24 June 2009
Patrick Le Fèvre, marketing director, Ericsson Power Modules
Patrick Le Fèvre, marketing director, Ericsson Power Modules, talks with Chris Shaw
CS: What are the main challenges of meeting MicroTCA specifications?
PLF: Developing a MicroTCA power module was full of new challenges, especially when considering the level of requirements such products need to comply with interoperability within a complex environment.
The MicroTCA specification actually contains three levels of designation, which in everyday language can be called "shall", "should" and "may". That is, it defines a level of absolute requirements along with two additional levels that are intended as recommendations and guidelines but which allow for some flexibility as to the implementation of MicroTCA component assemblies and systems so that they are best suited for the actual application.
This flexibility is an advantage to the system designer who must balance performance, reliability and cost and in most cases must make some sort of trade off between these attributes. The specification and design of MicroTCA power modules represents an example of this flexibility in that their parameters may vary and still meet the provisions of the MicroTCA specification.
Ericsson has undertaken a detailed study of some of these parameters in order to determine how the selected performance level affects other attributes of the power module, including cost. This information is essential for the OEM system designers when specifying and selecting MicroTCA power modules.
For power designers used to working with traditional platforms for board mounted products, a number of new parameters had to be integrated.
One example is the redundancy mode which is more complex to manage and requires that power modules communicate, exchanging information about load condition and potential fault signals generated by internal malfunctions.
The MicroTCA specification requires that any given power module be identified to the system as either a primary power module or a redundant power module.
A given power module within the system may change between these two roles as decided by the MCH, but one power module could not maintain both roles at the same time. In the event of a failure in any output channel of a primary power module, the redundant power module will take over responsibility for all output channels of that primary power module, not just the failed channel.
Automatic transition between a failed primary power module and the redundant power module is accomplished by the settings of their output voltages. Primary power modules are set to a higher output voltage than redundant power modules, the two nominal settings being perhaps 12.5 and 11.5V. This output ORing allows instantaneous and automatic transition in the event of a failure due to the power module with the higher output voltage delivering power to the loads.
This technique also imposes much more stringent voltage budgets and output regulation requirements on power modules (including the primary power modules) used in redundant systems.
Again, redundancy at MicroTCA power modules requires very stringent voltage and tight output regulation when delivering high power, while keeping power losses at the lowest possible level and including possible communication between the module, the MCH and the rest of the system. These were major challenges for our designers.
CS: And how were these challenges resolved?
PLF: First we formed two teams to split the work between hardware and software, and to get each team to focus on a specific part of the MicroTCA power module.
For the first time, the hardware team included systems architect experts sharing their knowledge about communication protocols and becoming the active link between the two teams.
As I detailed in my article published in New Electronics 24 February 2009, the first generation of MicroTCA power modules integrated a standard intermediate bus converter (PKM4304BPI) with the external complex analogue circuitry that was required to meet the tight specification inherent to redundancy.
Rapidly, considering the level of complexity added by the external circuitry and the challenging requirement we had to meet, but as well to lower energy consumption by optimising parameters to various load conditions while reducing cost, Ericsson considered the implementation of a brand new digitally controlled DC/DC converter, the BMR453.
Besides electrical benefits offered by the embedded digital control in the BMR453, the benefit to have an easy way communicate between system to board mounted power module has highly contributed to reduce debugging time and to conform with MicroTCA specifications in terms of interoperability with other equipment makers.
CS: What were the drivers behind Ericsson's decision to take a fully digital approach to power converter design?
PLF: Since 1983, when Ericsson launched its first integrated DC/DC converter, the PKA, the company has brought a number of technical innovations improving products' efficiency. Each innovation improved that efficiency to reach a level that, based on today's technology, is difficult to better. However, considering that fact, in 2003, Ericsson Power Modules started a project to figure out what systems and applications will require, in terms of energy efficiency and how to improve energy efficiency outside traditional 'full load' condition.
As I said earlier, analogue DC/DC converters, based on synchronous rectification and advanced topologies, already delivered unprecedented energy efficiency. But that project focused on improving products' performances and new ways to reduce power consumption at any point of their operation. This project included isolated and non-isolated power modules and consisted of two development teams.
Analogue experts worked on improving the power trains performance, as well as optimising layouts and circuitries to reduce power losses. The digital experts investigated different solutions to bring, at module level, a power management unit capable of controlling vital parameters such as switching, feedback and control loop without requiring an external controller.
As for the development of our MicroTCA power module, at a very early stage of that project we involved system architects to define the strategic parameters they consider a BMP power converter should be able to share with a system energy management host controller.
Step by step, we included more and more power intelligence into our prototypes, to reach an unprecedent level of efficiency characterised by a 'flat curve' from low power conditions to full load.
While improving electrical performances contributing to reduce energy consumption, the rapid development of the open source PMBus accelerated the second part of our project aiming to add a communication interface to BMP DC/DC converters.
Combining internal digital power control and PMBus, in June 2007, Ericsson launched a new generation of BMP DC/DC intermediate bus converters. The quarter brick BMR453 offers an extreme level of flexibility to systems' architect when integrating that module into the overall energy management chain.
CS: What is so special about the BMR454 DC/DC converter?
PLF: As we developed the 396W quarter-brick BMR453, we considered customers' applications requiring a similar level of flexibility but less power, while maintaining the option to upgrade their boards.
As BMR453 was the first intermediate bus converter to deliver such levels of power, the BMR454 delivers 240W which, considering the form factor, equals a power density of 18W/cm2 and is fully PMBus compatible.
By having the possibility to access real time information, it is easy to extract relevant information to report to a centralised energy management controller. Because of this new way of working, which combines hardware, software and systems experts, the time to market is shortened by four months, making it possible for our company to market release this product in June instead of October.
CS: Has the current downturn changed the priority from reducing carbon footprint to improving time to market?
PLF: Our company has always concentrated on reducing our carbon footprint as part of our overall process and developed products, systems and services. For many years companies have driven individual initiatives to reduce their carbon footprint but the current downturn – preceded by high oil prices, contributed to governmental agencies initiating global regulations and collaboration projects, such as The Smart Grid.
On a worldwide basis, we could say that the preservation of our environment has been a growing concern and, during the economic slowdown, companies are taking that momentum as an opportunity to reconsider portfolios and strategies contributing to reduce their carbon footprints.
We are seeing a number of announcements about boosting research on new materials and energy management, for example, President Sarkozy recently discussed increasing investment into the research and deployment of solar based solutions.
Personally, I believe the economic turmoil is not the driver behind these activities but, as I said before, the economical climate has probably contributed to an increase in the level of concerns. Many now realise that reducing the carbon footprint is very important, but their commitment shouldn't be limited to that.
I would like to share a sentence often used by Antoine de Saint-Exupery who said: "We haven't inherited the earth from our ancestors, we borrow it from our children."