First, most BLDC motors -- especially those in truly innovative, advanced products -- contain special functions that one-fits-all algorithms can't accommodate. Another issue is that matching motor and load dynamics to the algorithm calls for tedious bench-level tuning. Also, when the algorithm malfunctions (and it will in the course of development), the lack of insight into control-loop dynamics, as well as limited design control, can make diagnosing and resolving problems difficult at best. Another drawback is that this approach defers many design decisions to late stages of the development process, when changes are the most expensive. All of this usually results in increased development time, overengineered and complex motor controls, and, ultimately, products that cost more to produce.
Lack of practical engineering expertise
Compounding generic algorithms and other design issues for companies that make motors or integrate electronics and controls into motors is the fact that the kind of experienced control design know-how they often need isn't readily available. It's not taught in schools. To a considerable degree, it's an acquired skill -- one that's learned and honed in actual application development environments. "Practice makes perfect" is its credo.
Historically, Detroit's auto industry and related Midwest manufacturers have served as the de facto headquarters for this expertise. Now many control design veterans who once anchored this talent cadre have relocated or moved on to new challenges.
As a result, there is a dearth of people who have intimate working knowledge of controls theory, are familiar with the full spectrum of available capabilities and how to apply them, and have demonstrated proficiency in using the latest timesaving development tools. Familiarity with a wide range of physical system architecture components and the issues encountered with them further characterizes these professionals and, ultimately, ensures total system operation -- the electronics and the controls that go with them.
Breakthrough advances in custom-model design
More manufacturers are seeking out such veterans to fill critical control design roles. Others are finding the required expertise among independent specialists forging new avenues to achieve faster, more cost-effective product development. Eschewing canned algorithms, these engineers are making microchip choices that support customized modeling, which ensures clear insight into control-loop dynamics, along with critical design control and manipulation capabilities throughout the development process. Model-based design is nothing new, but these engineers are taking it to a new level, employing tactics that directly apply computationally intense models to highly constrained microelectronics systems, leveraging a unique approach to code export.
Using visible model coding tools such as Simulink, they avoid tedious, time-consuming hand coding (which can be bug laden). Also, architecture changes can be easily seen and implemented, even with several engineers collaborating on a development. More than that, exportable code allows model changes to be translated to the embedded domain in minutes.
Dramatically shortening overall development time, however, begins with another tactic: decoupling. Astute leveraging of modeling techniques allows hardware, firmware, and control algorithms to be developed separately. Thus, the engineer can focus on control system performance and evaluation without the constant distractions of managing and debugging details of actual implementation, problems with hardware (which often lags behind), compiler issues, memory constraints, and so on. Such details can be addressed more effectively and efficiently in a subsequent integration phase.
Tweet This
3 
