Which cell for long lasting battery

DrKraut

1 W
Joined
Jun 29, 2013
Messages
64
Location
alemania
Hey guys,

at the moment i have a 7P13S battery consisting of 91 LG MG1 and after only about 170 cycles it lost already about 40%. I tested it multiple times and its quite disappointing. My setup looks like following:

Charge overnight with about 3A or 0,3C.
Commuting to work and back, each way is about 35-40mins, with a about 0,75C or 15A.
Then i recharge again, with about 10% Capacity untouched.
Temperature is between 12°C and 23°C
With this scenario battey is toast after only 170 cycles.

I want a battery that lasts as long as possible, something like 10 years.
What kind of cell do you recommend? Sanyo GA, Panasonic PF (too old?) or Samsung 50E looked promising concerning cycle life.
 
If you can carry about 30%? more weight per Ah

LFP is around 220 Wh/L (790 kJ/L), 120 Wh/kg vs LCO 150-190 ?

At a .75C discharge rate, and ideally stopping at 15% SoC rather than 10%

you should IMO get 4000+ cycles

The larger size prismatics - CALB, GBS, Winston, Sinopoly let you build a larger pack with only one or two strings, 180Ah is considered a medium cell size.

A123 pouches are more difficult to DIY but also great quality.

Note the charge V is 3.5-3.6 rather than 4.1-4.2Vpc.

_____
LTO is the new hotness, over 10000 cycles claimed, but crazy low charge voltage, not sure about energy density.
 
I tought that too at my first design phase for the battery. But I'm limited to the triangle size of my bike. Which just fits about 91 cells plus the controller. So i have to stay high energy density cells, since I can't charge at work.
 
Are your battery cells assembled into plastic cavities, or just connected to eachother with glue?
Do you know, with which materail the cells are welded to eachother? Pure Nickel, or Hilumin?
 
For some reason, the factory specs on cells continue to list 4.2V as the max charge per cell. Although lithium cells will work at 4.2V per cell, their life can be roughly doubled if you only charge to 4.1V per cell..and 4.0V is even better (going down to 3.9V per cell provides no benefit).

It is useful to charge to the full amount once a month or so, to ensure that the BMS will balance all of the cells. However, I continue to be shocked that nobody is making a BMS that will balance the cells at 4.1V...

I guess if I would order 1,000 units and pay cash up front, China will make whatever I want, but...if the majority of your profits come from selling battery packs, why would they sell a pack (and charger) that lasts longer? They WANT your pack to die soon.
 
I just found something interesting while digging around:

Its about the Sammy 35E cell but anyway really suspicious or strange:

From this datasheethttp://downloads.cdn.re-in.de/14000...de-SAMSUNG_LI_ION_AKKU_18650__3500MAH_10A.pdf
Thats the cycle life stats:inr18650-35E cycle life per.png
So you can see its about 80% after 500 cycles, while noting the charging/discharging conditions.

Next datasheet:https://akkuzentrum.de/media/datasheets/Samsung__35E.pdf
Thats what it tells us now:
inr18650-35E cycle life-60p.png
So somehow they reduced the cycle life from 80% to 60% with the same conditions, if I'm not mistaken :confused:

What do you guys say to something like that? :idea:
 
spinningmagnets said:
It is useful to charge to the full amount once a month or so, to ensure that the BMS will balance all of the cells. However, I continue to be shocked that nobody is making a BMS that will balance the cells at 4.1V...

Someone correct me if I am wrong but I believe since em3ev has been using a smart BMS; their packs will balance from as low as 4.0V.
 
spinningmagnets said:
nobody is making a BMS that will balance the cells at 4.1V...
Wut?

I thought there were dozens of BMS that allow user-custom setpoints? Maybe only in LFP land? Only ones costing more? I dunno. . .

Of course, no need to rely on the BMS to balance anyway, plenty of units dedicated to balancing, do not do so at piddly sub - amp rates either.

Or use a hobby charger.

I see balancing as a regular maintenance thing, depending on the use case once a year can be more frequent than necessary.

Of course monitoring things more frequently is required, especially as the bank ages.

