Mask printing method and apparatus with optimized parameters

Abstract

In order to determine an optimized parameter value of the mask printing method, a quality value for the substrate is determined in a test method after the printing process, and an image of the upper side of the template is generated. A relationship between the parameter, the quality value and the template image is acquired, and an optimized value for the parameter is determined by means of the relationship. The optimized values are set during the printing process and the printing method is monitored using the corresponding template image, whereby a test method can be omitted.

Description

Technical Field

[0001] The invention relates to a method for printing a substrate with a medium using a printing mask and to a corresponding device. Background Art

[0002] Mask printing methods are printing methods in which the material to be printed (substrate) is covered with a template, i.e. a printing mask, and the medium to be printed is applied to the substrate through the openings of the printing mask by means of a scraper. For example, a screen or a plastic template or a metal template can be used as a printing mask; the term "template" is used for all types of printing masks in the following text. Various fields of application are known, for example, substrates can be printed with paints, lacquers or other pasty media. Substrates are also used in the production of electronic components, in particular printed circuit board substrates are printed with soldering materials in order to be able to produce connections to the printed circuit board substrate itself or to electronic components.

[0003] It is important that the pasty medium, i.e., for example, solder material, fills the openings (holes) of the template as completely as possible after the doctor blade application and does not protrude beyond the side of the template facing away from the substrate (template top side). Otherwise, the printing process and thus the printed circuit board to be produced are also defective; in the first case, the desired electrical connection cannot be established, in the second case, a thicker layer thickness than specified is produced and the printed image becomes inaccurate, and solder residues may also remain on the template top side. Since solder residues are destructive to the next printing process, a cleaning step must be performed if necessary.

[0004] The printing process is influenced by various parameters, such as the pressure and speed with which the squeegee moves over the stencil or the speed with which the stencil is lifted from the substrate. In addition to these controllable parameters, which can in principle be reset for each individual substrate, other parameters also influence the result, in particular those which are usually set only once per production day or process (e.g. the orientation of the squeegee relative to the stencil), those which are set once for a component type in a setup (e.g. squeegee material, squeegee thickness, squeegee length, solder paste product), and those which represent design data (e.g. the thickness of the stencil, the arrangement and shape of the stencil openings). In addition, boundary conditions of the printing process which are not or can only be indirectly influenced (e.g. temperature, humidity, vibrations, air pressure) also fall under the term "parameters".

[0005] Defects can be minimized by selecting suitable parameter values. For example, excessive layer thicknesses of soldering material and residues on the template top side can be avoided as much as possible by suitable values ​​for the scraper pressure and sufficiently parallel orientation of the scraper to the template surface.

[0006] In order to check the printing result, a so-called solder paste inspection step (SPI) is performed in the prior art. The printed substrate is led out of the printing machine chamber and optically inspected. For this purpose, at least one image of the printed substrate is analyzed, for example by direct comparison with a target image, or specific parameter values ​​are acquired and deviations from predetermined target values ​​are acquired. For many pads (Lot-Pad), specific characteristic parameters such as, for example, position and height in the x- and y-directions are acquired, and percentage deviations from the target values ​​are determined. The printed circuit board is discarded if a predetermined tolerance threshold is exceeded or fallen below (that is, a defect that is too large during the printing process). Alternatively or additionally, an optical inspection (so-called automated optical inspection (AOI)) is performed after further manufacturing steps.

[0007] It is also possible to optically check the underside of the template (that is to say the side of the template facing the substrate) after it has been lifted from the substrate in order to detect whether solder residues remain in the holes, so that the pads are not installed at the corresponding locations of the substrate. These known methods are expensive and require a more complex printing press chamber or additional process stations on the production line. In addition, although these methods can determine the defects that have occurred, they cannot provide any conclusions on how the printing process must be changed in order to avoid defects in the future. For the latter, it is necessary to make a decision by experts who, based on their experience, identify which parameters of the printing process must or can be changed in what way in order to eliminate the defects as much as possible. Summary of the invention

[0008] The object of the present invention is to detect defects occurring in mask printing methods in an easy manner and to provide options for avoiding these defects.

