Because fiber optic sensors are lightweight, compact and immune to electromagnetic interference, they have found numerous applications in monitoring industrial processes and mechanical structures. But reading the sensors’ data requires precision optical components such as narrowband, wavelength-tunable lasers and filters. The cost of these components makes fiber optic sensors impractical in many applications. Recently, scientists at the University of Strathclyde in Glasgow, UK, proposed a novel approach to interrogating fiber optic sensors that could drastically cut costs and increase the utilization of these measuring devices.Fiber Bragg gratings are probably the most common type of fiber sensor. They have narrowband reflectivities whose center wavelength changes with the sensor’s temperature or strain. An advantage of the gratings is that they can be strung in series along a single optical fiber, unlike electronic sensors that require separate leads for each sensor. Each sensor reflects a different wavelength, and each can be interrogated independently with its own wavelength (which passes unaffected through all the other sensors on the chain).But it is the requirement for producing and analyzing all the wavelengths that makes fiber-sensor interrogation systems so expensive, typically running well into thousands of dollars. Taking advantage of off-the-shelf telecom components, the scientists’ proposed system would lower that cost to less than $1000.They envision using an “8-skip-8” arrayed waveguide grating such as the one manufactured by Gemfire Corp. in Fremont, Calif. This device has eight output waveguides, or channels, and when the bandwidth of the input signal exceeds the arrayed waveguide grating’s free spectral range, multiple wavelengths emerge from each one (Figure 1). The “8-skip-8” designation indicates that the device transmits eight adjacent wavelengths of the standard telecom wavelength grid, one on each channel. It skips a bandwidth equal to eight additional grid wavelengths, and then transmits another eight in its next free spectral range. Figure 1. An arrayed waveguide grating is an interferometer fabricated in a planar waveguide (above). The different wavelengths in a broadband signal entering from the left are diffracted into different output channels on the right. The commercial arrayed waveguide grating used in this scheme has several free spectral ranges (shown here are four) in the telecom C-band between 1530 and 1565 nm (right). Within each eight-peak transmission band, the first peak from the left emerges in the arrayed waveguide grating’s channel one, the second peak in channel two and so on. Images reprinted with permission of IEEE Photonics Technology Letters.The other two critical components of the scheme are a commercial 4 × 1 optical switch and a commercial demultiplexer intended for coarse wavelength division multiplexing telecom systems. The demultiplexer selects four spectral segments from a broadband source and places each on one of its four output waveguides (Figure 2). The 4 × 1 switch then selects one of these four spectral segments to interrogate a string of fiber sensors.Figure 2. A commercial coarse WDM demultiplexer selects four spectral segments from a broadband source and places one on each of its output waveguides. In this case, the signals on each output waveguide correspond to the wavelengths transmitted in one of the free spectral ranges of the arrayed waveguide grating in Figure 1. In other words, λ1 = 1538.58 nm, λ17 = 1544.92 nm and so on.The interrogation system is illustrated in Figure 3. The broadband source — perhaps a laser diode with a 40-nm bandwidth centered at ~1550 nm — illuminates the demultiplexer. The 4 × 1 switch selects one of the demultiplexer’s spectral segments to send down the string of sensors. The sensors are chosen such that wavelengths in the selected segment can be reflected from only one sensor, so this is the only sensor interrogated. The light reflected from the sensor is routed through a circulator to the arrayed waveguide grating, where its exact wavelength is determined by the grating channel from which it emerges. A processor analyzes this information and converts it into a measurement of the sensor’s environment.Figure 3. The system sends light down the string to interrogate one sensor at a time. First, the 1 × 4 switch selects light whose wavelengths can be reflected only by the first sensor, and the return light is analyzed by the arrayed waveguide grating and the processor (top). The switch selects light to interrogate the second sensor, and the process is repeated (bottom).The 1 × 4 switch then selects from the demultiplexer a second spectral segment, whose wavelengths can be reflected only by a second sensor, and again the return signal is routed through the circulator to the arrayed waveguide grating. In the grating, the light resonates with one of the passbands in a second free spectral range and emerges from a telltale waveguide of the grating. And again, the processor converts the information about the exit waveguide into information about the sensor’s environment.The same process is repeated sequentially for each of the sensors in the string until all of them have been interrogated, and then the process begins again. If the system uses a solid-state optical switch, such as the one manufactured by Agiltron Inc. of Woburn, Mass., the individual sensors can be interrogated at speeds in excess of 5 kHz. The scientists calculate that a 1-mW laser diode would provide adequate power to interrogate a chain of sensors. IEEE Photonics Technology Letters, Sept.15, 2006, pp. 1904-1906.