Vision-enabled robot replaces 4 conventional flame-treating robots
Inside Machines: Robot with machine vision replaces four conventional robots in this Fitz-Thors Engineering Inc. design for a tier 1 automotive supplier. The robot operates continuously, eliminating robot idle time with conventional cells. The new robot approach also eliminates proximity sensors, making it possible to handle a wider range of parts.
Enabling a robot with machine vision and careful workcell layout improves asset utilization and frees three robots for use elsewhere. Volatile organic compound (VOC) free coating systems used for automotive interior components require flame treatment to increase the surface tension and improve coating adhesion. Traditional cells used for flame treating have four stations, with a robot in each station applying or flame treating parts mounted on a turntable fixture with proximity sensors. Fitz-Thors Engineering Inc. has improved on this approach by producing a cell that provides the same throughput with one robot serving all four stations.
The vision system views the part prior to flame treating to ensure it is present and properly positioned in the fixture. The robot operates continuously, eliminating the considerable amount of robot idle time seen with conventional cells. The new robot approach also eliminates the need for proximity sensors, making it possible to handle a much wider range of parts and eliminating the time required for wiring the sensors. When running at full production, the new cells require four operators, the same as traditional cells. However, when running at less than full production, the new cells can be serviced by two operators loading two cells each.
Traditional flame-treating methods
Most plastic parts used in automotive interiors are coated to improve their feel and acoustic performance. Recent regulations have mandated the use of VOC-free coating systems based on polyurethane or ultraviolet coatings that require the substrate to have a high level of surface tension. Flame treatment is an inexpensive and efficient way to provide the surface tension levels that can help promote coating adhesion and eliminate the need for primers. Flame treatment in the automotive industry is normally applied with a robot to ensure high quality levels. A typical vehicle interior might have several components that require treatment, such as steering wheels, dashboards, consoles, glove boxes, etc. A traditional cell used to flame treat these components has individual stations with each station served by an individual robot. Each different part has a custom fixture with proximity sensors to ensure that the part is properly positioned in the fixture. Each fixture is mounted on a turntable to isolate operators from the robot. Operators load each part, and then rotate the turntable to put the part and fixture into position for flame treatment by the robot.
Workflow process efficiency
Fitz-Thors engineers decided to take a fresh look at flame-treatment systems. They questioned why a robot was needed for each station when the robots spent much of the time sitting idle. They also asked whether machine vision might be a better approach than proximity sensors to ensure the proper positioning of the part. Fitz-Thors' engineers concluded that a robot enabled with a vision system could move quickly around the cell to each fixture that was ready for flame treatment, verify that the right part was present and properly positioned in the fixture, flame treat the part in the fixture, and then move on to the next.
Fitz-Thors engineers selected an entry-level vision system developed for inspection tasks where vision sensors are too limited and standard vision systems may not be cost effective. They configured the vision system with a software user interface by selecting the pattern find vision tool with an angle orientation limit that provides a percentage match to ensure that the right part is positioned correctly in the fixture. The pattern-find function can locate the part anywhere in the field of view so the program can consistently detect the presence or absence even when the fixture might be in a different position.
The self-contained vision system includes autofocus optics and integrated lighting in an IP67 rated industrial housing. With autofocus, users can set and save the focus values associated with the inspection of each part. Users also can fine-tune focus levels manually with the interactive software, enabling seamless part change over without any manual adjustment of the lens. Integrated white lighting is suitable for most vision applications. If a specific color light is required to highlight particular parts or features, four optional colored lights are available.
A robot fitted with a surface treater flame treats the parts. All process cables and hoses are routed inside the robot arm to decrease downtime caused by interference and wear. Integrated dressing also ensures that the maximum achievable acceleration is available at all times without restriction. The robot has a swing base radius of only 337 mm and a base width of only 511 mm. The surface treatment ignites flammable gas to form an intense blue flame and a plasma field. The flame changes the distribution and density of the electrons on the surface of the part and polarizes surface molecules through oxidation. Flame treatment also deposits other functional chemical groups that promote wetting and adhesion.
Cell for tier 1 supplier
Fitz-Thors Engineering recently built a flame-treatment cell for a major tier 1 automotive supplier. The four-station system uses a turntable to handle smaller parts while the other three stations are open to handle larger parts such as dashboards. A robot safety option enables operators to safely load parts on the three stations without turntables. Sensors monitor the location of the operators, and the robotic safety option enforces geometrical and speed restrictions to maintain automatic operations while restricting robot motion to maintain a safe distance from operators. A safe tool zone is continuously calculated based on operator position, and the robot motion is restricted to stay inside the defined zone. The operator loads the turntable or loads a part into one of the open fixtures and pushes a green button. The robot control controls the robot based on the availability of parts in fixtures and the position of the operators. Operators can walk in and place parts in cells while robots are working in a different station.
Using vision instead of proximity sensors makes it possible for each fixture to handle a wider range of parts. Design changes can frequently be accommodated without having to make any changes to the fixture. In addition, the elimination of proximity sensors eliminates the need for a substantial amount of wiring when installing the system and also for rewiring when changes are made to the design. The vision systems cost about as much as the proximity sensors, but their flexibility to handle future changes is much greater so they reduce the total cost of ownership.
One robot used in the cell continuously treats parts without downtime. The floor space occupied by the cell is considerably smaller than a traditional cell with four robots and four turntables. When the system runs at full capacity, it uses four operators to continuously load parts for maximum throughput. Another option is to run at a reduced production rate with two operators each loading parts for two stations.
Machine vision for test stand
Fitz-Thors Engineering uses machine vision in many other applications. For example, the company recently completed an automated custom vision application for another tier 1 automotive supplier. These test stands use vision systems and displays to verify the location and color of the screws and plastic clips on the inside of door panels. Machined fixtures ensure the door panel is positioned correctly by the operator. Once the panel is in position, the operator initiates the test by pressing two buttons. The display alerts the operator of a failed part by displaying a red box around the quadrant that failed as well as placing a red circle around the incorrect or missing fasteners. This visual indication allows the operator to quickly identify why the part has failed. The results are stored in a database, providing traceability in case future defects are found.
The first flame-treat cell shown here has been in successful operation for more than one year. Fitz-Thors Engineering recently built a second system for the auto supplier. This new cell is capable of functioning as a backup to the current cell and also will make it possible to treat new components and increase throughput.
- Ron Pulicari is Cognex Corp. marketing manager for the Americas. Edited by Mark T. Hoske, content manager, CFE Media, Control Engineering, mhoske(at)cfemedia.com.
- Consider workcell and workflow design to maximize robotic utilization.
- Machine vision allows one robot to work more efficiently and on more parts than four using proximity sensors.
- Four operators work the cell at full speed or two operators can for less throughput.
Before applying robotics, have you reconsidered workcell and automation design to maximize flexibility and throughput?
- See related machine vision story below.