Li-ion cells cycle ageing

Batteries, Chargers, and Battery Management Systems.
eMark   100 W

100 W
Posts: 278
Joined: Nov 02 2019 11:53am

Re: Li-ion cells cycle ageing

Post by eMark » Feb 23 2020 11:19am

Thanks! Maybe, i got a really good BG-8S or i don't have a good enough DMM ... or, maybe because the comparison was at 3.75V ... anyway close enuf for my scientific notation :wink:

john61ct   100 GW

100 GW
Posts: 6242
Joined: Dec 18 2018 2:06pm

Re: Li-ion cells cycle ageing

Post by john61ct » Feb 23 2020 11:20am

eMark wrote:If the disparity is as much as 30mV difference between one of the P-groups compared to the other P-groups that are within 10mV of each other it's not only time to balance charge, but there could be more serious problems down the road.
Nope sorry just not true.

I've seen many many packs go thousands of cycles just fine with huge imbalances, many for well over a decade.

So long as you just accept the overall pack capacity is limited by the weakest link (which is unavoidable anyway)

if the top and bottom stop-points are based on the first cell or group to get there, not the pack level voltage

or just calibrate your pack-level stop-points for the same result

then that "dire problem" you are warning about is a complete non-issue.

With "zero balancing ever", or middle balancing, then you will see imbalances both top and bottom, and that is just fine, no need to "do something" about it.


eMark   100 W

100 W
Posts: 278
Joined: Nov 02 2019 11:53am

Re: Li-ion cells cycle ageing

Post by eMark » Feb 23 2020 11:23am

docware wrote:
Feb 23 2020 10:52am
Citation from iSTD BG-8S manual :
„Voltage measurement accuracy : ± 0,005 V @ 4,2 V“

You have to also take into consideration that iSTD unit is supplied from some measured cells – another factor affecting the voltage. Resulting inaccuracy may be 10 – 40 mV, or more, who knows. Until you check the real voltages (first without BG-8S, then with BG-8S connected) with external voltmeter having sufficient resolution 0,1 mV and sufficient input resistance, you have no idea what´s truly going on.
Thanks! Maybe, i got a really good BG-8S or i don't have a good enough DMM or a little of both ... or ... maybe because the comparison was at 3.75V. Anyway close enuf for my scientific notation :wink:
Last edited by eMark on Feb 23 2020 11:24am, edited 1 time in total.

john61ct   100 GW

100 GW
Posts: 6242
Joined: Dec 18 2018 2:06pm

Re: Li-ion cells cycle ageing

Post by john61ct » Feb 23 2020 11:24am


docware wrote: I decided to use voltage reference MAX6341 also in this single purpose instrument to have all in one.
So the voltmeter accuracy is just being human-compared to the reference displayed, no self-calibrating going on?

If so can the voltmeter be manually calibrated if you see it's drifted?


eMark   100 W

100 W
Posts: 278
Joined: Nov 02 2019 11:53am

Re: Li-ion cells cycle ageing

Post by eMark » Feb 23 2020 11:50am

docware wrote:
Feb 23 2020 10:52am
Until you check the real voltages (first without BG-8S, then with BG-8S connected) with external voltmeter having sufficient resolution 0,1 mV and sufficient input resistance, you have no idea what´s truly going on.
Padja where are you when i could use your moral support :)
docware thinks he's talking to a newbie (not my opinion), but what do i know :wink:

How precise is precise enuf? It's not like we need Musk's precision for reading the remote voltages when his Tesla is cruising around on the Sea of Tranquility. Even then +/- 5mV may be close enuf between all the Tesla P-groups IF "P-groups" is still relevant nomenclature by then.

