pc keeps rebooting without warning

Various, home build / Corsair hx psu
January 8, 2009 at 04:02:32
Specs: Windows XP, intel E8600/3.5 GB ddr 2
my pc keeps rebooting randomly! it usually happens about 3 to 4 times in an hour then will run fine for a few weeks, giving me with a false sense of security, then it returns!
i was experiencing a particularly bad run of it last night and it shut down without rebooting at one point, i unplugged the multi adapter plugged it back in and heard the faithful humming of 'standby' then rebooted myself...

i've read up on similar problems, i've turned off auto reboot upon error, but get no error so think its definatley a power failure. i've been inside the chassis and checked every cable, power lead and firmly placed ram in the slots.

also note that a few weeks ago it failed to reboot totally, i swapped the power cable for another which worked fine.

i connected the pc straight to the wall socket and not the adapter, but the problem persists.

sometimes it reboots before windows has attempted log in, i entered bios and left it for a while, the pc rebooted even from there!

i have unplugged all peripherals, tv-out cables etc, to no avail.

also, would a surge protector help my situation or are they only useful against storms?

thanks in advance

windows xp pro SP3
nvidia 8800GTX
corsair HX psu
3.5 GB ram

See More: pc keeps rebooting without warning

Report •

January 8, 2009 at 05:57:28
in my experience it is often ram. have you more than one stick in there. can you try pc with just one and if still not good then try the other on its own, or get some ram from somewhere else to try. but recently i had a pc that kept rebooting for no apparent reason and after swapping parts around with 2 other PCS it was finally narrowed down to a faulty Pentium 4 2.6ghz HT Socket 478 CPU which is the last thing I thought it would be. Took a long time to establish that

Report •

January 8, 2009 at 05:59:00
Your Corsair PSU is a decent one but there's always the possibility that it's defective.

Have you checked the CPU & system temps?

Report •

January 8, 2009 at 06:09:08
Do you watch the screens when it is rebooting? Do you see a momentary blue screen filled with text, flash by?

Boot into the BIOS screens and go to PC Health. Check the voltages to verify they are close to what they are supposed to be.

A voltage drop in the AC current can cause the computer to reboot. Even a momentary one. That can be caused by a motor driven appliance in the house starting up. Can also be line voltage fluctuations coming from the power company.

Was all the RAM installed at the same time?

Do you have programs set to run at start up?

These are some of the hardest things to troubleshoot IMO.

Manage> System tools> Event viewer. Reviewing those logs may shed some light on what is happening. They are time stamped. Viewing right after a reboot may prove useful.

Report •

Related Solutions

January 8, 2009 at 09:30:25
You have reseated cables, but did you think about the motherboard power cable?

Report •

January 8, 2009 at 09:47:42
a surge protector will protect against sudden INCREASES in power ("surges"), but not against DROPS in power ("brown-outs" or "dirty power"). For this, you would need a battery backup and/or a "line conditioner."

test the ram like people suggested. I didn't see anyone suggest testing your RAM with memtest86, though, which would be more thorough than simply running the PC with each stick, one at a time.

and yeah, check voltages of your PSU to make sure they're not too far off. crummy power supplies will start to damage other components as well the longer you leave it in there.

Report •

January 8, 2009 at 14:39:27
ok, a pleasant plethora of replies indeed! thanks peeps

i'll be back home in the mornin so will check each thing individualy.

the CPU is the second i've had in the system and remember slipping around a bit when rushing it in, so that could be the problem.

i'm using 4 sticks of identical 1GB ddr 2 geil black dragon ram which were all installed together. seems like a good brand, but i guess using 4 sticks raises the chance of one malfunctioning.

i don't remember seeing a blue screen when it happened, but i doubt it's a common BSOD as it reboots even when i go straight into BIOS, sometimes before it even gets there!

i checked every cable, including the PSU power cable.

and about voltages... do you mean check what they are actually running at in bios/health against what they should be running at as setup in advanced BIOS setting?

thanks for help

Report •

January 8, 2009 at 15:20:44
The PSU puts out 3.3V, 5.0V, 12V. See how close the actual readings are to those numbers. Check in BIOS/ PC health.

Verify the voltage settings in Advanced BIOS or wherever the RAM voltage is set to verify it is set at the voltage required for your RAM. The Geil may be non standard (may require higher than JDEC).

Lastly, when changing the processor did you clean off all the old thermal compound and apply new? As was suggested watch the temps. I doubt that is the problem though. Temperature issues would be more consistent.

