Artefacts during POST, Windows + games.

[nadnerb]

I'm having a few problems with a PC I built recently.
I bought all the parts new about 2-3 months ago, built it and it has been working fine up until the last week or so.
I basically bought the same parts used in this review, with the intention of OCing it eventually. (however, everything is currently running stock)

Intel e2160
MSI P35 Neo2-FR
2GB Geil Black Dragon DDR2 800
PowerColor HD3850 512MB
320GB Seagate IDE
Samsung DVDRW SATA
Antec Earthwatts 500W
Antec Sonata III
Windows XP SP2

So, onto the problem...
I was playing Stalker for about 20 mins when suddenly some artefacts appeared in the sky and then the screen went black.
I could still run around and could still see the map and ammo etc on the screen. I went to save the game and reload but it just froze.
I reset my PC and clicked my name on the login screen but the monitor went black and displayed "no signal"
The PC was still running as I could plug in USB devices and hear the usual sounds, it was just as if the video cable had come out.
I reset the PC and managed to log in fine.
I checked the GPU temps and they were about 40'C which is about normal considering it idles at 35'c and I was just playing a game a few minutes earlier.
I then loaded up Stalker and on the menu screen were loads of artefacts - the PC then crashed.
Reset again but this time there was artefacts all over the screen right from POST and I don't think it would even load Windows.
Reset again and the same problem.

I thought that it's possible the GPU may have overheated (although I'm 99% sure it didn't)
Left it for a few hours, started it up and it asked for a BIOS password (I've never set up a BIOS password! wtf?)
Reset the BIOS using the switch on the mobo and the password prompt is gone, but everything is still garbled.

Contacted the seller and they said to RMA the graphics card.
I then noticed the IDE cable wasn't plugged all the way into my HDD so I unplugged it and put it back in.
This seemed to solve the problem as it booted fine with no artefacts! weird!
4 days passed with no problems. I spent quite a bit of time using the PC doing uni assignments (but no games) without any problems.
Left my PC for about 20mins and came back to a garbled Windows Flag screensaver and the resolution had been lowered.
Shutdown and restarted but it seems now there was no video output at all, the monitor just displayed "no signal" and didn't show anything, not even POST messages.
The monitor I'm using is an old Sony lcd via a DVI-> VGA adapter.
I then tried plugging my TV into the PC via a DVI-DVI cable.
It worked, well kinda....

Managed to get into Windows and load up ATI's Catalyst Control Centre which then told me the driver was not compatible and to reinstall it.
Since a corrupt driver wouldn't effect VGA output of POST and BIOS I didn't think reinstalling it would help and it didn't!
I then got a BSOD that mentioned ati2dv..... something... hmmm....

Someone mentioned to me it could be a dodgy power supply so I unplugged all the power adapters and tested them with a multimeter and found they were all pretty much spot on what they should be. Plugged everything back in and the damn thing works again!
Today I played a few hours of Stalker on highest settings and got no problems at all.
I scanned for artefacts with ATI Tools for 30mins and got nothing.
Ran a memory test and got no errors.

Any ideas what's causing this?
At first I thought it was the GPU but now I'm not sure.

Sorry for the long post and thanks for any help!
Cheers!

 

[DarthPiggie]

Yes this does look GPU related. You might want to try a different driver. Do you have a backup graphics card, or an integrated one? might want to try and see if you still get the error.

 

[nadnerb]

Oh yeah, I forgot to mention I haven't been able to swap any parts as I don't know anyone with PCI-E or PCI cards.
All my friends at uni have laptops or onboard gpus.
Cheers

 

[grieve]

Based on the information provided I also think it sounds like gpu issue.

Know anyone at all with a pci-e slot to try your PowerColor HD3850 512MB in? Maybe pop it out and go to some tech store and have them test it for you if you have no replacement and nothing to test it in.

 

[nadnerb]

Nope, don't know anyone with PCI-E - I'm not really a gamer TBH.

