NOOB uses Liquid Metal on an RTX 3080...
**The Great RTX 3080 Thermal Pad Experiment: A Journey of Trial and Error**
In Canada, it's a rarity to find thermal pads stocked on a regular basis. I've tried stacking various thicknesses of thermal pads on top of each other in different locations on my graphics card, thinking that might be the solution to the overheating issue. However, after repeating this process several times, with increased liquid metal application, and experimenting with different thermal pad sizes, I still couldn't get the desired results.
**The Thermal Pad Experiment Goes Awry**
I started with stacking multiple thermal pads on top of each other, thinking that would provide better contact between the cooler and the GPU. However, this approach didn't yield any significant improvements in temperatures. In fact, it seemed to make things worse, with the core frequency throttling more severely than before.
**The Liquid Metal Application Experiment**
Next, I decided to apply a decent coating of liquid metal on the GPU, thinking that might be the issue. I repeated this process multiple times, but even after extensive application, the temperatures remained high, and the fan speed was still maxed out.
**The Thermal Pad Size Experiment**
After the liquid metal experiment, I started playing around with different thermal pad sizes to see if that made a difference in contact between the cooler and the GPU. However, this also didn't yield any significant improvements.
**The Breakthrough: Removing All Thermal Pads**
It wasn't until I removed all the thermal pads and screwed the cooler back onto the card that I noticed proper contact was made between the die and the GPU cooler. This led me to conclude that there was indeed a thermal pad issue, and stacking multiple thermal pads on top of each other didn't provide any significant benefits.
**The Corrected Approach**
After this realization, I put just really thin thermal pads on everything, thinking that would be sufficient for good contact between the cooler and the GPU. While temperatures were still higher than ideal, they were much less severe than before, and the fan speed was not maxed out immediately upon gaming.
**The Final Solution: More Liquid Metal Application**
I took the GPU out again and applied even more liquid metal on the die, this time with very little bits at a time to avoid any potential issues. After an entire day of repeated tear-downs and reassemblies, I finally reached a point where the temperatures were much improved.
**The Results: A 2°C Improvement**
When I put the GPU back into my PC and started gaming with it, the core frequency was sitting at about the same point as before, and the fans weren't maxing out immediately. However, after half an hour of playing Battlefield 5, the temperatures were still a bit higher than ideal, at around 75°C Celsius. Despite this, I was able to get a 2°C improvement over the stock configuration, which I considered worth the effort.
**The Conclusion**
While some people have reported much bigger temperature improvements with liquid metal application, my experiment shows that even a small improvement of 2°C is significant considering the time and effort invested. The actual thermal interface didn't seem to be the bottleneck in this specific RTX 3080, which has a relatively small cooler on it.
**Thanks and Farewell**
I'd like to thank Neo for sending over the liquid metal and helping me out with all my noob questions during the experiment. If you enjoyed this video, please like and subscribe to the channel, and check out some other videos if you're interested! Until next time, bye!