A pilot scale pulse-jet baghouse was used to investigate the influence of several key pulse-jet design parameters on the overall effectiveness of the pulse. Firstly, the effect of pulse nozzle position on the acceleration of the filter fabric was measured using a lightweight accelerometer attached to the bag surface. Secondly, skim milk powder (SMP) was filtered in the baghouse, and the influence of pulse air supply pressure, nozzle position, and pulse duration on the cleaning effectiveness was determined by measuring the baghouse pressure drop before and after pulse cleaning. Results were compared to previous studies from other industries to provide guidelines for optimising baghouses for milk powder collection. Increasing the pulse air supply pressure was the most effective way to improve the pulse cleaning performance. Increasing the distance between the pulse nozzle and the bag opening was found to have no effect on the pressure drop after the pulse, despite previous studies with larger baghouses showing that a greater nozzle distance increases the entrainment of secondary air into the pulse. This difference was determined to be due to the small size of the clean air plenum in the baghouse used here, which restricted the entrainment of secondary air. Fabric deceleration was found to be a poor measure of pulse cleaning performance. Changes in the measured fabric acceleration did not correspond to changes in the pressure drop following the pulse. In addition, the tensile stresses on the cake caused by the acceleration were found to be insufficient to cause cake removal. While direct measurements of the bond strength between milk powder filter cakes and polyester filter fabrics are not available, estimates can be obtained from studies on similar systems. Effective pulse cleaning was achieved even when the tensile stress on the filter cake produced by the fabric deceleration was much lower than estimates of the bond strength, indicating that other mechanisms must be important to the cleaning process. Increasing the pulse duration from 0.1 s to 035 s had no effect on the baghouse pressure drop. This indicated that the cleaning effect of the pulse occurred during the initial expansion of the filter bag, while continued reverse airflow during the remaining duration did not contribute to the cleaning effect. The pulse duration should therefore be kept short to minimise the compressed air consumption. (C) 2015 Elsevier B.V. All rights reserved.