Integrated machine vision optimizes automated component assembly
- Aixemtec GmbH in Herzogenrath, Germany, develops automated solutions for the precision assembly of electro-optical systems.
- Technology advances helped Aixemtec saved more than 60% of space.
- Aixemtec retained modularity by using distributed clocks and fast control processes in the sub-millisecond range.
PC-based controls and integrated machine vision are among technologies helping Aixemtec GmbH in Herzogenrath, Germany, to develop automated solutions for the precision assembly of electro-optical systems. Founded in 2016 as a spin-off from the Fraunhofer Institute for Production Technology (IPT), the technology company offers customized modular systems. These systems range from material feeding and handling to micromanipulation and measurement for ultra-precise assembly, rounded off with quality assurance. PC-based control, including integrated machine vision software, has proven invaluable in helping ensure accurate and rapid process workflows.
Optical systems are used in a wide range of products, in part because of the increasing miniaturization of core technological components. Other application examples can be found in sensors and cameras for autonomous driving such as LiDAR and driver assistance cameras, for gesture recognition and for beam shaping of high-power lasers. The accuracy demands for the assembly of these optical systems range from a few micrometers to a few hundred nanometers.
Combined with cycle times of less than one second and the production of several million components per machine and per year, special demands are placed on the entire automation system.
Compact and flexible automation technology
Aixemtec offers assembly solutions for various application areas based on a comprehensive modular system platform while applying a scalable housing design concept to each machine. Precise and dynamic linear drives are frequently used. Automation components help optimize space utilization across the platform while minimizing the system footprint. One example is the compact and modular servo drives used to control some linear motors, requiring less space in the control cabinet than before.
The industrial PC (IPC) used is another key factor in reducing system size. Previously, a separate PC for image processing, human-machine interface (HMI) and sequential control was used alongside an embedded PC for machine control. Now an IPC performs all these tasks. Compact drive technology in an input/output (I/O) form factor and high-density EtherCAT terminals also save space on the DIN rail.
“The small dimensions of each axis module, the possibility of lining up the modules beside each other and the integration of the safety function, with no additional wiring, but rather EtherCAT, have saved us 63% of space requirements on the mounting plate in comparison with our previous solutions. Plus, we can also control third-party drives with it,” said Sebastian Sauer, head of machine development at Aixemtec.
IPC hardware and software interfaces
The central IPC controls all machine components with its powerful multi-core performance. On the software side, the integrated machine vision software is used in the universal engineering and runtime platform alongside programmable logic controllers (PLCs), NC PTP, functional safety and HMI, and it plays a key role in connecting up to eight industrial cameras. Certain preparations for the process chain are carried out outside of the real-time environment in a specially-developed high-level language program. Here, the PC-based control system allows the automation software and user-specific programs to be integrated on the same IPC. The automation software supports the necessary communication between different software systems with universal communication interfaces such as automation device specification (ADS) or OPC UA.
Sauer said there was another key aspect, the philosophy that “We want to be there for our customers as a partner throughout their entire product chain. The ‘friction losses’ we have saved as a result are tremendous.”
Pick-and-place process for micro-lenses
A prime example of high-demand automation is seen in the pick-and-place process for micro-lenses. To prepare the randomly fed in micro-lenses for assembly, they must first be arranged in a specific orientation on a tool carrier. The goal is about quickly positioning the fragile components in a workpiece magazine. In many cases, these components have a cross-section no wider than a few hairs.
The micro-lenses are fed in bulk on a backlit surface. This surface is scanned with a camera via XYZ kinematics. The result is a 2D panorama of the area being examined. By means of precise time synchronization of NC axis control with machine vision in the same software – with the help of the distributed clocks function of EtherCAT – image capture can be precisely aligned with axial positioning. Where before a time-consuming PTP process was used, integrated machine vision can reduce the setup time of this process by at least a factor of eight.
It is not necessary to stop for each individual image capture. The individual images created in this way during the “flyover” are inserted into an overall image with real-time pixel accuracy. The micro-lenses are identified by image processing and their current orientation is measured. This results in an efficient work plan for how the pick-and-place system should pick up the individual lenses, orient them in all spatial dimensions and place them. The linear axes controlled by the multi-axis servo drive system enable highly dynamic and yet highly precise kinematics.
The micro-lenses prepared for assembly are bonded with a light source in a subsequent step. For this purpose, an adhesive dispenser applies a fixed amount of adhesive to the micro-lens. Precise dosing is essential for correct assembly, which is why the drop flow is continuously monitored and adjusted by a camera system during the dispensing process. In the future, this task also will be taken over by the same machine vision solution, as drop triggering, image capture and exposure can be ideally synchronized with an EtherCAT terminal for LED strobe control and the distributed clocks function. Before the adhesive cures, the optical function of the system to be assembled is optimized with the help of a 6D manipulator in a closed control loop.
Maintaining modularity in real time
The complete hardware and software package, including fully integrated machine vision software is ideal for Aixemtex, according to Sauer. “We are able to retain our modularity by using distributed clocks, he said, adding this provides fast control processes in the sub-millisecond range. The integrated machine-vision software “allows us to remain completely in the real-time environment even in the context of image processing. One particular advantage for us and our customers is that we already rely almost entirely on OpenCV for applications outside the real-time environment.” Since the automation provider also uses OpenCV – an open-source library of computer vision programming functions and tools – Aixemtec engineers can build on their existing expertise.
The comprehensive product range allows uniform workflows during the engineering process and a high-performance yet compact machine design. Versatile interfaces for software and hardware connection help the engineers overcome all challenges in a consistent manner. The open system solution from Aixemtec offers maximum flexibility so any customer-specific process can be mapped.
“Industrial image processing has long been state of the art and one of the cornerstones of our machine architecture,” said Aixemtec managing director Sebastian Haag. With the machine-vision software, “This can now be integrated into process control in a cycle-synchronous manner.”
Keywords: machine vision, automation assembly
How have machine vision applications helped in your manufacturing facility?
For more information:
The post Integrated machine vision optimizes automated component assembly appeared first on Control Engineering.