Bonding Wire Inspection Apparatus and Method

1. Field of the Invention

The present invention relates to a bonding wire inspection apparatus and method for inspecting wires, particularly the shape of balls at the ends of bonded wires, between the pads of a semiconductor chip and the leads of a lead frame.

2. Prior Art

FIGS. 15 through 18 show a sample 6 that is an example to be inspected. The sample 6 includes wires 5 that are bonded between the pads 2 of a semiconductor chip 1 and the leads 4 of a lead frame 3. In these Figures, reference numeral 7 indicates squeezed out paste that is used for bonding the semiconductor chip 1 to the lead frame 3. FIG. 19 shows one of the conventional inspection apparatuses.

As shown in FIG. 19, a vertical illuminating means 11 is located above the sample 6, which is placed on an inspection stand 10. The vertical illuminating means 11 is installed at the bottom of an optical means 12. A CCD camera 13, that consists of a CCD (photoelectric converter element) and a drive section therefor, is mounted on the top of the optical means 12. The CCD outputs an image signal that is an image of the sample 6 take in by the optical means 12. The image signal from the CCD camera 13 is processed by an image-processing arithmetic unit 14, and the shape of the wire 5 is recognized and the bonded position of the wire 5 is measured.

The optical means 12, that includes the vertical illuminating means 11 and the CCD camera 13, is provided on an X-Y table 15. Inside the vertical illuminating means 11, a half-mirror 16 is installed so that the half-mirror 16 is beneath the CCD camera 13. Also inside the vertical illuminating means 11 are a condensing lens 17 and a light bulb 18 that are installed on one side of the half-mirror 16.

The bonding wire inspection apparatus of this type is described in, for example, Japanese Patent Application Laid-Open (Kokai) Nos. 2-129942 and 3-76137.

In the apparatus shown in FIG. 19, when the vertical illuminating means 11 is switched on, the light from the light bulb 18 passes through the condensing lens 17, is reflected by the half-mirror 16, and then vertically directed onto the sample 6. Then, the X-Y table 15 is moved so that the CCD camera 13 is positioned above the object of inspection or the sample 6, and the image signal (that represents microscopic parts of the object of inspection, such as the ball and crescent at both ends of a bonded wire) obtained by the CCD camera 13 is subjected to noise removal, edge enhancement in the inspection area and enlargement or reduction of the image, etc. by means of the image-processing arithmetic unit 14, thus enhancing (restoring) the image of the microscopic parts of the object of inspection so that the image is easier to see. Afterward, the inspection and measurement operations are performed.

When the microscopic parts of the object of inspection are thus optically imaged, the magnification of the optical means 12 is set at a high value. For example, the magnification is set to be 2 to 3 times larger when a 2/3-inch CCD camera is used.

In the prior art described above, inspection is performed using the vertical illuminating means 11 so that the ball 5a appears dark as shown in FIG. 17. Accordingly, the area where the wire 5A is and where the edges of the pad 2 are appear dark as shown in FIG. 18 and FIG. 19(b), respectively, and the correct diameter of the ball cannot be detected. Furthermore, as shown in FIG. 19(b), the shadow 95 of the wire is unnecessarily reproduced by the camera; though, on the other hand, as shown in FIG. 19(c), the area 96 of the ball 5a which is beyond the outer circumference edge of the pad 2 is not reproduced.

Accordingly, one object of the present invention is to provide a bonding wire inspection apparatus and method that obtain an image of the ball shape even though there are not-inspected items such as a wire, etc. between the ball and an optical means which takes in the image of the ball.

Another object of the present invention is to provide a bonding wire inspection apparatus and method that inspects the ball shape without being disturbed by the pattern shape nor brightness of the semiconductor chip surface.

