Link to home
Start Free TrialLog in
Avatar of crnz
crnz

asked on

PC keeps shutting down by itself!

Ok, I'm ready for some lateral thinking from my fellow EE-ers, because I've lost my sense of humour about this.

Situation: after a power related computer fault (no longer powers up), I replaced the motherboard, CPU, memory and hard drive.  Put Windows XP SBE 2003 back on the drive, put the data on and gave it back to the customer.

Once there, it started shutting down at about 1.5 to 2.5 hour intervals.  Here are some thoughts and what I did about them

1. overheating.  PC Health monitoring shows LGA775 never gets over 45 degrees C, still took off sides of computer and moved it away from other servers to have clear air around it.  NOT SOLVED

2. power supply.  Installed a new Belkin surge adapter, a brand new 400W PSU.  NOT SOLVED

3. The only parts no longer original are the case and the CDRW drive.  I unplugged the CDRW drive data cable and power cable.  NOT SOLVED

At this stage, apart from the case, everything is new.  

XP SBE complains about the server going down unexpectedly - so currently it doesn't seem to be XP itself doing it.  Syslogs don't show any errors apart from trying to connect to a PC on the network that's not there right now and retrying occasionally.  I still have some suspicions about that as the possible source, but for the moment I want to eliminate hardware as a cause.

I'm about to go and swap the new memory for another new one.  I suspect this will also won't fix it.

Which then leaves me with

- change new motherboard for another new one
- change new CPU (2.6GHz LGA 775 Pentium Celerom) for another new one
- change new hard drive for another new one

But by then there is NOTHING original but the case, and the motherboard, CPU and hard drive are only 1 week old now, so am only swapping them in case the fault lies with them.

Any brainy bod out there able to assist with some lateral thinking as to what is happeing here?  This is new to me.  

Unless the problem lies with XP SBE Server 2003.  I wish I had more detail on the error it is showing to hand, but don't right now - and that's probably an OS question not for this area.


Avatar of AndreDekolta
AndreDekolta

Have you tried re-seating the memory?  Try removing modules.  I had this happen to my PC and finally wouldn't start until I replaced a bad memory module.  Just a thought...hope you don't have to replace the MB!
Avatar of crnz

ASKER

Thanks - We'll find out today if it is the memory.  I'm about to go there within an hour to swap out the memory for a brand new module.  I very much hope this fixes it because I am not looking forward to swapping out the motherboard or having to set up a whole new HDD with SBE2003 :)  As this server is serving an accountancy practice, we have the added joy of the pressure of having people unproductive.  Maybe my luck will finally catch me up today :)
Hi crnz,
try these steps:
1) to eliminate the possible cause of hardware malfunction use a reference system,like Knoppix:
http://www.knopper.net/knoppix/index-en.html
boot the system from your master optical device and check out the boot process,if you can 'catch' the KDE desktop ,then work for about couple of hours under the KDE and check out if the problem can still resist...
1.1)regarding memory checkup,use the memtest 86+ tool:
http://www.memtest.org/ -->this is really strong utility regarding memory stress and check up ,test the ram for about 5~6 hours!
1.2)check out the motherboards manual for special jumper settings(could be you must once more time reset the BIOS,but don't be in a hurry of doing this) and special memory configutration,aka just study carefully the motherboards manual+ the graphics card manual,what kind of graphics card do you use?
2)check out the windows installation:
what about the energy settings of the system,the ACPI management,check out if there is something about the hardware going into suspend mode,set these features to never...
2.1)there are also a setting in the system under the advanced tab regarding the restarting behaveour of the pc,unmatch the setting 'restart automatically after crash':
http://www.windowsdevcenter.com/pub/a/windows/2004/06/08/stop_errors.html ,perhaps you're getting the whole time BSODs and the PC restarts automatically after it,so if this is an issue ,
write down the 'STOP:' error and we will se furthur(as you said that the event-log is clean ,i'd rather assume that there are no kind of BSODs ,but let's see) ,perhaps firstly a link/links:
http://www.aumha.org/win5/kbestop.php
http://www.microsoft.com/downloads/details.aspx?FamilyID=859637b4-85f1-4215-b7d0-25f32057921c&DisplayLang=en


Cheers!
Avatar of Mark Poirier
Just a longshot, but maybe it is the case. Shorting reset or start buttons,or such. Have you tried it outside the case, just in CASE. Sorry couldn't help it.
aslo a good suggestion,but why should it run for about 1~1,5 hours till it gets the shortcut;anyway it's worth to be tested!
everything out of the case,mobo on a thick cardboad,right wiring(+ once more time check-up) ,jumper settings check and go ahead with the furthur tests....
Avatar of crnz

ASKER

What is somewhat interesting is that it lasts ok for a few hours.  So my (current) hope is on a bit of memory getting accessed as the server is being used and memory use increases over time.

