Issue 5, p. 13 (2015)
Heterogeneity Tests have been very popular for the last 30 years and there are several versions of them such as the method of choice used by François-Bongarçon to quantify and minimize QFE1 which is a combination of the Fundamental Sampling Error and the Grouping, Segregation Error and sometimes Analytical Error. A more recent version called “segregation free analysis for calibrating the constant K and x” is used by Minnitt, and an older, obsolete version using fragments collected one by one at random from several size fractions to calibrate the constant K was used a long time ago by Gy and Pitard. All these methods have their merits and pitfalls. The common pitfall is that they all depend on the collection of a representative composite sample consisting of about half a ton of material. In Mineral Processing it is well known how difficult it is for geologists to provide a representative sample from a given geological unit to perform reliable metallurgical testing; the same difficulties are encountered in performing Heterogeneity Tests. Furthermore, experience clearly shows that for trace constituents such as gold, many tons should be collected to obtain a reliable composite. Perhaps there is a more representative way to collect the information necessary to calculate the variance of the Fundamental Sampling Error FSE, which can support and complement the method of choice referred to earlier. This paper suggests that all the necessary information can be obtained by slightly modifying the logging practices of geologists. From such observations, reliable histograms of the size distribution of particles of the mineral of interest can be made representing the properties of an entire geological unit. Such information can be obtained at an early phase of exploration leading to an unmistakable definition of the sampling constant K, and possibly an accurate definition of the mathematical model of the liberation factor leading to the constant x; using modern microscopy the mineralogist can define the evolution of the liberation factor as a function of increasing comminution better than anyone else. Furthermore, this paper suggests that the determination of the liberation factor is no longer a critical factor, though most certainly useful, if using the information from modified logging practices and two old formulas suggested by Gy in the 50’s instead of his famous formula using the liberation factor.