Component mounting machine and method for detecting abnormal state of suction nozzle

Abstract

A component mounting machine 10 according to the present invention comprises a mounting head 32, a camera 40, and an abnormal state detection device 71. In the mounting head 32, a suction nozzle 35 that is made of an elastic material and that suctions a component 2 is held by a holder 34. The camera 40 captures an image of a nozzle tip 36 of the suction nozzle 35 through a transparent member 41 in a state in which the nozzle tip 36 is pressed against the transparent member 41. The abnormal state detection device 71 detects an abnormal state of the suction nozzle 35 on the basis of an image G2 of the shape of the nozzle tip 36 captured by the camera 40.

Description

Component mounter and method for detecting abnormal conditions of suction nozzles

[0001] The technology disclosed in this specification relates to a component mounter and a method for detecting abnormal conditions of suction nozzles.

[0002] Conventionally, a component mounter having a mounting head for mounting components on a substrate is known. In this type of component mounter, a suction nozzle is held by the mounting head. Then, the component is sucked by the suction nozzle held by the mounting head, and the sucked component is mounted on the substrate. As this type of conventional technology, for example, the one disclosed in Patent Document 1 (Japanese Unexamined Patent Application Publication No. 2008-205318) is known.

[0003] When sucking a component with the suction nozzle, the tip of the suction nozzle is brought into contact with the surface of the component. In order to make the tip of the suction nozzle adhere closely to the component, it is preferable that the tip of the suction nozzle be made of rubber. However, a rubber suction nozzle is more likely to deteriorate compared to a metallic suction nozzle. For example, due to hydrolysis or the like, rubber hardening or breakage may occur. And even if a component is sucked with a suction nozzle in which hardening or breakage has occurred, since the force for sucking the component is weakened, a suction error in which the component cannot be sucked from the component supply device in the correct posture or a mounting defect in which the component is displaced during component conveyance or mounting may occur. Note that it is difficult to judge rubber deterioration by human eyes and there is no method for detecting it.

[0004] Further, since the tip of the suction nozzle is made of rubber having viscosity, dust such as pieces of paper tape may adhere and soil it. In such a state, since the component cannot be sucked normally, there is a possibility that it may lead to a suction error or a mounting defect.

[0005] Therefore, this specification provides a technology capable of preventing suction errors and mounting defects caused by deterioration and soiling of the suction nozzle by preventing the use of a deteriorated or soiled suction nozzle.

[0006] This specification discloses a component mounting machine. The component mounting machine comprises a mounting head, a camera, and an abnormal condition detection device. The mounting head holds an elastic suction nozzle for picking up components in a holder. The camera presses the nozzle tip of the suction nozzle against a transparent member and images the nozzle tip through the transparent member. The abnormal condition detection device detects an abnormal condition of the suction nozzle based on the image of the shape of the nozzle tip captured by the camera. Therefore, with the above configuration, by not using a suction nozzle that has been detected to be in an abnormal condition for picking up components, suction errors and mounting defects caused by abnormalities in the suction nozzle can be prevented.

[0007] This specification also discloses a component mounting machine. The component mounting machine comprises a mounting head, a transparent member, a camera, and an abnormal condition detection device. The mounting head holds a suction nozzle made of an elastic material for picking up components in a holder. The nozzle tip of the suction nozzle is pressed against the transparent member. The camera images the nozzle tip through the transparent member while the nozzle tip is pressed against the transparent member. The abnormal condition detection device detects an abnormal condition of the suction nozzle based on the image of the shape of the nozzle tip captured by the camera.

[0008] Furthermore, this specification discloses a method for detecting abnormal conditions in a component mounting machine's suction nozzle. This abnormal condition detection method includes the steps of pressing the nozzle tip of an elastic material suction nozzle used for suctioning components against a transparent member, and detecting an abnormal condition in the suction nozzle based on the shape of the nozzle tip as viewed through the transparent member while the nozzle tip is pressed against the transparent member.

[0009] This is a schematic cross-sectional view showing the component mounting machine of Example 1. This is a cross-sectional view taken along line A-A in Figure 1. This is a bottom view showing the mounting head. This is a schematic cross-sectional view showing the parts camera. This is a block diagram showing the electrical configuration of the component mounting machine. This is a flowchart showing the method for detecting abnormal conditions of the suction nozzle. This is a flowchart showing the abnormal condition detection process. This is a flowchart showing the abnormal condition detection process. (a) to (c) are schematic diagrams showing images of the shape of the nozzle tip surface. This is a block diagram showing the configuration of the learning system 100 according to Example 2. This is a schematic cross-sectional view showing the parts camera of another embodiment.

