Vacuum Belt Filter Selection: Sizing and Cake Wash

Last updated: 2026-11-01

A vacuum belt filter is the default choice when a hydrometallurgical plant must dewater and wash fast-draining cakes continuously, typically achieving 90-95 percent solute removal in two or three countercurrent washes at 40-70 kPa of vacuum. This guide covers belt filter selection and sizing: duty matching, design parameters, a worked area calculation, and comparisons with filter presses, rotary drum filters, and other solid-liquid separation equipment.

Key Takeaways

• A vacuum belt filter suits fast-settling, free-draining crystalline or granular solids where thorough cake washing drives downstream recovery.
• Size the effective area from a test-confirmed filtration rate, typically 200-2,000 kg/m2·h depending on the material.
• Cake wash efficiency of 90-95 percent solute removal is realistic with two or three countercurrent washes.
• Select cloth permeability, vacuum level, and wash ratio together; they interact strongly and must be test-verified.
• Against a filter press, a belt filter wins on continuous washing and availability; a press wins on final moisture.

Where a Vacuum Belt Filter Fits in a Hydrometallurgical Plant

Solid-liquid separation sits at the hinge of every hydrometallurgical flowsheet: it defines how much soluble metal value reports forward with the liquor and how much impurity stays behind with the solids. For an overview of the full equipment family, see our solid-liquid separation pillar page. Within that family, the horizontal belt filter occupies a specific and valuable niche: continuous operation, in-line cake washing, and moderate capital cost per square meter of area.

The classic duties are easy to recognize. In copper operations across Chile, Peru, Zambia, and the DRC, belt filters wash SX residue and neutralization solids so that acid and copper entrained in the cake return to the circuit rather than reporting to a tailings facility. Zinc plants in Australia, Kazakhstan, and China use them for gypsum from neutralization and for leach residue washing ahead of residue treatment. Nickel laterite and mixed hydroxide projects in Indonesia and the Philippines apply them to crystallized salts and washed intermediates, while gold operations in West Africa and Mexico use comparable horizontal belt machines for washed flotation concentrates and neutralized tailings streams.

The common thread is the material itself. This machine rewards feeds that settle rapidly, form a permeable cake, and drain freely: crystalline gypsum, washed leach residue with a coarse gangue fraction, crystalline salts such as sodium sulfate or magnesium sulfate, and granular flotation concentrates. In each case the cake supports airflow, which is what sustains both the filtration rate and the washing action. Where the feed is fine, slow-settling, or compressible, the economics shift toward pressure filtration, and that evaluation is covered later in this article.

Positioning within the flowsheet matters as much as the machine itself. Most installations receive thickener underflow or classifier sand as feed; sizing that upstream duty is covered in our thickener sizing guide. Because cake structure and permeability respond strongly to fine-particle agglomeration, flocculant conditioning ahead of the belt filter is often the cheapest performance lever available, as described in our flocculant dosage guide. A feed holding tank with agitation, a distribution box, and a headbox complete the wet end. Downstream, the filtrate splits into strong and weak wash liquors, while the washed cake discharges onto a conveyor or repulps into an agitated tank.

How a Horizontal Vacuum Belt Filter Works

A horizontal belt filter is a continuous filtration machine that forms a cake on a traveling filter cloth while vacuum is applied from below, typically operating at 40-70 kPa gauge. Three machine families dominate mining practice, and the differences are worth a brief note.

The rubber-belt filter is the most common heavy-duty design. A reinforced endless rubber belt with drainage grooves and transverse ribs carries the cloth; the belt itself slides over a vacuum box, sealed by low-friction wear belts or water seals. Filtrate passes through the cloth, down the grooves, and into the vacuum box ports. Rubber-belt machines handle the largest areas, widest belts, and most abrasive duties, which is why they dominate copper SX residue and gypsum service in plants from South Africa to Argentina.

The reciprocating tray or pan belt filter replaces the continuous rubber belt with a row of shallow trays that travel forward under vacuum, then retract rapidly to index back to the start of the circuit while the cloth continues moving. Filtrate is kept in discrete trays, which suits applications that require sharp separation between mother liquor and wash liquor, such as crystallized salt washing where filtrate purity grades matter.

The horizontal pan filter discards the belt altogether: a circular rotating table divided into wedge-shaped cells performs the same form-wash-dry sequence. Pan filters suit compact duties and are often seen on niche crystallization services.

Whichever family is selected, the process sequence is identical. Slurry is fed across the cloth in the form zone; vacuum draws liquor through while solids bridge on the cloth and build a cake. The cake then passes under wash weirs or wash bars, where countercurrent wash water displaces mother liquor from the pores. A dry zone pulls air through the cake to reduce residual moisture. At the discharge end the cloth wraps around a small radius roller and the cake cracks off by its own stiffness; the cloth is washed on both faces, tracked, and returned. Cloth tracking, wash weirs, and the vacuum box seal arrangement are the three mechanical details that most influence day-to-day performance, so they deserve close attention during vendor evaluation.

