Thermal paste kpt 8 impact on health. Big thermal paste testing

Due to the progress of modern microelectronics, the speed of central processors and other components of a modern computer is rapidly increasing. Often, the growth of computing power is accompanied by an increase in heat dissipation of one or another PC component.
We have to admit that today semiconductor technology faced with the problem of heat dissipation from the crystals of the most powerful chips. So, the central processors and cores of top-end video cards are those representatives of the consumer microelectronics segment, where heat dissipation per square centimeter approaches 100 watts. For particularly powerful chips, this figure is further increased.
As it turned out, it is very difficult to remove heat from such a small area... And so far it is impossible to radically reduce the heat release of the mentioned components without resorting to very expensive research in the field of semiconductor and nanostructure technologies.

Of course, manufacturers are taking adequate measures - they have improved and continue to improve the cooling of certain computer components, promote water cooling to the masses, and develop new designs of air CO. A vivid example of the expression of this movement in practice is the current "epoch of supercoolers", which literally overwhelmed store shelves and the minds of most users with masterpieces of technical art made of copper, aluminum and heat pipes.
A high-quality cooling system is the key to low temperatures of PC components, silence in operation, and the possibility of overclocking the system. However, in this case, it must be remembered that a “barrel of honey” can be easily spoiled by a “fly in the ointment”.
Schematically, heat removal from a heating component (for example, a central processing unit) can be displayed as follows: "processor - thermal interface - cooling system"(by the way, the heat-dissipating cover of a modern CPU contacts the core through another thin layer of the same thermal interface, but we will miss this point in this material, because the user cannot influence the characteristics of this factor). Most users forget about the binding component, which can be a fabric sticker impregnated with various substances, a small sheet of foil, paste, ointment, liquid, or use “what was in the box” - a free substance that comes with the purchased cooling system . And many beginners do not even suspect the existence of thermal interfaces and their use in modern computers!
Is such an approach to seemingly trifles justified? Not always, therefore, today's material is intended to demonstrate the importance of the topic under consideration and draw the attention of readers to one of the important aspects of cooling PC components - the effect of the thermal interfaces used on the quality of the heat sink.
Our goal is to study various substances that enthusiasts use in order to achieve the most efficient heat transfer from a processor die, graphics core, motherboard chipset to the base of the cooler or water block. This provides an additional "margin of safety" during overclocking, or simply reduces the overall temperature indicators of the components and facilitates the operation of a particular PC node.

Heat transfer: a bit of theory

For those who have forgotten or do not know what thermal interface, we will give the most understandable definition to most: this is the same layer, consisting of some special substance that exists between the processor and the base of an air cooler or water block.
As you understand, the surfaces of the chip itself and its cooler are not ideal in terms of absolute evenness. In conditions of mass industrial production, it is often impossible to provide a very high surface finish and its geometric plane. Even on visually very even bases, whole areas of microgeometry with non-ideal contact remain, which, without the use of thermal interfaces, turn out to be filled with air molecules. These may be miniature notches, bulges or micro-scratches that are not visible to the naked eye.

Heat is transferred between the contact surfaces by means of conduction. This term refers to the process of exchange of kinetic energy between the molecules of substances together with the diffusion of electrons in metals. Heat transfer by conduction will take place provided that the bodies are in contact with a temperature difference. In all cases, the heat flux will be directed towards the fall of the gradient of absolute values. Consequently, the main part of the thermal energy goes in the direction from the chip to its cooler.
Convection and radiation alone are not capable of removing huge heat fluxes over a small area of ​​a microchip, and only partially participate in the overall heat transfer.

If we touch upon theoretical physics a little, then we should remember that thermal conductivity of metals is determined by the vibrations of the crystal lattice and the movement of free electrons (the so-called "electron gas").
With an increase in temperature for all metals, the electrical conductivity, and, as a result, the thermal conductivity decreases (these two phenomena are interconnected and one does not occur without the other). Conversely, as the temperature decreases, the thermal conductivity increases.
The presence of free electrons determines the high electrical conductivity of metals.
Knowing this, it becomes clear why aluminum, copper, silver and their alloys are widely used in the manufacture of parts for cooling devices. These common metals have the highest electrical and thermal conductivity known to the mass industry. In addition, they are relatively easy to give the desired shape by appropriate processing. Here are brief characteristics of the thermal conductivity of the most accessible metals and some interesting materials that are used in various industries:

But let's get back to our "rams": we have two surfaces - a chip chip and the base of the cooling system, which is tasked with cooling it. The thermal interface displaces air and forms a film between them, consisting of a substance with low thermal resistance.
Various pastes also make it possible to mechanically separate the heat source and its cooler, which is necessary in case of replacing any PC component.
If fasteners for radiators are not provided, or a more rigid fixation of heat sink devices is necessary, then hot-melt adhesives and special stickers are used. These types of interfaces are not considered in this article, however, based on the data given in one of our earlier ones, one can roughly estimate the effectiveness and other characteristics of some products of this kind.

We hope that the readers have no more questions on the theoretical part, so we will move on.

Test methodology

When choosing a reference paste, we proceeded from the following considerations:

• mass availability of the test sample;
• high efficiency;
• ease of application and rinsing;
• low cost.

We think you have already guessed that we are talking about a rather old masterpiece of the domestic chemical industry - paste KPT-8. The key to total popularity for a huge number of users is the excellent price / quality ratio of this product.
But not everyone is satisfied with the parameters of the specified paste. Among those who use a PC intensively, there are so-called "racers", enthusiasts. They crave fame and records, boost iron operating modes by all available means, thereby squeezing out megahertz, parrot-power, and, as a result, creating more difficult operating conditions for various PC components, which invariably lead to increased heat dissipation. It is clear that in a state of record performance, the system will be very unstable. In this case, every degree and every extra watt of heat removed will be decisive.
In such conditions, any component and link of the cooling system is subject to increased requirements, and sometimes even exceptional ones to the thermal interface, because nothing worsens heat dissipation like poor-quality thermal paste.
As we have already said, powerful microprocessors of modern PCs are perhaps the only segment of consumer microelectronic equipment where the heat dissipation of a crystal often reaches more than 100 watts per square centimeter. As it turned out, it is very difficult to remove heat from such a small area, so many companies are engaged in research and development of devices and substances designed to effectively remove heat from central processors and video card cores.

Within the framework of one good test on a PC, everything seems extremely clear and understandable. However, when reviewing and comparing a significant number of reviews and articles published on the web, we sometimes found conflicting research data and ambiguous conclusions made by their authors.
In almost all cases, directly or indirectly, emphasis was placed on the processor on which the testing was carried out and the cooling system used.
This prompted Test Lab site to collect all the thermal pastes available to us and conduct their own independent investigation using a special test bench.
After reviewing the results of a study of the characteristics of thermal pastes carried out on the CPU, you can see that in the vast majority of cases it is difficult to feel the difference between samples with similar characteristics. Much depends on the architecture and TDP of the processor. As the heat dissipation of the heater increases, the difference between the investigated thermal pastes becomes more and more obvious.

