How to prepare parts for robotic welding: fit-up, gap and bevel
How far a part can deviate, what gap a fillet tolerates and how a bevel must be prepared for the robot to weld it right the first time.
A welder compensates by hand for what comes badly cut or badly fitted. A robot can compensate too, but within specific limits. Knowing those limits before buying avoids the most common surprise: the robot is there, but the parts coming from fit-up are not usable.
Part position and angle
With a laser scanner that searches for the joint, the system tolerates up to about 30 mm of difference between where the part should be and where it ended up, up to 8° of angle deviation in fit-up and up to 30 mm of difference in seam length. Without joint searching, that margin shrinks to what arc tracking absorbs: jigs and stops are recommended.
Gap between parts on fillet welds
The gap the robot can cover depends on fillet size: up to 3 mm when the leg is 8 mm or more, and up to 2 mm when it is between 5 and 8 mm. If the gap exceeds what weaving covers to one side, the weld fails. It is the point most worth controlling in fit-up.
Bevels and butt welds
On beveled butt joints, the recommended gap is 3 to 5 mm with the wire aimed at the center. The bevel must repeat within ±1 mm in size and 0 to 5° in angle, and not exceed about 25 mm in depth so the wire stick-out does not get too long. Tack welds inside the bevel must be at least 20 mm from the start of the seam and kept low.
Material and surface
Carbon and low-alloy steel from 6 to 30 mm is the ideal ground. Galvanized or polished parts reflect the laser, and aluminum has no arc tracking either. The part must arrive free of slag, oil and heavy rust around the joint.
Room for the torch
The robot needs to reach the joint at a good angle. On parts with cells or boxes, the weldable depth depends on the available width: as a reference, with a 36° torch, welding 200 mm deep needs about 150 mm of free width. Each case is confirmed with a simulation on the part's model.
