Archive for September 18th, 2018
A recent Squidwrench meeting produced a treasure trove of discarded LED lighting, including a shoplight-style fixture in a narrow, finned aluminum extrusion. It was in “known-bad” condition, so I extracted the four LED panels, connected each one to a widowmaker cord, and determined I had two good ones, a mostly working one sporting some dead LEDs, and a corpse.
The working panels showed the power supplies produced about 19 V across two parallel strings of six LEDs, with each string running at 350 mA for a total of 700 mA = 13 W. I wired up a quartet of 6 Ω power resistors to check out the power supplies from the suspect panels:
The supply in the background is truly dead. I can’t tell whether it killed the LEDs or the gaggle of failing LEDs dragged it down with them.
Some multimeter probing revealed enough live LEDs to restore the partially working panel. A rather sweaty interlude at the SqWr hot-air rework station transplanted the good LEDs, whereupon combining it with the live supply gave me a third fully functional panel:
I did the test firing in the Basement Laboratory, because I’m nowhere near crazy enough to deploy a widowmaker line cord on the SqWr Operating Table in public.
I bandsawed the last working LED from the gutted donor panel:
The SMD LEDs mount on traces applied to and electrically insulated from the aluminum sheet, so unsoldering them required way more heat than you (well, I) might expect at first glance. A snap-on condenser lens over each LED concentrates the light into a nice cone, producing a narrow sheet of light from each panel.
The elaborate aluminum extrusion seems much too heavy for the individual panels, but those open-frame supplies definitely need more than casual protection. Now that LEDs are more common than when these panels came off the assembly line, I should probably replace the supplies with enclosed constant-current drivers and be done with it.