GB2114329A

GB2114329A - Computers for use in sampling procedures

Info

Publication number: GB2114329A
Application number: GB08203001A
Authority: GB
Inventors: David John Fisher
Original assignee: Individual
Current assignee: Individual
Priority date: 1982-02-02
Filing date: 1982-02-02
Publication date: 1983-08-17
Also published as: AU1038983A; EP0086059A2; BR8300496A; EP0086059A3; JPS58176766A; GB2114329B

Abstract

The computer comprises a memory into which is entered tables of data identifying a sample code by batch size and inspection level, and a file table of data representing a sampling plan identifying sample size and acceptance and rejection parameters by the sample code and an acceptable quality level. The sample code is identified by the batch size and a predetermined inspection level. Data are entered representing a predetermined acceptable quality level and the identified sample code is used to identify the corresponding sample size and acceptance and rejection parameters. A visual display responds to the identified sample size to display this sample size. In response to the identified acceptance and rejection parameters a value is stored which is a function of such parameters. Sample data are entered representing the acceptance and rejection of each sample article tested and the entered sample data are compared with the derived and stored value to signal batch acceptance or batch rejection. <IMAGE>

Description

1

GB 2 114 329 A 1

SPECIFICATION

Improvements in or relating to computers for use in sampling procedures

5

This invention relates to improvements in computers for use in sampling procedures.

Sampling procedures are known in which a given number of samples of a batch of products is tested to 10 determine the extent of any defective attributes the samples may possess and the result of such a test is used as a measure of whether or not the batch is acceptable. Standards have been worked out and tables compiled for this purpose in which the 15 accept/reject conditions are statistically based upon expansion of the Binomial and Poisson probability distributions. In most mass/batch production companies in the United Kingdom, goods that are received in bulk are inspected in accordance with the 20 various charts and levels detailed in British Standard 6000,6001 and 6002. However, as will be apparent hereafter, the complications of these charts are such that a considerable degree of training is required to understand the principles involved and inspection is 25 generally held to the simplest plan available which is known as the "single sampling plan" although cost savings could be achieved using the more complicated plan known as the "sequential sampling plan".

The tables and charts referred to, and in particular, 30 B.S. 6001:72 will be well known to and understood by those skilled in the art. However a brief description will be given forthe better understanding of the invention.

In sampling procedures it is necessary for the 35 consumer or purchaser and the manufacturer to agree both the inspection level and the Acceptable Quality Level (AQL). BS 6001:72 provides for seven inspection levels of which four levels are special levels S-1; S-2; S-3 and S-4, used only for small 40 sample sizes where relatively large risks can be taken and three general levels LI, Lll and Llll of which Lll is the most commonly used, level LI being specified where less discrimination is needed and level Llll being specified for greater discrimination. 45 For a given batch size, the inspection level provides a code which defines the sample size to be inspected according to the sampling plan to be adopted.

The AQL represents the minimum quality which can be accepted and may be expressed in per cent 50 defective or in defects per hundred units for values of 10 or less but for values above 10 it must be expressed in terms of defects per hundred units. Inspection for per cent defective assumes that an article with one defect is to be rejected as much as an 55 article with three defects and inspection for defects per hundred units assumes that an article with three defects is of as much concern as three articles with one defect.

The sampling plans are arranged as columns of 60 numbers, the columns being identified by the agreed AQL and the rows being identified from the code derived from the batch size and the inspection level. For a single sampling plan the columns comprise two sub-columns which identify the condition for 65 acceptance or rejection of a batch.

The single sampling plan requires all samples to be inspected and the numbers in the sub-columns of the AQL columns define the maximum number of defective articles which can be accepted and the minimum number of defective articles leading to rejection of the batch.

For example for a batch size of 1201-3200 articles, the inspection level Lll code is K. The normal single sample plan identifies this code as requiring a sample size of 125 articles and the AQL column for 1.0% gives the acceptance figure of 3 and the rejection figure of 4. All 125 articles in the sample have to be inspected and if 3 or less are rejected, the batch is considered acceptable and if 4 or more are rejected, the batch is rejected. In addition to the normal single sample plan, there are tightened single sample plans and reduced single sample plans. Generally, the normal plan is used but when two out of five consecutive batches have been rejected it is normally required to move to the tightened plan until five consecutive batches have proved acceptable when a return may be made to the normal plan. Reduced inspection may be instituted under certain conditions which include:

a) 10 or more batches or a predetermined number of samples have been normally inspected and the total number of defectives is less than a given number defined on another chart;

b) production is at a steady rate, and c) the responsible authority considers reduced inspection desirable.

