Page 221 - Introduction to Mineral Exploration

204 M.K.G. WHATELEY & B. SCOTT

2
Global estimation error [S (GE)] ness. Following from (2) and (3) an optimum
sample is provided by diamond drilling which,
Global error (GE) includes all errors in sample with 100% core recovery, cuts almost perfect
extraction and preparation (total sampling cylindrical lengths of rock the full length of the
error or TE), and analysis errors (AE).
sampled area.
4 A correct preparation of the assay portion
2
2
2
S (GE) = S (TE) + S (AE)
from the original sample, as described later.
Selection errors are of two main categories,
The TE includes selecting the assay portion which arise from:
for analysis while the analytical error proper 4.1 the inherent heterogeneity of the sample
excludes this step. Gy (1992) maintains that and
analysts usually complete their work in the 4.2 the selective nature of the sample extrac-
laboratory with great accuracy and precision. tion and its subsequent reduction in mass and
In this strict sense, associated analytical error size to the assay portion.
is negligible when compared with the variance The first category comprises the funda-
of taking the assay portion. Consequently, for mental error [S (FE)] and is the irreducible
2
2
practical purposes S (AE) ≈ 0 and from this minimum of the total sampling error and
2
2
S (GE) ≈ S (TE). can be estimated from the sampling model of
Gy (1992) (section 10.1.4). Gy (1992) separates
2
Total sampling error [S (TE)] those of the second category into seven differ-
ent types, which are estimated from measure-
The total sampling error (TE) includes all ments (i.e. mass and grain size) from each stage
extraction and preparation errors. It includes of a sample reduction system.
errors of selection (SE) and preparation (PE): Preparation errors [S (PE)] are related to
2
the former are inherent in both extraction and processes such as weighing, drying, crushing,
preparation while the latter are restricted to grinding, whose purpose is to bring successive
mechanical processes used in sample reduction samples into the form required by the next
such as crushing, etc.: selection stage, and ultimate analysis. Sources
of possible error are:
2
2
2
S (TE) = S (SE) + S (PE) 1 Alteration of the sample’s chemical com-
position by overheating during drying as, for
Selection errors (SE) are minimized by: example, with coal and sulfides of mercury,
1 A correct definition of the number of point arsenic, antimony, and bismuth.
increments, or samples. Sample spacing is best 2 Alteration of the sample’s physical condi-
defined by geostatistics, described in section tion. If the grain size of the sample is important
10.4. this can be changed by drying and careless
2 A correct definition of the area sampled. handling.
Sampling of mineral deposits presents a three- 3 Once a sample is collected, precautions
dimensional problem but for practical pur- must be taken to avoid losses during prepara-
poses they are sampled as two-dimensional tion of the assay portion. Losing material
objects which equate to the surface into which always introduces error (i.e. increases the vari-
samples are cut. The long axis of the sample ance) because the various sized fractions differ
preferably should be perpendicular to the dip in grade and it is usually the finer grained
of the mineralisation, or at least at an angle to material which is lost. Losses can be checked
it, but not parallel. The cut in plan is ideally a by weighing samples and rejects at each reduc-
circle, sometimes a square or rectangle, and in tion stage.
length should penetrate the full length of the 4 Sample contamination must be avoided:
sampled area. sample containers and preparation circuits
3 A correct extraction and collection of the must be clean and free from foreign material.
material delimited by the above cut. Usually Dust produced by the size reduction of
this presents a problem with the planes of other samples should be excluded. Accidental
weakness in rocks, and their variation in hard- sources of added material include steel chips