Computer won't stay on...

P

PlainDave

Hi,

I usually have my pc on 24/7, but this past weekend I turned it off for
a little over 24 hours. When I tried to turn it back on, it booted up
and stayed on for maybe 2-3 minutes, then abruptly turned off almost as
though it lost power. So, I tried turning it on again, but it would
only stay on for about 1 second.

I immediately thought it might be the power supply, but the green LED
that indicates "power" stayed on. So, I found out that if I let it sit
for an hour or so, it will turn back on for that 2-3 minute time period
and then shut off.

Any ideas?

Thanks,
PlainDave
 
G

Guest

Us the search for option above and see my post "Computer Powers Off Randomly"
in Windows XP Hardware section.
 
W

w_tom

Problem could be anything in a power supply 'system'. A 'system';
not just a power supply. Green light will illuminate even when
voltages are defective. Two standard tools for fixing PC are the screw
driver and a 3.5 digit multimeter. In but minutes - many times faster
than any power supply 'shotgunning' - we can determine what is good and
what is suspect.

First measurement is voltage on purple wire from power supply to
motherboard by pushing meter probe into nylon connector body. That
voltage must measure 4.87 VDC even when computer is off but AC
receptacle connected. Next is green wire. It must measure more than 2
volts when off, then drop and remain at below 0.8 volts DC when power
switch is pressed. Next is gray wire. It must rise to well above 2.4
volts within seconds of power switch press.

I would bet the gray wire is your first problem detected when system
cuts out.

Next measure one of orange, red, and yellow wires as power switch is
pressed. Each must rise up and remain above 3.23, 4.87, and 11.7
volts. I suspect one is either low or is dropping out. Finally these
voltages must remain above those limits when everything (hard drive,
network, DVD player, heavy graphics on video, etc) are used
simultaneously.

If all numbers are in spec, then move on to other suspects. If those
numbers are posted here, other useful facts may be gleaned. One does
not start replacing anything until the suspect is first identified. The
meter means conclusions in but minutes.
 
P

PlainDave

Fan that cools your CPU working?

Yeah, I checked that, as well.

Thanks to you guys for the replies and the info. I think I've narrowed
it down to the power supply.

Thanks again,
PlainDave
 
P

PlainDave

Fan that cools your CPU working?

Yeah, I checked that, as well.

Thanks to you guys for the replies and the info. I think I've narrowed
it down to the power supply.

Thanks again,
PlainDave
 
G

Guest

A volt meter will tell you if voltages are out of range or missing. If a
power supply is intermittent it may not show up with a simple voltage check.
It only takes a drop for a millisecond or two on critical rails to initiate a
system shut down.
 
W

w_tom

A volt meter will tell you if voltages are out of range or missing. If a
power supply is intermittent it may not show up with a simple voltage
check. It only takes a drop for a millisecond or two on critical rails to
initiate a system shut down.

A power supply that is intermittent often has low voltage constantly
when under load. Remember, the world is not binary. The world is
ternary. Either voltages are good, so bad that a power supply does
not work, or in a mid-range where things such as intermittents occur.
Too many - especially shotgunners - assume a binary world.

Most computers will work most of the time when voltages are in that
mid-range - too low but just enough to work only most of the time.
These computers intermittent. We identify an intermittent before it
fails by measuring those voltages.

PlainDave describes a problem often seen when voltages were too low
months previously. However, the same problem can be due to other
problems. First establish what is working. Without that meter, we
are only guessing that a power supply 'system' (a system - not just a
power supply) is working. With a 3.5 digit multimeter when power
supply is under load, confirm a power supply 'system' is working.
Then move on to other suspects.

Let's assume a new power supply 'fixed' the problem. Maybe. First
do the same voltage measurements. That new supply could be
intermittent - again using ternary logic. Confirm what is observed.
We put that new power supply under maximum load and measure those
voltages using numbers from above previous post. New power supply is
working only after those voltage numbers exceed minimum values (and
other considerations). It's a ternary world.

How to detect intermittents before failures happen? A 3.5 digit
multimeter and an informed human. Just an example of how to work
smarter; not harder.
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Top