MCE: 1 die is dead but how come?

[rizky_p]

This is probably the time when you hate yourself!:hairpull:
I saved some money for this MCE, now i runied it!

Ok i bought this Bare emitter from KD Bin supposed to be M but i don't have means to verify it well it doesnt matter. While reflow soldering by heating the star from underneath with solder i damaged 1 of the the emitters somehow. I have done countless soldering time with P7 and not one is damaged so i am quite confident in doing it. Not this time though.

FYI. LED are connected in parallel. If i connect the 4 legs including that connected to the damaged led all LED would not work, it is like the one that died somehow short circuited. I cant see anything wrong with it but i took a picture just so you can see it.

i'll give you the shots. One is capture behind a welding glasses the other the emitter it self.

I guess now i have a 600 Lumens Max MCE

Sorry i lost track on which one is damage
i'll take a better picture tommorow with good sunlight.

Regards
Rizki P

 

[JohnR66]

So, forward biasing the bad die draws current but no light?
It could be the protection device has failed or the LED die it self has formed a conduction channel. I can't see how this would short just from soldering.

Since you have nothing to lose, try giving the die quick blast of high current. This sometimes will open the conduction channel and allow the die to work again. Start at 2 amps. I'm not sure when the bond wires will fail, but if 5 amps don't do it, it is probably toast.

 

[rizky_p]

thanks i'll try it.

So, forward biasing the bad die draws current but no light?
It could be the protection device has failed or the LED die it self has formed a conduction channel. I can't see how this would short just from soldering.

Since you have nothing to lose, try giving the die quick blast of high current. This sometimes will open the conduction channel and allow the die to work again. Start at 2 amps. I'm not sure when the bond wires will fail, but if 5 amps don't do it, it is probably toast.

 

[rizky_p]

IT WORKS NOW!! sorry for the caps i am too happy :laughing: :twothumbs
I used a CC/CV power supply and pushed 1.4Amps on the single die, that must have been something for that poor led.

thanks man you save my day...

I test mine with power supply now it reads 3.27v@ 2.8A which is quite good.

Best Regards
Rizki P

 

[Illum]

While its unlike CREE to have a sloppy gate precision, that may very well be an issue: mismatched Vfs

notice there's 4 little black squares on the sides of the dies, those are reverse voltage protection diodes...if there was a short, it would be across one of those though I'm not sure what could go wrong in a diode

 

[bshanahan14rulz]

Ok, that's one of the coolest things I've ever seen on here! So the die itself might have had a weird area that was letting current flow through it freely? and overdriving it a bit spreads it out or something?

 

[Vesper]

Since you have nothing to lose, try giving the die quick blast of high current. This sometimes will open the conduction channel and allow the die to work again. Start at 2 amps. I'm not sure when the bond wires will fail, but if 5 amps don't do it, it is probably toast.

Nice going. That was an impressive end-around and touchdown. I'm glad it's resurrected. :thumbsup:

 

[saabluster]

Hmm an LED resurrection. JohnR66 I think people may start to think you are the Jesus of the LED world. Good work.

 

[Illum]

we'll see about that when someone does a weevil test and it came out positive

 

[bshanahan14rulz]

forgot about those, so you can kill them with high burst of current, and once you kill them, they can't eat the lumens anymore?

 

[saabluster]

forgot about those, so you can kill them with high burst of current, and once you kill them, they can't eat the lumens anymore?

So I wonder if the real reason burning in LEDs to increase lumens and decrease Vf works is that you are killing off the weevils while still in the larval stage? Get em before they hatch folks. Left untreated they can reek havoc.

 

[csshih]

nah, unless you thoroughly crisp the emitter.. there will always be larvae.

 

[Illum]

forgot about those, so you can kill them with high burst of current, and once you kill them, they can't eat the lumens anymore?

the bad part is, if the larvae has crawled between the P/N junction and jolted to a crisp you have a potential short on your hands, which may explain why its drawing current but not lit up:nana:

 

[rizky_p]

Just like JohnR66 says i have nothing to lose so i try it. I didn't know what happened but it simply works, it may not be the high current burst that resurrect the LED maybe something else but for whatever reason i am happy that now it's working.

 

[saabluster]

MOD could you please lock this thread since it is now solved.

Although solved I think it would be a good idea to keep this thread open as it may be of interest to others in the future and there would be no way for them to continue this line of thought if locked. Some info may come to light later. They would then have to open a new thread and then searching becomes more difficult and our information base more spread out.

 

[jtr1962]

This technique worked for me on a defective P7. One die was partially shorted. The LED started drawing current at a few tenths of volt, and was drawing about an amp before the other dies came on. I hit it with 9 amps. This dropped the current the defective die drew a bit. I figured with nothing to lose, why not hit it with a little more? At 11 amps it turned angry blue but I only did that for a moment. Upon turning the LED on again, it drew NO current until about 2.5 volts, and then all the dies turned on at roughly the same time. One die was slightly dimmer, but at normal current levels they all appeared about equally bright (as observed through heavily tinted plastic).

So it looks like this idea is a winner! :twothumbs It's similar to the idea of hitting NiCd cells with heavy current spikes from a capacitor to blow away internal shorts. I'm assuming though that the shorts won't return with LEDs as they sometimes do with NiCd cells.

 

[TorchBoy]

Well that was a little surprising. Well done JohnR66.

 

[rizky_p]

Good idea...i'll delete the request.

@jtr1962: nice to know someone actually done the same thing and it worked but i wasn't that brave to push that much amps since i can address the die individually unlike P7.

please no more larvae

Although solved I think it would be a good idea to keep this thread open as it may be of interest to others in the future and there would be no way for them to continue this line of thought if locked. Some info may come to light later. They would then have to open a new thread and then searching becomes more difficult and our information base more spread out.

 

[JohnR66]

Glad this worked!:twothumbs

I was experimenting with ESD damage to 5mm LEDs and found what seems to be a conduction channel form in the die of the damaged LED. This would cause the LED to conduct in both directions and not to light at lower currents or at all. Sometimes blasting the LED with high current would open the channel and allow the rest of the die to function normally. Sometimes The LED was not recoverable. Perhaps the large area of power LEDs allows a better chance to recover it.

I'm not saying conduction channels only form from ESD. Sometimes channels formed in the 5mm LEDs after my 192 hour 30ma fade tests. And these were the super cheap ebay LEDs.

 

[mpteach]

This reminds me of defibrillator