<!–#set var="article_header" value="Can't Touch This! –
A Comparison of 46 CPU Coolers” –>
Cooler Trends – Complex Design & Precious Materials
There’s been a lot of water under the bridge since our last CPU cooler test. In this comparison, we’ve rounded up no fewer than 46 coolers (of which 23 are brand-new models), which all strive for an optimum in cooling performance and noise level.
There are basically two ways to maximize a cooler’s performance – you can increase its dimensions, with the “more is more” approach; or you can turn to an entirely new design. The first method is comparable to increasing a car’s performance by increasing its cubic capacity. The second method, on the other hand, requires using technical tricks to achieve better cooling results. We have numerous test candidates that have followed this path, offering some tasty construction and design tidbits. The material used in the heat sink is crucial, since an ideal material drastically improves cooling performance by offering the best thermal conductivity. Silver, copper and gold offer ideal properties, providing very high thermal conductivity. With aluminum, a compromise can be reached – it is very inexpensive to produce, but doesn’t offer ideal thermal conductivity.
Although the Zalman coolers look really slick, their thermal performance is disappointing.
Cooler Trends – Complex Design & Precious Materials, Continued
A global view of things reveals that a development is taking place away from the simple fan design of previous years. Technically high-quality models in this test rely on a copper core for their heat sink and are surrounded by thin fins. This core is complemented by a fan with ball bearings and a rotor designed to enhance airflow. The CPU contact surface has also been designed to enhance heat transfer. The ergonomic design of the coolers is also important – users should be able to install them easily, without needing a lot of elbow grease.
AMD’s new Athlon processors above 1200 MHz practically beg for a powerful cooling system. Using a cheap generic cooler is just not worthwhile in that scenario. Also, overclocking fans who want to eke the last ounce of performance out of their system won’t be able to get around buying a top cooler.
Complex construction – the Molex models are made up of many separate parts.
In a nutshell – this comparison shows a wider selection of coolers than the world has probably ever seen. The prices and features of all the coolers vary dramatically from one model to the next. In our last test, the Silverado from Noise Control emerged as the victor – and the result? Delivery times shot up to between four and five weeks and the manufacturer changed its business strategy.
Theory: Ideal Cooling
This table shows a comparison of the thermal conductivity of different metals.
The basic principle behind a CPU cooler can be reduced to a single mathematical equation which describes the heat flow of two media while ignoring external factors:
Iw = ??x (T1-T2) x (l/A)
A brief explanation is in order – Iw describes the heat flow created between two different bodies (materials) the moment a temperature difference (T1 and T2) exists. I describes the length and/or the distance between the two bodies, while A describes the surface (cooling or radiating surface). Gamma describes the thermal conductivity coefficient. A quick glance at this equation reveals which qualities an almost ideally designed cooler must have. The temperature difference (T1-T2) is the factor that drives heat flow. In addition, heat flow is substantially affected by the material’s thermal conductivity (gamma) and the radiating surface. That should explain the fundamentals.
Cooler Master EP5-6151: Plain Jane Cooler for Undemanding Tastes
Cooler Master EP5-6151.
The Cooler Master EP5-6151 is not so much designed to be a cooler for overclocking fanatics as it is an add-on for undemanding standard PC systems. The basic thought behind this cooler is that “a lot helps a lot.” Its classic design, combining a large aluminum heat sink with a cheap fan, and its average cooling performance put this model smack dab in the middle of the testing field.
Cooler Master DP5-6H51: Small Heat Sink – Large Fan
Cooler Master DP5-6H51.
Cooler Master DP5-6H51.
The DP5-6H51 is another model produced by Cooler Master. Its high-end Pabst fan really stands out with its ability not only to produce a strong airstream, but also to generate an attention getting 50 dB(A) of noise. The manufacturer combines the large fan with a smaller aluminum profile, which leads to the same cooling effect as was the case with the EP5 model.
Cooler Master DP5-6G11: Flatliner
Cooler Master DP5-6G11.
Cooler Master DP5-6G11.
The third Cooler Master model belongs to the “very inexpensive” category and, in our humble opinion, only lends itself to use with AMD Duron or Intel Celeron processors. The cooler is a rather average performer, only attaining 45 degrees, which is primarily generated by the combination of a high fan speed and a small heat sink.
