How to scope your first cobot cell without buying the wrong arm.

The brochure version of a collaborative robot is simple: bolt it to a bench, teach it a few points with your hands, and let it work beside your people. The version that shows up on a real shop floor is messier. The arm is usually the easiest part of the project. What decides whether the cell runs is everything around it: how parts arrive, how they leave, what the gripper has to hold and what the risk assessment says about the people nearby.

That is good news for a first-time buyer, because those things can be scoped on paper before anyone signs a purchase order.

Start with the task, not the robot

Pick one job that is repetitive, steady in volume and boring enough that nobody will miss it. Machine tending, palletizing finished cartons and simple pick-and-place between two fixtures are the classic first cells for a reason: the motions repeat, the parts are known and the cycle time is easy to measure.

Then write the job down as a sequence. Where does the part come from, in what orientation, how heavy is it, where does it go and how long does a person take today? That one page is worth more than any spec sheet, because it tells you which arm you need rather than the other way round.

1. Measure the part and the motion

   Record the heaviest part plus the gripper, the longest reach the motion needs and the cycle time a person achieves today. Payload and reach ratings are maximums, so leave a healthy margin on both.

2. Solve part presentation

   Robots are poor at finding parts in a jumbled bin. Fixtures, trays or a simple chute that present every part in the same place and orientation are cheaper than a vision system and far more reliable.

3. Choose the gripper early

   End-of-arm tooling often costs as much as a month of debugging. Test grip and release on real parts, including the oily, bent and out-of-tolerance ones, before you commit.

4. Run the risk assessment before you buy

   Collaborative operation is a property of the whole application, not the arm. The assessment may limit speed, add scanners or require guarding, and that changes your cycle time.

5. Plan the hand-off

   Decide who restarts the cell after a fault, who changes the program for a new part and who owns spare grippers. A cell nobody owns stops within a month.

Why collaborative does not mean unguarded

Standards such as ISO 10218 for industrial robots and ISO/TS 15066 for collaborative operation describe how a robot and a person can safely share space, usually by limiting the speed and force the arm can apply. In practice that means a cobot working right next to an operator often runs slower than the same arm would behind a fence.

“

The arm was the cheapest decision we made. The tray it picks from took three weeks.

If the cycle time only works at full speed, plan for an area scanner that slows the arm when someone approaches, or for light guarding around the parts of the cell where people do not need to be. Neither is a failure of the cobot idea. They are simply part of the cell.

Budget for the cell, not the arm

A useful rule of thumb from integrators is that the robot itself is often well under half of a first cell's total cost. The rest goes on tooling, fixtures, safety devices, a stand or base, programming time and the hours your own people spend learning to run it. Quotes that only price the arm are quoting the easy part.

Scoped this way, a first cobot cell stops being a bet on a robot and becomes a small, measurable change to one job. That is the version that pays for itself, and the version that earns you a second cell.