Welding sensors

Sensors for welding robot applications have unique advantages. Most technologies can be mixed and matched, and understanding sensor technologies can help robot users improve operations.
Editors note: Adapted from “Knowing When Welding Sensors Make Sense” blog post by Josh Leath, product manager for welding at Yaskawa America Inc., Motoman Robotics Division. 


Through-wire touch sense

The weld wire touches the conductive surface of a workpiece. A low-voltage circuit detects the weld joint. Slower than laser technologies, users must wait for the robot to move to the detection location.
Best for: Finding part orientation with simple joints, geometries
Not recommended for: Materials less than 3/16" with small joint thickness; Square butt-joints; Fast cycle times
Complexity: Low; Built-in pendant commands
Cost: Low


Through-arc seam tracking 

A sensor mounted near the power supply measures arc characteristics to determine variations between the robot’s taught path and the actual seam path – amending the programmed path during welding. Welding speed is regulated to 50ipm. 
Best for: Parts with long, curved seams; part-to-part variation
Not recommended for: Material less than 3mm; Non-weaving weldments; Weldments less than 6"; Large gaps 
Complexity: Low; Prewritten programs, algorithms provide easy operation
Cost: Low


Laser sensing 

A laser and sensor capture part location and orientation. Non-contact lasers are 2x-to-5x faster than touch sensing but cannot find square butt-joints and can struggle with reflective surfaces. Torch-mounted sensors 

can create access challenges.
Best for: Fast cycle times
Not recommended for: Square butt-joints; Large, inconsistent gaps; Reflective materials; Weldments with limited joint access
Complexity: Low/medium; Basic user training required with built-in commands
Cost: Medium

2D camera

A camera sensor mounted to the robot arm can quickly capture part location, confirm orientation, and identify the fixture to prevent job errors. More sensitive to lighting and surface conditions, stacked parts can be more difficult to program. 
Best for: Parts with higher placement variability; Demanding cycle times
Not recommended for: Applications with large depth, lighting, surface condition variations
Complexity: High; Additional user training required
Cost: Medium/high

3D camera 

Laser sensors and high-speed controllers process seam and part locations in real time. Compensation programs allow up to 100ipm simultaneous welding, enabling part automation with changing gaps. Power supply arc data combines with weld scans to track each part. 
Best for: Thin materials with varying seams; Very fast cycle times
Not recommended for: Wide gaps; Parts with limited joint access 
Complexity: High; Advanced user training required
Cost: High

Laser seam tracking 

A camera-driven robot system can automatically pick up a part, place it into tooling, weld 
using a robot, and remove the part. Complexity and cost increase compared to other solutions.
Best for: Randomly placed parts; Lights-out automation
Not recommended for: Simple jobs
Complexity: Very high; Advanced user training required
Cost: Very high