[0010] (Embodiment 1) In the component mounting machine disclosed herein, the transparent member may be a transparent component that is adsorbed by a suction nozzle. With this configuration, the state in which the suction nozzle has adsorbed a component can be accurately simulated.

[0011] (Embodiment 2) In the component mounting machine disclosed herein, the camera is a parts camera that images the component attached to the suction nozzle from below, and the transparent member may be a glass portion positioned above the parts camera. With this configuration, the nozzle tip of the suction nozzle can be imaged using the parts camera.

[0012] (Embodiment 3) In the component mounting machine disclosed herein, the abnormal condition detection device may detect an abnormal condition of the suction nozzle based on the roundness and area of ​​the nozzle tip. With such a configuration, the abnormal condition of the suction nozzle can be detected with high accuracy.

[0013] (Embodiment 4) In the component mounting machine disclosed herein, the suction nozzle may attract the transparent member by generating negative pressure. The abnormal state detection device may further detect the abnormal state of the suction nozzle based on the negative pressure value when the transparent member is attracted. With such a configuration, the abnormal state of the suction nozzle can be detected with even greater accuracy.

[0014] (Embodiment 5) In the component mounting machine disclosed herein, the suction nozzle may attract the transparent member by generating negative pressure. The abnormal state detection device may detect an abnormal state of the suction nozzle based on the negative pressure value when the transparent member is attracted. An abnormal state of the suction nozzle can also be detected with this configuration.

[0015] (Mode 6) In the component mounting machine disclosed herein, the abnormal state detection device may determine whether the target suction nozzle is in an abnormal state using a trained model generated by using an image of the normal shape of the nozzle tip as training data. With such a configuration, the accuracy of detecting the abnormal state of the suction nozzle can be improved over time.

[0016] (Embodiment 7) In the component mounting machine disclosed herein, the abnormal state detection device may determine whether the target suction nozzle is in an abnormal state using a trained model generated with an image of the normal shape of the nozzle tip and the normal negative pressure value when the transparent member is adsorbed as training data. With such a configuration, the accuracy of detecting the abnormal state of the suction nozzle can be improved over time.

[0017] (Example 1) The component mounting machine 10 of this embodiment will be described below with reference to the drawings. As shown in Figures 1 and 2, the component mounting machine 10 is a device for mounting components 2 onto a substrate 1. Specific examples of components 2 include semiconductor packages such as QFP (Quad Flat Package) and BGA (Ball Grid Array), and chip components such as chip resistors and chip capacitors.

[0018] The component mounting machine 10 also includes a feeder attachment / detachment unit 11, an XY robot 12, a substrate transport device 13, a head unit 31, a parts camera 40, a control device 71, and a touch panel 72. The feeder attachment / detachment unit 11 is configured to allow multiple feeders 20 to be installed side by side. Each feeder 20 is for supplying components 2 for substrate mounting. Each feeder 20 is equipped with a component reel, and a tape containing multiple components 2 is wound around the component reel. The component mounting machine 10 takes the components 2 from each feeder 20 attached to the feeder attachment / detachment unit 11 and mounts them on the substrate 1.

[0019] The XY robot 12 moves the head unit 31 between the area above the feeder 20 and the area above the substrate 1 by moving the head unit 31 in the X and Y directions. The XY robot 12 consists of guide rails that guide the head unit 31, a movement mechanism that moves the head unit 31 along the guide rails, and motors that drive the movement mechanism. The head unit 31 moves through the space from above the feeder 20 to above the substrate 1 by the XY robot 12.

[0020] The substrate transport device 13 is a device that performs the tasks of transporting the substrate 1 to a work position P1 in the component mounting machine 10, positioning the substrate 1 before component mounting at the work position P1, and transporting the substrate 1 from the work position P1 after component mounting. The substrate transport device 13 in this embodiment can be configured, for example, with a pair of belt conveyors 14, a support device (not shown) attached to the belt conveyors 14 and supporting the substrate 1 from below, and a drive device (not shown) that drives the belt conveyors 14.

