Electronic Scanned Array Design covers the fundamental principles of ESA antennas including basic design approaches and inherent design limitations. These insights enable better appreciation of existing and planned ESA systems including their application to earth observation. The material describes general design principles of aperture antennas applied to the specific case of ESA design. System applications are discussed to set the framework for requirements allocation and flowdown. Specific examples are cited throughout to relate theory to practice. The book begins by introducing the concept of electronic scanned arrays, giving a brief history of the technology and outlining its scope and applications. Further chapters cover antenna principles; synthetic arrays; antenna figures of merit; mutual coupling effects; errors and tolerances; grating lobes; thinned arrays; beam width and sidelobes; beam shaping and spoiling; reflector applications; design practice; radiating elements; T/R modules; assembly, packaging, power and thermal management; technology base and cost; and ESAs in space. The final chapter offers a comparison between an ESA and a reflector, exploring their benefits, detriments and design trades. The book will be invaluable for radar and antenna engineers and researchers, and advanced students studying ESA design.
Other keywords: R module; radiating element; space systems; microwave distribution; antenna feed; performance analysis; reflector antenna; requirements allocation; T module; aperture antenna; electronic scanned array; packaging; monolithic microwave integrated circuit; design engineering
Electronic scanned array (ESA) design is based on fundamental principles known almost since the time of Maxwell. They were only abstract concepts for many years because the component technology did not exist. ESAs offer a variety of benefits to the system designer to justify their additional cost and complexity. In recent times, the reduced cost of advanced electronics together with advanced systems has seen realization of many of these concepts. The development of ESAs took many years and the efforts of many talented individuals and organizations. It has its roots in military applications, and advances may be credited to World War II during which there was an emphasis on rapid development of critical technologies including a requirement for many and varied radar systems. ESA development parallels that of modern electronic devices and has relied greatly on those efforts. The invention of the transistor and the integrated circuit initiated a technological transformation replacing vacuum tubes with solid-state components. Both analog and digital domain applications benefited. The latter enabled revolutionary advances in computation, both sophisticated electromagnetic (EM) simulation for ESA design and enormous processing capability required to understand the resulting sensor data.
Antennas are devised for the purpose of transmission and reception. In transmission, the antenna objective is to create a current/voltage distribution at the aperture surface which will create a specified beam pattern. In reception, the antenna produces an output signal whose magnitude and phase correspond to the incoming signal's magnitude and phase according to the antenna beam shape. General antenna theory provides concepts and rules for ESA analysis. Antenna analysis assumes linearity and superposition to decompose solutions into frequency and angular components.
To this point, the discussion has dealt with physical antennas. Real beamforming uses samples collected at one point in time. Resolution and beam shape are limited by the size of the antenna and the number of elements. Real antenna angular resolutions are on the order of 1°, so resolution at 50 km is almost 1 km which is too large for many purposes. The array concept may be extended to arrays that are synthetic. Figure 3.1 illustrates such an arrangement. Synthetic arrays create the effect of a very long antenna with very small azimuth resolution by coherently combining data from a relatively short antenna which occupies multiple positions sequentially. Usually, but not necessarily, it operates in continuous motion. Target motion during data collection will degrade the image.
An ESA design is a collection of sampled gain and phase states of its elements. There are numerous design techniques to achieve a desired result. In principle, any specific beam pattern has only one parametric realization. Accordingly, various design techniques striving for the same performance should generate similar designs. This section will discuss optimization of the mainbeam and sidelobes. Improving one unfortunately tends to degrade the other. Mainbeam shaping is usually a matter of obtaining the desired beamwidth. Sidelobes are rarely beneficial as they are responsible for mainbeam gain reduction, small target masking, false target insertion, and other signal artifacts. Most techniques for sidelobe reduction are applicable to both transmit and receive antenna patterns but often penalize the transmit efficiency. Since the system performance is a function of the product of the transmit and the receive antenna patterns, it is desirable to apply as much of the sidelobe mitigation as possible to the receive function.
