Average Weight Control
Hello,
We do run a small bakery trying to get the BRC certification, hopefully this summer. We bake bread --> slice it --> and pack to a closed bag.
We do declare the average weight on nearly all of our products and we were doing the manual weight check by sampling 10 breads hour or per run (if run was shorter than 1hr). we then check the average per every 5 breads and do the rolling average for the next checks.
Following the bakers standard and EU weight legislation I have decided the weight checks must be increased to minimum 20 per run or 10 per 15 min (if run is longer). However, my staff is now struggling with manual weight calculations, because there is more to calculate than to pack. We have assessed the weeks worth documentation and we know that there is nearly no product which falls under declared weight. they happen occasionally once or twice within 24hr or not at all. I wonder then if I have to do the average check if all my checked weights are above the declared weight? the only thing i have found was:
14. The Directive sets out three rules with which packers must comply: - the actual contents of the packages should not be less, on average, than the nominal quantity; - the proportion of packages which are short of the stated quantity by a defined amount (the “tolerable negative error” or TNE) should be less than a specified level; and - no package should be short by more than twice the TNE.
We do have T1 and T2 established as well, but we never reach these values.
If i only check the avearge of 10 products, where one product fell under the declared weight, would i be ok?
Hi Leatcookies,
The short answer is yes, because the overall average is what the regulatory body would take into consideration. If 9/10 make weight then the 1/10 is clearly an outlier and if say all weighed 10 kg (and 10 kg is your control limit) but the one was 9.9 kg you would be compliant. Of course make sure to take action when your set average checks consistently fail, which sounds rare.
Cheers,
Dan
Hello,
We do run a small bakery trying to get the BRC certification, hopefully this summer. We bake bread --> slice it --> and pack to a closed bag.
We do declare the average weight on nearly all of our products and we were doing the manual weight check by sampling 10 breads hour or per run (if run was shorter than 1hr). we then check the average per every 5 breads and do the rolling average for the next checks.
Following the bakers standard and EU weight legislation I have decided the weight checks must be increased to minimum 20 per run or 10 per 15 min (if run is longer). However, my staff is now struggling with manual weight calculations, because there is more to calculate than to pack. We have assessed the weeks worth documentation and we know that there is nearly no product which falls under declared weight. they happen occasionally once or twice within 24hr or not at all. I wonder then if I have to do the average check if all my checked weights are above the declared weight? the only thing i have found was:
14. The Directive sets out three rules with which packers must comply: - the actual contents of the packages should not be less, on average, than the nominal quantity; - the proportion of packages which are short of the stated quantity by a defined amount (the “tolerable negative error” or TNE) should be less than a specified level; and - no package should be short by more than twice the TNE.
We do have T1 and T2 established as well, but we never reach these values.
If i only check the avearge of 10 products, where one product fell under the declared weight, would i be ok?
Short answer - I anticipate the Directive includes sampling requirements ?. Logically analogous to those for the EC's "weight control" Logo.
Hi leatcookies,
No idea if this article matches yr specific situation but maybe of some assistance -
We do have T1 and T2 established as well, but we never reach these values.
If i only check the avearge of 10 products, where one product fell under the declared weight, would i be ok?
Two main elements to consider with this:
1) How much "under" would make a difference to whether this is or isn't acceptable, as even if the average remains above the nominal weight, if the one errant result is too low then you'll potentially fall foul of the T2 limit.
2) There is a potential risk with fewer samples that one anomalous figure skews things in a direction that does not work in your favour. One T1 result out of ten samples gives you 10% T1s, which exceeds the 2.5% limit...
If it's the calculating rather than the physical act of weighing that is taking too much time, can you not make a simple Excel spreadsheet to do this for them, so they just need to enter the weight of each checked item and it does the rest?
Thank you for all comments. @pHfruit, yes, i did not consider the that 1 oos sample would give me 10% of t1s. I will incorporate this into the sampling procedure and if my runs are shorter (sometimes we pack 70 items of one product), we will have to check more if one falls under T1. . Looks like 40 products checked per hour permit 1 being T1. to stay compliant i must have max 2.5% T1 for the whole run of one product, which i am sure, I can provide.
