JPH03215706A - Pattern inspecting device - Google Patents

Abstract

PURPOSE:To detect a microdefect which is nearly as large as a picture element to be inspected and to improve the reliability of the inspection by coding a detected pattern signal as a signal indicating a signal intensity distribution in picture element units. CONSTITUTION:An image detector 6 measures the quantity of light and outputs a pattern signal corresponding to the quantity of light. The pattern signal is stored in a frame memory 12 temporarily through an amplifier 109 and an A/D converter 11. Then a coding circuit 13 codes the signal intensity distribution in object image picture element units. Then a binary-coding circuit 15 matches the coded signal intensity distribution with the contents of an LUT14 to decide whether or not there is a pattern. Namely, a binary-coded pattern is generated. Then a defect decision circuit 16 inspects whether or not the pattern has a defect according to the binary-coded pattern. Consequently, even the microdefect which is nearly as large as the picture element to be inspected can be detected and the reliability of the inspection can be improved.

Description

[Detailed Description of the Invention] [Summary] The present invention relates to a pattern inspection device, particularly a technique for inspecting the presence or absence of defects in the wiring pattern of a printed circuit board, which makes it possible to detect even minute defects of the same size as the pixels to be inspected, and furthermore. The purpose is to increase the reliability of inspection by irradiating light onto a printed circuit board, forming an image of the reflected light from the board at a predetermined position, and outputting a pattern signal according to the amount of light in the formed image. an image detection optical system for digitizing the output pattern signal; and a means for digitizing the output pattern signal; A means for encoding a signal strength distribution representing how the signal strength is changing according to a predetermined standard, and a dictionary code in which the coded signal strength distribution for each pixel to be inspected is created and registered in advance. and a means for performing comparison and binarization, and is configured to inspect the presence or absence of defects in the wiring pattern based on the binarized pattern.

[Industrial application field]

The present invention relates to a pattern inspection device, and particularly to a technique for inspecting the presence or absence of defects (chips, protrusions, disconnections, shorts, dents, etc.) in wiring patterns of printed circuit boards.

As electronic computers become faster and more dense, the width and spacing of the wiring patterns on the printed circuit boards used for them are becoming narrower to 100 μm or less, and the outer diameter is also becoming smaller.
00n+n+ The angle is getting thicker. This wiring pattern may have defects such as disconnections and short circuits, so it is necessary to inspect the presence or absence of such defects. However, this inspection is no longer possible with human eyes due to the reliability and speed of the inspection. Therefore, there is a demand for the development of a method for automatically inspecting patterns.

[Prior art and problems to be solved by the invention]

In conventional pattern inspection methods, for example, a level of approximately 50% of the amplitude of a detected pattern signal is set as a threshold level, and the threshold level is set as the level of the detected pattern signal. The pattern was binarized and inspected for defects based on the results.

In other words, since there is only one reference level (threshold level) for binarization, micro defects of the same size as the pixel to be inspected, for example, Fig. 6 (a) and (b).
As shown in (c) and (d) of the same figure, for short-circuit defects or disconnection defects, the intensity of the detected signal increases its threshold level vth (from "H" level to "L" level). level or vice versa). Therefore,
A problem arises in that the defect cannot be detected even though it is originally a defect. This is undesirable because it leads to a decrease in detection accuracy and, in turn, to a decrease in reliability of pattern inspection.

Another technique is to set a plurality of threshold levels, each of which is used to detect a specific defect. However, with this method, 3
Since different types of binary images are obtained, the scale of the circuit required for inspection increases, and there is also the disadvantage that even dust and scratches that are not originally defects are excessively (erroneously) detected as defects. ．． Like the above-mentioned method, this form is also not preferable from the viewpoint of reliability of the test.

The present invention was created in view of the problems in the prior art, and provides a pattern inspection device that can detect even minute defects of the same size as the pixel to be inspected, and that can further improve the accuracy of inspection. is intended to provide.

[Means to solve the problem]

In order to solve the above problems, the present invention provides an apparatus for inspecting the presence or absence of defects in the wiring pattern of a printed circuit board, which irradiates the printed circuit board with light and directs the reflected light from the board to a predetermined position. an image detection optical system that forms an image and outputs a pattern signal according to the amount of light of the formed image; a means for digitizing the output pattern signal; and a predetermined number of digits for the digitized pattern signal. means for encoding a signal strength distribution representing how the signal strength of other pixels changes with respect to the signal strength of a specific pixel in accordance with a predetermined standard in units of pixels to be inspected; A means for performing binarization by comparing the signal intensity distribution of each pixel to be inspected with a dictionary code created and registered in advance, and inspecting the presence or absence of defects in the wiring pattern based on the disulfide pattern. There is provided a pattern inspection device characterized by the following features.

