1923 - Wave Antenna - Beverage

given

given by by Mr.Mr.

P.

P. S

S..

CarteCarter and Mr. R.r and Mr. R. D.D. Greenman.Greenman. T THEHE

 N

 N

EEWW r y I c l x c v e r y I c l x c v e m m .IcII.IcII.c- .c--w -w fwwr

fwwr l l R cl l R cII-II- memetam.-tam.- a 4 ~a 4 ~

fwwr

fwwr *aLOo.*aLOo. ,z- ,z- 4-4- 0.-

0.-,vcv

BEVERAGE,

BEVERAGE, RICERICE _{AN ANDD TransactionsTransactions E.E.}

shows

shows aa photograph  photograph oof f papart rt oof f ththe e lineline.. AA flexibleflexible line construction

line construction waswas adopted whichadopted which o

off loopsloops or orseriesseriesoof f verticals connected throughverticals connected through aa transmission

transmission line, line, etc. etc. Two Two crosscross armsarms were provided,were provided, one

one atat

aa

height ofheight of 1188 feet afeet and thnd th e othere other 3030 feetfeet wires

wires and the lowerand the lower armarm lineslines were

were bro broken ken every every ten ten polpoles, es, and and dowdown n leads leads werweree

OF

OF RRIVERHEADIVERHEAD AANTENNANTENNA.. FFOUROUR

 A  A ==

 pro

 providvideded that connections could be readily changed.that connections could be readily changed. The line

The line ran ran seven miseven miles approximately southwesles approximately southwestt Riverhead along an unfrequented sand road, and Riverhead along an unfrequented sand road, and waswas later extended to Terrell River,

later extended to Terrell River, aa total lengthtotallength o

off nine nine milemiles. s. The The lineline iiss aassstraightstraight aassititwas feasiblewas feasible

to

to

 bu builild id it, and itt, and its direcs direction tion is substantially in lis substantially in line ine withwith the principal European long wave stations, the recep the principal European long wave stations, the recep

--tio

tion on of which wasf which wasaamatmatter oter of primary interest.f primary interest. The

Thefirstfirst

tests

tests

ononththe nee new line sw line showhowed thed thaat tt the hoped.he hoped.

ESTS

ESTS OO FF RRIVERHEADIVERHEAD AANTENNANTENNA.. _TTWWOO

M

MULTIPLEULTIPLE..  A A == 79007900 for

for results hadresults had realized. Instead orealized. Instead of f an an optimumoptimum o

off sixsix or seven kilometers, theor seven kilometers, the

stronger and stronger as the receiving set was moved stronger and stronger as the receiving set was moved towar

toward the d the southwest, southwest, and and the signathe signal strength l strength there wasthere was several times greater than had been obtainable with several times greater than had been obtainable with

rubber

rubber covcovereered ad antennas. ntennas. WhiWhile le ththe e strastray y ratiosratios observed on the rubber covered antennas had seemed observed on the rubber covered antennas had seemed excellent, the

excellent, the antenna fully met our expectationsantenna fully met our expectations Oscillator

Oscillator on the new antenna showed that theon the new antenna showed that the

was

was high and attenuationhigh and attenuation lowlow compared withcompared with

the

the

rubber coveredrubber covered wires,wires, andand thatthat itit hadhad nono seriousserious reflection points.

reflection points. Figs.Figs.

obtained with the new antenna.

obtained with the new antenna. AA comparison withcomparison with

similar

similarcurves, takencurves,taken ononthe the rubber coverrubber covereded  brings

 brings out thout the ie impromproved eleved electrictricalcal properties.properties. The failure

The failure

at

at

thethe startstart toto get good short circuitget good short circuit reflections when the

reflections when the farfar end oend of tf thhe linee line waswas groundedgrounded caused

caused e hige highh--frequency frequency resistanresistancece o

of tf the groundshe grounds waswasmuchmuch an wean wehadhadestimated.estimated. These grounds we

These grounds were made with lines of re made with lines of ironiron

wire

wire

laid inlaid in water.

water. ..The substitutionThe substitution ooff coppercopper removed thisremoved this

T

TESTSESTS OONN

= =

diffi

difficultculty. y. The The adjuadjustmestment nt oof resistance f resistance for for unidirecunidirec

--tional

tional effeeffectscts waswas now clean cut and in accord with thenow clean cut and in accord with the

A

Adeadead groud ground and a t thet the N. N. E.E. end gaveend gave aass bad bad

aa

str

stray ratio ay ratio on the average on the average as an open circuitas an open circuit..

T

THEHEAACTIONCTION OOFF TTHEHE

 W 

 W 

 A  A  VE VEAANNTETENNNNAA

.. The The wavwave antee antenna nna in in itits simplest form consists os simplest form consists off aa

hori

horizontzontal wire oal wire of tf thhe ordee order or off aawave length long pointwave length long point

--ing towa

ing towards the rds the transmitting station,transmitting station, asas pictured pictured 10.

10. Whe

When n the signal wavthe signal wave reaches the e reaches the endend “A”“A”an e. m.an e. m. f.f. is induced in the horizontal antenna wire, due

is induced in the horizontal antenna wire, due

to

to

thethe fact that the wave’front

fact that the wave’front iiss not perpendicularnot perpendicular toto thethe

ground, but has a tilt

ground, but has a tilt ooff 11 deg. todeg. to 1100 deg.,deg., depending

depending ground.ground.

Thus,

Thus,atatthe the endend aalittlittle wavele wavestartsstartsto run down theto run down the antenna

antenna towards the rtowards the receceiveiving station, and ing station, and iif f iit travelst travels with the same velocity

with the same velocity asasthe radio wave in space, thethe radio wave in space, the space wave f

space wave follollows ows righright t along with italong with it, , supplyinsupplying energ energygy to it

to it aassitit goes and building it up, untilgoes and building it up, until aatt the the endend ii tt has

hasreacreached a magnitude manhed a magnitude many times ty times thhat at whiwhich it hadch it had

at

at

 A A .. This is illustrated in Fig.This is illustrated in Fig. 1111 which showswhich shows aa singl

single wave e wave aa t t succesuccessive ssive time intervals.time intervals. IIff the the velvelococ

--ity o

ity of f the wave the wave on the wiron the wire ie is s not not equal to equal to the the spacespace

31s

31s KELLOGGKELLOGG A;A; I.I.E.E.

type

type ooff would-

would-m k e

m k e

it

it

 pos possibsible le tto tro try numerous y numerous experiments experiments suchsuch aa series

series

aboye aboye ground.

ground. TheupperTheupperarmarm car carrieriedd twotwo No No..1100BB

&S

&S

coppercopper four similar wires.

four similar wires. AllAll

FIO. FIO. 77

--

OSCILLATION TOSCILLATION T_{WWIRESIRES ININ}ESTSESTS_{MMULTIPLEULTIPLE.. 94009400}

so so &om &om making making FIG. FIG. 88

--

OSCILLATION TOSCILLATION T_{UUPPERPPER WWIRESIRES}

been been lengi’l

lengi’l signalssignals bec becameame

the the new new of ofimprovement.improvement. tests tests velociw velociw 7,8

7,8 andand 99showshow

typical

typical

curvescurves

groundwires, groundwires,

us

us

to

to

suspectsuspectthatthatthth

g&ater g&aterthth

Wires Wires

FIQ.