We noticed another interesting point. So, manufacturers indicate the thermal conductivity of pastes on the packaging of their products, but it is not enough to determine the winner by this indicator.
The reason is simple - different methods of measuring thermal conductivity give different values. Even conducting studies using a single method in several laboratories does not exclude inaccuracies in the final results. For example, the paste may have a different contact layer during the test, and this will directly affect the numerical expression of the subjective results of the study.
Undoubtedly, only empirically through a single _big_ comparison using a single method real differences can be found between test participants.

As a stable source of heat, we have chosen the MARK Sea Launch experimental test stand that has proven its right to life.

On this modification, the heater core has an adapter with a small area (less than 12x12 mm), which makes it difficult to transfer heat from the heat source to the cover. The top, polished part of the heater "emulates" the heat spreader of the processor. Its dimensions are 25 x 25 mm, thickness - 2 mm.
With a power output close to 100 watts, the heater becomes like a powerful overclocked processor, which would be very difficult to cool in real conditions. The microprocessor-based thermal sensor embedded in the core of the heater is capable of registering temperature changes in tenths of a degree.

The heater power was set to 100 W. This value fit perfectly. It's nice that the values ​​of the final temperatures turned out to be approximately the same as they take place on modern processors with average CO.

Accordingly, our powerful heat source will require an equally powerful cooler, and it is possible that it will be liquid. But on a water cooling system, testing thermal pastes is difficult. It is possible to introduce an error into the test due to the presence of an intermediate coolant (water) acting as a capacitor between tests. This means that the system will have a certain inertia. Such moments are always an inconvenient "bottleneck" of long and laborious research.
When testing air coolers, the test results are more stable, which is confirmed by testing control samples at longer intervals.
The basis of our cooling system is a radiator manufactured by Noctua, model NH-U12. This sample is assembled on four U-shaped heat pipes that are in contact with a copper base, and solid aluminum plates. We decided to "overclock" it a little, and equipped the radiator with two 120 mm Sunon KD1212-PMS1 industrial fans with a capacity of 181 cubic meters per hour each.
This configuration made it possible to achieve a record efficiency of the air-cooling system, which significantly exceeds the power of budgetary air-cooling kits.
The cooler was pressed with a pair of screws through standard mounting holes socket 939. During the tests, there were no shock-absorbing springs, the clamping force was not regulated. In each test, the screws were tightened to the limit, which guaranteed the formation of a thinner intermediate layer of thermal paste and, as a result, the most correct final result.

In the room where the testing was carried out, the air temperature was at the level of 27.5°C, monitoring was carried out continuously. If the threshold of this value was exceeded by 1 °C (in any direction), the stand automatically issued a warning signal, and the study was suspended.

Each paste, if possible, was checked at least twice. In this case, the contact layer was applied again, and the result obtained will be refined.
For pastes that showed unexpected, suspicious results, or took some time to fully acquire their optimal condition, the test was repeated after a few days. * .

Please pay attention to diagrams- they are deliberately built "incorrectly" to more clearly demonstrate the difference between the tested interfaces. So, the mark of 45 ° C was taken as the initial mark, so do not be alarmed by the relatively large visual difference between some substances on the graphs showing the final results.

* during the entire time of research, the same temperature was kept in the room

Thermal paste parameters

Regardless of the model and manufacturer's name, any samples of good pastes must meet the following requirements:

1) the lowest thermal resistance;
2) stability of properties in a fairly wide range of operating temperatures;
3) ease of application and ease of rinsing;
4) immutability of properties over time.

We believe that it is necessary to dwell on each of them in more detail.

Least thermal resistance of the applied layer will ultimately determine the limiting thermal conductivity of the paste for a given contact area. If the values ​​of operating temperatures are within reasonable limits and the substance does not lose or change its properties during the entire period of operation, then the thermal conductivity parameter will be the only and decisive one.

Working temperature range
All high-quality thermal pastes work fine in a home computer at standard temperatures. Recall that in PC, in most cases, we are dealing with values ​​of the order of 30–80°C at the point of contact.
Within this "positive" range, a comparison will be made.
Temperatures above 100°C, for obvious reasons, are not considered in principle. Also, everything below zero up to -200 ° C is already extreme, which is a topic for another conversation. We do not know how various pastes will behave in this case, and we will not conduct experiments in this direction today.

Ease of application is a very important factor, and if the paste is applied with great difficulty in a thin layer on the contact surfaces, or is washed off very poorly, polluting everything around, then this causes certain problems for the user and definitely reduces the overall score, even despite other high parameters.

Stability of properties over a wide time range determines the "survivability" of the paste. For example, we know a lot of cases of drying or partial drying of low-quality samples of KPT-8 during its operation even for one month! Naturally, a thermal interface that demonstrates similar performance for a given parameter can, at best, be used only for short tests.

Characteristics such as dielectric strength and dielectric constant, volume electrical resistivity and other special indicators for any PC user are mostly irrelevant.
In the process of getting acquainted with thermal pastes, we will not dwell on the description of physical and chemical properties, as others do, but focus only on the main criteria for us.

Acquaintance with thermal interfaces: general impressions

KPT-8

First, we will smear our reference paste, which we use with success in all tests. You probably already guessed that we are talking about the domestic KPT-8. One of the G8 samples was purchased at the Kiev radio market. The filling of a 10cc syringe usually lasts a long time, but we always take a margin of pasta. The true manufacturer of the paste is unknown, there are no identification marks.
In ordinary syringes, the paste is packed from a large container, and obviously not far from the place of their subsequent sale.
This sample of KPT-8 is squeezed out with some effort, but with frequent use, you can quickly get used to it.
In appearance, the paste is white, does not contain any inclusions, and is quite thick.
After application, for correct testing, the paste must be spread on the surface with a thin layer. For these purposes, a spent card for city payphones, or a clean user's finger, is well suited :)
Typically, manufacturers declare the thermal conductivity of this type of paste in the range of 0.5-0.8 W / (m x K) (hereinafter, in the characteristic of the thermal conductivity unit, degrees Celsius are replaced by a more common unit - Kelvin). It is she who will be present in all comparative tests on the diagrams under the designation "Standard".

KPT-8 is also present in the tests, but from a smaller syringe, on which a red sticker with the image of Mendeleev and the name of the contents flaunts (popularly nicknamed “Mendeleevskaya”).
Like the first sample, it is very common, but it is purchased elsewhere on the radio market :).
It applies and spreads a little better than the previous one, and is not as thick. It does not differ from our standard in any way.

The next sample is also a "figure eight", with the same "hacky" sticker. But now it is already called as kTP-8 - this is something new! I wonder if they are different? (looking ahead, let's say that nothing). Obviously, there was a problem with the name of the packers :).

Oh my God, the next test participant is also KPT-8! But this time the pasta is really special. The originality lies in the use of beryllium oxide in its manufacture, VeO. This sample in recent times actively advertised in some places of sale. True, its price and "packaging" are no different from Mendeleevskaya.
It's funny, but there are legends about the use of beryllium oxide (BeO) as a heat conductor on the Web. There are rumors that this is a rare military-space purpose paste with amazing characteristics.

