30-Gb/s 3 x 3 optical mode group-division-multiplexin system
with optimized joint detection
Citation for published version (APA):
Chen, H., Boom, van den, H. P. A., & Koonen, A. M. J. (2011). 30-Gb/s 3 x 3 optical mode group-division-multiplexin system with optimized joint detection. IEEE Photonics Technology Letters, 23(18), 1283-1285. https://doi.org/10.1109/LPT.2011.2158639
DOI:
10.1109/LPT.2011.2158639 Document status and date: Published: 01/01/2011
Document Version:
Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)
Please check the document version of this publication:
• A submitted manuscript is the version of the article upon submission and before peer-review. There can be important differences between the submitted version and the official published version of record. People interested in the research are advised to contact the author for the final version of the publication, or visit the DOI to the publisher's website.
• The final author version and the galley proof are versions of the publication after peer review.
• The final published version features the final layout of the paper including the volume, issue and page numbers.
Link to publication
General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain
• You may freely distribute the URL identifying the publication in the public portal.
If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please follow below link for the End User Agreement:
www.tue.nl/taverne
Take down policy
If you believe that this document breaches copyright please contact us at:
openaccess@tue.nl
providing details and we will investigate your claim.
IEEE PHOTONICS TECHNOLOGY LETTERS, VOL. 23, NO. 18, SEPTEMBER 15, 2011 1283
30-Gb/s 3
3 Optical Mode
Group-Division-Multiplexing System With Optimized Joint Detection
H. S. Chen, Student Member, IEEE, H. P. A. van den Boom, and A. M. J. Koonen, Fellow, IEEE
Abstract—This letter investigates a 30-Gb/s 3 3 optical mode group-division-multiplexing (MGDM) system over graded index-multimode fiber (GI-MMF). With mode-selective spatial filtering (MSSF), the optimized joint detection has been realized, which de-creases the complexity of the signal processing to recover received signals.
Index Terms—Mode group-division multiplexing (MGDM),
mode-selective spatial filtering (MSSF), multimode fiber (MMF), optical communication.
I. INTRODUCTION
F
ROM an information theory viewpoint, multimode fiber (MMF) has a larger information transport capacity than single-mode fiber (SMF). The capacity of each mode guided in an MMF basically equals that of an SMF. In the way MMFs are typically used up to now, all these guided modes are excited simultaneously by one light source carrying all the data infor-mation. The different propagation speeds of the guided modes, however, cause multimode dispersion of the optical signal pulses, and thus severely decrease the bandwidth of the fiber link. A considerable bandwidth improvement can be realized by exciting a subset of modes. Within this subset, propagation constants are close together so that modal dispersion is smaller [1].With this mode-selective launching, one may partition the whole set of guided modes in many subsets and launch each of them separately. Thus many parallel transmission paths, each with an improved bandwidth, can be created in a MMF, which coined Mode Group Division Multiplexing (MGDM; see [2]), can drastically multiply the capacity of the fiber.
In this letter, Mode-Selective Spatial Filtering (MSSF) is utilized to decrease the crosstalk among mode groups which decreases the complexity of signal recovery. With the aid of MSSF, a 30 Gb/s 3 3 optical MGDM system is experimen-tally demonstrated with optimized joint detection.
II. PRINCIPLES
A. MGDM Model
For a 3 3 MGDM system, shown in Fig. 1, let 3 1 vector represent 3 electrical signals that modulate the power of the 3 optical light sources. In the MMF, 3 optical signals
prop-Manuscript received February 03, 2011; revised May 11, 2011; accepted May 28, 2011. Date of publication June 07, 2011; date of current version August 19, 2011. This work was supported in part by the European Commission in the FP7 project MODE-GAP.
The authors are with the COBRA Institute, Eindhoven University of Technology, NL-5600MB Eindhoven, The Netherlands (e-mail: haoshuo.chen@gmail.com; h.p.a.v.d.boom@tue.nl; A.M.J.Koonen@tue.nl).
Color versions of one or more of the figures in this letter are available online at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/LPT.2011.2158639
Fig. 1. 3 3 MGDM system.
Fig. 2. Transfer matrix for a 3 3 MGDM system (a) without MSSF and (b) with MSSF.
agate through 3 different mode groups. At the receiving side, each spatially selective receiver detects a mixture of all mode groups. The 3 1 vector represents the 3 output electrical signals after photo-detection and can be written as:
(1) where is a 3 1 additive noise vector. Entries
of transfer matrix , see Fig. 2(a) describes the signal transfer from transmitter and to receiver . Besides the diag-onal items, all other items in mean crosstalk. To recover the received signals, estimated by sending some training symbols is inverted. Zero-Forcing (ZF) is applied to do joint detection.
