Related: Automotive Production
Spot welding? Of course. Diecasting machine tending? Natch.
Laser cutting? Uh-huh.
The issue here is robotic implementation. Painting and arc welding. Adhesive application and machine tool load and unload. These applications and those mentioned above are really the bread-and-butter, in effect, of robot applications. But then there's testing. This hasn't exactly been one of the primary applications for robots, says J. Scott Myles, vice president of Control Power-Reliance (Troy, MI), a firm that has been producing testing equipment since 1957, the vast majority of it for automotive applications.
Blinding Flash of the Obvious
Control Power-Reliance is a part of DCT (Detroit), the firm well known among OEMs and suppliers for its ability to produce assembly lines and other systems. Myles recalls how he, three years back, was visiting DCT. He was distracted. He was thinking about an order that he had to engineer for a customer, a test system for seats.
The object that he needed to build was to perform sliding entry tests: How people get into and out of cars. They'd been making systems of this type at Control Power-Reliance for a number of years. They were large superstructures that were fitted with a number of cylinders and slides that provided X and Y axes motion. These machines were special machines. Each one was built to meet specific requirements. Consequently, there were facilities within Lear Corp. and Johnson Controls that were full of dedicated testing equipment.
So, while walking through the DCT plant, something occurred to him: There was an extensive body assembly system right in front of him. There were countless robots on the line. "Why," he asked himself, "not use a robot to perform the motions?"
Heating Up Testing
Robots, so far as he was aware, hadn't been used for automotive component testing. But for the job at hand, Myles realized that the robot would be a natural. For one thing, one of the issues that robot users for more conventional applications need to take into account and make adjustments or compensation for is the weight of the end effector. On a large robot, this can be on the order of 100 lbs. But that's about the mass that is being sought for a seat testing application. In other words, the load is actually removed from the robot when it is in use.
He built the seat testing unit with the robot. The cost to the customer was about the same.
The seat manufacturer began to use the robot. They started using it at the same cycle rate as a dedicated machine. But as time went on, the testers realized that they could increase the speed. It seems that they got so interested in the possibility of cranking up the cycle rate that it actually got to a point where the seat was being heated by the end of arm tooling. They got to a point where they were getting a 300% increase in their testing throughput.
Myles talked with another seating company. Its engineers, too, were interested in the possibility of going for robotic applications. They wanted something a bit more extensive. There is a test that Ford requires that has a 57-step cycle. There are varying loads applied during the cycle. So this was going to be a more challenging task for Myles and his colleagues.
So he went to robot companies to get some help in terms of developing the application. But he discovered that many of them weren't interested in taking on the tasks. Which is understandable. Realize that the robot business in automotive tended to be about multiple units—big orders for things like body-in-white assembly. Motoman, Inc. (West Carrolton, OH) did agree to help the development of what is accurately, descriptively, but perhaps not tastefully known as the Ro-Butt testing unit.
The company also makes systems that check the integrity of welds and castings and the like. It got an order for a laser weld inspection system for transmission parts. There are 11 different parts that need to be checked. Historically, the job would be done with dedicated fixtures. There would be huge superstructures created to handle each of the parts.
But the system they devised makes use of a robot from Nachi Robotic Systems (Novi, MI). Welds are checked with ultrasound. The robot, using standard grippers, picks up the part to be checked, then immerses it in a tank of water. The robot moves the part in front of the ultrasonic transducer and rotates the part to make a 360° check. Doing this with dedicated automation is far, far trickier from the standpoint of mechanisms.
One interesting aspect is that although sometimes special systems can have a throughput advantage compared to more flexible devices, like robots, in this case, the cycle time (14 seconds or less) compares with what is achieved with the dedicated machinery.
Myles is thinking about additional applications for robots. But he observes, "They're not for every application. But if you can incorporate material handling and processing..."