In our case, vague pictures from the science fiction movie "Shadow", a beryllium sphere, ancient evil, and all that appear before our eyes;).
Be that as it may, but in the specified GOST 19-783-74 nothing is said about beryllium oxide at all, in fact, it does not say anything about the exact composition of the paste.
For those who do not know, we recall that in traditional KPT-8, the heat conductor is finely dispersed ZINC oxide. What about beryllium?
Raised information on the analytical chemistry of this metal suggests that, indeed, beryllium oxide combines high thermal conductivity and low electrical conductivity. It is used in special ceramics and in many branches of science and technology. It is quite possible that thermal pastes can also be made on the basis of BeO.
By the way, beryllium compounds are definitely poisonous, but the degree of this indicator depends on the specific compound. There was no reliable information about the toxicity of the oxide, as well as the very fact of the presence of BeO in the paste in question.
To establish the truth, it is necessary to conduct a chemical analysis of the paste, and this is already specific problem for any test lab even more and x Internet resources. Therefore, we will limit ourselves to the test.

AlSil-3

A very popular thermal paste among domestic users. Produced by the Moscow company "GM Inform". There are a lot of rumors on the Internet about the substance in question. Apparently, one of the reasons for this is its maximum declared thermal conductivity, which is approximately 2 W/(m x K), compared to 0.8 for KPT-8. On the forums, some users report about excellent results with the use of AlSil-3, in contrast to another domestic rival, while others do not feel any difference, or, on the contrary, approve of the G8 more. It is claimed that there are fake AlSil-3 based on toothpaste * . There are also suggestions that the manufacturer is experimenting / saving, and does not always guarantee consistently high performance of the products.
* for fun, we also tested toothpaste to find out if a fake can be made in this way; See research data at the end of the article.

Two samples of the substance in question came to us for tests - the original, branded AlSil-3, released in the second quarter of 2006:

And another slightly larger syringe marked AlSil-3:

Visual comparison showed that the pastes from both syringes are no different. The substance in each case has a characteristic gray tint. This feature of AlSil-3 is dictated by the presence of aluminum nitride in it, which acts as a heat conductor. There are no inclusions in the composition. The paste is squeezed out easily and smeared easily. Of our two samples, AlSil-3 in a larger syringe was released quite a long time ago, approximately in 2002. However, during testing, no difference was found between the pastes.

This thermal interface is supplied with akasa coolers.
The paste is in a small syringe, has a white color, compared to our standard, it is more liquid and easier to smear.

The thermal conductivity declared by the manufacturer is more than 7.5 W/(m x K). Theoretically, this is about 7 times more than that of KPT-8! And what will happen in practice? ... Testing will show!

AOS is a well-known manufacturer of thermal interfaces abroad.
Silicone paste, #54013, packed in a branded syringe, came to us for testing.

It is white in color and easy to apply. Washes off without any problem. The consistency is very runny.
The declared thermal conductivity of this sample is 0.73 W/(m x K).

Apus-TMG 301

We got this sample from the XC-801 cooler kit from LEXCOOL.

The paste has a slight grayish tint and resembles AlSil-3.
Consistency is quite runny. A thermal conductivity of about 4.5 W/(m x K) is indicated.

Arctic Cooling MX-1

This paste is one of the non-traditional products of the Swiss company Arctic Cooling, which specializes in the production of quiet and high-quality cooling systems. We have already talked about this product, so we will not dwell on the details.
The substance is in a branded syringe, which, by the way, changed its appearance. Ash colored paste. Squeezed out in small lumps. For proper application, it must be rubbed into the base of the cooling system and the processor cover. Note that on both surfaces you need to apply very little paste, remove the excess.
This is the "old" version of the packaging:

And here is the pasta new packaging in a thinner and longer syringe:

Arctic Alumina

This paste is probably the brainchild of the most famous and advertised foreign manufacturer of thermal interfaces - Arctic Silver.

Arctic Alumina is made from aluminum oxide. The paste is white, applied to the surface easily, just as easily smeared. The declared thermal conductivity is over 4.0 W/(m x K).

Arctic Ceramique

The heat conductor in the paste is a mixture of aluminum oxide, zinc oxide and boron nitride; The manufacturer does not indicate the proportion of substances.

Arctic Ceramique, like all Arctic Silver products we tested, is based on a proprietary highly stable polysynthetic base. There were no problems with applying and rinsing the product.

Arctic Silver 3

One of the most famous silver-based pastes. The composition is a dark gray substance with a greenish tinge.

The manufacturer specifies a content of approximately 70% fine silver by volume of the paste.
The substance is squeezed out and applied without problems, removed quickly and easily.

Antec Reference

Looking at the syringe, it is easy to guess where and by whom the paste was produced.

Interestingly, the package claims to reduce the processor temperature by 4°C to 15°C due to the use of this thermal paste. We still couldn't figure out in which case it is possible to achieve such outstanding results... Perhaps, the manufacturer's marketers mean the difference between installing a cooler without using any thermal interface, and using Antec Reference :)
The product in question has exactly the same characteristics as the Arctic Silver 3, and the tests carried out confirm this.

Arctic Silver 5

This product has replaced Arctic Silver 3 and has improved performance. This time, the presence of 88% finely dispersed silver of high purity is indicated in the composition of the paste.

The substance is dark gray in color, rather thick consistency. To spread the paste in a perfect thin layer, you need to spend certain time.
The declared thermal conductivity of this product is impressive - about 8.7 W / (m x K).
Many well-known companies use Arctic Silver products under their own brand, often with their own packaging. For example, Arctic Silver 5 is referred to as Thermal Grease #2 by Thermaltake.

This thermal paste comes with the Asetek WaterChill KT03A water cooling system.
The substance is contained in a dense white bag, which is enough for several applications.

The paste is white, sometimes liquid, but mostly comes in small clots. It spreads well and washes off easily.

Data Cooler

This thermal interface is supplied in bags with coolers produced under the brand of the same name.

The pasta is very reminiscent of the Polish W.P. - much more liquid than KPT-8. There were no problems with application.

Standard "silicone" thermal paste.

Abroad, DC-340 is found in many manufacturers of chemical products. Our paste is in a plastic tube. When squeezed out, it turns out that it is very thick, stretches, has a white color. Typical thermal conductivity of DC-340 is 0.42 W/(m x K).

fanner 420

This thermal paste is also known as Evercool 420, but in fact we have a product from Stars with the same digital designation - 420. As you can see, this thermal compound is very popular among many suppliers.

Paste white, very liquid. Specified thermal conductivity is 2.062 W/(m x K).

GeIL GL-TCP1b

Quite an interesting example. Recall that the company Geil produces RAM. A tube of thermal paste could once be purchased separately, or found bundled with some modules as a free bonus for the buyer.

The composition is very beautiful, so to speak, golden in color. The manufacturer indicates the presence of 5% copper and 5% silver (by volume).
I wonder what thermal conductivity this "mix" has? On the label of the syringe, you can find a value of 1.729 W / (m x K), which most likely looks like the truth. However, the actual effectiveness of GeIL GL-TCP1b will be determined by testing.
The composition of this paste is liquid, homogeneous, applied in layers, smeared easily. This substance is removed a little easier than the ever-memorable "silver".

gigabyte

We fished out this paste from the Gigabyte 3DGalaxy CBO kit.

Note that the manufacturer does not provide a full syringe, and the substance is only enough for one or two water block installations per processor.
The paste is white and very runny.

Coolance

We got this sample from the Koolance Exos CBO kit. In front of us - Stars 360, keep that in mind.

Ash colored paste. Thick but spreads relatively easily. A rather high thermal conductivity is declared - about 4.5 W / (m x K).

