JP4324516B2 - Ball detection method and ball detection apparatus - Google Patents

Description

  The present invention relates to a ball detection method and a ball detection device for detecting surplus balls adhering to balls other than regularly aligned balls.

  Conventionally, as a method for manufacturing a BGA (Ball Grid Array), a method of mounting a solder ball on a circuit board after adsorbing and aligning the solder ball in an adsorption hole or the like provided on a ball alignment surface for aligning the solder ball is widely used. Has been. Excess balls adhering to the ball alignment surface are usually removed by dropping the ball mounting head by applying vibration to the ball mounting head. However, the number of solder balls mounted at one time is several thousand to several tens of thousands. In addition, when the diameter of the ball becomes small, there is a case where not all the excess balls can be removed. If a solder ball is mounted on a circuit board in a state where surplus balls are attached, a serious problem that a circuit board including an expensive semiconductor integrated circuit becomes defective occurs.

  A method for detecting the presence or absence of a ball with a light-shielding optical sensor is disclosed in Japanese Patent Application Laid-Open No. 11-154690. This method uses a light-shielding type optical sensor to check for the presence of balls adhering to the lower surface of the suction head, and is very effective in detecting residual balls remaining on the ball alignment surface after mounting the balls on the circuit board. .

However, as can be seen from FIG. 3, the area that the surplus ball shields from light varies greatly (15 to 100%) depending on the position where the surplus ball is attached. Further, errors due to the level of the ball alignment surface, vibration, various drifts, and the like are added. In addition, even for a ball that is regularly aligned on the ball alignment surface, the amount of light that is blocked varies depending on the sampling due to the suction state, vibration, and the like. For this reason, the S / N ratio, which is the ratio between the voltage drop value and noise caused by the light shielding of the surplus balls, is small (see FIG. 9A), and the S / N ratio is significantly reduced when the ball diameter is further reduced. Thus, there arises a problem of erroneously determining the presence or absence of surplus balls.

On the other hand, a board on which a solder ball is mounted often has dimples at the ball mounting location. The dimple is a recess formed by etching away the resist in the external connection terminal region in order to expose the external connection terminal whose substrate surface is covered with the resist. In such a substrate, there is a problem that when the solder balls are mounted on the substrate, the light shielding plate becomes an obstacle and the solder balls do not reach the external connection terminals.
Japanese Patent Laid-Open No. 11-154690

  SUMMARY OF THE INVENTION An object of the present invention is to provide a method and apparatus for reliably detecting excess balls attached to a ball alignment surface.

  The present invention relates to a light-blocking member that shields a regular ball from light in a ball detection device that detects an excess ball attached to a ball alignment surface of a ball alignment unit by an optical sensor that photoelectrically converts light emitted from a light-emitting unit at a light-receiving unit. Is provided in the ball alignment portion.

  The present invention is a ball detection device for detecting surplus balls adhering to a ball alignment surface of a ball alignment unit by an optical sensor that receives light emitted from a light emitting unit by a light receiving unit, and a part or the whole of a regular ball is detected. A light shielding member that shields light from light, a first drive mechanism that moves the light shielding member, a fixing member that fixes the first drive mechanism to the ball alignment portion, and a second drive mechanism that moves the ball alignment portion. And

  The present invention is a ball detection method for detecting surplus balls adhering to a ball alignment surface of a ball alignment unit by a light-shielding type optical sensor including a light emitting unit and a light receiving unit, and a step of moving a light shielding member to a surplus ball measurement position Scanning the vicinity of the ball alignment surface with light, photoelectrically converting the scanned light with the light receiving unit, sampling the output voltage from the light receiving unit, and determining whether there is an excess ball based on the output voltage, And a step of moving the light shielding member to the retracted position.

  The present invention is also a ball detection method for detecting surplus balls adhering to the ball alignment surface of the ball alignment unit by a light-shielding type optical sensor comprising a light-emitting part and a light-receiving part, wherein the light-shielding member is moved to the ball measurement position. An operation, an operation of moving a ball aligning unit on which a ball is mounted between a light emitting unit and a light receiving unit, an operation of sampling n output voltages and (n + 1) th output voltage from the light receiving unit, and n operations The difference between the average value B and the average value A, the operation for calculating the average value A of the output voltages, the operation for calculating the average value B of the m output voltages including the (n + 1) th output voltage, It is characterized by comprising an operation for calculating the value D, an operation for comparing the difference value D with a predetermined threshold value to determine the presence or absence of surplus balls, and an operation for retracting the light shielding member from the ball measurement position.

