JP2633147B2 - Component mounting method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a component mounting apparatus for accurately positioning components with respect to a predetermined mounting position on a substrate by a recognition means in mounting components on a printed circuit board, and for determining whether the positioning is good or bad.

[0002]

2. Description of the Related Art Generally, when a component is automatically mounted on a board, the component is picked up by a component transfer head from a feeder for supplying the component, and the component is mounted on a board at a predetermined component mounting position. Carry to implement.

In particular, when a surface mounting component or the like having a very large number of leads is used as a component, the pitch between adjacent leads is small, and even if the position is slightly shifted, erroneous wiring occurs.

In order to prevent erroneous wiring, the position of the component and the position of the board on which the component is mounted are detected, and the positions of the components are aligned to accurately mount the component at a predetermined position on the board. Is required.

As a method for mounting such a component, as disclosed in Japanese Patent Application Laid-Open No. 60-1900, the center position and the rotation angle of the component are determined from the image of the component lead, and the mounting pattern and the mounting pattern are determined from the image of the board. Two I formed simultaneously
It is proposed to determine the center position and the rotation angle of the mounting pattern by the C mark, and to move and rotate the component transfer head to mount the component at a predetermined position on the board based on the obtained positional relationship between the component and the mounting pattern. (Hereinafter, Conventional Example 1)
).

Further, the positional relationship between the component and the mounting pattern is accurately and precisely defined, and even if the component can be mounted on the mounting pattern with high accuracy, even if the component and the mounting land are temporarily attached to each other, such as an adhesive or cream solder. There is a possibility that misalignment may occur immediately after mounting due to poor printing, etc., and it is conceivable to detect these misalignments and prevent the board with the poor mounting from flowing in the next process (reflow furnace, etc.). I have.

A method for determining the quality of component mounting is disclosed, for example, in Japanese Patent Application Laid-Open No. 63-90707. This method is based on the positional relationship between land images before and after mounting. JP-A-1-309190 describes that
There is disclosed a technique for making a determination based on the amount of overlap between land images before and after mounting (hereinafter referred to as Conventional Example 2).

Conventionally, components are mounted on a board by a component mounter, and the quality of the mounted components is determined by a mounted component inspection machine in the next process after component mounting of the entire board is completed. And a visual inspection (hereinafter referred to as Conventional Example 3).

[0009]

However, in the conventional example 1, when the mounting position of the component is corrected based on the obtained positional relationship between the component and the mounting pattern, the positional relationship between the mounting pattern and the IC mark and the component are corrected. If the positional relationship of the imaging means between the substrate and the substrate is not accurately and precisely defined, the component cannot be mounted on the mounting pattern with high accuracy.

[0010] Furthermore, a fine chip component or a component with a lead having a narrow lead pitch that requires mounting accuracy requires two IC marks for each component, so that high-density mounting cannot be performed.

[0011] Further, the conventional example 2 is a method for realizing an inspection device after mounting, and the reference for comparison is a taught one or a reference substrate, so that a substrate actually mounted is used. Absent. For this reason, it is necessary that the variation of the mounted substrate itself has a fixed value as an allowable range.

Further, in Conventional Example 3, when visually judging the quality of a mounted component, sufficient inspection accuracy cannot be expected for a fine chip component or a component with a lead having a narrow lead pitch. However, uniformity cannot be expected due to individual differences and fatigue.

Therefore, the mounted component inspecting machine is installed in the next process of the component mounting machine. However, both require input of data such as a component mounting position and a component inspecting position, and adjustment of the input data. The number of times when data is created increases, the burden on the operator is increased, and it is difficult to handle various types of mounting boards.

The present invention has been made to solve such a problem, and an object of the present invention is to specify a positional relationship between a component and a mounting land so that the component can be mounted with high accuracy. By defining at least one of the mounting lands and the recognizing means to be orthogonal to each other, and by generating the reference position so as to be projected from any camera, it is possible to easily perform the image processing operation for extracting the correction amount and the image synthesis. It is to provide a component mounting method that can be performed.

