Experimental Setup & results of bidirectional loop transmission

Fig. 4.5 shows the experimental setup for testing long-distance bidirectional transmission using four-port interleavers in a recirculating loop. The eight-channel laser sources are grouped into two categories: One with wavelengths between 1550.52 nm and 1551.72 nm, another with wavelengths between 1554.54 nm and 1555.75 nm; both are on standard ITU 50 GHz channel spacing grids. The even and odd channels stand for two traffic directions – even channels represent east (bond) traffics and odd channels represent west (bond) traffics; they are individually modulated by a LiNbO3 electro-optical (EO) modulator at 10 Gbps with a 2³¹−1 PRBS pattern. A polarization controller is used on the odd traffic to ensure that the polarization states of the odd/west channels are orthogonal to those of the even/east channels, reducing the deleterious nonlinear effects. An interleaver is placed at the input of the recirculating loop to split the odd/east and the even/west channels for bidirectional transmission and reverse-directed traffics (east and west traffics) are combined for unidirectional amplification [31]. The transmission fiber used in the recirculating loop has two spools of 50 km Corning LEAF fiber. A dual-stage EDFA with 5 km of Corning DCF is employed in the mid-stage of the loop to compensate the transmission loss and accumulated dispersion in LEAF fiber. The effective gain and noise figure of the dual-stage EDFA in all channels are around 23 dB and 5.5 dB, respectively. The fully compensated wavelength of this fiber loop is located approximately at 1553.2 nm. The two interleavers in the loop are specially arranged to reduce chromatic dispersion induced by the flat-top transmission band design of the interleaver [32]. A 3R receiver with -32 dBm sensitivity at BER of 10^-9 is utilized to evaluate the transmission quality.

Fig. 4.5. Experimental setup of bi-directional loop transmission

Fig. 4.6 displays the received optical spectrum after 500 km with an OSNR of over 31 dB for all channels with a 0.02nm resolution bandwidth setting on the OSA. This configuration effectively blocks the RB by using only one amplification section for two traffic directions. Fig. 4.7 shows the receiving power penalties of BER equals to 10^-9 at all channels.

All channels had power penalties of less than 2.2 dB and the penalty differential between them is less than 0.36 dB. Fig. 4.8 plots the BER curves and Fig. 4.9(a)~(d) show the corresponding eye diagrams at channel seven, for back-to-back, 100 km, 300 km and 500 km transmissions. The measured power penalties were about 0.4 dB, 1.1 dB and 2 dB for 100, 300 and 500 km transmissions, respectively, at a BER of 10^-9 with the optimal polarization condition. In the recirculating loop, the polarization controller is used to minimize the polarization effects -- such as polarization dependent gain (PDG) and polarization dependent loss (PDL). The penalties are attributed to ASE accumulation due to the SNR degradation

Transmitter

resulting from high link loss within the amplifier span.

Fig. 4.6. Received optical spectrum after 500 km bidirectional transmission

1550 1551 1552 1553 1554 1555 1556 -2

-1 0 1 2 3 4

Penalty (dB)

Waveolength (nm)

Power Penalty

Fig.4.7. Power penalty after 500 km bidirectional loop transmission

Fig. 4.8. BER curves of Channel seven after 0 km, 100 km, 300 km, 500 km traffic

(a) (b)

(c) (d) Fig.4.9(a)~(d). Corresponding eye diagrams of Channel 7 after 0 km, 100 km,

300 km, 500 km

-35 -34 -33 -32 -31 -30

10

^-11

10

^-10

10

^-9

10

^-8

10

^-7

10

^-6

10

^-5

10

^-4

10

^-3

Receiving Power (dBm)

Bi t E rro r R a te

0 Km

100 Km

300 Km

500 Km

In a recirculating loop experiment, if the optical data pattern length time is longer than the sampling window used in taking BER measurement, pattern-dependent errors arise from time to time [34]. An accumulated error measurement can verify the stability and ensure that the appropriate sampling window is used in the recirculating loop experiment. As a result of [34], the error counts accumulate almost continuously when system keeps accurate sampling window. Otherwise, the accumulated errors would be missed for long periods of time, and then would be over-sampled for certain periods of time. Fig. 4.10 plots the measured accumulated errors as a function of time (10-s intervals) at a BER of 2.46 × 10^-9 after 100 km and 500 km transmission. This figure demonstrates the robustness of the transmission system for BER measurement. Moreover, the authors believe that this configuration can accommodate more optical channels, 16 or 32 channels, within the C band because this interleaver is implemented to cover the whole C band (1530 nm to 1560 nm).

