Jump to content

  • Quick Navigation
Photo

Is Non-fe standard necessary to calibrate the metal detector?


  • You cannot start a new topic
  • Please log in to reply
30 replies to this topic

#26 Alan Johnson

Alan Johnson

    Grade - Active

  • IFSQN Associate
  • 15 posts
  • 7 thanks
1
Neutral

  • United Kingdom
    United Kingdom
  • Gender:Male
  • Location:South of England

Posted 06 July 2012 - 10:01 AM

Dear Alan,



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relevant data / accessible references which demonstrate detection limits for a variety of real food matrices are either unpublished or unreferenced.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


Hi Charles, I mean that the signal that the test samples give off with respect to the threshold is BIG so in essence the marginality between 304 and 316 in most practical instances isn't that relevant. The issue (para 2 above) is that there is a lack of understanding of what is achievable/practical/reasonable with the physics of metal detection and the food industry/audit regimes rely on information from the manufacturers to arrive at something that minimises risk but is workable, in terms of the technology delivering 'metal free' food I think we arrived at this point a while back. However, metal still reached the consumer, so it's either below the specification of the equipment or above, the latter indicates a system (line control) failure.

Regards

Alan


  • 0
Alan@S+S

#27 Charles.C

Charles.C

    Grade - FIFSQN

  • IFSQN Moderator
  • 12,590 posts
  • 3280 thanks
350
Excellent

  • Earth
    Earth
  • Gender:Male
  • Interests:SF
    TV
    Movies

Posted 07 July 2012 - 10:34 PM

Dear Alan,

Sorry for my misunderstanding.
I certainly agree there is a lack of general understanding regarding MD capabilities. i also agree that processors are probably quite happy to be able to avoid a detailed appreciation of all the twists and turns discussed in this thread. :smile: Unfortunately, the status quo may perhaps be changing somewhat in BRC's case.

I will try and simply illustrate the problem as i see it, below is a machine spec from attachment given. The manual appears to be ca. 2003 so maybe no longer reliable as to specific numbers, (or degree of information content?) –

Sensitivity Ratios
Dependent on frequency of coil system e.g. at 300 kHz
Non Ferrous x 1.1 to 1.3 Fe diameter (depending on the metal)
Stainless Steel x 1.2 to 1.5 Fe diameter (depending on type, most difficult to detect is type 316)

IMHO, the specs are reasonably (ie probably conservatively) stated from a relative sensitivity POV although I find it strange that no claim is made as to Fe capability. This prevents any absolute use of the preceding info. of course.

Taken at face value, from a presumably reliable source, this spec. gives a direct (theoretical) validation for GMO’s auditor query.

But there are zero actual examples in the 168 page manual (perhaps there was a separate book but i do not see any xreferences to such).
Caveat emptor.

Attached File  MD manual, specs.pdf   4.28MB   32 downloads

Rgds / Charles
  • 0

Kind Regards,

 

Charles.C


Thanked by 1 Member:

#28 Alan Johnson

Alan Johnson

    Grade - Active

  • IFSQN Associate
  • 15 posts
  • 7 thanks
1
Neutral

  • United Kingdom
    United Kingdom
  • Gender:Male
  • Location:South of England

Posted 09 July 2012 - 07:02 PM

Dear Alan,

Sorry for my misunderstanding.
I certainly agree there is a lack of general understanding regarding MD capabilities. i also agree that processors are probably quite happy to be able to avoid a detailed appreciation of all the twists and turns discussed in this thread. :smile: Unfortunately, the status quo may perhaps be changing somewhat in BRC's case.

I will try and simply illustrate the problem as i see it, below is a machine spec from attachment given. The manual appears to be ca. 2003 so maybe no longer reliable as to specific numbers, (or degree of information content?) –


IMHO, the specs are reasonably (ie probably conservatively) stated from a relative sensitivity POV although I find it strange that no claim is made as to Fe capability. This prevents any absolute use of the preceding info. of course.

Taken at face value, from a presumably reliable source, this spec. gives a direct (theoretical) validation for GMO’s auditor query.

But there are zero actual examples in the 168 page manual (perhaps there was a separate book but i do not see any xreferences to such).
Caveat emptor.

