By CHRIS ANDERSON Welding Engineer, Technology Leader for Welding, Motoman Inc.

edited by RICHARD MANDEL, senior editor

A laser tracker added to a robotic seam weld system improves production speed without a reduction in weld quality.

The camera views the seam just ahead of weld torch, responding to distortion of a projected laser line.

Product manufacturers continually strive to find ways to robotically weld parts faster while maintaining consistent, high-quality welds. Over the years, various touch-sensing, seam-finding and seam-tracking technologies have been developed to help welding robots locate the welding seam and accommodate variations in joint gap and part fit-up.

Because traditional touch-sensing/seamfinding methods require the robot to physically touch different points on the part with the welding wire prior to starting the arc, and because seam-tracking technology required the robot to weave across the weld joint, such technologies were only feasible for material thicknesses greater than 0.125 in. (3 mm).

Quality was improved, but cycle times suffered — traditional touch-sensing typically required 3-5 seconds per joint, while the maximum travel speed with through-arc seam-tracking was limited to 35-50 in. (899-1,270 mm) per minute.

Coordinated motion and adaptive control capabilities
Improved laser sensors can now detect lapped joints less than 0.040 in. (1 mm) thick. The high sensitivity allows robots to weld sheet metal parts with characteristically variable part fit-up — examples include automotive frame components and exhaust systems, as well as food industry equipment such as sinks and appliances.

In addition to service as a welding aid for thinner-gage materials like aluminum and sheet metal, laser seam-tracking technology can be incorporated with coordinated motion of positioners to expand applications to include welding of fuel tanks, cylinders and other contoured vessels.

Joint gap information from a laser seam-tracking system can provide a robot an adaptive welding capability that, according to users, increases production rates. Without the feature, a robot has to be programmed to fill the worst-case joint gap for the entire length of the weld at a reduced speed. With an adaptive welding feature, the robot only slows its speed when the joint gap is encountered, so the average overall travel speed is faster. The robot controller will quickly respond by varying travel speed, weld settings (such as current and voltage), or even wire position in the joint on the fly in order to provide additional fill capability for the joint gap. These conditions can be set up in a table on the robot pendant to correspond with anticipated gap conditions.

Tracking technology
Laser diodes on the tracker package project a line of light onto the workpiece. Use of a single wavelength laser light makes the sensor relatively immune to interference from the bright welding arc or ambient lighting in the plant. The joint seam distorts the laser line.

A camera, mounted at an angle relative to the projected line, detects both the top plate and bottom plate location via the distortion. The camera is typically installed about 1 in. ahead of the welding torch, which allows it to preview the weld joint just prior to the welding arc. This "look-ahead" distance, along with the use of a protective glass over the camera lens and a positive airflow, all help protect the camera from the welding arc, spatter and fumes as well as high ambient heat.

The camera's software interprets the depth of the joint line in the camera's field of view, and can provide feedback on the gap of a lap joint by subtracting the known material thickness of the top plate. Additional software in the robot controller includes a buffer to delay feedback from the camera until the torch reaches the position viewed by the camera.

Laser cameras are available with different optics that affect their resolution and laser power. The proper camera can provide reliable tracking even on shiny surfaces such as aluminum and stainless steel.

At initial system set-up, the camera and torch need to be calibrated for the exact look-ahead distance to enable the robot to properly track contoured shapes, as well as coordinate with the adaptive control to vary travel speeds and weld conditions.

It is advantageous for maintenance purposes if the system can provide a display of target information from the tracking camera for trouble-shooting communication and calibration information. A gage with a machined step located in the cell is also recommended, with the robot programmed to "find" the fixed gage. The result of detecting this known location can be used to determine if the camera has been moved and requires recalibration. It is important that the robot's Tool Center Point (TCP) or torch position also be monitored to verify that it does not deviate.

Faster process speeds
One manufacturer reports reliably welding 0.100 in. (2.54 mm) thick lap joints, on steel transformer housings with heavy oil deposits, at speeds of over 60 in. (1.5 m) per minute, using a laser tracking system equipped with a high-power camera. The result has been improved quality and reduced rejects of housings with oil leaks at the weld joint.

Using 3D laser seam-tracking, welding speeds of up to 100 in. (2.5 m) per minute are easily obtainable. This rate is generally faster than most normal welding speeds. Additionally, developments currently under way will soon allow tracking speeds of 120 in. to 200 in. (3 to 5 m) per minute for high-energy processes such as laser-hybrid welding..

Cost justification
While laser seam-sensors are being successfully applied in production situations, the new technology comes with a hefty price tag. Laser seam-tracking can add $35,000-$50,000 per robot to the price of your welding system, depending on the type of camera used and level of support required. However, when higher welding speeds and reduced rework are factored in, actual payback on the system can be achieved in less than a year.

Laser seam-finders can be implemented for about half the price of a seam-tracking system. Short, intermittent welds generally do not have enough angular variation to require seam-tracking, but may benefit from seam-finding. Additionally, in many welding applications, the stamped components are consistent, but the location in which they are placed is not. Seam-finding helps the welding robot locate these parts. Seam-finding can combine multiple searches to offset entire seams or part frames. Unlike touch-sensing, no robot motion is required for joint variations of ±0.59 in. (±15 mm).

To maximize the success of any vision application, it is important that users attend training classes on the laser sensor as well as the robot system. Vision parameters need to be established in the sensor control in order to reliably detect the joint. Any loss of tracking data from the laser sensor will cause the robot to track out of the welding seam unless the robot is programmed to signal an alarm if this condition occurs.