Power connectors can cause failure

[Sage]

Visited by Mr Sod and his Law, this morning. One of the optical drives didn't show in BIOS, and, after bypassing it, the OS stopped booting midway. This is not the first time, indeed, it is reasonably frequent on long service machines. Why should the Molex socket pins spread when they remain stationary (I can understand the phenomenon on my test rigs, which are connected and uncoupled several times a day!)?
For the inexperienced, this effect could be most perplexing. Wondering whether Mr Fleming's left hand has been busy!!! All the forces would be in the correct direction to cause spread and the current on the 5V line is not insubstantial. Any other comments/observations?

[fluxit]

Corrosion of the sort that is invisible to the naked eye is a possibility. What about uneven heat expansion?
The metal of the contacts in some molex connectors is of a higher gage(and quality) than it is in others. Many times the cheaper connectors really need to be "re-crimped" after only a single insertion/removal.

This problem can lead to excess anxiety for sure as no power connection to a drive can lead to not only a failed boot or unreliable operation of the affected drive, but a machine that will not even power up.

[Sage]

Yes, I suppose corrosion could be an issue, especially the nearer you get to the Gulf! Oxidation is a well-established issue on RAM strips. Thermal expansion/contraction could also cause problems. Strangely, I've not encountered contacting issues on any of the hundreds of cheap PSU that have passed this way, although it has happened on a couple of very expensive, proprietary units with heavier gauge pins!!! The offending contacts encountered this morning were on an extension piece.

[marksouth2000]

I suppose corrosion could be an issue....

Some gremlins are badly behaved and don't use the spitoons provided inside computer cases. Really, the only solution is to leave small editions of etiquette guides inside the case for them to read.