Design for Thermal Issues

Unit 17: Applications of thermal management

Over the past few Units we have considered thermal modelling from components to enclosures, looking both at the fine detail and the larger picture. In this final Unit, we are trying to pull together some of the threads of our discussion, setting the thermal management task in the context of the overall design process. Some of our thoughts go back to what was said in the Module Overview, and the consideration given in Unit 4 to the way in which thermal analysis can be used to maximise product performance and minimise risk without adding to the length of the design cycle.

So in this Unit we will be reviewing thermal management as a holistic activity, operating throughout the design cycle and at all levels, and then reflecting on some aspects of best practice, before looking at specific industry applications where thermal management problems are particularly important. Finally, recognising that thermal issues are becoming more important and challenging, we describe some of the more significant trends in the field.

Health warning!

You may find that our section on specific applications, which contains many links to case study information, takes too long to work through in detail. However, we would recommend that you at least skim-read this material, as it gives an insight into the many and varied problems of thermal management.

 

Unit contents

A holistic activity

The Intel paper Thermal performance challenges from silicon to systems has already been referred to (in Unit 13) as a source of information on thermal trends and semiconductor packages. But it is worth revisiting at this stage, as a reminder that thermal problems are becoming more severe, and that improvements in packaging, heat sink technology and system design have to work together to give solutions to considerable technical challenges.

Drawing on their thermal technology roadmap and discussion of some of the solutions, the authors conclude that:

And they finish with the statement: “The desired outcome would be to drive design and technology development concurrently at silicon, package, motherboard and system-level packaging to ensure that thermal solutions can support the need for increasing computing and communication needs”. This emphasis on taking an overall view of the thermal system is a theme that we feel compelled to stress during this final Unit.

Look at all levels

For this reason, when modelling an application in an enclosure, we need to look at all levels, from component, board and enclosure, to the whole room in which the equipment is sited. We could of course create a model in very high detail for the whole room, but then we would find the calculations beyond the scope of simulation programmes to solve. So we need to look at the wood, rather than individual trees, branches or twigs. Unfortunately, we also need to have some local focus, especially for those elements in the system that are especially significant or sensitive.

It is this need to be both wide-ranging and selective, yet within the scope of the software to compute, that puts the greatest demand on the thermal engineer. He/she has to decide whether the focus is on the component, the overall board, or the system. In practice, the key is to concentrate on those relatively few elements that are crucial to the computation, and make simplifying assumptions about the rest.

But don’t forget those key components. They are not always the ones that have the highest dissipation! For example, it is easy to overlook thermal gradients that may have an impact on the function of the circuit: the output of a balanced amplifier may be sensitive to temperature if one section of the amplifier is cooler than the other, causing an offset. So the thermal modeller needs to carry out selected detailed evaluation of the board assembly, using knowledge gained from the circuit diagram (possibly aided by simulation) to identify areas that are particularly susceptible to thermal problems.

Again, at the board assembly/card rack level, where some simplification will be needed, the modeller should look out for problems associated with high impedance flow paths. Read Palis and Sullivan Mezzanine card assemblies; where is all that cooling air going? for an insight into the way in which having a daughter board can affect cooling, with the air flow to hot components being substantially restricted.

At the enclosure level, in order to keep the problem manageable, we will probably treat the majority of the board assemblies as blocks of material generating more or less heat, but need to consider the “thermal wake” of any particularly hot boards or components.

At the whole room level, we have to remember that this is a linked system, and that “power dissipation and cooling techniques employed at one end of the size scale have cascading impacts at the other extreme. For example, ineffective airflow distribution or insufficient underfloor static pressure in a data center can reduce the supply airflow rate from cold-aisle floor tiles, adversely impacting inlet air temperatures and causing microprocessor thermal and reliability problems. Conversely, on the component level, the continued growth in microprocessor power dissipation, and the attendant increase in equipment heat load and cooling airflow requirements, creates challenges for facility air conditioning design, deployment, and operation.”

That quotation from de Lorenzo and Opdahl Server design challenges for the high-heat load internet data centre is an interesting reminder of this interdependence, and contains an interesting illustration of the way that system designs that work well in isolation, whether on the test bench in a wind tunnel or when simulated, may not work as well in situations where a number of cabinets with high dissipation are positioned close to each other.

[ back to top ]


Consider all aspects of the specification

As Wendy Luiten puts it in Sense and nonsense thermal requirements, “Unfortunately, good, workable thermal requirements are a rare commodity”. Her paper is well worth reading, if only as a reminder that there are many ambiguities associated with a specification that quotes “max. operating ambient”, because the flow conditions are not specified, and there is no information on the required local temperature.

