WIRELESS ENGINEER - WorldRadioHistory.Com · 2019. 7. 17. · RADIO ENGINEERING LABORATORIES INC....

WIRELESSENGINEERThe Journal of Radio Research & Progress

Vol. XX MAY 1943 No. 236

CONTENTS

EDITORIAL. Reflections from

Unmatched Feeder Terminals .. 215

SHORT-WAVE DIPOLE AERIALS.

By N. Wells, M.Sc. .. 219

F.M. COMMUNICATION SYSTEMS.

By D. A. Bell, M.A., B.S 233

WIRELESS PATENTS .. 243

ABSTRACTS AND REFERENCES 245-266

Published on the first of each monthSUBSCRIPTIONS (Home and Abroad)One Year 32/- 6 months 16/ -

Telephone: Telegrams:Waterloo 3333 (35 lines) Experiwyr Sedist London

ILIFFE & SO':S LTD. DORSET HOUSE STAMFORD ST. LONDON S.E.1

WIRELESSENGINEERENGINEER

FROM EVERY POINT OF VIEW"Frequentite " is the most suitable insulating material forall high frequency applications. Ten years ago we introducedthe first British -made low loss ceramic, and consultationwith us before finalising the design of new components is awise precaution.

M4y, 1943

STEATITE & PORCELAIN PRODUCTS LTD.Head Office: Stourport-on-Severn, Worcester. Telephone: Stourport Ill. Telegrams: Ste3tain, Stourport.

SPI9

May, 1943 WIRELESSENGINEER

These four colossal heads, occupying anarea of II acres of solid granite rock -faceand proportionate in size to full figures ofmen 465 ft. tall, will weather the stormsand fury of ages to come.

The ubiquitous " VARIAC " A -C Voltage Control

-a device with 1,001

uses, covering innu-

merable applications

in the electrical, radio

and "communica-tions" fields. In 15

models, with power

ratings from 170

low asoOTT

F OEMATIONAL M

TOICAL WOOSLOS ARTISTOVISIO 1.14C11005 APIO Fr

TO TO SALOLAO IC TIP OFAIM Araftonstvccr

al CALVIN CRI

CILYCLT P554us IT 500 B&RNLI IICIAK

A

I', May, 1943

BOONTONMODEL 170-A Q -METER

The latest and improved type, designedto extend range of measurement to

200 Mc.

BOONTONSIGNAL GENERATOR forFREQUENCY MODULATION

Standard frequency ranges I to 10 and41 to 50 Megacycles. (Special ranges

to order.)

Also products of :BALLANTINE LABORATORIES INC.MEASUREMENTS CORPORATIONTRIUMPH MANUFACTURING CO.

SOLE AUTHORIZED BRITISH AGENTS

A&H Radio LtaBuchanan Buildings, 24,Holborn,London,ECIPhone: Ho/born 7394 Groins Tun gsol Smith London

WIRELESSENGINEER

COSSORDoable Beam

CATHODE RAYOSCILLOGRAPH

Produced by the pioneers of CathodeRay Tube development and applica-tion, Cossor Model 339 Oscillographis the most comprehensive instru-ment of its type. In addition toremarkable performance, it is uniquein enabling the simultaneous exami-nation of two variables such as

voltage and current.

Further details may be had onapplication to :

A. C. Cossor Ltd_InstrumentDept., Cossor House,Highbury Grove, London, N.5.

'POINT' TO CONSIDER

TUBULAR CABLE MARKERSARE SELF -FIXING BY SHRINKAGE

NO roots' aftmiuo

No rubber used Supplied ready for use inall sizes and colours, indelibly printed in black

Imperishable Impervious to oils and petrol-eum Do not increase diameter of cable.

VISKRING

NN

CABLE MARKERS

VISCOSE DEVELOPMENT CO. LTD., Woldham Rd., Bromley. Kent

Telephone Ravensbourne 2641

WIRELESSENGINEER

A

WIRELESS TELEGRAPHY BOARDSPECIFICATION K.110 (Revised Nov., 1942)

Electrical

Telecommunication

Apparatusior use in

Damp, Tropical Climates

.)y testing your products in a chamber ofthis design you are assured of recording resultsidentical to those that will be obtained by thetesting authorities of all the services.Each equipment is supplied with its individualperformance curve superimposed on that of theprototype now in use at R.A.E.

We are actually in production with HumidityChambers to this specification, identical to thosesupplied by us to the parent research establish-ments of the services.Deliveries are steady though slow-materials aredifficult-considerable orders are already in hand,please help us to help you by advising us ingood time as to your requirements.

BELLING & LEE LTDCAMBRIDGE ARTERIAL ROAD, ENFIELD, MIDDX

08795

May, 1943WIRELESSENGINEER

epely

Condensers are at workPerfect communication by radio is essential

in all modern war operations, from grandstrategy down to the smallest tacticalmanoeuvre. Nothing but the very best ofapparatus is good enough.

ADVT. of A. H. HUNT, LTD., LONDON, S.W.I8. EST. 1901

TRADE

TheHome of Condensers

MARK

TRANSFORMERS FOR THE

MIIIIEN TRANSFORMER CO.Thornley Street, Wolverhampton

Tel.: Wolverhampton 22829MAKERS OP TRANSFORMERS, POWER PAC KS,N SPECIAL RECEIVING KTRAKSIIITTINO APPARATUS

FM EquipmentARMSTRONG iSYSTEM

TWO-WAY MOBILE INSTALLATIONVICTORY SERIES 565Y.

F.M. insures minimum interference-compact single -unit designI00",, crystal controlled. 25 30 watts.

NOW IN QUANTITY PRODUCTION BYRADIO ENGINEERING LABORATORIES INC.

America's Pioneer F.M. Manufacturers.Available on Priority Orders only.

Ciectronte Ongtneet-ing Jet -vices, oat.SOLE BRIT! ST.IPUT0.5

24, STANLEY ROAD, HEATON MOOR, STOCKPORT.TELEPHONE: HEATON MOOR 3107

WIRELESSENGINEER

HETERODYNE

OSCILLATORwith exclusive B.S.R. features

Never has a commercial Oscillator been produced with such a

a performance irrespective of price. Built to retain its superlativeperformance indefinitely.

8 Watts with under I % totalHarmonic distortion.

2-54" dia. Nickel Dials withslow motion.

4 Output Impedances selectedby switch.

Multi -range Rectifier Outputmeter.

All stages checked by milli -ammeter.

Hum Content less than 0.16%.

Alternative R.C. Output downto 5 c.p.s.

Output within + 0.5db. 20 to15,000 c.p.s.

110/200/250 volts, 50 cycles.

£67-10-0.Model L0800.

BIRMINGHAM SOUND REPRODUCERS LTD.,CLAREMONT WORKS, OLD HILL, STAFFS. 'Phone: Cradley Heath 621213. 'Grams : Electronic, Old Hill.

At eexfti dta*

TRANSFORMERSLine Voltage Variations of

+ 15 % reduced to 4-- 1 %Typical Specification :

Input Voltage ... 190-260 v. 50 c.Output Voltage ... ... 23o v. ± I %.Max. load ... 15o watts.

Input power factor over 90 %.Prices an application.

Write for details.

ADVANCE COMPONENTS LTDBACK ROAD, SHERNHALL STREET, WALTHAMSTOW, LONDON, E.17. Phone : LARkswood 4366 & 4367

WIRELESSENGINEER

Regd. Trade Mork

PRECISIONTESTING INSTRUMENTS

HIGH accuracy, simplicity, excep-tional versatility and proven reliabilityhave won for "AVO " Instruments aworld-wide reputation for supremacywherever rapid precision test work isdemanded. There is an "AVO " Instru-ment for every essential electrical test.

Some delay in delivery of Trade Orders is inevitable, butwe shall continue to do our best to fulfil your requirementsas promptly as possible.

Sole Proprietors sod Manufacturers :THE AUTOMATIC COIL WINDER & ELECTRICAL EQUIPMENT CO. LTD.,

The MODEL 7 46 -Range Universal

AVOMETERElectrical Measuring Instrument

A self-contained, precision moving -coil instrument, conforming toB.S. 1st Grade accuracy requirements. Has 46 ranges, providingfor measuring A.C. & D.C. volts, A.C. & D.C. amperes, resistance,capacity, audio -frequency power output and decibels. Directreadings. No external shunts or series resistances. Provided withautomatic compensation for errors arising from variations intemperature, and is protected by an automatic cut-out againstdamage through overload.

Winder House, Douglas Street, London, S.W.1. 'Phone: VICtoria 3404-7

BOONTON

A Signal Generator developed specificallyfor use in the design of F.M. equipment.

Full specification will be sent on request.

LE LANDINSTRUMENTS LTD

Frequency Ranges: 41-50mc and 1-10 mc. (Otherranges to order.)Output : .1 Microvolt to1 Volt.Deviation : From zero to2000 KC. Amplitude Modu-lation also available separ-ately or simultaneously.

2I.JOHN STREET. BEDFORD ROW. LONDON WOTELEPHONE: CHANCERY 8765

CLOUGH-BRENGLE. BOONTON-FERRIS. BALLANTINNIEWLETT-PACKARD

As soon as the war is wonwe hope to offer our usual

DAINITE " Service forMoulded Rubber Parts. In

the meantime, enquiriesfor permitted essential lineswill receive every attention.

THE HARBO 0' RUBBER CO. LTD.Hooker HARBOROUGM

TECHNICAL BOOKSsupplied from stock or obtained to order.

Particulars of books on Radio and Electrical Engineeringpest free on application.

Foreign books obtained under licence.

LEWIS'S POSTAL SERVICEPlease state interests when writing.

H. K. LEWIS &Co. Ltd.136GowerStreet,London,W.C.1

Telephone : EUSton 4282 15 lines)


UNITED INSULATOR C° LTD 9510012-20 LAYSTALL STREET, LONDON, E.C.1. a)kaitni"Phone. TERrrnous 7383 grams: Caloncl, Smith. London.

A7

Have you such a problem ?

60X/08 01-1R)p/c z

But for the war our research and design labor-atory staffed by experienced technicians couldsupply you with the answer and our manufac-turing department with the goods.We are not primarily interested in the simplemass-produced job, it's the difficult ones we liketo manufacture. To these we can bring anadvisory and production service able to coperesourcefully with problems of every kind

involving Small Electric Motors, Alternators,Generators and Electro-Mechanical apparatus.

In due course we shall be pleased to give allour attention and experience to any questionyou may submit to us.

Write for particulars of any special type of apparatus thatinterests you ; we will forward details and literature whenavailable.

SMALL ELECTRIC MOTORS LTD.,A subsidiary

BECKENHAM, KENT.of

BROADCAST RELAYSERVICE LIMITED

Rv WIRELESSENGINEER

The glory of the future..

GOODMANS INDUSTRIES

. . when the harsh

place to the merry'cello and big bassLoudspeakers will be

May, 1943

roar of battle gives

music of " fiddle,

drum," Goodmans

available again to

reproduce music's most transient

with a faithfulness rarely

achieved in the commercial

sphere.

note

GOODMANSMakers of

HIGH GRADE LOUDSPEAKERSPriority Orders only can be accepted

LTD., LANCELOT ROAD, WEMBLEY, MIDDLESEX

/5 the bas/c pr7nciAde of

CCIan LOW LOSS

CABLES

Ilheywaled/WropertiksThere is n CO -AX Cale forwAyftf:Purpre

TRANSRADIO LTD. Inc TREQUIPMENT CO. 16 HIGHWAYBEACONSHEID

WHEN YOU REQUIRE

machine -engraved, chemically -etched or die -

pressed calibrated dials and rack arms, instru-ment panels, labels etc., in metal or plastic (inEnglish, Russian or other languages) consult

THE STAR PROCESS ENGRAVING CO. LTDLower Mosley Street, Manchester 2

A.I.D FULLY APPROVED

[VOYLESLARGE DEPT. FOR WIRELESS BOOKS.

