Computer randomly shuts itself down.. help? :)

Hiya,

I've just started having a problem with my computer just deciding to randomly shutdown after 1-2-3-? hours time. I'll be up and doing stuff as normal, and for no reason I can see the computer just powers down.. I can push the power button straight away and the computer will turn on and stay of for another random 1-2-3-? hours b4 shutting down again.

Not really sure of the reason - this just started happening Yesterday evening (great timing.. long weekend! booo!).

I've tried using 2 different power boards (just in case the 1 was faulty) as well as 2 different power sockets. Still drops power - and only the comp does.. nothing else.. which made me think from the start the power board was not the issue (but had to test, just to make sure).

My PSU has a lil light that shines in it's on/off button - this doesn't go out.. so there is still power in the PSU (at least for that light.. LOL) when the power drops.. but everything else in the case acts if I'd hit the power button - ie case fans stop, cpu fan stops, etc.

Here's some specs, in case it's of any help:

AMD 64 3500+ (single core, not X2 or anything fancy :P)
Gigabyte GA-K8n Pro-SLI MB
2 GB Corsair Ram
1 * 250 GB Western digital SATA2 HD
1 * 320 GB WD SATA2 HD
1 * Sony DVD Burner
1 * XFX nVidia 7950GT 512MB Video card

The PSU is a 650W Atrix (Never heard of the name to be honest... but I've had bad luck with some known brands.. so decided to give this a shot.. hehehe) It's specs are:

+3.3Vdc = 35A (35A) / 115W
+5Vdc = 42A (58A) / 210W
+12Vdc1 = 17A (18A) / 204W
+12Vdc2 = 19A (22A) / 228W
-12Vdc = 1A (1.2A) / 12W
-5Vdc = 0.8A (0.9A) / 4W
+5VSB = 2.7A (2.7A) / 12.5W

Maybe that will mean something to someone .. (and maybe have some relevance to my problem?)

On the 1 plug rail I have: 1 Case fan & the video card
On the other rail I Have : both HD's & the DVD burner.

MB Power is using the full 24-pin plug. It can be done as a 20-pin connection only if I want/need to use that (both MB & PSU support that)

This PSU does not come with a PCIE power plug, so am using an adapter that came with the video card.

I've also had the PSU for 3 weeks now and not had any problems with it (b4 now, if it is the PSU causing this issue...)

I'm not too sure if the problem is with the PSU going on the blink.. or maybe an old MB going on the blitz, or could random power outages be caused by bad memory? CPU going bad?

BTW, don't anyone dare suggest I buy a Antec PSU.. I've had 3 Antec's.. and all 3 went *BOOM* on me.. hehee.. I will never, ever buy an Antec PSU again.

As I said, the power dropping is random.. I never know when *poof* I'll drop. I've had it drop after ~55 min, and as long as 4 hours on with no issues b4 *poof* it went. there is no indication it's going to happen that I can see or hear.. no flickering, power fluctuation's, no noise, etc.. one minute it's on and the next it's powered down.

If anyone has any idea's, or suggestion's of what I can do to try and isolate the problem.. (apart from trying a different PSU or MB or such.. I don't have a spare PSU, nor 1 I could borrow... same goes for the CPU and MB.. I do have a backup video card though.. hehehe). so yeah, any thoughts on this?

Mine does the same thing, but I have no idea about my computer specs and stuff :P

I actually think it's the video card but I'm no computer whizz

Does the computer shut down only when you stress the system (play a game, for example)? Also, what are your temperatures?

Never heard of Atrix, so I don't know how reliable those numbers are. Assuming it can actually supply 17A and 19A at normal temperatures, then the PSU is fine. But then, you won't know unless you swap it out.

http://www.memtest86.com/ for testing memory

To answer the question's (which should have been stated in my OP.. oh wells):

1. No, it 1st randomly shutdown when the computer was sitting idling... I was connected to the net, but had nothing open and it was sitting on the desktop (I was having some dinner and watching tv at the time). sometimes I'll be playing and it'll shutdown, other times I'll be browsing forums and such..

