Methods for analyzing and / or maintaining manufacturing systems

Abstract

A method of factory analysis and / or maintenance, preferably comprising receiving factory information and / or associating defects with factory components, and optionally taking action and / or operating factory machines based on the defect association, the method preferably being associated with one or more manufacturing systems and / or elements thereof.

Description

[Technical Field]

[0001] The present invention relates generally to the field of manufacturing, and more particularly to novel and useful methods for analyzing and / or maintaining manufacturing systems.

[0002] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims the benefit of U.S. Provisional Application No. 63 / 228,495, filed August 2, 2021, and U.S. Provisional Application No. 63 / 306,375, filed February 3, 2022, each of which is incorporated by reference in its entirety. [Brief explanation of the drawings]

[0003] [Figure 1] FIG. 1 is a schematic diagram of one embodiment of the method. [Figure 2] Figure 2A is a schematic diagram of one embodiment of a system for carrying out the present method, and Figure 2B is a schematic diagram of an example manufacturing system. [Figure 3] Figures 3A and 3B are schematic diagrams of example elements of operating factory machinery, and Figures 3C and 3D are schematic diagrams of first and second examples, respectively, of associating defects with factory components. [Figure 4] Figure 4A is a schematic diagram of an example of renumbering of array location designations, and Figure 4B is a schematic diagram of an exemplary mapping between the various array location designations shown in Figure 4A. [Figure 5] FIG. 5 is a schematic diagram of an example of identifying program links based on temporal considerations. [Figure 6] 6A and 6B are schematic diagrams of first and second particular examples, respectively, of identifying program links based on product identifiers. DETAILED DESCRIPTION OF THE INVENTION

[0004] The following description of preferred embodiments of the invention is not intended to limit the invention to those preferred embodiments, but rather to enable any person skilled in the art to make and use the invention.

[0005] 1. Overview Factory analysis and / or maintenance method 100 preferably includes (e.g., as shown in FIG. 1 ) step S110 of receiving factory information and / or step S120 of associating defects with factory components, and may optionally include step S130 of taking action based on the defect association and / or step S105 of operating a factory machine. However, method 100 may additionally or alternatively include any other suitable elements.

[0006] The method 100 is preferably associated with (e.g., directed to) maintenance (e.g., repair and / or predictive maintenance) for one or more manufacturing systems (or elements thereof), as described in more detail below. Additionally, the method may optionally be performed (in part or in whole) by and / or in association with one or more computing systems, manufacturing systems, and / or any other suitable systems (e.g., the systems shown in FIG. 2A).

[0007] 2.Technical advantages Aspects of the present technology may provide one or more advantages over the prior art. In some instances, these advantages may facilitate the analysis and / or maintenance of one or more manufacturing systems or elements thereof.

[0008] First, in some embodiments, the techniques may enable identification and / or quantification (e.g., automatic identification and / or quantification) of potential links between factory components (and / or other manufacturing elements) and defective products. For example, by attributing responsibility for detected product defects to factory components that may have caused the defects (e.g., factory components that interacted or may have interacted with defective areas or elements of the product), factory components associated with numerous product defects may automatically accumulate a large portion of such responsibility, thereby enabling and / or facilitating detection of those factory components as potential sources of manufacturing system errors.

[0009] Second, in some embodiments, the present technology can enable the determination (e.g., automated determination) of one or more data structures (e.g., matrices such as probability matrices, graphs such as directed and / or undirected graphs with edge weights, etc.) that represent and / or indicate links (e.g., indicators of importance such as causal relationships, partial responsibility associations, etc.) between features of a manufactured product (e.g., PCB assemblies) and manufacturing elements (e.g., factory operations, programming parameters, machines and machine elements, etc.). One or more of these data structures can provide one or more indicators of importance associated with those links, thereby enabling sensitivity analysis related to the manufacturing elements represented therein. For example, this can enable the analysis and / or quantification of manufacturing elements that are most likely related to (e.g., causing) product defects, low production rates, and / or other undesirable manufacturing system behavior.

[0010] Third, in some embodiments, the technology may enable and / or facilitate providing manufacturing system information (e.g., related to defects, defect responsibility, defect rates, causality, etc.) to one or more users. For example, the technology may include one or more user interfaces that enable exploration of links between factory components (and / or other manufacturing elements) and products, product components, and / or defects thereof.

[0011] Fourth, in some embodiments, the present techniques can enable and / or facilitate the automated determination of maintenance recommendations and / or the execution of manufacturing system maintenance tasks. For example, the present techniques can include identifying one or more potentially problematic manufacturing elements (e.g., factory components that may require inspection and / or maintenance and / or may benefit from inspection and / or maintenance, programming parameters that may require change and / or may benefit from change, etc.) (e.g., based on links between the manufacturing elements and products, product components, and / or defects thereof). Based on such identification, the present techniques can optionally include presenting information indicative of such identification, operating to effect changes based on such identification (e.g., changing programming parameters, performing inspection and / or maintenance, etc.), and / or taking any other suitable action.

[0012] Fifth, in some embodiments, the present technology can enable the identification and / or quantification (e.g., automatic identification and / or quantification) of defect rates and / or other production rates. For example, information reported by (and / or associated with) one or more manufacturing systems may not include successful production information (e.g., production of defect-free product), such as production numbers, rates, and / or timing information (but may instead include only information indicative of detected defects and / or their associated timing and / or rates), and in this example, embodiments of the present technology can function to identify such information related to successful production. As such, such information can be utilized to identify defect rates and / or other production rates.

[0013] Sixth, in some embodiments, the present technology may enable standardization (e.g., automatic standardization) of information received from (and / or associated with) one or more manufacturing systems. In some examples, this information may include inconsistent program names (e.g., assembly and / or inspection programs associated with the manufacture of a particular product), product location identifiers (e.g., array location indexes), and / or other information. In some such embodiments, this standardization may enable and / or facilitate further analysis and / or presentation of the relevant information (and / or derivatives thereof).

