Personal computing discussed

Moderators: renee, SpotTheCat, Nemesis

 
OldBob
Gerbil
Topic Author
Posts: 53
Joined: Wed Dec 01, 2004 12:18 pm
Location: St. Louis

PSU or MB failure, how know?

Wed Mar 02, 2011 9:09 pm

Antec Sonata case with Antec power supply. ASUS MB. Purchased in 2005. System normally on 24/7. Away for a week, shut system down and power switch on surge protector/strip turned off but remained plugged into household wiring. Reasonable chance of surge due to lightening storm and power flickers (same storm). When I first turned on power, case fan and front lights came on for about one second. Pushing power button again does nothing unless I first unplug the power supply. Held 'on' button and reset button in (with power disconnected) and still get only one brief power shot. Seems likely that either PSU or MB has a problem. How do I determine which?
OldBob
 
michael_d
Gerbil Elite
Posts: 569
Joined: Sun Dec 16, 2007 2:42 pm

Re: PSU or MB failure, how know?

Wed Mar 02, 2011 10:09 pm

If the machine refuses to power on I would check PSU first. Does the PSU fan spin?
 
OldBob
Gerbil
Topic Author
Posts: 53
Joined: Wed Dec 01, 2004 12:18 pm
Location: St. Louis

Re: PSU or MB failure, how know?

Wed Mar 02, 2011 11:58 pm

The PSU fan spins for about a second, just like the case fan.
OldBob
 
just brew it!
Administrator
Posts: 54500
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: PSU or MB failure, how know?

Thu Mar 03, 2011 12:04 am

Sounds like the PSU to me, but that is by no means certain. Your best bet is to try a different PSU to see if that fixes it.
Nostalgia isn't what it used to be.
 
anotherengineer
Gerbil Jedi
Posts: 1688
Joined: Fri Sep 25, 2009 1:53 pm
Location: Northern, ON Canada, Yes I know, Up in the sticks

Re: PSU or MB failure, how know?

Thu Mar 03, 2011 12:07 am

I had a sonata with a smart power psu, I can;t remember if it was a 400 or 450W but about a year ago it started making a high pitch whining noise so I swapped it out.

If I were you (and you can scrounge up the funds) I would get a new decent PSU, even if the old one ends up being ok, at least you have a back up and a new psu should last a good 5 yrs too.
Life doesn't change after marriage, it changes after children!
 
OldBob
Gerbil
Topic Author
Posts: 53
Joined: Wed Dec 01, 2004 12:18 pm
Location: St. Louis

Re: PSU or MB failure, how know?

Thu Mar 03, 2011 12:48 am

Replying to Just Brew It ...

Obviously life will be easier if it's the PSU, but what's the logic? If things run for a second, including the PSU fan, wouldn't that indicate that the MB wasn't sending the right signal to the PSU (such as saying RUN)?
OldBob
 
UberGerbil
Grand Admiral Gerbil
Posts: 10368
Joined: Thu Jun 19, 2003 3:11 pm

Re: PSU or MB failure, how know?

Thu Mar 03, 2011 4:48 am

OldBob wrote:
If things run for a second, including the PSU fan, wouldn't that indicate that the MB wasn't sending the right signal to the PSU (such as saying RUN)?
It doesn't really work like that. The signal to say "run" is the power switch; there's no communication beyond that. The rationale here is that the PSU isn't completely dead, but as the various components come to life and draw power, the PSU falls over and gives up. Of course that's not the only possibility, and it is possible for a bad motherboard to give these symptoms, but swapping out the PSU is easy and it's worthwhile having a known good spare around anyway.
 
just brew it!
Administrator
Posts: 54500
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: PSU or MB failure, how know?

Thu Mar 03, 2011 7:17 am

To expand on that just a bit, the signal to start (which is just the motherboard grounding a wire that leads back to the PSU) is obviously getting there, since it tries to start (as evidenced by the fans spinning for a second). The fact that it immediately powers down again is probably an indication that the PSU's internal failsafe circuits are kicking in and shutting things down.

The internal failsafe circuits can fire either because the PSU detects that its voltages are out of tolerance, or because the PSU believes it is overloaded. Out of tolerance voltages are a PSU failure; overload can either be a PSU failure or a shorted component (motherboard, hard drive, etc.) elsewhere in the system.

Which brings to mind one more test you can do: See if the system will POST with the power to the hard drive disconnected. The hard drive is not a likely culprit but it is possible, and very easy to rule out. If it still won't POST with the hard drive disconnected then the problem is not the hard drive; if it POSTs, then either the hard drive has failed (with a short circuit that is overloading the PSU), or the PSU has become so marginal that it can't supply the power surge required on the +12V rail to spin up the hard drive platters.
Nostalgia isn't what it used to be.
 
