Power Supply Diagnostics?

[RobV]

In message
<[email protected]>
[snip]

If you ARE suggesting that an average end user should poke around
inside a live system with a multimeter's probes, I'd suggest that
perhaps we're in completely different realities.

You certainly are in a different reality. Do you comprehend the name of
the news group to which you are posting? Hint: *pc-homebuilt*.

Anyone posting to a news group named *alt.comp.hardware.pc-homebuilt*
has, logically, enough knowledge to build a computer and certainly knows
how to use a DVM. If not, that person is in the wrong news group and
wrong reality.

 

[John Doe]

In message
<f472e962-b5b4-417e-b7d1-247b012546c0
e4g2000hsg.googlegroups.com>
[snip]

If you ARE suggesting that an average end user should poke around
inside a live system with a multimeter's probes, I'd suggest that
perhaps we're in completely different realities.

You certainly are in a different reality. Do you comprehend the
name of the news group to which you are posting? Hint:
*pc-homebuilt*.

Yeah, and if you've been here in this group for any length of time,
you would know that we get many questions from people needing to
troubleshoot their system or upgrade their system, whether or not
it's homebuilt.

Anyone posting to a news group named
*alt.comp.hardware.pc-homebuilt* has, logically, enough knowledge
to build a computer and certainly knows how to use a DVM. If not,
that person is in the wrong news group and wrong reality.

I've designed and built lots of small electronic devices, mainly
using 4000 series logic (and all sorts of LEDs). I've home built
(that is, constantly upgrading) my own computer for many years. The
only time I've had my PC case open while powered is to make sure
that the fans are running. My multimeter is a low end BK Precision.
If you have a modern mainboard, the voltage level readings shown by
a simple Windows utility or in the BIOS are the best way to figure
out whether your system is properly powered.

Do you have citations to back up your claim that the computer's own
voltage readings are "notoriously inaccurate"? I can understand it's
using a multimeter if you are a technician and you need to know the
voltage level readings on an outdated mainboard. If you're an
average user or homebuilder with an outdated mainboard, swapping the
power supply probably is a better idea.

Do you use a multimeter for anything PC besides checking power
supply voltages? Just curious.

 

[Strobe]

"RobV" <robv nowhere.invalid> wrote:

Do you have citations to back up your claim that the computer's own
voltage readings are "notoriously inaccurate"? I can understand it's
using a multimeter if you are a technician and you need to know the
voltage level readings on an outdated mainboard.

Many people come here because they already have a problem.
One possible cause of problems is that their mobo is misbehaving - thus,
it makes sense to *not* rely on that mobo for voltage information.

 

[DevilsPGD]

In message <[email protected]> "RobV"

In message
<[email protected]>
[snip]

If you ARE suggesting that an average end user should poke around
inside a live system with a multimeter's probes, I'd suggest that
perhaps we're in completely different realities.

You certainly are in a different reality. Do you comprehend the name of
the news group to which you are posting? Hint: *pc-homebuilt*.

Anyone posting to a news group named *alt.comp.hardware.pc-homebuilt*
has, logically, enough knowledge to build a computer and certainly knows
how to use a DVM. If not, that person is in the wrong news group and
wrong reality.

Unlikely. I know several people who have assembled their own computers
that wouldn't know what to do with a multimeter if it came with an
instruction manual. One of them was responsible for my initial
introduction to this group (more or less, rather, they stumbled across
it trying to solve another issue -- he read for a while and then
wandered off, now I build his PCs for him)

The days of actually "building" a computer are more or less over, at
best, most people just assemble 'em, and that definitely doesn't require
knowledge of how to use a multimeter in most cases.

 

[DevilsPGD]

In message
<07185418-a929-45ab-a008-37afd7ce8762@j20g2000hsi.googlegroups.com>

No minimally acceptable power supply must ever damage any computer.

Agreed. And yet, such supplies are sold on a daily basis.

No one who builds reliable computers would use a $10 or $20 supply.

Agreed. And yet, such supplies are installed on a daily basis all
around the planet. Some of those machines end up on my desk to be
repaired, which leaves me with needing some way to verify whether the
problem is the supply or motherboard is the cause of a no-power
situation.

As it turns out, a power supply tester can often point out that the
problem is the supply.

It can never reliably tell you that the problem isn't the power supply
though.

Where bad practices
are advocated, then also recommended is a power supply tester.

I'm suggesting STARTING with one, not ending with one.

