How to Select the Right Nanopositioning System
Nanopositioning device: Drive system & guiding system are frictionless and not limited in resolution.
Micropositioning device: Other drive and guiding systems, such as scew drives, inertial drives, friction drives, ball bearings, roller bearings, etc.
Piezo Nanopositioning Stages, Air Bearing Stages, Magnetic Bearings, and Traditional Micropositioning Stages
There are several ways to achieve nano-precision motion. The best positioning systems avoid friction all together, in both the drive system (motor) and in the guiding system (bearings). Frictionless bearings also avoid the bearing rumble caused by balls and rollers and provide vibration-free motion with highly constant velocity.
For short travel ranges, piezo drives with frictionless flexure guidance are the technology of choice. They combine fast response, extreme guiding precision, very long, maintenance-free service life and can easily achieve sub-nanometer step sizes. Due to the high stiffness and low inertia, piezo flexure stages can also achieve extremely fast step and settle times in the millisecond or microsecond range and high scanning rates with hundreds or thousands of Hz, both important features in optics, alignment and semiconductor test and manufacture.
For longer travel, positioning stages with frictionless air bearings and linear motors are available.
More on air bearing stages
Another option to go frictionless is known as magnetic levitation (magnetic bearings).
Watch a video demonstration of a multi-axis mag-lev stage by PI here
Positioning stages with conventional guiding systems (crossed roller bearings, ball bearings, etc.) and drive systems (leadscrews or ball screws) cannot provide the same performance in terms of geometric precision, responsiveness and resolution, but they also have many applications. For more examples on motorized stages, click here
For detailed information on piezo flexure-based nanopositioning and scanning systems, read on:
Low Inertia = Higher Speed
With piezo drives, capable of accelerations of up to 10,000 g, and their
low moved mass, such piezo stages can provide significantly higher
scanning speeds than motorized systems.
Flexure motion is based on the elastic deformation (flexing) of a solid
material. Friction and stiction are entirely eliminated, and flexures
exhibit high stiffness, load capacity and resistance to shock and
vibration. Flexures are maintenance free and not subject to wear. They
are vacuum compatible, operate over a wide temperature range and require
neither lubricants nor compressed air for operation.
The latest flexure guided nanopositioning stages provide travel ranges close to 2 mm (see P-629 PIHera® stages). For longer travel ranges, the novel
NEXLINE® and NEXACT®
nanopositioning motors are available. These motors were developed to
overcome the limitations of nanopositioning devices in the semiconductor
Zero-Runout Multilink Flexures: Excellent Guiding Accuracy
The multilink flexure guiding systems employed in most PI piezo
nanopositioners (Fig. ) eliminate cosine errors and provide
bidirectional flatness and straightness in the nanometer or microradian
range. This high precision means that even the most demanding
positioning tasks can be run bidirectionally for higher throughput.
Lifetime / Ceramic Encapsulation
PI nanopositioning systems employ the patented, award-winning PICMA® piezo actuators, the only actuators with cofired ceramic encapsulation. The PICMA® piezo technology was specifically developed by PI’s piezoceramic division to provide higher performance and lifetime in nanopositioning applications.
Multilayer piezo actuators are similar to ceramic capacitors and are not
affected by wear and tear. PI nanopositioning systems are designed to
be driven at lower voltages than most other piezo systems (100 V vs. 150
V). The research literature, PI’s own test data and 30+ years of
experience all confirm that lower average electric fields, lead to
In addition PI’s monolithic ceramic-encapsulated design provides better
humidity protection than conventional polymer-film insulation.
Diffusion of water molecules into the insulation layer is
greatly reduced by the use of co-fired, outer ceramic encapsulation
(fig. ). Humidity is the main influence on the longterm
reliability in low-dynamics or quasi-static operation modes,
where the piezo actuator is supplied with a DC voltage to
maintain a position for a long time.
100 Billion Cycles - Lifetime Tests Prove new Actuator Design
Comparative tests with both PICMA® and conventional
piezo actuators have proven the positive effects of the ceramic encapsulation.
