60/208,483 (Attorney Docket No. 34013-00029USPL), filed June 2,2000; U. S. Provisional Patent Application No.

60/208,478 (Attorney Docket No. 34013-00028USPL), filed June 2,2000; U. S. Provisional Application No. 60/208,482 (Attorney Docket No. 34013-00027USPL), filed June 2,2000; U. S. Provisional Application No. 60/208,477 (Attorney Docket No. 34013-00033USPL), filed June 2,2000; U. S.

Provisional Application No. 60/212,099 (Attorney Docket No. 34013-00036USPL), filed June 15,2000; which are incorporated by reference herein in their entirety. This application also incorporates by reference herein in their entirety U. S. Patent Application 09/724,803, entitled "Diffraction Grating for Wavelength Division Multiplexing/Demultiplexing Devices" (Attorney Docket No.

34013-28USPT), filed November 28,2000; U. S. Patent Application 09/724,804, entitled"Device and Method for Optical Performance Monitoring in an Optical Communications Network" (Attorney Docket No. 34013- 40USPT), filed November 28,2000; U. S. Patent Application 09/724,771, entitled"Athermalization and Pressure Desensitization of Diffraction Grating Based WDM Devices" (Attorney Docket No. 34013-27USPT), filed November 28, 2000; U. S. Patent Application 09/724,710, entitled "Athermalization and Pressure Desensitization of Diffraction Grating Based WDM Devices" (Attorney Docket No. 34013-48USPT), filed November 28,2000; U. S. Patent Application 09/724,638, entitled"Athermalization and Pressure Desensitization of Diffraction Grating Based Spectrometer Devices" (Attorney Docket 34013-33), filed

November 28,2000; and U. S. Patent Application 09/724,604, entitled"Athermalization and Pressure Desensitization of Diffraction Grating Based Spectrometer Devices" (Attorney Docket 34013-47), filed November 28,2000.

BACKGROUND OF THE PRESENT INVENTION Field of the Invention The present invention relates generally to spectrometry, and more specifically, an optical performance monitor for monitoring wavelength division multiplexed optical signals on a fiber optic network.

Description of the Related Art The telecommunications industry has grown significantly in recent years due to developments in technology, including the Internet, e-mail, cellular telephones, and fax machines. These technologies have become affordable to the average consumer such that the volume of traffic on telecommunications networks has grown significantly. Furthermore, as the Internet has evolved, more sophisticated applications have increased data volume being communicated across telecommunications networks.

To accommodate the increased data volume, the telecommunications network infrastructure has been evolving to increase the bandwidth of the telecommunications network. Fiber optic networks that carry wavelength division multiplexed optical signals or channels provide for significantly increased data for the high volume of traffic. The wavelength division multiplexed information is composed of narrowband optical channels. The wavelength division multiplexed optical channels are time division multiplexed for communicating information, including voice and data. Contemporary optical networks can include forty or more narrowband optical channels on a single fiber and each narrowband

optical channel can carry many thousands of simultaneous telephone conversations or data transmissions, for example. A narrowband optical channel is defined as being a narrowband optical signal, carrying modulated data.

These channels are typically at optical frequencies defined by the International Telecommunication Union (ITU) grid.

An important component of the fiber optic networks is an optical performance monitor (OPM) for monitoring the performance of the optical system. The OPM is essentially a smart spectrometer that provides a system operator the ability to monitor the performance of the individual narrowband optical channels. The optical performance monitor may measure the following metrics: power level, center wavelength, optical signal-to-noise ratio (OSNR), interference between channels such as crosstalk, and laser drift. By monitoring these metrics, the optical network operator can identify and correct problems in the optical network.

The OPM generally includes a dispersion engine and an optical sensor. The dispersion engine may include lenses and a dispersion device, such as a diffraction grating.

The dispersion device diffracts the multiplexed optical signal into its component narrowband optical signals, which are diffracted at angles as a function of the wavelength of each narrowband optical signal. Each narrowband optical signal forms a spot that is focused at distinct locations as a function of wavelength along the optical sensor.

Finesse is a metric used in spectrometer design to measure optical spectral dynamic range of the instrument.

Finesse equals free spectral range divided by spectral resolution. Free spectral range is the range of wavelengths in a given spectral order for which superposition of light from adjacent orders does not

occur. Spectral resolution is the minimum wavelength difference between which two wavelengths that can be resolved unambiguously. The width of detector array elements (i. e., pixels) of the optical sensor are generally larger than the optically resolvable spot size of the spectrometer system.

Optical spectral resolution is typically optimized by focusing the output beam to a minimum size onto the optical detector. However, utilizing a focused spot that is too small limits center wavelength resolution to the detector element size of the optical detector array because each spot is focused within a single detector element. Because each spot is focused on a single detector, centroid computations cannot be used to obtain sub-pixel knowledge of center of energy distribution and center wavelength resolution is limited to the detector element size of the optical detector array. The measured uncertainty of spot position is thus limited by the detector element size.

SUMMARY OF THE INVENTION To improve center wavelength resolution of the OPM, the spot size (s) formed on the optical detector array are optimized by increasing the spot size (s). By increasing or magnifying the spot size (s) formed on the optical detector array, optical spectral resolution is degraded, but center wavelength resolution is improved because centroid accuracy is improved. Magnification techniques may be used to increase the size of the spot while maintaining its quality. However, determining the magnification to properly adjust the spot diameter to be focused on a given optical detector array is complex because a properly sized spot is based on a number of factors, such as expected optical signal-to-noise ratio (OSNR) and intrinsic detector noise, for example.

Additionally, defocusing techniques may be used to

increase the size of the spot, but are less desirable than magnification for system performance reasons.

One embodiment according to the principles of the present invention is as follows. A monitor device and method for monitoring operating conditions of a wavelength division multiplexed optical signal. The monitor device includes an optical engine for receiving the multiplexed optical signal and generating a plurality of demultiplexed optical signals. An array of optical detectors is disposed to receive the demultiplexed optical signals from the optical engine. The demultiplexed optical signals form spots on the array of optical detectors. At least one spot is wider than a center-to-center dimension spanning two optical detector elements in the array of optical detectors.

A more complete appreciation of the principles of the present invention and the scope thereof can be obtained from the accompanying drawings which are briefly summarized below, the following detailed description of the presently-preferred embodiments of the invention, and the appended claims.

BRIEF DESCRIPTION OF THE DRAWINGS FIGURE 1 is a block diagram illustration of an optical engine and optical detector array as utilized in an optical performance monitor; FIGURES 2A-2H are exemplary optical paths of the optical performance monitor according to FIGURE 1; FIGURE 3 is a graph showing the relative x-axis and y-axis positions of an input fiber optic line, a fold mirror, and an optical detector array of the optical performance monitor according to FIGURE 2A; FIGURE 4 is a more detailed exemplary optical path of the optical performance monitor according to FIGURE 2A;

FIGURE 5 is an exemplary graph of positional error versus spot size based on a three-pixel centroid model calculation according to FIGURE 1; FIGURE 6 is an exemplary graph of positional error versus spot size based on a five-pixel centroid model calculation according to FIGURE 1; FIGURE 7 is an exemplary optical network having an optical performance monitor for monitoring system performance along the fiber optic lines according to FIGURE 1 ; and FIGURE 8 is a representative graph showing relative insertion loss for each narrowband signal measured by the optical performance monitor according to FIGURE 1.

DETAILED DESCRIPTION OF THE DRAWINGS The principles of the present invention will now be described more fully hereinafter with reference to the accompanying drawings, in which preferred embodiments of the invention are shown. This invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.

An optical performance monitor (OPM) is a specialized spectrometer that is used by fiber optic network operators to monitor the operation of a fiber optic network. The OPM allows the fiber optic network operator to monitor power levels, center wavelengths, optical signal-to-noise ratio (OSNR), interference, and laser drift, for example, for each narrowband optical channel of a wavelength division multiplexed optical signal carried on the fiber optic network. The OPM is coupled to a fiber optic transmission line via an optical splitter so that wavelength division multiplexed optical signals are input into the OPM along an input fiber optic line. The OPM

demultiplexes the wavelength division multiplexed (WDM) signals into demultiplexed optical signals using an optical engine. The optical engine may include lenses and a dispersion device, for example. It should be understood, however, that many other embodiments of the optical engine are capable of adhering to the principles of the present invention. The optical engine applies the demultiplexed optical signals to an optical detector composed of an array of optical detector elements.

The fiber optic transmission line is connected to an input port of the OPM and typically has a core diameter of approximately 10 micrometers (pm). Each optical detector element within the array generally has a diameter that is larger than 1ohm, such as 25pm or 50Am. If a narrowband optical signal having the diameter of the input optical fiber is applied to the array of optical detector elements, each having a width of 50Am, there would be a 50Am uncertainty in focus spot location. The center wavelength measurement resolution is limited by the linear dispersion produced by the dispersion device and the spacings of optical detector elements (i. e., pixel spacings).

To improve the center wavelength measurement resolution of the OPM, the spot size of the narrowband optical signals incident on the optical detector may be increased using a"spot sizing"technique. The spot sizing technique may include either a magnification or defocusing technique While a high finesse means that the spectrometer can resolve fine structure in the spectrum under observation, the principals of the present invention intentionally reduce the finesse or optical spectral resolution of the spectrometer to improve center wavelength measurement resolution. The center wavelength measurement of the OPM is increased as centroid (first moment) techniques are used, and accuracy is improved due to sub-pixel center of energy distribution determination.

Note, however, that reducing optical spectral resolution of the OPM is counter-intuitive as, in general, a high

finesse is a metric that is desired for spectrometers.

The reduction in finesse is performed due to the given limits in minimum detector size of available infrared optical detector arrays.

As will be shown hereinafter, the size of a spot to apply to the optical detector is non-trivial to determine as several system performance factors, such as noise, optical signal-to-noise ratio (OSNR), energy density profile, and spot intra-pixel position, for example, should be considered. Since optical channels or wavelengths have a known minimum spacing, the dispersion can be chosen so that spots can be sized without the loss of channel-to-channel resolution. Thus, optical spectral resolution can be traded-off for carrier center-wavelength measurement resolution in wavelength division multiplexed systems.

FIGURE 1 is a representative block diagram of an optics path 100 for an optical performance monitor (OPM).

In a fiber optic network, a fiber optic line 105 carries multiplexed wavelength division multiplexed optical signal having optical wavelengths X-. \. An optical splitter or tap coupler 110 is coupled to the fiber optic line 105.

The tap coupler 110 taps power from the multiplexed optical signals traveling through the fiber optic line 105 via the input fiber optic line 112. In the present embodiment, less than about 1% of the power of the multiplexed optical signals traveling through the fiber optic line 105 may be tapped. The amount of power being tapped, however, is not considered essential to the principles of the present invention.

An optical engine 115 receives the tapped multiplexed optical signal from the optical splitter 110 via the input fiber optic line 112. The optical engine 115 may include lenses and a dispersion device. Alternatively, a configuration of the optical engine 115 may include no lenses. The dispersion device may be a reflective diffraction grating, a transmissive diffraction grating, fiber grating or a prism. The dispersion device separates

the multiplexed optical signals into their component narrowband optical signals. Each narrowband optical signal is applied to an optical detector array 120 composed of a plurality of optical elements or pixels 120a-120i. The optical detector array 120 may be composed of indium gallium arsenide (InGaAs), which provides for low noise and infrared-sensitivity. It should be understood that solid-state photocathode or thermal optical detectors having different compositions (e. g., germanium) can be utilized for the optical detector array 120.

Further referring to FIGURE 1, a detection of each wavelength using the optical detector array 120 is indicated by spots 122. The optical detector array 120 may have 256 pixels or detector (photodiode) elements that are 20ohm high (lp) and 50pm wide (wp), where the center- to-center dimension or spacing spanning two (adjacent) detector elements is 5ohm. The center-to-center dimension defines pitch of the detector elements. Note that the height of the detector elements may be longer or shorter, but should be sufficiently longer than the diameter of the spot 122. As shown, six detector elements have been allocated for each 0.8 nm of optical spectral range. In other words, the dispersion means 115 has been arranged to create 30OAm of spot deflection at the optical detector array 120 for a 0.8 nm wavelength shift at the optical splitter 110. Note that 0.8 nm has been chosen in this example because it approximately equals 100 GHz at 1550 nanometers, typical of the optical carrier separation in wavelength division multiplexing communication systems.

Such fiber optic communication systems generally operate at the optical C-band (approximately 1520nm to 1566nm) and L-band (approximately 1560nm to 1610nm).

Shown on the optical detector array 120 are two spots 122 formed by two independent narrowband signals (i. e., demultiplexed from the multiplexed optical signal). The spots 122 are 100ym in diameter (du) and are spaced center- to-center by a distance (D) of 300ym (D=300. m = 100 GHz

(0.8nm AN)). The spots are defined by the diameter where the irradiance drops to 1/e2 of the peak of each spot formed by each narrowband optical signal. The spot size is defined at the 1/e2 point of a Gaussian energy distribution, which encompasses 86. 5% of the power of the Gaussian spot. The spacings between spots generally increase between successive wavelengths as the equation that defines the diffraction grating dictates such an increase in spacings as is understood in the art.

If the lOjUm spot received by the OPM from the input fiber optic line 112 is focused onto the optical detector array 120 without spot sizing, the spot would generally fall on only one optical detector element. Thus, the wavelength measurement resolution would be 0.133nm, as limited by the pixel size for the optical detector array 120, which has each pixel being 50Am wide. However, if the spot size is intentionally increased, or sized, to cover more than one pixel, the ratio of power (represented as photo-current) falling on adjacent pixels can be used to measure spot location finer than the pixel dimensions.

This is known as a sub-pixel measurement.

Higher quality sub-pixel position measurement techniques use more than two pixels to calculate the center of energy of a spot. One such calculation is a centroid calculation. The centroid calculation may calculate the center of energy using the following equation: spot_ center = where a is intensity at pixel n-1, b is intensity at pixel n, and c is intensity at pixel n+1. The centroid calculation can be thought of as power-weighted mean ordinate. Other methods, such as curve fitting, are available for obtaining sub-pixel spot location. Position measurements smaller than 0.1 times the resolution provided by a single pixel are readily obtained vielding

0.0133nm versus 0.133nm wavelength resolution in the present example.

An alternative to the magnification spot sizing technique is to increase the pixel count and reduce the detector element spacing of the OPM. However, detector array sizes above 512 pixels are not practical due to the expensive hybrid construction of optical detector arrays.

The optical spectral range required of the OPM is determined by the carrier wavelength range of deployed WDM systems, and hence, for a given optical system, the detector overall length is fixed.

Enlarging the spot size reduces the optical spectral resolution of the spectrometer. However, since the narrowband optical channels of a WDM system have a known minimum spacing, the dispersion can be chosen so that spots can be sized without the loss of carrier-to-carrier resolution. Thus, optical spectral resolution can be traded-off for carrier centered-wavelength measurement resolution for optimum deployment of wavelength division multiplexing applications.

FIGURE 2A is a block diagram representation of the dispersion engine or optics path 100a of the OPM. Input fiber optic line 112 carries multiplexed optical signals that have been tapped from a fiber optic line of a fiber optic network (not shown). The diameter of the core of the input fiber optic line 112 is approximately 1ohm.

A focusing lens 210a receives the output of the input fiber optic line 112 and focuses the multiplexed optical signal from the input fiber optic line 112 to an intermediate focal plane 215a. The focusing lens 210a may be a GRIN lens as is known in the art. The optics path is centered about the z-axis and the intermediate focal plane is located at the y-axis. The ratio of focal distances from the end of the input fiber optic line 112 to the focusing lens 210a (L2) and the focusing lens 210a to the intermediate focal plane 215a (L1) creates a magnification of the spot seen at the end of the input fiber optic line 112 (i. e., magnification = L1/L2). For example, if the

mode field diameter or core of the input fiber optic line 112 is approximately 10ym and the distance (L1) from the focusing lens 210a to the intermediate focal plane 215a is six times the distance (L) from the end of the input fiber optic line 112 to the focusing lens 210a, then the intermediate image of the spot at the intermediate focal plane 215a is about 60ym in diameter. The magnification factor selected is based upon system design factors, such as the optical signal-to-noise ratio.

The intermediate image at the intermediate focal plane 215a is collimated by a collimating telephoto lens 240, which may be a Cooke triple as known in the art, onto a diffraction grating 245 that is reflective and oriented at a near-Littrow condition relative to the wavelength division multiplexed optical signal incident from the input fiber optic line 112. Alternatively, the diffraction grating 245 may not be set at a near-Littrow condition. The diffraction grating 245 may be planar or curved. The multiplexed optical signals are diffracted by the diffraction grating 245 and the narrowband optical signals that comprise the multiplexed optical signal are diffracted as a function of wavelength. The narrowband optical signals are focused back at the intermediate focal plane 215a by the collimating telephoto lens 240.

The spots formed back at the intermediate focal plane 215a are focused along the y-axis of the intermediate focal plane 215a as a function of the carrier wavelength of the narrowband optical signals. If the beam formed by the focusing lens 210a is aimed below the central axis of the collimating telephoto lens 240, then the beams reflecting from the diffraction grating 245 are re-formed by the collimating telephoto lens 240 above its central axis.

A fold mirror 230 deflects the narrowband optical beams into a detector array 120. The fold mirror 230 is placed as an optical design choice to redirect the narrowband optical signals at the optical detector array 120.

The configuration of the optics path 100a achieves spot sizing by magnifying the multiplexed optical signal at the end of the input fiber optic line 112. Spot sizing, according to the principles of the present invention, may alternatively be achieved by using a defocusing technique. It should be understood that other embodiments of the optics path 100a are possible without departing from the principles of the present invention as presented in FIGURES 2B-2H.

Figure 2B is an alternate embodiment of an optics path 100b of the OPM. In this configuration, the input fiber optic line 112 is placed at a location where the focusing lens 210a creates a magnified virtual focal spot 215b behind the end of the input fiber optic line 112.

This virtual focal spot is re-imaged by the optical engine, including the collimating telephoto lens 240 and the diffraction grating 245, back to the intermediate focal plane 215b. The fold mirror 230 is again placed at the intermediate focal plane 215b to reflect the narrowband optical beam onto the optical detector array 120. Focus spot magnification is achieved as set by the ratio of L1 to L2- Figure 2C is another embodiment of an optics path 100c of the OPM. The input fiber optic line 112, which carries the wavelength division multiplexed optical signal, is optically coupled to a collimating lens assembly 250 having a focal length fl. The collimating lens assembly 250 is disposed to transmit the collimated wavelength division multiplexed optical signal onto the diffraction grating 245. The diffraction grating reflects/diffracts the collimated wavelength division multiplexed optical signal as narrowband optical signals onto a focusing lens 255 having a focal length f2. The optical sensor 120 composed of an array of detector elements 120a-120n is located at the y-axis and receives the narrowband optical signals as spots 122. The spots 122 are spaced according to FIGURE 1. Focus spot

magnification of the optics path is created by the ratio of the two focal lengths (i. e., magnification = f2/fl).

Figure 2D is another embodiment of an optics path 100d of the OPM. The input fiber optic line 112, which carries the wavelength division multiplexed optical signal, is optically coupled to a collimating lens assembly 250 having a focal length fl. The collimating lens assembly 250 is disposed to transmit the collimated wavelength division multiplexed optical signal onto a prism 260. The prism 260 disperses the collimated wavelength division multiplexed optical signal into narrowband optical signals as a function of wavelength.

The narrowband optical signals are focused by a focusing lens 255 having a focal length f2 onto the optical sensor 120 composed of an array of detector elements 120a-120n.

The optical sensor 120 is located at the y-axis and receives the narrowband optical signals as spots 122. The spots 122 are spaced according to FIGURE 1. Focus spot magnification of the optics path is created by the ratio of the two focal lengths (i. e., magnification = f2/fl).

Figure 2E is another embodiment of an optics path 100e of the OPM. The input fiber optic line 112, which carries the wavelength division multiplexed optical signal, is optically coupled to a concave diffraction grating 265. The concave diffraction grating 265 diffracts the wavelength division multiplexed optical signal into narrowband optical signals and focuses the narrowband optical signals onto the optical sensor 120 having an array of detector elements 120a-120n. The distance between the output of the input fiber optic line 112 and the concave diffraction grating 265 is L1. The distance between the concave diffraction grating 265 and the optical sensor 120 is L2. The concave diffraction grating 265 reflects/diffracts the collimated wavelength division multiplexed optical signal as narrowband optical signals onto the optical sensor 120 composed of an array of detector elements 120a-120n located at the y-axis. The optical sensor 120 receives the narrowband optical signals

as spots 122. The spots 122 are spaced according to FIGURE 1. The focus spot magnification of the optics path is created by the ratio of the two distances LI and L2 (i. e., magnification = L2/L1).

Figure 2F is another embodiment of an optics path 100f of the OPM. The input fiber optic line 112, which carries the wavelength division multiplexed optical signal, is optically coupled to a collimating lens assembly 250 having a focal length fl. The collimating lens assembly 250 is disposed to transmit the collimated wavelength division multiplexed optical signal onto the diffraction grating 245. The diffraction grating 245 reflects/diffracts the collimated wavelength division multiplexed optical signal as narrowband optical signals onto a focusing lens assembly 250 having a focal length fl. The optical sensor 120 composed of an array of detector elements 120a-120n is located at the y-axis, which is a distance Ex before the focal point of the focusing lens assembly 250. By placing the optical sensor 120 a distance Ex before the focal plane, the spot is defocused. Alternatively, the optical sensor 120 could be placed Ex after the focal plane of the focusing lens 250 and create the same defocusing results. It should be understood, however, that defocusing provides a reduced opto-mechanical stability. The optical sensor 120 receives the narrowband optical signals as spots 122. The spots 122 are spaced according to FIGURE 1.

Figure 2G is another embodiment of an optics path 100g of the OPM. The input fiber optic line 112, which carries the wavelength division multiplexed optical signal, is optically coupled to a collimating lens assembly 250 having a focal length fl. The collimating lens assembly 250 is disposed to transmit the collimated wavelength division multiplexed optical signal onto the diffraction grating 245. The diffraction grating 245 reflects/diffracts the collimated wavelength division multiplexed optical signal as narrowband optical signals onto a focusing lens 250 having a focal length fl. An

aperture stop 270 is placed before the focusing lens 250 to increase the focus spot size by diffraction.

Alternatively, an aperture stop 270a could be placed after the collimating lens assembly 250 to increase the focus spot size by diffraction. Yet another embodiment could place an aperture stop 270b at the surface of the diffraction grating 245. The optical sensor 120 composed of an array of detector elements 120a-120n is located at the y-axis. The optical sensor 120 receives the narrowband optical signals as spots 122. The spots 122 are spaced according to FIGURE 1. It should be understood that diffraction side lobes may be added and would be an acceptable tradeoff for center wavelength main lobe.

Figure 2H is another embodiment of an optics path 100h of the OPM. The input fiber optic line 112, which carries the wavelength division multiplexed optical signal, is optically coupled to a chirped Bragg fiber grating 275. Each wavelength A1-An of the wavelength division multiplexed optical signal is diffracted by the chirped Bragg fiber grating 275 and passed through a glass block 280 onto the optical sensor 120. Spots 122 are increased as determined by the distance (d) of the glass block 280 from the optical sensor 120. The spots 122 are spaced according to FIGURE 1.

FIGURE 3 is a graph 300 showing the relative x-axis and y-axis locations of the input optical fiber line 112, the fold mirror 330, and the optical detector array 120 according to FIGURE 2A. In general, the input fiber optic line 112 is placed on the x-axis such that the reflection from the diffraction grating (not shown), which is located along the z-axis, has a total of a 3° reflection across the z-axis back to the fold mirror 230. Other reflection angles are possible and it should be understood that multiple input fiber optic lines 112, multiple fold mirrors 230, and multiple optical detector arrays 120 may be utilized in a single optical performance monitor utilizing a single diffraction grating.

FIGURE 4 is a more detailed representation of the optical path 100a of the optical performance monitor. The input fiber optic line 112 carries the multiplexed optical signals, which are comprised of the narrowband optical signals Al-An- An input launch 405 supports the input fiber optic line 112 to maintain a precise and stable position of the input fiber optic line 112. A focusing lens assembly 210b optically coupled to the input fiber optic line 112 focuses the multiplexed optical signal to an intermediate focal plane 215a located at the y-axis. The focusing lens assembly 210b is disposed outside of the input launch 405 within the optical engine 115. Alternatively, the focusing lens assembly could be located within and/or attached to the input launch 405. The multiplexed optical signal is then passed through the collimating lens assembly 240a, which is composed of a lens structure. The collimating lens assembly 240a collimates the multiplexed optical signals, which then enter a prism 410. The prism 410 bends the multiplexed optical beam onto the diffraction grating 245 at a near-Littrow condition.

The diffraction grating 245 diffracts the multiplexed optical signals into their narrowband optical signal components as a function of the wavelength. The narrowband optical signals are reflected back through the prism 410 to the collimating telephoto lens assembly 240a.

The narrowband optical signals are focused by the collimating telephoto lens assembly 240a onto the optical detector array 120 located at the intermediate focal plane 215a. As shown, the optical detector array 120 is located at the intermediate focal plane 215a rather than a fold mirror (as shown in FIGURE 2A) for reflecting the narrowband optical signals onto the optical detector array 120. It should be understood that the fold mirror is a design choice.

It is important to understand the trade-off between spot size and centroid calculation accuracy. An optimum spot size provides accurate, high-resolution, sub-pixel

position measurement, while limiting optical spectral resolution as low as possible. It is possible to find the optimum spot size for a particular centroid by modeling the energy transfer from a Gaussian beam onto the rectangular optical detector elements and further modeling the centroid calculation using the received power from the Gaussian beam on the rectangular optical detector elements. This modeling permits a substantially optimal spot size to be chosen given a set of design constraints.

Determining a spot size that is substantially optimum and one that provides a robust solution for the optical performance monitor is non-trivial. A model to determine the spot size may take into account the center wavelength of the narrowband signals that the optical performance monitor measures, noise that the optical detector array generates, thermal constraints on the optical performance monitor, width of the elements of the rectangular optical detector array, modulation of the data on the optical channels, optical signal-to-noise ratio (OSNR), and a number of other optical, physical, and thermal parameters.

In performing this modeling, there are a number of equations that are utilized to generate the center of the spot based on a centroid calculation and the actual center of the spot. A more detailed equation for the centroid calculation is shown below: (N-l) z TT E V,,, [ (X-jwix"), P,.. I, rol (jw,.) (N-1) ' (2 XCentroia (X, Ptotal S ro, N) = (N-I) E vlixel [ (X-kwlll) lpllallr.] "2 2 where Xcentroid is the calculated centroid x-position, Vpixel is the voltage on a pixel centered at position x for a given total power Ptotal, Wpixel is the size of a pixel along the direction of pixelization, ro is the 1/ (e2) half-width

of the spot, and N is an integer value for setting the pixel numbers as identified by the indices j and k.

Optical signal-to-noise ratio (OSNR) and centroid calculation accuracy are affected by spot size and modulation frequency. OSNR is an optical layer measurement, which can be made by the OPM. OSNR is defined for each channel of a WDM signal as the ratio of the optical signal power associated with a channel to the noise power in that channel. The noise is usually estimated by finding the"valleys"between channels and assuming they represent the noise floor. Linear interpolation between the valleys is used to estimate the noise profile across a channel. Once the total channel power and the noise power are calculated, the OSNR is defined as: p-p OSNR = 10log (Pchannel ~ Pnoise) noise One common source of noise is the gain amplifier (e. g., an erbium doped fiber amplifier (EDFA)). As the spot size grows, the ability to measure the noise floor is degraded. The optimum spot size, balancing OSNR versus centroid accuracy, leads to an optimum l/(e2) spot size between 55Am and 70ym for a 50ym detector width.

FIGURES 5 and 6 provide graphical representations of the difference between a centroid calculated center of a spot and the actual center of the spot. The difference between the centroid calculated and actual center of the spot provides for the estimation error of the spot position. The effect of spot size on the centroid calculation is significant, even in the absence of noise.

Both FIGURES 5 and 6 show the differences based on spot sizes ranging from 40jim to 100jim. As shown, the 40 micrometer spot size creates the largest oscillatory variation between the actual and the centroid calculated center of the spot. The difference in accuracy illustrates the effect of using too narrow of a

calculation region (i. e., too few pixels). Note that the zero position on the horizontal axes of the plots correspond to a perfectly centered spot with respect to the center of the centroid calculation window.

FIGURE 5 is a three pixel centroid calculation window and FIGURE 6 is a five pixel centroid calculation window.

The three pixel centroid calculation uses three pixels of the optical detector array for computing the centroid.

The five pixel centroid window uses five pixels of the optical detector array to compute the centroid. It should be understood that the centroid calculation is one way of determining the center of the spot and that other methods for computing the center of the spot could be utilized.

Furthermore, the spot as presented herein (see, for example, FIGURE 1) is represented as being circular.

However, the spot could be shaped as a top hat, rectangle, oval, or other shape and that a centroid calculation could still be performed to find the center of the shape as focused on the optical detector array 120.

As shown in both FIGURES 5 and 6, the centroid calculation becomes more accurate as the spot is fluffed.

As the spot grows larger than the pixel size, the centroid error increases at either end of the plot for the three pixel centroid as shown in FIGURE 5. This is due to the fact that the signal moves into the pixels outside of the centroid calculation window. In practice, the three pixel centroid is kept properly centered. The five pixel centroid would show similar deviations if the spot were moved further away from the center of the centroid window.

FIGURE 7 shows a block diagram of an exemplary optical network 700. The exemplary optical network 700 includes two end points 705a and 705b. The two end points represent, possibly, two different cities that are in fiber optic communication. At each city, a network operator maintains the fiber optic network equipment. At each end point 705a and 705b, a plurality of fiber optic lines 710a-710n carry narrowband optical signals ranging from Each narrowband optical signal Xi-xi is a time

division multiplexed optical signal and is wavelength division multiplexed by a wavelength division multiplexer 715. The multiplexed narrowband optical signals are inserted into the fiber optic line 105.

The optical splitter 110 extracts and routes a percentage of the power of the multiplexed optical signal.

An optical performance monitor 720 receives the multiplexed optical signal from the input fiber optic line 112. The optics path 100a (FIGURE 5) demultiplexes the multiplexed optical signal into its narrowband optical signals and applies them onto an optical detector array (not shown). The optical detector array 120 (FIGURE 1) converts the power of each narrowband optical signal in parallel and electronics 725 prepare the measurements for a processor 730. The processor 730 performs the centroid and other monitoring calculations. A display 735 is utilized to display the results of the calculations as performed by the processor. Alternatively, the optical performance monitor 720 may be connected to a network and the results of the processing of the narrowband optical signals can be communicated over the network. The network may be a wide area network, such as the Internet, or a local area network. It should be understood that at least the other embodiments of the optics paths 100b-lOOh could be utilized in the OPM 720.

FIGURE 8 is a representative graph 800 showing insertion loss for each narrowband optical channel (i. e., narrowband optical channel) 810a-810n measured by the optical performance monitor. The central wavelengths of each narrowband optical channel are generally predefined, and may correspond with an industry standard, such as the standards set by the International Telecommunication Union (ITU). The insertion loss variation across the narrowband optical signals or channels 810a-810n across the C-band is less than 0.7 decibels (dB). An operator of a fiber optic network can determine if a problem exists on one of the narrowband optical signals 810a-810n by simply inspecting the relative losses via the OPM. Alternatively, the OPM

can be programmed to automatically determine if a"fault" condition on a narrowband optical channel 810a-810n exists.

The previous description is of a preferred embodiment for implementing the invention, and the scope of the invention should not necessarily be limited by this description. The scope of the present invention is instead defined by the following claims.