[Effect]

According to the above-described configuration, the pattern signal detected by the image detection optical system is not immediately binarized;
Once digitized, it is coded as a signal representing the signal intensity distribution in units of a predetermined number of pixels, and then binarized based on comparison with a predetermined dictionary code. Based on this binarized pattern, it is determined whether or not there is a defect in the wiring pattern.

In other words, since the presence or absence of defects is inspected after analyzing the distribution of signal intensity for each pixel, even minute defects of the same size as the pixel to be inspected can be detected with high precision. This contributes to improving the reliability of the test.

Note that other structural features and details of the operation of the present invention will be explained using the embodiments described below with reference to the accompanying drawings.

〔Example〕

FIG. 1 schematically shows a part of the configuration of a pattern inspection apparatus as an embodiment of the present invention.

In the figure, 1 is a printed circuit board including a wiring pattern;
2 is a table on which the printed circuit board is placed and is moved and adjusted in the X and Y directions by a drive mechanism (described later) during pattern inspection; 3 is a light source (for example, a halogen lamp); 4 is a table that guides the light from the light source for printing. an optical fiber that illuminates onto the substrate 1;
Reference numeral 5 denotes a lens that forms an image of reflected light from the printed circuit board 1 at a predetermined position, and 6 represents an image detector disposed at the predetermined position. The image detector 6 includes, for example, a CCD (charge-coupled device), measures the amount of light of an image formed at the predetermined position, and outputs a pattern signal corresponding to the amount of light.

Further, 10 is an amplifier that amplifies the detected pattern signal, 11 is an analog/digital (A/D) converter that digitizes the amplified pattern signal, and 12 is a pattern signal digitized by the A/D converter. Frame memory, l3, for temporarily storing
is a predetermined number of pixels to be inspected (in this example, 3×3 pixels) for the digitized pattern signal stored in the frame memory.
1 shows an encoding circuit that encodes a signal intensity distribution in units of pixels (pixel) according to a predetermined standard. This signal strength distribution will be described in detail later, but it represents how the signal strength of other pixels changes with respect to the signal strength of a specific pixel among the pixels to be inspected. 14 is a lookup table (LUT) storing dictionary codes created and registered in advance;
Reference numeral 15 denotes a binarization circuit that compares the output of the encoding circuit 13 with the contents of the LUT 14 to determine whether to output 1 or 0.

Reference numeral 16 indicates a defect discrimination circuit that discriminates whether or not there is a defect in the pattern based on the binarized pattern using, for example, radial matching logic; 17 indicates a device (for example, a CRT) that outputs the result of the defect discrimination; and 18 indicates a table 2. A drive mechanism for moving in the XY directions, 19 a drive mechanism for adjusting the movement so that the position of the image detector 6 matches the focal position of the imaging lens 5, and 20 a controller. The controller 20 controls the defect discrimination circuit 16 to determine the presence or absence of a pattern defect, controls the output device 17 to output a defect instruction, and controls the drive mechanism 18 . 19
have the function of controlling the position adjustment of the table 2 and image detector 6, respectively.

Next, encoding of the signal strength distribution will be explained with reference to FIGS. 2 and 3.

FIGS. 2(a) to (e) show various shapes of signal intensity curved surfaces. In this example, the characteristics of the signal intensity distribution are the curved surface characteristics around a specific pixel among the pixels to be inspected (3 x 3 pixels). It is classified based on the combination of signal strength and signal strength. In the figure, (a) shows the top of a single protrusion, (b) shows a part of a continuous ridge, (c) shows a part of a continuous valley, and (d) and (e) show parts of a continuous slope.

The binarization circuit 15 described above determines whether the pattern is normal or has a defect according to the combination of the shape of the curved surface and the signal intensity, and performs binarization.

At this time, the characteristics of the curved surface are expressed as codes, and the signal strength is also divided into a plurality of nodal regions (three in this embodiment). To determine which combination of surface characteristics and signal strength is 1 or O, use LUT1 as described above.
See 4.

For example, if the strength is low in the form of a single protrusion as shown in Figure 5(a), 0 (no pattern) is output, and if the strength is the same but in the form of continuous protrusions as shown in Figure 5(b), 1 ( (with pattern) is output.

FIGS. 3(a) to 3(c) show an example of encoding the signal strength distribution.

In this example, the pixels to be inspected are 3×3 pixels (p+~P,)
So, first of all. As shown in figure (a),
It is determined how the signal strength of each pixel changes with respect to the signal strength of the central pixel (ps in the illustrated example). The form of change is
High in the center (peak), low in the center (valley), flat (
flat), gradual rise (up), gradual decline (down)
There are five.

