Problem PC

Auraomega

Auraomega

Auraomega

Associate
Joined
22 Sep 2009
Posts
2,085
Location
Leicester
I made a post a few days back on the graphics section relating to an ongoing issue I've been having. They seem to believe it's another component and not the graphics card, so I'm posting here in the hopes someone might be able to impart their wisdom on me ;)

The original thread is here, I'm really frustrated but now that the seed of doubt has been planted I'm thinking it may well be hardware rather than software.

Just to outline the problems; graphics drivers crashing and restarting, applications crashing, PC restarting after screen messes up, hard drive corruption, after a while booting into Windows will give me just a black screen and a cursor, occasionally when powering on the fans spin but nothing else happens, Windows update fails to install certain patches, occasional graphical glitches without a crash. Another issue that is almost certainly unrelated (as it's happened to others as well) but might be worth throwing in, is that occasionally my sound card won't be recognised even by the driver installer (which reports no Xonar installed), as I say I'm not the only person to have suffered from this but might be worth noting.

The most recent issue I've had has been just now, playing Dirt the game kept locking up followed by the Windows pop up about changing to basic theme from Aero, obviously not as major as the other problems that crop up.

Something I read online was that a complete lock-up without a blue screen is indicative of faulty RAM, when my system crashes usually the picture screws up (imagine a smashed LCD with random splodges everywhere, that's what I get) then a hard reset.

Any suggestions let me hear them, I'll keep a list below of all the things I've attempted and the outcomes:

HDD - SMART; passed
Graphics - ATITool + Rthdribl; no artefacts or lockups
CPU - Prime95; stable as far as I can tell
RAM - Prime95 + Memtest; stable as far as I can tell
PSU - DMM; rails are well within tolerance and as far as I can tell don't fluctuate
 
Welcome to troubleshooting hell.

I've skipped through the other thread you've posted, i apologise if I repeat anything but there was a lot of advice being thrown around.

The file corruption is most likely due to the system not being shutdown properly repeatedly but we can't make that assumption. I use a program called spinrite (you boot it from disk) to check HDDs and repair them. Seeing as your system is unstable you might not be able to use other HDD diagnostic tools effectively. You could try running linux from the DVD rather than the HDD. If linux is running ok, you could question the HDDs health.

Does your mobo have an IGP? If you can remove your 5850 and use that or indeed beg borrow or steal another gfx card to test, we can establish if the gfx card seems ok.

If memtest is saying the memory is ok it probably is ok. I would CMOS reset to remove any overclock settings and ensure the memory timings are ok anyway.

I call it troubleshooting hell as for the most part you need spares to check if parts are faulty, can you source another PSU from a friend or family? It could well be the mobo that is flaky but it's best to remove all other possible causes before making that assumption.
 
Last edited:
I've used Spinrite before but no issue. Thing is it happens no matter what drive I use in the system (4 in total, 3 new), all 4 come back passing.

The system can work fine for months before crashing as described above at which point the whole thing falls apart, I'm inclined to say because of corruption of Windows files as sometimes using the basic tools (sfc, chkdsk, etc) I can get up and running again. Linux I generally have no problems with however one time attempting to restore my image of Windows my system kept crapping out on me, crashing, not finding the drives, generally not being able to put the image on the drive, etc, this was using Acronis. Again for the record I've reinstalled Win7 with a fresh install rather than the image to make certain nothing went dodgy on that first install.

No graphics inbuilt and I can't really use my other card as it's a baked one, so I'd never be able to tell if that card had just become faulty again or not, not to mention that it could be months until the system crapped out on me again and the 5850 is struggling to keep up with my gaming as it is so a cheap one is out the question.

The only other PCs in the house are all low powered systems so I can't borrow their PSUs, however if the PSU is suspect (OCZ ModXStream 700w, I know, I know) I might be inclined to grab a Corsair PSU from the MM to test with, if I was to do this what power would I need to run an i5 and 5850?

I've reset my overclocks more times than I can remember and my RAM is running at its rated timings or slower.
 
Oh it's an intermittant fault.....that makes things even worse. Even if you can return the mobo, gfx card etc there is no guarantee they will pick up the fault.

600W is the minimum I would spend out on. Although I'm not convinced it is the PSU. It doesn't hurt to have a good spare PSU for testing. So in my eyes it's not a wasted purchase even if your OCZ PSU is ok.

It is a case of swapping out parts till the problem is gone. Seeing as it happens randomly it's going to be hard to know if you have fixed it for sometime with each part.

It could well be the mobo. From experience I've seen plenty of different problems when they start to "pack up". I can't say for definate but it's interesting that the soundcard is failing to be recognised aswell.

I don't know the exact spec of your rig. If it was me I would upgrade the mobo to one that supports xfire properly (seem to recall in the other thread the other PCI-E slot was x4). Hopefully that would resolve the issue, if not you'd need to replace the GPU.

As you are saying that the 5850 isn't "man enough" for your needs it's going to be changed eventually. Atleast if you have a mobo that does xfire properly you have the option of adding another 5850 to improve performance if your old card is proven to be ok.......it all comes down to cost.

Good luck with it fella........it really does come down to luck as to finding which part is fubar
 
I've just had a worrying thought, my i5 and mobo were some of the first ones. Would I be looking at these kinds of issues with a damaged CPU? I'm aware that I could be sporting a Foxconn socket and if there's any possibility that the CPU is damaged I may well get the CPU out for a looking over.
 
How long did you run memtest for?

I had very similar symptoms on one of my older computers and it was caused by faulty RAM. Try running memtest overnight and maybe try one stick of a RAM at a time to see if it makes any difference.

Do you have access to any other components which you could swap to help diagnose the problem? I'd be tempted to try different RAM and GFX card if you can.
 