Fig. 4.10. Accumulated Errors after 100 km and 500 km bidirectional transmission at BER = 2.46 × 10^-9

0 100 200 300 400 500

0 200 400 600 800 1000 1200 1400 1600 1800 2000

Number of errors

Time (s)

100 km 500 km

BER @ 2.46E-9

Chapter 5

Simulation of bidirectional transmission systems

According to chapter tree and four, we have successfully demonstrated bidirectional transmission system by exploiting four port interleavers. Table II lists the power penalty obtained form the two experiments.

Table 2. Power penalty of bidirectional transmission experiments

Distance \ Configuration Straight Line Loop

100 km - 0.4 dB

160 km 0.8

210 km 1.5dB -

300 km - 1.1 dB

500 km - 2.1 dB

From the results of experiments, the penalties of “Loop” experiment increase steadily with the transmission distance. However, the mean penalty is 1.4 dB in the “Straight Line” case after 210 km transmission; while the penalty of the “Loop” case is 1.1 dB after 300 km transmission distance. This discrepancy exhibits that “Straight Line” and “Loop” case may not be identical even though they are both bidirectional transmission configurations. This chapter will focus on the simulation for comparing the two schemes.

5.1.1 Analysis of bidirectional loop transmission system

As we know, the recirculating loop is a convenient and economic method to examine the performance for long transmission distance without actually constructing the whole setup.

However, in a recirculating loop, some of the concerns in a bidirectional loop system are different from those in a unidirectional loop system. For instance, the loop time control, the setup of the bidirectional recirculating loop, the input noise in EDFA, etc.

Tx Rx

Coupler

Fiber

EDFA

Tx RT RT RT

Uni-directional transmission Round Trip

Analysis

Rx

Transmitter Switch (AO1)

Loop

Switch (AO2)

Error Gate

On Off

On Off

On Off

τ

loop

RT: Round Trip

Fig. 5.1(a) Uni-directional loop and corresponding round trip analysis (b) loop time control

In the unidirectional transmission system, as long as the appropriate loop time is executed, all of the concerns are irrelevant whether in loop or straight line system. The performance of a unidirectional loop transmission is approximately the same as that of a straight line transmission, which means the recirculating loop could represent unidirectional straight line system appropriately. On the contrary, due to the lack of the adequate loop time control, it is inappropriate to demonstrate the long distance bidirectional transmission just directly applying the recirculating loop to the bidirectional system. The detail of the structure and loop time control is displayed in the following Fig. 5.2.

Tx Rx

* IL:IL:InterleaverInterleaver

West

Fig. 5.2(a) seems be to a typical type of bidirectional loop configuration. However, while considering the timing control of such scheme, the lack of proper switching time and the unpredicted power balance issue is inevitable. All of the qualms lead to the consequence that either abandon this configuration or adjust it to be a more practicable structure.

Fig. 5.3(a) adjusted bidirectional loop configuration (b) loop time control

After modifying the loop configuration, as shown in Fig. 5.3, this bidirectional loop scheme seems more achievable. However, the concerns mentioned above, mostly the setup of the bidirectional recirculating loop, the input noise in EDFA, may lead to the consequence that the conditions of a bidirectional loop are not the same as a bidirectional straight line

Figure5.2. (a) Configuration of original bidirectional loop (b) Loop time control of original bidirectional loop

IL IL IL IL

system. Under this consideration, is bidirectional loop still adequate to stand for bidirectional straight line transmission? The following Fig. 5.4 and Fig. 5.5 depict the differences among bidirectional loop and straight line transmission systems.

Tx 1 Opposite w est traffic East Even W est Odd IL : Interleaver

From Fig. 5.4, we observe that the “same conditioned” bidirectional traffics co-propagate in the CO-EDFA route. Furthermore, during the forward propagating traffic flow, the opposite signals in transmission fiber (50 km LEAF fiber) are the older/less noisier signals which are ahead of the present signals half of the round trip span, 50 km.

5.1.2 Analysis of bidirectional straight transmission system

As shown in Fig. 5.5, the basic architectures of loop and straight line case are dissimilar.