Key Design Parameters for Vacuum Belt Filter Sizing

Belt filter sizing is the exercise of matching three coupled quantities: the solids load the machine must pass, the filtration rate the slurry can sustain, and the geometry that turns those into belt width and length. The machine is therefore never sized from a catalog alone; the parameters below must be fixed by test work

Three parameters deserve elaboration. First, zone layout: the effective length divides into form, wash, and dry zones, and the proportions are a process decision, not a mechanical default. A duty with two wash stages might allocate roughly 30-35 percent of the length to formation, 40-45 percent to washing, and the balance to drying, with the cloth tracking and reclaim section outside the vacuum box. Getting this split wrong is the most common cause of a machine that filters well but washes poorly.

Second, vacuum level and air flow define the vacuum system. Air flow per square meter is the more demanding number: it drives liquid ring pump or blower selection, and it varies enormously between a coarse crystalline salt (high airflow) and a fine residue (low airflow but higher pressure drop). The filtrate receiver, barometric legs or filtrate pumps, and separator vessel must be designed around the worst-case airflow, not the average.

Third, cloth selection closes the loop. A monofilament cloth is a woven fabric that resists blinding because its smooth, round synthetic fibers shed particles at the weave surface, typically offering permeabilities of 300-2,000 L/m2·h. A multifilament cloth is a woven fabric that captures finer particles by providing a denser pore structure, typically trading lower permeability for improved filtrate clarity. In hydrometallurgical service, monofilament polyester or polypropylene cloths dominate because of chemical resistance and long cloth life; the choice between a permeable open weave that maximizes rate and a tighter weave that protects filtrate quality is exactly the trade-off pilot testing exists to resolve. Cloth life in abrasive service ranges from roughly three to twelve months depending on feed abrasiveness and cloth-wash provision.

Filtration Rate and Area Calculation With Worked Example

The sizing method is short: fix the dry solids yield, apply a filtration rate from test work, correct for availability, and select standard geometry. Filtration rate is expressed as dry kilograms of cake solids deposited per square meter of effective area per hour, and it is the single most site-specific number in the whole selection. The table below gives indicative ranges for the materials most often filtered on horizontal belts; every value is typical and must be confirmed on representative, well-conditioned samples.

The calculation proceeds in four steps.

1. Fix the dry solids yield. Take the design tonnage of dry cake solids from the mass balance; for this example, 15 t/h (15,000 kg/h) of washed SX residue solids.
2. Apply the test-confirmed filtration rate. Pilot work on the actual residue gave 800 kg/m2·h at 60 kPa, so the net required area is 15,000 ÷ 800 = 18.75 m2.
3. Correct for availability and turndown. Continuous belts need cloth washing time and tolerate occasional feed upsets; applying an 85 percent availability factor gives 18.75 ÷ 0.85 = 22.1 m2.
4. Select standard geometry and verify. The nearest standard selection is a 2.4 m wide belt with 10.0 m of effective vacuum length, giving 24.0 m2. The actual design rate is then 15,000 ÷ 24.0 = 625 kg/m2·h, a comfortable margin. At a 12 mm cake and a wet bulk density of about 1,600 kg/m3, the cake load is roughly 19.2 kg/m2 per pass, so the belt speed needed is 625 ÷ 19.2 ≈ 33 m/h, or about 0.55 m/min, comfortably inside the typical 0.5-8 m/min range and leaving headroom to slow down for thicker cakes or speed up during upset duty.

The vacuum system follows from the area: at an indicative 0.9 m3/m2·min, 24 m2 demands about 22 m3/min of air at 60 kPa, pointing to a liquid ring vacuum pump of roughly 50-60 kW installed power. Every number in this chain is indicative until confirmed on your material; our pilot-scale vacuum filtration unit and lab-scale rotary drum vacuum filter exist precisely to turn these ranges into project-specific design data before an equipment commitment is made.

Vacuum Belt Filter vs Filter Press vs Rotary Drum: What to Choose?

Process engineers rarely get to choose the machine in isolation; they choose a flowsheet behavior. The table below compares the three principal options on the criteria that actually decide projects.

