Below-size and light beans in a consignment are a direct consequence of inadequate size and density separation, partly during primary processing but mostly during dry or export processing. Not only do too many smalls and lights spoil the coffee's green appearance, but large and small, or heavy and light beans also do not roast well together. This is because smalls and lights will over-roast during the time it takes for the roasting of the larger, heavier beans to be completed. There are strict limits to the proportion of smalls and lights roasters may tolerate in whole bean coffee; if a coffee exceeds their in-house tolerance for smalls and lights, then out it goes.Not all size grading is accurate. Opinions differ on the accuracy of different size grading techniques (vibratory or flatbed versus rotary or cylinder graders for example) and this is not the place to argue for or against any of them. But, often, when operated at full design capacity, graders do not necessarily produce accurate separation, so the throughput must be regulated. This can be especially troublesome if a grader is directly auto-fed by a preceding processing unit, or if the product quality is quite variable. It is always advisable therefore to have a supplier commission any new milling, grading and sorting equipment, using the actual product that is to be handled. Regulating the intake flow by placing a buffer silo or feed hopper ahead of the grader can improve grading accuracy quite considerably, but constant supervision will always be necessary. The grading accuracy should be verified regularly, using hand or sample screens that should be kept handy, near the grader.When grading whole bean type coffee bear in mind that some very large beans may not be particularly attractive as they are often soft or misshapen. Such beans become especially noticeable in the roast appearance. They can be easily removed by the insertion of a large size screen (number 20 screen for example) ahead of the regular screens. This is also helpful when elephant beans are present (beans which have become inter-twined in the cherry and which nearly always break up, if not during milling then during roasting).One easy way to quickly verify whether a shipment corresponds to the selling sample is to check the coffee's size and density composition. Pass 100 g or 200 g of the original sample and the shipment sample over the appropriate size screens and compare the percentages. Do the same with the lights by counting them.Many shipments appear visually to be a match but turn out not to be when this simple test is applied. Buyers know this, and so should the exporter.Coffee is graded by size using rotating or shaking screens, replaceable metal sheets that have round holes in them that retain beans over a certain size and allow smaller beans to pass. Screen sizes are expressed as numbers (e.g robusta grade one screen 16), or by letters (e.g. arabica grade AA -indicating a bold bean), or by descriptions (e.g. bold, medium or small bean). It all depends on the trade custom in any given country. Intermediate screen sizes (e.g. 16.5), are important in some producing countries but disregarded in others. However, nearly all coffee for export is graded to exclude the largest and smallest beans, as well as broken beans and other particles.
Standard coffee round screen dimensions
ISO dimensions (mm)
It is not always easy or possible to achieve a 100% accurate screen (e.g. nil passing through screen 16). Where a 100% accurate screen is required then marginally increasing the size of the holes to give a small tolerance in the screen may provide the required result.Slotted screens with oblong slits (usually 4.00 or 4.50 mm wide) are used in some countries to remove peaberries (single oblong beans in a cherry, the result of a genetic aberration because normally there are two beans in a cherry), which are sought after in some consuming countries.