The South African Institute of Mining and Metallurgy

DMS and Gravity Concentration Operations and Technology in South Africa

R A Heins, P M Grady and R L Langa

MEASUREMENT SYSTEMS FOR GRAVITY CIRCUIT

PERFORMANCE: A NEW APPROACH.

R A Heins
Gekko Systems (SA)

P M Grady
Gekko Systems (Canada)

R L Langa
Gekko Systems (SA)

ABSTRACT:

The use of differences in specific gravity between minerals to separate them has long
been utilized in the extractive industries. The environmental and cost benefits of the
commercialized forms of these processes are well understood and widely used.

Gravity concentration is used inter alia in the primary beneficiation of gold, diamonds,
coal, tin, ferrous metal ores and andalusite. The potential for economic loss due to poor
performance of these gravity processes is significant, but diverse in manifestation and
regular process monitoring can prevent or at least reduce these losses.

The measurement of performance of gravity processes has always presented a challenge,

due primarily to the masses of material involved and the physical nature of the processes,
generally typified by long reporting time, and hazardous materials (such as
Tetrabromoethane) for float and sink type tests.

The development of density tracers has improved this situation, however recovery of
these tracers from the process streams created the next set of challenges, requiring
significant manpower to physically remove tracers from product dewatering screens.
Additions to these tracers (magnetic or x-ray fluorescent) eased the recovery mechanism
issues but made them much more expensive to produce, hence operators became
concerned about losses during process testing. In addition, magnetic recovery of smaller
size tracers by a magnet suspended above a loaded screen panel is still problematic.

The development of new low cost magnetic tracers has improved statistical significance
of tracer tests, as the tracers themselves can be bought and used in much larger quantities,
whilst still remaining economically viable. Coupled with the development of effective
magnetic recovery technology, such as the Gekko MagScreen, for even fine sized tracers,
on-line, continuous tracer testing is now becoming a reality.

205
The South African Institute of Mining and Metallurgy
DMS and Gravity Concentration Operations and Technology in South Africa
R A Heins, P M Grady and R L Langa

Whilst the development of high technology radio frequency detection tracer systems
continues, enabling the real time measurement of the process and interpretation of results,
miniaturization and recovery of these tracers, which is critical given their costs at present,
remains an issue.

Keywords: Gravity, process measurement, tracers, magnetic recovery

INTRODUCTION

DMS and Gravity Processes

The use of gravity (or differential particle settling rates in a medium) has been exploited
as a means of recovering or upgrading valuable minerals for thousands of years, with the
process occurring even in nature, typical examples being placer gold deposits, alluvial
diamond deposits and the heavy mineral rich dunes of the KwaZulu-Natal North Coast
of South Africa.
The process utilizes the density difference between two (or more) mineralogical
components of an orebody and hence their different settling velocities in a medium of
given density. This may be practically illustrated by considering the rate of settling of,
for instance, 5mm lead beads (density 11.35) in water compared with the same size of
aluminium beads (density 2.71).
The ease of separation is dictated by the degree of liberation of the minerals to be
separated, (hence how close they are to their theoretical specific gravity) and how close
their specific gravities are to both each other and to the medium in which they are to be
separated.
This measure is defined as the concentration coefficient (cc) and for a binary mineral
system is given by the following relationship:
Concentration coefficient (cc) = H - M after
Aplan [1]
L - M
Where H is the specific gravity of the dense or heavy component, M is the specific
gravity of the fluid medium in which the mineral mix is suspended, and L is the specific
gravity of the light component.
For the case of alluvial gold recovery in a water medium, H is 19.3, L is 2.7 (for a silica
sand type gangue material), and M is 1.0 for a water medium, giving a coefficient of
10.76. As the density of the media is increased, so the value of the denominator is
reduced, thus increasing the concentration coefficient. Under these conditions, the ease of
separation not only increases, but the sharpness of the separation is also improved.
Gravity processes are widely used due to their environmentally friendly and benign
nature. In addition, they are generally significantly cheaper to install and operate than
their hydro or pyrometallurgical alternatives.

206
The South African Institute of Mining and Metallurgy
DMS and Gravity Concentration Operations and Technology in South Africa
R A Heins, P M Grady and R L Langa
Typical applications of gravity processes are as follows:
a) Coal Processing Plants:
The coal industry has long utilised many forms of gravity concentration as a means to
reject low value non-combustible inorganic components (generally shales) which are
co-mined with the coal and reduce the calorific and hence financial values of the
resulting coal product. This industry utilises spirals and jigs (water medium) and also
uses dense media extensively as a means to control product ash content.
b) Diamond Recovery:
The use of gravity processes is well established in the diamond industry, with dense
media separation forming the main recovery mechanism to reduce non-valuable bulk
to a minimum ahead of specialist low tonnage recovery processes such as grease or x-
ray sorting. The use of jigs is finding more acceptance as a more economic way of
reducing the bulk fed to the dense media plant with commensurate benefits in terms
of both capital and operating cost.
c) Ferrous Metal Ores:
The use of dense media technology to reject light low value gangue components of
the ore is common practice. This has the effect of increasing feed grade and reducing
bulk to downstream smelting processes, examples of this may be found in the iron,
manganese and chrome beneficiation industries.