The objects of the present invention are accomplished by a unique structure for a bonding wire inspection apparatus that inspects wires bonded between the pads of a semiconductor chip and the leads of a lead frame, and the structure includes a ring-form illuminating means that consists of plural LED's installed in a ring-form configuration and illuminates the object of inspection at an inclined angle; optical means that include diaphragms and receive reflected lights created by the illumination provided by the ring-form illuminating means and then image the ball shape; and cameras which photograph the images obtained by the optical means.

In addition, the object of the present invention is accomplished by a unique method that inspects wires bonded between the pads of a semiconductor chip and the leads of a lead frame, in which the inspection is performed under a dark-field illumination circumference created for the object of inspection so that only such an object (which is a ball at one end of the bonded wire) can appear bright.

When the ball is illuminated by a low angle of inclination by the ring-form illuminating means, the surface of the semiconductor chip gives out only a directly reflected light, because the surface of the semiconductor chip is more or less in a mirror surface state. Accordingly, reflected light from the surface of the semiconductor chip does not enter the optical means. This means that the surface of the semiconductor chip appears dark. On the other hand, the ball which is on the chip reflects a considerable amount of lights that enter the optical means; thus, the ball appears bright. In other words, by creating a dark-field illumination circumstance around the ball, all the areas except for the ball that projects from the surface of a semiconductor chip appear dark, resulting in that the ball alone can appear bright.

Furthermore, by using a diaphragm means that can alter the opening diameters of the optical means, the ball alone can further be reproduced as a bright image. In this case, the diaphragm means is opened large so as to create a shallow focal depth, and the focal length of the optical means is set matched on the surface of the semiconductor chip by raising and lowering (that is, adjusting) the height of the optical means. By doing this, since there is little reflected light from the wire, the wire is not imaged and the ball alone is illuminated bright.

FIG. 1 is a cross section of the bonding wire inspection apparatus according to one embodiment of the present invention;

FIG. 2 is a perspective view of the Z-direction driving means for the objective lens group used in the apparatus of FIG. 1;

FIG. 3 is a cross section of the ring-form illumination assembly used in the apparatus shown in FIG. 1;

FIG. 4 illustrates the layout of the LED's used in the ring-form illumination assembly shown in FIG. 3;

FIG. 5 is a block diagram of the illumination switching circuit for the ring-form illumination assembly;

FIG. 6 is a block diagram of the control circuit of the bonding wire inspection apparatus shown in FIG. 1;

FIG. 7 illustrates the illumination of light onto the ball of a bonding wire;

FIG. 8 illustrates a problem (shadow) that would occur in the illumination shown in FIG. 7;

FIGS. 9(a), 9(b), and 9(c) show diagrams of images obtained under the problem of FIG. 8;

FIG. 10 shows the illumination of light by the ring-form illumination assembly onto the ball of the bonded wire;

FIG. 11 shows the shape of the ball detected under the illumination shown in FIG. 10;

FIG. 12 shows the illumination of light onto the crescent of the bonded wire;

FIG. 13 illustrates the wire height inspection, wherein the focus is on the semiconductor chip;

FIG. 14 illustrates the wire height inspection, wherein the focus is on the lead;

FIG. 15 is a top view of a wire-bonded semiconductor device;

FIG. 16 is a front view thereof;

FIG. 17 is an enlarged front view of the bonding wire bonded to the pad of a semiconductor device;

FIG. 18 is a top view thereof; and

FIG. 19 is a structural diagram of a conventional bonding wire inspection apparatus.

One embodiment of the present invention will be described with reference to FIGS. 1 through 6. In these Figures, parts that are the same as those in FIGS. 15 through 18 are given the same symbols, and a detailed description of those parts is omitted.

As seen from FIG. 1, an opto-head 23 is mounted on an X-Y table 22 which is driven in the X and Y (or horizontal) directions by an X-direction drive motor 20 and a Y-direction drive motor 21, respectively.

