Bidirectional long-reach PON using Kramers-Kronig-based receiver for Rayleigh Backscattering noise and SSBI interference elimination

Di Li; Quan Yu; Lei Deng; Liang Huo; Songnian Fu; Ming Tang; Mengfan Cheng; Minming Zhang; Deming Liu

doi:10.1364/OE.26.019020

1. Introduction

To comply with the development of 100 Gigabit Ethernet (100 GbE), there are growing interests in the 100 Gbps passive optical network (PON) for longer transmission distance and more users [1–4]. For example, Huawei has unveiled its 100 Gbps PON technology using a hybrid architecture of both time division and wavelength division to support 4x25 Gbps downstream and 4x10 Gbps upstream [3]. By using ultra dense wavelength division multiplexed (UDWMD)-PON architecture, 320 Gbps (256x1.25 Gbps) PON has also been demonstrated in the COCONUT project [4]. However, these PONs are not intended to deliver 100 Gbps service to each subscriber, which is also called maximum per wavelength transmission speed. To meet the requirements of high-speed per wavelength and large cover range, not only coherent detection technology [5–7] which could provide for increased power budget and much higher channel density, but also amplitude and phase modulation technique based on optical IQ modulator [8] which could effectively increase the spectral efficiency (SE) can play a significant role in the access network. However, compared to long haul and metro, the corresponding design of coherent transceivers must be redefined and optimized for low-cost and power-efficient requirements in access network. First, low-cost and small-footprint lasers rather than expensive narrow linewidth external cavity laser are desired as transmitter and local oscillator (LO). Second, a miniaturized and low-cost optical IQ modulator is another key component, and silicon photonics based modulator is applicable for small optical modules and suited for low-cost mass production [9,10]. Third, low-cost receiver based on single photodiode (PD) rather than costly standard coherent receiver composed of one local oscillator, one optical 90° hybrid, four PDs is preferred. Moreover, as a cost-effective solution, wavelength reuse can avoid using wavelength-specific lasers in optical network units (ONUs) and achieve colorless ONUs.

However, there are some issues need to be addressed to realize low-cost coherent receiver mentioned above. For example, when only one PD is used for receiver, the transmission speed and cover range are significantly restricted by nonlinear distortion induced by square-law detection, low tolerance to chromatic dispersion, and signal-to-signal beating interference (SSBI) [11,12]. To solve this problem, a low chirp electro-absorption modulator (EAM) is used to relieve the power fading problem induced by fiber dispersion [13]. By increasing the launch power and the corresponding self-phase modulation (SPM) effect to combat dispersion-induced power fading, the transmission bandwidth can be extended and the loss budget can be improved in [14]. To eliminate SSBI, a blank spectral band is usually used to separate the SSBI terms and the desired signal in frequency domain [15,16], but the SE is reduced by half. To avoid the SE loss, there are two types of SSBI elimination methods [17–25]. In the first kind of SSBI mitigation schemes, the SSBI is estimated and then subtracted from the received signal. By this means, the SSBI estimation and cancellation techniques are realized based on the single-stage linearization filter in [17], the two-stage linearization filter in [18], and the iterative linearization filter in [19]. Another perspective is to use the Kramers-Kronig (KK) coherent receiver [20–26], which can reconstruct the phase of the received optical signal from the detected intensity when the minimum phase condition is satisfied. It has been proved that the KK scheme could reconstruct the transmitted single-sideband signal with minimum SSBI influence [25]. More importantly, it is easy to compensate the accumulated dispersion of the fiber link digitally due to the use of KK scheme [20].

Moreover, when a single fiber is used for bidirectional transmission in optical access networks, Rayleigh backscattering noise (RB) becomes another big obstacle. Currently, many schemes have been proposed [27–29] for the RB noise mitigation. In [28], direct detection is adopted for downlink and coherent detection is used for uplink to relieve the impact of RB noise. Carrier suppressed single-sideband signals are produced for upstream transmission by using a dual-drive Mach-Zehnder modulator (MZM) [27] or a dual-parallel MZM [29], and the RB noise could be effectively mitigated due to the fact that the spectral overlap between the distributed carrier and the upstream signal is reduced. However, in these schemes radio frequency (RF) signal synthesizer is indispensable and the frequency gap between optical carrier and the upstream signal will cause the SE loss.

On the other hand, multicore fiber (MCF) offers an alternative solution to increase the fiber density and end users in optical access networks [30–32]. A 7-core fiber based optical access network using 2.5 Gbps TDM-PON technologies has been proposed in [30]. In our previous work [31], hybrid wavelength-space division multiplexing optical access network with wavelength reuse technique has been demonstrated for bidirectional transmission. To eliminate the RB noise, the traffic for downstream and upstream are transmitted via independent physical channels, resulting in the increase of system complexity. Moreover, it has been proved that self-homodyne coherent detection (SHCD) is a feasible way to mitigate the stringent requirement of narrow linewidth LO and complex DSP [33]. MCF obviously has some advantages for SHCD [32], for example, one core could be used for transmit the LO and the others can still be used for signals, resulting in negligible SE penalty.

To simultaneously cope with these issues mentioned above and realize high-capacity, spectral-efficiency and low-cost optical access network, a novel bidirectional long reach PON (LR-PON) has been experimentally proposed and demonstrated by using KK-based receiver and MCF in this paper. For low-cost consideration, self-homodyne coherent detection using only one PD is adopted at both OLT and ONU. A single 7-core MCF is adopted in this system, where the middle core is used to deliver the seed light to ONUs for colorless upstream transmission, and the upstream and downstream signals are transmitted simultaneously over the same outer core for each ONU. Therefore, the signals and LOs for the receivers at both OLT and ONU all originate from the same laser which is located at OLT for each wavelength. Thanks to the KK-based receiver, the impact of SSBI is effectively eliminated and fiber dispersion can be digitally compensated, resulting in the increase of SE and cover range. Moreover, with the help of single-sideband Nyquist-shaped subcarrier modulation with 16-ary quadrature amplitude modulation (SSB-Nyquist-16QAM), the upstream and downstream signals can occupy the left and right sideband of the optical carrier respectively, in this way the RB noise can be easily removed by using a simple optical filter in the receiver. The SSBI mitigation and fiber dispersion compensation scheme based on KK receiver is analyzed and verified by a simulation of 100 Gbps SSB-Nyquist-16QAM signal transmission over 80 km standard single mode fiber (SSMF). Furthermore, 60 Gbps SSB-Nyquist-16QAM signals are experimentally transmitted over 50 km 7-core fiber in both uplink and downlink. The key factors such as the carrier-to-signal power ratio (CSPR), the residual RB noise and the required frequency gap between the upstream and downstream signals are investigated in our experiment. The experimental results show that a frequency gap of 3 GHz is enough for bidirectional transmission, and 2.3 dB and 1.4 dB optical signal to noise ratio within signal bandwidth (OSNR_BW) penalty due to 50 km 7-core fiber is observed for upstream and downstream transmission respectively.

