Diffusion type gas detectors are used in fixed, machine mounted and handheld applications in underground mines for the real time detection of flammable and toxic gases. The calibration of a gas detector – typically by a ‘bump’ or ‘challenge’ test – sets the reference against which all subsequent measurements are taken. Failure to properly calibrate a detector will introduce errors into every subsequent measurement, regardless of the claimed accuracy of the instrument. The vulnerability of the calibration process to poor calibration mask design and to ambient air velocities has been established. The real problem, however, is knowing if and how the calibration process has been compromised. This paper discusses recent experiments showing the susceptibility of gas detection and calibration to poor calibration mask design and to ambient air velocities, and outline new research utilising a computational fluid dynamics approach to analysing calibration mask performance.