by Sergey Lepilov
08/26/2010 | 02:36 PM
You may remember that the Thermalright IFX-14 used to be an unrivalled leader among air coolers for CPUs. It was the best in terms of sheer performance. Then it was overtaken by the Noctua NH-D14 after which a number of effective coolers came out such as the Cogage Arrow and the ThermoLab Baram 2010. The release of the Venomous X with its unique fastening mechanism and mirror base could not bring the leadership back to Thermalright but now the company makes another attempt with two new products: HR-02 and Silver Arrow.
We will discuss them in two separate reviews. Why two rather than one review? The fact is the HR-02 and the SilverArrow are designed for different applications. The HR-02 is meant to deliver its best performance in passive mode or with low-speed fans. Such coolers should be tested in a special way and we guess you’ll like this model if you prefer silent computers. The SilverArrow ambitiously aspires to be the best overall in both quiet (but not exactly silent) modes as well as with high-speed fans. Considering this difference, we won’t throw both coolers into the same review but will focus on each of them individually. First goes the Thermalright HR-02.
The product box is very simple and bears little information. It keeps the cooler from any danger well enough, though. Inside the thick cardboard wrapper there is a polyethylene-foam corset that fixes the cooler in place. Manuals, a sticker and a screwdriver can be found on top of that corset.
A cardboard box with accessories is next to the cooler. It contains everything you need to install the cooler and equip it with fans.
Fans are not included into the box. This is rather odd since the recommended price of the Thermalright HR-02 is rather high at $79.95. The cooler is manufactured in Taiwan.
The Thermalright HR-02 can be characterized as monumental and huge. This tower-design heatsink weighs 860 grams and measures 112x140x160 millimeters.
It is but slightly smaller in comparison with the Noctua NH-D14.
The total heat dissipation area of the Thermalright HR-02 is 10,690 sq. centimeters which is 11% smaller in comparison with the Noctua NH-D14 (12,020 sq. centimeters).
Consisting of thick fins, the HR-02 is robust and seems monolithic, even though it is actually a composite structure. It is based on six copper heat pipes, 6 millimeters in diameter, which go through a copper base. Perforated aluminum fins, 0.5 millimeters thick and 3 millimeters apart from each other, are put on those pipes and soldered to them.
The perforation, the thickness of the fins and the large spaces between them indicate that the HR-02 is meant for fan-less mode or for low-speed fans. As you can see in the right photo, the heatsink is somewhat shifted relative to the cooler’s base. Thermalright says this moves the top of the heatsink away from the back-panel system fan for easier installation but enthusiasts say that this slanting design may facilitate the airflow at the bottom of the heatsink (below the fins) and improve the cooling of the mainboard’s power components. The heatsink is nickel-plated.
Each of the aluminum fins is perforated and has 62 slits with one edge curved up. There is also a large intricately shaped central opening.
This perforation makes the airflow inside the heatsink more turbulent to improve the transfer of heat from the fins. The only thing we are confused about is the size of the large opening. Judging by the user manual, it is meant for the long screwdriver included into the box: you insert it in there to tighten one screw of the cooler’s fastening. But why is it so large and figured? This may serve an aesthetic purpose but the loss of some of the heatsink area is obvious. There is also a loss in the total heat dissipation area due to the large clearance: the distance from the bottom fin of the heatsink to the cooler’s sole is 50 millimeters (57 millimeters to the mainboard surface). Of course, this is meant to ensure compatibility with memory modules equipped with tall heatsinks but the Thermalright HR-02 is kind of leaning forward and does not hang above the memory modules if you orient it properly (as explained below).
Some 4 or 5 fins might have been added to increase the heatsink’s total area and improve its performance. It would also be good to get rid of the large hole for a screwdriver.
