Sign in with
Sign up | Sign in
Your question

Shutting down without warning, driving me crazy!!

Last response: in Systems
Share
May 9, 2006 8:40:23 AM

Hi,

I have been experiencing some serious problems with my Pc for quite some time now.

My computer just shuts down without any warning. Ive tried swapping out every component apart from the motherboard.

Once it shuts down once it will tend to not turn back on again reliably untill it appears to have cooled down though it doesnt really get a chance to become hot as I am not taxing it when it shuts down.

The problem is quite random although one way to almost garauntee a shutdown is to play a dvd from the hard drive or DVD drive.

When I say it shuts down I mean its like having a powercut. No warning it just turns completely off.

I have had my graphics card replaced recently and was using a low-end card as a replacement whilst I was waiting for it to come back. Even with the other card it was shutting down although it did seem to last longer when I had that card in.

I have tried a different 580W power supply and it still had the same problem.

Does anyone have any ideas?

Thanks

Spence

MOTHERBOARD - Abit N18 SLI
CPU - Pentium D Dual Core 3.0GHz
CPU fan - Coolermaster Hyper48 Cooler
COMPOUND - Arctic Silver 5
POWER SUPPLY - Hyper HPU-4B580 Type R 580W Modular ATX2.2 PSU
MEMORY - 4 x 512Mb Corsair DDR2 XMS2-5400 675Mhz memory
GRAPHICS CARD- NVIDIA GeForce 6800GT Series GPU
BOOT DISK - SATA Western Digital Raptor 36Gig (WDC WD360GD-00FLA2)
OTHER BOOT DISKS - 2 x Western Digital Caviar SE16 250GB 2500KS SATA-II
CASE - Lian-Li PC-60 with 120mm front fan, exhaust blower fan + 2 exhaust bearing fans.
FLOPPY -1 x Floppy Drive
DVD RW - 1 x Sony DW-U14A
OP SYSTEM -Windows XP Pro SP2
May 9, 2006 9:38:44 AM

Could be a heat problem but as you have mention that it is not tasking hard to be overheated. Mmmm.

Go to the bios and give me all info about your ram like the timings and clock cycle.
May 9, 2006 9:40:21 AM

By the way if you haven't done it yet, go to the bios and set the ram clock cycle from 1T to 2T. You can do this by setting the ram controller from auto to manual or expert mode.
Related resources
May 9, 2006 9:47:31 AM

Hi Chuck!

Im not at home with my PC at the moment but I can say that I haven't changed any bios settings if that helps.

Ive also run a temp check with the UGURU utility that ships with the mobo and the system is pretty cool:

CPU =<56 when rendering 3D, =<34 normally
SYS=<40
PWR SPLY=<70 when rendering, much lower normally

The graphics card runs around 65 but I don't know if thats bad or not tho it has just been replaced so it shouldn't be a problem with the card unless its just incompatible with the mobo.
May 9, 2006 9:49:15 AM

Righto Ill try those settings when I get home thanks!
May 9, 2006 10:03:02 AM

Another thing that is worth mentioning is that last night once it had shut itself down I couldn't turn it back on for at least half an hour.

If I tried to turn it back on it would either get a few seconds through the power cycle then shutdown or the instant I touched the power switch it would seem to 'short out' and turn off instantly.

Its as if something has to cool down before it will turn back on again..
May 9, 2006 10:07:02 AM

I have an 840 with 4x512Mb Corsair DDR2 XMS2 533 and I had changed the clock cycle from 1T to 2T for it is stable that way. 4x512Mb on Intel aren't stable at 1T for reasons I don't not comprehend, something about dual channel and memory controller. Anyways when I set my ram to 1T then I get something like your pc is doing and just shutdown or restarts randomly. So you need to manually change it and it is set to 1T at default. Give me some update of you don't mind and I hope this would help.
May 9, 2006 10:14:32 AM

OK Ill let you know how it goes. Thanks a million!
May 9, 2006 10:31:23 AM

I had the same problem before with my Abit SLI S939 board. I found out that the uGuru software both in Bios and in Windows enable auto shutdown by default when it doesn't sense the rotation of some critical fans ( CPU, PW, NB etc..).

If you watch the fans rpm sensors, you will see they occasionally display as 0 rpm or very low then back up to 3000 or so. Obviously, it must be the fault in sensing.

