Talk:List of LED failure modes

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Reorganizing LED-related pages[edit]

There are 23 different LED-related pages, and I think there should be less than half that many. I've started a discussion on this at Talk:Light-emitting_diode#An_absurd_number_of_articles_involving_light-emitting_diodes; please come and join in. -- Dan Griscom (talk) 03:07, 3 March 2009 (UTC)[reply]

Re. COD in LEDs[edit]

Hi, as I have experimentally observed the above especially in off the shelf COB and other modules:

Essentially due to the way photonic lattice LEDs are constructed a single point failure can lead to significant internal leakage and die heating: this leads very shortly afterwards to a catastrophic short circuit if mitigation is not present in the form of active cooling via Peltier module.

I accidentally duplicated this effect in a red Phlatlight which already had some damage: it can also be induced by reflecting light back into an LED at the right angle as this is why non reflective (AR) coated optics are typically used. In this case the effect was not caused by heat directly as diode temperature never went above 30c but when it broke down there was little to no warning: output power went from >11W down to less than 0.7W with 7A drive current which is not a good sign!

The same effect also happens in other high brightness emitters especially blue and UV (390nm) InGaN and GaN diodes. I can reliably induce a failure simply by running them at maximum current and "spiking" the diode with very short pulses leading to it losing most if not all of its output power; this does not happen as easily if at all with SiC based blue. Microscopic inspection of the emitter shows barely visible darkening of the surface as seen on damaged LDs which incidentally can be patched up by controlled application of reverse voltage below the zener breakdown point at <0.1A and also tested this experimentally on defunct N**h*a Bluray and BRIR PS3 diodes sourced from LaserPointerForums. Even got a dead diode with B-R low resistance to lase again with 85% of original rated power albeit not for long (3-4 days) — Preceding unsigned comment added by 185.3.100.30 (talk) 03:19, 8 March 2019 (UTC)[reply]