Taking direct airflow measurements from a PCB during operation is infeasible due the small size of most boards.
Because measurements are basically impossible to gather, CFD analysis relies on simulations to evaluate the effectiveness of a cooling strategy.
With transient and steady-state simulations, you can quickly evaluate how your cooling strategy affects the temperature distribution in your electronics.
CFD analysis and simulation can help you determine how to remove heat from this IC during operation.
Anytime you need to evaluate a production board, you need to gather some measurements and check your data against your design requirements. Electrical measurements need to be gathered at multiple points during fabrication and assembly, as well as during field tests. Once you get your measurements and compare them with simulation data, you can quickly identify problem areas and propose some redesigns for your product.
One set of measurements that are very difficult to gather in small devices are airflow measurements. Far away from the board, airflow measurements can be gathered, but accurately regressing back to airflow behavior close to small electronics does not yield accurate results. The problem with gathering accurate measurements motivates the use of simulations to see what happens in the interior of integrated circuits, as well as how airflow moves heat away from components. Before and after you perform CFD analysis, use an integrated simulation tool to examine heat transfer away from hot components and qualify your cooling strategy.
Why Use CFD Simulations?
To properly evaluate your cooling strategy, you need to examine how heat moves throughout the board and, ultimately, the steady-state temperature during operation. Because of the difficulties in testing, you’ll need to rely on some simulations to examine how the steady-state temperature is reached. This gives you a more complete view of airflow and heat transfer within your PCB and components.
Why use simulations at all, rather than measurements at the boundary of a system near a PCB? If you’re using cooling fans or relying on convection to remove heat from your system, it is true that as long as boundary conditions are well defined, airflow measurements at the boundary of a system can be used to infer the airflow inside the boundary. However, these techniques are only accurate for laminar flow, while they are highly sensitive to boundary and initial conditions for convective and turbulent flow.
Because of these problems, massive amounts of data are needed for estimating sensitivity and developing averaging methods. In addition, using these techniques in thermal-CFD co-simulations involves decoupling heat transfer from fluid flow, which creates a non-physical problem. Because of decoupling, fluid flow can only be examined at the steady-state temperature, which does not yield useful results for a real system. A better approach is to use a commercial-grade 3D solver for CFD problems. These problems can be time-consuming in complex systems, but they give very accurate results that can be used to qualify a design.
Getting Your CFD Data
To get the data you need for simulation-based CFD analysis, here are the simulations you need to use:
Transient simulation: This shows you how the temperature of a system changes over time as it approaches the steady-state, assuming time-independent heat sources and sinks. If your sources and/or sinks suddenly change at some point in time, you can also examine how airflow carries heat away from the system, as well as how this affects the approach to a new steady-state temperature.
Steady-state simulation: Simulating the equilibrium temperature directly reduces the number of independent variables in a CFD simulation from 4 to 3. This then reduces the simulation time and allows more accurate simulation near complex geometries. Note that this requires defining steady (i.e., time-independent) sinks and sources in the system.
Steady-state simulations can show you the temperature distribution you might measure with an infrared camera.
CFD Simulation Inputs
The basic inputs into any CFD simulation are the system geometry, material parameters, boundary conditions, initial conditions, sources, and sinks, both for temperature and fluid flow. For more complex simulations involving airflow, density can also be defined as an initial or final state, although these simulations are much more complicated and are not used in electronics design. Sources and sinks can be quite complex and define how heat/airflow is generated/removed from the system.
How to Use CFD Simulation Results for Analysis
Simply running a CFD simulation is not enough to qualify your system’s behavior and functionality. You’ll need to go further and analyze the velocity and temperature fields in your CFD simulation results. When you examine your simulation results, airflow is normally shown with striplines, while the temperature field is normally shown as a heat map. These sets of data are shown in 3D to give an idea of the temperature distribution in the PCB and how it relates to airflow.
Going further, it’s important to analyze how airflow carries heat away from the system and components, as well as whether airflow even reaches components. This where both types of CFD analysis simulations become useful for evaluating your cooling strategy in your system.
Example Results in a Power Supply PCB
Looking at streamlines and heat flow together in 3D CFD simulation results helps reveal the correlation between airflow and heat transfer. In the power supply PCB below, we can see that airflow helps cool components in the back-right corner, as illustrated from the warm airflow streamlines along the back of the board. However, the large resistor in the front-left corner does not receive much airflow and dissipates significant heat into the substrate.
Streamlines and heat flow shown together in 3D CFD results.
Similar results could be seen in a transient simulation, which will show how the temperature and airflow fields develop over time. This can help you evaluate whether airflow needs to be increased in the system to help bring down temperature during operation. It can also help you identify whether components should be moved slightly to aid cooling, such as the situation shown above. All this can be done before field tests with the right simulation tools in your PCB design software.
Anytime you need to examine heat transfer and airflow in your board, you need simulation-based CFD analysis tools. Additionally, you’ll want to be able to perform transient and steady-state CFD simulations. These processes will add a level of confidence and security to your designs you wouldn’t know otherwise