Motion systems, in the most basic sense, convert energy from one mode or form to another. In the process, some energy is rendered unusable, typically lost as heat. In some cases, as in mobile or battery-powered systems, concern for this sort of inefficiency often drives the design process. In other cases — industrial equipment, for example — efficiency may not seem to matter as much; that is, until you consider the relationship between excess heat and imprecision, vibration and premature wear, thermal cycles and lifetime, energy consumption and profitability.
Efficiency matters, and in a motion system, it is a collective property, resulting from many components working together. It may start with the motor, but also encompasses interface, drive, and control dynamics. Like any design variable, there is a price associated with efficiency, but the greatest costs are usually incurred when efficiency is overlooked.
Start with the right motor
Motors convert electrical power to an electromagnetic field that, in turn, produces mechanical energy. Making them more efficient requires decreasing losses at each point in the conversion process. “Basically this means reducing iron losses and resistance losses,” says John Malinowski of Baldor Electric Co., Fort Smith, Ark.
“Iron losses are largely determined by the type of steel used in the stator and rotor laminations, as well as the annealing process the steel undergoes,” says Chris Medinger of Leeson Electric Corp., Grafton, Wis. Other motor design variables linked to efficiency include insulation materials, winding methods, and cooling fan parameters. “Cooling fans consume energy, and must be optimized to reduce friction and windage losses,” explains Medinger.
Perhaps the biggest bang for the buck related to motor efficiency, however, is to “right size” the motor for the application. In a recent survey, the New York State Energy Research and Development Authority (NYSERDA) found that 55% of all installed motors are oversized. The “sweet spot” for peak motor efficiency, says Malinowski, is about 80 to 100% of full load. “Below this, efficiency falls off, as does power factor.” And the problem is compounded further. “An underutilized motor has a high amount of ‘lost watts,' which turns into higher operating temperature that could lead to premature failure,” adds Medinger.
Besides selecting a motor designed to run at its rated load, designers should also look at speed. “Some applications, such as pumps and fans, are sensitive to speed variations,” says Medinger. “Motors with an rpm rating higher than the pump or fan's cause higher flow, reduced efficiency, and increased energy consumption.” Along those lines, another energy saving idea is to use an adjustable-speed drive instead of a fixed-speed, on-off drive. “For variable-torque loads, this can reduce energy consumption by 50% or more,” says Malinowski.
Designers can also save energy by ensuring that motor-drive combinations operate at the most efficient point along the speed-torque curve. “The efficiency of a motor and drive package is not a fixed number,” says John Chandler of Automotion Inc., Ann Arbor, Mich. “Rather, package efficiency is a curve starting and ending at 0%.” The curve peaks, says Chandler, where the motor-drive package operates very close to the mid-point of the speed-torque range. One way to lock in there is by adding power factor correction to the drive's ac input, thus eliminating adverse effects of variable ac-line conditions.
Though motor commutation and control may seem to be transparent at the system level, the consequences are painfully obvious when motor and drive fail to efficiently convert electrical current to magnetic flux, and ultimately, torque.
According to Chuck Lewin of Performance Motion Devices, Lincoln, Mass., the greatest losses usually stem from current ripple. “The biggest gains in system efficiency are achieved by reducing switching-current ripple in the motor,” Lewin explains. Here, he's talking about cyclic current fluctuations that occur at the pulse-width modulation (PWM) frequency. “Ripple current increases copper loss in the stator as well as induced magnetic loss in the rotor,” adds John Chandler.
The solution, says Lewin, is to increase motor inductance or PWM frequency, or to apply switching techniques that leave motor ‘legs' unswitched whenever possible. Each approach, however, has a drawback. “Higher inductance motors react more slowly,” says Lewin. “The problem with higher PWM rates, on the other hand, is that they decrease switching efficiencies, thus increasing heat generation in the drive. The third approach, modifying switching schemes, provides only modest increases in performance, while greatly increasing electronic complexity.” Lewin thinks the problem will eventually be solved by new generations of MOSFET and IGBT drivers that switch more efficiently at higher frequencies.
