A cheap interface for a general sensor, resulting in high selectivity and linear output signals: application to a two dimensional array of tin oxide gas sensors to detect carbon monoxide and methane

Abstract

It is demonstrated that the direct data conversion (DDC) algorithm in combination with diode convertor networks provides an easy way to build n-dimensional sensor systems, able to detect the concentrations of n gases. The interface described substantially improves the selectivity of an array of tin oxide gas sensors. In contrast to pattern recognition techniques, where it is difficult to predict concentrations of gases, it becomes possible not only to predict the gases present in a mixture but also to determine their concentrations. The output signals of the interface vary linearly with the gas concentrations to be measured. The interface also makes it possible to solve the problem of drift. The only requirement is that the drift of the characteristics of the sensor as a function of time are known. A prototype has been built for the detection of carbon monoxide and methane concentrations in mixtures of these gases and in the presence of ambient humidity.<>