Why measure?
In all forms of gravity concentration, the efficiency of the separation of the minerals from
one another is important, but for different reasons depending on the industry.
Measurement of separation efficiency is critical to diamond processors to ensure that
immensely valuable diamonds are not discarded through process inefficiency. However,
for coal, measurement is crucial to ensure the finished product grade and hence value is
not eroded. In ferrous metal ore processing, gravity inefficiency will lead to an improved
recovery of valuable minerals but at a reduced grade, thus adversely affecting either sale
value or downstream processes such as smelting.
All gravity processes have an efficiency of separation (or partition), which will be
examined in more detail later.
In order to effectively manage a unit process such as a gravity or dms plant, it is
necessary to measure its operation and efficiency. Measurement provides feedback to
process operators and managers, allowing control of the process and reaction to changes
such as variations in the quality of an orebody.
This process control, in turn, prevents or minimizes value losses and improves the
economics of the process by finished product control or operating cost optimization.
The old adage of If you dont measure it, you cant manage it could have been written
with this application in mind!

Forms of Measurement
There are two crucial measurements in the control of a gravity process, namely the
separation density or cut-point, and the error of the separation or the amount of material
misplaced to either the floats or sinks streams.

207
The South African Institute of Mining and Metallurgy
DMS and Gravity Concentration Operations and Technology in South Africa
R A Heins, P M Grady and R L Langa

There are currently three main mechanisms for the determination of the efficiency of a
gravity process as follows:
1) Float and Sink Analysis
Samples are taken from the floats and sinks streams from the gravity process. These
samples are then immersed in a heavy liquid (normally organic) which has a density
equivalent, or close to the theoretical density of separation of the process. Material
above the density of the liquid in both streams will sink, while lighter material will
float, by varying the density of the liquid the amounts of misplaced material in both
streams may be established. This process is extremely accurate, but is time
consuming, with reporting times after sampling being measured in weeks. In addition,
the liquids used are generally extremely environmentally unfriendly and in some
cases extremely hazardous (such as perchlorethylene (1.6sg) and dibromomethane
(2.48 sg) for coal and lead sulphamate for diamonds)
2) Fractional Density Analysis
This method of analysis will generate a partition curve for the process (from which
can be derived the error and also the cut point), but is tedious, time consuming and
labour intensive. It entails subjecting each of the product streams to separations at a
range of different densities above and below
the theoretical separation density and
capturing floats and sinks at each density.
This process is carried out in either a Multi-
Gravity Separator or Ericsson Dense Media
Cone type such as the Gekko Viking Cone
(figure 1) apparatus.
The floats and sinks fractions at each density
are then dried and each individual particle
checked for its individual density using a
gravimetric flask. Figure 1: Gekko Viking DMS Cone

3) Tracer Testing
Tracers are particles of known density, size and
shape, which may be introduced into a separation
process, their passage and final disposition in the
process allowing calculation of both process cut point
and also the error in the process (see figure 2). They
can also be used to calculate the size recovery
efficiency of a process for particles of a known
density. This method is used in the diamond industry
where tracers simulating the density of diamonds are
Figure 2: Example of density tracers used to assess size-by-size recovery efficiency of the
process. The drawback to tracers is that their
disposition needs to be measured, which is generally done by collection of the tracers
from floats and sinks streams and then manual reconciliation. There have been a
number of methods of tracer recovery developed, and it is the future of these systems
on which this paper will focus.

208
The South African Institute of Mining and Metallurgy
DMS and Gravity Concentration Operations and Technology in South Africa
R A Heins, P M Grady and R L Langa

Theory of Measurement
The performance of a specific gravity concentration device is measured using a Tromp or
Partition curve, which depicts the percentage of product reporting to sinks at different
particle densities.
The most practical application of this involves the feeding of a known quantity of tracers
of known sizes, shapes and densities, and measuring the recovery of these particles to
floats and sinks products.
A typical partition
curve is illustrated in
Fig 3. It is obvious that,
for a perfect split at a
given density, the cut
point line should be
vertical. However,
production processes
are not ideal and as
such a measure of this
imperfection is given
by the EPM or Ep (ecart
probable moyen),
which is
Figure 3: Typical Partition or Tromp Curve
the error or misplaced

material present in each of the process streams.

The EPM is defined as half of the specific gravity difference between the 25% and 75%
cut points on the partition curve, hence the lower the EPM value obtained, the closer to
vertical the separation line and the more accurate the density cut.

