Heat is an awkward parameter to control in an avionics system, especially if normal operation generates large amounts of it. And every part, wire and circuit board trace that has energy flowing in it generates some waste heat. Removing it from a system requires art, planning and insight to avoid early system failure. It’s easy to make heat, but hard to make cold.
It sometimes is difficult to relate power to heat. It helps to remember that you cannot hold even an illuminated 40-watt light bulb in your hands, and the human body cannot touch surfaces over 60 degrees C without pain or damage.
Small amounts of dissipated power give rise to very high temperatures. Small electronic components, such as transistors, integrated circuits, resistors, inductors and transformers, will generate more than 125 degrees C if they are run at their full ratings with no heatsinking.
On an operational level, the parts that rise to the highest operating temperature during normal operation (excluding tubes) are usually resistors. Semiconductors can rise only to an internal temperature of about 125 degrees C before failure. However, resistors work by generating heat, and many types, especially metal film and wire-wound parts, routinely go to 200 degrees C at their full rating, even as low as 1 watt, which can desolder the part from a board or connecting wire.
Remember that for the same dissipation, resistor operating temperature is inversely proportional to size. As parts shrink, they must operate at higher temperatures to achieve the same wattage rating. This often is ignored in surface-mount device work, leading to failures in assemblies and to the desoldering of parts from the board while it is operating.
When power dissipation is large (over 2 watts), the part also must be large (or have excellent heatsinking) to avoid high localized temperatures. The military derating system for increasing resistor reliability is based on this concept. It derates the part’s power-handling capability to give longer life, higher stability and lower temperature rise. Essentially, a larger part, or one with heatsinking, is used to achieve better overall operation at a lower power level by reducing its operating temperature.
It is also important to understand that a part’s thermal rating frequently depends on a significant heatsink (equaling low thermal resistance), or even the notional but very hard-to-find "infinite heatsink." Virtually no board-mountable semiconductor can dissipate more than about 2.5 watts without a heatsink, and many can achieve only 0.5 watt before rising to their failure temperature. These devices depend on external heatsinking to achieve their 10-, 20-, 50-watt or higher ratings.
When parts get hot, a sequence of problems begins at the board level that ultimately leads to an assembly failure. The extreme cyclic temperature rise and fall cracks package seals and bond wires, and causes solder joints to become crystallized and fail. Some power resistors have the ability to reach very high temperatures (over 500 degrees C) before internal failure, and thus can ignite fires, melt other components and board substrates, and trigger insulation failures.
Because of our everyday association with heat, we assume that convection (thermal exchange with air molecules) will remove heat from an assembly. But this works only at high atmospheric pressures (low altitudes), and only if a significant moving air path exists. If the air is trapped or at low pressure (high altitude or in a vacuum in space), this mechanism will fail, and temperatures will start to soar at every localized hot spot.
Heat conduction via heat spreaders, plates, internal board cores and other (generally metallic) conduits spreads heat and avoids hot spots. But the heat must eventually be ducted away from the system by radiation or the physical exchange of material, such as active liquid cooling.
Many systems use convection as the primary means to transfer heat within a unit. But this is inefficient, as conduction through metallic surfaces has far lower thermal resistance. Generally, it is wise to assume you will get little or no internal air exchange; this is the reality of most installations and at most operating altitudes. You have to assume that 10,000 feet is a system’s normal operating environment.
Fans often are employed to move air and improve convective cooling, but they have a limited range of operation and may fail. Their reliability and effectiveness may be compromised because of the installation. A fan’s effectiveness becomes meager when a system is mounted in a congested radio stack, surrounded by other hot units, all trying to exhaust hot air into the same limited space.
Air contaminants (particulates) trapped inside an internal fan, along with water and other solvents, create another critical problem. The resulting electrochemistry can cause catastrophic unit failure or erratic operation and often is associated with extensive surface corrosion.
Walter Shawlee 2 may be reached by e-mail at firstname.lastname@example.org.