[0009] This object is achieved by a method having the features of claim 1, by a method having the features of claim 5 and by a device according to claim 8. Advantageous embodiments of the invention are specified in the dependent claims.

[0010] The invention provides for obtaining an optimized value for at least one parameter of a printing method by means of a machine learning method. For this purpose, a printing process is carried out with the assigned value for the parameter and an image of the top side of the template is recorded before and / or after the template is lifted from the substrate surface. A quality value is obtained for the printing result, that is to say for the printed substrate.

[0011] The printing process is re-run with the changed values ​​of the parameters, an image of the template top side is recorded before and / or after the template is lifted, and the associated quality value of the printed result is re-determined. A relationship, i.e. a relationship between the parameters, the template image and the quality value, is determined in a machine learning method. The machine learning method is then used to determine an optimized value of the parameter, i.e. a parameter value that leads to an improved quality value of the printed result.

[0012] According to one embodiment, the printing process is performed several times with the changed values ​​of the parameters and / or with the changed values ​​of the further parameters, an image of the template top is created in each case, and a quality value is assigned to the corresponding printing result. As a result, the machine learning method or the acquired correlation between the parameters and the quality value becomes more reliable. Based on the correlation between the parameters and the quality value acquired by the calculation unit, it can be determined which corrective measures should be carried out in the case of poor printing results, that is, which parameters should be changed in what way in order to increase the quality value.

[0013] In order to obtain the quality value, an optical inspection of the printed substrate is preferably carried out. For example, the image of the printed substrate can be compared with a target image, and alternatively or additionally, deviations, in particular percentages, can be evaluated in the case of one or more predetermined characteristic variables (parameter values). The smaller these deviations, the higher the quality value. In the case of a plurality of characteristic variables, the deviations can optionally be weighted differently and combined with one another in a suitable manner.

[0014] The method can be advantageously used in the production of electronic components, wherein a printed circuit board substrate is used as substrate and an electrically conductive solder material is used as medium. The quality value can then be determined by means of an SPI step. In a conventional SPI station, the printed substrate is optically evaluated using known methods in such a way that the deviation, in particular percentage, from the corresponding target value is determined for one or more characteristic variables.

[0015] For example, the following controllable parameters can be used as parameters: squeegee speed, squeegee pressure, cleaning cycle, travel path, lifting speed (lifting of the template after the printing process), run-out (ie any distance between template and substrate).

[0016] Alternatively or additionally, for example the following parameters can be used as parameters which are more or less controllable or selectable and / or represent boundary conditions of the process: relative orientation of scraper and template surface, scraper material, scraper thickness and scraper length, batch number, type and manufacturer of medium (paste), precision of metal particles, hole geometry, thickness of template, air temperature, air humidity, vibrations, air pressure.

[0017] When the template is still placed on the substrate, an image of the template can be taken. In this image, it can be recognized in particular whether all holes are filled with solder material or whether some are empty or only partially filled.

[0018] Alternatively or additionally, after the template is lifted from the substrate, a template image can be captured. In such an image, it can be particularly recognized whether soldering material remains in the hole after lifting and is therefore not or only insufficiently printed on the substrate.

[0019] In both cases it is possible to detect whether solder material remains on the template surface at locations without openings.

[0020] Machine learning methods make it possible to capture the complex interrelationships between parameters and quality values, identify the parameters that decisively influence the results, and determine the parameter values ​​for the best results (referred to as the learning phase). Known optimization methods can be used as machine learning methods, such as Bayesian optimization. By adding further parameters, better models with higher reliability can be generated with the aid of machine learning methods.

[0021] The invention also includes a method for producing a substrate printed with a medium, wherein an optimal value is determined for at least one parameter of the printing process and wherein a relationship between the parameter and the quality value is determined as described above. A learning phase can be integrated into the actual production of such a manufacturing method, i.e. a printed printed circuit board or, in general, a printed substrate, in other words, the printed substrate produced in the learning phase can already represent the product to be manufactured. The required change of one or more parameter values ​​is achieved by fluctuations that actually occur in the case of the set parameter values, i.e., the changed values ​​are generated when the method is carried out and are not set specifically.

