CoolIt Systems Freezone Peltier CPU Cooler Review

Testing

Testing an active TEC based cooler is more difficult than testing a standard heatsink fan or water-cooling system.  I tested the Freezone CPU Cooler at two different ambient air temperatures to see how the cooler would perform with relatively cool temperatures (21°C) and how it would perform under very warm ambient air conditions (36°C).  On the cool side, I wanted to see if the chiller would get too cold and possibly allow condensation to form.  And on the warm side, I wanted to know if the Freezone would be able to cool my test bed’s red-hot 955 CPU.

The Freezone CPU Cooler was tested on an open frame test rig consisting of the components listed below.  The ambient room air temperature was maintained at 21°C ±0.5°C during the first series of tests and artificially increased to 36°C for another set of tests.  The CPU was loaded by running multiple instances of CPUBurn. Tests were conducted with the TCM potentiometer set for three different cooling levels: full CCW (maximum cooling), full CW (minimum noise), and ½ way in the middle.

Test Rig Configuration

• Asus P5N32-SLI Deluxe motherboard
• Pentium 4 Extreme Edition dual core 955 @ 3.46 GHz
• (2) Corsair CM2X512-8000UL DDR2
• (2) nVIDIA 7800 GTX 512 MB video cards in SLI
• Western Digital WD1200JD S-ATA HDD
• SilverStone Zeus ST75ZF 750W PSU
• Windows XP Pro with SP2
• nVIDIA 91.31 nForce driver

(click to enlarge)

Standard Test Setup (21°C Ambient Air Temperature)

The key measure of a cooler’s performance is how well it can cool the CPU.  This is typically measured by subtracting the ambient air temperature from the CPU temperature, which results in a differential temperature (referred to as the Delta T).  The closer the CPU temperature is to the ambient air temperature the better (smaller Delta T).  And when some form of active cooling is employed, there is the potential for temperatures to drop below the ambient temperature (-Delta T). 

The performance of most traditional heatsink fans and water-cooling systems that rely on convection cooling changes linearly with the ambient room air temperature (at least within the temperature range of most CPU operation).  If the ambient air temperature goes up 10°C, then so will the CPU temperature (Delta T remains the same) and vise versa.  However, with active TEC cooling this is NOT the case.  Testing the true potential of the Freezone CPU cooler required a slightly modified testing setup… 🙂

(click to enlarge)

Modified Test Setup (36°C Ambient Air Temperature)

In this configuration, a 4′ duct is used to route the warm exhaust air coming out the back of the Chiller/Pump Module around to the front where it is sucked back into the central heatsink.  By carefully positioning the duct, just the right combination of warm recycled air and cool ambient air can be mixed to simulate a much warmer operating environment.  Doing this to a normal HSF would kill its performance, but when I fed the Freezone warmer air it just kept on chillin’… 🙂

Test Instruments

To facilitate taking more accurate CPU temperature measurements, a very small Omega thermocouple was attached to the side of the P4-955 IHS with Arctic Alumina thermal epoxy.  I also placed a thermocouple on the top chiller waterblock (next to the Freezone’s temperature sensor) and placed two more thermocouples around the chiller to measure the air temperature going into (ambient) and coming out of the central heatsink.  A FLUKE ammeter was used to measure the total current going to all six TECs and a FLUKE voltmeter was used to monitor the voltage sent to the cooling fan by the TCM.

• CPU/IHS — Barnant Model 115 digital thermometer (accuracy +/- 0.4º C)
• Air-in — FLUKE 52 II digital thermometer (accuracy +/- 0.05% + 0.3º C)
• Air-out — FLUKE 52 II digital thermometer (accuracy +/- 0.05% + 0.3º C)
• Water — Barnant Model 115 digital thermometer (accuracy +/- 0.4º C)
• TEC current — FLUKE 87 III digital multimeter (accuracy +/- 0.2% of reading)
• Fan voltage — FLUKE 87 III digital multimeter (accuracy +/- 0.05% of reading)
• Extech Model 407736 digital sound level meter (accuracy +/- 1.5 dB)

Software Utilities

• Lavalys Everest Ultimate Edition 2006 (hardware monitoring)
• Asus PC Probe II (hardware monitoring)
• CPUBurn (load the CPU)

Note: Four instances of CPUBurn were executed at the same time (two to load both physical cores, and two to load the two virtual HyperThreading cores), which resulted in 100% CPU usage.

The following data is presented for comparative purposes only.  Your actual results may be different depending on the variables unique to your system (CPU, overclock, ambient temperature, case air flow, temperature monitoring, etc).   Tests were performed under two different sets of testing conditions: first with a 21°C ambient temperature and then with the ambient temperature artificially raised to 36°C.

Amb. (°C) — Ambient air temperature

CPU-IHS (°C) — Temperature obtained with calibrated thermocouple attached to IHS

∆T (°C) — Fully loaded Tc temperature rise above ambient

Fan (V) — Voltage sent to cooling fan by TCM

SPL (dBA) — Sound pressure level recorded 3′ away (background ~30 dBA)

Water (°C) — Water temperature measured at chiller waterblock

TEC (A) — Current flow to TECs (controlled by TCM)

Air-in (°C) — Air temperature going into chiller heatsink

Air-out (°C) — Air temperature coming out of chiller heatsink

As you can see, the performance of the Freezone CPU Cooler was very good during the first set of tests with an ambient room temperature of 21°C.  With the TCM adjusted to minimize noise (full CW), the system was very quiet and still kept the CPU relatively cool.  Cranking the TCM up for maximum cooling resulted in significantly more fan noise along with excellent cooling.  At idle, the 955 CPU temperature actually dropped a full degree C below ambient, which isn’t enough to cause concern about condensation.  Note the average water temperature was 16.8°C, which is 4°C below ambient.  This is the coolest temperature I recorded and again, does not impose much of a risk for condensation.

Artificially increasing the ambient room air temperature to 36°C also resulted in excellent cooling.  Remember, a traditional HSF or water-cooling system would yield significantly higher average water and CPU temperatures.  But thanks to active TEC cooling, the Freezone system kept the average water temperature well below ambient and the CPU temperature within 4.3°C of ambient!  The chiller fan noise does become very noticeable at high speed.  Worst case was 44.7 dBA at full voltage (3,200 rpm) measured 3′ away with an ambient background noise level of 30 dBA.

The key thing to note here is that the Freezone CPU Cooler operates more or less independently from the ambient air temperature.  Even with very warm environmental conditions, the Freezone just keeps on chillin’.