Panasonic/Sanyo NCR20700B vs Samsung INR21700-50E
- Thread starter Jil
- Start date
Pajda
10 kW
I think I owe some actual informations about cycle life tests of Samsung 21700 cells. At first I must repeat that I can not show you graphs but I can tell you numbers. 
At this time I already reched milestone of 1000 cycles with INR21700-50E under several conditions. I also have 1000 cycles with INR21700-40T under one particular test 0.5C-1C at 100% DoD and another test conditions are already running.
I personally use dividing cells into three categories:
HE (High Energy) - LG MJ1,M36,M50, Samsung 35E,50E, Sony VC7, Sanyo GA...
"BP" (Balanced Parameters) - LG HG2, Samsung 30Q, 40T, Sony VTC6...
HD/HP (High Drain/Power) - Samsung 30T, 25S...
I am not focusing on HD cells, so I have only limited database/knownledge of their cycle life test. But I have tested all above mentioned cells and many others from HE and BP categories in small cylinder format. Here are some general notes
1) Absolute best cycle life is domain of HE cells!!! but the continous discharge rate must be lower than 1C and charge rate 0.5C. I personally recomend to design traction battery with HE cells up to 0.5C average discharge and up to 0.3C average charge rate. HE cells at higher continuous discharge rates than 1C are going hot as hell and their cycle life falls rapidly to ca 150 cycles at 3C continuous discharge, also 1C continuous charge rate usually have significant negative impact on their cycle life.
2) Almost all tested BP cells have significantly worse cycle life than HE cells. But their cycle life is usually almost the same under 1C or even 3C continuous discharge rate. Also their cycle life in not affected if they are charged at 0.5C or 1C continuous rate. Also BP cells are much cooler at higher continuous charge/discharge rates.
Some data for Samsung cells:
50E: holds 83% of its nominal capacity after 1000 cycles at 0.5C-1C 100% DoD
40T: holds 63% of its nominal capacity after 1000 cycles at 0.5C-1C 100% DoD
40T: holds 77% of its nominal capacity after 350 cycles at 0.5C-3C 100% DoD (it is the same number as for 0.5C-1C at 350cycles)
At this time I already reched milestone of 1000 cycles with INR21700-50E under several conditions. I also have 1000 cycles with INR21700-40T under one particular test 0.5C-1C at 100% DoD and another test conditions are already running.I personally use dividing cells into three categories:
HE (High Energy) - LG MJ1,M36,M50, Samsung 35E,50E, Sony VC7, Sanyo GA...
"BP" (Balanced Parameters) - LG HG2, Samsung 30Q, 40T, Sony VTC6...
HD/HP (High Drain/Power) - Samsung 30T, 25S...
I am not focusing on HD cells, so I have only limited database/knownledge of their cycle life test. But I have tested all above mentioned cells and many others from HE and BP categories in small cylinder format. Here are some general notes
1) Absolute best cycle life is domain of HE cells!!! but the continous discharge rate must be lower than 1C and charge rate 0.5C. I personally recomend to design traction battery with HE cells up to 0.5C average discharge and up to 0.3C average charge rate. HE cells at higher continuous discharge rates than 1C are going hot as hell and their cycle life falls rapidly to ca 150 cycles at 3C continuous discharge, also 1C continuous charge rate usually have significant negative impact on their cycle life.
2) Almost all tested BP cells have significantly worse cycle life than HE cells. But their cycle life is usually almost the same under 1C or even 3C continuous discharge rate. Also their cycle life in not affected if they are charged at 0.5C or 1C continuous rate. Also BP cells are much cooler at higher continuous charge/discharge rates.
Some data for Samsung cells:
50E: holds 83% of its nominal capacity after 1000 cycles at 0.5C-1C 100% DoD
40T: holds 63% of its nominal capacity after 1000 cycles at 0.5C-1C 100% DoD
40T: holds 77% of its nominal capacity after 350 cycles at 0.5C-3C 100% DoD (it is the same number as for 0.5C-1C at 350cycles)
Fantastic stuff.
Personally, I do not do 100% DoD cycling, cutoff usually well above 3V even at high C rates, and stop-charge well before 95% SoC.
I believe that use pattern should get at least 50% more cycles, likely even double?
And then C rates, 0.2-.3 for charging, anything over 0.6C discharge would be momentary only.
What would you recommend (anyone can respond of course) for top **value per $** for this use case, longevity being top priority?
Looking at 20-30Ah groups, so bigger capacity cells would be ideal.
