PIERS ONLINE, VOL. 3, NO. 7, 2007 1071 Lightwave Technique of mm-wave Generation for Broadband Mobile Communication B. N. Biswas 1, A. Banerjee 1, A. Mukherjee 1, and S. Kar 2 1 Academy of Technology, Hooghly 712121, India 2 Institute of Radio Physics and Electronics, Kolkata 700009, India Abstract In future generation pico-cellular mobile communications FM-sideband injection of two laser sources are used to generate mm-wave signals. This paper considers some overlooked issues with particular emphasis on the loss of coherence between the two lasers due to cycle slipping phenomenon and suggests a new tracking system, that considerably improves the performance. DOI: 10.2529/PIERS061004061503 1. INTRODUCTION Increasing demand for broadband mobile communication and limited atmospheric propagation at mm-waves has resulted in the need for high density of pico-cells. And as such future cellular broadband mobile communication systems will comprise mm-wave components for radio link between the mobile station (MS) and the numerous base stations (BS), which are remotely controlled by the central station (CS). Moreover, the base stations are widely separated from the central station, optical transport of the mm-wave signal is the choice owing to the inherently low transmission loss coefficient of the optical fibers. This, in turn, requires the use of lasers and photo-detectors. The cost of numerous BS s should be kept as low as possible. Therefore, generation and control of mm-wave signals should be optically carried out at the control station, making use of the proposed optical devices needed for the purpose of transport of the mm-wave signals. This avoids the need for mm-wave oscillators and modulators in the numerous base stations. In achieving this purpose two approaches are adopted, viz., (1) single optical source technique and (2) multiple optical source technique. In this we present a method based on multiple optical source technique using sideband separation principle. This is depicted in Fig. 1. Frequency Modulated Master Laser I Frequency Modualted Master Laser Isolater I Photo Diode MW Signal Figure 1: Coherent mixina of two optical sources. The single optical source technique is the simplest approach for impressing microwave signal on an optical carrier. It can be realized either by direct current modulation of semiconductor laser or with an electro-optic modulator (Mach Zehnder Modulator MZM). Direct current modulation
PIERS ONLINE, VOL. 3, NO. 7, 2007 1072 is limited to frequency range below 15.0 GHz and accompanied by a large microwave noise floor due laser intensity noise RIN and large harmonic content due to laser diode non-linearity. Another disadvantage is non-flat frequency response. Although indirect intensity modulation method MZM enjoys the advantage of larger bandwidth over the direct modulation scheme, it suffers from nonlinear response, limited modulation depth, optical insertion loss, cost and complexity. Moreover, the link loss is dependent on the half-wave voltage V π [typically 17+20 log(v π ) db]. Incidentally for a polymer MZM, built at the University of Southern California, V π has a value of 1.2 V at 1310 nm whereas it is 1.8 V at 1550 nm. This indicates that the minimum link loss is on the order of 23.0 db (the laboratory stage). 2. TOWARDS A SOLUTION In order to overcome the limitations of direct and external intensity modulation methods, heterodyning of two frequency-offset-low-intensity-noise lasers in a fast photo-detector has been proposed [1, 2]. A typical version of this concept is depicted in Fig. 2. A DFB laser is modulated and its output consisting of spectral lines separated by the modulating frequency is fed to two optical filters to isolate two sideband components (Alcatel within the EU FRANS Project). After this the two outputs from the two optical filters are heterodyned to generate the required microwave or millimeter wave signal (say, 60.0 GHz). This is possible only when a double sideband signal whose sideband components are separated by at least 50.0 GHz as the presently available optical filters based on Bragg Grating technology has a bandwidth of about 0.2 nm at 1550 nm, i.e., 25.0 GHz. Even then mm-wave signal generated at the output the photo-detector will a large spectral width, as the outputs from the two optical filters are phase incoherent. Moreover, double modulation at 25.0 GHz is difficult to generate. Carrier at (2nf) BASE STATION DFB Laser 1550 nm Double Sideband Generation Optical Filter Bragg Grating or ILO +nf -nf Mod EDFA f Sub-carrier X Modulaton Data CENTRAL STSTION Figure 2: Optical generation of mm-wave signal. This problems and limitations can be overcome by replacing the FG optical filters by injection locked lasers working as narrow band tunable filters. The injection locked optical filters (ILOF) have the following advantages over the Bragg Grating optical filters, being much larger and more complex than ILOF. 1. An injection locked optical filter (ILOF) can have a bandwidth on the order of 300.0 MHz only as compared to 10.0 to 25.0 GHz for a BG optical filter. 2. It is difficult to maintain phase coherence between the signals at the outputs of the two separate BG optical filters. Whereas the outputs of the two injection locked tunable filters (slave lasers SL) are are phase coherent. 3. Since an ILOF is narrow band, the master laser (ML) can be modulated at a much lower microwave frequency making it easier to modulate as well as to isolate the required higher sideband components.