 
DrKraut said:
I just found something interesting while digging around:

Its about the Sammy 35E cell but anyway really suspicious or strange:

From this datasheethttp://downloads.cdn.re-in.de/14000...de-SAMSUNG_LI_ION_AKKU_18650__3500MAH_10A.pdf
Thats the cycle life stats:View attachment 1
So you can see its about 80% after 500 cycles, while noting the charging/discharging conditions.

Next datasheet:https://akkuzentrum.de/media/datasheets/Samsung__35E.pdf
Thats what it tells us now:

So somehow they reduced the cycle life from 80% to 60% with the same conditions, if I'm not mistaken :confused:

What do you guys say to something like that? :idea:
Not same termination profile.

Each factory run differs.

Each lab test run will differ.

Point is no on should be using this data to set their own usage patterns in real life.

Don't charge that fast, Do not use that high a V setpoint. Do not hold CV that long. Do not discharge that low.
 
Been running Headway 40152S 15Ah LiFePO4 for 6 1/2 years. Over 800 cycles of daily use. They are at 80% of original capacity but sag really bad now. 10 years is probably out of reach for today's 18650's. LTO might do it for you, but it hasn't been used enough around here enough to know for sure. It is also big and heavy. But all the other numbers on it look great.

:D :bolt:
 
Its been a while since I researched the termination voltage where the BMS performs the balancing. I know cell-balancing can be done at 4.0V-4.1V with RC chargers on a pack with no BMS.
 
I've looked at spec sheets for hundreds of BMS, and have yet to see a fixed-settings one where the "start-balance V" wasn't higher than where I'd want my bank to sit for the length of time it would take for the too-puny mA "balance-current rate" to do much of anything.
 
DrKraut said:
Hey guys,

at the moment i have a 7P13S battery consisting of 91 LG MG1 and after only about 170 cycles it lost already about 40%. I tested it multiple times and its quite disappointing.

A) Can you please describe in details how did you measure battery capacity at the beginning and after 170 cycles ?

B) Where did you buy the cells ?
 
DrKraut said:
Hey guys,

at the moment i have a 7P13S battery consisting of 91 LG MG1 and after only about 170 cycles it lost already about 40%. I tested it multiple times and its quite disappointing. My setup looks like following:

Charge overnight with about 3A or 0,3C.
Commuting to work and back, each way is about 35-40mins, with a about 0,75C or 15A.
Then i recharge again, with about 10% Capacity untouched.
Temperature is between 12°C and 23°C
With this scenario battey is toast after only 170 cycles.

I want a battery that lasts as long as possible, something like 10 years.
What kind of cell do you recommend? Sanyo GA, Panasonic PF (too old?) or Samsung 50E looked promising concerning cycle life.

Hi,

Thanks for detailed destription of your experience with LG MG1 based battery. I can just say that the MG1 cells performs in your application exactly as I expected and it confirms my laboratory cycle life tests of this particular LG chem cell. LG MG1 is my personal favourite for the most shitty cell ever made by LG chem. :( It was some time ago replaced by the LG M29 cells which I hope will perform much better as It is expected from the LG chem production line.

Under my standard test procedure 0.5C charge 1C discharge, 100% DoD the results for MG1 are:
100 cycle: 90% of nominal capacity
200 cycle: 82% of nominal capacity
300 cycle: 70% of nominal capacity
few cycles after 300 and the cell is dead.

For brief comparison:
all LG MH1, MJ1 and M36 are after 1000!!! cycles under the same conditions still at ca 86% of their nominal capacity.
 
john61ct said:
Not same termination profile.

Each factory run differs.

Each lab test run will differ.

Point is no on should be using this data to set their own usage patterns in real life.

Don't charge that fast, Do not use that high a V setpoint. Do not hold CV that long. Do not discharge that low.

What do you mean with termination profile?
I think you should be able to trust a manufacturers datasheet, right?
 
docware said:

A) Can you please describe in details how did you measure battery capacity at the beginning and after 170 cycles ?

B) Where did you buy the cells ?

A rough measure always was that I could ride to work, which are 52km total and another 15km until it was empty. As time passed I could barely go to work, so I had to reduce the speed on my way home. And it's getting worse and worse.
First measurements took place with a resistive load and a hobby grade coulomb meter. Same now. So the inaccuracy might be high, but they are the same all the time.

