Optical & Fiber

Fiber optics carries information as pulses of light instead of electrical signals — immune to electrical noise, capable of enormous bandwidth, and able to run for kilometers between repeaters. The Associate CET exam covers the core physics (how light stays in the fiber), the parts of a cable, and the optoelectronic devices that convert between light and electricity. It also covers display devices, LEDs and seven-segment displays, photo devices and solar cells, cameras, where opto-isolators are used, light-activated controls and their symbols, and fiber loss budgets.

Why light, not electricity?

A fiber-optic link sends data as modulated light down a glass or plastic strand. Compared to copper, it offers far greater bandwidth, very low loss over distance, complete immunity to electromagnetic interference (light isn't affected by nearby motors, RF, or power lines), no radiated signal to pick up (tapping a fiber takes physical access, such as bending it to leak light), and electrical isolation between the two ends. The trade-offs are more delicate cable, more expensive terminations, and the need for optoelectronic converters at each end.

Total internal reflection — how light stays in

The whole trick of a fiber is total internal reflection. The fiber has a core of one glass surrounded by cladding with a slightly lower refractive index. Light traveling down the core hits the core-cladding boundary at a shallow angle and reflects completely back in — bouncing along the fiber with almost no loss instead of leaking out the sides.

The cone of acceptance

Light only stays trapped if it enters the fiber within a certain angle. That range forms the cone of acceptance — light entering inside the cone hits the boundary shallowly enough for total internal reflection; light entering at too steep an angle hits the boundary too directly, refracts into the cladding, and is lost. The width of this cone is described by the fiber's numerical aperture (NA).

Parts of a fiber cable

Single-mode vs. multi-mode

Optoelectronic devices

A fiber link needs devices to convert electricity to light at the transmit end and light to electricity at the receive end:

Handling & safety

Display devices compared

Competency 18.1 asks you to list the common electronic display devices and explain how they operate, with their advantages and disadvantages. The Audio & Video guide (section 3) explains how each TV display makes its picture; this table weighs them against each other.

LEDs and seven-segment displays

Competency 18.2.1 asks you to describe the operation and application of component LEDs and seven-segment displays. An LED is a PN junction that glows when forward-biased: electrons and holes recombine in its active region and give off photons (electroluminescence). The semiconductor sets the color: gallium arsenide phosphide glows red, and other compounds give orange, yellow, green, blue, ultraviolet and infrared (Kuphaldt). Its symbol is a diode with two arrows pointing away, for light leaving it (NEETS Module 7).

A standard LED runs at around 20 mA (Kuphaldt; NEETS gives about 10 mA for indicator LEDs). Fed from a supply higher than its forward voltage, it needs a series dropping resistor to set that current. Its reverse rating is low, typically only 5 V, so on AC it needs a rectifier diode connected anti-parallel across it (Kuphaldt). LEDs serve as power-on indicators and as the displays in calculators, digital voltmeters and frequency counters (NEETS), as fiber-optic sources (section 6) and in infrared remote controls (section 13).

A seven-segment display is seven LED bars, a to g, lit in patterns to form the digits 0 to 9, in common-anode or common-cathode form. The decoders that drive them, and why a common-anode segment lights on a low, are in the Digital Electronics guide (section 8).

Photo-electronic components and solar cells

Competency 18.3 asks you to describe how photo-electronic components operate, including how photovoltaic cells are activated (18.3.1) and the materials they are made from (18.3.2). All of them work the same way inside: light that reaches the junction frees electron-hole pairs, and the junction's electric field sweeps the electrons into the N region and the holes into the P region, which is the photocurrent (Kuphaldt).

How a photovoltaic cell is activated. A photon frees an electron-hole pair only if its energy is greater than the material's band gap; longer wavelengths, with less energy, pass through without doing so (Kuphaldt). Photon energy in electronvolts is about 1239.84 divided by the wavelength in nanometers (from Planck's constant, the speed of light and the electron charge). The cell then operates in photovoltaic mode, forward-biased by the voltage it develops across its load; a photodiode used as a sensor is usually run in photocurrent mode, with reverse or zero bias, because its current then follows the light linearly (Kuphaldt).

Materials. Most solar cells are monocrystalline or multicrystalline silicon. Amorphous silicon thin film uses far less silicon, works better in dim light and powers calculators, at about half the efficiency. Thin-film cadmium telluride, with a cadmium sulfide buffer layer (Kuphaldt), and copper indium gallium diselenide, CIGS (US Department of Energy), are the other thin films. The most efficient cells, used in space, stack gallium indium phosphide, gallium arsenide and germanium layers, each absorbing a different part of the spectrum (Kuphaldt). The band gaps are in the Semiconductor Basics table.

Electronic cameras and image sensors

Competency 18.4 asks you to explain the basics of electronic cameras and sensors. A lens focuses the scene onto an image sensor, a two-dimensional array of identical photodiodes called pixels. During the exposure, light frees electrons in each photodiode in proportion to its brightness, and they collect in a potential well. Then a CMOS sensor converts each pixel's charge to a voltage, while a CCD transfers the charge out to a register for readout; either way an analog-to-digital converter turns the result into the digital image (Evident Scientific).

