US20010029674A1 - Abbe error correction system and method - Google Patents

Abstract

The present invention preferably employs non-contact, small-displacement, capacitive sensors to determine Abbe errors due to the pitch, yaw, or roll of a near linear mechanical stage that are not indicated by an on-axis position indicator, such as a linear scale encoder or laser interferometer. The system is calibrated against a precise reference standard so the corrections depend only on sensing small changes in the sensor readings and not on absolute accuracy of the sensor readings. Although the present invention is preferred for use in split-axis positioning systems with inertially separated stages, the invention can be employed in typical split-axis or stacked stage systems to reduce their manufacturing costs.

Description

RELATED APPLICATION
• This application derives priority from U.S. Provisional Patent Application No. 60/175,993, filed Jan. 11, 2000.[0001]
TECHNICAL FIELD
• This invention relates to systems or methods for positioning one or multiple “tools,” such as laser beams or other radiation beams, relative to target locations on one or multiple workpieces and, in particular, to a system that accurately compensates for Abbe errors associated with the movement of one or more stages of such a beam positioning system.[0002]
BACKGROUND OF THE INVENTION
• A variety of technologies employ tools to micro-machine, or deposit patterns or materials on target locations on a workpiece. For example, a micro-dimensioned punch may be used to punch holes in a thin metal plate; a laser may be used to precisely machine or selectively erode metallic, crystalline, or amorphous specimens; and ion beams may be used to selectively implant charged particles into an integrated circuit. All of the above-mentioned processes share a common requirement for accurately and rapidly positioning a pertinent tool to target locations on the workpiece.[0003]
• The following background is presented herein only by way of example to laser beam positioning systems, but skilled persons will appreciate that the description is applicable to tool positioning systems in general. Conventional tool positioning systems, and particularly beam-positioning systems, typically provide movement within a three-dimensional coordinate system and can be characterized in several ways.[0004]
• Traditional positioning systems are characterized by X-Y translation tables in which the workpiece is secured to an upper stage that is supported by a lower stage. Such systems typically move the workpiece relative to a fixed beam position and are commonly referred to as stacked stage positioning systems because the lower stage supports the inertial mass of the upper stage and the workpiece. These positioning systems have relatively good positioning accuracy because interferometers are typically used along each axis to determine the absolute position of each stage.[0005]
• In U.S. Pat. No. 4,532,402 of Overbeck, a high-speed short-movement positioner (“fast positioner”), such as a galvanometer, is supported by the upper stage of an X-Y translation table (“slow positioner”) and the upper stage and the workpiece are supported by the lower stage. The combined movement of the two positioners entails first moving the slow positioner to a known location near a target location on the workpiece, stopping the slow positioner, moving the fast positioner to the exact target location, stopping the fast positioner, causing the tool to operate on the target location, and then repeating the process for the next target location.[0006]
• However, the combined system of Overbeck is also a stacked stage positioning system and suffers from many of the same serious drawbacks as the aforementioned fixed beam system. The starting, stopping, and change of direction delays associated with the inertial mass of the stages and fast positioner unduly increase the time required for the tool to process the workpiece. Overbeck's system also imposes a serious drawback upon a computer-based machine tool control file or “database” that typically commands the tool to move to a series of predetermined target locations across the workpiece. The database positioning the tool across the workpiece must be “panelized” into abutting segments that each fit within the limited movement range of the fast positioner when the size of large circuit patterns exceeds this movement range.[0007]
• U.S. Pat. Nos. 5,751,585 and 5,847,960 of Cutler et al. describe split-axis positioning systems, in which the upper stage is not supported by, and moves independently from, the lower stage and in which the workpiece is carried on one axis or stage while the tool is carried on the other axis or stage. These positioning systems have one or more upper stages, which each support a fast positioner, and can process one or multiple workpieces simultaneously at high throughput rates because the independently supported stages each carry less inertial mass and can accelerate, decelerate, or change direction more quickly than can those of a stacked stage system. Thus, because the mass of one stage is not carried on the other stage, the resonance frequencies for a given load are increased. Furthermore, the slow and fast positioners are adapted to move, without necessarily stopping, in response to a stream of positioning command data while coordinating their individually moving positions to produce temporarily stationary tool positions over target locations defined by the database. These split-axis, multirate positioning systems reduce the fast positioner movement range limitations of prior systems while providing significantly increased tool processing throughput and can work from panelized or unpanelized databases.[0008]
• Such split-axis positioning systems are becoming even more advantageous as the overall size and weight of the workpieces increase, utilizing longer and hence more massive stages. At the same time, feature sizes are continuing to decrease, causing the need for dimensional precision to increase, and split-axis systems are more likely to exhibit rotational errors that introduce Abbe errors, which are errors indicative of the physical separation between the effective position of a stage and the indicated position of the stage. Abbe errors are typically caused by imperfections or thermal variations in the bearings upon which the stages slide and/or alignment or acceleration imperfections of the drive mechanisms that provide movement to the stages.[0009]
• FIG. 1 shows three mutually perpendicular translational motion axes, such as[0010]X axis10,Y axis12, andZ axis14 that define a three-dimensional coordinate system16, and three mutually perpendicular rotational motion axes (hereafter referred to as aroll axis18, apitch axis20, and a yaw axis22). Skilled workers typically refer to roll as an angular rotation aboutX-axis10, pitch as an angular rotation about Y-axis12, and yaw as an angular rotation about Z-axis14.
• Although laser interferometer systems can be used to indicate and compensate for certain Abbe errors, such systems are costly and heavy because they typically require reference mirrors that are nearly as long as the combined stage length plus the length of travel, e.g. as much as two times the travel distance. Such mirrors are difficult, if not impossible, to procure for the long travel dimensions of large stages, such as with a lengthwise dimension of 76 to 92 cm (30-36 inches), needed to accommodate larger workpieces. Furthermore, split-axis systems would require at least two interferometers for each stage and/or a very complex system of optics to rat -indicate angle and position, and the additive weight of the interferometers would increase the inertial load on the stages at the expense of frequency response time to changes in momentum.[0011]
• U.S. Pat. No. 5,699,621 of Trumper et al. discloses the use of small range displacement transducers to indicate pitch, yaw, and roll angle errors. Trumper et al. correct angular errors by controlling the bearing gap with electromagnets that require the use of a highly compliant magnetic or air bearing system. The correction speed of the Trumper et al. system is limited to the bandwidth of the linear stage system and therefore has similar mass versus bandwidth limitations as stacked stage positioning systems.[0012]
