In both the geometrical optics (GO) and physical optics (PO) formulations, the ultimate goal is to determine equivalent currents, which can then be integrated to obtain the far-field patterns, a process well described in the literature on aperture antennas (20). We focus our attention on the assumptions and approximations inherent in each of these formulations, as well as on their intrinsic differences.

The GO technique yields the aperture fields, assuming equal angles of incidence and reflection. The far-field patterns can then be calculated using a Fourier transformation directly, which is equivalent to obtaining equivalent currents and then integrating, as described later in this subsection. With the use of image theory, it is necessary to know only the electric field distribution over the reflector projected aperture, Er, which is computed from the incident electric field E{ (i. e., the feed radiation), with (20)

Er = 2(n • E;)n — E; (31)

where A is the unit vector normal to the surface; see Eq. (5). Equation (31) assumes that at the point of reflection the reflector is planar and perfectly conducting. In addition, the incident wave from the feed antenna is treated locally as a

plane wave. These same assumptions are also used by the PO

technique to determine the surface currents, Js, over the reflector as follows:

Js = 2n x H; (32)

where H is the incident magnetic field from the feed antenna and can be computed from Eq. (14), recalling that in the far field H = (r X E)/^ (where ^ is the free-space characteristic impedance). The PO approximation assumes that currents exist only over the side of the reflector directly illuminated by the feed antenna.

Table 6. GBT Dual Offset Reflector Configuration and Computed Performance Values

Main Reflector Configuration

Shape: Offset paraboloid Projected diameter D: 100 m Parent reference diameter Dp: 208 m Focal length F: 60 m Offset of reflector center, H: 54 m Angle fi, 5.58°

Subreflector Configuration

Shape: Offset ellipsoid Projected height DS: 7.55 m Parameter c of ellipse: 5.9855 m Parameter fS of ellipse: 5.3542 m Eccentricity e: 0.5278

Feed Configuration (On Focus; GRASP Calculation)

Polarization: Linear (xf)

Pattern shape: Gaussian, Eqs. (14) and (15)

Gain Gf: 21.31 dBi 10 dB beamwidth: 30°

Angle a: 17.91°

Angle y: 12.33°

System Performance (GRASP Calculation)

Gain G: 82.83 dBi

Cross-polarization (XPOL) level: —43.01 dB Sidelobe level (SLL): —22.56 dB Aperture efficiency eap: 77.76%

The far-field pattern can then be determined by summing the individual contributions of each current point over the surface, taking into account the different amplitudes and phases due to the excitation and spatial location. Antenna theory shows that a unit point source of current radiates a spherical wave, which is normally referred to as the free — space Green’s function (e—jkr/4wr); see Ref. 20 for further details. In the limit as the current distribution becomes continuous, such as the one given by Eq. (32), the weighted sum of spherical waves becomes an integral, yielding the radiated patterns.

Note that the integration process for obtaining the patterns is the same as the one employed by the GO technique, given that once the aperture distribution is determined from Eq. (31), equivalent currents can then be obtained and integrated over the reflector aperture. This process is equivalent to computing the Fourier transform of the aperture distribution given in Eq. (31). One difference between GO and PO is that PO currents are determined over the reflector curved surface and the GO equivalent currents over the planar projected aperture, with the latter already in a format more appropriate for integration through a Fourier transform. However, the use of a Jacobian transformation (3,4) maps the PO currents over the reflector curved surface to the planar aperture, yielding the possibility of also using Fourier transformations for performing the integration. Analytical integration is only possible for symmetrical reflectors (4,8), and numerical techniques are normally required to evaluate offset reflectors, as discussed in the next subsection.

The PO formulation is generally considered more accurate than GO to evaluate offset reflectors, especially if XPOL assessment is a main concern. However, pattern accuracy as determined from both techniques degrades beyond the main beam and near-in sidelobes. The pattern in the far-out region is dominated by diffraction effects, especially scattering from the reflector and/or subreflector edges. This is taken into account by augmenting GO with the geometrical theory of diffraction (GTD) or augmenting PO with edge currents through the physical theory of diffraction (PTD); see Refs. 20 and 30 for details. However, the near-in pattern region is, most of the time, the region of interest when analyzing high — gain antennas such as the reflector antennas considered here.