Second Attempt: Scythe Setsugen 2 VGA Cooler Review

A new cooler from the Japanese Scythe Company is their second attempt to design a universal and at the same time pretty efficient graphics card cooling system. Let’s see if they succeeded.

by Sergey Lepilov
03/02/2011 | 12:30 PM

Almost a year ago we reviewed Scythe Setsugen cooler, which turned out not a great success, so to speak. That was probably one of the reasons why it had been discontinued so rapidly. Note that they only announced Setsugen’s replacement this past September, and the new Setsugen 2 started selling just recently. Today we are going to find out if the new cooler is indeed that good compared with the reference cooling systems and competitor products.

Package and Accessories


The cooler is sealed in a compact plastic package, which is covered in all sorts of information, typical of Scythe:


The box is pretty heavy, although the cooler doesn’t weigh that much: only 435 g. Scythe Setsugen 2 is accompanied by the following items:

I would like to add that Scythe Setsugen 2 is made in Japan and is priced at about $45. It comes with a two-year warranty.

Design and Functionality

The second version of the Setsugen cooler is dramatically different from the first one. They modified just about everything, starting with the heatsink base and finishing with the fan:


The new cooling system measuring 176x33x139 mm and weighing 435 g consists of only 68 aluminum plates pierced by four copper heatpipes 6 mm in diameter. There is a 120x120x12 mm fan on top of it:


For added compactness of the cooler design the fan is pushed inside the heatsink, which reduces the effective cooling surface size of the latter substantially making it only 2,360 cm2. The heatsink plates are 0.25 mm thick and are spaced out at 1.3 mm from one another. These plates seem to be simply pressed against the heatpipes, because we saw no traces of soldering anywhere.

I would like to draw your attention to the peculiarity of the Scythe Setsugen 2 design, namely, to the fact that the heatpipes do not pierce the heatsink body evenly and are grouped in its lower part:


I dare suppose that the fan will not cool the heatsink too well in this area.

I was even more surprised to see how thick the cooler base plate was: 13 mm!

As you may understand with a base plate that thick the heat transfer between the heat source and the heatsink plate array leaves much to be desired. Moreover, there were no grooves of any kind where the heatpipes contact the base. However, there were a few in the top miniature heatsink, although they are hardly needed there. I would also like to add that the heatpipes are glued to the cooler base using thermal glue instead of soldering, which we would definitely prefer.

At the same time, we have to give Scythe due credit for making the base very even and finishing it nicely:

However, the base plate is only 30x29 mm big, which will be enough to completely cover the bare AMD Barts die and will be barely enough for AMD Cayman:

AMD "Barts" (Radeon HD 6870/6850)

AMD "Cayman" (Radeon HD 6970/6950)

The base of Scythe Setsugen 2 cooler will be too small for the majority of graphics processor heat-spreaders on Nvidia graphics accelerators, although they are all listed among the compatible products. Another (rhetorical) question is: do we really need to overlap the entire heat-spreader on these GPUs?

Scythe Setsugen 2 is equipped with a slim 12 mm Scythe Slip Stream Slim fan (SY1212SL12H-V), which we have already discussed in our Fan Roundup recently:

The “H-V” letters in the model name indicate that it is equipped with a variable rotation speed controller designed as a rear-panel bracket:

The fan itself is connected to the free three-pin connector on the mainboard. The fan rotation speed may vary in the interval from 800 (±30%) to 2000 (±10%), the airflow may vary between 19.4 and 45.47 CFM, and the noise – from 19.53 to 33.67 dBA. According to the results obtained during our fan tests, the thin Slip Stream model doesn’t look too convincing against the background of regular 25 mm models of 120 mm fans, in both: noise as well as airflow. However, in case of Scythe Setsugen 2 its job is simpler, because the heatsink to be cooled is only 10 mm thick, so a slim fan like that should be quite enough. Nevertheless we will check the cooling efficiency of Scythe Setsugen 2 also with a standard 120x120x25 mm fan. But before we proceed to the tests, I would only like to add that the fan is attached to the heatsink using two compact wire clips, which can also be used for a regular fan, unless it has bushes around the retention holes.

Compatibility and Installation

Scythe Setsugen 2 is compatible with a lot of contemporary as well as pretty outdated graphics accelerators (which, in fact, doesn’t necessarily mean that it can cope with their cooling alright). It comes with detailed step-by-step instructions, which facilitates the installation process. Scythe Setsugen 2 is fastened using two types of retentions screwed on to the cooler base with four screws:


As we see, these aluminum retention plates have a bunch of small threaded holes, which makes Setsugen 2 compatible with so many graphics accelerators. Also, since the base is very thick, the cooler heatsink is located far enough for the graphics card PCB and therefore doesn’t interfere with any of the electronic components or other heatsinks on it.

