|
Yes, apply a thin even coat.
|
|
|
|
|
There are many school of thoughts.
some will put a pea size amount in the center and let the cooler spread it out when you screw it down; it works well for smaller CPU size.
other will put it in an X shape and let the cooler spread it out; maybe better for bigger cpu.
others will manually spread it out.
I'm on the X team (I've done it twice in my life,soooo ... )
youdoyou.
CI/CD = Continuous Impediment/Continuous Despair
|
|
|
|
|
Last time I found a place with a nice (and apparently serious) test and explanations about it:
Thermal Paste Application Techniques | Puget Systems[^]
M.D.V. 
If something has a solution... Why do we have to worry about?. If it has no solution... For what reason do we have to worry about?
Help me to understand what I'm saying, and I'll explain it better to you
Rating helpful answers is nice, but saying thanks can be even nicer.
|
|
|
|
|
This is from 2012, most CPU at the time would work without a cooler (kinda lol)
I'd like to see a more recent test with a comparable chip (intel 12700 or 13700)
CI/CD = Continuous Impediment/Continuous Despair
|
|
|
|
|
Funny thing is that although the article is dated 2012, the first comment is marked '15 years old' - but it seems to be unrelated to the subject of the article. So something is messed up in the comment system of that site!
To the subject: Do you think that different ways of spreading the thermal paste depends greatly on the CPU model? That the newer chips can benefit from a different way of applying the paste than the chips of ten years ago? Why would that be?
Maybe you are right, that we today should do it in a different way than we did ten years ago, but I do not immediately see any reason for why it would be different today.
After all, that is what the article really is about - not specific temperature measurements of specific CPUs!
(And: You are free to follow the instructions in the article to yourself repeat the tests with 12700 and 13700 and publish the results  ).
|
|
|
|
|
the tests has nothing to do with the CPUs.
It is just a plain comparison of different methods of applying the thermal paste, and what happens once you put the cooler on top.
As long as the new CPUs do no have a different surface...
M.D.V. 
If something has a solution... Why do we have to worry about?. If it has no solution... For what reason do we have to worry about?
Help me to understand what I'm saying, and I'll explain it better to you
Rating helpful answers is nice, but saying thanks can be even nicer.
|
|
|
|
|
I did what the instructions with the kit said and it worked.
"Before entering on an understanding, I have meditated for a long time, and have foreseen what might happen. It is not genius which reveals to me suddenly, secretly, what I have to say or to do in a circumstance unexpected by other people; it is reflection, it is meditation." - Napoleon I
|
|
|
|
|
I just put a few drops in the center. Also, my last cooling fan had thermal paste pre-applied on the radiator's contact surface.
There is only one Vera Farmiga and Salma Hayek is her prophet!
Advertise here – minimum three posts per day are guaranteed.
|
|
|
|
|
Don't get it on the contacts - everything will stop working. Ask me how I know.
You can clean it off with white spirit if you do and then start again.
Paul Sanders.
If I had more time, I would have written a shorter letter - Blaise Pascal.
Some of my best work is in the undo buffer.
|
|
|
|
|
Quote: does it matter how I apply it?
Yes
how much does it matter, meh
if it covers the whole cpu, not so much, cpu heats at different points.
that 2012 article posted above, well that simply testing IDLE temp, and how well spread is. well that more about how the contact of the heatsink. bits being missed on the edge already likely to be low heat areas, also if decent test would repeat each multiple times to remove issues with human pushing heat sink on.
unless overclocking to the extreme, drop, or multiple drops tiny, or X if doing thin amounts plus the heatsink push down will contact where heatsink actually contacts. too much will spill out side
too little, will still work, again unless doing extreme 100% load 24/7, ok coverage will be just as good as attempting to get that 1/1000 perfect fit.
|
|
|
|
|
Thanks. This is helpful! I probably won't OC the core frequency and voltages - the thing already runs at 5.1GHz, but I may tighten up the RAM timings (my CPU is unlocked) just to maximize my computing power, but I'm not going to get too crazy with OC anyway since I'm on air. That said, my air setup should handle some amount of OC.
To err is human. Fortune favors the monsters.
|
|
|
|
|
Are you sure it isn't already applied? Most heatsinks now will already have a thin layer applied to them and you just push it on carefully.
But if not, I've always just put a blob in the middle and let it push out.
|
|
|
|
|
I'm not sure yet. I considered that, but I haven't ordered the parts yet. I'm pulling the trigger on everything all in one shot in December after a contract pays out. I just want to have my ducks in a row for when everything gets here so I've been covering all my angles - cooling, wattage, clearances, etc and this has always been part of it for me, so I wanted to get good info on it. I started searching online and saw a bunch of conflicting information as I said, so here I am.
To err is human. Fortune favors the monsters.
|
|
|
|
|
Well if you go for any of the Corsair all in one liquid coolers(which I would recommend) then it would already have thermal paste applied.
|
|
|
|
|
My chassis is a thermaltake Level 20 VT, and while it's set up in theory to take either air or liquid, it's partially an open air chassis - it's enclosed, but there are gaps between the glass panels and chassis itself.