2. Principle

2.1 The proposed LR-PON using KK-based receiver and multi-core fiber

Figure 1 illustrates the proposed bidirectional LR-PON using KK-based receiver and N + 1 core fiber, where N represents the number of outer cores of the used MCF. At the central office (CO), a continuous wave (CW) laser with wavelength λ₁ is first divided into N + 1 equal parts by a power splitter (PS). One of these N + 1 parts is delivered to ONUs as the seed light by using the middle core (Mid-core), the optical spectrum of this seed light is shown in the inset (a) of Fig. 1. There are N OLTs at the CO, and in each OLT the input optical signal is then split into two parts by an optical coupler (OC). The first part is modulated by the downstream signal, and the other part is used as the local oscillator (LO) for KK-based upstream signal detection. In our scheme, an optical in-phase and quadrature (IQ) modulator and SSB-Nyquist-16QAM modulation format is used for downstream transmission in each OLT. The optical spectrum of the generated optical downstream signal is shown in the inset (b) of Fig. 1, and obviously the downstream signal is located at the right sideband of λ₁ and its optical carrier is suppressed. After an optical circulator (CIR), the generated downstream signal is launched into an outer core with the help of fan-in device in each OLT. It should be noticed that by using the wavelength division multiplexing (WDM) technique and devices (λ₁~λ_k) for each outer core, the transmission capacity of each OLT could be further improved.

Fig. 1 Schematic of the proposed bidirectional LR-PON using KK-based receiver and multi-core fiber.

Download Full Size | PDF

After MCF propagation and N + 1 core fiber fan-out device, the distributed seed light is split into N parts and sent to ONUs in remote node (RN). In each ONU, the input seed light is split into two parts by an OC, one part is served as optical carrier for upstream transmission, and the other part is used as LO for KK-based downstream signal detection. For upstream transmission, the same SSB-Nyquist-16QAM modulation format is used to realize symmetrical transmission, and unlike the downstream signal, the upstream signal is modulated on the left sideband of λ₁ by another optical IQ modulator to relieve the impact of RB noise as shown in the inset (e) of Fig. 1. Subsequently, the generated carrier-suppressed upstream signal is launched into the same outer core as the downstream signal after a CIR.

For low-cost consideration, the CW laser at OLT is used as the seed light for wavelength reuse at ONU, remote LO for KK receiver at ONU, optical carrier for downstream and LO for KK receiver at OLT in our scheme. Taking 50 km 7-core fiber for example, the optical loss of seed light from OLT to ONU is about 35 dB (8.4 dB for 1:7 optical splitter, 10 dB for 50km 7-core fiber, 6 dB for Fan-in and Fan-out devices, 7.7 dB for 1:6 optical splitter, and 3 dB for 2x2 optical coupler). Therefore, optical amplifier is indispensable. In our scheme, two erbium-doped fiber amplifiers (EDFAs) are used for seed light distribution as shown in Fig. 1, and one is located at OLT as booster amplifier with gain of 10-20 dB, the other is located at RN as in-line amplifier with gain of 20-30 dB. Obviously, these two EDFAs are not located at ONU, and thus the introduced cost can be shared by all the connected ONUs. Moreover, with the help of seed light, no wavelength-specific lasers, complicated wavelength management module and cooling module are required at each ONU, resulting in lower cost and power consumption. For the upstream signal at ONU, the optical loss from ONU to OLT is about 21 dB (2 dB for two optical circulators, 10 dB for 50km 7-core fiber, 6 dB for Fan-in and Fan-out devices, and 3 dB for 2x2 optical coupler), and the same optical loss is achieved for the downstream signal at OLT. Therefore, only one stage optical amplifier is required for both upstream and downstream transmission at ONU and OLT respectively. Low-cost SOA with the gain of 10-20 dB instead of EDFA could be used at ONU and OLT.

Moreover, due to the use of remotely seeded scheme and wavelength reuse technology in our scheme, the polarization of the remotely supplied optical carrier is random at ONU. Obviously, automatic polarization control is useful but expensive. For low-cost consideration, polarization-independent silicon-based bidirectional IQ modulator has been reported and is proved to a good solution for solving this problem [34]. Moreover, a silicon photonic integrated circuit (PIC) based IQ modulator which is copackaged with a conventional SOA is also realized [35] and the PIC has the same functionality a Faraday mirror and thus enables a self-polarization stabilization scheme [36,37]. Furthermore, photonic wire-bonding technology [38] could further greatly reduce cost and footprint of co-packaging of SOA and PIC.

More importantly, fiber dispersion can be digitally compensated and SSBI could be effectively mitigated due to the use of KK-based receiver. Because the same outer core is used for upstream and downstream transmission, the RB noise would occur as shown in the insets (c) and (d) of Fig. 1 and degrade the system performance severely. Fortunately, due to the fact that different sidebands are assigned for upstream and downstream signals, the RB noise can be easily removed by using an optical filter.

2.2 SSBI in homodyne coherent detection system using SSB-Nyquist-16QAM modulation

The operating principle of SSB-Nyquist-16QAM modulation is shown in Fig. 2(a). The 16QAM baseband signal is first filtered by a digital root-raised-cosine (RRC) filter for Nyquist pulse shaping, and the roll-off factor of this RRC filter is 0.1. The generated complex signal is then up-converted to RF frequency of $f_{0}$ to produce a real-valued double sideband (DSB) signal. Subsequently, by using a SSB filter based on digital Hilbert transform, the negative frequency sideband of this DSB signal is filtered out and the SSB-Nyquist-16QAM signal is generated. After high speed digital-to-analog converter (DAC) and optical IQ modulator, the optical SSB-Nyquist-16QAM signal is obtained. As mentioned in the section 2.1, the transmitted optical SSB-Nyquist-16QAM signal $E_{s} (t)$ is combined with a LO $E_{c} (t)$ driven from seed light before PD to demodulate the signal in receiver, and the combined optical signal with optical spectrum shown in the inset of Fig. 2(b) can be expressed as:

Fig. 2 The principle of SSB-Nyquist-16QAM modulation (a), and the generation of SSBI in homodyne coherent detection system (b).

Download Full Size | PDF

E_{x} (t) = E_{c} (t) + j \cdot E_{s} (t) .

Thus the received RF signal after PD can be written as:

\begin{matrix} r (t) \propto {| E_{x} (t) |}^{2} \\ = {| E_{c} (t) + j E_{s} (t) |}^{2} \\ = {| E_{c} (t) |}^{2} + 2 Re {j E_{c}^{*} (t) \cdot E_{s} (t)} + {| E_{s} (t) |}^{2}, \end{matrix}

where

Re {x}

represents obtaining the real part of

x

. It could be clearly observed that the first term of Eq. (2) is the direct current (DC) component which can be removed by a DC block, and the second term is the desired carrier-signal beating products. The third term is the unwanted SSBI term, whose spectrum is overlapped with the signal spectrum as shown in the inset of Fig. 2(b). Therefore, the transmitted signal is disturbed severely by this SSBI.

