Part 2 of series appears in April 1994 issue of PT Design. Electronic, optic, computer, inspection, automation, and laser industries require diverse positioning-system specifications. No one system is right for all.
To ensure that a high-accuracy positioning system works optimally, the components that make up the system — bearings, position-measuring system, motor-and-drive system, and controller — must all work together as well as possible to meet the application criteria.
Base & bearing
To decide on the optimum system configuration, consider the mechanical part of the system first. For linear stages, these are the four common base-andbearing design choices:
• Aluminum base and slide with bolton ball-bearing ways.
• Aluminum or steel base and aluminum or steel side with four recirculating roller-bearing blocks on steel rails.
• Meehanite cast iron base and slide with integral roller-bearing ways.
• Granite guides with granite or cast iron slide and air bearings.
Aluminum is lighter than meehanite or steel but less stiff, less stable, less able to take a beating, and less stress-resistant. In addition, aluminum is much more sensitive to temperature changes. Cast iron is 150% stiffer than aluminum and 300% better in vibration damping. Steel is durable and stronger than iron. However, it suffers prolonged ringing, which is detrimental to fast move and settle times. See Figure 1.
Granite guides with air bearings provide the stiffest, most durable combination. Granite can be polished for flatness and straightness in the submicron range. The drawback to a granite table is that, because of the granite’s mass, it has a larger space envelope and weighs more than a steel or iron-based positioning system. However, because there is no contact between bearings and granite guide surfaces, there is no wear, and air bearings are largely self cleaning. Also, granite has excellent vibration damping characteristics and thermal stability.
In addition, design of the table itself is important in the table’s overall performance. Tables come in a variety of configurations from bolt-together units with many parts to simple cast bases and slides. Use of one material throughout the table generally provides more uniform response to temperature variations, leading to a more accurate system. Features such as ribbing provide damping, which enables swift settling.
Integral ways have an advantage over bolt-on ways in that even after a long time, no adjustment of ways for preload is needed.
Crossed roller bearings, Figure 2, have line contact between roller and raceway, whereas ball bearings, Figure 3, have point contact between ball and raceway. This generally results in smoother motion for roller bearings. There is less surface deformation (and wear) over the rolling surface and there is a greater contact area, so load is distributed more evenly. Loads up to 4.5 to 14 kg/roller are standard, along with high mechanical stiffness of about 150 to 300 Newtons/micron. Disadvantages include inherent friction from the line contact.
The small contact area that limits the ball bearing’s friction, however, also limits its load capacity. Roller bearings generally have longer lives than ball bearings. However, roller bearings cost more.
Standard table sizes of one manufacturer include 25 to 1,800-mm length and 100 to 600-mm slide width.
An air bearing configuration, Figure 4, consists of lift and guide bearings preloaded by opposing air bearings or by high-force rare-earth magnets imbedded in the guiding members. This noncontact design avoids the friction of other bearing designs. Also, air bearings suffer no mechanical wear. Furthermore, air bearings can be spaced widely apart. Thus, resulting geometric errors are averaged, producing angular deviations of less than 1 sec of arc and straightness of better than 0.25 micron over 200 mm.
Continue on Page 2
Figure 5 presents a graphical representation of general characteristics of bearing systems, because numerical values are difficult to provide — they depend on many factors. For instance, positioning accuracy depends not only on the bearings or guides, but also on the position- measuring system and the controller. Friction in a positioning system depends not only on which drive system you chose, but also on bearing adjustment, table sealing, lubrication, and so forth. Therefore, the exact values that can be reached depend very much on the combination of all components, which in turn depends on the application.
Of the many types of drive systems — belt, rack-and-pinion, lead screw, precision-ground ball screw, and linear motor — only the last two are considered for most high-accuracy positioning systems.
Ball screw drives come in a range of resolution, precision, and stiffness characteristics, and can provide high velocities (above 250 mm/sec). However, because the ball screw drive is limited by the screw’s critical rotary speed, higher velocity requires a lower pitch, with less mechanical advantage and a higherpower motor. This usually means changing to a higher-power motor drive with higher bus voltage. Ball screw drives, though widely used, may also suffer mechanical backlash, wind-up, pitch cyclic errors, and friction. Also overlooked is stiffness of the mechanical coupling that joins motor and drive.
With the linear servomotor, the electromagnetic force directly engages the moving mass with no mechanical connection. There is no mechanical hysteresis or pitch cyclical error. Accuracy depends entirely on the bearing system and the feedback control system.
Dynamic stiffness indicates how well a servo system maintains position in response to an impulse load. In general, greater bandwidth and higher gain provide greater dynamic stiffness. This can be quantified by dividing the measured impulse load by the deflection distance:
Dynamic stiffness = ΔF/ΔX
The high stiffness and high natural frequency result in excellent servo behavior with short settling times. The slide reacts quickly to change in position commands because there is no mechanical linkage between motor and slide. Also, because there is no ball screw “ringing,” fast move and settle times can be achieved.
A brushless linear motor consists of a permanent magnet assembly fixed to the machine base, and a coil assembly fastened to the slide. A gap of about 0.5 mm is maintained between the coil assembly and the magnets. There is no physical contact between the two assemblies.
The core of the moving coil assembly houses a series of overlapped and insulated copper coils. These are precision wound and pitched for three-phase operation. A series of Hall Effect sensors is used for the electronic commutation. The design of the commutation electronics provides motion with negligible force ripple. Because commutation is electronic rather than mechanical, commutation arcing is eliminated.
Those properties make a linear servomotor useful in applications requiring high acceleration (say 2.5 m/sec2 or more), high velocity (say 2 m/sec or more), or precise velocity control, even with very low speed (say just a few mm/sec). Moreover, such a motor needs no lubrication or other maintenance and has no wear. As with any other motor, because of heat dissipation the rms value of continuous force or current must not exceed allowable values for long periods.
You can get linear servomotors in continuous drive forces of 25 to more than 5,000 N. Most larger motors have air or water cooling. Multiple linear motors can be connected in parallel or series arrangement to get higher drive forces.
Because there is no mechanical linkage between motor and slide, there is no mechanical reduction such as there is with a ball screw. The load transfers in a 1:1 ratio to the motor. With a ball screw drive, load inertia on the slide to the motor is reduced by the square of the reduction ratio. This makes the linear motor drive less suitable for applications with frequent load changes unless you choose a controller that you can program with different sets of motor control parameters corresponding to different loads to get effective servo compensation.
Continue on page 3
For many vertical applications, a ball screw is easier and more cost-effective — the linear motor must be continuously energized to offset gravity. Also, an electromechanical brake can lock table position when power is off. You can use a linear motor, however, if you offset the motor and load weight with a spring, counterweight, or air cylinder.
In initial cost, there is little difference between a linear motor drive and a ball screw drive that includes motor, couplings, bearings, bearing blocks, and ball screw. In general, a brush type linear motor is slightly cheaper than a ball screw drive, and brushless versions are usually somewhat more expensive.
There is more to consider than initial cost. A more realistic comparison includes maintenance, reliability, durability, and replacement costs, including labor. Here, the linear motor shows well.
Figure 6 presents the most important characteristics of drive systems.
Part 2 will cover position-measuring systems.