PC powers off at random times

Question

Short Version

After experiencing some problems with Mobo batteries my PC started to power off at random times, the power off is instant and sudden and does not restart afterwards, need help figuring out the cause.

Facts:

• Powers off when PC is playing games
• Powers off when PC is idle
• Powers off when PC is in safe mode
• Powers off when PC is in BIOS
• Powers off when PC is booted through a Windows installation USB
• Replaced the motherboard battery several times
• Replaced the 650W PSU with a 750W PSU
• Replaced the RAM
• Swapped the RAM between slots
• Re-applied thermal paste to the CPU
• Checked if the motherboard touches the case
• Nothing is overclocked

PC Specs

PC specs:

OS: Windows 7 Ultimate SP1
RAM: klingston 1333MHz 4GB stick
CPU: AMD Phenom II x4 955
Mobo: Gigabyte 88GMA-UD2H rev 2.2
Motherboard battery: CR2032 3v
HDD: 500GB Seagate ST3500418AS ATA Device
Graphics: ATI/AMD Radeon HD 6870

Very Long version

Around 10 months ago I built a brand new gaming PC.

Around 6 months ago it's time setting in windows started resetting to the year 2010. I swapped the Motherboard battery for a new one of the exact same size and shape and voltage, and the problems disappeared...for around 2 weeks. Then the same problem happened again, time gets reset, I swapped the battery again, and the problem was gone for good and everything was great for about 3 months.. then another problem started happening, the PC started to power off suddenly and without warning at completely random times, sometimes the PC works for and hour, sometimes 5 minutes. So I read on the forums that it might be either the PSU or the motherboard Battery or RAM or HDD or the Graphics card or the CPU or the motherboard or the drivers or a Virus or Grounding issues, or something short circuiting, basically it can be anything...

I spent some days researching, and decided to remove the possibility of a virus. I reset the CMOS, cleared all BIOS settings and reinstalled windows 7 after a full format of the HDD, but the random power off kept happening.

I then disabled the restart on error option in windows and looked at the event log for error events, but they did not help me figure out the problem.

Network list service depends on network location awareness the dependency group failed to start
Source Kernel Power Event 41 Task Category 63
Source Disk Event ID 11 Task Category None The driver detected a controller error on device disk

I took apart the PC, every little piece, re-applied some expensive thermal paste to the CPU, and double checked that none of the pieces are touching the PC case.

The problem was gone, the PC no longer powered off randomly I re-attached the graphics card and all was good for 4 months... then the power off problem appeared again, but was happening at high intervals, the PC would shutdown once in 2 days on average, at random points in time, sometimes when it's idle all day long, sometimes when it's running CRYSIS 2.

I checked the CPU temperature, because I know that AMD CPU's have a built in protection mechanism that switches off the PC if the CPU gets too hot, and the Temp was 50C system temp, and 45C CPU after running the PC all day long (I did not do tests to see if there are any temperature spikes, don't know how to do them)

Originally the PSU that powered the PC was 650Watts and had one 4 pin cable to power the CPU, I replaced it with a new 750Watts PSU which has two 4 pin cables for the CPU, but the problem remained.

I removed the graphics card and let the motherboard use the built in one, but the PC kept suddenly powering off at random times.

I took apart the PC completely again, and re-applied thermal paste to the CPU, added lots of insulation, and checked for any type of short-circuit possibility again and again, but the problem remained.

The problem was like that for some months. I replaced the Battery a couple of times over the time, changed lots of options in windows, and tried everything I could, but it kept powering off, so I stopped using the PC as much as I used to, just living with the random power offs from time to time, until a couple of days ago, when the power off happens almost immediately after powering on the PC.

I replaced the RAM with a brand new one, but that did not help.

Took apart the PC again, checked for anything anywhere that might cause it, found some small scratches on the very edge of the motherboard to the left of the PCI express x16 slot. This might cause the problem, I thought, but the scratch looks very superficial, not deep at all, and if the scratch did harm the motherboard, wouldn't it cause it to not start at all? And why did it start to power off a while ago, and then suddenly stop powering off? The scratches could not have vanished???

did chkdsk \d but it powered off when it was at 75%

I removed the hard disks, the graphics card, while I fiddled with the BIOS settings, and suddenly the PC shut down while I was looking at the BIOS version.

