11 replies to this topic

[Anca86]

I’ve recently started working as a supplier quality technician at a frozen raw dog food manufacturer, and I’m encountering issues with one of our raw materials. The product arrives as large frozen blocks (40 × 60 × 15 cm). At intake, we carry out a visual inspection for colour changes or off‑odours, check the temperature, then temper the blocks and run them through a metal detector. This step is not a CCP, as our CCP metal detection is performed on the finished product. The detector sensitivity is set to ferrous 6 mm, non‑ferrous 6.5 mm, and stainless steel 8 mm, and the machine was calibrated specifically for this raw material by metal detector engineers.

 

Because the material is beef or lamb tripe, it naturally contains traces of hay and other environmental debris. Recently, we’ve been experiencing a high number of metal detector rejections. When the quality team defrosts and inspects some of these rejected blocks, they occasionally find small wires, grit, or bolus. However, most rejected blocks are not inspected — they are simply returned to the supplier. This has created tension, as we are rejecting product without physical evidence and relying solely on metal detector alarms. The supplier maintains that their tripe is thoroughly cleaned and metal‑detected three times, and they argue that the high iron content of tripe can lead to false positives.

 

I recognise that we need to implement visual verification on our side and provide evidence with each rejection. The difficulty is that the quality team is pushing back, as defrosting every rejected block is time‑consuming and sometimes involves 30 or more blocks in a single run. If no foreign body is found, the rejection is likely a false positive.

I’m looking for ideas on how to reduce false positives with this type of raw material and how to improve our metal rejection process so we can maintain a good working relationship with the supplier. Any suggestions would be greatly appreciated.

 

[SHQuality]

Because the material is beef or lamb tripe, it naturally contains traces of hay and other environmental debris. Recently, we’ve been experiencing a high number of metal detector rejections. When the quality team defrosts and inspects some of these rejected blocks, they occasionally find small wires, grit, or bolus. However, most rejected blocks are not inspected — they are simply returned to the supplier. This has created tension, as we are rejecting product without physical evidence and relying solely on metal detector alarms. The supplier maintains that their tripe is thoroughly cleaned and metal‑detected three times, and they argue that the high iron content of tripe can lead to false positives.

If you sometimes find small wires, your supplier's metal detector either isn't working as intended, or the wires are too small to be detected with the required sensitivity. Are they using the same limits you are? 

 

Since you're also finding grit and bolus, I question their cleaning methods. Have you (or anyone else from your company) seen their procedures or visited for an in-person audit?

[MDaleDDF]

"When the quality team defrosts and inspects some of these rejected blocks, they occasionally find small wires, grit, or bolus"

I guess I'd be curious how often they find foreign material in defrosted blocks.   "Some" and "Occasionally" doesn't sound like a lot, so it would be interesting to see how many.   Although I know that it can be difficult at times to locate whatever caused a kickout as well.

I agree with SHQ that a meeting and visit may be in order, and this may be an issue that can be rectified together with your supplier with some discussion and overview of the processes of each location, and just tell them you don't want to accuse them of anything, but just work together to get to the bottom of it.   It's a pita for both parties no doubt, so working together to figure out root cause would benefit you both, make safer food, and save money.

[Anca86]

If you sometimes find small wires, your supplier's metal detector either isn't working as intended, or the wires are too small to be detected with the required sensitivity. Are they using the same limits you are? 

 

Since you're also finding grit and bolus, I question their cleaning methods. Have you (or anyone else from your company) seen their procedures or visited for an in-person audit? 

The supplier’s detector sensitivity settings differ from ours; they are currently set to 4 mm for ferrous and 5 mm for both non‑ferrous and stainless steel. I am planning a site visit, subject to the supplier’s agreement, so that we can resolve the issue. It will also be useful to understand how they manage false positives, as we currently have no procedure in place for this. Clarifying their approach should help us reduce the quantity of material being unnecessarily rejected, although this may be a little tricky to address

[Anca86]

"When the quality team defrosts and inspects some of these rejected blocks, they occasionally find small wires, grit, or bolus"

I guess I'd be curious how often they find foreign material in defrosted blocks.   "Some" and "Occasionally" doesn't sound like a lot, so it would be interesting to see how many.   Although I know that it can be difficult at times to locate whatever caused a kickout as well.

I agree with SHQ that a meeting and visit may be in order, and this may be an issue that can be rectified together with your supplier with some discussion and overview of the processes of each location, and just tell them you don't want to accuse them of anything, but just work together to get to the bottom of it.   It's a pita for both parties no doubt, so working together to figure out root cause would benefit you both, make safer food, and save money.

I am new to the company, and I understand that the quality team previously carried out a short trial in which frozen blocks of tripe from two different suppliers were defrosted and assessed. The findings were presented to the suppliers. However, no further tripe has been defrosted since, meaning we are now rejecting blocks without any physical evidence, which the suppliers are understandably unhappy about. One of my objectives is to gain a better understanding of each supplier’s process and to reduce unnecessary tripe rejections. I also need to encourage the quality team to resume defrosting and checking for evidence, as well as recording their findings and any false positives.

[SHQuality]

I would first try to quantify how often "some" and "occasionally" was back then. Are there any people who were around at the time you can ask about that? Did your predecessor leave any documentation?

 

I find the situation strange. Your supplier has a lower limit for their metal detection, yet YOU are the one having a system that triggers even though they are supposedly less sensitive.

 

I agree with MDaleDDF. Focus on working together. They argue that the high iron content of tripe can lead to false positives. How are they dealing with that in their own system? Do they have other systems in place to prevent metal contamination earlier in their system than the metal detector? Is there a chance of post detection contamination in their setup?

[Anca86]

I would first try to quantify how often "some" and "occasionally" was back then. Are there any people who were around at the time you can ask about that? Did your predecessor leave any documentation?