I am also planning to take the average if at least one sample falls under declared weight, so I do not exclude averages totally. This exercise would be the assurance, that we provide stock above the average.
We cannot provide them with excel, we dont have a spare computer to put there (neither the space :oops2: ) it was already discussed :ejut:
yes, because the overall average is what the regulatory body would take into consideration.
Thank you for all comments. @pHfruit, yes, i did not consider the that 1 oos sample would give me 10% of t1s. I will incorporate this into the sampling procedure and if my runs are shorter (sometimes we pack 70 items of one product), we will have to check more if one falls under T1. . Looks like 40 products checked per hour permit 1 being T1. to stay compliant i must have max 2.5% T1 for the whole run of one product, which i am sure, I can provide.
I am also planning to take the average if at least one sample falls under declared weight, so I do not exclude averages totally. This exercise would be the assurance, that we provide stock above the average.
We cannot provide them with excel, we dont have a spare computer to put there (neither the space :oops2: ) it was already discussed :ejut:
Ah yes, IT/space/CapEx challenges can scupper many an otherwise good idea :ejut:
You may feel comfortable reducing the sampling frequency once you have plenty of data that demonstrates your process is routinely above the declared average - just expect it to be something that Trading Standards ask about as it's an easy one to get caught out on!
yes, because the overall average is what the regulatory body would take into consideration.
With the greatest respect, this is most definitely not the case for the EU/UK position.
I do not know what the equivalent regulatory requirement is in the US, but here it is intended to consider both the average and the distribution, or at least the part thereof that sits below the nominal average value.
Consider an example where one packs 1000 units of a product with a nominal weight of 1kg. If 999 of them have a mass of 1.01kg and one has a mass of 0.005kg, the arithmetic mean of the batch is 1.009kg, and thus it exceeds the nominal figure. Is the unlucky consumer that buys the pack labelled as 1kg but only actually contains 5g likely to be happy about this? ;)
I would set up a specific weight parameter for each product. Each loaf, boule, etc. would only be allowed to be under weight by a certain percentage. Anything under the minimum weight would be rejected, and a larger sample size taken if multiple samples are under weight. If the product is underweight, but still within the allowable limit, then a specific number based on the sample size for the average weight.
I would create a chart of calculations for each specific product that has the math already laid out for them.
Have each product with min and max weight parameters and then a number of allowable fails based on the sample size.
Example:
So when the staff is taking samples, they can look at the chart, see that if the loaves should weight 1 kg and you take 20 samples, only 2 of them can be below 1kg, but still must be within 980g to pass the lot. (this was just easy math, you'll have to set up your own parameters based on regulations and a risk analysis for sample size)
Ah yes, IT/space/CapEx challenges can scupper many an otherwise good idea :ejut:
You may feel comfortable reducing the sampling frequency once you have plenty of data that demonstrates your process is routinely above the declared average - just expect it to be something that Trading Standards ask about as it's an easy one to get caught out on!
yes, i am afraid that the standards require to do minimum 40, so probably will not be able to reduce. but thanks a lot! helped a lot today!
I would set up a specific weight parameter for each product. Each loaf, boule, etc. would only be allowed to be under weight by a certain percentage. Anything under the minimum weight would be rejected, and a larger sample size taken if multiple samples are under weight. If the product is underweight, but still within the allowable limit, then a specific number based on the sample size for the average weight.
I would create a chart of calculations for each specific product that has the math already laid out for them.
Have each product with min and max weight parameters and then a number of allowable fails based on the sample size.
Example:
So when the staff is taking samples, they can look at the chart, see that if the loaves should weight 1 kg and you take 20 samples, only 2 of them can be below 1kg, but still must be within 980g to pass the lot. (this was just easy math, you'll have to set up your own parameters based on regulations and a risk analysis for sample size)
Production has T1 and T2 values provided, however, we dont have products under declared weight. therefore the question, do i have to calculate the average?
E.g. Declared weight of the loaf is 800g, and I have the values between 840-866. They check 10 products per hour currently, but i want to increase it to 40ph, therefore the calculation of average takes them around 10min, which is a bit from their valued hours :D
Production has T1 and T2 values provided, however, we dont have products under declared weight. therefore the question, do i have to calculate the average?
