Cobot versus industrial robot: choosing the whole cell

Separate the robot label from the operating arrangement

The familiar cobot-versus-robot comparison mixes two decisions: which arm and controller to buy, and how people will interact with the finished application. An arm marketed as collaborative may work inside guarding because its tool, part or process needs separation. Conversely, ABB describes SafeMove as supporting collaborative arrangements with other industrial robots and compatible controllers. That manufacturer capability is a reason to investigate a design, not a guarantee about a proposed installation.

Write two concrete cell concepts before ranking them. Identify the robot model, tool, input and output stations, protective measures and operating modes in each. State where people are expected to stand and what they need to do during production. Comparing a bare cobot with a fully engineered conventional cell produces an attractive but misleading purchase comparison because the boundaries differ.

References: SafeMove robotic safety solution

Map human access by task and frequency

Observe replenishment, unloading, quality checks, tool changes and recovery. Record why each person approaches the process, how often and for how long. Distinguish access to an external loading drawer from entry into the robot's working area. A process that only needs occasional replenishment may suit a different arrangement from one requiring a person to handle each assembly between robot operations.

Then ask whether the interaction can be redesigned. A second fixture or exchange station may allow useful separation without making the operator's work awkward. A supposedly open layout can still lose output when normal pedestrian traffic repeatedly triggers its protective response. Review material routes and adjacent workstations with the people who use them. The practical question is how production and human work coexist over a shift, including busy periods.

Review what the process brings to the cell

Bring the actual workpiece and tool to the safety review: edges, gripping surfaces, hot material, stored energy and potential dropped objects matter. Universal Robots' UR5e PolyScope 5.19 documentation requires assessment of the application, including its tools and connected machinery. Ask a competent integrator to establish the required protective measures and validation for each proposed concept before using estimated production speeds in the comparison.

Keep the process hazards visible even when the arm has contact-sensitive features. A welding torch, rotating cutter or heavy unstable pack creates questions beyond contact with a robot link. Equally, fencing does not answer every setup or maintenance question. The review should define allowed modes, authorised access and restart arrangements. Purchasing staff should obtain the resulting design evidence, rather than choosing their own force, speed or separation values.

References: UR5e Risk Assessment, PolyScope 5.19 manual

Compare output under the proposed operating conditions

Request representative timing with the intended tool, parts and protective configuration. Separate robot motion from grip acquisition, machine waiting, inspection and operator interaction. A robot with higher nominal speed may produce little additional output when another machine dominates the cycle. A cobot concept may meet demand comfortably even if its isolated transfer takes longer. The comparison needs accepted production over the period that matters to the workshop.

Include the effect of ordinary access in the test or validated timing model. State whether replenishment pauses production, whether another station keeps working and how long authorised restart takes. Supplier estimates should identify their assumptions. Do not combine the fastest motion from one configuration with the most convenient access from another; those results may not be achievable together in the installed cell.

Use a comparison that can reveal either winner

Use the same requirement rows for both concepts and demand evidence against each. The table is a buyer's evaluation structure, not a scoring system that predicts which architecture is safer or cheaper. Mark unanswered requirements as unresolved rather than assigning an optimistic average. If a material constraint fails, such as inaccessible workholding or inadequate output, resolve it before calculating a weighted preference score.

ABB's SafeMove page identifies controller support as part of its offering, including IRC5 and OmniCore families. It does not establish that every robot, option and controller combination is supported. This is a useful example of why a proposed safety capability needs a named configuration. Apply the same discipline to hand guiding, simulation, external axes and tool interfaces: purchase the demonstrated option set, not an assumed family-wide feature.

References: SafeMove robotic safety solution

Work through a hypothetical access tradeoff

Imagine a fictional packing task requiring 240 accepted transfers per hour. Concept A has a 10-second transfer and loses 15 minutes per hour to its assumed access and replenishment pattern. With no rejects, that leaves 270 transfers. Concept B has an 8-second transfer and loses 5 minutes per hour under a different proposed loading arrangement, leaving approximately 412 completed transfers after rounding down. These are arithmetic illustrations, not measured cobot or industrial-robot capabilities.

Both concepts meet the stated demand on those assumptions. Concept B's spare capacity is valuable only if it addresses expected demand, resilience or another evidenced need. Concept A might require less investment or easier changeovers, but those advantages must be quoted and demonstrated. Recalculate using observed stop time and accepted yield. The point of the example is to make access assumptions visible instead of selecting the fastest brochure figure.

Compare changeover work and ownership responsibilities

Ask the intended operator to perform an approved part-family changeover on each concept. Record fixture exchange, tool checks, recipe selection, first-part inspection and production release. Ease of hand guiding is only one possible contribution. A well-organised conventional program with reusable fixtures can be practical for repeat work; an approachable cobot interface can still sit behind a complicated feeder or fragile tooling setup.

Compare complete installed scope and recurring support using the existing quotation and ownership-cost guides. Include protective equipment, conveyors, utilities, training, maintenance access, spare tooling and software entitlements. Give floor space a measured operational meaning: aisle access, replenishment routes and room for service are more useful than a bare base footprint. Obtain responsibility for the whole application, especially where several suppliers divide the arm, machine and safety work.

Record why the chosen cell fits this job

Document the part families, output range, access arrangement and approved configuration on which the decision rests. State the principal reason for selection, such as reliable throughput with occasional loading access or practical reuse across recurring short batches. Keep remaining uncertainties with an owner and resolution date. The decision should be understandable to a production manager who did not attend the demonstrations.

Revisit the assessment when the work changes. A different end effector, sharper part, new loading position or substantially altered protective strategy can change the suitability of the original concept. HSE's guidance on modified machinery explains that changes in design, function or safety need assessment and may alter obligations. Have the responsible specialists review such changes before treating the original acceptance as permission for a new application.

References: Refurbished and modified machinery

Checklist

• Define two complete cell concepts with matching purchase boundaries.
• Record human tasks, access frequency and material routes.
• Review actual tools, parts and process hazards with the integrator.
• Compare accepted output under the proposed protective configuration.
• Demonstrate the intended operator's approved changeover.
• Verify controller options and support for the exact equipment.
• Document the selection reason and conditions that require reassessment.

Common questions