Traditionally, designers had the choice of using expensive DC-brush motors with low-cost controllers, or low-cost brushless AC motors with complex, expensive controllers.
Today, however, digital signal processing is enabling cost-effective and energy-efficient motor-control designs using TMS320 16-bit, fixed-point DSPs. Designs can now use low-cost motors, smaller memories, and fewer sensors, with the DSP providing the real-time MIPS and integrated peripherals.
When considering variable-speed controller designs, advanced control techniques must be used, which require computational speeds beyond the capabilities of most microcontrollers. This is because AC, DC, and switched-reluctance motors, which use cross-coupled three-phase currents, are nonlinear, multi-variable systems that cannot be effectively controlled by the proportional-integral-derivative (PID) method.
With these advanced techniques coupled with DSPs, designs can use modern brushless motors which reduce system cost through lower maintenance and the elimination of mechanical parts associated with variable-speed control of fixed-speed DC motors.
Adaptive control, with its high computational demands, has become viable only through the use of DSPs. As the name suggests, adaptive controllers update their information regularly, allowing optimum performance in variable systems.
In many industrial control applications, sensing system variables such as position, velocity, or current can be very difficult because of cost, imperfect sensors, or environmentally-induced noise. In these scenarios, optimal performance can only be achieved by using an "observer" model to estimate unavailable system variables.
One of the major benefits of adaptive and observer control methods is significant cost reductions. Observer models eliminate the need for sensors and sensor-interface circuits, and because DSPs can compute parameters as needed, the expense of memory for lookup tables is also eliminated.
In addition, the higher raw computing power of DSPs allows the use of more advanced control algorithms such as state space models, adaptive controllers, Kalman filtering, and observer modeling techniques for more robust and reliable systems.
These advanced control algorithms also reduce a motor's torque ripple, resulting in a more efficient electro-mechanical conversion of power, reducing the overall electricity consumed by a motor.
With electric motors in everything from industrial applications to appliances in the home, it has been estimated that if every integral-horsepower electric motor in the U.S. increased its efficiency by 1 percent, overall power consumption would decrease to the equivalent of an entire 100-MW power plant. The only way to accomplish these efficiency gains is through advanced control algorithms.
Designing low-cost, energy-efficient motor-control systems is facilitated by platforms available through TMS320 Third-Party members. These development and prototype systems are being used by many developers, including TI corporate research and DSP engineering labs.
Spectrum Digital has recently released the Motor Development System (MDS), a programmable motor controller based on the TMS320C52, that provides all of the functions needed to develop customer designs (AC or DC motors). The system is housed in a forced-air-cooled industrial enclosure and is comprised of two plugable modules: the Power Module and the DSP Servo Controller. The Power Module resides on a 6U-sized card and accepts 90-265 VAC control power to generate +5 V, and ±15 VDC for the rest of the system. The DSP Servo Controller module consists of the SBC320C52 CPU, which provides 128K words static RAM and 32K words of Flash EPROM, and a Power Amplifier/Interface Board.
As the cost of DSP solutions decrease, DSP-based motor control designs can be used to increase the performance and energy efficiency of low-cost motor-control applications.