a buck-boost coil which in turn changed the magnetic field impressed on the magnetron and therefore the current flow through the magnetron itself. The modification of the magnetron with a buck-boost coil is shown in Fig. 4. It is well known that the free running frequency of the magnetron is a function of the anode current. It is also known that when the free running frequency of the magnetron coincides with the drive signal, there will be zero phase shift of the magnetron's output from the drive signal. Therefore, when the tube is tuned exactly to the frequency of the drive signal by an appropriate change in the anode current, the resulting phase shift between input and output will be zero (3,4). The experimental results of operating the magnetron directional amplifier in this manner were amazing in terms of keeping the phase shift through the device to a low value even though the drive frequency was being varied over a wide frequency range and a very high nominal power gain of 32 dB was being achieved. Figure 5 shows the results in contrast with that of the same magnetron directional amplifier without phase lock. Inclusion of lead and lag compensation in the feedback loop which would improve performance was not used. No instabilities were exhibited over the range of available system gain. The output spectrum remained of high quality at all times but this was expected because the spectrum of the free running oscillator is of high quality. Obtaining simultaneously such high gain and phase control motivated looking at its adaptation to the SPS problem which was basically that of feeding two high gain amplifiers into a Magic T with the two outputs so well matched in phase and amplitude that the combined power flows almost completely to the radiating antenna and very little toward the drive source. If it could be adapted to the SPS, it would allow the use of a low level drive source of the order of a few watts and a ferrite circulator that does not have to handle more than a few watts of reflected power and that could be mounted on the front surface of the slotted waveguide. However, some substantial modifications are necessary in the adaptation process. Although the phase locked loop circuit described in Fig. 3 had given both phase
RkJQdWJsaXNoZXIy MTU5NjU0Mg==