 |
Description  |
|
|
This invention relates to method and system for calibrating a coordinate
measuring machine (CMM) and the like and, in particular, to method and
system for calibrating a CMM whereby, as a result of the calibration, the
CMM is compensated during its operation for its entire measuring volume.
With the advent of numerically controlled machine tools, the demand has
grown for a means to support this equipment with faster first-piece
inspection and, in many cases, 100% dimensional inspection. To fill this
need, the CMM was developed in the early 1960's. A CMM can also be used as
a layout machine before machining and for checking feature locations after
machining. In many cases the CMM plays a vital role in the mechanization
of the inspection process.
Since its development, the CMM has been increasingly used throughout the
automotive and aerospace industries. Although it was once considered an
exotic tool for ensuring quality control, the CMM is now becoming a
mandatory piece of equipment for both the large manufacturing plant and
the small job shop. This is primarily due to the need for an accurate
measuring instrument and detailed documentation of the components being
produced.
Currently, the CMM is being used in one of three ways in manufacturing. The
simplest approach is to place the CMM at the end of the production line or
in an inspection area. With this approach, the CMM is used to inspect the
first part of a production run to verify the machine setup. Once the setup
is verified, it then measures parts on a random basis. For many
applications, this permits the best approach to inspection.
Another approach is to incorporate the CMM between two work centers and
then measure 100% of the parts produced at the first center before any
secondary operations are performed at the second work center. This
approach is possible because CMMs are capable of measuring
three-dimensional geometry and making many different measurements within a
short period of time. When this approach is used, the CMM indirectly
controls the production process. In this setting, however, the CMM must be
"hardened" to perform in the shop environment for part inspection.
A third approach integrates the CMM into the production line. This permits
the CMM to directly control the production process. In operation, an
integrated system would measure the workpiece, compare the measurements
with required dimensions and, if necessary, automatically adjust the
machine controls so that the part is manufactured within the required
specifications.
A basic CMM consists of four elements: (1) the machine structure, which
basically is an X-Y-Z positioning device; (2) the probing system used to
detect part surfaces and provide input to a control system; (3) the
control system including a machine control and computer hardware; and (4)
the software for three-dimensional geometry analysis. The measuring
envelope or volume is defined by the X, Y and Z travel of the machine.
Although a variety of machine designs and configurations exist, all designs
incorporate the same fundamental concept of three coordinate axes. Each
axis is ideally square in its own relationship to the reference plane
created by the other two axes. Each axis is fitted with a linear
measurement transducer for positional feedback. This allows position
displays within the envelope to be independent of any fixed reference
point.
The most common reference systems in use are steel and glass scales Both
systems utilize noncontact, electro-optical reader heads for determining
the exact position of the machine. Steel reference systems are widely used
in shop environments because the difference in the coefficient of
expansion between the steel scale and workpiece is minimal. Glass scale
reference systems are generally used in controlled environments because of
the difference in the coefficient of expansion between glass and the metal
workpiece.
The worktable of the machine generally contains tapped holes to facilitate
the clamping and locating of parts. It may be made from granite or steel
because of its stability in various environments.
Electronic or solid probes are inserted into the probe arm or shaft&, which
is supported by cantilever, bridge gantry, column members or other CMM
types. Probe arm movement is guided by means of frictionless air bearings
or mechanical bearings.
Coordinate measuring is typically a two or three-dimensional process that
determines the position of holes, surfaces, centerlines, and slopes. Up to
six sides of a cube-shaped part may be inspected without repositioning the
part.
In a typical operation, the part is placed on the table of the CMM at a
random location. Generally, this location is approximately central to the
machine axes to access all of the part surfaces to be inspected with the
probe. Depending on the size of the part and the type of probe used, the
part may need to be clamped to the machine table. If multiple inspections
of similar parts are required, a reference location point may be
established with a reference precision cube or sphere. The probe is then
moved, manually or under machine control, until contact is made with
desired part features. Reader heads, traveling on each axis along built-in
axis measuring scales, transfer the instantaneous machine position through
the digital display and to the computer interface. The dimensional and
geometric elements may then be calculated, compared, and evaluated, or
stored, or printed out as required.
Some of the advantages of using CMMs over conventional gaging techniques
are flexibility, reduced setup time, improved accuracy, reduced operator
influence, and improved productivity.
CMMs do not need to be dedicated to any single or particular measuring
task. They can measure practically any dimensional characteristic of
virtually any part configuration, including cams, gears and contoured
surfaces.
Establishing part alignment and appropriate reference points are very time
consuming with conventional surface-plate inspection techniques. These
procedures are greatly simplified or virtually eliminated through software
available on computer-assisted or computer-controlled CMMs.