This product is included in the set of coolers manufactured by Noctua. The paste is in a small syringe filled to capacity.

The substance is white, unremarkable, liquid and slippery.

Pasta Siliconowa

This paste is quite common in the market. Manufactured in Poland. We hope you understand that in “silicone” pastes, the heat conductor is not the substance that is used to increase certain parts of the body of female representatives, but mainly metal oxides :).

The paste is contained in a tin tube. White colour; thick, like our standard, but applied and smeared easily.
Note that squeezing the paste out of such a tube is extremely inconvenient.

The next thermal paste is also Polish, packaged in disposable bags. There are no identification marks, but the seller managed to find out the abbreviation of this substance - W.P.

The paste is very liquid, applied very well, in a thin layer.

Panasonic

Do not be surprised that the well-known company Matsushita Electric Co. (owner trademark Panasonic, among other things, produces thermal pastes for use in its own production.
The substance intended for retail sale is packaged in a small round jar with a red lid.

The pasta itself turned out to be similar to whipped cream, “airy”. As soon as the cooler is installed on the processor, it will instantly squeeze out the excess amount of the substance, so in this case you don't have to worry about the thin working layer.

Perhaps this is a Stars product. Many manufacturers use thermal pastes from this vendor, often "releasing" them under their own brand.

The thermal conductivity of the composition indicated on the syringe is 0.88 W / (m x K), which is very similar to the characteristics of our standard. The paste is white, very runny and easy to spread.

Shin Etsu

We cannot name the exact model of this substance, but buying it will not be a particular problem. In some cases, the user may receive a package that will be provided with a sticker. If you believe the statements of the sellers, these syringes are filled with thermal paste from Shin-Etsu MicroSi, Inc.

It was not possible to find out the nominal characteristics of the product. The paste is white, very similar to Mendeleevskaya KPT-8. Applied normally, a little "slippery".

Stars (soft pack)

Another product produced by Stars. Perhaps it is no different from other similar substances.

In appearance, the paste is white, somewhat slippery, with a consistency reminiscent of the substance from the Data Cooler cooler kit.

stars silver

And this is non-standard thermal grease from Stars, very similar to Titan TTG-S104. The substance is well applied and smeared on the surface of the heat-distributing cover of the processor.

True, during its washing off, the same problems arise as with the "silver".

Stars 700

Like other relatives from this manufacturer, it is also a very common pattern. Interestingly, the manufacturer indicates the presence of 25% silver by volume in the composition of the paste. Unfortunately, we will not be able to verify this statement as part of today's test.

Specified thermal conductivity is 7.5 W/(m x K). The composition is silver in color, lays down in layers. Very reminiscent of "titan silver".

Aero 700

Paste from a set of coolers from Aerocool.

In fact, we have Stars 700 in front of us, but in a different package: all the same 25% silver by volume and thermal conductivity at the level of 7.5 W / (m x K).

Sil more

The paste comes in a transparent plastic bag. Appears white, very liquid.

When squeezing out on the processor cover, in addition to the paste, some other transparent substance appears. This thermal interface is applied easily, washed off simply.

Shin-Etsu MicroSi G-751

There was nothing else on the syringe, except for the identification mark in the form of the name of the manufacturer, but we managed to find out the true name of the product - G-751.

The paste was included in one of the coolers for server Intel processors xeon. The composition has a gray color, rather thick, is in a thin and long syringe. The thermal conductivity declared by the manufacturer is 4.5 W/(m x K).

Shin-Etsu MicroSi MPU-3.7

This sample of thermal paste has been preserved with us since the era of processors AMD Athlon XP (K7)!

I wonder what result this substance will demonstrate. The paste itself is dark gray in color, very thick.
MPU-3.7 is not smeared in the best way, if I can put it that way. Similar to Arctic Cooling MX-1, it needs to be rubbed into the surface for normal application in a thin layer.

Titan TTG-S104, -S103 (silver)

This substance was previously supplied in a small bag or in a syringe with coolers manufactured by Titan. With us, it is one of the most famous and common thermal pastes. For the specific color and composition, she received the nickname "silver".

The paste is really silvery in color, but nothing more: it seems to us that silver is absent in the composition by definition, although the manufacturer claims some percentages. It seems that the heat conductor is a finely dispersed aluminum powder.
The paste is squeezed out easily, lays on the surface in layers, smears well. The packaging in a syringe is more convenient, so do not get lost when choosing between S104 and S103 - there is no difference between them, except in the package, we have the same substance. The peculiarity of the "silver" appears at the moment of washing off this interface - the composition quickly, as if expediently and spontaneously, appears on some parts of your body, and on objects that have undergone the slightest contact with the paste or the user's dirty hands.
We, perhaps, have never seen such a "dirty" thermal interface.

Titan Nano Blue

One of the options for replacing the classic "silver". In the form of a small syringe, it is included in the set of coolers and water cooling systems from Titan. It is a very common model, but how successful - tests will show.

The composition of the syringe itself is of a radical blue color, lays down in layers, and is not smeared in the best way. Rated thermal conductivity - more than 2.5 W/(m x K).

Titan Nano Grease TTG-G30010

This thermal interface is the newest product of its kind from Titan. Apparently, it will replace the well-known Nano Blue paste.
The tested product comes in a small flattened syringe that comes with new coolers from the manufacturer in question.
The composition is grey. The paste is very thick, viscous and dense, so it will take some time to apply evenly. Declared thermal conductivity - 4.5 W/(m x K).

It is worth noting that the same thermal paste is available separately in retail:

The only difference from the sample we tested is the delivery in a syringe of a noticeably larger volume and, as a result, the marking TTG-G30030.

Thermopox

Before us are the products of the Amepox company, well-known in certain circles.

The substance in question was taken from a two-component kit designed for gluing heatsinks to memory chip and/or power transistor cases. The heat conductor is a rather original mixture, the basis of which is liquid finely dispersed copper.
The specified thermal conductivity of the composition is 6.4 - 6.8 W / (m x K).

Zalman CSL 850

A very common pattern. This paste is included in the vast majority of coolers manufactured by Zalman, which determines its mass availability and wide popularity.

The composition is in a miniature tin tube, which is enough for two or three uses. The paste is white, relatively liquid, easy to apply. The declared thermal conductivity is 0.837 W/(m x K). Many people constantly use CSL 850 and speak about its good properties, better than those of KPT-8. However, these thermal pastes are very similar, and, most likely, their effectiveness is approximately on the same level. Like it or not, testing will show.

47 D90T8-010 GFC-M1

Before us is a dark ash-colored paste. No identifying marks, except for labeling, and the origin of the substance could not be determined.

The product in question was part of one of the kits for self-assembly of a laptop. But since she turned up under the arm - why not test it ?!

Coollaboratory Liquid Pro

This substance is the first serial thermal interface based on liquid metal. Those who had an interesting childhood must have beaten the thermometers behind the garages and rolled balls of mercury. So, this composition inspires nostalgia for past inventions and experiments with liquid metals. The substance has a characteristic shiny metallic color.

