Motion control 101
April 9, 2001
Electronic motion is fast replacing gearing, line shafts, cams, and clutch-brake units in packaging, converting, machine tool, and robotic applications. The benefits include increased machine speed, accuracy, throughput, process flexibility, and reduced product changeover times.
This special report presents an overview of the basic motion system elements, their operation, and interaction. While the primary focus is on closed-loop digital positioning systems controlling dc motors using incremental encoders for feedback, many of the concepts explored apply to other types of motion systems.
Servo motion is just a tiny slice of the multi-billion-dollar electronic motion control pie. But it is an area of rapid growth. Closed-loop servo motion implies the ability to control position at high bandwidths, and such applications typically cycle a motor rapidly from one position to another.
Motion systems are made up of individual components and subsystems that interact to move a load along a specified path, in much the same way as a human moves an arm or leg:
In terms of function, the controller is the brain. It processes motion algorithms, generates command signals, and requests status updates.
An amplifier and motor combine to make up the muscle. The amplifier boosts the controller's signals to a power level sufficient for driving the motor, so the motor can generate the actual movement.
Signals travel through a network of wires, analogous to the central nervous system.
Feedback transducers are the eyes of the system. They allow adjustments for process changes that are based on comparing a measured output with the input.
The mechanical stage is the skeletal structure, supporting the load and/or actuators.
The brain
As the intelligent element that commands the motion, the controller may be visualized as the motion system brain that generates commands and requests status reports. "Motion controllers today are predominantly digital devices that calculate motion trajectories, and compare them with the actual motor/actuator position," explains Jacob Tal, president of Galil Motion Control Inc. (Rocklin, CA). These calculations occur at 2-5 kHz or greater, and the result is information (the reference input to the amplifier) that commands the amplifier to move the motor. Whether packaged as a stand-alone controller or as a board that resides inside the system PC, the controller:
Receives inputs in the form of system commands from a PC, terminal, or PLC
Interprets the inputs to create a profile of the required motion
Stabilizes the system using various techniques (lead/lag, PID, etc.)
Generates command outputs for torque or velocity to the amplifier.
The muscle
Although there are too many types of motors, both ac and dc, to go into much detail here, the most commonly used motors in high-performance motion control systems include brush and brushless rotary motors, and brushless linear motors.
In creating torque or velocity commands, the controller compares the required motion profile, as determined by the PC, terminal, or PLC, with the state of the system as supplied to the controller by the feedback device. |
In general terms, dc motors convert electrical energy to mechanical energy. The key dc motor parameters include:
Armature resistance (r), in Ohms, is the total resistance of the armature winding and the brushes.
Torque constant (Kt), expressed in units of Nm/A or inch-oz/A, indicates the amount of torque a motor can generate for a unit of current.
Moment of inertia (Jm), expressed in kg-m2or oz-inch-sec2, is the sum of the moments of inertia of the rotating parts of the motor.
The servoamplifier takes the reference information from the controller and amplifies this into electrical power to the motor. Historically, the most common reference signal was analog and ranged from +10V to -10V to command the full power output of the servo amplifier. With dc brush motors, analog tachometers were widely used as velocity feedback, but as digital signal processing techniques developed in sophistication, engineers could reduce costs by replacing the tachometer with calculations. Now the controller outputs a signal that commands not velocity, but acceleration. In a rotary motor this is expressed as torque; in a linear motor (See Design News 11/6/2000, pg. 61) it is force. The amplifier senses the motor current that produces acceleration, compares it to the input command, and the difference is processed and amplified to power levels sufficient to accelerate and decelerate the motor to follow the trajectory in the controller. Viewed as a "black-box", this torque amplifier is just a velocity amplifier without the tachometer connection and velocity-loop filter, according to Jim Woodward, applications manager servo products at Copley Controls Corp. (Westwood, MA).
"Since DSPs (Digital Signal Processors) have now migrated from the controller into the structure of the amplifier," explains Woodward, "the servo amplifier of potentiometers and analog circuits is evolving into an industrial network appliance, or a system block that installs simply and is controlled entirely by information." The reference inputs are no longer simply plus or minus 10V analog, but now include information in the form of digital signals in a variety of formats.
