Long Seams: Locate First, Then Talk About Reach
Direct answer: Steel structure long seam welding starts with holding the part, not with programming. Clamp position decides contact, contact decides whether the gap stays constant, and a wandering gap makes one program behave differently part to part. Access matters at the end plates and flanges, but it is answered inside the fixture decision rather than before it, which is why the clamping concept is settled first.
Who this is for: Fabrication engineers and production managers automating long structural seams on beams, girders and similar members.
Scope: This EVST guide is written from one filmed station: a long steel beam lowered into a row of locating clamps, with a robot welding along the seam and the part rotated where access runs out at the end plates. Several takes of that one station are used. Nothing here states a cycle time, a weld quality result or an acceptance decision for the filmed work.
Automation does not start at the program
In the reference footage a long beam is lowered into a row of locating clamps before anything else happens. That row is what gives the station a stable reference; without it there is nothing for a program to be written against.
This is where long seam projects most often stall. A program can be written to any level of detail, but if the part lands in a slightly different place each time, the result differs each time, and the difference is read as a programming problem when it is a fixturing one.
A practical way to judge a fixturing concept is to load ten parts in succession and look at the spread of where they land, rather than to check how precisely a single part can be set up.
According to ISO 13920:1996, general tolerances for welded constructions are stated against defined reference features, which is why steel structure long seam welding begins by establishing those features on the fixture.
Clamping decides contact, and contact decides the gap
Clamp position determines whether the workpiece actually sits against its locating faces. Where it does not, the joint gap varies along the length of the beam.
That variation is what makes one program behave differently on different parts: the same travel speed and the same wire feed produce different penetration and different bead shape when the gap underneath them changes.
So clamps belong where they hold the distortion, not where they are convenient to reach. Those two positions are frequently in conflict, because the convenient position is on the outside of the beam while the effective one is nearer the seam. The conflict is resolved in favour of the seam, and access for loading is solved through fixture design instead.
Contact can be measured rather than judged: feeler checks at a few defined positions along the beam give a number, and a gap outside the set band is corrected by moving clamps rather than by adjusting welding parameters.
According to ISO 5817:2023, imperfection limits apply to the finished joint, so in steel structure long seam welding a gap held by clamping is worth more than a parameter set adjusted to compensate for one that is not.
Holding attitude along the length
Once the part is stable the torch can travel the full length of the seam, and the requirement becomes consistency: the working attitude has to hold from one end to the other.
The difficulty of a long seam is not the arc start. It is whether the same angle and stand-off are maintained across several metres of travel.
The beam itself moves during the operation. Heat from the first part of the seam changes the geometry the second part is welded against, which is one reason attitude drifts even when the program is correct.
According to ISO 15614-1:2017, procedure qualification fixes the conditions a joint is welded under, and distortion during a long run is one of the conditions steel structure long seam welding has to keep inside that range.
Where access runs out at the ends
At end plates and flanges the arm’s access begins to run out. There are two responses: force an attitude at the edge of the working range, or rotate the part so the seam presents itself.
The reference station uses the second. Rotating the workpiece returns the torch to a comfortable attitude, which is more repeatable than operating at the limit of the envelope.
This is also why locating and access are one question rather than two: whether the part can be rotated at all is a property of the fixture, decided at the same time as the clamping.
Where the structure does not allow rotation, the answer moves to robot mounting or an external axis, and the cost structure of the proposal changes with it.
A decision table for a long seam station
Each row below is settled from the customer’s own drawing and parts. None of them is answered by a robot specification alone.
| Decision | Settled from | What goes wrong if skipped |
|---|---|---|
| Locating faces | Beam section, which surfaces are machined or true | No stable reference; results vary part to part |
| Clamp positions | Where distortion has to be held, not loading convenience | Gap wanders along the seam |
| Rotation capability | Structure at end plates and flanges | Torch forced to work at the limit of its range |
| Travel and attitude plan | Seam length, expected distortion during welding | Attitude drifts across the length |
| Inspection points | Weld quality level from the drawing | Deviation found only after the assembly is complete |
What the footage cannot show
Weld quality level, inspection method and whether this batch of beams is acceptable cannot be read from a moving picture. ISO 5817 defines the quality levels and ISO 17637 covers visual testing; both are settled against the drawing.
Heat input, interpass temperature and distortion compensation are equally outside what footage carries.
The segments used here are several takes of one station, not several production lines.
The order EVST works in is the same each time: establish the locating faces, place the clamps where they hold distortion, then answer access.
Frequently asked questions
Should the fixture or the robot be specified first?
The fixture. Once the part position, clamping and rotation capability are fixed, reach becomes a calculation. Specifying reach first leaves the fixture to fit around a constraint that was chosen without it.
How is a wandering gap identified before welding?
By checking contact at defined positions along the beam with feeler gauges. A gap outside the set band is a clamping problem and is corrected by moving clamps, not by changing welding parameters.
Is rotating the part always better than reaching?
Where the structure allows it, yes, because it returns the torch to a working attitude instead of the limit of its range. Where it does not, robot mounting or an external axis is the next option.
Does distortion during welding need to be compensated in the program?
It has to be accounted for somewhere. Sequence and clamping usually carry most of it; residual movement that remains is handled in the welding plan rather than absorbed silently.
Project inputs for an application review
To assess a long seam station, an EVST application review opens from the following inputs:
- beam length, section form, plate thickness and part weight
- joint type and seam layout, including whether both sides are welded
- existing fixture and clamping arrangement, and whether clamp positions can move
- structural limits at end plates, and whether the part can be rotated
Send the beam length, plate thickness and joint type, and we will work through locating and access together against your own seam list. Related reading: welding robot workstation presentation, smart robotic factory solution.