Stereolithographic beam profiling

- 1 334052 1 75358-4 D~-~CRTPTION StereollthG~La~l,lc Beam Proflllnq Background of the Inventlon 1. Cross-Reference to Related APpllcatlons Thls appllcatlon relates to Canadlan Patent Appllcatlon Nos. 596,850; 596,837; 596,825; 596,838 and sections of No.

596,827, all filed on Aprll 17, 1989 in the name of 3D Systems, Inc.

2. Fleld of the Inventlon This inventlon relates generally to apparatus and methods for measuring beams of electromagnetic radiation or particles and, more particularly, to a new apparatus and method for measurlng the intensity profile of beams and related appllcations 'C 1 334052 -2- 60724-1890 ln connectlon wlth the productlon of three-dlmenslonal ob~ects by stereolithography.

4. Backqround of the Invention In recent years, "stereolithography" systems, such as those described in U.S.

Pat.

No. 4,575,330 entitled "Apparatus For Production of Three-dimensional Ob~ects By stereollthography," have come into use. Basically, stereolithography is a method for automatically bulldlng complex plastic parts by successlvely printing cross-sections of photocurable polymer or the like on top of each other untll all of the thln layers are ~olned together to form a whole part. Wlth this technology, the parts are llterally grown in a vat of llquld plastlc. Thls method of fabrlcatlon ls extremely powerful for qulckly reduclng deslgn ldeas to physlcal form and for making prototypes.

Photocurable polymers ("photopolymers") change from llquid to solid ln the presence of light and their photospeed with ultraviolet light (UV) ls fast enough to make them practical model bulldlng materlals. The materlal that ls not polymerized when a part ls made ls still usable and remalns ln the vat as successlve parts are made. An ultravlolet laser may be used whlch generates a small lntense spot of UV llght. Thls spot ls 1 334052 moved across the liquid surface with a galvanometer X-Y mirror scanner.

The scanner is driven by computer generated vectors or the like.

Precise and complex patterns can be rapidly produced with this technique.

The laser scanner, the photopolymer vat, and an elevator, along with a controlling computer, combine together to form a stereolithography apparatus, referred to as an "SLA." An SLA is programmed to automatically make a plastic part by "drawing" one cross-section at a time, and building it up layer by layer.

Stereolithography represents an unprecedented way to quickly make complex or simple parts without tooling.

Since this technology depends on using a computer to generate its crosssectional patterns, a natural data link to CAD/CAM exists.

To be effective, a stereolithography system must have information about the focus, laser beam oscillation mode, beam power, intensity distribution or profile, and scanning system drift of the drawing laser in order to carry out the accurate and efficient production of parts (objects made by stereolithography are known as "parts").

The beam must be in relative focus at the surface of the working photopolymer fluid. The laser mode, intensity distribution, and beam power are important to the depth and width of cure of the working fluid, as weli as the scan speed. The "drift" of the scanning system must be measured and corrected periodically.

Beam profile (a profile of the intensity of the beam) measurements provide useful information about the beam because they can help accomplish the following purposes:

1. Focus optics and correct astigmatism and other aberrations; 2.

Measure the power of the beam (needed on a day to day basis); 3.

Study the laser mode and changes of the mode; 4.

Compensate for drift of the laser scanning system; 5.

Allow recording of the drift for later analysis of changes; 6.

Automatically calibrate the scanners; 7.

Allow easy control of beam position for making other 4 1 334052 measurements (e.g. to independently measure the beam power in order to find the system's power calibration factor) and 8.

Permit the prediction of the size and shape of the cured plastic trace.

Accordingly, a need exists for an apparatus and method to rapidly determine the profile of the beam, particularly in connection with a stereolithographic apparatus, to calibrate and normalize a stereolithographic apparatus, and to correct for the drift of the mirror positioning system of the stereolithographic apparatus.

.

Summary of the Invention Briefly, and in general terms, the present invention provides a new and improved apparatus and method for profiling a beam. The apparatus comprises means for measuring the intensity of a portion of the beam when the beam is incident on the measuring means and means for changing the relative perpendicular displacement of the measuring means from an optical path followed by the beam in order to measure the intensity of some or all portions of the beam along a surface substantially perpendicular to the optical path followed by the beam. The method for profiling a beam comprises the step of measuring the intensity of a crosssectional portion of preselected size of the beam along a surface substantially perpendicular to the optical path followed by the beam and repeating the measuring step for other portions of the beam along the surface.

The apparatus and the method develops an intensity map of the beam along a surface substantially perpendicular to the optical path followed by the beam.

The intensity map gives the intensities for each of the portions of preselected size of the cross-section of the beam. The intensity profile so gained can be used to determine and adjust the focus of the beam as well as to calculate the power of the beam (given a known power conversion factor). The profile of the beam may be used - 1 334052 to predict the cure depth and width of plastic created on the working fluid. The beam profiling apparatus may be used to detect drift of the apparatus that scans the beam, by serving as a fixed reference point and determining the scanning apparatus coordinates of the center of the beam, which can be used to recalibrate the positioning "map" or table that directs the scanning apparatus in translating computer generated designs to actual dimensions on the surface of the fluid that solidifies to form the object.

The presently preferred version of the beam profiling system has a significant advantage of economy because it uses the computing system and light-beam positioning system which are already present in the stereolithographic apparatus. Although the present system refers to "laser beam" and "X-Y galvanometer scanning system" it is apparent that these advantages also apply to other possible systems with different energy sources or positioning means or combinations of these.

The present invention also provides a related new and improved apparatus and method for normalizing and calibrating a stereolithographic apparatus.

In an apparatus for the production of parts by stereolithography, it is useful to have an apparatus and a method for normalizing and calibrating the projection of the reaction means (in a preferred embodiment, a laser beam) upon a working medium (in a preferred embodiment, a designated surface of a photopolymer). Accurate positioning of the reaction means upon the working medium is made possible according to the present invention by providing at least one sensor capable of being positioned at a predetermined location in a plane defined by the designated surface of the working medium. The sensor is sensitive to the presence of the reaction means. Memory means are provided to store information which includes the positioning information which causes the reaction means to be pointed accurately at the sensor. In a preferred embodiment, a memory look-up table or map is prepared 1 334052 containing specific positioning information for each of a number of specific locations on the surface of the working medium. A standard linear interpolation technique is utilized to determine positioning information useful to point the reaction means at points intermediate to those in the look-up table.

Accordingly it is an object of this invention to provide an apparatus and method for accurately calibrating and normalizing a stereolithographic apparatus.

lo It is a further object of this invention to provide an improved and more accurate method and apparatus for pointing a beam.

The present invention also provides a related new and improved apparatus and method for correcting for drift of a mirror positioning system so as to provide improved accuracy and repeatability.

In a stereolithographic device, a three dimensional object or "part" may be produced from a medium capable of solidification upon exposure to a reaction means operating in a preselected manner upon a designated surface of the medium. In this manner successive adjacent laminae are built up of solidified medium to form the part.

Thermal effects, wear effects and other effects can cause the positioning means for positioning the reaction means to "drift" so that the reaction means can no longer be repeatably directed to the same location. In order to alleviate such drift errors, the present invention utilizes at least one sensor means capable of detecting application of the reaction means to the sensor means, said sensor means being mounted in a predetermined location fixed with respect to the designated surface of the medium. Positioning means position and apply the reaction means and can apply the reaction means to said sensor means. Current sensor location output means provide current sensor location information indicating the location of the sensor means when the reaction means is applied to the sensor means. Likewise a memory means is 7 1 334052 75358-4 utilized to contain past sensor location information which includes a prior apparent location of said sensor means. Means for comparing said current sensor location information and said past sensor location information is provided and is capable of determining a drift correction for use in adjusting the positioning means so as to eliminate drift error effects in the pointing of the reaction means.

In summary, one exemplary aspect of the present invention provides a method for calibrating and normalizing a stereolithography apparatus for the production of a threedimensional object from a medium capable of selective physical transformation when exposed to a reaction means operating in a prescribed manner upon a designated working surface of the medium defining a working region of the apparatus to form successive laminae of the object, the improvement comprising the steps of:

defining a plurality of known locations near the working surface with known spatial relationships between each location; directing the reaction means using desired positioner information to a plurality of locations near said working surface; recording positioner information used to direct the reaction means; sensing positions of the plurality of locations to which the reaction means is directed; using the sensed positions, the known locations, the known spatial relationships and the recorded positioner information to calibrate and normalize the apparatus.

According to another exemplary aspect of the invention, there is provided in a stereolithography machine for the production of a three-dimensional object from a medium capable of - 7a 1 3 3 4 0 5 2 75358-4 selective physical transformation when exposed to a reaction means operating in a prescribed manner upon a designated working surface of the medium defining a working region of the machine to form successive laminae of the object, the improvement of an apparatus for calibrating and normalizing the machine prior to operation of the machine to make the three-dimensional object, comprising:

means for defining a number of known locations near the working surface with desired spatial relationships between each location;

means for directing the reaction means using desired positioner information to a plurality of locations near said working surface;

means for storing positioner information used to direct the reaction means; means for sensing positions of the plurality of locations to which the reaction means is directed; and means for using the sensed position, the known locations, the desired spatial relationships and the recorded positioner information to calibrate and normalize the apparatus.

According to yet another exemplary aspect of the invention, there is provided in a machine for the production of a three-dimensional object from a medium capable of selective physical transformation when exposed to a reaction means operating in a prescribed manner upon a designated working surface of the medium defining a working region of the machine to form successive laminae of the object, the improvement of an apparatus for calibrating and normalizing the machine prior to operation of the machine to make the three-dimensional object, comprising: at least one sensor located at a plurality of predetermined locations within the working region prior to forming the three-dimensional 7b 1 334d~3~-4 object, for sensing when the reaction means is pointed at each predetermined location; a positioner for positioning the reaction means in response to positioner information; a first memory for receiving selected positioner information and corresponding predetermined location information for each of the plurality of predetermined locations, the selected positioner information for a predetermined location corresponding to the positioner information when the positioner directs the reaction means on to the at least one sensor when the at least one sensor is located at the predetermined location; a second memory for receiving data representing a laminae of the three-dimensional object to be formed wherein the data represents desired spatial relationships between a plurality of points to be transformed; and a processor connected to the at least one sensor, the positioner, and the first and second memories and programmed to use the selected positioner information and the corresponding predetermined location information to map the data which represents desired spatial relationships between the plurality of points to be exposed to calibrate the positioner information to accurately position the reaction means on the working surface.

According to a further exemplary aspect of the invention, there is provided a method for calibrating and normalizing an apparatus for the production of a three-dimensional object from a medium capable of selective physical transformation when exposed to a reaction means operating in a prescribed manner upon a designated working surface of the medium defining a working region of the machine to form successive laminae of the object, - 7c 1 3 3 4 0 5 2 75358-4 the improvement comprising the steps of: temporarily locating at least one sensor at a plurality of predetermined locations within the working region prior to forming the three-dimensional object to sense when the reaction means is pointed at each predetermined location; positioning the reaction means in response to positioner information; receiving in a first memory selected positioner information and corresponding predetermined location information for each of the plurality of predetermined locations, the selected positioner information for a predetermined location corresponding to the positioner information when the positioner directs the reaction means on to the at least one sensor when the at least one sensor is located at the predetermined location; receiving in a second memory data representing a laminae of the three-dimensional object to be formed wherein the data represents desired spatial relationships between a plurality of points to be transformed; and using selected positioner information and the corresponding predetermined location information to map the data which represents desired spatial relationships between the plurality of points to be exposed to calibrate the positioner information to accurately position the reaction means on the working surface.

The above and other objects and advantages of this invention will be apparent from the following more detailed description when taken in conjunction with the accompanying drawings of illustrative embodiments.

Brief Description of the Drawinqs Figures 1, 2 and 3 are flow charts illustrating the basic concepts employed in practising the method of 1 334052 _- 7d 75358-4 stereolithography;

Figure 4 is a combined block diagram, schematic and elevational sectional view of a stereolithography system;

Figure 5 is a block diagram of a stereolithography system;

Figures 6 and 7 are exploded perspective views of the major component groups in a stereolithography system;

Figure 8 is a perspective view of the laser and optical system in a stereolithography system which utilizes the preferred embodiment of the present invention;

Figure 9A is a cross-sectional schematic view of a beam profiler sensor of the preferred embodiment of the present invention;

Figure 9B is a top plan view of a pinhole plate for a beam profiler sensor of the preferred embodiment of the present invention;

Figure 10 is a block diagram showing the apparatus of a preferred embodiment of the present invention;

Figure llA is a functional block diagram of the preferred embodiment of the method of generating an 1 334052 `_ intensity profile of a beam according to the present invention;

FIG. llB is a functional block diagram of a method of moving a beam in carrying out the method described in FIGURE llA;

FIG. llC is a functional block diagram of a method of reading the intensity of a portion of a beam in carrying out the method described in FIG. llA;

FIG. 12 is a functional block diagram showing process and analyses which may be coupled with the method described in FIG. llA;

FIG. 13 is a chart showing a sample intensity profile for a beam generated by a preferred embodiment of the invention;

FIG. 14 shows predicted cure depth profiles along two axes generated from beam profile information from a preferred embodiment of the present invention;

FIG. 15 depicts a cross-section of a cured photopolymer trace caused by exposure to a beam;

FIG. 16A depicts a perspective view of a calibration plate; and FIG. 16B depicts a cross-sectional elevation of a calibration plate.

Description of the Preferred Embodiment The stereolithographic system with which the apparatus and method of the preferred embodiment of the present invention is used generates three-dimensional objects by creating a crosssectional pattern of the object to be formed at a selected surface of a fluid medium, e.g., a W curable liquid or the like, capable of altering its physical state in response to appropriate synergistic stimulation such as impinging radiation beams, or electron or other particle beam bombardment. Successive adjacent laminae, representing corresponding successive adjacent crosssections of the object, are automatically formed and integrated together to provide a step-wise laminar or thin 1 334052 layer buildup of the object, whereby a three-dimensional object is formed and drawn from a substantially planar or sheet-like surface of the fluid medium during the forming process. The technique is generally described in the flow charts and diagrams of FIGS. 1-5.

A stereolithographic system is shown in elevational crosssection in FIG. 4. A container 21 is filled with a W curable liquid 22 or the like, to provide a designated working surface 23. A programmable source of ultraviolet light 26 or the like produces a spot of ultraviolet light 27 in the plane of surface 23. The spot 27 is movable across the surface 23 by the motion of mirrors or other optical or mechanical elements (not shown in FIG. 4) used with the light source 26. The position of the spot 27 on surface 23 is controlled by a computer control system 28.

The system 28 may be under control of CAD data produced by a generator 20 in a CAD design system or the like and directed in PHIGS format or its equivalent to a computerized conversion system 21 where information defining the object is specially processed to reduce stress, curl and distortion, and increase resolution, strength and accuracy of reproduction.

A movable elevator platform 29 inside container 21 can be moved up and down selectively, the position of the platform being controlled by the system 28. As the device operates, it produces a three-dimensional object 30 by step-wise buildup of integrated laminae such as 3Oa, 30b, 30c.

The surface of the W curable liquid 22 is maintained at a constant level in the container 21, and the spot of W light 27, or other suitable form of reactive stimulation, of sufficient intensity to cure the liquid and convert it to a solid material is moved across the working surface 23 in a programmed manner. As the liquid 22 cures and solid material forms, the elevator platform 29 that was initially just below surface 23 is moved down from the surface in a programmed manner by any suitable 1 334052 actuator. In this way, the solid material that was initially formed is taken below surface 23 and new liquid 22 flows across the surface 23. A portion of this new liquid is, in turn, converted to solid material by the -5 programmed W light spot 27, and the new material adhesively connects to the material below it. This process is continued until the entire three-dimensional object 30 is formed. The object 30 is then removed from the container 21, and the apparatus is ready to produce another object. Another object can then be produced, or some new object can bè made by changing the program in the ~ computer 28.

The light source 26 of a stereolithog-aphy system according to a preferred embodiment of the invention is typically a helium-cadmium ultraviolet laser such as the Model 4240-N HeCd Multimode Laser, made by Liconix of Sunnyvale, California.

A commercial stereolithography system will have additional components and subsystems besides those previously shown in connection with the schematically depicted systems of FIGS. 1-5. For example, the commercial system would also have a frame and housing, and a control panel. It should have means to shield the operator from excess W and visible light, and it may also have means to allow viewing of the object 30 while it is being formed.

Commercial units will provide safety means for controlling ozone and noxious fumes, as well as conventional high voltage safety protection and interlocks. Some commercial units will also have means to effectively shield the sensitive electronics from electronic noise sources.

The commercialized SLA is a self-contained system that interfaces directly with the user's CAD system. A commercialized SLA containing the preferred embodiment of the apparatus of the present invention, as shown in FIGS.

6 and 7, consists of four major component groups: the SLICE computer terminal, the electronic cabinet assembly, -- I 334052 11 the optics assembly, and the chamber assembly. A block diagram of the commercialized SLA is shown in FIG. 5.

The electronic cabinet assembly includes the process computer, keyboard, monitor, power supplies, ac power distribution panel, and control panel. The computer assembly includes plug-in circuit boards for control of the terminal, high-speed scanner mirrors, and vertical (z-stage) elevator. Power supplies for the laser, dynamic mirrors, and elevator motor are mounted in the lower portion of the cabinet.

The control ~ panel includes a power on ~ switch/indicator, a chamber light switch/indicator, a laser on indicator, and a shutter open indicator.

Operation and maintenance parameters, including fault diagnostics and laser performance information, are also typically displayed on the monitor.

Operation is controlled by keyboard entries.

Work surfaces around the keyboard and computer are covered with formica or the like for easy cleaning and long wear.

Turning to FIG. 8, the helium cadmium (HeCd) laser 100 and optical components are mounted on top of the electronic cabinet and chamber assembly 102.

The laser and optics plate may be accessed for service by removing separate covers.

For safety reasons, a special tool is required to unlock the cover fasteners and interlock switches are activated when the covers are removed.

The interlocks activate a solenoid-controlled shutter to block the laser beam when either cover is removed.

As shown in FIG. 8, the shutter assembly 104, two ninety degree beam-turning mirrors 106, 108, a beam expander 110, an X-Y scanning mirror assembly 112, and precision optical window 114 are mounted on the optics plate. The rotary solenoid-actuated shutters are installed at the laser output and rotate to block the beam when a safety interlock is opened. The ninety degree beam-turning mirrors 106, 108 reflect the laser beam to the next optical component. The beam expander 110 1 334052 12 enlarges and focuses the laser beam on the liquid surface.

The high speed scanning mirrors direct the laser beam to trace vectors on the resin surface. The 2- mirror, 2-axis galvanometer scan heads sold by General Scanning Inc. of Watertown, Massachusetts have been found to be satisfactory for this purpose and in a preferred embodiment their Model DX-2005 servo and Model XY-0507 galvanometer X-Y scanning heads are used. A quartz window 114 between the optics enclosure and reaction chamber allows the laser beam to pass into the reaction chamber, but otherwise isolates the two regions.

The chamber assembly contains an environmentally controlled chamber, which houses a platform, resin vat, elevator, and beam profilers.

The chamber in which the object is formed is designed for operator security and to ensure uniform operating conditions.

The chamber may be heated to approximately 40C (104F) and the air is circulated and filtered. An overhead light illuminates the reaction vat and work surfaces. An interlock on the access door activates a shutter to block the laser beam when opened.

The resin vat is designed to minimize handling of the resin.

It is typically installed in the chamber on guides which align it with the elevator and platform.

The object is formed on a platform attached to the vertical axis elevator, or Z-stage.

The platform is immersed in the resin vat and it is adjusted incrementally downward while the object is being formed.

To remove the formed part, it is raised to a position above the vat.

The platform is then disconnected from the eleva~or and removed from the chamber for post processing.

Handling trays are usually provided to catch dripping resin.

Two beam profiler sensors 116, 118 according to the preferred embodiment of the present invention are mounted at the sides of the resin vat and the focal point of the laser optical system is adjusted to match radial sensor positions (i.e., they are mounted at a radial distance ~ 13 1 334052 from the galvanometer scanners equal to the distance from the galvanometers to a point 0.3 inches below the surface of the liquid). (See FIG. 7). The scanning mirror is periodically commanded to direct the laser beam onto the beam profiler sensors, which measure the beam intensity profile. The data may be displayed on the terminal, either as a profile with representation of intensity values or as a single number representing the overall - (integrated) beam intensity. This information is used to determine whether the mirrors should be cleaned and aligned, whether the laser should be serviced, whether the scanner~mirrors have drifted, and what parameter values will yield cured vectors of the desired thickness and width.

The beam profiler sensors 116, 118 are symmetrically placed relative to the center of the resin vat. (See FIG.

7).

They should preferably have similar X and Y offsets as measured from the vat center (of opposite values) although this is not required; that is to say, they are on a diagonal of the stereolithographic apparatus. In FIG. 7, the beam profiler sensors 116, 118 are seen in the corners of the chamber assembly. The distance from the second scanning mirror on the optical plate above the chamber assembly to each beam profiler sensor aperture is the focal length which equals the desired liquid to scanning mirror length plus a small increment. In the SLA-l sold by 3D Systems, Inc. (which is shown in FIGS. 6 and 7), this liquid to scanning mirror length is approximately 27 inches and the small increment is an additional 0.3 inches, therefore the focal length is approximately ~7.3".

