7 replies to this topic

[granola]

Hi all. We're in the beginning stages of setting up inspections on our packaging. We currently only have pouches of 2,8,and 12 ounces for our granola. We need a written procedure and I've read up a bit on AQL sampling but I am unsure on how much to sample. I noticed that General inspection II seems to be the most common. It seems like a lot to ask for of my small crew to inspect 500 pouches on our bigger orders of 60,000 pouches. Do I just have them do the 500 or is it acceptable to go to an S4 inspection. I can't find anything that really states what inspection level I should use. We've received orders in before and our current inspection process is a little lacking and I want to make it more robust but 500 seems like a lot for us right now. Product information if helpful is granola that goes in product contact pouches. I imagine we're going to use the typical 1.0 and 4.0 sections but I'm not sure what sampling plan is too little or too much. Any ideas? Thank you!

[Scampi]

This calculator should help

https://www.intouch-.../aql-calculator

 

got you down to 200 pouches

 

 

Because I'm curious............why add this level of inspection?  What are you looking for and have you had issues?

 

There may be another way to address this

Edited by Scampi, 01 March 2021 - 07:29 PM.

[granola]

This calculator should help

https://www.intouch-.../aql-calculator

 

got you down to 200 pouches

 

 

Because I'm curious............why add this level of inspection?  What are you looking for and have you had issues?

 

There may be another way to address this

We haven't had any major issues, we need to have an incoming inspection process in place and as of right now we just open a few boxes and look at a couple pouches. I am looking to put into place something a little more official and am unsure how to go about getting the actual amount we should be sampling.

 

As for what I'm looking for is mostly cosmetics(colors/print quality), product information, and defects or damaged pouches.

Edited by granola, 01 March 2021 - 08:19 PM.

[Scampi]

Ah, ok that simplifies things considerably

 

 

Let say each shipment is 9 pallets. 60cs/pallets-assuming ALL from the same lot/production run

 

You could choose to sample the square root of 9-3 cartons from 3 different pallets, 3 different locations on the pallet and then ALL the pouches within those 3 cartons

 

That would still give you an AQL sampling plan, but one that is more manageable

 

 

If you notice that defects are larger than 1% overall (or whatever you deem acceptable), then you could expand and sample twice as much following the same system

 

This should allow you to still get a sample that is representative of the whole without spending tons of time sampling

 

Also, make a binder with pass/fail photos  that will speed up inspections

 

Don't forget about weights........just sayin

[granola]

Thanks for the guidance! Since our cases are anywhere from 700-1500 pouches per case I'm thinking of square root of cases rounded up x 20.

 

e.g 60 cases of 1000 pouches would equal 160 inspected( 60 squared rounded up is 8 and 20 per case). If the failure rate is 2% or higher of our initial sample we'd jump up to a regular AQL sample size. 

 

This seems decent enough without going to overboard. If we were to see more defects often then maybe up our sampling plan.

 

Thanks for the help!

[Charles.C]

Amazing !

 

I always thought that the idea of AQL is to first decide what AQL is required, ie what is the specification ?.

 

I don't think sampling based on square roots is equivalent to AQL ?.

[Scampi]

No it's not equivalent Charles

 

But given this posters requirements I think it should be sufficient

 

using even the lowest AQL sampling plan in this scenario = finding another supplier

 

I've only EVER used true AQL in sampling Cannabis post harvest, pre grind.  And those samples get analyzed for 114 pesticides, as well as heavy metals and then purity and potency, so alot more riding on that sample, than inspecting packages for defects

[Charles.C]

No it's not equivalent Charles

 

But given this posters requirements I think it should be sufficient

 

using even the lowest AQL sampling plan in this scenario = finding another supplier

 

I've only EVER used true AQL in sampling Cannabis post harvest, pre grind.  And those samples get analyzed for 114 pesticides, as well as heavy metals and then purity and potency, so alot more riding on that sample, than inspecting packages for defects

 

Hi Scampi,

 

^^^ You're probably right.

 

I have previously used Plans based on S4 levels since they do offer a solution to avoid prolific sample sizes. At a cost of course but then other small size Plans have similar limitations.

It always depends on what level of accuracy you are happy with (or sometimes prefer not to know).