Chip-Based Motion/Motor Control Saves Time, Cost, and Space

You would expect leading-edge digital signal processors (DSPs) to be reserved only for sophisticated, high-end motion applications. However, some of the latest processor chips now provide single-axis motor control in washing machines, refrigerators, and related commercial products.

By Frank J. Bartos October 1, 2002
  • Motors, drives & motion control

  • Motion processors

  • Chip-level control

  • Digital signal processors

  • Software development tools

  • Servo motion systems

Motor control vs. motion control

Added on-line coverage:
Chip-based Motion/Motor Control: The Way To Go for Some Users


You would expect leading-edge digital signal processors (DSPs) to be reserved only for sophisticated, high-end motion applications. However, some of the latest processor chips now provide single-axis motor control in washing machines, refrigerators, and related commercial products. Falling prices of chip-level devices and user demands help put this emerging trend-which is counterintuitive at first glance-into perspective.

Today’s motion-control chips are a distinct breed, characterized by programmability via development tools, high data-sampling rates, and ability to incorporate more functions on one chip. Their advanced features are becoming available to a wider range of commercial and industrial products.

Why chip-level control?

”Motion-control chips are the fastest growing segment of the motion control industry because they represent a low-cost solution,” says Chuck Lewin, chairman of Performance Motion Devices Inc. (PMD, Lincoln, MA). Growth of chip-level motion control, albeit from quite a small user base, was substantiated in a recent Control Engineering survey (see ‘Megatrends in Motion Control,’ a special supplement to CE , March 2002).

This technology is available directly from chip manufacturers or as off-the-shelf motion-chip solutions, offered by PMD and others. Mr. Lewin sees further benefit in using off-the-shelf chips-also known as motion processors-to reduce the substantial software tasks inherent in general-purpose processors. He includes here designing profile routines and servo filters, plus handling various ”intricacies of programming digital signal processors,” and states, ”eneral-purpose DSPs do not natively understand motion.” This contrasts with off-the-shelf chips, ”which natively speak a motion instruction set.” Using off-the-shelf chips results in ”substantial size, space, cabling, and cost savings versus buying standard bus cards and trying to squeeze the application into predefined dimensions and connection schemes.”

A chip-based approach costs much less than a standard bus or module approach, according to Mr. Lewin, because design parameters (card size, power consumption, wiring, etc.) can be more tightly matched to a specific application. ”And, by using a standard off-the-shelf motion processor that’s well documented and can be reused in a number of applications, the learning curve for software development is lowered. The software stays substantially the same, since the motion processor instruction set is the same,” he adds.

Chip-based motor/motion control users gain real benefits from ease of use, small size, and reduced system cost, according to Analog Devices Inc. (ADI, Norwood, MA). Ease of use comes from single-chip integration of CPU performance with ADC (analog-to-digital converter), application-specific peripherals, and other functions. Small device sizes follow without the need for a separate DSP controller, external ADC, and special circuits (e.g., FPGA or ASIC) to implement the encoder and PWM interface functions, explains Finbarr Moynihan, product line manager at ADI. Also, a more integrated solution minimizes the number of external parts needed, reducing system cost.

Todd Solak, C2000 DSP marketing manager at Texas Instruments Inc. (TI, Houston, TX), sees two main benefits in using chip-level controls: space savings and more possibilities for innovation. Ability to shrink circuit board size translates to real cost savings for manufacturers when implementing a full motion system. It goes beyond mere part counts and design changes. Software programmability makes DSPs and motion processors more versatile than alternative controls. They offer more potential for design innovation in analog and digital chips, and digital ASICs.

Faster, more capable

PMD’s Mr. Lewin notes several trends in chip-based motion technology. Along with overall advancements in silicon, motion chips are becoming steadily more capable. He cites servo-loop times shrinking from 333e dimensions.” These latest products also have fine-pitch, surface-mount packaging, such as quad flat pack. ”In looking at motion-chip developments over the last 10 years, one important trend is continual improvement in performance both from a function and size standpoint,” says Mr. Lewin.

