Can 2.8 volts burn out an LS?

I'm intrigued by the Luxeon Star, but before I order an expensive amber Lambertian, I would like to make sure it will do the simple job of surviving 2.8 volts---the maximum voltage of 2 NiMH cells.

I've studied the spec sheet. In particular, the maximum ratings. It specifies max current and junction temperature, but not a maximum voltage.

Digging deeper, I find characteristics data, which leads me to believe that a properly heatsinked unit would draw well under the rated maximum current at 2.8 volts.

The key here, of course, is "properly heatsinked". Included in the spec sheet is an ominous chart that derates the maximum current at high junction temperatures, but no chart correlating current with junction temperature, implying there's at least a possibility of thermal breakdown if the device fails to follow its characteristic behavior for any reason.

And vendors selling the device "as-is" doesn't instill confidence.

I just want to run the thing off two AA cells at normal room temperature, and not have to build a relatively complicated current limiting circuit, or find a 350 mA fuse!

Has anyone here managed to sizzle one of these things with only 2.8 volts?

I run my LS mods at 2.9 and 3.0 volts. No problems except for some heat buildup. Heatsinking took care of that relatively minor problem. Each LS runs differently, though. Some will draw 350+ mA at lower voltages, some at higher voltages. You'll need to test the LS when you get it. However I HIGHLY suspect that you will have no problem whatsoever at 2.8 volts. But I think 2 NiMH cells will only put out about 1.2 volts each (2.4 for 2).

BTW: welcome aboard "Duggg"!
- Doug P.

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by Duggg:
I've studied the spec sheet. In particular, the maximum ratings. It specifies max current and junction temperature, but not a maximum voltage.
<HR></BLOCKQUOTE>

We all learned that the spec sheets do not give accurate data, or the units have a pretty high tolerance at this state of production. Although it is unlikely that they overheat at 2.8V it is not completely impossible. You have to try your actual sample.

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>
The key here, of course, is "properly heatsinked". Included in the spec sheet is an ominous chart that derates the maximum current at high junction temperatures, but no chart correlating current with junction temperature,<HR></BLOCKQUOTE>

You have to make your own calculations depending on your setup. They just can't give you a current versus junction temperature diagram.

Thanks Doug, glad to be here!

The 2.8-volt figure is a maximum for two freshly charged NiMH cells, and I figure if the LED is going to sizzle, it'll sizzle then.

The voltage then drops quickly to somewhere around 2.6 volts. At C/10, I'm hoping the the large cells should provide that voltage for many hours before eventually dropping to 2.4 volts.

It'll be interesting to see, though. There's a HUGE performance difference between 2.4 and 2.6 volts.

I have been using a LS in a 2 AA headlamp with no problems. I modified a Coleman 2AA headlamp that I got from ****'s Sporting Goods for $9.00 with a LS. No resistors. No boards. No sinks. No nuthin'. Just soldered wires from the LS to the battery leads and it has performed flawlessly. The lamp does not heat up much at all. I would guess a couple of degrees. I am even running AA lithium batteries, which have been known to put out as much as 1.7V each.

I have a total of 10 hours of burn time on the current set of batteries, Li AA's. I did a burn time test with a set of cheapo Alkaline AA's and got about 35 hours of usable light. (When I say "usable light" I mean I could safely travel thru a cave with it, brighter than a photon.)

I am quite pleased with light.

I suppose I should qualify my need for heatsinking - I use LM2621 voltage regulation boards to power the LS. Since the board will feed as much amps to the LS as it possibly can draw, they heat up a little. The LS won't heat up much at all on 2-AAs as the AAs have enough internal resistance that they won't feed out the amps the LM2621 will. The stock aluminum heatsink PCB should be fine for a 2-AA application.

Caver - If you can get digital photos of your mod, I'd be happy to post them on my site to share with everyone. Contact me via e-mail if you're interested.

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by Quickbeam:
The LS won't heat up much at all on 2-AAs as the AAs have enough internal resistance....
The stock aluminum heatsink PCB should be fine for a 2-AA application.<HR></BLOCKQUOTE>

Again: This all may be true for your sample, but not for everyones. We learned that Luxeon LEDs have pretty high tolerances and data sheets are usually behind the state of the art. It is unlikely that you may overheat a LED with 2 AA LiFe but not impossible (but when you get such a low volatge one it would be easy to sell it at a higher price, as several people are looking for it).
For example the Luxeon's datasheet has a given minimum voltage of 2.25V, so it will easily be possible to fry it at 2.6V. If you get a LED from the lower end of the tolerance band, it will draw some 500mA at 2.6V at 25°C junction temperature. As junction temperature will rise fast when we follow your recommendations (not external heatsink), current will even be higher. For example when you use no external heatsink but you have the LED vertical in free floating air, junction temperature will be at some 80-90°C, maybe even more.

And don't forget you loose a lot of light when using the LED at higher temperatures (down to less of one third of initial efficiency at our 80-90°C). So it is smart to add a heatsink anyway.

The good thing is, that it is unlikely that you get a sample from the lower voltage end and that your NiMH cells will be down at 2.45V very fast (or even lower).

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>This all may be true for your sample, but not for everyones... <HR></BLOCKQUOTE>

Hence why I said in my first post: "Each LS runs differently, though. Some will draw 350+ mA at lower voltages, some at higher voltages. You'll need to test the LS when you get it."

And in my second post that: "The stock aluminum heatsink PCB should be fine...", not that it is always fine "...for a 2-AA application."

Testing and experimentation are always the key. I did not and would not make a difinitive statement about the current draw or heat tolerances of the LS. I would state, however, that it is very probable (we're talking probabilities here, not possiblities) there won't be any significant damage done to an LS by running it off of 2 AA batteries without supplimental heatsinking. Testing the sample would verify it's tolerance of a 2 AA configuration, if you felt it was worth the bother.

Personally, I'd just hook it up and let 'er rip! The idea is to have fun, not turn everything into a scientifically valid and repeatable experiment. We don't want to scare off the new folks!

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by Quickbeam:
Personally, I'd just hook it up and let 'er rip! The idea is to have fun, not turn everything into a scientifically valid and repeatable experiment. We don't want to scare off the new folks!<HR></BLOCKQUOTE>

Yes, we have completely different approaches. I'm still thinking in a technical way (not scientific). If the main idea is to have fun, it's better just to try it and not caring about some lost parts.

Personally I (and all the other people I work with at lights) would not even think about making their own circuits with such a low knowledge as some people here do. These behavior may cause many missed opportunities..

Sorry for scaring anybody off (if ever).

Actually I was wrong with the voltage, the lower limit is 2.31V.

Don't worry, you won't scare me off.

I know the LS's have been available for several months now, and that's plenty of time to build a base of practical experience, which is far more significant to me than any spec sheet.

I appreciate the different scientific and technical points of view and I understand both the risks and benefits of experimentation.

Doug

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by Duggg:
I know the LS's have been available for several months now, and that's plenty of time to build a base of practical experience, which is far more significant to me than any spec sheet.
<HR></BLOCKQUOTE>

And I actually know one knowledgeable person who killed two of them .-)

But they are advancing pretty fast, so specs are outdated all the time.

But it is no problem at all: When you get your LED, just connect it in series with an incadescent bulb to a voltage source. Take for example a 6V 400mA bulb (usual bicycle bulb here) and some 4 cells. Then check voltage and current. And look at any change in voltage (which means that the junction temp gets too high). The bulb works as some kind of constant current source.