Li-ion cells cycle ageing
- Thread starter docware
- Start date
Probably everybody here know negative impact of low temperatures on capacity and internal resistance of Li-ion cells, apart from aging during operation.
As I noticed some potential effect of temperature on DCIR during measurements, was wonder to know more. Not having temperature test chamber I made some provisory measurement to get some data on two cells, M36 and 29E7.
It´s obvious that mere 1 °C temperature difference has notable effect on resulting DCIR. These data are probably explaining sometimes little bit incoherent DCIR values of M36 and 29E7 during cycling testing where I use ambient air temperature measurement.
However the cell real temperature may be little bit different for various reasons, therefore for DCIR measurement is better to place thermocouple directly on the cell surface as I usually did for DCIR mapping.
As I noticed some potential effect of temperature on DCIR during measurements, was wonder to know more. Not having temperature test chamber I made some provisory measurement to get some data on two cells, M36 and 29E7.
It´s obvious that mere 1 °C temperature difference has notable effect on resulting DCIR. These data are probably explaining sometimes little bit incoherent DCIR values of M36 and 29E7 during cycling testing where I use ambient air temperature measurement.
However the cell real temperature may be little bit different for various reasons, therefore for DCIR measurement is better to place thermocouple directly on the cell surface as I usually did for DCIR mapping.
Yes too bad even the big cells they don't build in a temp sensor wire internally.
You can see why well-funded tests use a liquid bath.
Internally generated heat would really throw high-C rate results way off!
I guess do a test, wait an hour, do the next…
You can see why well-funded tests use a liquid bath.
Internally generated heat would really throw high-C rate results way off!
I guess do a test, wait an hour, do the next…
I think if a cell costs hundreds of dollars, at least giving an internal tube into the interior if nit an actual sensor would not be cost prohibitive.
Even just a couple of locations, one likely to hit overtemp first, another undertemp, would be useful in dynamically changing use cases.
Even just a couple of locations, one likely to hit overtemp first, another undertemp, would be useful in dynamically changing use cases.
flippy
1 MW
- Joined
- Aug 12, 2015
- Messages
- 2,351
but a cell is less then a cup of coffee at the local McD, so thats not going to happen.
bigger cells are not cyclinderical and flat cells can very simply be equipped with a sensor.
your protection setup and use case should be such that any overtemperature behaviour is mitigated as such to prevent core temperatures to exeed the stated limits. usually this is done by giving the cell time to soak the heat to the outser cell.
this is also why you keep the temp logging going AFTER the discharge (or charge) test has completed so see the heat soaking process and you can make modifications to your planned battery setup to ensure you dont exeed any limits by figuring out at what temperautures and current limits the cell can hit said temperature limits.
i would never build a battery that runs that close to its limits unless its for something like drag racing and lifespan/saftey is secondary to performance.
bigger cells are not cyclinderical and flat cells can very simply be equipped with a sensor.
your protection setup and use case should be such that any overtemperature behaviour is mitigated as such to prevent core temperatures to exeed the stated limits. usually this is done by giving the cell time to soak the heat to the outser cell.
this is also why you keep the temp logging going AFTER the discharge (or charge) test has completed so see the heat soaking process and you can make modifications to your planned battery setup to ensure you dont exeed any limits by figuring out at what temperautures and current limits the cell can hit said temperature limits.
i would never build a battery that runs that close to its limits unless its for something like drag racing and lifespan/saftey is secondary to performance.
Of course I did not mean cells under say 60Ah.
And I have no expectation this will come to pass, but it might! no one can predict the future, mass market batteries already on the market with internal heating capability thermostatically controlled, so why not a sensor?
Just saying, it would be very useful to "more instantly" see the impact of various C-rates, peak or continuous on internal temperatures, without having to wait for the delta to dissipate until it is detectable from the outside.
And I have no expectation this will come to pass, but it might! no one can predict the future, mass market batteries already on the market with internal heating capability thermostatically controlled, so why not a sensor?
