RECONFIGURABLE MULTI-CHANNEL TRANSMITTER FOR DENSE WAVELENGTH DIVISION MULTIPLEXING (DWDM) OPTICAL COMMUNICATIONS
CROSS-REFERENCES TO RELATED
 The present application is a continuation of U.S. patent application Ser. No. 09/610,312 filed Jul. 5, 2000, which is a non-provisional patent application of and which claims the benefit of priority from U.S. Provisional Patent Application No. 60/212,431 as filed Jun. 16, 2000, the full disclosures of which are incorporated herein by reference.
FIELD OF THE INVENTION
 The field of the present invention relates in general to optical communication using inexpensive, stable and accurate devices for creating channels for wavelengths of light. More particularly, the field of the invention relates to a wavelength reconfigurable, multiple channel transmitter for optical Dense Wavelength Division Multiplexed signals.
BACKGROUND OF THE INVENTION
 Optical fiber has nearly unlimited bandwidth, yielding 20 to 50 times more bandwidth than copper cable. Optical fiber is now the solution of choice at the wide area network (WAN) level. Implementation of fiber optic communication at the local area network (LAN) level will enable users to break current bottlenecks in the last mile of information transfer. A further attraction of fiber-optic technology is its scalability. Most current fiber LAN products are Ethernet-based in a range of 100-Mbit/second up to 1-Gbit/ s. Fiber optical communication is easily scalable up to 10 Gbits/s, and several equipment vendors have announced fiber-optic links that can support transmission up to 1 terabit/s utilizing more than 128 DWDM.
 Optical telecommunications networks require significantly increased bandwidth to handle current and projected communications traffic. Most optical networks use time division multiplexing (TDM) with a single laser transmitter as a means of combining many separate transmissions, allowing data rates of up to 10 Gbits/sec. The current market trend is toward systems that use many individual transmitters, each of a different wavelength to increase channel capacity (an approach known as wavelength division multiplexing, WDM). For example, a transmitter consisting of 8 distinct wavelengths provides 8x the capacity of a single channel network (e.g. 80 Gbits/sec). Furthermore, WDM systems are scalable to 16, 32, 64, 128 etc. distinct channels, and make the most efficient use of the extremely high bandwidth of optical fiber communication networks. Therefore, what is needed is a system for providing a stable, rapidly reconfigurable, multi-channel source for WDM and DWDM (dense-WDM) optical communications networks which can meet anticipated bandwidth demand.
 A fiber optic communication link provides a virtually noise free medium for transporting complex signals without distortion and interference. Fiber optic cable has losses as low as 1-2 dB per kilometer, which is much lower than the 1-2 dB per 100 ft. for coaxial cable. Due to high frequency of laser light, fiber optic cables provide a bandwidth in which many channels of information can be trans
ported across a single fiber cable. The laser diodes' high efficiency, small size, high reliability, and low cost ($5-$10 for infrared, $40-$50 for visible light) make them the ideal choice for communication devices. However, other components in these devices escalate the cost of present optical DWDM transmitter devices.
 Low cost optical DWDM optical communication devices are not available in the market due to the current trend toward integrating costly solid state semiconductor lasers and associated optical structures, and because optical solid state control structures are new and changing rapidly. The latest integrated optical components are created lithographically on a semiconductor chip substrate or on glass. However, it is very difficult and expensive to make multiple lasers lithographically (in VLSI) which are locked to a reference grid.
 Therefore, what is needed is a way to reduce the cost of building state of the art DWDM optical communication devices. What is also needed is a method for shifting the burden of providing a high performance DWDM optical communication system from costly precision optical components to inexpensive solid state control structures. It would be desirable to achieve a high performance DWDM transmitter using low cost optical components having a wider range of tolerances than is presently possible
 Conventional WDM/DWDM sources lack stabilized frequency referencing and rapid reconfigurability in the event one or more lasers fail or are disrupted. FabryPerot interferometers (FPIs) have been used to attempt to lock a laser to a stable frequency. Fabry-Perot Interferometers used as scanning interferometers can sense extremely small wavelength shifts when piezo-electric actuators (PZTs) are used for tuning the multi-pass dual mirror optical cavity. However, these interferometers require adequate wavelength references for long-term stability.
 Molecular absorption line sources from various gases have been tried, in particular acetylene, and offer a potential medium for multi wavelength referencing. However, their unevenly spaced absorption lines and absorptive operation makes them difficult to tune for FPI-based spectral applications. FPIs with narrowly—spaced uniform transmission peaks(~100 GHz) have often been considered for an absolute frequency reference when locked to atomic or molecular absorption lines. Glance, B. S. et al. (1988), "Densely Spaced FDM coherent Star Network with Optical Signals Confined to Equally Spaced Frequencies," IEEE J. Lightwave Technol. LT-6:1770-1781. and others. While there have been some applications for stabilizing laser arrays, such methods are restricted in wavelength positioning and require active feedback control in conjunction with costly precision optical structures, thereby driving up costs.
 Fiber Bragg gratings (FBGs) can produce a narrow band response around a single wavelength. However, their narrow response band and over wavelength span poses limitations. Some of the current alternatives for optical transmitters provide only partial solutions. Examples of some conventional solutions and their disadvantages are described below. Most describe an optical reference source. None use a gas spectral line reference source.
 U.S. Pat. No. 5,892,582 discloses a fiber Bragg grating (FBG) source which provides spectral output at a