18650 battery test with capacity curves for many cells

HKJ

Well-known member
Joined
Mar 26, 2008
Messages
9,698
Location
Copenhagen, Denmark
18650 battery test 2011

A 2012 test can be found here


DSC_6200.jpg




In this test I have measured many different 18650 batteries and collected the data in bar graphs and curves. The above picture shows most of the cells I have tested.



Summary.png




Capacity: Measured capacity at 0.2 ampere discharge rate (This is the average from the two tested cells). This is not measured the same way as the manufacture specified capacity, i.e. it is not supposed to be exactly the same value.

Top: SB: Small button top, LB: Large button top, Flat: Flat top (See here for more information).

Bottom: A protected battery has a PCB at the bottom, on some batteries this has been strengthen with a metal plate.

Protected: Does the battery have a protection circuit? IMR batteries do not.

Max. Dis.: Maximum allowed discharge rate. This is only listed for batteries where it is specified, for all other batteries assume 2C.

Source: Where I got the battery.

Date: The month I got the battery.



Capacity.png




This chart show the average capacity of the tested cells.



Energy.png




This chart show the average energy of the tested cells.



EnergyDrop.png




In this chart I have taken the difference between 0.2 and 2 ampere energy bar and calculated the change in percent of total energy (([email protected] - [email protected]) /[email protected]). This high lights the difference on handling loads.







Physical dimensions



18650 batteries are not all the same size, they are supposed to be 18 mm in diameter and 65 mm long, but due to the added protection circuit they are up to 5 mm longer. These charts show the size of the largest of my test batteries (The difference is usual not more than 0.1 mm to 0.2 mm for a cell type).



SizeLength.png




SizeDiameter.png




Weight.png






Combined discharge curves



These curves can be very hard to read, due to the many cells in one chart. When in doubt about which cell a specific curve is, use the charts under each cell to confirm the cell type.



0.2 ampere discharge



Capacity-0.2A.png




Energy-0.2A.png








0.5 ampere discharge



Capacity-0.5A.png




Energy-0.5A.png








1 ampere discharge



Capacity-1A.png




Energy-1A.png








2 ampere discharge



Capacity-2A.png




Energy-2A.png








5 ampere discharge



Not all batteries can sustain 5 ampere discharge and few batteries are rated for it. Because some curves are missing the color of the curves are different from above.



Capacity-5A.png




Energy-5A.png








Curves and pictures for each cell





The individual curves are here





Notes



For all battery types I have tested two different cells, they are called A and B in the graphs. All cells have been charged on a hobby charger with half the marked capacity as charge current and the charge is stopped when the current is 1/10 of the charge current. I.e. if the battery is marked 3000mAh it is charged with 1500mA and the charge is stopped when the current is down to 150mA. (Note: For Samsung at 4.35 volt I used another charger).

This charge method will not reach the maximum capacity of the batteries, there might be space for a few mAh more. Batteries that are marked with too high capacity will be penalized with this charging method.



When I only use one curve for a battery, it is the curve from battery A.



All the discharging is done on a CBA and I frequently verify that the reading from the CBA matches readings taking with a voltmeter directly over the cell.



Both capacity and energy is done in one discharge, capacity is a summation of the current, energy is a summation of current*voltage.



Cytac batteries were supplied by Cytac for a review.
 
Last edited:

Corvette6769

Well-known member
Joined
Apr 28, 2009
Messages
130
Location
Central Illinois USA, In a van down by the river
This is Google's cache of https://www.candlepowerforums.com/threads/308451&page=1 as it appeared on Feb 26, 2011 17:26:26 GMT.

The posts that were on PAGE 2 are still missing.

================================================================================================

Re: 18650 battery test with capacity curves for many cellsWritten by jasonck08 on 01-26-2011 02:28 PM GMTGreat overall test comparison. I would like to point out that the tests of the 2900mAh cells were only discharged to 2.80v (not 2.50v). At high discharge currents like 5A there is about 200-250mAh capacity left @ 2.8v.

Re: 18650 battery test with capacity curves for many cellsWritten by HKJ on 01-26-2011 02:38 PM GMT
jasonck08 said:
Great overall test comparison. I would like to point out that the tests of the 2900mAh cells were only discharged to 2.80v (not 2.50v). At high discharge currents like 5A there is about 200-250mAh capacity left @ 2.8v.
But not all light will be able to use that capacity. Some lights will not even be able to drain the battery down to 2.8 volt, mostly lights with build in protection for LiIon batteries! But not all light will be able to use that capacity. Some lights will not even be able to drain the battery down to 2.8 volt, mostly lights with build in protection for LiIon batteries!

