by Sergey Lepilov
11/20/2009 | 06:24 PM
The CPU cooling solutions market is currently the most diverse of all other market segments connected with personal computer components. I believe only mainboard market could probably compare with the cooling solutions market in terms of variety richness and that would be primarily due to the fact that there exist multiple modifications of the same mainboard model. Unfortunately, this enormous number of cooling products is mostly the result of aggressive marketing rather than need, because no matter ho many new coolers come out, the cooling efficiency doesn’t improve beyond its current level any more.
The introduction of new platforms with new processor sockets and retention types encourages the release of new cooling solutions. For example, the recent announcement of new CPUs and mainboards with LGA1156 socket was a real gift for the companies’ marketing specialists. By simply adding a new retention kit to the accessories bundle and freshening up the cooler model name they can claim having a new solution that is absolutely worth your attention. And users do buy them. Although, there are also positive examples, when the manufacturers provide the owners of their old cooler models with new retention kits for free thus winning more respect and appreciation. In fact, it is really not about retention kits but about something we have long been trying to draw your attention to: the absence of any improvement in the cooling efficiency of existing solutions.
Nevertheless, there are some interesting findings in this sphere, and today we are going to discuss and test one cooler like that. I am talking about CoolAge X120TF. And to make sure that it performs in good company we decided to add a couple of “new” solutions represented by Evercool Transformer 4 and Zalman CNPS10X Quiet as well as five competitors with similar design and features. As a result, the initial single-product review turned into a pretty large-scale comparison including eight air cooling solutions, which is not something we do that often. So, let’s start with the detailed review of the three new products introduced on our site for the first time.
In the past we couldn’t regard CoolAge Company as a manufacturer of high-end cooling solutions, because they were primarily offering products similar in design and functionality to the coolers included with the boxed CPU models. Now CoolAge is proud to offer a very interesting heatsink, X120TF, and a cooler built with it – FO-X120. We received the heatsink alone and we are going to take a real close look at it right now.
The packaging is a small box with a cut-out window in the front:
The information provided on the package is pretty typical for such products. It includes the description of key features, official specifications, the list of supported processor sockets and bundled accessories. The latter look as follows:
CoolAge X120TF comes with everything necessary for successful cooler installation on all contemporary platforms. They also included a cute little polar bear figure as a souvenir :)
The cooler is made in China and is recommended to sell at $45. There is no mention of the warranty period anywhere, but it is definitely at least a year long.
So, we see a not very traditional tower heatsink measuring 126x63x150 mm and weighing 750 g:
Six copper nickel-plated heatpipes 6 mm in diameter are soldered to the copper nickel-plated base. They hold 65 aluminum plates of pretty unique shape:
The unique thing about CoolAge X120TF heatsink is the so-called honeycomb structure of the heatsink array, which was created by pressing two neighboring plates against one another. The plates are profiled in such a way that they create honeycomb shaped cells when pressed against one another:
As a result, we see a very dense array of 0.4 mm plates with variable gaps between them ranging from 0.2 to 0.4 mm. CoolAge call this technology “Wind Tunnel” and believe that in this case heatsink plates form a lot of channels for uninterrupted and efficient airflow.
As you may see, the contact spots between two heatsink plates form about 0.8 mm of metal. So, we assume that far not every heatsink will be able to push the airflow through such a dense structure successfully, which means CoolAge X120TF is very unlikely to perform efficiently in quiet fan modes.
Besides the uniquely shaped heatsink plates CoolAge X120TF heatsink also uses non-linearly distributed heatpipes. Moreover, the heatpipes are lined up in two rows on each side of the heatsink forming kind of a wedge along the airflow direction:
CoolAge doesn’t specify why they decided on this particular heatpipes placement that is why we can only suppose that this has been done not only for more even heat distribution over the heatsink plates, but also for lowering the airflow resistance. Overall, CoolAge X120TF heatsink reminds us of the Thermalright Ultra-120 eXtreme (not only by its design but also by the color of the metal parts).
