A new software tool is developed to evaluate the

A new software tool is developed to evaluate the measured/simulated transmission characteristics of optical multiplexers/demultiplexers. D. Seyringer* and P. Schmid University of Applied Sciences, Research Centre for Microtechnology, Hochschulstraße 1, 6850 Dornbirn, Austria ABSTRACT A new software tool, called AWG-Analyzer, is developed to evaluate the simulated/measured transmission characteristics of optical multiplexers/demultiplexers based on arrayed waveguide gratings (AWG). The output of the calculation is a set of the transmission parameters like: non-uniformity, adjacent channel crosstalk, non-adjacent channel crosstalk, background crosstalk, insertion loss, polarisation dependent loss (PDL), etc. calculated for each output channel first and then for the whole AWG - the worst case value of each parameter over all the output channels. This set of the parameters is then taken as the AWG specification. The parameters are calculated for a particular channel bandwidth (also known as the channel passband or ITU passband), that is also an input parameter for the calculations. Additionally, the developed software tool, having a user friendly interface, offers the help where all calculated transmission parameters are explained and exactly defined. The tool also includes a brief overview about AWG functionality with a small animation and the information about various AWG types (CWDM and DWDM AWGs, Colourless AWGs). Keywords: arrayed waveguide gratings, CWDM, DWDM, AWG transmission parameters, transmission characteristics, software-aided evaluation

1. INTRODUCTION Nowadays the steadily increasing data volume in communication networks is driven by a rapid proliferation of home-based and business computers, storage capacities, processing capabilities, and the extensive availability of Internet. The challenge is to transmit high data volumes in short periods of time over high distances as lossless as possible. The optical technologies offer several advantages compared to others1 which enable larger and more flexible bandwidths. In optical data transmission this is provided by the use of wavelength division multiplexing (WDM) technique. This technique enables that several optical signals, each propagating with a specific wavelength, can be simultaneously transferred over one single optical fiber without any interferences. Due to the fact that optical fibers do not let light of every wavelength propagate at the same loss-level only specific wavelength ranges also called operating windows can be used for a low-loss optical data transmission. Moreover it should be possible for the demultiplexer on the receiver side to separate the optical signals, also denoted as wavelength spectrum, combined by the multiplexer on the sender side. To achieve this a standard called ITU-T Recommendation2 defines exact wavelength values and spacings to be used. Each of these wavelengths can be seen as a separate channel usable for data transmission therefore the ITU definitions are also denoted as ITU-channel center wavelength and channel spacing. In DWDM applications special optical multiplexers/demultiplexers based on arrayed waveguide gratings are used to combine respectively separate the single wavelengths on or from the optical fiber. The quality of such a multiplexing or demultiplexing process can be determined by the evaluation of the transmission characteristics obtained from simulation or measurement. The evaluation is done for each channel of the transmission characteristics first which provides a set of parameters each channel denoted as transmission parameters. The worst case values over all channels then provide a set of parameters denoted as AWG transmission parameters. These AWG transmission parameters state the performance of the simulated or measured AWG. * [email protected], www.fhv.at

Optical Design and Engineering IV, edited by Laurent Mazuray, Rolf Wartmann, Andrew Wood, Jean-Luc M. Tissot, Jeffrey M. Raynor, Proc. of SPIE Vol. 8167, 81671D · © 2011 SPIE · CCC code: 0277-786X/11/$18 · doi: 10.1117/12.892462 Proc. of SPIE Vol. 8167 81671D-1

