Hi at 5:25 you apply the thermal paste in a double "zig zag" pattern. Doesn't this lead to high hot spot temps due to the trapped air? Or is this not a problem because of high cold plate convexity or is paste pattern importance just exaggerated by people?
All the paste lore I believe is heavily exaggerated by people who are honestly trying to make the best decisions for their investments they can. Working about which synthetic oil is best for your car, people loose their heads over and send used oil analysis in to Blackstone labs etc... The important part is just to change the oil regularly with a grade/spec that meets the criteria/requirement. I think the same with paste. It's a pseudo consumable item in enthusiast computing. In my experience of repasting maybe 75-100 different cards, some cards used for testing were pasted hundreds of times each, is that it does not matter that much. The internet as a whole puts so much focus on little details, then magnifies those details. Air will be squeezed out. You can have too little paste. And if the paste is thick and there is not any room to squeeze out around the die package you can have too much, but it would have to be a lot of paste and stuff that is very tough like KOLD-01. I just got that card (well both my titan Xps) into a new production plex server last night. They are currenting stacked tight together in a Lenovo p520 workstation tower. So not ideal, not a ton of air flow and the temps are: 28c for the bottom card, 30c for the top card with restricted air flow. Under load they are still pretty similar. I just installed a new AIO and I used the 9 dot pattern on the processor. Again, I think some of this stuff is blown up a bit too much, it's just not that serious for 99% of builds. I do keep the application consistent for testing because I want to try to minimize variables though, hence why I spread each paste roughly the same way on the test card's die in my other videos.
Hi at 5:25 you apply the thermal paste in a double "zig zag" pattern. Doesn't this lead to high hot spot temps due to the trapped air? Or is this not a problem because of high cold plate convexity or is paste pattern importance just exaggerated by people?
All the paste lore I believe is heavily exaggerated by people who are honestly trying to make the best decisions for their investments they can. Working about which synthetic oil is best for your car, people loose their heads over and send used oil analysis in to Blackstone labs etc... The important part is just to change the oil regularly with a grade/spec that meets the criteria/requirement. I think the same with paste. It's a pseudo consumable item in enthusiast computing.
In my experience of repasting maybe 75-100 different cards, some cards used for testing were pasted hundreds of times each, is that it does not matter that much. The internet as a whole puts so much focus on little details, then magnifies those details. Air will be squeezed out. You can have too little paste. And if the paste is thick and there is not any room to squeeze out around the die package you can have too much, but it would have to be a lot of paste and stuff that is very tough like KOLD-01.
I just got that card (well both my titan Xps) into a new production plex server last night. They are currenting stacked tight together in a Lenovo p520 workstation tower. So not ideal, not a ton of air flow and the temps are:
28c for the bottom card, 30c for the top card with restricted air flow. Under load they are still pretty similar.
I just installed a new AIO and I used the 9 dot pattern on the processor. Again, I think some of this stuff is blown up a bit too much, it's just not that serious for 99% of builds. I do keep the application consistent for testing because I want to try to minimize variables though, hence why I spread each paste roughly the same way on the test card's die in my other videos.