Personally, I do not do 100% DoD cycling, cutoff usually well above 3V even at high C rates, and stop-charge well before 95% SoC.
I believe that use pattern should get at least 50% more cycles, likely even double?
And then C rates, 0.2-.3 for charging, anything over 0.6C discharge would be momentary only.
What would you recommend (anyone can respond of course) for top **value per $** for this use case, longevity being top priority?
Looking at 20-30Ah groups, so bigger capacity cells would be ideal.
Pajda said:I think I owe some actual informations about cycle life tests of Samsung 21700 cells. At first I must repeat that I can not show you graphs but I can tell you numbers.
I personally use dividing cells into three categories:
HE (High Energy) - LG MJ1,M36,M50, Samsung 35E,50E, Sony VC7, Sanyo GA...
"BP" (Balanced Parameters) - LG HG2, Samsung 30Q, 40T, Sony VTC6...
HD/HP (High Drain/Power) - Samsung 30T, 25S...
I am not focusing on HD cells, so I have only limited database/knownledge of their cycle life test. But I have tested all above mentioned cells and many others from HE and BP categories in small cylinder format. Here are some general notes
1) Absolute best cycle life is domain of HE cells!!! but the continous discharge rate must be lower than 1C and charge rate 0.5C. I personally recomend to design traction battery with HE cells up to 0.5C average discharge and up to 0.3C average charge rate. HE cells at higher continuous discharge rates than 1C are going hot as hell and their cycle life falls rapidly to ca 150 cycles at 3C continuous discharge, also 1C continuous charge rate usually have significant negative impact on their cycle life.
2) Almost all tested BP cells have significantly worse cycle life than HE cells. But their cycle life is usually almost the same under 1C or even 3C continuous discharge rate. Also their cycle life in not affected if they are charged at 0.5C or 1C continuous rate. Also BP cells are much cooler at higher continuous charge/discharge rates.
Some data for Samsung cells:
50E: holds 83% of its nominal capacity after 1000 cycles at 0.5C-1C 100% DoD
40T: holds 63% of its nominal capacity after 1000 cycles at 0.5C-1C 100% DoD
40T: holds 77% of its nominal capacity after 350 cycles at 0.5C-3C 100% DoD (it is the same number as for 0.5C-1C at 350cycles)
Pajda
10 kW
flippy said:
Actually from the data I have already collected from 21700 testing I can confirm both issues.
LG M50(with bottom vent, without marking on shrink tube) seems to have the best cycle life in 21700 format under low load test (0.5C-1C 100%DoD), but it is unusable at higher continuous discharge rates (at 3C continuous is the cell dead after 50 cycles). The similar problem is with 1C continuous charge rate.
Samsung 50E is slightly worse at low load test, but it performs significantly better at higher charge/discharge rates. It was a surprise for me that 50E is practically usable even with 1C continuous fast charging. Also it performs relativelly well even at 3C continouos discharge rate.
Unfortunatelly, I have encountered a problem with samples of newer LG M50T cells(without bottom vent, with standard LG marking) from one German distributor. Because those samples perform as total scam even at low load test. I must reorder them from another distributor and run the tests again.
However, at low loads all above mentioned cells are outperformed by 18650 LG MJ1 and M36 in cycle life tests.
___
If I compare actual prices then:
LG M50 is on pair with MJ1 per Wh, but M36 is slightly cheaper.
Samsung 50E is ca 15% more expensive than MJ1 per Wh
Very valuable stuff.
For me, dollars **per Ah capacity** is required for comparisons.
Fewer cells per 100Ah unit is important.
Dollars per Ah capacity **per year** in daily cycling is really the bottom line metric
assuming replacement EoL at 70-75% SoH.
Ideally at 60-70% DoD not 100%, and low C-rates
For me, dollars **per Ah capacity** is required for comparisons.
Fewer cells per 100Ah unit is important.
Dollars per Ah capacity **per year** in daily cycling is really the bottom line metric
assuming replacement EoL at 70-75% SoH.
Ideally at 60-70% DoD not 100%, and low C-rates
flippy
1 MW
- Joined
- Aug 12, 2015
- Messages
- 2,351
the Ah is a meaningless value in general. it is dependant on voltage. a 100Ah 3.2v battery is something completly different then a 100Ah 300V battery....
Wh is the metric that is used for calculating capacity. Ah is only useful when comparing batteries with identical voltage like a car battery.
Wh is the metric that is used for calculating capacity. Ah is only useful when comparing batteries with identical voltage like a car battery.
mAh and Ah are more universal for a reason, that's a more accurate measure of battery capacity and energy usage.
