Note: Descriptions are shown in the official language in which they were submitted.
1 BhCKGROUND OF THE INVENTION
2 Field of the Invention
3 This invention relates to array antenna
4 systems and particularly to such systems wherein the
antenna element pattern is modified by providing a
6 lossless spatial filter between the antenna input ports
7 and the antenna elements so that the effective element
8 pattern associated with each input port is primarily
9 within a selected angular region of space.
_escription of the Prior Art
11 An array antenna system may be designed to
12 transmit a desired radiation pattern into one of a
13 plurality of angular directions in a selected region of
14 space. In accordance with the prior art designs of such
array antennas, each of the antenna elements has an
16 associated input port. By variation of the amplitude
17 and/or phase of the wave energy signals supplied to the
18 input ports, the antenna pattern can be electronically
19 steered in space to point in a desired radiation
direction or otherwise controlled to radiate a desired
21 slgnal characteristic, such as a time reference beam
22 scanning pattern. When it is desired to have an array
23 antenna radiate its beam over a selected limited region
24 of space, it is preferable that the radiation pattern of
the individual antenna elements also be primarily within
26 the selected angular region. This permits maximum
2-
1 element spacing while suppressing undesired grating
2 lobes.
3 In certain systems, control of the element pattern
4 by modification of the physical shape of the antenna
element may be impractical because of a desired element
6 pattern may require an element aperture size which
7 exceeds the necessary element spacing in the array. A
8 practical approach to overcome the physical elements size
9 limitation is to provide networks for interconnecting
each antenna input port with more than one antenna
11 element, so that the effective element pattern associated
12 with each input port is formed by the composite radiation
13 of several elements. These networks can be realized by
14 printed circuit techniques using a single substrate
layer.
16 One prior art approach to this problem has been
17 described by Nemit in U.S. Patent No. 3,803,625. Nemit
18 achieves a larger effective element size by providing
19 intermeidate antenna elements between the primary antenna
elements and coupling signals from the primary antenna
21 element ports to the intermediate element ports. This
22 tapered multielement apertyure excitation produces some
23 measure of control over the radiated antenna pattern.
24 A more effective prior art antenna coupling
network is described by Frazita et al. in U.S. Patent No.
26 4,041,501 assigned to the same assignee as the present
-3-
, ~,,, ,.,, ~
~: . ' ' :
1 invention. According to the technique oE Fra7.ita, the
2 antenna elements are arranged in element modules, each
3 module is provided with an input port. Transmission
4 li.nes are coupled to all of the antenna element modules
in the array. The transmission lines couple signals
6 applied to any of the ports to selected elements in all
7 the antenna element modules of the array. This antenna,
8 herein referred to as a COMPACT antenna, provides an
9 effective element aperture which is coextensive with the
array aperture.
11 Still another effective prior art antenna coupling
12 ne-twork is described by Wheeler in U.S. Patent No.
13 4,143,379 assigned to the same assignee as the present
14 invention. According to the technique of Wheeler, cross
coupling ports are employed to couple wave energy signals
16 to modules which are contiguous to each module.
17 Yet, another technique is shown in U.S. Patent No.
18 4,168,503 which describes an antenna array with a printed
19 circuit lens in a coupling network. A radiated signal,
received by each one of a plurality of spatially
21 separated antennas forming a directive array, is
22 coherently recovered by the lens. The lens comprises a
23 plurality of verticlaly standing and circularly arranged
24 printed circuit panels, each of which includes a
conductor strip connected at one end to each antenna. A
-4-
1 pluralit,y of semi-el~iptical circuit panels are affixad
2 to the vertical panels at a predetermined angle. Metal
3 strips plated on the semi~elliptical panels provide the
4 desired time delay to the antenna signa]s. A combining
strip couples the time delay strips and provides a
6 combined output signal at one end of the semi-elliptical
7 pattern. The angle at which the semi-elliptical boards
8 are a~ixed to the vertical boards corrects for time
9 delay distortion caused by the placement of the
combining strip. This configuration cannot be
11 implemented using printed circuit techniques on a single
12 substrate layer.
13 U.S. Patent No. 4,321,605 describes an array
14 antenna system ~laving at least a 2:1 ratio of antenna
elements to input terminals interconnected via primary
16 transmission lines. Secondary transmission lines are
17 coupled to and intersecting a selected number of the
18 primary transmission lines. Signals supplied to any of
19 the input terminals are coupled primarily to the
elements corresponding to the input terminal, and are
21 also coupled to other selected elements.