 

I find the situation strange. Your supplier has a lower limit for their metal detection, yet YOU are the one having a system that triggers even though they are supposedly less sensitive.

 

I agree with MDaleDDF. Focus on working together. They argue that the high iron content of tripe can lead to false positives. How are they dealing with that in their own system? Do they have other systems in place to prevent metal contamination earlier in their system than the metal detector? Is there a chance of post detection contamination in their setup?

Unfortunately, all I have are photographs of the foreign body that was found. No other information was recorded, such as how many blocks were defrosted or whether each block was contaminated (referring to the rejected one). There’s nothing that gives me a solid background to identify the root cause.

At this stage, the only option is to start from scratch: review and update the existing metal‑detector rejection procedure, gather proper information, and visit the supplier so we can work together on this.

Thank you all for your input — I’m hopeful we’ll get to the bottom of the issue.

[Scampi]

I'm going to play devil's advocate here

 

As tripe would be quite difficult to process, and you've stated you have metal detection as your CCP in the final product...................is your company making a mountain out of a molehill?

 

Get your MD engineer back in, something may have changed on your unit   and Was it calibrated on thawed or frozen material?

[Anca86]

I'm going to play devil's advocate here

 

As tripe would be quite difficult to process, and you've stated you have metal detection as your CCP in the final product...................is your company making a mountain out of a molehill?

 

Get your MD engineer back in, something may have changed on your unit   and Was it calibrated on thawed or frozen material?

The metal detection step (ccp) is carried out on the finished product. The metal detector positioned at the start of the process-protects the equipment—was  calibrated using frozen material. We use the frozen blocks in their original state throughout processing (cutting, mixing, forming, etc.). The difficulty we now face is that, over the past  months, we have rejected several batches without being able to provide the supplier with any physical evidence of contamination. As a result, the suppliers are increasingly unwilling to accept the nonconformances or to collect the rejected material.

[GMO]

Hmm. I'm doubtful about their "high iron" in the ingredient claim. It's not iron in ferrous form but compounded as haemoglobin.

 

But that aside, this is a problem and for any problem, I'd use a structured root cause analysis technique. But when the problem is between a supplier and customer, finger pointing can start.

 

Firstly I'd like both parties to recognise you're both having problems with this. The supplier thinks they've metal detected the product to a tighter sensitivity than you're using and had no issues. Now they're getting returns. You are getting loads of rejects and defrosting the material takes time which you don't have. But when you've done that in the past you've found contaminants which makes you doubt their processes.

 

I'd be tempted to do a two way visit with a metal detection expert, for example, whoever you use to calibrate your metal detectors (ideally they should be familiar with whatever brand the supplier is using too). Check all the inputs as you would for any fishbone. What is the training like? The process? Is it correctly set up? What is the product signal? Does that change if frozen, partially frozen or chilled etc.

 

I'd also check their process to remove contaminants to see how robust it is.

 

I would do this as quickly as possible because I'd also suggest in the short term you suck up defrosting the rejects until this is sorted. You can't really expect a supplier to accept rejects on the basis of the metal detector rejected it. We all know false rejects do happen (even if I doubt their explanation). Longer term the ideal will be you won't be getting so many rejections so defrosting the blocks will be less of a chore. But perhaps whatever claim you make for a foreign matter finding needs to be commensurate with the cost and hassle of defrosting. Or better still, get your procurement involved. If a reject means wasting that block each time, perhaps there needs to be recosting to account for that?

 

I don't think you can continue though rejecting on the basis of your metal detector. I don't think that's fair on your supplier. It's vague, it could be influenced by poor behaviours by your staff. A foreign matter find is unequivocal. I can't see a way even if it's only occasional in the future that you'd be able to revert to rejecting the whole batch unless you actually find something.

 

Or, get an x-ray. For one, that will give you an image of the foreign material and then it's easier to say "huh, that's a wire" and reject with a copy of the image. 

[Scampi]

The metal detection step (ccp) is carried out on the finished product. The metal detector positioned at the start of the process-protects the equipment—was  calibrated using frozen material. We use the frozen blocks in their original state throughout processing (cutting, mixing, forming, etc.). The difficulty we now face is that, over the past  months, we have rejected several batches without being able to provide the supplier with any physical evidence of contamination. As a result, the suppliers are increasingly unwilling to accept the nonconformances or to collect the rejected material.

 

Understood

you need to have the MD company back in         Something at one end or the other has changed over time

[GMO]

Unfortunately, all I have are photographs of the foreign body that was found. No other information was recorded, such as how many blocks were defrosted or whether each block was contaminated (referring to the rejected one). There’s nothing that gives me a solid background to identify the root cause.

At this stage, the only option is to start from scratch: review and update the existing metal‑detector rejection procedure, gather proper information, and visit the supplier so we can work together on this.

Thank you all for your input — I’m hopeful we’ll get to the bottom of the issue.

 

Just one idea. Either yourself or the MD engineer, see if you can get the data out of your machines on number of items going through and number of rejects. It might be your metal detector retains this over a long time period (many do) and you can see if there was a change or if you have more than one detector if one is rejecting more than another. Or if there's a certain time of year etc.

 

If you can't get this data, it's great to start collecting it now. Then if the supplier makes changes you can start to see if it changes your data.

 

I should have also asked. Is every rejected block put back through the detector 3 times in 3 different orientations? Reason I ask is for a non CCP metal detector, this is an easy way to screen out (possible) false rejects. If it goes back through without rejection 3x then you can be fairly confident it didn't contain metal. It's not 100% foolproof and possibly not recommended for CCP metal detectors but worth doing here. Better still if you can do it on an offline detector so you don't "contaminate" the data collection I've suggested above.