APPLICATIONS OF MEASUREMENT SYSTEMS

Coal Industry
Coal, an organic-rich sedimentary rock, is different from most minerals, in that the
valuable component occurs in large quantities (particularly when compared with hard
rock mineral recovery systems). In addition the low value component (in this case the
inorganic minerals) are generally widely distributed but as discrete components. This
allows coal to be beneficiated at a much coarser size than most minerals and using
primarily gravity processes (typically 90% of the recovery process utilises gravity) [2].
The Run-of-Mine coal is generally crushed and sized, and selected size ranges are
subjected to different gravity processes such as jigs, spirals and dense media plants.
Inefficient operation of these plants leads to value loss for the operating company through
reduced value of the washed coal product (ash content increases due to inefficient
removal of mineral impurities). Value loss also results if the coal is washed too clean,
since saleable coal will report to the gangue/discard stream from the gravity circuit. It is
thus evident that operational control and management of coal washing plants is critical.

209
The South African Institute of Mining and Metallurgy
DMS and Gravity Concentration Operations and Technology in South Africa
R A Heins, P M Grady and R L Langa

CURRENT TESTING MECHANISMS:

The commonly accepted method for
Springlake Colliery : Norwalt Plant Feed - Underground
testing gravity process efficiency in the
1.40
coal industry is the standard float and sink
35

analysis, which involves taking significant

1.45 quantities of representative sample from
30

both float and sink screens (typically

1.50 500kg of each for a Wemco Drum type
25

R
DMS plant) [3].
e
l
1.55
These samples are then manually 20

D
e
subjected to a float and sink analysis Ash
n
s
using organic heavy liquids (such as %

I 1.60 15
t
y
tetrabromoethane diluted with benzene),
modified to cut at a range of closely
1.65
spaced densities. This technique measures
10

the amount of material above and below

1.70 density for a range of density fractions.
5

The float and sink products from each

1.75 range are then subjected to determination
0
0 10 20 30 40 50 60 70 80 90 100

Yield %
of the so-called ash content. The fractional
Rd / Yield Ash / Yield
analyses are used to calculate a
Figure 4: Typical Coal Plant Float Sink results reconstituted feed analysis. The
percentage in each fraction reporting to the clean coal component is then expressed as a
percentage of the total coal in each density fraction of the whole.[4]. Figure 4 presents a
washability curve for anthracite produced from the underground orebody at Springlake
Colliery in Kwa-Zulu/Natal[5]
This technique can be used for process control purposes. However, the turn-around-time
for the results usually reduces its effectiveness for plant monitoring purposes, and thus is
more often used as a routine check, or for specification acceptance purposes on new
plants.
The results of these tests can also be entered into a simulation programme for the actual
coal being washed, and the day-to-day results of the operation of the washing plant
compared against the model as an empirical means of checking the process operation.
An alternative is the use of tracers as a supplement to the float and sink analysis. This has
the benefit of not requiring chemical analysis to obtain results, but is still very labour
intensive and time consuming.
A known quantity of tracers of a predetermined range of densities is introduced, to allow
collection from the product streams and hence the construction of a partition curve based
on tracer disposition. There is some debate about the number of tracers required for a
significant test, however consensus is that ten to twelve density ranges is sufficient with
the increment between each range being determined by the sharpness of the split [6]. The
recommended number of tracers per density increment varies between 30 and 100
according to authors [4,6,7,8] to obtain a statistically meaningful result
The tracers report to either float or sink streams and are generally collected by hand from
the product flowing across the screens. This process is relatively straightforward for large

210
The South African Institute of Mining and Metallurgy
DMS and Gravity Concentration Operations and Technology in South Africa
R A Heins, P M Grady and R L Langa

size tracers (32mm tracers are most commonly used for cobble and nut coal fractions) as
the tracers float to the surface of the stream on top of the screen, allowing for easy and
accurate recovery. However, for smaller fractions, recovery of the tracers becomes more
difficult, as they tend to remain in the mass of the product, and also their physical size
makes them more difficult to both see and recover.
Tracers have more recently been manufactured in different colours per density, both to
ease recovery and also to identify density fractions. This is also useful in establishing
residence time and hang-up through the process.
The recovery process for tracer testing has been the subject of much research, with
development of colour sensing technology to read tracer recovery, and also magnetic
flux technology whereby a magnetic sensor registers the flux from an implant within each
tracer as it passes. Any technology such as this implies that the tracers are lost, as they
report to the product streams and are not recovered.

LIMITATIONS TO CURRENT TECHNOLOGY:

Float and sink analysis is a slow and laborious procedure requiring much manual labour.
The results of a single analysis are often only available to plant management weeks after
the initial sampling exercise was performed, as a result, its value as an on-line process
control measurement is limited.
There are also a number of hazards associated with this analysis, manhandling of large
bulk samples, and, as previously mentioned, the hazardous nature of the liquids used for
separation (TBE and Benzene are both known carcinogens)
Tracer testing has a quicker turnaround time than float and sink analysis, but is still
labour intensive in the recovery and counting of the various tracers. In addition if tracers
are not recovered as a result of being missed by the collector, the statistical accuracy of
the test is compromised, and hence its absolute value is diminished.
The other systems discussed, such as magnetic flux and colour recognition have not to
date been commercialised, however the technology associated with these systems will, in
all likelihood, be expensive and may be impractical for the coal industry.

