Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
1;~369~7
METHOD AND APPARATUS FOR PATTERN FORMING
Background of the Invention
The present invention relates to methods and apparatus for pattern
generation, and more particularly concerns improved methods and apparatus for
rapid formation of patterns of high precision by a combination of mechanical andelectrical drive.
In the manufacture of printed circuit boards, a circuit pattern is often
produced as a high quality black-and-white image on photographic film or plnte
and transferred to a coated substrate. Features in such an image may have a
positional accuracy of 0.001 inch over a distance of a much as 30 inches. The
image on such film may be produced by a photoplotter comprising a light source,
a lens system, a shutter, and a set of apertures for use in exposure of the film.
A photohead of the photoplotter is moved relative to the film by high precision
mechanical positioning devices in two axes and in two modes. In a "flash" mode,
the shutter is opened and closed while the photohead is stationary to produce animage of a selected aperture at a particular position on the film, and then the
photohead is moved to the next position, with closed shutter, where a second
flash is caused. In a "draw" mode, the shutter is left open while the photohead is
moved to draw a line of a width determined by the diameter of the selected
aperture.
A major problem with such photoplotters is that they are very slow. If the
image is of high density (more than 100 features per square inch) and large
(several square feet), the image may take hours to produce. Many applications
require several pieces of film for one job, and some jobs may require several
days of continuous plotting for completion. A second problem is lack of
versatility. The photohead can carry only a limited set of apertures of differcnt
sizes and shapes and is, therefore, unsuited for applications requiring a broad
range of image features, such as phototypesetting. A third problem is the high
degree of mechanical perfection required. In order to achieve desired position-
ing accuracy of features in the image, position of the photohcad must be
carefully maintained whenever the shutter is open. Maintenance of the position
in the draw mode is even more difficult, because the photohead is in motion.
There have recently been developed precision photoplotters employing
lasers to scan the film. The laser beam is rapidly turned on and off to produce
,
~2~3~i9~'7
_.
dark and light areas as it scans, frequently in a rectangular raster. Since the
image is produced in a raster mode rather than a vector mode, as is commonly
used in photoplotters, the digital data must be converted into the raster format,
a process that may take considerably longer than the plot itself. The amount of
5 prior data processing required for this conversion varies with the square of the
resolution of the raster. Precisely pointing the laser beam over a large area and
modulating it on and off at exactly correct times may present problems as
difficult as the mechanical problems of the traditional photoplotter.
Phototypesetters will generally expose a single line of print at a time, with
10 the paper moving transversely to the line for exposure of successive lines. In a
phototypesetter, one is concerned primarily with producing a high quality image,but geometric precision and precision positioning of the exposed images are of
little concern and are not available.
Accordingly, it is an object of the present invention to provide a precision
15 image-forming method and apparatus that avoids or minimizes above-mentioned
problems.
Summary of the Invention
In carrying out principles of the present invention in accordance with a
preferred embodiment thereof, an energy beam generator, which may be a
20 cathode ray tube having a small fiber optic faceplate, is successively positioned
at each of a plurality of adjacent areas of an energy sensitive medium upon
which a pattern is to be exposed. At each area of the medium, the energy
generator stops, and a section of the overall pattern is caused to be traced on a
target area or faceplate of the generator. The overall pattern to be exposed
25 upon the medium is divided into a number of small pattern sections, and the
beam generator is operated at each position to produce an individual one of suchpattern sections in a position on the medium that is precisely aligned with other
pattern sections that have been exposed on the medium. Mechanical positioning
error of the energy beam generator, which tends to cause misalignment of one
30 pattern section with respect to previously exposed pattern sections, is compen-
sated for by measuring mechanical positioning error when the energy bcam
generator has come to a stop, and employing the measured error to electroni-
cally displace the entire pattern section, as traced by the energy beam on the
beam generator target area, in a direction to decrease the measured mechanical
35 positioning error. Static distortions in the image or pattern section traccd by
the beam on the target area are also corrected by operating the image gencrntor
. .
'7
_,
to form a test pattern and measuring deviations of selected points of the tcst
pattern from the nominal test pattern configuration. The measured deviations
are used to compensate beam deflection signals during formation of an image by
the energy beam.
5 Brief Description of the Drawings
Figure 1 is a pictorial illustration, with parts broken away, of an exemplary
embodiment of apparatus incorporating principles of the present invention;
Figure 2 illustrates an exemplary etched circuit pattern;
Figure 3 illustrates a typical path of the energy beam traced in the course
10 of forming elements of a pattern section;
Figure 4 is a functional block diagram of an embodiment of the invention;
Figure 5 is an enlarged illustration of portions of the pattern of Figure 2,
illustrating mechanical positioning error;
Figure 6 illustrates one type of static image distortion; and
Figure 7 illustrates interpolation of static distortion corrections.
