GTX 280 SLI Technical Issue

[Archangel1977]

I will try to be short here are my specifications:
I7 980x 3.33Ghz Gulftown socket 1366
Asus Rampage III Extreme 1601 BIOS
Corsair Dominator 6GB (3x2GB) 2006Mhz 8-8-8-24 1.65v
X-25M Intel SSD 80Gb
Asus ENGTX280 1GB DDR3 x2
Silverston Olympus 1200W PSU
The problem is that I get a BSOD when both Video cards are enabled on the motherboard( The asus rampage III extreme has jumper switches that turn on and off every PCI-E slot).When i close the card placed on slot 3 and stay with the card on slot 1,the PC boots into windows just fine...Things i have tried:
1.Reinstalled Windows from scratch
2.Tried 15 versions of Nvidia Drivers
3.Reflashed the motherboard BIOS
4.I have closed the card(working one) on the board and connected the monitor on 2nd card (the faulty one propably) and get no video signal but i can see the USB's posting and the lights on the PC working...everything except the video post...
The PC boots into safe mode just fine with both cards and recognizes them and i can install the drivers for them,but then when i try with both enabled to enter normally into windows it results in a BSOD,with the message "Page_fault_in_nonpaged_area",or sometimes I get the same error message with a failure about nvlddmkm.sys.I just want to know if there is a way to test if the 2nd card has gone bad...
I cannot uninstall the card to check it to another PC,cause the whole system is watercooled,so that means need to find a way to check it on this board...Anyone any ideas on the matter...?

 

[lavaheadache]

I am assuming this is a "new" build and you can't confirm individually that everything works ?

 

[BrightCandle]

The problem you face is that the issue could be with SLI (you didn't mention the bridge at all), could be with the individual GPU or it could be the motherboard slot the second card is in.

So you need to start whittling it down so that you can work out which of these it is. Turning off the PCI-E slot eliminates all 3 so you need to remove one of the cards, swap the cards to test if one is bad, try different slots etc. Only then will you be able to narrow it down to something particular. I know how that feels as a watercooler and how this is going to take (to either re plumb or reinstall air coolers), and if I had another option I would provide it but I don't.

 

[Archangel1977]

I am assuming this is a "new" build and you can't confirm individually that everything works ?

The main PC (CPU/Motherboard/RAM/SSD/Watercooling system) was made at 2010,since then only thing i have changed was the Video Cards...
2010...was using 2xGTX280
2011...was using HD 5990 if I remember the code of the card correctly
2012...was using a Gainward Phantom GTX 580 3GB DDR5
The HD 5990 and the Gainward though never were overclocked were gone bad from extreme weather conditions (lightning struck the cable network) and both of them were fried.Though I was able to RMA the HD5990 due to the fact it was on waranty..the GTX 580 lost waranty for 3 months or so...
Since at the momment i cant replace the card...I thought to install the old GTX 280's I had on the shelf...As I recalled the cards were working great of old,without any issues but they were on the shelf (boxed) for like almost 5 years.
I must also add...that i have already done check of RAM with mem86+ test and CPU with Prime95,running them for about 24h each without any mistakes being returned.The PSU was checked used the Voltometer and the outlets on the board (the Rampage III extreme has Voltimeter outputs)to check the working voltages and all seem to be fine on that regard...However the only issue is the 2nd GTX 280...which when enabled no matter what driver being used,(even tried the original CD boxed Driver)
returns a bsod with the code Page_fault_in_nonpaged_area and sometimes it also adds in the BSOD screen that "nvlddmkm.sys error"
Is there a way to really test a bad card other than indications of it being non operational on your system?

 

[Archangel1977]

The problem you face is that the issue could be with SLI (you didn't mention the bridge at all), could be with the individual GPU or it could be the motherboard slot the second card is in.

So you need to start whittling it down so that you can work out which of these it is. Turning off the PCI-E slot eliminates all 3 so you need to remove one of the cards, swap the cards to test if one is bad, try different slots etc. Only then will you be able to narrow it down to something particular. I know how that feels as a watercooler and how this is going to take (to either re plumb or reinstall air coolers), and if I had another option I would provide it but I don't.

Already tried the cards in slots 1-3 and 2-4 thankfully i had thought before to install longer tubing a bit than it was needed in case i need to move the cards...however i cant remove one of them to swap them in place,without messing with my whole loop..thing which i hate cause it involves draining the loop and refilling it..!

