DESIGN, INSTALLATION AND TESTING OF THE NO 250 WEMCO

SMARTCELL™ FLOTATION MACHINE AT MINERA LOS PELAMBRES

A Weber , L MacNamara, H Schreiber

ABSTRACT

Design and development of the world’s largest flotation cell was completed in 2002. This
3
was the first 257 m flotation cell installed in the mineral industry and represents a major
development in the design of larger flotation cells. This cell has a 50% increase in volume
3 ® ™
over the widely installed 160 m WEMCO Smartcell flotation machine. In the next few
3
years it is anticipated that 257 m flotation cells will be incorporated into rougher-scavenger
flotation circuits in many more world class copper sulphide concentrators.
3
The prototype 257 m cell, No.250 Smartcell™ flotation machine was installed at Minera Los
Pelambres in the front of Row 1 in the existing rougher flotation circuit. This cell is followed
by nine No. 130 Smartcell™ flotation cells. After the successful development phase, this unit
has been incorporated as part of a row of No. 250 Smartcell™ flotation machines at Minera
Los Pelambres. A key to the successful development of the equipment has been the emphasis
placed on fundamental hydrodynamic scale -up criteria. Hydrodynamic factors include
superficial gas velocity, water (pulp) circulation through the draft tube, and pulp mixing
(distribution) characteristics.

This paper will review the scale-up criteria of the No. 250 flotation cell, and present a case
study of the in-plant equipment testing. The case study will document the metallurgical and
hydrodynamic performance of the machine. It will also illustrate the economic benefits
derived by incorporating larger flotation cells in a modern copper sulphide concentrator.

-1-
 INTRODUCTION
rd
J. Graham Eby presented a paper on large capacity flotation cells at the 73 Annual General
Meeting of Canadian Institute of Mining and Metallurgy. The meeting was held in April
1971, and his presentation described the latest development in large flotation equipment with
3 3
volumes of 8.5 m and 14.1 m . Increasing numbers of low-grade ore bodies, higher tonnage
plants, and larger grinding mills (Eby, 1972) were the driving factors cited for the introduction
of large capacity machines. Thirty-five years later, these same factors, namely economies of
scale, have lead to the development of the No. 250 flotation cell.

Development costs have increased with higher capacity equipment. This increase in cost is
not only incurred by the machine manufacturer, but also by the mineral producer who agrees
to have the machine tested at their plant. In addition to the obvious increase in fabrication and
component costs, costs are also incurred for the foundation, infrastructure, interconnecting
piping, sample systems for the feed / concentrate / tailing streams, and equipment installation.
Once the machine is in operation, modifications to the equipment equates to additional costs in
terms of time and expense. The plant operator must contend with interruptions in daily
operations during testing, and more importantly potential loss of production. These real costs
are offset by the introduction of a new product for the machine manufacturer. Benefits
derived for the plant operator include an increase in flotation residence time, lower installed
costs for future expansion, and mitigation of the risk involved with an expansion of similar
size cells.

Recognizing the economic factors involved with new equipment development, emphasis is
placed on proven scale -up methodologies and fundamental hydrodynamic scale -up criteria.
Hydrodynamic factors included in the machine’s design include superficial gas velocity, water
(pulp) circulation through the draft tube, and pulp mixing (distribution) characteristics.

CELL DESIGN AND CHARACTERISTIC HYDRODYNAMIC PARAMETERS

™
The basic design of the No. 250 Smartcell flotation machine is consistent with previously
designed WEMCO flotation equipment. The mechanism is designed to simultaneously induce
ambient air into the pulp through an air control valve, and suspend / mix the contents of the
cell, thus drawing the pulp into contact with the induced air. This dual pumping action is
accomplished by the rotor generating a rotating liquid vortex which extends from the upper
portion of the standpipe to the mid section of the rotor. A low pressure zone (5-10 cm. water)
generated in the core of the vortex ingests ambient air into the flotation region. Primary pulp
circulation initiates at the bottom of the cell and travels upward through the draft tube.
Secondary pulp circulation occurs in the mid-section of the cell and enters the flotation region

-2-
through the area delineated by the periphery of the draft tube and disperser’s inner diameter.
These flow patterns are illustrated in Figure 1.

