Measurement of Resistance and Capacitance of MOX Sensors With High Sampling Rate

Resistive chemical sensors, such as metal oxide (MOX) devices, usually exhibit resistance values within a wide range, from tens of kilohms to tens of gigohms. This is due, for example, to the different sensor material, manufacturing, working temperature, and so on. Electronic interfaces based on the resistance-to-time conversion (RTC) technique are widely used to handle such sensors, thanks to the low-cost, low-noise, and wide-range characteristics; in fact, time intervals are easier to measure within several decades without the need of scaling factors. However, the main limit of the RTC-based schemes is the variable and long measuring time, from microseconds (tens of kilohms) to quite a lot of seconds (tens of gigohms), impeding, for instance, a fine analysis of fast transients. This work proposes an innovative circuit based on the combination of the RTC method with the use of the least mean square interpolation algorithm. The implemented prototype allows the sensor resistance to be estimated with a constant measuring time of 10 ms within the range 10 kΩ ÷100 GΩ. The proposed implementation has shown a relative estimation error less than 10% (about 1% within 100 kΩ ÷100 GΩ). In addition, the system is capable to estimate the parasitic effect of the sensor (modeled with a capacitance up to 50 pF in parallel with the resistive component), showing a linearity error around 0.3% full scale (FS). A new tool has been also provided to estimate the uncertainty of the proposed approach due to rapid resistance changes, such as fast sensor transients. The simulation tool has been also used to prove performance improvement in the capacitance estimation, with respect to previously proposed solutions. Experimental results conducted using different real MOX sensors show the suitability of the proposed system for applications in which the sensor behavior need to be accurately analyzed, even in presence of fast dynamics.