Evaluating heat in vehicles' instrument panels

Various strategies can be applied to the design to address these challenges early in the design process.

First, thermal issues can be addressed when the hardware engineer creates the cluster architecture. The mechanical group then ensures the created architecture is thermally acceptable: Does it contain unnecessary high-power density areas? Is the generated power being used efficiently throughout the system?

Using computational fluid dynamics (CFD) simulation upfront allows designers to gain a better understanding of these strategies and their most efficient implementation.

Modelling the PCB assembly
With the PCB being used as an optimised heatsink for the component conduction path, it is crucial to model it accurately, especially when considering the copper content of the board.

CFD software such as FloTHERM XT seamlessly leverages an existing board layout from a CAD program such as Xpedition PCB via a *.cce file. The direct import of this file (figure 2) allows for the definition of a thermally comprehensive 3D model of the PCB. This model contains the traces in each layer based on the predefined routing as well as the placement, size, and names of each component.

Parts can be linked to an existing library of predefined thermal models, ranging from a simple block representation to a more detailed one, along with thermal networks such as a two-resistor or a DELPHI network. The components are then swapped automatically with the most appropriate fit based on the user's definition and match of either the package name or the part number.

Figure 2: 3D thermal representation of a PCB assembly (front and rear). Images courtesy of Visteon.

For this cluster design, the component thermal models used were the simple 2R and detailed models.

Figure 3 demonstrates a detailed model of the processor sitting on the cluster board. This model was provided directly by the component manufacturer. Once all the components are properly defined, the power dissipation can simultaneously be allocated to each part from the import of a *.csv file containing the power per reference designator.

Figure 3: Modelling at the component level. Images courtesy of Visteon.

Alternatively, the PCB can be modelled depending on how refined the thermal results need to be.

The CFD simulation tool can create an equivalent compact model of the board based on the copper content for each of the four layers transferred from CAD. An orthotropic in-plane and through plane conductivity is then computed based on this data.

This approach works well for the instrument cluster overall but is insufficient for components relying on copper patches to disperse their heat. In this case, the image of the traces can be used to draw the outline of the desired copper shape and extrude the corresponding copper patch (figure 4).

Placing the newly created copper pad lower in the CFD model makes it overwrite the PCB area of interest. Thermal resistance values were also defined on the edges of the patch to represent the buffer that the FR4 creates between the detailed shape and the rest of the copper plane.

Figure 4: Detailed thermal model. Images courtesy of Visteon.

   1 • 2 • 3 • 4