Looking for Information on Sub-Lumen LED efficiency~~Will smaller run longer?

I'm a longtime flashaholic and I recently started messing with simple mods and 5mm LEDs. I made something very similar to anapplesnail's minimag forever mod.

Just out of curiosity I started a runtime test on Sept01 mainly as an experiment and it made a handy nightlight-at the time it was drawing 2-3ma and I wanted to see how long it would run. It has now been 4 months and it is running strong @2.456v <1ma. I can no longer detect current on my DMM. I'm not sure at what voltage it will quit emitting visible light.

I decided to do some 4 month in beam shots and it occurred to me when comparing it to my new ZL H502

It seems to me I'm getting MUCH higher lumen per watt with this 5mm LED. (EDIT:at the sub lumen level)I know a part of that is due to direct driving it, and also the efficiency of running twice the voltage and halving the load per cell, but I would think being a high end CW LED the XM-L would still out perform the $.55 HCRI LED.(EDIT:at the sub lumen levels)

I know I can't lump all 5mm LEDs together with all the different makes and models, but is there any general information or even some rating I should be looking for when it comes to underdriving?

I've listed the tail cap current I've measured on the different modes and the manufactures rated output and runtime. Brand new gen 2 eneloops, opened and charged on a BC-700.

Low: L1 2.7 Lm (4 days) or L2 0.34 Lm (3 weeks) / 0.06 Lm (2 months) / 0.01 Lm (3 months)

Click to expand...

I get 2.7lm=17ma--.34lm=5ma--.06lm=2ma--.01lm=1ma(solid 1ma)

Is there a informative thread or something I've missed? Am I missing something obvious in my assumptions?

One of my next projects will be making a light box or sphere, I see some different DIY threads if you want to recommend one, please do.

beamshots

Re: Looking for Information on 5MM LED Efficiency~~More or less than Cree at low lows

Cree has developed their big LEDs for big current. I think that individual variation will have more impact than inherent differences in output... But I haven't done the research you've done. It's very tough to compare outputs with such different beam patterns. The short version is, you're in the optimum range of the one LED, and ten thousand percent below the optimum of the other.

Re: Looking for Information on 5MM LED Efficiency~~More or less than Cree at low lows

Not sure what the big deal is. Lumens per watt for the 5mm is less than that for an X-L.

Take your output at 17mA for example. You get 2.7 lumens. What you seem to have forgotten is forward voltage at 17mA drive. It is probably around 3V. So lumens per watt is ~2.7/(0.017*3) = 53.

If you look at the XM-L T6 bin, it is rated at 280 lumens at 700mA drive. Forward voltage at that current is rated at 2.9V. Thus, efficiency is 280/(2.9*0.7) = 138 lumens/watt, or almost 3X that of the 5mm LED.

If we go to lower drive currents for the XM-L, we get for example:

~170mA drive, ~70 lumens, ~2.7V Vf --> 70/(2.7*.17) ~150 lumens/watt

Re: Looking for Information on 5MM LED Efficiency~~More or less than Cree at low lows

Hi justin, I'm more interest at the sub lumen level, if you have info on any CREE LEDs at those levels I'd love a link.

apple snail: I see what you mean. I guess when the XM-Ls with 3 month runtimes came out I expected it would be the most efficient source of low ultra long runtime-I expected a race horse to also be a mule.

Re: Looking for Information on 5MM LED Efficiency~~More or less than Cree at low lows

Looking at the XM-L datasheet, we can linearize the relative flux vs forward current curve at low currents. The two relevant points are (0,0) and (25%, 170mA), approximately. So at 1% relative flux, we have about 170/25 = 6.8mA forward current to generate 2.8 lumens for a T6 bin XM-L. The Vf is probably about 2.6V, so efficiency is 2.8/(2.6*0.0068) = 158 lumens/ watt. There are lights on the market that probably run the XM-l at such low drives to obtain the so-called "moonlight mode".

So maybe not quite sub-lumen, but I still fail to see what the big deal is with the 5mm LED. At a power draw of 2.6V*6.8mA = 0.018W, an XM-L can deliver about 2.8 lumens. The 5mm LED needs about 0.05W to deliver the same output, or about 3X the power draw.

Re: Looking for Information on 5MM LED Efficiency~~More or less than Cree at low lows

I'm not thinking this is a big deal. I was just asking if there was any real data on underdriven LEDs. IMO some light manufacturers advertise their sub lumen levels and long runtimes as much as the turbo modes.

To me this comparison shows at this level(.06lm) XM-L or H502 to be very inefficient, based on one runtime test no means conclusive, but not what I expected. Although like anapplesnail said they are not the same beam profile, to me the H502 and the flat or cut LED in candle mode are both pure wide flood.

Also if I had to give up either anything over 200 lumen or under 2.7 lumen I would give up the 200+ lumen modes. To me efficiency at sub lumen levels is a valid topic.

