Portland, Ore. - A novel electronics architecture enabled a seismic sensor costing nearly one-fiftieth that of more elaborate ones to detect the recent tsunami-causing earthquake in South Asia. While today's seismic detectors cost upward of $10,000 each, making them impractical to deploy in poor or rural areas such as those devastated by the recent tsunami, the new detector can be built for under $200.
"My detector does not replace those expensive seismic detectors, which have many more functions than my sensor. But my design can sense earthquakes and volcanic activity much less expensively," said Randall Peters, chairman of the physics department at Mercer University (Macon, Georgia). "Right now my detector is set up to try to sense the Earth's acceleration to test a theory of James Shirley's [a scientist at NASA's Jet Propulsion Laboratory]. I was monitoring the instrument for that purpose when it jumped significantly and continued above the background noise for a period of more than two hours."
By checking the exact timing, Peters confirmed that even though his sensor was almost halfway around the world, the device provided early warning just before the Indian Ocean earthquake that caused the region's devastating tsunami. His sensor, based on a precision pendulum-like architecture, moved two thousandths of an inch, with a period of oscillation of 30 seconds for two hours.
Standard seismometers are much more complicated, with many precision mechanical parts and complicated electronics. With their delicate feedback networks and scientific functions, the best of them cost $10,000 or more. On the other hand, Peters predicts that his design could be manufactured for as little as $200 if a custom ASIC was fabricated to hold the relatively simple electronics on a single chip.
The technological key to Peters' patented sensor is a novel method of varying the surface area of a capacitor. Today's capacitive sensors work by a widening and narrowing of a capacitor's gap, causing a dropoff in sensitivity. But the capacitor gap is constant in Peters' sensor. Variations in surface area do not compromise sensitivity.
"Most capacitive sensors depend on gap variation, but I have a patent on several variations of what I call a symmetric differential capacitive sensor-what microelectromechanical systems [MEMS] designers call 'fully differential.' Instead of gap variation, my sensor varies the area of the capacitor," said Peters.
In traditional capacitive sensors, the sensitivity and the dynamic range must be treated as a design trade-off. But the benefit of Peters' architecture is that the dynamic range is determined by changes in the surface area, while the sensitivity is set independently by the constant size of the gap. Since the mechanical area to be changed can be arbitrary and set independently of its sensitivity, its dynamic range as well as the sensitivity can be independently set to anything an application requires.
In Peters' current laboratory setup, the detector's capacitors have a constant gap of about 1 millimeter. The moving pendulum of the sensor that varies the capacitor area only has to move 25 nanometers to register a signal. Even at this lower limit, Peter's device was able to detect an aftershock last week that was centered 60 miles away from Banda Aceh (Indonesia).
"I was able to barely see the very low-frequency motions associated with the 6.2-magnitude quake," said Peters.
The deceptively simple sensor comprises four capacitors wired in a diamond similar to a Wheatstone bridge, but with series diodes and parallel resistors that rectify the bridge's output into direct current. A high-frequency sinusoidal alternating current is pumped across two opposing capacitor leads in the diamond-shaped bridge, while the differential inputs to the sensing operational amplifier are wired to the other two opposing capacitor leads.
Nominally, all four capacitors have the same surface area, thus giving a zero-rectified direct-current output from the bridge as sensed by the operational amplifier. With this setup, even nanoscale changes in the area of the capacitors are sensed by the operational amplifier, making the seismic detector exceedingly sensitive.
"In standard differential capacitive detectors, such a bridge only lets two of its components change, but in my detector all four components change, giving it twice the sensitivity of other designs," said Peters.
The physicist used a novel physical architecture to make the capacitors that varies the area of all four capacitors simultaneously. Instead of opposing two parallel plates and letting the gap vary in proportion to external vibrations in the environment, Peters used a physical phenomenon called Faraday shielding. The Faraday shield is a grounded metal plate that freely moves in between the stationary plates of the capacitors.
"This architecture works because it is not possible to induce charge through a Faraday shield. Thus the moving electrode shields various parts of the capacitors' plates, thereby changing their capacitive coupling even though they remain stationary," said Peters.
Physically, all four capacitors are fabricated as equal-sized rectangular areas on two opposing printed-circuit boards. Then a square grounded Faraday shield is inserted in between the opposing rectangular plates of the capacitors in such a way that it shields an equal area of each capacitor. When the Faraday shield is attached to a long plumb bob, similar to those used by carpenters to ensure their vertical lines are true, the plumb bob picks up any vibration in the environment, which causes a slight movement in the Faraday shield. A hinge at the top of the pendulum only allows the Faraday shield to move parallel to the plane of the capacitors' plates.
Nominally, the grounded Faraday shield is calibrated so that the surface area of the plates in all four capacitors in the bridge are equal, yielding a zero output from the sensing operational amplifier. Then any vibration in the environment that causes the Faraday shield to move will change the surface of all four capacitors in such a way that an increase in area for opposing capacitors causes a simultaneous shrinking of area in the other two opposing capacitors. This doubles the sensitivity of the device independently of its dynamic range.
In his current experimental setup, the plumb bob is dangled from a long quartz rod, to ensure that any temperature changes do not affect the length of the plumb bob. However, that requirement is for the Earth acceleration application Peters is pursuing to test Shirley's theory. Thermal stability is not necessary for a pure seismic sensor, according to Peters. Thus an inexpensive steel wire can replace the expensive quartz rod.
Peters said that the plumb bob, which measures tens of inches high, cannot be scaled down to an on-chip MEMS unit, so a single-chip sensor is impossible. All the electronics can be put on one chip, however. Thus the only external assembly is the plumb bob that dangles the moving Faraday shield between the capacitor plates.