Review: Xtar 3100mAh 18650 protected cell

As this is the first full battery review that I have conducted, I think it is worthwhile to quickly list the equipment used, and briefly explain some of the measurements taken.

Equipment

For consistency of results, batteries were always charged on the same charger, which was a Jun-si iCharger 106B+. Internal resistance and discharges were both carried out on the same charger, in this case, an iCharger 208B. Two chargers were used for convenience, in order to speed up the testing process.

Dimensions were measured with a Precision Gold digital caliper, and weights measured with a Neva 50gram digital scale, said to be accurate to 1/100 gram.

Voltages were measured with a Precision Gold WG 020 Multimeter. While I do not have the means to calibrate this, I also have a UNI-T UT60 multimeter, and measurements on the two correlate.

Internal resistance measurements, and discharges, were carried out using approximately 12" lengths of 16AWG cable, with either a crocodile clip hardwired to the cable, or a 10x10mm high strength, rare earth magnet soldered to the wire. Each wire is terminated with a gold plated banana plug for connection to the charger. Shorter wires would have reduced IR further, but could have placed strain on the wires, and increased likelihood of the magnet or clip being pulled off the battery.

Measurements

Some of the measurements that I will be taking may seem a little pointless, but add to the overall information available, and certainly do no harm. For instance: -

Initial voltage - this is the voltage of the cells as I received them. It doesn't mean a whole lot, as the person who sold you a cell could take a 0V cell, with a tripped protection circuit, give it a storage charge, and then send it to you, and you would be none the wiser. Still, this would at least indicate that the person knew something about li-ion cells. Certainly, no-one wants to actually receive a cell with a very low voltage or a tripped protection circuit.

Capacity from initial charge - even less important, but with todays high capacity cells, I personally find it interesting just how much capacity is available even from the storage charge. In the case of cells with lower than normal termination voltage, I can also use this step to get an idea of the difference in capacity when discharging the cell down to 3.0V (this will be my normal termination voltage when running a discharge), and when discharging it to the manufacturer's stated minimum termination voltage.

With no further ado, lets do some testing.

Cell: Xtar 3100mAh 18650

Base cell: Panasonic NCR18650A

Basic chemistry: ICR (LiCo)

Charge voltage: 4.2V +/-0.05

Discharge voltage: 2.5V

Purchased direct from Xtar, for $26 including shipping.

Results at a glance - initial (please check my final results further down)

Cell 1

Cell 2

Initial voltage

3.625V

3.630V

Quoted length

68.5mm +/- 0.5

Quoted width

18.4mm +/- 0.1

Measured length

69.02mm

69.00mm

Measured width (max)

18.68mm

18.58mm

Weight

46.48 grams

46.37 grams

Capacity from storage down to 3.00V

1131mAh

N/A

Capacity from storage down to 2.50V

N/A

1282mAh

Internal resistance after storage charge

123mOhm (magnets)

126mOhm (magnets)

Capacity from 4.2V down to 2.5V @ 0.6A

2984mAh

2983mAh

Capacity from 4.2V down to 3V @ 0.6A

2877mAh

2855mAh

Capacity from 4.2V down to 3V @ 1.0A

2806mAh

2761mAh

Capacity from 4.2V down to 3V @ 3.0A

2284mAh

2399mAh

Capacity from 4.2V down to 3V @ 5.0A

1485mAh

1562mAh

As it happens, I wasn't happy with these results, and ended up repeating them. The cell that initially had the highest capacity ended up having a lower capacity at the higher discharge rates, despite having the lower internal resistance. This didn't seem right at all. More on this later.

Construction

As can be seen above, the cells have a standard, mid-sized button. This means that using the cells in series should not be an issue, and there should be no problems using the cells in lights that have use physical reverse polarity protection, or have issues with flat top cells for any other reason.