Not only the lack of the extra boost EDFAs within transmission path, but also the traffic flows of the bidirectional signals does the “Straight Line” scheme reveal the differences between the

Fig. 5.4 Bidirectional Loop System

“Loop” configuration. To be more specific, take the bidirectional traffics input CO-EDFA at span I for example, the west-to-east (odd channel) signal are practically noiseless signals, Podd, which need to be amplified in order to continue propagation, while the east-to-west (even channel) signals, Peven+ (n-1)PASE, are the older and noisier that need to be amplified to complete the transmission. The opposite signals in the transmission fiber are not just older than the present signals for 50 km transmission. It usually concerns with the total transmitted length and at which part of the signal we might be interested. All in all, the conditions of the west-to-east and east-to-west signal flows for straight line case are not the exactly same as those of the loop case.

T x 1

For comparing the two bidirectional systems, straight line and loop system, we ought to build both schemes as many spans as we can. Unfortunately, the lack of the crucial devices, such as Interleavers, obstructs us for doing so. Therefore, we use the simulation tool, VPI, to

Fig. 5.5 Bidirectional Straight Line System

simulate both loop and straight line bidirectional transmission structures. To simulate the loop and straight line schemes, we decompose and analyze the signal path in a recirculating loop and in straight line system, separately. The detailed casual relation of bidirectional traffics will be discussed later in Fig. 5.7 ~ Fig. 5.9.

However, some preliminary conditions should we be aware of before starting to simulate and compare the two models. Due to the basic structures of bidirectional loop and straight line are not the same; some of the key factors should be held the same or approximate the same for comparing the two systems.

1. Input power before coupler and circulators, which means the total input power to the whole system.

2. Total power before dispersion compensated fiber.

3. Maxima total power in transmission fiber.

Fig. 5.6 Dispersion map used in simulation

Besides the preliminary conditions, the dispersion management should be taken into account for evaluating both schemes. Generally speaking, for a recirculating loop system, if the dispersion flattened method is not applied, as long as the transmission distance is

elongated, the accumulated chromatic dispersion would be severe in the transmission channel which center wavelength is away from compensated wavelength/frequency. On the other hand, for a straight line system, the dispersion slope and compensated wavelength/frequency vary from one to one DCF. Therefore, the accumulated chromatic dispersion in this circumstance is smoother than that in the loop structure. Nonetheless, the dispersion management is not the crucial discussion in comparing the bidirectional loop and straight line traffics, we apply the dispersion flatten method to eliminated the accumulated dispersion issue. Fig. 5.6 shows the dispersion map for sixteen channels after applying the dispersion flatten method.

5.2.1 Simulation setup of bidirectional loop system

In this simulation, we consider two types of bidirectional transmission, bidirectional transmission using a recirculating loop (bidirectional loop) and the simulated real world bidirectional transmission (the straight line transmission). The setup configuration for bidirectional loop transmission that used in our simulation is shown in Fig. 5.7. We use a four port interleaver, with insertion loss of 2.2 dB and 50/100 GHz channel spacing on standard ITU grid, and a two stage EDFA to function as a bidirectional amplified EDFA in testing the bidirectional traffics. The whole set contains an interleaver and a dual stage EDFA with appropriate length of DCF is so called a CO-EDFA, which main function is a bidirectional amplifier with high gain, low noise figure and elimination of Rayleigh Backscattering characteristics [35]. The sixteen input laser sources are grouped into two categories, west/odd traffic arranges from 1550.52 nm to 1556.15 nm and east/even traffic arranges from 1550.12 nm to 1555.75 nm, each group matches the standard ITU 50 GHz channel spacing grid. The west and east channels are individually modulated by an electro-optical (EO) modulator at 10 Gb/s with a 2³¹-1 PRBS pattern.

Tx Rx

Spread into line structure

IL IL

Fig. 5.7 Decomposed bidirectional loop system

The transmission consists of 8×100-km (50 km /50 km) LEAF fiber, the detail parameters are listed in Tab.3. In the bidirectional loop scheme, with adequate route selection [35], the first Interleaver is used to separate east and west signals for bidirectional traffics; the second Interleaver is used to combine opposite direction signals into uni-directional amplification. A dual–stage EDFA and 4.6-km dispersion compensated fiber (DCF) are employed in the uni-amplification routing for compensating the transmission loss and accumulated dispersion arise from transmission (LEAF) fiber. The chromatic dispersion is assume to be fully compensated at each channel in the two cases, bidirectional loop and bidirectional straight line systems. The noise figure in each amplifier is assumed to be 5 dB. In accordance to the former bidirectional loop experiment [35], the maximum launch power input transmission fiber is +4.1 dBm, whereas the power launched into the DCF is kept low (-5.6 dBm) to avoid introducing nonlinear effects due to the highly nonlinear behavior of this type of fiber.

