Measuring with Light Part 4: Sensor and Communications Component Compatibilities

Measuring with Light

  Part 1
  Part 2
  Part 3
  Part 4

The previous parts of this series examined the foundations of light propagation in fibers, the adaptation of the technology to sensing a wide variety of physical properties, and the potential of dense wavelength division multiplexed (DWDM) fiber-optic systems. This final part attempts to ascertain if photonic components meant for one area—be it sensing or communications—can be used in the other area.


In many respects, the comparison is quite lopsided. The annual sales of fiber-optic communications components are measured in the billions of dollars while those associated with fiber-optic sensing are counted in the tens of millions. From the view of designing a fiber-optic sensor, this isn't necessarily a bad thing. For example, I've been able to buy fiber-optic Bragg gratings for 1/100 the normal price simply because a fabrication error caused the gratings' nominal operating wavelength to miss the fiber-optic window at 1550 nm, rendering them useless in a communications system. But the new wavelength was certainly acceptable for fiber-optic sensing.

The key elements of either a fiber-optic sensor or a communications system are similar: fiber, a light source, a detector, any beam-conditioning optics, and perhaps some special components (e.g., modulators).

The Fiber

Communications. High-speed fiber-optic communications, especially in DWDM installations, operate in the 1550 nm window. The fiber of choice may be dispersion-shifted or dispersion-flattened, but it will most assuredly be single mode. There are, however, instances where multimode fiber communications links are used—typically at data rates <100 Mbps—at wavelengths other than the 1550 nm window. There are even instances where a large plastic fiber is used, such as in the short-distance transport of 400 Mbps FireWire (IEEE 1394) communication signals.

Sensing. The different forms of fiber-optic sensors (e.g., intensity-modulating, wavelength-encoding, and interferometric) can use different types of fiber. For example, a simple extrinsic, intensity-modulating sensor operates most effectively with more light delivered to the sensing location. Therefore, while a telecommunications-grade, single-mode optical fiber can be used, it's wise to use a multimode or plastic fiber—with its large core size—allowing more light to be coupled into it. For an interferometric sensor, it's desirable to use a single-mode fiber because you avoid some of the signal-fading problems that arise as a result of the multiple light paths in a multimode fiber. In other words, a single-mode fiber acceptable for communications could be used. Finally, fiber-optic sensors that rely on polarization-modulation may require polarization-maintaining single-mode fiber, so once again, that type of communications-grade fiber is quite acceptable.

Summary. Fiber-optic sensors tend to be much less picky about the type of fiber used.

Figure 22. Output powers from three different, yet similar, semiconductor light sources can radically differ. The laser diode exhibits the highest electron-to-photon conversion efficiency—and the fastest response—followed by a super-radiant diode (SRD). The slowest and lowest output power device is a generic LED—hey, what do you expect for a component that costs 10 cents?
Figure 23. The spectral width of a super-radiant diode (SRD) is shown along with that for an unregulated laser diode. The SRD bandwidth, which approaches 150 nm, is too broad to make it useful in almost every communications application, but this width makes it an attractive source for use in fiber-optic Bragg strain sensors. The unregulated laser diode output, while considerably narrower than the SRD, is also too broad to be used in dense wavelength division multiplexed systems.

Light Sources

Communications. In these applications, the restrictions placed on a DWDM light source are quite severe. It must operate at a fixed wavelength within the single-mode fiber's attenuation minimum (~1550 nm), as well as within the wavelength range in which erbium-doped fiber can provide reasonable levels of amplification (~1540– 1565 nm). The separation between adjacent DWDM channels is now 100 GHz (0.8 nm), but there is a move under way for 50 GHz (0.4 nm) channel separation.

An InGaAsP laser diode exhibits a wavelength variation with respect to temperature change on the order of 0.1nm/°C. Thus, a DWDM transmitter must be temperature stabilized to keep adjacent channels from wandering into each other. Direct modulation causes the laser's wavelength to line-width broaden (chirp) from 0.2 nm to 1.2 nm. For data rates above 2.5 Gbps, external optical modulators are frequently used to lessen the line-width broadening problem. With these factors and others, the DWDM transmitter exhibits low levels of wavelength variation but comes with a price tag in the $4500–$10,000 range. LEDs? No way—they're too spectrally broad for a DWDM channel.