Cycles are pretty easy. The bike is almost purely used for commuting. Which are 52km, my odometer is just shy of 9000km. After every day I recharge the bike overnight. That gives me ~170 cycles, which are not even 100%DoD cycles!
 
Pajda said:

Hi,

Thanks for detailed destription of your experience with LG MG1 based battery. I can just say that the MG1 cells performs in your application exactly as I expected and it confirms my laboratory cycle life tests of this particular LG chem cell. LG MG1 is my personal favourite for the most shitty cell ever made by LG chem. :( It was some time ago replaced by the LG M29 cells which I hope will perform much better as It is expected from the LG chem production line.

Under my standard test procedure 0.5C charge 1C discharge, 100% DoD the results for MG1 are:
100 cycle: 90% of nominal capacity
200 cycle: 82% of nominal capacity
300 cycle: 70% of nominal capacity
few cycles after 300 and the cell is dead.

For brief comparison:
all LG MH1, MJ1 and M36 are after 1000!!! cycles under the same conditions still at ca 86% of their nominal capacity.
Thanks for the info! Can you specify your charge and discharge method, like Voltage, waiting time and so on.
What cell would you recommend to me, it doesn't need to be super high energy. I'm only concerned about longevity under relatively mild conditions. I think about the Sanyo GA or Sam. 35E, or is the Panasonic PF still any good?
 
e-beach said:
Been running Headway 40152S 15Ah LiFePO4 for 6 1/2 years. Over 800 cycles of daily use. They are at 80% of original capacity but sag really bad now. 10 years is probably out of reach for today's 18650's. LTO might do it for you, but it hasn't been used enough around here enough to know for sure. It is also big and heavy. But all the other numbers on it look great.

:D :bolt:

Thanks for the input, I still have some headways laying around! But there to heavy for me :mrgreen:
 
DrKraut said:
john61ct said:
Not same termination profile.

Each factory run differs.

Each lab test run will differ.

Point is no on should be using this data to set their own usage patterns in real life.

Don't charge that fast, Do not use that high a V setpoint. Do not hold CV that long. Do not discharge that low.

What do you mean with termination profile?

The formula for when charging is stopped.

C-rate is a critical factor, that changes the voltage setpoint if doing "charge to" CC-only.

If adding CV time, holding absorb, then the tapering amps setpoint determines how close to theoretical maximum capacity you push to.

Which is harmful if used in normal daily cycling.

> I think you should be able to trust a manufacturers datasheet, right?

Those numbers are "do not allow your bank to get near", not what you should actually use if you value longevity.

 
Pajda said:
DrKraut said:
Hey guys,

at the moment i have a 7P13S battery consisting of 91 LG MG1 and after only about 170 cycles it lost already about 40%. I tested it multiple times and its quite disappointing. My setup looks like following:

Charge overnight with about 3A or 0,3C.
Commuting to work and back, each way is about 35-40mins, with a about 0,75C or 15A.
Then i recharge again, with about 10% Capacity untouched.
Temperature is between 12°C and 23°C
With this scenario battey is toast after only 170 cycles.

I want a battery that lasts as long as possible, something like 10 years.
What kind of cell do you recommend? Sanyo GA, Panasonic PF (too old?) or Samsung 50E looked promising concerning cycle life.

Hi,

Thanks for detailed destription of your experience with LG MG1 based battery. I can just say that the MG1 cells performs in your application exactly as I expected and it confirms my laboratory cycle life tests of this particular LG chem cell. LG MG1 is my personal favourite for the most shitty cell ever made by LG chem. :( It was some time ago replaced by the LG M29 cells which I hope will perform much better as It is expected from the LG chem production line.

Under my standard test procedure 0.5C charge 1C discharge, 100% DoD the results for MG1 are:
100 cycle: 90% of nominal capacity
200 cycle: 82% of nominal capacity
300 cycle: 70% of nominal capacity
few cycles after 300 and the cell is dead.

For brief comparison:
all LG MH1, MJ1 and M36 are after 1000!!! cycles under the same conditions still at ca 86% of their nominal capacity.

Is HG2 in that 1000cycle club?
 