22 In time reference scanning beam systems such as
23 microwave landing systems (MLS)~ there may be a
24 linearity requirement for the glide path guidance i.e.,
:
~' ~
-5-
.
-
1 the difference between the actual and indicated angle
2 must be within a limited range. There is also a
3 requirement to minimize the field monitor distance for
4 the glide path antenna. Particularly in MLS, this
invention provides a non-thinned or fully filled array
6 which may be used to achieve linearity and minimize the
7 field monitor distance.
8 SUMMARY OF THE INVENTION
9 It is an object of the present invention to
provide an alternate array system having an antenna
11 element pattern formed by a spatial filter between the
12 antenna element input ports and the antenna elements.
13 It is another object of this invention to provide
14 a non-thinned antenna system i.e., an antenna system
wherein the number of antenna input ports equals the
16 number of antenna element output ports so that there is
17 no reduction ratio in the number of radiators to the
18 number of phase shi~ters.
19 It is another object of this invention to provide
an antenna system which does not generate grating lobes.
21 It is still another object of this invention to
22 provide a lossless spatial filter having a 1:1
23 input/output ratio which employs a minimum number of
24 couplers and terminations.
It is another object of this invention to provide
; 26 a lossless spatial filter having flexibility in
~ -6-
: ::
1 controlling the spatial filter radiation pattern,
2 meeting linearity requirements and minimizing field
3 monitor distances.
4 In accordance with the invention, the antenna
system radiates wave energy signals into a selected
6 angular region of space and into a desired radiation
7 pattern. The system includes a lossless spatial filter
8 having N input ports and N output ports. The aperture
9 of the system comprises a plurality of N antenna
elements. The antenna elements are arranged along a
11 predetermined path and each element is connected to only
12 one output port of the spatial filter.
13 A beam steering unit controls the direction of
14 radiation and includes N phase shifters and means for
controlling of phase shif-ters. Each phase shifter has a
16 phase shifter input port and a phase shifter output port
17 which is connected to only one input port of the spatial
18 filter. The antenna also includes a supply means for
19 supplying wave energy signals. The supply means
includes a signal generator supplying a power divider
21 having N output signal ports, each output port connected
22 to only one phase shifter input port.
23 For a better understanding of the present
24 invention, together with other and further objects,
reference is made to the following description, taken in
26 conjunction with the accompanying drawings, and its
27 scope will be appointed out in the appended claims.
7-
. ' :. . : ' ,
- , '
~D~
1 BRIEF _ SCRIPTION OF THE DRAWINGS
2 Figure 1 i9 a conceptual diagram of an antenna
3 system including a three level spatial filter wherein
L~ signa]s applied to an antenna input port are provided to
5 the antenna element associated with the port and to the
6 antenna elements adjacent to the associated element.
7 Figure 2 is a plan view of a printed circuit
8 coupling network of the three level spatial filter
9 illustrated in Figure 1.
Figure 3 i9 a conceptual diagram of an antenna
11 system in accordance with the present invention
12 including a three level spatial filter cascaded with a
13 four level spatial filter.
14 Figure 4 is a plan view of a printed circuit
coupling network of the cascaded spatial filters
16 illustrated in Figure 3.
17 Figures 5A, 5B and 5C are antenna patterns for
18 antennas according to the invention employing spatial
19 filters having two level, three level and four level
coupling, respectively.
21 Figure 6A illustrates a schematic diagram of a
22 coupler and its relative inputs and outputs.
23 Figure 6B is a listing of the formulas which
24 define the coupler values and the termination values.
Figure 6C illustrates a schematic diagram of a
26 series coupler network.
:
.
~$~
l Figure 6D is a generalized schematic
2 representation of a five level spatial filter.
3 Figure 7 illustrates a prototype network for an
4 infinite spatial filter antenna to be employed with the
invention.
6 Figure 8A is a schematic diagram of an antenna
7 system of two cascaded 8-coupler spatial filters
8 according to the invention.
9 Figure 8B is a table of the optimum excitations
for an 8-port spatial filter according to the invention.
11 Figure 8C is a schematic diagram of a unit cell
12 of a modular antenna system of two cascaded 4-coupler
13 spatial filters according to the invention.