This alloy does not evaporate, is not as toxic as mercury, and does not form such dangerous compounds. This thermal interface consists of rare earth metals alloyed in a certain proportion. Its melting point is below room temperature. But that doesn't mean that you can do whatever you want with Liquid Pro. Like mercury, this metal reacts chemically with many other metals. So, oxide flakes grow on aluminum parts after a while, and they themselves literally decompose and dissolve at the point of contact (this behavior is typical for gallium). In this case, transmetallic compounds are formed. On copper, this process will also occur, but not so quickly and far from being so obvious.
Unfortunately, Liquid Pro is also very difficult to apply.
All attempts to smear liquid metal will be futile unless several conditions are met to ensure the desired effect. The contact surfaces of the chip and the cooler must be clean and smooth, copper must not have oxides. It is best to pre-treat the sole of the cooling device with sandpaper with a fine grain (zero), and then degrease with alcohol. The processor cover should also be degreased.
Prepare a cotton swab. Squeeze a small ball of Liquid Pro out of the syringe onto the surface, press the ball with a cotton swab. The metal will enter the wool fibers, and will be held there. Now you need to rub it into the surface with a little effort. If the surfaces are really clean, then the result will not be long in coming. Other methods, such as smearing with a brush or rag, rarely bring results. In most cases, you will roll the metal in the form of balls until they roll down somewhere, under the processor substrate or just onto the board's textolite (checked).
And when you rub cotton wool over the surface, you remove the thinnest oxide film from copper, which promotes adhesion.
It should be noted that Liquid Pro is a metal, and it is simply an excellent conductor of electricity. Neither the Arctic Silver 5, nor even any “silver” in this regard, can be compared with it at all. You need to handle this substance very carefully, because one small ball, imperceptibly rolled onto the contacts of a chip, can create a short circuit and permanently disable your entire system. If you work carefully and slowly, and follow the simplest recommendations, precautions, everything will be fine.
For Liquid Pro, the manufacturer specifies a thermal conductivity of more than 80 W/(m x K).

Test results

Depending on the data obtained, we divided all the samples into five categories, based on the level of thermal conductivity they demonstrated:

1) worst thermal conductivity (The Worst Thermal Conductivity)
Pastes that fall into this group are not recommended for use in a PC.

2) average thermal conductivity (Medium Thermal Conductivity)
This category includes relatively simple and inexpensive thermal pastes that are able to satisfy the needs of most users for whom a couple of "extra" degrees on the processor are not decisive.

3) good thermal conductivity (Good Thermal Conductivity)
Thermal interfaces are recommended for demanding users who prefer to use well-known brands of proven products. For this category, the exceptionally high quality and stability of the characteristics of the pastes are in the first place.

4) excellent thermal conductivity (Very Good Thermal Conductivity)
Samples of pastes that fall into this category have impressive characteristics and can be recommended to those who are seriously interested in overclocking or who want to reduce the temperature of the processor, graphics chip, memory in any way, even by a relatively small amount.

5) outstanding thermal conductivity (Outstanding Thermal Conductivity)– the highest, excellent performance among all thermal interfaces.
The substances presented in this category are a worthy choice for those who rightfully consider themselves a real enthusiast.

Worst thermal conductivity

Only a few pastas fell into the category of losers. They are the worst of what we have tested, but in comparison with various exotic alternatives to thermal interfaces, they do not look so bad and hopeless:

Frankly, we did not expect such a result, at least from the Titan product. It turns out that the "free" Nano blue turned out to be simply hopeless ... For the accuracy of the results obtained, this paste was tested several times and consistently showed the worst result.
It is up to each user to decide whether to use the two substances presented in the diagram, but there are a sufficient number of significantly better products on the market, which can often be found in a set of inexpensive central processor cooling systems or in a separate sale, and use them.

Average thermal conductivity

This group is the most numerous. Our standard, KPT-8, also fell into it. The paste as a whole shows satisfactory characteristics, however, it should be noted that it has practically no obvious competitors in its price range.
As it turned out, the viscosity and thermal conductivity of KPT-8 may vary slightly, depending on the specific sample and the place of its production. However, this affects the final result very, very little.
In our case, the difference between the available pastes was only 1°C, which is really very small.
I would like to say a few words about AlSil-3. They say that this paste has b about greater thermal conductivity than another product of the domestic chemical industry, and is positioned as a replacement for KPT-8. But as a result of the tests, there were no qualitative differences between AlSil-3 and good KPT-8, neither in terms of final thermal conductivity, nor in terms of ease of application and removal. The laboratory site is forced to state the fact that AlSil-3 cannot potentially compete with the "kpt-shka", since it has absolutely no advantages in performance over the latter. In addition, she has b about higher cost and less common, making the KPT-8 a better buy.

This test group included many foreign thermal pastes, which showed satisfactory characteristics and were on par with our standard, and in some cases were slightly better.
All of them are just normal "workhorses", which in no case should be thrown out of the kit of a brand new cooling system and immediately look for a replacement. These thermal interfaces are recommended for those who do not strive to set world records, but still moderately overclock their PC components.
Also, many metal-based pastes fell into the group of "middle peasants". The samples shown in the diagram do not justify the hopes placed on them (remember the declaration of the presence of silver in the composition of individual substances and the high declared thermal conductivity). They turn out to be no better than the high-quality "eight", but the contamination of everything and everything when working with such pastes is guaranteed to you.

Good thermal conductivity

As you know, Arctic Silver products are almost idolized on foreign sites, and in each test they respond with the most flattering words. Recently, there has been a total admiration of users for a new idol in the face - Arctic Silver-5 ...
We've done a detailed test to reveal the true benefits of this reputable brand's thermal pastes.

It turns out that Arctic Alumina is no better than the "Mendeleev" KPT-8.
Alumina entered the Good Thermal Conductivity group solely as a stable product High Quality.
Silver-based Arctic Silver 3 really wins 2 degrees from the standard.
Arctic Silver 5 gains already as much as 3 degrees, which is truly an achievement for thermal pastes in this series.
Everything would be fine… But Arctic Ceramique brings a little chaos to our ranks! It exhibits almost the same characteristics as Arctic Silver 5, but is much easier to apply. And this result is not a mistake, because tests carried out even after a few weeks showed the same result.
In this case, we can definitely say that Arctic Ceramique is a very good buy.
As for Arctic Silver 5, it sells effectively, partly due to the total faith of users in the power of noble silver;). It is one of the highest quality and most successful thermal pastes on the market. In addition, the product in question does not cause any difficulties during application and removal, and can be safely recommended to those users who do not seek to save a penny on the purchase of a thermal interface. This group also included some other pastes from lesser-known manufacturers, which for most ordinary users will not be an easy task to get.

Excellent thermal conductivity

First of all, we were pleased with Titan's TTG -G30010 thermal grease - it not only showed one of the best results in the test (even better than Arctic Silver 5), but also does not suffer from "childhood diseases" inherent in Nano Blue and Silver Grease. If, in addition to everything, we take into account its retail price, then we have a real killer not only for KPT-8, but also for many more effective thermal pastes, regardless of the price of the latter! Gigabyte's nondescript stuff from a syringe, as well as Apus-TMG 301 and Shin-Etsu MicroSi G-751, also showed impressive results, but they are less common than the aforementioned product from Titan, so it's not worth making much effort to find them at the point of sale. .

Outstanding thermal conductivity

The last group contains the best of the best - thermal pastes, which neither mass products nor other widely publicized and expensive substances could compete with.