[0014] However, additional or alternative benefits may be provided by the systems and / or methods described herein.

[0015] 3. Manufacturing System As noted above, the method may be associated with one or more manufacturing systems or elements thereof (e.g., as shown in FIG. 2B ). For example, in some embodiments, the method may be associated with one or more factories, manufacturing lines, and / or other manufacturing systems (and / or elements thereof) for printed circuit board (PCB) assembly (e.g., one or more surface mount technology (SMT) PCB manufacturing lines including multiple assembly and inspection machines).

[0016] In some such embodiments, the factory may include one or more of: PCB assembly machines, e.g., pick-and-place (PNP) machines, solder printers, ovens, and / or elements thereof; PCB inspection machines, e.g., optical and / or X-ray inspection tools (e.g., automated optical inspection (AOI) and / or automated X-ray inspection (AXI) machines), solder paste inspection (SPI) tools, in-circuit test (ICT) tools; and / or any other suitable entities. For example, PCB assembler elements may include one or more PNP elements, such as a placement head, a nozzle holder on the placement head, a nozzle attached to the nozzle holder, a camera (e.g., a camera associated with the nozzle holder that can be used to detect component placement on the nozzle holder; a separate camera such as a "flying" camera associated with the placement head that can be used to detect panel placement, warpage, and / or other attributes), a feeder, a component reel, a vacuum pump and / or a feed; solder printer elements, such as a stencil, a squeegee, a camera (e.g., to inspect board position and / or align board position with stencil position); and / or any other suitable element.

[0017] For example, an SMT line can include one or more modules (typically multiple modules). Each module can hold multiple feeders (with the same or different components loaded into each feeder, or a subset thereof). Each module preferably has a placement head with multiple nozzle holders, each capable of holding one of multiple different nozzles (e.g., each nozzle having a particular size and shape). In a specific example, an SMT line (e.g., with Fuji NXT modules) includes 18 modules, each with 40 feeders, 24 nozzle holders, and more than 24 nozzles (e.g., 40-100 nozzles, about 60 nozzles, etc.).

[0018] Thus, for any particular component defect, there may be many (e.g., equal to the product of the number of modules, feeders per module, nozzle holders per module, and the number of nozzles) distinct placement paths, each describing the path along which the component is fed from a feeder, picked up by a nozzle held in a nozzle holder, and then placed on a substrate; any of these system components may potentially be involved in the cause of the defect; and there may also be a variety of different sensors involved in the placement of the component, and thus potentially involved in the cause of the defect. In a typical manufacturing system, defect information (e.g., identified using an inspection device, as described in more detail below) does not indicate which of these placement paths (or a subset, such as any system components within a placement path; e.g., which nozzles, nozzle holders, and / or feeders were used) was used to place the component, nor which system components (or types of system components, such as nozzles versus feeders) were responsible for causing the defect.

[0019] Furthermore, factory operations are typically controlled by numerous programming parameters, some of which may be specific to particular component types, assembly tools, placement paths, and / or other aspects. Thus, for any particular component defect, the defect may be attributable to one or more programming parameters rather than (or in addition to) attributable to one or more system components.

[0020] For example, programming parameters can include PNP parameters, printer parameters, oven parameters, and / or any other suitable parameters. PNP parameters, in some examples, can include the expected dimensions and / or shape of the component, the location to pick up the component, the vacuum pressure used to hold the component, the air pressure used to release the component onto the substrate, the pick-up height, the drop height above the substrate, the motion profile used, the wait time (e.g., wait N milliseconds before picking up from the reel to compensate for slow feeder motion; wait M milliseconds before dropping the component onto the substrate to allow lateral motion to stop, etc.), the force used to press the chip into the solder paste, etc. Printer parameters, in some examples, can include the force applied to the squeegee to squeeze the paste through the stencil, the cleaning cycle interval (e.g., cleaning every 5 panels, cleaning every 10 panels, etc.), the motion profile and / or speed used to lower the stencil onto the panel, the motion profile and / or speed used to lift the stencil off the substrate after paste application, the speed at which the squeegee moves, the speed at which the squeegee retracts to the starting point, the amount of solder paste to apply onto the stencil, etc. The oven parameters may include, in some examples, the temperature profile of the oven, the velocity of the substrate through the oven, the dwell time at each temperature, etc. However, the programming parameters may additionally or alternatively include any other suitable parameters.

[0021] In a typical PCB factory (e.g., SMT factory) operation, manufacturing defects may be detected during production inspection by automated inspection machines such as automated optical inspection (AOI), automated x-ray inspection (AXI), solder paste inspection (SPI), in-circuit test (ICT), etc. Additionally or alternatively, defects may be detected by manual inspection processes and / or any other suitable inspection tools and / or processes. These inspection machines typically take measurements of the product (e.g., a panelized array of one or more circuit boards that are typically processed as a single unit) and use those measurements to detect manufacturing defects. For example, such manufacturing defects may include missing components, incorrect components (e.g., the wrong type of component placed in a location intended for a different component), incorrect component orientation (e.g., rotated, "tombstoned," inverted, or other incorrect polarity), misaligned components (e.g., laterally displaced from their proper location by more than a threshold, too close to another component or trace, etc.), electrical defects (e.g., open circuits, short circuits, out-of-tolerance resistance such as high or low resistance, out-of-tolerance capacitance such as high or low capacitance, analog measurements of incorrect circuit node voltage values, etc.), solder defects (e.g., solder bridges, improperly formed solder joints, too much or too little solder paste, solder paste not deposited in the correct location, incorrect solder paste dimensions, e.g., solder paste width, length, and / or height outside of range, etc.), OCR failures, etc.

