Coffee Cherry Classifier

How does flotation actually separate ripe from unripe cherry?

The separation is based on the density of the whole cherry relative to water (specific gravity 1.00 g/mL). Ripe coffee cherry has a specific gravity of approximately 1.05–1.10 g/mL — it is denser than water and sinks. Unripe green cherry has not yet completed cellular development; its specific gravity is below 1.00 g/mL — it floats. Overripe and fermenting cherry has undergone internal breakdown that reduces its density; it also floats. This density difference exists because the ripe cherry bean (seed) is fully formed and dense, and the ripe pulp is fully sugar-developed and firmer. A coffee berry borer-damaged cherry has internal voids and also floats. The flotation channel simply provides a water medium in which this density difference becomes a spatial separation — dense cherry sinks and is carried along the bottom; light cherry rises to the surface and is diverted off.

What percentage of cherry typically floats, and what should I do with the floaters?

Under good selective harvesting conditions (cherry picked at peak ripeness), floater rates of 5–10% are typical. In mixed or strip-harvested lots, floater rates of 15–25% are common; in poorly managed or late-harvest lots, floater rates above 30% have been recorded. For the floater stream: (1) Green unripe floaters — best held separately for 5–10 days and re-assessed; in some estates, these are re-submitted after a rest period. (2) Overripe floaters with good cherry integrity — process as natural/dry on raised beds; their higher initial sugar may produce interesting cup notes. (3) Damaged, fermented, or burst floaters — compost. The key principle is to treat floaters as deliberate by-product streams, not as mixed waste. Each category should have a documented handling decision made before the harvest season begins.

Does cherry classification improve the outturn (percentage of premium grade) from a lot?

Yes, measurably. The primary improvement is in quaker (unripe bean) defect counts in the green coffee. Quakers form when unripe cherry is processed through the wet mill — the immature, starchy seed does not roast correctly and appears as a pale, underdeveloped 'quaker' bean in the roasted lot. SCA and Coffee Board defect standards both include maximum quaker counts for specialty and premium grades. Removing unripe cherry at the classifier stage directly reduces quaker frequency. Secondary improvements include lower partial-fermentation defect counts (fewer sour beans) and a more uniform final lot that performs better at gravity separation and colour sorting — each downstream machine works more effectively on a cleaner, more uniform input. Estates that install cherry classifiers typically report a 5–12% improvement in AA/specialty outturn by weight across a season.

How much water does a cherry classifier use?

Gross water flow is high — a continuous water channel requires sufficient flow to carry cherry and maintain flotation, which means several hundred litres per minute on a commercial machine. However, net water consumption (fresh water drawn in to replace losses) is much lower when a recirculation circuit is used. With the water recycling sedimentation channel included in CC-S2000 and larger models, net water consumption is typically 30–60 litres per tonne of cherry processed. Without recycling, fresh water consumption can be 150–300 litres per tonne. Water conservation matters in Indian coffee regions, particularly in areas where dry-season water access is restricted. The sedimentation channel also reduces organic load in the mill's wastewater circuit — cherry and mucilage solids settle in the sedimentation channel and are removed as semi-dry organic matter rather than being flushed to the effluent pond.

Can the cherry classifier replace the trommel grader, or do I need both?

They perform different and complementary functions — both are needed in a complete wet-mill intake line. The trommel grader separates by physical size: it removes undersized pin-cherries, soil, and large debris that would interfere with flotation accuracy. The cherry classifier separates by density (maturity): it separates ripe from unripe cherry within the size range that passed the trommel. If you run cherry directly to the classifier without trommel pre-grading, small stones and pin-cherries will sink along with ripe cherry (stones are denser than water) and enter the pulper — which is exactly what the trommel was designed to prevent. Conversely, the trommel alone does not distinguish ripe from unripe cherry of the same size — that requires flotation. The standard intake sequence is: receiving hopper → trommel grader → cherry classifier → pulper.

What is the difference between a cherry classifier and a cherry flotation tank?

They refer to the same functional concept — using water flotation to separate cherry by density/maturity. A flotation tank is a simpler, older implementation: a static tank of water where cherry is fed in batch or continuously and floaters are skimmed manually. A cherry classifier (as supplied by VMAC) is a more engineered, continuous-flow version: it includes a mechanical pre-screen to remove debris first, a controlled-flow channel rather than a static tank, an adjustable overflow weir for consistent floater removal, and a water recycling circuit. The classifier delivers more consistent and repeatable density separation than a simple flotation tank, handles higher throughput, and requires less manual intervention. For commercial-scale wet mills, the engineered classifier is preferable; for very small micro-mills with low throughput, a simple flotation channel may be sufficient.

Does cherry classification affect cup quality, and how would I verify this?

Yes — the cup impact is measurable. The most direct evidence is comparing quaker frequency (count of pale underdeveloped beans in the roasted lot) in classified versus unclassified lots from the same farm block. Classified lots typically show quaker counts 30–60% lower than unclassified lots from the same harvest. At the cupping table, classified lots generally show cleaner, sweeter profiles with reduced grassy or starchy notes that characterise lots with unripe cherry contamination, and reduced sour or fermented notes from overripe cherry. For specialty or GI-certification purposes, submitting paired cupping sets — classified and unclassified from the same block and harvest date — provides clear evidence of the classification benefit. Most Indian specialty roasters and international green buyers now expect cherry classification as a baseline for Arabica lots submitted for scoring above 83–84 SCA points.

How do I calibrate the weir gate depth for accurate floater removal?

The weir gate depth (overflow height) in the flotation channel determines the boundary between the sinking product fraction and the floating reject fraction. Set it too shallow — barely below the water surface — and you capture only the clearest, most buoyant floaters; marginal floaters may be carried along with the sinking fraction. Set it too deep — at mid-channel depth — and you risk drawing sinking ripe cherry over the weir, reducing ripe cherry yield. The correct starting point is to set the weir at approximately 70–80% of the water depth. Then observe the sinking cherry stream: if you see obviously unripe green cherry arriving at the pulper discharge, the weir is too deep and the channel is not diverting floaters adequately. If cherry yield from the classifier is significantly lower than expected, the weir may be too shallow and is capturing ripe cherry as floaters. Adjust in 20 mm increments and allow 10–15 minutes of steady-state operation before re-assessing. VMAC's adjustable stainless steel weir plates allow fine depth adjustment without tools.