Metal Detector Validation with Test Wands

We are getting ready to put a "new to us" metal detector in service.  We've had a OEM technician in to balance the machine, train us and determine sensitivities.

My question is this:

We run non-packaged chocolate products through the metal detector (chocolates are later wrapped in foil, so this is the last point when we can verify).  I have test wands.  Is it enough to run the test wands alongside - on top - underneath - on each side -- of the products?  I cannot physically put the test wands inside of the product. 

We verify all of our metal detectors at shift start, every 2 hours and at shift end using this method.

Thank you!!

Dear Op,

We manufacture frozen burritos that are passed through a metal detector naked and then packaged in poly bags as bulk (naked no wrap).

Our normal procedure was to place the wands alongside the naked product. During our SQF GAP audit, the auditor did not approve this procedure and said we had to figure a way to place the wand on the actual burrito, our burritos our rock frozen, there is no way. Therefore we decided to use clear sandwich bags to place the wands and the frozen burrito inside every time we test the metal detector.

Hope this helps :)

"Therefore we decided to use clear sandwich bags to place the wands and the frozen burrito inside"

So are you actually putting the wands inside the burrito OR putting the burrito and wand in the same package?  If the latter, how is that different from putting the wands alongside the burrito?

Thank you!

"My question is this:

We run non-packaged chocolate products through the metal detector (chocolates are later wrapped in foil, so this is the last point when we can verify).  I have test wands.  Is it enough to run the test wands alongside - on top - underneath - on each side -- of the products?  I cannot physically put the test wands inside of the product."

I had similar situation with frozen pastry products going through metal detector. We could not insert the test piece into the product. We simply placed the test piece between the two products which was almost at the center of the magnetic field where it is comparatively harder to detect. With this customer and auditor both were satisfied. Fundamental worry of customers and auditors is how is the test piece detected through the product.

Dear ksullivan,

I think there are, at least, 2 routine operational aspects to validate -

(1) the sensitivity

(2) the correct "viewing" of the detector of the product flow and thereby proper detect/reject functioning.

(1) involves choices such as for (a) top/bottom/middle in respect to position of minimum sensitivity (typically at [or near] the detector aperture centre line). As you will see from link below, various height factors can be involved for which you have not yet given any details of actual setup.

(2) involves testing front/back of product within product flow.

Maybe try this post/thread for some "light" theoretical reading, although your product is somewhat different matrix -

http://www.ifsqn.com...ge-2#entry68189

Rgds / Charles.C

are the sizes of the chocolate pieces consistent?  or you run multi-sized products-if these are consistent, here is one way I had done it in a similar product and didn't meet any resistance from auditors

pre-make the chocolate by embed the test  pieces inside- colour code the pieces(mistake-proof) and label them . this ensures if real  pieces are in the middle  then you can catch them .

make new samples  every month and keep the metal test piece calibration records etc handy. If there is chances the choc will melt then this  possibly wont work

To answer your validation question:  validation you generally don't do on a daily basis and here is steps to do  one

 for each test run the test pieces 20 times and they should reject all 20 times.

Pick your worst case scenario- eg SS is difficult to pick up, metal detector # 3 had more issues this year etc , used your smallest box compared to aperture

1) place test pieces on the top of the product

2) place test pieces on the middle of the product

3) place on the  bottom of the product

that's your vertical validation

repeat this on the horizontal side

then do a leading, trailing check

check your buzzer, your andon lights,  lock on your rejection box, cleanliness of the belt, vibration and voila- you are done

if you get a single failure then - this is a fail- adjust and repeat.

To answer your validation question:  validation you generally don't do on a daily basis and here is steps to ........

Although I agree with a lot of your post, this isn't a validation question it is a verification question, see below.

My question is this:

We run non-packaged chocolate products through the metal detector (chocolates are later wrapped in foil, so this is the last point when we can verify).  I have test wands.  Is it enough to run the test wands alongside - on top - underneath - on each side -- of the products?  I cannot physically put the test wands inside of the product.

 

Thank you!!

First of all you need to validate your equipment is capable of detecting as specified and so in commissioning you will need to put the test wands in the product to do this. After all you need to prove that metal can be detected inside the product.

For verification that the metal detector is functioning correctly during production I see test wands placed on top of product where it is impractical to place inside. As you want to run production with a test wand outside the product I would also validate that there is no significant difference in detection/rejection when outside of the product during your commissioning.

Regards,

Tony

You are right Tony

I saw his dilemma as 2-pronged; the heading says validation and the question is verification  that is why I suggested both

We use the same set up , we place the "wands", which are actually cards, under the product.  When I started working here and we conducted a validation, we were placing them on top of the product, and the detector was not set at a sensitivity high enough to  detect the cards placed on the bottom, so we corrected.  The procedure as is hasn't been challenged by an auditor yet  (AIB/ISO)

The rejected bars are then weighed as part of our weight checks, and then destroyed as waste they touch the dirty reject pile and the dirty metal detector wands) .

