The Following item is a document produced by the weighing machine specialists, Oertling, and although in its present form it doesn’t include the use of computer controlled scales and weighing equipment, it DOES provide the basic requirements and technology needed for those involved in Average Weight.
AVERAGE WEIGHT THE SIMPLE PRACTICALITIES INTRODUCTION
CHAPTER 1 CONTROL BY WEIGHING EACH PACK ON A STAMPED SCALE CHAPTER 2 CONTROLBYSAMPLING CHAPTER 3 IMPORTERS CHAPTER 4 CONTROL BYAUTOMATIC CHECKWEIGHER CHAPTER 5 CONTROL BY MEASURING CONTAINER BOTTLES APPENDIX I- GLOSSARY APPENDIX II- SOURCES OERTLING LTD. Cray Valley Works Telephone: Orpington (0689) 25771 St. Mary Cray, Orpington, Telex: 896459 Kent BR5 2HA OERTLING LTD. Oaktree Place, Telephone: 051 645 0639 St. Pauls Road, Rockferry, Telex: 628050 Birkenhead,L42 1NF
1
INTRODUCTION AVERAGE WEIGHT LEGISLATION This is an introduction to the average weight legislation. We hope when you have read this through, you should have a grasp of the implications, and at least know where to obtain further information if you need it. It has been written in a basic manner, but a glossary of some of the technical terms is included. The whole basis of the law is the Weights & Measures Act 1979. It concerns packaged goods. Up to now a package was supposed to contain at least the amount printed on the label. This could be checked at any time, especially at retail level, and if the amount in the package was not as printed on the label, a prosecution could result. Overfilling was necessary. Now control and checking will happen at the factory, or when goods are imported. (Details of the checks necessary by an importer are shown in Chapter 3). This may mean a worthwhile reduction in overfill is possible, and a more efficient process. What the Law says From 1st January, 1980, the packer must ensure, if a Weights and Measures' Inspector carries out a Reference Test on a group of packages, that a test is passed. (It is worth noting that Oertling have supplied the majority of Weights and Measures' Authorities with their Reference Test scales). This means that the packer must have a good knowledge of the performance of his filling machine, and be sure that he is always putting enough in each package. This implies he must check his production to see this is happening. So he can prove this, records will have to be kept, in most cases. The Three Rules for Packers sum up the principles of the Average Quantity System: 1) The average contents of packages shall be not less, on average, than the Nominal Content (Qn). (What this means is that on average, packs will have in them/at least, the amount shown on the label). 2) Not more than 1 in 40 packages shall be non-standard, i.e. contain less than T1. (This means that not more than 2½% of production shall lie below a certain limit which is known as T1 - this limit is shown in Table 1, and varies between 1% of label weight for the heaviest packs and 9% for the lightest packs) 3) No package shall be inadequate, i.e. contain less than T 2. (This means there is an absolute bar on selling any package which falls below a second limit known as T2 specified in Table 1 - this limit varies between 2% of label weight for the heaviest pack and 18% for the lightest packs.)
TABLE1 Tolerable Negative Errors Nominal Quantity (Qn) g or ml 5 to 50 to 100 to 200 to 300 to 500 to 1000 to 10000 to above
50 100 200 300 500 1000 10000 15000 15000
Tolerable Negative Error (TNE} As % of Nominal Quantity (Qn) 9 4.5 3 1.5 1
(TNE as % should be rounded to nearest 0.1 g or ml above) T1 is the nominal quantity minus the tolerable negative error. So for 300gT1 = 300g — (3% of 300g) = 300g—9g =291g. T2 is the nominal quantity minus twice the tolerable negative error. So for 300g, T2 = 300g — 2 x (3 of 300g) = 300g — 2 x 9g = 300g — 18g = 282g. 2
g or ml 4.5 9 15 150 -
What the packer must do is to find out what target weight to aim at to ensure that he meets these three rules, and then carry out checks to be sure this is in fact happening, and then in most cases record and retain the results of these checks for at least 1 year for possible examination by a Weights and Measures' Inspector (also known as a Trading Standards Officer). There are several ways of controlling production: 1. Control by weighing out each pack on a stamped scale (See Chapter 1) 2. Control by sampling (See Chapter 2) 3. Or are you an importer ? (See Chapter 3) 4. Control by automatic checkweighers. (See Chapter 4) 5. Control by measuring container bottles. (See Chapter 5) Whatever scales you are using must comply with Table 2, which gives the largest permitted scale intervals.