But this paper also reminds us that there are three levels of thermal requirement, relating to reliability, functionality and safety.

Of these, the reliability thermal requirement is the most difficult to test in the final product, but is obviously important because unacceptable field failures lead to major problems for the manufacturer.

Functionality is much more easily tested by testing the system in a worst-case situation and observing its performance. But even here there will be variations between pieces of equipment, resulting from manufacturing variability, so some statistical evaluation is advised.

The third classification is of temperature limits that relate to safety issues, from standards such as UL6500 and IEC60065, which describe the limits, the circumstances for which they are applicable, and suggest measuring methods. Outside these safety limits, the situation is potentially unsafe.

So we have to think about what happens in the event of failure – will the system shut itself down? Or catch fire? Will the lack of functionality and failure result in consequential damage?

Whilst safety is not one of the focuses of this module, everyone involved in electronics must appreciate that there are aspects of an equipment that have an impact on safety, depending on the nature of the enclosure and the application. For many consumer products there will be limits on the accessibility of unsafe voltages, requiring the enclosure to be designed to prevent accidental contact. There may also be constraints on cable entries and other points of access, and on the acceptable outer surface temperature of the enclosure. Even where there is no legislative requirement, and the equipment is designed for maintenance only by trained personnel, it is still good practice to indicate the presence of high voltages and hot surfaces.

In Unit 16 we made the point that an equipment specification contains many elements other than those that directly impact on the thermal requirements for the product. Some of these will be part of the formal specification; others will be implied, rather than stated. And often these other aspects of the specification will be at odds with the thermal requirements, so that some compromise needs to be reached. For example, power supplies frequently generate substantial amounts of heat, but need to be mounted towards the base of the enclosure in order to give the box sufficient mechanical stability – top-heavy can easily equate to dangerous! This is one case where effective practice may be to use louvres, baffles or similar to direct cooling air to other areas, and perhaps provide a bypass opportunity for the cooling air needed by the power supply. Alternatively, one could consider blowing air downwards, but then one loses the benefits of “going with the flow” in terms of the inherent natural convection.

Specifications generally focus on the steady state, but most practical applications involve a changing environment, with changes that are occasionally gradual, but often either step changes or transients. As examples, most equipment will be turned on and off, which may lead to power surges, whilst some circuitry may be designed to work in an overload condition for restricted periods. These transients need careful modelling, because the strains induced in the structure by the resultant temperature cycling can lead to premature failure. It is for this reason that most simulation packages have the ability to model the thermal response to transients, though the increased computational burden might require a greater degree of simplification.

[ back to top ]


Think about the need for a thermal safety margin

As well as designed-for transients, there is always the possibility that things can go wrong, and few specifications deal adequately with maintaining satisfactory operation in the event of partial failure. For example, when fan motors fail or filters become clogged, or when the cooling water system becomes clogged with algae. This last is regrettably common, especially with closed-circulation systems.

So, in any consideration of cooling, we have to think of what might happen thermally in the event of system failure. And we have to think laterally as to potential causes of failure. For example, what is the effect of the direction of airflow being affected by a panel being removed during maintenance?

To guard against system failure, whenever we specify a heat sink, for example, should we use a larger heat sink, and less (or no) fan assistance? And how should we position the heat sink to take maximum advantage of any natural cooling there may be in the event of fan failure?

Typically, designers will adopt a conservative approach, allowing for more cooling than is strictly necessary, in order to provide a safety margin. But doing this may involve additional cost, making the product uncompetitive. Simulation is one way in which a designer can be more certain of the potential effects of failure. And the holistic approach that simulation encourages might also lead to other ways of dealing with unreliability, such as arranging for the safe shut-down of vulnerable components or building in over-temperature detection.

Similar issues of reliability, cost and over-engineering reflect whether our design is to be for steady-state or transient conditions. Bearing in mind that overload is a function of both time and over-stress, and that heat takes time to flow, as we have seen in Unit 7, the best strategy may be to have effective heat spreading close to heat-generating components, and a relatively undersized final heatsink.

Comment

Good thermal design depends critically on having an understanding of the circuit function and of the likely dissipation experience of key components. There is no substitute for the thermal engineer working closely with the designer to achieve maximum effect for minimum cost.

