Most Efficient >100lumen 1xAA light?

[RedForest UK]

It seems that as led lights get more and more efficient, it is becoming more and more possible to carry a single AA edc light which gives over the magic number of 100 lumens with a very useful runtime as well. My question is, which single AA light gives the best efficiency mode (read: output x runtime) at an above 100 lumen minimum output?

I am currently thinking that the Fenix LD10 and the Zebralight SC50 are currently frontrunners, with the Quark AA just behind.. Can anyone else throw any more light on the subject?

Examples: (check light-reviews.com/selfbuilt for some graphs) Figures are estimated on eneloop/2000mah Ni-mh cell. All numbers are approximated.

1. Fenix LD10 Turbo: 170 lumens, 50 minutes = 8500lm*min
2. Fenix LD10 High: 92 lumens, 100 minutes = 9200lm*min
3. Zebralight SC50:122 lumens, 100 minutes = 12200lm*min
4. Zebralight SC50w: 107 lumens, 100 minutes = 10700lm*min
5. Zebralight H501:96 lumens, 100 minutes = 9600lm*min
6. Quark Mini AA: 80 lumens, 120 minutes = 9600lm*min
7. Quark AA R2: 90 lumens, 120 minutes = 10800lm*min
8. Quark AA R5: 109 lumens, 80 minutes = 8720lm*min
9. Nitecore D10 GDP:105 lumens, 70 minutes = 7350lm*min
10. iTP A2: 75 lumens, 120 minutes = 9000lm*min

If anyone has any to add or wishes to clarify anything then please feel free..

NB: While the methodology, accuracy and validity of this test may be questionable, my point is that while output and runtime are what we look for in a light, no-one as yet seems to have given us any sort of comparable number of the two multiplied together for different lights. For those for whom efficiency is a top consideration I thought that it would be really good to at least attempt to provide that sort of information. This is an admittedly rather basic attempt at that, but even if all it does is prompt someone else to do the same but better then great, it has achieved it's purpose! :candle:

 

[red02]

lm/m = 170lm/50m = 3.4 lm/m, aren't lumens adjusted for time like watts already?

wouldn't that be lm*m? not lm/m? argggg the units are getting me confused.

I'm surprised to see the SC leading the pack. I wonder if the iTP numbers are solid... I was considering the Olight I15 and the LF5XT for a while, I think they are worth a look despite lacking lm/m numbers.

EDIT: unrelated question...

 

[RedForest UK]

Erm.. I'm not really sure what you mean at the start, I dont mean lumens divided by minutes for the l/m part, I just meant l/m as a unit of measurement. In a similar way, a watt could be called an a/v. Sorry if it's confusing, it's my fault because I dont really any better terminology for it. I have seen similar comparisons for watt hours and things like that I thing, which would be w/h.. Well yeah, sorry about that, but I think the basic premise makes sense.

One of the biggest problems is the different sources the numbers are taken from and variability between samples. For example, I have never measured an SC50 or seen any runtime graphs or even actual output readings, I just took the specs for output and then extrapolated their runtime figure to a 2000mah cell in a similar way to their H501 model, which is rated the same runtime by them on a 2700mah cell, but gives 100minutes on an eneloop. Whereas the Fenix results and the Quark are taken from an independently verified source. Therefore, all of these results, in any comparative sense are approximate, but still pretty interesting.

 

[Quension]

No, / signifies "per" as a divisor, not a multiplied combination. The closest equivalent to watt you're thinking of would be VA (although that's technically not the same measurement -- long story), the same way you see battery capacity in mAh meaning milliamp-hour, and watt-hour as Wh.

For your unit of lumen-minute, use lm*min.

 

[daimleramg]

It seems that as led lights get more and more efficient, it is becoming more and more possible to carry a single AA edc light which gives over the magic number of 100 lumens with a very useful runtime as well. My question is, which single AA light gives the best efficiency mode (read: output x runtime) at an above 100 lumen minimum output?

I am currently thinking that the Fenix LD10 and the Zebralight SC50 are currently frontrunners, with the Quark AA just behind.. Can anyone else throw any more light on the subject?

Examples: (check light-reviews.com/selfbuilt for some graphs) Figures are estimated on eneloop/2000mah Ni-mh cell. All numbers are approximated.

