Robot Payload and Reach: Count the Tool and the Layer
Direct answer: Rated payload is not the product weight. It is the total mass at the wrist flange, including the gripper, quick changer and service lines, and it is bounded by an allowable moment as well as an allowable mass. Robot payload and reach is therefore one combined check, resolved at the far top corner of the final layer.
Who this is for: This guide is written for manufacturing engineers and buyers sizing a robot for palletising or full-layer handling, where the load is not a single item and the stack grows through the cycle.
Scope: It covers how gripper mass, centre of gravity and the geometry of the final layer decide whether a given arm is adequate, and what each judgement still owes in evidence. It does not cover conveyor design, pallet quality, stretch wrapping, or the products being handled.
What rated payload actually refers to
A datasheet payload figure describes what the wrist flange is permitted to carry, not what the product weighs. Everything mounted beyond the flange counts: the gripper frame, the vacuum or clamp elements, the quick changer if one is fitted, valve blocks and the section of hose and cable that travels with the tool. On full-layer handling these items are rarely trivial, because the tool has to span the whole layer.
The second half of the rating is usually printed nearby and read less often. Manufacturers publish an allowable moment at the wrist alongside the allowable mass, because a load hung further from the flange applies more torque for the same weight. A wide layer gripper is, by construction, a long lever. It is entirely possible to sit inside the mass limit and outside the moment limit at the same time.
Acceleration is the third term. A payload that is comfortable at low speed can exceed the permitted wrist load during aggressive acceleration or deceleration, which is exactly where cycle time is usually won. Sizing that ignores the motion profile tends to produce an arm that is adequate on paper and marginal in the cell.
In practice the figure that decides a project is not the headline payload but the load diagram read at the real offset. According to ISO 10218-2:2025, the robot application rather than the robot alone is what has to be assessed, and robot payload and reach behave the same way: they are properties of the installed cell, not of the arm in isolation.
The gripper is part of the load, and it sets the geometry
A full-layer tool has to hold every item in the layer reliably at pick, keep the layer intact through the move, and release it so the items land in position. Those three requirements pull the design in different directions, and they generally make the tool heavier and wider rather than lighter and more compact.
The width has a second consequence that is easy to miss during selection. A large tool reduces the poses the robot can actually use, because the tool itself becomes an obstacle near the stack, near the conveyor guard and near any structure above the pallet position. The usable working volume is smaller than the volume the arm can technically reach.
For that reason the tool concept belongs in the sizing exercise, not after it. Even a preliminary mass estimate and an approximate centre-of-gravity offset are enough to tell whether the candidate arm is comfortable or borderline, and it is far cheaper to discover that on paper.
Centre of gravity and allowable moment
Robot manufacturers express the permitted load as a combination of mass and the distance of its centre of gravity from the flange, often as a load diagram rather than a single number. Reading that diagram with the real tool geometry is the only way to know whether a selection is valid, and it is a different exercise from comparing headline payload figures between models.
Full-layer tools tend to place their centre of gravity well forward, because the gripping elements are distributed across the layer footprint rather than concentrated near the flange. Two tools of the same mass can therefore produce very different wrist loads, and the lighter-looking option is not automatically the safer one.
Where the load diagram shows the selection sitting near a boundary, that is a result to carry into the specification rather than to round away. It usually points either to a lighter tool concept or to the next arm in the range, and both are cheaper decisions before purchase than after.
Decision table: what the layer pattern justifies
The table maps what a layer pattern and stack geometry can justify on their own against the evidence a supplier still has to produce.
| Evidence from the application | Selection it justifies | Evidence you still owe |
|---|---|---|
| Single layer pattern, one product, modest stack height | A single arm sized on layer mass plus tool mass, with the tool concept fixed early | Load diagram checked with the real centre-of-gravity offset, and the motion profile the cycle actually needs |
| Mixed products with different layer masses and stack heights | Sizing driven by the heaviest layer and the tallest stack, not by an average case | Pose-by-pose reach check at the far top corner of the tallest configuration |
| Pallet position constrained by conveyor, columns or guarding | Layout review before arm selection, since usable volume is smaller than reachable volume | Clearance study with the real tool envelope, including service lines at full extension |
| Layer pattern still under discussion | No final arm selection; a shortlist with the sensitivity of each option recorded | Confirmed layer pattern, product mass, and the target stack height before commitment |
Reach is bounded by the last layer, not the datasheet radius
The published reach of an arm is a radius measured in free space. The reach a palletising cell needs is a set of specific poses at specific heights, bounded by the floor, the pallet, the conveyor and any guarding around them. The two are related but they are not interchangeable.
