Modular approach to an embedded future
14 October 2008
The idea of embedding components, or circuits, within circuits has always been an attractive one to designers as it maximises use of the space available.
As a technology it has suffered from the Sinclair C5 syndrome; something that seems like a genuinely good idea but has insurmountable obstacles.
In the case of the C5 the problems were that lorries couldn’t see you and so death on your eco-friendly and cheap route to work was all the more likely. And it looked very silly.
The two equivalent problems of embedded components both revolve around the fact that once a component has been embedded, it stays embedded. One problem with this is that if that component goes wrong it cannot be reworked and the other problem is that design changes are fundamental rather than superficial, particularly if they need to be made after the board has been manufactured. Rework or change is not an option.
There is also another problem that is more of a practical issue. Who actually makes these embedded circuits? Is it something board assemblers should do, or is it a PCB fabrication issue? Either way, there is no supply chain in place to cater for it and the average PCB fabricator is unlikely to want to invest in the equipment and expertise in SMT to be able to offer this complete solution.
I recently spent some time with a company called Imbera, who believes that it has made such progress that its technology overcomes the problems above and is now ready to take the mainstream by storm.
Firstly, argues Imbera, like all technology, people get better at it the more they do it. The parallel here might be the use of ICT at the end of every SMT line – often a necessity in the past but now increasingly rare as yields improve. For all that this year marks the start of high volume manufacturing for Imbera, the concept has been around for a decade and prototype manufacturing has been going on for three years. The founder, Risto Tuominen, developed the concept as part of his thesis at the University of Helsinki, Finland in the late 90s. The point is that manufacturing problems in Imbera’s process have been ironed out and it claimed a yield of 99.97% in one recent project for a four-layer motherboard.
Also, what Imbera makes are modules – a bit like a modern equivalent of an old fashioned hybrid circuit – that can then be mounted on a PCB in the same way as any other component. I assume that the only limitations on the size of these modules is in the confidence in the technology. If a customer was happy with the process he would be happy to embed more expensive silicon and the module (called an Integrated Module Board – IMB) could become the whole board itself. It is anticipated that this gaining of confidence would be a gradual process. But as a component it can be discarded before assembly if faulty and therefore not result in having to add the whole assembly to the bone pile.
Unlike traditional surface mount technology, using Imbera’s IMB technology, components are embedded inside the core layer of the PCB while being electrically connected directly to the conductors on top of the core layer. Using the technology it is possible to embed all kinds of components including discrete passive components, application specific integrated passive (ASIP) components, bare dies (Silicon, GaAs) and wafer level CSPs. Since the manufacturing process utilises standard PCB and SMA equipment and processes, adopting the technology does not require any heavy investments or exotic techniques.
At the very least this would therefore result in a re-allocation of duties for the SMT line, if not setting up a whole new line, so how would this technology fit into an existing manufacturing process? Imbera’s solution to this is a simple one, it does the manufacturing either itself or through a partner. In fact the company is launching a high volume facility in Asia at the same time as announcing a five year tie up with Ibiden, the world’s largest PCB manufacturer. This effectively means that the modules are supplied as a component, the same as any other, to be placed as part of the ordinary SMT process.
Imbera’s founder and CTO Risto Tuominen commented: “During 2008 we have been building a strong foundation for our manufacturing operation and will soon be announcing a major production line investment, which will enable us to provide first production volumes in H1 2009. The target is then to ramp-up the technology into higher volumes starting from the last half of 2009.”
Any company using the technology does not therefore need to invest in it – the supply chain issue taken care of.
There are a host of technical reasons in terms of electrical and RF performance, power and thermal management, and of course in terms of miniaturisation why electronics engineers should start looking at this sort of technology, but it strikes me that Imbera has taken away the reasons why design engineers would stop looking at it – and this is an important step forward in its long term viability.
Jeff Baloun, CEO at Imbera commented: “IMB technology has enormous potential to satisfy the electronics market’s requirement for thinner, smaller products with enhanced electrical performance and functionality. The robust technology allows significant reductions to be made to the physical size of products whilst improving the cost of ownership through what is a very easy to adopt and versatile process. Our goal is to establish IMB technology as the industry standard for embedded packaging going forward.”
Contact Details and Archive...
Most Viewed Articles...