FINEST STOCK IN THE WORLD OF NEW ANDSECONDHAND BOOKS ON EVERY SUBJECT.Quick Postal Service. Books Bought.119-125, CHARING CROSS ROAD, LONDON, W.C.2.Tel.: GERrard 5660 (16 lines). Open 9 a.m.-6 p.m., including Saturday.

CSSLERS (zondon)13Machining and Turning of Plastic Material.

Signal Lamp Caps, Screw -cutting in Bakelised Fabric invited.ALBION HOUSE, 201-3, CHURCH STREET,

LONDON, N.16.Clissold 6247

Your

enquiries

are


Fl 9

AL- R

An electrical instrument may embody everyassembled component necessary for an accuratemeasuring device. But it will remain ineffec-tive for its purpose until its own individualperformance has been checked and provedand calibrated. It is on the test bench that it

X\finds itself : the inquisitor will be content withnothing less than the truth.

`C The testing of every measurement device pro-duced by Marconi Instruments-and thereare 6o of them in production - is as uncom-promising as is every phase in its design andproduction. For these are instruments madefor the measurement of very fine limits ofperformance. They apply the experience ofa lifetime to the demands of a new technicalage. In every requirement of to -day they liveup to the traditions of accuracy and resourcethat have belonged for nearly half a centuryto the name Marconi,

C.C3 :-.111 Instruments LtdST. ALBAN S HERTS TEL E P HONE 4 3 2 3

communication

industrial

ence

great

INDUSTRIALThe

DESIGNERS

applications

The

which

measure

developments.

development

of radio

to many

will be fully

to an increasing

A

& MANUFACTURERS

Victoria Station

of

subsidiary

frequency

fields

Rediffusion

at the

of

House,

engineering

of industry

service

efficiency

of BROADCAST

RADIO

Victoria

is

within its

of industries

of operation.

COMMUNICATION

Street, London,

ELECTRONICS

writing

own

RELAY

and

field

and

S.W.I

the

a

of the

SERVICE

INDUSTRIAL

general

new

provides

future

L 1.111

(Phone

chapter

LIMITED

Victoria

technique

knowledge

and

fl'E il

ELECTRONIC

in

will

6131)

of

the

and

contribute

EQUIPMENT

electronic

history of

experi-

in

ft I() WIRELESSENGINEER

witches?From small Q.M.B. to heavy duty

pattern switches-millions havebeen supplied and have

given o.nuplete satisfac-tion. State the specifi-cation - give your

priority Ct. Nos.-and your order

will receiveimmediate

atten-tion

ELECTRICAL AND RADIO LABORATORY APPARATUS ETC,

180 TOTTENHAM COURT RD., LONDON, w.iand 76 Oldhall Street,r_Liverpool, 3, Lanes.

LOND EX for RELAYS

[-Telephone: SYDEN H AM 6258 & 6259.

For A.C. and D.C.

1116

STEP RELAY LF/FS(Heavy:Silver Contacts).

Also Aerial Change -over Relays.Second Impulse "OFF."First Impulse " ON."

Ask for Leaflet 88A WE.

LONDEX LTD.Manufacturers of Relays.

Anerley Works, 207, Anerley Rd., London, S.E.20.

TESTO5COPE:Make; So .m-

portant tests, 100to 750 volts, A.C. or

D.C. Used everywhere by' Electricians, Wiremen, etc. From

Wholesalers or direct. Send for leaflet .G24"RUNBAKEN -- MANCHESTER-- I

Piezo QUARTZ CRYSTALSfor all applications.

Full details on request.

QUARTZ CRYSTAL CO., LTD.,(Phone : MALden 0334.) 63-71, Kingston Rd., New Malden, SURREY.

May, 1943

H1VAC LIMITEDGreenhill Crescent.

Telephone: Harrow 0895 Harrow on the Hill. M iddx.

MINELECTRIC Write for book-

let on lifting andCHAIN shifting or separ-

PULL EYate catalogue ofconveyors, cranes,

BLOCK i=lea°1theflar7lhi;equipment.

GEO. W. KING LTD.,HARTFORD WORKS HITCHIN

HITCHIN 960H ERTS.

Appointment VacantDirector of research and development, to takecomplete charge of these departments, required by,a London firm of Electrical Engineers manufacturingradio frequency, acoustic and electrical instruments.First class qualifications in Physics or ElectricalEngineering essential, coupled with wide industrialexperience. Write, stating full particulars, to"Wireless Engineer" Box 829.

\TA Si NOVSpecialists in Sound Amplifying Equipment, Microphones,

Power Amplifiers, Loudspeakers, etc.Also makers of Scientific Instruments.

TANNOY PRODUCTS (Guy R. Fountain Ltd.) W.NORWOOD, S.E.27

411


POWER LEVEL INDICATORS

Power Level IndicatorsTYPES D -152-A and 11-152-B

* Embodying a simple resistance bridge,self - balancing at a predetermined powerlevel - Robust -Accurate - Inexpensive.

* Type D -I52 -A indicates a fixed powerlevel; D -152-B is variable over a limitedrange.

* Equally accurate on D.C. or A.C. Accuracyis independent of wave form and frequencyup to 53 K/cs.

STANDARD TYPES

* With A.C., telephones will indicatebalance to better than 0.1 db; a reasonablysensitive galvanometer (e.g. 100-0-100 uA.)gives equivalent discrimination with D.C.

* Cannot be damaged by overloads up to1000%.

* Indispensable to Transmission Engineers,Meter Manufacturers, and others requiringa standard power level.

Full particulars are given in Bulletin B -540-Aa c py of which will gladly be sent on request.

TYPE NO. POWER INDICATIONREF. I mW

INPUTRESISTANCES DIMENSIONS WEIGHT

D- 152-A + 20 db ± 0.1 db 1 6302 ± I% 6" x 5" x 5j" 3 lbs.

D -152-B + 10 to ± 22 db 6002 ± 5%+ 0.2 db

8" x 5" x si" 4 lbs.

MUIRHEAD & COMPANY, LTD.ELMERS END, BECKENHAM,KENT, Beckenham 0041-0042FOR OVER 60 YEARS DESIGNERS

AND MAKERSOF PRECISION INSTRUMENTS

MUIRHEADC.R.C.40

Electrical Standards forResearch and IndustryTesting and Measuring Apparatusfor Communication Engineering

BRIDGES

RESISTANCES

INDUCTANCES

CONDENSERS

OSCILLATORS

WAVEMETERS

ALL TYPES-ALL FREQUENCIES-ALL ACCURACIES

_ CapacitanceInductanceResistance

H. W. SULLIVAN- LIMITED -London, S.E.15

Tel. New Cross 3225 (Private Branch Exchange)

Fl 12

COMMERCIALINDUSTRIALSCIENTIFIC

MEDICALEXPERI MENTAL

DOMESTIC

WIRELESSENGINEER

May, 1943

The accurate timing of watches and clocks is normallya lengthy procedure and involves keeping the instrument under

observation for a considerable period.

The development of the Thermionic Valve and Cathode Ray

Tube have made possible the design of special apparatus by

means of which any timepiece can be regulated with greatprecision in a matter of minutes.

This is yet another example of the important part whichthe valve plays in solving specialised problems of controlwhich arise in almost every industry.

MULLARDTHE MASTER VALVE

A Valve for Every PurposeTHE MULLARD WIRELESS SERVICE CO. LTD., CENTURY HOUSE, SHAFTESBURY AVE., LONDON, W.C.2 (47)

6

WIRELESSENGINEER

Editor HUGH S. POCOCK, M.I.E.E.

2CS

jechnical Editor Prof. G. W. 0. HOWE, D.Sc., M.I.E.E.

VOL. XX MAY, 1943 No. 236

EditorialReflections from Unmatched Feeder Terminals

ALTHOUGH this subject is far fromnew and has been dealt with in a num-ber of articles and text -books, it has

lately become of increasing interest to agreat number of people, and we thefeforepropose to explain a simple graphical methodof representing the conditions that existunder various circumstances and of solvingThe design problems that arise.

If a pair of Lecher wires or the feeder toa short-wave aerial has a terminal load ofimpedance Z1 the resultant voltage wave isthe sum of the incident or forward wave andthe reflected or backward wave. If theformer is represented at the load end of theline by the vector Vf, it will be representedat any point along the - line by V fejsx where/3 = 2-7-r/A and x is the distance from the end.If the reflected or backward wave is repre-sented at the end by Vb, it will be repre-sented at a distance x by Voe-isx or by(pVfosY) e-isx where /3y is the phase displace-ment suffered by the wave on reflection andp the magnitude of the reflection coefficientVb/Vf. If the line be assumed to havenegligible losses, its characteristic impedanceZo will be a resistance R, equal to V(L/C),and the currents of the two waves will bein phase with the voltages and, of course,proportional to them. For the resultantvoltage and current at any point, we have

V = V feigx pV fe - 96(x - y)

I = I feigx - pI fe-the negative sign in the current equation

being due to the fact that the two wavestransmit power in opposite directions. Atthe load end, x = o,

andV z = V7 + PV fej(33' (a).

Ii = If PI feiPY (b)

V, V f VfZi R,- P Roe'

or V1-° = V1 - pV fejsYZi

From (a) and (c)

V/ (1 Z°) = 2Vf

and

V z(r = Z°/ 2pVf-P:IPY . .

(c)

. (d)

(e)

These equations are shown graphically inFig. i for a specific case. The whole diagramrotates, of course, at an angular velocityw = 27rf, but it is drawn for the momentwhen the voltage across the load is passingthrough zero. As one proceeds back alongthe line Vf rotates anti -clockwise and Vbclockwise as indicated by the arrows, untilat a distance x, = y/2 they come into phaseand give an antinode with a maximumresultant voltage V, equal to their sumVf + Vb. Proceeding further along the linewe reach a minimum V2 equal to Vf Vb

at a distance x2 = Y- -4

, since the two2

vectors will then have turned through go°

216 WIRELESSENGINEER

but in opposite directions from the maximumposition. After another distance of A/4 theywill come into phase again and so on.

If

minimum V2 Vf -Vb- k

then VfVb- k - 1'

in which the symbols refer merely to themagnitudes of the vectors.

An ExampleAs an illustration, we assume that a line

for which VLIC = Ro = 600 ohms is fed

maximum V1 Vf+Vb.= =

Fig. i.

by an oscillator at a frequency of ioo Mc/sand terminated by an impedance Z1, thenature of which we wish to determine. Itis found that a maximum occurs at a distancex1 of 3o cm and that the voltage V1 at thispoint is 3.7 times the voltage V 2 at thefollowing minimum.

Vb k-i 3.7 - I 2.7P Vf- k I- 3.7 + - 47- 0.575

and, since A = 30o cm,zirRy=zx1fl=2 X 30 X300

= 0.47r= 72°

A line of any suitable length is drawn in Fig.to represent Vf, and Vb is drawn at an angle

fly of 72° and equal to 0.575Vf. Equations(d) and (e) show that by simply doublingthe lengths of these. lines we get VA' +RolZ1) and Vi(i - Ro/Z/). On completingthe triangle and joining the origin 0 to the

May, 1943

mid -point C of the base AB we have Vi, theresultant of Vf and Vb, and to the same scale

CA = V1R0/Z1.OCHence Z1 = Ro , which

we find from the figure to be 790 ohms.This is then the niagnitude of Z1 ; its phaseangle is 5805°. It will be seen that I/Zi,operating on OC, turns it through a negative

angle to CA, but since -P = -27-(1,/ the

load impedance is Z1/58.5° = 790/58.5°= 412 + 5673 ohms.

