17 replies to this topic

[AitorMenta]

Good morning all, 

 

I start a topic because I did not find (me or my colleagues) any legislation that help us to determinate whether we are performing a correct checking of the CP of metal detector or nor.  

 

My question is, there are some legislation or directions that specifies that the verification of the function of the metal detector with tests must be carried with product? 

 

In other words, it is possible to perform a verification of the correct function of the equipment without the presence of the actual product in the conveyor¿? Like putting the actual tests of the MD in the conveyor and check if there are an actual detection + stopping + alarm + rejection or whatever the corrective action established takes? 

 

We already ask at the FSIS, and they where quite vague with their answer:

 

FSIS does not have a guideline or Directive specific to metal detector validation. It depends on the manufacturer's specifications, the establishment's hazard analysis decisions, and how the establishment supports its hazard analysis decisions. Establishments often use seeds or wands to determine whether the metal detector is functioning correctly for certain sizes or composition of metal. Depending on the products, packaging, and processes, the seeds or wands may be placed in a package or container with products or used by themselves for testing to ensure a metal detector is functioning properly.

 

I am not an expert, I supposed that the signal that the metal detector receive, when the tests are detected, there are quite few difference if there are product or not, as the product should be seen by the metal detector as a "sample blank". If there are any signal (as a product/matrix always produce a background noise), it is not significant and the metal detector must work in all the circumstances. 

 

Thanks in advance, 

[Scampi]

To answer what i think are your 2 questions:

 

1. No regulatory rules to follow

 

2. you must have the MD set to each different product (it should be programed as such) as things such as density, total weight, packaging type etc all affect how the aperture will perform

[kingstudruler1]

If you want to make verify / validate that your metal detector can remove metal that is IN the product, wouldn't it make sense test with the standards in the product?

 

There could be product signal issues as well as test piece orientation issues if testing is completed outside of the package.  In some cases, placing the test piece on the side bottom or top of a package will trigger a response from the metal detector that is not consistent with how the metal detector will react with the test piece in the product.   It depends on the product, package, and MD.   

 

In the product, as close to the center of the aperture as possible is the ideal place to test a MD.   

[G M]

Whether or not you benefit from having the product present depends on the physical traits of the product (and packaging if you run packaged goods through it), and how it affects the sensitivity of the instrument.

 

If product dampens the distortions to the magnetic field caused by the metal sample seed, then yes the product should be present during testing.  Alternatively, if the product amplifies the liklihood of detection or triggering the system then you would be better off not having it present.

 

 

Does your product occasionally trigger a false positive due to its moisture or salt content?

[Charles.C]

Good morning all, 

 

I start a topic because I did not find (me or my colleagues) any legislation that help us to determinate whether we are performing a correct checking of the CP of metal detector or nor.  

 

My question is, there are some legislation or directions that specifies that the verification of the function of the metal detector with tests must be carried with product? 

 

In other words, it is possible to perform a verification of the correct function of the equipment without the presence of the actual product in the conveyor¿? Like putting the actual tests of the MD in the conveyor and check if there are an actual detection + stopping + alarm + rejection or whatever the corrective action established takes? 

 

We already ask at the FSIS, and they where quite vague with their answer:

 

FSIS does not have a guideline or Directive specific to metal detector validation. It depends on the manufacturer's specifications, the establishment's hazard analysis decisions, and how the establishment supports its hazard analysis decisions. Establishments often use seeds or wands to determine whether the metal detector is functioning correctly for certain sizes or composition of metal. Depending on the products, packaging, and processes, the seeds or wands may be placed in a package or container with products or used by themselves for testing to ensure a metal detector is functioning properly.

 

I am not an expert, I supposed that the signal that the metal detector receive, when the tests are detected, there are quite few difference if there are product or not, as the product should be seen by the metal detector as a "sample blank". If there are any signal (as a product/matrix always produce a background noise), it is not significant and the metal detector must work in all the circumstances. 

 

Thanks in advance, 

 

Hi Aitor,

 

Unfortunately a substantive answer to yr OP probably demands a reasonable understanding of the basic functional theory of metal detectors together with terms like Validation and Verification.
I suggest you get a copy of the MD Guidelines Manual freely offered by Mettler, a well-known Metal Detector manufacturing Company. This i predict  will answer all the elements mentioned in your OP although, depending on your technical background,  you may need additional technical backup on some aspects. Most of yr queries are also already discussed on this forum but will involve  a certain amount of searching.

[AitorMenta]

Good morning,

 

Thank you all for your answers. 

