Two Robots, One Member: Interference Before Cycle Time
Direct answer: Two robots on one workpiece share a part, a fixture and a time window, so output does not simply double. The first thing to settle is interference: where the working envelopes overlap, motion has to be staggered in time or the seams divided in space. Every positioner angle then has to be checked for both arms, and the segment plan has to spread arc-out points so heat does not accumulate in one zone.
Who this is for: Manufacturing engineers evaluating whether a second welding robot on the same workpiece is justified, and what has to be settled before it is.
Scope: This EVST guide is written from a single filmed station: two robots welding opposite sides of one long structural member carried between a single-axis positioner and a fixture table. It is one continuous observation of one station, so no cycle time, throughput gain or acceptance result is stated for the filmed work.
Two arms is not twice the output
Replacing one robot with two suggests doubled output. In practice the two share one workpiece, one fixture and one time window, and whether they can move simultaneously is conditional.
In the reference footage the two arms weld opposite sides of the same long member, each running its own seams and neither handing work to the other. That division is the most robust form of twin-robot working.
Where two arms have to alternate on one seam, or relay within one region, coordination complexity rises sharply and the benefit becomes uncertain.
Multi robot cycle coordination starts with interference
Two arms serving one part necessarily have an overlapping envelope. Only one of them can occupy that region at a time, and that is resolved during path planning rather than discovered during commissioning.
There are two ways to resolve it: stagger in time, at the cost of waiting, or divide in space, at the cost of fixing which arm owns which seams. Mixed schemes are common – divide the bulk, stagger the few seams that cross.
In the footage the envelopes clearly overlap and the arms are separated by timing, which means their programs were written together rather than independently.
According to ISO 10218-2:2025, a robot application is assessed as an integrated system, so two arms sharing one workpiece are a single safeguarding and interference problem rather than two separate installations.
| Question | What decides it | What to prepare |
|---|---|---|
| Can both arms work at the same time? | Whether the seams divide cleanly by side or region | A seam list with side and length |
| How is the overlap resolved? | Stagger in time, or divide in space | A path plan written for both arms together |
| How many index angles are needed? | The union of what both arms require | A reach check at every angle for both |
| Will the member stay straight? | Segment plan and distribution of arc-out points | Section form, thickness and straightness requirement |
One index moves both envelopes
When the positioner turns the workpiece, the spatial relationship changes for both arms at once. Every index angle has to be verified for both.
Checking only one arm is how a project discovers, on the floor, that at some angle the other cannot reach or the two collide. Corrections at that stage are expensive.
The number of index positions is therefore the union of what both arms need. If that union grows large, the seam distribution is probably not suited to twin-robot working in the first place.
Segments and arc-out points
Long members are not welded end to end. Segmenting spreads heat and distortion, and the segment plan largely decides whether the member is still straight afterwards. ISO 13920 provides the tolerance framework for judging that.
With two arms working simultaneously, the arc-out points also have to be distributed. If both finish in the same region, heat accumulates there and the member pulls to that side.
Segment length follows the section: smaller sections that shed heat slowly need shorter segments and longer intervals between them.
According to ISO 13920:1996, straightness and angular deviation on welded constructions are specified as tolerance classes, which is how the target for a long member should be written before the segment plan is fixed.
According to ISO 12100:2010, the assessment has to cover setting, loading and fault finding as well as automatic operation, which for two arms includes what happens when only one of them stops.
Decide from the part, not from the budget
What justifies a second arm is the seam length, its distribution and the target rate – not whether the budget stretches to two robots.
Where seams concentrate on one side, a second arm contributes little. Where they are distributed across two sides or two ends, the benefit is visible.
A rough but useful test: split the seams into the two reachable groups. If the total lengths differ by more than about a third, utilisation of the second arm will stay low.
What the footage cannot establish
Actual cycle time, throughput gain and how much of the cycle is spent waiting for the other arm are measured on the floor, not read from footage.
Weld quality, distortion and inspection outcomes come from procedure and measurement, with ISO 5817 defining quality levels.
The material used here is a continuous observation of one station and does not represent the output of a line.
The order EVST works in is the same each time: divide the seams, resolve the overlap, then check every index angle for both arms.
Frequently asked questions
Will a second robot halve my cycle time?
Not by itself. The arms share one part and one time window, so the achievable gain depends on how evenly the seams divide between them and how much waiting the overlapping envelope forces.
Should the two arms share a seam?
Usually not. Relaying on one seam adds tie-in handling and parameter matching between arms; dividing the seams by side or by region is simpler and more robust.
How many positioner angles will twin-robot working need?
The union of what both arms require. If that union is much larger than what one arm would need, the seam distribution is probably better served by a different station layout.
Why do arc-out points matter with two arms?
Because both arms deposit heat into the same member. If their finishing points cluster in one region, the accumulated heat pulls the member toward that side even when each individual weld is sound.
Project inputs for an application review
To judge whether a second arm is justified, an EVST application review opens from these inputs:
- member length, section form, plate thickness and material
- total seam length and its distribution across sides and ends
- target rate and the measured time of the current single-arm method
- existing fixture and positioner configuration, and available index angles
Send the member length, thickness and seam layout, and we will work through interference, index angles and heat distribution before quoting a second arm. Related reading: robot welding workstation, smart robotic factory solutions.