DPower (DP) Batteries at High Current Levels

[PaulW]

I bought some of these 3v lithium CR123As to include with Arc LS gifts for family members. I also got some extras for testing to see how they would perform with High Output Lamp Assemblies. There are two CPF dealers that I know of who sell them: arcflashlight.com and batterystation.com.

Being frugal, I decided to test these at three levels of load, using only one cell at a time. The loads were incandescent bulbs or combinations of them in parallel. These bulbs were designed to operate at 3 volts or at larger voltages. Because the batteries are not matched with bulbs, the tests in themselves say nothing about the battery life in any practical configuration. But using the same load on all kinds of batteries tested provides a comparison useful for incandescent HOLAs.

Run Times to Half-Power Points.

The following summarizes the tests by showing the run times in minutes to the point that power to the bulb(s) decreased to one half of "initial level" (defined below):

__________________SureFire or Duracell____DPower
Load 1 (2.4 Amps) ___ 29 minutes ________ 21 minutes
Load 2 (1.5 Amps) ___ 48 minutes ________ 40 minutes
Load 3 (.75 Amps) ___112 minutes ________ 86 minutes [corrected]

The more detailed charts below show the run time until voltage decreases to indicated values. For reference purposes I also show "initial levels" of voltage and current early in the life of the run.

Test 1. Heavy Load.

Initial levels (at 4 minutes):
* Surefire . . . 2.17v 2.33A
* DPower . . . 2.27v 2.41A

Run-time values:
Volts___Amps___SureFire__DPower
2.00____2.26___14 min.___13 min.
1.55____2.04___26 min.___18 min.
1.11____1.81___32 min.___26 min.


Test 2. Medium Load.

Initial levels (at 6 minutes):
* Surefire . . . 2.39v 1.53A
* DPower . . . 2.41v 1.53A

Run-time values:
Volts___Amps___SureFire__DPower
2.16____1.44___28 min.___25 min.
1.68____1.30___45 min.___36 min.
1.20____1.15___54 min.___46 min.


Test 3. Light Load.

Initial levels (at 9 minutes):
* Duracell . . . 2.60v 0.75A (I ran out of SureFires)
* DPower . . . 2.55v 0.74A

Run-time values:
Volts___Amps___Duracell__DPower
2.50____0.74___54 min.___30 min.
2.32____0.71___85 min.___57 min.
1.80____0.64__106 min.___77 min.
1.29____0.57__115 min.___91 min.

Disclaimer: I did these tests to compare performance at current levels for high current lamps. I don't believe it's useful or accurate to extend the results to low-current, regulated lights. I mention this because, as far as I know, the discussion of DP batteries began here, when Peter Gransee announced availability of the DP cells. In that thread, Kevin of batterystation discussed his tests in an Arc LS where the DP and SF batteries performed almost identically.

Paul

 

[batterystation]

Thank you for the heavy burn tests. I hope this still makes them worth a buck each from Peter and I. They appear to fall closer together on LED lights. Thanks again.

 

[Gransee]

I am logging some tests on the DPs, Duracells and a new cell by Yuntong this weekend. I am using a LSH-P hooked to a light meter/data logger to provide nice graphs of each run. I will post these in the Arc forum.

Peter

 

[newg]

I don't quite see where the summary 138 minute number for the Duracells at light load comes from. From the data it seems that at 88 minutes, the DP should be putting out more power than the Duracell at 115 minutes, no?

If it is true that the Duracell lasts 50 minutes (138-88) longer at this load, that would be pretty significant.

 

[PaulW]

Newg,

You are absolutely correct. Those times at the half-power points should be 112 minutes for the Duracell and 86 minutes for the DPower.

It turns out I made two errors: one in interpolation of the raw data and one in division. I have rechecked the charts and believe they are now correct. I'll edit the original post to show this correction.

Thanks very much for pointing this out. /ubbthreads/images/graemlins/thumbsup.gif

Paul

 

[Ginseng]

Paul,

Thanks again for sacrificing all that lithium to advancing our knowledge of battery performance. I have a question. Do you ever see recovery after rest periods? By that, I mean if you run a SF123 at 2V, 2.26A for 8 minutes, cut the load, rest the battery for some period of time and then run the test again. Do you get more than 6 minutes until the half power point?

I am interested because all my incandescent usage is intermittent and I'd be happy to find that short burst usage allows one to wring more of the useable capacity out of a cell.

Wilkey

 

[PaulW]

Wilkey,

I have not run any such tests in a scientific fashion, taking data in my little book. But I do remember a few experiences that may be of some use. After I do a battery run, I disconnect the load first. Although the ammeter drops to zero, the voltmeter is still recording. My impression is that the voltage rises, following an exponential curve to a final value that may take a day or two to reach.

