High failure-rate for desktop PC power-supplies (Antec)?

[David H. Cook]

I've built quite a few desktop PCs systems, from motherboards, etc.
And, I've also repaired quite a few machines (e.g. Compaq, HP,
Gateway, etc).

It seems that power-supplies are the failure point much more often
than
I had expected.

Once again, a couple of days ago, on a six-month old machine I own and
built,
(with good fans, no overclocking, etc), I inserted a 2nd ethernet-PCI
card into
it, and when I brought it up for the first time, it wouldn't boot.
(No smoke or noises.)
The little activity light seemed to indicate random action, just as
though it was
seeking the disk and booting, but nothing ever showed up at the
console...not
even initial BIOS output.

So, I took the case-side off, and hooked up my little power-supply
tester-unit.
Sure enough...one of the tester's multiple GREEN lights did NOT light
up, and that
RED 'danger' light on the tester DID light up! [It's a fairly nice
Antec 400-watt supply,
which is PLENTY of wattage for this system. It is an Antec model
SP-400 'SmartPower'.]

Any advice or comments about how often power-supplies fail, or
thoughts about
why this one might have?

TIA...

Dave

 

[JS]

Did you remove the second NIC card and try again to boot?

JS

 

[David H. Cook]

Yes, of course, I did remove the 2nd ethernet card and rebooted.
That wasn't the problem. The power-supply is clearly now BAD.
[Not sure yet WHETHER that brand-new ethernet-card CAUSED
the power-supply failure yet. Or, whether the motherboard might
ALSO now be bad. I'll update this note soon, once I get the new
power-supply and learn the extent of any other possible issues.]

Dave

 

[JS]

Sorry didn't mean any disrespect.
One thing to check if you haven't already done so is to inspect that new
card and or motherboard for any discoloration and or a strong smell of
something burnt.

Good luck
JS

Yes, of course, I did remove the 2nd ethernet card and rebooted.
That wasn't the problem. The power-supply is clearly now BAD.
[Not sure yet WHETHER that brand-new ethernet-card CAUSED
the power-supply failure yet. Or, whether the motherboard might
ALSO now be bad. I'll update this note soon, once I get the new
power-supply and learn the extent of any other possible issues.]

Dave

Did you remove the second NIC card and try again to boot?

JS

 

[David H. Cook]

Hmm...I don't think you're answering my question at
all. Please re-read the base note.

I don't care at all right now, about what condition the
2nd ethernet card is in. It's only worth $12.

My question is about why I'm seeing so many power-supplies
failing. So, even if the ethernet card is flawed (e.g. a dead short
where it shouldn't have one), is that capable of permanently
damaging a power supply? If so, that would explain alot.

I'm needing to understand the answer to that question. Because
if it's reasonably possible that the $12 ethernet card will
blow out my NEW power supply, I'll just throw the ethernet
card away and not even risk putting it back in the motherboard.

Hope this clarifies...

Dave

 

[David H. Cook]

Just realized I didn't answer your last question. The
2nd ethernet PCI-card looks just fine. No sign of
burns or arcing or whatever. My assumption has
been that the power supply just failed because its
"time had come". {e.g. just like a light-bulb burns out)
But, that's naive, right? Another way to view what I'm
asking is: "Do power-supplies just burn out after
a time? What is their normal life expectancy?"

Anyway, my new power-supply (a different brand) has
arrived. It's a 430-watt unit (from NewEgg). I'm
gonna hook it up to the motherboard (without the
2nd ethernet card), and see if the motherboard
and the original system will work again with NEW
power-supply. I'll post again with the results from
that.

Cheers...

Dave

 

[Curt Christianson]

Hi David,

In *OEM* built systems, the p.s. is often the most "compromised" component
of a PC. New computer prices are so low, that the power supply is one of
the places they like to use what I would consider to be sub-standard
assemblies. IMO, power supplies are often responsible for a myriad of
problems--not just a dead machine. They may be enough for what is shipped
in the computer, but inadequate for any expansion. The supply may be rated
high enough, but the power it supplies may not be as "clean" as it
could/should be.