Blizzard Thunderbird: Shooting Star
Blizzard Thunderbird.
Blizzard Thunderbird with a heat sink made of solid copper.
The Blizzard Thunderbird’s test caused quite a stir. Of the 25 new coolers we tested, this one had the best cooling performance. In the overall placement, it placed third among a total of 46 coolers tested. A very good performance! A goodly chunk of the high cooling performance can be attributed to the heat sink of pure copper, which, however, also makes for a high overall weight. Other peculiarities: the fan rotates at an unusually high 4900 RPM. The soft CPU clamps make installation child’s play.
TRIG T40-6EA: Powerful Fence-Straddler
Trig T40-6EA.
Trig T40-6EA.
The TRIG cooler’s roomy design stands out immediately. It is made up of a solid aluminum heat sink and a large fan. The polished CPU contact surface is a positive feature. This giant’s cooling performance of 41 degrees is impressive, although, because of its mammoth measurements, it can’t be mounted on all boards. Its operating noise level is a conspicuously loud 60dB (A), which should bring your local safety representative onto the scene. During the last test, we criticized the clamps for being much too stiff. The manufacturer has responded to our criticism, and the new model can be installed much more easily.
TRIG T40-6EAC: XXL and Case Fan
Trig T40-6EAC.
Trig T40-6EAC with an inlaid copper core.
This cooler is unique in one crucial area – the CPU contact surface consists of an inlaid copper core and a heat sink made of aluminum. A case fan operating at 3000 RPM is attached to the heat sink by means of springs. The extra-large cooler is a good performer, with a cooling performance of 40 degrees.
Global Win WBK 38: Behind Bars
Global Win WBK 38.
The main attraction of this model is the new high-speed Pabst fan, which operates at 6000 RPM and has protective bars for a good reason. Not only that, but the cooler is attached to the heat sink by means of a clasp, which means that it is securely fastened. The cooler used quite a lot of electricity during the test – for that very reason, the package includes an additional voltage adapter for a 5.25″ power outlet. Its cooling performance of 40 degrees is good, which puts the cooler in the lead.
Global Win WBK 68: Wacky Little Brother
Global Win WBK 38.
Global Win WBK 38.
The small model from Global Win is based on the same heat sink that the WBK 38 uses, although the former’s fan is smaller and spins much more slowly. This cooler gave up the ghost towards the end of the testing procedure, which doesn’t really say a lot for the quality of its manufacturing. Its cooling performance of 44 degrees is substantially worse than that of its big brother WBK 38.
Global Win CAK 38: Top Model with High Cooling Performance
Global Win CAK 38 with a heat sink made of pure copper and a Pabst fan.
Global Win CAK 38 with a heat sink made of pure copper and a Pabst fan.
The CAK 38 is clearly one of Global Win’s high-end models. The manufacturer has come up with a good combination: a high-speed fan that rotates at 6000 RPM and a heat sink made of pure copper ensure that cooling performance is a very good 38 degrees. However, this high performance comes with a price – a high sound level of 59 dB. This is sure to set the cooler up for some run-ins with trade unions and professional associations. Weighing in at 488 grams, this is the heaviest cooler among the newcomers.
Elan Vital FSCUG9C: Ribbed Structure Made of Sheet Steel
Elan Vital FSCUG9C.
This edge on the Elan Vital FSCUG9C prevents use with Pentium III processors.
Compared to the FSCUG3C we examined during the first test, the dimensions of this cooler have grown a little. However, it doesn’t fit on an Intel Pentium III, since one of the two embedded copper runners gets in the way during assembly. Cooling performance measured during the test is good, averaging at 43 degrees. The heat sink consists of a ribbed structure made of sheet steel that is attached to a large copper plate acting as the contact surface.
Elan Vital FSCUG9C-6FC: Unsuccessful Thermal Trigger
Elan Vital FSCUG9C-6FC with an integrated, temperature-triggered fan control.
Elan Vital FSCUG9C-6FC with a solid copper plate as the CPU contact surface.