Ambient Air

Secondary Pulp Circulation

Primary Pulp Circulation

Figure 1 – The WEMCO® Smartcell™ flotation machine

Improvements in the design of flotation equipment which are also incorporated in the No. 250
machine are radial launders and vertical baffles. These features and other standard
components are depicted in Figure 2. The use of vertical baffles in round tanks prevents
swirling of both the pulp and the froth. This enhances cell mixing and improves pulp
residence time distribution. Also froth transit times are reduced when froth swirling is
minimized. Radial launders increase lip length and further reduce froth transit time,
minimizing froth zone recovery losses and maximizing concentrate recovery. It has been
shown in plant test work that the inclusion of these features significantly increases recovery
for only a small reduction in copper grade, (Weber, et. al, 2001).

-3-
 Radial Launder
Standpipe

Vertical Baffle
Disperser Hood Rotor

Disperser
Draft Tube

Figure 2 – Radial launders and vertical baffles

WEMCO flotation machines, specified by a given rotor diameter, are characterized in terms of
hydrodynamic parameters. As used in this paper, hydrodynamic parameters are defined as the
relationship between air ingestion and liquid flow as a function of rotor submergence, rotor
engagement, and rotor speed. Rotor submergence is the vertical distance from the froth
overflow weir to the top of the rotor. Rotor engagement is defined as the vertical distance of
the rotor which is encased within the draft tube. These hydrodynamic parameters are used in
subsequent calculations to verify predicted hydrodynamic scale -up criteria. Table 1 shows a
™
comparison of the hydrodynamic criteria for the No. 130 and No. 160 Smartcell flotation
™
machines vs. the scale -up criteria for the No. 250 Smartcell flotation machine.

Table 1. Hydrodynamic Scale-up Criteria

Machine Size m3

Criteria (Water Only) Definition No. 130* No. 160* No. 250**

Superficial Gas Velocity Qa / Ac (cm/sec) 2.1 1.65 2.1

3
Specific Cell Power P / Vc (kW/m ) 0.92 0.89 1.07
3
Liquid Circulation Vl *Ad (m /min) 95.4 108 197
3
Air Capacity Qa / ND 0.15 0.14 0.10
3 5
Power No. Pg / ?N D 5.7 6.0 6.9

* Actual Values: ** Predicted Value

-4-
In Table 1, symbols are defined as follows

Qa = Air Ingestion Rate P = Absorbed Power

N = Rotor Rotational Speed g = Acceleration due to Gravity

? = Liquid Specific Gravity Ac = Cell Surface Area

D = Rotor Diameter Vc = Cell Volume

Vl = Liquid Velocity Ad = Draft Tube Cross Sectional Area

Based on the predicted scale -up criteria, it was determined that a 1.40 meter diameter rotor
operating at rotational speeds in the range of 90 – 110 rpm was required. The No. 250 tank is
8.25 meters in diameter with an overall tank height of 6.14 meters. The operating height,
including the power drive, is 7.89 meters and the total working height required to lift the
complete mechanism clear of the cell is 11.27 meters. The total weight of the machine
including mechanism, motor, and drive unit is 31,109 kg. Table 2 summarizes the No. 250
tank and drive specifications and shows a comparison to No. 130 and No. 160 Smartcell.™

Table 2. Tank and Drive Specifications

Tank Specifications Drive Specifications

Water Water Only
Liquid
Diameter Liquid Rotor Installed Only Specific
Model Vol.
3 (m) Height Drive Dia., Power Power Power
(m ) 3
(m) Type (m) (kW) (kW) (kW/m )
130 130 6.62 4.47 V-Belt 1.09 150 119 0.92*
160 160 6.86 4.95 V-Belt 1.14 185 141 0.89*
250 257 8.25 5.53 Gear 1.40 340 275 1.07**
Reducer
* Actual Values: ** Predicted Values

AUXILIARY EQUIPMENT AND INSTALLATION

In 2002, an agreement was reached with Minera Los Pelambres to install and test the machine
at their concentrator. Los Pelambres is the world’s fifth largest copper mine and is located in
Chile’s Choapa Province, 200 km north-east of Santiago. The ore body is of porphyry copper
mineralization containing chalcopyrite, chalcocite, bornite, covellite and molybdenum.
Original plant design capacity was 85,000 mtpd with current production rates in excess of
125,000 mtpd.