PIERS ONLINE, VOL. 3, NO. 7, 2007 1073 Notwithstanding these possibilities, there still remains an important question that needs to be critically examined, and this is: Is it possible to realize the required phase coherence between the two outputs of the slave lasers with help commercial DFB lasers with linewidth on the order of few MHz (in order to make the system economically viable)? 3. A NEW APPROACH This paper presents a modified optical injection phase frequency locked oscillator (OIPFLO) shown in Fig. 3. It consists of the arrangements by means of which instantaneous frequency of the slave laser is controlled by two mechanisms, namely, through injection locking and optical phase locked principle using an additional arrangement for controlling the output phase of the VCO laser in correspondence to a measure of the instantaneous phase error. The outputs from such two systems are heterodyned to generate the required mm-wave signal (cf. details on such novel system has been given in a companion paper by B N Biswas entitled Optical generation of mm-wave signal with wide linewidth lasers for broadband communications ). M o d u l a t i n g S i g n a l w i t h appropriately adjusted drive voltages (v ) with transfer characteristic T v 2 =1.5V π +xv π cosωt 1 v T = 1+ sin π 2 V π v 1 =0.5V π +xv π cosωt MZ Modulator MZ Modulator Master Laser Optical Signal (E out ) 0.43 0.43 0.114 0.114 Phase Det Filter Isolater Phase Mod VCO Out Figure 3: Injection phase frequency locked oscillator. Phase coherence between the two outputs from the slave lasers means phase coherence between the master laser and the slave laser also, and cycle slipping is an annoying problem. Improving the phase coherence means two things simultaneously, namely, increasing the tracking bandwidth of the locked oscillator and improving the noise squelching property of the system. Unfortunately an attempt to increase one deteriorates the other [3 5]. However, the tracking system as shown in Fig. 3, achieves this property (i.e., to increase the locking range and to decrease the noise bandwidth at the same time) to a great extent. Before we come to the question of phase noises of the master laser and the slave laser, that disturb phase tracking, it is also necessary that there is no unwanted modulation component close to the center frequency of the slave laser. Otherwise, this exerts pulling and pushing force on the slave laser. Thus this will cause serious disturbances in the synchronization process. 4. REDUCTION OF UNWANTED PULLING AND PUSHING EFFECT To solve this problem let us take the following example. Let the modulation bandwidth of the master laser is close to 10.6 Hz, and it is desired to generate 60.0 GHz mm-wave signal. Referring to the new modulation scheme depicted in Fig. 3 and using the drive voltages with
PIERS ONLINE, VOL. 3, NO. 7, 2007 1074 xπ = 3.843 it is observed that the modulated laser field is [α(t) being laser phase noise]. E(t) = 0.43 cos [(ω o + 3ω) t + α] 0.43 cos [(ω o 3ω) t + α] +0.114 cos [(ω o + 5ω) t + α] + 0.114 cos [(ω o 5ω) t + α] (1) The third harmonic components have been picked up, because the output mm-wave signal is required to be of 60.0 GHz with modulating frequency of 10.0 GHz. Disturbing components are away by 20.0 GHz and 6.0 db less in amplitude causing almost no deleterious effect. 5. MINIMIZING FREQUENCY OF SLIPPING CYCLES In the following we neglect the shot noise contribution to the phase error, as it is small over the range of the laser linewidth (say, 10.0 MHz) with the laser power on the order of 100 microwatt and the detector sensitivity of about 0.5 [6]. In view of this noise bandwidth gives an estimate of the phase noise contribution at laser output. In this connection it is important to note that noise output is considered at the output of the laser VCO from where the reference signal is taken and not at the output of the phase modulator after the laser VCO (Fig. 3). Here K i is the injection locking range and K is the open loop gain of the OPLL. T P is the time constant of the phase control loop and T is the loop delay parameter. To demonstrate the validity of the concept in the simplest form, we consider only first order system. In this connection it is important to note that and injection locked oscillator is a first order phase locked system and its locking range K i is limited to about 360.0 Mrad by the stability condition [5]. The loop phase error variance is inversely related to K i. Because of the phase noise, the slave lasers will slip cycles even when the phase error variance is small, and in doing so the phase coherence is lost during the period of slipping cycles. The average time between cycle