Chandler suggests a more immediate remedy based on improvements in modulation. “Newer digital drives with flexible modulation schemes significantly reduce current ripple,” he says. “Center-aligned modulation is one such technique, particularly for high-speed or low inductance motors.”
Beyond modulation, another area of concern at the motor-control level, says Lewin, is the phase angle between phase currents and rotor. “Any misalignment means it takes additional current to achieve the same torque, resulting in more losses in the motor.”
Dave Wilson of Freescale Semiconductor, Austin, Tex., agrees. “For optimum motor efficiency, the most critical task from a control standpoint is making sure that the applied stator field vector is always properly aligned with the rotor flux. It's like opening a kitchen drawer,” he explains. “The most efficient use of your force is to pull the drawer straight out. Components of force at any other angle are wasted. The same is true with electric motors, except that the optimum angle through which to apply magnetic force constantly changes as the rotor turns. So, you must quickly process the rotor angle feedback, and alter the stator currents to create a magnetic field that properly orients to the rotor flux angle at all times. For commutated applications at higher motor speeds, commutation latency can have a devastating effect on motor efficiency.”
To maintain optimum commutation at high speeds, Lewin suggests using field-oriented control rather than traditional sinusoidal commutation. “Field-oriented control forces the current into alignment better than sinusoidal commutation,” he explains. Other high-speed techniques worth exploring, says Lewin, include phase advance and the use of high-speed torque loops.
At lower speeds, however, sinusoidal drives are likely more efficient. “For identical I2R losses on a sinusoidal back-EMF motor, sinusoidal current waveforms will produce about 5% more torque compared to commutated current waveforms” says Dave Wilson. “That's in addition to the dramatic improvement in torque ripple.” Meanwhile, sinusoidal back-EMF motors are becoming more common than true trapezoidal back-EMF motors because the latter are difficult to construct.
Getting in tune with efficiency
At first glance, tuning may not seem to be a major concern when it comes to efficiency, but it is one of the biggest potential sources of loss in servo-driven systems. “Many types of losses relate to tuning,” says Lee Stephens of Danaher Motion, Wood Dale, Ill. “Wear from vibration, premature component failure, poor product output, and loss of reliability and uptime are just a few.”
Poorly tuned systems also waste energy. “A system that resonates and vibrates may settle within performance specifications, but it will expend more energy than necessary and likely convey a lack of quality,” says Stephens. “Stiffness is one of the best friends of a motion control engineer, while resonance is one of the biggest enemies. On systems that are well-damped and stiff, the tuning can be accomplished in a textbook manner,” he explains. “By keeping resonances above the control range, designers can more freely manipulate gain, filtering, damping, and feedback parameters to reject disturbances and optimize efficiency.”
Sometimes few tuning improvements can be made to the system being controlled, and optimization must come from the controller. “One of the more innovative features in this regard, found on newer servodrives, is the use of observers for tuning,” says Stephens. Observers have the dynamic capability of a filter without the phase lag and potential for instabilities. Observers allow velocity commands to be optimized, which maximizes efficiency without risking instability.
Controllers also need to know the dynamics of the motor they are trying to drive. “Some controllers automatically recognize attached motors, and set motor parameters accordingly,” says Bill Leang of Yaskawa Electric America Inc., Waukegan, Ill. “Tuning is always easiest and most efficient when motor inertia is properly proportioned to load inertia,” adds Leang. Here, motors with high torque to inertia ratios have an advantage.
Motion profiles and control
Another place to look for efficiency in a servo-driven motion system is in the motion profiles commanding the motor on its path from one point to the next. “Losses are likely to reside in the acceleration and deceleration of the motor and load,” says Jacob Tal of Galil Motion Control Inc., Rocklin, Calif.
“Such losses can be reduced using trapezoidal velocity profiles where the three intervals of acceleration, slew speed, and deceleration are equal,” Tal explains. “Strictly speaking, a trapezoidal velocity profile provides good, efficient control. But when the system structure is not rigid, this may cause settling overshoots that increase motion time and reduce process efficiency. In that case, some velocity smoothing many give the best results.”
