7 replies to this topic

[cthomp]

Hello! 

 

I'm at a facility where a MD reject stops the line until the source is found.

 

I recently visited a company that uses xray. They put rejects aside and revaluate them at the end of the run. Xray false rejects are way more common than MD false rejects, so I guess that makes sense.

 

Does it make sense to do the same with MD rejects? just put them aside till the end of the run? I know, you're exposing yourself to potentially unpacking pallets.But the metal hits we get are almost always from secondary packaging, and we never find "metal". There's just enough metallic dust somewhere in the cardboard, or they're false rejects. Pretty much the same scenario with the xray.

 

What are best practice procedures for responding to a MD hit? I'm haven't found any sort of standard guidance.

I know the answer is it depends on your process and risk.  

 

I'd really apprecriate hearing the community's procedures for MD rejects.

Thanks in advance for any insight!

 

[SQFconsultant]

I don't know I may be too anal about risk, or so I've been told - but I've always found putting stuff aside until later meant that somehow it will get overlooked - as in best to it now.

[Tony-C]

Hi cthomp,

 

 

Welcome to the IFSQN forums.

 

From an efficiency perspective stopping the line is not ideal unless this is a rare occurrence.

 

The best system I have seen that enables an efficient line and rejects to be investigated is having rejected product diverted to a locked cage then having QA check/investigate the rejected product (this could be done say as part of hourly line checks).

 

The line would only be stopped if an unusual amount of rejections were encountered or the hourly line check discovered a rejection of concern.

 

The QA check would involve checking rejected product through the metal detector again and examining the products to identify the cause of the initial rejection. Any causes should be investigated and can used as the basis for analysing trends. False rejects should also be logged as they may indicate an error or that the settings are too sensitive for a particular product and require adjustment.

 

Kind regards,

 

Tony

[jfrey123]

At the spice plant I started QA in, we would evaluate MD rejections but the critical limit for our CCP was 10 rejections per hour.  If that was reached, we had to document an investigation to determine if there was an inherent problem within the product stream or the equipment feeding the packaging line.  We always evaluated all rejections as part of our release program (open stream MD before packaging), but having a limit within a time period gave our CCP teeth.

[MOHAMMED ZAMEERUDDIN]

You can recheck the material if no rejection is observed then release the lot.

[MDaleDDF]

Our rejects are kicked into a container, and after the lot is run, they're run through again.   A kickout must pass thru the MD 3 times clean to be used again.   If it kicks out again even once, it's sent to the lab for sifting to try and find the metal.   This includes tearing up the packaging and sending it thru the MD if needed to zero in on what's setting off the MD, even though yes, at times you can't find the metal if it's in the packaging, which has happened to me a few times.  

But basically after it's for sure kicking out, the lab takes over searching for the metal hit so the production line can continue on to the next lot of product.  

We don't have a hard and fast number for what is too many kickouts, but in my experience it's obvious when the machine is doing something wonky and kicking out every bag, etc.   I'm lucky in that my MD crew in production have been dealing with the machine a long time and know it well.   Our MD training guy is really good as well so our annual testing on the machine keeps us tight and runnin' right...

 

We have seen weird issues over the years.   Last year we had bags kicking out like crazy for a few days, found no metal after sifting through hundreds of pounds of mix.   It ended up being too much glue on the bag labels.   That one drove us nuts till we figured it out, lol. 

[G M]

 

Hello! 

 

I'm at a facility where a MD reject stops the line until the source is found.

 

I recently visited a company that uses xray. They put rejects aside and revaluate them at the end of the run. Xray false rejects are way more common than MD false rejects, so I guess that makes sense.

 

Does it make sense to do the same with MD rejects? just put them aside till the end of the run?...

 

Getting more/less false rejects from MD / X-ray is going to depend on the physical characteristics of your product.  Our product chemistry means we see far fewer false rejects from x-ray inspection equipment.

 

Only the FM equipment doing CP duty that precedes another processing step like dicing are programmed to stop the line, to prevent generating more problems -- the packaging line CCP FM instruments reject sealed packages into a bin for review.  The FM rejects for packaging are reviewed periodically throughout the shift, at roughly the same frequency that verification checks are performed on the inspection equipment (1-3h of operation).

[Planck]

 

Hello! 

 

I'm at a facility where a MD reject stops the line until the source is found.

 

I recently visited a company that uses xray. They put rejects aside and revaluate them at the end of the run. Xray false rejects are way more common than MD false rejects, so I guess that makes sense.

 

Does it make sense to do the same with MD rejects? just put them aside till the end of the run? I know, you're exposing yourself to potentially unpacking pallets.But the metal hits we get are almost always from secondary packaging, and we never find "metal". There's just enough metallic dust somewhere in the cardboard, or they're false rejects. Pretty much the same scenario with the xray.

 

What are best practice procedures for responding to a MD hit? I'm haven't found any sort of standard guidance.

I know the answer is it depends on your process and risk.  

 

I'd really apprecriate hearing the community's procedures for MD rejects.

Thanks in advance for any insight!

 

Hello,

 

I am a technical support person directly responsible for commissioning MD sites and I have been working in this field for over 10 years.

In most cases, this status can be directly identified as the metal detector not being set up well.

 

In the majority of areas where a metal detector can be used, it can work extremely consistently if the MD is set up correctly.

But a bad reality is that it is very challenging to try to get a metal detector to be very stable.

 

For a metal detector to achieve a basically satisfactory state, it should have a false detection rate and a missed detection rate of less than 1 in 10,000 respectively.

 

There are so many things that need to be checked on the metal detector site, the general order of checking is:

1, The inductive signal of a metal detector at the rest state must be low enough!

If the MD's induction signal exceeds the machine's normal signal fluctuations, the cause must be found.

You can't proceed to the next step until you find it.

 

2, The inductive signal must also be low enough for idle operation. The cause must be found and then the next step.

 

3, When passing through the product, you must try to minimize the product's own inductive signal.

At the same time, view the inductive signals of the 3 standard test pieces.

Must try to get them to pull as wide a gap as possible, the bigger the better.

Generally speaking, it is best if the inductive signal of the test piece can reach more than 2 times of the product signal.

 

But that's just in general, it's not a must. This needs to be evaluated on a case-by-case basis. This is actually another separate complex issue that you can ignore for now.

 

If you can let me know your product, packaging, weight, temperature and your test pieces etc, and if using a LOMA metal detector, I should be able to advise a little more.

 

Earlier I mentioned that it would be best to go to 1 in 10,000, and in fact our highest record goes to 1 in a million.

In this case, MD will be extremely reliable.

In a typical case, one of our customers complained to us that his metal detector had 9 false alarms in 3 consecutive days.

He asked me to arrive at his facility for further checks as this was unacceptable.

 

Any questions about metal detectors are welcome!

 

Thanks.

Edited by Planck, 05 April 2024 - 07:12 PM.