What's the best 120mm case fan?

Introduction

Although the choice of heatsink or waterblock/radiator has a huge effect on the temperature of the components they cool, without a suitable fan blowing air over said components and fans mounted in the case, your PC will soon overheat, wither and die.

Fortunately, bit-tech is here with a well-timed group test of 120mm fans, ideal for cooling and reducing the noise signature of an existing PC, or including as part of a new build. We've tested 22 120mm fans, so that we can give a definitive verdict on which fans are best for either an air-cooled PC or for a 120mm-wide radiator.

We've scientifically measured the performance of each fan by measuring its airflow and noise. Airflow is crucial, as this figure determines how effective the fan is at moving air through your PC, while a noisy fan is an irritant everyone could do without. So, to find out which fans are worth installing in your PC and which are best left on the shelf, read on!

How we Tested

Unlike most other PC components, a fan only has to perform a single task, which is to move air through your PC to help keep the components cool. Whether the fan is mounted on a CPU heatsink or radiator, blowing over the motherboard or acting as an input or output chassis fan, moving air is all that it has to do.

However, human beings are slightly more complex and demanding than your average fan. While a fan may be content with simply moving air around, we’d also like it to do this quietly. So, for this group test, we devised a method of measuring both the airflow and noise of a fan to find out which fans provide the ideal balance between the two.


Measuring airflow accurately is one of the hardest tests to perform, as the most common measuring device, the anemometer, only measures the air flowing over its surface area. This is a problem, as the velocity of the air leaving the fan will vary depending on how close the anemometer is to the centre of the fan. To measure the volume of air moved by the entire fan, not just a small proportion of it, we used the so-called ‘black sack’ method, espoused by several PC cooling hardware and air-conditioning manufacturers.

This incredibly simple device, which costs around 12p to assemble (compared to several hundred pounds for a decent anemometer) acts as a funnel that directs the airflow from a fan into a black rubbish sack. The volume of the sack is a known quantity, so by timing how long it takes each fan to fill the bag with air, we can calculate the airflow using the following formula:

This tells us how many cfm (cubic feet per minute) of air the fan is moving. Using this method also allowed us to compare the airflow of fans with different diameters, simply by adjusting the size of the hole through which the fan blows air into the bag.

Obviously, the ‘black sack’ method introduces a degree of human error, which is why we tested each fan three times, and used the average time (to the nearest second) to calculate the airflow. Also, even though we used a black sack made from lightweight plastic, it was still too heavy to fill completely for the fans that have a particularly low airflow or low static air pressure, so we were unable to measure their airflow accurately. The lowest-performance fan that completely filled the sack was the Revoltec RL037, which has a calculated airflow of 7cfm; therefore, any fan that couldn’t fill the bag provided a level of airflow that was less than this.

How we Measured Noise Levels (or, More Accurately, Sound Pressure)

As few fan manufacturers state specifically how they measure sound level, it would be extremely foolish to compare the claimed figures of each manufacturer. Obviously, the only way we can really compare the sound level of each fan is by testing them in exactly the same way. For this reason, we sent all the fans to the experts at Intertek, a company that specialises in comparative testing of consumer electronic devices.

Testing was carried out in Intertek’s home-cinema listening room, which is designed to meet the IEC 286-13 standard. The measurements were taken on a calibrated Brüel & Kjær 2260 sound level meter positioned 50cm away from the intake side of the fan being tested. We chose a distance of 50cm, as this is the typical distance that most people will sit from their PCs at home.

Also, if we took the measurement from any closer than 50cm, the results would be distorted. This is incredibly important, as the further away the sound level meter is from the sound source then the lower the reading will be. The fans were powered by a passively cooled PSU, so as not to increase the sound level in the room, and held in place by two metal clamps with a reflecting surface behind.

The sound pressure level of each fan was recorded between 80Hz and 20KHz over a 10-second period. The background noise level of the room averaged 19.3dBA during the test period, although it dropped below this level several times. For this reason, any fan that measured less than 19.3dBA has been graphed with a sound level of 19.3dBA.

After testing all the fans, it can be said that any fan producing 25dBA or less was barely audible, while fans emitting between 26dBA and 30dBA were audible but not annoying. Fans measured above 31dBA were perceived to be too loud for comfort.

Not shown on the graphs, but also important, were the subjective observations we made about any particularly unpleasant noises produced by the fans. This is crucial, since the motors of some fans can produce sound at irritating frequencies, which can be very distracting, even if they’re not particularly loud.

How We Scored the Fans

Each fan has three separate scores: airflow, acoustics and value, which combine to provide the overall score.

The airflow and acoustics scores are derived from the measured airflow and noise levels of the fans, while the value score is calculated by dividing the airflow, acoustics and features score by the price.

The features score takes into account the accessory bundle, awarding points for power adaptors, resistor cables, rheostats and whether the fan is mounted with traditional metal screws or anti-vibration rubber fastenings.