Wait a minute. I now remember a test of the Carley 612 bulb in a Mag 2C. Because the reflector is not rated for such a high wattage bulb, I ran three 15-minute segments, separated by 15-minute recovery periods. In fact, I entered the data in my little book, and I documented it here. Admittedly this is a measurement of light intensity, rather than voltage or current or power. However, you can see that a recovery period does give a measureable, albeit short-lived, boost. It would be interesting to plot these data.

I guess a test could be done to measure exactly what happens and whether more of a cell's capacity could be wrung out. I think I'll elect to save myself some time and not do it. But I'm more than willing to speculate on the answer to your question. It seems, from intuition (read guessing) and a cursory inspection of the data, that for short recovery times one cannot wring any significant amount of extra juice. In reconnecting the load after such a recovery, the voltage is higher, but then drops again assymptotically toward values that would have been experienced without any recovery time -- maybe to even slightly lower levels.

Capacity after a longer recovery period is another question, and perhaps it is this you are really interested in. It seems that someone familiar with the chemistry of batteries could give us some good thoughts based on theory.

Perhaps all of this is just a complicated way of saying: "Gosh, I really don't know." /ubbthreads/images/graemlins/icon3.gif

Paul /ubbthreads/images/graemlins/grin.gif

 

[treek13]

[ QUOTE ]
Gransee said:
I am logging some tests on the DPs, Duracells and a new cell by Yuntong this weekend. I am using a LSH-P hooked to a light meter/data logger to provide nice graphs of each run. I will post these in the Arc forum.
Peter

[/ QUOTE ]
PaulW, thanks for sharing the results of your tests.
And Peter, I can't wait to see your results especially since you will be using the 123 light I use most frequently (plus I love graphs /ubbthreads/images/graemlins/grin.gif).

Pat

 

[PaulW]

Peter,

I too love both graphs and my Arc LSs. Thanks for offering to show us your results. Both Pat and I will be waiting.

Paul

 

[newg]

[ QUOTE ]
Do you ever see recovery after rest periods?

[/ QUOTE ]

Energizer's data shows a significantly longer run time for a 123 cell powered repeatedly for 3 seconds with a 27 second rest than one powered repeatedly for 3 seconds with a 7 second rest.

http://data.energizer.com/batteryinfo/other_available/datasheet.htm

I seem to get ridiculous amounts of recovery in my amber Arc LS. It will switch into moon mode, then I will pick it up the next day and get 20 minutes of regulated power again.

 

[PaulW]

Thanks Newg for finding this information. I wonder what it really means. When I look at the chart here, what I see is:
<ul type="square">[*] 3 sec. on and 27 sec. off at 900 mA causes voltage to drop to 2.5 after ~1600 cycles.
[*] 3 sec. on and 7 sec. off at 1200 mA causes voltage to drop to 2.5 after ~1150 cycles.
[/list]
That's a quite complex relationship among three variables: duty cycle, current, and number of cycles. The run time for Test A with the 10% duty cycle is quite a bit longer than the run time for Test B with the 30% duty cycle. But the current for Test B is quite a bit greater. This clouds the issue, and I don't quite know how to make sense of it.

Your experience with your Arc, on the other hand, seems to suggest quite strongly that some recovery time boosts the current capability out of the Arc's regulator. Perhaps it helps in the "wringing out" process, although I have read that cycling a battery in the Arc -- although it increases time available in sun mode -- actually decreases time in moon mode.

. . . a lot of stuff to think about.

Paul

 

[newg]

I think I remember somewhere reading a technical brief on how lithium batteries work and coming away with the impresion that "rest" rejuvenates 123s because during normal operation "gunk" accumulates on the anode, reducing the available area for the chemical reaction.

Apparently, the "gunk" slowly disperses at rest. This would mesh with PaulW's contention that cycling increases time in sun mode at the cost of moon mode time.

I think it could be pretty huge if someone could figure out a circuit that would manage the battery power so as to reduce "gunk" accumulation during constant use.

 

[James S]

I played with some circuits that would do that /ubbthreads/images/graemlins/smile.gif I used a slow cycle timer that switched the battery on and off once or twice a second, or even less. When it was on it charged a capacitor that ran the LED while it was off. With very small loads, like a single LED it did seem to make a measurable difference, but my patience wasn't such that I could do too many tests. With higher loads it did not make any difference, you were letting the battery recover a bit, but you were draining at double the amperage during the ON cycle to charge the cap and so this reduced the life of the battery.

But you need to increase the output of the battery enough to also cover the timer circuit and any losses in the extra circuitry. It might be possible to carefully tune to get a little bit more runtime this way, but not significantly so because of those things.