2. MOTIVATION AWG is a key component in the dense wavelength division multiplexing (DWDM) technique. When defining its performance and also determining its suitability for a particular application, a set of transmission parameters must be considered.3 These parameters are extracted by analyzing AWG transmission characteristics - the simulated/measured AWG spectral response for both the transverse electric (TE) and the transverse magnetic (TM) polarization states. While the measurement method adhered by most AWG vendors is the deployment of Mueller Matrix method4, 5 the way that vendors specify the performance of a device from the measured curves differs widely. Additionally to this, it is also important to note that the output from all commercially available AWG design tools is the simulated transmission characteristics only and none of them supports (or supports only partially) the software-aided evaluation of AWG parameter calculation. The only possibility is to manually determine the AWG transmission parameters directly out of the transmission characteristics graph. Major problems of the manual evaluation are that it is very unhandy, time intensive and can be imprecise due to reading errors especially for high channel number, narrow channel spacing or colorless AWGs. As a consequence of the manual evaluation the design, simulation and fabrication of AWGs can become a complicated matter because modifications of layout or fabricated device always require a detailed evaluation of the current simulation or measurement result data. Beside the mentioned disadvantages of the manual evaluation a further problem is that the evaluation of AWGs is not underlying a uniform standard.3 This means that the definitions of the AWG transmission parameters by now are not standardized. This ends in the fact that without enclosed definitions no one can exactly determine the performance of a present AWG excepting the people involved in design respectively manufacturing of that particular AWG. Therefore a sustainable evaluation should always imply clear definitions of how the AWG transmission parameters were determined. In summary there are two major AWG evaluation problems: 1. Manual evaluation is very unhandy, time consuming and can be imprecise due to reading errors. 2. Definitions of the AWG transmission parameters are not standardized. To solve these problems a new software tool is developed which provides the following AWG evaluation solutions: 1. Evaluation of all relevant AWG transmission parameters out of simulated and measured transmission characteristics. 2. Provision of help on each calculated AWG transmission parameter by showing its textual and graphical definition. This paper presents the AWG-Analyzer software tool which is intended to bridge the gap between design, simulation and manufacturing by offering performances which provide powerful support in every stage of the AWG life cycle.

3. AWG TRANSMISSION PARAMETER DEFINITIONS The AWG transmission parameters state the performance of simulated design or fabricated AWG. They can be determined by manual or software-aided evaluation of simulated or measured transmission characteristics. As described in Section 2 a sustainable evaluation only can be performed if the AWG transmission parameters are clearly defined. Table 1 presents all relevant AWG transmission parameters by showing their names, specific abbreviations and units. For each of these AWG transmission parameters a clear textual and graphical definition is determined.6 All transmission parameter calculations implemented in the evaluation algorithm of the AWGAnalyzer software tool are performed according to these definitions. Moreover these textual and graphical definitions are implemented in the software’s help system and therefore available every time the user needs specific information about how the AWG transmission parameters are calculated.

Proc. of SPIE Vol. 8167 81671D-2

Table 1. AWG transmission parameters.

4. SOFTWARE PERFORMANCES The software is intended to be used on windows computers and is written in C (C - C sharp). Due to high demands7 on the graphical user interface the software is developed as a Microsoft Visual C Windows-Forms application.8 The graphical user interface (see Figure 1) is based on a main window which provides file handling options like open and save, and the help system. Each file opened causes the generation of a new file specific window offering 2 different representations of the opened file each located in a separate sub window. The raw data window shows the original content of the opened result file and the diagram window the graphical representation of the transmission characteristics. For starting the evaluation procedure each file specific window provides a button named analyze. The analyze-button causes, after the user was encouraged to set the desired passband value by use of the enter passband window, the evaluation algorithm to calculate the AWG transmission parameters out of the opened transmission characteristics. The parameter window automatically updates its content whenever new evaluation results occur. Via save-button the evaluation results can be stored as a file in portable document format (PDF).

4.1 File Input The transmission characteristics obtained from simulation or measurement are usually available as text files. The AWG-Analyzer software tool is capable to open the following 4 special kinds of text files:


Figure 1. Graphical user interface of AWG-Analyzer software tool.

1. Optiwave WDM-Phasar software tool simulation result files. 2. Apollo APSS software tool simulation result files. 3. Measurement result files from a specific manufacturer. 4. General result files. If simulation or measurement result data files produced by other software tools shall be evaluated the result data can be rearranged to match the predefined general data structure.

4.2 Input File Visualizations The AWG-Analyzer software tool offers 2 different views of an opened files data. Both can be used for a visual pre-evaluation of the input data and post-evaluation of the calculated transmission parameters. 1. Graphical representation of raw data The graphical representation of the raw data is displayed in the diagram window (see Figure 1) within the file specific window by the use of ZedGraph9 charting tool. The diagram shows insertion loss (ordinate) in dB as function of wavelength (abscissa) in nm and offers features like zoom-function, data cursor and save diagram as image function. 2. Textual representation of raw data The content of the original input file is displayed in the raw data window (see Figure 1) within the file specific window. If the graphical representation of the raw data shows any basic problems (e.g. missing channel), the raw data window enables the user to check the original data set. The data has a read only property, changes therefore can only be performed in the original text file.


