Open Access
Add to BibSonomy
Abstract
In this paper, we demonstrate a cost-effective solution for fiber-optic frequency transfer. By employing the commercially available small form pluggable (SFP) transceivers and other components for telecom, the scheme is compatible with the existing telecom networks. An experiment testbed based on the common round trip frequency transfer is carried out to investigate the corresponding performance in detail. Compared with the conventional analog electro-optic interconversion scheme, the optical modulation and demodulation employing SFP transceiver has an inferior performance in terms of additive phase noise. However, the scheme enables a much larger operation range of receiving optical power for an optimal system performance, which indicates the certain capability of adapting to different fiber transmission links. On the basis of the established testbed, fiber-optic frequency transfer with a frequency of 1 GHz is demonstrated over a 50 km dispersion-compensated fiber link. Furthermore, the additive phase noise of ${-}{{84}}\;{{\rm dBc}/{\rm Hz}}$ @1 Hz and ${-}{{130}}\;{{\rm dBc}/{\rm Hz}}$ @10 kHz is achieved. Simultaneously, the Allan deviation (ADEV) of ${9.6} \times {{1}}{{{0}}^{- 14}}/{\rm s}$ and ${8.4} \times {{1}}{{{0}}^{- 17}}/{{{10}}^4}\;{\rm s}$ is reached, respectively. The scheme can provide a relatively feasible solution to implement fiber-optic frequency transfer over the existing telecom network infrastructure.
© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. INTRODUCTION
Atomic clocks have been instrumental in social and scientific developments, leading to innovations such as navigation, very long baseline interferometry (VLBI), and radio astronomy [1,2]. Owing to the high cost of advanced atomic clocks and the demanding need of clock synchronization, frequency transfer is rather indispensable. During the past decade, optical fiber has been considered to be a superior medium to the satellite for its outstanding propagation characteristics in terms of low loss, high reliability, excellent electromagnetic interference (EMI), and so on [3]. Hence, fiber-optic frequency transfer has been intensively investigated to seek for the most accurate and stable frequency dissemination method [2,4–12]. Suffering from temperature variation, mechanical vibration [13,14], etc., the performance of a directly transmitted frequency signal will be seriously deteriorated. In general, two-way or round trip method is adopted to set stringent limit on a possible violation of the fiber link propagation delay fluctuation. As for a certain field implementation, a dedicated dark fiber or channel is applied to guarantee the bidirectional transmission in most demonstration experiments. However, the expense is too high to be accepted for large-scale applications.
At present, the commercial fiber network infrastructure for telecom communications is quite mature and spread all over the world. Taking into consideration the cost, it is highly anticipated to implement frequency transfer over the active telecom fiber networks. However, the telecom communication networks are not compatible with the fiber-optic frequency transfer. There are several types of components installed [15] that impose restrictions on the practical application. One is employed for electro-optic interconversions, which is dedicatedly designed for digital signal communication rather than analog frequency transfer, such as the small form pluggable (SFP) transceiver. It is the key issue to be solved and evaluated in this paper. The second is the signal regeneration between each fiber span due to transmission attenuation. The inline optical amplifiers in the telecom communication networks feature unidirectional transmission. The last is the unidirectional optical devices, which are designed for path routing, such as the add/drop multiplexer.
In this paper, we demonstrate a cost-effective solution for fiber-optic frequency transfer over the existing telecom network infrastructure. To achieve the best performance as much as possible, the frequency signal is transmitted to the remote users directly, instead of clock discipline or data recovery [16]. An experiment testbed employing a commercially available SFP transceiver is carried out to investigate the performance in detail. As for the performance deterioration, the additive phase noise of optical modulation and demodulation is compared between the SFP transceiver and the conventional analog conversion scheme. Via the evaluation testbed, the system performance with different receiving optical powers is studied. Furthermore, 50 km fiber-optic frequency transfer with the frequency of 1 GHz is demonstrated with and without dispersion compensation fiber, respectively.