2. Not 100% sure of the temp's, but I know I can still touch the video card and cpu heatsink's with my finger's and not get burnt :P I do plan on next time I get a shutdown to quickly bootup and check temp's... but since I can touch the heatsink's.. not sure that's an issue..

Unless of course the heatpaste on the cpu is fried and I'm cooking the CPU.. I do have some arctic silver paste around.. might be worth considering re-applying the paste.. *shrugs* not sure.. will see what the cpu temp is after my next shutdown..

I might just try and do a swap on the video card and see.. My other card is an olg Gigabyte 7600GT SilentPipe II.. decent card, but not as good as my current one IMO.. but still, for the games I play.. it does the job just fine

Thanx for the link.. I'll give that a whirl and see what results I get

Edit - *sighs* course.. it needs to be booted from a floppy. and guess what? I have no floppy drive :P

Anyway to test the memory that doesn't require formating and booting from a floppy?

Download the ISO file instead, and burn it to a bootable CD.

Sounds like you don't have something hooked up correctly, or something on the board isn't getting the required voltage it needs to stay functioning. I would also check the BIOS and make sure there isn't anything setup under power management to auto shut off. There are a variety of things that can cause a PSU to automatically shut down, do you have an alternate PSU to try on the board? Could be a short, or bad jumper setting, or simply not enough power supplied to the PSU, bad connection, no connection, overheating problems, etc.

I'm hoping that i have solved the issue.. will see if my comp shuts down randomly again...

I decided to strip the comp down, clean any dust out, check connection's, etc.. (and took my ~2-3 hours to do so.. hehehe). what I found.. wasn't good, and hopefully is resolved..

1. Dust.

The dust on certain parts of the MB was shocking. Especially around the round power converter thingies on the MB.. the built up dust was not good - in fact it wouldn't surprise me if it was shorting out now and again and causing the shutdowns. There was also built up dust around the memory ports..

I carefully cleaned the dust away and carefully made sure the heavy buildup of dust was gone from all area's. I can't see any more dust anymore.. although the parts around where the dust was built up is still slightly stained looking ;/ Hopefully no permanent damage was done.

2. Connector's.

One thing I found while carefully disconnecting the cables was the power on/off connector was not seated properly - in fact the black plastic was sitting correctly, but one of the wires was half out (and I have no idea how that happened ;/).

I'm wondering if that was causing short's or something which was causing the shutdowns.. I've removed that and am now using my reset switch cable instead (it had no problems).. so I now don't have a reset switch.. but I do have a power switch where the plug isn't "broken".

4. The case / dust

I took the time to clean out the case.. the dust buildup on the floor of the case, as well as other parts was chocking.. I'm glad I've cleaned it out..

Matsumi - unless something has changed in my BIOS the only setting that is set to auto shutdown is if the CPU overheats. Next time I startup the comp I'll check the bios again to make sure nothing has changed.

Unfortunately, the only other working PSU I have is an old 400W PSU which doesn't have the 4-pin power connector that plugs in near the CPU... so couldn't use it in this comp ;/

Edit - This may sound silly..

But how do you make a bootable CD and burn a ISO file to it?

I've never done this before.. all I've ever done is either drag files to the CD drive and used windows burn, or used Nero to made DVD's that can play video's.. So any help in getting this done would be appreciated

Edit #2 - Well, turns out that I hadn't solved the issue.. 1 hr 20 min after booting up... shutdown again for no reason.

I've now swapped my video card with my other one, which apparently needs less power to run.. will see if that resolves my problem or not. If it does.. all it says is either the video card is bust, or the psu wasn't able to supply the correct power to it and started dying because of it..

I have also managed to get my hands on a 400w PSU which has the correct plugs for my MB... so if I get another random shutdown.. I'll try that and see.

The Memtest ISO should be bootable. All you need to do is find something called "Burn Image" in Nero (I think that's what they call it... should be under one of the menus. If not, then use Nero Express). Select the ISO file as the image and let Nero do its work.

I had that problem on another computer once and it was the power supply. They replaced it and it worked fine after that.

Final update, problem resolved.