Compared to the simpler FSCUG9C, the Elan Vital FSCUG9C-6FC has an additional thermal trigger system. Its heat sink is the same as the one in the previous model, but a thermal sensor has been integrated into the copper contact surface (copper plate) in order to measure the CPU temperature and adjust the fan speed accordingly. However, the cooling performance of 51 degrees lags quite far behind expectations. The manufacturer needs to re-work this cooler. Our take on this cooler is that the technical approach is good, but it doesn’t deliver the goods in practice. In terms of noise level, this Taiwanese product is relatively unobtrusive, in line with the saying, “When you’re cold, you’re cold. When you’re not, you’re – quiet!”
SPRING SPCT381 BAF-1B: Black and Quiet
Spring SPCT381-BAF-1B.
Spring SPCT381-BAF-1B.
The Spring cooler stands out with a very low weight of 168 grams and a black aluminum heat sink. The CPU clamps have been designed admirably, incorporating three gaps on both sides. This allows for a corner to break off on the CPU Socket without wrecking the motherboard or making it necessary to do some dangerous drilling. The Spring’s cooling performance doesn’t exactly shine with an attained temperature of 44 degrees. Be that as it may, an operating noise level of 43 dB(A) is a positive feature. In summary, we can say that this cooler isn’t a bad pick for OEM manufacturers who are not necessarily interested in peak cooling performance. Your local industrial health and safety representative would also be quite pleased with this cooler, which can be installed in office PCs without worries.
SPRING SPCT801 AAF-1B: Son of Flatliner
Spring SPCT801-AAF-1B.
Spring’s flatter model uses the same CPU clamps as the previous cooler. The heat sink is made of aluminum and has a larger surface base. However, the small Spring’s performance is considerably flatter than that of the large model – cooling performance 45 degrees at best.
SPACEWALKER: A Nobody in the Crowd
Spacewalker.
Spacewalker.
The well-known motherboard manufacturer Shuttle is behind the name of this fan. A real plus is the CPU clamp, which is produced by the Taiwanese manufacturer Taisol, who also has some models in this test. The very large aluminum heat sink with its generous dimensions catches the eye immediately. Interestingly, though, the Spacewalker cooler is still remarkably light at 217 grams. As far as maximum cooling performance goes, we need only say that we measured a temperature of 42 degrees in our test system under laboratory conditions.
AVC Mega-Cool: Compact and Powerful
AVC Mega-Cool.
AVC Mega-Cool.
Unique design parameters allow this cooler to stand out. For example, the heat sink consists of 40 separate cooling fins that create a large surface area to dissipate heat better. This allows the cooling temperature to reach an impressive 40 degrees. The fan used by the manufacturer is pretty much standard issue, such as can be found on numerous rival products. One noteworthy factory-standard feature is the dollop of thermal paste on the CPU contact surface, topped off with a protective cover made of transparent plastic.
Taisol CGK 744092: Unergonomic
Taisol CGK 744092.
Taisol CGK 744092.
This cooler has been used by AMD in innumerable reference systems. Just looking at it, you wouldn’t think that this conventionally designed cooler would perform that well. The cooler is attached to the aluminum heat sink with a clip. You will need some help to install this cooler, since the clip is much too tight. In a word, this cooler has earned the rating “unergonomic.”
Smart Cooler FSM 1168T: Buyer Beware!
Smart Cooler.
Smart Cooler.
One glance at the Smart Cooler doesn’t exactly leave an impression of solidity. The heat sink consists of separate fins, which are attached to the copper plate. By no means does this guarantee an ideal heat transfer scenario between the aluminum fins and the contact plate. The cooling performance hardly makes a blip on the screen – during the test, we measured a temperature of 56 degrees at 1000 MHz (AMD Athlon). The manufacturer has released the fan for use with AMD Athlon processors at 1200 MHz, a claim we consider to be more than utopian. Even its visually appealing design can’t hide its defective implementation. Our advice to the manufacturer – try your hardest not to offer this cooler on the market and instead, follow up with a new model with better properties!
Molex 371650008: Avant-garde Design
Molex 371650008.
Molex 371650008.