[0022] In a typical factory operation, these manufacturing defects are associated with (e.g., labeled and identified based on) one or more location fields, such as a defect type and / or array location, a reference designator, and / or a pin number. For example, each defect may be labeled with a defect type, an array location, a reference designator, and optionally a pin number. The array location preferably designates a circuit board within an array of circuit boards (e.g., top right, bottom left, location #2, etc.). The reference designator preferably designates a defective component (within the particular circuit board designated by the array location), such as C16, U1, R5, etc. The pin number may designate a particular component pin (of the component designated by the reference designator), such as pins 1, 2, 3, 4, etc., but may alternatively be omitted in some or all cases (e.g., for defects not associated with a particular pin). The defect type preferably designates the type of identified manufacturing defect, more preferably using one of a standardized set of defect type labels (e.g., “MISSING,” “OPEN,” “TOMBSTONE,” etc.). However, the defects may additionally or alternatively be associated with any other suitable information.

[0023] However, the manufacturing system may additionally or alternatively include systems for any other suitable manufacturing and / or other production tasks and / or may include any other suitable manufacturing and / or inspection tools.

[0024] 4. Method 4.1 Receiving factory information Step S110 of receiving factory information preferably functions to receive information regarding defects detected in factory products (e.g., defects as described above with respect to the manufacturing system), such as defects detected by an automated inspection machine. The factory information is preferably received (and / or stored) by one or more computing systems and / or analytical tools, but may additionally or alternatively be received by any other suitable entity.

[0025] The defect information received at S110 preferably includes a defect type (e.g., as described above) and one or more location fields for each defect (or a subset thereof). Additionally, the defect information may optionally include a timestamp (e.g., a manufacturing timestamp, an inspection timestamp, a defect detection timestamp, etc.) and / or other time-related information (e.g., a time window associated with the defect, such as a time window when the defective product was manufactured and / or a time window when the defect was identified), product information (e.g., a product identifier, a product serial number, a product type, CAD information, programming information, a program name, etc.), etc. However, the defect information may additionally or alternatively include any other suitable information.

[0026] Additionally, S110 may optionally be operable to receive other information regarding factory production (e.g., information regarding passed inspections, defect-free production counts, total production counts regardless of defect status, etc.), information regarding factory status (e.g., production and / or inspection tool status), information regarding machine programming (e.g., program identifiers running on various machines in the factory, preferably received in association with timing information indicating periods during which various programs have been running on the machines, etc.), and / or any other suitable information, which may be available and / or received by component, by product type, by production line, by production tool, by time window, and / or by any other suitable unit.

[0027] The factory information is preferably received in real time (or near real time) in response to the identification of the information (e.g., immediately or substantially immediately). In particular, the defect information is preferably received in response to the identification of the defect, and more preferably immediately (or substantially immediately) in response to the identification of the defect. However, the factory information may additionally or alternatively be received after a delay, in batches, periodically, sporadically, and / or at any other suitable time.

[0028] The factory information is preferably received from the manufacturing system (e.g., from an inspection machine and / or assembly machine, from one or more manufacturing system controllers such as a centralized controller), but may additionally or alternatively be received from any other suitable entity.

[0029] However, S110 may additionally or alternatively include receiving any other suitable factory information in any suitable manner. Further, although reference is made herein to receiving factory information, those skilled in the art will recognize that S110 may additionally or alternatively include receiving information from and / or associated with any other suitable manufacturing system(s).

[0030] 4.2 Associating defects with factory components Step S120 of associating defects with factory components preferably functions to enable identification of factory components (e.g., factory tools such as assembly tools and / or elements thereof, programming parameters, etc.) that may be causing the defects. S120 preferably includes step S125 of identifying potential links (e.g., links between defects and factory components) and may optionally include step S123 of determining data mapping and / or step S127 of determining defect rates. However, S120 may additionally or alternatively include any other suitable elements performed in any suitable manner.

[0031] 4.2.1 Identifying Potential Links The step S125 of identifying potential links preferably functions to identify potential links between a defect and its cause (e.g., as shown in Figures 3C and 3D). For example, S125 may function to identify (e.g., preferably automatically generate) links between defect occurrence, defect characteristics (e.g., elements of information received related to the defect, associated placement paths and / or aspects thereof, etc.), and / or factory components, e.g., particular machine components involved in operations affecting particular components (e.g., R3 of circuit 1, pin 2 of U1 of circuit 4, etc.), preferably components related to the defect.

[0032] In some embodiments, this may include attributing responsibility or potential responsibility for the defect to upstream factory components (e.g., any machine components and / or programming parameters, or a subset thereof, that affected the element associated with the defect before the defect was detected; any machine components and / or programming parameters that could have caused the defect, etc.) Thus, each defect may be tied not only to the associated product component (e.g., a component identified in the defect information received in S110), but also to specific factory components (e.g., nozzles, feeders, components, programming shapes, etc.) that interact with the associated component.

[0033] For example, for each defect (e.g., each defect represented by defect information received in S110), S125 may include noting links between the defect (and / or associated product components, such as, for example, U3 of Circuit 2, indicated by the defect information) and N machine components that interacted with the product component associated with the defect before the defect was identified (alternatively, noting links between the defect and only a subset of such machine components). Noting those links can function to attribute potential responsibility for the defect to those machine components.

[0034] For each machine component, noting the link can be performed by adding the value of the link to the sum of the fault links associated with the machine component. In a first example, this can involve assigning responsibility for the fault to each associated machine component (e.g., adding 1 to each of the associated fault link sums), while in a second example, this can involve assigning partial responsibility for the fault to each associated machine component (e.g., adding 1 / N to each of the associated fault link sums, as shown by way of example in FIG. 3C).