slips is approximately given by [4, 6]. T av = π [ 2 exp 4B n σ 2 ϕ ], where σ 2 ϕ = ( ) [1 H (j2πf)] 2 f πf 2 df and B n = 1 2π H 1 (j2πf) 2 dω (2) The relation clearly demonstrates that in order to increase T av (a strong function of σ 2 ϕ) it is necessary that B n as well as σ 2 ϕ should be reduced. Incidentally, H(s) denotes the closed loop transfer function and H 1 (s) denotes the transfer function when the output is observed at the VCO output. Let us see how does this concept apply to the PLL system. If one ignores the loop delay then these parameters are given by σ 2 ϕ = π ( α + β) K i + K 1 and B n = K i + K 1 + KT p 4 1 1 + KT p (3) The beauty of inclusion of the phase modulator (i.e., T p ) is that B n as well as σ 2 ϕ decreases with the increase of T p. This is in contrast to the operation conventional OPLL. Where α is FWHM (full width at half maximum) linewidth of the master laser and β is that of the slave laser. Had there been no limitation on the value of K, the summed linewidth can be increased to a large value. But due to presence of the loop delay the maximum permissible value of K is limited. 6. MAXIMUM PERMISSIBLE VALUE OF K For the sake of simplicity we consider the first order system with delay. We write the characteristic equation of the system as jω + b + (1 + jωp) exp ( jωd) = 0 (4) where Ω is the frequency normalized with respect to K, b is the injection locking range normalized with respect to K, T p is the time constant of the phase modulator and d (KT ) is the normalized delay time of the loop. The plot of critical value of KT against b is shown in Fig. 4. An appropriate choice of the phase modulation index p = KT p can improve the value of KT to a large extent. That is, either higher value of K can be used or larger value of loop delay can be accommodated. It is seen that the introduction of the injection signal for direct synchronization and the additional phase control loop, aside directly decreasing the phase error variance and noise bandwidth, it also increases the stability of the tracking system.
PIERS ONLINE, VOL. 3, NO. 7, 2007 1075 Figure 4: Variation of normalised delay with Kinj/Kopll. Figure 5: Variation of phase error variation with loop delay. To illustrate let us consider an OPLL with a loop delay of 1600 ps. Let us choose a value of 2.2 for KT with p = 0.35. Therefore, the permitted value of the K (open loop gain) is 1375.0 Mrad. Thus, b = 0.2618. Hence to realize the same phase error variance for the two cases, namely with p = 0.0 and p = 0.35. The ratio of the summed linewidths of the lasers with and without phase control can be found to 1.545. That is, with the OIPFPLO, 50% more linewidth of the laser can be allowed. When the loop propagation delay is taken into account the noise bandwdith and the phase error variance increase rapidly after a certain values of the loop delay. Simple relations for the noise bandwidth and phase error variance do not hold good. However, when the delay is small, say up to about 700.0 pico seconds, the variation nearly obeys the rule as given in (3). But it is important to observe that by properly adjusting the value of the phase modulation index, both the noise bandwidth and phase error variance can be kept small over a larger value of the loop propagation delay (Fig. 5). It is also interesting to note that the performance of the loop will become worse than that of the simple OPLL if a large value of the phase modulation (p) is taken. ACKNOWLEDGMENT The author is grateful to Professor J. Banerjee, Chairman of the Academy of Technology for his encouragement and financial assistance during the course of the work. REFERENCES 1. Braun, R. P., G. Grossskopf, D. Rhode, and F. Schmidt, Fiber optic millimeter-wave generation and bandwidth efficient data transmission for broadband mobile 18 20 and 60 GHz band communication, IEEE Intl Microwave Photon Topical Meeting, 235 238, Essen, UK, 1997. 2. O Reilly, J. J., P. M. Lane, and M. H. Capstick, Optical generation and delivery of modulated mm-wave for mobile communication, Analog Optical Fiber Communications, 229 256, IEE, 1995. 3. Kroupa, V. F., Phase Lock Loops and Frequency Synthesis, John Wiley, England, 2003. 4. Biswas, B. N., Phase Lock Theories and Applications, Oxford and IBH, New Delhi, 1988. 5. Berdonalli, A. S, C. Walton, and A. J. Seeds, High performance phase locking of wide linewidth semiconductor lasers by combined use of optical injection locking and optical phase locked loop, J. Lightwave Technology, Vol. 17, No. 2, 328 342, February 1999. 6. Ramo, R. T. and A. J. Seeds, Delay, linewidth and bandwidth limitations in optical phase locked loop design, Vol. 26, No. 6, 389 390, 15 March, 1990.