Page 399 - Maxwell House
P. 399

DISCONTINUITY IN FEED LINES                                             379



            analysis with normalized or denormalized - and -matrix since their elements have a tendency
            to take an infinite values around the unknown in advance circuit resonances. We hope that the
            reader will share our opinion that S-matrix provides the most natural and physically clear
            description of microwave network.  Nevertheless note that for lossless network  ℛ� � = 0
                                                                                 
            and ℛ� � = 0, ∀, , i.e. all elements of  these  matrices are purely imaginary. When two
                  
            networks are connected in series  =  +   and in parallel  =  +  . Both features
                                                                      1
                                                                          2
                                                  2
                                             1
                                         Σ
                                                                 Σ
            lets simplify the control of computer algorithms.
            ABCD-matrix.  Note  only  one  more  “beast”  named  ABCD-matrix  or transmission/cascade
            matrix relating the input and output voltages. For example, in case of 2-port network
                                           1       2
                                         � � = �     � � �
                                           1       2
            This  matrix is an analog of  T-matrix  being discussed  above and possesses the  same
            multiplication feature: the resultant ABCD-matrix of the cascade network presented in Figure
            7.3.3a is the product of matrices of networks in a cascade like (7.23).
            Meanwhile, there are several other matrix representations of network mostly inherited from the
            conventional circuit theory like hybrid H- and G-matrices. For example, H-matrix simplifies
            the analysis and synthesis of networks  whose input ports are connected in series while the
            outputs are in parallel. It links the column vectors ( ,  )  and ( ,  ) . G-matrices is often
                                                           
                                                                      
                                                         2
                                                                    2
                                                                 1
                                                       1
            selected in case of parallel-series configurations and  =  . More detailed discussion of this
                                                           −1
            matrix and its element physical interpretations are beyond the scope of this course. Just note
            that all of them can be expressed in term of the S-matrix. We refer the reader to relevant
            literature [2] for more information.
            7.3.6   S-Matrix of Complex Network
            In fact, the  majority of complex networks  can  be decomposed  into  some  combination  of
            interconnected  elementary  networks  of  the  much  more  simple structure.  If so, each of
            elementary network may be described by given Z- or S-matrix or any other matrices being
            considered above.  As an example assume that we intended to develop a coaxial low-pass filter
            consisting from multiple LC-sells as Figure 7.3.4 demonstrates.
















                      Figure 7.3.4 Low-pass filter equivalent circuit and its coaxial layout
            Looking back at Table 7.2 we should indicate the coaxial layout corresponding to each filter
            element as depicted in Figure 7.3.4, its equivalent circuit, and scattering matrix like (7.15). Then
            the question prompts how to find S-matrix of the whole filter after all elements are connected.
   394   395   396   397   398   399   400   401   402   403   404