Six-Axis Robot Selection: Posture, Reach and Tool Payload
Direct answer: Payload and reach are necessary, not sufficient. Three things decide the selection. Access: whether the wrist can still rotate inside a machine door with the gripper fitted. Posture: whether the tool holds the angle the process requires along the whole path, because reaching a seam is not welding it. Tool payload as a moment: end effector mass and its offset from the flange, checked against the load diagram rather than the headline figure. Fix the part and tooling first.
Who this is for: Written for engineers selecting a six-axis robot for machine tending, arc welding or handling where the end effector is not trivial.
Scope: This EVST guide covers the three things that usually decide a six-axis selection once payload and reach have been read off a datasheet: access at the point of work, posture along the path, and tool payload treated as a moment. It does not quote cycle times or accuracy figures for any specific robot.
Payload and reach are the easy half
Every selection starts with two numbers because they are the two numbers on the front page. They rule out the obviously wrong candidates and they rarely choose between the plausible ones.
The reference footage runs one arm family through three unrelated cells: loading a machine tool, arc welding a fabricated frame, and handling glass with a vacuum frame. In each case a different property is what binds the choice.
That is the useful observation. The binding constraint is a property of the application, not of the robot, which is why it cannot be read off a datasheet.
What six axis robot selection posture reach payload actually means
Reach is normally quoted to the wrist centre. The work happens at the tool centre point, which sits beyond it by the length of the end effector and usually off-axis as well.
Payload is normally quoted at a stated centre of gravity. A real end effector has its own mass distribution, and moving that mass away from the flange increases the moment on the wrist even when the mass itself is unchanged.
Posture is not quoted at all, because it is a property of the path rather than of the robot. It is the orientation the tool has to be held in, at every point where the process is active.
EVST checks all three against the real end effector, since that is the point where a datasheet stops being useful.
Access is about the wrist, not the arm
In the machine-tending cell the arm reaches through a door, into the working area, and back out. Whether the arm can get its wrist centre to that point is the easy question.
The question that binds is whether the wrist can still rotate once it is in there. A gripper that has to be presented at a particular angle needs clearance for the whole wrist assembly to turn, inside a space bounded by the door, the chuck, the tooling and any swarf guarding.
That is checked with the real gripper modelled, at the deepest reach, not at a nominal point in free space.
Posture decides process quality
In the welding cell the part is stationary and the arm sets the torch angle. Work angle and travel angle are process parameters, and they have to be held along the whole seam rather than only where the arm is comfortable.
This is why access and posture separate. A seam can be inside the envelope and still weld badly because the only pose that reaches it presents the torch wrongly, or because holding the correct angle drives a joint towards its limit part-way along the path.
The table below sets out which of the three usually binds in which kind of cell.
| Cell | What usually binds | What to verify first |
|---|---|---|
| Machine tending through a door | Access and wrist rotation | Wrist clearance at the deepest reach with the gripper fitted |
| Arc welding a fixed part | Posture along the seam | Work and travel angle at the worst point on the path |
| Handling a large flat part | Tool payload as a moment | End effector mass and offset against the load diagram |
| Mixed part family on one cell | All three, per variant | The worst case of each variant rather than the average |
Tool payload is a moment, not a mass
In the glass-handling cell the end effector is a full vacuum-cup frame. Thin, brittle sheet has to be held over an area, so the tool is necessarily large, and a large tool puts its centre of gravity well away from the flange.
The robot’s load diagram, not its headline payload, is what that has to be checked against. The same mass at a greater offset can exceed what the wrist can carry at speed even though the number on the datasheet looks comfortable.
Acceleration belongs in the same check. A tool that is within limits when moved slowly may not be when the cycle is run at the rate the application needs.
What only a real test piece can settle
Accuracy and repeatability figures are measured under defined conditions, and quoting them without those conditions makes them incomparable.
According to ISO 9283:1998, pose accuracy and repeatability are determined under stated load and velocity conditions, which is why a figure quoted without the tooling and speed it assumes cannot be carried into your application.
According to ISO 10218-2:2025, the integrated cell is what gets assessed, so the end effector and the machine it serves are inside the boundary rather than outside it.
According to ISO 12100:2010, the risk assessment starts from the tasks and hazards of the application, which is why the same arm in a different cell is a different assessment.
In practice an EVST application review starts from the same project inputs each time: part weight and geometry, the end effector concept, and the space the cell has to work in. The robot is chosen against those rather than the other way round.
Frequently asked questions
Is a higher payload rating always safer to specify?
Not necessarily, and it is not free. What matters is the load diagram at your tool’s actual centre of gravity and at the acceleration your cycle needs. A larger robot also needs more space, which can conflict with the access that bound the choice in the first place.
Why is reaching a seam not the same as welding it?
Reach is a position question and posture is an orientation question. The tool has to be held at the process angle along the whole path, and the pose that just reaches a point is often not a pose that can hold the angle.
Can a video show whether a robot suits our cell?
It can show the arrangement, the end effector type and the working envelope in use. It cannot show repeatability, cycle time or whether the wrist has clearance in your machine, which need a layout study and a test piece.
What should we send to get a useful answer?
Part weight and geometry, the end effector concept including its approximate mass and offset, and the space available including any opening the robot has to reach through.
Project inputs for an application review
To have a selection checked against your own part and tooling, send:
- part weight, geometry and how it is presented
- end effector concept with approximate mass and offset
- cell space, including any door or opening to reach through
- cycle requirement and the variants the cell has to cover
Send part weight, the end effector concept and the cell space and the selection can be worked through against them. Related reading: six-axis industrial robots, cell and line integration.