A motion control system typically includes a user interface or host, motion controller, amplifier, motor and encoder. |
Since the advent of the digital amplifier, many controller functions have migrated to the amplifier. As a consequence, the amplifier/controller becomes one part in a distributed control system, where the system controller (PC) communicates with outlying amplifier/controllers via RS-232 or RS-485 link, downloads the complete motion profile to each axis and then allows each axis to perform its operation, while polling and monitoring status. Alternately, high-speed bus structures put the system controller in continual bi-directional communication with all the digital amplifiers so it can perform some of the higher-level loop-closing functions, with the amplifier performing local functions, such as closing the current loop. The nerve
Controllers now send their commands via RS-232, RS-485, CAN, Ethernet, Sercos, DeviceNet, Profibus, or others in a growing list. And the amplifiers no longer close their internal loops with op amps and linear circuits, but use DSPs for these functions, too. In some cases, the amplifier is integrated into the motor housing. In other cases the amplifier becomes a simple power block that receives no reference signal from the controller, just digital signals that control the on and off times of the power switches that drive the motor.
The eyes
Feedback transducers are the eyes of a servo system. There are basically two types: digital and analog. Digital feedback transducers include incremental and absolute encoders, and laser interferometers. Analog feedback transducers include inductive and magnetostrictive resolvers.
Encoders and resolvers, the most common servomotor feedback sensors, are typically distinct physical components mechanically coupled to the motor. In large part, these sensors determine the overall capability of a motion control system. "Sensor error limits the steady-state accuracy of position or velocity," explains Edward Burk, technical sales manager at Renco Encoders Inc. (Goleta, CA). "And imperfections such as coarse resolution and cyclical error cause torque perturbations."
Encoders, used in both rotary and linear configurations, are electronic rulers that provide a position signal in a digital (square wave) or analog (sine wave) form. These
signals indicate position to the controller. Encoders work with a light source, a photodetector going through a movable disk or scale, and a fixed mask.
Amplifier/controllers may reduce the prevalence of distinct controller and amplifier elements in motion system design. |
A resolver is a special type of rotary transformer with three windings. The reference winding connects to the rotor, and two output windings oriented 90 degrees to each other mount on the stator. All three windings connect to a resolver to digital converter that interprets sine/cosine data to produce digital outputs as binary coded data and frequently A/B quadrature encoder signals.
While servo systems require feedback signals to close control loops, the lack of discrete feedback components doesn't mean the system is not a servo. Sometimes the feedback device isn't easily identified as in the head-positioning servo of hard-disk drives. Such systems integrate the feedback sensor into the disk platter, instead of using a separate device. Other applications, such as "sensorless" servo systems, use electrical signals from the motor itself. The terminology is somewhat misleading because, in fact, the position is still sensed using inherent motor properties instead of a discrete feedback device.
The bones
Drivescrews, speed reducers, bearings, bushings, guides, and slides make up the skeletal structure of a motion system. Engineers often leave out the machine structure and such mechanical systems, narrowly defining a precision servo system as a controller and amplifier connected to a motor and feedback device. But according to Alan Feinstein, VP of Technology at Bayside Motion Group (Port Washington, NY), the elements of a servo system actually range far beyond that, and into the realm of the structure and surrounding environment.
"Especially since the advent of linear motors," notes Feinstein, "engineers can't limit their view to just the motor and control system. Instead, they must consider the dynamics of the whole system, which has a significant effect on performance." Strength, thermal properties, and damping frequency are the key factors used in selecting materials used in motion systems.
Resolution is four times the line count for the two-channel encoders used in most servo applications. Encoders interface circuitry uses a technique called quadrature decoding to generate four pulses per encoder line. |
Failure mode and impact of material strength is critical. Engineers must determine the types of load/shock, bending, and shear stresses that will be encountered. "As most critical motion components have predetermined materials (such as bearings in hardened bearing steel, gears in various hardened steels, or drive screws in hardened tool steel), it is most often the housings or machine frames that are left to discretion," Feinstein explains. "Material housings must take into consideration if a device is fully supported or cantilevered, locations of load, and type of operation (high shock/vibration)."
With steel bearings, aluminum housings, and glass or various other metals in the feedback devices, it's no wonder that thermal errors are one of the most common in precision motion systems. The coefficients of thermal expansion within a motion system can range considerably from part to part. In dynamic applications that tend to generate heat, engineers have to consider thermal conductivity as well as thermal expansion.
For more fundamentals on motion control, see the motion control channel at www.designnews.com
About the Author
You May Also Like