Metal Detectors - Determining Critical Limits

My organisation is in the process of implementing HACCP system. I have been tasked to come up with critical limits on the undermentioned metals for the purposes of designing appropriate testers for metal detectors effectiveness.

Ferrors (Wrought iron)
Non-Ferrous (Tin)
Stainless steel(High speed steel)

However, current suggestions are as follows:
Ferrors 1.5mm

Non-Ferrous 2.0mm
Stainless steel 2.5mm

Are there international standards set elsewhere? Or can someone advice me on where to get this information?
I will be very grateful to have your contributions.

Thanks

Oscar M.

I am not aware of international standards for the limits of detection but they may exist. However from my experience I know that the sensitivity of a metal detection unit will depend on a number of factors including:

• The characteristics of the product itself. For example dry product have a high sensitivity than wet products.
• Aperture of the unit. The dead center is less sensitive than the outside areas.
• The type of metal is also important. S.S. can be difficult to detect and ferrous metal can be relatively easy.
• Orientation of the metal through the aperture can determine sensitivity
• The size of the test pack can be important
• The environmental conditions and the packaging material the product is in can also impact on sensitivity

So you can see there are a lot of factors that can determine the sensitivity of the unit. I have used the following guidelines in the past - see document. It might be helpful. My advice is to work closely with the vendor of the metal detection unit. They are experts in their field and will usually help you get the maximum sensitivity

Honestly, this is one of the many things that frustrates me the most.
Almost everyone has metal detectors, but hardly anyone has any hard and fast scientific info to explain why they choose certain test wands to challenge those detectors. It seems that every industry has a "standard" that they go by, basically because it's what everyone else in their field have been using for years.

If metal detectors have become de facto CCP's, because standard makers and customers seem to think they should be, why has there not been any demand for some specific detection levels?

The lack of any movement to make a standard for metal detection, even broken down by the many processors that use them, just pushes the debate closer to the side of "metal detection is monitoring". Not a CCP.

Marshall

Hi George

In the same boat, missing justification and validation for the test pieces we are using, ' have always used these pieces' and no one knows the reason behind the original decision. Have you got a source for your attachment.



Thanks

I am not aware of international standards for the limits of detection but they may exist. However from my experience I know that the sensitivity of a metal detection unit will depend on a number of factors including:

• The characteristics of the product itself. For example dry product have a high sensitivity than wet products.
• Aperture of the unit. The dead center is less sensitive than the outside areas.
• The type of metal is also important. S.S. can be difficult to detect and ferrous metal can be relatively easy.
• Orientation of the metal through the aperture can determine sensitivity
• The size of the test pack can be important
• The environmental conditions and the packaging material the product is in can also impact on sensitivity

So you can see there are a lot of factors that can determine the sensitivity of the unit. I have used the following guidelines in the past - see document. It might be helpful. My advice is to work closely with the vendor of the metal detection unit. They are experts in their field and will usually help you get the maximum sensitivity

Thanks George!


I think this information will play an important part when making a decision. However, do you have an authoritative source of this information(Attachment) for reference. Normally auditors want the basis (scientific research) why we consider a certain size of a metal particle found in our product to be safe if consumed by a customer.

Hi George

In the same boat, missing justification and validation for the test pieces we are using, ' have always used these pieces' and no one knows the reason behind the original decision. Have you got a source for your attachment.

Thanks

Dear classic,

George's data looks similar to last attachment in 2nd forum link below however you should be aware that as noted in his post, the (your) product presentation can be highly relevant as far as operational sensitivities go.

If you are interested in some comparative data / technical discussion of practicalities, can try this post and the surrounding thread -

http://www.ifsqn.com...dpost__p__54841

As per above, most of the raw data predictably comes from MD manufacturers. Most customer specifications are probably based on some sort of "average" of such limiting data, eg Tesco -
(http://www.ifsqn.com...dpost__p__48692 )

Some suppliers may have to tweak the exact numbers a bit if ageing equipment in use.

Rgds / Charles.C

The source is from the Tesco Standard. This is what i typically used over the years.

George

Thanks George!


I think this information will play an important part when making a decision. However, do you have an authoritative source of this information(Attachment) for reference. Normally auditors want the basis (scientific research) why we consider a certain size of a metal particle found in our product to be safe if consumed by a customer.

I wont pretend to be an expert on the area of metal detection but all my experience with MD as a CCP or detection step in general is that companies struggle with it on an almost daily basis. It needs to be managed and I think this ultimately falls to the company itself backed up and supported by the vendor of the unit.

