Metal Detection: The Boring Gate That Still Protects the Line

A frozen tray may pass through a plant looking ordinary: sealed film, clean label, correct weight, case packed and ready for the cold store. Then a small piece of wire, a shaving from a worn blade or a fragment from maintenance turns that ordinary pack into a recall risk. Metal detection is the use of electromagnetic inspection equipment to find and reject packs or bulk material containing ferrous, non-ferrous or stainless steel contamination. It is old technology by automation standards, but in food plants it remains one of the basic checks that cannot be allowed to become wallpaper.

The familiar machine at the end of the line is not a decoration

Metal detectors are so common that people stop seeing them. They sit after a bagger, before case packing, over a conveyor, around a pipeline, at the discharge of a bulk line, sometimes after final packing. Operators walk past them all shift. Auditors ask for records. Test pieces are passed through. The lights blink. The line keeps moving.

That familiarity is dangerous.

Food plants still contain metal. Cutting blades, screens, pumps, screws, forming plates, conveyors, mixer parts, clip fragments, maintenance tools, staples from the wrong place, wire from brushes that should not be there. Frozen food adds its own rough handling: hard pieces, low temperatures, vibration, washdown, and equipment that works under pressure for long hours.

A detector does not prevent metal entering the line. It catches what has already gone wrong. That makes placement, testing and rejection more serious than the modest hardware suggests.

In frozen vegetables, fries, seafood portions, bakery items, ready meals or ice cream multipacks, the detection point has to match the risk. Bulk inspection before packing may catch a fragment early. Finished-pack inspection checks what is actually sold. Both can be valid. Neither should be chosen because it was the easiest place to fit the frame.

Ferrous is easy, stainless is where optimism gets punished

Metal detectors work by sensing disturbances in an electromagnetic field. Ferrous metals, such as iron and mild steel, are usually easier to detect because they are magnetic and conductive. Non-ferrous metals, such as aluminium, copper or brass, behave differently but can still be detected well in many foods. Stainless steel is often the difficult one, especially non-magnetic grades used widely in food machinery.

That last point matters because stainless steel is exactly what many plants are built from.

Detection sensitivity is normally expressed by the smallest test sphere size that can be found reliably under defined conditions. The word reliably matters. A detector that finds a test piece once during a demonstration is not the same as one that detects it repeatedly, in the right position, at real line speed, with the actual food running through it.

Shape also matters. Test balls are convenient and standardised. Real contamination is rarely polite enough to arrive as a perfect sphere. A thin wire, a flat shaving or a twisted fragment may present differently depending on its orientation as it passes through the aperture. That is one reason plants should avoid treating the sensitivity number as a magic shield.

There is also aperture size. Larger openings usually reduce achievable sensitivity. A detector around a wide case conveyor cannot behave like a smaller unit checking individual sachets or narrow flow-wrapped pieces. Bigger packs may be necessary. They still come with physics attached.

Product effect is the quiet nuisance in the room

Some foods disturb the detector signal even when they are clean. Moisture, salt, temperature, acidity, mineral content and pack size can create what the industry calls product effect. In simple terms, the food itself can look slightly conductive or reactive to the detector.

Frozen lines can be easier than wet chilled foods in some cases, because freezing reduces certain conductive effects. But it is not a free pass. Partly frozen material, temperature variation, ice glaze, brines, sauces, marinades and dense ready meals can all complicate the reading. A block of frozen spinach is not the same challenge as a dry bakery item. Seafood in glaze behaves differently from a bag of potato wedges.

The detector must be set up with the real item, at real temperature, in the real pack. A test done with room-temperature samples in a quiet trial area does not tell the whole story. Neither does a setting copied from a similar line because the packs look almost the same.

Packaging can interfere too. Metallised film, aluminium trays or foil components may make conventional metal detection unsuitable or force different placement. In those cases, plants may need X-ray inspection, upstream bulk detection, ingredient controls or a different risk control plan. Pretending the detector can see through everything is poor engineering and worse food safety practice.

Sensitivity is not only about how low the number can go. It is about stable detection without false rejects becoming so frequent that operators begin to distrust the machine.

Industry misconception: passing a test piece means the risk is covered

A common mistake is to reduce metal detection to a test routine. Ferrous passed. Non-ferrous passed. Stainless passed. Initials on the sheet. Move on.

The paperwork may be complete while the protection is weak.

Validation and verification are not the same thing. Validation asks whether the metal detection setup is capable of controlling the identified risk in that line, with that food, that pack, that speed and that reject device. Verification checks that it continues to work during production. Both are needed under hazard analysis and critical control point (HACCP) thinking, especially where metal detection is treated as a critical control point or an important foreign body control.

The reject system deserves as much attention as the detector head. If the machine detects metal but the reject gate misses the pack, the plant has not controlled the hazard. Air blast, pusher, drop flap, retracting belt, stop-on-detect and lockable reject bins all have their place. The choice depends on pack weight, speed, spacing and line layout.

Rejected packs need control. They should not sit loose on a table where someone can return them to the line by habit. Bins should be secure where risk demands it. Reject confirmation, bin-full alarms and line-stop functions are not luxuries on high-risk lines. They exist because people get busy and shifts get messy.

Records also need to mean something. A failed check should trigger action: hold back affected goods, investigate the cause, repair or adjust the equipment, document the decision. A metal detector that fails a routine check after two hours of running creates a question about everything that passed since the last good check. Nobody enjoys that conversation. Avoiding it is not a control measure.

Questions buyers should ask suppliers

Metal detection should not be waved away with "we have detectors on the line." The useful discussion is more specific, and usually less comfortable.

• Where is metal detection placed: raw material, bulk flow, after forming, after freezing, after final packing or before case packing?
• What are the validated detection limits for ferrous, non-ferrous and stainless steel in the actual food and pack format?
• How is product effect managed for wet, salty, glazed, dense or temperature-variable frozen items?
• What happens when the detector rejects a pack, and how is successful rejection confirmed?
• Are rejected packs locked away or physically controlled until investigation?
• How often are test pieces run, and are they tested at the least sensitive positions in the aperture?
• What action is taken if a routine check fails during production?
• Does the packaging format contain foil, metallised film or trays that limit conventional metal detection?

These questions are basic. That is the point. Foreign body controls fail more often through weak basics than through lack of fashionable technology.

Metal detection is not a replacement for good maintenance, tool control, sieving, magnets, supplier checks or sensible line design. It is one barrier in a wider foreign body plan. But it is a barrier with a very clear job: find metal contamination before the pack leaves the factory.

Old equipment can still be critical equipment. The detector may not impress visitors in the way a robot cell or spiral freezer does. It will not make a better photograph for a presentation. It just stands there, waiting for the piece of metal everyone hopes never appears.

On the day it does, the plant will be judged by whether that familiar machine was treated as protection or background furniture.