Possible Failing PSU?

  • Thread starter Thread starter Dan Ricketts
  • Start date Start date
D

Dan Ricketts

Hi All,

My Dad's having problems with his PC, which I think are being caused by the
PSU, but am not 100% sure.

Basically the machine will run for a short while (10-15 minutes tops) and
will then cut out, not allowing it to be powered back on until the line cord
is removed from the machine and reinserted. The PC can then be booted only
for the same to happen again.

So far I have checked all power connectors inside the machine (removed and
reconnected), as well as reseating the processor, RAM, and PCI cards - all
to no avail. There is 240vac at the connector inside the psu, but no obvious
charring or burning smells.

the spec of the machine is as follows:

P3 550,
256Mb Ram,
Diamond Stealth Graphics Card,
4Gb Seagate HDD

Any light you guys could shed on this would be very much appreciated.

Thanks In Advance

Dan Ricketts
 
Sure, that's a reasonable theory. Have you also noted the temp of the cpu in
the bios? If the cpu is overheating that's reason as well for the machine to
be shutting down.
 
You could have your answer in but two minutes (the work
takes that long; but learning how to work and read the numbers
takes many times longer). Background posted previously in two
discussions: "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://www.tinyurl.com/2musa

No one can answer your question without numbers. Anything
less is wild speculation. Your numbers must lie in the upper
3/4 limits of chart in that second discussion.

When system cuts out, what happens to the Power Good
signal. If you don't know, then no one can provide a
responsible answer. Again, get those numbers.

Quite simply take numbers with system powered on. Then do
same while OS is accessing all peripherals simultaneously.
Read directories repeatedly from CD-Rom, diskette, hard drive,
and network drives while connected on modem to internet. Does
the voltage drop out of upper 3/4 range of limits? If yes or
no, then you have an answer (except when voltage sits within
limits but only in lower /14 range). Either fix the problem
or more on to other usual suspects. Meter gives you a direct
answer without wild speculation. Don't waste time or money.
Get the meter. Gets answers. Fix problem and move on. IOW
if you are not almost 100% sure, then you have not a clue -
which is why we always demand the numbers up front.
 
Hi All,

My Dad's having problems with his PC, which I think are being caused by the
PSU, but am not 100% sure.

Basically the machine will run for a short while (10-15 minutes tops) and
will then cut out, not allowing it to be powered back on until the line cord
is removed from the machine and reinserted. The PC can then be booted only
for the same to happen again.

So far I have checked all power connectors inside the machine (removed and
reconnected), as well as reseating the processor, RAM, and PCI cards - all
to no avail. There is 240vac at the connector inside the psu, but no obvious
charring or burning smells.

the spec of the machine is as follows:

P3 550,
256Mb Ram,
Diamond Stealth Graphics Card,
4Gb Seagate HDD

Any light you guys could shed on this would be very much appreciated.

Thanks In Advance

Dan Ricketts
Click to expand...

Having to pull the power cord (or otherwise cut the AC to
it) is typically a sign that the power supply's protection
circuitry has shut it off. That can be either a problem
with power supply itself OR a problem with connected
component(s), particularly some type of short-circuit.
"Most often" it is the power supply itself, but could simply
be a wire frayed on a sharp case edge or caught in a fan
blade. Might be something as simple as power supply fan
failure, check that too.

To isolate the power supply you can disconnect it from the
system, put a load on it enough so that similar current is
drawn to what the system would use, and monitor it, whether
it then also shuts off. A multimeter would preferribly be
used to verify that suitability of the load, that it's
keeping the power rails within spec such that PSU is in a
proper working voltage zone, rather than shutting off for a
different/unrelated reason- uneven load causing an unloaded
rail to drift too high or low outside of the PSU's
set-point, threshold. There could then be an entirely
separate and unrelated cause for shutdown.

So either the PSU passes or fails this simple test. If it
passes it may then be time to strip system down to bare
minimal parts, possibly substituting any parts if possible
to result in isolation of as many other parts/variables as
feasible. This may include unplugging drives, so it would
not be booting the OS, rather leaving it sitting in the bios
(at a health/hardware monitor screen if one is provided in
the bios menus).

Examine all capacitors on the motherboard, video, and in
PSU too. If all else fails source another power supply for a
trial or purchase.
 
It sounds like a component in the PSU is overheating over time and then
shutting the unit down. I would try replacing the supply.
 
Dan said:
My Dad's having problems with his PC, which I think are
being caused by the PSU, but am not 100% sure.

Basically the machine will run for a short while (10-15
minutes tops) and will then cut out, not allowing it to
be powered back on until the line cord is removed from
the machine and reinserted.
Click to expand...

That's a sign of a marginal power supply, and if it had been working
normally until recently, then the most likely cause is electrolytic
capacitors failing in the low voltage section. It runs OK for 10-15
minutes since it takes time for the inverter transformer to heat up and
lose capacity (20% is common), and the marginal capacitors make the
regulator drive the transformer harder than usual and closer to its
maximum.
 
Back
Top