2.3 SSBI elimination scheme in KK-based receiver

The KK-based receiver allows the phase to be reconstructed digitally from the detected intensity, thus the complex field of the SSB-Nyquist-16QAM signal before PD can be recovered. According to Eq. (1), the received signal before PD can be written as:

E_{x} (t) = E_{c} (t) + j \cdot E_{s} (t) = (A + E_{s} (t) \cdot e^{j (Δ φ + \frac{π}{2})}) \cdot e^{j ω t},

where

A

is the amplitude of the optical carrier,

Δ φ

is the phase difference between optical carrier and the transmitted optical signal, and

ω

is the angular frequency of optical carrier. If

A

is large enough,

E_{x} (t) \cdot e^{- j ω t} = A + E_{s} (t) \cdot e^{j (Δ φ + \frac{π}{2})} = | A + E_{s} (t) \cdot e^{j (Δ φ + \frac{π}{2})} | \cdot e^{j ϕ (t)} = \sqrt{r (t)} \cdot e^{j ϕ (t)}

will satisfy the condition of minimum phase, and we can further obtain:

\log (E_{x} (t) \cdot e^{- j ω t}) = l o g (\sqrt{r (t)}) + j \cdot ϕ (t) .

Obviously, the real and imaginary parts in Eq. (4) can be related through the famous Kramers-Kronig relations [39] as

\begin{array}{l} l o g (\sqrt{r (t)}) = - p . v . \int_{- \infty}^{\infty} \frac{ϕ (t^{'}) d t^{'}}{π (t - t^{'})} \\ ϕ (t) = p . v . \int_{- \infty}^{\infty} \frac{l o g (\sqrt{r (t^{'})}) d t^{'}}{π (t - t^{'})}, \end{array}

where

p . v .

stands for principle value, therefore the transmitted signal can be reconstructed as:

\begin{array}{l} E_{s} (t) \cdot e^{j (Δ φ + \frac{π}{2})} = \sqrt{r (t)} \cdot e^{j ϕ (t)} - A \\ ϕ (t) = \frac{1}{2 π} p . v . \int_{- \infty}^{\infty} \frac{l o g (r (t)) d t^{'}}{t - t^{'}} . \end{array}

The phase noise $Δ φ + π / 2$ in Eq. (6) can be easily estimated by carrier phase recovery algorithm. It is noted that the complex field can be linearly reconstructed by using KK algorithm, thus the impact of SSBI could be neglected. Moreover, the fiber dispersion can be compensated conveniently in digital domain thanks to the obtained complex field [40].

3. Simulation setup and discussions

The cost is the most important factor for PON system. To analyze the cost of the proposed PON, a simulation setup as Fig. 3 is built up based on VPI Transmission Maker 9.0, and four types of receivers are comparatively tested: traditional coherent receiver, heterodyne receiver using one PD, KK-based receiver with LO at ONU (KK-based Receiver #1), and KK-based receiver with remote LO at OLT (KK-based Receiver #2). In our simulation, the same optical transmitter composed of a CW laser with the optical power of 10 dBm and an optical IQ modulator is used. The wavelength of CW laser is $λ_{1} = 1550 nm$ and the linewidth is 1 MHz. 80 km SSMF is adopted for transmission, the loss of fiber is 0.2 dB/km and the dispersion value is 17 ps/nm/km. In all tests, the data rate and modulation format is fixed at 100 Gbps and 16QAM or SSB-Nyquist-16QAM respectively. The power of LO, whose maximum output power is 14 dBm, is optimized for the best transmission performance in each test. In addition, the frequency difference between the CW laser and LO is set to 40 GHz in the case of heterodyne receiver using one PD. For all the receivers, the used PD has a model of PIN and responsivity of 1 A/W. Thermal noise and shot noise are also considered, and the value of thermal noise is set to $10 \times 10^{- 12} A / H z^{\frac{1}{2}}$ . Finally, the received signal after PD is sent to Matlab for demodulation and performance evaluation. It should be noted that the same digital receiver algorithm is used in these four types of receiver except that additional KK algorithm is adopted in KK-based receiver.

Fig. 3 The simulation setup of 100 Gbps 16QAM signal transmission using four types of receivers.

Download Full Size | PDF

The simulated bit error rate (BER) performances of 100 Gbps 16QAM signal in terms of the received optical signal power before PD after 80 km SSMF propagation in these four cases are plotted in Fig. 4(a). The measured receiver sensitivity at the forward error correction (FEC) limit (BER of 3.8x10⁻³) of 100 Gbps 16QAM signal is −36 dBm, −32.6 dBm, −32.7 dBm and −31.4 dBm for traditional coherent receiver, heterodyne receiver using one PD, KK-based receiver with LO at ONU, and KK-based receiver with remote LO at OLT respectively. Obviously, 3.3 dB power penalty is observed between the traditional coherent receiver and the KK-based receiver with LO at ONU. The heterodyne receiver using one PD and the KK-based receiver with LO at ONU have the similar received sensitivity. 1.3 dB power penalty is observed when the LO is launched from OLT in KK-based receiver, and this could be attributed to the increased ASE noise in the LO. The electrical spectra of the received signals of these four types of receivers at the received optical power of −29 dBm are shown in the insets of Fig. 3 respectively, and the required number and bandwidth of electrical and optical components are listed in Table. 1.

Fig. 4 The measured BER performance versus received optical signal power using four types of receivers (a). The measured BER performance in terms of linewidth of the used CW laser based on heterodyne receiver using one PD, and KK receiver with LO at ONU (b).

Download Full Size | PDF

Table 1. The Required Number and Bandwidth of Electrical and Optical Components

View Table

Taking into account the achieved receiver sensitivity, the system cost and the required power consumption, the KK-based receiver with LO at ONU is the good choice. Although higher oversampling factor is needed in KK-based receiver for nonlinear calculations as shown in Eq. (4). The used oversampling factor in our simulation is 6 which is higher than the used oversampling factor in heterodyne coherent receiver (usually 3-4). However, it is reported in [41] that the issue of higher oversampling rate can be solved, and no additional digital oversampling operation is required. Moreover, if LO is located at OLT, no lasers are required at ONU, and the signal and LO originate from the same laser. Therefore, the carrier frequency offset is no longer needed to be tracked, and the inherent characteristic for phase noise cancellation can also reduce the complexity of DSP and power consumption. Therefore, by considering the system cost and the required power consumption, the KK-based receiver with remote LO at OLT is obviously a good choice, compared with the case of LO at the ONU, and 1.3 dB power penalty is acceptable.

Obviously, low-cost, small-footprint lasers with relatively large linewidth are preferred over the costly narrow linewidth external cavity lasers. Figure 4(b) displays the measured BER performance in terms of linewidth of the used CW laser based on heterodyne receiver using one PD, and KK-based receiver with LO at ONU. The analysis shows little impact on 100 Gbps 16-QAM signals when the linewidth is increased from 100 kHz to 2 MHz. Even when the linewidth is increased to 10 MHz, the measured BER performance below the FEC limit could still be achieved after 80 km SSMF. Moreover, single-side band modulation based KK receiver has higher tolerance to phase noise compared to the heterodyne coherent detection which agrees with the description in [42]. This will allow the use of cheap laser sources for the proposed PON system.

