**2. Evaluation of various channel spacings for increasing spectral efficiency of WDM-PON transmission system**

At the moment passive optical networks have been standardized to nextgeneration NG-PON2 accordingly to ITU-T G.989.2 recommendation standards and are widely investigated. Operators are widely deploying time-division multiplexing (TDM) based passive optical networks in urban areas with bitrates up to 10 Gbit/s, but WDM-PON's still are in stage of research [14, 15].

The ITU-T G.694.1 recommendation provides a frequency grid for (WDM) transmission systems and specifies inter-channel intervals. The same frequency grid or channel spacing is used for spectral effectiveness improvement of PON system in our research. Anchored to 193.1 THz (central channel frequency), it supports a variety of inter-mediate channel spacings ranging from narrowed 12.5 GHz to 100 GHz and wider. Depending on the selected step of the inter-channel interval are defined the following abbreviations and acronyms:

WDM—wavelength division multiplexing. CWDM—coarse wavelength division multiplexing. DWDM—dense wavelength division multiplexing.

There are two types of inter-channel interval definitions in (WDM) systems:

Fixed inter-channel interval (fixed grid). Flexible inter-channel interval (flexible grid).

According to ITU-T G.694.1 rec. the minimum step of a fixed channel interval is 12.5 GHz (please see **Table 1**). The flexible channel step is half of the 12.5 GHz, that can be used for the inter-channel interval like 6.25 GHz. Reducing the inter-channel interval leads to increase of crosstalk and non-linear effects (NOE) of transmitted optical signal [16–18].

For research of spectral efficiency increasing, the experimental 2-channel NRZ-OOK modulated 10 Gbit/s bit rate per channel transmission system model was created for Next-generation WDM-PON systems based on tunable wavelength transmitters, please see in **Figure 1**. First step of the research is based on various channel spacing impact on the end user transmitted signal with following fixed 10 Gbit/s transmission speed per channel.

As one can see in **Figure 1**. transmitter (Tx) part of our investigated transmission system model consists of two continuous wave (CW) laser sources—Agilent 81949A, with fixed central frequency 193.1 THz or 1552.524 nm in wavelength, and COBRITE DX-1 laser with tunable central frequency, which can be set the necessary channel spacing. Agilent 81949A continuous wave laser source was connected


**Table 1.**

*Nominal central frequencies grid of the DWDM grid [17].*

*Telecommunication Systems – Principles and Applications of Wireless-Optical Technologies*

Solution for telecommunication infrastructure upgrade and alternative solution for increase of the serial line rate of the NRZ link is to use multi-level signaling formats such as pulse-amplitude modulation (PAM), abbreviated as PAM-M or M-PAM, where multiple digital bits per symbol are encoded into M different signal amplitude levels. The four-level PAM modulation format is receiving significant attention because of its relative ease of implementation in comparison to higher-order modulation formats like quadrature phase-shift keying (QPSK), and m-ary quadrature amplitude modulation (m-QAM). It is clear that M-PAM offers a good trade-off between performance and complexity. Usage of PAM-4 format is effective way to double the data rate of NRZ link. Previously PAM-4 modulation formats have been investigated for application with traditional electrical networks [3, 4], but now researchers are focused on investigation of PAM-4 and M-PAM modulation formats for utilization in optical access networks as well as data center interconnections [5]. Also, there are very limited number of studies which are focused on spectrum slicing and stitching back method, which deals with bandwidth bottleneck problem by slicing the broadband signal in lower-bandwidth signal slices. This spectrum slicing and stitching back method or technique allows transmission of wide bandwidth signals from the service provider to the end user over an optical distribution network via low bandwidth equipment [6, 7]. It is ideally suited for cost sensitive fiber optical access networks where variable bandwidth and scalability as well as flexibility are important. It must be noted that this method is investigated for intensity modulated direct detection NRZ-OOK and duobinary systems, but there are no investigations on its usage together with M-PAM systems [8, 9]. It must be noted that multi-level signaling also changes some rules, which were used in NRZ coded transmission systems. For M-PAM systems it is important to implement more complex and precise level threshold detection for signal inputs, also signal-to-noise (SNR) requirements are higher than in case of NRZ. Eye time skew, amplitude compression in lower eye diagram eyes, intersymbol interference for M-PAM systems also is an issue which must be investigated. So, we can say that PAM-4 links are new science—still learning what impairments create errors in receivers [10, 11]. Significant efforts have been put on investigation of PAM-4 format in fiber optical transmission networks, however there are following aspects, which have not been studied or have been studied insufficiently. High-level PAM modulation techniques, like PAM-4, can dramatically improve the spectral efficiency and available bitrate by using the bandwidth of already existing optical, electro-optical or electrical devices. Minimal available channel spacing (which has direct impact on the utilization of resources like optical spectrum), maximal available number of channels, by wavelength division multiplexing (WDM) technique, maximal transmission distance (network reach) in dispersion compensated and non-compensated M-PAM modu-

