Power Electronics

Digital Control Needs Clarity on Design Tradeoffs

In the past couple of years, digital control in power supplies has become a frequent topic of discussion within the power electronics community. Numerous articles have appeared on this subject in the trade magazines, and members of the industry have debated this topic vigorously in panel discussions. For example, at last year's Applied Power Electronics Conference, one rap session posed the question, “Power Supply ‘Digital Control’: Real or Virtual?”

As speakers in that session pointed out, digital control technology is real in the sense that power supply products employing digital control already exist. However, there was much debate on whether the benefits of digital control are sufficiently real to warrant the widespread use of digital control techniques in power supply design.

Still other questions were raised about what it really means to say a power supply uses digital control. After all, there's no standard definition for this concept since power supply designers can implement different levels of digital control. For example, digital control could refer to something as simple as using a microcontroller to facilitate supervision of an otherwise analog supply. Or, digital control could refer to an almost completely digital power supply design built around a digital PWM controller.

If we focus on the latter, more invasive form of digital control, then there are a number of pros and cons for adopting this design approach. Advocates of digital control cite benefits such as the ability to easily configure, control and monitor the supply. They also mention the overall suitability of digital control to greater integration.

Skeptics counter that digital design adds unacceptable costs in terms of added silicon and associated hardware and software development. Plus, there's the learning curve for power supply designers steeped in analog design.

But both advocates and skeptics alike grapple with one overriding question: Is there a compelling reason to adopt digital control within power supplies?

The answer coming from parts of the power semiconductor and power supply industry these days seems to be “yes,” and that reason is system-level power management. Witness the recent efforts by Artesyn, Astec and a group of chipmakers to develop an open architecture for power systems control based on the I2C bus (see “Digital Power Management Protocol” in Data Points, page 12.) The need to communicate with and control the power supply also was frequently raised as a justification for digital control at the recent Digital Power Forum in San Jose, Calif.

Not long before digital control became the industry's hot-button topic, many power IC and supply vendors were fixated on power management issues. They kept searching for the best way to control the growing number of supply voltages on the pc board. So, it comes as no surprise that many of these same companies would marry their interests in digital power control and power management.

Judging by talks at the Digital Power Forum, many companies are now addressing the technical hurdles to digital power control while also trying to establish the value of the technology to their customers. To that end, it's been noted that the cost of a power system increases with system complexity, but much more so for a pure analog system than for a digitally controlled power system.

However, more than a causal observation of the relationships between cost and complexity will be needed. To gain acceptance of the new technology in the marketplace, digital power supply developers should clearly document the tradeoffs in performance and cost for digital versus analog power technology as a function of power system parameters.

An analogy can be made with the intermediate bus architecture (IBA). Announcements of the earliest IBA power converters were accompanied by detailed analysis of the cost, efficiency and space tradeoffs afforded by IBA (versus traditional distributed power schemes) as a function of the number of supply voltages, power levels, etc. If IBA, which represented an evolutionary step ahead in power system design, demanded that level of applications-oriented analysis to define its utility, then surely the leap ahead associated with digital power design requires no less.

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.