[0022] In a preferred embodiment, production is carried out after a learning phase, in which optimized values ​​for at least one parameter are acquired as described above. In this production phase, an image of the template top side before and / or after lifting off the substrate is then generated as in the learning phase. The printing method is monitored and / or the printed substrate is evaluated based on the template image. This advantageously allows larger and / or targeted changes in the parameter values ​​to be implemented in the learning phase, whereby the machine learning method provides faster and / or more reliable results.

[0023] If the production phase follows the learning phase, then in the production phase, it is preferred not to optically check the printing result, i.e. the printed substrate, but to create an image of the upper side of the template before and / or after being lifted from the substrate surface as in the learning phase. Because this relationship is learned by machine learning methods in the learning phase, a quality value can be assigned to the printing process or the printed substrate based on the template image. As a result, the SPI step can be omitted in continuous production operation or at least not performed after each printing process. The quality value of the printed substrate can be obtained based on the template image without performing an SPI step and / or an AOI step in the printed substrate. Therefore, production can be accelerated and the equipment structure can be simplified.

[0024] The parameters can be set automatically and printed substrates with predetermined quality values ​​can be produced. Alternatively or additionally, it is possible to correct the set parameter values ​​by evaluating the template image with the aid of the acquired relationship between the parameter values, the quality value and the template image, i.e., to change the set parameter values ​​in the event of defects and a drop in quality in such a way that defects are reduced or avoided in subsequent printing processes.

[0025] It is within the scope of the invention to continue the learning method continuously or discontinuously in the production phase in order to further improve the model. In this case, quality values ​​of the printed substrates produced are additionally acquired in the production phase and used for the machine learning method. As indicated above, a separate learning phase can also be omitted when the machine learning method is carried out in the production phase.

[0026] The invention also includes a device for carrying out the method. The device comprises a receiving device for a substrate to be printed, a support for a template, a support for a scraper, and an image recording device, which is arranged in such a way that it can record an image of the upper side of the template after the printing process has been completed. The image recording device can be connected to a computing unit so that the created image of the template can be communicated to the computing unit.

[0027] The image recording device can be arranged, for example, on the housing or on a holder for the scraper.

[0028] The device preferably comprises a computing unit connected to the image recording device, which is connected to the testing unit. The testing unit is designed to store one or more parameter values ​​for a printing process, at least one image of the template after the printing process, and the quality value obtained by the testing unit, and to obtain a relationship between these data. In this case, the quality value can also be obtained by the computing unit, wherein the testing unit provides data of the test result to the computing unit for this purpose.

[0029] The test unit is preferably a unit for an SPI step, which provides an image of the printed substrate and / or data resulting from an optical comparison of the printed substrate with a target image as test result. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Next, the present invention will be explained in more detail according to the embodiments. The schematic diagram shows:

[0031] Figure 1 and Figure 2 A device for carrying out the method according to the invention during and at the end of a printing process is schematically shown;

[0032] Figure 3 schematically shows an image of a template prior to the printing process;

[0033] Figure 4 schematically shows an image of the template on the substrate at the end of the printing process;

[0034] Figure 5 schematically shows an image of a template on a substrate at the end of a defective printing process;

[0035] Figure 6 schematically shows an image of a template lifted from a substrate at the end of a defective printing process;

[0036] Figure 7 A flow chart of the method is schematically shown. DETAILED DESCRIPTION

[0037] Figure 1 Schematically shown is a chamber 1 of a stencil printer having a receiving device 2 for a printed circuit board substrate 3. A stencil 4 is arranged on the substrate surface by means of a holder (not shown). The scraper device comprises a holder 5 for a scraper 6. During the printing process, solder is applied as a medium 7 to the surface of the stencil 4 by means of the scraper 6, as indicated by the arrows. In this process, the openings in the stencil are filled with solder. Furthermore, an image recording device 8 is arranged in the chamber in such a way that it can record the surface of the stencil 4.

