JPH0680954B2 - Device for mounting IC on printed circuit board - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Field of Application] The present invention enables effective mounting of a surface mounting package (flat package) IC on a printed circuit board, particularly positioning between the IC and the printed circuit board. Onboard equipment.

[Conventional technology]

Conventionally, mounting of an IC such as an electronic component on a printed circuit board is disclosed in, for example, Japanese Patent Laid-Open No. 60-1900.

In this conventional example, a special mark for positioning is provided on the substrate, the mark is picked up and then confirmed, and the IC is positioned and mounted on the substrate.

[Problems to be Solved by the Invention]

The above-mentioned conventional example has a problem that a design change is required to provide a special positioning mark on the substrate, or a special movement is required for the television camera to capture the special mark. . In addition, the field of view of the TV camera is in principle the entire printed circuit board, but special marks must be made small on a part of the board, and the TV camera must be moved to detect these small special marks. Then, it was necessary to take an image under the enlarged field of view.

On the other hand, in the "automatic mounting system for imposition components" (Japanese Patent Application No. 60-224852), which is the same applicant, a special mark for positioning is required on the board, and the TV camera captures the special mark. You don't have to make a special move for
With one field of view, it is possible to position and mount face-to-face parts such as ICs on the board with high speed and high accuracy.

The present invention further improves the present invention, and calculates the center coordinates of the IC lead row or the contact pattern row on the substrate even when the IC lead row or the contact pattern row on the substrate has asymmetry or partial omission. Provided is an apparatus for mounting an IC on a printed circuit board, which is capable of calculating the center and inclination of the IC or the center coordinates and inclination of a contact pattern on the board and aligning the IC with the board with high accuracy.

[Means for Solving the Problems]

In the present invention, attention is paid to the fact that the leads or contact patterns on the printed circuit board arranged in rows on two or four sides of the IC are regularly arranged, respectively. Even if there is a rule or a part missing, a means to calculate the center coordinates of each row considering the irregularity or part missing of the lead row or contact pattern row imaged by the TV camera, and the calculation result From the above, a means for calculating the center and inclination of the IC or the center coordinates and inclination of the contact pattern was provided in the onboard device.

Furthermore, when calculating the center coordinates, the memory that stores the correction data that is calculated for irregular or partially missing arrays and the lead or contact pattern of an IC that has irregular or partially missing arrays are The mounting device is provided with means for calculating the center coordinates, center position and inclination of the lead row or contact pattern row using the correction data.

[Action]

A plurality of leads forming a row of two or four sides of the IC and a plurality of contact patterns on the printed circuit board corresponding to the leads are imaged by a television camera, and each of the imaged leads of the IC or the contact pattern on the board is picked up. An image of a row is projected on a screen within a rectangular range set in advance in the image processing apparatus, a measurement line is set on each row of the lead or contact pattern within the rectangular range, and the lead or contact on the measurement line is set. The brightness data of the pattern is stored in the order of arrangement in a memory, and the center coordinates of the lead row or contact pattern row are calculated by specifying the boundary position between the lead or contact pattern and the background from the brightness data in the memory. The center and inclination of the lead or the center coordinate and inclination of the contact pattern can be calculated from the center coordinates.

Furthermore, paying attention to the fact that the lead or contact pattern is a regular array, and even if the array is irregular or partially missing, by replacing it with a regular array, Similarly, the center coordinates of the lead row or contact pattern row can be calculated, and the center and inclination of the lead or the center coordinate and inclination of the contact pattern can be calculated from the center coordinates.

Furthermore, in reality, since array irregularities or partial omissions are often intentionally provided, the center coordinates of each row of leads or contact patterns with irregular or partial omissions of the array are calculated first. By storing the correction data calculated in advance in the memory, the correction data can be used to calculate the center coordinates of the lead row or the contact pattern row having the same irregularity or partial omission.

〔Example〕

Figure 1 shows the configuration of the automatic mounting device (surface mounter) that mounts the IC on the printed circuit board.