The Molex’s innovative design, dominated by the radially arranged aluminum fins, is a real eye-catcher. A solid copper core forms the middle around which the fins are arranged. Its cooling performance of 41 degrees is very good; its low weight of 205 grams doesn’t create any problems. Attention to detail – the CPU mounting is ergonomically designed, with a plastic clip that makes installing the mounting very easy. And that’s not all! This cooler is one of the few we tested that were manufactured in the USA – or at least that’s what the package says.
Molex 371660002 u. 371610002: Little Brother
Molex 371610002.
Molex 371660002.
In contrast to the large Molex cooler, the other two we tested weren’t quite as efficient. Their cooling performance was considerably weaker. The cooler with the green CPU clamp (Molex 371660002) has a green, thermally conductive fillet on the CPU contact surface, while the other cooler has a small, flexible metal plate on the contact surface. On the upside, the operating noise level is minimal.
Zalman CCC Project: Much Ado about Nothing
Zalman CCC Project.
Zalman CCC Project.
Of all the coolers tested, this baby is especially eye-catching. A solid copper core surrounded by 60 copper fins arranged in a star puts in quite an appearance. But appearances can be deceiving – during our practical test, this cooler attained a temperature of 44 degrees. This cooler is difficult to install – the entire motherboard has to be removed no matter what you do – and has a disappointingly poor cooling performance. Attaching the processor is a complicated process involving screwing it down to the motherboard. There’s no way around this, since the cooler weighs in at a ponderous 467 grams.
Zalman CNPS2005 (Aluminum): Below Expectations
Zalman CNPS2005.
Zalman CNPS2005.
The unconventionally designed Zalman attracts attention prima facie. It is made up of a heat sink topped with 112 fins arranged in the shape of a fan and a contact surface also made of aluminum. However, the passive heat sink only works in conjunction with a fan that has to be installed in addition (included in the package). Although this cooler seems, on the face of it, to promise passive and thus silent cooling, flipping the switch on the active fan proves the opposite. The noise level is a relatively high 56 dB(A). Installing the fan is anything but easy – only after three attachment screws for the components have been removed can it be affixed. Its cooling performance is below the expectations created by the design – 46 degrees are really nothing to write home about!
Zalman CNPS3000 (Copper): Better than Aluminum
Zalman CNPS3000.
Zalman CNPS3000.
As a direct comparison to the CNPS2005, which has the same design, only in aluminum, the heat sink on the CNPS3000 is made of pure copper. This improves the cooling performance somewhat to an attained temperature of 46 degrees.
Zalman CNPS3100G (Gold): The Crowning Glory in Gold
Zalman CNPS3100G.
Zalman CNPS3100G.
It’s almost unbelievable! The rare precious metal gold has never before been used in a heat sink. The Zalman CNPS3100G consists of a heat sink with a 0.3 mm thick gold coating covering the pure copper core. This allows it to obtain the best cooling scores of all the Zalman coolers we tested (44 degrees), even though, objectively speaking, 44 degrees is not fantastic.
Zalman CNPS3100 (Cu-Au Alloy): Visual Blarney
Zalman CNPS3100.
Zalman CNPS3100.
In contrast to the other Zalman coolers, the CNPS3100 is manufactured using a gold-copper alloy. This cooler didn’t turn out to be the dream system during our practical tests. Its generated temperature of 45 degrees is on a par with multitudinous standard coolers that are substantially cheaper to produce. One thing that all the Zalman coolers have in common is a chintzy-looking CPU clamp, which proves to be not very durable. Not only that, but an incorrectly designed spring keeps this cooler from pressing down firmly enough. To add insult to injury, the burrs on the CPU clamp weren’t shaved off properly. The result? Very sharp edges and the inevitable injury.
Zalman Coolers: Poor Manufacturing Quality
Zalman fans for all passive heat sinks.
Installing the fan.
Installing the fan.
We were somewhat surprised to discover that Zalman, a company based in South Korea, uses a PC case fan from the People’s Republic of China. This continues the claim of passive cooling ad absurdum. The fan is included with the models sporting fan-shaped heat sinks.
