JPH02260600A - Method of forming reference patterns for detecting defects of printed internal wiring patterns or the like - Google Patents

Abstract

PURPOSE:To facilitate both 'strict inspection' and 'lerient inspection' for inspecting printed internal wiring patterns in the same visual field by a method wherein reference patterns of pattern elements and pattern element spacings for inspecting pattern defects are provided. CONSTITUTION:Data are extracted by design data from specific regions (mounting part (b) and wire bonding part (e)) and other regions. Relatively fine patterns (i) and (j) about the size of 3 picture elements (portion of width (n)) are formed for pattern elements (g) and pattern element spacings (h) from original pictures extracted from the specific regions. On the other hand, patterns about the side of 1 or more picture elements are formed for pattern elements and pattern element spacings from original pictures extracted from the other regions. By synthesizing them, patterns which facilitate strict inspection for the mounting part (b) and the wire bonding part (d) are obtained. The defect inspecting reference patterns obtained by making the synthesized patterns as described above facilitate accurate inspection of the regions requiring different judgement levels, i.e., the mounting part, etc., which require strict judgement and the other regions which allow lenient judgement.

Description

[Detailed Description of the Invention] Field of Application in CML Industry) The present invention relates to a method for creating a standard pattern for defect inspection of printed internal wiring patterns, etc., and more specifically, to a method for creating a standard pattern for defect inspection of printed internal wiring patterns, etc. Standard patterns for defect inspection, such as printed internal wiring patterns, that enable strict inspection and loose inspection within the same field of view without requiring input images from a camera. Regarding how to create.

[Conventional technology]

Conventionally, multilayer printed wiring boards have been widely used, and there is a demand for efficient manufacturing technology for this type of multilayer printed wiring boards. The multilayer printed wiring board is constructed by laminating a plurality of printed wiring patterns for electrical insulation and conduction. In recent years, these printed circuit boards have not only become smaller in order to achieve higher integration, but also the pattern elements and the spacing between the pattern elements have become narrower. Testing is also becoming more difficult.

In this quality inspection, each printed wiring pattern is usually inspected for defects separately, and after these are laminated together, the completed product is further inspected, but defects are discovered immediately after printing. Since it is considered that this greatly contributes to the manufacturing yield of the product, it is desirable to remove or correct defective products as much as possible at this point. Testing methods have been proposed. However, in the case of conventional inspection methods,
There are problems such as the inspection takes a lot of time and the defect shape cannot be accurately inspected.

Therefore, the present inventors first addressed this problem by r
A method for creating a standard pattern for defect inspection of pattern elements and pattern element intervals in an inspection pattern such as a printed internal wiring pattern, in which the inspection target is determined based on the planar spread of the viewpoint on the line of the wiring pattern from design data such as CAD. An inspection target range setting step that determines the range, a field of view division step that divides the inspection target range obtained in the inspection target range setting step for each field of view of the camera, and considering the balance between the inspection target range and the divided field of view range, The field of view coordinate setting process involves determining the center position of the camera's field of view and shifting the coordinates for each field of view, and extracting specific areas from each field of view, such as the area that will become the mount part of the tC chip and the area that will become the wire bond part. ,
A specific region coordinate setting step for determining the coordinates of the nuclear field of view and its specific region; In each field of view, areas other than the above-mentioned specific region are set to have a pattern width or pattern spacing that is less than or equal to a specific pattern prime width or pattern spacing. A defect inspection standard pattern is created with a width of at least one pixel, and the specific area has at least an I
Create a standard pattern for defect inspection with a width smaller than N elements,
a standard pattern creation process for defect inspection in which the two patterns are finally combined; and a print internal placement pattern that enables "strict inspection" and "relaxed inspection" within the same field of view. proposed a method 1 for creating a standard pattern for defect inspection (see the specification of Japanese Patent Application No. 164533/1983).

According to this production method, the above-mentioned problems can be solved and an extremely accurate inspection can be realized.

