Power technology this year is poised to get ahead of its standard beck-and-call applications for a change, transforming from servant to master of innovation in its own right. Personal-computer-based technologies will very likely be cast in a catch-up role with power management and control for the first time in a long while.

With increased competition and mergers likely among the already crowded field of vendors delivering power-related products, the major fallout will be delivery of a more consolidated meaning to power management as hardware and software forge a more seamless blend. Expect to see important developments in hardware at both the system and component levels, especially in surface mount.

With more than 300 companies competing for a share of the dc/dc-converter market-many with just one or two major customers-consolidation is inevitable. With that, look for the face and contour of power management to change.

On the software side, the Advanced Configuration and Power Interface looks to take its next big step. ACPI took off like a shot and on schedule. But it has been little heard of since, probably because of the "shakeout" phase, nondisclosure agreements among companies and proprietary schemes for selective turn-on of critical components on system buses.

It will be ACPI's year to leap into broadly integrated systems for software control of power decisions, spanning telecom networks to laptop PCs, all within the framework of total Internet control and monitoring. Tied in with that, expect good innovation from power-backup providers.

When it comes to basic hardware, dc/dc-converter technologies across the board will continue to do more in less space. One likely result is that the true differences between dc/dc converters and the so-called voltage-regulator modules as specified for such applications as the Pentium, already a somewhat dated comparison, will enter the end game.

Surface-mount capability will be honed for power components of all kinds as distributed architectures expand, and dc/dc converters in particular will extend their practical surface-mount limit from the current 25 W to perhaps 50 W. The 100-W "hybrid" chips are not yet a practical scenario.