Laboratory Tests Up Close & Personal – CPUs Maxed at 100%
We had already noticed during our last comparative tests that not many cooler models were a good compromise between the necessary cooling performance and an acceptable noise level. In fact, our readings showed that the majority of the coolers don’t cool optimally. The testing platform consisted of an AMD reference board (AMD EVTS-BX-004) that was furnished with an AMD-750 chipset from the same manufacturer. All our readings were based on an AMD Athlon 1000 that worked with a 100 MHz front side bus. We used Windows 98 SE as an operating system platform. In order to keep CPU capacity use at 100% as much as possible, we ran the well-known program Seti, which processes data packets from the Internet. At the same time, we ran Flask Mpeg to create an MPEG-4 film and tested its playback. This ensured that the processor was constantly running at maximum capacity, which in turn ensured that heat flow was maximized.
Cooling Temperature
We carried out all the measurements in our Munich laboratory at an ambient temperature of exactly 24°C, two degrees higher than during the last test as a consequence of the warmer season. We calibrated the results in order to ensure comparability with the last measurements at 22°C.
Noise Level
We carried out our sound measurements in our newly conditioned sound measurement box, which simulates an almost completely soundproof environment.
Fan Speed
We measured the fan speeds of the new 23 coolers using a stroboscope. Each rotor was marked in order to facilitate the measuring of the speed.
Weight
The coolers were weighed using a precision weighing machine as is used in chemical laboratories to weigh materials.
Cooling Temperature: Range from 30°C to 56°C
The Swiftech remains the undisputed champion as the cooler with the best cooling performance.
The most essential criterion to consider when buying a processor cooler is the temperature produced on the CPU surface. The coolers from our last comparative test weren’t shown up by the 23 new test candidates. The Swiftech MC462 continues to be la crиme de la crиme, cooling down the CPU surface of an AMD Athlon 1000 to a minimal 30°C. Of the new coolers tested, the Blizzard Thunderbird demonstrated very good cooling capabilities, producing a temperature of 37°C on the CPU surface. This puts it on a level with the exemplary Silverado from Noise Control, whose innovative technology set the standard during the last test.
In general, you can say that all models that cool down to under 40°C are very good candidates for extreme overclocking of AMD Athlon/Duron and Intel Pentium III/Celeron CPUs.
The difference between these separate coolers and the standard el cheapo solution – the no-name serial cooler – is telling, as the chart indicates. The no-name cooler produces a temperature of 54°C, whereas the leader of the pack – the Swiftech MC462 – manages to produce a sensational 30°C. This fact alone reflects the vast range in cooling capacity among the models – a range largely invisible to the naked eye.
Noise Level – Is that a Construction Site or an Office?
The Smart Cooler’s score of 37 dB (A) is by no means positive – it is the cooler with the worst cooling performance of all 46 coolers tested.
While the Silverado from Noise Control was the quietest cooler in the last group tested, the newcomers have changed the rankings somewhat, even if the change is of no practical consequence. Although the Smart Cooler can serve up an operating noise level of a mere 37 dB (A), it still scores last in cooling performance among a total of 46 coolers. This allows the Silverado from Noise Control to hang onto the lead, since it offers very high cooling performance in addition to low noise. The Silverado’s design with its encapsulated construction remains in a class by itself. Among all the Johnnys-come-lately, there’s not a comparable cooler that’s even close to the Silverado. Although the Molex cooler’s design is relatively innovative and it can serve up a low noise level, its scores in the more crucial area of cooling performance are rather disappointing.
In general, a background noise rating between 35 dB(A) and 38 dB (A) is a great result, considering that the majority of the coolers tested produces 50 dB(A) of sound.
There is a scattering of new additions to the group of the noisiest coolers – the TRIG T40-6EA and the Global Win WBK 38 generate a noise level of 60dB(A), which puts them in direct violation of official guidelines governing health and safety in the workplace. This high noise level can be slightly reduced by using an appropriate case. According to readings we took of a standard PC case, however, the level doesn’t drop below 55 dB(A). This makes the coolers too loud to install in a PC at the office. In general, one thing can be said about the noise level – each person’s perception of noise level is subjective and hence different from the next person’s. Be that as it may, noise levels above the 50 dB(A) limit are always disruptive, no matter what range of frequencies create them. Noise levels below 45 dB(A) are generally acceptable (depending on the frequency) and are perceived as less disruptive by the majority of people.