Re: 18650 battery test with capacity curves for many cellsWritten by jasonck08 on 01-26-2011 03:50 PM GMT
HKJ said:
But not all light will be able to use that capacity. Some lights will not even be able to drain the battery down to 2.8 volt, mostly lights with build in protection for LiIon batteries!
True, but most lights that accept 2 or 3 cells, will be able to drain the cell down to 2.50v. True, but most lights that accept 2 or 3 cells, will be able to drain the cell down to 2.50v.

Re: 18650 battery test with capacity curves for many cellsWritten by etc on 01-26-2011 05:20 PM GMTMost excellent job here.

Re: 18650 battery test with capacity curves for many cellsWritten by Vbeez on 01-26-2011 06:28 PM GMTGreat review, as always. Thanks

Re: 18650 battery test with capacity curves for many cellsWritten by boboweb on 01-29-2011 08:23 AM GMTOutstanding reviewHKJ ! You done a lot of hard work - thank you very much !

Re: 18650 battery test with capacity curves for many cellsWritten by Shinpu on 01-29-2011 08:17 PM GMTexcellent post, thanks for the massive time invested.btw what happened to the Cytac 2600 (blue)?

Re: 18650 battery test with capacity curves for many cellsWritten by Loben on 01-29-2011 10:39 PM GMTThanks for this review!BTW, where can I get the cytac red 2900 battery?Please show me the link of it!ThanksBen

Re: 18650 battery test with capacity curves for many cellsWritten by HKJ on 01-30-2011 01:22 AM GMT
Shinpu said:
excellent post, thanks for the massive time invested.btw what happened to the Cytac 2600 (blue)?
They will be added in a few weeks. They will be added in a few weeks.
Loben said:
Thanks for this review!BTW, where can I get the cytac red 2900 battery?Please show me the link of it!ThanksBen
I believe that the batteries are new for Cytac and that they will show up in some shops soon, but due to Chinese New Year I can not ask them just now. I believe that the batteries are new for Cytac and that they will show up in some shops soon, but due to Chinese New Year I can not ask them just now.

Re: 18650 battery test with capacity curves for many cellsWritten by Colonel Sanders on 01-30-2011 04:20 PM GMTVery nice! And very interesting. Thanks for taking the time to do this. :thumbsup:

Re: 18650 battery test with capacity curves for many cellsWritten by jorinr on 01-30-2011 08:46 PM GMTThank you for the detailed charts! I have been trying to decide on some new batteries to buy for several of my lights and this helps a lot!

Re: 18650 battery test with capacity curves for many cellsWritten by radellaf on 01-30-2011 09:51 PM GMTGreat data - thanks for running these tests. The energy loss graph is a new one to me.Any chance of full resolution versions of these pics? Even the 1024 from a free flickr account would help.Is it the Redilast 2600 the one that is maintaining such a high voltage on the 2A curves? There are so many colors it is a bit hard to tell. If so, anyone know if the cell in there is the LG Chem 2600? (since I have some of those, unprotected) For my SST MG P-Rocket I need a slight edge on voltage (to maintain 2.5A on Hi), especially if I can't find a way to charge the special 4.35V cells up past 4.2V. Right now best I could do is buy the 4.35 version of the Microchip 500mA charger IC and the Sparkfun USB charger, but that's $40+, bunch of work, and only 500mA. I'm hoping the Yezl XM-L light is less picky.Overall, if I had to buy more cells, the Tenergy 2600 seem to come off well here, though the diameter might not work in the SC60.

Re: 18650 battery test with capacity curves for many cellsWritten by HKJ on 01-30-2011 10:50 PM GMT
radellaf said:
The energy loss graph is a new one to me.
Also to me. I wanted a way to visualize the load handling ability of the cells, that was more obvious than the capacity and energy bar. Also to me. I wanted a way to visualize the load handling ability of the cells, that was more obvious than the capacity and energy bar.
radellaf said:
Any chance of full resolution versions of these pics? Even the 1024 from a free flickr account would help.
What is it that you want in higher resolution? What is it that you want in higher resolution?
radellaf said:
Is it the Redilast 2600 the one that is maintaining such a high voltage on the 2A curves? There are so many colors it is a bit hard to tell.
Looking at the chart for Redilast, it confirms that they are good at holding a high voltage at 2A load. Looking at the chart for Redilast, it confirms that they are good at holding a high voltage at 2A load.