While polymer-coated piezos typically only survive 30 days
of continuous operation - PICMA® actuators are still working
after many years. Other tests by NASA have shown 100 Billion cycles can be reached with no failures.
To suit an application requiring
10 years minimum lifetime
under cryogenic conditions,
accelerated lifetime tests with
PICMA® piezo actuators have
been successfully performed.
Inserted in a cryogenic bath of
liquid nitrogen (75 K), the piezo
is placed in a vacuum chamber
(2 • 10-3 mbar) and subjected to
dynamic operation at 90% of
the maximum voltage range
(>105 V) with an operating frequency
up to 1000 Hz. After one
month of continuous operation
there were no degradations in
piezo performance to be measured,
neither mechanic concerning
the displacement, nor
electrical concerning electrical
capacitance or resonant frequency.
(Dr. Bosotti et al.,
University of Milano, Italy,
Video: Piezo Nanopositioning Systems Explained
Ultra-low-profile (6 mm), closed-loop multi-axis piezo scanning stage
for image resolution enhancement (pixel-sub stepping, dithering). The
parallel-kinematics design enables the compact dimensions and low
inertia (= fast response).
Fig. Simple parallelogram flexure guiding system with motion
amplification. This design is for run-out non-critical applications.
The amplification r (transmission ratio) is given by (a+b)/a.
Fig. For the highest precision, zero-arcuate-error (anti-runout) multi-flexure guiding systems are required.
Fig. Wire-EDM cutting process provides highest-accuracy flexure
guiding systems in compact nanopositioning stages. The picture here
shows a relatively simple guiding system
Fig. Example of a multi-dimensional, zero-arcuate flexure for long
travel ranges (much longer than feasible with classical piezo stack
drives). The Roberts-linkage flexure system depicted here provides
10x10x10mm XYZ multi-dimensional motion, while preventing parallelogram
Measuring Nanometers: Stage Metrology Selection
Achieving nanometer and subnanometer precision requires more than a
piezo stage capable of making moves on this precision scale. The stage
internal metrology system must also be capable of measuring motion on
the nanometer scale. The five primary characteristics to consider when
selecting a stage metrology system are linearity, sensitivity
(resolution), stability, bandwidth, and cost. Other factors include the
ability to measure the moving platform directly and contact vs.
noncontact measurement. Three types of sensors are typically used in
piezo nanopositioning applications—capacitive, film strain gauges, and PRS (piezo resisitve) strain gauges. Table 1
summarizes the characteristics of each sensor type.
PI capacitive sensors: measure the gap between two plates based
on electrical capacitance. These sensors can be designed to become an
integral part of a nanopositioning system, with virtually no effect on
size and mass (inertia). Capacitive sensors offer the highest
resolution, stability, and bandwidth. They enable direct measurement of
the moving platform and are noncontact. Capacitive sensors also offer
the highest linearity (accuracy). PI's capacitive sensors / control
electronics use a high-frequency AC excitation signal for enhanced
bandwidth and drift-free measurement stability (subnanometer stability
over several hours, see here).
PI’s exclusive ILS linearization system further improves system
linearity. If used with PI’s digital controllers, digital polynomial
linearization of mechanics and electronics makes possible overall system
linearity of better than 0.01%. Capacitive sensors are the metrology
system of choice for the most demanding applications.
Johnson Noise Free:
Capacitive sensors are not plagued by the intrinsic noise found in all
resistance type sensors (Johnson noise) such as piezo-resistive or other
resistive sensors. They offer advantages due to their well-understood
characteristics, and high linearity
Strain gauge sensors: There are two commonly used types of strain gauge sensors: piezo-resistive
(semicoductor strain gauges) and film type sensors. They derive
position information from a change in their resistance when stretched or
compressed. They need to be bonded to a piezo stack or—for enhanced
precision—to the guiding system of a flexure stage. Either type strain
gauge sensor offers high resolution and bandwidth and is typically
chosen for cost-sensitive applications. As a contact type sensor, it
measures indirectly, in that the position of the moving platform is
inferred from a measurement at the lever, flexure or stack. PI employs
full-bridge implementations with multiple strain gauges per axis for
enhanced thermal stability. PI's PICMA® drive technology also enables
higher performance of actuator-applied strain gauge sensors.