14 Figures 9 and 10 illustrate a computed antenna
pattern for the zero-thinned spatial filter shown in
16 Pigure 8A.
17 Figure 11 is a graph illustrating the linearity
18 requirements which limits the deviation from the ideal
19 linear relationship of the MLS guidance angle and the
actual angle.
21 Figure 12 illustrates the geometry and formulas
22 of a model of a flat horizontal surface used to quantify
23 the effects of sidelobe radiation on the performance of
24 an automatic flight control system.
Figures 13, 14, and 15 summarize the simulation
26 rasults of vertical acceleration, vertical velocity, and
g_
. - , ,
.,. , , ;.
~;~$i~
1 vertical attitude, respect;vely, with regard to the peak
2 MLS guidance error for 10 feet and 20 feet elevation
3 antenna phase center heights when passenger comfort is
4 considered.
DETAILED DE~CRIPTION OF l'HE INVENTION
6 Figure 1 is a schematic diagram illustrating an
7 antenna system in accordance with the present
8 invention. The diagram of Figure 1 includes a plurality
9 of antenna elements 1~8 arranged in a predetermined path
which, in this case, is a straight line. Each antenna
11 element is connected to one and only one output port
12 9-16 of spatial filter 17. The spatial fllter is
13 comprised o~ a plurality of modules A through H, one
14 module for each antenna element. Spatial filter 17
includes 8 input ports, 18-25 each connected to the
16 output of one and only one phase shifter 26-33. The
17 array of phase shifters 26-33 form beam steering unit
18 34. The inputs 35-42 of the phase shifters are
19 connected to one and only one output of power divider 43
20 which is fed by signal generator 44. The power divider
21 and signal genera'~or form a supply means for supplying
22 wave energy signals. Although filter 17 has been
23 illustrated as symmetrical, it is contemplated that
24 spatial filters according to the invention may be
unsymmetrical.
26 Referring to the signal path of wave energy
- 10 -
~:
"~. ~.. , ,^` ` .
~L~
1 signal supplied by signal generator 44, the original
2 signal is provided via line 45 to power divider 1l3 which
3 divides the signal into eight equal components. Fach
l~ component is provided via lines l~6-53 to only one input
of beam steering unit 3L~. For example, referring to the
6 left-most portion of the antenna system, line 46
7 provides the signal component to input 35 of beam
8 steering unit 34. The component then passes through
9 phase shifter 26 which may shift the phase of the
component according to instructions received from
11 control uni~ 54 via control line 55. The output of
12 phase shifter 26 is provided to input port 18 of spatial
13 filter 17. The signal component provided to input port
14 18 is provided to output port 9 which is connected to
antenna element 1 and is also provided by a coupling
16 arrangement to element 2 which is adjacent to antenna
17 element 1,
18 Spatial filter 17 couples component signals which
19 are provided to any input to the antenna element
associated with the input and to elements adjacent to
21 the associated element. 50uplers 56-62 couple signals
22 which are provided to an associated antenna element to
23 the antenna element which is to the left of the
24 associated antenna element. The component signal
provided to an input is transmitted to the antenna
26 element associated with the input by transmission lines
27 64-71. For example, the component signal pro~ided by
-1 1 -
:
. .~ . . .
branch 39 of the power divider 43 is fed through phase
2 shifter 30 and provided to input 22 of spatial filter
3 17. Input 22 is connected by transmission line 68 to
4 its associated output 13 and antenna element 5. The
component signal is also coupled by coupl,er 59 to
6 antenna element 4 which is to the left of and adjacent
7 to antenna element 5. Similarly, component signals
8 provided to an input are also coupled to antenna
9 elements adjacent and right of the associated antenna
element by couplers 72-80. For example, the co~ponent
11 signa.l provided by branch 49 of the power divider to
12 input 38 of phase shifter 29 passes through phase
13 shifter 29 and is provided to input 21 of the spatial
14 filter 17. The component signal is then provided to
output 12 by transmission line 67. Output 12 is
16 directly connected to antenna element 4. Element 5 is
17 adjacent to and to the right of antenna element 4 and
18 receives a portion of the component signal via coupler
l 9 76 . Element 3 is adjacent to and to the left of antenna
element 4 and receives a portion of the component signal
21 via coupler 58.
22 Spatial filter 17 is shown in modular form. As a
23 result, the input to coupler 72 is terminated by
24 termination 81 because there is no antenna element to
the left of antenna element 1. Similarly, the output
26 from coupler 56 is terminated by termination 82 because
27 there is no antenna element to the left of antenna
- 1 2 -
. .. ....
