V5 Ultimate
Guide

Foreign Material Control: Validating the CCP and Defending the Reject Log

Foreign material (metal, plastic, glass, stone, wood, bone, insects) is the #2 driver of Class II food recalls in the US — and one of the most-cited GFSI audit areas because the controls are mechanical and the evidence is binary: either the metal detector found the test piece in the validation run, or it didn't. Metal detection and X-ray inspection are usually the headline CCPs in a finished-product line; magnets, sieves, filters, screens, and glass/brittle plastic policy round out the system. This guide is the operating manual for validating the CCP, defending the reject log, and surviving the GFSI / FSMA inspection conversation.

Start free trial Free trial, no credit card, onboard in days, not months.

Metal detection vs X-ray inspection — when each is the right CCP

Metal detection finds ferrous, non-ferrous and stainless metal contaminants by electromagnetic field disturbance — fast, cheap, well-understood, but degraded by product effect (high-moisture, high-salt or warm product creates a signal that masks small contaminants). Detection sensitivity is typically 1.5–2.5 mm ferrous, 2.0–3.0 mm non-ferrous, 2.5–3.5 mm stainless on a wet product line; tighter on dry product. X-ray inspection finds metal, glass, stone, dense plastic and bone by density difference — slower, more capex, but unaffected by product effect and can find non-metal contaminants metal detection misses. The right CCP is usually metal detection on dry/packaged finished product, X-ray on wet/high-care or where bone or glass is the credible hazard. Some lines run both.

Validation: proving the detector finds what the hazard analysis demands

Validation is the science exercise that establishes the detector's reliable detection limit on a specific product on a specific line under worst-case conditions. The deliverable is a validation report with the test piece sizes (ferrous, non-ferrous, stainless), the product matrix tested, the detector settings, the conveyor speed, the position of the test piece in the product flow, the number of trials, and the false-reject rate. Validation runs at first introduction, at any change to the product formulation or packaging that affects detection, and on a defined frequency (annually is typical). A reject log without a current validation file underneath is a Major NC in BRCGS Issue 9 and FSSC 22000.

Verification: the routine check that proves the validated detector is still working

Verification is the routine confirmation, on every shift or every defined run, that the detector still detects the validated test piece sizes — typically at start-up, end-of-run, and at a defined interval (every 30–60 minutes is common, sometimes more frequent for high-risk products). Verification uses certified test pieces (NIST-traceable or supplier-certified spheres at the validated sizes). A failed verification implicates all product made since the last successful verification — that's the hold-and-test event most operators don't want to face. GFSI auditors will pull the verification log, count gaps, and look for failed verifications that didn't trigger a product hold.

Reject log and reconciliation — the discipline that catches the missed reject

Every reject must be logged, identified (visual inspection of the rejected unit to confirm whether it actually contains foreign material), and investigated if it does. Reject reconciliation — counting the rejected packs and matching against the reject counter — proves that the reject mechanism worked and that no contaminated pack continued down the line. Common 483 and audit findings: rejects not logged; rejects logged but not investigated; reconciliation absent (so a contaminated pack could have been recovered onto the line without record); positive finds (real metal in a rejected pack) not escalated into root-cause and CAPA.

Magnets, sieves, screens and the upstream defence

Metal detection is the last line of defence — the upstream programme catches contamination earlier and cheaper. Magnets (rare-earth, typically 10,000+ gauss) on ingredient feed and finished-product flow remove ferrous before it reaches the detector. Sieves and screens (per ingredient particle size) catch oversize and foreign objects. Filters on liquid streams catch finer particulates. Each requires its own inspection cadence, integrity test, and replacement schedule — magnets cleaned and inspected per shift, sieve integrity verified per batch, filter integrity tested per defined cycle. A broken sieve or a magnet found loose during routine inspection implicates production back to the last documented intact check.

Glass and brittle plastic policy

Glass and brittle plastic in the production environment are themselves a foreign-material risk — a broken light cover, a snapped polycarbonate guard, a fractured plastic scoop. The policy: register every piece of glass and brittle plastic in the production environment (often called the 'glass and brittle plastic register'), inspect each piece on a defined cadence, document breakages with implicated-product investigation, and replace breakable items with shatter-resistant equivalents where possible. Lights in production zones must be shatter-resistant or protected. A glass and brittle plastic register that hasn't been updated in 12 months is a frequent BRCGS minor.

Standards covered in this guide

Each standard, retailer code or assurance scheme referenced above has its own deep-dive page with scope, audit detail and common pitfalls.

Where this lives in V5 Ultimate

The clauses above aren't theoretical — every one maps to a shipped module and an industry profile. Jump to the parts of the product that turn this guide into evidence on a Monday morning.

Frequently asked

How sensitive does my metal detector need to be?
What the hazard analysis demands and what the validation supports. Common practice on wet/RTE product lines is 2.0–2.5 mm ferrous, 2.5–3.0 mm non-ferrous, 3.0–3.5 mm stainless; tighter on dry product. The right answer is the smallest reliably-detected size on your product on your line under your worst-case condition — that's the validation output, not the spec sheet.
How often do I have to run verification?
Most GFSI-aligned operations run verification at start-up, end-of-run, and at a defined interval (30–60 minutes is common). High-risk product (infant, allergen-sensitive, premium retailer programmes) may demand every 15–30 minutes. The verification cadence determines the size of the hold window on a failure — shorter verification = smaller hold.
Is X-ray always better than metal detection?
Not always. X-ray finds non-metal contaminants metal detection misses (glass, stone, bone, dense plastic) and is product-effect-independent. But it's slower, more capex, and on a dry product with no glass/bone hazard, metal detection is faster and cheaper. The right choice is hazard-analysis driven.
What's the reject log auditors look for?
A complete log per shift with reject count, reject-mechanism reconciliation (count rejected packs vs counter), visual inspection of rejected packs, and a closed loop on any confirmed positive find — root cause, CAPA, supplier or process implication. Missing reconciliation is the most-cited gap.

See it on your shop floor.

Free trial, no credit card, onboard in days, not months.

Spot something off? .