Overheating CPU? That's a thought.
Well, I guess my system is a year old and could use a bit of an
overhaul. Thank you very much for the help!
-- Darren
James Cameron wrote:
On Sat, May 23, 2009 at 09:50:45PM -0400, Darren
Landrum wrote:
Well, I decided to run memtest86, and it locked
up during the test.
Ah, good, that reduces the component set to something quite smaller.
The things I would consider as cause for that are:
1. dust in the CPU cooling fins, (I take my systems outside and run a
vacuum cleaning in reverse, with a thin nozzle, and play the air stream
over the various parts of the heatsink),
2. non-rotation or slow rotation of the cooling fans, (if the system
has a BIOS sensor display, check that it shows a reasonable rotation of
the fan, typical rotation rates are from 1000 to 3000 RPM, in my
experience, and is fixable by replacing the fan, or cleaning it)
3. drying out of the thermal conducting grease between the CPU and the
heatsink, (I recently had to remove and reapply the grease on a Pentium
4 3GHz desktop at home, symptom was CPU temperature consistently high
and random power downs),
4. failing power supply, (I unplug non-essential devices temporarily,
such as hard drive, to lower the average power draw, and see if the
memtest symptom goes away ... I also check the power supply voltages
with a meter),
5. corrosion or other damage to the memory DIMM socket or module, (I
wiggle the DIMMs during a memtest, with about the equivalent of up to
200 gram force ... if the memtest result changes in a cycle with my
wiggling, I know there's damage),
6. a specific memory DIMM failed, (remove it, see if memtest
completes).
Oh, and above all remember to use anti-static procedures, and try not to
unplug or replug things inside the unit while the power is on.
Static discharge damage is particularly annoying because it typically
happens months after the static discharge happens. The discharge causes
damage which then takes a long time before it begins to make the
component fail.
So "it works after I zapped it" isn't a reliable method of proving no
damage was done.