Movable components, essential to any optics laboratory, include translation stages, rotation stages, and positionable mirror and optics mounts. Movable mounts typically feature adjustment screws for positioning and springs to preserve alignment and remove backlash. Often these mounts may be automated by replacing manual adjustment screws with powered actuators.
Degrees of Freedom
Figure 1. The ball, groove, and flat construction of a simple kinematic mount constrains motion in all six axes.
Three linear motions (x, y, z) and three angular motions (Θx, Θy, Θz) are necessary to describe fully the motion and position of a solid body in space. When each of these degrees of freedom is singularly constrained, we refer to the mount as kinematic. There are three types of kinematic mounts in common use. Simple kinematic mounts use a cone or ball, a groove, and a flat in the front plate of the mount to constrain unwanted motion (see figure 1). The ball constrains motion in the x, y, and z axes. The groove constrains motion in Θy (pitch) and Θx (roll). The flat constrains motion in Θz (yaw). This principle is used for mirror mounts and mounting plates. A mirror mount typically accomplishes angular adjustment by drive screws that push against the groove and flat. Coil springs pull the plate against the drive screws.
Flexure mounts differ from simple cone-based mounts in that they use leaf springs (flexures) to constrain the mirror mounting plate. These springs constrain the motion in y, z, and Θx, leaving the drive threads free to control the remaining three degrees of freedom. Flexure mounts provide performance superior to that of simple kinematic mounts because the restraint on roll is more robust, and the weight of the mirror and mirror plate is supported directly by the flexures, eliminating all in-plane motion. In simple mounts, unless the mirror mount is pointing vertically, gravitational pull must be counteracted by the drive threads; consequently, any play, no matter how small, results in pointing error and wobble.
The chief disadvantage of both cone and flexure mounts is that some cross-coupling occurs between the linear and rotational axes because the rotational axes are not centered on the surface of the mounted optic. This generally presents no problems in a retroreflection application, but can cause significant translational error in a beam-steering application. Gimbal mounts provide angular adjustment without translation. The axes of rotation for gimbal mounts are orthogonal, noninteracting, stationary, and centered on the optic. Gimbal mounts are used for the most precise beam-control applications. Translation Stages
We refer to movement along a single planevertical, horizontal, or even rotationalas translation. There are several ways to move a stage or platform in this manner (see figure 2). Dovetail slides offer a simple and effective solution for systems requiring long travel, and are often found in low-cost systems intended for infrequent motion. With appropriate preloading, matching, and way-surface hardening, they are appropriate in a wide range of applications. Dovetail stages are inappropriate for high-precision systems because of their high friction and stiction (breakaway friction). In addition, because the slide depends on effective, whole-surface lubrication, the stage floats, with a consequent lack of microscopic definition.
Figure 2. The three most common flexure mechanisms are dovetail slides, ball bearings, and crossed-roller bearings.
Ball-bearing stages replace the friction of sliding motion with the lower friction of a rolling motion. A linear array of spherical balls is held between V-grooves or rails with a cage that prevents adjacent balls from touching one another. To minimize wobble, the rails are preloaded to apply pressure uniformly along the bearing. Because there is a very small area of contact between ball and rail, small microscopic bumps influence stage motion. To eliminate play and wobble, all small irregularities must be effectively smoothed out by applying preload. Ball bearings require lubrication to minimize metal-to-metal contact, thereby preventing seizing and reducing friction and stiction. They have a self-cleaning action whereby any tiny dirt or dust particles are simply pushed clear by the rolling and squeezing motions. Ball-bearing stages are not recommended for high loads. Because balls make point contact with the rails, at high loads the ball bearings will indent the rails, causing permanent damage.
Crossed-roller-bearing stages replace balls with small steel rollers that are held apart from one another by a cage to prevent adjacent rollers from touching. Because the axes of rotation alternate or cross at 90°, the stage can be preloaded and will operate at any angle. Cross-roller bearings have larger load-bearing surfaces than ball bearings (line vs. point contact) and can therefore tolerate a higher preload, carry greater weight, and meet very tight runout specifications. This is achieved, however, at the expense of higher friction and stiction.
Contamination is a particular problem with cross-roller bearings. Dirt is rolled over and trapped on the rollers and tracks, continuously building up and grinding into the surface. The resulting increase in friction and stiction further limits ultimate resolution, and, with time, leads to erratic motion.