The distancing of the beam profiler sensors 116, 118 from the second mirror by the desired focal leng~h has the effect of detecting the best average focus for the designated surface of the photopolymer in the SLA-l resin vat. At the center of the resin vat, when the photopolymer is at the desired level, the focal length of the laser beam will be 0.3 inches below the surface of the 14 l 334052 photopolymer. The focus of the beam at the surface of the photopolymer in the center of the resin vat will not vary much. At the corner of the 12 inch resin vat of the SLA-l, the focal length will be approximately 1 inch above the surface of the photopolymer. The focal length will be at the surface of the photopolymer at a circle with a 4.2 inch radius about the center of the surface of the photopolymer. The placement of the beam profiler sensor at the focal length is intended to obtain an optimal beam profile considering that the surface of the photopolymer will mostly not be at the focal length of the laser.

FIG. 9A is a cross-sectional view of a beam profiler sensor 35 of a preferred embodiment of the apparatus of the present invention and FIG. 9B is a top plan view of a pinhole plate used in the beam profiler sensor. The beam profiler sensor has a thin stainless steel metal plate 40 having four etched pinholes 45 of varying size. In a preferred embodiment these holes have diameters 0.0005", 0.001", 0.002", and 0.004". The pinholes each permit a small fraction of the laser beam 50 incident upon the pinhole to fall on a photodetector 55 underneath the plate 40. The purpose for providing several pinholes is to permit profiling of beams having a wide range of incident power. One of the pinholes will be best suited for 2~ measuring the intensity profile of a beam of a given incident power. For the HeCd lasers used in the SLA-l, a pinhole of 2 mil (0.002 inch) diameter has been found to be satisfactory. The beam is scanned across a selected pinhole in an X Y array to build up a two dimensional profile of the beam intensity.

As may be seen in FIGS. 7 and especially 9A, the beam profiler sensor 35 has a two part housing 60. Light beam 50 enters from the right and moves toward the left in FIG.

9A. The beam profiler sensor is mounted in the corners of the chamber assembly compartment in such a way as to prevent the resin vat from hitting the beam profiler 60724--1890 1 334052 sensor when it is moved into and out of the compartment (see FIG. 7).

Turning to FIG. gA the beam profiler sensor 35 comprises a split two part housing 60, pinhole plate 40, ultra-violet transmitting filter 70 that absorb~ visible light and prevents spurious readings due to visible light.

Filter 70 is a two millimeter thickness of Schott UG-ll* filter glass which has been found to be acceptable for this purpose in a preferred embodiment. The characteristics of this filter provide reasonable transmission of light in the 300-370 nanometer wavelength region with considerably less tranemittance at other wavelengths. A one millimeter thic~neF~ of HOYA U-350* filter material would also be acceptable.

Underneath the filter 70 in the beam profiler housing is a photodiode sensor 55 which detects the ultraviolet light which passes through the filter 70 from the pinhole 45. An EEG Vactec*VTS 3072 super blue enhanced photodiode has been found to be acceptable. The output from this photodiode is passed to a current to voltage amplifier (not shown). An OP07 current to voltage amplifier whose implementation i6 well known to those skilled in the art has been found to be acceptable.

The pinhole plate 40 of the beam profiler sensor 35 is covered with a quartz filter (not shown). The quartz filter is cleanable and protects the beam profiler sensor from dust and photopolymer drips. The quartz filter should be coated to prevent internal reflections when the photoeencor is not perpendicular to the beam, in order to prevent false shape measurements. Optionally a diffuser (not shown) can be used between the pinholes to aid the filter to protect the optical components from damage.

FIG. 10 is a block diagram showing the apparatus of the preferred embodiment of the present invention.

Fundamental to the invention is a control and analysis computer. This computer réceives inputs from a program, a keyboard or the like and may display results through a *Trade mar~ 16 1 334052 printer or terminal or the like. The control and analysis computer sends positioning commands to a mirror positioning system which controls the X Y scanner mirrors.

The laser beam is focused by the optics shown in FIG. 8 to reach the X Y scanner mirrors and is directed by those mirrors to one of the beam profiler sensors. The use of two beam profiler sensors is recommended for the purpose of drift correction. The sensor signal from the beam profiler sensors is converted to a signal readable by the computer which is then sent back to the control and - analysis computer to be manipulated as described hereafter.

In physical terms, the beam profile method according to the present invention causes the beam to be moved to each of the points of an array on the pinhole plate centered on the best known position of the pinhole. As a result, different sectors of the beam will fall on the pinhole and will be transmitted through the same to be detected by the photodiode and converted into a numerical signal that can be analyzed by the computer. A profile of the intensity of different sectors of the beam will be built up by the computer (see FIG. 13). This is the "intensity profile" of the beam.

FIG. llA is a functional block diagram showing how a beam profile is developed according to the preferred embodiment of the invention. The best known location of a pinhole on the beam profiler sensor is called up from memory by the control and analysis computer and sent to the mirror positioning system to position the X Y scanner mirrors to direct the beam at this best known location.

The control and analysis computer, through the beam positioning system, moves the beam to the first row in the first column of a square array centered on the best known location. The intensity of the portion of the beam entering the pinhole as detected by the beam profiler sensor is then read and is saved as well as the mirror position commands which are associated with that 17 1 334052 intensity. The beam is then moved in sequence from the first to the last array points on a particular row or column and the reading and saving intensity value steps are repeated. The beam is then moved in sequence from the first to the last array columns or rows and the moving and reading steps are then performed down each column or row.

The result is that beam intensity readings are taken for each of the positions on the array (a "position" is known to the computer as a set of mirror positioning commands).

A standard analysis, performed by the control and analysis computer of the array of intensity values is usually performed to generate a new best known location of the pinhole (for use in performing the first step of the scan profile the next time) regardless of the detailed function actually being analyzed (see FIG. 12). This calculated best known position can be found with great precision by this means, to an accuracy much finer than the size of the pinhole. When an appropriate number of locations have been found and stored, the control system can use these values with two-axis linear interpolation to move the beam to those or any intermediate locations.

FIG. llB is a functional block diagram of a method of moving a beam and carrying out the method described in connection with FIG. lOA. To move the beam, the first step is to send the beam positioning information to the servomechanism of the X Y scanner mirrors concerning the desired location. The servomechanism (which may be analog or digital) then send a signal to the mirror drivers to position the X Y scanner mirrors to a new location The servomechanism of the X Y scanner mirrors measures the actual position of the mirror drivers and compares the actual position with the intended position and adjusts the drive signals accordingly.

The adjustments are continued within specification values of the intended location.

FIG. llC is a functional block diagram of a method of reading the intensity of a portion of a beam and carrying out the method of the preferred embodiment of the present 1 334052 18 invention. The first step is to convert the total amount of light which passes through the pinhole into a signal which is proportional to that amount of light. In the preferred embodiment, this process is performed by the photodiode which measures the light coming through the pinhole and the filter. The current from the photodiode is sent to a current to voltage amplifier which generates a signal proportional to the amount of light received by the photodiode. Amplification of the signal (which will be proportional to intensity) to provide wide dynamic range of the measurement is important to obtain small but significant readings for the edge of the beam which would otherwise be lost. The next step is to measure the signal, which is proportional to the amount of light received, after the signal is converted into a digital form for numerical analysis.

FIG. 12 is a functional block diagram showing the processes and analyses which may be coupled with the method described in FIG. llA. As may be observed from that figure, a number of different processes and analyses may be selected from a menu, the first five of which are connected to the scan profile routine of FIG. llA. The first step is to scan the intensity profile of the beam according to the method described in connection with FIG.

llA. The intensity profile may be displayed numerically or in the form of a graph. As an option, the power may be calculated from the intensity profile as well as a new best known location of the pinhole used. Another possible process is to add the data generated in connection with the beam intensity profile to a history file with an option of displaying the history file. A further possible process is to calculate and display drift information for the mirror positioning system which generally involves scanning a second separate sensor (in the case of the preferred embodiment, another beam profiler sensor) and then calculating and displaying the offset and gain terms of the drift.

Another process is to calculate and display 1 334052 19 the power of the beam, which involves summing up the intensities of a profile and multiplying by a power conversion factor. The power conversion factor can be determined, for example, by utilizing the process with a S beam of known power or by comparing calculated power to that of a calibrated sensor and determining the required gain factor. A further function is to calculate and display focus information, with a possible option of using a special transformation of the intensity data used to calculate focus information and the use of known resin properties to predict the shape and size of the cured traces of photopolymer. Another possible function is to move the beam to the desired location for set up (to make parts), tests and so forth with an option for searching for the sensor or scan profile from this new location. A useful function is to search for sensor pinholes in a spiral move-read format. This may be necessary if the best known location of the pinhole is not accurate, in case when the array is traced over the best known location of the pinhole is not detected. A further step may be to verify findings (of pinhole location) by test or scan profile. A still further function is the use of the beam profilers in calibration, which involves the measurement of the drift while obtaining a calibration map for the 2~ surface corresponding to the surface of the photopolymer.

A final function is to store in the machine the parameters of best known location, scale factors, resin properties, and so forth.

FIG. 13 is a chart showing a sample intensity profile for a laser beam as generated by a preferred embodiment of the present invention.

The numerical values correspond to the measured intensity of a beam from a preferred embodiment of the present invention. The numbers have been converted to integers for ease of reading the display.

The intensity profile generated by the present invention may be used to calculate the power of the beam 1 334052 and to predict the shape and dimension of a cured trace of photopolymer (the solidified photopolymer due to exposure to the beam of UV light), as the following discussion will demonstrate.

The intensity of the beam is measured by the beam profiling apparatus when the beam is directed at each point of an array on the pinhole plate, which is a surface generally perpendicular to the optical path taken by the beam.

The X and Y directions on this surface correspond to the directions taken by the beam when one or the other of the scanning mirrors rotate.

The X and Y array coordinates are 1 to i~x and 1 to j~x, respectively (typically i~x and i~x are each equal to 22).

The beam is typically stepped or moved from point to point in the array, the time for the movement being much less than the time the beam remains at a point. The distance between points is:

s (mm) = scanstep / scalefactor [Eq. 1] The scanstep is typically 4 "bits" and the scale factor is normally 140 bits/mm. The scanning mirrors can each take on 65535 (64K) different positions over a 40~ optical beam rotation. This in turn means that 1 bit along the X or Y axis corresponds to a rotation of 6.104 x 10-4 degrees.

Since mirror to liquid distance is approximately 27" this angular rotation corresponds to a translation at the liquid surface of 2.875 X 10-4 inches or equivalently, 137 bits/mm cr approximately 140 bits/mm.

The array area must cover the full beam (measurement of the entire beam is needed for power calibration as well as to produce the maximum information concerning the beam), and must have a number of points sufficient to resolve the desired beam profile. Typically, the spacing between points in this array is less than one tenth of the beam width. The pinhole diameter should be smaller than this resolution limit.

21 1 334052 An "element" is the portion of a beam measured when the beam is directed at a point (m,n) of the array. Each element (m,n) has an intensity reading I(m,n). The letters m,n refer to positions or points in the X, Y directions in the array, respectively. FIG. 13 shows intensity readings in an array much as discussed here.

The beam power is measured independently and the power calibration factor k is derived from the following equation:

k * ~ I(m,n) = Power (watts) m,n [Eq. 2] The power calibration factor k only applies to the selected pinhole and measuring system and laser wavelength. The independent power measurement must be made on the beam after the beam has traversed the same number of optical surfaces in the beam path.

These calculations also assume that the background light signals have been eliminated and amplifier scaling compensated for.

The power/unit area at element (m,n) is given by:

Intensity (at element m,n) = k * I(m,n) / 52 (watts/mm2) [Eq. 3] This is the instantaneous intensity experienced by a small area in the (m,n) element, regardless of whether the beam is stationary or moving.

When the beam is moving uniformly at the speed v (mm/sec) along the Y axis, then each element takes a time equal to s/v to pass, and the exposure energy absorbed per unit area from the element (m,n) is:

Exposure from element (m,n) = k * I (m,n) / 52~.

( s / v) (Joules/mm) [Eg. 4] This is the energy absorbed per unit area from a particular element (m,n) of the beam.

The total exposure (beam energy absorbed) as the entire beam passes an area equivalent in size to an element, as defined above, is:

22 l 334052 Exposure at m = k * ~ I(m,n) / s2) * (s / v) (Joules/mm2) [Eg. 5] Physically speaking, this equation represents a situation in which the beam traverses in the Y direction over an area equivalent in size to an element of the beam as the term element is used above.

The area is traversed by the elements of the beam corresponding to the X coordinate m, so the element-sized area at m is exposed to all elements (m,n) of the beam as n varies between O and i~x~ The calculations described above are based on - discrete elements, but could obviously be generalized to use integrals.

The movement is assumed along the Y axis for convenience. Other angles can be simply derived, and may be necessary if the beam is asymmetrical.

The speed v corresponds to parameters SS and SP as follows:

v = (SS / ScaleFactor) (SP / 100, 000) (mm/sec) [Eq. 6] Where:

SS = Step Size in bits/step;

20 ScaleFactor normally is 140 bits/mm;

SP/100,000 = Step Period in seconds (SP units are units equal to 10 microsec); and lE6 = 1,000,000 is a conversion factor between Joules / mmZ and Joules / m2 or microJoules / mm2 Equations 5 and 6 are combined to calculate the total exposure (power incident, i.e., energy) at the sensor or at the liquid (photopolymer) surface represented by Z=O at a small area at position m as the beam is moved in the Y direction:

Exposure at (m,z=0):

E(m,z=0) = ~ I(m,n) * k * SP * ScaleFactor * lE6 n s * SS * 100,000 [Eg. 7a] = ~ I(m,n) * SP * ScaleFactor * power * 10 n uJ / mm ~ I(m,n) * SS * s [Eg. 7b] m,n ` 1 334052 23 Finally, the absorption of the beam as it penetrates into the liquid may now be compensated for in accordance with Beer's law:

E(m,z) = E(m,0) * exp (-z/lambda) [Eq. 8] 5 Where:

lambda is the penetration depth (mm);

E(m,0) is the summed exposure at the surface; and E(m,zj is the exposure at depth z below the surface (mm).

Attenuation is assumed to have no nonlinearities or time-dependence in absorption, being represented simply by I(z)=I(z=0) * exp (-z/lambda). It is clear that appropriate modifications can be made to the foregoing calculations to allow for deviations from the above simple absorptive behavior.

lS The photopolymer has been shown experimentally to cure to a gel if the exposure is greater than a critical value Ec, so for any given system the shape of a trace of cured plastic can be predicted by calculating the locus of points having exposure Ec. Ec can be accurately and separately measured for each photopolymer.

The "gel point" gives only the "cured" versus "not cured" boundary, and disregards the gradient in exposure (related to penetration depth) at resin depths other than the Ec boundary depth. Part strength seems to relate to the 2~ higher exposure, and so absorption characteristics should be chosen to give the best (highest) cure gradient. The gradient or penetration depth also limits the best available resolution in the Z direction, because some variation in exposure (crossing lines, etc.) is unavoidable and this results in cure depth changes with this variation in exposure.

For any X location (m) the cured depth zc(m) is found from:

zc(m) = lambda * ln ( E(m,z=0) / Ec ) [Eg. 9] A beam profile measured with sufficient reliability and precision can be used to predict cure depths dependent only on chemical properties of a resin. FIG. 14 shows two ~ 24 1 3 3 4 0 5 2 75358-4 examples of such predlctlons, along the X and Y axls respectlvely.

The proflle functlon (m,z) also permlts predlctlon of the trace wldth as a functlon of depth (and wlth approprlate modlflcatlons, "beam wldth" and "mlnlmum surface angle") automatlcally. Maklng and measurlng "ban~o tops," l.e., traces cured by the beam to dlrectly determlne the shape and dlmenslons of the cured photopolymer, wlll only be needed as an lnternal check of the system. FIG. 15 shows a cross-sectlon of a test trace from a ban~o top. FIG. 15 should be compared wlth FIG. 14.

To dlsplay the predlcted trace proflle, a scaled diagram of cure depth versus posltlon ls plotted across the beam. The scale factor for dlstance across the beam ls easy, where one slmply selects one column (or row, etc.) of the scan, wlth dlmenslon s, to correspond to a row plxel or graphlcs dlsplay block. The depth scale ls then lambda/s plxels (or blocks) for each lncrease ln exposure by a factor of e. The only arbltrary feature ls the zero of depth, related to the characterlstlc exposure Ec of Eg. 9 , or equlvalent factors from Eq. 7. The useful depth to be dlsplayed ls determlned by the dynamlc range of the lntenslty measurlng system, and truncatlon at I(m,n)s any approprlate value close to the nolse level.

Callbratlon and Normallzatlon As polnted out above, lt ls deslrable ln a stereollthographlc apparatus to provlde an apparatus and method for callbratlng the polntlng of a reactlon means on a worklng medlum for lmproved preclslon and accuracy.

Callbratlon procedures of a preferred embodlment of the present lnventlon allow the creatlon of a "map" from a deslgn ln a 24a 1 334052 75358-4 CAD space to drawlng lnstructlons on the real SLA. In any automated bulldlng system there wlll be - 1 334052 several different sources of error needing to be corrected by calibration procedures. The present system has a pair of scanning mirrors close together, and, if left uncorrected a simple mapping of CAD dimensions to the angles of the mirrors will result in pincushion distortion.

This is because the system is building on a flat surface where further from the point on the surface nearest to the mirrors, equal increments of angle will project to progressively greater distances on the surface.

This happens to be the dominant distortion in the present system, and is predictable from geometry so that its correction could be calculated. However there are numerous other errors and distortions which need to be compensated, and many of them are not readily predictable.

The calibration and normalization system of the present invention is useable in a broad range of applications and systems and automatically generates a "look-up table" to enable the conversion of CAD locations into instructions which are to be sent to the scanning system so that the intended pattern will be drawn on the working surface. The term "Normalization" may be used to indicate that more than one dimension at a time is being corrected, while "Calibration" could have connotations of applying a single scale factor to a system. In a preferred embodiment, the apparatus has a single position (beam profile) sensor which is moved automatically to an array of points on the working surface, and then a record is made of the corresponding mirror instructions needed to reach each of these points. In another preferred embodiment, a square array of sensor pinholes is utilized so that no movement of the sensor was required.

In yet another preferred embodiment, a linear array of pinhole sensors needing to be moved along only one axis is utilized.

Turning to FIGS. 16A and 16B the square calibration plate 200 of a preferred embodiment of the present invention is depicted.

An ultraviolet-opaque metallic 1 334052 26 coating 206 is provided by evaporation on a substrate 204 made preferably of quartz or pyrex of l/8-inch to 1/4-inch thickness. In a preferred embodiment an array of 49 by 49 holes 202 are etched in the W-opaque metallic coating 206 on 1/4-inch centers. Each etched hole has a diameter of 0.004" + 0.0005" although it is only important that the hole be of smaller diameter than the diameter of the projected beam at the plate for best resolution. Sensors (not shown) are mounted below the plate and are arranged when the calibration plate is in use to be in the precise location of the surface of the working medium.

In a presently preferred embodiment of the invention, an array of 5 x 5 or 25 photodiodes 208 sensitive to UV light are utilized in conjunction with the calibration plate. Because W light can only enter the plate through one of the pinholes and because the plate material tends to diffuse light that enters, light entering a pinhole will travel horizontally beyond the exact location of the pinhole so that in a preferred embodiment 25 sensors such as the sensors herein previously described are adequate to cover the entire 49 x 49 array of pinholes.

Typically a calibration procedure will be performed prior to shipping an SLA to a customer and after any physical trauma to the mirror control system which would tend to uncalibrate the SLA.

In operation the calibration procedure uses the same beam profiler methodology to obtain a "best location" of a pinhole from the centroid of the scan data.

A new "best location" is obtained for each pinhole in the case of the plate, each pinhole-column location in the case ~f the linear array, and each predetermined location in the case of a sensor positioned at certain predetermined locations. It is not always necessary to scan each pinhole in order to obtain a working and practically useable look-up table. In a preferred embodiment of the present invention only about 40 x 40 pinholes are located and mapped. In situations where 1 334052 27 there are smaller geometric distortions, or less precision desired, or the other sources of distortion are suitable for correction with more reliance on interpolation, fewer pinholes could be mapped.

Linear interpolation relative to the X and Y locations is used to determine mirror positioning for points intermediate the "best locations" stored in the look-up table in memory. The appropriate number of pinholes will be determined from these considerations, and from the time needed to make the calibration, and from the system memory available for storing the look-up table.

Optionally and preferably the drift correction apparatus and method described farther on are used in conjunction with calibration to obtain more precise and more accurate results. Similarly, during part building the same drift correction method and apparatus are used to improve precision and accuracy.

Briefly the method of a preferred embodiment of the present invention is set forth below:

STEP 1: User inserts calibration plate into SLA building chamber with its sensor pinholes positioned where the liquid surface normally stays; user specifies the delay times between reading sensors 1 and 2 (Fixed Beam Profile Sensors 116, 118) and the Calibration Plate (deemed "Sensor 3") locations.

STEP 2: Sensors 1 and 2 are located again as they are located when the first scan profile occurs to finalize and store their apparent coordinates in terms of mirror position information.

STEP 3: Center the calibration pl~te by determining if the center sensor 3 leading coordinates are within acceptable tolerances.

The center plate coordinates should correspond to the mirror positioning center - 1 334052 28 coordinates.