These require positive air pressure, which in turn requires 4 120mm fans mounted on top of my case blowing about 1cm away from the glass to push positive pressure over the edges of it.
It has kept dust and cat hair out for a year so far. I am loath to change it, and installing a radiator means ditching those fans for lack of space. At that point I lose the positive air pressure along one side, and potentially create a vacuum point for dust. Not a fan of that.
To err is human. Fortune favors the monsters.
|
|
|
|
|
Pea size dot on the die, if the die is square. Smuch the pea.
Rectangular requires a small line.
This is to avoid air bubbles, which you occasionally get when doing patterns like x's or such. In the ideal world, you'd want an even spread and as little paste as necessary, hence the pea comparrison for reference on what "little" means in this context.
Not much you can do wrong really.
|
|
|
|
|
Kate-X257 wrote: In the ideal world, you'd want an even spread and as little paste as necessary
That's why historically I've used the credit card approach - applying it and then using a credit card to smooth and thin it across the entire surface.
To err is human. Fortune favors the monsters.
|
|
|
|
|
Another vote here for "spread it evenly (eg with a card)". From an erstwhile physicist.
|
|
|
|
|
Even, thin layer. Don't let it get on other parts of the board as thermal paste is an electrical conductor.
|
|
|
|
|
As others have said - any way you apply it, you don't want it to get oozing out when you put the chip in. An even coat is sufficient.
It's a bit like a cooking recipe - follow the directions and you'll find it really is as simple as it sound.
As for the dangers of ruining your machine, there may safeguards built into your chip. There was one build where I was rushing and forgot to put the thermal goop in. When I went to power it up it worked great for about a minute, then it did a hard shutdown. I checked my connections, tried again, same result. Then I thought through the build step, and saw the unused tube and thought I'd ruined the machine. But after I added the goop, all was good.
I learned that day that there are thermal shutdowns built into many (most?) modern chips including the one I was using. It's an AMD and I'm still using it now, years later.
If it had been a few years earlier I might have recreated the exploding CPU chip scenario you often see on YouTube.
Cheers!
|
|
|
|
|
I just know that you need fresh paste, and it's expensive now. Using it provides absolute heat transfer to dissipate heat. Like with automotive parts, the stuff will spread evenly, but I can see using a spreader would provide a good even coat across the top of the CPU.The excess will just squish out.
In the future, like with the new Apple M1 chip, we won't need heat sinks anymore because the chip hardly gets that hot and uses way less electricity.
If it ain't broke don't fix it
Discover my world at jkirkerx.com
|
|
|
|
|
The biggest problem I have noticed in machines is too much paste as this reduces the conductivity of the heat. I have found the"credit card" technique results in too much paste. All you are trying to do is fill the micro pits on both the CPU and the heat sink surfaces.
|
|
|
|
|
i have studied physics at the university, but i doubt that will help... scientific progress is mostly of empirical nature and it will emerge that despite the theoretical beauty of the explanation to applying thermal paste, there is a nano-particle catch that yields a totally different result from intuition
but i can tell you what i am doing and what has worked for me in decades. i stopped using thermal paste. there was some tutorial about it that made sense to me. the tutorial was so long ago that maybe was from the previous century, but i remember that the problem i was researching was about how the thermal paste lose it's physical properties in time. after so many working hours at high temperatures
since then i would clean the surface of the heat sink and the CPU cap to a crystal mirror clear state and then mount the heat sink. that's it
the moral of the story is: either use some very expensive HQ thermal paste, either not use paste at all
cheers
ps - "Ivy Bridge(22 nm) temperatures are reportedly 10 °C higher compared to Sandy Bridge(32 nm) when a CPU is overclocked, even at default voltage setting. Impress PC Watch, a Japanese website, performed experiments that confirmed earlier speculations that this is because Intel used a poor quality (and perhaps lower cost) thermal interface material (thermal paste, or "TIM") between the chip and the heat spreader, instead of the fluxless solder of previous generations. The mobile Ivy Bridge processors are not affected by this issue because they do not use a heat spreader between the chip and cooling system. Socket 2011 Ivy Bridge processors continue to use the solder"
|
|
|
|
|
It works exactly the way that wet fingers on your hot porcelain coffee cup transfer more heat to your fingers. Provided it gives good contact between the two surfaces with no gaps/bubbles, it can be as thin as you like. If it squeezes out when you press down the heat sink - then you've done too much. Less is more once you have coverage.
Otherwise, everything else between is more or less between tolerances.
|
|
|
|
|
So it's like doing bodywork on a car, minus the block sanding. 
I get the concept. I've always used the credit card approach to applying it. It's just after reading some stuff I questioned my technique.
I have had paste harden over time, but it was by the time the system was useless anyway. I'm just triple checking my handiwork since these new Intels are like little furnaces.
To err is human. Fortune favors the monsters.
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