[0035] In some embodiments, if only n out of N machine components can (alternatively, are likely to) contribute to a defect of a flagged type, responsibility for 1 / n of the defect is assigned to each such machine component, and no responsibility is assigned to the remaining (Nn) machine components. For example, a wrong polarity defect type is typically not attributable to a nozzle or nozzle holder (e.g., because a nozzle would not accidentally flip a product component), and thus responsibility for a wrong polarity defect should not be assigned to any nozzle or nozzle holder (e.g., rather, responsibility for such a defect is potentially attributable only to the programming parameters and machine components involved in loading the product component), as shown by way of example in FIG. 3D .

[0036] In some embodiments, responsibility can be assigned unevenly, such as assigning greater responsibility to machine components that are more likely to have caused the defect (e.g., in such a way that the sum of the amounts added to all defect link sums equals 1), e.g., a first and second machine component can each be assigned 1 / 4 of the responsibility for the defect, and a third machine component can be assigned the remaining 1 / 2 of the responsibility. Those skilled in the art will recognize that the above values ​​can similarly be multiplied by a coefficient c (e.g., adding c to each associated defect link sum, adding c / N to each associated defect link sum, adding to the associated defect link sums so that the sum of the amounts added equals c, etc.), where c is any suitable value greater than or less than 1.

[0037] In some embodiments, different defects may be associated with different amounts of responsibility (e.g., more significant or critical defects may be associated with larger amounts of responsibility). For example, a first defect may be associated with A units of responsibility (e.g., 1 unit), and a second defect may be associated with B units of responsibility (e.g., 2 units). Thus, in a first particular example in which responsibility is allocated equally among potentially responsible factory components, N components associated with the first defect may each be assigned A / N units of responsibility (e.g., 1 / N units), and M components associated with the second defect may each be assigned B / M units of responsibility (e.g., 2 / M units). Alternatively, in a second particular example in which full responsibility for the defect is allocated to each potentially responsible factory component, N components associated with the first defect may each be assigned A units of responsibility (e.g., 1 unit), and M components associated with the second defect may each be assigned B units of responsibility (e.g., 2 units). However, different defects may additionally or alternatively be associated with different (and / or the same) amounts of responsibility in any other suitable manner, and / or responsibility may additionally or alternatively be assigned to factory components in any other suitable manner (e.g., partial responsibility may be assigned unevenly among various factory components).

[0038] Additionally or alternatively, S125 may include generating (and / or adding to) a web of links between product features (e.g., features on one or more circuit boards, such as individual SMT components or pins thereof) and upstream factory components (e.g., factory operations, programming attributes, machines and / or machine components, etc.). This web of links may be used to generate a matrix (e.g., a probability matrix) describing the strength of the links between product features and factory components. In this matrix, values ​​may represent a sensitivity analysis of which parameters and / or systems are the most frequent cause of defects over time (e.g., and therefore potentially worthy of maintenance and / or modification, as described in more detail below with respect to S130).

[0039] For example, S125 may include creating a list of the machine systems and parameters that were involved in placing a given reference designator on a given substrate at a given array location on a panel for all machines (or a subset thereof) upstream of the inspection machine, and then (e.g., when a defect is flagged) using the associated information to apportion responsibility for the defect (e.g., as described above) to all involved systems (or any suitable subset).

[0040] However, in some embodiments, the defect information received at S110 may not be sufficient to uniquely identify a particular product component associated with the defect (and / or identify a set of factory components that are responsible for the defect). For example, there may not be enough information to allow the array location reported with the defect to be mapped to a particular board (e.g., mapping between array locations reported by an inspection machine and an assembly machine), and thus the reference designator reported with the defect may represent any one of a number of different product components (e.g., one or a subset on each board, such as each board having such a component).

[0041] For example, array numbering may be inconsistent across different machines, such as machines from different machine vendors. In one example, all machines agree that there are 6 circuits in the array, such as arranged in a 2x3 rectangular array. However, a test machine may assign array positions in a snake fashion, ascending from 1 in the top right to 6 in the bottom left, while a placement system may assign array positions in a brick-laying fashion, ascending from 1 in the bottom left to 6 in the top right.

[0042] In some cases, it is possible to resolve this ambiguity (e.g., as described in more detail below with respect to S123). However, in other cases, the ambiguity may remain (e.g., S123 is not performed, no attempt to resolve the ambiguity is made, insufficient information is available to resolve the ambiguity, etc.). Thus, in some instances (e.g., when the ambiguity remains), the defects may be mapped to each of their potentially defective product components, and then, for each of those potentially defective product components, responsibility may be linked to the factory component as described above. In a first example, responsibility for the entire defect is allocated to each potentially defective product component, while in a second example, for B potentially defective product components (e.g., B different boards, each containing one such potentially defective product component), partial defect responsibility, such as 1 / B of the defect, may be allocated to each potentially defective product component (e.g., resulting in a reduction in the amount added to each associated defect linkage total by a factor of 1 / B compared to the approach described above in which there is no ambiguity regarding the identity of the defective product component).

[0043] S125 may additionally or alternatively include identifying time series associated with those defect links. In a first example, a factory component time series may include timestamps of linked defect detections (and optionally partial defect responsibility assignments). In a second example, a factory component time series may include a series of defect link sums, each accumulated over a short time window (equal to the time resolution of the time series). However, the time series may additionally or alternatively be identified in any other suitable manner.

[0044] However, S125 may additionally or alternatively include identifying any other suitable potential links in any suitable manner.

