Help! Sudden Computer Death!

[X-ray Doc]

My ultimate gaming computer suddenly died last night. I
built it several months ago and it has worked flawlessly,
except for about three spontaneous reboots while gaming
and one VPU recovery screen from my Radeon 9800 A-I-W
Pro. (I think the VPU episode might have been related to
the Catalyst 3.9 drivers and the game Call of Duty. This
is suppose to be fixed in the 3.10 drivers.) Anyway,
here are the bloody details. My USB wireless Joystick
stops working properly two days ago. I assumed it was
the joystick. Then yesterday, while trying to play
Battlefield 1942, my Microsoft USB Explorer mouse started
mapping keys incorrectly. (In game, I'd press the left
mouse button to map a control and it would be listed as
Joystick 17.) Shortly after this, the computer starts
to reboot spontaneously and then it emits a loud, fast
continuous beep. It no longer boots, or at least there
is no video signal at all. Here are my thoughts. Please
help if you can. My Gigabyte manual says that short
continuous beeps may indicate a power problem. I doubt
this because it is a high quality 550 watt PSU and all my
fans and drive lights work. The AWARD BIOS site said
that short continuous beeps may represent a CPU problem.
The fast beeps only occured during the computer's actual
death. It is silent now. I'm wondering if the CPU has
been overheating, causing the reboots, and is now fried.
Or did my video card somehow just die? Or did the USB
problems indicate a motherboard failure? Please help.
My life, (or at least part of it) is on that computer!
Thanks.

 

[Murray McNeill]

I would definitely suspect the CPU. The quickest way to check it is to
borrow a working CPU, and change it to see if the motherboard boots again.
The symptoms you describe are 'across the system,' so to speak, and indicate
a central issue. I doubt if it's the power supply, but IF you get it booted
again with another CPU, I would use the BIOS hardware monitor to see if all
voltages are within range.

 

[w_tom]

Your symptoms are more consistent with static electric shock
damage some months ago. A slow death from overstress by
static electricity.

Until voltage readings are taken with a 3.5 digit
multimeter, then no idea of that power supply quality.
Furthermore, if supply did not come with a long list of
numerical specifications, then it is likely not a high quality
supply. High quality supplies list at about $80 retail. A few
minutes with procedure listed in "Computer doesnt start at
all" at http://tinyurl.com/2t69q would provide more
important facts.

Of course, if this is a Windows NT based OS, then the system
keeps failure information in system (event) logs. What do
those logs say?

 

[Guest]

I have Windows XP. I paid $294.00 for what I thought was
a top of the line PSU, the PC Power and Cooling Turbo-
cool 510 Deluxe ATX. It has power conditioning and
Maxximum PC magazine always uses their power supplies in
their ultimate computers. I was very careful avoiding
static electricity when I built the computer. I didn't
wear a ground strap, but I grounded myself on the power
supply before touching any components. When you suspect
this damage, are you referring to the motherboard being
damaged, or what component in particular?

 

[w_tom]

Do as I didn't do. Was doing the same thing (grounding
human skin to chassis) recently. But then moved a shoe as I
went to change something. Did not 'static wrist strap' ground
myself to the case (not earth ground - the case). The
resulting static shock, I believe, entered via network cable.
Rather surprised me that it transversed NIC isolation
transformer with enough force to immediately damage input to
National CP83223 line transceiver. But then the shock so
massive that I even felt it. Static electric is that
destructive.

In another location, they did not bother installed a
humidifier. They have static electricity constantly. But
since damage is not immediately apparent, bean counter
mentality cannot associate cost reduction with cost of
humidifier. In this case, the USB port on a digital camera
failed. Virtually every USB camera connection to that
computer results in a static shock.

It's difficult to correlate static damage with a particular
event. But if a human ever shock equipment - even once - then
the many more 'less detected' static electric discharges may
be causing electronic failure later.

BTW, how to test that a computer system is properly
constructed? With the system buttoned up and on a glass table
(wood and other tables are too conductive), static shock every
corner of chassis, keyboard, etc. Static shocks so strong as
to be felt should not even cause an Operating System crash.
Any software crash due to static electric? Start with how
motherboard ground connects to chassis ground. Only one
conductive standoff for a single point ground. Static
electricity is also a good tool for testing system
construction.

 

[Guest]

I used a voltmeter and checked the 3.3V, 5V and 12V
connections to the motherboard. I did this while the
motherboard was connected so there should have been load
on the power supply. All of the voltages are within
spec. I did not check every power cable coming out of
the power supply, but I doubt the power supply is the
problem. I'm thinking either motherboard or CPU.
Unfortunately, I don't have spare components of either of
these to try a swap.

 

[w_tom]

Those voltage reading eliminate power supply as reason for
reboots. Now move on to other reasons. Don't recall whether
you said this was an NT or 9x based OS. If 9x based, then
virtually everything in the machine is suspect. If NT based,
then reason for reboot would be limited to video card, sound
card, memory, CPU, and associated software drivers.

Also if NT based, then the system keeps a system log of
events. What do they say?

Responsible computer manufacturers provide, for free, a
complete set of diagnostics. What did they say? Heat is also
a tool to find problems. For example, while running a memory
diagnostic, then heat memory chips with hairdryer on high.
This is normal operating temperature to memory. Any memory
that fails diagnostic while heated is defective.

Same heat test can be applied to video controller while
running video diagnostics.

Just some ideas on how to find a hardware based reason for
shutdown. But your shutdowns may be software created. Best
we can do is trace problem to software by first eliminating
hardware as reason for failure.