Mutual coupling is not an independent phenomenon. It describes the effects of incomplete boundary condition specification in the original problem. It operates at short-range and is not always apparent. In a very large array, the effect of the coupling for most elements is identical and its effect is incorporated into the element pattern leaving the array factor unchanged. The treatment of mutual coupling in this section is qualitative as analytic tech-niques have not been very effective. It is difficult to measure in physical arrays. Its practical consequences have diminished as improved electromagnetic models and solvers deal with its effects directly with suitable boundary conditions. Mutual coupling may introduce significant errors and uncertainties in the analysis. Of most concern, it may create unexpected nulls in the antenna pattern (scan blindness), but it also affects the element pattern and the array pattern. It is important to evaluate mutual coupling by various means including simulation and subscale tests. Frequently it is necessary to build test articles before accepting the design. The array size is a consideration: small arrays may be modeled in their entirety with numerical methods; large arrays may be partitioned or analyzed as infinite in extent.
In converting a design into hardware, there is seldom an exact match between as-designed and as-built, so performance will deviate from prediction. System requirements must be decomposed into subsystem specifications and tolerances compared with accuracy achievable with available components. Design and simulation with software tools utilize parameters that are exact to many decimal places. An essential step in the design process is to assess the sensitivity of predicted performance to inaccuracies, including temporal changes. This chapter describes common error sources and assesses their effects. The effect of errors is to degrade antenna performance metrics that include beam pointing, beamwidth, sidelobe magnitude, and location. Correlated errors such as phase and gain quantization tend to be more noticeable but their average effect is comparable to random errors. Errors may arise from component or fabrication variation or from partial or complete component failure. Intentional deactivation of elements, may also be analyzed as a failure whose effects vary. For instance, turning off one-half of the array has greatly different effects depending on whether those elements are on one side of the array, alternate rows or columns of the array, or randomly selected. It is necessary to assess the specific design options to determine the importance of these errors.
Grating lobes are a sampling phenomenon. They represent aliasing appearing from other Nyquist zones due to undersampling. They are detrimental because they divert power to unwanted directions in transmit and they introduce additional noise and spurious targets in receive. It is both expensive and usually unnecessary to totally eliminate grating lobes. Scanning far off-boresight is not a common expectation of a planar antenna. Many applications require only 100-30° scan capability. Instead, the optimum solution is usually a compromise of restricted scan volume and reduced grating lobes.
The analysis in this book primarily concerns conventional arrays with elements spaced uniformly across a specified aperture to achieve specified performance: radiated power, sidelobes, etc. In this chapter, we consider thinned (or sparse) arrays by which we mean conventional arrays from which some number of elements have been deliberately removed. Rocca et al. provide an overview of types of irregular arrays and associated design techniques. Thinned arrays are useful in cases where the design complies with most performance requirements while exceeding others. It may be possible to reduce the excess performance (and cost) by selectively removing some elements.
The designer has almost unlimited freedom to specify the amplitude and phase distribution of the signal across the aperture. However, the range of beam shapes is limited. A uniform amplitude produces the maximum gain and minimum beamwidth but relatively high sidelobes. Any attempt to lower sidelobes generally reduces gain and increases beamwidth. We may consider analysis and synthesis techniques for antenna optimization. In general, analysis produces a solution although perhaps not the best solution possible. Various optimization techniques have been used and are limited by computing resources, the existence of local optima, or both. Synthesis uses the desired solution and works backwards but does not necessarily produce a good result or any result at all. Stutzman and Licul provide an overview of synthesis techniques. This chapter considers analytic techniques. It describes a number of solutions and their beam characteristics.
The techniques described in the previous chapter dealt primarily with sidelobes without constraining the resulting mainbeam width and shape. It is sometimes desirable to form a broad beam, of constant or shaped amplitude. Examples include fan beams for increasing SAR swath coverage or cosecant beams for air surveillance where range varies with elevation angle. This chapter discusses the last three techniques.
Reflectors may be incorporated into an ESA design to increase the aperture area. By their nature, ESAs require considerable electronics with associated size, weight, power, and cost. High-gain, narrow-beam systems require many T/R modules. If other requirements such as electronic field of view allow, reflectors can provide increased effective area without additional electronics. The added passive mechanical structure is generally lower cost than the equivalent ESA area. This approach combines some of the benefits (and some of the disadvantages) of ESAs and reflectors. ESA feeds may be used with both cylindrical (one-dimensional curvature) and rotationally symmetric reflectors (two-dimensional curvature). The basic trade-off is to exchange electronic field of regard (EFOR) for fewer T/R modules; this is exactly analogous to a thinned array.
