An Electronic Designer's Guide To Thermal Management

March 17, 2005

DESIGN-TEAM ORGANIZATION IS CRITICAL FOR SUCCESS: Thermal management must start with the beginning of the project. It can't be as effective if it's started further downstream in the design. Too often, initial component selection is predominantly an el

Sam Davis 2

DESIGN-TEAM ORGANIZATION IS CRITICAL FOR SUCCESS:

Thermal management must start with the beginning of the project. It can't be as effective if it's started further downstream in the design.

Too often, initial component selection is predominantly an electrical engineering function. The packaging engineers then have to "make" it work, sometimes employing expensive solutions.

THERMAL-MANAGEMENT DESIGN OPTIONS START WITH COMPONENT SELECTION:

Component selection can make a significant impact on the thermal-management solution, so it should be considered early in the design.

Printed-circuit-board layout is critical. Copper thickness, via locations, and power plane designs often dictate the solution's performance.

Exploit component supplier support and design experience to improve system performance and lower design costs.

DESIGNS WITH FANS:

Convection heat transfer (fan airflow) covers a broad range of heat flux capacity, depending on the employed technique.

To optimize performance, select and design heatsinks and fans together, early in the design.

A given fan can deliver only one flow and one pressure in a given system.

Fan flow and pressure dictate the performance of the entire convective system.

Entertaining alternate techniques, such as heat pipes, confined flow, and closed-circuit liquid cooling, can reduce total system cost in some high-performance products.

DESIGNS WITH HEATSINKS:

The fin manufacturing process determines the cost and the vendor's ability to optimize the heatsink for the user's specifications.

If the fin efficiency is already 80% or better, you can make minimal improvements at best.

Often, modifying the fin geometry can improve performance.

Serrations boost fin performance by 10% to 20%, but add cost.*

Rounded geometries outperform similar sharp-edged fin shapes.*

Staggered fins outperform in-line fins.*

Interrupted or pin-based fins outperform parallel plate fins.*

In low-pressure drop systems, elliptical fins work best.*

In high-pressure drop systems, round pins offer the best performance.*

DESIGNING WITH TIMS (THERMAL INTERFACE MATERIALS):

As power levels and power densities rise, selecting the optimum interface material is as important as the optimum heatsink and fan.

Achieving optimum TIM performance requires tradeoffs between its thermal conductivity, rigidity, and thickness.

The optimum TIM solution will result in the selection of the highest thermal conductivity material, which enables the largest contact area with the thinnest bondline.

First-order interface material selection should be from the data sheet. However, selecting the optimum TIM requires empirical testing due to multiple interrelated variables.

IMPROVING OVERALL THERMAL SOLUTION PERFORMANCE MAY REQUIRE ONE OR ALL OF THE FOLLOWING:

Reconsider component selection and packaging, which can open up new avenues for heat removal.

Optimize the pc-board geometry and layout.

Engineer the fan and fin components for optimum per formance.

Entertain new cooling techniques that ultimately may lower system cost.

Empirically test TIMs to achieve optimum performance.

Improve power-management efficiency.

Source: M. Behnia, D. Copeland, D. Soodphakdee, "A Comparison of Heat Sink Geometries for Laminar Forced Convection:Numerical Simulation of Periodically Developed Flow," The Sixth Intersociety Conference on Thermal and Thermomechanical Phenomena in Electronic Systems, ITHERM '98, Seattle, Wash.

About the Author

Sam Davis 2

Sam Davis was the editor-in-chief of Power Electronics Technology magazine and website that is now part of Electronic Design. He has 18 years experience in electronic engineering design and management, six years in public relations and 25 years as a trade press editor. He holds a BSEE from Case-Western Reserve University, and did graduate work at the same school and UCLA. Sam was the editor for PCIM, the predecessor to Power Electronics Technology, from 1984 to 2004. His engineering experience includes circuit and system design for Litton Systems, Bunker-Ramo, Rocketdyne, and Clevite Corporation. Design tasks included analog circuits, display systems, power supplies, underwater ordnance systems, and test systems. He also served as a program manager for a Litton Systems Navy program.

Sam is the author of Computer Data Displays, a book published by Prentice-Hall in the U.S. and Japan in 1969. He also authored the book Managing Electric Vehicle Power. He is also a recipient of the Jesse Neal Award for trade press editorial excellence, and has one patent for naval ship construction that simplifies electronic system integration.

You can also check out additional articles on his other author page.