UNCLASSIFIED404IL90AD
DEFENSE DOCUMENTATION CENTERFOR
SCIENTIFIC AND TECHNICAL INFORMATION
CAMERON STATION. ALEXANDRIA, VIRGINIA
UNCLASSIFIED
NOTICE: Whcn government or other drawings, speei-fications or other data are used for any purposeother than in connection with a definitely relatedgovernment procurement operation, the U. S.Government thereby incurs no responsibility, nor anyobligation whatsoever; and the fact that the Govern-ment may have formulated, fturniaed, or in any waysupplied the said drawings, specifications, or otherdata is not to be regarded by implication or other-wise as in any manner licensing the holder or anyother person or corporatiun, or conveying any rightsor permission to manufacture, use or sell anypatented invention that may in any way be relatedthereto.
/ -ARCH'c5N_ LOG-PERIOD1• ARRAYS OF SLOTS
by
J. W. GREISER
Januar/ 1964
Technical Report No. 3
Contract No. NOBSR 85243
Index Number SS024001
Sponsored byBUREAU OF SHIPS
DEPARTMENT OF THE NAVY
Washington 25, D.C.
ANTENNA LABORATORYDEPARTMENT OF ELECTRICAL ENGINEERING
ENGINEERING EXPERIMENT STATION
UNIVERSITY OF ILLINOISURBANA, ILLINOIS
it
S •ASEARCH 01 LOG-PERIODIC ARRAYS OF SIDTS
by
J. W. Greiser
January 1964
Technical Report No. 3
Contract No. NOBSR 85243
Index Number S8024001
Sponsored by
BUREAU OF SBTPS
DEPARTMENT OF THE NAVY
Washington 25, D. C.
ANTENNA LABORATORYDEPARTMENT OF ELECTRICAL ENGINEERING
ENGINEERING EXPERIMENT STATION
UNIVERSITY OF ILLINOISURBANA, ILLINOIS
ACKNOWL1~eWMENTr
The author is pleased to acknowledwe the --zgestions and aid of Project
Director P. E. Mayes. Discussions with 'ýeveral other members of the Antenna
Laboratory staff were often enlightening. A special debt of gratitude is due
Messrs. G. Ohiesney, R. Griswold, and S. Guccicre for their efforts in theconstruction and testing of the antenna =odeis, and for their drafting of the
illustrations in this report. This work was supported by the Navy Bureau of
Ships under Contract NORSR 85243
TABLE OF CONTENTS
Page
1. Introduction 1
2. Miscellaneous and Preliminary Designs 1
2.1 Log-Periodic,Zigzag Slot Antennas 22.2 Log-Periodic Simple Slot Arrays 4
"2.3 Preliminary Log-Periodic Folded Slot Arrays 4
3. Log-Periodic Folded Slot Arrays and Their Relations 6
3.1 Coaxial Cable Phasing Lines 63.2 Printed Circuit Log-Periodic Folded Slot Arrays 63.3 Cavity Backed Slot Arrays 17
4. Conclusions and Future Plans 23
References 24
LIST OF ILLUSTRATIONS
Figure No. Page
la Log-periodic zigzag slot antenna 3
lb Zigzag slot with loop coupling 3
2a Log poriodic ariay of slots with loop coupl1t
2b Capacitively coupled log-periodic slot antenna 5
3a An early log-periodic folded-slot array 7
3b Log-periodic folded slot antenna with phasing lines 7
4 Radiation patterns of SA-2B, T = .81, a. = 22.51 8
5 Log-periodic folded element arrays 10
6 A family of log-periodic folded element arrays 11
o 07 Radiation patterns of FSA-l, T = .78j aE= 25 aS = 12.5 12
8 Radiation patterns of FSA-l, T =.78, a= 25 a = 12.5' 13
9 Impedanco of FSA-1, T = .78, aE -25, a. =12. 15
10 Radiation patterns of FSA-5, T =.82, aE 18°, as 9.5° 16
11 Radiation patterns of FSA-3E, T = .8, a = a5 aS =100 18
12 Effect of the addition of cavities 20
13a Impedance reduction with folded elements 22
13b Unilateral far-field cancellation of slot by loop 22
1. INTRODUCTION
Probably the last remaining application of log-periodic antenna principles
that bas not been really satisfactorily developed consists of slot arrays.
