Dense Medium Separation

Cut Point Control in Heavy Media Cyclone Circuits: SG Drift and Correction Our Dense Media Cyclone product page states a separation efficiency of 95% or better and an Ep of 0.03 or lower. Those numbers are real and achievable, but they describe what a correctly operating cyclone delivers, not a guarantee that holds regardless of how the circuit around it is run. Cut point and Ep both drift in service, and understanding what erodes them is what actually keeps that spec true on your plant, not just on the datasheet. What Ep and Cut Point Actually Measure Cut point d50 is the density at which a particle has an equal chance of reporting to either the clean coal or the reject stream. It’s the number your washability curve is built around, and it determines yield and ash at a given density target. Ep, the probable error, measures how sharp that separation actually is. Take the density at which 75% of particles report to sink and the density at which 25% report to sink, and Ep is half the difference between them. A lower Ep means less material is being misplaced near the cut point of coal that should report to product staying in reject, or higher-ash reject slipping into product. Two circuits can hold the same d50 and produce very different yield and ash outcomes if their Ep values differ. Both numbers matter together. A cyclone holding its design cut point with a widening Ep is still losing yield and ash performance, even though the density target on the gauge hasn’t moved. This is also why Ep alone, without cut point context, can be misleading when comparing circuits. A cyclone running a low cut point with a moderate Ep can still misplace more total tonnage than one running a higher cut point with a genuinely sharp Ep, simply because more of the feed sits near the separating density in the first case. Reading the two numbers together against your actual washability curve tells you whether a given Ep is actually good enough for your coal. What Erodes These Figures in Service Medium SG drift is usually the first point to consider separately against your actual washability curve. Density drift shifts the achieved cut point directly, even when the cyclone itself is mechanically sound. If your medium density isn’t holding at the target you designed the cut point around, no amount of cyclone geometry corrects for it. Medium rheology is the second factor, and it’s driven largely by magnetite granulometry and solids concentration rather than anything happening inside the cyclone body. Poor medium viscosity widens Ep directly because particles don’t respond cleanly to the density gradient the cyclone is trying to create; the medium itself is behaving inconsistently before separation even happens. We go into the specification numbers that hold this steady in our post on magnetite quality and granulometry. Cyclone wear changes the internal geometry the design assumes. Liner wear, particularly at the spigot, gradually alters the flow pattern inside the cone. A worn spigot that’s grown beyond its design diameter changes the underflow density the cyclone can hold, which widens Ep well before the wear becomes visually obvious from outside the unit. Feed pressure outside the design range works against separation from a different angle. Heavy media cyclones are sized for a specific feed head run below it and centrifugal force drops, widening EP run. meaningfully above it, and turbulence increases inside the cone, which also degrades separation sharpness rather than improving it. When spigot capacity is exceeded, roping becomes visible, where the underflow discharge loses its normal spray pattern and instead ropes out in a dense, coherent stream, which is the clearest sign that something upstream has pushed more solids through the spigot than it can properly discharge. Once roping starts, Ep degrades quickly, and it’s usually a downstream symptom of one of the other four causes rather than an independent fault. The vortex finder-to-spigot diameter ratio sits behind several of these symptoms at once. That ratio is what the cyclone’s design assumes when balancing overflow and underflow splits, and it’s fixed by the geometry John Finlay specifies for a given duty. Spigot wear changes the ratio directly. Feed pressure outside the range changes the effective split even with the physical geometry unchanged. Medium rheology issues change how the solids actually distribute across that split regardless of what the ratio nominally is. None of these show up as a single obvious fault; they show up as the ratio no longer doing the job it was sized for, and Ep is usually the first number to reflect that. Five factors that erode separation efficiency in service Diagnosing Which Factor Is Yours Check the medium SG trend first. It’s the cheapest and fastest check, and if density has been drifting, that’s very likely where cut point movement is coming from before anything else needs investigating. Check magnetite specification against your certificate of analysis next, particularly if SG is holding steady but Ep itself has been widening independent of cut point. A circuit can hold its density target while running a medium that’s fundamentally worse at supporting a sharp separation. Inspect spigot diameter directly rather than relying on visual wear alone. A spigot worn a few millimeters beyond spec can shift underflow density meaningfully without looking obviously different from a normal one. Confirm feed pressure against the cyclone’s design specification, since pump wear and circuit changes elsewhere can quietly shift operating pressure away from where the cyclone was sized to run. Watch the underflow discharge for roping during normal operation. If it’s present, treat it as an urgent signal rather than a cosmetic issue; it means the circuit upstream is already overloading the spigot’s capacity. Work through these in order rather than jumping to the one that seems most likely. Because several of these factors interact, worn spigot geometry compounds with feed pressure drift, and off-spec medium compounds with both a plant that assumes it knows the cause and jumps