Performance of the long distance transmission

At first, the transmission performance of the APSK system as a function of the transmission distance focusing on different ER of the ASK signal was measured. Figure 4.8 shows the results. A clear trade off between the ASK signal and the PSK signal in the APSK system was observed. This experimental result fits qualitatively well with the simulation results in chapter 2.

0 500 1000 1500 2000 2500 3000 3500

Transmission distance (km)

Fig. 4.8 BER as a function of the transmission distance 4.3.3 Performance of the APSK format using the zero-nulling method

The performance of the zero-nulling method was evaluated after 500km transmission.

Figure 4.9 shows the BER performance of the ASK signal in the APSK system as a function of the receiver input power with different ER. As seen in the figure, high ER showed better performance. After the transmission, the performance was degraded due to the ASE noise.

The BER performance of the PSK signal in the APSK system is shown in figure 4.10. As shown in this figure, the BER performance of the PSK signal using the zero-nulling method was clearly improved regardless to the back to back situation and after 500km transmission, even when the ER was equal to 10dB.

-37 -32 -27 -22 -17 -12 -7

Fig. 4.9 Performance of the ASK signal in the APSK system

-37 -32 -27 -22 -17 -12 -7

(b) after 500 km transmission

Fig. 4.10 Performance of the PSK signal in the APSK system

As shown in figure 4.10, the PSK signal using the zero-nulling method exhibited a clear improvement compared to the conventional APSK format. As mentioned in chapter 2, it is impossible to use the delay demodulation in this scheme, but a measurement technique using the delay demodulation was developed and the effectiveness of the zero-nulling method was confirmed experimentally. For the actual implementation of this method, it is needed to find some method to obtain the phase information imposed only when the ASK signal is equal to

“one”.

4.4 Conclusion

The performance of the long-haul APSK system as a function of the ER was confirmed through the experiment in this chapter. A clear trade off between the ASK performance and the PSK performance was observed. The qualitative measurement of the zero-nulling method was conducted in this chapter, and the performance of the APSK system was clearly improved by using the zero-nulling method.

References in this chapter

[1] Anritsu MU181040A Operation Manual

[2] G. P. Agrawal, Fiber-optic communication systems, Wiley Interscience, Third Edition.

Chapter 5 Discussion of simulation and experimental results

5.1 Introduction

This chapter discusses the results obtained so far. At first, the results of the numerical simulations are discussed in section 5.2. Section 5.3 discusses the results of the experimental investigations. Some phenomenon shown in the experimental results are pointed out and discussed what kind of mechanism causes this kind of result. Section 5.4 compares the results of numerical simulations and experimental investigations. Some results of the experiments agree the numerical simulations, but some are not. The reasons of this kind of discrepancies are discussed in this section.

5.2 Discussion of the simulation results

Figure 5.1 shows the simulation result of the transmission performance of the APSK signal as a function of the transmission distance with different ER of the ASK signal, which was already explained in chapter 2. As seen in this figure, the transmission performance degrades as the transmission distance extends. Because the chromatic dispersion was compensated by the dispersion map and the receiving terminal, the degradation was mainly due to the accumulated ASE noise. The accumulated ASE noise degraded the SNR, and the SNR degraded the transmission performance. Figure 5.2 summarizes the performance of the APSK signal after 1500km transmission. The ER was set from 3dB to 13dB. The transmission performance of the ASK signal became better when the ER increased, while the transmission performance of the PSK signal became better when the ER of the ASK signal decreased.

These results show a clear trade-off between the ASK performance and the PSK performance. The reason of the PSK performance degradation could be attributed to the degradation of the phase information in the space signal of the ASK format. In the definition, the optical power of the mark level was fixed. Therefore, high ER meant the optical power of the space level was small. As the space signal of the ASK format suffered the effect of the optical amplifier noise more severely, the information of the PSK format on this part was damaged, and the overall performance of the PSK signal was degraded. As shown in figure 5.2, the performances of the ASK signal and the PSK signal could be compromised. The optimum extinction ratio for 1500km transmission should be around 5 to 6dB.