The opto-head 23 has an objective lens group 24 which is installed so as to face an inspection stand 10. As shown in FIGS. 1 and 2, the objective lens group 24 is mounted to a linear slider 25. The linear slider 25 is installed so that it can move up and down along a pair of linear gauges 26 which are fixed to the opto-head 23. The objective lens group 24 is also attached to a roller-supporting plate 27, and a roller 28 is rotatably mounted on the roller-supporting plate 27. In addition, a Z-direction drive motor 29 is installed in the opto-head 23. A cam 30 is attached to the output shaft of the Z-direction drive motor 29, and the roller-supporting plate 27 is urged downwardly by a spring 31 so that the roller 28 is in contact with and presses against the cam 30.

As shown in FIG. 1, a first half-prism 35, which is used to split light, is installed above the objective lens group 24, and an illumination lens group 36 is installed above the first half-prism 35. In addition, a diffusing plate 37 is installed above the illumination lens group 36, and a vertical illuminator 38 is installed above the diffusing plate 37.

A second half-prism 39 is installed next (on the right side in FIG. 1) to the first half-prism 35, and a low-magnification imaging lens group 40 is installed above the second half-prism 39. Furthermore, a low-magnification camera 41 is installed above the low-magnification imaging lens group 40.

Reference numeral 42 refers to a high-magnification imaging lens group which is installed next (on the right side in FIG. 1) to the second half-prism 39, and a reflecting mirror 43 is installed next (on the right side in FIG. 1) to the high-magnification imaging lens group 42. Furthermore, a high-magnification camera 44 is installed above the reflecting mirror 43.

Diaphragms 50, 51 and 52 that are electrically operated are respectively installed beneath the vertical illuminator 38, between the first and second half-prisms 35 and 39, and between the second half-prism 39 and the high-magnification imaging lens group 42. These electrically operated diaphragms 50, 51 and 52 are respectively driven by an illumination diaphragm driver 53, a low-magnification diaphragm driver 54 and a high-magnification diaphragm driver 55, and the opening diameters of the diaphragms 50, 51 and 52 are altered accordingly. The electrically operated diaphragms 50, 51 and 52 are similar to diaphragms used in, for example, ordinary cameras.

The diaphragm 50 controls the angle of illumination of the vertical illuminator 38 with respect to the object of inspection or the sample 6. The diaphragm 51 adjusts the opening diameter of the objective lens group 24. In other words, the diaphragm 51 adjusts the focal depth for the objective lens group 24. The diaphragm 52 equalizes the brightness obtained at high and low magnification rates.

The opto-head 23 has a ring-form illumination assembly 60. The illumination assembly 60 is positioned so that it is under the objective lens group 24.

The ring-form illumination assembly 60, the detail thereof being shown in FIGS. 3 and 4, comprises an inner ring-form illuminator 63 that has a high illumination angle and an outer ring-form illuminator 65 that has a low illumination angle.

The inner ring-form illuminator 63 is made up of plural LED's 62 installed in the form of a ring or circle. The LED's 62 are mounted to an illumination holding plate 61 so that they are around the center opening 61a of the holding plate 61. The outer ring-form illuminator 65 is also made up of plural LED's 64 mounted to the illumination holding plate 61. The LED's 64 of the outer ring-form illuminator 65 are arranged in the form of a ring or circle around the inner ring-form illuminator 63.

The respective circular rows of LED's 62 and 64 face toward the optical axis of the vertical illuminator 38. The LED's 62 of the inner ring-form illuminator 63 are inclined at an angle of approximately 30 to 55 degrees relative to a horizontal level, thus having a high illumination angle. The LED's 64 of the outer ring-form illuminator 65 are inclined at an angle of approximately 5 to 15 degrees relative to the horizontal level, thus having a low illumination angle.

An attachment plate 67 is mounted on the illumination holding plate 61 via a supporting column 66, and the illumination holding plate 61 is fastened to the attachment plate 67 by a screw 68 that is inserted into the supporting column 66. A spacer 69 is fitted over the supporting column 66, and a light-screening plate 70 and a cover 71 are installed between the spacer 69 and the attachment plate 67.