PRS : Piezoresistive sensors provide very high resolution but cannot match the linearity and stability of capacitive sensors. PRS are for example employed in PI's P-545 PInano stage series for
optical microscopy applications. For many optical microscopy
applications the stability and linearity of piezoresistive sensors is sufficient.
Resolution of all resistance type sensors such as piezo-resistive or other strain gauge sensors is limited by the intrinsic Johnson noise.
||Stability* / Repeatability
||Direct / Noncontact
||Inferred ** (Indirect) / Contact
||Inferred ** (Indirect) / Contact
* Note. The ratings describe the influence of the sensor on the
performance of the whole nanopositioning system. Resolution, linearity,
repeatability, etc. specifications in the PI product data sheets
indicate the performance of the complete system and include the
controller, mechanics and sensor. They are verified using external
nanometrology equipment (Zygo Interferometers). It is important not to
confuse these figures with the theoretical performance of the sensor
** Strain type sensors (metal foil, semiconductor, or piezoresistive) infer position information from strain.
Dynamics, System Behavior, Frequency Response / Step Response
One of the advantages of piezo systems is their ability to react rapidly
to an input signal. As a rule of thumb, the higher the resonant
frequency of the system, the faster the response. For open-loop systems
the shortest rise time is approximated by
Tmin = time [s]
f0 = resonant frequency [Hz]
This theoretical value (basically switched operation) requires an
amplifier with sufficient output current and slew rate and would induce
significant amounts of overshoot and ringing. It is not the fastest way
to get to a stable position.
Read Methods to Improve Piezo Dynamics for more information.
For closed-loop systems, the performance of the servo controller
also plays a significant role. Digital controllers with advanced
filtering techniques can provide much faster response (while eliminating
keeping overshoot and oscillation) than simple analog controllers.
The phase response of a piezo actuator system can be approximated
by a second order system and is described by the following equation:
j = phase angle [deg]
Fmax = resonant frequency [Hz]
f = operating frequency [Hz]
More information on piezo actuators and motors can be found in the piezo motion tutorial. and in the Dynamic Operation section on the PI Ceramic Website
Fig. A nanopositioning system can only be as good as the equipment
that was used to test and optimize it. To test and measure runout,
dynamic crosscoupling and stability at the nanometer level, extensive
nanometrology equipment is required. PI has designed several
room-in-room nanometrology-labs with multiple thermal, acoustic and
seismic isolation. Location of the labs in the basement, with specially
designed, separated foundations is key to obtaining meaningful
sub-nanometer precision test results.
Fig. Industry-standard Zygo ZMI-2000 interferometers are used in
PI´s nanometrology calibration labs. Each stage is individually
calibrated and optimized for dynamic response.
Fig. Quality control: Fully randomized position repeatablity
measurement of a direct-metrology-equipped 100 µm PIHera™ stage. Data
taken with Zygo interferometer. The data show the exceptional precision
of these mid-level nanopositioning devices.
Fig. Closed-loop step response of a P-713 piezo flexure stage, measured with Zygo interferometer
Fig. Measured phase and amplitude response of a high-resonant
frequency (2.2kHz) piezo flexure stage. The phase response of a piezo
system can be approximated by a second order system.
Effects of the
interaction of the PZT ceramics, preload, flexures and couplings show up
in the graph and the result is different from a straight line.
Beware of perfect, straight line graphs. They do not represent the performance of a system but the wishful thinking of the designer.
Parallel and Serial Kinematic Stage Designs
There are two ways to achieve multi-axis motion: parallel and serial
kinematics. Serial kinematics (nested or stacked systems) are simpler
and less costly to implement, but they have some limitations compared to
parallel kinematics systems (see here for more information).
Serial Kinematics for Lower Cost / Standard Applications
In a multi-axis serial kinematics system, each actuator (and usually
each sensor) is assigned to exactly one degree of freedom. Serial
kinematic designs have advantages when it comes to cost and simplicity,
and work very well for most standard applications.