It's my understanding (could be wrong) that most all inexpensive BMSs only discharge balance when voltages between the P-groups is over 30mV. In other words a BMS won't even attempt to balance the P-groups in a battery unless disparity exceeds 30mV between the P-groups. So, how concerned does one really need to be if most all run of the mill BMSs don't even discharge balance unless there is at least a 30mV disparity ??

docware   1 kW

1 kW
Posts: 318
Joined: Sep 16 2018 5:36am

Re: Li-ion cells cycle ageing

Post by docware » Feb 23 2020 12:04pm

john61ct wrote:
Feb 23 2020 11:24am
If so can the voltmeter be manually calibrated if you see it's drifted?
No, volmeter can´t be calibrated. In fact, there is no reason as this simple-purpose instrument was built for comparative measuring of the parallel groups. So that absolutely accuracy was not the primary. However, thanks to built-in precise voltage reference, I know that measuring around 4,1 V is pretty accurate.

I don´t see any significant drift, just have to wait cca 8 – 10 minut after switch-on to see that reference voltage is stabilized at 4,0962 V.

john61ct   100 GW

100 GW
Posts: 6242
Joined: Dec 18 2018 2:06pm

Re: Li-ion cells cycle ageing

Post by john61ct » Feb 23 2020 12:19pm

Thanks.

is cca = "approx" ?

eMark   100 W

100 W
Posts: 278
Joined: Nov 02 2019 11:53am

Re: Li-ion cells cycle ageing

Post by eMark » Feb 23 2020 1:44pm

docware wrote:
Feb 23 2020 10:52am
Until you check the real voltages (first without BG-8S, then with BG-8S connected) with external voltmeter having sufficient resolution 0,1 mV and sufficient input resistance, you have no idea what´s truly going on.
You do realize that 0,1mV is actually 0,0001V. It isn't necessary to have a DMM that can read 4 places beyond the decimal point to know "what's truly going on". Scientific notation uses a period (not a comma) and generally three digits (e.g. 3.745V) is sufficient. For the purposes of even this ageing thread it's not necessary to know the cell voltage four digits beyond the decimal point ... as you've implied with "0,1mV" (e.g. 3,7453V). Three digits beyond the decimal point is sufficient for ebiker DIY builds when using a quality DMM and Cell Monitor for checking Lithium-ion cell voltages :thumb:

Also, glad to see your reference to Li-ion cell chemistry (e.g. NCA & NMC) in a recent post in your excellent thread. As further explanation the reason it's NCA and not NAC is because the recipe called for more cobalt than aluminum. Also the acronym ICR emphasized the importance of cobalt in the recipe which is no longer true. In part because of what some believe is the growing scarcity of cobalt and Nickel's importance. The relevance of Oxide to a Li-ion cell is either expressed as O2 or 04. Apparently 3 letters is sufficient ... so no need for LNMCO :wink:

The reason for N at front is because of its importance in the chemical recipe of NMCs 811 cathode (80% Nickel, 10% Manganese, 10% Cobalt). More recent hybrid recipes include 433 and 532 with 622 being the a more recent NMC recipe. There is still some discussion whether the acronym should be NMC or NCM depending on which ingredient they believe is more important and that depends in part on the application. NCA cathodes typically have 80% Nickel and 15% Cobalt, but are ‘doped’ with aluminium as opposed to manganese.
Blog-NMC_compositions-e1519742044936.png
Blog-NMC_compositions-e1519742044936.png (8.75 KiB) Viewed 2361 times

john61ct   100 GW

100 GW
Posts: 6242
Joined: Dec 18 2018 2:06pm

Re: Li-ion cells cycle ageing

Post by john61ct » Feb 23 2020 2:10pm

commas are canon for the decimal place in much of the world

and how much accuracy you desire is a personal choice, not much point arguing how much is "enough"

User avatar
flippy   1 MW

1 MW
Posts: 1708
Joined: Aug 12 2015 3:07pm

Re: Li-ion cells cycle ageing

Post by flippy » Feb 24 2020 7:38am

eMark wrote:
Feb 23 2020 1:44pm
You do realize that 0,1mV is actually 0,0001V. It isn't necessary to have a DMM that can read 4 places beyond the decimal point to know "what's truly going on".
you actually do in a lot of cases. if you do lots of measurements that require lots of range you will see lots of limitations with like a cheapo 6000 count meter. i personally have a quite expensive 500000 count meter. so i can have up to mV accuracy with batteries up to 50V and 10mV accuracy up to 500V.
Lithium beats liquid dinosaurs.

