Take the 2-minute tour ×
Super User is a question and answer site for computer enthusiasts and power users. It's 100% free, no registration required.

I just recently (last week) built a new PC. Upon initial assembly, I had issues getting it to power up. When I turned on the power supply nothing seemed to happen. I tried disconnecting and reconnecting all the power cables from the power supply to the mobo and this seemed to fix the problem. After getting Windows set up and Skyrim installed (I built this computer so I'd finally have a gaming system) I set about to immerse myself in sweet RPG goodness but soon encountered a new problem. The machine keeps turning itself off. Sometimes I can run the game for hours without incident before it shuts off. Sometimes it takes only minutes, but it will always happen eventually.

My first thought was that this might be an overheating issue (Actually my first thought was that my graphics drivers were out of date, but they are not). I tried running the game with maximum settings and a temperature monitor in the background. I never saw any evidence of overheating (the highest temperature I could reach was 55C) and my crashes did not appear to correlate at all with GPU load or temperature. I did the same with a CPU temperature monitor. Nothing.

I have, however, noticed a few patterns:

  1. I do not lose power while the system is idling. I have left it on all day without incident. The problem only seems to occur while running Skyrim. Though in fairness I've used the computer for little else, I've browsed the internet for hours on end and watched movies on DVD and Netflix without incident.

  2. After a shut down I cannot just power back up. I have to power cycle the power supply and then the system starts up with no problem.

  3. When I do lose power the power light on the mobo remains lit.

Here are my system specs: OS: Windows 7 Professional SP1 RAM: Patriot 1866Mhz 8Gb CPU: AMD FX-4100 3.6GHz Quad Core Mobo: Asus M5A97LER2.0 Graphics: MSI nVidia GTX-650ti Power Supply: Raidmax KY-600ATX 500W

My intuition is that this is an issue with my graphics cards or power supply, but I'd rather not start burning money I don't have replacing components if there are other things I should try first. If anyone could provide some guidance on what I should try next I would greatly appreciate it.

UPDATE: After walking away for a few hours I came back to find that the computer had apparently recovered from a BSOD in my absence. This is new! Here is the error message I received:

Problem signature: Problem Event Name: BlueScreen OS Version: 6.1.7601.2.1.0.256.48 Locale ID: 1033

Additional information about the problem: BCCode: 116 BCP1: FFFFFA8006AE6010 BCP2: FFFFF88011B63630 BCP3: FFFFFFFFC000009A BCP4: 0000000000000004 OS Version: 6_1_7601 Service Pack: 1_0 Product: 256_1

It also referenced two error log files which I cannot open because I get a "Permission Denied" error =/

@SmirksWhileWalkingWC: Thanks for the input and sympathy! I did not consult any compatibility lists before purchasing my hardware, though the specifications at face value indicate they should be compatible. I'm aware it's a bit of a gamble building from scratch though as I've encountered some sneaky compatibility issues with previous builds. I'll do some research on that and see if I've rolled the proverbial snake eyes with one of my choices.

I have a hard time believing it's an overheating problem for a couple reasons. First, when I've had that problem with other machines in the past, the affected part has been hot to the touch after a shutdown. In this case the CPU and GPU (I haven't checked the RAM) have been cold after many of the shutdowns. I also don't see any correlation between the time since the last crash and the next one. I've had it die within 15 minutes after a cold start and restart within 30 seconds of a crash after a couple hours of game play. I also don't need to give it any time to cooldown to restart, but no matter how long I wait, if I don't cycle the power it won't restart. I hadn't considered the RAM though, so that's an interesting thought. I can swap in some new RAM and see if it performs any differently.

UPDATE: I have now tried the following fixes with no success:

  1. Replaced the RAM - The motherboard user manual indicated a potential compatibility issue with the RAM I initially selected. I replaced it with RAM from the recommended memory list in the manual.
  2. Replaced the mobo - I had someone take a look at my system and they noticed that my memory frequency was being misread by the BIOS as 1333 instead of 1866. He suggested this may indicate a problem with the motherboard. After replacing both the motherboard and the memory, the memory frequency now registers as either 1333 or 1436 (it is seemingly random) rather than the correct 1866. Note that the mobo has misread the frequency on two different sets of RAM from different manufacturers. This has occurred with new and old mobo.
  3. Replaced the graphics card
  4. Replaced the PSU - Power supply voltages are correct in BIOS but I hooked up a different power supply anyway and replicated the same crash problem I've had.
  5. Flashed the BIOS. - The mobo manufacturer had a BIOS update from just a few days ago. After installing this, the memory frequency is still misread and the same crashing problem occurs. The new BIOS did however give me the option of setting the memory frequency to 1866 (this did not appear as an option in the old BIOS) but doing so caused my system to not boot at all.

