Robot buying & ownership

AYX / KNOWLEDGE BASE

New, used and refurbished robots, maintenance and ROI.

Source research checked: 2026-10-14

08

Used, refurbished and new robots

BUY-001

Should I buy a new, used or refurbished industrial robot?

Choose according to task suitability, condition evidence, support horizon and complete installed cost. New equipment can offer a clearer support path; a well-documented used or refurbished robot can be a practical option. The lower purchase price matters only if integration, reliability and lifecycle costs still make the project worthwhile.

Explanation, examples & sources

Compare equivalent working packages, not a bare used arm against a fully supported new system. Require the exact controller, cables, pendant, licences, declarations and service scope in each offer. Evaluate whether spare parts and engineering knowledge will remain available for your intended service life. For low utilisation, a supported older model may be adequate; for production-critical work, recovery and parts availability can outweigh the initial saving. No universal discount or payback period can be inferred from the word “refurbished.”

References

Explore next

BUY-002

What do used, refurbished, remanufactured and retrofitted actually mean?

“Used” mainly describes previous service. “Refurbished” usually implies inspection and corrective work, but the scope varies by seller. Remanufacturing can imply a more comprehensive rebuild to a defined specification. Retrofitting changes or adds capabilities. None of these words alone proves condition, remaining life or legal conformity.

Explanation, examples & sources

Replace labels with a written scope: inspected items, replaced components, measurements, software changes, test duration and acceptance criteria. Ask what was deliberately not rebuilt and why. A repainted robot with new grease is different from a documented mechanical and electrical overhaul, even when both are advertised as refurbished. Identify whether the work changes the intended use or safety behaviour; that is a separate substantial-modification question. Require before-and-after evidence rather than relying on appearance or an undefined “as new” promise.

References

Explore next

BUY-003

What should I inspect and request before buying a used robot?

Request a traceable identity, operating and service history, complete hardware list, original documentation and a functional test representative of the intended task. Inspect mechanical condition, cables, controller health and software options. Operating hours help, but they do not describe loading, collisions, environment or maintenance quality by themselves.

Explanation, examples & sources

Agree evidence before travelling: photographs of nameplates, diagnostic records, a backup inventory and a test plan. During inspection, check motion across the intended workspace, abnormal noise, play, overheating and repeated faults with qualified personnel. Verify that brakes and relevant protective functions are inspected under the correct procedures rather than improvising hazardous tests. Record unresolved defects and who pays to correct them. A short unloaded video is useful screening evidence, not a substitute for a documented acceptance test.

References

Explore next

BUY-004

Can missing software licences or documentation make a cheap robot expensive?

Yes. A robot may physically work while lacking the software options, interfaces, calibration data or documents needed for your project. Licence transfer and restoration rights depend on the manufacturer and agreement. Do not assume that an installed feature is legally transferable or recoverable after a controller failure.

Explanation, examples & sources

List required communications, external-axis support, process packages, simulation tools and backup utilities. Ask the supplier to demonstrate the functions and identify licence ownership, transfer restrictions and renewal costs. Confirm that the controller and mechanical unit are an appropriate documented combination. Missing assembly instructions or declarations can also complicate integration and conformity work. Obtain written prices and support confirmation for missing items before comparing the offer with another machine; otherwise the apparent saving may disappear during commissioning.

References

Explore next

BUY-005

What should a refurbished robot warranty and handover include?

A useful warranty states covered components, duration, exclusions, permitted use, claim procedure and who pays for labour, travel and transport. The handover should identify the supplied configuration, known limitations, service work, test results and recovery materials. A business-to-business warranty is not automatically the same as a consumer guarantee.

Explanation, examples & sources

Clarify whether the promise covers the arm only or also controller, pendant, cabling and software. Ask about response times, replacement parts, remote diagnosis and downtime exclusions. Agree who commissions the robot and how defects are distinguished from integration problems. Keep the signed acceptance record and baseline backups with the machine's documentation. Treat the suggested contents as a contract checklist, not a statement that AYXWORKS or any named seller provides these services or terms.