~JI~
element 1 to receive the component signal provided to
2 input 18. On the right side of spatial filter 17,
3 coupler 80 is terminated by termination 83 and coupler
4 ~3 is terminated by termination 84 because there is no
antenna element to the right of antenna element 8 to
6 receive the couple signal from coupler 80 or to provide
7 a coupled signal via coupler 63.
8 Figure 2 illustrates a plan view of a printed
9 circuit coupling network useful as the spatial filter 17
10 OI Figure 1. Network 17 includes input ports 18-25
11 connected to the outputs of beam steering unit 34.
12 These input ports are connected to a first series of
13 couplers C1 shown in detail in Figure 2A. Coupler C1 as
14 well as all other couplers may be standard microstrip
15 network couplers having a predetermined coupling
16 ratio. The specific coupling ratio depends on the
17 width, length and on the thicknesLs of the transmission
18 lines within the coupler. ~y convention, signals
19 provided to the inputs 101 and 102 of coupler C1 are
20 coupled to the outputs 103 and 104 according to a
21 predetermined ratio. In the case of coupler C1, input
22 102 is terminated by termination 105 resulting in any
23 component signal which is supplied to input 101 being
24 distributed to outputs 103 and 104 such that
C 1 2 + T 1 2 = 1 .
26 Following the firs-t array OI couplers C1 is a
27 second array of couplers C2 illustrated in more detail
': ~ ` ' , '';
3~
1 in Figure 2~. Signals provided to inputs 105 and 106
2 are combined and transmitted to output 108 at a ratio T2
3 and coupled to output 107 at a ratio C2 such that
4 T22 + C22 + 1. Completing the three level spatial
filter 17 is a third series of couplers 109-116.
6 According to the invention, these couplers have the same
7 configuration as coupler C1. Couplers 109-116 work in
8 the same manner as coupler C1 as shown in Figure 2A by
9 combining signals provided to their inputs to the
outputs 9-16 o~ spatial ~ilter 17.
11 As specified by the invention, spatial filter 17
12 is ideally lossless (except for dissipative losses) and
13 for that reason the relationships
14 Cl + T12 = 1 and C22 + T22 = 1
must apply to the power (voltage) passing through each
16 coupler C1 and T1, respectively. The following
17 relationship ensures the lossless condition for the
18 network:
19 C1 = 2 (1 + ~1 - C2 ~ (1)
This relationship can be derived by setting the
21 inputs at 18-25 equal to unity and the inputs to the
22 terminations 117-124 equal to zero.
23 As used in rega~d to the invention, a non-thinned
24 spatial filter is a filter formed by an array of
couplers. The array is essentially lossless in that the
26 power dissipated within terminations is minimized.
27 Figure 3 is a schematic diagram of an antenna
.
1 system in accordance with the invention including a
2 three/four level cascaded spatial filter 300. In
3 general, this spatial filter may be used in combination
4 with the antenna system as shown in Figure 1 by
replacing spatial filter 17 with spatial filter 300.
6 Each antenna element 1-8 would then be connected to one
7 and only one output port 301 of the spatial filter
8 300. Spatial filter 300 is comprised of a plurality of
9 modules A through H, one module for each antenna
element. Spatial filter 300 includes input ports 302
11 each connected to one and only one of the outputs of a
12 phase shift network.
13 Figure ll is a plan view of a printed circuit
14 coupling network of the cascaded spatial filter 300
illustrated in Figure 3. Network 300 includes input
16 ports 302 connected to the output ports of a beam
17 steering unit. These input ports are connected to a
18 first series of couplers C1 shown in detail in Figure
19 2A. Following the first array of coupler C1 is a second
array of couplers C2 illustrated in more detail in
21 Figure 2B. Following the second array of couplers C2 is
22 a third array of coupler C2. Completing the four level
23 spatial filter 300 is a fourth series of couplers C1.
24 According to the invention, for symmetrical excitations,
couplers C1 at the beginning and end of the array and
26 intermediate couplers C2 have the same configuration.
15-
'
.
l The following relationship ensures the lossless
2 condition for the networks
3 C12 = - + C2 ~1 - C2 (2)
4 Figure 5A illustrates an ideal antenna pattern
for an antenna according to the invention,employing
6 spatial filters having a two level coupling.