[0015]

In order to achieve the above object, according to the present invention, a rectangular mounting component is taken out from a component supply device and the mounted component is transferred to a predetermined component mounting position on a substrate. In a component mounting method for mounting on a mounting land, a component transfer head for transferring the mounted component is provided.
With at least two cameras that are symmetrical to each other
The configured recognition means recognizes the mounting component and the mounting land from a direction substantially perpendicular to at least one edge of the mounting component or the mounting land, and determines corresponding feature points of the mounting component and the mounting land. The positional relationship is defined based on the component transfer head.

[0016]

According to the present invention, in the present invention, a rectangular mounted component is taken out from the component supply device, and the mounted component is transferred to a predetermined component mounting position on a substrate and mounted on a mounting land. At the time of mounting, the mounting component or the mounting land is recognized from a direction substantially perpendicular to at least one of the edges of the mounting component or the mounting land by a recognition unit configured by at least two cameras symmetrical to each other. By recognizing, the relationship between the captured image and the pixels of each camera becomes a parallel or right-angled relationship, thereby simplifying image processing and image synthesis. In addition, by including this reference position in the field of view of the camera, a combining operation for combining a plurality of images can be easily performed. For example, the recognition means is, for example, arranged such that four cameras are displaced by about 90 ° around the component transfer head, or two cameras are displaced by about 180 °. Things.

[0017]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will be described below with reference to the drawings. In FIG. 1, the component transfer head 1 and the cameras 2-1, 2-2, 2-3, and 2-4 are mounted on the same base and can be transferred in the XY directions. The component transfer head 1 has a mechanism capable of transferring components in the vertical direction and the rotation direction. Four (or two) cameras are arranged around the component transfer head 1, and the imaging (or recognition) direction of each camera is orthogonal to the component, and the transfer head 1 is oblique to the component. Has a field of view in the movable direction.

As described above, by synthesizing the fields of view using a plurality of cameras, the field of view resolution of the cameras can be enhanced, the portion that becomes a shadow of the component transfer head 1 can be recognized, and the height information of the object to be recognized can be obtained. It is easy to obtain, and the reflection of the component transfer head 1 into the field of view can be suppressed (hereinafter, the unit of the component transfer head 1 and the recognition camera is referred to as a “transfer recognition unit”).

Since the transfer recognition unit moves in the X and Y directions but the camera does not move up and down, the camera is focused on a certain plane with respect to the movement of the transfer recognition unit (see FIG. 1). The focus plane 8 is shown). Therefore, the components and the pattern of the board 5 to be mounted must be on this surface. However, conversely, if the background or the like that is not desired to be imaged is not on this surface, no image is formed because the camera is out of focus, and the influence of the background can be suppressed.

The steps up to the recognition and mounting of components using such a configuration will be described below. FIGS. 2A and 2B show a state at the time of component recognition. The mounted component 3 is sucked from the feeder 4 by the component transfer head 1, lifted up to the focal plane 8 of the camera, captured by the cameras 2-1, 2-2, 2-3, 2-4 and stored in the frame memory. The composite image obtained at this time is shown in FIG.

Thereafter, the component transfer head 1 rises to a height at which it can pass and moves to the board recognition point 7 in FIG. In balance with the operation, the component position recognition processing is performed with the component transfer head 1 as a reference.

Next, the state at the board recognition point 7 is shown in FIG.
(A) and (b). Since the component transfer head 1 is located above, the focus of the cameras 2-1, 2-2, 2-3, and 2-4 matches the mounting land 9 on the board surface. FIG. 3C shows a composite image obtained by imaging in this state. Since this image is also obtained by the same camera as in FIG. 2C, the position of the mounting land can be recognized based on the component transfer head 1.

The component position information obtained by FIG. 2C and the substrate position information of the mounting land 9 obtained by FIG. 3C are recognized based on the component transfer head 1. However, the absolute accuracy of the entire device is not required, and the correction is performed so that the parts and the land are matched, so that high-precision mounting becomes possible.

That is, since the same camera can recognize components and lands, and the head, which is the center of transfer, can also be recognized by the same camera, error components can be reduced and high-precision mounting becomes possible. FIG. 4 shows a component mounting state).

As shown in FIGS. 7A and 7B, the head is lowered to a focus plane (hereinafter, referred to as a recognition plane) 8 of the camera, and an image is taken by each camera.
As shown in (f), the tip of the head imaged by each camera is stored as a reference position for each camera. This is because calculating the reference position every time of recognition leads to an increase in tact. By including the reference position in the field of view of each camera, the images of the cameras can be combined using the reference position, and the combining operation can be easily performed.