0

0 500 1000 1500 2000 2500 3000 3500 4000

3dB 4dB 5dB 6dB 7dB 10dB 13dB

ASK PSK

Transmission distance (km)

Fig. 5.1 Simulated transmission performance of the APSK signal

0

Extinction ratio of ASK signal

Q-factor (dB)

ASK PSK

Fig. 5.2 Simulated transmission performance of the APSK signal after 1500km transmission

In figures 5.1 and 5.2, the performances of the APSK signal as a function of the ER and the transmission distance are described. Then, next step is the discussion of the zero-nulling method. Figure 5.3 shows the transmission performance of the APSK signal with and without the zero-nulling method. For the ASK signal, the performance of the original ASK signal (7.5G) was slightly better than the zero-nulling APSK because of the noise reduction

corresponding to the bit-rate reduction. The performance of 7.5G ASK signal was slightly better than 10G ASK signal. On the other hand, the PSK performance of the zero-nulling APSK signal was greatly improved compared to the original APSK signal. This simulation result showed that the zero-nulling method was proved to be quite effective to improve the long-distance transmission performance of the APSK format. On the other hand, because there was no PSK information in the ASK space level, the transmission capacity was reduced compared to the original APSK system.

0

0 500 1000 1500 2000 2500 3000 3500 4000

Transmission distance (km)

Fig. 5.3 Transmission performance of the APSK signal with and without the zero-nulling method

5.3 Discussion of the experimental results

This section focuses on the results investigated experimentally. The phenomenon of the transmission performance is discussed in this section. Figure 5.4 shows the BER performance as a function of the transmission distance and the ER. As seen in this figure, the performance of the ASK signal was getting better as the ER was increased. On the other hand, the performance of the PSK signal was degraded as the ER was increased. This experimental result showed a clear trade off between the ASK signal and the PSK signal.

Next, the performance of the zero-nulling method was evaluated after 500km transmission. Because the bit window function was not compatible with the burst mode measurement required for the recirculating loop experiment, only the straight line transmission (500km) result was measured. Figure 5.5 shows the BER performance of the PSK signal in back to back situation and after 500km transmission. The performance was

evaluated as a function of the ER of the ASK signal. As seen in figure 5.5(a), the performance of the PSK signal degraded as the extinction ratio was increased. It should be noted that the power penalty of 6dB ER case was the largest among five cases. When the ER was high, the PSK information in the ASK space level suffered significant degradation due to the small optical power.

0 500 1000 1500 2000 2500 3000 3500

Transmission distance (km)

Fig. 5.4 BER performance as a function of the transmission distance

The performance after 500km transmission is shown in figure 5.5(b). The performance was degraded significantly compared to the back to back situation. Less than 10^-9 BER could be achieved only when 3 dB ER case and with the zero-nulling method case. The error floor was clearly observed when the ER was 4, 5, and 6 dB. It could be explained that the PSK signal in the ASK space level suffered signal to noise ratio degradation due to the accumulated ASE noise, and it caused significant transmission penalty to exhibit an error floor after 500km transmission.

As a matter of fact, we could only achieve synchronous loss both in back to back situation and after 500 km transmission when the extinction ratio was 10dB. On the other hand, we could achieve less than 10^-9 BER even when the extinction ratio was 10dB with the zero-nulling method. From this result, it can be said that zero-nulling method was quite effective to improve the transmission performance of the APSK system.

Figure 5.6 shows the eye diagrams of the PSK signal in different ER case. It was easy to observe the phenomenon of eye closure in high ER case. Those figures provided an evidence of the PSK information degradation in the ASK space level due to the small optical power.