This makes me realize, it is not caused by: HDD, Windows, Drivers or the Graphics card

I cleared the CMOS again, updated the BIOS from F5 to F6f beta, but that did not help, it might even seem that the PC powers off even sooner.

The shutdown even happened to me while I booted through a windows 7 installation USB and was in the repair console.

I removed one of the cables powering the CPU, now only one 4pin cable powers it, and it worked for 30mins after doing that, which makes me think that it's the CPU overheating, and because it gets less power, it overheats slower?

The things that I am still considering:

1. CPU overheating (does not seem to overheat, maybe false readings?)
2. Motherboard short circuiting (faulty motherboard?)

I desperately need some advice in what is faulty, is it a faulty Motherboard or an overheating CPU? or maybe something else?

I have been breaking my head over this problem over a span of 6 months. I'm not sure if this is a good place to ask this question, if it is not, then tell me where I can get some experienced help.

More info

• Looks like each time that it powers off the datetime gets reset
• I have also discovered a mysterious piece that seems to have fallenout of the motherboard What is it? (1 year later, it seems to be a tuning fork oscillator that is responsible for counting seconds, which explains the clock getting reset and may have been responsible for all of this.)
• I also found another forum post tomshardware.co.uk/forum/… except I don't have Integrated Peripherals>USB Keyboard Function option in BIOS :S
• http://answers.microsoft.com/en-us/windows/forum/windows_7-performance/no-blue-screen-kernel-power-event-id-41-task/bca045cb-db3d-40b4-abf9-3a6ba3772ab5 - criziot
• How to Diagnose a Pre-Operating System Load or Hardware Issue

Comments summary (asked by Random moderator)

Q. tell me, if the computer restarts, is it immediately? Does it take a second and then restarts? Do you see (BSOD) or hear (PSU, short circuit) any suspicious when it happens? After reading trough it, it remains the mainboard that is faulty. – JohannesM

A. Immediate power off, all the fans stop instantly, all the light turn off instantly, no sound or anything, and it remains off until I turn it back on. Thanks for the feedback, faulty motherboard is what I fear.

---

Q. Try stress-testing the system with Prime95 and see if errors or shutdowns occur when the CPU is under full load. – speakr

A. Prime95 heat stress test peaked CPU heat at 60C after 5mins, it powered off after 30mins of testing in the middle of the test with no errors, Prime95 Heat test or the stress-testing with low RAM usage (small or in-place FFTs) do not report errors while testing for 10-60 mins. The power off does not seem like it is affected by Prime95 at all Makes me wonder if it's a CPU or Motherboard issue at all.

---

Q. I had similar random/intermittent problems with my old board. It gave one of a few different symptoms: keyboard and/or mouse would die and/or the RAM wouldn't work and/or it would shut down. It was in bad shape. One problems was that my old PSU had literally burned the connector on it (browned around the pins), another was that a broken lead inside the layers of the PCB would work sometimes if it happened to be hot or if I bent the board—by jamming a hunk of wood behind it. I managed to keep the board alive for several years, but eventually nothing I did would make it work correctly anymore. – Synetech

A. I will try that as the last resort, ok? ;)

---

Q. Have you tried a different power cord, surge protector, outlet (on a different circuit). It's worth a shot just to ensure it's not subpar wiring or a week circuit (dips in power may cause shutdown if the PSU can't pull enough juice from the wall). – Kyle

A. yes, I attached the PC to an entirely different outlet on a different circuit and the problem persists. After connecting it to a different outlet after starting the PC it gave me 3 long beeps and 1 short one, then the PC immediately proceeded to boot up normally.

---

Q. Re-check your mainboard manual and all PSU connections to your mainboard to be sure that nothing is missing (e.g. 12V ATX 4-pin/6-pin connector). If you can provoke shutdowns with Prime95, then consider buying new hardware -- a stable system should run Prime95 for 24h without any errors. Prime95 mentions errors in the log when they occur and gives a summary after the stress test was stopped manually (e.g. "0 errors, 0 warnings", if all is fine) – speakr

A. Re-checked, there are no more PSU connectors that I can physically connect, except the one ATX 4-pin (there are 2 that power the CPU) that I disconnected on purpose, I have reconnected it but the problem persists.