This just leaves me with the CPU, which I'm replacing tonight. If this does not work I'm pretty much at a loss for what else I can do. Does anyone have a long shot idea for what could cause a consistent problem like this even after changing out every hardware component? Something I haven't considered?

share|improve this question
1  
If its a proper shutdown, event viewer may have something of use. –  Journeyman Geek Jan 5 '13 at 1:42
    
Did you customize any BIOS timing or voltage settings? –  David Schwartz Jan 5 '13 at 3:48
    
@ Journeyman: I checked event viewer and there's nothing helpful there. The only event generated is the one created when the computer recovers from an unexpected shutdown, which has not information other than 'hey, that happened' @David: Nope, I never got that far –  Josh M Jan 5 '13 at 8:19
    
If its possibly ram, hit it with a couple of rounds of memtest86+. You can also troubleshoot bluescreens easily with bluescreenview nirsoft.net/utils/blue_screen_view.html or whocrashed resplendence.com/whocrashed –  Journeyman Geek Jan 5 '13 at 15:06
    
I don't know for sure, and I'm not familiar with your exact motherboard, but it isn't entirely uncommon (unfortunately) for motherboards to downclock RAM when you max out the slots. So the downclocking of the RAM may be correct in some sense. If you still haven't figured this out (I see the question is pretty old), try removing all but one or two of the RAM sticks and see what happens to the memory bus frequency. –  Michael Kjörling Sep 25 '13 at 7:32

3 Answers 3

Did you assemble those components based on already verified compatiblity lists from others or bolted the best parts you could find within your budget ? Although it is a reasonable expectation to assume parts should work together, custom built systems based on desire instead of benchmarks tend to have these issues more than something from a WHQL verified system from Dell (bleeechhh)

You've hit the problem with a donut shaped hammer properly encircling it for checking your temps and the fact it is a custom built box throws a big variable into the cause. the way it shuts down sounds exactly like an overheated CPU or defective RAM and I'm more likely to guess RAM since it has a tendency to start from a cold state, then fail and fail sooner and sooner at every subsequent reboot until given a chance to cool down since RAM still isn't as adequately cooled the way CPU's are but is working nearly as hard and seems to keep getting hotter with every new generation, but you've checked the temps, lets just make sure since different motherboards have different thermostats and you've verified your CPU monitor is compatible with the latest and greatest motherboard. Verifying your temp ut works on that board is crucial to ensure you aren't missing temperature outputs since your symptoms sound exactly like a CPU going into temp panic, then waits to cool down before restarting.

Once you've verified your temps are inside the limits, try to EKG it using Procmon.exe from Microsoft's SysInternals and enable the log file feature and registry, disk and every other sensor it has, it will write alot of data out but its thorough and I'd focus on the last events logged at shutdown espeically the last 10 -20. collect a few of those and notice the consistencies such as a particular file being accessed at death, or syscall taking place and you'll gather the clues.

I'm assuming you aren't overclocking anything (unwaware if that AMD can be overlclocked) or have stopped overclocking after the 3rd shutdown that begain squishing your budding happiness for completing your build, what is happening sucks and I feel for you on a personal level because troublkehsooting it instead of just enjoying it feels as crappy as staring at your new ferarri stuck in the driveway because the alarm won't let you inside... bitter sweet and good thing you don't live next to a water tower with a footlocker full of ammo along with a collection of high capacity mags and that are about to be banned.