4.3 User defined passband value The parameter passband (also known as channel pass bandwidth, the channel passband or ITU passband), is the range of wavelengths surrounding the center wavelength for any given channel. Most of the AWG transmission parameters listed below this parameter in Table 1 are calculated within this passband. It usually equates to 25% of the respective channel spacing. However, to be able to calculate the transmission parameters for any passband value this parameter is implemented as a variable input quantity. The user is encouraged to set the desired passband value in the enter passband window (see Figure 2) after pressing the analyze-button. The user is only able to enter a passband value within a data set specific valid input range from 1 to 100% of the channel spacing. The lower limit of the input range is limited by the resolution of the data set which actually is the distance in wavelength between 2 data points. Due to the fact that e.g. for the calculation of the bandwidth at -0.5dB ([email protected]) a left and a right intersection point has to be determined, minimum 3 data points including point of peak transmission are required. The lower passband limit of 9% of channel spacing therefore represents the value which contains at least 3 data points. The upper limit is also dataset specific. It states the maximum value of the passband for which the existence of later on required data points is guaranteed.

Figure 2. Enter passband window including valid input range.

4.4 Calculation of AWG Transmission Parameters AWG-Analyzer software tool is capable to evaluate all AWG transmission parameters stated in Table 1 excepting polarization dependent loss (PDL) and polarization depended wavelength (PDW). Nevertheless PDL and PDW can be manually determined by the help of the software tool without circumstances.6 All channels respectively orders of a channel having peak transmission less than background crosstalk BX plus 25dB are detected as missing channels. The evaluation algorithm works for transmission characteristics containing channel spacings of minimum 0.1nm/12.5GHz. If the user defined passband value exceeds any channels or orders bandwidth at -25 dB from peak transmission the evaluation cannot be performed due to the possibility of not existing passband intersection points. If a missing channel, a narrower channel spacing or a passband exceeding bandwidth at -25dB is detected an error message appears which informs the user that the evaluation cannot be performed. The AWG-Analyzer software tool supports the evaluation of transmission characteristics containing single, multiple or colorless spectrum. 4.4.1 Single spectrum transmission characteristics Single spectrum transmission characteristics only show a single order of the phased array response as can be seen in Figure 3 by the example of the Optiwave WDM-Phasar simulation data of an AWG with 8 channels, 100GHz channel spacing and 122 phased arrayed waveguides denoted as simulated 8Ch100GHz122WG-AWG. For single spectrum transmission characteristics the evaluation of the transmission parameters is performed for each output channel. The AWG transmission parameters then show the worst values over all channels. To track the worst values the channels are numbered starting in the characteristics (see Figure 3) at the right outermost channel with channel 1.


Figure 3. Transmission characteristics containing single spectrum.

4.4.2 Multiple spectrum transmission characteristics Muliple spectrum transmission characteristics show multiple orders of the phased array response with a free spectral range FSR greater than Number of channels (Nr.Channels) times channel spacing (dLambda]) (see Table 1). A multiple spectrum characteristics can be seen in Figure 4 by the example of the manufacturer measurement result data of an AWG with 8 channels, 100GHz channel spacing, and 122 phased arrayed waveguides denoted as measured 8Ch100GHz122WG-AWG. For multiple spectrum transmission characteristics the evaluation of the transmission parameters is performed for each channel per order for each order. The AWG transmission param-

Figure 4. Transmission characteristics containing multiple spectrum.


eters then show worst values over all channels of all orders. To track all evaluated parameters the orders are numbered starting in the characteristics (see Figure 4) at the right outermost order with order 1. 4.4.3 Colorless spectrum transmission characteristics Colorless spectrum transmission characteristics show multiple orders of the phased array response with a FSR equal to Nr.Channels times dLambda. A colorless spectrum can be seen in Figure 5 by the example of the Optiwave WDM-Phasar simulation result of an colorless AWG with 8 channels, 100GHz channel spacing, 62 phased arrayed waveguides denoted as simulated 8Ch100GHz62WG-AWG. For colorless spectrum characteristics the evaluation of all parameters excepting adjacent channel crosstalk AX and non-adjacent channel crosstalk nAX is performed for each channel per order for each order. Because for the evaluation of AX and nAX the connection between the orders has to be considered they are calculated over all channels and orders. The AWG transmission parameters then show worst values over all channels and orders. To track all evaluated parameters the orders are numbered starting in the characteristics (see Figure 5) at the right outermost order with order 1 regardless if it is complete or not. For transmission characteristics containing colorless spectrum with a free spectral range FSR higher than 1.05times the channel spacing times number of channels, the characteristic is evaluated as containing multiple orders and the connection between the orders gets lost.

Figure 5. Transmission characteristics containing colorless spectrum.

4.5 Evaluation parameter visualization The evaluated parameters are displayed in the parameter window (see Figure 1). The parameter window provides 2 different views on the evaluated parameters. Figure 6 shows both the AWG and the all transmission parameters view including switching context menu and mouse hover options on AWG tranmission parameter LambdaCH. 4.5.1 AWG Transmission Parameter View (by default) This view shows a table containing all evaluated transmission parameters, their units, and their associated channels. Left hand side picture in Figure 6 shows the transmission parameters view of the evaluated transmission characteristics of the simulated 8Ch100GHz122WG-AWG.


Figure 6. AWG- and all parameter view available in the parameter window.

The AWG transmission parameter definitions can be obtained by performing a left mouse click on the desired AWG transmission parameter name. This feature works on all parameter names which show the value-tip-text Show Definition (left click) as mouse hovers on the parameter value as can be seen in Figure 6 by the example of AWG transmission parameter channel center wavelength LambdaCh. The AWG transmission parameter definition of LambdaCh appears as can be seen in Figure 7. The transmission parameter view offers the possibility to show the values of a specific transmission parameter over all channels by performing a left mouse click on the specific value. This feature works on all parameter values which show the value-tip-text ”Show All Parameters (left click)” as mouse hovers on the parameter value (see left hand side picture in Figure 6) by the example of AWG transmission parameter channel center wavelength LambdaCh. Figure 8 shows the appearing window. 4.5.2 All Transmission Parameters View By performing a right mouse click on the parameter window the view can be changed by checking the ”Show All Transmission Parameters” item of the context menu shown in Figure 6. The appearing view (see right hand side picture in Figure 6) displays all transmission parameters evaluated by the software tool as transmission parameters over all channels blocks containing channel numbers, values and units of the parameters.

4.6 File Output To summarize the evaluation results AWG-Analyzer software tool provides the generation of an evaluation sheet containing the evaluated transmission parameters and the diagram of the transmission characteristics. Figure 9 shows the evaluation sheet of the simulated 8Ch100GHz122WG-AWG. This evaluation sheet can be


Figure 7. Help on parameter LambdaCh.

Figure 8. Show specific transmission parameter over all channels feature.

saved in portable document format (PDF). The generation of the evaluation sheet is performed by taking the actual content of parameter (AWG transmission parameter view) and diagram window. This enables that an individually adjusted diagram can be inserted by saving the evaluation results.

5. MANUAL EVALUATION As described in the motivation section the manual evaluation is a very unhandy, time intensive and can be also an imprecise issue due to reading errors especially for high channel number, narrow channel spacing or colorless AWGs. Nevertheless the manual evaluation finally provides the transmission parameters which represent the performance of the simulated or measured AWG. The results of this manual evaluation are later on used for a comparison of manual and software-aided evaluation results. The manual evaluation was performed by the example of the simulated 8Ch100GHz122WG-AWG described in Section 4.4.1. The parameters are calculated for each output channel (according to the AWG transmission parameter definitions) separately however since the evaluation procedure is always the same it is shown by the example of channel 1. In order to perform the manual evaluation three details of the graphical representation of the simulated transmission characteristics (see Figure 3) can be taken as shown in Figure 10. Detail 1 can be used for the evaluation of the parameters pIL, B@-05dB, B@-1dB, B@-3dB, LambdaCh, LambdaITU, dLambdaITU, Ch.Spacing, PB, PBu, and IL. Detail 2 was used only for the B@20dB parameter and detail 3 for determination of BX and the maximum transmission each channel in channel 1 which later on is needed for the calculation of AX and nAX. The parameters have to be calculated for each channel to get the deviations pILu and ILu. The three details can be printed to DIN-A4 paper and evaluated by the help of a drawing board. The manual evaluation using the mentioned three details can be performed according to the following steps S1 - S16. (S1) Find point of peak transmission PPT which corresponds to pIL. (S2) Perform the sub-steps S2.1 and S2.2 with index variable i = 0.5, 1, 3, 20. (S2.1) Find left IPidB left and right IPidB right intersection point


Figure 9. AWG-Analyzer evaluation result sheet.

at a -idB drop from PPT. (S2.2) Calculate the width between the intersection points at -idB from PPT. B@idB = IP idB right − IP idB lef t which yield [email protected], B@1dB, B@3dB and B@20dB. (S3) Find center point P3dB center of B@3dB and calculate LambdaCh = IP 3dB center + B@3dB . (S4) Find nearest 2 ITU-channel center wavelength LambdaITU according to the ITU-grid.2 (S5) Calculate deviation to nearest ITU-channel wavelength dLambdaIT U = ABS(LambdaIT U − LambdaCh). (S6) To take a quarter of the ITU-channel spacing2 dLambda yields passband P B = dLambda/4 and find left IPPB left and right IPPB right passband intersection points at LambdaCh − P2B and LambdaCh + P2B . (S7) Find minimum value IPPB min out of IPPB left and IPPB right and calculate P Bu = pIL − IP P B worst (S8) Find worst transmission in passband (here IPPB worst) which corresponds to insertion loss IL of channel 1. (S9) Determine the maximum


Figure 10. Manual evaluation details of channel 1.

transmission points each channel in the passband of channel 1. (S10) Perform the previous steps (S1) - (S9) for each channel and determine worst case values over all channels. This provides the AWG transmission parameters pIL, LambdaCh, LambdaITU, dLambdaITU, dLambda, [email protected], B@1dB, B@3dB, B@20dB, PB, PBu and IL (see Table 1). (S11) Calculate the existing channel spacings between the channel center wavelengths e.g. dLambda i = c LambdaCh i − LambdaCh (i + 1) and dfi = LambdaCawg 2 ∗ dLambdai whereas c is the speed of light, i is the indexed variable for the channel number and LambdaCawg is the central wavelength of the AWG. (S12) Calculate peak insertion loss uniformity pIL through pILu = pIL min − pIL max. (S13) Calculate insertion loss uniformity ILu through ILu = IL min − IL max. (S14) Calculate adjacent channel crosstalk each channel first then determine worst value over all channels. E.g. RAX21 = IL2 − P 21 whereas RAX21 is the right adjacent channel crosstalk of channel 2, IL2 is the insertion loss of channel 2 and P21 is the maximum transmission of channel 2 in the passband of channel 1. (S15) Calculate non-adjacent channel crosstalk each channel first then determine worst value over all channels. E.g. RnAX31 = IL3 − P 31 whereas RnAX31 is a right non adjacent channel crosstalk of channel 3, IL3 is the insertion loss of channel 3, and P31 is the maximum transmission of channel 3 in the passband of channel 1. (S16) Find background crosstalk BX as the median value PmaxMedian between best PmaxBest and worst PmaxWorst maximum over all channels excepting the ones in the region of channel center wavelength +/- channel spacing. Performing steps S11 - S16 finally provides the remaining AWG transmission parameter pILu, ILu, Ax, nAX and BX (see Table 1). Table 2 shows the manual evaluation results of the simulated 8Ch100GHz122WG-AWG over all channels. The marked fields state the AWG transmission parameters which are the worst case values over all channels.


Table 2. All transmission parameters of manual evaluation.

6. SOFTWARE-AIDED EVALUATION The software-aided evaluation was performed by the help of new developed AWG-Analyzer software tool by the same example as used for the manual evaluation (see Section 4.4.1). The evaluated transmission parameters shown in Table 3 are taken out of the parameter window switched to ”all transmission parameters” view (see Figure 6) and rearranged to match the same representation as used for the manually evaluated transmission parameters (see Table 2). The marked fields state the evaluated AWG transmission parameters which are the worst case values over all output channels.

