Question Gigabyte RTX 3070 Black Screen Crashes

Dec 4, 2023
2
0
10
So I have a Gigabyte RTX 3070 OC (8gb)

For the last ~6 months I have been battling a black screen crash, all monitors lose signal, but the computer stays running. While in this state I cannot interact at all keyboard and mouse seem to do nothing, but I can still hear audio through my headphones. The only way to get it back to an operation state is to hard restart it with the button on the outside of the case.

In the journey of trying to solve this, I have:
- Safe mode DDU'd most current drivers, as well as some older ones to eliminate those as the issue
- Updated Motherboard BIOS
- Reinstalled Windows
- Reseated and dusted multiple times
- inspected and switched power connectors
- Tested power supply (everything seems stable always ~12v and ~5v)
- RMA'd through gigabyte (they "found" no issue :/)

So, I've eliminated everything I know how to eliminate, and have run out of things to try. The only thing that concerns me that I think could be the possible culprit is temps. During gaming, keep in mind this card has never been overclocked, the GPU temp is pretty normal 70-80C under heavy load, but what is concerning is the hotspot temp will hover from 102-110C (super consistently, not spiking), which from what I understand is a much higher delta than expected(from what I've read a ~15C delta is normal). Is this a possibility of my crashes? I don't know how gigabyte wouldn't have seen these during their stress tests unless I'm mistaken and they are normal temp gaps.

Additionally, it is a hard error to replicate, it doesn't really happen during benchmarks or stress tests. It happens under intensive games, Rainbow 6 with the ultra HD texture pack, and the newly released ARK: Ascended, which is very GPU heavy.

Here is a screenshot of my GPU-Z Log right as it crashed, as you can see GPU temp is ~30C lower than hotSpot, with the hotspot reaching 109C. (The dashes are presumable whilst the screen is black?)

Full PC Specs:
CPU: Ryzen 9 3900x
PSU: EVGA 750W
Ram: 32gb DDR4 @3000mhz (Corsair Vengence)
MB: Gigabyte B550 DS3H

Thanks for any help ahead of time :)
 
Last edited:

Lutfij

Titan
Moderator
Welcome to the forums, newcomer!

PSU: EVGA 750W
EVGA is the brand of the unit, while 750W is he advertised wattage of the unit. What is the model for the PSU and it's age?

Updated Motherboard BIOS
Did you clear the CMOS after verifying that the BIOS was successfully updated to the latest version?

You could also try and drop the GPU onto a donor platform and see if the exact same issue persists.
 
Dec 4, 2023
2
0
10
Welcome to the forums, newcomer!

PSU: EVGA 750W
EVGA is the brand of the unit, while 750W is he advertised wattage of the unit. What is the model for the PSU and it's age?

Updated Motherboard BIOS
Did you clear the CMOS after verifying that the BIOS was successfully updated to the latest version?

You could also try and drop the GPU onto a donor platform and see if the exact same issue persists.
PSU full spec: EVGA 750 N1 750 W ATX Power Supply its around 5 years old

I did not clear the CMOS, thats the small battery on the motherboard right?
 
Feb 3, 2024
4
0
10
Note, I was able to resolve my issues by repasting the GPU.

I was sitting at 85c, hotspot 105c, 100% fan, crashing
Now at 70-75c, hotspot 90c, 80% fan speed, no crashing

Note: in Australian summer without aircon, this is pretty good