Next, inspection after mounting is performed. Even if the above-described method is used to perform high-precision alignment, the state of the cream solder at the time of mounting, the cutoff of the air pressure of the transfer head, and the vibration at the time of mounting are also considered. Parts may move or parts may stand up. Therefore, inspection is performed after mounting.

FIGS. 5A and 5B show a state where the component transfer head 1 is lifted up after mounting, and FIG. 5C shows a composite image obtained at this time. Since the composite image is stored in the frame memory, the transfer recognition unit can go to the next mounting component.

FIG. 6 shows an image of image extraction. That is, an image of the mounting land 9 before mounting (FIG. 3
An image comparison process is performed between (c)) and the mounted image (FIG. 5 (c)) to obtain a mounted component image. From this image,
The position of the component after mounting can be recognized based on the component transfer head 1. By comparing the position information with the position information of the component before mounting obtained from FIG. 2C with reference to the component transfer head 1, the mounting state of the component after mounting can be recognized. .

Thus, the apparatus can determine whether or not the mounted components have been mounted normally, and can also teach an operator of an abnormality, and can return the positional deviation information to data for the next mounting. .

In this embodiment, as a mounting apparatus having a recognition function, this function is used not only for alignment but also as a function for checking the mounting state after mounting. In addition, since recognition is performed immediately after mounting, there is almost no time loss due to inspection, and an inspection device is not required in a post-process as in the related art. For this reason, teaching at a position necessary for the inspection machine is unnecessary, the number of steps before curing is reduced, and there is no possibility that parts are inadvertently displaced. That is, the size of the entire line can be reduced, the teaching time and the type switching time can be shortened, and the number of defective elements can be reduced.

In this way, high-precision mounting is performed by recognizing components and mounting lands with the same camera, and inspection is performed after mounting, so that it is not necessary to introduce a mounting inspection machine. The above ability can be exhibited.

Next, a series of flows from mounting to inspection will be described with reference to FIGS. In FIG. 8, step 1
At 01, rough teaching of the mounting position and input of the accompanying mounting data are performed, and then the process proceeds to a mounting data optimizing process. In this process, in step 102, movement to the component supply position and removal of the component based on the component supply position data are performed. In step 103, component image information is captured, and component position information is calculated based on the transfer head. I do.

At this time, FIG. 11 (c) shows an image taken by the camera 2-1. However, as shown in FIGS. 11 (a) and 11 (b), since each camera is orthogonal to the rectangular mounting parts, The relationship between the captured image and the pixels of each camera is parallel and perpendicular as shown in FIG. 11D, which simplifies image processing (extraction calculation of feature position) and image synthesis. The same applies to the mounting position corresponding to the component.

In step 104, movement to the component mounting position based on the teaching data is performed. Step 105
Then, the board image information is captured, the mounting land is searched and extracted for the mounted component based on the extraction result of the component information, and the mounting land position information is calculated based on the transfer head.

In step 106, a mounting position that minimizes the difference in the positional relationship between the corresponding feature points of the component position and the mounting land position with respect to the transfer head is calculated. Step 107
Then, the mounting position data is updated, and in step 108, movement to the optimum mounting position and component mounting are performed.

In step 109, it is determined whether or not the component requires mounting accuracy. If the determination is YES, the process proceeds to the actual operation process of step 110 (FIG. 9). If the determination is NO, step 119 is performed.
To the actual operation step (FIG. 10).

In this way, in step 110 of FIG. 9, movement to the component supply position and removal of the component based on the component supply position data are performed, and in step 111, component image information is captured and the transfer head is used as a reference. The component position information is calculated. The information on the component position calculated here is used for updating the data in step 117, and the updated data is used in step 110.
To be entered.

In step 112, the component is moved to the component mounting position based on the component mounting data. In step 113, the board image information is captured, and the search and extraction of the mounting land for the mounted component based on the extraction result of the component information are performed.
The mounting land position information is calculated based on the transfer head.