E.g. Declared weight of the loaf is 800g, and I have the values between 840-866. They check 10 products per hour currently, but i want to increase it to 40ph, therefore the calculation of average takes them around 10min, which is a bit from their valued hours :D
I personally do not calculate an average weight. I only have min and max parameters set up. We do have automatic check-weighers that will kick out out-of-spec product. If all of your samples fall within weight limits, then I feel like calculating an average would only give you an insight as to how much product you are giving away.
I have never had an auditor ask for a calculated weight average, only that if your weights are out-of-spec that there is a documented corrective action.
I work in the meat industry, so regulations may differ for bakery.
I personally do not calculate an average weight. I only have min and max parameters set up. We do have automatic check-weighers that will kick out out-of-spec product. If all of your samples fall within weight limits, then I feel like calculating an average would only give you an insight as to how much product you are giving away.
I have never had an auditor ask for a calculated weight average, only that if your weights are out-of-spec that there is a documented corrective action.
I work in the meat industry, so regulations may differ for bakery.
As noted above, there seem to be some difference with the regulatory requirements between the UK/EU and US on this. Average weight is defined as a specific legal requirement here, so not calculating/knowing/controlling it is not an option ;)
The link that Charles provided gives a good brief overview of the expectations for those who are curious.
Hi everyone,
I have 2 bakeries in my clients. None of them do any single bread weighing.
Why ?
The process is "industrialized". They make boxes of x kg for the first raise. Then they put the content of that box into a mecanical divider, which cuts the loaf in y parts. Each part (or sometimes 2 or 3 together) are then shaped to make single loafs. And they know, because they have validated the process, that such loafs will make 400g, 800g or any required weight.
And the weight, for product able to lose weight (like drying bread) must be ensure at the end of the shelflife.
So, what about the directive telling about 80 samples ? Well, this explains how authorities must control the weight. FBO must be sure they respect the law, but they are not obliged to do the 80 samples.
In any case, do not forget to keep records of your validation.
Hello,
We do run a small bakery trying to get the BRC certification, hopefully this summer. We bake bread --> slice it --> and pack to a closed bag.
We do declare the average weight on nearly all of our products and we were doing the manual weight check by sampling 10 breads hour or per run (if run was shorter than 1hr). we then check the average per every 5 breads and do the rolling average for the next checks.
Following the bakers standard and EU weight legislation I have decided the weight checks must be increased to minimum 20 per run or 10 per 15 min (if run is longer). However, my staff is now struggling with manual weight calculations, because there is more to calculate than to pack. We have assessed the weeks worth documentation and we know that there is nearly no product which falls under declared weight. they happen occasionally once or twice within 24hr or not at all. I wonder then if I have to do the average check if all my checked weights are above the declared weight? the only thing i have found was:
14. The Directive sets out three rules with which packers must comply: - the actual contents of the packages should not be less, on average, than the nominal quantity; - the proportion of packages which are short of the stated quantity by a defined amount (the “tolerable negative error” or TNE) should be less than a specified level; and - no package should be short by more than twice the TNE.
We do have T1 and T2 established as well, but we never reach these values.
If i only check the average of 10 products, where one product fell under the declared weight, would i be ok?
Hi leatcookies,
It rather depends on what you mean by "OK?"
(Definitive information on this topic requires a copy of the "Packers' Code (1979 !)" which is an elusive object albeit still commercially available (a summary of its contents are in file pk2 below). Other available publications, eg Welmec series contain, afaik, summaries of aspects in the Packers' Code document.).
IMEX of exports to UK/EEC some Companies simply avoid the control complexities discussed below by substantial overfill and accepting the financial hit.
I assume the OP comments relate to this document -
pk1 - UK weight control - Guidance_-_The_Weights_and_Measures__Packaged_Goods__Regulations_2006_v.4_December_2015.pdf 644.22KB 55 downloads
You did not mention if the EC "e" logo is directly involved so I assumed not. This has its own, relatively well-defined, expectations.