Older topic but still:

2 important thing needs to be kept in mind: Any metal detector will have it lowest sensitivity in the middle of the aperture. Also the product itself can give a big signal and have big influence on what the detector will be able to detect inside of the product. In the end this what we want to know. Always use the proper test wands that are mentioned on the calibration report. During maintenance or calibration of the units also test with bigger pieces to test against "saturation" error that some brands have.

These are the 2 reason you must always perform any testing with the product and at the center of the aperture. If this is not possible it is very important to test the setup and document why the testing is done differently.

A lot of company's underestimated rejecting as well. In my opinion a good test always involves the complete reject system as most errors show up here.

(front/back etc etc)

Older topic but still:

 

2 important thing needs to be kept in mind: Any metal detector will have it lowest sensitivity in the middle of the aperture. Also the product itself can give a big signal and have big influence on what the detector will be able to detect inside of the product. In the end this what we want to know. Always use the proper test wands that are mentioned on the calibration report. During maintenance or calibration of the units also test with bigger pieces to test against "saturation" error that some brands have.

 

These are the 2 reason you must always perform any testing with the product and at the center of the aperture. If this is not possible it is very important to test the setup and document why the testing is done differently.

 

A lot of company's underestimated rejecting as well. In my opinion a good test always involves the complete reject system as most errors show up here.

(front/back etc etc)

Dear Michiel Bouwmeester,

Yr comments (as per the OP) appear to be a combined mixture of validation and verification.

Regardless, thank you and Welcome to the forum :welcome:

Rgds / Charles.C

:welcome: Michiel Bouwmeester

It looks like you are referring to the Big Ball Metal Detection (and not the AC/DC song contrary to Magenta_Majors) that everyone here was trying to figure out a month or two ago.

During maintenance or calibration of the units also test with bigger pieces to test against "saturation" error that some brands have.

 

 

And you're 100% correct on the placement of the verification rods/strips/spheres. This is also mentioned in another post...http://www.ifsqn.com...ensitive-place/

:welcome:  Michael and thanks for the information. 

As far as I know, there are different designs of metal detector depending on the supplier.  Validation of metal detector must be in coordination with the manufacturer. 

Ok, I have observed heaps of chatter over the years on Metal Detector Validation here and was never really concerned as I haven’t had to deal with them… until now L

TO preemptively keep this discussion defined (as not an extension of the entire Val/Ver debate), my basis is "Codex Alimentarius CAC/GL 69 - 2008":

"Verification: The application of methods, procedures, tests and other evaluations, in addition to monitoring, to determine whether a control measure is or has been operating as intended."              

"Validation: Obtaining evidence that a control measure or combination of control measures, if properly implemented, is capable of controlling the hazard to a specified outcome."        

My concern is with the latter (Validation) applied to Metal Detection.    

My rational come down to this:

1. Given one has traceable certified standards for ones “detectable metal” wands (whatever those size and composition measurements may be).
2. If one places the known “detectable metal” wands through a metal detecting aperture (albeit by any manufacture) with the product pack and it “rejects” it (by whatever method your standard requires): it “is capable of controlling the hazard to a specified outcome”, therefore Validated as working.
   1. Say one places a product with product code setting “cat” with a “detectable metal” wand in Schrodinger’s Box and the “product” is rejected: it “is capable of controlling the hazard to a specified outcome”.

Therefore, say one has 15 “product sensitivity” codes for the metal detector, ones “unit” measurement for Validation is each setting for “each product code change” (15) no matter what type of metal detector manufacturer (because it is validated as capable).

Follow?

Can anybody poke holes in this reasoning?

I have more to say on the “unit measurement” for validating but I leave that for later.

Cheers with any input you may have.

Hi bacon,

No, I don't follow. Sorry.

Yr query looks suspiciously quantum mechanical?

The Schrodinger’s Box was snarky joke but illiterates, basically, every time one passes traceable certified “detectable metal” standards wands and (per adjustment made) and the “product” is rejected: it “is capable of controlling the hazard to a specified outcome”. So one's "daily verification checks" IS a validation.

The Schrodinger’s Box was snarky joke but illiterates, basically, every time one passes traceable certified “detectable metal” standards wands and (per adjustment made) and the “product” is rejected: it “is capable of controlling the hazard to a specified outcome”. So one's "daily verification checks" IS a validation.

Hi Bacon,

Thks for the clarification.

Afai can deduce, The Cat originated while Schodinger was in USA. So I guess NACMCF rules.

You might equally enjoy this one (if you can stay awake) –

http://rdn-consultin...-vs-validation/

Thank you Charles C., I have been down that "Software Validation" path before (at a medical device company with lifecycle management)... and fell asleep (but good link on it; thanks!)

However, any takers of the "daily verification checks are ones validation"?