Cycle life tests of High Energy density cylindrical cells

[Dunlop]

Hi Pajda,
I am out to build a 18S5P or 20S5P pack for a hot rod e-bike build, using the N.E.S.E. 21700 modules.... I was looking around to find the latest and greatest Samsung 50S cells (which as advertised were upgraded 50E's). Please forgive me, but I can't figure out what happened to the 50S cells you were testing.... Bottom line is what is the the best cell available at this particular time in history??
Thanks,
Keith

 

[silenec]

Hi Pajda,
I am out to build a 18S5P or 20S5P pack for a hot rod e-bike build, using the N.E.S.E. 21700 modules.... I was looking around to find the latest and greatest Samsung 50S cells (which as advertised were upgraded 50E's). Please forgive me, but I can't figure out what happened to the 50S cells you were testing.... Bottom line is what is the the best cell available at this particular time in history??
Thanks,
Keith

I think 50S got moved to other thread called "Cycle life tests of High Power density cylindrical cells".

It "died after 952 cycles in the 1C test" which is maybe not that surprising because it is rated as:

Capacity ≥ 2,940mAh @ after 250cycles (60% of the standard capacity)
- Charge : 6A, 4.20V, CCCV 100mA cut-off
- Discharge: 25A , 2.5V cut-off

Capacity ≥ 2,940mAh @ after 100cycles (60% of the standard capacity)
- Charge : 6A, 4.20V, CCCV 100mA cut-off
- Discharge: 45A, 2.5V cut-off
(80℃ cut-off, re-discharge release < 50℃)

Personally I went with the Samsung 50G because I don't need that many amps and price was ok.

50G:
Capacity ≥ 3,880mAh @ after 1,000cycles (80% of the Standard Discharge Capacity @ RT)
- Charge : 0.33C(1,600mA), 4.2V, CCCV 0.05C(242.5mA) cut-off @RT
- Discharge: 1C(4,850mA), 2.5V cut-off @ RT

 

[Pajda]

I apologize for the late replies and update.

In November it looked like my test room wouldn't be heated to 23°C from the new year due to the EU gas price crisis. This would have significantly affected the tests, so I have not started another one since then. Fortunately, the situation stabilized and the heating was maintained so I was able to start the planned tests of another HE cells (and some new HP):

BAK N21700CD-53
Samsung INR21700-50G2

Back to your questions:

Samsung INR21700-48X was also prepared for testing, so I will add it to the chart soon. But we probadly did not see the significant difference against 5000mAh competitors in first 1000 cycles.

Samsung INR21700-50S is a cell I have a problem with. All already tested HP cells (new Molicel P45B included) are rated as 8C continuous with discharge end temperature below 80°C. Unfortunately, 50S can't handle this. Yes, its power capability is definitely significantly better than HE cells, but a typical HE application (traction) will not utilize such performance and will rather suffer from the poorer cycle life associated with the HP cell design. And in a typical HP application this cell will suffer from frequent overheating in comparison with current HP standard cells.

I would attribute the sudden death of the 50S sample after 950 cycles to bad luck rather than being standard. I've seen this with other cells over the years, namely the HG2.

 

[silenec]

Samsung INR21700-48X was also prepared for testing, so I will add it to the chart soon. But we probadly did not see the significant difference against 5000mAh competitors in first 1000 cycles.

BTW I was deciding between 48X and 50G and I found 2 datasheets for the 48X. The newer one (marked as 05 Nov. 2019 V2) said:

Cycle life1
Capacity ≥ 70% @ after 500cycles (70% of the Initial Capacity @ RT)
- Charge : 1.0C(4,800mA), 4.2V, CCCV 0.05C(240mA) cut-off @ RT
- Discharge: 1.0C(4,800mA), 2.5V cut-off @ RT

Cycle life2
Capacity ≥ 70% @ after 1,500cycles (70% of the Initial Capacity @ RT)
- Charge : 1.0C(4,800mA), 4.15V, CCCV 0.05C(240mA) cut-off @ RT
- Discharge: 1.0C(4,800mA), 3.0V cut-off @ RT

While the older(?) one says:

Cycle life
Capacity ≥ 70% @ after 3,000cycles (70% of the Initial Capacity @ RT)
- Charge : 1.0C(4,800mA), 4.15V, CCCV 0.05C(240mA) cut-off @ RT
- Discharge: 1.0C(4,800mA), 3.0V cut-off @ RT

So the 50G 80% @ 1000 cycles 4.2V->2.5V cycle life actually looked better to me on paper.

 

[Pajda]

So the 50G 80% @ 1000 cycles 4.2V->2.5V cycle life actually looked better to me on paper.

My opinion is that the cell manufacturers in the Product Specification deliberately put only the most necessary information, probably forced to do so by some institution. They are probably comfortable giving cycle life values with a huge margin due to potential warranty claims by customers. It's quite a pity, because this thead show that almost every modern cell can handle 1000 cycles even at high 90% or even 100% DoD.

 

[Pajda]

small update after while

LG M58T finished 1000 cycles in both 100% and 90% DoD tests with impressive results.
BAK N21700CD-53 on the other hand show significant problems with its cycle life. I expected much better results here, so I will definitely try to test this model again on other samples in the future.