I was having some problems getting my magnets to attach to the button. The magnet was tending to tip over on to the shoulder of the cell, reducing contact with the button, and effecting the internal resistance. The magnet's greater attraction to the shoulder of the cell leads me to conclude that the button is aluminium rather than steel. Nothing wrong with that of course..... it just makes it harder for me!

The negative end uses a 'foil-type' contact - this could lead to wear and tear occurring quicker than with a cell with a metal plate base.

The wrap is a vinyl style heat shrink, rather than the brittle tape type that some cells have.

Light compatibility


Light

Characteristics

Fits

Functions

Olight M20

Wide tube, spring at head, sprung plunger at tail cap

Yes

Yes

SWM T20CS

Short tube, dual springs

Yes

Yes

Jet IIIM

Wide tube, tail spring only

Yes

Yes

Fenix TK15

Two-piece tube – narrow at head

Short tube, short spring at head

Yes – but snug

Yes

Fenix TK11

Narrow tube, tail spring only

No

N/A

Eagletac G25C2

Wide tube, sprung plunger at head (minimal travel), physical reverse polarity protection

Yes

Yes

Nitecore IFE2

Narrow tube, physical reverse

Polarity protection

Yes – but extremely tight

Yes

The wider of the two cells was used for this test.

The Nitecore Infilux IFE2 only just accepted this cell width wise, and required the head to be removed so that the cell could be pushed out.

Charger compatibility

Charger

Fits

Functions

4Sevens single bay

Yes

Yes

Trustfire TR-001

Yes – but tight

Yes

Ultrafire WF-139

No (1)

Yes (1)

Ultrafire WF-188

Yes

Yes

HXY-042V2000A

Yes

Yes

XTAR WP2 II

Yes

Yes

Jetbeam/Sysmax Intellicharge i4

Yes – but tight

Yes

Pila IBC

Yes Yes

(1) The cell will not sit flat in the cradle – the cradle is a fraction too short – but the cell’s positive button still makes contact with the charger’s positive terminal, and charging will proceed.

Internal Resistance

After my initial testing and the odd results it produced, and deciding that this may be related to the poor contact with the positive end magnet, I tried measuring the internal resistance with a crocodile clip holding on to the button via the vent holes. Despite the minimum area of contact this produces, and not being the most stable of contacts anyway, the internal resistance figures did lower in comparison to my initial measurements.

The below table shows the new IR figures, and shows comparisons to other batteries based on the Panasonic NCR18650A cell: -

Cell

Voltage

Internal resistance

Xtar 3100mAh 1

3.75V (storage charge)

101mOhm (crocodile clip)

Xtar 3100mAh 2

3.75V (storage charge)

104mOhm (crocodile clip)

AW 3100mAh 1

3.96V (as received)

112mOhm (magnets)

AW 3100mAh 2

3.97V (as received)

112mOhm (magnets)

Redilast 3100mAh 1

3.63V (as received)

106mOhm (magnets)

Redilast 3100mAh 2

3.62V (as received)

106mOhm (magnets)

Excell 3100mAh 1

3.75V (storage charge)

112mOhm (magnets)

Excell 3100mAh 2

3.75V (storage charge)

110mOhm (magnets)

Intl-Outdoor 3100mAh 1

3.62V (as received)

142mOhm (crocodile clip)

Intl-Outdoor 3100mAh 2

3.63V (as received)

150mOhm (crocodile clip)


Some of these cells have already seen a little use, some are brand new out of the packet, hence why some are shown as having received a storage charge, and some say 'as received'.

Discharge Cycles

These are the results that I measured after switching to connecting the positive end of the battery to the charger via a crocodile clip. Unfortunately, it is only possible to carry out the discharge from the initial voltage once, so I have carried those over from my original results.

I have carried 0.2C (0.6A) discharges down to 2.5V to measure the true capacity of the cell (0.2C is the discharge capacity that most manufacturers specify), and down to 3V to see what difference there is in capacity. I have then carried out discharges at 1A, 3A and 5A to see how the cells perform at higher rates. That should give an idea of how the cell will perform with moderately driven XR-Es, XP-Es and XP-Gs (1A), fully driven MC-Es, P7s, XM-Ls and moderately driven SST-50s (3A), and with heavily driven XM-Ls and fully driven SST-50s (5A).