I also assumed a checkweigher is not currently involved.(there is considerable discussion on using such option in various files attached below).
I noted/ignored some specific comments in pk1 related to bakers/bread packing.
I assume by "OK" you mean that your system will be accepted by a W&M Inspector as equally effective when compared to the "reference test" as discussed in pk1. (BRC acceptability is presumably a separate issue which I have ignored for present purposes). For example quoting pk1 -
16. The packers and importers checks need to be ‘sufficiently rigorous’ to ensure that the packages are packed to meet the three packers’ rules. This implies that the system should be as efficient as a reference test to control the average quantity and the proportion of packages with a deficiency of more than the TNE.
(Rule 16 appears to be interpreted as, inter alia, requiring compliance with Rules 17 - 19), ie -
17. In the design of the system, attention is recommended to the following issues:
a. for processes controlled by sampling, that the sampling plan is specified and that the target quantity and control limits are appropriately set,
b. for packing lines controlled by checkweighers, that the set points are appropriate,
c. that checking or measuring equipment is properly maintained and checked,
d. that the system should be formalised and reviewed regularly for its appropriateness, and include corrective actions and records that must be made,
e. that the system should specify the training for staff to ensure that the system is adequately implemented
18. Further advice on designing and setting up quantity control systems is contained in Appendices C and D of the Packers’ Code.
19. All products are required to meet the three packers’ rules at the time of packing. For desiccating products, which may lose weight after packing, the onus is on the packer to prove that any subsequent failure was due entirely to the desiccation of the product contained in the packages after they were made up.
A quite nice detailed example (pg 23 et seq) is given in pk1 to illustrate how you can ensure compliance with the reference test. This uses a specific sample size/methodology (ie 50/XYZ). Regardless, afaik, the choice of your control system/sample size is in Principle entirely up to you since the Inspector's reference test is independent of yr implemented setup. However I assume you realize that sample sizes/methodologies other than 50/XYZ require the use of additional "factors" such as noted in pk1 ..
In addition to Post 4, material in pk1 and other files (eg pk4,pk5,pk9) attached below enable a "success" prediction for the effectiveness of various alternative setups when compared to the " reference test" procedure. This involves hypothesis testing which requires some basic statistical capabilities (eg see AnnexD,pk5).
Some related files which I found useful -
pk2 - 2.2.3_Packers Guide_Version 4.01-1.pdf 74.71KB 23 downloads
pk3 - Welmec Guide_6.4_2015_Guide_for_packers_and_importers_of_e-marked_prepacked_products.pdf 934.23KB 17 downloads
pk4 - Average weight,compliance 3 rules for Packers pdf.pdf 863.97KB 26 downloads
pk5 - WELMEC_Guide_6.5,2012,-i_2_Cntrls_By_Competent_Depts_on_e_marked_Prepack.pdf 889.77KB 11 downloads
pk6 - Principles of Checkweighing,Mettler.pdf 589.83KB 19 downloads
pk7 - weights and measures (UK) 2006, No.659.pdf 239.04KB 11 downloads
pk8 - Average Quantity System,Queensland,2006.pdf 929.46KB 19 downloads
pk9 -NWML, quantity-average-guidance-2007.pdf 354.68KB 15 downloads
pk10 - UK,2006,No.659 weights and measures (packaged goods) Regs.pdf 148.45KB 11 downloads
pk11 - EEC Directive 76-211(,1976) weight prepackaged products.pdf 736.67KB 18 downloads
pk12 - Welmec Guide,6.6,2013,recognition of procedures.pdf 538.76KB 15 downloads
PS - I suspect that control characteristics of US Systems may not relate to this thread unless they incorporate T1/T2 concepts.
PPS - JFI these.3 previous (UK ) threads are vaguely related to this thread -
https://www.ifsqn.co...ion/#entry24701
2008
https://www.ifsqn.co...ion/#entry33484
2009
https://www.ifsqn.co...ce/#entry102583
2016
Predictably, all the above had difficulties with the complexities of the UK Weight control scenario.
Hi everyone,
I have 2 bakeries in my clients. None of them do any single bread weighing.
Why ?
