Optical Transmitters and Receivers
In an optical communication system, transmission begins with the transmitter, which consists of a carrier generator and a modulator. The carrier is a light beam that is usually modulated by turning it on and off with digital pulses. The basic transmitter is essentially a light source.
The receiver is a light or photodetector that converts the received light back to an electric signal. In this section, the types of light sources used in fiber-optic systems and the transmitter circuitry, as well as the various light detectors and the related receiver circuits, are discussed.
Conventional light sources such as incandescent lamps cannot be used in fiber-optic systems because they are too slow. To transmit high-speed digital pulses, a very fast light source must be used. The two most commonly used light sources are light-emitting diodes (LEDs) and semiconductor lasers.
A light-emitting diode (LED) is a PN-junction semiconductor device that emits light when forward-biased. When a free electron encounters a hole in the semiconductor structure, the two combine, and in the process, they give up energy in the form of light. Semiconductors such as gallium arsenide (GaAs) are superior to silicon in light emission. Most LEDs are GaAs devices optimized for producing red light. LEDs are widely used for displays indicating whether a circuit is off or on, or for displaying decimal and binary data. However, because an LED is a fast semiconductor device, it can be turned off and on very quickly and is capable of transmitting the narrow light pulses required in a digital fiber-optics system. LEDs can be designed to emit virtually any color light desired. The LEDs used for fiber-optic transmission are usually in the red and near-infrared ranges. Typical wavelengths of LED light commonly used are 0.85, 1.31, and 1.55 μm, more commonly designated 850, 1310, and 1550 nm where 1 micrometer (μm) equals 1000 nm. These frequencies are all in the near-infrared range just below red light, which is not visible to the naked eye. These frequencies have been chosen primarily because most fiber-optic cables have the lowest losses in these frequency ranges.
One physical arrangement of the LED is shown in Fig. 19-27(a). A P-type material is diffused into the N-type substrate, creating a diode. Radiation occurs from the P-type material and around the junction. Fig. 19-27(b) shows a common light radiation pattern.
The light output from an LED is expressed in terms of power. Typical light output levels are in the 10- to the 50-μW range. Sometimes the light output is expressed in dBm or dB referenced of 1 mW (milliwatt). Common levels are 215 to 230 dBm. Forward bias current levels to achieve this power level are in the 50- to 200-mA range. High output LEDs with output ratings in the 600- to the 2500-μW range are also available.
A typical LED used for lighting is relatively slow to turn on and off. A typical turnon/turn-off time is about 150 ns. This is too slow for most data communication applications by fiber optics. Faster LEDs capable of data rates up to 50 MHz are available. For faster data rates, a laser diode must be used.
Special LEDs are made just for fiber-optic applications. These units are made of gallium arsenide or indium phosphide (GaAs or InP) and emit light at 1.3 μm. Other LEDs with as many as six multiple layers of semiconductor material are used to optimize the device for a particular frequency and light output. Fig. 19-28 shows some LED assemblies made to accept ST bayonet and SMA fiber-optic cable connectors. These are made for PC board mounting.
The other commonly used light transmitter is a laser, which is a light source that emits coherent monochromatic light. Monochromatic light is a pure single-frequency light. Although an LED emits red light, that light covers a narrow spectrum around the red frequencies. Coherent refers to the fact that all the light waves emitted are in phase with one another. Coherence produces a focusing effect on a beam so that it is narrow and, as a result, extremely intense. The effect is somewhat similar to that of using a highly directional antenna to focus radio waves into a narrow beam that also increases the intensity of the signal.
The most widely used light source in fiber-optic systems is the injection laser diode (ILD), also known as a Fabry-Perot (FP) laser. Like the LED, it is a PN junction diode usually made of GaAs. See Fig. 19-29. At some current level, it emits a brilliant light.
The physical structure of the ILD is such that the semiconductor structure is cut squarely at the ends to form internal reflecting surfaces. One of the surfaces is usually coated with a reflecting material such as gold. The other surface is only partially reflective. When the diode is properly biased, the light is emitted and bounces back and forth internally between the reflecting surfaces. The distance between the reflecting surfaces has been carefully measured so that it is some multiple of a half-wave at the light frequency. The bouncing back and forth of the light waves causes their intensity to reinforce and build up. The structure is like a cavity resonator for light. The result is an incredibly high- brilliance, single-frequency light beam that is emitted from the partially reflecting surface.
