Investigation of wired and wireless services based on OFDM DSB-RC transmission in the presence of modulation chirp of a DEMZM

Investigation of wired and wireless services based on OFDM DSB-RC transmission in the presence of modulation chirp of a DEMZM Paulo Almeida1,* and Henrique Silva1,2 1Instituto de Telecomunicações, Department of Electrical and Computer Engineering, University of Coimbra, Coimbra 3030-290, Portugal 2[email protected] *[email protected] Abstract: In this paper, we investigate the performance of intensity modulation with optical carrier reduced by biasing a dual-electrode MachZehnder modulator (DEMZM) below its quadrature point, in the presence of controlled modulation chirp. The effects of the modulation chirp and of the bias point of a DEMZM on the received signal are analytically derived for small-signal operation. The interaction of these two effects is assessed in terms of optical signal to noise ratio (OSNR) required for a BER = 10−9, through a comparison between double sideband (DSB) modulation and double sideband - reduced carrier (DSB-RC) modulation, by numerical simulation. We found that the power of the optical carrier has impact on the optimum value of the α chirp parameter and that a positive value of the α chirp parameter can be the optimum value to drive a DEMZM, depending on the central frequency of the orthogonal frequency division multiplexing (OFDM) signal and on the DEMZM bias point. ©2013 Optical Society of America OCIS codes: (060.2330) Fiber optics communications; (060.4080) Modulation. References and links 1. M. Popov, “The convergence of wired and wireless services delivery in access and home networks” Proc. Optical Fiber Communication conference (2010), paper OWQ6. 2. X. Mazda and F. Mazda, Focal Illustrated Dictionary of Telecommunications (Focal Press, 2013). 3. R. Hui, Z. Benyuan, H. Renxing, C. T. Allen, K. Demarest, and D. Richards, “Subcarrier multiplexing for highspeed optical transmission,” J. Lightwave Technol. 20(3), 417–427 (2002). 4. M. Attygalle, C. Lim, G. J. Pendock, A. Nirmalathas, and G. Edvell, “Transmission improvement in fiber wireless links using fiber Bragg gratings,” IEEE Photon. Technol. Lett. 17(1), 190–192 (2005). 5. J. Leibrich, A. Ali, H. Paul, W. Rosenkranz, and K. D. Kammeyer, “Impact of modulator bias on the OSNR requirement of direct-detection optical OFDM,” IEEE Photon. Technol. Lett. 21(15), 1033–1035 (2009). 6. P. Almeida and H. Silva, “Impact of the modulation chirp of a DEMZM on the transmission of signals based on OFDM,” IEEE Photon. Technol. Lett. 25(3), 283–286 (2013). 7. J. L. Wei, E. Hugues-Salas, R. P. Giddings, X. Q. Jin, X. Zheng, S. Mansoor, and J. M. Tang, “Wavelength reused bidirectional transmission of adaptively modulated optical OFDM signals in WDM-PONs incorporating SOA and RSOA intensity modulators,” Opt. Express 18(10), 9791–9808 (2010). 8. 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). 9. H. Silva, R. Fyath, and J. O'Reilly, “Sensitivity degradation with laser wavelength chirp for direct-detection optical receivers,” IEE Proc. Optoelectron.– Pt. J,136(4), 209–218, (1989). 10. C. Sánchez, B. Ortega, J. L. Wei, J. Tang, and J. Capmany, “Analytical formulation of directly modulated OOFDM signals transmitted over an IM/DD dispersive link,” Opt. Express 21(6), 7651–7666 (2013). 11. P. Almeida and H. Silva, “Expressions of the chirp parameter components for intensity modulation with a dualelectrode Mach-Zehnder modulator, ” Proc. International Conference on Transparent Optical Networks (2012). 12. S. Walklin and J. Conradi, “Effect of Mach-Zehnder modulator DC extinction ratio on residual chirp-induced dispersion in 10-Gb/s binary and AM-PSK duobinary lightwave systems,” IEEE Photon. Technol. Lett. 9(10), 1400–1402 (1997). 13. P. Almeida and H. Silva, “Corrections to ”Impact of the modulation chirp of a DEMZM on the transmission of signals based on OFDM,” IEEE Photon. Technol. Lett. 25(11), 1087–1087 (2013). 