The SSBI mitigation and fiber dispersion compensation based on the KK-based receiver is also simulated and analyzed. Figure 5(a) shows the measured error vector magnitude (EVM) performance of the received 100 Gbps SSB-Nyquist-16QAM signal in terms of CSPR for 80 km SSMF transmission, and in this test GAP is fixed to 1.75 GHz and KK algorithm is used for demodulation. Here GAP indicates the frequency band between the optical carrier and the transmitted optical SSB-Nyquist-16QAM signal as shown in Fig. 5(b). Two types of PD with model of PIN and responsivity of 1 A/W are tested for comparison. In the first type of PD, thermal noise and shot noise are both considered, and the value of thermal noise is set to $10 \times 10^{- 12} A / H z^{\frac{1}{2}}$ . Thermal noise and shot noise of the second type of PD are not taken into account. Moreover, the optical signal at the input of optical coupler with different OSNR_BW is evaluated in these two tests. It should be noted that the CSPR is varied by adjusting the power of LO and signal while keeping the input optical power at PD constant in all tests. It could be observed in Fig. 5(a) that when CSPR is decreased and lower than 10 dB, the measured EVM performance will be severely worsen in both PDs, and this could be attributed to the fact that the condition of minimum phase is no longer satisfied for the KK receiver and SSBI occurs. When CSPR is larger than 10 dB, the measured EVM performance changes little as CSPR increases in the PD without noise. However, in the PD with noise, an increased CSPR would result in a degradation of SNR in the receiver due to the PD noise and slightly decrease the EVM performance. Therefore, the optimum CSPR value is 10 dB in our test.

Fig. 5 The measured EVM performance of the received 100 Gbps SSB-Nyquist-16QAM signals in terms of CSPR for 80 km SSMF with and without PD noise (a). The electrical spectrum of received signal at CSPR of 10 dB when GAP is 1.75 GHz (b). The measured EVM performance as a function of GAP with and without KK demodulation algorithm in OBTB case (c). The measured BER performance in terms of fiber dispersion value by using different demodulation algorithms when GAP is set to 1.75 GHz (d).

Download Full Size | PDF

Figure 5(c) shows the measured EVM performance of the received 100 Gbps SSB-Nyquist-16QAM signal as a function of GAP when the OSNR_BW at the input of optical coupler is fixed to 22.5 dB for optical back-to-back (OBTB) transmission, and the CSPR is set to 10 dB in the test. Obviously, the transmission performance of the proposed system is severely degraded by the SSBI without the help of the KK receiver, and the transmission performance will be improved as GAP increases. It should be noted that the electrical spectrum of the generated SSBI noise has a typical triangle shape [12], thus a minimum disruption by SSBI will be achieved when the GAP is close to the occupied bandwidth of the transmitted signal (27.5 GHz). However, when the KK receiver is adopted, SSBI has negligible impact on the transmission performance even if GAP is very small, and this illustrates that the KK-based receiver can effectively eliminate the SSBI.

As mentioned above, it is easy to compensate the accumulated chromatic dispersion (CD) induced by fiber link because the complex field of the transmitted signal can be reconstructed by using the KK receiver. Figure 5(d) shows the measured BER performance of the received 100 Gbps SSB-Nyquist-16QAM signal after fiber transmission with different dispersion values, the GAP is set to 1.75 GHz and the OSNR_BW at the input of optical coupler is fixed to 22.5 dB in the test. Without the help of the KK receiver, the chromatic dispersion compensation (CDC) algorithm cannot be used due to the loss of the optical phase in square-law detection. Therefore, the received 100 Gbps SSB-Nyquist-16QAM signal is severely affected by both SSBI and the accumulated CD, and no BER performance below FEC limit is achieved even if the accumulated CD value is close to 0 ps/nm. On the other hand, CDC could be used thanks to the KK receiver, and the BER performance with and without CDC algorithm are both measured and plotted in Fig. 5(d) to evaluate the impact of CD. The results show that if CDC algorithm isn’t used, the tolerance of CD is just 400 ps/nm, and this value is increased to beyond 3800 ps/nm which corresponds to 220 km SSMF with dispersion value of 17 ps/nm/km when CDC algorithm is adopted.

4. Experimental setup and discussions

The experimental setup of the proposed bidirectional LR-PON using the KK receiver and 7-core fiber is shown in Fig. 6(a). As mentioned above, the left and right sideband of the optical carrier is allocated for the upstream (US) and downstream (DS) transmission respectively in our experiment, and an optical bandwidth-tunable filter (TBF) is used to remove the RB noise for both upstream and downstream signals. At OLT, an arbitrary waveform generator (AWG, Keysight M8195A) with a sample rate of 60 GSa/s and 3-dB bandwidth of 20 GHz is used to generate the SSB-Nyquist-16QAM signals. As shown in Fig. 6(c), a data stream with a pseudo-random bit sequence (PRBS) length of 2¹⁵-1 is mapped into 16QAM symbols with symbol rate $f_{s}$ of 15 Gbaud, and the generated 16QAM symbols is up-sampled by a factor of 4. Subsequently, Nyquist pulse shaping is operated by using a RRC filter with a roll-off factor $β$ of 0.1. The filtered signals are up-converted to 9.75 GHz or 12 GHz ( $f_{s} \times 0.65$ or $f_{s} \times 0.8$ ) radio frequency to produce a DSB Nyquist-16QAM signal. A single sideband filter based on Hilbert transform is then used to generate the desired SSB-Nyquist-16QAM signal with the data rate of 60 Gbps ( $f_{s} \times \log_{2} 16 = 60 Gbps$ ). A 100 kHz-linewidth CW external cavity laser (ECL, λ = 1549.992 nm) with the output power of 13.9 dBm is split into three parts by two OCs. One part of this light is used as the optical source for downstream, and the other two parts are served as seed light for upstream and LO for the KK receiver in OLT respectively. The generated SSB-Nyquist-16QAM signal is then modulated onto the optical carrier for downstream transmission by an optical IQ modulator (Fujitsu FTM7962EP) with a half-wave voltage of 9 V at the bias port. After an EDFA, the output optical power is amplified to 6 dBm. Afterwards, the transmitted downstream signal and the seed light for upstream are coupled into 7-core MCF by the self-developed fan-in/fan-out devices [43]. The used 7-core MCF in our experiment has low crosstalk of −45 dB/100km between adjacent cores, and the cross section view of this MCF is indicated in the inset of Fig. 6(a). The middle core is used for seed light distribution, and one outer core is used for bidirectional signal transmission. After 50 km 7-core fiber propagation and fan-out device, the optical downstream signal is filtered by a TBF (Yenista Optics XTM50-standard) with a filter edge gradient of 500 dB/nm to remove the RB noise induced by the optical upstream signal. Meanwhile, the seed light from the middle core is amplified by an EDFA and then split into two parts. One part of seed light is used as the LO for the KK receiver in ONU and the other part is reused for upstream transmission. The LO and the filtered optical downstream signal is combined and detected by a single PD. As for the problem of polarization aligning for optical downstream signal and LO, it could be effectively solved by the method reported in [44], where cost-efficient optical polarization controllers driven by low-speed polarization tracking circuitry for both signal and LO path are used. The total optical power before PD is about 1 dBm in our experiment, and the detected RF signal is captured by a digital sampling oscilloscope (DSO, Tektronix DSA72504D) with a sampling rate of 100 GSa/s and 3-dB bandwidth of 25 GHz. Offline signal demodulation as shown in Fig. 6(c) is then performed by a DSP-based receiver consisting of KK reconstruction algorithm, CD compensation algorithm, down-conversion, RRC filter, constant modulus algorithm (CMA) for channel estimation, carrier-phase recovery, decision-directed least-mean square (DD-LMS) algorithm for channel equalization and BER counting. In our experiment, 3.4 × 10⁴ bits are calculated for BER counting.