Detailed Description
In carrying out principles of the present invention in accordance with a
preferred embodiment that is illustrated herein for purposes of exposition, a
cathode ray tube, which images one small area at a time, is stepped over a
20 photosensitive film in as many steps as needed to form a complete image or
pattern on a film. The system is very fast, because imaging of any small area isdone by electronically sweeping the electron beam of the cathode ray tube, and
the tube makes only one pass over the film, regardless of density of image
elements. The pattern forming capability is versatile, since almost any pattcrn
25 can be formed with a cathode ray tube. Mechanical requirements are simplified,
because imaging is performed only when the tube is stationary. Further, there isno requirement that the tube be moved to a precise position, but merely that itsposition, when achieved, be precisely known. Small positioning errors of the tube
are measured, and the image on the tube faceplate is electronically shifted to
30 minimize position error of the tube. No conversion to raster format is required,
since the mechanical positioning of the tube and control of the cathode ray tubebeam are readily achieved to provide the pattern generation by electronic
motion in vector mode of the cathode ray tube beam.
Presently preferred apparatus for practicing principles of the prcsent
35 invention is illustrated in Figure 1. A massive fixed table lO, having a prccisely
., .
369~
flat upper surface 12 in which are formed a plurality of parallel rows of
apertures 14 connected via passages in the table (not shown) to a vacuurn
manifold 15, provides a support for vacuum attachment of a photosensitive film
or plate upon which a pattern, such as a printed circuit board pattern, is to beexposed. An X axis beam 18 extending along an X coordinate axis is fixed to the
table and mounts one end of a Y axis beam 20 for slidable motion on the beam 18
in either direction along the X axis. Beam 20 is mounted on the X axis beam 18
by means of suitable bearings, which preferably are air bearings, such as those
indicated at 22 and 24. The other end of the Y axis beam is slidably supported on
10 the table by air bearing 25. Beam 20 is driven in either direction along the X
axis beam by means of a screw 26 fixed to the beam 18 and engaging a nut (not
shown) carried in beam 20. A motor 28 drives the screw, and an X axis position
pickoff is provided in the form of an optical gauge or scale 30, which is fixed to
and extends the full length of X axis beam 18, and a light emitting and light
15 sensing diode assembly 34 carried by the beam 20.
A cathode ray tube housing and carriage 36 is mounted for linear travel
along Y axis beam 20 by means of suitable bearings, such as air bearings 38, 40,and driven by means of a motor 42 fixed to the beam and a ærew 44 journalled
on the beam, with the screw threaded in a nut (not shown) carried by the housing20 and carriage 36. A Y axis optical pickoff, substantially identical to the X axis
pickoff, is provided in the form of an optical scale 46, fixed to and extending
substantially the full length of the Y axis beam 20, which cooperates with a light
emitting diode and a light sensitive diode assembly 48 carried by the housing and
carriage 36.
Mounted in the housing and carriage 36 is an energy beam generator,
preferably in the form of a cathode ray tube 60, having a fiber optic faceplate
62. To minimize effects of stray magnetic fields, the tube neck is circumscribedby one or two concentric magnetic shields 64. An additional outermost
concentric magnetic shield (not shown) circumscribes substantially the entire
30 tube, extending longitudinally from the faceplate to a point well pnst the center
of neck 64. The cathode ray tube includes conventional electron energy beam
generating electrodes and deflection devices such as magnetic yokes or electro-
static deflection plates (not shown). The cathode ray tube is restrained againstall horizontal motion relative to the housing and carriage 42, but is mounted
35 therein by means (not shown) for a limited amount of vertical motion under
control of a vertical drive motor 66 carried by the carriage housing nnd driving a
screw 72 that is threadedly engaged with the cathode ray tube.
~3~ 7
-
The face of the cathode ray tube is small, in the order of bout two to
three inches, havirlg a circular configuration and being precisely planar both
inside and out. As is common with a fiber optic faceplate, the cathode ray tube
phosphor is deposited upon the inner surface of the faceplate, which is formed by
the innermost ends of a large number of closely packed optical fibers. Prefer-
ably, the phosphor is of an extremely low permanence, so as to have an
extremely short retention of its state of optical exeitation. The vertically
shiftable mounting of the eathode ray tube enables the faeeplate to move
vertieally between an imaging position in whieh the faeeplate, in the present
embodiment, is in flat, direet faee-to-faee eontaet with the surfaee of the
photosensitive medium (sueh as film) seeured to the table, and one or more
raised positions in whieh the faeeplate elears the film to facilitate motion of the
eathode ray tube across the film, or to clear a thicker photosensitive plate that
may be used instead of the photosensitive film. Alternatively, the vertical
motion of the tube may be eliminated and the tube floated upon the photosen-
sitive film upon a thin cushion of inert gas.
It will be seen that the cathode ray tube can be moved in two dimensions
across the surface of the sensitive medium, sueh as a photosensitive film or
plate, that is seeured to the table. The eathode ray tube housing and earriage is
moved in diserete steps over the table surfaee, employing the X and Y drive
motors 28 and 42 in eonjunetion with optieal position feedbaek signals derived
from the optieal seales and sensor diodes. Although elosed loop position eontrolis employed, high preeision imaging is obtained with the use of relatively looseservo eontrol of physical position of the housing and carriage, and therefore ofthe cathode ray tube itself. Eleetronie image positioning compensates for loose
servos, and distortion controls eompensate for distortions due to cathode ray
tube characteristies as will be described more particularly below.
