Page 364 - Maxwell House
P. 364

344                                                                Chapter 6




        (right or left) of H-field. Clearly, each of these waves has the propagation coefficient  =
                                                                                 ±
         0�   meaning that the phase variations are defined by the factor   �− ± � . The simplest
              ±
        nonreciprocal WR phase shifter based on this effect schematically shown in Figure 6.8.3a.



















          Figure 6.8.3 WR phase shifter: a) Schematic drawing, b) RF H-field in WR (view from the
                   top), c) Two-stub design, d) Latching design, e) Twin-toroid design
        The G-shape electromagnet (shown in green with red coil) exerts static magnetic field that can
        be varied by changing up and down the current   from power supply (not shown). The force
                                                
        lines of bias field H0 are depicted in green. The thin longitudinal ferrite stub is located in the
        WR cross section where H-field (blue dot-line with arrows) of TE10-mode has CP, i.e.  ()
                                                                                +
        in Figure 6.8.3b, for example. Previously in Section 6.6.4 of this chapter, we learned that it
                                 −1
                           ⁄
        happens  when  = ( )sin (/2). If so, the forward  wave gets the phase shift   =
                     0
                                                                                +
         �     while the backward CP wave in the same WR cross section being polarized   ()
         0    +                                                     −
                                             , i.e.   ≠  .  Here  is the ferrite stub length
        gets the different phase shift   =  �  −  +  −
                                −
                                      0
        and      as  well as     are some equivalent  constants  of  ferrite  medium filling WR only
               ±
        partially. The best way to estimate the real phase shift and maximize its value is a numerical
        simulation because the presents of ferrite stub itself alters more or less TE10-mode field structure
        making it really complicated and obstructs the finding the optimal cross section with CP.
        Note that the phase shift  is  almost doubled  if  we put two ferrite stubs  symmetrically  and
        magnetize them in opposite directions by two electromagnets as Figure 6.8.3c demonstrates. It
        turns out that two  heavy  and  bulky  magnets  might be replaced  with  a  single  magnetizing
        conductor (see Figure 6.8.3d) passing through and carrying relatively large DC-current. To
        diminish the current magnitude, the two-plate stub should be replaced with single- or twin-
        toroid  assembly  thereby  creating the closed  magnetic path and  minimizing  the bias field
        scattering. The dielectric spacer regulates the EM energy density in ferrites and optimizes the
        phase shifter performance. Even more, the toroids produced from ferrite with magnetic memory
        are able to keep or “remember” their state of magnetization in the absence of external source.
        As such, they should be magnetized and re-magnetized by current pulses of different polarity
        or magnitude. Just note that WR and coaxial ferrite phase shifters are superior in average and
        pulse power handling than similar semiconductor ones, but they certainly lose the competition
        in weight and switching time. The nonreciprocal phase shifters are integral parts of so-called
   359   360   361   362   363   364   365   366   367   368   369