In FIG. 3, only the right and left side LED's 62 and 64 are shown; and in FIG. 4, some of the LED's 62 and 64 are omitted with those omitted LED's shown by broken lines.

The LED's 62 and 64 of the inner and outer ring-form illuminators 63 and 65 are operated by an illumination switching circuit that is shown in FIG. 5.

In FIG. 5, one end of each of the LED's 62 and 64 is connected to a programmable constant-voltage circuit 81 via a resistance 80, while another end of each LED is connected to the collector of a transistor 83. The emitters of the transistors 83 are grounded, and the bases of the transistors 83 are connected to a shift register group 85 via a latching register group 53. A load signal 86 is inputted into the latching register group 84, a high-low (H-L) input signal 87 is inputted into the shift register group 85, and a clock signal 88 is inputted into each shift register of the shift register group 85.

As shown in FIG. 6, the low-magnification camera 41 and the high-magnification camera 44 convert the obtained images into electrical signals and send those signals to an image processing unit 90. The images are processed by the image processing unit 90 and displayed on a monitor 91. The processed images are also sent to a main controller 93 which has a memory 92. Furthermore, via the signals from the main controller 93, the image processing unit 90 is operated.

The X-direction drive motor 20, the Y-direction drive motor 21, and the Z-direction drive motor 29 are all controlled by the main controller 93; and positional signals from the X-, Y-, and Z-direction drive motors 20, 21, and 29 are sent to the main controller 93. In addition, the illumination diaphragm driver 53, the low-magnification diaphragm driver 54, the high-magnification diaphragm driver 55, the vertical illuminator 38, the inner ring-form illuminator 63 and the outer ring-form illuminator 65 are all controlled by the main controller 93.

Next, the operation of the embodiment will be described below with referece to FIGS. 7 to 15.

Prior to the commencement of actual inspection, opto-head information is inputted, as described below, into the memory 92 and stored therein. First, a sample that is the same type as the object of inspection (sample 6) is selected, and alignment points and alignment images for calculating the positions of the pad 2 and the leads 4 are stored in the memory 92. In this case, the opening diameters of the respective diaphragms 50, 51 and 52 and the rotational amount of the output shaft of the Z-direction drive motor 29 are also stored in the memory 92, so that they will be used during the inspection. The "alignment points" includes a total of four points, which are as shown in FIG. 15, for example, two points 2a and 2b on the semiconductor chip 1 and two points 4a and 4b on the lead 4.

First, the inspection of the alignment points 2a and 2b on the semiconductor chip 1 and the alignment points 4a and 4b on the lead 4 in the sample 6 will be described.

When these alignment points are inspected, the ring-form illumination assembly 60 is not switched on, but the vertical illuminator 38 is switched on, thus creating a bright visual field for the sample 6 in order to stabilize the positional correction of the alignment points (2a, 2b, 4a, 4b) at a high level of precision. In addition, the high-magnification camera 44 is used. The opening diameter of the diaphragm 51 is set small, and the opening diameter of the diaphragm 50 is also set small, thus creating a large focal depth.

After these preparations, the sample 6 that is the object of inspection is placed on the inspection stand 10. Then, the X-Y table 22 is moved in accordance with the coordinates of the alignment point 2a of the semiconductor chip 1, which are stored in the memory 92 beforehand as described above, thus positioning the center of the objective lens group 24 roughly over the alignment point 2a so that the alignment point 2a of the sample 6 is located. At this time, the Z-direction drive motor 29 is also actuated to the stored position, thus causing the objective lens group 24 to be moved up and down so that the image of the alignment point 2a that is to be taken in by the objective lens group 24 is focused. In other words, by actuating the Z-direction drive motor 29, the objective lens group 24 is raised or lowered by the cam 30 via the roller 28 and the roller supporting plate 27. In addition, the light from the vertical illuminator 38 is directed onto the area near the alignment point 2a via the diffusing plate 37, the diaphragm 50, the illumination-use lens group 36, the first half-prism 35 and the objective lens group 24.