docware   1 kW

1 kW
Posts: 318
Joined: Sep 16 2018 5:36am

Re: Li-ion cells cycle ageing

Post by docware » Feb 24 2020 9:15am

It is simple. If my paralel groups are within 3 mV, resolution 0,1 mV is must for reliable and repeatable measurement. Requirement of ten times better resolution than measured values is valid not only for measurement of electrical values, but for measurement of all physical units generally.

Good news – you don´t need expensive DMM.

Today´s checking against 5 ½ 0,012 % precision DMM GW Instek revealed that this cheap chinese voltmeter is unbelievably accurate. It´s within 0,1 - 0,3 mV through the whole area of interest 2,7 – 4,2 V depending on the time this cheap voltmeter is running. Longer running time (one hour and more) means higher difference 0,3 mV. However, this cheap unit is nearly on par with 5 ½ DMM ! At least in the range 2,7 – 4,2 V.

eMark   100 W

100 W
Posts: 278
Joined: Nov 02 2019 11:53am

Re: Li-ion cells cycle ageing

Post by eMark » Feb 24 2020 12:10pm

docware wrote:
Feb 24 2020 9:15am
It is simple. If my paralel groups are within 3 mV, resolution 0,1 mV is must for reliable and repeatable measurement.
Realistically, doesn't that "If" become more of a challenge the more P-groups (e.g. 10S4P vs 14S8P) even with top brand name grade A cells as the battery ages.
docware wrote:
Feb 24 2020 9:15am
Requirement of ten times better resolution than measured values is valid not only for measurement of electrical values, but for measurement of all physical units generally.
Not disagreeing with the need for precise accuracy when testing cells whether by the manufacturer for their datasheet or by you for presenting in this thread. But here's my rub (if a rub) so would appreciate your take (Yes, No, Maybe with brief reasoning) as it possibly deserves further discussion in another thread. So, here goes FWIW ...

It's my understanding that even supposedly good enough BMSs (say $20-$30) only discharge balance when the P-groups variance is over 30mV. In other words most all BMS (maybe even smart bluetooth BMSs) don't begin discharge balancing until there is at least a 30mV disparity between at least one of the P-groups and the other P-groups.

If this is true then would you agree that it's a good idea on every DIY build to attach two monitoring-balance leads for at least cell monitoring and discharge balancing. A 14S8P would require two 7S balance leads with one going to each 7S8P. Two leads because there's no reasonably priced cell monitor/discharge balancing feature beyond 8S, (no current duo balance charger beyond 6S). Not only for monitoring P-group cell voltage, but also for discharge balancing and balance charging (if 10S and 12S) when necessary. Especially, if your goal is to maintain no greater variance than say 3mV, 5mV or even 10mV as the battery ages instead of waiting until the BMS begins discharge balancing at a variance greater than 30mV. By then it may be too late. Even then any discharge rebalancing via a BMS is only as good as 30mV P-group disparity ... if in fact that is true with majority of BMSs.

Realize this topic is for another thread, but would at least appreciate whether you think it's a good idea or not necessary to attach two monitor-balance leads (e.g. 2-5S, 2-6S, 2-7S) on a DIY build. According to your above post is it even possible to maintain such close tolerances (3mV) between P-groups on a DIY pack (either before or after charging) without the use of two balance leads for P-group monitoring an discharge balancing as needed? Hopefully keeping the P-groups within 3mV, 5mV or even 10mV as the battery ages as long as possible :thumb:

