by Sergey Lepilov
07/11/2007 | 07:13 PM
Modern CPUs have stopped to generate more and more heat, but cooler manufacturers are still rolling out new products, some being modernized versions of older models and others completely novel. There is no stagnation in this field, this you can be sure of. And I think it means more fun and excitement for an overclocker since we can expect something exceptional in terms of cooling performance and noise from each more or less serious product that comes out these days.
This review is special. Not just because you’ll find the Ultra-120 eXtreme cooler from Thermalright examined and tested here. This new cooler has already gathered awards from hardware websites and earned recognition among overclockers. What makes this review special is that I tested five more super-coolers (and there were two more models that didn’t pass the qualification test). Besides checking the coolers’ performance and noise level, I also tested them by overclocking the same CPU to the highest frequency. You’ll find all of that in this very special review.
Frankly speaking, I had expected to see a more interesting package than an ordinary cardboard box, considering the proud name of the product.
You can learn what exactly is inside only by looking at the sticker on one of the sides:
This unassuming package design can hardly attract the eye of a casual shopper, but the packaging performs the protective function quite well. The heatsink is wrapped into a plastic pack and fixed in a polyurethane-foam jacket:
The following accessories can be found in the small box to the right of the main compartment:
There is no fan in the box – you have to buy one separately.
The Thermalright Ultra-120 eXtreme is very tall, equaling the Scythe Infinity at 160mm. Its heatsink is 132mm wide and 63.4mm thick:
The Thermalright heatsink has a tower-like design. Six copper heat pipes, 6mm in diameter, go through the copper base and carry curiously shaped aluminum plates. The plates are curved, resembling a plane’s propeller.
This increases the heatsink area in comparison with straight plates, but the manufacturer doesn’t declare the total heat dissipation area. Every heatsink component is nickel-plated. The weight of the heatsink is 790g.
The top view shows that the pipes are placed not linearly, like in the Scythe Infinity for example, but with a shift, forming two ovals:
That is, having a less thick heatsink, this cooler accommodates as many heat pipes as the Infinity. I guess this solution also has an unobvious additional advantage that the shifted heat pipes distribute heat more uniformly in the heatsink than if they were placed in a line.
On the other hand, the airflow from the fan meets more resistance with this design.
The pipes lie in special grooves and the minimum thickness of the base is about 3 millimeters.
The heat pipes have contact with the base by means of some thermal glue. You can see traces of it on the edge of the plate. This is no solder.
The cooler’s base is not protected with anything:
No one took the trouble of finishing it. The structure of its base is made up by semi-rings you can easily feel with your fingers:
Well, some people think that a surface like this allows for a more efficient transfer of heat from the CPU heat-spreader to the heat pipes than a polished base does if you use thick thermal grease.
The fasteners included into the Thermalright Ultra-120 eXtreme kit allow to install it on LGA775 and Socket AM2 mainboards. Other platforms are not supported.
The mainboard has to be taken out of the system case for either of the CPU sockets. The installation is a simple and intuitive procedure. You won’t even have to look up anything in the manual. You just take the necessary fastener and back-plate and fasten the heatsink with four screws through the mainboard. The spring-loaded screws have limiters so you just cannot damage anything. The pressure is high and the cooler is immovable on the CPU. The back-plate prevents the mainboard from bending. The cooler’s fastening mechanism is overall a very thought-through thing.
Installed on the mainboard, the Thermalright Ultra-120 eXtreme won’t have conflicts with the near-socket components because the heatsink hangs high on the heat pipes.
You only have to stick rubber compensator strips to the heatsink and hitch one or two 120mm fans to the juncture of ribs on the heatsink sides. There are only two wire brackets in the kit, but that’s enough for two fans. You can hitch the brackets to one side of each fan (to the top side, preferably).
Inside a system case, the Thermalright Ultra-120 eXtreme looks like this:
This cooler costs about $60 but you should add the price of one or two 120mm fans ($10) to that.