Injection laser diodes are capable of developing light power up to several watts. They are far more powerful than LEDs and, therefore, are capable of transmitting over much longer distances. Another advantage ILDs have over LEDs is their ability to turn off and on at a faster rate. High-speed laser diodes are capable of gigabit-per-second digital data rates.
Although most of the light emitted by an FP ILD is at a single frequency, light is also emitted at frequencies from slightly below to slightly above the main light frequency. As a result, the light occupies a narrow spectrum on the fiber. In systems using multiple light wavelengths, FP lasers can interfere with one another. To overcome this problem, a distributed feedback (DFB) laser can be used. This laser is made with a cavity that contains an integrated grating structure that acts as a selective filter. The output from a DFB laser has a much narrower bandwidth and is nearer to a single light wavelength than any other type of laser. FP and DFB lasers are the highest-power lasers made and are used primarily in long-distance fiber transmission and in metropolitan-area networks.
Another type of laser used in fiber-optic systems is the vertical-cavity surface-emitting laser (VCSEL). Instead of emitting light from the edge of the diode, the VCSEL is made on the surface of silicon wafer–like transistors and integrated circuits. The cavity is made vertical to the wafer surface so that light is emitted from the surface. VCSELs are very easy to make and low in cost. Furthermore, they can be made in arrays on the wafer surface. The output power emitted by a VCSEL is greater than that of a LED but less than that of an FB or a DFB laser. Its bandwidth is less than that of a LED but wider than that of either the FB or the DFB laser. This makes VCSELs ideal for short-distance LANs or MANs. Most VCSELs are 850 nm, but recently 1310-nm VCSELs became available.
The most popular laser frequencies are 850, 1310, and 1550 nm. Some lasers are also made to operate at 980 and 1490 nm. The 850- and 1310-nm lasers are used mainly in LANs and in some MANs. The 1550-nm lasers are used primarily in long-distance fiber-optic systems because the fiber attenuation is less near 1550 nm than it is in the 850- and 1310-nm bands.
A recent development is a tunable laser whose frequency can be varied by changing the dc bias on the device or mechanically adjusting an external cavity. A wavelength range of more than 100 nm above and below a center wavelength is possible. Wider-range tunable lasers are being developed. Tunable lasers are used in applications where multiple frequencies are needed, as in dense wavelength-division multiplexing (DWDM).
Instead of having to inventory many expensive lasers, each on a separate wavelength, a single laser or at most several lasers are adequate to cover multiple frequencies. The lasers are modulated by the data in two ways, direct and indirect. In direct modulation, the data turns the laser off and on to form the binary pulses. Actually the laser is never turned off completely but is kept biased on at a low light level because it takes too much time to turn the laser on. This limits the speed of operation. To generate a maximum light pulse, the bias is then increased to its peak value by a binary 1 level.
This method works well up to about several gigabits per second, but for rates beyond this, an indirect modulator is used.
Lasers dissipate a tremendous amount of heat and, therefore, must be connected to a heat sink for proper operation. Because their operation is heat-sensitive, most lasers are used in a circuit that provides some feedback for temperature control. This not only protects the laser but also ensures proper light intensity and frequency. Many lasers use a thermoelectric cooler based upon the Peltier effect.
A light transmitter consists of the LED and its associated driving circuitry. A typical circuit is shown in Fig. 19-30. The binary data pulses are applied to a logic gate, which in turn operates a transistor switch Q1 that turns the LED off and on. A positive pulse at the NAND gate input causes the NAND output to go to zero.
This turns off Q1, so the LED is then forward-biased through R2 and turns on. With zero input, the NAND output is high, so that Q1 turns on and shunts current away from the LED. Very high current pulses are used to ensure very bright light. High intensity is required if data is to be transmitted reliably over long distances. Most LEDs are capable of generating power levels up to approximately several thousand microwatts. With such low intensity, LED transmitters are good for only short distances. Furthermore, the speed of the LED is limited. Turn-off and turn-on times are no faster than tens of nanoseconds, and so transmission rates are limited. Most LED-like transmitters are used for short distance, low speed, digital fiber- optic systems.
A typical laser driver circuit is shown in Fig. 19-31. Most of the circuitry is contained in a single integrated circuit designated the VSC7940 and made by Vitesse Semiconductor Corporation. It operates from 3.3 or 5 V and contains special automatic power control (APC) circuitry that maintains a constant laser output. This circuit can operate at data rates up to 3.125 Gbps.
