You pick the robot by reach: the farthest pickup point is 1300 mm from the base, so you buy a 1400 mm robot. It fits on paper. On the floor, the wrist won’t reach because the joint five configuration fights you, the end-of-arm tool collides with the fixture, and the maintenance guy can’t get to the teach pendant.
The reach envelope is a sphere, not a table
Robot reach is specified as the maximum horizontal distance from the joint-1 axis to the wrist center. That’s a single number on a catalog. The actual usable workspace is a distorted sphere: reachable at shoulder height, cramped near the floor, and almost nonexistent directly behind the base. A point at the edge of the envelope may only be reachable at one wrist orientation. Move 50 mm and the robot can’t get there without flipping a joint.
When you lay out a cell, plot the actual points the tool needs to reach, at the actual tool orientation, in the robot’s offline simulation. Don’t just check distances. Check joint angles. If a point forces joint 5 past 170 degrees or joint 2 near its limit, you’ve bought yourself a singular position, and the robot will slow down or fault there every cycle.
The tool changes the envelope
The catalog reach is to the wrist flange. Your gripper extends another 200 mm. That extension helps reach farther, but it also changes the moment of inertia and pushes wrist joints near their limits. Worse, the tool itself can collide with the fixture when the wrist rotates. A reach that works at the flange may not work once the gripper is attached and the part is oriented correctly.
Simulate the complete cell: robot, tool, part, fixture, conveyor, and safety fence. If the simulation doesn’t show the tool and the part, it isn’t a layout; it’s a wish.
Maintenance access is part of the layout
Integrators design the cell around the robot’s motion. They forget the person who has to change the gripper fingers, reload the feeder, or clear a jam. A cell packed tight against the fence looks efficient on paper, but when the line stops at 2 AM, the maintenance technician can’t get to the valve bank without climbing over the safety barrier.
Leave 600 mm behind the robot for encoder battery changes, cable dress pack inspection, and grease fittings. Leave a door or a removable panel at the tool change position. If the cell has to be opened every shift for part presentation, that access should be a listed design requirement, not an afterthought.
Safety fencing consumes space
The reach envelope plus the safety distance is the cell footprint. A robot that reaches 1400 mm needs a fence at least 1400 mm from the base in every direction, plus the minimum safety distance based on the stop time. That easily pushes the fence to 1800 mm or more. If you sized the cell to the robot footprint, you’ve already run out of floor space before you add the conveyor and the operator station.
Use a safety scanner instead of a fixed fence where the application allows it. It shrinks the cell footprint by replacing a physical barrier with a monitored area, and it lets the operator load parts from the same side without opening a gate. But it requires a slower approach speed when someone enters the monitored zone, which trades cycle time for floor space.
Light curtains vs gates
For part loading, a light curtain across the load station lets the operator reach in without opening a gate. That’s faster than a gate with a safety interlock, which requires a full stop and restart. But light curtains don’t protect against a robot reaching out toward the operator. If the robot’s workspace overlaps the load zone, a gate is safer than a curtain, because the gate physically blocks access during motion.
Pick the guarding based on what the robot is doing when the operator is there. Loading a stationary robot is different from loading a robot mid-cycle.
The floor slope and bolt pattern
This sounds trivial, until the robot is bolted down and the floor has a 1 degree slope. The base isn’t level, joint 2 fights gravity, and the accuracy drifts. Anchor the robot to a machined plate or a weldment that gets shimmed level during installation. Don’t assume a 10-year-old factory floor is flat enough to bolt directly to.
Bottom line
Layout a robot cell from the simulation outward, not from the catalog number inward. The reach number is the starting point; the joint limits, tool orientation, maintenance access, and safety distance are what actually decide the footprint. If the cell doesn’t fit in the simulation, it won’t fit on the floor.
