An artefact-based fast and automated volumetric error mapping solution for medium size 3-axis machine tools that enables the calibration of a 1m3 workspace in less than 1 h is proposed for characterizing how temperature variations affect the volumetric accuracy of the machine without a priori knowledge of the temperature variations. A continuous measurement during seven days is performed on a machine affected by different heat sources and a volumetric error variation model is identified. Residual errors remaining from the identification process are used to estimate the uncertainties of individual parameters and motion errors, and Monte Carlo simulations are used to propagate them to the TCP. This model is used to understand how the volumetric positioning error of the machine changes over time and how it is generated within the kinematic chain of the machine.