Diamond Industry

Diamond bearing ores are different to most other ore types in that the valuable
component, namely diamonds, are generally extremely high in value and discrete in their
occurrence. Diamond feeds are typically measured in carats per hundred tons (cpht), for
example, a typical diamond head grade of 50 cpht would correspond to 10 g of diamonds
per 100 tons of ore, or 0.1 ppm. [9] .
Due to the high value associated with each individual diamond in an ore stream,
establishing recovery efficiency is essential to maximizing revenue potential for a given
deposit. Diamond ore deposits typically fall into the categories of either kimberlitic or
alluvial (diamonds liberated over time from host rock and occur as discrete particles in a
gravel mix), and the characteristics of either type can vary greatly within a given ore
body. Variables such as grade, size and mineralogical distribution, ore hardness, and clay
content can each vary greatly over the life of an ore body. The effects of these variations

211
The South African Institute of Mining and Metallurgy
DMS and Gravity Concentration Operations and Technology in South Africa
R A Heins, P M Grady and R L Langa
on the diamond recovery process must be well understood to ensure optimum process
efficiencies.
A conventional diamond flowsheet consists of three sections: feed preparation, primary
concentration, and recovery. The feed preparation stage serves several purposes;
liberation of the diamonds from the gangue; crushing, screening and washing the feed
product into a size range suitable for downstream primary concentration, and finally to
remove screened material below the plant bottom cut size to tailings. This size has
historically been 1.4mm for the diamond industry as until recently, diamonds below this
size had no economic value.

After the material has been prepared, it is then fed to the primary concentration section.
Gravity separation is the most commonly used method of primary concentration, with
dense media separation in the form of cyclones (DMS), jigs, and diamond pans being the
most common equipment types employed in diamond flowsheets.

CURRENT TESTING MECHANISMS:

Similar to the coal and iron ore industries, diamond operations use methods such as
heavy liquids and tracer testing to monitor the performance of their gravity circuits.
Given the significantly higher density of separation for diamonds when compared with
coal, heavy liquid testing requires even more exotic and hazardous liquids, in the form of
lead sulphamate.
In some cases, large bulk samples of the DMS or jig tailings streams are procured, and
processed through separate DMS/Recovery Plants. While this approach does provide the
extractive metallurgist with statistically meaningful data, it is costly and time consuming,
meaning that audits are usually conducted on a non-routine basis.
Diamond spikes are another method used to monitor the performance of gravity circuits
in the diamond industry. This method relies on the use of adding a known amount of
marked, seeded diamonds to a gravity process, and recovering these diamonds in the final
sorting process. The diamonds are often painted or coloured in some way, alternatively
boart stones (diamonds of no commercial value) may be used. It is critical however that
the marked diamonds used can be recovered and their properties match those of the
normal stones. The obvious advantage of this method is that real diamonds are used
during feed on situations, the results obtained have much more credibility than tracer
tests performed without feed, in order to minimise risk of losing tracers. However,
downstream efficiencies must be well understood to account for potential diamond losses
that could occur in these processes prior to the final diamond hand sorting process. For
example, a concentration process using gravity alone could recover diamonds that due to
sorthouse inefficiency are not recovered, thereby creating a grey area as to which
process did not recover the diamonds
In recent joint trials between Gekko Systems and Namdeb Diamond Corporation, the
Gekko Viking Cone was used to evaluate the performance of the In-Line Pressure Jig
(IPJ). For these tests, samples of the floats and sinks products from the IPJ were
procured, and analysed using the Gekko Viking Cone pictured previously in Figure 1.

212
The South African Institute of Mining and Metallurgy
DMS and Gravity Concentration Operations and Technology in South Africa
R A Heins, P M Grady and R L Langa
The Gekko Viking Cone, is similar to a DMS cyclone in that it uses ferrosilicon as a
medium to manipulate cut-point. A variable speed pump is used to circulate the gravel /
ferrosilicon medium in a closed system. Floats and sinks products are recovered in
separate product baskets, sized at 2 mm. For the test work, the Viking Cone was
calibrated to recover 100% of + 3.2 s.g. material to sinks. Floats and sinks samples from
the IPJ were then separately processed through the Viking Cone to determine the
percentage of + 3.2 s.g. material present in each fraction. Results of the test work show
promise for the technology, as the Viking Cone was able to generate points for a partition
curve, without the environmental drawbacks of heavy liquids. Turn around time is
similar to that for heavy liquids.

