The experimental design consisted of ten repetitions per each combination of two wood species (locust or chestnut), two knife conditions (new or worn out) and two piece breaker options (with or a without). Each repetition consisted of a manually-fed log bunch, weighing about 30 kg. Average productivity with fresh material, w(H2O) = 36-42%, varied between 4 and 14 t h(-1), or between 3 and 9 oven-dry t h(-1). Productivity was 15-30% higher for locust than for chestnut, when working with new knives; it was 30-40% higher for chestnut than for locust, when working with worn knives. All else being equal, blade wear determined a 50% reduction of productivity. Installing a piece breaker resulted in a 30% decrease of productivity, but only when knives where new. Specific fuel consumption per unit of chipped biomass varied between 1.5 and 4.2 L t(-1),of fresh matter, or 2.3 and 6.4 L t(-1) of dry matter. Blade wear had a dominant effect, explaining between 70 and 75% of the total variability. Specific fuel consumption increased 1.4-2.8 times with blade wear, depending on species and piece breaker option. Knife wear and the use of a piece breaker determined a marked increase in the incidence of small chips (8-3 mm) and fine particles (<3 mm), and a parallel reduction in the incidence of large chips (45-9 mm). If the main goal of a chipper operator is increasing productivity and decreasing fuel consumption, managing knife wear should be a primary target. (C) 2013 Elsevier Ltd. All rights reserved.