( )
(t I i t
I = p + s , (2-13) where Ip=RPin is the average current and is(t) is a current fluctuation related to shot noise.
The shot noise is represented by a mean square shot noise current generator, f
qI is >= p∆
< 2 2 , (2-14) where ∆f is the effective noise bandwidth. The same bandwidth appears in the case of both shot and thermal noises.
2.6 Distortion characteristics
In above section, we explored one type of extraneous signals in links – noise. In this section, we will investigate the other type of extraneous signals in links – distortion.
Unlike noise however, distortion signals are deterministic. We begin as assuming that the modulation is a pure sinusoid. It generates light power modulations at the input frequency ω and also at the harmonics 2ω, 3ω, ···, nω. Harmonic distortion (HD) depends on the nonlinearity of a transmission system. The amplitude of the nth order harmonic is proportional to the nth power of the optical modulation index (OMI) and it decreases rapidly with higher order. The optical modulation is defined as
P^
P I
I m I
th b
= ∆
−
≡ ∆ , (2-15)
where ∆I is the variation of electrical driving current around a bias point in modulating a laser light source, Ib is the laser bias current, and Ith is the laser threshold current.
For a far common RF practice, two equal-amplitude sinusoids are used to characterizing distortion. The harmonic terms for each frequency are of course identical
simultaneously, additional distortion terms are generated. These are referred to as the intermodulation distortion. Two signals at ω1 and ω2, for example, are accompanied by second order distortions are frequencies 2ω1,2, ω1±ω2, and third order distortions at frequencies 3ω1,2, 2ω1±ω2, and 2ω2±ω1, etc. The third order modulation distortion (IMDs) at 2ω1-ω2 and 2ω2-ω1 are special interest since they are close to original signals and they might interfere with other signals in multichannels applications. The third order IMD increases as the cube of the OMI [29].
The power at which the fundamental and one of the distortion curves intersect is one measure of distortion. As we will see below, practical systems are always operated below this point to avoid serious distortion, but the intersection point is a useful measure of system performance. Clearly the higher the intercept point, the lower the distortion at a given power. In general there will be a separate intercept point for the second-order, IP2, and third-order, IP3, distortions, as shown in Fig. 2-14. In this figure, we can find another measure of distortion–intermodulation-free dynamic range (IM-free DR). It is also called spurious free dynamic range (SFDR). The IM-free DR is the function of both distortion and the noise of the component. Thus to define the IM-free DR we need to add the noise power to Fig. 2-14.
In multichannel applications such as CATV and cellular telephone, multiple, regularly spaced carriers are frequency multiplexed onto the optical link. In such cases, these multiple signals can interfere and add to levels that are greater than simple power addition, so there is a need for a new distortion measure that makes statistical assumptions. The common measures that are often used are composite second order (CSO) and composite triple beat (CTB) quantities.
Fig. 2-14 Plot of the output signal powers - fundamental, second- and third-order intermodulation - as a function of the inpur signal power.
(Ref. Charles H. Cox. III, Analog optical links, 2004)
2.7 Carrier-to-noise and distortion ratio
An RF lightwave system consists of transmitters, fiber link, and receivers. The performance of this system is determined by the performance of those active and also passive components. All these can be quantified by carrier-to-noise ratio (CNR), and second and third order distortion (CSO and CTB). This is shown in Fig.2-15.
Fig. 2-15 CNR, CTB and CSO in RF subcarrier system
(Ref. William S.C. Chang, RF Photonic Technology in Optical Fiber Links, 2002)
We can also combine these measures as carrier-to-noise-and-distortion ratio (CNDR) which directly determines the received signal quality at customer premises equipment.
CNDR is described as
1
The first three terms on the right of this equation represent the CNRs by receiver noise, shot noise, and RIN, respectively. These three terms can be written as
>
The fourth term on the right of Eq.(2-16) represents the carrier-to-distortion ratio (CDR), which can be written as [30]
2
In this equation, we ignored the second-order intermodulation term since most fiber optic networks for transporting wireless signals utilize less than one octave of bandwidth. Thus
as
where k=1 represent the first channel [30]. The highest number of intermodulation terms, which falls on the center of the signal band be comes to be (3N2-14N+8)/8.