FIQ.99

_-

_-

OSCILLATOROSCILLATOR W

WIREIRE,, RIVEBEEADARIVEBEEAD

ANTENNANTENNA..

 A

 A EENDND OOPENPEN.. EENDND

SINQLE

SINQLEUUPPERPPER

G

GROUNDEDROUNDEDTTHROUGHHROUGH

7500. 7500. 600 600 OOHH M  M SS.. theorS;. theorS;. EXFUNATION EXFUNATIONOOFF ia iaFig.Fig. F FIGIG. . ~~ ~~

--

SSI M P L EI M P L EFFORNORNWWAVEAVE AANTENNANTENNA

fnnrsr.ard fnnrsr.ard

or,

or,t b ~t b ~wavewave I-nrr(hI-nrr(h :kmact::kmact: -4,

Feb.

Feb. 19231923 THETHE WAVEWAVE ANTENNAANTENNA 219219

strung on

strung on aa pair of  pair of small small steel wsteel wiriresesaa hahalf inch apalf inch apart inrt in the

the cascasee ooff the the lolowewer line and an r line and an inch andinch and aa half aparthalf apart in the upper line.

in the upper line. The rockers oThe rockers of f ththe upper line aree upper line are loaded with lead.

loaded with lead. The inertThe inertia oia of f ththe rocking e rocking stick isstick is analogous

analogous

to

to

line inductance, while the elasticityline inductance, while the elasticityooff thethe connection between successive

connection between successivestickssticksisisanalogousto theanalogousto the capacit

capacity betweey between n ththe condue conductors octors of f an an electrical electrical linline.e. The velocity can be changed by varying the tension on The velocity can be changed by varying the tension on th

the e wiwireres. s. TheThe at the ends practically stopat the ends practically stop reflections.

reflections. The upper line providesThe upper line provides aa means of immeans of im

-- partin

 parting g energy progreenergy progressivssivelyely

to

to

the different portionsthe different portions wave

wave (velocity o(velocity of f light) light) interfeinterference effectrence effects devs develop,elop, the

the wave wave on thon the wire buie wire building up forlding up for aacertain distancecertain distance The velocities on The velocities on actual antennas, however, are nearly enough equal actual antennas, however, are nearly enough equal

to

to

that

that of lightof light soso that for considerable lengths the wavethat for considerable lengths the wave on the

on the wiwire builds re builds upup aass iit wout would onld onaa light vlight velocity line.elocity line. A

A signsignal coal cominming frg from the om the oppositopposite directioe direction to n to thth atat which we have been considering, will build up

which we have been considering, will build up aa wavewave on the wire in a similar way, from

on the wire in a similar way, from aa small valuesmall value atat

U

UPP OOFF WWIIRREE AASS

ttoo aa large valuelarge value

aatt

A.A.

If

If

now the line were open ornow the line were open or grounded

grounded

aatt

A,A, the wave would be reflected back overthe wave would be reflected back over the antenna to the receiver end

the antenna to the receiver end BB,,and would be heard.and would be heard. O

Onn other hand,other hand, ifif we damp the endwe damp the end AA in suchin such aa manner

manner aass to prevent reflections, the antenna becomesto prevent reflections, the antenna becomes unidirectional.

unidirectional. AA nonnon--inductive resistance, havinginductive resistance, having the value

the value RR == ohms, whereohms, where LL andand are theare the

inductance

inductance and capacity oand capacity of f ththe antenna per une antenna per unit length,it length, constitutes

constitutesaa practically perfect damper. practically perfect damper.

F

FIIGG.. MMODELODEL OFOF W A V EW A V E

Many mechanical analogies will occur to the reader, Many mechanical analogies will occur to the reader,

a

ass ththe building e building up up oof f water waves in the directionwater waves in the direction the wind.

the wind. interesting interesting experiment experiment isis toto run overrun over hin ice.

hin ice. IIff youyou runrunaa t just tt just the speehe speed od of f wavewave ion o

ion of f ththe ice e ice surfasurfacece youyou can build upcan build up aa larger wave.larger wave. ff you run too slow or too fast little effectyou run too slow or too fast little effect iiss produc produced.ed.

e manner o

e manner of f buildibuilding ng upup ooff waves on awaves on a rave antenna, Rice and Kellogg built the mechanical rave antenna, Rice and Kellogg built the mechanical

shown in

shown in Figs.Figs. 1212 toto 17.17. The upper lineThe upper line e space

e space wave wave and and the lowethe lower line r line ththe wave e wave on on ththee Each line consists of

Each line consists of series oseries of wooden sticf wooden sticksks

F

FIIGG.. MMODELODEL OFOF W W AA VV EE AA_NTENNANTENNA

E ENDND

o

of tf the lower line,he lower line, aassaa space wave imparts energy to thespace wave imparts energy to the antenna line.

antenna line. AA slightslight the two linesthe two lines iiss supplied by light rubbersupplied by light rubber

hooks on

hooks on ththe uppe upper sides er sides oof tf the lower he lower stistickscks toto correscorres

-- ponding hooks

 ponding hooks onon the under sidethe under side ooff the upper sticks.the upper sticks. Waves are imparted to the upper line by moving one Waves are imparted to the upper line by moving one end by

end by hand for an hand for an impimpulsulse or bye or by aa motor driven rockermotor driven rocker for

for continuous wavecontinuous waves.s. Fig.

Fig. 1212isisaastatstatioionanary ry vieview ow of tf the machine.he machine. Fig.

Fig. 1133showsshowsaa series series oof f views views taktaken en withwith

aa

movingmoving  picture

 picture whwh impimpulse ulse is is impartedimparted thethe

dashpots dashpots -and

-and then decreasing in amplitude.then decreasing in amplitude.

FIQ.

FIQ.INT BUILDINGINT BUILDING WAVEWAVEONON SPACESPACE WAVE

WAVEPPRROOGGRREESSSSEESS

12-MECHANICAL

12-MECHANICAL ANTENNA- ANTENNA-S STATIONARYTATIONARY V  V IIEEww nch nch An An propaga-?o ?odemonstrate thdemonstrate th iodel

iodel repre-

repre-mts mtsthth

13-MECHANICAL

13-MECHANICAL WITHWITH

DAMPED DAMPED

coupling between coupling between

banus

BEVERAQE, RICE AND I. E. E. The growth of th e wave on the lower line line (antenna) as the waves progress from left to right.

