The startling growth sweeping through the power-management industry can be traced to concerns surrounding energy costs and the environment. Depending on how you define power management, Exar Corp. estimates that it's a market of some $15 billion for silicon devices. The specific part of the market served by Exar currently comes in at about $1 billion, including analog and mixed-signal devices—some of which have a large digital element.
One of the most important developments in recent years has been the rise of programmable power controllers, which use digital technology. Naturally, analog designers were somewhat reluctant to embrace digital devices. However, the arguments in favor of digital power management have become so compelling that they’re hard to ignore.
Consider designs that use FPGAs, DSPs, applications processors, media processors, or microcontrollers. These usually need several voltage rails: 5 V for analog, 3.3 V for digital, 1.5 to 2.5 V for memory, and one for the processor core or cores.
Lots of components and board space is required to deliver these separate voltages via analog regulators. It ultimately adds complexity and is very restrictive during system design.
A programmable, digital power-management controller cuts both component count and manufacturing costs. For example, an Exar XRP77xx programmable switching controller can form the basis of a four-rail, flexible power-control circuit using just 33 components. The input voltage can range from 4.75 to 25 V with user-configurable outputs from 0.6V to 5.5 V. Similar functionality delivered from an analog circuit would need over four times as many components, and therefore render the circuit less reliable. On top of that, it still would deliver inferior performance and features. However, functional integration is only part of the story.
For example, when using analog regulators, if any aspect of the power circuit needs to be changed, it will require a board re-design, more component inventory, and time to implement these changes. It’s possible to eliminate these issues, though, with a programmable power-management controller (e.g., a software design tool with a simple GUI). That means quick adjustment of voltage rails, setting up precision power sequencing, implementing voltage and current monitoring, and programming conditional fault handling.
During system development, it’s very easy to adapt the power system to accommodate changes. As a result, programmable power management can be a catalyst in significantly reducing time-to-market for new products. But it goes further than that.
Exar has come across cases where it’s desirable to make small changes to equipment in the field to improve performance. Programmable power control facilitates this field serviceability. A simple firmware upgrade, delivered via an integrated I2C interface on the power controller, can be used to increase a supply rail from 1.2 to 1.3V, for example, to boost the performance of the part of the system it feeds. Many designers have learned that the nominal voltage-supply figure shown in a semiconductor data sheet is not always indicative of its highest-possible performance for a given application.
In addition to these advantages, programmable power controllers offer the opportunity to minimize system energy consumption under all load conditions. This is important because many systems often operate well below full load.
For example, a data center only approaches full-load operation during periods of maximum data throughput—perhaps just 10% of the time. Energy consumption must be minimized for the other 90%. Distributed power architectures using point-of-load converters and programmable controllers deliver that capability.
Power management isn’t just going digital, it’s going programmable. Coupled with the right tools to simplify the programming, analog designers are now learning to love digital, rather than shying away from it.
James Lougheed, VP for Power Management products, holds an MBA and an electrical engineering degree from the University of Southern Queensland, Australia.