The input data is in differential form, as are most logic signals above several hundred megahertz. A multiplexer is used to either pass the data directly to the laser driver transistors or select data that is clocked via a flip-flop and an external differential clock signal. Enable/disable signals are used to turn the laser off or on as desired.
The laser diode is connected to the driver by means of several resistors and a capacitor that set the current and switching response. Most laser packages also contain a photodiode that is used to monitor the laser light output and provide feedback to the APC circuit in the chip. If laser output varies because of temperature variations or decreases because of lifetime variations, the APC automatically corrects this, ensuring a constant brightness. The output driver transistors can switch up to 100 mA and have typical rise and fall times of 60 ps. External resistors and capacitors set the laser bias, modulation bias level, average current, and control loop response time.
Digital data is formatted in a number of ways in fiber-optic systems. To transmit information by fiber-optic cable, data is usually converted to a serial digital data stream. A common NRZ serial format is shown in Fig. 19-32(a). The NRZ format at A was discussed. Each bit occupies a separate time slot and is either a binary 1 or binary 0 during that time period.
In the RZ format of Fig. 19-32(b), the same time period is allotted for each bit, but each bit is transmitted as a very narrow pulse (usually 50 percent of the bit time) or as an absence of a pulse. In some systems, the Manchester or biphase code is used, as shown in Fig. 19-32(c). Although the NRZ code can be used, the RZ and Manchester codes are easier to detect reliably at the higher data rates and are thus preferred.
The most common way to modulate the laser beam is simply to turn the laser off and on in accordance with the binary data to be transmitted. This is called on-off keying (OOK). It is not desirable to turn the laser off completely, so the dc supply current to the laser is modulated by the binary data to produce amplitude-shift keying (ASK). While this works at rates below several Gbps, it is not desirable at higher data rates. Varying the dc power to the laser produces frequency variation in the laser light; chirp, a kind of frequency modulation that occurs with a dc current change; and limited depth (percentage) of modulation. Therefore, at higher frequencies an external modulator is used.
External modulators are referred to as indirect modulators or electro-optical modulators (EOMs). One type of EOM is a semiconductor device physically attached to a laser that is fully on at all times. The modulator sometimes called an electroabsorption modulator (EAM), acts as an electrically activated shutter that is turned off and on by the data. Data rates from 10 Gbps to more than 40 Gbps can be achieved with an EAM. Sometimes the EAM is packaged with the laser, but it is also available as a separate component.
A popular EOM is the Mach-Zehnder (MZ) modulator. The MZ modulator uses an external electric field to control the refractive index of a material to produce a phase or polarization shift. See Fig. 19-33. It is based on the Pocket effect, which uses an external voltage to induce a phase delay or phase shift. The voltage used to produce a π (180°) phase shift is called Vπ or the half-wave voltage. This voltage is a dc bias combined with a high-frequency sine wave that creates an electric field between the electrodes that causes the delay.
To produce an amplitude change, an interferometer is used. The interferometer uses two equal paths of light from the laser that are mixed to aid or interfere with one another to convert a phase shift into an amplitude variation.
An MZ modulator is shown in Fig. 19-34(a). The base is made of a material that exhibits the Pockel effect. Popular material is lithium niobate (LiNbO3). The laser is on continuously, and the light is split into two paths. In one path, the light just passes through. The other path is phase-modulated as described above. If the external voltage is not applied, no phase shift occurs. The light signals at the end of each path are combined in phase to form the output. If the two beams are in phase or add constructively, a strong light output occurs. This is a binary 1. If a large phase shift is introduced by applying the external modulating voltage, the two light beams will mix destructively,
causing a low-intensity light or no light output. This is a binary 0. By varying the phase shift in the controlled path, amplitude modulation (AM) is produced.
In some MZ modulators, both light paths are modulated by the external signals, which are inverted from one another as shown in Fig. 19-34(b). This permits improved control of the light output, allowing continuous variable phase shifts and/or amplitude variations.
The receiver part of the optical communication system is relatively simple. It consists of a detector that senses the light pulses and converts them to an electric signal. This signal is amplified and shaped into the original serial digital data. The most critical component is the light sensor.
The most widely used light sensor is a photodiode. It is a silicon PN-junction diode that is sensitive to light. This diode is normally reverse-biased, as shown in Fig. 19-35. The only current that flows through it is an extremely small reverse leakage current. When light strikes the diode, this leakage current increases significantly. This current flows through a resistor and develops a voltage drop across it. The result is an output voltage pulse.