I hope to add images of discharge curves at some point - once I work out how to use Logview! I have all the discharge data saved so that I can go back to it at a later date.

Cell 1

Cell 2

Initial voltage

3.625V

3.630V

Quoted length

68.5mm +/- 0.5

Quoted width

18.4mm +/- 0.1

Measured length

69.02mm

69.00mm

Measured width (max)

18.68mm

18.58mm

Weight

46.48 grams

46.37 grams

Capacity from storage down to 3.00V

1131mAh

N/A

Capacity from storage down to 2.50V

N/A

1282mAh

Internal resistance after storage charge

101mOhm (crocodile clip)

104mOhm (crocodile clip)

Capacity from 4.2V down to 2.5V @ 0.6A

2956mAh

2945mAh

Capacity from 4.2V down to 3V @ 0.6A

2832mAh

2810mAh

Capacity from 4.2V down to 3V @ 1.0A

2754mAh

2764mAh

Capacity from 4.2V down to 3V @ 3.0A

2448mAh

2429mAh

Capacity from 4.2V down to 3V @ 5.0A

1730mAh

1722mAh

Conclusions

After switching my method of connecting the cell to the charger, internal resistance was lowered. While the maximum capacity of the cells dropped in the testing that followed this change, and the extremely close capacity matching at 0.6A down to 2.5V was lost, capacity matching at higher discharge rates tightened up, and the usable capacity at 5A increased. Maximum variation in capacity was 22mAh.

Lesson learnt - find the connection with the lowest resistance prior to testing. This test should have taken half the time it did!

There was around 130mAh increase in capacity when discharging down to 2.5V, but you are still getting over 2800mAh when only discharging down to 3V, so owners of lights that cut out at 3V are loosing little, and still getting a cell with a high capacity. Capacity remains good, at over 2400mAh at 3A. A lot of voltage sag is occurring at 5A, leading to the battery hitting 3V early, and reducing usable capacity as a result.

The cells are rather wide, which could cause issues with some lights, but the button tops ensure no issues with fixed contacts or physical reverse polarity protection - if it fits in your light, it should work. It was possible to charge the cell in all the chargers I had to hand, though it was a tight fit in a couple of cases, and in one case, had to sit in the charger at an angle, though it did still charge.

I would have liked to have seen a solid metal base for longevity, but maybe this would lead to the cell being longer.

All in all though, the Xtar 3100mAh cells performed very well, providing a high capacity for low and medium power lights. High powered lights would have reduced run times due to the voltage sag that these cells suffered from.

Outstanding review Stephen

Detailed and informative .

Thanks for your responses.

If this review format (currently minus discharge curves) provides information that people would find useful, I will happily do some more testing. Future tests should only have the one results table though! And hopefully take half as long.....

Now to get on to the Logview forum and see if I can get some answers re the graphs I want to plot.....

Very nice! You say you paid $26 shipped...is that for a pair or each?

That was for a pair.

Having said that, although I purchased these, they are the same specification as the batteries that have been provided for free for some people to test. In much the same way that Xtar are selling some of the tester spec S1 lights at discount prices, it is possible that I got these cells at a lower cost than the retail spec cells will cost.

Nice review! Thank you for taking your time to do this for us.

Thanks very much Stephen for the careful review! Very well done. I just put one of your pictures at the top for the teaser to make it Frontpage and Sticky.

Thanks - much appreciated.

That is really splendid for a first review!

Nice work and good explanation on why you chose to do it with this format.

I echo the others - great job !

Have someone a link, i didn´t find it, where to buy.

I don't believe that these cells are freely available yet.