This configuration allows maximum equal mirror movements in all directions.

STEP 4: The gain of the calibration plate (Sensor 3) is set by reading the center hole of the plate (which is found with the find sensor algorithm defined in Beam Profile Fig. 1).

The actual beam intensity is derived by reading Se~sor 3 with and without the beam in the center hole.

This subtracts the background noise Sensor 3 receives.

The gain control is adjusted by the user until the sensor sensitivity is optimized.

STEP 5: The borders of the calibration plate are established by stepping the beam from plate holes to the borders of the plate (1 = West, 2 = North, 3 = South, and 4 = East).

A) Move in direction 1, locating holes along the way by moving a predetermined (bits/hole separation) value.

Bits refer to mirror coordinate change values.

B) When the known number of holes before the border are read, one more ~5 "bits/hole separation" movement is executed.

C) If a read finds a hole there, either the gain is set incorrectly because a false hole has been read, or the hole to the right of the center hole was where the search started.

Return to Step 3.

D) If no hole is detected, the left and right borders are now established.

E) The beam returns to the center hole and begins a search for the back border in a similar fashion to A-D.

`- 1 33~052 29 F) Once all borders are established through movements 1 and 2, the values of the hole locations for movements 1, 2, 3, and 4 are used to build a "rough map of hole separations" across the plate in mirror bits.

Movement 4 leaves the beam at plate hole (1,1).

STEP 6: Quick search for all plate hole locations.

If any hole cannot be located the beam is left at its most probable location, as determined by the "rough map of hole separations", and prompts the user to examine for dust around the hole location.

After the wait the find is resumed until the hole is found or the user aborts.

The gain may be reset in this process if the user desires.

If the gain is changed the user restarts the quick search at hole (1, 1) .

STEP 7: After the quick search locates all the plate holes a final search is performed with the requested delays from Step 1.

Also, on both auick and final searches the locations of sensors 1 and 2 are found to 2 determine GAIN and OFFSET corrections to be made at the intervals (end of each row) of mirror bit movements to the same locations.

These correction factors are applied proportionately to correct each of the calibration locations in a manner normalized to a single set of reference locations of the fixed sensors 1 and 2.

STEP 8: After the final search stores all the gain, offset, intensity, and locations date the procedure ends.

7 334052 Drift Correction Drift correction is a procedure which compensates for drift of, inter alia, the mirror positioning system by periodically checking on the apparent position of one or more beam profile sensors (herein referred to as "sensor means").

Measurements of changes in the apparent position of a single sensor allow for compensation for drift of the "zero setting" of the mirror system, two separate sensors allow additionally for correcting for otherwise uncompensated gain in the system and/or otherwise - uncompensated changes in size of parts of the SLA due to thermal and other effects. Other errors can be corrected by utilization of even more sensors although in a preferred embodiment of the present invention two beam profile sensors are deemed sufficient.

In a preferred embodiment of the present invention, a calibration operation is periodically performed. In the calibration procedure, a plate having a number of holes and sensors is utilized to generate in system memory a table of mirror position settings which correspond to fixed predetermined locations on the plate.

During the time a calibration is being run, the system periodically checks on the apparent locations of two sensors.

These measurements are used to correct the - 25 calibration measurements for this amount of drift, so that the values are all normalized to a "standard" pair of apparent positions of the two sensors. When later building a part, the same two sensors are again scanned periodically, and the apparent positions can be used to correct for changes in the zero and gain of the mirror system relative to the time at which the calibration was made.

This procedure has been found to remove 90% of the error caused by drift of the mirrors. A description of the calibration procedure is provided earlier.

In the drift compensation method and apparatus of a preferred embodiment of the present invention two beam profile sensors capable of detecting when the laser beam ~ 1 334052 31 is directed toward them by the mirror positioning system are mounted fixedly in predetermined locations fixed with respect to a designated surface of the solidifiable working medium upon which the reaction means impinges to solidify said medium.

Periodically the laser beam is directed toward the sensors, and sensor location output means provide a readout of the apparent location of the sensors. The current apparent location of the sensors are compared with a past apparent location of the sensors which is stored in memory and a difference indicates the need for a drift correction.

For example, when a single sensor No. 1 is in use, this sensor No. 1 had a past apparent position of X = 20, Y = 20 and a current apparent position of X = 22, Y - 22 then a drift of +2X and +2Y has occurred and the mirror positioning system can apply an appropriate correction factor in order to point the beam in the desired location.

As another example, when in addition to this first sensor, a second sensor is used, then the second sensor might be read at calibration to be X = 64000, Y = 64000 and have a current apparent position of X = 64004, Y = 64003.

Were this the case, in addition to a linear shift of +2X, +2Y across the entire system (the two sensors conveniently being located on a diagonal) there would also be a gain or stretching of +2X, +lY of the apparent distance between sensors No. 1 and No. 2 and we expect and correct for proportionately different stretching for different locations relative to the sensor No. 1.

In compensating, linear interpolation could be used to help the mirror positioning system compensate for the gain term in the drift error.

To achieve more accurate part-building using stereolithography, a special computer software algorithm was developed. This algorithm, known as the Drift Correction algorithm, can be considered having two parts.

- 1 334052 32 The first part is code that determines the value of the variable used in drift correction.

The second part is code for the application of these variables to correct for drift.

The drift correction variables are:

* the X-axis gain called DriftGainX in our examples * the X-axis offset DriftOffsetX * the Y-axis gain DriftGainY * the Y-axis offset DriftOffsetY - The variables are used to compute a correction factor that is applied to all X,Y coordinate positions of the laser mirrors.

This correction is applied to each axis separately and has a gain (multiplier) and offset (addition) term for each axis.

It is in the form:

corrected axis value = (gain * desired axis distance from first sensor) + offset of first sensor where the gain and offset variables map the ideal desired axis position to the corrected real-world position based on measurements.

The drift correction variables are determined at the start of each layer through the use of Beam Profiling algorithms very similar to those found in the Beam program. The position of two beam profile sensors at opposite corners of the vat is measured and calculated and compared to the ideal positions remembered in a disk data file. The differences between these two sets of profile sensors positions determines the value of the drift correction variables. Placement of the two profile sensors at the corners of the scanning field optimizes the baseline along both X and Y axes from which to measure changes in gain.

The Pascal function below is a sample for computing the values of the drift correction variables. The 'Ref' variables are the reference positions of the two beam profile sensors read from the disk data file. The 'Now' variables are the most recently determined positions of 1 334052 33 the two profile sensors.

Because of dynamic changes in the physical characteristics of the laser scanning system, the positions held in the 'Now' variables will differ slightly from the ideal reference positions. The goal is to use these position differences to adjust upcoming mirror positions.

{ CorrectForDrift -- compute value of drift correction variables - procedure CorrectForDrift;

var quotient: Float Type:

begin DriftGainx := (Sensor2_NowX - Sensorl_NowX) / (Sensor2_RefX - Sensorl_RefX;) DriftOffsetX := Sensorl_NowX;

DriftgainY (Sensor2_NowY - Sensorl_NowY)/ (Sensor2_RefY - Sensorl_RefY;) DriftOffsetY := Sensorl_NowY;

end;

The Pascal procedure below shows how a coordinate pair can be corrected for drift once the drift correction variables have been determined.

Note,that the correction for each ~5 axis is separate from the other.

{ DriftCorrect -- correct coordinate pair for drift } procedure DriftCorrect(oldX,oldY: FloatType;varnewX,newY:

FloatType);

begin newX := (oldX - Sensorl_RefX) * DriftGainX + DriftOffsetX; newY := (oldY - Sensorl_RefY) * DriftgainY + DriftOffsetY;

end;

The DriftCorrect procedure is used to correct the start and the end of each drawing vector before it is 1 334~ 75358-4 converted to a "feature" which is a highly optimized form of a laser drawing used within 3D Systems software. The CorrectForDrift is used only once per layer, after the new position of the two beam profilers has been determined.

In practice, these routines are located within a special memory resident driver (known as STERE0) that different stereolithography application programs can use. The application program is responsible for finding the profiler location, and it passes this information along to the special device driver.

BUILD STERE0 an applications program the memory-resident used to build parts; it --~ driver that performs passes along profile several Stereolithoposition info for each graphy related tasks, layer one of which is drift correction An equivalent drift correction procedure is performed on all of the calibration points during the calibration of the SLA.

In a presently preferred embodiment, drift is determined at the end of a scan of each row of predetermined fixed positions and the correction is applied through linear interpolation for that row.

A method of actual software implementation of such a routine for compensation would be well within the capabilities of one of ordinary skill in the art and therefore will not be further described herein.

Appendices A, B and C attached hereto are software code listings pertaining to beam profiling, drift correction and calibration.

It will be apparent from the foregoing that, while particular forms of the invention have been illustrated and described, various modifications can be made without departing from the spirit and scope of the invention. Accordingly, it is not intended that the invention be limited, except as by the appended claims.

34 APPENDIX A I To compile: use MKSTEREO Batch File } 1 3 3 4 0 5 2 { STEREO, Sterolithography Device Driver for 3D Systems' Sterolithography System Version 3.16 by Tarnz Technologies STEREO is a memory-resident laser controller device driver. It interfaces between the Background Task's Software Interrupt Handler and a program wanting to control the laser apparatus using easy text-based commands.

STEREO receives and processes simple (X,Y) laser position c~ n~c passed to it by some other program using a software interrupt vector.

The CLIFF Turbo Pascal program is just such a program; it can read ,_ n~C from either the computer's keyboard or from disk files and then pass them along to this program.

STEREO, like an MSDOS driver, can be called by programs written in any language, including assembly and BASIC. The interface will usually be a special subroutine that invokes the STERO software interrupt. Unlike an MSDOS driver, STEREO cannot be referenced directly by a filename.

STEREO's device driver format is simpler, faster, and more clean than an MSDOS device driver. STEREO processed a complete text string (a line~ at a time, as opposed to one character at a time for an MSDOS device driver. If desired, an official MSDOS device driver (STL:, LSR: or whatever) could be added in later.

STEREO understands the following c n~c:

BX 0-65535 Beginning X position (position) BY 0-65535 Beginning Y position (position) JX 0-65535 Jump to new X position (position) JY 0-65535 Jump to new Y position (position) NX 0-65535 Next X position (position) NY 0-65535 Next Y position (position) SS 1-65535 Step Size (rel incr) ! SP 5-6553,.. multi-pass Step Periods (10 us) SD 0-65535 Scanning Delay (step periods) JP 5-6553 Jump Period (10 us) JS 1-65535 Jump Size (rel incr) JD 0-65535 Jump Delay (step periods) LO 0-65535 Laser On delay (10 us) LF 0-65535 Laser Off delay (10 us) + NS eNter Sequence mode + AB Enter Absolute mode EX Execute EC Execute and Clear CL Clear table ! MD AS +/- set Auto-Shutter mode (on/off) ! MD BL +/- set BLock mode (on/off) ! MD PA +/- set Pos Ack mode (on/off) ! SR Soft Reset ! HR Hard Reset ! RC 1-10 Redraw pass Count (# passes) ~ 3~ 1 334052 ~ ! RD 0-65535 Redraw Delay (step periods) ! RS 0-65535 Redraw Size (distance) ! VC 1-10 riVet pass Count (# passes) ! VP 5-6553,... riVet step Periods (10 us) ! VR 0-65535 riVet Reduction amount (distance) ! VX 0-65535 riVet end X position (position) ! VY 0-65535 riVet end Y position (position) ! WI 0-32767 Wait for Idle (delay) - New Version 2.80 C~ ~c ! DC +/- set Drift Correction mode (on/off) ! R1 0-65535,0-65535 new sensor #1 Reference (X,Y) position ! R2 0-65535,0-65535 new sensor #2 Reference (X,Y) position ! D1 0-65535,0-65535 new sensor #1 Drift adjustment (X,Y) position ! D2 0-65535,0-65535 new sensor #2 Drift adjustment (X,Y) position ! BG 0-65535,0-65535 define BeGin (X,Y) position ! JM 0-65535,0-65535 JuMp to new (X,Y) position ( preferred over JX, JY combination ) ! MV 0-65535,0-65535 MoVe to next (X,Y) position ( preferred over NX, NY combination ) ! RV 0-65535,0-65535 make a RiVet to next (X,Y) position ( preferred over VX, VY combination ) where:

10 us is the # of 10s of microseconds resolution = 0.84 us (from 1.19 M~z clock) position is the absolute position in the field distance is the absolute distance in step size units rel incr is the relative # of position increments # passes is the number of redraw passes step periods is the time to cover that # of steps programmed by SP command (= tick counts) delay delay value in milliseconds (may be faster on faster computers) In the above list, following each two-character command is the range of numbers allowed as arguments to the co = nd. Each co = nd must be on a separate line, ending with a CR or CR/LF sequence. All spaces are ignored.

All numbers are in unsigned decimal.

The Control Instruction Set is intended to be compatible with General Scanning's DG Series Electronic Packages for Laser Control. Only the Vector Mode is supported. The Raster Mode is not supported. The step and delta functions of the Vector Mode are also unsupported. No Z-axis control is supported yet.

C~ n~C marked with an '~' are not fully implemented as per the DG specification for functional reasons. C~ n~c marked with '+' are included only for the sake of complete command compatibility. '!' denotes new c~ n~c not available on the DG devices.

DG Version 2.6 Vector Mode C~ n~c not supported:

ST enter STep mode RE REset to top of stack in step mode DL enter DeLta mode Q1 Query IB 1 Q2 Query IB 2 The co = nd line with invokes STEREO and causes it to be installed in memory takes the form:

STEREO [/options] 1 lo .', ~ '3 l 334052 ~ where /options are any number of user-selectable options:

/D - debug mode, display vector information /NV - no constant vector veloc:ty /NG - no geometric correction /NC - no version check at install time STEREO can only be installed once a'ter the computer has been powered up. If different options ase required, the computer system will have to be completely rebooted.

The SP (and VP) command has multiple arguments. Each of up to ten passes of redraw can have its own s~ep period. The SP c~ n~ can have 1 to 10 arguments, each a step period for a pass:

SP sp0,spl,sp2,sp3...

where sp0 is the step period for redraw pass 0 ~the original vector move), spl is the step period for redraw pass 1, and so on. /V121 option is not longer supported.

Geometric correction is automatical'y invoked. If geometric correction is not desired, the optaon /NG or /GEO NO should be specified on the command line. Gecmetric correction occurs on the endpoints of any vector STEREO passes to the software interrupt handler.

STEREO code came from Version 1.40 of CLIFF, which originally performed the command translation. This memory-resident STEREO device driver was formed to speed up 3D Systems' supervisory program.s, written in GWBASIC and Turbo BASIC. With the Geo Correct Table, STEREO takes up approximately 80K of memory, increasing to about 150K the amount of main memory used by the complete laser controller system ar.d its data structures.

Axes Control Options: DEFAULT /X +/- select X-axis direction + /Y +/- select Y-axis direction + /SWAPXY swap X and Y coordinates no /GEO XYIYXINO select geo correct sequence YX (used for old style) /AXESBEGIN to axes-control begin pos no Recent History:

8/18/87 Ver 2.02 STEREO released memory resident portion of Ver l.x CLIFF process WI Wait for Idle command added 8/22/87 Ver 2.03 Stuart's new Geometric Correction Y is the first, X i5 the second mirror /NC no version check flag jump vectors of 0 steps skipped CvtInt now works for 32768 ($8000) Step Periods supported for stream (non-block) mode 8/25/87 Ver 2.04 Geometric Correction delta bugs fixed 9/23/87 Ver 2.12 STEREO Axes Control added new options /X, /Y, /SWAPXY, /GEO 9/30/87 Ver 2.25 Table-Interpolatron Geometric Correction installed 10/26/87 Ver 2.26 Wait for Idle ar~lment is 100 mlni lm /NG prevents loading of STEREO.GEO 10/28/87 Ver 2.27 improved Geometrlc Correction formula 11/12/87 Ver 2.30 first Beta software release , 31 1 334052 11/18/87 Ver 2.31 allow axes control independent of geo correction /AXESBEGIN option added to axes-control BX,BY GeoCorrectTable size made programmable (70 elements/axis) added TableDivisor for GeoCorrectTable 12/11/87 Ver 2.40 STEREO upgraded to Turbo Pascal Version 4.0 STEREO always returns with Queue Size Change in CX GeoCorrectTable storage allocated on heap dynamically max GeoCorrectTable size increased to 90 by 90 Some variables changed to Nord or LongInt type 12/16/87 Ver 2.41 GeoCorrect overwrite / memory alloc debugged QueueCount substituted for QueueSize calculation 12/21/87 Ver 2.43 Check for Break turned off improved keyboard and Ctrl-Break Abort service 1/06/88 Ver 2.50 zero-length vectors ignored Ctrl-Break check looks at scan code now 1/27/88 Ver 2.60 better STEREO.GEO processing blanks allowed around commas separating numbers check for too few or too many entries in STEREO.GEO no mark and redraw around executes when redraw count is 1 OkayForRedrawCommand flag added to limit redundant mark and redraw c~ n~ 3/09/88 Ver 2.80 added copyright notice implemented drift correction logic for profiler sensors reduced ASCII overhead for begin, jump, and next commands installed support for binary STEREO c~ n~e new STEREO c~ n~e DC, R1, R2, Dl, D2, BG, JP, MV, RV 4/09/88 Ver 3.00 production release 4/18/88 Ver 3.01 fixed internal version number drlft correction defaults to on 4/18/88 Ver 3.10 Support Numeric Coprocessor 4/27/88 Ver 3.11 STEREO memory resident code size increased 5/02/88 Ver 3.12 Drift Correction Y-axis bug removed 5/13/88 Ver 3.13 JP renamed to JM, JP now reserved Jump Period 5/16/88 Ver 3.14 new multi-timebase LASER support 6/06/88 Ver 3.15 improved Jump technique using multi-timebase 7/01/88 Ver 3.16 correct Laser Unit for Co Processor check for Multi Timebase afer checking for LASER 7/06/88 Ver 3.20 debugged Numeric Coprocessor uses alternate LibTrunc and LibRound routines 7/18/88 Ver 3.20 update MinSystemVersion ~$R-,S-,I-,D-,T-,F-,V-,B-,L{$M 16384,0,65536 ~ { CoProcessor Support ) {$IFOPT N+~ {$DEFINE COP~ {$ELSE~ ~$UNDEF COP~ ~$ENDIF~ uses Dos, Crt, Laser, LibNC;

type ~$IFDEF COP~ FloatType = Single;

{$ELSE~ FloatType = Real;

{$ENDIF~ str4 = string[4];

strl6 = string[16];

1~ 3~ -- 1 334052 str80 = string[80];

BitArray = array[0..15] of Word;

GeoCorrectTableType = array[0..10000] of LongInt;

GeoCorrectTablePtr = ^GeoCorrectTableType;

FpPointerStructType = array[0..8] of Byte;

FpPointerType = ^FpPointerStructType PositionType = record X: FloatType;

Y: FloatType;

end;

SensorDataType = record Ref: PositionType;

Adj: PositionType;

end;

var FirstVar: Byte; ~ must be first variable defined } ch,key: Char;

temp,cmd,TextCommand,DefaultLine: String;

OperationMnemonics,Operand: String;

StereoVersion: strl6;

count,i,j,remline,Er,OpNum,code,code2: Integer;

SensorNum,IOerr,DefIndex: Integer;

GeoFile,DefaultFile: text;

tempX,tempY,Xl,Yl,X2,Y2,oldX2,oldY2: FloatType;

arg,beginXpos,beginYpos,Xpos,Ypos,oldX,oldY: FloatType;

FirstX,NegX,NegY,SwapXY,SwapBegin: Boolean;

first,VelocityFlag,DebugMode,Redrawing,done: Boolean;

OkayForRedrawC~ n~,MultiTimeBaseFlag: Boolean;

Step Size, Step Period, Scan Delay, Jump Period, Jump Delay: Word;

Jump Size, Rivet Reduction: FloatType;

Laser On Delay, Laser Off Delay: Word;

Redraw Count, Redraw Delay, Redraw Size: Word;

KeyAbort,VersionCheckFlag,GeoCorrectFlag,DoingRivet: Boolean;

DriftCorrectionFlag,BinaryFlag: Boolean;

changeX,changeY,ChangeV,BrokenX,BrokenY: FloatType;

deltaX,deltaY,steps,GeoDeltaX,GeoDeltaY: FloatType;

Distance,IntDeltaX,IntDeltaY: Integer;

LongDeltaX,LongDeltaY,LongSteps,TotalSteps,TempSteps: LongInt;

RivetMainPeriod,CurrentRedrawCount, Rivet Count: Word;

StoreCount,QueueCount: Word;

angX,angY,tanX,GeoScale,GeoFact: FloatType;

FieldX,FieldY,StartX,StartY,EndX,EndY: FloatType;

GeoEndX,GeoEndY,Xcoord,Ycoord: FloatType;

DiscreteX,DiscreteY: Integer;

LongDiscreteX,LongDiscreteY: LongInt;

DiscreteSteps: Word;

regs: Registers;

XC,YC,XCount,YCount,CoordCount,TableDivisor: Word;

StereoSize: LongInt;

C -n~fs,C ~eg: Word;

data,Return: Byte;

XGeoCorrectTable,YGeoCorrectTable: GeoCorrectTablePtr;

Sensor: array[1..2] of SensorDataType;

DriftGain: PositionType;

LastVar: Byte; I must be last variable defined const {$IFDEF COP) ThisStereoVersion = 3.20 C ;

{$ELSE~ ThisStereoVersion = 3.20 NC ;

{$ENDIF~ MinSystemVersion = 3.20 ;

CtrlBreak = #0;

HexDigits: strl6 = 0123q56789ABCDEF ;

39 1 334052 ~ sp40 = ' Indentl = 4;

Indent2 = 41;

AllBits = $FFFF;

BitValue: BitArray = ($0001, $0002, $0004, $0008, $0010, $0020, $0040, $0080, $0100, $0200, $0400, $0800, $1000, $2000, $4000, $8000);

Shutter = 0; ( control word bit 0 - shutter open/closed } BIOSVECTOR = $16;

STEREOVECTOR = $65;

MaxGeo = 100; { ~-xi 1- GeoCorrectTable Index ~ GeoMult = 4096; { GeoCorrectTable LongInt multiplier } StepRound = 0.02; { rounding value for # of laser steps } label ExitProgram;

function StartOfProyL ~sd~: String;

begin StartOfProgramCode := 'STEREO Device Driver Version 3.20 Copyright (C) 1988 by Tarnz Technologies ';

end;

var StepPeriodsArray,RivetPeriodsArray: StepPeriodsType;

function _StereoOk: Boolean;

type CopyType = array[1..20] of char;

var CopyPtr: ^CopyType;

CopyStr: string[20];

begin with regs do begin AX := ($35 shl 8) or STEREOVECTOR;

MsDos(regs);

CopyPtr := Ptr(ES,57);

end;

move(CopyPtr^~l],CopyStr[1],18);

CopyStr[0] := #18;

if CopyStr = 'Tarnz Technologies' then begin StereoOk := true;

with regs do CopyPtr := Ptr(ES,30);

move(CopyPtr^[l],StereoVersion[1],4);

StereoVersion[0] := #4;

end else begin StereoOk := false;

StereoVersion := '?.??';

end;

end;

function upper(text: String): String;

var i: Integer;

temp: String;

begin temp := text;

for i := 1 to length(temp) do temp[i] := upcase(temp[i]);

upper := temp;

end;

function many(Ch: Char;num: Integer): String;

var temp: String;

i: Integer;

begin O 1 334052 ~ temp := '';

for i:=1 to num do temp := temp+Ch;

many := temp;

end;

procedure Rea~C~ n~T.i n~ (var line: string);

var i: Integer;

begin line := ParamStr(l);

for i:=2 to ParamCount do line := line + ' ' + ParamStr(i);

end;

function CvtDelta(num: FloatType): Integer;

begin num := num * 256;

if num < 32768.0 then CvtDelta := LibTrunc(num) else if num > 32768.0 then CvtDelta := LibTrunc(num) else CvtDelta := $7FFF;

end;

function CvtLongDelta(num: FloatType): LongInt;

var tnum: LongInt;

begin CvtLongDelta := LibTrunc(num) shl 16 + LibTrunc(frac(num) * 65536);

end;

function sgn(num: FloatType): Integer;

begin if num<0 then sgn := -1 else if num>0 then sgn := 1 else sgn := 0;

end;

function CvtTicks(time: FloatType): Word;

begin CvtTicks := LibTrunc(time * 11.9); (119.0 / 10.0 end;

procedure SkipToNum(var str: String);

var i: Integer;

begin i := l;

while (i &lt;= length(str)) and not(str[i] in ['0'..'9','.']) do inc(l);

if i>1 then str := copy(str,i,99);

end;

procedure RealParse(var RealNum: FloatType);

var i: Integer;

begin i := l;

while (i &lt;= length(temp)) and (temp[i] = ' ') do inc(i);

delete(temp,l,i-1);

if copy(temp,1,1) = '.' then insert('O',temp,1);

val(temp,RealNum,code);

if code<>0 then val(copy(temp,1,code-1),RealNum,code2);

temp := copy(temp,code+1,99);

end;

procedure IntParse(var IntNum: Word);

var num: FloatType;

~1 1 334052 begin if length(temp)=0 then IntNum := 0 else begin val(temp,num,code);

if code<>0 then val(copy(temp,l,code-l),num,code2);

IntNum := LibTrunc(num);

if code=0 then temp := '' else temp := copy(te~p,code+1,99);

end;

end;

function IntStr(num: Integer): strl6;

var temp: strl6;

begin if num>=0 then Str(num,te~p) else Str(num+$8000,temp);

IntStr := temp;

end;

function RealStr(num: FloatType): strl6;

var temp: strl6;

begin if num&lt;0 then Str(65536.0+num:6:0,temp) else Str(num:6:0,temp);

while temp[1]=' ' do temp := copy(temp,2,6);

RealStr := temp;

end;

function spc(len: Integer): str80;

begin spc := copy(sp40+sp40+sp40,1,len);

end;

procedure msg(text: String);

begin (~ return text to calling program ~) end;

procedure msgln(text: String);

begin (~ return text + Cr to calling program ~) end;

function MemAbs(pt: Pointer): LongInt;

begin MemAbs := LongInt(seg(pt^)) shl 4 + ofs(pt^);

end;

( converts a Single type (4 bytes) number to a Real (6 bytes) procedure ConvertStoR(s,r: FpPointerType);

begin r^ 0 :=(((s^[3] shl 8) + s^[2]) shr 7) and $FF;

r^ 1 :=0;

r^ 2 :=0;

r^ 3 :=s^[0];

r^ 4 :=s^[1];

r^ 5::=(s^[3] and $80) or (s^[2] and $7F);

if r^[0]<>0 then Inc(r^[0],2); ( e offset } end;

function CharArgument: Char;

begin if BinaryFlag then begin 1 334052 CharArgument := char(Mem[r~ n~eg:C~ n~Ofs]);

inc(C~ -ndofs~;

end else begin if length(Operand) > 0 then CharArgument := Operand[1] else CharArgument := #0;

delete(Operand,1,1); { slow, avoid for long strings } end;

end;

function WordArgument: Word;

var WordArg: Word;

code,code2: Integer;

begin if BinaryFlag then begin WordArgument := Mem[C~ ~eg:CommandOfs] shl 8 + Mem[C~ eg:CommandOfs + 1];

inc(C~ n~fs,2);

end else begin val(Operand,WordArg,code);

if code<>0 then val(Operand,WordArg,code2);

WordArgument := WordArg;

if code = 0 then Operand := '' else delete(Operand,l,code);

end;

end;

function FloatArgument: FloatType;

var FloatArg: FloatType;

begin if BinaryFlag then begin ~$IFDEF COP} Move(Mem[C. ~.~eg:C~m~An~Ofs],FloatArg,4);

($ELSE) ConvertStoR(@Mem[C~ n~.Seg:C~ fs],~FloatArg);

~$ENDIF) FloatArgument := FloatArg;

inc(C~ -n~Ofs,4);

end else begin if copy(Operand,1,1) = '.' then insert('0',Operand,1);

val(Operand,FloatArg,code);

if code<>0 then val(copy(Operand,l,code-l),FloatArg,code2);

FloatArgument := FloatArg;

if code = 0 then Operand := '' else delete(Operand,l,code);

end;

end;

function TimeArgument: Word;

begin TimeArgument := CvtTicks(WordArgument);

end;

function PositionArgument: Word;

var PosArg: Word;

code,code2: Integer;

begin if BinaryFlag then begin PositionArgument := MemW[C~ eg:CI ndOfs];

inc(C nA~fs,2);

end else begin while copy(Operand,1,1) = ' ' do delete(Operand,1,1);

val(Operand,PosArg,code);

if code<>0 then val(copy(Operand,l,code-l),PosArg,code2);

. ~ , ~ ~3 `_ 1 334052 PositionArgumene := PosArg;

if code = 0 then Operand := '' else delete(Operand,1,code);

end;

end;

function PtrAdd(point: Pointer;offset: LongInt): Pointer;

var NewAddr: LongInt;

begin NewAddr := Mem~bs(point)+offset;

PtrAdd := Ptr(NewAddr div 16,NewAddr mod 16);

end;

function StatusKey: Char;

begin with regs do begin AX := $0100;

Intr( BIOSVECTOR,regs);

StatusKey := chr(AH); ( return scan code ) end;

end;

procedure SIHerror;

begin Return := Er;

end;

function TryCommand: Boolean;

begin if Er=3 (gueue full~ then TryCommand := true else TryC~ ' := false;

if keypressed (abortl then begin lf StatusKey = CtrlBreak then begin TryCommand := false;

KeyAbort := true;

end;

end;

end;

function TryBlockCommand: Boolean;

begin if (Er=3) (command queue full) or (Er=6) (exec queue full} then TryBlockCommand := true else TryBlockCommand := false;

if keypressed (abort~ then begin if StatusKey = CtrlBreak then begin TryBlockCommand := false;

KeyAbort := true;

end;

end;

end;

procedure newX(num: FloatType);

begin oldX := Xpos;

Xpos := num;

end;

procedure newY(num: FloatType);

begin oldY := Ypos;

Ypos := num;

end;

. ~ 1 334052 AxesControl -- adjust axes } procedure AxesControl(oldX,oldY: FloatType;var retX,retY: FloatType);

begin if SwapXY then begin retY := oldX;

retX := oldY;

end else begin retX := oldX;

retY := oldY;

end;

if NegX then retX := 65535.0 - retX;

if NegY then retY := 65535.0 - retY;

end;

procedure SetRedrawCount(RedrawCount: Integer);

begin if RedrawCount<>CurrentRedrawCount then begin CurrentRedrawCount := RedrawCount;

Er := 3;

while TryCommand do Er := _SetRedrawCount(CurrentRedrawCount);

inc~QueueCount,3);

SIHerror;

end;

end;

{ Laser Motion Routines:

* ResetRedraw - just mark for beginning of block * PerformRedraw - redraw and mark last section * MoveLaser - new move alg using redraw &amp; vector slicing * JumpLaser - controlled breakup of jump motion to limit overshoot } procedure ResetRedraw;

begin if OkayForRedrawCommand and _BlockModeOn then begin if DebugMode then msgln('Mark '+IntStr(_Redraw_Size));

Er := 3;

while TryC- ~ ~ do Er := _Mark;

inc(QueueCount);

SIHerror;

end;

Distance := 0;

Redrawing := false;

OkayForRed~ C- n~ := false;

end;

procedure PerformRedraw;

begin if OkayForRedrawCommand and _BlockModeOn then begin if DebugMode then msgln('Redraw '+IntStr( Redraw Size));

Er := 3;

while TryCommand do Er := _ Redraw;

inc(QueueCount);

SIHerror;

end;

Di~tance :z 0;

Redrawing := false;

OkayForRedrawCommand := false;

end;

~t5 1 334052 CorrectForDrift -- develop drift gain terms procedure CorrectForDrift;

var quotient: FloatType;

begin quotient := Sensor[2].Ref.X - Sensor[l].Ref.X;

if quotient <> O then DriftGain.X := (Sensor[2].Adj.X - Sensor[1].Adj.X) / quotient else DriftGain.X := O; ~(* writeln('CorrectForDrift-X:');

writeln('2Xadj=',Sensor[2].Adj.X:8:2,' - lXadj=',Sensor[l].Adj.X:8:2,' /');

writeln('2Xref=',Sensor[2].Ref.X:8:2,' - lXref=',Sensor[l].Ref.X:8:2);

writeln('DriftGainX=',DriftGain.X:8:2);

*) quotient := Sensor[2].Ref.Y - Sensor[l].Ref.Y;

if quotient <> O then DriftGain.Y := (Sensor[2].Adj.Y - Sensor[l].Adj.Y) / quotient else DriftGain.Y := O;

(* writeln('CorrectForDrift-Y:');

writeln('2Yadj=',Sensor[2].Adj.Y:8:2,' - lYadj=',Sensor[l].Adj.Y:8:2,' /');

writeln('2Yref=',Sensor[2].Ref.Y:8:2,' - lYref=',Sensor[l].Ref.Y:8:2);

writeln('DriftGainY=',DriftGain.Y:8:2);

*) end;

t GeoCorrect -- geometric correct coordinate pair procedure GeoCorrect(oldX,oldY: FloatType;var newX,newY: FloatType);

var GXll,GX12,GX21,GX22,GYll,GY12,GY21,GY22: FloatType;

Xf,Yf,XfYf: FloatType;

XIndex,YIndex,Lookup: Word;

begin if GeoCorrectFlag then begin XIndex := LibTrunc(oldX / TableDivisor);

if XIndex >= XCount then XIndex := XCount-l;

YIndex := LibTrunc(oldY / TableDivisor);

if YIndex >= YCount then YIndex := YCount-l;

Lookup := XIndex * YCount + YIndex;

GXll := XGeoCorrectTable^ Lookup];

GX12 := XGeoCorrectTable^ Lookup+l];

GX21 := XGeoCorrectTable^ Lookup+YCount];

GX22 := XGeoCorrectTable^ Lookup+YCount+l];

GYll := YGeoCorrectTable^ Lookup];

GY12 := YGeoCorrectTable^ Lookup+l];

GY21 := YGeoCorrectTable^ Lookup+YCount];

GY22 := YGeoCorrectTable^ Lookup+YCount+l];

Xf := (oldX - TableDivisor * 1.0 * XIndex) / TableDivisor;

Yf := (oldY - TableDivisor * 1.0 * YIndex) / TableDivisor;

XfYf := Xf * Yf;

newX := (GXll + Xf * (GX21 - GXll) + Yf * (GX12 - GX11) + XfYf * (GX22 + GX11 - GX12 - GX21)) / GeoMult;

newY := (GY11 + Xf * (GY21 - GY11) + Yf ~ (GY12 - GY11) + XfYf * (GY22 + GY11 - GY12 - GY21)) / GeoMult;

end else begin newX := oldX;

newY := oldY;

end;

if DriftCorrectionFlag then begin newX := (newX - Sensor[l].Ref.X) * DriftGain.X + Sensor[l].Adj.X;

1 334052 ~ newY := (newY - Sensor[1].Ref.Y) * DriftGain.Y + Sensor[1].Adj.Y;

end;

end;

procedure MoveLaser;

begin AxesControltXpos,Ypos,EndX,EndY);

AxesControl(oldX,oldY,StartX,StartY);

changeX := EndX - StartX;

changeY := EndY - StartY;

if VelocityFlag then changeV := sqrt(sqr(changeX) + sqr(changeY)) else if abs(changeX) &lt; abs(changeY) then ChangeV := abs(changeY) else ChangeV := abs(changeX);

steps := ChangeV / Step_Size;

if steps>=1 then begin deltaX := changeX / steps;

deltaY := changeY / steps;

LongSteps := LibTrunc(steps+StepRound);

IntDeltaX := CvtDelta(deltaX);

IntDeltaY := CvtDelta(deltaY);

TotalSteps := 0;

repeat BrokenX := StartX + deltaX * TotalSteps;

BrokenY := StartY + deltaY * TotalSteps;

TempSteps := LongSteps - TotalSteps;

if TempSteps>256 then TempSteps := 256;

if (CurrentRedrawCount > 1) and (_Redraw_Size>0) then if Distance+TempSteps >= _Redraw_Size then begin TempSteps := _Redraw Size - Distance;

Redrawing := true;

end;

TotalSteps := TotalSteps + TempSteps;

Distance := Distance + TempSteps;

GeoCorrect(BrokenX,BrokenY,FieldX,FieldY);

{ Geometric Correction addition ~ if GeoCorrectFlag then begin GeoCorrect(BrokenX+TempSteps*DeltaX,BrokenY+TempSteps*DeltaY, GeoEndX,GeoEndY);

changeX := GeoEndX - FieldX;

changeY := GeoEndY - FieldY;

ChangeV := sqrt(sqr(changeX) + sqr(changeY));

steps := ChangeV / _Step_Size;

GeoDeltaX := ChangeX / steps;

GeoDeltaY := ChangeY / steps;

DiscreteX := CvtDelta(GeoDeltaX);

DiscreteY := CvtDelta(GeoDeltaY);

DiscreteSteps := LibTrunc(steps+StepRound);

end else begin DiscreteX := IntDeltaX;

DiscreteY := IntDeltaY;

DiscreteSteps := TempSteps;

end;

Er := 3; { send out move command } while TryCommand do begin Er := _ AddVector(LibTrunc(FieldX),LibTrunc(FieldY), DiscreteX,DiscreteY,DiscreteSteps);

OkayForRedrawcommand := true;

end;

inc(QueueCount,13);

SIHerror;

if Redrawing then PerformRedraw;

until TotalSteps = LongSteps;

end;

end;

~ c "~,`................................... ~1 1 334052 ~ procedure SingleTimebaseJu~pLaser;

begin if Jump_Delay<>0 then begin AxesControl(Xpos,Ypos,EndX,EndY);

AxesControl(oldX,oldY,StartX,StartY);

changeX := EndX-StartX;

changeY := EndY-StartY;

if VelocityFlag then changeV := sqrt(sqr(changeX)+sqr(changeY)) else if abs(changeX) &lt; abs(changeY) then ChangeV := abs(changeY) else ChangeV := abs(changeX);

steps := ChangeV / Jump Size;

if steps>0 then begin deltaX := changeX / steps;

deltaY := changeY / steps;

LongSteps := LibTrunc(steps+StepRound) + l;

TotalSteps := l;

repeat BrokenX := StartX + deltaX*TotalSteps;

BrokenY := StartY + deltaY*TotalSteps;

TotalSteps := TotalSteps + l;

if TotalSteps >= LongSteps then begin BrokenX := EndX;

BrokenY := EndY;

end;

GeoCorrect(BrokenX,BrokenY,FieldX,FieldY);

Er := 3; ~ add discrete jump } while TryCo = nd do begin Er := _AddVector(LibTrunc(FieldX),LibTrunc(FieldY), 0,0,1); ( incl a 1 tick pause OkayForRedrawCommand := true;

end;

inc(QueueCount,13);

SIHerror;

until TotalSteps >= LongSteps;

end;

Er := 3;

while TryCommand do Er := _ AddPause( Jump_Delay);

inc(QueueCount,3);

SIHerror;

end;

end;

procedure MultiTimebaseJumpLaser;

begin AxesControl(Xpos,Ypos,EndX,EndY);

AxesControl(oldX,oldY,StartX,StartY);

changeX := EndX - StartX;

changeY := EndY - StartY;

if VelocityFlag then changeV := sqrt(sqr(changeX) + sqr(changeY)) else if abs(changeX) &lt; abs(changeY) then ChangeV := abs(changeY) else ChangeV := abs(changeX);

steps := ChangeV / Jump_Size;

if steps>=l then begin if MultiTimebaseFlag then begin Er := 3;

while TryCommand do Er := _SelectTimerChannel(2);

inc(QueueCount,2);

SIHerror;

end;

deltaX := changeX / steps;

deltaY := changeY / steps;

LongSteps := LibTrunctsteps+StepRound);

LongDeltaX := CvtLongDelta(deltaX);

1 334052 ~ LongDeltaY := CvtLongDelta(deltaY);

GeoCorrect(StartX,StartY,FieldX,FieldY);

GeoCorrect(EndX,EndY,GeoEndX,GeoEndY);

{ Geometric Correction addition } if GeoCorrectFlag then begin changeX := GeoEndX - FieldX;

changeY := GeoEndY - FieldY;

ChangeV := sqrt(sqr(changeX) + sqr(changeY~);

steps := ChangeV / _Jump_Size;

GeoDeltaX := ChangeX / steps;

GeoDeltaY := ChangeY / steps;

LongDiscreteX := CvtLongDelta(GeoDeltaX);

LongDiscreteY := CvtLongDelta(GeoDeltaY);

DiscreteSteps := LibTrunc(steps+StepRound);

end else begin LongDiscreteX := LongDeltaX;

LongDiscreteY := LongDeltaY;

DiscreteSteps := LongSteps;

end;

Er := 3; ~ send out jump as fast move command } while TryCommand do Er := _AddLongVector(LibTrunc(FieldX),LibTrunc(FieldY), LongDiscreteX,LongDiscreteY,DiscreteSteps);

inc(QueueCount,19);

SIHerror;

if _Jump_Delay > O then begin Er := 3; { send out jump as fast move command } while TryCommand do Er := _AddVector(LibTrunc(GeoEndX),LibTrunc(GeoEndY),O,O,1);

inc(QueueCount,13);

SIHerror;

if _Jump_Delay > 1 then begin Er := 3;

while TryCommand do Er := _AddPause(_Jump_Delay-l);

inc(QueueCount,3);

SIHerror;

end;

end;

if MultiTimebaseFlag then begin Er := 3;

while TryCommand do Er := _SelectTimerChannel(1);

inc(QueueCount,2);

SIHerror;

end;

end;

end;

procedure JumpLaser;

begin if MultiTimebaseFlag then MultiTimebaseJumpLaser else SingleTimebaseJumpLaser;

end;

_ procedure BeginX;

begin beginXpos := PositionArgument;

end;

procedure BeginY;

, X1,Y1,X2,Y2: FloatType; 1 3 3 4 0 5 2 begin beginYpos := PositionArgument;

if SwapBegin then AxesControl(beginXpos,beginYpos,X1,Y1) else begin X1 := beginXpos;

Y1 := beginYpos;

end;

GeoCorrect(X1,Y1,X2,Y2);

Er := _SetHomePosition(LibTrunc(X2),LibTrunc(Y2));

SIHerror;

end;

procedure BeginPosition;

begin BeginX;

BeginY;

end:

procedure JumpX;

begin newX(PositionArgument);

end;

procedure JumpY;

begin newY~PositlonArgument);

JumpLaser;

end;

procedure PositionJump(X,Y: FloatType);

begin NewX(X);

NewY(Y);

if Jump Delay<>O then begin JumpLaser;

Er := 3;

while TryCommand do Er := _AddPause( Jump Delay);

inc(QueueCount,3);

SIHerror;

end;

end;

procedure JumpPosition;

begin JumpX;

JumpY;

end;

procedure NextX;

begin newX(PositionArgument);

end;

procedure NextY;

begin newY(PositionArgument);

MoveLaser;

end;

procedure PositionMove(X,Y: FloatType);

begin newX(X);

newY(Y);

,_ S ~ MoveLaser;

end;

procedure MoveToPosition; 1 3 3 4 0 5 2 begln NextX;

NextY;

end;

procedure StepSize;

begin Step_Size := WordArgument;

lf _Step Size &lt; 1 then _Step Size := 1; { limit end;

procedure StepPeriods;

var i: Integer;

begin PerformRedraw;

if BinaryFlag then begin count := WordArgument;

for i:=0 to count-1 do StepPeriodsArray.periods[i] := TimeArgument;

end else begin count := 0;

while (Operand <> '') and (count < 10) do begin StepPeriodsArray.periods[count] := TimeArgument;

inc(count);

end;

end;

StepPeriodsArray.num := count;

if count>0 then begin (* if DebugMode then begin msg('New step periods array: ');

for i:=0 to count-1 do msg(StepPeriodsArray.periods[i],' ');

msgln('');

end; *) Er := 3;

if _ BlockModeOn then begin while TryCommand do Er := _SetStepPeriods(StepPeriodsArray);

inc(QueueCount,count*2+2);

end else begin while TryCommand do Er := ~ h~ngeTimebase(StepperiodsArray.periods[o]);

inc(QlleueCo-lnt,3);

end;

SIHerror;

end;

PerformRedraw;

end;

procedure ScanDelay;

begin Scan Delay := WordArgument;

end;

procedure JumpPeriod;

begin if MultiTimebaseFlag then begin Jump Period := TimeArgument;

Er := 3;

while TryC~ n~ do Er := _SelectTimerChannel(2);

inc(QueueCount,2);

SIHerror;

wh;le TryCommand do 1 3 3 4 0 5 2 Er := _AddChangeTimebase( Jump_Period);

inc(QueueCount,3);

SIHerror;

Er := 3;

while TryC~ n~ do Er := _SelectTimerChannel(1);

inc(QueueCount,2);

SIHerror;

end;

end;

procedure JumpSize;

begin _Jump_Size := WordArgument;

if _Jump_Size &lt; 1 then _Jump_Size := 1; 1 limit ) end;

procedure JumpDelay;

begin _Jump_Delay := WordArgument;

end;

procedure LaserOn;

begin _Laser_On_Delay := TimeArgument;

Er := _SetTurnOnDelay(_Laser_On_Delay);

SIHerror;

end;

procedure LaserOff;

begin _Laser_Off_Delay := TimeArgument;

Er := _SetTurnOffDelay(_Laser_Off_Delay);

SIHerror;

end;

procedure SequenceMode;

begin ( no-operation } end;

procedure AbsoluteMode;

begin { no-operation } end;

procedure Execute;

begin PerformRedraw;

Er := 3;

while TryBlockCommand do Er := _Execute;

SIHerror;

ResetRedraw;

end;

procedure ExecuteClear;

begin PerformRedraw;

Er := 3;

while TryBlockCo = nd do Er := _ExecuteClear;

QueueCount := 0;

SIHerror;

~ ResetRedraw end; 1 3 3 4 0 5 2 procedure ClearTable;

begin Er := _Clear;

QueueCount := O;

if Er<>5 then SIHerror; { not in block mode okay } ResetRedraw;

end;

procedure SetMode;

begin if BinaryFlag then begin case CharArgument of 'A': Er := _SetAutoShutterMode(CharArgument);

'B': Er := _SetBlockMode(CharArgument~;

'P': Er := _SetPosAckMode(CharArgument);

end;

end else begin if length(TextCommand)=5 then case TextCommand[3] of 'A': Er :=. _SetAutoShutterMode(TextCommand[5]);

'B': Er := _SetBlockMode(TextCommand[5]);

'P': Er := _SetPosAckMode(TextCommand[5]);

end;

end;

SIHerror;

end;

procedure SoftReset;

begin Er := _SetAutoShutterMode('-'); 1 off ) SIHerror;

Er := _SetBlockMode('-'); { off ) SIHerror;

Er := _SetPosAckMode('-'); ~ off ) SIHerror;

QueueCount := O;

end;

procedure HardReset;

begin SoftReset;

Er := _ ResetCmdQueue;

SIHerror;

end;

procedure RedrawCount;

begin Redraw Count := WordArgument;

if Redraw Count < 1 then Redraw Count := 1 else if Redraw Count > 10 then Redraw Count := 10;

SetRedrawCount( Redraw Count);

PerformRedraw;

Er := 3;

while TryCommand do Er := _SetRedrawCount( Redraw Count);

inc(QueueCount,3);

SIHerror;

end;

procedure RedrawDelay;

begin Redraw Delay := WordArgument;

Er := 3;

while TryCommand do 53 Er := _SetRedrawDelay(_Redraw_Delay); 2 inc(QueueCount,3); 1 3 3 4 0 5 SIHerror;

end;

procedure RedrawSize;

begin Redraw_Size := WordArgument;

end;

procedure RivetCount begin Rivet Count := WordArgument;

f _Rivet Count < 1 then _Rivet_Count := 1 else if _Rivet_Count > 10 then _Rivet_Count := 10;

_Rivet_Count := _Rivet_Count - l;

end;

procedure RlvetPeriods;

var i: Integer;

begin if BinaryFlag then begin count := WordArgument;

RivetMainPeriod := TimeArgument;

for i:=0 to count-l do RivetPeriodsArray.periods[i] := TimeArgument;

end else begin RivetMainPeriod := TimeArgument;

count := 0;

while (Operand <> '') and (count<10) do begin RivetPeriodsArray.periods[count] := TimeArgument;

inc(count);

end;

end;

RivetPeriodsArray.num := count;

(* if DebugMode then begin msg('New rivet periods array: ');

msg(' Main=',RivetMainPeriod);

for i:=0 to count-l do msg(' ',RivetPeriodsArray.periods[i]);

msgln('');

end; *) end;

procedure RivetReduction;

begin _Rivet_Reduction := WordArgument;

end;

procedure ResetFromRivets;

begin if RivetPeriodsArray.num>0 then begin ~ set standard step periods } Er := 3;

while TryC~ ~ ~ do Er := _SetStepPeriods(StepPeriodsArray);

inc(QueueCount,StepPeriodsArray.num*2+2);

SIHerror;

end;

end;

procedure RivetX;

begin newX(PositionArgument);

end;

procedure RivetY;

1 334052 var StartX,StartY,EndX,EndY: FloatType;

ReducedStartX,ReducedStartY,ReducedEndX,Re~llc~F.n~Y: FloatType;

ReductionX,ReductionY: FloatType;

DirX,DirY,ReducedDirX,ReducedDirY: Integer;

begin StartX := oldX;

StartY := Ypos;

EndX := Xpos;

EndY := PositionArgument;

{ set rivet step periods Er := 3;

while TryCommand do Er := ~ rh~ngeTimeBase(RivetMainperiod);

inc(QueueCount,3~;

SIHerror;

~ draw complete vector NextY;

DirX := sgn(EndX - StartX);

DirY := sgn(EndY - StartY);

ReductionX := Rivet_Reduction * deltaX;

ReductionY := Rivet Reduction * deltaY;

ReducedStartX .= StartX + ReductionX;

RP~ e~tartY := StartY + ReductionY;

~dllce~n~x := EndX - ReductionX;

ReducedEndY := EndY - ReductionY;

ReducedDirX := sgn(ReducedEndX - R~duce~CtartX);

ReducedDirY := sgn(ReducedEndY - ReducedStartY);

DoingRivet := true;

~ do rivet if reduced vector size is greater than O units ) if ((ReducedDirX = DirX) and (ReducedDirY = DirY)) and ( Rivet Count > O) then begin ~ end standard redraw if any PerformRedraw;

if RivetPeriodsArray.num>O then begin ~ set rivet step periods } Er := 3;

while TryCommand do Er := _ SetStepPeriods(RivetPeriodsArray);

inc(QueueCount,RivetPeriodsArray.num*2+2);

SIHerror;

end;

~ set rivet pass count } SetRedrawCount( Rivet Count);

~ draw reduced vector - the rivet } PositionJump(Re~llce~tartX,Re~uceACtartY);

PositionMove(R~ ce~Fn~X~ReducedEndY);

~ end rivet redraw } PerformRedraw;

~ restore original redraw pass count } SetRedrawCount( Redraw Count);

~ jump to end of full vector } PositionJump(EndX,EndY~;

end;

{ reset standard step periods } Er := 3;

while TryC~ d do Er := _ SetStepPeriods(StepPeriodsArray);

inc(QueueCount,StepPeriodsArray.num*2+2);

SIHerror;

DoingRivet := false;

end;

procedure RivetPosition;

begin RivetX;

5~ RivetY; 1 3 3 4 0 5 2 end;

procedure WaitIdle var time: Integer;

begin time := WordArgument;

if time &lt; 100 then time := 100; ( set mini delay time } delay(time); { delay long enough for background task to see EX } repeat until _ GetBkgndIdleStatus or keypressed;

if keypressed then if StatusKey = CtrlBreak then KeyAbort := true;

end;

{ Drift Correction C~ n~ } procedure SetDrift;

begin if BinaryFlag then DriftCorrectionFlag := (CharArgument = '+'} else DriftCorrectionFlag := (TextCommand[3] = '+');

end;

procedure Referencel;

begin Sensor[l].Ref.X := FloatArgument;

Sensor[l].Ref.Y := FloatArgument;

end;

procedure Reference2;

begin Sensor[2].Ref.X := FloatArgument;

Sensor[2].Ref.Y := FloatArgument;

end;

procedure DriftAdjl;

begin Sensor[l].Adj.X := FloatArgument;

Sensor[l].Adj.Y := FloatArgument;

CorrectForDrift;

end;

procedure DriftAdj2;

begin Sensor[2].Adj.X := FloatArgument;

Sensor[2].Adj.Y := FloatArgument;

CorrectForDrift;

end;

{ STEREO Command Parsers procedure CallOperation(OpNum: Integer);

begin case OpNum of 1: BeginX; { BX } 2: BeginY; { BY } 3: JumpX; ( JX } 5~ 1 334052 ~ 9: JumpY; ~ JY } 5: NextX; ~ NX } 6: NextY; ~ NY } 7: StepSize; ~ SS } 8: StepPeriods; ~ SP } 9: ScanDelay; ~ SD } 10: JumpPeriod; ~ JP } 11: JumpSize; ~ JS } 12: JumpDelay; ~ JD } 13: LaserOn; I LO } 19: LaserOff; ~ LF } 15: SequenceMode; { NS } 16: AbsoluteMode; { AB } 17: Execute; { EX } 18: ExecuteClear; ~ EC } 19: ClearTable; ~ CL } 20: SetMode; ~ MD } 21: SoftReset; { SR } 22: HardReset; { HR } 23: RedrawCount; { RC } 29: RedrawDelay; { RD } 25: RedrawSize; { RS } 26: RivetCount; { VC } 27: RivetPeriods; { VP } 28: RivetReduction; { VR } 29: RivetX; { VX } 30: RivetY; { VY } 31: WaitIdle; { WI } 32: SetDrift; { DC } 33: Referencel; { R1 } 39: Reference2; { R2 } 35: DriftAdjl; { D1 } 36: DriftAdj2; { D2 } 37: BeginPosition; { BG } 38: JumpPosition; { JM } 39: MoveToPosition; { MV } 90: RivetPosition; { RV } else Return := 128; { unrecognized command } end;

end;

procedure BinaryParse;

var OpNum: Integer;

begin { break apart multiple - n~ } done := false;

while (not KeyAbort) and (not done) and (Return = 0) do begin OpNum := Mem[C~ n~5eg:C~ -'Ofs];

if OpNum <> 0 then begin ~ cmd of 0 marks end ) inc(C- n~Ofs);

CallOperation(OpNum);

end else done := true;

end;

end;

procedure ParseTextCommand;

var OpName: str9;

begin OpName := copy(TextCommand,1,2);

Operand := copy(TextCommand,3,99);

i := pos(OpName,OperationMnemonics);

CallOperationt(i + 2) div 4); { good cmds give 1,2,3... ) end;

~1 1 334052 -;

~ procedure TextParse;

begin ~ break apart multiple command strings TextCI n~ done := false;

while (not KeyAbort) and (not done) and (Return = 0) do begin Ch := chr(Mem[C~ ~eg:C~ -ndOfs]);

done := Ch &lt; ' ';

if done or (Ch=';') then begin inc(C n~Ofs);

ParseTextC: ~n~;

TextC: n~ end else begin if Ch in ['!'..'~'] then TextCommand := TextCommand + upcase(Ch);

inc(CommandOfs);

end;

end;

end;

Process STEREO C~ n~e coming through the Software Interrupt - Software Interrupt $65 (101) Handler ES:BX points to first character of string. String should end with a character value of 0, 10 (LF), or 13 (CR). An error code will be returned in AL with 0 meaning no error occured.

_ procedure ReceiveC: nd(Flags~cs~Ip~Ax~Bx~cx~Dx~sI~DI~Ds~Es~Bp: Word);

interrupt;

begin ReyAbort := false;

C~ -ndOfs := BX;

C~ -n~Ceg := ES;

INLINE ($FB); ( STI ;allow interrupts } Return := 0;

QueueCount := 0;

case AX of 0: begin BinaryFlag := false;

TextParse;

end;

1: begin BinaryFlag := true;

BinaryParse;

end;

else Return := 129; { invalid STEREO request end;

CX := QueueCount;

AX := Return;

BX := C~ n~Ofs;

end;

Memory Resident Routines ,, S 1 334052 After the following routines install the memory-resident routines above, they are thrown away and not used again.

procedure ResidentTermination;

begin with regs do begin AH := $31;

AL := 1;

DX := LibTrunc(StereoSize) shr 4;

MsDos(regs);

end;

end;

procedure InitMemoryResidency;

begin SetIntVec( STEREOVECTOR,QReceiveCommand);

end;

function WhiteSpace(var str: String;loc: Integer): Integer;

begin while (loc C= length(str)) and (str[loc] in [#9,' ']) do loc := loc + 1;

WhiteSpace := loc;

end;

procedure CheckForOptions(var str: String);

var i: Integer;

begin str := upper(str);

while (length(str) > 1) and (str[1] = ' ') do delete(str,1,1);

i := pos('/D',str); { set debug mode } if i>0 then begin DebugMode := true;

delete(str,i,2);

end;

i := pos('/NV',str); I not constant vector velocity ) if i>0 then begin I larger of X and Y axes ) VelocityFlag := false;

delete(str,i,3);

end;

i := pos('/NG',str); ( no geometric correction if i>0 then begin GeoCorrectFlag := false;

delete(str,i,3);

end;

i := pos('/NMT',str); ( no multiple timebase if i>0 then begin MultiTimebaseFlag := false;

delete(str,i,4);

end;

i := pos('/NC',str); ( no version check ) if i>0 then begin VersionCheckFlag := false;

delete(str,i,3);

end;

i := pos('/SWAPXY',str); ( swap X L Y coordinates ) if i>0 then begin SwapXY := true;

~ 5~ - ` - 1 334052 delete(str,i,7);

end;

i := pos('/AXESBEGIN',str); t axes-control begin (home) coordinates ) if i>O then begin SwapBegin := true;

delete(str,i,10);

end;

i := pos('/GEO',str); t geo-correct axes order if i>O then begin j := WhiteSpace(str,i+4);

case str[j] of 'X': FirstX := true;

'Y': FirstX := false;

'N': GeoCorrectFlag := false;

end;

delete(str,i,j+2-i);

end;

i := pos('/X',str); ( X-axis orientation ) if i>O then begin NegX := (str[i+2]='-');

delete(str,i,3);

end;

i : = pO5 ( I /Y', str); ( Y-axis orientation ) if i>O then begin NegY := (str[i+2]='-');

delete(str,i,3);

end;

end;

procedure ShowAxis(Neg: Boolean;AxCh: Char);

begin if Neg then write('-') else write('+');

write(AxCh,' axis');

end;

procedure ShowGeometry;

begin write('Geometry is ');

if SwapXY then write('XY Swapped, ');

if SwapBegin then write('Begin Swapped, ');

if FirstX then begin ShowAxis(NegX,'X');

write(' first, then ');

ShowAxis(NegY,'Y');

end else begin ShowAxis(NegY,'Y');

write(' first, then ');

ShowAxis(NegX,'X');

end;

writeln('.');

if GeoCorrectFlag then writeln('Geometric Correction active.');

if MultiTimebaseFlag then writeln('Multiple Timebase active.');

end;

procedure ~eadDefaultFile;

begin while not eof(DefaultFile) do begin readln(DefaultFile,DefaultLine);

if (length(DefaultLine)>O) and (DefaultLine[1]<>'!') then while length(DefaultLine)>O do begin DefIndex := pos(';',DefaultLine);

if DefIndex=O then DefIndex := length(DefaultLine) + 1;

TextC~ -n~ : = copy (DefaultLine~l~DefIndex-l);

ParseTextCommand;

delete(DefaultLine,1,DefIndex);

end: (while) 1 334052 end; (while not eof} close(DefaultFile);

end;

~.

{ - - - M A I N P R O G R A M - - - } begin CheckBreak := false;

FillChar(FirstVar,ofs(LastVar)-ofs(FirstVar),#0); { clear variables ) VelocityFlag := true; l default to constant vector velocity ) GeoCorrectFlag := true; { default to geometric correction } DriftCorrectionFlag := true; { perform drift correction } VersionCheckFlag := true;

writeln;

writeln(#213,many(#205,45),#184);

writeln(#179,' STEREO, Stereolithography Device Driver ',#179);

writeln(#17g,' Version ',ThisStereoVersion,' ',#179);

writeln(#179,' Copyright (C) 1988 by Tarnz Technologies ',#179);

writeln(#179,' 3D Systems Laser Stereolithography System ',#179);

writeln(#212,many(#205,45),#190);

writeln;

cmd := StartOfProgramCode;

Rea~C~ 'T.;n~ (cmd);

if not _ LaserPresent then goto ExitProgram;

if _StereoOk then begin writeln('STEREO Device Driver Version ',StereoVersion, ' already installed.');

goto ExitProgram;

end;

CheckForOptions(cmd);

if VersionCheckFlag then if LaserVersion &lt; MinSystemVersion then begin writeln('Obsolete LASER system in use.');

goto ExitProgram;

end;

MultiTimebaseFlag := _ GetTimerCapability; { from LASER } Xpos := 32767;

Ypos := 32767;

DoingRivet := false;

RivetMainPeriod := 1190;

StepPeriodsArray.num := 0;

RivetPeriodsArray.num := 0;

OperationMnemonics := '*BX *BY *JX *JY *NX *NY *SS *SP *SD *JP *JS *JD ' + '*LO *LF *NS *AB *EX *EC *CL *MD *SR *HR *RC *RD *RS *VC *VP *VR ' + '*VX *VY *WI *DC *R1 *R2 *D1 *D2 *BG *JM *MV *RV';

Sensor 1 .Ref.X := 0;

Sensor 1 .Ref.Y := 0;

Sensor 2 .Ref.X := 65535;

Sensor 2 .Ref.Y := 65535;

{ read and/or set defaults } Assign{DefaultFile,'STEREO.DEF');

~$I-} reset(DefaultFile);

{$I+} if IOresult~>0 then begin writeln('Can''t find STEREO.DEF default file!');

Step Size := 2;

Step Period := 119;

Scan Delay := 64 Jump Period := 50;

Jump Size := 1024;

Jump Delay := 0;

Laser On Delay := 0;

_ ~1 1 334052 ~' Laser Off Delay := 0;

Redraw Count := l;

Redraw Delay := 0;

Redraw Size := 200;

Rivet Count := l;

Rivet Reduction := 100;

end else ReadDefaultFile;

CurrentRedrawCount := Redraw Count;

XGeoCorrectTable := HeapPtr;

if GeoCorrectFlag then begin ( read Geometric Correction Table Assign(GeoFile,'STEREO.GEO');

($I-} reset(GeoFile);

($I+} if IOresult<>0 then begin writeln('Can''t find STEREO.GEO Geometric Correction File!');

GeoCorrectFlag := false;

end else begin ( read geometric correction table from file } repeat readln(GeoFile,temp);

if upper(copy(temp,l,7)) = '!SENSOR' then begin val(temp[8],SensorNum,code);

temp := copy(temp,9,99);

if SensorNum in [1..2] then begin RealParse(Sensor[SensorNum].Ref.X);

RealParse(Sensor[SensorNum].Ref.Y);

end;

temp := '!';

end;

until ((length(temp)>0) and (temp[l]<>'!')) or eof(GeoFile);

SkipToNum(temp);

IntParse(XCount);

SkipToNum(temp);

IntParse(YCount);

YGeoCorrectTable := PtrAdd(XGeoCorrectTable,LongInt(XCount*YCount) shl 2);

SkipToNum(temp);

IntParse(CoordCount);

SkipToNum(temp);

IntParse(TableDivisor);

If TableDivisor < 1 then TableDivisor := 1024; I for V2.30 compatibility } if (XCount<>YCount) or (CoordCount<>2) or (XCount>MaxGeo) or (YCount>MaxGeo) then begin writeln('Illegal format for STEREO.GEO file.');

halt(l);

end;

XC := O;

YC := O;

StoreCount := 0;

while (XC&lt;XCount) and not eof(GeoFile) do begin while (YC&lt;YCount) and not eof(GeoFile) do begin repeat readln(GeoFile,temp);

until ((length(temp)>0) and (temp[l]<>'!')) or eof(GeoFile);

SkipToNum(temp);

RealParse(Xcoord);

SkipToNum(temp);

RealParse(Ycoord);

XGeoCorrectTable^[StoreCount] := LibTrunc(Xcoord*GeoMult);

YGeoCorrectTable^[StoreCount] := LibTrunc(Ycoord*GeoMult);

inc(StoreCount);

inc(YC);

end; (while YC) inc(XC);

YC := O;

1 334052 ~ end; ~while XC~ if XC &lt; XCount then begin writeln('Not enough entries in STEREO.GE0 file.');

halt(2);

end;

while not eof(GeoFile) do begin readln(GeoFile,temp);

if (length(temp)>0) and (temp[l]<>'!') then begin SkipToNum(temp);

if temp[1] in ['0'..'9','.'] then begin writeln('Too many entries in STEREO.GE0 file.');

halt(3);

end;

end;

end;

close(GeoFile);

end;

end;

for i:=l to 2 do Sensor[i].Adj := Sensor[i].Ref;

CorrectForDrift; ~ gain is l, offset is 0 ~ StereoSize := MemAbs(XGeoCorrectTable) + (LongInt(XCount*YCount) shl 3) - (LongInt(PrefixSeg) shl 4) + 4096;

StereoSize := ((StereoSize shr 4)+1) shl 4;

writeln('MemAbs XGeo=',MemAbs(XGeoCorrectTable));

writeln('MemAbs yGeo=~MemAbs(yGeocorrectTable));

writeln('X*Y shl 3=',10ngInt(XCount*YCount) shl 3);

writeln('PrefixSeg=',~ongInt(PrefixSeg));

writeln('StereoSize=',StereoSize);

writeln('STERE0 Device Driver installed using ', RealStr(StereoSize/1024),'K bytes.');

ShowGeometry;

InitMemoryResidency;

ResidentTermination;

ExitProgram:

end.

193 APPENDIX B 1 334052 3D Systems Stereolithoqraphy System Software CALOUT.PAS SLA-l Calibration Post-Proces~ing Program ______________________________________ _______________________ Recent History:

3/28/88 Ver 3.00 first released for production use.

5/16/88 Ver 3.10 upgraded to turbo pascal version 4.0 7/18/88 Ver 3.20 update VersionId 9/01/88 Ver 3.21 change 'Use temperature correction...' to 'Adjust calibration points to correct for measured mirror drift?' 9/13/88 Ver 3.22 allow escape if enter is input for filename _ _ _ _ } uses CRT,DOS;

const VersionId = '3.22';

tlo = 0;

thi = 7 4;

thiplusl = 75;

clo = 13;

chi = 61;

incint = 889;

CALdirectory = '\CAL\';

AnyFile = $3F;

type str25 = string[25];

str80 = string[80];

arrayptr = array[tlothi,tlo.... thi] of real;

var FirstVar : byte; ~ keep this in the first location DirInfo : qearchRec;

regs : registers;

FileOK : boolean;

i,j, rangeerror : integer;

key : char;

filel, file2, file3 : text;

DXstr, dateStr : str25;

InputStr, InputFilel, GEOfile, DRFfile, Astring : str80;

xl,yl,x2,y2 : array[tlo..thiplusl] of real;

x3,y3 : ^arrayptr;

x4, yq : real;

x3temp,y3temp : real;

xltot,x2tot,x3tot,x4tot,yltot,y2tot,y3tot,y4tot : real;

xof,xofd,yof,yofd,xqn,xqnd,yqn,yqnd : real;

xd,xdmax,xdmin,yd,ydmax,ydmin : real;

xldev,x2dev,yldev,y2dev : real;

ArrayIndexAlloc : Word;

LastVar : byte; ~ keep this in the last location ~$R+,S-,I+,F-,V-,B-~ procedure CursorOff;

begin with reqs do begin AH := 3;

L ~ BH := 0; 1 3 3 4 0 ~ 2 Intr($10,regs);

AH := l;

CH := CH or $20;

Intr($10,regs);

end;

end;

procedure CursorOn;

begin with regs do begin AH := 3;

BH := 0;

Intr($10,reg~);

AH := l;

CH := CH and $DF;

Intr($10,regs);

end;

end;

procedure ShowTop;

begin window(l,l,80,25);

ClrScr;

writeln;

writeln;

writeln(' CALOUT : SLA-l Calibration Adjustment Program ');

writeln(' version ',versionID);

writeln;

writeln;

end;

function RealVal(line: string;var codel: Integer):Real;

var code2 : integer;

num : real;

begin if line[l]='.' then line := '0' + line;

if (line[l]='-') and (line[2]='.') then begin line := '-0' + line;

for i:= 4 to length(line) do line[i-l]:=line[i];

line[length(line)] := ' ';

end;

val(line,num,codel);

if codel <> 0 then val(copy(line,l,codel-l),num,codel); { was code2 RealVal := num;

end;

procedure UserWait;

begin writeln;

writeln(' ... pre~s any key to C~nt; n~ repeat delay(l00) until keypressed;

key := ReadKey;

end;

procedure PromptForCalibFile;

begin ShowTop;

exec('c:\~ -~.com','/c dir \CAL\~.CAL');

writeln(' Existing Calibration Files: ');

write (' (none) ');

gotoxy(WhereX-16,WhereY);

i := O;

FindFirst('\CAL\~.CAL',anyfile,DirInfo);

~ DosError <> 0 then begin 1 3 3 4 0 5 2 writeln;

writeln;

UserWait;

Halt;

end;

while DosError = 0 do begin if (i - 4~ = 0 then begin writeln;

i := O;

end;

gotoxy(8+(15~i),WhereY);

i := i + l;

write(DirInfo.Name);

FindNext(DirInfo);

end;

writeln;

gotoxy(l,WhereY);

FileOK := false;

repeat writeln;

writeln(' Enter the calibration file in need of adjustment:');

CursorOn;

readln(InputFilel);

if InputFilel = '' then Halt(0); ~e~cape} for i := 1 to length(InputFilel) do InputFilel[i]:= upcase(InputFilel[i]);

i := pos('.',InputFilel);

InputStr := InputFilel;

if i > 0 then InputStr := copy(InputFilel,l,i-l);

if i = O then InputFilel := InputFilel + '.CAL';

InputFilel := CALdirectory + InputFilel;

InputStr := CALdirectory + InputStr;

writeln;

FindFirst(InputFilel,anyfile,DirInfo);

if DosError = 0 then FileOK := true;

if not FileOK then writeln(' Error: Invalid file name input.');

until FileOK;

GEOFile := InputStr + '.GEO';

DRFFile := InputStr + '.DRF';

end;

procedure GetCustData;

begin ShowTop;

writeln(' Program to convert data from Beam Profiler Calibration Run.');

writeln;

writeln(' Enter customer name and/or SLA-l serial number:');

readln(InputStr);

writeln(' Enter DX serial number:');

readln(DXstr);

writeln(' Enter date of calibration (mo/dy/yr):');

readln(DateStr);

end;

procedure FillUncorrectedTable;

var x3temp, y3temp : real;

begin ShowTop;

writeln(' Filling Uncorrected Table --------------------------------------------------');

for i:= tlo to thi do begin x3temp := i;

x3temp := incint ~ x3temp;

if x3temp&lt;5 then x3temp := 5;

if x3temp>65530 then x3temp := 65530;

1 334052 - for j := tlo to thi do x3^[i,j] := x3temp;

end;

for j:= tlo to thi do begin y3temp := j;

y3temp := incint * y3temp;

if y3tempc5 then y3temp := 5;

if y3temp>65530 then y3temp := 65530;

for i := tlo to thi do y3^[i,j] := y3temp;

end;

end;

procedure FileReadError;

begin writeln(' ERROR encounted in reading calibration file:');

writeln(' CALOUT execution aborted.');

UserWait;

Halt(l);

end;

procedure ParseToComma(var InString: string;var xoutput : real);

var i, valcode : integer;

begin i := pos(',',InString);

if i=0 then xoutput := RealVal(Instring,valcode) else begin xoutput := RealVal(copy(Instring,l,i-l),valcode);

InString := copy(InString,i+l,length(InString));

end;

end;

procedure R~T.;n~X3Y3(var x3,y3: real);

var i : integer;

begin repeat readln(filel,Astring);

until Astring[l]<>'!';

ParseToComma(Astring,x3);

ParseToComma(Astring,y3);

end;

procedure ReadFour(var x3,y3,x9,y4: real);

var i : integer;

begin repeat readln(filel,Astring);

until Astring[l]<>'!';

ParseToComma(Astring,x3);

ParseToComma(A~tring,y3);

ParseToComma(Astring,x9);

ParseToComma(Astring,y4);

end;

procedure ReadMeasuredValue~;

begin ShowTop;

writeln(' Reading Measured Values -------------- --------------');

~$I-~ assign(filel,InputFilel);

Reset(filel);

if (IOResult <> 0) then FileReadError;

for i:= clo to chi do begin ReadFour(xl[i],yl[i],x2[i],y2[i]);

for j := clo to chi do begin ~1 1 334~52 ~ ReadLineX3Y3(x3^[i,j],y3^[i,j]);

if (x3A[i,j]<8000) or (x3^[i,j]>57000) then begin writeln(' WARNING: X value out of range. X=',x3^[i,j]:6:2);

RangeError := -1;

end:

if (y3^[i,j]<8000) or (y3^[i,j]>57000) then begin writeln(' WARNING: Y value out of range. Y=',y3^[i,j]:6:2);

RangeError := -1;

end;

end;

end;

i := chi + l;

ReadFour(xl[i],yl[i],x2[i],y2[i]);

close(filel);

t$I+~ if RangeError <> 0 then UserWait;

end;

procedure AdjustMeasuredValues;

{ NOTE: we will assume 32767 is at center of range, to 1st order so we can use offset &amp; gain (of &amp; gn) independently First calculate offset &amp; gain for each set of readings, then take proportionate part of the ones for beginning &amp; end of the Y Axis (j) table v~.ent;

--assume drift evenly distributed over time intervals between readings, remember that profile measurements take time too.

Xnew = ( Xold - 32767 ) / Xgn + 32767 - Xof (how nice if numbers were centered on Zero!) begin Showrop;

writeln(' Adjusting Measured Values --------------- -----------------------');

for i := clo to chi do begin need the range of corrections from beginning &amp; end of Column } xof:=xl[i ; ~ start } xofd:=(xl i+l]-xl[i])/(chi-c1O+2); ~ delta } yof:=yl[i ; ~ etc... } yofd:=(yl i+l]-yl[i])/(chi-clo+2);

xgn:=x2[i ;

xgnd:=(x2[i+1]-x2[i])/(chi-clc+2};

ygn:=y2[i];

ygnd:=(y2[i+1]-y2[i])/(chi-clo+2);

for j := clo to chi do begin x3^[i,j]:=(x3^[i,j]-32767)/(xgn+xgnd~(j-clo+1))+32767-xof-xofd*(j-clo+l);

y3^[i,j]:=(y3^[i,j]-32767)/(ygn+ygnd~(j-clo+1))+32767-yof-yofd~(j-clo+l);

end;

end;

Adjusting Extrapolated Values -----------------do we really need to ? - yes for 12" pot at same level . . .

but should be calibrating 12 n by then: SO - DON'T ( Comment kept from the original basic program. ) } end;

procedure FindMirrorDrift;

begin ShowTop;

writeln(' Finding Mirror Drift -----------------------------------------------------');

for i:= clo to chi+l do begin xltot := xltot + xl i /(chi+l-clo+l ;

x2tot := x2tot + x2 i /(chi+l-clo+l ;

yltot := yltot + yl i /(chi+l-clo+ll;

y2tot := y2tot + y2 i /(chi+l-clo+ll;

end;

for i:= clo to chi+l do begin xd := xltot-xl[i];

.,, I ~I ~ ~ b D `- 1 334052 if xdmax&lt;xd then xdmax:=xd;

if xdmin>xd then xdmin:=xd;

xldev:=xldev+xd*xd/(chi+l-clo+l~;

xd:=x2tot-x2[i];

if xdmax&lt;xd then xdmax:=xd;

if xdmin>xd then xdmin:=xd;

x2dev:=x2dev+xd*xd/(chi+l-clo+l);

yd:=yltot-yl[i];

if ydmax&lt;yd then ydmax:=yd;

if ydmin>yd then ydmin:=yd;

yldev:=yldev+yd*yd/(chi+l-clo+l);

yd:=y2tot-y2[i];

if ydmax&lt;yd then ydmax:=yd;

if ydmin>yd then ydmin:=yd;

y2dev:=y2dev+yd*yd/(chi+l-clo+l);

end;

~$I-} a~sign(file2,DRFfile);

rewrite(file2);

{ begin ~ writeln(file2,'!CAL0UT ',versionID,' Drift Report File');

writeln(file2,'!SL~-1 ',inputstr,' DX ~erial # ',dxstr,' on ',datestr);

writeln(file2,'!Position values used for calibration, &amp; RMS deviations (bits)');write (file2,'!Sensorl ');

writeln(file2,xltot:6:2,' ',yltot:6:2,' ',SQRT(xldev):6:2,' ',SQRT(yldev):6:2);

write (file2,'!Sensor2 ');

writeln(file2,x2tot:6:2,' ',y2tot:6:2,' ',SQRT(x2dev):6:2,' ',SQRT(y2dev):6:2);

writeln(file2,' xltot x2tot yltot y2tot (bits) ');

writeln(file2,' ',xltot:6:2,' ',x2tot:6:2,' ',yltot:6:2,' ',y2tot:6:2);

writeln(file2,' xldev x2dev yldev y2dev (bits RMS)');

writeln(file2,' ',SQRT(xldev):6:2,' ',SQRT(x2dev):6:2,' ',SQRT(yldev):6:2,' ',SQRT(y2dev):6:2);

writeln(file2);

writeln(file2,' xl[i] x2[i] yl[i] y2[i] (bits) ');

for i:=clo to chi+l do begin writeln(file2,' ', (xl[i]-xltot):6:2,' ',(x2[i]-x2tot):6:2,' ',(yl[i]-yltot):6:2,' ', (y2[i]-y2tot):6:2);

end;

end; } ~ re-work the drift information for I:=clo to chi+l do begin xof:=((xl~i]+x2~i])-(xltot+x2tot))/2; ~ bits ~ yof:=((yl[i]+y2[i])-(yltot+y2tot))/2; ~ bits } xgn:=(xl[i]-x2[i])/(xltot-x2tot); { approximately = 1 ygn:=(yl[i]-y2[i])/(yltot-y2tot); ~ approximately = 1 } { re-u.~e the xl() etc. arrays } xl[i::=xof;

x2[i :=xgn;

yl[i::=yof;

y2[i :=ygn;

end, writeln(file2);

writeln(file2,' X offset X gain Y offset Y gain (bits, bits/10")');

writeln(file2);

for i:= clo to chi+l do begin writeln(file2,' ', xl[i]:7:1,' ',((x2[i]-1)*35560):7:1,' ',yl[i]:7:1,((y2[i]-1)*35560):7:1);

end;

close(file2);

{$I+) if ydmax&lt;yd then ydmax:=yd;

if ydmin>yd then ydmin:=yd;

xd := xdmax-xdmin;

yd := ydmax-ydmin;

i~ xd&lt;yd then xd:=yd; ~ select largest drift ) 1 334052 writeln;

writeln(' From measuring the apparent position of the beam profile sensors');

writeln(' during calibration, we find total effective mirror drift');

writeln(' ',round(xd/3.556),' mils ');

UserWait;

writeln;

if xd>500 then begin writeln(' Program found wrong profiler pinholes,');

writeln;

writeln(' File will be written without attempting to correct for drift');

UserWait;

exit;

end;

if xd>35 then begin writeln(' WARNING: Excessive drift detected. ');

writeln(' Possible causes are temperature drift, ');

writeln(' direct air currents from heating or cooling ');

writeln(' vents, and/or insufficient warmup time. ');

writeln;

if xd>100 then begin writeln(' File will be written without attempting to correct for drift');

exit;

end;

writeln(' Adjust calibration points to correct for measured mirror drift ?');

repeat delay(l00) until keypressed;

Rey := upcase(ReadKey);

if Key = 'Y' then AdjustMeasuredValues;

end;

end;

procedure WriteGEOFile;

begin Show~op;

writeln(' Writing Corrected File: ',GEOfile,' ---------------------------------');

~$I-} assign(file3,GEOfile);

rewrite(file3);

writeln(file3,'!SLA-l ',inputstr,' DX qerial # ',DXStr,' on ',datestr);

writeln(file3,'!Poqition valueq used for calibration, ~ RMS deviations (bits)');write (file3,'!Sensorl ');

writeln(file3,xltot:6:2,' ',yltot:6:2,' ', SQRT(xldev):6:2,' ',SQRT(yldev):6:2);

write (file3,'!Sensor2 ');

writeln(file3,x2tot:6:2,' ',y2tot:6:2,' ', SQRT(x2dev):6:2,' ',SQRT(y2dev):6:2);

writeln(file3,thi+1,',',thi+1,',2,',incint);

for i := tlo to thi do begin for j := tlo to thi do begin ( str(trunc(x3^[i,j]~10+0.5)/10,Astring);

Astring := copy(Astring,2,length(Astring));

write (file3,Astring,',');

qtr(trunc(y3^[i,j]~10+0.5)/10,Astring);

Astring := copy(Astring,2,length(Astring));

writeln(file3,A_tring); ) writeln(file3,((x3^[i,j]~10+0.5)/10):7:1,',', ((y3^[i,j]~10+0.5)/10):7:1);

end;

end;

writeln;

close(file3);

~I+~ end;

1~ pro~_~ure FinalWritelns; 1 3 3 4 0 S 2 ShowTop;

writeln(' Calibration correction complete.');

writeln;

writeln(' Two files have been generated: ',DRFfile,' and ',GEOfile,'.');

writeln;

writeln(' ',DRFfile,' has the profiler positions measured after each');

writeln(' column of the grid.');

writeln;

writeln(' ',GEOfile,' has the calibration data which will become');

writeln(' \3DSYS\STEREO.GEO if selected during the "install new');

writeln(' geometric correction table" option of program BEAM. ');

writeln;

writeln(' CALOUT program: normal termination.');

UserWait;

end;

begin { main program ) FillChar(FirstVar,ofs(1astVar)-ofs(FirstVar),#0); { clear variables ) ArrayIndexAlloc := Sqr(thi-tlo+1) * 6; { reals have 6 bytes } writeln('memavail =',Mem~vail,' Arrayin~PxAlloc=~arrayin~pxAlloc);

if (Mem~vail - 32768) &lt; (2*ArrayIndexAlloc) then begLn writeln(' ERROR: Insufficient memory available for CALOUT.');

Halt(1);

end;

GetMem(x3,ArrayIndexAlloc);

GetMem(y3,ArrayIndexAlloc);

PromptForCalibFile; ( Get user calibration file. ) GetCustData; ( Get user information such as serial number and date. ) FillUncorrectedTable; ( MAnllAlly fill the 74 x 74 real array with defaults. ) RPa~MeA~uredValues; ( Read calibration data from input calib. file. ~ FindMirrorDrift; ( *** Find the deviation between sensor readings. } Writer-E~file;

{ Build corrected .GEO file by altering the input from the calibration file. } FinalWritelns; ( Closing proceedings. } end.

( ADDITIONAL NOTES:

1.

The calibration table has a sensor every 889 bits ( 0.2500").

2.

The nearest sensor above 65536 is 74 x 889 = 65786 = 65535 + 251, thus our table goes from 0 to 74.

3.

The neare~t reading to center is 37 x 889 = 32767 + 126 (0.0354"), thus our readings (-6 to 6n) are from array locations 13 to 61.

4.

The X-axis is reversed, it is corrected by reading X from 6 to -6.

5.

Block zeros over variable declaration emulates BASIC setting all variables to zero as they are introduced.

6.