I wouldn't expect many problems from an Antec supply as long as it was rated
high enough.

Just an interesting FYI, as I have a great deal of respect for the author:

http://www.quepublishing.com/articles/article.asp?p=31105&seqNum=11&rl=1



--
HTH,
Curt

Windows Support Center
www.aumha.org
Practically Nerded,...
http://dundats.mvps.org/Index.htm

 

[JS]

I've had no problems, I'm currently using a six year old Dell with a 200W
power supply (Video card, 3 hard drives, sound card, NIC, and PCI based IDE
interface). I also had a Dell that was also 6 years old (put it to rest)
without problems and also have a 1 year old home built system with an Antec
SP-400 supply that is 5 years old. None have had problems. I did have one
cheap 350W supply that did cause problems.

A good supply usually cost about two or three times more than a cheap supply
of the same wattage. Also the load should not exceed more than 50 to 75% of
the rated wattage as a general rule.

Here is a link that should give your a rough estimate of what size PS you
will need.
http://www.journeysystems.com/?powercalc

JS

 

[Paul]

Just realized I didn't answer your last question. The
2nd ethernet PCI-card looks just fine. No sign of
burns or arcing or whatever. My assumption has
been that the power supply just failed because its
"time had come". {e.g. just like a light-bulb burns out)
But, that's naive, right? Another way to view what I'm
asking is: "Do power-supplies just burn out after
a time? What is their normal life expectancy?"

Anyway, my new power-supply (a different brand) has
arrived. It's a 430-watt unit (from NewEgg). I'm
gonna hook it up to the motherboard (without the
2nd ethernet card), and see if the motherboard
and the original system will work again with NEW
power-supply. I'll post again with the results from
that.

Cheers...

Dave

If you check the reviews on Newegg, and look at a wide
range of power supplies, that should give you some idea
how common power supply problems are.

If you knew nothing about a computer, had no symptoms,
and it didn't start, the most likely advice you'd
get, is to start by replacing the power supply. The
power supply is the most likely culprit, which should
tell you something of the relative reliabilities of
components.

Antec is not as high up my list as they used to be,
but that has more to do with how they treat customers
and how they deal with design defects in their products.
The computer I'm typing this on, has an Antec supply,
and I'm not concerned enough about that fact, to run
out today and change it. (Note that Antec doesn't build
supplies - they have companies like Channelwell build them.
That is fine, as long as Antec performs the appropriate
level of quality control, before the product ships to
customers. They should at least verify that the supply
works with a wide range of motherboards.)

On a much older system here, I did preventative supply
maintenance. My older system fan speed started to
vary a bit, which is a reasonably sensitive test that
something is up with the +12V rail (12v no longer holding
a steady value). In that case, rather than take a chance,
I did change the supply, before it had a chance to blow.
So you don't have to wait for a complete failure, to
change an ATX power supply.

Also, some people, when debugging their system, try too
hard to make a dodgy power supply work. Flicking the
power switch 50 times, like one guy tried, is spinning the
roulette wheel way too many times. From your first warning
of trouble, until it blows, there might not be that many
trials left in the thing.

Paul

 

[NoConsequence]

I've built quite a few desktop PCs systems, from motherboards, etc.
And, I've also repaired quite a few machines (e.g. Compaq, HP,
Gateway, etc).

It seems that power-supplies are the failure point much more often
than
I had expected.

Once again, a couple of days ago, on a six-month old machine I own and
built,
(with good fans, no overclocking, etc), I inserted a 2nd ethernet-PCI
card into
it, and when I brought it up for the first time, it wouldn't boot.
(No smoke or noises.)
The little activity light seemed to indicate random action, just as
though it was
seeking the disk and booting, but nothing ever showed up at the
console...not
even initial BIOS output.

