Engineers who work in motion control often start out using a relatively simple approach involving stepper motors to synthesize moves. Steppers are appealing in that their controllers can be simple and inexpensive, at least for undemanding tasks. And most steppers run open loop, needing no position or velocity feedback to hit their targets.
The problem comes when the application demands appreciable torque and relatively fast speeds. Stepper motors can handle a range of torques and speeds that overlap those of more-expensive servomotors, but the economics for the stepper approach often don’t work out.
It is useful to examine some of the key differences between stepper and servomotors as a means of understanding which applications each type of motor fits. Steppers and servos are both synchronous motors. In both cases, the rotation period exactly equals an integral number of ac cycles. Both motor technologies employ a rotor with permanent magnets and a stator with coiled windings. Both systems operate by applying a dc voltage to the stator windings in a specific pattern that results in movement of the rotor, and both technologies are capable of position and speed control.
Two key differentiators between stepper and servo systems involve the use of a feedback device and the complexity of the amplifier electronics. Servos, by definition, are closed-loop systems utilizing a feedback device. Steppers are open-loop systems with no feedback.
The amplifier electronics in a servosystem are typically much more complex than those of a stepper system. While a stepper amplifier simply sends full-rated current to each winding set, a servoamplifier regulates the current levels it delivers to the motor windings. In other words, a servosystem produces only the current the application needs.
Because current is proportional to torque, the control loop in the servoamplifier that regulates current is called the torque loop. The servoamplifier typically also employs velocity and position-control loops. The point of these feedback loops is to handle loads that vary within the design parameters. The servosystem might boost the torque supplied, say, if necessary to keep the servosystem moving at its programmed velocity. In contrast, a stepper system has no feedback and no control loops. So the stepper motor will stall when torque demand exceeds available torque at any given speed.
Stepper and servosystems also perform differently because of variations in their motor designs. Stepper motors have a large number of poles and a winding inductance exceeding that of servomotors. As a result, the torque available from a stepper motor drops off much more quickly with rising speed than with a servomotor, given the same dc-bus voltage. This behavior can be seen in graphs comparing typical stepper and servosystems. In the nearby example, both motors are of similar size, about 2.3-in. square. The servo is slightly longer because of the added feedback device.
Another inherent disadvantage of stepper motors is that they exhibit two distinct regions of instability because the motor behaves as a spring-mass system. One is at low speeds, typically between 100 and 300 full steps/sec or 30 to 90 rpm. It results from excitation at the natural frequency of the motor. When the motor operates in this region, there will be a large velocity ripple, a potential for lost steps, and a significant error in the system’s final position.
There is also a midrange instability that results from interactions between the drive electronics and the motor. This instability typically arises at the speed where motor output torque is half its full running torque. Midrange instability can lead to a stalled motor and, like low-speed instability, velocity ripple and loss of steps that cause inaccurate positioning.
A technique called microstepping can minimize low-speed instability by dividing each step into smaller increments of movement. And there are electronic-damping techniques that can minimize the effects of midrange instability. But the recommended practice is simply to steer clear of operating in these two speed ranges.
Of course, use of a servosystem avoids concerns about instability regions. There are other benefits as well. With a servosystem, output-torque capacity rises dramatically at higher speeds. The additional torque at these speeds can produce a desired movement much more quickly compared with that possible with a stepper system.
Moreover, the accuracy of a stepper system is limited by the number of physical full steps per revolution. For example, a typical two-phase stepper motor has 200 full
steps/rev, so its finest movement increment is 1.8° without gearing. The repeatability of a stepper system will vary with the amount of frictional load. But a servosystem employs a feedback device to hit levels of accuracy far exceeding 1.8°, with better repeatability as well. The position-control loop in the servoamplifier will assure the servo gets to the position that has been commanded, regardless of changing load conditions.
Stepper systems work well in applications where loads don’t change, but many motion-control scenarios are characterized by frequent changes and adjustments. The qualities of mechanical components can vary with temperature and time, and with changes in frictional loads. Users can introduce loads or duty cycles that fall outside the machine specifications. Stepper motors stall when the situation calls for more than their rated torque, so changing conditions can degrade machine throughput. In contrast, servosystems have the ability to warn the user when they sense changing conditions. The servoamplifier keeps track of torque, speed, and position via the control loops, and this information can be used to prevent production stoppages.
Servos also have advantages from the standpoint of energy efficiency. As noted previously, stepper systems send full-rated current in sequence to the motor windings to move the motor, regardless of the application requirements. In addition, stepper systems also draw current when the motor is not moving, though most have a current reduction setting which reduces the current level typically to 50% of the full-rated level when the motor sits idle. But servomotors don’t draw current when they are motionless. They only use the amount of current necessary to realize required motion at any given point in time.
One benefit of higher efficiency is that a servomotor generates less heat than a stepper motor executing the same motion profile. Consider an example situation where the motor accelerates for 85 msec, maintains a constant velocity for 1 sec (at 2,000 rpm), decelerates for 85 msec, and dwells for 1.17 sec. Assume the pattern repeats for 10 min and that the stepper current at dwell is half the motor’s rated stall current. A nearby thermogram depicts the temperature of a servo and stepper motor executing this move. The servomotor maintains a relatively low temperature of 30°C while the stepper hits 70°C.
There are also situations where a servo isn’t a practical alternative to a stepper. For example, sometimes cost is king. Servosystems are typically more expensive than steppers partly because of the additional feedback device and accompanying cable. Moreover, servoamplifiers are more complicated and, thus, pricier than those for steppers. In the same vein, steppers work best where conditions are predictable and won’t change much.
It also typically takes longer to get a servosystem up and running. A servo is a closed-loop system with multiple feedback loops, each of which must be tuned to ensure the system remains stable throughout all its possible operating modes. Consequently, there is setup time associated with adjusting the parameters of each loop.
However, newer servo technologies address some of these concerns. For example, low-cost servo options have become available for as little as a 10% premium over the cost of an equivalent stepper. And some servo products, like the Yaskawa Junma servo Series, have been designed specifically to replace stepper technology. To simplify things, the Junma servosystem detects load inertia automatically. Adaptive tuning algorithms optimize gains for the control loops with no user interaction. The amplifier comes with a pulse-and-direction input resembling that of stepper motors. So in many instances, the servo can use the same controller and motion program as the stepper system that it replaces.