Optical waveguides – With optical coupler – Switch
Reexamination Certificate
2001-05-11
2003-02-04
Palmer, Phan T. H. (Department: 2874)
Optical waveguides
With optical coupler
Switch
C385S016000, C385S018000
Reexamination Certificate
active
06516109
ABSTRACT:
BACKGROUND
1. Field of Invention
The present invention relates to free-space optical switches, and particularly to a scalable, non-blocking, low insertion loss, re-configurable optical switches.
2. Related Art
An optical switch is a device for controlling the routing of light from one input port to any one of a number of output ports. A free-space optical switch is one in which the optical path is predominantly in an unguided medium. A non-blocking switch is one in which any input port can be connected with any unused output port, e.g., there are no particular paths between any input port and unused output port that are blocked because of other connections supported within the switch.
There are many ways to control the propagation of light. One way is to change the propagation direction of the light within the switch to direct the light through free-space to a desired output port, such as by rotating refractive (e.g., prisms) or reflective (e.g., mirrors) elements. Typically, the switch size is described in terms of the number of input and output ports. For example, a 2×2 switch would allow each of two input ports to be connected any of two output ports. Typical switch sizes range from 2×2 to 32×32. Future projections, however, show a commercial demand for non-blocking switches as large as 2048×2048 or more.
One important performance metric of optical switches is their insertion loss, defined (in dB) as 10 log (I
i
/I
o
), where I
i
is the optical power applied to the input of the switch and I
o
, is the optical power transmitted through the switch (arriving at the desired output port). Achieving a low insertion loss is an important goal for any optical switch design. This requires, for example, careful control of the input position and propagation direction of the light as it is inserted into the free-space medium, accurate control of the propagation direction of the light within the switch, and efficient collection of the light into the output port. In general, the free-space beam must arrive at the desired output port at the proper position and propagation angle to be coupled efficiently into this port. The effect of even small translations or angular deviations from the nominal values can lead to a reduction in coupling, and hence an increase in insertion loss.
Typical insertion losses range from 7 dB for small switches (e.g., 16×16) to 20 dB for larger switches. Larger switches exhibit higher insertion loss due to both diffraction effects and pointing-error-induced translation errors of the propagating beam increase as the optical pathlength is increased. The ability of a particular switch's design technology to enable the fabrication of larger format switches with appropriately small insertion loss is defined generally as the scalability of the switch.
A related performance metric is the uniformity of the insertion loss over all of the possible connections within the switch. In a large switch, the insertion loss may vary between 3 dB and 7 dB over the entire range of switch connections. It is desirable to have a lower and more uniform insertion loss between any input to output optical path.
Currently, large format optical switches are developed using microfabrication techniques to create large arrays of small, movable, opto-mechanical two and three-dimensional structures. In a typical two-dimensional approach, a two-dimensional array of mirrors is used to deflect the light from any input fiber to any output fiber. For an N×M switch, the mirror array contains a total of N times M mirrors. Every mirror is used only for one particular switch connection, and it is moved out of the way for any other switch state. As such, the typical two-dimensional approach can be non-blocking, although many are not, such as disclosed in U.S. Pat. No. 6,072,923, entitled “Optical Switching, Routing, and Time Delay Systems Using Switched Mirrors”, incorporated by reference in its entirety.
Even if this approach is non-blocking, there is the problem of limited scalability, due mostly to the requirement of using the same number of mirrors as switch states. For an N×M switch, N×M mirrors need to be arranged in a two-dimensional plane to provide the desired optical paths. Larger size mirrors increase the optical path length and the size of the switch, while smaller mirrors increase the insertion loss. Thus, a typical two-dimensional switch will have unacceptably larger insertion losses and larger state-dependent losses (larger non-uniformity), as the number of input and output ports is increased. U.S. Pat. No. 6,097,859, entitled “Multi-Wavelength Cross-Connect Optical Switch”, which is incorporated by reference in its entirety, discloses a two-dimensional switch that has limited scalability. Further, because the two dimensional switch has N×M mirrors, issues with manufacturing yield and operating reliability also become important as the number of ports in the switch is increased.
Three-dimensional approaches allow smaller optical pathlengths and greater uniformity in optical pathlengths, due in part to the ability to distribute the number of N×M optical paths within a three-dimensional volume, instead of a two-dimensional plane. The effect of this is a reduction in insertion loss and insertion loss non-uniformity over that provided by a similar size, two-dimensional switch. However, current three-dimensional approaches also have problems, such as with size, scalability, and complexity.
In one approach, described in PCT Int'l Publ. No. WO 00/52835, entitled “Opto-Mechanical Valve and Valve Array for Fiber-Optic Communication”, incorporated by reference in its entirety, the three-dimensional switch still requires a large number of movable mirrors, with little description on reducing insertion loss. Other types of three-dimensional switches that reduce the number of mirrors necessarily increases the complexity, such as by requiring each mirror to have a large number of discrete pointing positions, e.g., corresponding to each input or output port. This requires precise alignment of the fibers and lenses with respect to each other and with respect to the micromirrors. Furthermore, positioning and control of the micromirrors can become complicated, thereby increasing the cost and complexity of designing and making such switches.
Other current types of optical switches utilize feedback to accurately adjust the propagation direction of the free-space beam, such as disclosed in U.S. Pat. No. 5,206,497, entitled “Free-Space Optical Switching Apparatus”, and complex beam pointing circuitry, such as disclosed in U.S. Pat. No. 6,005,998, entitled “Strictly Non-Blocking Scalable Matrix Optical Switch”, both of which are incorporated by reference in their entirety. Both these types of switches introduce added complexity and expense to the switch.
Current optical switches typically also have fixed fiber inputs and outputs, i.e., one set of fibers are used as light inputs and another set of fibers are used as light outputs. This fixed configuration of inputs and outputs limits the flexibility of the switch.
Accordingly, it is desirable to make and have a large free-space optical switch without the disadvantages discussed above with current optical switches.
SUMMARY
In accordance with one aspect of the invention, a free-space non-blocking optical switch includes two planar substrates facing each other, with each substrate having patterns of beam steering elements (e.g., moveable mirrors, either singularly or in an array) and fiber interfaces with collimating elements (e.g., lenses). Each of a plurality of input and output fibers has an associated lens and beam steering element. The fibers are held in lithographically-defined through-holes in the substrate, and the associated lenses are located by lithographically-defined kinematic points in the substrate. The kinematic points set the lateral position of the lens with respect to the fibers. The use of lithography to determine the relative placement of the critical optical elements increases th
Gutierrez Roman C.
Tang Tony K.
MacPherson Kwok & Chen & Heid LLP
Palmer Phan T. H.
Siwave, Inc.
LandOfFree
Low insertion loss non-blocking optical switch does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Low insertion loss non-blocking optical switch, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Low insertion loss non-blocking optical switch will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3123030