The wave Tha t the upper linecarries _{practically pure}

e waves has no return waves) is shown by the nearly uniform amplitude throughout its length.

with 13 shows the effectof

re

-

Fig. 16 shows the effect of removing the dashpot at

the dashpot, a free end re- the right hand end of th e lower line. Considerable We notice these a return wave movementatthe extreme left end now appears, due to which on reaching the near end, causesa movement of

MODEL O F W A V E

P

WAVE ND

the

tional properties if the end isnot damped.

Time exposures, with continuous waves of constant amplitude supplied to the upper (space wave)

of

the lower line, or equivalent of current distribution

Fie.

15showsthe increase in amplitude on the lower

MODEL O F

DOWN

the reflected wave. The forward wave (left to right)  built up on the line,isvery small near the left end, so we find the amplitude there nearly uniform, as would be expected with waves traveling in one direction only (right to left). On the other hand toward the right we see very clear standing wave effects, for here the

220 KELLOCfC3 Transactiom A.

upper h e .

approachesthefarend isreadily seen. traveling

on he lower line appears Passoffthe end leaving th (i. e.

line practically stationary. fig. 14, c0mpi.d

moving thus permitting

flectjon. in pictures

F I G . AKTESNA-FULL

VELOCITY,BUILDING U

FIG.MECHANICAL M O D E LOF ASTESSA-FULL

VELOCITY,EFLECTION E

~ ~ - M E C A A N I C A L ~ O D E LW A V EA N T E N N AWITH

FREE-E K D REFLECTIOK

first rocker. This illustrates the loss of

unidirec-line, k n g out the amplitudes developed on various parts

the in wave antenna.

FIG. 17-MECHANICAL W A V E ANTEKNh-SLOWED

for-Feb. 1923 THE WAVE ANTENNA

ward (left to right) wave has an amplitude more nearly equal to th at of the return wave.

In Fig. 17 the right hand end is again damped, and the lower

tension on the wires. This illustrates the building up and down of the waves on the antenna when its velocity is much below th at of the wave in space.

R EDUCTION TO PRACTICAL FORM

After the construction of the Riverhead antenna, considerable time was devoted to experiments of various with the new antenna, tests and com - parisons of different antenna arrangements, an

graphic study of static, and tests of station apparatus.

FIG. LOCATING RECEIVING SET AT

URGE RESISTANCE

S . Carter, R. D. A number

Circuit. In order that at the same end of the facilitating adjust-ment, the arrangement shown Fig. 18was proposed  by Kellogg. The two wires work in multiple as an antenna, but act as a balanced transmission line to  bring the signal currents back from the southwest end. This scheme obviates the use of extra wires for thereturn transmission line and th us avoids the problem of pre-venting detrimental effects of nearby conductors.

FIG. ALANCING CIRCUIT EMPLOYED FOR OBTAINING

RECEPTION FROM TEE BACK END

When the primary of the "reflection transformer " was opened the receiving set was quiet, showing that the

line although it was not transposed, not any undesirable electromotive forces. Thusa horizontal loop receives neither static or signal. Compensation for

Back

Wave. Beverage showed that while the resistance could be adjusted to give a mini-mum for static while listening to European signals, still  better stray ratios were obtained by combining with the signal currents brought in over the transmission line,

asmall amount of the currents flowing to groundatthe northeast end of the antenna. . _{This he accomplished}

with a phase adjuster and intensity coupler of the used by Mr. Alexanderson arrage receiver.

The two long wave stations New Brunswick (13,600

meters) and Annapolis (16,900 meters) were of

assistance in tests and adjustments.

station could t o

intensity adjustments of It was found when the adjustments were made for putting New Brunswick out, t he stray ratio was best on the European stations whose wave lengths were near that of

Brunswick, and that when Annapolis was put out the adjustments were such as to

ratio for the longer wave European stations. In other words, the best stray ratio was obtained when the end conditions were adjusted to put out the image of the European station. If a light-velocity wave antenna is an exact number of half wave lengths long,

it

 provided the true surge impedance is connected between antenna and groundat the end nearest the transmitting station. This is illustrated in Fig. 29. Whether the ,

true

or inductive component will depend upon the

FIG. O F BALANCING BACK END CURRENTS BY

FROM DAMPED END

characteristics of antenna and ground at the frequency under consideration. Kellogg pointed out that since the antenna was not an exact number of half wave lengths long, the back wave effect would prevent the antenna from being unidirectional even though the true surge impedance had been used, and itwas for this reason th at Beverage found that balances were required to obtain th e best stray ratio. Another method pro - posed by Kellogg of supplying this necessary compensa -tion was to insert a circuit consisting of inductance, resistance, and capacity in series in the neutral at the end e transmitting station, as shown in Fig. 20. By adjusting the resistance and varying the capacity through th e point where it tuned out the reactance of coil, a wave of any desired intensity and phase could be reflected down the antenna t o exactly compen

-sate

for the back wave effect and thus render theantenna unidirectional.

If only one station is to be received the circuit shown inFig. 20 is as satisfactory

found. For reception of

output transformer was used having a step-down ratio of

circuit of about 3/4

221

line has been slowed down by reducing the

kinds oscillo-IS-ARR'ANGEMENTFOR SAME END ASS - in FIRST B ZERO transmissiog inkoducing type inthe b

circuit arrangement is shown in Fig. 19. The

great

making Either

be entirely pu ut by using the phaseand

Fig. 19. that

New givethe best possible stray

desired

themathematical analysis shows that isunidirectional

m g e impedance is non-inductive or contains a

capacity ANT'MY n o e s 2GMETHOD REFLECTIONS nea,restth the

any that the writers have

longwave stations an antenna

200

to

10turns, and hamigacomplete iron magnetic

BEVERAGE, RICE AND TransactionsA. I. E. 0.0015enamelled sheet iron of the developed for

the alternators. The secondary was in series with the tuned circuit of the receiving set.

FIO.

 Mu lt ip le x Reception. The simultaneous reception of a number of stationswas one of the next objects of our work. If the surgeimpedanceis set at the best value

foramean wave length, and no finer adjustment of the

constitute so small a load on the tha t the  potential distribution on the line. The

dampedasit isatthefarend, actsasa

resistance in the antenna surge impedanc By adjustment of any one doe

I

. FIO. ANTENNA CIRCUITS.

 back -end compensation is attempted, then it is only adjustment of the series and shunt resistance boxes the necessary to provide more secondaries for the antenna antenna damping resistance may be set at the best

output transformer, Fig. 20. In order that one set average value, leaving only a small residual to be should not sap too much energy from another, trans- “cleaned

were designed with very slight reaction be- Fig. 22 is a diagram showing some tween the secondaries. Some data were taken of the changes which have been made since the original in - best resistance and reactance in the ground circuit, asa stallation.

networks were figured out Shielded Sets. The aperiodic nature of the wave which give the desired impedances atthe wave antenna and the success of the coupling tube

lengths of the stations which itwas most important to system made it clear that there would be call for receive.

