So what did system (event) logs say? Don't even think about
fixing anything yet. Smart technicians first collect facts.
For example, temperature is rarely a reason for failure.
Defective ICs tend to work less reliably at higher (also
called normal) temperatures. Then the shotgun repairman fixed
the symptom by installing 'more fans'.
Acceptable computer manufacturers also provide diagnostics -
for free. Best run the diagnostics with computer in a 100
degree F room (quite normal temperature for a properly working
computer) or by selectively heating computer components with a
hair dryer on high. Again, this is quite normal temperature
to every part of a good computer. But heat tends to identify
intermittent components faster.
Also necessary is information as defined in this procedure -
"I think my power supply is dead" in alt.comp.hardware on 5
Feb 2004 or at
http://tinyurl.com/yvbw9 Too many waste big
bucks and time wildly replacing things before they even know
why the intermittent happened.
These procedures assume the problem is hardware. Different
procedure for identifying a problem in software. Don't let
myth purveyors promote temperature as reason for failure. If
a computer is suffering temperature problems in a 70 degree
room, then it does not even work at 100 degrees - which is
normal temperature for any acceptable PC. Ignore any solution
that advocates 'more fans'.