Review: Generic LiMn (IMR) batteries

This week I received my order of Lithium Manganese (aka LiMn, IMR, or LMR) batteries from They had great prices (now a little higher) and carry a lot of different sizes of batteries. The advantage of LiMn batteries is they are supposed to be able to safely provide high currents of 5C to 8C based on their capacity. Their capacities are lower than standard li-ion batteries (Lithium Cobalt). So, for instance, whereas Trustfire flame LiCo 14500 batteries are 900 mAh, the LiMN batteries I got are 600mAh. AW sells IMR batteries in that size that are also 600mAh, but for $6.50 each whereas I got a pair for $5.99 (now they are $7.99, and bestinone has a $2.27 handling fee per order, providing a tracking number).

I don't have a scale, but these seem pretty lightweight. The 14500 and 18650 sizes do not have button tops and will not work in a couple of my lights that don't have springs on each end (like my Ultrafire X1). Because these fall pretty closely in line with their nominal sizes instead of being an extra couple of mm for the protection circuit (LiMn batteries don't need protection circuits), the 18650 isn't long enough to make contact with the little spring on the 2.8A driver in either my L2P or 504B hosts. The voltage of most of the batteries out of the package was just over 3.8V except one of the 14500's which was 1.6V. LiMn batteries are not supposed to be left at a low voltage, but it charged okay.

I'll include some pictures later, but here are some dimensions:

10440: 350mAh, 44.35mm long x 10.15mm diameter (has a button postive end)

14500: 600mAh, 48.60mm long x 14.05mm diameter (flush positive end)

18650: 1500mAh, 65.30mm long x 18.20mm diameter (flush positive end)

The challenge is in evaluating the high discharge rate of these batteries. My Turnigy hobby charger has a maximum discharge rate of 1A which wouldn't even adequately test the puny 350mAh 10440, since that is just a discharge rate of 3C.

So I devised a test using my XM-L driven by the 8x7135 NANJG driver from KD. This drop-in draws a steady 2800mA which is pretty good exercise for the 600mAh 14500 (that comes out to 4.67C). The battery started out with an at-rest voltage of 4.18V. Testing was tricky because I don't have a 1xAA P60 host and I wasn't crazy about doing an endurance test with a XM-L at this current. I used my Solarforce L2i as a host with the head down in a shallow metal pan of water (just a few mm deep to get good contact) with the pan on top of some ice water to keep it cool. I left the tailcap off and connected clips to the positive spring of the driver and the rear threads of the host, leading back to my DMM and the battery. You've got to love magnets for some of this stuff. Every three minutes I would check the current which I was monitoring continuously and eventually I started checking the voltage under load (I went back and picked up the starting voltage under load with a freshly charged battery), stopping when I got to 3V (3.04V actually, close enough). Due to the declining voltage, the light didn't stay in regulation for very long at all, just 3 minutes. But no explosions and the battery didn't seem to get hot (and the ice water kept the light cool). Despite not being in regulation and the current dropping with the voltage, the calculated capacity actually exceeds the nominal: 667mAh measured instead of 600mAh nominal. I calculated the mAh by taking the average of the current at 0 and 3 minutes, multiplying by 3 minutes, and dividing by 60 minutes do get mAh. I did this for each 3-minute interval and then added all of the mAh values to get the total capacity.

Minutes Voltage Current (mA) mAh
0 3.84 2850
3 NA 2830 142.00
6 3.38 2540 134.25
9 3.38 2330 121.75
12 3.31 2020 108.75
15 3.22 1670 92.25
18 3.04 1030 67.50
Total: 666.50

I may try something similar with the 10440 and 18650. I connected the 10440 to the same drop-in and measured 2500mA draw, so 7C seems to be its maximum output. AW doesn't even offer 10440 IMR's, so I kind of wonder about this size being true IMR. But 2.5A is pretty good. I might test it with a 1400mA driver, which would be a 4C draw.

I will add to this review, but for the 14500 at least, these generic LiMN batteries seem to be pretty decent and fairly well suited for driving an XM-L, albeit briefly, even though I don't have any AA-size lights with an XM-L (yet!).

I mailed these batteries to Mitro in June and he did some much better tests on them than I could do, testing them out at different current levels and creating graphs. Those results confirm what I found, which is the 14500's have higher than expected capacity, the 10440's have less, and the 18650's have somewhat less. But he was also able to test that the 14500's and 18650's hold up pretty well under higher loads like you would expect from true LMR batteries. So the conclusion I make is that you could do okay ordering the 14500 and 18650 batteries from bestinone.

Mitro's tests start here:

Thanks brted, this looks like an interesting new option for battery aficionados.

Thanks for the review brted.

Hmmm . . . got some pretty awful results with one of the 10440's (129 mAh capacity at 1A). Will check the other one.

129 mAh, ooh that realy isn't good. I assume that part of the advantage of the IMR's is that their capacity should be reduced less by high loads than with a conventional Li-ion?

