Most sink-float separation tanks are bought on a single number — a target throughput in kg/h — and that is exactly why so many of them underperform once real post-consumer material starts flowing. A tank that looks correctly sized on paper can still hand you contaminated flake, runaway water bills, or a sludge problem nobody planned for. Sizing a sink-float tank well means matching it to your feedstock behavior, your purity target, and the machines on either side of it — not just to a capacity figure. This guide walks through how to do that before you issue an RFQ.
What a sink-float tank actually does (and what it can’t)
A sink-float separation tank — also called a float-sink tank, float washing tank, or floating washer — separates plastics purely by density relative to water (1.0 g/cm³). Lighter-than-water polymers float and are skimmed off the surface; denser polymers and heavy contaminants sink and are dragged out at the bottom. There are no screens or sensors deciding the split — gravity and buoyancy do the work, which is why the tank is so reliable and cheap to run, and also why it has a hard limit.
| Material | Densidad (g/cm³) | Behavior in water |
|---|---|---|
| PÁGINAS | 0,90–0,92 | Floats |
| PE (HDPE / LDPE) | 0.91–0.96 | Floats |
| PD | 1.04–1.07 | Sinks |
| abdominales | 1.04–1.07 | Sinks |
| MASCOTA | 1.38–1.40 | Sinks |
| PVC (rigid) | 1.35–1.45 | Sinks |
The takeaway for sizing: a water tank cleanly splits floaters (PP/PE) from sinkers (PET/PVC/PS/ABS), but it cannot separate two materials that both sink. PET and PVC have overlapping densities, so both go to the bottom together. If your spec calls for sub-50-ppm PVC in rPET, you will need an electrostatic or NIR polishing step after the tank — size for that now, not later.
Step 1: Profile your feedstock before anything else
The single most useful thing you can do before sizing is characterize what is actually entering the tank. Three variables matter most:
- Polymer mix. A clean PP/PE float stream behaves completely differently from a mixed rigid post-consumer stream. The wider the density spread you need to capture, the more separation passes you want.
- Bulk density and form. Light film and fluff sit on the surface and need strong, well-placed paddles to submerge and convey; heavy rigid flake sinks fast and loads the bottom screw harder.
- Contamination load. Paper labels, glue, sand, fines, and organic residue all end up in the water and the sludge. Heavy contamination is the main reason a tank that is “big enough” on throughput still fails on purity.
If you can, run a float-sink test on a representative sample in a bucket of water. It tells you your real float:sink ratio, which directly drives skimmer and bottom-discharge sizing.
Step 2: Fix your real throughput band — not your peak
Throughput should be expressed as a band tied to a specific material, not a single headline number. A tank rated “2000 kg/h” on clean PP regrind will not hold that rate on wet, contaminated post-consumer PET, because residence time — how long each flake stays in the water — is what determines separation quality. Push material through too fast and floaters get dragged down before they surface.
As a planning reference, typical single-tank bands fall roughly into:
- 300–1,500 kg/h — most film and bottle wash lines, clean to moderately contaminated feed.
- 1,500–3,000 kg/h — higher-capacity lines and heavily contaminated post-consumer streams (usually a triple-row tank, see below).
- 3,500–5,000 kg/h — industrial bottle-to-bottle plants, typically with multiple or extended-length tanks.
Actual capacity always depends on feed geometry, bulk density, contamination, and how tight your purity target is — so treat these as starting points for an engineering conversation, not guarantees.
Step 3: Single-row vs triple-row — the decision that drives footprint
This is usually the real sizing decision. A single-row tank gives each flake one pass through the water. A triple-row sink-float tank makes material travel three times the distance and soak three times as long in one chassis, which is what lets it hit higher purity on dirty feed without lining up three separate tanks in series.
| Criterios | Single-row tank | Triple-row tank |
|---|---|---|
| Throughput band | 300–1,500 kg/h | 1,000–3,000 kg/h (one unit replaces ~3 single-row tanks in series) |
| Separation passes | 1 pass | 3 passes — 3× longer travel and soak time |
| Typical purity | 98–99% in one stage | 99.5%+ even on heavy post-consumer contamination |
| Best feedstock | Clean post-industrial regrind; pre-sorted PET/HDPE bottle flake | Heavily contaminated post-consumer PET, mixed rigids, WEEE, automotive ASR |
| Potencia instalada | ~5.5–15 kW | ~15–22 kW |
Rule of thumb: if your line target is above 1,500 kg/h, or your feedstock comes from curbside or DRS bales, the triple-row configuration usually pays back its premium within the first year through purity uplift and far fewer extruder screen-filter changes downstream.