The heatpipes lie in special grooves cut out in the cooler base where the minimal plate thickness beneath the heatpipe makes 2 mm. the contact surfaces between the heatpipes and the base plate are soldered together. The base surface finish is of pretty high quality although it hasn’t been polished:
Unfortunately, we can’t say the same about its evenness:
It is obvious that the base is slightly curved in two spots and you can see the result of this curving on the thermal compound imprint photo below:
We didn’t even out the base plate before the actual performance tests and used the cooler as is.
The cooler is installed in very simple but at the same time very reliable ways. It uses a retention swing-clip with a locking tab for AMD platforms (no need to remove the mainboard from the system case for this installation). As for Intel LGA775 and LGA1366 platforms, it uses steel retention plates that are attached to the cooler base with screws and have spring-screws of their own that in their turn get attached to the board.
Of course, you will need to remove the mainboard from the system in the latter case, because the spring-screws go through the mainboard PCB to the backplate on the bottom, but the hold will be extremely secure: way stronger than in case of a swing-clip.
During the cooler installation onto an LGA1366 mainboard the distance from the lowest heatsink plate to the mainboard surface will be 45 mm (38 mm to the base). The cooler won’t be hanging over the memory modules that is why there shouldn’t be any conflict of interest between any of the system components:
CoolAge X120TF is shipped without a fan. The users can choose and purchase the best fan themselves. X120TF is bundled with two wire clips for the fan attachment. They should catch on to the retention holes on the inside of the fan frame and then are placed into special grooves in the heatsink:
There are only two clips in the bundle. Besides, the heatsink is shaped in such a way that you can’t install a second fan the traditional way that is why you will have to come up with something. For example, you could fasten the additional fan like this:
Or you could also top CoolAge X120TF heatsink with a single 38 mm fan with high static pressure:
However, it turned out not as efficient as I expected it to be, but we are going to talk more about it in the performance chapter of our today’s review.
Zalman CNPS10X Quiet cooler is a second modification of the Zalman CNPS10X Extreme, which should already be familiar to you. There is also the third modification called Zalman CNPS10X Flex and we are also going to review it later. And now let’s see what the differences are between the Quiet model and the pretty successful Extreme.
The box with Zalman CNPS10X Quiet is primarily of blue color, while the Extreme version was mostly black:
You can see the cooler fan and part of its heatsink through the round cut-out window in the front of the box. The back of the box describes the features and peculiarities of this cooling solution. The official technical specs are also there: you can see them on one of the box side panels.
The accessories bundle has changed. Besides the same items, which you get with the Extreme version, there is now Fan Mate 2 rotation speed controller with a power cable:
All other accesosries remained the same. Only the installation instructions are slightly different now (the pictures are different). Zalman CNPS10X Quiet is made in Korea and its recommended retail price is $50.
The cooler is of the same size and weight: 135x100x160 mm and 750 g. Its design has been modified just slightly that is why the overall shape is exactly the same as that of the Zalman CNPS10X Extreme model:
The most noticeable changes are the absence of nickel-plating over the heatpipes and base and the absence of the top plastic housing with the built-in rotation speed regulator (that is why it comes with Fan Mate 2). Despite these changes, Zalman CNPS10X Quiet looks even more stylish in my opinion than the Extreme version due to the polished top plate with embossed manufacturer and cooler logos and a central part of the heatsink of 15 slightly bigger plates than the top or bottom sections colored bright blue:
By the way, the Quiet cooler model has fewer heatsink plates than the Extreme one: 45 instead of 56. Fewer plates at the same cooler height of 160 mm resulted in larger gaps between the plates: 2.4 mm instead of about 1.7 mm. As a result, Zalman CNPS10X Quiet allowed using a less powerful 120 mm fan (Zalman ZF1125ATM model) and adding “Quiet” to its model name. The rotation speed of this fan varies between 700 and 1500 RPM (compare against 1000-2150 RPM by Extreme). The maximum level of noise with this fan shouldn’t exceed 28 dBA, according to the manufacturer.
