Experimental review: Influence of fans speed on water cooling loop performance
By Daz, April 19, 2010
Introduction
The purpose of this review is to find out temperature changes dynamics related to changing rotating speed of water cooling radiator fans.
1. Lab system setup
For this review we have used Intel’s Core i5 750 based system overclocked to 4Ghz.
Hardware specs:
- Processor: Intel Core i5 750
- Main board: EVGA P55 LE
- Memory : 2x 2Gb G.Skill DDR3 9-9-9-24 1.5v
- Graphic card: ATI Radeon 4870 512MB
- Hard drive : 1 x 1.5T Seagate 7200rpm
- Power Supply: Seasonic 850W
- Operating System: Windows 7 64-bit
Overclocking BIOS specs:
- Multiplier @ 20
- Clock @ 200
- VCore @ 1.45v
- VTT @ 1.325v
- PCH @ 1.9v
- C-states: enabled
- RAM @ 1.55, Auto timing
Water cooling specs:
Our water cooling setup is a single CPU loop with externally mounted radiator with fans attached in pull position.
- Radiator: Feser TFC240 High Performance Radiator
- Fans: 2x San Ace 38x120mm 3000 RPM w/ sensor
- CPU block: EK Supreme LT, Acetal Top
- Pump: D5 Vario with EK D5 X-Top Rev.2, Max speed (5)
- Tubing: TFC 3/8”ID hose
- Reservoir: EK Multi-Res 100
- Coolant: Feser One Clear/UV Blue
Sensors and instruments:
In our test for each fixed fans speed we measured CPU cores temperatures, air temperature before entering radiator and after exiting radiator, coolant temperature before entering CPU block and after exiting CPU block. Each speed setting was run for 1 hour starting from 3000rpm speed.
- Koolance TMS-200 Management board
- Koolance Coolant Temperature Sensors (2)
- Koolance Temperature Sensors (2)
- Koolance FM-16 Flow Meter
- CPU Stress Program: OCCT 3.1.0
- CPU-Z 1.54
- Core Temp 0.99.5.27
2. Test Results
CPU Cores Temperature readings. Overclcok 4Ghz, ambient 23C.
Graph shows CPU cores temperatures in degrees C (Y-axis) vs. fans speed in RPM (X-axis).
Coolant Temperature readings. Overclcok 4Ghz, ambient 23C.
Graph shows coolant temperatures in degrees C (Y-axis) vs. fans speed in RPM (X-axis).
Air Temperature readings. Overclcok 4Ghz, ambient 23C.
Graph showsair temperatures in degrees C (Y-axis) vs. fans speed in RPM (X-axis).
3. Final Remarks
First of all, we need to note that our system kept failing at fans speed of 450rpm. The readings represented for 450rpm speed are the last results before system failure and most likely temperatures should continued to rise if system stayed on. Also, it was quite obvious that at that slowest speed radiator couldn’t handle the load and started to heat up dramatically affecting intake air sensor. For that reason we consider lowest acceptable fans speed is 720rpm for 38mm fans for our particular setup; the 25mm fans would need slightly higher speed to maintain the same radiator performance.
The largest performance gains were observed in 720-1700rpm range, which is good news for people who are looking for silent water cooling set up. The noise levels are quite bearable for speeds up to 1400-1500rpm and it will suite comfortably majority of water cooling enthusiasts.
The performance gains past 1700rpm range are not as dramatic with coolant temperatures dropping 1C with each speed increase in our test. Those gains still can be quite handy for adding more blocks to CPU loop or higher overclock.
As a side note, this test shows that dual thick radiator is quite adequate solution for an overclocked single Core i5 CPU loop.
You can discuss or make comment HERE