The present invention relates to the field of network optical interconnection and more specifically to apparatus and methods to facilitate transition from multimode fiber networks to single mode networks.
The apparatus and methods disclosed here, provide optimized mode couple functionalities that enable transmission of signals from single mode transceivers over multimode fibers.
Intensity modulated and direct detection (IM-DD) transceivers using vertical cavity surface emitting lasers (VCSELs), operating in the spectral range of 850 nm to 950 nm, are widely deployed in enterprise data centers. Those transceivers, which operate over multimode fiber (MMF) channels, can currently support aggregated data rates up to 100 Gb/sand 25 Gbps per lane).
As the demand for higher data rates continues to grow, the pursuit of cost effective and efficient methods to increase transmission capacity are actively underway. For VCSEL-MMF channels, transmission at aggregated data rates up to 800 Gb/s require the combination of several schemes, such as short wavelength division multiplexing (SWDM), spatial division multiplexing, (SDM) using parallel fiber optics, and pulse amplitude modulation (PAM). Utilization of VCSELs for higher data rates while maintaining useful reaches and low cost is challenging due to VCSEL bandwidth limitations, non-linear responses, laser mode dynamics, chromatic dispersion, and other effects.
On the other hand, IM-DD transceivers using single mode lasers, such as DFBs with relatively narrow linewidth, may face less difficulties to achieve higher transmission rates. Single mode transceivers are designed to operate over single mode fiber (SMF) using wavelengths over 1260 nm.
As any other asset, data center hardware ages over time and eventually need to be replaced. However, the replacing cycles for transceivers and media infrastructure are different, their schedules do not necessarily synchronize. In most cases, transceivers are changed at least two times more often than fiber infrastructure.
It is likely that in a mid-term or long-term future, data center using currently MMF will require data rates beyond the capabilities of the VCSEL transceiver technology and, therefore, will need to switch to SM transceivers. Since SM transceivers cannot operate over MMF deployed in data centers, maintaining reaches and reliable connections will require to change also the fiber and connectivity of the datacenter.
It will be highly desirable to have SM transceivers that operates in both MMF and SMF channels. However, this is not currently possible for reasons described below.
First, there is a mode mismatch between the SM transceiver (and SMF) mode to the fundamental mode of the MMF. Therefore, more than one mode is excited during coupling, and more are added along the channel due to multiple connections, fiber bending, patch cord vibrations, among other causes.
Second, the SM transceivers operate at much longer wavelengths, i.e. 1300 nm. Since the modal dispersion of deployed MMF has been optimized for a narrow spectral window around 850 nm (i.e. >4700 MHz-km), these fibers provide very low modal bandwidth at the SM transceiver operational wavelengths. Third, the receivers used in SM transceivers have a very small aperture relative to the MMF core diameter, increasing the losses of the channel.
Due to the reasons mentioned above, even small vibrations can be important sources of modal noise when trying to operate MMF with SM transceivers. Therefore, SM transceivers operating over MMF channels are subject to critical high modal dispersion (low modal bandwidth), increased noise, and increased attenuation.
A solution to minimize this large signal degradation is to optimize the coupling of the SM transceiver (or SMF) mode to the MMF fundamental mode. Several works towards that goal are described in the next section.
A way to overcome the limitations describe above is to use a non-standard fiber that is designed to optimize the coupling of the SM transceiver to the fundamental mode of the fiber as shown in patent application for Universal Fiber [Corning US 20150333829 A1], In addition, a hybrid SM-MMF that combine in some way the features of SMF and MMF, such as designs with double cores to have operation as MMF, where the smallest center core is designed to operate with SM transceivers as described in U.S. Pat. No. 9,563,012.
However, those approaches, require modification of the MMF core which impact on its performance. MMF used in VCSEL-MMF channels have a refractive index design, denominated alpha refractive index profile, which is very sensitive to changes on the geometry or concentration of the dopants. Therefore, a new fiber such as the proposed in prior art, can neither match the performance of high bandwidth MMF nor the performance of SMF. A decision to install a new fiber not optimized for SM or MM transmission, to enable SM transmission cannot be economically supported since it is significantly more efficient to install SMF directly.
A better approach is to use an external device that can operate with installed base MMF. This device will enable better coupling between the fundamental modes of the SMF and MMF. For example, the use of adiabatic couplers such the ones described in U.S. Pat. No. 7,606,452—2009, U.S. Pat. No. 7,184,623 Avanex (currently Oclaro) or U.S. Pat. No. 4,763,976—(1988) Corning—Connector.
Yet another approach is to use lenses to minimize the mismatch of the SMF to MMF modes. For example, in U.S. Pat. No. 8,218,928 GRIN lenses with pin elements are used. Other approaches with lenses are shown in U.S. Pat. No. 6,655,850 2003 (Corning Hybrid fiber expanded beam connector) or in paper Photonics Society Summer Topical Meeting Series, 2013 IEEE, pp. 256-257: Converting a Multimode Fiber into a Single-Mode Fiber.
A more practical approach is to use a SMF path cord with large core area such as fiber U.S. Pat. No. 6,185,346 (2001) and U.S. Pat. No. 6,487,338 (2002). However, it is difficult to make a large core diameter fiber while maintaining the single mode condition. The inventors of this application verified that even SMF designed for long haul applications do not have a core diameter large enough to optimize the coupling SMF-MMF fundamental mode.
A more recent approach is to use phase mask to modify the mode profile of the SM transmitter or SMF, to optimize the coupling with the MMF as shown in http://www.cailabs.com/.
In the next section of this application we disclose novel embodiments that provide optimum coupling between SM mode to MMF fundamental mode while overcoming cost, portability, or complexity limitations attributed to some disclosed prior art mentioned above.
A patch cord for transmitting between a single mode fiber (SMF) and a multi-mode fiber (MMFs) has a MMF, SMF, and a photonic crystal fiber (PCF) with a hollow core placed between the SMF and MMF, A mode field diameter (MFT)) of the PCF hollow core section is in the range of 16 to 19 microns, the length of the PCF is between 1 cm to 10 cm, the MMF has 50±2 microns core diameter, the SMF has a 6-9 microns core diameter, and the coupling between the PCF mode to the MMF fundamental mode is maximized.
An apparatus in the form of a fiber optic patch cord that optimize the excitation of the fundamental mode of a laser optimized multimode mode fiber (MMF) from a single mode fiber, or a single mode transceiver, is disclosed. The apparatus enables efficient coupling of SMF mode to MMF fundamental mode and MMF fundamental mode to SMF mode.
The apparatus was designed using fiber modeling, imaging, and temporal signal processing. The fiber modeled is MMF with refractive index often referred to as the α-profile. The refractive index profile of these MMFs inside the core is described by a function given by,
where Δ≈(n1−n2)/n1, n1 is the refractive index on the axis of the fiber, n2 is the refractive index in the cladding, r is the radial position inside the fiber core, a is the core diameter, and α is the exponent parameter which takes a value of ˜2 for fibers designed to support operation near 850 nm.
From theory described in [2], a simplified expression for the relative mode group delay, tg, can be derived from (1) as a function of the wavelength and the α-profile parameters as shown,
where c is the speed of light in the vacuum, g is the mode group (MG) index, (a mode group comprises those modes that have nearly equal propagation constants), vg is the number of modes inside the MG, which have a propagation constant larger than βg(v), vT is the total number of modes, N1 is the group refractive index of the core material at r=0 and, λ is the optical source wavelength.
The optimum alpha value that minimize group delay at a single operational wavelength λ and y the profile dispersion parameter is given by,
Using (3) and λ=850 nm the α-profile that optimize transmission at the 850 nm window can be obtained. Around 850 nm there are around 380 modes grouped in 19 mode groups (MGs) are obtained. At 1300 nm, the same fiber can have less than 160 modes distributed in 12 or 13 mode groups. In
The coupling ratios resultant from a SMF launch into a MMF can be obtained from the overlap integral of the MMF normalized field amplitude patterns, ψGi(r,ϕ) and at the field amplitude patterns of a SMF with core radius, R, as shown below,
where i is the index of the modes that are included in the mode group, MG=G, Δx and Δy represent misalignments of the SMF fiber with respect to the MMF and ε(R,x,y) is the normalized fundamental field pattern of a SMF of radius R.