The final step in the cooler installation process is to fasten the cooler on the PCB using four thin screws (there are two types of those bundled with the cooler) and a backplate:

I have to admit that it is very difficult to do, as it is pretty difficult to get into the right retention hole. However, if managed to accomplish that, then this is what you graphics card will look like with Scythe Setsugen 2 installed (we have an AMD Radeon HD 6970 here):


He photo on the right shows Scythe Setsugen 2 cooler with an alternative 25 mm thick fan.

And this is what a graphics card with Scythe Setsugen 2 on top will look like inside the system case:


As we see, only the closest PCI-Express slot will be blocked by the installed graphics card, but I am sure you all understand that you shouldn’t install any full-size expansion cards in it even if you could, because it will block the airflow to the graphics card and most likely will affect the cooling efficiency.

Now we should only install the rotation speed variator instead one of the back panel brackets, connect the fan power cable to the mainboard and we are ready to start. That’s pretty much the entire installation procedure for Scythe Setsugen 2. Of course, you should remember to stick all enclosed aluminum heatsinks onto the VRM components even before the Setsugen 2 cooler is in place.

In conclusion I would like to show you a photo of Scythe Setsugen 2 installed onto an AMD Radeon HD 6870:

On this graphics card (and most likely other similar ones), the cooler heatsink covers the additional power connectors, so it becomes extremely inconvenient to connect and disconnect additional power.

Technical Specifications and Recommended Pricing

Testbed and Methods

The tests were performed in a closed system case. Our testbed was configured as follows:

Both graphics cards used in this test session were originally equipped with reference cooling systems:



AMD Radeon HD 6970 worked at nominal frequencies throughout the entire test session, while Inno3D GeForce GTX 570 was pre-overclocked substantially above the nominal. It would be even better for our today’s cooler test.

The testing programs were installed under Microsoft Windows 7 Ultimate x64. We used DirectX End-User Runtimes libraries (from November 2010) and Catalyst 11.1a and Nvidia GeForce/ION 266.58 WHQL graphics card drivers. We used two 12-minute runs of FurMark version 1.8.2 launched using a renamed EXE-file in stability test mode with “Xtreme Burning Mode” option enabled and in 2560x1600 resolution. WE enabled anisotropic filtering 16x in the driver control panels in order to increase the GPU operational load:

We also tested the graphics cards in game mode using five runs of Aliens vs. Predator game in 2560x1600 resolution with maximum graphics quality settings but without antialiasing:

By testing the graphics cards in this mode we should be able to reveal their temperatures under typical gaming load.

We used MSI Afterburner utility version 2.1.0 Beta 7 to monitor graphics card temperatures and frequencies and GPU-Z version 0.5.1 utility:

The tests were run at least twice for each type of load. The temperature stabilization period between the two test cycles was about 10-12 minutes. 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 stayed around 23.8-24.5°C.

We are going to compare the cooling efficiency of Scythe Setsugen 2 cooler against that of the reference coolers on AMD Radeon HD 6970 and Nvidia GeForce GTX 570 as well as against a pretty efficient Arctic Cooling Accelero XTREME 5870 and Thermalright Shaman coolers: 

We tested Scythe Setsugen 2 not only with its default Slip Stream 120 Slim fan, but also with a regular 120x120x25 mm fan – Noiseblocker NB-BlackSilent XLP, which you may have already noticed on the photos above.

Moreover, we also performed a quick blitz-test of the new Scythe Setsugen 2 cooler on a reference AMD Radeon HD 6870 graphics card and compared it against the reference cooler and Arctic Cooling Accelero Twin Turbo Pro, which competed very successfully against the first Setsugen cooler about a year ago. Here I would like to add that we replaced the default thermal interface on all graphics cards with Gelid GC-Extreme and used VRM heatsinks from Thermalright Shaman kit in all cases.

