Exhausting all solutions before I RMA the GTX 680, Please Help! BSOD, Freeze, Etc...

[kohakyun]

Hello,

I built my system a couple of days ago,
The Specs are:

Windows 7 x64 Ultimate
i7-3770K (H60 Cooler)
ASUS Sabertooth Z77 Mobo
Corsair Vengeance 32GB (4x8GB 1600)
OCZ Vector 256GB
EVGA SuperNova 750W PSU
Dual Monitor Setup (S2330MX x2, One Analog Connection, One Digital)

And the graphics card I am in need of your help is
Gigabyte GV-N680OC-2GD NVIDIA GeForce GTX 680 2048MB GDDR5 PCIe 3.0 x16 Video Card
(Windforce, 3 fans)

Problems:
* Screen Blacks Out, then recovers, windows tells me "display driver stopped responding and has successfully recovered."

Then it evolved into:

* BSOD - The computer has rebooted from a bugcheck. The bugcheck was: 0x000000d1 (0xfffff8880b6e3ff8, 0x0000000000000002, 0x0000000000000001, 0xfffff8800497e103).

And was "Unable to produce a minidump file from the full dump file."

Then it evolved into:
http://i281.photobucket.com/albums/kk220/koguryeo/photo.jpg
http://i281.photobucket.com/albums/kk220/koguryeo/photo1.jpg

White boxes all across the screen (Analog monitor is the pink).
And the computer freezes on that screen, must manual restart.

This first happened while running the Windows Experience Index during the initial GPU test.
Then it happens when playing the game "League of Legends." It happens when 3+ players enter my screen and spam effects, no problem when just 1-3 players on my screen.

What I have tried already:
* Reinstalling Driver (Clean reinstall, removing all traces of old driver)
* MSI Afterburner to control fans (Still persists with all fan settings)
* It is not a temperature issue, GPU temp remains 30C-40C.
* MSU Afterburner underclock the factory OC'd GTX 680 by 100, still no good.
* Downgrade Nvidia Driver to older versions, tried the last 3 drivers with all clean reinstalls.
* No OC on CPU, Ram, anything, all stock, problem persists.
* Disable Powermizer (Still persists)

Is there anything more I can do?
Or do I have to RMA without question?
Anyone else have the same problem?

Thank you very much!

Locked per OP - solution found. Thanks all. -Admin DrPizza

 

[VirtualLarry]

Did you do a thorough RAM test? Download the Memtest86+ bootable ISO, burn it to a CD, and boot off of it. Let it run 12hours or more. If it throws errors, double-check your RAM settings in the BIOS, and if they are correct, then RMA the RAM first.

 

[kohakyun]

I was planning on doing that tonight while I sleep, but I just started it now and using my laptop for the time being. Thanks! I'll let you know if anything comes up.

If any one else has any other insights please contribute!!

 

[philipma1957]

do you have a spare low cost card? .

your psu may not supply enough juice.

other issues mobo could be bad.


is this your mobo?


http://www.newegg.com/Product/Produc...82E16813131821


if it is try the other video card slot. your mobo slot could have the problem. so

1) psu gives wrong power

2) bad ram

3) bad mobo slot

4) bad video card

 

[kohakyun]

Just installed it on the 2nd slot, but it's running at 8x and not 16x.
It should be running at 16x on that slot as well if i'm not mistaken.
Shrug, i'll just test to see if I get the same problems, will update.

Thanks!!

I'll restart the memtest when I go to sleep and while i'm at work so it should be running for 12+ hours.
Had no errors for the hour or two.

 

[kohakyun]

Yeah, I got error #1 again changing slots.
It worked for a few games of League of Legends, then it happened again. Thought it might have been the slot for a second there.
Not too sure how I would see if the PSU would give me wrong power.
Hmmm just have to check the ram, gonna let it run til tmrw

 

[Termie]

Those are clearly video errors. What's the point of running memtest for 12 hours? A better test would be to see if you get the errors with a single stick of memory. Unfortunately I assume you will.

Run the system off the on board video and see if you get similar video corruption.

By the way, I assume you have both pcie power connectors attached to the 680.

 

[kohakyun]

Might as well, test for the sake of testing since I have until Sunday before I can do any returning to vendor (work).

I've tried with onboard, works perfectly fine.
I'm most likely just going to return and exchange it come Sunday, probably for something else and not this model. I know it was probably just the one, but my feelings were hurt.
Just going to try whatever I can until then.

Thanks for the input Termie!! Oh, and yes both connected.

 

[3DVagabond]

If this is new enough to return things my advice is to dump that PSU. Every Supernova I've seen tested has had voltage regulation issues. Get an (in no particular order) Antec, Corsair, Enermax, or Seasonic.

 

[DoeBoy]

I would point my finger at the ram on the motherboard as well before I would point a finger at the video card or PSU. Make sure your seating the card in all the way. Your PSU has plenty of power for that 680.

 

[ShintaiDK]

Tried downclocking the GFX card? Its not unhead of that factory OCed cards fail.

 

[nOOky]

From what you've said and done it really sounds like you got a bad video card. As mentioned before, make sure the card is fully seated and all power connectors attached (which it sounds like you've confirmed). It happens, probably a cooling issue with the card? I'm guessing.

 

[philipma1957]

WELL the problem is?


1)psu don't know

2)mobo don't know

3) ram don't know

4) GPU don't know


So far You have not pinpointed which of the above.

I found it interesting the gpu worked for few games in the second slot.

Could be a micro short in the mobo.


You need to tease out the real problem.

