Authored by:
Scott Jordan
Stefan Vorndran
Physik Instrumente L.P.
Auburn, Mass.
Edited by Robert Repas
robert.repas@penton.com
Key points:

• The default choice for precision positioning usually falls to piezoelectric actuation.
• Piezoelectric actuators come in a wide assortment of linear and rotary-motion configurations.
• Hybrid systems combine standard servodrive motors with piezoelectric actuators for more precise incremental motion control.
Resources:

HyperBit demo video

Physik Instrumente (PI) L.P.

Today’s modern machines have to move fast over long distances in many different directions. They must also maintain a high degree of positional accuracy and repeatability. These conflicting requirements place high demands on the motion-control equipment that drives these systems.

Applications like cellphone cameras, medical and industrial endoscopy, and even fluid delivery need small but stiff, responsive, and reliable positioning systems for the optics, probes, shutters, and other associated small loads. For example, sophisticated optic assemblies need multiple axes of nanoprecise alignment and must remain aligned for months of around-the-clock use. Emerging nanoimprint lithographies demand precise positioning and trajectory control. They must also retain alignment integrity under significant physical and thermal stresses.

Until recently, there was no way to resolve these conflicting requirements. Fortunately, a confluence of new piezo-based approaches has breathed new capability into the nano and micropositioning world. Some of these represent significant advancements in traditional mechanisms. Others create forks in the road of positioning technology, adapting the divergent technology for greater suitability to different applications.

Nanometer precision over millimeters
Piezoelectric actuation is usually the default choice for precision positioning. It represents the gold standard for high force and high dynamics. Piezo actuators have a long history in semiconductor manufacturing and testing applications due to their extreme speed and precision. Microscope and medical-device manufacturers as well as the automotive industry have warmed up to the benefits of this actuator technology.

Traditionally, maximum travel distance for most piezostack actuators is on the order of 0.1% of the stack’s length. Methods devised over many years to amplify the flexure lever can extend those travel lengths up to 2 mm.

The downside is that the stiffness of lever mechanisms diminishes as the square of the lever ratio. This limits the bandwidth and, thus, the tracking performance of traditional servocontrol techniques. Demands for longer travels, faster actuation, and higher holding force are at cross-purposes. Furthermore, at some point the granularity of the driving digital-to-analog converter (DAC) will be seen, limiting coarse position resolution or introducing bit-transition noise into the position waveforms.

Novel control technologies can now address these situations. A nontraditional algorithm known as Digital Dynamic Linearization virtually eliminates following errors in repetitive-motion patterns and scanning of long-travel nanomechanisms. The algorithm lets the controller optimize the internal command generated for a repetitive waveform according to the error signal detected via the internal sensor.

A brief auto-optimization commanded by the user when application changes occur typically takes less than 1 sec. Auto-optimization could also be triggered by a reload of previously saved optimizations. The optimization reduces following errors to approximately the noise level of the system. Support for this capability is handled by internal waveform generators and software drivers including COM objects, DLLs, and LabView libraries.

A DAC drives an amplifier to create the voltage applied to the piezo stack. The voltage applied is roughly proportional to the amount of motion. Because DACs operate in discrete digital steps, the number of addressable positions for a piezo mechanism is 2b, where b = the bit-width of the DAC’s digital input. For example, an 8-bit DAC would provide 28 or 256 possible positions. Traditionally, the resolution of the nanopositioner can be no better than its total travel length divided by this number. Achieving tight resolution goals with long-travel flexure nanopositioners means using DACs with high bit counts.

Until recently, a DAC’s limitations were permanent characteristics of the specific chip chosen by the designer. Popular PC analog I/O interfaces like those from National Instruments, Austin, Tex., typically top out at 16 bits, providing 65,536 possible output voltage states. Update rates can reach several million per second with useful waveform and synchronization capabilities. Compatibility also exists with popular programming languages like C++, Visual Basic, and LabView. OEM engineers designing their own controls might choose a 20 or 24-bit DAC for their custom circuit design, but these require careful design because they can raise issues over drift, noise, or other drawbacks.

Sometimes a new design is not economically practical. Others may need to use the low bit-width of legacy DACs, a potential problem when used with highly leveraged mechanisms made from lead-zirconate titanate or PZT.

A recently introduced technology known as HyperBit capitalizes on the underutilized time-domain capability of today’s DACs. HyperBit modulates the DAC’s least-significant bits (LSBs) to virtually extend a DAC’s bit-width, effectively adding as many as 11 extra bits to the width.

For example, the LSBs can be dithered using pulse-width modulation at a high rate. The DAC creates a discrete step voltage for position 7, and a slightly higher voltage for position 8. Rapidly switching the DACs input between position 7 and 8 generates an average voltage between the voltages for the two positions and, thus, a new position between positions 7 and 8. The new position is based on the proportional amount of time the DAC remains in each discrete position. Changing the proportion of time between positions creates many new positions between the two discrete values. The result is higher positioning resolution without trade-offs. This technology has now been built into the industry’s first LabView library for analog nanopositioner interfacing.