Have commercial terms reviewed for the actual jurisdiction and contract; no universal warranty duration is asserted here.

References

Explore next

09

Maintenance, reliability and lifecycle

MNT-001

What preventive maintenance does an industrial robot need?

Follow the maintenance plan for the exact robot, controller and application. Typical attention areas include lubrication where specified, cables, connectors, brakes, cooling, fastening, cleanliness and inspection of protective functions. The interval depends on the manufacturer's instructions and operating conditions, not a universal annual checklist.

Explanation, examples & sources

Turn the manual into an asset-specific schedule with responsible people, approved materials and recorded results. Track abnormal noise, repeated alarms and changes in cycle performance instead of waiting for a complete failure. Distinguish inspection tasks from service work requiring isolation, specialist tools or authorised personnel. Do not apply grease or replace a component merely because another robot model uses that procedure. Include tooling, air preparation and cell peripherals; a reliable arm cannot compensate for an unreliable gripper or fixture.

Follow the machine's hazardous-energy isolation procedure before maintenance; a stopped program does not establish a safe working condition.

References

Explore next

MNT-002

What robot data should be backed up, and why do batteries matter?

Back up more than the motion program. Depending on the controller, recovery may require configuration, calibration or mastering data, safety settings, option licences and associated PLC or vision files. Some systems use batteries for position retention or controller functions; their role and replacement procedure are model-specific.

Explanation, examples & sources

Maintain a versioned recovery inventory linking files to the robot serial number and software baseline. Store protected copies away from the controller and test restoration on a suitable setup. Record battery warnings and replacement dates according to the manual; do not improvise a universal powered-on or powered-off replacement instruction. Before retiring an old controller, verify that the data and licences can actually be restored. A directory full of unidentified backup files is not a usable disaster-recovery plan.

References

Explore next

MNT-003

When does a robot need mastering, calibration or TCP recalibration?

These are different tasks. Mastering establishes the relationship between joint position sensing and mechanical reference positions. Calibration refines a model or measurement relationship. Tool centre point (TCP) calibration defines the tool relative to the robot flange. A collision, component replacement or tool change can affect different parts of this chain.

Explanation, examples & sources

Diagnose the pattern before adjusting anything. An error limited to one tool suggests a different problem from errors across all tools or poses. Preserve the baseline and check fixtures, tool mounting and coordinate frames before changing controller data. Use the manufacturer's required procedures and references for mastering. After correction, validate the actual production points and relevant protective settings. Re-teaching a few points can hide an underlying mechanical or coordinate problem rather than fix it.

References

Explore next

MNT-004

How can I distinguish normal wear from a robot reliability problem?

Look for a reproducible change from the machine's healthy baseline, not a single isolated symptom. Increasing play, unusual noise, temperature changes, repeat-position errors or recurring alarms can indicate a problem, but they do not uniquely identify its cause. Loading, mounting, tooling and program changes can produce similar symptoms.

Explanation, examples & sources

Record when the issue occurs, with which load, at what speed and after which recent changes. Compare diagnostic logs and measurements under controlled conditions. Avoid clearing repeated protective stops and continuing production without investigation. Build trend data that can distinguish a worn component from an overloaded motion or loose fixture. Qualified inspection should determine whether continued operation is appropriate. Predictive analytics can help prioritise attention, but a model score is not a substitute for a safety-related maintenance decision.

References

Explore next

MNT-005

How do I plan for spare parts, obsolescence and downtime?

Plan around the time needed to restore the process, not merely the age of the robot. Critical parts, compatible replacements, recoverable software and available expertise determine supportability. An older supported system can be more manageable than a newer one with undocumented dependencies or inaccessible specialist tools.

Explanation, examples & sources

List credible failure scenarios and their operational impact. For each, identify spare availability, delivery time, required skills and restoration evidence. Consider holding selected critical parts where downtime justifies it, rather than buying every possible spare. Record controller and licence dependencies before a supplier changes its product range. Define an exit plan for unsupported components and test it early. Include the cost of lost output and recovery labour in the lifecycle model instead of treating purchase price as the whole investment.

