Introduction

Today I have roundup of 17 different 120mm fans for you, with a focus on finding a good noise/performance I’ve measured noise production using a dBA meter, a high end microphone and also tested each fan on a large CPU heatsink to see how much air it pushes (low/high CPU temperatures).

This is a to the point roundup offering you many facts and figures and I’ll try to sort through the results with you by splitting up the test results in different sections.

First it’s time to meet the contestants:



Specifications

These are the official specifications of the different fans, clicking on the fan name will take you to the product’s webpage.

	Airflow (CFM)	Noise (dBA)	Input (A)	Power (W)	Bearing	Height (mm)	RPM @ 12V	Mounting Holes
AC Ryan Blackfire4	77.70	28.95	0.23	2.80	Ball	25	2000	Flange
AcoustiFan AF120C	50.00	34.00	0.18	2.16	Sleeve	25	2000	Flange
AcoustiFan AFDP-12025	56.30	24.70	0.13	1.56	Dual Ball	25	1500	Flange
Aerocool Turbine 1000	37.44	19.66	0.25	3.00	Sleeve	25	950	Flange
Arctic Fan 12	56.30	23.50	0.15	1.80	Fluid Dynamic	38	1500	custom
Coolermaster ALU AAF-B12-E1	56.00	23.00	1.00	12.00	Dual Ball	38	1200	Flange
Coolermaster LED Silent TLF-S12	42.73	22.00	0.15	1.80	Sleeve	25	1220	Flange
Coolermaster Ultra Silent SAF-S12-E1	11.27	13.00	0.05	0.60	Sleeve	25	720	Flange
Coolink SWiF-1201	38.30	17.00	0.07	0.84	Dual Ball	25	1200	Flange
GlobalWin 1202512L	41.70	19.00	?	?	Nanometer Ceramic	25	1300	Rib
mCubed X12	55.00	27.00	0.30	3.60	Sleeve	25	1750	Flange
Nexus D12SL-12	36.87	22.80	0.15	1.80	Sleeve	25	1000	Rib
Papst 4412 F/2GLL	41.00	18.00	0.06	0.72	Ceramic Sleeve	25	1300	Flange
Scythe S-FLEX SFF21D	33.50	8.70	0.10	1.20	Fluid Dynamic	25	800	Flange
Scythe S-FLEX SFF21E	49.00	20.10	0.15	1.80	Fluid Dynamic	25	1200	Flange
Scythe S-FLEX SFF21F	63.70	28.00	0.20	2.40	Fluid Dynamic	25	1600	Flange
Spire FD12025C1E3	44.84	22.00	0.60	7.20	Dual Ball	25	1200	Rib

Most fans use either Sleeve or (Dual) Ball bearing, there are a few with rather unique bearings like the offerings from Scythe, Coolermaster and GlobalWin.

The difference between Flange Type and Rib Type can be seen here:


Image from eminebea

Features

If you are going to use one of these as a CPU heatsink fan you might want to know how much the fan weighs, as there can be quite a difference. If you have large case a longer cable length will give you more freedom. Most fans come with mounting screws and 3-to-4 pin adapter, but not all. To help reduce vibration noise some manufacturers are providing rubber rings, or custom rubber mounting kits. The “life” value for each fan is taken directly from the product page and has to be taken with a grain of salt. Last is the visual aspect of the fan, some come with LEDs and light up when powered.

	Weight (gram)	Cable Length (mm)	Mounting Screws	3-to-4 Pin	Anti Vibration	Life (Hours)	UV/LED
AC Ryan Blackfire4	112	300	yes	yes	no		UV&LED
AcoustiFan AF120C	116	520	yes	no	no	26000	no
AcoustiFan AFDP-12025	162	410	yes	yes	no	60000	no
Aerocool Turbine 1000	107	500	yes	yes	no		LED
Arctic Fan 12	120	400	yes	no	yes		no
Coolermaster ALU AAF-B12-E1	375	600	yes	yes	yes	100000	no
Coolermaster LED Silent TLF-S12	167	305	yes	yes	no	30000	LED
Coolermaster Ultra Silent SAF-S12-E1	172	220	yes	yes	no	50000	no
Coolink SWiF-1201	94	405	yes	yes	no		no
GlobalWin 1202512L	124	415	yes	no	no	80000	no
mCubed X12	116	400	yes	yes	yes		no
Nexus D12SL-12 (black/white)	114	410	no	yes	no		no
Papst 4412 F/2GLL	157	325	no	no	no		no
Scythe S-FLEX SFF21D	164	300	yes	yes	no	150000	no
Scythe S-FLEX SFF21E	164	300	yes	yes	no	150000	no
Scythe S-FLEX SFF21F	164	300	yes	yes	no	150000	no
Spire FD12025C1E3	132	240	yes	no	no	70000	no