The image of the alignment point 2a and its surrounding area are picked up by the high-magnification camera 44 via the objective lens group 24, the first half-prism 35, the diaphragm 51, the second half-prism 39, the high magnification imaging lens group 42 and the mirror 43. This image is processed by the image processing unit 90, displayed on the monitor 91, and also stored in the memory 92 by the main controller 93.

The other alignment points 2b, 4a, and 4b are also detected in the same manner as the point 2a is detected as described above.

As a result, the positional coordinates of the respective pads 2 and leads 4 are corrected by the main controller 93 based upon the positional coordinates of the detected alignment points 2a, 2b, 4a, and 4b.

When the inspection of the alignment points 2a, 2b, 4a, and 4b are performed, the conditions (that is, the opening diameters) of the electrically operated diaphragms 50, 51 and 52, the intensity of the vertical illuminator 38 and the rotational amount of the (output shaft of the) Z-direction drive motor 29 (in other words, the position of the objective lens group 24), etc. that are used are those that are initially set and stored in the memory 92 as described above.

Usually, due to the presence of the paste between the semiconductor chip and the lead frame, the heights of the semiconductor chips vary (chip to chip) by approximately 20 to 30 microns. In addition, the surfaces of the semiconductor chips are slightly inclined with respect to the horizontal direction. In order to detect this height discrepancy, the X-Y table 22 and/or the Z-direction drive motor 29 are/is actuated so as to perform automatic focusing. In this case, the X-Y table 22 is driven so that the objective lens group 24 is moved to three previously inputted points (e.g., 2a, 2c and 2b) on each semiconductor chip 1 and the Z-direction drive motor 29 is actuated so that objective lens group 24 focuses the chip 1. Then, the respective detected height values (i.e., rotational amount of the Z-direction motor 29) are stored by the main controller 93.

There are various different methods in automatic focusing referred to above. In the present embodiment, the automatic focusing is accomplished by the use of a combination of pattern matching method and a histogram method. Thus, the opening diameter of the electrically operated diaphragm 51 is set large so that the focal depth can be shallow. Then, using the pattern matching value (that is obtained while moving the objective lens group 24.+-.50 microns from the stored rotational position of the Z-direction drive motor 29) and a histogram of the image obtained at this time, the maximum value is taken as a focus matching 21.

The focus matching at the position where the ball 5a is bonded (i.e., the rotational amount of the Z-direction drive motor 29) is ascertained from the differences in the heights detected between the three points on the surface of the semiconductor chip 1 and the size of the semiconductor chip 1. Accordingly, when the ball 5a is inspected (described below), it is accomplished by moving the objective lens group 24 to the position of the ball 5a via the X-Y table 22 and the Z-direction drive motor 29.

The ball, the crescent and the wire of a bonded wire in a sample will be described below. As an example, the inspection of the ball 5a, the crescent 5b and wire 5A will be described.

When the ball 5a is inspected, the X-Y table 22 is driven so that the center of the objective lens group 24 is brought to coincide with the corrected position of the ball 5a. Then, the ring-form illuminator 65 is switched on, but the vertical illuminator 38 and the inner ring-form illuminator 63 of the ring-form illumination assembly 60 are not switched on. In other words, a dark visual field circumstance is created so that only the ball 5a is made to appear bright. Furthermore, the opening diameter of the diaphragm 51 is set large so that a shallow focal depth is created.

As shown in FIG. 7, when the ball 5a is illuminated only by the outer ring-form illuminator 65 from an angle of approximately 5 to 15 degrees, there is only a directly reflected light on the semiconductor chip 1, since the surface of the semiconductor chip 1 is in a more or less mirror surface state. Accordingly, the light reflected by the surface of the semiconductor chip 1 does not enter the objective lens group 24, and as a result, the surface of the semiconductor chip 1 appears dark. However, the ball 5a provides a considerable amount of directly reflected light that enters the objective lens group 24; as a result, the ball 5a appears bright.