The mounting land position information calculated here is
It is used for updating data in step 118, and the updated data is input to step 112 described above. Subsequently, in step 114,
A mounting position that minimizes the difference in the positional relationship between the corresponding feature points of the component position and the mounting land position with respect to the transfer head is calculated. In step 115, movement to the optimum mounting position and component mounting are performed. The image information is taken to extract the positional information of the mounted component, the component mounting information is calculated based on the transfer head, and the quality of the mounting of the component based on the transfer head is determined.

Next, when mounting accuracy is not required, FIG.
In step 119, the component supply position is moved and the component is taken out based on the component supply position data, and in step 120, the component image information is captured and the component position information is calculated based on the transfer head. The component position information calculated here is used for updating the data in step 126, and the updated data is input to step 119 described above.

Subsequently, in step 121, the component mounting position data is corrected based on the component position information based on the transfer head, and movement to the component mounting position corrected in steps 122 to 124 and taking of board image information are performed. In step 125, the board image information is captured to extract the position information of the mounted components, calculate the component mounting position information with the transfer head as a reference, and set the transfer head as a reference. Whether the mounting of the component is good or not is determined by comparing the obtained information.

The board image information in step 123 and the component mounting position information calculated in step 125 are used for updating the data in step 127, and the updated data is input to the previous step 121. It has become.

When mounting leads extending in four directions as shown in FIGS. 1 to 6 at corresponding mounting positions,
Since the precision for defining the X and Y directions is required, it is necessary to recognize the leads from four directions and the mounting positions corresponding thereto, as shown in FIGS. 12 to 17 (corresponding to FIGS. 1 to 6). When recognizing a mounting position corresponding to a lead extending in two directions as described above, only the accuracy that defines the Y direction is required.
The X direction does not require high precision. Therefore, in this case, as shown in FIGS. 13 (a) to 13 (c) and FIGS. 14 (a) to 14 (c),
The correction amount is calculated by recognizing only two cameras whose two leads or lands corresponding to the symmetrical positions look linear.

FIGS. 18A to 18F show details of the component recognition method. In FIGS. 18A and 18B, each camera 2-
Since the field of view is different for each of 1,2-2,2-3,2-4,
These captured images are also different, and these FIGS.
A recognized image is obtained by synthesizing the image of (f).

FIGS. 19A to 19F show details of the board recognition method. In FIGS. 19A and 19B, each camera 2-
Since the field of view is different for each of 1,2-2,2-3,2-4,
These captured images are also different, and these FIGS.
A recognized image is obtained by synthesizing the image of (f).

FIGS. 20 to 25 show a mounting method of a mounting component having no lead legs, which correspond to FIGS. 1 to 6 already described.

FIG. 26 is a flow chart of switching a recognition camera without using a camera when mounting a component having leads extending in two directions. That is, the direction of the mounting land is stored, the land direction is compared with the terminal direction of the component, and if they match, the component is recognized by the camera in the same direction as the terminal, and if they do not match, the component is perpendicular to the direction of the terminal. Recognize parts with the camera in the direction.

FIG. 27A shows switching using a camera, which is determined based on the presence or absence of a lead. That is, the component is imaged by four cameras, image processing is performed, a camera in the direction in which the lead exists is selected, and the component is recognized by the selected camera. FIG. 27B shows the switching using a camera. If the external dimensions of the component are longer or shorter, the determination is made based on the direction in which the longer side exists. For example, a component is imaged by four cameras, image processing is performed, a camera in a direction in which a long side exists is selected, and the component is recognized by the selected camera.

[0049]

According to the present invention, the following effects can be obtained. (1) Since the same camera can recognize the mounted component and the mounting land with reference to the component transfer head, the accuracy of the entire device is not required, and the correction is performed by matching the mounted component and the mounting land. Positioning is possible. In addition, the mounting state of the component can be inspected by comparing the positions of the image after mounting and the image obtained for mounting, and an inspection machine is not required in a post-process. Therefore, there are no elements such as teaching of the inspection machine, switching of types, occurrence of defects by the inspection machine, and the like, and a failure determination line can be connected immediately after the mounting apparatus. Further, the mounting deviation data can be accumulated and returned to subsequent mounting data, so that mounting with higher accuracy is possible. (2) Further, in the present invention, since the mounted component and the mounted land are recognized from a direction substantially perpendicular to at least one edge of the mounted component or the mounted land, the captured image is recognized with respect to the image. Parallel, right angle,
Image processing operations and image synthesis are simplified. (3) Further, according to the present invention, the corresponding feature points of the mounted component and the mounting land are defined in a positional relationship with respect to the component transfer head. For example, for each camera, the tip of the component transfer head is set as the reference position. By storing this reference position in the field of view of the camera, an increase in recognition calculation can be prevented, and calculation for combining a plurality of images can be easily performed.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating a recognition mounting operation.

