Maximizing the lifespan of 18650 cells

Disciple · Oct 23, 2013

I recently made the jump into lithium ion powered lights using 18650 cells. I knew in advance that these cells have a finite lifespan (OK, what doesn't) but I don't think I gave it as much thought as I should have before purchasing. I only have a few flashlights :faint:

and I intend to keep my cells charged and in the lights as they are of little use to me otherwise. I am learning that Li-ion cells may age faster than I expected under these circumstances. The article How to Prolong Lithium-based Batteries from BatteryUniversity.com gives as a typical example a loss of 20% capacity after one year stored fully charged at 25°C. I do not know how much charge is lost the following year but it sounds like cells stored in this fashion will only have a useful lifespan of 2-3 years, or less if the aging accelerates.

How much difference it would make to store cells charged to perhaps 85-90% capacity versus 100%? (That site only shows 40% and 100%.)

What would the lifespan of Samsung or Sanyo 4.35v cells charged only on a 4.2v charger be? (I know that initial capacity would be derated by this.)

Knight_Light · Oct 23, 2013

My advice to you would be to keep your batteries charged at 100%, and just simply enjoy them for whatever purpose you got them. 18650 batteries are fairly cheap nowadays (I get the Panasonic NCR18650B Rechargeable 3400mAh 3.7V 18650 for $7.50 each, and that is delivered). What I would suggest is that you measure the capacity of your batteries when you 1st purchase them as a reference point, and then periodically check capacity, when it drops by 20% or greater just buy new batteries. There really is no reason to make it more complicated than this. Hope this helps.

Justin Case · Oct 23, 2013

For storage, the usual advice is to keep the cells at about 40% charge to minimize permanent capacity loss. But if you also have a need for run time and perhaps can't charge up the cells otherwise, then 40% may not be the right state of charge for your needs.

I would probably charge the cells to 4.1V instead of the usual 4.2V (still giving you 90% capacity), and store your lights that way. In use, I'd probably recharge the cells when state of charge falls to about 70% (i.e., a resting voltage around 3.9V). A lower float voltage combined with shallow depth of discharge should greatly increase the cycle life of the cells.

poeee · Oct 23, 2013

General rule from the RC world with lipo's is if you aren't using it for 2 or more weeks put it in a storage voltage.

Norm · Oct 23, 2013

poeee said:
General rule from the RC world with lipo's is if you aren't using it for 2 or more weeks put it in a storage voltage.

Flashlights are a completely different case you know when you're going to use your RC models flashlight use is less predictable.

Norm

samgab · Oct 23, 2013

Take a look at the results of these tests carried out by NASA on Panasonic NCR 18650 Li-ion cells.
Some interesting data can be gleaned about how to make them last as long as possible.
For instance the chart on Page 10; the capacity over 200 cycles doesn't change too greatly when the discharge rate is anything from 1C and below, eg C/5 or C/10 or 1C, as long as the charge rate is 1/2 C or less than 1C. When the charge rate was at 1C, the lifespan of the cell really diminished rapidly.
On Page 11, the discharge rate was always 1C (except for the C/10 charge, C/10 discharge baseline), but they changed the charge rate, and again, you can see that the higher the charge rate, the bigger the impact on lifespan.
So the key is to charge at as low a rate as you have time to wait; eg, if you can, leave the cell to charge overnight at a low rate, 1/2 C, 1/5 C, or 1/10 C (for the CC part of the CC/CV charge algorithm of course). The slower you charge the better for the life of the cell.
For LONG TERM storage only, 75% SOC is better than 100% SOC. If it's only for a few days or weeks of storage between usage cycles, it won't make a lot of difference in comparison to the difference made by the way (rate) you charge and discharge the cells.
Temperature also plays a role, as shown on page 12. Normal room temperature is ideal.

Disciple · Oct 24, 2013

Justin Case said:
I would probably charge the cells to 4.1V instead of the usual 4.2V (still giving you 90% capacity), and store your lights that way. In use, I'd probably recharge the cells when state of charge falls to about 70% (i.e., a resting voltage around 3.9V). A lower float voltage combined with shallow depth of discharge should greatly increase the cycle life of the cells.

This is what I am thinking of doing, but I want to know if the 4.1v/90% charge is actually adding significant life; for all I know (or knew) it could be that the aging rate is the same from 50% to 100% charge, since I had only seen figures for 40% and 100%.

samgab said:
Take a look at the results of these tests carried out by NASA on Panasonic NCR 18650 Li-ion cells.

Thanks for the reference, and your summary.

poeee · Oct 24, 2013

Norm said:
Flashlights are a completely different case you know when you're going to use your RC models flashlight use is less predictable.

