HACCP, BRCGS and metal detection
What fails an audit is rarely the metal detector. It is the documentation around it. Here is what actually gets asked — and where people usually come up short.
This is practical experience from four decades in the industry, not a legal interpretation. Standards are revised, and your certification body is the authority on current wording.
The detector as a critical control point
In most food production the metal detector is designated a CCP in the HACCP plan. That has a consequence which is easy to overlook: from that moment it is not enough that the detector works. You have to be able to prove it worked, for the whole production period.
A CCP requires a critical limit, a monitoring procedure, a corrective action, verification and records. For a metal detector that means, concretely:
| Critical limit | The smallest test sphere in ferrous, non-ferrous and stainless steel the detector must find in that product. It has to be set on the basis of a test on the product — not copied from a datasheet |
|---|---|
| Monitoring | The test routine: who tests, when, with which pieces, and where the result is recorded |
| Corrective action | What happens when a test fails — including how product run since the last passed test is identified and handled |
| Verification | Confirmation that the routine is actually followed and the detector still meets the limit. Usually an external calibration |
| Records | Documentation that can be produced for an auditor and for customers |
Test frequency: how often is often enough?
There is no single number that applies to everyone. The frequency has to be set by you and be justifiable. What the standards look for is that the choice was made deliberately and is followed — not that it is unusually high.
Common practice is to test at:
- Production start-up
- Product change, because sensitivity is product-dependent
- Shift change
- End of production
- After any adjustment to the detector or intervention on the line
Consider the consequence when you choose. If you test once every 24 hours and the test fails, a full day of production has to be handled. If you test every two hours, it is two hours of production. Test frequency is as much a commercial decision as a food safety one.
The five most common findings
- The reject is not verified. The test shows the detector signalled — but not that the item was actually removed. This is the most common one. The fix is reject confirmation sensors and a lockable reject container.
- Test pieces are not traceable. Unmarked or worn spheres with no documented material and diameter. Cheap to replace, expensive to argue about. We supply traceable test pieces →
- No procedure for product since the last passed test. There has to be a written description, and the people on the line have to know it.
- Sensitivity can be changed without a record. If an operator can turn it down without it being logged, the control point is not under control. Password protection and logging solve it.
- The critical limit is not justified. A number copied from a datasheet does not survive the question of how it was established for this specific product.
The practical point: if there are false rejects on the line, sooner or later somebody turns the sensitivity down — and then the critical limit in the HACCP plan is no longer the one actually being run. That is the most serious and most widespread failure in industrial metal detection. The answer is not discipline. It is making detection stable at the required sensitivity. Read the causes of false rejects →
We can help with the documentation
Our technicians calibrate and validate metal detectors — including other manufacturers' equipment — and leave a written report you can put in front of an auditor. If you need a critical limit established for a new product, we can test it in our test centre and give you the figure the limit can be justified with.