This paper is organized as follows: Section 2 presents the principle of fiber-optic frequency transfer employing SFP transceiver. Section 3 gives the experimental system structure. Via the experimental system, different kinds of characteristics are investigated and presented in detail. Section 4 draws the conclusion.
2. PRINCIPLE
To mitigate the influence from the ambient environment, the common round trip frequency transfer method is adopted. From the perspective of engineering, electrical phase shifters (PSs) are used as fiber delay compensators. Figure 1 schematically illustrates the principle of the fiber-optic frequency transfer scheme employing SFP transceiver. The RF signal to be transmitted at the local station can be denoted as
where ${\omega _0}$ and ${\varphi _0}$ represent the angular frequency and initial phase, respectively.
Fig. 1. Schematic structure of fiber-optic frequency transfer employing small form pluggable (SFP) transceiver. CU, control unit; PD, phase detector; PS, phase shifter; EPS, electrical power splitter; BPF, bandpass filter; WDM, wavelength division multiplexer.
After passing through a PS, the analog RF signal is modulated on an optical carrier of ${\lambda _1}$ with the help of the SFP transceiver. Although designed for digital communication, the SFP transceiver can still convert the analog RF signal to the optical domain according to a certain signal level. Then the optical carrier is launched into the fiber link via a wavelength division multiplexer (WDM). The RF signal sent from the local station can be given as
where ${t_1}$ stands for the transmitting delay introduced by the SFP transceiver and WDM in the local station, and ${\varphi _1}(t)$ signifies the additive phase induced by the PS.At the remote station, the RF signal is picked out by another WDM and recovered back by another SFP transceiver. For a sine wave signal, an electrical bandpass filter (BPF) is applied on the output square wave signal, which results from the clipping effect due to the limiting amplifier in RX-part. Then, the acquired frequency signal can be expressed as
To sense the fiber propagation delay variation, part of the received frequency signal is fed back to the local station with another wavelength of ${\lambda _2}$. Similarly, the round trip RF signal, going through another PS at the local station, can be shown as
With the aid of a phase detector, an error signal of ${V_5}$ can be obtained to drive the PSs at the local station. The assumption that the variations of forward and backward propagation delays, $\Delta {t_{{fi}}}$ ($i = {{1}}$, 2), are identical should be applied. For a long time operation, the variation difference of bidirectional wavelengths from two SFP transceivers has been observed to be within 1 pm [17]. Over a standard 50 km G.652 fiber link with a dispersion coefficient of ${{17}}\;{{\rm ps}}/({{\rm km} \cdot {\rm nm}})$, the value of $\Delta {t_{{fi}}}$ ($i = {{1}}$, 2) can fluctuate in the range of 0.5 ps in practice. Further, for a fiber dispersion compensation with a residual link dispersion of less than 3 ps/nm, the fluctuation of bidirectional propagation delay difference will be even within 2 fs. On the other hand, the sending or receiving delay variations between both stations can be neglected as usual, that is to say, $\Delta {t_i}$ $(i = {{1}},{{2}},{{3}},{{4}}) = {{0}}$. Therefore, once the value of ${V_5}$ keeps to a constant, the achieved RF signal at the remote station is a perfect replica of the transmitted one, which can be reformulated as Eq. (6),
3. EXPERIMENT AND RESULTS
We construct an evaluation testbed in a normal laboratory, which has a diurnal temperature variation of about 3°C. As shown in Fig. 2, the RF signal from a signal generator AV1464A at the local station is split into two branches. One branch is carried on an optical carrier by an off-the-shelf SFP transceiver with the features of 2.5 Gb/s and 120 km. The inset distributed feedback (DFB) laser emits an optical power of 3 dBm with a central wavelength of 1550.12 nm. Characterized by an optical wavelength meter, the 20 dB spectra width is less than 0.3 nm. After a forward fiber transmission, the RF signal is recovered back by another SFP transceiver and a BPF with a bandwidth of less than 50 MHz at the remote station. To block the effects from the ambient environment, half of the receiving frequency signal is directed back to the local station over another wavelength of 1549.32 nm. At the local station, the round trip frequency signal is compared with the other branch from the frequency source in a phase detector to follow up the fiber link propagation delay variation. Accordingly, two commercially electrical PSs are applied for forward and backward, respectively, to stabilize the fiber link by a control unit (CU). The CU is a field programmable gate array (FPGA)-based hardware module with a full proportion-integration-differentiation (PID) controller. For both PSs, the resolution is below 0.01º, and the operation range is beyond 420º. And the control bandwidth is several Hz. When the range limit is reached, the PS is stepped by 360º. At both stations, commercial electrical amplifiers (EAs) are adopted to form usable RF levels. And a pair of WDMs with a channel spacing of 0.8 nm is employed to separate and combines the transmitted wavelengths back and forth. To mitigate the optical reflection, angled physical contact (APC) connectors are adopted in the experiment.