If the problem had been reversed, that is,if we had been given the terminating impe-dance Z1/0 and the line characteristicR0, and were required to find the positionsand relative magnitudes of the maxima andminima along the line, the line OC = V,would be drawn to any suitable scale andACB set out at the correct angle # andlength zViRolZ,. On completing the triangleVf, Vb and the angle fly can be read off ;then

k- maximum Vf Vb I p. .

mimmum V f Vb I -pand x, = y/2 = A . -I3Y

It is interesting and instructive to considerwhat happens to the diagram as the nature

A

Fig. 2.

of the terminating load is varied. If .themagnitude of Z1 is maintained constant butits phase angle 4 gradually reduced, the lineBCA will swing round left-handedly until,


when the load is purely resistive, OBCAwill lie on a straight line, Vf will reach itsmaximum and Vb its minimum value, andthe antinode will occur at the end of the line.

If the purely resistive load is made equalto R0, BC and CA become equal to OC ;the point B thus falls on 0 and Vb vanishes.The line is then correctly matched.

If the load is now made capacitive, thatis, if # is made negative, BCA will swingfurther round, until, when has the same

Fig. 3.

value as in Fig. i but of opposite sign, we. get Fig. 2, which we can also get by lookingthrough the paper at Fig. i upside down.Vf is still bigger than Vb but it will benoticed that as one proceeds along thefeeder, they get further out of phase, and aminimum will occur before the first maxi-mum is reached. If one still denotes theacute angle AOB by fly then the aboveformula for x1 will give the distance beyondthe end of the line to a fictitious maximum,but if fly denotes the large external anglethen the formula will give the distancex1 to the real maximum, and the firstminimum will occur at a distance x2 = x1- A/4.

If the load is purely inductive or purelycapacitive, the triangle AOB becomesisosceles, Vf = Vb, the maxima = 2V, andthe minima = o.

If the line is open, that is, if Z, is infiniteA and B coalesce at C and Vf = Vb = V4/2.If Zi is made progressively smaller but V1decreased proportionately so that II remainsconstant, the line ACB will remain ofconstant length but C will approach 0 untilwhen Z, = 0, that is, when the line is short-circuited, C and 0 coalesce, and V,. = - Vb= IiR012.

217

Current WavesUp to this point we have confined our

attention almost entirely to the voltagewaves, but we saw that in finding theresultant current wave, the sign of thebackward or reflected current wave must bereversed. This is illustrated in Fig. 3 inwhich the scales of voltage arid current aresuch that the same vector serves for both ;this is possible since both Vf/./f and 17,//bare equal to R0. If is drawn in the samedirection as V1, but I, is drawn in theopposite direction to Vb. VI is the result-ant voltage and II the resultant current atthe end of the line. Taking Vf as ourstandard of reference, i.e., keeping VI,

horizontal, as we proceed along the lineaway from the load, the line Vb/b rotatesas shown through an angle 2/3x = 72o° xx/A where x is the distance.

It will be noted that a complete revolutioncorresponds to only half a wave -length. Acircle of radius Vb is drawn about each endof Vf and it is seen that the ends of thevectors V and I (they cease to be V, and Itas soon as one leaves the end of the line)move around the right-hand circle, alwaysbeing at the opposite ends of a diameter.When x = 3o cm, 2flx = 72° and V and Iwill be horizontal, Va maximum and Iaminimum. For the next 180°, i.e. A/4, thecurrent 'will lead, and they will then comeinto phase again with V a minimum and Ia maximum.

Fig. 4.

Stub MatchingFigs. 4 and 5 illustrate another example

of a different type. Here thq 600 -ohmfeeder is assumed to be terminated by a non -inductive resistance R1 of ioo ohms. SinceR0/R1 = 6, it will be seen from (d) and (e)that Vf = 3-5V/ and Vb = - 2.5V/. InFig. 4 V1,Vf and Vb are drawn horizontallyto any suitable scale. Using as before the

218 WIRELESS May, 1943ENGINEER

same vectors to represent- either currents orvoltages, It will be represented by the sumof Vf and Vb reversed. As one proceedsback along the line, Vb rotates as shown,and the ends of the vectors V and I movearound the right-hand circle. When thedistance x1 from the end is 18.3 cm, theangle 2f3x1 =7, 44° and V is exactly tangentialto the circle. Fig. 4 is drawn for this point.V is at right -angles to the diameter and 1and V form the hypotenuse and base of aright-angled triangle. The current I maybe regarded as the resultant of two com-ponents in parallel, one in phase with Vand represented by the same vector, and theother in quadrature with V and representedby the diameter of the circle. The in -phasecomponent being represented by the vectorV must be equal in magnitude to V/Ro, i.e.to V/600, and if the quadrature componentof the current could be eliminated thefeeder would appear to be perfectly matched.The elimination of the quadrature com-ponent is achieved by connecting at thispoint a short-circuited stub or length ofline, which we assume to be of the samecharacter as the feeder, viz. with R, = 600ohms as shown in Fig. 5. If this is madeof the correct length it will take a wattlessleading current equal to the lagging quadra-ture component in the feeder and thus cancelit, leaving only the in -phase component VIR,.

X

Fig. 5.

Fig. 6 shows the conditions in the short-circuited stub. In this case we have alreadyseen that, Vf = = Vb = 11R0/2 ; II is there-fore made twice as long as Vf;Vlis, of course,zero. When Vb begins to rotate the currentis lagging, but after 18o° the current passesthrough zero and becomes a leading current.When the angle 2flx2 is equal to 309°, i.e.when x2 = 129 cm the conditions are as

shown in Fig. 6 in which the right-angledtriangle IV is similar to that in Fig. 4, butwith. the phase -angle reversed, so that thecapacitive reactance of one balances the induc-tive reactance of the other. The stub in Fig. 5must therefore have a length x2 of 129 cm.

In practice it is not necessary to drawFig. 6 since the length x2 can be obtaineddirectly from Fig. 4. Since the two right -

Fig. 6.

angled triangles must be similar the linejoining the right-angle to the mid -point ofthe hypotenuse must make the same anglewith the hypotenuse. In Fig. 6 this angle 0is 27 - 2flx2. Hence if in. Fig. 4 a line bedrawn from the right angle to the mid-point .of the I vector, the angle 0 will beequal to 27 - 2#x2 from which x2 can becalculated. Hence Fig. 4 gives us straightaway both the distances x1 and x2 withnothing more than a protractor and a pairof compasses.

If the stub is not short -Circuited butopen, the initial positions of the I and Vvectors in Fig. 6 will be interchanged, butthe same result is then obtained by in-creasing or decreasing the angle 213x2 by18o degrees, i.e., by making the open stublonger or shorter than the short-circuitedone by A/4.

Attention should be drawn to the factthat in this example we have assumed theload to be purely resistive and the stub tohave the same characteristic resistance asthe feeder. Any departure from these twoassumptions necessitates some modificationof Fig. 4.

It need hardly be emphasised that thissimple treatment of the problem is basedon the assumption that the transmissionlosses are negligible over the short lengthsof line involved, so that we have to dealwith circles and simple trigonometricalfunctions instead of spirals and hyperbolicfunctions. G. W. 0. H.


Notes on

Short-wave Dipole Aerials*By N. Wells, M.Sc.

IntroductionTHE horizontal half -wavelength aerial,

usually known as a dipole, is the mostubiquitous and popular type of aerial

unit employed in the entire range of short-wave radio communication ; this may bebecause of its symmetrical and technicallysatisfying appearance, coupled with easeof erection, but whatever the reason a dis-cussion of the various factors affecting itsperformance, and the performance of alliedS.W. aerials, may be of interest: Actuallyin the course of the discussion it will soonbecome apparent that the half -wavelengthis only important in the case of singledipoles, where exactitude of dimension, i.e.of tuning, permits the Y form feeder couplingto be standardised and applied with easeand certainty : when it is a case of twinaerials, either employed alone as in Fig. 4,

or grouped in the form of beam arrays, thereare standing waves along the feeder whetherthe twin aerials are dipoles, or differ appre-ciably from the half -wavelength ; in fact,

azo

L0475 X

OZD

BUFFERLENGTH

600 OHM FEEDER

Fig. 1.-Transmitting dipole.

the precise half -wavelength dimension hasno longer any particular virtue. As sug-gested above, the discussion centres mainlyaround the horizontal form of S.W. aerialbut, because of its importance, the verticaldipole has been given a certain amount ofattention. The notes are descriptive ratherthan analytical, being based partly ontheory and partly on experiment, as this

* MS. accepted by the Editor, August, 1942.

219

appeared to be the most useful form inwhich to present them.

Dimensions and Coupling for Single DipoleTo some extent the effective length of a

dipole is dependent on the added capacitanceof its end insulators, though for most casesit is sufficiently accurate to make the physicallength 95 per cent. of half the workingwavelength. As to diameter, for low powersthis is simply a matter of tensile strengthto withstand the stress in the triatic, andgenerally No. io S.W.G. or 7/o.o44in. strandis sufficiently strong ; for the higher powerslarger diameters are essential, not from thepoint of view of strength but in order toneutralise the risk of brushing and " torch "discharge, of which more will be said whendiscussing feeder dimensions.

Some notes on coupling between aerialand feeder are given at the end, under theheading " Line Termination " ; in the mean-time it may be said that in the case of singledipoles the usual coupling is in the form of asymmetrical Y tapping, on either side of thecentre, the length included between thetapping points being adjusted to throw back .

on to the feeder a load equivalent to someboo ohms for transmitters, and some 500ohms for receivers. The correct tappinglength varies slightly with the height ofdipole above earth, but for transmission areasonable compromise is- to take 24 percent. of the overall length, while for receptionthe figure is 20 per cent. Appropriatedimensions are indicated on Figs. i and 2.

Directivity of Horizontal DipolesThe ordinary single dipole has appre-

ciable directivity, as shown by the half polardiagram A of Fig. 5. When greater direc-tivity is required a simple and effective planis to employ twin dipoles in line, as shownin Fig. 3 (a), for which the corresponding halfpolar diagram is curve B of Fig. 5 : in this ar-rangement there is no attempt to match the

220 WIRELESS May. 1943ENGINEER

feeder, the latter being simply connected tothe inner ends of the twin dipoles. Stillgreater directivity can be realised by in-creasing the twin dipoles until they are twinaerials each with a physical length of o.625A ;

L -o 475X

1

Fig. 2.-Receiv-ing dipole.

03L

500 OHM FEEDER

in this case the half polar diagram is givenby curve C of Fig. 5, and it will be noticedthat two side loops have appeared. Roughlyspeaking, the field strength due to twindipoles is 20 per cent. greater than that fora single dipole, while that due to twin Iwavelength aerials is 45 per cent greater,figures which correspond to gains of i.6 db.and of 3.3 db. respectively. The three dia-grams are intended to indicate relative fieldstrength on one side of and in the plane ofthe aerial when this plane is inclined tocoincide with the major axis of emission,as discussed a little farther on.