  Further, the present invention is characterized in that m = 1.

  Furthermore, the present invention is characterized in that the presence or absence of surplus balls is determined using a difference value curve in which the difference values D are arranged in time series.

The present invention has the following effects.
The ball detection device of the present invention can eliminate the influence of the regular ball by disposing the light blocking member that shields the regular ball from the light, so that the presence or absence of the surplus ball can be determined with high accuracy.

  The present invention can provide a ball detection device in which the light shielding plate does not obstruct the operation in other processes (for example, a ball alignment process or a ball mounting process) by providing a drive mechanism that moves the light shielding member up and down.

The ball mounting device of the present invention can remove fluctuations caused by regular balls by scanning the vicinity of the ball alignment surface with light after moving the light shielding member to the surplus ball measurement position. Presence / absence can be determined with certainty.

According to the ball mounting method of the present invention, the ball can be reliably mounted on the external connection terminal of the substrate by detecting the surplus ball using the light blocking member and then moving the light blocking member to the retracted position.

Since the ball detection method of the present invention uses the difference value D between the average value A of the n output voltages from the light receiving unit and the average value B of the m outputs including the (n + 1) th output, the background noise It is possible to easily determine the presence or absence of surplus balls by offsetting the drift.

  In the ball mounting method of the present invention, the difference value D corresponding to the surplus ball can be increased by setting m = 1.

  In the ball detection method of the present invention, the presence / absence of surplus balls is determined using a difference value curve obtained by connecting the difference values D in time series, so that the presence / absence of surplus balls can be determined more accurately.

The present invention realizes a method and apparatus for reducing background noise of an optical sensor in light detection of surplus balls.

  FIG. 1 shows an optical sensor composed of a light emitting unit 1 and a light receiving unit 2, a light shielding plate 12 movably attached to a ball alignment unit 4, and solder balls (6, 7, The relationship of 8) is shown. FIG. 2 is a bottom view thereof. FIG. 3 is an enlarged view of a part of the cross section of the ball alignment surface 5. In order to emphasize the shading part of the surplus ball, it is shaded. FIG. 4 is a front view of the slit plate 3 and shows the relationship between solder balls and light. The ball detection device of the present invention will be described with reference to FIGS. 1, 2, 3 and 4.

The light emitting unit 1 has a built-in laser element, and an emitted light beam 16 is indicated by an arrow 16. The wavelength of the laser beam is 670 nm, and the diameter of the light beam is 1 mm. The light receiving unit 2 incorporates a phototransistor, and a slit plate 3 is externally attached near the entrance. The light emitting unit 1 and the light receiving unit 2 do not move in conjunction with the movement of the ball mounting head. The ball mounting head includes the ball alignment unit 4 shown in FIG. 1 and has a function of sucking and mounting the ball on a circuit board or the like.

The slit plate 3 has a slit 15 processed at the center. It is shown with the laser light beam 16 shielding plate lower end 12a by a broken line. The regular balls 6 are exposed on the light shielding plate lower end 12a, and the surplus balls 7 are partially exposed from the light shielding plate lower end 12a. The diameter of the ball is 0.3 mm, the slit width is 0.3 mm, the slit length is 3 mm, and the diameter of the light beam is 1 mm. The diameter of the ball 6 can take various sizes (for example, 0.3, 0.2, 0.1, 0.05 mm), and the diameter of the ball is 0.3, 0.2, 0. .1, 0.05 mm, the slit widths are 0.3, 0.2, 0.1, and 0.05 mm, respectively. It should be noted that there is a disadvantage that the output voltage is reduced when the width of the slit is 1/3 or less of the ball diameter. On the other hand, if the slit width is too larger than the ball diameter, the background noise increases, so it is preferably less than twice the ball diameter.

The air cylinder 9 is fixed to the side surface of the ball alignment portion 4 via a fixing member 11. And the light-shielding plate 12 is fixed to the slide part 10 of the air cylinder 9, and can move up and down by pneumatic pressure. If the solder ball mounting substrate is a semiconductor wafer and is large, such as 8 inches or 12 inches, it is difficult to mount the solder balls at one time, and the ball mounting must be divided into a plurality of times. In this case, if the light shielding plate 12 is not retracted, the light shielding plate 12 is brought into contact with the solder balls already mounted on the semiconductor wafer.

