The pressure drop resulting from polyamide 6 flow through industrial spinnerets and wire-mesh filters was examined as a possible parameter for improving spinning process constancy with experimental techniques and a numerical approach. The rheological characterization of the polymer melt was performed with a capillary rheometer and a controlled-stress rotational rheometer equipped with a high-temperature oven cell. Measurements in a nitrogen atmosphere were carried out at different temperatures and at various moisture contents to determine the effect of the postcondensation process on the rheological properties of the polymer melt. These experiments were used to collect all basic material information necessary to fit the data with the purely viscous Cross model and the viscoelastic Kaye, Bernstein, Kearsley, Zappas (K-BKZ) model. A spinning pilot plant (consisting of an extruder, a gear pump, a pressure sensor, and a spin beam with several spin packs installed) was used to measure pressure drop values through industrial spinnerets and through two types of filters: (1) Dutch twilled weave filters and (2) sintered filters. Pilot plant tests on filters showed that in the examined range of melt throughputs, the pressure drop increased linearly with an increase in the melt flow rate for all the filters considered. The results with respect to the spinneret geometry led to the conclusion that the numerical simulations gave satisfactory predictions even for experimental data coming from complex systems such as spinning plants, as long as extensional properties were accounted for by the model. On the contrary, pressure drop predictions obtained from the Cross model underestimated the pilot plant values by approximately 20% because of the inability of the model to consider the extensional component of the flow. (c) 2006 Wiley, Periodicals, Inc.