Question What's wrong with my RTX 3070? nvlddmkm and error 43 in device manager

[kosmaaa]

Hello. I've encountered an issue with my GPU a couple of months ago where after playing a game (CS:GO) I've encountered bugged textures but thought nothing of it because it's a source game. But then my 2nd monitor started blinking black and I've reinstalled the drivers and ever since then the drivers are unable to install or make the GPU crash. The GPU displays just fine until drivers are loaded. I'm 99% certain it's a hardware issue because I've tried another PC and Linux Ubuntu to no avail as the other machine showed the same symptoms as my machine and Ubuntu just failed to install the drivers and the screen was yellow. Is this faulty memory chips?

I didn't overclock my GPU rather I undervolted it, could that be the cause?
I sent it into my local repair shop and they've only reballed the chip to no avail.
I can't RMA the GPU as it's a second-hand GPU and I don't have the original proof of purchase of the first owner.
My Windows and BIOS are up-to-date.

Specs:
CPU: Ryzen 5 2600
GPU: MSI RTX 3070 Ventus 3X
Motherboard: MSI Aorus Elite B450
RAM: Adata XPG Gammix D10 16GBx2
PSU: be quiet! Pure Power 11 FM 850W 80+ Gold


I've attached files of how the GPU looks, if you want me to zoom onto an area of interest please let me know.
Any and all help is greatly appreciated.

https://imgur.com/ms7vnnM

View: https://imgur.com/ms7vnnM

https://imgur.com/jB5O2OH

View: https://imgur.com/jB5O2OH

https://imgur.com/zNtFLlM

View: https://imgur.com/zNtFLlM

https://imgur.com/mFWxso0

View: https://imgur.com/mFWxso0

https://imgur.com/7SH9hF0

View: https://imgur.com/7SH9hF0

https://imgur.com/Pb1iLjx

View: https://imgur.com/Pb1iLjx

https://imgur.com/ygCn2lo

View: https://imgur.com/ygCn2lo

 

[kosmaaa]

Should this be a cause for concern? There's a mark on my chip.

https://imgur.com/D2Rh5d3

View: https://imgur.com/D2Rh5d3

 

[Phaaze88]

Could you post screenshots of Gpu-Z's main tab and sensors tab?

 

[kosmaaa]

Could you post screenshots of Gpu-Z's main tab and sensors tab?

https://imgur.com/XwRP6EB

View: https://i.imgur.com/XwRP6EB.png

https://imgur.com/r0C7fBZ

View: https://i.imgur.com/r0C7fBZ.png

GPU immediately resorts to a black screen once Windows finds drivers for it. I'm pretty certain it's the VRAM chips because it shows 0mb on memory.

 

[Phaaze88]

I'm pretty certain it's the VRAM chips because it shows 0mb on memory.

Hmm, I think it's worse than that. There are numerous readings missing from the sensors tab - gpu & hot spot temperature, core voltage, rail voltage, PCIe slot voltage, fan speed, video load, perfcap reason... all missing, as well as a few specs on the main tab.

This card needs to have another look at by the shop, 'cause something else needs a reball, or traces are broken.

 

[kosmaaa]

Hmm, I think it's worse than that. There are numerous readings missing from the sensors tab - gpu & hot spot temperature, core voltage, rail voltage, PCIe slot voltage, fan speed, video load, perfcap reason... all missing, as well as a few specs on the main tab.

This card needs to have another look at by the shop, 'cause something else needs a reball, or traces are broken.

You think so? Do cards usually show those values despite there being no drivers installed? Could it be a faulty vBIOS causing this?

 

[kosmaaa]

The big question is it even worth repairing? I can't RMA the card since I'd have to contact the original shop the first owner bought it from but I don't have any proof of purchase from the first owner and also I broke the factory seal and removed some thermal pads. If it's really that bad should I just sell it as a broken card and move on?

 

[Phaaze88]

You think so? Do cards usually show those values despite there being no drivers installed? Could it be a faulty vBIOS causing this?

Yes, they absolutely do.

Never heard of a corrupt vbios without having attempted a vbios flash first, but I'm just one guy on the interwebs.

The big question is it even worth repairing?
...

