Does this sound like a bad video card?

sornywrx

Member
Jun 16, 2010
175
0
76
I am helping a friend with a PC problem. I am kind of stumped because I usually don't get to work on gaming machines but just business class PC's at work. At first he said that the computer powered up but no display (no BIOS screen or anything). When I powered it up it did that once or two but then suddenly kicked on and started loading Windows 7 but then blue screened. I rebooted and it almost loaded Windows then just powered off. I was going to check memory but booting a CD with MemTest and it ran about 20 seconds before it shut off so I'm ruling out hard drive/Windows errors.

I thought maybe the GPU is overheating but the fan seemed to be running ok, not clogged with dust, and it was cutting off after the machine was only on for 15-20 seconds max so I didn't think it was long enough for it to overheat. I checked and tightened the heatsink screws too, didn't help. The card is a Radeon HD 6770 and I'll be honest I don't even know if it's a top of the line card or anything fancy or not. I know it has the 6 pin power socket on it and I made sure it was plugged in. The power supply is a XION brand 700 watt.

I pulled the Radeon out of the PC and used the onboard graphics and it booted and ran fine for hours and hours with NO problems at all. I shut down, replaced the video card and booted and it shut off or blue screened immediately. I was sure it was the video card and am about to tell him to buy another one but I hate to be wrong and him order one when it may be something else. Since I'm not used to even dealing with power supplies with SLI/PCIE 6 pin power connectors I am wondered if it's possible that it's the power supply. The computer has ran fine for hours without the card plugged in so logically I think it's the card but in the back of my head I'm wondering if it's running fine w/o the card just because the additional load of the card isn't there.

Sorry for the long post but I just needed some opinions. I have spare PSU's and video cards laying around but I don't use my computer for gaming so I my PSU's are plain old 300w models and my video cards are old AGP cards :p. Nothing that needs a direct power source to the PSU.
 

zaydq

Senior member
Jul 8, 2012
782
0
0
Sounds like the PSU is failing to provide the GPU the juice it needs. A faulty GPU shouldn't theoretically make a PC power off.
 

sornywrx

Member
Jun 16, 2010
175
0
76
Sounds like the PSU is failing to provide the GPU the juice it needs. A faulty GPU shouldn't theoretically make a PC power off.

Well that makes sense I just wasn't sure if since it was doing it right after boot (or sometimes before even booting) if that would be it. Not disagreeing with you, just explaining my ignorant thought process. If the computer shut off during a game then I would think the power supply couldn't keep up but figured the power draw (to video at least) during boot would be minimal. And I thought that the Radeon, when not playing a game, wouldn't draw much more than the onboard graphics. But you're right, a GPU normally wouldn't make a PC just shut off immediately, well from what I've seen.
 

zaydq

Senior member
Jul 8, 2012
782
0
0
The system should boot with a malfunctioning gpu and simply make beeps to communicate that no display device is being found. I could be wrong, but i doubt the gpu is at play here. Try running the system without the 6 pin plugged in and see what happens. If the system carries on to boot then i'd be pretty sure its the PSU.
 

Greenlepricon

Senior member
Aug 1, 2012
468
0
0
I had a friend whose computer ran just fine up until he bought a serious game to play, then it would shut off pretty similarly. I thought it was just a bad driver since it ran fine before, but nope. It was the psu overheating. Pretty similar symptoms. If you could plug that gpu into another computer that would be awesome. If you can plug in another gpu to that specific computer that would be awesome too. The easiest way to tell psu is probably to stress it out at this point (just be careful!). 700w won't have any problem running a 6770 if it's working, so I would try adding anything extra and seeing if it causes the computer to crash again.
 

sornywrx

Member
Jun 16, 2010
175
0
76
The system should boot with a malfunctioning gpu and simply make beeps to communicate that no display device is being found. I could be wrong, but i doubt the gpu is at play here. Try running the system without the 6 pin plugged in and see what happens. If the system carries on to boot then i'd be pretty sure its the PSU.

Well any time I've ever seen a graphics card (either on board or card) go bad it would just do the "no video" beeps at start up or just be a black screen and no boot (in laptops usually). I tried booting without the 6 pin plugged into the card, on accident actually, and I got power but not POST or beeps or anything.


I had a friend whose computer ran just fine up until he bought a serious game to play, then it would shut off pretty similarly. I thought it was just a bad driver since it ran fine before, but nope. It was the psu overheating. Pretty similar symptoms. If you could plug that gpu into another computer that would be awesome. If you can plug in another gpu to that specific computer that would be awesome too. The easiest way to tell psu is probably to stress it out at this point (just be careful!). 700w won't have any problem running a 6770 if it's working, so I would try adding anything extra and seeing if it causes the computer to crash again.


I wish I had another card to try but none of my cards require external power and are powered by the bus because I don't have any decent video cards laying around. I also do not have any high power PSU's and all of mine are around 300-350W 24pin ATX PSU's without the 6 pin video connectors.

I DO have a few old SCSI hard drives, think plugging up 2-3 of those monsters would draw as much as this video card?
 

sornywrx

Member
Jun 16, 2010
175
0
76
Well beforehand the computer wouldn't usually even boot Windows let alone be usable but today I ran the machine for over and hour including running Prime95 and playing a few games like Black Ops at high settings. GPU averaged about 150 degrees but the system never shut off or rebooted.

