One Flexible Line, Five Different Robot Structures
Direct answer: One line rarely uses one robot type because each station constrains something different. If the motion stays in a single plane, a planar-jointed structure is sufficient and extra axes cost speed and stiffness. Once posture has to change it becomes a degrees-of-freedom question rather than a payload one. A protected environment filters models on ingress and materials rather than on load. And a feeding station is usually limited by supply pace, not by payload at all.
Who this is for: Written for manufacturing engineers specifying robots across several stations of a mixed assembly line.
Scope: This EVST guide covers the criteria that decide robot structure station by station: whether motion stays in one plane, whether posture has to change, whether the environment demands protection, and whether the limit is feed pace or payload. It does not quote cycle times, repeatability values or changeover times; those come from calculation and acceptance testing on your own line.
Five stations, deliberately unrelated
The reference footage puts five stations side by side. They are separate stations on separate parts from separate projects, shown together to compare selection criteria rather than to describe one continuous line.
Comparing them this way is more useful than a single line would be, because the criteria that separate them barely overlap. Each station is constrained by something the others are not.
The four criteria that appear are plane of motion, posture change, environmental protection and feed pace. Payload appears in none of them as the deciding factor, which is the point worth taking away.
When motion stays in one plane
The first station inspects parts carried on an index table under a backlight. The arm picks and places above the table, so the motion is a horizontal path plus a vertical stroke.
That kinematic profile is what a planar-jointed structure is built for: stiff and fast in the horizontal plane, with the vertical axis handled by a single actuator. Adding two more axes provides motion the station never asks for, and the cost is paid in speed and stiffness.
The test is straightforward. Between picking and placing, does the part’s orientation have to change? If it does not, a planar structure is the first thing to size.
When posture has to change
The second station feeds glass sheets into an edge-grinding line. The vacuum cups have to follow the angle of the sheet, and the attitude at pick differs from the attitude at place.
Once that is true, a planar structure is not merely slower – it cannot reach the required attitudes at all. This is a degrees-of-freedom question, and no amount of payload margin substitutes for a missing axis.
It is worth separating the two ideas explicitly, because they are routinely conflated: mass decides the size of the robot, attitude decides how many axes it needs. A light part can still require six axes and a heavy one can be fine with four.
The order EVST works in is the same each time: plane of motion, then posture, then environment, then feed pace – and the structure follows from whichever binds first.
| Station type | Deciding criterion | What to verify first |
|---|---|---|
| Planar pick and place under vision | Motion stays in one plane | Whether orientation changes between pick and place |
| Loading a machine with angled parts | Posture has to change | Attitudes required at pick and at place |
| Work inside a paint or coating booth | Environmental protection | Ingress rating, materials, and the atmosphere assessment |
| Continuous feeding of light parts | Feed pace and supply stability | Arrival consistency, not payload |
When the environment does the filtering
The third station sits inside a paint booth. The first question there is not speed or accuracy but whether the robot body tolerates continuous exposure to atomised paint and solvent.
Protected models answer that with sealing, surface treatment, cable encapsulation and in some cases internal pressurisation. None of those relate to payload or reach, but they eliminate most of a catalogue before performance is compared at all.
According to IEC 60529:1989+A2:2013, IP codes describe protection against solid objects and water ingress. They do not describe resistance to solvents and they say nothing about explosive atmospheres, so a booth application needs the materials compatibility and the atmosphere assessment named separately rather than an IP number alone.
The fourth station feeds small light parts to sewing positions continuously. Payload is not close to a limit there; what constrains it is the pace and consistency of supply, which is again not a robot datasheet property.
Where a flexible automation line actually gets decided
The fifth station stands beside an injection moulding line and hands trays to a screwdriving position. It is a handling task defined by the interfaces on either side of it rather than by the arm.
Taken together, the five stations show that structure selection is a per-station decision. In practice, standardising on one model across a mixed line optimises for spares and training, and pays for it in speed at the planar stations and in capability at the others.
Where standardisation is worth having, the useful version is usually narrower: one controller family, one set of interfaces, one safety architecture – with the structures themselves chosen per station.
Every EVST application review opens from the same project inputs – the part list, the motion required at each station and the environment – so selection rests on evidence and acceptance is agreed against your own line rather than a datasheet.
What belongs to testing, and safety
Cycle time, repeatability under load and changeover time come from calculation and acceptance testing. None of them can be read from footage, and none are quoted here.
According to ISO 9283:1998, pose repeatability is measured under stated load and speed conditions, so a repeatability figure quoted without those conditions does not transfer between stations.
According to ISO 10218-2:2025, the safeguarded space follows from the movement of the whole integrated system at each station, and a line built from several different structures has several different safeguarded spaces to define.
According to ISO 12100:2010, the interfaces between stations – conveyors, index tables, booths and feeders – belong in the same risk assessment as the robots themselves.
Frequently asked questions
Should we standardise on one robot model across the line?
Rarely at the structure level. Standardising helps with spares and training, but a mixed line has stations with genuinely different constraints, and one structure will be slow at some and unable at others. Standardising controller family, interfaces and safety architecture usually gives most of the benefit.
How do we decide between a planar structure and six axes?
Ask whether the part’s orientation has to change between picking and placing. If it does not, a planar structure is the first thing to size. If it does, the extra axes are a capability requirement rather than an upgrade.
Is an IP rating enough for a paint booth?
No. IP codes cover solid object and water ingress. Solvent compatibility of seals and coatings is a separate question, and any explosive atmosphere is governed by its own assessment. All three have to be named.
What should we send to get a useful answer?
The part list per station, the motion required at each – including whether orientation changes – the environment at each station, and the arrival rate where parts are fed continuously.
Project inputs for an application review
To have this checked against your own line rather than a generic layout, send:
- part list and weights per station
- motion required at each station, including orientation changes
- environmental conditions at each station
- arrival rate where parts are fed continuously
Send the part list and the motion required at each station and the structure selection can be worked through against them. Related reading: industrial robot product range, automated factory integration.