TABLE 2 Least nominal package weight (or weight corresponding to least nominal volume)For which the scale is used.
LARGEST PERMITTED SCALE INTERVAL
digital 5g inclusive to 10g exclusive 10g to 20g 20g to 50g 50g to 200g 200g to 1kg 1kg to 2.5kg 2.5kg to 5kg 5 to 20kg 20kg to 40kg 40kg to 100kg inclusive
0.1g 0.25g 0.5g 1g 2.5g 5g 10g 25g 50g 100g
analogue 0.2g 0.5g 1g 2g 5g 10g 20g 50g 100g 200g
digital 1 grain 2.5 5 1 /32 oz ½ dram 1 /8 oz 1 /8 oz ¼ oz 1 oz 2oz 4oz
analogue 2 grains 5 10 1 /16 oz 1 dram ¼ oz ¼ oz ½ oz 2 oz 4oz 8oz
There are many cases when the Packers' Code recommends the use of scales with tighter tolerances, in order that recommended values may be achieved. If scales are not DOT stamped (or do not have the EEC initial verification mark) they must be checked every working day using stamped weights at maximum capacity and the nominal quantities used.
CHAPTER 1 Control by weighing out each pack on a stamped scale If a packer uses a stamped scale in accordance with Table 2 for making up every package, with the intervention of an operator, this is continuing to operate on the minimum system. There is no need to carry out checks and keep records. A check that the minor divisions on your current scales conform to those in Table 2 for the weights you are packing to is recommended as the requirements are now generally more stringent. This is particularly true for weights up to 1 kg. The Inspector will still carry out a Reference Test on the average system, but as the use of this equipment implies each package will contain at least the label weight, the Reference Test should always be passed with ease. Oertling have stamped scales available. Further information, if needed, is in the Packers' Code, Chapter 2.
CHAPTER 2 Control by Sampling If control by sampling is to be used, the scales must be at least as good as in Table 2. What must first be established is the variability of fill of each product where packed by each machine used for this product. A detailed procedure for establishing this process characteristic is given shortly. 3
A target weight must then be calculated. This consists of some or all of the following: a. The weight stated on the label b. The container weight (tare) c. Tare variability allowance (for gross weights only) d. Process variability allowance (i.e. if there is a wide spread in the output of a filling machine) e. Additional allowance for wandering average f. Sampling allowance (the more samples taken, the smaller this is) g. Storage allowance (for desiccating products) h. Miscellaneous factors Details are given later on how to do this. Having established the target weight, the checks are now made by weighing packs and recording the weights. Provided the results prove that the rules for packers are being obeyed, all is well. If the results drift away, to the point where even 1 of the 3 rules will be broken, then corrective action must be taken and it must be recorded that this action has been taken. In addition, any production which has been packed since the last check will need to be specially screened if inadequate (below T 2) packs have been produced. It is very obvious that there is a trade off between overfilling and less inspection, and running exactly to the label weight, and doing more inspection. To carry out the checks it is possible to use scales and hand-written records, but this can be labour intensive and subject to human error. Alternatively, over-filling with reduced inspection can take place, but this may be costly. If a packer wishes to maximise any possible reduction in overfill, without drastically increasing labour costs, it may be helpful to use more precise weighing equipment linked to simply operated calculators to provide filling machine adjustment information; this can be presented via control charts or calculator printout. Oertling have available well proven systems to meet these requirements. Details are now given step by step of the component parts of establishing that target weight.