Illustrated in Figure 2 is an exemplary rectangular pattern 70 to be used in
making a printed circuit board, and having a plurality of pad areas 72, 74, 76
interconnected by lines 78, 80, and 82, among others. According to principles ofthe present invention, the overall pattern 70 is subdivided into a number of
square sections or subsections, sueh as the 16 square sections illustrated in
Figure 2 for purposes of exposition. Eaeh pattern seetion is in the order of oneineh square in a presently preferred embodiment to correspond to and be
eongruent with a one ineh square pattern that will be formed on the faee of the
cathode ray tube. Each of the exemplary 16 squares is designated as being in oneof four columns, labeled 1 through 4, and in one of four rows, identified as A
,
l~t3~ '7
through D. Since an actual machirle embodying principles of the prescnt
invention is capable of forming a single overall pattern in the order of severalfeet square (depending upon table size), it will be understood that a pattern ofsuch larger size would be divided into many more than 16 one inch square
subsections.
In forming the overall pattern with the apparatus illustrated in Figure 1,
the cathode ray tube is first moved to position it over an area of the
photosensitive medium that corresponds to the pattern section identified as A1
(for row A and column 1). The illustration of Figure 2 may be considered to be
an illustration of the desired pattern as it is to be directly exposed upon the face
of the photosensitive medium. Thus, in employing the apparatus of Figure 1 to
form a pattern such as that illustrated in Figure 2, the cathode ray tube, with its
beam blanked out, is raised slightly above the surface of the medium, moved to
position it at section A1, and then lowered to place the fiber optic faceplate in
direct face-to-face contact with the film surface. When the cathode ray tube
motion has stopped and the apparatus is perfectly still, the cathode ray tube isoperated to unblank its beam and to cause its deflection control circuits to move
the projected electron beam over the tube target area formed by the inner
surface of its fiber optic faceplate, so as to trace that portion of the overallpattern that is defined within the boundaries of section A1. Impingement of the
scanning electron beam upon the inner surface of the faceplate causes a spot of
light to scan the faceplate in the desired pattern section. The light is
transmitted by the fiber optic faceplate directly to the surface of the photo-
sensitive medium upon which the pattern section of section A1 is thus exposed.
Although many types of image forming data may be employed for image
generation, in a presently preferred embodiment a vector mode is used in which
each line of a pattern is divided into short segments of which each end is
identified by x, y coordinates that are set forth in successive input data
commands, together with a command to turn the beam on (when tracing a line
segment) or off (when shifting to another pattern element). To trace a circular
pad 89 of 60 mils diameter, for example, as illustrated in Fig. 3, the beam (which
is focused to form a spot of one mil diameter on the faceplate) is first moved to
a point such as point 90 on the pad periphery, and then moved in successive
straight line segments completely around the circular pad, to points 91, 92, etc.,
each of which is defined in the image forming data by a pair of x, y coordinatesthat are fed as analog signals to the tube deflection controls. After completingthe trace of the pad perimeter, the beam is caused to fill in the intcrior (for a
.
1~36~3~.7
solid pad) by successively tracing mutually displaced lines such as those indicated
at 95, 96, 97, which are shown as distinct, separated lines only for exposition. In
an actual device, the fill-in lines are all mutually contiguous to provide the
desired solid pad. A similar procedure is followed for .racing a straight line 98
of ten mils in width, for example. The beam first traces each of the short line
segments making up one edge 99 of the line and then traces the segments of the
other edge 100. Line 98 is then filled in by causing the beam to trace segments
forming interior lines such as those designated at 101, 102, 103. During all of
this operation of the tube, the tube and its carriage and housing remain
stationary. Only the electron beam is moving to trace an image only within the
boundaries of a single pattern section A1.
Upon completion of the exposure of the pattern section A1, the beam is
blanked, the cathode ray tube is raised, moved to the next area of the film
(which may be section A2, for example), lowered to provide the desired face-to-
face contact between the fiber optic faceplate and the film, and when motion of
the cathode ray tube has been completely stopped, the cathode ray tube is
operated to expose upon the optically sensitive film that pattern section
contained within the square section A2. This step and expose cycle is repeated
for each pattern section, successively moving the cathode ray tube to each area
of the film and, at each such area, exposing the film to the corresponding section
of the overall pattern, until all sections of the overall pattern have been exposed
upon the film. Only a single pass is required, since all elements of each section
of the pattern are exposed on each exposure step. The cathode ray tube beam is
disabled so that the faceplate is dark during all motion of the cathode ray tube.
Only after motion of the cathode ray tube has been stopped is its electron beam
deflected to a selected position and its intensity control energized to cause the
beam to form a light spot on the faceplate.