Why is the OP waiting for a power supply tester to be delivered?

Because it's easier to use a power supply tester AFTER it is delivered
then BEFORE it is delivered?

Because most people don't waste money on foolishness and don't buy
miracle cancer curing drugs based upon a TV commercial. The meter is
available in most every town, does everything the power supply tester
would do - and much more. Whereas the power supply tester might see a
defective supply, the meter may identify the defect every time. Why?
Metert provides numbers. Meter can even test a power supply that is
not connected to any computer. DevilsPGD lied when he said,
"multimeter won't even power up an ATX basedpower supplyon". He also
lied when posting, "nor will it help you test voltages under any sort
of load".

Okay -- How would a multimeter help you power up an ATX based power
supply and test each of the individual voltages without connecting the
PSU to a motherboard?

. Why waste money on a tester that has no other function? The OP
must wait for it to be delivered. Tester cannot even provide a useful
answer. After using the power supply tester, a meter must still be
used.

Not if the power supply tester indicates a problem -- If so, you're done
troubleshooting.

 

[spodosaurus]

In message
<07185418-a929-45ab-a008-37afd7ce8762@j20g2000hsi.googlegroups.com>


Agreed. And yet, such supplies are sold on a daily basis.


Agreed. And yet, such supplies are installed on a daily basis all
around the planet. Some of those machines end up on my desk to be
repaired, which leaves me with needing some way to verify whether the
problem is the supply or motherboard is the cause of a no-power
situation.

As it turns out, a power supply tester can often point out that the
problem is the supply.

Actually, the one I purchased (and returned, at my expense for postage)
registered the bad rails as good and the good rails as bad. It also only
went to one decimal place, and rounding up or down can make a big
difference!! This is typical of power supply testers. Most of the
voltages were off by at least 0.2V as well. I dug out my old $10AUD
multimeter and tested all my supplies after that (paper clip jumper to
switch it on with at least one optical and one hard drive attached).

The multimeter was very informative, and isolated the bad PSU which was
replaced under the Antec three year warranty. By the way, the retailer
was Simline.com.au and they dealt directly with Antec for me even after
2.5 years. THAT's good service!

Ari



--
spammage trappage: remove the underscores to reply
Many people around the world are waiting for a marrow transplant. Please
volunteer to be a marrow donor and literally save someone's life:
http://www.abmdr.org.au/
http://www.marrow.org/

 

[w_tom]

Okay -- How would a multimeter help you power up an ATX based power
supply and test each of the individual voltages without connecting the
PSU to a motherboard?
...
Not if the power supply tester indicates a problem -- If so, you're done
troubleshooting.

Short a power supply's PwrOn pin to ground even with a paper clip
(green wire to black wire). Measure voltages on any orange, red,
yellow, purple, and gray wires. Meter performs exactly what a power
supply tester would do. Some supplies require (but often power up
anyway) a small load. Any old peripheral or even a burned out hard
drive will do as a load. The tool that has numerous functions has also
done everything the power supply tester will do.

Only useful measurements are 'definitive'. "It might be" is wasted
labor. A defective supply can be reported good by a tester (see other
posts). But troubleshooting is complete only when the answer is
either 'defintively good' or 'definitively bad'. That means using a
meter. Meter is required, again, if the supply is replaced. A new
and defective supply may still boot the computer. An intermittent
created, instead, by a marginal driver on power supply controller is
still not detected or fixed. Both old and new supply were OK and that
intermittent problem still exists to cause failures again later. Just
another example of failures not detected by a tester. So the
misinformed repairman starts replacing other perfectly good parts?
Just another example of failure made more complex because a tester was
being used; because the answers were not definitive.

The OP has problems that could easily be any part of the power
supply 'system'. Which 'system' part does a tester test? Only one
component, not definitively, and not other parts of the power supply
'system'. Three of so many reasons why a tester is not useful. Also
missing are numbers necessary for useful replies. Troubleshooting
cannot be completed with the tester. Every 'fix' would only be
speculative. Meter is required to complete troubleshooting - with or
without using the tester.

 

[Strobe]

Not if the power supply tester indicates a problem -- If so, you're done
troubleshooting.

We wish it were so!

Murphy says that the clear and obvious failure is masking 3 tricky ones...

 

[John Doe]

Many people come here because they already have a problem.
One possible cause of problems is that their mobo is misbehaving -
thus, it makes sense to *not* rely on that mobo for voltage
information.

But seriously.