LG M36 vs Samsung 36G capacity test - 3600mAh or just marketing?

Original article: https://www.thunderheartreviews.com/2018/09/lg-m36-vs-samsung-36g-capacity-test.html

The first 3600mAh Li-ion battery in 18650 format appeared 4-5 years ago. It was Panasonic NCR18650G. But it disappeared from the market very quickly and since then there was no 3600mAh cell you could buy (except fakes on eBay). Now two of top manufacturers, LG and Samsung, offer their cells, which are being sold as 3600mAh ones - LG INR18650 M36T and Samsung INR18650-36G. Both cells were bought from Queen Battery.

The video version of this test for those who prefer watching rather than reading:


As always the tests were done using ZKETECH EBC-A20, which supports up to 20A discharge, 4-wire measurement and is PC-connected.


I've upgraded my battery holder to version 3.0


This time i've used thicker (0.5mm) pure copper strips (9mm and 11.5mm wide). The parts which contact with the cell are rised a bit to provide reliable connection when a cell with deep placed contact(s) is held.


I've followed all the prescriptions of the IEC61960-2003 standard concerning battery's capacity measurement. Before each discharging cycle each battery was charged at standard current mentioned in its datasheet to charge end voltage (4.2V) (cut-off at 0.1A, which is the lowest supported by EBC-A20). Before each discharging or charging i've held a 1-1.5hrs pause. The environment temperature was 23.5-24.5°C. To be sure in results i've done each test minimum twice.

When i was testing these cells there was no datasheet available for either of them so i decided to use the following values for testing both guys:
Standard charge current: 1.80A
Charge end voltage: 4.2V
Charge cut-off current: 100mA
Max discharge current: 10A
Discharge cut-off voltage: 2.5V


LG INR18650 M36T
My cell has the following marking: LGDBM361865 Q248I053A1. According to BatteryBro Q248 means that the production date is Sep 2017.


Specifications from LG M36 datasheet (pdf):
Nominal energy: 12.50Wh
Minimum energy: 12.10Wh
Nominal voltage: 3.63V
Standard charge current: 1000mA (0.3C)
Charge end voltage: 4.2V
Charge cut-off current: 50mA
Max. charge current: 1000mA (0.3C) at 0-24°C, 2330mA (0.7C) at 24-50°C
Max. discharge current: 5000mA
Discharge cut-off voltage: 2.5V
AC impedance at 1KHz: ≤30mΩ
Maximum weight: 48.2 ± 1.0g

I remind that the datasheet i found (thank Pajda) after testing and the specs above are just for information.

Pay attention that there is no capacity mentioned in the datasheet, but the 0.2C current is mentioned as 670mA, so it's not hard to find the C (capacity): 670/0.2=3350mAh. It corresponds to the "minimum capacity" i think, because 3.35x3.63=12.16Wh. The nominal capacity would be 12.50/3.63=3.444Ah ≈3450mAh.

Also pay attention to the Max discharge current in the datasheet - it's 5A, not 10A as the suppliers tell. Though i'm not sure it's not a mistake.

The measured weight of my M36 cell is 48.40g


Capacity test results:

At 0.72A which is higher than 0.2C mentioned in the datasheet (0.67A) M36 showed 3471mAh and 12.56Wh - more than the nominal energy declared. All the curves, including at 8A and 10A, look excellent and that's why i think that the max discharge current is not 5A as mentioned in the datasheet.


Samsung INR18650-36G
The cell which i've tested has the following marking: INR18650-36G SAMSUNG SDI 2H24. H24 means that the cell was manufactured in Feb 2017.


The measured weight of my cell is 47.97g


Samsung 36G capacity test results:

At 0.72A capacity is below 3400mAh and at 10A the curve shows a deep voltage sag which is not a good sign. I would not recommend discharging 36G at current higher than 8A.


COMPARISON
Let's start with 2A discharge:

M36 is the winner. The difference is about 30mAh in capacity and 0.3Wh in energy.

At 5A discharge:

Picture didn't change - M36 is the better one but the difference is already about 120mAh / 0.75Wh.

At 8A discharge:

At 8A the gap reached 190mAh and 1Wh...

At 10A discharge:

M36 outperformed 36G by 200mAh and 1.1Wh but 36G doesn't look confident at 10A.