Parameters used in bidirectional loop simulation are listed in Table 3.

T x 1

Table 3. Parameters used in bidirectional loop architecture.

Number of spans 8 Fiber Length (LEAF) 100 km (50/50)

Dispersion (LEAF) 4.1639 Dispersion slope (LEAF) 0.06 Loss Coeff. (LEAF) 0.2 Nonlinear Coeff. (LEAF) 2.6e-20

Bit rate 10 Max. launch power +4.1

Power input DCF -5.6 Fiber Length (DCF) 4.63

Loss Coeff. (DCF) 0.6 Dispersion (DCF) -90

Dispersion slope (DCF) -0.06*4.1639/90 Nonlinear Coeff. (DCF) 4.0

5.2.2 Simulation setup of bidirectional straight system

The long distance bidirecitonal straight line configuration is illustrated in Fig. 5.8.

Fig. 5.8 Decomposed bidirectional straight line system

From Fig. 5.8, we observe that “west-to-east traffic” encounters monotonic decreasing amount of ASE within “east to west traffic” during the transmission process, and vice versa.

However, constructing the transmission system according to Fig. 5.8 configuration is irrational due to the causal relationship between the west-to-east/east-to-west traffic flows. On the other hand, one cannot take the future incident as present information to manipulate. It’s contradictory to logical operation in our simulation tool, VPI. Therefore, three stages of preparation signals are utilized to simulate the real bidirectional transmission circumstance Fig. 5.9 shows the decomposed bidirectional straight line system as well as some tips we used in the simulation, particularly the causal relationship in the transmission flow is concerned.

West traffic P*_even1+(n-1)P_ASE

P_even1+P_ASE P_even1+nP_ASE

P*_even2+(n-1)P_ASE

P_odd3+P_ASE

Fig. 5.9 Simulated bidirectional straight line system

The circulators at every input end are used to maintain the same input power at each stage. In stage I, unidirectional signals with different amount of noise are created, these preparation signals could be the opposite input signals for stage II. Stage II focuses on the west-to-east transmission, various opposite signals input to fiber and EDFA are considered.

Stage II is a bidirectional system with actual conditions considered, nevertheless, the sources of the opposite data are not generated under bidirectional situation. To simulate the actual bidirectional transmission, not only the various opposite signals should be pondered, but also the opposite signals should be a bidirectional system suffered from different amount of noise in reverse traffic. As a result, a bidirectional transmission, stage III, which opposite signals

originate from a bidirectional system, stage II, is taken into account as a relatively literal bidirectional transmission system.

Considering the optimized setup for straight line structure for comparing with the loop scheme, the maximum launch power input transmission fiber and the power launched into DCF are +4.6 dBm and -5.1 dBm, respectively. Parameters used in simulation are listed in Table I and Table II for loop case and straight line case.

Table 4. Parameters used in bidirectional straight line architecture.

Number of spans 8 Fiber Length (LEAF) 100 km (50/50)

Dispersion (LEAF) 4.1639 Dispersion slope (LEAF) 0.06 Loss Coeff. (LEAF) 0.2 Nonlinear Coeff. (LEAF) 2.6e-20

Bit rate 10 Max. launch power +4.6

Power input DCF -5.1 Fiber Length (DCF) 4.63

Loss Coeff. (DCF) 0.6 Dispersion (DCF) -90

Dispersion slope (DCF) -0.06*4.1639/90 Nonlinear Coeff. (DCF) 4.0

5.3 Simulation results of bidirectional transmission

The simulation results of two types of bidirectional transmission systems, “Loop” and

“Straight Line”, are listed below. Each of the case contains the results such as, the received optical spectrum, the BER curves and the corresponding eye diagrams at channel ten, penalties of all channels at 100-km, 200-km, 400-km, 600-km and 800-km transmission distance, respectively.

5.3.1 Case I: Bidirectional loop system

Fig. 5.10 shows the received optical spectrum after 800 km transmission and an OSNR of 30.5 dB with 0.01-nm bandwidth resolution setting is observed. The BER curves of Channel ten after different transmission distance is shown in Fig. 5.11. A more detailed penalty distribution for sixteen channels after various numbers of loops is depicted in Fig.

5.12. The penalty mean of the sixteen channels at transmitted distance 100, 200, 400, 600 and

800 km, are 0.4, 0.7, 1.5, 2.5 and 3.7-dB, respectively. The descended performance is mainly attributed to the accumulated ASE after long transmission distance.

