32 replies to this topic

[bala_s]

Dear members,

we are using a metal detector and calibrating with Fe and SS standard test pieces. we do not have Non-fe test piece.

is it necessory to have non-fe standard
i had checked with 1mm aluminium piece and it is detecting.
what are all will come under Fe and Non-fe.

please can any one reply to this mail


regards/bala

[Simon]

Can anyone with experience of metal detectors share experience with Bala?

[Robert Rogers]

The testing for the various metal types should be based on the potential hazards discovered through the hazard analysis stage.

Typically, Fe is easiest to detect then NFe then the various grades of SS. NFe materials copper, aluminum and brass are most common NFe materials.

The size that will be detected depends on several factors including the size of detector, operating frequency of the detector, installation factors and product conditions.

The best situation would be the detector is able to completely operate at maximum sensitivity with no product and have no effect from the "environment". This way the only factor affecting sensitivity would be the product itself, then setting the detector to successfully inspect the product with little to no false rejects. Once this setting is found placing metal samples as close to the absolute center of the detector (the weakest sensitivity point) as possible will show what is possible.

[jglopez]

Typical size of test pieces ( spherical diametre) are: 1 mm Fe, 2.2-2.5 Non ferrous such as Cu and Al and 3mm for stainless steel.

[Charles.C]

Dear balas,

you don't mention if a standard / audit is involved.?
For BRC food, IMEX, the absence of a non-Fe standard may/may not lead to a NC. Perhaps that answers yr direct query.

The technical logic (ie relative sensitivities) is from memory of discussions on this forum less simple, you might try a little searching here.

Rgds / Charles.C

[infoiqc]

Clause 4.10.3.5 of BRC v.6
Metal detector checking procedures shall be based on best practice and shall as a minimum include:
...• tests carried out using separate test pieces containing ferrous metal, stainless steel and typically non-ferrous metal, unless the product is within a foil container.

So under BRC unless you have a good reason you need 3 test pieces.

Gail
infoiqc

Dear balas,

you don't mention if a standard / audit is involved.?
For BRC food, IMEX, the absence of a non-Fe standard may/may not lead to a NC. Perhaps that answers yr direct query.

The technical logic (ie relative sensitivities) is from memory of discussions on this forum less simple, you might try a little searching here.

Rgds / Charles.C

[Brian Meek]

Hi Balas

From a metal detection perspective you dont actually require the stainless sample because strictly speaking that is also a non ferrous sample.

Metal detectors look for two things, magnetic properties for ferrous metals and conductive proerties for non ferrous properties in very simple terms.

As posted earlier BRC require the use of all three test samples in there requirement so you must use all three.

There is also another overlooked factor within the stainless steel category regarding the grade of stainless to use. There are quite a few different grades of stainless steel available but the hardest to find is 316 grade and the easiest to find is 304 grade. Your factory and its machinery will have 316 grade in it so you should test for this.

All the best

Brian

[Charles.C]

Hi Balas

From a metal detection perspective you dont actually require the stainless sample because strictly speaking that is also a non ferrous sample.

Metal detectors look for two things, magnetic properties for ferrous metals and conductive proerties for non ferrous properties in very simple terms.

As posted earlier BRC require the use of all three test samples in there requirement so you must use all three.

There is also another overlooked factor within the stainless steel category regarding the grade of stainless to use. There are quite a few different grades of stainless steel available but the hardest to find is 316 grade and the easiest to find is 304 grade. Your factory and its machinery will have 316 grade in it so you should test for this.

All the best

Brian

Dear Brian Meek,

Appreciate yr various recent inputs.

I can follow yr opening sentence but, as stated, I didn't quite get the logic in yr last paragraph. Steel properties definitely not my direct area of expertise but based on other posts here and a little digging i did for a previous thread, the practical 304/316 usage situation can vary considerably. This has certainly been the case in food factories IMEX also. Additionally the detailed aspects of steel composition vis-a-vis metal detectors seem to be somewhat confused / confusing.

Please refer my (amateurish) review / post/ (included)links and surrounding thread here -

http://www.ifsqn.com...dpost__p__52433

Rgds / Charles.C

@Balas - any response would be appreciated.

[GMO]

Hi Balas

From a metal detection perspective you dont actually require the stainless sample because strictly speaking that is also a non ferrous sample.