EXAMPLES & TYPICAL RESULTS:

During the past year, Gekko Systems have

conducted several test programs to evaluate
their IPJ in various diamond applications.
As is the case for any new technology, the
effects of individual operating parameters
must be well understood and demonstrated
before acceptance can be gained within the
industry. Although the main purpose of
these trials was to gain a better
understanding of the IPJ technology, a side
benefit was the experience gained in the
design and use of tracer recovery systems.
For each of the trials, magnetic tracers of
various sizes and densities were used to Figure 5: Example of Gekko Manual Tracer Recovery Magnet
generate partition curves for the IPJ. High
intensity rare earth magnets were used to recover the magnetic tracers from the
concentrate and tailings streams, under feed on conditions. An example of such a system
is depicted in figure 5 above. This particular installation was designed to allow for both
horizontal and vertical movement of the magnet to optimize tracer reconciliation. Tracer
reconciliation averaged 99% for 12 mm tracers, 95% for 8 mm tracers, and 90% for 4
mm tracers.
Tracer testing in the diamond industry generally is performed using only 4mm and 8mm
tracers at sgs ranging from 2.50 to 3.53 (diamond simulant). Sampling plants are
generally tested daily due to the high profile nature of the application, however operating
plants are more likely to be tested either weekly or monthly[10]. An example of the
results of a tracer test from a diamond plant dense media plant audit is presented as figure
6.

213
The South African Institute of Mining and Metallurgy
DMS and Gravity Concentration Operations and Technology in South Africa
R A Heins, P M Grady and R L Langa
4mm 8mm
Densit Tracer Floats Sinks Cum Floats Sinks Cum
y Colour
g/cm3 No. No. % No. No. %
2.50 Dark Pink 48 0 0.00 50 0 0.00
2.70 Black 42 0 0.00 50 0 0.00
2.80 Lime Green 50 0 0.00 50 0 0.00
2.90 Yellow 47 0 0.00 50 0 0.00
2.95 Lilac 47 0 0.00 50 0 0.00
3.00 Blue 46 0 0.00 50 0 0.00
3.05 Buff 44 5 10.20 48 2 4.00
3.10 Green 25 24 48.98 37 13 26.00
3.15 Olive Green 2 47 95.92 6 44 88.00
3.20 Cinnamon 0 50 100.00 0 50 100.00
3.30 Brown 0 50 100.00 0 50 100.00
3.53 Orange 0 50 100.00 0 50 100.00
Figure 6: Example of Diamond Plant Tracer Test Results

LIMITATIONS TO CURRENT METHODS

The turn around time associated with densimetric (or float/sink) type analyses remains an
issue for diamond operators, particularly given the potential for loss with a non-
performing plant.
The cost of diamond tracers (magnetic or x-ray fluorescent) and the resultant tendency to
perform tests in feed off environments remains the major limitation to the accuracy and
value of diamond tracer testing

Ferrous Metal Ore Industry

The run-of-mine ore in the ferrous metal industry generally contains significant
proportions of non-valuable waste rock such as granites, dolomites or limestones. After
crushing to obtain some degree of liberation, it is generally possible to separate much of
this waste from the valuable ore components by means of gravity separation prior to
downstream beneficiation.
In many cases, the next step in the process for ferrous ores is either commercial sale or
pyrometallurgical processing (such as roasting or smelting). The inclusion of excess
gangue material is obviously undesirable in either of these options.
Gravity processing is particularly applicable to ferrous ore beneficiation due to the
density difference of the valuable mineral (in the range 4.3 to 5.1 for manganese, chrome
and iron ores) against the waste (2.5-2.7 quartz/calcite/dolomite) components.
The specific processes favoured in this industry are conventional dense media separation
(using cyclones and drums), spirals and jigging.

214
The South African Institute of Mining and Metallurgy
DMS and Gravity Concentration Operations and Technology in South Africa
R A Heins, P M Grady and R L Langa
CURRENT TESTING MECHANISMS
Due to the density of separation for iron and ferrous ores (greater than for diamond
processing and typically around an sg of 4.0), float/sink testing using heavy liquids is not
practical. As a result, dense media separators such as the Mintek Multi Gravity Separator
and Ericsson cone are required.
By changing the density of the suspension through a range of densities, the floats and
sinks at each density can be established. However the final assessment of misplaced
material at each density is still a manual process.

Tracers have also been used in the ferrous industry, although recovery remains manual
due to the generally magnetic nature of most of the ores.

LIMITATIONS TO CURRENT METHODS

The non-availability of heavy enough liquids to perform a float/sink analysis means that
synthetic suspensions must be made up generally using Ferro-silicon. It then becomes
necessary to have equipment which circulates and maintains this suspension, not allowing
the medium to settle and hence change the density. As a result of this, the equipment
required and hence difficulty and manpower implications of performing these tests once
again become restrictive.

Tracer testing has been performed and is effective in larger size fractions for the same
reasons as in coal processing, however magnetic recovery of tracers is often not an option
due to the magnetic properties of the ore. Thus tracer recovery remains manual.

