New look for closed loop steppers

The stepper motor is a well-known way of translating electrical energy into precise mechanical motion. Each electric impulse sent to the motor moves the shaft one increment corresponding to a defined angle of rotation. For example, sending 100 impulses to a 1.8° stepper motor would turn it exactly 180°.

One of the attractions of stepper motors is that they generally can operate open loop, with no position feedback for correcting shaft-position errors, because the pulse count defines their shaft position. This scheme can work if a stepper motor is never overloaded or if it never operates near its resonance frequency. But there are problems in many real-world applications that are associated with these two conditions.

To assure the stepper motor never overloads and loses track of its shaft position, many engineers overspecify the motor. So they may end up using a larger and more-expensive model than would be necessary to handle normal conditions.

Moreover, stepper-motor resonances frequently occupy the same band needed to handle application needs, so they may be difficult to avoid in practice. At lower frequencies the resonance behavior is more pronounced; a small increase in load or a torque peak makes the motor lose steps or fall out of step and even come to a standstill. One technique for mitigating resonances is to divide step commands into smaller microsteps, but this entails use of filtering and more complicated controls.

Another way of handling temporary overloads and loss of shaft position is to add a position encoder to the motor shaft in a manner analogous to that used with servomotors. The encoder reads the actual position of the shaft and sends this information back to a controller which compares it to the commanded position. The controller uses the resulting error signal to manage pulses in a manner that either accelerates or decelerates the motor shaft to position it at the right spot.

Unfortunately, this position feedback scheme has no effect on issues associated with oversizing the motor and resonances. But there is a way to formulate a closed-loop stepping-motor system that eliminates difficulties around resonances and allows less-conservative motor sizing.

The technique combines position feedback with a special way of controlling the current in the stepper-motor windings. It is analogous to the vector control of ac-induction motors. It orthogonally aligns currents of the field and armature (or torque component) currents such that, when torque is controlled, the field flux linkage is not affected, thus enabling dynamic torque response. However, this scheme needs no Clark transformation ordinarily used to project threephase currents into two vectors because ordinary stepper motors have only two phases.

In the accompanying diagram, I_d is the flux generating part of the current vector, I_q is the torque generating part. A velocity controller changes the torque-generating current vector. The flux-generating vector is only influenced by field weakening. (Because the controller tries to have the optimum 90° angle between the stator and rotor magnetic field, it always tries to keep I_d at zero. Only if we are using field weakening, implying the speed is above the constant-torque range of the motor, do we need the I_d component.)

A point to note is that the motor- drive waveform in this case is sinusoidal rather than steps or microsteps. One advantage of this technique is that motors powered this way have no problem with temporary load peaks. Their field-oriented current regulation lets them adjust in real time, a feature ordinarily available only through use of servomotors which are more expensive.

Dissecting closedloop feedback

Readers are probably aware there have long been stepper-motor controls claiming to use closedloop schemes with position encoders. Such systems merely check the step position and add or subtract steps to compensate. They cannot correct step angle errors on the fly.

In contrast, sinusoidal commutation coupled with fieldoriented control can indeed compensate for step-angle and load-angle errors within a full step. Vector control of the magnetic field ensures that the stator magnetic field is always perpendicular to the rotor magnetic field and the field intensity corresponds precisely to the required torque. This maximizes torque, optimizes energy efficiency and dynamics, and minimizes torque fluctuations while letting the motor operate more quietly than with ordinary open-loop control.

Closed-loop stepper motors offer dynamics similar to those of servomotors up to about 2,000 rpm. At higher speeds, however, servomotors continue to dominate. The ideal range for stepper motors is below 1,500 rpm where they boast much higher torque than servomotors. They can retain good dynamics even at higher speeds.

These qualities prove beneficial where there’s a need to cover small distances or make small movements during an extremely short reaction time. Applications where sine-wave-commutated, closed-loop stepper motors can replace servomotors include wafer positioning in die-bonding machines, welding-wire feeds, and nozzle control for regulating the throughput of pumps. Additionally, such stepper motors often need no gearbox, which makes them interesting for domains where drive components must be small.

Finally, it should be noted that stepper motors with field-oriented control receive only as much current as they need. They run cooler and deliver the same performance as an ordinary stepper while dissipating less electric power. The advantage becomes clear when considering belt-driven applications which normally use induction motors coupled with flange-mounted gearboxes. It is not uncommon for induction motors in these situations to be replaced by closedloop steppers weighing one-third as much and which occupy about one-third the space.

© 2012 Penton Media, Inc.