Random Power Outtages

[dldeadbolt]

Hi Folks

Was told to post here by a friend that you've helped previously.

For the past month or so (since mid December), my computer has been randomly shutting down. No BSOD, Just shutting down as if someone had pulled the power cable out.

At first I thought that something was overheating, but after checking it a few times when it shuts down, there are often no parts that are warm, let alone hot.

Please can you help me sort this out... I'm just hoping its nothing too serious.

Regards,
DeadBolt

 

[ICIT2LOL]

Hello and welcome deadbolt mate run Speccy ands see what the CPU and motherboard temps are - you can't just feel the outside outside of the case for those.

You should find the CPU and mother board temps around 10C above room temp if that


http://www.piriform.com/speccy/download

Another possible cause could be the PSU is on the blink.

Some system specs mate would be very helpful

 

[dldeadbolt]

Hey Icit

Currently it is sitting at:

Operating System
Microsoft Windows 7 Home Premium 64-bit SP1
CPU
Intel Core i5 750 @ 2.67GHz 72 °C
Lynnfield 45nm Technology
RAM
8.00 GB Dual-Channel DDR3 @ 761MHz (9-9-9-24)
Motherboard
ASUSTeK Computer INC. P7P55D LE (LGA1156) 65 °C
Graphics
PLE2472HD (1920x1080@60Hz)
ATI Radeon HD 5800 Series (ATI)
Hard Drives
699GB Western Digital WDC WD7500AADS-00M2B0 ATA Device (SATA) 30 °C
1863GB Western Digital WDC WD2002FAEX-007BA0 ATA Device (SATA) 33 °C
1863GB Western Digital WDC WD2002FAEX-007BA0 ATA Device (SATA) 33 °C

Been chatting to a couple guys over Mumble, and they are pretty much going with if its not a power surge (which the pc is protected against) then its more than likely my PSU (Corsair TX850W, 3yrs old)

 

[ICIT2LOL]

Hiyya amte the CPU and mobo temps are touch on the warm side if that is at idle.

Having said that as I mentioned too the PSU is a prime suspect although it is a Corsair (I prefer that brand myself HX's) but that doesn't mean it isn't going to bed.


If you want you can check the volts very easily but please be careful these devices pack a lethal dose a bit like a hospital defibrillator if you mess around inside them. We have just dome a fair few debates and tutorials on these and one that you can try is this one and you do not need to take them apart to test the volts anyway
http://www.sevenforums.com/tutorials/86648-psu-jump-start.html

but do be careful to use just the first part of the tutorial to start it and then test the leads for volts which will be
YELLOW = 12v RED = 5v and ORANGE = 3v +/- 0.5v on all typically you will find 12.4v 5.3v and 3.3v any less than that will indicate some very amiss with the PSU.

Having said that just because there are volts doesn't mean there is sufficient current!

AGAIN BE CAREFUL!!

 

[dldeadbolt]

Hey.

Those temps are from right after running BF3 on high/ultra. (though even now after web browsing and a couple power outs the temp seems to stay aroung 70/60* respectively)

i did a RAM test earlier and those all came back clean.

I have just tried to run a Prime95 torture test a couple times (In-Place Large FFTs - Max Heat, power Consumption, some RAM tested), but each time right after you click OK to start, the PC shuts down... Its really looking to be the PSU, but theres a slight chance it could be the mobo

 

[Britton30]

Those temps should be in the 30C-40C range at idles and max out at maybe 75C under stress. Try clearing out dust from your fans and CPU cooler fins. Check the Corsair site to see if that PSU is still under warranty and describe the issues.

 

[ICIT2LOL]

Those temps should be in the 30C-40C range at idles and max out at maybe 75C under stress. Try clearing out dust from your fans and CPU cooler fins. Check the Corsair site to see if that PSU is still under warranty and describe the issues.

Absolutely mate all my Corsairs came with a five year warranty so at three years old you should get a replacement . Plus you should just check those volts too like I said any volts below 12 5 and 3 are not good.

 

[Layback Bear]

Just a note; your power surge protection (???) could also be at fault.

 

[dldeadbolt]

Heya

Those temps should be in the 30C-40C range at idles and max out at maybe 75C under stress. Try clearing out dust from your fans and CPU cooler fins. Check the Corsair site to see if that PSU is still under warranty and describe the issues.

The comp had a full cleanout about a ~10 days after this first started happening.

Absolutely mate all my Corsairs came with a five year warranty so at three years old you should get a replacement. Plus you should just check those volts too like I said any volts below 12 5 and 3 are not good.

Yeah it is still under the 5year warranty.
I would check them but I don't have any devices to do so (Unless it shows in BIOS?)


Sorry, not really tech savvy. I can do general fixes, but when something like this happens I have no clue what I'm doing.