WHAT ARE THE ALTERNATIVES?

Low Cost Tracers, Batch Systems

Tracers are generally plastic or polymeric particles with materials added to them to give
them physical properties allowing them to both simulate valuable ore components and
also be recovered relatively easily. These properties
include density, magnetic susceptibility, fluorescence
under specific X-Ray illumination and colour.
Tracers were commercialised and have been used in
diamond operations since the early 1970s, with the
coal industry following in the early 1980s [6], are
currently available from 1mm to 64mm and sgs from
1.24 to 4.50 and generally cubic in shape. Recent
technology however is increasing these envelopes for
Figure 7: Examples of Gekko Low Cost size, sg and also shape at reduced costs and an example
Tracers is presented in figure 7.
Tracer testing usually involves the addition of a known
quantity of same sized tracers at known density intervals, with the density range centred
at the estimated separation density of the unit process. According to suppliers of
conventional tracers, approximately 10 density fractions containing 40-60 tracers in each

215
The South African Institute of Mining and Metallurgy
DMS and Gravity Concentration Operations and Technology in South Africa
R A Heins, P M Grady and R L Langa
is sufficient to develop a partition curve for the process and hence establish its
operational efficiency.
After the tracers are added to the process feed, it is necessary to recover the tracers. The

MANUAL
TRACER
FEED

MAGNET
FEED BIN
FLOATS SCREEN

MOTIVE WATER
PUMP

JET PUMP

GRAVITY PROCESS MAGNET

SINKS SCREEN

SINKS
BIN

Figure 8: Manual Tracer Test Flow Diagram

recovery of the tracers is generally done manually from the lip of the floats and sinks
screens downstream of the process being tested.
The recent addition of ferromagnetic powders to the tracer polymer mix has meant that
the recovery system can be automated to some extent by the installation of a plate magnet
above the floats and sinks dewatering screens. This process flow is depicted in figure 8.
This has increased the complexity of production of the tracers and hence the cost. The
effectiveness of this magnet at recovering tracers (particularly in the smaller size ranges)
is restricted by the depth of the bed on the product screen during the test, which increases
the risk of losing tracers. As a result of this and the increased cost, operators have tended
to perform tracer tests, particularly for dms plants, without feed to the plant. This
obviously produces a partition curve under ideal conditions, without the interparticle
interferences and viscosity effects which occur under feed conditions. Thus, the value of
the data generated by the test is limited.
The risk of loss of the tracers becomes even more important in the diamond industry
where compounds are added to the tracers which make them fluoresce at the same
frequency as diamonds under X-Ray illumination, thus allowing process efficiency
measurement for sorting plants. This addition to the tracers makes them even more
expensive, and operators even less willing to risk losing them.

Low Cost Tracers, Continuous Systems

The development of extremely low cost magnetically susceptible tracers has opened up
the potential for real time statistically significant plant process control.

216
The South African Institute of Mining and Metallurgy
DMS and Gravity Concentration Operations and Technology in South Africa
R A Heins, P M Grady and R L Langa
The process flow is depicted in figure 9.

TRACER
FEED BIN

FEED BIN FLOATS SCREEN

MOTIVE WATER
TRACER DETECTION
PUMP

JET PUMP GEKKO

MAGNET
GRAVITY PROCESS
SINKS SCREEN

TRACER
DETECTION SINKS
BIN
TRACER
RECOVERY
BIN

Figure 9: Continuous Tracer Test Flow Diagram

A hopper containing a known mix of size and density tracers sits above the process feed
bin, and on a preset cycle batches these tracers into the process feed. Due to the low cost
of these tracers when compared with conventional hand cut tracers, the number per test
will be of the order of hundreds or thousands per density fraction.
These tracers will travel through the process and will report to either floats (lights) or
sinks (heavies) dependant on their density and the process efficiency.
The floats and sinks streams containing the tracers will feed over a drum magnet ahead of
the product dewatering screen. The drum contains an alternating pole tapered high
strength Neodymium Iron Boron rare earth magnet configured to extract even the
smallest tracers from the stream, and then discharge them into a separate tracer collection
bin.
Once the tracers have been recovered from the process stream, they need to be counted
and their size/density recorded prior to recombination ahead of recycle to the tracer feed
bin.
The technology to identify the tracers reporting to each stream still needs to be
commercialised for full continuous on line measurement, but could use a combination of
physical sizing and optical methods (such as small optical sorters etc). These techniques
would allow instant assessment of the process partition and feedback to the process
control system, whereby action could be taken in real time to rectify process
inefficiencies.
The technology is currently available to recover large quantities of low cost tracers in the
form of the Gekko MagScreen (figure 10). This will produce two separate hoppers of
tracers, one originating from the floats stream and one from the sinks stream. The tracers
will still need to be reconciled and counted manually and an assessment of process
efficiency made from this exercise, however the risk in terms of poor manual tracer
recovery and the statistical accuracy of the test are both addressed.