Besides slower rotating fan, silicon shock-absorbing pads stuck to the corners of the fan frame should also help lower the noise. I would like to add that the fan uses a dual call bearing with 50,000 hours claimed MTBF.
The evenness and finish quality of the base plate are impeccable:
The thermal compound imprint from the processor heat-spreader on the cooler base turned out practically ideal:
Zalman CNPS10X Quiet installs exactly the same way as Extreme that is why we are not going to dwell on it here. In conclusion I would only like to show you the way this cooler looks inside the system case:
Note that if you decide to install a second fan you will have to slightly raise the fan above the heatsink because of the rounded heatsink shape on one side. We used rubber bands and plastic spindles that are normally used to install a mainboard into a system case. If you don’t lift the fan above the heatsink its blades will be sitting right on top of it and won’t spin.
The third and last new cooler for today is called Transformer 4 (HPJ-12025) and comes from Evercool Thermal Co., Ltd. The cooler is packed into a clear plastic blister box with colorful cardboard inserts:
The cooler and the following accessories are inside:
The cooler is made in China. Its recommended retail price is only $40 and it comes with a 1-year warranty.
Evercool Transformer 4 is a tower-cooler based on four copper heatpipes 8 mm in diameter that form part of the cooler base (heatpipe direct touch technology):
The distinguishing feature of this cooler is the use of two 120 mm fans installed for air intake and exhaust.
The heatpipes hold a total of 51 aluminum plates, each 0.35 mm thick, that are spaced out at 2.0 mm distance from one another:
In the base the heatpipes are 2 mm away from one another and each has a 7 mm wide effective contact surface:
This is the thermal compound imprint taken off an LGA1366 processor:
Evercool Transformer 4 is equipped with two EC12025SL12EA fans measuring 120x 120x25 mm. they are attached to the heatsink with wire clips so that one of them sucks air in and another one – ousts it out. The fans are of sparkling silver color. They are so bright that no LEDs are necessary:
They rotate with constant speed of 1000 (±10%) RPM. The claimed noise is at 23 dBA and the created airflow should be 38.14 CFM. Improved slide bearings in the fans should ensure up to 50,000 hours of worry-free operation. Both fans together consume no more than 4 W of power. The fans have rotors with 43 mm diameter and use three-pin power connectors on the end of a 310 mm long cable.
Evercool Transformer 4 is compatible with all contemporary platforms except the latest addition to Intel family – LGA1156. However, taking into account the installation procedure for this cooler we are certain that the corresponding retention plate for the new socket LGA1156 will be available shortly. In all cases the cooler is installed through the mainboard PCB by tightening the screw-nuts over the backplate retention spindles. This is what it looks like:
This retention ensures extremely strong hold and overall, we consider this type of retention to be the most reliable. There are about 37 mm from the lowest heatsink plate to the base of the cooler, but it is important to remember that the fans hang 8 mm below the last plate (although you can easily shift them up if you want to). Evercool Transformer 4 sparkles inside the system case:
The cooling efficiency barely depends on the way the cooler is turned during installation (the difference between two ways the heatpipes can be positioned didn’t exceed 1.5-2 °C under peak load).
All tests were performed inside a closed system case. Our testbed was identical for all coolers throughout the test session and featured the following configuration:
During this test session we managed to overclock our 45nm quad-core processor (with polished heat-spreader) with the multiplier set at 21x and “Load-Line Calibration” enabled to 3.95 GHz using the weakest cooling system of the today’s testing participants with the fan in quiet mode. The nominal processor Vcore was increased to 1.35 V in the mainboard BIOS.
