[muddocktor]

All right, here are some concrete numbers from the testing.

Test setup:

Gigabyte GA-P35-DS3L
E6750 running at 3600 MHz with 1.44 volts vcore at load
EVGA 8800GTS
2 X 1 GB Crucial Ballistix DDR2-5300 at 5-5-5-15 2T timings, 2.1 v vdimm and running at 450 Mhz
Corsair VX450 psu
The case is an old Chieftec clone of the Antec sx600 series. It has been modified for better cooling with a 92 mm Panaflo installed in the back, replacing the 80mm fan that came with the case. Also, I added 2-80mm fans to the case side that blow in roughly into the area of the vid card and pci slots.


Testing methodology:

After installing the fan to be tested, I booted it up to the desktop and left it idle for around 15 minutes. Then I started taking idle temp measurements with Core Temp 0.96.1 for at least 15 minutes to get a good average idle temp of each core. I then shut down Core Temp to separate the readings in the log and started running Prime95 version 25.6 build 6 using In-place large FFT's for maximum heat production by the processor. After starting Prime95 I restarted Core Temp. I also monitored case temps with Speedfan version 4.34 build 38, which is the latest beta of Speedfan and seems to read temps correctly with this motherboard/processor combo. I ran Prime95 for 30 minutes minimum and then shut down Core temp before quitting Prime95. I then took the resulting log files and set them into Excel spreadsheets and broke out the idle temps from the load temps and averaged each core for idle and load temps. I then combined each averaged core temps together to come up with an average idle temp and an average load temp for each fan run. BTW, this processor must have very good core to IHS contact as temps on both cores stayed very close together both at idle and at load.


Here is a list of the fans I tested and the manufacturer's posted specs of their fans.

Delta EFB1212LE
RPM-2000
Max Airflow - 75.93 CFM, 2.150 Cu. meters/minute
Max air pressure - .157 In. water, 4.00 mm water
Noise - 34.0 dB-A

Panaflo FBA12G12M-1BX
RPM - 2100
Max Airflow - 86.5 CFM, 2.45 Cu. meters/minute
Max air pressure - .1891 inches of water, 4.80 mm of water
Noise - 35.5 dB-A

Panaflo FBA12G12L-1BX
RPM - 1700
Max Airflow - 68.9 CFM, 1.95 Cu. meters/minute
Max Air Pressure - .1301 inches of water, 3.30 mm of water
Noise - 30.0 dB-A

Scythe SFF21F S-Flex
RPM - 1600
Max Airflow - 63.7 CFM, 1.8 Cu. meters/minute
Max Air Pressure - not rated on website
Noise - 28.0 dB-A

Scythe SY1225SL12M Slipstream
RPM - 1200
Max Airflow - 68.54 CFM, 1.94 Cu. meters/minute
Max Air Pressure - not rated on website
Noise - 24.00 dB-A

Noctua NF-P12
RPM - 1300
Max Airflow - 54.3 CFM, 1.54 Cu. meters/minute
Max Air Pressure - .0662 inches of water, 1.68 mm of water
Noise - 19.8 dB-A

Yate Loon D12SM-12
RPM - 1650
Max Airflow - 70.5 CFM, 2.0 Cu. meters/minute
Max Air Pressure - not rated on website
Noise - 33 dB (website doesn't show which weighting scale used for the sound rating)

Sunon KD1212PMB1-6A
RPM - 2650
Max Airflow - 132.7 CFM, 3.76 Cu. meters/minute
Max Air Pressure - .26 inches of water, 6.6 mm of water
Noise - 43.7 dB-A


Observed RPM's of the fans in bios while installed on the heatsink:

Delta EFB1212LE - 1865 RPM
Panaflo FBA12G12M-1BX - 2115-2130 RPM
Panaflo FBA12G12L-1BX - 1745 RPM
Scythe SFF21F S-Flex - 1550 RPM
Scythe SY1225SL12M Slipstream - 1323 RPM
Noctua NF-P12 - 1375 RPM
Yate Loon D12SM-12 - 1535 RPM
Sunon KD1212PMB1-6A - No RPM monitoring, 4 pin molex powered

A couple of observations here on the RPM monitoring: All fans that RPMs could be monitored on except the Delta showed a little higher than manufacturer specs. The Delta showed lower speeds than the manufacturer's claim. And the Panaflo M1BX showed more variability than any other fan, with the RPM's being read bouncing between 2115 to 2130. I don't know if it was actually varying it's speed by that or if it was just an anomaly of the RPM sensor in the fan.


Now to the meat of the matter. Here is a screenshot of the temps posted by each fan. I also made this into a chart on this spreadsheet too for you to compare them.



Here is a picture of the test rig:




Conclusion:

While the Panaflo M1BX came out as the best at cooling the processor down, none of the fans did a terrible job and temps were actually much closer than I would have thought they would be. The Thermalright Ultra 120 must be pretty efficient even with lower airflow through the heatsink than I thought it would be. There is only a 2 1/2 degree difference between the Panaflo M1BX and the Noctua NF-P12 with this combination of case, heatsink and heatload. There might be a larger temperature delta under a higher heatload though, such as an overclocked Q6600 for example. One thing that did strike me in this testing was the total lack of better cooling efficiency when going from the Panaflo M1BX to the Sunon fan. The testing showed no gain at all, with the differences between the 2 runs being in the realm of testing variation. This tells me that the Ultra 120 does not respond to higher airflow or air pressure beyond around 85-90 cfm and is a waste of time to try to run a focused flow Delta or other high cfm, high pressure, high noise fan on this heatsink.

One other thing that I will try later: Since the Noctua fan comes with some resistor adapters that will let me run it at 1100 RPM and 900 RPM, I will make a couple of runs like that and see how low the airflow and pressure can go on an Ultra 120 before starting to see a large loss of cooling efficiency.