[0045] 4.2.2 Data Mapping Decisions The step of determining data mappings S123 may function to determine mappings between elements of data received from (and / or with respect to) the manufacturing system (e.g., as described in more detail above with respect to S110), e.g., mappings that may further enable and / or facilitate defect analysis (e.g., mappings that may further enable and / or facilitate identifying potential links as described herein with respect to S125, identifying defect rates as described herein with respect to S127, and / or analyzing such information as described herein with respect to S130, etc.). In some examples, S123 may include transforming the data (e.g., based on those mappings), e.g., converting it into a form that may further enable and / or facilitate this defect analysis.

[0046] In some embodiments, S123 may include determining a mapping between array position indications (e.g., array position numbers) associated with various manufacturing system machines (e.g., various assembly and / or inspection machines), as shown by way of example in FIGS. 4A and 4B . Such mapping may be determined based on CAD data (e.g., associated with products manufactured by the manufacturing system), based on machine-specific reference information for array position numbering, and / or based on any other suitable information. In some examples, such information may be determined in an automated manner (e.g., in real time or substantially real time), such as by requesting information from factory machines. Additionally, S123 may optionally include transforming the data (e.g., received in S110) based on those mappings, such as by standardizing all array positions in the data so that each array position indicator (e.g., number) consistently represents a particular array position (e.g., array position “1” always represents the top-left board, array position “2” always represents the top-right board, array position “6” always represents the bottom-right board, etc.). In these embodiments, such mapping and / or renumbering can serve to resolve ambiguities such as those described above with respect to S125, thereby facilitating a more accurate determination of potential links between defects and factory components.

[0047] Additionally or alternatively, S123 may include identifying links (e.g., for each product for which information was received, or a subset thereof) between program identifiers (e.g., program names) associated with various factory machines (e.g., various assembly and / or inspection machines) corresponding to the production of the product (e.g., each linked program identifier corresponds to one or more aspects of the production of the same product). For example, this may include identifying that a set of program identifiers (e.g., program “X12_a” is associated with a first machine, program “X12” is associated with a second machine, program “X12_insp1” is associated with a third machine, and program “a5” is associated with a fourth machine) are all associated with part of the production of the same product. These links may facilitate further understanding of the production process, may enable identification of product counts (e.g., as described in more detail below with respect to S127), and / or may be used in any other suitable manner.

[0048] In a first example, such program links can be identified based on the temporal overlap (and / or proximity) of programs running on various machines (e.g., within a single manufacturing line). For example, given an inspection program identifier and one or more associated time windows during which it was executed on an inspection machine in a particular manufacturing line, program links can be identified for the inspection program identifier based on the associated time windows, as shown by way of example in FIG. 5. For all other machines in the manufacturing line (or a subset thereof, e.g., each machine upstream of the inspection machine), program identifiers that are commonly executed during (and / or near) the associated time windows are likely to correspond to the same product and can therefore be linked to the inspection program identifier; for any machine, no link may be identified if there is no sufficiently dominant program identifier during the associated time window (e.g., if it is not executing for at least a threshold portion of the window duration). Additionally or alternatively, temporal correlations between program identifiers can be identified, and any set (or subset of such a set) of program identifiers with sufficiently strong (positive) correlations can be linked.

[0049] In a second example (which can be performed in addition to or instead of the first example), such program linkage can be identified based on product serial numbers (e.g., as shown in FIGS. 6A and 6B). For example, given a set of one or more serial numbers (and / or other identifiers) associated with a particular product (e.g., a single board, a set of boards with the same design, etc.), the time the product was present on a machine (e.g., as indicated by the serial number information) can be correlated with one or more programs executed on that machine while the product was present, and the associated program identifiers can be linked to the product. This can be repeated for each machine (or a subset thereof) on which the product is known to have been present. Thus, a program identifier linked to a product can be linked to all other program identifiers (e.g., associated with other machines) that are also linked to the product.

[0050] However, S123 may additionally or alternatively include determining any other suitable data mapping (and / or transformation) in any suitable manner.

[0051] 4.2.3 Identifying Defect Rates Identifying the defect rate S127 can function to provide additional context to the defect information (e.g., received in S110) and / or potential defect links (e.g., potential links and / or total defect links identified in S125). In some situations, the defect rate can provide important context that cannot be discerned from the defect count alone. For example, in various situations, 100 defects may be associated with a high defect rate (e.g., 10%, corresponding to 100 defects out of 1,000 total units produced), a low defect rate (e.g., 0.001%, corresponding to 100 defects out of 10 million total units produced), an intermediate defect rate (e.g., 0.1%, corresponding to 100 defects out of 100,000 total units produced), and / or any other suitable defect rate. Thus, it may be beneficial to identify the defect rates associated with various defects.

[0052] In many embodiments, the factory information received in S110 includes a report of detected defects but does not include information indicating successful production (e.g., the number of defect-free units produced, such as units of a particular product type or total units produced) or total production (e.g., total units produced, production "opportunities" including both produced units and defective units discontinued, etc.). This can make it difficult to determine a defect rate (e.g., the ratio of detected defects or the number of defective units produced to the number of defect-free units produced, total units produced, or production opportunities) based on the received information. To overcome this difficulty, S127 can include determining one or more metrics related to total production (e.g., determining the number of production opportunities ("opportunity count") for a particular product type or a particular potential defect).

[0053] For example, the opportunity count of a potential defect may be determined based on the inspection program that can detect the defect. In response to receiving information indicating a particular defect (e.g., in response to receiving information indicating a defect with a particular reference designator for the first time in association with a particular product type), the defect (and, e.g., all defects with the same reference designator associated with the same product type) may be matched to and / or added to a defect catalog (e.g., if not already present) that maps the defect to the associated inspection program responsible for detecting the defect. In one aspect, the defect catalog may further identify defects based on defect type (e.g., defects in a particular reference designator for a particular product type may be mapped to different inspection programs depending on the defect type). The defect catalog may include multiple defect entries mapped to each inspection program (although, additionally or alternatively, it may include an inspection program associated with only one defect entry). Thus, the opportunity count of a defect may be determined based on (e.g., equal to) the number of times the associated inspection program has been run (e.g., during a specified time window, either counted only after the defect was initially added to the catalog or counted to include the number of times the defect was run before the defect was added).

