The application of an ideal model to the scaling up of a laboratory single screw food extruder is analysed. The method proposed by Harper (case 3) with a scale up factor of 2.25 was used. Experiments were carried out using a commercial corn grits sample. To compare the extruders performance, the following parameters were measured in each experiment: output (Q), measured as the solids (at feeding moisture level basis) coming out from the extruder; the die pressure (P), measured at the cylindrical head which supports the die and degree of cooking (DC), which is taken as product characteristic. DC was estimated by two parameters, the water solubility (S) and the amylographic retrogradation (R). Two grits feeding moisture levels were selected for the extrusion (15% and 30 %) in order to produce products having two different degrees of cooking levels, high and low, respectively. Three screw speeds for each extruder and two different die I/d ratio: I/d = 10 and I/d = 20. were used in the experiments. The results show that the scaled up extruder (SE) gives lower P values than the predicted by the model. This is due, mainly because a higher DC is obtained with the SE. The differences in DC are more significant in the case of 15% moisture and even more noticeable for the longer die and higher rpm. We conclude that, considering the simplicity of the model used, it could be applied to the scaling up of a single food extruder providing that a new set of processing condition are found for the SE so as to obtain the same DC as that of the small one. Among the different alternatives used (change of the die lld ratio or adjusting moisture level or the change of rpm), the decrease in the die lld ratio seems to be the best, because the higher cooking capacity of the SE allows to increase the output by reducing the die flow restriction (lld). (C) 2001 Elsevier Science Ltd. All rights reserved.