Computer randomly hard freezing, shutting off.

[Greatbacon]

So for the last three weeks my computer has randomly hard freezing (image freezes, m+k become unresponsive, audio halts) while playing pretty much any video game and sometimes just when I have netflix running and am surfing the web.  The weird thing is though is that the amount of time it takes to freeze is utterly random.  Sometimes I can play for hours before it acts up, sometimes it happens within the first thirty seconds of the game.  Not only that, the case lights and fans remain on, and I can still hear some of the HDs spinning.

This started around the time I put in a few more sticks of RAM (same style as the pair that were in there before) and I've run memtests on the full suite with no errors and I've also been getting the hard freezing with my original, stable set of RAM in.

I also added a new HD in the process although I've been getting freezes even from games that wouldn't need to be accessing it.

I've been monitoring my CPU temps as well and they've mostly been hovering around 50 C so I don't think that's an issue.

Mostly I guess I've either got it down to issues with my PSU or my motherboard and I'd just like some external input on that.

[HeK]

System specs?

[Greatbacon]

MOBO: ASUS P8Z77-V LX LGA 1155 Intel Z77 HDMI SATA 6Gb/s USB 3.0 ATX Intel Motherboard
CPU:  Intel Core i5-2500K Sandy Bridge 3.3GHz LGA 1155 95W Quad-Core
GPU:  EVGA 01G-P3-1556-KR GeForce GTX 550 Ti (Fermi) FPB 1GB 192-bit GDDR5
RAM:  G.SKILL Ripjaws Series 4GB (2 x 2GB) 240-Pin DDR3 SDRAM DDR3 1600 x 4 (2 sticks new, 2 sticks old)
PSU:  Thermaltake TR2 W0379RU 500W ATX 12V v2.2 Power Supply

[Greatbacon]

Okay, so I ran memtest86+ last night in fail safe mode.  It made 8 passes with no errors over the course of 12 horus.
I did notice however that if I tried to run it in multicore mode around the time my cpu hit 50c the whole thing would stop and if I just ran it normally it seemed to be subject to the same random freezing and shutdowns as normal operation.

[HeK]

50C is within normal operating temperatures, but it wouldn't hurt cleaning and reapplying paste to the heat-sink, especially if you haven't done so since you initially built the system.

While we're here, clear the CMOS with the jumper on the mobo and verify that you are running the latest bios, version 2104.

Are you mixing different series of memory? Are the voltages (1.5v for red Ripjaws and 1.6v for blue) and latency matching?

To confirm, did you have freeze-up during the memtest, or has it only been within Windows?

[WaR1761]

Run HDD error checker to rule out and clean up any potential bad sectors on it. Can take a while depending on how big the drives are so maybe worth running overnight.

Also event viewer is worth skimming through to see if there is any info to be gained. I have used this in the past and found out I was getting blue screens due to a generic NIC driver from windows update.

Since you mentioned you got new ram, make sure its installed in the correct slots. Can check your motherboard manual to verify positions.

Side note if you have a 2nd monitor and windows 7, I would recommend getting some of these gadgets installed. They are very useful in identifying issues when things slow down as its real time and can see what is getting hammered.

[Greatbacon]

So to address questions in order:

I figured that was a reasonable operating temp for my CPU, it just seemed sort of consistent to the freezing.

I haven't cleared my CMOS and I am currently not running the latest BIOS (from what I've read it can be sketchy to do this without a UPS, especially if power fluctuations are actually the source of all this hullaballoo)

All my memory is of the same size, timing, and voltage (red 1.5v) but one pair of sticks is new and one I bought when I built my box in 2012

Yes, there appeared to be freeze-ups during the memtest when run outside of failsafe mode.

I ran HDD check on both my drives, came back fine.

I've combed over the event viewer and there are absolutely no events being logged when my computer freezes up.  The only way to tell it happened is that there's a ten minute gap and an event logged that says Windows didn't shut down properly.