[0038] The computing unit 10 is connected to the camera 8 so that the image of the camera can be transmitted to the computing unit. In addition, the computing unit 10 is connected to a control unit 11. The control unit controls the printing process in the following way: the control unit sets parameter values ​​and controls the movement of the scraper unit, for example. For this purpose, the control unit receives data from the computing unit 10.

[0039] The computer unit 10 is connected, at least in a learning phase, to a test unit 12 in the form of an SPI station.

[0040] Figure 2 The arrangement is shown at the end of the printing process, the scraper having reached the end of the template 4. An image of the upper side of the template is now created using the camera 8. In an embodiment, the template 4 is on the substrate 3 when the image is taken, alternatively or additionally the image can be taken after the template has been lifted from the substrate. The image is communicated to the computing unit 10. The printed circuit board 3 printed with solder is moved into the testing unit 12 and tested there; in particular, an image of the substrate surface is taken and the parameter values ​​thus obtained (among others, the position, area and volume of the pads) are compared with the target values. The resulting data, for example the percentage deviation from the target value, are communicated to the computing unit 10, which determines the quality value therefrom.

[0041] Figure 3 The surface of a template 4 is shown. The surface has openings 40 (holes) through which the medium is brought onto the substrate surface during the printing process. The thickness of the template corresponds here to the desired coating thickness.

[0042] Figure 4 The image of the surface of the template 4 still on the substrate after the printing process has ended for a defect-free print is shown. The holes 40 are completely filled, the scraper removes the solder cleanly from the template surface, so that no solder residues can be seen outside the holes.

[0043] Figure 5 An image of the surface of the template 4 still on the substrate after the printing process has ended for the case of defective printing is shown. Some holes are not filled (40a) or not completely filled (40b). In other holes 40c, solder residues remain outside the holes on the template surface.

[0044] Figure 6 Shows the corresponding Figure 3 The image of the surface of the template 4 lifted from the substrate after the printing process has ended for the case of defective printing. Due to the lifting, further defects occur, because the soldering material remains hanging in the holes 40d and 40c. Here, the desired contact cannot be established on the printed circuit board. On the other hand, it cannot be seen in this image of the template that the opening 40a was not filled during the scraping, and therefore a defect also occurs here. It is therefore advantageous to create an image of the template surface before and after lifting from the substrate in order to reliably detect defects.

[0045] Figure 7An embodiment of the process flow for the method is schematically shown. In the learning phase 20, the printing process is carried out in step 30 using preselected values ​​for the parameters of the printing process. For example, the scraper material, scraper speed, scraper pressure and template type can be selected. The control unit 11 stores the parameter values ​​and controls the printing process accordingly. After the printing is completed, the control unit 11 controls the recording of an image of the template surface by the camera 8 (step 31), preferably recording images before and after the template 4 is lifted from the substrate 3. The image and the associated parameter values ​​are communicated to the calculation unit 10.

[0046] In step 32, the printed substrate is analyzed in the test station 12, in particular at least one image of the printed substrate is recorded in the SPI step and the geometry of the printed deposit or the pad is measured. After the data are communicated to the computing unit, a quality value for the printing result is acquired in the test station 12 or in the computing unit 10. For example, the percentage deviations from the target values ​​for the position and thickness of the printed pad can be weighted and combined in order to calculate the quality value. In step 33, the quality value, the template image and the used parameter values ​​are stored in the computing unit.

[0047] In step 34, the value of at least one parameter is changed. The printing process is carried out with the changed parameter value and the steps 30-34 are carried out in a loop until a sufficient amount of data is present. In step 35, a machine learning method is used in order to extract the relationship between the parameter, the template image and the quality value from the data. The learning phase 20 is thus ended.