A Real Heavyweight – Swiftech’s 789 Grams!
The Swiftech MC462 continues to be the heaviest cooler tested. It weighs in at 789 grams, making it seven times heavier than the cheap serial fan.
The correlation between a cooler’s weight and its cooling capacity and efficiency is a direct result of the materials used to manufacture it. While gold, silver and copper are remarkable heat conductors, they are also very dense: hence the coolers’ impressive total weight. Furnished with a solid copper plate, the Swiftech MC462 weighs in at almost 800 grams! AMD’s guidelines only allow for a total mass of 300 grams maximum, however, if the cooler is to be affixed to the socket using metal clamps. Among the new coolers we tested, there were some models that weigh in at well over 400 grams. The heaviest of these is the Global Win CAK 38 with a solid copper heat sink, weighing in at 500 grams. And even the Zalman CCC-Project and the Blizzard Thunderbird are anything but light. The Global Win CAK 38 draws some flak for its classic CPU clamp; it really should have special screws to attach it to the motherboard. You have no choice if transporting your PC – you have to remove the fan. Leaving it attached could seriously damage the CPU.
Specific Cooling Performance – Swiftech Duo Still in the Lead
Specific cooling performance C/W is a gauge for the quality of a cooler.
Specific cooling performance denotes the quality of a cooler – the smaller the value is, the better the cooling effect is. In general, specific cooling performance scores are related to the temperature readings. The winner in previous tests, the Swiftech MC462, attains a C/W score of 0.16, putting it once again squarely at the top of the pack. This positions it a considerable distance ahead of its brother, the Swiftech MC370-0A, which was la crиme de la crиme in the first group tested. Of the newcomer coolers, the Blizzard Thunderbird has entered very high on the performance totem pole. Its C/W of 0.31 puts it on a par with the excellent Silverado from Noise Control. The Global Win CAK 38’s cooling performance, a C/W of 0.33, is equally good. In general, we can say that
If specific cooling performance is somewhere between 0.35 and 0.4, this indicates a really decent cooler-fan system. Scores below 0.35 are very good; below 0.20, they’re top-notch.
The upper echelon still includes the two Swiftech models, which placed first and second. The MC462 was designed for use with the new Pentium 4, which has to meet entirely new requirements for thermal behavior.
The scores of the last models in this chart put them entirely out of the question – the Smart Cooler is the new laggard with a C/W score of 0.69. If it comes to this, then it’s hardly even worthwhile installing a cooler. All coolers with C/W scores above 0.57 are fully unsuitable for both overclocking and for AMD Athlon CPUs faster than 1000 MHz. At this point, we would like to mention that all scores were obtained using our reference configuration. Although using these coolers in another system could lead to slightly different results, these deviations shouldn’t be significant.
Fan Speed – 2850 RPM to 6700 RPM
High fan speeds automatically produce a higher noise level.
The chart above lists the fan speeds for all the coolers. The bars lined with yellow indicate the 25 new cooler models. The fans on the Swiftech MC370-0A and the Global Win FNP50 continue to spin the fastest at 6700 RPM. As we did during the last test, we measured fan speed using a stroboscope and checked these readings against readings we took using a sensor on the motherboard. However, this second method of measuring fan speed is often somewhat imprecise, resulting in discrepancies of at least 10 percent. The passively designed Zalman coolers, which only cool tolerably well when combined with the active fan, have the lowest fan speeds. The victor from the last test, Noise Control’s Silverado, continues to have a relatively low fan speed of 3300 RPM. The two Global Win coolers (CAK 38 and WBK 38) rely on the new Pabst fans that are designed to burn some serious rubber at 6000 RPM. The majority of the remaining coolers tested have an average fan speed of 4000 RPM.
Measurement procedure – We used a stroboscope to measure the fan speeds of all the fans.
