Inspection performance
Run agreed good, bad and borderline samples through every approved recipe at representative rates.
URS, FAT and SAT planning
Build acceptance evidence into the seal inspection project
Define defects, samples, operating conditions, reject challenges and records before hardware is committed.
Quick answer
How should an inline seal inspection system be accepted?
Acceptance should challenge the complete decision chain against agreed products, defects, speeds and fault conditions—not simply prove that the camera switches on.
A useful plan starts with a user requirement specification (URS), proves the sensing method with representative samples, checks the built system during factory acceptance testing (FAT), and confirms performance on the production line during site acceptance testing (SAT).
The exact documents and regulated validation route remain the customer’s responsibility unless specifically included in the project scope. Oxford Vision Systems can provide test evidence and validation support appropriate to the supplied system.
URS checklist
Define what must be true before asking for a machine
Clear requirements reduce quotation assumptions and give later acceptance tests an objective basis.
| Requirement area | Questions to answer | Evidence to supply |
|---|---|---|
| Products and formats | Which packs, sizes, lanes, films, artwork and changeovers are included? | Drawings, samples, SKU list and format range |
| Defect definition | What must fail, at what size or severity, and what can pass? | Labelled good, bad and borderline samples |
| Performance | Required rate, pitch, resolution, coverage, false-reject expectation and miss risk? | Line data, video and current quality records |
| Line interface | Trigger, encoder, PLC signals, ready/fault states, recipes and stop logic? | Electrical drawings, network standard and control philosophy |
| Reject control | How is the pack tracked, removed, confirmed and secured? | Pack behaviour, reject destination and risk assessment inputs |
| Data and users | Which images, counts, events, audit information and access levels are required? | Retention, reporting, backup and user-role requirements |
| Environment | Temperature, washdown, dust, glare, vibration, hygiene and available space? | Site survey, layout and cleaning procedure |
Feasibility evidence
Prove that the proposed sensor sees the real defect
A feasibility study should include production variation, known failures and ambiguous examples. The report should state the setup, sample set, observed separation, limitations and unresolved risks.
Promising laboratory images are not the same as installed performance. Product motion, speed, temperature, lighting isolation and pack presentation have to be included in the engineering basis.
Plan a sample trialSample identityProduct, format, defect class and expected result
MethodCamera, lens, lighting, thermal timing or other sensor
ResultDetected, missed, ambiguous and false-reject cases
BoundaryWhat was not tested or cannot be claimed
FAT challenge
Test the complete machine before it reaches site
Factory acceptance should combine normal operation, known failures and deliberate fault simulation.
Tracking and rejection
Challenge consecutive failures, spacing changes, reject confirmation, full-bin and bin-missing conditions.
Controls and records
Verify user access, recipe selection, alarms, counters, image records, data export, backups and recovery.
SAT challenge
Confirm the accepted behaviour on the live line
Site acceptance closes the gap between the build environment and real production.
Install and interface
Verify utilities, guarding, safety functions, line signals, access and physical integration.
Run production
Challenge approved products at agreed normal and peak conditions with real operators.
Test faults
Prove loss of trigger, air, communications, reject confirmation, bin status and other agreed failures.
Handover control
Complete training, documentation, backups, spares, support and change-control responsibilities.
Ongoing assurance
Acceptance is a baseline, not the end of control
Routine challenge tests, golden samples, reject checks and change assessment help keep the system within its approved operating space.
At start-upCamera health, lighting, recipe, reject and reference checks
During productionFault trends, false rejects, low-confidence cases and line events
After changeProduct, material, artwork, software, model, lighting or conveyor review
After maintenanceReconfirm image formation, calibration, tracking and reject performance
Practical answers
Seal inspection validation questions
Clear, application-led answers before you specify an inspection project.
What should a seal inspection URS include?
It should define products and formats, defect classes, minimum relevant feature, line rate, lanes, presentation, recipes, reject response, records, interfaces, environment, user access and acceptance conditions.
What is tested during FAT?
A FAT can challenge image acquisition, inspection recipes, known defects, line-rate simulation, tracking, rejection, fault states, data records, access levels, backups and agreed documentation before shipment.
What is different during SAT?
SAT confirms performance after installation with the real line, products, utilities, controls, operators, speeds, changeovers and upstream or downstream interactions.
Does FAT or SAT prove every future product?
No. Acceptance applies to the agreed sample space, formats, conditions and criteria. New products, materials or process changes may require assessment, recipe work or revalidation.
Confirm what the seal challenge samples actually prove
A visibly wrinkled seal and a confirmed leaking pack are not interchangeable test samples. A camera may detect a visible feature without establishing leakage, while a leakage path may lack the appearance used by that camera. Label challenge samples by the property that has actually been established so the trial does not claim more than its evidence supports.
For each relevant fault, record how the sample was created and how its expected condition was confirmed. Keep visual appearance, seal strength and package leak integrity as separate attributes where they matter to the product. Agree any supporting test method with the customer's packaging and quality specialists; the appropriate method depends on the pack and the required decision.
Compare the proposed inline result against that independently established reference, including cases where appearance and integrity disagree. Record the sensing method's limitations and which remaining risks are controlled elsewhere. This makes a seal-inspection FAT more informative than demonstrating that obvious marks are rejected. Send actual production faults with their available quality-test results, rather than relying entirely on simulated defects that only change how the seal looks.
Does a visible seal defect automatically prove that a pack leaks?
No. Appearance, seal strength and leak integrity are distinct properties and need the appropriate evidence.
What information should accompany seal challenge samples?
Record the intended fault, how it was created, how its condition was confirmed and which property the inline system must assess.
Compare seal-inspection methods and limits Define the relevant seal defects