JPS603777B2 - Wire bonding method - Google Patents

Description

[Detailed description of the invention] The present invention relates to a wire bonding method.

When manufacturing semiconductor integrated circuits (ICs) and large-scale integrated circuits (Mr. 1), semiconductor pellets S are used as shown in Figure 1.
After positioning the lead frame L with the attached lead frame L on the processing table T, the wire W is moved by displacing a tool (not shown) that holds the wire W relative to the lead frame L and the semiconductor bellet S. Wiring is performed by guiding and bonding to the dot post LP provided on the lead frame L and the pad SP of the semiconductor pellet S, respectively.

Generally, the actual bonding points of the hard frame L and pellet S that are positioned on the processing table T for bonding are shifted from the predetermined regular bonding points.

Therefore, when bonding, for example, Tsubaki Kaisho 51-419
As shown in Publication No. 6, the position of the bonding point is corrected by a four-point correction method. In this correction method, two fixed points on the lead post are determined in advance, and the actual bonding position of each point on the lead post is calculated by detecting the actual positional deviation of these two fixed points. It is assumed that there is no variation in the relative position of the post. Now, the variation in the relative position of the lead posts is the number of lead pins (number of lead posts)
This does not occur much in ICs with a small number of lead pins, but becomes large in ICs with a large number of lead pins.

For this reason, fully automatic bonding of ICs with multi-pin lead frames has not been possible. The present invention was made in view of the above-mentioned drawbacks of the prior art, and an object of the present invention is to provide a wire bonding method that can automatically correct misalignment of lead posts and automatically perform bonding even in multi-pin lead frames. shall be.

Hereinafter, the present invention will be explained with reference to illustrated embodiments. FIG. 2 is a block diagram of a correction value detection device used in the method of the present invention. The timing circuit 10 includes, for example, an oscillation circuit 11 that transmits a standard signal of KHz, and a synchronization pulse generation circuit 12 that outputs constant horizontal and vertical signals using the pulses of the oscillation circuit 11. The output is input to a television camera 14 through a deflection circuit 13.
In this manner, the horizontal and vertical deflection outputs of the television camera 14 are produced by one timing circuit, so that the screen is completely synchronized with the signal of the timing circuit 10. The television camera 14 images the lead frame L and the pellet S attached to the lead frame. The output of this television camera 14 is supplied to a binarization circuit 16 through an amplifier 15.
This binarization circuit 16 digitizes the bright part of the object, ie, the lead post part, to "1" and the dark part, ie, the part around the lead post, to "0". The binarized digital signal is displayed on a television monitor 17 as an image.

On the other hand, the horizontal and vertical signals output from the timing circuit 10 are input to a detection range creation circuit 20, and from this circuit 2, a rectangular detection range signal 20a of a constant size is created.

The detection range of this detection range signal 20a is created at the time of bonding on the regular bonding point output based on the bonding data inputted in advance to the computer 21,
The image is displayed on the monitor 17 as a rectangular image. Now, assume that bonding is to be performed on one lead post extending in the vertical Y direction as shown in FIG. 3a.

Position 1 indicated by the two-dot chain line is the normal position of the lead post during bonding, and the normal bonding point is A.
shall be.

If position 0 shown by the solid line is the actual lead post position, bonding must be made to point A, so
It is necessary to detect the amount of deviation between A and A and correct the positional deviation of the bonding point. In this case, it is ideal to bond to a fixed point A on the lead post, but it is not necessarily necessary to bond to a fixed point, and the bonding point should be positioned within the shaded area as shown in Figure 4. It is sufficient to correct the deviation, and there is sufficient margin in the longitudinal direction of the lead post LP. Returning to FIG. 3a again, as explained in FIG. The actual lead post LP located here appears white.

On the other hand, the vertical width y created by the detection range creation circuit 20
, the center point of the rectangular detection range B with horizontal width x moves to the regular bonding point A that is bonded according to the bonding data signal of the microcomputer 21 that has been input in advance, and detection is performed with this regular bonding point A as the center. Range B is created. Figure 3b shows the case where the lead post LP is shifted to the right of the normal bonding point A, Figures c and d are the cases at the captain's lead post, where c shows the case where the lead post LP shifts above point A, and d shows the case where the lead post LP shifts to the right side of the regular bonding point A. This shows the case where the position shifts downward from point A.

