Thermal Paste applying metods

[westman 257]

Hi

Applying thermal paste , i see a lot diskussions her mostly recomends center ricecorn as the right metod.

I prefer the spread metod , Done it for more than 15years and it always worked good.

Here is a video from Der8uer Word class Overclocker builder of special pc at Caseking and lot of more.

Enjoy

[martin-w 3,234]

A few years ago the company Innovation Cooling, who make the diamond paste, did lots of research on this.

 

Their research concluded that the "spread" method resulted in air pockets and compromised cooling. Not vastly, would probably only make a couple of degrees difference, but it was compromised.

 

Their research suggested that the best method is the blob in the middle method. Perfectly even coverage, no air pockets.

 

http://www.innovationcooling.com/application.html

 

If you are happy with your method, then so be it.

 

 

Regarding your video, the guy may be barking up the wrong tree in some respects as he assigns too much importance to covering "the entire" heat spreader. It's not necessary, the die is small and in the centre of the heat spreader. Covering the "entire" heat spreader makes very little difference at all, in my experience undetectable.

[westman 257]

Have read this from IC,,

Iam happy with my as you with your.

 

The Guy on the video has Done much research on thermal paste.

Her is his latest tests.

Go to the meny , there you have some nice OC- guides

http://overclocking.guide/thermal-paste-roundup-2015-47-products-tested-with-air-cooling-and-liquid-nitrogen-ln2/

[superglide17 11]

Use this and you won't go wrong. 

 

http://archive.benchmarkreviews.com/index.php?option=com_content&task=view&id=170&Itemid=38&limit=1&limitstart=5

 

This link puts you on page 6 of 6. Scroll back to page 1 for the entire report.

[martin-w 3,234]

At the end of the day, unless you do something utterly stupid and slap a ton of TIM on the heat spreader, you wont see a huge difference.

 

So if anyone is happy with their method and results so be it.

 

My point is though, that covering the "entire" heat spreader is unnecessary and makes very little difference if any.

 

The reason covering the "entire" heat spreader makes little if any difference, is that the "hot spot" is directly above the die, and the die is much smaller than the heat spreader. In addition, between die and heat spreader is either solder, epoxy solder, or Intel's thermal paste. Thus the vast majority of the heat is in the centre.

 

Cooling the heat spreader, beyond the region in the centre is unnecessary, as there is no direct heat spreader/TIM/die contact. Beneath the outer regions of the heat spreader there's merely an empty void.

 

@westman Couldn't see anything in his test where he looked at covering the entire heat spreader as opposed to most of it. If it is there, can you show me where?

[martin-w 3,234]

A Philosophical Debate: The Application Method

 

It’s tough to pick a technique for applying paste. Any method only works well if paste quantity and viscosity is absolutely correct for the particular application. In light of the hot spot discussion, however, we believe that smearing paste on the whole CPU is quite pointless and a thing of the past. Instead, we want to focus on the particularities of the CPU, its heat spreader, the heat sink, and the mounting method (in particular the mounting pressure).

 

 

 

Trying to spread a high-viscosity paste with a credit card is a fool’s errand. You'll waste a lot of time and won’t achieve a thin, smooth layer. Yes, you can try to put a latex glove on your hand and use your index finger. But even with this method, the risk of applying too much paste is significant, especially if you have no practice. The higher the viscosity, the less successful you can anticipate being trying to "paint the wall".

 

 

http://www.tomshardware.com/reviews/thermal-paste-heat-sink-heat-spreader,3600-5.html

[PaulGR 168]

#1. ArcticSilver.

#2. Match "head" sized droplet right in the centre of the cpu.

#3. Tighten the cooler screws little by little (one turn on each screw then next screw across centre).

#4. Take into consideration break in period which is about 50-90 hours.  Temps might be few degrees higher in this time. Many people do not know this and disassemble everything and reapply paste for no reason.

#5. Enjoy.

 

Using this method (not the first time I did it ) I'm a running Devil's Canyon at @4.5GHz with Zalman air cooler and temps mid 60 celsius under heavy load. The paste spread method just doesn't work as good. Micro air bubbles develop when applying paste this way.

[martin-w 3,234]

#1. ArcticSilver.

#2. Match "head" sized droplet right in the centre of the cpu.

#3. Tighten the cooler screws little by little (one turn on each screw then next screw across centre).

#4. Take into consideration break in period which is about 50-90 hours.  Temps might be few degrees higher in this time. Many people do not know this and disassemble everything and reapply paste for no reason.

#5. Enjoy.

 

Using this method (not the first time I did it ) I'm a running Devil's Canyon at @4.5GHz with Zalman air cooler and temps mid 60 celsius under heavy load. The paste spread method just doesn't work as good. Micro air bubbles develop when applying paste this way.

 

I agree with you, as I said in my posts.

 

The only difference is I use an NH-D14. And I use MX4 paste, so no break in period.

 

The other negative about the spread method of course is that you end up with a thicker layer than the blob method.

 

 

http://www.arctic.ac/uk_en/mx-4.html

[westman 257]

Mayonnaise is good to not sure if it blob or spread but in some tests on pair with MX4.

[TheFamilyMan 279]

Wow...this topic never seems to get enough bandwidth...for how many years now has TIM application been beaten (err...smeared ) to death???

 

I'm suprised that people still recommend using Artic Silver, for it dries out over time which substantially reduces its effectiveness. There are now other TIMs which are just as effective that don't have this problem.

 

I'd wish that posts such as westman's above would use a smiley to point out it's in jest (even if true).  I always figure that there are some readers out there who are clueless and may take it as advice.

[OneWhoKnocks53 51]

Gelid, very good test results and I swear by it. A Haswell takes a stripe down the center along the die profile about 1mm wide. Excellent results.

[PaulGR 168]

Wow...this topic never seems to get enough bandwidth...for how many years now has TIM application been beaten (err...smeared ) to death???

 

I'm suprised that people still recommend using Artic Silver, for it dries out over time which substantially reduces its effectiveness. There are now other TIMs which are just as effective that don't have this problem.

 

I'd wish that posts such as westman's above would use a smiley to point out it's in jest (even if true).  I always figure that there are some readers out there who are clueless and may take it as advice.

It's because it is very effective and anybody that actually oveclocks CPU's will replace paste every year or two which is still waaaay under its drying out phase.

I tried others but they are not better so I stick with the proven.

[martin-w 3,234]

I've been overclocking for years, and have never found the need to reapply TIM every year or two.

 

There's a comparison chart here...

 

http://www.tomshardware.com/reviews/thermal-paste-performance-benchmark,3616-18.html

 

As you can see, MX4 beats Arctic silver, and importantly MX4 has no break-in period and doesn't dry out as far as I recall.

 

You'll also notice that there's hardly anything in it. just a degree or so.

 

 

http://www.tomshardware.com/reviews/thermal-paste-performance-benchmark,3616-18.html

[stans 702]

I've been overclocking for years, and have never found the need to reapply TIM every year or two.

 

There's a comparison chart here...

 

http://www.tomshardware.com/reviews/thermal-paste-performance-benchmark,3616-18.html

 

As you can see, MX4 beats Arctic silver, and importantly MX4 has no break-in period and doesn't dry out as far as I recall.

 

You'll also notice that there's hardly anything in it. just a degree or so.

 

 

http://www.tomshardware.com/reviews/thermal-paste-performance-benchmark,3616-18.html

 

Same here, never felt the need to reapply once I had a good application of paste and even after years of use, I have not seen a change in temperatures.

[PaulGR 168]

I just ordered MX-4 since I'm upgrading to liquid and hopefully you guys are right