I was thinking of taking it to a computer shop to test it, but I'm not sure if that would prove anything.
I mean it's currently working in my PC now but I'm kinda expecting things to go wrong again sometime soon.

I can RMA it, but they said they'll charge me if it's not faulty and since it's working now it will probably work for them when they test it too.

Cheers.

Could it possibly be a PSU issue?
Both times I managed to 'fix' it was by unplugging and plugging in other pieces of hardware.

 

[DarthPiggie]

It might be i guess, but i doubt it. 500W is more than enough for that system

 

[romulus47plus1]

The PCI-E socket on the MB. It may be broken.
Or the PCI-E plug of your GPU.
My friend experienced a similar problem.

 

[nadnerb]

Could be I suppose, although I tried the second PCI-E slot on the motherboard and had the same issues.
I'll make sure to clean the contacts on the GPU card next time I open it up though.

BTW, I've just been running ATI Tool artefact scanner for an hour and got no artefacts at all, 45c average temp with 100% load.
Doesn't that suggest the card is in fact ok?
Are there any other kind of tests/software I can run to test the card?

 

[romulus47plus1]

Send it to the repairman, they know the best.

 

[grieve]

If you don’t have any other "test beds" your pretty much SOL guessing what’s causing the issue...
Hate to tell you but RMA or get it tested somewhere, as romulus47plus1 mentioned it very well could be the MB also.

-I highly doubt it is a PSU issue, those aren’t really the symptoms of a bad power supply.

 

[nadnerb]

It's been working for the past 4 or so days now so fingers crossed.
If it happens again then I guess I'll take it to a computer shop and see if they can test the parts for me.

 

[gomerpile]

have a card with the same thing a 5700 and it is the memory timing 0 has lost its timing, I tested this with nibitor and when I set the timing an error that said timing 0 failure. This was caused by ocing and heat. I put the card in my second computer and never set the clock speed back. That computer did not have the cooling I had
or psu
Many things can cause this, inproper heatsink app., psu,heat,ocing,power,memory,fastwrites,drivers/dx, try to reseat the card or disable directdraw and fastwrites.

 

[nadnerb]

So it's been working fine for the last week or so including yesterday when I was playing STALKER for several hours on high settings on my 1680x1050 monitor/TV via DVI and the other monitor connected via VGA showing ATI Tools and CPU temps.
No problems at all after a few hours gaming.
(I recorded all temp data to a log and the maximum temp reached was 56'c)

Today I was working on a uni assignment so I only had Word, Firefox and Winamp open and some horizontal artefacts appeared.(nowhere near as bad as the above photos though, it was more like specs and dashes)
I opened up ATI Tool and saw the GPU temp was 35'c.
Before I could do anything it froze but the music continued playing. It stayed like this for a few seconds before going black and the music stopped.

So playing games at hi res for several hours did nothing, but having word and firefox open seemed to trigger something and start messing up.
After several restarts (kept crashing at first) I managed to get into Windows (with artefacts similar to the first photo) long enough to recover my assignment and shut down.

So confusing and such a pain in the arse!!

 

[spoonboy]

The screens u posted look ALOT like what my 2900pro does. I was using a badly-out-of-its-depth psu though on a super power hungry card and an oc'd cpu. Im hoping a well over the top 750w pc power & cooling psu im gonna order and a bios flash will do the trick. Its a faint hope but u never know. Just thought id say that. Any progress with fixing your problems? its faintly relevant to mine u see...

 

[lbakyl]

Hello guys,

many thanks for this thread. I had the exactly same problem, even in BIOS and POST. I found out after replacing the graphic card that it was faulty memory on the graphic card. This explains why the GPU was ok (e.g. temperatures, that it was doing the artifacts whenever, whether in games or when working in Word). In other words, the graphic card consists of GPU and memory and it is the memory that is faulty. If I understand it correctly, BIOS or the system (e.g. Windows) load stuff into the memory and because its faulty, it shows defective images. This is not the operation memory on the motherboard, this one is ok. Hope this helps.