Norm

Yeah definitely, but I guess it depends on how many cells you have and the way you use them in your lights.

Wiggle · Oct 25, 2013

samgab said:
Temperature also plays a role, as shown on page 12. Normal room temperature is ideal.

I'd be interested to see some temperatures between room temp and 0C. I was under the impression a cooler storage temperature can greatly reduce the gradual loss of capacity while holding a charge. Or perhaps they actually tested the cell while in a cold environment?

hunterblue · Oct 25, 2013

Hi,

Cycling a lithium ion cell at a 1C charge and 1C discharge at room temperature between the voltages of 4.0V and 3.0V increase the number of potential cycles you can get out of the cell, the tradeoff is obviously capacity as you would be getting something on the order of 70% ish of the maximum capacity of the cell, but what you get in return would be atleast a 4X increase in number of cycles before having to recycle the cell. Depending on chemistry and construction of the specific cell that number could be greatly increased.
EV, HEV, BEV and LEV battery manufacturers use this kind of formula when constructing a battery so that they can increase the calendar life of the battery pack so the consumer is not changing out a battery every year but more like every five to ten years. Seven years seems to be the current average lifespan of automotive batteries.

Also, Samgab is correct, if you charge and discharge at less than 1C rates you can also increase the cycle life, tradeoff being time.

Not all Lithium-ion cells work the same way, so for some this type of usage profile works better and for others make less of a difference.

HB.

Gauss163 · Oct 25, 2013

hunterblue said:
Cycling a lithium ion cell at a 1C charge and 1C discharge at room temperature between the voltages of 4.0V and 3.0V increase the number of potential cycles you can get out of the cell, the tradeoff is obviously capacity as you would be getting something on the order of 70% ish of the maximum capacity of the cell, but what you get in return would be atleast a 4X increase in number of cycles before having to recycle the cell.

@HB It is not clear how to evaluate these claims, since there is no mention of the other (dis)charge rates, temp's, etc that you are comparing to. Do you claim significantly better results than the above-linked NASA tests?

samgab · Oct 25, 2013

I've heard of people keeping cells in fridges or freezers based on hearsay and conjecture, spread around on forums and by word of mouth, that it has some beneficial effect on cell longevity; but I haven't seen any empirical evidence or studies to support those theories.
In fact, the opposite appears to be the case with Li-ion cells -- they don't seem to do as well in low temps as in mild temps.
I prefer to base my treatment of cells on the results of conclusive studies and the results that I've observed in my own tests than on ideas opined online without credible references.
Fortunately I personally use the same Panasonic cells that were used in that NASA testing (as well as the "A" 3100 version), so their findings should apply in my case.
Unfortunately that NASA summary PDF doesn't state all of the details of the parameters/protocols they used in their tests; such as with the 25 cycle temperature comparison tests, whether they were stored between cycles for any period of time at the stated temp; but as they haven't given any details, it seems reasonable to conclude that they ran all 25 cycles for each sample at the stated temperature contiguously (charge-discharge-repeat 25x) for the purposes of the test.
So it doesn't factor in how the cells would do if stored long term at any given temperature.
But considering the charted results of that 25 cycle test at 40, 23, 0, and -10 deg.C; we can see a clear degradation in cell life and performance as the temperature decreased, so my conclusion would be that it's more likely that the cells would last better stored at room temperatures than at low temps, although the NASA test didn't specifically test that parameter.

It would be interesting to do some testing, by getting 9 samples of at least 3 matching Panasonic cells, such as NCR18650A or B.
Then label, measure, and record the capacity of all cells (by cycling each at 1/2C).
Then store 3 sets of samples in the fridge (at about 2 deg.C) at 100%, 75%, and 50% SOC respectively.
Store 3 sets of samples in the freezer (at about -18 deg.C) at 100%, 75%, and 50% SOC respectively.
And store 3 sets of samples at normal room temp (about 20 deg.C) at 100%, 75%, and 50% SOC respectively.
Then, at say, monthly or quarterly intervals, run a cycle on each of the cells (at 1/2 C) and return them to their SOC and storage condition, and keep repeating for an extended period of time, such as 5 years.
Such a test would prove once and for all what effect storing at various temperatures and states of charge has on these cells.
Of course, to do the test would require the cost of purchasing at least 27 brand new cells, plus one would need the equipment, space, and time to do all of the testing and recording of results etc... Who has that kind of time? Not me, unfortunately.