[0048] Next, the production phase 21 for manufacturing the printed substrate is performed. In step 36, the computing unit acquires the best quality value and the associated parameter values ​​and template image. The control unit uses these best parameter values ​​to perform the printing process (step 37). After the printing process, at least one template image is photographed in step 38, and it is checked whether the template image is consistent with the desired best template image or whether it is within a predetermined tolerance range. In the case of excessive deviations, suitable measures can be performed, which are acquired by machine learning in the learning phase or derived from the acquired relationship. In particular, it can be derived from the relationship in what way to change which parameters of the printing method, that is, how the value of the parameter must be corrected, in order to achieve the desired quality value. The SPI step can be omitted. However, it is also possible to analyze the printed substrate with the help of SPI, and the obtained data can be used for further improvement. Steps 37 and 38 are performed in a loop to print another substrate.

[0049] The features and aspects of the present invention described in the embodiments can of course be combined with each other in different ways. In particular, the features can be used not only in the described combinations, but also in other combinations or alone.

[0050] List of reference numerals:

[0051] 1 Chamber

[0052] 2 Receiving device

[0053] 3. PCB Substrate

[0054] 4 Templates

[0055] 40a-d holes

[0056] 5. Scraper device / bracket

[0057] 6 Scraper

[0058] 7 Brazing material

[0059] 8. Camera

[0060] 10 Computational Units

[0061] 11 Control Unit

[0062] 12 Test Unit

[0063] 20 Learning Phase

[0064] 21 Production Phase

[0065] 30-35 Steps in the learning phase

[0066] 36-38 Steps of the production phase.

Claims

1. A method for determining optimized parameter values ​​for a printing method, in which a substrate to be printed is printed with a medium using a template, The method has the following steps: - performing a printing process with the assigned value for at least one parameter, - creating an image of the upper side (4) of the template before and / or after said template is lifted from the substrate surface (3), - obtaining a quality value of the substrate (3) printed during said printing process, - performing said printing process with changed values ​​for said parameters, - after the printing process with the changed parameters, recreating the image of the upper side (4) of the template before and / or after the template (4) is lifted from the substrate surface, - re-determining the associated quality value for the substrate (3) printed with the changed parameters during the printing process, - obtaining the relationship between the parameter, the quality value and the template image by means of a machine learning method, - obtaining an optimized value for the parameter according to the relationship.

2. The method according to the preceding claim, characterized in that The quality value is obtained by optical inspection of the printed substrate (3).

3. The method according to any one of the preceding claims, characterized in that A printed circuit board substrate is used as the substrate (3), and a conductive solder material is used as the medium (7), and the quality value is acquired through a solder paste inspection step.

4. The method according to any one of the preceding claims, characterized in that Parameters used are squeegee speed, squeegee pressure, cleaning cycle, movement path, template lifting speed, relative orientation of squeegee and template surface, squeegee material, batch number or properties of the medium, geometric data of the template opening and / or thickness of the template.

5. A mask printing method for manufacturing a substrate printed with a medium, -in, According to any one of claims 1 to 4, a relationship between a quality value and a template image is obtained for at least one parameter, - wherein optimized values ​​for said parameters are set, - wherein an image of the upper side of the template (4) is created before and / or after the template is lifted from the substrate surface, - wherein the template image is used for monitoring the printing method and / or for evaluating the printed substrate.

6. A mask printing method according to the preceding claim, wherein: In order to determine the relationship between the parameter and the quality value, the parameter is first set to a value and a changed value for the parameter is derived from fluctuations in the parameter.

7. The mask printing method according to any one of the preceding claims, wherein: The set values ​​for the parameters are modified based on the evaluation of the template image.

8. A device for performing the method according to any one of claims 1 to 7, the device comprising: - a receiving device (2) for said substrate (3), - a support for said formwork (4), An image recording device (8) which is arranged in such a way that it can record an image of the top side of the template (4) after the printing process has been completed.

9. Device according to the preceding claim, characterized in that The device comprises a computing unit (10) which is connected to the image capturing unit (8) and to a testing unit (12), wherein the computing unit is designed to acquire, in a machine learning method, a relationship between at least one parameter of a printing process, a quality value assigned to the printed substrate with the aid of the testing unit (12), and an image of the template (4) captured after the printing process.

Images