loop that is frequency controlled. When the free-running frequency of the VCO drifts, its closed loop frequency does not shift, which is proper. However, the loop bias voltage needed to maintain the correct frequency will change to compensate for the drift. This will reflect itself directly as a phase-shift change in the path through the phaselock loop. Work should be undertaken to look into the effects of aging on any system performance. All contractors should examine the aging tolerances of their systems. Separation of Steering and Phase Control The one sigma error of the attitude control system is several times larger than the required steering accuracy. The gimbal angle steering error could be reduced to acceptable levels by using a monopulse aiming system. A secure coded transmitter on the ground would be received by a line of a few horns dispersed across the face of the spacetenna. The baseline should be roughly 200 m. This link margin should allow a monopulse improvement factor of about 50, plus anti-jamming capability. The actual power module phasing could then be performed independently by a system that only needs to keep the surface of the module phase center flat. Thus, beam steering would be accomplished with one system, and phase control and surface flattening is achieved with another. The pointing of the beam in the axis normal to the gimbal axis may also need some improvement. For this another monopulse set may be used. The monopulse outputs could be used either to control the phase of the power modules, or the output could be used to drive small correction drive motors or thrusters. If the two functions of beam steering and phase control could be thus separated a wider range of options would be available for the choice of the phase control system. Phase and Amplitude Error Estimation The SPS spacetenna array of power modules will have a statistical error distribution over each sub-array in the system. However, the phase at the center of each sub-array will have a different statistical error distribution. A probabilistic random error theory should be developed to analyze this problem so that accurate tolerance values can be developed. Monte Carlo computer simulations are risky for two reasons. First an array with over a hundred thousand elements cannot be realistically simulated. Second, many computer runs must be made to establish the probability distribution function value. With only a few dozen runs, only smoothed values are obtained. Monte Carlo should be used as an adjunct to the probabilistic formulas. Ear Sidelohes The beam patterns generated thus far may be somewhat optimistic in the far-out sidelobes. Excluded from the analysis of the radiation pattern has been the effect of position errors in the phase centers of the subarrays and modules. In particular, the effect of these errors on the distant sidelobes has not been calculated. There is some angle away from the main lobe beyond which the sidelobe pattern begins to deviate from the design pattern and tends toward the sidelobe properties of a random array. This angle should be calculated, and the effect determined. If the subarrays are rigid and ideally manufactured, and if all the in-plane distortion is associated with their relative positions, then the average sidelobe level beyond the angle mentioned above
RkJQdWJsaXNoZXIy MTU5NjU0Mg==