The per-scenario verdicts follow directly from that comparison. For SX residue and neutralization gypsum where soluble loss drives reagent consumption, a vacuum belt filter is preferred because countercurrent washing happens continuously on the belt, at a wash efficiency that batch presses struggle to match without repulping circuits. For fine, slow-settling leach residues or feeds that need low final moisture below roughly 10 percent, a membrane filter press is preferred because pressure filtration overcomes the high cake resistance that vacuum cannot, as detailed in our filter press for mining guide and the pressure filter vs filter press comparison. For high-tonnage crystalline salts where washing is light and rate matters most, a rubber-belt vacuum filter is preferred because it delivers the highest airflow and rate per unit area with minimal operator attention. Where the duty is only primary dewatering ahead of a CCD circuit or tailings disposal and no cake washing is needed, a thickener-only flowsheet is preferred because settled underflow at 50-65 percent solids avoids buying filtration area at all. Finally, when a vertical flowchart decision is genuinely close, run the numbers per tonne: a belt filter’s lower mechanical complexity usually wins on operating cost, while a press wins where moisture spec, not washing, is the binding constraint. Machines do not substitute for flowsheet logic; our quoting process starts from your mass balance, not from a catalog.

Why Does Cake Wash Efficiency Fall Off?

A horizontal belt filter is sized to wash, and most field disappointments trace to a handful of recognizable mechanisms. If solute removal drops below the 90-95 percent design expectation, work through this list in order.

1. Cake cracking. Fast-draining cakes shrink and crack as they dewater in the form zone, and wash water then short-circuits down the cracks instead of displacing liquor from the cake body. Remedies include slowing the belt to build a thicker, more plastic cake, feeding ahead of the first wash weir, and gentle cloth-side drainage control. Cracking is most severe on coarse gypsum and crystalline salts.
2. Wash maldistribution. Wash liquor must arrive as an even, low-energy curtain across the full cake width. A single open pipe, a scaled wash bar, or a tilted weir starves one side of the belt and over-washes the other, so composite wash efficiency collapses even though average wash ratio is on target. Inspect the wash weirs during every cloth change.
3. Insufficient wash ratio. Displacement washing has a physics limit: below roughly 1.2 kg of wash per kilogram of dry solids, most of the wash simply fills pore volume without displacing mother liquor. If the mass balance squeezes wash ratio below that, no mechanical fix will recover the efficiency; add a wash stage or accept the soluble loss.
4. Cloth blinding. A cloth that has lost permeability drains unevenly, so the cake above it forms at inconsistent thickness and washes unevenly. Because blinding is a large topic in its own right, it is treated separately in our filter cake moisture resource and cloth lifecycle notes; the short version is that sustained cloth-wash spray pressure and periodic chemical cleaning are mandatory, not optional.
5. Air bypass at the vacuum box. Wear belts, seals, and tray lips age. When the vacuum box draws air around the cake instead of through it, wash vacuum collapses locally and the wash zone loses displacement energy. Rising vacuum pump load with falling cake vacuum is the diagnostic signature; re-shimming or replacing wear components restores it.
6. Zone layout drift. Plants that later increase throughput often speed the belt, silently shrinking residence time in the wash zones. If a machine that once washed at 94 percent now holds 88 percent at the same wash ratio, measure the actual zone lengths at current speed before blaming the cloth or the water.

What Does Ownership Cost?

The operating cost of a horizontal belt filter concentrates in three lines: vacuum energy, wash water, and cloth replacement. Vacuum pumping is the dominant consumer. At the airflow levels typical of mining service, expect an indicative 2-5 kWh per tonne of dry solids for the vacuum system; a 24 m2 machine filtering 15 t/h at 50 kW installed pump power lands at about 3.3 kWh/t. Energy cost therefore tracks airflow more than area, which is another reason coarse crystalline duties need careful pump selection.

Wash water consumption follows the wash ratio directly: 1.0-2.5 t of water per tonne of dry solids, typically drawn from process water of controlled quality so that wash efficiency is not degraded by the water itself. In countercurrent layouts, most of that water returns to the circuit with the filtrate, so net consumption is far lower than the gross figure suggests. Cloth life spans roughly three to twelve months depending on feed abrasiveness, cloth-wash provision, and chemical environment, and a cloth set for a large belt is a meaningful line item worth negotiating into the supply package.

Maintenance is dominated by wear components: belt wear strips, vacuum box seals, cloth tracking rollers, and spray nozzles. Well-maintained rubber-belt machines in Chilean and Zambian copper plants routinely exceed twenty years of service, and availability of 85-90 percent is a realistic planning figure. Labor is light; one operator can oversee several belts from a control room. Set against the alternative of a filter press with its batch cycles, cloth change labor, and plate maintenance, the belt filter usually wins the cost comparison wherever the feed suits it, and the decision framework in the filter press guide linked above shows where that logic reverses.

Operating Data From Installed Belt Filters

The following anonymized figures are drawn from horizontal belt filters supplied into mining circuits. All values are typical and indicative, intended to calibrate expectations rather than to serve as design data; test work on your own material remains the only defensible basis for sizing.

