Power problem

Do you have a question? Post it now! No Registration Necessary.  Now with pictures!

Threaded View
After running fine for years, my Pentium 4 PC would not power up the other
When I hit the on switch nothing happened - no fan noise, no beep, non
lights, nothing.
After determining the outlet was  providing juice, I replaced the power
supply and it ran fine for two days.
Then once again it would not power up.
I disconnected cables connecting the the power supply to the motherboard,
then reconnected them and, voila, the PC started up once more.
It ran for another day, and now it's dead again?
Any thoughts on what is going on here?

Posted via a free Usenet account from http://www.teranews.com

Re: Power problem

Quoted text here. Click to load it

Check the battery on the mainboard.

-- Bob Day

Re: Power problem

On Sat, 21 Apr 2007 10:44:20 -0400, "Tom Hill"

Quoted text here. Click to load it

Did you watch the fans very carefully at the moment of
hitting the power button, just in case they had stuttered a
bit?  It would be a sign the PSU had tried to turn on but
then quickly off again because it couldn't keep the power

If you have a multimeter, check voltages on the PSU,
including while plugged into the system measure the 5VSB and
PS-On pins.  Also measure PS-On pin while momentarily
pressing the power button to confirm it pulls down the
voltage.  Following is a diagram of the ATX connector.

Quoted text here. Click to load it

Could you provide a concise description of all major
components in this system so we can have an idea of the
power load, including make & model of motherboard and power

Details on both the old and new PSU might help, including
the 12V current ratings of both.

Had you adding anything to the system, even an external
peripheral like something USB, or moved the system just
prior to these problems?

Quoted text here. Click to load it

One of the typical things to check for is vented capacitors
on the motherboard, particularly the larger ones around the
CPU socket area.

If the power supply has marginal capacity, including cases
where the manufacturer overrated it (as is often the case
with generics), that alone could cause this... that your
original PSU might have failed and the replacement merely
isn't up to the job of powering the system.

One thought is the motherboard is dying, perhaps a cold
solder joint has taken awhile to be a problem and the slight
mechanical stress of plugging/unplugging the system provides
slight flexure and a temporary contact on intermittent
joints, but when turned off the thermal cycle and cooldown
degrades the joint again.  

If all else fails, begin stripping the system down to bare
essentials, leaving only essential parts (first, only enough
to POST, then if that succeeds reliably, add back enough to
run the operating system).  If you have another PSU you
might try that as well.  Certain parts are more failure
prone than others for various reasons like hard drive or
video card (after motherboard) and you might substitute
spares for these if you have any.

Check all the basic things as well like that no fans are
siezed up and the battery hasn't ran down.

Re: Power problem

Quoted text here. Click to load it

  I would check that all the power connections to the MB
were tight and clean, sometimes corrosion can develop on
power contacts.  Check that the wires going into the
connector(s) are secure.  Look for green corrosion and

 Try disconnecting and reconnecting the connector a lot
of times, with the power off, of course.  Then see if it
takes longer for the problem to reappear.


Re: Power problem

  Your intent should be to first see a problem rather than just
randomly replacing things. As you have learned, shotgunning the power
supply solved nothing AND you still have everything unknown.

  If connector was bad, then one wire and contact is obviously loose.
Furthermore, first disconnect and reconnect would clean contacts.
Connectors are designed to be self cleaning. Multiple disconnecting
and connecting a connector is often recommended when basic electrical
concepts have not yet been grasped.

  Kony provided one of first things to check.  That procedure is made
even more useful when measurements are made with all peripherals
accessed simultaneously (multitasking). (''Better' test possible only
when computer is working.)  Defective computers may even appear to
work fine.  But numbers from a 3.5 digit multimeter will expose a
problem today that will also cause failures tomorrow.

  An equivalent procedure to what Kony recommends is "When your
computer dies without warning....."  starting 6 Feb 2007 in the
newsgroup  alt.windows-xp  at:
What numbers are measuring (should you better want to understand what
is happening) was described in "PC doesn't start" on 15 Feb 2007 in
microsoft.public.windowsxp.hardware at:

  Those measurements in less than 2 minutes could eliminate loose
connector, bad AC power, power switch problems, etc all at once.  Once
a power supply 'system' is confirmed (without disconnecting or
replacing anything), then move on to other 'suspects'.  Kony describes
some of those suspects.

  My speculation is a problem with your power supply 'system'.  But
that is only speculation - not sufficient for action.  As long as a
power supply 'system' remains 'unknown', then nothing else (ie
hardware diagnostics) can report useful information.  First establish
a power supply 'system' as good or bad.  Even after shotgunning a
power supply, your power supply 'system' is still unknown.  You
Shotgunning is a bad diagnostic procedure.  First find the problem.
Then fix it.

  Another suggested replacing a CMOS battery.  Again, replace only
when a defect is identified.  Meter will measure that battery without
removing the battery.  A 3 volt coin cell is starting to become old
when voltage is 2.8 volts DC.  But coin cell might only create
problems if voltage was much lower - maybe 2.4 volts.  Replace the
battery if at or less than 2.8 volts.  But at 2.8 volts, that is not a
reason for your failure.  Again, fix things by first seeing the
problem.  Don't just wildly replace parts.

  Also post here what was discovered - especially the numbers.
Numbers may tell others what you don't yet realize.  Your replies are
only as good as information in your post.

Quoted text here. Click to load it

Site Timeline