- Oct 15, 2021
- 3
- 0
- 10
My specs are as follows:
AMD Ryzen 5950x
Zotac 3080 ti Trinity
Gigabyte Auros Master x570S
Corsair 2x32GB 3200MHz RAM
Corsair 1000w PSU
Windows 10 for Workstation
For the above system, I built a Corsair based custom-cooling system and used the Corsair XG7 Reference water-block for the GPU (that comes with pre-applied paste & pads). My first day of running tests (Cinebench, Unigine Heaven, 3D Mark TimeSpy) I noticed that while my GPU temps hovered between 60-65C (never exceeded this), I was getting a hotspot temp of 105C, as mentioned here. I also saw clock fluctuations ranging from 300 to 1900 MHz. I decided to open the GPU again and replaced them with fresh thermal pads and thermal paste, from Deep Cool and Thermal Grizzly. Since then I've only been facing booting issues, where my motherboard either doesn't POST and shows 'no signal to monitor' after 10 seconds of loading, or reaches a Windows blue screen. After multiple attempts at trouble shooting (including: disassembling & reassembling the PC, replacing the PCIE cables that connect to the PSU, changing RAM configuration, changing monitor etc.) I was able to pinpoint the issue directly to 3080 ti when a lower-end GPU worked perfectly fine with the rest of the system, enabling it to boot up completely.
I took the 3080 ti to the Zotac service center (India) and they informed me that it will take at least 2 weeks to process the RMA, and that RMA might not even go through because I've installed a water-block on it myself. I have a lot of professional projects on hold because my system is dysfunctional and 'at least 2 weeks' is a very long time for to wait and be possibly rejected for the RMA. I also definitely can't afford to buy another card for now. I want to try solving this issue one last time before I try the RMA since I've already opened the GPU up. So I've been doing research on different forums, and multiple threads/articles suggest that some 3090s, 3080s and 3080 ti's come with insufficient padding that don't cover all possible heating parts of the GPU. I can imagine that to be a possible reason why my GPU shuts off within 10 seconds of functioning, and why I was getting those hotspot temps + thermal throttling before.
I want to thoroughly understand the GPU board w.r.t. the following:
AMD Ryzen 5950x
Zotac 3080 ti Trinity
Gigabyte Auros Master x570S
Corsair 2x32GB 3200MHz RAM
Corsair 1000w PSU
Windows 10 for Workstation
For the above system, I built a Corsair based custom-cooling system and used the Corsair XG7 Reference water-block for the GPU (that comes with pre-applied paste & pads). My first day of running tests (Cinebench, Unigine Heaven, 3D Mark TimeSpy) I noticed that while my GPU temps hovered between 60-65C (never exceeded this), I was getting a hotspot temp of 105C, as mentioned here. I also saw clock fluctuations ranging from 300 to 1900 MHz. I decided to open the GPU again and replaced them with fresh thermal pads and thermal paste, from Deep Cool and Thermal Grizzly. Since then I've only been facing booting issues, where my motherboard either doesn't POST and shows 'no signal to monitor' after 10 seconds of loading, or reaches a Windows blue screen. After multiple attempts at trouble shooting (including: disassembling & reassembling the PC, replacing the PCIE cables that connect to the PSU, changing RAM configuration, changing monitor etc.) I was able to pinpoint the issue directly to 3080 ti when a lower-end GPU worked perfectly fine with the rest of the system, enabling it to boot up completely.
I took the 3080 ti to the Zotac service center (India) and they informed me that it will take at least 2 weeks to process the RMA, and that RMA might not even go through because I've installed a water-block on it myself. I have a lot of professional projects on hold because my system is dysfunctional and 'at least 2 weeks' is a very long time for to wait and be possibly rejected for the RMA. I also definitely can't afford to buy another card for now. I want to try solving this issue one last time before I try the RMA since I've already opened the GPU up. So I've been doing research on different forums, and multiple threads/articles suggest that some 3090s, 3080s and 3080 ti's come with insufficient padding that don't cover all possible heating parts of the GPU. I can imagine that to be a possible reason why my GPU shuts off within 10 seconds of functioning, and why I was getting those hotspot temps + thermal throttling before.
I want to thoroughly understand the GPU board w.r.t. the following:
- What are the various chipsets on the GPU board?
- Which parts of the GPU board heat up and which don't?
- Where all do I need thermal padding?
Facebook
Twitter
Instagram