There is a need for an efficient, frequency-independent, flush-mounted design
to be used in aircraft, missiles, and hardened-site situations. A few designs
have been reported in the literature '2 but they suffer from either excessive
size, or a rather narrow bandwidth. Research on a promising design is being
carried out at the present time and may, in the end, provide one solution to
the problem.3 The philosophies with which these investigators have approached
the problem are somewhat varied. Schemer and Isbell begin by considering the
strict electromagnetic dual of a LP dipole array but are forced to deviate when
faced with the problem of constructing a satisfactory dual of the dipole array's
twisted feed line. Mittra and Wahl begin with a nniformly-periodic corrugated
surface antenna---- then attempt to tap(-r it ioto a log-periodic device. Mikenas
and Mayes start out with a . log-periodic array of slots with individual cavity
backings and study the prohlem of coupling to the cavity fields in a frequency-
independent manner. None of these designs has been able to reach the standards
of performance achieved by the free-space LP dipole array over large bandwidths.
Nor do the designs to be discussed in this report reach that dtandard of
pprformance in a cavity-backod form. Howey,', a symmetrical slot design (i..e.,
one that radiates on both sides of the ground plane) has shown excellent wide
band pattern performance for a range of physical parameters. In addition, this
new slot design has an exact electromagnetic dIal in free space; and further,
- - - the free spaco dual has image symmetry along .ts axis, thus allowing one half
of the structure to be operated over ground as a LP monopole array. No 6ther -
known frequency-independent design possesses this high degree of symmetry and
the resulting wide versatility. All three forms of the new design have demon-
strated similar wide band pattern and impedance characteristices.
2. MISCELLANEOUS AND PRELIMINARY DESIGNS
Many a designer of tog-periodic antennas has expressed the wish for a
catalogue oi unsuccessfv: devices since it Is virtually impossible for the
2
enugineer to predict with certainty whether a given new design will. be a success
or' a fai lure. And because authars seldom list their failures in technical Papers,
soce of these are apt-to be repeasted several times. Therefore, a few marginal
designs and iailures will be reported here.
One of the major ideas which has guided the present work on slot aryays is
-Int vicwpoitnii tat the fcwcr t~he varrabico ur parameters in a probliem, the easier
thet solut ion w~tl he, For this reason all initial work was done on bilaterally
symmetric slot arrays. it was felt that if these devices were successful, the
addition of a cavity backing would be a secondj less difficult atep. Another
guiding principle, was the belief that. approximately the same physical parameters
satisfactory IV,- existing LP designs would also be suitable for the !,lot arrays.
This- lat ter prinvtpb', hns htpon fouind true-for the new slot arrays and their related
duals. The designs discussed below are not taken up in chronological order; although
it might be meentiotted that the earliest of theta was conceived in December 1961.
2A1 Log-Periodic Zigzag Slot Antennas
The ztgvag slot angennia shown in Figure la wns inspired by the successful
fret-space Lp Vigzauf. The new device did provide good backfire patterns over most
titt its design banid; however-, it had two ttajor defects: 1. Serious pattern breakup
OCCttrrted uver narrow bands of frequency at otto or two poitnts in the designt band;5
2 Beam waJggtttg Was U Persistent problem. Thu shorting straps, whose endpoints
alt indteautd by dots itt the drawing, were found necessary fur back-fire operatiotn.
thecy astcnt jul Ly are short lengths of conductour whnichi break the- zigzag slot into
,4igeetits while forctng the opposite sides of the ground plane to the same potential.
fit spitfe of Ithe desigttt
s marginal performaince, an open sloping cavity was added in
otto oi the models, The distance f roe each slot to the bottom of the cavity was
adittated it) kA. at the slut resunanens. Untortunately, the open cavity (i.e., tio
itittritor partitions) almost completely destroyed the backfire characteristic of
111L 711,zag slot Clemnttt,
A stceond form of the; Vigzag slot antentit, which was suggested by P. E. Mayes,
uscd a small hoop to couple to the magnetic field of fthe slot as shown in Figtitre lb1.
The1se deVices -xhibited very similar patterns to the models just discussed while not
suhiertog as muchl from the occastonal pattertn breakup. Beam wagging 'still remained
t_%cti tit the best models, and a cavity backing again obliterated the backfire effect.
