Rather interestingly, and rather suddenly, I came home to find one of my PAP2T ATA’s that served the home phone line to be offline. A closer inspection showed no LEDs were lit, and a quick exchange of power adapters proved that it had failed. Luckily, the ATA still remained functional.
The adapter itself is dated to late 2007 or early 2008 when I bought it, and is a Linksys/Cisco AD 5V/2F. The adapter itself has a model number of PSM11R-050 and is manufactured by Phihong (Dongguan) Electronics Co. Ltd. It only has an efficiency level of III.
The adapter itself has a sliding plug adapter plate, and is otherwise fully sealed together. Opening it up was quite difficult, and prying with a screwdriver got me nowhere as the plastic was brittle and liked to crack. Instead, I had to resort to using a hot knife (soldering iron with a blade attachment) to cut through the seam to open it up and see what was going on inside.
Internally, the electronics were all nicely bundled together, with copious amounts of silicone rubber glue holding parts in place, and insulating mylar tape over transformers and heatsinks.
Looking at the unit from above, it seems like it’s a little bit of a mixed kettle of fish. There are recognizable good-quality Nippon Chemi-Con capacitors on the output and a mix of other off-brand units elsewhere.
The underside seems to look as expected with a clear separation between LV and HV. It seems there is evidence of a little accumulated heat in the browned solder, but otherwise, it is clean and the underside components seem well mounted. There is no cabled connection to the mains as it comes in via two spring contacts on the rear of the case.
The view from this side shows that the product is indeed a Phihong product, and the off-brand capacitors are Ltec … eugh. The transistor is mounted under a heatsink … and it seems to have violently let go.
It must have been over-stressed, internally arced, which caused a build up of hot molten silicon to burst out of a crack in the package and melt the heat-shrink around the capacitor.
As usual, it seems the culprit is a bad primary side capacitor. This Ltec capacitor was ever so slightly bulged with no signs of venting. It was very suspicious, so I measured its capacitance on the LCR meter.
That’s right. It measured just 49.56nF when it should be 22uF, so it only had about 0.225% of the capacitance it should have. At least it didn’t fail internally shorted, otherwise it would have been a hot goopy mess, but it seems the increased ripple on the high voltage side had stressed the transistor into failure. While the capacitor was failing, it was likely that the power adapter ran hotter and less efficiently than initially designed for.
That being said, it served about 8 to 9 years, which is adequate for consumer electronics. That represents about 74,500 hours. Assuming this was a 105 degrees C rated capacitor, and the ambient temperature is 55 degrees C (50 degrees C margin), it would have had to have a 2,328 hour lifetime at 105 degrees C rating. However, our definition of failure was when the power adapter failed, rather than when the capacitor drifted out of the tolerance (20%) margin, so chances are, this capacitor had failed a long time earlier than that.
The unit was rendered safe with a T1A potted fuse which opened as designed.
The cause of failure of this power supply is now known, and it seems the PAP2T attached to it escaped unscathed. As I do have another one or two of these supplies, which are likely to also fail in a similar way, I have a difficult choice of either replacing the capacitor pre-emptively, or letting them fail and hoping that they fail safely. As the case is not easy to cleanly open, I think I will have to just let them fail and replace them as necessary, but it would be nice if the case was designed to be serviceable in some way.