Sensing. As with the fibers, different categories of sensors possess different criteria for light sources. An intensity-modulating sensor operating with a response speed in the Hz to MHz range can use an LED. In contrast, an interferometric sensor is better served by a light source with a line-width much narrower than a 100 nm wide LED. The narrower the source width, the longer the coherence length, which implies that the sensor can operate with greater physical path separation between the interferometer's two arms.

Packaging issues may dictate that you use a semiconductor light source, resulting in your choosing a laser diode, an LED, or a super-radiant diode. As shown in Figure 22, the power output and electrical power input requirements differ for these components—as does their spectral width (see Figure 23). The light source can be as simple as an incandescent light (an optical proximity sensor) or as complicated as a injection-locked, frequency-doubled solid-state laser (in an atmospheric aerosol detector).

Intensity-based sensors and interferometric sensors represent the two ends of the spectrum of fiber-optic sensors. Typically, intensity-modulating devices can use almost any light source. While variations do exist (e.g., a sensor may perform better using a polarized light signal, which is provided by a laser diode operating above lasing threshold), the sensor is just looking for photons. Interferometric sensors operate most efficiently with a narrow line-width (long co her ence length) source. Therefore a DWDM transmitter would serve just fine, but usually there are less expensive alternatives.

Summary. Different types of fiber-optic sensors allow for a much more varied suite of light sources.


Communications. There aren't too many choices available for the high-speed, low-noise detection of photons. Communi cations systems typically use semiconductor detectors (e.g., PIN diodes and avalanche photodiodes). Tradeoffs between frequency response, wavelength responsivity, amplification, and noise performance dominate the decision-making process.

Sensing. Although semiconductor detectors are frequently used, the relaxed frequency response criteria mean that optical detectors ranging from photomultiplier tubes to pyrometers can be used in low-rate sensing situations. By allowing the sensors to operate over a wide wavelength—as opposed to the DWDM window of 1540–1565 nm—many different detector materials can be used.

Summary. Photodetection is photodetection. But fiber-optic sensors offer the designer much more flexibility in choosing a detector.

Beam-Conditioning Optics

Communications. In essence, this is a very easy game: Put the photons into and get the photons out of the single-mode fiber efficiently. The applicable beam-conditioning optics can range from graded-index lenses to specialized lens systems placed between the source/detector and the fiber. Additional components can include astigmatic lenses (which compensate for the elliptical beam emitted from the laser diodes), polarization components included in optical circulators (used in optical add/drop multiplexers—see Part 3 of the series), and Bragg grating filters. You can add coatings to within the 1540– 1565 nm window.

Sensing. Beam-conditioning optical components used in fiber-optic sensors include all those just discussed for fiber-optic communications. But consider the case of a fiber-optic extrinsic sensor used to examine CO emissions in the constellation Orion. In this case, the fibers are used as light pipes to take the signal from the telescope to a signal processing location. From one perspective, the extrinsic fiber-optic sensor's optical components include the 55 in. telescope and the IR transmitting fibers. In that case, as in many other highly specialized sensing applications, the optical components are many and sometimes unique. In other cases, such as with a fiber-optic, Bragg grating–based strain gauge, the components can be the same as those in a communications channel, just operating at different wavelengths.

Summary. The beam-conditioning optical components found in fiber-optic communications systems follow traditional lines, while those used in fiber-optic sensors may be specialized. In other words, the sensor application can probably use most of the optical components used in fiber-optic communications. The opposite, however, is not true.

And in the End

This four-part series has provided an over view of fiber-optic sensing and communications. The material has ranged from an analysis of electromagnetic equations to descriptions of specific fiber-optic sensors and communications systems. An examination of the types of components used in each area provides the background for an assessment of the ease with which components usually used in one area can be used in the other.

So what's the conclusion? DWDM fiber-optic communications provides clear and detailed guidelines for the choice and operation of specific components. The fiber-optic sensor world is much more varied in components and operational guidelines—representative of the vast array of parameters that the sensors can measure. In certain cases, components aimed at the communications market may migrate over to the sensor world, but rarely will the opposite migration path exist.

Editor's Note


The figures, equations, and references in Part 4 are numbered consecutively from those in Part 1. The information in this series of articles will be elaborated on in Dr. Fuhr's presentation at the Sensors Expo Conference Program in Detroit, MI.


Suggested Articles

Revenues overall hit $3.82 billion, up 1% from third quarter of 2019, as auto plants reopened and personal electronics revenues grew

MIT Sloan and Boston Consulting Group call for expanding organizational learning to gain better financial rewards of AI deployments

Originally a 1960s memory manufacturer, Intel wants out of NAND following the market decline in 2018