Pointers were required due to PASCAL limiting the number of variable definitions. ******~***~********* ) . A~?PENDIX C - 1 334052 3D Systems Stereolithography System Software BEAMCALI.INC SLA-1 Calibration Process Driver for the BEAM program ______________________________________________________________________________ procedure UserWait;

begin writeln;

writeln;

writeln(' ... press any key to cont;n repeat delay(l00) until KeyPressed;

junkkey:=Readkey;

end;

{ ****************************** * Option 8: Calibrate System * ****************************** } procedure Calibr8;

const CalibHoleSeparation = 890;

DEFAULTchs = 890; ~ 3556 vat units per inch * 1/4 inch = 889 } XVatCenter = 32767i YVatCenter = 32767;

{ chs = calibration hole separation } NormSpiralDelay = 100;

NormColsDelay = 10;

NormRowsDelay = 60;

var CalibKey : Char;

i,code,CalibAttempts: Integer;

tableX,tableY,FullPart,ADC: Real;

RowStartXsp,RowStartYsp: LongInt;

row,col,Xcount,Ycount: Integer;

CalibCoords: array[1..50,1..50,1..2] of Real;

CalibTimeout,GoodRow,GoodCol,~TT~OT~OK: Boolean;

SpiralDelay: Word; { in milliseconds } BetweenRowsDelay,BetweenColsDelay: Word; { in tenths of seconds } temp,Indent: String;

passOption,passReturnC: integer;

passColX,passRowY : integer;

Xl,Yl,X25,Y25,oldXsp,oldYsp : longint;

Xchsl,Xchs5,XchslO,Xchsl5,Xchs20,{Xchs25 is default} Xchs30,Xchs35,Xchs40,Xchs45,Xchs49:integer;

Ychsl,Ychs5,YchslO,Ychsl5,Ychs20,{Ychs25 is default} Ychs30,Ychs35,Ychs40,Ychs45,Ychs49:integer;

{ label AbortCalib; } label RePCP; { Repeat Process Calibration PinHoles } procedure CalibrationDelay(time{in 1/10 secs}: LongInt);

var sec,TotalSec: LongInt;

begin if time > 0 then begin delay(999); { keep the calibration hole message up for a second } ClearLine(23);

gotoxy(33,23);

sec := 2; { start at two seconds to account for first delay } TotalSec := time div 10;

TotalSec > 0 then write('Waiting ... '); 7 ~ A ~ C ~~~hile (sec <= TotalSec) and not keypressed do begin gotoxy(95,23);

write(TotalSec + 1 - sec,' ');

delay(999);

inc(sec);

end;

delay(100 * time mod 10);

while keypressed do key := ReadKey;

ClearLine(24);

end;

end;

function StartCalibFile: Boolean;

beqin StartCalibFile := true;

assign(CalOutFile,CalibFileName);

{$I-} rewrite(CalOutFile);

if IOresult<>0 then begin Center('Cannot write to Calibration File',24);

WaitKey;

StartCalibFile := false;

exit;

end;

{$I+} writeln(CalOutFile,'!

Calibration File V' + VersionId + ', started ',TimeStamp);close(CalOutFile);

end;

function UpdateCalibFileRow: Boolean;

var i: Integer;

Folln~Rei 1,FoundBeam2: Boolean;

SensorPosX,SensorPosY: array[1..6] of Real;

begin UpdateCalibFileRow := true;

assign(CalOutFile,CalibFileName);

{$I-} append(CalOutFile);

if IOresult<>0 then begin Center('Error writing to Calibration File',24);

WaitKey;

UpdateCalibFileRow := false;

exit;

end;

{$I+~ MoveBeam(trunc(Xsp[1]),trunc(Ysp[1]));

CalibrationDelay(BetweenRowsDelay);

SetupFindBeam(FindBeamVariables,1{Sensor},-1{Xsp[1]},-1{Ysp[1]},40{SpiralSideMax}, 1{SpiralTries},4{ScanStep},trunc(Bk[1]+0.5){BkgndThreshold}, 20{SpiralDelay},1{NumCenterPasses});

FoundBeaml := FindBeam(FindBeamVariables,1,NormalFind,0{no mssgs});

SensorPosX[1] := FindBeamVariables[1].Xavg;

SensorPosY[1] := FindBeamVariables[1].Yavg;

MoveBeam(trunc(Xsp[2]),trunc(Ysp[2]));

CalibrationDelay(BetweenRowsDelay);

SetupFindBeam(FindBeamVariables,2{Sensor),-1{Xsp[2]},-1{Ysp[2]},40{SpiralSideMax}, 1{SpiralTries},4{ScanStep},trunc(Bk[1]+0.5){BkgndThreshold}, 20{SpiralDelay},1{NumCenterPasses}};

FoundBeam2 := FindBeam(FindBeamVariables,2,NormalFind,0{no mssgs});

SensorPosX[2] := FindBeamVariables[2].Xavg;

SensorPosY[21 := FindBeamVariables[2].Yavg;

for i:=3 to 6 do begin SensorPosX[i] := 0;

73 SensorPosY[i] := 0;

l;

not (FoundBeaml and FoundBea~2) then write(CalOutFile, T ); 1 7 7 A for i:=1 to 6 do begin I ~ ~ ~ 0 5 2 if i > 1 then write(CalOutFile, , );

write(CalOutFile,RealStr(SensorPosX[i],2), , , RealStr(SensorPosY[i],2));

end;

writeln(CalOutFile);

close(CalOutFile);

end;

function UpdateCalibFileCol: Boolean;

var Resist,Temper: Real;

begin UpdateCalibFileCol := true;

ReadTemperature(Resist,Temper);

assign(CalOutFile,CalibFileName);

{$I-} append(CalOutFile);

if IOresult<>0 then begin Center( Error writing to Calibration File ,24);

WaitKey;

UpdateCalibFileCol := false;

exit;

end;

~$I+) GetTimeStamp;

if CalibTimeout then write(CalOutFile, T );

writeln(CalOutFile,RealStr(Xsp[Sensor],2), , ,RealStr(Ysp[Sensor],2), , , RealStr(Power[Sensor],2), , ,RealStr(Temper,2), , ,CurrentTime);

close(CalOutFile);

CalibTimeout := false;

end;

function QuickCalibSearch: Boolean;

begin SetupFindBeam(FindBeamVariables,Sensor,Xsp[Sensor],Ysp[Sensor], 40(SpiralSideMax},l{SpiralTries~,16~ScanStep},trunc(Bk~Sensor]+0.5), SpiralDelay,3(NumCenterPasses));

if SpiralSearchForHole(2,FindBeamVariables,Sensor) then begin QuickcAlihsearch := true;

Xsp[Sensor] := FindBeamVariables[Sensor].Xval;

Ysp[Sensor] := FindBeamVariable~[Sensor].Yval;

end else QuickCalibSearch := false;

end;

function CalibFind: Boolean;

( CalibFind starts its search at Xsp[Sensor], Ysp[Sensor], and resets these values when the beam is found. } begin SetupFindBeam(FindBeamVariables,Sensor,Xsp[Sensor],Ysp[Sensor], 40(SpiralSideMax~,3(SpiralTries},9(ScanStep},trunc(Bk[Sensor]+0.5), SpiralDelay,2~NumCenterPasses});

if FindBeam(FindBeamVariables,Sensor,NormalFind,2{no mssgs,noverify~) then beginCalibFind := true;

Xsp[Sensor] := trunc(FindBeamVariables[Sensor].Xavg+0.5);

Ysp[Sensor] := trunc(FindBeamVariables[Sensor].Yavg+0.5);

end else CalibFind := false;

end;

procedure UserNait;

begin writeln;

writeln;

writeln(' press any key to continue '):

~peat delay(100) until KeyPressed;

nkkey:=Readkey;

end; 1 3 3 4 0 5 2 procedure GetBackgroundNoiseLevels(Xcoord,Ycoord:longint);

begin with BeamProfileVariables[Sensor] do begin SensorBkgnd[Sensor] := BkReadFunction(Sensor,Xcoord,Ycoord);

( true background } BkgndThreshold := trunc(Bk[Sensor]+0.5); { high level mark { Bk[Sensor] is used by FindBeamVariables end;

end;

procedure CenterCalibPlate;

var i : integer;

~h~ : boolean;

begin Xsp[Sensor] := XVatCenter;

Ysp[Sensor] := YVatCenter;

MoveBeam(XVatCenter,YVatCenter);

ClrScr;

gotoxy(1,8);

writeln(' 1.

The beam is positioned in the middle of the vat (32767,32767).');

writeln;

writeln(' 2.

Center the calibration plate by the using beam position.'~;

writeln(' Position plate with mirrored section centered at 7.25" x 7.25n.');

writeln;

writeln(' 3.

Find the actual coordinates of the center pinhole by pressing');

writeln(' ENTER to begin the search.');

writeln;

writeln(' ( or press X to exit )');

repeat delay(100) until KeyPressed;

junkkey := readkey;

writeln;

writeln(' Please wait while the centerhole is located. ');

~N~K~ : = false;

while not ~ do begin Xsp[Sensor] := XVatCenter;

Ysp[Sensor] := YVatCenter;

MoveBeam(XVatCenter,YVatCenter);

if CalibFind then begin ClrScr;

gotoxy(1,9);

writeln(' *** Calibration plate center hole located at: ',Xsp[Sensor]:5:0,' writeln;

if (Abs(32767-Xsp[Sensor]) &lt; 80) and (Abs(32767-Ysp[Sensor]) &lt; 80) then begin writeln(' These values are within the acceptable tolerances for the calibration');

writeln(' plate center hole.');

~ KE~ := true;

userwait;

end else begin writeln(' These values are not withing the acceptable tolerances for the ')';

writeln(' calibration plate center hole (32767 +/- 80 for X and Y).');

writeln;

writeln(' Take the following corrective action(s):');

if (32767-Xsp[Sensor] > 80) then writeln(' X coordinate too small - slide plate to the left. ');

if (Xsp[Sensor]-32767 > 80) then writeln(' X coordinate too large - slide plate to the right.');

if (Ysp[Sensor]-32767 > 80) then writeln(' Y coordinate too large - slide plate forward.');

if (32767-Ysp[Sensor] > 80) then writeln(' Y coordinate too small - slide plate backwards.');

userwait;

1 334052 w~iteln;

writeln(' Please wait while the centerhole is relocated. ');

endi end else begin writeln('~ Center Hole not found.

Reposition plate and/or increase gain. ~*~');

userwait;

writeln;

writeln(' Please wait while the centerhole is relocated. ');

end;

end; ~ while not X25 := trunc(Xsp[Sensor]); { positions for hole(25,25) Y25 := trunc(Ysp[Sensor]);

end; ~ procedure CenterCalibPlate ~ procedure AdjustGain(keychar:char);

{ 'C' is center hole, otherwise do not reset X25 and Y25.

var key: Char;

v,code : integer;

leave, newsensor, goodgain : boolean;

begin goodgain := false;

ClrScr;

gotoxy(l,9);

writeln(' The next step is adjusting the gain on the calibration plate:');

writeln;

writeln(' Raise the gain if the MAXIMUM reading is below 40.');

writeln(' Lower the gain if the MAXIMUM reading is above 200.');

writeln;

writeln(' Press any key to leave the profiler after the gain is set.');

userwait;

OpMode := GraphicsOp;

Sensor := 3;

Display := true;

repeat Leave := false;

TextMode(C80);

SetScreen;

gotoxy(1,25);

write(' GATHERING PROFILE for Sensor 3 (Calibration Plate), press Q to Quit');

NewSensor := false;

repeat Display := true;

BeamProfileVariables[Sensor].ScanStep := 2;

GoodRead(l);

if TotWtAvg=0 then begin GoodGa;n := false;

Center(' ERROR: Sensor #' +IntStr (Sensor) + ' was not found: Increase gain and press ENTER to continue.',24);

WaitKey;

end;

if GoodCount > 0 then begin GoodGain := true;

Xsp[Sensor] := Xaverage;

Ysp[Sensor] := Yaverage;

if keychar = 'C' then begin X25 := trunc(Xsp[Sensor]);

Y25 := trunc(Ysp[Sensor]);

end;

end;

PrintInfo(Sensor);

if keypressed then begin key := ReadKey;

v := IntVal(key,code~;

if (v > 0) and (v <= NumSensors) then begin Sensor := v;

MoveBeam(trunc(Xsp[Sensor]),trunc(Ysp[Sensor])); 1 3 3 4 0 ~ 2 ~ delay(200); J NewSensor := true;

end else Leave := true;

end;

until NewSensor or Leave;

until Leave and GoodGain;

Display := false;

TextMode(C80);

ClrScr;

end; ( procedure AdjustGain ) procedure PutCalibHoleSeparation(XHoleNum,YHoleNum,XchsLast,YchsLast, NumSteps:integer; Axis:char; var XchsNew,YchsNew:integer);

Determines the calibration hole separation from the average of NumSteps holes for one axis v~ t. } var SaveX,SaveY: LongInt; ( coordinates of pinhole SumXchs,SumYchs,i : integer;

AXISX : boolean;

begin if not BORDERSOK then exit;

SumXchs := 0;

SumYchs := 0;

XchsNew := XchsLast;

YchsNew := YchsLast;

SaveX := trunc(Xsp[Sensor]);

SaveY := trunc(Ysp[Sensor]);

if Axis = 'X' then AXISX := TRUE else AXISX := FALSE;

for i := 1 to NumSteps do begin if (XHoleNum &lt; 25) or (YHoleNum &lt; 25) then i := -i;

if AXISX then begin write(' Pass for hole ',X8OleNum+i,',',YHoleNum,':');

SaveX := trunc(Xsp[Sensor]);

end else begin write(' Pass for hole ',XHoleNum,',',YHoleNum+i,':');

saveY := trunc(Ysp[Sensor]);

end;

if i &lt; 0 then begin ( stepping down towards 1,1 Xsp[Sensor] := SaveX - XchsNew;

Ysp[Sensor] := SaveY - YchsNew;

end else begin Xsp[Sensor] := SaveX + XchsNew;

Ysp[Sensor] := SaveY + YchsNew;

end;

if CalibFind then begin writeln(' FOUND at ',Xsp[Sensor]:6:0,',',Ysp[Sensor]:6:0);

if AXISX then begin XchsNew := abs(trunc(Xsp[Sensor] - SaveX));

SumXchs := SumXchs + XchsNew;

end else begin YchsNew := abs(trunc(Ysp[Sensor] - SaveY));

SumYchs := SumYchs + YchsNew;

end;

end else begin writeln;

writeln(' ~ ERROR : calibration plate hole not found');

writeln(' Borders of plate are not established. ');

BORDERSOR := false;

userwait;

writeln(' Please wait... ');

exit;

end;

i := abs(i);

end;

if AXISX then XchsNew := abs(SumXchs div NumSteps) ~ 7~ else YchsNew := abs(SumYchs div Nu~Steps); 4 e I procedure PutCalibHoleSeparation } I ~ J U J L procedure MakeRoughGeoCorrect;

{ X and Y coordinates for center hole 25,25 are established.

Now establish coordinates for the remainder of the holes in increments of 5.

var junk : integer;

begin BORDERSOK := true; { until proven false ) Xchs30 := 0;

Ychs30 := 0;

ClrScr;

writeln;

writeln(' Establish borders of calibration plate:');

writeln;

{ The beam is at the center hole from procedure adjust gain. ) { X walk up border ~ PutCalibHoleSeparation(25,25,DEFAULTchs,0,5,'X',Xchs30, unk);

PutCalibHoleSeparation(30,25, Xchs30,0,5,'X',Xchs35, unk);

PutCalibHoleSeparation(35,25, Xchs35,0,5,'X',Xchs40, unk);

PutCalibHoleSeparation(90,25, Xchs40,0,5,'X',Xchs45, unk);

PutCalibHoleSeparation(45,25, Xchs45,0,4,'X',Xchs49, unk);

Xsp[Sensor] := Xsp[Sensor] + Xchs49; { one past border ) if CalibFind and BORDERSOK then begin BORDERSOK := false;

writeln(' *** ERROR : Cannot determine X border. ');

writeln(' Check calibration plate positioning. ');

userwait;

exit;

end;

{ Y walk up to border on the left side ) Xsp[Sensor := X25;

Ysp[Sensor: := Y25;

PutCalih~o eSeparation(25,25,0,DEFAULTchs,5,'Y', unk,Ychs30);

PutCalibHoleSeparation(25,30,0, Ychs30,5,'Y', unk,Ychs35);

PutCalibHoleSeparation(25,35,0, Ychs35,5,'Y', unk,Ychs40);

PutCalih~oleSeparation(25,40,0, Ychs40,5,'Y', unk,Ychs45);

PutCalih~nleSeparation(25,45,0, Ychs45,4,'Y', unk,Ychs49);

Ysp[Sensor] := Ysp[Sensor] + Ychs49; { one past border if CalibFind and BORDERSOK then begin BORDERSOK := false;

writeln(' *** ERROR : Cannot determine Y border. ');

writeln(' Check calibration plate positioning. ');

userwait;

exit;

end;

{ X walk down to the front border ) Xsp[Sensor] := X25 - Xchs30;

Ysp[Sensor] := Y25;

PutCalibHoleSeparation(24,25,DEFAULTchs,0,4,'X',Xchs20,~unk);

PutC~lih~QleSeparation(20,25, Xchs20,0,5,'X',XchslS, unk);

PutCalibHoleSeparation(15,25, Xchsl5,0,5,'X',XchslO, unk);

PutC~lih~oleSeparation(10,25, XchslO,0,5,'X', XchsS, unk);

PutCalibHoleSeparation( 5,25, XchsS,0,4,'X', Xchsl, unk);

( Y walk down to the right side : NOTE - X is not reset to center ) Ysp[Sensor] := Y25 - Ychs30;

PutCalibHoleSeparation(1,24,0,DEFAULTchs,4,'Y',~unk,Ychs20);

PutCalibHoleSeparation(1,20,0, Ychs20,5,'Y', unk,YchslS);

PutCalibHoleSeparation(1,15,0, YchslS,5,'Y', unk,YchslO);

PutCalibHoleSeparation(1,10,0, YchslO,5,'Y', unk, YchsS);

PutCalib~HoleSeparation(1, 5,0, YchsS,9,'Y',junk, Ychs1); 1 334052 ~-writeln;

if BORDERSOK then begin X1 := trunc(Xsp[Sensor]); ~ Save the coordinates of hole 1,1. ) Y1 := trunc(Ysp[Sensor]);

writeln(' Borders of Calibration Plate are established. ');

writeln;

end;

end;

function GetCalibHoleSeparation(axis:char;aisle:integer):integer;

var ivalue : integer;

begin if axis = 'X' then begin if aisle ~ 5 then ivalue := Xchs1 else if aisle < 10 then ivalue := Xchs5 else if aisle < 15 then ivalue := XchslO else if aisle < 20 then ivalue := Xchsl5 else if aisle < 25 then ivalue := Xchs20 else if aisle > 45 then ivalue := Xchs99 else if aisle > 40 then ivalue := Xchs95 else if aisle > 35 then ivalue := Xchs90 else if aisle > 30 then ivalue := Xchs35 else if aisle >= 25 then ivalue := Xchs30;

end else begin { axis = 'Y' ) if aisle < 5 then ivalue := Ychs1 else if aisle < 10 then ivalue := Ychs5 else if aisle < 15 then ivalue := YchslO else if aisle < 20 then ivalue := YchslS else if aisle < 25 then ivalue := Ychs20 else if aisle > 45 then ivalue := Ychs49 else if aisle > 40 then ivalue := Ychs45 else if aisle > 35 then ivalue := Ychs40 else if aisle > 30 then ivalue := Ychs35 else if aisle >= 25 then ivalue := Ychs30;

end;

GetCalibHoleSeparation := ivalue;

end; { function GetCalibHoleSeparation } procedure ProcessCalibrationPinHoles(var iOption,iReturn,Rowx,Colx: integer);

{ iOption = 0 is regular calibration.

= 1 is a quick scan.

iReturn = 0 is all holes found = 1 could not locate a pinhole = 2 pinhole ADC reading over saturation level iReturn <> 0 returns BadX and BadY } { Calibration:

Alg moves out each row (Y-axis), then out each column (X-axis) lookingfor a hole. The current alg looks for a set of 49 x 49 holes. For each position, whether it finds a hole or not, it makes an entry in the CA~IB.DAT file. Entries in this file for holes for which the laser beam wasn't found after 10 attempts are preceded by a 'T '.

} var row,col : integer;

begin if iOption = 0 then if not StartCalibFile then exit;

iReturn := 0;

CalibTimeout := false;

Xsp[Sensor] := X1; { Start at established corner hole 1,1.} Ysp[Sensor] := Y1;

..~ for row~:= 1 to 49 do begin 1 3 3 4 ~ ~ 2 GoodRow := CalibFind;

~ if iOption = 0 then if not UpdateCalibFileRow then exit;

MoveBeam(trunc(Xsp[Sensor]),trunc(Ysp[Sensor]));

if iOption = 0 then CalibrationDelay(BetweenRowsDelay);

for col := 1 to 49 do begin colx := col;

if AbortFlag then exit;

oldXsp := trunc(Xsp[Sensor]);

oldYsp := trunc(Ysp[Sensor]);

CalibAttempts := 0;

repeat GetBackgroundNoiseLevels(trunc(Xsp[Sensor])-445,trunc(Ysp[Sensor])-445);

( Xsp[Sensor] and Ysp[Sensor] are used to position CalibFind.

GoodCol := CalibFind;

if AbortFlag then exit;

if not GoodCol then begin inc(CalibAttempts);

beep;

iReturn := 1;

if iOption = 0 then Center('Calibration hole ('+IntStr(row)+','+IntStr(col)+ ') not found.

