A pump-cap inspection system should answer a defined quality question for each bottle and pass a traceable result to the line controls. It should not be selected from a generic list of sensors. Start with the defects that matter, decide which can be detected online and use controlled sample testing for the conditions that cannot be proven non-destructively.
Start with the defect, not the sensor
Create a defect list using approved good packs and representative reject examples. Separate missing pump, tilted seating, incomplete thread engagement, wrong actuator direction, damaged overcap, cosmetic marking, leakage and functional failure. Each condition may need a different detection or sampling method.
A sensor can only confirm what its measurement is capable of proving. A pump-present sensor does not prove the thread is engaged. A height check does not automatically prove torque or seal integrity. Define the decision and the acceptable uncertainty before choosing hardware.
Confirm pump presence and bottle state
Presence checking can prevent an uncapped bottle leaving the cell, but the detection point must account for actuator shape, bottle height variation and reflected surfaces. The line should also know that the expected bottle was present at the capping station and remained in sequence through inspection.
Where the line indexes bottles, link the inspection result to the corresponding pocket or position. On a continuous conveyor, use a tracking method that remains valid through accumulation, speed changes and controlled stops.
Use height and seating checks carefully
Closure height can reveal a missing pump, a visibly high collar or an incomplete seating condition when the bottle datum and expected range are stable. The measurement needs a defined reference surface and must tolerate normal bottle-height, shoulder and decoration variation without accepting a tilted or cross-threaded closure.
Trial known good and known defective samples at normal line speed. If one measurement cannot separate the defect reliably, combine it with another feature or move the decision to controlled sample testing.
Check actuator direction and overhang
Directional lotion pumps, fine-mist nozzles and trigger sprayers may need to face a label panel, carton datum or defined customer presentation. A vision or position check can compare the actuator to a bottle reference when both features are visible and consistently presented.
Define the allowable angle, the bottle datum and the response to an unreadable pack. The same check should not be assumed to work across clear, coloured, printed and reflective bottles without trial evidence.
Separate online checks from sample testing
Online inspection is strongest for visible or measurable conditions that can be assessed without disturbing the pack. Removal torque, leakage under defined conditions and dispensing function may require destructive or off-line sample tests. The quality plan should state which checks are continuous, which are periodic and what action follows a failed sample.
Use the pump cap torque and quality guide to define the measurement method before using a result to tune the machine.
| Defect or condition | Possible evidence | Important limitation |
|---|---|---|
| Missing pump | Presence sensor, height or vision check. | Must distinguish the intended pump from bottle or background features. |
| High or tilted collar | Height profile, vision or multiple-point measurement. | Bottle height and neck variation can affect the datum. |
| Wrong actuator direction | Vision or position measurement against a bottle datum. | Requires visible, repeatable features and a defined tolerance. |
| Loose or over-tight result | Validated process setting plus scheduled torque or removal-torque testing. | Online presence or height alone does not prove torque. |
| Leakage or dispensing failure | Defined sample test or application-specific inspection. | Test conditions must represent the pack requirement. |
Track and confirm every reject
The inspection result must remain linked to the correct bottle until the reject point. Define the distance, conveyor behaviour, accumulation and what happens if the line stops between inspection and rejection. The reject device must handle the actual bottle without creating leakage, breakage or a secondary jam.
Reject confirmation should prove that the failed bottle left the saleable flow. A full reject bin, failed actuator or missing confirmation should create a defined line response rather than allowing production to continue silently.
Define data and fault behaviour
Agree which counts, images, measurements or fault codes are retained and how operators distinguish a machine fault from a genuine pack defect. Data should support production control, not create a misleading record from an unvalidated measurement.
Define the result for unreadable packs, sensor contamination, loss of tracking, blocked downstream flow and restart after an emergency or controlled stop. The line should return to a known sequence before new pumps are released.
Prove the complete system during FAT and SAT
Factory acceptance testing should challenge the agreed good and defective samples, tracking, reject actuation, reject confirmation, full-bin or failed-reject response and controlled stop/restart sequence. Site acceptance should repeat the checks with the connected conveyors, production lighting, actual bottle flow and downstream accumulation.
Record the method, sample set, result and limitation for each defect class. Use the FAT and SAT guide to keep factory and site evidence separate.
Information to provide for an inspection review
- Approved good packs and representative defective packs.
- Defect list, action and tolerance for each format.
- Bottle materials, colours, labels and reflective surfaces.
- Conveyor speed, pitch, accumulation and reject location.
- Required records, counts and operator messages.
- Sample-based torque, leak and functional test methods.