Metal detectors look for two things, magnetic properties for ferrous metals and conductive proerties for non ferrous properties in very simple terms.

As posted earlier BRC require the use of all three test samples in there requirement so you must use all three.

There is also another overlooked factor within the stainless steel category regarding the grade of stainless to use. There are quite a few different grades of stainless steel available but the hardest to find is 316 grade and the easiest to find is 304 grade. Your factory and its machinery will have 316 grade in it so you should test for this.

All the best

Brian

I disagree with the first point because IME Stainless Steel is almost always harder to detect than other Non Fe metals and as other posters have said, BRC asks for all three.

The last point though is really interesting and almost ended up being a Non Conformance on our BRC audit. We have both 304 and 316 stainless on site and our auditor was trying to claim that we should test for both 304 and 316 stainless. Do you have any evidence / validation info that 316 is harder to detect so if it's raised next year I can defend our position more comfortably?

[Alan Johnson]

I disagree with the first point because IME Stainless Steel is almost always harder to detect than other Non Fe metals and as other posters have said, BRC asks for all three.

The last point though is really interesting and almost ended up being a Non Conformance on our BRC audit. We have both 304 and 316 stainless on site and our auditor was trying to claim that we should test for both 304 and 316 stainless. Do you have any evidence / validation info that 316 is harder to detect so if it's raised next year I can defend our position more comfortably?

Hi,


It's complicated. If we assume that a standard metal detector detects the magnetic and conductive properties (where present) in a metal then the higher these properties are the more likely that the metal can be detected. Both the conductive properties of 304 and 316 can be viewed practically as a constant with 316 being 9% 'less conductive' than 304, so in this sense 304 will be more detectable. With regard to the magnetic quality both are classed as austenitic and therefore have a little response in the presence of a permanent magnetic field. However, 304 stainless when cold worked can appear to take on magnetic properties and potentially be more 'detectable'; the 304 samples provided by reputable metal detector manufacturers should be magnetically inert. However, there have been instances where stainless test samples have been supplied to the market of a 440 grade (easily detectable/magnetic), merely labelled as stainless and 'magnetic' 304. The complexity arises because general purpose metal detectors operate with an oscillating magnetic field and the frequency of operation will largely be dependent on the application, lower frequencies tend to boost the magnetic response to metals whereas higher frequencies boost the conductive response. You tend to find lower frequencies with conductive products (mix of salt & free water) or conductive packaging like metallised film (not foil) and higher frequencies with non conductive products crackers, cereals, deep frozen (-18C) etc. There has been a general trend in the market over the years to increase operating frequencies to boost detection of stainless steels both for conductive and non-conductive products, so there may be an age issue to consider. If the product/packaging is non conductive then there is little for the metal detector to compensate for as the product is relatively inert (there are exceptions) and the detection of non ferrous metals (you can include 304/316 in this as being 'weak' non ferrous) is relative predictable. However, if the product is conductive then the metal detector has to compensate for this by effectively tuning it out, this compensation point can in some instances tune out 'weak' non ferrous samples, so you in theory could be in a situation where you could tune out either a 316 or 304 test sample for a particular product. How much separation between detectability of 316 and 304 for a given application is difficult to predict, especially as test samples tend to increment in 0.5mm steps. You may had a situation where a 2.0mm 304 is readily detectable and the same size 316 is on the margin, and 1.5mm in both materials would not be detectable. 2.0mm would likely be adopted as a standard and the assumption made that both materials are equal in terms of detection. Theoretically, then you cannot assume that a certain 304 mm diameter size test sample will result in the same size 316 test sample being detected and vice versa although it is more likely that detection of 316 will result in the same level of detection of 304, due diligence and compliance aside.

Hope this helps

Best regards

Alan

[Charles.C]

Dear Alan Johnson,

Thks for yr input but I hv to confess that yr post got a bit too complicated / tortuous for my brain.

Apologies if my lack of understanding but I get the impression that you hv concluded that the relative ferrous / non-ferrous etc detection claims which manufacturers provide for their instruments are often meaningless. However this seems contrary to user reported experiences with supplied test strips. (Or perhaps you were referring to real product situations as compared to standardised testing [eg axially placed spheres] scenarios).
Perhaps one needs to talk about specific response magnitudes / baselines to meaningfully compare. Such data is never published AFAIK, except in the operator manuals maybe (?).