In prior systems employing vector mode, a single coordinate system (a
"pattern" coordinate system) is established for the entire pattern. Coordinates
of all points for the entire pattern are expressed in the one pattern coordinatesystem. According to an important feature of the present invention, a number of
"section" coordinate systems are established, there being a different section
coordinate system for each pattern section. Coordinates of all points are
expressed in the respective section coordinate system.
Referring to the pattern of Figure 2, beam deflection coordinates are
initially expressed in an X, Y pattern coordinate system having an origin at point
90, so that a point 92, for example, on the periphery of pad 76 has coordinates xl,
~36g~7
Yl in the overall pattern coordinate system X, Y. As previously mentioned, each
point at an end of each line segment is identified by its x, y coordinates in the
pattern coordinate system X, Y. To enable the image generator to form an
image corresponding to each of the individual pattern sections, such as sectionsAl, A2, etc., the pattern coordinate data for each pattern point in the overall
pattern are transformed into section coordinate data for each point in an
individual pattern section. The section coordinate data for each individual
pattern section are expressed in a section coordinate system that is unique to the
particular section. For example, to identify points in the pattern section B2, asection coordinate system X', Y' is established with its origin at a point 100,
which is one corner of the pattern section. The section coordinate system X', Y'has its axes parallel to the axes of the pattern coordinate system X, Y, but itsorigin 100 is displaced from the pattern coordinate system origin 90 in X and Y
by the extent of the length of one pattern section in each of the X and Y
directions. The origin 100 of the section coordinate system X', Y' has
coordinates xO, yO in the pattern coordinate system X, Y. Accordingly, the X
coordinate x'1 of point 92, expressed in the section coordinate system, is
x'1 = x1 - xO, and the Y coordinate Y'l of point 92, expressed in the section
coordinate system, is Y'1 = y' - yO. Thus, data transformation of the line
segment coordinates is simply a matter of calculating coordinates in the patternsystem X, Y of a like corner (lower left corner, for example) of each pattern
section to thereby determine the pattern coordinates of all of the section
coordinate system origins. Then, from these section coordinate system origins
and the coordinate data expressed in the pattern system X, Y, the coordinates ofeach point expressed in each section coordinate system may be readily deter-
mined. Coordinates of the section coordinate system origins, as expressed in thepattern coordinate system, also are employed as positioning commands to the
drive motors 28, 42 for stepping the cathode ray tube to its successive positions
on the film.
With the cathode ray tube positioned at subsection B2, for example, the
pattern section at this area may be formed on the cathode ray tube faceplate.
This is achieved by feeding to the tube deflection control circuits coordinate
information locating each of the points of this pattern section in the section
coordinate system. For point 92, the cathode ray tube deflection circuits are
energized by analog signals corresponding to coordinates x'i and Y~i. All other
points within the pattern subsection B2 are likewise defined and the beam
deflected by coordinates in the section coordinate system X', Y'. Similarly, an
unique section coordinate system is employed for each pattern section.
:,_
1~3~ ;2 7
The system, including the cathode ray tube, its mechanical and electronic
positioning, and all of its motions and intensity, focusing and deflection controls,
are controlled by a conventional computer programmed to calculate from input
coordinate data the necessary commands for positioning the cathode ray tube in
s its step~y~tep traverse over the film medium and for controlling the scanning
of the electron beam over the cathode ray tube faceplate at each of the
successive positions of the image generator. Input data also includes commands
for blanking and unblanking, varying intensity and changing focus, if needed.
Input data may all be specifically prepared for a given pattern to be exposed,
10 including prior calculation of coordinate system origin coordinates and line
segment coordinates in the section coordinate systems, with the data aU
organized in blocks, each of which is individual to a single pattern section. The
computer stores the input data, as blocks of section coordinate data and relatedcommands, and reads out the various commands for control of the machine
15 motors and cathode ray tube to perform the operations described herein.
Various vector mode formats of data bases for photoplotters are known and
have been used. Thus, as an alternative to calculating the above described inputdata directly from a pattern to be exposed, there may be employed a pre-
existing data base in a common photoplotter format, such as, for example, in the20 so-called Gerber format which is employed for control of Gerber photoplotters.
Such data may be readily converted to the type of input data required for
operation of the present invention. A typical series of commands embodied in a
Gerber format data base for a Gerber photoplotter may be as follows: Select a
small round shape of .012 inch in diameter. Move the photohead to a first point
25 (identified in a pattern coordinate system) with aperture closed. Move the
photohead with the selected aperture open from the present point to a second
point (thereby drawing a line of the selected width between such points). Selecta large round shape of .062 inch in diameter. Move to a third position with the
shutter closed, and open the shutter at such third point (to thereby draw a solid
30 round shape of .062 inch in diameter at such third point).