**88**

lated WDM-PON optical access systems.

Another way to improve capacity of limited bandwidth is by using duobinary modulation format. Transmission capacity will be increased in comparison with NRZ, utilization of DB will increase the transmission capacity by improving the bandwidth efficiency and reducing channel spacing with this modulation format [12]. Duobinary modulation format is type of proficient pseudo-multilevel modulation format, and therefore is the area of interest due to its increased spectral efficiency. It has been already used to increase the channel capacity by improving the bandwidth utilization in commercial links. The most important feature of duobinary modulation format is its usage for longer transmission distances where it

At first, in the paper we investigate the performance and minimal channel interval of 10 Gbit/s per channel NRZ-OOK (which is basically PAM-2) modulated transmission system, then we investigate PAM-4 and raise the transmission speed up to 20 Gbit/s per wavelength and in the end compare it to NRZ and duobinary modulation formats.

has high tolerance to the influence of chromatic dispersion (CD) [13].

#### **Figure 1.**

*2-Channel NRZ-OOK modulated optical transmission system with 10 Gbit/s transmission speed per channel and flexible channel spacing.*

to the 40G intensity Mach-Zehnder (MZM) modulator, COBRITE DX-1 laser light source was connected to the second MZM intensity modulator. Both laser sources were used with minimal output power +9 dBm for Agilent 81949A and +6 dBm for COBRITE DX-1. To provide the same level of optical power for both optical channels, after the PHOTLINE 40G MZM, an optical attenuator of 3.05 dB insertion loss was additionally attached to the modulator's optical output. Pulse Pattern Generator (PPG) with Pseudo random bit sequence (PRBS9) was used for generation of NRZ coded electrical signals. The external 10 GHz clock signal generator was used in this experiment for as a clock signal source for PPGs. Two electrical PPG non-inverted RF data signal outputs were connected to each of MZMs electrical signal inputs. The data rate for each of the PPGs was 10 Gbit/s throughout the experiment.

ITU-T G.652 standard single mode fiber (SSMF) with dispersion coefficient of 16 ps/(nm × km), and 0.2 dB/km attenuation coefficient was used in optical distribution network. Depending of SSMF fiber span length (20 or 40 km), an Erbium doped fiber amplifier (EDFA) with additional gain was used to provide sufficient optical power level before the PIN photoreceiver.