Three patterns in this data are instructive. The crystalline salt, being coarse and extremely free-draining, filters at double the gypsum rate and dries to low moisture with a single wash. The SX residue, containing the finest solids of the three, needs the deepest vacuum and accepts the highest moisture, yet still reaches 93 percent wash efficiency with two countercurrent stages because the cake remains permeable after flocculant conditioning. In every case the machines ran near 87-90 percent availability with cloth life between five and nine months, confirming that the ranges quoted earlier in this guide are achievable in routine industrial service rather than only on test pads.

Frequently Asked Questions

How much vacuum does a vacuum belt filter need?
Most hydrometallurgical duties run at 40-70 kPa of gauge vacuum. Coarse crystalline salts and gypsum filter well at 40-55 kPa, while finer leach residues need 55-70 kPa to sustain rate. Deeper vacuum rarely helps, because beyond the cake’s capillary entry pressure the extra energy only pulls more air through cracks. Confirm the level with test work, then size the pump for airflow, not vacuum depth.

How is belt filter sizing done before test results exist?
Use an indicative rate from the table in this guide to build a budget-size area, then confirm on representative samples. Treat catalog or handbook rates as plus or minus 50 percent until pilot work on your actual slurry, at your target pH and temperature, replaces them. Area is cheap to overbuild slightly; a machine 20 percent undersized constrains plant throughput for its whole life.

How many wash stages does cake washing require?
Two countercurrent washes typically achieve 90-93 percent solute removal, and three stages reach 95 percent or better on permeable cakes, at indicative wash ratios of 1.0-2.5 kg per kilogram of dry solids. One stage rarely exceeds 80 percent. The right number comes from washing test curves: if the third stage adds little, the soluble loss is better attacked upstream.

What vacuum filter cloth should be selected?
Start with monofilament polyester or polypropylene for chemical resistance and blinding resistance, then choose weave permeability against the finest particle fraction in your feed. A cloth is a compromise that balances filtration rate, filtrate clarity, and cloth life; the correct weave is the one pilot testing shows forms a cake reliably at design speed. Multifilament weaves suit fine feeds that need clearer filtrate and accept lower rates.

Can a belt filter handle fine or slow-settling solids?
Poorly. If the slurry does not settle within seconds and form a permeable cake, vacuum cannot overcome cake resistance, and the machine washes unevenly. Such feeds belong on filter presses, where pressure supplies the driving force, or on thickener and CCD circuits. Aggressive flocculation sometimes rescues marginal feeds, but only test work tells you whether the cake holds together at belt speed.

How long does the filter cloth last in mining service?
Indicative life is three to twelve months. Coarse, abrasive gypsum and salt duties with good cloth washing sit at the long end; sharp, angular residues with poor washing sit at the short end. Track cloth condition by filtrate clarity and cake release, budget one full cloth change per inspection interval, and insist on two-sided cloth washing sprays in the machine specification.

What cake moisture can this machine achieve?
Expect roughly 5-12 percent on crystalline salts, 15-25 percent on gypsum, and 18-28 percent on fine leach residues, all indicative at 40-70 kPa. Moisture falls with thinner cake, deeper dry-zone time, and higher airflow, and rises with fine solids. If your specification demands below about 10 percent on a fine residue, plan for pressure filtration instead, since vacuum cannot reach that moisture on such feeds.

Conclusion

Selecting a vacuum belt filter is fundamentally an exercise in honest test work: fix the filtration rate on your own slurry, respect the interaction between vacuum level, cloth permeability, and wash ratio, and only then choose belt width, length, and zone layout. The machine rewards fast-settling, free-draining feeds with continuous, high-efficiency cake washing that no batch filter matches, and it punishes optimistic rate assumptions with chronic under-capacity. Where washing drives your flowsheet economics, the belt filter is usually the right answer; where moisture spec or fine solids dominate, let the comparison guide you to a press instead. Lexin-tech supplies horizontal belt filters, pilot units, and complete solid-liquid separation packages for mining operations worldwide. Send us your mass balance through our request-a-quote page, or explore more engineering guides in our resources library.

References

• Rushton, A., Ward, A. S., and Holdich, R. G., Solid-Liquid Filtration and Separation Technology, 2nd edition, Wiley-VCH.
• Svarovsky, L., Solid-Liquid Separation, 4th edition, Butterworth-Heinemann.
• Wakeman, R. J. and Tarleton, E. S., Filtration: Equipment and Modelling, Elsevier.
• Dahlstrom, D. A., et al., Liquid-Solid Filtration and Separation, SME Mineral Processing Handbook section.
• Havlíček, J., et al., countercurrent cake washing theory papers in Minerals Engineering journal.
• Tarleton, E. S. and Wakeman, R. J., Solid/Liquid Separation: Equipment Selection and Process Design, Butterworth-Heinemann.

Changelog

• 2026-11-01: First publication.