[0021] As shown in Figures 1 to 3, the head unit 31 is a movable unit that performs the operation of mounting components 2 onto the substrate 1, and is equipped with a mounting head 32. The mounting head 32 is attached to the lower side of the head unit 31 and is a device that mounts the held components 2 onto the substrate 1. The mounting head 32 is equipped with a rotary head 33, a plurality of holders 34, and a plurality of suction nozzles 35. Note that in Figure 1, only one suction nozzle 35 is shown for the sake of explanation. The rotary head 33 is a cylindrical member on which a plurality of holders 34 are arranged. Each holder 34 is a cylindrical member and is arranged at equal angular intervals with respect to the central axis of the rotary head 33. A suction nozzle 35, made of an elastic material such as rubber, is detachably held at the lower end of each holder 34. The suction nozzle 35 is a member that attracts components 2 by generating negative pressure. Note that Figure 3 shows a state in which four suction nozzles 35 are holding components 2. Furthermore, each suction nozzle 35 is a cylindrical member that forms a circular shape when viewed from below, and is arranged at equal angular intervals with respect to the central axis of the rotary head 33.

[0022] The rotary head 33 rotates in the direction of arrow F1 (see Figure 3) around its central axis. Each holder 34 (including the suction nozzle 35) rotates in the direction of arrow F2 (see Figure 3) around its central axis. Consequently, the part 2 held by the suction nozzle 35 also rotates in the direction of arrow F2. Furthermore, each holder 34 moves vertically (Z direction in Figures 1 and 2). Consequently, each suction nozzle 35 and the part 2 held by the suction nozzle 35 move up and down.

[0023] To mount component 2 onto substrate 1 using mounting head 32, first, the suction nozzle 35 is moved downward until it contacts component 2 contained in feeder 20. Next, component 2 is picked up by the suction nozzle 35, and the suction nozzle 35 is moved upward. Once the process of picking up component 2 with the suction nozzle 35 is complete, the XY robot 12 is driven to move mounting head 32 above substrate 1. Then, the suction nozzle 35 is lowered towards substrate 1 to mount component 2 onto substrate 1.

[0024] As shown in Figures 1 and 2, the head unit 31 is equipped with a mark camera 37. The mark camera 37 is mounted on the lower side of the head unit 31 near the mounting head 32 and is configured to be movable together with the mounting head 32. The mark camera 37 moves above the substrate 1 that has been transported to the work position P1 by the substrate transport device 13 and captures images of the marks (not shown) attached to the substrate 1. The mark camera 37 is configured using an image sensor such as a CCD (Charge Coupled Device) or CMOS (Complementary Metal Oxide Semiconductor).

[0025] The parts camera 40 is located between the feeder attachment / detachment unit 11 and the substrate transport device 13, and is positioned below the movement path of the mounting head 32. The parts camera 40 captures images of the parts 2 that are held in place by the suction nozzle 35 from below.

[0026] Furthermore, as shown in Figure 4, the parts camera 40 presses the nozzle tip surface 36 (nozzle tip) of the suction nozzle 35 against the glass part 41 (transparent member) and images the nozzle tip surface 36 from below through the glass part 41. The glass part 41 is located on the upper part of the parts camera 40 and is attracted by the suction nozzle 35. A negative pressure measuring sensor (not shown) provided on the mounting head 32 measures the pressure (negative pressure value) inside the suction nozzle 35 when the glass part 41 is attracted by the suction nozzle 35.

[0027] The parts camera 40 also includes an illumination unit 51 and an imaging unit 61. The illumination unit 51 irradiates light onto the nozzle tip surface 36 of the object to be imaged. The illumination unit 51 includes a housing 52, a connecting unit 53, a first illumination unit 54, a half mirror 56, and a second illumination unit 57. The internal space of the housing 52 gradually increases from the lower end to the upper end. The connecting unit 53 is a cylindrical member that connects the housing 52 and the imaging unit 61. The first illumination unit 54 has a plurality of LEDs 55. The half mirror 56 reflects the light irradiated horizontally from each LED 55 upward. The half mirror 56 also transmits light from above toward the imaging unit 61. The second illumination unit 57 has a plurality of LEDs 58. Each LED 58 irradiates light in a direction inclined from the optical axis 61a. Specifically, each LED 58 irradiates light toward the central part of the glass portion 41 (the part where the suction nozzle 35 adheres).

[0028] The imaging unit 61 generates an image G2 (see Figure 9) based on the received light. The imaging unit 61 includes an optical system such as a lens (not shown) and an image sensor such as a CCD or CMOS. When light emitted from the illumination units 54 and 57 and reflected from the nozzle tip surface 36 passes through the glass part 41 and the half mirror 56 and reaches the imaging unit 61, the imaging unit 61 receives this light and generates an image G2 of the shape of the nozzle tip surface 36.