At the receiver part (Rx), the incoming optical signal was divided by 50% power splitter with 3.5 dB insertion loss. One output of optical power splitter was connected to the optical spectrum analyzer (OSA). Second output of power splitter was connected to the optical band pass filter (BPF) OTF-350 with a tuned 35 GHz 3-dB bandwidth. After BPF filter, fiber Bragg grating dispersion compensation module (FBG DCM), with 3 dB insertion loss was connected for post-compensation purposes of chromatic dispersion (CD). To avoid the maximum optical input optical power level rating of +3 dBm before the 10G PIN photoreceiver (PD) a monitoring power splitter with a power ratio of 10–90% and power meter was used. First channel was filtered out by using optical BPF. As one can see in **Figure 2(a)**, optical spectrum with central channel frequency 1552.560 nm (193.096 THz in frequency)

**91**

*Research of M-PAM and Duobinary Modulation Formats for Use in High-Speed WDM-PON…*

*Central channel spectrum of 2-channel NRZ-OOK modulated optical transmission system with 10 Gbit/s per* 

*channel: (a) after BPF and (b) measured amplited frequency response of BPF.*

is slightly shifted relative to ITU-T G.694.1 rec. Grid central frequency of 193.1 THz. By obtained results from the optical spectrum analyzer (OSA), the BPF pass band is Δλ = 0.280 nm equal to 35 GHz, where λ0 = 1552.564 nm and λ1 = 1552.424 nm.

*Simulation scheme of 2-channel NRZ modulated optical transmission system with 10 Gbit/s transmission speed* 

An eye analyzer was used for measurements of received electrical signal quality. The eyes of received signals for both channels were open, therefore leading to error free transmission. As the eye pattern analyzer for quality measurement use special masks to determine if the signal is above or below necessary quality. We continued our research in OptSim simulation environment by creating relevant simulation

For more precise expected Bit-error-rate (BER) values of received signal the simulation model was created in OptSim simulation software environment. The model used BER estimator based on statistical signal analysis. As one can see in **Figure 3**, simulation scheme implemented in OptSim simulation software for BER measurements has the same setup as experimental system. In the OptSim simulation environment, it is necessary to perform the assembly of used electrical-optical components in order to repeat the 2-channel NRZ-OOK modulated 10 Gbit/s per channel transmission system to research impact of various channel spacings.

According to ITU-T G.694.1 rec., see **Table 2**, during the experiment, the interchannel interval for transmission system was changed from 100 GHz to 25 GHz. We started the experiment at a 20 km long fiber ODN distance with 100 GHz channel spacing. Firstly, the measurements was carried out without the chromatic dispersion (CD) post-compensation, at 20 km fiber link. For transmission over 20 km fiber span we observed negligible chromatic dispersion impact on 10 Gbit/s signal,

The 12.5 GHz channel spacing interval was not obtained in this step of research. The reason for that was too wide filter pass-band, as a result photoreceiver captured

model and using the previously obtained experimental data.

received signal is mainly insignificant impact of dispersion [19].

*DOI: http://dx.doi.org/10.5772/intechopen.84600*

**Figure 2.**

**Figure 3.**

*per channel with flexible channel interval.*

*Research of M-PAM and Duobinary Modulation Formats for Use in High-Speed WDM-PON… DOI: http://dx.doi.org/10.5772/intechopen.84600*

#### **Figure 2.**

*Telecommunication Systems – Principles and Applications of Wireless-Optical Technologies*

to the 40G intensity Mach-Zehnder (MZM) modulator, COBRITE DX-1 laser light source was connected to the second MZM intensity modulator. Both laser sources were used with minimal output power +9 dBm for Agilent 81949A and +6 dBm for COBRITE DX-1. To provide the same level of optical power for both optical channels, after the PHOTLINE 40G MZM, an optical attenuator of 3.05 dB insertion loss was additionally attached to the modulator's optical output. Pulse Pattern Generator (PPG) with Pseudo random bit sequence (PRBS9) was used for generation of NRZ coded electrical signals. The external 10 GHz clock signal generator was used in this experiment for as a clock signal source for PPGs. Two electrical PPG non-inverted RF data signal outputs were connected to each of MZMs electrical signal inputs. The

*2-Channel NRZ-OOK modulated optical transmission system with 10 Gbit/s transmission speed per channel* 

data rate for each of the PPGs was 10 Gbit/s throughout the experiment.

optical power level before the PIN photoreceiver.