A plurality of suction holes 13 are opened in the ball alignment surface 5, and the balls 6 are sucked into the suction holes 13. The position of the suction hole 13 corresponds to the position where the solder bump is formed on the circuit board. When the ball diameter is 0.3 mm, the suction hole has a diameter of 0.13 mm, and the outer diameter of the chamfer 14 is 0.34 mm.

When the ball 6 attracted to the suction hole 13 is referred to as being distinguished from the surplus ball (7, 8), it is referred to as a regular ball in view of the fact that the ball 6 is attracted to a regular position. The surplus ball is not a ball having a defect but a ball that is unfortunately attached to a place other than the suction hole 13 of the ball alignment surface 5.

Since the regular ball 6 is vacuum-sucked in the chamfered suction hole 13, the regular ball 6 is illustrated as being slightly recessed from the ball alignment surface 5. Since the surplus balls (7 and 8) are not attracted to the regular position but are attached on the ball alignment surface 5, the end of the surplus ball 7 is exposed below the regular ball 6. Yes. The surplus balls may adhere to the regular balls, a plurality of balls may gather, or may adhere in various modes.

  The air cylinder 9 can set two stop positions of the slider portion 10. One stop position is a position where the light shielding plate 12 is retracted. The other stop position is a position for measuring the surplus ball 7. This position is set by teaching. The retracted position is a position where the light shielding plate 12 is retracted so as not to interfere with the operation when the ball is mounted on the circuit board. Therefore, the position accuracy is not particularly required. Note that the air cylinder 9 may be a single-action air cylinder in which one stop position can be set because the accuracy of the retracted position may be rough.

When detecting the surplus ball 8, the height of the surplus ball 8 exposed from the light shielding plate is 75 microns when the ball diameter is 0.3 mm, and when the ball diameter is 0.2 mm and 0.1 mm. The exposed heights are 50 microns and 25 microns, respectively, and both are as small as about 1/4 compared with the ball diameter.

  For this reason, when the ball diameter is reduced, it is necessary to accurately control the height position of the light shielding plate 12. Specifically, the height position accuracy of the lower end 12a of the light shielding plate is preferably 10 microns or less, and more preferably 5 microns or less. The vibration due to the vibration of the head or the optical sensor is preferably 2 microns or less, and in addition, the level of the ball alignment surface 5 is preferably 5 microns or less. The level of the ball alignment surface 5 is the height difference of the ball alignment surface 5 in the moving direction of the ball mounting head.

  The dimensions, the number, and the positions of the balls 6 shown in FIGS. 1 and 2 are shown for explaining the detection of surplus balls and are different from the actual examples. An example of a circuit board on which solder balls are mounted is shown in FIG. The circuit board 17 is composed of six module boards 18, and is cut for each module board 18 for practical use. When the circuit board 17 is a semiconductor wafer, the module board 18 is a semiconductor integrated circuit chip, and the solder balls 6 are mounted on the bumps of the semiconductor integrated circuit chip.

FIG. 6 is an example of a BGA having 18 × 18 balls, and is an enlarged view of the solder balls mounted on the back surface of the module substrate 18. The circuit board on which the solder balls are mounted has dimples (recesses) by removing the polyimide tape and resist extending to the ball mounting locations. The external connection terminal is exposed at the bottom of the dimple. Further, flux is applied to the dimples. It is important that the ball is fully pushed into the flux until it reaches the bottom of the dimple in order to prevent the remaining load when the ball is mounted and to prevent the ball from moving after the ball is mounted.

The solder balls are reflowed to form solder bumps on the external connection terminals of the module substrate. A semiconductor integrated circuit chip is mounted face down on the surface of the module substrate 18 and connected to solder bumps via via holes by multilayer wiring. The module board is mounted on another electronic circuit board using solder bumps. Such a mounting method is widely used in CSP (Chip Size Package).

  Various ball alignment methods and flux coating methods have been developed for ball mounting devices. Examples of the ball alignment method include a fluid immersion adsorption method, a jump adsorption method, a reverse adsorption method, and a transfer method. Further, as a flux application method, there are a screen printing method, a dipping method, a pin transfer method, and the like.

FIG. 7 shows an example of a ball mounting apparatus that combines a jumping suction method in which a ball jumped by vibration is sucked into a suction hole and a pin transfer method in which a flux is applied to a transfer pin tip and transferred to a circuit board. The outline will be described below.

