Why should the middlings from a tantalum-niobium ore shaking table be collected separately-for regrinding, further separation, or processing via a three-disk magnetic separator?
When conducting tests on tantalum-niobium ores, the component most frequently overlooked is not the concentrate or the tailings, but rather the "middlings" band-that strip in the middle of the shaking table deck which appears neither particularly dark nor particularly light. If this fraction is simply returned to the feed, the circulating load will continue to grow; conversely, if it is incorporated directly into the concentrate, it may introduce unwanted tin, iron-bearing minerals, and gangue. Whenever I examine a shaking table in operation on-site, my first step is to isolate the middlings; they provide the clearest insight into whether the process bottleneck lies in mineral liberation, classification, or the final cleaning stage.

"Middlings" do not refer to a fixed, specific material, but rather to a heterogeneous mixture of mineral particles. Middlings recovered from a shaking table may contain a combination of minerals such as columbite-tantalite, cassiterite, garnet, iron-bearing minerals, and composite grains (intergrowths) of quartz and feldspar. Because the composition varies significantly depending on the ore deposit, one cannot determine the appropriate downstream processing route based on color alone. Instead, the middlings should be classified by grain size-coarse, medium, and fine-and examined using tools such as a handheld magnet, a microscope, or mineralogical analysis. The objective is to determine whether the valuable minerals exist as liberated (monomineralic) grains, as intergrowths, or as particles encapsulated within fine-grained gangue (slimes).

First, determine whether the coarse-grained middlings require regrinding.
If the valuable minerals in the middlings remain tightly intergrown with quartz or feldspar, repeated cleaning stages will merely cause the material to shift back and forth between the concentrate and the tailings. In such cases, small-scale regrinding tests should be conducted to evaluate the degree of mineral liberation and the increase in fine-slime content at various grinding durations. Tantalum-niobium minerals, while dense, can also be brittle; over-grinding may cause recoverable particles to report to the fine-slime fraction. Therefore, it is more appropriate to regrind only those middlings that specifically require further liberation, rather than subjecting the entire gravity-separation product stream to regrinding.
Adjust Shaking Table Operating Parameters When Significant Amounts of Liberated Particles Are Present
If microscopic examination reveals a substantial quantity of liberated tantalum-niobium minerals in the middlings, it is necessary to re-examine the shaking table's feed size distribution, pulp density, deck transverse slope, and wash-water flow rate. A wide feed size distribution can cause coarse gangue particles and fine heavy minerals to crowd into the same intermediate zone, while fluctuations in feed density can cause the boundaries between separation zones to shift. By classifying the feed into narrower size fractions and stabilizing the flow rate-followed by another three-product separation run-it is often possible to determine whether the issues stem from the operating parameters.

The three-disk magnetic separator is designed for processing dry, narrow-size-fraction products.
Magnetic separation using a three-disk unit relies on differences in mineral magnetic susceptibility to achieve further fractionation. This process requires the material to be dry and loose, with a thin, uniform feed layer. Wet middlings from a shaking table cannot be fed directly into the magnetic separator; otherwise, particle agglomeration will mask the differences in magnetic properties. Such material must first be dewatered and dried, then tested in batches according to specific size fractions, with the magnetic field intensity and the positioning of the discharge splitters adjusted for each disk. Magnetic separation is not a substitute for gravity separation; rather, it is a method for processing complex heavy-mineral concentrates that have already been enriched and contain minerals with varying magnetic properties.
The three branch circuits should be evaluated through closed-circuit testing. A reliable testing approach involves dividing the middlings into three separate streams: (1) regrinding followed by further beneficiation, (2) direct further beneficiation, and (3) dry magnetic separation. For each stream, data should be recorded regarding product yield, Ta-Nb grade, and the distribution of associated components such as tin and iron. If direct further beneficiation is sufficient to significantly narrow the middlings fraction, then additional grinding is unnecessary. Conversely, if regrinding improves mineral liberation but results in excessive losses in the fine-slimes fraction, the grinding parameters must be re-evaluated. Ultimately, the process flow should ensure that the quantity of middlings stabilizes-converging to a steady state-rather than continuously accumulating within the system.