Fig. 6 Experimental setup of the proposed bidirectional LR-PON (a), the optical spectra of the transmitted signals at corresponding points of OLT and ONU (b), and the offline DSP algorithms of the SSB-Nyquist-16QAM modulation and demodulation (c).

Download Full Size | PDF

For upstream transmission, another 60 Gbps SSB-Nyquist-16QAM signal with the same modulation scheme as the downstream signal is produced in the same AWG, and the upstream signal is produced by another two ports of this AWG to realize full-duplex transmission. After two electrical amplifiers (EA) with 27 dB gain, the upstream signal is used to modulate the seed light in another optical IQ modulator. The optical upstream signal is amplified by another EDFA and then sent to the same outer core with the help of CIR and fan-out device. After the same MCF propagation, the optical upstream signal is filtered by another TBF and then be combined with the LO in OLT by using an OC. The combined optical signals are detected by a PD and captured by the same DSO, and the same digital signal process as used in downstream transmission is performed. The optical spectra of the received upstream signal in OLT and the downstream signal in ONU are shown in Fig. 6(b).

Figure 7(a) shows the measured EVM performance of the received downstream signal in ONU in terms of CSPR with and without 50 km MCF transmission. In this test, only downstream transmission rather than bidirectional transmission is considered, thus no RB noise from the upstream signal is existed and TBF is not used. Different frequency gap of 1.5 GHz and 3.75 GHz between optical carrier and the desired downstream signal is adopted for evaluation. Here the value of GAP is calculated as $f_{s} \times 0.65 - f_{s} \times (1 + β) / 2 = 1.5 G H z$ and $f_{s} \times 0.8 - f_{s} \times (1 + β) / 2 = 3.75 G H z$ . The OSNR_BW at the input of optical coupler is fixed to 25 dB in each measurement. The CSPR value is varied by adjusting the gain of EDFA for the light seed. As shown in Fig. 7(a), the measured EVM curves have the same trend in all cases. When CSPR is lower, the EVM performance is limited by the SSBI which is induced by dissatisfaction of minimum phase condition. When CSPR is too higher, an increased CSPR leads to a degradation of SNR at a given optical input power due to the PD noise, resulting in slight degradation of the measured EVM performance. The optimum CSPR value is 18 dB and 17 dB for OBTB and 50 km MCF transmission respectively. The measured EVM performance with GAP of 3.75 GHz is slightly better than that of 1.5 GHz in both OBTB and 50 km MCF cases, and this could be attributed to the imperfect SSB modulation in the optical IQ modulator. Compared to the OBTB case, the EVM performance is degraded after 50 km MCF transmission whatever the GAP is, and this could be explained by the incomplete CD compensation and the cross-talk between different cores in MCF. The electrical spectra of the received downstream signal after PD in ONU are shown in Fig. 7(b). It should be noted that the optimum CSPR value (17 dB) in our experiment is obviously greater than those reported values (usually 11 dB), and this could be attributed to the residual SSBI induced by imperfect optical SSB signal. The imperfect SSB modulation could be caused by the power difference of the electrical IQ signals or the drift of bias point of optical IQ modulator.

Fig. 7 The measured EVM performance of the received downstream signal in ONU in terms of CSPR under different GAP and fiber distances (a), and the electrical spectral of the received RF signals (b).

Download Full Size | PDF

To evaluate the impact of the RB noise, the EVM performance of the received downstream signal in ONU is measured by varying the optical power of the upstream signal which is launched into the MCF in ONU. Figure 8(a) presents the measured EVM performance of the received downstream with and without the TBF in both GAP of 1.5 GHz and 3.75 GHz cases. The CSPR value is fixed at 17 dB in this test. It could be clearly observed that without the help of TBF, the induced RB noise becomes larger as the input optical power of the upstream signal increases, resulting in the EVM performance degradation in both GAP of 1.5 GHz and 3.75 GHz cases. However, the measured EVM performance with GAP of 3.75 GHz is slightly better than that of 1.5 GHz whether or not the TBF is used, and this could be attributed to the imperfect SSB modulation in the optical IQ modulator. When the TBF is adopted, the RB noise can be effectively eliminated and the measured EVM performance is improved observably. Moreover, the EVM performance difference between the case of 1.5 GHz and 3.75 GHz GAP becomes bigger as the input optical power of the upstream signal increases. This could be explained by the fact that it is easier to filter out the RB noise when the GAP is bigger. To further analyze the effect of RB noise elimination by using the TBF, the optical spectra of the received downstream signal before and after the TBF are plotted in Figs. 8(b)-8(e) with different input optical power of the upstream signal. Figures 8(f) and 8(g) is the shape of the used TBF in the case of 1.5 GHz and 3.75 GHz GAP respectively. The induced RB noise increases evidently when the optical power of the upstream signal is increased from 2.5 dBm to 10.5 dBm as shown in Figs. 8(b) and 8(c). When the TBF is used, the RB noise could be effectivley filtered out, but there is still some RB noise residual when the optical power of the upstream singal is 10.5 dBm, as shown in Fig. 8(d). However, when the GAP is increased to 3.75 GHz, all the RB noise could be completely filtered out in both optical power of 2.5 dBm and 10.5 dBm, as shown in Fig. 8(e). It should be noted that too large GAP will increase the required bandwidth of electrical and optical components, and the requirement for the sharpness of the optical filter will be higher if the frequency gap is too small. Moreover, the frequency drfit of laser will also degrade the transmission performance. Fortunately, the impact of laser frequency drift can be effectively weakened by using frequency stability operation, and this actually is acceptable because only one laser is needed and located at OLT in our scheme. However, a frequency gap of 3 GHz (corresponding to GAP of 1.5 GHz) is enough for bidirectional 60 Gbps SSB-Nyquist-16QAM signals transmission in our experiment, and the used optical filter has an edge gradient of 500 dB/nm.