7 Essentially this coupling creates lobes 501, 502 and
8 503. Figure 5B illustrates a typical antenna pattern
employing a three level spatial filter which for~s a
single lobe 504. Figure 5C illustrates a typical
11 antenna pattern for a four level spatial filter
12 generating a more well defined single lobe 505.
13 Synthesis Procedure For Five
14 Level Non-Thinned Spatial Filter
Step 1: Referring to Figures 6A, 6B, 6C, and 6D,
16 determine initial values for couplers C1--C5
17 (a) specify desired excitations A1-A5
18 (b) specify C1
19 (c) compute C2-C5 Using Figure 6C
Step 2: Compute actual excitations A1'-A5' according
21 to the following formulas:
22 (a) A1' = T5C4C3C2C1
23 (b) A2' = C5T4C3C2C1 - T5T4T3C2C1
- T5C4T3T2T1 - T5C4C3T2T1
:
~ -16-
.
'
,
t~
1 (c) A3' = C5C4T3C2Cl - T5C4T3C2T1
2 - T5T4T3T2Tl - T5T4C3T2C1
3 - C5T4C3T2Tl - C5T4T3T2Cl
4 (d) A4' = C5C4C3T2C1 - T5T4C3C2T1
- C5T4T3C2T1 - C5C4T3T2Tl
6 (e) A5i = C5C4C3C2T1
7 Step 3: Adjust values for couplers C2-C5
8 (a) adjust C5 such that
A2' A2
A1' A1
(b) adjust C4 such that
11 A3' A3
A2' A2
12 (c) adjust C3 such that
13 A4' A4
A3' A3
14 (d) adjust C2 such that
A5' _ A5
~, _ ~
16 Step 4: Recompute actual excitations A1' - A5'
17 (see Step 2 for formulas for A1' - A5')
18 Step 5: Normalize actual excitations by computing
19 A1" - A5"
(a) Let A1" = 1 . Then,
-17-
'' ":
.
$~
(b) A2" = A1,
2 (c) A3n = A3
3 (d) Al~n = A4
4 (e) A5" = A1
5 Step 6: Compute deviation S between normalized
6 actual excitations A1" - A5" and desired
7 excitations A1 - A5
8 S ~ (AN" - AN) N = 1,2,.......... ,5
9 Step 7: Repeat steps 3-6 until deviation S is within
an acceptable limit
11 Step 8: Repeat steps 1--7 until ratio of power in
12 terminations PT to radiated power PR is a
13 minimum i.e., minimize PT
R
PT ~(TN) ; PR = ~[AN)2 N = 1,2,...,5
:: ~
15~ For example, consider the case o~ a five
16 element aperture as illustrated in Figure 6A. Assuming
: :
17 the desired excitation (from step 1a) is:
18 A1 = 1.0000
-18-
.~.,
A2 = 1.6086
2 ~3 = I.93156
3 A4 = 1.6086
4 A5 = 1.0000
Let C1 = 0.979 (from step lb); then, the
6 values of the other couplers (from step 1c) are:
7 C2 = 0.9502
8 C3 = 0.9366
9 C4 = 0.9600
C5 = 0.9852
11 The normalized actual excitations (steps
12 2-5) result in:
13 A1 = l
14 A2 = 1.3755
A3 = 1.6478
16 A4 = 1.5449
17 A5 = 1.1957
18 The db loss (from step 8) between the
19 normalized actual excitations (from step 5) and the
::
~; 20 desired excitations (from step 1a) is:
21 LOSS = 7.12db
22 Table 1 below continues the synthesis
23 procedure.
~;~ T l C1 C2 C3 C4 C5 A11' A2'' A3'' A4'' A5'' loss
.979 .9225 .8401 .9042 .979 1 1.6061 1.932 1.6061 1 6.72db
2 .98 .9285 .857 .9132 .98 1 1.608 1.932 1.611 1 6.59db
24 3 .985 .953 .9155 .9461 .985 1 1.608 1.933 1.604 1 6.69db
4 .99 .971 .9523 .9685 .99 1 1.608 1.931 1.609 1 7.68db
5 .981 .9343 .8718 .9212 .98 1 1.6085 1.932 1.6085 1 6.53db
-19-
,
~ .
