Metal detection is often treated as a checkbox on a supplier questionnaire. In practice the equipment only finds what its aperture, frequency and reject mechanism allow, and a cream or gel with high conductivity can mask a fragment that a dry powder would never hide. Knowing the limits is what makes the control meaningful.
1. What the equipment actually finds
A balanced coil detector looks for a disturbance in an electromagnetic field. A metal fragment disturbs it; so does salt, water and any conductive product. The detector is therefore always a compromise between sensitivity and false rejects, and the setting that works for a dry powder will not work for a wet emulsion.
2. Ferrous, non-ferrous and stainless
Sensitivity is quoted separately for each metal type. Ferrous is easiest to detect, non-ferrous such as aluminium and brass is harder, and stainless steel is hardest because many grades are only weakly magnetic. On a typical cosmetic line, quoted figures might be 1.5 mm ferrous, 2.0 mm non-ferrous and 2.5 mm stainless in the centre of the aperture.
3. Aperture size and product effect
Sensitivity falls as the aperture grows, and it falls further at the aperture edges. Size the aperture to the largest pack that will pass, not to the line’s maximum. Wet or salty products create a product effect that can be phased out, but only within limits; heavily conductive fills may need a smaller aperture or a different technology.
4. Where on the line it sits
The detector belongs after the last step that could introduce metal and before case packing. That usually means after filling and capping, on the individual unit rather than the shipper. Detecting a case tells you a case is contaminated, which is a much less useful piece of information.
5. Testing with challenge samples
A detector is only proven by passing test pieces through it. Standard practice is a set of ferrous, non-ferrous and stainless samples at or near the quoted sensitivity, passed at the start of the shift, at the end, and at intervals in between. The test piece must go through the aperture, not be waved near it.
6. What metal detection cannot see
Glass, stone, hard plastic, insects, hair and foil-laminated fragments are all outside its range. Neither does it see metal that is part of the pack by design, such as an aluminium seal or a metallised sachet. For those packs, detection has to happen before the metal enters the product stream, or not at all.
7. X-ray as the alternative
X-ray inspection finds metal at similar or better sensitivity and also catches glass, dense plastic and missing components. It costs considerably more and needs a radiation safety regime. Where the pack itself contains aluminium, X-ray is often the only workable option, which is why metallised sachet lines frequently rely on it.
8. Reject confirmation and quarantine
A detector that signals but does not physically reject is a monitoring device, not a control. The reject bin should be locked, the reject confirmed by a sensor, and every rejection recorded with time and batch. Rejected units are quarantined and investigated, never returned to the line by hand.
9. Keeping records that hold up
What a buyer wants to see is the challenge test log with times, the reject confirmation record and the maintenance history, aligned to batch records. A single passing certificate proves nothing about the shift in which the batch was packed. Records stitched to batch numbers are the only version that survives an audit.
10. Where it fits
Foreign-body control matters most on filled and sealed formats such as facial masks and creams. See operations notes and the catalogue.