The last term on the right of Eq.(2-16) represents CDR degradation due to clipping distortion which occurs when the bias current falls below threshold current instantaneously [31]. It can be expressed as
2 ) where μis the total rms modulation index m(N/2)1/2. The clipping distortion becomes important when the total rms modulation index is greater than ~25%.
Chapter 3 Simulations of Radio over Optical FSK and DPSK Schemes
3.1 Introduction
The demand on broadband services in both fixed and wireless access network is increasing. The millimeter-wave band is considered to be a promising solution because of its spectrum availability, compact size of radio frequency (RF) devices. There is an emerging need to integrate the fixed wired network and the Rof wireless network with low cost infrastructure. Simultaneous transmission of 10Gbps baseband and 60-GHz-band radio signals on a single wavelength is proposed [3]. Simultaneous three-band modulation of a 2.5Gbps baseband, 5.2GHz microwave and 60-GHZ-band signals is also experimentally and theoretically studied, recently [4]. However, the interference and nonlinear distortion among multiple radio channels should be carefully considered at a practical point of view, and these had not yet been investigated so far.
In this chapter, we propose two radio over optical schemes which load the analog signals on the differential phase-shifting-keying (DPSK) and frequency-shift-keying (FSK) modulated baseband data, respectively. In these schemes, baseband data is modulated through FSK and DPSK without affecting intensity so analog signals can be transmitted by conventional intensity modulation and direct detection [5].
In section 3.2, we introduce the simulation tool simply. The operation principles and performance of the FSK and DPSK schemes are described in section 3.3 and 3.4, respectively. In section 3.5, we compare the results of the two schemes. Finally, the future work is given in section 3.6.
3.2 Simulation Tool
Theoretical analysis in this chapter is implemented by using the commercial simulation software, VPItransmissionMakerTM (VPI). VPI is an integrated design environment for individual collaborative design teams and it offers:
． Design Assistants: these capture common simulation tasks and provide automated synthesis and verification tools. They are written in a simple scripting language, mimicking keyboard and mouse interactions with the simulator, so enabling you to incorporate your own design rules and test specifications,
． vertically-integrated design process: with layered simulation technologies that enable concurrent analysis across network levels,
． co-simulation capability: so designers can integrate third party tools and their own models alongside a comprehensive library of component and system models,
． tailored GUI: with easy to use optimizers, parameter and module sweep capability,
．extensive library of photonic modules covering the latest photonic technologies over several levels of abstraction, from detailed physical models, to Black-Box, measured, and Data Sheet Models.
3.3 Radio over optical FSK scheme
In this section, we demonstrate a radio over optical FSK scheme which simultaneous transmits 10Gbps baseband data and 60-GHz-band radio signals on a single wavelength.
Fig. 3-1 The proposed radio over optical FSK scheme
Our proposed modulation scheme consists of a distributed feed-back (DFB) laser, an electroabsorption (EA) modulator, and a dual-driven Mach-Zehnder modulator (DD-MZM) as shown in Fig. 3-1. A DFB laser with large frequency chirp for FSK encoding is used as the light source. The simplest method to implement optical FSK is to directly modulate a semiconductor laser. The frequency output from the laser will change with the variation of the injected data current as shown in Fig. 3-1(a), where f0 and f1
represent data bit “0” and “1”, respectively. The residual intensity modulation of the baseband data makes the carrier unsuitable for radio application. To overcome this problem, we put an EA modulator after the DFB laser. The EA modulator with inverted baseband data and appropriate time delay is used to equalize the power fluctuation caused by the parasitic intensity modulation. The equalized power is shown in Fig. 3-1(b). The radio signals are, then, loaded on the envelope of the baseband in single-side band (SSB) format by a DD-MZM baised at the quadrature point, shown in Fig. 3-1(c). The SSB
FSK BASEBAND DATA
Chirp Laser EA modulator
Radio Signal Bias
900 A
B
C
DATA DATA
Electrical Delay Line
f1 f0 f1 f0 f1
1 0 1 0 1
A
f1 f0 f1 f0 f1
1 0 1 0 1
B
1 0 1 0 1
C
f1 f0 f1 f0 f1
fiber [6]. This technique will provide a better linearity and the potential to accommodate more channels than the previous work [3-4], and thus is more suitable for hybrid radio/digital transmission or optical labeling applications.