Re: 18650 battery test with capacity curves for many cellsWritten by radellaf on 01-31-2011 11:20 PM GMTThe charts - could email me if you have them readily at hand. I'm just trying to I.D. the lines by color, and at this res the JPEG artifacts make it so I can't do it with the eyedropper tool in Photoshop (to match colors).Mostly, on "Capacity at 2A", I'd like to know which one is holding a higher voltage than all the rest from just past 500mAh out to 2Ah. It stands out. Color could be Redi 2600, but it's so far off from AW 2600 that I don't think that's right.

Re: 18650 battery test with capacity curves for many cellsWritten by HKJ on 02-01-2011 09:43 AM GMT
radellaf said:
The charts - could email me if you have them readily at hand. I'm just trying to I.D. the lines by color, and at this res the JPEG artifacts make it so I can't do it with the eyedropper tool in Photoshop (to match colors).Mostly, on "Capacity at 2A", I'd like to know which one is holding a higher voltage than all the rest from just past 500mAh out to 2Ah. It stands out. Color could be Redi 2600, but it's so far off from AW 2600 that I don't think that's right.
There are no jpg artifacts in the curves (They are not in a compressed image format). The steps in the curve is the resolution from the CBA (10mV). Here is a chart with a few of the cells: There are no jpg artifacts in the curves (They are not in a compressed image format). The steps in the curve is the resolution from the CBA (10mV). Here is a chart with a few of the cells:
Capacity-2A-4.png


Re: 18650 battery test with capacity curves for many cellsWritten by radellaf on 02-01-2011 07:15 PM GMTOK, if I download the image, enlarge it, and use the color sampler then yes, color #5555FF is the RediLast 2600. Cool.Anyone know what cell is inside that?

Re: 18650 battery test with capacity curves for many cellsWritten by darklord on 02-02-2011 05:00 AM GMTDid you ever manage to get a graph for the CYTAC 2600 BLUE....?? These are available here in the UK so I'm interested.You say they were supplied by Cytac for a 'review'....is this the review, or is the review somewhere else??Anyone know or rate this battery?

Re: 18650 battery test with capacity curves for many cellsWritten by HKJ on 02-02-2011 06:14 AM GMT
darklord said:
Did you ever manage to get a graph for the CYTAC 2600 BLUE....?? These are available here in the UK so I'm interested.
The curves will be added later, but Cytac supplied the wrong batteries (There was 2900mAh in my 2600mAh wrappers). The curves will be added later, but Cytac supplied the wrong batteries (There was 2900mAh in my 2600mAh wrappers).
darklord said:
You say they were supplied by Cytac for a 'review'....is this the review, or is the review somewhere else??
This is part of it, I am also working on something like this with only Cytac batteries, but including some other sizes. This is part of it, I am also working on something like this with only Cytac batteries, but including some other sizes.There will not be any review with test in different lights, but if anybody want to know if a battery fits in a specific light that I have (Many of them can be seen on my website), just ask, and I will check.

Re: 18650 battery test with capacity curves for many cellsWritten by RepProdigious on 02-02-2011 06:43 AM GMT
HKJ said:
There was 2900mAh in my 2600mAh wrappers
I wouldn't complain if that ever happened to me I wouldn't complain if that ever happened to me ;)

Re: 18650 battery test with capacity curves for many cellsWritten by LEDAdd1ct on 02-02-2011 01:17 PM GMTThank you for all your hard work!:)

Re: 18650 battery test with capacity curves for many cellsWritten by jsr on 02-04-2011 11:02 AM GMTExcellent info HKJ! I don't know where to get most of those cells, but the Tenergy cells are readily available here and lower priced than the Redilast and AW cells and they seem to perform quite well.

Re: 18650 battery test with capacity curves for many cellsWritten by 357mag1 on 02-06-2011 09:45 AM GMTI have Tenergy cells that perform very well and I have Tenergy cells that perform poorly (though better than most of my Ultrafire). I have some AW cells that don't perform that well though they are consistently better than other Chinese cells. Poor quality control plagues all Chinese cells that I have used including Nimh AA - D size.I now only buy Panasonic, Sanyo or LG 18650s and the quality from cell to cell is great. I know they don't have protection but half of my protected cells circuits don't appear to work consistently anyway.I feel safe using quality cells and taking proper precautions. Safer than I do using questionable quality cells with band-aids (protection circuits). It would be awesome if some of the above mentioned unprotected cells could be included in the test.

