Someone gave you bad advice. If there is any question at all about a power
supply, replace it.
That's rather ridiculous advice, since myriad problems could
be caused by a poor or failing PSU but one can't go and
replace the PSU every time Windows/etc hiccups. They would
need to reduce the variables, using the tools at their
disposal. A multimeter is one of those tools. Perhaps more
important is this question:
What did you think is a more valid indicator of PSU
malfunction? Ok, a 'scope, but who that had one, would have
an aversion to even using a 'meter?
Don't **** around with a multimeter, or "range holding
multitester". Voltage dips that can cause instability will not necessarily
register on a multimeter, as the multimeter display won't update as fast as
the voltage regulator switches.
Actually, yes they will register. If the event was so short
that a reasonable meter can't register it, it was also short
enough to be buffered by the bulk electrolytics that just
about any and all parts have. When a PSU problem becomes
evident through system malfunction, more often it can be
realized though a meter reading.
Plus, not all PSUs fail gracefully. They
often destroy other connected components when they go. So while you are
dinking around trying to get voltage readings, the PSU is a ticking time
bomb that is ready to fry your mainboard (etc.) at any moment. So while you
might "capture" a voltage drop, this is like taking a picture of an airplane
as it explodes. Good to satisfy your own curiosity, but not good for the
airplane.
Same is true of a brand new replacement, if you have a PSU
that might be working fine and introduce another PSU that is
not yet qualified or proven working properly (a multimeter
is nice here, too), simply adding that additional variable
is a potential to cause damage or just unnecessary expense
while not resolving anything. Replacement should be a last
resort if other methods of determining PSU fitness can't be
absolute.
If you ever have a PSU that will show a bad reading on a multimeter, then
the multimeter will be redundant at that point,
obviously not, because if one was measuring the PSU voltage
there was still the suspicion that the PSU might not be the
problem. One does not tend to replace parts that they feel
might not be the problem unless they happened to have
exhausted all other ideas and had a PSU lying around. How
do they know the PSU lying around is good? I suppose they
guess that it is if they don't take voltage readings. How
many system owners happen to have that spare known-good PSU
just lying around instead of having to order, wait, put the
$ into that? The percent is close to zero, if the # of
systems brought to me with a PSU problem is any indicator
then they did not have the spare, and did not know it was a
PSU problem else they would've just replaced it had they
known exactly what was needed. A meter is a source of info
so one does approach knowledge instead of guesses.
as there will be NO DOUBT in
your mind that the PSU is bad
No, at that point of a bad reading you have exactly what you
wanted, evidence to support your suspicion that it was, or
refute that it wasn't, a PSU problem. You could have the
ideal situation, KNOWING you have a psu problem based on the
bad reading instead of only guessing. Guessing can be
expensive if the system is anything more than a low-end box,
or time consuming if this system is set up for specific
needs/desires/etc and stores important data.
If there is a strong suspicion that the PSU is bad but that
the other parts are ok, it could be helpful to use an
(invaluable) load instead of critical parts, but having
tried the system with the PSU (pretty much manditory, that's
what was running the system up until that point), odds are
low that the PSU would have caused a problem, but be working
enough that it's much of a risk to
. . . even if the nearest multimeter is in the
next zip code.
Multimeter testing is never redundant, it is always the
first source of real information. OTHER guesses, anecdotal
evidence, software readings or even replacement of the PSU
and finding it then resolving the situation, can only be
redundant to actual determination though measurement, the
multimeter being the obvious cost-effective tool to do it.
This is all within a certain context though, that if one
needed to be told to measure a supply voltage with a meter,
they may not have the experience (or even the meter) to do
it. In such cases it would be better to just take the
system to an experienced technician if fast resolution is
important. They could instead buy the replacement PSU for
about the cost of the technician's bench fees, but they may
not get the system working either... and if the owner isn't
able to even use a multimeter, I'd have second thoughts
about having them muck around inside a system (that already
had a problem) TOO much. This may introduce yet another
problem, instead of keeping things static except only to
reduce the # of variables in troubleshooting.
