By Kati Hope, global key account manager, Mettler Toledo
Metal detector testing can look like a simple process. Pass a test piece through, check the result, record it and carry on. However, this is an area that frequently catches manufacturers by surprise and can leave them exposed to unnecessary risk. The value of a test is not speed. It is whether it proves the detector can still protect the line and meet the standards of retailers and regulatory requirements.
Why does testing matter?
A metal detector is qualified to meet a specific challenge. At installation, it should be validated for the product, pack format, packaging, and expected contaminant risk. If those factors change significantly, re-validation may be needed. An initial performance verification process will confirm the system is capable of meeting the specified sensitivity standard.
Routine performance monitoring proves the initial metal detection standard has not drifted. This could be from product changes, line changes or detector functionality. Without that proof, operators may believe that their products are being protected when in fact the detector is no longer finding the smallest contaminant it was specified to detect. The result can be costly recalls, failed audits, and wasted product, all of which can lead to significant costs.
How often is often enough?
Testing frequency is not a generic daily task. It is set by the application, HACCP plan, food standard and customer or retailer Code of Practice. Some major UK retailer codes go beyond a simple pass or fail check, requiring defined routines such as consecutive pack tests, memory tests, large metal tests, or specific leading edge, middle, and trailing edge positions.
Quarantine is also key. Product should not leave site until there is evidence that the detector was working during the period in which it was made. If a line produces one tonne an hour and the site can hold only two tonnes in quarantine, testing every two hours may be right. Testing once at the start of the day may look tidy on paper, but it may not protect product already in the supply chain.
Where do tests go wrong?
The most common problem is not that manufacturers forget to test. It is that the test does not challenge the detector in the right way. On horizontal conveyors, the least sensitive point is normally the geometric centre of the aperture. For packed products, many GFSI-recognised Standards and Retailer Codes of Practice may suggest testing at defined points in the product flow, commonly the leading edge, middle, and trailing edge.
That can be difficult on a busy line, particularly with larger products or packs that sit close to the aperture. To reduce testing time, some manufacturers use prepared test packs or samples, with the test piece already positioned in the right place which is typically at points of the product closest to the centre of the aperture. Used properly, this can make testing quicker and more repeatable. But the sample must still represent the worst case. Simply placing a test piece on top of a pack can create false confidence, because it may pass through a more sensitive part of the aperture.
Testing routines can differ depending on whom manufacturers are supplying. Some may only need to follow GFSI-recognised standards where others may need to do further testing if they have to follow specific Retailer Codes of Practice. Â This means one test method will not suit every line, product or manufacturer.
What else should the test prove?
Detection is only half the job. The system must also remove contaminated product and stop it being reintroduced to the line. That means the reject mechanism, reject confirmation, bin full monitoring, lockable reject bins, alarms, and failsafe systems all matter.
For example, if the system detects contaminants in a pack but does not reject the product, then the system must not allow that pack to pass further down the line. Furthermore, a rejected pack must not be able to be lifted from the reject bin and put back into production as it leaves the site exposed. Measures should be put in place to prevent this from happening. Rejected product should also be investigated.
The question is not only what was found, but how it entered the process and what should change.
Can automation improve the process?
Manual testing can be slow and inconsistent. It may need one or two operators, interrupt production and create paper records for audits. It also leaves room for human error when test pieces need accurate placement. In addition, manual testing carries a greater risk that records could be falsified. For example, to save time, an operator relying solely on manual test processes could simply tick a box on a checklist to indicate a test has been run even when it hasn’t.
Automated testing can make checks more consistent and less disruptive. The Mettler Toledo INFINITY-Electronic Test System, for example, injects a signal equivalent to the geometric centre of the aperture, equivalent to predetermined positions on the pack, supporting repeatable tests at challenging positions.
Automation is not a shortcut around good practice. Manufacturers still need a documented routine, clear quarantine rules, rejection checks, and records that show the system was working as intended. However automated testing is more accurate, saves time and money, and will record all activities digitally.
Why are records important?
Auditors need evidence of what happened, when it happened, and what action followed. Manual records can be difficult to manage across multiple lines, but Wi-Fi data transfer can simplify this process. Â Digital data capture can store results centrally and make them easier to retrieve. Â ProdX data management software from Mettler Toledo, for example, can collect and store compliance-relevant data from connected inspection equipment.
As due diligence expectations become tighter, the best testing routines combine practical line protection with clear evidence; testing in the right place, at the right frequency, with the right records to prove it.
For more information, download the Mettler Toledo Metal Detection Guide on Testing Processes by clicking here Â