[0029] As shown in Figure 5, the control device 71 consists of a computer comprising a CPU, memory, etc. The control device 71 is connected to the XY robot 12, substrate transport device 13, head unit 31, mark camera 37, parts camera 40, and touch panel 72 via a bus so as to be able to communicate alternately. Based on the production program stored in memory, the control device 71 controls the operation of each part of the component mounting machine 10 (XY robot 12, substrate transport device 13, head unit 31, mark camera 37, parts camera 40, and touch panel 72). The touch panel 72 is a display device that shows various information of the component mounting machine 10, as well as a user interface that receives instructions and information from the operator.

[0030] Next, the method for detecting abnormal conditions of the suction nozzle 35 of the component mounting machine 10 will be explained. First, the control device 71 of the component mounting machine 10 controls the suction of components 2 supplied from the feeder 20 and mounting them onto the substrate 1. Specifically, the control device 71 first drives the substrate transport device 13 to transport the substrate 1 to the work position P1 inside the component mounting machine 10. Then, the control device 71 drives the XY robot 12 to move the mounting head 32 to the component suction position (i.e., above the feeder 20) and picks up the component 2 with the suction nozzle 35 of the mounting head 32. Then, the control device 71 moves the mounting head 32 to the substrate 1 while passing it above the parts camera 40, and drives the parts camera 40 to take an image of the component 2 during this movement and check the condition of the component 2.

[0031] Next, the control device 71 drives the mark camera 37 to capture images of marks (not shown) on the substrate 1, and recognizes the position where the component 2 should be mounted based on the captured data. Once the component mounting position is recognized, the control device 71 lowers the suction nozzle 35 and releases the component 2 that it was holding. This operation mounts the component 2 onto the substrate 1, and a component-mounted substrate 3 (see Figure 2) is produced. The process of picking up the component 2 and mounting it onto the substrate 1 is repeated until the production number of component-mounted substrates 3 is reached.

[0032] Furthermore, the control device 71 performs a process to detect an abnormal state of the suction nozzle 35 each time the number of component mounting boards 3 produced reaches a predetermined number. Specifically, in step S10 shown in Figure 6, the control device 71 moves the mounting head 32 above the parts camera 40 and lowers the suction nozzle 35, which is in a predetermined position, and presses it against the glass portion 41 of the parts camera 40. The control device 71 then drives the suction nozzle 35, which is pressed against the glass portion 41, and generates negative pressure to adsorb the glass portion 41. As a result, the nozzle tip surface 36 of the suction nozzle 35 is pressed against the glass portion 41. In the following step S20, with the nozzle tip surface 36 pressed against the glass portion 41, the control device 71 causes the parts camera 40 to capture an image G2 (see Figure 9) of the shape of the nozzle tip surface 36.

[0033] In the subsequent step S30, the control device 71 detects an abnormal state of the suction nozzle 35 based on the shape of the nozzle tip surface 36 when viewed through the glass portion 41 (i.e., performs abnormal state detection processing). Specifically, the control device 71, which is an abnormal state detection device, detects an abnormal state of the suction nozzle 35 based on the image G2 captured in step S20.

[0034] More specifically, in step S30, the control device 71 executes the subroutines shown in Figures 7 and 8. The subroutines in Figures 7 and 8 consist of the processes from steps S110 to S220. In step S110, the control device 71 calculates the area of ​​the nozzle tip surface 36 based on the image G2 of the shape of the nozzle tip surface 36. In this embodiment, the area of ​​the nozzle tip surface 36 is calculated by counting the number of pixels in the region constituting the nozzle tip surface 36 in image G2.

[0035] In the following step S120, the control device 71 calculates the roundness of the nozzle tip surface 36 based on the image G2 of the shape of the nozzle tip surface 36. In this embodiment, the roundness of the nozzle tip surface 36 is calculated using conventional methods such as the least squares center method (LSC), minimum area center method (MZC), maximum inscribed circle center method (MIC), and minimum circumscribed circle center method (MCC). In the following step S130, the control device 71 measures the negative pressure value inside the adsorption nozzle 35 when the glass part 41 is adsorbed by driving the negative pressure measuring sensor described above.