Report •

January 8, 2009 at 23:04:34
You to me have the exact symptoms of a failing power supply. As was already mentioned, power supplies from a decent brand can malfunction too. If you let it go too long without replacing it, it may damage the motherboard.

WinSimple Software

Report •

January 13, 2009 at 18:47:19
ok, soon after my original post, the reboots become more and more frequent the

longer i had the pc turned on, so naturally put it down to overheating.

i tried one ram stick at a time (from 4 1gb sticks) in different slots - no


all fans are running well.

i took the entire system apart, cleaned everything, re-applied arctic MX-2 to

the cpu and north bridge. south bridge has a thermal mat which seems fine.

after putting it back together again, the reboots happen less often, but are

stil happening.

i tried memtest86 with each stick in for about an hour while in different

slots and all reported fine without rebooting. when i used 3 or 4 sticks

together it started rebooted after a about 2 hours, but still reported no

errors to the ram, so again i think something is overheating. i had the HDs

and dvd drive unplugged during the mem tests.

my cpu temps are in the low 30s, the system temp is about 40-ish i think.

the voltages are reported as normal by the mobo.

the pc reboots more often when in windows due to more stress on system, but

i'm convinced its not software related due to reboots during startup and


pcmark ran once without problems, and again with a reboot.

3dmark caused many reboots often, which i guess would point toward the GFX,

which brings me to another thing.... i recenlty started using an s-video lead

to play old games through a crt tv. i noticed sparks when the adapter touched

the metal exposure of the GFX card at the back of the computer!
could that have damaged the card with electrostatic charges?
also, getting the lead plugged in required some wobbling and movement of the

gfx card while the system was running....

well so far i think its down to the GPU or PSU and i'm trying to get hold of

someone's spares to test them.

what are other people's thoughts on this?


Report •

January 13, 2009 at 18:51:56
oh, and i can't make much sense out of event viewer, but figure that would be more important for a software problem....?

Report •

January 13, 2009 at 21:49:41
"i recenlty started using an s-video lead to play old games through a crt tv. i noticed sparks when the adapter touched the metal exposure of the GFX card at the back of the computer! could that have damaged the card with electrostatic charges?"

Most definitely yes. Swap out your graphics card. Electronics are very sensitive to static electricity. You don't have to see or feel a static discharge to damage them. It takes atleast 400v to kill a piece of electronic equipment. If you feel it, it's atleast around 3,000v. If you can plainly see sparks, it's much higher.

WinSimple Software

Report •

January 14, 2009 at 03:20:23
thought so, but i had hoped the metal that's exposed at the rear would have some sort of protection against these things, plus i've read about others experiencing sparks, so guess there's a few peeps out there electrocuting their cards!

and i most definitely did feel the static charges!

i was hoping it would be something a cheaper to replace...

Report •

January 14, 2009 at 07:47:43
would a damaged pci slot have any adverse affects on a replacement card? i don't wanna borrow someone elses to test it and ruin theirs...

Report •

January 14, 2009 at 10:12:22
something i forgot to mention - when i was cleaning the old arctic MX-2 thermal grease from the cpu, i noticed i had used alot last time, and a little bit got between the heat spreader and the actual cpu, could this be a problem?

Report •

January 14, 2009 at 12:00:46
That shouldn't be a problem, but I don't know.

I don't think you will ruin a PCI card by putting it in a damaged slot, but I don't know that for sure either, lol.

WinSimple Software

Report •

January 14, 2009 at 16:12:51
If you know a slot is damaged why would you even attempt to use it?

Report •

January 14, 2009 at 16:18:00
ok, i tried a different video card and everything's fine. the ram passed memtest 7 times (took hours!)
so that's a geforce 8800 GTX down the drain.

thanks for all the help in pinning this problem down!

i might try a radeon this time, any suggestions for under £200 (around $300) or am i better off sticking with nvidia

Report •

January 15, 2009 at 09:09:12
i dont know if the slot is damaged

Report •

January 15, 2009 at 09:18:55
OK, I thought you had examined it after the arcing and found some burn marks or such.

Report •

January 15, 2009 at 11:11:38
i did see a little crumb of something, looked like it could have been a tiny bit of metal or copper. maybe off the contact of the video card....

anyway, i've ordered a geforce gtx 260 core 216, 55nm with a 650w corsair PSU. the card is a PCI-E 2.0. will that be ok in a x16 slot?

Report •

January 15, 2009 at 12:16:17
It should be fine.

WinSimple Software

Report •

Ask Question