Designs are never ideal, and compromises are necessary to reconcile requirements with size, weight, power, and cost. ESA antennas provide unique benefits to improve performance which must be weighed against their greater cost and complexity The first consideration in the design process is to devise an architecture consistent with system constraints and capable of satisfying system requirements. This requires a functional decomposition and partition, bearing in mind practical considerations such as component cost and availability. It is typically an iterative process, as each design iteration reveals opportunities for improvement and difficulties in realization.
Radiating elements are the most visible and in many respects the most important constituent of an ESA. Their function is to create a desired electromagnetic field on the surface of the aperture to generate the desired farfield pattern. To this end, the T/R module provides an RF signal to the antenna port (connector) with the necessary amplitude and phase which the radiating element, as a linear device will impart to its radiated signal.
T/R modules are the heart of an ESA and their continued improvement has been the key to ESA development. T/R modules provide distributed gain and phase control, typically at each radiating element. They provide flexibility enabling the attractive performance of the ESA. There are many good examples in systems and many more concepts described in the literature. The cost of T/R modules has been the most important restriction on their wide use. Since the 1990s, costs have declined precipitously, leading to the vast increase in ESA applications, primarily because of commercial demand for MMICs, application-specific integrated circuits (ASICs), and other components.
The previous chapters have dealt mainly with performance in terms of antenna beam characteristics and quality. This chapter deals with the necessary provisions and infrastructure for the ESA to perform to its potential at the lowest cost. There are a few papers covering the general subject including Kemkemian et al. and Renard et al.
Cost is a dominant consideration with ESAs and was historically the principal reason that ESAs were not more prevalent. Cost represents a significant competitive discriminator, so this information is rarely published. Accordingly, the cost discussion in this chapter is necessarily inferential. Large investments in R&D have mitigated most of these issues to the point that ESAs are cost competitive with other antenna types, particularly in view of their additional capabilities difficult or impossible to achieve with traditional designs. Historically, T/R modules constituted about 50% of the cost of an ESA and components were about 50% of the cost of the T/R module. Accordingly, considerable effort was expended on both component and module cost. These estimates were predicated on high reliability (military) requirements and standards. Commercial systems relaxe these constraints with commensurate cost savings. Cost is a strong function of both production rates and quantities. In 2001, a US Defense Science Board Study titled Future DoD Airborne High-Frequency Radar Needs/Resources [1] forecast a requirement for a total of 6 million T/R modules for current and planned US military programs through 2021. Ultimately, the demand for ESAs was sufficient to support dedicated production lines, using similar components, processes, and equipment as their commercial counterparts and achieving commensurate cost savings.
ESAs are widely used in space applications, and numerous satellite SAR use ESAs. These examples provide an interesting opportunity to compare and contrast alternative design concepts. ESAs are a good match to the application because their costs do not seem as daunting compared to the cost of satellites. Their reliability and graceful degradation are attractive for missions expected to last many years. Communications satellites, discussed in Section 1.4.2, were quickly realized after the first satellite launch in 1958. Radar satellites, discussed in Section 1.4.3, were developed some years later.
There does not appear to be a conclusive advantage for ESA or ESA-fed reflectors for the earth observation mission. At X-band, ESAs seem to be the dominant choice, in part because the antennas are smaller and also because of the huge technology base developed for airborne and ground-based radars. At lower frequencies, the reflector benefits are more compelling as the antennas are larger. In many cases, the experience base at the manufacturer will determine the outcome for cost and risk benefits. Relative merits will continue to evolve.
Books and Web-based material are considered in the appendix, covering topics relevant to electronic scanned array design.
MATLAB® scripts were used to create most of the figures in this book and for the associated short course. The name of the script is shown beneath its associated figure. The programs are available for download from the publisher's website at https://digital-library.theiet.org/. The scripts do not include a user interface or extensive error checking. Parameters are hard-coded but may be altered to assess other cases. The underlying engineering formulae represent relatively few lines of the program; much of the rest relates to graphical output and format. The program run times range from seconds to hours. Run times depend, in large part, on the size of arrays and the number of iterations (cases).
Antenna analysis uses polar coordinates. The array is shown in this paper. It lies on the xy-plane (z = 0) and radiates in the +z-direction (hemispheric). Radiation in the -z-direction is usually undesirable and is prevented by a ground plane in the plane of the array. Generally, the ground plane need extend only a few wavelengths away from the array. For analysis purposes, these coordinates may be used directly or converted to an orthogonal representation.
Matlab scripts for Electronic Scanned Array Design
Click on the icon below to download the Matlab scripts used to generate the figures in this book.