suggestions for higher power white LEDs to testing?

As it turns out, Lumileds quotes a projected lifetime until fading to

For example:

That one claims average of 70% of initial output at 50,000 hours at 1 amp and junction temperature not exceeding 135 degrees C.

I believe this is only a projection, since I doubt many of these have been made 50,000 or more hours ago as of 8/27/2010.

Well, as to specific causes, cooking is a big one. Pay a lot of attention to cooling, and/or don't run at peak power. While it's nice to get everything it's got to give, getting the heat out is a big issue with the high power LEDs, and not getting the heat out is a major cause of degradation at higher than expected rates.

That seems very good advice to remember. I'll keep it very much in mind.

Thanks, Jon

... not cooking?

snip

Google up the youtube videos and sites and links surrounding modding the Logitech G15 and G19 keyboards. They use an open source stage lighting management program to individually address every key's LED color and brightness.

Pretty cool keyboards too.

You're an idiot, again, as usual.

Of course new is better. Unless you are bulding a race engine. Give me a "seasoned block" any day.

I do not expect you to understand which and when.

They have a hardware page at:

But it really looks to me like this is a special keyboard with all the LEDs and drive circuitry already included when you buy it. I'm not sure what is being modded, except perhaps some software? I certainly do not see any schematics there and the web site is ... well, terse.

Jon

The G19 is fully addressable. The G15 required hacking to get to run under Linux, and modding to get to run in banks of colors, and some haked down to individual key level, which is a hardware level hack.

For me, simply changing the current limit resistors on the LED boards would have been enough to make both 'brightness' settings a bit brighter.

Absolutely right. As with any semiconductor device, operation at higher temperatures reduces device life.

These big LEDs cost a lot, so it makes sense to run them as 'hard' as you can to get value for money. This not only reduces lifetime but degrades efficiency too.

Graham

Check Ebay for LED heatsinks.

Graham

Higher than what? Talk about a overly broad, vague statement. :(

You could fall back on life halved for each 10'C temp rise? Vague, but I've seen the idea expressed many times over the years.

Besides, we work with stuff that goes pop sometimes for no apparent reason.

Murphy lives! And that magic smoke is difficult to keep in place ;)

Grant.

Grant, I've worked in places that the equipment was operated well below zero. A 10° C increase would have made the equipment perform better and not affect it's life enough to notice. The record low at that location was -69°F and on windy winter days it could be below freezing inside the control room.

I've also used equipment that had strip heaters & a thermostat to maintain a minimum temperature to prevent condensation. That's why i said it was a vague statement.

Fair enough :) I don't live where it snows :)

Once I worked for AU agents for Tenney environmental test chambers, had to fix one where Tenney forgot to put in stainless fixings to survive their own environments. Customer certainly not impressed with the rusting heater assembly mounts.

Closest I've been to extreme electronics locales, at least it was in boxes and not the control room :)

Grant.

to

This was a military TV station in Alaska. The colder it was outside, the less heat was availible for the station from the boiler plant at the other end of the complex. At one time it heated the entire W.W.II base of wood buildings, but that had been diverted to newer, larger concrete buildings. There were days I couldn't get the test pattern anywhere near linear, or the cameras to settle down. They were designed to operate in the 60 to 80 F range.

Respectable manufacturers will provide the relevant data. Junction temperature will depend greatly on how good the heatsinking is too.

Graham

You're a troll again as usual.