THE INSTITUTION OF ELECTRICAL, ENGINEERS PHASED ARRAY FUNDAMENTALS: MS SMITH ‘STC TECHNOLOGYPHASED ARRAY FUNDAMENTALS MS Smith ‘STC Technology Ltd, Harlow, Essex Contents 1. Introduction 2. Radiating apertures 2.1 Linear apertures 22° Planar apertures 3. Linear array analysis 3.1 Introduction 3.2 Radiation patterns 33. Beam steering 34° Nulls 4. Amray realisation 4.1 Amay elements 42 Beamforming 43° Systematic and random error effects 44° Calibration and BITE. 45 Limited sector scanning 5. Planar arrays Sal Planar array analaysis, 52. Beamforming 53 Munual coupling 61 62 63 Active arrays - a brief introduction ‘Transmitheceive modules LF. beamforming Baseband beamforming wa1. Introduction Phased arrays are a major class of antennas which offer many important features such as electronic beam scanning. Their flexibliry has to be assessed against their complexity and cost. In this lecture the fundamental principles of a phased array are described, to form a ‘basis forthe later lectures on applications. ‘The propenies of radiating apertures are given first, as these are basic to both aperrure antennas and arrays as sampled apertures. Linear array analysis then introduces beam steering and null steering. Array beamforming shows how array excitation is controlled, This is {followed by discussion ofthe effects of systematic or random excitation errors. Planar array analysis includes different array grid structures and the important topic of mutual ‘coupling. Finally active arays are briefly introduced. 2.1 Linear apemunes ‘The basic result of aperture theory for linear apertures is that the radiation pattem P(sing) is the Fourier transform of the aperture disibuton E,(,0). Mathematically snk) = Lo | Bg (x,0). exp (jhe covet) de (i) The calculation of radiation panems for some stndard aperture distbutions will both itsstrmte the technique and derive several very imporant antenna properties. The frst case isa uniform distribution, that is f maa a/a ‘Then the radiation pattem is P(S = sire) E.. exp (jx 5) da cin (ho S/a) ha 8/2 ree x Figure 1(a) shows P(S). The radiation pattem peak is at $ = 0, normal to the aperture, and the main beam extends from S = to s = +a. The first sidelobe peak is at S = 223A, and zeros of wethe pattem occur at $ = gr). Negative values for P(S) indicate a phase difference of 180° relative tothe main beam, Figue 10) shows he same pater expressed on 8 sl, a 20 fog 1S an] The fi (dle) sidelobe level i a 13.2 elie tothe sin beam peak ‘The two main fers of dectiv ration pater a th Dearth (oly athe Bevel) nd the sidelobe levels. Here the zero to zero beamwidth is 2/a on a sinecscale, while the 3 dB beamwidth is ~0.9Va. ‘The above case illustrates one very important antenna property, namely the inverse relationship between aperture size in wavelengths and the beamwidth. Thus narrow beams require large apertures. ‘A second example isa half-cosine distribution: betas) = | cack) i “The radiation pater i: Ms Ee (xe (jeas), 292. Grs/e) = 2x0 jnads) gy x Soe a 2 - n/a_ces (kas/a) (2-5) ~% — wats fy (after some algebra). This radiation pattem is plotted in figure 2 (linear field scale). The first zero of the pattem is not where kaS/2 =1/2, as both numerator and denominator become zero simultaneously. The 2er0-to-zero beamwidth is now 34a on a sin scale, $O per cent larger than for a unifonm aperture distribution. The 3 dB beamwidth is also increased correspondingly. ‘The sidelobe levels have decreased significantly - the first is now -23 4B (compared with -13 @. Ey‘This second case illustrates another important antenna propery; using a tapered apemure iswibution (an ‘amplirade taper’) tends to reduce sidelobe levels (there are limits), at the expense of an increase in main beam width. Low sidelobes are often a very desirable property in various applications. ‘Table 1 lists four standard aperture distributions and the corresponding radiation partes. nate a rn ee zoe Sa (rset o_o Wea ny = fue : ‘As a further example, consider a general amplimde distribution A(x). The corresponding radiation parter is = tf Ale) axp Gkxs) dx. (2.6) ‘Now ‘add’ a linear phase variation across the aperture - that is, the aperture distribution ‘becomes A(x) exp (pr). Then the new radiation patter is e(s) = 4 Als) exp(5 (aS + p)=) dx 2 P (s+ pA). ey) ‘The new pater therefore has an identical shape to the original one (on a sinsscale) but is shifted by Sisinac) = -fk. (This is an application of the Fourier shift theorem). If the vafirst partem has its main beam peak at $ = 0, the second pattem has its peak where S ik. ‘Asan example, i=, the beam maximum istered from o'= 0 /= 30° ‘Because the pattem has the same shape on a sine scale, its beamwidth is constant on this scale. ‘This is not constant on ans'scale. Let the beamwidth be 2on the sine scale, and let the beam steering angle be oj. Then the beamwidth on anqscale is x, — where ‘Assume that Sis small (anarrow beam), andlet og, = of, + € ‘Thea St oy = sh Cate Stn oy Cel E + Cot oy SHE HS ye + COS. EL Then Gz ffnd the beamwidth on afeale becomes 2 b/us x, ‘The beamwidth therefore increases as the stering angle increases. Physically this corresponds to the reduction in projected aperture as the beam is steered (a smaller aperture giving a wider beam), ‘The principle of beam steering by applying a phase gradient to a radiating aperture is very important, as it allows electronic control of beam direction. 