I disable auto shut down on fan rpm, and change it to temperature only. It works for me with that setting.
May 9, 2006 10:49:15 AM

hi,

enter system properties and there goto properties of "start and restore" (I hop it´s called that way on the english version). There uncheck the option "automatic restart".

I´m not english native so I´m hoping a direct translation fits for these menus.

After disabling that option XP should not restart on errors but show you the errormessage which would be stored at the event manager anyways.
May 9, 2006 11:06:55 AM

The thing is that it is not restarting it is just shutting down although I will give it a try just in case thanks.

I havent played with the auto-shutdown features in uguru yet so I will try that too thanks guys.

The only thing that bothers me is the way it behaves once it has shut down and I attempt to turn it back on. It seems like something has to cool down before it will start up properly....could this tally with the auto-shutdown theory??

As for checking all the system temps and voltages before it cuts out I believe there is an app that ships with the mobo called Blackbox that could do it. I will look into it, thanks.
May 9, 2006 11:46:01 AM

try to look for some thermal shutdown temperature in bios. it might be set too low
May 9, 2006 11:49:55 AM

With everything else you've done, it is a strong possibility that something is wrong with the motherboard. It could have taken a jolt from the wall circuit at some point, or has a faulty capacitor on it. I've been told by an engineer friend that if you can remove it, take out the motherboard, or at least look closely, for any capacitors with bulging tops, or that have burst and are leaking electrolytic fluid.

also seem to remember that Abit had quality issues with capacitors failures only a year or so ago, so check that closely.
May 9, 2006 11:59:08 AM

I use to have this problem, but it was due to my own fault, i extended the power cable of a case fan to fit around the case, i forgot to put some insulted tape over the copper wire at the joints, and it was hitting against my sound card, which cause the computer to turn off. Maybe you have a striped wire somewhere? just a guess, you never know :roll:
May 9, 2006 11:59:13 AM

I have checked the mobo for any signs of leakages at the feet of the capacitors or any fluids that look like they shouldn't be there but alas no joy with that either.

Ive even bought a power surge plug adapter to rule out voltage spikes.

There is a good chance it could be a faulty mobo but ive just got a feeling its a config problem although I have done everything by the book in terms of installation.

My PNY 6800GT graphx card might have some compatibilty issues with the mobo I guess. I have got a few tings to try tonight anyway so I will post the results of everyones suggestions back up here tonight.

Big thanks to everyone for the help!!

Spence
May 10, 2006 8:34:35 AM

OK I ran some tests last night based on some of the advice I was given yesterday.

The first thing I did was turn off 'Shutdown Enable' in both the bios and the windows UGURU utility under fan speed and voltage monitoring.

I tested the computer by running a dvd (as I usually do) and the system still shut itself down.

The next thing I tried was changing the memory timings in the bios as Chuckshissle suggested:

I couldnt see a simple option to change from 1T to 2T so instead I changed:

Memory timings from Optimal to expert
X T (LAS) from auto to 2
X T (RCD) from auto to 2
X T (RP) from auto to 2
X T (RAS) from auto to 2
X T (RC) from auto to 2

When I rebooted the computer it made a constant beep noise and so I turned it off and restarted it and got a message saying the computer is in safe mode and I should reset CMOS, press F1 to continue.

I pressed F1 to continue. I assume when you get this message that it loads the bios fail safe defaults???

Anyway I have left it running the dvd software when I left for work so I can see if it has shut itself off when I get home later.

Does anyone have any other helpfull advice....?

Thanks

Spence
May 10, 2006 9:42:20 AM

So it's still going off after playing dvd for while huh? So it's not a memory problem then, that's scratch. Did you run a Memtest on that memory for errors? It's not likely the psu problem cuz it only happens even during a light task like playing a dvd.

So something about a task like playing dvd that makes your pc turn off. Im thinking again that it could be a heat problem. Do you happened to monitor the heat by the way using not just the software but hardware when playing dvd or any task?
May 10, 2006 10:10:09 AM

jesus .. take is easy my friend ...those timings ..are too low.....just leave them to auto......that`s not the problem....or try spd.....

try this: use memtest........if you have a linux live cd you can use it as well and type memtest in the command line.......this little program will test your memory.............

if it shows any error...in the window......you have a problem most probably with one of the memory module ..if that`s the case try them individualyy and see which one can generate the error.......
good luck my friend...
May 10, 2006 10:12:48 AM

I have tried to run memtest from the microsoft website but only for about 45mins.

