'Over-driving' an LED , some measures

After a bad experience with an XM-L which lost a significant quantity of flux because was drived at 5A for around 2 seconds a close look here.

Note; the measures of this box are only for comparative between the same kind of emitter, mean , given the different spacial distribution the XR-E perform way better than XP-G or XM-L

Note;stock's emitters are in green, bought from vendors in blue, trouble emitters in red.

Note; all emitters are feed with a power supply and are mounted in a proper 55 grs heat sink with fins , a lot of area , which give the thermal dissipation of an average good torch.That heat sink has a place in the window of the box as a way to repeat the conditions.

As the measures show overdriving is a bad idea, the stock P10's which were never badly treated perform almost like the virgin R5 meanwhile the G10's feeded at 2A very little time , maybe half an hour , show less flux...

There are more in the graphic....

Excellent work, Arenet! I'm not sure how I missed this post. I noticed it now when you linked to it from another "technical" thread.

This goes hand in hand with what I observed during a multi-xml build..which was signifigant dimishing returns after 3A.

Great information up there. Thanks for doing this. I wonder if /how much the UF980Ls life is greatly reduced as it is driven over 4 amps on high.

Arenat- On your test bench, i wonder if youre able to cool that xml die with lets say a peltier cooler to keep temps really cool if it can handle the higher amps without loss of flux.

Do you have a pic of the setup you use to measure this?

Or maybe a description of your thoughts on the design/build?

Granted the heatsink used is a little small to deal with 10 Watts but is way better than a P60 setup can do.The general idea of this test were find what can practically achieved. Notice that the numbers are at start only especially at high power.

I'm not totally against pushed hard ,in fact I've an R2 at 1,4A gaining very little if any at high cost but this is under 6W and the heat could be managed at least 4 minutes before kill the deal entirely.

"Arenat- On your test bench, i wonder if youre able to cool that xml die with lets say a peltier cooler to keep temps really cool if it can handle the higher amps without loss of flux."

I would love to made that test but lately I've had a lot of 'casualties' , 3 xm-l , 1 xp-g 2 xr-e ...our currency worth cents at change therefore aren't cheap and the few still in the mail system has a place assigned .

"Do you have a pic of the setup you use to measure this?

Or maybe a description of your thoughts on the design/build?"

Nothing to show up but work . Is a box the size of shoes one with a window and some kind of tunnel one third the way in . At one side receiving very few light is located the sensor of the meter . The box has a rectangular hole where the sensor can slide to calibrate it.

As patron an ITP A3 was used . The readings 1 Lux= 1 Lumen in the three modes are consistent and plausible as are all the readings of the emitters tested .

This box do not work with flashlights because many reflection issues but with bare emitter work fine. I know the many flaws of the thing but for the intended purpose, more or less light, work.

ThankĀ“s for test it

Excellent info! It's what scares me about the 980L. it's been known to peak at 4.7-4.8A and it's just not worth the risk to me especially at $45.

Many thanks guys..

title corrected