An automated disk infiltrometer was developed to improve the measurements of soil hydraulic properties (saturated hydraulic conductivity and sorptivity) of soils affected by wildfire. Guideline are given for interpreting curves showing cumulative infiltration as a function of time measured by the autodisk. The autodisk was used to measure the variability of these soil hydraulic properties in three different sample sets: (1) a reference soil consisting of a non‐repellent, uniform, fine sand; (2) soils with the same soil‐textural classification derived from the same bedrock geology but having different initial burn severities; and (3) soils from different bedrock geology but having the same burn severity. The autodisk infiltrometer had greater sampling rates and volume resolution when compared to the visual mini‐disk infiltrometer from previous studies. There was no statistical difference in the mean values measured using the autodisk and visual mini‐disk, but the variability of the autodisk measurements was significantly less than the visual mini‐disk for a given set of samples. The greatest variability of soil hydraulic properties in reference samples with uniform particle size was attributed to different pore geometries (coefficient of variation, COV = 0.28-0.34). Unburned field samples (same soil type) with heterogeneous particle sizes had greater variability (COV=0.57-0.78) than the reference samples. However, this basic variability decreased or remained constant in these field samples as burn severity increased. Additional sources of variability (COV=0.53-1.99) were attributed to multiple‐layers resulting from ash or sediment deposition. Results indicate that resolving differences in soil hydraulic properties from different sites requires more than the common 10 random samples because of the multiple sources of variability.