Rechargeable CR123 3.7V - highest available capacity?

[Foxel1]

Hi,

i recently got a Pila GL2 light with XR-E LED and now looking for CR123 rechargeables with high capacity. I already bought 4 of them on ebay, rated 1000mAh, but in my discharge rig all of them only deliver about 620mAh at 660mA discharge current, even after cycling them more than 5 times, and also about the same on half the current. I also contacted buyers of cells that another seller offers, also rated at 1000, and one buyer said to me they only bring about 800 at best, after cycling a lot (dont know at which current he tested though). So im wondering if 1000 is even a practical value for this battery size at the moment? I saw the reviews on rechargeable CR123 in the sticky thread, but they all are rated sub-1000, making me believe that this isnt really possible right now anyway.
Does anyone know more about this?

I know that those Hongkong sellers on ebay arent really trustworthy (just look at all the negative feedback) and the stuff they sell is most likely rubbish, but until now i believed that at least the batteries would live up to their specs, since only experienced manufacturers can make those cells, so one would think theyd be honest and print the true capacity the cells were designed for, but apparantly i was mistaken and they really fool people this way

I just did a quick calculation by volume, using the 18650 with 2400mAh as reference, and using the mAh/cubic mm to get to the 16340(about CR123) cell, and its only 990mAh

but this means it should be possible right

 

[mdocod]

as the cell increases in size, there is less "dead" space wasted. 18650 is where the use of space gets pretty nicely balanced... There are some components in cells that don't change much in size as the cell size increases (like the thickness of the canister, or the PCB)... So the amount of space available in a small cell to dedicate to actual "capacity" is a smaller percentage than on a larger cell...

Seems about the best performing RCR123 cells are still around 600-700mAH. Anything claiming 1000mAH in that size, is exaggerating. AWs high current RCR123 cells seem to deliver about 600mAh into a hefty load, so that's not bad. At lower drain rates (like around 600mA) I'd expect them to perform around 700mAH

 

[SilverFox]

Hello Foxel1,

Sorry, 600-650 mAh is about the capacity of the R-CR123 cells. There have been a couple of 700 mAh cells reported during the first charge, but their capacity quickly drooped off on subsequent charges.

Tom

 

[SilverFox]

Hello Mdocod,

Aw's cells also seem to fall in the 600-650 mAh capacity range.

Tom

 

[DM51]

That would suggest that using RCR123s with 1.5A LAs is pushing the cells too hard, as at the figure of 600mAH which you have calculated for a heavy load like this, the discharge rate is 2.5C, not the 2C calculated from AW's figure of 750mAH. Obviously 2.5C is not a problem for short bursts, but would you now say that for extended use this is a bad idea?

Let's take one example of a combination which I and I think many others now use, which is 2 x RCR123s driving a Lumens Factory HO-9. Are you now saying we should be cautious of this and similar high-current-load set-ups? Or can we continue to use manufacturer's specs when calculating suitable combinations?

 

[mdocod]

this is a common question that is brought up when discussing drain rates for li-ion cells... think about it, for every application you'd have to re-measure and calculate capacity and determine if it is safe. It's like a continually changing vicious cycle.

luckily, when a reputable manufacture states a "C" rating for current drain, they are talking about label capacity, not measured capacity into a load. So while the current drain is technically more like "2.5C" in a situation like you describe, it is still considered a 2C load by the manufacture, and still within spec. The losses from the high-drain-rate are expected and acceptable.

You'll see that in the compatibility chart, almost all the configurations that push cells to 2C, have a 20 minute runtime, (instead of 30 min like you'd get if cells actually delivered stated capacity into high loads)

When dealing with UN-reputable manufactures that are claiming 1000mah on an RCR123, I would take their other ratings with a grain of salt. If they say 2C is fine, I'd probably still only recommend 1.5A maximum.

 

[SilverFox]

Hello Mdocod,

You must also keep in mind that battery manufacturers recommend recycling the cell when its capacity drops below 80% of its labeled capacity...

Tom

 

[mdocod]

then we should throw most cells out before we even use them. lol

 

[SilverFox]

Hello Mdocod,

That brings us back up to the original point...

If a battery manufacturer is confident that the maximum continuous draw from a cell is 2C, and is expecting to get 30 cycles at that rate, what happens when we start out drawing 2.5C? Also, as the cell ages and the capacity goes down, we can find ourselves drawing over 3C. Cells wear out rapidly at higher discharge rates.

Let's look at an example. Suppose you have a cell that is labeled 750 mAh, but only is capable of 600 mAh. Let's further suppose that you have an application that runs at 1.5 amps that you use the cell in. At 1.5 amps, the cell is only capable of around 500 mAh.

You start off with a 3C actual draw (2.5C nominal draw), and by the time the cell drops to 80% of its original capacity you are at 3.75C (3.125C nominal). At these high rates you may find that it only takes 10-20 cycles before your cell needs to be replaced.