I will leave it on for longer tonight. Probably all night in fact.

I tried to turn the ram cycle to 2T but couldn't see an option so turned all the memory timings to 2 as described above. Is this what you meant?

Yes its strange that it crashes when playing a dvd. Its not the software by the way and not the dvd drive as it crashes with more than one dvd software and when playing direct from the hard drive.

It also crashes when Im working through a very simple 3d scene that is not at all taxing.

I monitor the cpu, system and power supply temps but they all seem relatively low when it shuts down.

The funny thing is, once its gone off it wont turn back on again for a while - as if something needs to cool down before it will turn on properly.

If I try to switch it straight back on after a shutdown sometimes it turns off the instant I have depressed the power switch (as quick as that!)
May 10, 2006 10:18:25 AM

Fainis, thanks for your input I will try memtest again later. The only thing that bothers me is that I have tried the system with the old dvd test with each mem stick in by themselves and got the same crash with all of them individually - surley they cant all be faulty??
May 10, 2006 10:19:14 AM

nononononononono...my friend...you made a littel confusion......not the timings....but the command rate....

the timings are related to the memory module and they should stay on SPD......usually fo a ddr 2 module the timings are samething like 4 4 4 12...or a little lower...but that is a memory module problem....

the command rate ...is 1T or 2T ...and it`s related to the memory controller...1T..is fast and 2T is stable....

all these timings 4 4 4 12.....for example...are multiplied by 1 or by 2...this is the command rate....


the easy way it will be to reset the bios ..and run memtest....

good luck
May 10, 2006 10:23:28 AM

here is the little confusion you are making my friend.......

on intel boards....most of them.....command rate is 1T........
if you want advanced options .....do ctrl+f1 (it works on award bios)...but that`s not the case.......

be carefully my friend you`re confusind the command rate with the memory timings .....if you lower the timings ..that much.....2....your sistem will crush......and in same cases could damage the modules....

here it is another solution...

install cpu-z......in the mem section...you will have the supported timings for your corsair modules........make sure you don`t lower them too much
May 10, 2006 10:38:00 AM

Oh, I see!

Thanks for clearing that up for me.

I have downloaded cpuz and ran it on the my computer at work.

Under the memory tab>timings it gives reading like 4 clocks, 8 clocks and 16 clocks. I assume these all correspond to the timing options in the BIOS?

As you said I will reset optimised defaults and then rum mem test. On 1 gig at a time I have been told.

What if it the tests come back fine? The only thing I havent changed is the cpu maybe that could be causing the problem. Unless I was unlucky with the new mobo that was replaced back when I first built the PC?
May 10, 2006 10:44:02 AM

if the memory test turns all right.......the only problem i see (you said my friend ..you have a good psu)..is the mobo........perhaps ..a shortcut with the case...try to take it out of the case and take a look...........


all this time you check the stability of you sistem make sure you have all the settings default-ed......reset the bios.......in this way you will eliminate a wrong software setting

...dissable spread spectrum..for the hdd..try to run only one of it ..the one with the os........

good luck ...came back with more details...i wish you good luck my friend
May 10, 2006 11:00:18 AM

I have had a good root round and taken the mobo off the tray and put it back on again to eliminate a screw that might be causing a short but this didnt seem to help.

Not sure how to disable spread spectrum for the hdd. Could you explain how this works?

Thanks again for your help with this Fainis.

Spence
May 10, 2006 11:08:53 AM

well .. i understood you have wd drives........wait ...to take a look at one of it...i have a wd2500js ..in my store......


on my wd2500js...i see a table with jumpers.....on the label it saids...short the jumper 1 and 2 to enable ssc...spread spectrum clocking.......this means that the channel ..the sata channel..is tricked...it means that the channel reserves more for the hdd......the bandwith is enlarged..and the clock is more stable.......this is usefull for many situations...like overcloking ...

i`m sure this is not the case but ..take this into consideration.......you`ll never know


you could try only one hdd.....one memory module ....no expansion cards like modems ..or sound cards...just the video card...and see what happens...
good luck.....
May 10, 2006 11:32:52 AM

Right I get it. OK then Ill give it a go.

Thanks so much for your help!
May 10, 2006 11:40:42 AM

Sounds like a heat problem to me. The give-away is the fact that you can't restart immediately - sounds like something is overheating.