We have only three past champions, and it is difficult to call the most outstanding of them pasta. Coollaboratory Liquid Pro is truly the best thermal interface available today. He demonstrated maximum efficiency and once again confirmed the glory, those legends that are already roaming the Internet about this product.
However, it has a number of significant drawbacks - a very high cost, difficulties in applying to contact surfaces, electrical conductivity, a relatively narrow geography of distribution (mainly large metropolitan cities). Those who care about every tenth of a degree on a processor, a video card chip, can easily turn a blind eye to all the shortcomings inherent in Liquid Pro, but more rational buyers should pay attention to Arctic Cooling's product - MX-1 thermal grease.
What the Swiss company does often demonstrates higher efficiency than competitors' products, and the thermal interface is no exception. If you look closely at its packaging, then on the back you can see a comparison table of MX-1 with common samples, including Arctic Silver 3. Some of us found it hard to believe that this paste could compete so well with more eminent competitors, but the testing puts everything in its place.
MX-1 showed a stable result from the first application - the separation from the standard was at least 5°C!
And what will happen after the indicated 200 hours, which are necessary for the paste to acquire the optimal condition? To do this, the cooler remained pressed on the stand for exactly 200 hours, every 24 hours the performance of the Swiss product was measured. Unfortunately, in the process of testing on a test stand, the paste only slightly improved its result - by a few tenths of a degree, which does not cause much enthusiasm. However, the obvious benefits of MX-1 are on the face!
The only annoyance associated with the Arctic Cooling product is the relative difficulty of applying it to the CPU cover and/or cooling system base. However, these disadvantages are much less than those of Coollaboratory Liquid Pro.
Shin-Etsu MicroSi MPU-3.7 also demonstrated a very good result, but there is one "but" - an ordinary user will probably not be able to find such a product. When searching for this substance, you can only hope for fortune, nothing more, so we advise you to pay more attention to other thermal interfaces provided on the Outstanding Thermal Conductivity substance diagram.

Bonus: test for substances that are not thermal interfaces

The natural interest of enthusiasts to search for adventures has crept into the field of cooling as well - many overclockers use (or at least try to use) non-standard and exotic substances instead of the usual pastes. Someone reports getting very high thermal conductivity, others simply use original substances to stand out from the crowd or avoid going to the market :) In any case, this trend exists. That is why we decided to check how successfully some popular and exotic substances can replace real thermal paste.
Here are the test results of those substances that were tested:

We think it would not be superfluous to comment on the figures obtained, because the harsh reality spoils the relatively rosy picture depicted in the diagram.

Mustard "Russian"
Yes, the temperature was set exactly at such a digitally interesting value - 66.6 ° C. However, if you wait a certain amount of time, it becomes clear that moisture is slowly evaporating from this sharp product, and a dry concentrate remains between the heat-distributing cover of the processor and the cooler base. After the test, it can be turned back into normal mustard by adding a little water :).
We hope that none of the sane readers will engage in such experiments on a home computer.

Nefras С4-155/200(White Spirit)
Solvent. With it, the sensor of the test stand at a certain moment recorded a relatively stable temperature of the heater in the region of 65.5 ° C. True, it is not entirely correct to indicate the resulting value here. The fact is that the temperature of the heater reached 65.5 ° C, and its growth slowed down, but over time, the readings of the digital thermometer gradually increased. The reason is simple - the solvent evaporated easily, and after a certain time, instead of a heat-conducting substance, we would get an air gap between the processor cover and the cooler sole.

Isopropyl alcohol
It turned out to be strange that the temperature when using this substance stopped at a value of 63 ° C (the solvent showed 65 ° C, and their physicochemical properties are very similar). However, after a while the temperature began to rise slowly. As expected, the alcohol evaporated.

Machine oil TP-22
It is used for lubrication of tape drives. We will try to use it as a thermal interface. Moreover, it is various types of machine oils that often use overclockers instead of the usual thermal pastes.
Given that this is an ordinary mineral oil, the result in terms of thermal conductivity turned out to be very modest and expected. True, this substance does not evaporate at such temperatures, and besides, it is a good insulator.
Bottom line: as a thermal interface for the TP-22 processor, it is no good.

"Hado"
Reminds Litol, but has a little the best performance; It is used to lubricate various mechanisms, reduces their friction and wear.
On overclockers.ru, many of us have probably read in which the author used Litol instead of paste in his computer.
63.6°C stable. The result is indeed better than that of mineral oil, but it does not even reach the level of the worst thermal pastes, and therefore it cannot be recommended for use in a PC.

Unrefined sunflower oil "Yamran" :)
Very interesting. The result was a very stable (albeit bad) result. Heater temperature - 62°С stable.

Petrol
62.5°C.
Gasoline evaporates, the temperature rises slowly, as is the case with alcohol.

Watch oil low-temperature MN-30
60.5°C stable. Already better. So we will soon catch up with the worst thermal pastes :)

Pilot Gold, gold colored marker
For application as a thermal interface, the impregnation contained in the inner fibrous "rod" was used. 57.5°C is a very good result, but since the marker dye is alcohol based, the test temperature is unstable and slowly rises as the substance evaporates.

Toothpaste "Pearl"
So, we are trying to fake the classic white thermal paste. They say that KPT-8 and AlSil 3 are replaced with cheap toothpaste. The comparison will show how true these beliefs are!
A clear smell of menthol, and the consistency is not the same. You would surely distinguish any toothpaste from CPT-8 :)
The test result is 55.5°C! We could not believe our eyes - this is a true KPT-8! In terms of efficiency - even a little better than our standard.
No, wait. Do not run to smear processors with toothpastes! The result in any case will be unstable, because any dentifrice contains water, and it evaporates over time, and the temperature of the heater slowly rises. The paste becomes a heat insulator, turning into a kind of chalk.

Drinking water
54°C.
Look, the water turned out to be 2 degrees better than our standard! The H20 is truly capable of doing wonders. Without water, there would be no life on our planet. True, one cannot escape the laws of physics: the eternal circulation of moisture in nature spoils everything ... Water evaporates and the temperature of the heater rises over time. Therefore, it cannot be used as a thermal interface. In addition, even when conducting tests “for the sake of sports interest” in a real computer, there is a risk of flooding the near-socket space, which can lead to a short circuit and failure of PC components.

Summing up the intermediate result, it should be noted that we got very interesting results. Nevertheless, in no case do not rush to change the standard thermal paste in your computer for toothpaste, or, worse, fill the processor cover with water! On special equipment protected from short circuits, and for short-term tests, we could afford it.
In addition, there is one important point: the vast majority of the substances discussed in this section contain alcohol or water (or are). When the heat sink is heated, they evaporate very intensively, which leads to complete “self-liquidation” of the used thermal interface!
Also, some substitutes may contain chemically active substances that cause corrosion of the sole of the cooler or water block! A vivid example confirming this is a proven toothpaste. At first, it demonstrates a result that is better than that of KPT-8, but after an hour or two during computer operation, the moisture contained in it evaporates almost completely, and the Pearl turns into a heat insulator! After removing the cooler from the processor after such a test, you will see that its copper sole is covered with a dark coating, which can only be removed by grinding. Therefore, in order to avoid trouble, in no case repeat tests like ours at home!