Fig. 8 The measured EVM performance of the received downstream signal in ONU in terms of the input optical power of upstream signal (a). The optical spectra of the received downstream signal before TBF with upstream signal of different input power when GAP is 1.5 GHz (b) and 3.75 GHz (c). The optical spectra of the received downstream after TBF with upstream signal of different input power when GAP is 1.5 GHz (d) and 3.75 GHz (e). The shape of the used TBF when GAP is 1.5 GHz (f) and 3.75 GHz (g).

Download Full Size | PDF

Figure 9 shows the measured BER performance of the received downstream signal in ONU in terms of OSNR_BW with and without upstream transmission after OBTB and 50 km MCF transmission. It is not easy to fix at the optimum CSPR value when varying the received optical power in the experiment, thus OSNR_BW is used for performance evaluation. It should be noted that, in this test the ratio of the combined optical power (including LO and signal) at the output of optical coupler to noise power within the signal bandwidth is used for OSNR_BW measurement. The OSNR_BW is varied by adding ASE noise into the optical downstream signal, and the CSPR value is fixed at 17 dB in each measurement. Moreover, the GAP is set to 1.5 GHz to improve the system SE, and the input optical power of downstream and upstream signals launched into the MCF are both set to 6 dBm to take both the impact of the RB noise and the fiber link loss into consideration. Obviously, the required OSNR_BW after 50 km MCF transmission at the FEC limit is decreased from 40.7 dB to 36 dB thanks to the effective RB noise elimination by using TBF. To verify the proposed RB noise elimination scheme, the BER performance of the received downstream signal without the upstream transmission is also tested, and negligible OSNR_BW penalty is observed compared with the measured EVM performance with TBF in bidirectional transmission case. Moreover, the required OSNR_BW without the upstream transmission at the FEC limit is 34.6 dB and 36 dB for OBTB and 50 km MCF transmission respectively, and this 1.4 dB OSNR_BW penalty could be attributed to the incomplete CD compensation and the cross-talk between different cores in MCF. The demodulated constellations of the received downstream signals after PD in ONU are presented in the insets of Fig. 9.

Fig. 9 The measured BER performance of the received downstream signal in ONU in terms of OSNR_BW with and without upstream transmission after OBTB and 50 km MCF transmission.

Download Full Size | PDF

The transmission performance of the upstream signal is also evaluated in our experiment. The input optical power of upstream signals launched into the MCF is 6 dBm, and the GAP is set to 1.5 GHz in our test. Figure 10 illustrates the measured EVM performance of the received upstream signal in OLT without downstream transmission in terms of CSPR after OBTB and 50 km MCF transmission, and the OSNR_BW at the input of optical coupler is fixed to 25 dB in each measurement. Similarly as the downlink, the measured optimum CSPR value is 18 dB and 17 dB for OBTB and 50 km MCF transmission respectively, and the EVM performance different between the case of OBTB and MCF transmission can also be explained by the incomplete CD compensation and the cross talk between different cores in MCF.

Fig. 10 The measured EVM performance of the received upstream signal in OLT without downstream transmission in terms of CSPR for OBTB and 50 km MCF transmission.

Download Full Size | PDF

Figure 11 shows the measured BER performance of the received upstream signal in OLT in terms of OSNR_BW with and without downstream transmission after OBTB and 50 km MCF transmission, and the input optical power of downstream and upstream signals launched into the MCF are both set to 6 dBm. Similarly, the ratio of the combined optical power (including LO and signal) at the output of optical coupler to noise power within the signal bandwidth is used for OSNR_BW measurement in this test. The CSPR value is also fixed at 17 dB. Obviously, the required OSNR_BW after 50 km MCF transmission at the FEC limit is decreased from 41.5 dB to 37 dB with the help of TBF. The BER performance of the received upstream signal without downstream transmission is also tested, and the results show that negligible OSNR_BW penalty is observed compared with the measured EVM performance with the TBF in bidirectional transmission case. This proves that the RB noise is effectively eliminated due to the use of TBF. Furthermore, the required OSNR_BW without downstream transmission at the FEC limit is 34.7 dB and 37 dB for OBTB and 50 km MCF transmission respectively. The demodulated constellation of the received downstream signal after PD in ONU are presented in Fig. 11. The slight transmssion perforamnce difference between the upstream and downstream signals in Figs. 9 and 11 could be attributed to the different performance of optical and electrical components used in uplink and downlink.

Fig. 11 The measured BER performance of the received upstream signal in OLT in terms of, $O S N R_{B W}$ with and without downstream transmission, after OBTB and 50 km MCF transmission.

Download Full Size | PDF

5. Summary

We have experimentally demonstrated a bidirectional LR-PON based on the KK receiver and 50 km 7-core fiber. In our experiment, the middle core is used to deliver the seed light to ONU for colorless upstream transmission, and the upstream and downstream signals are transmitted simultaneously over the same outer core for each ONU. By using the SSB-Nyquist-16QAM modulation technique, the upstream and downstream signals can be allocated to the left and right sideband of optical carrier. By this way, not only the information spectral densities can be increased but also the RB noise can be easily removed by a simple optical filter in the receiver. Moreover, due to the use of a single PD with KK detection in both OLT and ONU, the impact of SSBI could be eliminated at the extreme, and chromatic dispersion induced by fiber link can also be compensated digitally. By this means, bidirectional transmission of 60 Gbps SSB-Nyquist-16QAM signals over 50 km 7-core fiber are successfully achieved, and the frequency gap between the upstream and downstream signals is only 3 GHz (corresponding to GAP of 1.5 GHz). These results show that the proposed scheme has potential application in future high-capacity, spectrally-efficiency and cost-effective optical access networks.

Funding

National “863” Program of China (2015AA016904); National Nature Science Foundation of China (NSFC) (61675083, 61505061); Fundamental Research Funds for the Central Universities HUST (2017KFKJXX010, 2017KFXKJC002); Key project of R&D Program of Hubei Province (2017AAA046).

References

1. K. Curtis, “IEEE 100G-EPON,” in Optical Fiber Communication Conference (Optical Society of America, 2016), paper Th1I.6.

