We know that much of the analog world is related to sensors for physical variables such as temperature, pressure, or acceleration. The physical realization of these sensors takes on many forms, of course, depending on the specifics of the application. As a result, there's incredible diversity in the sensors themselves. Can there be a need for yet more?
That's why I was curious when I read about a sensor developed at NASA which uses no wires but is wireless, though not in the conventional RF-link way. This may seem like a contradiction, but it's not. NASA call it the SansEC, (sans , or without, electrical connection), and the idea is remarkably simple — always a good thing.
The SansEC is really almost nothing. It's an open-circuit conductive spiral designed to self-resonate and couple to a near-field magnetic probe antenna. It works on the underlying physics principle that changes in surrounding physical factors (such as a stress) cause subtle but detectable changes in the resonating spiral's fundamental characteristics.
“All materials in nature have an intrinsic characteristic to them, an intrinsic electrical storage characteristic known as permittivity and an intrinsic magnetic storage characteristic known as magnetic permeability,” Kenneth Dudley, a researcher in the Electromagnetics and Sensors Branch at NASA's Langley Research Center, said in a NASA Tech Briefs article. “If a change in a material's permittivity or permeability occurs, due to, say, fatigue of the material or damage to the material or stresses within the material, that electrical storage characteristic or magnetic storage characteristic can be sensed through the electromagnetic field.”
What's interesting about the SansEC approach is that it is small, costs little, and can be embedded into the material you wish to monitor. Clearly, it's not for many applications, and it may not work out for various reasons, but it might be a viable solution in some situations, such as detecting cracks in an aircraft's skin due to flexing or lightning.
The brief article doesn't say it, but we all know that the transducers by themselves are not the whole story for successful sensing. In most cases, the change in a sensor's characteristic due to the change in the sensed parameter is quite small — much smaller than the original signal level itself, which is already at a fairly low level. This means that any practical system needs low-noise, high-resolution analog interface circuitry that works well with fairly low magnitudes.
Coincidentally, the same issue of NASA Tech Briefs also reported on progress in yet another common yet very difficult sensing situation: measuring humidity (specifically, for car air conditioning systems).
We all know what humidity is, and we can all sense relative humidity, but it is surprisingly difficult to assess accurately. Factors such as temperature affect sensor range and performance. Thus, many humidity sensors have been developed, each targeted at a niche application. (Capacitance is being tested elsewhere to look for hidden tire flaws, but that's being done a via a wired interface; though it's simple enough in concept, it is obviously more challenging in practice.)
In many ways, sensors are an area of great technical challenge. When you peel back simple terms such as temperature, pressure, or acceleration, the real-world challenges of building and installing reliable, accurate sensors for a given application class, range, and situation are often quite complicated. That's why there are so many very different implementations for each parameter. Just look at all the ways you can measure air velocity — from a basic bendable cantilever with mounted strain gauge to a more sophisticated hot-wire anemometer circuit or the late Jim Williams' design for an acoustic thermometer.
Have you ever seen or used sensors for apparently simple physical parameters but still required cleverness and subtlety?