Modifying such a data base for use as input data to the described
apparatus, a command to draw a line of a given width between given points is
changed into a series of commands to draw a plurality of line segments, first
along the outside edges of the line and then to fill in the line as previously
35 described. In a like manner, a command of the prior data base to draw a solidround shape of a given diameter at a specified location is changed to a plurality
of commands to draw a series of successive line segments extending around the
1~369~
-
perimeter of the solid shape, all centered upon the center of the so]id shape, and
an additional series of commands to fill in the interior of the perimeter thus
drawn. The prior data identified in single pattern coordinate system is divided
into data blocks, each corresponding to an individùal one of the pattern sections
5 in which the overall pattern is to be divided. Thus, it will be apparent that input
data for operating the described machine may be derived either directly from a
pattern to be exposed or from some type of previously prepared data base for
such pattern.
Portions of a controlling computer are functionally illustrated in the
10 functional block diagram of Figure 4. The cathode ray tube 60 is moved along
the table 12 under control of motors 28, 42 which are energized from a moving
position error comparator 110 in response to motor commands stored in X, Y
position command storage 112 of the controlling computer.
The X, Y cathode ray tube position commands, which define successive
15 cathode ray tube positions in terms of coordinates of the origins of the
respective section coordinate systems, are extracted from X, Y position com-
mand memory 112 and compared with X, Y feedback signals from the optical
pickoffs 34, 48 in the moving position error comparator 110 which feeds X and Y
error signals to the motors 28 and 48 to control their operation and move the
20 cathode ray tube to the desired positions. Signals for raising and lowering the
cathode ray tube are not illustrated in Figure 4.
Section coordinates of each point of the pattern, and related blanking
commands, are stored as data blocks, each of which is specific to an individual
one of the subsections of the pattern. Deflection cGntrol signals corresponding
25 to each pair of subsection coordinates x, y are stored in deflection tables 124,
which are entered, for each point, with x, y section coordinate data from the
data blocks in storage 122 to provide beam deflection commands that are added
to other deflection control signals (to be more particularly described below) in an
adder 126. From the output of the adder the final deflection control signals are30 fed to digital-to-analog converters 128 which send appropriate analog signals to
the deflection controls, such as the conventional magnetic yokes 130 of the
cathode ray tube. All line segments are traced in the same fixed time interval,
regardless of segment length. Therefore, the deflection control circuits for theX and Y beam deflections each includes an analog computer (not shown),
35 responsive to the difference between coordinates of the two ends of a segment,
that computes the rate of change of voltage required to deflect the electron
beam from one end of the segment to the other in the common fixed time
1~3~ 7
11
interval. Intensity control of the cathode ray tube is provided by controls 132
via computer commands, which, as previously described, blank the electron bcam
during motion of the cathode ray tube and control tube intensity in accordance
with known techniques during the forming of an image on the cathode ray tube
faceplate. Beam focus controls 134 are employed to vary the focus as necessary
or desirable. For example, the focused beam that produces a one mil spot on the
faceplate for tracing a pattern element outline may be slightly defocused or itsintensity slightly increased to fill in traces with a larger spot within the outline.
Even with electromechanical servos of exceedingly high precision, it may
be difficult, if not impossible, to physically position the cathode ray tube to
enable it to trace an image section that is precisely aligned and registered with
pattern sections that have been previously exposed upon the same film. How-
ever, neither precision servos nor precise mechanical positioning are required.
According to a significant feature of the invention, the electromechanical
positioning servos may be made relatively loose (subject to relatively large
positioning error), and therefore of increased efficiency and speed, and extremeprecision of pattern forming is still achieved.
Figure 5 illustrates an exaggerated mechanical positioning error. The
cathode ray tube is operated to first form the image or pattern section within
the boundaries of section Al, including pad 72, line 78, and a first section 80a of
the line 80, which extends across the imaginery boundary between image sections
A1 and A2. After exposing the image of the pattern section of section Al upon
the film, the cathode ray tube is moved to the position of section A2. But, as
previously mentioned, the physical positioning may be such that when the
cathode ray tube comes to rest it is displaced both in X direction and Y direction
from the position in which section A2 would be precisely registered and in
alignment with the adjoining section Al. If at this time the cathode ray tube
were operated to expose the image section A2, pad 74 and a second section 80b
of the line 80 that interconnects pads 72 and 74 would be displaced, as illustrated
in solid lines in Figure S, from their nominal or desired positions (shown in
phantom in Figure 5). According to a feature of the invention, this physical
displacement of the cathode ray tube from its nominal position when it has come
to rest (purportedly but not actually at its nominal position) is corrected by
electronically shifting the entire image section of section A2 in a sense to
decrease the misalignment. This is achieved by using the same position pickoffs
34, 38 that are used for controlling the drive motors in the motor servo loop.
With the cathode ray tube at rest, X and Y position error signals from the optical
-
l~?s3~ 7
12
pickoffs are compared, respectively, with the X, Y cathode ray tube position
commands in a static position error circuit (Figure 4), which provides a second
input to the adder 126 in the form of static position error signals in X and Y that
are each proportional to the difference between commanded and measured
5 positions. This is achieved independently for each of the measured X and Y
positions, although the drawing shows but a single static position error circuit.