So, I took the case-side off, and hooked up my little power-supply
tester-unit.
Sure enough...one of the tester's multiple GREEN lights did NOT light
up, and that
RED 'danger' light on the tester DID light up! [It's a fairly nice
Antec 400-watt supply,
which is PLENTY of wattage for this system. It is an Antec model
SP-400 'SmartPower'.]

Any advice or comments about how often power-supplies fail, or
thoughts about
why this one might have?

TIA...

Dave

This has nothing to do with the OS and is out of place here.

Ask elsewhere.

 

[housetrained]

microsoft.public.windowsxp.hardware. You mean to say this has nothing to do
with HARDWARE on a machine running XP?

--
John the West Ham fan
(e-mail address removed)
<><

I've built quite a few desktop PCs systems, from motherboards, etc.
And, I've also repaired quite a few machines (e.g. Compaq, HP,
Gateway, etc).

It seems that power-supplies are the failure point much more often
than
I had expected.

Once again, a couple of days ago, on a six-month old machine I own and
built,
(with good fans, no overclocking, etc), I inserted a 2nd ethernet-PCI
card into
it, and when I brought it up for the first time, it wouldn't boot.
(No smoke or noises.)
The little activity light seemed to indicate random action, just as
though it was
seeking the disk and booting, but nothing ever showed up at the
console...not
even initial BIOS output.

So, I took the case-side off, and hooked up my little power-supply
tester-unit.
Sure enough...one of the tester's multiple GREEN lights did NOT light
up, and that
RED 'danger' light on the tester DID light up! [It's a fairly nice
Antec 400-watt supply,
which is PLENTY of wattage for this system. It is an Antec model
SP-400 'SmartPower'.]

Any advice or comments about how often power-supplies fail, or
thoughts about
why this one might have?

TIA...

Dave

This has nothing to do with the OS and is out of place here.

Ask elsewhere.

 

[JohnO]

microsoft.public.windowsxp.hardware. You mean to say this has nothing to

do with HARDWARE on a machine running XP?

Have you ever yelled at a brick wall in an attempt to make it move? Same
thing here.

-John O

 

[Guest]

I fully support the statements / views given by JS regarding rating of Power
Supplies (PS) vs. the nominal load. From my work with heavy duty electronics
for 35 years, it is quite interesting to experience the peak current surges
generated when a system is "powered up". These "spikes" are added up from
electrolytic capacitors charging, - motors starting (i. e HD-drives in PCs),
- chips shifting states and so on. And these "peak loads" takes place when a
system is "powered down" as well.

My experience tells me that especially electrolytic capacitors tends to dry
off when working in high ambient environments, i.e. when a PS is working at –
or close to its rated load, and at room temperatures exceeding 85F. Thus is
providing cooling, either forced – or by other means, as required for the PS
of significant importance.

And as remark to Davids "Any advice or comments about how often
power-supplies fail –", I would say that electrolytic capacitors in PSs may
dry off in five years or so when working at worst case conditions. Secondary
failures may then be blown fuses, crowbar shut down (when such circuitry is
provided) or other burned components in the PS.

 

[Guest]

Yeah - it has no consequences

Regards,
Ka2H

Have you ever yelled at a brick wall in an attempt to make it move? Same
thing here.

-John O

 

[w_tom]

Just realized I didn't answer your last question. The
2nd ethernet PCI-card looks just fine. No sign of
burns or arcing or whatever. My assumption has
been that thepower supplyjust failed because its
"time had come". {e.g. just like a light-bulb burns out)
But, that's naive, right? Another way to view what I'm
asking is: "Do power-supplies just burn out after
a time? What is their normal life expectancy?"

Life expectancy of a power supply means the manufacturer has
provided that MTBF number in a full page of numerical specs. If those
specs are not provided, then suspect that supply is 'missing' things.
Yes, a full sheet that includes references to hi-pot testing, PFC
harmonics, safety compliance, etc

On that spec sheet would be a reference to overpower protection.
Why? Because all outputs from a power supply must be shorted together
and still power supply is not harmed. IOW if that second ethernet
card caused power supply failure, then he power supply was defective
before it failed.

