JPS613417A - Method for bonding - Google Patents

Abstract

PURPOSE:To effect bonding to the plural bonding positions having differences in height automatically with accuracy at a high speed by focusing a TV camera so as to make a quantity of discrepancy of focus zero which is calculated by collation of an image pick-up output signal with a reference signal. CONSTITUTION:A pattern recognition circuit 23 collates a binary signal pattern of an image of a substance to be bonded 16 picked up by a TV camera 11 with a reference pattern which is stored previously and calculates a quantity of discrepancy in position. From this quantity, a real position is detected and bonding is effected automatically. In this case, the control process is included which includes the program that a quantity of position discrepancy is calculated by collating an image pick-up output signal of the predetermined position of substrance 16 with the previously stored reference signal at a time of when an image of substance to be bonded 16 is picked up and positioning of the TV camera in Z-axis direction and focusing of an optical lens of the TV camera 11 are carried out so that the above quantity becomes zero.

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to a bonding method that can accurately detect bonding positions even if they have heights.

[Technical Background of the Invention and Problems Therewith] Bonding devices include, for example, die bonders that attach semiconductor chips to printed circuit boards, wire bonders that connect semiconductor chips and leads of die-bonded printed circuits, and the like. Although many ideas have been proposed regarding these bonders, the following are the ones that have been put into practical use as fully automatic die bonders and wire bonders. In a die bonder, the heights of electrode pads to which semiconductor chips are attached are the same. Therefore, for automation purposes, the mounting position of semiconductor chips is recognized using an ITV camera, and the ITV camera was able to process a large number of printed circuit boards at high speed and accurately in a fixed state. Regarding the wireponder, the ITV camera, which moves together with the capillary, alternately moves along the heights and lows of the semiconductor chip to automatically wire the wires, but there are only a few As technology progresses, the size of a single printed circuit becomes larger, and as a result of the development of various types of circuit configurations and three-dimensional laminated structures, the heights of parts of a single printed circuit become uneven. There is a demand for technology that can automatically bond such large-scale printed circuit boards.To meet this demand, conventional automatic bonding equipment can automatically bond a position recognition processing system to one mounting position. When I executed the bonding, I found that the bonding was not accurate. A closer look at the cause of this revealed that accurate bonding was performed on the bonding pads at the same height as this setting position.

However, it was found that bonding errors were relatively common at different height positions.

[Purpose of the invention]

The present invention has been made in view of the above points, and it is an object of the present invention to provide a bonding method that can automatically perform bonding to objects having heights at a plurality of bonding positions accurately and at high speed.

[Summary of the invention]

The amount of positional deviation is calculated by comparing the binarized signal pattern obtained by imaging the object to be bonded with a television camera and a pre-stored standard pattern. The amount of defocus is calculated by comparing the output signal with a standard signal stored in advance, and the Z-axis direction position of the television camera or the focusing of the optical lens of the television camera is executed so that the amount of defocus becomes zero. The present invention provides a bonding method comprising a control procedure including a program.

[Embodiments of the invention]

Next, an embodiment in which the method of the present invention is applied to die bonding of semiconductor chips will be explained with reference to FIGS. 1 and 2. FIG. 1 shows a component that automatically recognizes the bonding position. FIG. 2 is a diagram showing a die bonding apparatus having the functions shown in FIG. 1.

That is, a motor such as a near motor (1) K is provided with an X-axis table (2) that moves in the X-axis direction, and a motor such as a linear motor (3) K is provided with a Y-axis table (4) that moves in the Y-axis direction. A motor such as a DC servo motor or a pulse motor (
5), upper and lower blocks (6) that move in the Z-axis (vertical direction) direction are provided. This upper and lower block (6) uses a belt (7) to rotate the motor (5) using a ball screw (8).
drive to rotate. Therefore, the diameter of the part that engages with the belt (7) on the motor (5) side and the block (6)
By selecting the diameter ratio of the side belt (the part that engages with the force), you can control high speed and accuracy.

A mount head a1 that supports a mount tool (9) is fixed to the upper and lower blocks (6). The upper and lower blocks (
6) is fixed with a television camera, for example, an ITV camera αυ. The mounting tool (9) is a semiconductor chip (
A replacement tool mounting table a3 is provided within the moving range of the mount head 01, and a tool storage hole α4 is drilled in the table a3 so that 120 kinds of tools can be selected in advance according to size, weight, etc. Another tool 9 is removably housed in this hole (14). An optical lens αη is attached to the camera (Ill) to take an image of the object to be bonded, for example, a printed circuit board.The circuit board αe
In FIG. 1, the difference in height of the bonding position is exaggerated. An example is shown in which a mark (19) is provided for recognizing the height of each Kt'i.
) by moving the focusing upper and lower blocks (6) in the vertical direction. This block (6) is configured to slide along the slide (a).

The above engineer T■ Camera aDi11. ITv drive circuit (21)
It is driven by Ic, and the output of camera α1) is output from the monitor (A).
K is displayed and also supplied to the pattern recognition circuit (c). This circuit (c) obtains a binary signal from the output signal of the ITV camera a1), and compares and verifies this signal with a standard pattern. On the other hand, the motor (5)K is provided with an encoder, and the amount of movement in the Z-axis direction by the motor (5)K is controlled by controlling the rotation angle using the output of the encoder (c). This rotation angle is determined according to the number of pulses. This is done by checking the difference from the target position (height) stored in advance in the waveform shaping/synchronization, circuit 5~ circuit (c), Z-axis control circuit (to), and memory (to) by the arithmetic processing circuit. The deviation position is calculated from the error counter. This deviation position signal is converted into an analog quantity by the conversion circuit (30),
A command value for the biaxial movement motor (5) is output, and based on this, the KZ-axis motor drive circuit 01) controls the current of the motor (5). This is digital control in the Z-axis direction.

A pattern at the target position, for example, a mark on a board, is imaged, and the output of the camera αJ is passed through the optical system αD, and is subjected to pattern matching processing in the pattern recognition circuit (c) via the ITV drive circuit 0υ, and the pass/fail of this is determined. The arithmetic processing circuit (5) compares and determines the data with the data in the memory (c), and when the camera is in focus, the position of the camera αυ is detected. The XY horizontal movement adjustment is performed when the camera (111) is in focus. In other words, the X and Y components of the arithmetic processing circuit (5) output are controlled via the control circuit 03 and the drive circuit (to). Grasp each motor (1) chip α and move it up and down using the Z-axis/digital control described above.
Mount via Y table. Tool conversion and ITV
Move camera I up and down and look at tool storage hole (I4) and tool a9.
Find the tool (9) and replace aω.

The arithmetic processing circuit (5) is connected to a Chessman e-reader and is used for teaching which manually sets data in the memory (to) in advance. For focusing, the focal length of each position on the circuit board αe is stored in advance in the memory @, the camera C11l is automatically set, and the vertical adjustment of that position is performed using the mark αg mentioned above.
1 (1 (J) may be used for focusing.Furthermore, a line sensor may be used to adjust the focus by the number of sensors projected by a predetermined electrode pad, or ultrasonic waves may be generated. Then, the camera C is adjusted to the focal length using the Dotsupura effect.
11l may be set.

In FIG. 1, the ball screw (8), motor (5), and encoder may be installed on the same axis. Further I
The main body of the TV camera αυ is fixed on the envelope C33 of a two-axis direction drive device provided on the Y-axis table (4), the optical image is guided by an optical fiber, and only the optical system is connected to the camera circle shown in Fig. 1. It may be installed in the position of Furthermore, the teaching of focusing of the ITV camera circle is carried out by manually controlling the Chessman 0 Enoki so that the imaging output of the mark αF3dl provided on the circuit board αυ becomes a clear image on the monitor (c). Can be executed. -Next, a series of steps of the diponda will be explained with reference to FIG. A thick film circuit board (for example, a ceramic circuit board with a width of 3α and a length of 5cIr) in which a wiring pattern is formed on a ceramic substrate (
(not shown) is placed on the carrier α and the magazine (6
) are loaded in multiple stages. When this magazine (6) is installed, the magazine (6) is removed one by one from the bottom.
6) is controlled by the downward moving drive of the loader (43).

That is, the belts located at the lower stage of the magazine (6) and arranged at regular intervals are driven into a conveying state, and the carrier (41) is driven by the downward movement of the magazine (4) by the loader (43). When the magazine (6) is placed on the belt (44), the downward movement of the magazine (6) is stopped and the carrier (44) is placed on the belt (44).
m)t! It is conveyed linearly to the dispensing position below the dispense head (c) and stops. In this position the circuit board resting on the carrier (6) is placed on the dispensing head (
After focusing a television camera mounted on a, for example, a black and white television camera 0e, a semiconductor chip (
A standard pattern for automatically attaching phrases is manually stored in advance in a memory (not shown). Therefore,
-! : Semiconductor chips for circuit boards arriving each time (4?)
The mounting position of the semiconductor chip is imaged with the above-mentioned TV camera), and the image output is compared with the binary signal and the above standard pattern to calculate the positional deviation amount, and from this positional deviation amount, the actual semiconductor chip mounting The position is recognized and the semiconductor chip Q is placed in this area.
7) Attach a material for fixing, such as silver paste having adhesive properties, for example, by dripping. To carry out these steps, the dispense head (c) is placed on an X-Y axis drive table (not shown). This completes the dispensing process. Thereafter, the carrier (9) is conveyed by the bell 144 and stopped at the mounting position.

At the mounting position, the mounting position of the semiconductor chip αη is automatically detected by the imaging output after focusing the television camera (41) mounted on the mounting head (c). Similar to position detection, this is done automatically using pattern recognition technology. When this automatic position detection is completed, the semiconductor whose position has been detected is picked up from the mount head f41 K and θ table trowel, and lifted up using suction means, for example. The semiconductor chip (4η) is placed on the adsorption position of the chip 0η, and the above head (411K) is continued to be pressed with a pressing force to mount the semiconductor chip (4η) on the circuit board. Y@
It is placed on a moving table (not shown). This completes the mounting process.

After completing this mounting process, the carrier (40) is transported forward in the transport direction and further transported in a vertical direction in a U-shaped path, and is inserted into the receiving magazine on the 6th floor. so that the magazine is inserted from the top.
It is installed at the deepest position at first, and the unloader f is installed for each installation.
51) to lift it upward. Intermittent driving is performed to stop at each position of the first stage of the apex. This loader/unloader 143, 5
All of the intermittent operations (m) are executed under the control of the conbeater, and the control programs for conveying the carrier (41), such as dispensing position, stopping at the mounting position, starting and ending each operation, and conveyance, are all controlled by the conbeater. controlled by

On the other hand, the semiconductor chip αη placed on the θ table 151 is transported in the following manner. In other words, semiconductor chips (
For example, there may be a large number of tray openings provided with a predetermined arrangement of recesses corresponding to the size of an IC chip, or different types may be used. In addition, other trays (I
It is constructed in a tray format so that the size and type may be different from that of C. A mark is attached to the storage container □□□, and this mark is manually detected in advance and stored in the memory as the amount of positional deviation of the storage container. Based on this amount of deviation, a pickup head (b) K driven in the X-Y direction sequentially picks out IC chips from one tray 62. For example, the pickup head (b) is attracted to K and placed on the θ table (4).

The θ table (A) is a circular plate provided with many suction holes. This θ table (A) is fixed to a rotating disk (To).

The rotation is controlled independently.

Therefore, the transferred IC chip corrects the direction of the C chip by rotating the θ table 6G. After this correction operation, the rotary disk 6!9 is rotationally transferred toward the mount head, for example, rotated by 180 degrees. In this 180 degree rotation, the θ table (A) is a disk I! Two pieces are provided on a straight line passing through the center of i9. In the case of a rotation of 0120 degrees, which is determined by the transport position to the mount position, three θ tables 61 are required. The direction correction means for the IC chip calculates the amount of positional deviation by means of comparison between the imaging output of the television camera and standard position information stored in advance;
Based on this amount of positional deviation, the amount of rotation of the θ table Cl1l is calculated and corrected.

A TV monitor 6η158 is also installed, and a monitor 67)
displays the imaging output of the television camera (4), and the monitor (to) switches and displays the imaging output of the television camera (4). In other words, when the dispense position is detected, the camera automatically focuses on the TV camera and then reproduces the image output.
When detecting the mount position, the computer controls the video camera (4I) to play back the imaging output.

In this way, dispensing work, mounting work,
- is automatically carried out on the circuit board transport path, so everything can be carried out automatically, which is effective in saving manpower.

This has the effect of enabling fully automatic and accurate bonding of bonding objects.

[Brief explanation of the drawing]

FIG. 1 is a configuration diagram for explaining an embodiment of the method of the present invention,
FIG. 2 is a perspective view for explaining an embodiment in which the apparatus shown in FIG. 1 is applied to a die bonder. 2...X-axis table, 4...Y-axis table, 6
...Upper and lower blocks, 11...ITV camera, 10.
...Mount head.

Claims (1)

[Claims] (1) The amount of positional deviation is calculated by comparing the binarized signal pattern obtained by imaging the object to be bonded with a television camera and the standard pattern stored in advance, and the actual position is detected from this amount of positional deviation and automatically In the method of performing bonding, the amount of defocus is calculated by comparing the imaging output signal at a predetermined position of the object to be bonded with a standard signal stored in advance when taking an image of the object to be bonded, and the amount of defocus is calculated when this amount of deviation is zero. and a control procedure including a program for executing the Z-axis direction position of the television camera or focusing of the optical lens of the television camera so that