Furthermore, if the opening diameter of the diaphragm 51 is set large and thus creating a shallow focal depth, and if the focal length is set so as to match the surface of the semiconductor chip 1, since the reflected light from the wire 5A is minimal, the wire 5A, as shown in FIG. 8, is not imaged, and only the ball 5a appears bright.

In other words, by creating a dark-field illumination circumstance for the ball 5a, it is possible to darken everything, including the pad 2, pattern, etc., but the ball 5a that projects from the surface of the semiconductor chip 1; thus, only the ball 5a appears bright.

In addition, as mentioned above, by setting the opening diameter of the diaphragm 51 large and thus creating a shallow focal depth, it is possible to eliminate the shadow of the wire 5A that is seen above the ball 5a. Also, since the illumination is provided only by the monochromatic light emitted by the LED's of the outer ring-form illuminator 65, the focal depth can be as shallow as approximately 10 to 20 microns. Thus, the shape of the ball 5a can be accurately ascertained via image processing, and the accurate ball diameter is obtained.

Peelings, floating, etc. of the ball at the end of the bonded wire have been considered impossible to detect in the past. However, such defects can appear as an indistinct image of the edge of the ball in the present invention. Accordingly, they can easily be detected by the apparatus and method of the present invention.

Brcause of the illumination provided by the outer ring-form illuminator 65, the shape of the ball 5a is clearly reproduced as a bright image by the high-magnification camera 44 via the objective lens group 24, the first half-prism 35, the diaphragm 51, the second half-prism 39, the diaphragm 52, the high-magnification lens group 42 and the mirror 43.

When the shape of the ball 5a is inspected, the bright boundary is located using a commonly used image processing technique, so that the surface area, the center of gravity and the radius of the ball 5a are measured.

Next, the inspection of the crescent 5b of the wire 5A will be described.

In this case, the X-Y table 22 is driven so that the center of the objective lens group 24 is brought to coincide with the corrected position of the crescent 5b. The inspection of the crescent 5b is performed under a bright visual field as in the same manner as the inspection of the ball 5a. In other words, the inspection is performed with the vertical illuminator 38 switched on. However, since the lead 4, on which the crescent 56 exists, has more surface irregularities than the semiconductor chip 1, not only the lead 4 but also the crescent 5b tend to appear dark. Accordingly, in the crescent inspection, the electrically operated diaphragm 50 is opened wide so that the angled illumination from the vertical illuminator 38 is also used as shown in FIG. 12.

If necessary, the inner ring-form illuminator 63 of the ring-form illumination assembly 60 is switched on, and the outer ring-form illuminator 65 is switched off. When illuminated perpendicularly by the vertical illuminator 38, the crescent 5b provides lesser irregular reflected light that the lead 4; therefore, the crescent 5b appears dark. However, since there are indentations and projections on the lead 4, illumination by the vertical illuminator 38 creates brightness irregularity on the lead 4. Accordingly, the crescent 5b is illuminated from an angle of approximately 30 to 55 degrees by the use of the inner ring-form illuminator 63 of the ring-form illumination assembly 60, so that the brightness irregularity on the lead 4 caused by the illumination by the vertical illuminator 38 is eliminated

However, the illumination provided by all the LED's 62 of the inner ring-form illuminator 63 from the direction indicated by arrow A in FIG. 12 (i.e., from the direction of the lead 4 to which the crescent 5b to be inspected is bonded) brightens only a portion of the crescent 5b. Thus, a correct entire shape of the crescent 5b is not imaged. In order to avoid this, some LED's 62 which illuminate the crescent 5b from the direction indicated by arrow A are switched off. In other words, the LED group 62A shown in FIG. 4 is turned off. This is accomplished via the illumination switching circuit shown in FIG. 5.