Personally, I see it just as one element in a family of control systems required for metal contamination. I never subscribed to the idea that putting a MD at the end of a production line was a licence to forget about metal contamination. It is a useful detection method and should be in place, but other methods are just as important including:

1. Regular inspection of moving machine parts based on a risk assessment for metal contamination
2. Knife and sharp control procedure
3. Supplier / Ingredient risk assessment and control
4. Magnets

MD is not an absolute control. You need to strive to gain the maximum sensitivity for the given products and conditions, known what that is, and ensure the gap is bridged by specific and well designed controls as above.

I agree with and reinforce George's comments:
"My advice is to work closely with the vendor of the metal detection unit. They are experts in their field and will usually help you get the maximum sensitivity"

However a word of caution George, regarding the table you added as an attachment...

There are a number of factors that effect sensitivity but WRT aperture size...

The size of the aperture (the metal detector opening) has a larger effect on the achievable sensitivity than the size of the product. You may be familiar with this table, but for other readers / users I suggest that you modify any such similar tables you create so that it reads "metal detector opening size" rather than "product size"

To give an example:
A 50mm high block of cheese can be put through a 100mm high opening (possible result 2.5mm S/S detected), or through a 350mm high opening (possible result 6.5mm S/S detected) . The difference in the opening size will affect the achievable result and as you can see from the figures the difference is significant. Therefore a 100mm high opening is capable of detecting a far smaller contaminant than a 350mm high opening. Thus, the product is not the key factor, rather it is the size of the opening in the MD.

If you are using the same size opening but putting different size products through, THEN the product size will affect the result.

I also read through some previous posts WRT 304 vs 316 and I add the following comments:
Test pieces for these two metals can be purchased for the same price - so you can choose which ever one you want with no additional cost to your business.
Both these metals give very similar (usually indistinguishable) results - so use the one that you think is appropriate. I usually advise people to use the one that represents most of their production machinery i.e. if their equipment is made from 316 then use a 316 test piece. However I repeat, there is no notable difference in the result between the two.

I wont pretend to be an expert on the area of metal detection but all my experience with MD as a CCP or detection step in general is that companies struggle with it on an almost daily basis. It needs to be managed and I think this ultimately falls to the company itself backed up and supported by the vendor of the unit.

Personally, I see it just as one element in a family of control systems required for metal contamination. I never subscribed to the idea that putting a MD at the end of a production line was a licence to forget about metal contamination. It is a useful detection method and should be in place, but other methods are just as important including:

1. Regular inspection of moving machine parts based on a risk assessment for metal contamination
2. Knife and sharp control procedure
3. Supplier / Ingredient risk assessment and control
4. Magnets

MD is not an absolute control. You need to strive to gain the maximum sensitivity for the given products and conditions, known what that is, and ensure the gap is bridged by specific and well designed controls as above.

Hi George,

Your phrase

MD is not an absolute control

really makes me smile.

How many food producers therefore have MD as a CCP?

I assume there will be very few who don't, where metal contamination is recognized as a hazard.

Taking into consideration the debate shown in these posts it seems like MD is something of a "black art" and this is also my experience of this when you try to decipher a rationale behind the use of specific test pieces and not others.

I agree 100% with your view that MD is only one part of an overall defense strategy against metal foreign body contamination. This is only one link in the chain (albeit maybe the last link in a process) that includes all of the elements 1 - 4 in your post and others not listed eg engineering maintenance procedures.

Why are the different size test pieces chosen and what is the relationship with hazard to the consumer?

All the test pieces are significantly smaller than the sizes quoted as being injurious to health. See The Physical Hazards of Foreign Materials from the FSIS website in the US. Clearly using a smaller test piece compared to the size that could be injurious to health gives an extra margin of safety but why 2.5mm and not 3.5mm for example?

Surely there must be someone from the MD manufacturing industry on this forum who can give us their views, even if this is done anonymously

Good luck with your MD calibration everyone.... from my experience you will need it!

DP

There are several factors that can affect detection capabilities including detector size, operating frequency and other physical installation factors such as vibration and interference from other equipment.



The product conditions such as moisture percentage, temperature and packaging material will also potentially affect capabilities.


The metal contamination itself will have an affect due to the physics involved. Materials that have both conductive and magnetic properties are easiest to detect in most situations (Fe) and those that have reduced conductive and magnetic properties are more difficult (SS). There are also orientation factors to consider when dealing with non-spherical contamination (wire, swarf, welding slag)


The detector will have a maximum sensitivity capability based on its physical characteristics (size and frequency). That does not mean this will be the same as the production capabilities.

Detectors must be set to successfully inspect production (no or limited false rejects). The variations within production must be taken into consideration (temperature, moisture, acidity, etc.).

These variations make it difficult to identify a one size fits all standard.


A detector that continually rejects good product may be sensitive however, it will not add any value to a foreign material prevention program and will certainly add aggravation frustration and cost to the process.

Thanks for the clarification MD. This is very useful.