Fig. 2. Experimental testbed of fiber-optic frequency transfer employing SFP transceiver. CU, control unit; PD, phase detector; PS, phase shifter; EPS, electrical power splitter; EA, electrical amplifier; BPF, bandpass filter; WDM, wavelength division multiplexer.
Fig. 3. Fiber-optic transmission link employing (a) SFP transceiver or (b) conventional analog scheme (incorporating with electro-optic modulator and photodetector). DMLD, directly modulated laser diode; LD, laser; EOM, electro-optic modulator; OPD, photodetector.
Fig. 4. Performance comparison in terms of additive phase noise for fiber-optic transmission link employing SFP transceiver or conventional analog scheme. The transmitted frequency is (a) 100 MHz and (b) 1 GHz, respectively.
Figure 4 gives the performance comparison of fiber-optic transmission link with the SFP transceiver and conventional analog scheme (see Fig. 3). The additive phase noise is measured by a signal source analyzer. The optical light from the transmitter is directly fed into the corresponding receiver with a power as high as possible. With a 100 MHz frequency signal [see Fig. 4(a)], the observed additive phase noises are similar for the offset frequency varying from 100 Hz to 10 kHz. Nevertheless, the conventional analog scheme, which incorporates with a 40 mW DFB laser, a 10 GHz electro-optic modulator (EOM), and a 10 GHz photodetector (OPD), still achieves a better performance to some extent for other offset frequencies. Still, the additive phase noises of ${-}{{110}}\;{{\rm dBc}/{\rm Hz}}$ @1 Hz and ${-}{{153}}\;{{\rm dBc}/{\rm Hz}}$ @10 kHz are achieved for the SFP transceiver. Figure 4(b) shows the measured results for a 1 GHz signal. The additive phase noise for conventional analog scheme is almost always better than the one of the SFP transceiver. It can be attributed that the frequency of 1 GHz is close to the permissible upper frequency limit of the used 2.5 Gb/s SFP transceiver. For a 1 GHz signal transmission, the additive phase noises of ${-}{{94}}\;{{\rm dBc}/{\rm Hz}}$ @1 Hz and ${-}{{145}}\;{{\rm dBc}/{\rm Hz}}$ @10 kHz are achieved.
Based on the experimental testbed (illustrated in Fig. 2), fiber-optic frequency transfer is implemented with the frequency of 1 GHz. To avoid the performance deterioration from the fiber transmission link [18], the higher frequency of 1 GHz is chosen, rather than 100 MHz in the experiment. When both stations are connected by a 2 m optical fiber patch cord, the measured system results in terms of additive phase noise for different receiving optical powers are depicted in Fig. 5. They can be regarded as the system noise floor, which is mainly determined by the electro-optic interconversions by SFP transceivers and other electrical devices. When the optical power of ${-}{15}\;{\rm nm}$ is applied, the additive phase noise of ${-}{{88}}\;{{\rm dBc}/{\rm Hz}}$ @1 Hz and ${-}{{140}}\;{{\rm dBc}/{\rm Hz}}$ @10 kHz is observed. As the optical power increases from ${-}{25}\;{\rm nm}$ to ${-}{15}\;{\rm nm}$, the receiving signal-to-noise ratio (SNR) or additive phase noise gets a substantial performance enhancement. However, when the optical power keeps going up to ${-}{5}\;{\rm nm}$, a small optimization is observed, which may be operated in a saturation mode. It will be investigated further by our next work. At the same time, it is noteworthy that a much larger range of receiving optical power can be enabled for an optimal system operation, which is quite different from the conventional analog scheme. Hence, the scheme can support different fiber transmission links with a minimum parameter adjustment or component addition.