Well, again, where did you buy the cells ?

Let´s assume the cells are from trustfull source:

Quick 40% loss after 170 cycles indicates that something in the conditions is not relatively mild. Despite of probably not optimal cells like MG1 (I trust to Pajda´s measurement).

Something is wrong about handling with the battery. So discovery what is it is of the upmost importance.

Cells sitting on 4,2 V during night is not OK, you can use timer to have cells charged just 15 – 20 minutes before your departure.

Operating in the temperature about 12 °C means increased stress to cells. You probably were running the battery during this winter. That means much lower temperature then 12 °C.
By the way, what is ambient temperature during charging ?

What is charging current, 3 A or 0,3 C (= 6A) ?
What is battery voltage after charging ?

26 km in 35 – 40 minutes means speed 39 – 45 km/h. Average current is probably somewhere between 9 – 12 A, not 15 A. From your description I understand that you are using the whole range 4,2 - 3 V.
You probably pedal, not using throttle only. Is it flat terrain ? What is total weight (bike, rider, load).
What motor do you have ? What is max current of the control unit ? How many months is battery in the usage ?
Measuring the battery capacity at the start means that you are not novice.
We can assume that your measurement have some inaccuracy, just don´t know how large.
 
It seems to me like you didn't really test each cell, did you?
Are you sure that there is not one weak paralleled group?

Could be that you just got one or a few bad cells and that the rest is fine, so the BMS shuts down/indicates a low SOC because of these while the rest is still fine.
 
Dak77 said:
Is HG2 in that 1000cycle club?

Sure it is. But HG2 also has its own issues. The problem with HG2 is that it loses capacity very quickly during the first 50 cycles (ca 8-10% of nominal capacity). Between 50-100 cycles the capacity drop start with stabilization and after ca 150 cycle there is only minor capacity loss (86% of nominal capacity at 1C discharge 100% DoD after 1000 cycles). The good thing is that HG2 behave practically the same even under 3C continuous discharge (82% capacity drop at 3C discharge 100% DoD after 1000 cycles). For example LG MH1, MJ1 and M36 dies very quickly at 3C discharge torture.

So this is the reason why many users prefers Samsung 30Q or Sony VTC6 before LG HG2. Because for flashlight, powertools and many other applications the 50-100 cycles is the whole life. I am charging my cordless drill no more than 10 times per year. So it is useless that LG HG2 in cycle life totally outperforms both above mentioned cells after ca 400 cycles, when particularly 30Q dies.
 
DrKraut said:
Thanks for the info! Can you specify your charge and discharge method, like Voltage, waiting time and so on.
What cell would you recommend to me, it doesn't need to be super high energy. I'm only concerned about longevity under relatively mild conditions. I think about the Sanyo GA or Sam. 35E, or is the Panasonic PF still any good?

Sure, I am using for cycle life test calibrated battery testers with sufficient accuracy. Cells are tested in cell holders with true 4-wire connection. The temperature during measurement is 23°C +/- 2°C. I am designing my own test procedures. Each 50th cycle is inserted nominal capacity and DCIR 10s test.

The most common test have those steps:
0.5C charge (4.2V @ 100mA cut-off) - 10 min rest time - 1C discharge (LVC by manufacturer) - 10 min rest time / 100% DoD

You should choose from the equivalent category of cells:

In high-end category: LG MJ1, M36, Sony VC7, Samsung 35E and Sanyo GA I would recommend both LGs and then Samsung 35E (35E is good choice for the first 500 cycles when the fast charging is not needed = 0.3C charge rate is maximum for 35E otherwise it dies very soon). If you insist on longer cycle life (500 full cycles should be enough for most applications) or faster charging like 0.5C then take only LGs. Sanyo GA suffers with steep initial capacity loss and the most important problem is its huge DCIR rise. VC7 is very expensive and inferior to both LGs.

in best price category: Samsung 29E, Panasonic PF, LG M29, Sony NC1 I would recommend 29E as the best chioce. Panasonic PF is slightly worse under almost all cycle life test. PF is only better in fast charging (like 0.7-1C). I do not already have data for M29 and NC1 because I started with their cycle life tests only few days ago.
 
Back
Top