Inspection means have increasingly been incorporated into typical manufacturing of boards, substrates and/or systems. A significant number of automatic inspections rely on the analysis of images that are acquired by a multitude of means such as Optical, X-Ray, Infrared, Acoustic microscopy. In contrast to automatic inspections, traditional visual inspection is performed manually by humans based on images and can be laborious and inaccurate. Detection of “indeterministic” defect types such as cracks and/or scratches is quite challenging since such defects may have a variety of shapes, locations and severity. Deep Learning, a subfield of Machine Learning, has recently advanced the state-of-the-art learning from images and become the standard approach for computer vision tasks. This paper presents a case study for automating visual inspection of boards by as much as 40%. The application can be extended to identify specific types of defects and to support root cause analysis.

Digitalization changes everything, everywhere. It is inevitable, new business drivers are forcing the Electronics industry to rethink every element of their business. Virtually every company is talking about innovation and digitalization. And a major driver is the so-called Digital Twin. While talking about this Industry 4.0 enabler, most people have one benefit in mind: Aggregating the data in a cloud and integration of artificial intelligence for future enhancements which lead to an optimization of the operation. Another one is the simulation of a product and derive the behavior in certain conditions. Both are covering only one certain aspect of the product lifecycle from product design to production execution while there is so much more possible by utilizing the concept of the digital twin in the production engineering phase via Virtual

Commissioning. Streamline the activities of all disciplines involved in the physical commissioning of their automated production systems, reducing errors and increasing the speed in which they bring automated manufacturing systems online by writing the controller program and building the machine at the same time. The paper will review the different digital twin concepts in production:

-Digital Twin of the product: Product design, Hot Spot simulations and adoptions, simulate PCB circuit design-Digital Twin of production workflow: Machinery design, production layout, simulation of production steps & continuous improvement including manual labor-Digital Twin of manufacturing: The emphasis of the presentation will be within this discipline of the digital twin. Steps

1.Write PLC (Programmable Logic Control)/Controller code for machine operation

2.Define sensor and actuators (conveyors, robots, grippers,...) in the CAD model as input and output signals according to the PLC code

3.Define collision models and kinematics in the CAD model (e.g. a product falls over at the end of a conveyor belt or a robot arm must not collide with the chassis of the production cell)

4.Test your machine functionality in the digital twin.

Since physical behavioral model is defined within the CAD model the integration of the PLC code allows to detect failures in the construction of the machine (e.g. robot arm cannot reach a defined position in the process or conveyor belt cannot transport defined parts smoothly) which can be corrected prior to the actual construction of the machine. As a result, the generated PLC code is validated before the machine is even built and reduces the real commissioning time dramatically.

In the provided use case the example will show the collaboration between the end customer and a Chinese OEM, where the machine was constructed in China and PLC code written in Germany with the integration into the CAD model. After the shipment of the machine to Germany the validated PLC code was downloaded smoothly and no structural changes were necessary. Needed changes were recognized beforehand in the Digital Twin model of the machine and communicated to the OEM during construction phase in order to amend the machine design.

The last Digital Twin use case will be the Digital Twin of performance with closed-loop innovation by aggregating data in a cloud and run analytics for further improvement of future production machines or prescriptive maintenance mechanisms.

As the reflow soldering process is the main energy consumer in a SMT assembly line, there is a high potential for cost reduction by significantly reducing the amount of energy needed. Melting the solder only needs a tiny percentage of heat, most of the energy is wasted to heat up the machine itself. If it would be possible to heat up only the solder joints to the required temperature the energy reduction will be substantial.

The crucial factor for such a process is to use a conductive heating material layer inside the printed circuit board (PCB) and generate the heat from inside by joule heating[2,3]. For this process it is necessary to have an electric current flow inside the heating material. This flow must be controlled as the heating layer is a carbon-based material and changes its resistance as a function of the temperature. The control circuit must be able to regulate the produced heat and generate a reflow like profile at the joints of every component on the PCB. The structure of the heating layer must be flexible as a result of having the possibility to adjust the amount of heat in a specific area of the PCB. So for a common PCB it will be necessary to have several connections from the current controller to the PCB.

In the joint research project “ERFEB” (Energy and Resource Efficient Production of Electronic Assemblies)[1], a current control unit will be designed and tested for different numbers of connections and different structures of heating layers to achieve a reliable reflow profile for a PCB soldering process.

As miniaturization trends continue in the electronics industry, System in Package (SiP) technology is gaining more and more traction. In many ways, some SiP modules are just surface mount technology in a high-density package. Component to component spacings are being driven even closer, pcb technology are migrating to package type substrates and more bare die are being used in these modules instead of conventional SMT (surface mount technology) packages.

This paper will discuss the flip chip process that has been developed for a 150μm pitch package, 10x10mm die size and greater than 3700 I/O full array solder flip chip. The test vehicle design will be shown, challenges that were faced during the process development, impact of warpage of the substrate on the flip chip yield and other factors that influence the successful assembly of these types of packages. This paper will describe processes from SMT and underfill and will touch on reliability test runs and the results of those testing.

The advance in technology and its relentless develop mentis delivering yet another surface mount assembly challenge. To meet the market demand for products with higher functionality whilst reducing the overall product size, the next generation of chip package is being readied upon the surface mount community. The Metric 0201 will have dimensions in the order of 0.25mm x 0.125mm,as a result the entire assembly process will be questioned as to its ability to deliver high volume/quality product.

This paper will look at the challenges of assembling the M0201component in a high-volume manufacturing environment. The investigation will start with the printing process, with close attention to the impact of aperture and pad designs. The placement and reflow process will likewise be studied in detail.  The resultant assemblies will be reviewed to determine their suitability for a high-volume manufacturing environment. Discussion and conclusions will be directed at possible Metric 0201 assembly rules and the future challengers that exist.

Polymer coating materials protect electronics from harsh environments, assuring safe and reliable performance. To reach the highest level of reliability, coatings must be selectively applied to avoid keep out zones and critical components, such as connector pins, test points, and relays.

To gain better control over selectivity and eliminate the need for masking, manufacturers often choose to automate their conformal coating process. Air assisted atomization is commonly used in automated processes because it accommodates a wide range of material viscosities. The spray created when external air pressure is applied to the fluid stream provides excellent coverage on the sides of components and is cost effective. When atomization benefits are combined with properly characterized selective coating equipment, yield and throughput can also be increased.

Successfully atomizing the coating fluid relies on multiple factors including fluid chemistry and the amount of air assist applied. Fluid property requirements depend on the desired coating protection and are often defined before the coating equipment is selected.

An optimized spray pattern is determined by the results that appear on a coated board. And identifying the air assist settings to achieve that optimized spray pattern is an iterative process. Issues such as overspray, cobwebbing, and inconsistent pattern thickness can occur. When attempting to resolve these issues, air assist is typically increased or decreased. Unfortunately, these adjustments can introduce an imbalance that leads to amplified complications. If a manufacturer is unable to resolve issues quickly, they might choose to reintroduce manual activities –forgoing the cost savings, consistent quality, and other advantages that come with automation.