More specifically, the respective shift registers of the shift register group 85 are successively switched by the clock signal 88; and in each case, a high (on) or low (off) input signal 87 is inputted; the high or low input signal 87 is successively stored in each register of the shift register group 85 via the clock signal 88; and then the signal stored in each register is transferred to the latch register group 84 via the load signal 86. As a result, the LED group 62A that illuminate the crescent 5b from the direction indicated by arrow A is not turned on. Which LED's 62 are to be switched off is designated beforehand by a program stored in a control circuit (not shown) in accordance with the object to be inspected.

Accordingly, as described in the inspection of the ball 5a, the crescent 5b is clearly reproduced by the high-magnification camera 44 via the objective lens group 24, the first half-prisnm 35, the diaphragm 51, the second half-prism 39, the diaphragm 52, the high magnification imaging lens group 42 and the mirror 43.

Lastly, the inspection of the presence, linearity and height of the wire 5A will be described.

The presence and linearity of the wire 5A is inspected after the ball 5a and the crescent 5b have been inspected by the methods as described above. This presence and linearity inspection of the wire 5A is accomplished by the image processing unit 90; in other words, the image processing unit 90 processes a bright line that appears between the ball 5a and the crescent 5b. In this inspection, the vertical illuminator 38 and the inner ring-form illuminator 63 of the ring-form illumination assembly 60 are not switched on, and only the outer ring-form illuminator 65 is switched on. The opening diameter of the diaphragm 51 is set small.

If the vertical illuminator 38 is on, the reflected light from the area surrounding the wire 5A (i.e., the semiconductor chip 1, lead 4 and paste 7) will enter the low-magnification camera 41 to a greater extent than the reflected light from the wire 5A, thus brightening such a surrounding area, resulting in that the wire 5A does not appear bright. The reason for this is that the surface of the bonding wire is in an Au mirror surface state, and there are various bright and dark spots in the area surrounding each wire. Accordingly, when the wire is inspected, it is necessary, in order to increase the precision of inspection as well, to darken the entire area that surrounds the wire and brighten the wire only. For this reason, the vertical illuminator 38 is not turned on nor used. In addition, the angle of incidence of the light from the LED's 62 of the inner ring-form illuminator 63 is as large as approximately 45 degrees; as a result, there is considerable reflected light from an area other than the wire 5A if the inner ring-form illuminator 63 is on. For this reason, the inner ring-form illuminator 63 is not turned on nor used.

Experiments show that illumination at an angle of incidence of light of approximately 5 to 15 degrees is the most appropriate for making the wire alone appear bright. When such a small angle of incidence of light is used, directly reflected light from the area other than the wire is extremely small.

However, if the wire 5A is illuminated by all the LED's 64 of the outer ring-form illuminator 65, some portions of the wire 5A are obscured by the paste 7. Accordingly, the LED group 64A that provides illumination from the direction indicated by arrow A (i.e., from the side of the lead 4 to which the wire 5A being inspected is bonded) is turned off as in the case of inspection of the crescent 5b (in which the LED group 62A is turned off). As a result, the wire 5A can be clearly reproduced by the low-magnification camera 41. The positions of the ball 5a and the crescent 5c are already known and stored in the image processing unit 90 as described above; thus a bright line that appears between the ball 5a and the crescent 5c and represents the wire 5A can be easily traced. Thus, the presence and linearity of the wire 5A is inspected.

The height of the wire 5A is inspected as described below. In the wire height inspection, the electrically operated diaphragm 51 is opened, and the low-magnification camera 41 is used.

By opening the diaphragm 51, the focal depth of the objective lens group 24 becomes shallow. Accordingly, when the focus of the objective lens group 24 is brought into the semiconductor chip 1, the image of the wire 5A is obtained by the low-magnification camera 41 via the objective lens group 24, the first half-prism 35, the diaphragm 51, the second half-prism 39 and the low-magnification imaging lens group 40 and then processed by the image processing unit 90. The thus processed image appears as shown in FIG. 13. On the other hand, when the focus of the object lens group 24 is brought into the lead 4, the image of the wire 5A appears as shown in FIG. 14; in other words, the wire 5A appears indistinctly and differently in the width. For this reason, the widths that are at right angles to a (imaginary) line 5c, that represents the presence of the wire, between the ball 5a and the crescent 5b (which has been confirmed by the inspection of the wire as described above), in other words, the indistinct-image widths B.sub.1, B.sub.2, B.sub.3 . . . , are measured by the main controller 93.