I'd make you main priority finding out the temps of your components and checking the fan RPMs.

Also, trying running your computer with the case open. This will help cool it, so it should run longer before shutting down. Also listen out for any weird noises that might indicate a faulty fan.
May 10, 2006 11:47:39 AM

When you say check the temps of the components do you mean other than the ones I have already checked?

Is there a way to check temps of other components on the mobo?
May 10, 2006 11:52:35 AM

Sorry, I missed that post where you listed your temps. They all look OK to me - GFX cards tend to run quite hot.

Did you check what the shutdown temp in your BIOS settings is?
May 10, 2006 12:02:48 PM

Yes the shut down temp was about 80 for the cpu.

By default auto-shutdown is disabled for power supply and system temperatures.
May 10, 2006 12:06:08 PM

Well I'm stumped then. If it was anything other than temperature, I'd expect the computer to be able to restart immediately.

Probably a fault with the motherboard. :( 
May 10, 2006 12:38:29 PM

hopefully Ill work it out, thanks
May 10, 2006 12:48:38 PM

check this out ...my friend i have another ideea........

you said your cpu is 80 degrees.....
check whether you have a shut down temperature set in bios.........maybe a default setting is arround 75 degrees......set it a little higher to isolate the problem...anyway 80 degrees for a cpu is quite high........


clean the dust of the heatsink or ...make sure is properly seatted
May 10, 2006 12:51:57 PM

No the temp is usually between 28-56 depending on how much processing is going on. The shutdown temp is set at 80 in the bios.
May 10, 2006 12:55:08 PM

ihh.........than that`s out of question.......ahhh..keep searching
May 10, 2006 2:17:01 PM

Had the same problem but without the apparent temperature sensitivity. System would inexplicably shut-down/re-boot while running DVD or even an MPEG, also when running programs like AutoCAd and MS Office.

Drove me crazy for almost two weeks before I slicked the disk and reinstalled everything one thing at a time and I found it was the Realtek HD Audio Driver that came during an update session.

Although my board is a Gigabyte GA-8I955X Royal vice an Abit, they both use Realtek's HD Audio chip, don't know if its the same but it may use the same bogus driver provided by Microsoft Update.

My original installation CD had a driver dated 5/4/2005, version 5.10.0.5125, that worked perfectly. Did the normal update thing with Microsoft which installed a "signed" driver version 5.10.0.5127 dated in SEP 2005 and that's when things went south. Uninstalled the updated driver and no reboot!! Reinstalling the original driver from the installatio CD fixed the "New Hardware Found", everything works and I'm happy!

Sent an email to Microsoft about a month ago - no response! Not surprising I guess.

Hope this helps. If it works, be sure to check the "Don't Tell Me Again" box for the update before it reinstalls itself again.

Good luck!
May 10, 2006 2:19:38 PM

ihhhmmm ... it worths trying this one....... :D 
May 11, 2006 9:13:35 AM

If the problem was caused by a driver wouldn't it only crash in windows though?

I used the mircosoft diagnostic tool which checks the RAM from DOS last night and this morning as we discussed; the pc still shutdown of its own accord after between 5mins and 30mins on 3 consecutive occassions. I left it off overnight so it had a good chance to cool down and so that any previous system temperatures could not attribute to another crash.

When I tested it again this morning it lasted about 10mins I think before it shutdown.

I had unplugged all the uneccessary components leaving only the graphics card, ram and cpu plugged into the mobo although it was still inside the case so I suppose there is an outside chance that something is shorting it out from within the case. It is worth mentioning though that I have stripped and rebuilt the whole pc only a couple of days ago to check for this possibility.

What I am wondering now is what these crashes during the memtest mean....does the memtest test rule out the cpu leaving only the ram and mobo in question (since I have a brand new graphics card in replacement of the old one, therefore ruling out it out)?

If so I will borrow some ram from my associate Joe, which is compatible with my mobo but a different spec (corsair value pc4200 dual channel 533Mhz) and if it still shutsdown then I know its not the ram.

Thanks again for any help with this.
May 11, 2006 10:33:23 AM

Quote:
The only thing that bothers me is the way it behaves once it has shut down and I attempt to turn it back on. It seems like something has to cool down before it will start up properly....could this tally with the auto-shutdown theory??