Re: 18650 battery test with capacity curves for many cellsWritten by HKJ on 02-07-2011 12:01 AM GMTAW uses Panasonic cells in their 2900 mAh batteries.I would like to add more cells to the test, but I need a place to get them, without paying a fortune in shipping! If anybody has some good ideas, use PM to tell me.

Re: 18650 battery test with capacity curves for many cellsWritten by Helmut.G on 02-07-2011 11:34 AM GMTanother awesome test HJK!

Re: 18650 battery test with capacity curves for many cellsWritten by HKJ on 02-08-2011 01:00 AM GMTI got a good link to get unprotected "Panasonic NCR18650 2900mAh" and "Samsung ICR18650-30A 3000", they will be added when I have receive and tested them.Other links or ideas for batteries are welcome.

Re: 18650 battery test with capacity curves for many cellsWritten by jasonck08 on 02-08-2011 03:14 AM GMTDo you have the equipment to charge a 4.35v cell?

Re: 18650 battery test with capacity curves for many cellsWritten by HKJ on 02-08-2011 03:48 AM GMT
jasonck08 said:
Do you have the equipment to charge a 4.35v cell?
I have some advanced power supplies, but they would not charge in a predictable way (I would not be able to automatic stop at a minimum current). I have some advanced power supplies, but they would not charge in a predictable way (I would not be able to automatic stop at a minimum current). The best predictable solution I have is some iChargers that can charge to 4.3 volt

Re: 18650 battery test with capacity curves for many cellsWritten by jasonck08 on 02-08-2011 03:34 PM GMT
HKJ said:
I have some advanced power supplies, but they would not charge in a predictable way (I would not be able to automatic stop at a minimum current). The best predictable solution I have is some iChargers that can charge to 4.3 volt
I saw that the iCharger can charge to 4.20-4.30v. I was wondering if you can also calibrate it incorrectly so it would read 4.30v on the screen but actual charge voltage might be 4.33 or something. 4.30v vs 4.35v there might be 3-5% difference. I saw that the iCharger can charge to 4.20-4.30v. I was wondering if you can also calibrate it incorrectly so it would read 4.30v on the screen but actual charge voltage might be 4.33 or something. 4.30v vs 4.35v there might be 3-5% difference.Also, are there different iCharger models? Can all charge to 4.30v? What model do you have?

Re: 18650 battery test with capacity curves for many cellsWritten by HKJ on 02-08-2011 03:53 PM GMT
jasonck08 said:
I saw that the iCharger can charge to 4.20-4.30v. I was wondering if you can also calibrate it incorrectly so it would read 4.30v on the screen but actual charge voltage might be 4.33 or something. 4.30v vs 4.35v there might be 3-5% difference.
It is possible to change the calibration, but I do not know if I want to do that. It is possible to change the calibration, but I do not know if I want to do that.
jasonck08 said:
Also, are there different iCharger models? Can all charge to 4.30v? What model do you have?
There a a couple of different models, and the software is nearly equal in all models. I believe all can be adjusted to 4.3 volt, but I have not checked all the manuals. There a a couple of different models, and the software is nearly equal in all models. I believe all can be adjusted to 4.3 volt, but I have not checked all the manuals.I have two 208B:
DSC_6175.jpg
 
Last edited:

HKJ

Well-known member
Joined
Mar 26, 2008
Messages
9,698
Location
Copenhagen, Denmark
More great work HKJ. Thank you much.

And I must thank you for restoring the discussion in the thread.

Just curious though, with the new Cree XM-L T6 being driven at 3A, why did you skip the 3A test this time?

Time and it is always possible to interpolate between the 2 and 5 ampere trace. You must also remember that most lights uses a buck or boost driver that will change the current, i.e. running the ML-L at 3A with two 18650 in series (or parallel) would only draw about 1.5A.

Also, I picked up a CBA II and am running the CBAII & CBAIII Software Version 2.1. What software did you use this time?

I am using the CBA III with the supplied software, then I save the data in CSV format and uses my own software to make the curves. For the actual curve drawing I uses a package called JFreeChart. This way all my curves has exactly the same size and format and it is easy to make global changes or plot any combination of cells.
 

tobarger

New member
Joined
May 16, 2009
Messages
4
Thanks HKJ
The most comprehensive battery comparison thread I've seen
Nice work providing the information
 

HKJ

Well-known member
Joined
Mar 26, 2008
Messages
9,698
Location
Copenhagen, Denmark
I have updated the comparison with Panasonic 2900, Samsung 3000 cells and gray TrustFire 2400. The Samsung is tested at both 4.2 volt and 4.35 volt charge.
The missing test for Cytac 2600 has also been added.
 

