As broadband networks become more complex, operators will need to become familiar with a wider variety of test equipment.
Historically, the analog-broadband industry (cable television, etc.) has developed relatively simple and straightforward optical networks allowing the industry to maintain fiber-optic circuits using only optical power meters, optical time-domain reflectometers (OTDRs), and fusion splicers. With the advent of dense wavelength-division multiplexing (DWDM), however, the level of optical complexity is increasing, suggesting a growing need for more sophisticated optical test equipment to keep pace with this increased complexity.
Optical noise tends to be one of the more significant limiting factors in the size of analog fiber-optic networks. Increasing the capacity by adding wavelengths or increasing the number of modulated channels will decrease the distance the link can span; assuming the optical parameters remain unchanged. Two of the primary effects that limit optical links are attenuation and dispersion.
Attenuation is the loss of optical energy the signal experiences as it travels through optical fiber. Analog link budgets are considerably shorter than their digital counterparts because they usually need much higher signal-to-noise ratios. The passive losses through a DWDM multiplexer/demultiplexer can take a significant portion of the original link energy. Adding "colors" also affects the loss budget by dividing the overall power among multiple channels.
Dispersion affects the optical signal by causing it to "stretch" and become less sharp. Ideally, a laser would emit light at a single precise wavelength, but that is not physically possible. What actually occurs is the laser source transmits a band of wavelengths centered on a desired frequency. Each of these wavelengths propagates at a slightly different speed, so the receiver sees a wider pulse than was transmitted, making the signal harder to resolve.
In the analog realm, fiber dispersion along with laser chirp generates composite second order (CSO) and composite triple beat (CTB) in multiple RF channel optical-transmission systems. The dispersive effect is dependent on the number of RF channels over a single wavelength, RF frequency of operation, fiber link length, and laser chirp. Higher chirp lasers over longer distances of fiber have worse CSO and CTB, and so on. In 1550-nm analog networks, fiber dispersion is one of the most severe distance-limiting factors. Externally modulated transmitters, which generate almost nonexistent chirp, have been widely used to extend optical reach. You can also use dispersion compensating fiber (DCF) and fiber Bragg gratings to reduce this effect.
Nonlinear noise effects become more problematic as links reach tighter optical spacing and higher channel loading. The following nonlinear distortion products, however, become less significant as optical attenuation and dispersion increase. A well-designed analog network thus becomes a balancing act between offsetting phenomenon to achieve the optimum signal quality.
Four-wave mixing (FWM) is a particularly troublesome result of this effect in WDM systems. Three wavelengths on a fiber will mix to create a fourth wavelength at a frequency Fnew , where (Fnew = + flambda 1 + flambda 2 - flambda 3). This formula will be very familiar to anyone who has calculated CTB distortions in a broadband network. These fourth wavelengths appear right on top of an existing channel because typical WDM systems space the transmitted wavelengths at regular intervals, causing constructive and destructive interference. Fiber types like nonzero dispersion-shifted fiber reduce the effects of FWM but cost more than traditional fiber.
Crosstalk in fiber links appears in multiwavelength systems where the longer wavelength channels are amplified at the expense of shorter wavelength channels. It depends upon RF frequency, fiber link, erbium-doped fiber amplifier (EDFA) spacing and output power, and polarization states. The only acceptable way to characterize stimulated Raman scattering (SRS) in DWDM systems is by design and calculations confirmed by testing worst-case SRS scenarios. SRS is measured in a 2-wavelength system, and the SRS effect is prorated for multiple wavelengths and for different EDFA fiber links. The SRS effect changes rapidly with the RF frequency. FWM is less than the SRS effect in most multiwavelength 1550-nm systems, due to high fiber dispersion.
Of the many nonlinear optical processes that can occur in singlemode fibers, stimulated Brillouin scattering (SBS) is the one most likely to limit the maximum power that may be launched into the network, limiting the usefulness of the higher-power EDFAs. Launching too much optical power into a fiber will generate acoustic waves, which cause a periodic variation of the dielectric constant that scatters the lightwave. In simpler terms, if you put too much power into a narrow optical bandwidth, the SBS effect will cause severe distortions. SBS distortions can be generated at optical power levels below 10 dBm using a very narrow line-width continuos-width laser. The power threshold where SBS starts can be extended by spreading the optical power over a broader bandwidth, effectively lowering the optical power per hertz.
In currently available analog 1550-nm equipment, most vendors can launch up to 16 dBm into a fiber. Because of differences in fiber losses, dispersion, mode diameter fluctuations, etc., the exact power level at which SBS starts varies from fiber link to fiber link. If the power being launched is just over the threshold limit, the first sign will be higher-than-predicted CSO in the lower end of the RF spectrum. As more power is applied, the effect will quickly bring distortions to unacceptable levels.