FIG. 2A is a front view showing a state at the time of component recognition;
(B) is a plan view, and (c) is a diagram showing a composite image of a component.

3A is a front view showing a state at the time of recognition at a board recognition point, FIG. 3B is a plan view thereof, and FIG. 3C is a view showing a composite image of the board.

FIG. 4 is a front view showing a component mounting state.

FIG. 5A is a front view showing an inspection state after component mounting;
(B) is a plan view thereof, and (c) is a diagram showing a composite image in a component mounting state.

FIG. 6 is a diagram showing an image of image extraction.

7A is a front view when a reference position is generated, FIG. 7B is a plan view thereof, and FIGS. 7C to 7F are views showing images captured by a camera.

FIG. 8 is a diagram showing a flowchart from component mounting to inspection.

FIG. 9 is a diagram showing a flowchart from component mounting to inspection.

FIG. 10 is a diagram showing a flowchart from component mounting to inspection.

11A is a front view showing a positional relationship between a mounted component and a camera, FIG. 11B is a plan view thereof, FIG. 11C is a view showing a captured image of the camera, and FIG. FIG. 3 is a diagram illustrating a relationship between pixels of each camera.

FIG. 12 is a diagram showing a recognition mounting operation of a component with two-way leads.

FIG. 13A is a front view showing a state at the time of component recognition,
(B) is a plan view, and (c) is a diagram showing a composite image of a component.

14A is a front view showing a state at the time of recognition at a board recognition point, FIG. 14B is a plan view thereof, and FIG. 14C is a view showing a composite image of the board.

FIG. 15 is a front view showing a component mounting state.

16A is a front view showing an inspection state after component mounting, FIG. 16B is a plan view thereof, and FIG. 16C is a view showing a composite image of the component mounting state.

FIG. 17 is a diagram showing an image of image extraction.

18A is a front view at the time of component recognition, FIG. 18B is a plan view thereof, and FIGS. 18C to 18F are views showing images taken by a camera.

19A is a front view at the time of board recognition, FIG. 19B is a plan view thereof, and FIGS. 19C to 19F are views showing images taken by a camera.

FIG. 20 is a diagram illustrating a recognition mounting operation of a mounted component having no lead leg.

FIG. 21A is a front view showing a state at the time of component recognition;
(B) is a plan view, and (c) is a diagram showing a composite image of a component.

22A is a front view showing a state at the time of recognition at a board recognition point, FIG. 22B is a plan view thereof, and FIG. 22C is a view showing a composite image of the board.

FIG. 23 is a front view showing a component mounting state.

24A is a front view showing an inspection state after component mounting, FIG. 24B is a plan view thereof, and FIG. 24C is a diagram showing a composite image of the component mounting state.

FIG. 25 is a diagram showing an image of image extraction.

FIG. 26 is a diagram showing a flowchart of switching recognition cameras without using a camera.

FIG. 27 is a diagram showing a flowchart of switching a recognition camera using a camera.

[Description of Signs] 1 Component transfer head 2-1, 2-2, 2-3, 2-4 Camera 3 Component 5 Board 6 Component recognition point 7 Board recognition point 8 Camera focusing surface 9 Land

Claims (1)

(57) [Claims] 1. A component mounting method for taking out a rectangular mounted component from a component supply device, transferring the mounted component to a predetermined component mounting position on a substrate, and mounting the mounted component on a mounting land. each other provided on the component transfer head
Composed of at least two cameras symmetrically positioned
The recognition means performs the recognition of the mounting component and the mounting land from a direction substantially perpendicular to at least one edge of the mounting component or the mounting land, and identifies a corresponding feature point of the mounting component and the mounting land. A component mounting method comprising defining a positional relationship based on a transfer head.