The effect of turbulence and mechanical impaction on dry powder aerosol deaggregation was tested using a novel powder deagglomeration rig, with fine particle fraction (FPF(ED<5.6 microm)), defined here as particles sized smaller than 5.6 microm, measured using an Anderson inertial impactor. Powder from GlaxoSmithKline Ventodisks was deaggregated either using turbulence generated with a ring of impinging jets, or by impacting the powder on bars of a wire mesh. This deaggregation was compared with deaggregation achieved with the GlaxoSmithKline Diskhaler. The turbulence levels in the test rig and at the exit of the Diskhaler were quantified using laser Doppler velocimetry (LDV). In addition, the Ventodisk powder's auto-adhesion properties were altered by introducing the powder into a high humidity environment (25 degrees C and 25% R.H.) and then deagglomerated by both the rig (using turbulence as the primary deagglomeration mechanism) and the Diskhaler. Fine particle fractions were found to increase from 13 to 24% as the level of turbulence in the rig was increased. However, fine particle fractions found with the Diskhaler were 35%. Turbulence levels found in the rig at the highest jet flow rate were significantly higher than that at the outlet of the Diskhaler, leading to the conclusion that turbulence is not the only method of deaggregation in this inhaler. The humidified powders were significantly more difficult to deaggregate, giving a FPF(ED<5.6 microm) of 9% when using the rig and 15% when using the Diskhaler. Fine particle fractions produced when deagglomerating the powder with the wire meshes were similar to those produced without a mesh, showing that mechanical impaction had little effect. The results underline the utility of having a rig that can explore the ability of a powder to deagglomerate with controlled variations in the deaggregation forces.