Rgds / Charles.C

[GMO]

Hi,


It's complicated. If we assume that a standard metal detector detects the magnetic and conductive properties (where present) in a metal then the higher these properties are the more likely that the metal can be detected. Both the conductive properties of 304 and 316 can be viewed practically as a constant with 316 being 9% 'less conductive' than 304, so in this sense 304 will be more detectable. With regard to the magnetic quality both are classed as austenitic and therefore have a little response in the presence of a permanent magnetic field. However, 304 stainless when cold worked can appear to take on magnetic properties and potentially be more 'detectable'; the 304 samples provided by reputable metal detector manufacturers should be magnetically inert. However, there have been instances where stainless test samples have been supplied to the market of a 440 grade (easily detectable/magnetic), merely labelled as stainless and 'magnetic' 304. The complexity arises because general purpose metal detectors operate with an oscillating magnetic field and the frequency of operation will largely be dependent on the application, lower frequencies tend to boost the magnetic response to metals whereas higher frequencies boost the conductive response. You tend to find lower frequencies with conductive products (mix of salt & free water) or conductive packaging like metallised film (not foil) and higher frequencies with non conductive products crackers, cereals, deep frozen (-18C) etc. There has been a general trend in the market over the years to increase operating frequencies to boost detection of stainless steels both for conductive and non-conductive products, so there may be an age issue to consider. If the product/packaging is non conductive then there is little for the metal detector to compensate for as the product is relatively inert (there are exceptions) and the detection of non ferrous metals (you can include 304/316 in this as being 'weak' non ferrous) is relative predictable. However, if the product is conductive then the metal detector has to compensate for this by effectively tuning it out, this compensation point can in some instances tune out 'weak' non ferrous samples, so you in theory could be in a situation where you could tune out either a 316 or 304 test sample for a particular product. How much separation between detectability of 316 and 304 for a given application is difficult to predict, especially as test samples tend to increment in 0.5mm steps. You may had a situation where a 2.0mm 304 is readily detectable and the same size 316 is on the margin, and 1.5mm in both materials would not be detectable. 2.0mm would likely be adopted as a standard and the assumption made that both materials are equal in terms of detection. Theoretically, then you cannot assume that a certain 304 mm diameter size test sample will result in the same size 316 test sample being detected and vice versa although it is more likely that detection of 316 will result in the same level of detection of 304, due diligence and compliance aside.

Hope this helps

Best regards

Alan

I think I can conclude "it's not an easy answer" and "I should get advice from our MD contractor and probably do some validation testing as it will be product dependent". Is that fair?

[Brian Meek]

Hi GMO

Yes it is all product relative and dependant on you having a stable and consistant product with little or no variability. The test samples whatever they are will always be the constant.

The key to all of these problems is to be realistic with your expectations and not try for that extra sensitivity because a metal detectors will soon turn into a very good product detector.

Kind regards

Brian

[Alan Johnson]

Dear Alan Johnson,

Thks for yr input but I hv to confess that yr post got a bit too complicated / tortuous for my brain.

Apologies if my lack of understanding but I get the impression that you hv concluded that the relative ferrous / non-ferrous etc detection claims which manufacturers provide for their instruments are often meaningless. However this seems contrary to user reported experiences with supplied test strips. (Or perhaps you were referring to real product situations as compared to standardised testing [eg axially placed spheres] scenarios).
Perhaps one needs to talk about specific response magnitudes / baselines to meaningfully compare. Such data is never published AFAIK, except in the operator manuals maybe (?).

Rgds / Charles.C

Hi Charles, metal detection is a competitive market and the sensitivity quotations provided by suppliers tend to be best case, unless product testing has been carried out beforehand, when a more application specific sensitivity can be provided. There is often a rider on quotations 'subject to product effect and environmental conditions' the latter being less of an issue with modern equipment that tends to be more stable across a wide range of operating environments. Suppliers generally provide a range of test strips with new equipment, there being the realisation that what is quoted may not actually be achieved during production conditions due to the factors mentioned above. Generally depending on how adverse the effect of the product is on sensitivity will determine the margin between what is quoted and what is achieved during production. The sensitivity claims are not meaningless as such, the closest analogy I can come up with is car manufacturers that quote fuel consumption figures, all pretty much best case, hardly ever achieved in the real world but you have to start somewhere.