The grinding circuit utilizes (2) 10.7 meter diameter, 12.67 megawatt SAG mills in
combination with (4) 6.4 meter diameter, 7.08 megawatt ball mills. Each ball mill has a

-5-
dedicated cyclone cluster consisting of (14) 838 millimeter cyclones. The cyclone overflow
particle size distribution is approximately 80% passing 250 microns. Overflow material from
each cyclone pack is combined in a common distribution box and diluted to approximately 38-
42 wt. % prior to rougher flotation. The Rougher flotation circuit consists of four parallel
rows of (9) No. 130 Wemco Smartcell™ flotation cells.

The prototype No. 250 machine was installed in the rougher circuit ahead of the first cell of
Row 1. A diversion box was installed upstream of the No. 250 cell which allowed slurry to
either enter the No. 250 cell or be diverted to the existing distribution box. This flexibility
allowed test work to be conducted and modifications to be made with a minimum of
interruption to the concentrator’s day-to-day operation. This feed configuration resulted in
less than a 0.05% copper variation in the feed assay of Row 1 as compared to Rows 2, 3 and 4.

Dorr-Oliver Eimco and Minera Los Pelambres’ scope of supply for this project is summarized
in Table 3. A photograph of the installation is shown in Figure 3.

Table 3. Scope of Supply

Dorr-Oliver Eimco’s Responsibility Minera Los Pelambres Responsibility

No. 250 Cell with Platforms Site for No. 250 Cell

Feed and Connection Boxes Feed Slurry as Required for Testing

Detailed Site Engineering Utilities

Civil Works Metallurgical Analysis

Interconnecting piping On-site Accommodations

Feed, Concentrate and Tail Samplers Operation Support

Instrumentation and Electrical

Installation

Operation, Maintenance and Test Support

-6-
Figure 3. Photograph of No. 250 Cell

HYDRODYNAMIC AND METALLURGICAL RESULTS

After installation and mechanical testing of the No. 250 cell, the first phase of design
validation was water only testing. The purpose of this work was to confirm that the machine
operating parameters produced the predicted hydrodynamic characteristics consistent with
historical scale-up criteria. As such, the test program was designed to exceed both the upper
and lower limit of the design criteria. Factors considered during this phase of the investigation
include submergenc e, engagement, and rotor speed. Measured responses include d power
consumption, water circulation rate through the draft tube, and aeration rate.

Electrical power measurements were taken directly from the output of the Allen-Bradley
model ACS 600 - 380v variable frequency drive. Water circulation through the draft tube was
measured using a Marsh-McBirney Flomate™ 2000 magnetic probe and the aeration rate was
measured by a Sierra Mass Flow meter, model 640S. During each test series, data for all the
response factors, was continuously measured and stored for computer tabulation and
evaluation.

Six test series completed are listed in Table 4. Each series consisted of 5 tests conducted
randomly at rotor speeds of 80, 90,100,110, and 120 rpm.

-7-
Table 4. Summary of Factors and Levels of Hydrodynamic Experimental Program

Factor

Test Series Submergence, mm Engagement, mm

1 635 146

2 635 267

3 889 267

4 889 146

5 380 146

6 380 267

Data reduction and statistical analysis produced a Superficial Gas Rate model, Model 1. The
2
model is statistically significant with an F-value of 235 and a predicted R of 0.95. Both rotor
speed and submergence proved to be statistically significant factors. Model 1 is presented
below and the response surface diagram is shown Figure 4.

Figure 4. Superficial Air Velocity Response Surface

X = Submergence
Y = Rotor Speed

Engagement = 20 0.94
(cm) 0.64
0.34
Log10(Sup. Air Vel.)
(cm/sec) 0.04
-0.26

120
89
110
76
Rotor Speed 100
63
(rpm)
90 51 Submergence
80 38 (cm)

Model 1: Log10 (Superficial Air Velocity) = -1.061 – 0.0211 * Submergence + 0.1929 * Rotor
Speed.

Rotor speed, submergence, and engagement proved to be significant factors for cell absorbed
power and liquid circulation. The model developed for Cell Power, Model 2, resulted in an F-

-8-
 2
value of 429 and a Predicted R of 0.97. Liquid circulation through the draft tube, Model 3,
2
proved to be significant with an F-value of 25 and a Predicted R of 0.67 which is in
2
reasonable agreement with the Adjusted R of 0.71. Equations for Models 2 and 3 are
presented below, with Figures 5 and 6 representing the respective response surfaces.

