Global Sources
EE Times-Asia
Stay in touch with EE Times Asia
EE Times-Asia > EDA/IP

Manage thermal optimisation during PCB design

Posted: 09 Jul 2015 ?? ?Print Version ?Bookmark and Share

Keywords:computational fluid dynamics? CFD? simulation software? response surface optimisation? PCB?

The results from the simulations are shown in figures 5 and 6. From the results, it is impossible to determine what would be the optimal design but it does indicate the range of output ranges that are achievable. It takes approximately 30 minutes for the software to analyse the 50 designs using CFD.

The next step to fine-tuning our design is response surface optimisation.

Response surface optimisation
For each of our 50 design cases analysed in the DOE, a response surface optimisation (RSO) is performed. The RSO fits a response surface to the data and mathematically predicts the optimum design. An additional 1-2 minutes is needed to perform the RSO. For this cPCI card, the following output constraints are used for each of the RSO optimisations:
???Case 1: Maximum allowed U8 junction temperature = 100C.
???Case 2: No output constraints applied.
???Case 3: Maximum allowed U8 junction temperature = 100C. Maximum junction temperature difference between U7 and U8 = 1C.

The following are results for each optimisation case considered. Both the mathematically predicted optimum and the actual results from the CFD simulation are shown for comparison. Also included is the error in the RSO predicted optimum, which is part of the software's standard output (table).

Figure 7: U8 junction temperature vs. fin count for each fin height in a 2.5D plot.

In addition to the predicted optimum design point, the RSO calculation produces 2D/2.5D and 3D plots. Plots from Case 1 are shown here as an example. Figure 7 shows U8 junction temperature vs. fin count for each fin height in a 2.5D plot, with all other parameters set to the RSO predicted optimal value.

Figure 8: 3D plot of U8 junction temperature vs. base length and base width.

Figure 8 shows a 3D plot of U8 junction temperature vs. base length and base width. These plots are useful in understanding the sensitivity of the design to various inputs. In this example, we see that increasing the base length beyond 35 mm is much less effective as increasing the base width.

What to avoid
In the results for this case study, the RSO predicted junction temperature for U8 was 100C, which was the value used in the output constraint; however, the CFD result showed the junction temperature to operate at 102.6C. Further refinement through additional DOE simulations is recommended to satisfy this constraint. The RSO optimal base width Case 1 and Case 2 is 40 mm, which is the maximum value in the allowed range. Additional margin or an improved design might be possible if this base-width maximum size is increased. The RSO 2D/2.5D and 3D plots are essential in understanding the sensitivity of the thermal design to an individual input when allowing multiple changes to occur in any given simulation.

Essentially, this CFD optimisation process lets us know which variables are important to design around and which ones are not. Using this latest technology, we only have to run CFD once, and then we can run many simulations with different variables. It also illustrates the reasons for making a particular design choice.

About the author
John Wilson is the FlowTherm technical marketing engineer at Mentor Graphics with 10 years doing thermal/airflow design.

?First Page?Previous Page 1???2???3

Article Comments - Manage thermal optimisation during P...
*? You can enter [0] more charecters.
*Verify code:


Visit Asia Webinars to learn about the latest in technology and get practical design tips.

Back to Top