Summary – A Comparison of all 46 Coolers
Fan and Heat Sink | CPU Platform | Heat Sink Material |
Heat Sink Profile | CPU Contact Surface |
Fan type | Weight in Grams (meas- ured) |
Revol- utions per minute (meas- ured) |
Temper- ature (meas- ured) |
Noise level (meas- ured) |
3DfxCool Socket-AHO-Fan | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Alum- inum |
Tech Fan | 238 g | 4400 | 39 °C | 55 dB |
Alpha PAL6035MFC | Socket 370 Socket 462 Socket 7 |
Steel | Grid structure | Copper | Sanyo Denki | 299 g | 3700 | 40 °C | 46 dB |
Alpha PEP66T | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Copper | Sanyo Denki | 373 g | 3800 | 39 °C | 49 dB |
AVC Mega-Cool | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Alum- inum |
no name | 147g | 4850 | 40 °C | 53 dB |
Blizzard Thunderbird | Socket 370 Socket 462 Socket 7 |
Copper | Grid structure | Copper | no name | 469g | 4900 | 37 °C | 52 dB |
Cooler Master DP5-6G11 | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Alum- inum |
no name | 143g | 4950 | 45 °C | 50 dB |
Cooler Master DP5-6H11 | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Alum- inum |
no name | 163 g | 4690 | 45 °C | 48 dB |
Cooler Master DP5-6H51 | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Alum- inum |
no name | 186 g | 4650 | 42 °C | 51 dB |
Cooler Master DP5-6H51 | Socket 37 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Alum- inum |
Pabst | 176g | 4250 | 45 °C | 48 dB |
Cooler Master EP5-6151 | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Alum- inum |
no name | 247g | 4850 | 45 °C | 50 dB |
EKL 20704001006 | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Alum- inum |
no name | 240 g | 4800 | 44 °C | 51 dB |
EKL 20704001007 | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Alum- inum |
no name | 138 g | 5700 | 51 °C | 48 dB |
Elan Vital FSCUG9C | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Copper | no name | 249g | 5500 | 43 °C | 48 dB |
Elan Vital FSCUG9C FSCUG 3 | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Copper | no name | 178 g | 5100 | 43 °C | 44 dB |
Elan Vital FSCUG9C-6FC | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Copper | no name | 268g | 3300 | 51 °C | 42 dB |
Global Win CAK 38 | Socket 370 Socket 462 Socket 7 |
Copper | Grid structure | Copper | Global Win | 488g | 6000 | 38 °C | 59 dB |
Global Win FNP50 | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Alum- inum |
no name | 154 g | 6700 | 41 °C | 50 dB |
Global Win FOP32-I | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Alum- inum |
no name | 220 g | 4600 | 38 °C | 52 dB |
Global Win WBK 38 | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Alum- inum |
Global Win | 289g | 6000 | 40 °C | 60 dB |
Global Win WBK 68 | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Alum- inum |
Global Win | 231g | 4600 | 44 °C | 50 dB |
Intel (Boxed Socket 370) | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Alum- inum |
no name | 179 g | 4000 | 41 °C | 46 dB |
Kanie Hedgehog-238M | Socket 370 Socket 462 Socket 7 |
Copper | Grid structure | Copper | Pabst | 475 g | no sensor | 45 °C | 55 dB |
Molex 371610002 | Socket 370 Socket 462 Socket 7 |
Copper, Alum- inum |
Ribbed structure | Copper | no name | 136g | 4500 | 48 °C | 39 dB |
Molex 371650008 | Socket 370 Socket 462 Socket 7 |
Copper | Ribbed structure | Copper | no name | 205g | 4300 | 41 °C | 46 dB |
Molex 371660002 | Socket 370 Socket 462 Socket 7 |
Copper, Alum- inum |
Ribbed structure | Copper | no name | 135g | 4500 | 48 °C | 39 dB |
NoiseControl Silverado | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Silber | Noise Control | 416 g | 3300 | 37 °C | 38 dB |
P Alpha | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Alum- inum |
no name | 228 g | 5000 | 46 °C | 56 dB |
Serienlьfter (No name) | Socket 370 Socket 462 Socket 7 |
Steel | Ribbed structure | Steel | no name | 108 g | 4900 | 54 °C | 42 dB |
Smart Cooler FSM1168T | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Copper | no name | 187g | 3700 | 56 °C | 37 dB |
SPACE WALKER | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Alum- inum |
no name | 217g | 4700 | 42 °C | 50 dB |