Normal inspection is generally resumed if one of these conditions is breached.

Thus in the above example code K for a tightened single sample plan still requires 125 samples but the figures for acceptance or rejection for an AQL of 1.0% are now 2 and 3. In the same example but for a reduced single sample plan the code K only requires 50 samples to be tested and the corresponding AQL numbers are 1 and 4 respectively. This means that if examination of the 50 samples reveals not more than 1 reject the batch is accepted and reduced inspection continues. If the number rejected in the sample is less than 4 but greater than 1 the batch is accepted but the normal sampling plan must be resumed. If the number rejected in the sample is 4 or more, the batch is rejected and the normal sampling plan must be resumed.

In addition to the single sampling plan there are two other similar plans, each with its own normal, reduced and tightened versions. These are known as the "Double Sampling Plan" and the "Multiple Sampling Plan". Doubling sampling is a system in which a first sample is taken which is of a lesser number than it would be for a single sampling. If the quality is found to be sufficiently good, the batch may be accepted straight away and, if the quality is found to be sufficiently bad, the batch may be rejected. If, however, the quality is intermediate these two conditions, a second sample of the same size as the first must be taken and inspected before a decision is made, the number of defective articles in the first and second sample sizes being cumulative. Thus, in the above example, the code K gives a first sample size of 80 and a second sample size of 80 and

70

75

80

85

90

95

100

105

110

115

120

125

130

2

GB 2 114 329 A

2

the acceptance and rejection figures for an AQL of 1.0% are respectively 1 and 4 for the first sample and 4 and 5 for the cumulative first and second samples.

The Multiple Sampling Plan is essentially the same 5 as the Double Sampling Plan except that more than two samples may be called for. The plans generally provide for seven sampling stages all of equal size, although it is rare for a seventh stage to be required.

The above described sampling plans require all 10 samples in a sample size to be examined and each sample rejected counts as one, the rejections being cumulative.

Another plan which is of importance is known as the "Sequential Sampling Plan" and operates on the 15 basis of a starting handicap and a penalty number. In this plan account is taken also of acceptable articles and, for each acceptable article, 1 is added to the handicap number and, for each rejected article, the penalty number is subtracted from the running total 20 of the handicap number. The batch is deemed acceptable when the running total of the handicap number reaches a value which is twice the initial handicap number and the batch is rejected if the running total reaches zero or less.

25 The tables for the Sequential Sampling Plans differ from those of the Single Sampling Plans in that the AQL columns are single columns, i.e., there are no sub-columns representing acceptance and rejection numbers, and it is the rows which comprise two 30 sub-rows, one sub-row representing the handicap (H) figure and the other sub-row representing the (b) penalty number.

Referring once more to the above example, the table for Normal Sequential Sampling shows that 35 the code K requires 224 articles in a sample. For an AQL of 1.0% the H figure is 54 and the b figure is 36. For each acceptable article is the sample 1 is added to the handicap figure and for each rejected article in the sample 36 is subtracted from the running total. It 40 will be seen that, if the batch is good, only 54 articles need to be inspected to reach the acceptance figure of 2 x 54= 108 as compared with the normal Single Sampling Plan which requires 125.

There is also a tightened version of the Sequential 45 Sampling Plan but there is no reduced version.

It will be appreciated from the above discussion that considerable staff training is required to enable full use to be made of the sampling plans.

It is an object of the present invention to provide a 50 computer for use in sampling procedures which shall be capable of being produced in portable form and which shall be operable by relatively unskilled staff.

According to the present invention there is pro-55 vided a computer for use in sampling procedures for testing samples of a batch of articles comprising a memory into which is entered at legst a first table of data identifying a sample code of batchsize and inspection level and a file comprising at least one 60 sampling plan as a table of data identifying sample size and acceptance and rejection parameters by the sample code and an acceptable quality level (AQL), means for entering into the computer data representing the batch size and a predetermined inspec-65 tion level to identify in the memory said sample code, means for entering into the computer data representing a predetermine AQL and for utilising the identified sample code to identify in the memory the corresponding sample size and acceptance and 70 rejection parameters, a visual display responsive to the identified sample size to display said sample size, means responsive to the identified acceptance and rejection parameters to derive and store a value having a magnitude which is a function of said 75 parameters, means for entering into computer sample data representing the acceptance and rejection of each sample article tested, comparison means operable to compare said entered sample data with said value to derive a first signal representing batch 80 acceptance or a second signal representing batch rejection.