(Problems to be Solved by the Invention) However, in the case of the above-described method for creating a defect inspection standard pattern such as a printed internal wiring pattern, for example, when creating a defect inspection standard pattern for a multilayer printed wiring board, the method shown in FIG. 14(a) As shown in , ■CAD data extraction process, ■Pattern identification information extraction process, ■Layered printing pattern superimposition process, ■Positive creation process, ■Sfreefleaf creation process, ■Tape printing preliminary test implementation process, ■Prototype and Multiple steps are required: tape selection process, camera insertion process, image capture process for each field of view, and [phase] standard pattern creation and storage process.

More specifically, as shown in FIG. 14(b), first, an arbitrary pattern is extracted from the design data through a CAD data extraction process, and then pattern information is extracted through a pattern identification information extraction process. After extraction, the design data is converted into cutting data in the process of superimposing the laminated printing pattern, and red negative cutting is performed using a cutting block, and this is repeated for 2 to 3 layers to create superposition positive information. Then, in the next positive creation step, a positive is created using the positive information. Here, the positive is usually created by a positive manufacturer, and is attached to one sheet.
10,000 to 20,000 yen is required, and an acceptance inspection is required.
Next, plate making is performed in the screen creation process using the created positive, and in the next tape printing preliminary test implementation process, printing conditions are first set, preliminary printing is performed on the tape, and whether the printing is normal or not. If it is normal, print several tapes, and then use the tape selection process that will become the next prototype to extract the tape that will become the standard for creating the standard pattern. , and in the image capture process for each field of view, each field of view is captured by a camera, and in the next standard pattern creation/storage process, pattern elements and standard patterns between pattern elements are created using information about all fields of view. do. Here, a mock inspection is performed on the pattern. This mock inspection uses the pre-printed tape as the standard pattern creation standard, but it often deviates from the actual design data depending on the printing conditions (40
(about 60 μm), depending on the results of the mock test, we changed the printing conditions and performed the following steps: ■ Preliminary tape printing test implementation process, ■ Tape selection process to become a prototype, ■ Camera insertion process,
■Image capture process for each field of view, 0 standard pattern creation,
It is necessary to repeat each step of the storage process (actually, it is repeated about 10 times).

Therefore, in the case of the above-mentioned method, when creating a defect inspection standard pattern such as a printed internal wiring pattern, the identification information and position correction information of the internal wiring, chip mounting part, and wire bonding part are Although it is obtained from information, the image itself of the pattern, which is the prototype of the standard pattern created inside and outside the pattern, is
Since it is necessary to use the information input by sensors such as cameras, we need to: ■ Overlay process of laminated printing patterns, ■ Positive creation process, ■ Screen creation process, ■ Preliminary tape printing test implementation process, ■ Tape selection as a prototype. Processes: ■Inserting the camera; [F]Multiple steps are required to capture images for each field of view, which increases the number of creation steps, and also conducts mock inspections, and based on the results, ■Preliminary tape printing test implementation process;■ It is recognized that there are inadequacies in that each process has to be repeated, including the process of selecting a tape to serve as a prototype, (1) loading the camera, (2) capturing an image for each field of view, and (1) creating and storing a zero standard pattern.

The present invention has been devised in response to the above-mentioned problems, and its purpose is to provide a printed interior that can be created with fewer steps and without the need for input images from a camera. An object of the present invention is to provide a method for creating a standard pattern for defect inspection such as a wiring pattern.

[Means for solving problems]

As a means for achieving the above object, the method for creating a standard pattern for defect inspection of printed internal wiring patterns, etc. of the present invention is for defect inspection of pattern elements and pattern element intervals in inspection patterns such as printed internal wiring patterns. A standard pattern creation method that includes an inspection target range setting process in which the test target range is determined from the planar spread of the viewpoint on the line of the wiring pattern from design data such as CAD, and the inspection obtained in the test target range setting process. A field of view division process in which the target range is divided into each field of view of the camera, and a field of view coordinate setting that determines the center position of the camera's field of view and shifts the coordinates of each field of view, taking into account the balance between the inspection target range and the divided field of view range. Extract specific areas such as the area that will become the mounting part of the IC chip and the area that will become the wire bonding part from the process and each field of view,
A specific area coordinate setting step for determining the coordinates of the field of view and its specific area, and converting the vector information of the pattern portion within each field of view into raster information, enlarging/reducing this, and inputting the input image from the camera. In each field of view, in areas other than the above-mentioned specific area, at least one A standard pattern for defect inspection is created with a width equal to the size of a pixel, and a standard pattern for defect inspection is created in the above specific area with a width at least one pixel smaller than the pattern element width or pattern element interval, and both patterns are finally and a step of creating a standard pattern for defect inspection, which is synthesized with the standard pattern.