Since we understand that the cell labels are often "optimistic," I think it is good to establish a baseline run so we have something to compare to later on. If your light runs for 20 minutes, you know that you are drawing over 2C. When the runtime drops to 18 minutes, it is time to recycle your cells.

There are several ways to tell if your cells are aging. The charge time will lengthen. The ending voltage after you pull the cell off of the charger will get lower and lower. The cell will get hot during the discharge. Finally, the capacity, and runtime, will drop off. In some cases, the mid point voltage will also drop off.

These cells seem to be capable of handling high loads, but the question is how many cycles do you expect from them?

Tom

 

[mdocod]

ok, I see your point... as a cell ages and wears out, what's remaining is working harder to deliver the same load...

but... you have explained a phenomenon very well, but haven't really answered the question at hand...

do you think that when a manufacture says "2C" on a 750mah cell, they mean 1.5A max drain rate, or do you think they mean "2x measured capacity at that drain rate."

I'm pretty sure they mean 1.5A. That the cell is designed to last a reasonable number of cycles at 1.5A before 80% capacity is reached.

Obviously it will wear out faster, but the manufacture is saying that 1.5A is considered safe?

 

[SilverFox]

Hello Mdocod,

I believe the manufacturer says 2C for the chemistry mix. The label maker does not always consult with the manufacturer when picking a label for the cell.

I do not know of any Li-Ion manufacturer that condones continuous run times with standard Li-Ion cells of under 30 minutes. There are special high current cells available for electric vehicles and space travel that can handle higher loads, but they are very costly. You have to go to a different chemistry make up, such as the Molicell, or A123 cell, to safely handle currents that drain the cells in under 30 minutes. Many manufacturers limit their cells to 40 minute discharge rates.

Please note that this is for continuous runtime. Pulse discharging is completely different. Li-Ion cells can handle huge pulses, but are not up to high continuous discharge rates.

Tom

 

[mdocod]

hmmm... I could swear we had a discussion like this already.. and the conclusion was that they were talking about 2xlabel capacity drain rates. I seem to recall discussing this with AW.

but like you say, in pulses, a 20 minute runtime, used 2 minutes at a time? 5 minutes? pretty reasonable?

You've got me thinking I should include a disclaimer in the "configuration" chart. something to the effect of ..
"Do not use configurations with under 30 minutes estimated runtime in continuous runs of more than 5 minutes" or something to that effect.

I already mentioned that anything with a "20 minute" runtime is borderline safe, and to select a longer running configuration to maximize safety. Most of the runtime estimates are purposely low-balled in hopes of getting people to select longer running setups, and in hopes of getting people to plan on putting the cells on the charger sooner rather than later.(shallower cycles) I'm now thinking more should be said on this topic. Every time you have anything to say about a battery, I learn more, and worry about how much more I might not know, lol.

 

[SilverFox]

Hello Mdocod,

I believe the key to Li-Ion chemistry is the midpoint voltage under load.

The cells are designed to maintain a midpoint voltage of 3.7 volts. If you use the cells with a load that holds that midpoint voltage, you can hope for around 500 cycles. Looking at the manufacturers discharge specifications, we find that they list the maximum draw as the load that brings the midpoint voltage down to about 3.5 volts. The cycle life is reduced at these loads, but the manufacturers don't specify how much is lost. I usually see > 500 cycles when used within the specifications (loads that maintain a 3.7 volt midpoint voltage), and then there are references to 300 - 500 cycles in "normal" use. One might speculate that if you run a load that drops the midpoint voltage to 3.5 volts, you loose 40% of your cycle life. Running your cells at higher loads will further reduce your cycle life.

Now, let's take a look at the graphs in the Li-Ion Shoot Out. You can see that cells are capable of higher loads. What you don't see is the effect the higher loads have on cycle life. I generally stop testing when I notice the cell temperature rising, and when there is a dip in the voltage that recovers when the cell chemistry heats up.

You may also notice that for cells larger than the R-CR123 cells, the 1.0 amp discharge capacity is reasonably close to the labeled capacity...

OK, so far we don't have any problems, we just have cells that wear out a lot faster than "normal." If you want to add a disclaimer, I think you should advise people that when their cells drop below 80% of their initial capacity, they should stop using them.

You can also point out that high "C" rate discharges = greatly reduced cycle life.

The real problem surfaces when you use "aged" cells that have dropped below 80% of their initial capacity. I have observed tests on damaged cells where a 1.0C discharge was terminated because the cell temperature rose above 160 F. This particular cell charged somewhat normally (charge time was extended and ending voltage was low, but cell temperature remained cool), but crapped out during the discharge.

On the subject of pulse duration, I think of pulses from a very short duration up to 30 - 60 seconds. I believe Alkaline cells for digital camera use are pulsed at 10 - 15 seconds/minute for a certain number of minutes per day. I am not sure what a good pulse duration would be for Li-Ion cells.

Tom

 

[mdocod]

Tom: excellent. thank you for the massive wisdom here..

