Eeyore said:
I remember one from a Motorola SMPS that was one of the start up resistors
(IIRC 10 - 20 K ohm, about 5W rating, about 1.25" long by 0.3", ~165 VDC max
across it.
This seems to be a common failure in some SMPS.
I've also saw it happen to the small (1/4W through hole) current limit
resistors going to optocouplers on the mains side, used for zero crossing
detection. It's bad enough with some older brand name servo amps, that I
stock spare resistors for them! I only had to refurb one amp so far, but it
lost something like 5 resistors before it quit. The one that stopped it was
for the regen resistance (to pull the power supply down to reasonable levels
from the energy dumped into it from regen), which is only used for short
periods of time, but potentially repeatedly (pick and place machine)
Electro-migration presumably.
Possible. I presume from the very localized damage that one it started, the
voltage potential / power in that area increased, driving the failure mode
faster, thus increasing the voltage differential and localized power in a
runaway fashion. Towards the end, it may have been more of an evaporation
process.
It actually looked like the resistive material corroded away in the small
area leaving little stains on the ceramic form..