Analysis of SRS-Induced Power Depletion and CNR Degradation

We consider the AM (the pump light) and OTDR signals (the signal light) with the same polarization co-propagated in the SMF. The OTDR pulses are rectangular periodic signal with repetition cycle T = 2.63 ms (= 1/380) and a pulsewidth τ = 10 µs. We started with the coupled equations in [17] describing the SRS interaction in optical fiber, and the AM signal power along the SMF is derived and approximately given by

α τ

where Pp0 is the average AM optical power, m is the modulation index, Ωi is the subcarrier

angular frequency, vg2 is the group velocity of the transmitted signal, α is the fiber loss coefficient, g (= gr/Aeff) is the standard Raman gain coefficient, gr, divided by the fiber effective area, Aeff, and Ps0 is the peak power of OTDR pulse. The first term represents the 1.55-µm AM power along the SMF, and the second term in the square bracket represents the power depletion due to the SRS interaction between the AM signal and OTDR light. The values of the above parameters are of Pp0 = 14 dBm, α = 0.25 dB/km, gr = 8.6×10^-12 cm/W, Aeff

where CNR1 is the CNR of the system using DEMUX with ideal channel isolation, POTDR is the crosstalk power from the OTDR channel in RX, and P av-p(1.55µm) is the average received AM signal power. m is the modulation index (0.03), ρ is the responsivity (0.85 mA/mW) of the used photodiode, q is 1.6×10^-19 Coulomb, RIN is the relative intensity noise of laser source (-155 dB/Hz), In is the receiver equivalent noise current (7pA/√Hz), and B is the bandwidth (5 MHz).

Figure 2.6 shows the calculated CNR as a function of the channel isolation of DEMUX.

We find that the minimum isolation required to completely exclude the OTDR pulse power at the RX is about 53 dB. Hence, the 55-dB isolation of the DEMUX in the experiment is satisfied for not introducing pulse-induced distortion. There is 1-dB instantaneous CNR degradation, due to the SRS-induced power depletion, in the presence of the OTDR pulse. Fig.

2.7 shows the calculated CNR evolution along the SMF, where the “#” and “*” represent two measured CNR data of 53 dB at 0 km and 49.5 dB at 37.6 km in fiber link, respectively. Note that the calculated results agree well with the experimental data. Again, the SRS-induced instantaneous CNR degradation is 1 dB as compared with the OTDR off while employing the DEMUX with either 55-dB or ideal channel isolation. Then two reasons for why almost no 1-dB CNR degradation observed are obviously. First, most of the SRS-induced noise components as shown in Fig. 2.4(c) or 2.4(f) would be averaged out by the averaging measurement in the HP8591C analyzer. Second, the duty-cycle of OTDR is so small of 3.8 × 10^-3 (= 10 µs/2.63 ms), then the SRS-induced averaged CNR degradation is < 0.1 dB, which also agrees with the experimental data.

2.4 Theoretical Model of the SRS-Interaction in the CATV System

In the on-line monitoring AM-VSB CATV transmission system, we consider both the 1.55 µm AM-VSB signal and 1.65 µm OTDR probe pulses with the same polarization copropagating in an SMF. The 1.55 µm AM-VSB signal launched into the SMF can be expressed as

))

where PAM0 is the average optical power, m is the OMI, and Ωi is the subcarrier angular frequency. We assume that the OTDR probe pulse is a rectangular periodic signal with a repetition cycle T and a pulsewidth τ. Therefore, we can express the OTDR pulse npulse(t) , as the sum of repeated pulses

)

The equation using the complex form of the Fourier series is given by

2)]}

OTDR probe pulse. eqs. (2.4) and (2.6) yield the coupled equations that describe the SRS interaction in the optical fiber [18]. By expanding the exponential term of the above expression to the first order in g, which is the standard Raman gain coefficient divided by the fiber effective area, (g= gr/Aeff), we evaluate the integral and obtain (see in Appendix A)