Figure 5. Cell Power Response Surface

X = Submergence
Y = Rotor Speed

Actual Factor
1.04

0.88
0.72

Cell Power 0.56

3
(kW/m ) 0.40

120
89
110
76
100
63
Rotor Speed
(rpm) 90 51
Submergence
80 38 (cm)

Model 2: Cell Power = -0.271 + 0.004 * Engagement – 0.133 * Submergence + 0.007 * Rotor
Speed + 0.00021 * (Submergence * Rotor Speed)

-9-
Figure 6. Liquid Circulation Response Surface

X = Engagement
Y = Submergence

Actual Factor
Rotor Speed = 105 202
(rpm) 185

Liquid Circulation 169

3
(m /min) 152

136

89
25
76
22
Submergence 63
19
(cm)
51 16 Engagement
38 13 (cm)

Model 3: Liquid Circulation = 150.71 +5.303 * Engagement + 2.17 * Submergence – 0.780 *

Rotor Speed

The aforementioned models were then used to optimize the machine’s hydrodynamic
characteristics while minimizing the cell’s specific power. This exercise resulted in a
submergence of 889 mm, 177 mm of rotor engagement and a rotor speed in the range of 100 -
108 rpm. A comparison of the predicted scale -up factors versus the actual hydrodynamic
factors is summarized in Table 5.

Table 5. No. 250 Predicted vs. Actual Hydrodynamic Design Factors

Design Cri teria Definition Predicted Values Actual Values

Superficial Gas Velocity Qa / Ac (cm/sec) 2.1 2.1

3
Liquid Pulp Circulation Vl / Ad (m /min) 197 187
3
Cell Specific Power (Water P / Vc (kW / m ) 1.07 0.88
Only)
3
Air Capacity Qa / ND 0.10 0.11
3 5
Power No. Pg / ?N D 6.9 7.1

As shown, both the predicted and measured superficial gas velocities are 2. 1 cm/sec. The
measured liquid pulp circulation was approximately 5% less than the predicted value of 197

- 10 -
 3
m /min. The measured air capacity, (0.11) and power number (7.1) were 10% and 7% greater
than the predicted values respectively. All of these hydrodynamic factors are in reasonable
agreement with the machine’s design criteria. By optimizing the cell’s design, the measured
3
cell specific power (0.88 kW/m ) is approximately 20% less than the predicted value.

The final phase of the test program include d residence time distribution tests (RTD) and
metallurgical tests. Dorr-Oliver Eimco contracted the Chilean Commission of Nuclear Energy
(CCHEN) to conduct residence time distribution (RTD) measurements in the No. 250 cell.
CCHEN determined 15 milicuries of Na-24 were sufficient to properly describe the RTD of
82
each size fraction in the investigation. For the liquid tracer, 15 milicuries of Br was
employed. Both the tailing and concentrate streams were included in the analysis. Table 6
identifies the particle size distribution of each size fraction. The irradiated samples were
introduced into the feed box, while the tailings and concentrate streams were monitored to
determine the cell’s residence time distribution. Collected data was corrected for natural
background radiation and the half-life of the irradiated material

Table 6. Activated Size Fractions

Tracer Size Designation Particle Size

Global No Separation

Fine -325 mesh

Intermediate -50 + 325 mesh

Coarse + 50 mesh

The graphs, Figures 6 and 7 display no early peaks or long drawn out tails that would indicate
short-circuiting of the slurry or dead zones in the cell (Levenspiel, 1993). The curves are
characteristic of a perfectly mixed reaction vessel. Experimental mean residence time was
calculated from the RTD by,

?tC(t)dt
Texp =
?C(t)dt

where: Texp = experimental residence time

t = measured time
C(t) = tracer concentration at time t

- 11 -
Figure 6. Global RTD, Tailing Stream Sample

800
700
Radiatiion Intensity (cps)

600
500
400
300
200
100
0
0 500 1000 1500 2000
Time (sec)

Figure 7. Coarse RTD, Tailing Stream

900
800
Radiaton Intensity (cps)

700
600
500
400
300
200
100
0
0 500 1000 1500 2000
Time (sec)

- 12 -
The measured residence time is presented in Table 7.