Addendum: Here is a summary of another NASA study on Li-ion cells, this time comparing how they are affected by long term storage at various temperatures. Six different storage temperatures (-20, 0, 10, 25, 40 and 55 deg.C). Two cells for each temperature. They took periodic measurement of capacity after each 3 months at 25 and 0 deg.C.
The problem is, although the tests are on Li-ion chemistry cells, they are special prototype aerospace prismatic cells, and not the normal 18650 cylindrical type cells we often use in laptops and flashlights.
So while the results probably (or possibly?) give an indication of the results in other cell types of similar chemistry, it's not conclusive.
But FWIW, they found very little difference in the capacity of the cells when they were stored at low or moderate temps, but when stored at high temps such as 55 deg.C, there was a noted degradation of cell capacity over time (3-12 years). If you're talking about storing for a few weeks or even a month or two, it's probably not enough of a difference to be worth worrying about.

samgab · Oct 27, 2013

Disciple said:
This is what I am thinking of doing, but I want to know if the 4.1v/90% charge is actually adding significant life; for all I know (or knew) it could be that the aging rate is the same from 50% to 100% charge, since I had only seen figures for 40% and 100%.

*I copied this from another thread, to save retyping it all again, but I thought you might be interested, and it pertains to your original question; although it doesn't touch on long term storage, it does get into the effect that the charge termination voltage has on life of the cells:

I just read an interesting article here...
It states that:

"...An interesting thing to notice is that the cycle life goes up at lower voltages, the equation is roughly Ef (Vch)= 2^[10*(4.2-Vch)] where Ef is the enhanced life cycle factor (Ef = 2 would mean that the battery will survive twice as many charge-discharge cycles as Ef = 1), and Vch is the charge voltage..."

Now they haven't given their references for this statement, but if it is accurate, then:

Code:

[TABLE="width: 334"]
[TR]
[TD][SIZE=1]Term. V[/SIZE][/TD]
[TD][SIZE=1]Ef[/SIZE][/TD]
[TD][SIZE=1]@rated 500 cycles[/SIZE][/TD]
[TD][SIZE=1]@ rated 1000 cycles[/SIZE][/TD]
[/TR]
[TR]
[TD="align: right"][SIZE=1]4.10[/SIZE][/TD]
[TD="align: right"][SIZE=1]2.00[/SIZE][/TD]
[TD="align: right"][SIZE=1]1000.00[/SIZE][/TD]
[TD="align: right"][SIZE=1]2000.00[/SIZE][/TD]
[/TR]
[TR]
[TD="align: right"][SIZE=1]4.11[/SIZE][/TD]
[TD="align: right"][SIZE=1]1.87[/SIZE][/TD]
[TD="align: right"][SIZE=1]933.03[/SIZE][/TD]
[TD="align: right"][SIZE=1]1866.07[/SIZE][/TD]
[/TR]
[TR]
[TD="align: right"][SIZE=1]4.12[/SIZE][/TD]
[TD="align: right"][SIZE=1]1.74[/SIZE][/TD]
[TD="align: right"][SIZE=1]870.55[/SIZE][/TD]
[TD="align: right"][SIZE=1]1741.10[/SIZE][/TD]
[/TR]
[TR]
[TD="align: right"][SIZE=1]4.13[/SIZE][/TD]
[TD="align: right"][SIZE=1]1.62[/SIZE][/TD]
[TD="align: right"][SIZE=1]812.25[/SIZE][/TD]
[TD="align: right"][SIZE=1]1624.50[/SIZE][/TD]
[/TR]
[TR]
[TD="align: right"][SIZE=1]4.14[/SIZE][/TD]
[TD="align: right"][SIZE=1]1.52[/SIZE][/TD]
[TD="align: right"][SIZE=1]757.86[/SIZE][/TD]
[TD="align: right"][SIZE=1]1515.72[/SIZE][/TD]
[/TR]
[TR]
[TD="align: right"][SIZE=1]4.15[/SIZE][/TD]
[TD="align: right"][SIZE=1]1.41[/SIZE][/TD]
[TD="align: right"][SIZE=1]707.11[/SIZE][/TD]
[TD="align: right"][SIZE=1]1414.21[/SIZE][/TD]
[/TR]
[TR]
[TD="align: right"][SIZE=1]4.16[/SIZE][/TD]
[TD="align: right"][SIZE=1]1.32[/SIZE][/TD]
[TD="align: right"][SIZE=1]659.75[/SIZE][/TD]
[TD="align: right"][SIZE=1]1319.51[/SIZE][/TD]
[/TR]
[TR]
[TD="align: right"][SIZE=1]4.17[/SIZE][/TD]
[TD="align: right"][SIZE=1]1.23[/SIZE][/TD]
[TD="align: right"][SIZE=1]615.57[/SIZE][/TD]
[TD="align: right"][SIZE=1]1231.14[/SIZE][/TD]
[/TR]
[TR]
[TD="align: right"][SIZE=1]4.18[/SIZE][/TD]
[TD="align: right"][SIZE=1]1.15[/SIZE][/TD]
[TD="align: right"][SIZE=1]574.35[/SIZE][/TD]
[TD="align: right"][SIZE=1]1148.70[/SIZE][/TD]
[/TR]
[TR]
[TD="align: right"][SIZE=1]4.19[/SIZE][/TD]
[TD="align: right"][SIZE=1]1.07[/SIZE][/TD]
[TD="align: right"][SIZE=1]535.89[/SIZE][/TD]
[TD="align: right"][SIZE=1]1071.77[/SIZE][/TD]
[/TR]
[TR]
[TD="align: right"][SIZE=1]4.20[/SIZE][/TD]
[TD="align: right"][SIZE=1]1.00[/SIZE][/TD]
[TD="align: right"][SIZE=1]500.00[/SIZE][/TD]
[TD="align: right"][SIZE=1]1000.00[/SIZE][/TD]
[/TR]
[TR]
[TD="align: right"][SIZE=1]4.21[/SIZE][/TD]
[TD="align: right"][SIZE=1]0.93[/SIZE][/TD]
[TD="align: right"][SIZE=1]466.52[/SIZE][/TD]
[TD="align: right"][SIZE=1]933.03[/SIZE][/TD]
[/TR]
[TR]
[TD="align: right"][SIZE=1]4.22[/SIZE][/TD]
[TD="align: right"][SIZE=1]0.87[/SIZE][/TD]
[TD="align: right"][SIZE=1]435.28[/SIZE][/TD]
[TD="align: right"][SIZE=1]870.55[/SIZE][/TD]
[/TR]
[TR]
[TD="align: right"][SIZE=1]4.23[/SIZE][/TD]
[TD="align: right"][SIZE=1]0.81[/SIZE][/TD]
[TD="align: right"][SIZE=1]406.13[/SIZE][/TD]
[TD="align: right"][SIZE=1]812.25[/SIZE][/TD]
[/TR]
[TR]
[TD="align: right"][SIZE=1]4.24[/SIZE][/TD]
[TD="align: right"][SIZE=1]0.76[/SIZE][/TD]
[TD="align: right"][SIZE=1]378.93[/SIZE][/TD]
[TD="align: right"][SIZE=1]757.86[/SIZE][/TD]
[/TR]
[TR]
[TD="align: right"][SIZE=1]4.25[/SIZE][/TD]
[TD="align: right"][SIZE=1]0.71[/SIZE][/TD]
[TD="align: right"][SIZE=1]353.55[/SIZE][/TD]
[TD="align: right"][SIZE=1]707.11[/SIZE][/TD]
[/TR]
[TR]
[TD="align: right"][SIZE=1]4.26[/SIZE][/TD]
[TD="align: right"][SIZE=1]0.66[/SIZE][/TD]
[TD="align: right"][SIZE=1]329.88[/SIZE][/TD]
[TD="align: right"][SIZE=1]659.75[/SIZE][/TD]
[/TR]
[TR]
[TD="align: right"][SIZE=1]4.27[/SIZE][/TD]
[TD="align: right"][SIZE=1]0.62[/SIZE][/TD]
[TD="align: right"][SIZE=1]307.79[/SIZE][/TD]
[TD="align: right"][SIZE=1]615.57[/SIZE][/TD]
[/TR]
[TR]
[TD="align: right"][SIZE=1]4.28[/SIZE][/TD]
[TD="align: right"][SIZE=1]0.57[/SIZE][/TD]
[TD="align: right"][SIZE=1]287.17[/SIZE][/TD]
[TD="align: right"][SIZE=1]574.35[/SIZE][/TD]
[/TR]
[TR]
[TD="align: right"][SIZE=1]4.29[/SIZE][/TD]
[TD="align: right"][SIZE=1]0.54[/SIZE][/TD]
[TD="align: right"][SIZE=1]267.94[/SIZE][/TD]
[TD="align: right"][SIZE=1]535.89[/SIZE][/TD]
[/TR]
[TR]
[TD="align: right"][SIZE=1]4.30[/SIZE][/TD]
[TD="align: right"][SIZE=1]0.50[/SIZE][/TD]
[TD="align: right"][SIZE=1]250.00[/SIZE][/TD]
[TD="align: right"][SIZE=1]500.00[/SIZE][/TD]
[/TR]
[/TABLE]