This

system of multiplexing, however, did not operatinga number of receiving setsin the same build  -appeal

to

the writersas the most satisfactory solution ing. The artificial line and antenna output transformer of the problem. A system was wanted in which all had been designed for operating four sets, but this was the adjustments for each station

to

 be received could be not necessarily the limit. In fact, later, when the made without reacting on the adjustments fortheothers. Riverhead station out a total capacity of

222 KBLLOGG

kind Alexanderson

connected

21-MULTIPLEX RECEIVINOSYSTEM

The 21 was worked out, and

 proved entirely satisfactory. Each coupling tube

areceiving set, and the antenna output transformer and artificial line, with its sliding contacts and potentio-meters, serve to impress the desired potentials on the grids of the tubes. Since the load is negligible there is no reaction between the different sets. Grounded shields between the secondaries of the antenna output transformer prevent electrostatic reactions. The de -sired component of the currents or potentials in the ground circuit can be obtained in any desired phase by moving the sliding contact along the artificial line, and in the needed intensity by adjusting the potentiometer.

The artificial line has a characteristic impedance of

about 400-ohms and reflections are prevented by a resistance of about this value. This results in a phase adjustment which gives practically constant intensity. Five thousand ohms potentiometers are these

mngementshown in Fig.

feeds uscd,and artificialjliline aniL,,.C2;’line, pr9rtically pure s not wit. appreciabh ‘ ~ b bthe

2-wAVE PREfiENT ATRANGEMENT

up” by the artificial line adjustments.

formas more.detailed

functionof wavelengtl, and

would multiplex

new

Feb. 1923 THE WAVE ANTENNA .

receiving setswas planned, six of which are now in daily

operation. The amplifiers, detectors, and tuners  previously used in the receiving sets of the Radio Cor 

- poration were unshielded, and the practise had been,

where two setswere in use in t he same station, t o keep

them well separated and operate from separate

 bat-teries. Tomeet the new situation,anew line of appara

-tus was developed. Each piece of appara-tus was in a metal lined box, the metal lining being grounded, and connections between the boxes were made through shielded cable. All tuned inductances consisted of rs of coils, of compact form, thus reducing chances of magnetic coupling. The radio amplifier was

shielded between stages as well as externally. The

amplifiers and detector have individual plate and fila

-ment filters in the supply lines. A two stage filter was

introduced in the circuit between the detector and the audio amplifier in order to prevent radio frequency currents and potentials from getting into the audio

circuits where the y might cause back coupling or inter 

-ference between sets. Low resistance telephones were used to minimize electrostatic coupling. These

FIG. STATION, RIVERHEAD, LONG ISLAND

cautions made it  possible to operate th e several sets in close proximity, and from th e same plate and filament  batteries, without an y interference between sets, and to employ high radio and audio amplification without trouble from back coupling.

It was usual in receiving the high power European

stations t o develop ahigh-frequency potential of about

7voltsatth e plate of the last tube in the radio amplifier,

with a useful current of ab out 0.2 milliamperes. By

connecting the outp ut of t he audio amplifier to a good telephone line, satisfactory tone signals were received

and copied in New Ydrk City. Beverage arranged a

rectifier for the audio frequency currents and operated

a

telegraph sounder, obtaining very

in this way when static was moderate or light. During

the springof 1921, considerable commercialtrafficwas

received directly in New York in this manner, using th e  private telegraph wire which connected th e Riverhead

experimental station with the Broad Office of the

Radio Corporation. One demonstration which aroused considerable interest consisted in putting

signal on th e telegraph line a t Riverhead, and

matically repeating at  New York into th e

control line, so that the operatars

British station heard their signal coming back

 New .

When the success of the wave antenna had

demonstrated at Riverhead similar antennas

constructed at

stations. These new antennas gave satisfactory per 

-FIG. RECEIVING STATION

FIG. RECEIVING SETS, RIVERHEAD. LONG

formance in commercial service until the long wave

traffic was finally concentrated in the new Riverhead

station which the Radio Corporation constructed

during the summer of 1921. Fig. 23shows th e present

Riverhead Station. Fig. 24 shows an output trans

-former, artificial line and potentiometers, Fig. 25 shows

shelves with two receiving sets. The building does not

 provide space for operators, since all signals are trans

-ferred to telephone lines and copied directly by operators

astatic pai pre-23-RECEIVING satisfaetory signals St. Carnarvon’s auto-New Brunswick in the own on Brunswick’swave. been were Chatham, Mass.and Belmar, N.J.where

the Radio Corporation was already operating receiving

%ANTENNA P A N ~ L ,RIVERHEAD

BEVERAGE,RICE AND . E. E. 232

 purpose of the directive curve is

to

show the relative 32

to

38 bring out the

effects

of length, velocity

currents for different directions, the and on the

of

a wave

antenna. The value0.06 per a,and 0.8

factor

omitted (or in other words assumed to haveavalue of 1) and the relative currents shown in column XI are

\ . _W AVE H A LF = OF WAVE ANTENNA. FIQ. OF WAVE ANTENNA LENGTHS =24

CURRENT O F IDEAL WAVE ANTENNA, ONE

AVE ENGTH LONG. n = 1, 1 = _12,  A = 12 served for long waves (7000to 25,000 meters) on bare

0.102 inch (0.259cm.) diameter copper wire supported

on poles. Fig. 36 shows the directive curve for an antenna one wave length long on theassumptionof zero attenuation and light velocity. By analogy with transmission line practise, we have referred to a wave the largest one of those, which corresponds to

= 0. The corresponding directive curve is shown in

Fig. 35.

K E L L O W TrsnsttotionsA.I.

Figs.

receiver signal attenuation directive properties

-

2 2 which is the same for all directions, is for 36,are mean values

kilometerfor

u,used in calculating Fig.

ob-FIG. 3'J-DIRECTIVE CURVE OF ANTJCNNA

WAVELENGTH LONQ. = 12KM., 6 M., 0.05,u 0 . 8 ~

FIQ. &DIRECTI~ECURVE

12KM., a 0.05, 0.8

~S-DIRECTIVECURVE

LONG. = 12KM.; 1 KM., =0.05,T w o=W_0.8AVE

FIQ.%--DIRECTIVE

W L =0,

tained by dividing all the calculated currents of column  by

Peb. 1923 THE WAVE ANTENNA

signaldirection deg. Both I, asgiven

(22) and (23)have negative signs, and the

negative signs are retained throughout rather than

reversing the vectors in t he .diagrams, which would

obscure the angular relations.

If

th e entire

diagram is turned 180deg. (lookedatup side down) the

vectors appear in their t rue positions.

Short. Antennas. Th e question of t he possibility of obtaining in less space directive properties approaching those of a full wave length antenna is of considerable

0.5, 0 . 8

interest. Fig. 47shows th e directive curve of a quarter

wave antenna without compensation. It is obvious

that so long as there is considerable inequality in the

size of the two lobes of a directive curve

such as Fig. 47, a back end zero can be obtained by

compensation without entirely sacrificing the signal,

or front end reception, but how favorable a total

directive curve would result is not apparent until

detailed calculations are made.