Obviously when comparing the above values between two build options the voltages will be comparable.
Converting to W or Wh is required only when voltages are different, say DCDC conversion, or the two sides of an inverter, or completely different pack builds.
The conversion process is so trivially easy of course if you prefer watts suit yourself.
Obviously when comparing the above values between two build options the voltages will be comparable.
Converting to W or Wh is required only when voltages are different, say DCDC conversion, or the two sides of an inverter, or completely different pack builds.
The conversion process is so trivially easy of course if you prefer watts suit yourself.
flippy
1 MW
- Joined
- Aug 12, 2015
- Messages
- 2,351
john61ct said:
Ah differs depending on the load, Wh does not.
it actually has problems even depending on chemistry, some have a 3.5V as their nominal value, most are 3.6 and 3.65 or even 3.2 in case of LFP. that can give quite a difference in Wh if you dont catch that. for that reason Wh is a much more usable metric to compare capacities, even when they have the same voltage, it also allows for simpler math if you use the battery to calculate your range.
Matador
100 kW
- Joined
- Jun 29, 2016
- Messages
- 1,045
flippy said:
I agree that Price per Wh is higher. However I think that to be fair on comparing cycle life of two different formats, you have to compare cells taht have the same C-rating and the same chemistry. For example, compare the Samsung 25R 25A 2500mAh 18650 (10C) to the Samsung 40T 35A 4000mAh (9 C). Moreover, fair comparison needs to cycle both cells at the same C-discharge rate to compare cycle life. For example, dicharge the 25R at 5A (2C) and the 40T at 8A (also 2C)...
You cant really compare a 1C cell to a 10C cell... The 10C cell has a bigger cathode for current, but les litthium for capacity... The 1C cell has a miniscle electrode and more apace for lithium, but if ypu discharge it to fast, it will become very hot!
High discharge cell are generally pricier, because they tend to last much longer. Why? Well, less internall resistance means a cells runs cooler, which become proportionnally more evident as the load increase. Cell that run cooler live longer lives. So low diacharge cells are fine if you never need to upgrade to a higher power setup, but higher discharge cell can do both low and high power.
To me saying that a the 21700 cell are weaker than that in 18650 is as meaningless as saying that the wine is a magnum bottle is not as good as that in a 750 mL bottle... Well, I would enjoy a Cote Rotie in a a 1.5 mL magnum format much more than a cheap supermarket wine in a 750mL format. Quality of the internal is much more relevant.
Matador
Matador
100 kW
- Joined
- Jun 29, 2016
- Messages
- 1,045
The price on 21700 will continue to decline as it becomes more and more popular...
But to compare cycle life of two different cells, you need to discharge with the same ratio (ratio=cycling C-discharge rate choosen / max rated C-discharge rate of a given cell). For example discharge both different cells at 50% of max C-rate of each cell and measure cycle life till you reach 70% of initial capacity... Then we'll see.
That is, compare the cycle life by discharge the Samsung 25R 18650 at 12.5A (50% of max C dschg rate) and the Samsung 40T 21700 at 17.5A (50% of max dschg rate) and the compare wich yields best cycle life. Both cells are 10C so they should be comparable since they are both NMC chemistry.
Matador
But to compare cycle life of two different cells, you need to discharge with the same ratio (ratio=cycling C-discharge rate choosen / max rated C-discharge rate of a given cell). For example discharge both different cells at 50% of max C-rate of each cell and measure cycle life till you reach 70% of initial capacity... Then we'll see.
That is, compare the cycle life by discharge the Samsung 25R 18650 at 12.5A (50% of max C dschg rate) and the Samsung 40T 21700 at 17.5A (50% of max dschg rate) and the compare wich yields best cycle life. Both cells are 10C so they should be comparable since they are both NMC chemistry.
Matador
Hillhater
100 TW
Using manufacturers “C” ratings for any comparason is not very reliable.
I have never seen a “standard” methodology for establishing “C” rates, so the proceedure can vary considerably between companies to suit their specific products
Hobbyking lipo are probably the best know example of exaggerated “C” rates, whilst the cells are technically capable of the discharge rates , it has significant negative impact on other cell properties (life cycle, = cost)
I have never seen a “standard” methodology for establishing “C” rates, so the proceedure can vary considerably between companies to suit their specific products
Hobbyking lipo are probably the best know example of exaggerated “C” rates, whilst the cells are technically capable of the discharge rates , it has significant negative impact on other cell properties (life cycle, = cost)
Similar threads