1. Fenix LD10 Turbo: 170 lumens, 50 minutes = 8500l/m
2. Fenix LD10 High: 92 lumens, 100 minutes = 9200l/m (I know this one is slightly under the 100 lumen minimum)
3. Zebralight SC50:122 lumens, 100 minutes = 12200l/m (Based on Zebralights specs adjusted to a 2000mah cell. Pretty impressive, I'd love to see an actual test of this!)
4. Zebralight H501:96 lumens, 100 minutes = 9600l/m
5. Quark Mini AA: 80 lumens, 120 minutes = 9600l/m
6. Quark AA: 90 lumens, 120 minutes = 10800l/m
7. Nitecore D10 GDP:105 lumens, 70 minutes = 7350l/m
8. iTP A2: 75 lumens, 120 minutes = 9000l/m

If anyone has any to add or wishes to clarify anything then please feel free..

Some of these lights are emitter lumens and some are OTF, you need to adjust the emitter lumens to OTF lumens to get an accurate comparison.

 

[RedForest UK]

No, / signifies "per" as a divisor, not a multiplied combination. The closest equivalent to watt you're thinking of would be VA (although that's technically not the same measurement -- long story), the same way you see battery capacity in mAh meaning milliamp-hour, and watt-hour as Wh.

For your unit of lumen-minute, use lm*min.

Okay, sorry, I know that it is generally used as a divisor and that my original way of writing it was technically incorrect and that this might be causing confusion. I will now edit the original post to make it more clear.

 

[RedForest UK]

Some of these lights are emitter lumens and some are OTF, you need to adjust the emitter lumens to OTF lumens to get an accurate comparison.

Thanks for the input, I had actually considered that and all of the lumen ratings that I have used so far are based on OTF values recorder from sources such as 'light-reviews' which are independent from the manufacturer. However you are correct that for the zebralight outputs I have not been able to source any actual OTF lumens and so have simply listed the specs they give, although I had always thought that zebralight gave very accurate and generally true lumen ratings for their lights based on OTF ratings.. :thinking:

If you can point me to a source where there are accurate OTF lumens given for any of the zebralight models then I would really appreciate it

 

[Ragiska]

the current quark is 109 otf

zebralight does list otf also.

 

[daimleramg]

Thanks for the input, I had actually considered that and all of the lumen ratings that I have used so far are based on OTF values recorder from sources such as 'light-reviews' which are independent from the manufacturer. However you are correct that for the zebralight outputs I have not been able to source any actual OTF lumens and so have simply listed the specs they give, although I had always thought that zebralight gave very accurate and generally true lumen ratings for their lights based on OTF ratings.. :thinking:

If you can point me to a source where there are accurate OTF lumens given for any of the zebralight models then I would really appreciate it

Because some of your listings are tested by different sources why not ask bigchelis to test these lights. If all are tested with the same sphere it will give us a more accurate reading(some people get different lux readings from the same light with different lux meters).

 

[AnAppleSnail]

Examples: (check light-reviews.com/selfbuilt for some graphs) Figures are estimated on eneloop/2000mah Ni-mh cell. All numbers are approximated.

1. Fenix LD10 Turbo: 170 lumens, 50 minutes = 8500lm*min
2. Fenix LD10 High: 92 lumens, 100 minutes = 9200lm*min (I know this one is slightly under the 100 lumen minimum)
3. Zebralight SC50:122 lumens, 100 minutes = 12200lm*min (Based on Zebralights specs adjusted to a 2000mah cell. Pretty impressive, I'd love to see an actual test of this!)
4. Zebralight H501:96 lumens, 100 minutes = 9600lm*min
5. Quark Mini AA: 80 lumens, 120 minutes = 9600lm*min
6. Quark AA: 90 lumens, 120 minutes = 10800lm*min
7. Nitecore D10 GDP:105 lumens, 70 minutes = 7350lm*min
8. iTP A2: 75 lumens, 120 minutes = 9000lm*min

As noted above, every '/' should be a *. We speak of "Miles per hour" because it's miles/hours, while amp*hours is amps times hours. As a side note, 'm' is taken by 'meter' and I abbreviate lumen as 'lm,' so I would write lm*min.