3.3.1 Optical single-side band modulation
Double side band modulation in a typical radio over fiber transmission has been explored and proved that the reach is strongly limited by chromatic dispersion as the increasing radio frequency. Signal degrades due to the inevitable coherent interference between the two different phase shifts of the two band signals [7]. The detected RF power varies as
W ∝ cos 2(πlcD(f / f0)2) [8] (2-1)
where
D fiber group velocity dispersion parameter (ps/km);
C velocity of light in vacuum;
L link length;
f0 optical carrier frequency;
f frequency of the subcarrier.
If one of the subcarrier sideband is removed, this situation is improved and the signal power remains constant. Accordingly, Optical SSB modulation is proposed as a solution for improving the transmission quality of high-frequency subcarrier optical systems [9-11].
In order to obtain SSB, using a DD-MZM is regarded as a simple and wise method.
The simplified DD-MZM configuration is shown as in Fig. 3-2, composed of two waveguides on the LiNbO3 substrate. The phase of light wave going through each waveguide is modulated by each RF signal.
Φ RF signal:Sin ωs t RF signal:Sin (ωs t+Φ)
Opt IN Opt OUT
Fig. 3-2 The DD-MZM configuration
In Table 3-1, we summarize three different types of optical modulation, such as IM, SSB, and Double Side Band with Suppressed Carrier (DSB-SC).
Optical RF ψ
Φ -π
2
π 2 π
-π 2
π π 2
ωc ωc+ωs
SSB
ωc-ωs ωc
SSB
ωc ωc+ωs
SSB ωc-ωs ωc
SSB ωc-ωs ωc
IM
ωc+ωs ωc-ωs ωc
IM ωc+ωs ωc-ωs
DSB-SC ωc+ωs
Table 3-1 Different types of optical modulation
3.3.2 Simulation results and discussion
Fig. 3-3 The FSK simulation diagram
The system diagram is shown in Fig. 3-3. The characteristics of the DFB laser are shown in Fig. 3-4(a), including laser chirp and power related to the input current. The transfer function of the EA modulator in simulation is shown in Fig. 3-4(b). It can be seen in Fig. 3-4(a) that both the power and wavelength change with the variation of the injection current. We drive the laser at 10Gbps data rate and the space and mark current levels are tuned to 50 and 80 mA, respectively. Thus the frequency difference between data bits “0” and “1” is about 23GHz. The extinction ratio (ER) of the output signal from the laser is about 1.6-dB and the eye pattern is shown in Fig. 3-5(a).
LD 10Gbps
inverter EA M od
SSB M Z M od
60GHz 61GHz
Optical Attenuator
Optical Filter
Frequency Discrim inator
Optical Detector Scope
Baseband Data
Optical Detector
EDFA
RF Spectrum
3dB Coupler
Power amplifier
Power amplifier
Pre-amplifier
(a)
(b)
Fig. 3-4 (a) The laser chirp and output power characteristics related to
input current. (b) The EA modulator relative output power vs. driving voltage.
To compensate the intensity fluctuation, we drive the EA modulator with -0.03965V and -0.06345V which can attenuate the power of bit “1” and bit “0”. Consequently a uniform power is achieved and shown in Fig. 3-5(b).
0.0 0 0.0 2 0.0 4 0 .06 0 .0 8
In put C urrent o f D F B L aser (A )0 .1 0
Relative Output Power (dBm)
1 55 2.4 1 5 52.3 1 5 52.2 15 5 2.1 1 55 2.0 1 55 1.9 1 5 51.8
Wavelength(nm)
14 12 10 8 6 4 2 0
0 -1 -2 -3 -4 -5 -6
0.00 0.02 0.04 0.06 0.08
Bias Voltage (V)
Relative Output Power (dBm)
(a)
(b)
Fig. 3-5 (a) The output eye diagram after laser (b) Eye diagram after EA compensation
The optical spectrum output from the EA modulator is shown in Fig. 3-6(a), where the power of the two peaks (bit “1” and “0”) located at 1551.97 and 1552.15 nm. A DD-MZM is used for SSB radio modulation. Signals with frequency of 60GHz and 61GHz are selected for demonstrating the radio link. The output optical spectrum after radio SSB modulation is shown in Fig. 3-6(b), where the radio side band is apart from the center carrier about 60GHz.
