Random power shut offs (what from?)

  • Thread starter Thread starter SP2captive
  • Start date Start date
S

SP2captive

6 months after building my own computer with an Asus motherboard (P5P800)
and an Intel Pentium 4 processor with hyperthreading, I've been getting
random power outages.

The trouble started last month after I tried installing an old burner into
my computer as a slave to my CD/DVD player. I could never get it to work in
there, so I didn't use it. Then the first of these power shut offs occurred.
I was able to power up right away, but from then on out this became a twice
a week occurrence. As of last Wednesday, its been happening twice a day or
more.

This usually happens when I either have a p2p program running, or I'm
converting files though Musicmatch Jukebox. But just 30 minutes ago it
happened when all I had on was winamp while surfing myspace.

When my computer starts, I get this message before Windows XP loads that
says "CPU Fan error! Press F1 to continue". Sometimes my fan is more noisy
than others (?) Sometimes, if I try to start back up immediately after one
of these shut offs, it shuts back off just as the Windows black screen
starts loading.

What I'm wondering is whether its my Pentium processor that's doing this, or
if it's my Asus motherboard? I'm also wondering if its the power supply in
my case? There have been some power blackouts in my building, and I'm
limited to one power outlet for all my computer components (in a very old
building.)

Sorry for the length of this post, I'm just trying to get as many bases
covered as possible.:)

Thanks in advance,
TS
 
Start by eliminating suspects. For example, if a power
supply does not have the ratings claimed, then it will be
obvious using a multimeter. Some procedures to study a power
supply are provided in two previous posts. In the second post
group is a chart with voltages for each wire. Most important
in your case are voltages on the red, orange, yellow, and
purple wires. Measured voltages must be in the upper 3/4 of
these limits.

Perform these reading when the computer has started. Then
get the computer to access all peripherals simultaneously.
For example, play music on sound card, while a Command Prompt
task constantly reads directory from hard drive; another
commmand prompt reads from diskette drive; a third reads from
a network drive on another computer. Another program accesses
the CD-Rom. And you access the internet. Now power supply is
under full load. What are those same voltages again? Are
they still in spec?

Post those numbers. The values may provide further
information. The procedure demands a 3.5 digit multimeter.
You cannot use the on-board voltage monitor which is only a
monitor; not yet calibrated.

Once those numbers are determined 'in spec', then the power
supply 'system' (a 'system' and not just a power supply) is
eliminated as a suspect.

Then move on to other diagnostics. In your case, the
manufacturer (you) does not provide them. So you must
download component manufacturer and third party diagnostics.
If using XP, then more likely suspects are video card, sound
card, memory, and CPU diagnostics. For example Memtst86 or
Docmem can be used to test memory. The memory test is
performed at normal temperature; then repeated with memory
heated by a hair dryer on high. A hair dryer can never
overheat any electronics inside a computer. Memory is best
tested hot enough to be uncomfortable to touch; but does not
leave skin. Intermittents are more often found when memory is
at a perfectly good temperature - heated by a hair dryer on
high.

As diagnostics check out, then move on to other suspects.

Once you have eliminated hardware, then move on to software
suspects such as component drivers.

Process to find an intermittent is to eliminate the suspects
one by one. First verify hardware. Then move on to verify
software.
 
Does the message, "CPU Fan error! Press F1 to continue" that you keep
seeing, not indicate to you what the possible problem is? Monitor the CPU
temperature to confirm the message. CPU could be overheating and shutting
down to protect itself. Even your statement, "Sometimes my fan is more
noisy than others", seems to point to this as the problem.

The noisy fan is caused by the lack of bearing lubrication. You could try
to dismount the CPU fan it and clean the dust out of it and lubricate the
bearings and then reinstall it. If that fails replace the CPU fan. The
replacement fan must have the third wire for fan speed detection.

Your statement, "Sometimes, if I try to start back up immediately after one
of these shut offs, it shuts back off just as the Windows black screen
starts loading.", indicates that the CPU is still too hot to be stable
enough to load the operating system.
 
w_tom said:
Start by eliminating suspects. For example, if a power
supply does not have the ratings claimed, then it will be
obvious using a multimeter. Some procedures to study a power
supply are provided in two previous posts.
Click to expand...

