Motor and Control Advancements Save Energy

Over the last few years, many equipment designers have moved away from low-efficiency motors, such as single-phase induction motors or brushed dc motors. But often this transition results in a higher cost for motion systems that comprise the motor, an inverter and a controller, and require development of new control algorithms. But enterprises must balance the higher initial cost of efficiency with the long-term overall cost of ownership. Over its lifetime, a more efficient motor can save the consid­erable cost of otherwise-wasted electricity.

"You can buy a one-horsepower 1,800-rpm ac-induction motor for $40 to $50, but a comparable industrial-grade perma­nent-magnet brushless motor could cost as much as $400," says Scott Evans, an appli­cation engineer with Kollmorgen. "Large machine manufacturers have started to include energy-efficiency criteria such as an upper current limit, power use and a high power factor in their design specifications. And, of course, companies want their customers to see them as ‘green.'"

"So rather than using less-expensive, less-efficient motors to save on the initial purchase price of, say, a million-dollar packaging line, they'll pay a fair price for highly efficient motors that help save electrical costs long term," he continues.

"Just because you specify a permanent-magnet brushless motor doesn't ensure a more-efficient system," Evans stresses. "A permanent-magnet motor could use a ferrite magnet typical of those in older brushed-dc motors. Those motors are a bit more efficient than a comparable ac-induction motor but you don't get much for the increase in cost. On the other hand, samarium-cobalt or neodymium magnets produce a much stronger field, which manifests itself in higher thrust or torque. Specify one of these rare-earth-type magnets in a brushless motor to improve efficiency."

"When you use BLDC motors or PMSMs (permanent-magnet synchro­nous motors), you must control them efficiently to generate the highest torque," says Jorge Zambada, senior applications engineer in the High Performance Micro­controller Div. at Microchip Technology. "So you use field-oriented control rather than the basic six-step control that just spins a motor. In field-oriented control, or FOC, the controller measures current in the stator coils and uses that information to precisely ‘position' the fields so you ob­tain the highest torque possible with the same current you would use for a six-step controller." For more information about field-oriented control. (Ref. 1)

"Engineers refer to the current mea­surement technique as ‘sensorless,' but you can use sensor-based techniques instead," says Zambada. "In theory, a sensor, such as an encoder, could give you better position information, but unless you have excellent motor coil-to-sensor alignment, you won't get good feedback information. So in many cases, engineers use sensorless techniques,

although they increase the complexity of control algorithms."

"In a sensor-based design, the engi­neers could use motors with built-in Hall Effect sensors that detect magnetic fields," says Raju Karingattil, MCU business development manager for mo­tor-control applications at Texas Instru­ments. "But this approach adds wiring complexity and cost because the motor requires more wires for the sensors and if a sensor burns out, you must replace the motor. Sensorless designs simplify wiring and mechanical challenges, but they in­crease the challenge of knowing the exact position of the rotor."

"You can apply FOC to 3-phase induc­tion motors, too," continues Zambada. "Those motors traditionally have used what we call volts-per-hertz control that employs a basic configuration of MOS­FETs or IGBTs in an inverter circuit to control the stator-coil currents. Engineers can use a basic 8-bit microcontroller (MCU) to create a dedicated volts-per-hertz controller that doesn't use feedback information. If they want to change to FOC, they must measure field positions and choose a more-capable MCU that can run the FOC algorithms. But they can use the same MOSFET- or IGBT-based inverter circuit to drive the stator fields." In steady-state operation, for a motor under volts-per-hertz control, the air-gap flux is approximately related to the ratio Vs/fs, where Vs represents the amplitude of motor phase voltage and fs represents the synchronous electrical frequency applied to the motor. Thus the volts-per-hertz designation. (Ref. 2.)

"Engineers might say, ‘our brushless application is 70-percent efficient now, what efficiency can we expect from FOC?'" says Zambada. "It depends on their algorithm, the type of motor, the type of load and other factors. Engineers can use a dynamometer to analyze efficiency from electrical-energy input to mechanical-energy output."

"You can always look at the pulse-width modulator (PWM) signals and the type of inverter circuit used to control a motor for ways to save energy," says Zambada. "If you operate the inverter at, say, 20 kilohertz you might get comparable results by dropping to 16 kilohertz. That represents a small change, but it eliminates some switching loss at the MOSFETs or IGBTs. As a rule of thumb, use as low a PWM frequency as you can to get the performance you need. The lower frequencies can increase audible noise, but if you build the motor into a noisy compressor, for example, a bit of added noise won't make much dif­ference but you do save power."

"Put special emphasis on inverter de­sign," stresses Zambada. "The slower the MOSFETs turn on, the higher the switch­ing losses due to the IR drop through the MOSFET. Aim for a low gate resistance on the inverter MOSFETs so they turn on quickly. But if you go too low, you can create switching noise that can adversely affect other electronic circuits."