3
SHORTINGG -
~~STRAPS , ,
• STRAPS1
Figure Ia. Log-periodic zigzag slot antenna
Figure lb. Zigzag slot with loop coupling
4
2,2 Lo~g-Pernj';(-c Simple Slot Arrays
it is possible to s'tilaie thle twisted feed line of the log-pcriodir, dipole
arrays by coup' up to the magncl ic fields of the slots with properly oriented
,.oups, Fight.- 2a illustrates oie. such design using alternately wound square loaps.
liutod loops have alIso heen tri~ed. bu for loops of small area the difference is
pOblh~bfy s.light . Rathfer good backfire patterns were measured over parts of the
ccsign bandwidth. but, serious break-tip appeared in other portions. Although the
results were etiouraging, study of these loop-coupled arrays was discontinued in
favor of more promising designs.
The only .1 he~r lug-periodic array of simple slots that will he discussed here
ts the capacti ivcly coupled array of Figure 2b. It is analogous to the monopole6
arrays of Wirkci-fiham. Blasically it was hoped that sufficient voltage differences
.ýoufd lie deve~oI~ed across the gaps in (ihe leg-periodically segmented outer conductor
ol hL coaxial fued line to excite the slots, The line was moved off the uxsof
luit slut arrasy tar impedance matching. All models of this type showed persistent
bidirectional ttatterna. indicatinug insufficient coupling of energy from tile feed
l ive to thu- slot fields. lIi addition thie unsymmetrical feed system causued beam
squint at numerous fre-quencies.
1.3 Preliminary La.g-Puri odic Folded Slot Arrays
Past expeltience in tiug-puriod~c aiteunta design indicates that the-impedance of
hei radiat ing e emeulas and tile tmpedance of the fead syster, should ftat deviat~e too
-.zdely if goutd vtitpting, and suhauquent low VSWRt are to be achieved. This criterion
t,; difficultl, al best. to satisfy when arrays of simple slots are used because
¶hiit-soaotil epedance is abo-.c. 400 42 withoul a cavity backing and 800 f with a
resoinant cavity backing. It is well-known that the iepedancc,Z ,of a two-
dhicensional siot) antf Z of its free space dual beat tile following relation
a i 4
t~hus, if thu fretý space diptilf has anl impedance of Z d =72 thle dual. slot, impedanice
-Il heb Z 19:3~ Ou the ether bond, ii the electromagnetic dutal of a folded
iitIpult- ort a foldited tilt ipu h is t atket, litfietatices of 125 Q and 55 2~sftutld result..
Thseiattcs havt hi,1 ti iqIer imritaf ly Verified in thu laboraitory . A pai r of dttn i
foldetd slit and tttof ded t poft. alict Oinls are depi ct ed iii Figure 5d, wherein the dots
tndivate theirr fertt it- ilsinusf .
5
Figure 2a. Log-periodic array of slots with loop coupling
Figure 2b. Capacitively coupled log-periodic slot antenna
Initial attempts wil~h models incorporating the folded elements were encouraging.
One of the early models is sketched in Figure 3a. Note that the segmented coaxial
feed line is grounded along the central axis of the array, and that its center
conductor is periodically connected t.o the center conductors of the slots. Here
another adiatiiege of the folded slot, elements over ordinary slot elements appear;
it. is possible to attach a whole series of elements with varying resonant fre-
quencies to a single feed line with no danger of shorting out the low frequency
energy in the high frequency portion of t~he antenna.
Radiation patterns of several models like the one of Figure 3a were in general
good -- in fact the heat- so far obtained--except for persistent inexplicable
anomalies. such as were found in the zigzag slot antennas. The models were also
rather insensitive, making patitern measurement difficult suggesting that impedance
problems may trove still rcmained.
3. L.OG-PYIlODIC FOLOSI) 81017 ARRAYS AND) THEIR RElATIONS
1*11 turtitng poitnt in the log-periodic slot array. research came with the testing
ofi an sitictina of f'he type shownt in Figure 3b. Two points of difference from the
array of Figure 3a should be noted: 1. 'the segmented outer condutieor of the feed
lint is itot grounded, 2. the individual segatents have been lengihcned in a fog-.-
puitodic fashion to provide a controlled amount of extra phase shift per period.
nft effects oil ibis exitra phase shift on the radiation patterns were much the same
ais those previously expejirienced with bent log-periodic zigzag antennas. InI any
case, 'ihe anomalies in -'he patterns disappeared and and an average of 20 db more
signal was available wititn taeasurbms,nts were carried out.