If it's really that bad should I just sell it as a broken card and move on?

I'm not really in the position to say if it's worth repairing... your local repair shop may or may not be able to resolve it.
I guess you need a new gpu urgently(a project, or work)? The Ryzen 2600 doesn't have integrated graphics, so you won't be able to use the PC regardless.

Selling the card for parts is certainly an option.

 

[kosmaaa]

Yes, they absolutely do.

Never heard of a corrupt vbios without having attempted a vbios flash first, but I'm just one guy on the interwebs.



I'm not really in the position to say if it's worth repairing... your local repair shop may or may not be able to resolve it.
I guess you need a new gpu urgently(a project, or work)? The Ryzen 2600 doesn't have integrated graphics, so you won't be able to use the PC regardless.

Selling the card for parts is certainly an option.

I did flash the vBIOS before and after sending it in to the local repair shop so it could be corrupted, unsure. I have a spare GTX 970 so I haven't really been making fixing the card my priority but it's been laying collecting dust so I'm just thinking if it's worth repairing or not. I guess I'll just visit my shop when I have the money for it. Thank you for your answers.

 

[Avro Arrow]

https://imgur.com/ms7vnnM

View: https://imgur.com/ms7vnnM

Ok, so... Here we have three side-by-side groupings of what appear to be resistors, capacitors and diodes. In the top-left corner of the third group, there appears to be a stretch of pretty severe thermal damage between the top-left corner of the third group (on the right) and the little string of resistors above it. The PCB itself appears to be cracking and flaking away because of it. There also appears to be some less severe thermal damage above the first group on the left.

This kind of thermal damage is usually fatal to a card. I'm actually surprised that it still functions. I'm sorry to be the bearer of bad news but I think that your card is toast. If it's still under warranty, you can try contacting MSi because undervolting doesn't cause thermal damage, it actually guards against it. My experience with MSi "customer service" has been terrible but you never know...

 

[kosmaaa]

Ok, so... Here we have three side-by-side groupings of what appear to be resistors, capacitors and diodes. In the top-left corner of the third group, there appears to be a stretch of pretty severe thermal damage between the top-left corner of the third group (on the right) and the little string of resistors above it. The PCB itself appears to be cracking and flaking away because of it. There also appears to be some less severe thermal damage above the first group on the left.

This kind of thermal damage is usually fatal to a card. I'm actually surprised that it still functions. I'm sorry to be the bearer of bad news but I think that your card is toast. If it's still under warranty, you can try contacting MSi because undervolting doesn't cause thermal damage, it actually guards against it. My experience with MSi "customer service" has been terrible but you never know...

Are you sure? The card never had any heat issues. That could be just debris. I never saw any flacky parts of the PCB. I don't think all that matters though because the card is still toast. I don't think contacting MSI will help me with anything as I've removed thermal pads to see the memory chips so that automatically voids my warranty. Thank you for your answer.

 

[Avro Arrow]

Are you sure? The card never had any heat issues. That could be just debris. I never saw any flacky parts of the PCB. I don't think all that matters though because the card is still toast. I don't think contacting MSI will help me with anything as I've removed thermal pads to see the memory chips so that automatically voids my warranty. Thank you for your answer.

I suppose that it could be debris but it definitely looks like thermal damage to me. If you look closely at the little line of resistors above the three main groupings, there are definitely cracks in the PCB at the right end of it.

As for your warranty, if you just put the pads back on, it would be difficult for them to tell.

 

[kosmaaa]

I suppose that it could be debris but it definitely looks like thermal damage to me. If you look closely at the little line of resistors above the three main groupings, there are definitely cracks in the PCB at the right end of it.

As for your warranty, if you just put the pads back on, it would be difficult for them to tell.

I think it's too late for that the pads fell apart and I threw them away anyway. I'll just send it in for repairs when I have the time and money. Kinda sucks because I probably could've RMA'd it had I not opened it.

 

[Avro Arrow]

I think it's too late for that the pads fell apart and I threw them away anyway. I'll just send it in for repairs when I have the time and money. Kinda sucks because I probably could've RMA'd it had I not opened it.

I hope that it can be repaired. When you get cracks in the PC Board itself, it's never a good sign.