14. G. P. Agrawal, Fiber-Optic Communication Systems, (NJ, 2002). #196157 - $15.00 USD Received 20 Aug 2013; revised 4 Nov 2013; accepted 21 Nov 2013; published 6 Dec 2013 (C) 2013 OSA 16 December 2013 | Vol. 21, No. 25 | DOI:10.1364/OE.21.030764 | OPTICS EXPRESS 30764 15. L. V. T. Nguyen and D. B. Hunter, “A photonic technique for microwave frequency measurement,” IEEE Photon. Technol. Lett. 18(10), 1188–1190 (2006). 16. High rate ultra wideband PHY and MAC standard (2007). European Computer Manufacturers Association International Std. ECMA-368. 17. A. B. Carlson, P. B. Crilly, and J. C. Rutledge, Communications systemsAn introduction to signals and noise in electrical communication, (Mc Graw-Hill, 2002). 18. A. Ali, J. Leibrich, and W. Rosenkranz, “Spectral efficiency and receiver sensitivity in direct detection opticalOFDM,” Proc. Conference on Optical Fiber Communication conference (2009), paper OMT7. 19. W. Peng, K. Feng, A. E. Willner, and S. Chi, “Estimation of the bit error rate for direct-detected OFDM signals with optically preamplified receivers,” IEEE/OSA J. Lightwave Technol. 27(10), 1340–1346 (2009). 20. T. Alves and A. Cartaxo, “Analysis of methods of performance evaluation of direct-detection OFDM communication systems,” Fiber Integr. Opt. 29(3), 170–186 (2010). 21. T. Dennis and P. A. Williams, “Chirp characterization of external modulators with finite extinction ratio using linear optical sampling,” IEEE Photon. Technol. Lett. 22(9), 646–648 (2010). 22. F. Devaux, Y. Sorel, and J. F. Kerdiles, “Simple measurement of fiber dispersion and of chirp parameter of intensity modulated light emitter,” J. Lightwave Technol. 11(12), 1937–1940 (1993). 23. M. Morant, T. Alves, A. Cartaxo, and R. Llorente, “Transmission impairment compensation using broadband channel sounding in multi-format OFDM-based long-reach PONs,” Proc. Optical Fiber Communication Conference,(2012), paper OW3B.2. 1. Introduction Convergence of wired and wireless access networks into a single hybrid optical access infrastructure based on radio-over-fiber (RoF) networks has gained considerable momentum in recent years, due to the possibility of offering the final user a single simplified access to broadband services. Furthermore, the merging of two different access networks allows reducing upgrade and management costs of future broadband access networks [1]. This convergence, applied on long-reach passive optical networks (LR-PONs), allows reducing the network cost of operation further, since a larger geographic area can be covered with a single central office (CO) and without optical-electrical-optical conversions at the middle of the network. Intensity modulation direct-detection (IMDD) systems are largely used in optical access networks due to their simplicity and reduced cost. However, IMDD systems present some drawbacks, such as power fading induced by chromatic dispersion, which occurs after photodetection due to the inherent double sideband (DSB) spectrum, and low modulation efficiency, since at least 50% of the optical power is wasted in the optical carrier, which does not carry any information. To overcome the latter drawback, the modulation efficiency has been improved with double sideband - reduced carrier (DSB-RC) modulation [2], obtained by reducing the power of the optical carrier with an optical notch filter, such as a Fabry-Pérot filter in reflection mode [3] or a narrow band fiber Bragg grating [4]. More recently, a cost-effective method based on biasing the Lithium Niobate (LiNbO3) Mach-Zehnder modulator (MZM) below its quadrature point (QP), to reduce the power of the optical carrier, has been demonstrated [5]. To mitigate the former drawback, the power fading induced by chromatic dispersion in orthogonal frequency division multiplexing (OFDM) IMDD systems, the introduction of controlled negative modulation chirp by a dual electrode MZM (DEMZM) has been shown to mitigate the dispersion