This shows the effect on cycle life that a few millivolts of overcharge can have.
For instance, if a cell is rated for an expected 500 cycles (when charge is terminated correctly at 4.20V), that would mean you may expect to get about 400 cycles if you have a charger that terminates at 4.23V.
Conversely, if your charger is the opposite way inclined, and terminates charge at 4.17V, one might expect to get about 600 cycles out of the same type of cell. And the effect on capacity per charge if termination voltage is 4.17V is not really that much, as shown in the chart on that linked article. Given what I've read, I think I'll set my iCharger 206B to terminate at 4.15V from now on, to extend the life of my 18650 cells.
I'd like to see studies backing up those numbers, but I think it's worth doing anyway, for the sake of a few mAh per cycle.

li%2520ion%2520termination%2520voltage%2520vs%2520cycle%2520life%2520factor%25202.jpg

SilverFox · Oct 27, 2013

Hello Disciple,

As you can see there are many ideas about how to maximize the life of these cells. I don't know which method is the best, but I can tell you what I have been using.

I charge my cells to 4.1 volts. While giving me a little less runtime (I have spares) it seems to give enough runtime for immediate use. Unfortunately this is a long term limited test. It has only been going on since 2006 but so far there has been little degradation. Perhaps that was a good year for Sanyo cells...

I am due to check them again in January so we will see how they do.

Tom

markr6 · Oct 28, 2013

I'm still unsure how to treat these, but I've decided to charge them to around 4.10 or so and just store them however I want. One torch/18650 sits in my car, one in the house, one in the garage and another in a backpack that gets moved around. I also decided to skip the protected versions, saving me a lot of money. With a pair of 18650B around $15, I've decided to use them and forget it.

It still bugs me that these seem so fragile. Charging, storing, temperature, etc. I was camping over the weekend and kept anything Li-Ion in the sleeping bag with me. The NiMH in my GPS seem bullet proof so they just sit in my pack even though it got down to 27°F. With Li-Ion, I'm always worried about waking up and finding my batteries drained (seen it before with cell phones that go from 90% to 50% over a cold night)

samgab · Oct 28, 2013

Modern Li-ion cells are actually extremely robust and hardy. These things we discuss are just the ideals for maximizing the life-span of the cells. They'll still survive well and perform excellently if they're used and abused, kept and used at all sorts of temps, etc. It's just that then they might last a few less cycles than if they're babied. Think of the type of treatment many cellphones and laptops receive, and the batteries in those have the same kind of chemistry as the 18650 cells we use in flashlights.

T0rch · Oct 28, 2013

poeee said:
General rule from the RC world with lipo's is if you aren't using it for 2 or more weeks put it in a storage voltage.

Norm said:
Flashlights are a completely different case you know when you're going to use your RC models flashlight use is less predictable.

Norm

Doesn't change that the battery is still a lipo and storage voltage would be the better choice when not in use.

KITROBASKIN · Oct 28, 2013

T0rch said:
Doesn't change that the battery is still a lipo and storage voltage would be the better choice when not in use.

Having an RC car I can tell you that my use of it is very different than the way I use flashlights. Never is my RC car "on standby" 24/7 like a flashlight. Or is it put away as one would with a torch. Like you said though, it is still LiPo we're talking out, just a different mission.

reppans · Oct 28, 2013

Just to add one more data point, this chart was taken from a military study of 18650s on different charging habits, and it seems to substantiate what BatteryUniversity says. The main points I take away from it are:
- shallower depths of discharge
- lower average voltages, and
- refraining from full 4.2V charges
all tend to be less stressful to the Li-ion cells.

I personally just use lithium primaries for ready-use storage, and keep Li-ions at half capacity in a fridge if not being used, and then only charge them to 4.1v when I am using them. I'm not too bothered about flashlight batts though, they're cheap and easily replaced, but I take more care of my Apple products with non-removable Batts. For example, I tend to use my iDevices about 40% per day, so I'll charge using a light timer between 5-7am striving for a 70-30-70 cycle.

Maximizing the lifespan of 18650 cells

Enlightened

Enlightened

Flashlight Enthusiast

Newly Enlightened

Retired Administrator

Flashlight Enthusiast

Enlightened

Newly Enlightened

Flashlight Enthusiast

Newly Enlightened

Flashlight Enthusiast

Flashlight Enthusiast

Flashlight Enthusiast

Flashaholic

Flashaholic

Flashlight Enthusiast

Newly Enlightened

Flashlight Enthusiast

Flashlight Enthusiast

Similar threads