Nanometer precision over centimeters
There are some applications that need longer travels than lever amplification of piezo-stack actuators can provide. These needs have led to an array of new technologies based on novel configurations and actuation modes of piezoceramics.

Resonant piezomotors, otherwise known as ultrasonic motors, consist of a ceramic wafer configured in such a way that any high-frequency excitation drives one or more resonant nodes in its material. The material vibrates, creating oscillations in a friction pusher tip affixed to the material at an antinode. The pusher tip oscillates like a miniature pogo-stick, though with nanoscale motions and “bounce-rates” of hundreds of kilohertz. The tip is preloaded against a guided workpiece, pushing it along its path. The dynamic behavior is somewhat similar to that of dc servomotors.

A wide assortment of linear and rotary-motion configurations have been commercialized by providers in many countries over several decades, both open and closed-loop. By eliminating leadscrews, leadscrew inertia, and its associated linkages and structures, these mechanisms can be compact and responsive. For example, an off-the-shelf linear stage can provide 19 mm of travel at up to 500 mm/sec and 10-g acceleration with a 0.1-mm resolution linear encoder, all in a package 35-mm square.

The actuator acts as a brake when quiescent, making its in-position stability superior to conventional stages. Fieldlessness, vacuum-compatibility, long life, and other signature advantages of piezo actuation also apply.

The principle drawback to these piezomotors has traditionally been their difficulty of control. Any motor exhibits a linear stimulus-to-velocity curve: The more power applied to the motor, the faster the motor spins. Small amounts of friction in most dc motors means a small stimulus may still have zero velocity — a departure from the desired linearity. Conventional motor control technology overcomes this stiction with an integrating filter which escalates the stimulus as needed.

Resonant piezomotors have a similar behavior, though from a different cause. Just as a child on a pogostick must achieve a certain threshold “bounce” before he moves, a certain threshold of stimulus must be applied to the resonant piezomotor before motion begins. This deadband behavior is great for in-position stability, but conventional proportional-integral-derivative (PID) controls tend to be unhappy with it. So much integrator gain must be applied to accommodate the deadband that the setting is incompatible with running at speed. The controller oscillator frequency also needs to be in tune with the resonant frequency of the ceramic slab in the piezomotor. In the past this usually required initial tuning and frequent touch-up adjustments during operation.

Both problems were solved with the latest generation of ultrasonic motor controllers developed by Physik Instrumente, Auburn, Mass. An autotuning circuit constantly keeps the oscillator frequency in the optimum range; and a faster processor now switches between gainsets, letting the user take full advantage of a constant high velocity and the stability friction motors provide at rest.

Piezo walking drives
The foundation for a walking actuator combines piezo elements acting in longitudinal and transverse directions. These elements are pressed against a longitudinal rod to confer motion. A familiar progenitor of this family of mechanisms is the Burleigh Inchworm. More recently developed mechanisms offer stiffness and holding force. They are optimized for reliability in applications that hold position over long periods of time while providing centimeters of travel with picometer-class resolution.

A newer design uses cost-effective bender-type piezo elements. Though size and cost are substantially reduced in this motor, power-off stiffness is still high with a 10-N holding force. This design also provides centimeters of travel range with picometer-class resolution. This makes the motor a good choice for applications like optic positioning in microlithography and sample positioning in electron microscopy.

Motion in these mechanisms occurs in two modes: a long-travel stepping mode, and a fine-motion analog mode in which the actuator elements are sheared but not stepped. The shearing action provides submicron positioning capability.

Hybrid motorized/piezo mechanisms
While the mechanisms discussed so far can replace magnetic motors, piezo actuation can also be combined with them to develop a coarse/fine motion. Traditional “stacked” coarse and fine mechanisms operate independently; the piezo mechanism greatly improves the minimal incremental motion capability of the system, but the overall repeatability is no better than the motorized stage alone.

The advent of linear encoders with nanometer-scale resolution has let both the coarse and fine mechanisms share the same feedback sensor. Coordinated by advanced controls, this hybrid provides piezo-class repeatability and fast dynamics over many centimeters of travel.

The field of piezo motion control has expanded rapidly in recent years. The wealth of new concepts introduced aim at eliminating piezos’ former travel limitations while preserving their unmatched resolution, responsiveness, and throughput.

Piezo actuation is becoming more suitable for applications formerly addressed by magnetic motors. In addition, piezo motors add benefits in terms of size, speed, fieldlessness, reliability, vacuum compatibility, resolution, dynamics, and reliability, enabling significant advances in existing and new applications.

© 2012 Penton Media, Inc.