Present-day machining systems make use of rotary tables to tilt and index a part. The positioning accuracy of the rotary table is an integral element of the accuracy of a system. If the rotary table is out of position, even by a measure of 10 arc seconds, then a to-be-machined part with a radius of 20 cm (8") could be subject to an error of 0.01 mm (0.0004") to a feature location.
Similar difficulties are faced while using a rotary table for inspection. Specifically, if the rotary table is positioned on a coordinate measuring machine (CMM), then the accuracy of the table highly contributes to the accuracy of the entire system. In order to make sure that the system accuracy is high, the angular errors of the rotary table have to be carefully verified while carrying out machine qualification trials. An autocollimator is a handy and economical tool to quantify errors of CMMs and machine tools. It can verify encoder and positioning errors as well (Figure 1).
Figure 1. An autocollimator is a very useful economical tool to quantify errors of machine tools and CMMs.
An autocollimator is an easy-to-operate optical device that is designed to achieve high-resolution measurement of angles with a higher accuracy. It is equipped with angle masters, such as polygons, to measure the deviation from a nominal angle. Support software packages can also be used to apply plot results and polygon calibration deviations.
When an autocollimator is employed to measure the angular error of a rotary table, it measures the deviation from nominal angle determined by the angular master, which is typically an index table or a precision polygon mirror.
When either is placed on the table to be measured, it should be concentric with the axis within 0.5 mm (0.02"). Both include a reference diameter that can be centered using an electronic indicator. The angular masters should be positioned parallel to the rotational axis within 120 seconds or better. This way, the autocollimator can be employed to measure the parallelism deviation.
Despite the fact that polygons come with as many as 72 faces, the ones utilized for rotary tables usually have 8, 12, or 16 faces, and are regular polygons - they have equal angles between the faces. Due to the fact that polygons are not perfectly regular, a calibration chart including a list of deviations is provided. These angular masters have a typical calibration accuracy of 0.2 seconds.
The inner diameter is used as the reference while mounting the polygon on the rotary table so that proper alignment is ensured. The inner diameter center line is positioned square to the base and parallel to the faces. Once proper alignment is ensured, one mirror face on the polygon is rotated toward the autocollimator and zeroed. Subsequently, the rotary table readout is also zeroed.
While carrying out the inspection, the rotary table is rotated until its readout equals the nominal angle of the polygon - 45° increments for an eight-sided polygon. The subsequent mirror face must be aligned with the autocollimator, otherwise the error is read on the autocollimator. Similarly, the rotary table must be rotated to each face of the polygon until all positions are inspected. At 0°, the table must return to zero deviation.
A precision indexing table is an alternative to the polygon. The indexing table has a typical angular accuracy of 0.25 seconds. A resolution of 1° is yielded by a 360 position indexing table. Moreover, indexing tables come with any number of positions for one revolution. While using an indexing table, a plane mirror is positioned parallel to the axis of rotation and at the center of rotation.
The alignment of the indexing table is similar to that of a polygon. While carrying out inspection, for example, the rotary table is rotated to an angle of 23°, and the indexing table is counter-rotated by the same angle (Figure 2). Again, if the mirror is not aligned, then the error is read on the autocollimator.
Figure 2. Principle of checking an indexing table with Ultra Autocollimator.
The same inspection can be carried out by employing a high accuracy clinometers such as the Taylor Hobson TB100 (Figure 3) in place of the precision indexing table.
Figure 3. Checking a machine tool indexing head using a Taylor Hobson TB clinometer on its back, with a small reflector to enable any position from 0° to 360° to be measured.
Most of the machine tool rotary tables include optical encoders mounted on the axis of rotation of the tables. As the principal sources of error of the optical encoders are the eccentricity and the tilt of the grating with respect to the axis of rotation and not the lines on the grating, they are considered to be highly accurate devices. As both these errors are sinusoidal in shape, an 8- or 12-sided polygon can be used to verify the accuracy.
Another example is a rotary table that is driven by a gear and worm, where an optical encoder is positioned on the worm shaft. Here, the error pattern is the product of the gear errors, the worm errors, and the encoder errors. In this case, when the pitch on the ring gear is 2°, then the periodic worm and encoder errors can be explored by testing eight places in the table that are 15 minutes apart.
Here, periodic errors are errors that are repetitive in a fixed increment. Then, the ring can be checked in 2° increments. If the table is in use, then the error is the sum of the ring and worm errors.
Two types of periodic error are exhibited by the Inductosyn rotary scale. The first type of error occurs due to the eccentricity and/or tilt of the scales with respect to the rotational axis. Similar to errors caused by optical encoders, these errors occur once per revolution and can be read by inspecting greater than or equal to eight points. The second type is periodic in one pole of the Inductosyn scale.
Since there are 180, 360, or 720 poles per revolution, the scale should be tested accordingly at least at eight points in one pole. For qualifying Inductosyn scales, an index table can be employed by stacking a 360 position table and a 375 position table. The use of a counter rotation can enable one table to go forward by 1° and then come back by 0.96° (360/375) on the other table, generating multiples of 144 seconds.
All of the processes discussed work equally well for tilting tables. A polygon or index table can be employed to inspect the accuracy of tilt motion of the main axis. One must exercise caution in case an offset arises between the point of inspection of a tilt axis and the point of use. As no structure is known to exhibit infinite stiffness, the offset could be a source of error. In such cases, it is highly desirable to use a precision clinometers to verify the angle of the worktable with the fixture and part in place.
This information has been sourced, reviewed and adapted from materials provided by Taylor Hobson Limited.
For more information on this source, please visit Taylor Hobson.