Most likely point of failure is...

[Ruri]

A few weeks ago I picked up a 60 gig OCZ Vertex 3 SSD for about $50. Formatted it and installed Win7 64 bit, things were groovy for about two weeks, then I started experiencing random shutdowns. No BSOD, no lockup, just powering down like the plug had been pulled. Some days it would run fine, other days it would shit itself multiple times. I ran stress tests, unplugged and reseated every internal component, booted and ran it from my backup HDD, updated chipset drivers - the whole shebang. Even removing the SSD entirely didn't solve things, so that's been eliminated as the cause.

Problem is, I'm not entirely sure where to go from here. My power supply is one of those bedrock-solid 650w Corsairs, the kind of fucker that has 5,000 reviews on Newegg and a 98% approval rating. In my experience if a PSU goes, it basically blows up and never works again. So my question to the hivemind is, what's the most likely culprit for causing this sort of random power failure?

This has never happened to me before, it's kinda sad really

[Vortex]

Probably the BSOD bug. You could try to isolate this by switching out the drive for a week and seeing if the issues stop, but that is also a PITA. First run the obvious memtest to check for faulty ram (its probably not, but memtest is one of the few diagnostic tests available for hardware so just do it). If you don't have a PSU available to swap in then diagnosing PSU death is basically impossible until the system fucks itself. So really, start with checking for the BSOD bug.

[Ruri]

Probably the BSOD bug. You could try to isolate this by switching out the drive for a week and seeing if the issues stop, but that is also a PITA.

Even removing the SSD entirely didn't solve things, so that's been eliminated as the cause.

[noobcake]

I know you might not think its the power supply, but if you have the ability to swap out to another one for a week, give it a shot...especially since now you've concluded it isn't the hdd.

Also listing the rest of your systems configuration would be a good idea.

Sent from my HTC EVO using Tapatalk (pissing FHC off since 2011)

[Daneel Trevize]

Defo consider the PSU, it's most able to cause such a shutdown (think how many bought it based on those reviews, and then the realistic failure rate because they ain't perfect every time). You could try memtest86+ for the RAM, and/or just run with each invididual stick for a bit, but it'd have to a lucky flaw that kept twiddling the bits that trigger a shutdown rather than subtle corruption and varying errors.

It could be something stupid, like the case/power button is shorting, so maybe try a different/without the case, or look for exposed or damaged wires or any hint that the button or mobo's tripping the regular shutdown signal. Also I think you can set it to ignore the power, sleep & reset buttons, might suddenly stop crapping though you'd have to shut down from menu.

[Zekk Pacus]

A straight shutdown with no warning is almost always either PSU or motherboard. Swap the PSU for a spare - if that doesn't fix the problem then it's motherboard, either it's shorting against the case or one of the capacitors has gone.

[Hast]

the most likely point of failure is as always between the chair and the keyboard.

[Bombcrater]

A straight shutdown with no warning is almost always either PSU or motherboard.

[Ruri]

Groovy, that's what I was leaning towards but I wanted to get a feeling for other folk's experiences. Will try my roommate's PSU today and see if I can reproduce the problem. Though part of me secretly hopes it's the mobo so I have an excuse to finally upgrade to a core i5.

[Belid Hagen]

you know something is about to die in a ball of fire when it takes 20 minutes to install and start SiSoft Sandra.

current suspects prior to testing:
RAM
CPU
OS HDD

[Belid Hagen]

even better, 1 hour 15 minutes into the "burn in" test, and the computer locked up completely, all peripherals became completely unresponsive. the last progess status i saw (i had it running in the background) was 38% complete, and it was 3 minutes into the GPU bandwidth test - which signaled it's start by crashing the display drivers 3 times in 40 seconds \o/

I don't think i can wait until the end of the month for those bastards to get home that special case.

[Ruri]

Update: here's the system specs for kicks, the main components are about 2 years old and inexpensive.

GIGABYTE GA-EP43-UD3L LGA 775 Intel P43 ATX Intel Motherboard
Intel Pentium E5300 Wolfdale 2.6GHz LGA 775 65W Dual-Core Desktop Processor BX80571E5300
OCZ Flex XLC 2GB (2 x 1GB) 240-Pin DDR2 SDRAM DDR2 1200 (PC2 9600) Dual Channel Kit Desktop Memory Model OCZ2FX12002GK
PNY RVCGGTX470XXB GeForce GTX 470 (Fermi) 1280MB 320-bit GDDR5 PCI Express 2.0 x16 HDCP Ready SLI Support Video Card
CORSAIR Enthusiast Series CMPSU-650TX 650W ATX12V / EPS12V SLI Ready CrossFire Ready 80 PLUS Certified Active PFC Compatible with Core i7 Power Supply

So here's something cool. Running Prime95 to keep a load on the CPU for several hours has no effect. However, playing a game that requires any amount of 3d processing seems to reliably cause the power fault. Is it even possible for a video card to cause that kind of problem, or does that point to the PSU somehow failing to handle the load?

EDIT:

even better, 1 hour 15 minutes into the "burn in" test, and the computer locked up completely, all peripherals became completely unresponsive. the last progess status i saw (i had it running in the background) was 38% complete, and it was 3 minutes into the GPU bandwidth test - which signaled it's start by crashing the display drivers 3 times in 40 seconds \o/

I don't think i can wait until the end of the month for those bastards to get home that special case.

Get your own thread squatter.

[Zekk Pacus]

I would say the power supply's crapping out under load, fermis are notoriously power hungry. A good 650w shouldn't do so though, so it's definitely pointing towards that.

[Ryas Nia]

Try running Furmark it will put your GPU under serious load, but its sounding like bad gfx memory

[Verizana]

I know you did not list it in your setup but do you by any chance have a soundcard? If yes, try removing it, if no, its the PSU or memory. Try memtest and a friends psu.

[Ruri]

Try running Furmark it will put your GPU under serious load, but its sounding like bad gfx memory

I poached a different gfx card and not five minutes into my test - BAM shutdown. Funny story though, as I was swapping in a replacement PSU as my next step I must have shorted something on the board because now all it does is blink the HDD light and beep at me. FRIED LOL.

At this point I like Zekk's thesis, so in my list of new components will be a higher rated PSU. It's a shame, but at least this way I finally upgrade to Sandy Bridge.

[Ryas Nia]

I poached a different gfx card and not five minutes into my test - BAM shutdown. Funny story though, as I was swapping in a replacement PSU as my next step I must have shorted something on the board because now all it does is blink the HDD light and beep at me. FRIED LOL.

At this point I like Zekk's thesis, so in my list of new components will be a higher rated PSU. It's a shame, but at least this way I finally upgrade to Sandy Bridge.

Man that sucks

[Brullig]

Another thing it may have been is overheating.

[Navigator Six]

...or your replacement PSU isn't powerful enough (that's what the problem was last time I had those symptoms).

[TZeer]

Gratz on a new computer!!