Experiments show that there is a constant relationship between the indistinct-image widths and the height of a wire as obtained by the Equation below:

In the above Equation, the "Height to be Determined" refers to the height of wire from, in this embodiment, the focal position of the objective lens group 24 that is calculated from the rotational amount of the Z-direction drive motor 29 at the time of the inspection. In addition, "K" in the Equation is a proportionality constant.

Accordingly, by measuring the indistinct image with the main controller 93, the height at any desired point of the wire 5A can easily be measured.

However, if the indistinct image width becomes larger, the difference in the relative brightness between the indistinct image and its surrounding area becomes smaller, thus making the recognition of the wire difficult. In this case, the Z-direction drive motor 29 is actuated, and the focus of the objective lens group 24 is shifted from its position on the semiconductor chip to a position on the lead in order to take in an indistinct image of the wire 5A based upon the focal position on the lead 4.

By combining the indistinct image of the wire with the focus brought into the semiconductor chip and the indistinct image of the wire with the focus brought into the lead frame, a precise recognition of the wire is obtained. That is to say, the image in the vicinity of the focal point has greater contrast; accordingly, the difference between the indistinct image and its surroundings (i.e., the edges of the indistinct image) can be more easily determined.

As described above, depending upon the object of inspection and in order to obtain a clear reproduction of the image of the object, the vertical illuminator 30, the inner ring-form illuminator 63 and/or the outer ring-form illuminator 65 of the ring-form illumination assembly 60 are/is turned on and off, and also some of the LED's of the inner and outer ring-form illuminators are turned on and off.

In the embodiment described above, a double ring-form illuminator that consists of the inner and outer ring-form illuminators 63 and 65 is used. However, it would be possible to install three or four ranks of ring-form illuminators that are respectively made up of plural LED's. Furthermore, when either the wires 5 or the crescents 5b (but not both) are inspected and measured, it would also be possible to use only the inner ring-form illuminator 63 or the outer ring-form illuminator 65.

As described above, since the ball 5a is illuminated from a low angle by the outer ring-form illuminator 65 that has a low illumination angle, the reflected light from the surface of the semiconductor chip 1, which is in a mirror surface state, does not enter the objective lens group 24, and accordingly, the surface of the semiconductor chip 1 appears dark. However, from the ball 5a, a considerable amount of directly reflected light enters the objective lens group 24; accordingly, the ball 5a appears bright. In other words, by creating a dark-field illumination circumstance for the ball 5a, all the areas including the pad 2, pattern, etc. but not the ball 5a that protrudes from the surface of the semiconductor chip 1 can be darkened. Thus, only the ball 5a can appear bright. Furthermore, since the opening diameter of the diaphragm 51 is set large so that a shallow focal depth is created for the surface of the semiconductor chip 1, there is little reflected light from the wire 5A. Accordingly, the wire 5A is not imaged, and only the ball 5a appears bright and clearly imaged.

As described above, the apparatus of the present invention includes: a ring-form illuminating means that consists of numerous LED's installed in a ring-form configuration and illuminates the object of inspection (ball) at an inclined angle; optical means that receive reflected light, which is created via the illumination provided by the ring-form illuminating means, and image the ball shape; and a camera that photographs the images obtained by the optical means. Accordingly, the image of the ball is obtained clearly without any fail even though items such as wires, etc. which are not to be inspected exist between the ball and the optical means. In addition, the ball shape can be inspected without being affected by the pattern shape nor the brightness of the semiconductor chip surface.