- Why Sortability Testing Matters
Sortability testing is the structured process by which ore sorting potential moves from theoretical possibility to technical proof. It answers the questions that every mining engineer and project team must resolve before committing to ore sorting technology: can the valuable fraction genuinely be separated from waste under dynamic sensor-based conditions? What grade and recovery is realistically achievable? And which sensor principle is best aligned with your specific ore body?
Unlike standard metallurgical test work, sortability testing for ore sorting must account for three distinct levels of performance - what is theoretically achievable based on particle composition alone, what a given sensor is capable of detecting under ideal conditions and what is actually delivered by a commercial ore sorting platform under realistic feed conditions. Understanding all three, and the gaps between them, is what separates a well-designed ore sorting evaluation from one that produces misleading results.
ROKKSTA's sortability testing services cover all three levels, using a structured, material-first methodology that ensures sensor and equipment selection is driven by your ore's properties - by the contrast that actually distinguishes ore from waste, rather than by assumption or convenience.
- Three Levels. One Complete Picture.
Sortability testing for ore sorting is not a single test - it is a progressive programme that builds technical confidence at each stage. ROKKSTA structures all ore sorting evaluations around three distinct levels of assessment, each addressing a different question and delivering a different type of evidence.
Level 1 — Intrinsic Sortability: What Is Theoretically Possible?
Intrinsic sortability defines the theoretical upper limit of ore sorting performance for your material. It assumes perfect detection and perfect separation - a mathematical ideal - and quantifies what could be achieved if every particle were correctly classified based solely on its composition.
The output of intrinsic assessment is the grade/recovery curve for your ore: a continuous relationship that shows, for any given mass pull to tailings, what mineral recovery and gangue rejection is theoretically achievable. This curve sets the benchmark against which all real sensor and machine performance must be evaluated.
Intrinsic sortability is calculated from particle composition data - typically from ore characterisation results - and accounts for:
• The distribution of valuable minerals and gangue across particles of different sizes
• The degree of compositional overlap between ore classes
• The variability of grade within and between ore types and lithological domains
A strong intrinsic sortability result does not guarantee that ore sorting will work in practice - but a weak intrinsic result is a reliable indicator that it will not. This is why intrinsic assessment is always the first step.
Level 2 — Sensor Amenability: What Can a Sensor Actually Detect?
Sensor amenability testing asks whether the intrinsic contrasts identified in Level 1 are detectable by a specific sensing technology. It is conducted under controlled, static conditions - typically single-particle tests - to isolate sensor detection performance from the mechanical and presentation variables that affect full-scale ore sorting machines.
This is the stage at which sensor principle selection is made. ROKKSTA evaluates multiple sensor technologies against your material before a sorting configuration is fixed, using a structured decision framework that rates each principle against the specific contrast drivers in your ore:
XRT (X-Ray Transmission / Dual Energy XRT): Detects density and atomic number contrast; highly effective for silicate–carbonate discrimination, sulphide detection and materials where compositional contrast is mineralogical rather than surface-expressed. Works on dry/unwashed feed - an important operational advantage.
Color / VIS optical: Detects visible surface color and reflectance contrast; fast, low-cost, effective where ore and waste have consistent visual differentiation.
NIR / SWIR: Detects mineralogical and surface-chemical contrasts in the near and short-wave infrared spectrum; particularly relevant for carbonate-bearing systems, clay minerals, and hydrothermal alteration zoning
Laser : Detects surface texture, reflectance, scattering, UV and structural features; often used in combination with optical sensors
XRF: Detects elemental composition at the particle surface; high selectivity but constrained by throughput and detection limits
Our sensor amenability assessments rate every relevant sensing principle against the physics of your ore, on a common and transparent set of criteria. The result is a sensor selection that is fundamentally aligned with your ore's contrast drivers - chosen because it suits the material, not because it was the most readily available option.
Level 3 — Dynamic Sortability Testing: What Does the Ore Sorting Machine Actually Deliver?
Dynamic sortability testing validates sensor amenability results under realistic ore sorting conditions: continuous feed, belt transport, singulation, sensor scanning, classification, and (pneumatic) ejection, the full mechanical and control system of a commercial ore sorting platform.
At this level, total ore sorting performance is the result of two interacting components:
Classification Effectiveness (CE) - how accurately the sensor and its image processing system classify particles as ore or waste. This reflects detection sensitivity, signal quality, threshold stability and algorithm robustness.
