· EQUIPMENT & PROCESS LIBRARY

Visual inspection system (manual and automated)

100% inspection · Automated visual inspection (AVI) · Manual inspection booth · Semi-automated inspection

E13 · INSPECTION, INTEGRITY AND IN-PROCESS TESTINGAseptic processingChecked against primary sources by SPEQ

The booths and machines that inspect every filled parenteral container for particles and defects, and the qualification, defect library and trending that show they find what they should.

What this page does not claim

An equipment family page teaches how a kind of equipment works and how it is assured. It is not a qualification of any site’s equipment, sets no universal limits, and was checked by SPEQ against the sources it lists, not by an independent engineer.

WHAT IT DOES

Annex 1 asks that "all filled containers of parenteral products should be inspected individually for extraneous contamination or other defects" (8.30). Manual inspection uses controlled lighting and backgrounds; automated systems spin and image each container; semi-automated systems present containers to an inspector.

Defects are classified by criticality, a defect library covers all known classes and is used to train inspectors, and batch defect levels are trended and investigated when unusual. A critical defect found after inspection should trigger an investigation, because it shows the inspection may have failed (Annex 1, 8.30).

Where it sits: After filling, capping and any terminal sterilisation, before labelling and packaging.

WHERE YOU FIND IT

  • Vials, ampoules, syringes and cartridges of parenteral products
  • Lyophilised products, inspected for cake and container defects

In the reference facility: Aseptic filling & isolator line, Packaging, labelling & serialisation

Not covered on this page: Container-closure integrity, which visual inspection does not test (Annex 1, 8.22); Sub-visible particle testing in the laboratory.

MAJOR COMPONENTS

  • Manual inspection booth. Controlled illumination, black and white backgrounds, and a defined inspection rate (Annex 1, 8.31).
  • Automated inspection machine. Spins, stops and images each unit; rejects defects.
  • Defect library and challenge sets. Known defects used for training, qualification and start-up challenges (Annex 1, 8.30, 8.32).
  • Reject and segregation bins. Keep rejected units apart by defect class for counting and trending.

Connects to: Filling and capping line; Packaging; Batch record and trend reports; Training records.

INSTRUMENTS AND WHAT THEY MEASURE

  • Automated visual inspection system. Particles and defects in each unit, challenged at start-up and during the batch (Annex 1, 8.32). Watch for: validated against too few or unrepresentative defects.

PARAMETERS

SETPOINT · MEASURED · LIMIT · ATTRIBUTE
  • Inspection illuminance (limit, lux). At the manual inspection point, set by the site in qualification (Annex 1, 8.31).
  • Detection probability (measured, % of known defects detected). By defect type, in inspector qualification and automated system validation (Annex 1, 8.31, 8.32).
  • Inspection defect rate (limit, % by defect class or units per batch). By defect class per batch, against the site’s routine and trend data (Annex 1, 8.30, 8.33).
  • Belt speed or dwell time (setpoint, mm/min or min). Inspection rate or line speed, controlled and qualified (Annex 1, 8.31).

SPEQ publishes no universal alarm or acceptance values: they depend on the process, the product and the site’s own qualification.

CONTROLS, ALARMS AND INTERLOCKS

  • Procedure: Inspector qualification. At least annually, with corrective lenses if normally worn, using defect library sets and worst cases such as fatigue and line speed (Annex 1, 8.31).
  • Procedure: Start-up and in-batch challenges. Automated systems challenged with representative defects before start-up and at intervals (Annex 1, 8.32).
  • Procedure: Defect trending. Defect types and numbers recorded and trended; adverse trends assessed for product on the market (Annex 1, 8.33).
  • Procedure: Breaks and distractions. Frequent breaks and minimal distraction for manual inspectors (Annex 1, 8.31).

An alarm asks for a response; an interlock stops an action; a safety function protects people and has its own risk basis.

HOW IT FAILS

  • Inspector fatigue or too fast a rate. Lower detection of particles and defects.
  • Automated system drift. Defects pass between start-up challenges.
  • Defect library missing a new defect type. Inspectors and machines not trained to find it.
  • Rejected units re-inspected into acceptance. Real defects released.

ASSURANCE ACROSS THE LIFECYCLE

EXAMPLES, NOT A MANDATORY LIST
  • Qualification. Defect classification by criticality (Annex 1, 8.30); Manual inspector qualification with defect sets; Automated system validated as equal to or better than manual (Annex 1, 8.32).
  • Each batch. Reject counts by defect type; Automated challenge results; Acceptance sampling of passed units.
  • Trending. Defect and reject trends with statistically based reject levels (Annex 1, 8.33); Investigation of unusual batches.

WHAT THIS DOES NOT PROVE

  • A unit that passes inspection is not shown to be sealed. Instead: Container-closure integrity testing; visual inspection is not an integrity test (Annex 1, 8.22).
  • A low reject rate does not prove the inspection is effective. Instead: Detection rates from qualified defect sets and acceptance sampling results.
  • Inspection does not prove the product is sterile. Instead: Sterility assurance from the aseptic process, APS and sterility testing.

SOURCES

1 CITED
  • EU GMP Annex 1: Manufacture of Sterile Medicinal Products — European Commission (EudraLex Volume 4), 2022 revision, in operation 25 August 2023. EU / EEA; revised jointly with the PIC/S Committee; in force. Supports: Individual inspection, defect classification, defect library and investigation (8.30); manual inspection conditions and qualification (8.31); automated inspection validation and challenges (8.32); trending (8.33); visual inspection is not an integrity test (8.22).

HOW THIS PAGE WAS CHECKED

Researched and checked against the sources above by SPEQ (Claude) on 2026-10-21. It has not been reviewed by an independent engineer. It teaches how this kind of equipment works and is assured; it does not say any site, asset or process is qualified, validated or compliant.

  • Pharmacopoeial chapters on visible particles and inspection (for example USP <790> and <1790>) are not yet cited; they were not checked here.

Spotted something wrong? Tell us.