Bolster

Well-known member
Joined
Oct 7, 2007
Messages
1,542
Location
Mexifornia
Great thread. Wish this forum had "a sticky of stickies," one file at the top that linked to important reference files within. I'd nominate this thread on the sticky-of-stickies list.
 

teefix

New member
Joined
May 2, 2011
Messages
2
Thank you for that great thread. Makes it a bit easier to find the right battery...

One question arose for me when I had a look at the graph:

Any idea why the RediLast 2900 and the Trustfire TF2400 are so similar. The capacity, at least from the specs, is quite different...

thank you
teefix
 

HKJ

Well-known member
Joined
Mar 26, 2008
Messages
9,698
Location
Copenhagen, Denmark
One question arose for me when I had a look at the graph:

Any idea why the RediLast 2900 and the Trustfire TF2400 are so similar. The capacity, at least from the specs, is quite different...

Similar? You must be mixing the curves (That is very easy with the number of curves in the chart).
Try looking at the individual charts, there it is much easier to see the actual capacity curves.
 

nodoubt

Banned
Joined
Mar 15, 2008
Messages
131
ok... so all those pretty graphs, and hard work are to be commended..........
my questions is......
which one is the best !! ???
how did they rank ?
 

HKJ

Well-known member
Joined
Mar 26, 2008
Messages
9,698
Location
Copenhagen, Denmark
ok... so all those pretty graphs, and hard work are to be commended..........
my questions is......
which one is the best !! ???
how did they rank ?

That depends on what parameters you look at, for high capacity protected cells the AW2900 and RediLast2900 are some of the best.
 

Budda

Well-known member
Joined
Nov 30, 2009
Messages
559
Location
Italy
Thanks HKJ, that's a lot of useful data.

I can't decide which cheap battery buy to use in low power flashlight.
 

LEDite

Well-known member
Joined
Dec 6, 2002
Messages
318
Location
Dallas, TX
I have used hundreds of #18650 cells in various UV lights I have built.

The "no-name" cells with protection circuits were UNRELIABLE.

I now only use "name brand" cells (Sanyo, LG, Samsung, Panasonic) which have passed the U.L. #1642 testing for safety.

Capacity after a hundred cycles or more is much better.

LEDite
 
Last edited:

LuxLuthor

Well-known member
Joined
Nov 5, 2005
Messages
10,409
Location
CT
I have used hundreds of #18650 cells in various UV lights I have built.

The "no-name" cells with protection circuits were UNRELIABLE.

I now only use "name brand" cells (Sanyo, LG, Samsung, Panasonic) which have passed the U.L. #1642 testing for safety.

Capacity after a hundred cycles or more is much better.

LEDite

A sensible policy. My understanding is that AW & Redilast are using Panasonic cells with their PCB circuit added.
 

LuxLuthor

Well-known member
Joined
Nov 5, 2005
Messages
10,409
Location
CT
HJK, thanks for all the work. I don't mean to sound petty, but putting the legends for the curves on a white background (vs. the color of the graph plots above), makes it impossible for me to identify which curve belongs to which battery.

I know I can look at individual cells, but seeing the comparison plots together also gives easier ability to do cross comparisons. I wonder if whatever discharge software you are using can change the legend background color and/or thicken the lines for better contrast?

Essentially what I am trying to respectfully say is that as much time as I know you spent getting all these results, I really want to be fully able to use and appreciate your results--especially the comparisons.
 

HKJ

Well-known member
Joined
Mar 26, 2008
Messages
9,698
Location
Copenhagen, Denmark
HJK, thanks for all the work. I don't mean to sound petty, but putting the legends for the curves on a white background (vs. the color of the graph plots above), makes it impossible for me to identify which curve belongs to which battery.

I know I can look at individual cells, but seeing the comparison plots together also gives easier ability to do cross comparisons. I wonder if whatever discharge software you are using can change the legend background color and/or thicken the lines for better contrast?

Essentially what I am trying to respectfully say is that as much time as I know you spent getting all these results, I really want to be fully able to use and appreciate your results--especially the comparisons.


I can probably change the background for the legend, but I do not believe that will help much, there are way to many curves in these comparisons charts.
I have been thinking about make a extra document where I split the batteries up in maybe 3 groups and make a set of charts for each group.
 
Top