[0036] In the subsequent step S140, the control device 71 determines whether the area of ​​the nozzle tip surface 36 calculated in step S110 has decreased from the normal state. If it is determined that the area of ​​the nozzle tip surface 36 has decreased, the control device 71 proceeds to the process in step S150. On the other hand, if it is determined that the area of ​​the nozzle tip surface 36 has not decreased (is the same), the control device 71 proceeds to the process in step S180.

[0037] In step S150, the control device 71 determines whether the roundness of the nozzle tip surface 36, calculated in step S120, has decreased from the normal state. If it is determined that the roundness of the nozzle tip surface 36 has decreased, the control device 71 proceeds to the process in step S160. In step S160, the control device 71 controls the touch panel 72 to inform the user that the suction nozzle 35 is in an abnormal state of being torn. Specifically, the touch panel 72 displays an image G2 (see Figure 9(b)) of the shape of the nozzle tip surface 36 with the tear A1 on the display screen 73. On the display screen 73, an image G1 (see Figure 9(b)) of the nozzle tip surface 36 in a normal state is displayed to the left of image G2. The touch panel 72 then displays the words, for example, "The nozzle is torn," on the display screen 73. This prompts the operator to replace the suction nozzle 35. After that, the control device 71 terminates this subroutine and ends the process here.

[0038] On the other hand, if it is determined that the roundness of the nozzle tip surface 36 has not decreased (has not changed), the control device 71 proceeds to step S170. In step S170, the control device 71 controls the touch panel 72 to inform the user that the suction nozzle 35 is in an abnormal state due to dust A2 (see Figure 9(c)) adhering to it. Specifically, the touch panel 72 displays an image G2 (see Figure 9(c)) of the shape of the nozzle tip surface 36 with dust A2 adhering to it on the display screen 73. On the display screen 73, an image G1 (see Figure 9(c)) of the nozzle tip surface 36 in a normal state is displayed to the left of image G2. The touch panel 72 then displays text on the display screen 73, for example, "There is dust on the nozzle." This prompts the operator to replace the suction nozzle 35. After that, the control device 71 terminates this subroutine and ends the processing here.

[0039] Furthermore, in step S180 described above, the control device 71 determines whether the roundness of the nozzle tip surface 36 calculated in step S120 has decreased from the normal state. If it is determined that the roundness of the nozzle tip surface 36 has decreased, the control device 71 proceeds to the process in step S190. In step S190, the control device 71 controls the touch panel 72 to inform the user that the suction nozzle 35 is in an abnormal state due to hardening. Specifically, the touch panel 72 displays an image G2 (see Figure 9(a)) of the shape of the hardened, elliptical nozzle tip surface 36 on the display screen 73. On the display screen 73, an image G1 (see Figure 9(a)) of the nozzle tip surface 36 in a normal state is displayed to the left of image G2. The touch panel 72 then displays the words, for example, "The nozzle has hardened," on the display screen 73. This prompts the operator to replace the suction nozzle 35. After that, the control device 71 terminates this subroutine and ends the process here.

[0040] On the other hand, if it is determined that the roundness of the nozzle tip surface 36 has not decreased, the control device 71 proceeds to the process of step S200 shown in Figure 8. In step S200, the control device 71 determines whether or not the negative pressure value measured in step S130 has decreased. If it is determined that the negative pressure value has decreased, the control device 71 proceeds to the process of step S210. In step S210, the control device 71 controls the touch panel 72 to indicate that the suction nozzle 35 is in an abnormal state of being torn. In this embodiment, the control device 71 performs the same process as in step S160. This prompts the operator to replace the suction nozzle 35. After that, the control device 71 terminates this subroutine and ends the process here.

[0041] In addition, when it is determined that the negative pressure value has not decreased, the control device 71 proceeds to the process of step S220. In step S220, the control device 71 performs control to cause the touch panel 72 to guide that the suction nozzle 35 is in a normal state. Specifically speaking, the touch panel 72 displays only the image G1 (see FIGS. 9(a) to 9(c)) of the normal state shape of the nozzle tip surface 36 on the display screen 73. Then, the touch panel 72 displays, for example, the characters "The nozzle is normal" on the display screen 73. After that, the control device 71 ends this subroutine and ends the processing here.

[0042] As described above, in the component mounter 10 of the present embodiment, the parts camera 40 images the nozzle tip surface 36 through the glass part 41 in a state where the nozzle tip surface 36 of the suction nozzle 35 is pressed against the glass part 41. The control device 71 detects an abnormal state of the suction nozzle 35 based on the image G2 of the shape of the nozzle tip surface 36 imaged by the parts camera 40. Thereby, the suction nozzle 35 in which an abnormal state is detected can be prevented from being used for sucking the component 2, and suction errors and mounting defects caused by the abnormality of the suction nozzle 35 can be prevented in advance.

