Articles: Graphics
 

Bookmark and Share

(1) 
Pages: [ 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 ]

Cooling System Design and Performance

Besides the packaging, accessories bundles, PCB design and frequencies, all three GeForce GTX 560 graphics cards have their own unique coolers:

 

The cooler on EVGA GeForce GTX 560 Superclocked consists of a copper base, three 6 mm heatpipes coming out of it and three aluminum heatsink arrays, one of which is soldered directly to the base of the cooler and the other two hang at the ends of the heatpipes. Gigabyte GeForce GTX 560 Ultra Durable features a cooling system with four 6 mm heatpipes forming part of the base (heatpipe direct touch technology) that hold a heatsink of aluminum fins pressed firmly against the heatpipes. Nickel-plated heatsink of the MSI GeForce N560GTX Twin Frozr II-OC consists of a classical copper base, two 6 mm and two 8 mm heatpipes and thin and long aluminum fins.

The fans are also very different in each of the coolers:

  

EVGA graphics card uses one 11-blade fan 75 mm in diameter made by AVC. This particular DASA0815R2U model is based on a fluid dynamic bearing and can be PWM controlled in the interval from 1000 to 4000 RPM. Gigabyte card uses two 9-blade fans 96 mm in diameter tied together using WindForce 2X technology, which implies certain placement on the PCB and the rotation direction. The GTX 560 Ultra Durable has both fans installed at 0 angle along the same axis. They both rotate in the same direction (counterclockwise). Gigabyte uses Everflow T129215SM fans with a slide bearing. They rotate at 1000-1950 RPM. The MSI card is equipped with a pair of 75 mm 11-blade fans from PowerLogic. This PLD08010S12HH model also uses a slide bearing. Their rotation speed is PWM controlled between 1500 and 4500 RPM.

Now let’s see how efficient and noisy these coolers are and how well they cope with their task. To accomplish this, we used a test from Aliens vs. Predator (2010) game. We ran it five times with maximum graphics quality settings in 1920x1080 resolution and 16x anisotropic filtering, but without the FSAA (which allows heating up the GPUs even more):

Besides this mode we also warmed up the testing participants by FurMark version 1.9.9 stability test in 1920x1080 resolution:

We used MSI Afterburner utility version 2.2.0 Beta 5 to monitor graphics card temperatures, frequencies and fan rotation speeds. We also sued GPU-Z version 0.5.4 utility. All tests were performed inside a closed system case with the ambient temperature at 27.5-28.0°C.

First let’s check out the temperatures of our tested cards during the gameplay with automatic fan rotation speed control:


EVGA

Gigabyte

MSI

Now let’s continue with the tests in the same gaming mode with the fans at their maximum rotation speed:


EVGA

Gigabyte

MSI

Now we will repeat the test one more time with the load created by FurMark in automatic fan mode:


EVGA

MSI

…and at maximum fan rotation speed:


EVGA

MSI

The FurMark charts have no Gigabyte results on them for a reason. The thing is that this particular graphics card behaved really strange: three minutes into the test the temperature almost fully stabilized at 74°C, which is lower than by the other two participants, but then the card suddenly stopped sending the signal to the monitor and it shut down. We suspect that one of the electronic components on the PCB or maybe even a specific unit of the GPU could get overheated (which may often happen with heat-pipe direct touch coolers). As a result, we couldn’t check the cooling efficiency of the Gigabyte cooler under ultimate load.

All other results are summed up on the diagram below for your convenience:

The cooling efficiency of three GeForce GTX 560 graphics accelerators reviewed today turned out pretty predictable. The EVGA cooler works very well, but in automatic mode it yields to the other two competitors. At the same time, GeForce GTX 560 Superclocked remained the coolest graphics card at maximum fan rotation speed. Gigabyte GeForce GTX 560 Ultra Durable product with two largest fans could have taken the leader’s crown in the efficiency test, if it hadn’t failed the FurMark test. Note that the maximum rotation speed of its fans doesn’t exceed 2000 RPM. As for MSI, their Twin Frozr II once again proved superb, that is why it will be available for purchase separately very soon. However, it is impossible to draw the final verdict about the graphics card coolers efficiency without checking out their acoustic readings. So, let’s move on to the noise tests.

The noise level of each cooler was measured between 1:00 and 3:00 AM in a closed room about 20 m2 big using CENTER-321 electronic noise meter. The lowest noise reading our noise meter device can register is 29.8 dBA and the subjectively comfortable noise level in these testing conditions was around 36 dBA (do not mix it up with low noise level). The noise level for each cooler was tested outside the system case when the only noise sources in the lab were the cooler and its fan. The noise meter was installed on a tripod and was always at a 150 mm distance from the cooler fan rotor. The tested cooling systems were placed at the edge of the desk on a sheet of polyurethane foam. The fans rotation speed was adjusted in the entire supported range with the new controller by changing the voltage with 0.5 V increment.

Here are the obtained results (the dotted lines show where fans worked in automatic mode during the gaming load tests):

Overall I can conclude that Gigabyte’s graphics card ahs the quietest cooler of the three, which is exactly what we have expected judging by its design and its fans speed modes. In 2D mode you can’t hear Gigabyte fans at all, just as the EVGA cooling fan. However, we can’t say the same about the MSI fans: the noise from two 75 mm fans rotating at 1800 RPM can be clearly distinguished against the background of a quiet system case. Although in my subjective opinion, during the gaming tests Twin Frozr II works in a more acoustically comfortable noise range than the EVGA cooler.

We performed all overclocking experiments using the graphics cards’ proprietary coolers and without changing the GPU voltage. The results showed that EVGA GeForce GTX 560 Superclocked can remain stable at up to 930/1860/4640 MHz, Gigabyte GeForce GTX 560 Ultra Durable – at up to 960/1920/4880 MHz and MSI GeForce N560GTX Twin Frozr II-OC reached the maximum frequencies of 940/1880/5020 MHz:

  

This is a pretty typical GeForce GTX 560 overclocking result. Gigabyte graphics card managed to achieve the highest GPU frequency, while MSI card hit the maximum memory speed. Of course, it is important to remember that overclocking success depends a lot on a specific graphics card piece.

The temperatures of all three testing participants didn’t change much during overclocking (they only gained about 2°C in GPU temperature):


EVGA

Gigabyte

MSI

However, it only happened because the graphics cards cooling fans sped up in automatic mode and prevented the temperature from growing any higher.

Temperatures in SLI mode seem to be much worse:

As you can see, but the GPU of the Gigabyte graphics card located at the top of the 2-way SLI bundle heated up to 100°C during common gaming tests, even though this card has a very efficient cooler, as we have just seen above:

Of course, it could be the summer heat or the closed system case, or the close proximity of the second graphics card or the sound card in the system. But aren’t these the usual conditions. In which our system components normally work? Besides, not everyone has a roomy system case with well-organized internal ventilation, like Antec Twelve Hundred, and a versatile mainboard with multiple options for PCI-Express slot configurations like Gigabyte GA-X58A-UD9. It is obvious that you need graphics cards with airflow directed along the PCB and then pushed out of the system case if you want to have you SLI or CrossFireX systems running smoothly. A liquid-cooling system could be an even better alternative, but this would be a topic for a completely different article.

 
Pages: [ 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 ]

Discussion

Comments currently: 1
Discussion started: 07/26/11 11:48:32 PM
Latest comment: 08/02/11 12:52:15 PM

View comments

Add your Comment