This invention relates to the field of arrayed waveguide grating devices and, more specifically, to arrayed waveguide gratings with improved filter accuracy.
In order to meet the ever-increasing demand for transmission bandwidth in communication networks, the development of techniques for Wavelength Division Multiplexing (WDM) is increasing in importance. In particular, in dense WDM (DWDM), the use of the available optical bandwidth is optimized by increasing the system spectral efficiency (i.e., the ratio between signal bandwidth and channel spacing). This is achieved by employing many closely spaced carrier wavelengths multiplexed together onto a single waveguide such as an optical fiber and/or by increasing the signal modulation speed of every single data channel.
Because the channels' spectra are more closely packed, the ability of separating or merging channels without introducing any additional signal penalty is becoming of great importance. Therefore, optical filters are being developed with filter response and frequency alignment that provide for a better match to the signal frequencies.
An important class of such optical filters is represented by Arrayed Waveguide Gratings (AWGs). An arrayed waveguide grating (AWG) is a planar structure comprising a number of array waveguides whose arrangement emulates the functionality of a diffraction grating (see e.g. M. K Smit and C. van Dam, “Phasar based WDM devices: Principles, design, and applications,” IEEE J. Select. Topics Quantum Electron., vol 2, pp.236-250, 1996). AWGs are commonly used as multiplexers or demultiplexers (i.e., devices that can merge a multitude of optical frequencies from multiple inputs to a single output port or separate a multitude of optical frequencies from a single input to multiple output ports, respectively). Furthermore, AWGs can also be designed to perform as passive, wavelength selective, strictly non-blocking cross-connects for sets of optical channels. For example, AWGs can simultaneously operate as a multiplexer and demultiplexer by distributing and recombining multiple frequencies entering any of a multitude of input ports into any of a multitude of output ports (see e.g. C. Dragone, “An N×N optical multiplexer using a planar arrangement of two star couplers,” IEEE Photon. Technol. Lett., vol. 3, pp. 812-815, 1991). Because of this property and uncommon versatility, AWGs are attracting an increasing interest for large optical cross-connect systems.
The use of AWGs, however, presents some limitations. Due to the intrinsic diffraction characteristics of an AWG, the maximum channel count or maximum spectral width of this kind of devices may be limited. Additional limitations arise if the AWG is designed to cross-connect channels that are equally spaced in frequency to be compliant with the industry International Telecommunications Union (ITU) standard frequency grid as opposed to being equally spaced in wavelength (see e.g. P. Bernasconi, C. Doerr, C. Dragone, M. Cappuzzo, E. Laskpwski, and A. Paunescu, “Large N×N Waveguide grating routers,” J. Lightwave Technol., vol. 18, pp. 985-991, 2000).
The present invention advantageously provides for an arrayed waveguide grating with improved transmission efficiency.
In one embodiment of the present invention a method for improving the transmission efficiency of an arrayed waveguide grating (AWG) comprising P input ports and Q output ports (P×Q AWG), includes a novel positioning of the ports at the input of a first P×M star coupler and at the output of a second M×Q star coupler, M being the number of waveguides connecting the star couplers. The novel method used to determine the optimum position of the ports ensures that the maximum frequency deviation between the signal frequency and the center of the corresponding transmission passband is minimal for all possible connections among input and output ports.
In an embodiment of the present invention, an AWG with improved transmission efficiency includes an input P×M star coupler, an output M×Q star coupler, and a plurality of waveguides of unequal length connecting the input and output star couplers. A first input port of the AWG is initially positioned near the center of the input star coupler. The initial positions of the output ports of the AWG are as such defined by the images of the wavelengths dispersed by the first input port. The remaining P−1 input ports of the AWG are initially positioned as defined by the images of the wavelengths dispersed by a substantially central output port. An angular spread, Δq, for each of the Q output ports is then determined, where Δq is the angular spread of the images of wavelengths of all of the input ports expected to converge on each of the output ports. Final positions of the input ports are then defined by the subsequent variation of the initial positions of the input ports, such that the value of ΔMAX is minimized, where ΔMAX is a maximum one of respective angular spreads, Δq, of the images of the wavelengths expected to converge for each of the output ports. The output ports of the AWG are finally positioned such that each output port is positioned in substantially the center of its respective final angular spread, Δq, where the respective final angular spreads are defined by the images of the wavelengths dispersed by the finally positioned input ports.
In an alternate embodiment of the present invention, the center axis of the output ports of an output star coupler of an AWG are again repositioned to maximize the weakest wavelength signal transmission coefficient for any wavelength dispersed from any of the input ports of an input star coupler, instead of being in the center of their respective angular spreads Δq. As such the transmission coefficient for the AWG is increased.
The teachings of the present invention can be readily understood by considering the following detailed description in conjunction with the accompanying drawings, in which:
To facilitate understanding, identical reference numerals have been used, where possible, to designate identical elements that are common to the figures.
The present invention advantageously provides for an arrayed waveguide grating with improved transmission efficiency. Although embodiments of the present invention are being described herein with respect to P×Q arrayed waveguide gratings (AWGs), it will be appreciated by those skilled in the art informed by the teachings of the present invention, that the concepts of the present invention may be applied to AWGs of various other input and output port configurations and even further to other optical cross connect switches and devices.
The diffraction properties of the AWG 100 may be described in the paraxial approximation by the following relation:
anp(λin−λout)+nwΔL=mλ (1)
where a is the grating period, in/out are the incident and the diffracted angles, respectively, for the wavelength λ, ΔL is the optical path difference between neighboring grating arms, np/w are the refractive indices in the planar and grating regions, respectively, and m is an integer corresponding to the order of diffraction. From equation (1) above, the inventors determined that:
From equation (1), it may be determined that if the same set of wavelengths is launched from two distinct input ports, two nominally identical combs will appear shifted with respect to each other at the output of the star coupler 1102 as depicted in FIG. 2.
The image combs are depicted as arrays of arrows pointing at the image locations along the x-axis that characterizes the image plane. Two different situations a represented: Channels evenly spaced in wavelength (λq=λo+qΔλ, Δλ is the channel spacing, 1<q<Q), and channels evenly spaced in frequency (νq=νo+qΔν, Δν is the channel spacing, 1<q<Q). The wavelength combs have evenly spaced teeth so that when they originate from two distinct input locations A and B they appear shifted to each other but their teeth may still coincide in the overlap region. The frequency combs have unevenly spaced teeth, therefore when shifted with respect to each other, no two consecutive points coincide within the overlap region.
The full cross-connectivity between input and output ports will be achieved when the teeth of all the image combs locally overlap. Therefore, an optimal configuration would require a constant distance between the comb's teeth and, consequently, the ports of the AWG to be equally spaced.
Because WDM systems are moving toward higher channel counts often combined with a higher spectral efficiency, the constraints of the filter response or the ability of the devices to separate channels, is getting tighter. As such, a need for better alignment of the filter transmission passband to the channel frequency, a wider and flatter passband, lower losses, and lower crosstalk are needed. Typically, P×Q AWGs (with P,Q>1) are characterized by Gaussian intensity responses, which enhances the required alignment accuracy to the optical channels in order not to magnify the transmission penalty due to additional insertion losses, asymmetric signal filtering, and increased crosstalk.
Current P×Q AWGs are affected by an intrinsic misalignment between the center of the transmission passband and the ITU frequency standards. Although this discrepancy may by negligible in AWGs handling a relatively narrow frequency span, this problem is becoming increasingly important in AWGs with large P and Q and narrow channel spacing.
The original P×Q AWG design, and still the one most commonly applied, possesses the property of having the same number of input and output ports (P=Q) and can provide full cross-connectivity by using a unique set of P different wavelengths. This arrangement takes advantage of the multiple images generated by an AWG in different diffraction orders to provide a fully periodic frequency response. Such AWGs are designed with the output ports covering an entire Brillouin zone of the grating, so that as soon as one channel falls out on one side of the sector occupied by the output ports, the next diffraction order moves in from the other side replacing the “lost” channel with its copy in the contiguous diffraction order.
However, referring back to the equation (1) above, to provide the periodic response as depicted in
A frequency misalignment consequent to the use of three diffraction orders may, however, be partially mitigated by an appropriate global shifting of the output ports as proposed in “Large N×N Waveguide Grating Routers”, P. Bernasconi, C. Doerr, and C. Dragone, U.S. Pat. No. 6,381,383, filed on Apr. 30, 2002, which is herein incorporated by reference in its entirety. The Bernasconi patent, U.S. Pat. No. 6,381,383, teaches a novel method for arranging the spacing of the output ports of an output star coupler of a waveguide grating router so as to increase a signal transmission coefficient from each of the input ports to the output ports. The Bernasconi patent further describes how transmission coefficients vary for wavelengths originating from a given input port. The misalignment increases linearly with the distance between a central wavelength and a wavelength diffracted in the next higher or in the next lower diffraction order.
Another effective improvement in the frequency response is obtained by operating the AWG within a single diffraction order. This is achieved by doubling the number of output ports and eventually merging them into pairs to recover the original number of output ports as depicted in FIG. 4.
Output ports i and i+P are complimentary to each other and by merging these pair of output ports, new output port combinations are obtained comprising all of the initial P=Q outputs on the now four combined output channels. For example, in
Consequently, misalignments at the output ports are drastically reduced and nominally eliminated in cases where the optical channels are equally spaced in wavelength. In addition, improved loss uniformity can be guaranteed because of the lack of stringent constrains on the free spectral range of the AWG. Note that the channel mapping between input and output ports in FIG. 3 and
As such, one embodiment of the present invention provides a novel AWG where the P input and Q output ports positions are modified to correct for the limitations presented when optical channels are evenly spaced in frequency. Embodiments of the present invention also provide a method to determine the optimal position of the input and output ports for increasing a signal transmission coefficient and, as such, improving the transmission efficiency of a subject AWG. It should be noted that although the concepts of the present invention are being described herein with reference to a P×Q AWG, the concepts of the present invention may be applied in substantially any AWG with any number of input ports and output ports, and even further, to other optical cross connect switches and devices.
The method of one embodiment of the present invention begins with the design of a grating that will support the P×Q AWG. That is, an AWG with a sufficient number of needed input and output ports is chosen. The initial positions of the input ports and output ports of the AWG are then determined. More specifically, the initial positioning of the ports is accomplished by choosing an input position (usually substantially at the center of the input of a first star coupler) and finding where the different optical channels of an input signal with channels evenly spaced in frequency are imaged at the output of a second star coupler. The output ports of the second star coupler are, as such, positioned at positions defined by the images of the wavelengths dispersed by the substantially central input port. The position of the remaining P−1 input ports are then determined by following the same procedure above, but using a substantially central output port as a source. More specifically, a substantially central output port, determined by the input images, is chosen as a source and the remaining P−1 input ports of the AWG are positioned at positions defined by the images of the wavelengths dispersed by the substantially central output port.
In a second step, for every channel originating from every input position, the exact location of every image is calculated. In a third step, all the images that are expected to appear at any output port q (1<q<Q) are considered collectively, regardless of their diffraction order, and their angular spread Δq determined. The fourth step consists of the identification of ΔMAX corresponding to the maximum value among Δq (1<q<Q). In the next step, the position of the input ports is varied to minimize the value of ΔMAX by iteratively accomplishing steps two to four. When the value of ΔMAX is minimized, the final positions for the input ports have been determined. That is, the minimization process will deliver the optimum position of the input ports. The output ports will be finally positioned in the middle of respective Δq intervals, defined by the images of the wavelengths dispersed by the finally positioned input ports. This method results in a decrease of ΔMAX by approximately a factor of two. This allows for the doubling of the number of channels that an AWG may efficiently handle for a given tolerated ΔMAX, thus improving the filtering performance for a given number of channels.
In alternate embodiments of the present invention, the center axis of the output ports of the output star coupler of an AWG are repositioned to maximize the weakest wavelength signal transmission coefficient for any wavelength dispersed from any of the P input ports, instead of being in the center of their respective angular spreads Δq. As such the transmission coefficient for an AWG in accordance with one embodiment of the present invention is increased.
As evident in
It will be appreciated by those skilled in the art informed by the teachings of the present invention that in alternate embodiments of the present invention, the method and concepts described above for the optimization of the positioning of the input and output ports of an AWG may be implemented for both, optical channels that are equally spaced in wavelength and optical channels that are equally spaced in frequency. Even further, the method and concepts described above for the optimization of the positioning of the input and output ports of an AWG may be implemented with optical channels that are not equally spaced.
While the forgoing is directed to various embodiments of the present invention, other and further embodiments of the invention may be devised without departing from the basic scope thereof. As such, the appropriate scope of the invention is to be determined according to the claims, which follow.
This application is related to U.S. Pat. No. 6,381,383, issued Apr. 30, 2002, which is incorporated herein by reference in its entirety.
Number | Name | Date | Kind |
---|---|---|---|
5136671 | Dragone | Aug 1992 | A |
5745612 | Wang et al. | Apr 1998 | A |
5822481 | Mestric | Oct 1998 | A |
6058233 | Dragone | May 2000 | A |
6181849 | Lin et al. | Jan 2001 | B1 |
6351583 | Bergmann et al. | Feb 2002 | B1 |
6381383 | Bernasconi et al. | Apr 2002 | B1 |
6418249 | Nakamura et al. | Jul 2002 | B1 |
6490395 | Nara et al. | Dec 2002 | B1 |
6549698 | Bernasconi et al. | Apr 2003 | B2 |
6597841 | Dingel | Jul 2003 | B1 |
6678446 | McGreer et al. | Jan 2004 | B1 |
6741772 | Ide | May 2004 | B2 |
6768842 | Bulthuis et al. | Jul 2004 | B2 |
6810167 | Tabuchi et al. | Oct 2004 | B2 |
6810177 | Kaneko | Oct 2004 | B2 |
20020089721 | Nicolas | Jul 2002 | A1 |
20030063858 | Bulthuis et al. | Apr 2003 | A1 |
Number | Date | Country | |
---|---|---|---|
20040264857 A1 | Dec 2004 | US |