penalty [6]. Up to now, the transmission of optical signals in the presence of modulation chirp has been demonstrated only for intensity modulated signals with full optical carrier, through DEMZMs [6], semiconductor optical amplifiers [7], electroabsorption modulators [8], or laser diodes [9, 10]. In [11], it has been analytically demonstrated that, inside a bias voltage range around the QP of a DEMZM, the expressions derived for the α chirp parameter components can be used to estimate the modulation chirp, whereas at the maximum transmission point (MATP) or minimum transmission point (MITP) of the intensity transfer function of the DEMZM the expressions derived are not valid due to singularities in the expression of the α chirp parameter. Unlike the other aforementioned devices, the expression of the α chirp parameter of the DEMZM does not depend on the average optical power, i.e. power of the optical carrier. This opens the possibility of improving the modulation efficiency of IMDD systems, by employing DSB-RC modulation #196157 - $15.00 USD Received 20 Aug 2013; revised 4 Nov 2013; accepted 21 Nov 2013; published 6 Dec 2013 (C) 2013 OSA 16 December 2013 | Vol. 21, No. 25 | DOI:10.1364/OE.21.030764 | OPTICS EXPRESS 30765 along with controlled modulation chirp to mitigate the power fading induced by chromatic dispersion. In this paper, we investigate analytically and numerically the performance obtained with an optical hybrid signal (OHS) composed by a custom OFDM Gigabit Ethernet (OFDM-GbE) signal and three independent OFDM ultra-wideband (OFDM-UWB) signals, when modulation chirp is introduced in the intensity modulation with reduced carrier based on a DEMZM. The optical carrier is reduced by biasing the DEMZM below its QP, in a bias point determined as the best value for all signals considered. The remainder of this paper is organized as follows. A mathematical description of the effect of modulation chirp on the received signal and the effect of the DEMZM bias point on the modulation efficiency is presented in Section 2. In Section 3, the modeling of a wavelength division multiplexed passive optical network (WDM-PON) operating with a single wavelength is described, based on the equivalent block diagram of the system used to evaluate the performance obtained with the received OHS, when transmitted in the presence of modulation chirp. In the same section, the main parameters required to generate the baseband OFDM signals are presented. The numerical simulation results comparing the effect of modulation chirp on DSB and DSB-RC modulations are presented in Section 4, and the main specifications required to drive the DEMZM and limitations of the transmission system found are pointed out. Finally, in Section 5 the main conclusions of this work are drawn. 2. Theoretical analysis 2.1 Effect of modulation chirp of a DEMZM on the received signal A complex-valued baseband OFDM signal up-converted to the central frequency, frf (ωrf = 2πfrf), can be expressed by: () () () () () cos sin irfqrf x txt txt t ωω =− (1) where xi(t) and xq(t) are the baseband in-phase and quadrature components of the OFDM signal with zero mean and energy σ2 xi and σ2 xq, respectively, at the output of the OFDM modulator, before the RF up-converter. The optical field at the output of the LiNbO3 DEMZM with finite extinction ratio (ER) can be expressed by [12]: () () () () () 11 2 2 2, 2 cbias bias jt jVxtV jVxtV VV MZM Pe Et e e ππ ππ ω γ −+     =+      (2) where the amplitude drive voltages for any value of the α chirp parameter are expressed as [13]: () () () 2 12 cos sin 22cos 1 bias bias bias VV VV VmV VV ππ π π γπ γαγπ γπ γ +− =++ (3) () () () 22 cos 1 sin 22cos 1 bias bias bias VV VV VmV VV ππ π π γπ αγπ γπ γ ++ =− ++ (4) where Vbias = - (Vbias1 + Vbias2) is the bias voltage of the DEMZM, determined by the DC voltages Vbias1 and Vbias2, Vπ is the half-wave voltage of the DEMZM, m is the modulation index, P is the average