SPRING SPCT38I BAF-IB | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Alum- inum |
no name | 168g | 4050 | 44 °C | 43 dB |
SPRING SPCT80I AAF-IB | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Alum- inum |
no name | 174g | 4150 | 45° C | 46 dB |
Swiftech MC370-0A | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Grid structure | Alum- inum |
Papst | 429 g | no sensor | 35 °C | 57 dB |
Swiftech MC462 | Socket 462 | Alum- inum /Copper |
Grid structure | Copper | Pabst | 789 g | 4900 | 30 °C | 63 dB |
Taisol CGK 747092 | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Alum- inum |
no name | 165 g | 4800 | 41 °C | 49 dB |
Taisol CGK 747092 | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Alum- inum |
no name | 136g | 4700 | 41 °C | 49 dB |
Thermaltake Aircooler | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Alum- inum |
no name | 206 g | 5500 | 43 °C | 49 dB |
Thermaltake Super Orb | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Alum- inum |
no name | 273 g | 4600/6400 | 39 °C | 55 dB |
Titan TTC-D2TB | Socket 370 Socket 462 Socket 7 |
Steel | Ribbed structure | Steel | no name | 150 g | 5100 | 50 °C | 46 dB |
Titan TTC-M5AB | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Alum- inum |
no name | 207 g | 5500 | 45 °C | 49 dB |
TRIG T40-6EA | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Alum- inum |
ADDA 1025 | 282g | 4400 | 41 °C | 60 dB |
TRIG T40-6EAC | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Copper | ADDA 8025 | 323g | 3380 | 40 ° C | 55 dB |
Zalman CCC-Projekt | Socket 462 | Copper | Ribbed structure | Copper | no name | 467g | 4900 | 44 °C | 52 dB |
Zalman CNPS2005 | Socket 370 Socket 462 Socket 7 |
Alum- inum |
Ribbed structure | Alum- inum |
no name | 129g | 2850 | 46 °C | 46 dB |
Zalman CNPS3000 | Socket 370 Socket 462 Socket 7 |
Copper | Ribbed structure | Copper | no name | 260g | 2850 | 45 °C | 46 dB |
Zalman CNPS3100 | Socket 370 Socket 462 Socket 7 |
Copper /Gold |
Ribbed structure | Copper /Gold |
no name | 310g | 2850 | 45 °C | 46 dB |
Zalman CNPS3100G | Socket 370 Socket 462 Socket 7 |
Gold | Ribbed structure | Gold | no name | 313g | 2850 | 44 °C | 46 dB |
Conclusion – Just Say “No” to Cheap Generics
Tom’s Hardware pulled out all the stops – no fewer than 46 coolers underwent a comprehensive testing procedure in our Munich laboratory. The progress made in the area of coolers is absolutely mind-boggling. Hardly any manufacturer can afford to limit itself to offering simple, two-bit coolers if he wants to survive in the market these days. Having said that, it’s still utterly incomprehensible to see that among the slew of coolers we tested, there are so few innovative ideas that engender well-engineered products. In particular, the South Korean manufacturer Zalman, despite its offering of fan coolers with unusual and interesting designs in all kinds of materials, the doesn’t exactly emerge in glory here. When combined with a large case fan, these designer items still only provide below-average cooling performance.
One of the real highlights in the group remains Noise Control’s Silverado, which combines two encapsulated rotors (arranged like turbines) with a CPU contact surface made of pure silver. It’s the best of both worlds – a very low noise level and a high cooling performance. Of the newcomers, the Blizzard Thunderbird doesn’t cut that bad a figure – while its cooling performance is very good, its noise level is a whisker away from unacceptable. Real hard-core overclockers, who really don’t care about the noise level, ought to take the cooler that produces the lowest temperature – the Swiftech MC462. The CPU core is kept at a remarkably cool 30 degrees, creating a lot of leeway for overclocking. On the downside, this 800 gram heavyweight can easily turn into a projectile if isn’t screwed down on the motherboard. Having said that, in cooling performance, the Swiftech MC462, the Blizzard Thunderbird and the Silverado from Noise Control beat anything this side of water coolers, as our highly comprehensive tests have shown once again.
Preview – Liquid Cooling
For this reason, we’ll soon be taking a look at water cooling systems, which produce even more arctic temperatures.