Among the first experiments tried on the Riverhead

antenna was loading to give various propagation

velocities. Itwas pointed o utatth at time by Rice that

reducing the velocity on a full wave length antenna to

something less than the velocity of light would result ina sharper directive curve, but tests showed that the

natural velocity of the line was so nearlp equal

to

the

 best velocity th at there was little t o be gained by load 

-ing. The urgent problem a tthe time was to obtain th e

best possible reception of European signals, utilizing

as much space as seemed conducive to this result.

Testshad shown the half wave antenna, with back end

compensation, to be definitely inferior to th e full wave

length antenna. The question of short antennas was therefore not investigated until some months later when the principal engineering problems connected with

multiplex reception with the wave antenna had been

worked out.

At

the

suggestionof H. Ranger, of the Radio

Corporation, Kellogg of the

directive properties of short antennas on which the velocity had been reduced by loading

for the wave length and antenna length in question. Mr. Ranger reasoned that since reducing the velocity

sharpened the directive curve of a full wave length

antenna, it might be possible, by sufficiently reducing

the velocity, to compensate for reduced length and

 perhaps obtain a good directive curve in very small

space. The most favorable velocity was found

to

 be

that which gave zero (or minimum) reception at the

 back end, without compensation. The condition for this is tha t

=

in which is the length of the antenna

v is th e velocityof light u is t he antenna velocity

Figs. 48 and 49show directive curves for quarter and

eighth wave length antennas with velocities equal to one third and one seventh of t ha t of light respectively in accordance with the above equation.

FIG. C U R V E OF SLOWED DOWN QUARTERWAVE

NTENNA n =_0.333, 1 =_3, = 12

By way of comparison, th e directive curves Figs. 50

and 51have oeen calculated for compensated

without loading. The back end areas of these curves are seen to be slightly greater.

Multiplex reception with different compensation for

each wave length is possible with the unloaded, com

- pensated antennas. On the other hand, with the slowed

down antenna, the velocity is right only for one wave

length. In order to receive other wave lengths,

com- pensation would be employed in addition t o the loading. 239 = 120 _and ljy formulas somewhat P” FIQ.4 7 4 =3,_{A . =} 12,a figure-eight Mr. R.

calculations were made by

to the bestvalue

VIU X / l

-isthe wave length

48--DIRECTIVE

LENQTEA =0 . 5 ,

BEVERAGE, RICE AND KELLOGG I. E.E. 62 shows the directive

an

length of the slowed down antenna, whose

for is shown in

The dotted line A th e directive properties without compensation and the solid line is the directivecurve with compensation. Fig. 53shows th e directive curve for the same antenna receiving an eight

Compared with the simple unloaded,

antenna,

th e slowed down antenna has

a

 better directive curve at one wave length, but a narrower range of satisfactory multiplex reception. The signal strength

CURVE OF SLOWED DOWN

FIG. - A UNCOMPENSATED. B COMPENSATED. 1

=0.333

EM.,

is of t he same order of magnitude on both

-

considerably

LENGTH ANTENNA, OMPENSATED.

240 Trmssctione A.

fig. curve for 18kilometer

wave

directivecurve = 12 kilometers Fig.

48.. Fhows

compt3nsated

FIQ.&-DIRECTIVE EIQHTH W  A  V E

LE NQTHANTENNA. = 1.5, = 12 FIQ. a =0 . 5 , n =0 . 8 = 0.5,n = 0.143, 52-5 1 4 = 1.5, 12, i8, = .05,n

WAVE URVEC VELoCrTY1 = 3, less tha n on an uncompensated full velocity ant enna of

= _0.8 the same length. The intensity factor or receiver

wave. at th e slowed down current per unit value of th e quant ity

-

equation

antenna can be multiplexed  by compensation and receive longer waves

t that for shorter waves i t is unfavorable.

is shown in Table VI for th e shortantennaswhose directive curves are given in Figs. 47 t o 53. The

QUARTER= 12, 0.5, n

k h n e t e r We observe th in

than th at for which its velocity is

. Feb. 1923 THE WAVE ANTENNA 243

55 shows the values of X/Zfor various values

specific resistance and dielectric

. With long waves and low-resistance soils

i snegligible compared with g’, in which case

Thisisafunction of wave length and specific resistance,

and is shown in the sloping lines of Fig. 55. On the

other hand if th e waves are so short and the soil re

-sistance so _{high that g’ is negligible compared with}

which is independent of wave length orsoil conductivity

and is shown for several values of  K in the horizontal

lines at the to p of the figure. To find the value of

X/Zfor a certain wave length and a soil of a given

resistivity and dielectric constant we use whichever

curve (the sloping line of equation (33)orthe horizontal

line of equation (34) gives th e lower value for

near th e intersection of t he two straight lines weusethe

=

RONT TILT, FORMULA

transition curve which is shown dotted. To illustrate,

if = 1000meters and K = 4,we find

X/Z= X

2 . 5 X

3 . 9 10for p = 1

x

. (on curve)

4 . 9 10 for = ₄

horizontal line for = 4).

transition (On

It will be noted that in formula

ratioX/Zis avector qua ntity whose phase angle ranges

‘from0 deg. to 45 deg. By far the most common

con-dition is that greatly exceeds the earth

carries current by conduction ra ther than by capacity)

in which case the phase angle is nearly 45 deg. The

 phase difference means t hat th e vertical and horizontal  potential gradients do not become zero simultaneously,  bu t the electric field is a rotating one, Under these

circumstances if a straight conductor is held in various

 positions in.a vertical plane parallel to the direction of wave propagation, there will be no direction of the

con-ductor in which the electromotive force induced in it

is zero. This may account for the doubtful results

obtained in attempt stomeasure wave til t b y observing

the angle of zeroor minimum electromotive force in a

straight conductor rotated in a vertical plane parallel

toth e signal direction. Wherever the tilt isconsiderable

so th at i t might be readily th e minimum is

correspondingly dull.

More satisfactory as a test of t he theoretical con

-clusions, would be quantitati ve measurements of the relative magnitudes of t he electromotive forces induced in horizontal and vertical conductors, for vari ous wave lengths and ground conductivities. Determination of

the phase relations a further check.

An observation of relative signal intensities on a

large loop and a wave antenna, indicated a wave tilt

of th e order of magnitude called for by

formula, but little data of thi s have been taken.