22° Planar apemures We now consider the three-dimensional generalisations of the formulae for two dimensions ‘There the fields had no ydependence, and an aperture of width a was infinitely long in the y direction. We now have an aperture such as shown in figure 3, and Cartesian coordinates x, y, 4, or spherical polar coordinates r, @, dl, may be appropriate. As in the two-dimensional case, there are two independent polaisaions, and we shall give revats for the Rycase (E, = 0 in the aperture). vs‘The two-dimensional and three-dimensional radiation pattems of linear and planar apertures are Ex (22) exp (jks) doc, (aa) P(S., 52) = Ex (26, 4.0) exp (jhe $+ ySe)) doe dy, (2.2) where S,2 sin Ec 3. ‘n6siusing conventional spherical polar coordinates as in figure In the far field, in two-dimensions and three-dimensions respectively: Ee Ve) axp (—j (hr-w/e)). ue. P(r) (a.te} E (r, 6,8) = \ exp (-j (br—m/a)) (cord ue — sing ces ug) x (5, S2), (Gu ‘The dependence on r in the three-dimensional case has now changed to that ofa spherical, rather ‘than cylindrical, wave. The w/2 phase shift is half the r shift through a spherical focus. The factor (ue. u,\efines the polarisation of the wave. ‘The evaluation of radiation patterns for some important cases can now be made. consider a ‘uniformly illuminated rectangular aperture, as in figure 3, such that » ele ala, by) < b/s 7 (aa) FH) Oahu Be sale ‘The radiation pattem is ?(s,,S2) = Ee| [err ines e9S)dety Mily dap t/a bo. = Eo [ axp(jhxSi)de Pep thys)dy (2.1) Map Jape 16(Some simple aperture distributions are separable - that i, E, (x, y, 0) = E, (x). E,(y), which allows the x and y integrations to be separated asin this case.) Ait f= y ‘and the pattem in the y-z plane is p F(0,%) = akEo. ra ‘These puters are ofthe same form as fora uniform aperture in two dimensions. Note thatthe first zero in the x-2 plane is where sin @= 2}/a, and the first zero in the y-z plane is where sin @ = yb. The beamwidths in the two principal planes are separately controlled by the aperture dimensions in each plane. ‘Asa second example (a hom antenna), let [Bo ces (y/o), Ixixa/a, iyle ¥* Ce, cthendse Ex (=. 419) (505) pySe)) dx dy (2-12) wrThis is again separable, with the two integrals similar to the two dimensional examples given In the x plane (S, = 0), the fet zero of the panem is where sin6=}/a, wheres inthe y-z plane (S, = 0), the first zero is where sin © = 3/2 Nb. ‘The sidelobes in the y-z plane are at 8 lower level than in the x-z plane. A third example is a uniform illumination over a circular aperure with diameter D. The radiation patter in any plane is a Bessel function, shown in figure 4. The first sidelobe level is -17.6 dB, and the 3 4B beamwidth is 1.02\/D. The (normalised) pattem function is BT, (aa ein) (a0) ba sn e (where D = 2a), If an amplinude taper is superimposed on this field distribution, the bearnwith increases and the sidelobes are (generally) reduced, just as fora linear aperture. ‘The directional characteristics of an antenna are frequently expressed in terms of a gain function G(@, 9), whose maximum value is the gain. (Here andgare spherical polar angular coordinates.) This is now discussed for the transmit case, for convenience. The gain is defined as the ratio of the maximum radiation intensity from the antenna to the maximum from a reference antenna with the same input power. The reference is usually a hypothetical lossless isotropic radiator, and the gain can then be expressed in dBi. At a distance r from an isotopic source, the power ransited is evenly spread over aspherical surface of area 47 For an antenna with gun, i follows thatthe power P incident on an area A ata distance rin the direction of mos intense radiation is P S. %. A/ (ure?) (aal) where Py isthe transmitter power, and G is now used forthe gun, the maximum valve of G (@, @). Disectivity, D, is defined as the ratio of the maximum radiation intensity to the average radiation intensity (averaged over all angles). For an antenna which is 100 per cent efficient va