In effect, the system measures the positioning error caused by the lack of
precision of the X and Y servos, after the cathode ray tube has come to rest, and
provides fixed biases to each of the X and Y deflection controls. Each bias is of
10 a magnitude and sense that tend to shift the entire pattern section relative to
the cathode ray tube faceplate so as to precisly align the second pattern section
A2 with the pattern section A1 that was exposed upon the film when the cathode
ray tube was at a previous position.
In effect, the system employs coarse and fine image positioning to
15 precisely locate each pattern section upon the film. Coarse positioning is
achieved by mechanical operation of the X and Y servos which physically
position the cathode ray tube. Fine positioning, which is accomplished after thecoarse positioning has been achieved, is thè electronic positioning or shifting of
the entire pattern section relative to the cathode ray tube faceplate, which is
20 carried out without motion of the cathode ray tube itself. This combination of
electromechanical and electronic positioning of the pattern section not only
enables use of faster, more efficient servos of less precision, but also minimizes
or eliminates other positioning problems. Because the fine positioning takes
place while the cathode ray tube is at rest, physical vibrations have settled and
25 faded to substantially negligible proportions, and mechanical resonance and other
problems of mechanical dynamics are avoided. The final positioning is accom-
plished in the absence of resonant vibrations and other dynamics of machine
motion. It may be noted that the relative size of the cathode ray tube faceplateand the maximum pattern section to be traced thereon are such that a buffer
30 area is provided on the faceplate entirely around the pattern section to enable
bodily shifting of the pattern section as a whole relative to the pattern faceplate
in a magnitude sufficient to compensate for expected and predetermined
electromechanical positibning errors.
Although the several pattern sections are positioned by the coarsc and fine
35 positioning described above, each pattern section is subject to static distortions
due to characteristics of the tube. The several elements of each pattern section,
such as various line segments that make up a circular pad or a straight line for
~369;~7
example, are electronically drawn by controlling electron beam deflection to
cause the beam spot upon the faceplate to traverse the desired path. However,
position of the point of impingment of the electron beam upon the cathode ray
tube faceplate is not necessarily a predetermined or linear function of analog
5 deflection voltages applied to the deflection control, because the cathode raytube is subject to a number of what may be termed static errors of the bcam
positioning, which result in a distorted image. In such static distortion, there are
included errors in cons`truction of the tube, its magnetic deflection yoke, the
tube support hardware, and its analog electronics. For example, the axis of the
10 beam, namely the beam path with zero deflection applied to both X and Y
deflection yokes, may not be precisely normal to the flat surface of the
faceplate. Further, the two deflection coils may not be precisely perpendicular
to a normal to the faceplate, nor precisely perpendicular to each other or to the
physical X and Y axes of the machine carriage and beams. An ideal structure is
not physically possible to manufacture, and therefore, deflection signals fed tothe cathode ray tube deflection controls that are calculated to produce, for
example, a square pattern on the faceplate may actually produce a pattern that
is trapezoidal or otherwise distorted. Further, as previously mentioned, the
faceplate, which is the target area of the beam, is flat on both interior and
20 exterior surfaces. It is not curved, as in a common configuration of cathode ray
tube display screens. Accordingly, when deflection signals that are fed to the
deflection controls are produced of a magnitude to form a square, such as square150 in Figure 6, they will actually produce on the faceplate a distorted square
having curved sides 152, rather than the straight sides required of the desired
square. The illustrated distortion is exaggerated for exposition. For example,
considering Figure 6 to represent a view of the interior of a cathode ray tube
faceplate from a point at the electron beam emitting cathode of the tube, one
would attempt to trace a straight line 154 by using a fixed y deflection and
varying the beam x deflection to sweep the beam across the faceplate along line
154 at a fixed distance from the faceplate center. However, because of the y
deflection, the distance along the electron beam from the cathode to the
faceplate at a midpoint 159 of the beam sweep (where x deflection is zero) is
less than the length of the beam from the cathode to the faceplate when the x
deflection is at either of its maximum values or at any value other than zero.
This is inherent in the geometry when using a flat faceplate. Accordingly, when
the beam is at a large x deflection, such as the deflection x illustrated in Figure
6, it will impinge upon the faceplate at a point 156 rather than at a point 158.
69:~7
14
Since this is true of all x deflections other than zero (for a given relatively large
y deflection), the beam will trace a curved line, such as line 152, on the flat
faceplate rather than a straight line. This is another one of the distortions
inherent in characteristics of the tube that are termed static distortions.
For increased precision of the pattern section to be formed on the
faceplate, compensation is also provided for these static distortions. Static
distortions are individual to specific tubes, and each cathode ray tube is
individually calibrated for its own unigue static distortion compensation.