According to the above configuration, in the conventional example, ■positive creation process, ■sleeve creation process, ■tape printing advance test implementation process, ■prototype tape selection process, ■camera insertion process, and image capture for each field of view. Since the process can be replaced with one step of the image information setting process for changing vector information to raster information, there is no need for a camera to capture input images or to create a positive screen.
The production process can be reduced and the production cost can be reduced.

〔Example〕

Hereinafter, embodiments embodying the present invention will be described with reference to the drawings.

Here, FIGS. 1 to 13 are detailed explanatory diagrams of the present invention, in which FIG. 1 is an explanatory diagram of a standard divided state, FIG. 2 is an explanatory diagram of a protruding state of the included part, and FIG. 3 is a diagram of a designed state. An explanatory diagram of a design example in data, Fig. 4 is an explanatory diagram of data extraction of the mount part, Fig. 5 is an explanatory diagram of wire bonding detection, Fig. 6 is an explanatory diagram of a pattern in the extracted data, and Fig. 7 is an explanatory diagram of overlapping data. An explanatory diagram showing the combined state, Fig. 8 is an explanatory diagram of raster quantization in a specific field of view, Fig. 9 is an explanatory diagram to explain the development of the quantized area, and Fig. 10 is an explanatory diagram expressed as a focus row. , Fig. 11 (a) and (b) are explanatory diagrams for explaining the development from temporary information to real information, Fig. 12 is an explanatory diagram of the synthesis order of standard patterns for defect inspection, and Fig. 13 is standard pattern creation. It is a flow diagram for explaining a process.

This example describes a method for creating a standard pattern for defect inspection used for detecting defects in an ICC70 board. It consists of six steps: a setting step, a zero image information setting step, and a standard pattern creation step for defect inspection.

One information extraction process One process is a process of extracting necessary information from design data such as CAD. That is, this is a step of extracting the minimum value, maximum value, and design reference vector center from the design data regarding the finger point coordinates for designing wiring patterns such as straight lines and circular arcs.

- Inspection Range Setting Step - This step is a step of setting an inspection range based on the minimum and maximum values of the X and Y coordinates of the wiring pattern extracted in the information extraction step and design criteria. In other words, design data such as CAD is often vector data with respect to a reference point, and the data extracted in the previous process is extracted as relative position information. Minimum value (XIIIN, Yw+w)
, maximum value (Xnall * YNIk weight), the center of the design standard vector, that is, the in-plane slope of the center position of the wiring pattern is detected, and thereby the range to be inspected as the design standard is determined.

One field of view division process The one field of view division process is a process in which the inspection target area is divided at regular intervals depending on the field of view size of the inspection. By dividing the inspection target range in response to the difference, it is possible to image the entire inspection target range with the camera, and the division of the camera's field of view is based on the , the minimum value of the y coordinate (XNIN, Yg
+w), the maximum value (XIIAII, Ynam) is divided as shown in Figure 1, that is, XNA
The difference between l and XNIN, Y, 11. Divide the difference between I'm trying to go.

-Visual field coordinate setting step- This step is a step of determining the center position of the field of view of the camera and shifting the coordinates for each field of view. This is done in consideration of the balance between the inspection target range and the divided visual fields. And X
The adjustment of the balance in direction will be explained using FIG. 2, which is an enlarged view of the final inclusion part a of the division in the X direction in FIG. Xo, Ye) are stored, so the amount of protrusion ΔX in the X direction is Δx = Xo +5IYAX / 2 + DA[
It is calculated as 1UR11xxxax. Similarly, the amount of protrusion Δ in the y direction
y is A y = Ye + 5tvay / 2 + DA
It is determined by the BURI amount yYNAm. However, 5IYAXSSIY^Y is the size of the division amount in the X direction and y direction, DABURI l x
, DABIIRlliy is the overlap size with the adjacent field of view.

