Nanometer Positioning Solutions
How integrating an Air-Bearing, a Piezo Stage, a laser Interferometer & EtherCAT -- Creates Long-Travel Precision
Three technologies that each solve a different aspect of nanometer-positioning — and why EtherCAT is the best interface to connect them.
Nanometer-level positioning accuracy over millimeter-scale travel ranges is achieved by understanding which technology does which job well and designing the system so that each technology operates within its optimal range while compensating for the limitations of the others.
Three technologies form a particularly coherent combination: air bearing motion stages, piezoelectric fine positioning actuators, and laser interferometers. Each is mature and well-understood in isolation. Together, they address the full performance envelope that neither could reach alone. Understanding why this combination works so well is the starting point for designing systems that perform at the highest level physically possible.
Using An Air-Bearing for The Long-Travel
The coarse stage determines the quality of the entire stack built upon it. Air bearings make nanometer-level correction possible by eliminating contact friction and wear.
Precision motion starts with the coarse stage. The coarse stage carries the load, provides travel range, and must reject external disturbances well enough that the fine correction stage can function optimally. The choice of bearing technology here has consequences that propagate through the entire system.
Ball-screw or recirculating ball-bearing stages, while cost-effective for many applications, introduce three problems that are difficult to compensate downstream. The first is friction and stick-slip: as the stage reverses direction or moves at very low velocities, static friction creates discontinuous motion, small jerks that appear as positioning errors at the nanometer scale. The second is mechanical noise: rolling contact elements generate vibrations as the balls recirculate. The third is wear and repeatability drift: contact mechanics change over time and often require periodic maintenance, and component replacements to maintain precision standards.
Air bearings eliminate all three. The moving carriage rides on a thin film of pressurized air, maintaining a non-contact relationship with the guide surface. There is no static friction, no stick-slip, no recirculating contact noise, and no wear. The motion profile is smooth and continuous in a way that contact bearings simply cannot achieve.
For a fine positioning piezo actuator to correct residual positioning errors, those errors must be predictable and small enough to fall within the piezo’s stroke range. Stick-slip events and contact noise make errors both unpredictable and large, exactly what a piezo cannot compensate for efficiently.
Air bearings are also extremely stiff in the directions they constrain. This stiffness means the coupling between the air bearing stage and the payload is mechanically rigid and well-defined. This is a prerequisite for the fine correction stage to work against a predictable mechanical interface.
Using A Piezo Stage for The "Final Mile"
Servo drives in combination with air bearings cannot eliminate all positioning error – their control bandwidth is limited by mechanics. Piezo actuators operate at a fundamentally different speed, correcting what the servo cannot, at resolutions the servo cannot reach.
Even a well-designed air bearing stage with an excellent servo controller leaves residual positioning errors. These come from: vibration transmitted through the machine structure, thermal gradients causing slow drift in the stage guidance geometry, and the finite bandwidth of the coarse motion control loop. No servo system, regardless of controller quality, can fully eliminate these at nanometer scale.
Piezoelectric actuators address this through a fundamentally different physical mechanism. The piezoelectric effect allows direct conversion of electrical energy into mechanical displacement with no moving parts, friction, and zero backlash. This means:
- Control bandwidths up to tens of kilohertz, orders of magnitude above what a servo drive can achieve through a mechanical drive train.
- Sub-nanometer resolution. Solid state piezo actuators have virtually infinite resolution, limited only by the voltage noise of the electronics.
- Deterministic and repeatable response where the displacement-voltage relationship is predictable and stable.
The practical limitation of piezo actuators is stroke: a typical stack or flexure-based piezo stage offers travel in the range of micrometers to a few hundred micrometers. This is why piezos do not replace the coarse stage — they complement it. The air bearing stage positions the payload to within a few micrometers of the target; the piezo actuator covers the remaining gap with sub-nanometer resolution.
The air bearing positions to within the piezo’s range. The piezo positions to within the measurement system’s resolution. Each technology sets the stage for the next.
Critically, an air bearing stage is an ideal host for a piezo actuator. Because air bearing motion is frictionless and smooth, the residual errors the piezo must correct are small, well-behaved, and within its stroke range. The piezo is not fighting against stick-slip events or mechanical noise, it is making fine adjustments to an already-stable platform.
Using an interferometer to close-the-loop
An encoder measures where the stage mechanism is. An interferometer measures where the payload is. These are not the same thing — and at nanometer scale, the difference is not negligible.
The most common approach to position feedback in precision motion is to read an encoder — a linear scale or rotary scale attached to the motor or stage mechanism. This is practical and cost-effective, but it measures something subtly different from what we actually care about.
An encoder measures the position of the stage mechanism. The position we care about is the position of the workpiece — or the point of the tool, or the optical element being aligned. Between the encoder and the workpiece lies: the structural compliance of the stage body, the thermal expansion of the scale and its substrate, any angular errors in the guidance geometry (which project onto the measurement axis via the Abbe effect), and any mechanical hysteresis in the coupling.
The Abbe principle states that the measurement axis and the axis of interest must be collinear to avoid amplified angular errors. In most encoder-based systems they are not — there is an offset. A 10 µrad angular error (a very small tilt) at an Abbe arm of 50 mm produces a 500 nm position error. This is directly visible at the nanometer-accuracy level.
Laser interferometers bypass this problem by measuring the absolute distance between the interferometer head and a reflector mounted directly on the payload — the actual object whose position matters. The measurement is based on the laser wavelength, which is a physical constant traceable to the SI definition of the meter. There is no scale to thermally expand, no mechanical contact to wear, and no accumulated counting error.
