Robot frees 2 operators for inspection tasks

A chucking machine that produces elbow- or tee-shaped pipe fittings was difficult to automate because the fittings, when loaded, had to match the angle of rotation where the chuck happened to stop spinning.

The challenge was solved by using a Cognex In-Sight vision system to determine the angle of rotation of a marked disk that was attached to the opposite end of the spindle.

The angle of rotation is passed to a Mitsubishi RV-12 robot that twists its wrist to insert a new part into the chuck in exactly the right orientation. One robot now loads and unloads two machines, freeing up two operators who are now able to perform inspection tasks.

The company that owns the chucker produces high-end industrial pipe fittings in diameters ranging from ¼-inch to ½-inch from steel forgings or brass castings. The chucker has seven spindles that operate simultaneously.

When the machining operation on any spindle is finished, the spindle indexes upward into a common position where it can be more easily accessed for unloading and loading. Loading the chuckers is a dirty job–the machine sprays cutting oil even during loading because the seven other spindles are in operation.

The operators that load these machines quickly become covered with oil, making it difficult to find people who are willing to perform this task day in and day out.

Challenge Of Automating Loading Operation

The fitting manufacturer was interested in automating the loading and unloading operation, but ran into an obstacle. Because the elbow and tee fittings are not radially symmetrical, they cannot be held in a chuck with a cylindrical opening.

Instead, the chuck has two flat jaws opposed to each other that move in a radial direction to clamp the double-ended section of a tee fitting, or either arm of an elbow fitting, so that the opposed end can be turned, faced, and threaded.

When the power is turned off after the chucker finishes making a part, the chuck continues spinning due to centrifugal force and then stops in a random orientation. Typically on newer equipment applications automated robotic chucker  loaders do not require vision because the chuck’s position remains constant and radial symmetry allows the part to be loaded in any orientation.

But the lack of radial symmetry in this application, or on any older Chucker machine, means that the fitting must line up precisely with the angle of rotation of the jaws during loading. A conventional “blind” robot has no way of knowing the chuck’s orientation and is, therefore, unable to grip the part.

Steven Douglas Corporation (SDC), a designer and manufacturer of custom and special factory automation machines based in Newbury, Ohio, was contracted to auto load two Chucker machines.

SDC, in turn, enlisted the support of RAF Automation, a Cognex distributor and integrator specializing in automation applications based in Solon, Ohio. The two companies’ engineers worked together to develop the innovative vision-guided robot application.

An Innovative Solution

“We first noted that it would be difficult to have a vision system looking directly at the chuck to determine its orientation,” says Steve Blair, Sales Engineer for RAF Automation.

“The chuck is continually being sprayed with cutting oil that could interfere both with the vision system and its ability to view the chuck.”

However, the opposite side of the machine spindle protrudes in an area that is much cleaner and free of cutting oil. The engineers came up with the idea of mounting a disk on the back spindle and marking the disk with a distinctive pattern of dots that the vision system uses to determine which spindle is in position to be unloaded.

The robot is mounted in an inverted position centered between the two chucker machines where it can equally service both. RAF Automation recommended two Cognex In-Sight 5100 vision systems for this application. These systems contain all the processing power needed for complex vision applications in a compact 30 mm by 30 mm by 60 mm enclosure.

“The wide selection of available In-Sight vision systems made it possible to select a very economical camera while still meeting the speed and accuracy requirements of this application,” Blair says.

Integration Between Robots & Vision Systems

SDC engineers selected a Mitsubishi RV-12 series 6-axis robot with a 12 kilogram payload. The robot is surrounded by wire mesh guarding to protect personnel in the area. Mitsubishi robots and Cognex In-Sight vision systems have been integrated through a joint development effort between Cognex and Mitsubishi Electric.

Mitsubishi Melfa-Vision software incorporates robotic programming software and Cognex In-Sight Explorer software to set up the robot and the vision system simultaneously as an integrated solution.

“This new approach makes it much easier to create vision-guided robotic applications with capabilities such as locating and inspecting parts that go far beyond what robots alone can accomplish,” says Paul Belliveau, Vice President of SDC.

RAF Automation and SDC engineers worked together to develop the vision application. They used the Melfa-Vision application to configure the camera and wrote the rest of the application using Cognex In-Sight Explorer software.

The application also includes a Mitsubish Q-series programmable logic controller (PLC) that operates the robot and a human machine interface (HMI) used by operators to switch from one part to another and view information on the performance of the automation system.

The robot initiates communications with the vision systems and delivers commands to capture an image and perform the calculations needed to determine which chuck is in position to be loaded and what angle it has stopped at.

The robot controller, PLC and the vision system use a proprietary protocol developed by Mitsubishi that is included in Melfa-Vision software and all communicate via Ethernet.

The HMI is used to calibrate the robot to the vision system. The robot is jogged into position to grip a part located in the chucker. The vision system then captures an image to determine the angle of the chucker. The difference between the angle determined by the vision system and the angle of the robot’s wrist is then entered as the offset to calibrate the robot. Prior to beginning operation, the operator loads parts into Delrin pucks (blocks of Delrin with features to hold the parts in a specific orientation).

The operator places the pucks on an infeed power-and-free conveyor that locates the next part to be loaded on a precision escapement. The automation provides a signal when the infeed conveyor is running short on parts.

Step By Step

When the PLC identifies that one of the machines is ready to be loaded, it notifies the robot using the Modbus TCP protocol. The robot, in turn, sends a command to the vision system monitoring that particular machine to capture an image and determine the spindle number and its angular position.

The Cognex In-Sight Explorer circle fit tool is used to locate the center of each disk. In-Sight Explorer blob tools are used to identify the dots on the disk and determine which disk is in position for unloading. Cognex’s pattern tool locates the unique marks on each disk that are used to determine its angular position. Custom code is used to determine which disk is in the load/unload position and determine its angular position.

The vision system sends the spindle number and angle of rotation to the robot which in turn relays the ID number to the PLC. The robot then moves to the chuck and attaches its gripper to the part. The robot does not require the angular position for the unload operation because the end of the part being machined is radially symmetrical.

The robot signals to the PLC that it has gripped the part and the PLC instructs the machine to open the chuck jaw. The machine then indicates that the chuck jaw is open and the PLC passes the information along to the robot.

The robot, then, removes the part from the jaw and places it into a Delrin puck on the outfeed conveyor. It picks a part from the escapement on the infeed conveyor and twists its wrist to the orientation provided by the vision system so the part fits precisely between the jaws of the chuck.

The robot then signals the PLC that the part is in position and the PLC instructs the machine to close the jaws of the chuck. The machine signals the PLC that the chuck jaws have been closed and the robot then releases the part and moves away from the part, back to its rest position.

The completed system is now in operation at the customer’s site and is working exactly as designed. The vision-guided robot application has substantially increased production, and reduced labor costs associated with these two chucker machines.

For more information visit and