Blending of motor control functions into motion controllers is another trend (see ‘Motor vs. Motion Control’ sidebar). One example is motor commutation that now routinely takes place in the motion controller.

”Products appearing in the next several years will provide digital current control with high-/low-side digital outputs and shoot-through protection,” explains Mr. Lewin. ”Bringing these functions under the same ‘hood’ as the motion control engine adds advantages, but this level of integration also lowers part count when building a motion controller, particularly where the amplifier is integrated with the motion DSP.”

PMD also notes the beginnings of single-axis chips able to ”natively speak in various network protocols.” For example, Control Area Network (CAN) serial bus connectivity now appears on many DSPs. This has led to such higher level protocols as a CANopen Device Profile for motion control, allowing basic movements via a standard protocol (see the online version of ”Motion Control Standards,” Back to Basics, CE , July 2002). ”More chips designed to be linked on a network will be appearing in the next few years,” adds Mr. Lewin.


Analog Devices views progress of chip-based control as having reached the point where very high-performance, mixed-signal ”system-on-a-chip” designs are possible. Dr. Moynihan points to the importance of mixed-signal capability for chip products because sensor inputs in the real world begin as analog data. ”Accuracy to acquire data and perform precise analog-to-digital conversion results in better position determination,” he says.

International Rectifier’s PI – IPM (Programmable Isolated-Intelligent Power Module) combines the power stage (EMP) with an embedded control board (EDB) that includes a 40-Mips digital signal processor. All components fit within an Econo2 package outline (see graphic).

Single-chip integration with digital control and analog precision is the direction at ADI. For example, ADI has introduced in sample quantities its ADSP-2199x family of mixed-signal DSPs intended for robotics, servo drives, production machines, and other high-performance motor-control applications (see photo). However, such high-performance single-chip products present a challenge. The digital clock running at 160 MHz in the DSP is ‘intuitively at odds’ with the high-performance analog-to-digital conversion process, according to Dr. Moynihan. ”How to make these elements more compatible is a major technology challenge.”

Texas Instruments notes that progress in chip-based control is solidly underway, and at a stage where its adoption is ongoing both in industrial drives and the consumer appliance segment. Fueling further growth will be the need for still faster algorithms and processing speeds, more onboard features-in short, a system-on-a-chip, explains Mr. Solak,

TI’s new processor products (TMS320F2810 and F2812) fulfill demand from users wanting more on-chip capability. This mixed-signal 32-bit DSP offers 150 million instructions per second (Mips), up to a quarter MB on-chip Flash memory, and 12-bit analog-to-digital converter.

To differentiate itself, TI focuses on development tools that enhance DSPs. Among key elements are a comprehensive development environment, debugging tools, and a downloadable digital motor-control software library.

Improving motor performance

Chip-based motor control is becoming more common in devices and products of all kinds, according to Renee Mitchell, marketing development manager at Motorola Semiconductor Products Sector (SPS, Austin, TX). ‘Chip-based motor control is going to become even more pervasive over the next several years. Semiconductors continue to decline in cost and increase in performance, making it cost-effective to add digital motor control to many devices.

Ms. Mitchell lists among advantages of this type of control improved motor performance, extended motor life, and more precise control. Digital control offers another benefit as well. ”Once you have added digital control to a motor, it’s easy to take that next step and connect the motor to an industrial network, perhaps through CAN, to enable better monitoring and management,” she says.

Motorola SPS (and other manufacturers) also attribute energy efficiency to digital controls. However, the dramatic savings figures cited are not compared to another type of variable-speed control, but to constant-speed or on/off-type control. [Energy efficiency of chip-based versus other electronic motor/motion control needs careful evaluation and discussion elsewhere.]

Still, chip-level devices’ sophisticated speed and torque control can offer other benefits, for example, in spindle drives of machine-tools using ac induction or reluctance motors. ”[It] permits an unprecedented degree of tool-speed control and flexibility, while reducing mechanical complexity and machine cost,” adds Ms. Mitchell.

PMD’s Mr. Lewin refers to drive ”smoothness” as one measure of efficiency when DSP methods control brushless or stepper motors. If the actual input current is closely matched to the desired current (torque) flowing through the motor, it does operate more efficiently.

