20 laboratories took part in the experiment from 9 European countries. The laboratories have been given numerical codes. For the purposes of this report they have also been assigned letter-codes (because single-character codes are needed in the graphs).
Samples of three materials were prepared and distributed by Partner 1. Partner 3 supplied Partner 1 with a bulk sample of the aggregate used for Level 2, and Partner 4 supplied Partner 1 with a bulk sample of the high-PSV aggregate used for Level 3. The three materials were chosen so that they would give a wide range of results in the PSV test.
The material used for Level 2 came from a quarry that had also been used as the source of material for another cross-testing experiment, involving 13 laboratories from Austria, Germany, the Netherlands and Switzerland (Zieger, 1990). In that experiment it had given an average of 55.2 PSV.
The requirement that the test has to be carried out on particles that pass a 10.0 mm test sieve but are retained on a flake sorting sieve with slots of width 7.2 mm causes the rejection of a large proportion, perhaps as much as 85 %, of aggregate as produced. To avoid having to distribute large quantities of materials to laboratories in different countries, only for much of these materials to be wasted, Partner 1 carried out the sieving and flake sorting of the materials prior to preparation of the laboratory samples. Apart from this, the samples were prepared, for each level of the experiment, as if they were laboratory samples all taken from one bulk sample - Partner 1 used fractional shovelling to reduce the bulk samples of the three levels to the laboratory samples. The participants were required to prepare and test duplicate test portions from each sample. Hence the measures of repeatability and reproducibility given by the experiment are consistent with the definitions of r1 and R1 used here.
Figure A, below, shows the sample reduction procedure that the participants were asked to follow, and Table B shows how the specimens were allocated to runs of the standard polishing machine.
|Test result||Run||Level 1||Level 2||Level 3||Control stone||Other specimens|
|1||1||1a 1b||2a 2b||3a 3b||a b||6 other specimens|
|1||2||1c 1d||2c 2d||3c 3d||c d||6 other specimens|
|2||3||1e 1f||2e 2f||3e 3f||e f||6 other specimens|
|2||4||1g 1h||2g 2h||3g 3h||g h||6 other specimens|
Data for individual specimens are recorded in "Data". Uncorrected PSV measurements for Levels 1 to 3 and for the control stone (i.e. values of S and C) are also given there, together with PSV test results for Levels 1 to 3 (i.e. values calculated using PSV = S + 52.5 - C).
The PSV test method requires test results to be rounded to the nearest 1 PSV units. However, for the purpose of the cross-testing experiment, all PSV results were recorded to the nearest 0.1 PSV units. This was to prevent rounding of the data affecting the assessment of the repeatability and reproducibility of the test method. Data from Laboratory 95 arrived too late to be included in the calculations, however, their results are shown in the data tables, and in the histograms.
Laboratory averages are used to calculate the reproducibility of the test method, and to assess laboratory biasses. Between-test-portion ranges are used to calculate the repeatability of the test method, and to assess the repeatability of tests from individual laboratories. The averages and ranges are shown in the histograms, and laboratory averages are plotted in the Mandel plots.
The averages and ranges are also used to test for stragglers and outliers. Where these have been found, they are indicated throughout using a single question mark (?) to indicate a straggler, and a double question mark (??) to indicate an outlier.
Uncorrected PSV measurements for the three levels (i.e. values of S) are shown in Figures C, D and E, below, plotted against PSV measurements for the control stone (i.e. values of C).
These graphs also show the limits of 52.5 ± 3.0 PSV units within which the results of tests on the control stone are required to fall by prEN 1097-8. One laboratory (J) reported results for the control stone outside the permitted range of 52.5 ± 3.0 units.
According to the formula that is used to calculate PSV test results:
PSV = S + 52.5 - C
so the plotted points of S against C for one level of the cross-testing experiment should fall along a line:
S =PSVaverage - 52.5 + C
PSVaverage = the average PSV test result for the level
These lines are also shown in Figures C, D and E. It can be seen in these figures that the points are not uniformly scattered about the solid lines. There is a tendency for laboratories that obtain high results on the control stone to give points above the solid line, and for laboratories that obtain low results on the control stone to give points below the solid line. Thus the results depart from the pattern that would be expected from the correction formula (PSV = S + 52.5 - C). If the reason for this could be found it would allow the reproducibility of the test method to be improved.
Standardised values of the averages and ranges are shown in the Mandel plots. These figures are used to identify laboratories that give rise to large laboratory biasses, or large between-test-portion ranges, in more than one level of an experiment. The horizontal broken lines in these graphs show the critical values of the "h" and "k" statistics at the 5 % and 1 % significance levels, taken from the revised ISO standard on precision (ISO 5725, 1994).
Samples of the emery flour and corn emery used by the participants have been collected together, together with examples of the specimens made using the control stone. These are being retained for further studies.