The reverse current in a diode is extremely small even when exposed to light. The resulting voltage pulse is very small and so must be amplified. The base-collector junction is exposed to light. The base leakage current produced causes a larger emitter-to-collector current to flow. Thus the transistor amplifies the small leakage current into a larger, more useful output (see Fig. 19-36). Phototransistor circuits are far more sensitive to small light levels, but they are relatively slow. Thus further amplification and pulse shaping are normally used.
The sensitivity of a standard PN-junction photodiode can be increased and the response time decreased by creating a new device that adds an undoped or intrinsic (I) layer between the P and N semiconductors. The result is a PIN diode (Fig. 19-37). The thin Player is exposed to the light, which penetrates to the junction, causing electron flow proportional to the amount of light. The diode is reverse-biased, and the current is very low until light strikes the diode, which significantly increases the current.
PIN diodes are significantly faster in response to rapid light pulses of high frequency And their light sensitivity is far greater than that of an ordinary photodiode.
The avalanche photodiode (APD) is a more widely used photosensor. It is the fastest and most sensitive photodiode available, but it is expensive and its circuitry is complex. Like the standard photodiode, the APD is reverse-biased. However, the operation is different. The APD uses the reverse breakdown mode of operation that is commonly found in Zener and IMPATT microwave diodes. When a sufficient amount of reverse voltage is applied, an extremely high current flows because of the avalanche effect. Normally, several hundred volts of reverse bias, just below the avalanche threshold, are applied. When light strikes the junction, breakdown occurs and a large current flows. This high reverse current requires less amplification than the small current in a standard photodiode. Germanium APDs are also significantly faster than the other photodiodes and are capable of handling the very high gigabit-per-second data rates possible in some systems.
Fig. 19-38 shows a representative light receiver circuit. This integrated circuit, the VSC7969 by Vitesse Semiconductor Corporation, uses an external PIN or APD photodiode and can operate at rates to 3.125 Gbps. The input stage, generally known as a transimpedance amplifier (TIA), converts the diode current to an output voltage and amplifies it. The following stage is a limiter that shapes up the signal and applies it to a differential driver amplifier. The output is capacitively coupled to the next stage in the system. A signal detect circuit provides a CMOS logic output that indicates the presence of an input signal if the diode current exceeds a specific lower limit. A photodiode current monitor is also provided. The circuit operates from either 3.3 or 5 V dc, and an onboard regulator operates the circuitry and provides bias to the photodiode.
Optical transceivers or transponders are assemblies called optical modules into which both the light transmitter and light receiver are packaged together to form a single module.
See Fig. 19-39. These modules form the interface between the optical transmission medium and the electrical interface to the computer or other networking equipment. These modules are made up of the transmit optical subassembly (TOSA) and the receive optical subassembly (ROSA). Each is provided with an optical connector to get the signals into and out of the unit. These subassemblies connect to the interface circuits that supply the transmit signals and receive the input signals. The entire unit is housed in a metal enclosure suitable for mounting on a printed-circuit board of a router line card or other interface circuit.
Fig. 19-40 shows a block diagram of a typical transceiver module. The optical fiber cable connections are on the right, while the electrical connections to the networking equipment are on the left. The optical input fiber connects to a PIN diode or APD IR detector and transimpedance amplifier (TIA). The photodiode and TIA are a single package called the receive optical subassembly. This is followed by additional amplification in a postamplifier (PA). If electronic dispersion compensation (EDC) is used, it appears here in the signal flow. The EDC output then connects to a clock and data recovery (CDR) unit.
The CDR is a phase-locked loop (PLL) that extracts the clock signal from the incoming data. The recovered clock signal is then used to time the serial input data and related operations. The PLL voltage-controlled oscillator operates at the clock frequency and is locked to the incoming data frequency. The PLL serves as a filter and signal regeneration circuit to produce a clean clock signal from the degraded received signal.
A serializer/deserializer (SERDES) circuit converts the serial data to parallel and demultiplexes the individual data words and sends them to the interface, where they connect to a host computer or other equipment. The electrical interface circuits are sometimes parallel but may also be serial. A 16-bit word is common.