I prefer that over the case if you don't mind :)  Be that as it may, I'l pull out any other non-critical wires such as hdd light, reset button, etc (can't do the power switch - they need it to turn it back on! ;)

Problem is that it is in a production environment and even with the 3-5 reboots a day they need it to be there.  However, if the next few steps don't solve it, we'll limp towards the weekend and I'll take it off-site for a thorough seeing to. With previous suggestions about various burn-in programs and seeing how it fairs under linux, etc...

And sparkmaker, you're a hard CASE. :)
You are assuming (as I would have to begin with) that this problem has it's origin in the hardware.  Why not the OS?

You mention "XP SBE complains about the server going down unexpectedly...".  What server?  Is it the DCOM server?  Maybe the customer has a virus, like the Blaster worm and its variants, that mucks around with the DCOM server and shuts down the RPC service, causing the computer to shut itself down?

Usually the message is something like:

   This system is shutting down. Please save all work in progress and log off. Any unsaved
   changes will be lost. This shutdown was initiated by NT AUTHORITY\SYSTEM. Time
   before shutdown: 00:00:XX

   Message: Windows must now restart because the Remote Procedure Call (RPC) service
   terminated unexpectedly.

Is the customer behind a firewall?  Is the DCOM/RPC exploit closed by installing service packs?  Does the customer run an antivirus program with an up-to-date definition file?
Avatar of crnz

ASKER

Hi mjcoyne

Yes, I do assume it is hardware.  The reason is that XP SB Server 2003 asks on the first login for an explanation for the sudden loss of service.  So it's not XP shutting down, it's the hardware pulling the rug on the OS.  Or so I still believe.

The server was freshly loaded with all the updates and protections, so a virus appears unlikely at this stage.  Also, the message log has no entries relating to this event.  It's just business as usual and then as if someome pulled the power lead from the back.

Behind firewall?  Yes.  All Service Packs installed.  Antivirus in place.  
Avatar of crnz

ASKER

Ok, read the motherboard manual - no stiches or jumpers relating to memory types, sizes or CPU speeds, voltages that appeared relevant to me - but thanks for the suggestion.

Been there to put in a new memory module and swap the other one out.  This is where I managed to pick up a little hope...  There are two memory slots, and the module was in DDR1 with DDR0 empty.  I have replaced it by putting the new one (by 1 week ;) in the DDR0 slot.  Other than the fact the module may have a fault, I also have hope that its position may have been possibly the cause of trouble.

Anyway.  It's a waiting game now.  See if it stays up until the end of the day to start with.


Yeah, I guess the most likely culprits are RAM and heat -- but it looked from your initial post that you'd ruled them out.  DDR0 versus DDR1 slot swap might be it...  If not, I guess we have to look more closely at heat (and the hard drive).  Could it be a bad IDE or SATA cable?  Doesn't seem like this would lead to a power shutdown...

Are the rest of your components integrated (network, audio, video), or are these cards?
Avatar of crnz

ASKER

Heat won't be it.  It's running at near-room temp now with the case open.  None of the surfaces feel warm, the heat sinks are slightly warm to the touch and the CPU temp is about 43 degrees C.  

Left to check are the new motherboard and the new hard drive.

I do take your cable suggestion on board - I reused the original, so a fresh one is called for when I visit next.  

Luckily the client is in good spirits - the result of some goodwill in the bank no doubt, but this has to come to an end sooner or later or it will get ugly :)
what's the make and model of the motherboard?
Avatar of crnz

ASKER

ECS 661GX-M7

Just had a call - no fail since putting in the new module and putting it into the DDR0 slot.