Step 4: Size the water loop and sludge handling, not just the tank
The tank is the cheap part. The water and sludge system around it is where operating cost is won or lost, and it is routinely under-specified.
- Water consumption. Without a treatment loop, a tank typically draws 20–30 m³ of fresh water per ton of throughput. Adding a DAF (dissolved air flotation) and sedimentation loop typically cuts that to 3–5 m³ per ton — up to ~85% reuse. Over a year that difference dwarfs the price gap between tank models.
- Sludge. Paper fiber, label glue, fines, and grit have to go somewhere. A dedicated screw press de-waters the sludge into a 60–70% solid cake that is cheap to haul away, instead of leaving you with a liquid-waste headache.
Specify the water loop and sludge press in the same RFQ as the tank. Bolting them on afterward almost always costs more and fits worse.
Step 5: Place the tank correctly in the wash line
A sink-float tank is a separation stage, not a standalone machine. It needs the right feed condition coming in and the right handoff going out:
- Upstream: material should already be size-reduced and pre-washed. A arandela de fricción o hot wash system ahead of the tank removes the bulk of glue and surface dirt so the float-sink split is clean.
- Downstream: separated flake still carries water. Plan for dewatering and drying before pelletizing, and — if you are chasing food-grade rPET — the electrostatic or optical PVC-removal step mentioned earlier.
If you are scoping a full line rather than a single machine, start from the washing system level and let the tank size fall out of the line target.
Quick-reference sizing checklist
Before you send an RFQ, have answers to these:
- What polymers are in the feed, and what is the approximate float:sink ratio?
- What is the realistic throughput band for that material — not your best-case peak?
- What purity does your buyer or extruder actually require?
- Is the feed clean post-industrial, or contaminated post-consumer? (This usually decides single- vs triple-row.)
- Are water reuse (DAF) and a sludge press included in the scope?
- What machine feeds the tank, and what machine receives the separated streams?
Five sizing mistakes that cost you later
- Sizing on peak kg/h. Tiempo de residencia, no capacidad de placa, determina la pureza. Tamaño para una tasa sostenida en tu alimentación más realista de suciedad.
- Esperar que el agua sola elimine PVC de PET. No puede — ambos flotan. Planifique la etapa de pulido por adelantado.
- Ignorar el ciclo de agua. Un tanque sin DAF puede costar más en agua fresca y tarifas de descarga que el tanque en sí.
- Olvidar el lodo. Sin desague, significa un costo recurrente de residuos líquidos y tiempo de inactividad.
- Comprar el tanque en aislamiento. El mal lavado upstream o el secador downstream hace que un tanque perfectamente dimensionado parezca que ha fallado.
Preguntas frecuentes
¿Puede un tanque de hundimiento-flotación separar completamente PET de PVC?
No. PET (1.38–1.40 g/cm³) y PVC (1.35–1.45 g/cm³) ambos flotan con densidades superpuestas, por lo que un tanque de agua los envía al fondo juntos. Para menos de 50 ppm de PVC en rPET, agregue un separador electrostático o un clasificador óptico NIR como etapa de pulido después del tanque.
¿Cuánta agua consume un tanque de hundimiento-flotación?
típicamente 20–30 m³ de agua fresca por tonelada sin filtración. Con un DAF y un ciclo de sedimentación, esto generalmente baja a 3–5 m³ por tonelada (hasta ~85% reciclado), lo que es donde se generan la mayoría de los ahorros en costos operativos.
¿Cuándo un tanque de triple hilera vale el costo adicional?
Generalmente cuando su línea opera por encima de ~1,500 kg/h o su alimentación es material contaminado post-consumidor (baldes de acera, bales DRS, materiales rígidos mezclados). Las dos pasadas adicionales aumentan la pureza y reducen los cambios en el filtro downstream, generalmente pagando de vuelta en el primer año.
¿Dónde se sitúa el tanque en la línea?
Después de la reducción de tamaño y lavado previo/fricción o lavado caliente, y antes del desague, secado y granulación. Una alimentación limpia significa una división de flotación-hundimiento limpia.
¿Necesita ayuda para ajustar un tanque a su alimentación y objetivo de línea específicos? Compartir su mezcla de polímero, nivel de contaminación y objetivo de rendimiento, y nuestros ingenieros pueden recomendar una tanque de separación hundimiento-flotación configuración — de una o triple hilera, con el ciclo de agua y el manejo de lodo adecuados — dimensionada en torno a su proceso real en lugar de un número en grandes letras.