 

I already got a background of intense research because it is a topic I discussed with the autorities, but the problem is that there are no law, legislation or direction that states what is the correct way to perform this cp/ccp. 

 

We already asked to mettler and CEIA and they also told us that there are any directions of the correct way to perform the controls, each company stablish their directions based on their validation of the process.

 

So, if anyone knows some rule or basic legislation about this topic, it would be usefull to stand up for our process.

 

Many thanks,

[Scampi]

A) we have already weighed in on legislation, there isn't any-anywhere

 

B) we can help you more (perhaps) if we know what country you sell in/to

 

C) the manufacturer has already told you what to do-----what more are you looking for?

[juanolea1]

Try:

FSMA Final Rule for Preventive Controls for Human Food
https://www.fda.gov/...A/ucm459719.htm.
 
CPG Sec. 555.425 Foods, Adulteration Involving
hard or Sharp Foreign Objects.
 https://www.fda.gov/.../71953/download
 
 

[Scampi]

Neither of those links tell a person HOW to verify a MD which is what the poster is asking

[AitorMenta]

A) we have already weighed in on legislation, there isn't any-anywhere

 

B) we can help you more (perhaps) if we know what country you sell in/to

 

C) the manufacturer has already told you what to do-----what more are you looking for?

 

Hello Scampi, 

 

We are selling frozen pork meat for Europe, Japan, USA, China... 

 

I genuinely asking because I did not know if I am losing some simple rule/law, because it is strange that, to this important control point, there are no legislation of the way to verify the correct function of the MD (which normally involves a CP or a CCP).

 

Thanks in advance,

[DeadPresident]

Hello Scampi, 

 

We are selling frozen pork meat for Europe, Japan, USA, China... 

 

I genuinely asking because I did not know if I am losing some simple rule/law, because it is strange that, to this important control point, there are no legislation of the way to verify the correct function of the MD (which normally involves a CP or a CCP).

 

Thanks in advance,

Hi,

 

I think you're thinking about it too ,uch. I had this issue when i was first introduced to metal detectors. All YOU have to do is verify that the metal detector goes off when running test pieces provided by the manufacturer ( I would recommend testing it every two hours as that seems to be what a lot of companies do ) and having your equipment validated at least annually by a third party.

[G M]

Hello Scampi, 

 

We are selling frozen pork meat for Europe, Japan, USA, China... 

 

I genuinely asking because I did not know if I am losing some simple rule/law, because it is strange that, to this important control point, there are no legislation of the way to verify the correct function of the MD (which normally involves a CP or a CCP).

 

Thanks in advance,

 

The performance of a metal detector is very dependent on the physical properties of your product, and the packaging, orientation and size of the orifice, etc.  So accordingly the regulations are vaguely written and put most of the onus for ensuring the output is safe on you the manufacturer.

 

You validate the process.  Even in temperature based lethality you don't need to follow government guidelines if you perform your own validation study -- many companies just don't bother to, because of expense or lack of expertise.

[AitorMenta]

Hi,

 

I think you're thinking about it too ,uch. I had this issue when i was first introduced to metal detectors. All YOU have to do is verify that the metal detector goes off when running test pieces provided by the manufacturer ( I would recommend testing it every two hours as that seems to be what a lot of companies do ) and having your equipment validated at least annually by a third party.

In our company, we perform the MD test every two hours. At the end of the process, we do a final test without product, to ensure the correct function of the metal detector. To this day, no authority or auditor told us it was a wrong practice. 

 

The performance of a metal detector is very dependent on the physical properties of your product, and the packaging, orientation and size of the orifice, etc.  So accordingly the regulations are vaguely written and put most of the onus for ensuring the output is safe on you the manufacturer.

 

You validate the process.  Even in temperature based lethality you don't need to follow government guidelines if you perform your own validation study -- many companies just don't bother to, because of expense or lack of expertise.

We already perform our own validation of the metal detector. Maybe, when the question comes, we can defend the procedure with the validation and the historic. 

[marco8963]

Good morning, first of all, let me say that I am by no means an expert in this field.

Therefore, I would like to ask you...

 

Do you think it's excessively simplistic to perform the following steps:

Perform the test by placing metal samples in various positions inside the metal detector and recording the results. Of course, testing different sizes of metal and repeating the tests multiple times to ensure "minimally" reliable results.

 

Calculation of the POD and FRR using the collected data: To find the POD, one could calculate the ratio of "correct alarms" to the total number of "contaminations" actually present, while for the FRR, one could calculate the ratio of "false alarms" to the total number of nonexistent "contaminations" (false positives).