Unlike dovetail slide, ball-bearing, and cross-roller bearing stages, flexure stages rely on elastic deformation of a solid material. As a result, both friction and stiction are negligible, and resolution and repeatability can be as small as a few nanometers. Flexures are without equal when ultimate performance is the goal and the application can accept two significant constraintsarcuate motion and limited translation. When used in translation stages, simple flexures approximate straight-line movement with a circular path, so there can be a second-order cross coupling between the horizontal and vertical axes (see figure 3).
Figure 3. Flexures provide quasilinear translation without stiction and friction, but flexure-based-translational motion is inherently arcuate.
In automated multiaxis systems, this approximation is rarely a problem because the amount of arcuate motion can be accurately calculated and compensated for. Compound-flexure stages are somewhat more complex, but essentially eliminate arcuate motion. A more pressing issue with flexure stages is their limited translation, often only a few millimeters. One common compromise is to combine a precision flexure stage (for resolution) with a bearing stage (for extended travel).
Air-bearing stages are often used when ultraprecise resolution and repeatability must be combined with a relatively long travel range (on the order of inches). In these stages, the moving platform floats on a layer of pressurized air that is a few microns thick, essentially eliminating friction and smoothing out any mechanical imperfections in the bearing surfaces.
Components can be translated vertically (elevated) in a variety of ways. Adjustable post holders using rotating threads or rack-and-pinion gears are commonly used to move optical elements, such as lenses, vertically. For higher resolution, a horizontal translation stage can be mounted vertically with an "L" bracket. The most popular precision vertical translator uses a micrometer head to drive a pivoting cam. This cam converts the horizontal motion of the micrometer head to a vertical motion. The cam lifts a top plate that is constrained by a bearing assembly to travel only vertically.
In the laboratory, it is common practice to build up a multiaxis stage by combining the appropriate number of single-axis stages, one on top of the other. This is often acceptable for a quick experiment or a temporary setup, but should be avoided for precision applications. With the addition of each stage, the stiffness, accuracy, susceptibility to vibration, and overall performance of the combination degrades. Furthermore, the dimensions of the combined system can be become quite cumbersome. Compact, integrated multiaxis stages provide travel over six orthogonal degrees of freedom, each with the precision, accuracy, and repeatability of a single-axis stage. Manual and Powered Actuators
Actuators are the devices that physically apply the force that causes the moving plate on a translation stage or mirror mount to move. In the absence of friction, stiction, and travel imperfections, the actuator defines the resolution and repeatability of a device. There are three common types of manual actuators in usethumbscrews, standard micrometers, and differential micrometers. The resolution of these devices is determined by the quality and effective pitch of the thread (the smaller the pitch, the higher the resolution) and the size of the adjustment knob (the larger the knob, the higher the resolution). Thumbscrews with fine threads can typically provide resolution of a few microns; repeatability is less precise. Standard micrometers integrate fine internal threads with an outside scale, allowing them to produce precise, repeatable displacements. Resolution is typically less than a micron with repeatability on the order of a few tens of microns. Differential micrometers use reverse-spiral threads of different pitch. The effective pitch of the micrometer is the difference of the two pitches, providing resolution as low as a 50 nm and repeatability of less than 1 µm.
When it becomes necessary to automate a process, powered micropositioners (high-resolution stepper motors) can be used to drive a precision leadscrew to generate a linear or rotary motion. Stepper motors provide higher precision than conventional dc servo motors, since the motor is driven by a series of current steps that can be counted so that the final position is known well within a resolution of one step. By using microstepping techniques, these steps can be accurately subdivided into many smaller increments of equal size, increasing resolution. Powered micropositioners can achieve resolution and repeatability of 25 nm or better.
Piezoelectric actuators provide even higher resolution motion. The main drawback of these devices is the extremely low distance of travel (e.g., 20 µm), which means they usually must be coupled with another actuator. Although piezoelectric actuators exhibit some hysteresis and nonlinearity, proper drive electronics can often compensate for these effects, yielding resolutions of 20 nm or less for devices without feedback, and 5 nm or less for devices with feedback.
Positioning is a fundamentaland frequently criticaltechnology in many applications. For optimal results, match the positioning equipment to your application parameters and performance requirements. oe
Rich Sebastian is the product manager for optomechanical hardware at Melles Griot, Carlsbad, CA.