Assessment of the process characteristic of a filling line This must be established by collecting data and analysing the result. One method is now given. 1. Collect data From the end of the production line, take 5 consecutive packages, and measure the net contents. Repeat at regular intervals till 40 sets of readings have been obtained, over at least 3 production periods. (The Packers' Code recommends that the scales used be at least as accurate as Table 2. Elsewhere better accuracies are suggested). Record the weight of individual empty containers (tares) for each sample in every set of five weighings. 2. Summarise data _ Use a calculator to calculate the arithmetic mean (x) and standard deviation (S) for each of the sets of 5 net weights. Having done this, calculate the short term variability So by squaring each value of S, adding them together, dividing by the number of sets of samples, 40, and taking the square root. The result Is So - the short term standard deviation. (An alternative method is to add up all the values of S, divide this by 40, and multiply by 1.064). The medium term process variability (Sp) is the standard deviation of all 200 readings of net weights taken as one complete run of the figures. Oertling have weighing and calculating equipment to help make this as painless as possible. 3. Test the data Test A. Check to see if the distribution is Normal. (This means that a check is being made to see if the scatter of the results agrees with the theoretical scatter expected in the legislation). _ From the overall mean subtract twice the short term standard deviation ( x - 2 So). If more than 4 of the 200 weights are below this figure. Packers' Rule 2 is in danger of being broken even if overfilling takes place. Reason for the scatter must be sought. 4
_ Now from the overall mean subtract 3.72 times the short term standard deviation, (x - 3.72 So). If any of the 200 weights lie below this. Rule 3 will be broken and the reason for it must be sought, to prevent recurrences. (Otherwise 100% check weighing will be necessary). Test B. Check on the variability of fill If the medium term standard deviation Sp divided by the short term standard deviation So (Sp — So) is more than 1.056 then it must be concluded that the fill varies appreciably from time to time. It is possible the process is over controlled, which means that too many machine adjustments are being made. In this case it may be more economic to increase the target weight and make less adjustment. Sp should be used for setting the target quantity, and a check should be made for wandering average. Test C. Check on the medium term process variability (Sp) to see that Packers' Rules 2 and 3 are not being broken. In Table 3 two figures are given. If Sp is less than the first figure, there is unlikely to be any difficulty in meeting Rules 2 and 3. If Sp is between the two figures. Rule 2 is likely to be broken. Target weight should be set to at least T1 + 2Sp. So for a pack of 300g nominal weight (T1 = 291 g), with a medium term standard deviation of 5g (Sp = 5g) the target weight must be at least 291 g + 10g = 301g. If Sp is above the higher figure, then Rule 3 is likely to be broken. Target weight must be set to at least T 2 + 3.72 Sp. So for a pack of 500g nominal weight T2 = 470g) with a medium term standard deviation of 10g (Sp = 10g), the target weight must be at least 470g + 37.2g = 507.2g.
TABLE 3 Critical values of Sp (the medium term standard deviation) If Sp is equal to or less than in Column 1, probably all is well. If it is more than in Column 1 but less than in Column 2, Rule 2 (T1) is at risk. If it is more than in Column 2, Rule 3 (T 2) is at risk. Overfilling will then be necessary. (The figures in Column 1 are based on the relevant Tolerable Negative Error, divided by 2, and in Column 2 by TNE divided by 1.72). Nominal Quantity g or ml 5 10 20 30 40 50-100 150 200-300 400 500-1,000 2,000 3,000 4,000 5,000 10,000-15,000 Above 15,000
Critical Values of Sp Column 1 Column 2 0.2 0.3 0.4 0.5 0.9 1.0 1.3 1.6 1.8 2.1 2.2 2.6 3.4 3.9 4.5 5.2 6.0 7.0 7.5 8.7 15.0 17.5 22.5 26.2 30.0 34.9 37.5 43.6 75.0 87.3 (TNE/2) (TNE/1.72)
Additional allowance for wandering average This may be necessary if the sample average wanders significantly from one set of samples to the next, and the process cannot be controlled, or it would be impractical to do so. If this clearly does not apply to you, go on to Tare Variation. 5
This additional allowance is made by increasing Sp by a factor based on a variability component known as S1. The method of doing so, given below, is based on 2 estimates of the standard error (Se). _ Estimate A. Take the 40 averages (x) from the process characteristic, and calculate the difference between each adjacent value. Add up these differences. Divide by 39. Multiply this result by 8/9. This gives estimate A. Estimate B. Using a calculator, calculate the standard deviation of the 40 values of x. This gives estimate B. Now divide A by B. If the result is between 0.8 and 1.2, no additional allowance is necessary. If A/B is not between 0.8 and 1.2, the larger value of A or B should be taken and the following calculation performed. The new medium term standard deviation Sp is calculated by squaring the larger of A or B, squaring Sp, adding them together, and taking the square root. ______ Sp’= √So2+S1 . Sp’ should be used for calculating the process variability allowance. Please note that if this section is applicable to your process the Packers' Code should be consulted, as this section is condensed in the interests of simplicity. Tare Variation In establishing the process characteristic, 40 sets of average tares were found. In precisely the same way as finding the short term standard deviation of the process characteristic, find the standard deviation of the tares (St). If the standard deviation of the tares is less than one tenth of the Tolerable Negative Error (see Table 1), an average constant tare can be used. If it is more than 1/10 of TNE, the target fill should be increased by an allowance. This allowance is calculated by squaring the medium term standard deviation (Sp), squaring the tare standard deviation (St), adding the two together and taking the square root of the result. ________ Sp(t)=√Sp2+St2 This should now be used wherever Sp would otherwise be used. (If an allowance is also necessary for a wandering average, then the tare variations should be added in as well, along the following lines. _________ Sp’(t)=√So'+S12+St2 It is of course possible, to pre-tare containers by weighing each container empty and then deducting this weight from the filled container to give the net weight. No allowance is then necessary. Tare variability should be checked from time to time, and particularly when a new source of containers is used.