Just saying, it would be very useful to "more instantly" see the impact of various C-rates, peak or continuous on internal temperatures, without having to wait for the delta to dissipate until it is detectable from the outside.
spinningmagnets
100 TW
docware, thank you for posting this info. The heat that is generated by a given discharge is an interest of mine. I have the money, but I simply don't have the time to post the tests that I want to do. The world (and ES) is a better place because of data like this.
Wow. I knew colder temps raised the IR , but I had no idea how drastically each deg. affected it. Very good info . Does continued use at low temps cause permanent accelerated degredation? I mean, you are effectively operating the cells at an increasingly higher C rating with increasingly higher resistance as temps drop, correct?
flippy
1 MW
- Joined
- Aug 12, 2015
- Messages
- 2,351
Dak77 said:
tesla heats up the battery to nearly 50 in order to enable ludicrous mode/track mode.
otherwise it wont get up to the amps needed (nearly 1500A) without causing massive sag and uncontrolled heating of the cells.
it also does this if you drive to a supercharger it will try to heat the battery so it can accept the 150+kW of charge it can get.
so yes, a hot battery is a happy battery. to a certain point. unless you have spent millions in battery temperature managment i would just stick to ambient temps. going over 50 without thermal control is not a good idea.
As flippy wrote, contrary to storage mode, in operation mode higher temperature above 30 °C
are better for Li-ion cells. Of course, lower temperatures bellow 15 °C mean worse operation conditions. All because temperature has significant impact on the chemical conversions, diffusion of the ions through the electrolyte, intercalation and deintercalation of the ions in the electrodes, …..
Want to remind that at our comparative cycling test are cells on friendly temperatures. In the real life is situation different. Not to speak about temperature differences between the cells at the battery wall and internal cells or temperature gradient through the cells itself.
are better for Li-ion cells. Of course, lower temperatures bellow 15 °C mean worse operation conditions. All because temperature has significant impact on the chemical conversions, diffusion of the ions through the electrolyte, intercalation and deintercalation of the ions in the electrodes, …..
Want to remind that at our comparative cycling test are cells on friendly temperatures. In the real life is situation different. Not to speak about temperature differences between the cells at the battery wall and internal cells or temperature gradient through the cells itself.
flippy
1 MW
- Joined
- Aug 12, 2015
- Messages
- 2,351
here is a view on capacity from a samsung 29E, the cell had a easy load so not much internal heating so it takes ambient temps changes quite well and thus bigger changes on capacity.
the difference was about 5c between heating and aircon. you see the wear as a constant but the big changes are from changes in ambient during the course of a day. the cell remained basically at ambient at all times.
same time but much higher load, and lots of internal heating so the cell is much more even as the cell is constant ridden hard to keep the temperature up.
the difference was about 5c between heating and aircon. you see the wear as a constant but the big changes are from changes in ambient during the course of a day. the cell remained basically at ambient at all times.
same time but much higher load, and lots of internal heating so the cell is much more even as the cell is constant ridden hard to keep the temperature up.
Thats all very interesting thanks for sharing docware :thumb:
I have a small question, as LG seems to have so much better cycle life, couldn't it be because of the higher LVC?
LG was 4,1 - 3,4V, and all other down to 3,3V.
If find it quite surprising that if you use high current cells at very low discharge currents that cylce life is still worse as on high capacity cells.
I always had the meaning that they will live longer when used far below the specs but it seems not.
Does your equipment allow for a 5-10A cylce test, with something like 2A charge current? That would be nice to have and also a cylce test between 4,15V and 3,2V or so as comparison.
I have a small question, as LG seems to have so much better cycle life, couldn't it be because of the higher LVC?
LG was 4,1 - 3,4V, and all other down to 3,3V.
If find it quite surprising that if you use high current cells at very low discharge currents that cylce life is still worse as on high capacity cells.
I always had the meaning that they will live longer when used far below the specs but it seems not.
Does your equipment allow for a 5-10A cylce test, with something like 2A charge current? That would be nice to have and also a cylce test between 4,15V and 3,2V or so as comparison.