 

Risk and reliability are related, and system performance can be enhanced by appropriate choice of technology, but risk and reliability can also be improved by starting early in the process, and trying to reduce the heat generated within the system. [See the Technical Brief by Vance Poteat, A few design techniques on how to reduce the power]

Although primarily within the remit of the circuit designer to influence, trying to keep circuits cool benefits everybody. Unfortunately, when a circuit is supplied from the mains, it is easy to overlook the many strategies eagerly adopted by designers of portable equipment in order to extend battery life. Reducing power consumption helps keep the system cool, and reduces the complexity and cost of the cooling arrangements, as well as saving energy. Don’t forget the requirement of the Energy-using Products Directive, that many designers will have to take into account the energy needs of their designs throughout life.

Supplementary information

For information and links about Directive 2005/32/EC on the eco-design of Energy-using Products (EuP), visit http://ec.europa.eu/enterprise/eco_design/index_en.htm.

[ back to top ]


Using thermal modelling

As we originally discussed in Unit 4, thermal analysis is used throughout the product development process from concept to detailed design. Figure 1 is a simplified alternative way of describing the activities by which a design is first optimised using modelling techniques, and then validated, to produce a final design that meets the specification constraints.

Figure 1: Using thermal modelling

Using thermal modelling

 

We don’t apologise for having included a reminder that, as far as possible, we want the multiple iterations to be carried out in software, rather than requiring too many incidences of “cut and try” in order to optimise the design. In the next section, we have included a reminder of two of the key processes, using an appropriate model, and validating the design.

Modelling approaches

In studying Units 8, 9, 11 and 12, and in your work on Assignment 1, you will have seen the use of different modelling approaches, varying from simple computation to the most advanced CFD analysis. Typically, those without access to modelling tools will use whatever is available, whereas those with the good fortune to own modelling tool licences will use the tool for everything! But we would encourage the use of different tools at different stages, if only to ensure that the CFD tool has given answers that are broadly in line with engineering estimates developed from scratch. The thermal design engineer should always seek in this way to check the results of simulations, and be able to explain any significant differences.

Whether using hand calculation, spreadsheets or full CFD analysis, we would still recommend starting with simple models, and only adding complexity when required, as was done in the walk-throughs. You should only include features that are important to the model in question, and in particular avoid over-complicating the geometry of the model. For example, features such as the fillets in box corners have no bearing on the thermal simulation, and serve only to increase the solution time.

We would also commend the practice of concurrent design and analysis, where thermal design is carried out in parallel with the electronic design, starting at the concept phase. As an example of how this was done, this paper from Hewlett Packard describes the development of a desk-top workstation. Using the Flotherm tool enabled the thermal engineers to evaluate different cooling scenarios without incurring the cost of models. Commencing with simple blocks, the design was progressively detailed as the board design and components were finalised.

The benefits of concurrent design and analysis are further supported by the case study of the design of a Motorola RISC microcomputer that was referenced in Unit 11. The paper is interesting for the comparison drawn between a fan blowing situation, and a configuration in which a fan is used to exhaust the enclosure. This shows clearly the contribution of air leakage outside the affected cooling area, and the importance of the positioning of air vents.

Both case studies involve the development of products that are dependent on earlier work as regards electronics design, the components used, and the enclosure configurations. Whilst this prior art certainly simplifies the approach, this is not to say that an early involvement between design and thermal engineering, even at the stage when the circuit is just blocked out, cannot be of advantage in creating a thermally efficient design.

Refining the design

Having created a model that represents the real situation to a satisfactory degree of accuracy, we will usually need to make modifications to the design. These might be to modify the circulation system with baffles, move the relative positions of vents, fans and the enclosures internal assembly, or re-specify the fan/fans.

Most engineers with access to simulation software will examine the results of the initial simulation carefully, especially with respect to airflow and make judgements about strategies that might enhance performance. But this process requires experience, and may still not produce an optimal solution. However, there are techniques from the mathematical optimisation and design of experiments stable that can be used to examine the effect of changes, and “home in” on the best solutions.

Some of these are discussed by Peter Stehouwer in Design of Experiments for Numerical Parameter Studies of Electronic Systems. Note the use of the formal DoE processes, but especially the use of the response surface concept. In a more recent paper, Improved Thermal Design of PCBs Using Surface Optimization Modeling, Robin Bornoff shows how a structured approach can combine with appropriate software to provide an insight into the thermal behaviour of a proposed design. Again, Bornoff uses a “response surface” to give an overall representation of the interaction between aspects of the design and the resulting thermal performance.Though shown in just two dimensions, the concept can be extended to analysis in multiple dimensions, although this can be difficult to visualise.

