ISO TR 7821:1982 download.Tobacco — Preparation and constitution of identical samples from the same lot (Code of practice, for collaborative studies for evaluating methods of test).
The repeatability and the reproducibility of a particular laboratory method are defined by ISO 3534. These characteristics may be deterrnned by means of inter-laboratory tests as described in ISO 5725.
In addition to the descnption of the equipment and the methods of test to be used, often by reference to existing standards, the nrgrniition of aith tnsts also indudes the preparation of test materials. i.e. samples which will serve as medis for the studies to be carried out.
It should be noted that, in this type of study, there is rarely any interest in the characteristics of the test material itself, provided that such characteristics are within the typical range for the materials commonly tested, and values are determined in this way only in order to obtain the quantitative date neccasary for compiling a statistical report concerning the method of test.
Thus, in most cases, every effort is made to obtain lots or samples which are as skirilar to each other as possible, ‘in such a way as to reduce the residual variance of the tests and, in the same way, to increase the accuracy of the conclusions which may be drawn from the global analysis of the results.
Since the individual test samples do not necessarily need to represent the population from which they are taken, but only have to be identical to each other, the method for thair preparation can be quite different from those needed when the aim of the study is to Obtain a knowledge of the original population.
In the case of certain industries, the chemical industry for example, it is relatively easy to make up samples which are practically identical to each other, as the structure of the material allows It to be divided up so as to conserve the microscopic or macroscopc hoeiioystieuy of the product, for example solutions, powders, etc.
In the case of tobacco, however, it is completely different, because the i-aw materials, and even the finished products, are affected by an intrinsic heterogeneity likely to cause considerable differences between samples if special precautions are not taken.
Tobacco, as a test medium, may occur in the following torms:
powdor:
— cut tobacco (scaterlati);
— cigarettes;
— leaves or strips.
— )eavesorstrips.
The method for the preparation of samples of test media depends, in most cases, on the type of test to be carried out. Nevertheless, it seems possibde to lay down a few general guidelines applicable to a large number of collaborative studies, depending on the form in which the material occurs.
1 Scope and field of appliciitlun
This Tectw,iosI Ropori provide. a pracdcel l,ain.wotk for the operation. of dI.ideig a lot (not flce.eWSP homogaiwovel of tobacco or
tobacco producte nfo .ev.ral skodar s.n’gites intended lot boretive atua ether wIthin or between or,tories.
2 References
ISO 3634, St.t,arica — VocabcdVy and iymOc
ISO 5725. Prec,sior, of as! ,naøtods — D.tarnskebo,’, of fafastabrny and , dutttdity by w w4or,tory tests,
3 Preparation of samples
3.1 General
Th. division of a tot into wrple. which are iiar to each other is vy ee.y if the tot Is lor can be made. according to Its physical state) homoganeot. but may be more dltfktiIt liii b i.4. In thai c, en q rqJiaiL. sampling proc.dta. he. to be used In order that the irOluenca of heterogeneity is as sm as ple.
It i olMoudy not po1l. to de.alb. in detail avery cae. which Irey ale..
Th, aim of th fcaowuiq ewmple.. therefore, isto thistrate the praiciple. of, and to glvs guid.nce on. tI* operations to be carried Out. which. moreover, appear Irrçb to conceive beasig in mind
— on the one Pid, it. given aiim. which is to malt. up tample. which e sitnilar e. poeIle;
— on the oilier hand, to take into account the fact that t tars wd ptobbty have charactefleilca which are most slimier if they are saropied from point. which are physically dose together.
3.2 Powder
The flhld lot comprises N g.
N
It h to be dlvided up into ii sanaslas of — g.
a
3.2.1 PrInciple
Fact .en,b consist, 01 a ,iu,mi)er tO inrn.vwrs
Each Inoanient is ten aeperst.y froim the lot end shotuld. is theory, be docated ai random to eech of the a sample., In practice, it
* iuf%clait to efocate the increments to sect of 1*1.4 sample. is rotation, provided th.t the mxrter of cydes of rotation lie. m,,*r of Incrernews is each eamplel is lerge.