Dust?',24);

if CalibAttempts &lt; 10 then GoodCol := false else begin GoodCol := true; ~ go on after 10 attempts ) CalibTimeout := true; ~ set calibration timeout flag } end;

Xsp[Sensor] := oldXsp;

Ysp[Sensor] := oldYsp;

end;

if GoodCol then begin if col = 1 then begin RowStartXsp := trunc(Xsp[Sensor]);

RowStartYsp := trunc(Ysp[Sensor]);

end;

Ysp[Sensor] := Ysp[Sensor] + GetCalibHoleSeparation('Y',col);

end;

until GoodCol;

if iOption = 0 then if not UpdateCalibFileCol then exit;

writeln('Calibration hole ('+IntStr(row)+','+IntStr(col)+')');

if col &lt; 49 then begin MoveBeam(trunc(Xsp[Sensor]),trunc(Ysp[Sensor]));

if iOption = 0 then CalibrationDelay(BetweenColsDelay);

end;

if (iOption = 1) and (iReturn = 1) then begin RowX := row;

ColY := col;

exit;

end;

end; {col~ XsptSensor] := RowStartXsp;

if row &lt; 49 then begin oldXsp := trunc(Xsp[Sensor]);

oldYsp := trunc(Ysp[Sensor]);

CalibAttempts := 0;

repeat Xsp[Sensor] := Xsp[Sensor] + GetCalibHoleSeparation('X',row);

Ysp[Sensor] := RowStartYsp;

if iOption = 0 then GetBackgroundNoiseLevels(trunc(Xsp[Sensor])-445,trunc(Ysp[Sensor])-445);

~ Xsp[Sensor] and Ysp[Sensor] are used to position CalibFind. ) GoodRow := CalibFind;

"~ i~ AbortFlag then exit;

if not GoodRow t~en begin 1 3 3 4 0 5 2 ~ inc(CalibAttempts);

beep;

iReturn := 1;

if iOption = 0 then Center('Can''t locate row '+IntStr(row)+'.

Dust?',29);

if CalibAttempts &lt; 10 then GoodCol := false else begin GoodRow := true; { go on after 10 attempts ) CalibTimeout := true; { set calibration timeout flag ) end;

Xsp[Sensor] := oldXsp;

Ysp[Sensor] := oldYsp;

end;

until GoodRow;

end;

if (iOption = 1) and (iReturn = 1) then begin RowX := row;

ColY := col;

exit;

end;

end; {row) if iOption = 0 then begin if not UpdateCalibFileRow then exit;

Center('Calibration completed.',24);

end;

WaitKey;

exit;

end; {procedure ProcessCalibrationPinHoles ) {*********************** BEGIN MAIN CALIBRATION PROGR~M *********************) begin window (1,1,80,25);

ClrScr;

Calibrate := true;

GetTimeStamp;

CalibFileName := '\CAL\' + FNTimeStamp + '.CAL';

t Center('Beam Calibration Function',1); } gotoxy(1,8);

Indent := many(' ',16);

writeln(Indent,'Type Fast Medium Slow');

writeln(Indent,'-------------- ------ ------ ------');

writeln(Indent,'Spiral ',CalibDelays[1,1]:4,' ', CalibDelays[2,1]:4,' ',CalibDelays[3,1]:4,' ms');

writeln(Indent,'Between Points ',CalibDelays[1,2]/10:5:1,' ', CalibDelays[2,2]/10:5:1,' ',CalibDelays[3,2]/10:5:1,' secs');

writeln(Indent,'Between Rows ',CalibDelays[1,3]/10:5:1,' ', CalibDelays[2,3]/10:5:1,' ',CalibDelays[3,3]/10:5:1,' secs');

writeln(Indent,' ----- ----- -----');

writeln(Indent,'TOTAL ', (2.5*10*49*49+CalibDelays[1,2]*49*49+CalibDelays[1,3]*49*3)/36000:5:1,' ', (2.5*10*49*49+CalibDelays[2,2]*49*49+CalibDelays[2,3]*49*3)/36000:5:1,' ', (2.5*10*49*49+CalibDelays[3,2]*49*49+CalibDelays[3,3]*49*3)/36000:5:1,' hours');

writeln;

Center('Fast, Medium, Slow or Different Calibration?',18);

CalibKey := upcase(ReadKey);

case CalibKey of 'F': begin SpiralDelay := CalibDelays[1,1]; in milliseconds BetweenColsDelay := CalibDelays[1,2]; in tenths of seconds BetweenRowsDelay := CalibDelays[1,3];

end;

'M': begin SpiralDelay := CalibDelays[2,1];

BetweenColsDelay := CalibDelays[2,2]; 3 3 4 0 5 2 ~ BetweenRowsDelay := CalibDelays[2,3];

end;

'S': begin SpiralDelay := CalibDelays[3,1];

BetweenColsDelay := CalibDelays[3,2];

BetweenRowsDelay := CalibDelays[3,3];

end;

'D': begin writeln;

writeln;

write('Enter Spiral Delay (in ms): ');

readln(temp);

if temp='' then exit else SpiralDelay := LongVal(temp,code~;

write('Enter Between Points Delay (in secs): ');

readln(temp);

if temp = '' then exit else BetweenColsDelay := trunc(RealVal(temp,code)*10+0.5);

write('Enter Between Rows Delay (in secs): ');

readln(temp);

if temp = '' then exit else BetweenRowsDelay := trunc(RealVal(temp,code)*10+0.5);

end;

else exit;

end;

SetScreen;

case CalibKey of 'F': Center('Fast Beam Calibration, ~tarted '+TimeStamp,1);

'M': Center('Medium Beam Calibration, started '+TimeStamp,1);

'S': Center('Slow Beam Calibration, started '+TimeStamp,1);

'D': Center('Different Beam Calibration, started '+TimeStamp,1);

end;

Search7;

Display := false;

Sensor := 3;

with FindBeamVariables[Sensor] do begin TextMode(C80);

GetBackgroundNoiseLevels(Xrest,Yrest);

CenterCalibPlate;

AdjustGain ('C');

MakeRoughGeoCorrect; ( Establishes borders and builds separation table end;

TextMode(C80);

ClrScr;

if not BORDERSOK then begin writeln;

writeln(' ***** Aborting Calibration ***** ');

userwait;

exit;

end;

RePCP: {nix me} (**** passOption :=1; ( quick check for pinholes } repeat RePCP: window(1,1,80,25};

ClrScr;

gotoxy(1,12);

writeln(' Calibrate: Quick search for all pinholes');

writeln;

writeln(' Please wait for twenty minutes to verify');

writeln(' that all pinholes in the calibration plate ');

writeln(' can be located and that no readings are above');

4~ u wr-teln(' the saturation level.');

writeln;

ClrScr; I 3 3 4 0 5 2 writeln(' Xchs01=',Xchsl );

writeln(' Xchs05=',Xchs5 );

writeln(' Xchs10=',Xchs10 );

writeln(' Xchs15=',Xchs15 );

writeln(' Xchs20=',Xchs20 );

writeln(' Xchs30=',Xchs30 );

writeln(' Xchs35=',Xchs35 );

writeln(' Xchs40=',Xchs40 );

writeln(' Xchs45=',Xchs45 );

writeln(' Xchs49=',Xchs49 );

writeln(' XchsO1=',Xchsl );

writeln(' Xchsl5=',Xchs5 );

writeln(' Xchs10=',Xchs10 );

writeln(' Xchs15=',Xchs15 );

writeln(' Xchs20=',Xchs20 );

writeln(' Xchs30=',Xchs30 );

writeln(' Xchs35=',Xchs35 );

writeln(' Xchs40=',Xchs40 );

writeln(' Xchs45=',Xchs45 );

writeln(' Xchs49=',Xchs49 );

userwait;

Proces~CalibrationPrinholes (passOption,passReturnC,pas~ColX,passrowY);

if passReturnC = 0 then begin writeln(' *** All pinholes have been located with acceptable');

writeln(' readings. ');

writeln;

writeln(' Press enter to c, ce the last phase of the calibration writeln(' sequence which builds the STEREO.GEO file. ');

WaitKey;

end else begin ClrScr;

if passReturnC = 1 then begin ~ a pinhole was not found } { it could be dust or a gain level set too low } wrlteln(' *** Pinhole ',passColX,',',passRowY,' was not found.');writeln(' This is probably due to the gain being set too low,');

writeln(' and/or dust on the calibration plate.'};

writeln;

writeln(' The beam is positioned near the pinhole it could not read.');

writeln(' Adjust the gain until the pinhole reads correctly. ');

userwait;

AdjustGain('X'); ( not the center hole } goto RePCP;

end;

if passReturnC = 2 then begin ( saturated reading received } writeln(' *** Pinhole ',passColX,',',passRowY,' was saturated.');writeln(' Please lower the gain control. ');

userwait;

goto RePCP;

end;

end;

until passReturnC = 0; ***) I***} passOption := 0; ~ normal calibration : after all holes are found } ProcessCalibrationPinholes(passOption,passReturnC,passColX,passRowY);

{ use row number and column number } end;

{ } procedure CALIMENU;

a~3 co~st rlC = ^C; 1 3 3 4 0 5 2 type str80 = string[80];

var key: Char;

regs: Registers;

i, indent : integer;

InputStr : str80;

FileOR : boolean;

{ Cursor on / off routines } procedure CursorOff;

begin with regs do begin AH := 3;

BH := 0;

Intr($10,regs);

AH := 1;

CH := CH or $20;

Intr($10,regs);

end;

end;

procedure CursorOn;

begin with regs do begin AH := 3;

BH := 0;

Intr($10,regs);

AH := 1;

CH := CH and $DF;

Intr($10,regs);

end;

end;

procedure beep;

begin sound(1000);

delay(500);

NoSound;

end;

function many(Ch: Char;num: Integer): string;

var temp: string;

i: Integer;

begin temp := '';

for i:=1 to num do temp := temp+Ch;

many := temp;

end;

procedure centerln(text: String);

begin gotoxy((79-length(text)) div 2,whereY);

writeln(text);

end;

procedure Abort;

var key: Char;

be~in ile keypressed do key := ReadKey; 1 3 3 4 0 5 2 halt(127);

end;

procedure HandleReyboardRequest;

begin while keypressed do begin key := ReadKey;

if key = CtrlC then Abort;

if (key = #0) and keypressed then key := ReadKey;

end;

end;

procedure DrawBox(xl,yl,x2,y2: integer;BoxType: integer);

type BoxMode = array[1..4] of Char;

{ Box Types:

1: Single Line all around 2: Double Line all around 3: Single Line horizontal, Double Line vertical 4: Double Line horizontal, Single Line vertical ) const { Type 1 2 3 4 UpperLeftCorner : BoxMode = (#218,#201,#214,#213);

UpperRightCorner: BoxMode = (#191,#187,#183,#184);

LowerLeftCorner : BoxMode = (#192,#200,#211,#212);

LowerRightCorner: BoxMode = (#217,#188,#189,#190);

HorzBar : BoxMode = (#196,#205,#196,#205);

VertBar : BoxMode = (#179,#186,#186,#179);

var i: integer;

FullHorzBar: String;

begin gotoxy(xl,yl);

if x1&lt;x2 then FullHorzBar := Many(HorzBar[BoxType],x2-x1-1) else FullHorzBar := '';

if x1<>x2 then write(UpperLeftCorner[BoxType]);

write(FullHorzBar);

if x1<>x2 then write(UpperRightCorner[BoxType]);

for i:=yl+1 to y2-1 do begin gotoxy(xl,i);

write(VertBar[BoxType]);

gotoxy(x2,i);

write(VertBar[BoxType]);

end;

gotoxy(xl,y2);

if xl<>x2 then write(LowerLeftCorner[BoxType]);

write(FullHorzBar);

if xl<>x2 then write(LowerRightCorner[BoxType]), end;

procedure GetSelection;

begin repeat HandleKeyboardRequest;

key := upcase(key);

if key in ['1'..'5','X'] then exit;

until false;

end;

procedure ShowCaliMenu;

4~ va~ ~ 1 3 3 4 0 5 2 integer;

be ~ window(1,1,80,25);

textmode(C80);

writeln;

writeln;

centerln('SLA-1 Calibration Control Program');

centerln('Version '+VersionId);

centerln('Copyright (C) 1988 by 3D Systems, Inc.');

centerln('3D Systems Laser Stereolithography System');

for i:=1 to 7 do writeln;

DrawBox(Indent-1,8,80-Indent,20,1);

window(Indent,9,80-Indent-1,15);

ClrScr;

writeln(' 1.

Calibrate SLA-1 Machine');

writeln;

writeln(' 2.

Process Calibration File');

writeln;

writeln(' 3.

Install New Geometric Table');

writeln;

writeln(' 4.

Explanation of Calibration Process');

writeln;

writeln(' 5.

Initialize Digital Potentiometer');

writeln;

write (' ( X to Exit )');

window(1,1,80,25);

gotoxy(1,22);

centerln('Please make a selection.');

end;

procedure reboot;

begin port[$305]:= port[$305] or 7;

Inline($B8/$40/$00/$8E/$D8/$BB/$72/$00/$C7/$07/$34/$00/$C7/$47/$01/$12/$00/ $EA/$00/$00/$FF/$FF/ $01/$04/$43/$4F/$50/$59/$03/$61/$26/$05/$50/$41);

end;

procedure UserWait;

begin writeln;

writeln;

writeln(' ... press any key to continue ');

repeat delay(100) until ReyPressed;

junkkey:=Readkey;

end;

procedure InstallNewGeometricCorrectionFile;

const AnyFile = $3F;

var i : integer;

DirInfo : SearchRec;

begin ClrScr;

gotoxy(1,13);

writeln(' Warning: Installing the geometric correction table will delete');

writeln(' the existing \3DSYS\STEREO.GEO file by copying the');

writeln(' selected \CAL\~.GEO file over it. ');

writeln;

writeln(' Continue ? ');

repeat delay(100) until keypressed;

key := upcase(readkey);

if key <> 'Y' then exit;

ClrScr;

~riteln, 1 3 3 4 0 5 2 -iteln;

~~~riteln(' SLA-l Calibration Control Program - Install Geometric Table');

writeln;

{ exec('c:\command.com','/c dir \CAL\*.GEO'); } writeln(' Existing Geometric Correction Files~ write (' (none)'):

gotoxy(WhereX-16,WhereY);

i := O;

FindFirst('\CAL\*.GEO',anyfile,DirInfo);

while DosError = 0 do begin if (i - 4) = 0 then begin writeln;

i := O;

end;

gotoxy(8+(15*i),WhereY);

i := i + l;

write(DirInfo.Name);

FindNext(DirInfo);

end;

writeln;

gotoxy(l,WhereY);

writeln;

CursorOn;

FileOk := false;

repeat writeln(' Please enter the file to install or X to exit.');

readln(InputStr);

for i := 1 to length(InputStr) do InputStr[i]:=upcase(InputStr[i]);

if InputStr[1]='X' then begin CursorOff;

exit;

end;

if pos('.',InputStr)<>0 then InputStr := '\CAL\' + InputStr else InputStr := '\CAL\' + InputStr + '.GEO';

FindFirst(InputStr,AnyFile,DirInfo);

writeln;

if DosError = 0 then FileOR := true else writeln(' Error: Invalid file name input.');

until FileOR;

exec('c:\command.com','/c copy'+InputStrf' \3DSYS\STEREO.GEO');

delay(S00);

if DOSerror = 0 then begin writeln;

writeln(' The system will now automatically reboot.');

UserWait;

reboot;

end else begin CursorOff;

writeln(' Error: Invalid file name input.');

UserWait;

end;

end;

procedure Explain;

begin ClrScr;

writeln;

writeln(' The calibration process provides data to translate the absolute');

writeln(' laser ~o~ ---t coordinates to the actual locations in the SLA-l');

writeln(' vat. ');

writeln;

writeln(' The first step in calibrating utilizes a plate with many sensing');

writeln(' holeq which the beam moves across. The absolute coordinates for');writeln(' each sensing hole are stored in a *.CAL file. Many calibration');

writeln(' files can be generated. They are stored in the \CAL directory');

~1 1 334052 writeln~(' as *.CAL where the ~ is a time stamp. A calibration file run on');riteln(' May 1 at 1:06 PM would be named 05011306.CAL (modytime.CAL).');

`-~riteln;

writeln(' The ~.CAL file does not compensate for hardware mirror drift.');

writeln(' Between each sensor hole reading on the calibration plate the ');

writeln(' laser recalculates the (static) sensor 1 and Z positions. By');

writeln(' comparing the drift of sensors 1 and 2 the drift of the entire');

writeln(' mirror apparatus can be computed. These results are in contained');writeln(' in the ~.DRF file.');

writeln;

writeln(' The geometric (correction) table is the culmination of the ');

writeln(' translation data. This table only becomes active when the ~.GEO');

writeln(' file from the \CAL directory is moved to the \3DSYS directory');

writeln(' as STEREO.GEO and the system is rebooted. (press any to exit)');

repeat delay(100) until keypressed;

key := Readkey;

end;

procedure InitDigitalPot;

var ADC : real;

i : integer;

passwordl : string;

WiperSettings : array[1..100] of real;

procedure ShowWiperMenul;

begin ClrScr;

gotoxy(1,8);

writeln(' Digitally controlled potentiometer settings should only be ');

writeln(' executed by 3D systems field engineers. ');

writeln;

writeln;

write ('Please key password:');

end;

procedure ShowWiperMenu2(ADC : real; Status : char);

var StatusStr : string;

begin if Status = 'U' then StatusStr := 'UP else if Status = 'D' then StatusStr := 'DOWN else if Status = 'S' then StatusStr := 'STABLE';

ClrScr;

gotoxy(1,8);

writeln(' Voltage Wiper Element Status =',StatusSTR);

writeln;

writeln(' Current sensor ADC x100 =',ADC:4:6);

writeln;

writeln;

writeln(' ( Initalization is complete only when status = STABLE ) ');

delay(500);

end;

procedure MoveWiper(direction:char;times:integer);

var iwa~tel,iwaste2,i : integer;

begin ~ port[$305] bit 4 = CS, 5 = U/D, 6 = INC { MSB 76543210 LSB ===> CS is bit 4 = 00010000 = $10 } { NMI is bit 0-2 = $07 port[$305] := port[$305] or $F0; { set all hi initially~ for iwastel := 1 to 10 do inc(iwaste2);

port[$305] := port[$305] and $EF; { set CS lo to set-up for iwastel := 1 to 1 do inc(iwaste2); { 0 delay for CS to INC for iwastel := 1 to 1 do inc(iwaste2);{ 100 n ~ ~f direction = ~U~ l 3 3 4 0 5 2 then port[$305] := port[$305] or $20 ~ set for upward increment ~ else port[$305] := port[$305] and $DF; { set for downward Lncrement~ for iwastel := 1 to 1 do inc(iwaste2); ~ 4 us AC timing delay for i := 1 to times do begin port[$305] := port[$305] and $BF; ~ use INC to move wiper } for iwastel := 1 to 10 do inc(iwaste2);~ 50 u delay before voltage x } ~ this covers the 2u INC to CS} port[$305] := port[$305] or $gO; ~ set incl hi } for iwastel := 1 to 1 do inc(iwaste2};

end;

port[$305] := port[$305] or $10; ~ set cs hi to save setup} delay~50); ~ 50 ms delay for CS deselect } end;

procedure DecWiperlOOtimes;

var i : integer;

begin ShowWiperMenu2(0,'D'};

MoveWiper('D',100};

end;

procedure IncWiperlOOtimes;

var i,j: integer;

sum: real;

begin sum := O;

for j:= 1 to 10 do begin ADC := ReadADC~SensorChannel[Sensor]+10};

sum := sum + ~200 - ADC};

ADC := sum / 10;

end;

WiperSettings[1] := ADC;

i := 2;

repeat inc(i};

MoveWiper('U',1);

sum := O;

for j:= 1 to 10 do begin ADC := ReadADC(SensorChannel[Sensor]+10~;

sum := sum + (200 - ADC);

ADC := sum / 10;

end;

WiperSettings[i] := ADC;

ShowWiperMenu2(ADC,'U');

until ~ADC &lt;= O) or (i = 100);

MoveWiper('D',1);

if i = 100 then i := 99;

WiperSettings[i+1] := 999;

end;

procedure OutputValues;

var junkkey : char;

i : integer;

begin ShowWiperMenu2(ADC,'S');

repeat delay(100) until keypressed;

junkkey := ReadKey;

ClrScr;

for i := 1 to 25 do begin write (' W[',i ,']= ',WiperSettings[i ]:3:3);

write (' W[',i+25,']= ',WiperSettings[i+25]:3:3);

write (' W[',i+50,']= ',WiperSettings[i+50]:3:3);

~, ~q write (~ W[~,i+75,~]= ~,WiperSettings[i+75]:3:3); 1 334052 if i <> 25 then writeln;

~end;

repeat delay(100) until keypressed;

junkkey := ReadKey;

end;

begin ShowWiperMenul;

readln(passwordl);

if passwordl = '3dsys' then begin Sensor :z 2;

for i := 1 to 100 do WiperSettings[i] := O;

DecWiperlOOtimes; { start our search for the lowest ADC at bottom } IncWiperlOOtimes; { and save the ADC reading for each one } OutputValues; { this is temporary for print screen } end;

end; { procedure InitDigitalPot begin CursorOff;

indent := 18;

repeat ShowCaliMenu;

GetSelection;

ClrEol;

case key of '1': Calibr8;

'2': begin exec('c:\command.com','/c CALOUT'};

delay(3000};

end;

'3': InstallNewGeometricCorrectionFile;

'4': begin Explain;

key := 'z';

end;

'5': InitDigitalPot;

'X': begin key := 'z';

Exit;

end;

end;

key := ~z.;

until false;

end;

q~