Quality control, or lack thereof...

[TheMG]

Thought some of you would find this interesting.

First of all, soldering quality that looks like it was done by a 5th grader (actually, I'm pretty sure with a little practice, a 5th grader could do better). Yes, this power supply did let out the magic smoke, and I wouldn't be the least bit surprised if a dry solder joint caused it. "Dry as a dead dingo's donger" as Dave would put it!

Secondly, the OEM for this power supply, U-San Electron Co Ltd, can't even spell their own name properly: U-SAN ECECTRON silkscreened on the PCB!

Lastly, check out R16. It's a 2W resistor, not the flameproof type, across the 48V output (minimum load resistor). Color bands have changed color but it is indeed a 1k resistor. If we do the math, that's 2.3W of power dissipation. FAIL!!!! The placement of this resistor next to the output filter caps is also not a great design decision.

The most surprising thing is, this is not a cheap consumer product at all, it is in fact enterprise level telco gear. Power supply for an HDSL (T1) line card cage. It's supposed to be high-reliability stuff.

[Hornethandler]

Pretty scary looking soldering.  I was a NASA quality inspector for electronic assembles and the workmanship on that board gives me the chills.

If an operator turned in something like that for in-process or final inspection they would immediately be sent to a “re-education camp” for retraining and an attitude adjustment.

[TimFox]

Perhaps you did not follow the instructions on the silk screen starting "FOR CONYINUED PROTECTINO..."

[SeanB]

Probably an outsourced power supply, and the first batch worked as designed, so the manufacturer decided that they will cost optimise the design a little with each subsequent order, and thus some enterprising person figured out that simply by getting a cloned version made he could cut the cost by half in procuring it.