However, the manufacturing precision of even the best machines
(and technicians) does not allow for crosstalk-free mechanics, at least
not at the nanometer level, even in quasistatic operation. For many
applications this is not a problem at all. In a parallel kinematics
multi-axis system, all actuators act directly on the same moving
platform (relative to ground), enabling reduced size and inertia, and
the elimination of microfriction caused by moving cables (Fig. ). This
way, the same resonant frequency and symmetrical dynamic behavior can be
obtained for both the X and Y axes. The advantages are higher dynamics
and scanning rates, better trajectory guidance as well as better
reproducibility and stability. Still, without the proper test metrology
/ instrumentation these advantages can easily be overlooked, and a
serial kinematics system may appear to be perfect. For an inexperienced
designer it could then be very tempting to draw the conclusion there is
no cross talk (see also "Coupled and Uncoupled Motion..." below).
Parallel Kinematics for Highest Multi-Axis Precision / Dynamics.
For the ultimate in dynamics straightness/flatness/orthogonality and for
coordinate transformation situations (such as rotating about an
arbitrary point in space), parallel kinematics is required. Parallel
Kinematics systems, if designed well, are superior to serial kinematics
systems. A comparison that comes to mind is the electronic computer and
the slide ruler. Both can be used to solve many problems, but a computer
can do more, faster. To reap the benefits of parallel kinematics, you
must have high-bandwidth, direct drivetrain-output metrology, so the
workpiece can be simultaneously observed and controlled in multiple
degrees of freedom.
Because it takes more knowledge, experience and more advanced
controllers to produce a good parallel kinematics precision positioning
system, inexperienced designers may avoid that challenge dismissing the
technology as unpredictable, uncontrollable. Everyone who has seen a
6-axis motion system compared to a classical stack of translation and
rotation stages will immediately understand the benefits of parallel
Coupled and Uncoupled Motion, Static and Dynamic
Serial kinematics are sometimes referred to as uncoupled motion systems
vs. coupled motion systems for parallel kinematics. When it comes the
nanometer realm, there is no uncoupled motion. By nesting or stacking a
second and/or third axis onto a translation stage, there will be an
influence on the first axis. Any load attached will have a further
influence on all axes, even when at rest. Since we are talking about
positioning, things are not always at rest, actually most
nanopositioning / scanning applications require very high dynamics. The
seemingly uncoupled multiaxis-system then provides coupled (unwanted)
motion in many degrees of freedom, that cannot be detected by its
internal serial metrology sensors and hence goes partially uncontrolled.
These errors may not be critical in many applications, but can be
detrimental in some others.
Direct Parallel Metrology: All Motion Inside the Servo Loop
Multi-Axis Measurements Relative to a Fixed Reference
Parallel kinematics facilitates implementation of Direct Parallel
Metrology—measurement of all controlled degrees of freedom relative to
ground. This is a more difficult design to build but it leads to clear
The parallel-kinematics / parallel metrology system “sees" motion in all
controlled degrees of freedom and will respond to it. This means that
all motion is inside the servo-loop, no matter which actuator (or
unwanted external force / crosstalk) may have caused it, resulting in
superior multi-axis precision, repeatability and flatness, as shown in
Fig. . Direct parallel metrology also allows stiffer servo settings for
faster response. Off-axis disturbances—external or internal, such as
induced vibration caused by a fast step of one axis—can be damped by the
Fig. Response of a PI Nanopositioning stage to a square wave
control signal clearly shows the true sub-nm positional stability,
incremental motion and bidirectional repeatability.
Measured with external capacitive gauge, 20 pm resolution.
Fig. Principle of a PI XY-Theta-Z, minimum-inertial-mass,
monolithic, parallel kinematics nanopositioning system. Accuracy,
responsiveness and straightness/flatness are much better than in stacked
multi-axis (serial kinematics) systems.
Fig. Flatness of an active-trajectory-controlled nanopositioning stage over 100 x 100 µm scanning range is about 1 nm.
Controller Choice / Interfacing
Analog and Digital Controllers
PI manufactures a large variety of analog and digital nanopositioning controllers (see here).