Best regards

Alan
Alan

[Alan Johnson]

I think I can conclude "it's not an easy answer" and "I should get advice from our MD contractor and probably do some validation testing as it will be product dependent". Is that fair?

I would say that is fair. The goal is to achieve the best level of sensitivity (without false rejects) for a given product, and a robust test regime that validates not only sensitivity but also stress tests the reject system. This is why the various retailer COP's focus on the overall system capability not just sensitivity, as a significant number of consumer complaints exceeded the metal detector test sample size for a given material. You can have the most sensitive metal detector available but if the reject does not remove the contaminated product virtually any size can reach the consumer.

Best regards

Alan

[Charles.C]

Dear Alan Johnson,

Thks yr response.

As you (and the Walrus?) noted, one has to start somewhere.

After a beer and another read of yr previous post, my conclusion to yr interesting (alarming?) list of caveats is that (1) a well-defined test sample/ “neutral” procedure is a necessary starting point for any comparison of (intrinsic) machine capabilities. (2) The “field” performance for relevant matrices will obviously be an equally important practical aspect and probably dominant if other basic specs are similar across brands (are they ??).
The above general approach is fairly typical IMEX for measurement equipment and others (eg disinfectants).
Post#2 of this thread more or less spells out a procedure to (1) above.
My previous referred review link overviews composition aspects also, eg paras (1,4).

Accordingly, I assume, although it is rarely stated, that the caveats are allowable / allowed for in commercial machine specs. Otherwise any published basic sensitivity data would indeed seem to be meaningless.

As far as matrix effects go, I do appreciate yr points that ultimately such may well be crucial. Unfortunately this thread has no specific data on application so one can maybe only comment on the “pure” case for the moment. As discussed in other references posted on this forum, it behooves the user to evaluate (or the seller to supply?) the consequences for their own matrix, for example using dummy procedures as discussed in various other threads on this forum.

So I would like to re-pose one simple but typical query from this thread(s), eg –

For any well-defined materials / procedures, are there any typical quantitative data available regarding relative detector sensitivities for a progression of “stainless steel”, “ferrous”, etc materials. And different brands of detector ? (I appreciate that machine parameters such as aperture size may well restrict available comparisons, even if I don’t exactly know why [ ].)

Hope the above is not too scrambled and i appreciate yr patience. TGISa.

Rgds / Charles.C

[Alan Johnson]

Dear Alan Johnson,

Thks yr response.

As you (and the Walrus?) noted, one has to start somewhere.

After a beer and another read of yr previous post, my conclusion to yr interesting (alarming?) list of caveats is that (1) a well-defined test sample/ “neutral” procedure is a necessary starting point for any comparison of (intrinsic) machine capabilities. (2) The “field” performance for relevant matrices will obviously be an equally important practical aspect and probably dominant if other basic specs are similar across brands (are they ??).
The above general approach is fairly typical IMEX for measurement equipment and others (eg disinfectants).
Post#2 of this thread more or less spells out a procedure to (1) above.
My previous referred review link overviews composition aspects also, eg paras (1,4).

Accordingly, I assume, although it is rarely stated, that the caveats are allowable / allowed for in commercial machine specs. Otherwise any published basic sensitivity data would indeed seem to be meaningless.

As far as matrix effects go, I do appreciate yr points that ultimately such may well be crucial. Unfortunately this thread has no specific data on application so one can maybe only comment on the “pure” case for the moment. As discussed in other references posted on this forum, it behooves the user to evaluate (or the seller to supply?) the consequences for their own matrix, for example using dummy procedures as discussed in various other threads on this forum.

So I would like to re-pose one simple but typical query from this thread(s), eg –

For any well-defined materials / procedures, are there any typical quantitative data available regarding relative detector sensitivities for a progression of “stainless steel”, “ferrous”, etc materials. And different brands of detector ? (I appreciate that machine parameters such as aperture size may well restrict available comparisons, even if I don’t exactly know why [ ].)

Hope the above is not too scrambled and i appreciate yr patience. TGISa.