2 )]}}

where α is the fiber loss coefficient , ₍ ^{1 1} _{) ( )} group velocities for the transmitted AM light (λAM) and OTDR light (λSV), respectively. In eq.(2.7), the first term in brackets corresponds to the carrier power of the AM light. The second term in brackets corresponds to the interaction between the optical carrier of the OTDR light and the subcarrier channel of the AM light. This power loss resulted from the SRS effect but it does not contribute to the crosstalk. The third term in brackets is the crosstalk due to SRS interaction between 1.65 µm OTDR probe pulses and the 1.55 µm AM signal channel. The SRS-induced noise is received by the photodiode and then passed through a tunable radio frequency (RF) band-pass filter (BPF). The transfer function H(f) of BPF is assumed to be

)

where t0 is the group delay time of BPF, fc is the subcarrier frequency of CATV channels, and BW is the bandwidth of BPF. The SRS-induced noise power within a subcarrier signal band is given by [19]

where ρ is the responsivity of the photodiode in the receiver, RD is the input resistance of the optical receiver, and G is the electrical power gain of the receiver. nSRS(t) is the temporal response function of the SRS-induced noise passing through one subcarrier band and can be obtained using

4

The simulated system structure and parameters are the same as the experimental setup in Fig. 2.1. Both the 1.65 µm OTDR light with a repetition rate of 380 Hz and the 1.55 µm AM-VSB optical light have the same polarization copropagating in SMF. The system parameters are a fiber length (z) of 37.6 km, a fiber attenuation (α) of 0.25 dB/km, a fiber dispersion (D) of 17 ps/nm/km, and a fiber effective area of 80 µm². The Raman gain coefficient (gr) is about 8.6×10^-12 cm/W for the 1.65 µm OTDR on-line monitoring 1.55 µm transmission system since its frequency shift is about 11.7 THz (100 nm). For the 1.31 µm OTDR on-line monitoring 1.55 µm transmission system, a Raman gain coefficient of 0.2×10^-12 cm/W is used since its frequency shift is about 35.4 THz (240 nm). For the Raman gain coefficient as a function of the frequency shift for silica fiber, it is shown in Fig. 1.2. Moreover, a transmission fiber attenuation of 0.35 dB/km is used for the 1.31 µm OTDR on-line monitoring 1.55µm transmission system. For the receiver, its parameters are as follows:

responsivity (ρ) of 0.85 mA/mW, input resistance (RD) of 1 kΩ, and electrical power gain (G) of 1000.

2.5 Simulation Results and Discussions

Figures 2.8(a)-2.8(c) describe the simulated SRS-induced noise level of the CATV channel 2 (55.25 MHz), channel 24 (223.25 MHz), and channel 78 (547.25 MHz), respectively, using an OMI of 3%, an OTDR peak power of 10 dBm, and a pulse width of 10 µs. We found that the SRS-induced noise level for these channels of different frequencies has the same level of about -52 dBm, and then the SRS-induced baseband video distortion of “faintly white horizontal thin streaks lines” was experimentally observed on a TV picture for these low-, intermediate-, and high-frequency CATV channels. This is because SRS interaction instantaneously depletes the optical power of the 1.55 µm signal light and then generates a crosstalk between the OTDR pulses and the AM optical carrier. Therefore, the SRS effect resulted in the instantaneously degraded CNR and baseband noise in each CATV channel.

In the experiments, we found that the above-mentioned SRS-induced baseband video distortion may vanish by either attenuating the peak power of OTDR pulses to below 4 dBm, shortening the pulse width of OTDR pulses to below 0.5 µs, or reducing the OMI of the AM transmitter. Figures 2.9(a)-2.9(c) show the simulated SRS-induced noise level of CATV channel 2 (55.25 MHz) induced by (a) attenuating the OTDR peak power to 4 dBm, (b) shortening the OTDR pulsewidth to 0.5 µs, and (c) reducing the OMI to 0.7%. As compared with Fig. 2.8(a), the SRS-induced noise level obviously decreases from -52 dBm to -65 dBm in Fig. 2.9(a) and 2.9(c), and to –73 dBm in Fig. 2.9(b). Apparently either a lower peak power, a narrower pulsewidth of OTDR probe pulses, or a lower OMI gives rise to a lower SRS-induced noise level. The trend of the calculation results accurately reflects that of the experimental results. Therefore, it is experimentally and theoretically demonstrated that the SRS-induced noise level can be reduced and therefore the baseband video distortion of “faintly white horizontal thin streaks lines” on a TV picture may be eliminated by reducing the OTDR peak power, pulse width of OTDR pulses, and/or the OMI of the AM transmitter.

