White LED Failures

Well, that didn’t take long:

The two dim LEDs to the left are actually very faintly lit, so I think the dark one has failed nearly open.

When I installed those nine central LEDs, I didn’t notice that the bag (from the usual eBay source, IIRC) contained two different types of white LEDs. The difference shows up clearly under UV that lights up the yellow phosphor:

By random chance, each of the three groups has one non-fluorescing LED. If I can extricate them from their epoxy tomb, maybe I can figure out which one failed.

Rather than replace those, I’ll try a new-fangled chip-on-board light source, even though that might require a current limiter and maybe a heatsink. Obviously, this is getting out of hand, but maybe the same folks who can’t make a white LED can make a functional COB assembly for a buck… [sigh]