You need a second psu that is known to work.


repeat all the tests with the second psu.


if it does not work your problem is not the psu.

what is a good way to get a second psu for you to test with.

 

[kohakyun]

I'm actually going to Micro Center today, I just checked their store hours and they have special holiday hours going on.

I'm going to replace both the PSU and the GPU in one go.
If the problem persists, you will be right about the motherboard, which I would even more distraught about. (I used my last stick of paste, took forever tinting perfectly, worse I would have to make another trip to MC)
I already mailed the rebates, am I supposed to call and cancel them?

 

[Termie]

I'm actually going to Micro Center today, I just checked their store hours and they have special holiday hours going on.

I'm going to replace both the PSU and the GPU in one go.
If the problem persists, you will be right about the motherboard, which I would even more distraught about. (I used my last stick of paste, took forever tinting perfectly, worse I would have to make another trip to MC)
I already mailed the rebates, am I supposed to call and cancel them?

You better call MicroCenter to see if they'll accept a return on a product where you've cut out the UPC symbol. My guess is that they will not.

In that case, you'll need to RMA the products. I'd start with the GPU. The PSU wasn't a great pick, but it's unlikely that it's causing this problem. You'll still have the chance to RMA it later if replacing the video card doesn't fix the issue.

 

[mango123]

Those are clearly video errors. What's the point of running memtest for 12 hours? A better test would be to see if you get the errors with a single stick of memory. Unfortunately I assume you will.

Run the system off the on board video and see if you get similar video corruption.

By the way, I assume you have both pcie power connectors attached to the 680.

Yeah, what is the point of running a memtest check?

Like others have said, this is clearly a video card issue.

 

[AdamK47]

The nVidia "display driver stopped responding and has successfully recovered." message doesn't necessarily mean that the graphics card is at fault. Back when I was using 3x GTX 580s I would get this message if my memory was overclocked too far or the timings were too tight. Really, any instability with the system can cause the nVidia drivers to do this.

 

[nOOky]

Looking at the images he posted, they clearly look like video card issues instead of system memory issues imho.
System memory issues generally cause system instability, not corrupt graphics output in Windows.

 

[philipma1957]

the problem I have with his situation is he can't determine what is good.

I have seen that symptom with my bit coin rigs (14 hd7970's and 2 hd7950's).


It never was the video card.

it was some bad riser ribbons one time it was a bad psu another time, it was a bad mobo slot the third time.

Of course most of the time it was undervolting the cards. 1.12 volts and 925mhz was stock. I would go for .95 volts and 925 mhz. this gives me the best watt to hash rate.

some gear would allow this setting some gear wants .96 volts or .993 volts and 1 setup won't give me 925ghz unless I do 1.012 volts.

He got the gear to run 3 games on the second slot then it did not want to work any more. My guess is a lack of power /volts to his clock of the card.

I have not built pc's for as long as I have done mac upgrades. But if you diy a know good psu helps. I have a fanless seasonic 400 watter as my bench psu. so if a build fails I try it with the seasonic. just 2 weeks ago I was doing a build with antec platinum ea 450 watter .. the gear turned off 10 seconds then powered off. My first move was test all connections for tightness no go. I then swapped the psu with the seasonic and all worked.

I rma'd the antec time lost to solve problem under 20 minutes.


If he rmas the psu and the gpu he can setup again and it may not work. guess what that puts him right were he started. is the gpu bad? is the psu bad? new replacements can be bad. or is it the mobo slot. I do not think it is the ram. I have done around 30 machines the last 14 months. So having extra parts is easy for me. I just hate to see him do the rma and it still does not work.

 

[Will Robinson]

I have seen that symptom with my bit coin rigs (14 hd7970's and 2 hd7950's).

14 x HD7970 ?

That much mining could destabilize the Earth's rotation or at the very least affect Global Warming

Suffice it to say...I think this is a First World problem you are dealing with here.^_^

 

[philipma1957]

14 x HD7970 ?

That much mining could destabilize the Earth's rotation or at the very least affect Global Warming

Suffice it to say...I think this is a First World problem you are dealing with here.^_^

it heats one floor of my house.

 

[kohakyun]

Hey guys, so I ended up replacing both the gpu and psu.
From GTX 680 -> MSI HD 7970
EVGA 750 PSU -> Corsair TX750M

I have zero problems now, so it was definitely one of the two.
Most likely the GPU, I just changed the PSU because 3DVagabond suggested it.

Thanks for all your inputs, greatly appreciated!
Wish I could have pinpointed the problem with 100% accuracy but just didn't have much to work with, ie: spare PSU.

 

[Termie]

Hey guys, so I ended up replacing both the gpu and psu.
From GTX 680 -> MSI HD 7970
EVGA 750 PSU -> Corsair TX750M

I have zero problems now, so it was definitely one of the two.
Most likely the GPU, I just changed the PSU because 3DVagabond suggested it.

Thanks for all your inputs, greatly appreciated!
Wish I could have pinpointed the problem with 100% accuracy but just didn't have much to work with, ie: spare PSU.

Great to hear you were able to exchange the parts and that that system is up and running. And by my estimation, you saved some money in the process.

 

[peonyu]

Hey guys, so I ended up replacing both the gpu and psu.
From GTX 680 -> MSI HD 7970
EVGA 750 PSU -> Corsair TX750M

I have zero problems now, so it was definitely one of the two.
Most likely the GPU, I just changed the PSU because 3DVagabond suggested it.

Thanks for all your inputs, greatly appreciated!
Wish I could have pinpointed the problem with 100% accuracy but just didn't have much to work with, ie: spare PSU.

Nice trade, you saved some money and got a faster card.