[0043] In the component mounter 10 of the present embodiment, based on not only the shape (roundness and area) of the nozzle tip surface 36 but also the negative pressure value at the time of suction of the glass part 41, an abnormal state of the suction nozzle 35 is detected. Thereby, even when the state cannot be judged by the appearance of the nozzle tip surface 36, such as when invisible dirt or deterioration of the suction nozzle 35 occurs, the detection accuracy of the abnormal state can be improved by using the change in the negative pressure value as a judgment criterion.

[0044] (Embodiment 2) Next, the component mounter 10 of Embodiment 2 will be described. In this embodiment, mainly the configuration different from that of Embodiment 1 will be described. For the configuration common to Embodiment 1, detailed description will be omitted by attaching the same member numbers.

[0045] This embodiment differs from Embodiment 1 in its method for detecting abnormal conditions of the suction nozzle 35. Specifically, the control device 71 first generates a trained model for detecting abnormal conditions of the suction nozzle 35, and then uses the generated trained model to detect abnormal conditions of the suction nozzle 35. For example, the learning system 100 shown in Figure 10 is used in the process of generating the trained model. The learning system 100 includes an image input unit 102, a negative pressure value input unit 104, a trained model 106, and a state output unit 108. The image input unit 102 receives an image G1 of the nozzle tip surface 36 captured by the parts camera 40. The negative pressure value input unit 104 receives the negative pressure value inside the suction nozzle 35 measured by a negative pressure measuring sensor. The learning model 106 is a machine learning model (hidden layer) for determining whether the suction nozzle 35 is abnormal or not, using the image G input by the image input unit 102 and the negative pressure value input by the negative pressure value input unit 104. It has a multilayer structure consisting of multiple layers of nodes provided between the input units 102 and 104 (input layer) and the state output unit 108 (output layer). The state output unit 108 outputs the result output from the learning model 106 (i.e., whether the suction nozzle 35 is abnormal or not). For machine learning using the learning system 100, for example, a suction nozzle 36 in a normal state is used. Specifically, the control device 71 presses the nozzle tip surface 36 of the suction nozzle 36 (normal state) against the glass part 41 and controls the part camera 40 to capture an image G1 (see Figures 9(a) to (c)) of the normal shape of the nozzle tip surface 36. Furthermore, the control device 71 drives a negative pressure measuring sensor provided on the mounting head 32 to measure the normal negative pressure value inside the suction nozzle 35 when the glass part 41 is adsorbed. The control device 71 then stores the captured image G1 and the measured negative pressure value in memory. Next, the control device 71 uses the data stored in memory, specifically the image G2 of the normal shape of the nozzle tip 36 and the normal negative pressure value when the glass part 41 is adsorbed, as training data and inputs this training data into the learning system 100 to have the learning model 106 perform machine learning and generate a trained model. In other words, the learning model 106 becomes the trained model. The control device 71 then stores the generated trained model (learning model 106) in memory.

[0046] On the other hand, in the process of detecting an abnormal state of the suction nozzle 35, first, the control device 71 presses the nozzle tip surface 36 of the suction nozzle 36 (the suction nozzle 36 to be inspected) against the glass part 41 and controls the parts camera 40 to capture an image G1 of the shape of the nozzle tip surface 36. Also, when the parts camera 40 captures the image G1, the control device 71 measures the negative pressure value inside the suction nozzle 35 using a negative pressure measurement sensor provided on the mounting head 32. Next, the control device 71 uses a trained model to determine whether or not the target suction nozzle 35 is in an abnormal state. Specifically, the control device 71 inputs the captured image G to the image input unit 102 and the measured negative pressure value to the negative pressure value input unit 104. Then, the trained model 106 (trained model) uses the input image G and negative pressure value to determine whether or not the target suction nozzle 35 is in an abnormal state and outputs the determination result to the state output unit 108. The output value output to the status output unit 108 (for example, "1 (abnormal state)" or "0 (normal state)") determines whether the suction nozzle 35 is in an abnormal state or not.

[0047] As described above, in the component mounting machine 10 of this embodiment, the control device 71 determines the state of the suction nozzle 35 using a learned model (learned model 106) that has been learned using the image G and negative pressure value. By not using the suction nozzle 35 that has been determined to be in an abnormal state for picking up the component 2, it is possible to prevent suction errors and mounting defects caused by abnormalities in the suction nozzle 35.