References

Explore next

10

Costs, ROI and project acceptance

ROI-001

Why does a complete robot application cost more than the robot arm?

The arm is only one part of a working application. Tooling, fixtures, feeders, safety measures, controls, integration, utilities, commissioning, training and support can all be necessary. Comparing arm-only prices with turnkey quotations gives a misleading picture of the actual investment.

Explanation, examples & sources

Ask for a scope breakdown showing included hardware, software, engineering, acceptance tests and exclusions. Separate one-time engineering from repeatable hardware and recurring service costs. Clarify who supplies site preparation, extraction, networks and machine interfaces. Use the same production requirement for every bidder, including expected quality and recovery from common faults. This file deliberately gives no universal cell price because costs depend heavily on process and scope; the numerical example in ROI-002 is illustrative arithmetic, not a quotation or market benchmark.

References

Explore next

ROI-002

How do I calculate robot payback without exaggerating savings?

A simple payback estimate is installed investment divided by credible annual net benefit. Net benefit should subtract recurring operating costs and count only savings or contribution the business can actually realise. Faster theoretical cycle time or reassigned labour is not automatically cash saved.

Explanation, examples & sources

Illustrative example, not market pricing: €80,000 installed investment excluding VAT, €38,000 annual realised benefit and €8,000 additional annual operating cost produce €30,000 net benefit and approximately 2.67 years simple payback. This excludes financing, tax, depreciation, ramp-up and residual value. Recalculate for lower utilisation, slower commissioning and higher downtime. Distinguish cash savings from capacity, ergonomics and resilience benefits rather than double-counting them. Use measured baseline output and contribution margins, not a vendor's generic payback calculator defaults.

This is an engineering planning example, not financial or tax advice. A final investment case should use the organisation's accounting assumptions.

References

Explore next

ROI-003

How do I know whether a task is ready for automation?

A task is a good candidate when its inputs, required outputs, variation and exceptions can be described and tested. Stable presentation, measurable quality and a clear recovery process matter as much as robot capability. Automating an unstable upstream process can simply reproduce its failures faster.

Explanation, examples & sources

Observe real shifts and collect representative good, marginal and failed parts. Record cycle distribution, replenishment, changeovers, operator judgement and unplanned work. Identify which constraints are physical, organisational or software-related. Run a small proof of concept against the hardest credible cases, not only the easiest sample. Define what would make the project fail before choosing equipment. Where a fixture or process redesign removes variability cheaply, do that first and reassess whether the robot is still the best solution.

References

Explore next

ROI-005

What should FAT, SAT and production handover verify?

Factory Acceptance Testing (FAT) checks the agreed system before delivery; Site Acceptance Testing (SAT) checks it in its installed environment. Neither acronym defines a sufficient test by itself. The contract must specify process performance, interfaces, recovery, documentation and the separate safety-validation evidence required for the application.

Explanation, examples & sources

Test representative part variants, sustained throughput, quality, replenishment, changeover and credible faults. Include power loss, communication loss and restart under controlled procedures. Define pass criteria, sample sizes, unresolved-defect categories and who signs each result. At handover, deliver current programs, backups, instructions, training records and support contacts. Do not confuse a successful production demonstration with proof that protective functions have been validated. Retain the tested software and hardware baseline so later changes can be assessed against a known state.

References

ROI-004

What belongs in a robot's total cost of ownership?

Total cost of ownership includes acquisition, integration and the costs of operating, maintaining, changing and eventually retiring the application. Energy and consumables matter, but support, downtime, licences, product changes and recovery effort can also be significant. Use a stated time horizon and consistent assumptions.

Explanation, examples & sources

Build a year-by-year model rather than hiding everything in one payback figure. Include planned maintenance, likely replacement parts, training turnover, backup and security work, software subscriptions and revalidation after changes. Show a base case and a downside case. Treat uncertain future resale value cautiously. Compare purchase and service contracts on identical uptime and scope assumptions, including who carries breakdown risk. The model should remain understandable enough to update when actual production data becomes available.

References

Explore next

← Back to all subjects