You can get to the bottom of this, and its impossible to determine how much time it will take but just following the money, everything at this moment hints at defective / incompatible RAM, then a CPU temp event that might be getting missed and consider using three different temp monitors to verify them against eachother for the temp levels and whether they all detect the same thermistors.

share|improve this answer
    
Thanks for the input and sympathy! I did not consult any compatibility lists before purchasing my hardware, though the specifications at face value indicate they should be compatible. I'm aware it's a bit of a gamble building from scratch though as I've encountered some sneaky compatibility issues with previous builds. I'll do some research on that and see if I've rolled the proverbial snake eyes with one of my choices. –  Josh M Jan 5 '13 at 7:42
    
stop that downward thinking, without it we wouldn't have the jet engine. I don't know what Skyrim is but can you use it in LINUX ? If you had to narrow the cone of possibilities between lets say Europe or North America, but in your case is it a HW issue or windows driver issue, can you determine if the reboots occur in LINUX ? You can use a LIVE LINUX booter (research it if you don't already know, fascinating stuff) and it will help us isolate which continent to start on. You're better off with a HW and non windows issue, windows has terrible device error interpretation and will just confuse –  SmirksWhileWalkingWCabaretGirl Jan 5 '13 at 7:44
    
But, if this is a thermal issue, LINUX is less likely to cause it because it's just better at resource manaement, if you hadn't the chance to use both, run a laptop from a bootable LINUX cd and then on a stripped up, tuned out XP box and that XP box even if using less RAM and CPU will still scorch your knees and the LINUX box will feel un-naturally cool even if the CPU and RAM drain is higher. It doesn't make sense, but what if anything about Microsoft Windows has made sense in the last 20 years since the commercial availability of the NT kernel still used today in Win8 and Server 2012 ??? –  SmirksWhileWalkingWCabaretGirl Jan 5 '13 at 7:50
    
Not enough space to comment, but as a long standing debate remains immortal about fixing windows drivers is harder than diagnosiing physical hardware.... visit the Microsoft technet forums at any time of any day and you'll find it replete with the same issues of hw errors, driver vomit and the same stuff that hasn't gotten better. Windows still takes too long to shut down and always refuses to turn on when needed the most and thats from a guy who has wasted the last three days, solid, on heavy doses of aderall and still unguckinhly able to get Windows 8 to properly install. I'm entitled. –  SmirksWhileWalkingWCabaretGirl Jan 5 '13 at 7:55
    
I cannot run Skyrim on Linux, at least not without an emulator (I used Wine in the past, but it was a nightmare to configure) but I could burn an Ubuntu live boot disc and give that a shot. I haven't used Linux in a good ten years though so it could be a real adventure! I'm sure I can find other ways besides Skyrim to tax the system, especially since my issues seem to now extend beyond power failure while gaming (see update) –  Josh M Jan 5 '13 at 8:05

Did you install MSI Afterburner? NOT FOR overclock, but GPU FAN CONTROL.

Following is download link for N650Ti-1GD5/OC

http://www.msi.com/product/vga/N650Ti-1GD5-OC.html#/?div=Utility&os=All

Base on my personal experience and poking around various forums, the base Nvidia video drive does not control the fan. When playing 3D games, you WANT TO HEAR THE FAN ROARING!!

I have a EVGA nvidia GTX-285, before I installed EVGA Precision X (equivalent to MSI Afterburner), whenever I play 3D game, the graphic may go choppy after some time, then it will either froze completely or just turn complete black. The only thing I can do is reboot.

After I have installed Precision X, I discover that when I was playing 3D games, the GPU temperature can shot up from 50C/60C to 90C+, even over 100. That is when the problem start. After I set the fan speed to step up proportion to GPU temperature, the problem go away completely. It keep the temperature below 75C. Non gaming temperture of my GPU is ~50C now.

share|improve this answer
    
It might also keep track of what the core temperature is. Asus has an identical software - and all of them seem to be based off rivatuner. Didn't realise that was the main point of them. –  Journeyman Geek Jan 5 '13 at 5:22
    
As I said John, there is no evidence that the GPU is overheating. I have MSI Afterburner installed (indeed, not for overclock) and it is what I have been using to monitor my GPU temperature while trying to reproduce the shutdown problem. I have yet to see it climb north of 55. Good to know I can use Afterburner to improve my fan control if I do start seeing overheating problems though. –  Josh M Jan 5 '13 at 7:39
    
arrrr, some how I missed that piece of info >.< –  John Siu Jan 5 '13 at 13:51

I've had this issue twice before.

My first video card was corrupting wireframe models. I deemed it to be a faulty video card, probably a loose solder and exchanged it. The next model of the video card had arrived in stock, so I upgraded to that.

This video card worked great, until I had the sudden shutdown during games, like you are experiencing. It was only during games, and the more intense the game, the sooner it occurred. I realised this card required more power than its predecessor and was overloading my PSU.

Back to the store I went. I exchanged it for another video card the same as the first. This one worked properly ever since.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.