Such software allows the operator to define the part's orientation on the
CMM, and all coordinate data are subsequently automatically corrected for
any misalignment between the part reference system and the machine
coordinates. A CMM with sophisticated software can inspect parts in a
single setup without the need to orient the part for access to all
features even when a fourth axis (rotary table) is employed.
All measurements on a CMM are taken from a common geometrically fixed
measuring system, eliminating the introduction and accumulation of errors
that can result with hard-gage inspection methods and transfer techniques.
Moreover, measuring all significant features of a part in one setup
prevents the introduction of errors due to setup changes.
The use of digital readouts eliminates the subjective interpretation of
readings common with dial or vernier-type measuring devices. Operator
"feel" is virtually eliminated with modern electronic probe systems All
CMMs have canned software routines for typical part features, such as
bores or center distances. In the part-program-assisted mode, the operator
positions the machine; once the initial position has been set, the machine
is under the control of a program that eliminates operator choice. In the
computer numerically controlled (CNC) mode, motor-driven machines run
totally unattended by operators. Also, automatic data recording, available
on most machines, prevents errors in transcribing readings to the
inspection report. This all adds up to the fact that less skilled
operators can be readily instructed to perform relatively complex
inspection procedures
All the factors previously mentioned help to make CMMs more productive than
conventional inspection techniques. Further dramatic productivity
improvements are realized through the computational and analytical
capabilities of associated data handling systems, including calculators
and all levels of computers.
A variety of machine configurations are available from the manufacturers of
CMMs. Each configuration has advantages that make it suitable for
particular applications. A total of 11 different machine configurations
exist; however, some of these configurations are modifications of one of
the five primary configurations: cantilever, bride, column, gantry, and
horizontal arm.
The utility of a CMM depends largely on the nature of the probing device.
Three types of probes are commonly used: (1) hard; (2) electronic, and (3)
noncontact. A probe is selected according to the dimensional and
geometrical requirements of the inspection process.
Various accessories used in conjunction with the probes enhance the
capability of CMMs. For example, indexable probe heads permit orienting
the measuring probe in horizontal and vertical planes to keep the probe
normal to the plane desired. This feature gives the CMM the capability to
reach and inspect geometrical elements that are not aligned to the machine
axes. In addition, the use of indexable heads is generally required when
inspecting and scanning complex surfaces. Indexable probe heads, however,
tend to shrink CMM measuring volume.
A microprocessor control system is usually supplied with indexable heads to
operate as a power drive and intelligent interface between machine control
and indexing heads.
Rotary tables are especially useful when inspecting complex, multifaced
parts or workpieces with a rotation axis such as cams, gears, and rotors.
A variety of sizes are available to accommodate different size workpieces.
Rotary tables expand CMM measuring volume.
Rotary tables can be controlled manually or automatically. When
automatically controlled tables are used, special software programs
interact with the machine controls to control table movement and provide
misalignment compensation.
Besides their physical configurations, CMMs can also be classified
according to their mode of operation: manual, manual computer-assisted,
motorized computer-assisted, and direct computer controlled. Manual
machines have a free-floating, solid or electronic or non-contact probe
that the operator moves along the machine's coordinate axes to establish
each measurement. Digital readouts, associated with each axis, provide the
measurement values that the operator notes and records manually. In some
instances, a simple digital printout device may be used to record the
readings.
Manual computer-assisted CMMs use a data processing system to manipulate
the measurements which are still made by manually moving the probe through
a series of measurement locations. Solid or electronic or non-contact
probes may be used on this type of machine. The data processing may be
accomplished by a special microprocessor-based digital readout, a
programmable calculator, or a full-fledged computer.
Depending on the sophistication of the data processing system and
associated software, computer-assisted CMMs perform functions ranging from
simple inch to millimeter conversion to automatic three-dimensional
compensation for misalignment and a host of geometric and analytical
measuring tasks. Storing of predetermined program sequences and operator
prompting are also available to create part programs. The part program is
generated and stored in the computer, which determines the inspection
sequence and compares measured results with nominal values and tolerances
for automatic GO, NOT GO decision making.
In effect, the computer system can carry out all the calculations and
analyses required to arrive at dimensional and tolerance evaluations and
can lead the operator through a prescribed series of positioning and
measuring moves. Data recording is usually included with computer-assisted
CMMs.