docware   1 kW

1 kW
Posts: 318
Joined: Sep 16 2018 5:36am

Re: Li-ion cells cycle ageing

Post by docware » Feb 24 2020 1:04pm

eMark wrote:
Feb 24 2020 12:10pm
Realistically, doesn't that "If" become more of a challenge the more P-groups (e.g. 10S4P vs 14S8P) even with top brand name grade A cells as the battery ages.
I again ask administrators to ban this trolling member acces to this thread. eMark is responsible for weed infestation also in other threads, not only this one.
Thank you

eMark   100 W

100 W
Posts: 278
Joined: Nov 02 2019 11:53am

Re: Li-ion cells cycle ageing

Post by eMark » Feb 26 2020 2:04am

eMark wrote:
Feb 24 2020 12:10pm
Realistically, doesn't that "If" become more of a challenge the more P-groups (e.g. 10S4P vs 14S8P) even with top brand name grade A cells as the battery ages.
docware wrote:
Feb 24 2020 1:04pm
I again ask administrators to ban this trolling member acces to this thread. eMark is responsible for weed infestation also in other threads, not only this one.
Thank you
Is it not a legitimate ageing question? Don't see why you think it's "weed infestation."

Maybe, it's a dumb question as the answer is probably Yes with the same kind of cells, same application, etc. Possibly something to consider when deciding to build an 8P or 6P for maintaining P-groups within 3mV of each other as long as possible for battery health. Then thinking about it some more thought maybe just the opposite with the advantage going to 8P for longevity (Tesla battery has 96 P-groups). So, maybe the number of P-groups in a DIY battery makes no difference when it comes to whether or not more P-groups can shorten battery life (if that were the only variable) ??

Just want to say i appreciate your quest for precision in this ageing thread. It got me to making comparisons between three DMMs, Capacitor Controller, Tenergy 5-in-1 Cell Meter, two different digital balance chargers, two LVA cell logs and iDST BattGO 8S. Anyway will not post anymore on your thread. Keep up the good work :thumb:

docware   1 kW

1 kW
Posts: 318
Joined: Sep 16 2018 5:36am

Re: Li-ion cells cycle ageing

Post by docware » Feb 27 2020 4:30am

New data : Samsung 30Q 536 cycles, LG MJ1 500 cycles, SONY VTC6 267 cycles, Samsung 50E 216 cycles, Samsung 35E 182 cycles, LG HG2 127 cycles.
Wrong DCIR data LG HG2 at 50 cycles corrected (originally wrongly measured with filter on).
Samsung 30Q   536 cycles.jpg
Samsung 30Q 536 cycles.jpg (200.02 KiB) Viewed 2080 times
LG MJ1   500 cycles.jpg
LG MJ1 500 cycles.jpg (199.09 KiB) Viewed 2080 times
SONY VTC6   267 cycles.jpg
SONY VTC6 267 cycles.jpg (195.71 KiB) Viewed 2080 times
Samsung 50E   216 cycles.jpg
Samsung 50E 216 cycles.jpg (215.1 KiB) Viewed 2080 times
Samsung 35E   182 cycles.jpg
Samsung 35E 182 cycles.jpg (194.49 KiB) Viewed 2080 times
LG HG2   127 cycles.jpg
LG HG2 127 cycles.jpg (193.16 KiB) Viewed 2080 times
Capacity decay comparison.jpg
Capacity decay comparison.jpg (179.89 KiB) Viewed 2080 times
Capacity decay comparison  zoom.jpg
Capacity decay comparison zoom.jpg (191.99 KiB) Viewed 2080 times
DCIR comparison.jpg
DCIR comparison.jpg (156.73 KiB) Viewed 2080 times
Adding new picture of M36 DCIR with corrected curve ( originally wrong data on 95 % SOC on the page 4 - typo error)
LG M36 DCIR b.jpg
LG M36 DCIR b.jpg (171.98 KiB) Viewed 2080 times

User avatar
serious_sam   1 kW

1 kW
Posts: 426
Joined: Mar 05 2017 8:07am
Location: Australia

Re: Li-ion cells cycle ageing

Post by serious_sam » Feb 27 2020 4:46am

Hi docware, firstly, thanks for this data you generously share with us.