Many of these lights have imperfect regulation, and others claim emitter lumens. What you'd want to use is one of the regulation vs runtime charts that shows %output over time. Consider a flashlight rated at 100 lumens for 2 hours. If it had perfect regulation, it would grant (100 lm * 120min) = 12000 lm*min. But let's say that it drops linearly from 100 lumens to 50 lumens over its claimed 2-hour runtime. Then it would have fewer lm*minutes to claim.

50lm*120 minutes (the minimum output) + (100-50)lm * 120 minutes * 1/2 (triangle of decreasing output) = 9120lm*min. This is rather less than 12000 lm*min.

 

[RedForest UK]

Yes, okay guys I'm sorry.. I know that this can never really be an accurate measure of efficiency, I prefer runtime graphs as well. I acknowledge that the lights output will drop off as hardly any single AA light is perfectly regulated. However, I am perhaps slightly naively assuming that most lights will drop off relatively evenly, and have so far only included pretty well regukated lights up to where they start to drop out of regulation.

Also, yes I think the results for the quark that I used were for the R2 version, as I couldnt find an independent source for the R5 one. I will add those as spec from 4sevens now.

It would be great if Bigchilis or someone did some tests of genuine output with a more varied number of production lights but to put it simply I don't think he has them available.

Selfbuilts runtime graphs are always my go-to info for efficiency, but as he hasnt reviewed all of the lights available, I was hoping to set up some rough comparison as to how others might compare efficiency wise.. :shrug:

 

[kramer5150]

Interesting...
Isn't Lumens per Watt (Lumens/Watt) a better way to quantify overall efficiency?

Lumens should be measured OTF in bcs and MrG's sphere, as opposed to emitter specs and manufacturer data sheets. Surefire lumen specs could be used instead of CPF measurements... IMHO.

Watt being Vin x Iin

I have never seen efficiency calculated as a multiplicative product of 2 variables. Its always what you get OUT divided by what you put IN (OUTPUT / INPUT). In this case it would be Lumens OUT / Wattage IN

 

[RedForest UK]

Yes, that would also be a good way of measuring the efficiency, if applied to account for the entire circuits efficiency including the driver etc, as it is only normally applied to the efficiency of the emitter itself. The lm/w could probably be extrapolated somehow from the lm*min value and the 2000mah @ 1.2v capacity of the cell, but I do not have time right now. Maybe tomorrow when I have done my final exam I will have a bit more time and feel a little less stressed..

I am not currently calculating the efficiency from an input and an output, I am merely calculating overall output over time from two seperate output variables. This when coupled with the fact that all of the lights are running on the same power source means that it can imply efficiency of the system.

 

[kramer5150]

I think you can perform a basic calculation... like this. I ended up with Lumen*Hours / Watt. A surefire E2L (KX2) is shown. I don't know what value this has. Someone double check my math.

I suppose you could omit the Wattage consumed variable and it simplifies down to the product of Lumens x Hours.

 

[AnAppleSnail]

I think you can perform a basic calculation... like this. I ended up with Lumen-Hours / Watt. A surefire E2L (KX2) is shown. I don't know what value this has. Someone double check my math.



I suppose you could omit the Wattage consumed variable and it simplifies down to the product of Lumens x Hours.

A good measure of efficiency is lumen*hours per watt*hour out the front of the flashlight. This includes driver efficiencies, heating, LED choice, and max level choice. Do we want to get into measuring this for different levels?

Kramer, I think you want to use the average voltage for RCRs rather than the maximum voltage. As I recall it's 3.7v?

8.4 v *Not average voltage! Avg is 2x3.7 = 7.4v)
110 lm
.165 amp
.5 amp-hours (RCR123)


110lm/(.165amps*7.4v) * (.5 ah/.165amp)
110lm/1.22watts * 3.03 hours
273.2 lm*hours per watt (Slightly different answer from voltage choice)

Does this make sense?
The light is handed (2*3.7v * .5Ah) watt-hours = 3.7 watt-hours
It delivers (Ask BigChelis) lumens for (actual runtime).

110 lumens for 3 hours. But SureFire lights dim over time, making this complex to nail down.

 

[kramer5150]

A good measure of efficiency is lumen*hours per watt*hour out the front of the flashlight. This includes driver efficiencies, heating, LED choice, and max level choice. Do we want to get into measuring this for different levels?

