Re: System Failure - CPU Task

We don't know if power supply is properly constructed which
is why a minimally sufficient supply is sold with a long list
of numerical specs. No such written (numerical) claim
suggests a power supply - that appears to work just fine for
months - was actually defective when designed.

The OP asked if he was doomed. Well, is power supply
minimally sufficient? Where are the page long list of
numerical specifications? No specs, then suspect the worst.

But this still does not answer the OPs question. Even if
the power supply was defective, MBM5 readings would tell him
nothing useful ... without first having calibrated those
reading with a 3.5 digit multimeter. Furthermore, MBM5
readings will not record spikes nor AC mains brownouts. It is
a monitor that can report slow voltage changes. To determine
proper power supply voltages, first those MBM5 readings must
be calibrated with a meter. The meter could report symptoms
of an 'aging' power supply (ie diminishing voltages or
excessive ripple) - an upcoming weakness that is not yet
caused complete computer failure. A weakness that the
multimeter can identify but the MBM5 software cannot (without
first using the meter).

Properly noted is a clone computer market chock full of
marginal supplies. So many 'computer assemblers' don't even
know how electricity works. First symptom of an inferior
supply is manufacturer does not even claim to provide
essential functions. If manufacturer did not provide a long
list of numerical specifications, then some functions are
probably missing. Second symptom is a supply selling for less
than $65 full retail. Computer assemblers too often only buy
on two numbers - price and watts. Those other essential
functions be damned.

Can an AC mains loss of only a few ms cause power failure?
No. That is also specifically stated in the standards. A UPS
that typically takes 10+ milliseconds to switch from or too
battery backup does not crash computers. Computers can
tolerate short millisecond power losses. If not, even a UPS
switchover would crash a computer. Just another power supply
function IF power supply was built by a responsible
manufacturer who provides numerical specifications.

If power is crappy, then does a online UPS provide benefits
that the standby UPS does not? Yes. But then it better since
an online UPS costs about five times more money. Meanwhile,
those on-line UPS advantages are totally lost on a computer.
Computer power supply standards even 30 years ago makes those
benefits irrelevant.

Is the power supply insufficient? Another good indicator is
the price. A power supply selling for less than $65 full
retail typically is missing essential functions. It provides
just enough functions so that the 'computer assembler' feels
it is good - because it still works six months later.
Furthermore, $65 for a power supply does not guarantee a good
power supply. The reverse logic is not true. A power supply
selling for less must forget essential functions. A power
supply selling for more need not contain essential functions.
Just another reason why those numerical specs are important.

How do I describe my answer here and previous? Right on the
money. MBM5 to measure a defective power supply is bogus.
Obviously bogus because motherboard monitors are not
sufficiently calibrated. Called a "monitor" to detect voltage
changes; not measure voltage levels. To determine power
supply integrity, one must measure voltage levels - the 3.5
digit multimeter. Furthermore, noisy AC mains power does not
adversely affect a properly constructed computer nor is it
eliminated by a typical plug-in UPS. Either computer has
sufficient noisy power to keep running, or computer no longer
gets a Power Good signal - shuts down. Power is fully
sufficient or fully insufficient - according to computer.
This another function of the power supply system.

Information as posted earlier answers the OP's question.
Provided earlier were even AC mains standards. Power supply
must work just fine - not even indicate low voltage in MBM5 -
when incandescent bulbs are at 40% intensity. MBM5 could
report a defective supply that did not maintain voltage. But
then the computer would crash so that MBM5 could not be read.
Neither on-line UPS nor MBM5 provide useful assistance nor
useful information - as was stated previously. The power tool
for such analysis is a 3.5 digit multimeter.

Mercury wrote:
> The OP has given a sketchy description. However food for thought
> often leads to using appropriate terminology and gaining feedback
> that drills down to the fault. Are you unfamiliar with conversation
> perhaps?
>
> How do you know the PSU is correctly constructed and operating? You
> don't, nor do I and with the dominance in the market of marginal PSU
> it is a reasonable place to check.
>
> The PSU could be faulty, it could be producing diminishing voltages
> due to aging capacitors or have developed some other fault. The PSU
> may have been marginal from day 1, may never have adhered to
> specifications, and may not behave per specs with low mains voltage.
> Or, it may have been near 100% to spec originally and have developed
> a fault... It may not be the source of the fault at all, the mains
> could be, or the OP could have described things in a misleading
> manner, so get off your high horse and contribute something
> positive.
>
> The mains may be marginal: There may be power fails lasting from a
> few ms to longer - enough to trip the PSU even if the PSU were on
> spec and is supposed to handle it.
>
> Plotting voltages from MBM 5 is not always conclusive - out of tolerance
> voltages are however an indicator that there is something away from the
> ideal, and possibly of a degrading PSU. If the PSU has degraded then the
> likelihood of power related issues increases, so MBM 5 can be helpful,
> so could sung a DVM or a plotting device on the mains, or a lab...
>
> PSU's do fail. They do age.
>
> Mains can do all sorts of wonderful things. Here, the power can fail 6
> times in a day - 4 out of those six times may be so brief that they are
> inconsequential, and other times it can go off for hours. The power is
> here crappy, but the true online UPS I have is good (1100va) and hooked
> up correctly.
>
> So excuse me for talking about reality. People don't live in testing
> labs, they consume made products which are sometimes marginal, and have
> to use them on mains that is sometimes also a problem.
> ...
.

Loading