A motorized computer-assisted CMM has all the features of a
computer-assisted CMM, but uses power-operated motions under the control
of the operator, who uses a joystick. Most motorized CMMs also provide
means for disengaging the power drive to permit manual manipulation of the
machine motions. Some machines use direct-current servomotors and
pneumatically operated friction clutches to reduce the effect of
collisions, and most permit drive disengagement for manual movement
Direct computer controlled (DCC) CMMs are equivalent to CNC machine tools.
A computer controls all the motions of a motorized CMM. In addition, the
computer also performs all the data processing functions of the most
sophisticated computer-assisted CMM. Both control and measuring cycles are
under program control. Most DCC machines offer various programming
options, including program storage and, in some instances, off-line
programming capability.
Beyond the microprocessor-based digital readouts, which were initially
developed to provide basic measurement data processing capabilities for
manual coordinate measuring machines, there is also a need to solve
sophisticated measuring problems involving three-dimensional geometry and
to provide more flexible general-purpose programming capabilities to solve
special measuring problems. Many CMM manufacturers offer a series of data
processing equipment for such purposes, including full DCC capability.
The key to the productivity of all forms of computer-assisted CMMs lies in
the sophistication and ease of use of the associated software. Software is
the most important element in any coordinate measuring system because its
power determines how many part features can be measured and its ease of
use determines the extent to which the machine is used.
The functional capabilities of CMM software depend on the number and type
of application programs available. Virtually all CMMs offer some means of
compensation for misalignment between the part reference system and the
machine coordinates by probing selected points. Some machines are limited
to alignment in one plane, while most machines provide full
three-dimensional alignment. Once the designated points have been taken,
the program calculates the misalignment and applies the appropriate
correction to all subsequent measurement readings.
Conversion between Cartesian, polar, and, in some instances, spherical
coordinate systems is also commonly handled. Most systems also calculate
the deviation of measurements from nominal dimensions of the part stored
in memory and flag out-of-tolerance conditions.
Geometric functions handled by the CMM software define geometric
elements--such as points, lines, planes, cylinders, spheres and
cones--from a series of point measurements and solve measurement problems
dealing with the interaction of such geometric elements. Such software can
determine, for example, the intersection of two circles established on the
basis of a selected number of measurements or it can establish the angle
of intersection of two surfaces.
Many software packages also provide a means for evaluating geometric
tolerance conditions by determining various types cf form and positional
relationships (such as flatness, straightness, circularity, parallelism,
or squareness) for single features and related groups of features
Best-fit programs can identify the location of a part finished to size
within a rough part from which it is to be made, to optimize the
machining-allowance distribution: maximum material condition (MMC)
programs evaluate features dimensioned according to MMC principles.
Other application programs include automatic part scanning for digitized
profiles and a variety of special programs to handle the inspection of
special shapes such as gears and cams. Statistical analysis software
available provides for graphic data display, including histograms.
In the simplest form of CMM, a single transducer mounted parallel to each
axis is able to determine the position of the probe tip relative to the
datum point, which may conveniently be the point at which the axes
intersect, or any other suitable location.
There are several possible sources of error if such a technique is
employed. Lack of straightness in movement and of orthogonality of the
axes are major sources of such errors. A further cause of error is the
angular rotation of the carriages about axes perpendicular to their
directions of movement. Such errors, often referred to as Abbe errors,
depend not only upon rotation, but also upon the lateral offset between
the probe tip and the transducer measuring in that dimension, and are
obviously variable with that offset. Other sources of error exist, such as
errors in the linear transducers themselves.
Many attempts have been made to compensate for error. For example, it is
known to introduce a deliberate and known error into the transducers by
various means. However, such corrections only apply for a given location
in the measuring volume. An alternative technique is to "calibrate" the
machine, measuring the errors existing at various points when the machine
is actually used. As may be imagined, such a calibration process can be
extremely lengthy, especially for a large machine and an enormous amount
of storage is necessary.
One prior method for determining axis misalignment is as follows:
(a) positioning a granite cube on the CMM table with a first side aligned
with the CMM X axis and then measuring the variation in the CMM-generated
Y coordinate as the CMM probe is moved over the first side, then adjusting
the cube position until no Y variation is produced.
(b) Move the CMM probe over a second side (perpendicular to the X axis) and
measure the variation in the CMM-generated X coordinate. The ratio of the
X coordinate variation to the Y coordinate variation is a measure of the
misalignment between the CMM X and Y axes.
(c) Measure Y and Z axis misalignment by repeating steps (a) and (b), using
appropriate sides of the granite block and substituting Y for X and Z for
Y in steps (a) and (b).
(d) Measure X and Z misalignment by repeating steps (a) and (b), using
another pair of sides and substituting Y for X and Z for Y.
In addition to being time-consuming, this granite square method is subject
to errors caused by imprecise positioning of the granite square on the CMM
table.