Test Setups and Methodology

JMke's Test Setup
CPU	A64 3200+ @ 2200Mhz 1.5v
Cooling	Thermaltake Big Typhoon
Mainboard	DFI NF3 Lanparty
Memory	1 * 256Mb PC3700 OCZ
Other	Antec Sonata II ATI R9000 Passive Cooling Silverstone EFN-300 300W Passive Cooled PSU Maxtor 120GB IDE HDD

The tower case was placed on its side and the temperature tests were done without the side panel installed; I placed a small thermistor probe above the CPU cooler to measure in-take temperature.

in-take temperature was measured at 25°C for all tests, but temp fluctuations, and user error can account up to 1-3°C of inaccuracy in the obtained results. Please keep this in mind when looking at the results. Each fan was tested repeatedly; if I got questionable results the test was restarted.

Noise level of each HSF combo was recorded with SmartSensor SL4001A, the sensor was placed ~10cm away from the fan. The lowest dBA reading in the test room was 34dBA without fan running in the system.

System was stressed by running K7 CPU Burn for 30min ; this application pushes the temperature higher then any other application or game we’ve yet encountered. Speedfan was used to log maximum obtained temperatures.

Noise samples were recorded with a high quality directional microphone approx. 1cm away from the fan’s motor.

Official Specs vs Own Measurements

With an ambient noise of 34dBA I recorded the fan noise 12v from ~10cm while the fan was installed on top of a Thermaltake Big Typhoon. The RPM was measured through the fan header on the DFI NF3 motherboard.

	Official dBA	Noise@10cm	Official RPM	Measured RPM
AC Ryan Blackfire4	28.95	61.00	2000	1918
AcoustiFan AF120C	34.00	59.30	2000	1962
AcoustiFan AFDP-12025	24.70	53.30	1500	1520
Aerocool Turbine 1000	19.66	41.70	950	893
Arctic Fan 12	23.50	52.20	1500	1592
Coolermaster ALU AAF-B12-E1	23.00	50.30	1200	1288
Coolermaster LED Silent TLF-S12	22.00	47.70	1220	1180
Coolermaster Ultra Silent SAF-S12-E1	13.00	41.30	720
Coolink SWiF-1201	17.00	47.00	1200	1260
GlobalWin 1202512L	19.00	44.70	1300	1172
mCubed X12	27.00	54.00	1750	1607
Nexus D12SL-12 (black/white)	22.80	41.10	1000	917
Papst 4412 F/2GLL	18.00	46.70	1300	1241
Scythe S-FLEX SFF21D	8.70	40.60	800	844
Scythe S-FLEX SFF21E	20.10	46.70	1200	1223
Scythe S-FLEX SFF21F	28.00	52.70	1600	1577
Spire FD12025C1E3	22.00	46.90	1200	1214

As you can see from the results, the official dBA readings don’t quite reflect those from the manufacturer, this is of course due to the testing environment, theirs is tested in a soundproof room where the ambient noise is <15dba and the dba meter is placed 1 away at a 45°c angle. fan also hung up in mid-air – this makes it not even quite close to real operating environments where ambient noise most definitely higher, as you can see here. my result for nexus comes closest official one, only ~18dba difference ;-).

RPM wise the difference between my results and official specs is a lot smaller, the mCubed X12 gave the largest fluctuation, close to 150rpm less then officially specified. While the Arctic Fan runs almost 100rpm faster than advertised.

I’ll try to summarize the results of each fan so this review won’t be a 17-page beast. On the next page I’ll round up the 12v results for all fans, with enough details and sound clips to please the information-hungry.

name of the fan @ voltage
Small click-able thumbnails to let you know what the fan looks like as well as show any specialties	CPU temp dBA reading and RPM monitoring	a wave chart which can be clicked to download a small .mp3 noise recording
Subjective noise rating where I placed my ear next to the fan: more “+” means louder and more irritating motor noise -- other comments and remarks --

For the 7 and 5v result I’ll use a table (since you know what the fans look like by then) to display things more clearly.

So let’s start of by taking a closer look at the fans and see how they do at 12v ->