FAQ

Bulk Ore Sorting & Magnetic Resonance Technology

Will Magnetic Resonance-based sorting work for my operation?

The main characteristics to assess when determining the applicability of Magnetic Resonance-based sorting are ore mineralogy and the variability of grades within the ore, called heterogeneity. The very first stage of analysis by NextOre will include an assessment of the minerals for detection, the in-situ heterogeneity of the ore, and the expected ore “mixing” that will occur as a result of operations prior to sorting.

What benefit do I get from bulk ore sorting?

Bulk ore sorting delivers a combination of benefits that together can fundamentally improve the economics of a mining operation. Increased process plant feed grade leads to higher metal production per tonne processed. Reduced tonnage to the plant lowers operating costs across comminution, flotation, and tailings management. Lowering the mining cut-off grade allows previously uneconomic material to be captured, expanding the effective resource. Improved grade control reduces misclassification of ore and waste, recovering metal that would otherwise report to waste dumps or low-grade stockpiles. A reduced environmental footprint follows naturally less energy, less water, and less tailings per tonne of product.

One published case study modelling a large copper open pit showed a 13% increase in annual copper production, a head grade improvement from 0.39% to 0.45%Cu, and a 23% reduction in copper sent to waste. All within the same fixed plant capacity.

How do I estimate the benefit of bulk ore sorting?

The benefit of bulk ore sorting is best estimated through a heterogeneity study an analysis of the grade variability in your orebody at the point where sorting would take place. The key driver is the natural grade variability (heterogeneity) of the orebody and how well it is preserved through mining and crushing before reaching the sorter. The output of these studies are a series of ‘sorting curves’, showing the mass and metal recovery in bulk sorting for a given feed source, typically showing the impact of varying degrees of operational mixing.

Beyond the sorting curves, an economic benefit assessment should consider practically achievable implementation scenarios, for example in an operating mine these scenarios may include: increasing head grade to the processing plant within fixed plant capacity; reducing the mining cut-off grade to mine more tonnes and recover previously uneconomic material; and reducing dilution and ore loss through improved grade control. NextOre’s services include performance of heterogeneity studies using MR analyser data, delivering the granular, tonne-by-tonne grade measurements required to properly model these outcomes.

Can you estimate grade uplift and waste material rejection for a target mining project?

Yes, and this is a core part of NextOre’s engagement process. The starting point is a heterogeneity study, an assessment of the in-situ grade variability of your orebody at the intended sorting point. NextOre draws on geological information in the form of drill core data, and comparable deposit analogues to produce an initial estimate of sorting potential. Every orebody is unique, so the results must be evaluated on a case-by-case basis, but published results from NextOre operations give useful benchmarks: 15–40% tonnage reduction is targeted for copper operations with suitable heterogeneity, with corresponding metal retained of 97%–89%. Spodumene deposits have been observed to have even higher heterogeneity and prospectivity for bulk ore sorting, with the same 15–40% tonnage reductions calculated retaining 99%–93% of feed metal, respectively.

How much can bulk ore sorting increase the grade of my copper ore?

This varies by orebody and depends on the in-situ grade variability, and the degree to which it is preserved to the sorting point. In a Metso case study of a large copper open pit, bulk ore sorting increased head grade from 0.39%Cu to 0.45%Cu — a 15% relative improvement. Modelling at the Phu Kham operation showed that net metal production increases of 13–40% are achievable when cut-off grade optimisation is combined with sorting.

Critically, the full benefit is not just grade upgrade. Metal losses from sorting are more than compensated by recovery of marginal tonnes that would otherwise be wasted. At the Cozamin underground mine trial conducted with NextOre, sorting delivered a net copper metal improvement of 6.3–7.8% versus conventional truck assignment. The headline grade improvement number alone risks understating the total value delivered.

What are the dominant sensing technologies used for preconcentration?

For bulk ore sorting on conveyor belts, only fully penetrative sensing technologies are suitable. Surface-only sensors like XRF, colour cameras, or LIBS cannot provide a representative measurement of a large, heterogeneous mass of rock. The three dominant penetrative options are Magnetic Resonance (MR), PGNAA (Prompt Gamma Neutron Activation Analysis), and PFTNA (Pulsed Fast and Thermal Neutron Activation).

MR directly measures the target mineral at the atomic level using tuned radio frequency pulses. It is fast (seconds per measurement), highly accurate, and mineral-specific. PGNAA and PFTNA measure elemental composition and are penetrative, but their measurement cycle times are in the order of minutes rather than seconds, currently too slow for effective pod-by-pod bulk sorting decisions and are susceptible to calibration drift. They are more suitable for ore characterisation than active sorting. NextOre’s MR analysers are the only commercially deployed technology capable of delivering grade measurements fast and reliably enough for real-time bulk ore sorting at high throughput.