Furthermore, as shown in FIG. 6(b), the signal strength of the central pixel P is a relatively high (H) level disconnection detection level (see FIG. 6(d)) and a relatively low (L) level. It is classified into three levels: a short detection level (see FIG. 6(C)) and a pattern width measurement level (M) intermediate therebetween.

Combining the above, there are 4 types of signal intensity distribution types, 5 types of shapes, and 3 types of signal strong motions at the center pixel, and the number of combinations thereof is 5' x 3 = 1875 types.

FIG. 4 shows a flowchart of pattern inspection processing performed by the apparatus shown in FIG.

First, in step S1, the image detector 6 measures the amount of light, and a pattern signal is output in accordance with the amount of light. This pattern signal is temporarily stored in the frame memo 2 via the amplifier 1o and the A/D converter 11. In step s2, the encoding circuit 13 encodes the signal intensity distribution for each pixel to be inspected (3×3 pixels).

In step S3, the binarization circuit 15 converts the coded signal strength distribution into LUT. 14, and further, in step S4, it is determined whether there is a pattern (1 or O). In other words, a binarized pattern is created. In step S5, the defect determination circuit 16 inspects the pattern for defects based on the binarized pattern.

Finally, in step S6, it is determined whether the inspection has been completed for all pixels (YES) or not (NO), and the determination is Y.
In the case of ES, this flow becomes "end", and if the determination is NO, the process returns to step S1 and the above-described process is repeated.

As explained above, according to this embodiment, the pattern signal detected by the image detection system is once digitized and then coded as a signal representing the signal intensity distribution in units of a predetermined number of pixels. It is now binarized based on comparison with Based on this binarized pattern, the wiring pattern is inspected to see if there is a defect.

For example, as shown in FIG. 5(a), even if the binarized pattern has a certain signal strength, if the pattern has the form of a single protrusion, the circuit Since it does not affect operation, it is determined to be normal. In other words, they are considered dust or scratches that are not defects.

On the other hand, as shown in Fig. 5(b), the signal intensity is the same as the pattern of single protrusions in (a), but when the pattern is continuous (when it has the form of continuous protrusions), teeth,
It is determined that a short-circuit defect exists (see FIG. 6(a)).

In this way, since the presence or absence of defects is inspected after analyzing the distribution of signal intensity for each pixel, it is possible to increase detection sensitivity and accuracy compared to the conventional type. Therefore, even minute defects of the same size as the pixel to be inspected can be detected with high precision, and the reliability of the inspection can be increased.

〔Effect of the invention〕

As described above, according to the present invention, only practical defects such as short circuits and disconnections can be detected with high precision as defects, without excessively treating patterns such as dust as defects. This leads to improved reliability of the test and is extremely useful.

[Brief explanation of drawings]

FIG. 1 is a block diagram partially schematically showing the configuration of a pattern inspection device as an embodiment of the present invention, and FIG.
(e) is a diagram showing various shapes of the signal strength curved surface, Figures 3 (a) to (c) are diagrams for explaining coding of signal strength distribution, and Figure 4 is a pattern inspection performed by the equipment in Figure 1. Flowchart showing the process; Figures 5(a) and (b) are signal intensity distribution diagrams to explain the operation of the apparatus in Figure 1; Figures 6(a) to (d) illustrate problems in the conventional pattern inspection method. This is a diagram for explaining the points. (Explanation of symbols) 1... Printed circuit board, 2... Table, 3... Light source, 4... Optical fiber, 5... Imaging lens, 6...
- Image detector, 10... Amplifier, l1... A/D converter, l2... Frame memory, 13... Encoding circuit, l4... Lookup table (LUT)
, 15... Binarization circuit, 16... Defect discrimination circuit, 1
7... Output device, 18, l9... Drive mechanism, 20.
··controller.

Claims (1)

[Scope of Claims] An apparatus for inspecting the presence or absence of defects in the wiring pattern of a printed circuit board (1), comprising: irradiating the printed circuit board with light and focusing the reflected light from the circuit board on a predetermined position; An image detection optical system (5, 6) that outputs a pattern signal according to the light intensity of the formed image; and a means (1) that digitizes the output pattern signal.
1), and the signal strength distribution representing how the signal strength of other pixels changes with respect to the signal strength of a specific pixel in units of a predetermined number of pixels to be inspected for the digitized pattern signal. A means (13) for encoding according to a predetermined standard; and a means (13) for comparing the coded signal intensity distribution of each pixel to be inspected with a dictionary code (14) created and registered in advance to perform binarization. 15) A pattern inspection apparatus characterized in that the pattern inspection apparatus is characterized in that it inspects the presence or absence of defects in a wiring pattern based on the binarized pattern.