ITU-T G.652 standard single mode fiber (SSMF) with dispersion coefficient of 16 ps/(nm × km), and 0.2 dB/km attenuation coefficient was used in optical distribution network. Depending of SSMF fiber span length (20 or 40 km), an Erbium doped fiber amplifier (EDFA) with additional gain was used to provide sufficient

At the receiver part (Rx), the incoming optical signal was divided by 50% power

splitter with 3.5 dB insertion loss. One output of optical power splitter was connected to the optical spectrum analyzer (OSA). Second output of power splitter was connected to the optical band pass filter (BPF) OTF-350 with a tuned 35 GHz 3-dB bandwidth. After BPF filter, fiber Bragg grating dispersion compensation module (FBG DCM), with 3 dB insertion loss was connected for post-compensation purposes of chromatic dispersion (CD). To avoid the maximum optical input optical power level rating of +3 dBm before the 10G PIN photoreceiver (PD) a monitoring power splitter with a power ratio of 10–90% and power meter was used. First channel was filtered out by using optical BPF. As one can see in **Figure 2(a)**, optical spectrum with central channel frequency 1552.560 nm (193.096 THz in frequency)

**90**

**Figure 1.**

*and flexible channel spacing.*

*Central channel spectrum of 2-channel NRZ-OOK modulated optical transmission system with 10 Gbit/s per channel: (a) after BPF and (b) measured amplited frequency response of BPF.*

#### **Figure 3.**

*Simulation scheme of 2-channel NRZ modulated optical transmission system with 10 Gbit/s transmission speed per channel with flexible channel interval.*

is slightly shifted relative to ITU-T G.694.1 rec. Grid central frequency of 193.1 THz. By obtained results from the optical spectrum analyzer (OSA), the BPF pass band is Δλ = 0.280 nm equal to 35 GHz, where λ0 = 1552.564 nm and λ1 = 1552.424 nm.

An eye analyzer was used for measurements of received electrical signal quality. The eyes of received signals for both channels were open, therefore leading to error free transmission. As the eye pattern analyzer for quality measurement use special masks to determine if the signal is above or below necessary quality. We continued our research in OptSim simulation environment by creating relevant simulation model and using the previously obtained experimental data.

For more precise expected Bit-error-rate (BER) values of received signal the simulation model was created in OptSim simulation software environment. The model used BER estimator based on statistical signal analysis. As one can see in **Figure 3**, simulation scheme implemented in OptSim simulation software for BER measurements has the same setup as experimental system. In the OptSim simulation environment, it is necessary to perform the assembly of used electrical-optical components in order to repeat the 2-channel NRZ-OOK modulated 10 Gbit/s per channel transmission system to research impact of various channel spacings.

According to ITU-T G.694.1 rec., see **Table 2**, during the experiment, the interchannel interval for transmission system was changed from 100 GHz to 25 GHz. We started the experiment at a 20 km long fiber ODN distance with 100 GHz channel spacing. Firstly, the measurements was carried out without the chromatic dispersion (CD) post-compensation, at 20 km fiber link. For transmission over 20 km fiber span we observed negligible chromatic dispersion impact on 10 Gbit/s signal, received signal is mainly insignificant impact of dispersion [19].

The 12.5 GHz channel spacing interval was not obtained in this step of research. The reason for that was too wide filter pass-band, as a result photoreceiver captured


#### **Table 2.**

*Experimentally used channel interval according to ITU-T G.694.1 rec.*


#### **Table 3.**

*Channel spacing dependence on the channel interval.*

both channels simultaneously. They did not appear on the Eye Analyzer because it was not possible to synchronize between the transmitter and receiver. After obtaining the results at fixed inter-channel intervals from 100 to 25 GHz, the smallest inter-channel interval at which transmission is possible was found. The step used to search for the inter-channel interval is 6.25 GHz and half of the found step 6.25/2 = 3.125 GHz. Result of channel spacing impact was obtained from channel with fixed central frequency of 193.1 THz = 1552.524 wavelength corresponding to the laser source used by Agilent 81949A. Our transmission system has only two channels, it is not possible to choose a central channel, both channels have mainly the same effect of crosstalk. The channel interval was changed by changing the central wavelength of the second CW laser source with 6.25 and 3.125 GHz step. Instead of experiment for 2-channel NRZ-OOK modulated optical transmission system with 10 Gbit/s transmission speed per channel previously calculated flexible channel interval was used in our research, please see **Table 3**.

