Automatic alignment system for coordinates measuring machines.
Stanciu, Valerian ; Dan, Stefan ; Stanciu, Mariana 等
1. INTRODUCTION
The measurement systems for dimensional measurements that are on
the market nowadays rely on the experience and the precision of the
manual setup of the operator. A complex 3D measurement conducted with a
Coordinate Measuring Machine is subject the errors inducted by
misalignments during the manual alignment of the part in the working
volume of the CMM can lead to false measurements and, consequently, to
misjudgments of the production process involving unnecessary costs.
These kinds of misalignments are often done by inexperienced operators
and the training costs of the personnel for operating/ programming the
CMM's are high. For mass production, the vendors offer dedicated
solution for each production line, which can run semiautomatic in
conjunction with unique fixtures for each type of product. The aim of
this paper is to present an automatic alternative alignment system to
the actual CMM architecture that assumes manual alignment. Using simple
laser system built on the portal of the CMM, that can measure the
distance from the emitter to any solid obstacle in the working volume
and return this data as 3D point, the machine can execute at the
beginning of the inspection sequence an automatic alignment routine.
2. ACTUAL CMM STRUCTURE
The actual commercialized CMM's are based on two major methods
of point acquisition.
First class, based on the direct contact of the touch-probe with
the part to be measured is the most used solution in steel and plastic
industry and basically in each application for solids with a reasonable
elasticity. The principle is to transform the deflection on the solid in
reading triggers of the optical rulers and of the normal vector of the
deflecting surface into vector components on each axis depending on the
coordinates system (Pettersson, 2004). The alignment sequence is
manually performed by the operator by indicating with successive
deflections (points) the areas that give the coordinate system of the
part: plan--three points; line--two points on a normal direction to the
defined plan's vector; one point with the normal vector
perpendicular to the plan created by the two vectors defined before
(initial plan and the line). The errors of the manual alignment are
propagated throughout the entire measurement with very low possibility
of detection. The main producers of contact CMMs are Brown & Sharpe
Inc., Carl Zeiss and Mitutoyo.
The second class of coordinate measuring machines is so-called
non-contact CMM. Using an optical projector and a CCD camera to acquire
information regarding dimensions and position of the part to be
inspected, these measuring devices are the top of the arrow of the
dimensional measurements industry.
There are several methods to identify the position of a spatial
point in the working volume without mechanical contact.
The first is to send a laser beam from the projector (optical spot)
and with a CCD camera to read, based on the triangulation principle the
spatial coordinates of the measured point. This technology is used by
the Carl Zeiss CMM with a VEST non-contact probe.
The second solution (optical scanning) provided by Perceptron (US)
is to project a slit lighting beam onto the surface to be measured and
the beam image is recorded to calculate the x, y, z coordinates from the
beam central line on the surface.
Probably the most revolutionary non contact measuring solution is
offered by OPTON using the Moire algorithm. A grating pattern is
projected on the surface to be measured. The projection lighting is
provided by the light source of the projector (OPTON, 2008). The
original grating pattern is deformed along the curved surface. (Figure
1) Then this deformed grating pattern is taken into a computer via a CCD
camera and saved as a digital image.
Regardless the sensor system and the algorithm chosen for
calculation a manual alignment sequence is need it to indicate the CMM
exactly where is the part to be inspected in the entire volume of the
machine. Another inconvenient of the noncontact CMM is related to the
material of the part. The transparent or translucent surfaces like glass
or plastic, the mirror surfaces--high finished metal surfaces, cast iron
surface in black color or even opaque materials that have a color
similar to the grating pattern can not provide reliable measurements
data.
[FIGURE 1 OMITTED]
3. SELF ALIGNMENT CMM ARCHITECTURE
The system proposed to eliminate manual alignments that could lead
to false measurements is simple and can be used by both types of
coordinate measuring machines without major changes of the architecture.
A bridge/ portal CMM consist of three major elements: a solid table
(preferable made of granite), a light metal structure in shape of a
portal that travels along the length of the table and a beam that slides
on the portal left-right, up and down (Fan et al., 1998).
Laser pointers that are able to measure the distance between the
emitter point and the focus point are common tools used in construction
or any other application that require straight measurements. The costs
of this kind of laser pointers is relatively low and for short distance
(up to 3 meters) the precision is acceptable (about 2-3 mm).
In the figure 2 is presented a CMM with set of laser pointers built
in each foot of the portal. The pre-hit (safety distance between surface
of the part and the point where the machine changes from traveling speed
to measurement speed) is normally defined at 5 mm and consequently a
precision of 5 mm is enough for a rough estimation of the part position
on the table, before the automatic alignment sequence is performed. The
sensors are evenly distributed at each 5 mm, vertically on the foot
detecting the presence of any object by measuring the distance to the
object. If there is no object, the sensors measure the distance between
the feet. To increase the vertical resolution from 5 to 2.5 mm the sets
of sensors can be shifted with 2.5 mm as is shown in the picture 2.
The pre-alignment sequence (identification of an object position in
the working volume) is a simple algorithm that is performed at once by
traveling the portal along the entire Xaxis of the CMM. The position of
the object on the X-axis is given by memorizing the coordinates from the
optical ruler when the sensors detect first and last an obstacle. To
avoid overcharging of the lasers, overheating of the part or of the
environment the lasers are inactive by default and trigger every 5 mm a
reading; at a speed of 5mm/s this means every second a measurement and
for a standard table with 1500 mm length the entire pre-alignment
sequence is finished in about 5 minutes.
The position and the boundaries of the part on the Y-axis are
determined directly from the measurements of the lasers and the height
of the part (Z-axis) is given by the sensors vertical pattern. Each time
the sensors trigger a reading two sets of points are created in the
computer. These points can be void if the sensors measured the standard
distance between the portal feet or data points containing (x,y,z)
coordinates of the surface detected.
[FIGURE 2 OMITTED]
[FIGURE 3 OMITTED]
[FIGURE 4 OMITTED]
After a complete pre-alignment cycle a cloud of points determining
the boundaries surfaces of the object is stored in the computer with a
minimum resolution of 5 mm on each direction--figure 3.
A shape recognition algorithm (Hofer et al., 2005) overlaps the
points cloud on the 3-D model and determines automatically the set of
rotations and translations between a local coordinate system and an
absolute system--figure 4.
The machine is ready in less then 7 minutes to run the standard
automatic alignment.
4. CONCLUSION
Automatic alignment feature gives the opportunity to eliminate
manual operation on a CMM, this way the possible false readings of the
operator are eliminated as well. Additionally, the supervision of the
CMM can be done from distance (e.g. operating room) given the
possibility to perform measurements in unfriendly environments.
Combining the automatic alignment with features like fuzzy path planning
algorithm (Fiorentini et al., 1992), that calculates the optimum path
between given inspection areas/ points and with automatic loading/
unloading conveyor systems several production lines can be served only
by one supervisor. The mechanics of the CMM need no change to run the
automatic alignment. Minimum electronics must be implemented (laser
system) supported by an intelligent software, transforming the CMM into
an semi-independent machine.
5. REFERENCES
Fan, K. C.; Chen, M. J.; Huang, W. M. (1998). A
Six-degree-of-freedom Measuring System for the Motion Accuracy of Linear
Stages, Inst. J. Mach. Tools and Manufact., Vol. 38, No. 3,
Fiorentini, F.; Moroni, G.; Palezzato, P. & Semerano, Q.
(1992). Probe path generation for an automatic inspection system, 8th
International Conference of CAPE
Hofer, M.; Odehnal, B.; Pottmann, H.; Steiner, T. & Wallner, J.
(2005). 3D shape recognition and reconstruction based on line element
geometry, Institute of Discrete Mathematics and Geometry, TU Wien
OPTON, OPTON's non contact 3D Moire measurement machine
Available from: http://www.opton.co.jp Accessed: 2008-07-15
Pettersson U. (2004). Identification and Adaptive Control of a
Coordinate Measuring Machine, Linkoping
