Title: A statistical study of the performance of a heat-flow calorimeter: Calibration and error analysis
Abstract: From an analysis of more than three years of calibration data we derived a new calibration data for the ``small` calorimeter. Using this calibration we should be able to measure the thermal output power of unknown radiological sample to better than 0.3% of the true value (based on one standard deviation) when the unknown lies in the power range 0.25 - 4.0 watt. We learned that the calibration data taken before August 1987 showed considerably more variability than data collected subsequently. We also detected a long-term periodic fluctuation in the data on the scale of about one year. The cause of this fluctuation is unknown. This fluctuation does significantly reduce the accuracy of our measurements. The accuracy of the calorimeter is degraded by the random error of the entire measurment system, the periodic fluctuation, and the statistical error in estimating the calibration coefficients. We recommend that the cause of the periodic fluctuation be determined by monitoring the environmental conditions of the calorimeter laboratory, and a continuous program of recalibration to assure continued accuracy of these measurements.