Embodiments of the invention will now be described by way of example, reference being made to the accompanying drawings in which:-85 Figure 7 is a perspective view of a hand-held calculator-type computer embodying the present invention,

Figure 2 is a plan view of the lower part of the computer case shown in Figure 1, the top part being 90 removed,

Figure 3 is a plan view of the inside of the top part of the computer case,

Figures 4A and 4B together represent a flow diagram illustrating one embodiment of the inven-95 tion, and

Figures 5A and 5B together represent a flow diagram illustrating another embodiment of the invention.

The computer illustrated in Figure 1 is similar in 100 configuration to a hand-held calculator and comprises a substantially rectangular case indicated generally at 1 having an upper part 2 and a lower part 3 which can be secured together in any suitable manner. A number of function buttons or keys are 105 mounted on the upper part 2 and comprise a "cancel entry" button 4, accept and reject buttons A5 and R6 respectively, ten digit buttons 7 representing the digits 0 to 9, a decimal point button la, three buttons 8,9 and 10 identified as the AQL button, the Level 110 button and the Batch button respectively, three buttons 11,12 and 13 representing the normal sampling plan, the reduced sampling plan and the tightened sampling plan respectively, and four buttons 14,15,16 and 17 to identify the plan selected as 115 Double Sampling Plan, Multiple Sampling Plan,

Single Sampling Plan and Sequential Sampling Plan. Also mounted in this upper case part 2 is a Texas TIL 398.3 Type 8 digit 7 segment LED display module 18 which is viewed through window 19. The lower case 120 part 3 houses a printed circuit board 20 on which are mounted a Texas TMS 1000 Series Micro-Processor 21 and a memory microchip 22, connections being made between the circuit board 20 and the function keys and the LED display module 18 by a ribbon 125 cable 23. A rechargeable battery power pack 24 is connected to the circuit board 20 and also to pins 25 opening to the exterior of the casing for recharging purposes. An audible alarm 26 incorporating an amplifier is also connected to the circuit board 20 130 and the lower part 3 of the casing is vented as at 27

3

GB 2 114 329 A

3

to allow the alarm to be heard. An on-off switch 28 is provided and may be located as required; in Figure 3 it is shown mounted in the lower part 3 of the casing, and in Figure 1 it is shown mounted in the upper part 5 2 of the casing. It could equally be provided as an on-off press button.

The embodiment of the invention illustrated in Figures 4A and 4B is intended to provide for both Single and Sequential Sampling Plans including the 10 normal, reduced and tightened versions of the Single Sampling Plan and the normal and tightened versions of the Sequential Sampling Plan. The tables or charts representing these plans are stored in the memory microchip 22. Each of the function buttons 15 of the computer, when operated, generates its own unique code in binary or equivalentform, e.g., hexadecimal, and the computer is designed, by operation of these function buttons, to identify the selected table in the memory, to identify the code 20 corresponding to the batch size and, by this code and the agreed AQL, to identify the appropriate acceptance and rejection numbers or the handicap and penalty numbers as the case may be. Operation of the rejection button R6 in the case of a Single 25 Sampling Plan or operation of the acceptance button A5 and the rejection button R6 in the case of a Sequential Sampling Plan then generates values which are automatically compared with the acceptance and rejection values of the plan to generate an 30 alarm when the batch is to be rejected.

Referring to Figures 4A and 4B, an operation is commenced by switching the computer on by means of switch 27. The operator then operates the selected mode or plan button, let us assume the 35 Single Plan button 16, which then sends its code to a comparator unit 30 which compares the received code with stored codes and, if it identifies the code as acceptable, transmits the code at its Y (Yes)

output or, if it is not acceptable, produces an N (No) 40 output, the latter being used to operate the alarm 26 to indicate a malfunction. (All outputs referenced 100 are used to operate the alarm 26.) The coded Y output from comparator unit 30 is supplied to a comparator unit 31 arranged to identify the Single 45 Plan code and to produce a Y output when it identifies that code or to pass the code to its N output when it fails to identify that code. In the latter case, the code is passed to the input of a comparator unit 32 arranged to produce a Y output if it identifies 50 and accepts the code as representing a sequential sampling plan or an N output if it fails to do so in which case the N output actuates the alarm 26. However, assuming a Single Sampling Plan is selected, the Y output from comparator 31 (theY31 55 output) opens a gate 33 which is arranged to be supplied with the codes generated by operation of the buttons 11,12 and 13 representing a normal plan, a reduced plan or a tightened plan respectively.