How reliable are magnetic resonance analysers?

They are extremely dependable. At Kansanshi Mine (First Quantum Minerals, Zambia), the NextOre MRA functioned continuously throughout a complete commercial trial at 2,800 tonnes per hour, marking the largest publicly documented installation of sensor-based bulk ore sorting to date. At Lundin Candelaria in Chile, the MRA has served as a crucial operational data source since its commissioning in early 2023, operating on a 6,000 TPH crusher discharge belt. In blind validation testing at Cozamin, the MRA achieved an R² of 0.96 compared to laboratory assay results. At Kansanshi, static testing yielded an R² of 0.98.

The MR signal is responsive solely to the tuned mineral, without interference from other rock matrix minerals, moisture levels, variations in particle size, or changes in ore type. This characteristic ensures its robustness across the highly variable conditions present in actual mining operations, where conventional sampling techniques or inferential sensors are significantly more vulnerable to noise and calibration drift.

Why doesn't a magnetic resonance analyser need to be recalibrated?

There are two main reasons:

First, MR assesses a fundamental physical characteristic, the quantitative resonant response of atoms within the tuned mineral phase. Instead of deducing grade from a material’s property. The correlation between the signal and the mineral quantity is determined by the number of nuclei present in the target mineral phase.

Second, due to the highly distinct response frequencies, resonant frequencies for minerals span a wide range (from 1MHz to 100MHz) and rarely overlap. This means that only the mineral for which the analyser is tuned is excited and responds to a specific resonant frequency. This behaviour can be contrasted with spectroscopic methods like XRF or PGNAA, where energy peaks often overlap among elements, necessitating adjustments for the presence of other elements.

MR analysers come pre-calibrated from the factory. A single proportionality factor linking MR signal to mineral mass is established once during the commissioning process and verified with a known standard on-site. After this initial setup, no further recalibration is necessary.

What does the Magnetic Resonance Analyser "sense" in ore?

Whereas some other sensors measure characteristics of ore like density, surface composition/colour, or permittivity/conductivity, the Magnetic Resonance Analyser is a direct, quantitative measurement of the weight of the target mineral in the ore. This is combined with data collected from a weightometer or belt scale to deliver accurate grade measurement of the ore stream on a metre-by-metre basis.

What precision can Magnetic Resonance measurements achieve?

The Magnetic Resonance sensor has delivered sensing resolutions (1-sigma standard deviation) less than 0.025% copper. Sensor performance quoted for copper as chalcopyrite at full production rates.

Are any studies available that prove bulk ore sorting works and how it can be implemented?

NextOre believes high-quality, comprehensive and transparent reporting of bulk ore sorting projects is essential to advance adoption of the technology by the mining industry. Detailed studies have been published, either by the mining company implementing bulk sensing and sorting at their mine, or by NextOre personnel, showing methodologies and results of bulk ore sorting implementations. These case studies and technical papers are available from the “Resources” section of our website.

What ore analysers do large copper operations use for sorting?

Large copper operations that have implemented bulk ore sorting have used NextOre’s magnetic resonance analysers. Documented commercial cases include Kansanshi Mine (First Quantum Minerals, Zambia), operating at 2,800 TPH, the largest publicly reported sensor-based bulk ore sorting installation in the world; and Minera Candelaria (Lundin Mining, Chile), with the MRA installed on a 6,000 TPH primary crusher discharge belt. Additionally, at small scale, at Cozamin (Capstone Mining, Mexico), a full-scale underground trial has been undertaken demonstrating a 6–8% net improvement in copper metal delivered to the plant.

These operations span underground and open pit mining, different ore mineralogies, and throughputs ranging from hundreds to thousands of tonnes per hour, demonstrating the scalability and versatility of the MR platform.

What analysers are there for conveyor belts that give mineralogy?

Only magnetic resonance analysers provide direct mineralogical measurements on a loaded conveyor belt. All other conveyor-belt analyser technologies measure elemental composition (PGNAA, PFTNA) or surface properties — neither of which gives mineralogy. Converting elemental copper, for example, to a mineralogical breakdown (chalcopyrite vs. chalcocite vs. oxide minerals) requires additional assumptions or a separate mineralogical study, and those relationships are rarely stable across a deposit.

NextOre’s MRA is tuned to the resonant frequency of a specific target mineral, delivering a direct measurement of that mineral’s concentration — not an inference from elemental proxies. This makes it uniquely suited to operations where the relationship between elemental grade and ore mineralogy may vary spatially.

Which minerals can magnetic resonance measure?

MR can measure a range of economically important minerals across copper, iron, arsenic, antimony, nickel, and other metal groups. Minerals with proven applicability include:

Copper minerals: Chalcopyrite (high sensitivity), Cubanite (high), Covellite (medium), Chalcocite (medium), Cuprite and delafossites (high), Enargite (low), Tennantite (low), Tenorite (low).