Sampling allowance. An increase must be made to the target weight to account for the length of time it takes a sampling system to detect a change in the filling process. The amount of this allowance depends on the type of control chart used, and the number of samples taken during a production period. A production period is defined in the following manner: (1) If the production rate is over 10,000 packages in an hour, the production period is one hour. (2) If the production rate is under 1250 in an hour, the production period is one 8-10 hour shift. (N.B. For production rates under 500 in a shift consult your Inspector). (3) If the production rate lies between these figures, the production period is the time taken to produce 10,000 packages. As an example, for someone using a Shewhart control chart, with both warning and action limits, taking 2 sets of samples of 5 packages each period, the allowance to be added is 0.29 times S o (see Table 4). For a greater frequency of sampling the allowance is less, and vice versa. Using simpler charts the allowance is greater, and vice versa. Oertling have systems to help you reduce your sampling allowance, or even eliminate the use of control charts.
6
TABLE 4 Sampling Allowance (expressed in units of S o) Sample Size 2 2 3 4 5 6 10
A B A B A B A B A B A B
0.84 0.58 0.70 0.43 0.61 0.35 0.54 0.29 0.47 0.25 0.26 0.15
Number of time samples taken per hour or production period 3 4 5 6 8 10 0.65 0.43 0.53 0.32 0.46 0.25 0.37 0.20 0.31 0.17 0.15 0.08
0.54 0.35 0.44 0.25 0.35 0.19 0.27 0.15 0.21 0.12 0.07 0.03
0.46 0.34 0.37 0.20 0.27 0.21 0.20 0.11 0.15 0.08 Nil Nil
0.40 0.25 0.31 0.17 0.21 0.12 0.15 0.08 0.10 0.06 Nil Nil
0.32 0.19 0.21 0.12 0.13 0.07 0.07 0.03 0.03 Nil Nil Nil
0.26 0.15 0.15 0.08 0.07 0.03 Nil Nil Nil Nil Nil Nil
A=Action limit chart only.
B =Warning and action limit chart.
Storage Allowance This is only necessary for products which desiccate, although an allowance may also be made for commercial reasons. It is normally set at so many per year.
CALCULATING THE TARGET WEIGHT We need some or all of the following: a Nominal weight (This is known as a commercial decision or by law) b. The container weight (This is known from the tare variability check—if applicable) c. Tare variability (This is known from the tare variability check—if applicable) d. Process variability allowance (Known from process characteristic) e Additional allowance for wandering average (if no controls applied) f. Sampling allowance (Known from type of chart and sampling frequency) g Storage allowance (Known from previous tests if applicable) h. Miscellaneous factors (Density factors etc. known from previous tests) Take a 300g pack, which desiccates at 1% per year, with a tare of 100g which has a standard deviation of 1g; a short term standard deviation (So) of 4.7g and a medium term standard deviation (Sp) of 5.0g. (TNE is 9g, T1 is 291 g and T2 is 282g - see Table 1). Sp First Test A is applied, which is found satisfactory. Then Test B is applied and So is found to be 1 063, which is more than 1.056, so Sp should be used for calculating the target weight. As we intend to adjust the process, which does not show significant variation of the sample average no additional allowance will be made for wandering average. The tare standard deviation is 1g, which is greater than 1 /10 of TNE. So a new medium term standard deviation is calculated as: ______ _____ ____ Sp (t) = √Sp+ St = √52 +12 = √26 = 5.1g TNE TNE Test C is performed, and 2 is 4.5g, and 1.72 is 5.2g (see Table 3) Sp (t) is 5.1 g, which lies between these figures. So target must be at least T1 +2Sp (t) = 291 + (2 x 5.1) = 301.2g. So the variability allowance is 1.2g. The packer is packing at the rate of 5000 per hour, and wishes to take 2 samples of 5 each hour. i.e. 4 samples of 5 in a production period. He will use a Shewhart Chart with action and warning limits From Table 4 this gives a factor of 0.15 times, in this case, Sp(t) = 0.15 x 5.1 = 0.77g The container weight will be deducted in each case by the calculator linked to the scale storage allowance of 1% must be added to this final result, as there are no miscellaneous factors such as density to consider:
7
Qt = nominal
g 300 (Qn)
+ variability allowance + allowance for wandering average + sampling allowance
1 -2 0 0.77 301.97 3.02 304.99 (Qt)
+ storage allowance 1% Qt Will be set at 305g
