-
Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!
For Air cooling what TIM is best?
- Thread starter wingman99
- Start date
- Joined
- Sep 13, 2015
Plenty of TIM roundups all over the web... including Loch's own, awesome internal testing. 
The short of it........... there is barely a difference. 1-3C from highest end metal to *most* cheap pastes. Take the metal pastes out and many more are within that 1-3C. Get something middle ground/bang for the buck and focus more on application and mount.
http://overclocking.guide/thermal-p...d-with-air-cooling-and-liquid-nitrogen-ln2/6/
The short of it........... there is barely a difference. 1-3C from highest end metal to *most* cheap pastes. Take the metal pastes out and many more are within that 1-3C. Get something middle ground/bang for the buck and focus more on application and mount.
http://overclocking.guide/thermal-p...d-with-air-cooling-and-liquid-nitrogen-ln2/6/
Last edited:
I can't agree with some results on this list, not to mention that is clear that der8auer is promoting products which are branded his nick. For instance on air I had better results using gelid extreme than kryonaut, on cold kryonaut was a bit better. Alphacool TIM is about the same as Gelid Extreme but on this list is much lower ( on a smaller package of this TIM added to AIO is even info powered by Gelid Extreme or something ). Bequiet TIM was much worse when I was testing it. Literally 3-4°C worse than MX4, dryish and harder to apply. Noctua TIM is almost as good as Gelid Extreme on air ... well, many these TIMs give results close to each other like EarthDog said so I just wouldn't spend too much.
Gelid Extreme or Supreme ( a bit worse but easier to spread and cheaper ), MX4 or Noctua is what I'm using because of reasonable price or because I get it for free ( in coolers or other things ).
Extreme benching is where TIM counts much more. On air proper mount and apply give better results than differences between some TIMs.
Many forum users will stick to pea/rice grain method of applying but it depends on used TIM. Generally if you make it this way then it won't cover whole IHS and it will spread well only in the center. On a large IHS I'm just spreading it some more. Best if you try it yourself 2-3 times and check how it's spreading.
Gelid Extreme or Supreme ( a bit worse but easier to spread and cheaper ), MX4 or Noctua is what I'm using because of reasonable price or because I get it for free ( in coolers or other things ).
Extreme benching is where TIM counts much more. On air proper mount and apply give better results than differences between some TIMs.
Many forum users will stick to pea/rice grain method of applying but it depends on used TIM. Generally if you make it this way then it won't cover whole IHS and it will spread well only in the center. On a large IHS I'm just spreading it some more. Best if you try it yourself 2-3 times and check how it's spreading.
Last edited:
- Thread Starter
- #6
Plenty of TIM roundups all over the web... including Loch's own, awesome internal testing.
The short of it........... there is barely a difference. 1-3C from highest end metal to *most* cheap pastes. Take the metal pastes out and many more are within that 1-3C. Get something middle ground/bang for the buck and focus more on application and mount.
http://overclocking.guide/thermal-p...d-with-air-cooling-and-liquid-nitrogen-ln2/6/
So your saying going from ARCTIC MX-2 to Gelid gc extreme will only net 1-2c lower in temperate? If that is the case I will still use what I have left of the Arctic MX-2.
Im not saying that, my link/roundup does (did you read it? Others?)...any of them on the web. Youll see the difference isnt much.
I have a big tube of mx-2 i still use (on review pc)..i also have some kryonaut (thanks loch!) I use on my daily driver.
- Joined
- Mar 7, 2008
IMO unless you need to gain every last fraction of a degree in extreme conditions, they're much of a muchness. If MX-2 is good enough, stick with it. I personally use the generous blob method, more generous for bigger CPUs. Enough for some to spill over the sides after squishing so most of the IHS will be covered, maybe missing corners at worst.
- Thread Starter
- #10
Im not saying that, my link/roundup does (did you read it? Others?)...any of them on the web. Youll see the difference isnt much.
I have a big tube of mx-2 i still use (on review pc)..i also have some kryonaut (thanks loch!) I use on my daily driver.
I read the link you posted before in another forum and I don't know how Kelvin drop equates to what I will get for reduced core temperature. The poster in the other forum said I would see 10c drop in temperature with Gelid gc extreme. I did not believe believe what he was saying, it sounded to good to be true, so I started a thread here.
The review temperature scale did not answer my questions, I started the thread so I would not waste time and money trying different TIM.
- Joined
- Jan 1, 2008
- Location
- Colorado Springs, CO
Kelvin = deg. C + approximately 273. C and K are essentially the same, with the difference being the 0 point.
- Thread Starter
- #12
So according to the chart LINK: http://overclocking.guide/thermal-p...d-with-air-cooling-and-liquid-nitrogen-ln2/6/ Arctic MX-2 and Gelid gc extreme is ~1c difference?
- Joined
- Jan 1, 2008
- Location
- Colorado Springs, CO
As far as I can see. The chart is in Delta temps (temperature difference) in this case delta-K = delta-C as they merely subtract the lower number from the higher number.
Example:
50C-43C=7C
Now converting those numbers to K using C + 273 = K:
323K-316K=7C
There is no scaling factor like in C->F there Delta K should always equal Delta C.
Example:
50C-43C=7C
Now converting those numbers to K using C + 273 = K:
323K-316K=7C
There is no scaling factor like in C->F there Delta K should always equal Delta C.
- Joined
- Jul 26, 2004
I still use AS5 its one of the few kinds that I can find local and has worked for me for 10+ years .
- Joined
- Sep 13, 2015
So I was curious and ran 1 test set of data on the MX-2 paste for comparison. It seems to be performing slightly behind the Gelid and Aeronaut and on par with the Aeronaut. I want to run 2 more test sets to confirm before I include it in my results but this basically confirms the results that ED linked.
Which article is in Kelvin?? Wasnt the one i linked earlier...and the (same one) you linked...thats in C...
How did we get Kelvin??????????? All that graph shows with a bit of subtraction, C minus C - no conversions needed... is the difference in temps, CELSIUS. It absolutely answers your question. If i missed where they mention Kelvin (hint, a k-type probe is a type of probe....does not mean Kelvin!), my apologies...but i dont see it...
So to answer this, its ~1.3C. What you see is what you get.
In other words what this graph shows is incredibly simple. Its the difference between the pastes. If the difference is 2c, then when using that paste, your temps will be 2c lower. Simple.
Last edited:
- Joined
- Sep 13, 2015
I think "Kelvin" got thrown in to the conversation based on this sentence that immediately follows the chart.
As others have said though since we are talking delta's here you can almost use Kelvin and Centigrade interchangeably. It does not help that the article uses both measurement methods as it just adds confusion.
As others have said though since we are talking delta's here you can almost use Kelvin and Centigrade interchangeably. It does not help that the article uses both measurement methods as it just adds confusion.
- Joined
- Jan 1, 2008
- Location
- Colorado Springs, CO
A lot of physics equations use Kelvin... Maybe he was using some of those?Then my apologies...
That (should have been) an easy google though...and was covered well anyway.
Why the hell hes using K is also beyond me....lol. seems more like a typo since everything else in C.
Either way ∆C = ∆K, so moot point (triangle thingy is math-speak for Delta). If we were to plot a time (x-axis) vs temp graph of something being heated, the Kelvin and C lines would be exactly the same, except the Y intercept. There is no slope or scaling factor to account for in converting from C to K.
