LCTUR 18: Horn Antennas (Rectangular horn antennas. Circular apertures.) 1 Rectangular Horn Antennas Horn antennas are popular in the microwave band (above 1 GHz). Horns provide high gain, low VSWR (with waveguide feeds), relatively wide bandwidth, and they are not difficult to make. There are three basic types of rectangular horns. The horns can be also flared exponentially. This provides better matching in a broad frequency band, but is technologically more difficult and expensive. The rectangular horns are ideally suited for rectangular waveguide feeders. The horn acts as a gradual transition from a waveguide mode to a free-space mode of the M wave. When the feeder is a cylindrical waveguide, the antenna Nikolova 14 1
is usually a conical horn. Why is it necessary to consider the horns separately instead of applying the theory of waveguide aperture antennas directly? It is because the so-called phase error occurs due to the difference between the lengths from the center of the feeder to the center of the horn aperture and the horn edge. This makes the uniform-phase aperture results invalid for the horn apertures. 1.1 The H-plane sectoral horn The geometry and the respective parameters shown in the figure below are used often in the subsequent analysis. l H R x a a H R A z R H H-plane (x-z) cut of an H-plane sectoral horn a l H H A = R +, (18.1) A = arctan lh RH = ( A a) A R, (18.) 1 4. (18.3) Nikolova 14
The two fundamental dimensions for the construction of the horn are A and R H. The tangential field arriving at the input of the horn is composed of the transverse field components of the waveguide dominant mode T 1 : where Z g = η λ 1 a λ βg = β 1 a π y( x) = cos x e a H ( x) = ( x)/ Z x y g jβ z is the wave impedance of the T 1 mode; g is the propagation constant of the T 1 mode. (18.4) Here, β = ω µε = π / λ, and λ is the free-space wavelength. The field that is illuminating the aperture of the horn is essentially a spatially expanded version of the waveguide field. Note that the wave impedance of the flared waveguide (the horn) gradually approaches the intrinsic impedance of open space η, as A (the H-plane width) increases. The complication in the analysis arises from the fact that the waves arriving at the horn aperture are not in phase due to the different path lengths from the horn apex. The aperture phase variation is given by e j β ( R R ). (18.5) Since the aperture is not flared in the y-direction, the phase is uniform in this direction. We first approximate the path of the wave in the horn: x 1 x R= R + x = R 1+ R 1 R + R. (18.6) The last approximation holds if x R, or A/ R. Then, we can assume that Nikolova 14 3
1 x R R. (18.7) R Using (18.7), the field at the aperture is approximated as π a ( x) cos x e y A β j x R. (18.8) The field at the aperture plane outside the aperture is assumed equal to zero. The field expression (18.8) is substituted in the integral I y (see Lecture 17): I = ( x, y ) e dx dy, (18.9) jβ( x sinθcosj y sinθsin j) y a + S A y + A/ b + b/ j x I = x e R e jbx sinθcosjdx e jby sinθsinjdy y π cos A. (18.1) (((((( (((((( A/ b/ I ( θj, ) The second integral has been already encountered. The first integral is cumbersome and the final result only is given below: where and bb sin sinθsinϕ 1 π R Iy = I( θϕ, ) b, (18.11) b bb sinθsinϕ R π j βsinθcosj+ β A R π j βsinθcosj β A [ 1 1 ] I( θj, ) = e C( s ) js( s ) C( s ) + js( s ) [ ( ) ( ) ( 1 ) ( 1 )] + e C t js t C t + js t 1 βa πr s1 = Rβ u πβ R A ; 1 βa πr s = + Rβ u πβ R A ; (18.1) Nikolova 14 4