Thanks for the info. Could you give me the link to that post?
 
Homer J. Simpson said:
Does the message, "CPU Fan error! Press F1 to continue" that you keep
seeing, not indicate to you what the possible problem is? Monitor the CPU
temperature to confirm the message. CPU could be overheating and shutting
down to protect itself. Even your statement, "Sometimes my fan is more
noisy than others", seems to point to this as the problem.
Click to expand...

Thanks, the CPU does seems to be the likely suspect. How do I monitor the
CPU temperature and fan speed while in windows?
 
SP2captive said:
Thanks, the CPU does seems to be the likely suspect. How do I monitor the
CPU temperature and fan speed while in windows?
Click to expand...
You can use either ASUS Probe or Speedfan. Motherboard Monitor will
work for everything except the voltages with this motherboard. I
currently have the same motherboard and am using Speedfan as it lets me
display the CPU temperature in the taskbar.
 
Michael W. Ryder said:
You can use either ASUS Probe or Speedfan. Motherboard Monitor will work
for everything except the voltages with this motherboard. I currently
have the same motherboard and am using Speedfan as it lets me display the
CPU temperature in the taskbar.
Click to expand...

Thanks. Asus Probe informed me that my CPU temperature has gone from 80 to
90c, when the threshold is at 72c. Meanwhile, my CPU fan, power fan and
chasis fan all have a value of 0.:(
 
SP2captive said:
Thanks. Asus Probe informed me that my CPU temperature has gone from 80 to
90c, when the threshold is at 72c. Meanwhile, my CPU fan, power fan and
chasis fan all have a value of 0.:(
Click to expand...

Are the readings for the fans "enabled"? There is a place in both the
Bios and Probe to not monitor them. If your temperatures are as high as
you say they are then the computer will definitely power off. The
problem could be that the heat sink is not properly attached. I had a
hard time getting all the pins tight at the same time. Another problem
I ran into that raised my temperatures was the lack of clearance between
the heat sink and the power supply. A new case solved that problem and
dropped my temperatures over 10 degrees Centigrade.

 
"SP2captive" said:
Thanks. Asus Probe informed me that my CPU temperature has gone from 80 to
90c, when the threshold is at 72c. Meanwhile, my CPU fan, power fan and
chasis fan all have a value of 0.:(
Click to expand...

Take the side cover off the computer case and verify the fan is
spinning. When the fan makes that bearing noise, it won't be
able to spin at its normal speed.

Did you use thermal paste between the heatsink and the CPU ?
If there is no thermal interface material, that would help.
More than likely, the heatsink is not properly clamped to
the surface of the CPU.

This is one of the reasons I like to install the heatsink
while the motherboard is outside the case (assuming there is
room to fit it when fully assembled). With the motherboard
outside the case, it gives you an opportunity to tell whether
the assembly is sitting flat or not, or whether it is hanging
up on something.

I had an old motherboard, where the heatsink was resting on
the lever for the socket. The heatsink was not milled
correctly, and it took 30 minutes with a file to fix it.
That would have been enough to make the CPU overheat.

CPUs can also overheat for other reasons - excessive Vcore
voltage, too high a clock speed, a short inside the CPU,
and so on.

Paul
 
Previous discussions were in two posts; chart is in the
second group:
"Computer doesnt start at all" in alt.comp.hardware on 10
Jan 2004 at
http://tinyurl.com/2t69q and
"I think my power supply is dead" in alt.comp.hardware on 5
Feb 2004 at http://tinyurl.com/yvbw9 .

However if your CPU is really going to 90 degree C, then
the CPU is shutting down to protect itself. This voltage
stuff performed only to confirm power supply system integrity
and to calibrate motherboard voltage monitor since the CPU
temperature is clearly the #1 suspect. If CPU was an AMD at
those temperatures, then you would probably have CPU and maybe
motherboard damage.

For CPU to have that large temperature, the heatsink would
need be improperly mounted to CPU or the heatsink fan is
impeding airflow. If the first, then heatsink would feel
cool. If the second, then heatsink also would feel hot.
 
Back
Top