I think I'll go with something like this:

1. if cells are heating up abnormally during discharge or charge, discard. (slightly warm to touch probably normal, HOT to touch=very bad)
2. if cells drop below 80% initial capacity (read runtime), discard.
3. if cells are spending a long time on charger to reach 4.200V, especially in conjunction with heating up, discard.
4. If cells are coming off charger, and loosing voltage within a few hours of resting. (like dropping from 4.19 to 4.11), discard.

 

[SilverFox]

Hello Mdocod,

Excellent summary, however I would like to pick a couple nits...

#2 A good way to initially check this is to pull the cell from the charger, let it rest for 10 - 15 minutes, then measure the voltage. The voltage should be over 4.000 volts per cell. 4.000 volts resting is roughly 80% of the capacity of the cell.

This assumes that your charger is a true Constant Current/Constant Voltage Li-Ion charger with a hard current shut off at the end, and has the voltage clamped at a maximum of 4.200 volts. Chargers that "trickle charge" may throw this measurement off a little.

#3 actually works differently.

Aged cells take longer to charge because the voltage ramps up to 4.200 volts very quickly. Once the cell reaches 4.200 volts, the current drops off and the cell is charged at a slower rate. A healthy cell spends roughly 25 - 30% of it charge time in the Constant Current phase of the charge. An aged cell only spends 5 - 10% in this CC phase. Once the cell enters the Constant Voltage phase, the charging current is tapering off and the charge time is extended.

I think it would be better to say that an increase in the typical charge time is a cause for concern.

Here is an example. A normal 1C charge with a healthy cell takes around 90 minutes. I have a "crap" cell that took over 6 hours to charge. This cell is very sick, so this is an extreme, but if your charge time starts dragging out, this is a signal that the health of the cell is in question.

I don't have numbers that I can give because people are charging at a variety of charge rates. If you are using a 0.5 - 0.7C charge rate, the charge should be completed in around 2.5 hours. If it stretches out to 3 hours, that would be cause for concern. The same goes for 1C charging. The normal charge time is roughly 1.5 hours. If that time extends to 2 hours, that would signal concern.

Tom

 

[mdocod]

HMMM.


part of the problem, Is trying to sum up these safety issues for various chargers..For exmaple: My WF-139 always "goes green" on my RCRs at 4.16V, while it ramps other cells to 4.19-4.21 when it "goes green"... so on a cell that comes off at 4.16(they have done this since brand new)... would the "above 4.000V" rule still apply the same, or would we lower it by a few hundredths to account for the lower "off-the-charger" voltage. hehe. hehe... this is tough, obviously, more to say on this topic than we all wish there were.

I've modified the guide a bit, if you have a chance to look it over, I'd appreciate it. I *think* I have things covered reasonably well now.
Thanks again Tom
Eric

 

[SilverFox]

Hello Eric,

4.16 volts is about 96% charged, so that looks good. On the other end, 4.0 volts is actually about 79%, so there is some wiggle room already built in.

I'll take a look at your guide and see how it is shaping up.

Tom

 

[Foxel1]

About the Ultrafire WF139, i dont know if any of you ever looked inside, but i did:
The negative contacts are spring loaded as one might guess, but here is the problem: the springs themselves are not really good choice, they have about 0.1 ohm resistance or more (dont remember exactly) and cause a voltage drop in charging of around 0.1 Volts. This means the voltage at the cell is always a bit different than whats at the regulation on the PCB, and the offset is basically dependent on the current flow (Ohms law

) so the PCB never really knows what voltage the cell exactly has, i figure they built in some fixed offset, eg. PCB output is 4.3 (it actually is around 4.35 if i remember correctly) so cell get roughly 4.2V on CC mode, but when the current drops, this slowly starts to equal

the ~4.3 on the PCB (less current -> less voltage drop across resistor)

Now one might think hey just bridge the springs with a small cable - but then you would give the cells 4.3 V from PCB directly, youd have to modify the PCB regulation as well to make it put out actual 4.2, but i couldnt be botherd to try doing that (first have to analyze how it works with all its custom ICs and all) and simply went with the Pila charger

. It uses a different spring method and works ok.

 

[mik1111]

Hi all,
this is my first post,new to flashlights so if i make any mistakes or have posted in wrong section im very sorry!
my question is that i just got an ultrafire c5 which works great on aa nimh bats.
i also ordered a li-ion 123 charger and 4 x 3.6 volt 123 bats off ebay which im sure are the unprotected type,i was reading somewhere that the accepted voltages on the c5 are 1.2-3.7v, i also read that straight off the charger the 123 bats are at about 4.1 or 4.2 volts would this be bad for the light and possible blow it?
if so what can i do to drop the voltage on the batteries a little(maybe leaving it off the charger for a few hours before use?)


any help would be greatly appreciated.
thanks regards and hopw everyone is well.

 

[Forlix]

If the light accepts "3.7" or "3.6" V batteries then there is no problem, as those are nominal voltages, and cells with those nominal voltages have a charge ending voltage at 4.2 which is normal.