~::
:::
~ 3
l Table l
2Five Coupler Synthesis
3As shown in table l, trial 5 illustrates an
4 optimum arrangement with minimum power loss. As shown
in table 2, trial Ll illustrates an optimujm arrangement
6 for a five coupler structure where the symmetry of the
7 excitation is invoked to set C5 = Cl and C4 = C2.
Trial C1 C2 C3 A1 A2 A3 loss
l .981 .91506 .85575 1 1.6086 1.932 6.93db
8 2 .979 .8823 .7849 1 1.6086 1.9318 7.90db
3 .982 .92425 .8739 l 1.6086 1.932 6.80db
4 .984 .93866 .90095 l 1.6086 1.9321 6.75db
.986 .95011 .92131 1 1.6086 1.932 6.90db
9 Table 2
Five Coupler Synthesis, C5 = Cl, C4 = C2
11 Although the above procedure has been applied to
12 develop a symmetrical filter, the procedure is general
13 in nature and can also be used to develop non-
14 symmetrical filters. Symmetry is generally preferred to
maintain simplicity and reduce complexity. Symmetrical
16 filters usually employ redundant couplers and other
17 structures which minimizes design efforts.
18 The design of a spatial filter involves the
19 determination of coupler values ~or a multilayer
circuit. No closed form solution is readily apparent to
21 the synthesis of a network that produces a specified
22 output voltage distribution. However, analysis of any
-20-
:
., ~,......
a~'~9~3
1 network is possible. There~ore, synthesis involves the
2 iterative trial and error procedure described above in
3 which coupler values are gradua1ly adjusted until the
4 desired outputs are achieved.
Since the analysis of a complex network requires
6 significant computer time, it is desirable to formulate
7 an iterative algorithm that converges to the desired
8 solution within a reasonable time. Analysis of every
9 possible combination of coupler values could take weeks
or months to e~aluate on the computer. Furthermore, an
11 infinite number of solutions exist that produce the
12 desired amplitude distribution. The difference in
13 solutions is the insertion loss of the resulting
14 network. Therefore, it is necessary to
determine by theoretica] means the minimum possible
16 loss, so that it will be known when an optimum solution
17 has been achieved.
18 The theoretical loss of a spatial filter network
19 is determined by conservation of power considerations.
The network prototype is shown in Figure 7. The network
21 is symmetrical and continues to infinity in both
22 directions. Each input excites a sub array with N
23 outputs. The sub array outputs, resulting from adjacent
24 inputs, overlap. The network shown in Figure 7 has an
equal number of inputs and outputs. Therefore, the
26 input and output spacings are equal and, when all inputs
27 are excited, each output port will be the sum of
-21-
1 contributions from N input ports. There must be an
2 internal termination for each output port.
3 The output excitation that results from input 1
4 is designated A1(N), whereas the output excitation
resulting from input 0 is designated AO(N?. Because the
6 networ~ is symmetrical, A1(N) = AO(N) = Aj(N).
7 Similarly, the power terminated, designated as Bj(N),
8 must also be equal.
9 The network is realized with N layers of
directional couplers. To achieve the desired symmetry,
11 all coupler values in a given layer must be equal.
12 Furthermore, a symmetrical output excitation (Aj(1) =
13 Aj(N), Aj(2) = Aj(N-1), etc.), requires that the coupler
14 values in the first layer be equal to those in the Nth
layer, etc. Therefore, as an example, an 8-output
16 network has 8 layers of couplers. If the 8-element
17 excitation is symmetrical, C1 (coupling value for all
18 couplers in first layer) must equal C8, C2 = C'7, C3 =
19 C6, and C4 = C5. Therefore, there are only 4 different
coupler values or unknowns that must be determined for
21 an 8-output network.
22 When input power is delivered to port one,
23 conservation of power dictates the sum of powers in
2ll A1(N) added to that internally terminated (B1(N)) must
equal the input power. A normalization to an input
26 power of 1 watt yields the equation:
-22-
N 2 N 2
~ A1(N) + ~ B1(N) = 1
1 i=1 i=1
2 The A~s and B's are voltage coefficients. The
3 power at each output port is equal to the square of the
4 voltage coefficierlt when the system imped,ance is
normalized to one ohm.
6 ~hen all input ports are excited with equal power
7 and in phase, the output at each port is the sum of N
8 voltages. From symmetry and conservation of power, the
9 sum of the power at one output port and its internal
termination must equal one watt. All output ports will
11 be equal.