Sorry if I missed it, but did you mention what SOC you perform the IR measurement (for the bar graph @50 cycles) ? Thanks again.

docware   1 kW

1 kW
Posts: 318
Joined: Sep 16 2018 5:36am

Re: Li-ion cells cycle ageing

Post by docware » Feb 27 2020 5:01am

DCIR of cycled cells is measured at approx 3,75 V, which is around 55 % SOC for most cells. Original intention was 50 % SOC, but I missed data I have now.
SOC% versus voltage  14.1.2020.jpg
SOC% versus voltage 14.1.2020.jpg (390.79 KiB) Viewed 2067 times

User avatar
serious_sam   1 kW

1 kW
Posts: 426
Joined: Mar 05 2017 8:07am
Location: Australia

Re: Li-ion cells cycle ageing

Post by serious_sam » Feb 27 2020 5:04am

Understood. Thanks.

john61ct   100 GW

100 GW
Posts: 6242
Joined: Dec 18 2018 2:06pm

Re: Li-ion cells cycle ageing

Post by john61ct » Feb 27 2020 6:54am

eMark wrote:
docware wrote:
Feb 24 2020 9:15am
It is simple. If my paralel groups are within 3 mV, resolution 0,1 mV is must for reliable and repeatable measurement.
Realistically, doesn't that "If" become more of a challenge the more P-groups (e.g. 10S4P vs 14S8P) even with top brand name grade A cells as the battery ages
Problem is your questions are often too wordy and just not comprehensible. Besides being needlessly specific and unecessary, argumentative to the point of constituting harassment.

e.g. you've already been told it is pointless if not counterproductive to argue against another member's desire for accuracy in their instrumentation, hence the OP's expression of extreme frustration.

Please try to take in this feedback as friendly and constructive, which is my intention here.

_____
In this case I can't tell if you mean "more groups" which would mean, 10S goes to say 14S

or "10S4P vs 14S8P" which is just doubling the # of cells in each group.

The ability to keep the cells well balanced (in this case delta < 3mV) does not change in either case.


Last edited by john61ct on Feb 27 2020 6:54am, edited 1 time in total.

john61ct   100 GW

100 GW
Posts: 6242
Joined: Dec 18 2018 2:06pm

Re: Li-ion cells cycle ageing

Post by john61ct » Feb 27 2020 7:02am


docware wrote:Image
Hadn't noticed before, how much the low end of the 30Q data makes obvious there's a very different chemistry there.



goatman   10 MW

10 MW
Posts: 2086
Joined: Jun 23 2019 6:50pm
Location: Surrey, B.C.

Re: Li-ion cells cycle ageing

Post by goatman » Feb 27 2020 8:31pm

the above picture of 30q at about 525 cycles that docware posted. is that about 1900mah? from 4.1v to 3.3v

okashira did a cycle test of 30q and after about 540 cycles at 7 amp discharge it had about 2100mah but his test was 4.15v to 2,8v I think

viewtopic.php?f=14&t=68556&start=100#p1148592

do you have the graph of the discharge line for 1 cycle? im just curious if the capacity is starting to drop sharply like a hockey stick or if its still discharging on a straight line.

docware   1 kW

1 kW
Posts: 318
Joined: Sep 16 2018 5:36am

Re: Li-ion cells cycle ageing

Post by docware » Feb 28 2020 1:51am

Here is Samsung 30Q No 3 condition before the cycling :

viewtopic.php?f=14&t=103092&p=1513064#p1513064

Cycle No 1 had 2180 mAh, average at cycle 535 is 1880 mAh.

leffex   1 kW

1 kW
Posts: 304
Joined: Apr 08 2012 4:39pm
Location: Sweden

Re: Li-ion cells cycle ageing

Post by leffex » Mar 04 2020 9:50pm

I also noticed the Samsung 30Q had some dfifference. However the capacity graph on the side states 100% to 80% and is very similar to Panasonic, Sanyo ga, Sony vtc6 and possibly samsung 35e in the future. I have and like alot of cells like Samsung 29E, Panasonic PF. They seem to come in three versions a b and c which say that they are high performance, medium and long endurance cell.

