[SOLVED] 1080 Ti flashes, crashes and is no longer recognized by the system - what can I still salvage?

[wrathyee]

Greetings!

For the past 5 years or so, I've had a setup with a Gigabyte AORUS Xtreme 1080 Ti graphics card. Yesterday, after installing Dying Light and trying to play it, my entire system snapped, shut down and released a burning odor.

After rebooting the system, it would no longer recognize the GPU and the two screens attached to it. The backplate LED logo does not light up, fans don't spin and every utility from BIOS to NVIDIA's own software fails to identify the card.

My setup has three screens, one of which is Full HD and connects directly to the motherboard while the two other screens, a 165Hz 1440p screen and a 75Hz 1080p screen, are connected to the GPU. Both screens seem to light up, indicating that they're working properly, luckily.

What I suspect is that thermal overload killed the card, since running games at 1440p for the last 9 months has put a lot of stress to the card, which isn't exactly new to begin with. Any moderately demanding game would crank the temperatures up to 80 degrees C and even beyond in some cases, drawing full stress from the fans. Alas, the aforementioned happens and the card doesn't work anymore. The smell of smoke (although no smoke or fire was visible) indicates that the card did indeed burn out.

The actual question is, what's safe to salvage from the system? The components I'm hoping to preserve are my M.2 SSD units and my power supply. Now, considering what happened, calling the supply faulty would be sensible. However, the PSU in question is a 5-month-old Seasonic PRIME PX-1000, which I highly doubt being faulty taken the overall build quality, extensive wattage and certification of the unit.

As I'm writing this, my system is running with the GPU off. None of the fans in the system appear to be particularly loud at the moment. Especially the PSU is running very silent, blowing out air that feels cold against the back of my hand in a room temperature of 18 degrees C. Having watched HWMonitor for a while, it doesn't look like anything odd is going on with the voltages. Would it be safe to repurpose the PSU, should it be tested first, or scrapped altogether when putting together a new rig?

Cheers in advance

 

[punkncat]

I would consider pulling the graphics card as well as the motherboard to inspect them very closely to see if you can identify what 'burned' when it went out. Hopefully, you will see something obvious on the board for the GPU. I would check the PCI slot "pins" as it were to see if anything looks burned on the card. You probably won't be able to tell on the motherboard.

If you don't see issue with the motherboard, I still think it would be advisable to try a graphics card of old/low cost to see if you can get an output that way. I would really hate not to see any damage on the motherboard, insert some expensive new graphics card, and blow it up as well.

Since you didn't leave system spec, it would be a stab in the dark to suggest replacing anything other than the mentioned graphics card.

 

[JeffreyP55]

Greetings!

For the past 5 years or so, I've had a setup with a Gigabyte AORUS Xtreme 1080 Ti graphics card. Yesterday, after installing Dying Light and trying to play it, my entire system snapped, shut down and released a burning odor.

After rebooting the system, it would no longer recognize the GPU and the two screens attached to it. The backplate LED logo does not light up, fans don't spin and every utility from BIOS to NVIDIA's own software fails to identify the card.

My setup has three screens, one of which is Full HD and connects directly to the motherboard while the two other screens, a 165Hz 1440p screen and a 75Hz 1080p screen, are connected to the GPU. Both screens seem to light up, indicating that they're working properly, luckily.

What I suspect is that thermal overload killed the card, since running games at 1440p for the last 9 months has put a lot of stress to the card, which isn't exactly new to begin with. Any moderately demanding game would crank the temperatures up to 80 degrees C and even beyond in some cases, drawing full stress from the fans. Alas, the aforementioned happens and the card doesn't work anymore. The smell of smoke (although no smoke or fire was visible) indicates that the card did indeed burn out.

The actual question is, what's safe to salvage from the system? The components I'm hoping to preserve are my M.2 SSD units and my power supply. Now, considering what happened, calling the supply faulty would be sensible. However, the PSU in question is a 5-month-old Seasonic PRIME PX-1000, which I highly doubt being faulty taken the overall build quality, extensive wattage and certification of the unit.

As I'm writing this, my system is running with the GPU off. None of the fans in the system appear to be particularly loud at the moment. Especially the PSU is running very silent, blowing out air that feels cold against the back of my hand in a room temperature of 18 degrees C. Having watched HWMonitor for a while, it doesn't look like anything odd is going on with the voltages. Would it be safe to repurpose the PSU, should it be tested first, or scrapped altogether when putting together a new rig?

Cheers in advance

I would suggest checking the power supply first. Stick your nose down there. Do you smell any ozone?

