A robot tool-change time is not a production changeover

Retain the output endpoint when reporting elapsed time — method added 16 October 2026

A lock indication and an acceptable production result are separate events. A review packet should make it possible to recover both timestamps without letting the shorter interval stand in for the whole changeover. Agree the run boundary and the acceptable-output rule before collecting the evidence; preserve the rule's revision rather than reconstructing it after seeing the elapsed time.

Lean Enterprise Institute's public Changeover entry defines the interval from the last piece in the completed run to the first good piece after changeover. It does not explicitly call the outgoing piece good. The page displays no publication date; its official metadata records July 2020, and it was checked on 16 October 2026. RobotAtom's proposed last-acceptable-A to first-acceptable-B start is an agreed evidence convention, not an exact quotation of that definition.

ATI's undated QC-310 page, rechecked 16 October 2026, separately describes ready-to-lock proximity sensors and lock/unlock sensing. These are configured coupling signals. Our inference is that a coupling indication alone cannot establish whether the next production output meets the buyer's quality rule. Do not label a ready-to-lock event as locked, or a locked event as acceptable output.

For an illustrative run, suppose the agreed start is 22:00:00, the lock event is 22:00:10 and the first accepted B output is 22:01:00, all on the same dated, synchronised timeline. The entered start-to-lock interval is 10 seconds and the start-to-output interval is 60 seconds. The remaining 50 seconds is elapsed time after locking; those timestamps alone do not tell you whether it involved travel, setup, rejection, recovery or another activity. These are invented example values, not deployment performance.

If the first B output is rejected, retain it as a rejected event and continue to preserve the agreed endpoint. If no acceptable output is evidenced, the full interval is unresolved. Do not substitute the lock timestamp, delete the failed run or treat a missing output as zero. If the quality rule or run scope changed, identify that change and its reviewer before comparing the result with an earlier run.

The separate clock record now accepts dated, whole-second timestamps with explicit timezone and retains the start/run, lock and output references. It checks calendar validity and event order, and withholds the full interval unless the quality-rule revision, output evidence and an explicit confirmation are present. Changing an input resets that confirmation. This is a record-completeness gate: neither the reference nor the checkbox inspects the product or proves that the rule was applied correctly.

Keep controller and quality records alongside the export, including clock synchronisation or uncertainty where relevant. The worksheet does not compensate for clock error, infer an overnight date, analyse repeated-run performance or approve production. Use the existing transition register to preserve intermediate events, failures and responsible owners. Qualified production, integration and safety owners still decide what the evidence supports and what operation is permitted.

Transition-record method added 2 October 2026

Record the event that actually ends each stage. A ready-to-lock indication, a locked indication and a first acceptable output answer different questions. Keep the exact changer, control module, storage station and tool identities with the test. ATI's QC-310 page lists module-dependent identity and interlock options; Zimmer's WPR5000 page documents optional storage-station sensing. Ask which checks the installed configuration provides, rather than copying a generic checklist as if it were a validated interlock design. Both pages were checked on 2 October 2026; neither states a publication date.

Our proposed evidence boundary starts at the last acceptable output with tool A and ends at the first acceptable output with tool B under a stated quality rule. This is a coordination method, not a manufacturer performance claim. Keep utility restoration, controller tool data, recipe confirmation and process verification as separate milestones where applicable. Record the actual event definitions, timestamps with timezone, evidence references and responsible owners.

If a check is missing or fails, retain the exception; do not silently end the clock at mechanical locking. In a hypothetical run where the changer locks promptly but the first part is rejected, mechanical exchange may have succeeded while production changeover remains unresolved. Keep that failed run in the evidence set. A not-applicable disposition needs a reason from the responsible application owner, not an empty field.

The linked local register preserves these transitions and a withheld release disposition. It does not command the robot, diagnose a fault, validate a safety function or approve production. Filling every field does not establish that the configured cell may resume operation. Qualified application owners define the applicable checks and decide whether the evidence is sufficient.

A changer action is only one step in the changeover

A tool changer can complete its mechanical lock or unlock action quickly without proving how long a production cell takes to change work. The robot still has to travel to storage, align, control utilities, verify the tool, switch configuration, return to the process and produce an acceptable result.

Define the buyer's boundary from the last acceptable output with tool A to the first acceptable output with tool B. Report the changer action inside that boundary rather than presenting it as the whole changeover.