  Reference numeral 15 is a circuit board, 19 is a loader, 20 is a transfer head, 21 is a transfer unit, 23 is a ball mounting head, 24 is a ball supply unit, 25 is an illumination device for inspecting the ball adsorption state, and 26 is a light emitting unit 1. An optical sensor comprising the light receiving unit 2, 27 is a ball mounting inspection device for inspecting whether or not the ball is correctly mounted on the circuit board, 28 is an unloader, 29 is a repair unit, 30 is a board guide for guiding the movement of the circuit board, 31 is a Y-axis slider, 32 is a transfer pin cleaning device that heats and cleans the tip of the transfer pin, 33 is an X-axis slider, 34 is a CCD camera, 35 is a squeegee, 36 is a parts feeder, 37 is a ball vibrating container, Reference numeral 38 denotes a ball amount detection sensor.

  A typical process for mounting a ball includes a process of adsorbing the ball to the ball mounting head, a process of inspecting that the ball is adsorbed in all the adsorbing holes, and a surplus ball detecting process of detecting adhering surplus balls. And a step of mounting the ball on the substrate coated with the flux, a step of inspecting the ball mounted on the substrate, and a step of reflowing the ball mounted on the substrate. FIG. 8 is a flowchart of the ball mounting method, which mainly describes an operation for detecting whether or not an excess ball is attached. With reference to FIGS. 7 and 8, the ball mounting method will be described focusing on the excess ball detection step.

  In step 1 (S1), the ball vibrating container 37 is vibrated to cause the ball to jump 1 to 2 cm, and the ball mounting head 23 is placed in the jumping ball group, and the suction hole 13 provided in the ball alignment surface 5 is placed. The ball 6 is vacuum-sucked. A ball amount detection sensor 38 is attached to the ball vibrating container 37. The ball is supplied from the parts feeder 36 by a signal from the ball amount detection sensor 38 so that the ball occupies approximately half of the bottom surface of the ball vibrating container 37. When the balls come in contact with each other on the bottom surface of the ball vibrating container 37 and interfere with each other, the balls do not jump normally. It is desirable that the amount of the ball take a maximum value so that the balls do not interfere with each other in order to shorten the time for the balls to be attracted to all the suction holes. Since the ball vibrating container 37 is set horizontally and vibrates up and down by a vibrator provided below the ball vibrating container 37, the balls are distributed almost uniformly on the bottom surface of the ball vibrating container 37.

  After the ball is attracted to the ball mounting head 23 for a predetermined time, the ball mounting head 23 is moved on the illumination device 25 along the Y-axis slider 31 to inspect whether or not the ball is attracted to all the suction holes. To do. The ball mounting head is configured to detect light leaking from the suction holes by a plurality of photosensors disposed therein. If no light leakage is detected, the process proceeds to a surplus ball detection process.

  In step 2 (S2), the air cylinder 9 attached to the side surface of the lower portion (ball alignment portion 4) of the ball mounting head 23 is operated to move the light shielding plate 12 to the measurement position (light shielding plate lower end 12a shown in FIG. 4). Move. The light flux 16 is shielded from the lower end 12a of the light shielding plate. The surplus ball 7 blocks light in the shaded area in FIG. When the light shielding plate 12 is provided, the surplus balls 7 attached to the ball alignment surface 5 have the same light shielding area regardless of the attachment position.

When the ball diameter is 0.2 mm or less and the position accuracy of the light shielding plate 12 is insufficient in the air cylinder, an AC servo motor and a ball screw may be used as the driving device.

In step 3 (S3), the ball mounting head is moved horizontally along the Y-axis slider 31. The moving speed is 40 cm / second. The light beam 16 of the laser light emitted from the light emitting unit 1 is set so as to pass through the approximate center of the surplus ball 7. Since the ball alignment surface moves relative to the light emitting portion, an output voltage generated by photoelectric conversion from the light receiving portion is generated in substantially the same manner as the light beam 16 scans the ball alignment surface.

  In step 4 (S4), the output voltage from the light receiving unit 2 is sampled. The sampling time (speed) is 500 μs. Sampling data (output from the light receiving unit) is 0.5 volt at maximum. This sample value is stored in the memory. When the ball diameter is as large as 0.3 mm, it is possible to detect the presence / absence of an excess ball without error by using the output voltage shielded by the light shielding plate.