Rgds / Charles.C

Hi Charles, there exists quantitive data for different brands of detectors, both held by the metal detector manufacturers (in-house bench-marking) and those carried out by the larger food manufacturers when evaluating prospective suppliers, which tend to be product relative. Both sources are pretty much confidential i.e. not in the public domain. Nearly all suppliers of detectors will have internal documents (graphs/databases) that will provide best case sensitivities for ferrous/non-ferrous/stainless steel for given aperture sizes, used for quotation/estimation purposes that may or may not be moderated with real-world data. These also tend not to be in the public domain also, as they can be open to mis-interpretation due to the high degree of variability that can be found in real-world applications. If there is no product effect then as a rule of thumb the detection level for ferrous metals will be the same for non ferrous and stainless steel will be 20 - 30% less detectable. Unfortunately, there is no practical rule of thumb that can be applied where the product/packaging is conductive (or magnetic) due to the magnitude and variability present in some cases. Yes you can apply a general downgrade performance factor that will produce a guideline specification. The safest way to proceed with product that may adversely affect a metal detector's performance is to get it tested by the prospective manufacturer(s) and obtain a definitive spec. All the major players have in-house test facilities and the ability to produce test reports.

Here is an example of how it can go wrong. Take a ready meal (TV dinner) meat, sauce, rice. Frozen below -18C, no (or virtually no) effect from the product, a supplier quotes according to the above i.e. sensitivity determined by the aperture size alone, Ferrous, Non Ferrous equal and Stainless in proportion. This quoted specification is achieved on start up but the production line is then stopped with packs in transit, the product thaws out on the surface (big effect) and the detector rejects everything until normality is resumed. If this was a frequent occurrence then some form of compensation would need to be applied to the detector in order to accommodate frequent line stoppages, this would (most likely) result in a fall off in sensitivity. Alternatively the speed of transit through the freezer could change resulting in the core of the product not freezing, similar (but less adverse) effect.

Are there significant differences in performance between the major players? In terms of sensitivity I would say not (OK there will be someone out there that will have data saying on product A we found supplier X was better than Y) but if you line them up in a test lab there's not a lot in it. The key differentiators tend to be Robustness, Reliability, Ease of use, Ability to integrate into the production line, Availability of support if things go wrong. You can have the most sensitive machine on the market but it's no use if it dies every time someone waves a wet cloth at it or you need a degree in computer science to set it up. I think it is worth remembering that these metal detector (performance) test strips are not really representative of real world metallic foreign bodies, which tend to be irregular in shape. A metal detector set to detect 2.0mm stainless steel for example could quite easily pass a 20mm long contaminant if the shape and orientation was worse case.

Recent improvements in 'metal detector' sensitivity have been incremental and this it likely to be the case in the future, in terms of measurable improvements (especially to stainless steel) X-ray is the technology to look at, but obviously there is cost to consider.

I hope this has been useful, and has answered more questions than it has posed!

Best regards

Alan

[Charles.C]

Dear Alan Johnson,

Thks for the helpful comments. Yr post prompted me to do a little IT searching.

I now slightly see just how complex this topic actually is. I also appreciate how much a part is (electronically) played by the product itself.

Based on my own ability to understand their content, I hv attached 4 documents (seemingly all different companies) below, mainly for other people who might still be interested in this subject.

I had some uncertainty whether all the accessible refs (mostly 2000-04) were still current in these technical times. However most texts implied recent changes hv mainly been in the later digital signal handling capabilities rather than fundamentals so c'est la vie.

I hv extracted from the above and others (a) some text and (b) some data on sensitivities and posted into sheets 1,6 of excel document attached. As you stated, raw quantitative data is invisible so far.

The text sheet is primarily only illustrative however I noted that one “expert” (ref1) seemed to disagree somewhat with the other 2 over the “similar” ease of detection between Fe/non-Fe contaminants. Although the table I added which appears in other texts also does seem rather confident.

I was interested to see some support (again ref1) for my previous post regarding the possibility of a brand(s) gaining ascendancy. A critical review on the subject of metal detectors might reveal such, but so far hv not seen any.

