Follow Us

Passive wireless strain and temperature sensors based on SAW devices

V. A. Kalinin.

Proc. of IEEE Radio & Wireless Conference, Atlanta, USA, Sept. 19-22, 2004, pp. 187-190. (invited)

Abstract

Wide availability of relatively cheap RF IC transceivers and microcontrollers have made it possible to merge wireless technology and smart sensors. A rapid development and beginning of high volume production of wirelessly interrogated sensors for tire pressure monitoring in automotive industry clearly illustrates a high demand for such systems. Most of the wireless sensors contain a sensing element (often in the form of a MEMS) with signal conditioning circuitry, a microcontroller and a transmitter or transceiver obviously requiring a local dc energy source. Very often it is a battery that limits the sensor lifetime and the operating temperature range. "Virtual batteries" powered by RF radiation would be an alternative to chemical batteries; this approach is used in RF identification tags. However it would require a substantial level of RF power of the order of 1 W to achieve read range of the order of 1 m. There is another way of building batteryless sensors that can be interrogated at the same distance using ten to hundred times lower RF power. These sensors are based on SAW devices and work as passive back-scatterers.

The aim of the paper is to give an overview of existing approaches to wireless strain and temperature measurements that employ two types of SAW devices, reflective delay lines and one-port resonators. The work on the wireless SAW sensors was going on for more than 15 years and a lot of interesting systems have been proposed and developed at the prototype level. Few of them found relatively small-scale industrial application but the objective is to reach the state when these sensors may go into high-volume production.

The first aspect of the development of the wireless sensor system is the design of the SAW sensing element itself. The choice has to be made between the reflective delay line and the one-port resonator. Both types of the devices rely on the fact that the phase delay of the SAW propagating along the strained substrate depends on strain and temperature. Reflective delay lines offer a possibility to utilise TDMA and even CDMA for multi-sensor operation. Apart from measuring strain and temperature they can also transmit an ID number that would facilitate calibration procedure. On the other hand reflective delay lines usually have high insertion loss and they need to be interrogated in a wide frequency band to achieve good resolution. At the moment the 2.45 GHz ISM band is the best frequency range for operation of the reflective delay lines and the interrogation electronic unit working in this range is still quite expensive.

SAW one-port resonators have lower insertion loss. They are narrow-band devices and can operate in a number of bands below 1 GHz that are allowed by US and European regulations. However it is difficult to store ID information in the sensor itself and only FDMA can be used for multi-sensor operation. One should bear in mind that antenna matching requirements for the resonant SAW sensor are not exactly the same as those of the SAW delay line sensor. Perfect conjugate matching of the resonator to the antenna would lead to a very strong but very short pulse response and strong influence of the antenna impedance on the resonant frequency which needs to be avoided.

Design of the sensor package is a very important part of the sensor development. It belongs to the realm of mechanics and material science rather than electronics. However it is the package that mostly determines repeatability and reproducibility of the sensor. In the case of the contactless torque sensor it is also necessary to provide a good bond between the sensor substrate and the shaft. The choice of the material, the cut of the piezoelectric substrate and the SAW propagation direction determine the sensitivity of the device. There is always a room for optimisation in this field. For example, replacement of the traditional ST-X cut quartz by Y+34° rotated cut allowed to increase the torque sensitivity of the resonant SAW sensor by a factor of 2.6.

The next aspect of the wireless SAW sensing system design is selection of the interrogation method. Both the reflective delay lines and the SAW resonators can be interrogated either in time or in frequency domains and there is a number of different methods that can be used in each case. For instance, resonant sensors can be interrogated in the time domain by a short RF pulse that excites natural oscillations in the resonators (usually there are more than one resonator per sensor). Then the receiver picks up the re-transmitted exponentially decaying signal. The frequencies of the natural oscillations can be measured by (a) zero crossing method, (b) spectrum estimation and (c) gated PLL. Frequency domain interrogation of the SAW resonators can be performed the same way as it is done in the network analyser. Alternatively a continuous tracking of the minimum of S11 of the interrogator antenna can be performed by the automatic frequency control loop. All these methods have their advantages and disadvantages. As an example, a short-range continuous tracking interrogator can provide a standard deviation of the resonant frequency measurement of approximately 0.3 ppm for a typical SAW resonator with the unloaded Q factor of 10000. The pulsed interrogator using DFT has three times lower accuracy for the same measurement update period but it can perform interrogation at larger distances and can easily cope with the interrogation of more than one resonator. Careful analysis of trade-offs between the interrogation power, the sensor read range, its accuracy and the update period is required at the interrogator design stage.

In conclusion, successful development of the SAW wireless sensing system requires efforts of specialists in a number of areas such as mechanical engineering, material science, SAW design, smart sensing systems, RF ASIC and DSP design and embedded software development. It is a challenging task but it is worth all the efforts because the passive SAW wireless sensors promise to become the cheapest and the lightest ones.