As the pat 'erns of Figure 4 show, something is still to be desired in the
opeitaiolt of this type, --> folded slot array. In the actual models, sections of
iniciodult coax providvd the teq tired phasing between slot~s, and inaccuracies in
these cables may have Utei-e part of the problem. Their lengths generally averaged
2-5 times the distance beitwern slotIs.
3.2 _Printt-d Circuit Log-Periodic Folded Slot Arrays
A considerablfe ititlt-avemeni in accuracy and simplification of construction wis
madL possible by adttp'ing cimpletely printed anil~taiia construction. This improve-
m(III wasA also re-fle~etdi in ithe radiationt pa~t-erns, The method of reducing lhiO
7
Figure 3a. An early log-period'ic folded-slot array
Figure 3b. Log-periodic folded slot antenna with phasing lines
HPLANE £ PLANE
I I
525 MC 8' 525 MC 77'
H PLANE £ PLANE
Lt
t00MC 76' Soo $I*
H PLANE E PLANE
1500 MC 586 1500 mc 710
H PLANE C PLANE
1900 MC 67' 1900 me
Figure 4. Rjadiation patterns of SA-2B, r 8,aE =22.e
9
phasing cable sections to a two-dimensional structure is illustrated in Figure 5a.
Antenna currents flowing on the ground plane on the array axis are parallel to a
line bisecting the elements (i.e., the axis of the array). Thus, the ground
plane can be cut along the array axis and sections of printed circuit transmission
line inserted therein. An adjustable phasing mechanism is. provided by the smaller
slots which are symmetrically cut intu the center conductors of the outer slots.
It has been determined experimentally that the optimum length for these phasing
slots is in the neighborhood of one half the length of the outer slots.
Appreciable deviations from the optimum length are accompanied by pattern and
impedance disturbances.
The unique symmetry properties of the printed circuit version of the LP slot
array -ere not recognized immediately since attenLiunI was principally directed
towards flush mounted designs. However, a fortuitous mistake in the specification
of the type of negative to be made from a folded slot drawing drew attention to the
unusual symmetry of the new design. To this writer's knowledge there is no other
single log-periodic antenna design that will operate in all three basic modes:
1. As a slot array, mounted flush with a ground plane and producing vertically
polarized radiation; 2. As a linearly polarized free space antenna; and 3. As a
vertically polarized monopole array ever ground. Figures 5a, b, and e illustrate
how the various forms of the design are related. A family of folded element
antenna arrays which were all made from the same original drawing is shown in
Figure C. Since this report is primarily concerned with flush-mounted antennas,
nothing more will be said of the folded dipole and folded monopole arrays except
to state that they have excellent impedance and pattern characteristics which are
implied by duality and symmetry considerations.A series of patterns measured on the first printed circuit folded slot array,
FSA-1, is shown in Figures 7 and 8. This design had an average directivity of 7.5 db/
isotropic with average beawmidths of 58.80 and 58.50 for the E and H plane respectively.
Kraus's relation was used to calculate the directivity value. When specifying the
low frequency cutoff of a log-periodic antenna, care must be taken to first choose a
reasonable criterion and second, to properly relate this critical wavelength to the
physical size of the antenna structure. In the present work, the frequency at which
the front-to-back ratio reached 10 db was chosen as the low frequency criterion.
This wavelength corresponding to the low frequency was then compared to the largest
10
Se Log-Periodic Folded MonopoieArray
S b Log-Periodic FoldedDipole -Array
xt
j,.,. on,.\
..... .r .
-~~~4 - -- * - --.
- -.- - -- --
5G. .o -Periodj4~ .4t-add.ltAra 5 ua ode lo.nFoledDiol
.iur 5.Lgproi oddeeetar~
~1I-
Figu..6.A l of l
Figre6. fmil o lo-prioicfoledeleen aray
HPLANE E PLANE
600M yo BOO$00MC 83*
H PLANE if PLANE
70,0MC V7' Too MC 0
II- PLANE E PLANE
IO00 uc 660 100011 MC
H PLANE Ef PLANE
1500 mc 1500 56
Figure 7. Rad iation patterns of ISA-1, 'r =.78., a 1 25",
a8 25
13
H PLANE E PLANE
1800 mc 54v 1800 mc 660
H PLANE E PLANE
2000 MC 490 2000 >)C 48
H PLANE E PLANE
2500 e 51 2500mU 8250 250 > 48
HPLANE E PLANE
3000 MC 550 3000 MC 486
Figure 8. Radiation patterns of MS-1 , r .78, CIS 250as=12.50
14
iroinevrst. dimenstion of the structure. IL. would be most desirable it the aintenna
had a 10 db F-l raltio at Itno frequency where the largest slot was )Lo/2 in length.