In step 5 (S5), background noise is removed to further improve the detection accuracy. Eight consecutive sample values are set as calculation targets. First, an average value A of seven sample values is calculated to obtain a background (background) noise value. Next, a difference value D is obtained by subtracting the average value A from the eighth sample value B. The above example is an example. For example, the average value B may be the average of the seventh sample value and the eighth sample value, and similarly, the difference value D may be calculated by subtracting A from B. Further, the average value B may be an average of seven sample values.

In step 6 (S6), it is determined whether or not there is a surplus ball. The difference value D is used as data for detecting the presence / absence of surplus balls, and the presence / absence of surplus balls is determined by comparing the data with a threshold (0.15 volts) determined separately. In some cases, a plurality of difference values D are calculated for one surplus ball and it is determined that there are a plurality of surplus balls. If the difference value D exceeds the threshold value once or more, it is determined that there is an excess ball, so no problem occurs.

When it is determined that the absolute value of the difference value D is smaller than the threshold value and there is no surplus ball, the subsequent ball mounting operation is continued. When it is determined that there are surplus balls, the vacuum of the ball mounting head is broken, and all the adsorbed balls are discarded. And it returns to step 1.

In step 7 (S7), the ball is mounted on the circuit board, and the light shielding plate 12 is moved to the retracted position by the air cylinder 9 so that the light shielding plate 12 does not get in the way.

In step 8 (S8), the ball is mounted on the circuit board 15. A circuit board is used in which the tip of the pin transfer head 20 is dipped in the flux formed in the transfer unit 21 and the flux attached to the pins is transferred. The circuit board is moved to the ball mounting position. The ball mounting head 23 adsorbing the ball is lowered, and the ball is pushed into the flux until the ball contacts the circuit board. Thereafter, the vacuum of the ball mounting head is broken to raise the ball mounting head, the ball is separated from the ball mounting head, and transferred to the circuit board. In addition, since the weight of the ball mounting head is counterbalanced by a spring, even if the ball is pressed against the circuit board, a large force is not applied to the ball. However, the ball mounting head is pressed and fixed to the fixing member by an air cylinder during movement, and is not oscillated due to disturbance caused by movement.

The CCD image processing apparatus inspects whether there is any missing ball on the circuit board on which the ball is mounted. If it is determined that the ball is correctly mounted, the circuit board is sent to the unloader and further sent to the reflow process. If it is determined that the ball mounting is insufficient, the repair device adds the ball. Further, the ball remaining on the ball mounting head may be detected in the same manner as the method of detecting the surplus ball with the optical sensor.

In the flowchart of FIG. 8, it is described that the determination step is omitted after step 8 and the process returns to step 1. However, when the ball mounting is not continued, the process ends. The next process after step 8 is a reflow process.

  Next, the output voltage photoelectrically converted by the light receiving unit will be described. When the ball mounting head 23 is moved between the light emitting unit 1 and the light receiving unit 2, the output voltage at the point where there is no ball on the ball alignment surface is 0 volts. Where there is a light-shielding object such as a regular ball or an extra ball, the amount of light received decreases, so the output voltage of the sensor decreases and becomes negative.

FIG. 9 is a graph obtained by sampling the voltage photoelectrically converted by the light receiving unit and setting the vertical axis and time the horizontal axis. The ball alignment surface 5 used for the measurement corresponds to the circuit board shown in FIG. In the ball alignment surface 5, a group of suction holes 36 are formed corresponding to the module substrate 18. FIG. 9 shows a measurement in which balls are aligned in all the suction holes and one surplus ball is attached to each of the suction hole groups. These module boards correspond to a set of balls that differ in the number of balls and the number of balls shown in FIGS. The number of surplus balls attached to the module substrate is one. The position of the surplus ball is not set at a specific position. In FIG. 9 (a) and FIG. 9 (b), the position of the surplus ball does not accurately correspond to the experimental error.

FIG. 9A shows the output of the light receiving unit when the light shielding plate 12 is not used. The wide drop portion 39 of the output voltage does not show a waveform in which one row and one row of the balls are soundly separated, but corresponds to a set of regular balls (module 18). The narrow drop 40 of the output voltage is a voltage drop caused by excess balls. In addition, the area 42 where the ball is not mounted between the module boards has nothing to shield, so the voltage is 0 volts, which is the maximum.