CONCLUSION
None of two cells can be called a 3600mAh battery... Samsung 36G doesn't even reach the 3400mAh point... LG M36 is comparable with 3500mah cells which i've tested but doesn't outperform them all.

Queen Battery sells Samsung 36G for USD 2.90 and LG M36 for USD 3.00 (shipping not included).
UPDATE: Queen Battery doesn't sell Samsung 36G due to its poor performance. It turned out they stopped selling them a month ago.

Check out my YouTube channel for batteries, chargers and other stuff reviews.
I've recently launched my blog where you can find all my reviews in one place. Every new test/review will be first published on YouTube and in the blog. I'll be happy to see new subscribers, comments, suggestions and just your thoughts.

DC IR test results done with EBC-A20 (5 iterations with 15-30s pause @3600mA)

M36:
28.6, 29.7, 31.1, 30.0, 28.9mOhm

36G:
31.4, 31.9, 30.8, 33.3, 30.8mOhm

Nice review mate. Thank you.

rojitor, thank you too, dude!

In fact it was the biggest disappointment for me... Especially the 36G... In M36 datasheet they didn't even mention the nominal capacity directly!

UPDATE: Queen Battery doesn't sell Samsung 36G due to its poor performance. It turned out they stopped selling them a month ago.

As promised I am posting my LG M36 test results under the same testing conditions as thunderheart. Test parameters are described in the picture title below. I can share .dat files from this test.

Thank you guys.
It seems like it isn't possible to squeeze more than 3400-3500mAh into an 18650 size cell. Everything else is marketing or better name it sheer fraud similar to those "4000mAh" 18650 cells from unknown brands you will find on ebay..
I am also afraid that robustness and cycle life of those high mAh cells will be worse as it is with older generations, less mAh cells.
I remember that this was the case with latest generations of NiMh cells 10 years ago (for powering RC models before LiPo was popular). They had a bit more mAh, but lifetime and robustness (regarding abuse) was bad.

Pajda said:

As promised I am posting my LG M36 test results under the same testing conditions as thunderheart.

Click to expand...

thunderheart's graph shows about 0,1V more sag over the entire curve or 4-5% less Wh, so your test bench seems to measure the voltage more accurate.
Do you have a photo of your cell holder or test setup?

yours @ 10A:

11Wh
3,36V @ 1400mAh
3,2V @ 2100mAh

thunderheart's @ 10A:

10,5Wh
3,26V @ 1400mAh
3,1V @ 2100mAh

madin88 said:

thunderheart's graph shows about 0,1V more sag over the entire curve or 4-5% less Wh, so your test bench seems to measure the voltage more accurate.

Click to expand...

I'm always very careful to talk about accuracy when comparing battery measurements even if I always checking my tester calibration before running test with at least 5 1/2 digit precision multimeter. Absolute measured values can be tricky but the trend is usually ok. In this particular case we should focus on the voltage sag. Both measurements at low load (0.72A) shows almost the same sag, but measurements at 10A shows significant difference almost 0.1V in voltage sag. This is definitely caused by the thunderheart misunderstanding of the 4-wire measurement principle.

madin88 said:

Do you have a photo of your cell holder or test setup?

Click to expand...

I can make a photo, but as I mentioned before I am now using for all my test only BF-2A test fixture with carefully aligned cell. Photo of this fixture can be found in Samsung 30T thread or on fasttech/aliexpress.

madin88 said:

It seems like it isn't possible to squeeze more than 3400-3500mAh into an 18650 size cell. Everything else is marketing or better name it sheer fraud similar to those "4000mAh" 18650 cells from unknown brands you will find on ebay..
I am also afraid that robustness and cycle life of those high mAh cells will be worse as it is with older generations, less mAh cells.
I remember that this was the case with latest generations of NiMh cells 10 years ago (for powering RC models before LiPo was popular). They had a bit more mAh, but lifetime and robustness (regarding abuse) was bad.

Click to expand...

Increasing the absolute value of energy density starts to be problematic as early as 2012 when the first 12Wh cell NCR18650B appears on the mass market. Newer high energy cells brings only about 5% increase in absolute energy density but at least they have significantly lower internal resistance.

But back to the LG M36. Despite the fact that the datasheet contains very conservative parameters, so according to my actual measurements the M36 has very similar parameters to MJ1. Especially cycle life under high DoD seems after first 300 cycles to be very similar, where the MJ1 is currently the best high energy cell when we are looking at this particular parameter.

