- Date:
- Tuesday , April 07, 2009
- Author:
- Marc Adams
- Editor:
- Kyle Bennett
- Share:
Thermal Paste Shootout - Q209
We take 10 high profile thermal pastes used for mating heatsink surfaces to CPUs and other components and subject those to over 1200 hours of testing to find out which one reigns supreme. Drop or spread? And we tried American cheese too, just in case you find yourself in a pinch.
Introduction
When it comes to setting up a new rig, most of us if not all of us, understand that each component plays an important role in the larger system. Skimping on one item can have drastic effects on another, as in the case of an inadequate power supply and a high end GPU. The same also goes for the type of thermal paste we use. Most enthusiasts know to stay away from the generic stuff that comes packaged with CPUs and heat sinks but there are also a lot of inconsistencies when it comes to retail thermal paste. "How much is too much?" "How should it be applied?" "Does it need time to reach maximum performance?" These are just a few of the questions that many of us ask. Searching the Internet only serves to make these issues even more unclear further as manufacturers and users alike all have their own ideas when it comes to dealing with thermal paste.
System Setup
We will test all thermal pastes on our tried and true QX9650 platform whose days are numbered as we move towards testing with our new Core i7 system coming on-line. Given the tremendous time we have spent testing our pastes, we just did not have the resources to scrap our previous testing once we got our Core i7 thermal testing system online. For this article our setup consists of the ASUS Maximus Formula paired with an Intel QX9650.
Test Methods
CPU
When we concluded the first roundup of CPU heat sinks way back when, Kyle and I both came to the same conclusion. Software temperature monitoring just isn't going to cut it. We need to step it up the [H]ard way. That is just what we did. We reached out to Intel about using a hardware monitor drilled into the IHS to measure the CPU temperature and wouldn't you know it, this is the exact way Intel tests their own Core 2 processors. (You can get the specs on how Intel does it starting on page 77 of this PDF document.) Now we knew we were on the right path. A few weeks later and a little planning along with some specifications and we were able to do it just like Intel does. Any advanced machine shop should be able to get you there too.
Temperatures for the thermal paste will be measured using a Sperry Digital 4 Point thermometer.
Now you may be asking why we didn’t just use a quad core CPU and continue monitoring temperatures through software. Take a look at this table. These measurements were recorded using the Intel stock cooler with the CPU at default settings. Core-Temp reports the individual temperature from each core.
With large discrepancies and variations between programs, this is why we chose hardware monitoring.
GPU
For this article the GPU will be kept at stock speed to keep any excess heat away from the CPU that could impact the results. The 9500GT is a low end card that puts out very little heat. To help keep heat away from the CPU it will be placed in the PCIe slot furthest from the CPU socket.
Heat Sink
For this review we will use Thermalright's Ultra-120 Extreme paired with a Noctua NF-P12 fan set to maximum speed. This combination gives us a solid cooling solution with low noise output. If you're wondering why we didn't use a higher output fan it's because we want there to be more pressure on the thermal paste to conduct the heat away and give us a bigger delta between pastes. A super high-output fan would minimize the differences between pastes.
Temperatures
As you will see from our charts, we recorded temperatures with no curing time and with 120 hours of curing time. Our curing process and temperature collection specifics are discussed further in the article. Ambient temperature will be kept at 25C for the duration of the tests and measured with a MicroTemp EXP non-contact infrared thermometer and cross referenced with the Sperry Digital 4 Point thermometer. Any variance greater than 0.2C during out temperature recording phase will halt the testing until temperatures return within spec for fifteen minutes.
Idle
Idle temperatures will be recorded after a fifteen minute period of inactivity. Any fluctuation during the last sixty seconds will reset the timer for an additional five minutes.
Load
Load temperatures will be recorded after a fifteen minute period of 100% load. To obtain this we will be using Prime95 v25.3. Any fluctuation during the last sixty seconds will reset the timer for an additional five minutes.