Given that Design of Experiments is a well-proven approach, it is not surprising that attempts have been made to automate the optimisation process. The way in which repeated runs can be used for optimisation is discussed by Garron Morris in DFSS for thermal management: Introduction to optimization.

Comment

In Automated analysis guarantees optimum thermal design CAS claim that their Coolit CFD software can guarantee an optimal thermal design at the click of a button. A promising approach, but one where it will be difficult to introduce all the options. There is an additional complication that running a number of computer simulations is expensive in computing time – the one reported in their paper ran 73 cases – so this type of thorough analysis will only be feasible with simplified models.

 

[ back to top ]


Validating the model

Whether a model is built up using flow network modelling or computational fluid dynamics, it is still just that, a model. And models are highly dependent on the assumptions made during their generation. An example of the effect that can come from considerations of turbulence can be found in CFD prediction of electronic component operational temperature on PCBs by Rodgers and Eveloy. So, at the very latest at the prototype stage, we need to check that the real situation actually corresponds to our model.

Activity

The article by Clemens Lasance, CFD simulations in electronic systems: A lot of pitfalls and a few remedies is worth reading, as it focuses on the ways in which CFD models may be misleading.

 

Lasance suggests that the difference between simulation and experiment is the result of inaccuracies in the input parameters and boundary conditions, combined with the complexity of real geometries. He concludes that, in a complex system, it is not possible to predict the junction temperature of critical components with sufficient accuracy, unless the system has been both calibrated and validated.

Lasance draws a distinction between three key words, using them in a specific way that is worth noting:

So ‘validation’ in this context relates to the form of the relationship; ‘calibration’ to the actual temperatures predicted.

Sources of error

Many of the uncertainties arise from the way the model is built (we refer to these as ‘numerical’ uncertainties), and to the information provided, both about the thermal parameters of the materials and the component dissipation. Kordyban has made the comment that electronic engineers often tend to quote a worst-case dissipation rather than a typical dissipation, but assuming that this is always the case is obviously dangerous. [Note in the Lasance paper that, even if the information is correct, the root mean square of other error percentages indicates a final error of the order of 20%.]

Of course the difference between practice and theory may be affected (in either direction) by measurement error. Particular care has to be taken to ensure that the act of making a measurement does not affect the thermal performance of the system, as was discussed in Unit 7. And Kordyban is particularly scathing about system tests that affect the performance of the unit under test and at the same time fail to emulate the application!

Quote

Two points are clear as a result of this exercise:

Modeling Thermocouples with FLOTHERM, Kathy Biber, Flomeric Applications web site, undated

 

[ back to top ]


Selected applications of thermal management

General issues

Although our discussion so far has been generic in terms of the suggested methodology and the issues to be considered, it has inevitably focused on mainstream equipment practice, where multiple boards are fitted into one or more racks, and the overall equipment enclosure is designed to an international specification that is commonly used within a wide range of industries. Elsewhere in the electronics industry, a wide variety of enclosure solutions have been developed for specific applications, from small single boards through to complex and thermally challenging situation. Our aim in presenting typical applications is to reinforce the statement that not everything is a 19-inch rack, and also to expose some general issues that apply in varying measure to almost all these applications.

The first pair of considerations is commercial; How much will the product cost? How soon can the product be brought to market? Both aspects can be improved by a concurrent engineering approach, in which thermal design is initiated as soon as outline information is available on the proposed electronic configuration and package. By allowing changes to be made early in the process, when design considerations are still flexible, considerable savings can be made in the cost of thermal management. At the same time, the risk associated with the development cycle and the eventual performance of the product can be reduced.

But risk is related also to issues of reliability in the field, and the decision that has to be made as to the degree of planned maintenance intervention that is acceptable in order to give the specified up-time. Here there are significant differences between applications, ranging to the “fit and forget” philosophy for a commercial product that is associated with a 10% expectation of failure by end-of-life, to the frequent planned maintenance in a military environment that is coupled with an expectation that there will be no unexpected failures.

Apart from such general themes, all the applications involve conscious thermal management, and a structured approach to resolving the problem, often involving simulation, and usually testing the results.

Specific applications

Our first example is of a mobile telephone in Langari and Hashemi, A system level cooling solution for cellular phone applications. Here the problems are very local to the heat-generating devices, and the problem is of conduction to the exterior. The solution uses micro heat pipes as a system-level cooling solution, a method that you will also find used within portable computers.

