8 replies to this topic

[The Food Scientist]

hey there everyone! 

 

I just joined a new company recently and am being tasked to improve their MD program. I have never really worked with MD before, especially at the level of role I am in now. 

 

One of the things they previously done by my predecessor was a 30-pass, so they would pass each wand 30 times to "verify" its properly working. I have no clue where this came from, I havent found any basis for it nor a document justifying it. I called the manufacturer and said its odd to do 30 pass but its not uncommon, and that every company can decide what to do. 

 

what is everyone's experience/procedure?

[AltonBrownFanClub]

30-pass sounded familiar to me from my past metal detector work.

I found this reference on a UK site about CCPs.

https://techni-k.co....etal-detection/

 

This reference also says it is up to the company to set the limits.

Sorry I can't help more.

 

P.S. I love the quote in your bio. One of my favorites from A.B.

[G M]

30 passes is more akin to what I'd expect from validation.  Most verifications will be 5 or less.

 

Depending on the extent of the maintenance, I suppose re-validation would be reasonable.

[jfrey123]

Agree with GM.  In almost all sites I've been around, 3 passes each with an fe, non-fe, and stainless test piece was done for the hourly verifications to challenge the MD.  Something akin to 30 passes sounds like a validation you might do with either new products or new settings.

[Brothbro]

I have similar experience to GM/jfrey, MD verifications were a 5-pass system. I'm a little surprised the manufacturer didn't give clearer guidance, with some more questioning they should be able to give a better indication on how to properly verify your unit if you give them info on your product. 

[Planck]

This verification is essentially used to check the reliability of the sensitivity of the metal detector (there is actually an equally important rejection reliability at the same time, which will not be discussed here).

If you deal with metal detectors on a regular basis and follow this CCP very closely, you should be able to notice a phenomenon:

A certain validation on a certain day, and your test stick passes straight through without being rejected.

( Only cases that are not detected by MD, rather than detected but not rejected, are discussed here. )

When you feel confused and revalidate at this moment, this phenomenon no longer occurs! And it may not happen for a long time thereafter.

 

This is actually detecting reliability, so we need verification.

 

Therefore, different companies may take different programmes.

 

For example, the 30 times you have here (which I believe should be after a new product or MD has been adjusted), as well as the hourly verification that almost all companies have.

 

As of now there doesn't seem to be a more advanced technology to confirm the reliability of MD, so almost all users can only check it by repeated testing, including relevant industry standards with similar requirements.

 

We've been working in the extremely challenging field of MD service support, so we noticed this situation many years ago, and after further research and testing with our customers, we actually have the technology, which we currently call DPA ( Digital Performance Analysis technology ).

 

Here I'll upload another simple brochure with a simple example there.
 

Simply put, if the risk can be controlled by technical analysis in advance, it is almost impossible to encounter the risk of a missed test piece again.

This means in practice extremely reliable detection reliability and sensitivity.

The highest record we've tracked is 260,000 consecutive reliable detections without a missed detection.

 

We look forward to seeing more interesting professional issues and relish the challenge!

Attached Files

•  DSN802-00045 Devo DPA EN v1.4.pdf   165.49KB   15 downloads

Edited by Planck, 24 January 2025 - 02:19 AM.

[Planck]

This verification is essentially used to check the reliability of the sensitivity of the metal detector (there is actually an equally important rejection reliability at the same time, which will not be discussed here).

If you deal with metal detectors on a regular basis and follow this CCP very closely, you should be able to notice a phenomenon:

A certain validation on a certain day, and your test stick passes straight through without being rejected.

( Only cases that are not detected by MD, rather than detected but not rejected, are discussed here. )

When you feel confused and revalidate at this moment, this phenomenon no longer occurs! And it may not happen for a long time thereafter.

 

This is actually detecting reliability, so we need verification.

 

Therefore, different companies may take different programmes.

 

For example, the 30 times you have here (which I believe should be after a new product or MD has been adjusted), as well as the hourly verification that almost all companies have.

 

As of now there doesn't seem to be a more advanced technology to confirm the reliability of MD, so almost all users can only check it by repeated testing, including relevant industry standards with similar requirements.

 

We've been working in the extremely challenging field of MD service support, so we noticed this situation many years ago, and after further research and testing with our customers, we actually have the technology, which we currently call DPA ( Digital Performance Analysis technology ).

 

Here I'll upload another simple brochure with a simple example there.
 

Simply put, if the risk can be controlled by technical analysis in advance, it is almost impossible to encounter the risk of a missed test piece again.

This means in practice extremely reliable detection reliability and sensitivity.

The highest record we've tracked is 260,000 consecutive reliable detections without a missed detection.

 

We look forward to seeing more interesting professional issues and relish the challenge!

In fact, to ensure that the CCP is as reliable as possible, the POD, FRR and CL need to be calculated, which is fundamentally different from the annual metal detector calibration.
This should be done after the MD manufacturer's calibration service.

The annual calibration service for metal detectors involves proprietary technology from different manufacturers and therefore only the manufacturer or a licensed company has this technical capability.

The POD, FRR and CL are completely independent quality controls, and the control standards implemented can vary significantly from one user to another, all of which are beyond the direct control of the manufacturer.

 

Please let me know at any time if there are any further questions!

[siskos]

Hi

Metal detector verification is a tricky business … You must take into account several parameters. Such as dimensions, weight, temperature etc. I have also noticed that in my procedure air moisture pays a role in detection sensitivity.

The same product but different pieces per bag changes the results. Also, the placement of the sample pays an important role. You must understand physics and the way the MD works. Passing the same product 30 or 100 times does not help you. My suggestion is to pass several similar products using different angles and distances between them and check if the machine works properly. In other words, to simulate the process as much as you can.

[GMO]

It's been a while since I've been in the detail on metal detection but agree with others 3-4 sticks is normal.  Retailers in the UK seem to be the biggest sticklers on this about how you then pass it through.  Often (from memory) they request a "large bolt" test which was a 20mm Ferrous stick to replicate the situation where a large metal contaminant could pass.  I believe there were incidents of this causing the MD to "blind" in some way or go into fault.  That was always first.

 

Normally there would then be the three other sticks passed in normal pack spacing (assuming belt metal detector).  I'm too old and it's been too long for me to remember the order but it would be

 

Large Fe - Fe -Non Fe -SS.

 

Then, again apologies, from memory, they also insisted I believe on another check with previously metal detected packs in-between the test packs (apologies if this is the wrong order, I do think Morrisons particularly were really specific about which stick should be first), eg:

 

Fe - good pack - non Fe - good pack - SS 

 

All at normal pack spacing.  This was to check that only the metal containing packs are rejected.  The rationale being that if you have "good" rejects then you may have poor behaviours on the lines disbelieving a reject is real.

 

There was also another check, on the failsafe where a pack was shown to the detector and then removed before reject.  That should stop the line and alarm.

 

Apologies, the manufacturers I've been working in for the last 8 years haven't used belt systems due to the product so if the above is misremembered or out of date please correct me!  But either way, 30 sounds bonkers.