It was the motherboard (or maybe cpu or ram.. still not 100% sure..). I did have an old MB/CPU/Ram that was able to boot up.. and ran it on my current PSU for 20 hours.. so I'm fairly confident it was the MB/Ram/CPU which was causing the shutdowns.

I've replaced the MB, CPU & Ram (couldn't find a replacement MB.. so had to upgrade).. and had the comp on for ~6 hours today, so I'm fairly confident the problem has been resolved.

Thank you all for the help and advice

Short answer || In short, as you say, that device is toast.

Long answer ||
Before this is closed up, id like to state that there is a possibility that your system was producing too much heat.

This is especially an issue in acrylic cases, and "positive flow" air flow design, or could also be a problem if one of your fans has suddenly died, causing a hot air buildup in the system, which would also cause a failure.

In a liquid cooled system, there are any number of any other possible causes for cooling failuire, and I am sure you can seach the web for them. Two of the most common are :

Using tap water or filtered water, or any other liquid that causes build up in the pump or hoses.

Pump set too high, or too low, after a maintenance, causing not enough water to linger over the copper, and not allowing it to absorb heat properly.




The dust buildup also does not help much. You probabbly know this, but for others that may view this page, that dust build up more than likely added to additional heat problems, causing either a capacitor along that gold trace near the over heated device, or a transistor inbedded in the board, to fail.

If this is the case, the system board or that effected device in particular, is now broken and almost impossibly unrepairable.

I hope that helped to explain a bit of a possible cause!

Karatta - I don't use a liquid cooled system. It's pure air cooled, but thats definitely some interesting points about the common problems with liquid cooled system.

I'm currently using a Antec Black Sonata case - it's a metal (well, feels and sounds like metal) case with 2 120mm case fans and a flow-through vented PSU (has a fan sucking air into the psu and another fan blowing air out the back). I did check all 4 fans, and all 4 are working fine. The CPU fan was working fine as well.

i've currently had this comp running non-stop for 3? days or so now :P

BTW, I'll definitely advise people to try and keep the dust from building up on their MB's & other component's. I'm not sure which build up on my MB was the cause, but as I stated in an earlier post - there was a considerable amount of dust that had built up on 2 major portion's on my MB, 1 was near the cpu/power and the other was around the memory. the dust was hard caked to the mb and thick...

While the cpu & video card were not overheating.. I hadn't even considered the heat the dust could cause where it was built up. I considered the possibility of it causing shorts in the wiring (though wouldn't that fry the comp when it happened?).. hadn't thought of the heating issue. Since it was built up a lot around the base of some transistors/capacitors.. quite possibly it was causing a rapid heat build up and that is what was shutting down the system.

I've had that problem before, both in my own computer, and in ones I've fixed for other people. I never did find out what exactly caused the problem, but it appears to have been some sort of bad driver or corrupt system file that causes the computer to shut down under some conditions (which may appear random). (The system can, as you know, be shutdown through software - it doesn't have to be a hardware problem. But it can be - see below) What I did in all those cases was to re-install windows, but that may not be entirely necessary.
You could try booting the computer from the WinXP CD. If you follow the normal series of prompts, you will get to a point where the install says it has found an existing Windows installation on the hard drive. It should give you the option to "Repair the current installation". Try that.

About the power switch - obviously, the switch was working before you noticed one lead was partially out, so the switch itself is OK. It's not likely that the partially exposed wire was causing any "shorts", and you can probably plug the switch back in and physically push the expose pin back into the plastic part. You should try to check if the connector has been damaged first though - the connector is a 'female' - it's designed to slide over the 'male' pin on the motherboard. Make sure it is not collapsed so that it won't fit over the pin. (You can also reseat the connector in the plastic part before you try to reinstall it.)

Oh yeah, I also just remembered that this problem can be caused by bad capacitors on the motherboard. The caps look like little metal cans (with plastic around the outside) standing upright on the motherboard. If any of these appear to be swollen, or have brown stuff oozing out, well... I'm afraid it's terminal. (Unless you can replace the capacitors yourself - the cost of buying new capacitors and getting someone to install them can be more than the price of a new Mobo.)