The several sets of sensitivity data I located are amazingly (suspiciously?) similar IMO. I can roughly summarise them as –

Dry (= Non-Conductive) Product
Aperture
<=50mm, Fe/Non-Fe = 0.8mm – 1.0, stainless steel (SS)(Limited Data) = 1.2
<=125mm, Fe/Non-Fe = 1.2mm – 1.5, stainless steel (SS) = 1.5-1.6
<=200mm, Fe/Non-Fe = 1.2mm – 2.0, stainless steel (SS) = 2.0-2.2

Wet (=Conductive) Product (Limited SS Data)
Aperture
<=50mm, Fe = 1.5, Non-Fe = 2.0, SS = 2.5
<=125mm, Fe = 2.0, Non-Fe = 2.5, SS = 3.5
<=200mm, Fe = 2.5, Non-Fe = 3.0, SS = 4.0

One publication was more wary and only gave ratios, similar to some textual comments –

Non-Fe = (1.1-1.3)Fe
SStypeA = 1.2Fe
SStypeB = 1.3Fe
SStypeC = 1.5Fe

The only obvious comment is that it is not too clear (to me anyway) just how meaningful these figures really are in the absence of any parallel matrix information. Though no problem for auditors presumably.

As noted previous posts, GMO’s specific query remains floating, ie is the detection capability for “316” always < “304”?, regardless of the matrix. Validatable measurements on specific product are inevitable.

I noticed this comment in one of my attachments –

Processing plants in the food and pharmaceutical industries use the two most common grades, 304(L) and 316. The detectability of these grades are further hindered when the product is either wet, containing a high salt content or both, thus contributing to a high product signal.
As the properties of stainless can be modified by machining ( increasing the magnetic effect ) specific sensitivity figures are difficult to quote. In general it can be expressed as a ratio to ferrous, at best 1:1.5 rising to 1:2.5

Next auditor question - Are all the 316/304 locations of (known) same species ?

 s1 - Metal-Detection-Basics (fortress) .pdf   98.29KB   224 downloads
 s2 - metal detection basics (Loma).pdf   86.22KB   182 downloads
 s3 - Guide_to_Metal_Detection-1(Loma).pdf   408.88KB   186 downloads
 s4 - Metal-Detection-Guide (Advanced Det.Sys.).pdf   794.03KB   210 downloads
 Metal detections2.xls   409.5KB   262 downloads

Rgds / Charles.C

[Alan Johnson]

Dear Alan Johnson,

Thks for the helpful comments. Yr post prompted me to do a little IT searching.

I now slightly see just how complex this topic actually is. I also appreciate how much a part is (electronically) played by the product itself.

Based on my own ability to understand their content, I hv attached 4 documents (seemingly all different companies) below, mainly for other people who might still be interested in this subject.

I had some uncertainty whether all the accessible refs (mostly 2000-04) were still current in these technical times. However most texts implied recent changes hv mainly been in the later digital signal handling capabilities rather than fundamentals so c'est la vie.

I hv extracted from the above and others (a) some text and (b) some data on sensitivities and posted into sheets 1,6 of excel document attached. As you stated, raw quantitative data is invisible so far.

The text sheet is primarily only illustrative however I noted that one “expert” (ref1) seemed to disagree somewhat with the other 2 over the “similar” ease of detection between Fe/non-Fe contaminants. Although the table I added which appears in other texts also does seem rather confident.

I was interested to see some support (again ref1) for my previous post regarding the possibility of a brand(s) gaining ascendancy. A critical review on the subject of metal detectors might reveal such, but so far hv not seen any.

The several sets of sensitivity data I located are amazingly (suspiciously?) similar IMO. I can roughly summarise them as –

Dry (= Non-Conductive) Product
Aperture
<=50mm, Fe/Non-Fe = 0.8mm – 1.0, stainless steel (SS)(Limited Data) = 1.2
<=125mm, Fe/Non-Fe = 1.2mm – 1.5, stainless steel (SS) = 1.5-1.6
<=200mm, Fe/Non-Fe = 1.2mm – 2.0, stainless steel (SS) = 2.0-2.2

Wet (=Conductive) Product (Limited SS Data)
Aperture
<=50mm, Fe = 1.5, Non-Fe = 2.0, SS = 2.5
<=125mm, Fe = 2.0, Non-Fe = 2.5, SS = 3.5
<=200mm, Fe = 2.5, Non-Fe = 3.0, SS = 4.0

One publication was more wary and only gave ratios, similar to some textual comments –

Non-Fe = (1.1-1.3)Fe
SStypeA = 1.2Fe
SStypeB = 1.3Fe
SStypeC = 1.5Fe

The only obvious comment is that it is not too clear (to me anyway) just how meaningful these figures really are in the absence of any parallel matrix information. Though no problem for auditors presumably.