Fur FSA-I a fi qoeicy of 593 Me. was calculated from the structure size, while
pat tern measurumunts govc- 70-0 Mce. Thus., %.he actual cutoff (based oil the 10 db
F-fl criterion) is Ie4 higher than the structural cutoff. Other types of log-
Pt~rindic antennas exhibit a simtilar behavior.
fto any case, the low frequency cut-off of at given structure is somewhat dependent
on Ithe paramfeters 'r ;iid a . For slot arrays at least, which are most likely to be
usted at microwave I Ieieqoenctes,. the smallI increase in physical size over the X/2
dititousioni is of nit consequence.
Returning again to the patterns oi Figures 7 and 8, it is noted that thu
aviirage irvet ivity calcuilated from the beamwidt~hs Is just. about the same as would9
be (,xiueted f row a LP' dipoleA array with the same 'r and a.. This is a reassuring
-rusulI , Thu Itaticin at itiOO Mc is included biccause it is the worst one that- was
wt-asurud, however, at both 1700 and 1900 Me very clean and unidiructional patterns
we-re uto[Ijiotd.
Aui imuuttlnvt ptid -. 1 HSA- I is given tin Figure 9. A VSWIt of 2.3- 1 with etuautic
lit 93 Qw,. tStt4 u st ittt I .. HSA-I liver the range 600-2000 Mc. It shottid be 1ooted
that imapudatnceausui& o ini this f reqtuenty range are diff icult to do accurately,
itsome tiof t ¶he tivatiluru'd pointsa may be due, to that. cause. Blocks of Eccusurb
pl aced undo rnita tilit gi oiid sccreen served it) absorb any uiy donwa id i-at!iatiOtt I roe
tlithi blaternri v scmiulw itiia'ns
All of Itho prtnmed i ircuit antunnas wore constructud of 1/32 in. teflonl
titpr'idt'toatctl ui-gloac Ionlad, wi lb 2 o4. copper on one side. Drawings were
Itcitiural ly math: t,%V 131 fi hiut- timus li fu-size, then photographically reduced so its
Iii pr I tCite uioga I I %.t', t thlit dc trod st ze thits, gotid accuracy vnit d h: otah i nedt
u-i 11i1iimat-a I IV
lttolot ciu rout itdng fiItus Sot I ion, a fow other folded slot designs sad tiodift -
u-at nistn ltl hit d ,-tu bc hitcliv. Al though not a lot of work has bretn done on I Ito-
nc.w lmodil III! &I i sltr 1ut-, tflow that t he choice of parametecrs is iiot port icularly
Cvii it-al. tild that.t wilt- grid ott:Iititt ot the cetiter conductor is a saltisfact ory
ttt flt ti - lo, *b h i l pii it.--d Vertsion.
h tit Sia It lus eF jigj- tf thi'Woon array axits 'aid ouit)t exi vi till y ofi Slits)' used
'ti'1. It,,'.ý1 Ii,~ It 18 at .1b a ' , of 1t2. All a9ý t 9.5" thuftod the lvngtli of 11th
phdSiit 1: t,, Sti'i, , pit 'ti iaittit I),,,titits of model I-SA-5 oit. shown iii Figuii- M 0
15
NORMA.IZED TO S0 5 2.0
Figure 9. Impedance of FSA-1, T = .8, aE B 25 0 a. 12.5 a
16
HPLANE E PLANE
TOO MC 87' 70OOMC 55'
H PLANE E PLANE
1100M 8I1I' 1100 mc 52'
H PLANE E NPLANE
1500 MC 56' - 150.MG 43'
H PLANE E PLANE
1800 MC 430 1'000 45'
Figr I Ivl. Iltidiatiott patturns of FSA-5, T .82, c 180, a 9.5,
17
For this design t~he actu~al cutoff wns about 17", above the Structural ciutoff of
590 Me. Due to tihe longthi ot the array (because of 1i;G low Er angle.) aod the
limited size of the pat tern aperturo plate (14 in. x 14 in.) the smallest slot
resonated at. --bout 2010 Me. At 1800 Mc effects of the front truncation were
beginning to appeal-, while at 2000 kMc thc pattern began to split into two distinct
lobes, Cross-polarized cncrgy (hort700ntal) was measured at several frequencies
and found to be below -35 db. The overage E and H plane bearrwidt.lis were 53 0and
600 respectively for the 700-1800 'Mc range.
