By EVST Editorial Team · Reviewed by EVST Editorial Team · Updated
Robotic Welding: Will Seam Tracking Solve Your Part Variation?
A seam-tracking option is worth evaluating when the joint moves away from the programmed path. It does not, by itself, establish that the welding process can handle every gap or fit-up condition in your parts. Before adding sensing to a robot quotation, separate where the joint is from what joint the process has to weld. That distinction can change whether the next investment belongs in the fixture, part preparation, or welding system.
What does the cell footage actually show?
The robot welds a plate assembly on a worktable, stops the arc, moves the torch and begins welding from another position. The pause makes the torch and surrounding plates easier to see than the bright arc does.

This is a useful starting point for a welding enquiry: several robot movements have to bring the torch to the required joints. The footage does not establish whether this cell uses touch sensing, seam tracking or adaptive welding. Those are separate equipment questions, not capabilities that can be inferred from a torch following a seam.
Our sample welded successfully. Why does the next part need adjustment?
First establish what changed. A correctly taught path can miss the next assembly because the entire part sits differently. Alternatively, the part can be in its intended location while its edges or assembly gap differ. Both problems may be described as “inconsistent parts,” but they call for different responses.
| Variation seen in the trial | First question | Where the result points |
|---|---|---|
| The whole joint is displaced after reloading | Does the part locate and clamp in the same way each time? | Correct the locating problem; evaluate a suitable position-finding function for residual variation |
| The joint position changes along its length | Can the proposed sensing method follow that joint during welding? | Evaluate tracking over the actual joint geometry and welding conditions |
| The joint is found, but its gap or fit-up changes | Can the welding procedure produce the required joint across that variation? | Improve part preparation/assembly, or qualify suitable process adaptation within defined limits |
These are diagnostic branches, not a way to identify a defect from video alone. A trial can reveal more than one of them. For example, an assembly may be loaded inconsistently and also contain variable gaps; solving its location does not eliminate the second problem.
Are touch sensing and seam tracking interchangeable?
They address different stages of locating the weld. In one common touch-sensing arrangement, the wire probes the workpiece before the program applies a position correction. Through-arc tracking instead uses feedback during a suitable weaving weld to adjust the path. The selected method must suit the joint and welding process.
For a buyer, the practical distinction is the demonstration to request. If an otherwise consistent assembly is displaced after loading, show how the proposed system locates it before welding. If the joint wanders along the weld, show the system following that variation while welding. A demonstration of one does not establish the other.
Ask the supplier to identify the sensing method in the quotation and show its complete sequence on your joint. Include the approach and search movements when comparing cycle time. A correction that helps the weld can still add work before the arc starts; its value should be evaluated with that time included.
If the robot finds the gap, can it simply fill it?
Finding a joint and producing an acceptable weld are different capabilities. Increasing deposition, changing travel speed or using another process strategy cannot be treated as an unrestricted remedy for poor fit-up. The chosen joint, material and acceptance requirement still define what the welding procedure must achieve.
For instance, suppose two assemblies present the seam at the same coordinates, but one has a visibly larger gap. Re-teaching the coordinates does not make their geometry identical. The useful comparison is whether the proposed welding procedure passes the specified inspection on both, and what preparation or parameter changes are needed. This is an application-testing example, not a claim about the filmed assembly.
Consistent part preparation and fit-up belong in the automation project. If the variation can be reduced economically before welding, compare that option with the extra equipment and process development needed to accommodate it.
What would make a trial useful for the purchase decision?
Start by reloading the same representative assembly and checking its location before welding. This helps distinguish a locating problem from variation between different assemblies. Then introduce production-representative parts with the fit-up differences that cause concern. Keep those differences identifiable, so a successful result is tied to an actual condition rather than to an unspecified “good sample.”

During that comparison, watch whether the torch and its services can reach the whole joint without clamps obstructing the approach. A clamp that improves location but blocks the weld exchanges one problem for another.
Interpret the result before selecting options. If reloading moves the joint, address the locating arrangement. If location is repeatable but the joint varies along its length, investigate suitable tracking. If the torch follows the joint yet the weld fails the agreed inspection, investigate fit-up and the welding procedure. The inspection method and acceptance criteria should come from the part’s requirements; a neat surface or a bright arc is not a substitute.
Bring the difficult joint into the discussion
For an EVST robotic-welding enquiry, include photographs or drawings of the joint in its normal and difficult fit-up conditions, together with material and thickness. This lets the discussion begin with the variation that matters and the options for part preparation, fixturing and sensing. Discuss your robotic-welding application with EVST.
Method: Application-footage review, manufacturer documentation and conditional trial-design examples. Editorial policy · Corrections policy · Terms of use