The SRS-induced spectral noise of a CATV channel, which resulted from the SRS interaction between the OTDR probe light and the 1.55 µm signal light, is defined as the total noise power divided by the channel bandwidth of 6 MHz. Figure 2.10 shows the simulated SRS-induced spectral noise with respect to the OTDR peak power for the 1.31 µm and 1.65 µm OTDR on-line monitoring systems with an OTDR pulse width of 10 µs and an AM transmitter OMI of 3%. The simulated SRS-induced spectral noise level decreases as the OTDR peak power decreases. By using an OTDR peak power of 10 dBm, the simulated SRS-induced spectral noise of 8.5×10^-12 mW/Hz for the 1.65 µm OTDR case is about three orders larger than that of 6.56×10^-15 mW/Hz for the 1.31 µm OTDR case. By reducing the 1.65 µm OTDR peak power to 4 dBm, the experimentally observed picture distortion vanishes, and the spectral noise is about 5.31×10^-13 mW/Hz, which is still larger than that of the 1.31 µm OTDR monitoring system operating at a peak power of 10 dBm. Figure 2.11 shows the simulated SRS-induced spectral noise with respect to the OTDR pulse width for the 1.31 µm and 1.65 µm OTDR on-line monitoring systems with an OTDR peak power of 10 dBm and an AM transmitter OMI of 3%. The SRS-induced spectral noise level decreases markedly as the OTDR pulse width is shortened to less than 1 µs for both OTDR on-line monitoring systems. By shortening the 1.65 µm OTDR pulse width to 0.5 µs, the SRS-induced spectral noise, which is still larger than that of the 1.31 µm OTDR monitoring system, is about 3.23×10^-13 mW/Hz and the experimentally observed picture distortion vanishes. Moreover, the OMI of the AM transmitter will affect the

level of SRS-induced noise. Figure 2.12 shows the simulated SRS-induced spectral noise level with respect to the OMI of the AM transmitter for the 1.31 µm and 1.65 µm OTDR on-line monitoring systems with an OTDR peak power of 10 dBm and a pulse width of 10 µs. The spectral noise level decreases as the OMI decreases. By decreasing the OMI to 0.7%, the SRS-induced spectral noise is about 5.57×10^-13 mW/Hz, which is the same as that observed when the picture distortion vanished.

Based on the above simulation and experimental results, we find that the noise level of the 1.31 µm OTDR monitoring system is about three orders less than that of the 1.65 µm OTDR case. Consequently, for the 1.31 µm OTDR on-line monitoring1.55 µm AM-VSB system, such SRS-induced baseband video distortion will not occur. This is because the Raman gain coefficient encountered by the wavelength set with a channel separation of about 240 nm between the 1.55 µm AM channel and 1.31µm OTDR channel is quite small. For the countermeasure approaches of the 1.65 µm-monitored system, the reduction of either the peak power or width of the OTDR pulse may give rise to a reduced OTDR dynamic range from 18 dB to less than 15 dB, and thus will limit the allowable supervisory link length of less than 60 km. On the other hand, a lower OMI for the transmitter also degrades the CNR performance of all the received CATV channels, and thus is unable to satisfy the system requirements.

Consequently, utilization of the 1.31 µm OTDR for the on-line monitoring of the 1.55 µm AM-VSB CATV system is the best choice not only to eliminate the SRS-induced baseband video distortion, but also to provide a longer supervisory link length. The dynamic range of today’s 1.31 µm OTDR is about 40 dB, and therefore the allowable supervised link length is about 100 km.

2.6 Summary

We have investigated the distortion of baseband video picture arising from the 1.65-µm OTDR-induced SRS effect in 1.55 µm AM-VSB CATV system. The baseband video-picture distortion in appearance with “faintly white horizontal thins lines” on the TV picture is observed. Such instantaneous distortion could not be examined through the averaged-process measurement of CNR. The SRS-induced baseband video distortion may vanish by either attenuating the peak power of OTDR pulses to below 4 dBm, shortening the pulse width of OTDR pulses to below 0.5 µs, or reducing the OMI of the AM transmitter. The countermeasure approaches entail the reduction in the OMI of the AM transmitter, OTDR peak power and/or