To accomplish static distortion correction for an individual cathode ray
tube employed in carrying out principles of the present invention, the tube is
operated by deflection signals that would produce a predetermined nominal test
or calibration pattern (if there were no static distortions) extending through anumber of different points on the cathode ray tube faceplate. An image of this
test pattern is exposed upon a photosensitive film, and, under high magnification,
the distortions or deviations of the actual test pattern from the nominal test
pattern are measured at each of a selected number of points, such as, for
example, the points of a 128 x 128 matrix. It is not necessary to measure
distortions at each of the possible points of impingement of the electron beam
upon the pattern section. A number of points are selected and coordinate
positions and deviation magnitude and sense are measured for each and stored in
a correction map table 160. The correction map table is entered with the x, y
coordinate positions of a given point, and the appropriate deviations or inter-
polated values thereof are extracted or computed, as functionally indicated at
162, to provide X and Y static distortion corrections to the adder 126. If
interpolation is carried out for determining distortion corrections, it may be
accomplished in any suitable manner. For example, as illustrated in Figure 7, todetermine distortion corrections for a point 166, the measured distortions at
each of four points 167, 168, 169, and 170 at the corners of a small square
surrounding point 166 (and at each of which x and y components of distortion
have been measured) are applied for a linear interpolation of the x and y
components of distortion at point 166. Thus, each set of beam deflection
commands extracted from deflection tables 124 is modified by a set of static
distortion corrections and also by a set of static position error corrections bysimply adding or subtracting the appropriate digital values to the digital
deflection commands from the deflection table. The resulting deflection
command signals fed to the digital-to-analog converters 128, and thence to the
analog dejection controls 130, are thus compensated for measured electro-
~236~32~7
mechanical positioning errors of the cathode ray tube and for static distortionsdue to characteristics of the individual cathode ray tube. Accordingly, the
pattern section, such as section A2 shown in Figure 5, is not only correctly
positioned, registered, and aligned with previously imaged pattern sections,
regardless of the servo positioning errors, but distortions in the pattern caused
by individual tube characteristics are substantially decreased.
An alternative arrangement for measuring static distortions may take the
form of an array of photosensitive diodes mounted below the surface of the tablewith a lens mounted in an aperture formed in the table for focusing a light spotfrom the cathode ray tube faceplate upon the diode array. Prior to forming a
desired pattern, static distortions would be measured at each of a number of
points at predetermined section coordinates by physically positioning the tlJbe
faceplate over the table aperture (with no film or other medium on the table
surface) and moving the electron beam, point by point, to illuminate each of a
selected group of programmed nominal points on the diode array. For each such
programmed point the deviations of the actual point of iIlumination are
measured by noting the particular diode or diodes il1uminated and storing these
deviations in the correction map table 160.
PATTERN FORMING METHOD
Pattern forming, according to the presently preferred form of the inven-
tion, is carried out by performing steps substantially as follows:
l. Compute the section origin coordinates X0, Y0 for each section.
2. Separate the X, Y data (in the pattern coordinate system) into
section data blocks for each section, utilizing the calculated origin
coordinates for identification of the individual data blocks.
3. For each block compute and store the x'j, Y'i coordinate data (in the
individual section coordinate system) for all points within the section,
from the section origin coordinates and the x, y pattern coordinnte
data for each point.
30 4. Compute deflection control data for each pair of section coordinates
of a pattern section and store in the X, Y deflection tables at the x, y
address. (The same X and Y deflection tables are used for all pattern
sections.)
5. Store a list of section coordinate system origins as X, Y position
commands for positioning the cathode ray tube.
1~369;~7
",
16
6. Form a test pattern, measure and store distortions at a selected
number of points over the area of the test pattern, and store in the
eorrection map table.
All of the above may be accomplished before beginning operation of the
5 pattern forming process and may provide input data for operation of the system.
7. For operation of the cathode ray tube, feed a first set of X, Y
position commands to the motors, employing the optical feedback to
position the cathode ray tube at a first pattern area of the film.
8. When the cathode ray tube has stopped, as indicated by signals in the
positioning servos, select the data bloek for the present cathode ray
tube position and turn on the deflection eircuits and intensity
controls (although beam intensity is not turned up until the beam has
been deflected to a point at whieh a line segment is to be drawn on
the pattern seetion).
9. Read the x'i, Y'i section eoordinates point by point from the seleeted
data bloek.
10. Use eaeh x'i, Y'i to feteh the deviation signals from the eorreetion
map table and interpolate to obtain the statie distortion eorreetion
for the partieular besm defleetion set.
11. Read the beam defleetion command set from the deflection tables.
12. Read the optical pickoff signal and compute the static position error.
13. Combine the static distortion correction, the static position error,
and the beam deflection commands to obtain the eorreeted defleetion
signals for eaeh point.
14. Use the eorreeted defleetion signal to defleet the eleetron beam.
15. Repeat steps 9 through 14 for eaeh point in the data block corres-
ponding to the image section being formed.
16. After extraeting and employing the last set of beam defleetion
eommands, blank the tube and return to step 7 to move the eathode
ray tube to the next position.
