Pc keeps black screening and crashing. Needs help.

MumboJumbo

MumboJumbo

MumboJumbo

Soldato
Joined
2 Dec 2006
Posts
8,204
Hello, well i'm having a lot of problems all of a sudden so I will break it down:

1) I had these problems before with my old 8800GTX. I used to have the screen randomly turn black as if the driver had crashed and it was rebooting it. Then after about 3-5 seconds everything would be normal again. This would happen in day to day useage. The graphics card died about 1.5-2 years after.

2) New card 285GTX. Quickly developed the same problem again where it would randomly give me black screens for 3-5 seconds. Now about 1.5 years old.

3) New card has developed a new symptom I never had before. This has been happening in games now for about 1-2 weeks where the screen goes black and the audio loops. I have to turn the pc off and back on. It's worth noting that my monitor actually states that it has no gpu connection. So the gpu is actually cutting out completely and the audio is freezing.

However just now it happened in day to day useage too. My pc is clearly on a crash course to dieing very soon.

Im thinking its not a graphics card problem as such because this has happened twice and the problems happened so quickly after getting a new card that it's something else causing the problems. Possibly powersupply?

Spec:

-W7
-q6600 overclocked @ 3ghz now as that became unstable at any higher overclockers a few years back having been run at 3.6ghz previously.
-285GTX
-4gb ram
-Seasonic m12 600w powersuply

thanks.

:EDIT:

I've tryed updating the drivers. I uninstalled the old ones and nothing could work out what graphics card I had. The nvidia drivers wouldn't install. So I had to let windows install what ever it wanted and install the latest nvidia drivers over teh top.
 
Last edited:
i'd have to steal my mates powersuply. I built his pc and it has the same one. But I some how suspect he's quite attached to it haha. I spose i'd have to lug my pc round to his and try run a game on it off his powersuply. If it doesn't crash within 20-30 min its the power suply i recon.

It worries me as it probably means its the powersupply thats been killing my graphics cards and this one is gona go too at this rate. May also be why my overclocks are more unstable now.
 
you should google you power supply and see if any one else out theres having issues

with it also..... may just be your one though....sorry i cant be more help fella
 
Well seasonic do some of the best powersupplies and its now 4 years old so if they had a dodgy batch it would be hard to track down now if it was mine. This is annoying, my pc is like a ticking time bomb it seems. Already specing up a new one as I figure its inevitable lol.
 
Defo PSU issue and it's not a fault...... Its well known that PC's that run under the advised power rating for each component over time that part will degrade overtime and fail..... one of the signs is a black display for a few seconds then goes back to normal.
I've had years and years of this issue in the company I work for, years ago I was head of the PC production manufacturing and orders came in for cheap PC's for mass orders and bigger bosses being profit happy decided to cut corners by buying in cheap PC cases with low rated PSU's in them Where talking 150 - 200 Watt PSU's and they where trying to run them with Intel All in M/B's running P4 CPU's socket 478 from 2.6Ghz to 3Ghz (remember the ones that ran hot and liked to heat your room LOL)
Anyway to keep me going on, Intel issued an advisory with all Intel branded M/B's and today still do stating the correct rating of PSU for that M/B they also stated that they will work with less rated PSU's but overtime of Under Voltage/Wattage the components will degrade and reduce the expected life of the M/B and the rest of the components.
anyway after 6 months of being overruled by my bosses they finally listened to me and started to use the correct PSU's..... by this time they had sold 1000's of units and as I predicted over the next 12 months 76% of the PC's that ran with under rated PSU's came back as faulty with no display as apose to the ones that ran with the correct rated PSU's only 2% came back as faulty, My bosses were not happy but as i stated to them well before a "I told you so" they did not find funny.....LOL

anyway sorry to bore you with the above but basicly you need to test this theory by either testing your rig with a high rated NEW PSU or you can drop the amount of load is on your PC by unplugging unneeded items in your PC..... like CD/DVD drives extra hard drives, extra PCI or PCI Express cards even FANS just run it bare with PSU, M/B, CPU, Ram, Video Card & the HD with your O/S on it.... run your game and see what happens.....

another thing you can do is use a PSU calculater..... you will need to know the required power rating of your M/B, CPU, Video Card, ram, HD etc....... also your PSU spec will also have wattage and ampage ratings for each of it's connectors... the ATX connector the P4 connector plus the ones used for optical drivers molex connectors.
Im not talking about Voltage as they all run the same at 12V and 5V it's that Watts and Amps we need to know if your PSU meets the requirement.

For Example a my new nVidia GTX460 states in the manual the below
"The ASUS GeForce GTX 460 video card requires a 450 Watt or greater power supply with a minimum of 24 Amps on the +12 volt rail"

At first I ran this on a 500Watt PSU that only gave 20 Amps on the +12 Volt rail, and guess what? I had the same issue black screen whlie running games, videos or even just at windows desktop for a few secs...... then I bought and installed a new PSU that did the was was 650 Watt that gave 24 amp on the 12 Volt rail and bingo PROBLEM SOLVED.....

Hope this helps..... Note M/B's also state an Wattage and or AMPage rating so check this also..... as My M/B required at least a 650Watt PSU too.
 
Mister Nutts may be right, and he may be wrong. Don't believe him when he says defo.

