A geostatistical study of diamond-saw sampling versus chip sampling

This paper describes a geostatistical study of the relative merits of diamond-saw sampling and conventional chip sampling, using results from experimental work carried out on Randfontein Estates Gold Mine and Western Areas Gold Mine. The diamond-saw sampling procedure is described, and it is concluded that this is a very efficient operation and leads to good-quality sampling that is clearly an improvement over conventional chip sampling. The study principally involved the computation of the errors of estimation of the mea gold value of mining panels for the two sampling methods, revealing that, in the normal production environment, the diamond-saw sampling method is a significant improvement over conventional chip sampling, as evidenced by the reduced estimation errors achieved. The question of optimum sample spacing is discussed, and it is concluded that the required spacing for chip sampling needs to be smaller than for the diamond-saw method.