Fiber optical transmission system made by the optical components affected by various factors caused by higher attenuation mentioned in specification insertion loss. To create same simulation model in OptSim simulation software environment, it was necessary to adapt model optical elements of the actual loss. In **Figure 4**. we can see BER estimated from the data obtained in OptSim simulation according to different channel intervals.

The BER threshold of 10<sup>−</sup><sup>9</sup> for our investigated transmission system was used to evaluate maximal crosstalk impact between the channels. According to the

**93**

BER < 10<sup>−</sup><sup>3</sup>

**Figure 5.**

**Figure 4.**

*Research of M-PAM and Duobinary Modulation Formats for Use in High-Speed WDM-PON…*

*BER dependence on the channel interval for a 20 km long 2-channel NRZ-OOK modulated optical* 

obtained results channel interval effect up to 30 GHz can be evaluated, higher than used value of BPF filter. Deterioration of the BER used for channel interval less than 30 GHz in our research, can be explained by adjacent channel overlapping. At 20 km long SSMF fiber optical link minimal channel spacing was achieved ensuring

*Comparison of experimental and simulative results: eye diagrams of 20 km 2-channel NRZ-OOK modulated optical transmission system with 10 Gbit/s transmission speed per channel without CD post-compensation: (a) 100 GHz channel spacing, (b) 50 GHz channel spacing, (c) 25 GHz channel spacing, (d) 100 GHz channel spacing in the environment of OptSim, (e) 50 GHz channel spacing in the environment of OptSim, and* 

cal (simulation data) eye diagrams of received signal for second channel with 100,

In second part of our research the length of ODN was increased from 20 to 40 km, by adding 20 km SSMF fiber span. The effect of chromatic dispersion was observed in upgraded transmission system. Fiber Bragg grating dispersion compensation module (FBG DCM) with −640 ps/nm was used for dispersion compensa-

channel spacing according to the obtained results of OptSim simulation software. By performing experiment, the 31.25 GHz inter-channel spacing was the last interval at which mask testing with eye diagram analyzer for received eye diagrams was

50 and 25 GHz channel spacing crosstalk impact, please see **Figure 5**.

tion. The BER value exceeded our defined BER threshold of 1x10<sup>−</sup><sup>9</sup>

threshold at 25 GHz. In **Figure 5**, we can see experimental and theoreti-

at 31.25 GHz

*DOI: http://dx.doi.org/10.5772/intechopen.84600*

*transmission system with 10 Gbit/s transmission speed per channel.*

*(f) 25 GHz channel spacing in the environment of OptSim.*

*Research of M-PAM and Duobinary Modulation Formats for Use in High-Speed WDM-PON… DOI: http://dx.doi.org/10.5772/intechopen.84600*

#### **Figure 4.**

*Telecommunication Systems – Principles and Applications of Wireless-Optical Technologies*

both channels simultaneously. They did not appear on the Eye Analyzer because it was not possible to synchronize between the transmitter and receiver. After obtaining the results at fixed inter-channel intervals from 100 to 25 GHz, the smallest inter-channel interval at which transmission is possible was found. The step used to search for the inter-channel interval is 6.25 GHz and half of the found step 6.25/2 = 3.125 GHz. Result of channel spacing impact was obtained from channel with fixed central frequency of 193.1 THz = 1552.524 wavelength corresponding to the laser source used by Agilent 81949A. Our transmission system has only two channels, it is not possible to choose a central channel, both channels have mainly the same effect of crosstalk. The channel interval was changed by changing the central wavelength of the second CW laser source with 6.25 and 3.125 GHz step. Instead of experiment for 2-channel NRZ-OOK modulated optical transmission system with 10 Gbit/s transmission speed per channel previously calculated flexible