9 10
8 7
5
6
4 3
1 2
0
20 ps/div
Power(mW)
0.348m W
Z ero level 10 ps/div
(a)
(b)
Fig. 3-6 (a) Optical spectrum of EA output. (b) Output spectrum of SSB DD-MZM.
The hybrid signals are transmitted through 75-km standard single-mode fiber (SSMF) with dispersion parameter of D= 16-ps/(km.nm). At the receiver node, the power is divided into two branches, while one is for baseband data and the other is for radio signals. At the baseband branch, an optical filter (1st order Gaussian profile) with
1551.2 1551.6 1552.0 1552.4 1552.8
W avelength (nm)
Power(dBm)
-20
-80 -50
-110
Power(dBm)
-20
-80 -50
1551.2 1551.6 1552.0 1552.4 1552.8
Wavelength (nm)
bandwidth of 15GHz is used to retain the information at 1552.15 nm and convert the signal to an on-off keying (OOK) signal for easy detection. The spectrum of the converted signal is shown in Fig. 3-7, where the power of 1551.97 nm is attenuated about 25-dB than that of 1552.15nm.
Fig. 3-7 Optical spectrum after the optical filter
At the analog branch, the signals are directly injected into the photodiode and its performance is observed by a RF analyzer. Fig. 3-8 shows the picture of carrier- to-noise and distortion ratio (CNDR) related to the optical modulation index (OMI) of radio signal.
As indicated in Fig. 3-8, the scheme without EA compensation has less CNDR than that with compensation at the points of the same OMI. This is due to that the intensity variation of baseband data has interference influence on the radio signals. To see this situation clearly, we show the two spectrums with and without compensation scheme in Fig. 3-9 for comparisons.
Power(dBm)
-20
-120 -70
-170
1551.9 1552.0 1552.1 1552.2 1552.3 W avelength (nm)
Fig. 3-8 The carrier to noise and distortion ratio in both scheme cases with and without EAM
-50
Fig.3-9 Comparison of the spectrums with and without EAM
With 300k thermal and shot noise considered, the baseband data performance with and without EAM is shown in Fig. 3-10(a). Fig. 3-10(b) are the eye diagrams with and without EAM at BER = 10-9. It can be seen that the scheme with EA compensation is
0.1 1
1.4-dB better than that without EA compensation in the back-to-back case. And after 75-km SSMF transmission, sensitivity of the scheme with EA compensation can be improved about 1.8-dB than that without EA compensation. The transmission penalties for the schemes with and without EAM are 2.9 and 3.1-dB, respectively.
-32 -31 -30 -29 -28 -27 -26 -25 -24 -23 -22 -21 -20 -12
-11 -10 -9 -8 -7 -6 -5 -4 -3
LOG(BER)
Received power(dBm)
back to back(w/o EAM) 75km SMF(w/o EAM) back to back(with EAM) 75km SMF(with EAM)
back to back (with EAM) back to back (w/o EAM) 75km SMF (with EAM) 75km SMF (w/o EAM) (b)
Fig. 3-10 (a) Probability of error of the received baseband data (b) The eye diagrams of the schemes with and without EAM at BER
= 10-9
3.4 Radio over optical DPSK scheme
DPSK carries the information in optical phase changes between bits. This modulation format is important for fiber-optic communication systems. It has a
(a)
significant benefit of requiring ~3dB lower optical-signal-to-noise ratio (OSNR) than OOK to get the same bit-error rate. In this section, we demonstrate a radio over optical DPSK scheme that simultaneously transmits 10Gbps baseband data and 60-GHz-band radio signals on a single wavelength. The simulation diagram is shown in Fig. 3-11.
Bias
Fig. 3-11 The simulation setup of DPSK scheme
In this setup, we perform the phase modulation by a Dual-Driven Mach-Zehnder modulator (DD-MZM). The output optical electric field of DD-MZ modulator is expressed as where Ein is the input optical electric field, V1 and V2 are the two applied voltages, and Vπ is the switching voltage. According to the above equation, we design the radio over
optical DPSK scheme with the intensity-modulated RF signals and DPSK baseband data.
3.4.1 DPSK demodulation
delay interferometer (DI) with a bit period delay between two arms and converted into intensity modulation for easy detection. There are two types of the optical delay line demodulator, one is wrapped around a piezoelectric transducer (PZT) to allow active locking of the interferometer operating point to a fraction of a wavelength and the other is silica integrated optical waveguide with a integral thermal heater which allowed active trimming of the optical delay [13].
