Important Ways to Approach Heat Management

When you watch a movie on your laptop, the fan almost immediately starts to sound like a jet engine, and feels warm enough to start a fire. If you leave your smartphone in the glove box in the summer, it won’t work until it cools down. Keeping electronics a reasonable temperature is not magic, although it may feel like it at times as the rules and situations are not very familiar to electrical engineers. The effects of heat on electronics are also somewhat difficult to predict and often can be better understood using statistics, such as mean time between failures to show the lifetime averages.

When current flows through a resistor, power is dissipated in the form of heat. Resistors are rated for the amount of power they can dissipate before they burn out or burn up. Electronics are filled with transistors that turn on and off at literally unimaginable speeds but in some ways act as resistors. Transistors are typically understood to be in two states, open, wherein no current flows, or closed, wherein the resistance is nominally zero and current flows unimpeded. However, sometimes reality does not exactly match the concept. For example, when a transistor is open, there is a small leakage of current through what is considered a very high resistance burning a small amount of power. When the transistor is closed, the resistance is not truly zero, and that minor resistance also adds to the total used power. However, the greatest amount of power used is during the transition between the two states.

Although assumed to be a perfect square wave for most considerations, clock signals have finite edge widths where a transistor is partially on and partially off. During this time, a current is flowing yet the resistance is much higher than when the transistor is in the closed state. In the large view, the amount of power in one clock cycle for that transistor is minuscule. However, when you multiply that one transistor by the hundreds of millions, or even billions of transistors in a processor, it jumps phenomenally. Multiply that by the billions of cycles per second and suddenly the trivial amount of power consumed has become a serious concern. Designers have tried many different methods to reduce the power consumed by making the clock edges sharper, at the expense of increased cost and electromagnetic emissions, or by decreasing the voltage level, which increases the susceptibility to spurious signals or digital errors. Many processors now are able to shut down portions of the chip when not in use to decrease power usage. This is very helpful, yet does not decrease the peak power usage.

In addition to avoid burning holes in desks, there are many reasons why we care about how much heat is being generated by electronics. The effects of heat on electronics vary greatly, and they are not always negative. For example, the chemical reactions within batteries tend to work better at room temperature. Yet batteries store better at cooler temperatures and becoming too hot while running can lead to catastrophic and explosive results.

Temperature Ratings 

Most electronics have a minimum and maximum temperature rating, typically 0C to 70C for commercial applications. The minimum temperature rating of electronics is usually not an issue except during the winter when electronics are left in the car or attic. Most of the concern related to cold is either the condensation that could be formed when bringing the electronics into the warmth, much like glasses fogging when coming inside, or the rapid expansion of traces and connections when electricity flows reheat them. Fortunately, these problems are easily avoided by either keeping electronics at a reasonable temperature or letting them come to room temperature before using them.

The damage caused by heat has some similarities with those with cold. Much as a cold part coming to room temperature causes unequal expansion of the different materials, changing from room temperature to levels better suited for baking cakes also causes different expansions. As different materials have different coefficients of thermal expansion, some materials will expand much more than others and break connections. Another issue created by excessive heat is the increased resistance of copper, which creates a spiral effect and can contribute to runaway conditions. Even if the circuit remains functional and the temperature levels off at acceptable levels, the increased temperature and resistance means that more power is being used by the circuit to accomplish the same things, thus decreasing the overall efficiency. While copper increases its resistance with increasing temperatures, the silicon substrate of most integrated circuits is also susceptible to heat, causing it to lose its semiconductor properties over a certain threshold. This leads to completely unpredictable behavior changes in the circuit. Finally, excessive heat can also encourage whiskering, a concern that has already increased due to the transition back to lead-free solder.

Heat Management 

There are a myriad of ways to approach heat management. The simplest could be to reduce the amount of heat created by your product. Microprocessor manufacturers over the last two decades have struggled with keeping the temperature of their products at a reasonable level as they produce a large amount of heat in a very small area. As mentioned, decreasing the voltage, and therefore power consumption, has been very successful in the past. There are other ways to decrease the power output at more of a board versus IC level. Any power conversion is a great opportunity to find savings in heat and efficiency, such as replacing a low drop-out voltage regulator with a switching regulator. This has its own trade offs, efficiency for cost and electromagnetic noise, however, it is a viable option. Other ways to increase efficiency are to reduce any bit-banging that may be required on any embedded microcontrollers. By choosing systems that are able to handle complex or difficult data management in hardware, the amount of required instruction cycles is reduced, allowing more time to go into sleep mode, wherein the chip has time to cool.