power of the continuous wave (CW) laser emitting at the frequency νc (ωc = 2πνc) fed into the optical input of the DEMZM, and γ is the scaling factor, between 0 and 1, used to account for the unbalance between the arms and related with the ER of the DEMZM [12]. For balanced DEMZMs, this scaling factor is 1. Under small-signal modulation conditions (m << 1), the higher order sidebands are small and can be ignored, and thus the optical field signal given by Eq. (2) can be simplified to: #196157 - $15.00 USD Received 20 Aug 2013; revised 4 Nov 2013; accepted 21 Nov 2013; published 6 Dec 2013 (C) 2013 OSA 16 December 2013 | Vol. 21, No. 25 | DOI:10.1364/OE.21.030764 | OPTICS EXPRESS 30766 () () () { } 0 2, 2 crf crf c jt jt jt MZM P Et Ee Ee Ee ωω ωω ω −+ −+ ≅++ (5) with [] [] ** 11 2 2 01122 11 2 2 ) ) ) E EAA E γ γ γ − +   =− ΓΛ + ΓΛ   =Γ +Γ =ΓΛ+ΓΛ (6) with Γ1 = exp(-jπVbias1/Vπ), Γ2 = exp(jπVbias2/Vπ), Λk = J1(θkq(t)) + jJ1(θki(t)), Ak = J0(θki(t)) + J0(θkq(t)), where Jn(·) is the Bessel function of the first kind and order n (n = 0,1), and θki(t) = πVkxi(t)/Vπ, θkq(t) = -πVkxq(t)/Vπ are the modulation phase shifts of the baseband in-phase and quadrature components, respectively, when applied to the electrode k = 1,2. The superscript * denotes the complex conjugate and 1j=−.When this optical signal is transmitted over a dispersive single mode fiber, a phase shift is induced into each optical sideband, relative to the optical carrier. The Taylor’s expansion of the dispersion propagation constant of the dispersive fiber can be expressed as [14]: () ( ) () 2 01 2 2, cc βω β β ω ω β ω ω ≈+ − + − (7) where () |c mm mdd ωω ββωω = =. The effect of higher order fiber dispersion at 1550 nm is neglected. For the sideband centered at the frequency ωc ± ωrf, the dispersion propagation constant is given by: () 2 01 2 2. crf rf rf βω ω β β βω β ω ± ±==± + (8) After transmission over a standard single-mode fiber (SSMF) of length L, the optical field can be written as: () () { () } 0 0 2 ,. 2 crf crf c jt jt jLjt jL jL P EtL Ee e Ee e Ee e ωω ωω βω ββ −−+ − −−+ −+ ≅++ (9) Considering a limited bandwidth positive-intrinsic-negative (PIN) photodiode with responsivity ℜ, the photocurrent is expressed as: () ()() { } () * ,, DC s it EtLEtL i i t=ℜ = + (10) where iDC the DC photocurrent and is(t) the signal photocurrent. Considering the small-signal approximation, Jn(x)≈1/n!(x/2)n, we derived the following expression for the signal photocurrent as a function of the system parameters: () ( ) () () [] () () 22 2 21cos arctan cos crf sbias iq rf DfL it PmF xt jxt t c πλ π γ ααωτ ≈ℜ + + + −    (11) where Fbias = sin(π(Vbias1 + Vbias2)/Vπ), D is the dispersion parameter, λc is the carrier wavelength, and τ = β1L is the time delay suffered by the optical signal propagating a distance L through optical fiber with a group delay per unit length β1. In the derivation of Eq. (11), a very narrow-bandwidth OFDM signal was assumed, such that the amplitude factor cos(·) present in Eq. (11) for the central frequency frf can be assumed constant over all the signal bandwidth. It should be noted that, in the case of a real-valued OFDM signal, the expression of Eq. (11) continues valid, with xq(t) = 0. The average power of the photodetected signal can be approximated by: #196157 - $15.00 USD Received 20 Aug 2013; revised 4 Nov 2013; accepted 21 Nov 2013; published 6 Dec 2013 (C) 2013 OSA 16 December 2013 | Vol. 21, No. 25 | DOI:10.1364/OE.21.030764 | OPTICS EXPRESS 30767 () ( ) ()() () 2 22 22 22 2 2 21cosarctan crf rf s L L bias xi xq DfL PitR RPmF c πλ πγ σ σ α α =≈ℜ + + +    (12) where ⋅ denotes the averaging operator in the time domain and RL is the load resistor where the RF power is measured. The RF power given by Eq. (12) is in accordance with Eq. (1) in [15], when normalized by the RF power in optical back-to-back (B2B) configuration, considering as modulating signal a real sinusoidal signal (xq(t) = 0). Thus, Eq. (12) is the