Fiber optical transmission system made by the optical components affected by various factors caused by higher attenuation mentioned in specification insertion loss. To create same simulation model in OptSim simulation software environment, it was necessary to adapt model optical elements of the actual loss. In **Figure 4**. we can see BER estimated from the data obtained in OptSim simulation according to

to evaluate maximal crosstalk impact between the channels. According to the

for our investigated transmission system was used

channel interval was used in our research, please see **Table 3**.

**92**

**Table 3.**

**Table 2.**

*Channel spacing dependence on the channel interval.*

*Experimentally used channel interval according to ITU-T G.694.1 rec.*

different channel intervals. The BER threshold of 10<sup>−</sup><sup>9</sup>

*BER dependence on the channel interval for a 20 km long 2-channel NRZ-OOK modulated optical transmission system with 10 Gbit/s transmission speed per channel.*

#### **Figure 5.**

*Comparison of experimental and simulative results: eye diagrams of 20 km 2-channel NRZ-OOK modulated optical transmission system with 10 Gbit/s transmission speed per channel without CD post-compensation: (a) 100 GHz channel spacing, (b) 50 GHz channel spacing, (c) 25 GHz channel spacing, (d) 100 GHz channel spacing in the environment of OptSim, (e) 50 GHz channel spacing in the environment of OptSim, and (f) 25 GHz channel spacing in the environment of OptSim.*

obtained results channel interval effect up to 30 GHz can be evaluated, higher than used value of BPF filter. Deterioration of the BER used for channel interval less than 30 GHz in our research, can be explained by adjacent channel overlapping. At 20 km long SSMF fiber optical link minimal channel spacing was achieved ensuring BER < 10<sup>−</sup><sup>3</sup> threshold at 25 GHz. In **Figure 5**, we can see experimental and theoretical (simulation data) eye diagrams of received signal for second channel with 100, 50 and 25 GHz channel spacing crosstalk impact, please see **Figure 5**.

In second part of our research the length of ODN was increased from 20 to 40 km, by adding 20 km SSMF fiber span. The effect of chromatic dispersion was observed in upgraded transmission system. Fiber Bragg grating dispersion compensation module (FBG DCM) with −640 ps/nm was used for dispersion compensation. The BER value exceeded our defined BER threshold of 1x10<sup>−</sup><sup>9</sup> at 31.25 GHz channel spacing according to the obtained results of OptSim simulation software. By performing experiment, the 31.25 GHz inter-channel spacing was the last interval at which mask testing with eye diagram analyzer for received eye diagrams was

#### **Figure 6.**

*Comparison of experimental and simulative results: eye diagrams of 40 km 2-channel NRZ modulated optical transmission system with 10 Gbit/s transmission speed per channel with CD post-compensation: (a) 50 GHz channel spacing, (b) 31.25 GHz channel spacing, (c) 25 GHz channel spacing, (d) 50 GHz channel spacing in the environment of OptSim, (e) 31.25 GHz channel spacing in the environment of OptSim, and (f) 25 GHz channel spacing in the environment of OptSim.*

#### **Figure 7.**

*Comparison of experimental and simulative results: eye diagrams of 40 km 2-channel NRZ modulated optical transmission system with 10 Gbit/s transmission speed per channel without CD post-compensation: (a) 50 GHz channel spacing, (b) 31.25 GHz channel spacing, (c) 25 GHz channel spacing, (d) 50 GHz channel spacing in the environment of OptSim, (e) 31.25 GHz channel spacing in the environment of OptSim, and (f) 25 GHz channel spacing in the environment of OptSim.*

