretty accurate.

 

[john61ct]

store below 25c if you want it to last, period.

If you are looking for hard numbers and simple easy to follow "rules", fine.

But reality is greyscale and complex. Those specific graphs' numbers are only relevant for those cells. Some other similar chemistries, maybe similar numbers. But others, different numbers, different sloped curves, etc.

Better to just adapt **guidelines** to your use case, preferences, budget, etc.

Cooler is **always** better. Same with low SoC. Either factor helps longevity, but they work best together.

30% SoC at 10°C is much better than 80% at 24°

There is no harm in going to **your** definition of normal-cycling 100%, e.g 4.07Vpc at 0.2C

as long as there are loads ready to bring SoC% back down, ideally within the next hour or two.

IOW, do not sit at or even near Full - don't charge until just before discharge is needed.

Accuracy in specific numbers is irrelevant.

Design of the system for healthy habitual usage patterns are what leads to longer lifespan.

 

[Jan-Erik-86]

If you are looking for hard numbers and simple easy to follow "rules", fine.
But reality is greyscale and complex.

Got it, thanks!

My unwanted remarks

Your remarks are not unwanted at all.

OK, as 30Q is also very popular at the opposite side of the planet, it is next cell to test.

I'm looking forward to see how this HP cell compares with the other less powerfull/HE cells under the same conditions.

 

[docware]

I have some new data from cycling. Panasonic PF at 250 cycles keeps trend both in capacity decreasing and DCIR rise. LG M36 continue with very slow capacity decline. M36 DCIR got reverse trend, at 250 cycles is lower than 150 and 200 cycles. Can´t exclude my measuring error, trying to find out what is happening. Sanyo GA capacity decreasing trend continue, however DCIR rise stopped. Unfortunately thanks to today´s dropout of the ZKETECH control software I lost lot of Samsung 29E last batch data (cca 36 cycles) and also some Panasonic PF and Sanyo GA data. At least I measured 29E DCIR and continue cycling. 29E keeps DCIR still the same around 29,3 miliohm.

 

[eMark]

I am also interested in such chemistry details. Few others here are. It would be a useful contribution for you to take that aspect on, do the required detective work to find out and post that info into the threads where the batteries you know about are being discussed.

They're "hybrid" chemistries (proprietary secret recipes) e.g. 35E, MJ1, 30q, and that's basically one reason why doc evaded my question and brushed me off by telling me to go look at all of Padja's posts. Instead doc should have picked up on my post as an opportunity to tell others why he (and Pajda) considered the chemistries of the test cells outside the scope of his thread.

Like others that have so implied it would have been more beneficial if doc would have included two of the most popular cells used today and going forward by DIY builders (30q and MJ1 available since 2015) in his "ageing" cell tests. Being that for whatever reason(s) he should've at least explained WHY instead of ignoring and brushing aside ES inquiring minds.

What doc should have expected and therefore clarified from the get-go is why he choose not to include these two cells in his ageing cell tests being 30q and MJ1 are reasonably priced, easily available, and a popular choice by DIY ebikers. If doc thinks this post and my previous posts are arrogance on my part he is sadly mistaken.

Bottomline: At least doc should've briefly explained (as intro info thread clarification) instead of just talking down to me with his comment to go look at all of Padja's posts. In actuality i was providing doc an opportunity to explain why he felt is wasn't necessary to discuss the pros and cons of the chemistries of the cells in his ageing tests.

Was i being arrogant? No!

 

[john61ct]

Add my vote to banning this creep.

Certainly should at least be ignored by all in the meantime.

No clue wrt being a constructive member of the community.

 

[eMark]

Disconnect cells from each other, and any potential parasitic loads

The answer is, you design your connection methodology with that functionality in mind, not just "possible", but so that it is easy.

When i read your first post and your subsequent post the easiest method that comes to mind is using one or more Vruzend V2.1 kits depending on the size of the battery pack as one DIY design build for ease of disconnecting cells.

Perhaps what you're also referring to is when it comes time to salvage cells from a pack that has seen better days. It would be helpful if you could give us an example of a DIY build design that is both SECURE/SAFE during it's useful life; while still designed "so that it is easy" to disconnect cells.

We would appreciate your insight into an easier method of disconnecting cells that are spot welded to the buss bars?