Archived from groups: alt.comp.hardware.pc-homebuilt (
More info?)
Rarely does a damaged power supply component look damaged.
Typically not possible to fix electronics by visual
inspection. Again, the 3.5 digit multimeter is something
everyone needs to solve electrical problems. A tool as
necessary as a screwdriver. Alternative is to just keep
swapping things until something appears to work. 'Not having
a meter' is not a valid excuse because multimeters are so
inexpensive and because the human cannot either see electrons
nor remove screws using teeth. Get the meter.
In the meantime, caution: Capacitors inside a power supply
that can shock with over 300 volts (if the discharge resistors
have failed - and yes I have seen or suffered same because
those resistors do sometimes fail). Power supplies are sealed
for good reason.
Without a meter or an oscilloscope, in most every case, you
are not going to fix the power supply.
If purchasing another supply, do not use watts as an
indication of quality. In fact, often those inferior power
supplies (that don't provide a long list of numerical
specifications) cannot even output the required wattage. In
those other posts (if I remember) are many necessary functions
that as missing in power supplies sold to bean counters.
Power supplies that only provide one valid numerical spec -
the price. A typically acceptable power supply is about $60
full retail. For example, one missing function called OVP can
result in damage to everything else inside the computer.
"Bean counter' supplies are routinely missing OVP. A failed
power supply must never cause other damage; as was defacto
standard even 30 years ago.
Dust rarely causes a problem - including overheating - in a
properly designed machine. But many only see dust - don't see
the electrons - and are quick to blame dust. A more likely
possibility is an intermittent connector. Again, with that
multimeter and not yet removing cables, one can see a problem
which may still exist and later become intermittent. Only
time
will tell.
Connectors are self cleaning. By breaking and making a
connection, you may have cleaned a connector contact. If the
problem happens again, don't fix or disconnect anything. Use
the multimeter. Even a power supply that works on the power
supply tester can still be defective. That suspect power
supply could have excessive ripple which is why it works
sometimes and not others. Ripple induced problems may not be
obvious except with a multimeter and the numbers - voltages
that
are in the lower 1/4 of limits or less. That meter provides
much
more information that even a swapped power supply does.
IOW just because a power supply works today does not mean it
is working properly - which is again why the meter is so
informative. Is the power supply sufficiently sized for the
load? System would still work normally most of the time even
is power supply was marginal. But meter would identify a
marginal power supply.
So, good luck with the solution. But should symptoms
return, get the so inexpensive and useful meter (sold even in
Sears, Home Depot, and Radio Shack).
Kevin Collins wrote:
> Follow-up to the problem: I decided to swap power supply from
> another system to help troubleshoot and the system came up
> clean! So, after reading the links provided by w_tom, I decided
> I was going to open up the "bad" PSU and see if any components
> looked damaged. Everything appeared fine, but I had to dust it
> off before I could tell. On a whim, I decided to hook it up to
> the computer I had stolen the other PSU from and to my amazement,
> that system powered up fine, too!
>
> Now, I'm wondering if there was just a short or something caused
> by dust and the inadvertant shutdown was a red herring?
>
> But I'm too lazy to swap them back and see if everything still
> works, so I'm leaving them swapped...
>
> Kevin