OldBob
Gerbil
Topic Author
Posts: 53
Joined: Wed Dec 01, 2004 12:18 pm
Location: St. Louis

Re: PSU or MB failure, how know?

Thu Mar 03, 2011 12:33 pm

Weird! This morning, when I went to pull the power connectors from the hard drives I noticed the 'power to the MB' led was on. It turned on and off with the main power switch. So I tried the on/off button and the system powered up and is generally working fine. Only problem is that the external hard drive is not working properly. I have two internal hard drives backing up a RAID, but if either the MB or PSU is dead I'm screwed and won't be able to do my taxes :(

So it's off to BestBuy for a new external drive ..too much data to use CD's ... at least in the short term.
OldBob
 
just brew it!
Administrator
Posts: 54500
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: PSU or MB failure, how know?

Thu Mar 03, 2011 2:59 pm

I sure would not count on that system being stable enough to do your taxes on at this point. Better get that Plan B in place!
Nostalgia isn't what it used to be.
 
OldBob
Gerbil
Topic Author
Posts: 53
Joined: Wed Dec 01, 2004 12:18 pm
Location: St. Louis

Re: PSU or MB failure, how know?

Thu Mar 03, 2011 3:28 pm

Is it possible that one of the drives was 'stuck,' and was jarred free when I moved the box? That would only work if a stuck drive increased the demand on the PSU significantly.
OldBob
 
wibeasley
Gerbil Elite
Posts: 952
Joined: Sat Mar 29, 2008 3:19 pm
Location: Norman OK

Re: PSU or MB failure, how know?

Thu Mar 03, 2011 3:45 pm

OldBob wrote:
Is it possible that one of the drives was 'stuck,' and was jarred free when I moved the box?
That's a scarier thought. If you haven't already, can you make a backup before you turn off your computer? Even if it's only your most important data on a 4GB usb drive.
 
OldBob
Gerbil
Topic Author
Posts: 53
Joined: Wed Dec 01, 2004 12:18 pm
Location: St. Louis

Re: PSU or MB failure, how know?

Thu Mar 03, 2011 8:08 pm

Got a Seagate external drive which is currently backing up all my files. Supposedly it runs in the background and keeps all files always backed up. So I guess I'll have no excuse to avoid the IRS.

Thanks to all for help and suggestions. It's nice to have somewhere to sort out the unexpected problems. :)
OldBob
 
westom
Gerbil
Posts: 69
Joined: Tue Jun 16, 2009 2:47 pm

Re: PSU or MB failure, how know?

Thu Mar 03, 2011 9:35 pm

OldBob wrote:
Antec Sonata case with Antec power supply. ASUS MB. Purchased in 2005. System normally on 24/7. Away for a week, shut system down and power switch on surge protector/strip turned off but remained plugged into household wiring. Reasonable chance of surge due to lightening storm and power flickers (same storm).

Flicker causes no damage. If a surge exists (typically once every seven years), then what could not be stopped by thee kilometers of sky was not stopped by millimeters in a switch. Power off (open switch) is virtually zero protection.

Either a surge connects harmlessly outside the building to earth. Or you have all but invited it to go hunting destructively inside. A most common incoming path is AC electric. Outgoing to earth via phone lines. In part, because all phone lines already have an effective, superior, and properly earthed protector (not ineffective products located too close to appliances).

Without effective protection on every incoming wire (ie AC electric wires) where wires enter the building, then virtually no protection exists (except inside every appliance). Once surge energy is inside, then nothing will stop a destructive hunt for earth. A protector adjacent to an appliance can sometimes make surge damage easier.

Computer contains a power controller that makes many decisions. If it turns power on and fails some checks, then it powers off the machine. Those are your symptoms. Reasons for power off means either replacing parts until something works (shotgunning). Or obtaining numbers in one minute from six wires with a multimeter (a $17 tool in Wal-Mart). The reply provides an immediate and complete answer in a next post. Your choice. Keep posting repeatedly to have 'it could be this' answers. Or get numbers to have an immediate answer and a solution. You have two choices.

I see nothing in your posted that even imply a defective hard drive.
 
OldBob
Gerbil
Topic Author
Posts: 53
Joined: Wed Dec 01, 2004 12:18 pm
Location: St. Louis

Re: PSU or MB failure, how know?

