What is the best method of splitting a sample from a reverse circulation drill rig? How important is it to obtain a correctly split sample? What is the impact of an incorrectly split sample? These are just a few of many questions asked daily in the mining industry with regard to sampling._x000D_
Modern sampling theory is well described in the literature in texts and is taught in short courses throughout the world, however little has been published on the relative performance of sample splitting devices used in the field. This paper compares the performance of the Jones tiered riffle splitter with a cone splitter manufactured by SGS Metal Craft in Western Australia._x000D_
The testwork was carried out at Anglogold's Sunrise Dam Gold Mine, 55 km south of Laverton in Western Australia. Mineralisation at Sunrise Dam is hosted in a structurally complex framework of shear zones with dips that range from steep (70) to shallow (25). The deposit is characterised by significant amounts of coarse gold, which presents a challenge in sampling, analysis and grade estimation._x000D_
The three major areas where sampling errors can occur are in drilling, sampling, sample preparation and analysis. Sampling errors attributable to drilling and sample preparation often far outweigh the errors made at other stages in the process, however historically most effort to identify sampling errors has concentrated on the laboratory and analytical procedures. Splitter design and cleanliness can have a large impact on sample collection and preparation. Splitter performance was tested with respect to sample weights, particle size distribution, and gold grade consistency. The aim was to determine if each particle has an equal probability of being sampled, the resulting sample has integrity, and if any segregation problems occurred. The cone splitter was found to provide the best split in terms of particle size distribution, with no apparent size bias. Duplicate samples had more comparable masses than those provided by the riffle splitter. Most importantly, the cone splitter gave less grade variability between duplicate samples than the riffle splitter. Testwork also confirmed that, like the riffle splitter, the cone splitter must be oriented vertically and the design of the rig or cone splitter trailer should accommodate this requirement.