• A less expensive and/or less massive and very accurate Abbe error correction system or method is therefore desirable.[0013]
SUMMARY OF THE INVENTION
• An object of the present invention is to provide a method or apparatus that employs non-contact small displacement sensors, such as capacitive sensors, to determine Abbe errors due to mechanical stage pitch, yaw and roll that are not indicated by an on-axis position indicator, such as a linear scale encoder or laser interferometer, and a means to compensate for such Abbe errors.[0014]
• Another object of the invention is to employ such sensors to determine and correct Abbe errors due to linear bearing variability or distortions associated with acceleration or temperature gradients.[0015]
• The present invention provides a cost effective means to determine and compensate for linear stage positioning system Abbe errors that are errors at the effective position of the system that are not indicated by a position indicator such as a metal or glass scale encoder or laser interferometer due to pitch, yaw, or roll of the linear stage and the resulting physical distance between the effective position and the indicated position of a stage. To minimize cost, the system is calibrated against precision X and Y position reference standards so the corrections depend only on sensing small changes in the sensor readings and not on absolute accuracy of the sensor readings. Although the present invention is preferred for use in split-axis positioning systems, it can be employed in stacked stage systems to reduce their manufacturing costs. Although a linear scale encoder can be employed to indicate the nominal on-axis stage position to reduce costs further, a laser interferometer can be used when a greater level of accuracy and/or resolution is desired.[0016]
• Additional objects and advantages of this invention will be apparent from the following detailed description of preferred embodiments thereof which proceeds with reference to the accompanying drawings.[0017]
BRIEF DESCRIPTION OF THE DRAWINGS
• FIG. 1 shows six axes, including three mutually perpendicular translational motion axes, X, Y, and Z, and three mutually perpendicular rotational motion axes, roll, pitch, and yaw.[0018]
• FIGS. 2A and 2B provide a pictorial block diagram of a multi-stage laser beam positioning system of this invention.[0019]
• FIG. 3 is a fragmentary pictorial side view showing a prior art galvanometer-driven mirror positioner of a type suitable for use with this invention.[0020]
• FIG. 4 is a plan view showing preferred positions of Y-stage Abbe error sensors mounted on a Y-axis stage (workpiece stage) relative to a reference surface.[0021]
• FIG. 5 is an end view showing preferred positions of the sensors mounted on the Y-axis stage of FIG. 4.[0022]
• FIG. 6 is a side elevation view showing preferred positions of X-stage Abbe error sensors mounted on an X-axis stage (tool stage) relative to a reference surface.[0023]
• FIG. 7 is an end view showing preferred positions of the sensors mounted on the X-axis stage of FIG. 6.[0024]
• FIG. 8 is a plan view showing preferred positions of the sensors mounted on the X-axis stage of FIG. 6.[0025]
• FIG. 9 is an oblique pictorial view showing a multi-head laser machining system employing the present invention.[0026]
• FIG. 10 is a simplified electrical block diagram of a digital signal processing system including multiple fast stage signal processors employed in the multi-head laser machining system of FIG. 9.[0027]
• FIG. 11 is a simplified electrical block diagram of one of multiple fast stage signal processors employed in the digital signal processing system of FIG. 10.[0028]
DETAILED DESCRIPTION OF PREFERRED EMBODIMENT
• FIGS. 2A and 2B (generically FIG. 2) show a multi-stage[0029]tool positioner system50 having positioning command execution capabilities in accordance with this invention.Positioner system50 is described herein only by way of example with reference to a single-head, laser-based hole cutting system that employs a digital signal processor (“DSP”)52 to control a fast galvanometer positioner stage54 (scanner or “fast stage54”), a slow X-axis translation stage56 (“slow stage56”), and a slow Y-axis translation stage58 (“slow stage58”) to direct alaser beam60 to target locations on asingle workpiece62, such as an etched circuit board.
• With reference to FIG. 1, in a preferred split-axis embodiment, the[0030]X-axis translation stage56 is supported by bearings onrails46 and generally moves along an X-Z plane, and the Y-axis translation stage58 is supported by bearings onrails48 and generally moves along an X-Y plane. Skilled persons will appreciate that bothstages56 and58 could alternatively be adapted to move in parallel planes and be inertially separated or dependent. In a preferred embodiment,positioner system50 employs high stiffness re-circulating or cross-roller bearing systems to support and direct the movement ofstages56 and58.
• A[0031]system control computer63 processes a tool path database stored in adatabase storage subsystem64. The database contains the desired processing parameters for cutting holes and/or profiles withlaser beam60 inworkpiece62. The database is conventionally compiled using a tool path generating program, such as I-DEAS Generative Machining provided by Structural Dynamics Research Corporation located in Milford, Ohio.System control computer63 conveys parsed portions of the stored database to alaser controller68 and position control portions of the database as a data stream to adelta process70.Delta process70 resolves the data stream into x and y components for delta position (“dp”), delta velocity (“dv”), and delta time (“dt”) for each intended change in the path oflaser beam60 acrossworkpiece62. Consequently, each movement oflaser beam60 is defined in dp, dv, and dt components that are further processed by aposition profiler72 into move profiles including acceleration and/or constant velocity segment position signals.
• [0032]Delta process70 preferably generates the dp, dv, and dt components in accordance with a preferred BASIC language signal processing procedure described in U.S. Pat. Nos. 5,751,585 and 5,847,960 of Cutler et al., which are assigned to the assignee of this application.
• Referring again to FIG. 2, the dp, dv, and dt components generated by[0033]delta process70 are further processed byposition profiler72 into the move profile positioning signals required to movefast stage54 andslow stages56 and58 as commanded by the database. Ideally, positioner acceleration is proportional to motive force, and motive force is proportional to electrical current supplied to a positioner driver such as a linear or rotary servo motor or a galvanometer coil. Therefore, the positioning signal produced byposition profiler72 is a series of “full-spectrum” half-sine profiled acceleration-inducing and constant velocity-inducing positioning steps that cause system movements. The full-spectrum bandwidth need only be about 250 Hertz, a bandwidth sufficient to drive a typical galvanometer-driven mirror positioner at its maximum frequency.
• Instantaneous values of the full-spectrum positioning signal are generated by[0034]DSP52 at a rate of about 10,000 points per second by employing the dp, dv, and dt components generated bydelta process70 as variables for a sine value generation program running inDSP52. Alternatively, the dp, dv, and dt components may be employed to address and fetch associated sinusoidal waveform values stored in a sine value lookup table that is incorporated withinDSP52.
• The resulting full-spectrum positioning signal has acceleration and position components that are received by a[0035]profiling filter78 having a constant signal propagation delay and adelay element79 that compensates inDSP52 for the constant signal propagation delay ofprofiling filter78. For example,delay element79 delays the laser triggering pulses generated byposition profiler72 to coincide with the delayed movements offast stage54 andslow stages56 and58.Profiling filter78 anddelay element79 also cooperate, as described below, to moveslow stages56 and58 smoothly over the average position profile while limiting their acceleration to ±1 g and cooperate to limitfast stage54 positioning movements to ±10 millimeters.