Table 7. Measured Residence Time

Tracer Size Designation Tail Sampler (min) Concentrate Sampler (min)

Liquid 6.6 6.2

Global 6.7 6.2

Fine 6.1 6.0

Intermediate 6.2 5.8

Coarse 6.3 5.5

Concentrate 5.3 7.1

The mean residence time was 6.17 minutes with a standard deviation of 0.5 minutes. The data
set range d from 5.3 to 7.1 minutes and all values fell within two standard deviations (Oblad,
2004). There were no outliers in the data set. Over the time of the tests, the average feed rate
was 1203 mtph and feed density was calculated at 40 wt%. For WEMCO flotation cells
3
greater than 100 m , empirical data indicates flotation cell volume should be de-rated by 10%,
which equates to a theoretical residence time of 6.13 minutes. The difference in observed and
theoretical residence time may be attributed to variations in process conditions during the test
period. The residence time data indicates the machine is a well mixed vessel with no stagnant
areas or short circuiting.

METALLURGICAL TESTS RESULTS

The benefit of installing the No. 250 cell at Los Pelambres was an increase in plant flotation
residence time. At current production rates of 125,000 mtpd, the addition of one No. 250
flotation cell increased the residence time of rougher Row 1 from 21.8 minutes to 26.8
minutes. Residence time calculations are based on a feed solids concentration of 38 wt. % and
a 2.6 solids specific gravity.

During the in-plant testing of the No. 250 cell, the plant’s automatic sampling system was
configured to collect a composite feed sample for each of Rows 2, 3 and 4. Combined tails
samples were taken for Rows 1 & 2, and also for Rows 3 & 4. A combined concentrate
sample was then taken from the entire plant. Although not ideal, this automatic sample system
allowed for continuous samples to be obtained over several months. The 127 data points
collected and subsequent analysis were deemed far superior to the limited number of manual
samples which could have been economically taken on each stream. The analysis of this data
indicated that the combined recovery of Rows 1 and 2 was 1.58% higher than the combined
- 13 -
recovery of Rows 3 and 4. This variation in recovery is statistically significant with a P-value
of less than 0.005.

Ten metallurgical samples were also taken around the No. 250 cell. During the sample period,
the machine’s submergence and engagement were 890 mm and 180 mm, respectively. Rotor
speed was held constant at 100 rpm. Process test conditions are outlined in Table 8. Stated
values were averaged over the entire data set as analysis indicated there were no outliers. The
results from this data set produced an average copper recovery of 59.1 %. Concentrate
averaged 33.88 % Cu and the ratio of concentration average was 59.7 %.

Table 8. Test Conditions

Feed Rate Feed Grade Pulp Density Froth Depth
(m3 /hr) (% Copper) (% solids) (mm)

2769 0.96 39 117

Minera Los Pelambres evaluated the effectiveness of the No 250 Smartcell™ by measuring
the increase in recovery associated with an increase in residence time for Row 1. They then
assumed an equivalent residence time for rows 2, 3 & 4 and projected a recovery increase
based on the Klimpel Flotation Model. As reported by Fernando Villanueva, 2006, the
projected increase in recovery for rows 2, 3 & 4 would be 1.49% as compared to the actual
increase of 1.7 % measured from Row 1.

Fernando goes on to report that, compared to (2) No 130 Wemco flotation cells, the No 250
flotation cell recovered 0.9 wt% more copper. The majority of which was in the +100 mesh
size fraction. Table 9 summarizes recovery by particle size for the (2) 130 machines as
compared to the No. 250, (Fernando 2006.)

Table 9 Recovery by Particle Size

Tyler Mesh No 130 Cells No 250 Cell
Recovery % Cu Recovery % Cu
+48 3.2 6.9
48 x 65 25.2 44.7
65 x 100 49.1 70.2
100 x 200 67.2 76.6
200 x 325 74 75.9
-325 67.6 62.6
Total 62.4 63.3

- 14 -
 ECONOMIC CONSIDERATIONS

Driven by economies of scale, machine manufacturers continue the development of higher

capacity flotation machines. As part of this development, it is incumbent on the manufacturer
to rationalize the economic benefit to the plant operator. A partial list of factors, in this
cost/benefit analysis; include identifying applicable mineral applications, production rates,
equipment sizes, and the commodity market value. Additional factors include cost parameters
such as equipment, building, installation, maintenance, operational, power, mill product
transportation, refining, and final delivery. It is not the intent of this paper to review such an
exhaustive economic evaluation, but to provide the reader with an installed cost comparison of
the No. 250 flotation cells versus No. 160 flotation cells, the previous largest size
commercially available.