Last edited:
I ran for a good few hours but I guess I've nothing to lose by doing it when I'm at work. First things first though, updated the BIOS so I need to make sure my overclock is stable again, no point in testing the RAM when the CPU is unstable.

The RAM is some G.Skill 2000MHz 9-9-9-27 sticks (forget which ones), I've never run them beyond those.
 
Just a few things I have not seen mentioned,probably not but...
Kettle lead-try changing these ,computer and monitor
Have you tried in a different 13 amp socket,preferably not on the same ring i.e. if upstairs try downstairs or reverse
Can you try monitoring your 13 amp socket voltage with one of them plug in power meters,maybe you are getting power spikes or drops?
The above are most likely the least chance of being right but it looks like you haver tried most everything else :)
 
Funny you should say that... I get some minor issues with an amp and the whole house is getting rewired in a few weeks time because the electrics are so old ;)
 
This is the best chance you're going to have to suggest things to me; Windows OS fubar and trying to repair via Windows disk caused a BSoD. System was running P95 all last night until 30 minutes ago without issue, now it won't do a thing without crashing.

Voltages appear normal and my BSoD while trying to repair Windows from the dvd is a stop 24. The system will likely be unstable again unrolled I power down so going to get some tests started.

EDIT:
Don't like the look of memtest, getting fails across the board, it's just a constant flow of red (fails) ranging from between 2316.2mb and ~2800mb. Faulty RAM or faulty mobo? Going to have to try and do some CPU test now via a flash drive (netbook has no optical drive).
 
Last edited:
Check if there is any semi faulty capacitor on the Mobo.

Try reinstall the chipset driver for the mobo and must restart for the new chipset driver to take effect.

Hope this helps.
Bill
Tech manager
 
You mention it is overclocked.

Drop all the overclocks and check everything is seated properly, inc. re-seating the cpu/heatsink.

Then do a fresh windows install and see how long it takes to go wrong.

Also, a full spec. would be helpful.
 
All solid caps so I can't see if any have popped, the mobo isn't that old though and the computer isn't used 24/7 so I wouldn't imagine they'd go. As above issues are occuring outside of Windows also, so drivers aren't the issue.

Already reset back to stock and issues still arise so overclocking isn't the issue, besides the CPU is the only component not running at what it's rated to do. I've reseated the CPU in the past too, but it made no difference apparently.

Just about to run a bootable CPU stress test on the system now it's failing constantly. If it comes back passing I'll assume the RAM is at fault, if it fails then would I be safe in assuming that the mobo is actually the faulty component?

EDIT:
CPU tests didn't show anything, running memtest again gave a bunch of fails. Hard powered off the system, tried again and memtest is running perfect again. ARGH
 
Last edited:
I'm far from an expert on this with only 1 custom build to my name but I had similar occasional problems with my build, though mine always manifested as blue screens.

I too passed memtest numerous times even though my hunch was it was the ram all along. It eventually failed on the 4th overnight pass some months after I'd first encountered the problem after trying to troubleshoot almost everything else and new Corsair XMS3 1600 has totally sorted the problem, 100% stable for months ever since with no issues whatsoever.

You mention in your last post that your ram had multiple failures on memtest, even though it subsequently passed again. Even if it only failed once it means that you have problems with the ram, that and the wide range of issues that you're having really seems like the ram is at fault here.


.
 
OK... I'm going to admit. I haven't read all of this thread, so if you've done this or it's been suggested... just ignore me.

However, I did read the last post and he says you failed memtest once and then passed.

Have you tried testing the modules individually as one working one and one failed one can cause incosistencies.
 
Whenever theres a problem that's random and doesn't look like it's down to anything in particular it's usually the RAM.
 
Do you have a couple of spare sticks of RAM you could throw in to test out that you know for a fact are perfectly fine? If so then do that and run Memtest again and see if the sticks fail because if they do then I would say the motherboard is on it's a$$ or certainly very nearly.

It's the only thing I can think of :(

Stoner81.
 
Auraomega said:
All EDIT:
CPU tests didn't show anything, running memtest again gave a bunch of fails. Hard powered off the system, tried again and memtest is running perfect again. ARGH
Click to expand...

Sounds like a memory problem - i've had faulty sticks that would work fine on first boot but then would fail on subsequent boots. If i reset the CMOS they would work perfectly again until reboot...

As said - concentrated evil wrapped up in heatsinks.

As has already been mentioned, test each stick individually, and different slots, just to be sure...
 
Last edited:
I'm uploading a video I took of the failed memtest and will add when it's completed uploading. In the space of 20 minutes I had over 115000 errors. I'd have expected the sticks to have shown at least a single error after an hour or so were they this faulty.

Unfortunately I don't have any spare sticks, it is however one of the few components I'm willing to buy more of (Photoshop + Lightroom takes up a hefty chunk). If I can find a fault with the RAM installed and no error after swapping out then I will RMA these sticks.

Right now I'm running Windows Memory Diagnostic with the extra tests, the system is crawling on the cache disabled tests, not sure if this is normal or not?

Would testing RAM individually indicate an error that would be missed during normal testing considering it's 2x2gb sticks and I believe memtest is 32bit some of the memory would remain untested. The failed range was in the 2gb range, does this help pinpoint which module is faulty?

One final thing to mention which has started again; my keyboard (Razor Lycosa) keeps cycling through it's LED settings on it's own, on, WASD and off. There doesn't appear to be any interference from other cables as I've removed them from the vicinity of the other cables. Might not even be related but I figured worth mentioning.

My short time as a technician I found RAM to be one of the easier components to test, but we did have stacks of the stuff so it was easy to grab another stick and prove one way or the other.

EDIT:
Link for video.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom