The opinion of the court was delivered by: Chin, D.J.
In this patent case, plaintiff Scanner Technologies Corp. ("Scanner") alleges that defendant
ICOS Vision Systems Corp., N.V. ("ICOS") infringes the claims of two of Scanner's patents (the "Patent's"):
U.S. Patent No. 6,064,756 (the "'756 Patent") and U.S. Patent No. 6,064,757 (the "'757 Patent"). The
parties waived their right to a jury and the case was tried to the Court. My findings of fact and conclusions
of law follow.
Scanner is a New Mexico Corporation, with its principal office in Minneapolis, Minnesota. ICOS is a Belgian corporation with its principal office in Belgium. (Pl. PFF ¶¶ 1, 2).*fn1
Ball grid arrays ("BGAs") and solder bumps on wafers and dies ("Bumps on Wafers") are electronic components that have small solder balls mounted on them in rows and columns that serve as electrical contact elements. (Tr. 31; PX 1 at col. 1; Def. PFF ¶ 1). BGAs are used in computer chips and can be found in devices such as personal computers, cellular telephones, electronic organizers, and compact disc players. Tens of billions of BGAs are produced every year. All the solder balls in each array must be positioned precisely at the same height, for a minute difference in height in any one ball in the array can render the BGA useless. Because the economics involved render repairs impractical, a defective BGA usually means the entire electronic device must be discarded. As a result, the industry has sought to develop an inspection machine to enable manufacturers of ball array devices to inspect BGAs and Bumps on Wafers in a fast and efficient manner. (Tr. 31-32).
The industry, including ICOS and Scanner, began searching for an apparatus and method for the three-dimensional inspection of ball array devices in the early 1990s. The Patents pertain to such an inspection device and method. (See PX 1; Tr. 33-36, 165, 450-51; Pl. PFF ¶ 20). The concept is to take two different views of the BGA and then to extrapolate three-dimensional information that will confirm the height of each ball. (Tr. 38).
Various kinds of inspection methodologies and types of inspection equipment have been available for some years for inspecting the solder balls on BGA devices. In particular, prior art to the Patents included laser range-finding technology, moiré interferometry, structured light pattern systems, and two-camera systems. (Def. PFF ¶ 2; PX 1 at col. 1).
The '756 Patent is an apparatus patent entitled "Apparatus for Three Dimensional Inspection of Electronic Components." (PX 1). The '757 Patent is a method patent entitled "Process for Three Dimensional Inspection of Electronic Components." (PX 2). The Patents relate to the three-dimensional inspection of BGAs and Bumps on Wafers. (PX 1 at col. 1; Tr. 62).
Applications for the '756 and '757 Patents were filed on May 28, 1999, and the Patents were issued on May 16, 2000, to Elwin M. Beaty and David P. Mork -- the two inventors. (Pl. PFF. ¶ 6; PXs 1, 2; Tr. 61). Mork assigned his rights in the Patents to Beaty, the CEO and majority shareholder of Scanner. Beaty then granted Scanner an exclusive right to the Patents. (Pl. PFF ¶ 6; Tr. 61; PXs 145, 146).
D. The ICOS Projector System
ICOS has been working on the three-dimensional measurement of BGAs since 1993. (Tr. 569). As early as November 1993, ICOS recognized that a market was emerging for BGA technology. (Tr. 569; DX 31 at 016800). It recognized that three-dimensional inspection of BGAs was a major development, and that "stereovision" --using two or more views of a camera --looked "most promising" as a technology. (Tr. 571; DX 31 at 016805). It was obvious that the use of two or more cameras would require calibration of the cameras to obtain measurements. (Tr. 571-72). In addition, as early as November 1993, ICOS recognized that different triangulation techniques could be used to obtain three-dimensional information. (Tr. 573-74; DX 31 at 016802). The use of triangulation in two-camera vision systems has been well known since before 1997. (Def. PFF ¶ 3; Tr. 406, 757).
ICOS began developing a two-camera system for the inspection of BGAs, the "Projector" system, in 1993 or 1994 and began selling it in November 1996, with publicly available brochures. (Tr. 372, 586-88, 596; Def. PFF ¶¶ 9, 10, 11; DX 12). The Projector system used a first camera in a normal position and a second camera in a position angled from the normal, as well as structured light from a projector. (Def. PFF ¶ 4; Tr. 588). The Projector system was prior art to the Patents. (Def. PFF ¶ 5; PX 1 at col. 1).
The image taken by the first camera in the Projector system produced a donut-shaped image because of the use of a ring light. (Tr. 588; DX 12). The Projector system contained a processor coupled to receive data from the two cameras. (Tr. 589; DX 12). A calibration reticle then calibrated the two-dimensional camera. (Tr. 590; DX 12). A manual from October 1997 for the Projector system explained the use of a triangulation principle to perform three-dimensional measurements, including the use of Z calibration and bilinear interpolation. (Tr. 593-94; DX 33 at 013517, 013520).
E. The ICOS CyberSTEREO System
The Projector system had problems with speed and reliability because of its projector illumination source. In the summer of 1998 ICOS began to consider removing the projector from the Projector system and changing to something different. (Tr. 378-89, 408-09, 597-600). By the fall of 1998, ICOS's efforts to improve the three-dimensional measurement of BGAs were well under way. (DX 21; Tr. 380-81, 436, 451; see also Tr. 597-615; DXs 34-36, 38-43).
In the fall of 1998, ICOS began testing a prototype of a new BGA inspection system that would eventually become the CyberSTEREO. (Tr. 388-91, 451, 612-13; DXs 24, 25; see also Tr. 615-16; DX 44). The CyberSTEREO was announced on January 26, 1999, and ICOS described it as a system different from its Projector system. (Tr. 49; PXs 4, 101). ICOS started converting existing Projector systems to the new CyberSTEREO system by mid-February 1999. ICOS was shipping new CyberSTEREO modules by mid-1999. (Tr. 393-95; PX 4). Subsequent generations of the product were called CyberSTEREO II and 3D Stereo. (Pl. PFF ¶ 37).
The CyberSTEREO products are an outgrowth of the Projector system. (Tr. 374, 377). The Projector system was designed to measure the absolute value of the top of the balls of BGAs. (Tr. 376). When the projector was taken out of the Projector system for the CyberSTEREO, the computer code was changed so as to measure coplanarity -- whether the balls were lying in the same plane -- instead of measuring the absolute value of the top of the balls. (Tr. 377, 408-10, 450-52). In addition, the CyberSTEREO measures a point inside the ball rather than the top of the ball. (Tr. 581-82).*fn2 This measurement, which involves certain assumptions, is simpler and faster. (Tr. 582). The CyberSTEREO is also faster and more reliable than the Projector system. (Tr. 409-10, 596-97).
The CyberSTEREO uses two digital cameras: the first (the 2D camera) looks straight down at a glass surface and the second (the 3D camera) looks at the surface from an angle or side view. (Tr. 715-16; DX P at 2). The cameras observe fiducial marks (points of reference) on the surface and record their locations on the surface, in pixels, horizontally and vertically. For example, a mark at (50, 40) would be located at the 50th pixel on the X axis and the 40th pixel on the Y axis. (Tr. 715-17; DX P at 3). The two cameras, however, will "see" different locations in each camera's field of view because of the differing angles. Hence, four fiducial marks with locations of (30, 30), (70, 30), (30, 70), and (70, 70) as seen by the 2D camera would be seen, for example, as locations of (40, 40), (70, 30), (40, 80), and (70, 70) by the 3D camera. (Tr. 717-18; DX P at 4, 5). The shift is the result of the distortion caused by the differing perspective of the side camera. (Tr. 717).*fn3 The locations for both cameras are stored in computer memory and the CyberSTEREO computes the distances between the marks and some ratios of distances between the marks. This is the calibration process. (Tr. 717-18).*fn4
The inspection process begins with the placing of a ball on the surface to be viewed by the cameras. The 2D camera sees the ball, for example, between two fiducial marks. The location as seen by the 2D camera is recorded, e.g., (40, 30). (Tr. 721; DX P at 6). The 3D camera sees the ball as being in a different location, e.g., (55, 35). (Tr. 721-22; DX P at 7).
The computer records this location as well and also computes the distances from the first fiducial mark to the center of the ball (D1) and from the center of the ball to the second fiducial mark (D2). (Tr. 722; DX P at 7). The ball is then "transferred" to the 2D camera, maintaining the distances D1 and D2 (respectively, from the first fiducial mark to the center of the ball and from the center of the ball to the second fiducial mark). "Transferred" means that the ball is placed in the approximate position where it should appear in the 2D camera, using the information obtained in the calibration process in a bilinear interpolation process. (Tr. 721-23).
The locations of the actual ball and the "transferred" ball in the 2D camera are compared as a ratio of distances, the distance between (1) the first fiducial mark and the center of the "transferred" ball and (2) the center of the "transferred" ball and the second fiducial mark. (Tr. 723-24). The comparison of the proportional distances is an approximation, not a trigonometric calculation. (Tr. 724). The distance between the centers of the two balls is the "shift," which is designated "dx" (DX P at 9), and ICOS assumes that the shift is proportional to the relative height of the ball. (Tr. 724-25). The higher the ball, the closer it is to the field of the view of the camera, in which case the "shift" appears bigger. (Tr. 725).*fn5 The dx is not measured at the top of the ball, but at the center of the ball, and even this is an approximation. (Tr. 733; DX P at 20).
All the calculations are done in pixel units. (Tr. 729). Because ICOS's customers prefer to deal in "real world coordinates" rather than "pixel coordinates," the pixel coordinates are converted to real world coordinates using "scale factors." (Tr. 730; DX P at 17, 18). Using the scale factors, the dx is converted from pixels to microns. (Tr. 731-33). The dx and scale factors are used to calculate -- or approximate --the relative height, designated "dz." (Tr. 807; DX P at 18).*fn6
The dx and dz are relative or differential values, not referenced by any particular point. (Tr. 734). If the dz values for all the balls in a BGA are the same, then the balls are all the same height. (Id.). If the dz values are not the same, that would mean the balls in the BGA vary in height and there would be a lack of coplanarity. (Tr. 734-36; DX P at 21).
The formulas for the calculations and functions used in the CyberSTEREO calibration and inspection processes have been written into the computer source code. (Tr. 736-49; DX 49; PX 75A). Much of the computer code was drawn from the code used for the Projector system. (Tr. 738-39, 752-54; DX 72).
Scanner invented an apparatus and method for the three-dimensional inspection of ball array devices in mid-1997, and filed a patent application on the apparatus and method in January 1998. (Pl. PFF ¶ 22; PXs 55, 176; Tr. 36-39). Scanner shipped its first embodiment of this invention, an inspection module called the Ultra Vim Plus ("UV") in July 1998, to its distributor, Yamatake-Honeywell, which delivered the module to Syntax in Japan. (Pl. PFF ¶ 23; Tr. 41).
Also in July 1998, Scanner displayed a UV module at the Semicon West trade show in San Jose, California, and again in December 1998 at the Semicon Japan trade show. (Tr. 41, 93, 820).
G. ICOS's Inquiry About the UV
Following the December 1998 Japan trade show, ICOS asked its various representatives to find out more about Scanner's UV system. (Tr. 399, 460-61; PXs 97, 159; Pl. PFF ¶ 29). In December 1998, DeProft suggested a meeting with Scanner's President, Beaty, to investigate a business relationship between the two companies. (Pl. PFF ¶ 33; Tr. 43-44, 436; DX 26). ICOS eventually sent Scanner a proposal to license the Scanner technology. (Tr. 48, 56, 384-85; PXs 3, 162; DX 23, 26). Scanner did not respond. (Tr. 218, 387).
Even as ICOS was approaching Scanner about a license, ICOS was continuing its efforts to modify its Projector system. (Tr. 385, 438-39). ICOS never, however, was given the opportunity to examine Scanner's technology, and ICOS did not have detailed knowledge of the inner workings of the Scanner machine or its mathematics or calibration process. (Tr. 386-87, 398-99; see PX 97). ICOS never saw Scanner's source code and could not have known Scanner's mathematical calculations. (Tr. 219). In developing its CyberSTEREO, ICOS did not have in its possession any Scanner systems, notes, internal manuals, or pictures. (Tr. 618).
In February 1999, at the Anaheim trade show, DeProft and Verjans on behalf of ICOS met with Beaty and Mork on behalf of Scanner. The meeting lasted only a few minutes, and DeProft and Verjans told Beaty and Mork that ICOS was no longer interested in the proposal that it had sent Scanner earlier, as ICOS had introduced its own new system in the meantime. (Tr. 448-49). DeProft and Verjans did not ask Beaty and Mork whether Scanner had filed any patent applications. (Id.; see Tr. 215-16).
H. The Petition To Make Special
Shortly after it filed its patent application in May 1999, Scanner filed a "Petition To Make Special," alleging that ICOS was infringing on at least one of the claims in Scanner's then-pending patent application, and asking for "special handling" of the application because of ICOS's infringing conduct. (DX 3; Tr. 221).*fn7
In a supporting statement, Scanner's attorney alleged that "I have made a rigid comparison of the . . . CyberSTEREO Ball Inspection System with the claims of the application, and . . . in my opinion, some of the claims are unquestionably infringed." (DX 5). In fact, the attorney had not seen the ICOS device at the time he made the statement. (Tr. 231, 543-44). Indeed, the attorney admitted at his deposition that he had not seen any ICOS products, not ...