One more day without fails and you're all invited to a party at my place (air fares not included, sorry :)
when then..party at my place in Hawaii!  Lot's of mai tai''s and beautiful women.

aside from that.. I would suggest for a server platform that you use an Intel board.
Avatar of crnz

ASKER

irwinpks... the customer concerned is extremely price sensitive (a hint as to the industry concerned is embedded in previous posts :).  My preferred brand is Asus or Gigabyte as a close 2nd.  Intel don't have that good a presence here, and are more expensive than any other top board by 50-70%.  Anyway.  Server.  That's a joke.  All it does, and I mean ALL it does is share 3 folders.  That's it.  And for that they purchased (or were to told by the previous - now dumped) service company to buy XP Small Business Server.  The previous hardware succumbed to a power related event, and when I opened it up I found a motherboard with ISA slots on it.  I asked when they bought it.  They told me it was bought new 18 months earlier.  When I expressed my surprise about the apparent age of the motherboard, they kinda admitted that they "probably got what they paid for".  So trust me, the ECS is a major victory and a substantial upgrade for these people ;-)

As for the mai tais... I guess the more I have, the more beautiful the women are! :)   And Mrs crnz sisn't going to permit me to be anywhere that sort of a party anyway ;)

Just read your profile.  First program I sold was a IBM PC text-character-only version of Loadrunner, as at that time a Hercules graphics card was beyond the budget of most.  Acutally got some shops to accept it on indent and sold a few!  

Kia ora from New Zealand.
"And Mrs crnz sisn't going to permit me to be anywhere that sort of a party anyway ;)"

Hey...don't bring sand to the beach and it will all be good ;-)

I had a Hercules at one time...it worked with my 10mb MFM Seagate drive, but we have to go back a little further for me when APPLE-TRS-80 and Commodore PET computers were abundant.

"ooh"

Aloha from Hawaii!
Avatar of crnz

ASKER

My first hard drive was a 10MB Plus "hard card" that was a 10MB hard drive attached to its controller so it fit into the expansion slot as one unit.  I was SO proud.  I remember doing "dir" over and over again to see the free drive space expressed in bytes.  That was a BIG number!!! My 2nd was a Seagate 5MB 5.25" FULL height boat anchor.  I /so/ wish I kept that one.  I still have an original IBM PC motherboard with ram and cpu for posterity (I intend to frame it and put it behind glass for display), but that 5MB Seagate bugs me every time I think of it :)  And you must remember RLLing the Seagate (or was it Maxtor?) 80MB MFM drive to 120's and feeling you cheated the system?  Who cared about the bad sectors!  You got 49% more space free!!! :)

Roger on the sand to the beach thing.  I'm just too well trained I guess :)
"I'm just too well trained I guess :)"

uh..."whipped" :X

:-)

Well before that I had a TRS-80 Level II computer with 16mb of ram and CASSETTE tape recorder at 300 baud.
Avatar of crnz

ASKER

Still remember the joy of going from a 300 baud acoustic coupler to a 1200/75 baud modem.  Xmodem suddenly FLEW along at a 400% increase in speed!  Definitely one of the defining moments of my personal computer history :)  Others?  Getting a 3% speed increase by usinga V20 CPU instead of a 8088.  Logging into a computer in Finland in 1989 and FTPing some Turbo Pascal source code.  Ok, back to work.

Yeah, whipped.  No doubt about it.  But Mrs crnz v2.0 is a serious upgrade on v1.0, and as a result I am quite happy to stay true to the cause :)
"But Mrs crnz v2.0 is a serious upgrade on v1.0, and as a result I am quite happy to stay true to the cause :)"

Ha!!..

Well Mrs. Hawaiian Superman v1.0 sucked...but there were other fulfilling qualities.  However, after 16 years, I've grown out of the lackluster performance of old software.  Currently I'm not seeking Mrs. HS v2 as I'm enjoying the prospect of leasing girlfriends and at v27, I quite haven't found a flawless model. Though GFv27 has been good to me for about 2 years and shows promise to become Mrs. HS v2, the thought of no longer leasing is going to be hard to bare a longer term of committment.  GFv27 was once wife v1.0, and from her position, I'm BF v3.0.