Fig. 3-12 A typical DPSK receiver
But the common demodulation method is complex and expensive. In order to solve the problem, a fresh demodulation idea is proposed—a discriminator filter followed by direct detection [12]. The discriminator filter has many advantages: simplify the demodulation design, decrease the cost, and hold the advantage of ASE noise-limited performance over OOK. In Fig. 3-13, we compare the spectrums of DI and narrow band filter and 10Gbps NRZ-DPSK signal spectrum at input and output of BPF [12].
Therefore, in our proposed scheme, we use a discriminator filter as DPSK demodulator.
Optical signal
1bit delay
DATA Output T
0 π 0 π π
Delay Interferometer
Fig. 3-13 (a)The spectrum of DI (b) The spectrum of narrow band filter (c) 10Gbps NRZ-DPSK signal spectrum at the input and
output of BPF
3.4.2 Simulation results and discussion
The proposed configuration is shown in Fig. 3-11. It is based on the use of a DD-MZM biased at the quadrature point, which externally modulates a CW laser source.
The 10Gbps baseband data and two 60-GHz-band radio signals are first separated into 5 GHz
10 dB
FW HM = 6.2 GHz
After filter NRZ-DPSK (a)
(b)
(c)
the lower arm signals to drive the DD-MZM. One part of radio signals is inverted. Then we can generate hybrid analog/digital signals with baseband data of DPSK modulation and radio signals of intensity modulation.
After the modulation, the optical signal is transmitted through an optical attenuator and divided into two branches. At the receiver node of RF signals, the signals are directly injected into the photodiode and its performance is observed by a RF analyzer. At the baseband branch, an ultranarrow optical bandpass filter (1st order Gaussian profile) with 3-dB bandwidth of 5GHz is used to reject ASE noise and demodulate DPSK [12]. The optical signal is directly converted to the electrical signals by a photodiode with responsibility of 0.7, 300k thermal noise and shot noise. The purpose of the electrical low pass filter after photodiode with 3dB bandwidth of 7GHz is to reduce the noise power.
Finally, the signal is feed to an error detector to estimate the performance. The results are shown as BER in Fig. 3-14(a). And Fig. 3-14(b) show the eye diagrams before and after filter at BER=10-9 in back to back situation. The power penalties of 75-km and 150-km SMF transmission with reference to back-to-back at BER = 10-9 are 0.1dB and 0.9dB, respectively.
-42 -41 -40 -39 -38 -37 -36
-11 -10 -9 -8 -7 -6 -5 -4
LOG(BER)
Received Power(dBm)
back to back 75km SMF 150km SMF
(a)
(b)
Fig. 3-14 (a) Probability of error of the received baseband data (b) Eye diagrams before and after filter at BER=10-9
in back to back situation.
3.5 Discussion and comparison
Fig. 3-15 shows the plot of analog CNDR related to OMI in DPSK and FSK schemes. It can be seen the CNDR performances of the two schemes are almost the same.
This is because that the difference between these two configurations is the modulation method of baseband data.
0.1 1
15 20 25 30 35 40 45 50 55
CNDR
OMI
DPSK scheme FSK scheme
Fig. 3-15 CNDR vs. OMI for DPSK and FSK schemes
Zero level
Before filter After filter
Zero level Zero level 10ps/div
10ps/div
The measured receiver sensitivities for DPSK and FSK schemes are shown in Fig.
3-16. In FSK scheme, we use an optical filter with bandwidth of 15GHz to convert the signal to an OOK signal. And in DPSK scheme, the demodulator filter has 5GHz bandwidth. The comparison is almost fair: the electrical bandwidth of Be= 9GHz and optical BPF bandwidth of B0 = 15GHz are close to theoretical optimum for NRZ-OOK [14]. The receiver sensitivity of DPSK is -38.3dBm, in comparison, the receiver sensitivity of FSK is -36.75dBm. The DPSK scheme achieves 1.55-dB better than the FSK scheme since ASE noise is dominated in these two systems and the 5GHz filter in the DPSK scheme filter out more ASE noise.
-42 -41 -40 -39 -38 -37 -36
-11 -10 -9 -8 -7 -6 -5 -4
LOG(BER)
Received Power(dBm)
back to back
back to back