On the transmit side of the transceiver, the data to be transmitted is usually received from a network device in parallel form via the interface. The parallel data is converted to serial data by a serializer/deserializer circuit that serves as a multiplexer where the parallel words are put into a serial sequence for transmission. The serial data is clocked out of the multiplexer by a clock signal at the desired transmission rate.
A reference clock oscillator, usually external to the transceiver, is multiplied up to the desired clock rate by a clock multiplier unit (CMU). The CMU is a PLL with a frequency divider in the feedback path used as a frequency multiplier. The serial data is then sent to the laser driver (LD), where it operates the directly modulated laser (DML) diode. Direct modulation simply means that the data turns the laser off and on to transmit. In some modules that use higher-power, higher-frequency lasers, the serial data drives an external modulator that in turn interrupts the light path of a continuously operating laser. Note that the laser driver and modulated laser form a unit called the transmit optical subassembly.
Finally, most transceivers include a serial port called the I2C port that is used to monitor specific conditions in the module (laser temperature, supply voltage, etc.) and to control some aspects of the module.
Over the years, a variety of such subassemblies have been developed by different manufacturers. This has led to interconnection and interoperability problems, meaning that units of different manufacturers cannot be used with one another. Lack of standardization meant that no second sources of products were possible. The result is that manufacturers of optical transceivers and network equipment have come together to standardize on optical transceiver sizes, mechanical characteristics, electrical characteristics, and connectors. A set of standards known as a multisource agreement (MSAs) has emerged. Referred to as optical modules, the most common transceiver types are listed below.
300-pin. The most widely used format and standard. Converts between 16-bit 622.08-Mbps electric signals and 10-Gbps optical signals. See Fig. 19-40. For SONET system, 16-bit words are supplied at the 622.08-MHz rate and converted to serial data at a rate 16 times greater, or 9.953 Gbps. This serial rate is usually rounded off and expressed as 10 Gbps. The electrical interface is a standard called SPI-4 developed by the Optical Internetworking Foundation (OIF) or one called XSBI, developed by the IEEE. The data paths are for 16-bit words, and two-wire differential connections are used.
QSFP. This is a quad small-form factor pluggable module that is designed to support four channels of 10 Gbps, making it suitable for 40-Gbps Ethernet, SONET/SDH, Fibre Channel, and Infiniband. A QSFP+ module provides four channels of 28 Gbps.
SFF. Small form factor. A module developed for lower-speed optical applications in the 1-, 2-, and 4-Gbps range.
SFP. Small form factor pluggable. A module of 1-, 2-, or 4-Gbps applications but is “pluggable,” meaning that it uses fiber-optic cable connectors rather than the short fiber links used with SFF modules. The modules are also “hot-pluggable,” meaning that they may be put into or taken out of the system with power on. Their circuits are protected from transients and surges that
may occur during hot pluggability.
SFP+. This is an enhanced small-form factor pluggable module. It supports data rates to 10 Gbps and is compatible with 10-Gbps Ethernet, 8-Gbps Fibre Channel, and OTN OTU2.
XENPAK. A standard for a 10-Gbps optical system using the standard Ethernet 10-Gbps attachment unit interface (XAUI) electrical interface. (Note: The X is a roman numeral for 10). The XAUI interface uses four 3.125-Gbps serial electrical channels to achieve a maximum data rate of 4 x 3.125 = 12.5
Gbps. This is the gross serial optical rate in 10-Gbps Ethernet systems. The net rate is 10 Gbps, but because of the 8B/10B error correction feature of this protocol, the additional bits cause the gross rate to be higher by a factor of 10/8: 10 x1.25 = 12.5 Gbps.
X2. This standard is similar to XENPAK but is a smaller and less complex package because it does not have to deal with the high heat dissipation common in XENPAK modules. X2 is used in shorter-reach applications. XFP. A 10-Gbps small form-factor pluggable. It does not use serial-to-parallel or parallel-to-serial conversion. The electrical interface is a standard 10-Gbps serial interface referred to as XFI. The input and output frequency range is 9.95 to 10.7 Gbps.
XPAK. This is a smaller version of a XENPAK module for reaches to 10 km. Most of these modules focus on 10-Gbps applications, which include SONET, Ethernet, and Fibre Channel. Each has a slightly different data rate depending upon the standard, use of FEC or not, type of FEC, and other factors. However, the modules are essentially “protocol-agnostic,” meaning they can handle any standard or protocol. The modules are also generally classified by the range or reach of the optical signals. The basic categories are
Very short reach (VSR)—300 to 600 m or less
Short reach (SR)—2 km
Intermediate reach (IR)—10 to 40 km
Long reach (LR)—40 to 80 km
Very long reach (VLR)—120 km
The VSR and SR modules use 850-nm lasers; some SR and LR modules use 1310-nm lasers. The LR and VLR modules use 1550-nm lasers.