Thu Mar 03, 2011 11:02 pm

to westom

Thanks for the input. Right now it's running again and I'm set to keep it backed up (externally) automatically. As long as it keeps running I'll ignore the problems. Having the data it the important part. When it fails again I may just upgrade the whole system. A portion of the MB (usb ports) failed a couple of years ago (static discharge) and I almost torched the whole thing about three years ago (shorted the PSU to ground) so it's just a matter of time. I've obviously learned that using multiple internal drives for back up has serious limitations!
OldBob
 
westom
Gerbil
Posts: 69
Joined: Tue Jun 16, 2009 2:47 pm

Re: PSU or MB failure, how know?

Thu Mar 03, 2011 11:57 pm

OldBob wrote:
Right now it's running again and I'm set to keep it backed up (externally) automatically. As long as it keeps running I'll ignore the problems. ... I almost torched the whole thing about three years ago (shorted the PSU to ground)

Shorting a power supply to ground must not cause PSU failure. In fact, Intel specs define how thick (or thicker) a wire must be to short that supply - and have no damage. A test that every supply - even those in the original IBM PC - must pass.
Good to hear it is working.
 
Rectal Prolapse
Gerbil
Posts: 54
Joined: Fri Jul 28, 2006 11:39 am

Re: PSU or MB failure, how know?

Fri Mar 04, 2011 1:15 am

Strangely enough, almost all my hardware issues with PCs have been because of the PSU. With the exception of one case where the motherboard BIOS needed to be updated. :)

My OpenSolaris server was having trouble with hard drives on eSATA suddenly losing connection or generally acting badly. 2 weeks later it got worse and then I decided to try a different PSU and it worked again!
 
just brew it!
Administrator
Posts: 54500
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: PSU or MB failure, how know?

Fri Mar 04, 2011 9:15 am

westom wrote:
I see nothing in your posted that even imply a defective hard drive.

I noted that it was a long shot. But I have seen at least one instance where a fried hard drive was shorting out the +12V rail and preventing the system from powering up.

westom wrote:
Shorting a power supply to ground must not cause PSU failure. In fact, Intel specs define how thick (or thicker) a wire must be to short that supply - and have no damage. A test that every supply - even those in the original IBM PC - must pass.

I agree in theory; but in practice this is overly optimistic, especially when dealing with cheap PSUs. I have seen PSUs fail catastrophically when one of the rails is shorted or overloaded. We're talking exploding MOSFETs in the PSU's switching regulators, sparks, and smoke.
Nostalgia isn't what it used to be.
 
westom
Gerbil
Posts: 69
Joined: Tue Jun 16, 2009 2:47 pm

Re: PSU or MB failure, how know?

Fri Mar 04, 2011 8:45 pm

just brew it! wrote:
I agree in theory; but in practice this is overly optimistic, especially when dealing with cheap PSUs. I have seen PSUs fail catastrophically when one of the rails is shorted or overloaded.

If shorting a power supply causes any damage, then locate the reason for damage. Education and knowledge of the purchaser.

Decades before PCs existed, a power supply was shorted without damage. That old, that required, that standard, and that well understood. A problem when so many computer ‘experts’ do not even know how electricity works. An example of why some domestic industries move overseas where technical knowledge is better.

For example, no power supply dumped into this market need meet UL, CEC, Energy Star, EU Code of Conduct, VDE, CSA, or any other standards. Meeting obligations lie 100% on a computer assembler. Who typically has no electrical knowledge (no electrical knowledge is required to pass an A+ Certified Computer exam). So uniformed as to not even know those obligations are 100% on him. A market ripe for dumping. One test that any consumer might do before installing a supply is to power it up shorted. If it fails, his retailer should admit to selling a defective product.

No load must damage any supply. That one sided and obvious. And not just big supplies. All single chip power supplies routinely contain that protection. Damage from shorting was that much unacceptable so many generations ago when the first single chip power supply was marketed. Saying it can happen is completely unacceptable. But many are educated by observation rather than generations of well proven knowledge.

Any supply that fails when shorted means the person who installed that supply should be looking at the failure in a mirror. Another reason why brand name machines tend to be more robust and reliable. And why the Silicon Valley must find 60% of its new employees from better educated immigrants. Its not theory. Reality complete with the most common reason for that failure – a technician with insufficient knowledge and education.
 
just brew it!
Administrator
Posts: 54500
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: PSU or MB failure, how know?

Fri Mar 04, 2011 9:10 pm

Again, theory vs. practice.

Yes, the official ATX PSU spec says a short or overload cannot cause a failure. Some PSUs don't meet spec; that's a fact of life. Just because it isn't supposed to happen doesn't mean it doesn't happen. We, as consumers who assemble our own PCs, need to be cognizant of that, and vote with our wallets when there is evidence that a vendor is screwing us over. This will tend to weed out the dodgy vendors over time, but will never completely eliminate them.
Nostalgia isn't what it used to be.
 
westom
Gerbil
Posts: 69
Joined: Tue Jun 16, 2009 2:47 pm

Re: PSU or MB failure, how know?

