Eduardo J. Magri, Professor; Mining Engineering Department, University of Chile; Av. Tupper 2069; Santiago-Chile; +(56-2) 3344226 / +(56-2) 9784503; +(56-8) 4498971; +(56-2) 2336328 / +(56-2) 9784985; firstname.lastname@example.org.
The author, sometimes in conjunction with Francis Pitard, has carried out a number of heterogeneity tests forexploration or mining operations around the world. The aim of such tests is to establish sampling constants for the various geological units and elements of interest with the following objectives: • • Optimize sampling and sample preparation protocols and recommend appropriate equipment and methodology to carry out these operations. Study the distribution of the various elements in different sizefractions in order to assess possible errors due to grouping and segregation.
This paper presents the results of 57 tests carried out on different types of ore bodies, including several large porphyry copper / gold deposits in Chile, Greece and the Philippines, gold-silver vein type deposits, poly metallic vein deposits, exotic copper deposits, volcanic massive sulﬁde deposits in northernGreece, etc. Results for each test include all variables, for example, for some porphyry copper deposits, the variables of interest are: total copper, soluble copper, molybdenum and arsenic. The following information is presented for each test: • • • • • Brief geological and mineralogical description of the geological units where the tests were carried out. Size fraction where the tests were done andthe average size of fragments, usually tests were done on the –1/2” + ¼” with an average size of fragments of 1.04 cm. Relevant statistics and sampling constant for each variable. Ratio between the maximum and minimum size fraction grade for each variable. Usually the ﬁne fragments (-70 mesh) have much higher grades than the coarse fragments. This can lead to serious segregation related problemsunless the necessary precautions are taken. For some gold deposits, screen ﬁre assays were done for each size fraction. In these cases, the weight percent and the percent of the total gold left above the screen (150# or 170#) are reported.
Some experiences in blast hole sampling in a large porphyry copper and muck sampling in a vein type gold mine are discussed.
Atvery early stages of geological exploration, sampling and sample preparation protocols can be established using Pierre Gy’s “Preferred Nomogram” which was derived using average results for materials showing three different levels of heterogeneity. However, as exploration progresses, it is recommended to apply Gy’s theory to the project at hand and not rely on average results, specially if dealingwith important secondary minerals such as molybdenum, gold or arsenic in a porphyry copper deposit, for example. The fundamental sampling error at any particular stage of a sampling protocol is given by Pierre Gy’s well known formula: S2 FE = C * d3 * (1 / Ms – 1 / ML) where: Ms ML d C mass of the sample at any given stage (g); mass of lot at any given stage (g); maximum diameter of particles (d95in cm); sampling constant.
The sampling constant C is composed of the following factors: C=f*g*c*l where: f g c where: shape factor of the fragments (generally around 0.5); particle size distribution factor (generally 0.25); mineralogical factor = λm (1-aL)2 / aL + λg (1-aL). λm and λg are the densities of the mineral and the gangue respectively and aL is the average grade of the lot (a gradeslightly below cut-off can be chosen to be conservative). liberation factor.
Calculating these f actors directly is known as the parametric approach, its major difﬁculty lies in the estimation of the liberation factor for each stage of a sampling protocol. The liberation factor may vary between 0.0, for perfectly homogenous very course material and 1.0 when the particles of interest are...