N 2 N 2
12 ~ A1(N) + ~ B1(N) = 1 (4)
i=1 i=l
13 A co~bination of equations (3) and (4) gives:
N 2 N 2 N 2 N 2
14 ~ A1(N) + B1(N) = A1(N) ~ ~ B1(N) (5)
i=l i=1 i=1 i=l
If the network is to be lossless when a single
16 input port is excited, no power can be delivered to the
17 internal terminations (all B~s = O). If that condition
18 exists,
N 2N 2
19 A1(N) = ~ A1(N) (6)
~ ~ i=l i=l
-23-
:~;
.,
6~
1 There are few output excitations that satis~y
2 equation 6. The least loss occurs for an excitation
3 that does not satisfy equation 6 when
N 2 N
1~ ~ Bl(N) = 0 or ~ B1,(N) = 0 (7)
i=1 i=1
When that condition is met, the network will be
6 lossless when all input ports are excited with equal
7 amplitude and phase. The 105s, when a single input port
8 is excited and the sub array pattern has a maximum in
9 the in-phase direction, is given by:
N 2 N 2
loss = ~ A1(N) ~ ~1(N) (8)
i=1 i=1
11 ~hen the sub array pattern has a maximum in a
12 direction other than the in-phase direction, the lower
13 bound on the loss is increased by the difference in the
14 sub array gain in the two directions. The optimum
networ~ is one that provides the least loss. The loss
1:6 that can be expected is the difference between the
17 computed network loss and the theoretical value. Thus,
18 if one computes the theoretical minimum loss to be 3.1
19 dB when a single input port is excited using equation 8,
and the least loss that can actually be achieved with a
21 realizable network is 4.6 dB, it will be found that the
22 loss, when all inputs are excited in phase, is 1.5dB.
23 This 1.~ dB loss results from the consideration of the
-24-
.
'
1 center of the sub array pattern. When the array is
2 scanned to the sub array peak the theoretical loss is
3 reduced to zero.
Il The basic spatial ~ilter network topologies are
well-known. A preferred implementation requires 17
6 layers and is nearly impossible to synthesize. A
7 practical network, that closely approximates the
8 per~ormance of a l7-layer network, uses two cascaded
9 8-layer networks as illustrated in Figure 8. The
pattern characteristics for this network are shown in
11 Figures 9 and 10 for a radiating element spacing of 0.79
12 wavelengths.
13 Figure 11 describes the linearity requirement for
14 MLS glide path guidance. The discussion of linearity
concentrates on the ele~ation guidance per~ormance,
16 however, linearity is also a requirement for the azimuth
17 guidance. Linearity is a subject that has generated
18 much discussion in the MLS community. The invention
19 provides a phased array antenna which meets the
elevation linearity requirement. The spatial ~ilter
21 network is a practical way to satis~y the low effective
22 sidelobe requirement which is directly related to the
23 linearity requirement.
24 The linearity ~autopilot) requirement limits the
deviation ~rom the ideal linear relationship of the MLS
26 guidance angle and the actual angle (see Figure 11). It
~: :
-25-
.
¢ ~
1 specifies the transverse accuracy characteristic of the
2 angle guidance signal as opposed to the longitudinal
3 characteristics of PFN and CMN. The longitudinal
4 characteristic causes the aircraft to deviate from the
glide path (bends~ or generates noise-lik,e action of the
6 controls. The transverse characteristic is capable of
7 causing instability in an automatic flight control
8 system.
9 After several years of discussion within the MLS
community it is now generally accepted that PFN, CMN and
11 linearity for the EL guidance equipment are all
12 dependent on the effective sidelobe level of the
13 antenna. The issue has been which one of the three
14 characteristics (PFN, CMN or linearity) is the driver
with respect to the specification of the effective
16 sidelobe level. The Path Following Noise (PFN) relates
17 to the path following mean course error and is caused by
18 any frequency component that an aircraft can follow.
19 The Control Motion Noise exists in situations where
there is no PFN but the scanned MLS signal indicates a
21 bounce or deviation which an aircraft cannot follow.
22 initially it was argued that PFN was the driver. The
23 effective sidelobe level required to ensure that the PFN
24 for a 1.5 beamwidth antenna does no~ exceed 0.083 is
-25 dB (a 0 dB ground reflection coefficient is assumed,
26 the 0.0830 PFN limit is derived t'rom the ICA0 standard
27 that the PFN shall not be greater than plus or minus l.3
-26-
1 feet). After some analysis by the FAA, it was
2 recogni2ed that with the antenna phase center 20 feet.