FIG. 9B shows a case where the light shielding plate 12 is provided. The light shielding plate 12 removes the change in the output voltage caused by the regular ball. However, the background noise 41 resulting from the vibration of the apparatus is superimposed on the light shielding effect by the surplus balls. The descending portion 40 with a narrow output voltage corresponds to light shielding by surplus balls. When the moving speed of the ball mounting head is 10 cm, the narrow descending portion 40 is a set of 5 to 6 sampling voltages.

As described above, by using the light shielding plate 12, fluctuation due to the presence or absence of a normal ball is deleted from the output voltage photoelectrically converted by the light receiving unit, so that the output voltage caused by the surplus ball is detected with a small background noise. Can do. However, if the ball alignment surface is inclined with respect to the ball mounting head feed direction, the background noise level drifts even if the light shielding plate 12 is provided.

A generalized method for removing background noise drift, which is one of the present invention, will be described.
Let the number of samples sampled in order be (n + 1). The average value A is obtained by averaging the n-th sample data. Next, m (an integer equal to or less than n) average values B including the (n + 1) th sample data are obtained. A difference value D between the average value B and the average value A is obtained. The difference value D corresponds to an output voltage drop due to the extra ball shielding the laser beam. If there is no surplus ball, D = 0 volts, but background noise is superimposed. The background noise level drifts due to various factors. However, it can be almost eliminated by subtracting the average value.

  In this way, by adopting the above-mentioned multipoint average value method, it is possible to eliminate the drift of the output voltage from the light receiving element, so that the horizontality of the mounting hole mounting head and the horizontal movement of the ball mounting head can be set with high accuracy. Is no longer required. Specifically, even if the light shielding plate lower end 12a moves 0.2 mm with respect to the light beam 16 at the beginning and end of the excess ball detection, the drift of the output voltage can be eliminated by the multipoint averaging method. On the other hand, when the ball diameter is reduced to 100 μm or less, the diameter of the light beam 16 may be reduced to about 0.3 mm to improve the sensitivity.

A difference value curve is created by connecting the difference values D in time series. Since this curve is subjected to the multipoint average value processing, it is similar to FIG. 9B, but has a uniform voltage waveform. By passing the voltage of this difference value curve through the pulse shaping circuit, the voltage waveform corresponding to the surplus ball is shaped into a rectangular wave pulse, and the background noise can be cut. The presence / absence of a surplus ball is determined using the pulse-shaped difference value curve.

The number of times of sampling the output voltage that has decreased due to light shielding by one surplus ball depends on the moving speed of the ball mounting head and the sampling speed, but it must be one or more times. Usually, it is set to 1 or more times, and more preferably 2 to 4 times. If the number exceeds 10 times, the accuracy is improved, but it is not preferable because it takes too much measurement time or the sensitivity of the measuring device needs to be increased.

In order to shorten the time required for detecting the presence or absence of excess balls, it is preferable to arrange a plurality of optical sensors each including a light emitting unit and a light receiving unit. In that case, the photosensors are arranged so that the phases to be sampled are different, or are sampled at a timing when the phase difference is appropriate. When a plurality of optical sensors are installed, the sampled data is rearranged in order of phase. The number of sets of optical sensors is preferably 2-4. With 5 or more sets, the effect diminishes.

In the present invention, it is apparent in principle that the balls are not limited to solder balls or conductive balls. Also, the ball diameter is difficult to detect when the ball diameter is smaller than 20 microns. A diameter greater than 30 microns is preferred. In addition to the laser diode, the light emitting unit may be a laser or an LED, and the light receiving unit may be a photodiode or a phototransistor.

As for the ball alignment, there is a method of arranging by transfer other than the adsorption to the adsorption hole, and it is obvious that the arrangement method is not limited in principle. In this case, since the ball is transferred onto the circuit board, the ball is temporarily fixed on the ball alignment surface or the board. The positional relationship of the light shielding plate 12 described above is inverted upside down. In the present invention, the light shielding member has been described as a light shielding plate, but the shape is not limited to a plate shape, and may be a shape such as a cylinder or a prism.

In the present invention, the ball detection method and the ball detection device are not necessarily linked. That is, the ball detection method of the ball detection apparatus of the present invention may use the ball detection method of the present invention, or may use another ball detection method. Similarly, the apparatus for performing the ball detection method of the present invention is not limited to the ball mounting apparatus of the present invention.