Typical life expectancy for a power supply is 100,000 hours. As
Ka2H notes, that number of hours is defined by ambient temperature.

So you are experiencing numerous failures. Nobody can say why
because internal failed component has not been identified. Best
evidence is a dead body. Your post does not even list which voltage
has failed or why. Voltage too low or no voltage at all? That's a
problem with those power supply testers. Go-Nogo testing reports
woefully too little useful information.

Did power supply come defective? Use a 3.5 digit multimeter with
power supply under maximum load - to see some defects before it
results in a failure. A new power supply is measured as soon as
connected. Notice ternary conditions: Power supply can be good.
Power supply can be defective but not yet cause computer failure.
Power supply causes computer failure.

 

[David H. Cook]

JS -

Good info...now I'm liking it.
Ok, the failed Antec does have dual-12v rails...one rated at 14A and
other
at 15A. My new replacement is a 'Termaltake' 430-watt unit and it has
18A on the 12v rail, according to sticker on its side-panel.

Yeah, I'll go back and do a mini-autopsy on the failed Antec, and see
if
the missing green light and the red-danger light map onto something.
Probably have to just get out my meter and put power to the Antec
and see which leg(s) check out, right?

Cheers...

Dave

 

[JS]

Thanks, glad this helped in some small way.
Also what Ka2H mentioned is true as there is a startup surge which only adds
to the problem.
One thing I forgot to mention is a surge protection or UPS between the AC
line power and the PC, without which makes the power supply bear the brunt
of any AC power issues.

Again
Good Luck
JS

 

[JS]

Every computer and the OS that it runs is dependent on hardware reliability
..... bad or failing hardware leads to down time either on a home PC,
Business workstation or Server(s).

It is of SomeConsequence!
JS

I've built quite a few desktop PCs systems, from motherboards, etc.
And, I've also repaired quite a few machines (e.g. Compaq, HP,
Gateway, etc).

It seems that power-supplies are the failure point much more often
than
I had expected.

Once again, a couple of days ago, on a six-month old machine I own and
built,
(with good fans, no overclocking, etc), I inserted a 2nd ethernet-PCI
card into
it, and when I brought it up for the first time, it wouldn't boot.
(No smoke or noises.)
The little activity light seemed to indicate random action, just as
though it was
seeking the disk and booting, but nothing ever showed up at the
console...not
even initial BIOS output.

So, I took the case-side off, and hooked up my little power-supply
tester-unit.
Sure enough...one of the tester's multiple GREEN lights did NOT light
up, and that
RED 'danger' light on the tester DID light up! [It's a fairly nice
Antec 400-watt supply,
which is PLENTY of wattage for this system. It is an Antec model
SP-400 'SmartPower'.]

Any advice or comments about how often power-supplies fail, or
thoughts about
why this one might have?

TIA...

Dave

This has nothing to do with the OS and is out of place here.

Ask elsewhere.

 

[w_tom]

One thing I forgot to mention is a surge protection or UPS between the AC
line power and the PC, without which makes the power supply bear the brunt
of any AC power issues.

If that 'brunt' created a problem, then a UPS in battery backup mode
would destroy a computer's power supply. Review a comptuer grade UPS
output in battery backup mode. 120 VAC can be two 200 volt square
waves with a spike of up to 270 volts between those square waves.
This is called a modified sine wave. That voltage may be destructive
to some small motors. But computer power supplies must be so robust
as to be powered by this computer grade UPS - without damage. ATX
power supplies are required to be so robust as to even make that
'dirtiest' UPS output irrelevant; also make what are lesser AC power
issues irrelevant.

But again, no definitive answer is possible without knowing what
inside the power supply has failed.

 

[NoConsequence]

microsoft.public.windowsxp.hardware. You mean to say this has nothing to do
with HARDWARE on a machine running XP?

The problem described was purely HARDWARE and had nothing to do with
it's interaction with the OS. Hence, it is being brough up in an
improper group.