Hi Pajda,

there is a mention of 23°C ambient temperature in your LG M36 test results.

What is ambient temperature during your cycle life testing and what is tolerance of this ambient temperature during whole testing ?

Do you measure also temperature rise of cell during the cycle life testing ?

I have ran LG M36 test in temperature chamber, but for cycle life tests I am using only basement room with stable long term average temperature 22°C (it very slowly changes from 20 to 24°C during the year seasons - I am logging temperature in this room).

I do not measure temperature of cells during cycle life.

OK, thank you.

Pajda said:

I have ran LG M36 test in temperature chamber, but for cycle life tests I am using only basement room with stable long term average temperature 22°C (it very slowly changes from 20 to 24°C during the year seasons - I am logging temperature in this room).

I do not measure temperature of cells during cycle life.

Click to expand...

Hi Pajda,
Do you have any results you can share?
Thanks,

Already finished with M36 samples from Nkon my standard 0.5C-1C 100% DoD 1000 cycle life test. In this particular test M36 performs exatly the same as MJ1. In another tests with low loads up to 0.5C charge and 1C discharge with lower DoD, the results are also almost identical. But the significant difference begins at higher loads. MJ1 performs relatively well even at 1C continuous charge rate where M36 quickly losses the capacity. Also at 3C continuous discharge test, the MJ1 hold its capacity unexpectedly well (MJ1 performs 400 cycles at 100% DoD at 3C continuous discharge - but unfortunately this is useless for battery pack use, because it is going hot as hell during 3C discharge). M36 barely make a half of this.

So M36 probably uses newer chemistry with lower manganese and cobalt than MJ1 and so it is cheaper. The drawback is that it has lost some of its robustness but still it has very good parameters for traction applications. M36 seems to be the cell with the best price performance ratio for highend e-bikes and maybe also it is the last in 18650 format because the 21700 format is coming up very quickly.

Pajda said:

Already finished with M36 samples from Nkon my standard 0.5C-1C 100% DoD 1000 cycle life test. In this particular test M36 performs exatly the same as MJ1. In another tests with low loads up to 0.5C charge and 1C discharge with lower DoD, the results are also almost identical. But the significant difference begins at higher loads. MJ1 performs relatively well even at 1C continuous charge rate where M36 quickly losses the capacity. Also at 3C continuous discharge test, the MJ1 hold its capacity unexpectedly well (MJ1 performs 400 cycles at 100% DoD at 3C continuous discharge - but unfortunately this is useless for battery pack use, because it is going hot as hell during 3C discharge). M36 barely make a half of this.

So M36 probably uses newer chemistry with lower manganese and cobalt than MJ1 and so it is cheaper. The drawback is that it has lost some of its robustness but still it has very good parameters for traction applications. M36 seems to be the cell with the best price performance ratio for highend e-bikes and maybe also it is the last in 18650 format because the 21700 format is coming up very quickly.

Click to expand...

Very useful info - thank you very much for sharing. Do you have a site or location where you have these results. The community would benefit greatly from having these readily available.

Very interesting and surprising results! Thanks a lot, Pajda!

Thank you Pajda for sharing your info. As LG 18650 M36 seems to be interesting cell, I also have made some measurement.

LG M36 specification :
- standard charge 1 A , 50 mA cut off
- standard discharge constant current 0,2 C 670 mA is giving capacity 3350 mAh
- nominal energy 12,5 Wh and minimum energy 12,1 Wh is giving nominal capacity 3,444 mAh and minimum
capacity 3333 mAh at nominal voltage 3,63 V.

So according the LG specification we can expect capacity somewhere about 3400 mAh.

I agree with Pajda that accuracy of Li-ion cell measurement is pretty tricky. However, we should do our best to have usable results.
True four – wire measurement have been mentioned here many times. Keeping stable specified temperature is also important. Very important factor is calibration, especially current calibration.

Nowadays, I am checking and calibrating the real current during whole measurement with help of precise shunt resistor 0,01 Ohm, tolerance 0,1 %, RIEDON RSN – 10.




Voltage at precise shunt is measured by 5-4/5 handheld multimetr Brymen BM869s, resolution 1 µV, accuracy 0,02 % + 2d.