Confirmed that ram is in the correct slots and paired accordingly by make and slot order.

Thanks for all your advice on this guys.

[WaR1761]

I doubt its due to cpu temp, most have to get close to 90c to start throttling back and shutdown all together they don't usually freeze. Temperature monitoring app I use.

It sounds like this usually happens under graphical load or am I misinterpreting that?

Also remove ram and just run one set at a time and see if you still keep having the issue. Was either of the sets on the QVL list of the motherboard or ram vendor? Gskills configuration tool to verify compatibility.

Only other thoughts are loose cables/parts from when you installed the ram. Verify everything is seated.

[HeK]

Yes, there appeared to be freeze-ups during the memtest when run outside of failsafe mode.

Right there you can rule out any software impact, or HDD issues. If you are freezing during memtest 86, you have a hardware problem.
Did these issues start shortly after your move?

Typically, I've seen this problem from bad video, bad ram, or a bad mobo. Do you have another videocard to test with? Doesn't have to be anything impressive.
Try running memtest with one stick at a time.

[Greatbacon]

So most crashes seem to have occured under graphical load, but it's also happened more or less at idle with Pandora playing and during memtest execution as well.

I don't have another videocard, but I think the mobo has some onboard graphics, would that help narrow it down?

I'll try the memtest per stick, how many passes are usually enough to know you are set for that stick?

[FlyingMongoose]

So most crashes seem to have occured under graphical load, but it's also happened more or less at idle with Pandora playing and during memtest execution as well.

I don't have another videocard, but I think the mobo has some onboard graphics, would that help narrow it down?

I'll try the memtest per stick, how many passes are usually enough to know you are set for that stick?

The onboard graphics CAN narrow it down, if that part of the motherboard isn't the problem.

[Greatbacon]

So some updates.

Ran memtest one stick at a time and while I was able to run one stick for about 7 passes without incident, the others experienced the same freeze and shut down.

This was also with the GPU out, so I think I can rule that out.

I also (re)discovered that the outlet my power strip is plugged into isn't grounded, and I already trust the wiring in this house about as much as I'd trust a pederast to babysit for me so my guess is at this point it is either some serious external electrical issue or my mobo has gone and got itself fucked in up in a particularly special way.

[Elder]

use a volt meter to check your psu.
That part is most likely to fail at some point in your comp.

[Greatbacon]

So I think I may have resolved the issue.  It seems to have been based around the lack of ground for my comp.

Anyway, I looked around the house and found a grounded outlet (for some reason my room is the only one without one in the entire place.) plugged my computer in and ran memtest86+.  It went for almost 24 hours with no errors and no freeze shutdown.  Right now I ran an extension cord from my powerstrip to a grounded outlet and am currently sitting at a 5 hour uptime with some intensive benchmarking.

Anyway, for those who are curious, my research into grounded outlets indicated that 1.) A non grounded outlet will mean that your computer will have a high tendency to build up static charge and 2.) A surge protector that is not grounded is more or less useless as a surge protector.  My guess is that between the de-humidifying effect of the cold and the construction around here everything sort of culminated in a perfect storm of frustration.

[FlyingMongoose]

So I think I may have resolved the issue.  It seems to have been based around the lack of ground for my comp.

Anyway, I looked around the house and found a grounded outlet (for some reason my room is the only one without one in the entire place.) plugged my computer in and ran memtest86+.  It went for almost 24 hours with no errors and no freeze shutdown.  Right now I ran an extension cord from my powerstrip to a grounded outlet and am currently sitting at a 5 hour uptime with some intensive benchmarking.

Anyway, for those who are curious, my research into grounded outlets indicated that 1.) A non grounded outlet will mean that your computer will have a high tendency to build up static charge and 2.) A surge protector that is not grounded is more or less useless as a surge protector.  My guess is that between the de-humidifying effect of the cold and the construction around here everything sort of culminated in a perfect storm of frustration.

Adding a ground to existing outlets is not hard, just remember to turn off power to that room before you do it.