[0048] Although the embodiments have been described above, the specific embodiments are not limited to the above embodiments. In the above Embodiments 1 and 2, the suction nozzle 35 adsorbed the glass part 41 with the nozzle tip surface 36 pressed against the glass part 41 of the parts camera 40, but the configuration is not limited to this. For example, in other embodiments, the suction nozzle 35 may adsorb the transparent part with the nozzle tip surface 36 pressed against a transparent part (such as a transparent member like a glass plate) of an inspection table installed at a position different from the parts camera 40. Also, as shown in FIG. 11, the suction nozzle 35 may adsorb the glass part 81 with the nozzle tip surface 36 pressed against a glass part 81 (transparent part), which is a transparent member different from the parts camera 40. Note that the glass part 81 has a rectangular plate shape, but may have other shapes such as a disc shape, an elliptical plate shape, or a block shape as long as it can press the entire nozzle tip surface 36. Also, the supply method of the glass part 81 is not particularly limited. For example, a feeder 20 storing a plurality of glass parts 81 may be prepared instead of the part 2, and the suction nozzle 35 may be made to adsorb the glass part 81 stored in the feeder 20. Also, the glass part 81 may be stored in a recess provided in a conveyor rail (not shown) of the substrate transfer device 13, and the suction nozzle 35 may be made to adsorb the glass part 81 stored in the recess. Further, a dedicated inspection jig made of glass may be supplied into the component mounter 10 by the substrate transfer device 13, and the suction nozzle 35 may be made to adsorb the inspection jig.

[0049] In the above Embodiment 1, the control device 71 detected the abnormal state of the suction nozzle 35 based on the roundness and area of the nozzle tip surface 36 and the negative pressure value at the time of adsorbing the glass part 41, but the configuration is not limited to this. For example, in other embodiments, the control device 71 may detect the abnormal state of the suction nozzle 35 based only on the roundness and area of the nozzle tip surface 36.

[0050] In the above embodiment 2, the control device 71 determined whether the target suction nozzle 35 was in an abnormal state using a trained model generated with image G1 of the normal shape of the nozzle tip surface 36 and the normal negative pressure value during suction of the glass part 41 as training data. However, the configuration is not limited to this. For example, in another embodiment, the control device 71 may determine whether the target suction nozzle 35 is in an abnormal state using a trained model generated with only image G2 of the normal shape of the nozzle tip surface 36 as training data. That is, it is not necessary to use the normal negative pressure value during suction of the glass part 41 as training data. Alternatively, image G2 when the nozzle tip surface 36 is in an abnormal state may be used as training data to generate a trained model (multi-output trained model) that determines the type of abnormality (for example, a torn suction nozzle 35, a dust adhering to the suction nozzle 35, etc.). By determining the type of abnormality using a trained model, it becomes easier for the operator to take appropriate action.

[0051] In the above embodiment 1, the abnormal state of the suction nozzle 35 was detected based on threshold judgments of roundness, area, and negative pressure value, and in the above embodiment 2, the abnormal state of the target suction nozzle 35 was detected using a trained model, but the system is not limited to these configurations. For example, in other embodiments, the control device 71 may detect the abnormal state of the suction nozzle 35 based on both the above threshold judgments and the trained model.

[0052] In the above embodiments 1 and 2, the parts camera 40 was used as a camera to image the nozzle tip surface 36 via the glass portion 41, but the configuration is not limited to this. For example, in other embodiments, a mark camera 37 or a camera provided separately from the parts camera 40 and the mark camera 37 may be used as a camera to image the nozzle tip surface 36.

[0053] In the above embodiment 1, the control device 71 performed control to notify the operator that the suction nozzle 35 was in an abnormal state, but the configuration is not limited to this. For example, in other embodiments, the control device 71 may perform control to stop the use of the suction nozzle 35 when it is in an abnormal state.

[0054] In the above embodiment 1, the status of the suction nozzle 35 was displayed on the display screen 73 of the touch panel 72 to notify the operator of the abnormal condition, but the configuration is not limited to this. For example, in other embodiments, the operator may be notified of the abnormal condition by outputting sound from an audio output means such as a speaker.