The process is "industrialized". They make boxes of x kg for the first raise. Then they put the content of that box into a mecanical divider, which cuts the loaf in y parts. Each part (or sometimes 2 or 3 together) are then shaped to make single loafs. And they know, because they have validated the process, that such loafs will make 400g, 800g or any required weight.
And the weight, for product able to lose weight (like drying bread) must be ensure at the end of the shelflife.
So, what about the directive telling about 80 samples ? Well, this explains how authorities must control the weight. FBO must be sure they respect the law, but they are not obliged to do the 80 samples.
In any case, do not forget to keep records of your validation.
Thats interesting. I always thought that dividers in bakeries (but i may be wrong) work based on volume and not on weight. And bread is live product with live cultures, so it may vary batch by batch depending on weather. I would not reply on validation purely. but i guess it may be the product you are doing, or you have better dividers :D
Hi leatcookies,
It rather depends on what you mean by "OK?"
(Definitive information on this topic requires a copy of the "Packers' Code (1979 !)" which is an elusive object albeit still commercially available (a summary of its contents are in file pk2 below). Other available publications, eg Welmec series contain, afaik, summaries of aspects in the Packers' Code document.).
IMEX of exports to UK/EEC some Companies simply avoid the control complexities discussed below by substantial overfill and accepting the financial hit.
I assume the OP comments relate to this document -
You did not mention if the EC "e" logo is directly involved so I assumed not. This has its own, relatively well-defined, expectations.
I also assumed a checkweigher is not currently involved.(there is considerable discussion on using such option in various files attached below).
I noted/ignored some specific comments in pk1 related to bakers/bread packing.
I assume by "OK" you mean that your system will be accepted by a W&M Inspector as equally effective when compared to the "reference test" as discussed in pk1. (BRC acceptability is presumably a separate issue which I have ignored for present purposes). For example quoting pk1 -
Hi Charles,
Yes, we do apply the e-mark, but we do not sell abroad. We do not have a checkweigher. I will have to go through all the legislations and links attached to see if not providing the calculation will be a problem.
Basically I wondered if i need to calculate the average of sampled product if my final product-loaf does fall under declared weight. And if one of them falls under declared weight, can I only average the 10 samples taken in one 15 min time range.
Generally in my opinion I will be compliant to 3 packers rule, because my product's average value will be above declared weight. I will just physically not calculate it. Any feedback, please let me know :)
Hi Charles,
Yes, we do apply the e-mark, but we do not sell abroad. We do not have a checkweigher. I will have to go through all the legislations and links attached to see if not providing the calculation will be a problem.
Basically I wondered if i need to calculate the average of sampled product if my final product-loaf does fall under declared weight. And if one of them falls under declared weight, can I only average the 10 samples taken in one 15 min time range.
Generally in my opinion I will be compliant to 3 packers rule, because my product's average value will be above declared weight. I will just physically not calculate it. Any feedback, please let me know :)
Hi Ieatcookies,
I am not legal (or Brexit[?]) expert but I noted this -
Using the ‘e’ mark
Packages that meet the requirements of the Packaged Goods Regulations and are between 5 g to 10 kg and 5 ml to 10 l can apply the ‘℮’ mark.
Packaged goods labelled with the ‘℮’ mark are declaring their compliance with the requirements of the ‘average system’ under Directive 76/211/EEC and are not subject to further weights and measures regulation. It is an optional choice for the packer whether to display the ‘℮’ mark.
Packages which do not display the ‘℮’ mark and are to be traded outside the UK must meet the regulations set by the destination country.
The ‘℮’ mark is a metrological passport to trade allowing free access within the EEA and its respective markets.
etc, etc
https://www.gov.uk/g...sing-the-e-mark
Offhand, the above mentioned Directive looked statistically somewhat more "explicit" than some of the UK-specific documents although there are obvious conceptual overlaps.
Hi Ieatcookies,
I am not legal (or Brexit[?]) expert but I noted this -
https://www.gov.uk/g...sing-the-e-mark
Offhand, the above mentioned Directive looked statistically somewhat more "explicit" than some of the UK-specific documents although there are obvious conceptual overlaps.
Great! thanks a lot!