You may have already noticed the way the heat pipes are positioned in the heatsink body. Thermalright was the first manufacturer to place the heat pipes in the heatsink alternately in order to distribute the heat uniformly in the fins. Other makers have adopted this solution, too. The HR-02 shows a more fundamental approach of this problem: the heat pipes go through the base and alternate and pierce the heatsink in two pairs of separate rows, each consisting of three pipes. Thus, neighboring heat pipes have a dedicated non-overlapping heat dissipation zone in the heatsink fins and work more effectively than if they were placed alternately or linearly.
The cooler’s sole is typical of Thermalright products:
There is no mirror shine but the surface is finished well. The pipes are soldered in the base, the minimum thickness of the copper plate below the pipes being 1.7 to 1.9 millimeters.
The cooler’s sole is far from ideally flat.
With the new high-pressure fastening mechanism the HR-02 left a good-looking trace on our CPU.
The cooler’s base is 42x40 millimeters large.
The Thermalright HR-02 is compatible with all modern Intel platforms and is equipped with the same fastening mechanism as the Venomous X. It consists of a back-plate with plastic tips, steel pins with spacers and a fastening plate:
The fastening mechanism has a variable pressure (from 40 pounds (or 18 kilos) to 70 pounds (32 kilos)) which is adjusted by means of the central screw and key.
To install the HR-02 on mainboards for AMD processors, you have to purchase a special fastening mechanism which is not included into the box.
This is how large the Thermalright HR-02 looks compared to our mainboard:
Although the HR-02 kit does not include any fans, there are four wire brackets (two for a 120mm and two for a 140mm fan) and four rubber angles that you should glue to the heatsink to suppress vibrations.
The HR-02 looks even more massive with a 140mm fan.
Take note that a large part of the fan’s airflow will miss the fins, which is not good in terms of CPU temperature but may have a positive effect on the temperature of mainboard’s near-socket components.
According to Thermalright, the HR-02 should be installed in one position on mainboards for Intel processors:
That is, the heatsink must be as close as possible to the back panel of the case and slanting away from it. In this position the HR-02 does not block the memory slots even when you install a fan on it. And it can use the exhaust system fan to improve its own performance. Besides, the HR-02 is a one-way heatsink. It means that the airflow should go in one direction only to achieve maximum performance: from the straight pipes to the slanting ones. If the airflow goes in the opposite direction (i.e. when you install a fan on the other side or turn the heatsink around on the socket as in the right photo), the peak CPU temperature grows by 3°C. That’s not much, but even such trifles may be important.
You can try to install the heatsink perpendicularly to the memory slots. Then you can install a second fan on it:
Despite the better orientation of the heat pipes on an LGA1366 processor, this positioning lowers the cooler’s performance in comparison with the two installation variants mentioned above. Thermalright engineers do not recommend this orientation, either.
The HR-02 allows using 120mm fans instead of rather rare 140mm ones.
However, the HR-02 is 140 millimeters wide, indicating that 120mm fans are not the best option for this cooler. Indeed, the peak CPU temperature is about 5°C higher with one 120mm fan and 3-4°C higher with two 120mm fans compared to one and two 140mm fans, respectively.
We ran all tests inside a system case with a removed side panel. Here is our testbed configuration:
When we tested the coolers with fans installed, we overclocked our six-core processor (with its default protuberant heat-spreader) with the multiplier set at 24x and “Load-Line Calibration” enabled to 4.27 GHz. The nominal processor Vcore was increased to 1.38125 V in the mainboard BIOS:
Turbo Boost and Hyper-Threading technologies were disabled during our test session. The memory voltage was at 1.64 V and its frequency was 1.71 GHz (8-8-8-16_1T timings). All other parameters available in the mainboard BIOS and related to CPU or memory overclocking remained unchanged.
All tests were performed under Windows 7 Ultimate 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 CPU core for the results charts. Moreover, we will also provide a table with the temperature readings for all cores including their average values. 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. The room temperature during our test session varied between the annoying 27.6-28.0 °C.
We are going to compare Thermalright HR-02 against Noctua NH-D14 super-cooler:
Both coolers were tested in passive (fanless) mode as well as with one and two Thermalright TY-140 fans at 600, 800, 1000 and 1200 (±3%) RPM. We decided not go for Noctua’s default fans for a reason that will be explained in our next article.