pulse width. However, a low peak power or a short OTDR probe pulse width degrades the OTDR dynamic range, and a low AM transmitter OMI also degrades the CNR of the transmitted system. Moreover, the 1.31 µm OTDR wavelength for on-line monitoring of 1.55 µm AM CATV systems due to almost bare of SRS interaction, occurred in the same experiment, is a good candidate since the in-between channel separation of 240 nm (about 35.8 THz for 1.31 µm ) results in a very small Raman gain coefficient of the used SMF. The theoretical analysis of the SRS interaction and these countermeasures in the transmission system are also demonstrated. The simulation results agree with the experimental data. Consequently, the SRS-induced baseband video distortion due to the 1.65 µm OTDR monitoring makes the 1.31 µm OTDR wavelength a good candidate for in-service supervision of 1.55 µm AM CATV systems and networks.

Chapter 3

SRS-Induced OTDR Distortion Traces in On-Line Monitoring 1.55 µm Fiber Raman Amplifier Transmission

System

The SRS effect can be used as direct optical amplification to extend the repeater spacing in optical fiber transmission systems [20, 21]. The FRAs pumped at multiple wavelengths are attracting much attention in DWDM systems since they offer the advantages of greatly extended gain bandwidth, distributed amplification, and low noise performance with the installed fibers as the gain media [21, 22]. In general, the OTDR is a popular tool to provide on-line monitoring of optical fiber links carrying live signals in 1.55 µm transmission systems [23-25]. With the growing deployment of distributed FRA, the system should facilitate the on-line fault-location monitoring capability to enhance system reliability. However, the SRS effect between the 1.31 µm or 1.625 µm OTDR light and 1.48 µm-band pumped lights will distributed deplete 1.31 µm OTDR light or increase 1.65 µm OTDR light in 1.55 µm FRA transmission system.

In this chapter, the on-line monitoring impact using 1.31 and 1.625 µm OTDR on-line monitoring 1.55 µm distributed FRA transmission systems are investigated and the SRS effect between the OTDR wavelength and pump wavelengths of FRA or/and the signal wavelength is discussed.

3.1 Experimental Setup

Figure 3.1 shows the experimental setup of the OTDR-monitored 1.55 µm distributed FRA transmission system, in which the forward- and backward-pumping schemes are separately examined. At transmitter site, a 1.55 µm distributed feedback (DFB) laser diode with a continuous-wave (CW) output power of 0 dBm was externally modulated by a LiNbO3

modulator with 2³¹-1 NRZ pseudo-random bit-sequence (PRBS) data at 10 Gb/s. The modulated signal was amplified by an EDFA with an output power of about 17 dBm and a noise figure of 5 dB. The power level of data channel was adjusted by a variable optical attenuator (VOA1) to 0 dBm at position A. The data channel combined with the 1.625 µm or

1.31 µm Anritsu OTDR channel that operated with 20 µs pulsewidth, and then launched into the fiber link through a 1.55/1.65 or 1.55/1.31µm wavelength selective coupler (WSC1). The peak power for 1.625 and 1.31 µm OTDR pulses with 20 µs are about 11 and 13 dBm, respectively, and the corresponding single-way dynamic range (SWDR) OTDR for 1.625 and 1.31 µm OTDR are about 29 and 34 dB, respectively.

The FRA consists of the pump module and the 50 km large effective area fiber (LEAF) link as the gain medium. The pump module is composed of a pump power combiner and four fiber-grating-stabilized pump laser diodes (LDs) with the central wavelengths of 1460, 1470, 1480, and 1490 nm, respectively. The output power of each pump LD is about 160 mW, and the total pump power at the output port of the pump combiner is about 500 mW. In forward-pumping scheme, the pump lights combined with both data and OTDR channels through a 1.48/1.55 µm coupler (WSC2), and then launched into the LEAF link, in which the pump lights and the data and OTDR channels have the same propagating direction. In backward-pumping case, which is operated in most distributed FRA transmission systems, the pump lights are launched into the LEAF link through a 1.48/1.55 µm coupler (WSC3). In this case, the pump lights have opposite propagating direction with both data and OTDR channels.