Fig. 5.10. Received optical spectrum after 800 km bidirectional loop transmission

Fig. 5.11 BER curves @ ch.10 after 0, 200, 400, 600 and 800-km bidirectional loop transmission

192.6 192.7 192.8 192.9 193.0 193.1 193.2 193.3 193.4 0

1 2 3 4 5 6 7 8 9 10 11 12

Penalty (dB)

Frequency (THz)

100 km 200 km 400 km 600 km 800 km

Bidirectional Loop Transmission

Fig. 5.12. Power penalty distribution of bidirectional loop transmission

Fig. 5.13 (a) ~ (f) demonstrates the corresponding eye diagrams after 0, 1, 2, 4, 6 and 8 loops of bidirectional transmission.

(a) Eye diagram of 0-km transmission (d) Eye diagram of 400-km transmission

(b) Eye diagram of 100-km transmission (e) Eye diagram of 600-km transmission

(c) Eye diagram of 200-km transmission (f) Eye diagram of 800-km transmission Fig. 5.13 Correspond eye diagrams after (a) 0 km (b) 100 km (c) 200 km (d) 400 km (e)

600 km (f) 800 km bidirectional loop transmission

5.3.2 Case II: Bidirectional straight line system

In Fig. 5.14, the optical spectrum after 800 km transmission with an OSNR of 29.5 dB, which bandwidth resolution sets on 0.01-nm, is observed. Fig. 5.15 plots the BER curves at channel ten after different bidirectional traffic length. The detailed power penalty distribution for sixteen channels after several numbers of spans is illustrated in Fig. 5.16. The mean of receiving penalties for sixteen channels at 100, 200, 400, 600 and 800 km transmission length are 0.4, 0.8, 2.1, 4.1 and 7.1-dB, respectively. The performance degradation is due to the accumulated ASE increasing rapidly as long as the transmission distance is extended.

Fig. 5.14. Received optical spectrum after 800 km bidirectional transmission (straight line)

Fig. 5.15 BER curves @ ch.10 after 0, 200, 400, 600 and 800-km bidirectional transmission (straight line)

192.6 192.7 192.8 192.9 193.0 193.1 193.2 193.3 193.4 0

1 2 3 4 5 6 7 8 9 10 11 12 13 14

Penalt y (dB)

Frequency (THz)

100 km 200 km 400 km 600 km 800 km

Bidirectional Transmission (Staight Line)

Fig. 5.16 Power penalty distribution of bidirectional transmission (straight line)

Fig. 5.17(a) ~ (f) shows the corresponding eye diagrams after 0, 100, 200, 400, 600 and 800-km of bidirectional transmission.

(a) Eye diagram of 0-km transmission (d) Eye diagram of 400-km transmission

(b) Eye diagram of 100-km transmission (e) Eye diagram of 600-km transmission

(c) Eye diagram of 200-km transmission (f) Eye diagram of 800-km transmission Fig. 5.17. Correspond eye diagrams after (a) 0 km (b) 100 km (c) 200 km (d) 400 km (e)

600 km (f) 800 km bidirectional transmission (straight line)

5.3.3 Comparison of the “Loop” and “Straight Line” cases

Fig. 5.18 shows the BER and corresponding eye diagrams after 0 km and 800 km transmission distance for both cases. Furthermore, a more detail power penalty comparison is exhibited in Fig. 5.19. From simulation results, we obtained the penalty after 800 km bidirectional transmission of “loop” and “straight line” cases are 3.7 dB and 7.1 dB, respectively. The average penalties are represented by symbols and the fitting curves for the two cases are displayed as well. Clearly, from Fig. 5.18, the fitting curve for loop case is a 1^st order polynomial function while fitting curve for straight line case is a 2^nd order polynomial function. The results manifest that the power penalty difference between the two cases increases along with the transmission distance.

Fig. 5.18 BER curves and corresponding eye diagrams @ ch.10 for “Loop” and

“Straight Line” cases after 0 km and 800 km bidirectional transmission.

Fig. 5.19. Penalty distribution for “Loop” and “Straight Line” cases.

0 200 400 600 800 1000

0 1 2 3 4 5 6 7 8

Po w er Pe nalty (dB)

Distance (km)

AveragePenalty (Loop) AveragePenalty (SL)

Fitting curve (Loop) Fitting curve (SL) SL: Straight Line

Penalty Distribution

Nonetheless, due to the different times of passing through interleavers in one round trip,

Nonetheless, due to the different times of passing through interleavers in one round trip,

在文檔中 利用無色散間隔器探討雙向光纖傳輸系統 (頁 55-0)