7. DISCUSSION As described in the motivation section a manual evaluation offers several drawbacks as it is very unhandy, time intensive and offers wide range for human mistakes along the evaluation procedure. The manual evaluation of the simulated 8Ch100GHz122WG-AWG took about 8 hours of concentrated evaluation to finally obtain the transmission parameters whereas the software-aided evaluation algorithm requires about 126ms for evaluating the same transmission characteristics. The software evaluation algorithm follows a straight computational line


Table 3. All transmission parameters of software-aided evaluation.

e.g. if there is an intersection point at -20dB from peak transmission there is only one with a specific insertion loss (y-axis) and wavelength (x-axis) component. In manual evaluation the location of this intersection point may depend on format of the paper which determines the basic resolution, orientation of the printed diagram on the paper, thickness of the used pencil, the incidence angle of the light on the paper, parallelism and perpendicularity of the self-made lines, or the actual concentration of the evaluator. Nevertheless if one takes care of those problems the manual evaluation provides reliable results. Table 4 shows the deviations of the manual (see Table 2) and software-aided (see Table 3) evaluation results. To concentrate on the highest deviations over all channels Table 5 summarizes the worst case values of the parameters shown in Table 4. As can be seen in Table 5 parameters in dB-scale show larger deviations than the ones in wavelength-scale. This can be explained by the fact that for the manual evaluation the resolution of the ordinate (in dB) is much lower than the abscissa (in nm). E.g. for manual evaluation detail 1 (see Figure 10) a step of 1 dB on the ordinate has about the same length in mm as 0.023 nm on the abscissa which apparently enables a more accurate reading of wavelength values. Through changing the representation of the deviations to percent of its nominal value (manual evaluation results are taken as nominal values) it appears where the largest deviations are located.


Table 4. Deviations of all manual and software-aided evaluated parameters.

All in all the transmission parameters in Table 4 and 5 only show slight deviations therefore manual and softwareaided evaluation results can be stated as consistent.

8. CONCLUSION The strength of AWG-Analyzer software tool is a fast and comprehensible calculation of the transmission parameters. The exact and extensive evaluation results support that modifications in design match the desired effect more exactly. The provided transmission parameter definitions enable the user to be informed about the implemented evaluation algorithm and how finally the AWG transmission parameters were calculated. The implemented evaluation algorithm is capable to calculate the AWG transmission parameters of transmission characteristics containing single, multiple or colorless spectrum over all channels respectively orders of the channels. Additionally the channels of the calculated transmission parameters are tracked over the whole evaluation procedure to finally provide the AWG transmission parameters in combination with their associated channels.


Table 5. Worst deviations of manual and sofware-aided evaluation over all transmission parameters.

The software tool can be used for the following applications. • As support during the whole design phase by providing the transmission parameters of the successively adapted layouts. This enables the qualitative adaption of the design by the use of solid evaluation results. • Comparison between different design respectively simulation tools in order to find advantages and drawbacks. • Verification of designed AWGs through comparisons between their simulated and measured transmission characteristics. • Implemented help system enables usage for educational and training purposes.

REFERENCES [1] [2] [3] [4] [5] [6]

Keiser, G., [Optical Communications Essentials], McGraw-Hill (2003). ITU, [ITU-T Recommendation G.694.1, Spectral grids for WDM applications: DWDM frequency grid] (2002). Rahman, A., [AWG-Characterization ] (2001). Shurcliff, W. A., [Polarized Light: Production and Use], Harvard University Press (1966). Derickson, D., [Fiber Optic Test and Measurement ], Prentice Hall (1998). Schmid, P., [Development of a software tool to evaluate simulated and measured transmission characteristics of optical multiplexers/demultiplexers based on arrayed waveguide gratings.] (2011). [7] Schmid, P., [Software Requirements Specification - AWG-Analyzer Prototype] (2011). [8] Moessenboeck, H., [Kompaktkurs Csharp 4.0], vol. 3. Auflage, dpunkt.verlag (2009). [9] ZedGraph, [Charting library for .Net applications ].