The coolers were tested on an open testbed as well as in a system case with the following configuration:
All tests are performed in Windows XP Professional Edition Service Pack 2. SpeedFan 4.32 is used to monitor the temperature of the CPU, reading it from the CPU sensor. The CPU is heated up by means of OverClock Checking Tool version 1.1.0 in a 60-minute test during which the system remains idle in the first and last 4 minutes.
The mainboard’s automatic fan speed management is disabled for the time of the tests. The thermal throttling of the Intel Core 2 Duo processor is controlled with RightMark CPU Clock Utility version 2.25. Our CPU begins to skip clock cycles on reaching a temperature of 82°C and higher.
I perform at least two cycles of tests and wait for 25-30 minutes for the temperature to stabilize during each test cycle. The maximum temperature of the hottest CPU core in the two test cycles is considered as the final result (if the difference is not bigger than 1°C – otherwise the test is performed once again). Despite the stabilization period, the result of the second cycle is usually 0.5-1°C higher.
The ambient temperature was monitored by means of an electric thermometer and remained at 22-22.5°C during the tests. The fan rotation speed is shown in the diagram as reported by monitoring tools.
Now, about the opponents. Most super-coolers hadn’t been tested on the quad-core CPU before. Moreover, I got a noise-level meter the super-cooler hadn’t been tested with, either. So, the review of the Thermalright Ultra-120 eXtreme was the occasion to perform an extensive comparative test of high-performance coolers. In order to qualify for participation in this review each cooler had to pass twice an hour-long test with the OCCT program on an overclocked processor. I’m aware that this test doesn’t promise a 100% guarantee of stability, but I just had neither time nor opportunity to test the coolers for days.
After the quad-core processor was overclocked to 3304MHz (+37.7%) with a voltage increase to 1.5V, I had to leave out the OCZ Vindicator (an analogue of the Scythe Ninja) and, oddly enough, the Cooler Master GeminII. The latter had behaved very well on a dual-core CPU, being only 2°C worse than the best air coolers, but would let the overclocked Intel Core 2 Quad begin to skip clock cycles. I made sure there was proper contact between the cooler and the CPU heat-spreader, but the results didn’t change: over 82°C under peak load and frequency throttling as the consequence.
So, there are only the strongest contenders left. I am going to compare their performance and noise level with those of the Thermalright Ultra-120 eXtreme. Here’s the full list of the super-coolers:
The tower-like coolers were tested with a stock intake 120mm fan as well as with two fans for intake and exhaust (I used the fans from the Scythe coolers set at 1200rpm). The Scythe Andy Samurai Master and the Thermaltake Big Typhoon were tested with their stock fans as well as with a fan from the Enzotech Ultra-X at the maximum speed of 2530rpm. The Zalman CNPS9700 LED and the Enzotech Ultra-X were tested without any fan modifications.
You won’t see the highly efficient Tuniq Tower 120 in this review because I didn’t had it in the lab at the time of the tests. The ASUS Silent Knight and Titan Amanda TEC are missing, too, unfortunately. Feel free to check out the corresponding reviews of these coolers for their complete performance report.
And now we can go right to the test results.
The results are presented in the following diagram:
It’s perfectly clear that the Thermalright Ultra-120 eXtreme is one of the best super-coolers available. Just consider the fact that this cooler with one 120mm fan is as good as the Scythe Infinity with two such fans. The installation of a second 120mm fan on the Thermalright cooler brings about a small performance increase but also makes it the best cooler in terms of performance to noise. Note that the Enzotech Ultra-X can keep the CPU a couple of degrees cooler at the maximum speed, but the noise level is too high.