Fig. 5. Additive phase noise for the fiber-optic frequency transfer testbed when different optical powers are applied. The fiber connecting both stations is 2 m and the transmitted frequency is 1 GHz.
Figure 6 plots the results for the fiber-optic frequency transfer over 50 km spooled fiber transmission links. And the transmitted frequency is 1 GHz. In contrast with the short fiber link, a significant additive phase noise deterioration is acquired for the transmission link without fiber dispersion compensation. Nevertheless, the additive phase noise is optimized to ${-}{{84}}\;{{\rm dBc}/{\rm Hz}}$ @1 Hz and ${-}{{130}}\;{{\rm dBc}/{\rm Hz}}$ @10 kHz via fiber dispersion compensation, which features the residual link dispersion coefficient of less than 3 ps/nm. It reveals that the wide bandwidth of the emitted optical light from the SFP transceiver strongly affects the receiving SNR by fiber dispersion [19,20].
Fig. 6. Additive phase noise of the fiber-optic frequency transfer over 50 km transmission links. The transmitted frequency is 1 GHz. The receiving optical power is ${-}{15}\;{\rm nm}$. Dispersion-compensated fiber is employed for dispersion compensation.
For a long time operation, the Allan deviation (ADEV) is presented in Fig. 7. The transmitted frequency is 1 GHz. For a better performance, dispersion compensation is employed for long-haul transmission. For a 2 m fiber link, the ADEV is ${7.7} \times {{1}}{{{0}}^{- 14}}/{\rm s}$ and $1.2 \times {10^{- 17}}/{10^4}$ s. It is mainly dominated by the adopted electrical and optical devices for RF signal processing and optical modulation/detection at the two stations. With active stabilization, the ADEV of ${9.6} \times {{1}}{{{0}}^{- 14}}/{\rm s}$ and ${8.4} \times {{1}}{{{0}}^{- 17}}/{{1}}{{{0}}^4}\;{\rm s}$ is achieved over a 50 km fiber dispersion-compensated link, which is significantly improved compared to the ones without stabilization. However, in contrast to the stability floor of 2 m, the evident performance aggravation beyond the averaging time of 200 s is attributed to the residual bidirectional propagation delay variation induced from the wavelength fluctuation, etc., which may be strongly affected by the air conditioning.
Fig. 7. Measured frequency stability in terms of Allan deviation (ADEV) for fiber-optic frequency transfer employing SFP transceivers. The transmitted frequency is 1 GHz.
It should be pointed out that the SFP transceiver is commonly applied in signal electro-optic interconversion and is compatible with the telecom communication networks [21]. Table 1 gives the performance comparison of fiber-optic frequency transfer employing commercially available telecom components. It could be observed that the demonstrated fiber-optic frequency transfer scheme features certain application advantage.
Table 1. Performance Comparison of Fiber-Optic Frequency Transfer Employing Commercially Available Telecom Components
To exploit the existing telecom communication networks for fiber-optic frequency transfer to a great extent, the inline unidirectional devices should be replaced or reconstructed with the same function over long haul. With the balance of noise elimination and delay symmetry maintenance, various kinds of bidirectional amplification schemes for signal regeneration have been proposed and demonstrated up to now [22–25], especially the low-cost ones with only passive component modification (as shown in Fig. 8) [24,25].
Fig. 8. Bidirectional amplification schemes with only passive components modification. EDFA, erbium-doped fiber amplifier; FBG, fiber Bragg grating.