general expression of the RF power after photodetection, in the presence of the modulation chirp. From Eq. (12) we can conclude that: (1) When the DEMZM is biased at its null intensity point, (Vbias1 and Vbias2 equal to Vπ/2), there is no signal photocurrent located at the frequency ωrf;this is clear since, in this case, at the output of the DEMZM only odd-order harmonics of ωrf are present, and there is no optical carrier to beat with the 1st order harmonics to generate the photocurrent signal at ωrf. (2) The negative values of the chirp parameter factor arctan(α), within the argument of the cosine function, have the effect of generating a phase shift opposite to that caused by the cumulative dispersion along the fiber; thus, the first null occurs for a greater fiber length in the presence of negative chirp, allowing mitigation of the power fading induced by chromatic dispersion for a given distance. (3) The chirp parameter has also impact on the maximum received power, through the factor (1 + α2). In order to validate Eq. (12), Fig. 1 depicts the RF power of the received signal normalized by its maximum, obtained analytically from Eq. (12) and from numerical simulation of an OFDM-UWB signal following the ECMA-386 standard [16] centered at the frequency frf = 3.960 GHz, which corresponds to the 1st of 14 channels. The DEMZM is biased at the quadrature point. More details about the OFDM-UWB signal and the optical system parameters considered are provided in Section 3. From Fig. 1 it is concluded that the analytical model derived for the signal photocurrent and its RF power shows excellent agreement with the numerical simulation results obtained, with a small discrepancy. This small discrepancy between the curves can be due to the high peak-to-average power ratio (PAPR) of the components xi(t) and xq(t). Hence, the approximation of the Bessel functions in Eq. (11) may not be accurate since xi(t) and xq(t) have high-peaks at some points. Furthermore, it is observed that the first null for α = −1 occurs for a larger fiber length than for α = 0. 2.2 Effect of DEMZM bias point on the modulation efficiency For the small-signal analysis of the effect of DEMZM bias voltage on the optical powers and photodetected RF power, the DEMZM is considered without insertion losses, balanced (γ = 1), and the modulation is assumed chirpless (α = 0). For these conditions, from Eq. (5) and Eq. (6), the power of the optical carrier is given by: () 2 2 012 241cos , 2 OC bias bias P PEPVV V π π    ==++         (13) and the average optical power of the signal sidebands is given by: #196157 - $15.00 USD Received 20 Aug 2013; revised 4 Nov 2013; accepted 21 Nov 2013; published 6 Dec 2013 (C) 2013 OSA 16 December 2013 | Vol. 21, No. 25 | DOI:10.1364/OE.21.030764 | OPTICS EXPRESS 30768 015 30 45 60 75 90 105 120 135 150 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Fiber length (km) Normalized RF power = 0 =-1 Analytical Numerical Fig. 1. Normalized RF powers as a function of fiber length, for an OFDM signal centered at 3.96 GHz. () () () () 22 22 2 222 12 22 2 22 =1cos. 2 OSB xi xq bias bias PP PEE E PmVV V π π πσσ −+ +   =+=       +− +      (14) The RF power in optical B2B configuration, given by Eq. (11) for L = 0 and under the aforementioned conditions, can be rewritten taking into account the product of Eq. (13) by Eq. (14): 2. rf L OC OSB PRPP=ℜ (15) Figure 2 depicts the optical power of the carrier, the optical power of the signal sidebands, the total optical power (POC + POSB) and the photodetected RF power, normalized by their maximum values, as functions of the bias voltage normalized by the half-wave voltage of the DEMZM. The normalized power of the optical carrier (POC) is equal to the intensity transfer function of an ideal DEMZM for negative bias voltages. In Fig. 2 it can be observed that, when the bias voltage of the DEMZM decreases from the QP towards the MITP, the power of the optical carrier decreases