2. V. Houtsma, D. V. Veen, and E. Harstead, “Recent progress on standardization of next-generation 25, 50, and 100G EPON,” J. Lightwave Technol. 35(6), 1228–1234 (2017). [CrossRef]

3. S. Li, Z. Ye, N. Cheng, and X. Liu, “Demonstration of a real-time 25-Gb/s TDM-PON system with 25-Gb/s downstream based on optical duobinary and 10-Gb/s burst-mode upstream based on NRZ,” in Optical Fiber Communication Conference (Optical Society of America, 2016), paper Th1I.3. [CrossRef]

4. J. Prat, I. N. Cano, M. Presi, I. Tomkos, D. Klonidis, G. V. Llosera, R. Brenot, R. Pous, G. Papastergiou, A. Rafel, and E. Ciaramella, “Technologies for cost-effective udWDM-PONs,” J. Lightwave Technol. 34(2), 783–791 (2016). [CrossRef]

5. N. Yoshimoto, J. Kani, S. Kim, N. Liyama, and J. Terda, “DSP-based optical access approaches for enhancing NG-PON2 systems,” IEEE Commun. Mag. 51(3), 58–64 (2013). [CrossRef]

6. D. Lavery, R. Maher, D. Millar, B. Thomsen, P. Bayvel, and S. Savory, “Demonstration of 10Gbit/s colorless coherent PON incorporating tunable DS-DBR lasers and low-complexity parallel DSP,” in Optical Fiber Communication Conference (Optical Society of America, 2012), paper PDP5B.10. [CrossRef]

7. G. Pandey and A. Goel, “100 Gbps long reach coherent PON downstream transmission using dual polarization-QPSK with digital signal processing,” Opt. Quantum Electron. 47(11), 3445–3453 (2015). [CrossRef]

8. H. C. Chien, Z. Jia, J. Zhang, Z. Dong, and J. Yu, “Optical independent-sideband modulation for bandwidth-economic coherent transmission,” Opt. Express 22(8), 9465–9470 (2014). [CrossRef] [PubMed]

9. S. Wolf, M. Lauermann, G. Ronniger, P. C. Schindler, R. Palmer, S. Koeber, W. Bogaerts, J. Leuthold, W. Freude, and C. Koos, “DAC-less and amplifier-less generation and transmission of 16QAM signals using a sub-volt silicon photonic modulator,” in Proceedings of European Conference on Optical Communication (Institute of Electrical and Electronics Engineers, 2014), pp. 1–3. [CrossRef]

10. P. C. Schindler, A. Agmon, S. Wolf, R. Bonk, L. Meder, M. Meltsin, A. Ludwig, R. Schmogrow, M. Dreschmann, J. Meyer, J. Becker, M. Nazarathy, S. B. Ezra, T. Pfeiffer, W. Freude, J. Leuthold, and C. Koos, “Ultra-dense, single-wavelength DFT-spread OFDMA PON with laserless 1.2 Gb/s ONU ready for silicon photonics integration,” J. Lightwave Technol. 33(8), 1650–1659 (2015). [CrossRef]

11. D. Z. Hsu, C. C. Wei, H. Y. Chen, W. Y. Li, and J. Chen, “Cost-effective 33-Gbps intensity modulation direct detection multi-band OFDM LR-PON system employing a 10-GHz-based transceiver,” Opt. Express 19(18), 17546–17556 (2011). [CrossRef] [PubMed]

12. W. Peng, B. Zhang, K. Feng, X. Wu, A. E. Willner, and S. Chi, “Spectrally efficient direct-detected OFDM transmission incorporating a tunable frequency gap and an iterative detection techniques,” J. Lightwave Technol. 27(24), 5723–5735 (2009). [CrossRef]

13. D. Z. Hsu, C. C. Wei, H. Y. Chen, J. Chen, M. C. Yuang, S. H. Lin, and W. Y. Li, “21 Gb/s after 100 km OFDM long-reach PON transmission using a cost-effective electro-absorption modulator,” Opt. Express 18(26), 27758–27763 (2010). [CrossRef] [PubMed]

14. H. Chen, C. Wei, I. Lu, H. Chu, Y. Chen, and J. Chen, “High-capacity and high-loss-budget OFDM long-reach PON without an optical amplifier,” J. Opt. Commun. Netw. 7(1), A59–A65 (2015). [CrossRef]

15. B. J. C. Schmidt, A. J. Lowery, and J. Armstrong, “Experimental demonstrations of electronic dispersion compensation for long-haul transmission using direct-detection optical OFDM,” J. Lightwave Technol. 26(1), 196–203 (2008). [CrossRef]

16. W. Peng, X. Wu, V. R. Arbab, K. Feng, B. Shamee, L. C. Christen, J. Yang, A. E. Willner, and S. Chi, “Theoretical and experimental investigations of direct-detected RF-tone-assisted optical OFDM systems,” J. Lightwave Technol. 27(10), 1332–1339 (2009). [CrossRef]

17. S. Randel, D. Pilori, S. Chandrasekhar, G. Raybon, and P. Winzer, “100-Gb/s discrete-multitone transmission over 80-km SSMF using single-sideband modulation with novel interference-cancellation scheme,” in Proceedings of European Conference on Optical Communication (Institute of Electrical and Electronics Engineers, 2015), pp. 1–3. [CrossRef]

18. Z. Li, M. S. Erkilinc, R. Maher, L. Galdino, K. Shi, B. C. Thomsen, P. Bayvel, and R. I. Killey, “Two-stage linearization filter for direct-detection subcarrier modulation,” IEEE Photonic. Tech. Lett. 28(24), 2838–2841 (2016). [CrossRef]

19. K. Zou, Y. Zhu, F. Zhang, and Z. Chen, “Spectrally efficient terabit optical transmission with Nyquist 64-QAM half-cycle subcarrier modulation and direct detection,” Opt. Lett. 41(12), 2767–2770 (2016). [CrossRef] [PubMed]

20. A. Mecozzi, C. Antonelli, and M. Shtaif, “Kramers-Kronig coherent receiver,” Optica 3(11), 1220–1227 (2016). [CrossRef]

21. C. Antonelli, M. Shtaif, and A. Mecozzi, “Kramers-Kronig PAM transceiver,” in Optical Fiber Communication Conference (Optical Society of America, 2017), paper Tu3A.5. [CrossRef]

22. Y. Zhu, K. Zou, X. Ruan, and F. Zhang, “Single carrier 400G transmission with single-ended heterodyne detection,” IEEE Photonic. Tech. Lett. 29(21), 1788–1791 (2017). [CrossRef]

23. L. Shu, J. Li, Z. Wan, F. Gao, S. Fu, X. Li, Q. Yang, and K. Xu, “Single-lane 112-Gbit/s SSB-PAM4 transmission with dual-drive MZM and Kramers-Kronig detection over 80-km SSMF,” IEEE Photonics J. 9(6), 1–9 (2017). [CrossRef]

24. T. M. Hoang, M. Y. S. Sowailem, Q. Zhuge, Z. Xing, M. M. Osman, E. E. Fiky, S. Fan, M. Xiang, and D. V. Plant, “Single wavelength 480 Gb/s direct detection over 80km SSMF enabled by Stokes vector Kramers Kronig transceiver,” Opt. Express 25(26), 33534–33542 (2017). [CrossRef]

25. Z. Li, M. S. Erkılınç, K. Shi, E. Sillekens, L. Galdino, B. C. Thomsen, P. Bayvel, and R. I. Killey, “SSBI mitigation and the Kramers-Kronig scheme in single-sideband direct-detection transmission with receiver-based electronic dispersion compensation,” J. Lightwave Technol. 35(10), 1887–1893 (2017). [CrossRef]

26. Z. Li, M. Erkılınç, K. Shi, E. Sillekens, L. Galdino, B. C. Thomsen, P. Bayvel, and R. I. Killey, “Joint optimisation of resampling rate and carrier-to-signal power ratio in direct-detection Kramers-Kronig receivers,” in Proceedings of European Conference on Optical Communication (Institute of Electrical and Electronics Engineers, 2017), pp. 1–3. [CrossRef]