This mounting device includes two TV cameras 1 and 2 ((B) and (E) in FIG. 3), an orthogonal robot 3 for mounting an IC, a substrate supply device 4, an IC supply device 5, a rough positioning device 6, The visual processing device 7, the robot control device 8, the sequencer 9, and the operation unit 10 are provided.

The sequencer 9 performs a process of mounting the IC on the printed circuit board according to a predetermined sequence program. First, the robot control device 8, the substrate supply device 4, and the rough positioning device 6 operate according to an instruction from the sequencer 9. Further, since the sequencer 9 and the orthogonal robot 3 need cooperative operation, the sequencer 9 takes in a signal from the robot controller 8.

The robot control device 8 controls the orthogonal robot 3 and the IC supply device 5.

The visual processing device 7 captures images of a plurality of contact patterns on the printed circuit board from the TV camera 1, captures images of ICs from the TV camera 2, and forms contact patterns in a row from the captured images and the center of each row of ICs. The coordinates, position and inclination can be calculated and obtained. This calculation result is sent to the robot controller 8, which directly controls the orthogonal robot 3 and the IC supply device 5, and the sequencer 9
The substrate supply device 4 is controlled through.

The operation unit 10 includes a keyboard and a CRT, and performs man-machine operations between the visual processing device 7 and the robot control device 8.

The IC supply device 5 supplies the IC from the IC storage location to a position where the robot hand (FIG. 3C) of the orthogonal robot 3 can adsorb. Rough positioning device 6 for robot hand (C)
Is provided to roughly position the IC.

FIG. 2 is an external view of the mounting device, in which the orthogonal robot 3, the substrate supply device 4, the IC supply device 5, the rough positioning device 6, the robot control device 8, the sequencer 9, and the visual processing device 7 are installed on the base 11. Indicates that it has been done. TV camera 2
The operation unit 10 is omitted.

In the present invention, the robot may be a joint type or a scalar type as long as it has sufficient positioning accuracy. As shown in FIG. 3, the robot hand (C) has an imposition component IC.
Adsorption part for holding (D) and TV camera 1 (B)
Is mounted downward, and the TV camera 2 (E) is mounted upward on the base 11.

Next, the basic operation of the mounting position will be described with reference to FIG. 3 and the processing procedure with reference to FIG.

(A) First, the substrate (A) is mechanically fixed on the substrate supply device 4, and the television camera 1 (B) fixed to the robot hand (C) is moved onto the substrate (A) in that state. Then, the television camera 1 captures an image including a plurality of contact patterns on the substrate (A).

(B) Next, the robot hand (C) moves to the IC supply device 5, sucks and grips the IC (D), and roughly positions the IC (D) by the rough positioning device 6.

(C) After that, the robot hand (C) is the TV camera 2.
Moving to (E), the television camera 2 images the IC (D) including a plurality of leads.

(D) The image information picked up by the television cameras 1 and 2 is sent to the visual processing device 7, where the position and inclination of each are calculated from the position information of the contact pattern row and the lead row of the IC, and the result is calculated by the robot. Robot 3 sent to controller 8
The IC leads are accurately mounted on the contact pattern on the board.

Next, a positioning processing method performed by the visual processing device 7 will be described.

In the following description, it is assumed that the leads of the IC or the contact pattern on the substrate does not have asymmetry or partial omission.

FIG. 5 shows an example in which leads are provided on four sides of the IC, and a plan view and a front view are shown in FIGS.

When the contact patterns and the leads of the IC are imaged by the reflected light by the television cameras 1 and 2, they are imaged as a bright portion and have a contrast difference with the background.

An image of the contact pattern 16 on the substrate corresponding to the IC 15 is shown in FIG.

As shown in the figure, corresponding to the lead lines 15-1 to 15-4 on the four sides of IC15, the pattern lines of the contact pattern 16 are 16-1 to 16-4.
In the image, elongated rectangular patterns are aligned in the direction perpendicular to the pattern longitudinal direction. Hereinafter, the elongated rectangle on the image formed by either the lead row or the contact pattern row will be generically described as a pattern.

