Dead Graphics Card

[#1]

I think I know what I need to do here - but just asking in case someone has any better ideas than me.

I will try and keep a long story short. So bear with me 

So I have a Mac Mini and a PC. I mainly use the Mac as now I have gotten used to it and use it for general internet and Microsoft Office.

Not used the PC in a while. I decided to renew my World Of Warcraft subscription. It is the only game I really play. Switched on PC which is a monster (big mobo and case - yes I learnt the hard way that EATX = very big) which used to work fine.

So got a SSD C drive with W7 on it and a standard D drive for additional storage. Played Warcraft stored on the D drive. Would work fine for about 10 minutes, then graphics would break up. Then the PC would BSOD and the PC would beep at me.

That's not good me thinks - also was getting Windows messages outside of Warcraft saying the graphics card driver had stopped responsing and has now recovered.

Fan is running on graphics card. Checked that.

Re-installed graphics card drivers - no difference.

So I had a spare blank SSD - swapped it over and installed a brand new Windows on the alternative blank C drive. This is to rule out software - no difference.

Then I took the card out, and tried my usual hope-for-the-best-put-some-WD40-on-something approach and see if it makes any difference, which seems to have made the problem worse.

I've attached pictures of my Windows 7 login screen, but notice the pink dots? That's not right is it?

So what I know is:

It was not just affecting Warcraft - but Windows. So rules out the game being corrupt.
Happens on a new Windows install - rules out software.

The Mobo has a built in graphics card - but like all of the built in ones it is [-Censored-]. Tried to play Warcaft on that and it took me a matter of minutes to work out it was unplayable. The screen refresh was about as fast as one of those CCTV videos you see on Crimewatch.

The only thing I can think of then it is hardware, either the graphics card or the Mobo. I am hoping graphics and I am just going to have to buy another and hope it fixes the problem. So if anyone can recommend a PCI-E card for around £100. Going to be a bit of a gamble but there you go. 

Annoyingly it is a nice (read that as expensive) Motherboard and card - typical it happens to expensive items.

I'll post back more detailed specs of the PC if needed.

[#2]

Hi
This might be a bit of a long shot but have you checked that the BIOS is configured correctly. Try altering the graphics section in bios to see if it makes a difference. Double check to make sure it's set to PCI-E and not onboard vga, that's if the motherboard has onboard vga

[#3]

You said that when you took the card out and replaced it, it seemed to make matters worse. Can I suggest you remove the video card again and clean the fan entry and exit points using a paint brush and vacuum also clean the slot the card uses. If you have some Isopropyl alcohol use some of it to clean the contacts used on the card, if you’ve no alcohol then rub them with a pencil eraser.

When you replace the card ensure that it is locked into place and all power inputs are seated correctly. Do this to all of the MB connections as well as it sounds like something has or is working itself loose. Once this is done let us know the outcome please.

[#4]

Simplest solution.......Buy another video card.

If your guess is correct then the problem is solved, if however that does not solve the problem you have 14 days to return the card.

Problems like this can take hours to sort out and can often be the final thing you end up checking, could be faulty ram, cpu overheating or a number of other things.

[#5]

Thanks for the help, I'll try what has been suggested later on this evening.

[#6]

WD40 is the last thing that you want around the connectors, as are fingers due to the risk of static causing problems, so clean the connector pins off very carefully as already suggested. This could be an overheating problem. I would try operating with the PC case sides temporarily removed to see if the fans are running, it would also run cooler under these conditions.

Many graphics cards come with utility software that enables you to monitor the temperature of the vital chips. It might e worth a Google search to see if any such compatible software is available.

The fact that it is worse since you ‘oiled’ it could be due to you introducing a second  problem, so start with the careful cleaning of the connection pins.

[#7]

Do you still get the spots when you use the on-board video?

[#8]

No it is fine. Will check it again when I get home tonight but it does not seem to not occur on there.

Got the alcohol and a mini vaccum on order so will check it when it arrives on Tuesday.

[#9]

Thanks all for your help,

Tried cleaning it (with the items which had been suggested), and disassembling the GFX card to get to the GPU was not easy. Think of a more sweary version of that remark . Did it in the end, but getting to the GPU was like an episode of the Crystal Maze.

Got to it, tried some cleaning - assembling it back was even better. Try putting around 20 tiny screws back. Them bouncing off onto your carpet and finding them when they fell on the floor.

So basically there was about two hours of Alex swearing to himself in that process, managed to assemble it eventually.

My theory is the GPU had overheated, as I tried gaming for a bit. Worked fine for 10 minutes then started to go wrong, and then windows started to go wrong.

So it appears to me:

1. Did on-line gaming for about 10-15 minutes - worked fine.
2. Windows was fine.
3. Graphics started to go wrong in game.
4. Then the game started to crash, and BSOD. Okay me thinks probably just the game. Whatever.
5. Got worse, then even a boot to Windows started showing the start-up login screen got bad areas.

So it's had an overheat from #1, which has broken the GPU permanently given that the problems got worse from there. I think.

Anyway not much else I can do, picking up a new card from a friend tomorrow and see how I go.

[#10]

Such is life!

Many years ago I had an AMD twin core cpu that went the same way. I was running monitoring software on it and could watch the problem develop. It would work normally for a bit then when the cpu internal temperature got up to about 50-60 degrees one of the cores would go into thermal runaway, shooting up to over 120 degrees, at which point the computer crashed. The other core remained at a sensible temperature. If I used the pc with no covers on the box then the cpu core temperature remained below the tipping point and worked without fault. 

[#11]

Could always try baking it in the oven, it's probably got cracked solder joints under the GPU die which as they warm up lose contact and the GPU goes bananas & fails, baking the card in the oven is a rudimentary way to re-flow the damaged solder which often can restore functionality...

[#12]

Can I add that this should only be tried if the next step if it doesn't work is to put it in the recycling bin.

To get the temperature up high enough temperature to reflow the solder the absolute maximum storage temperature of components will probably be exceeded. It may be worth a try but only as a last resort.

[#13]

That's basically why I recommended it after determining it's already a goner, so, can't make it any worse than it already is...

[#14]

Oooh the baking in the oven idea sounds interesting.

Trouble with me is I am so bad even with the oven, I managed to burn a supermarket pizza a few months back which was impressive for me.

Better check my house insurance is still valid before I go down that road