[0054] Additionally or alternatively, the opportunity count for a potential defect may be determined based on links identified (e.g., in S123) between various programs (e.g., assembly and inspection programs) associated with that potential defect's product type. Thus, even if information indicating the number of runs of an inspection program is not available, the opportunity count may similarly be determined based on (e.g., equal to) the number of runs of an assembly program linked to that inspection program. In a particular example where a link is known (e.g., identified in S123) between assembly program "prod17" and inspection program "prod17_insp," which are both associated with the production of product type 17, the opportunity count for all potential defects for product type 17 may be determined to be equal to the number of runs of assembly program "prod17."

[0055] Additionally or alternatively, S127 may include receiving the opportunity count directly (e.g., receiving information indicative of the opportunity count in S110). However, S127 may additionally or alternatively include determining the opportunity count in any other suitable manner.

[0056] Once the opportunity count for a set of defects (e.g., all defects for a product type, all defects for a single reference designator within a product type, defects of a particular defect type in a single reference designator, and / or any other suitable set of defects) is known, S127 may include determining an associated defect rate by dividing the number of defects in the set by the opportunity count.

[0057] Additionally or alternatively, S127 can include determining defect rates associated with the factory components. In some embodiments, the defect rates for the factory components can be determined in a manner similar to that described above with respect to S125 (e.g., with respect to defect link sums and / or defect link time series). For example, in addition to identifying links between defects and upstream factory components, the method can also include identifying links between each production opportunity (and / or each non-defective product, etc.) and the upstream factory components (e.g., determining an associated production opportunity sum for each upstream factory component in a manner similar to that described above with respect to defect link sums in S125). In other embodiments, the defect rates for the factory components can be determined based on estimated opportunity counts. For example, the opportunity counts can be estimated based on production line configuration and / or based on observed workload sharing of various components (e.g., during a portion of the total time for which the opportunity counts are determined). This estimate is preferably determined in conjunction with a known or estimated total opportunity count (e.g., for a particular product type, for a particular production line); for example, an estimated percentage of total opportunities performed by a particular factory component may be multiplied by the (known or estimated) total opportunity count to determine an estimated opportunity count for that factory component. In some example of these embodiments, the factory component's defect rate is determined by dividing its defect link sum (e.g., determined in S125) by the total number of production opportunities (e.g., known or estimated), although the factory component's defect rate may additionally or alternatively be determined in any other suitable manner. In some example of these embodiments, the method may additionally or alternatively include taking an action (e.g., presenting information, performing inspection and / or maintenance, etc.) based on one or more of the factory component defect rates (e.g., in a manner similar to actions taken based on the defect link sum, as described in more detail with respect to S130) and / or performing any other suitable action based on the factory component defect rate.

[0058] However, S127 may additionally or alternatively include determining the defect rate in any other suitable manner and / or may additionally or alternatively include any other suitable elements performed in any other suitable manner.

[0059] 4.3 Actions Based on Defect Association The method may optionally include a step S130 of taking action based on the defect associations, which may function to diagnose manufacturing system defects (e.g., resulting in high defect rates and / or low production rates) and / or to make, enable, and / or facilitate improvements to manufacturing system functionality (e.g., reducing defect rates and / or increasing production rates). The action taken in S130 is preferably informed by the defect associations identified in S120 (e.g., defect link totals, defect rates, etc.) and may additionally or alternatively be informed by any other suitable information identified and / or received during performance of the method. In some embodiments, S130 may include determining one or more elements of manufacturing system maintenance information (e.g., information determined based on the defect associations, e.g., information related to and / or indicative of manufacturing system defects and / or potential solutions, etc.) and taking action based on the one or more elements of the manufacturing system maintenance information (e.g., providing information to a user, performing inspections and / or maintenance based on the information, etc.). For example, manufacturing system maintenance information may include information related to and / or indicative of factory components, maintenance actions, recommendations and / or prioritization that may be contributing to manufacturing system failures, and / or any other suitable information (e.g., such as other information described herein with respect to S130).

[0060] In some embodiments, S130 may include presenting information to one or more users (e.g., derived information, such as information received in S110 and / or information determined based on decisions made in S120). For example, S130 may include presenting potentially problematic factory components to a user (e.g., providing a user with information indicating one or more potential problems).

[0061] For example, S130 may include generating and / or presenting a list of factory components that may be contributing to an unacceptable (e.g., increasing) defect rate in the production line (and / or may be otherwise problematic). This list may be used to inform decisions regarding maintenance to be performed (e.g., preventive maintenance performed during a scheduled production line shutdown, corrective maintenance performed during a production line quality shutdown to eliminate the root cause of a high defect rate, etc.). For example, this list may include factory components (e.g., feeders, nozzles, nozzle holders, cameras, programming parameters, etc.) that are most overall responsible for defects (e.g., highest link defect totals and / or link defect rates), which may be particularly useful for preventive maintenance, and / or factory components most responsible (e.g., link defect totals and / or link defect rates) for particular types of defects and / or defects associated with a particular component or set of components (e.g., one or more defect types and / or one or more components with a particularly high defect rate). Additionally or alternatively, the list may include factory components whose defect liability (e.g., total count, defect rate, etc.) is greater than a threshold (e.g., a uniform threshold, a threshold determined per component or per component type, etc.; a fixed threshold and / or a threshold dynamically determined based on factory-wide performance, etc.). Additionally or alternatively, factory components may be selected in the list based on an automated anomaly detection system (e.g., a machine learning system that accepts as input a time series of factory component defect rates, etc.).