Another time-consuming method is used to measure axis scale errors and
involves the use of a laser and the following steps:
(a) A reflector for a laser in&interferometer is attached to the CMM in
place of the CMM probe.
(b) The Y and Z axes of the CMM are locked so that only movement along the
X axis is allowed.
(c) A laser interferometer is aligned so that its beam travels parallel to
the X axis and strikes the reflector.
(d) The reflector is then moved along the X axis of the CMM and
CMM-generated X axis readings and the interferometer readings are
obtained. From these readings the scale error in the CMM X axis can be
determined.
(e) Steps a-d are then repeated for the Y and Z axes.
Also known are CMM inspection procedures which involve the use of artifacts
such as the barbell and the Bryan Gauge. The data generated by the CMM
during these procedures is used on a pass-fail basis. In other words, if
use of the above artifacts indicates that adjustment is required, then the
previously described granite block or laser interferometer procedures are
used in making the needed adjustments.
Another method used to calibrate a CMM includes the steps of installing a
CMM artifact on a CMM table, coupling the CMM probe to the artifact and
storing a plurality of CMM-generated cartesian coordinate data points for
a plurality of positions defined by the artifact. A data processor is
programmed to generate a set of distance equations in terms of the CMM
generated cartesian coordinates, a known diameter of the artifact and a
plurality of unknown CMM axis alignment error and scale error factors.
This set of equations is then solved for the unknown error factors, from
which can be determined the necessary CMM adjustments. The CMM can then be
properly aligned by making the indicated adjustments. In one version of
this procedure, the artifact may be a ball bar. In another version, the
artifact may be a Bryan Gauge. An example of this method is found in U.S.
Pat. No. 4,437,151 issued Mar. 13, 1984 and entitled Coordinate Measuring
Machine Inspection and Adjustment Method. Another prior art approach for
applying software error compensation to a coordinate measuring machine is
disclosed in the article entitled: "Error Compensation of Coordinate
Measuring Machines" dated January, 1985 and published in the Annals of the
CIRP by G. Zhang et al. The Zhang system is based on a rigid body model of
workpiece motion in the machine coordinate frame. By taking a relatively
small set of data for each axis, errors are computed throughout the full
workzone. Squareness data is determined using linear displacement
measurements along the machine diagonals. The error compensation
computation is incorporated into the machine position reading subroutines
to automatically produce compensated readings.
SUMMARY OF THE INVENTION
One advantage of the present invention is to provide an improved method and
system for calibrating a CMM and the like in a reliable, accurate and
cost-efficient manner in a manufacturing environment.
The CMM calibrating method of the present invention includes the steps of
correlating feedback data and calibration data to obtain axis correction
data for each predetermined position along the entire axis of motion. The
axis correction data is then stored in a form which can be utilized by a
control means during operation of the CMM to compensate the CMM for the
entire measuring volume of the CMM.
The CMM calibrating system of the present invention includes data
processing means, including means for correlating feedback data and
calibration data to obtain axis correction data for each of the
predetermined position along the axes of motion of the CMM. Storage means
stores the access correction data in a form which can be utilized by a
control means during operation of the CMM to compensate the CMM for the
entire measuring volume of the CMM.
Preferably the CMM includes a probe shaft having three degrees of freedom
representing axes of motion of the CMM. Also, preferably, all geometry
errors (i.e. 21) of the CMM are corrected electronically prior to actual
use of the CMM, so that the corrections can be performed in real time. As
a result, the need to manufacture and assemble a totally accurate machine
is eliminated.
Other advantages of the method and system include the reduction in time in
final assembly of the CMM and allowing increased manufacturing tolerances
for many of the components of the CMM.
The above advantages and other features of the present invention are
readily apparent from the following detailed description when taken in
connection with the accompanying drawings.
DETAILED DESCRIPTION OF THE INVENTION
FIG. 1 is a perspective view of a typical CMM with which the present
invention is utilized;
FIG. 2 is a representation of the three-dimensional coordinate system of
the CMM of FIG. 1;
FIG. 3 is a block diagram of the system for automatically calibrating the
CMM;
FIGS. 4A through 4C is a flowchart illustrating the various steps taken by
the system of FIG. 4 to calibrate the CMM;
FIGS. 5A through 5C illustrate two different positions of an artifact, such
as a ball bar, in each of the coordinate planes of the CMM;
FIG. 6 is a flowchart illustrating the various operating steps taken to
determine the squareness of the CMM;
FIGS. 7A and 7B illustrate the three-dimensional coordinate system and
measuring volume of the CMM of FIG. 1, together with a probe subcoordinate
system;
FIG. 8 is a block diagram illustrating the method and system for
determining position within the measuring volume of the CMM; and
FIG. 9 is a set of equations to be solved to determine the corrected
position within the measuring volume for a machine described in FIG. 1
whose axis coordinates are defined in FIGS. 7A and 7B; similar equations
could be developed for other types of CMM configurations.
Referring to FIG. 1, there is illustrated a coordinate measuring machine
(CMM) collectively indicated by reference numeral 10. The CMM generally
includes an X-Y-Z positioning device, generally indicated at 12; a probe
14; and a control system, generally indicated at 16. The control system 16
includes a machine control, generally indicated at 18; computer hardware,
generally indicated at 20; and software for programming the computer
hardware 20.
The probe 14 is inserted into a Z-axis probe arm or shaft 22 of the device
12. The device 12 includes a base or work table 26 which contains tapped
holes to facilitate the clamping and locating of parts.
The device 12 also includes a backrail 28 which slidably supports an X-axis
carriage 30 by preloaded air bearings, which also guide the carriage 30.
An overhead Y-axis carriage 32 moves relative to the X-axis carriage 30 and
is also supported and guided by preloaded air bearings on the X-axis. The
probe shaft 22, in turn, moves relative to the Y-axis carriage 32 and is
supported and guided thereon by preloaded air bearings.
While a conventional cantilever CMM has been described, it is to be
understood that other types of CMMs may be utilized with the present
invention based on a set of equations defined for the particular CMM being
corrected.
Referring to FIG. 2, the X-Y-Z coordinate system depicted therein
illustrates typical errors caused by angular rotation of the carriages 30
and 32 and the shaft 22 about the axes cf the three axis system. Three
such errors exist for each of the axes. Consequently, nine such errors
exist for the coordinate system of FIG. 2 even though the mechanical
components of the device 12 are manufactured and assembled in a highly
accurate fashion. Because of these rotation errors and other errors,
position error still exists as the device 12 is moved to different spots
within its measuring volume 102 as shown in FIG. 7. Angular rotation
errors about each of the axes are defined as follows: A(x)=X-roll;
A(y)=Y-pitch; A(z)=Z-pitch; B(x)=X-pitch; B(y)=Y-roll; B(z)=Z-yaw;
C(x)=X-yaw; C(y)=Y-yaw; and C(z)=Z-roll.
Lack of straightness in movement along the axes of FIG. 2 is also a source
of error. For each axis there typically exists a lack of straightness with
respect to the other two axes, thereby resulting in six errors with
respect to straightness of the device 12. Straightness errors are defined
as follows: X(y)=Y-straightness in X direction; X(z)=Z-straightness in X
direction; Y(x)=X-straightness in Y direction; Y(z)=Z-straightness in Y
direction; Z(x)=X-straightness in Z direction; and Z(y)=Y-straightness in
Z direction.
Although oftentimes a less serious source of error, axis scale errors can
become significant. Such errors are defined as follows: X(x)=scale errors
in X; Y(y)=scale errors in Y; and Z(z)=scale errors in Z.
Another possible source of error is lack of orthogonality of the X,Y, and Z
axes. Such error is typically given as the angular deviation from 90
degrees as follows: P.sub.yx =y-x squareness; P.sub.zx =z-x squareness;
and P.sub.zy =z-y squareness.
Consequently, it can be seen that there are 21 different geometry errors of
the device 12, all of which, except for the squareness errors vary
depending on the position of the device 12 within its measuring volume
102.
Finally, another possible source of errors are probe offset errors, which
are determined from the angular errors and probe offsets, S.sub.x, S.sub.y
and S.sub.z of FIG. 7B. The probe 14 is used to measure the center
position of a sphere 100. Probe offsets are then determined by vector
subtraction of vectors A and B of FIG. 7A.
Referring now to FIG. 3, there is illustrated in block diagram form a
system for calibrating the 10. The system 10 includes calibration
equipment 32, such as a laser interferometer with linear, angle and
straightness optics. Preferably, the laser comprises an HP 5528A laser.
The calibration equipment 32 also includes electronic levels with a level
meter and an A to D HPIB interface The calibration equipment 32 further
includes a ball bar as well as a vertical straight edge and probe.
Preferably, the electronic levels comprise Wyler electronic levels and the
probe comprises a Renishaw TP-2, PH-6 probe, including a 200 millimeter
extension. Finally, the calibration equipment 32 includes probe adaptors
for the laser optics, the level and the PH-6 probe.
The calibration equipment 32, in general, is used by factory personnel to
measure | | |