Whichever plan is selected (assume normal, i.e., 60 button 11 is operated) the corresponding code is passed to a comparator unit 34 which produces an N output to operate the alarm 26 if the received code is not one of the three permissible codes or, if acceptable, passes the code to a string of comparator units 65 35,36 and 37 each arranged separately to identify one of the three permissible codes and when so identified to produce a Y output. The Y output from comparator 35 identifies in the memory 22 the stored Single Sampling Plan (normal inspection) -70 sayTable 1 -; the Y output from comparator36 identifies in the memory 22 the Single Sampling Plan (tightened inspection)-sayTable2-andthe Y output from the comparator 37 identifies the Single Sampling Plan (reduced inspection). An N output 75 from comparator 37 indicates that the received code has not been accepted by any one of the comparators 35,36 and 37 and operates the alarm 26. At this stage, the computer has identified the table to be used in the memory 22.

80 The second stage is to identify the required column in that table by entering the agreed AQL. A Single Sampling Plan usually provides for 26 AQL values ranging from 0.010 to 1000 and an appropriate value must be entered to select the appropriate 85 AQL column. This value is entered to select the appropriate AQL column. This value is entered by the digit keys 7 and is displayed by the LED display 18. If a mistake has been made and the AQL is wrong, it can be cancelled by operating the "cancel 90 entry" button 4. However, assuming it to be correct, the AQL button 8 is operated to enter the number which is checked by comparator unit 38 and, if acceptable, is passed to the string of 26 comparator units 391 to 3926 and, if one of these accepts the 95 number it produces a Y output which identifies the column in the selected table.

The third stage is to identify the sample code from the batch size and the level selected from levels LI, Lll and Llll.

100 To this end, the digit buttons 7 are operated to display the level required, e.g. 1,2 or 3, in the display 18 and the level button 9 is actuated to enter the displayed level. Actuation of button 9 also instructs the computer that an operation is to be performed 105 on the Level/Batch size table stored in the memory 22. This entered level is checked by comparator unit 40, and passed to a string of three comparators 41, 42,43 arranged to identify levels I, II and III respectively. Whichever one of the comparators 110 identifies the level, it produces a Y output which identifies the required column 44\ 442 or 443 in the Level/Batch size table. The batch size is then entered by digit keys 7, displayed on display 18 and entered by operating batch button 10. This entered size is 115 checked by comparator unit 45 to see if it is acceptable and, if so, is passed through the Y45 output to a string of fifteen comparators 461 to4615 which by determining whether the batch size is below a predetermined value determine whether or 120 not the batch size is within a given predetermined range. Thus comparator461 determines whetherthe entered batch size is less than 8, comparator 46s whether it is less than 500, comparator 4614 whether it is less than 500,000 and comparator 4615 whether it 125 is greater than 500,000. The comparator 46 which identifies the range of the batch size produces a Y outout which in turn identifies the line or row471 to 4715 in the Level/Batch size table to identify the sample code. This sample code when applied to the 130 selected table does two things, firstly it identifies in

4

GB 2 114 329 A

4

the memory 22 the sample size and, secondly, it identifies in the memory 22 the accept/reject numbers for a Single Sampling Plan or the handicap and penalty numbers for a Sequential Sampling Plan. 5 These are read from the memory into a temporary store 48 which transmits the sample size to the display 18 and which also transmits the accept/reject numbers or handicap/penalty numbers to a switch 49 which has inputs 50 and 51 derived respectively 10 fromY31 and Y32 outputs to identify whether a Single or Sequential Sampling Plan has been selected and to route the numbers accordingly. With a Single Sampling Plan selected, the switch 49 routes the accept/reject numbers via lines 52 and 53 15 respectively to comparator units 55 and 56 to enter the acceptance number into comparator units 55 and the rejection number into comparator unit 56.

The operator now selects from the batch the displayed number of samples and commences test-20 ing. For each article of acceptable quality, the A5 button is operated and for each article to be rejected, the R6 button is operated. The operation of buttons A5 and R6 generates a corresponding coded signal which is supplied to switch 57 which is also control-25 led by the Y31 and Y32 outputs along lines 50 and 51 to distinguish between the Single and Sequential Plans and to route the coded signals accordingly.

