Transient signal analysis of step temperature modulation in metal oxide sensor response

Abstract

Staircase waveforms of metal oxide sensor response modulated heater temperature were analysed and the alternative transient signal processing was proposed to improve the selectivity and sensitivity of gas sensor in order to identify various types of volatile organic compounds. Twentyone steps of ramping voltage up and down were applied to the sensor heater and the step responses were investigated by wavelet transform for both the step levels and decay characteristics in each step. For the step decay signal analysis, Haar and the fourth order daubechies wavelets were used to decompose and analyze the signal into the seventh level resolution. The approximate coefficients at the seventh level can be used to represent the step level response. The unequal transient response during heating up and cooling down processes of each sensor for different vapours can be determined by the difference between approximate coefficient and its next step value. The detail coefficients of the level less than the fourth level were neglected to clean out unimportant detail and noise. The detail coefficients at the fourth level were employed to illustrate the unique decay feature at each step response. The wavelet coefficients containing major transient characteristic of each vapour can be selected and used as input parameters for identifying species of odour by principal component analysis without problem of voltage offset shifting.