I just know that behavior with power fluctuations. Once the board gets too much juice it turns off and needs 5 - 15 minutes (depending on the board) to get it to work again. But with this error you would have to unplug the power cable as well to make it work again, without unplugging the power cable it never worked, so it might not fit to your problem, except they changed it so that you dont have unplug the power cable in case of overvoltage.
May 11, 2006 10:42:50 AM

Yeh I have also got a surgemaster power adapter which might minimse the chance of power fluctuations. Also when I bought my computer into work recently it suffered from the same problem so I doubt if its the power in house causing the problem.
May 11, 2006 10:59:59 AM

What I would really like to know is; does the microsoft diagnostic memory test that runs from dos use the cpu? If it doesnt then I know my problem is down to ram or mobo...
May 11, 2006 1:17:44 PM

It does use the CPU, although very lightly. It's not processor-intensive at all. Very memory intensive though. So that leaves you with RAM and motherboard being your main possibilities (although the CPU is still possible. Not likely, just possible).

-J
May 11, 2006 1:20:08 PM

Righto thanks bud
May 11, 2006 3:18:50 PM

I'm going to suggest the all time default of hardware trouble shooting. Take out everything (not cpu) and only plug in video card, HDD (the one with OS) and one memory module.

Then download memtest86 (it's free), can't stand microsoft diagnostic stuff, doesn't work well (just like their defrag). Run that for at least 4 hours but probably 8 and see what the results are.

Do you have the most up to date bios and drivers for everything?

I actually don't think the amount of power is the issue, even if that PSU sucks and is only 70% efficient it should still have enough juice for your system.

After you know it's not the memory it has to be either your MB, CPU or HDD, don't think your vid card should shut down your system that way. Not easy to test those except to try installing your OS on another HDD so you can rule it out. After that only the CPU and MB are left, unless you know someone with another exact CPU to test in your setup you may be stumped.
May 11, 2006 3:20:36 PM

you said your power supply reaches 70C... how do you measure that? does our psu has some shutdown temperature?
May 11, 2006 3:26:47 PM

Yes the psu does reach 70 sometimes and my uguru mobo utility measures that but it is rearly on long enough to get that hot it usually shuts down within the first 30mins if im using it.

If have swapped out the hdd already and the same problem occurs.

I have a mates ram to use tonight so I can rule that out hopefully.

Yes I will run memtest86 with the mobo outside the case with only the gfx card, ram and cpu in.

Im off to ABIT tomorrow and they said they will test/swap my mothrboard for if necessary.

I will post what happens tomorrow incase anyone is interested

Thanks

Spence
May 11, 2006 3:58:53 PM

I agree, with the lag in the startup it is heat related and to get a heat related shutdown that prevents immediate restart it is the mobo or cpu.

snape: you mentioned earlier that just touching the power switch (not acutally pushing it) causes freaky things too. That could be a short somewhere, and not just a ground on the mobo. (I saw you checked for that) It could be a faulty ground on the outlet you are plugged into for the system. That ground is for the psu, but also grounds out the case and all connected components. If you are shorting out on the case that could be a bad ground in the outlet or the psu. I realize the psu seems to be good, but it could be failing to ground. If the casing of the psu is not contacting the case right, or if the outlet is not properly wired that could cause the shorting out.

This could also be resulting in the mobo capacitors or other components overheating.

Check that the psu is seated well with bare metal on it and the case touching each other. (even just the screws are fine) also try plugging it into another outlet, preferably at another house or if you know for certain that a particular outlet in yours is good...

Finally, try re-seating your processor... I know you said it is cool, but just before the shutdown it could be spiking its temp. and a reseat w/ thermal grease could fix that...

Just thoughts man... good luck
May 11, 2006 4:30:39 PM

Right, well I guess that stripping down the pc and testing it out of the case with just the mobo, cpu, floppy, psu, and graphics card on a wooden table should eliminate any shorts due the case or nearby environment.

Also when I mentioned the power turning off when I touched the power button it is only after I have depressed the button, therefore turning on, then off - in an instant.
May 11, 2006 4:47:56 PM

Quote:
Also when I mentioned the power turning off when I touched the power button it is only after I have depressed the button, therefore turning on, then off - in an instant.


right, which could be from static build up in the case from improper grounding.

I would still try to plug in the system in a known-grounded outlet, even if you have it out of the case... but the reseating the processor is prob a more heat-related option to try first. It could be that the heatsink is just not fully in contact w/ the proc and after a short time the temp spikes...
!