Also each time the A5 and R6 buttons are operated, a signal is supplied to the display 18 to deduct 1 from 30 the displayed number. Assuming a Single Sampling Plan, the coded signals are routed along line 58 to a comparator unit 59 operable to identify each "Accept" code which produces a Y output for each "Accept" code which is supplied to a store 60 and N 35 output for each "Reject" code which is identified in comparator unit 61 to produce a Y output which is supplied to the store 60 and to the comparator units 55 and 56. The store 60 is in effect a counter loaded or biased with the sample size number from the 40 temporary store 48 and counts the number of acceptances and rejections to produce an output along line 62 when this count reaches the sample size number.

If before the count has been completed, the 45 number of rejects equals the reject number, this is detected by the comparator 56 to produce a Y output which is used to cause the display unit 18 to display "FAIL" to show that the batch is not acceptable. If, however, the full sample size is tested without this 50 occurring, the output from counter 60 along line 62 is passed to gate 63 which is a gate normally held open by the N output from the comparator 55 when the reject number is not greater than the acceptance number which has been supplied along line 52. As 55 the acceptance and rejection numbers only differ by 1 unit in the tightened and normal versions of the Single Sampling Plan, the display, for such a selection, will either show FAIL or PASS. However, for the reduced version of the Single Sample Plan, the 60 difference between acceptance and rejection is greater than 1 unit and the corresponding numbers supplied along lines 52 and 53 to the comparators 55 and 56 will differ accordingly. Therefore, at the end of a sampling procedure, if the reject number has not 65 been exceeded, it is necessary to know whether the acceptance number has been exceeded. If it has, the N55 output will disappear leaving gate 63 closed. The Y55 output indicates that the number of rejects exceeds the acceptance number and this is supplied 70 to a gate 64 which is only open when it receives a first input from the N56 which is only open when it receives a first input from the N56 output indicating that the number of rejects is less than the reject number and a second input from line 62 indicating 75 that the count is complete. Under such conditions the Y55 output is passed through the gate 64 to cause the display 18 to display "PASS/F" to indicate that the batch can be accepted but that the normal version of the plan must be resumed. The PASS or 80 FAIL display may be accompanied by an audible signal from the alarm 26 such as, for example, a continuous sound representing a PASS and an intermittent "bleep" representing a FAIL.

The description of Figures 4A and 4B has so far 85 largely been concerned with the selection of a Single Sampling Plan. If a Sequential Sampling Plan is selected, the corresponding code is not accepted by the Comparator 31 which passes it through the N output to the comparator 32 where it is accepted to 90 generate a Y output which opens gate 73 which is arranged to be supplied with the codes generated by operation of the buttons 11 and 13 representing a normal or tightened plan respectively. Such a code is passed to comparator unit 74 which produces an N 95 output at line 100 if it does not identify one of the two acceptable codes or, if the code is acceptable, passes the code through its Y output to comparator unit 75 intended to identify a normal plan and produce a corresponding Y output or pass the code through its 100 N output to comparator 76 intended to identify selection of a tightened plan. The Y75 and Y76 outputs identify that a Sequential Sampling Plan has been selected and whether or not the normal or tightened version is required, thereby enabling the 105 computer to identify the required table in its memory 22, - say tables 10 and 11, respectively.

The subsequent entry of the AQL, the selected level and the batch size are as previously described with the result that the display 18 displays the 110 required sample size, and the temporary store receives this sample size and, in this case, the appropriate handicap and penalty numbers H and b. As in this case, the line 51 is energised from Y32, the switch 49 supplies the H and b numbers along lines 115 77 and 78 respectively, the b number being fed to a store 79 and the H number being supplied to a counter 81 and also being doubled in multiplier 80 and supplied as 2H to comparator 82. The Y32 output along line 51 operates the switch 57 to send the 120 signals from the A5 and R6 buttons along line 82 to comparator unit 83 intended to identify an acceptance and produce a corresponding Y output. If it is a rejection signal, the comparator 83 passes the signal to comparator unit 84 intended to identify a rejection 125 signal and produce a corresponding Y output. Each Y83 output adds 1 to the number in counter 81 and each Y84 output causes the store 79 to transmit the penalty number to the counter 81 where it is subtracted from the running total. The running count 130 in counter 81 is monitored by the comparator 82 and

5

GB 2 114 329 A

5

a further comparator 85 which is set to identify when the running count is less than 1, i.e., zero. The comparator 82 produces an output on line 86 when running count attains the value 2H causing the 5 display unit 18 to display PASS. The comparator 85 produces an output on line 87 when the running count falls below 1 or passes through zero to cause the display unit 18to display FAIL.