In a CMOS active-pixel sensor each pixel has its own transistors: one resets the photodiode, one amplifies (buffers) its voltage and one selects its row for readout (Simon Fraser University ENSC 895 notes). CMOS sensors draw little power and run from a single supply voltage, whereas CCDs often need five or more supply voltages and more power (Evident Scientific). For color, a mosaic of red, green and blue filters, the Bayer pattern, sits over the pixels, one filter per photodiode, with a tiny lens over each to gather light (Evident Scientific).

Where opto-isolators are used

Competency 18.5.1 asks you to list some locations or circuits where opto-isolators are used. Section 6 shows what an optocoupler is: an LED and a photodetector in one package, so the signal crosses as light. Each place it is used is a place where two circuits must not share a connection.

An optocoupler's gain is its current transfer ratio (CTR), the output current divided by the LED current, expressed as a percentage. It varies from part to part, so designers select narrow CTR ranges, and it falls slowly as the LED ages, typically by less than 10% after 10000 hours (Vishay Application Note 55). NEETS adds that optical couplers suit frequencies into the low megahertz range.

Light-activated controls and their symbols

Competency 18.6 asks you to list uses for light-activated controls and how photo devices are built into them, and 18.6.1 to identify the symbols for photoresistors, photodiodes and phototransistors.

Reading the symbols. Every optoelectronic symbol carries arrows: pointing toward the symbol if the device uses light, and away from it if the device produces light (NEETS Module 7; Kuphaldt).

Fiber links in use: wavelengths and loss budgets

Competency 18.7 asks you to describe how broadband RF and optical links are used. Fiber carries light in three low-loss windows, centered at 850, 1300 and 1550 nm (NEETS Module 24). Multimode fiber, with its 50 or 62.5 µm core, runs at 850 and 1300 nm: LED sources for slower local networks, and lasers at 850 and 1310 nm for gigabit networks. Single-mode fiber, with its 9 µm core, carries telephone and cable TV traffic from lasers at 1310 and 1550 nm, because its loss is lower and its bandwidth far higher (FOA, which notes that 1300 and 1310 nm name the same window).

Broadband RF over fiber. Cable TV carries analog signals up to 1 GHz, which coax attenuates heavily, so a city-wide system once needed 15 or more amplifiers in a row, and one failure cut off every subscriber beyond it. Highly linear distributed-feedback (DFB) lasers made hybrid fiber-coax (HFC) possible: the cable company runs fiber from the headend into each neighborhood, converts the electrical TV signals to light and back, and uses coax only for the last part, typically fewer than four amplifiers deep. The fiber also gives the return path for Internet and telephone service (FOA).

How much light gets in. The cone of acceptance in section 3 is measured by the fiber's numerical aperture, NA = √(n1² − n2²), where n1 is the core's index and n2 the cladding's; launched from air, NA is the sine of the acceptance half-angle. Glass fibers typically have an NA of 0.20 to 0.29, and plastic fibers can exceed 0.50. The critical angle at the core-cladding boundary, measured from the normal, follows from Snell's law: its sine is n2 / n1 (NEETS Module 24).

The loss budget. To judge an installed link, you measure its insertion loss with a light source and power meter and compare it with a loss budget: the sum of the losses of the fiber, connectors and splices. If the measured loss is well over the budget, test segment by segment to find the cause. The FOA gives typical values and the TIA-568 maximums: about 0.3 dB for an adhesive-polished or fusion-spliced connector (0.75 dB maximum), 0.3 dB maximum for a multimode mechanical splice (about 0.15 dB for a single-mode fusion splice), multimode fiber about 3 dB/km at 850 nm and 1 dB/km at 1300 nm (3.5 and 1.5 dB/km maximum), and single-mode fiber about 0.5 dB/km at 1310 nm and 0.4 dB/km at 1550 nm (1.0 dB/km maximum for premises, 0.5 dB/km for outside plant). When the 0 dB reference is set with one reference cable, the connectors at both ends belong in the budget (FOA).

An optical time-domain reflectometer (OTDR) sends pulses down the fiber and measures the light scattered back by the glass and reflected by connectors, so it can show where along the fiber each splice, connector or break is and what it loses. It is used to verify individual events and to troubleshoot; the standards still require an insertion-loss test to qualify the link, because an OTDR can underestimate the loss of multimode fiber considerably (FOA).

On the job: checking a fiber link against its loss budget

Job practice, not exam content. A typical sequence for testing a newly installed multimode link, such as the 500 m example above. The cable plant documentation and the test standard your customer specifies come first.

1. Make it safe. Confirm that no source is connected at the far end before looking at any fiber end, and use an inspection microscope with an infrared filter (Safety, section 13).
2. Clean and inspect every connector. A speck of dust on an end-face causes significant loss (section 7), and it would show up as a failed link.
3. Set the reference. Set the meter's 0 dB reference with one reference cable between the source and the meter; the link's end connectors then count in the result, as they do in the budget (FOA).
4. Measure and compare. Test at the link's wavelengths, 850 and 1300 nm for multimode. If the 850 nm reading is 5.2 dB against a 3.55 dB budget, it is 5.2 − 3.55 = 1.65 dB over budget. The FOA calls for judgment when a reading is close to the budget; this one is well over it, so the link needs work.
5. Find the event. Test segment by segment, or shoot the link with an OTDR to locate the bad splice, connector or bend, repair it, retest and record both results (FOA).