Life is good, and I'm virus free. ;-)
So it was the memory as I suggested?  Cool, well not.  But I'm glad it's working out!
just in case your problems is not solved yet (i hope it is), did you consider having 2 bad parts? In this case, i suggest testing the parts on another known good PC, like disks, ram, cd drive, cards etc...
Hey crnz & irwinpks, great dialogue there.
@crnz--> I understand your situaton with my CASE study here, what with it being in use and all ,But, The only empty CASE I like  is a beer case...You've probably nailed it with the ram slot anyway.
And if you like parties come on up to western Cape Breton in the summer time( in winter we switch to Rum for the antifreeze effect and tend to get a little wooly)
Interestingly enough we have had an outmigration to New Zealand in the 1800's from here but I don't believe we got to Hawaii yet as we seem to get to Toronto and no further.(We Capers get homesick easily)
As for my computer experience, I believe I used an Abacus one time there, way back,and not very well. Wait,I think that was just a necklace one of my old GF's had that I liked to play with, never seemed to get past the count of 2 for some reason  ;-)
  I kind of fell into this computer stuff when I was building my telescope and needed to get info on it. Pulled a bonehead move by deleting the user file in Win98 one day when using file checker(Now a known problem in the MS knowledgebase) And ended up with a blank monitor screen. Had to get some kid to fix it, and said never again.

Enough of that,back to your problem crnz. Have you inspected the contacts on the Ram for any discoloring or maybe the slot itself needs a bit of housecleaning. Is it dusty inside that CASE? couldn't help myself.
@sparkmaker...I'm not geographically adept...where is Cape Breton?

Oh yeah... I used rocks to count
Eastern Canada http://en.wikipedia.org/wiki/Cape_Breton_Island 
Hey, us islanders gotta stick together,eh.
This may be more helpful http://www.unb.ca/csd/nbpei/provmap.html We're just northeast of Maine.
cool..... here is where I am at in Hawai...courtesy of Google Earth
http://www.ctrlaltdel-usa.net/yabb_2_1_superman/YaBB.pl?num=1132409419
Hey thats cool. Can't get high res where I am on Google earth. We're so far out of the way the only Satellites we see are  old Dodges.
bummer...did you see the one with Eva Longoria for Maxim?
http://maximonline.com/maximusa/Maxim_Cover.kml
Avatar of crnz

ASKER

It's only just gone 7:50am so I'm calling at 8:30 to see if it survived the night.

nobus: TWO faulty parts?  Geez, you're the life of the party, aren't you :)  Let's cross that bridge when we come to it, especially since everything in there is now straight from the stock shelf, and some of it twice so...

As for dust and discolouring - these parts haven't even had a chance to see dust.  To them dust is the Great Myth they were told about while still in their anti-static packaging.  It was used to make sure they ate their dinner or "else we'll tell the Dust about it and it will come and get you".

As for Google Earth - the best you can see about where I live is about at 1:50,000 resolution.  That's what happens when you are of "no strategic interest" to the US military.  A good thing, I suppose!

Oh, and I learned how to count by the number of times I was clubbed over the head and dragged back into the cave!
Avatar of crnz

ASKER

Update:  gone down overnight.  <sob>

I'll get them to limp along until Friday afternoon and take it off-site over the weekend.

Bugger.

As they say here.
<insert>
Intel
</insert>
Avatar of crnz

ASKER

I am inclinded to agree!
Avatar of crnz

ASKER

inclined even.
argh.

Avatar of crnz

ASKER

Ok, here's a question to you experienced installers (supposedly like myself - only better ;)

When you put the motherboard in the case for the first time, do you just put it in and screw the motherboard in place, or do you do something else first?

Reason is - and I didn't want to muddy the waters by revealing this earlier - I have another system with the same model brand new motherboard, and that was refusing to boot up consistently.  Goodness knows, it may in fact have shut down as well if left on long enough.

I decided in the end that the motherboard may be shorting, so I placed insulation tape over the little metal whatyemecallits that the motherboard sits on, and then screwed through it.  That seems to have solved the "occasionally won't boot" problem.  So that could mean that the problem this topic is about may have a similar cause.

Which got me to thinking - I have never needed to prepare the whatyemecallits with insulation tape in the past.  Have I just been dumb lukcy or have I struck a brand/model here that's extremely sensitive to shorts from the whatyemecallits?

Do you prepare the motherboard or case in any insulating way before you put a motherboard in?

Before you prepare it in an insulting way by putting in an ECS of course ;)

And of all the boards I have ever taken out of cases assembled by others, especially the official ones like HP, Dell, Packard-Bell, etc, they never had any either.