Attached File  MD manual, specs.pdf   4.28MB   32 downloads

Rgds / Charles


Hi Charles, i think the quote/table needs to be treated with a degree of caution. The NFe (Non-ferrous) and Stainless specs are multipliers of the Fe (Ferrous) values so Safeline are essentially saying whatever you get with Fe expect the other test samples to degrade (increase size) by the factors indicated. If you apply the reverse, whatever you get NFe & St/St can effectively predict Fe. However, around the time it was written 300KHz was pretty much the top end of frequency operation for standard conveyor based metal detectors and this level of frequency was predominantly used for 'dry' products i.e. those with little conductive effect. Start to lower the frequency and pass through conductive products (the two tend to go hand-in hand) and these ratios become less 'accurate'. Newer metal detectors will tend to operate at higher frequencies for a given product than older ones and this has an impact on sensitivity especially to poor conducting (non-magnetic) metals like 304/316. In order to operate at higher frequencies metal detectors need to have improved signal processing, screening from the environment and product, and better overall stability, this being where the majority of the design effort has been focussed.

Best regards

Alan
  • 0
Alan@S+S

#29 Charles.C

Charles.C

    Grade - FIFSQN

  • IFSQN Moderator
  • 12,590 posts
  • 3280 thanks
350
Excellent

  • Earth
    Earth
  • Gender:Male
  • Interests:SF
    TV
    Movies

Posted 11 July 2012 - 05:37 AM

Dear Alan,

Thks above.

I subsequently did some googling for food metal detectors in general, 15 pages or so.
The google output ca. 90% consisted of links to manufacturers (good for google anyway).

Provision of integrated, multiple frequency scanning, seemed to be a popular selling point these days.

Did not find a single example containing meaningful product data although a few shots of anonymous bar graphs of LED responses were visible.

I noticed one manufacturer who offered a free preliminary service involving the submission of yr typical finished product after which "guaranteed" sensitivities for their proposed unit(s) would be submitted. Seemed quite a worthwhile offer based on this thread.

There was also one nice, cereal, publication where discussions between MD supplier and processor culminated in, from memory, selection of a minimum 6mm detection size as an extension of the "famous" FDA 7mm recommendation. The perfect customer perhaps.

I did deduce from the various application lists on show that GMO's specific product area may well be problematic, eg "warm bread" was a frequently mentioned example of a difficult matrix.

Sorry if the above may sound a little cynical but it reflects my frustration as to lack of comparative data. I can only assume that, as you initially intimated, the manufacturers simply prefer to not release any raw data. Frankly, I am surprised that no-one has seemingly analysed this lack in print (hint, hint... :oops: )

Rgds / Charles


  • 0

Kind Regards,

 

Charles.C


#30 redwards

redwards

    Grade - Active

  • IFSQN Active
  • 3 posts
  • 0 thanks
0
Neutral

  • United States
    United States

Posted 09 February 2017 - 03:16 PM

We just had our SQF audit.

One of our Non-conformances was on our justification in our metal detector testing.

We have 2 non-ferrous wands

1-Brass

1-Bronze

He claimed we should have a justification of why not aluminum. Or be testing for aluminum.

My justification was non-ferrous is non-ferrous and if you capture one you capture them all.

  The detector is not detecting a specific metal but the magnetic or conductive properties that are associated with non-ferrous metals.

 

I found near the bottom of this chain a group of attachments, the last of which is an excel spreadsheet "Metal Detector2".

4.2C has a graph that might be my answer.

Does anyone know where this is from?

 

Any help would be welcome.

We did pass the audit.

Ray

 


  • 0

#31 Charles.C

Charles.C

    Grade - FIFSQN

  • IFSQN Moderator
  • 12,590 posts
  • 3280 thanks
350
Excellent

  • Earth
    Earth
  • Gender:Male
  • Interests:SF
    TV
    Movies

Posted 09 February 2017 - 05:23 PM

Hi Ray,

 

The graph is from Pg 4.17 of file s2 which is attached in same post (18) as the Excel file.

 

Congratulations on passing the audit !


  • 0

Kind Regards,

 

Charles.C





1 user(s) are reading this topic

1 members, 0 guests, 0 anonymous users