So I'm playing a little game this afternoon (World of Warcraft if you must
know) and all of a sudden withut warning or anything my computer shuts off.
No power.  I push the power button; Nothing.  Nada.  I begin to become a
bit concerned.  I assumed it was the PSU (ThermalTake Purepower 420W, more
on hardware details shortly), so I took it out and put in an older one only
plugging in the hard drives.  Nothing.  I took it to a different plug in
another part of the house.  Nothing.  I changed power cables.  Nothing.  I
am now very concerned.  I'm thinking this means a motherboard failure, no?
And so my question is, what are the chances that other parts of my box were
fried during this crash.  It wasn't an electrical spike; there were two
other computers running in the same room (1 desktop; 1 laptop plugged in)
and there was no disruption with them. 

The computer is: P4-2Ghz, 512MB ram, ECS P4S5A Motherboard, ThermalTake
Purepower 420W PSU, Chaintech nVidia FX 5500-256MB, 2 Maxtor HDs (40G,
160G), Floppy, 1 Sony DVD-Burner, 1 Creative DVD-ROM, 1 NIC, 1 Creative
Soundcard (3+ years old).  I leave the box on pretty much all the time, but
the actual case is smaller and airflow was probably a problem.  That's why
I recently finally bought a new case designed for airflow; the only fans
were in the PSU and one of those PCI-slot Cyclone blowers.  But with all
the flat IDE cables and all the wires from the PSU, there was definitely
congestion in that poor case.  Could it have just finally overheated?  I
haven't had reset problems, as in the CPU overheating.  Whenever I checked
the Hardware Monitor in the BIOS the CPU ran at about 114F and the "System"
at about 70F.  I'm prepared for the castigation for mistreating my computer
so poorly.  But my main concern is that other parts (hard drives, gfx card)
were damaged in the failure.  I fortunately just ordered a new MB and CPU a
couple of days ago, so I will shortly be able to find out, but I just
thought I would relay this tale so that anyone who had any
advice/admonition to offer me would have their chance :).

Sorry for the length.  I probably still left some stuff out.  Thanks for any
information you can provide concerning how I operated and had the computer
set up.  Was the graphics card too much for the MB?  That Chaintech is only
about 3 weeks old; previously I had a Jaton nVidia 3DForce-2 MX 400 (64MB).


Thanks,
jab3

On Sun, 27 Mar 2005 20:28:26 -0500, jab3
<jab3@_del_linuxmail._del_org> wrote:

"Nothing" as-in, no signs of life at all, no LEDs light and
the fans don't spin?


It would tend to mean a short somewhere if "nothing" really
means "nothing" above.  Motherboard is prime candidate but
you'll need to pull parts and do it by process of
elimination.  Strip system down to bare minimums for POSTing
(CPU, video, 1 memory module), disconnecting EVERYTHING
else.   Keep in mind that you may need to disconnect AC from
power supply for a few seconds between each test to reset
it.


Depends a lot on what's happened.  Typical failure by a
shorting capacitor will just cause a shutdown with no
damage, usually.

Very small cases can have problems with cables, but
generally the overall fan and unimpeded intake, exhaust area
is much more important.

Continually running at high temps will tend to wear out
parts faster though without knowing the cause yet, we can't
know if this played any part in the demise of ??? yet.
Examine motherboard for failed capacitors... swollen vented
leaky residue on top or bottom.


Probably not, BUT the same (potential) overheating that
"could" have caused a failure in (whatever part may've
failed for this reason) will likewise cause similar
reduction in lifespan to other parts.  This specific event
may not have killed multiple parts but the circumstances
leading up to the event may be progressively deteriorating
other parts' function.


Best advice is to buy a case ahead of time, taking your time
to mod it if necessary.  Better too good an airflow for the
parts you'd use then have reduce fan RPM (making system
quieter, less dusty and longer fan lifespan) than to have to
deal with problems later and strip down system to combat the
issue.


After removing all parts non-essential to posting, also try
clearing CMOS.

kony wrote:

Show quoteHide quoteNothing as in Not a thing.  No attempt at any life at all.  As though there
were nothing plugged in the case at all. 

[ snipped most of rest ]

Hope you didn't mind the snip.  I disconnected everything but the CPU,
video, memory.  Again, nothing as in nothing.  So I decided to just pull
the motherboard out and look at it.  I actually did see swollen tops to 3
or 4 capacitors.  1-2 near the CPU, 2 near the parallel connection.  I even
took the MB in to the other room to show my wife, pointed out what a
capacitor was and asked if she noticed anything about any of them.  She
almost immediately pointed out the same capacitors and said their tops were
bulging and the others weren't.  There is no residue anywhere; but the tops
are definitely swollen.  So I'm assuming that the MB did just fail.  Again,
do you think this a heating/overloaded issue or just faulty parts (this MB
is a little over 2 years old; again, ECS P4S5A from Elitegroup)? I remember
reading about this problem with some ABit boards and a stolen formula, but
it seems odd that this happened abruptly after 2 years.  I've never had any
configuration/lost CMOS data problems or strange reboots.  Can 3-4
capacitors really cause it to just flat out die? :)  I know that probably
is a silly and ignorant question; it's just the more I think about it the
more it seems crazy that those 3 little parts can take out the whole MB in
a flash. :)


Thanks for all your help and patience,
jab3

jab3 <jab3@_del_linuxmail._del_org> wrote in news:Yo-dnXwC4bRw4NrfRVn-
3*@comcast.com:

one of those 'kick myself in the head' thoughts that sometimes happens to
me. But does the outlet work? the cord?