The embodiment described with reference to 10 Figures 4A and 4B may additionally incorporate the facility of using either one or both of the Double Sampling Plan and the Multiple Sampling Plan in anyone or all of the versions reduced, normal and tightened. It will be described as incorporating both 15 plans and all versions. To this end, the comparator 30 (Figure 4A) would additionally store the codes representing the Double and Multiple Sampling Plans which would be selected by the function buttons 14 and 15 and if a Y output is produced 20 would supply this code to the comparator string 31,32 which would be enlarged by the addition of comparators D31 and M31 (Figure 5A). The N32 output would no longer go to the 100 line but to the D31 input as is represented by the line 101 (Figures 25 4A and 5A). As the Double and Multiple Sampling Plans are very similar to the Single Sampling Plan, particularly the reduced version thereof, the references used forthe Double and Multiple Sampling Plans will, where appropriate, be the same as those 30 used forthe Single Sampling Plan prefixed by the letter D or the letter M to denote "Double" or "Multiple".

Figure 5A shows the arrangement for selecting the appropriate table for a Double or Multiple Sampling 35 Plan from the memory 22 and as it corresponds in substance to that described for the Single Sampling Plan, it need not be further described. Also, the entry of the AQL is to identify the column inthe selected table, the entry of the level and the batch size to 40 identify the sample size code are all as previously described and are therefore omitted from this description. The derived sample code will identify in the selected table the sample size which, as has been mentioned, is of identical size for each of the two 45 sample sizes which may be required for the Double Sampling Plan and is of identical size for each of the seven sample sizes which may be required forthe Multiple Sampling Plan. This sample size is read into the temporary store 48 (Figure 4B) and displayed in 50 the display 18. The derived sample code also identifies the accept/reject numbers for each sample size so that for the Double Sampling Plan there will be two pairs of accept/reject numbers and forthe Multiple Sampling Plan there will be seven such 55 pairs of accept/reject numbers.

The comparators D31 and M31 provide a Y output along line 50, if selected, and this output operates the switches 49 and 57 as previously described. The first pair of accept/reject numbers are fed through 60 switch 49 along lines 52 and 53 respectively to the respective comparators 55 and 56 and the sample size is supplied to store 60. However, in order to accommodate the Double and Multiple Sampling Plans, the block A shown in dotted lines in Figure 4B 65 needs to be modified and one such possible modification is shown in Figure 5B. In this modification, the output from gate 64 is supplied to a switch 110 the output of which is normally supplied along line 111 to the memory 22 or is switched along line 112 to the 70 display 18to display PASS/F. In orderto identify whether a Double Sampling Plan or a Multiple Sampling Plan has been selected and hence the possible number of sample sizes which may require to be tested, the comparator D31 (Figure 5A) has a Y 75 output along line D50 and the comparator M31 has a Y output along line M50 and these outputs are supplied to a counter 113 which is set by a D50 input to count two sample sizes and by the M50 output to count seven sample sizes and when such a count is 80 completed, the counter 113 produces an output along line 114to operate the switch 110 and pass the output from gate 64 to the display 18 to display PASS/F.

Assuming a Multiple Sampling Plan is selected, 85 the store 60 will be supplied from the memory with the sample size for the first testing and the comparators 55 and 56 will be respectively supplied with the accept/reject numbers of the first pair. As with the reduced version of the Single Sampling Plan if, at the 90 end of testing the first sample size, the acceptance number has not been exceeded, the N55 output will cause the display 18to indicate PASS and, if the rejection number has been equalled or exceeded, the Y56 output will cause the display 18 to indicate 95 FAIL. If, however, the number is intermediate the acceptance and rejection numbers, there will be an output from gate 64 and a second identical sample size will need to be tested. The output from gate 64 will be supplied to the counter 113 and through the 100 switch 110 along line 111 to the memory 22. This will cause the memory 22 to redisplay the sample size on the display 18, to re-enter the sample size in the store 60 and to supply the second pair of accept/reject numbers to the comparators 55 and 56 respectively. 105 Each time a sample size of a given number of samples is tested, the display 18 will either show PASS or FAIL or there will be an output from gate 64 which passes through switch 110 and along line 111 to the memory 22 to set up the apparatus for the next 110 sample size test. In the event that the display 18 does not previously show PASS or FAIL, the final sample size to be tested will be the seventh. Forthe normal and tightened versions of the Double and Multiple Sampling Plans, the accept/reject numbers for the 115 final test differ by only one unit so that there can be no intermediate value to operate gate 64 and the display 18 will either show PASS or FAIL. However, forthe reduced version, the accept/reject numbers for the final sample size differ by more than one unit 120 and can produce an output at gate 64 if the actual number is intermediate the given values. In this case the output from gate 64, in this case the seventh output, is recognised in the counter 113 which has been set up by the M50 input thereto, and this 125 produces an output along line 114 to operate the switch 110 and direct the output from gate 64 along line 115 to the display 18 to produce the display PASS/F.