whereas the optical power of the signal sidebands increases in the same proportion, and consequently the total optical power becomes dominated by the power of the signal sidebands. Otherwise, when the bias voltage is increased from the QP towards the MATP, the power of the optical carrier increases and the total optical power is dominated by the carrier optical power. The performance and reach of the optical communication system are determined by POSB. By analogy with amplitude modulation in RF communication systems [17], the modulation efficiency is expressed by η = POSB/(POC + POSB) × 100. Therefore, the efficiency reaches 100% if the optical carrier is totally suppressed, which requires the addition of an optical local oscillator at the receiver side to perform coherent detection. In direct detection systems, the most used systems in optical access networks, a fraction of the optical carrier must be transmitted to enable direct-detection. Therefore, the efficiency reaches 100% if the optical carrier is totally suppressed, which requires the addition of an optical local oscillator #196157 - $15.00 USD Received 20 Aug 2013; revised 4 Nov 2013; accepted 21 Nov 2013; published 6 Dec 2013 (C) 2013 OSA 16 December 2013 | Vol. 21, No. 25 | DOI:10.1364/OE.21.030764 | OPTICS EXPRESS 30769 at the receiver side to perform coherent detection. In direct detection systems, the most used systems Fig. 2. Relationship between the optical powers and the RF power as functions of the Vbias voltage. in optical access networks, a fraction of the optical carrier must be transmitted to enable direct-detection. Typically, the optimum performance is reached when the power of the optical carrier is reduced and equal to the optical power of the sidebands [4]. Therefore, with POC = POSB, the optimum modulation efficiency is 50% for direct-detection systems. Additionally, in Fig. 2 it can be observed that the maximum RF power is obtained when POC = POSB, for the bias voltage equal to −0.5Vπ, because the optical powers are normalized. Typically, at the output of the DEMZM biased at the QP, the power ratio between the optical carrier and the sidebands is approximately 20 dB [4]. 3. Simulation setup Figure 3 represents schematically a hybrid long-reach WDM-PON (LR-WDM-PON) for a single wavelength transmitted from the optical line terminal (OLT) located in the CO to the hybrid optical network unit/base station (ONU/BS) located at the customer’s premises. In the CO, the optical transmitter based on the DEMZM has the modulation chirp controlled through the amplitudes of the drive signals [6]. MZM CW LASER Chirp Controller V1 Adjustment V2 Adjustment Noise Loading OF PIN EDFA VOA ONU/BS EHS RX EA VEA SSMF SSMF RN 80 km 0-20 km CO EDFA () 1 Vt () 2 Vt UWB #1 TX UWB #2 TX UWB #3 TX GbE TX 528 MHz 528 MHz #1 #2 #3 #1#2#3 0f BOFDM Bg UWBs GbE Fig. 3. Simulation transmission scheme of an OHS over a hybrid long-reach FTTH network with controlled modulation chirp. #196157 - $15.00 USD Received 20 Aug 2013; revised 4 Nov 2013; accepted 21 Nov 2013; published 6 Dec 2013 (C) 2013 OSA 16 December 2013 | Vol. 21, No. 25 | DOI:10.1364/OE.21.030764 | OPTICS EXPRESS 30770 An electrical hybrid signal (EHS) is created by subcarrier multiplexing four independent OFDM signals, a custom quadrature amplitude modulation (QAM) OFDM-Gigabit Ethernet (GbE) signal with 1 GHz bandwidth, and three independent quadrature phase shift keying (QPSK) OFDM-UWB signals following the ECMA-368 standard [16]. The three OFDM-UWB signals, UWB#1, UWB#2 and UWB#3, employing QPSK modulation with 528 MHz bandwidth, are up-converted to the frequencies 3.432, 3.960 and 4.488 GHz, respectively. These three channel signals perform no frequency hopping, with the three channels at time frequency codes (TFC) TFC5, TFC6 and TFC7, respectively. The subcarriers allocation in the OFDM-GbE signal is similar to the UWB signals. The OFDMGbE signal is the unique that is not yet