• The position component is received by profiling[0036]filter78 to produce filtered position command data for drivingslow stages56 and58.Profiling filter78 is preferably a fourth-order low-pass filter.
• Because profiling[0037]filter78 produces filtered position command data having a constant time delay with respect to the half-sine positioning signal position component, the constant time delay is compensated for bydelay element79. Delayelement79 is preferably implemented inDSP52 as a programmed delay in conveying the half-sine positioning signal acceleration and position components fromposition profiler72 tofast stage54 signal processing elements, the first of which areadders80 and82. Thereby, half-sine positioning signals directed tofast stage54 are time synchronized with the filtered position commands directed to slowstages56 and58.
• The acceleration component from[0038]position profiler72 is also filtered by profilingfilter78 to provide a filtered acceleration command to adder80 and afeed forward process94.Adder80 functions as a high-pass filter by subtracting the filtered acceleration command from the acceleration component of the full-spectrum positioning signal to form a galvo acceleration feed forward signal, which is conveyed to afeed forward process86. Likewise, the filtered position command from profilingfilter78 and the delayed position component of the half-sine positioning signal are conveyed respectively toadders90 and82 for processing and distribution, respectively, to slowstages56 and58 andfast stage54. Agalvo filter97 and aservo filter98 are conventional loop compensation filters that function to keepfast stage54 andslow stages56 and58 stable.
• Profiling[0039]filter78 is implemented by cascading two or more second-order filters having critical damping ratios. As the number of cascaded filters increases beyond two, their cutoff frequencies increase by about the square root of the number of filters (e.g., two filters have cutoffs that are 1.414 times the cutoff for a single filter). Preferably two filters are cascaded to provide good smoothing while keeping the overall filter implementation simple.
• For[0040]profiling filter78, the preferred38 radian per second cutoff frequency (about 6 Hertz (Hz)) is a very low frequency compared to the 10 kHz rate at whichDSP52 updates positioning data forslow stages56 and58. If profilingfilter78 runs at the 10 kHz slow stage update frequency, the discrete filter coefficients become sensitive to roundoff errors because the poles of the discrete filter move close to the unit circle.Profiling filter78 also receives the acceleration command fromposition profiler72 and generates the filtered acceleration command that is conveyed to servo feedforward process94 and to adder80.
• The desired move profile commands are preferably calculated at the 10 kHz updating rate, and the slow stage acceleration and actual (not commanded) position is subtracted therefrom at[0041]adders80 and82 to produce, respectively, the fast stage acceleration and position command signals.
• The fast stage acceleration command signal is processed through[0042]adder80 and feedforward process86, while the fast stage position command signal is processed throughadder82 andgalvo filter97. The processed fast stage signals are combined in anadder84 and conveyed to agalvanometer driver88.
• Likewise, the slow stage filtered acceleration command is processed through a[0043]feed forward process94, while the slow stage filtered position command is processed throughadder90 andservo filter98. The processed slow stage signals are combined in anadder92 and conveyed to a linearservo motor driver96.
• [0044]Galvanometer driver88 provides deflection control current to a pair of mirror deflecting galvanometers infast stage54, andservo motor driver96 provides control current to linear servo motors that control the positioning ofslow stages56 and58.
• FIG. 3 shows a prior art galvanometer-driven[0045]mirror positioner100 of a type suitable for use asfast stage54. Galvanometer driver88 (FIG. 2) provides rotational control current onconductors102 to respective X-axis and Y-axis high-speedresponse D.C. motors104 and106 that rotateshafts107 inbearings108 to selectively pivot a pair ofmirrors110 and112 that deflectlaser beam60 through anoptional lens114 to a predetermined target location onworkpiece62.
• Alternatively, a nonbearing motion positioner, such as a piezoelectric element, a voice coil actuator, or other limited angle high-speed positioner device could be used in place of galvanometer-driven[0046]mirror positioner100 inpositioner system50.
• Likewise with reference to FIG. 2, alternative accurate rotary or linear positioner mechanisms may be substituted for the linear servo motors driving[0047]slow stages56 and58. However, inpositioner system50, linear motors that preferentially respond to the slow stage position command are preferred.
• Two signals are combined with the slow and fast stage position commands to reduce positional errors between the commanded position and the actual position of[0048]laser beam60 onworkpiece62. The delayed fast stage position command atadder82 and the filtered slow stage position command atadder90 represent the ideal signal values required to cause proper positioning ofstages54,56, and58. However, practical factors such as gravity, friction, mass, and inaccuracies in the full-spectrum positioning signal generated byposition profiler72 are not contemplated in the unmodified position commands.
• The practical factors are accounted for by sensing the actual positions of[0049]stages54,56, and58 withposition sensors120 and122 to provide predictive position feedback data toadders82 and90 inDSP52. Note thatadder82 in the fast stage positioning path receives position feedback data from bothposition sensors120 and122.Position sensors120 and122 may be well-known types employing rotating capacitor plates, linear and rotary encoder scales, or interferometer motion detectors together with appropriate analog-to-digital and/or digital-to-analog conversion techniques.
• As[0050]laser beam60 undergoes movement acrossworkpiece62, the sensed beam position is continuously compared to the commanded beam position, with the positional difference representing a degree to which the practical factors have caused positioning errors. In particular, sensed position data offast stage54 andslow stages56 and58 are generated byposition sensors120 and122 and subtracted from the commanded position atadder82 to generate positional difference data that are combined inadder84 with acceleration data from feed forwardprocess86. Likewise, sensed position data ofslow stages56 and58 are generated byposition sensor122 and subtracted from the commanded position atadder90 to generate positional difference data that are combined inadder92 with acceleration data from feed forwardprocess94.
• Coordinated positioning is particularly beneficial for applications such as laser beam hole cutting that requires rapid movement between target locations along a tool path combined with pauses at each target location to fire the laser to cut a hole but, of course, is not limited to that application. Other features and preferred processing parameters of a conventional laser drilling system are disclosed in U.S. Pat. No. 5,841,099 of Owen et al.[0051]
• FIGS. 4 and 5 are respective plan and end views showing preferred positions of Y-[0052]stage Abbe sensors124 mounted on Y-axis translation stage58 relative to yawreference surface126 in accordance with an aspect of this invention, and FIGS.6-8 are respective side elevation, end, and plan views showing preferred positions ofX-stage Abbe sensors128,130,131, and132 mounted onX-axis translation stage56 relative to yaw and roll reference surfaces134 and pitchreference surface136 in accordance with an aspect of this invention.
• With reference to FIGS.[0053]2B and4-8,Abbe sensors124,128,130,131, and132 are preferably non-contact, small and lightweight displacement sensors. The most preferred sensors measure capacitance as a function of distance from a given reference surface. In a preferred embodiment, the Abbe sensors have a gap range (distance between sensor and reference surface) of 50 μm plus or minus 25 μm and a resolution of less than 50 nm and preferably less than or equal to 10 nm. Skilled persons will appreciated that numerous other ranges are possible including a wider or narrower gap range and better resolution when the technology becomes cost effective. Non contact sensors are preferred because they eliminate wear that might lead to inaccuracies. Preferred Abbe sensors include Model PX405H series probes available from Lion Precision of St. Paul, Minn. Other suitable capacitance probes or sensors are available from ADE Technologies of Westwood, Mass and Micro-Epsilon of Ortenburg, Germany.
• Reference surfaces[0054]126,134, and136 may be formed on appropriate sides of bearingrails46 and48 as shown in FIG. 2B or may be otherwise positioned near but separated from translations stages56 and58 as shown in FIGS.4-8. (In FIGS.4-8, four Y-stage bearings138 and threeX-stage bearings140 are depicted instead ofrails46 and48). The reference surfaces are preferably the same length as the base for the stages or at least as long as the movement ranges along bearingrails46 and48. The reference surfaces are preferably stable but do not need to be perfectly straight because the sensors are calibrated against the entire length of the surfaces so the corrections depend only on sensing small changes in the sensor readings and not on absolute accuracy of the sensor readings or stage positions.
• Although[0055]stages56 and58 could be adapted to move in parallel planes and be inertially separated or dependent, the following description is, for convenience, presented herein only by way of example to addressing X and Y axis position errors in split-axis positioning system70 where substantially flat (100 to 10,000 times larger in the X and Y dimensions than in Z dimension) workpiece62 is carried onY stage58 and the tool (laser76) is directed byX-stage56.
• With reference again to FIGS. 2A, 2B,[0056]4 and5, the nominal on-axis position ofY stage58 is indicated bysensor122a, which is preferably a glass or metal scale encoder or a laser interferometer depending on desired positioning accuracy specifications. In a split-axis configuration, Y-stage yaw typically produces the most significant X and Y Abbe errors. The yaw error is indicated by preferably a pair of Y-stage Abbe sensors124aand124b(generically sensors124) that are preferably mounted as far apart as possible alongY axis12 and as near to the top of the side ofY stage58, or a chuck that it may support, as practical.Reference surface126 is preferably integrated intorail46 or the base of the Y stage assembly in a manner that results in as stable an indication of the stage yaw as possible as a function of other effects including bearing repeatability, temperature, and stage acceleration.
• The capacitances indicating the X components of the distance from the reference surface of the Abbe error detected by[0057]sensors124 due to yaw are preferably converted by a Y-stageyaw probe driver145 into a DC voltage suitable for processing into Abbe error correction signals. These signals may be directed to separate X-Abbe and Y-Abbe error adders142 and144 before being routed to adder82 and incorporated into scanner position commands.
• With reference again to FIGS. 2A, 2B, and[0058]6-8, pitch, yaw, and roll ofX stage56 can also cause significant X and Y position errors. The preferred split-axis configuration, as shown in the figures, hasX stage56 oriented on edge, such that the planes defined by thestages56 and58 are transverse and such that stages56 and58 are inertially separated. In a most preferred embodiment,X stage56 is oriented vertically whileY stage58 is oriented horizontally. Thus, pitch, yaw, and roll in this context are defined with respect to the actual plane of movement ofX stage56 and not with respect to a more typical horizontal orientation.
• The nominal on-axis position of[0059]X stage56 is indicated bysensor122b, which is preferably a glass or metal scale encoder or a laser interferometer depending on desired positioning accuracy specifications. TheX-stage Abbe sensors128,130,131, and132 may all be the same types as or different types from Y-stage sensors124.Sensors128aand128b(generically sensors128) are preferably mounted as far apart as possible alongX axis10. Similarly,sensors130 and131 are preferably mounted as far apart as possible alongX axis10.Sensor132 is preferably mounted to be planar with and as far apart as possible alongZ axis14 fromsensor131.
• Because[0060]X stage56 is preferably kinematically mounted on threebearings140 as shown in FIGS.6-8, changes in distance fromreference surfaces134 and136 detected byX-stage Abbe sensors128,130,131, and132 will result predominately in movement of a plane associated withX stage56 and not from distortion ofX stage56.X-stage Abbe sensors130 and131 detect distances from X-stageyaw reference surface134aand indicate changes in the yaw angle of the plane ofX stage56.X-stage Abbe sensors131 and132 detect distances from X-stage role reference surfaces134aand134b, respectively, and indicate changes in the roll angle of the plane ofX-stage56.X-stage Abbe sensors128 detect distances from thepitch reference surface136 and indicate changes in the pitch angle ofX stage56.
• The capacitances indicating the X and Y components of the distance from the[0061]reference surface134aof the Abbe error detected bysensors130 and131 due to yaw are preferably converted by an X-stageyaw probe driver146 into a DC voltage suitable for processing into Abbe error correction signals. Similarly, the capacitances indicating the X and Y components of the distances from the reference surfaces134aand134bof the Abbe error detected bysensors131 and132 due to roll are preferably converted by an X-stageroll probe driver147 into a DC voltage suitable for processing into error correction signals. Similarly, the capacitances indicating the X and Y components of the distances from thereference surface136 of the Abbe error detected bysensors128 due to pitch are preferably converted by an X-stagepitch probe driver148 into a DC voltage suitable for processing into error correction signals. Skilled persons will note thatsensor131 feeds bothyaw probe driver145 and rollprobe driver146. Suitable probe drivers are well known to skilled persons; however, the Compact Probe Driver manufactured by Lion Precision is preferred. These yaw, roll, and pitch Abbe error correction signals may be directed to separate X-Abbe and Y-Abbe adders142 and144 before being routed to adder82 and incorporated into scanner position commands.
• The X and Y position components that correspond to these Abbe errors are calculated in real time as[0062]positioner system50 moves andprocess workpiece62 and are added to or superimposed on the scanner position commands to compensate for the Abbe position errors. These angular changes are combined with the geometry of the optics (including location of the beam path (or beam paths) relative to the stage and distance of the work from the stage) to indicate associated changes (errors) in effective beam position on the work. Fast response is achieved by adding the Abbe error corrections to the scanner position atadder82 because the bandwidth of thefast stage54 is significantly higher than the bandwidth of thelinear stages56 and58.
• However, the Abbe error corrections resulting from the system of sensors could be added to the linear stage position servo loop directly at[0063]adder90. This implementation would be appropriate whenfast stage54 is replaced by a fixed beam positioner. A fixed beam positioner would typically provide more precise beam positioning than is provided byfast stage54 and would be employed in applications where greater accuracy might be desirable such as in severing micron or submicron sized links. Skilled persons will appreciate thatX-axis stage56 could be adapted so thatfast stage54 may be interchangeable with a fixed beam positioner, or thatX-axis stage56 may support bothfast stage54 and a fixed beam positioner simultaneously. In the latter case, Abbe error corrections would be fed to adder82 wheneverfast stage54 is employed and fed to adder90 whenever a fixed beam positioner is employed.
• FIG. 9 shows a[0064]multi-head positioner150 embodiment of this invention in whichmultiple workpieces152A,152B,152C, . . .152N are simultaneously processed. (Hereafter multiple elements are referred to collectively without the letter suffix, e.g., “workpieces152”).Multi-head positioner150 employs one each ofslow stages56 and58 configured such that workpieces152 are fixtured and carried on Y-axisslow stage58 and multiplefast stages154A,154B,154C, . . .154N are carried on X-axisslow stage56. Of course, the roles ofslow stages56 and58 may be reversed, or two or morefast stages154 may be carried by one or more X-axisslow stages56 while Y-axis58 carries asingle workpiece62.
• As the number of[0065]fast stages154 carried onslow stage56 increases, their accumulated mass becomes increasingly difficult to accelerate. Therefore, the number N offast stages154 carried onslow stage56 is preferably limited to four, although N may vary with positioner types and applications.
• Each of workpieces[0066]152 has associated with it one or more processing tools, preferably alaser156A,156B,156C, . . .156N that directs processing energy toward associatedfast stages154A,154B,154C, . . .154N by way of associatedmirrors158A,158B,158C, . . .158N.Fast stages154 deflect the processing energy to target locations in substantially square, such as 20 by 20millimeter processing fields162A,162B,162C, . . .162N located on associated workpieces152.
• [0067]Video cameras160A,160B,160C, . . .160N are positioned onslow stage56 for viewing associated processing fields162, sensing the alignments, offsets, rotations, and dimensional variations of workpieces152, and aiming and focusing lasers156.
• In the preferred embodiment, the same processing pattern is duplicated on workpieces[0068]152 by each of lasers156 andfast stages154. However, in some processing applications, processing pattern variations may be required to match the pattern to variations among workpiece geometries, scale factors, offsets, rotations, distortions. Alternatively, it may be desirable to have one or more lasers156 simultaneously processing different, but preferably nominally identical or repetitive (slave to a single Y stage), patterns on the same workpiece152. It may also be necessary to correct for fast stage nonlinearities and mounting inconsistences introduced by mounting position variations among workpieces152 mounted onslow stage58. Unlike prior multi-spindle drilling machines,multi-head positioner150 can compensate for the above-described variations by employing programmable correction factors, described with reference to FIGS. 10 and 11, when driving each offast stages154. Similarly, the Abbe errors, indicating the degree to which a commanded tool position does not match a sensed target location, can be compensated for in a manner similar with that described with respect to FIGS. 2A, 2B, and4-8.
• FIG. 10 shows how multi-rate positioner DSP[0069]52 (FIG. 2) may be adapted to coordinate the positioning of multiplefast stages154 andslow stages56 and58, resulting in amulti-head DSP170. In like manner toDSP52,multi-head DSP170 receives fromsystem control computer63 dp, dv, and dt components that are further processed byposition profiler72 into half-sine profiled positioning signals.DSP170 also includes some of the same signal processing elements asDSP52, namely profilingfilter78,delay element79, feedforward process94,servo driver96,slow stage56, andposition sensor122. Because FIG. 10 is simplified, only X-axisslow stage56 processing elements are shown. Skilled workers will understand that corresponding Y-axis elements are implied.
• Only a single[0070]system control computer63 is required to driveslow stages56 and58 and N fast stages154. Multiple faststage signal processors172A,172B,172C, . . .172N each receive fast stage correction data fromsystem control computer63. In this way, fast stage position commands and current slow stage position data are received by each of faststage signal processors172 such that each offast stages154 is directed to a common set of target locations that are further positionable by unique error correction data. If only a singleX-axis stage56 is employed to carry multiplefast stages154 and a single Y-axis stage58 is employed to carry one or multiple workpieces152, then the Abbe error detection system shown and described in connection with FIGS. 2A, 2B, and4-8 can be employed without modification, and the Abbeerror correction data190 can be fed to adder80 of FIG. 10 or faststage correction processor180 of FIG. 11 as shown.
• FIG. 11 shows a representative one of fast[0071]stage signal processors172 receiving fast and slow stage positioning data fromDSP170 and correction data fromsystem control computer63. The correction data include slow stage and workpiece related correction data that are conveyed to ageometry correction processor180 and fast stage linearity and scale factor correction data that are conveyed to a faststage correction processor182. Skilled persons will appreciate that iffast stages154 are mounted on separateX-axis stages56, which preferably have synchronized movement but may be unsynchronized, then each suchX-axis stage56 may be commanded by itsown processor170 or subprocessor. Furthermore, eachsuch stage56 would preferably be equipped with itsown position sensor122 and five X-stage Abbe sensors to compensate for any Abbe errors associated with the individual stages.
• The correction data may be equation- or lookup table-based. However, correction data employed by[0072]geometry correction processor180 and fast stage correction processor are preferably equation-based along lines described in U.S. Pat. No. 4,941,082 of Pailthorp et al. (“the '082 patent”), which is assigned to the assignee of this application and is incorporated herein by reference.
• Fast stage linearity and scale factor errors are relatively constant and depend mostly on the individual characteristics of[0073]fast stages154. Therefore, faststage correction processor182 requires relatively small and infrequent correction data changes. Generating this correction data entails, for example, directing each offast stages154 to at least 13 calibration points on an associated calibration target as described in the '082 patent. A reflected energy detector senses any differences between the directed and actual target point locations and provides difference data to system controlcomputer63 for processing. The resulting correction data are conveyed to and stored in each faststage correction processor182. Also, any differences between the directed and actual target point locations sensed by associated video cameras160 are calibrated and compensated for. Slow stage linearity and scale factor errors are also relatively constant and do not, therefore, require frequent correction data changes.
• On the other hand, workpiece-related errors are relatively variable and depend mostly on workpiece placement, offset, rotation, and dimensional variations among workpieces[0074]152. Therefore,geometry correction processor180 requires relatively large correction data changes every time workpieces152 are changed. Generating this correction data entails, for example, directingslow stages56 and58 to at least two, and preferably four, predetermined calibration targets on each associated workpiece152. Alternatively, in an embodiment where the vision system is working through the fast positioner, both theslow stages56 and58 andfast stages154 are directed toward the calibration targets. These calibration targets may be, for example, corners, tooling holes, or photoetch targets of an ECB. Each video camera160 senses differences between the directed and actual calibration target locations and provides difference data to system controlcomputer63 for processing. The resulting correction data for each workpiece152 are conveyed to and stored in the associatedgeometry correction processor180.
• For each fast[0075]stage signal processor172, corrected positioning data for the Y-axis are conveyed fromcorrection processors180 and182 to feedforward process86,galvo driver88, andfast stage154. Position feedback data are generated by position sensor120 (as in FIG. 2A) and combined for correction inadders184 and84. Skilled workers will understand that the same process applies to X-axis fast positioning.
• In applying the correction data to[0076]fast stages154, each fast stage is preferably limited to an 18 by 18 millimeter positioning range within its 20 by 20 millimeter maximum linear positioning range. The remaining 2 millimeters of positioning range is employed for applying the above-described corrections.
• The foregoing describes signal processing for a single axis of motion for each of the fast and slow positioner stages. Skilled workers will readily understand how to replicate the signal processing to coordinate the motion of both axes, both stages, and single or multiple fast positioners.[0077]
EXAMPLE
• A typical tool application employing[0078]positioner system50 and including Abbe error correction is laser cutting of holes, such as blind via holes, in multilayer ECBs orother workpieces62. Multilayer ECBs are typically manufactured by registering, stacking together, laminating, and pressing multiple 0.05- to 0.08-millimeter thick circuit board layers. Each layer typically contains a different interconnection pad and conductor pattern, which after processing constitutes a complex electrical component mounting and interconnection assembly. The component and conductor density trend of ECBs is increasing together with that of integrated circuits. Therefore, the positioning accuracy and dimensional tolerances of holes in ECBs is increasing proportionally.
• Unfortunately, the pressing step causes expansion and dimensional variations that lead to scale factor and orthogonality variations among the ECBs. Moreover, when multiple ECBs (workpieces[0079]152) are attached to slowstage58, fixturing variations can cause dimensional rotation and offset errors among the ECBs. Adding to that, ECB thickness variations make it difficult to mechanically drill holes having an accurately predetermined depth.
• [0080]Positioner systems50 or150 solve the above-described problems as follows. Two to four calibration targets can be etched at predetermined locations, preferably one at each corner, on each ECB. Video cameras160 sense differences between the commanded and actual calibration target locations and provide difference data to system controlcomputer63 for processing. The resulting correction data are conveyed to and stored ingeometry correction processor180.
• Two calibration targets provide sufficient difference data to system control[0081]computer63 to correct for rotation and offset variations among the ECBs. Three calibration targets provide sufficient difference data to system controlcomputer63 to correct for rotation, offset, scale factor, and orthogonality variations among the ECBs. Adding a fourth calibration target further allows for correction of trapezoidal distortion in each of the ECBs.
• ECB thickness variations are readily accommodated by the ±0.13-millimeter (±0.005 inch) laser depth of field.[0082]
• Processing blind via holes presents a difficult challenge for any hole processing tool because of the tight depth, diameter, and positioning tolerances involved. This is because blind via holes are typically processed through a first conductor layer (e.g., copper, aluminum, gold, nickel, silver, palladium, tin, and lead), through one or more dielectric layers (e.g., polyimide, FR-4 resin, benzocyclobutene, bismaleimide triazine, cyanate ester-based resin, ceramic), and up to, but not through a second conductor layer. The resulting hole is plated with a conductive material to electrically connect the first and second conductor layers. Blind via processing windows are presented in detail U.S. Pat. No. 5,841,099 of Owen et al.[0083]
• Referring again to FIG. 9,[0084]multi-head positioner150 is configured as an ECB blind via cutting apparatus in which N equals an even number, such as 2, 4, or 6, but preferably 4.Lasers156A and156C are UV lasers (wavelength is less than about 400 nanometers and preferably about 355 or 266 nm), andlasers156B and156N are IR lasers (wavelength is in a range from about 1,000 nanometers to about 10,000 nanometers, preferably about 9,000 nanometers). Because the UV and IR lasers have substantially different wavelengths, mirrors158 and optics forfast stages154 are configured for compatibility with each associated laser's wavelength.
• [0085]UV lasers156A and156C are capable of cutting both the first conductor layer and the dielectric layer in a suitable manner. However, the laser power levels and pulse repetition rates are carefully controlled to prevent unacceptable damage to the second conductor layer. This results in a narrow “process window.” Therefore,UV lasers156A and156C are preferably employed to cut through only the first conductor layer and a portion of the dielectric layer, a process that has a wide process window. Once the first conductor layer is removed by the UV lasers156,IR lasers156B and156N, which have a wide process window for cutting through the remaining dielectric layer without cutting through or damaging the second conductor layer, are employed to remove the last portion of the dielectric layer. Thus, the ECB blind via cutting apparatus employsUV lasers156A and156C to cut through the first conductor layers ofworkpieces152A and152C andIR lasers156B and156N to cut through the dielectric layers onworkpieces152B and152N.
• The time required for[0086]UV lasers156A and156C to cut through the conductor layers is typically longer than the time required byIR lasers152B and152N to cut through the dielectric layer. Therefore, the longer processing time dictates the processing throughput. Because the target locations are substantially identical for all tools onmulti-tool positioner150, the different processing times are accounted for by providing appropriately different laser power levels and pulse repetition rates for the UV and IR lasers.
• Some applications require cutting relatively large hole diameters of about 200 micrometers or less. Because[0087]UV lasers154A and154C have a beam diameter of only about 20 micrometers,multi-tool positioner150 must cause the UV beam to follow a spiral or circular path to cut such holes in a conductor layer. Therefore, cutting these relatively large holes takes a proportionally longer time. However,IR lasers154B and154N have a beam diameter of about 400 micrometers, which is about 20 times the UV laser beam diameter. Therefore, when cutting these relatively large diameter holes through the dielectric layers, at least some portion of the IR laser beam will cover the entire hole while the UV beam follows the spiral or circular path to cut a hole in a conductor layer. Under these circumstances, the IR laser beams are on the target locations for a relatively longer time and the different effective processing times are again accounted for by providing appropriately different laser power levels and pulse repetition rates for the UV and IR lasers.
• If suitable laser power is available, a single laser may be shared among multiple workpieces by employing suitable power splitting devices. It is also envisioned that switchable-wavelength lasers may be employed in this invention.[0088]
• This invention provides an improved combination of positioning accuracy, positioning speed, minimized or eliminated stopping time, nonpanelized tool path databases, and minimized fast stage movement range that dramatically improves processing throughput while reducing workpiece rejects caused by dimensional and orientation variations.[0089]
• Skilled workers will recognize that portions of this invention may be implemented differently from the laser beam micro-machining implementation described above. For example, a wide variety of tools, in single or multi-headed configurations, may be moved by the fast positioner stage, such as micro-dimensioned drills, punches, lasers, laser beams, radiation beams, particle beams, beam producing devices, microscopes, lenses, optical instruments, and cameras. Also, many different positioning devices may be employed in different combinations drawn from among galvanometers, voice coils, piezoelectric transducers, stepper motors, and lead screw positioners. The DSPs need not be completely digital and may, for example, include any suitable combination of analog and digital subcircuits. Of course, the positioning signal profiles, spectral bandwidth and amplitudes, and filter characteristics described herein may all be modified to suit the requirements of other positioning applications.[0090]
• It will be obvious to skilled workers that many other changes may be made to the details of the above-described embodiments of this invention without departing from the underlying principles thereof. The scope of the present invention should, therefore, be determined only by the following claims.[0091]

Claims (46)

1. A positioning system for positioning a tool relative to a target location on a workpiece in response to a positioning command, comprising:

a slow positioner for effecting a large range of relative movement between the tool and the workpiece, the slow positioner including a translation stage capable of movement generally along an axis;

a fast positioner for effecting small ranges of relative movement between the tool and the workpiece;

a positioning signal processor for deriving from the positioning command slow and fast movement-controlling signals;

a slow positioner driver for controlling the large range of relative movement of the translation stage in response to the slow movement-controlling signal;

a fast positioner driver for controlling the small ranges of relative movement of the fast positioner in response to the fast movement-controlling signal;

a pair of spaced-apart displacement sensors in communication with the fast positioner driver and coupled to move with the translation stage along the axis; and

a reference surface positioned in proximity to the translation stage and parallel to the axis, the translation stage being capable of moving along the reference surface and the displacement sensors being capable of acquiring information concerning their relative distances from the reference surface and conveying the information to the fast positioner driver to correct for an Abbe error associated with off-axis or rotational movement of the translation stage.

2. The positioning system of

claim 1

in which the rotational movement is yaw.

3. The positioning system of

claim 1

in which: the translation stage is a first translation stage that supports a workpiece; the slow positioner further comprises a second translation stage capable of moving along a second axis that is substantially perpendicular to the first axis; and the fast positioner is mounted on the second translation stage.

4. The positioning system of

claim 3

in which first and second translation stages carry respective mutually exclusive first and second inertial masses.

5. The positioning system of

claim 3

in which the fast positioner is a first fast positioner, and in which the second translation stage supports a second fast positioner.

6. The positioning system of

claim 3

in which multiple workpieces are mounted on the first translation stage.

7. The positioning system of

claim 3

in which the second translation stage comprises a second pair of spaced-apart displacement sensors in communication with the fast positioner driver and coupled to move with the second translation stage along the second axis, and in which a second reference surface is positioned in proximity to the second translation stage and parallel to the second axis, the second translation stage being capable of moving along the second reference surface and the second displacement sensors being capable of acquiring second information concerning their relative distances from the second reference surface and conveying the second information to the fast positioner driver to correct for a second Abbe error associated with second off-axis or rotational movement of the second translation stage.

8. The positioning system of

claim 7

in which the second rotational movement is pitch or yaw.

9. The positioning system of

claim 7

in which the second translation stage comprises a third pair of spaced-apart displacement sensors in communication with the fast positioner driver and coupled to move with the second translation stage along the second axis, and in which a third reference surface is positioned in proximity to the second translation stage, parallel to the second axis, the second translation stage being capable of moving along the third reference surface and the third displacement sensors being capable of acquiring third information concerning their relative distances from the third reference surface and conveying the third information to the fast positioner driver to correct for a third Abbe error associated with third off-axis or rotational movement of the second translation stage.

10. The positioning system of

claim 9

in which the third rotational movement is pitch or yaw.

11. The positioning system of

claim 9

in which the second translation stage comprises a fourth displacement sensor in communication with the fast positioner driver and coupled to move with the second translation stage in a plane including the second axis, and in which a fourth reference surface is positioned in proximity to the second translation stage, parallel to the second axis, and in a second plane generally including the second reference surface, the second translation stage being capable of moving along the fourth reference surface and the fourth displacement sensor in cooperation with one of the second displacement sensors being capable of acquiring fourth information concerning their relative distances from the respective fourth and second reference surfaces and conveying the fourth information to the fast positioner driver to correct for a fourth Abbe error associated with fourth off-axis or rotational movement of the second translation stage.

12. The positioning system of

claim 9

in which the fourth rotational movement is roll.

13. The positioning system of

claim 1

in which the displacement sensors comprise capacitive sensors.

14. The positioning system of

claim 1

in which the displacement sensors are capable of discerning relative distances as small as 10 nm.

15. The positioning system of

claim 1

in which the displacement sensors are capable of measuring relative distances as large as 50 μm.

16. The positioning system of

claim 1

in which the tool is a laser beam.

17. A positioning system for positioning a laser beam relative to a target location on a workpiece in response to a positioning command, comprising:

a first positioner effecting a large range of relative movement between the laser beam and the workpiece, the first positioner including a first translation stage capable of supporting the workpiece in plane and moving generally along a first axis;

a fixed optical head including a fixed optical path directed at, and transverse to, the plane of the workpiece;

a second positioner effecting a large range of relative movement between the laser beam and the workpiece, the second positioner including a second translation stage capable of supporting the fixed optical head and moving generally along a second axis that is transverse to the first axis;

a positioning signal processor deriving from the positioning command movement-controlling signals;

first and second positioner drivers controlling the large range of relative movement of the respective first and second translation stages in response to the movement-controlling signals;

a pair of spaced-apart first displacement sensors in communication with the positioner driver and coupled to move with the first translation stage along the first axis; and

a first reference surface positioned in proximity to the first translation stage and parallel to the first axis, the first translation stage being capable of moving along the first reference surface and the first displacement sensors being capable of acquiring information concerning their relative distances from the first reference surface and conveying the information to the positioner driver to correct for an Abbe error associated with off-axis or rotational movement of the first translation stage.