Cells that reach me are

Chinese cells(found out they kind of die straight after 400-800 cycles), Panasonic PF(yah, thanks for all the info)

However one cell I have alot of. That comes in many iterations are Sanyo Red cells. They have many. I get alot of these called Sanyo 18650AA cells with a 4,3A power at 2150mah if I recon correctly.

Another cool cell I'd like to know more about is Sony 18650V3 2150mah 10A capable older type or model of cell if you compare to the newer medium power cells and can be spotted by the lower voltage at load throughout the test in a graph showing in essence less power or area in the graph than newer cells. The are also reflects the total capacity of the cell, this can be fun to know.

Also the first cell I came into first contact with was Panasonic CGR18650 2250mah. Used ones. Pretty weak cell if you ask me(mine were about 1500mah) and I just noticed they commit suicide charged under zero C* conditions ifs cell is higher than 65 m-ohm.(note: https://secondlifestorage.com/showthread.php?tid=1732)
Salestech in an ebike store

Jan-Erik-86   100 W

100 W
Posts: 100
Joined: Jan 23 2019 4:57am

Re: Li-ion cells cycle ageing

Post by Jan-Erik-86 » Mar 05 2020 5:22am

leffex wrote:
Mar 04 2020 9:50pm
.
Pretty weak cell if you ask me(mine were about 1500mah) and I just noticed they commit suicide charged under zero C* conditions ifs cell is higher than 65 m-ohm.
Some studies have shows you can charge below freezing at an extremely slow charge rate, but in practice any liion cell charged normally below 0c will die instantly, regardless of the dcir. This is because charging below freezing causes plating inside the cell.

I have several of those old CGR18650 as well and i noticed they often have a lot of capacity left, but only at a slow discharge rate.

docware   1 kW

1 kW
Posts: 318
Joined: Sep 16 2018 5:36am

Re: Li-ion cells cycle ageing

Post by docware » Mar 05 2020 11:21am

New batch of data : Samsung 30Q 585 cycles, LG MJ1 550 cycles, SONY VTC6 312 cycles, Samsung 50E 258 cycles, Samsung 35E 227 cycles, LG HG2 181 cycles.
Samsung 30Q   585 cycles.jpg
Samsung 30Q 585 cycles.jpg (200.68 KiB) Viewed 1745 times
LG MJ1   550 cycles.jpg
LG MJ1 550 cycles.jpg (199.83 KiB) Viewed 1745 times
SONY VTC6   312 cycles.jpg
SONY VTC6 312 cycles.jpg (196.13 KiB) Viewed 1745 times
Samsung 50E   258 cycles.jpg
Samsung 50E 258 cycles.jpg (215.8 KiB) Viewed 1745 times
Samsung 35E   227 cycles.jpg
Samsung 35E 227 cycles.jpg (195.06 KiB) Viewed 1745 times
LG HG2   181 cycles.jpg
LG HG2 181 cycles.jpg (193.53 KiB) Viewed 1745 times
Capacity decay comparison  zoom 5.3.2020.jpg
Capacity decay comparison zoom 5.3.2020.jpg (193.91 KiB) Viewed 1745 times
DCIR comparison 5.3.2020.jpg
DCIR comparison 5.3.2020.jpg (157.7 KiB) Viewed 1745 times
Because Samsung 50E got affected by the tester, I have made some trial to estimate how much. Result is aprox 1,7 %, so I made another chart with 50E offset 1,7 % from 150 cycle point in effort to estimate what normal course of this cell woud be like.
Capacity decay comparison  zoom offset 5.3.2020.jpg
Capacity decay comparison zoom offset 5.3.2020.jpg (195.2 KiB) Viewed 1745 times
Pajda, could you please comment on my attempt ? Do your 50E tests have similar trend of the capacity ?

Post Reply