Cycle Life tests or experience with Panasonic NCR18650GA vs LG MJ1 vs Samsung INR18650-35E

Joined
Jan 10, 2017
Messages
54
Hi,

Anyone who has done cycle Life Tests of 3500mAH batteries, Panasonic/Sanyo NCR18650GA vs LG MJ1 vs Samsung INR18650-35E for example.

I plan to do a pack, but prefer to use long lasting ones if differences are really important.

Have a Nice Day.

Thierry
 
I plan on setting up a life cycle test on some cells shortly. But the list of cells to test is long and I need more power supplies and Arduino boards. Plus it can take a month to get good data...
 
jonescg said:
I plan on setting up a life cycle test on some cells shortly. But the list of cells to test is long and I need more power supplies and Arduino boards. Plus it can take a month to get good data...

That's normal.

I've tested some LiPo cells (just to know the leakage current), and all the Kokam are very good.

Thierry
 
It is hard to answer even if i have all three cells measured for their cycle life. When you plan to use only 50% DoD average, all cells do the job well. I will personally prefer LG MJ1 for this. But if you are looking for >80% DoD average and you are thinking about more than 500 cycles, the only way is to go with NCA chemistry cells and so Sanyo GA.

Under my standard 0.5C charge - 1C discharge cycle life test at 100% DoD, the Samsung 35E is "dead" after 300 cycles and LG MJ1 after 500 cycles. "Dead" means reaching 70% initial capacity inside the very steep capacity drop. This behavior is typical for all NMC cells in 18650 format which I have measured. But until then they shows only relatively small capacity drop.

Only Sanyo GA is holding its capacity above 70% even after 1000 cycles under 100% DoD. All NCA cells I have measured shows opposite behaviour than NMC cells. They starts with a relatively steep loss of capacity, but there is visible a slow settlement, where around 1000 cycle the capacity loss is almost stopped.
 
Pajda said:
Only Sanyo GA is holding its capacity above 70% even after 1000 cycles under 100% DoD.
Any real data points to demonstrate those claims? Even GA official data fly in the face of those claims.
Thanks!
 
tomjasz said:
Pajda said:
Only Sanyo GA is holding its capacity above 70% even after 1000 cycles under 100% DoD.
Any real data points to demonstrate those claims? Even GA official data fly in the face of those claims.
Thanks!

Yes I've seen the Panasonic datasheet, about 300 cycles @100% DOD.

But Panasonic is often very conservative.

I'd be curious to know if the NCA cells are better in real file tests.
 
I've got NCR18650 Panasonics but they're the older 2700mah ones. Hoping cycle life is great even with it being slowly discharging in storage down to 2.1V...
 
ThierryGTLTS said:
I've just noticed that the cells are NCR, not NCA :!: :!:

Not the same technology.

Thierry

NCR is the "bussines name" used by Panasonic for their cell production line. The base chemistry principle (cathode active material) is known as NCA which is shortcut for – Lithium Nickel Cobalt Aluminium Oxide (LiNiCoAlO2). So NCR chemistry principle does not exist.
 
tomjasz said:
Any real data points to demonstrate those claims? Even GA official data fly in the face of those claims.
Thanks!

Yes it is based on my long term laboatory measurement with calibrated multichannel 4-wire testers. For an idea, the 1000 cycle test at 100% DoD (0.5C charge - 1C discharge) lasts ca 6 months. Unfortunately I can not present the graphs in public at this time, I can only comment the results.
 
I can also say that Sanyo NCR18650GA has very similar cycle life with Sanyo NCR20700B and this two NCA cells have the worst cycle life (under the same 100% DoD test) from the Panasonic 18650/20700 actual production line. Cells like NCR18650BM, Tesla 85kWh or NCR18650PF are much better in this parameter. The best performer under this particular test is "an old" NCR18650B with 77% of initial capacity after 1000 cycles. But as most of you already knows the practical use of this cell is very limited by its high internal resistance. With GA, BM or PF cells you can easily discharge 3C continuous without major heating problems, but this is not possible with the original "B" cell.
 
jonescg said:
I plan on setting up a life cycle test on some cells shortly. But the list of cells to test is long and I need more power supplies and Arduino boards. Plus it can take a month to get good data...

Can you share your setup and arduino code? I would love to build a cycle test setup for the batteries i use. I do have the time and can build a fairly large setup to test many cells.
 
https://endless-sphere.com/forums/viewtopic.php?p=1197080#p1197080

I had the voltage divider in there as I was testing 2s2p batteries, however I have since started using it as a single cell tester by wiring the battery sense wires directly to the analog input. Don't change the mV limits though as it calculates the correct voltage regardless. I set the Vmax to 8550 mV which cuts the charge off at about 4.10 V. The lower limit of 6000 mV ends the discharge at about 3.20 V.

Tweak these values to suit your discharge regime.

The trouble with testing multiple cells is that you need multiple Arduino boards, relays, and frustratingly, multiple constant current power supplies.
 
pG2038P.jpg


f6VK0qM.jpg


13s4p Sanyo GA pack. discharged w/ 6.8ohm load. pack was used on a BBS02 w/ over 4000km

i'm still getting 95% of original capacity
 
Thanks for the link and the arduino code.
I will go over it soon and scrape parts together. Charging it is no problem. I got plenty of cccv power boards and a couple of arduino megas. I might see if i can use a mega's multiple inputs to use more them one cell per arduino and make a more efficient setup.

