Experimental investigation of the effect of the SPM

In the previous section, the XPM effect is confirmed through the experiment. In this section, the SPM effect is evaluated by experiment. It is very difficult to eliminate the SPM effect in the actual experiment because the SPM is caused by the signal intensity itself. To investigate the effect of the SPM upon the transmission performance, following experimental configuration was used. The repeater output power was set to +8dBm, but the power of the selected channel to be measured was reduced by 1dB step compared to the other channels. If the channel power is reduced by 1dB, the measurement channel has 1dB smaller signal power, and this signal power corresponds to the power of +7dBm repeater output power. Figure 3.20 shows a simple explanation of this configuration. The measured channel power was reduced by 0dB to 6dB with 1dB step, and it corresponded to +8dBm to +2dBm repeater output power. In this condition, we can confirm the SPM effect because the signal power was reduced and the SPM effect also reduced.

Figure 3.21 shows the measured OSNR for different repeater output power. The dash line was the original experimental results, it means the power of every channel was the same and used the different repeater output power from +2dBm to +8dBm.

The direct line was using fixed +8dBm repeater output power but the measurement channel power was reduced by 1dB step to realize the same signal power as the different repeater output power, and the other channel powers were maintained to the

original power. From this figure, it can be observed that the OSNR of the measured channel was improved for the fixed repeater output power case shown by the direct line. This is because the repeater output power was increased to +8dBm and other WDM channels had the corresponding power. Figure 3.22 shows the performance after 5000km transmission. The direct line shows the performance of the fixed repeater output power, and the dash line shows the performance of the variable repeater output power. From this figure, it can be observed that the performance of the fixed repeater output power case was improved significantly when the signal power was suppressed by a few dB. The best performance was occurred when the signal power was decreased by 4dB. This condition corresponded to +4dBm repeater output power. The OSNR at +5dBm repeater output power was better than that of this case, but the Q-factor of this case was significantly better than that of +5dBm repeater output power with flat WDM signal power case. This result implied that the reduction of the SPM could improve the transmission performance significantly. It can be said that the simulation result have been confirmed by the experimental result.

Figure 3.20 Explanation of the SPM effect experimental setup

1dB reduced

Measurement channel Other channels

8

Figure 3.21 Measured OSNR after 5000km transmission

7

Figure 3.23 Measured Q-factor after 5000km transmission

3.3.5 Discussion

This experiment used the conventional map of the simulation in chapter 2. From these experimental results, there were a few important observations. First, the optimum repeater output power was around +5dBm. Second, the performance was degraded near the system zero dispersion wavelength, but it was improved when the wavelength was shifted from that wavelength. Third, the XPM effect did not cause so significant performance degradation in the transmission. Forth, it looked like that the SPM effect caused significant performance degradation in the transmission.

In the experiment, 5000km transmission was achieved using the conventional map with the RZ-DPSK format. Through the experiment, the simulation results of chapter 2 was qualitatively confirmed.

3.4 Conclusion

Experimental study of the transmission performance of the long-haul RZ-DPSK system was conducted using the recirculating loop setup. The obtained results were basically reasonable compared to the theoretical simulation results in chapter 2.

Therefore, it can be concluded that the nonlinear effects are important factors for the long-haul RZ-DPSK system.

Chapter 4 Conclusion

In this master thesis, the difference of the transmission performance of the long-haul RZ-DPSK system due to the different dispersion map and repeater output power was studied through the numerical simulation and experiment.

In the simulation, the numerical simulator which was designed to isolate the nonlinear effects was used to evaluate the transmission performance of the long-haul RZ-DPSK system. The simulation was done to use conventional dispersion map and new dispersion map for the different repeater out put power. First, the effect of different repeater output power for the conventional map and the new map was investigated. The repeater output power was varied between +10dBm and +15dBm with 1dB step. The best averaged Q-factor for the conventional map and the new map was occurred with +12dBm repeater output power. For the conventional map, the Q-factor was degraded faster than the new map at more than +12dBm repeater output power. This result implies that the new map has better tolerance to the optical fiber nonlinear effect and also shows the clearly evidence that the transmission performance of the long-haul RZ-DPSK system could be improved by adopting the new map instead of the conventional map. Second, transmission performance of +12dBm repeater output power was investigated for both dispersion maps because it was the best repeater output power. From the result, it was observed that the SPM was the major reason of the performance degradation and the XPM was not. For the conventional map, the performance dip was observed near the center wavelength but there was no significant performance degradation in the edge region. For the new map, there was no significant performance degradation near the center wavelength.

To confirm the theoretical simulation, the experiment was conducted. The repeater output power was varied from +2dBm to +8dBm with 1dB step. The best Q-factor was 8.0dB at +5dBm repeater output power. The simulation result using the conventional map was confirmed through the experiment. The result showed that the performance had the same tendency with the simulation. The Q-factor of near the edge wavelength was better than the center wavelength. For the optical fiber nonlinear effect, the XPM and the SPM effects were investigated. The Q-factor without XPM effect was a little better than that with the XPM effect but almost the same. For the SPM effect, the performance with reduced SPM was improved significantly than that with the SPM. From these experimental results, it can be concluded that the SPM is the most important factor in the conventional map.

In this master thesis, followings were achieved.

[1] The transmission performance of the long-haul RZ-DPSK system with the conventional dispersion map and the new dispersion map was evaluated through the numerical simulation.

[2] The new map was confirmed to have better transmission performance than the conventional map especially in higher repeater output power region.

[3] The major reason of the performance penalty of the long-haul RZ-DPSK system was the SPM effect and the minor reason was the XPM effect.

[4] The coincidence of the theory and the experiment was confirmed.

As these achievements are significant enough to contribute the progress of the optical fiber communication system, this master thesis was successful.