Yet it seems to make sense to me that as a preparatory step, a few small pieces of insulating tape wouldn't be a bad thing to do as part of a standard install.  After all, the board sits hard on the whatyemecallits and it's just begging to short out that way.

(Why am I trying ECS?  Well, because we can't buy socket A and socket 478 boards here anymore, so when we get a fail for those, the new board also drives a new CPU, and sometimes new DDR instead of the SDRAM they have.  If the repair job gets too expensive, people will just buy new.  We don't sell new - part of our philosophy - we just fix.  So to ensure we get the fix job we need to come in with cheap parts - that's a 2.6GHz Celeron P4 (60 USD) on a ECS board (40 USD) and that's before charging for time.  The ECA board has on-board VGA as well as AGP, PATA as well as SATA, so it's a great entry level repair board in principle... :)

whatyemecallits.... capstans?

And before you are amazed at the prices we pay - we're an island nation.  Everything is more expensive because it needs to be brought in via expensive boats and/or planes :)

Ok, back to work (while you all sleep tight :)
ASKER CERTIFIED SOLUTION
Avatar of Irwin Santos
Irwin Santos
Flag of United States of America image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Whatchamacallits-->standoffs<--
When you get a new motherboard they sometimes add insulating washers made of what looks to be paper and are a primer red color. Not all manufacturers use them and I have never used them.
 Just make sure you're not tightening the screws into the board so hard they fracture it.
And its only 7:30pm here,and  I don't wanna go to bed, unless of course the wife says so.(Wink, wink , Nod, Nod, know what I mean)
Avatar of crnz

ASKER

Argh - FFS!!!!

Just found the problem on the 2nd system

Reset switch occasionally shorting the motherboard.

Not my week it is... NOT my week...

Oh hang on a cotton picking minute - swithing it around fixes it.

Does that make any sense?  it's just a shorting switch, the "polarity" has nothing to do with it.

Step away from this topic now, in case what I have is catching....
"Does that make any sense?  it's just a shorting switch, the "polarity" has nothing to do with it."
no

oh yeah RISERS are synonymous to STANDOFFS
Perhaps the movement of the wires has undone the short. You are correct on the polarity , it shouldn't matter. I'd replace that reset wire and switch just in CASE.
Avatar of crnz

ASKER

Risers and standoffs.  

Yeah, the red ring things.  As noted, not used by other manufacturers, are they.  And not always included as part of the case screw kit either.  And I was more concerned with the board touching the risers/standoffs in this case.  But that turned out to be a red herring, seeing as the short came from the reset switch.  Doh!

Heh.  We even spell whatyemecallits differently.   I guess it depends on how we pronounce it phonetically locally...
Avatar of crnz

ASKER

now, come on sparkmaker... that's a different CASE to the last CASE's CASE.  So no win for you there... yet. :)  
Looked over this question again and I'm wondering if maybe the integrated network adapter is causing the pain. The SIS chipsets have been known to have some flakey issues with gaming and such. Would it take much to disable that onboard nic and use a card in there.
My hopes are high because the reset switch is part of the case, if so ,case closed.
Avatar of crnz

ASKER

The way this is going, I may have to create a fake topic for 500 points to award both of you :)
(You didn't read that Mr/Mrs/Ms Moderator).  Which reminds me, once you pay for access, you can spread 500 points around like there isn no tomorrow.  What's THAT all about? :)  Is there sucha  thing  as point fraud? ;)  In this CASE the topic morphed into TWO topics, so I don't see that as being particulary unfair.

As for the other CASE, picking up the server at 5pm business close tomorrow (Fri) to beat it to submission over the weekend.  I'll holler with a different topic if I need y'all then :)
Avatar of crnz

ASKER

Sparkmaker, re: NIC... I'll try that over the weekend, just in CASE, thanks for the heads-up.
BTW: you're a head CASE.  ;)
cool!

....and there is point fraud... only GOD knows how many incidents there are...I wonder what I would do if someone stole my points (valued at squat).
WOOHOO!!!!

Thank you!

We should get together someplace for a virtual party...
Damn, this was ugly!  Thanks irwinpks, I leanred something here!  Cool!
Really?  as long as we party..I'm good with that!
Good work Irwin, Aloha

crnz --> Kia ora
Aloha...Malama Pono...Ahui Ho!

(loosely translated--Blessings to All!)