Fri Mar 04, 2011 10:03 pm

just brew it! wrote:
Again, theory vs. practice.

Reality standard for so many generations that short circuit damage says a human is dumb stupid or in denial. Every layman can do the test. A supply that fails is immediately bought back by the retailer. A test so easy that no 100% defective supply should be in service. Unfortunately, we often call them cost controllers, bean counters, spin doctors, cheapskates, finance people, or uneducated. That is the reason why a shorted supply destroys itself.

All computer assemblers must meet FCC, UL, and other standards. Even most A+ Certified Computer techs don’t know that. And have virtually no electrical knowledge. So few are ‘voting’ responsibly. The domestic market remains ripe with scam supplies. A problem completely traceable to a person – not to hardware.

If a supply does not sell for at least $60, then one immediately has reason to suspect. That does not say that a $60 supply is sufficient. The second sentence so important because so many computer techs have insufficient technical knowledge. Will assume the first sentence says a $60+ supply is sufficient.

If a supply does not come with a long list of numeric specs, then an informed consumer immediately suspects the worst. ‘Voting with your wallet" is what free markets are supposed to do. But only if the consumer demands facts with numbers. Unfortunately too many computer 'experts' are so technically naive as to encourage all layman to be uninformed. And to stay uninformed.

Theory is ‘voting with a wallet’. It should work - in theory. Reality is what hardware must always do. Every layman can identify 100% defective supplies that still boot a computer. Simply short its outputs. If the supply fails, then it was always defective. And may also destroy other computer parts – motherboard, disk drives, etc. That is not theory. A requirement that most computer techs do not even know to, instead, dump inferior computers on naïve consumers. Who then deny they – not hardware – are why failures exist. May even deny reality using bogus ‘theory’ accusations.

The market has inferior power supplies because most computer techs do not even know how electricity works. Defective hardware directly traceable to that human failure.
 
OldBob
Gerbil
Topic Author
Posts: 53
Joined: Wed Dec 01, 2004 12:18 pm
Location: St. Louis

Re: PSU or MB failure, how know?

Sat Mar 05, 2011 1:52 pm

OK ... re : problem with the shorting the PSU a few years ago ... it was the smoke emanating from the case that gave it away. Leads to a floppy and its connector at the drive were burnt as was a nearby spare hd lead,. The power supply seemed fine afterward.

Meanwhile, the original problem returned: another big storm coming through last evening prompted me to shut it down and remove all conducting connections to the rest of the world (this was also an experiment). I had all the files backed up to an external drive by that point. This morning similar problem: a second of power then it quit. It would only do that much if I unplugged it AND held the on/off switch and reset buttons in for about 20 seconds. After several tries it actually started (normal POST) , but it had a blank screen after the Windows screen (before sign in). I reset it and an error screen came up suggesting I do a safe start. Which I did. Device manager showed no malfunctioning hardware. After a few tries I did a system restore (one of Windows options) to late the afternoon before shutting down for the storm. It restarted itself after the restore and is now (apparently) running fine.

My guess is that there is something wrong with the MB (vs PSU). Opinions solicited.

Also, if I go for a new machine, will Windows 7 read the files on the external drive (Seagate)?

Would an Apple read them?
OldBob
 
westom
Gerbil
Posts: 69
Joined: Tue Jun 16, 2009 2:47 pm

Re: PSU or MB failure, how know?

Sat Mar 05, 2011 6:56 pm

OldBob wrote:
This morning similar problem: a second of power then it quit. It would only do that much if I unplugged it AND held the on/off switch and reset buttons in for about 20 seconds.

You ignored everything about the power controller. If a controller finds a defect, it powers off the system. Exactly as you see. Exactly. Power controller is doing exactly what it is supposed to do. How many times was that posted and you ignored it?

Posted was how to have an immediate solution. Obtain "numbers in one minute from six wires with a multimeter (a $17 tool in Wal-Mart). The reply provides an immediate and complete answer in a next post. Your choice. Keep posting repeatedly to have 'it could be this' answers. Or get numbers to have an immediate answer and a solution. You have two choices."

Why complain about something you refuse to solve? Why post while ignoring the simplest and least expensive solution? Obviously, a power controller is doing exactly what it is suppose to do. You have zero reasons to suspect only a motherboard. Buy or borrow a tool sold in stores that also sell something more expensive – a hammer. Learn how to have a useful answer by doing a minute of labor. Or trash the entire system without further comment. Do you want to fix it? Or just complain while entertaining your fears?

Who is online

Users browsing this forum: No registered users and 1 guest
GZIP: On