Some like it hot, others do not. And those others for sure include the designers of products that contain electronics. There is a growing need to educate designers who are involved with the undesired side-effects of heat-dissipating components.
We often see the following situation. At some moment in time, it is ascertained that a product is getting too hot, and if it turns out that redesigns must be done time after time, it may be quickly concluded that the designers are not sufficiently equipped to avoid this sort of problem from the start. Hiring an expert each time is too expensive and/or too time consuming, and a demand for education quickly arises. In the case of electronics cooling, this is usually accompanied by a request for simple rules of thumb or, one step further, design rules. But is this demand, formulated in this manner, reasonable? Let’s look for a parallel with another discipline, and let’s compare thermal management with EMC management. With both disciplines, we are concerned with an effect that is not desired, and can negatively affect functionality, reliability and safety. One would expect that teaching how to manage the two effects would be approached in roughly the same way. I suspect, however, that most people who are responsible for a product design would look awry if someone were to suggest avoiding EMC problems by letting a couple of mechanical engineers attend a one-day short course, with the emphasis on rules of thumb.
One starts to wonder why solving thermal problems is treated differently. A major factor is the question of responsibility. The opinion generally is that those who cause the problem have to solve it. In itself, this is a nice philosophy, which fits at this juncture in time. However, if by ‘problem’ we mean dissipation, then the electronics engineer is indeed the culprit. It is forgotten that often not the cause (the dissipation), but the consequence (the temperature rise) causes the problem. Generally, the mechanical engineer has a main role here. After all, the final geometry (the cabinet, ventilation, 3D layout) helps to determine how high the temperature rises. Also, from his/her background, the mechanical engineer is mostly better suited to bear the responsibility for thermal analyses. This is why it is useful to ask the question: do we educate the correct people? And is it not so that the demand for an education based on simple design rules is an indication of an underestimation of the importance of thermal expertise to prevent thermally-related problems? And let us, thermal experts, also admit that most of our courses start from what I call ‘academic’ heat transfer principles, often containing a plethora of Nu-correlations ‘for engineering convenience’. It should be realised that correlations cannot cope with a world consisting of a combination of complex geometries and complex physics, such as frequently occurs in electronic system thermal management (see also my May 2007 editorial on ‘correligion’).
One often thinks that it is possible to find answers in simple representations of complex matters, as tendered by popular/technical magazines, data from suppliers, or by application of design rules for which the origin and the field of application can generally no longer be traced.
It should be clearly understood that contrary to common belief perhaps, no simple and at the same time general design rules exist (or can be formulated) for component thermal data, board thermal data, the board area that should be allocated to a component, and especially the boundary conditions. The rule for good design practice is to use the supplier’s compact model data, to perform dedicated measurements when information is missing, to use appropriate software, and, first of all, to collect prior knowledge. This is a standard approach when we talk mechanical engineering. And just as is the case for EMC, nobody should consider it a good idea to teach EE’s a few mechanical tricks and then expect that they are able to prevent mechanical failures.
A final issue that should not be forgotten is that the lack of proper thermal design tools and awareness of new developments could adversely affect the search for alternative designs, because designers tend to be conservative and rely on the availability of prototypes and prior experience.
In short, the business requirements of the future ask for a major change in technical qualifications of the designer who is responsible for thermal management, because precious time is lost when outside support is to be called upon. And this calls also for a major change in course content philosophy.
