Direct answer: A robot reach and posture envelope is not a distance. Whether a seam can be welded is decided by the arm’s posture at every point along it, and posture changes along one seam then starts again at the next. Turning the part into position is often cheaper than a longer arm, and the margin that matters is margin for posture and safety.
Who this is for: Engineers selecting robots and external axes for welding cells on heavy plate weldments and similar multi-seam assemblies.
Scope: This EVST guide is written from one filmed station: a heavy plate weldment held by a single-axis positioner that turns it, with a robot welding successive seams from changing postures. Several takes of that one station are used. Nothing here states a cycle time, a repeatability figure or an acceptance result for the filmed work.
A radius is a number; access is a set
Selection tables give a working radius as one figure, and it is natural to compare candidates on it. On a real station that figure resolves into a whole family of postures, because the same distance reached from different arm configurations gives very different access.
Selecting on the figure alone tends to strand a few seams after installation: not out of reach, but reachable only in a posture from which the process cannot be run.
That failure is almost invisible at proposal stage, because drawings carry distances. It appears during teaching, when a particular angle turns out not to be enterable at all.
According to ISO 10218-2:2025, an industrial robot application is assessed as an integrated cell rather than as a machine on its own, which is why the robot reach and posture envelope is read from the seam list and not from a catalogue figure.
Moving the part instead of stretching the arm
In the reference footage a heavy plate weldment is held by a positioner and can be turned. That gives the first choice available on any multi-seam assembly: move the part, or stretch the arm.
The positioner brings one seam into a favourable position so the robot works from a comfortable posture. Compared with selecting a larger machine to reach the same seam awkwardly, it is usually both cheaper and more stable.
It also changes what has to be verified. With a positioner in the cell, the question becomes how many index positions the seam list needs, which is answered from the drawing rather than from a catalogue.
According to ISO 12100:2010, risk reduction begins with the design of the installation itself, so the robot reach and posture envelope and the safety zone are settled together rather than one after the other.
Robot reach and posture envelope along one seam
Even on a single seam the arm’s posture changes for the whole length of the travel. Access at the arc start says nothing about access at the middle or the end.
So access is verified along the trajectory, not at a handful of sampled points. A posture that exceeds limits anywhere on that trajectory means the seam has to be rearranged: either the positioner angle changes, or the seam is split and welded in two parts.
Splitting has a cost of its own. It adds an arc start and stop, and the resulting joint is one more place where quality has to be verified, so a seam that can be answered by rotation is normally not split.
According to ISO 5817:2023, every joint is assessed against the same imperfection limits, so in practice an extra start and stop is an extra place the robot reach and posture envelope has to be right.
A new seam is a new posture set
An assembly carries more than one seam. Moving to the next one starts the posture question again, and access has to be verified again with it.
The union of the posture sets for every seam is what the station actually requires. The margin worth keeping is margin in that union, plus the space the safety zone needs; it is not headroom at the furthest point.
In practice this is a table: seams reachable in the current configuration, seams that need the positioner turned, and seams that cannot be reached in any orientation. The third group is what decides whether an external axis or a different mounting is required, and it belongs in the proposal rather than in commissioning.
The safety zone is drawn from the same analysis. Swept volume at extreme postures differs from swept volume in normal working postures, and the enclosure has to be set from the former. Torch and cable pack outlines are added, because at extreme postures they frequently extend further than the arm itself.
A decision table for reach and posture
Every row below comes from the customer’s own assembly and seam list. None of them is answered by comparing published working radii.
| Decision | Settled from | What goes wrong if skipped |
|---|---|---|
| Seam list | Assembly drawing, seam positions and directions | Access assessed on distance instead of on seams |
| Rotation capability | Whether the part can be held and turned | A longer arm bought to solve a fixturing question |
| Posture check along each seam | Trajectory, not sampled points | Seams reachable at the start and blocked mid-way |
| Index positions | Union of posture sets across all seams | Positioner angles improvised during teaching |
| Safety zone | Swept volume at extreme postures, torch and cable outline | Enclosure sized from normal postures only |
What the footage cannot show
Interference checking is not visible in a moving picture. It is run seam by seam in the 3D model and includes the torch, the cable pack, the fixture and the positioner, not only the arm.
Cycle time, repeatability and acceptance results are equally outside what footage carries; ISO 10218-2 and ISO 12100 cover integration and risk assessment, and neither is settled by watching a cell run.
The segments used here are several takes of one station, not several cells.
The order EVST works in is the same each time: build the seam list, decide whether the part turns, then verify posture along every seam before any machine is selected.
Frequently asked questions
Is a larger robot the usual answer to a seam that cannot be reached?
Rarely. Most unreachable seams are a posture problem rather than a distance problem, and turning the part answers them at lower cost and with more stable results.
How many index positions does an assembly need?
As many as the union of posture sets requires. It is read off the seam list, not taken from a comparable project.
Does inverted or gantry mounting simply increase reach?
No. It changes the shape of the usable posture range rather than scaling it, so it has to be checked against the same seam list.
Can the safety enclosure be sized from normal operating postures?
No. Swept volume at extreme postures is larger, and torch and cable outlines have to be included, so the enclosure is sized from the worst case.
Project inputs for an application review
To assess reach and posture for a welding cell, an EVST application review opens from the following inputs:
- assembly envelope dimensions, weight and centre of gravity
- seam layout with the position and direction of every seam
- whether the part can be turned, and the acceptable range of rotation
- existing robot model and mounting arrangement, and the site safety boundary
Send the part envelope and the seam layout, and we will work through the posture set seam by seam before any machine is proposed. Related reading: large workpiece robotic welding pose and reach, automatic welding system solutions.