- Description of the Crushed and Broken Surfaces test method.
- Details of the precision experiment (materials, participants, data obtained).
- Results of the precision experiment (repeatability and reproducibility, functional relationships, graphs).
- Return to Home Page.

This report gives the results of a cross-testing experiment, involving 17 laboratories from 10 European countries, on a method for determining the percentage of crushed and broken surfaces in coarse aggregates. Three coarse aggregates with differing levels of crushed and broken surfaces were used.

Thirteen laboratories reported percentages of crushed particles that incorrectly excluded totally crushed particles from the calculated results.

The reproducibility coefficient of variation of the test is substantially higher than the level of 8 % that is considered to be desirable for a test that is to be used to check compliance with specifications. In most cases encountered in the experiment, the repeatability coefficient of variation also exceeded 8 %.

Using statistical theory, the repeatability of the test is predicted to be:

r_{1} = 0.0028 ( X (100 - X) D^{3} ÷ M )^{½} %

This equation agrees with the experimental results for the determination of the percentages of crushed, rounded, totally crushed, and totally rounded particles. An equation of the same form

R_{1} = 0.0051 ( X (100 - X) D^{3} ÷ M )^{½} %

was fitted to the reproducibility results for the determination of percentages of crushed, rounded, and totally rounded particles. The estimates of the reproducibility of the determinations of the percentage of totally crushed particles did not fit the above equation for R_{1} - they were much too high.

It is recommended that laboratories should take part in proficiency tests. Participants in proficiency tests could be sent examples of sorted test portions to help them learn how to classify particles correctly. A number of ways of improving the test method description are recommended, as follows.

- To prevent other users making the same mistake as made by most of the participants in this experiment, the test method should describe the initial separation of particles into crushed and rounded classes as "
*Stage 1*" of the procedure, and the further separation into totally crushed and totally rounded classes as "*Stage 2*". The calculation section of the method should then describe the calculations required by the two stages separately. - The definitions of the classes should be amplified, and drawings or photographs added to help the user classify particles correctly.
- The equation for repeatability should be used to calculate the test portion masses that are needed to give a satisfactory level of repeatability in the test.
- The user should be given the option of using small test portions with aggregates that easily meet their specifications, or large test portions with aggregates that may give results close to a specification limit.
- For aggregates with D > 2d , the method of BS 812 Part 104 should be used to calculate how much of each size fraction to test, so that a chosen level of repeatability can be achieved from a minimum amount of work.