In the printed antennas discussed so far, the center, conductor of each slot
element coveree a rather large portion of the area with copper. This causes no
problem when bilaterally symmetric arrays are considered; btll., if attempts arc
made to eliminate radiation from one side of the ground plante by the use of
cavilty backings, it would seem that. lite extensive surfaces of the slot cent-er
conductors could tinrder lihe exit of energy frum the cavity volumes. For this
reason, the slot, center conductors were simpty outlined with a wire grid ( so as
to reduce t~heir blocking chaiact.ci'isl rcs to a minimum) onl model FSA-39. Its
patternis, shown in Figure 11., afic not tcatbly beutter t.half FSA-3C's whichi used thle
printed circutit type conductor. However, fin every case when a cavity backing
was added, partital pat tern breakup took place. More will be said of this iii tile
next sectiorn. The ave-rage btusmwidt~ts of FSA-3E were 580 and 64l fox, the E and Hl
planes respectiveliy.
One last lot dud stlot anut dma shioul d be mertit tund. Model SA-6t3 (originral
numbf-rmor -ýcheomm) ermployed an aE of 300 -- t(re largest so rar at tempted. The other
signi ficant paaat~i:ters were a. = 15 a ndt rT .71. Becautse of tihe low vuluec of 1r
the artlennia cuesi stud of only six widely spaced e lementsa. Even so, ttne pattern
variatitons werc not, too great arid fth- dlesign carl be nonsidered piacticlc for matny
alpplicat iotns. Undoubiti dty. models wit Iii von hiigiher a. aingles call be bulmt.i, but
their directtvi ty would probably sillfor ladle than lthe addi tionial size reductluji
mould be wvr hr. The paric ulivl model tested lere( hiad average E anti If plante beam-
wildth!. of 66.1o and 36.9o I'LSPectIVUt1y tot fliti 700-2000 Me range.
3.3 CavixtyBacked Slut Array5
Perhaps it is hos ii t it) soe at lite our se that sto far cavity hackedt des igns
have been, for thec motis part, uirsuecessi ii. Two general approaches arc being
pursued In thec pieseiti work 1. Tire simple adudition ut of X/4 deep jindividiual
H PLANE E PLANE
PL
600 MC 63' 600 MC 63*
H PLANE E PLANE
IKI
1100110 72' 110011055
H PLANE E PLANE-
1600 MC 73' 1600110 62'
H -PLANE E PLANE
I I
2000 MC 52' J2000 MC 49'
Figure 11. Radiation patterns of FSA-3E, r = .8, aE= 20o a 103Si~
I
calvities to eacth elemtent of a Successful bilaterally Symmetric Pfrtd.c slta''y
2. The des'go of a single folded Slot clement plus cavity to stecure the desired
resonant fr~niiency and impedance; thent the combination of a number of these
elements to formi a log-periodic array.
Thc. most promising approach so far i~s the first one mentioned--on the basis
of radiation pattern performance. SomE typical examples of pattern de~gradlation
are show~n in Figure 12. Itdidvidual cavitlies., Au/4 (free space) deep were added
to one side of cach sLot element of a reasonahl-y wcll performing LP slot array.
It should be stated that the phasing cables of SA-2A and SA-217) were considerably
different in length: both arrays represent the best phasing Adjustment that was
found. At 600 Mc tihe act ion of cavities is t~o reduce the front-to-back ratio from
8 db to 2.06 db. This effect was prcueni ott all cavit~y-backed designs that have
been tcsted to date. Again at 900 Me thec front-to-hack runti is reduficed front
21.8 db Lo 8.2 db, Although the pat tern of SA-2A is not. Partacnuitirly good at
3300., the damaging elfects of thc cavities can be clearly .3een in the p~attetrn of
SA-2D). Thi~s is repeated at' 1800 Mc. Hoat showni in Fipuro 12 are some of the
bettIer fiat t 0(0 of SA-2l) wtt icti lend eticoutragemrent. to t~his appraacti to unilateral
I lush-usoun ted LP am toiniaa
The second approach cottsistcd of studying a single folded slot. element plus
cavity attd adjusting il.s rusonatit. frequency said impedance level berore incorporation
int~o i log-periodic irrany. 11 was z ocognti.d fiat- a bit at~era I folded slot clement
culn have otto rcsfutnatt ';juotclly Luind impttdanuic level by itseolf', and an altut;Lther
different resonant frcqic~ncy atid imfeiiodetci levet tillor the additLion of a cavity
backing. Front pait cxpci leiwe it was dfecided thai1 -, real impedatcec of 100 R I
Waiuld be sat istfacloty . It was uafso duc idted that the resonant. frequecuty (wher'e tile
imnpedanice was t1t0 02 ca I) of thev slot plus cavilIy Should be the same. as that of
1itie slot afonc. No%,: if thu (evi t b ackintg IS UssuImed loSS leSS it can add only
puare react-etice to Itic sloti impedatice--outi atido of doubtin tg the radialion res istlance.