[0062] In one aspect of this example, the list may be filtered and / or prioritized based on the expected time, cost, and / or impact of performing maintenance associated with various factory components. For example, the list may be filtered and / or sorted based on the expected return on investment of maintenance resources (e.g., maintenance costs, factory or factory line downtime, etc.). In some embodiments, S130 may include providing a recommended maintenance list for current or upcoming (e.g., planned, required, or desired due to poor performance, such as low yield and / or production rate) factory or factory line shutdowns. In one example, based on the planned downtime duration (e.g., 5 minutes, 15 minutes, 2 hours, etc.), the list is filtered to include only maintenance tasks that are predicted to fall within the planned downtime duration (e.g., if each task is expected to require less time than the planned downtime duration, if the total expected time required for a set of tasks associated with a particular maintenance resource, such as a maintenance worker or team, is less than the planned downtime, etc.), and / or sorted to prioritize tasks with the highest expected value (e.g., in terms of expected increase in product yield and / or production rate) and / or tasks with the highest expected return on investment (e.g., in terms of the ratio of expected value to expected maintenance time).

[0063] Additionally or alternatively, S130 may include taking action in response to determining that a factory component may have a problem. For example, S130 may include alerting a user, modifying manufacturing line operations (e.g., to avoid or reduce use of the problematic factory component), and / or initiating maintenance (e.g., of the problematic component and / or other related factory components). Determining that a component may have a problem may include determining that a defect liability has exceeded a threshold (e.g., as described above), determining that operation of the component is abnormal (e.g., based on an automated anomaly detection system such as the machine learning system described above), etc.

[0064] Additionally or alternatively, S130 may include presenting the derived information to one or more users in a manner that enables the users to explore various aspects of the information, such as via a user interface (UI) that enables viewing of data slices associated with various defects, product components, factory components (e.g., tools, tool elements, programming parameters, etc.), and / or any other suitable aspects.

[0065] For example, the UI may display defect heat maps for upstream machine components (e.g., nozzles, feeders, feeder mounting locations, etc.), component types, shape codes (associated with programming parameters shared by one or more component types), etc.

[0066] Additionally or alternatively, the UI allows a user to filter defect data. For example, by reviewing defects associated with component U1 (of a particular product type), the UI may show that U1 is placed by two machines from three feeders, but that the majority of the defects originate from only one of the machines (even though the same parts are fed into both machines). Based on this information, the user may conclude that the problem is related to that one machine. The user may then review additional defects related to components of that one machine (e.g., remove the U1 filter and instead filter by machine components of that machine), thereby discovering a machine component (e.g., related to various SMT components) that is related to many defects, where most or all of the defects are likely to be problems with that machine component. Alternatively, the defects of the filtered components may not be primarily attributable to one machine (e.g., may be substantially evenly distributed across the machines). Rather, there may be many defects across multiple machines for a product component (e.g., an SMT component such as a particular resistor, capacitor, IC package, etc.) and / or its shape code, which may indicate a programming error (e.g., an inappropriate selection of component placement parameters).

[0067] As one skilled in the art will recognize, additionally or alternatively, some or all of such functions may be automated, in which case this analysis (or a subset thereof) may be performed without human intervention.

[0068] Further, S130 may additionally or alternatively optionally include performing maintenance (and / or modifying programming parameters) to reduce or eliminate the defect rate (e.g., in response to identifying one or more factory components associated with a high defect rate and / or a particular defect). In a first example, in response to a problematic defect rate, machine components and / or programming parameters associated with the defect may be modified (e.g., via immediate or prior maintenance of the associated component). In a second example, the method may include determining a prioritized list of machine components on which to perform predictive maintenance (e.g., the highest priority components are associated with the highest defect rate or the highest ROI, e.g., the ratio of defect rate to normal maintenance time), where the prioritized list can be used to perform predictive maintenance during line downtime (e.g., shift changes, product changes, etc.).

[0069] However, S130 may additionally or alternatively include taking any other suitable action based on the defect association (and / or based on any other suitable information identified and / or received in performing the method).

[0070] 4.4 Factory Machine Operation The method may optionally include step S105 of operating factory machines (e.g., as shown in FIGS. 3A-3D ), which may function to manufacture and / or inspect products (e.g., SMT boards) designated to be manufactured by the factory machines. S105 may include operating assembly machines (e.g., based on programming associated with one or more products), operating inspection machines (e.g., based on programming associated with one or more products), and / or providing information for analysis (e.g., manufacturing system status, defect detection, etc.) (e.g., provided to one or more analytical tools such as a computing system). This information is preferably received as described above with respect to S110, but may additionally or alternatively include any other suitable information used in any suitable manner.

[0071] Although omitted for the sake of brevity, preferred embodiments include any combination and permutation of the various system components and various method processes. Furthermore, the various steps of the preferred methods may be embodied and / or performed, at least in part, as a machine configured to accept a computer-readable medium storing computer-readable instructions. The instructions are preferably executed by computer-executable components integrated with the system. The computer-readable medium may be stored on any suitable computer-readable medium, such as RAM, ROM, flash memory, EEPROM, an optical device (CD or DVD), a hard drive, a floppy drive, or any suitable device. The computer-executable components are preferably general-purpose or application-specific processing subsystems, although any suitable dedicated hardware or hardware / firmware combination device may additionally or alternatively execute the instructions.