Samsung 50G(1) also finished 1000 cycles in both 100% and 90% DoD tests with outstanding results. Actually very similar to the 50E2 model and most other competitors in 5000 mAh category.
Samsung 48X(1) finally added to the test based on your wishes.

 

[Mitsoskart]

Very useful thread, thanks for your work

 

[Kajman]

Are 48x, 50g, 50LT and 58T tests completed and available to see and compare somewhere? Anyone seen LG 58 without "T" datasheet?

 

[Pajda]

Sorry for the long pause

48x is still running 4.15-3V test, after 850cycles it is at about 94% of nominal capacity. 50G and 50G2, 50LT and 58T are finished. Samples of EVE INR18650-35V and Tenpower INR21700-50ME should be soon added.

Probably the most interesting news this year are the INR18650-F38 and N40 cells from Vapcell. I don't know yet who is the original cell manufacturer. According to the tests available so far, the cells actually have a corresponding capacity of 3800mAh and 4000mAh. I have ordered samples and will hopefully be able to add them to the test during the summer.

 

[Kajman]

Will you update the table with 58T and 50g?

 

[Pajda]

Will you update the table with 58T and 50g?

Yes, but it will take some time. Btw. Vapcell N40 samples arrived today, first results from nominal capacity test tomorrow.

 

[Pajda]

Damn, they made it!

There is no marking at all on the naked body of the Vapcell N40 cell.

The 35V mass is 46.9g, dimensions with tube: D: 18.30mm h: 65.00mm
The N40 mass is 47.9g, dimensions with tube: D: 18.31mm h: 65.00mm
The F38 mass is 47.0g, dimensions with tube: D: 18.30mm h: 65.00mm

More tests to follow.

 

[NiOOH]

First results look very promising. What's the cell under the wrap? Do Vapcell use their own?

 

[Pajda]

updated data for Vapcell N40

 

[NiOOH]

updated data for Vapcell N40

One thing makes an impression. The discharge capacity drops very significantly at 1 C discharge rate, compared to 0.2 C. I don't think I've seen such a big drop on any other LiIon cell. It goes from 4000 mAh to about 3200 at 1 C. That's also strange, given the not so high IR. It behaves almost like a NiMH cell, strange. Obviously, it is not a high current cell, but it should be holding better, at least up to 1 C.

 

[riba2233]

One thing makes an impression. The discharge capacity drops very significantly at 1 C discharge rate, compared to 0.2 C. I don't think I've seen such a big drop on any other LiIon cell. It goes from 4000 mAh to about 3200 at 1 C. That's also strange, given the not so high IR. It behaves almost like a NiMH cell, strange. Obviously, it is not a high current cell, but it should be holding better, at least up to 1 C.

Look at the last picture in the post, it is a different situation if you discharge to 2.5V

 

[Stealth_Chopper]

One thing makes an impression. The discharge capacity drops very significantly at 1 C discharge rate, compared to 0.2 C. I don't think I've seen such a big drop on any other LiIon cell. It goes from 4000 mAh to about 3200 at 1 C. That's also strange, given the not so high IR. It behaves almost like a NiMH cell, strange. Obviously, it is not a high current cell, but it should be holding better, at least up to 1 C.

Were there unusual chemical emissions during this event ?

 

[Pajda]

As for the Vapcell N40, on Friday I started a standard cycle life test for two DoD ranges: 100% (4.20-2.50V) and ~95% (4.15-3.00V).

I asked Vapcell who the manufacturer of the N40 cell is and was told that the manufacturing company did not wish to be disclosed at this time.

So it forced Pajda to do his homework and after about 10 seconds of googling he came across the company FEB (Far East Battery Co.,Ltd.) which on its website presents 4000mAh cells in 18650 format, but also 6000mAh in 21700 and possibly 30Ah in 4680. I have no idea if this is the original manufacturer of the Vapcell cells, but I assume that the 6000mAh version in the 21700 is coming soon.

On their site: FEB if you click on the DOWNLOAD button, you will get their nice Roadmap on page 9.

 

[Pajda]

Attached a teaser of what the cycle life of N40 looks like after the first 50 cycles (6 posts above)

 

[Pajda]

Samples of Vapcell INR18650 F38 (3800 mAh) have just arrived . By the end of the week I hope to have the first test results.

 

[Pajda]

Added first data of F38 results in the table (the post above with N40 results), more will follow.

 

[Pajda]

The data for F38 are now updated, standard cycle life tests started yesterday.

 

[riba2233]

they look pretty good so far!

 

[Kajman]

Pajda, you have probably some mistake in M58 test capacity by 1C discharge in your table, I always get results of 5500+mAh with 8A. And 50G 1C test is underrated too, maybe some cell clamp problem?

 

[Pajda]

Pajda, you have probably some mistake in M58 test capacity by 1C discharge in your table, I always get results of 5500+mAh with 8A. And 50G 1C test is underrated too, maybe some cell clamp problem?

ES Table "test capacity" results at 1,2 and 3C are measured at ~90% DoD (4.15-3.00V range), but "nominal capacity" are 0.2C at 100% DoD (4.20-2.50V). I guess this could be the cause of the discrepancy? And yes, I admit it's a little bit confusing.