Kramer, I think you want to use the average voltage for RCRs rather than the maximum voltage. As I recall it's 3.7v?

8.4 v *Not average voltage! Avg is 2x3.7 = 7.4v)
110 lm
.165 amp
.5 amp-hours (RCR123)


110lm/(.165amps*7.4v) * (.5 ah/.165amp)
110lm/1.22watts * 3.03 hours
273.2 lm*hours per watt (Slightly different answer from voltage choice)

Does this make sense?
The light is handed (2*3.7v * .5Ah) watt-hours = 3.7 watt-hours
It delivers (Ask BigChelis) lumens for (actual runtime).

110 lumens for 3 hours. But SureFire lights dim over time, making this complex to nail down.

I don't think this light dims much at all. it generates no heat and during the 3-4 minute sphere measurements we did it was stable at 110L OTF.

**EDIT**
But yeah that makes sense to average the Vin. FWIW E2L tailcap current increases to .22A with a pair of SF-CR123 cells, that each measure 3.12V, so the regulation circuit is doing what its supposed to.

Sorry Redforest for derailing your thread away from AA and using a 2xCR123 light. I merely used the one light that I KNOW to be a fixed 110L OTF, and highly regulated. I think you can calculate the performance of any light this way.

 

[entoptics]

I think the simplest way to get a definitive answer to this question is for one person to test all the lights in the same lightbox, graph their results as output vs time (e.g. Selfbuilt), and then calculate the area under the curve.

Calculus FTW!

Well, not really calculus, since the curves probably won't have an easily identifiable equation and thus won't be suitable for integration, but the concept is the same.

Oh yeah, and I guess "Simplest" is relative term...:shrug:

 

[ZebraLight]

I think the simplest way to get a definitive answer to this question is for one person to test all the lights in the same lightbox, graph their results as output vs time (e.g. Selfbuilt), and then calculate the area under the curve.

Calculus FTW!

Well, not really calculus, since the curves probably won't have an easily identifiable equation and thus won't be suitable for integration, but the concept is the same.

Oh yeah, and I guess "Simplest" is relative term...:shrug:

You don't have to find (curve fitting) the equation to do the integration. You can simply add up all the readings from the lightbox at fixed intervals (say every 5 seconds). If you take the very first reading at 2.5sec, then it's a Riemann's Middle Sum, a common numerical integration method.

 

[kramer5150]

Just for the halibut (drum roll please) i did similar calculations using highly regulated lights that I have measured in BC's sphere and got the following results

SF-A2 Incan = 5.35 Lumen*Hours / Watt:mecry:
10.7 Lumens/Watt:candle:

Vin = 8.4V
Lumens = 90Lu
Cell Ampere Hours = .500 AH
Tailcap Current = 1Atc

Zebralight H501 = 119.9 Lumen*Hours / Watt
56.29 Lumens/Watt

Vin = 1.32V
Lumens 85Lu
Cell Ampere Hours = 2450AH (using energizer 2450, which I have seen others measure at or above 2450 AH)
Tailcap Current = 1.15Atc

Malkoff M60-Q5 = 35.18 Lumen*Hours / Watt
42.93 Lumens/Watt

Vin = 8.4V
Lumens = 220
Cell Ampere Hours = .500AH
Tailcap Current = .61Atc

The formula is heavily influenced by the cell capacity. Up-ing the Malkoff M60 to 2x18650 (Sanyo 2400mah) jumps to 168 Lumen*Hours / Watt. Along those same lines if you could lego a 2x18650 E series body and drive the E2L-KX2 head you'd be humming along for 15 hours at 1153.96 Lumen*Hours / Watt.

This is a GREAT thread. We've all heard it before..."Lumen output, small size, run time... pick any TWO". This formula kind of mathematically puts that concept down on paper... in a geeky Grant Imahara kind of way.

*EDIT*
I think the biggest sources of error are in the cell capacity assumptions I had made.

 

[PeaceOfMind]

Cool thread. Would definitely be neat to have a giant table of lights and their efficiency (lumen hours per watt, or whatever we agree on) for all the lights for which we have good independent data on their actual output/output curves. It'd be a big job to compile, but very useful.