Then, the division start point of the reference point is the coordinate (XMIN
Since the starting point is IYNIII), a well-balanced inspection visual field can be determined by shifting the center position of the visual field by the coordinates (-Δx/2.-Δy/2). In particular, it is an effective means even when inspecting only the field of view.

Specific area coordinate setting step - This step is a step of extracting a specific area to be inspected from each field of view of the camera, and setting the field of view and the coordinates of the specific area. Here, the specific area refers to the area that will become the mounting part (cavity part) of the IC chip and the area that will become the wire bonding part.

By the way, if the design of CA[) is standardized, data extraction of only the mount portion (cavity portion) can be easily performed. However, as a practical matter, when data is modified, added, deleted, etc., it is difficult to extract only the cavity data from among the data.

Therefore, the coordinate setting method will be explained with reference to FIGS. 3 to 5. In the design example in Figure 3, originally, finger point 1
4 was supposed to form one closed surface (including circular arcs), but after creating the closed surface β on the way, the closed surface α is again formed using the finger points 8 to 10. In order to deal with such chaotic data, artificial intelligence type language is usually used to judge the data of the mount (cavity). The structure of the knowledge database is as follows.

■ Examine the start and end points of all finger points to determine whether they have a possibility of passing through the same point and form a closed surface; ■ whether the points exist between the minimum and maximum values obtained in the information extraction process;
The points intersecting the straight lines in the X and Y directions are subjected to R processing; (2) The number of R is four; FIG. 4 shows how cavity data is extracted. In other words, design data ('IJ14 diagram left)
The data of the mount part (cavity part) b is extracted from the mount part (cavity part). Note that C is the outer shape and d is the center line.

Now, in order to recognize the wire bonding part, after extracting the data area of the mount part (cavity part) b, measurements are taken from the boundary, and the coordinates of the existing finger points are determined for the first time. This state is shown in FIG. 5, that is, the finger points in the X and Y directions are determined and stored. Since the distance of the wire bonding part e from the tip of the mount part (cavity part) b is determined for each user's word C chip product, its position is memorized.

Then, for each field of view including these data, its f(+
x point) is calculated and stored as a strict standard pattern creation area.

-Image information setting process- This process is a process of converting the vector information of the pattern portion into raster information within each field of view, and enlarging/reducing this to make it the same information as the input image from the camera. It is.

In each of the above-mentioned steps, each data extracted from design data such as CAD is stored in a data storage area called a layer for each upper and lower pattern to be stacked.

By the way, when a printed green sheet is viewed from above with a camera, the upper and lower patterns appear to be combined. Therefore, in this process, first, the identification information of the necessary pattern is extracted from the above layer.
(see figure), the laminated printing patterns are superimposed (see Fig. 7), and then this vector information is converted into raster information.

Here, a method of developing the superimposed diagram of FIG. 7 obtained by the above superimposing process into raster information will be described.

By the way, data created from vector information forms boundaries of a certain specific color, and is like an independent island of a closed surface. Then, the inside of the closed surface is filled with a specific color up to the boundary line, and is developed into pixel information in the X direction and the y direction as shown in FIG. Note that FIG. 8 is a diagram within the specific field of view after being filled in. When FIG. 8 is viewed microscopically, it can be seen that it is a collection of bits in the X direction and the Y direction. Here, the process of extracting information from the area where information for each bit is stored (hereinafter referred to as the quantization area) is shown in Fig. 9. In Fig. 9, the square divisions indicate the quantization area. The black visual part therein is the pattern part, and has dot information 1 (dot information outside the pattern is 0). Therefore, when this information is expressed by a bit string, it is expressed as follows.

x+y+-0,XtVt-01...XIIFI-
OxrVm-Or XtVt-Or...X
syt-0 Then, in the part expressed as a figure, this focus row is developed into figure information at the 0th order, which is pit information 1.

The size of this quantization area depends on the size of the image memory where the image captured by the input camera is stored.
Normally, if this size is 512 pixels x 512 pixels,
The area shown in FIG. 10 may be approximately 256 pixels by 256 pixels. The temporary information of the actual bit string is enlarged by the method shown in FIG. 11 to become actual information. That is, ll+Y+
In the case of = 1, it expands to XI-+ 7+-+ = 1

According to this method, processing can be performed four times faster than when expanding to 512 pixels x 512 pixels. Also, when looking at the actual printed pattern microscopically, there will be some unevenness at the border due to the mesh of the screen for screen printing, so if you take this into consideration, it will be difficult to
This is enough raster information.