It is a front view of an optical sensor and a ball alignment part. It is a bottom view of a photo sensor and a ball alignment part. It is the cross section which expanded a part of ball alignment surface. It is a front view of a slit board. It is a top view of a circuit board. It is a figure of a module board. It is a top view of a ball mounting apparatus. It is a flowchart centering on a ball | bowl detection process. It is a figure of the output-time curve of a light-receiving part, (a) without a light-shielding plate, (b) is a figure at the time of arranging a light-shielding plate.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Light-emitting part, 2 ... Light-receiving part, 3 ... Slit board, 4 ... Ball alignment part, 5 ... Ball alignment surface, 6 ... Regular ball, 7, 8 ... Excess ball, 9 ... air cylinder, 10 ... slide part, 11 ... fixing member, 12 ... light shielding plate, 12a ... lower end of light shielding plate, 13 ... suction hole, 14 ... Chamfering, 15 ... slit, 16 ... luminous flux, 17 ... circuit board, 18 ... module board, 19 ... loader, 20 ... pin transfer head, 21 ... transfer unit, 23 ... Ball mounting head, 24 ... Ball supply unit, 25 ... Illumination device, 26 ... Optical sensor, 27 ... Ball mounting inspection device, 28 ... Unloader, 29 ... Repair unit 30 ... substrate guide, 31 ... Y-axis slider, 32 ... Axis slider, 33 ... Slider, 34 ... CCD camera, 35 ... Squeegee, 36 ... Parts feeder, 37 ... Ball vibration container, 38 ... Ball quantity detection sensor, 39 ... Output voltage drop due to normal balls, 40 ... Output voltage drop due to surplus balls, 41 ... Output voltage background noise, 42 ... Gap between modules

Claims (4)

The surplus balls attached on the ball alignment surface of the ball alignment portion of the ball mounting head, the end protruding below the regular ball opened on the ball alignment surface and adsorbed by the chamfered suction holes In a ball detection device that detects the presence or absence of an excess ball with a light sensor comprising a light emitting part and a light receiving part,
A light-shielding plate that is fixed to a slide portion of an air cylinder that is fixed to a side surface of the ball alignment portion via a fixing member, and is movable up and down;
A Y-axis slider for moving the ball mounting head horizontally;
The light emitting unit and the light receiving unit that do not move in conjunction with the movement of the ball mounting head, the light emitting unit incorporating the laser element, the phototransistor, and the slit 15 being processed at the center. The light receiving unit to which the slit plate is externally attached;
A ball detector comprising: The surplus balls attached on the ball alignment surface of the ball alignment portion of the ball mounting head, the end protruding below the regular ball opened on the ball alignment surface and adsorbed by the chamfered suction holes In the ball detection method for detecting the presence or absence of an excess ball with an optical sensor comprising a light emitting part and a light receiving part,
The air cylinder attached to the side surface of the ball alignment portion is operated to move the light shielding plate to a measurement position where the regular ball is above the lower end of the light shielding plate and a part of the excess ball is exposed from the lower end of the light shielding plate. Process,
The light emitting unit and the light receiving unit that do not move in conjunction with the movement of the ball mounting head, and set so that the light beam of the laser light emitted from the light emitting unit passes through the approximate center of the surplus ball. Process,
Moving the ball mounting head horizontally by a Y-axis slider;
Sampling the output voltage from the light receiving unit;
Determining the presence or absence of the surplus balls from the output voltage;
And a step of moving the light shielding plate to a retracted position. In ball detection method for detecting by an optical sensor a surplus balls adhered to the ball alignment surface of the ball aligning portion ing from light emitting portion and a light receiving portion,
An operation of moving the light shielding plate to the ball measurement position;
An operation of moving between the light emitting unit and the light receiving unit so that the ball alignment unit on which the ball is mounted crosses the light emitted from the light emitting unit;
Sampling n output voltages and (n + 1) th output voltage from the light receiving unit;
an operation of calculating an average value A of n output voltages;
An operation of calculating an average value B of m output voltages including the (n + 1) th output voltage;
An operation of calculating a difference value D which is a difference between the average value B and the average value A;
An operation for comparing the difference value D with a predetermined threshold value to determine the presence or absence of the surplus ball;
A ball detection method comprising: an operation of retracting the light shielding plate from the ball measurement position. 4. The ball detection method according to claim 3, wherein m = 1.