B. Mode-Selective Spatial Filtering (MSSF)
MSSF utilizing the numerical aperture (NA) of the lens to mitigate the crosstalk in MGDM channels was proposed in [3]. However, an additional multisegmented PD is needed in this case. With the help of the pigtail of the PD, the MSSF function is realized in [4], where three rays corresponding to different mode groups are launched by three transmitters with different radial offset, see Fig. 3. The angle of the ray which denotes one mode group in a MMF is becoming larger, when the ray is closer to the MMF’s axis. If the NA of the fiber pigtail of the PD is smaller than that of the output optical ray, this ray can be filtered out. For instance, in position A in Fig. 3, the angle of Ray3 , is larger than the NA of the 50/125 m GI-MMF pigtail used in our experiment, so Ray3 would be filtered out and the crosstalk only exists for two mode groups. In position B there is no power from Ray1, so only Ray2 and Ray3 are detected. In position C only Ray3 is detected with the same rule. So with the MSSF function shown in Fig. 3, can be expressed as Fig. 2(b) shows.
The main benefit for minimizing crosstalk through MSSF is to decrease the complexity of signal recovery, since the calcu-lation for matrix inversion, especially in case of high
1284 IEEE PHOTONICS TECHNOLOGY LETTERS, VOL. 23, NO. 18, SEPTEMBER 15, 2011
Fig. 3. Realization of MSSF with fiber pigtail.
sions is expensive. The rough number of computations for in-version of a real matrix is [5]. For a basic joint detec-tion scheme with , a high speed transmission optical MGDM system would place a heavy burden on the computation ability of the circuits. The detailed information about optimized signal processing work will be covered in Section III.
C. Modal Noise and Modal Group Interference
Modal noise arises due to changes of mode patterns over time with power variation from light sources such as vertical cavity surface-emitting laser (VCSEL) and environmental fac-tors. When a single mode source with a Gaussian profile is used to realize selective launching, the power for all launched modes changes linearly with the power fluctuation of the source. Com-pared to overfilled launching with using a VCSEL whose mode pattern is data pattern dependent [6] so that power changes for different modes are different, the laser influence to modal noise can be basically neglected assuming phase changes of modes at the output are slow, which can be satisfied when the trans-mission fiber is kept stable. The maximum power fluctuation for different radial offset launching over a time period of 20 s is 0.2 dB in the case a 185/250 m GI-MMF is coupled with a 50/125 m GI-MMF, smaller than the case when a SMF is used [4]. This can be explained by the larger number of modes de-tected by MMF which average out sharp interference variation. Therefore, increasing detection fiber’s core size can minimize not only coupling loss but also modal noise.
Besides modal noise due to the coherence of mode groups, modal interference could also influence system’s performance because one source with a small linewidth (100 kHz in our case) is used to launch all mode groups. However, the inter-ference pattern with 2 mode groups launched with 0 m and 20 m offset over 20 m GI-MMF in Fig. 4(a) is stable during the measurement which can be proved by the power fluctuation in Fig. 4(b). Test 1 and test 2 denote two 20 s tests (interval is half hour). It can be seen that the largest power intensity vari-ance is smaller than 0.25 dB. Only a two coherent input case is measured, since in our experiment the interference happens just between two mode groups due to MSSF. Certainly, modal interference can be greatly minimized, if different light sources are used for different mode groups.
III. EXPERIMENT
A. Power Distribution of Mode Groups
Fig. 5 gives the power distribution of mode groups after a 20 m long 185/250 m GI-MMF under selective launching with
Fig. 4. (a) Modal patterns. (b) Power fluctuation from mode group interference.
Fig. 5. Normalized power distribution under selective launching.
different radial offset. It can be seen that the power distribution of mode groups in the fiber is fixed and most of the power of the different mode groups is distributed at the radial position close to its launching offset. To limit the crosstalk between neigh-boring mode groups, a PD with a standard 50/125 m GI-MMF pigtail is chosen to couple the light power at the receiving side. It is experimentally demonstrated that MSSF using a 50/125 m GI-MMF pigtail with an NA of 0.2 could constrain the mode group crosstalk only into neighboring mode groups.
B. Experimental Results
The experimental setup of the 30 Gb/s 3 3 optical MGDM system with the optimized joint detection is shown in Fig. 6(a). A Mach-Zehnder modulator (MZM) is driven by a 10 Gb/s random pattern to modulate the light of a 1550 nm wavelength laser diode. The modulated light is amplified by the EDFA before it is split into three different arms. Three streams are decorrelated by the transmission through fibers with dif-ferent length. These fibers are connected to a Fiber Concentrator (FC). The distance between two adjacent waveguides of the FC is 30 m. Considering the large power loss of the large radial offset launching, the positions of selective launching are chosen as depicted in Fig. 6(b). A 20 m long 185/250 m GI-MMF is used as the MGDM transmission medium. Mode mixing be-tween mode groups is not observed in the experiment due to the short length of the transmission fiber. The output light from the 185/250 m GI-MMF is coupled by a 50/125 m GI-MMF to realize MSSF. The positions of selective detection are shown
CHEN et al.: 30-Gb/s 3 3 OPTICAL MGDM SYSTEM WITH OPTIMIZED JOINT DETECTION 1285
Fig. 6. (a) Experimental setup of a 30-Gb/s 3 3 optical MGDM system with joint detection; selective (b) launching and (c) detection; (d) estimated transfer matrix and (e) eye diagrams for received signals , , and recovered , .
in Fig. 6(c). Fig. 6(e) gives the eye diagrams of received sig-nals and . It can be seen that the eye diagram of is much more open and the interference from the other mode groups can be neglected. This contributes to the small power of the mode groups yielded by and with small and medium offset at the position where is detected. The crosstalk exists in the detected and due to the NA and core diameter of the 50/125 m GI-MMF.
