EEVBlog for Power Supply Health?

[Xebec]

Hi folks - I've been watching EEVBlog occasionally for a little while and love the format/information.

I was curious if there is an episode or two folks would recommend to a beginner that would help me determine the relative health of a small power supply, and maybe understand it's likelihood of failure over time?   

I understand this is probably as much a design question as a health of individual parts question.  I'm asking as I (and others) are starting to use modified PC power supplies and the like for things such as powering vintage computers, and I'd like to make sure I don't end up selecting a power supply that will kill the much harder to replace 'old hardware'.

Thanks!
John H

[3roomlab]

im not pro, but the best indicator is operating temperature of critical components. hot things tend to break prematurely/faster than the well cooled. it sometimes helps if you can "FLIR" it.

[Ian.M]

Also, if the hardware you are powering is high value, learn about crowbar circuits.   http://axotron.se/index_en.php?page=26

[DanielS]

Also, if the hardware you are powering is high value, learn about crowbar circuits.   http://axotron.se/index_en.php?page=26

Assuming your load does not mind its input voltage suddenly getting yanked to ground. How many linear regulators and opamps have died from their power supply getting turned off or shorted with a significant capacitive load attached back-feeding power through the device's output?

[Ian.M]

Good designers allow for the possibility of the supply being shut off rapidly e.g. by using reverse diodes across linear regulators.   If the high value hardware was not well designed in this respect, series Schottky power diodes between the crowbar circuit and the load can be used, or -worst case- if their Vf drop isn't tolerable, it may need a protection circuit using high-side MOSFET load switches rather than a smple crowbar circuit.

Even so, without such protection, the damage due to crowbar activation is likely to be far less than that caused by a SMPSU output rising due to catastrophic loss of regulation.