Well, I again recommend to read all Pajda´s post for understanding the excelent result of LG M36 as also high power cells poor results.
LVC 3,4 V …... LG M36, Sanyo GA, LG MJ1, Samsung 50E, Samsung 35E
LVC 3,3 V ....... Panasonic PF, Samsung 29E, Samsung 30Q , SONY VTC6
Max. charging current is 5A, max. discharging current 10A.
Currently we are running tests at these particular conditions. Maybe in the future we can run test with higher loads. It is possible that after reading Pajda´s post no higher load testing will be required.
LVC 3,4 V …... LG M36, Sanyo GA, LG MJ1, Samsung 50E, Samsung 35E
LVC 3,3 V ....... Panasonic PF, Samsung 29E, Samsung 30Q , SONY VTC6
Max. charging current is 5A, max. discharging current 10A.
Currently we are running tests at these particular conditions. Maybe in the future we can run test with higher loads. It is possible that after reading Pajda´s post no higher load testing will be required.
Pajda
10 kW
Such small difference does not have significant impact on cycle life in SoC window below 70%.madin88 said:
madin88 said:
The principle of this issue was already formally described in this topic. So yes HP cells have significantly worse cycle life than most HE cells under low loads up to ca 1-2C continuous discharge. Generally(Although I do not like generalizations) HP cells cycle life is almost independent of the load. That means that HP cells should be used only in applications which really needs HP cells. From my point of view there are three fundamental categories of the cells on the market if we take a cycle life into account.
HE cells (up to 1C continuous) M36, MJ1, GA, 35E
"BP" (balanced parameters) cells (up to 3C continuous) 29E, M29
HP cells (more than 3C continuous) 30Q, HG2, VTC6
Pedrodemio
100 W
docware, thanks a lot for undertaking this endeavor, doing a whole lot of cell cycling tests is something I always wanted to do, but unfortunately I have no conditions at the moment or in the foreseeable future
One suggestion for the future of the tests
I’ve mainly build electrical skateboards, and a guy named Mathias Recalde Koller did a study where he gathered and bunch of logs from all over the world and built a standard riding cycle, very much like what is used to test consumption for bigger vehicles, I’m almost certain that someone must have done it for ebikes. If a discharge and charge curve is derived from a cycle like that, the cycling would be way more in line with real life use. It comes with a new set of challenges since the loading would vary with battery size, so a mean pack configuration would have to be determined
Maybe this doesn’t really change the results, probably Padja and others can share some knowledge if it’s a worth pursuit, also I don’t know if the tester you use would allow for a highly complex discharge curve
Find his paper on the link bellow
https://drive.google.com/file/d/1GMKLe-Vx0Fy3uRytyUZrYPWHZMp57IzD/view?usp=drivesdk
One suggestion for the future of the tests
I’ve mainly build electrical skateboards, and a guy named Mathias Recalde Koller did a study where he gathered and bunch of logs from all over the world and built a standard riding cycle, very much like what is used to test consumption for bigger vehicles, I’m almost certain that someone must have done it for ebikes. If a discharge and charge curve is derived from a cycle like that, the cycling would be way more in line with real life use. It comes with a new set of challenges since the loading would vary with battery size, so a mean pack configuration would have to be determined
Maybe this doesn’t really change the results, probably Padja and others can share some knowledge if it’s a worth pursuit, also I don’t know if the tester you use would allow for a highly complex discharge curve
Find his paper on the link bellow
https://drive.google.com/file/d/1GMKLe-Vx0Fy3uRytyUZrYPWHZMp57IzD/view?usp=drivesdk
Low C-rates were the norm, lifespan and **energy** density the goals.madin88 said:
Then the market demanded higher **power** density, and the chemists have been tweaking formulations ever since.
Turns out there is an inherent conflict between the two, if you want huge C-rates capability and lighter weight you sacrifice Ah capacity and longevity.
So, only buy the power density you actually need if value for money is important to you.
Similar threads
