Abstract In this paper we study the role of redundant sensory information to prevent the performance degradation of a chemical sensor array for different distributions of sensor failures across sensor types. The large amount of sensing conditions with two different types of redundancy provided by our sensor array makes possible a comprehensive experimental study. Particularly, our sensor array is composed of 8 different types of commercial MOX sensors modulated in temperature with two redundancy levels: (1) 12 replicates of each sensor type for a total of 96 sensors and (2) measurements using 16 load resistors per sensors for a total of 1536 redundant measures per second. We perform two experiments to determine the performance degradation of the array with increasing number of damaged sensors in two different scenarios of sensor faults distributions across sensor types. In the first experiment, we characterize the diversity and redundancy of the array for increasing number of damaged sensors. To measure diversity and redundancy, we proposed a functional definition based on clustering of sensor features. The second experiment is devoted to determine the performance degradation of the array for the effect of faulty sensors. To this end, the system is trained to separate ethanol, acetone and butanone at different concentrations using a PCAâ€“LDA model. Test set samples are corrupted by means of three different simulated types of faults. To evaluate the performance of the array we used the Fisher score as a measure of odour separability. Our results show that to exploit to the utmost the redundancy of the sensor array faulty sensory units have to be distributed uniformly across the different sensor types.