If you’ve ever used an LCD projector, you will know that a high light output equates to a level of noise that is not acceptable in a domestic situation. Read Wendy Luiten, Cooling of a Flat TV Monitor, for Philips’ approach to heat management within a flat-screen TV, and note the use of a formal experimentation method to ensure that the design was worst-case.

Computing has already been identified as an application with hot components and high packing density that equates to very high dissipation. In Paul Teague’s One Cool Machine there is a discussion on the design of a supercomputer, with 8kW dissipation per rack. The solution here used air cooling within the system, but with heat exchangers in the door to extract heat from the equipment and transfer it to the building’s water cooling system.

A similar challenge was faced by Intel in Numerical Modeling and Experimental Verification of High-density Servers. Here David de Lorenzo describes a simulation of a range of options. Notice the work carried out on the thin (1U) chassis of the voltage regulating module, that incorporated an integral fan.

Data centres represent a very considerable thermal challenge, and Roger Schmidt describes in Hot Spots in Data Centers a rack that dissipates almost 30kW. Schmidt shows that the primary focus for the operator of the centre is to maintain the inlet air temperature to the rack within the specifications for the equipment. Unfortunately, the flow patterns within the room are markedly affected by the geometry of the room in relation to the cabinets and by variations in power dissipation between adjacent racks.

Avionics applications have always provided challenges to technology, primarily because of the adverse environment, which includes high temperature operation. Operating in a high ambient temperature reduces the temperature increase that is allowable, so more attention has to be given to reducing thermal gradients. And at the same time the structures need to be both mechanically robust and light. Traditionally boards for avionics have been metal-cored, or mounted on metal heat spreaders, and the more thermally demanding applications have used water cooling. As you will see in Sarno and Moulin Thermal management of highly integrated electronic packages in avionics applications, micro heat pipes have now been developed to meet these requirements. Liquid cooling of a more conventional type is shown in operation in Potentials of advanced server cooling by Peter Koch of Knürr.

One of the difficulties associated with avionics applications is that, as altitude increases and pressure reduces, air reduces in its heat capacity and its ability to remove heat, so anything that relies on natural or forced convection will experience greater temperature rises for the same amount of power. A fan-cooled system may experience double the temperature rise at a mere 5,000m altitude! Whilst it is possible to locate some electronic equipment in the pressurised parts of the cabin, at least for commercial aircraft, additional cooling will typically need to be provided. One way of doing this is to use a “slip” fan, which operates at higher volume flow rates at the density decreases. This is discussed by Rhee and Azar in Adjusting temperatures for high altitude.

Our final example is not, as you might have expected, a space application, but the much more mundane area of automotive electronics. This is because the car represents not only a growth area for electronics and a major market, but also a very considerable challenge, as you will see from Figure 2 in Cooling Issues for Automotive Electronics. In many respects the environment under the bonnet is substantially more severe than in an aircraft! Bruce Myers describes the problems, but offers no generic solutions, although it seems that advanced techniques and material will form part of the “thermal engineer’s kit bag”. This area offers scope for innovation, as well as good engineering practice supported by simulation.

Supplementary information

There is a lot of useful information on the Flomerics web site, and their applications page gives links to many examples of the applications of thermal modelling. A number of these are just short promotional leaflets, but others have a more complete discussion of the modelling process and its results. On our supplementary sheet Selected Flotherm applications we have included links to some of the items that we found interesting or helpful.

 

[ back to top ]


Trends in thermal management

In Unit 1, at the very beginning of the module, we gave some data on trends in power dissipation, and referenced an article by Majahan. If you haven’t yet read that paper, it is still worth study, though there are more recent papers covering one or more aspects of current thermal challenges. For example, the November 2005 issue of Electronics Cooling has two articles looking both back and forwards.

Heat generation is a frequent issue at the device level, given that many products have energy-hungry graphic chips as well as main processors and Ron Iscoff’s paper in Chip Scale Review Online indicates how thermal issues are becoming the limiting feature in chip design. Figure 2 is taken from that paper, and shows how even an old and slow CPU can be used to fry an egg, though admittedly it took 11 minutes! [For the sake of accuracy, note that the experiment used bronze pennies to replace the original heat sink.]

Figure 2: A British technician fried an egg using the heat from his PC’s 1500 MHz CPU

A British technician fried an egg using the heat from his PC’s 1500 MHz CPU

Source: Chip Scale Review Online, March 2003

The changes are driven by applications, not just for PCs, but in particular by the telecoms industry and by the demand for extra capacity in the data infrastructure. Also, at the component and board level, at the same time as total power is increasing, we find that dissipation density is increasing, and these challenges are complicated by other changes, such as larger board sizes, increased layer counts, and the development of many specialised designs of device packages and enclosures.