Thtus, the IF roq~uonty at wJtu::h ftie reacl'ance of the cavity goes toe zero Is-of primary
Initerest, sinlce It weLst. COItIcido Wii ftithe slot. rcsotiatt frequency. A cavity which
is Ao/4 deep ItI tIII( Stot reosuitnce will tiot have zuto reactance there becatuse bf the
wavcg~tidc efiec c. *That ii., 'ict civity backitngs Sfiould be cotisiderted o'a sectiotis of
wavc[guide which atre excited in the dtominani mode hy the slot.* The guide wavelength
in such a sy5stI cal cit: beSLvolat I Inuts the t rcL. pa~c, wavelecntigh, implying thal., for
20
-NO CAVITIES CAVITIES ADDED
60 1C SAA 60 C
900 MC SA-2A 900 MC ISA-ZO)
NO CAVITIES CAVITIES ADDED
9300 MC SA-ZA 1300 MC SA-ZD
NO CAVITIES CAVITIES ADDED
1800 MC SA-Z2A 1800 MC SA-20
Figure 12. Effect of the addition of cavities
21
* rosonlanc',-', the civil ies must be Xgf.1 deep instead of Xo/4. Bly using a cavity
of adjLIstable depth, t~he cc~rrect length wds easily determi ned.
The usco of a folded slot eleaent. rather than a simple slot. allows the
rcsonnnt impedance to be widely varied without appreciably affecting the resonant
I requcecy. flaw this is achieved is illustrated in Figure 13a. By smaking the
V IotI Ic cu Q~loUt1 cUtId14tUtS of un1iequal tz in the fuldud dipole oi- the magnietic
current ConducIItors ini the folded slot. u wide range of input fitprlndnccs tire avail-
able. In the actual alim, plus cavitiy tested the center conductor of the folded
slot radiator was of the wire grid type already described, and was considerably
displacce from the center of the slot. After very few adjustments the desired
Limpedance level and resonant frequency of the slot-cavity combibnationi were
scoured. It was al!;u gratifying that. the dcpth of thc cavity necessary for
resonance turned M11 In ho almost exactly what. was predictcd theoretically
(abouit .45 Ao for the size waveguide chosen).
Uisingp siml ar proport ions a whole series of slot-cavity combi nations were
then inlcorporntLdcito1IL a single lug-periodic array. The parameters usedi were:
T -7, a. 250, ard a. 1P.5o However, the resulting radial ion pa ttei-ts and-
Impedu lce wuts ve2ry pour,. lIn fact the antenna tended mtore Of teitn ho li! ndf it s
then bach! i c. And., although i-he mean impedance of the array was 55 2 as deosired,
I lie accompanyting VSWII iatitgod frtomi 1:1 to 8 :1. Because of these discoutraging results
alid beCa1S iseOf I Ito large a sue o 1 of work required ;ii the cons trute .ion of a set a s
01t J~g-periudic CaVI iiUa, Ifurthler work has ceased for the time being.
Dcot il ulclcIud ing, a Su~reLwhltn. dif fercnt li nt- of alt.tack oil the problem of
coiiveri-ing ii bilaterally radialiing LP' slot array into a unilateral de-vice will lie
ment I ited. 'PTe idea basi cally resul ts uopn the fact that the far field patterns o)r- a ~~sm-allI louD aind a slot ML k/2) tire identical. If,* 1tben a tool) antd ilocod
be astilt Ia lit-ott sIy excifted wivthl Ihu ii:properI phas anIItd mog~itu ide re la.tions, i t would
tie puss lilctoli calitru I lie fotr fc iel of the sloL everywhere in the chosoen liao-space.
IPgure 131) shows an a liangelnutlo ithat gives partial cancellation iii the hal t-space
cuitteitintg 'lfie t ool). An allay or Ithree sheet miet",I lool)! wiith sean ci rcumferences
Of.65 A Li Slot retticglivi! Ito (Iit average cancellainott over a, frequtlicy range of
.100-1I2001 Mri C (>lt i gor:t ion'is withI fewer loops, Or' withi lo(ops of tI i aWire have, so
tat-, jhowtIc~ canelc atiton. *'n f0o'tI~tieteY the first at tempt at combtinin ti anumltber
of such utt tlftiat slots Iinto a ltig-Itortod~c array failod.
22
........*K:,"-ii
LOW IMPEDANCE FOLDED SLOT LOlW IMFTWCE FOLDED DIPOLE
Figure 13a. Impedance re~duction with fulded elements
F'igure 13b. Unllatcrai rf-r-field c--ncel lation of s'tot by loop
23
4. CONCLUSIONS AND FUTURE PLAN8
Al though the ultimate goal has not. yet beeni achieved, certainly the present
work is a subsiawiial step along the road to a satisfactory flush-mounted IP
antenna. Ildend, if onv werv willitig to accept P 2 1c. ii , el;iciency, a
successful LP slot array can already be built by providing one of the folded sal5
designs wit• l at, absorbing cavity. In addition, a valuable new family of log-
periodic designs has been discovered as a result of this program.
Fhltiure plants includu a contititing study of the folded slot arrays and
associated work on cavity backilgs. Whether the piece-meal appronch of dot-
stgning a singic coil, or modificaition of an operating bilateral antenna by
means of resofant cavities or loops will lead to success, only further work can
lull!.
24
IFFEIRENCES
1. R. Mittra and M. Wahi, "hr" L&tter-llack Antenna - A Wide-Band Flush MoUnied
Antenna of Log-Periodic Design", Convention Record, First IEEE Internationai
Convention, New York, 1963, R.iso University of illinois Techniical ReportNo. 71, Contract AF33(657)-W1,74, Jane 1963.
2. J. W. Suborner & D. E. Isbel l. "The Development of the Slot Complemernt of tlheLP Dipole Array", Proceedings of lhe 10th UJ. S. Air Force Antenna Sympnijj,
Allerton Park., October L960, University of Illinois, Urbana, Illinois.
3. V. A. Mikutias and P. E. Mayes. "Log-Periodic Cavity-Backed Slot Antt•inu",
Proceedings of the 13th U. S. Air Force Antenna Sympositur.. Allerton Park,
October 1963, IUniversi ty of Illinois, Urbana, illinois.
4, P.E. Mayuc;, and It. L. Carrel, 'Hiigh Gain tng-Periodic Anteanas", ibid.
5. Interim Enginvering Report No. 1, C0otract( NO•SR 85243, Antenna Laboratory,University of Illinois, U.rbana Illinois, March 30, t962.
6. A. F. Wickoetsham, J r., "Rtcent. I-vetop•tents in Ve, ry Broadband End-FtreArrays", (corrosptindence), Proc. II|F. V-.!, 48, April '960, pp. 794-795.
7. .1. W. Greisvr, "The tflut Log-Puer:'dtc Zigzd;, Anttflnn". SUpp)reient to
it erim Fnug I oer tnk Report go. .1, NORSI1 85243. May 31, 1962, mniversil-y oft1Iirmi s, Artunnu lhouraltory, ir;:.ýna, 11i nulmas. A[sot, J. W. Gre-iscr and
P. Et. Mayes, "Thu Bent Backrire Zigzag-.-A Vertically Polartized Frequency-
Indep.endent AntuLna", to be pub!isiud in ,he May 1964 IEEE, P'IGAP
Transact i tin.
8. J. I). Kraus, AntL-inas, MNe;j dS- ti, NIew ytirk, 195u, p. 26.
1). It. L. Carrel, "Analysis sod }esgign of oh&JLog-Periodic Dipole Antenna",
Tech. RVeVUVrl No. 52, Ctintrect No. AM'33(616)-6079, University 61' llitnoisAttinne IAlatratory, Urbana, Illitimes.
lUNCLASSI FIED
UNCLASSIFIED