Th e large values of horizontal voltage gradient found in the measurements mentioned below by Beverage and Weinberger were at first considered greater than could be accounted for by wave tilt. Assuming  probable values of ground resistance, the ratio of horizontal t o vertical potential gradient according to Zenneck’s formula is of t he order of magnitude of one or two per cent, whereas the measured horizontal

gradient was about 30 per cent of the vertical gradient

calculated by Austin’s formula. Th e space potential

theory of action however is still less capable of account

-ing for the potentials observed. If we assume ground

water to be 100feet (30 meters) below the surface of

the ground, and the earth above ground water level to

have aspecific resistance of 2

x

cube, which isabout the value found by measurement,

we find for a 15,000meter wave length tha t t he poten

-tial difference between ground water and surface would

 be less than that corresponding t o a difference of

eleva-tion of two feet in th e space above ground. Consider 

-ing the ground as constant potential, and expressing

the vertical potential gradient as G

expresses the change of phase

with distance tion)

th e potential of t he wire at a height h with respect to

ground would be hG and the potential

d gradient along th e wire would be

-

d x hG

wire is10meters and t he wave length 15,000meters the

magnitude of the horizontal potential gradient would be

= 0.0042G or 0.0042of that 15,000

of the vertical gradient, which is less than Zenneck’s formula gives for the horizontal potential

due to wave tilt. Theoretical analyses agree

Fig. of wave length &nstant  K c’

x/z

_8‘ (33) c’ we have

x/z

= (34) X/Z. If d/C/G. dTpT FIG.5 g W A V E F x / z BY ZENNECK’S 1 . 3 1 0 f o r p = 1

lo5

10forp = 4 lo5 -+ 106

lo6

or greater. Zenneck’s (32),the g‘ w e ’ e. measured, wourd provide Zenneck’s kind

lo6

ohms per centimeter

E - ’ ( ~ r z ’ X )

(in which e - l ( 2 r Z / h )

measured in direction of propaga

j ( ’ ~ - f ( ~ * ~ / ’ ) .If th e height of the - - j T 2 n h 2 7 r x 1 0 G gradient

more-R ICE AND I.E. E.

electromotive force induced in

a

over a conducting earth, and

the space operation is

untenable.

a

direct

-ive

curve

obtained experimentally. A transmitting operated, supplying about

5

kilowattstoasmall antenna, at wave length. On

a

field about

system of wave antennas

twenty-four lines each 55 meters long and about one meter above ground, radiating from a central point like the spokes of

a

wheel. By joining two opposite spokes together at the center,

a

wave antenna was

-obtained 110 meters long. Using th e next pair gave a similar antenna from the last. A ground of

about each end of

= iio. =

each antenna. Th e current at one end of t he antenna was measured with a thermoccuple and galvanometer, th e opposite end being damped by a resistance. Meas

-urements were taken successively on the several an

-tennas, while the radiation was kept as nearly constant as possible. From

Fig. 56 was plotted.

 A number of factors to muse a difference  between the shape of the experimental directive curve

56and th e curve A of Fig. 57 calculated by equa-tion (22). The resistance used at the end opposite e tr ue surge impedance of the

as determined by lat er measurements. Th e end were high enough in comparison with the

of the antenna to cause currents of considerable relative magnitude, t he ground

the divergence of the waves was appreciable on

account of the nearness of th e sending station. Addi

-tional observations but the workwas

interrupted before another directive curve could be obtained.

 As it

stands the directive curve shown in Fig. 56 serves as

a

qualitative check on th e theory.

During th e sameseries of short wave tests,readings were taken to show the building u p of th e curren t inan antenna. A wire sectionalized every ten meters was

used in the antenna which pointed

toward

the

DIRECTIVE CURVES FOR METER

=0.13, a =

FIG. 58

-

OBSERVED CURRENT D_ANTENNAISTRIBUTION_. I N METER

station. line was broken successively at the ints and the current read with the thermocouple and galvanometer. The values of cur 

-rent are plotted in Fig. 58. Humps such as appear in th e curve might be due in pa rt to imperfect damping,  bu t such humps are t o be expected from the theory. The total current at any point X in the antenna Fig. 26 is the resultant of t he forward wave built up on t he  pa rt of the antenna between A and X, and the back

wave from the part between X and B.

244 BEVERAGE, KELLOOG Tmnmtions A.

thqt there no

h o k o n t a l

Wire

perfectly

therefore potentid picture of E~pmh.e&zl Directive Curve. Fig.56 shows

was

v&i& 12Ometers

600meters from the transmitting station was erected consisting of

15deg.

20ohms resistance was provided at

FIQ.56-2 120

the series of readings thus obtained was present

Fig.

the ammeter was not th line,

\‘erticals length

was not perfectly level, and

hadbeen planned,

s;

FIG. 57-cALCULATED 120

AXTENNA.A

_-

SIMPLEWAVE ANTENNA.

REFLECTIONS AND END EFFECTS.

n =0.865. 1 =0.12,HORR FOR 2.0, 120 mitting The sectionalizing po

. 1923 THE WAVE ANTENNA 245

at

a point

on an antenna of length then becomes, _{The electrical constants of a n antenna or line which} are of most immediate interest, are the wave velocity

u,attenuation constant Z.

These may be ascertained  by measuring the input

X 2

- a x .

(1- ncos

Fig. 59 shows the current distribution in an ideal antenna one wave length long, calculated by equation Thus t he building-up curve found by measuring the current at various points in th e line, is of different form from th at found by changing th e length of th e antenna and measuring th e end currents. The latter shows a continuous increase as shown in Fig. 29.

A measurement was made in May 1921 of the inten-si ty of the received inten-signals on the Riverhead antenna. Mr. Weinberger of the Research Department of the

FIG. IN WAVE

Radio Corporation, brought to Riverhead a calibrated oscillator, by which a known voltage at the desired frequency could be supplied to a circuit. By this means a voltage of signal frequency was introduced in series with the damping resistance at the north east end of th e antenna, and adjusted to giveasloud atone in the receiver as the European signal which was being measured. The results of Mr. Beverage’s and Mr . Weinberger’s observations were, P. 0. Z.  Nauen, Germany, 80 millivolts; M. U. U., Carnarvon, Wales,

These correspond to about 9 and 4 microwatts respectively of received energy on th e antenna.

Since th e antenna is long the voltage readings indicate a horizontal potential gradient of 5.5 millivolts per kilometer for Nauen and 3 . 7 milli-volts per kilometer for Carnarvon, These values represent. normal receiving conditions. During fading  periods th e signals a re much weaker.

impedance- _{of the line through}

_a

_{sufficient range of}

frequency, first with the far end of t he line open, and then with it short-circuited, (or grounded, if we are dealing with a ground return circuit). The ground connection must be of low resistance, for the equations

OFTWELVEKILOMETERANTENNA.

A ND OPEN.

RICAL EAN OF

which follow are based on the assumption of a short

-circuit reflection, and all losses will therefore beattrib -uted to t he line attenuation.

As the frequency of th e current supplied to t he line is varied, a series of maximum and minimum current values are observed, corresponding to standing wave conditions which cause current loops and current nodes.

A current maximum corresponds to an impedance minimum and a current minimum to a maximum impedance. The impedance may be determined from

FIG. FIGURE FOR DISCUSSION OF LINE IM

-PEDANCE

the supplied voltage

method, which will be described. Only the maximum and minimum values of impedance are required for the  present purpose.

Fig. 60 shows the i nput impedance of a 12-kilometer two-wire antenna as a function of frequency. A maxi

-mum input impedance when the far end isopen is seen to occur a t the same frequency at which the input impedance is a minimum when the far end is

Feb.