I didn’t limit myself to the temperature factor in this test session, though. I also found the highest CPU frequency I could achieve with each cooler. The minimum temperature is not so important after all whereas the maximum frequency would point definitely at the best cooler. I performed this test on an open testbed to avoid the influence of the airflows that would be inside a system case. The CPU voltage was set at 1.5V. As had been verified on the best cooler, reducing or increasing this value did not improve the overclockability of the CPU. Besides the highest frequency achieved the diagram shows the CPU temperature under peak load for each cooler.
Here are the results:
The Scythe Andy Samurai Master in the standard version is working at its limit and the CPU can’t be overclocked further on it. Installing a higher-performance fan improves the result by 46MHz. As for the others, note that the CPU is indifferent to the installation of a second exhaust fan when it comes to the tower-like coolers, Infinity and Ultra-120 eXtreme. There is a reduction of temperature, but no increase in frequency.
The Zalman CNPS9700 LED performs well at last, being second best in the CPU frequency achieved (at the maximum speed). It is also obvious that the Thermaltake Big Typhoon is one of the most optimal solutions in the total of its characteristics (performance, noise, price, availability) even though I couldn’t reach the highest CPU frequency with it. The Enzotech Ultra-X is the leader in the CPU frequency achieved.
I can also tell you that my test of the weakest and strongest cooler of this review – Scythe Andy Samurai Master and Enzotech Ultra-X respectively – for reaching the highest frequency of an Intel Core 2 Duo showed that the latter cooler was better by 90MHz (in quiet mode).
I have to acknowledge that the difference in the max frequencies of the quad-core CPU is so negligible between the tested coolers that you should consider other factors in the first place, e.g. the amount of noise. I will consider the noise factor right now.
The amount of noise generated by the coolers was measured with a digital noise-level meter Center-321 (0.1dB accuracy) using A-curve weighing. The measurements were done at night in a totally quiet apartment – the level of ambient noise was as low as 31.6dBA. The amount of noise was measured:
The noise level of the system case can be viewed in this diagram. The subjectively comfortable level of 36-37dBA is marked with a dashed line in the diagram. The coolers are sorted in the diagram in the order of increasing loudness. Here are the results:
So, the Scythe Andy Samurai Master, the weakest cooler of this review in terms of performance, proves to be the quietest of all. The leader is closely followed by the Scythe Infinity and the Thermalright Ultra-120 eXtreme that have one 1200rpm fan. Next go the Thermaltake Big Typhoon and Zalman CNPS9700 LED together with the pair of tower-like coolers with two 120mm fans. The fan of the Enzotech Ultra-X is the loudest of all but the noise is compensated by this cooler’s highest performance. It should be acknowledged that the noise of the Enzotech Ultra-X is quite comparable to the other coolers at a distance of 3 meters from the system case.
A curious fact, the Scythe Andy Samurai Master and the Thermaltake Big Typhoon were tested with the same fan from the Enzotech Ultra-X at the max fan speed (2530rpm), yet the Big Typhoon proved to produce more noise. This must be due to the difference in the heatsink design because it is the noise from the flow of air that is measured at such a high speed rather than anything else. The Enzotech Ultra-X has the narrowest heatsink with minimum resistance. The Scythe Andy Samurai Master, on the contrary, has a heatsink with dense plates and also has an auxiliary heatsink placed right above the heat pipes. I guess this explains the difference in the noise level.
The Thermalright Ultra-120 eXtreme is the best air cooler I have ever tested in the total of its characteristics. When cooling an overclocked quad-core CPU it is only second to the Enzotech Ultra-X, but the difference is negligible. And the Ultra-120 eXtreme is better in terms of noise. The new cooler is superior to the Scythe Infinity in every parameter: it performs better by 4-5°C at peak CPU load and is more compact and does not bend the mainboard. Moreover, the Ultra-120 eXtreme doesn’t call for a second fan. Installing a second fan on it doesn’t increase its performance much and doesn’t help achieve a much higher CPU frequency at overclocking.
As a result of our today's test session, we are proud to award Thermalright Ultra-120 eXtreme with our Editor’s Choice as the today’s best processor air cooler.