Collaborative robots: shared workspace, real risk assessment, slower cycles

The arm is not the safe thing; the application is

A force-limited arm restricts the energy it can transfer through its own structure. It has no opinion about the blade, hot tool, needle or sharp-edged casting mounted at the wrist, nor about the trapping point created when it moves a part towards a fixed table. Assessment therefore covers the complete installation: tooling, workpiece, layout, approach directions and the tasks people perform nearby. Occupational safety authorities in most jurisdictions are explicit that the duty applies to the assembly, not the component. Where the risk assessment ends up requiring guarding, sensing or a restricted zone, that is a legitimate result, but it changes the economics you assumed.

Safe speed is slow speed

Limiting contact energy means limiting velocity and acceleration, and the effect on cycle time is substantial rather than marginal. Many installations mitigate this by varying behaviour with proximity: full speed while the workspace is clear, reduced speed as someone approaches, stopped on close approach, using area scanners or similar sensing. That recovers much of the throughput and adds cost and complexity, which is exactly the point — you are rebuilding a graded safety system rather than avoiding one. Judge the business case on the speed you will actually be permitted to run, measured on the real path with the real tool, not on the free-space figure in the catalogue.

Where the format genuinely wins

The strongest cases are low-volume, high-mix work where a fenced cell cannot be justified and floor space is tight: machine tending in a jobbing shop, bench assembly steps, test fixture loading, laboratory sample handling, palletising small cartons at modest rates. The other real advantage is deployment speed. Lead-through teaching and simplified programming let a competent technician set up a new job in hours, which suits a shop where the arm moves between tasks. That flexibility, not the absence of fencing, is usually what makes these arms pay in a small business.

Easy to program is not the same as easy to keep running

Guiding an arm by hand through a sequence is genuinely quick, and it creates an expectation that the whole installation is equally simple. It is not. Gripper choice, part presentation, fault handling, integration with the machine being tended and the safety configuration all take the same engineering they would on any cell. Shops that buy an arm expecting it to be an appliance often find it sitting in a corner after the person who set it up moves on. Whoever will keep it running needs training, documentation and time, and that is a smaller investment than the arm itself but an easy one to skip.

People have to accept sharing the space

An arm moving unpredictably beside a bench changes how the person at that bench works, and their reaction determines whether the installation is used. Practical measures matter: predictable motion paths, clear indication of what the arm is about to do next, a reachable stop within the operator's normal posture, and involvement of the people concerned during layout rather than presenting them with a finished cell. Where an installation is imposed, operators find reasons to move it back or leave it switched off, and the equipment quietly stops being part of the process without anyone recording why.

Frequently asked questions

Does a force-limited arm remove the need for a risk assessment?

No. The assessment obligation attaches to the installation you create, and the arm is one component of it. Tooling, the workpiece being carried, the layout, and the tasks people perform in the vicinity all contribute risk that the arm's own limits do not address. Assessments frequently conclude that additional measures are needed, such as rounding tool edges, restricting the working zone, adding area sensing, or guarding a specific trapping point while leaving the rest of the workspace open.

Why did our shared-workspace arm end up behind a fence?

Usually one of two reasons. Either the assessment identified a hazard from the tooling or workpiece that could not be reduced any other way, or the reduced speed made the cycle uneconomic and fencing allowed full-speed operation. Both are common outcomes and neither means the purchase was wasted, but they do mean the original justification no longer holds. If full speed behind a fence is where you have landed, compare the arm against a conventional one before buying the next.

Are these arms suitable for high-volume repetitive work?

Rarely as the primary choice. Their advantages are flexibility, small footprint and quick redeployment, none of which matter much on a dedicated high-volume line where a fenced conventional arm will be faster, more rigid and often less expensive for the same payload. Where they do appear on high-volume lines it is usually for a specific task requiring proximity to an operator, or as an interim solution while a permanent cell is designed and built.