We propose and demonstrate a novel method to improve the long-term stability of the wideband tunable optoelectronic oscillator (OEO) where a dynamic compensation scheme is achieved to offset the parameter variation of the fiber. In our method, a 900 MHz calibration signal transmits in the fiber link of the OEO’s feedback loop for establishing a servo system which can extract the time delay of the feedback loop. The time delay varies with the external environment because refractive index and length of the fiber fluctuate with ambient temperature variations. Taking the extracted information as the reference, the wavelength of the tunable laser used in the OEO can be controlled precisely and continually to offset the random delay fluctuation in the fiber. Consequently, the long-term stability of the microwave signal generated by the OEO can be optimized. The experiment results show that Allan deviation achieved in 1000-s averaging time is improved more than two orders of magnitude when the tunable OEO worked at 2.4 GHz.
© 2015 Optical Society of America
High performance oscillator plays a key role in modern communications, navigation, radar and precise metrology . In the past two decades, the optoelectronic oscillator (OEO) has aroused great interests among the microwave photonics community because of its capability to generate microwave signal with ultra-low phase noise .New structures, such as multi-loop OEO [3, 4], coupled OEO [5–7], and OEOs with photonic filters  have been demonstrated to further improve the phase noise performance and the frequency tunability of the generated signal. The phase noise as low as −163 dBc/Hz at 6 kHz offset frequency from the 10 GHz carrier has been realized . Meanwhile, the frequency tunable range larger than tens of GHz has also been reported . Unfortunately, the free-running OEOs suffer from the frequency drift problem because the long optical fiber using in the OEOs is sensitive to environmental perturbations, which greatly degrade the long-term stability . Injection locking [12, 13] and phase-locked loop  that widely used for frequency synchronization have also been adopted to optimize the long-term stability of the OEOs. In such scheme, a highly stable reference signal provided by the atom clock  or optical comb  is used to stabilize the OEO working at fixed frequency position. Although the frequency stability is improved remarkably, the tunability of the OEO is sacrificed. A phase self-locking OEO was proposed and demonstrated recently , where the phase error in the fiber loop is extracted by a simple phase frequency discriminator. Yet, the frequency drift with the temperature still needs to be further improved. A new architecture is investigated to enhance the long-term stability of OEO , in which a wideband pseudo-random sequence is injected into the feedback loop of the OEO to monitor the long fiber loop delay. In principle, the phase errors occurring in the fiber will be canceled by tuning an electrical driven optical delay line, therefore, a stabilized OEO will be achieved regardless of its operating frequency. However, the feedback scheme is difficult to realize due to the relatively low calculation precision and limited tunable range of the piezoelectric translator (PZT) and the optical delay line. Therefore, no experiment setup based on this scheme has been established yet .
In this study, we introduce a dynamical compensation scheme into the traditional OEO in order to improve the long term stability of the generated microwave signal. In our approach, a low-frequency sinusoidal signal is introduced into the feedback loop to measure the phase variation signal induced by the fiber loop delay variation. The delay variation signal is extracted by the phase discriminator, and then used to compensate the fiber loop delay variation by altering the wavelength of a tunable laser. A wideband tunable OEO with enhanced long-term stability is achieved. Our scheme can greatly improve the monitoring and compensation precision, because the phase discriminator can usually sense the delay variation much more accurately than the time interval counter [18, 19]. With this approach, we obtained a highly stable tunable OEO oscillating from 2.4 GHz to 10 GHz. The long-term stability of 2.4 GHz oscillating signal has reached 5 × 10−9 after 1000-s averaging time, which is two orders of magnitude better than free-running OEO. Besides, the impact of the calibration signal on phase noise performance is also discussed.
