Soldato
- Joined
- 29 Aug 2010
- Posts
- 8,667
- Location
- Cornwall
These days I tend to use AMeCh SGT-4, MX-4 or what came with the cooler depending on what I'm feeling at the time.
Thermal Paste Recommendations?
- Thread starter Vidar
- Start date
Associate
- Joined
- 19 Mar 2009
- Posts
- 1,647
- Location
- Leicester
At some point an elderly person will suggest AS5 even though it hasn’t won any benchmarks for years.
Ha I must be elderly! While I’m not going to recommend AS5 as such I will say that I just used it recently and given that it’s been sat in my Fridge for about a decade it still works perfectly well
I must admit, I did try to buy this straight away when I built a new PC
I was surprised it wasn't the claimed champ anymore but then I've tried whatever came with a HSF ever since, and nothing made any difference in terms of a brand beating another. I think it's a fad personally, and nothing is better than the other, not in a % that will actually benefit anyone enough to care, and lets face it, it's just thermal paste, so all the brand loyality is pointless, it's not like you can show it off
Associate
- Joined
- 19 Mar 2009
- Posts
- 1,647
- Location
- Leicester
Agree I think its likely to be such a minor difference between brands and compounds as to be irrelevant, application thickness and technique probably plays more of a factor. Certainly I cant see changing it will do much to drop my current 27oC idle temp on my 10600kI must admit, I did try to buy this straight away when I built a new PC
I think it's a fad personally, and nothing is better than the other, not in a % that will actually benefit anyone enough to care, and lets face it, it's just thermal paste, so all the brand loyality is pointless, it's not like you can show it off
Definitely! I agree that you'd be more likely to see placebo results from the application trying to replicate it exactly how you just did it, or as you say from various different techniques 
I concur. Other than complete dog **** TIM, the key differentiator is applying it correctly versus everything else.
There is no correct application method, past using as little material as possible. Everything depends on the mating surfaces between the hot and cold side as to how much of a temperature gradient can be achieved. The ideal case being two immaculate clean, completely flat pure copper surfaces in an inert environment. The best TIM is zero TIM as anything you sandwich between the two surfaces is acting as thermal insulation compared to copper.
I have found on AM4 for example, that an X isn't as effective as just spreading it out nicely with some patience with a credit card or blade. The temps definitely reflected so, as I had to repaste it, and the latter method, yielded better results.
That ideal case is ridiculous. No retail CPU and cooler are going to be remotely close to that scenario. This is why TIM exists. Show us your perfectly flat surfaces in an inert environment without TIM on your home desktop. I'll be waiting.
There is absolutely a method of applying it correctly. Using the right amount (not spreading it all over the CPU socket and making a mess for no reason) and ensuring even pressure is applied at the end. Sure, you could debate the amount of paste, how to spread it, what the ideal mounting mechanism is, blah blah blah. That doesn't mean there isn't a general consensus on a reliable and repeatable way to cleanly mount a cooler with TIM. Keep it clean, mount it evenly. Simples.
I have lapped water blocks and ISH to 1500 and 2000 grit (as flat as possible) and used no TIM at with impressive results. The issue is, after a while the copper will oxidise and reacts with the other trace metals.
The best solution I’ve found, is lap to 600-800 and use as little thermal grease as possible. The goal is to make both surfaces flat and going past 800 grit is pointless unless you want to try running without TIM.
The job of any TIM is fill in micro scratches and help prevent reactions on surfaces metals. At that point the brand of TIM is pretty much irrelevant. Spread the TIM as evenly and thinly as possible.
There is true no ideal scenario past using as little TIM as possible. All TIM has a much higher thermal resistance than something like or silver copper.
Copper (99) 400 W/m·K
Silver (99) 430 W/m·K
Best TIM 10 W/m·K
The “best” will depend on CPU or GPU used. More importantly is longevity, as some of the best ones may not keep performing as expected for too long. I used Kryonaut for a while, but noticed pump out issues quite often. Not an issue as I was changing parts quite often. The Gelid is good, but for me didn't keep performance for long. Using the Arctic MX6 which is great and cheap. Tried the TF9 (I think that’s the name) from Thermalright and was similar to the MX6. More important than the very best thermal paste is contact. Some AIOs may favour Intel, others AMD. As long the mounting pressure is good, even, and there is enough thermal paste to cover the CPU, unless you get some really nasty thermal paste, most will perform close enough. As mentioned before, longevity is the main difference.