In this way, in this step, the vector information of the pattern portion is converted into raster information to obtain information equivalent to the input image from the camera.

1. Standard pattern creation process for defect inspection - This process creates standard patterns for defect inspection corresponding to each of the specific area and the area other than the specific area, and
This is a process of combining both patterns to create a standard pattern for defect inspection. In other words, data within the field of view is extracted in advance based on the data about the strict region obtained in the previous process, and areas other than the specific area are filled with a specific pattern element or a pattern element with a pattern element spacing or less. A standard pattern for defect inspection having a width of 1 pixel relative to the pattern element interval is created, and a standard pattern for defect inspection having a width of 3 pixels smaller than the pattern element or pattern element interval is created in the specific area, and This is the process of finally combining both patterns to create a standard pattern for defect inspection.

Standard patterns for defect inspection, which are created as stricter patterns than previously extracted data, are usually
The original pattern is created as a thin pattern of about 3 pixels (this can be set arbitrarily depending on the target pattern). In the process mentioned above, vector information is developed into raster information, so each The logical OR of the patterns is taken to obtain a standard pattern for defect inspection (see FIG. 12).

That is, in FIG. 12, data of a specific area (mount part b, wire bonding part e) and areas other than the specific area are extracted using the design data of FIG. From the original image shown in (b), as shown in FIG. 12 (b'), each pattern element g1 has a narrow pattern of approximately 3 pixels (width n portion in the drawing) for each pattern element interval (this pattern is
(Can be set arbitrarily depending on the target pattern) 1. 12 (C) for areas other than the specific area.
), regarding the pattern element and pattern element interval,
Create a pattern of 1 pixel or more (pattern 1 with 1 pixel interval in a pattern with pattern pixel width), and apply this to the 12th
As shown in Figure (d), they are synthesized to obtain a pattern with a stricter pattern for the mount portion b1 and the wire bonding portion d.

The standard pattern for defect inspection created in the above-described embodiment is obtained by enlarging the IC chip inspection pattern with a camera, processing the image at the optimum level of brightness, and storing it as a binarized image on the image memory. By comparing this binarized image with combiator processing, etc., it is possible to accurately and stably determine different judgment areas, such as the mount part, which requires a strict judgment, and areas where a loose judgment is acceptable, depending on the synthesized pattern. It works so that it can be inspected under suitable conditions.

In addition, when creating a standard pattern for defect inspection, we have previously required ■Positive creation process, ■Sfreefle creation process, ■Tape printing preliminary test implementation process, ■Prototype tape selection process, ■Camera insertion process,■ Since the image capturing process for each field of view can be replaced with a single image information setting process for changing vector information to raster information, there is no need for a camera or positive screen creation for input image capturing. The manufacturing process can be reduced and the manufacturing cost can be reduced. In addition, the standard pattern creation standard is based on vector information extracted from design data such as CAD, and is not based on the previously printed tape as in conventional methods, so the occurrence of "misalignment 1" due to printing conditions is taken into account. This works to eliminate the need for mock tests.

It should be noted that the present invention is not limited to the above-described embodiments, but includes modifications that can be made without departing from the gist of the present invention. Incidentally, in the above-mentioned embodiment, the pattern element and pattern spacing in the specific area (mounting area of the internal wiring pattern, area of the wire bonding area) are 3 times smaller than the binary image in the original image.
A standard pattern for defect inspection is created for strict inspection, consisting of pixels (width) as narrow as a pixel, and a pixel of one pixel size is used for pattern guidance in areas other than specific areas and between patterns. Although the configuration has been described in which a standard pattern for defect inspection is created to perform a loose inspection with a width of In this case, create a pattern consisting of pixels (width) that are at least one pixel smaller than the original pattern element and the original pattern element interval (binarized image), and for the pattern elements and pattern element interval in areas other than the specific area. , depending on the width of the original pattern element and the original pattern element interval (binarized image), but it is 2 m.
! Pixel size of 1 pixel or more, such as pixel, 3 pixel, etc.
It is clear that it may be created as a width).