If a Double Sampling Plan is selected, it will be 130 appreciated that there is only a first and a last

6

GB 2 114 329 A

6

sampling size, i.e., two sampling sizes of identical numbers of articles, to be tested and the operation is otherwise as described above, the counter 113 being set up by the D50 input to operate the switch 110 at 5 the end of the second sample size test.

The tables representing the sampling plans and their alternative versions may be considered as a file of plans stored in the memory 22. For each such file plan, it is necessary to identify the sample code from 10 a table which determines the sample code by the batch size and the inspection level which is determined. It will be appreciated that, if desired, the file may only contain one version of one plan so as to materially simplify the device and in any given 15 application prevent an unskilled operator from having the opportunity of selecting an incorrect plan or the version thereof required. The file could, in fact, be in a separate memory capable of being plugged into the apparatus to provide the facility required for 20 any given application. In this way, the apparatus could be issued to an employee set up forthe purpose required and the employee would not have to make any selection of the plan or version required.

25 Where the apparatus holds less than the four plans and the versions thereof described, the number of function buttons can be reduced accordingly.

It will be appreciated that the apparatus may hold any one or more of the plans described and the 30 respective versions thereof.

It will be further appreciated that the invention is not limited to the embodiments described byway of * example and that various equivalents or modifications will be apparent to one skilled in the art and 35 may be employed without departing from the scope of the invention. Such modifications and equivalents are to be considered to be within the scope of the invention disclosed in this specification.

Claims

40 CLAIMS

1. A computer for use in sampling procedures for testing samples of a batch of articles comprising a memory into which is entered at least a first table of 45 data identifying a sample code by batch size and inspection level, and a file comprising at least one sampling plan as a table of data identifying sample size and acceptance and rejection parameters by the sample code and an acceptable quality level (AQL), 50 means for entering into the computer data representing the batch size and a predetermined inspection level to identify in the memory said sample code, means for entering into the computer data representing a predetermined AQL and for utilising 55 the identified sample code to identify in the memory the corresponding sample size and acceptance and rejection parameters, a visual display responsive to the identified sample size to display said sample size, means responsive to the identified acceptance 60 and rejection parameters to derive and store a value having a magnitude which is a function of said parameters, means for entering into the computer sample data representing the acceptance and rejection of each sample article tested, comparison 65 means operable to compare said entered sample data with said value to derive a first signal representing batch acceptance or a second signal representing batch rejection.

2. A computer according to Claim 1 in which the file table of data represents a single sampling plan in which the rejection parameters represent the minimum number of rejected articles in a sample size requiring batch rejection, the means responsive to said parameters is operable to derive said value as a function of said rejection parameters and said comparison means is operable to compare the entered sample data representing rejection with said value to generate said second signal representing batch rejection when said entered sample data representing rejection corresponds to said value and to generate said first signal representing batch acceptance when said entered sample data representing rejection has not attained said value and said entered data representing both sample acceptance and rejection identifies completion of testing of the full sample size.

3. A computer according to Claim 2 in which the file table includes at least the normal and tightened versions of the single sampling plan, and means is provided for entering into the computer a code identifying the version of the single sampling plan to be used.

4. A computer according to Claim 2 or 3 in which the file table includes the reduced version of the single sampling plan in which the acceptance parameters represent the maximum number of rejected articles allowable in a sample size to allow batch acceptance and continued use of the reduced version and any value intermediate said minimum and maximum values allows batch acceptance and denotes that use of the reduced version should be discontinued, said means responsive to the identified acceptance and rejection parameters being arranged to derive and store a first value which is a function of the acceptance parameters and a second value which is a function of the rejection parameters and said comparison means being operable to compare said entered sample data representing rejection of each rejected sample article tested with said first and second values to derive said second signal representing batch rejection in the event said entered sample data representing rejection attains said second value, said first signal at the end of the sample size test representing batch acceptance in the event said entered sample data representing rejection does not exceed said first value and a third signal at the end of the sample size test representing batch acceptance and discontinuance of the reduced version in the event said entered sample data representing rejection is intermediate said first and second values.