3.2.2 Preceutlons to b. taken
During the preparation of the sample., the foiowlng precaullon. shotid be taken:
a) Won is an atmospher, which has a relative humidity as dose sa possible to that corre.pondlng to it. equilibrium htariIdlty 01 the n*tlrid b.lng handled.
bi Work qulc*Iy se poaule In order to limit the exchange of molatute b.tw..n the atmoephesa and the material it it is not puseitib to satisfy tumpletalir çondlhiun .1.
3.2.3 Example
A grosa sample of 1 kg of ponder Is to be divided into 10 ewoplee of 100g. The condition to be luIfiU.d for the cabboratie. studee is not that loaan’vlee of exactly 100 g each alsOirki it. obtained, but that each e.mpleot ipproudmately 100g dtouk) be similar, (totn the phy,icl and chermicI pobit 01 view, to e.ch of the others.
3.2.3.1 M,nu.I rn,thod
After mixing the woes sanpIe as theroughty as poasibia, place the pile on a tray or in a racepearJa frnm which the matreied can be easlly scooped out. Use a spoon, ladle or measure, of capacity between 5 and 10 g for example. Provide 10 boxes capable of holdeig albeit bOg each.
Taking a erKcsaann of iimg the spoon. ails or measure, feed. at ttNn. box 1, box 3 box ID. Each simple thus made up of at least 1011010 increments.
Mien the initial pile is akiosr exhauated, a. when there is only enough left for one round of 10 ertienle. It le rec’Ornoiended that dititbullon be finished w+th a meense of smeller capeaty.
This method guarantee. diet the 10 samples thus constituted are as stnhlar as possible My mitdng of the groas sample will cause vadetion of ertide a. and denSIty through th pile from top to bottom, end the sI’ioukl be atcar, Wgo account when taking atcrernents.
32.3.2 Automatic method
Us. a mechanical device audi that. at each revolution, th, quantity attributed to each sample le lwe than or equal to one’lenth of the total owes of the final serople.
3.3 Cut tobacco (scat enlati)
3.3.1 General
Exactly the now jmipts as far powder Ia applicable. In practice, haidlig Is more difflailt the lobsoco rmsey form tan.., eeoecly II It is in long strands.
The temptation to make up the itcrements from Waite or knps should be resisted, and, Instead, they should be brolcan down so met they can be divided between all the samples.
As a general ride, eedi iample should be med tap of it least 10 lr,cremenls (and more if poesltile), In addition, kicriments shtwld always be distributed to each sample in a standard rotation.
Depending on the size of the kit, the iicn,m,nt may be of the order of a large handful, a handful or a pinch
3.3.2 SpecIal precautions
Cut tobacco lacaferlatli always tends to settle, i.e. handlatg causes e cartim amaimf of separation of the short strands which collect it lit, bottom of the receptacle.
In order to avoid cauaing en artificial heterogeneity between samples when the long strangle have bean ashaustad, it is necaseery to distribute the shail slrwdi .qua#y between eanlea using an instrument audi as. scoop, spoon, etc.. which prevents the lee. of these strands at the time of dittibutiOn.
3.4 Cigarette.
3.4.1 CIgarettes to be sampled from the production machate.
Ills knowfl that ther, are serial corteletlons between the cheract,rlstlcs of cigarettes I.e., on average, the probabdity of Cigarette. being Identical is greater the shorter the keerval between their fabrication by die cigarette making macins.
For example, if 10 samples, each of 100 agerettas. are to be prepared. procsied follow.:
a) take a nes of 10 consecutive ogarettas and distributa therm to the 10 simple. to be made lap:
in repeat the opsietion 100 dine..
NOTES
1 There Is rio ileethenage if stew lea pep Deiween talOng Increments lot 10 clperetteel, Le. If ate tenth ogerette of one increment and the first cigarette at his mai atciin’W as not suava.
2 To save time, it is permissible to take increments successively. For example 20, 50. or even 100 successive cigarettes can be taken. In ths case, It is undesirable to keep the cigarettes In bulk, and they should be kept in individual increments, labelled in the order of manufacture, and should be thctrihiitad In that order, rotating round the tan vimpkc tn he conetituted in the ..siii manner For this purpose, automatic sampling devices can be used.
experience shows that making up samples in this way gives results as good as, or even better than, selection by mass or draw resistance as far as comparability of samples is concerned.
3.4.2 Cigarettes sampled from a bulk cigarette storage before packing
For the same reasons as described in 3.4.1, it s necessary to avoid constituting the samples by taking each of them in a group from the case.
