something killed my LuxV...

I have a Mag modded with V4T/DD. It was on its third 2xPila150s cycle.

Last night, we had a relatively long power outage at home. Being the most 'lumen'ous light I have, it was my preferred stationary light. I turned it and left it with beam bouncing on the ceiling at my desk while my boy was eating his supper. It maybe about thirty minutes later when I check it out and it very hot, almost scalding temperature. I turned it off to let it cool down. About an hour later, I turned it back on again and was ok. After about 3 minutes runtime, I noticed flickering and reduced output. I thought that Pila was draining out and is asking for recharge.

I charged the Pilas one by one when power came back on at around 9pm.

This morning, I powered the Mag back on and notice the same flickering. I noticed when I did a beam test, that one section was missing. Using almost-drained 3x123, I saw the problem. One of the 4x4 die is not lighting up (dead). This is very evident on the beam shots.

Head focus fully twisted down{left} || focus turned back about 2x {right}
|

According to the previous owner of this Mag, the LuxV was drawing about 1A. Is it possible that too much heat caused one of the die to give up???

Sounds like you had a thermal runaway situation. Is that LuxV resistored or is it a pure DD?

Just speculation - but direct drive (with no resistor) is a kind of classic way to kill an LED. I have personal experience in this area /ubbthreads/images/graemlins/frown.gif - mostly on Lux IIIs.

The combination of high current flow and high temperature are the almost perfect storm for killing a Luxeon, or any other power LED for that matter.

Can you check the current now with a DMM? How about the voltage across the LED?

It's possible that the Vf dropped, causing current to rise. The excessive heat may have caused damage to one of the dies on the emitter.

A Luxeon V is 4 Luxeon dies in a series/parallel arrangement. It sounds like that die failed open circuit - if it failed as a short, the other good die wouldn't be lighting up (so 1/2 of the Luxeon would be dead).

What this means is that the remaining die in the series direction has all of the current flowing through it, instead of current being divided between it and its parallel counterpart. Expect that one to fail in short order as well.

Of course, since the one die has twice the current flowing through it, its Vf will be higher, so the Vf of the whole LED will be higher than right before the one die expired. Thus, current may be lower as well, and it may be difficult to asertain the exact cause of failure. Given that the light was scalding hot, the junction temperature was at least 50-60C higher than the body of the light. Its very likely that excessive heat caused the demise of the darkened die.

If Vf dropped and caused excessive current to damage the dead die, it's one more reason to avoid DD if possible.

I got this light about half a year ago.

Blinded> I don't really know if this is resistored or without one.

Evan9162> I'm electronically challenged. I don't have a DMM but an analog one that's only up to 250ma. The light is output is way reduced, much like a low output of a 2-stage light. It's flickering. I think you maybe right, EXCESSIVE HEAT was the main culprit. My other suspect was that the Pilas was putting out much more consistent high current. This light was said to have been designed for 3x123, but I thought it really doesn't matter, as I was playing with the voltage numbers /ubbthreads/images/graemlins/frown.gif

akula,
The internal resistance of the 3x123 is a saving grace with the LuxV as the current is limited. Putting it on 2x Li-Ion would allow the LED to pull all the current it chose to! /ubbthreads/images/graemlins/icon15.gif

I agree it's the li-on that killed it. It overdrived it probably well beyond 1.5amps and it got hot just sitting around, and thermal run-away, kablewy!!

If it got super hot, then it was being over-driven too hard!! To heat up a massive mag body to scalding hot means it must have been really really overdriven!!

/ubbthreads/images/graemlins/thanks.gif guys,

Lesson learned... /ubbthreads/images/graemlins/frown.gif
No more Pila or LiIon on DDs. For now, all things must go regulated or at least with a resistor.

LARGE Mag got scalding hot from one emitter? That's SERIOUS. Even my Mag3X driven at full power warms up, stays warm for its 2 hour run, then cools down towards the end again without becoming that hot. Smells like thermal runaway.

Proper design and understanding of the Luxeon can combat that, but you're right, regulated solutions offer the most control. DD is not without risk. But it sure is bright /ubbthreads/images/graemlins/grin.gif

Sorry for your mishap. I agree with Harry, this is a clasic way to kill a good LED. Direct Drive is the problem, IMO. It's just not a good way to go.

I'm also taken about loosing *one* die of the four. I thought we had two parallel pairs of series connected dies, how can one go open and the other still light? Or if it goes to short, how does the other one survive? How does the failed die still have the correct Vf? What does it look like, is it a phosphor issue?

Doug Owen

Doug, I've known of another case of this happening. Someone reported years ago about a 5W cyan and one die was not lighting up. This was an EL modded light, and Wayne made it right though!

This light must have been overdriven very hard. I'm surprised that the bond wires didn't melt!!!

Thinking about this is making my head hurt.
I'd say the dead die is dead short, and it's partner is getting 700ma, no matter which way it's wired.
If it was open circuit, then 2 would be out, or all of them depending on wiring...

Edit: I think in the 2nd pic the upper right die looks brighter. Can this be confirmed? Perhaps put run-down batteries to make it barely light up and examine the die directly would be a good experiment.

I'm pretty sure it's a series-parallel, but the "middle" of the circuit (where the anode of one die meets the cathode of the other), all 4 dies are connected to one common electrical plane. Somone once posted pictures of a Lux V with the die ripped off the silicon submount, and there were only 3 electrical pads on the submount, instead of 4 that you would expect for an "isolated" series-parallel.

Thus, if the failed LED is shorted, its parallel partner would be dark. However, if it went open circuit, its parallel partner would be getting all of the current instead of half (and would be brighter than the remaining dies)