Testing

Test Setup:
Asus P4P800 SE, Pentium 4 2.8C at 2800mhz, 2x512mb Corsair XMS 3200 Pro, ATi Radeon 9600XT, CoolerMaster Hyper6 heatsink fan on Low.

Since we projected only a slight (1-4° C) difference between the two compound's average temperatures we made sure the surroundings we kept as consistent as possible. To do this the test room was completely shut down and kept at 24.5° C (76° F) due to fluctuations beyond our control there was a ±2° C. Humidity was kept constant by moving a dehumidifier from my basement and putting in in the test room.

Arctic Silver tells us that a thermal compound should be allowed to set for about 200 hours before it will be able to be completely set and the thermal conductivity will be at its peak. Unfortunately we could not wait 8+ days for each test cycle so we burned in each installation at 100% CPU utilization for 48 hours and then let it run for 24 hours at idle. This means each application was able to set for at least 72 hours before it was tested.

Test Setup:
Asus P4P800 SE, Pentium 4 2.8C at 2800mhz, 2x512mb Corsair XMS 3200 Pro, ATi Radeon 9600XT, CoolerMaster Hyper6 heatsink fan on Low.

Since we projected only a slight (1-4° C) difference between the two compound's average temperatures we made sure the surroundings we kept as consistent as possible. To do this the test room was completely shut down and kept at 24.5° C (76° F) due to fluctuations beyond our control there was a ±2° C. Humidity was kept constant by moving a dehumidifier from my basement and putting in in the test room.

Arctic Silver tells us that a thermal compound should be allowed to set for about 200 hours before it will be able to be completely set and the thermal conductivity will be at its peak. Unfortunately we could not wait 8+ days for each test cycle so we burned in each installation at 100% CPU utilization for 48 hours and then let it run for 24 hours at idle. This means each application was able to set for at least 72 hours before it was tested.

As you can see from chart above the temperature changes are slight but noticeable. At Idle there is a 1° difference between the two compounds. At load (measured after 1 hour) the difference is more pronounced- there is an average difference of 2.5° C.

Also you can see that the difference between the Idle and Load temperature or AS5 is 14° C while that of the Ceramique is 15.5° C. This further proves that AS5 is a more efficient thermal compound.

Our testing has concluded that Arctic Silver 5 is a more effective conductor of heat that Ceramique. This is most likely due to the conductive capabilities of silver over that of aluminum oxide (etc.) of the Ceramique. What we did notice is that how the compound is installed is very important. When applied the compound cannot be too thick or thin or have air pockets in it or there will be a disruption in the thermal conduction. This will ultimately lead to less efficient cooling and higher CPU temperatures.