Conclusion

Testing is over - it's time to sum up. For greater clarity of the results obtained, the indicators of all thermal pastes are displayed on one summary diagram:

Absolute test leader, liquid metal based thermal interface Coollaboratory Liquid Pro, for outstanding performance indicators is awarded a badge of honor Certified Hardcore website:

Remembering a number of its features, which can easily be called shortcomings, we decided to note another thermal interface, Arctic Cooling MX-1, a similar award, Certified Hardcore website:

It is much more reminiscent of the usual thermal paste than "liquid metal", is widespread and has almost no drawbacks.
Product TITAN COMPUTER CO. Ltd., Nano Grease TTG-30030, due to mass availability, affordable cost and very high efficiency, deserved an award King of the Hill website:

Finally, it is necessary to focus on the fact that before you is a clear comparative test of many thermal interfaces using a single method, on a stable heat source under the same conditions.
We cannot speak with 100% certainty about the truth or objectivity of the results obtained, just as we cannot speak about the reliability of tests on real CPU. On many real systems, the results differ slightly due to the variability of conditions and the influence of many concomitant factors on the research process itself, so it is not possible to make an unambiguous and only correct conclusion.
Be that as it may, the results obtained clearly demonstrate the difference between individual thermal interfaces, and they should not be ignored. We have made every effort to present you with a subjective reflection of the objective truth about the effectiveness of various thermal pastes!

Readers should remember that in many respects the repetition of the results obtained on the test bench in the case of tests on the central processor will depend on the architecture of the latter, the features of the built-in thermal sensor, and, first of all, on the level of heat release. So, at TDP=35 W the difference between the best and worst pastes will be very small(one or two degrees) and only with an increase in this indicator (especially when overclocking powerful CPUs) will it manifest itself to the maximum.

We hope that now those who previously did not even imagine the existence of a substance called thermal paste in their computer have some reasons for reflection, supported by an appropriate test base.
True, it is not at all necessary to remove the cover immediately after reading this material. system block, dismantle the cooling system and change the substance that was originally smeared on the processor heat dissipator. It is necessary to soberly weigh all the pros and cons, and think about whether there is really a practical need for such a move. Those who operate their computer in nominal mode have nothing to worry about, even if the assembler used the worst thermal interface (although there are times when a seemingly qualified engineer from a reputable company not only does not apply thermal paste to the processor cover, but even forgets to remove the protective plastic film from the base of the cooling system!).
Those who are really worried about every extra degree on the processor, and / or are busy squeezing the last megahertz out of their favorite hardware, when looking for a specific thermal interface, they must first of all take into account their needs and the actual availability of any paste. Factors that will contribute to the purchase are the ease of application and rinsing, and, of course, the price.

I would also like to note that working with thermal interfaces is not dangerous if you only occasionally use these substances, and do not deal with them around the clock. When applying / removing pastes, it will not be superfluous to have alcohol and wipes on hand. It is advisable not to smear the thermal interface on the skin, because in some cases too much of a certain substance can cause an allergic reaction, but rejuvenation of the body is unlikely :)
Repeating the remark that has already become a classic, let’s finally say: if you have a desire to change the cooling system, you should first ask yourself: “... maybe, for a start, it would be better to just change the thermal interface?”.

Arctic Ceramique, Arctic Cooling MX-1 and Coollaboratory Liquid Pro thermal pastes are provided by the PCForsage online store.

Feedback, suggestions and comments on this material are accepted in the forum website.

ALL ORDERS IN THE STATUS "WAITING FOR PAYMENT" WILL BE AUTOMATICALLY CANCELED WITHOUT PRIOR NOTIFICATION AT THE END OF THE DAY.

In our online store, the price of goods indicated on the pages of the site is final.

The procedure for payment by electronic money, bank card, from a mobile account:

• After placing an order, your order will be placed in your Personal Area with status " Awaiting verification"
• Our managers will check the availability in the warehouse, and put the goods you have chosen in reserve. This changes the status of your order to " Paid". next to the status " Paid"link will be displayed" Pay", by clicking on which you will be redirected to the page for choosing payment methods of the Robokassa website.
• After choosing a method and making payment for the order, the status will automatically change to " Paid". Further, as soon as possible, the goods will be sent to you by the delivery method selected during the ordering process.

1. Cash payment

In cash, it is possible to pay for the goods you bought to the courier (delivering your goods), or in the store (for pickup). When paying in cash, you are given a sales receipt, cashier's check.

ATTENTION!!! Cash on delivery WE DO NOT WORK, so payment upon receipt postal parcel impossible!

2. Payment by bank transfer

For legal entities, we have provided the opportunity to pay for purchases using a bank transfer. In the process of placing an order, select the payment method cashless payment and enter the data for invoicing.

3. Payment via payment terminal

ROBOKASSA - allows you to accept payments from customers usingbank cards, at any electronic currency, using servicesmobile commerce(MTS, Megafon, Beeline), payments viaInternet bankleading banks of the Russian Federation, payments through ATMs, throughinstant payment terminals, as well as usingiPhone apps.

As you know, computers and laptops get very hot during long-term operation. Moreover, the most vulnerable parts are just the processor and video card. In this regard, all manufacturers are trying to provide the most maximum cooling for these two elements. Moreover, when assembling a processor, not a single conscientious manufacturer can do without the use of such a substance as thermal paste.

In general, this part is used not only on the conveyor in the manufacture of PC and laptop elements, but is also used without fail when repairing a processor for better thermal conductivity. And today we will talk about the characteristics of KPT-8 thermal paste and how to apply it correctly.

What it is?

Thermal grease is a multi-component element that is in a flexible plastic state. According to its properties, it serves as a good thermal conductor, therefore it is often used to reduce the temperature resistance between the touching surfaces of computer parts. In other words, KPT-8 thermal grease fills all the bumps in the heatsink and processor, thereby reducing their heating (by several degrees).

It is worth noting that outwardly this cavity may seem absolutely flat. But it's not. If you look closely (preferably through specialized tools), you can see many depressions and microcracks. During operation, they create some that interferes with the normal removal of heat. As a result, it fails. To prevent this from happening, all existing microcracks are sealed with a special KPT-8 thermal paste. The removal of heat to the radiator is the key task that this substance performs.

Thermal paste KPT-8: reviews and specifications

Outwardly, it is a homogeneous white substance with a thermal conductivity of the order of 0.7-0.95 W / m x K. At the same time, its specific level electrical resistance is 1014 ohms (sometimes this figure may be slightly higher than the specified norm). Working temperature, at which the thermal paste does not lose its properties - from minus 60 to plus 300 degrees Celsius. Thus, KPT-8 (reviews also note this point) provides excellent thermal conductivity under any load of the processor or video card.

About cost

It should be noted that this thermal paste can be sold both in a tube and in a small jar. As for the cost, a 10-gram jar of KPT-8 costs about 55 rubles. For a tube, you will have to pay about 2 times more, since its weight is 18 grams.

KPT-8 thermal paste is also sold in 20-kilogram buckets, but there is simply no point in purchasing such. For the processor, it is enough to use only a few grams of the substance.

How to properly apply KPT-8 on the processor surface? Nuances

We note right away that this does not require special skills and expensive tools. The main thing to remember is that the layer of thermal paste must be applied in such a way that the surface of the heatsink and the processor fit snugly against each other. In other words, the paste should only fill depressions and microcracks. The rest of the space should not be processed by it.