George

I agree with and reinforce George's comments:
"My advice is to work closely with the vendor of the metal detection unit. They are experts in their field and will usually help you get the maximum sensitivity"

However a word of caution George, regarding the table you added as an attachment...

There are a number of factors that effect sensitivity but WRT aperture size...

The size of the aperture (the metal detector opening) has a larger effect on the achievable sensitivity than the size of the product. You may be familiar with this table, but for other readers / users I suggest that you modify any such similar tables you create so that it reads "metal detector opening size" rather than "product size"

To give an example:
A 50mm high block of cheese can be put through a 100mm high opening (possible result 2.5mm S/S detected), or through a 350mm high opening (possible result 6.5mm S/S detected) . The difference in the opening size will affect the achievable result and as you can see from the figures the difference is significant. Therefore a 100mm high opening is capable of detecting a far smaller contaminant than a 350mm high opening. Thus, the product is not the key factor, rather it is the size of the opening in the MD.

If you are using the same size opening but putting different size products through, THEN the product size will affect the result.

I also read through some previous posts WRT 304 vs 316 and I add the following comments:
Test pieces for these two metals can be purchased for the same price - so you can choose which ever one you want with no additional cost to your business.
Both these metals give very similar (usually indistinguishable) results - so use the one that you think is appropriate. I usually advise people to use the one that represents most of their production machinery i.e. if their equipment is made from 316 then use a 316 test piece. However I repeat, there is no notable difference in the result between the two.

Hi George,

Your phrase really makes me smile.

How many food producers therefore have MD as a CCP?

I assume there will be very few who don't, where metal contamination is recognized as a hazard.

Taking into consideration the debate shown in these posts it seems like MD is something of a "black art" and this is also my experience of this when you try to decipher a rationale behind the use of specific test pieces and not others.

I agree 100% with your view that MD is only one part of an overall defense strategy against metal foreign body contamination. This is only one link in the chain (albeit maybe the last link in a process) that includes all of the elements 1 - 4 in your post and others not listed eg engineering maintenance procedures.

Why are the different size test pieces chosen and what is the relationship with hazard to the consumer?

All the test pieces are significantly smaller than the sizes quoted as being injurious to health. See The Physical Hazards of Foreign Materials from the FSIS website in the US. Clearly using a smaller test piece compared to the size that could be injurious to health gives an extra margin of safety but why 2.5mm and not 3.5mm for example?

Surely there must be someone from the MD manufacturing industry on this forum who can give us their views, even if this is done anonymously

Good luck with your MD calibration everyone.... from my experience you will need it!

DP

For me, metal detection as a CCP is an open question. As mentioned by a number of members, you will put in place a set of well designed control and detection measures to reduce this hazard to an acceptable level. Call MD what you will... the various factors that affect sensitivity will determine the practical limits of the control, these factors will vary day to day and so you need to manage it well. If your customer wants it as a CCP then make it a CCP. If they want the metal detection unit painted pink with yellow polka dots then do that too. In the final analysis MD is not absolute and my long held opinion is that it struggles to meet the full definition of a CCP. Let's put it this way... if a product heat treatment step (CCP) displayed the same variation and drift (Statistically) as metal detection no customer would accept this. In the spirit of HACCP principles they would tell you to 'redesigned' the process step. I would love to see some capability studies on metal detectors over time. This would make for excellent reading and dare I say validation. The obvious point of course is that metal detection is not an intrinsic process step like cooking. It is simply a detection / inspection activity. More evidence that it cannot be a CCP???

Final word. I'm not down on MD. It is a great piece of technology which has contributed significantly to the safety of the food we produce. I just think we have tried to ram it into the CCP hole. This is not the fault of MD but rather the understandable insistence of retailer technical standards, auditors and some failings in current HACCP principles.


George

Dear All,

All the test pieces are significantly smaller than the sizes quoted as being injurious to health.

With all due respect, this is somewhat selective quoting. The linked article has caveats, eg consumer, and the USFDA related document (from memory) has further details. Not sure how well this data is accepted outside USA ?
Some authorities apparently simply choose to be conservative since the technical capability (arguably) readily exists, eg Belgium (?).
Some auditors are apparently quite happy to go along with the upper values quoted.
The net result is that a variety of acceptable critical limits exists. Spice of Life.

I have never had an auditor challenge my (sensitivity) selection of 2-3 mm "wands" or "spheres". IMEX most auditors do not demand the exhaustive "calibration" procedures as contained within, say, the Tesco manifesto, impressive though they undoubtedly are (eg shea quays elegant post #6).

IMO it is "reasonably" well accepted (eg Codex 2008, et al) that for certain processes, a MD does not have to be a CCP, either conceptually or by risk assessment.

As per George's last paragraph, the MD's continuing popularity as a CCP is surely due to customers, auditors and a tree. And in some cases, the potential auditorial embarrassment of a process with no CCP.

Rgds / Charles.C