I sometimes tend to skip things like that and dont even consider it could
be the problem ;o)

Kev wrote:

Yeah I tried a different outlet and cord.  Nada.  Something is definitely
blown.  (Also tried 2 different PSUs)


-jab3

On Sun, 27 Mar 2005 22:45:48 -0500, jab3
<jab3@_del_linuxmail._del_org> wrote:


Nope, 'tis good to keep redundancy to a minimum.



Well technically the caps had probably "failed" to operate
within their specs long ago (IF they ever did), it's just
that they finally degraded enough that they shorted out.


Yes, no, and maybe.  They're usually rated for 105C, but
it'd be a bad idea to let them run at even 2/3 of that temp.
I try to keep 'em cool enough they don't feel more than
mildly warm after/during an extended full load stress test,
which in some cases/situations <cough>overclocking</cough>
even means a small fan dedicated to, directly overtop of the
regulation/capacitor area on the board.  For example,
http://69.36.189.159/usr_1034/nb_sinks/1.jpg
allows very high overclock with fans still nearly inaudible.

Hot-running caps fail sooner, there is a clear lifespan
reduction based on temp (many manufactuers spec 50% life for
each 10C temp rise).  BUT, ECS has been suspected (a gentle
way of putting it) of cutting corners and using inferior
(relative to other brands) caps, numbers of them and the
whole circuit might cause more ripple too, heating caps
further than they otherwise would.  "Defective" designs can
include caps that aren't defective, or are poor, or are
defective in addition to the board design, then compounding
that with poor case airflow... well the combination will
tend to make them die sooner, and any of those variables you
change would tend to promote longer life (ignoring whether
the caps were specifically defective as individual
components failing to meet THEIR manufacturer's specs).



Abit's problems were on earlier socket 370/slot1 boards,
these days Abit's caps are among the best, Abit is not
longer any more likely to have bad caps than any of the
other major brands.

What brand of caps were yours, including color, voltage, mfd
rating?  The most common failures I've seen on ECS boards
were G-Luxons which were typically turquoise (w/silver
stripe?) and sometimes (ofter older boards) black w/gold
stripe.   Tayeh and GSC have been notably problematic too,
in cases not particularly hot nor any other obvious problems
external to the caps themselves.



Sure, even 1 could.  To you it will seem like "a flash" but
they were probably swollen for a while already.  That your
system "may've" ran a little on the hot side could have
actaully made them perform a little better than they
otherwise would've, more stable for a time- and yet still
fail all that much sooner.

Show quote Hide quote Hey men....... all this talk about caps got me to wondering. It's a
routine repair to replace caps in tv's, I do at least half a dozen
everyday. I have yet to troubleshoot a mb at the component level but is
there something that makes cap replacement impossible or especially
difficult in this setting or are the right type of caps just hard to
come by?

Newfdog

On Mon, 28 Mar 2005 10:17:30 GMT, Newfdog
<newfdog2nospam@warp.nfld.net> wrote:


Well ideally you need a hot iron (higher wattage relative to
many fine electronics soldering due to large copper area
around the caps) with a small tip, steady hand... seems
pretty simple to someone used to soldering.  Solder pump or
sucker is nice to have too.    Even though these are
multi-layers boards, the problem caps are almost always on
plated through-board holes so the multiple layers aren't
much of an issue. 

Video cards are another story, much more likely to be
surface mount.  IIRC some of the newer Intel boards are also
more likely to have surface-mount caps, though I haven't
really studied any of them with an eye towards how hard it'd
be to replace those caps amidst surrounding components.

Otherwise the caps themselves aren't difficult to come by if
you don't insist on particular brands but they're
particularly low-impedance, low ESR relative to most
electrolytics.  Most of the major name-brands make suitable
families, Panasonic FM or Rubycon MBZ series come to mind
but are certainly not the only candidates.  Digikey sells
small quantities of the Panasonic FMs but if you do the
volume maybe you can justify >1000 of them, leaving you more
purchasing options from manufacturers or distributers.

Thanks for the info kony.

Newfdog
kony wrote:
Show quoteHide quote

Newfdog wrote:
A couple of factors.  Most users with the problem never saw a hot
soldering iron in their lives, let alone solder wick.  Even if they
did, MBs are likely to have more layers and be more delicate than
TV boards.

