Why would anyone need to replace analog power? The answer is, of course, a combination of factors, with conversion efficiency often at the top of the list.
Expectations for digital power have run high when one thinks of the explosive growth that the digital revolution has stimulated elsewhere. According to our own data and also independent research, digital power is one of the fastest-growing power management areas, particularly in the information and communications technology (ICT) area, but it is also moving into markets such as lighting and consumer applications.
However, uptake for the digital power sector overall has been firm rather than explosive, despite first-generation modular solutions being at least equal and typically bettering the performance of highly developed analog devices. Not so many years ago, digital power remained at best a slightly obscure idea for many power systems architects. And not least its definition demanded clarification, as
"digitally aware" power-management ICs were well accepted while the latest digital power converter solutions struggled to penetrate mainstream consciousness.
"Digital power" is of course something of a misnomer, as much of a digital power supply mirrors analog best-practice and uses mostly similar or identical components in key areas such as the power switches and output filter. In addition, digital power supply controllers are almost certain to be built in an analog and digital mixed-signal construction rather than pure bulk CMOS logic.
Fundamentally, the difference is represented by the inner control loop changes to implement digital's flexible methodology via the PMBus-based measurement-and-control subsystem. This is a technology that adds little to the cost of a power supply, but one that can hugely influence the outcome.
So why would anyone need to replace analog? The answer is of course a combination of factors with conversion efficiency often at the top of the list. Energy consumption is reaching extremely high levels worldwide and particularly so in the ICT area and in data centers especially. Another factor is the increasing complexity of power distribution systems, where the flexibility of digital power can help with the sheer number of different voltage rails required for microprocessors, ASICs, FPGAs and other digital logic, for example.
Looking at efficiency for a moment, unlike a regular analog converter, digital converters can adapt to changes in line and load conditions in real time, due to network traffic demand for example. We took the decision several years ago to move forward with a digital power project, and after two years of intensive effort that included inventing software tools to ease firmware development, we delivered our first advanced intermediate bus converter (IBC). This has been most recently followed by the BMR456 in 2012, which embedded the Ericsson DC/DC Energy Optimization firmware. This new product not only doubled the power density of a tightly regulated bus converter, but also bettered key efficiency and regulation performance parameters.
Of course, overall system savings can be further maximized via the use of dynamic bus voltage architectures by backing off the bus converter's output voltage to minimize dissipation in the point-of-load regulators during low-load periods. But this is another story for another day...