View Full Version : PC turns off then on then off then on

18-02-2013, 01:15 PM
This started yesterday, my daughter was on Lego Batman 2 and it froze, I couldn't alt-tab or get task manager or anything, so turned the pc off. Then when I turned it back on all the fans started up and just when it normally beeps it turned off for about 2 secs then came back on and the same thing happens again and again!

I thought it might be the CD drive as kids are always opening and shutting it. So unplugged that and tried pc again and this time it started up, but message came up saying last overclock settings had failed and to reset to default. But this cpu has never been overclocked! After doing that pc started fine and was running fine for 2 hours. Then I put a video on and about 15-20mins in the whole computer froze again! Same thing as well with startup off on off on.

After all that I went looking on google but it seems anything can trigger it, with the main things seeming to be cpu, psu and motherboard. It was getting late by this point and pc still wasn't starting 3 hours later so called it a night.

Roll on this morning and turned pc on and it started fine but got last overclock settings had failed and to reset to default again. And it's now been running fine for 4 hours, through I have just been surfing net and not played any videos or games. I even plugged the cd drive back in and its still working.

Does anyone have any idea what going on? I thought maybe it overheated but as I write this the cpu is at 30c give or take. And my Radeon HD 5670 is 26c. That seems fine to me but not sure what it goes up to when gaming

Intel Core i5 2500K
4.00 GB Dual-Channel DDR3
Radeon HD 5670
2x1TB SAMSUNG Hard drives

18-02-2013, 01:36 PM
First guess - the BIOS battery is dead or loose, so it's resetting the BIOS settings which is why you're getting the warnings about Overclock settings on a restart!?

It's also possible that in resetting, incorrect settings are being defaulted which don't match your current hardware and that could result in crashes.

It's also possible the motherboard is failing of course - that would have the same symptoms.

I'd say you're also describing overheating tho. The temps you're reporting are actually VERY low, 30c for a CPU is stone-cold - 26c for a GPU similarly - if you're seeing those temps with a PC 'in use' (e.g. running a game) I'd say they're wrong.

If the thermal solution (or fan) on one of those has failed, they can overheat quite quickly when stressed but MIGHT work OK otherwise. Heat doesn't always creep-up, a machine can actually run at a stable temp but when you make it work harder it can them climb quickly until it hits it's shutdown threshold (or the PC just crashes)

To check that, just download a benchmarking solution like Kombustor or FurMark - that will report temps as well as creating them!

18-02-2013, 02:11 PM
Although RAM errors seem to be accompanied by BSOD's, I would check the RAM. It's very easy to do with memtest.

18-02-2013, 06:57 PM
My bios has a memory check which I ran and it came back fine, I'll try memtest later just to be on safe side.

Im not sure if it is down to over heating as it's been on all day now and I've played a couple of games in that time, I've just sent 2 odd hours on crusader kings 2 and it's been fine, no lagging or anything unusual.

I'm thinking its maybe the motherboard... guess I'll see how it goes...

Oh and the bios warning dont come up on every restart just when the pc crashes and it finally turns on

18-02-2013, 07:41 PM
Here's my piece of stock advice, probably won't achieve anything, but on rare occasions reseating the memory and graphics card can cure a problem caused by them working loose. Why might this only go off after use? Some strange artifact of heat expansion. It only costs you five minutes (if you are careful), so its not as if you lose anything. While you are at it, check cables for a loose connection (apologies if this was the first thing you'd tried).

19-02-2013, 01:57 AM
Another thing on overheating - it's entirely possible that you've just got a build-up of 'fur' (dust and crud) on the heatsinks of your GPU/CPU.

I might have - erm - once, perhaps, when playing the Space Marine demo - erm - forced my machine into a hard reset thanks to not having properly cleaned those for about 3 months.


Can be tricky to get it out from behind fans/around heatsinks sometimes - but it will, eventually, coat them enough to make a gradual heat increase into a slightly greater one and the difference between a GPU which rises into the 50s/60s into one which rises into the 70s/80s/OFF

19-02-2013, 10:01 AM
I just remembered a similar problem on an older PC (AMD XP2800+ old). I think one of the electrolytic capacitors on the mainboard was leaking. But those should be solid on recent mainboards...
If there happens something funny with the graphics card, you could just remove it and use Intel's integrated one instead.

19-02-2013, 08:45 PM
First guess - the BIOS battery is dead or loose, so it's resetting the BIOS settings which is why you're getting the warnings about Overclock settings on a restart!?

That would have been my first guess too. Reseating RAM and GPU can't hurt either.