Subreflector Developed for Simultaneous S and X-Band from a Single Antenna.
In order to achieve high system G/T's in the dual bands, a dichoric subreflector has been developed as an integral part of its S and X-band antenna system. The S-band feed is located at the prime focus of the main reflector; the X-band feed operates in a Cassegrain configuration. The subreflector thus must be transparent at S-band and reflective at X-band. This frequency selective behavior is called "Dichronic," a term borrowed from the optics.
Dichroic literally means two colors. In this case the two colors are the two separate bands, S-band (2.2 to 2.29 GHz) and X-band (8.025 to 8.5 GHz). The dichroic subreflector design described here has been implemented in two S and X-band Landsat antenna systems. The antenna systems are currently installed and operational in Sweden and Kenya.
A main reflector, a dichroic subreflector, and S and X-band monopulse tracking feeds comprise the system. With a focal length of 150 inches, the main reflector is a 9.14-meter paraboloid of revolution. The reflecting surface is manufactured from 24 stressed-sheet aluminum panels attached to a rugged space frame. Measured RMS surface error is 0.019 inch.
The subreflector is a 57-inch diameter hyperboloid of revolution constructed of two frequency-selective surfaces supported by a hollow fiberglass mounting ring.
Printed on a dielectric substrate, the dipole lattice structure adheres to hyperbolic panels made of 0.060 thick polyester resin fiberglass. The front and back panels are resin coated. The front panel is both resin coated and painted to protect the dipoles from weathering.
Measured losses due to the frequency selective subreflector are less than 0.2 dB in both bands of operation. Monopulse autotrack feed patterns are unaffected by the subreflector.
The cassegrain configuration for the X-band portion of the antenna is an adaptation of a 17th century optical telescope design. Invented by William Cassegrain, a contemporary of Isaac newton, the folded-optical-bath approach uses a two-reflector system. By placing a small hyperboloid mirror (subreflector) in front of a larger paraboloid primary mirror, higher magnification is obtained with a physically shorter telescope.
The microwave application of this established optical principle allows the use of shorter transmission lines in the antenna system and significantly, in a dualband design, makes the volume at the prime focus of the main reflector available for installation of a second feed.
Benefits provided by a Cassegrain configuration include:
* Greater flexibility in design of the primary feed.
* Spillover past the subreflector is directed at cold sky.
* Low spillover past the paraboloid toward the ground at high elevation angles.
Design tradeoffs at X-band required that the subreflector be at least 57 inches in diameter. The manufacture of this size of dichroic subreflector is challenging, because it is extremely difficult to maintain acceptable tolerances. Fabrication with a lathe and mill was out of the question. Instead, special vacuum molding tooling and processes were developed to fabricate the large dichroic surfaces.
The length and spacing of the resonant elements are determined by the characterisitcs of the transmitted RF energy. The Landsat-D application uses circular polarization, requiring the symmetrical placement of resonant crossed dipoles.
Subreflector Changes Properties
The surface of the subreflector must change its transmission and reflection properties with the frequency. This property is called a frequency selective surface (FSS). An ideal dichroic surface is perfectly reflective at one microwave frequency while completely transparent at another frequency.
For this application, the dichroic, frequency selective surface is composed of a two dimensional array of printed circuit, crossed dipoles. The dipole resonance is designed for 8.2 GHz, making the surface highly reflective. At the lower frequencies of S-band, the dipoles are virtually invisible.
Initially, a design with only one FSS was considered. In order to meet mechanical requirements at dielectric thickness of 125 mils was needed. Analysis showed electrical performance at X-band would be acceptable.
At S-band there was a 1 to 2 dB transmission loss due to the dielectric material. To reduce S-band losses to 0.5 dB or less the dielectric thickness would have to be reduced to approximately 10 mils. The single FSS design waws therefore discarded based on mechanical considerations. A dual FSS design was then addressed, in this approach a second layer is placed behind the first layer. Separation between layers was held to a 1/4 wavelength at S-band.
With this geometry the two surfaces appear as a near perfect match in S-band because the reflections in the two layers cancel. The two-layer approach meets all requirements. Analysis indicated that the subreflector would exhibit less than a 0.25 dB loss at both bands.
Each of the two surfaces in the dichroic subreflector are made of crossed dipoles. The center of the reflection band depends on the dipole length and the dielectric on which the dipoles are placed. Bandwidth is controlled by the spacing of the dipoles.
Array spacing is expressed as the ratio of element gap to element length (G/L) for bandwith comparisons. The G/L ratio affects bandwidth (defined as the band of frequencies over which the reflection coefficient of the dichroic surface is higher than -0.5 dB).
Acting like a metallic surface, the subreflector becomes a mirror near the resonance frequency of the dipoles. The design length of the dipoles make the subreflector highly reflective in the X-band. On the other hand, the subreflector is a transparent to the s-band. At this much lower frequency, the dipoles appear short and almost invisible.