In the actual image, in addition to this pattern, the wiring and the background are also captured.

In order to align the leads of the IC 15 with the contact pattern 16 on the board, the pattern interval is as small as several pixels, so mechanical alignment is not enough and position detection can be performed with high precision using visual means. It is necessary to do. As this visual means, the television camera 1 mounted downward on the robot hand (C) and the television camera 2 mounted upward on the base 11 are used, and the visual processing device 7 is used for position detection. .

Next, a method for detecting the position from the images by the television cameras 1 and 2 will be described below by taking the contact pattern 16 on the substrate as an example.
This will be described with reference to FIGS. 7 to 10.

The image information from the visual processing device 7 is displayed on the screen of the operation unit 10, and the seventh image is displayed on each of the pattern rows 16-1 to 16-4.
As shown in the figure, the measurement line 17
(Also called sample line) is set. The setting is made on the operation unit 10.
It may be done while looking at the CRT screen, but pattern 16-1
If the position, size, and inclination of ~ 16-4 are almost known, the measurement line 17 is set in advance so that it will always intersect with the rectangular pattern displayed on the CRT image, even if it shifts slightly for each pattern. Can be fixed.

This setting will be described in some detail. While looking at the image on the screen of the CRT, operate the keyboard of the operation unit 10 to move the cursor, and measure lines 17 vertically and horizontally at the position intersecting the pattern.
To set. If the pattern position, size, and inclination are almost known, the measurement line 17 may be set by inputting a numerical value from the keyboard.

When the setting line 17 is viewed as the horizontal axis, the brightness of the image on the setting line 17 becomes a wavy pattern corresponding to the arrangement of the contact patterns as shown in FIG. Here, when the threshold value is set to a value approximately midway between the maximum value of brightness (the portion of the pattern where the reflected light is strong) and the minimum value (the portion of the background where the reflected light is weak), comparing with the above wavy pattern , Boundary values k ₁ , k ₂ , k ₃ , k ₄ , ... ki,… on the boundary between the pattern on the measurement line 17 and the background.
Is required. Hereinafter, for convenience of explanation, an example of the pattern sequence 16-1 will be used.

FIG. 9 (I) shows an ideal case, and the pattern sequence 16-1
The reflected light waveform from is easy to detect the boundary value such as rectangular,
When the rectangle is collapsed as shown in FIG. 9 (II), a method of obtaining the center position of the pattern from the maximum points a ₁ , a ₂ , ... Ai of the wave may be adopted. However, the actual waveform is more distorted, and a method of obtaining the boundary value from this distorted waveform will be described.

FIG. 10 shows three cases (III), (IV) and (V). In either case, FIG. 10 shows the position of the pixel (xi) of the image on the horizontal axis. When the position is obtained in pixel units, the measurement accuracy is 1 pixel or more. For example, when the field of view is 25 mm and the number of pixels is 250, the minimum measurement unit is 0.1 mm.

When the lighting condition shown in FIG. 10 (III) is good and the brightness of the screen is uniform, the boundary position threshold is appropriately set and the boundary values x ₁ , x ₂ , ... xi, ... are set. Obtainable.

FIG. 10 (IV) shows that the boundary position threshold value is calculated by the interpolation calculation by taking into consideration the brightness gradient of the pixels (xi ₋₁ , xi + ₁ ) on both sides sandwiching the pixel (xi) having the brightness close to the boundary position threshold value. The position where the brightness is the same as is calculated with a real value including the decimal point. In this case, since the space between two pixels (xi ₋₁ , xi + ₁ ) is calculated after the decimal point, the accuracy is improved.

When the illumination shown in FIG. 10 (V) is bad and the brightness is not uniform, the pattern cannot be separated into a bright portion and a dark portion with one boundary position threshold value, but adjacent pixels (xi ₋₁ and x
The position where the brightness gradient between i, xi and xi ₊₁ ) is the maximum is used as the boundary value, or the weighted average method described below is used.

