What is the Best Way to Apply Thermal Grease? - Part 1
By Rafael Otto Coelho on June 8, 2011
If there is one thing about computer hardware where there is no consensus, it’s the right way to apply thermal compound on a CPU. Today, we present seven ways to apply it, showing how the CPU looks after removing the cooler and how each way affects cooling performance.
We already explained how the thermal compound (a.k.a. thermal grease or thermal paste) works and a “how-to-apply-it” in our How to Correctly Apply Thermal Grease tutorial. Now, we will try seven different ways of applying it, testing each way, and comparing the cooling performance. Let's see if what we always mention - that less thermal compound is better than more - is true.
Each different method is presented in three pictures, one of the thermal compound applied to the CPU before installing the cooler, one of the CPU after the test with the cooler removed, and one of the base of the cooler after the test.
Figures 1, 2, and 3 present the test with a tiny drop of thermal compound in the middle of the CPU. As you can see, the amount was not enough to cover the entire CPU surface.
Figures 4, 5, and 6 illustrate the test with a slightly more thermal compound applied at the center of the processor. The CPU surface was totally covered; however, a very small amount of thermal compound exceeded its border.
Figures 7, 8, and 9 display the test using a larger amount of thermal compound. It resulted in more compound leaking from the borders of the CPU.
In Figures 10, 11, and 12, you see the test with a single stripe of thermal compound applied on our CPU.
Figures 13, 14, and 15 illustrate the test done with two stripes of thermal compound.
Figures 16, 17, and 18 show the test done by spreading the thermal compound uniformly on the CPU surface with a plastic spatula.
Finally, we applied a large amount of thermal compound to our processor. Notice the mess it made around the processor.
We tested the thermal compounds using the same testbed system that we currently use to test CPU coolers and thermal compounds, which is fully described below. Our Core i7-860 (quad-core, 2.8 GHz) CPU, which is a socket LGA1156 processor with a 95 W TDP (Thermal Design Power), was overclocked to 3.3 GHz (150 MHz base clock and 22x multiplier), and we kept the standard core voltage (Vcore).
We used a Zalman CNPS9900 MAX CPU cooler. The thermal compound we used was the Arctic Silver Céramique, which we tested some time ago. We chose this compound because of the large sample supply that we have, enough for making all the tests with the same thermal compound. Note that this thermal compound is very thick and viscous; therefore, a more fluid compound can behave differently.
Room temperature measurements were taken with a digital thermometer. The core temperature was read with the SpeedFan program (available from the CPU thermal sensors), using an arithmetic average of the core temperature readings. During the tests, the left panel of the case was open.
Operating System Configuration
Since both room temperature and core temperature readings have 1 oC resolution, we adopted a 2 oC error margin, meaning temperature differences below 2 oC are considered irrelevant.
The table below presents the results of our measurements.
|Thermal Compound Quantity||Room Temp.||Core Temp.||Difference|
|Tiny dot||15 °C||50 °C||35 °C|
|Small dot||15 °C||50 °C||35 °C|
|Bigger dot||15 °C||52 °C||37 °C|
|One stripe||15 °C||51 °C||36 °C|
|Two stripes||15 °C||51 °C||36 °C|
|Spread||15 °C||52 °C||37 °C|
|A lot||15 °C||53 °C||38 °C|
In the following graph, at full load you can see how many degrees Celsius hotter the CPU core is than the air outside the case. The lower this difference, the better is the cooling performance achieved.
We always say that less thermal compound is better than more, and the results of these tests confirm it. The best results we got were with a small (and a tiny) amount of thermal compound over our CPU. It seems to be irrelevant that the thermal compound covers the entire CPU surface, since most of the heat is produced at the middle of the processor heatspreader. In addition, it is a well-known fact that a CPU heatspreader is a little concave rather than really flat.
The quantity of thermal compound seems to be the key variable that defines a good application. A drop about the size of a grain of rice is the best quantity, while a drop about the size of a bean is too much paste. More than that can only make a mess around your processor.
Please keep in mind that these results apply to the hardware we are using, and a different CPU, cooler, or thermal compound can behave differently.
Answering the question made on this article title: Putting a small drop of thermal compound at the center of the CPU is the best way to apply thermal compound.