Issue link: https://iconnect007.uberflip.com/i/913789
14 The PCB Magazine • December 2017 In the wider electronics industry, thermal management has become increasingly impor- tant due to rapid growth in markets such as LED lighting and renewable energy. However, the is- sue remains a secondary concern to many en- gineers; a recent industry survey [1] showed that thermal management is often considered rel- atively late in the design cycle, and minimal thermal testing is carried out simply to confirm that the system will not overheat. If a problem is discovered at such a late stage, potentially ex- pensive re-engineering may be required. In the automotive market, not only is brand image at stake but the extra costs and delays as- sociated with re-engineering to ensure compo- nent reliability can be unacceptable. Design for reliability, encompassing proper thermal man- agement for electronic components, is extreme- ly important from the earliest stages of product development. On the other hand, over-engineered ther- mal management can be excessively expensive, not to mention heavy and bulky. None of these are acceptable in the automotive business, so engineers need to know how to specify thermal management accurately, using optimal mate- rials and assembly techniques to limit compo- nent temperature rise while also meeting tight constraints on size, weight and cost. Getting to Grips with Thermal Management Surface-mount power semiconductor pack- ages such as quad flat no-lead (QFN) or land grid array (LGA) are typically designed to max- imize extraction of heat from the die to the package underside. Here, a large, exposed, iso- lated metal heat spreader (Figure 1), or en- larged electrical connections such as metallic drain or source terminations of a power tran- sistor (IGBT or MOSFET), is presented, to be soldered directly to a circuit board or substrate. If space is tight in relation to power demand, such as in a high-power or highly miniatur- ized traction inverter, devices such as IGBTs or MOSFETs may be sourced as bare dies with metallized terminations and soldered directly on the substrate. Managing the heat from this point is critically dependent on the substrate properties. Designers have a number of techniques at their disposal, utilizing various combinations of metals and engineered thermal materials to achieve the desired thermal performance and mechanical properties such as size, weight and strength, for the right overall cost. To do this effectively, it is important to un- derstand how to visualize the thermal behavior of the assembly. This can be modelled in a way analogous to an electrical circuit, as a collection of thermal resistances connected in series repre- senting each of the substrate's constituent parts. Figure 2 shows how the stack of materials that thermally couple the transistor die to the sub- strate are viewed as series-connected thermal re- sistances. Note also that components have an associ- ated thermal capacitance, which is defined as the amount of heat energy absorbed or released by unit volume of a material per unit tempera- ture change. This can have an important bear- ing on the dynamic thermal performance of an assembly. In the context of a heatsink, or heat spreader, it can govern the rate of temperature rise when power components are turned on, the rate of temperature fall when turned off, and the time to reach steady-state temperature in continuous operation. Referring to Figure 2, the thermal resistance of any of the elements shown is a function of the thickness and surface area of the compo- UNDERSTANDING THERMAL MANAGEMENT AND MATERIALS TO BOOST POWER ELECTRONICS RELIABILITY Figure 1: Power packages typically conduct heat to the underside, to be extracted into the circuit substrate.