  • Heat Mediation – There are times that the amount of heat generated cannot be reduced, at which point the heat must be removed. The first step of heat remediation is to set goals. Establish what levels are acceptable, then identify which devices are most susceptible to heat, as well as what devices create the most heat. If these items are not documented, it is difficult to balance the different trade-offs that will be necessary in the design process. Also, without a set goal, there is no way to know when it has been achieved.
  • Environmental Conditions – Directly after setting those goals, establish your constraints. While all electronics ideally would be placed directly in the flow of dry, cold, clean air, the reality is that most electronics are in very tough environments – e.g. cell phones reside next to warm bodies, computers are placed in dusty office corners with restricted airflow, and embedded systems are placed in airplane locations with wildly oscillating pressure, temperature, and humidity conditions. Defining these constraints requires flexibility. Make sure you know the space availability, any I/O requirements, whether or not these items will be in contact with air, and whether there will be a place to mechanically and thermally connect to the enclosure. Also, determine if there will be encapsulating material such as potting on any of the ICs. Write these items down along with the items that you can change, so that all your variables are clearly defined. While this document should be well-organized and clear, it is also a working document that will certainly change as requirements are refined.
  • Heat Simulation Tools – Simulation tools are more prevalent today, and there are many heat simulation tools available to get general ideas of how devices will heat up and where that heat will go. These are powerful but limited by the same issue plaguing any computer program, that of garbage in, garbage out. When properly used, these can provide a general concept of where to start and what needs changing; however, they are only as accurate as the information provided.
  • Natural Convection – Natural convection is an inexpensive solution that allows your electronics to cool, and it should be the first choice. If the device is not producing enormous amounts of heat and there is flexibility in the spacing of the enclosure, you may be able to passively cool it. This is a fantastic option as there is no fan to power or to get clogged. However, it requires even more exactness in other portions of the heat management technique. Placement and orientation of the heat-producing and heat-sensitive components needs to be managed, taking into account that heat rises. To oversimplify the matter, do not put all of the heat-producing components directly below all of the heat-sensitive components. Also, if in an enclosed space, determine if the passive method will continue to work after the ambient temperature has increased due to the heat producing ICs.
  • Heat Sinks – To increase the effectiveness of passive or active heat removal, heat sinks are sometimes used. A heat sink can be the oddly shaped metal piece attached to a computer processor, but it can also be the circuit board attached to the IC. A properly laid out and drilled PCB can quickly move heat away from the IC and spread it throughout the board. It is important to remember that heat sinks do not make heat disappear; all they do is expedite the movement of heat from one point to another. In a tight enclosure, where the ambient temperature will rise with the heat put out by an IC, a heat sink will not change the overarching issue of the ambient temperature getting too warm. The idea is that a heat sink that is touching a very large heat mass, for example, the atmosphere, will cause such an inconsequential increase of temperature to that mass that it can be assumed to have no effect whatsoever on the temperature.
  • Fans – If the heat needs to be actively drawn away from something, fans are tried and true solutions, but carry their own concerns. They draw their own power, which may negatively affect other aspects of the design, plus they can be noisy, need to be cleaned on occasion, and will leave your product susceptible to damage if they fail. If possible, avoid fans, but if not possible, use them with caution.

When testing for temperature reliability, try to emulate real-world conditions as much as possible. Temperature testing is not purely about surviving certain temperatures, it includes other variables such as air flow, humidity levels, and concurrent physical strains. A product that has been tested up to one hundred degrees Celsius with forced air does not indicate whether or not that product will, in an enclosed environment without forced air, heat itself up to even higher temperatures. Actual tests, like computer simulations, are subject to the same rule of garbage in, garbage out. If the parameters are not correctly setup, then the test results will not be helpful.

Yet another aspect of engineering that must be taken into consideration when designing is that temperature control and testing are not the be all, end all of any product design. A well-designed product will balance all of the needs of the project in an efficient, cost-effective solution to best provide a solution for your customer. Creating a device that always runs well below the temperature threshold but is too bulky to be conveniently used is a failure compared to a device that runs barely within the temperature threshold and has a convenient form factor. However, creating a device that has a great interface but survives less than three days under normal operating conditions is an absolute failure. Keep heat management in your personal toolbox and pull it out every once in a while during the design to make sure that you are on the right track, not just designing another fancy looking brick.

Copyright © 2017 Advanced Assembly