t1 = 1 βa πr Rβ u+ πβ R A ; t = 1 βa πr + Rβ u+ πβ R A ; u = sinθcosϕ. Cx ( ) and Sx ( ) are Fresnel integrals, which are defined as x π ( ) cos τ τ ; ( ) ( ), Cx= d C x= Cx x ( ) sin ; ( ) ( ). π Sx= τ d τ S x= Sx (18.13) More accurate evaluation of I y can be obtained if the approximation in (18.6) is not made, and is substituted in (18.9) as or ay ( ) π π a ( x) = cos x e e cos x e y = A A + jβ R β jβ R + x R j R + x The far field can be now calculated as (see Lecture 17): e jβ r (1 cos )sin θ = jβ + θ j Iy, 4π r e jβ r j (1 cos )cos I j = β + θ j y, 4π r bb sin sinθsinj πr j r e b 1 cosθ + = jb b b 4πr bb sinθsinj ( θˆ + φˆ ) I( θj, ) sinj cos j.. (18.14) (18.15) (18.16) Nikolova 14 5
The amplitude pattern of the H-plane sectoral horn is obtained as Principal-plane patterns -plane ( ϕ = 9 ): bb sin sinθsinϕ 1 cosθ + = I( θϕ, ). (18.17) bb sinθsinϕ F bb sin sinθsinϕ 1 cosθ + ( θ ) = bb sinθsinϕ (18.18) The second factor in (18.18) is exactly the pattern of a uniform line source of length b along the y-axis. H-plane ( ϕ = ): 1+ cosθ FH( θ) = fh( θ) = (18.19) 1+ cos θ I( θϕ, = ) = I( θ =, ϕ = ) The H-plane pattern in terms of the I( θϕ, ) integral is an approximation, which is a consequence of the phase approximation made in (18.7). Accurate value for fh ( θ ) is found by integrating numerically the field as given in (18.14), i.e., + A/ π x fh ( θ ) cos e e dx A jβ R + x jβx sinθ. (18.) A/ Nikolova 14 6
- AND H-PLAN PATTRN OF H-PLAN SCTORAL HORN Fig. 13-1, Balanis, p. 674 Nikolova 14 7
The directivity of the H-plane sectoral horn is calculated by the general directivity expression for apertures (for derivation, see Lecture 17): D 4π = λ S a S A ads A. (18.1) ds The integral in the denominator is proportional to the total radiated power, + b/ + A/ π ds x dx dy ηπ rad = a = cos = A S b/ A/ A Ab. (18.) In the solution of the integral in the numerator of (18.1), the field is substituted with its phase approximated as in (18.8). The final result is where ε t = 8 ; π D H b 3 A 4p = εh ph = εε t λp λ λ H ph {[ ( ] [ ] } 1 ) ( ) ( 1 ) ( ) p ε H ph = C p C p + S p S p ; 64t 1 1 p1 = t 1 +, p = t 1+ 8t 8t ; 1 A 1 t =. 8 λ R / λ ( Ab), (18.3) The factor ε t explicitly shows the aperture efficiency associated with the aperture cosine taper. The factor ε H ph is the aperture efficiency associated with the aperture phase distribution. A family of universal directivity curves is given below. From these curves, it is obvious that for a given axial length R and at a given wavelength, there is an optimal aperture width A corresponding to the maximum directivity. Nikolova 14 8
R = 1λ [Stutzman] It can be shown that the optimal directivity is obtained if the relation between A and R is or A= 3λR, (18.4) A R = 3. (18.5) λ λ Nikolova 14 9
1. The -plane sectoral horn l R y b a R B z R -plane (y-z) cut of an -plane sectoral horn The geometry of the -plane sectoral horn in the -plane (y-z plane) is analogous to that of the H-plane sectoral horn in the H-plane. The analysis is following the same steps as in the previous section. The field at the aperture is approximated by [compare with (18.8)] Here, the approximations and π a = cos x e y a β j y R y 1 y R= R + y = R 1+ R 1 R + R. (18.6) are made, which are analogous to (18.6) and (18.7). (18.7) 1 y R R (18.8) R Nikolova 14 1
The radiation field is obtained as 4 j r βr βb a π R e β j sinθsinj ˆ ˆ ( sin cos ) = jβ e θ j + φ j π β 4πr βa cos sinθcosj (1 + cos θ ) [ C( r) js( r) C( r1) + js( r1) ]. βa 1 sinθcosj The arguments of the Fresnel integrals used in (18.9) are r β B βb = R sinθsin ϕ, 1 π R r β B βb = + R sinθsin ϕ. π R (18.9) (18.3) Principal-plane patterns The normalized H-plane pattern is found by substituting ϕ = in (18.9): βa cos sinθ 1 cosθ + H ( θ ) =. (18.31) βa 1 sinθ The second factor in this expression is the pattern of a uniform-phase cosineamplitude tapered line source. The normalized -plane pattern is found by substituting ϕ = 9 in (18.9): [ Cr ( ) Cr ( 1) ] + [ Sr ( ) Sr ( 1) ] (1+ cos θ) (1+ cos θ) ( θ) = f ( θ) = 4 C( r θ= ) + S ( rθ= ) Here, the arguments of the Fresnel integrals are calculated for ϕ = 9 :. (18.3) Nikolova 14 11
and r β B βb = R sin θ, 1 π R r β B βb = + R sin θ, π R (18.33) B β rθ = = r( θ = ) =. (18.34) π R Similar to the H-plane sectoral horn, the principal -plane pattern can be accurately calculated if no approximation of the phase distribution is made. Then, the function f ( θ ) has to be calculated by numerical integration of (compare with (18.)) B/ jβ R ( ) + y f j sin y θ e e β θ dy. (18.35) B/ Nikolova 14 1
- AND H-PLAN PATTRN OF -PLAN SCTORAL HORN Fig. 13.4, Balanis, p. 66 Nikolova 14 13
Directivity The directivity of the -plane sectoral horn is found in a manner analogous to the H-plane sectoral horn: where ε t 8 =, ε π ph D = a 3 B 4p ε ph εε t phab λp λ = λ, (18.36) + = =. λr C( q) S( q) B, q q A family of universal directivity curves λd / a vs. B/ λ with R being a parameter is given below. R = 1λ Nikolova 14 14
The optimal relation between the flared height B and the horn apex length R that produces the maximum possible directivity is B= λr. (18.37) 1.3 The pyramidal horn The pyramidal horn is probably the most popular antenna in the microwave frequency ranges (from 1 GHz up to 18 GHz). The feeding waveguide is flared in both directions, the -plane and the H-plane. All results are combinations of the -plane sectoral horn and the H-plane sectoral horn analyses. The field distribution at the aperture is approximated as π a cos x e y A β x y j + R H R. (18.38) The -plane principal pattern of the pyramidal horn is the same as the -plane principal pattern of the -plane sectoral horn. The same holds for the H-plane patterns of the pyramidal horn and the H-plane sectoral horn. The directivity of the pyramidal horn can be found by introducing the phase efficiency factors of both planes and the taper efficiency factor of the H-plane: where ε t 8 = ; π D 4p = εε ε ( AB), (18.39) λ H P t ph ph {[ ( ] [ ] } 1 ) ( ) ( 1 ) ( ) p ε H ph = C p C p + S p S p ; 64t 1 1 p1 = t 1 +, p = t 1 +, 8t 8t ε ph C( q) + S( q) B, q q λr = =. 1 A 1 t = ; 8 λ R / λ The gain of a horn is usually very close to its directivity because the radiation Nikolova 14 15 H
efficiency is very good (low losses). The directivity as calculated with (18.39) is very close to measurements. The above expression is a physical optics approximation, and it does not take into account only multiple diffractions, and the diffraction at the edges of the horn arising from reflections from the horn interior. These phenomena, which are unaccounted for, lead to only very minor fluctuations of the measured results about the prediction of (18.39). That is why horns are often used as gain standards in antenna measurements. The optimal directivity of an -plane horn is achieved at q = 1 [see also (18.37)], ε ph =.8. The optimal directivity of an H-plane horn is achieved at t = 3/8 [see also (18.4)], ε H ph =.79. Thus, the optimal horn has a phase aperture efficiency of ε P = ε Hε =.63. (18.4) ph ph ph The total aperture efficiency includes the taper factor, too: εp = εεhε =.81.63 =.51. (18.41) ph t ph ph Therefore, the best achievable directivity for a rectangular waveguide horn is about half that of a uniform rectangular aperture. We reiterate that best accuracy is achieved if ε ph H and ε ph are calculated numerically without using the second-order phase approximations in (18.7) and (18.8). Optimum horn design Usually, the optimum (from the point of view of maximum gain) design of a horn is desired because it results in the shortest axial length. The whole design can be actually reduced to the solution of a single fourth-order equation. For a horn to be realizable, the following must be true: R = RH = RP. (18.4) The figures below summarize the notations used in describing the horn s geometry. Nikolova 14 16
b α R y B α α H H R x A z R R H It can be shown that RH A/ A = = R A/ a/ A a, (18.43) H R B/ B = = R B/ b/ B b. (18.44) The optimum-gain condition in the -plane (18.37) is substituted in (18.44) to produce B bb λr =. (18.45) There is only one physically meaningful solution to (18.45): ( 8λ ) 1 B= b+ b + R. (18.46) Similarly, the maximum-gain condition for the H-plane of (18.4) together with (18.43) yields Since R H R H A a A ( A a) = A A 3λ =. (18.47) 3λ = R must be fulfilled, (18.47) is substituted in (18.46), which gives Nikolova 14 17
1 8 AA ( a) B= b b + +. (18.48) 3 Substituting in the expression for the horn s gain, 4p G = ε ap AB, (18.49) λ gives the relation between A, the gain G, and the aperture efficiency ε ap : 4p 1 8 Aa ( a) G = ε ap A b+ b +, (18.5) λ 3 + 3bGλ 3G λ =. (18.51) 4 A4 aa3 A 8pε 3 ap p ε ap quation (18.51) is the optimum pyramidal horn design equation. The optimum-gain value of ε ap =.51 is usually used, which makes the equation a fourth-order polynomial equation in A. Its roots can be found analytically (which is not particularly easy) and numerically. In a numerical solution, the first guess is usually set at A () =.45λ G. Once A is found, B can be computed from (18.48) and R = RH is computed from (18.47). Sometimes, an optimal horn is desired for a known axial length R. In this case, there is no need for nonlinear-equation solution. The design procedure follows the steps: (a) find A from (18.4), (b) find B from (18.37), and (c) calculate the gain G using (18.49) where ε ap =.51. Horn antennas operate well over a bandwidth of 5%. However, gain performance is optimal only at a given frequency. To understand better the frequency dependence of the directivity and the aperture efficiency, the plot of these curves for an X-band (8. GHz to 1.4 GHz) horn fed by WR9 waveguide is given below ( a =.9 in. =.86 cm and b =.4 in. = 1.16 cm). Nikolova 14 18
The gain increases with frequency, which is typical for aperture antennas. However, the curve shows saturation at higher frequencies. This is due to the decrease of the aperture efficiency, which is a result of an increased phase difference in the field distribution at the aperture. Nikolova 14 19
The pattern of a large pyramidal horn ( f = 1.55 GHz, feed is waveguide WR9): Nikolova 14
Comparison of the -plane patterns of a waveguide open end, small pyramidal horn and large pyramidal horn: Nikolova 14 1
Note the multiple side lobes and the significant back lobe. They are due to diffraction at the horn edges, which are perpendicular to the field. To reduce edge diffraction, enhancements are proposed for horn antennas such as corrugated horns aperture-matched horns The corrugated horns achieve tapering of the field in the vertical direction, thus, reducing the side-lobes and the diffraction from the top and bottom edges. The overall main beam becomes smooth and nearly rotationally symmetrical (esp. for A B). This is important when the horn is used as a feed to a reflector antenna. Nikolova 14
Nikolova 14 3
Comparison of the H-plane patterns of a waveguide open end, small pyramidal horn and large pyramidal horn: Nikolova 14 4
Circular apertures.1 A uniform circular aperture The uniform circular aperture is approximated by a circular opening in a ground plane illuminated by a uniform plane wave normally incident from behind. x z a y The field distribution is described as a = x ˆ, ρ a. (18.5) The radiation integral is ˆ I j x e β ds S = a rr. (18.53) The integration point is at ρ = xˆρ cosϕ + y ˆρ sinϕ. (18.54) In (18.54), cylindrical coordinates are used, therefore, ρˆ ρ = ρ sin θ(cosϕcosϕ + sinϕsin ϕ ) = ρ sinθcos( ϕ ϕ ). (18.55) Hence, (18.53) becomes a π a I j sin cos( ) x = e βρ θ j j dj ρ dρ = π ρ J( βρ sin θ) dρ. (18.56) Here, J is the Bessel function of the first kind of order zero. Applying the identity Nikolova 14 5
to (18.56) leads to xj ( x ) dx = xj 1( x ) (18.57) a I x = π J1( βasin θ) βsinθ. (18.58) Note that in this case the equivalent magnetic current formulation of the equivalence principle is used [see Lecture 17]. The far field is obtained as ( ˆ cos ˆ cos sin ) e jβ r = θ j φ θ j jβ I x = π r j r 1( sin ) ( ˆ e β J βa θ = θcosj φˆ cosθsin j) jβπa. πr βasinθ Principal-plane patterns -plane ( ϕ = ): θ (18.59) J1( βasin θ) ( θ ) = βasinθ (18.6) H-plane ( ϕ = 9 ): ϕ J1( βasin θ) ( θ) = cosθ βasinθ (18.61) The 3-D amplitude pattern: J 1( βasin θ) ( θϕ, ) = 1 sin θsin ϕ (( βa sin ( θ ( f ( θ ) (18.6) The larger the aperture, the less significant the cosθ factor is in (18.61) because the main beam in the θ = direction is very narrow and in this small solid angle cosθ 1. Thus, the 3-D pattern of a large circular aperture features a fairly symmetrical beam. Nikolova 14 6
xample plot of the principal-plane patterns for a = 3λ : 1.9 -plane H-plane.8.7.6.5.4.3..1-15 -1-5 5 1 15 6*pi*sin(theta) The half-power angle for the f ( θ ) factor is obtained at βasinθ = 1.6. So, the HPBW for large apertures (a λ ) is given by 1.6 1.6 λ HPBW = θ1/. arcsin 58.4 βa =, deg. (18.63) βa a For example, if the diameter of the aperture is a = 1λ, then HPBW = 5.84. The side-lobe level of any uniform circular aperture is.133 (-17.5 db). Any uniform aperture has unity taper aperture efficiency, and its directivity can be found directly in terms of its physical area, 4p 4p Du A p a λ λ = p =. (18.64) Nikolova 14 7
. Tapered circular apertures Many practical circular aperture antennas can be approximated as radially symmetric apertures with field amplitude distribution, which is tapered from the center toward the aperture edge. Then, the radiation integral (18.56) has a more general form: a I = π ( ρ ) ρ J ( βρ sin θ) dρ. (18.65) x In (18.65), we still assume that the field has axial symmetry, i.e., it does not depend on ϕ. Often used approximation is the parabolic taper of order n: a ρ ( ρ ) = 1 a n (18.66) where is a constant. This is substituted in (18.65) to calculate the respective component of the radiation integral: n a ρ Ix ( θ) = π 1 ρ J( βρ sin θ) dρ a. (18.67) The following relation is used to solve (18.67): 1 n n! (1 x) nxj( bx) dx = J 1 n 1( b) bn+ + a b= ba θ. Then, I x ( ) π a I x ( θ) = f( θ, n) In our case, x= ρ / and sin where. (18.68) n + 1 n+ 1( n+ 1)! Jn+ 1( βasin θ) n+ 1 ( βasinθ) θ reduces to, (18.69) f( θ, n) = (18.7) is the normalized pattern (neglecting the angular factors such as cosϕ and cosθsinϕ). The aperture taper efficiency is calculated to be Nikolova 14 8
εt = C 1 C C + n + 1 C(1 C) (1 C) + + n+ 1 n+ 1. (18.71) Here, C denotes the pedestal height. The pedestal height is the edge field illumination relative to the illumination at the center. The properties of several common tapers are given in the tables below. The parabolic taper ( n = 1) provides lower side lobes in comparison with the uniform distribution ( n = ) but it has a broader main beam. There is always a trade-off between low side-lobe levels and high directivity (small HPBW). More or less optimal solution is provided by the parabolic-on-pedestal aperture distribution. Moreover, this distribution approximates very closely the real case of circular reflector antennas, where the feed antenna pattern is intercepted by the reflector only out to the reflector rim. Nikolova 14 9
Nikolova 14 3