After the transmission, the degradation became more significant due to the accumulated ASE noise in the transmission line. When the ER was 10dB, the eye diagram was almost fully closed. This was the reason why we could only achieve synchronous loss both in back to back situation and after 500 km transmission when the extinction ratio was 10dB.

Figure 5.7 shows the eye diagrams of the PSK signal with the zero-nulling method.

These eye diagrams correspond to the special pattern of 32 bits sequence. As the pattern has ten consecutive mark bits, ten consecutive space bits, and six consecutive mark and space bits pair in the ASK signal, a clear eye opening of nine consecutive bits was observed in figure 5.7(a). In addition, an intermediate eye opening area was observed before the large eye opening area. As 1-bit delay demodulation scheme were used in this experiment, consecutive mark bits of the ASK signal showed the large eye opening area, and ten consecutive mark bits of the ASK signal corresponded to nine consecutive bits of the PSK signal after the demodulation. The intermediate eye opening area was generated by six consecutive mark and space bits pair in the ASK signal. Consecutive space bits of the ASK signal showed the worst eye opening area. Figure 5.7(b) shows the eye diagram after 500km transmission. Because the ASE noise caused significant signal to noise ratio degradation, only large eye opening area and no eye opening area were observed. Those figures also show the degradation of the PSK information in the ASK space level clearly.

-37 -32 -27 -22 -17 -12 -7

(b) after 500 km transmission

Fig. 5.5 Performance of the PSK signal in the APSK system

(a) ER 3dB, back to back

(b) ER 3dB, after 500km transmission

(c) ER 5dB, back to back

(d) ER 5dB, after 500km transmission

(e) ER 6dB, back to back

(f) ER 6dB, after 500km transmission

(g) ER 10dB, back to back

(h) ER 10dB, after 500km transmission

Fig. 5.6 Eye diagram of the PSK signal in different ER

(a)back to back

(b) after 500km transmission

Fig. 5.7 Eye diagram of the PSK signal using zero-nulling method

5.4 Comparison of the results of the theoretical studies and the experimental investigations

Comparison of the results of the simulations and the experiments is the most important issue to discuss. There were several phenomena of the experiments that agree to the theoretical simulations. At first, the effect of the ER was confirmed both theoretical simulations and experimental studies. High ER degraded the performance of the PSK signal.

It was because the PSK information suffered damages in the ASK space level. After the transmission, the degradation was more significant due to the accumulated ASE noise. As high ER meant the difference between the mark level and the space level was large, it was easy to distinguish each other compared to low ER case. Therefore, the ASK signal had a good performance when the ER was high. On the other hand, when the ER was small, the performance of the PSK signal was improved, and that of the ASK signal was degraded due to the small difference between the mark level and the space level. Therefore, it was easy to observe a clear trade off between the ASK signal performance and the PSK signal performance in the APSK system. Those results were confirmed both in theoretical simulations and in experimental studies and they were shown in figures 5.1, 5.2 and 5.4.

Second, because the small power of the ASK space level causes the degradation in the PSK signals, the zero-nulling method is proposed to improve the transmission performance.

The performance improvement was confirmed through the numerical simulations, and it was also confirmed in the experiments. As seen in figures 5.3 and 5.5, it was easy to observe a clear improvement for both in theoretical simulations and in experiments using the zero-nulling method. It should be noted that the zero-nulling method used in the experiment was the case when the ER was equal to 10dB. In original APSK format, the performance could only achieve to synchronous loss when the ER was 10dB. Even though, the difference of the ER did not affect the performance of the PSK signal using the zero-nulling method.

Because the optical power of the ASK mark level was fixed, different ER meant the optical power in the ASK space level was different. As the zero-nulling method ignored the PSK information in the ASK space level, different ER did not affect the performance of the PSK signal using the zero-nulling method.