〔Effect of the invention〕

As is clear from the above explanation, the method of creating standard patterns for defect inspection such as printed internal wiring patterns of the present invention extracts the spread of finger points from design data such as CAD, and divides the pattern according to the field of view of the set camera. At the same time, the coordinates of the field of view are moved, and based on this, for example, a pattern is created that can detect areas where severity is required and areas where severity is not required, and when creating the pattern, each field of view is The vector information of the part that becomes the pattern is converted into raster information, and this is enlarged/reduced to make it the same information as the input image from the camera. There is no need to create positive screens, which reduces the creation process and production costs.In addition, during inspections, instructions in strict areas that could not be automated in the past can be automated instead of being instructed by humans. It has this effect.

[Brief explanation of drawings]

1 to 13 are detailed explanatory diagrams of the present invention, and the first
The figure is an explanatory diagram of the standard division state, Fig. 2 is an explanatory diagram of the protruding state of the included part, Fig. 3 is an explanatory diagram of a design example in design data, Fig. 4 is an explanatory diagram of data extraction of the mount part, and Fig. The figure is an explanatory diagram of wire bonding detection, Fig. 6 is an explanatory diagram of patterns in extracted data, Fig. 7 is an explanatory diagram showing a superimposed state, Fig. 8 is an explanatory diagram of raster quantization in a specific field of view, Figure 9 is an explanatory diagram for explaining the expansion of the quantization area, Figure 1θ is an explanatory diagram expressed as a bit string, and Figures 11 (a) and (b) are explanatory diagrams for explaining the expansion from temporary information to real information. Explanatory diagram, FIG. 12 is an explanatory diagram of the synthesis order of standard patterns for defect inspection, FIG. 13 is a flow diagram for explaining the standard pattern creation process, and FIG. 14 <a>
(b) is a flow diagram for explaining a standard pattern creation process in a conventional example. a...Inclusion part, b...Mount part (cavity part)
%C...Outline, d...Center line, e...Wire bonding part, f...Reference point, g...Pattern element, h
...Pattern element interval, 1j...Pattern, k...
Pattern of pattern convergence, l... Pattern patent of pattern disjoint interval Applicant: Narumi Seito Co., Ltd. Agent Patent attorney: Hiroshi Yoshimura Figure 1 Figure 4 Figure 511! ! I 10χ Figure 6 Figure 3 Figure 13 Procedural amendment (support) 1. Indication of the case 1 year patent application No. 81746 2. Name of the invention n- of standard pattern for defect inspection of printed internal wiring patterns, etc. 3. Relationship with the person making the amendment Patent applicant Kunio Nishioka 4. 4 major J-persons 5. Date of amendment order July 4, 1999 (old shipping) 6. Subject of amendment Figures 6, 7, and 12 of the drawings (7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7)) (I will revise the attached drawing clearly drawn in black.) Figure 6 Figure 7 (a) Figure 12

Claims (1)

[Claims] (1) A method for creating a standard pattern for defect inspection of pattern elements and pattern element intervals in an inspection pattern such as a printed internal wiring pattern, which is based on design data such as CAD, and the planar spread of the viewpoint on the line of the wiring pattern. an inspection target range setting step in which the inspection target range is determined by a field of view division step in which the inspection target range obtained in the inspection target range setting step is divided into each field of view of the camera, and a balance between the inspection target range and the divided field of view range. A field of view coordinate setting process in which the center position of the camera's field of view is determined and the coordinates for each field of view are shifted, and specific areas within each field of view, such as the area that will become the IC chip mount part and the area that will become the wire bond part. A specific area coordinate setting step in which the coordinates of the field of view and its specific area are determined, and in each field of view, the vector information of the part that becomes the pattern is converted to raster information, which is enlarged/reduced, and the camera An image information setting step to make the information equivalent to the input image from A standard pattern for defect inspection is created with a width of at least one pixel size, and a standard pattern for defect inspection is created in the specific area with a width of at least one pixel smaller than the pattern element width or pattern element interval. A method for creating a standard pattern for defect inspection such as a printed internal wiring pattern, comprising: a step of creating a standard pattern for defect inspection in which both patterns are finally combined.