Discharge 2 A checking with Brymen 869s.
View attachment 4


LG M36 cell production date is July 2018, cell is from nkon. I also have measured temperature during discharging.




Cell surface warming at 5 A discharge (24 °C at the end ) indicates why LG specification denotes max 5 A discharge current. Take into account that BF-2A fixture and ambient air have some cooling ability. Warming of the cell in the closed e-bike battery pack would be significantly higher.


View attachment 2


For capacity measurement I used the same conditions as thunderheart and Pajda :
Charge 1,8 A to 4,2 V , 0,1 A cut off, 1 hour rest, discharge to 2,5 V.
Ambient temperature 24 – 25 °C.
Equipment : ZKETECH EBC-A10H, BF-2A fixture.





DCIR measured at 0,68 A/3,4 A according IEC 61960-3 : 2017, par 7.7.3 in 4 consecutive measurements : 30,15 miliohm, 30,15 miliohm, 30,15 miliohm, 28,9 miliohm, average 30 miliohm.

Equipment : ZKETECH EBC-A10H, BF-2A, PowerLog 6s (voltage drop recording), LogView software. Ambient temperature 25 °C, cell voltage 3,82 V.





For comparison, DCIR of other high energy cells at similar conditions : Panasonic 18650B cca 47 miliohm, Sanyo GA cca 35 miliohm, LG MJ1 cca 36,5 miliohm, Samsung 35E cca 29 miliohm.

As always excellent review from docware. I really appreciate your precision when preparing and performing the test procedure.

He also point on important parameter, that LG in their own datasheet specification for M36T did not use capacity but energy for nominal specification and so they did not say that this cell have 3600 mAh as could be expected from M36T marking. If we compare datasheets for MJ1 and M36T we can conclude that MJ1 is actually better in almost all parameters. But for the standard traction use the difference in almost negligible.

I am also interested in the new specific DCIR measurement method from LG. They are now specifying "10s" or "30s" DCIR at 50% SoC. I found this value in another LG cell datasheets. I must say I like it and I started using it some time ago, because it can be programmed in to the ZKEtech tester routine.

edit: just found that Samsung is using this "10s" 50% SoC DCIR method as well so it maybe a new industry standard.

Just small addition. I simulated cell warming in the closed battery pack with help of standard plastic material fixture and expanded polystyrene isolation. The plastic fixture was trimmed to decrease mass by cca 30 - 40%. Temperature was again measured by thermocouple K, wires diameter 0,127 mm, fixed to cell by Kapton tape 0,05 mm thick.





The 5 A discharge test was run again, this time with M36 cell isolated in the expanded polystyrene.





Here is comparison of the cell warm up in BF-2A and in the thermal isolation.






I finished this second testing after 30 minutes because cell wall temperature achieved 63 °C ( 37 °C warm up). Temperatures inside the cell were surely even higher.

And last postcript - simulation of cell warming in the closed battery pack during normal e-bike running. 2 Amp discharge was chosen as an average current per cell representing standard e-bike operation.

All cells were treated at the same conditions:
- charging 1A to 4,1 V, 0,1 A cut off, 1 hour rest. Fixture BF-2A, ambient temperature 24 – 25 °C
- discharging 2A, 60 minutes in the expanded polystyrene isolation











DCIR values of the cells :

Panasonic 18650B cca 47 miliohm
LG M36 cca 30 miliohm
LG MJ1 cca 35 miliohm
Samsung 35 E cca 29 miliohm
Sanyo GA cca 34 miliohm
Samsung 25R cca 19 miliohm
Samsung 30Q cca 19 miliohm

Excellent work again. If I may, I would ask for the same comparison test graph at 5A and please add Samsung 29E to the competition. :wink:

I am now running non-standard 3C continuous discharge rate "torture" cycle life tests and Samsung 29E together with Panasonic PF shows excellent results in this particular discipline. On the other side GA is practically dead after first 50 cycles.

And hopefully also add the DCIR value of 29E.
It would be much appreciated.

You can find some DCIR values including 29E here in the table :

http://budgetlightforum.com/comment/1429409#comment-1429409

Sorry Pajda, I´m not about to do 5A comparison test.

No problem , I now how the results will look like. It will be a good example for community that designing battery pack for more than 1C continuous discharge rate is problematic even with HD/HP cells in 18650 format like 30Q or HG2.