On the other hand, some results of the experiments did not agree to the numerical simulation results. For the numerical simulations, the performance of the APSK system could be calculated when the ER was between 3dB and 13dB with wavelength division multiplexing (WDM) system of 16 channels. Even though, the performance of the experiments was so degraded that only single channel measurement was actually possible. In addition, when the ER was larger than 8dB, it was very difficult to measure the APSK system performance experimentally. As seen in figure 5.6, the eye diagram was almost fully closed

when the ER was 10dB, and the performance could not be measured even in the back to back situation. The reason might be attributed that the simulation was assuming the ideal performance of the transmitter and the receiver.

Then, an optimized point of the ER was obtained to balance the transmission performance of the ASK signal and the PSK signal. For the results of the numerical simulation, the ER of 5.5dB was the optimized point. On the other hand, the optimized point was the ER of 4dB for the experimental results. The reason might also be attributed that the simulation was assuming the ideal performance of the transmitter and the receiver.

When the zero-nulling method was adopted, both the numerical simulation and the experiment showed the improvement compared to the original APSK system. As seen in figure 5.3, the PSK signal of the APSK system using the zero-nulling method could transmit the signal more than 3000km in the numerical simulation. In the experiment, as the experimental setup was limited by the test equipment, the measurement of the zero-nulling method was only possible without the recirculating loop. In other words, only 500km transmission was possible for the measurement of the zero-nulling method. Based on the result of 500km transmission, the transmission distance of the PSK signal in the APSK system using the zero-nulling method should be longer than 500km, but qualitative evaluation was impossible at this moment.

Finally, according to the result of previous chapters, the effect of the ER was confirmed both in the numerical simulation and the experimental investigation. The zero-nulling method could improve the transmission performance of the APSK system based on the numerical simulation, and the possibility to improve the transmission performance of the APSK format using the zero-nulling method was confirmed in the experiment.

Chapter 6 Conclusions

The APSK format is one attractive modulation format because the spectral efficiency can be doubled compared to the conventional ASK format. In this master thesis, the investigation of the APSK format as a function of the ER in the long-haul optical fiber communication system was conducted. The numerical simulation results showed that the ASK signal had better performance when the ER was high, and the performance degraded as the ER decreased. High ER meant the difference between the mark level and the space level was large, and it was easy to distinguish each other compared to low ER case. In other words, eye opening was clearer in high ER case. On the other hand, high ER degraded the performance of the PSK signal. The degradation of the PSK information occured in the ASK space level. Considering that the optical power of the mark level was fixed to the same amount for any ER, high ER had less optical power in the space level. Small optical power degraded the transmission performance of the PSK signals. The small optical power not only degraded the transmission performance, but also suffered more significant effect from the accumulated ASE noise. Therefore, high ER degraded the transmission performance of the PSK signal, and low ER had better performance compared to high ER case. This kind of phenomenon was also showed in the results of the experimental investigations. The experimental results agreed to the numerical simulation results. A clear trade off between the ASK signal performance and the PSK signal performance in the APSK modulation format was clearly observed, and an optimized point of the ER was found in the APSK system. The optimized point of the ER could balance the system performance of the ASK signal and the PSK signal in the APSK systems.

In order to improve the transmission performance of the APSK system, the issues caused by the ER should be solved. A method named “zero-nulling“ was proposed to solve this problem. Because the PSK information suffered degradation in the ASK space level, a method to ignore the PSK information in the ASK space level was proposed. In other words, only when the ASK signal was the mark, the PSK information was transmitted. When the ASK signal was the space, the PSK signal was not transmitted. Therefore, the degradation of the PSK signal due to small optical power in the ASK space level could be solved. As already mentioned, the performance of the PSK signal suffered degradation when the ER was increased. This problem disappeared after using the zero-nulling method. Even though the PSK signal suffersed degradation due to small optical power in the ASK space level, this PSK signal was ignored after using the zero-nulling method. Assuming the same optical power of the ASK mark level, different ER changed the optical power of the ASK space level.

Therefore, different ER did not degrade the PSK signal using the zero-nulling method. The