Thermal management in power electronics designComments Off on Thermal management in power electronics design

Thermal management in power electronics design

This article was first published in New Electronics Design, 28th January 2014 – you can download it here.

Don’t forget your thermals!

Thermal management is one of the key components in successful electronics design, particularly power electronics design, so Tim Fryer turned to Ian Loader, director of Universal Science, for a few pointers

Thermal design should not be an afterthought. That was the recurring message from Loader, obviously based on experience of when his company’s services would have been used more productively if called on earlier. “As a thermal specialist, we get involved at two stages in the design process. Hopefully at the start of the design, to make sure that the power design and the thermal design are as one. But we often get involved at the second stage, where the customer’s power design has gone ahead, but they have encountered problems and are going back to try and resolve them from a thermal perspective.” If the design is bad, no amount of thermal tinkering will cover it up, said Loader. “Thermal design will impact on the overall design, so perhaps if you are looking at designing a system in which components run to the absolute maximum of their operating temperature, then maybe you are backing yourself into a corner. I think that, eventually, you will get found out because if you try and drive things towards or beyond their thermal limits, you will notice that things starting to run slowly. The processor operates slowly on a hot day or whenever the temperature increases – it is in the periphery of the system spec is where you start seeing problems. So you need to look at the thermal characteristics of the components at the start before you go into the overall design; you shouldn’t separate the two.”

There is an adage that a 10°C rise in temperature will halve the operating lifetime of an IC. This may not be strictly accurate, but it does indicate that a comparatively small increase in temperature can result in significant reduction in the lifetime of the product. Loader noted this only becomes a consideration after a component exceeds its specified operating range, but cautioned: “There is a tendency to try to get as much as possible out of the system; engineers will design up to a certain limit and then drive the system and the components right up to that limit. The longevity of the system will suffer when you do that. You can look at the efficiency at which something is operating; a PC, for example, doesn’t work as well when it is getting hot. But what is probably a more important factor is in longevity and fall off in lifetime of a device or system when it is driven too hot.” Understanding thermal dynamics is, of course, a discipline in itself. Big engineering companies, particularly in sectors such as aerospace and automotive, may have dedicated thermal labs; others may start with thermal modelling and simulation tools, and there are several options on the market. Loader believes this can only become part of the process and, even then, one that is not available to everyone as investment in thermal analysis is often limited. “You need to design something, simulate it and then build and test it to see what comes out the other end. The more complicated or challenging the environment becomes, the more important it is to have built something and tested it, rather than just doing a thermal analysis.”

Thermal management comes in many forms, from the big components – heatsinks and fans, often which need to be designed or tailored for the application – to thermal materials. These materials come in different formats, such as pads, putty, paste, tape and phase change materials, to suit different applications, and, according to Loader, should not be simply regarded as devices for solving problems late in the design. “Hopefully, materials are included in the design as well. When engineers are designing a big system, they are going to look at the biggest, most expensive items first. A custom heatsink is often accounted for to within a few percent of the cost in that respect, the heat sink comes first. “But everything does need to be bound together and it is important to have good thermal interfaces between heatsink and pcb to get that heat away as quickly and efficiently as possible. A good thermal interface material giving low thermal resistance can actually give rise to a smaller heat sink, which can therefore cost less money in the long term.” Concluding, Loader said: “Good power design and good thermal design are inseparable – you can’t do one without the other.”