120mm Fan Roundup - 35 Models Compared

Cooling/VGA & Other Cooling by jmke @ 2009-02-13

We compare the performance/noise balance of 35 different 120mm fans from AcoustFan, Aerocool, Arctic-Cooling, Cooler Master, Coolink, CoolJag, Delta, Gelid, GlobalWin, mCubed, Nexus, Noctua, NoiseBlocker, Papst, Revoltec, Scythe, Sharkoon, SilenX and Xthermal.

  • prev
  • next

Test Results

Test Results

When faced with almost 300 results, coming from dBA meter readings, CPU temperatures and subjective motor noise rating we discussed how to best present the obtained data to our readers.

On the previous page you’ll find the raw data sheet for those who’re interested. Balancing the noise and performance results we decided that for most users out there a quieter fan is more important than a high performing fan. Meaning: a drop of 1dBA should be more important than a 1°C CPU temp drop.

We came up with the following formula:

Efficiency Rating = (10000 / ((2/3 * CPUtemp) + dBA))) – (MotorNoiseRating)

Subjective Moto Noise Rating can be found on previous page in the data table, to penalize those fans which are really irritating and definitely not silent, those rated 3 & 4 had their scores doubled so their Efficiency Rating drops noticeably; the dBA meter and CPU temp tells us some fans did fairly good, but subjective noise rating detected very noticeable and irritating motor noise, resonance.

After running this formula on the results, the Efficiency Rating gave us a single value between 92~140; the higher the score the better the fan is at delivering good performance at low noise levels. The subjective weighing is just that, subjective, so in chart below we provided both values, with/without the Motor Noise Rating deducted from the Efficiency Rating.

Here are the results:

Madshrimps (c)

For those who like a more classic X/Y scatter chart:

Madshrimps (c)
(click image for huge edition, 1577x1451)

Not very clear…
  • prev
  • next
Comment from Rutar @ 2009/02/15
How comes the Silent Eagle performs so badly compared to the previous roundups, especially in the noise/temperature ratios? Did fan tech advance that much?

It would be nice to have multiple scatter graphs with different groups of performers so differences can be seen better. I still prefer them over the MER.
Comment from jmke @ 2009/02/15
The Sharkoon Silent Eagle is better known as the « golf ball » fan, as the surface of its fan blades look exactly like that. We’re testing the 2000rpm model here, unfortunately the 1000rpm model we no longer have in our labs. Sharkoon does not label their Silent Eagle model so we added our own “2000rpm” tag to the label.

2000 model best effort is starting point of the 1000rpm model... I don't have the 1000rpm model in house anymore, otherwise would have included it

regrading scatter chart: what groups would you make?
Comment from Rutar @ 2009/02/15
Still, here the SE 2000 RPM did excellent vs the Globalwin which beat all other fans in the 17 fan review. The Noctua S12 1200 outperformed it with 51°/42.9 dBa vs 50.5°/43.6 dBa when there were -3.8°/+0.9 dBa difference in the 4 fan article.

I think it could have to do with the fin spacing difference between the Infinity and the Zipang and the fact that it wasn't mounted directly on the heatsink in the first roundups. I think at least the difference due to the fin spacing should be ruled out and the conflicting results explained.

Eventually 2 test methods have to be used to find out which one is the best casefan and which one is the best CPU fan.