First, look at the wavy pattern shape displayed on the screen of the operation unit 10, and from the brightness of the pattern part (mountain) and the brightness of the background part (valley), artificially approximate the brightness of the peaks and valleys. The allowable value Dth of the difference is set, and the brightness at the position of the pixel xi is set to f (xi). Next, paying attention to whether the wavy pattern type is monotonically increasing or monotonically decreasing, the starting point is defined as a valley if the monotonous increase and the mountain is defined as a mountain if the monotonous decrease. This is shown in Figure 10 (III).
Looking at the case of ~ (V), the valley is the pixel xi _-1 , the mountain is the pixel
At the position of xi ₊₁ and the brightness f (xi _-1 ) and f (xi
₊₁ ) is examined and compared with the allowable value Dth is the equation (1).

| f (xi ₋₁ ) −f (xi ₊₁ ) |> Dth (1) Next, equation (2) for obtaining the coordinate position xk of the boundary position by the weighted average is shown.

xk is obtained for each set of valleys and peaks. When there are n pairs of valleys and peaks as in the pattern sequence 16-1, the average value U by the weighted average of the n sets is calculated from the equation (3). The obtained average value U is the center coordinate position of the pattern row 16-1.

The above calculation can be similarly performed for the other pattern rows 16-2 to 16-4.

Furthermore, the lead lines 15-1 to 15-4 on the four sides of the IC15 are also elongated rectangular patterns on the image, and the leads of the IC are bright when reflected light is picked up, the background is dark, and there is a contrast difference. Since the image is captured, the above calculation can be performed similarly.

In the above calculation, the memory (not shown) in the visual processing device 7 is positively used. First, in the memory, the image information picked up by the television cameras 1 and 2 is sent to the visual processing device 7, fetched and stored as data. next,
The brightness data on the measurement line 17 set on the CRT screen of the operation unit 10 is read out and temporarily stored in a work area (table) in the memory. This stored content is the 10th
It is the wavy data as shown in the figure.

Next, the contents of this data table are read, the peaks and valleys are calculated by the equation (1), the weighted average is calculated by the equation (2) for each pair of peaks and valleys, and the coordinate xk obtained by this weighted average is calculated. To the table. Next, this table is searched to obtain the average value U for the coordinate xk by the equation (3). This calculation is performed for each lead row 15-1 to 15-4 and pattern row.
16-1 to 16-4 will be described. The case shown in FIG. 7A will be described.

For the four pattern rows 16-1 to 16-4, the measurement line 17 is the average value of the pattern rows 16-1 and 16-4 in the horizontal direction, respectively.
Pattern rows 16-2 and 16-3 in which U1 and U2 and measurement line 17 are oriented vertically
Let V1 and V2 be the average values of, respectively, and connect U1 and U2, V1 and V2 with a straight line, and define the intersection (U, V) as the overall position. Is to be defined as the overall slope.

Next, the case of the bidirectional pattern shown in FIG. 7B will be described. In this case, only the accuracy in the direction of the measuring line 17 is important, and mechanical accuracy is sufficient for the accuracy in the direction orthogonal to this. The overall position (U, V) is U (U1 + U2) /
2, V is fixed, and the overall slope is U1 and U
Direction connecting with 2 Is defined as.

The same applies to the case of the IC lead row.

The position information obtained in this way is used by the visual processing device 7
From the robot controller 8 to the orthogonal robot 3, the IC as an imposition component can be correctly mounted on the contact pattern on the substrate.

In the explanations given so far, it was assumed that the IC leads or contact patterns on the substrate were symmetric or missing, but the actual leads or contact patterns are
As seen in FIGS. 15, 15 and 16 (a), it is often intentionally asymmetrical or partially lacked. In such a lead train or pattern train, the center position cannot be calculated correctly even if the weighted average method described so far is introduced.

An object of the present invention is to correctly detect the center position of a lead row or a pattern row that is asymmetrical or partially missing so that the IC can be correctly mounted on the contact pattern. This will be described below.