5. A computer according to Claim 1 in which the file comprises at least one sequential sampling plan table in which the acceptance and rejection parameters represent handicap and penalty numbers, the means responsive to the parameters is operable to derive and store a first value representing the penalty number, to derive and store a second value representing the handicap number, and to derive and store a third value which is a multiple of said

70

75

80

85

90

95

100

105

110

115

120

125

130

7

GB 2 114 329 A 7

handicap number, and said comparison means is operable to deduct said first value from said second value for each entry of sample data representing rejection of a sample, to add a predetermined value 5 to said second value for each entry of said sample data representing acceptance of a sample to derive a running total and to compare said running total with said third value to generate said first signal representing batch acceptance when said running total 10 attains said third value and said second signal when running total falls below or attains a predetermined minimum value.

6. A computer according to Claim 1 in which the file table of data represents at least one sampling

15 plan providing for the test of at least two sample sizes of a batch in which the rejection parameters for each sample size represent the minimum number of rejected articles in the sample sizes tested requiring batch rejection, the acceptance parameters for each 20 sample size represent the minimum number of rejected articles allowable in the sample sizes tested to allow batch acceptance and any value intermediate said minimum and maximum numbers for all but the last sample size to be tested denotes that a 25 further sample size is to be tested, said means responsive to the identified acceptance and rejection parameters being arranged to the identified acceptance and rejection parameters being arranged to derive and store, for each sample size, a first value 30 which is a function of the acceptance parameters and a second value which is a function of the rejection parameters, said comparison means being operable to compare said entered sample data representing rejection of each rejected sample arti-35 cle tested with said first and second values to derive said second signal representing batch rejection in the event said entered sample data representing rejection attains said second value, said first signal at the end of each sample size test representing 40 batch acceptance in the event said entered data representing injection does not exceed said first value and a third signal at the end of each but the last sample size to be tested representing the requirement for testing a further sample size, said third 45 signal being operable to identify the further acceptance and rejection parameters for such further test said means responsive to said identified acceptance and rejection values being operable to derive and store a third value which is a function of the further 50 acceptance parameter and a fourth value which is a function of the rejection parameter.

7. A computer according to Claim 4 modified in that the file table represents a sampling plan allowing for the testing of at least two identical sample

55 sizes in a batch in which at least all but the last sample size to be tested have acceptance and rejection parameters allowing forthe entered sample data to be intermediate said first and second values for that sample size, characterised in that said 60 third signal is operable to identify that a further sample size is to be tested and to identify the further acceptance and rejection parameters for such further sample, the means responsive to the identified acceptance and rejection parameters being arranged 65 to derive and store values corresponding to such further acceptance and rejection parameters for the testing of a further sample size.

8. A computer according to Claim 7 in which the sampling plan is a double sampling plan allowing for

70 the testing of up to two identical sample sizes in a batch.

9. A computer according to Claim 7 in which the sampling plan is a multiple sampling plan allowing for the testing of up to seven identical sample sizes

75 in a batch.

10. A computer according to Claim 1 in which the file includes at least one version of at least one of a single sampling plan, a sequential sampling plan, a double sampling plan and a multiple sampling plan

80 constructed and arranged to operate substantially as herein described.

11. A computer for use in sampling procedures having a memory in which is tested at least one version of a sampling plan, a table for identifying a

85 sample code, means operable to enter the batch size to be tested and the inspection level required to identify in the table the sample code, means for entering an acceptable quality level, means responsive to the identified sample code and the entered

90 acceptable quality level to identify the sample size to be tested and to identify acceptance and rejection parameters from the sampling plan, means for entering acceptance and rejection characteristics of samples tested in the identified sample size and

95 means for comparing such entered characteristics with the identified acceptance and rejection parameters to provide a signal representing batch acceptance or rejection.

Printed for Her Majesty's Stationery Office, by Croydon Printing Company Limited, Croydon, Surrey, 1983.

Published by The Patent Office, 25 Southampton Buildings, London, WC2A1 AY, from which copies may be obtained.