Results of an accuracy test of the Mod-II Site Data Acquisition System

Abstract

The National Solar Data Network collects and analyzes data on the performance of active and passive solar-energy systems throughout the United States. The accuracy of the specially designed Site Data Acquisition System (SDAS) used in the network has not been adequately characterized. Uncertainty about the SDAS's accuracy has become one of the major limitations in evaluating the accuracy of the network. The results of SDAS testing for basic calibration accuracy and for errors caused by changes in ambient temperature, changes in a-c line voltage, drift with time, changes in loading on the excitation power supplies, channel-to-channel cross-talk, and maintenance. System accuracy was found to depend on the configuration of the SDAS channel. Over one year of operation and with the expected changes in temperature and ac line voltage, an accuracy of 0.6% of the input signal plus one count, with error-free initial calibrations assumed, was typical. For a full-scale signal, this error is seven counts. Most of the error is introduced through the power supplies used for sensor excitation; with improved power supplies, this seven-count error could be reduced to two counts. Most of the accuracy testing was conducted on only one ADAS, a Mod-II unit. Therefore, generalizations to other SDAS units are difficult. However, a Mod-II SDAS probably will not have errors less than half or more than twice as great as those measured and reported here.