Control limits All that remains to be done is to set warning and action limits for this target. For the mean the warning limit is set at a recommended 0.88 of the short term standard deviation (So) and the action limit at a recommended 1.38 So If the mean value falls below the action limit, the level of fill must be raised, and production since the last sample must be checked for inadequate (below T2) packages. Similarly, if the warning limit is breached, and an immediate second sample is also below the warning limit, the same action should be taken. To prevent overfilling the upper limits should also be watched. So for our target weight of 305g the warning and action limits are: upper lower warning 309.1 300.9 action 311.5 298.5 being 305 ± (0.88 x 4.7) 305.9 ± (1.38 x 4.7) For the standard deviation the control limit is a recommended 1.92 times the short term standard deviation. So for this target it is 9.0g If this limit is exceeded a check should be made of recent production to see no inadequate packages have been produced, and the short term standard deviation should be re-calculated. (This is done by taking the square root of the squares of the last ten vale of S divided by 10) _____________ i.e. So = √S12 + S22….S1 o2 10 If So has changed significantly the minimum target must be re-calculated, and new limits set. Summary What has been indicated is one way to meet the law. Oertling have scales, calculators and charts to help you do this. Further information is available in the Packers Code. An example of an Oertling Shewhart control chart, duly completed, is shown on the next page.
CHAPTER 3 Importers Several questions need to be asked. 1) Who is the importer? 2) What is a "relevant package ? 3) What must the importer do ? 4) How to carry out checks. 1.
The importer is the one who orders the customs entry to be made. So it depends on circumstances. What this means is that if an importing company buys products on its own account, brings them in and subsequently decides to re-sell them then they are the importer. But is a supermarket chain had asked the importing company to buy products on its behalf then the supermarket chain would be the importer,
8
9
2)
A "relevant package" is defined in Schedule 1 to the Weights and Measures (Packaged Goods) Regulations, 1979. Basically it means any closed package up to 100kg with a weight or volume declaration (except 'e' marked packages coming from an EEC country).
3)
The importer must (a) either make checks on each consignment, and make records of those checks, and keep the records for a year, or (b) get "certification documents" from his supplier. These documents must either be the actual production records relating to the consignment, or a statement from the Weights and Measures authorities of his country that they control his production to the same level as U.K. packers are controlled, or a written declaration from the overseas packer that each package has in it at least the nominal quantity declared. (The Packers' Code says that if importers use certification documents, they should still do some checks, in order to have a cast iron defence if one of their consignments fails a Reference Test).
How to carry out checks The cheapest way of checking the net contents of packages is by sampling. The sampling plans suggested are double sampling plans. This means that the test may be carried out in two parts; the second part is only used if the first part does not give a clear cut decision. The sample sizes are as follows:
TABLE 5 Consignment size
100-500 501-3200 3201 or more
Sample size Part1 8 12 20
Part 2 12 18 28
Total 20 30 48
A check follows these lines. 1) 2) 3) 4) 5) 6) 7) 8)
Obtain samples Weigh them Open some, and determine container weights (tares) Calculate net weights Check weight against criteria Accept, reject or refer for part two test If necessary, do part two test Accept or reject
If the consignment is rejected, it must be rectified, or returned to the supplier. 1) Obtain samples. This should be done in a reasonably random manner. For example, to get a sample of 30, take 10 cases at random, and take 3 packages at random from each case. 2) Weigh them. The scales used should conform to Table 2. 3) Determine container weights. The best way is to open the two lightest and two heaviest packages, and to then weigh the empty, clean containers. Two checks can then be done. If the standard deviation of the containers is less than one tenth the tolerable negative error (see Table 1), use the heaviest container weight as a constant tare. If this check fails, provided the net weight of the four packages was above the label weight, the average weight of the 4 containers may be used as a constant tare. If this check also fails, then all packages must be opened to check the actual contents.