State-of-the-art PI digital control systems offer several advantages
over analog control systems: coordinate transformation, real-time
linearity compensation and elimination of some types of drift. Digital
controllers also allow virtually instant changes of servo parameters for
different load conditions, etc. However, not all digital controllers
are created equal. Poor implementations can add noise and lack certain
capabilities of a well-designed analog implementation, such as fast
settling time, compatibility with advanced feed-forward techniques,
stability and robust operation.
PI digital controllers can download device-specific parameters and
calibration information from ID-chip-equipped nanopositioning stages,
facilitating interchangeability of nanomechanisms and controllers.
All PI nanopositioning controllers (analog and digital) are equipped
with one or more user-tunable notch filters. A controller with notch
filter can be tuned to provide higher bandwidth because side-effects of
system resonances can be suppressed before they affect system stability.
For the most demanding step-and-settle applications, PI’s exclusive
Mach™ InputShaping® implementation is available as an option.
Improved Piezo Control: Digital Dynamic Linearization (DDL) Firmware Upgrade
Conventional piezo controllers cannot completely eliminate phaseshift
and tracking errors in applications with rapid, periodic motion. This is
due in part to the non-linear nature of the piezoelectric material, the
finite control bandwidth and the inherent limitations of P-I-D
(proportional integral derivative) servo-control, which cannot react
before a position error is detected.
The DDL firmware upgrade option (ordering number E-710.SCN), available with most digital piezo controllers such as the E-753 (single-channel) or the E-712 (multi-channel, modular),
solves this problem. This technology, developed by PI, reduces the
error between the current and desired position to imperceptible values.
The dynamic linearity and effectively usable bandwidth are thus improved
by up to three orders of magnitude (1000-fold). DDL is of benefit to
single- and multi-axis applications where motion follows a given
trajectory repeatedly (see measurement graphs, (Fig. 6a, 6b, 6c, 6d).
Choice of Interface: Digital or Analog or Both?
Analog interfacing provides high bandwidth and remains the most common
way of commanding piezoelectric motion systems. It is usually the choice
when the control signal in the application is provided in analog form. A
key advantage of analog interfacing is its intrinsic deterministic
(real-time) behavior, contrasted to the difficulty of accurately timing
high-bandwidth communications on present-day multitasking PCs.
However, when analog control signals are not available, or when a
significant distance between the control signal source and the
nanopositioning controller would affect signal quality, digital
interfacing, which must not be confused with digital control, is the
Digital signals can be transferred through copper wires, or for complete EMI immunity, through optical fibers.
Several types of digital interfaces are typically used in
piezo-nanopositioning applications: TCP/IP, parallel-port, USB, RS-232,
Fiber Link, IEEE 488 and, with some digital controllers, direct DSP
links. For dynamic, high-precision applications, the exact timing of an
interface can be significantly more important than the data transfer
Interface Bandwidth vs. Timing / Througput, NanoAutomation®
Piezo-driven stages can respond to a motion command on a millisecond or microsecond time scale.
Many throughput-driven applications require nanometer precision in milliseconds, or NanoAutomation®,
That is why synchronization of motion commands and data acquisition have
a high impact on the quality of many applications, like imaging or
micromachining. The USB, for example, was designed to transfer huge
blocks of data at high speeds, but exact timing was a much lesser
concern. While insignificant in less responsive positioning systems,
this kind of non-deterministic behavior may not be tolerable in
high-speed tracking or scanning applications. Each motion
command—comprising just a few bytes—must be transferred instantaneously
and without latency. A lower-bandwidth bus with higher timing accuracy
may perform better in many applications.
There are several factors that affect the response of a digital
interface: the timing accuracy of the operating system on the
controlling computer; the bus timing protocol; the bandwidth of the bus;
and, the time it takes the digital interface (in the piezo controller)
to process each command. Parallel-port interfaces do not require command
parsing and offer the best combination of throughput and timing
In addition, to the interface properties, the bandwidth of the
nanopositioning system (mechanics and servo) matters. A slow system
(e.g. 100 Hz resonant frequency) will not benefit from a responsive
interface as much as a high-speed mechanism.