I also have a 1.5kW capable load for draining large packs. I will see if i can change a few bits and make it capable of automating a large pack cycle test. Some packs i build for customers are 6~15kWh at 72v so i need some automation to test them before i can ship them out the door so i can prove they can get their capacity rating under load. Also hand for incoming warranty claims. Not that i have ever had one so far but doing a comparison test after a few years is very helpful.
 
Pajda said:
It is hard to answer even if i have all three cells measured for their cycle life. When you plan to use only 50% DoD average,...

Why buy a 3500mAh cell if you only use 1750mAh?


the only way is to go with NCA chemistry cells and so Sanyo GA.

Imho the cycle degradations main factor is the silicone in the anode, less the cathode material.

All 3500mAh cells use Si-C anodes, whcihs significantly reduces cycle performance.

(Afaik the newer 21700 high capacity cells do not use Si-C anodes, that's why their energy densiotiy is lower than the enegey density of 3500mAh 18650 cells)

Also charging power has huge Impacts of cycle Performance. Updated data sheets for Samsung 35E say, that for "cycle performance" maximum charge rate should only be 1000mA (I asume this is at 25°C, charging should be much even more problematic at lower temperatures)
 
Pajda said:
NCR is the "bussines name" used by Panasonic for their cell production line. The base chemistry principle (cathode active material) is known as NCA which is shortcut for – Lithium Nickel Cobalt Aluminium Oxide (LiNiCoAlO2). So NCR chemistry principle does not exist.

Yes.

NCR = Panasonic internal shortcut for (n)ickel+(c)obalt based (NCM or NCA or more complex) and (r)echargeable. (or (r)ound ???)

It's a member of the Panasonic NNP battery familiy: (n)ickel oxid based (n)ew (p)latform.
 
Cephalotus said:
Imho the cycle degradations main factor is the silicone in the anode, less the cathode material.

All 3500mAh cells use Si-C anodes, whcihs significantly reduces cycle performance.

(Afaik the newer 21700 high capacity cells do not use Si-C anodes, that's why their energy densiotiy is lower than the enegey density of 3500mAh 18650 cells)

Also charging power has huge Impacts of cycle Performance. Updated data sheets for Samsung 35E say, that for "cycle performance" maximum charge rate should only be 1000mA (I asume this is at 25°C, charging should be much even more problematic at lower temperatures)

Thanks for all those usefull infos.

What kind of cells do you suggest to use to have a good balance between energy density, acceptable fast recharge, and long life :?: :!: :wink:

Thierry
 
Cephalotus said:
Pajda said:
It is hard to answer even if i have all three cells measured for their cycle life. When you plan to use only 50% DoD average,...

Why buy a 3500mAh cell if you only use 1750mAh?

The " 50% DoD average" principle means that most of the time you are using only less than 50% capacity of battery but you have still available 100% of capacity for occasional longer trips. A good example of this principle are Tesla vehicle Batteries. Very few users drive 400 km (90% DoD) in single trip without charging each day. This is one of major reasons why Tesla never released 40 kWh or less capacity battery based on their standard 18650 cells.

Cephalotus said:
Imho the cycle degradations main factor is the silicone in the anode, less the cathode material.

All 3500mAh cells use Si-C anodes, whcihs significantly reduces cycle performance.

(Afaik the newer 21700 high capacity cells do not use Si-C anodes, that's why their energy densiotiy is lower than the enegey density of 3500mAh 18650 cells)

I can only say that INR21700-48G cells from Samsung have the far worst cycle life at 0.5C-1C from all cells I have already tested. (no matter on chemistry)


Cephalotus said:
Also charging power has huge Impacts of cycle Performance. Updated data sheets for Samsung 35E say, that for "cycle performance" maximum charge rate should only be 1000mA (I asume this is at 25°C, charging should be much even more problematic at lower temperatures)
I think you are right in this point. Some of Samsung 18650/21700 cells, which I have already tested, have cycle life issue when they are charged at 0.5C rate. Even the 30Q High Power cell suffer with this issue too. The only two Samsung cells from my tests without significant problem with 0.5C charge rate are 25R and 29E.
 
the PF is my default battery of choice because it is so cheap compaired to it's capacity and 10A discharge rating.
i will try and do testing with the above mentioned setup but with loads more in line with what my batteries usually see. i always stop charging at 4.05V and charge at well under 0.5C. usually around 250~400mA per cell and float at 4.05V and discharge is around 1~2A or less with peaks to 5A per cell at best. that should extend the lifespan considerably.
 
DVDRW said:
What about even cheaper Samsung INR18650-29E 2.14€/qty? As good as PF?
https://i.imgur.com/4iKFhMc.png

I think Tumich is cheaper. My favored battery for <25A.
 
DVDRW said:
What about even cheaper Samsung INR18650-29E 2.14€/qty? As good as PF?
https://i.imgur.com/4iKFhMc.png

the 29E is more expensive right now at nkon then the PF. it has been for a while. the PF also has a higher current rating.
i have 1 large battery with 400 29E's and it is holding up very well even when being (seriously) overcharged once and having to deal with decent peak loads. but i do keep it at 4.05~4.1V at best and only drain it to about 3.6V every day.
 
Back
Top