Robots For Material Removal Maintain Their Cutting Edge
by James F. Manji, Contributing Editor
Robotic Industries Association Posted 10/16/2000
Robots used for laser or weld cutting of materials ranging from plastics, titanium, steel, aluminum, silver and gold, and even wood are seeing some of the most sophisticated advancements ever. In the forefront of these advances are Off-Line Programming (OLP) and Cloning of robotic workcells used in laser cutting of steel and other metals for the automotive industry.
Some of the major robot manufacturers are reporting fair success with OLP and Cloning. One of these robot-makers is Motoman Inc. (West Carrollton, OH). 'We think of robot cells as having three major components,' explains Carl Traynor, director of marketing for Motoman. 'The first component is the robot itself. The second is the end of the arc tool, the weld torch, or the laser cutting torch. And the third is the fixture or positioner upon which the part is located. By identifying some 40 parameters of that complete work cell-the robot, end effector and fixture for each of the cells-we're able to mathematically adjust for filter positions in the cutting program to reflect the differences between the cells. So, we can teach a program on one cell and, through the calibration and filter process-transfer the program to another robotic cell without losing production time or having downtime. This is off-line programming and cloning.'
The twin processes of OLP and Cloning are spurring interest for hydroform components in the automotive industry. 'We've seen and been successful in cutting with coated materials,' says Eric Nieves, manager of technology advances for Motoman. 'Where previously, the emphasis was on whether the robot could execute these parts, now it has become a process-intensive issue and we're actually seeing the parts themselves. Working with galvanized materials places much of the onus on the automation integrators to understand the process, sometimes even beyond the scope of the actual laser suppliers.'
Hydroform components are basically a means to form steel tubing, according to Nieves. Previously, two blanks of sheet metal would be formed into two U-channels, which would be nested one inside the other. These would then form the box assembly or car or truck frame. In this manufacturing process, all the holes for the connections for the suspension system or routing or cabling would be stamped in place before they were formed into a channel. This channel would then be stamped, bent, nested, welded, etc.
Hydroforming, on the other hand, forms the frame or chassis in a sophisticated water pressure press, which squeezes the square tube into the shape of the chassis with water pressure, Nieves continues. This, however, makes it impossible to knock out holes in a conventional press for any suspension or cable screws. OLP and Cloning are very critical in applications like these, so as to avoid re-teaching of waterjet programs. Robotic cells are equipped with OLP and Cloning because they are necessary to perform the cutting required on hydroformed components.
'Customers need tools like off-line programming and cloning in order to make effective use of these flexible cells, and make them truly flexible so they don't have to reteach water jet programs over and over again,' says Traynor.
Calibration is the principal tool behind OLP and Cloning. 'One of the things we've been able to demonstrate to some of our users is the ability of our systems to basically program themselves,' explains Nieves. 'We've developed some intelligence in our software that can make some assumptions as to the angles we want to process a particular part. Basically, we define some vectors. Once these vectors have been defined with, for example, some semi-cylindrical components, we can place that model in our software and assume that we have 300 points that would have to be taught because of all the different contours that we have to process. We can superimpose a cylinder on that program, define the vectors towards the center of that cylinder and the trim line that has been defined by the program from a solid model, and that program will be taught to our controller.'
OLP, Cell Cloning, and Calibration have all been successfully implemented at Johnson Controls. Some of the findings at four JCI facilities indicated that correlation between positions in simulation and the Actual Robot Cell were within 0.25 mm at up to 1,200 mm./sec. JCI used a key technology of Realistic Robot Simulation (RRS), which allowed accurate representation of robot motion in simulation, which, in turn, allowed OLP of path-intensive applications at JCI facilities.
Engineers at ABB Automation (New Berlin, WI) also report success with OLP. 'Offline Programming also speeds up cutting of parts,' says Chandra Patel, director of business development at ABB Automation.
Other advances in cutting robots are focused on specialized systems for trimming of blow molded plastic parts, finish de-flashing of plastic parts mainly for the automotive industry, a custom-designed jewelry grinding and polishing cell, and robotic sanding of wooden kitchen door cabinets.
FANUC Robotics North America Inc. (Rochester Hills, MI) has developed certain robots for specific applications, patenting certain technologies in the process. Blow-molded plastic parts,like cooler lids or storage shed covers, and automotive molded fuel tanks can now be easily trimmed with FANUC Robotics' M-710i. The robot dedicated to this task does not require expensive sensors/vision to detect a parting line and can achieve a Class A finish. Only one pass is required to remove flash in most cases. The robot meets mold cycle time and handles part to part variation up to 12mm.
FANUC Robotics developed the deflashing and trimming process by using force control technology. The deflashing head mounted on the robot can adjust to changing part geometry automatically, thus removing flash consistently and uniformly.
'Another development from FANUC Robotics concerns gate removal on jewelry through a combination of lesion and force controlled techniques,' explains Laxmi P. Musunur, engineering manager for material removal and off-line programming at FANUC Robotics. 'The robot picks up the piece of jewelry or other part, presents the part to the camera which will then inform the system of the exact location of the gate, and also gives the height measurement, so that the robot will know how much material needs to be ground.'
Said to be the first of its kind, FANUC Robotics' LR Mate 200I robot and its visLOC vision system creates a flexible material removal application. The FANUC Robotics integrator that developed the system is Robotic System Integrations Inc.(Rapid City, SD). This jewelry grinding and polishing cell gives jewelry manufacturers the following benefits: higher throughput, consistent part quality and production flexibility. The robot cell also takes over monotonous and injury prone jobs and may help to reduce manufacturers' sensitivity to labor fluctuations dictated by market irregularities.
The VisLOC vision system is used to locate parts after they are being loaded onto a conveyor. Based on this position information, the LR Mate 200I robot picks up pieces with predictable consistency. The vision system is also used to measure part-to-part size variations. The measurement is then used to calculate a position offset, which, in turn, enables the robot to take away the same amount of material.
The LR Mate 200I is a sophisticated piece of machinery working in a highly expensive environment. 'We do screw removal after the casting of the jewelry piece,' says Vojislav Kalanovic, CEO of RSI, Rapid City, SD. 'There is a piece of metal that needs to be removed right off the piece itself, because that's left from the casting process. Another machine then does a contour grind of the jewelry piece, for example, a ring. After this robot has done the contour grind, it then hands it over to another robot, which does the inside grind.'
Cycle times on the jewelry robot system is only 5 sec., but that time would vary depending on the size of the ring or other jewlery and tbe amount of material to be removed, according to Kalanovic. Comparisons between man and machine can be invidious, but, on average, a manual grinder takes up to 20 sec., a manual operator takes about 4.5 mins., and the robot can accomplish a truly consistent product in about 5 secs., according to Kalanovic.
When asked about the benefits of a specialty robotic cell like the one described above, Kalanovic sketches three major benefits. 'You're going to have consistent quality, because robots are very consistent,' he says. 'You're also going to have the saving of the precious metal, because if a manual operator does the jewelry grinding, and if he starts with a size nine. He might finish with a size 10. The robot, however, will still end up with a size nine because of the grinding operation.
'And third, you're actually going to relieve the humans from very liability prone jobs,' continues Kalanovic. People are sitting there doing very monotonous work and they're exposed to gold dust and rouge dust, which are all health hazards.'
James F. Manji is a free-lance writer in Brnswick, OH.