Over the years bricks have become smaller to save space. Sixteenth brick modules have now been on the market for two years and DOSA established a sixteenth brick specification more than a year ago. Yet most indications are that the package style hasn’t taken off. There has been little or no discussion of further size reductions. At the other end of the spectrum, full bricks offering 500+W have become too powerful for most applications.
Distributed Power Architectures (DPAs), such as the Intermediate Bus Architecture (IBA) are eroding brick opportunities by providing isolation in a separate stage from the power conversion. Even Vicor, the original power mason, is pushing a new form factor based on its Factorized Power Architecture.
Additionally, power semiconductor manufacturers are increasing their presence in the market and we are seeing more discrete solutions finding their way into new telecom applications, such as ATCA based boards.
So, have bricks hit the wall?!
The evolution of the BMP (board mounted products) industry has been driven by several complex factors, all tightly related to end-users’ applications but also related to macro-economics and business globalization. Another very important factor that influenced the emergence of new architectures and bricks is the speed of technology-shift that results from high competition pushing system integrators to offer higher capacity to customers and operators. That trend, in combination with different environments (e.g. systems’ cooling moving from still air to forced convection and now liquid cooling) and new topologies (e.g. from diodes to synchronous rectification) created room for development and innovation.
Among the many trends, three are dominating the BMP world that have, and will continue to contribute to next generation of bricks profiled for 2010 applications and highly integrated functionalities.
DPA architectures followed customers’ developments and the migration from analog (1980) to digital (1990) and to intelligent power (2000), pushing new developments for innovative bricks. Throughout, the trend has been for higher power density and smaller size.
One of the most important steps in bricks’ evolution has been the migration from free air convection to forced air cooling, which combined with new topologies – for example synchronous rectification - and smaller components has made it possible for manufacturers to increase power density while reducing size.
Size reduction has also strongly been pushed for by end-users to decrease the amount of board space allocated to power conversion and to increase room for core applications (e.g. ASIC, DSP and other processors).
Cost reduction is a very important factor that has forced bricks’ evolution to adopt new mechanical platforms and innovative topologies e.g. 1999 1/2 brick 100W $95, 2003 ½ brick 100W $40. At the same time product development had to take into consideration new legislations such as RoHS and new ways of working related to manufacturing delocalization e.g. US and EU to Asia, and wide outsourcing to various CEMs.
In parallel with the migration from DPA-L3 to DPA-L4 and intermediate bus architectures, bricks associated to POL have gained in functionality, for example sequencing, tracking, etc.
The production of new generations of bricks and POL devices that include such functionalities and trends in order to increase the level of embedded functionalities will continue, until bricks and POL become part of the digital loop.
Due to globalization, increased competition from all over the world, and new developments from the semiconductor industry, bricks will develop by increasing the level of integration. Close relationships between silicon and bricks manufacturers will result in new generations of products that are highly integrated into the end-customers digital loop.
The 2010 challenge is to create standardized bricks with embedded functionalities that make it possible for any customer to have full control over power management and the interoperability between the processor and BMP.
The bricks market is not a monolith. It will continue to address different applications, segment and customers’ applications. Embedded functionalities will continue to develop to support certain applications though significant number of others will not require such level of sophistication and focus on other areas (e.g. Cost, low profile, etc.).
Whatever, with or without digital control, it is clear that power integration will continue and that bricks will continue to bring certain benefits compared to other alternatives. The challenge for Bricks’ manufacturers is to move forward on standardization and to find a route to offer, without compromise, high quality (e.g. 100% Yield concept) and high specification (e.g. power density and efficiency) bricks all at a competitive price as required by end-users.
From Full-Brick to Pico-Brick
Based on 25 years of evolution and technical innovations, bricks have reached a level of maturity and now the challenges facing BMP designers are very exciting and promising.
Bricks are now ready to move to embedded functionalities and controls. The next step will be to make bricks more efficient at any point of load by adding a digital control loop in association with programmable magnetic components. As it has been for VRM, bricks and POL will gain in sophistication, opening new opportunities after decades of evolution.
So, to the question: “have bricks hit the wall?”, the answer is “NO” though tomorrow’s bricks will probably have different form factors than the full bricks that we know now.
What do you think? We want to know. To comment, please go to the PMDL Forums.
Patrick Le Fèvre, Marketing Director
Ericsson Power Modules AB
LM Ericssons vägen 30
126 25 Stockholm - Sweden
Phone: +46-8-568 69507