In a product such as a washing machine, engineers can design a direct-drive or a belt-and-gear-drive tub. The latter often puts restraints on spin speeds. "In that sit­uation, engineers can use a field-weakening algorithm to drive a motor at higher-than-normal speeds, but at a loss of some ef­ficiency," says Zambada. "They can reduce the size of the motor and eliminate the gears and belts, however." Motor speeds can exceed 1.5 times the rated speed. Up to that rated speed, a motor operates in constant-torque mode. But above its rated speed, it operates in constant-power mode. That means the increased speed comes at the price of reduced torque. But in a wash­ing machine, for example, the spinning tub does not need a lot of torque to keep it going. (Ref. 3.)

"We have seen a four- or five-fold increase above a motor's rated speed," says Zambada. "If a customer has used an 8-bit MCU to control the motor they would need to change to a higher-per­formance MCU, perhaps a digital signal controller. The algorithm requires a lot of feedback information and the MCU must continuously adjust the fields to en­sure the right commutation of the motor to raise it to those higher speeds."

Semiconductor companies realize these design techniques sound complicated, so they offer engineers development kits, software libraries and reference designs to give them a head start. "Motor control involves knowledge of the motor - an electromechanical device - analog components, the MCUs, knowledge of

the software and power electronics. So engineers have a lot to think about and a development kit or reference design lets them start with from 50 to 75 percent of the electronics and software already tested and ready to go," says Karingattil.

"Sometimes engineers need some help understanding the different approaches they can take to control motors efficiently," he says. "They might not understand how to compare field-oriented control with 6-step BLDC motor control. For example, you can control a BLDC motor with either trapezoidal or sinusoidal signals. But by us­ing a sinusoidal drive you get better control and less vibration. Vibration means less efficient control of a motor."

"Engineers must understand, though, that the motor that comes with a reference design is usually not the same one they will use in their equipment," cautions Karingattil. "TI provides the parameters for the motors in its kits, so when designers change to a different mo­tor, they must characterize it to obtain its associated parameters and characteristics. Motor suppliers will give you a data sheet, but you need the characteristics of the motor for your load conditions and your operating environment so you can determine the control parameters for your equipment. Those parameters lead to the values you use in the control soft­ware you create. So you must know what happens when the motor starts, stops, stalls and so on. Don't take the data sheet as the ‘last word' on a motor."

"When you develop a control algorithm, the motor exists as a mathematical model," continues Karingattil. "If you plug in in­correct values, the motor will vibrate; have a short life and waste power. So, character­ize a motor as accurately as possible."

With that mathematical model in mind, engineers could look to model­ing and simulation tools to lend a big hand. "Our simulation tools let these engineers model more and more of the motor down to nonlinearities," says Tony Lennon, industry marketing manager at The MathWorks. "Modeling supports the capability to create algorithms that provide better motor control and reduce the amount of power a motor uses."

Engineers may find, though, that a vendor's prototype motor might have somewhat different characteristics than the motors assembled later on a manu­facturing line. Also, the characteristics of a production motor can change due to variations in the wire used to wind the coils and magnetic materials used in rotors. So, those small changes can cause significant variations in the overall system performance.

"Experienced engineers have learned about these variations through a lot of lab testing and they understand the

need to include those variations in their simulation models," says Lennon. "Simulation tools can incorporate these kinds of tolerances in the model and you can run parameter sweeps to see how the system performance will change with the variations in the motors that you receive. Being able to assess the motor variation in the simulation lets you develop more robust motor control algorithms that can be tested against operational profiles. It helps you take a systematic and repeat­able approach to making more informed design and cost trade-offs."

"If you put a motor on a dynamom­eter and you excite it, you can measure inputs and outputs for many operating conditions," says Lennon. "Then you have data that represents the dynamics of motor operation. Optimization routines can use this data to tune parameters, such as rotor inertia and torque constant, in the motor model, resulting in a more accurate model. "A more accurate model lets you make better estimates of how the motor will perform with a variety of real operating loads, which gives you better insight into the kind of controller needed to achieve the desired efficiency of the motor-drive combination," says Lennon.

"Then in a simulation you drive the model with ‘real' current so you can de­termine how many kilowatts the motor will use. And you can put different loads on the simulated motor to examine how it behaves."

In addition to a thorough analysis of electronics and control algorithms, engineers must realize improvements in efficiency don't stop at the end of a motor shaft. "People will spend a lot of money on their motors and controller," says Kollmorgen's Evans. "Then they purchase an inexpensive 50:1 worm-gear box, for example, instead of a more expensive and high-efficiency helical-bevel, cycloidal or planetary gear box. Instead of getting 0.95 horsepower from a 1 horsepower motor, they only get a bit more than 0.5 horsepower after the worm gearbox. The wasted energy goes into heat and audible noise, resulting in premature motor failure."

For more information:

• "Induction Motors," Baldor Electric Corp. http://designnews.hotims.com/27744-503

• "Motor Operation above Base Speed: Field Weakening," Technical Note 128. Emerson Industrial Automation http://designnews.hotims.com/27744-504

• "AC Induction Motor Slip: What It Is and How to Minimize It," http://www.plantservices.com/articles/2002/48.html

• "Developing a Permanent Magnet Synchronous Motor Controller using Model-Based Design," The MathWorks http://www.mathworks.com/wbnr39764