#### **Figure 8.**

*BER dependence on channel interval for a 40 km 2-channel NRZ-OOK modulated optical transmission system with 10 Gbit/s transmission speed per channel.*

**95**

**Figure 9.**

*per wavelength.*

*Research of M-PAM and Duobinary Modulation Formats for Use in High-Speed WDM-PON…*

possible [20]. By obtained experimental and simulation results it can be concluded that the model of optical transmission created in the simulation environment corresponds to the experimental fiber optic transmission system. Channel overlaps at 40 km long fiber section, with use of dispersion compensation, see **Figure 6** and without dispersion compensation see **Figure 7**. Results, with BER below our inter-

**3. Evaluation of PAM-4 modulation format use in WDM-PON systems**

(FEC) code for 10 Gbit/s PONs [21, 22]. The theoretical FEC relationship restores

in **Figure 9**, the PAM-4 modulated WDM-PON simulation scheme was created in OptSim simulation software environment. Here the Matlab software was used for BER estimation of received PAM-4 signals. WDM-PON simulation model consists of 4 channels, with central frequency 193.1 THz for second channel and chosen 50 or 100 GHz, according to the previously mentioned ITU G.694.1 rec. According to our previously channel interval research of flexible channel spacing like 37.5 and 25 GHz also was realized. However, the quality of received signal was low, with crosstalk impact and error-free transmission was not possible, performance was

. We evaluated the performance of WDM-PON architecture in terms of maximal transmission reach. Optical line terminal (OLT) is located in central office (CO) and consists of four transmitters (OLT\_Tx). Each OLT\_Tx transmitter consists of two pseudo-random bit sequence (PRBS) generators and NRZ drivers, as a result two

*Simulation scheme of 4-channel PAM-4 modulated WDM-PON transmission system operating at 10 Gbaud/s* 

pre-FEC BER to a 10<sup>−</sup>12 post-FEC in the PON standards. As it is shown

In our research we investigated the 4-channel 10 Gbaud/s (20 Gbit/s) per channel PAM-4modulated WDM-PON access system with minimal allowable channel spacing, which has a direct impact on the utilization of resources like optical spectrum. The research was made with and without fiber chromatic dispersion (CD) fiber Bragg grating compensation module (FBG DCM). We evaluate system performance and found the maximal transmission distance for multichannel PAM-4 modulated WDM-PON transmission system operating at 20 Gbit/s per channel. In OptSim simulation software we created transmission system model to evaluate the performance of 4-channel PAM-4 modulated WDM-PON transmission system operating at 10 Gbaud/s or 20 Gbit/s per channel under the condition with BER

, by use of Reed Solomon (RS 255,223) forward error correction

was exceeded at the 31.25 GHz channel

*DOI: http://dx.doi.org/10.5772/intechopen.84600*

channel interval, please see **Figure 8**. Our defined BER threshold of 1 × 10<sup>−</sup><sup>9</sup>

above our defined BER threshold 1 × 10<sup>−</sup><sup>3</sup>

threshold of 10<sup>−</sup><sup>3</sup>

1.1 × 10<sup>−</sup><sup>3</sup>

interval where the BER of received signal was 7.4 × 10<sup>−</sup>11.

*Research of M-PAM and Duobinary Modulation Formats for Use in High-Speed WDM-PON… DOI: http://dx.doi.org/10.5772/intechopen.84600*

possible [20]. By obtained experimental and simulation results it can be concluded that the model of optical transmission created in the simulation environment corresponds to the experimental fiber optic transmission system. Channel overlaps at 40 km long fiber section, with use of dispersion compensation, see **Figure 6** and without dispersion compensation see **Figure 7**. Results, with BER below our interchannel interval, please see **Figure 8**.

Our defined BER threshold of 1 × 10<sup>−</sup><sup>9</sup> was exceeded at the 31.25 GHz channel interval where the BER of received signal was 7.4 × 10<sup>−</sup>11.