Mechanical Effectiveness (ME) - how faithfully the machine translates classification decisions into physical separation. This encompasses feed singulation (preparation effectiveness), particle presentation to the sensor (presentation effectiveness) and the accuracy of the (pneumatic) ejection system (separation effectiveness).
Understanding the relative contribution of CE and ME to the total performance gap - the difference between intrinsic potential and observed machine output - is critical for optimising ore sorting systems and for making defensible performance projections during plant design.
ROKKSTA conducts dynamic sortability testing at laboratory and pilot scale, using bench-scale single-particle tests that can be upscaled to full-machine performance predictions using our validated simulation methodology. Where larger-scale trials are required, we design and supervise them using a reference sample set built to reflect true ore heterogeneity - whether the trial runs on ROKKSTA equipment or at an external test facility.
- Our Ore Sorting Testing Methodology
ROKKSTA's ore sorting sortability testing programme follows a structured, phase-aligned workflow. Each step builds on the previous one and no stage is bypassed without explicit technical justification.
1. Intrinsic Sortability Assessment
Using ore characterisation data (particle composition, size distribution, density fractions), we calculate theoretical grade/recovery relationships and define the intrinsic sortability curve for your ore. This establishes what ore sorting can achieve under ideal conditions and provides the baseline against which all subsequent test results are evaluated.
2. Grade–Recovery Modelling
We generate complete grade/recovery and mass-rejection curves across the full range of sorting threshold settings. This allows the economic optimum, the threshold at which ore sorting delivers maximum value, to be identified before physical testing begins.
3. Sensor Principle Screening
We screen all relevant sensor principles against your ore's intrinsic contrast drivers using a transparent, weighted decision matrix. Criteria include alignment with the primary separation driver, robustness to surface variability and particle orientation, scalability to industrial ore sorting throughputs and evidence strength from any existing test work. This step selects the sensor principles to be advanced to amenability testing.
4. Sensor Amenability Testing
Static single-particle sensor measurements are conducted under controlled laboratory conditions to quantify detection effectiveness, signal contrast, threshold stability and misclassification behaviour for each shortlisted sensor principle. Results are compared directly against the intrinsic sortability baseline to define the sensor-to-intrinsic performance gap.
5. Dynamic Ore Sorting Trials
Laboratory or pilot-scale dynamic sorting trials are conducted using feed material representative of your ore body. We measure mass rejection, mineral recovery, concentrate and tailings grades, and process effectiveness indicators. We design the test programme ourselves - including reference sample sets, test conditions and evaluation criteria - to ensure results are interpretable against the primary sorting objective, not just operationally convenient.
6. Performance Comparison & Gap Analysis
We systematically compare results across all three levels - intrinsic, sensor amenability, and dynamic - to identify where performance losses arise and whether they are attributable to the material, the sensor, the image processing logic or mechanical presentation and ejection factors. This is the most important analytical step in any ore sorting evaluation and the one most often skipped when a single dynamic trial is treated as the whole answer.
7. Threshold Optimisation
We define the ore sorting threshold settings, the sensor response cut-off values, that deliver the best balance of mineral recovery, gangue rejection and mass pull for your specific project objectives.
8. Economic Evaluation
We translate ore sorting test results into grade, recovery and mass rejection projections and frame these within an economic assessment that quantifies the value of ore sorting for your operation - including waste rejection savings, downstream processing cost reduction and capital payback context.
- Rigorous Method. Honest Results.
A sortability test is only as valuable as the framework that surrounds it. A single dynamic trial, run without a baseline to measure it against, produces a number — but not an answer. That number might reflect the ore's true potential, or it might reflect a sub-optimal sensor choice, a non-representative sample, or a presentation problem on the day. Without a way to tell these apart, a flattering result can drive over-investment, and a poor one can cause a viable ore body to be written off in error.
ROKKSTA's testing methodology is designed to remove that ambiguity. Every test programme is built around four principles:
Sensor selection driven by the ore
We identify which sensor principle is fundamentally best aligned with your ore's contrast drivers before a test configuration is fixed - so the programme tests the right question, rather than confirming whatever sensor was easiest to reach for.
Test design driven by the decision
The reference sample set, size fractions, test conditions and evaluation criteria are defined by your project's objectives and your material's properties - built to answer the specific question your project faces at its current stage, not to suit operational convenience.
Results interpreted against a baseline
Every dynamic ore sorting result is evaluated against the intrinsic sortability curve and sensor amenability data established in the earlier stages. This allows a clear distinction between material limitations and machine limitations - the single most important, and most frequently missed, step in ore sorting evaluation.