The estimated transfer matrix is determined by the received power through training symbols in which transmitters send signals in turn. The rows in in Fig. 6(d) are normalized. It can be seen that the crosstalk, represented by the entries of which are circled, only occurred at neighboring mode groups. In the process of the signal recovery, is recovered based on the 2 2 submatrix at the low right corner of through ZF algorithm. The recovered eye diagram of is shown in Fig. 6(e). With the aid of the recovered could be recovered in the same way based on the 2 2 submatrix at the up left corner of . Fig. 6(e) gives the recovered result of . The 3 3 transfer matrix is reduced into two 2 2 submatrices due to limited mode group crosstalk. Matrix di-mension reduction means less computation and less calculation error. Optimized joint detection which separates the transfer matrix into small submatrices through MSSF is also applicable to MGDM systems with higher dimensions when radial offsets, detection positions and NA of detection devices are carefully chosen.
C. Discussion
Compared to modal noise and coherent interference, the MGDM system is more sensitive to the movement of the trans-mission MMF due to the power transfer between mode groups. However, in practical applications, by sending training symbols periodically, a real time transfer matrix can be calculated to cancel out the influence of fiber movements. Also the expecta-tion for long distance transmission with MGDM is high. With offset launching, it is possible that modes with large differential mode delays differences are launched together in one mode group, so the transmission distance is constrained. Techniques like adaptive optics [7] which can launch modes with small delay differences in one mode group could be helpful to extend
the transmission length. Certainly, the complexity of system is increased. Moreover, MGDM system’s capacity improvement can be expected with Wavelength Division Multiplexing and Polarization Division Multiplexing [8], although the perfor-mance for their applications in MGDM systems needs to be investigated further.
IV. CONCLUSION
In this letter, Mode Group Division Multiplexing (MGDM) over MMF has been proved as an efficient technique to enhance the MMF’s capacity without having additional bandwidth. The 30 Gb/s 3 3 optical MGDM system with optimized joint detection was realized using Mode-Selective Spatial Filtering (MSSF). The simplicity and optimization has a positive effect to minimize the complexity of signal processing and improve the system’s performance.
REFERENCES
[1] L. Raddatz, I. H. White, D. G. Cunningham, and M. C. Nowell, “An experimental and theoretical study of the offset launch technique for the enhancement of the bandwidth of multimode fiber links,” J. Lightw. Technol., vol. 16, no. 3, pp. 324–331, Mar. 1998.
[2] A. M. J. Koonen, H. van den Boom, I. T. Monroy, and G.-D. Khoe, “High capacity multiservice in-house networks using mode group di-versity multiplexing,” in Proc. OFC/NFOEC 2004, Feb. 2004, vol. 2, pp. 23–27.
[3] C. P. Tsekrekos and A. M. J. Koonen, “Mitigation of impairments in MGDM transmission with mode-selective spatial filtering,” IEEE Photon. Technol. Lett., vol. 20, no. 13, pp. 1112–1114, Jul. 1, 2008. [4] H. S. Chen, H. van den Boom, and A. M. J. Koonen, “30 Gbit/s 3 3
optical mode group division multiplexing system with mode-selective spatial filtering,” in Proc. OFC/NFOEC 2011, Los Angeles, CA, 2011, Paper OWB1.
[5] G. Strang, Linear Algebra and Its Applications, 3rd ed. Pacific Grove, CA: Brooks/Cole, 1988.
[6] S. Al-Sowayan and K. L. Lear, “Data pattern dependence of VCSEL far-field distributions,” IEEE Photon. Technol. Lett., vol. 16, no. 10, pp. 2215–2217, Oct. 2004.
[7] M. B. Shemirani, J. P. Wilde, and J. M. Kahn, “Adaptive compensation of multimode fiber dispersion by control of launched amplitude, phase, and polarization,” J. Lightw. Technol., vol. 28, no. 18, pp. 2627–2639, Sep. 15, 2010.
[8] E. Rochat, S. Walker, and M. Parker, “C-band polarisation orthogo-nality preservation in 5 Gb/s, 50 m multimode fibre links up to 3 km,” Opt. Express, vol. 11, no. 6, pp. 507–514, Mar. 2003.