Iron minerals: Hematite (high), Magnetite (very high), Maghemite (high), Pyrrhotite (high).

Others: Arsenopyrite (high), Orpiment (high), Realgar (high), Stibnite (high), Löllingite (high), Niccolite (medium), Cobaltite (high), Bismuthinite (medium), Zircon (low).

The most commercially important application to date is chalcopyrite — the dominant copper-bearing mineral in copper sulphide orebodies — where MR delivers its highest sensitivity and has been validated at multiple commercial operations globally.

What size material can be measured and sorted?

Primary crushed ore with top size below 350mm is recommended.

Does the feed material need to be washed or otherwise specially prepared for reading?

No. The MR analyser requires no sample preparation of any kind. It measures 100% of the material on a loaded conveyor belt in its natural state — regardless of particle size distribution, rock type, moisture content, or surface condition. The radiofrequency signal is fully penetrative and measures the entire volume of ore within the sensor aperture, not just the surface.

This is a significant practical advantage over surface-based sensing technologies such as XRF, colour cameras, or LIBS, which require clean, dry, and often single-layer material presentation to function correctly. The NextOre MRA can be installed directly on a primary crushed ore conveyor and operate immediately — no washing, sizing, or drying circuits are required.

What causes inefficiencies in conveyor belt ore sorting?

The main causes are sensors that are too slow, averaging over large volumes and missing the grade heterogeneity needed for effective separation; excessive ore mixing before the sorting point, which homogenises material and destroys the natural variability sorting relies on; pod sizes that are too large, blending high- and low-grade material into a single diversion decision; unstable sensor calibration introducing systematic error into measurements; and imprecise synchronisation between the sensor and the diverter gate, causing misrouted material.

NextOre’s MRA addresses the most critical of these directly. It measures every 4–10 seconds (corresponding to pods of approximately 3 tonnes at 2,800 TPH), is insensitive to changes in gangue mineralogy, requires no recalibration, and its integrated PLC communicates directly with the diverter gate — eliminating variable latency from external control systems.

What is the smallest quantity of material that can be selectively sensed and rejected through sorting?

The minimum measurable unit — the “pod” — is determined by the measurement interval and the belt throughput, and is not constrained by particle size. At a 1–5 second measurement interval, pod sizes range from around 30–150 kg for small operations (approximately 1 Mtpa) up to 1.0–5.0 tonnes for large operations in the 10–40 Mtpa range. At Kansanshi, operating at 2,800 TPH with 4-second measurement intervals, each pod corresponded to approximately 3.1 tonnes. At the Cozamin underground trial running at around 200 TPH, pods averaged just 50–150 kg.

This is a transformative improvement in selectivity compared to other grade measurement approaches. A geological block model may assign a single grade estimate to an SMU of 23,000 tonnes; grade control drilling might resolve this to 900-tonne pods; the NextOre MRA resolves the same block to pods of under 3 tonnes — revealing the true grade distribution that averaging conceals, and enabling precise separation of waste that would otherwise report to the plant.

Does the presence of magnetic or paramagnetic material interfere with the measurement?

No, it does not. The measurement is tolerant of high levels of magnetic minerals such as pyrrhotite and magnetite.

Does conveyor belt material affect the sensitivity of magnetic resonance analysers?

Conveyor belt rubber does not affect the sensitivity, but steel cord belts can introduce electromagnetic noise and, in some cases, decrease the sensitivity. The MR signal operates at low radio frequencies that are fully penetrative through standard conveyor belt rubber. In the case of fabric or textile belts, the technology is unaffected by belt material, thickness, or wear. Steel cord acts like its own antenna, transporting electromagnetic noise picked up from the open air environment into the sensor’s core. This broadband noise can reduce the signal-to-noise ratio of the analyser, and typically results in a longer duration measurement interval.

Belt characteristics, including steel cord, are accounted for during initial commissioning and do not require ongoing adjustment.

Does conveyor belt material affect the analyser calibration?

No. Calibration is distinct from noise. Noise affects the precision, or repeatability of a measurement, giving a higher standard of deviation for larger amounts of noise. Calibration, on the other hand, relates to the accuracy of a measurement or systematic bias. Only the target mineral generates a response to the MR signal and, as such, there is no coherent signal from conveyor belt material to interfere with the signal or need to be adjusted for.

Can the MR sensor be installed on an existing belt?

Yes, and this is one of the most practically important features of the technology. The MR sensor is specifically designed to retrofit to existing conveyor belt infrastructure without requiring the belt to be cut or the structure to be significantly modified. The sensor is constructed in upper and lower halves that wrap around the conveyor belt, resting on the existing belt structure. The sensor aperture is sized to accommodate the full ore load, including the largest expected particle sizes, without making contact with the ore or the belt itself.