 

[wrathyee]

I would consider pulling the graphics card as well as the motherboard to inspect them very closely to see if you can identify what 'burned' when it went out. Hopefully, you will see something obvious on the board for the GPU. I would check the PCI slot "pins" as it were to see if anything looks burned on the card. You probably won't be able to tell on the motherboard.

If you don't see issue with the motherboard, I still think it would be advisable to try a graphics card of old/low cost to see if you can get an output that way. I would really hate not to see any damage on the motherboard, insert some expensive new graphics card, and blow it up as well.

Since you didn't leave system spec, it would be a stab in the dark to suggest replacing anything other than the mentioned graphics card.

Indeed, I forgot to mention the rest of the specifications, but here goes:

Intel i7-8700k, not overclocked, be quiet! Dark Rock Pro 3
Gigabyte AORUS Z370 Ultra Gaming
Corsair Vengeance LPX 32gb (2x16) 3000MHz
M.2 drives are from Western Digital (1 Tb Black) and a Samsung 850 Evo 512gb.

If the card has indeed died, I'm intending to replace the entire motherboard and all of its contents anyway (sans the SSDs).

I'm currently asking around if anyone has an old, functioning GPU lying around to test with the PC.

 

[wrathyee]

I would suggest checking the power supply first. Stick your nose down there. Do you smell any ozone?

No, it doesn't put out any kind of a distict smell. Only the slightly cooler air.

 

[JeffreyP55]

No, it doesn't put out any kind of a distict smell. Only the slightly cooler air.

I still suspect the supply. Do you have extra parts like a PS? Usually discreate parts don't blow up in dramatic fashion. Power supplies holding all the current does.

 

[wrathyee]

I still suspect the supply. Do you have extra parts like a PS? Usually discreate parts don't blow up in dramatic fashion. Power supplies holding all the current does.

The old power supply I used (be quiet! Power Zone 750w) should still be lying around somewhere.

I'm only suspecting the GPU is at fault because everything else within the system appears to work as intended right now. Thermals have been an issue to the system ever since I changed to a 1440p display, for which the GPU really isn't meant for as far as I'm aware.

 

[punkncat]

Personally would question a possibly near 7 yo card that was already running hot before thinking it was a 5 month old power supply. I mean, things happen, but it just seems unlikely IMO that the power supply failed in that manner only on one "plug" (as it were).

 

[Maxsupertweaker]

I agree with (punkncat) on this one.

 

[wrathyee]

Personally would question a possibly near 7 yo card that was already running hot before thinking it was a 5 month old power supply. I mean, things happen, but it just seems unlikely IMO that the power supply failed in that manner only on one "plug" (as it were).

Hello,

after two weeks, I can confirm your suspicion on the topic was right (let this reply exist and hopefully be found in the future should this prove to be a common problem concerning the part in question).

Together with a friend of mine, we dismantled my old build before putting together a new one. Suspecting the GPU is the part that is damaged, and knowing the warranty has long since expired, we decided to dismantle the cooling solution around the Gigabyte Aorus GTX 1080 Ti Xtreme Gaming 11 GB in question.

Below, I've attached an image of the back side of the graphics card showing the damage in question:

https://imgur.com/a/6UXXpAg

TechPowerup reviewed the card back in the day, and a good indicating image of the card's backside is available on page 3: https://www.techpowerup.com/review/gigabyte-gtx-1080-ti-xtreme-gaming/3.html

The burnt-out parts, which have seemingly almost exploded, are coded R1631 and R1651, a short distance right and down from the CE sign on the board. Whatever their function is, I do not know, but they're the root of the damage in the question.

My Seasonic PSU, on the other hand, appears to be working without a problem, currently nurturing Ryzen 7 7800x3d and MSI's 4070 Ti Gaming X Trio edition.

Thanks to everyone participating in the thread and showing care over my concern!

 

[punkncat]

Yes, I figured you would find something like this based on your description of the event. Glad you got it sorted. Heck of a loss a 1080ti is even still, but it seems you did pretty well for the upgrade. Enjoy!

 

[vov4ik_il]

Below, I've attached an image of the back side of the graphics card showing the damage in question:

https://imgur.com/a/6UXXpAg

TechPowerup reviewed the card back in the day, and a good indicating image of the card's backside is available on page 3: https://www.techpowerup.com/review/gigabyte-gtx-1080-ti-xtreme-gaming/3.html

The burnt-out parts, which have seemingly almost exploded, are coded R1631 and R1651, a short distance right and down from the CE sign on the board. Whatever their function is, I do not know, but they're the root of the damage in the question.

The VRM is cooked (on the other side of the PCB). Those resistors are outcome, not a cause. Most likely the Pascal chip is shot too. 50$ parts card now.

 

[Avro Arrow]

You can always take the cooler off. It might be usable on another card (if you ever actually need to).