I installed Speed Fan to measure the temps and voltages and it showed some odd numbers (not sure if it's accurate though):

+3.3V: 3.39V
+5V: 5.16V
+12V: 12.48V
-12V: -5.94V
-5V: -2.04V
+5V: 3.63V
Vbat: 3.12V
GPU Vddc: 0.95V

Those few numbers that are WAYYYY off make me thing it's not accurate.. I mean would it even boot if -12V was -5.94V??

Still leaning towards PSU but I hate to say that and him spend $100 on one and then it still mess up.
 

SpeedTester

Senior member
Mar 18, 2001
995
1
76
Wow some of those volts are way outta whack. Pick up or borrow a multimeter and get some accurate reading s before spending any money.
 

Greenlepricon

Senior member
Aug 1, 2012
468
0
0
Wow some of those volts are way outta whack. Pick up or borrow a multimeter and get some accurate reading s before spending any money.

If you can then measuring the voltage accurately is the easiest way to tell, or I think you can take it into a store and they might do it for you. I think you can get away with a $60 or so psu if you need to replace it. Good luck to you and your friend on this. Luckily it seems as if everything else is working fine.
 

sornywrx

Member
Jun 16, 2010
175
0
76
Thanks guys I have a multi meter here somewhere I dunno why I didn't think to try that. I understand how to test the +12v and +5v but how do I do the -12 and -5? If its not an easy answer I'll just Google it instead of wasting your time.. posting from my phone or I wouldve already done it.
 

sornywrx

Member
Jun 16, 2010
175
0
76
Well I'm pretty ignorant about testing them I guess... just googled a little and I thought you tested just the 12 and 5v on a 4 pin molex connector. Duh, wasn't even thinking about the 24 pin.
 

sornywrx

Member
Jun 16, 2010
175
0
76
Well I finally found my multimeter and tested the PSU. Everything was pretty good (3.2V for a 3.3, 5.2V for a 5.0, and 12.28V for 12V, etc) but then when I tested the -12V I kept getting -11.08V. That seems pretty far out of range to me. I also did a little more research on this model PSU and it seems that it's not a total junk unit (like a $30 chepao) but it's definitely nothing to brag about. Lots of people saying it's junky or not worth putting in a gaming sysem and lots of forum posts of people replacing them.
 

sornywrx

Member
Jun 16, 2010
175
0
76
The best test in this situation is to try the video card in another PC that's otherwise known to be stable, that's pretty much the sanity test I use for any hardware errors I encounter.

That's normally what I do when troubleshooting but the problem is I do not have anything but plain 300W ATX power supplies, which don't even have the 6 pin PCIE plugs, and I do not have a video card that's near what this one is. I think the fanciest video card I have is an old 64MB card lol. I do not play PC games anymore (quick that back in the 3DFX days 10 years+ ago).
 

sornywrx

Member
Jun 16, 2010
175
0
76
Well I told him to get a new power supply and he picked up a Corsair GS700 700w PSU. Machine booted up and ran fine for hours and hours. I loaded prime95 and let it run and the system got warm but didn't overheat and ran fine during the stress test for about 5 hours. Shut it down and started up Skyrim (I'm not sure what the settings were originally, maybe 1920x1280.. anyways, after no more than 5 min of play I could hear the 6770's fan get louder and louder and I checked Speed Fan and the GPU was at 180 degrees F (~80c). Kept bouncing between 178-180 and before long it would just shut the entire system down instantly. I didn't give it time to cool off and fired it back up. As soon as it'd be loading the Windows 7 "Welcome" screen and asking for a password it'd shut off again. So I tried again, same deal. This time I took the card out, checked the heatsink screws (tight), and took the heatsink off. Shouldn't there be thermal paste on this thing? There wasn't. The AC was running so I laid the card on the AC vent in the floor for about 2 min where it got pretty cold and stuck it back in and it booted up fine.

Went back into Skyrim and it hit 180 soon and shut off again.

So I'm not sure what to do now. Will thermal paste help much? I set the CPU fan at 100% and it still over heated and shut off. Left the side panel off the case and it still over heated. In Black Ops it seemed to be ok. Is 180 hot enough to make it cut off or is it a coincidence that it's cutting off around then? Made sure that there wasn't any cables blocking anything but theres only 2 fans besides the GPU and since it's water cooled one is on the radiator (I guess it's called that, I don't have experience with water cooling but understand the principal), and the other is a cheapo NZXT side panel fan that doesn't blow much.
 

pwoz

Member
Aug 27, 2012
43
0
0
There should be thermal paste and adding some would be a vast improvement on the cooling.
 

AnonymouseUser

Diamond Member
May 14, 2003
9,943
107
106
Yes, it should have thermal paste, but I find it hard to believe it's overheating and shutting down at 80C. I have a 9600GT that would overheat to 130C before crashing the system, and blowing out the clogged heatsink dropped temps to max 70C. It's possible that the temps aren't reading correctly, though.

How are the NorthBridge temps?
 

sornywrx

Member
Jun 16, 2010
175
0
76
Well I added some cheap $1.99 thermal paste I got at Radio Shack. I was afraid it would be for small electronics and not a gaming system but the package said it was for "high performance CPU's and water cooled systems". I put some on and the card went from idling at like 130 to 118F. Fired up Skyrim again and this time it lasted for a lot more than 2 minutes (like 2 hours straight before I quit) and never got above 165F. I gave it back yesterday and haven't heard anything yet, fingers crossed.