Where to start work?

First, we need to clean the surface of the chip from the remnants of the old thermal paste. You can do this with a regular soft cloth. Alternatively, you can pick up a jar of ethyl alcohol and treat the surface of the parts. In the latter case, it is better to use cotton swabs.

By the way, these are not the only examples of how you can clean the surface of the processor from the remnants of the old thermal paste. There is another effective method. It consists in the use of ordinary clerical gum. With it, you can effectively clean the old layer of paste from the metal surface of the element. True, for this you need to have very strong fingers. Despite the physical complexity this method, the rubber band method is the safest for the CPU itself and the heatsink.

Next, remove the remnants of thermal paste from the base of the cooler radiator. Everything, now the processor is completely freed from the old paste. But do not rush to apply a new one on it. First you need to rub a "zero" layer of thermal paste into these two surfaces (that is, remove it after preliminary application). So the substance will remain only in the places we need - depressions and microcracks.

Now we move on to the main stage - application. Gently squeeze a little thermal paste from the tube onto the main chip. Many computer scientists recommend applying it diagonally crosswise. In this way, you will achieve the best lubrication of the elements around the entire perimeter of the surfaces. Remember that the layer of thermal paste must be very even and, most importantly, thin (even slightly translucent). It is very easy to distribute the product evenly over the metal surface - just swipe a plastic card over the elements. You should not make a centimeter layer between the processor and the heatsink, using the principle "the more the better". All this will only aggravate the quality of conductivity and heat removal from the processor to the heatsink. As a result, you will simply overheat and burn an expensive element that will no longer be subject to any recovery.

The final stage

After the layer of thermal paste is evenly distributed over the surface, and the excess is removed, you can safely fix the heatsink over the processor and press it with fastening locks. That's all, this is the whole repair is over. As you can see, everything is very simple and not too complicated (except for the hassle of removing the old layer of paste). And most importantly - now your computer will be less exposed to various overheating and certainly will not fail due to unnecessary overload.

So, we found out what KPT-8 thermal paste is and how to properly apply it to the metal surface of the processor and heatsink. Remember that overheating the processor can certainly lead to its failure. Take care of your computer and apply thermal paste correctly.

Alpenfohn

AMA

cooler master

CoolerMaster ThermalFusion 400 (RG-TF4-TGU1-GP)

Cool IT

deep cool

Ice Hammer

Nexus

Prolimatech PK-1

Scythe

SilMORE

Thermalright Chill Factor and Chill Factor II

Chill Factor Chill Factor II

Thermaltake

ThermoLab

Titan Nano Blue and Nano Grease

Zalman ZM-TG2

Zaward

Test results and conclusions

Hello Dear friends, with you Artyom.

Let's talk about the effectiveness of thermal paste for cooling the CPU.

There will not be many thermal pastes in today's test, but I will definitely test the popular KPT-8.

In addition to it, the test will include Zalman ZM-STG2 and Cooler Master E2.

Thermal pastes have quite a few different properties. Starting from the price, the size of the container (bottle, syringe) and ending with the consistency.

Depending on the consistency, some thermal paste needs to be applied a little more, and some a little less.

This fact may also affect the ability to evenly distribute the paste on the processor cover and cooler heat sink (provided that the processor cover and cooler base are perfectly even).

However, the most important thing in any thermal paste is the thermal conductivity. This parameter is measured in W / m * K (Watt per meter times Kelvin).

The higher this number, the more efficiently the thermal paste will be able to remove heat from the source, in our case, the processor.

Zalman ZM-STG2 has a thermal conductivity of 4.1 W / m * K, Cooler Master E2 3.5 W / m * K, and KPT-8 0.7-0.8 W / m * K.

For example, the top thermal paste Arctic Cooling MX4 has a record thermal conductivity of 8.5 W/m*K.

So it will be doubly interesting how KPT-8 will behave with a record low level of thermal conductivity of 0.7-0.8 W/m*K.

P. S. All thermal pastes were applied using the same technique.

Thin horizontal line on the processor, and the cooler substrate itself distributed the paste.

Thus, equal test conditions are achieved for all thermal pastes.

The first thermal paste temperature test will be done using the Deepcool Assassin 2 tower cooler.

Deepcool Assassin 2 is a massive and highly efficient dual-fan tower with eight heatpipes on board.

In the second test, I will use a noticeably simpler and smaller cooler. This is a compact tower from .

How can the effectiveness of thermal pastes change with coolers of different levels?

The test bench consists of:

CPU: Intel Core i5 2500K overclocked to 4.4 GHz.

RAM: Kingston HyperX DDR3 1866MHz (KHX1866C9D3K2/4GX).

Motherboard: Gigabyte GA-Z68-D3H-B3

Storage device: HDD WD Blue 1 TB (WD10EZEX).

Power Supply: FSP SPI600 600W.

As a load on the Core i5 2500K processor, I used the Linpack profile OCCT program with AVX instructions enabled.

Let me remind you once again that the processor was overclocked to 4.4 GHz.

P. S. Unfortunately, the processor ran stably at a frequency of only 4.4 GHz.

Most likely, further overclocking was hindered by the power supply system of the motherboard, which does not have a normal radiator for cooling.

The test itself lasted 12 minutes, and 1 minute at the beginning and at the end was allotted for fixing the minimum temperatures on the processor cores.

The temperature in the room was at the level of 21 degrees Celsius.

I showed several screenshots with temperatures, then there will be ready-made temperature graphs.

Zalman ZM-STG2 turns out to be the most effective thermal paste, although the rest of the pastes are only a couple of degrees behind.

I was most surprised by KPT-8, which shows results similar to other pastes, and this despite the extremely low thermal conductivity of 0.7-0.8 W/m*K.

However, final conclusions are yet to be drawn.

Now it's time for a smaller and simpler cooler - this is the Arctic Cooling Freezer 33.

When using processor coolers with a lower coefficient of heat dissipation, the efficiency of the KPT-8 drops significantly.

The difference in this case can be up to 13 degrees Celsius compared to the same Zalman ZM-STG2.

The temperatures for the two cores of the processor are already beyond the comfortable 80 degrees Celsius.

Perhaps further overclocking of the processor will further reduce the efficiency of the KPT-8 when using the Arctic Cooling Freezer 33.

A similar effect could also be seen with the Deepcool Assassin 2 cooler.

However, I cannot verify this, but I can only assume with some certain level of certainty.

Thus, the effectiveness of thermal paste depends not only on the level of thermal conductivity, but also on the CPU cooler itself.

The more massive the cooler, the more it is possible to neutralize the low thermal conductivity of the paste. Of course, this effect also has its own certain level and limit.

Again, it should be understood that the effectiveness of thermal paste will also depend on the TDP level of the processor and on the area where thermal paste is applied.

It would be doubly interesting to see the effectiveness of thermal pastes on processors with a large heat-distributing cover.

These are, for example, chips from Intel for the LGA2066 connector or huge AMD Threadripper. Perhaps in the future I will have such an opportunity.

Leave your comments and suggestions for future thermal paste tests. What thermal paste do you prefer to use?

Also, do not forget to join the Vkontakte group and subscribe to the YouTube channel.