Supplementary information

Emerging directions for packaging technologies, Ravi Mahajan, et al, Intel Technology Journal, Volume 06, Issue 02, 16 May 2002 (PDF, 670KB)

Cooling Solutions In The Past Decade, Wilson and Guenin, Electronics Cooling, November 2005

Advances In High-Performance Cooling For Electronics, Lasance and Simons, Electronics Cooling, November 2005

Thermal Management Tips: If You Can’t Stand the Heat, What Are You Going to Do About It?, Ron Iscoff, Chip Scale Review Online, March 2003

 

A recent widely-reported example relates to the X-box™, where the “Red Ring of Death” indicating hardware malfunction has cost Microsoft dear. Despite the inevitable red herring, that the problem was due to the failure of lead-free solder joints, it seems clear that the root source of the problem is over-temperature, although the actual failure mechanism may be related to thermally-induced flexure of the board. The change to enhanced thermal management techniques in the redesign, despite the higher cost, strongly suggests that thermal management issues were not sufficiently explored during the initial design.

Supplementary information

More information on X-box™ failures at:

Thermal Crisis, Kathy Nargi-Toth, Printed Circuit Design and Manufacture, August 2007

Microsoft: “No Comment” on Upgraded Xbox 360 Heatsinks, Marcus Yam, DailyTech, 15 June 2007

 

As well as creating challenges at component and small enclosure level, thermal management has also become more of a concern for larger systems. In fact, it has been reported that large data centres, such as the server farms run by the likes of Google, need to be sited close to sources of cheap power, and that much of the power is used in cooling the equipment, rather than running it.

Supplementary information

In the data center, power and cooling costs more than the IT equipment it supports, Christian Belady, Electronics Cooling, February 2007

 

As well as using older technologies such as heat sinks and fans more effectively, and developing better designs and materials, solutions for these problems increasingly use heat transfer methods such as heat pipes at the enclosure level, and liquid cooling for larger enclosures and systems, despite some concern that “liquids and electronics don’t mix”. This is just one of the three primary approaches identified in the Thermal Management chapter of the 2007 iNEMI Roadmap as being “currently being pursued to provide improved thermal management solutions for tomorrow’s semiconductor chips and the electronic devices that use them”. That document sees the need for developments at all levels:

Activity

Read Thermal Management in Advanced Microelectronics, Chuck Richardson, Printed Circuit Design and Manufacture, August 2007

 

The article demonstrates the need both for improvements in materials, technology, and software, and for extended collaborative research effort, and Richardson ends with a comment that reinforces our proposition that thermal management needs to be holistic in its view of the problem.

Quote

“. . . evidence is growing that a global thermal management solution strategy can be more effective than a series of localized solutions.”

Comment on the Thermal Management chapter of the 2007 iNEMI Roadmap
Chuck Richardson, op.cit.

 

[ back to top ]


Review

Congratulations on reaching the end of this module. We hope that it has given you insight into the deleterious effects that heat can have on electronic systems, and the greater importance that is becoming attached to effective thermal management. At the same time, you will be in a better position to appreciate:

Only by choosing the most appropriate solution for your application, designing for an adequate performance at minimum cost, and verifying the thermal performance of the product, will you achieve the aim of creating a design that will operate for its design life and over its entire design environmental envelope, yet be sufficiently cost-effective to win market share. Good thermal design can keep your company in business, and you in work!

Finally, don’t forget that, in common with most electronics topics, thermal management is not “set in tablets of stone”, though the basic physical principles may be. In order to create the best designs, you need to keep your knowledge updated. Some of this will come from reading books, and from talking to suppliers, but we would strongly recommend keeping in touch with practitioners in the field, by means such as reading the Electronics Cooling magazine. As with all continued professional development, be critical in your reading, and try to link the new material that you acquire to the knowledge that you have gained through experience and study.

Of course, you may come across new ideas, or applications that make particularly good learning points, and the course team would be grateful to hear of these. We have already included one such illustration in these units, and would encourage you to share further examples of good thermal practice with your AMI colleagues.

[ back to top ]


Resources for this Unit

Each of these lists is in the order in which the material is referenced in the Unit text.

Needed for activities

Specific applications

Optional links and information

[ back to top ]