The memory voltage was at 1.64 V and its frequency was around 1.5 GHz (7-7-7-14_1T timings). All other parameters available in the mainboard BIOS and connected with CPU or memory overclocking remained unchanged (set to Auto).
All tests were performed under Windows 7 RTM x64 operating system. We used the following software during our test session:
So, the complete screenshot during the test session looks as follows:
The CPU was loaded with two consecutive Linpack test runs with the settings as indicated above. The stabilization period for the CPU temperature between the two test cycles was about 8-10 minutes. We took the maximum temperature of the hottest processor core of the four for the results charts. The ambient temperature was checked next to the system case with an electronic thermometer with 0.1 °C precision that allows monitoring the temperature changes over the past 6 hours. During our test session room temperature was at 19.8-20.3 °C.
For comparison purposes we included the results for five additional super-coolers. Thermalright Ultra-120 eXtreme-1366 RT and ThermoLab BARAM to estimate the cooling efficiency of the CoolAge X120TF. Zalman CNPS10X Extreme – to compare against Zalman CNPS10X Quiet. Xigmatek Thor’s Hammer with heatpipe direct touch technology to compare against Evercool Transformer 4 with the same technology. Here are the photos of above listed coolers and their thermal compound imprint from the Intel Core i7 base:
The fifth cooler included is our ultimate reference Thermalright IFX-14 (with evened out base). We added it to compare against everyone else. So, we ended up with eight cooling solutions participating in our today’s test session.
Besides the default fans (if there are any included with the coolers), all testing participants were tested with one and two Thermalright TR-FDB-2000 fans working in quiet mode at 820 RPM, moderate mode at 1320 RPM and at maximum rotation speed of 2040 RPM. The fan rotation speed was adjusted and controlled by our special controller.
At first let’s check out the cooler performance with the default fans:
Click to enlarge
The only cooler in this part of our test session with heatpipe direct touch technology – Evercool Transformer 4 – can’t boast superior efficiency yielding over 10 °C even to the least efficient cooler of the remaining three. And all that despite the fact that Transformer 4 is equipped with two fans. Zalman CNPS10X Quiet in quiet mode is 2 °C more efficient than its “elder brother” – Zalman CNPS10X Extreme, due to larger gaps between the heatsink plates and the absence of nickel-plating on the cooler heatsink, base and heatpipes. However, at medium fan rotation speed of ~1500 RPM (which is the maximum speed for the Quiet model) the cooling efficiency of both Zalman solutions levels out, and at maximum rotation speed Extreme cooler dashes forward. In this case only Thermalright Ultra-120 eXtreme can compete against it.
The next part of our test session will be devoted to cooler performance when they were all equipped with identical Thermalright fans in three different rotation speed modes. Since we couldn’t fit all results into the same diagram we split them into three groups according to the fan rotation speed mode. The first diagram shows the coolers efficiency in quiet mode at 820 RPM:
The first thing that catches your eye in this diagram is the low efficiency of CoolAge X120TF and Evercool Transformer 4. And while we didn’t expect the latter to work any wonders, the new CoolAge solution turned out an unexpected disappointment. Looks like Tunnel Wind technology doesn’t really work that well at low fan rotation speeds, because the honeycomb heatsink structure features very high static resistance. Low-speed fans simply can’t push the airflow through the dense heatsink structure that is why this cooler didn’t do too well in the quiet fan mode. At the same time it is quite logical to assume that CoolAge X120TF cooling efficiency should increase significantly at high fan speed, and we are about to check it out shortly.
Speaking of other testing participants I can’t help pointing out that Xigmatek Thor’s Hammer has fully rehabilitated heatpipe direct touch technology, since we seriously doubted that it was fit for LGA1366 processors after the Evercool Transformer 4 tests. In a single-fan mode all coolers except the two solutions mentioned in the previous paragraph perform very close to one another. Even IFX-14 has no advantage here. However with two fans installed it gets far ahead of the competitors, although the processor is not overclocked too high and low room temperature helps all testing participants to be efficient. I would also like to specifically point out Zalman CNPS10X Quiet, which again turned out more efficient than the Extreme model proving that its name is well deserved. Note that with two fans the difference in peak temperature between them reaches 5 °C.