Put approach and storage into the task graph

ISO 11593:2022 defines vocabulary for automatic end-effector exchange systems used as part of robot systems in accordance with ISO 10218-2. That scope helps teams describe the exchange system consistently; it does not provide a universal changeover-time formula or verify a product or application.

• Station location, orientation, access envelope, approach and departure path.
• Alignment and insertion tolerances, compliant features and permitted plate separation.
• Clearance for every tool, adapter, service loop and dress pack during coupling and withdrawal.
• Tool support and retention, occupied or empty station logic, presence sensing and wrong-station prevention.
• Controls for debris, fluid, dust and wear on coupling faces, contacts and media ports.

Verify the mechanical and utility transition

ATI documents product options that distinguish ready-to-lock, locked and unlocked sensing, as well as tool identity, tool-stand interlock, network control and downstream utility connections. These are first-party capabilities for configured ATI products, not independent proof that an installed cell prevents every wrong-tool, dropped-tool or utility fault.

• Ready to unlock, unlocked, ready to lock, locked, tool present and correct tool identity.
• Electrical power, pneumatic, vacuum, hydraulic, cooling, process media, communications and downstream devices.
• The exact signal combinations that permit robot motion, process operation or another change attempt.
• Pressure, vacuum, leak, network, diagnostic and device-ready checks required before production resumes.
• A defined response to contradictory, missing, late or marginal signals.

Switch the robot and process configuration

The new physical tool also needs the correct payload, centre of gravity, inertia, tool-centre point, calibration, process recipe, motion limits, I/O mapping and any affected safety-related configuration. Record the owner, version, activation rule and evidence for each change.

A correct lock signal does not establish that the controller loaded the right tool data or that the process is ready. Treat mechanical coupling, tool identity, utilities, robot configuration and process readiness as separate states, then define the combination needed for production permission.

Make the storage station part of the configured solution

Zimmer's WPR5000 documentation lists storage variants, loose-part monitoring, piston-position sensing, teaching aids and compensation for approach deviations. Its ALWR1-50 manual separates tool-in-storage sensing, locking position and signal contact when components join. These are manufacturer descriptions for named Zimmer products, not universal sensor semantics or proof of a completed cell design.

The useful requirement names the station, installed geometry, stored tool, sensor meanings and inspection or cleaning rules. A tool changer should not pass a comparison while the storage arrangement, approach path or dress-pack behaviour remains unknown.

Design the failed-change path

ATI's manufacturer troubleshooting guidance identifies debris between plates and a tool not properly seated in its stand as possible causes of failed locking or an unexpected lock signal. That supports explicit cleaning, seating and recovery requirements; it does not establish a universal fault list or maintenance interval.

• Tool absent, wrong tool, occupied station, incomplete seating and failed lock or unlock indication.
• Contaminated interface, damaged contact, utility leak, low pressure, lost vacuum or communication fault.
• Where the robot stops, what remains mechanically retained and how hazardous energy is controlled.
• Retry limits, manual release or retrieval, authorized intervention and inspection after a collision or failed coupling.
• The confirmed state required before automatic changeover or production may resume.

Validate first acceptable output, not only a successful lock

ISO 10218-2:2025 covers industrial robot applications and cells across integration, commissioning, operation and maintenance. Its public scope supports treating the tool changer as part of the integrated application, but it does not certify a configured cell or prescribe this test plan.

• Freeze the robot, changer, tools, utility modules, storage, software, recipes, load and TCP data and safety configuration.
• Exercise every production tool sequence plus representative temperature, contamination and wear conditions where material.
• Predefine start and finish states, repetitions, acceptable output, retries, faults, interventions, leaks and invalid runs.
• Record complete elapsed time, abnormal changeover and recovery rather than only successful automatic exchanges.
• Reopen affected evidence after changes to tools, station, utilities, dress pack, path, software, recipe or safety controls.

Keep every claim inside its evidence boundary

ISO is authoritative for the public scope and status of its standards. ATI and Zimmer are authoritative for their own product options, procedures and troubleshooting guidance. Their data remains manufacturer evidence until the exact configured changeover is tested under the buyer's representative conditions.

RobotAtom can record the sequence, conditions, component claims, configuration, evidence type, failures and open validation. It does not certify a tool changer, safety function, production rate or regulatory compliance. The responsible buyer, integrator, production owner and qualified safety specialists must decide what applies and whether the evidence is sufficient.