The most important specification in a fiber-optic communication system is the data rate, i.e., the speed of the optical pulses. The very best systems use high-power injection laser diodes and APD detectors. This combination can produce data rates of several billion (giga) bits per second (bps). The rate is known as a gigabit rate. Depending upon the application, data rates can be anywhere from about 20 Mbps to 40 Gbps.
The performance of a fiber-optic cable system is usually indicated by the bit-rate– distance product. This rating tells the fastest bit rate that can be achieved over a 1-km cable. Assume a system with a 100-Mbps·km rating. This is a constant figure and is a product of the megabits per second and the kilometer values. If the distance increases, the bit rate decreases in proportion. In the above system, at 2 km the rate drops to 50 Mbps. At 4 km, the rate is 25 Mbps and so on.
The upper data rate is also limited by the dispersion factor. The rise and fall times of the received pulse are increased by an amount equal to the dispersion value. If the dispersion factor is 10 ns/km, over a 2-km distance the rise and fall times are increased by 20 ns each. The data rate can never be more than the frequency corresponding to the sum of the rise and fall times at the receiver.
A handy formula for determining the maximum data rate R in megabits per second (Mbps) for a given distance D in kilometers of cable with a dispersion factor of d, given in microseconds per kilometer (μs/km), is
Assume a cable length of 8 km and a dispersion factor of 10 ns/km, or 0.01 μs/km:
R = 1/5(0.01)(8) = 1/0.4 = 2.5 Mbps
This relationship is only an approximation, but it is a handy way to predict system limitations.
A measurement is made on a fiber-optic cable 1200 ft long. Its upper frequency limit is determined to be 43 Mbps. What is the dispersion factor d?
A power budget, sometimes called a flux budget, is an accounting of all the attenuation and gains in a fiber-optic system. Gains must be greater than losses for the system to work. A designer must determine whether an adequate amount of light power reaches the receiver. Does the receiver get as much input light as its sensitivity dictates, given the power output of the transmitter and all the cable and other losses?
There are numerous sources of losses in a fi ber-optic cable system:
- Cable losses. These vary with the type of cable and its length. The range is from
less than 1 dB/km up to tens of decibels per kilometer.
- Connections between cable and light source and photodetector. These vary
widely depending upon how they are made. Today, the attachment is made by
the manufacturer, who specifies a loss factor. The resulting assembly has a
connector to be attached to the cable. Such terminations can produce attenuations
of 1 to 6 dB.
- Connectors. Despite their precision, connectors still introduce losses. These typically run from 0.5 to 2 dB each.
- Splices. If done properly, the splice may introduce an attenuation of only a few
tenths of a decibel. But the amount can be much more, up to several decibels if
splicing is done incorrectly.
- Cable bends. If a fiber-optic cable is bent at too great an angle, the light rays will
come under a radically changed set of internal conditions. The total internal reflection will no longer be effective, for angles have changed due to the bend. The result
is that some of the light will be lost by refraction in the cladding. If bend radii are
made 1000 times more than the diameter of the cable, the losses are minimal.
Decreasing the bend radius to less than about 100 times the cable diameter will
increase the attenuation. A bend radius approaching 100 to 200 times the cable
diameter could cause cable breakage or internal damage.
All the above losses will vary widely depending upon the hardware used. Check the manufacturer’s specifications for every component to be sure that you have the correct information. Then, as a safety factor, add 5 to 10 dB of loss. This contingency factor will cover incidental losses that cannot be predicted.
Calculating the Budget
The losses work against the light generated by the LED or ILD. The idea is to use a light power sufficient to give an amount of received power in excess of the minimum receiver sensitivity with the losses in the system.
Assume a system with the following specifi cations:
- Light transmitter LED output power: 30 μW
- Light receiver sensitivity: 1 μW
- Cable length: 6 km
- Cable attenuation: 3 dB/km, 3 3 6 5 18 dB total
- Four connectors: attenuation 0.8 dB each, 4 3 0.8 5 3.2 dB total
- LED-to-connector loss: 2 dB
- Connector-to-photodetector loss: 2 dB
- Cable dispersion: 8 ns/km
- Data rate: 3 Mbps
First, calculate all the losses; add all the decibel loss factors.