I think is far far more likely to be faulty memory and I'd suggest that you run a memtest ISO before you do anything else.

Your PSU can supply way above what your system needs, so unless it's faulty it won't be the PSU. It's quite true that PSUs do degrade slowly over time, but to degrade that far it's unlikely.
 
Yeah I wonder if its the psu but if it is then its not because it doesn't have enough wattage or bad quality. It is an extremely good quality psu and cost £120 at the time. But it may have broken as some things do.

Do you really think memory could cause this? I will check it tomorrow tho.
 
Memory is very easy to check. Just grab the ISO image, burn it and let it run for a few hours. Usually it;'ll crap out in the first hour if there's a problem.
 
Had to go out and buy a cd as couldn't find the ones we have. Going to run it in a bit. Failing this I guess its like the psu. Though tbh my mobo is pretty dodgy too I think so I wonder if its possibly that. I'm not really sure how to tell which is malfunctioning without being able to put my psu into another pc.
 
56% of the way through so far all ok. Good signs as I beleive if you have a serious ram problem it won't make it through the first test let alone others and my problems do seem to be rather serious. Though come to think of it i'd rather it was a ram problem.

:EDIT:

now with a test and ahalf complete. Looks like its on to the next idea, shall leave it to finnish it's second test.
 
Last edited:
Nvidia did some faulty (basically designed to break up because of material choises) manufacturing during 8000-serie so its failure could come also from that.

Also CPU's overclocking could have overstressed voltage regulation circuitry in motherboard if you raised voltage lot.
Four DIMMs puts more stress on motherboard's northbridge (memory controller) along with rising FSB especially if there's been also overvolting so also that part could have degraded.


If it's failing PSU which is struggling under load causing those crashes running Linpack (Intel Burn Test, Linx or OCCT) and FurMark at the same time should make it fall on its face by causing lot higher load than games.

Memtest isn't that good at detecting instability.
Prime's Blend/Large-FFT tests whole "memory system" better and Small-FFT is good for testing internals of CPU (and its voltage regulation) as when data fits to cache CPU can keep crunching numbers instead of often pausing to wait data from RAM.
 
Well my mobo defo should have been able to take the voltages I put through it as it wasn't anyting too high but it has always been tough to keep stable at higher FBS and personaly felt it was a dodgy board right from the start but was never broken so not really able to RMA it.

Also i get the black out and repeating audio sometimes just from the first few minutes of l4d2 so I don't think I need to stress test it to comfirm that it crashs.

Memtest managed 2.5 passes before I stoped it and I beleive we can asume it is not the ram causing this.
 
Currently running prime 95 to check my cpu/ram and temps. Think this shall possibly comfirm gpu/psu. Interestingly speed fan seemed miles off with the temperatures by 10-15 degrees. At least I trust coretemp to give me the right readings and it disagrees with speed fan, which is giving low readings of 57-62 degrees while running max heat/power prime 95.


Do you think that this draws enough power from the psu to suggest its not a psu problem? I'm thinking not. But the only way to draw the power for that is through 3dmark, which will crash also if my gpu is fubar.
 
Last edited:
8igdave said:
Memtest managed 2.5 passes before I stoped it and I beleive we can asume it is not the ram causing this.
Click to expand...

Memtest ideally should be run overnight especially as you have an intermittent fault, a couple of passes is not really enough to guarantee stability.

How long to test: Unlike other memory checking software, MemTest is designed to find all types of memory errors, including intermittent problems.

To catch intermittent errors, however, you need to run it for several hours. You can also run MemTest Pro while you use your computer for other tasks, which will help identify memory errors which only show up while the computer is under load.

100% coverage represents one full pass of testing your memory. In general it is better to run multiple passes. Here are three typical lengths of testing you might use:

1: Test until 100% coverage (a quick test to make sure your RAM is functioning reasonably)
2: Test for 1 hour (this will catch everything except intermittent of errors)
3: Test overnight (recommended; your computer is not doing anything else at night anyway, why not be absolutely sure your RAM is good?)
Click to expand...

taken from the Memtest manual
 
Well considering my serverty of my problems I do not believe that I would need to run memtest over night just to catch them out considering I can't play a game for longer than 3 minutes sometimes before a crash and its getting worse each day it appears.

I'm also currently running prim 95, which would be stressing the ram also and have yet to have any problems.

I will run it over night if I can't get anything solid by the end of tonight. I'm thinking 3dmark is going to crash quickly.
 
Just had BSOD 20min in. Last time I had checked the temps they were at 70 but that was only 10min in so I guess they could have risen a bit more but I don't think enough to have caused the crash. At least when I originally overclocked I tested them to not peak above 75 under full load.

So what does this mean? Does an unstable cpu overclock cause these symptoms or are we back to ram? I will run memtest over night to make sure either way.

:EDIT:

Going to put my cpu back to stock settings, the ram was never overclocked anyway and infact slightly underclocked to begin with. I'm then going to play some l4d2 as I know this can cause crashs very quickly and see if it fixes the problem.
 
I notice there's been no mention of hard drives in this thread, what hard drives are you running 8igdave? Are you running RAID?

I take it you've run a full check disk on them as well as any manafacturers software/smart test?
 
You must log in or register to reply here.
Back
Top Bottom