The value of the total coupled power from the SMF with core radius RTx, to each mode group is given by,
where the power of all the modes inside the MG=G are added. The value of PG ranges from 1 for maximum efficiency coupling to 0 no coupling. To compute the signal after the detector we assume a SMF or RRx core radius placed between the MMF and the detector. The power coupled to the detector given,
The objective now is to find the optimum radius of the SMF that maximize the PG for G=1, which represents the fundamental mode of the MMF, while minimizing PG for the sum of all other Gs different than one.
The optimizing metric to find optimum RTx or RRx is estimate the power in the fundamental divided by the power in the other modes as shown below.
Considerations for a known range of mechanical tolerances can simplify (8) as follows,
M(RTx,RRx)=minf(Δx
where, f(ΔxTx, ΔyTx, ΔxRx, ΔyRx) represent the tolerance space for fiber misalignment during the fabrication of this device. For each RTx, RRx the minimum value in the tolerance space represents the worst case operation for that combination RTx, RRx as shown in the following example.
In 500, it is shown a SM source connected to a SMF fiber or waveguide, 510. This fiber or waveguide is connected to a 50 micron MMF, 520. The other side of that fiber is connected to another SMF or waveguide, 530. The SMF or waveguide is connected to a SM photodetector, 540.
In
The figures show the coupling power in the vertical axis vs the launch fiber mode field diameter. In this figure, we assume that there is not misalignment between the SMF to the MMF or from the MMF to the detector. Traces 600 to 620 are related to the configuration that use a MM photodetector shown in the configuration 400-410-420-430 shown in
Results for the configuration shown in
Numerical simulation using the methods described above and the tolerance range reduction restrict the optimum region to 16 to 19 microns.
It should be noted that the disclosed method does not include the mode coupling caused by the connectors of the channel. Those effects were evaluated experimentally and are described in the next section.
The calculation methods described in previous section indicates that MFD between 16 to 19 microns are needed to enable operation of SMF transceivers over a MMF of 50 micron diameter. The inventors realized that the required. MFD values cannot be achieved by standard commercially available fibers. Also, that it is difficult to increase the MFD in standard fibers without increasing the number of modes or without a high reduction of numerical aperture (NA). Lower NA can increase the coupling and losses between the laser and the fiber. In this application we propose to use a small section of a specific type of photonic crystal fibers to provide the large MFD without increasing the number of propagating modes in the MMF.
Alternatively,
A critical part for the fabrication method is the control of misalignment during splicing.
To test the properties of the disclosed patch cord we compare the intensity profiles of light launched to a MMF with and without the described apparatus of SMF to the fundamental mode of MMF adaptor as shown in
The performance differences between MMF channel using direct launch from a SMF and using the disclosed apparatus were measured. Several channel configurations were evaluated. Part of the channel was subject to small amount of motion using a fiber shaker. This simulates movement of the patch cord in the data center that can produce modal noise.
There were always performance advantage using the disclosed apparatus.
However, the eye diagram resultant of the propagation over the channel with six connectors is nearly close, due to modal noise. On the other hand, the channels using the disclosed patch cord shows smaller signals amplitude (due to the patch cord attenuation) but with improved eye opening. The disclosed patch cord helped to reduce modal noise penalties by optimizing the coupling to the MG.
While particular embodiments and applications of the present invention have been illustrated and described, it is to be understood that the invention is not limited to the precise construction and compositions disclosed herein and that various modifications, changes, and variations may be apparent from the foregoing without departing from the spirit and scope of the invention as described.
This application claims priority to U.S. Provisional Application No. 62/726,636, filed Sep. 4, 2018, the subject matter of which is hereby incorporated by reference in its entirety.
Filing Document | Filing Date | Country | Kind |
---|---|---|---|
PCT/US2019/049496 | 9/4/2019 | WO | 00 |
Number | Date | Country | |
---|---|---|---|
62726636 | Sep 2018 | US |