---

Q. With one PC I had a short curcuit. The power button on the front plate had its cables soldered, but not isolated, and the contacts were very close to the metal case. A heavier touch was enough to cause a shutdown. The PC's vibration could be enough – ott--

A. yes, it seems to switch off with even the lightest touch, I switched on the PC, then pulled out the front panel power cable that connects to the motherboard so the power button does not work anymore, after 5 mins of working like that, with the power button completely disconnected, just sitting idle, the PC powered off again, I don't think it's the power button.

---

Q. I wonder if you dare to operate components without the case, that is remove motherboard, power, disk ( just put the motherboard on a wooden desk). Don't bend the adapters when running like that. – ott--

A. yes, I do dare to do that, but only tomorrow, too tired/late right now.

Conclusion

Replaced the motherboard and all is fine again. Thank you all so much for your help.

1 year later Update

The strange piece that fell out seems to be a tuning fork oscillator/resonator that is responsible for counting seconds, which explains the clock getting reset and may have been responsible for all of this.

Solution 1

I had a motherboard that did that. The cause was I accidentally installed a 1394 cable into a USB spot and it permanently damaged the motherboard by putting voltage where voltage doesn't belong. It exhibited the exact same symptoms. I lucked out because it reset seldom enough that my computer was still mostly usable. It's a common error, and if your computer has been acting up since you built it, this is a likely cause.

Unfortunately, no amount of tinkering will be able to fix it, you might be able to permanently workaround the issue by completely disabling USB/1394 in BIOS though.

No way to know if this is what's causing your issue though, unless you can recall accidentally doing the same thing during build.

Solution 2

I believe that I've found the answer. Notsosure's link to the Toshiba thread really helped in figuring out the logic behind this mess.

What's happening is that certain capacitors on the motherboard that regulate the current of the CPU have started to fail. So whenever the CPU changes its clock speed or multiplier, the capacitors do not deliver the required voltage to it and the CPU malfunctions. This causes the system to reboot. Some users have fixed the issue by replacing the faulty capacitors (they are located directly behind the CPU socket on the motherboard).

Intel refers to this throttling technology as Speedstep. Windows Safe Mode seems to not use this CPU function which is why the system doesn't reboot in that mode. Disabling all the processors in Device Manager prevents normal Windows mode from using it as well. As long as the CPU doesn't use Speedstep, no spontaneous reboots will occur.

Comments

[Phillip R.]

I had a system that did the same. Turned out to be the motherboard. I first started noticing it when one of the nic ports stopped working. I tore it apart and had it running without a case and noticed that if I plugged a wired cable into the working nic it would die, but if I used the onboard wireless it would work fine. I replaced the board and it has worked fine since.

[Timo Huovinen]

Don't remember having anything to do with 1394 ever

[OmnipotentEntity]

This will also happen if you accidentally install a USB cable into the 1394 slot. FYI. So even if you don't have firewire you can still accidentally screw up your mobo.

[Timo Huovinen]

I looked at the firewire port, and I am very sure that I never connected any USB's there.

[Timo Huovinen]

I asked everyone in the house if they did that. Yes someone did connect the USB to the firewire port once -_-, will disable all USB/1394 in the BIOS and see what happens.

[Timo Huovinen]

I completely disabled USB and 1394 in BIOS, the pc proceeded to boot into windows, and after a while it shutdown... my attempt at the workaround failed

[Timo Huovinen]

@OmnipotentEntity you never said port, but I see what you mean, the internal 1394 and USB ports look very similar, might be the cause for the problems. Is there really no way to make sure that this is the reason?

[OmnipotentEntity]

Generally but not always, if you plug them in backwards and turn it on you'll smell some magic smoke, it's been a while, but I'll wager if you stick your nose down near the USB controller or firewire controller you might smell some.

[Timo Huovinen]

Thank you, it was the motherboard, probably because of the reason you stated.

[Timo Huovinen]

Maybe every time the CPU changed clock speed it launched the tuning fork and crashed because it was not there?