4. CONCLUSION
In summary, we demonstrate a cost-effective fiber-optic frequency transfer scheme. An experiment testbed of round trip frequency transfer is carried out to investigate the performance in detail. The 1 GHz frequency transfer is demonstrated over a 50 km dispersion-compensated fiber link. The additive phase noise of ${-}{{84}}\;{{\rm dBc}/{\rm Hz}}$ @1 Hz and ${-}{{130}}\;{{\rm dBc}/{\rm Hz}}$ @10 kHz is achieved. Moreover, the ADEV of ${9.6} \times {{1}}{{{0}}^{- 14}}/{\rm s}$ and ${7.1} \times {{1}}{{{0}}^{- 17}}/{{1}}{{{0}}^4}\;{\rm s}$ is realized, respectively. The attained results imply that the scheme possesses the potential of implementing high-performance fiber-optic frequency transfer over the existing telecom communication network infrastructure with a reasonable cost.
Disclosures
The author declares no conflicts of interest.
REFERENCES
1. H. Marion, F. Pereira Dos Santos, M. Abgrall, S. Zhang, Y. Sortais, S. Bize, I. Maksimovic, D. Calonico, J. Grünert, C. Mandache, P. Lemonde, G. Santarelli, P. Laurent, A. Clairon, and C. Salomon, “Search for variations of fundamental constants using atomic fountain clocks,” Phys. Rev. Lett. 90, 150801 (2003). [CrossRef]
2. K. Predehl, G. Grosche, S. Raupach, S. Droste, O. Terra, J. Alnis, T. Legero, T. Hänsch, T. Udem, and R. Holzwarth, “A 920-kilometer optical fiber link for frequency metrology at the 19th decimal place,” Science 336, 441–444 (2012). [CrossRef]
3. D. Piester and H. Schnatz, “Novel techniques for remote time and frequency comparisons,” PTB-Mitteilungen 119, 33–44 (2009).
4. C. Clivati, A. Tampellini, A. Mura, F. Levi, G. Marra, P. Galea, A. Xuereb, and D. Calonico, “Optical frequency transfer over submarine fiber links,” Optica 5, 893–901 (2018). [CrossRef]
5. C. Daussy, O. Lopez, A. Amy-Klein, A. Goncharov, M. Guinet, C. Chardonnet, F. Narbonneau, M. Lours, D. Chambon, and S. Bize, “Long-distance frequency dissemination with a resolution of 1E-17,” Phys. Rev. Lett. 94, 203904 (2005). [CrossRef]
6. Ł. Śliwczyński, P. Krehlik, A. Czubla, Ł. Buczek, and M. Lipiński, “Dissemination of time and RF frequency via a stabilized fibre optic link over a distance of 420 km,” Metrologia 50, 133–145 (2013). [CrossRef]
7. F. Yin, A. Zhang, Y. Dai, T. Ren, K. Xu, J. Li, J. Lin, and G. Tang, “Phase-conjugation-based fast RF phase stabilization for fiber delivery,” Opt. Express 22, 878–884 (2014). [CrossRef]
8. M. Fujieda, M. Kumagai, S. Nagano, A. Yamaguchi, H. Hachisu, and T. Ido, “All optical link for direct comparison of distant optical clocks,” Opt. Express 19, 16498–16507 (2011). [CrossRef]
9. M. T. Hsu, Y. He, D. A. Shaddock, R. B. Warrington, and M. B. Gray, “All-digital radio-frequency signal distribution via optical fibers,” IEEE Photon. Technol. Lett. 24, 1015–1017 (2012). [CrossRef]
10. M. Amemiya, M. Imae, Y. Fujii, T. Suzuyama, and S.-I. Ohshima, “Simple time and frequency dissemination method using optical fiber network,” IEEE Trans. Instrum. Meas. 57, 878–883 (2008). [CrossRef]
11. J. Wei, F. Zhang, Y. Zhou, D. Ben, and S. Pan, “Stable fiber delivery of radio-frequency signal based on passive phase correction,” Opt. Lett. 39, 3360–3362 (2014). [CrossRef]
12. B. Wang, C. Gao, W. Chen, J. Miao, X. Zhu, Y. Bai, J. Zhang, Y. Feng, T. Li, and L. Wang, “Precise and continuous time and frequency synchronisation at the 5 × 10-19 accuracy level,” Sci. Rep. 2, 556 (2012). [CrossRef]
13. T. J. Pinkert, O. Böll, L. Willmann, G. S. M. Jansen, E. A. Dijck, B. G. H. M. Groeneveld, R. Smets, F. C. Bosveld, W. Ubachs, K. Jungmann, K. S. E. Eikema, and J. C. J. Koelemeij, “Effect of soil temperature on optical frequency transfer through unidirectional dense-wavelength-division-multiplexing fiber-optic links,” Appl. Opt. 54, 728–738 (2015). [CrossRef]
14. N. Ashby, D. A. Howe, J. Taylor, A. Hati, and C. Nelson, “Optical fiber vibration and acceleration model,” in Joint IEEE International Frequency Control Symposium with the 21st European Frequency and Time Forum (2007), pp. 547–551.