Further, the detection range B is set as follows.

It is determined that bonding is possible when the entire detection range B is within the lead post LP, and the center A of the detection range at this time is set as the bonding point. Therefore, if the detection range is as large as possible, the bonding point will be the lead. It is good because it is close to the center of For example, in the case of Figure 5a, the x direction is made as large as possible, and in the case of figure 5b, the Y direction is made as large as possible.
In addition, when using one detection range for all lead posts, set the minimum detection range in both the . Of course, when setting the detection range for each lead post, the optimum detection range is set for each lead post. Returning to FIGS. 2 and 3 again, the white part of the actual binary lead post LP and the detection range B are determined by the AND circuit 22 by the AND circuit 22 to determine the overlap C between the two, as shown in FIG. ×,, Δy, are detected. The output of this AND circuit 22 is input to the microcomputer 21,
, by subtracting in the Y direction, the movement amount △x: x - △x,
, Δy=y−Δy· are obtained. Further, the direction of deviation of the lead post is calculated based on the white portion of the lead post LP and the position of the detection range B. Since the position of the detection range B is fixed on the TV monitor 17, the processing table T (see FIG. 1) is driven by the microcomputer 21 so that the lead post LP moves by Δx and Δy in the direction of the detection range. output to the processing table drive mechanism 23. The predetermined coordinates of the bonding point are corrected by the number of pulses corresponding to the movement amounts Δx and Δy outside the microconveyor. In this way, each bonding point is corrected and bonded each time.
In the above embodiment, the case where the lead appears white has been described, but the same procedure can be applied to the case where the lead appears black. The same applies to the detection range B. Further, in the above embodiment, a case has been described in which a deviation of a lead-side lead post is detected, but a case where a deviation of a pad on a bullet side is detected can be similarly performed. As is clear from the above explanation, according to the wire bonding method of the present invention, the detection range set at the regular bonding point and the actual bonding position are compared to calculate the amount of deviation, and the amount of deviation is calculated by comparing the detection range set at the regular bonding point with the actual bonding position. Since the tool is guided to the position for bonding, variations in the lead posts in the X and Y directions are automatically corrected, and fully automatic wire bonding is also possible for multi-pin lead frames.

[Brief Description of the Drawings] Fig. 1 is a schematic plan view showing the bonding state of a semiconductor positioned on a processing table, and Fig. 2 is a block diagram showing an embodiment of a correction value detection device used in the method of the present invention. , 3rd
Figures a, b, c, and d are explanatory diagrams for explaining the correction value detection state, Figure 4 is an explanatory diagram of the lead post showing the bonding range, and Figure 5 a, b, and c are for setting the detection range. It is an explanatory diagram showing a method. 10...timing circuit, 14...TV camera, 16...binarization circuit, 17...
...Monitor, 20...Detection range creation circuit, 2
2...AND circuit, S...semiconductor bellet, SP...pad, L...lead frame, LP...lead post, T...processing Stand, W...
...Wire. Figure 1 Figure 2 Figure 3 Figure 4 Figure 5

Claims (1)

[Claims] 1 Place the lead frame to which the semiconductor pellet is attached on a processing table, and displace the tool that holds the wire relative to the lead frame and the semiconductor pellet to connect the wire to the lead post provided on the lead frame. In the wire bonding method in which lead posts are guided and bonded to respective spots on the semiconductor pellet, a lead post imaged by a television camera that is deflected by the output of a timing circuit is binarized, and from the regular bonding point of this binarized image, In order to detect the deviation, a detection range creation circuit that creates a detection range based on the output of the timing circuit creates a detection range of a certain size, and this detection range is used as a bonding data signal of the microcomputer that has been input in advance. The lead post is projected on the regular bonding point corresponding to the binary image of the lead post, and the lead post is moved so that this detection range matches the image of the lead post, and the regular bonding point is determined by the amount of movement. A wire bonding method characterized in that wire bonding is performed by correcting this and guiding a tool to the corrected bonding point.