The next diagram shows the cooler performance in a moderate acoustic mode when the fans rotate at 1320 RPM:
True, when the fans speed up, CoolAge X120TF catches up with the leaders showing much better results, which, unfortunately, is not the case for Evercool Transformer 4. Although this cooler demonstrated 5 °C better performance than in the previous test, it is still the last one in our ranking. ThermoLab BARAM with a single fan, a cooler with a heatsink optimized for low airflow resistance, looks a bit more appealing than the others. Even Thermalright IFX-14 with polished off base surface can’t oppose BARAM here. At the same time, a second fan helps the latter get unattainably far ahead of the rivals.
Finally here are the cooler results at maximum fan rotation speed (at 2040 RPM):
Click to enlarge
There is not such a big gap between CoolAge X120TF and the rest of the pack (4..5 °C), but it is still less efficient than even Thermalright Ultra-120 eXtreme. As for Evercool Transformer 4, looks like I have already said everything about it, as well as about the advantage of Thermalright IFX-14 with two fans over five other super-efficient cooling solutions. That is why at this point we should only check out the results of maximum CPU overclocking under each of the testing participants (with two fans at 2040 RPM):
Judging by the results, the difference in cooler efficiency is minor. CoolAge X120TF is again the least efficient of all, however, it still copes well with cooling of an overclocked quad-core processor. As we have expected, we managed to hit the maximum clock speed with Thermalright IFX-14. This super-cooler is the only solution that could ensure stability at 4095 MHz and 1.3825 V Vcore.
Only two new coolers of the three testing participants are equipped with fans: Evercool Transformer 4 and Zalman CNPS10X Quiet. The fan of the latter demonstrates the same acoustics as the fan of the recently reviewed CNPS10X Extreme cooler. The difference is only the maximum rotation speed, which is lower by the Quiet model. At low rotation speed we can barely hear some rustling of the fan ball bearings, there is no electrical crackling in the entire rotation speed range. The subjectively comfortable acoustic mode lies within 1250-1350 RPM. As for the Evercool solution, its fans work extremely close to the level beyond which they become totally noiseless. This is, in fact, absolutely normal for quality fans using slide bearings and rotating at 1000 RPM or less.
Our today’s article told you about a few new air coolers for CPUs. Are there any solutions among them that would be suitable for overclocking systems? Yes, there is one – Zalman CNPA10X Quiet, which turned at least as good as the previous Extreme version, but costs $10 less and generated less noise during work. We can compare the efficiency of this solution against the best tower-coolers, like Thermalright Ultra-120 eXtreme or ThermoLab BARAM. Moreover, the new cooler is universal and very easy to install (if we disregard the fact that the mainboard needs to be removed from the system case, like with most other contemporary coolers).
Two other coolers are not so exciting. Despite the super high-tech heatsink design, CoolAge X120TF can only be efficient at high fan rotation speeds and high noise levels that is why it will barely suit any home users. It is pretty strange that a dense heatsink is designed to accommodate only one fan. We would also like to draw CoolAge’s attention to the quality of their cooler bases, unless it is a unique drawback of our particular sample. As for Evercool Transformer 4, we have to give this cooler due credit for low level of generated noise and the lowest price ($40) of all eight models discussed today. However, unfortunately, it turned out not too efficient, which diminishes the advantages mentioned above.
Coming back to the discussion we started in the beginning of this article, we can conclude that the new coolers tested today have once again failed to raise the efficiency bar to a new level and have simply joined the existing army of solutions. Nevertheless, we still hope that very soon really efficient coolers will appear that is why we will continue reviewing new solutions. Especially since the remaining months of the year 2009 promise to be especially exciting in this respect.