Singing strain gauge needs no power

Singing strain gauge needs no power Vibrating steel forks, mounted on weldable platforms, look to be the current most attractive alternative to replace strain gauges. The output frequencies allow measurement of strain to extreme accuracies, while consuming only about a microwatt. They can be powered remotely, requiring no wires or batteries and can be made immune to external interfering vibrations. First test applications are in pressure sensors but they could soon be quite ubiquitous, potentially cheaper than surface acoustic wave sensors in electric power steering systems and robust enough to go onto jet engine compressor blade roots. The devices were first conceived six years ago by Professor Barry Jones, director of the Brunel Centre for Manufacturing Metrology, part of Brunel University at Uxbridge in West London. However, given the usual speed in which new research ideas are funded in the UK, their further development has only recently attracted further funding and is being supported under an INTErSECT (INTElligent SEnsors for Control Technology) EPSRC Faraday Partnership. The project has been dubbed REMISE, which stands for REsonant MIcroSEnsor modules for measurement of physical quantities, although it has unofficially been named, the singing strain gauge. Development is proceeding under the auspices of research fellow Dr Tinghu Yan, in conjunction with a parallel study at Southampton University. The basic idea is to take a piece of thin sheet and etch out three narrow beams, joined to the original sheet at each end. Brunel is making its devices out of steel, while Southampton is investigating even smaller designs etched out of silicon. A thick-film area of piezoele ctric material is deposited on each end and the beams are driven in such a way that when the centre of the middle beam is rising, the centres of the outer beams are falling and vice versa. The output from the thick film piezoelectric transducer at one end is passed to an amplifier to provide slight positive feedback to the driving piezoelectric transducer at the other end, so that the device oscillates at its resonant frequency. The exact frequency of resonance depends on the tension applied externally to the beams, as is the case with violin and guitar strings. Frequency can be measured very accurately using digital techniques. This makes strain measurements more accurate than analogue changes in conventional strain gauge resistance and consequential current. Q factors of the first devices made in air have been found to be about 1,000. Because the beams need only vibrate with an amplitude of about 1micron, power consumption, even in air, is only about a microwatt. In initial experiments, the first devices – about 12mm long and with a natural frequency of 6 KHz – could be powered remotely by an inductive loop. Professor Jones says that it should easily be possible, in future designs, to reduce the average power consumption further, by making the devices smaller and running them only when required, with standby power available from a capacitor. The capacitor could be kept charged by a GHz electromagnetic signal from a source up to 3m away, while stil keeping well within the limits imposed by licence-free, low-power radio regulations. The next intended designs are to be 6 to 8mm long and are expected to resonate at 15 to 20kHz. Finite element analysis is used to determine resonant frequencies that will be well away from any interfering frequencies expected from any particular application. "We have never found external vibration to be a core issue," explained Professor Jones, when Eureka enquired about the state of play at this year's mtec show. The beauty of the steel products is that they have the potential to be made even in small quantities, at prices which should compete favourably with conventional strain gauge devices. They can, as in the prototype designs, be spot welded onto supporting plates, which can in turn be welded onto the devices to be monitored. The silicon devices being investigated at Southampton are around 3mm long, and are expected to have even smaller power requirements. Their advantage is that driving and amplifying electronics can, in theory, be incorporated into the same piece of silicon required to perform sensing. Their disadvantages are that they would have to be mounted in evacuated and sealed chambers, and initial tooling and set-up costs are likely to be higher. Either type of device will probably have to be mounted in protective containment if used for monitoring deflections on jet engine fan blade roots, but automotive applications might not require such a high degree of protection. The intention behind mounting the sensors on fan blades is to continuously monitor compressor torque, so as to optimise performance and drive components harder. If loading is inferred, a margin of safety has to be left in the design to allow for uncertainties in what might actually happen. If loading is actually measured, a component can safely be driven right up to just below the load at which it is expected to start to fail or exhibit an unacceptable fatigue life. Applied to electric power steering systems, reliable torque measurement prevents unintentional overloading on wheel bearings or other steering system components. It is thus highly desirable, if not essential, in such systems, however present designs, based on surface acoustic wave sensors, such as that first revealed in Eureka's September 1993 cover story, are more expensive than automotive customers would like. The research is expected to continue for another three years, but should anyone have a more immediate application, Professor Jones and his colleagues would like to hear about it. Since the technology has been successfully demonstrated and tested in prototype form, it could doubtless be implemented quite quickly if required. (More information at Brunel University and ) Design Pointers Vibrating beam sensors will yield much more accurate strain measurements than conventional strain gauges Power consumption is minuscule and the first devices are powered inductively. Future designs will probably be powered by low-power radio signals Steel tuning-fork devices can be supplied in weldable form. Silicon devices are likely to be supplied encapsulated and bonded on with adhesive