FIG. 11 shows the case where the pattern widths at both ends of the contact pattern row on the substrate are different. An example of this case will be described with reference to FIG. The brightness on the measurement line 17 shown in FIG. 12 (a) is as shown in FIG. 12 (b). In this case, the visual processing device 7 is provided with a data table for storing the brightness data of the pixel xi,
As shown in FIG. 12C, the detected values of both outer end patterns of the contact pattern row are ignored, and 0 is put in the data table where x ₁ and x ₁₀ should be stored. If the first point of each pattern is the falling edge, as shown in Fig. 13 (a) and (b),
As shown in Fig. 13 (c), the first point is regarded as within the pattern, x ₁ is put on the second data table, and the data table is created in the same manner as above.

Further, in the actual pattern, as shown in FIG. 14 (a), the brightness may change within one pattern width.
In the case of a rising edge, the last rising edge detection point x ₂ and in the case of a falling edge the last falling edge detection point x ₅ are regarded as the boundary values (edge points) of the pattern, and x ₁ and x ₆ are the boundaries (edges). ) Is not considered. This is shown by x ₂ to x ₅ in Fig. 14 (b).

As described above, the value of the rising detection point is stored in the odd-numbered table, the value of the falling detection point is stored in the even-numbered table, and the state of the pattern sequence is correctly registered. It becomes easy to correspond to a similar pattern row or lead row.

In the present embodiment, the case where the pattern row or the lead row is brighter than the background has been described. Conversely, when the pattern row or the lead row is darker than the background, the definition of storage in the data table may be reversed.

Furthermore, the case where the arrangement of the IC leads is asymmetrical as shown in FIG. 15 will be described with reference to FIG. The brightness waveform and data table on the measurement line 17 shown in FIG. 16 (a) are shown in FIG. 16 (b) and (c). Since this is regarded as the lead column A portion is missing in Figure 16 (b), the detection value x ₃ Part B, x ₄
Is ignored, and that portion of the data table is re-created as 0 as shown in FIG. The above is the case of the lead row, but the center of the pattern row can be obtained in the same manner as described above. In this way, the brightness data stored in the data table is sent to the visual processing device 7 and used for controlling the orthogonal robot equipped with an IC.

In the present invention, since the omission correction is often required in this way, in order to speed up the processing at the time of measurement, the correction amount Δp shown in FIG. The temporary center point p ′ calculated from the data table of FIG. 16C,
The center point p of the column is calculated by performing correction with the registered correction amount Δp. The correction amount Δp may be obtained by measuring from an actual image or by calculating when the lead row or pattern row is known.

A case of finding the center point of a lead row or a pattern row with a gap will be described with reference to FIG. As I mentioned before,
In the case of 4 directions (a), the intersection of the straight lines connecting Pi and P ₃ and P ₂ and P ₄ is the center of the whole and the direction of the bisector of two straight lines Is defined as the overall inclination, and in the case of two directions (b), P ₁ and
The middle point of the straight line connecting P ₃ is the center of the whole, and the inclination of the whole is
Direction connecting P ₁ and P ₃ Define as.

The processing flow at the time of teaching in the visual processing device 7 of the present embodiment described above is shown in FIG. 19, and the processing flow at the time of measurement is shown in FIG.
Shown in Figure 20.

〔The invention's effect〕

According to the present invention, it is not necessary to put a special mark on the printed circuit board, and the center position and the inclination can be detected even for the IC of the lead or the contact pattern of the board having an asymmetric or partially missing array. This has the effect of enabling position detection with high accuracy without depending on the shape of both ends of the lead or contact pattern row. Furthermore, since the boundary can be obtained even when the brightness changes within one lead or within the pattern, stable position detection can be performed, and the automatic mounting machine for imposition components with high reliability and adaptability. Can be realized.