217
The South African Institute of Mining and Metallurgy
DMS and Gravity Concentration Operations and Technology in South Africa
R A Heins, P M Grady and R L Langa

The continuous tracer system has a number of

benefits, specifically in terms of the number of
tracers which are used, the fact that minimal labour
is required and thus tests are easier and hence more
likely to be routinely performed.
As the tracers are low cost, large numbers of tracers
are used, as a result of which losses due to hangup
or non recovery become statistically much less
significant, for instance if 2 tracers from 2000 are
lost, this is much less significant than losing 2 from
50. A potential disadvantage is the cost of the
system, particularly around the fully automated
tracer identification systems, however this cost
needs to be viewed over the whole life of the plant
Figure 10: Gekko MagScreen
and offset against the potential for losses by running
an out of specification process for two or three weeks whilst waiting for a float and sink
analysis or worse still never doing meaningful process checks.

Radio Frequency Tracers, Continuous Systems

The CSIR in South Africa is currently in the process of negotiating to commercialise a
system of radio frequency identification (RFID or
Supertracers) tracer technology.
This involves implanting a transponder chip into each
individual tracer without affecting either its size or density,
this transponder is tagged with the data specific to each
tracer such as its size, density and any other information
relevant to the process.
These tracers are introduced into the process, and an antenna
is placed either around product discharge pipes or above
product screens, the emission from the antenna powers the
transponder as it passes, causing it to transmit its unique

Figure 11: Examples of Radio identification and data to the receiver. This collates the
Frequency Tracers information relating to which tracers have passed which
antenna (floats or sinks) and passes it to a computer which
interprets the information, deriving a partition curve in real time. Thus process control of
the gravity process can finally be automated based on real measurements of process
performance.
Currently these tracers are recovered manually, they are not disposable due to the cost of
the chip and manufacturing process, and cannot be made to utilise existing recovery
systems as magnetic or x-ray fluorescent additives interfere with the radio transmission.
Once commercialised the intention is that the tracer is cheap enough to not need to be
recovered from the process.

218
The South African Institute of Mining and Metallurgy
DMS and Gravity Concentration Operations and Technology in South Africa
R A Heins, P M Grady and R L Langa

The process flowsheet for this measurement system is presented as figure 12.

TRACER
FEED BIN

FEED BIN
FLOATS SCREEN

MOTIVE WATER
PUMP

JET PUMP

GRAVITY PROCESS
SINKS SCREEN

RADIO SENSORS

SINKS
BIN

Figure 12: Radio Frequency Tracer Test Flow Diagram

Currently RF tracers are only available as a 23mm diameter sphere, this is dictated by the
size of the transponder, however miniaturisation is being investigated, which will allow
the tracers to be made in smaller sizes.

Due to the addition of the chip, the sg is currently also limited to 4.0, which is still
enough for both the coal and diamond recovery industries.
Pilot testwork on the use of RF tracers has allowed calibration of 23mm tracers against a
float/sink analysis as reported by de Korte [4]. It should be noted that the test results
shown below were obtained using a limited number of tracers in a limited number of
density fractions. Larger quantities of tracers and more density increments would have
improved the correlation between the two results presented in figure 13.

219
The South African Institute of Mining and Metallurgy
DMS and Gravity Concentration Operations and Technology in South Africa
R A Heins, P M Grady and R L Langa

Figure 13: Supertracers vs Float & Sink Results (de Korte Jan 2005 [4])

A further potential benefit in the use of RF tracers if recovery technology can be

developed is the removal of the statistical limitation of the tests by using the tracers in
closed circuit. The testing is thus performed continuously. This would also allow a batch
of tracers to be made up to simulate the density distribution of the feed ore and a direct
reading of process efficiency would be obtained.
There are a few limitations to the technology as it currently stands, being primarily the
sizes which can be achieved and that the tracers are still relatively expensive, and as such
single pass testing with disposable tracers is not feasible. Current investigation into
interference if more than one tracer passes the antenna simultaneously is also being
carried out [11].

The benefits of this technology are apparent in that real time, immediate process control
is possible, in addition, the opportunity to add a number of different size and density
tracers simultaneously is also a benefit as the test mechanism thus starts to simulate the
actual process being measured.
The fact that the detection of the tracers is electromagnetic also means that the reading
cannot be masked by material depth on product screens and reduces the potential for
inaccurate results.