 

Thank you very much.

[DeadPresident]

Good morning, first of all, let me say that I am by no means an expert in this field.

Therefore, I would like to ask you...

 

Do you think it's excessively simplistic to perform the following steps:

Perform the test by placing metal samples in various positions inside the metal detector and recording the results. Of course, testing different sizes of metal and repeating the tests multiple times to ensure "minimally" reliable results.

 

Calculation of the POD and FRR using the collected data: To find the POD, one could calculate the ratio of "correct alarms" to the total number of "contaminations" actually present, while for the FRR, one could calculate the ratio of "false alarms" to the total number of nonexistent "contaminations" (false positives).

 

Thank you very much.

 

Excuse my ignorance but what is POD and FRR ( false reject ratio? )

 

thank you,

[Planck]

Good morning all, 

 

I start a topic because I did not find (me or my colleagues) any legislation that help us to determinate whether we are performing a correct checking of the CP of metal detector or nor.  

 

My question is, there are some legislation or directions that specifies that the verification of the function of the metal detector with tests must be carried with product? 

 

In other words, it is possible to perform a verification of the correct function of the equipment without the presence of the actual product in the conveyor¿? Like putting the actual tests of the MD in the conveyor and check if there are an actual detection + stopping + alarm + rejection or whatever the corrective action established takes? 

 

We already ask at the FSIS, and they where quite vague with their answer:

 

FSIS does not have a guideline or Directive specific to metal detector validation. It depends on the manufacturer's specifications, the establishment's hazard analysis decisions, and how the establishment supports its hazard analysis decisions. Establishments often use seeds or wands to determine whether the metal detector is functioning correctly for certain sizes or composition of metal. Depending on the products, packaging, and processes, the seeds or wands may be placed in a package or container with products or used by themselves for testing to ensure a metal detector is functioning properly.

 

I am not an expert, I supposed that the signal that the metal detector receive, when the tests are detected, there are quite few difference if there are product or not, as the product should be seen by the metal detector as a "sample blank". If there are any signal (as a product/matrix always produce a background noise), it is not significant and the metal detector must work in all the circumstances. 

 

Thanks in advance, 

I am a service engineer for metal detectors, I have been working in this field continuously for 12 years and for close to 10 years before that I was also working in R&D, so the total technical work time is over 20 years.

I'm happy to share a little bit of my opinion here:

1, I don't know much about the various regulations or standards, but I don't think that as an implementation standard you should be too specific in requiring technical details.

Because you can't get a handle on the technical details, and there are design differences in products from different companies, and there are patented technologies, and technology is always moving forward, etc.

 

2, As an end user, with different production lines, different products, different models of metal detectors, different detection methods and possibly different rejection methods.

As a prime example:

If your product is a packaged product rather than bulk, and it is relatively long, and the metal detector is set to package detection mode, meaning that the metal detector can accurately reject only packages with metal in them, it is a must that your test piece must be tested with your product (at least the complete package).

If you don't, then it's almost impossible to validate your machine. Unless you have an extremely slow production line, which is almost impossible.

Even in this case, you may need to verify the front, middle and end of the package.

 

3,Even if you don't look at the specification / standard at all, then in fact if you use a metal detector carefully, then you are still certain to find an anomaly.

In the simplest terms, your good product shouldn't trigger a rejection.

Similarly, when you try to verify a metal detector with a test piece, it has to trigger a reject.

And most importantly, in both of the above cases, whenever a failure occurs, you should look for the cause, especially in the case of verifying metal test pieces.

 

I'm sure if you can handle the above process smoothly, you won't be bothered anymore.

 

Thanks. FYI

[Planck]

Excuse my ignorance but what is POD and FRR ( false reject ratio? )

 

thank you,

Probability of Detection (POD)

If you test your metal detector a thousand or even ten thousand times in a row using your metal test piece, will it all be rejected?

 

False Reject Rate (FRR)

How many packs were rejected in total when your metal detector went through 1,000 packs, 10,000 packs, and 100,000 packs in a row?

If there are no metal contaminants in the package at all, then this is a False Reject.

 

Both of these numbers can go up to 100,000 or more for the customers we directly support.

Of course, this process is complex and affected by many factors.

 

https://www.ifsqn.co...-rate-analysis/

 

thanks...FYI

[AitorMenta]

Thank you all for your comments. 

 

Still wishing for a legislation, maybe it is only a mater of time....

 

Anyway, best wishes for all