The radiation pattern of a central dipole in a planar array is calculated by a moment-method computer program. Dichroic surfaces are evaluated by analyzing a single diple in a cluster-by-moment program. Parameters are dipole length, width and spacing.
A purely analytical approach must be complemented by experimentation before actually constructing the subreflector. The theory may be accurate, but critical parameters such as the dielectric constant of a given complex material are only approximately known and may depend on construction techniques.
Making Hyperboloid Is Expensive
Manufacture of hyperboloid shapes is expensive. Testing that dichroic panels is a cost-effective beginning. Made from the same material as the subreflector, panels can be easily adjusted until the required performance is achieved. The panels are rotated from the perpendicular to 40 degrees to simulate the hyperboloid range of incident angles.
The development plan has five phases:
* Build a single layer FSS and test it at the X-band. Iterate the dipole length until the surface is resonant at the desired frequency.
* Build a second surface identical to the first.
* Space the two surfaces approximately 1/4 wavelength at the lowest frequency of operation. Measure the S-band performance as a function of the incident angle.
* Adjust the spacing until the acceptable performance (insertion loss) is achieved at S-band.
* Compute the appropriate subreflector spacing so that the mean path through the two layers approximates the spacing determined in Phase 3.
Reflection bandwitdth depends on both the G/L ratio of the dipoles and the incidence angle. Arrays with G/L ratios of 1.2, 1.4 and 1.6 were tested for reflectivity at several angles of incidence. Higher G/L ratios are preferred for the S-band. But the E-plane reflection (parallel polarization) coefficient is very frequently sensitive for larger incident angles.
Reflectivity tests indicate that G/L ratios of 1.2 to 1.6 are acceptable. A reflected power versus frequency pattern is plotted using a sweep generator and pattern recorder.
Two such patterns, one with the FSS in place and the other with the metal reference surface in place, are recorded on the same chart. The difference between the two patterns is the measured reflectivity loss of the FSS panelin dB. Because a G/L ratio of 1.6 is susceptible to parameter tolerance, a G/L ratio of 1.4 was chosen.
Incident Angle Important
Also important for the FSS design is the maximum incident angle for both reflection and transmission. Reflection properties are a function of the angle of incidence. The FSS is designed for the worst angle of incidence. This ensures that the total loss of the FSS will be less than the design value. For the subreflector selected, the maximum incidence angle for both reflection and transmission is 40 degrees.
The angle of incidence is determined by calculating the incident ray and the normal to the subreflector. The FSS subreflector must maintain high reflectivity over the entire range of incident angles.
The bandwidth of any incidence angle is the band over which bot the E (parallel polarization) and H-plane (perpendicular polarization) coefficients are greater than -0.5 dB. The bandwidth drops to zero for larger incidence angles because of the shift of the E-plane reflection band away from the H-plane reflection band.
For design purposes, the subreflector FSS was simulated as a series of flat FSS test panels. Test samples of the dichroic surface were fabricated with copper crosses printed on the dielectric fiberglass. For each sample, a photomask was produced by computer, yielding a high accuracy in the spacing, length and thickness of the dipoles. The mask was then used for photo etching of the copper-clad dielectric test sheet.
RF Performance Evaluated
The RF performance of the panels was evaluated by making transmission tests. The panels were inserted between a transmit and receive horn in an anechoic chamber. The reference value was set by free space (without panels). Insertion amplitude and phase were recorded at S-band. Only insertion amplitude was recorded at X-band.
No matter how refined the computer model, the behavior of a material cannot be known in detail until empirical data are examined. For example, the influence of protective coatings on dipole resonant shift was determined experimentally.
Coating the dipoles made their effective length longer and lowered the reflection center frequency. The dipoles were designed shorter to compensate for this resonant shift. With a thickness of 0.002 inch, each dipole is arranged in an approximately square lattice over the hyperbolic surface.
The low loss, low voltage standingwave ratio of the production dichroic subreflector in conjunction with a high-efficiency monopulse tracking feed produced a system G/T in excess of +31 dB in X-band.
G/T was determined by measuring net antenna gain and system temperature separately, then calculating G/T. Gain ranged from 54.5 to 55 dBi. Tests were repeated after substituting a conventional solid surface hyperboloid for the dichroic reflector. Test data proved the G/T was degraded by less than 0.2 dB across the operating band by the dichroic subreflector.
|Printer friendly Cite/link Email Feedback|
|Author:||Seck, G.; Briscoe, H.|
|Date:||Jun 1, 1985|
|Previous Article:||Earth Stations Bring an End to White Alice.|
|Next Article:||FCC Chief Strives to Assure Mobile Communications Users Get More for Less.|