Alongside solid and liquid, gaseous is one of the three classic states of aggregation. Well-known gases are oxygen, helium, hydrogen and propane - to name just a few. However, there are also complex gas mixtures such as cigarette smoke, alcohol vapors or various gases that are produced in fires. In the event of danger, each gas requires a special reaction in order to limit the damage: In the event of a fire, for example, windows must be closed tightly to prevent oxygen from providing additional fuel for the fire. Gas escaping from gas pipes, on the other hand, requires all windows to be opened as wide as possible as an immediate measure in order to neutralize the escaping gas.

Gas sensors are used to measure and evaluate the respective gases. On the one hand, these are powerful components that perform their tasks quickly and reliably. At the same time, however, they are extremely complex components that often only achieve their true performance potential through more complex circuitry. Well-known concepts use so-called sensor arrays, for example, which are used simultaneously for measurements. Each sensor generates an individual measured value for all gases - as a result, these different measured values generate a characteristic gas imprint for the respective gas type. This creates unique patterns that can be stored in a library, similar to a fingerprint file, and compared for matches as required.

The disadvantage of this physical array technology is that the individual sensors react differently to changes in humidity, temperature and gas concentration as well as their long-term drift behavior. This sometimes results in distortions, meaning that sensor arrays often have to be recalibrated. The maintenance effort and costs involved are high. Added to this is the considerable power consumption of these multiple sensor systems.

Fingerprint for every gas

Relationship between sensor signal and gas concentration in a double-logarithmic diagram.

© Figaro Engineering

The development team at Unitronic has come up with a more modern alternative to the conventional method: The 'virtual multifunctional gas sensor array' (VGSA) uses just a single miniature gas sensor based on an oxide semiconductor, which can differentiate between the various gases by means of a special evaluation. The sensor determines the type of gas based on the gas-induced distortion of periodic temperature jumps: semiconductor sensors have the property of reacting with varying degrees of sensitivity to gases in the event of temperature fluctuations. To increase the measuring accuracy, the 'USM-VGSA' uses intelligent temperature control. There is an optimum temperature environment for each gas, which provides optimum measurement results. In addition to temperature modulation, the module evaluates the conductivity (impedance) of the sensor that a gas causes. Until now, signal processing has generally only involved evaluating the ohmic resistance of a sensor.

By using the Unitronic solution, the calculated signals are free from the effects of humidity, drift of the absolute value and the memory effect. The module uses a special, highly sensitive evaluation process with innovative algorithms and techniques. The technology uses a single semiconductor sensor to generate several individual sensor signals and thus corresponds to a virtual sensor array. The sensor parameters obtained contain complex gas-typical and long-term stable patterns that represent a kind of fingerprint for each gas.