As noted previous posts, GMO’s specific query remains floating, ie is the detection capability for “316” always < “304”?, regardless of the matrix. Validatable measurements on specific product are inevitable.

I noticed this comment in one of my attachments –



Next auditor question - Are all the 316/304 locations of (known) same species ?

 s1 - Metal-Detection-Basics (fortress) .pdf   98.29KB   224 downloads

 s3 - Guide_to_Metal_Detection-1(Loma).pdf   408.88KB   186 downloads
 s4 - Metal-Detection-Guide (Advanced Det.Sys.).pdf   794.03KB   210 downloads
 Metal detections2.xls   409.5KB   262 downloads

Rgds / Charles.C

Hi Charles, a couple of comments. The "quote" is a bit misleading in that for conductive products that produce a product effect they are generally classed as ionic compounds, which conduct electricity when molten or in solution, but not as a solid. Hence you can have a bag of salt, which will produce little in the way of product effect but place the contents into water and the signal will be huge. Conversely, pure/distilled water when passed through a metal detector will not produce a signal.


In addition, (as mentioned earlier) 304 stainless is prone to being magnetised when cold worked, to my knowledge it it not possible practically to magnetise 316 in the same manner. 316 could therefore offer more certainty in terms of the signal it produces during the test process. 304 samples should be sourced 'magnetically inert' or annealed prior to being placed on the market otherwise potentially there could be variable signals depending on how magnetised it was. If 304 stainless test samples are being used then I would recommend that they are sourced from reputable suppliers with at least some guarantee that they are annealed.


The Safeline download is actually an old Loma USA publication.


Best regards


Alan

[Charles.C]

Hi Alan,

The Safeline download is actually an old Loma USA publication

.

Thks for the tip. Shouldn't click and read at the same time.

I actually particularly liked this one since it was the first time I saw a discussion of why they recommend to also place the test wand at front / rear. So far I have never seen anyone actually do it though.? Same as the brilliant idea to do 5 repeat runs I saw somewhere else.

Regarding the 304/316 - yes i also picked up the possible 304 magnetic acquisition in an earlier thread but this would presumably only make it even easier to detect than a "true" non-magnetic form? Or not ?
It seemed to me the initial question is whether one can expect that a "properly prepared" 304 wand of a given size can be consistently more easy to detect than the same size of "pp" 316 ? I suspect no way to predict. And probably no easy way to trial/error validate either unless one gets lucky.?

Rgds / Charles

[Alan Johnson]

Hi Alan,

.

Thks for the tip. Shouldn't click and read at the same time.

I actually particularly liked this one since it was the first time I saw a discussion of why they recommend to also place the test wand at front / rear. So far I have never seen anyone actually do it though.? Same as the brilliant idea to do 5 repeat runs I saw somewhere else.

Regarding the 304/316 - yes i also picked up the possible 304 magnetic acquisition in an earlier thread but this would presumably only make it even easier to detect than a "true" non-magnetic form? Or not ?
It seemed to me the initial question is whether one can expect that a "properly prepared" 304 wand of a given size can be consistently more easy to detect than the same size of "pp" 316 ? I suspect no way to predict. And probably no easy way to trial/error validate either unless one gets lucky.?

Rgds / Charles

In terms of placing metal test wand at the front and rear of the pack, most UK retailer codes of practice recommend this in one form or another. Primarily it is because a metal detector on its own does not know the position of the test wand in the pack. It can only sense (not very accurately) the position of the metal sphere as it passes through the coils. To ensure accurate rejection of the product on a conveyor based system there normally has to be a photo eye that senses the leading edge of the pack and this is wired so that the pack is positively rejected wherever the test wand is positioned. Metal detectors have been known not to reject positively without photo eyes or if the reject timing has been set badly, so the test is worse case. There are even additional test routines that check the functionality of these additional failsafe devices at the start of a production shift.

316 is marginally less conductive than 304 so in theory (magnetised issues aside) 304 should be slightly easier to detect. A 'magnetised' 304 will invariably be easier to detect, and probably more so in a product that is conductive compared to 316.

