When it is necessary to evaluate a new gold deposit, composite construction is a key aspect in the flowsheet design process and it possible that discrepancies or differences between predicted metallurgical results obtained from a testing program and full scale plant performance can be attributed to the use of wrong samples. In this way, the single most difficult task in any testing program is to obtain a representative sample of gold ore for use in metallurgical tests. For example, representative three dimensional samples can be acquired by diamond drilling in a relatively tight pattern encompassing the mineralized zones.
Basically, the drill core intercepts can be classified according to lithological characteristics and different ore type or master composites are formed. It is often assumed that an ore sample from a given lithological type within a gold deposit will produce constant metallurgical response, regardless of the source of the sample. For example, in volcanic deposits, well defined zonations occasioned by remobilization may produce the lithological horizons resulting in marked differences in sulphide mineral crystal size from one are to another. If the crystal contains gold particles, variations in average crystal size between samples produce different metallurgical performance.
It has been noted that in the construction of a testing program composite much importance is accorded by management to approximating the average gold content of the deposit or horizon. Then, in order to ensure conformity with this, composites are prepared with disproportionate amounts of the higher grade intercepts. This problem is created by the tendency of most exploration geologists to increase the drill pattern density in zones of higher grade mineralization. At the moment of comparing to the overall gold deposit, these higher grade intercepts often contain relatively large gold grains and hence impart a bias into the composite construction.
A common error in sample selection is made during core logging. Essentially, differences between lithologies by core logging is the assigned task of the least experienced members of the geologists. It is not unusual to find mineralogically similar material being classified quite differently by two different people, leading to evaluate composite that present accidental mixtures of lithologies. This mistake gives unrealistic results and the creation of any geometallurgical mode is only a waste of time. The only solution available to offset the effects of these problems is to exercise increased diligence in identifying the lithologies within the gold deposit and taking special care in selecting samples to prepare master composites. In order to do this, intercepts from any given lithology from several parts of the ore body must be ground in the metallurgical laboratory and the sized fractions subjected to mineralogical study. In this way, it will possible to obtain precise information on gold mineralization and fragmentation characteristics by location within the ore body and the preparation of master composites is done with realistic data and samples.