I left Speccy running for a while whilst idle and the temp didn't change.. CPU stayed around 70-73, MB 60-62.

I've spoken to the tech guy at work, and he has said thatit might be an issue with the heat control chip?

 

[ICIT2LOL]

Still looks pretty hot to me DL I would be not happy if it were my temps my friend I personalty would be redoing the thermal compounds. Doing that is relatively simple and the readings for the volts you only need a cheap digital multimeter to check those - probably $10 -15 in a hardware store.

That fellow could be right too and it might be giving you false readings so do that temps correspond to the BIOS readings? Having said that I think I am right in saying the temp control is within the CPU but I stand to be corrected. So the best reading you can get will be in the BIOS because the CPU reports back to the BIOS what temp it is at.

Therefore if the temp sensor is wrong then it follows the CPU may be out of whack.

Hum just remembered this CPUID - System & hardware benchmark, monitoring, reporting and there is a pic of mine
John

 

[dldeadbolt]

Hey Icit

Just got in from work. I've had the PC manually shut down throughout the day.
*Room temp is only around 20-25*C
I checked the temps in BIOS at startup, and Mobo temp was at 20*, whilst CPU was showing at 70*C (158F)!

By the time it loaded into windows the Mobo was up to +60* again.


Ran that CPUID programme, here are the results:

 

[ICIT2LOL]

Hum DL Looks like the CPU is definitely out of whack the temp is still too high (for my liking) the power usage (watts that is close to max) and amps high too 60+ amps a bit like a runaway train.

Now for starters the thermal compound needs replacing because if the heat is dissipated properly then obviously the temp will come down. Why it is drawing that much current is odd to me and if that doesn't work then maybe the chip itself is going home or there is something going on with it and the motherboard to make it pass that amount of current.

All of the other volts and stuff look ok my friend.

 

[Britton30]

Have you tried anything yet, cleaning dust, renewing the thermal compound, checked PSU volts? I suspect either the CPU or the board is defective. No good way to test other than replacing parts unless someone knows a way.

 

[ICIT2LOL]

Have you tried anything yet, cleaning dust, renewing the thermal compound, checked PSU volts? I suspect either the CPU or the board is defective. No good way to test other than replacing parts unless someone knows a way.

Yes mate there is no way cos as I think I have said many times before (not here) you might find volts but if there is no current it ain't goin to work.

An almost flat car battery is a good analogy - volts may be ok but try cranking the engine and it just is not going to come up with the goods.

 

[Layback Bear]

Follow this site and it will show you how to install thermo compound,(paste)

Arctic Silver, Inc. - Intel® Application Methods

Make sure your bios are at default. Your vcore volts are to high. The max volts should be in your manual.

Please post your specs. This will help.

http://www.sevenforums.com/tutorials/180324-system-info-see-your-system-specs.html

You shut down is probably caused by temps and not a power outage.

 

[dldeadbolt]

Have you tried anything yet, cleaning dust, renewing the thermal compound, checked PSU volts? I suspect either the CPU or the board is defective. No good way to test other than replacing parts unless someone knows a way.

Follow this site and it will show you how to install thermo compound,(paste)

Arctic Silver, Inc. - Intel® Application Methods

Make sure your bios are at default. Your vcore volts are to high. The max volts should be in your manual.

Please post your specs. This will help.

http://www.sevenforums.com/tutorials/180324-system-info-see-your-system-specs.html

You shut down is probably caused by temps and not a power outage.

The comp got a full cleanout just after Xmas, along with new tharmal paste.

I've updated my specs page.

Just had the pc on, checked temps in BIOS (CPU 70+, mobo 20), logged in windows, ran Speccy (CPU 75+, Mobo ~70). Let it sit running for an hour, restarted and checked BIOS temps again... CPU at 70, mobo at 20.

SOmething is definitely wrong there.

BIOS to my knowledge has been set at default since i got the PC. My mindset previously with it has been "I don't know what I'm doing in there, so I don't touch it or i will break it"

I recall RMAing the PC a while back, and having checked my emails it was in mid Nov 2010 for a faulty CPU. The original purchase order also had a Coolermaster 212+ heatsink, but the current is a Thermaltake (not sure what model) that was changed then. Don't recall what the full outcome of the RMA was.

 

[Dwarf]

It sounds as though you could have used too much thermal paste. You will need to clean it off and reapply fresh paste, but don't use too much. A blob the size of a pea spread evenly and thinly over the top of the processor is plenty.

 

[Britton30]

In addition be certain the CPU cooler's screws are tightened in a cross sequence,that is top left-lower right, etc. until all are secure and don't over tighten them.

 

[dldeadbolt]

The paste was put in after the crashing/temp problems started.

I'll redo it and give it a go, otherwise ill replace test the parts

 

[Britton30]

While you're :under the hood: look for any indications of burn marks or swollen capacitors on the board.