All wavelength selective couplers (WSC1, WSC2, and WSC3) are the thin-film-filter-based devices. In the 1.31 µm OTDR monitoring FRA system, the WSC1 is a 1310/1550 nm wavelength selection coupler whose insertion loss is about 1.4 dB in the region of 1300-1330 nm, 3.6 dB around 1480 nm, about 0.5 dB in the region of 1530-1570 nm, and the channel isolation is about 55 dB at 1.55 µm port and 26 dB at 1.31 µm port. In the 1.625 µm OTDR monitoring FRA system, the WSC1 is a 1625/1550 nm wavelength selection coupler whose insertion loss is about 1.8 dB around 1480 nm, 1.6 dB around 1554 nm, about 0.8 dB around 1625 nm, and the channel isolation is about 63 dB at 1.55 µm port and 59 dB at 1.625 µm port. The insertion losses of both WSC2 and WSC3 are the same with about 0.8 dB in the region of both 1420-1500 nm and 1510-1670 nm. The channel isolations of both WSC2 and WSC3 are about 65 dB at 1.55 µm port and 64 dB at 1.48 µm port. Since both WSC1 and WSC2 have low flat spectral characteristics and high channel isolations, the OTDR probe lights and also the Rayleigh backscattering lights can properly pass through both WSC1 and WSC2, and the on-line monitoring for this system can be achieved. At receiving site, a DWDM demultiplexer (DEMUX) with a channel spacing of 1.6 nm with a 3 dB bandwidth of 0.88 nm, an averaged channel insertion loss of 1.8 dB, and a channel isolation of 40 dB was used to

DEMUX. The PINFET receiver (Rx) with a receiver sensitivity of –18.5 dBm is used for BER measurements.

3.2 Experimental Results and Discussions

Figures 3.2 and 3.3 show the OTDR traces of (a) the forward-pumping and (b) the backward-pumping case for 1.625 and 1.31 µm OTDR on-line monitoring 1.55 µm FRA transmission system, respectively. The healthy trace corresponding to the case while switching off all pump lights, and the measured loss coefficient of the 50 km LEAF is about 0.248 and 0.35 dB/km, respectively. In contrast, for both wavelength OTDR supervisory FRA transmission systems, while switching the pump light on, the measured OTDR traces being distorted in both pump schemes. Such distorted traces in Figs. 3.2 and 3.3 give rise to an un-accurately measured fiber-loss coefficient and reflection profile along fiber link. When there is a fiber break occurred, the fiber fault location could be still easily located and identified in such distorted OTDR trace. Note that the last reflecting spike at about 50 km coincides with the total link length of the system.

In 1.625 µm OTDR on-line monitoring 1.55 µm FRA transmission system, the Rayleigh backscattered source of the amplified 1.625 µm OTDR probe light along the fiber link may cause distortion to the healthy OTDR trace. The mechanism of the amplified 1.65 µm OTDR probe light is mainly due to the Raman gain, pumped by the 1.46~1.49 µm pumped lights of 1.55 µm FRA, in the distributed amplification process along the fiber. The 1.55 µm data channel relatively to the 1.625 µm OTDR channel gives a separation of 75 nm (about 8.93 THz), which matches higher Raman gain coefficient of about 5.7×10^-14 m/W for the used fiber, but the power of 1.55 µm signal is much less than that of the 1.46~1.49 µm pumped lights which give a separation of 165~135 nm (about 20.8 ~ 16.7 THz) relatively to the 1.625 µm OTDR channel, which match the lower Raman gain coefficient of about 1×10^-14 ~ 2×10^-14 m/W for the used LEAF. Therefore, the 1.46~1.49 µm pumped lights of 1.55 µm FRA are mainly sources to provide Raman gain to the 1.625 µm OTDR probe light and result in distorted OTDR trace, which are higher level than the healthy OTDR trace, as shown in Fig.

3.2. In forward-pumping case, the 1.55 µm optical Raman amplification gain is depleted from 12.6 to 12 dB while switching the OTDR on. In backward-pumping case, the Raman gain of 1.55 µm data light is depleted from 14 to 13.8 dB while switching the OTDR on. The gain

3.2. In forward-pumping case, the 1.55 µm optical Raman amplification gain is depleted from 12.6 to 12 dB while switching the OTDR on. In backward-pumping case, the Raman gain of 1.55 µm data light is depleted from 14 to 13.8 dB while switching the OTDR on. The gain