Instead of raising the eathode ray tube from the table during the eourse of
its traverse and then moving it back down to contaet the film when its selected
position has been obtained, a ring may be fixed to the periphery of the cathodc
ray tube faceplate having a plurality of apertures through which a pressurizcd
35 inert gas, such as nitrogen, is projected to form a thin gas-bearing cushion
between the faceplate and the optieally sensitive film. In sueh an arrangement,
the vertieal lifting and lowering of the eathode ray tube between each pattern
- 1~369~7
17
section image forming is eliminated, and thus speed of the entire operation is
considerably increased. However, the presence of the gas cushion, however thin,
decreases resolution of the image exposed upon the sensitive film, and thus
choice between the gas cushion or the vertically shiftable cathode ray tube
5 involves a trade-off between pattern forming precision and pattern forming
speed. At present, it is believed that pattern forming speed is adequate with a
vertically shiftable cathode ray tube, and thus the latter is preferred because of
its increased precision. Where the vertically movable cathode ray tube is
employed, the vertical drive motor of the cathode ray tube is operated to raise
10 or lower the tube at appropriate points in the above-described series of steps.
It will be readily understood that intensity of the beam may be varied at
different points in the formation of an image section on the cathode ray tube
faceplate, as may be deemed necessary or desirable. For example, where the
machine provides for a fixed time interval in which to move the beam along a
15 given line segment, regardless of the length of the line segment, the speed of
travel of the point of impingement of the beam over the faceplate is greater fora longer line than for a shorter line. Accordingly, beam intensity may be
decreased for the shorter line because of the longer dwell time of the beam uponthe faceplate. Further, beam intensity may be slightly decreased for increased
20 precision of forming a relatively fine line or for forming the outline of a solid
pad, whereas, for filling in the interior of a solid pad, beam intensity may be
increased (or the beam may be slightly defocused), since precision of illumination
of the sensitive film within the area of the pad is not of great significance.
It will be readily appreciated that many embodiments and alternative
25 arrangements of the described apparatus and system may be employed to carry
out principles of the present invention. Other two dimensional drive structures
for achieving relative motion of the cathode ray tube and film may be employed.
For example, the cathode ray tube may be fixed and the table moved in X and Y
relative to the tube, or the tube may be moved in one dimension and the film and30 its support in the other. Thus, the table may take the form of a rotating drum
carrying the film, with the cathode ray tube being shiftable parallel to the drum
axis. Instead of employing a fiberglass faceplate, a plain glass faceplate may be
used on the cathode ray tube by projecting its image through a lens onto the
film. Such an arrangement would provide higher resolution, but poor optical
35 efficiency, because the lens is able to capture only a small percentage of the
total light. This would greatly decrease plotting speeds, reducing a primary
advantage of the described plotter, which is its high speed. In a presently
9 ;~ %~7
18
_. .
preferred embodiment, employing a vertically shiftable cathode ray tube, the
apparatus is capable of exposing one pattern section (one inch square) per
second. Physical positioning of the cathode ray tube, including raising, lowering
and laterally sh;fting the tube, requires about one half second, and the forming 5 of a single pattern section on the cathode ray tube faceplate similarly requires
about one half second. Accordingly, the formation of a complete image 20
inches by 20 inches can be performed in less than seven minutes.
Principles of the present invention are also adaptable for use with an image
generator in which an electron beam impinges directly upon a workpiece, which
10 itself forms the beam target area, with the beam being deflected to form an
image section by conventional beam deflection controls. In the absence of a
faceplate, both the cathode ray tube and the workpiece, relative to which the
cathode ray tube is movable in two dimensions, are contained within a vacuum
chamber during pattern forming.
There have been described methods and apparatus for rapid precision
forming of patterns by employing an energy generator stepped over the surface
of a sensitive medium in a number of discrete steps to form at each medium area
a single section of an overall pattern. Means are provided for coarse and fine
positioning of the image generator to provide precision of alignment of each
20 image section with each other image section and to compensate each image
section for various distortions inherent in the beam generator. Although an
image generator having a larger area of faceplate or target area would increase
speed of operation of the machine in the forming of a complete pattern, it is
presently preferred to employ an image generator with a faceplate or target
25 area that will form image sections of one to one and a half inches square. Animage generator capable of producing an image larger than about one and a half
inches square is extremely difficult to control for achievement of desired
precision. As diameter of the faceplate increases, the difficulty of manufac-
turing a faceplate sufficiently free of defects greatly increases. Similarly, the
30 magnitude of other errors increases with increase in tube size, and cost of the
tube rises with increase of faceplate diameter, because of severe manufacturing
problems. Use of a fiber optic faceplate is preferred at present to the use of alens with a plain glass faceplate, because the fiber optic faceplate is as much as
thirty times more efficient in transferring light from the faceplate to the
35 sensitive medium.
~L~,369~7
19
The foregoing detailed description is to be clearly understood as given by
way of illustration and example only, the spirit and scope of this invention bcing
limited solely by the appended claims.
What is claimed is:
,