Analog In- and Outputs
Optionally available analog inputs can be configured in a flexible way:
either as a control input—the applied voltage can be connected with one
of the available axes and is interpreted e.g. as a position value—or as
an external sensor—e. g. used as an autofocus signal instead of an
Optional analog outputs allow the control of external power amplifiers
or can be used as monitor signal interfaces. The individual data sheets
for the controllers inform about number, voltage range (usually ±10 V)
and availability of the analog I/Os.
Data Recorder: Data Acquisition and Output
The flexibly configurable data recorder enables simultaneous recording
and read-out of input and output signals, such as for sensor positions
or control voltages depending on time stamps or using trigger signals.
Wave and Profile Generator: Pre-Defined and Programmable Trajectory Profiles
Trajectory profiles of arbitrary, user-defined mathematical functions
enable complex 2-axis motion. Depending on the controller used, either
time and position data value pairs can be saved (Wave Generator) or
complete trajectory profiles with velocity, acceleration and jerk (rate
of change of acceleration) can be specified (Profile Generator). The
- Programming of complex functions
- Quick access to common functions (e. g. sine, ramps, triangle and square waves ...)
- Coordination of two axes, e. g. for applications requiring circular motion
- Saving of defined functions in the controller
All controller specific functionalities are available as DLL function
calls and LabVIEW VIs, which enables their simple integration in
external programs. Additional graphical user interfaces allow convenient
selection and customization.
CE Compliance CE & RoHS Compliance
All standard PI nanopositioning systems are fully CE and RoHS compliant.
Extensive Software Support saves Time
Software support is becoming increasingly important in nanopositioning
applications. There is a need for flexibility in the vendor's software
to enable the user to switch hardware without rewriting the code and to
upgrade or switch between various interfaces -USB, TCP/IP, RS-232, real
time parallel interfaces, even analog -without starting from scratch.
Synchronization between commanded motion and execution of the command
can be paramount in leading-edge applications.
Digital controllers come with a large variety of software tools, such as
the PIMikroMove™ graphic user interface or the NanoCapture™ system
analysis program. Additional LabVIEW drivers and DLL’s for easy setup,
system optimization and integration in application-specific programs are
included. Comprehensive documentation, Online Help and sample code
offer added value. Even analog systems can be controlled with PI’s
LabVIEW driver set, optionally with HyperBit functionality.
PI controllers and software are based on the novel GCS Universal Command
Set. This platform independet architecture decouples positioning
hardware and controller and allows seamless integration of motion
systems into application control software.
More on GCS and PI Software Tools.
Fig. PI controllers are available with a number of different
interfaces for highest flexibility. In addition to the modern Ethernet
(TCP/IP) and USB, many industrial customers still appreciate the robust
Fig. 6a. Six-axis digital piezo controller with Super Invar 6D-
nanopositioning stage. All PI nanopositioning systems and controllers
are fully CE compliant.
Fig. 6b. Rapid scanning motion of a P-621.1CD (commanded rise time 5
ms) with the E-710 controller and DDL option. Digital Dynamic
Linearization virtually eliminates the tracking error (<20 nm) during
the scan. The improvement over a classical PID controller is up to 3
orders of magnitude, and increases with the scanning frequency.
Fig. 6c. Elliptical scan with a XY piezo scanner and conventional
P-I-servo controller. The outer curve shows the desired position, the
inner curve shows the actual motion
Fig. 6d. The same scan as before but with a DDL controller. The
tracking error is reduced to a few nanometers, desired and actual
position cannot be distinguished in the graph.
Fig. Complex trajectories can be created, stored and executed with the Wave Generator tool
Fig. Six-axis parallel-kinematic nanopositioning piezo stage (Hexapod) based on NEXLINE® high-load piezo nanopositioning motors for use in strong magnetic fields.
Fig. Compact piezo ceramic (non-resonant) nanopositioning motor providing very high holding force, sub-nanometer resolution, 1 kHz resonant frequency and 20 mm travel.