One solution to this problem, says Tal, is to make velocity profiles flexible, so they can be shaped to suit particular applications. Smart controllers with self-optimizing PID coefficients also remedy the situation, resulting in a stable response with less settling time and greater efficiency.
Stability plays a critical role in profile-related losses. “Overly aggressive motion profiles tend to excite system vibrations, extending settling time and reducing efficiency,” says Boaz Kramer of ACS-Tech80 Inc., Maple Grove, Minn.
One way to avoid vibrations while still using aggressive profiles is a method called input shaping. “The original motion profile is convolved with impulse functions separated in time to form the shaped profile,” says Kramer. “Impulse functions are calculated based on the vibration frequencies.” This method improves efficiency and throughput because it eliminates settling dwells, while allowing more aggressive motion profiles. It also improves efficiency because the shaped profile has very little energy at the system's natural frequencies.
Kramer also suggests using field-oriented control, space vector technology, and power-factor correction techniques. “Field-oriented control maximizes the ratio between motor force (torque) and current. In effect, it maximizes dynamic performance, while consuming less current. Space-vector technology, on the other hand, utilizes 100% of the bus voltage. Standard modulation techniques, by contrast, use only 86%, reducing efficiency as well as the maximum achievable velocity.”
Friction — good and bad
Ballscrews operate with relatively little friction, making them very efficient — 95% or better. Even more efficient are ballscrews employing nonuniform balls. “Alternating ball sizes within the ball nut assembly greatly increases efficiency,” says Andy O'Connell of Rockford Ball Screw Co., Rockford, Ill. “The technique reduces friction by changing the rotational direction between succesive balls.”
There is one caveat, however. Backdriving occurs when the lead angle tangent exceeds the coefficient of friction at the thread interface. Because ballscrew assemblies have a friction coefficient very close to zero, they tend to backdrive under almost any load in vertical applications. In other words, less resistance to motion means they're pulled by gravity out of position. Backdriving on a ballscrew can be prevented with a brake on the screw or motor — useful if the additional cost of a brake is worth the efficiency gained with ballscrews.
More often, however, leadscrews are better suited for vertical applications. Unlike ballscrews, leadscrews have distinct friction, which is also variable by design. “Think of a block and a sphere on a flat surface,” says Tom Solon of Kerk Motion Products Inc., Hollis, N.H. “Any inclination causes spheres to roll, but blocks remain stationary until the surface is inclined such that the forces parallel to the surface exceed the friction forces,” Solon explains. In this case, the lower efficiency of leadscrews actually prevents them from backdriving — so they can function as self-locking mechanisms.
Gear up with synthetics
Lubricants can increase the efficiency of just about any mechanical motion component. Synthetic lubricant in particular, though more expensive, is often justified by lower energy consumption. Darren Lesinski, a technical service and OEM compressor-fluid development manager at Anderol Inc., East Hanover, N.J., explains: “Synthetics can be customized to environments, and they typically maintain their original properties longer — sometimes reducing energy consumption by 15% — 10% more than with mineral varieties. On the other hand, mineral-based lubricants are subject to things like fluid churning, friction wear, and an increased need for relubrication.”
Synthetic polyglycol-based lubricants (PAGs) are especially beneficial for gears. “Gears operating with PAGs experience less friction because of the lubricant's engineered viscosity and high molecular weight,” says Lesinski. Synthetics also reduce temperature and change-out frequency while preventing rust and corrosion. “The cling properties of non-asphaltic synthetic lubricants help form a protective coating that actually cools equipment,” Lesinski explains.
Lubricants with an appropriate balance of base oil and additive reduce power consumption most. Some include antioxidants or extreme pressure additives for components that must withstand heavy loads. Outdoor machinery in colder climates often run at -25° to -50°C or lower. The low viscosity and pour point of synthetics are especially beneficial to such machinery because it improves pumpability. “On cylindrical gears, PAG and polyalphaolefin oils in particular exhibit less churning losses at low temperatures and less gear erosion, reducing costly maintenance,” says Lesinski. Other benefits include reduced start-up torque and lower current draw.