CBFalconer <cbfalco***@yahoo.com> wrote in
for OP, there are places that will do it for you, and use good quality
parts. But in your case it might be better simply to go get a new one,
unless you really need that one in particular. Doesnt have to be big time
if you dont need it. But better in the long run.

Kev wrote:

I'll keep that in mind.  The only thing that I would want this board for is
to get it running one more time so I can get into Windows and uninstall the
system/cpu drivers for the new MB/CPU I'm about to put in, hoping that
Windows XP will recognize the new set and load drivers without me having to
reinstall, but I'm doubtful that would work anyway.  Last time I changed
MB/CPU Windows 98 wouldn't even begin to load.  Of course Linux didn't skip
a beat, but that's a different story.  (I'm interested to see how Linux
handles a totally different architecture (AMD-64), though my previous
upgrade from PII to P4 was pretty much totally different architecture)  And
unfortunately I need Windows to play this damn game I've become addicted
to. :)

The whole thing is moot though because the new board comes in tomorrow, so
we'll soon see what happens.

Thanks,
jab3

jab3 <jab3@_del_linuxmail._del_org> wrote in
I have no way of knowing, but maybe it can be done from Linux... I imagine
brain surgery would be easier tho ;o)

kony wrote:

Show quoteHide quoteThat's an interesting placement for a fan.  Never seen that before.  Though
I doubt I stress my components enough to need it. :)  Maybe one day though;
I'll file it in the back of my brain.

Show quoteHide quoteYeah I shouldn't have bought the ECS in the first place; it was a couple
years ago, and at the time I couldn't resist the price (I was upgrading a
400mhz PII; I was dying :)).  Gotta learn somehow. :)  I just hoped it
would work, and it did work up to 2 days before getting a new MB,
so.....It's just that I was going to use this MB/CPU as another
server/something.  I'll have to get a new MB sometime soon now.  Not the
end of the world.

Cool.  I went ahead and bought the Abit assuming that something like what
you said was the case, since I've heard so may great things about Abit.

Yes, they are in fact black with gold stripe, and like I said this board is
over 2 years old with me - who knows when manufactured. 

Show quoteHide quoteThat's interesting (the running a little better because of heat but dying
sooner), and I've now learned my lesson.  :)  (I'm hardheaded and unwise;
have to learn myself)  But thanks for all the info and help and advice
Kony.  You've been most helpful and informative.  Now I'll just have to sit
back and wait on FedEx to deliver the new AMD/Abit and see how that works
with the ThermalTake PSU. :)  But this new case I bought is designed for
cooling, so hopefully that won't be as much an issue now. (and it's much
larger than the old one)


Thanks again,
jab3

On Sun, 27 Mar 2005 20:28:26 -0500, jab3
<jab3@_del_linuxmail._del_org> wrote:

Your card isnt exactly cutting edge and the other specs -- I really
doubt you had an overheating problem since you say you checked the CPU
and system  temps which are pretty low. Its also pretty rare that a
board just goes bad , unless it has bad caps or theres some power
surge etc.  Theres a chance your system may be OK since your PS seems
decent etc. The only questionable part is you say your MB is an ECS
---- a really cheapo brand. Though many people post that theyve bought
the FRYs ECS combo specials and had no problems , Ive seen others say
they had fairly high problems with ECS so the odds of it being a bad
board are probably a lot higher.

J***@Smith.com wrote:

Show quoteHide quoteYeah, I was afraid of the ECS part - that's why I mentioned the brand.  I
bought that a couple years ago.  As I told Kony, there _are_ some swollen
capacitors so....y/our suspicions may be correct.  This time I went with
ABit.  Even though they may have had a capacitor problem also, hopefully
that's been taken care of. :) 


Thanks,
jab3

jab3 <jab3@_del_linuxmail._del_org> wrote in news:Yo-dnX8C4bQfHdrfRVn-
3*@comcast.com:

yes it was a problem, had quite a few on mine recapped. except for an
occational hiccup when the thing just wont start (needing a quick turnoff
and turnon) no problems. But... when they were bad it wasnt a complete
shutdown, but RAM errors. Guess it all depends on where they are at. I have
been told that a cap that is swollen doesnt mean that it has already
failed. But its not at %100 either.

I guess if you can try it in another case (just in case its a switch
problem, or one of its wires) to see if it works. Or if you have another
board put that in the case you have and see if it even turns on.

Does LinkSys BEFSX41Router/Switch Work As a DNS Server?
External Firewire/USB HD drive enclosure problem
Computer Setup Question-2
Socket 939 Athlon 64 3500+, Newcastle, Winchester, Clawhammer. What is the difference between them?
Using 2 battery backups
Where can I get?
help, please- broken pin on HD...
Gigabyte GA-K8NS (socket 754) USB?
Wireless keyboard and mouse woes, XP Pro SP2.
b-grade monitor?