27. J. Prat, M. Omella, and V. Polo, “Wavelength shifting for colorless ONUs in single-fiber WDM-PONs,” in Optical Fiber Communication Conference (Optical Society of America, 2007), paper OTuG6. [CrossRef]

28. S. Amiralizadeh, A. T. Nguyen, C. S. Park, and L. A. Rusch, “Single-fiber lightwave centralized WDM-OFDMA-PON with colorless optical network units,” J. Opt. Commun. Netw. 8(4), 196–205 (2016). [CrossRef]

29. C. H. Wang, C. W. Chow, C. H. Yeh, C. L. Wu, S. Chi, and C. Lin, “Rayleigh noise mitigation using single-sideband modulation generated by a dual-parallel MZM for carrier distributed PON,” IEEE Photonic. Tech. L. 22(11), 820–822 (2010). [CrossRef]

30. B. Zhu, T. F. Taunay, M. F. Yan, J. M. Fini, M. Fishteyn, E. M. Monberg, and F. V. Dimarcello, “Seven-core multicore fiber transmissions for passive optical network,” Opt. Express 18(11), 11117–11122 (2010). [CrossRef] [PubMed]

31. Z. Feng, B. Li, M. Tang, L. Gan, R. Wang, R. Lin, Z. Xu, S. Fu, L. Deng, W. Tong, S. Long, L. Zhang, H. Zhou, R. Zhang, S. Liu, and P. P. Shum, “Multicore-fiber-enabled WSDM optical access network with centralized carrier delivery and RSOA-based adaptive modulation,” IEEE Photonics J. 7(4), 1–9 (2015). [CrossRef]

32. C. Xiong, M. Tang, C. Ke, Z. Feng, Q. Wu, L. Xu, S. Fu, W. Tong, P. P. Shum, and D. Liu, “Experimental demonstration of ultra-dense WDM-PON with seven-core MCF-enabled self-homodyne coherent detection,” IEEE Photonics J. 9(2), 1–7 (2017). [CrossRef]

33. G. W. Lu, M. Nakamura, Y. Kamio, and T. Miyazaki, “40-Gb/s QPSK and 20-Gb/s PSK with inserted pilot symbols using self-homodyne detection,” Opt. Express 15(12), 7660–7666 (2007). [CrossRef] [PubMed]

34. P. C. Schindler, M. Lauermann, S. Wolf, D. Korn, R. Palmer, S. Koeber, W. Heni, A. Ludwig, R. Schmogrow, D. L. Elder, L. R. Dalton, W. Bogaerts, H. Yu, W. Freude, J. Leuthold, and C. Koos, “Ultra-short silicon-organic hybrid (SOH) modulator for bidirectional polarization-independent operation,” in Proceedings of European Conference on Optical Communication (Institute of Electrical and Electronics Engineers, 2014), pp. 1–3. [CrossRef]

35. A. Agmon, M. Nazarathy, D. M. Marom, S. B. Ezra, A. Tolmachev, R. Killey, P. Bayvel, L. Meder, M. Hübner, W. Meredith, G. Vickers, P. C. Schindler, R. Schmogrow, D. Hillerkuss, W. Freude, C. Koos, and J. Leuthold, “OFDM/WDM PON with laserless, colorless 1 Gb/s ONUs based on Si-PIC and slow IC,” J. Opt. Commun. Netw. 6(3), 225–237 (2014). [CrossRef]

36. K. Y. Cho, K. Tanaka, T. Sano, S. P. Jung, J. H. Chang, Y. Takushima, A. Agata, Y. Horiuchi, M. Suzuki, and Y. C. Chung, “Long-reach coherent WDM PON employing self-polarization-stabilization technique,” J. Lightwave Technol. 29(4), 456–462 (2011). [CrossRef]

37. D. Kim, B. G. Kim, and H. Kim, “60-km transmission of 28-Gb/s QPSK upstream signal in RSOA-based WDM PON using SBS suppression technique,” in Optical Fiber Communication Conference (Optical Society of America, 2018), paper W4G.4. [CrossRef]

38. N. Lindenmann, G. Balthasar, D. Hillerkuss, R. Schmogrow, M. Jordan, J. Leuthold, W. Freude, and C. Koos, “Photonic wire bonding: a novel concept for chip-scale interconnects,” Opt. Express 20(16), 17667–17677 (2012). [CrossRef] [PubMed]

39. A. Mecozzi, “A necessary and sufficient condition for minimum phase and implications for phase retrieval,” IEEE Trans. Inf. Theory 13(9), 1–9 (2014).

40. G. Goldfarb and G. Li, “Chromatic dispersion compensation using digital IIR filtering with coherent detection,” IEEE Photonic. Tech. L. 19(13), 969–971 (2007). [CrossRef]

41. T. Bo and H. Kim, “Kramers-Kronig receiver without digital upsampling,” in Optical Fiber Communication Conference (Optical Society of America, 2018), paper W4E.2. [CrossRef]

42. L. G. Guerrero, H. Shams, I. Fatadin, M. Fice, M. Naftaly, A. Seeds, and C. Renaud, “Spectrally efficient SSB signals for W-band links enabled by Kramers-Kronig receiver,” in Optical Fiber Communication Conference (Optical Society of America, 2018), paper Th2A.

43. B. Li, Z. Feng, M. Tang, Z. Xu, S. Fu, Q. Wu, L. Deng, W. Tong, S. Liu, and P. P. Shum, “Experimental demonstration of large capacity WSDM optical access network with multicore fibers and advanced modulation formats,” Opt. Express 23(9), 10997–11006 (2015). [CrossRef] [PubMed]

44. M. Morsy-Osman, M. Sowailem, E. El-Fiky, T. Goodwill, T. Hoang, S. Lessard, and D. V. Plant, “DSP-free ‘coherent-lite’ transceiver for next generation single wavelength optical intra-datacenter interconnects,” Opt. Express 26(7), 8890–8903 (2018). [CrossRef] [PubMed]

Type of Receiver	Number of PD	Number of ADC	Coupler	Bandwidth
Traditional Coherent receiver	4	2	900 Optical Hybrid	B/2
Heterodyne Coherent Receiver	1	1	2 × 2 Optical Coupler	2B
KK-based Receiver	1	1	2 × 2 Optical Coupler	B
B is the baud rate of transmitted signal (B = 25 Gbaud in the simulation)

Bidirectional long-reach PON using Kramers-Kronig-based receiver for Rayleigh Backscattering noise and SSBI interference elimination

Abstract

1. Introduction

2. Principle

2.1 The proposed LR-PON using KK-based receiver and multi-core fiber

2.2 SSBI in homodyne coherent detection system using SSB-Nyquist-16QAM modulation

2.3 SSBI elimination scheme in KK-based receiver

3. Simulation setup and discussions

4. Experimental setup and discussions

5. Summary

Funding

References

Cited By

Figures (11)

Tables (1)

Equations (6)

Optics Express