Figure 1 shows the schematic diagram of the tunable OEO with the proposed dynamic feedback compensation scheme and the operational principle is described as follows. The proposed scheme contains a traditional single-loop OEO and a feedback compensation loop. Light wave from a wavelength tunable laser (WTL) is sent to a dual-arm Mach-Zehnder modulator (DA-MZM), which is modulated by the self-oscillating signal and the MHz-level sinusoidal calibration signal at each arm. The modulated light carrying these two signals is then launched to a spool of standard single-mode fiber (SMF) and detected by the high-speed photodetector (PD). The low-noise amplifier (LNA) following the PD is used to provide the required open-loop gain to achieve the oscillation condition. Then, the obtained electrical signals are divided into two beams by a 3-dB RF power divider. One of the beams is filtered by a tunable band-pass filter (T-BPF) and used to modulate the DA-MZM to ensure self-oscillation. Meanwhile, the calibration signal in another beam will be selected out by the BPF and compare with the original signal in the phase discriminator to extract the delay variation caused by the fiber link perturbation. Given that the main delay variation in the OEO loop originates from the fiber link , the additional perturbation in the electrical link is ignored. Hence, we can agree that the calibration signal has the same delay variation with the self-oscillating signal due to the fact that they share the same fiber link. Theoretically, the chromatic-dispersion-induced true time delay can be expressed as20]. Due to only several GHz frequency difference between the calibration and oscillating signals, it is worth noting that the high order dispersion induced delay difference is negligible in our scheme . Moreover, for the SMF used in the experiment , the dispersion coefficient D is 17 ps/km/nm at 1550 nm wavelength. The temperature coefficient α of standard telecom SMF28 fiber is as large as 35 ps/km/°C . The tunable range of our WTL is from 1525 to 1568 nm. Therefore, more than 20 °C temperature change can be compensated by the proposed scheme. If we use fibers with higher dispersion coefficient and WTL with larger tunable range, larger temperature change range can be compensated. Another advantage of our proposed scheme is that the frequency tunability of OEO will not be limited, as the compensation loop is independent with the oscillating frequency.
3. Experiment and results
To verify the proposed scheme, an experiment based on the configuration in Fig. 1 is conducted. The parameters of the OEO used in the experiment are as follows: The light source is a computer-controlled WTL with 1-pm tunable step. A DA-MZM (EOspace AZ-DD-1X2) biased at quadrature point, is used to convert oscillating signal and calibration signal from electrical domain to optical domain. A 900-MHz sinusoidal wave with power of −20 dBm is used as calibration signal which is provided by a microwave signal generator (Agilent E8267D). A 5 km spool of SMF is placed after the MZM. The fiber is placed at the open laboratory environment instead of the chamber with constant temperature. A high-speed PD has responsibility of ~0.8 A/W and bandwidth of 20 GHz. Two cascaded LNAs (Marki AP0120EZP) are used to compensate the power loss of all the components in the system. The T-BPF with 3-dB bandwidth of 30 MHz and tunable frequency range from 2 to 12.4 GHz (OMNIYIG M1022) is used to fulfill the tunability of the OEO’s oscillating frequency. The T-BPF will suppress the calibration signal significantly. In the loop of feedback compensation, a BPF (with 3-dB bandwidth of 30 MHz at 900 MHz) is used to filter out the transferred calibration signal. In the experiment, we only set the tunable filter working at 2.4 GHz and 10 GHz to demonstrate the stabilization of the proposed wideband tunable OEO scheme. The performance of generated microwave signal is measured by a frequency counter (Agilent 53230A) and an electrical spectrum analyzer (Agilent N9030A). The OEO and measurement equipment are placed in a laboratory room without temperature control.
In our experiment, we first set the center frequency of the T-BPF at 2.4 GHz. The wavelength of the WLT is fixed at 1550 nm when the feedback compensation loop is open. The delay variation information extracted by the phase discriminator is recorded. Meanwhile, the frequency variation of the oscillating microwave signal is measured by the frequency counter. The two sets of data are recorded once per second synchronously. To avoid the frequency drift between the frequency counter and the calibration signal generator, they are synchronized by a 10 MHz reference, which is provided by the calibration signal generator. Figure 2 shows the comparison between the two variations in a 13,000-second measuring time frame. It is clear that the oscillating frequency (blue curve) varied along with the fluctuation of the fiber delay (red curve). For each delay variation value, the corresponding frequency is almost one-to-one matched, which demonstrates that the frequency drift is mainly introduced by the fiber perturbation. Also, Fig. 2 indicates that the fiber delay variation is as large as 550 ps along with the 3.6 hours temperature change in the room, which is almost beyond the tuning range of commercially available PZT and tunable optical delay line. The corresponding oscillating frequency fluctuation is about 11 ppm (part per million), which is equal to about 25 kHz frequency shift at 2.4 GHz.
When feedback compensation loop is closed, we record these two variations again under the same measured conditions. The measured results are described in Fig. 3. As expected, the frequency variation is within ± 0.1 ppm during the whole 13,000-second recording time (red curve), which indicates that the stability of oscillating frequency is improved. Residual frequency fluctuation after compensation can still be observed, which is believed to be caused by the electrical filter out of the compensation feedback loop.
Allan deviation is commonly used to evaluate the long-term frequency stability. The overlapping Allan deviation calculated from the oscillating frequency measurement is shown in Fig. 4. In the free-running OEO, the frequency stability is 3.4 × 10−9 at 1-s averaging time (blue curve). Due to the thermal drift, the curve monotonically increases to 10−6 after 1000-s averaging time. When the compensation is on, the frequency stability is 2 × 10−9 at 1-s averaging time and slowly increases to 5 × 10−9 after 1000-s averaging time (red curve). More than two orders of magnitude improvement is achieved. The 10−9 level long-term stability is achieved without using the bulky thermal-controlled chamber, which is greatly needed for the practical applications such as satellite system. In fact, long-term stability can be further improved if a temperature-insensitive filter with ceramic technology substitutes for the T-BPF is used in the experiment .
The phase noise curves of the 2.4 GHz oscillating microwave signal with and without the compensation are also measured by the electrical spectrum analyzer (See Fig. 5). It can be seen that the phase noise performance has no conspicuous deterioration. Less than −120 dBc/Hz at 10 kHz offset frequency from 2.4 GHz carrier in both cases is obtained respectively. These results prove that the short-term stability of the generated microwave signal does not deteriorate while its long-term stability is improved by our proposed compensation mechanism.
In order to evaluate the wideband compensation ability of the proposed OEO, we tune the center frequency of the T-BPF to 10 GHz. In this case, the performance of the generated microwave signal is measured and analyzed. Due to the bandwidth limitation of the frequency counter (maximum bandwidth: 6 GHz), the Allan deviation of the generated microwave signal cannot be obtained. Therefore, we use an electrical spectrum analyzer working at “max-hold” mode with 50 kHz frequency span to measure the frequency drift for 10 minutes. The frequency drift of generated microwave signal at 10 GHz with or without the compensation is measured. Both curves are illustrated in Fig. 6. It can be seen that the frequency spectrum of the free running signal is about 10 kHz drift (blue curve). In contrast, the microwave signal keeps stable when the compensation loop is closed (red curve). Such experiment results confirm our compensation scheme is suitable for a wideband-tunable OEO, which is greatly needed in radar and wideband communication system.
In conclusion, we introduce a dynamical compensation scheme into a traditional OEO in order to improve the long term stability of the generated microwave signal. A low-frequency sinusoidal signal is added into the feedback loop to measure the phase variation signal induced by the fiber loop delay variation. The delay variation information is extracted by the phase discriminator, and then used to compensate the fiber loop delay variation by altering the wavelength of a tunable laser. Based on this approach, a wideband tunable OEO with enhanced long-term stability was achieved. We obtained a highly stable tunable OEO oscillating from 2.4 GHz to 10 GHz. The long-term stability of 2.4 GHz oscillating signal has reached 5 × 10−9 after 1000-s averaging time, which is two orders of magnitude better than the free-running OEO. The experiment results show that the proposed OEO is capable to keep good short-term and long-term stabilities simultaneously.
This work was supported in part by NSFC Program (61302016,61431003,61271042 and 61335002) and NCET-13-0682.
