Resumen: Factors determining the variability of air-pulse pressure to determine the thresholds of laryngeal-pharyngeal reflexes, which are related to swallowing and airway protection, were explored. Potential factors affecting the reproducibility of air-pulses were experimentally evaluated and included in a multiple linear regression model. A novel device controlling these factors and minimizing variability was designed. Its reproducibility was assessed by the coefficient of variation (CV) of the pressures and duration of air-pulses, and its validity was assessed by comparing obtained pressures and durations with desired pressures and durations. Differences in the pressures of airpulse categories were assessed by a one-way ANOVA of repeated measures, a Tukey test and a box and whisker plot. The distance and angle between the exit of the tube conducting the pulses and the surface to be impacted, the diameter of the tube, the feeding pressure of the system, and the duration of air-pulses significantly affected the accuracy of air-pulses. The novel device incorporated electronic valves and a telemeter for use during the fiberoptic endoscopic evaluation of swallowing. The differences between the desired and obtained pressures and durations were below 3%. The CV of the air-pulse pressures of the novel device was 0.02. The CV of air-pulse duration was 0.05. The oneway ANOVA, Tukey test and box and whisker plot showed that the outlet pressures of air-pulse categories had statistically significant differences between them without overlap between categories, which helps to obtain an accurate threshold.