Teams often spec a robot cell from the end-effector backwards — pick the gripper, then find a robot that fits. The expensive mistakes show up later: a robot that cannot reach the far fixture, or a payload margin so thin that a heavier part tips the cell into fault. Here is the order we size in.
Start with the load, including the "invisible" mass
Payload is not just the part. Add the gripper, any tooling, cables and the worst-case part weight, then apply a safety factor (commonly 1.2–1.5×). The moment of inertia at the wrist matters as much as mass for long or off-centre payloads.
| Factor | Often forgotten |
|---|---|
| Gripper + cables | 1–5 kg on a small arm |
| Part mass variation | A "2 kg" part can be 2.4 kg with fixtures |
| Inertia | Long parts swing the wrist even within payload |
| Speed vs. payload | Higher speed needs more margin |
A cell quoted at 90% of rated payload leaves no room for a heavier variant or a worn cable.
Reach is a working envelope, not a number
A robot's stated reach is the maximum from the base. Real working reach is smaller once you account for:
- The fixture and part height — the wrist often needs to come down into a jig, not just out to the edge.
- Approach angle — some welds or pick points need a near-vertical entry the arm cannot hit at full extension.
- Singularity avoidance — motions near the arm's singular posture slow down or fault.
Always validate the full motion path in the cell's actual layout, not in a catalogue reach diagram.
Axes and mounting
- 6-axis articulated: the default for flexibility; watch wrist singularities and floor-space footprint.
- SCARA: excellent for fast, horizontal, pick-and-place in a plane; useless for out-of-plane work.
- Delta: high-speed light payloads above a fixed point; limited reach and payload.
- Gantry/Cartesian: large working volume, very stiff, but needs floor space and a frame.
Mounting (floor, wall, inverted) changes the envelope dramatically and the cable-management risk.
The integration details that burn schedules
- Cables and dress packs — the most common unplanned downtime source. Specify routing, minimum bend radius and a service interval.
- Vision/alignment — if parts are not fixtured perfectly, the cell needs a locating step (camera or touch probe) that must be sized and timed.
- Safety — light curtains, interlocks and reduced-speed zones add cycle time; model them honestly.
- Controller and I/O — confirm the fieldbus matches the line PLC before ordering.
Acceptance: size against the cycle, not the brochure
Write the cycle-time target into the acceptance test with the real part, the real gripper and the real fixture. A cell that meets spec with an empty gripper and a light dummy part will disappoint on go-live.
FAQ: should we buy headroom in payload?
Yes, modestly. 20–30% payload and reach headroom is reasonable; beyond that, a larger frame is usually better value — provided the layout supports it. Over-sizing the whole arm for one edge case wastes money and floor space.
Wrapping up
Size the cell from load (plus tooling and inertia), then working reach in the real layout, then axis type and mounting. Treat cables, vision and safety as cycle-time contributors, and accept against the real part. That prevents the late-stage "the robot can't quite reach" surprise.
For robotic cells, fixture design and process qualification, see TrueSyn Robotic. Company and application details are at truesynrobotic.com.
Top comments (0)