Power modules, development tools

DSP-based intelligence also is trending toward motor power-stage control. New DSP-based power modules cut development time and risk for motor-control designers, according to Davide Giacomini, chief engineer for Intelligent Power Solution at International Rectifier (El Segundo, CA). In International Rectifier’s (IR) approach, traditional switching devices (IGBTs and companion diodes) are combined with current sensing, gate driving, protection, and isolation functions-along with memory-in one off-the-shelf package, reported to be the first on the market. IR’s iNTERO Programmable Isolated-Intelligent Power Module (PI-IPM) is a recent product example (see graphic).

The PI-IPM design takes care of matching analog components with power components and digital control, which otherwise is a sizable design task. It also handles such design issues as gate-drive circuit, thermal management, and electromagnetic interference (EMI). ”Ultimately, the benefits of chip-based motor/motion control are efficiency and higher performance in a drastically reduced space,” says Mr. Giacomini.

All chip devices depend on software tools to unleash their high performance. Especially general-purpose chips require appropriate development tools for motion system designers to realize savings in time-to-market, training, and reuse of code. Some chip manufacturers develop tools directly; others often rely on third-party software specialists to provide the tools.

One notable example in the latter category is by Visual Solutions Inc. (VSI, Westford, MA), a member of Texas Instrument’s third-party network. The company’s VisSim development tool is a visual-block language for defining embedded control algorithms. Connecting the blocks and configuring their parameters is all that’s needed to create a controller, according to Jim Webb, VSI’s product manager. ”Then you can both simulate the behavior of the controller and generate target DSP code,” he states. ”Simulation will find fixed-point overflow and precision-loss problems, as well as perform auto-scaling to prevent overflow and maximize dynamic resolution. Other blocks support on-chip peripherals such as PWM, CAN, serial, encoder, digital, analog, etc.” VisSim reportedly handles quick turnaround of edit-compile-debug tasks using one button (called ”code gen, compile, and link”). Software tools from Visual Solutions support TI’s newly released 28X family of DSPs (F2810 and F2812) for motor, motion, and power control.

Ms. Mitchell of Motorola SPS likewise mentions the importance of software and development tools. Such tools are vital to speed the application of chips into designs and at lower cost, and are part of the company’s ”systems solution approach” for motion-control products.

Chip-based motor/motion control is here to stay. As with all newer technologies the adoption rate is subject to the whims of the marketplace. However, the benefits of blazing processing speed, reprogrammable flash memory, and diagnostic capabilities offered by these devices are hard to ignore.

– Comments? E-mail

Motor control vs. motion control

Definitions of ”motor control” and ”motion control” have blurred over the years, however some distinctions still remain. Whether we take a technology or marketplace view also makes a difference.

Technically, motion control deals with path planning, servo-loop closure, and all other ‘machine behavior’ related to movement, but ends in a single-phase command to each motor or actuator-usually a desired torque or velocity-according to Chuck Lewin, chairman of Performance Motion Devices (Lincoln, MA). ”Motor control is everything that happens afterwards,” Mr. Lewin says. Given a desired torque or velocity, motor control is the mechanism that handles commutation, current (torque) control, low-/high-side signal generation, and feedback. Motion control, in this sense, doesn’t care about the specific motor type, phasing, etc., he explains.

From a market standpoint, ”motion control” refers to high-accuracy applications that usually involve position control. Examples include semiconductor manufacture, medical automation, and machine tools because of the high degree of precision required and the attendant cost. ”Motor control” in a market sense entails powering a motor at the level of torque or approximate speed needed in applications that do not require great precision. For example, a motor control application that spins a fan 1% below its optimal speed, or closes a car window 5% faster on one day than another is acceptable, says Mr. Lewin.

At Texas Instruments (Houston, TX), motor and motion control are regarded as different, with distinct chips applying to each. Motor control relates to spin and speed control, with no real concern about positioning. In motion control, exact positioning of a motor and accuracy of movements are important, according to Todd Solak, C2000 DSP marketing manager.