220
The South African Institute of Mining and Metallurgy
DMS and Gravity Concentration Operations and Technology in South Africa
R A Heins, P M Grady and R L Langa

CONCLUSIONS
Gravity processes are utilised extensively in most sectors of the extractive industry and
measurement of the efficiency of separation and operation of these processes has long
been under investigation.
Current processes are generally costly, resource hungry and unable to produce accurate
results in real time, with the turnaround time for some tests such as Float/Sink analyses
being in weeks.
The development of tracer technology has improved the situation, however tests are still
time consuming and labour intensive, the introduction of recoverable components into
these tracers such as magnetic compounds and X-ray fluorescent materials has allowed
some automation of the recovery process. However these same developments have
complicated the manufacturing process and hence increased the cost of the tracers to the
point that the risk of losing tracers by failure to recover has caused operators to conduct
tests without feed on to the plant to ensure recovery.
This step has obviously limited the usefulness and validity of the data generated.
More recently the development of significantly lower cost magnetically recoverable
tracers is about to open a whole plethora of new opportunities for process operators.
These include the potential to perform statistically meaningful tests on a routine (per shift
or per day) basis and reliably recover the tracers in order to measure and record process
efficiency, and take action as required to maintain optimum process efficiency.

The next step in using these low cost tracers, is the installation of automated magnetic
recovery and tracer identification systems which will allow real time process
measurement and control down to very fine sizes of tracer, typically 1mm. The
possibility further exists to close the circuit and then recycle these tracers back into the
process in order to obtain continuous real-time monitoring.

Radio frequency tracers provide the ultimate measurement technique and whilst still not
commercialised, present the potential for a completely computerised and automatic
measurement and control system in real time, the cost and physical size issues of the
tracers need to be resolved and the practical implications addressed.

REFERENCES:
1) Aplan FF, Gravity Concentration, SME Mineral Processing Handbook, Vol 1,
Kingsport Press, 1985
2) Babu & Muller, Coal Processing, SME Mineral Processing Handbook, Vol 1,
Kingsport Press, 1985
3) Pinheiro HJ,, Private Communication, May 2006
4) De Korte GJ, Comments on the use of tracers to test dense medium plant
efficiency,January 2005
5) Pinheiro HJ, Springlake Anthracite, Characterisation and Potential Industrial
Applications, unpublished thesis, University of Porto, April 2006.
6) Davis JJ, Wood CJ, Lyman GJ, Use of density tracers for determination of Medium
cyclone partition characteristics, JKMRC, January 1985
7) Partition Enterprises Pty Ltd, website www.partitionenterprises.com.au, May 2006

221
The South African Institute of Mining and Metallurgy
DMS and Gravity Concentration Operations and Technology in South Africa
R A Heins, P M Grady and R L Langa

8) Cresswell M, A selected review of DMS application and engineering design

practice, DMS Powders 9th Technical Dense Medium Symposium,2005
9) Teirlinck P, et al, Ore Dressing Studies, A tool for Design and Treatment Risk
Reduction for a Diamond Mine, Diamonds Source to Use, October 2003
10) Van der Walt H, Private Communication, June 2006
11) De Korte GJ, CSIR, Private Communication, May 2006

222
The South African Institute of Mining and Metallurgy
DMS and Gravity Concentration Operations and Technology in South Africa
Hendriena van der Walt and Dr Kurt R P Petersen

DMS CYCLONE SIZE VS. CYCLONE EFFICIENCY:

THE DE BEERS PERSPECTIVE

Hendriena van der Walt and Dr. Kurt RP Petersen

De Beers Group Services

ABSTRACT

With the decline in fine diamond production from the Argyle mine in Australia, there is a
greater emphasis on efficiently treating the finer size fractions through a DMS circuit.
Although it is generally well understood that smaller diameter cyclones are more
appropriate for this application, practical results do not necessarily meet this expectation.

In this paper the operational performance of particular sized cyclones throughout the
group was compared. Based on the results there is evidence that under certain practical
operating conditions, small cyclones are not necessarily best for fine size particle
treatment, particularly given the production pressures faced by all mines. However, there
seems to be considerable merit in using larger sized cyclones, which tend to exhibit less
performance variability, irrespective of operation conditions.

1. INTRODUCTION

The efficient treatment of diamondiferous ore by dense medium separation has been
standard practice in De Beers for many decades. With the decline in fine diamond
production from the Argyle mine in Australia, there is a greater emphasis on efficiently
treating the finer size fractions such that the supply gap can be addressed. DMS certainly
has a role to play in fine diamond production and it is generally well understood that
smaller diameter cyclones are more appropriate for the application.

However, while the theory is well entrenched, the practical results do not necessarily
meet expectations. This paper discusses the issue of recovery performance for particular
sized cyclones and highlights that size is not a silver bullet when considering recovery
efficiency for a particular sized material. There are many other setup and operational
aspects that can completely overwhelm the basic function of the cyclone, which seems to
have led to a wide variability in performance throughout the group. This is despite the
generally consistent approach to adhering to best practice standards.

This paper presents details on the performance of particular sized cyclones throughout the
group and highlights the basic operational conditions that can detract from this
performance. Moreover, there is evidence that given practical operating conditions, small
cyclones are not necessarily best for fine size particle treatment, particularly given the
production pressures faced by all mines.

2. SUPPLIER CYCLONE SELECTION CRITERIA

223