First, let’s see what the coolers can do without any fans.
For our comparative test to be correct, we lowered the speed of the system fans to 600 RPM so that they didn’t help the coolers much. The side panel of the system case was open, which worsened the airflows inside the chassis and made it harder for the coolers. There is one change in our test methods: the amount of memory for the Linpack x64 test was increased to 4750 megabytes (almost all of the free system memory). This makes the test longer, which is important for passive heatsinks.
First, let’s see how the coolers cope with a non-overclocked six-core CPU (these are the results of the second test cycles):
It is clear that we have no winner at the default CPU frequency. The coolers produce identical results. Next, we overclocked our CPU to 3.8 GHz and tested the coolers once again.
Surprisingly enough, we’ve got identical results again! Both coolers make the CPU stable at a clock rate of 3.8 GHz. This is good already but can we overclock our six-core CPU even more with these passive heatsinks? Yes, we can!
We’ve got impressive results indeed. Even some coolers with fans fail to reach a CPU frequency of 4.1 GHz and we achieve it with passive heatsinks! The new model from Thermalright is 1°C better than its opponent at peak load. Moreover, that frequency was the maximum we could achieve with the Noctua whereas the Thermalright HR-02 could perform one more feat (the amount of memory allocated to the test program was increased a little more):
Now let’s put all the results into a single diagram:
The Noctua NH-D14 loses its title of the best cooler for fan-less mode even though it is but slightly inferior to the Thermalright HR-02.
Now let’s see what our coolers can do if they are equipped with fans. The Noctua NH-D14 was installed as in the photo from the Testbed and Methods section of the review. The Thermalright HR-02 was installed in the proper way so that its airflow was directed towards the back panel along the heatsink. For the back-panel 120mm system fan not to intervene with the competition, we removed it. This also made it simpler for us to install a second exhaust fan on the HR-02. You can see the results of our test in this table and in the following diagram:
Click to enlarge
Thermalright engineers were right to claim that the HR-02 is invincible with one fan. This cooler beats the Noctua NH-D14! And the lower the speed of the 140mm fan, the larger the advantage of the HR-02 is. The Noctua NH-D14 is better with two fans, though.
When equipped with two fans, the coolers are equal at 610 RPM, but the Noctua goes ahead at higher speeds, winning by 1-2°C and remaining the best cooler in active mode with two fans.
And finally, let’s check out both coolers with the CPU overclocked to 4.46 GHz at a voltage of 1.4625 V. We ran this test with two fans at 810 and 1240 RPM.
Although the room temperature remained rather high during this test (about 28°C), both coolers coped with their job and made the well-overclocked Intel processor perfectly stable. The Noctua NH-D14 wins here, outperforming its opponent by 3°C in quiet mode and by 1°C at the maximum speed of the two 140mm Thermalright TY-140 fans.
We guess the Thermalright HR-02 cooling system is up to users’ expectations. It is especially good for everyone who prefers very, very quiet computers. The HR-02 is the best cooler available today in fan-less mode as well as with a single low-speed fan. You should take this into account when evaluating this model. An ordinary user may wonder if it wouldn’t be better to add $5 and buy a Noctua NH-D14 ($85) which comes with two fans, supports multiple platforms, and is even somewhat more effective from an overclocker’s point of view. The choice would be obvious for everyone, save for people who do not want to hear any sound from their system case. And for such people, the Thermalright HR-02 is unrivalled. That said, we still think that the recommended price of $80 is rather too high for the HR-02 and we wish the manufacturer included one Thermalright TY-140 fan and a fastening mechanism for AMD processors into its box. This would make the cooler even more appealing.
Thermalright HR-02 receives our Editor' Choice award for offering superb cooling efficiency at a very low level of noise:
And as a kind of teaser, we can tell you that our next review will feature the SilverArrow vs. the Noctua NH-D14.