Re: Looking for Information on 5MM LED Efficiency~~More or less than Cree at low lows

Oh dang I posted this in the wrong forum

I meant to post in "LED" discussions

Edit: Thanks for the move:wave:

Re: Looking for Information on 5MM LED Efficiency~~More or less than Cree at low lows

Justin Case said:

Looking at the XM-L datasheet, we can linearize the relative flux vs forward current curve at low currents. The two relevant points are (0,0) and (25%, 170mA), approximately. So at 1% relative flux, we have about 170/25 = 6.8mA forward current to generate 2.8 lumens for a T6 bin XM-L.

Click to expand...

That may be the case, I'm far from an expert, but on that order of magnitude? Even the extrapolation to 2.8lm is 47 times the brightness of the .06lm comparison. Also you can not draw a linearize for increase in brightness with any accuracy after just one order of magnitude or we could have 2800 lumen at 1.7amp right?

Re: Looking for Information on 5MM LED Efficiency~~More or less than Cree at low lows

The cree led's should trump 5mm efficiency, as has been mentioned. Here is a good thread for reference.

Re: Looking for Information on 5MM LED Efficiency~~More or less than Cree at low lows

Thanks hoop that is sorta what I was looking for, even though I'm looking for levels much much lower. I have no doubt it should be more efficient, even with driver losses.

I was surprised that after 4 months of decreasing brightness the 5mm LED still puts out more than the H502 on the .06lm setting with rated runtime of 2 months(1AA not 2). I know this is a poor comparison, I'm working with what I got and trying to learn.

Edit:
Reading a bit more of that thread it seems the XR-E may be the way to go for efficiency at lower drives
Maybe smaller is better for low lows.

Re: Looking for Information on 5MM LED Efficiency~~More or less than Cree at low lows

The problem is the data sheets don't lend much information to really low drive currents. You might give cree a phone call and see what they have to say. It is my belief that the crees are going to still be around 150 lumens per watt at extremely low drive currents. Probably can't go wrong with the latest gen 2 offerings.

Re: Looking for Information on 5MM LED Efficiency~~More or less than Cree at low lows

Hoop said:

You might give cree a phone call and see what they have to say.

Click to expand...

That had never even occurred to me. I'll probably try e-mail.

For me this is curiosity, having the sub lumen level is good enough, do I need it for 4+ months? No.

Re: Looking for Information on 5MM LED Efficiency~~More or less than Cree at low lows

Direct drive test on a Cree XP-G R4 using a bench supply to light up the LED (mounted on a 10mm diam, 2mm thick MCPCB):

At 2.601V, I get an If of 4.96mA. Output is bright. At 2.503V Vf, I get an If of 0.58mA. Output is medium (at least as bright as any LED indicator light on some panel). At 2.449V, If=0.17mA. Output is very low.

As my previous calculations showed, your efficiency figures are skewed because you are going through some driver electronics. Pure LED direct drive is going to be more efficient. The Cree XM-L datasheet shows that efficiency could be ~3X better than trying to drive the LED through a driver and measuring current draw through the tail. I would expect an XP-G to behave exactly the same as an XM-L.

Thus, I doubt that the 5mm LED is more efficient than your typical Cree XP-G or XM-L even at low drive.

Re: Looking for Information on 5MM LED Efficiency~~More or less than Cree at low lows

Thanks Justin that is exactly what I was looking for. It does seem from the thread hoop posted the xr-e does become more efficient at lower driver currents well above the sub lumen level.

AFAIK a comparison of different cree, nichia, and rebel leds has not been done on the sub lumen levels. I for one would be curious which is more efficient. Maybe I need a bench power supply more than another light...

Re: Looking for Information on 5MM LED Efficiency~~More or less than Cree at low lows

ToyTank said:

Thanks Justin that is exactly what I was looking for. It does seem from the thread hoop posted the xr-e does become more efficient at lower driver currents well above the sub lumen level.

AFAIK a comparison of different cree, nichia, and rebel leds has not been done on the sub lumen levels. I for one would be curious which is more efficient. Maybe I need a bench power supply more than another light...

Click to expand...

All InGaAn LEDs I have seen a curve for have an efficiency peak. Go above that drive current and the efficiency drops. Go below that drive current and the efficiency drops. Hence a good 5mm LED , ie Nichia, could have a better efficiency at a low current, say 1ma than an xml or similar.

Semiman

Re: Looking for Information on 5MM LED Efficiency~~More or less than Cree at low lows

You would be better off with a variable resistor and a good milliamp meter for sub-lumen measurement. You would need to spend a lot on a bench supply for good low current control.

Re: Looking for Information on 5MM LED Efficiency~~More or less than Cree at low lows

ToyTank said:

Thanks Justin that is exactly what I was looking for. It does seem from the thread hoop posted the xr-e does become more efficient at lower driver currents well above the sub lumen level.

AFAIK a comparison of different cree, nichia, and rebel leds has not been done on the sub lumen levels. I for one would be curious which is more efficient. Maybe I need a bench power supply more than another light...

Click to expand...

I have made some measurements using high end sourcemeters down to the nanolumen range, but I do not have the time to crunch the numbers at the moment. I made an older post with some preliminary numbers, though, in the forum.

I can tell you, though, that at sub-lumen, smaller is better. But you really have to go VERY low in order for 5mm leds to get better than XB-Ds, for example. If you want to get a millilumen, for example, a Nichia 5mm will be better...