Signs of failing hardware?

Toggle sidebar Toggle sidebar
J

james9120

Honorable
Mar 23, 2012
34
0
10,530
I've had my home-built PC for over a year now, and some serious problems are arising.
My PC will just instantly lock up, when doing the simplest tasks. It seems to do it most while browsing the web (such as Facebook). I have to do a hard-reset to restart my PC. But, sometimes it doesn't always stay frozen. If I wait 20-30 seconds, the mouse will sometimes "shudder" around the screen (bounce around) before coming back to life. Also, if I'm listening to music, it'll make a terrible noise when it freezes. After it comes back, I'll get an error message in the corner saying my graphics card (Nvidia) drivers have been restarted because they were unresponsive.
Also should note it's a guaranteed freeze when browsing the internet with Firefox, and when the computer turns off the monitor, it it will not turn back on. I'd have to do a hard-restart.

Any help would be great, here's my hardware:
ASRock 970 Extreme4
AMD FX 6100
Nvidia GTX 560 Ti
 
Sort by date Sort by votes
Maxx_Power

Maxx_Power

Distinguished
Jul 17, 2012
1,861
0
20,160


It sounds like a possible GPU failure. Check the GPU temperatures while the computer is just idling around (with say, GPU-Z).

Other than that, if the message is consistently related to the Nvidia display drivers, I would try to first clean the GPU heatsink/fan combo with some compressed air, and make sure it is getting adequate air flow, then uninstall and reinstall drivers.
 
Upvote 0 Downvote
J

james9120

Honorable
Mar 23, 2012
34
0
10,530

Thanks, I took your advice and cleaned out the years worth of dust that settled inside. The idling temps were roughly between 35-40C.


I'll give this a shot, thanks. I know all my important drivers (GTX, MB) are all up to date, but I'll keep checking!


I should also note that I'm not longer able to even reach closer to the OC numbers I previously had. And it seems if I try to go just barely above the factory settings, (like 3.5ghz) 2 of the cores would fail after a few minutes of a stress test. But they don't seem to fail with stock settings.

EDIT:
Spoke too soon... 10 minutes after a good cleaning, starts freezing and sputtering again.
 
Upvote 0 Downvote
Maxx_Power

Maxx_Power

Distinguished
Jul 17, 2012
1,861
0
20,160
The next best course of action is to try a different GPU in the same computer (spare one, or borrow one from a friend) OR try the same GPU in a different computer to see if you can reproduce the problem to narrow down what is causing the problem(s).
 
Upvote 0 Downvote
J

james9120

Honorable
Mar 23, 2012
34
0
10,530

It's a 1000W, I'm not sure the brand.


I wish I could, I'm the only one I know of that actually built a PC, not bought a pre-made.

I ran Prime95 for a bit, no problems (besides little freezes) everything passed. Then I ran 3Dbenchmark, and failed horribly. When it was about to begin the test, the GTX card just shut down, monitor turned off.


EDIT: Nevermind now... Restarted my PC, and was able to successfully run 3Dbenchmark and all it's tests with no problems.
 
Upvote 0 Downvote
marshallbradley

marshallbradley

Honorable
Sep 24, 2012
746
0
11,060
Sounds A LOT more like a dodgy power supply to me. As a general rule of thumb:

Bad graphics card -> artifacts in game/weird visual bugs, but doesn't usually just turn off (unless it's a heat problem, which doesn't seem to be the case, but have you checked the temperature under load? -- idle temps are useless dude)

Bad power supply -> computer just shuts down

It's a very general rule of thumb though, but in the most common cases I think it holds.

Also the fact that it's unstable when over-clocking after such a short period of time leads me to believe its something that affects both the graphics card/CPU (such as the power supply).

The fact that you don't know the brand is a REALLY BAD SIGN. Unbranded, generic power supplies have a tendency to do this. My guess is that it's probably not even delivering a third of its rated power at the moment, which is causing the issues. It may have permanently damaged components due to bad voltage regulation (which would explain the inability to over-clock), but let's hope not. The best course of action is to get that time bomb out of your PC, and replace it with a decent high quality unit, from a respectable brand like Corsair/Seasonic. This will do you very nicely: Corsair CX 500W.

