14 replies to this topic

[Mariann]

Hi

 

If the metal detector alarms for a product, how do you mark the product, and do you place it somewhere specific to wait for further investigation?

[MDaleDDF]

We write "METAL" on the bag and take it immediately to the lab for sifting/investigation.

[kconf]

Yes, write the lot #, date & time, line. Then like MDaleDDF said, take it immediately for further investigation. In my experience, there has been 50-50 chance of finding something. 

[GMO]

It has depended on the site but in some sites the person doing the detector check does the investigation themselves, in others it's QA but it's immediate because it could signal an issue further upstream.

If you don't find anything it's worth considering if it was just too small or multiple tiny pieces and still potentially an issue if you suddenly start getting lots of "false" rejects.  (Bitter experience...)

[TimG]

I'm not sure about BRC but when I ran an SQF program that had a metal detector we were sure to put the procedure in the metal detector document, which referenced the NCP doc. Those products got placed on NCP; we had a laminated NCP placard, and any hits went right to its own NCP pallet to then be re-ran by the shift supervisor. It had to make 2 clean passes by the shift supervisor who would do it immediately after his accuracy checks. If it flagged again, he/she would rip it open and fill out disposition paperwork that I would then review when I was back in that plant.

You want to make sure any product you set aside is clearly separated/sequestered and has no chance or going through as passed product.

[kconf]

So you just separate what was rejected, right? The line that is running the lot keeps running?

[GMO]

So you just separate what was rejected, right? The line that is running the lot keeps running?

 

For one reject, keep running and investigate.  For two rejects, stop the line and assess.  For two rejects of the same type, stop the factory...

Or at least that was the M&S protocol back in the day and you can see the sense in it but it probably is site dependent.  You shouldn't have metal routinely cropping up in your product.

[kconf]

Stop the line and stop the factory even if they are discovered to be false positive later? 

[Tony-C]

Hi Mariann,

 

Ideally you will be rejecting into a locked cage and then QA will collect the product, mark it as rejected with date/time, check the rejected product by running it through the detector again and then opening and inspection/sieving as appropriate.

 

Kind regards,

 

Tony

[GMO]

Stop the line and stop the factory even if they are discovered to be false positive later? 

 

No, only if metal is found which is one of the reasons why we always did it immediately but also it's the right food safety thing to do.  If you have two rejects that look similar, e.g. pieces of metal swarf, you need to find out what is going on to cause it as you're probably missing some in your metal detector (they're not 100% effective.)

 

But that was M&S protocol, not saying you have to adhere to that but it's a good indicator of what some people think is the right way.

 

If you're having loads of false rejects, what I'd suggest you do is:

 

• Check if you've recently had any settings changed.
• See if the false rejects have increased suddenly.
• If they have, do a thorough check of your line in key areas where metal on metal contact is possible, grinding and mixing (are there any points of grinding or fresh scratches?)  Also ask key employees (e.g. ops, hygiene and maintenance) for ideas on where it might be from and what has recently changed.
• If the line is clear check ingredients.  it might even be worth metal detecting some.
• If both are clear, look at reducing your sensitivity (but only as a last resort).  Loads of false positives is a sign that you might be trying to detect at a level which is too close to your product signal.

[Mariann]

Thank you all for response. In our plant when metal detector alarms the operators investigate the product. Usually they investigate the product immediately. They write non conformance form and attached there also the metal findings and bring it to the quality. 

                                                                                                           

We had recently audit and the auditor mentioned that the product should be marked somehow, or it should go to some locked box or specific area etc. I am interested to know what kind of solutions other have. 

[GMO]

Thank you all for response. In our plant when metal detector alarms the operators investigate the product. Usually they investigate the product immediately. They write non conformance form and attached there also the metal findings and bring it to the quality. 

                                                                                                           

We had recently audit and the auditor mentioned that the product should be marked somehow, or it should go to some locked box or specific area etc. I am interested to know what kind of solutions other have. 

 

It would be interesting (or would have been) to interrogate the auditor's concerns.  The pack will immediately be rejected into a locked bin so if the operator doesn't for some reason get to it immediately then there's no risk, albeit that shouldn't be left there long so it can be investigated.  If it's removed and left for some reason, why?  Why not investigate straight away?  Lastly if you're putting that pack through the same metal detector, not an off line one, then their concerns are valid.  You want to make sure it's not going to go back into normal production.  Some retailer standards in the UK allow a pack to be put back in if it's been re metal detected 3 times in 3 different orientations and not set off the detector but I'd never advocate this as good practice personally.

 

So I guess if you do have that concern, just get some labels near your metal detector, e.g. labels marked with "metal reject" in red and stick them on anything taken out of the bin.  That's one solution if the above concerns are what they had?

[SHQuality]

Yes, write the lot #, date & time, line. Then like MDaleDDF said, take it immediately for further investigation. In my experience, there has been 50-50 chance of finding something. 

 

Hi Mariann,

 

Ideally you will be rejecting into a locked cage and then QA will collect the product, mark it as rejected with date/time, check the rejected product by running it through the detector again and then opening and inspection/sieving as appropriate.

 

Kind regards,

 

Tony

Surely the lot and time codes are already on the package?

[Tony-C]

Surely the lot and time codes are already on the package?

 

Possibly, but this is a pack of strawberries, it should have a lot code that corresponds to the date, but not always, and it is easier to record the date at the time.

 

The time packed isn't always recorded and that might not even be the time it went through the metal detector.

 

Kind regards,

 

Tony

[jfrey123]

Hi

 

If the metal detector alarms for a product, how do you mark the product, and do you place it somewhere specific to wait for further investigation?

 

Anything flagged by MD needs to be isolated, period.  If it's a case detector, my plants have alarms/lights and the operator has to isolate the case onto a "reject" labeled pallet or area for evaluation.  If it's an inline MD of unpackaged product (belt or gravity feed) you'll have a diverting system that rejects the steam into an isolation area that is also labeled.

 

Everywhere I've worked had policies about continuing processing when MD is triggered depending on frequency: they can continue at maybe 1-3 rejects in say maybe an hour, with the caveat that all MD rejections are evaluated prior to positive product release for shipping.  But even with that, when the MD start going bananas, it should trigger shutdown and investigation mid-processing.  All scenarios of how you will react to MD rejections should be covered in the SOP.