10
4) If a constant tare could be used, calculate the net weights. If not, open all packages, weigh the empty containers and then calculate the net weights. 5) Check the net weights against the criteria. Here we must give the criteria. They may look fearsome at first sight; however, provided a calculator is used, which has a standard deviation key, they are not so fearsome. Oertling have electronic scales linked to calculators available which make the whole procedure very simple.
TABLE 6 Group or consignment size
Test for nonstandard packs Preliminary Referred Combined Accept Refer
100-500
Sample
Sizes
8
501 – 3200
12
3201 or more
20
12
20
18
30
28
48
0 1 0 2 1 3
1 1 2 -
Reject 2 2 2 3 3 4
Criterion for Average content k (see below) 0.599 0.640 0.430 0.503 0.297 0.387
For test for average: _ _ On first sample accept if x = Qn – k.s. refer if x < Qn - k.s. _ _ _ On second sample, calculate x and s for combined sample: accept if x = Qn - k.s. reject if x < Qn - k.s. For test for non-standard packages; operate as indicated in worked example. Key _ x is the mean (average) quantity of the sample = means "is equal to or greater than" < means "is less than" Qn is the nominal quantity, stated or to be stated on the container k is the factor from the table above s is the standard deviation of the sample 6)
Having checked the weights against Table 6, it will be seen whether the consignment has been accepted, rejected or referred for Part 2 of the test. Once an importer has obtained a pattern of accepted consignment for a commodity from an overseas packer, he may switch to a reduced inspection. This level of inspection is also recommended even when certification documents are held.
The Code suggests that the sample size should be 8. The acceptance level is set at, for non-standard products (i.e. below T1), not more than 1, and for the contents, label weight minus 1.237 times the standard deviation. If this test is not passed, then the checks shown in Table 6 should be carried out. Oertling have scales available, which together with a pocket calculator, can cope with these checks. For those importers doing a lot of checks, scales linked to electronic printing calculators can do all the arithmetic work, and give the accept, refer and reject information. Further information on the legislation can be found in the Packers' Code. A worked example follows.
11
TABLE 7 Worked example Product: 'X' Canned pears Nominal Quantity: 500g Consignment size: 2,400 Sample sizes from Table 6: (1) 12 (2) 18 TNE from Table 1 =15g Qn-T1 = 485g Qn-T2 = 470g Sample No. 1 2 3 4 5 6 7 8 9 10 11 12
Gross Wt{g) 571 565 562 575 572 561 568 570 546 564 561 573
Tare Wt{g) 68 68 68 68 68 68 68 68 68 68 68 68
Net Non-standard Wt{g) or inadequate 503 497 494 507 504 493 500 502 478————————NS 496 493 505 _______
Total: 5972g *The tare weight was taken as constant because it was found that the weight of the 4 empty tins initially tested did not vary by more than 1.5g (1 /10 TNE). Applying the criterion for non-standard packages in Table 6, the sample is referred because 1 is present. _ 5972 The mean weight x is 12 = 497.7g and the standard deviation s = 7.87g and from Table 6, k = 0.430 Therefore, Qn - k.s = 500 - (0.430 x 7.87) =500-3.38 _ =496.6 x = 497.7g Therefore the sample is accepted for this test. But the second part of the test for non-standard packages must be carried out, and a further 18 cans must be weighed. If no more than 1 non-standard is found, the consignment is acceptable - otherwise it is rejected.
CHAPTER 4 Control by Automatic Checkweighers (It is worth interjecting that Oertling do not see a commercial conflict from their viewpoint between a sampling system and a checkweighing system . There are certain cases where it is clearly obvious that a checkweigher is the right answer - most noticeably in the case of long runs Equally for short runs, because of the required setting up 12
and checking procedures, checkweighers are uneconomic. This also applies where tares are highly variable. In this case, control by sampling is more economic.) All types of checkweigher must be checked initially to establish the "zone of indecision'; Test packs must be prepared using scales at least as good as Table 2. However, the Packers Code recommends (Appendix D6) using scales reading to one tenth of these values. There are 3 types of checkweigher, each of which must be used in a specified manner: 1) 2) 3)
Simple equipment to reject underweights at a specified preset weight, known as the set point. As 1, but which also count and display numbers of packages above and below the set point. As 1, but also checks and records the average gross or net weight.