References and links
1. X. S. Yao, L. Maleki, and D. Eliyahu, “Progress in the opto-electronic oscillator - a ten year anniversary review,” in Proceedings of IEEE MTT-S International Microwave Symposium Digest (Fort Worth, TX, 2004), pp. 287–290. [CrossRef]
2. X. S. Yao and L. Maleki, “Optoelectronic oscillator for photonic systems,” IEEE J. Quantum Electron. 32(7), 1141–1149 (1996). [CrossRef]
3. X. S. Yao and L. Maleki, “Multiloop optoelectronic oscillator,” IEEE J. Quantum Electron. 36(1), 79–84 (2000). [CrossRef]
4. J. Xiong, R. Wang, T. Fang, T. Pu, D. Chen, L. Lu, P. Xiang, J. Zheng, and J. Zhao, “Low-cost and wideband frequency tunable optoelectronic oscillator based on a directly modulated distributed feedback semiconductor laser,” Opt. Lett. 38(20), 4128–4130 (2013). [CrossRef] [PubMed]
6. E. Salik, N. Yu, and L. Maleki, “An ultralow phase noise coupled optoelectronic oscillator,” IEEE Photon. Technol. Lett. 19(6), 444–446 (2007). [CrossRef]
7. A. B. Matsko, D. Eliyahu, and L. Maleki, “Theory of coupled optoelectronic microwave oscillator II: phase noise,” J. Opt. Soc. Am. B 30(12), 3316–3323 (2013). [CrossRef]
8. F. Jiang, J. H. Wong, H. Q. Lam, J. Zhou, S. Aditya, P. H. Lim, K. E. Lee, P. P. Shum, and X. Zhang, “An optically tunable wideband optoelectronic oscillator based on a bandpass microwave photonic filter,” Opt. Express 21(14), 16381–16389 (2013). [CrossRef] [PubMed]
9. D. Eliyahu, D. Seidel, and L. Maleki, “Phase noise of a high performance OEO and an ultra low noise floor cross-correlation microwave photonic homodyne system,” in Proceedings of 2002 IEEE International Frequency Control Symposium (2008), pp. 811–814. [CrossRef]
10. X. Xie, C. Zhang, T. Sun, P. Guo, X. Zhu, L. Zhu, W. Hu, and Z. Chen, “Wideband tunable optoelectronic oscillator based on a phase modulator and a tunable optical filter,” Opt. Lett. 38(5), 655–657 (2013). [CrossRef] [PubMed]
11. D. Eliyahu, K. Sariri, M. Kamran, and M. Tokhmakhian, “Improving short and long term frequency stability of the opto-electronic oscillator,” in Proceedings of 2002 IEEE International Frequency Control Symposium (2002), pp. 580–583. [CrossRef]
12. C. Williams, J. Davila-Rodriguez, D. Mandridis, and P. J. Delfyett, “Noise characterization of an injection-locked COEO with long-term stabilization,” J. Lightwave Technol. 29(19), 2906–2912 (2011). [CrossRef]
14. Y. L. Zhang, D. Hou, and J. Y. Zhao, “Long-term frequency stabilization of an optoelectronic oscillator using phase-locked loop,” J. Lightwave Technol. 32(13), 2408–2414 (2014). [CrossRef]
15. D. Hou, X. P. Xie, Y. L. Zhang, J. T. Wu, Z. Y. Chen, and J. Y. Zhao, “Highly stable wideband microwave extraction by synchronizing widely tunable optoelectronic oscillator with optical frequency comb,” Sci. Rep. 3, 3509 (2013). [CrossRef] [PubMed]
16. L. Bogataj, M. Vidmar, and B. Batagelj, “A feedback control loop for frequency stabilization in an opto-electronic oscillator,” J. Lightwave Technol. 32(20), 3690–3694 (2014). [CrossRef]
17. W. H. Tseng and K. M. Feng, “Enhancing long-term stability of the optoelectronic oscillator with a probe-injected fiber delay monitoring mechanism,” Opt. Express 20(2), 1597–1607 (2012). [CrossRef] [PubMed]
19. A. Zhang, Y. Dai, F. Yin, T. Ren, K. Xu, J. Li, and G. Tang, “Phase stabilized downlink transmission for wideband radio frequency signal via optical fiber link,” Opt. Express 22(18), 21560–21566 (2014). [PubMed]
20. A. Zhang, Y. Dai, F. Yin, T. Ren, K. Xu, J. Li, Y. Ji, J. Lin, and G. Tang, “Stable radio-frequency delivery by λ dispersion-induced optical tunable delay,” Opt. Lett. 38(14), 2419–2421 (2013). [CrossRef] [PubMed]
22. F. Narbonneau, M. Lours, S. Bize, A. Clairon, G. Santarelli, O. Lopez, C. Daussy, A. Amy-Klein, and C. Chardonnet, “High resolution frequency standard dissemination via optical fiber metropolitan network,” Rev. Sci. Instrum. 77(6), 064701 (2006). [CrossRef]