 

[BrightCandle]

Already tried the cards in slots 1-3 and 2-4 thankfully i had thought before to install longer tubing a bit than it was needed in case i need to move the cards...however i cant remove one of them to swap them in place,without messing with my whole loop..thing which i hate cause it involves draining the loop and refilling it..!

Oh I know, I hate having to do it as well. But the only way I can see of really testing if that card is faulty or if its an SLI issue is to have that second card in PCI-E 1 on its own with no other cards in the system.

 

[Archangel1977]

Oh I know, I hate having to do it as well. But the only way I can see of really testing if that card is faulty or if its an SLI issue is to have that second card in PCI-E 1 on its own with no other cards in the system.

When I close the PCI slots on the motherboard the only card that remained was the one on SLOT 3,should had given video signal during post but it did not,so it must have malfuctioned...I will dismantle the loop tommorow and test once more but i think already that 99% the card is bad!

 

[Archangel1977]

When I close the PCI slots on the motherboard the only card that remained was the one on SLOT 3,should had given video signal during post but it did not,so it must have malfuctioned...I will dismantle the loop tommorow and test once more but i think already that 99% the card is bad!

I did the test today one GPU is totally a fluke the other one is not working properly..I guess I have to buy new cards....

 

[ocre]

Are you saying both 280s are are bad? The worked perfectly when you pulled them out?

And the 5990 was bad and the gtx580 also bad?

When you had this electrical storm that you say damaged the GPUs you didnt bring anything else up. Just the GPUs went out? I find that incredible if that is the case. Other possibilities is the motherboard is damaged and flaky. Also and something that is very susceptible to power surges from storms is you PSU. Just because its giving out voltage doesnt mean its able to sustain. Everything in a PC is sensitive to power surges and a flaky PSU has dips and spikes.

Finally, just to let you know what your messing with here. There was a bad electrical storm that took out my cable modem. I got it replaced but there was something strange going on with my PC. The sound. I had none. Of course I had onboard sound and although i suspected motherboard damage I deleted and reinstalled drivers over and over. I tried ever darn driver i could and some would act weird and crash, some wouldnt install at all. Some looked like they installed but when i booted I had no realtek audio output device at all in my device manager.
Here is where it gets crazy. I finally came to the conclusion that the onboard audio was shot. Everything else seem to work fine. So i thought, well Nvidia has sound through their HDMI so i would just use it. I had an old Hines G Monitor that was HDMI and it was connected the whole time as a rarely used 2nd monitor. So i installed the nvidia audio driver (which i thought was already installed) and done...........so i thought. Tried to use the sound and no audio output error pops up. Now this time it shows nvidia HD AUDIO in the device manager but its throwing up some weird stuff and no sound whatsoever.
I was like, what the heck. I must have a virus. so i ran multiple virus and malware apps, i mean so really powerful tools......the most powerful. I ran these things multiple times over and over. They turn up nothing. So then i got a new HDMI cable....nope, no good.
Finally, at my wits end I dug out my old soundblaster 128 (or something). Found drivers that should work in Win7 and what do you know...........
It didnt work

Do you know what it finally ended up being. After weeks and weeks of total frustration and complete confusion?
A harddrive sector. One were windows system files for sound. I am not sure how all of that works but the issue wasnt the HW at all but with my Operating systems management of sound and its devices.

The fix was Western Digital HD tools. I ran it and it fixed my sound. I had no other issues at all. No indication of a corrupt OS or bad data. None. The sad part is that i am not a noob when it comes to issues with PCs. This was just a very crazy case. I am still using the same motherboard and HD going on 2 years later. I have checked the drive multiple times and its never had another issue. Just a surge at the right moment, in the right way caused me a very specific issue. There was no identification that it was corrupt data. It was very bizarre.

My hints were that all these different sound devices i tried couldnt have all been bad. And when i tried a sound card that i knew was good cause i pulled it out of a PC years back, a PCI card that couldnt possibly be damaged in the storm, thats what made me start looking further into this being a crazy SW thing. It could have been the motherboard, but i have worked on so many PCs you start to get a real feel for it. And even still, this threw me for a loop.

I am not saying your is your OS from corrupt data or a damaged HD. I am saying that what you are telling me sounds awful fishy. I would try to verify some of your HW in other PCs if you can. I just find that both the 280s being messed up a little to far fetched. Especially if you pulled them out working fine.