In this example, the scope is defined by a 100,000 mtpd porphyry copper concentrator with the
following process parameters:
3
Solids Concentration - 32 wt% Solids Specific Gravity - 2.65 t/m
3
Slurry Feed Rate - 11,000 m /hr Retention Time - 25 minutes

Plant Availability - 94%

From this data it directly follows that the plant requires either three rows of (8) No. 250
flotation cells, or four rows of (9) No. 160 flotation cells. Costs allocated to equipment
installation (i.e. foundation, structural, utilities), and flotation equipment are itemized in Table
10. The data has been normalized and presented as percent of the cost of the No. 160 cells.

Table 10. Equipment Installed Cost Summary

Machine Size (m3 ) Installation (%) Equipment (%) Total (%)

160 100 100 100

257 79. 7 94.2 85.7

As shown, a rougher circuit consisting of No. 250 cells would save approximately 15% in
installed costs. Another parameter, which is easily verifiable , and often factored into the
installed cost of the equipment , is equipment power cost. Plant data indicates that a circuit of
No. 250 cells, configured as described in the metallurgical test program, would save 7% on
3
power costs when compared to a circuit of 160 m flotation cells.

- 15 -
 CONCLUSIONS
3
Over the past four decades, an individual flotation cell’s capacity has increased from 8.5 m to
3
over 250 m . At Dorr-Oliver Eimco, each new cell size has been based on a set of
hydrodynamic scale -up criteria. The hydrodynamic data collected during the in-plant test
program, enabled the No. 250 machine’s configuration to be optimized in such a manner as to
maintain consistent hydrodynamic characteristics and minimize the cell’s specific power.
Residence time distribution tests indicate the No. 250 cell is a well mixed vessel even while
operating with an actual specific power of approximately 18% lower than predicted.

Metallurgical samples taken around the No. 250 cell produced an average copper recovery of
59.11% with an average concentrate grade of 33.88 % and a 59.66 ratio of concentration.
Data collected by Los Pelambres indicate the No 250 flotation cell recovers significantly more
coarse material when compared to two No. 130 flotation cells. This increase in coarse particle
recovery resulted in higher copper recovery at an equivalent residence time.

A rougher circuit comprised of No. 250 cells would save approximately 15% on an installed
basis and 7% in power consumption costs compared to a rougher circuit consisting of No. 160
cells. Larger cells also benefit from fewer required spare parts for the plant and reduced plant
down time for maintenance.

- 16 -
 References

Degner, V. R., and Treweek, H. B., 1976. Large Flotation Cell Design and Development. in
Flotation Vol. 2 ed. M.C. Furstenau, AIME, New York, NY, 816-837.

Degner, V. R., 1987. 3000 Cubic Foot Flotation Machine Development. Transaction. AIME

Eby, J.G., WEMCO Large -Capacity Flotation Cells. 1972 The Canadian Mining and
Metallurgical. 50-55

Levenspiel, O., 1993. The Chemical Reactor Omnibook. Corvallis, OR: OSU Book Stores,
Inc.

Lelinski, D., J. Allen, and A. Weber, 2003. Analysis of the Residence Time Distribution in
Large Flotation Machines, Minerals Engineering, 15(7), 449-505.

Nelson, M. G., and D. Lelinski, 2000. Hydrodynamic Design of Self-Aerating Flotation

Machines. Minerals Engineering, 13:10-11, 991-998.

Oblad, H. B. (Consulting Engineer), personal communication, December, 2004.

Weber, A., C. Walker, L. Redden, D. Lelinski, and S. Ware, 1999. Scale-Up and Design of
Large-Scale Flotation Equipment, in Advances in Flotation Technology, ed. B. K. Parekh and
J. D. Miller, Littleton, Colorado: SME, 353-369.

Villanueva, F., E. Tobar, A. Nunez, J. Tapia, Implementacion en Projecto Repotenciamienta 1,

3
Celdas de Flotacio 257 m Wemco® Smartcell™, Presentation, Santiago, Chile November
2006

- 17 -