The demanding position is almost always the far top corner of the final layer, because distance from the base, height above the pallet and a constrained approach direction all arrive together. Sizing against an average position quietly assumes the easy case and defers the hard one to commissioning.
Where more than one product is handled, the check has to be run against the tallest stack the cell will ever build. Leaving headroom in the selection is cheaper than discovering after installation that the last layer cannot be placed with an acceptable tool orientation.
Pose limits inside a real cell
Manufacturers state allowable payload for defined wrist orientations. In some poses the permitted load is lower than the headline figure, and full-layer handling frequently requires exactly those poses, because the tool has to stay level while the arm is extended and elevated.
Service lines add a further limit. A hose and cable package that is comfortable in the middle of the working envelope can reach its bend radius at full extension and height, and that limit is mechanical rather than programmable. It shows up as wear rather than as an error message.
The practical test is to review the worst combination of extension, height and orientation together, rather than each separately. If the combination has not been checked, the selection has not really been checked.
According to ISO 9283:1998, published performance figures for manipulating industrial robots are measured under defined test conditions, so they describe the manipulator and not the placement result of a wide layer tool. EVST treats those figures as a comparison basis between arms, never as a predicted stacking accuracy.
Safeguarding and access around a palletising cell
A palletising cell has a moving load at height, a pallet exchange that requires human or truck access, and a conveyor interface. Those three facts drive the safeguarding concept, and the concept in turn constrains where the robot can be placed and how large the tool may be.
ISO 10218-1:2025 and ISO 10218-2:2025 cover the robot and the robot application respectively, and ISO 12100:2010 sets the general risk assessment framework. Pallet exchange in particular deserves an explicit decision early, because retrofitting access to a completed layout is disruptive.
Repeatability figures published under ISO 9283:1998 are measured under defined conditions. They describe the robot, not the placement accuracy of a layer through a wide tool, and they should not be quoted as an expected stacking result.
According to ISO 12100:2010, risk reduction begins at the design stage rather than at the guarding stage, which is why pallet exchange access belongs in the layout discussion. When EVST reviews a palletising layout, the pallet exchange method is settled before the arm is fixed, because retrofitting that access is the change that disrupts everything else.
What to put in the capital request
State the layer pattern, the mass of a single item and of a complete layer, the pallet position and the target stack height, including the tallest configuration the cell will ever build. Add the preliminary tool concept with an estimated mass and centre-of-gravity offset.
Then state what still has to be proven by the supplier: the load diagram evaluated at the real offset, a pose check at the far top corner of the tallest stack, a clearance study with the real tool envelope, and the safeguarding concept for pallet exchange. A request written this way is comparable across suppliers, which a headline payload comparison is not.
Frequently asked questions
Is the rated payload the weight we can palletise?
No. It is the total mass permitted at the wrist flange, which includes the gripper, the quick changer and the service lines travelling with the tool, and it is limited by an allowable moment as well as by mass. On full-layer handling the tool is a significant part of that total.
Can we compare robots on payload alone?
Not reliably. Two arms with the same headline payload can differ in allowable moment and in the payload permitted at particular wrist orientations. The comparison that matters is each candidate evaluated against your real tool geometry and the poses your cycle requires.
Where should reach be checked?
At the far top corner of the final layer of the tallest stack the cell will build. That position combines maximum distance, maximum height and a constrained approach, and it is where an otherwise adequate selection usually fails.
Does the footage show what payload we need?
No. The footage shows a robot lifting a complete layer with a wide end effector and building a stack. It is evidence of a visible process. Any payload, reach or cycle figure has to come from your own layer pattern, masses and layout.
Project inputs for an application review
Send the following and the selection can be reviewed against the real application instead of a headline payload figure:
- the layer pattern, with the mass of a single item and of a complete layer
- the pallet position, pallet size, and the target stack height for every product
- your preliminary tool concept, with estimated mass and centre-of-gravity offset
- the cell layout, including conveyor position, columns, guarding and pallet exchange access
- the cycle requirement, so the motion profile can be checked rather than assumed
If you are sizing a robot for palletising or full-layer handling, send the layer pattern and masses, the pallet position and target stack heights, your preliminary tool concept with its estimated centre-of-gravity offset, the cell layout including access for pallet exchange, and the cycle requirement. The selection can then be checked at the worst pose rather than at an average one. Related reading: robot product range, wider robotic factory integration, robot workstation configuration context.