Interferometers do have dependencies of their own: the laser wavelength in air depends on temperature, pressure, and humidity. Modern interferometer electronics include environmental sensors and apply the Edlén correction in real time, keeping the wavelength uncertainty well below 100 nm/m under controlled conditions. For applications requiring better than this, external meteorological sensors can feed the compensation model directly.
The role of the interferometer in a multi-layer motion system is therefore not just measurement — it is closing the control loop on the physical quantity that actually matters. When the interferometer feedback reaches the motion controller, positioning errors that would otherwise be invisible become correctable.
Using EtherCAT to Synchronize The Timing
Three independently capable subsystems running on unsynchronized clocks will underperform relative to their specifications. EtherCAT gives them a shared time base — the prerequisite for coordinated control.
Each of the three technologies described above has its own control electronics: the servo drive for the air bearing stage, the piezo amplifier, and the interferometer signal processor. In a conventional system, these operate largely independently, connected through analog signals or proprietary digital interfaces with no shared time reference.
This creates a practical problem. If the servo drive, the piezo amplifier, and the interferometer are all sampling and updating at different rates with no synchronization, the system cannot combine their data. The interferometer measures position at time T₁; the servo drive applies a correction based on data from time T₂; the piezo acts on a setpoint computed at T₃. These small timing offsets, typically in the range of microseconds to milliseconds, translate directly into positioning errors at high speeds and make system identification, calibration, and diagnostic analysis significantly harder.
What EtherCAT Provides
EtherCAT (Ethernet for Control Automation Technology) is an industrial Ethernet protocol designed specifically for deterministic, high-speed control applications. Two features are particularly relevant here.
Deterministic cycle times: EtherCAT exchanges data between the master controller and all connected devices in a single network frame that traverses every node in sequence. This gives predictable, bounded latency — a property that conventional Ethernet does not provide. Typical bus cycle times in precision motion applications range from 250 µs to 2 ms, depending on the number of devices and the data volume.
Distributed Clocks: EtherCAT’s Distributed Clock (DC) mechanism allows every device on the network to synchronize its internal hardware clock to a shared master reference, with sub-microsecond accuracy. In practice, this means the interferometer’s measurement trigger, the servo drive’s encoder latch, and the piezo amplifier’s setpoint update can all occur at the same moment in physical time — even though they are implemented in completely different hardware.
It is worth being precise about what EtherCAT does and does not do. It synchronizes data exchange, but it does not run control loops. The inner control loop of the piezo amplifier typically operates at 50 kHz or higher, executing locally on the amplifier hardware between EtherCAT updates. The servo drive’s current loop runs at similar speeds. EtherCAT’s role is to coordinate these independently fast subsystems.
SIOS Laser Interferometer EtherCAT Interface
Practical consequences of synchronization
When all three systems share a time base, several things become possible that were previously difficult or impossible:
1) Interferometer-guided positioning becomes practical in a real-time control loop. The motion controller receives a new interferometer measurement every bus cycle, with a known timestamp, and can issue a corrective setpoint to either the servo drive or the piezo amplifier based on the actual payload position not the encoder reading of the stage mechanism.
2) Cross-correlation of datasets becomes reliable. In process control, manufacturing, or metrology applications, it is often necessary to correlate the position of the stage with a measurement from a sensor (force, optical, electrical). When all data shares a hardware timestamp accurate to less than a microsecond, this correlation is exact. When it does not, it is an approximation that introduces uncertainty proportional to the speed of motion.
3) System diagnostics and tuning become more tractable. A servo engineer tuning a cascaded coarse-fine system needs to understand how the stages interact dynamically. Synchronized data capture, where all signals are sampled coherently, makes this analysis possible. Asynchronous data makes it much harder.
Architecture Principles, Not Product Prescriptions
The combination described in this article represents a class of architecture, not a single fixed design. The specific choice of servo drive, piezo amplifier, interferometer model, and EtherCAT master controller will vary with the application. What does not vary is the underlying logic:
- Separate the motion tasks: coarse travel, fine correction, and absolute measurement are physically different problems requiring different solutions.
- Match the measurement to the error source: encoder feedback cannot correct Abbe errors, thermal expansion of the scale, or structural compliance. Interferometric feedback can.
- Synchronize the control loops: the performance of a multi-layer system is limited by how coherently the layers communicate. A shared time base is the prerequisite for coordinated control.
- Design for the error budget: each layer’s residual error must fall within the correction range of the next. This requires analysis, not just component selection.
The practical challenge is that these principles require multidisciplinary expertise: servo control engineering, piezo system design, optical metrology, and real-time network architecture. This expertise is rarely concentrated in a single component supplier, and the gap between individual component performance and integrated system performance can be significant.
Summary
Air bearings, piezo actuators, and laser interferometers form a complementary hierarchy based on fundamental physics. The frictionless, high-stiffness motion of an air bearing stage provides the mechanical stability necessary for a piezo actuator to operate optimally. By leveraging the high bandwidth and sub-nanometer resolution of piezos, the system achieves performance levels far exceeding the capabilities of a standalone servo drive. The laser interferometer serves as the primary feedback mechanism, providing the traceable measurement required to close the control loop at the payload level.
EtherCAT provides the necessary infrastructure to harness the capabilities of the independent systems in a unified solution. Through deterministic communication, a synchronized distributed clock, and a unified data plane, EtherCAT allows disparate components to function as a singular, cohesive unit. Without this shared time base, these devices would operate in isolation, limiting the system’s overall dynamic performance.
The primary engineering challenge lies in the management of the interactions of the individual components. Successful implementation requires a rigorous error budget analysis, careful tuning of cascaded control loops to ensure stages cooperate rather than compete, and a robust validation strategy against traceable standards. This integration process represents the core engineering effort in high-precision system design.
Want to learn more?
Address
1 Cabot Road, Suite 210 Hudson, MA 01749
Call Us
1-508-634-6688