[0055] In the above embodiment 1, when it was determined that the suction nozzle 35 was in an abnormal state, text indicating the state of the suction nozzle 35 (for example, "The nozzle is torn," "There is dust on the nozzle," or "The nozzle is hardened") was displayed on the display screen 73 of the touch panel 72, but the configuration is not limited to this. For example, in other embodiments, when it was determined that the suction nozzle 35 was in an abnormal state, text such as "Please replace the nozzle" may be displayed on the display screen 73. In other words, the same text may be displayed when it is determined that the nozzle is in an abnormal state.

[0056] In the above embodiment 1, the control device 71 performed step S110 to calculate the area of ​​the nozzle tip surface 36, then step S120 to calculate the roundness of the nozzle tip surface 36, and after step S120, step S130 to measure the negative pressure value when the glass part 41 is adsorbed. However, the configuration is not limited to this. For example, in other embodiments, the order of steps S110, S120, and S130 may be changed as appropriate. Also, the process of step S110 may be performed immediately before step S140 to determine whether the area has decreased, the process of step S120 may be performed immediately before steps S150 and S180 to determine whether the roundness has decreased, and the process of step S130 may be performed immediately before step S200 to determine whether the negative pressure value has decreased.

[0057] Although specific examples of the present invention have been described in detail above, these are merely illustrative and do not limit the scope of the claims. The technologies described in the claims include various modifications and changes to the specific examples illustrated above. The technical elements described in this specification or drawings exhibit technical usefulness individually or in various combinations, and are not limited to the combinations described in the claims at the time of filing. Furthermore, the technologies illustrated in this specification or drawings can achieve multiple objectives simultaneously, and achieving even one of these objectives itself constitutes technical usefulness.

[0058] 2: Component 10: Component mounting machine 32: Mounting head 34: Holder 35: Suction nozzle 36: Nozzle tip surface as nozzle tip 40: Part camera as camera 41: Glass part as transparent material 71: Control device as abnormal state detection device 81: Glass part as transparent material and transparent component G1, G2: Image

Claims

1. A component mounting machine comprising: a mounting head in which a suction nozzle made of an elastic material for adsorbing components is held in a holder; a camera that images the nozzle tip of the suction nozzle through a transparent member while the nozzle tip of the suction nozzle is pressed against the transparent member; and an abnormal condition detection device that detects an abnormal condition of the suction nozzle based on the image of the shape of the nozzle tip captured by the camera.

2. The component mounting machine according to claim 1, wherein the transparent member is a transparent component that is adsorbed by the adsorption nozzle.

3. A component mounting machine comprising: a mounting head in which a suction nozzle made of an elastic material for adsorbing components is held in a holder; a transparent member against which the nozzle tip of the suction nozzle is pressed; a camera that images the nozzle tip through the transparent member while the nozzle tip is pressed against the transparent member; and an abnormal condition detection device that detects an abnormal condition of the suction nozzle based on the image of the shape of the nozzle tip captured by the camera.

4. The component mounting machine according to claim 3, wherein the camera is a parts camera that takes an image of the component attached to the suction nozzle from below, and the transparent member is a glass portion positioned above the parts camera.

5. The component mounting machine according to any one of claims 1 to 4, wherein the abnormal condition detection device detects an abnormal condition of the suction nozzle based on the roundness and area of ​​the nozzle tip.

6. The component mounting machine according to claim 5, wherein the suction nozzle attracts the transparent member by generating negative pressure, and the abnormal state detection device further detects an abnormal state of the suction nozzle based on the negative pressure value when the transparent member is attracted.

7. The component mounting machine according to any one of claims 1 to 4, wherein the suction nozzle attracts the transparent member by generating negative pressure, and the abnormal state detection device detects an abnormal state of the suction nozzle based on the negative pressure value when the transparent member is attracted.

8. The component mounting machine according to any one of claims 1 to 4, wherein the abnormal state detection device determines whether the target suction nozzle is in an abnormal state using a trained model generated by using an image of the normal shape of the nozzle tip as training data.

9. The component mounting machine according to any one of claims 1 to 4, wherein the abnormal state detection device determines whether the target suction nozzle is in an abnormal state using a trained model generated with an image of the normal shape of the nozzle tip and the normal negative pressure value when the transparent member is adsorbed as training data.

10. A method for detecting an abnormal state of a suction nozzle of a component mounting machine, comprising the steps of: pressing the nozzle tip of a suction nozzle made of an elastic material for adsorbing components against a transparent member; and detecting an abnormal state of the suction nozzle based on the shape of the nozzle tip when viewed through the transparent member while the nozzle tip is pressed against the transparent member.

Images