The carrier-to-noise ratio that we commonly understand for analog networks is the ratio of the "average carrier power" to the "average cumulative noise" within the video bandwidth. This concept works very well for analog video signal transmission over the optical links. But 64-QAM and 256-QAM signals are usually described as a constellation where there is considerable variation between their peak powers and their average powers. This variation is particularly severe when a "large number" of RF channels (such as 32 QAM channels in 200-MHz loading) modulate the laser. While the combined average video carrier power does not clip the laser, the peak powers of the individual channels can clip lasers several times. In analog video transmission, these "clips" are so small they do not affect the picture quality because the eye averages out minor glitches in the picture. For digital systems, including digital signals that are modulated on an RF carrier, however, bit errors occur. The total RF input to the laser must be held 2 to 3 dB below the normally understood value of clipping when modulating an analog laser with a "large number" of QAM channels. This is designated as the clip margin.
The development of the EDFA has allowed analog-based optical networks to be built past 40- to 50-km limits to distances well over 100 km with acceptable performance criteria. The following specifications need to be considered when evaluating the performance of the various fiber-optic components described for EDFA applications. Most of these distortion criteria will be familiar to anyone with RF broadband experience.
Traditional EDFAs do not have a "flat" gain bandwidth with respect to wavelength. A "flat" multiwavelength input will not result in a flat output. Large wavelength tilts are undesirable, not only because of SRS and FWM considerations, but also because system design in complex DWDM networks is considerably complicated since receiver input powers cannot be calculated and guaranteed accurately. The standard EDFA has a relatively flat-gain region that spans approximately 20 nm. Wider bandwidths are available from several vendors, and rapid developments in this area are underway. One method for simulating gain flatness is to use computer programs to estimate the pre-emphasis as the gain flatness for EDFAs. In a typical system employing several wavelengths, it is almost always necessary to pre-emphasize at least a few wavelengths. This method may be restrictive in broadband system designs where trying to maintain precise optical input levels may not be possible.
Center wavelength and operating bandwidth defines the wavelength region for which the optical component optimally performs in accordance with its other specifications, while component insertion loss is the difference in a signal's output port power relative to its input port power measured in decibels.
Component isolation is the difference in power between two consecutive input signals, S1 and S2, measured at port P1 in decibels. Component directivity is the difference in power between an input signal, S1, measured at two consecutive ports, P3 and P2, in decibels. Directivity is typically better than -55 dB. Maximum optical power is the highest signal power that can be transmitted through the component without damaging it-usually in the 300-mW range. For components that do not have epoxy in the optical path, maximum optical power can exceed 1W.
Back in the early ྌs, we had our first experience with a fiber link in Kansas City. The 4.5-km run used multimode fiber and Grass Valley electronics to deliver one video channel over a 24-dB optical link budget. At the time, the Kansas City cable system had no optical test equipment and depended on the status lights for all optical information. This system worked for approximately three years with no problems until a garbage truck tore out two spans of the fiber. After we had the fiber spliced, the system would not operate because the extra splices put us over the maximum optical budget. The result was that we spent a lot of wasted time and energy on a problem that, had we the proper test equipment, would have been obvious.
Though the cable industry is far better equipped to maintain its optical networks today, our Kansas City story serves as a good reminder for those deploying increasingly complex optical networks today. Included is a short list of some types of optical test equipment available to support these networks.
- Wideband power meters: The standard power meter uses a calibrated photo diode with a large-diameter surface area to convert light energy to a voltage. Most of these devices are very broadband from 850 to 1600 nm. Internal circuitry calibrates the sensitivity to the frequency under test. This device will treat all optical-signal sources in its range as a total power output.
- OTDRs: This device works by sending an optical pulse down the fiber under test and examining the timing and intensity of the resulting reflections. OTDRs are a must for looking for optical-cable faults.
- Frequency-selective power meters: This new type of power meter integrates an optical filter system with a power meter to maintain DWDM systems. Many of these devices will sweep through the optical passband and display both frequency and power. These devices cost less than a spectrum analyzer, and many of those available are built to withstand field usage.
- Optical spectrum analyzers: As with their RF counterpart, these analyzers are much more sophisticated than an single longitudinal mode laser but are capable of a much wider range and more precision. They also come with a much higher price tag. There will be situations where the need justifies the expense.
Don Gall has been involved with the cable-TV industry for the last 28 years. He was an integral part of the team at Time Warner that developed the first practical applications of analog fiber and hfc networks. He is currently a consultant with Pangrac & Associates (Port Aransas, TX) and can be reached at [email protected]
Mitch Shapiro has been tracking and analyzing the broadband industry for more than 12 years. He is currently a consultant with Pangrac & Associates, which this summer will publish the first of a series of in-depth reports on clustering, network upgrades, and new service strategies in the cable industry. He can be reached at [email protected] or via the P&A Website at http://broadbandfuture.com.