CPU cooling is a critical aspect of a functioning computer system, and for this reason,
forced-air cooling is a significant factor that should be determined at an early stage in system design. Good airflow to heat-generating components, and adequate space and power for the cooling fan are a critical design requirement for any system. One of your first steps in system design should be to estimate the required airflow. This will depend on the heat generated within the enclosure, and the maximum temperature rise permitted.
The question becomes, how can we best determine the optimal design for computer cooling, particular, for CPU cooling? The answer comes in the use of thermal simulation.
Using thermal simulation technology, we can anticipate design flaws, and evaluate the behavior of the various components that will ultimately become an important part of the final computer cooling design.
The following is a quick view of how thermal simulation can assist in predicting and improving the airflow of an average CPU application in order to produce optimal CPU cooling.
The structural design of an average CPU will look like this:
The axial fan sits on top of the cover of the heatsink structure. Under the heatsink structure you will find the slug (in yellow) and the contact area to the CPU surface.
Between the heatsink fin tips and the cover it is necessary to provide a gap in order to allow air leaving the axial fan.
With a thermal simulator we can see the distribution of the surface temperature on the heatsink and CPU.
The highest temperature is seen directly at the contact area between the CPU and heatsink.
The heat is transported inside the heatsink material via conduction and from the surfaces to the air via convective heat transfer. Depending on the balance between conduction and convection, a temperature profile inside the heatsink material is observed.
As the cooling air from the fan does not provide a uniform flow pattern, there are regions in the heatsink which show a remarkable temperature field. This information provided by the thermal simulator can reveal exactly what computer cooling measures should be taken for optimal results.
In general, if high temperature gradients are observed in heatsinks, the conduction effect is subject to improvement. This can be done by choosing another material (alloy) with better conductivity or larger cross sections for the heat conduction mechanism.
In the following graphic, the distribution of the surface temperatures is shown.
Looking at the flow vector field in a center cross section, above left, the thermal simulation reveals a zone below the fan motor with no flow condition can be identified. This behavior is seen for all axial fans used in this type of application.
Thermal imaging shows the highest temperature directly at the contact area between CPU and heatsink.
The heat is transported inside the heatsink material via conduction and from the surfaces to the air via convective heat transfer.
As a consequence of still air below the fan motor, no heat is removed from the heatsink fins in this area and no cooling occurs. The temperature is increasing as shown on the right hand side of the graphic above.
Many factors, including material, fin design, air velocity and surface treatment, will influence the thermal performance of a heatsink.
Obstructions in the airflow path cause static pressure within the enclosure. To achieve maximum airflow and ideal cooling conditions, obstructions should be minimized. However, obstructions in the form of baffles may be necessary to direct the airflow over the components that need cooling.
Your final system design should show a continuous airflow and optimal thermal management for all your heat generating components.
The use of thermal simulations at NMB has become part of our design offerings, and has been proven to facilitate the design process of many of our customers by providing key information and solutions for their thermal cooling applications.