15. A. Norouzi, A. H. Zaim, and B. B. Ustundag, “An integrated survey in optical networks: concepts, components and problems,” International Journal of Computer Science and Network Security 11 , 10–26 (2011).
16. E. F. Dierikx, A. E. Wallin, T. Fordell, J. Myyry, P. Koponen, M. Merimaa, T. J. Pinkert, J. C. J. Koelemeij, H. Z. Peek, and R. Smets, “White rabbit precision time protocol on long-distance fiber links,” IEEE Trans. Ultrason. Ferroelectr. Freq. Control 63, 945–952 (2016). [CrossRef]
17. H. Zhang, G. Wu, L. Hu, X. Li, and J. Chen, “High-precision time transfer over 2000-km fiber link,” IEEE Photon. J. 7, 1–9 (2015). [CrossRef]
18. S. M. Foreman, K. W. Holman, D. D. Hudson, D. J. Jones, and J. Ye, “Remote transfer of ultrastable frequency references via fiber networks,” Rev. Sci. Instrum. 78, 021101 (2007). [CrossRef]
19. R. Bouchand, D. Nicolodi, X. Xie, C. Alexandre, and Y. Le Coq, “Accurate control of optoelectronic amplitude to phase noise conversion in photodetection of ultra-fast optical pulses,” Opt. Express 25, 12268–12281 (2017). [CrossRef]
20. S. Yamamoto, N. Edagawa, H. Taga, Y. Yoshida, and H. Wakabayashi, “Analysis of laser phase noise to intensity noise conversion by chromatic dispersion in intensity modulation and direct detection optical-fiber transmission,” J. Lightwave Technol. 8, 1716–1722 (1990). [CrossRef]
21. H. Zhang, G. Wu, X. Li, and J. Chen, “High-precision joint time and frequency transfer over a fiber-optic telecom testbed,” arXiv:1707.06327 (2017).
22. Ł. Śliwczyński and J. Kołodziej, “Bidirectional optical amplification in long-distance two-way fiber-optic time and frequency transfer systems,” IEEE Trans. Instrum. Meas. 62, 253–262 (2013). [CrossRef]
23. M. Amemiya, M. Imae, Y. Fujii, T. Suzuyama, and S.-I. Ohshima, “Time and frequency transfer and dissemination methods using optical fiber network,” in Proceedings of the IEEE International Frequency Control Symposium and Exposition (IEEE, 2005), pp. 914–918.
24. X. Yi, “Optical amplification method and apparatus for single fiber bidirectional transmission,” Chinese patent 02124992.X (26 February 2002).
25. H. Zhang, “Bidirectional optical amplification apparatus based on wavelength division multiplexing transmission,” Chinese patent 201910890155.X (20 September 2019).
26. T. Jurij, P. Leon, B. Bostjan, L. Primoz, R. Patrik, F. Mario, and V. Matjaz, “Fiber length compensated transmission of 2998.01 MHz RF signal with femtosecond precision,” Microwave Opt. Technol. Lett. 53, 1553–1555 (2011). [CrossRef]