[0072] The drawings illustrate the architecture, functionality, and operation of preferred embodiments, exemplary configurations, and possible implementations of systems, methods, and computer program products according to aspects thereof. In this regard, each block in a flowchart or block diagram may represent a module, segment, step, or portion of code, including one or more executable instructions for implementing one or more specified logical functions. It should also be noted that in some alternative aspects, the functions noted in the blocks may occur out of the order noted in the drawings. For example, two blocks shown in succession may, in fact, be executed substantially simultaneously, or the blocks may sometimes be executed in the reverse order, depending on the functionality involved. It should also be noted that each block of the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented by a dedicated hardware-based system that performs the specified functions or operations, or a combination of dedicated hardware and computer instructions.

[0073] Those skilled in the art will appreciate from the above detailed description, as well as the drawings and claims, that modifications and variations can be made to the preferred embodiments of the invention without departing from the scope of the invention as defined in the following claims.

Claims

1. 1. A method for maintaining a manufacturing system, comprising: receiving, at a manufacturing system, information related to the manufacture of a plurality of products, the information indicating a set of defects detected in products in the plurality of products, the information including, for each defect in the set, a respective set of location information indicating a respective location, in each product of the plurality of products, of a respective part associated with the defect; For each defect in said set: - automatically identifying a respective set of components of the manufacturing system based on the respective part, each component of the respective set having interacted with the respective part; - for each component in each set, incrementing a respective defect link sum associated with the component based on the links between the component and the defects; and controlling maintenance of the manufacturing system based on the defect link sum.

2. 10. The method of claim 1, controlling the maintenance of the manufacturing system - identifying a set of components of the manufacturing system that have a high probability of defective operation based on the defective link sums; and - responsive to identifying the set of components, selecting the set of components for testing.

3. 3. The method of claim 2, The method, wherein identifying a set of components includes selecting components of the manufacturing system that have the highest defect link sum.

4. 3. The method of claim 2, Identifying a set of components Selecting a subset of problematic defects from the set of defects; and - A method comprising: identifying a set of components based on an association between the set of components and defects of the problematic subset.

5. 10. The method of claim 1, controlling the maintenance of the manufacturing system - identifying components of the manufacturing system that have a high probability of defective operation based on the defective link sums; and - in response to identifying the component, selecting the component for repair.

6. 10. The method of claim 1, wherein a product in the plurality of products is a printed circuit board assembly (PCBA) and the component is a surface mount technology (SMT) device, and for each defect in the set, a respective set of location information comprises: - an array position indicating the position of one PCBA within an array of PCBAs; a reference designator indicating one SMT device of the PCBA.

7. 1. A method for maintaining a manufacturing system, comprising: receiving, at a manufacturing system, information relating to the manufacture of a plurality of printed circuit board assemblies (PCBAs), each PCBA including a respective set of surface mount technology (SMT) devices, said information indicating a set of defects detected in a PCBA in the plurality of PCBAs, said information including, for each defect in the set, a respective set of location information indicating a respective SMT device associated with the defect; For each defect in said set: automatically identifying a respective set of components of the manufacturing system based on the respective SMT device, each component of the respective set having interacted with the respective SMT device; - for each component in each set, incrementing a respective defect link sum associated with the component based on the links between the component and the defects; - determining maintenance information for the manufacturing system based on the defect link sum; - providing maintenance information for the manufacturing system.

8. 8. The method of claim 7, For each defect in the set, a respective set of location information is - an array position indicating the position of one PCBA within an array of PCBAs; and a reference designator indicating one SMT device of the PCBA, the SMT device being associated with the defect.

9. 9. The method of claim 8, The method, wherein for each defect in the set, the information further includes a label for a respective defect type.

10. 10. The method of claim 9, For a first defect in the set, the respective defect type label indicates a tombstone defect type.

11. 9. The method of claim 8, 10. The method of claim 9, wherein for at least one defect in said set, each set of location information further includes a pin number of the SMT device.

12. 9. The method of claim 8, determining a defect rate associated with the first reference designator, the step comprising: determining a number of defects associated with the first reference designator based on the information; Determining that a defect associated with the first reference designator was identified by a first inspection process; receiving test information indicating the number of times the first test process has been performed; and determining a defect rate based on the number of defects and the number of times.

13. 13. The method of claim 12, The method, wherein the defect rate is equal to the ratio of the number of defects to the number of times.

14. 8. The method of claim 7, 20. The method of claim 19, wherein for each defect in the set, a total defect link amount is allocated among each component in the respective set, and each respective defect link sum is increased by a respective fractional amount of the total defect link amount.

15. 15. The method of claim 14, the respective fractional amounts being equal, each of the respective fractional amounts being equal to the total defective link amount divided by the number of components in the respective set.

16. 15. The method of claim 14, The method of claim 1, wherein the total defect link volume is equal for each defect in the set.

17. 8. The method of claim 7, The method, wherein for each defect in the set, automatically identifying a respective set of components includes identifying each component of a manufacturing system that interacted with the SMT device.

18. 18. The method of claim 17, For a first defect in the set, the first component of the manufacturing system interacts with an SMT device; the information further comprises a label of the defect type; Automatically identifying each set of components - determining that the first component is not the cause of the defect based on the defect type label; and - In response to determining that the first component is not the cause of the defect, the method includes removing the first component from the respective set of components.

19. 20. The method of claim 18, Regarding the first defect, the defect type label indicates a polar defect type; The method, wherein the first component is a nozzle.

20. 8. The method of claim 7, For each defect in the set: automatically identifying a respective set of programming parameters based on each SMT device, each programming parameter of each set affecting the interaction of the manufacturing system with the SMT device; for each programming parameter in the respective set, incrementing a respective defect link sum associated with the programming parameter based on the link between the programming parameter and the defect; The method, wherein maintenance information for the manufacturing system is determined based on the defect link sum further associated with a programming parameter.

21. 21. The method of claim 20, Based on the total defect linkages associated with the programming parameters, - selecting programming parameters; - changing the value of said programming parameter.

Images