standardized, and thus can be freely translated in frequency, in order to maximize the guard band (Bg) between the DC frequency and the lower edge of the GbE signal spectrum and minimize the aggregate bandwidth of the OFDM signals (BOFDM), as illustrated in Fig. 3. Thus, by allocating a guard band of 100 MHz between the GbE signal and the lower frequency UWB signal, the GbE signal is up-converted to the central frequency equal to 2.568 GHz, resulting in a BOFDM = 2.684 GHz and a Bg = 2.068 GHz, which produce the ratio Bg/BOFDM = 0.7705. Therefore, Bg is insufficient to accommodate all intermodulation distortion products generated upon direct-detection, since it has been shown that Bg should be at least equal to BOFDM [18]. The main parameters used to generate the baseband OFDM signals used in the simulations are gathered in Table 1. Table 1. Main Parameters Used to Generate the Baseband OFDM Signals GbE UWB1,2,3 Nominal Bandwidth [MHz] 1000 528 Sampling frequency [MHz] 1056 528 FFT size (Total of subcarriers) 256 128 Data subcarriers 216 100 Pilot subcarriers 24 12 Guard subcarriers 15 10 Cyclic prefix [samples] 64 37* Symbol time[μs] 0.303 0.312 Constellation mapping QAM QPS K Central RF frequency[GHz] 2.568 3.432;3.960 and 4.488 Net bit rate [Mbit/s] 1425.7 641 References ~ [16] [16] *The cyclic prefix in the OFDM-UWB signals are null. The EHS involving four separately amplified OFDM signals modulates externally a continuous wave (CW) laser with a linewidth of 5 MHz (conventional DFB laser) emitting at 1550 nm, through a DEMZM with 25 dB DC extinction ratio, half-wave voltage Vπ = 5 V and 6 dB of insertion loss. The modulation index m is adjusted through a variable electrical attenuator (VEA). The model of the DEMZM is based on Eq. (2). The resulting OHS at the output of the DEMZM is transmitted through SSMF to the costumer’s premises. The SSMF is modeled as the low-pass equivalent of a linear bandpass system. The SSMF is characterized by an attenuation of 0.22 dB/km and a chromatic dispersion parameter D = 17 ps/km/nm. After transmission through 80 km of SSMF, at the remote node (RN), the optical signal is amplified by a noiseless erbium doped fiber amplifier (EDFA) to compensate for the transmission losses. From the RN, the OHS is transmitted through fiber segments of length 0 km, 10 km and 20 km, to the hybrid ONU/BS located at the costumer’s premises, in order to simulate the different locations of the customers relatively to the RN. At the input of the hybrid ONU/BS, a variable optical attenuator (VOA) along with an EDFA are used to adjust the optical signal-to-noise ratio (OSNR) of the system, defined in a reference optical bandwidth of 0.1 nm. The OSNR is evaluated considering optical noise over the two perpendicular directions of polarization. In the hybrid ONU/BS, a second-order super-Gaussian optical filter with 12.5 GHz full width at half maximum (FWHM) bandwidth, centered at 1550 nm, is used to reduce the amplified spontaneous emission (ASE) noise power, and a PIN photodetector with responsivity of 0.9 A/W is considered. The ASE noise is modeled as Gaussian noise. The #196157 - $15.00 USD Received 20 Aug 2013; revised 4 Nov 2013; accepted 21 Nov 2013; published 6 Dec 2013 (C) 2013 OSA 16 December 2013 | Vol. 21, No. 25 | DOI:10.1364/OE.21.030764 | OPTICS EXPRESS 30771 thermal noise of the receiver is not considered, since at the output of the PIN photodetector its effect is negligible, compared with the signal–ASE beating noise [19]. The received EHS is demodulated in the EHS receiver (EHS RX), where each OFDM signal is sent to the respective receiver to perform down-conversion and OFDM demodulation. The OFDM demodulation includes removal of cyclic prefix, serial-to-parallel conversion, FFT, subcarriers equalization, symbol de-mapping and parallel-to-serial conversion. The performance for each of the received OFDM signals is assessed through the required OSNR (ROSNR) for a bit error rate (BER) equal to 10−9, estimated by the error vector magnitude (EVM) method applied to the constellation of the received signal at the equalizer output [20]. 