3 above the reflecting ground, CMN could be generated when
4 the aircraft was within 2000 feet. of the runway
threshold. Consequently, in the draft sp,ecifications
6 for the FAA second MLS procurement, the effective
7 sidelobe level is specified such that the CMN does not
8 exceed 0.0ll5. This requires an effective sidelobe
9 level of -30 dB for a 1.5 beamwidth antenna.
Based on the results of simulations of an actual
11 automatic flight control system in service it has been
12 concluded that linearity is the most stringent
13 requirement with respect to the specification of the
14 effective sidelobe level. The results of the
simulations indicate that the angle error limit must not
16 exceed 0.02~ to ensure performance of an automatic
17 flight control system within passenger comfort levels.
18 This error limit corresponds to a -36 dB effective
19 sidelobe level for a 1.5 beamwidth antenna.
The discussion on the linearity requirement has
21 raised the issue of the measurement methodology for
22 determining compliance with specifications. With regard
23 to this issue it should be recognized that effective
24 sidelobes can be measured on an antenna range and that
design approval by an authority can be based on these
26 antenna range measurements.
::~
-27-
~ ' , .
~ "
.
1 The ~idelobes radiated by the elevation antenna
2 in the direction of the ground are folded back on the
3 main beam because of specular reflection. The sidelobe
ll radiation distorts the beam and causes PFN, CMN and
linearity errors. The specification of P,FN and CMN
6 limits the magnitude of the angle guidance error. The
7 linearity error, however, depends on the product of the
8 maximum angle guidance error and the height of the
9 antenna phase center above the reflecting ground
surface. A large error-height product is capable of
11 causing substantial degradation of the guidance loop
12 gain of an automatic flight control system to the point
13 where the automatic flight control system becomes
14 unstable. For example, a maximum error of 0.045 and a
phase center height of 20 feet can cause the loop gain
16 to vary between +6 dB and less than -40 dB (at the "max
17 gain spot" and the "dead spot", see Figure 11).
1~ The model of a flat horizontal surface is used to
19 quantify the effects of sidelobe radiation on the
performance of an automatic flight control system. The
21 geometry and formulas are presented in Figure 12. For
22 the case of a constant glide path, the magnitude of the
23 error remains essentially constant and the phase
24 variation is that attributed to the path difference
between the direct signal and the indirect signal
26 emanating from the ground image of the EL antenna.
-28-
~ ' ' , ' '
' ' ' ' ''
1 The model was used as a perturbation input to a
2 simulation of an automatic glide slope control systeM
3 for a small jet aircraft. The criteria for the
ll acceptability of the automatic flight control system is
passenger comfort. Figures 13, 14 and 15~provide a
6 summary of the simulation results with respect to the
7 allowable peak MLS guidance error, elevation antenna
8 phase center height and passenger comfort. The
9 simulations start at a distance of 3 r~M from the
elevation antenna.
11 Figure 13 shows that for a 20 feet phase center
12 height and a peak error of 0.0830 the automatic control
13 system is unstable. The vertical accelerations exceed
14 the passenger comfort level by a factor of 2. 4 :1 . For
peak error of 0.045, the system is marginally stable;
16 for larger phase center heights, say 37 feet, it is
17 expected that the system would be unstable (the error-
18 height product, 0.045 X 37', is equal to that of the
l9 0.083~ maximum error and 20 feet height case). Figures
l 4 and 15 exhibit the same trends; they show that for a
21 20 feet phase center height and a peak error of 0.0830
22 the vertical velocity and attitude exceed the passenger
23 comfort levels by factors of 4:1 and 2:1 respectively.
24 The following conclusions are based on a study of
the available information, with respect to the
26 specification of the effective sidelobe level and
27 autopilot performance within passenger comfort levels:
-29-
: :
,
1 1. the present PFN error limit (0.083) is not
2 acceptable;
3 2. the present CMN error limit (0.045) is
4 marginal (especially if higher than 20 feet
antenna phase center heights ,are
6 contemplated);
7 3. a limit of 0.0240 appears to be acceptable
8 ~or tlne case studied;
9 4. linearity is the dominant system requirement
with respect to the speci~icaticn of the
11 effective sidelobe level; and
12 5. the error-height product should not exceed
13 0.45 degrees-feet.
::;
:: ~
-3-
~:
:
:
