Fowler Associates Labs

Response on CPU's from Jeremy Smallwood
http://www.static-sol.com/

This is a very interesting question.

My first thought - as you mention storms - is, could the failure be related to lightning? If there is a close correlation between occurrence of lightning and failure, then this may be the answer. Lightning could induce large electrical transients on power or data cables. If such a transient could get onto a cable running into the cpu board, this could cause failure. Be particularly suspicious of long cable runs which act as good antennas.