I've tested ICR 10440s and gotten similar results. I think they just change the printing on the wrapper. :)

I like your method of testing a lot. I think you're getting a fairly accurate estimation of the capacity. Those 14500 look great. Better than the IMRs I got from lighthound (see: AW vs. Lighthound IMR) at 3A. I think I'm going to grab some of the 22650's to test.

I'd love to test the cells you got so that you could verify your measurements. If you have any interest in that, PM me.

The 14500's seem the best of the bunch so far.

On the Turnigy at 1A, the first 18650 tested out at 1330mAh. vs. nominal of 1500. In a flashlight runtime test, I calculated 1265 mAh. This isn't very impressive. But I'm wondering if maybe the Turnigy isn't fully charging the batteries.

I wouldn't advise anyone to order these until I can do a little more testing, including some tests in flashlights of the 10440 and 18650 batteries.

Once I'm done, maybe we can work out a deal where mitro can test the cells too. I realize I don't really need these because I don't have any small high-drain lights and the 18650's I have can take the 2800mA draw of the XM-L I have.

I wonder if those 22650 fit in a L2i body.

Did you manage to test the other IMR 10440? I'm trying to decide whether to buy protected 10440's or these. 130mah on the first one is just too low.

Also, the l2i will fit the 22650's. They're 24mm inside.

I did a test at 1000mA of both of the batteries on the Turnigy and got 188 and 190 whereas the label says 350mA.

Ran the test again, topping them off at 600mA this time to squeeze a little more charge into them and they hadn't lost much charge at all from a couple of weeks ago when I was last messing with them. Discharging at 1000mA one battery now has a capacity of 204mAh and the other a pathetic 46mAh (see below). They sag way, way down under load (down to about 3.7V right away). That's enough for me to say avoid these. In fact, I'd avoid all of the "IMR" batteries from BestInOne just to be safe.

Later: whoops. I may have accidentally tested one battery twice. So the 46mAh reading may have come from the same battery that had just been discharged. I charged the battery I thought was 46mAh and it took just a few mAh, starting out almost immediately at 4.20V. A discharge then gave me 203mAh. So both of these match pretty closely, but well below nominal and I still don't recommend them.

Thanks for the info Brted...I think you're right, best to avoid all of these.

Really? The 10440 and 14500's only seem to be about 2/3 of the capacity of the li-cobalts which doesn't seem too bad.

But doesn’t the cell also have to fit into the tailcap of the L2i? Thats ~22mm.

Good point :D.

Well I got the 18350s and 22650s today so I'm charging in preparation for testing, but I tested the 22650 in the L2i and it DOES fit inside the tailcap until it comes to the threads for the switch retaining ring and then it stops. This results in the cell being a little too long to screw down fully. It does function but the o-ring doesn't engage the body. I think I can find a way to grind down the threads that are holding it up so that it fits properly. (I'm thinking just a sanding drum on the Dremel should do it.)

How would these be compared to Sony, Samsung 18605 LMR s I have?

Here's the 18350s at 3 Amps: (click thumbnail for graph)

I tested both cells 3 times. I'm not real thrilled with the way #2 is lagging behind and just the consistency in general. Still... they perform acceptably. I just wish I had some of the Ultrafire 18350s to compare them to. The #2 cell is getting more capacity with each run, but I wish I knew what I did to get 915mAh out of #1. :) Here they are compared to the AW IMR 18350s:

I took the middle run from each of the generics and here is a single (only) run a pair of the AW IMRs. I like the curve of the AWs a lot better. That would give you higher current reading from the MF UF dropin or the like. The generics look like they're running out of wind at 3A and I bet there would be a big drop in capacity at 5A. But I don't think I'm going to try 5A. They are already getting warm at 3A. I'm doing 1.5A tests now, but I figured that 3A would be most relevant.

I'm going to try them both with the Manafont Ultrafire dropin and get some readings just to compare further.

EDIT: ok... done :)

(MAX from turn-on - tested with my Vichy VC99 with leads from my hobby charger)

AW - 4.325 amps

Generic - 3.506 amps

As long as they hold up, I think they are a good buy. I wish I had a larger sample size, but its safe to say that if your light will hold them, they're going to do better than ANY 16340.

Are you testing the 22650's? I think they're fit a lot of lights that are 3aaa with 18650 option, but most of those aren't terribly great. I have two though.

Yup... I will. I just haven't gotten to it yet.

The 22650 are pretty good. Between 2000 and 2100mAh @ 3A. Here is a graph of an AW IMR 18650, the generic 22650, and a Sanyo 2600 18650 (not IMR):

The only problem it that they don't fit in an L2i like I had hoped. It's possible to make them work, but the machining involved is beyond me. I murdered an L2i tailcap with a Dremel trying to make it work. What we need is solarforce to crank out a 26650 host for 4Ah of goodness. Since the uses for these cells are limited at this point, I'm not going to test them in depth.