.The complete expression for the current

 X hTEN”N

CONSTANTS

a,and surge impedance

CO8 ) / A

cos

- j , ( l - n c o s B )

(35).

5 CALCULATED CUkREN DISTRIBUTION

ANTENNA

CI

54miilivolts.

14.5kilometers

-Y

FIG.INPUTIMPEDANCE

-

FAR E_M ND G_AROUNDED_ANDB.. B

-

FAR E

C-GEOMET-. X ‘

‘g%J

61-REFERENCE

a?$,:anrentor by the rubstitution

short-THE

1

+

a I _'

-Fro. FOR DETE M R  NI  NGI ATTENUATION CONSTANT

Fig. 64 shows as a function of the impedance

ratio m, and Fig. 65 shows a1as afunction of m.

If the frequency at which it is desired to determine

the attenuation, is not such as to make the line an

exact number of quarter waves long, th e values of maximum and minimum impedance may be found by

interpolation, using the envelopes of the curves as

shown in Fig. 60. For example, in Fig. 60 the

FIG. CONSTANTS O F

ratio at 25,000 cycles is m =

Using this in Fig. 64 we find

T h e n

_-

a 1 log,

1.85 = = 2 . 3

x

0.2672 = 0.617 which

may also be found directly from Fig. 65.

In this case = 12kilometers so that a =

ANTENNA 249

=0.0513 Fig. 66 shows th e values of th e attenuation

constant

curvesshownin Fig. 60.

again to equations (46) and (47) we

that ifwe multiply the two together we have (Z

(Zmin,) =

Using the values of impedance for

obtained from Fig. 60 we find that th e surge impedance

of the line at this frequency is Z =

= The surge impedance or geometrical

mean of t he maximum and minimum impedance values

is indicated in curve of Fig. 60.

For finding the line velocity it is necessary to know

the mode of oscillation (or number of quarter waves on

the line) for each frequency at which a maximum or

minimum impedance occurs. A curve like the one

marked A in Fig. 67 can then be plotted showing the of quarter waves on th e line, as a function of

freauencv. A similar curve B is plotted on the same

.

of

if the length of the line is 12 kilometers and its

Thus

FIG. OF OF TWELVE KILOMETER AN

-TENNA. A

-

. LINE

velocity were equal to that of light, i t would show a

4/4 wave oscillation at = 12, or =

Since the line B is straight and passes through the origin, it may be drawn by calculating and  plotting a single point. The ordinate to the line B is

and t he ordinate to curve A is whence the

velocity ratio for any frequency is

Ordinate to B

Ordinate to A

Fig. 68shows the velocity ratio n

correspond-ing to the input impedance curves of Fig. 60. The

values shown represent fair average of thoce

observed on consisting of two bare

inch (0.26 diameter wires in multiple, seven to

nine meters above ground. Curves are also shown

which give an idea of the effect of using adifferent num

- ber of wires. The line constants depend not only on

the type of construction but on the character and moisture of th e soil and therefore will vary from place

to place and change somewhat with th e season.

.Feb. 1923 WAVE 2 e - 2 a 1

e-l - & i

= RRrxr m=sm & CURVE imped-66--ATTENUATION A NTX NN M ance 220/740 = 0.297. E - ~= 0.54. = l o g e 0 . 5 4 a n d + a Z = l o g e - = 2.31og,,1.85 0.617/12 1 .54

a, corresponding to the input impedance

Referring see max.) Zzor (51) 25,OOOcyCles 4740 220 (Z min.)

-

-435ohms. mimber = (Zmax.)

she& foralight velocity line thesame length.

wave

67-MODE OSC1LLATION

ACTUALLINE B-IDEAL

25,000cycles. l/X l/n 3

x

105 12 u/v or SG far antenhas 0.102 em.)

BEVERAGE, RICE AND TransactionsA.I.E. E. . .

fig. shows the apparatus which t he writers have for measuring line input impedance. At either voltage or current node the line is substantially a

factor load a n d ,its impedance may be determined by finding the resistance which will give

should preferably have a low reactance compared with th e impedanceto be When acurrent

maxi- mum isobserved, the resistance is substituted for the line and t he condenser

ance of t he pick -up coil. The circuit is then switched  back to the line and the oscillator frequency read 

- justed to give current maximum. If the change of

oscillator frequency has been considerable a repetition

of th e process will be needed. In th ecaseof alow loss line it is desirable to use a relatively high-reactance  pick up coil and high sensitivity meter for measuring the impedance maxima and. a lower impedance coil

FIG. VELOCITIESOF

ANTENNA FOR

and less sensitive meter for measuring the low values of

indicating coil which have been found suitable

a to

Current transforme rs are permissible. Owing tothe rapid change from low to high -current values precau

-tions should be taken to avoid meter burn outs. Oscil

-lator harmonics should be minimized by using

C, and high-efficiency, low-inductance coils e oscillator with loose coupling t o t he pick coil.

in all cases necessaryto take the impedance open and short-circuited.

a good set of readings for either condition, the of the drawn in and th e surge attenuat ion determined approximately. not necessary to use a damping resistance at

of th e line witha view t o preventing

reflection. For a given frequency the line presents a

definite impedance at its terminals, and adding re -sistance in the supply circuit merely reduces th e current and voltage supplied to the line without the rati o of voltage to currentatth e line terminals.

Itis desirable in some cases to check th e values of line constants asdetermined from the input impedance, by direct measurement. Arrangement is made for tele- phone communication over th e line asshownin Fig.71

the circuits being designed to have negligible effect on the radio frequency currents. Various resistances are tried at the far end of t he line, until a value

which gives constant impedance a t the oscillator end,

FOR IMPEDANCEOF

over a considerable range of frequency. A small amount of reactance in additio n to th e resistance may  be required to give perfectly constant impedance at the oscillator end, since the surge impedance is n ot neces

-sarily a  pure resistance. If the surge impedance changes with frequency a new resistance setting will be

required for a different frequency range. Leaving the surge impedance as found in this way, in the far end of t he line, simultaneous readings of t he currentsatthe two ends are taken atanumber of different frequencies, and t he average ratio of received t o supplied current gives the attenuation, Measurementsatdifferent frequencies of th th e line at th e oscillator end give a check on the surge impedance as found by trial at thefarend. The presence of partial reflections

ox

FIG. FOR DIRECT

on the line, which may result from changes in ground conditions, may cause considerable error in the values ofth e constantsasfound by this method.