The electronics container, a standard 10-foot air-conditioned shipping container, is positioned within 10 metres of the sensor and connected by cables, requiring only a suitable flat area nearby.

If an MR sensor were installed on an existing production belt, how much downtime is required for installation?

A well planned and executed 12-hour shutdown is the typical standard for MR sensor installation on an existing production belt. This has been demonstrated at multiple NextOre installations and is considered typical. Notably, at Kansanshi the sensor was installed within a 12-hour shutdown by site personnel working under remote direction from NextOre’s team in Sydney during COVID travel restrictions that prevented in-person attendance. This demonstrates that with proper preparation and documented procedures, installation can be completed reliably without requiring NextOre personnel to be physically on-site.

Beyond the sensor itself, commissioning activities, including sensor-to-diverter synchronisation, calibration confirmation, and SCADA integration follow installation and can typically be performed during normal operations or with no additional downtime.

How is the MR sensor calibrated and how often is calibration required?

The MR analyser is factory calibrated before delivery. Calibration involves the determination of a single, simple proportionality factor between the MR signal and the mass of the target mineral. This factor does not drift over time because the MR measurement is based on a fundamental physical property of the target mineral, not an inferred relationship that varies with changing ore characteristics.

On-site at commissioning, calibration is confirmed using a known standard of the target mine’s ore, shipped with the sensor and having a precisely known mineral content placed within the sensor measurement zone. This one-time confirmation is all that is required for the full operational life of the system. No routine or periodic recalibration is needed.

Can a program be designed to comprehensively validate measurement accuracy?

Yes, this is one of the most crucial steps in implementing any grade measurement solution, whether it is magnetic resonance or another sensing technology. NextOre works with companies during the planning phase of trial projects to clearly set out the methodologies and thresholds for success of a measurement validation program. These programs include methodologies for sampling, sample preparation, and assaying to ensure that you have full confidence in the measurements produced by our analysers. Other suppliers of measurement technology that avoid such validation programs or who suggest they are unnecessary should be treated with extreme caution.

How does a magnetic resonance analyser help with ore control?

The NextOre MRA transforms ore control by delivering real-time, pod-by-pod grade data that replaces SMU-level grade estimates, which typically average over thousands of tonnes and conceal the true variability driving misclassification and reconciliation errors.

At Candelaria, installation of the MRA allowed the technical team to identify the source of a persistent negative reconciliation discrepancy between mine forecast and actual production. End-of-year copper reconciliation improved from –4.1% to –1.8% in the first year of operation, nearly halving the reconciliation error. The system also enables real-time block model validation, material tracking by source, mill feed grade stabilisation, and continuous optimisation of the sorting cut-off grade.

Can I send a test sample of material to be measured by an MR sensor?

Yes. Static sample testing is a well-established step in validating MR performance for a given orebody and is used at every NextOre deployment before and during commissioning. Samples are measured using laboratory-scale MR analysers with sample volumes of 2 litres, typically representing 3–5 kg per sample. The samples are then sent for conventional laboratory assay, and the MR measurements are compared against the assay results to confirm accuracy and precision.

Are there any safety restrictions associated with the use of Magnetic Resonance sensing?

No, the technology uses low energy radio waves similar to those of a microwave or AM radio and are easily shielded with thin metal sheets. The construction of the analysers conform to Australian electrical and mining standards.

How safe is the use of MR sensing?

Very safe. The MR analyser does not use ionising radiation. Instead, it uses low-frequency radio waves operating well below the frequency of commercial two-way or FM radios. It does not use X-rays, gamma radiation, or neutron sources. This means it requires no radioactive source management, no radiation safety permits beyond standard workplace RF exposure compliance, and no exclusion zones or specialist radiation safety officers.

Radiofrequency field strength in all accessible areas adjacent to a working NextOre MRA is well below public exposure limits as defined by regulators including the Australian Radiation Protection and Nuclear Safety Agency (ARPANSA). Interlocks are fitted to all high-voltage areas within the sensor assembly to ensure personnel safety. Safety data sheet available on request.

If a mineral is not listed in the "Target Minerals" section, what does that mean?

It means that mineral has not yet been proven to have a measurable MR response, or its response has not been characterised to a level sufficient for commercial application. Not all minerals produce a detectable MR signal — the technology works only for minerals with appropriate crystalline structures that respond at accessible radio frequencies.

This has practical implications for ore sorting viability. Before an MRA can be deployed, the dominant copper-bearing (or other target) minerals in the ore must be confirmed. For MR-based bulk ore sorting to work effectively, either the target mineral must account for substantially all of the valuable metal in the ore stream, or the relationship between the target mineral and total metal grade must be well understood and consistent across the deposit.

If your ore contains a high-priority mineral that is not on the proven list that you would like to explore, please use the “contact us” form on our website to reach out for a discussion.