Total loss, dB= 18 + 3.2 + 2 + 2 = 25.2 dB
Also add a 4-dB contingency factor, making the total loss 25.2 + 4, or 29.2, dB. What power gain is needed to overcome this loss?
dB= 10 log Pt/Pr
where Pt is the transmitted power and Pr is the received power
29.2 dB = 10 log Pt/Pr
Pt/Pr=10dB/10 = 102.92 =831.8
If Pt is 30 μW, then
Pr = 30/831.8 = 0.036 μW
Accordingly, the received power will be 0.036 μW. The sensitivity of the receiver is only 1 μW. The received signal is below the threshold of the receiver. This problem may be solved in one of three ways:
- Increase transmitter power.
- Get a more sensitive receiver.
- Add a repeater.
In an initial design, the problem would be solved by increasing the transmitter power and/or increasing receiver sensitivity. Theoretically, a lower-loss cable could also be used. Over short distances, a repeater is an unnecessary expense; therefore, using a repeater is not a good option.
Assume that the transmitter output power is increased to 1 mW or 1000 μW. The new received power then is
Pr= 1000/831.8= 1.2 μW
This is just over the threshold of the receiver sensitivity. Now, we can determine the upper frequency or data rate.
R = 1/5dD = 1/5(0.008)(6) = 4.1666 Mbps
This is higher than the proposed data rate of 3 Mbps, so the system should work.
Regeneration and Amplification
There are several ways to overcome the attenuation experienced by a signal as it travels over a fiber-optic cable. The first is to use newer types of cable that inherently have lower losses and fewer dispersion effects. The second method is to use regeneration. Regeneration is the process of converting the weak optical signal to its electrical equivalent, then amplifying and reshaping it electronically, and retransmitting it on another laser. This process is generally known as optical- electrical-optical (OEO) conversion. It is an expensive process because in most systems regeneration is necessary about every 40 km of distance, even with the newer lower-attenuation cables.
This process is especially expensive in multifrequency systems such as dense wavelength-division multiplexing (DWDM), in which many signals must undergo OEO conversion. A third method, and the best, is to use an optical amplifier. Optical amplifiers boost signal levels without OEO conversion. A typical optical amplifier is the erbium-doped fiber amplifier (EDFA) shown in Fig. 19-41. The weak optical input signal is applied to an optical isolator to prevent signal reflections. This signal is then applied to an optical combiner that linearly mixes the signal to be amplified with one generated by an internal laser diode pump. The laser pump operates at a higher frequency (lower wavelength) than the wavelength of the signals to be amplified. If 1550-nm signals are being amplified, the pump laser operates at 980 or 1480 nm.
The combined signals are then fed into a coiled length of optical fiber that has been heavily doped with erbium ions. (Erbium is one of the rare earth elements.) The laser pump signal excites the erbium atoms to a higher-energy state. When the photons produced by the input signal pass through the erbium fiber, they interact with the excited erbium atoms, causing them to relax to their normal state. During this process, additional photons are released at the same wavelength of the input signals, resulting in amplification. A single EDFA produces a gain in the range of 15 to 20 dB. When two amplifiers with independent pump lasers and doped fiber coils are cascaded, a composite gain of up to 35 dB is possible. With such amplification, the signals can be transmitted over a distance of up to almost 200 km without OEO regeneration.
Optical Communication | Reflection | Connectors | Types of Fiber-Optics ( Optical Transmitters & Receivers | Regeneration & Amplification )
Facsimile | Fax | Image Processing | Modems | VoIP | Internet Telephony ( Optical Transmitters & Receivers | Regeneration & Amplification )
Telephones | Types | Cordless Telephones | Telephone System | Dialing ( Optical Transmitters & Receivers | Regeneration & Amplification )
Microwave Antennas | Horn | Parabolic | Helical | Bicone | Patch | Radar ( Optical Transmitters & Receivers | Regeneration & Amplification )
Satellite Orbits | Satellite Communication Systems | Satellite Subsystems ( Optical Transmitters & Receivers | Regeneration & Amplification )
Click Here To Learn ( Optical Transmitters & Receivers | Regeneration & Amplification )
Click Here ( Optical Transmitters & Receivers | Regeneration & Amplification )