Instead, it is recommended that one of the two following procedures be used in order to constitute, for example, 10 samples of 100 ctgdrtrttes:
a) take successive cgarettes for each sample at random, for example using a table of random numbers to which, for example, the (approximate) coordinates are made to correspond to a point from where each cigarette is taken;
b) take each horizontal row of the storage bin or case as a primary sample and take from It one increment for each of the samples being made up. It is not necessary in this case for the whole row to be allocated; the main point is that each sample should contain cigarettes from every row selected as a primary sample.
Procedure a) is theoretically the most satisifactory, but it can be unwieldy to apply when there is a large number of samples (greater than 3 or 4), comprising a large number of cigarettes, to be constituted.
In current practice, procedure b), which is more rapid, is amply sufficient, because of the partial stratification which is involved, aixi thus compensates for the fact that the sampling is not totally at random.
3.4.3 Cigarettes in packets
Initial lots of cigarettes in packets may, as appropriate, be made up of:
— packets in bulk;
— cartons or units (for example units of 200);
— boxes.
The procedure to be used for sampling the lot for the purposes ld down in this Technical Report relies on observabon of the following principles
a) Take into account the fact that two cigarettes are, probably, most similar in their characteristics, it they originate from the same package unit.
h) Map? the method of sampling to the volume of the samples to be made up.
Thus, in certain cases, the packet itself may be considered as an increment, whereas, in other cases, the packets themselves must be opened.
3.4.3.1 Example 1
The initial lot is mado up of two boxes oach containing 2.5 units of 10 pockets of 20 cigarottos (boxes of 20 x 10 x 25 = 5 000 cigarette units, i.e. 250 packets, or a total of 10 000 cigarettes, 500 packets, 50 cartons).
If it is wished to make up, for example, 5 samples (as &milar to each other as possible) of 20 packets from the initial lot of 500 packets, constitution of the samples by systematically sampling 2 units of 10 packets to make up each sample will obviously be avoided, since internally homogeneous samples would be obtained, but these would be greater diffeiences betwetsri them than within (hem. The best procedure would be the following:
From each box, take at random 1 unit in 5. Thus 10 units (2 x 5) each of 10 packets will be obtained.
Each unit can be divided into 5 increments each of two packets, und each sumple will ti.wi receive one increment from each unit.
In this case, the 5 samples of 2 x 10 packets will be replicates.
3.4.3.2 Example 2
From the initial lot described in 3.4.3.1. select 10 samples of 50 cigarettes each.
For obvious reasons, it is not recommended in this case that the packet should be kept as the unit of sampling.
Moreover, it is again recalled that the aim in view is not to make up samples representing the original lot but samples which are as similar as possible to aarh nthAr.
Starting from these pnnciples, and with a total number of cigarettes in all samples of 500, or 25 packets, the method to be used is as follows:
a) Select 3 units from one box. Keep two whole units, but rcrnovo half of the third unit to give a total of 25 pockets.
b) Open each packet and distribute the cigarettes into 10 samples (2 cigarettes per sample).
C) In this manner, 10 samples each of 50 cigarettes are constituted and these will be as similar to each other as possible.
3.5 Leaves or strips
This case is called upon only as a last resort, because it poses difficult specific problems and because it is rare that leaves or strips
themGelves will constitute the material on which the collaborative, physical or chemical studies will be carried out.
Similar samples of leaves or strips are usually used only for tests concerning changes of industrial processes. These tests require the use of industrial production machinery (for example mixing boxes) and the sampling procedure is outside the scope of this Technical Report.
However, if the lot consists of leaves or strips, analysis of the actual requirements shows that it is possible, without disadvantage to the method of test, tt mndify the phycir.aI aspect of the material and to use one of the methods described in the above examples.
An example will illustrate this point, as it was an actual case which occurred some years ago.
There were various pi-ocedures to be tested for the determination of the moisture content of tobacco for dill erent varieties of leaves11 end each initial lot to be sampled Was made up of loavos. Detailed analysis of the problem showed that sampling should not be carried Out Ofl the leaves, but after treating them and reducing them to cut tobacco. Collaborative tests were carried out in good conditions with correctly sampled lots of Cut tobacco.