Cooling Efficiency Tests

AMD Radeon Graphics Cards

First of all let’s see how well the new Scythe Setsugen 2 and its competitors will cope with a hot graphics accelerator like AMD Radeon HD 6970 2 GB in the gaming mode:

Overall, the situation looks very good for Setsugen 2, because even in quiet mode of the standard fan at 1000 RPM, the cooler works better than the reference one in automatic mode. Setsugen 2 starts cooling even better if you replace the slim default fan a regular 255 mm fan: the peak temperature in this case dropped by 9°C. At maximum fan rotation speed this difference reduces to 4°C. I have to say that even though Setsugen 2 proved to be more efficient than the reference cooler in the gaming mode, it lost to both alternative competitors from Arctic Cooling and Thermalright. However, it was something we had expected to happen. Here I would only like to add that Arctic Cooling Accelero Twin Turbo Pro has totally failed on a powerful graphics card like that:

At the same time, we shouldn’t get too excited about the success of Setsugen 2 in the gaming test just yet, because it still has to pass the “brutal” FurMark:

Scythe Setsugen 2 fails to cool AMD Radeon HD 6970 2 GB in this test. And it is not really about the super-high GPU temperature. The GPU overheating protection system kicks in, because even at maximum rotation speed of the cooler’s default fan and with the max overheating threshold increased by 20% in the Catalyst driver, we still go the following graph under heavy load in the FurMark test:

The overheating protection starts kicking in after 4-5 minutes into the test, and replacing the slim fan with a standard one didn’t help, although it did lower the peak temperature by 8°C. At the same time, other testing participants, including the reference cooling system produced straight graphs without any short-term drops in graphics card clock speed. I would also like to add that Arctic Cooling Accelero XTREME 5970 on Radeon HD 6970 proved to be more efficient than Thermalright Shaman.

Well, since Scythe Setsugen 2 failed to cool a super-powerful AMD Radeon HD 6970 2 GB, let’s see how well it deals with a less hot AMD Radeon HD 6870 1 GB at nominal frequencies. Unfortunately, neither Arctic Cooling Accelero XTREME 5970 nor Thermalright Shaman were compatible with this graphics cards, so Setsugen 2 will only be competing against the reference graphics card cooler and Arctic Cooling Accelero Twin Turbo Pro in two speed modes. Here are the results obtained in FurMark:

As we can see, Scythe Setsugen 2 does way better with the less hot Radeon HD 6870 than it did with AMD’s most powerful single-GPU graphics accelerator Radeon HD 6970. At the same time, the cooler loses 5°C to its competitor, Arctic Cooling Accelero Twin Turbo Pro, in quiet mode although it does outperform the latter at maximum fan speed. Replacing the default Slip Stream Slim fan with a 25 mm one was considerably less rewarding than on Radeon HD 6970, but the cooling efficiency did improve. As we have expected, the reference Radeon HD 6870 cooler was the ultimate outsider this time, unless you don’t really care about the noise from the VGA cooler at all.

Now let’s see how the new Scythe cooling solution and its competitors will cope with an Nvidia GPU.

Nvidia GeForce Graphics Cards

Here we only have one competitor left, besides the reference cooler - Thermalright Shaman, because Arctic Cooling Accelero XTREME 5970 cannot be installed onto Nvidia GeForce GTX 570 and we didn’t have any other cooler available to us at this time. Here are the obtained results:

Just like with Radeon HD 6970, Setsugen 2 failed to cope with cooling our GeForce GTX 570: the GPU temperature was way too high. The interesting thing is that the GPU temperature in Aliens vs. Predator game also reached 87°C under Scythe Setsugen 2 at 1000 RPM of its default fan, which is not an acceptable temperature under an alternative cooler, especially in a regular gaming test.

Since we have already tested the acoustic performance of all the fans participating in our today’s session, we won’t discuss it here again.


Although Scythe Setsugen 2 can cool not very hot graphics accelerators from the mainstream and entry-level price segments and at the same time produce very little noise, we still feel that this cooler is not a success for several reasons.

First of all, there are obvious issues with the cooler heatsink design, such as dense heatpipe packs inside the heatsink body, extremely thick base plate without the grooves for the heatpipes, use of thermal glue instead of soldering and last but not least small effective heatsink surface size (it is too thin). Secondly, universal cooler is a great thing, but the installation procedure (which is identical for all graphics cards) is so inconvenient that at times you get really mad about those engineers who designed this retention featuring thin screws and tiny holes, which require sniper-precision and extreme patience during assembly.

Thirdly, we consider the MSRP of Scythe Setsugen 2 to be a little too high for the cooling efficiency it provides. Had it been priced at $30 or less, it could have been regarded as an alternative to reference coolers on mainstream and entry-level graphics cards. However, today you cannot find Setsugen 2 for less than $40. So, summing everything up, Scythe Setsugen 2 is in our opinion an inconveniently universal cooler with average efficiency, adjustable acoustics and high price.