Gaps reported transparently
If results are ambiguous, non-representative or constrained by sample availability, we say so. Our recommendations are always qualified by the evidence actually available, never by optimistic extrapolation. A defensible "not yet" is worth more to your project than an unsupported "yes."
What protects the integrity of a sortability result is not where the test is run - it is how the programme is designed, what it is benchmarked against and how honestly it is reported. This methodology holds whether testing is conducted on ROKKSTA's own equipment, on our bench-scale test rig or at an external facility.
- What You Receive
Deliverable | Description |
|---|---|
Economic Assessment | Grade, recovery, and value projections; sorting vs. no-sorting comparison for feasibility evaluation |
Optimised Threshold Parameters | Defined sensor response cut-off settings for ore sorting operation |
Intrinsic vs. Sensor Performance Comparison | Quantified analysis of performance gaps between theoretical potential and measured machine output |
Dynamic Test Results | Laboratory or pilot-scale ore sorting performance: recovery, mass pull, concentrate grade, tailings grade, process effectiveness |
Sensor Response Dataset | Static amenability test results: signal distributions, contrast margins, detection limits, misclassification behaviour |
Sensor Principle Decision Matrix | Structured evaluation of sensor technologies against your ore's contrast drivers, with selection rationale |
Intrinsic Sortability Data | Theoretical grade–recovery and mass-rejection relationships across sorting threshold settings |
Sortability Report | Summary of intrinsic, sensor amenability and dynamic test outcomes with interpretation in the context of ore sorting decision readiness |
- Testing Aligned to Your Project Stage
The depth and scope of sortability testing appropriate for an ore sorting evaluation depends on the project phase. ROKKSTA designs test programmes that match the level of evidence required for each decision point - avoiding under-testing (which leaves critical questions unanswered) and over-testing (which consumes resources on precision not yet needed at that stage).
Scoping Phase - Intrinsic sortability assessment using characterisation data or existing assay information. Establishes whether ore sorting is technically plausible and justifies further investment. Requires no new physical testing if characterisation data is already available.
Pre-Feasibility Phase - Full three-level programme: intrinsic sortability modelling, sensor principle selection and amenability testing and dynamic trials using a representative reference sample set. Outputs a preliminary ore sorting mass balance and Class 4 cost estimate. Closes the decision-critical technical gaps needed to shortlist ore sorting as a process option.
Feasibility Phase - Representative bulk ore sorting trials under dynamic conditions, validated against the pre-feasibility amenability and intrinsic baselines. Outputs a ROM-representative mass balance and ore sorting performance dataset suitable for Class 3 (bankable) cost estimation and final process design.
Operational Optimisation - For existing ore sorting installations where performance is below expectation: targeted amenability and threshold testing to identify whether the root cause lies in sensor calibration, threshold settings, feed preparation or an intrinsic change in the ore being mined.
- Benefits to Your Operation
.png)
Confirm ore sorting viability before committing capital - establish whether ore sorting is technically justified for your ore body with structured, phase-appropriate evidence
Select the right sensor technology from the start - structured sensor principle evaluation prevents misdirected investment in sub-optimal ore sorting equipment
.png)
Understand the full performance envelope - grade/recovery modelling at multiple threshold settings gives project teams the complete picture, not just a single test point
Separate material limits from machine limits - know whether ore sorting performance constraints are intrinsic to your ore or addressable through equipment choice and process design
.png)
Reduce feasibility study risk - test programmes designed against your project objectives deliver evidence that holds up under engineering and investment scrutiny
Support planning and investment decisions - sortability data feeds directly into process simulation, plant sizing, and economic modelling at every project phase
- Part of a Structured Ore Sorting Workflow
Sortability testing sits at the centre of ROKKSTA's ore sorting evaluation workflow:
provides the particle composition, size distribution, grade variability and mineralogical data that make intrinsic sortability modelling possible. Sortability testing without prior characterisation is always limited by the quality of the available feed data.
uses particle data and sorting threshold results from sortability testing to generate plant-level mass balances, equipment sizing and ore sorting performance projections - the engineering inputs needed for pre-feasibility and feasibility studies.
use the sensor response and performance baseline data from sortability testing as reference benchmarks when evaluating whether an operating ore sorting installation is performing at its potential.
for XRT and other sensor systems can be developed using material characterised and tested during the sortability programme, ensuring calibration references accurately reflect your ore's physical and compositional properties.
Frequently Asked Questions
- Get in Touch With ROKKSTA
Ready to optimise your ore sorting process? Our team of experts is here to provide tailored solutions and independent insights.
