Seemingly Random Reboots on New System

[rockaday]

Hello,

I'm getting seemingly random reboots on my new PC. I set it up and it worked fine for about two months, maybe a couple or few reboots in that time but no big deal. Then in the last month I've had reboots of increasing frequency, occasionally the comp reboots so frequently it becomes unusable.

They seem random to me because they have now happened from Windows, Linux, and even while running memtest86+ outside any OS. They have happened while under load or idle, after the computer has been running for hours or right when I boot it up in the morning after it has been cold for hours. One time I was staring at Real Temp when the reboot happened and saw all core temperatures were in the low 30s. Occasionally Windows has blue-screened, but sometimes the comp is suddenly rebooting as if reset was pressed, and last night the comp was shutting itself off and unable to turn back on for several minutes.

The only hardware change I made around the time the reboots increased in frequency was to install a second SSD (OCZ Agility3).

Here are steps I've taken to try to resolve the issue so far:
- Unplug reset cable from front panel to motherboard in case of random pressure from case door or other weirdness
- Reset Bios to defaults (and then turn on AHCI so I can boot to windows)
- Run at stock speeds instead of overclock
- Ran memtest86+ v4.20 for over nine hours, 8 completed passes, 0 errors
- Disconnected new SSD
- Ran prime95 for over eight hours, 0 warnings, 0 errors

In the last week, reboot frequency has been kind of all over the place. I went three days with no reboot, and other times it has rebooted every few minutes or every hour. Not being able to consistently reproduce the reboots has made it a more lengthly process to diagnose.

Here were the specs of my rig starting in November 2011:

Cooler Master Storm Enforcer Mid Tower SGC-1000-KWN1
Cooler Master 200mm Red LED Fan R4-LUS-07AR-GP
Intel Core i5-2500k Quad-core 4.2GHz @ 1.310v LGA1155 BOX80623I52500K
Noctua NH-D14 CPU Cooler
ASRock P67 Extreme6 (B3 Chipset) LGA 1155 ATX Motherboard Rev. G/A 1.04 P1.60-15B
Mushkin Enhanced Redline 2x4GB DDR3 2133 9-11-10-28@1.65V SDRAM 996996
Seasonic X-Series 80 Plus Gold 650W Modular PSU SS-650KM Active PFC F3 (X-650)
Kingston HyperX SandForce Driven 120GB SSD SH100S3B/120G
Sapphire Radeon HD 6950 2GB DDR5 Dirt3 Edition (unfortunately not unlockable) @ 880/1375 model 100312-3SR
Mionix Naos 3200DPI Wired optical mouse with custom LED

February 2012 Updates to that:
Now I am running at stock speeds / bios defaults, which affects both the ram and the CPU. So now the CPU is at 3.3GHz and the RAM is at 9-9-9-24. I have also updated the ASRock P67 Extreme6 BIOS to version 1.80 (latest). I have occasionally tried overclocking my graphics card for benchmarks / fun for just a few minutes or hours at a time, but none of my games have needed it so I stopped overclocking it. Temperatures of the GPU have always been below 80C when I've looked (don't remember exact numbers, but they always seemed relatively cool compared to stuff I read online), even under load, but I should check them again.

I've run Prime95 for 8 hours straight with RealTemp open the entire time - at no time did any of the four cores get above 63C in RealTemp.

All the voltages looked in range when I was looking in my bios the other day, but I think for my next step I am going to install CPUID hardware monitor and monitor the voltages more. I also might try running with the case open for a couple days.

Do you have any other suggestions or thoughts as to the most likely cause?

Thanks for your help,

Rockaday

 

[geekapproved]

Well first of all your not supposed to run 1.65v memory with Sandy Bridge processors or you can damage the cpu.

Have you tried running your ram at say 1600mhz/9-9-9-24/1.5v?

Re-booting is usually faulty ram or cpu overheating, possibly power supply.

You may try running memtest on 1 dimm at a time.

 

[rockaday]

Well first of all your not supposed to run 1.65v memory with Sandy Bridge processors or you can damage the cpu.

Have you tried running your ram at say 1600mhz/9-9-9-24/1.5v?

Re-booting is usually faulty ram or cpu overheating, possibly power supply.

You may try running memtest on 1 dimm at a time.

Thanks for your message, and good point - I hadn't realized I was exceeding the default + 0.5V.

Woops, at first I thought you were talking about downclocking the CPU to 1600MHz, then I realized that was all RAM specifications.

Okay, I've set it to 1600MHz/9-9-9-24/1.5v in the bios. CPU-Z now reads DRAM Frequency 798.2 MHz, FSB: DRAM 1:6, 9-9-9-24.

I'm now wondering if my RAM was overheating, so I'm running with the case open for the moment. I'm thinking of buying an infrared thermometer. Any other suggestions for finding RAM temperatures?

Thanks,

Rockaday

 

[rockaday]

No new reboots so far. Thanks for pointing out the voltage. From what I read, people don't worry about RAM temperature as long as they use standard voltage, so at 1.5V I should just assume it's OK - does that sound right?

I may still get an infrared thermometer out of curiosity (found an inexpensive one), but I have no idea what temperature would be considered "bad".

Once I go a little longer without a reboot I'll start getting my system back to normal - add in the other SSD, close up the case, and start overclocking the CPU again.

Thanks,

Rockaday

 

[rockaday]

Best answer selected by rockaday.

 

[geekapproved]

You can probably speed that memory up to 8-8-8-20/1.5v/1600 without issues. If you do have issues, just drop back to 9-9-9-24.

 

[Mousemonkey]

This topic has been closed by Mousemonkey