XM-L on a star was attached to heatsink using just thermal grease while testing a driver (3A one)… thermal grease failed at being sticky enough, so the star detached from heatsink, and few seconds later LED was dead (I didn’t noticed it in time to do anything about it 
)
Raccoon City, you just had to do that. 
Shadow, what color does the LED turn? I heard it turns blue or something when overheated.
When my SST-90 got detached, it melted the solder right off the star. Amazing how it didn’t damage the LED.
It’s tint changed to deep-blue indeed. No idea what causes it, but it’s fascinating to watch (for those 3-5 seconds that it lasts).
I've been lucky. I haven't seen the death of any LED.
[quote=ryansoh3]
Raccoon City, you just had to do that. 
[/quote]
Yes, I had to. :evil:
EDIT: Fixed some poor grammar.
Yeah, but it was intentional. I fried a SSC P7 just for the heck of it. I poured some water onto it and it started to boil. :0
I’m not saying it can’t happen, but this poll may not be an accurate reflection of the presumed reliability of an LED. There are other factors such as “appropriate” heatsinking to take into account.
I for one have not experienced the demise of an LED while being driven within spec and with appropriate heatsinking.
Yeah, I forgot to add “appropriately heatsinked” as well.
I guess you can edit your vote if you want to. 
I said yes, but there’s a caveat to my response…
I’ve never seen a modestly driven and well heat-sinked NAME BRAND power led fail. And by fail, I assume we’re talking about being completely non-functional. I have seen them get old enough to experience a significant loss in brightness, but I still have yet to see one FAIL.
I have, however, seen lots of off brand .5w and smaller LED’s fail after only a few hundred hours. The ‘cheap’ no-brand 0805 and 5050 LEDs that are available on every chinese website do not last, even when being far UNDER driven.
PPtk
I have seen my oldest ones start to dim out and yellow slightly but never a properly setup and driven one completely burn out.
Having tortured lots of leds more than most, I must say that no...I've never lost one that was appropriately heatsinked, even at very high current levels. Then again, my idea of appropriate heatsinking isn't the norm.
Actually, only one comes to mind that I purposely killed: It was the star mounted xp-g that I did the emitter test on. Took it to 5amps (I think) and just watched it fade to blue then fizzle. A single tear soon followed. (I've got a strong suspicion that had it been directly mounted to my copper test post that wouldn't have happened.)
All white LEDs are actually royal blue on the inside. The phosphor coating turns the blue light into white. If the LED overheats, the phosphor coating can burn off and voila… blue-a-mundo.
I had several hundred of them die! The story is in the first post here: The Tao of Chinese LED Strip Lights
Never seen one die under appropriate usage conditions. When I was an intern at a lighting company and was working on the development of their first LED light I did manage to kill many LEDs, mostly by human error or not knowing what I was doing. Definitely dedomed a few, accidentally melted the dome with a soldering iron (tip directly into dome, not too much heat around it), tried soldering some Seoul P4s to star boards and didn’t realize that the part of the package under the LED needed to be reflowed as well, so saw some big time color and intensity shifts etc. Also didn’t know that the P4 was not isolated from the star so shorted some out, overall the P4 was an extremely shitty design. Also killed some LEDs hotplugging them because thats how we had to switch the lamp, we did devise a fix for that though.
So, like I said, in proper usage, no, in dumbassery, big yes.
Didn’t care to start a new thread so posted in the first appropriate thread Search showed me
An XML died on me yesterday. It was in a Convoy C8 and driven by a 2400mA driver. Technically, the LED isn’t DEAD, it lights up but at 0.00001 lumens. lol. (bright enough to distinguish on from off and no more.) It’s strange because there are no black spot on the die. Can anyone guess what went bad? Thanks.
Global warming, of course. :party:
- Double-check the bondwires under the dome to the LED. You’ll need a magnifying glass for this… the aspheric lens from a zoomie flashlight works well. It could be that one or more of the tiny bondwires wires broke.
- Another possibility is a bad solder joint. I had something like that happen very recently on a modded light. The problem turned out to be a bad solder joint under the LED when I reflowed it onto the star. You might try reflowing it to see if this fixes the problem.
- The problem could also be in the driver and not the LED. To check this, the first thing I’d do is disconnect the star and then test the LED using a single cell in direct drive from a battery carrier, with wires attached. If the LED lights up with no driver when outside the light then I know the problem isn’t in the LED or star.
Huh, what key words did you use to find this thread, I’m curious.
Did you check if there was enough thermal paste below the PCB?
What you’re describing seems like what happens if you put too much current through an LED, 6A+ and it becomes very dim. If you you increase the voltage, it does get a little bright, but not by much. Can you see if the three bond wires are intact?
The driver works. The LED has been in the Convoy for a year now with a Qlite and only used occasionally to play. I looked at the bond wires with a 45x glass but didn’t see anything strange. Will look again. WRT solder joint, to test, i connected a LiFePO4 cell to the little bits of exposed solder just outside the chip and got the same result.
I thought that if a wire has burnt out, at least part of the die would stop working?
Death of an appropriately driven LED that’s properly heatsinked drew me here.
If the led was bought preloaded it’s possible that the solder joint under the die is iffy with a partial cold joint that failed. If you haven’t reflowed an led this might be a good opportunity to try your first by removing/reinstalling the led. A broken bond wire might show as a very tiny flaw in the dome or a very tiny dark spot. If there was solder on the +/- pads but not on the center thermal pad the die would still overheat.