Best regards

Alan

[Charles.C]

In terms of placing metal test wand at the front and rear of the pack, most UK retailer codes of practice recommend this in one form or another. Primarily it is because a metal detector on its own does not know the position of the test wand in the pack. It can only sense (not very accurately) the position of the metal sphere as it passes through the coils. To ensure accurate rejection of the product on a conveyor based system there normally has to be a photo eye that senses the leading edge of the pack and this is wired so that the pack is positively rejected wherever the test wand is positioned. Metal detectors have been known not to reject positively without photo eyes or if the reject timing has been set badly, so the test is worse case. There are even additional test routines that check the functionality of these additional failsafe devices at the start of a production shift.

316 is marginally less conductive than 304 so in theory (magnetised issues aside) 304 should be slightly easier to detect. A 'magnetised' 304 will invariably be easier to detect, and probably more so in a product that is conductive compared to 316.

Best regards

Alan

Dear Alan,

Thks comments. I re-read yr earlier post #10. It seems remarkable that the tuning can selectively eliminate a 304 detection whose response was presumably (in the no product scenario) greater than 316’s using the same size sphere. as you say, I guess it all depends on the relative amplitudes (eg original closeness) / their proximity to the theshhold for detection and what the tuning is actually doing .

I did (previously) notice this advertising blurb –

Test your metal detectors at regular intervals, every hour is common practice, use three test samples made from Ferrous (Chrome Steel), Non-Ferrous (Brass) and grade 316 Stainless Steel. Grade 316 Stainless is used as it has Non-Magnetic properties, which means that it is one of the most difficult grades of Stainless Steel for a metal detector to find. So the theory is, if it can find this grade then it will certainly find the others.

http://www.fastecservices.com/test-standards/

It would be interesting to see their validation, and also their definition of “others”.

One wonders if the next BRC will have 4 test samples (now looking the likeliest conclusion for GMO's auditor [trailblazer?] IMO). I suppose one alternative would be to avoid 304 but IMEX this is not so easy $wise.

Rgds / Charles

[GMO]

Yeah. If you think how many packs are wasted every year and the insistance of some retailers to do three consecutive packs (so three packs are wasted) increasing to four would be a significant waste for some companies (into the £10,000s potentially) for what? Does anyone actually think testing for 304 and 316 will improve safety? I certainly don't.


I could do without trailblazing auditors thank you...

[Alan Johnson]

Yeah. If you think how many packs are wasted every year and the insistance of some retailers to do three consecutive packs (so three packs are wasted) increasing to four would be a significant waste for some companies (into the £10,000s potentially) for what? Does anyone actually think testing for 304 and 316 will improve safety? I certainly don't.


I could do without trailblazing auditors thank you...

The discussion has been to a large degree theoretical, which kind of like doesn't bear out in practice. It depends on how close the test sample sizes are to the threshold of detection. In my experience rarely is this the case. There are two reasons, one is food manufacturers tend to use the same test samples irrespective of the improvements of metal detector performance over the years and the other is that it is feasible to have a number of different sample specifications for different products on one production line if one is testing close to the limit. Product A would have X range of samples, Product B Y range etc., this could be complex to manage. The threshold example; a situation where a metal detector for a given product is capable of detecting 1.0mm Ferrous, 1.2mm Non Ferrous and 2.0mm Stainless (forget 304/316); QA test to 1.5 Ferrous, 2.0mm Non Ferrous and 3.0mm Stainless as it is their 'standard', the Stainless sample is going to be detected whether it is 304 or 316 as it is BIG compared to the threshold. This situation may not be the the case if it is a new line/plant or the result of a supplier evaluation, but it tends to be the norm. Complex test processes do not necessarily improve food safety in this respect, in my opinion the Non Ferrous test sample could be ditched and testing carried out with just the Ferrous sample and 316 Stainless, simpler system, less wastage, same level of security. One of the major reasons why metal contaminants reach the consumer is failure of the system; not the margin between threshold of detection and the samples used, as metal detectors calibrate themselves for the best outcome i.e. just to not detect the product NOT to just detect the samples.

Best regards

Alan

[Charles.C]

Dear Alan,

Stainless sample is going to be detected whether it is 304 or 316 as it is BIG compared to the threshold.

I think you mean the contaminant physical size is big ? You may well be correct but the validation of such logic will surely be problematic, both conceptually and practically ?.

IMO the basic, current, (auditor / processor) problem is that relevant data / accessible references which demonstrate detection limits for a variety of real food matrices are either unpublished or unreferenced. ? More likely the latter I suspect.

Rgds / Charles