In all cases records of the fixing and checking of the set point must be kept; In the case of 1, unless the rejection is at the nominal gross weight (the minimum system) back up checks on a suitable scale complying with Table 2 must be done. In the case of 2, the "count ratio" must be checked against a "reference ratio". In the case of 3, a warning must be given, or automatic action taken if the weights go below target. In both cases 2 and 3, if things are going wrong, the setting up procedure must be repeated, using a suitable scale. In all cases, therefore, where a checkweigher is used, a scale as shown in Table 2 is needed. In order to be able to establish "set points" as accurately as possible, for small packs (up to say 200g) it is desirable to read to 0.01 g, and for larger packs (up to say 3kg), 0.1 g is desirable. (See examples in the Packers' Code). Oertling have scales to meet the requirements of the above checks. Further information, if needed, is in the Packers' Code, Appendix D.
CHAPTER 5 Control by Measuring Container Bottles A Measuring Container Bottle (MCB) is manufactured to a nominal volume within certain tolerances. Provided that a packer fills an MCB to a given level, and checks and records these checks using a template, he meets the law. However quasi- MCB are now permitted under certain specified criteria. Scales, perhaps linked to calculating equipment, are desirable for checking these criteria. For establishing process variability the Packers' Code recommends measuring the contents of bottles in the most accurate way available, (i.e. pretaring, gross weighing and converting to volume at 20°C.) Oertling have scales and calculators available for these purposes. Further information if needed is in the Packers' Code, Appendix E.
13
APPENDIX 1 GLOSSARY OF TERMS ABSOLUTE TOLERANCE LIMIT T2 The nominal quantity minus twice the tolerable negative error, i.e.T 2=Qn-2TNE.
ACTUAL CONTENTS, X The net quantity of goods in a package.
AVERAGE The arithmetic mean of a set of numerical values, i.e. their total divided by their number. The average of values represented by a symbol, e.g. x, is denoted by a bar above that symbol, i.e. x.
AVERAGE CONTENTS. X The arithmetic mean contents of a number of packages.
BATCH A number of packages of the same type, forming a homogeneous collection for the packer's quantity control purposes.
GROUP A number of packages of the same type and production run, forming the subject of a reference test.
INADEQUATE PACKAGE A package whose contents are less than the absolute tolerance limit, T2.
NEGATIVE ERROR The amount by which the actual contents of a package fall short of the nominal quantity.
NOMINAL QUANTITY Qn The quantity marked on the container of a package.
NON STANDARD PACKAGE A package whose contents are less than the tolerance limit T1.
PACKAGE A container, together with the predetermined quantity of goods it contains, made up in the absence of the purchaser in such a way that none of the goods can be removed without opening the container.
PRODUCT LINE Packages of the same type, appearance and nominal quantity.
PRODUCTION RUN The process whereby packages are made up in the same place during the same period or periods in which the conditions do not materially alter.
REFERENCE TEST The procedure employed by an inspector to check whether a packer is complying with his duty under section 1 (1) of the 1979 Act. 14
TARGET QUANTITY Qt The average contents which a packing or filling operation is intended to produce.
TOLERABLE NEGATIVE ERROR, TNE The negative error in relation to a particular nominal quantity, as defined by the 1979 Regulations.
TOLERANCE LIMIT T1 The nominal quantity minus the tolerable negative error, i.e. T 1 = Qn-TNE.
APPENDIX 2 Sources 1. Code of Practical guidance for packers and importers. Weights and Measures Act1979 Issue No. 1. ISBN 0 11 5129229 HMSO £3.50 2. The Weights and Measures (Packaged Goods) Regulations 1979. Sl 1979 No. 1613 ISBN 0 11 0946138 HMSO £1.75 3. Weights and Measures Act 1979. ISBN 010 544579 7 HMSO £1.00
15
AUTHORS NOTE: The above item has been included in full and some parts may exceed that which is needed for the ‘run of the mill’ meat trader but it was felt that it would be better to include the full document for the benefit of those who could use it. A later version may be available by contacting the company who prepared the document.
16