18. The positioning system of

claim 17

in which the rotational movement is yaw.

19. The positioning system of

claim 17

in which: the translation stage is a first translation stage that supports a workpiece; the slow positioner further comprises a second translation stage capable of moving along a second axis that is substantially perpendicular to the first axis; and the fast positioner is mounted on the second translation stage.

20. The positioning system of

claim 19

in which first and second translation stages carry respective mutually exclusive first and second inertial masses.

21. The positioning system of

claim 19

in which the fast positioner is a first fast positioner, and in which the second translation stage supports a second fast positioner.

22. The positioning system of

claim 19

in which multiple workpieces are mounted on the first translation stage.

23. The positioning system of

claim 19

in which the second translation stage comprises a second pair of spaced-apart displacement sensors in communication with the fast positioner driver and coupled to move with the second translation stage along the second axis, and in which a second reference surface is positioned in proximity to the second translation stage and parallel to the second axis, the second translation stage being capable of moving along the second reference surface and the second displacement sensors being capable of acquiring second information concerning their relative distances from the second reference surface and conveying the second information to the fast positioner driver to correct for a second Abbe error associated with second off-axis or rotational movement of the second translation stage.

24. The positioning system of

claim 23

in which the second rotational movement is pitch or yaw.

25. The positioning system of

claim 23

in which the second translation stage comprises a third pair of spaced-apart displacement sensors coupled to move with the second translation stage along the second axis and in communication with the fast positioner driver, and in which a third reference surface is positioned in proximity to the second translation stage, parallel to the second axis, and transverse to the second reference surface, the second translation stage being capable of moving along the third reference surface and the third displacement sensors being capable of acquiring third information concerning their relative distances from the third reference surface and conveying the third information to the fast positioner driver to correct for a third Abbe error associated with third off-axis movement of the second translation stage.

26. The positioning system of

claim 25

in which the third rotational movement is pitch or yaw.

27. The positioning system of

claim 25

in which the second translation stage comprises a fourth displacement sensor coupled to move with the second translation stage in a plane including the second axis and in communication with the fast positioner driver, and in which a fourth reference surface is positioned in proximity to the second translation stage, parallel to the second axis, and in a second plane generally including the second reference surface, the second translation stage being capable of moving along the fourth reference surface and the fourth displacement sensor in cooperation with one of the second displacement sensors being capable of acquiring fourth information concerning their relative distances from the respective fourth and second reference surfaces and conveying the fourth information to the fast positioner driver to correct for a fourth Abbe error associated with fourth off-axis movement of the second translation stage.

28. The positioning system of

claim 27

in which the fourth rotational movement is roll.

29. The positioning system of

claim 17

in which the displacement sensors comprise capacitive sensors.

30. The positioning system of

claim 17

in which the displacement sensors discern relative distances as small as 10 nm.

31. The positioning system of

claim 17

in which the displacement sensors measure relative distances as large as 50 μm.

32. A method for positioning laser output relative to a target location on a workpiece, comprising:

providing slow and fast movement-controlling signals from a positioning signal processor;

controlling with a slow positioner driver a large range of relative movement of a translation stage, generally along an axis and along a reference surface positioned in proximity to the translation stage and parallel to the axis, in response to the slow movement-controlling signal;

controlling with a fast positioner driver a small range of relative movement of a fast positioner in response to the fast movement-controlling signal;

effecting the large range of relative movement between the laser output and the workpiece on the translation stage;

acquiring, with a pair of spaced-apart displacement sensors coupled to move with the translation stage along the axis, information concerning their relative distances from the reference surface;

conveying the information from the displacement sensors to the fast positioner driver;

effecting with a fast positioner the small range of relative movement between the laser output and the workpiece including a correction for an Abbe error associated with off-axis or rotational movement of the translation stage; and

generating laser output to impinge the target location on the workpiece.

33. The method of

claim 32

in which the rotational movement is yaw.

34. The method of

claim 32

in which: the translation stage is a first translation stage that supports a workpiece; the slow positioner further comprises a second translation stage capable of moving along a second axis that is substantially perpendicular to the first axis; and the fast positioner is mounted on the second translation stage.

35. The method of

claim 34

in which first and second translation stages carry respective mutually exclusive first and second inertial masses.

36. The method of

claim 34

in which the fast positioner is a first fast positioner, and in which the second translation stage supports a second fast positioner.

37. The method of

claim 34

in which multiple workpieces are mounted on the first translation stage.

38. The method of

claim 34

in which the second translation stage comprises a second pair of spaced-apart displacement sensors in communication with the fast positioner driver and coupled to move with the second translation stage along the second axis, and in which a second reference surface is positioned in proximity to the second translation stage and parallel to the second axis, the second translation stage being capable of moving along the second reference surface and the second displacement sensors being capable of acquiring second information concerning their relative distances from the second reference surface and conveying the second information to the fast positioner driver to correct for a second Abbe error associated with second off-axis or rotational movement of the second translation stage.

39. The method of

claim 38

in which the second rotational movement is pitch or yaw.

40. The method of

claim 38

in which the second translation stage comprises a third pair of spaced-apart displacement sensors in communication with the fast positioner driver and coupled to move with the second translation stage along the second axis, and in which a third reference surface is positioned in proximity to the second translation stage, parallel to the second axis, the second translation stage being capable of moving along the third reference surface and the third displacement sensors being capable of acquiring third information concerning their relative distances from the third reference surface and conveying the third information to the fast positioner driver to correct for a third Abbe error associated with third off-axis or rotational movement of the second translation stage.

41. The method of

claim 40

in which the third rotational movement is pitch or yaw.

42. The method of

claim 40

in which the second translation stage comprises a fourth displacement sensor in communication with the fast positioner driver and coupled to move with the second translation stage in a plane including the second axis, and in which a fourth reference surface is positioned in proximity to the second translation stage, parallel to the second axis, and in a second plane generally including the second reference surface, the second translation stage being capable of moving along the fourth reference surface and the fourth displacement sensor in cooperation with one of the second displacement sensors being capable of acquiring fourth information concerning their relative distances from the respective fourth and second reference surfaces and conveying the fourth information to the fast positioner driver to correct for a fourth Abbe error associated with fourth off-axis or rotational movement of the second translation stage.

43. The method of

claim 40

in which the fourth rotational movement is roll.

44. The method of

claim 32

in which the displacement sensors comprise capacitive sensors.

45. The method of

claim 32

in which the displacement sensors are capable of discerning relative distances as small as 10 nm.

46. The method of

claim 32

in which the displacement sensors are capable of measuring relative distances as large as 50 μm.

Images