In the meantime I'd recommend against using the PC, unless you really have to (and definitely don't use it for gaming).

All the best,

M
 
Upvote 0 Downvote
J

james9120

Honorable
Mar 23, 2012
34
0
10,530


Haha, I was too busy at the moment to look at the PSU, but I torn it down this morning. http://www.newegg.com/Product/Product.aspx?Item=N82E16817174028
That's what it is. It's probably at least 7 years old now, it was passed down, that's why I didn't know what it was.

As for the GPU, I hate this card and was ripped off, so it saw temps peaked at 90C and steady between 70-80C when gaming. Plus some pretty heavy overclocking before my PC started acting up.
 
Upvote 0 Downvote
Maxx_Power

Maxx_Power

Distinguished
Jul 17, 2012
1,861
0
20,160


Did you OC the GPU during your heavy overclocking ? What did you overclock exactly ?
 
Upvote 0 Downvote
J

james9120

Honorable
Mar 23, 2012
34
0
10,530

I only had the GPU OC'd, and possibly the CPU just barely. But only when I played games.

I should also note I had different problems months before this, when my monitor would go to into sleep mode, I couldn't get a picture to resume (moving the mouse) and hard resets didnt work either. I tried different ports, resetting the MB, unplugging things, and countless amounts of hard-resets, and nothing worked. Then suddenly it just decided to start working again. It was starting to get pretty frequent that this would happen.
 
Upvote 0 Downvote
Maxx_Power

Maxx_Power

Distinguished
Jul 17, 2012
1,861
0
20,160
Can you test the GPU in a different computer ? That's the only way we can know for sure.

On a different note, after you test the GPU, and if the problems are still there, you should try a different power supply. Your existing power supply is actually reputed to be okay to quite good at JohnnyGuru (gurus with PSUs):

http://www.jonnyguru.com/forums/showthread.php?t=5319

Still, it can fail, so if you test the GPU in a different computer, and you don't get issues, then try a different PSU.
 
Upvote 0 Downvote
Maxx_Power

Maxx_Power

Distinguished
Jul 17, 2012
1,861
0
20,160


Hmm, that makes it hard to "divide and conquer", literally. Would it be possible to hop down to the local computer/parts store (Bestbuy, TigerDirect, whatever) and buy a quick video card to test ? I suppose/think you can return it if it wasn't the problem, legitimately, or keep it if you prefer.

Other than this, you are kind of gambling on what part to replace without thoroughly testing it first. You can replace both the PSU and the GPU (already a bit investment) and keep your fingers crossed that things will be fixed, but that's blind faith that your problem isn't with the PCI-E slot on the motherboard, for instance.
 
Upvote 0 Downvote
J

james9120

Honorable
Mar 23, 2012
34
0
10,530

Are there any PSU tests I can run? Or stick a voltage meter in one of the available cables and see if there's a jump or dip in voltage?
Also, if the GPU was causing problems, would that affect SOUNDS coming from the PC? For example, when I'm listening to music, and my PC starts to lock up, the sound will freeze making that horrible ERRRRRRR noise.
 
Upvote 0 Downvote
Maxx_Power

Maxx_Power

Distinguished
Jul 17, 2012
1,861
0
20,160


I'm not sure why the GPU freeze would cause sounds to lock up, but in many situations I have observed a freezing PC or frozen PC playing looping sounds of whatever was playing at the time of the crash. It may have something to do with the way Windows handles real time content, as in chunking.

As for thoroughly testing a PSU, this isn't easy. You can try a multi-meter and test the yellow against the black and red against the black for voltage tests of 12V and 5V rails respectively, but that isn't going to tell you much, unless the rails are seriously below or above spec. Multi-meters are not usually fast enough to capture transients, you need an oscilloscope for that to obtain voltage as a function of time (when you load the PC, or whatever).

Do you have onboard video ? If you do, you can remove the GPU and see if the problem goes away with the onboard.