[Brief description of drawings]

FIG. 1 is an embodiment diagram of the present invention, FIG. 2 is an external view of a mounting machine,
FIG. 3 is a diagram showing a processing procedure, FIG. 4 is a processing flow chart, FIG. 5 is a top view of a printed circuit board, FIG. 6 is a top view of an IC, and FIG. 7 is inclination calculation by a measurement line in a contact pattern. FIGS. 8 to 10 are various contact pattern example diagrams, FIG. 11 is an actual contact pattern example diagram, and FIGS. 12 and 13 show the relationship between the edge coordinates of the contact pattern row and the brightness. Fig. 14, Fig. 14 is a diagram showing continuous edge points in the concentration gradient in the same direction, Fig. 15 is a three-sided view of an example of a flat package IC that requires missing correction, and Fig. 16 shows missing correction. FIG. 17 is a view showing the correction amount Δp in the missing correction measuring method of the present embodiment, and FIG. 18 is a position of the entire contact pattern having contact pattern examples in four and two directions, FIG. 19 is a diagram showing a posture measuring method, and FIG. 19 is a flowchart showing a processing procedure at the time of teaching in one embodiment of the present invention. Chart diagram, FIG. 20 is a flowchart showing a measurement time of processing procedure in an embodiment of the present invention. 1,2 …… TV camera, 7 …… Visual processing device.

 ─────────────────────────────────────────────────── ─── Continued Front Page (72) Inventor Shigechi Takezen 7-1, 1-1 Higashi Narashino, Narashino City, Chiba Prefecture Hitachi Keiyo Engineering Co., Ltd. (72) Inventor Kenji Suzuki 292 Yoshida-cho, Totsuka-ku, Yokohama-shi, Kanagawa (56) References JP-A-60-1900 (JP, A) JP-A-62-86789 (JP, A) JP-A-55-48945 (JP, A) JP-A-SHO 57-154851 (JP, A)

Claims (2)

[Claims] 1. An image of a plurality of leads provided in a row on two or four sides of an IC and a plurality of contact patterns provided on a printed circuit board in accordance with the leads are taken by a television camera. An image of each column of the lead or contact pattern is projected on a screen within a preset rectangular range of the image processing device, and a measurement line is set on each column of the lead or contact pattern within the rectangular range, The brightness data of the lead or contact pattern on the measurement line is stored in a memory that stores the array order, and the boundary position between the edge of the lead or contact pattern and the background is specified from the brightness data in the memory. From the border position on the screen
The center coordinates of the lead rows on the 2 or 4 sides of the IC are calculated, and then the center and tilt of the IC are calculated. Also, the center coordinates of the contact pattern are calculated for each pattern row on the screen from the boundary position. In addition, the center coordinates and the inclination of the contact pattern are calculated, and the center and the inclination of the contact pattern of the printed circuit board are aligned with the center and the inclination of the IC to form an apparatus for mounting the IC on the printed circuit board. And means for calculating the center coordinates of each row of the lead or each row of the contact pattern in consideration of the asymmetry or partial omission, even when the plurality of leads or contact patterns provided with asymmetry or omission are present. , From the center coordinates of each row of the lead or each row of the contact pattern calculated by the means, the center and inclination of the IC with asymmetry or missing or Characterized in that a means occupied subjected to alignment of IC so that the center coordinates and the inclination to the IC center and inclination of the contact pattern of the printed circuit board to calculate the center coordinates and inclination of the serial contact pattern matches
Device for mounting the IC on the printed circuit board. 2. The center coordinates of each row of the lead row or contact pattern row previously imaged with respect to the plurality of lead or contact pattern rows forming the row having the asymmetry or a part lacking are asymmetric or partially. The correction data to be calculated in consideration of the loss is stored in the storage memory, and for the substrate having the IC or the contact pattern having the same asymmetry or a partly missing lead, the correction data calculated using the correction data is stored. The center and inclination of the IC or the center coordinate and inclination of the contact pattern are calculated from the center coordinates of the lead row or each row of the contact pattern row to calculate the center coordinate and inclination of the contact pattern of the printed circuit board and the center of the IC. The device for mounting an IC on a printed circuit board according to claim 1, further comprising means for aligning the IC so that the inclinations match each other.