They will probably be available from here first, so you may want to keep an eye on that site: -

http://www.szwholesale.com/xtarlight-accessories-battery-c-243_245.html?osCsid=eb2acc3ca265aa747f44fff2be0759a8

No doubt, once cleared for retail sale, they will be available from any Xtar dealer. For that matter, as Xtar have dealt with us before, it's entirely possible that they will become the subject of a group buy in future.

Great review, thanks!

so is this a recommendable cell or not lol?

From what i gathered it is the best one for use with linear regulators (the famous AMC7135 based ones). As the voltage drops goes DD so on a 2,8A driver you get around 2h of decent output (800 down to 400ish lm). Very decent runtimes. There is no real advatage to go below 3V but the 150mAh down o 2.5V are there if really needed.

Excellent review! I would have liked a few hundred mAh more at 3A discharge. Are all the 3100 panasonic equal?

As you can see from the internal resistance table, I have a few other 3100mAh cells to hand, and I intend to test them all, but testing takes a long time. For instance, carrying out a 0.2C discharge test to measure the maximum capacity is always going to take 5 hours, give or take a little, if the capacity is correct. Then over three hours to re-charge at 1A. Obviously, as the stages progress, the higher the discharge rate the shorter the time, but always another three hours to re-charge. Double that for two cells, and you can see how long these tests take! So, don't hold your breath for quick results, but I'll post more tests whenever I can.

Once all the tests are done, we'll be able to see if all the NCR18650A based cells are pretty equal, or if one brand comes out ahead of the others.

With regards to capacity at relatively high current draws, these cells are known for sagging more under load than some other, lower capacity, cells out there.

Did you notice any excessive heat when using them for a high amp LED?

I have noticed something unusual about these cells. I purchased two Citipower X7-T6 from Ebay. If I use a Ultrafire 3000 Mah, the light stays cool but is not extreamly bright. When I switch to the XTAR 18700, the light is 50% (a guess) brighter at the highest setting but the light gets too hot to hold after 20 minutes. I tried switching the batteries to the other light and I can repeat the process.

Does this mean there is a problem with the 18700 or is it just that the LED pulls so many amps that the battery gets hot trying to supply these amps? I have no way to measure the draw of the LED. Could it also just be the lack of a great heatsink on an $18.00 light?

It's hard to say for sure without knowing the exact details of the light, measuring the internal resistance of the batteries, measuring the tail cap current and so on, but if I had to hazard a guess, I would suggest that the light is able to draw more current from the Xtar battery than from the Ultrafire.

Can someone elaborate on L-M-H current draw? I mean, what constitutes a high current draw? I am planning on using these in a 3 cell 3-XML torch. Recommended?

It is generally recommended not to exceed a 2C discharge, unless using cells that you know are rated to be safe with a higher discharge rate.

So, if we were to assume that these cells did have a genuine 3100mAh capacity, and that any loss in capacity I have observed is due to internal resistance in my test equipment, power consumption in the protection circuit, etc. then a 2C discharge would mean that these cells could handle a 6.2A discharge rate.

However, as you can see from the results above, even at 5A, these particular cells were seeing a noticeable degree of reduction in measured capacity, as the voltage was sagging, and hence the voltage dropped down to 3V sooner, and the dicharge was terminated before the cell would actually have been fully discharged.

While these cells may well have been safe to discharge at 6.2A, the capacity would have suffered even more at this high discharge rate.

Some LiCo cells are known not to sag as much as the Panasonic cells under load, and while their capacities may be lower (and hence the 2C discharge rate would be lower), as their voltage remains higher under load, they can run just as long or longer before the voltage drops down to 3V.

The high setting of a light is the only one that should really dictate what cells are suitable, and going back to what I said at the beginning, for most common LiCo cells, you do not want to exceed 2C. If your cells can deal with the high setting of the light, then the medium and low settings should not be an issue. That is assuming of course that the medium and low settings are not likely to risk over discharging the cells (one reason for selecting a light with low voltage warning/cut-out, or using protected cells).