4. Numerical simulation results and discussion The simulation results presented in this paper were produced with Matlab®, considering 528 OFDM-GbE symbols and 512 OFDM-UWB symbols. Since the OFDM signals have a spectrum similar to Gaussian noise and different bandwidths, the electrical power of each signal was adjusted for a power spectral density equal to −139.6 dB/Hz. Initially, the optimum bias voltage of the DEMZM required to maximize the modulation efficiency of the system was investigated for an optical B2B configuration. The DEMZM was driven in balanced push-pull mode. In Fig. 4, contour plots of the ROSNR for a BER = 10−9 are shown as a function of the modulation index m for different values of the bias voltage normalized by the half-wave voltage of the DEMZM. Note that a BER = 10−9 is enough to ensure that the EVM limit of −14.5 dB after transmission, defined in the standard for QPSK OFDM-UWB signals, is respected. Biasing the DEMZM below the QP, it operates in the non-linear region of its intensity transfer function, which degrades the transmitted signals due to the generation of harmonics and intermodulation distortion products. From Fig. 4, it can be concluded that the ROSNR can be kept constant when the DEMZM is biased below the QP, by decreasing m. The ROSNR for the UWB #2 signal shown in Fig. 4(c)) presents a degradation of nearly 2 dB relative to the UWB #1 and UWB #3 signals, due to linear crosstalk suffered from the adjacent channels. It can also be concluded that the lowest bias voltage of the DEMZM is limited by the GbE signal and is equal to −0.8 Vπ. Therefore, the optimum bias voltage equal to −0.8Vπ was selected to define the DSB-RC modulation, as the best value between all signals considered. In Fig. 5, the ROSNR is presented for all signals as a function of m when the DEMZM is biased at the QP, with Vbias = −0.5Vπ, and below the QP, at Vbias = −0.8Vπ. Biasing the DEMZM at the QP, all signals present approximately the same ROSNR as m increases, whereas when the DEMZM is biased at −0.8Vπ the GbE and UWB #3 signals exhibit degradation relatively to the other signals, due to spurious spectral components resulting from the nonlinearity of the DEMZM transfer function and the square-law of the direct-detection. The GbE signal is degraded because the ratio Bg/BOFDM is lower than 1, and thus a fraction of the GbE signal is distorted by intermodulation spectral components near DC, after photodetection. When the DEMZM is operated below the QP, the 2nd harmonic power of the GbE signal at 5.136 GHz increases, affecting the UWB#3 signal centered at 4.488 GHz. Furthermore, from Fig. 5 it is observed that the optimum values of m to drive the DEMZM are 0.35 and 0.45, for the DEMZM biased at −0.8 Vπ and −0.5 Vπ, respectively. The average OSNR improvement of DSB-RC (Vbias = −0.8Vπ) relative to DSB with full carrier (Vbias = −0.5Vπ) for the optimum values of m, among all signals, is nearly 6 dB. The drive voltage amplitudes expressed by Eq. (3) and Eq. (4) were derived from the small-signal approximation chirp parameter components presented in [11], and depend on the bias point of the DEMZM. Therefore, it is important to evaluate the accuracy of Eq. (3) and Eq. (4) for the optimum values of m obtained for different bias points of the DEMZM selected. The α chirp parameter value at the DEMZM output was measured as described in [11,21]. Figure 6 presents the absolute error between the desired α chirp parameter at the output of the DEMZM and the value obtained as a function of m, for the wired and wireless signals modulated in DSB-RC and DSB. #196157 - $15.00 USD Received 20 Aug 2013; revised 4 Nov 2013; accepted 21 Nov 2013; published 6 Dec 2013 (C) 2013 OSA 16 December 2013 | Vol. 21, No. 25 | DOI:10.1364/OE.21.030764 | OPTICS EXPRESS 30772