The process of linear motor commutation is not well understood and is often glossed over, but it has a significant effect on performance. Motor force (torque in a rotary motor) is proportional to the magnitudes of the magnetic fields of the magnet track (stator in a rotary motor) and forcer (rotor in a rotary motor), multiplied by the sine of the angle between them. Therefore, maximum force occurs when the magnet fields are oriented at 90 degrees to each other. The goal of commutation is to maintain this alignment and achieve optimal force production.

Early linear motors were brushed versions, which achieved commutation via a commutation bar that ran the length of the brushes in the forcer. But the majority of linear motors used today are brushless servo types, so there is no mechanical means to indicate which phase needs to be energized. Therefore, commutation must be accomplished electronically.

Two methods of commutation dominate most AC servo applications: trapezoidal (sometimes referred to as six-step commutation) and sinusoidal (also referred to as sine-wave commutation). Trapezoidal commutation uses three Hall effect devices that detect the passing rotor (forcer) pole. Each time a Hall signal transition occurs, the phase current is changed, and motor commutation occurs. This is a simple and inexpensive method, but its reliance on Hall sensors causes torque ripple (disturbance forces), which prevents smooth motion and produces higher operating temperatures.

Sinusoidal commutation is the preferred method for linear motors because it minimizes torque ripple and produces very smooth motion. Sinusoidal commutation can be achieved in two ways. One method is to use Hall effect devices that generate a sinusoidal signal as the motor passes over the magnet track. However, the Hall devices can pick up electrical noise, which will affect commutation. Another method of achieving sinusoidal commutation is to use the feedback from an incremental linear encoder to monitor the position of the magnet track.

In either case, sinusoidal commutation requires an initialization sequence to determine the initial commutation phase. There are several ways to accomplish this, but the technique of choice is typically to use Hall effect sensors for initialization, and then rely on the linear encoder for commutation. Three Hall effect devices are used—one for each phase. The state of the Hall effect sensors allows the controller to estimate the forcer position and energize the correct phase sequence. Commutation is then achieved by generating two phase command signals, sin(Θ) phase A and sin(Θ + 120^o) phase B, and multiplying them by the current command. (The third signal is generated by balance loops within the amplifier, based on the principle that the sum of the currents is always zero.) In this method, the noise in the Hall devices is easier to filter, smoother motion is produced, and the motor is driven more efficiently with less heat generation.

Although the methods for commutation and initialization discussed here are typically regarded as the most common for AC brushless servo motors, there are other methods that may be more suitable for specific applications or control schemes. Both linear motor manufacturers and control system manufacturers are knowledgeable about the various methods and factors that should be taken into consideration when selecting and configuring the motor, amplifier, and controller.

Feature image credit: Aerotech