Without a spare part/computer to test against, it is really difficult to isolate a given problem, but if you want to try an isolated approach, your first bet is to un-overclock everything, including the CPU. Put every setting at its optimal (for stability) setting. Then do a MemTest86, and let that run for overnight to see if there are issues with the memory. Then try another 24 hours of IntelBurnTest on max settings to see if there are CPU/thermal/VRM issues on the motherboard. After that, you can confidently test the GPU, with say, OCCT for a few hours. You can then try the OCCT's PSU test (ONLY AFTER you have individually tested all other components, to ensure that the failure is only due to the PSU, and not something else).
 
Upvote 0 Downvote
J

james9120

Honorable
Mar 23, 2012
34
0
10,530


Thanks, I will do each one of those things, starting tonight.
Weirdest thing... I left my PC unplugged all day, and hooked it up after I got home. Haven't had a problem yet... And it's been running for 4 hours. This is really getting frustrated!
 
Upvote 0 Downvote
Maxx_Power

Maxx_Power

Distinguished
Jul 17, 2012
1,861
0
20,160


That means there is some problems with your memory AND/OR memory controller.

To break that down, it can mean a few things, since the memory controller is embedded in the CPU these days, and settings in the BIOS affect it.

1) Your memory is/has gone bad.
2) Your settings for the memory (timings, voltages, etc) are incorrect.
3) The memory controller clock (can be independently controlled on AMD systems) is too high, or voltage too low, likely due to raising the base clock on an AMD system.
4) An occasional, rare defect in the motherboard.

Is the CPU still overclocked ? Set all the voltages for the CPU/memory in the BIOS to whatever default values they are supposed to be (wouldn't hurt to use the BIOS option to reset all to defaults).
 
Upvote 0 Downvote
J

james9120

Honorable
Mar 23, 2012
34
0
10,530


Everything has been at default, and when I ran the test.
I did the intelburn test through out the day, I saved the log and everything appears to be OK.

So it's looking like my RAM might be messed up?
 
Upvote 0 Downvote
Maxx_Power

Maxx_Power

Distinguished
Jul 17, 2012
1,861
0
20,160


It very well could be. IntelBurnTest dominantly tests the CPU and its stability. You MIGHT notice test failures if the test strength is set on max (to use nearly all RAM), but because the RAM in windows is shared between the application (IntelBurnTest in this case), and the operating system, the portion used by the operating system is not accessible to be tested. MemTest86 specifically tests the memory, and because it does so with a very low presence in the RAM (unlike windows+background applications), nearly all of the RAM is tested, and failures should be considered conclusive evidence that something is not right with the memory.

Usually if you pass 24hr of MemTest86, you can consider the memory configuration to be stable. If you have any errors (particularly in your case, where you have repetitive errors), you should consider testing the sticks 1 at a time to see if it is specific to each stick, or to the combination.

If the errors remain with each SINGLE ONE of the 2 sticks (or 4, not sure how many you have) in any slot, you probably have a memory module (stick) that has issues, such as degradation (to the point they can nolonger run at specs), or hard failures (where certain addresses are always faulty, regardless of speed).

If the errors go away with single sticks and does not return FOR ALL STICKS TESTED ONE AT A TIME, across the memory slots, you have a dual-channel specific issue, and could be due to memory controller/motherboard/BIOS.

Once you determine what the situation is, you can go further to fix it.
 
Upvote 0 Downvote
J

james9120

Honorable
Mar 23, 2012
34
0
10,530

I haven't had a chance yet to test each individual stick, but I did come across something interesting. Every time I used firefox, or had it running in the background, I would have the issues. So I started using IE and Chrome, and haven't had a single issue in day's. I've played games, music, photoshop, etc all with no problems. But as soon as I open firefox, the fun begins. Has anyone heard of such a thing? I've been using firefox all my life lol.
 
Upvote 0 Downvote
Maxx_Power

Maxx_Power

Distinguished
Jul 17, 2012
1,861
0
20,160


Firefox can be more memory intensive (larger memory footprint) than Chrome or IE, based on experience. If you have plugins installed, that memory footprint grows a bit too.

Still, if you have MemTest errors, that's a hardware (or configuration) issue independent of software. Some software might be more sensitive to the underlying error and crash, while others may not be as sensitive. Ultimately, if you have RAM issues, don't trust the computer to do anything important like video/audio encoding (you will get errors in the end files) and anything work related until you have fixed the underlying problem.
 
Upvote 0 Downvote
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom