Note: Descriptions are shown in the official language in which they were submitted.
CA 02717291 2010-04-30
T811 {
-- 1 -
[Name of Document] Specification
[Title of the Invention] NUMERICALLY CONTROLLED
MACHINE TOOL AND NUMERICAL CONTROL DEVICE
[Field of the Invention]
[0001)
The present invention relates to a numerically
controlled machine tool which has a linear feed axis and
a rotational feed axis and in which a main spindle and a
table are movable relative to each other.
(Background Art]
[0002]
In a machine tool which has a linear feed axis and a
rotational feed axis, it is generally difficult to
position a tool at a desired position since an error may
occur when the feed axes are moved in accordance with a
movement command. Therefore, when highly-precise
processing is desired, correction has to be carried out
depending on the machine error. In order to carry out
the correction, the machine error needs to be measured
accurately as a preliminary step of correction. The
following documents disclose the prior art of measurement
and correction of errors.
[0003]
In Japanese Patent Publication of Examined Patent
{ 25 Application (Kokoku Publication) No. H06--088192 (JP-H06-
088192 B1), a method is disclosed in which deviations of
two rotational feed axes (deviation of the position of
the center of axes) of a machine tool which has two
rotational feed axes (A, B) that are orthogonal to each
other are measured in advance,'and coordinates of the two
rotational feed axes are determined by taking these
deviations into account.
[0004]
In Japanese Patent Publication of Unexamined Patent
Application (Kokai Publication) No. 2004-272887 (JP-2004-
272887-Al), a method is disclosed in which, in a machine
tool which has three linear feed axes (X, Y, Z) that are
CA 02717291 2010-04-30
- 2 -
orthogonal to each other and two rotational feed axes (A,
C) that are orthogonal to each other, the position to
which the machine has to be actually moved is determined
based on deviations of the center of a rotating shaft or
the center of turning of a spindle, and by using drive
control means, the linear feed axes and the rotational
feed axes are moved to the determined position so as to
= correct the position of the tool tip.
[0005]
In Japanese Patent Publication of Unexamined Patent
Application (Kokai Publication) No. H09-2377.12 (JP-09-
237112-Al), a method is disclosed in which an error of a
tool unit of a machine tool having a parallel link
mechanism is corrected based on an error map. The error
map has error data corresponding to each lattice point
that have been calculated by computation from the
difference between the command value of a position and a it
posture for tool tip of the tool unit and the measured
value of them.
[0006]
in the pamphlet of WO 2004/034164, a system and a
process for measuring, correcting and testing numerically
controlled machine tool heads and/or tables are disclosed
which are automated and integrated in a numerically
~. 25 controlled system. This system comprises at least one
support base that is equipped with a plurality of
distance sensors, and at least one device of gauge tool
type that is equipped with connection means for the heads
at one of its ends and with a ball at another opposite
end. The ball is placed next to the distance sensors, so
that the distance sensors are able, at any time and in
any position, to move to any position in order to measure
the distance that separates them from the ball, and to
thereby determine the position in the Cartesian space.
(0007]
In the correction methods disclosed in JP-H06-088192
B1 and JP-2004-272887 Al, deviation of the rotational
CA 02717291 2010-04-30
- 3
axis is corrected. Thus, there is a problem that errors
or the like which vary depending on the undulation of the
rotational axis itself or on the position of the linear
feed axis cannot be corrected. The error map disclosed
in JP-H09-237112 Al is the error of the tip of a tool
unit that is driven by a parallel link mechanism obtained
as table data.. Thus, there is a problem that this method
is not applicable to a machine tool which has a linear
feed axis and a rotational feed axis. In the measurement
method disclosed in the pamphlet of WO 2004/034164, only
the center position of a reference ball is measured, so
that there is a problem that deviation of the position of
the tool tip produced due to an error of the posture of
the main spindle relative to the table due to variation
of the tool length or projecting length of the tool
cannot be corrected.
[Disclosure of the Invention]
(0008]
The present invention aims to resolve above-
described problems associated with prior art, and
therefore, it is an object of the present invention to
provide a numerically controlled machine tool which has a
linear feed axis and a rotational feed axis, and a {
numerically controlled machine tool having error map
{ 25 preparation function.
[0009]
In accordance with the present invention, there is
provided a numerically controlled machine tool which has
a linear feed axis and a rotational feed axis and which
has functions of measuring in advance errors of a
relative position and a relative attitude of a main
spindle relative to a work table by positioning said
linear feed axis and said rotational feed axis to
specified position and attitude, and of correcting a
movement command based on measured error data, said
error data being multi-dimensional data containing a
position error and an attitude error, said machine tool
CA 02717291 2010-04-30
4 -
comprising: a error data storage means which store an
error map that is prepared by collecting a plurality of
said error data corresponding to positions and rotation
angles of said linear feed axis and said rotational feed
axis, and a correction data computing means which compute
correction data for correcting said movement command from
a command position for said linear feed axis and said
rotational feed axis and said error data stored in said
error data storage means.
[0010]
Also, in accordance with the present invention,
there is provided the numerically controlled machine
tool, wherein said position error is an error of relative
position of said main spindle relative to said work
table, an error being expressed as 3-dimensional
coordinate values when said rotational feed axis is
positioned to specified rotation angle, and wherein said
attitude error is an error of relative attitude of said
main spindle relative to said work table, an error being
expressed as inclination angle when a first rotational
feed axis and a second rotational feed axis that are
orthogonal to each other are positioned at specified
rotation angles.
[0011]
Also, in accordance with the present invention,
there is provided the numerically controlled machine
tool, wherein a measurement spacing of said rotation
angles in said first rotational feed axis and said second
rotational feed axis is set such that a difference of
said attitude error between adjoining measurement points
is a specified value
[0012]
Also, in accordance with the present invention,
there is provided the numerically controlled machine
tool, wherein, when a tool tip position of a rotary tool
mounted to said main spindle is to be positioned in
accordance with said command position, said correction
CA 02717291 2010-04-30
- 5 -
data computing means compute an amount of a deviation of
3-dimensional coordinate values of said tool tip position
from said attitude error and a dimension of said rotary
tool, and thereby correct said movement command for said
linear feed axis
[0013]
Also, in accordance with the present invention,
there is provided a numerical control device which
controls a machine tool having a linear feed axis and a
rotational feed axis and which has a function of
correcting movement command for feed axes based on error
data that have been measured in advance, said device
comprising: a error data storage means that store an
error map prepared by collecting a plurality of said
error data corresponding to a position and rotation angle
of said feed axes, said error data being data containing
a position error and an attitude error, and a correction
data computing means which compute correction data for
correcting said movement command from a command position
for said feed axes and said error data stored in said
error data storage means.
[0014]
In accordance with the numerically controlled
machine tool and the numerical control device, an error
map can be prepared by measuring the position error and
the attitude error of the numerically controlled machine
tool which has a linear feed axis and a rotational feed
axis. In an error map prepared in accordance with the
present invention, error data for the position error and
the attitude error that change as the feed axes are moved
are stored separately, and the position command is
corrected based on the error data. Therefore, in
accordance with the present invention, even when the tool
length or the tool projecting length varies, the tool tip
or the processing point of the tool can be positioned to
the target position with high precision. In case where
measurement points are set such that the coordinate
CA 02717291 2010-04-30
- 6 -
position of the linear feed axis is the same in adjoining
measurement regions, influence of mounting error of the
measurement device can be eliminated. In case where
separation between adjoining measurement points is set
such that the difference of error is constant everywhere,
the amount of data of the error map can be reduced while
maintaining the precision of correction. In case where
an error map is prepared by measuring the processed test
piece or work piece, errors produced due to deflection of
the main spindle or the tool due to rotation of the main
spindle, flexure of the machine or the tool due to load,
or the bike, can be corrected.
[0015]
In the present invention, the term "command
position" refers to the position of the destination point
of the feed axis commanded by a processing program, and
the term "position command" refers, from among the
command pulses issued from the interpolation section to
the servo section based on command position, command
velocity, and the like, to the command for controlling
the position of the feed axis.
[Brief Description of Drawings]
[0016]
Above and other objects, features and advantages of
the present invention will become more apparent from the
following description of preferred embodiments with
reference to appended drawings, in which:
Fig. 1 is a side view showing a numerically
controlled machine tool according to the present
invention;
Fig_ 2 is a block diagram showing a numerical
control device of the numerically controlled machine tool
according to an embodiment of the present invention;
Fig. 3 is a view of lattice points in 3-dimensional
Cartesian coordinate space;
Fig_ 4 is a view of 2-dimensional data sheet (map
data) associated with each lattice point of Fig. 3;
CA 02717291 2010-04-30
- 7 -
Fig. 5 is a view of a reference ball attached to the
tip of a tool as it is measured with a measurement device
mounted on a palette;
Fig. 6 is a view showing the measurement range of
reference balls having support shafts of different length
as seen from the direction of Y-axis;
Fig. 7 is a view of a method for determining plural
measurement regions;
Fig. 8 is a flow chart of a first measurement method
for measuring a position error and an attitude error;
Fig. 9 is a detailed flow chart showing M3 of the
flow chart of Fig. 8;
Fig. 10 is a view of the attitude error expressed in
two variables;
Fig. 11 is a view showing an example of a spindle
rotation type machine having a reference ball attached to
the palette side and a measurement device mounted on the
main spindle side;
Fig. 12 is a view showing an example of a table
rotation type machine having a measurement device mounted
on the table and a reference ball attached to the main
spindle;
Fig. 13 is a flow chart of a second measurement
method for measuring a position error and an attitude
error;
Fig. 14 is a view of a state in which each surface
is processed only with operation of the linear feed axes;
Fig. 15 is a developed view of five surfaces of a
rectangular parallelopiped showing the area to be
processed for each indexed angle;
Fig. 16 is a view of a state in which a work piece
is processed into a lattice-like face at the indexed
angle for the rotational feed axes B, C;
Fig. 17 is a view of a state in which each
measurement surface indexed at a specified angle is being
measured;
Fig_ 18 is a view of a method for determining an
CA 02717291 2010-04-30
8 -
intersection point of three planes; and
Fig. 19 is a flow chart showing an example of
correction method using an error map.
[Best Form for implementing the Invention)
[0017]
The present invention will be described below with
reference to appended drawings showing preferred
embodiments thereof. A numerically controlled machine
tool according to the present invention comprises a
numerical control device for controlling an operation of
a machine in accordance with a processing program. In
Fig. 1, the construction of a five-axis horizontal type
machining center having two rotational feed axes on the
main spindle side, is shown. Referring to Fig. 1,
machining center 1 comprises bed 2 provided on the floor,
column 3 that is erected on bed 2 and is capable of being
moved linearly in a direction of Z-axis, and main spindle
stock 5 that is capable of being moved linearly in a
direction of Y-axis which is perpendicular to the
direction of column 3. On main spindle stock 5, a
bracket 5a is supported rotatably about C-axis which is
parallel to Z-axis. On bracket 5a, main spindle head 4
is supported rotatably about A-axis which is parallel to
X-axis. On main spindle head 4, a main spindle for
holding a tool is rotatably supported.
[0018]
Machining center 1 is erected on bed 2 at a position
opposed to main spindle head 4, and comprises table 6
that is capable of being moved linearly in a direction of
X-axis which is perpendicular to the paper surface. work
piece 7 is held via angle plate (equerre) 8 on table 6.
[0019]
Fig_ 2 is a block diagram showing the construction
of numerical control device 20 that controls the position
of feed axes of the machine tool.
[0020]
Numerical control device 20 shown in Fig. 2 has the
CA 02717291 2010-04-30
- 9 -
function of correcting position error and attitude error
of the machine tool, and comprises reading and
interpreting section 22 that reads and interprets
processing program 21 to compute a command velocity and a
command position for each feed axis, interpolating
section 23 that computes command pulses based on the
command position, command velocity, and the like to
perform linear interpolation or circular interpolation of
feed for each feed axis, position command recognition
means 24 that obtains the command pulses and recognizes a
position command for each feed axis, a computation
section that computes a position error and an attitude
error for a measurement point based on measurement data
measured with measurement device 50 and the coordinates
of the measurement point, error data storage means 25
that store the position error and the attitude error
computed by the computation section in correspondence to
the position of the linear feed axis and the rotation
angle of the rotational feed axis, correction data
computing section 26 that computes correction data for
correcting the position command from the position command
and the error data stored in error data storage means 25,
correction pulse computation means 27 that determine a
correction pulse for correcting the position command
~. 25 based on the correction data, and addition means 28 that
output the pulse obtained by adding the correction pulse
to the command pulse to servo section 29.
[0021]
Motor 30 for each feed axis is driven by the drive
current which is amplified by servo section 29 to thereby
move the feed axis. Servo section 29 controls based on
the velocity feedback from motor 30 and the position
feedback from an unshown position detecting device such
that each feed axis is moved at desired velocity to
desired position.
[0022]
The present invention also includes a device that is
CA 02717291 2010-04-30
- 10 -
constructed such that the command position is obtained
from reading and interpreting section 22 for correction,
and the corrected command position is inputted into the
interpolation section for the motor to move the machine
to desired position.
[0023]
Next, a method for preparing an error map will be
described. In an error map, as shown in Fig. 3, lattice
points 31 are set at desired positions in a direction of
each axis of linear feed axes X, Y, Z in Cartesian
coordinate system, and 2-dimensional array data 33
corresponding to the rotation angle of the rotational
feed axes are associated with respective lattice points
as shown in Fig. 4. Thus, an error map is composed as 5-
dimensional array data of X, Y, Z, A, C.
[0024]
An error map is composed of plural error data 34
obtained by measurement with each feed axis positioned at
desired measurement points. Error data 34 are composed
of position error 34a and attitude error 34b.
[0025]
Position error 34a refers to an error in relative
position of the main spindle relative to the table, that
is, an error of the position expressed as 3-dimensional
coordinate value (x, y, z) produced when the feed axes
are positioned at specified positions or rotation angles.
That is, a position error is the difference between the
theoretical position commanded by the position command
and the actual position.
[0026]
Attitude error 34b refers to an error in relative
attitude of the main spindle relative to the table, that
is, an error expressed as inclination angles produced
when the feed axes are positioned at specified positions
or rotation angles. Thus, an attitude error is the
difference between the theoretical inclination commanded
by the position command and the actual inclination.
CA 02717291 2010-04-30
- 11 -
[0027]
Measurement spacing of error data 34 is set such
that the difference of position error-34a or attitude
error 34b between adjoining measurement points becomes
equal to a specified value. In other words, if the
difference of error between adjoining measurement points
is small, the measurement spacing is broadened, and if
the difference of error is large, the measurement spacing
is narrowed. By broadening the measurement spacing in
the portion of small difference of error, amount of data
can be reduced and load on memory can be thereby reduced.
By narrowing the measurement spacing in the portion of
large difference of error, precision of the correction of
error can be kept high.
[0028]
Next, an example of measurement method for measuring
position error 34a and attitude error 34b of a machine
tool which has rotational feed axes A, C on the main
spindle side, will be described. As shown in Figs. 5 and
6, measurement device 50 is mounted via support shaft 40
to the main spindle of the spindle rotation type machine
tool, and comprises reference ball 52 which has known
values of outer dimension and of distance L1, L2 from the
control point to ball center P1, p2, and sensor bracket
53 which is mounted on palette 54 fixed to the table and
has non--contact type sensors 55 in X-direction, Y-
direction, Z-direction. The non-contact type sensor 55
can measure the distance to the reference ball 52 in each
direction in non-contacting manner. The sensor of the
present invention includes not only non-contact type
sensor but also contact type sensor.
[0029]
Measurement is performed by dividing the measurement
range of rotational feed axis A, C in even pitch or in
uneven pitch, and operating, at the same time, the linear
feed axis so as to maintain the center position of
reference ball 52 at each division point (measurement
CA 02717291 2010-04-30
-- 12 -
point). As used herein, the term "even pitch" means that
a measurement point is defined at every specified angle
so that an angular separation between adjoining
measurement points is everywhere equal, and the term
"uneven pitch" means that error data are obtained only at
points where error exceeds a certain defined value so
that an angular separation between adjoining measurement
points is generally unequal.
[0030]
As shown in Fig. 9, center position P1 of reference
ball 52 is measured in directions X, Y, Z that are
orthogonal to each other using measurement device 50
equipped with non-contact type sensors 55. In order to
determine actual relative attitude and actual control
point, the reference ball having support shaft 40 of
different length is attached as shown in Fig. 6, and
center position P2 of reference ball 52 is measured
again. By attaching support shafts 41a, 41b of different
lengths for measurement, relative attitude of the main
spindle and the table can be determined.
[0031]
The present invention includes the use of a support
shaft whose length can be adjusted. In the present
embodiment, the control point is set at the intersection
t 25 point of the rotation center of first rotational feed
axis C ,and the rotation center of second rotational feed
axis A. As used herein, the term "relative attitude"
refers to the relative inclination of the main spindle
and the table.
[0032]
Sensor bracket 53 of measurement device 50 is
mounted rotatably about an axis that is parallel to Z-
axis. Thus, when measurement is to be performed over
entire 360 degrees, the sensor bracket can be rotated
about the axis that is parallel to Z-axis by 90 degrees
four times to perform measurement.
100331
CA 02717291 2010-04-30
-- 13 -
As shown in rigs. 7 and 8, when the region to be
measured is large, measurement can be performed by
dividing the region into plural measurement regions. In
this case, in first measurement region 70e, which is to
be used as a reference, operating ranges of linear feed
axes X, Y, Z are measured with a laser gauge, an
indicator, or the like, and adjustment is performed to
obtain adequate precision as required. The present
invention includes a case where adjustment of the
precision of the operating ranges of linear feed axes X,
Y, Z in first measurement region 70a is not performed,
and error is computed taking into account the result of
the measurement. This is intended to restrict the result
of measurement in first measurement region 70a to the
error produced when rotational feed axes A, C are
rotated.
[0034]
The measurement points in measurement regions 70a,
70b are defined such that there exists at least one
measurement points 71 having the same coordinate value of
the linear feed axes as the measurement points in the
adjoining measurement region. This is because an error
in mounting measurement device 50 should not influence
the result of measurement between first measurement
region 70a and other measurement region 70b-
[0035]
The error in mounting measurement device 50 can be
determined by subtracting the error due to difference of
the rotation angle of the rotational feed axes from the
difference of the result of measurement between
measurement points having same coordinate value of the
linear feed axes. By subtracting this error in mounting
from the result of measurement in each measurement
region, same result of measurement can be obtained as if
all the measurement regions are measured in one proper
arrangement.
[0036]
CA 02717291 2010-04-30
- 14
Next, a method for computing the position error and
the attitude error will be described- First, the
attitude error is determined as follows. Commanded
relative inclination of the main spindle relative to the
table is determined from the command value of the
rotation angles of rotational feed axes A, C. Here, the
angle formed by the rotation axis of the main spindle and
the normal (perpendicular line) to the work piece
attaching surface of the angle plate (equerre) is taken
as the relative attitude of the main spindle and the
table. From center positions P1, P2 of reference ball 52
at two measured points, a line is determined that passes
through points P1, P2, and the angle formed by this line
and the normal to the work piece attaching surface of the
angle plate (equerre) is taken as the actual relative
inclination of the main spindle relative to the table.
Difference between the commanded relative inclination of
the main spindle relative to the table and the actual
relative inclination of the main spindle relative to the
table is determined as the attitude error. The attitude
error is expressed by the angle difference i relative to
Z-axis as seen from X-axis direction, the angle
difference j relative to Z-axis as seen from Y-axis
direction, and the angle difference k relative to Y-axis
as seen from Z-axis direction. The present invention
includes a case where, as shown in Fig. 10, the attitude
error is expressed by two angles I, J-
[00371
Next, the position error is determined as follows.
In the present embodiment, the control point is set at
the intersection point of the rotation center of first
rotational feed axis C and the rotation center of second
rotational feed axis A. Thus, to whatever rotation angle
the rotational feed axes are moved, the theoretical
position of the control point does not change. The
commanded position of the control point is determined
from the command values of linear feed axes X, Y, Z. The
CA 02717291 2010-04-30
- 15
position of the control point refers to the relative
position of the reference point of the table relative to
the control point of the main spindle. On the line
passing through the points P1 and P2 that is determined
in the above-described step of determining the attitude
error, the position of a point at a distance L2 in the
direction from P2 to P1 is determined and is taken as the
actual position of the control point. A vector from the
position of the commanded control point to the position
of the actual control point is determined and is taken as
the position error. The position error vector can be
split into components in the direction of X-axis, X-axis,
Z-axis, and is expressed in the form of (x, y, z). The
present invention includes a case where the position
error vector is expressed in other form.
[0038]
Fig. 11 is a view showing an embodiment in which, in
a spindle rotation type machine, reference ball 52 is
attached to the side of palette 54 and displacement
detection probe 58 is mounted on the spindle side.
Displacement detection probe 58 is constructed such that
it is displaced in a direction of the normal at the
measurement point of the object to be measured, and the
amount of the displacement is detected.
[0039]
Fig. 12 is a view showing an embodiment in which the
present invention is applied to a table rotation type
machine having rotational feed axes B, C on the table
side. In the embodiments shown in Figs. 11 and 12, the
error of the feed axes can be measured based on the same
principle as in the embodiment shown in Fig_ 5.
[0040]
Next, an example of measurement method for measuring
position error 34a and attitude error 34b of a machine
tool having rotational feed axes B, C on the table side
will be described. Fig. 13 is a flow chart showing this
measurement method. In this measurement method, without
CA 02717291 2010-04-30
- 16 -
using any special measurement device, a test piece or a
work piece is processed on the machine, and the processed
test piece or work piece is measured with a touch probe
mounted on the main spindle to thereby determine the
position error and the attitude error. A cubic test
piece is used in the present embodiment.
(0041]
As shown in Fig. 13, rotational feed axes B, C are
indexed to sufficiently small rotation angle (in this
case, 0 degree for axis B, 0 degree for axis C) as
required for sufficient precision of the position error
and the attitude error, and as shown in Fig. 14, each
surface of test piece 60 (frame-like reference processing
surface 61) having X, Y, Z-axis directions as the normal
is processed without operating the rotational feed axes.
(00423
The reason why reference processing surface 61 is
frame-like is that the attitude error can be thereby
accurately determined even when a large number of
measurement points are used, and the attitude error can
be determined more accurately by using the entire length
of the test piece to measure the inclination. Here, a
ball end mill is used as cutting tool 63. Reference
processing surface 61 is used as the reference for
measuring the attitude error at a specified rotation
angle of the rotational feed axes.
[0043]
Then, as shown in Fig. 16, the rotational feed axes
are indexed to each measurement point, and three surfaces
of the test piece that are orthogonal to each other are
processed by operating only the linear feed axes. As
shown in Fig. 15, a specified area is allotted as the
area to be processed in accordance with the indexed angle
of the rotational feed axis.
[0044]
Next, as shown in rig. 17, the rotational feed axes
are indexed to each measurement point, and P10-x'14 of
CA 02717291 2010-04-30
- 17 -
reference processing surface 61 are measured with touch
probe 64, and actual inclination of the line passing
through P10 and P11, actual inclination of the line
passing through P10 and P12, and actual inclination of
the line passing through P13 and P14 are determined.
Difference between three inclinations thus determined and
three theoretical inclinations computed from position
command of the rotational feed axes at the time of
measurement is taken as the attitude error.
[0045]
Then, as shown in Fig. 18, the rotational feed axes
are indexed to the reference rotation angle of 0 degree
for B-axis and 0 degree for C-axis, and processed surface
P15-P20 that are processed at each rotation angle are
measured, and difference between the positions of
processed surfaces P18-P20 that are processed with the
rotational feed axis indexed to 0 degree for B-axis and 0
degree for C-axis and the positions of processed surfaces
P15-P17 that are processed with the rotational feed axis
indexed to other rotation angle is determined.
[0046]
In the present invention, the difference of
positions and/or difference of inclinations between the
processed surfaces that are processed at one rotation
angle and the processed surfaces that are processed at
other rotation angle are referred to as the displacement
of the processed surfaces.
[0047]
From measurement data of the processed surfaces P18-
P20, intersection point P21 of three planes containing
processed surfaces P18-P20 when it is assumed that there
is no attitude error is determined. From measurement
data of processed surfaces P15-P17 and the determined
attitude error, intersection point P22 of three planes
containing processed surfaces P15-P17 is determined.
Difference between intersection point P21 and
intersection point P22 is taken as the position error.
CA 02717291 2010-04-30
- 18 -
The present invention includes a case where a test piece
or a-work piece is processed with a machine tool having
rotational feed axes on the main spindle side, and the
position error and the attitude error are determined from
measurement result of the processed surfaces.
[0048]
Error determined by the above-described method is
associated with the positions of linear feed axes X, Y, Z
and the rotation angles of rotational feed axes B, C, as
shown in Fig. 4, and is stored as an error map.
[0049]
Next, a correction method for correcting the
position command using an error map including the
position error and the attitude error will be described
taking a spindle rotation-type machine which has
rotational feed axes A, C (see Figs_ 1 and 2) as an
example.
[0050]
First, the command position of processing program 21
is read by reading and interpreting section 22, and
interpolation section 23 determines the command pulse for
each of feed axes X, Y, Z, A, C for every specified
interpolation period.
[0051]
Then, position command recognition means 24
recognize from the command pulses the position command
for each of feed axes X, Y, Z, A, C for every specified
interpolation period.
[0052]
If the position of each feed axis in the position
command is the same as the position of the measurement
point stored in error data storage section 25, error data
34 are obtained, and correction data are determined based
on the obtained error data. If the position of each feed
axis in the position command is not the same as the
position of the measurement point stored in error data
storage section 25, the error data are obtained from the,
CA 02717291 2010-04-30
- 19 -
error data of nearby measurement point using known
interpolation method, and correction data are determined
based on the error data thus interpolated. Correction
data thus determined are added to the position command to
obtain new position command for every interpolation
period. The position command is corrected in this
manner, and each of the feed axes can be positioned with
high precision.
[0053]
Next, a correction method for correcting the
position command by expressing in 3'-dimensional
coordinate values the correction value that has been
computed by the correction data computing section 26,
will be described. In case, for example, where, with C
axis set at the rotation angle of 0 degree, an attitude
error appears in a direction of B-axis which is not
proper to the machine, there is a problem that the
rotational feed axes need to be rotated considerably in
order to correct the attitude error in a direction of B-
axis. This problem is referred to as singular point
problem in the present invention. The correction method
described below is a correction method for circumventing
this singular point problem. B--axis is an axis of a
rotational feed axis that is parallel to Y-axis.
10054]
Fig. 19 is a flow chart showing this correction
method. Computation formula for determining the position
correction vector based on the attitude and attitude
error of a tool, the position and position error of a
tool, and the projecting length of a tool according to
this method is shown below, where:
L: distance from the command point to the position
of tool tip,
[I, J, K]; command attitude of the tool,
[dl, dJ, dK1: attitude error,
[dXl, dYl, dZl]: position error,
[dX2, dY2, dZ2]: position error of the tool tip
CA 02717291 2010-04-30
- 20 -
produced by the attitude error,
[dX3, dY3, dZ3]: position error of the tool tip,
dX2 = L x (tan (J'+dJ) / ((tan (I+dI)) 2 + (tan (J+dJ)) 2 +
1) 1/2
- (tan (J)/((tan (I))2 + (tan(J) )2 + 1)1/2)
dY2 = L x (tan (I+dI) / ((tan (I+dI)) 2 + (tan (J+dJ)) 2 +
1) 1/2
(tan(I)/((tan(I))2 + (tan(J) )2 + 1)1/2)
dZ2 = L x (1/((tan(I+dI))2 + (tan(J+dJ) )2 + 1) 1/2
(1/((tan(I) )2 + (tan (J) ) 2 + 1) 12)
dX3 - dXl+dX2
dY3 = dYl+dY2
dZ3 = dZl+dZ2
[0055]
First, at step SO, the command position and the
command attitude that are commanded by the position
command outputted from interpolation section 23 are
recognized. At step Si, error data 34 corresponding to
the command position are obtained from the error map. At
step S2, the position correction vector for correcting
the position error is computed from position error 34a of
error data 34.
[0056]
On the other hand, from attitude error 34b of error
data 34, at step S5, the attitude correction value is
computed. At step S6, the attitude correction value
obtained at step S5 is added to the command attitude read
out at step S3 to determine the attitude after
correction. At step S7, from the attitude after
correction obtained at step S6 and the projecting length
of the tool, the command point after correction is
determined.
[0057]
At step S4, from the command attitude read out at
step S3 and the projecting length of the tool, the
command point before correction is determined. At step
S8, position correction vector of the command point for
CA 02717291 2010-04-30
- 21 -
correcting the attitude error is computed by subtracting
the command point before correction obtained at step S4
from the command point after correction obtained at step
S7. This is referred to as the attitude correction
vector.
[0058]
The attitude correction vector is a vector
representing the magnitude and direction of the
displacement of the tool tip when, with the basal end of
the tool held by the main spindle as the control point,
the rotational feed axes are rotated so as to correct the
attitude error with the control point as the fulcrum.
[0059]
Finally, at step S9, the attitude correction vector
obtained at step S8 and the position correction vector
obtained at step S2 are added.
[0060]
In the present invention, the command point refers
to the position of the tool tip (tool tip position), and
the tool tip position means the actual position of the
tool tip, the position of the processing point of the
tool tip portion, the center of the semi-sphere of the
tip portion of a ball and mill, etc.
[0061]
As has been described above, the error of the tool
tip position is corrected only by the movement of the
linear feed axes, so that the rotational feed axes are
not rotated during the correction of attitude error 34b,
and the singular point problem can be circumvented.
[0062]
Thus, in accordance with the present embodiment, the
position error and the attitude error of a machine tool
having plural rotational feed axes can be measured and an
error map can be prepared. Since, in the error map thus
prepared, the position error and the attitude error are
stored as separate error data, the position command can
be corrected based on the error data, and the tool tip
CA 02717291 2010-04-30
- 22 -
can be positioned to a target position with high
precision to permit high precision processing.
[0063]
The present invention is not limited to the
embodiments described above, but can be implemented in
various modifications without departing from the scope
and spirit of the invention. For example, in the present
embodiment, numerical control device 20 comprises a
computation section that computes position error and the
attitude error at the measurement points based on the
measurement data obtained by measurement device 50 and
the coordinate values of the measurement points, and
error data storage means 25 for storing the position
error and the attitude error computed by the computation
section in correspondence to the position of the linear
feed axes and the rotation angle of the rotational feed
axes. It is also possible, in place of numerical control
device 20 to use personal computer or other device for
the computation section or the error data storage means.
C