More complete and reliable information on the  behavior of the line is obtained by supplying

constant frequency and amplitude, measuring the current at intervals along the line, with different circuit conditionsatthefarend. If the end is damped with the true surge impedance, and there are no points of partial reflection on the line itself, t he cur 

-rent will show a continuous decrease, following the exponential law. If the currentis plotted asafunction of distance on "semi log" paper, t he points will fall ona

straight line and th e slope of the line will show the

250 KELLOGG

69

unity power

nan-inductive

the-same current. The pick -up coil

measured.

adjustedtotune out thereact

-8 - w A V E ANTENNA0

F I G . GQ-CIRCUITFORMEASURING

LONG WAVES

JNPUT IMPEDANCE

lineimpedance. Fig. 70shows such an arrangement,

values

for 10,000 20,000 meter range of wave lengths.

large capacity

in th circuit,

up

It is not

characteristics of the line, both

With

envelope curves can be

impedanceand

It

the oscillator end

altering

found

. 7&cIRCUIT MEASURINGNPUT L O W Loss LINES e- OL e impedance of _ _ _ - - -_{_ _ _ _ _ _} =+% -O S S ~ 7 0 ~ ~~ C l r c r

2j.j

e e d IN-0-MI W e € z . T x c I * o o

71-cIRCUIT MEASUREMENTOF ATTENUATION

oneend

THE WAVE ANTENNA 251

attenuation constant.

If

reflections occur either

at

the end or at any other point on the line there will be humps or hollows in the curve. With the end

or grounded, this method of study shows th e standing on the line, from which the velocity and attenuation can be calculated. Curves of this

areshown in

Figs.

3;

 Interpretation of   Line Constants. The explanation of the manner in which velocity and at-tenuation are affected by frequency, is

to

 be found in the varying depth

of

 penetration of the return currents

into

the ground; Fig. 72 shows the general shape of the  path of the ground current. There is a effect” which tends

to

concentrate the earth currents near the surface. If

it

were not for this effect mean depth of the earth currents in ground of uniform con

-ductivity would be a considerable fraction of a wave length, probably between one and two thousand

with a twelve thousand meter wave. As it is, most of the earth current is within one hundred meters of t he surface, with waves of this length and soil of moderate conductivity. analysis gives the depths of  penetration of earth currents for the case of space waves

of plane wave front from which we can obtain a rough idea of the order of magnitude for the case of waves ona wire supported a short distance above earth.

0 . 8 FIG OF GROUND C U R R E N T S U N D E R CARRYING WAVES  m f perfect ground... ohms km. D-C. resistance of wires,

-

... resistance ofwires at

If the earth carries current almost entirely by con

-duction rather than by capacity, which is true for all except short wave lengths or extremely high-resistance ground, Zenneck’s formula for penetration may be stated in the following form.

in which

current density isreduced to = 0.368 of its value

a t the surface.

p = specific resistance of the earth in ohms per =frequency

This gives for a 12,000meter wave = 25,000) and

specific resistance = a  penetration

D =

In Fig. 72 it is seen that there are both vertical and horizontal earth currents, but for the vertical currents the distance is relatively small and the cross section great. If the drawing were more nearly to scale this difference between the vertical and horizontal earth currents would be still more apparent. The conditions may be approximately represented by Fig. 73in which a

19. Translation given in Fleming “Principles f Electric W

centimeter cube.

-3 . -3 4.44 5.50 6 . 6

small resistanceisshown in series with th e line

to

ground capacity and

a

higher resistance in th e horizontal return conductor, which is

at

a depth corresponding

to

the mean depth of the earth currents. The capacity of such a line would be substantially the same as for a wire of the same height, over a perfectly conducting earth. The inductance would be th at corresponding

to

a

wire

forms the return conductor. There would be

a

small added charging current

loss

due

to

the resistance in series with the capacity and a much larger loss due

to

the resistance of the horizontal return conductor. Since the depth of penetration increases with wave length we should expect greater inductance and therefore lower velocities on long waves. The greater the  penetration the lower the resistance

to

the earth cur 

-rents. Therefore the lossesareless and the attenuation less on long waves. High-ground resistance increases the penetration and loss

at

reduces the velocity and increases the attenuation. I. a specific resistance of about 2

x

centimeter cube. Since ground water occur

of something less than 100 feet (30 meters) the excess resistance and inductance of the Riverhead antenna are materially less than those corresponding to this value of soil resistance.

Table IX shows the calculated inductance and capacity of a one, two, and a four wire line based on  perfectly conducting ground. The wire spacing is taken as4feet (1.3 meters) each way with 20 feet (6.6

meters) clear above ground.

Beverage found for the sandy soil near Eastport,

L.

.

1.64

2.22

2.75

3 . 3

TABLE IX.

CONSTANTS OF LINE OVER PERFECT GROUND

0.82 1 . 1 1 1.38 1 wire  m h k m inductance, ,perfect 1.86

The observed attenuation and velocity shown in

Figs. 66 calculating the effective

inductance and resistance provided the capacity is known. The actual capacity is _{somewhat greater than}

 proximity of trees. An audio-frequency bridge measure

-ment showed 0.011 microfarads per kilometer for two wires a t a height of thirty feet. As this value seems

20. Calculated by on page 135 of Principles f Electric Wave Telegraphy and Telephony J. A. Feb, 1923 1:14 open-circuited wav& kind 4,6,7,8and 9. Observed “skin skin the meters Zenneck’s’s O c ~ C r o U npnr“TlO,+ _ _ -. 72-DISTRIBUTION W I R E capacity, k m Rw 12,000

...

20,000

...

30.000

...

= 5 0 . d p / f

is the depth in meters at which the earth

1/e

cf

a- lo5 ohms,

100meters.

0.006

+

D meters above a conducting plane which

thesametime, and therefore

lo6 ohms per s

at

adepth

o.on98 0.0139

1.64

wound...

and G I , _givea basisfor

as given above owing to insulators, poles, and the

BEVERAGE, RICE AND KELLOGG Transactions A. I. E. E. from certain directions. If static, while not confined

to

a

specificdirection, comes largely froma

it is important

to

have an antenna system whose directive curve has

a

small within the angle from which the heaviest static comes.

opposite

to

the desiredsignal. The spacingofthe loops

is preferably between an eighth and a quarter wave length. Compared with systems which obtain their directivity in asmall space, th e full length wave antenna

has advantage mentioned in with the

discussion of short wave antennas, namely, that the signal currents developed are strong in comparison with residuals, and therefore there is

a

 better chance of

CURVE O F LOOP

FIG. L O O P AND VERTICAL OF LOOP)

A number of antenna arrangements have

which are considerably more directive than the loop and vertical. The calculated directive curves of some of these are very similar th at of the one-wavelength, full-velocity Fig.77 shows the directive curve of a pair of loops spaced apart in the direction

of signal propagation, the currents from the two loops  being combined in the receiving circuit in such phase as

FIG. C U R V E FOR TWO L O O P S A N EIGHTH WAVE LENGTH

realizing in practise the directive properties predicted  by calculation.

Reference

to

Figs. 35 to 41 and 44 shows that the areasof the directive curvesofwave antennas are small, that the areas of the back and lobes of the curves are

254

certainquarter,

th’e connectioli

FIQ. 6 D I R E C T I V E

7 6 D I R E C T I V E CURVEFOR GIVING HALF INTENSITY

(VJ~RTICAL

beenused to

waveantenna.

neutralize for disturbances comingIfromza:direction

77-DIRECTIVE