The Multi - wave Aerial and HorizontalOmni -aerial

The dipole of Figs. r and 2 is a continuousconductor, and constitutes a fairly selectivecircuit, hence when working over a wide

0475A 0.475ATO TO625A-.1 r4- 0 625A-01

AS CONVENIENT

(a)

Fig. 4, we have at once an aerial that is lessselective and, because of balance, suitablefor multi -wave working. For this type ofaerial the upper frequency limit is critical,being governed by the 5/8 A relation given inthe previous section, and is thus such that :-

Overall Length = 1.25 (A min.)If the wavelength is shorter than given bythe above relation the side,lobes, as seen oncurve C of Fig. 5, increase at the expense ofthe major loops until the latter disappear-see curves C and. D of Fig. 6. The maximumwavelength is determined by considerationsof radiation efficiency, and is not at allcritical ; very roughly it can be taken asbeing from twice the overall length, giving

2

DOW NLEADS(ELEVATION)

22

Fig. 4.-Horizon-tal omni-direc-tional aerial

(spot wave).

thus a waverange of 2.5/1.0, to as much as4.o times the overall length, when the wave -range becomes 5.0/1.0 : a further note onthis matter will be found in the Appendix.Actually the form of aerial shown in Fig.3 (b) is sometimes preferable for multi -waveworking to that in Fig. 3 (a), as it is less direc-tive at the upper frequencies because of a slightdegree of vertical polarisation. It is impor-

-1 °A

B s.- 1 °

(b)

Fig. 3.-(a) Directive twin aerial. (b) Multi -wave aerial. A, 1.25 (Amin) ; B, 0.15 (A,,) ;C, 0.20 (Amin) ; D, Buffer lengths about loft.

band it is necessary to employ an aerial ofdifferent and less selective form : by thesimple expedient of splitting the conductorinto two halves, somewhat on the lines of

taut to appreciate that the multi -wave aerialhas marked directivity on the upper fre-quencies. The distance between aerial andtransmitter, or receiver, should be kept as


low as possible because of standing waves onthe feeder.

For omni-directional working there arevarious types of horizontal aerial, but un-doubtedly the simplest and most easilyerected is the right angle V pattern, in whichthe length of arms is made equal to one-half of the working wavelength, see Fig. 4 :provided the arm length is as indicatedthe radiation pattern of this form of twinaerial is a compromise between a squareand a circle, an excellent pattern for omni-directional transmission or reception.

Vertical Radiation DiagramsIt will be appreciated that short-wave

communication over long distance is due toinclined rays which are " reflected " betweenthe ionosphere and the earth, and thereforeshort-wave radio transmission should bedirected at an inclination to the horizontal.It is known that for any given service thereis an optimum inclination,* which largelydepends on the distance to be covered andon seasonal conditions ; it follows that inorder to radiate efficiently, the aerial em-ployed should, have its direction of maximumemission coincident with the optimum in-clination for the service in question. Forthe purpose of these notes we may take thisinclination as varying, very roughly, be-tween a minimum of from 7 deg. to 10 deg.,suitable for the longest distances and theshorter short waves, and a maximum of45 deg., suitable for local services, when. thelonger short waves are employed.

In practice the zenithal direction of maxi-mum radiation of a horizontal aerial isdetermined by its height above earth : thisis clearly shown in the nine polar diagramsof Fig. 6, which indicate the zenithal distribu-tion of field strength, i.e. the field strengthin the vertical plane, for nine differentheights of aerial. For each height thereare two conditions of earth, the full line curvebeing drawn for ideal perfectly conductingearth, the dotted line curve being calculatedfor rather poor, rocky earth. Upon ex-amination of the diagrams it becomesapparent that the modifying influence ofthe ground, while appreciable, is not somarked at lower inclinations ; at first sight

* In reality this is an average angle, the inclina-tion varying appreciably at times on most services.

221

this would seem to give the horizontaldipole an advantage over the vertical dipole,where these conditions are reversed, but thematter will be discussed a little farther on.Let us now consider the merits of the ninediagrams. At a height of ix or IA radiationis somewhat flat, with a maximum in avertical direction in the case of IA, andaround 45 deg. inclination in the case of

: such diagrams, but more particularly

7o°

to°

so scr

0 0

Fig. 5.-Horizontal polar diagrams.

the latter, are suitable for local communica-tions, that is, for distances up to 30o kilo-metres, when the lower frequencies wouldbe employed. On rising to the height ofx there is a well-defined maximum at 3o deg.,

suitable for medium -distance communica-tions when the intermediate short-wavefrequencies would be employed : this isobviously an excellent general height. Ato.6oA the maximum angle drops to 25 deg.,which is somewhat more suitable for thelonger mid -distances, though the improve-ment is offset by the appearance of a lobepointing upwards and dissipating an appre-ciable proportion of the total radiatedenergy.. At heights of 0.75A and o.90 mostof the energy is dissipated by unwanted highangle radiation, and it is clear that heightsintermediate to o.6oA and i.oA are to beavoided. At i.oA ' there are two pairs ofwell-defined lobes, the lower with axes at15 deg. inclination, which should be excellentfor.long-distance communications. The lasttwo diagrams for heights of 1.5A and 2.oAare given more as a matter of interest,

19.45


May, 1943

though cases may arise, such as abnormal coming rays, i.e. due to rapidly fluctuatingconditions, or a difficult circuit passing ionospheric conditions.close to the magnetic pole, when their It may be mentioned here that the dotted

(e)

900

90°

h.cy2sA

1111111111!!!!"

02

h.0 -75A

08

06 N

OA

02/

90°

Milil

01

OA

0/

CT6-

9o°

9

( k) ii° 61° h- 2 AFig. 6.-Earth constants. Full line diagrams,

22.

7720

diagrams have been. calculated for poor

length chosen is 32.0 metres, the difference

conductivity ground constants of K = 5E.S.U. and 0- = io-14 E.M.U. : the wave -

a = oc ; dotted line diagrams, a = lo -14E.M.U., K = 5 E.S.U.

S'111-7°

22°

at other frequencies within the short-waveband being insignificant. A second series of

multiple lobes might give a useful diversity diagrams for good conductivity ground,effect against the type of fading that is due with constants of K = 25, and a = 10-43,to a rapidly varying inclination of down- has been calculated, but it was found that


they came just about half -way betweenperfect earth and poor earth, and so, for thesake of clarity, have been omitted.

Generally speaking when heights of onewavelength and upwards are available, it iscustomary to employ stacked dipoles, be-cause by this means we have more controlover the shape of the vertical polar diagram,and can concentrate the greater portion ofthe energy into a single pair of major lobes ;but we are now getting beyond the scopeof these notes, except perhaps, that it maybe both useful and pertinent to describethe simple two-tier stack.

The Two-tier StackBroadly speaking there are three variants

of the two-tier stack, corresponding tooverall heights of o.75A, 1.0A and 1.25A, thegeneral arrangement being indicated inFig. 8, while the corresponding three verticalpolar diagrams are shown in Figs. 8 (a),(b), and (c) : these are _useful diagrams, theinclinations being lower, and the energydistribution generally preferable, than forsimilar overall heights of single dipoles.Clearly where a more elaborate aerial systemis justified, and poles of sufficient height areavailable, the stacked dipoles should be in-stalled. An overall height of o.75A, see Fig. 8(a), is more suitable for the longer waves andshorter distances, or for when overall heightis limited ; arrangement (b), with an overallheight of i.oA is suitable for the mediumshort waves and medium distances, beingthe most frequently adopted, -though arrange-ment (c) with an overall height of 1.25A, ismost suitable for the shorter wavelengthsand longer distances. Reverting to thesketches, it will be seen that there is a con-stant distance of o.5oA between the twotiers, the variations being in the distance ofthe lower tier above ground. Note that inorder to maintain the tiers in phase witheach other there is a cross -over in the verticalfeeder that links them together. As before,the length of each branch aerial may befrom o.25A to 0.625A.

Horizontal versus Vertical Polarisation.Vertical Dipoles

Although an analysis of the effect ofreflection at the earth's surface upon thespace component of vertically polarisedwaves and the space component of hori-

10°

223

zontally polarised waves is favourable tothe latter, it should be admitted that anysuch analysis is based upon an imperfectknowledge of all the subsequent conditions.This statement is due to a fairly closeacquaintance with the results over manyyears on long-distance communication ser-vices, results which indicate that there isnothing to choose between the two systems,provided each is properly installed; whenconditions are good both systems are com-pletely satisfactory, but when conditionsare poor each system causes anxiety to thecontrolling engineer, an anxiety that finds

I.0

(c)

2cP414'1 FP*kr4 tlIP Too

O.4 0'2 02 0.4 0 6 0 10

50° 60° 70° 80° 90° 80° 700 60° 50°

0.6 0.6 0.4 02 0.2 04

50° 60° 70° 80° 900 80° 70° 60°

0.6 0-8 1.0

so°

50° 60P 70° 80° 90° 80° 70° 60° So°

1.0 0.8 0-6

60°

300

20°

10°

Fig. 7.-(a) Vertical dipole close to earth. (b) Ver-tical dipole (3.3A clear of earth. (c) Verticaldipole 1.23A clear of earth. Earth constants.Ideal (full line), a = oc ; good land (dottedline), a = Io le E.M.U., K = 25 E.S.U. ;poor land (broken line), a = 1014 E.M.U.,

K = 5 E.S.U.

no substantial lessening in changing overfrom vertical to horizontal aerials, or vice -versa. The question has become largelyone of predilection, or prejudice, witheconomic factors slowly tilting the scale infavour of the horizontal systems. Un-


doubtedly from an economic point of viewthere is much to be said in favour of hori-zontal suspension for smaller installationsand for low in -line arrays, because in thefirst place they do not require radial earthsystems, while in the second place theeffective height of a horizontal aerial ispractically the height of the supportingmasts. On the other hand, for local short-wave broadcasting services the verticalaerial is generally the - better proposition,because the surface -wave attenuation ofvertical S.W. polarisation is appreciably

0.075107625 X1Fig. 8.-Two-tier

stack. Verticalpolar diagramsof the overallheights of 0.75.1,TA and 1.25A aregiven in (a), (b)and (c) respec-

tively.

0.50n

d = 0.25 )1.b a osoXc = 0.75 )1

(b) 40° 50° 60° 70° scP 90P6o°70° 60° so° 40°

..°A1 Ile

itlittO*art -**10°

2.0 1.6 1.2 0 0.4 0.11 1.2 1.6 2.0

0)40° 50° 60° 70°80°90°80°70° 60° 50° 40°

1PeAlkilietWr 11111111..,lift" I /14

411raZ !Pi Atetegi1 AL

2.0 1.6 I.2 0.111 04 0.4 Oa 1.2 1.6

30° 30°

20°

less than that of horizontal polarisation ;while a more obvious but less importantreason is the perfect omni-directional patternof a vertical aerial.

A discussion of the difference betweenhorizontal and vertical polarisation will befound in Appendix z ; in the meantime

May, 1943

this may not be an inappropriate place tosay something about vertical dipoles. Asindicated previously, the attenuation of thespace wave is greater than with horizontalpolarisation, an effect that is most markedwhen the aerial is close to the earth's surface ;with increasing height, however, the differ-ence in attenuation between vertical andhorizontal is less marked, and generallyspeaking a safe rule is to erect a verticaldipole as high as is economically possible,whatever the nature of the service in view.Broadly speaking, an overall height of onewavelength, giving one-half wavelengthground clearance, is a good working heightfor a vertical dipole. To go a little moreinto detail, in the case of yertical polarisationthe outline of the vertical radiation changeswith changing earth conditions, and for thisreason diagrams for both poor conductivityand good conductivity earth have been in-cluded in Figs. 7 (a), .03), and (c), whichindicate the effect of the earth upon avertical dipole ; Fig. 7 (a) when close toearth, 7 (b) when the centre of the dipole israised o.75A above earth, and 7 (c) when itis raised to i wavelengths above earth.The full Tine curves represent perfectly con-ducting earth, the dash line curves representthe effect of poor conductivity earth, whilethe dotted line curves represent the effectof good conductivity earth ; in the dia-grams for the two higher dipoles the dashline and dotted line curves have beenseparated to avoid confusion. Apart alto-gether from the usual effect of height on thebasic ideal curves, it will be observed thatas regards the effect of physical earth thegain due to height is considerable, par-ticularly in the case of poor conductivityearth.

Coming to purely practical considerations,when a vertical dipole is fed at its base, oris- otherwise close to earth, it is of majorconsequence to install a good earth system.This may consist of half -wavelength radialwires every 3 deg., centred immediatelybelow the aerial and comprising 120 radialsaltogether ; the wires should be laid justbelow the surface, if possible not more thanthree inches deep, and pegged at their farends to keep them in place. The gauge ofwire is preferably No. i6 or No. 14 S.W.G.To the uninitiated 120 radials may seem anunnecessary number, but, unless the'service


is of minor importance, consistency ofresults justifies the labour and expense.For minor services, where variation inresults is of little consequence, 36 radials orless will always work. For small powers avertical dipole can be conveniently sus-pended from a single wooden pole, with agallows arm, or arms, supporting the aerialwire about three feet clear of the pole.

Base fed vertical aerials may vary inoverall height between 1), and IA ; perhapsbecause the quarter -wave is the simplestand most readily comprehended aerial it ispopular even for short waves, but this formof S.W. aerial is most inefficient, sinceground losses rise with frequency, and withthe comparatively heavy earth current ofthe quarter -wave aerial the ground lossesare very heavy ; on the other hand, the IAhigh aerial, though seldom employed, isefficient, because the vertical radiationpattern is excellent, the earth current isreasonably low, and also the terminalpotential is much lower than, say, for themore popular half -wave high aerial.

Directivity of Twin Vertical Dipoles

Where horizontal directivity is desirabletwin (parallel) vertical dipoles form a simplebut very useful combination ; when fed inphase with equal currents the equation forthe shape of the resulting horizontal radiationpattern can be expressed in the simple form:-

Fe = V di I cos (7r - cos 0)

A

where Fe is the force at a distant station,o is the direction of the station measuredfrom the centre line connecting the twindipoles, d is their distance apart, and I isthe current in each. From an inspection ofthe equation it is clear that the maximumvalue of F is V'2 / ; but there is mutualcoupling between the dipoles and, therefore,I varies with d, a point not disclosed by theapparently simple equation, hence the pro-blem of determining maximum F becomesone of determining the spacing which givesmaximum I. It is a complex problem, andis best tackled by determining the value of dfor minimum radiation resistance, when, ofcourse, I will be maximum. The curve ofFig. 13 (a) gives the variation of R,. againstd, and indicates that Rt. reaches a minimum

225

when d is *A ; at this spacing the gain overa single dipole is about 4.8 db. when 0 is90 deg.: the appropriate radiation diagramis shown on the right of Fig. 13 (b). Whend is 2 A we get the " figure of 8 " diagramseen to the left of Fig. 13 (b) : for dual work-ing on one site this is an admirable pattern,because, when both transmitter and receiverare equipped with twin aerials, each pairbeing spaced of the transmitted wave,and the two pairs erected in line, there isremarkably little pick-up in the receiveraerial during transmission. The gain of the" figure of 8 " is 3.5 db.

Reverting to Fig. 13 (b), it may be ofinterest to note that the curve is derivedfrom the argument in the latter part of anarticle by L. Lewin on the radiation resis-tance of dipoles, in the Marconi Review,No. 39 of 1939. By drawing a dotted hori-zontal line at a height corresponding to theR,. of a single dipole in free space it can beseen how the mutual coupling between thetwin dipoles adds to and subtracts from thisvalue ; further, the spacing where the R,.value of the single dipole is coincident withthe R,. of each of the twins (at about 0.45 A,etc.) is indicated, and this is obviously thespacing where the effect of mutual couplingupon the loading component (resistance) isneutral, and therefore where the individualcurrents may be. varied without mutualeffect : this may be a digression, but it isof some interest when special radiationpatterns are required.

Arrangement and Dimensions of BalancedFeeder Lines

Where possible it is obviously better tosuspend the feeder directly between aerialand lead-in insulators, but when some dis-tance intervenes it is usual to come down towithin about 3.o metres of the ground, andthence to carry on with a horizontal feederline. In the case of twin aerials there is anunavoidable mis-match between feeder andaerial, with consequent standing wavesalong the feeder ; for short lengths this isimmaterial, but on long runs the horizontalfeeder should be matched as close to wherethe downleads join it as possible. A shortsection on the subject of matching has beenadded at the end of these notes.

The choice of feeder dimensions is governedby both practical and technical considera-

-226 WIRELESSENGINEER

tions. As to practical, it is clear that in apower transmission line the twin wires, whenstretched between supports, must be farenough apart to avoid the risk of swingingtogether in high winds ; about 6in. separa-tion is the safe minimum on bo feet spans forthe smaller conductors, while loin. or izin.is the normal limit for the larger conductors.

The main technical consideration is theavoidance of the phenomenon known astorch discharge, peculiar to the higher <fre-quencies, and the avoidance of the morecommon corona effects and flash-overs ;these unwanted phenomena are, of course,due to high voltage, therefore the problemonly arises on high power. The directapproach is to reduce the voltage by reducingthe line impedance, and for very high powerswe do this by increasing the number of con-ductors from two to four, thereby roughlydoubling the capacitance and halving theimpedance ; but the practical lower limit ofimpedance is still around 300 ohms, andthis precaution alone is not really enoughfor the surges that may occur on longlines. The indirect and most effective ap-proach to the problem is to increase thediameter of the conductors ; of course thisreduces the impedance, but a far morevaluable effect is that it reduces the volt-age gradient at the surface of the con-ductor, where it is steepest, and, there-fore, most critical. There is a certainamount of literature on this subject, butbriefly it may be accepted that for a safeworking telephone load of 50.0 kW. a twinline should be constructed of not less thanNo. 6 S.W.G. (4.9 mm. dia.) conductorsseparated 30.0 cm. ; provided the line hasbeen properly terminated this gives a reason-able margin for safety. At the other endof the scale No. 14 gauge (2.0 mm. dia.) isabout the smallest wire that is mechanicallyrobust enough for transmission lines ; whenseparated 15 to zo cm. such a gauge is goodup to 5.o kW. telephone load. It might bethought that the distance between feedersis also a matter for technical consideration,but while this is true to some extent, boththeory and experiment make It clear thatwhen the initial ratio between spacing anddiameter is not less than, say, 3o, the effectof doubling the spacing only reduces thebreak - down voltages slightly, whereasdoubling the diameter increases the break -

May, 1943

down voltages in almost the same pro-portion.

The torch discharge mentioned earliershould not be confused with corona orflash -over ; it takes a flame -like formupwards into space, and is started by somepoint, such as a speck of dust, or surfaceroughness ; it is dangerous because it burnsaway the wire.

Feeder lines employed for reception aloneare in a different category, because here theconductors are brought as close together aspossible, in order to reduce extraneous pick-up ; normally 4.0 cm. is a satisfactoryseparation, but as this is small it becomesnecessary to keep the conductors apart byintroducing separators at regular intervals,say every two metres, more or less accordingto circumstances. In order to furtherneutralise unwanted pick-up it is usual totranspose a receiving feeder by means ofspecial transposition insulators, which, ofcourse, also act as separators. Care andpatience should be exercised to maintainreceiving feeders at even length ; this ismore important than appears to be appre-ciated, because a discrepancy of only onecentimetre on a long run may permit quitean appreciable pick-up. The gauge ofreceiving feeders is preferably No. i8, asthis lends itself readily to the bending attransposition insulators ; No. i6 can beused where lines are not transposed.

Load Circuit Tuning

In the case of transmitters, provided thefeeder is correctly terminated at the aerialend the load it offers approximates to600 ohms of resistance, a value that willcouple up well across a shunt circuit. Whenthe feeder is not terminated the load at thetransmitter end will perhaps vary between200 and 2,000 ohms of more or less re-active impedance ; its value can becalculated, or measured, but in most casesthe form of transmitter coupling circuit bestadapted to the load is most convenientlydetermined by experiment.

For reception the conditions are reversed,as discussed under " Line Termination," thefeeder being terminated in the receiver :this implies that the input circuit of thereceiver should offer some 500 ohms to theincoming feeder. Such a value is, perhaps,


best dealt With by a series circuit, unless,as is usually the case, the receiver is fittedwith a transformer specially wound to offerthe appropriate termination.

Reflectors and Directors

On one-way services, or substantially uni-directional services, reflectors may be em-ployed with marked advantage, as givinga gain in forward signal strength whilst,what is sometimes more important, reducingback radiation. There seems to be a certainamount of misunderstanding about reflectorspacing, due, perhaps, to overlooking themutual coupling between driver and driven ;on the whole for single dipoles a spacing ofo.r5A between aerial and reflector is best,the reflector element being exactly o.5oAlong physically ; this combination of spacingand dimension gives an increase in forwardsignal strength of 6o per cent. to 7o per cent.,say 4.o to 5.o db., whilst there is a reductionof 6o per cent. to 70 per cent., say minus8.o to minus 9.o db., in back signals, Aspacing of o.2oA with a reflector o.475Ain physical length gives somewhat betterresults, but both spacing and length are

024

0 22

2020

>, 0 1 8

0 0 1 6

5014

20120~0100

0a.Cr 008

5006004

0 02

LENGTH OFCLOSED STUB

POSITION ON LINEMEASURED FROM -

MIN.

LENGTH OFOPEN STUB

0I00 02 03 04 05 0.6 07 08 09 10RATIO OF MIN TO MAX. CURRENT

Fig. 9.-Line stub curves. For closed stubsposition is towards transmitter from Imin,for open stubs position is towards aerial fromI. The length curve applies when stub and

feeder have similar impedances.

227

then more critical, and for this reason itis not generally recommended. For twindipoles and parallel vertical dipoles, alsofor stacked dipoles, the spacing is o.2oAin the case of stacked dipoles the gains arenot so marked, but are still important.For twin aerials o.625A long the spacing iso.25A, the reflectors and aerials being in lineat their outer ends.

Directors have the same effect as reflectors,except,' as the title would indicate, the actionis in reverse. A combination of director infront and reflector behind a single dipolecan be made to give a forward gain in signalstrength of ioo per cent. For this com-bination the reflector should "be o.5oA longphysically, and spaced o.2oA behind thedipole, whilst the director should be 0.50(A - ro per cent.) long and spaced o.2oA infront of the dipole. There is a weakness inthese combinations in that they are, quitenaturally, sensitive to frequency, and onceerected it must be appreciated that they aredefinitely spot wave aerials.

Line Termination

The sequel to a mis-matched feeder linetermination is the presence of standingwaves ; in the case of transmission this con-dition leads to dissipation of otherwiseuseful energy, and also to abnormal voltages,while from the point of view of receptionthe bad effects are loss of signal strengthand the pick-up of unwanted signals. Ineither case it is desirable to have a properlyterminated feeder, a statement which doesnot imply perfection but rather such amatch, between the surge impedance of thefeeder line and the load impedance of theaerial termination, as will reduce the ratiobetween the standing wave minimum andmaximum currents to a residual figure ofabout 0.85.

Notes on the process of terminating twinfeeders are given below, but it may be advis-able to write a few preliminary words on thedifferent aspects of transmission and re-ception. In the case of transmission, whenthe coupling between aerial and feeder offersto the latter a load equal to its surge im7pedance, all the energy conveyed by thefeeder is absorbed, hence there is no reflectedenergy travelling backwards, and hencethere are no standing waves. But in the


converse case of reception the aerial is thesource of energy, and correct couplingbetween aerial and feeder simply impliesthat all the available energy is being fed intothe feeder, while the absorption of energyis now at the receiver end, i.e., in one casethe feeder feeds, or is terminated at, theaerial, in the other case it feeds, or is ter-minated at, the receiver. Fortunately con-ditions for optimum coupling are identicalwith conditions for correct termination, sothat the same form of adjustment at theaerial -feeder transformer applies in eithercase.

In order to bring about a satisfactorytermination it is necessary to measure theratio' between minimum and maximumcurrent on the feeder line, also to determinethe location of the point of minimum currentnearest to the aerial end : with this informa-tion it is possible to locate the whereaboutsof a point on the feeder where the resistancecomponent equals in magnitude the surgeimpedance of the feeder ; let us call thispoint P. There will also be a reactancecomponent at P, but - and this is thecrux of the matter-if the reactance isneutralised it follows that the resistancecomponent becomes isolated, as it were, andwill act as the correct termination for thefeeder line. The principles by which thepoint is located and its reactance determinedare well known-for instance, see an articleby Sterba and Feldman in the Proceedingsof the I.R.E. for July, 1932-and will not beconsidered here ; the technique consists ofemploying an indicating ammeter that canbe slid along one or other of the twin feederlines in order to give an indication of relativeminimum and maximum current values, asdistinct from actual current values. Havingobtained the ratio between minimum andmaximum in the vicinity of the end of thefeeder, and having also located the positionof current minimum nearest the end, we arein a position to consult the curves of Fig. 9,and from them to read off the distance of Pfrom current minimum, and also to computethe length of neutralising reactance to beapplied at P. This latter subject requiresfurther explanation and so will be dealt withnext.

A length of feeder less than one -quarterwavelength will behave either as an inductiveor as a capacitive reactance, according to

May, 1943

whether it is either closed or open at itsfar end ; this form of reactance is known asa stub, and it is obviously a particularlysuitable form for neutralising unwantedreactance at any point along a feeder line.Usually the stub has the same dimensions,

rriI 0.16A -.-IIIII

'MAX.

0 5A

A

0

+MAXB

DISTANCE ALONG FEEDERA

TOWARDS TOWARDSTRANSMITTER AERIAL

Fig. Io.

that is, the same surge impedance as themain feeder line, in which case the requiredlength can be read directly off the appropriatecurve of Fig. 9. To determine whether astub should be open or closed the rule isthat, looking towards the load end of theline, where the aerial is attached, any pointwill have inductive reaction if current isrising, while reaction will be capacitive ifcurrent is falling ; thus at a point wherecurrent is rising towards the load an openstub is indicated as neutralising reactance,whereas if current is falling a closed stub isindicated. It will be appreciated that on along line there will be a number of possiblepoints, spaced at regular intervals on eitherside of each current minimum, see Q, Pand Q' on Fig. 1o. Note that currentminimum is chosen as the datum becauseit is definitely sharper and easier to locatethan current maximum.

It will be easier to follow the foregoingremarks with the help of Fig. To, whichrepresents a feeder line along which thereis a succession of " humps " of standingwave current, and to which the down -coming aerial feeder is connected either atAA or, alternatively, at BB. If the con-nection is at AA, and the point is located atP, the reaction at P is capacitive, and hencea closed stub is required ; but were the con-nection to be at BB the correspondingnearest point would be Q, where the reactionis inductive and an open stub would be


necessary. Normally, however, the closedstub is preferable, so even in this latter casethe point P would be chosen, unless it wereparticularly important to reduce the amountof unterminated feeder line. The closedstub is preferable for two reasons, firstbecause its closed end can be made adjustableby means of a sliding clamp bar, and second;because the voltage drops along a closedstub, whereas it rises along an open stub.Note that in the case of the closed stub thevoltage at the closed end is neutral, that is,at earth potential, and for this reason it ispossible to anchor the end directly to earth,always provided it is close enough to ensurea short non -inductive connection.

It may be helpful to take a concreteexample ; suppose the relative values ofminimum and maximum currents are foundon the sliding meter to be 0.17 and 0.41amps, giving a ratio of 0.415 ; upon con-sulting the curves of Fig. 9 we find that thedesired point is at a distance of a shade lessthan o.16A from minimum current ; if thefeeder line ends at AA, as indicated on Fig.10, the point will coincide with P, but if theline ends at BB the point will be at Q. Onagain consulting Fig. 9, we find that for Pthe length of closed stub is o.132A, while forQ the length of open stub is 0.118A. Auseful check is that the addition of o.132A'and 0.118A totals one -quarter wavelength,which is as it should be.

Having ascertained the current ratio andlocated the current minimum which is nearestto the end of the feeder, both the location of,and length of, the stub should be prettyaccurately determined ; none the less slightvariations in positions, with slight adjust-ments in stub length for each trial position,are often necessary before the residualstanding wave current is reduced to thestipulated ratio of o.85, or better still,towards the ratio of 0.90. With the objectof leaving a margin for adjustment, thelengths of the stubs should be cut longerthan the figure as computed from the curve.

The following additional notes are givenas being of practical interest to those en-gaged on terminating feeder lines.

(a) In the case of the closed stub therewill be idle tails behind the sliding bar ; itis undesirable to leave these longer thanabout one-half of a metre.

(b) The stubs may be carried in two or

229

three different ways, as indicated on Fig.(c) On a long feeder line the cumulative

effect of the capacitance of the supportinginsulators may easily cause a recurrence ofstanding waves farther down the line, to-wards the transmitter, and this despite thefact that there is a good termination at theaerial end. The presence of such waves iseasily ascertained by running the ammeterfarther back along the line; should it be

(a)

(b)

P FEEDER LINE

Fig. ii.-Feederstubs.

(C)

found that standing waves have grown toany appreciable extent, the remedy lies inlocating another terminating point P in somepreceding section, and neutralising its react-ance as described above by means of a stub.

(d) The sliding ammeter may have a scaleof about one-half ampere, depending on thepower available. It is fitted with twohooks made of stout wire and firmly clamped


to prevent lateral movement : obviously thelength of feeder line between the hooks willvary the sensitivity. The meter may beattached to a wooden handle for con-venience. It should be observed that thepotential difference baween the two hooksis very small, and, in consequence, the con-tacts are very sensitive to pressure ; forthis reason it is of the utmost importanceto clean the line under test, and to recleanit at short intervals of time. To overcomethe difficulties inherent with a simple slidingmeter, difficulties which often beconie seriouson the higher frequencies, the MarconiCompany have developed a new form of duo-coupled Line Ratio Meter.

APPENDIX IThe Limits of Multi -wave Aerials

If the upper sketch in Fig. 12 is taken to representa horizontal twin aerial, 1 being the length of eitherlimb measured in wavelengths, then it can be shownthat the shape of the horizontal polar diagram isgiven by the expression:

F00( [COS 2 in - cos (2 In cos 8)].where7F9 is the horizontal field strength at any con-stant radial distance subtending an angle 8 withthe line of the aerial.

-1A( b ) L=o 6X

50°

40°

30°

7 o °

1011i **gillOW.1 I:

4 2

i 11111111.11.1*Altiltillillini

* 11011W 1004/141111111111

Irage*:Alb

30° 4 "00.4. 20°

40

30°

50°50.40°

60° #41 I 4 °I 0*.0.75A

70° SlailliOW 60°

(c) 1.

60°0° 30° tic).

600 go. 80°

0

(a) 1.=0.125A

60°50°

40°

30°

200

10°

( d ) I .i.o

Fig. i2.-Vertical polar diagram for multi -waveaerial.

This formula is common to numerous papers,e.g. see Appendix I of the author's paper " AerialCharacteristics," Journal I.E.E., 5942, 89 part 3,

May, 1943

p. 76. Moreover it can be shown that the fieldstrength normal to the aerial, when 8 is 90 deg.,reaches a maximum when I becomes IA in length,see Fig. 2 anff text of the same paper ; it followsthat when 1 exceeds this length the polar diagrambecomes unsuitable for either ordinary receptionor ordinary transmission. For instance, applyingthe above formula to the four respectivevalues of 1 = 0.125 A, 1 = 0.6 A, 1 = 0.75 A, and

= I.o A, as has been done in the lower part ofFig. 12, we see in quadrant (a) a quarter of thediagram for I = 0.125 A ; in quadrant (b) for/ = 0.6 A ; in quadrant (c) for 1 = 0.75 A, and inquadrant (d) for 1 = 1.0 : these curves are con-firmed by experiment, and represent very closelythe shape of the radiation pattern actually givenwhen the aerial is in proximity to earth, hence wecan deduce from them that when the limbs exceedft A, as in quadrants (c) and (d), the shape of theradiation pattern is quite unsuitable for generalpurposes. It is this consideration that determinesthe overall length in terms of the shortest wave-length, thus overall length = 2 x 1 A = 1.25 A,where A is the shortest wavelength.

As regards the longest wavelength, the distinctionis not so clear cut, being simply one of radiationefficiency ; for example, if each of the twin limbswere one-half wavelength long, and the aerial wereone wavelength above ground, the radiationresistance would be of the order of 170 ohms ; if,however, the wavelength were doubled each limbbecomes one -quarter wavelength long, the heightabove ground is reduced to one-half wavelength andthe resistance drops to some 8o ohms ; * hence fora given power the current would rise in the pro-

portion N,/ or roughly 5.4. If we again double

the wavelength the resistance drops to somethingthat we may guess to be of the order of io ohms,allowing for an equivalent height of one -eighthwavelength above ground, and current rises in the pro-_portion of

\/170 or roughly four times its originalIO

value, when it is pretty certain that the un-matchedfeeder losses and other losses will be mounting toappreciable proportions. It is this form of reason-ing that causes us to fix the limiting value of longwave as being from 2.0 to 4.0 times the overalllength of the aerial.

Although we are here only concerned with theshape of the radiation patterns, it may be mentionedin passing that the relative values of field strengthare shown pretty accurately in Fig. 12. It will berecalled that half polar diagrams for 1 = 0.25 A,/ = 0.5 A and l = 0.625 A (ft A) have already beengiven in Fig. 5.

APPENDIX HThe Effect of Ground upon Horizontal and upon.

Vertical PolarisationThe difference between horizontal and vertical

polarisation is perhaps best discussed in connectionwith the image. -Just as the eye demonstrates

* L. Lewin, " Radiation Resistance of a Hori-zontal Dipole above Earth," Marconi Review,1939, No. 73, p. 13.

V

May,- .1943 WIRELESSENGINEER

the existence of a virtual image " behind " thesurface of a mirror, so do measurements of the shapeof the vertical radiation pattern of an aerial confirmthat it is of the nature of an interference patternas between the aerial and its virtual image " below "the surface of the ground. But since the compositionof the ground departs from the ideal of perfect con-ductivity the image is blurred, and it is preciselythe difference in the effect of ground upon the -

14

12

.M1111111M111111MMIIIIIIIII

° 8°MEEEMESEEOSEEMEHMEHEIE1111111MIMMIIIMMIRIIIMM1111111MMICIIM1111111116:Eilll

02 04 0-6 08

60

40

20

90°A

Fig. 13.-(a) Ra-diation resistanceof one of a pair ofin -phase dipolesin free space,.when the distancebetween them, d,is varied. Note :73.2 ohms istaken as the R,of a single dipolein free space, (b)Horizontal polardiagrams for twinvertical dipoles.

1.0d

(a)

'20°

130°

140°

150°

160°170°

180°

190°

200°

210°

220°

230°

240°

12 14 16 8 20

110° 100° 90° 80° 70°

250P 260° 270° 280° boo

d- o -5X

image that marks the differenceand vertical polarisation.

Consider the images of vertically polarised and ofhorizontally polarised aerials in relation to theirrespective vertical radiation diagrams, when tracedfrom o deg. to 90 deg. In the case of horizontalpolarisation over perfect earth it will be apparentthat the virtual current amplitude of the image isunity in relation to that of the aerial, while thephase difference is constant at 18o deg. ; for phy-sical earth the phase difference of the image shifts

between horizontal

231

slightly onwards as the vertical radiation diagramis traced from o deg. to 90 deg., and this is accom-panied by an almost uniform drop in virtual currentamplitude : both of these effects are easily compre-hended, and are indicated by dotted line curveson Fig. 14. In this figure, which indicates thevariation of image phase, and also of image virtualcurrent, as the inclination of the vertical radiationdiagram is traced from o deg. to 90 deg., it will beseen that there are two sets of curves, the dottedline set covering horizontal polarisation, and thefull line set covering vertical polarisation. Thetwo sets of curves correspond, respectively, to poorground conductivity and to good ground con-ductivity.

In the case of vertical polarisation over a per-fectly conducting earth the image is in phase withthe aerial, and its current amplitude is in unityratio ; but when we come to the physical earth,an earth with resistivity, this simple pictiire nolonger holds, and a little discursive explanation isnecessary. In order to analyse the effect of aphysical earth it is necessary to split the wave intotwo components, a surface component and a spacecomponent ; this is also true with horizontalpolarisation, but here the surface component(though real) is so insignificant as to be ignored,whereas in the case of vertical polarisation the

0

50°

20°10°

0°350°

340°

330°

320°

310°

300°

(b)

120°

130°

140°

150°

160°

170°

180°

190°

200°

210°

220°

230°

240°

110° 100° so° 80° 70°

2500 260° 270° 280° 290°

d .2/3 X .

60°

40°

30°

20°I0o°35CP

340°

330°

320°

310°

300°

surface component is significant and can pre-dominate for some distance along the surface, andup to a very limited (but real) height above it : forlocal S.W. broadcasting, say of waves beyond 30metres, the property is very useful, but in this notewe are mainly concerned with long-distance effects,that is,. with the space component. In order totreat this component mathematically it is necessaryto conceive an abrupt change to 18o deg. in thephase of the image, corresponding to zero inclina-tion of radiation ; it is, perhaps, correct to say that


the abruptness is only part of the picture, and thatsince surface and space components merge, thecomplete picture is of a modulus phase shift that israpid, but not abrupt. AU this by way of explainingthe full line curves of Fig. 14, which indicate theeffect of two grades of physical earth upon theimage of a vertical aerial : the effect is obviouslycomplex, and it is interesting to note the dip invirtual current amplitude at the Brewster angle,where the phase difference has ad-vanced through a quadrant, and is at27o deg.

For further information the articlescited at the end of this appendix shouldbe consulted, though the reader iscautioned beforehand that the signof the phase angle as indicated inFig. 14 departs from that adopted byFeldman, and that this in turn affectsone sign in the equations for the vertical

Fig. 14.-Virtual current ampli-tude and phase of aerial image.Curves 1, good conductivity,

E.S. U., 0=10-'3E.M.U. ;curves 2, poor conductivity, e5 E.S.U., a = 1014 E.M.U.

A = 32 metres.

O

0

O

May, 1943

aspect is that, while the working difference betweenthe space components of the two polarisations canbest be discussed in terms of the ground effect upontheir respective images, the actual vertical radiationpatterns are a result of the combination of aerialand image, and that, in addition to the phase andamplitude to be ascribed to the image as a result ofground characteristics, there is the phase differencewhich varies with the distance separating aerial

...-..I... AMPLITUDE

-..,..\,2\-. ....... -. "" HORIZONTAL

POLARISATION.--....

a

\., --..-\ ......, VERTICAL POLINHORIZONTALPOLARISATION

AR ISAT 106 ..k.\ .............

a -.........-....-....-__CALVERT

POLARISATION2

?...,/.[n

9.340°

zc.)

g woe

0ki 260°an

I4

ig 22 0°

4Wc° 160°

140°

PHASEVERTICAL POLARISATION111111

I

._1HORIZONTAL VERTICAL

2

e1i- _

2---zi-

HORIZONTAL POLARIS --1ATION111111

radiation diagrams. The equations employed here,for Figs. 6 and 7, are :

I (vertical) = cos 8 -VI 2,4 cos (2h sin d + 0) + A'

I (horizontal) = + 2A cos (2h sin B + 0) + A°where I is the current (or voltage) value of thevertical radiation diagram at any inclination 8,while h is the angular height of the centre of theaerial above earth, A the current amplitude of theimage, and 0 its phase, both relative to the aerial :these latter two values are shown on Fig. 14 forearth of good conductivity and for 'earth of poorconductivity. Actually the formulae apply todoublets of no magnitude, and therefore the re-sultant diagram for vertical polarisation is not anaccurate trace of the diagram of a vertical half -wave dipole, but it is quite close enough for practicalpurposes.

Two aspects of this discussion may require ,a littleemphasis ; the first is that it is only concerned withthe image effect on the space component, i.e. withthe rays reflected from the earth. The second

10° 20° 30° 5 60° 70° 80° 90°

INCLINATION OF VERTICAL RADIATION PATTERN

and image, i.e. the effect of height, while there isalso the fundamental difference between the twopatterns in free space, that for horizontal polarisa-tion being a circle, while that for, say, a verticalhalf -wave dipole is the figure of 8, one-half ofWhich is shown in Fig. 5, curve A .

Finally, it must not be overlooked that there isappreciable ground absorption below a horizontaldipole : writing from memory of test data (U.S.A.)the absorption varies only slightly for heightsbetween is and, perhaps, 2.0.1, and is of the orderof 25 per cent. of the total.

(1) K. A. Norton, " Physical Reality of Space and SurfaceWaves." Proc. Inst. Rad. Eng., 1937, Vol. 25, p. 1192.

(2) R. M. Wilmotte, " The Radiation Distribution of AntennaSystems." Jour'', I.E.E., 1930, Vol. 68, p. 1174.

(3) J. S. McPetrie, " The Magnitude and Phase of the ElectricField in the Neighbourhood of an Antenna." Journ. I.E.E.,1931, Vol. 69, p. 290.

(4) J. S. McPetrie, " A Method for Determining the Effect ofthe Earth on the Radiation from Aerial Systems." ProceedingsWireless Section I.E.E., 1932, Vol. 7, p. 54.

(5) C. B. Feldman, " The Optical Behaviour of Ground." Proc.Inst. Rad. Eng., 1933, Vol. 21, p. 764.

.z

1. IntroductionTHIS paper is not intended to be either

a theoretical analysis of frequencymodulation, or a comprehensive survey

of the subject, but rather a summary of theknown data on a subject which has receivedless attention in this country than in theU.S.A. In this spirit, the bibliography hasbeen restricted to a short list of criticallyselected references ; they were chosen as themost useful papers for the reader who hasnot specialised in frequency modulation, butrequires a more detailed treatment of someof the points mentioned here. The omissionof very many papers does not imply anyjudgment on their intrinsic merit or priority.

2. Definition of Frequency Modulation andPhase Modulation

The method of telephonic modulation ofa radio -frequency signal which was firstinvented was amplitude modulation, probablyby analogy with simple land -line telephonywhere the carrier (if any) was D.C. andtherefore capable only of amplitude Modula-tion and not of phase or frequency modula-tion. But if a mathematician were toapproach the problem of carrier -wave signal-ling without previous practical experience,he would see at once that there are threeways of modulating a sinusoidal oscillationof the form

V = E sin (ad + 00)- (I)

One could operate on E, giving amplitudemodulation, on 0, giving phase modulation,or on 0), giving frequency modulation.Amplitude modulation is too well known toneed further discussion, and since there isno objection to negative values of 0, phasemodulation can at once be written down inthe *form

V = E sin (wt ciS sin pt) . . (2)

* MS. accepted by the Editor, December, 1942.


F.M. Communication Systems*Interf6rence and Propagation Characteristics

By D. A. Bell, M.A., B.Sc.(A. C. Como.. Led.)

where p/27r is the modulation frequency. Infrequency modulation the frequency cannotbe allowed to become negative, but willnormally vary by only a small fraction aboutits mean value ; by analogy with the expres-sion for amplitude modulation, namely

V = E m sin pt) sin (cut ± 0) (3)

one is tempted to write for frequencymodulation

V = E sin [(0) 80) sin pt)t +"O.] (4)But this can be re -arranged as

V =Erin [ad t 80) sin. pt + 00] (5)when it becomes clear that the factor whichoscillates at modulation frequency nowincreases in magnitude with time, and is infact a modulation of cot, not of cu. To avoidthis one writes

V = E sin [wt 80) sin pt + 00] (5a)or more simply

V = E sin (0)t + sin pt) . . (6)

since there is no point in retaining the fixedphase angle 00 of equation (5a), and 80) sin ptbeing in effect a variable phase might as wellbe written as q sin pi. Thus the frontalattack on frequency modulation has led usback to phase modulation.

In order to understand the expression forfrequency modulation,we must have a cleareridea of what is meantby " frequency," es-pecially when it isnot constant, and forthis case a conventioncan be derived with 0

233

- di3dt

the aid of the rotating Fig. 1.vector representationof alternating quantities. In Fig. I, OAdefines a fixed direction of reference, and OBthe, direction at a given instant of therotating vector under consideration. Theangle AOB, denoted by fl, is called the phase


of the vector, and the angular velocity Co ofOB at any instant is 21T times the instantaneousfrequency ; there is then a pair of reciprocalrelations between FL and /3 :

diw

fl = dt .. (7b)

.. (7a)

From equations (7a) and (7b) it is clear thatany phase modulation must produce someassociated frequency modulation, and viceversa.

Returning to equation (2) for phasemodulation, V is represented by a vectorwhose amplitude is E and whose phase (13 inthe notation of Fig. I) is wt + # sin pt ; fromequation (7a) it follows that its instantaneousfrequency is

=dt

(wt + ¢ sin pt) p cos pt

.. (8)

This represents a frequency modulation, butby an amount varying with the value of p,the modulation frequency. The equationfor an oscillation which is frequency modu-lated by an amount independent of themodulation frequency can be deduced byinspection from (8), or obtained by putting

= co - k sin pt in (713) :

/3 =/(0) -k sin pt) dt = wt -k cos pt 00.

Omitting the arbitrary initial phase co, theexpression for a frequency -modulated oscilla-tion is

V = E sin /3 = E sin (wt -k cos pt)

(9)

The frequency -modulated oscillation (9) isnot sinusoidal, but, as shown in Appendix I,is equivalent to a carrier and an infiniteseries of sidebands whose amplitudes areBessel functions with argument klp :

May, 1943

The distribution of amplitudes among thesidebands is thus a function of k/p, a quantitywhich is called the modulation index ; thequantity k, which represents the maximumpositive and negative departure of instan-taneous frequency from the carrier frequency,is called the frequency swing.

If a carrier is frequency modulated bymore than one sinusoidal signal simul-taneously (i.e. by a complex wave form) thespectrum analysis of the complex modulatedwave includes sidebands corresponding tocombinations of the modulating frequenciesas well as those corresponding to the sinu-soidal modulating frequencies taken separ-ately. (See references i and 2.) The ampli-tudes of these sidebands, however, are givenby the products of two or more Besselfunctions, and since Bessel functions are lessthan unity, these sidebands of compositefrequency tend to have smaller amplitudesthan those corresponding directly to thesinusoidal components and having ampli-tudes which are given by single Besselfunctions. The general effect is that as themodulation wave -form becomes more com-plex, the energy distribution tends tobecome more uniform over the frequencyrange covered by the frequency swing.

3. Signal/Interference Characteristics3.1. Basis of F.M. Superiority.-Of the

three possible modes of modulation, ampli-tude modulation (abbreviation A.M.) was atfirst uniersally adopted. Frequency modula-tion (abbreviation F.M.) was later con-sidered, in the hope that it would result ina narrower bandwidth than that requiredfor A.M. if it used only a very narrowfrequency swing, but mathematical analysissoon showed that this hope was unjustified,and F.M. was shelved for some years. Itwas then thought that since various formsof " noise," e.g. atmospherics and inter-ference from electrical machinery, representsudden changes of amplitude of current in

sin (cot + LI cos pt) = J0(p). sin wt

(- I)" J2n+ (p-.) {cos [w + (2n + i)pit cos [w - (2n + 1)p]t}= 0

+ E (- I)n _An (7,k) (sin ((.0 2np)t + sin (w - 2np)t)

ufr


the receiver, a gammunication system relyingonly on change of carrier frequency, insteadof change of amplitude, would be immunefrom these disturbances. Experiment showedthat a substantial improvement was obtainedby the use of F.M., though the reasoning justoutlined was faulty.

A typical " noise " E.M.F. is of com-pletely random and arbitrary wave form,and so is just as likely to be modulated inphase or frequency as in amplitude. Theadvantage of communication by means ofF.M. rather than A.M. depends largely onthe fact that if two oscillations are combined,the relation of the frequency modulation ofthe resultant to the frequencies and ampli-tudes of. the components is different fromthe relation of the resultant amplitudemodulation to the component amplitudes.The actual laws concerned are as follows :if we take two components of the formA sin wit and B sin wet, and express theresultant in the form R sin wt we find

R = [A2 + B2 + 2AB cos (W2 -w i)t]t... (ii)

=A20)1+ B2 0 + AB ((di+ (02) cos (w 2- wi)t

A2 + B2 + 2AB cos (co, - wi)t. . (i2)

In view of the close relationship betweenfrequency and phase modulation shown inequations (7a) and (7b), they have identicalcharacteristics as regards signal/interferenceratio ; but the contrast between (II) and (12)shows that F.M. and A.M. differ markedly inthis respect.

The two important factors in the signal/interference performance of F.M. systems areas follows.

(I). When the combined signal and inter-ference have been passed through a limiter,the detector will respond only to the phaseor frequency of the resultant, and theresultant of two signals A sin wit andB sin wet such that the difference frequency(w2 - w1) is supersonic, has effective fre-quency

= ((-02 (Ai) B2/(A2 + B2) (is)

so that if B becomes small compared with A,the frequency shift which it causes decreasesas the square of the ratio of the two ampli-tudes. (Equation (i3) is derived from (12)

235

by omitting the terms in cos (w2.-wi)t,which being supersonic will have no ultimateeffect in the receiver.)

(II). The band width of the receivercan be limited to the frequency swing whichhas been chosen to represent ioo per cent.modulation ; then even if B> A , so that itcauses the resultant frequen.cy w 0 to swingover from co, to w2, it can never producemore than ioo per cent. modulation, andif w2 - wi is less than corresponds toioo per cent. modulation, i.e. w2 is not atthe edge of the received band, the maxi-mum depth of modulation due to the in-terference is reduced. Further, the post -detector response of the receiver may belimited to the desired audio -frequencybandwidth, usually less than the radio -frequency half bandwidth, so that an inter-fering E.M.F. of such frequency as to producedeep modulation also produces a beat noteof so high a frequency as to be rejected bythe A.F. amplifier.

But equation (13) shows that of two signalsthe stronger predominates, and it is thereforenecessary that the signal should be at leastequal to the noise before factor (I) becomeseffective.* Increasing the bandwidth of theR.F. portion of the receiver increases theamplitude of noise at the limiter,t andtherefore increases the threshold signal levelat which factor (I) becomes effective. Oncethis threshold has been passed, however,,theextended bandwidth does not add to theA.F. noise, because the additional compo-nents are at supersonic frequency. Thesignal/noise 'improvement of a F.M. receiveris usually expressed in terms of the differencebetween the output (A.F.) signal/noise ratioof the F.M. receiver and that of an A.M.receiver having the same A.F. bandwidth.

3.2. Fluctuation Noise.-In F.M. work itis necessary to distinguish between " fluctua-tion noise " and " impulsive noise." Theformer is defined as noise made up ofnumerous components of similar amplitudeoccurring at random intervals, but occurringsufficiently frequently to ensure that a large

* Factor (II) is effective even for signal belownoise level ; see below, under " Impulsive Inter-ference."

t In examining F.M. characteristics, the effectiveinput signal/noise ratio is that at the limiter, be-cause this is (or should be) the first non-linear stageof the receiver, the point at which linear super-position of two signals is no longer valid.

236 WIRELESS May, 1943ENGINEER

number of the components occur within aperiod equal to the time -constant of thereceiver ; the response of the receiver is thento the resultant of many components inrandom phase. Using the symbol [S/N] forsignal/noise ratio expressed in db., theperformance of a F.M. receiver on fluctuationnoise is given by

[S/N]F.M. = [S/N]Am. 20 lOgioD + 4.75 P.. (14)

provided the input [S/N] is above thethreshold of improvement. D is the "tion ratio," i.e. the ratio of the R.F. halfbandwidth to the A.F. bandwidth, and theterm 20 log10D is the expression of the factthat the noise components so situated as toproduce the highest audible output frequencyproduce less than Too per cent. modulation(whatever their amplitude) by a factor of D ;the addition of 4.75 db. represents the effectof the " triangular noise spectrum,"", i.e.the fact that even within the narrower R.F.band corresponding to twice the A.F. band,the effectiveness of the noise componentsdecreases linearly from a maximum at theedges of the band to zero at the carrierfrequency. Since the strongest componentsof noise in the A.F. output are those ofhighest frequency, a top cut in the A.F.circuits is even more effective with F.M. thanwith A.M. ; it is therefore customary toapply a top boost or " pre -emphasis " in theA.F. circuits of . the transmitter, so that acorresponding cut, or " de -emphasis," maybe applied in the receiver. The resulting gainin [S/N] is represented by the symbol P inequation (14). The standard degree of pre -emphasis used in American F.M. broad-casting (with an A.F. band extending up to.15 kc/s and -deviation ratio of 5) is thatcorresponding to a circuit having a time -constant (CR or L/R) of loo ps, which has anegligible effect on frequencies below 1 kc/s,but gives a gain of about 20 db at 15 kc/s ;if the programme fed to the transmittercould be distorted in this way withoutincreasing the total depth of modulation, weshould have P = 7.35 db. for these standards,but in practice the pre -emphasis necessitatesa 'slight reduction in the average programmelevel at the transmitter to avoid over -modulations, and P 4 db. for Americanbroadcasting conditions.

To find the threshold value of signal/noise

ratio, note that equation (13) is derived interms of instantaneous values of amplitude,and when signal and interference havedifferent wave forms, the peak values shouldbe used, not R.M.S. Now ideally the peakvalue of fluctuation noise is indeterminate ;but a peak value can be defined in terms ofthe probability of its being exceeded s, anda reasonable value is 4 : i for the ratiopeak : R.M.S. But the peak : R.M.S. of asinusoidal signal is only V2 : I, so that forequal peak values the R.M.S. value of asinusoidal signal must be greater than thatof a fluctuation noise in the ratio 4 : A/2, orabout 9 db. The R.M.S. fluctuation noisevoltage increases in proportion to the squareroot of the bandwidth, so that by comparisonwith the minimum bandwidth which wouldsuffice for A.M. reception, the noise in theR.F. side of the F.M. receiver is greater by10 log10D db., and whereas 9 -I- 10 log10Dgives equality of peak values, some superi-ority of signal over noise is necessary beforethe output is acceptable. The acceptabilityof the output signal/noise ratio is an arbitraryjudgment, and depends upon the purpose forwhich the signal is required ; but thepublished data suggest two values of thethreshold, [S/2\]0, in terms of the R.M.S.signal/noise ratio measured in the A.M.bandwidth :-(a) For high-fidelity entertainment,

[SA, = 22 + I0 lOgioD (I5)

(b) For speech of high intelligibility,[S/N]o = 15 + 10 logD .. (16)

If the signal is initially below noise level,it is practically eliminated in the F.M.receiver. Theoretically, therefore, the A.M.receiver is superior to the F.M. receiver overa range of [S/N] up to 9 + io log10D db., ifthe interference consists solely of fluctuationnoise ; but in fact this represents a rangeover which the signal is not usable.

3.3. Impulsive Noise.-This is defined asnoise made up of discrete impulses, each ofshort duration compared with the time -constant of the receiver, and having aseparation long compared with the time -constant of the receiver. The typicalimpulsive interference is that due to motor-car ignition, and may have a peak amplitudemany times that of the signal, though theR.M.S. energy is small.

May. 1943 WIRELESSENGINEER

If the impulse is analysed into a series ofsinusoidal components, they will be found toconstitute a uniform continuous spectrumthroughout the bandwidth accepted by thereceiver ; the spectrum therefore has thesame frequency distribution as that offluctuation noise, but there is the importantcondition that for each single impulse allcomponeits are in the same phase, ascompared with the random phase forfluctuation noise. It follows that the peaknoise voltage is proportional to the band-width, as compared with fluctuation noisewhich is proportional to the square root ofthe bandwidth. The special phase conditionalso increases the gain due to the " triangularnoise spectrum " and to pre-emphasis4. Forstrong signals the gain on impulsive noise isthen

[SIN],. = [S/N]A.M. + 20 logD + 6 + P'(17)

The " triangular noise spectrum " gain isincreased to 6 db. for impulsive noise, andthe pre -emphasis gain P' for i5 kc/s A.F.band and ioo is correction circuit is 9.5 db.if no adjustment to the programme level atthe transmitter is necessary, or about 6 db.in the practical case of high-fidelity broad-casting.

Since the sinusoidal components repre-senting an impulse are uniformly distributedthroughout the received frequency band,and are in a fixed phase relationship, theresultant instantaneous frequency is themid -frequency to which the receiver istuned. (Another way of looking at thepractical result is that the receiver containsa number of tuned circuits, and the effect ofan impulse will be to cause these circuits to" ring " at the frequency to which they aretuned.) The effect of a very strong impulseis then to shift the instantaneous frequencymomentarily to the centre of the band, whichshould coincide with the unmodulated fre-quency of the received carrier ; in otherwords, the impulse can momentarily wipeout the modulation of the received signal,but cannot generate any modulation of itsown. This is a very important feature offrequency modulation, since an impulse,even of much greater amplitude than thesignal, cannot produce an A.F. outputgreater than that due to the signal at theinstant at which it occurs. We have, how -

237

ever, assumed that the receiver is accuratelytuned to the desired signal, and that theselectivity characteristic of the receiver issymmetrical about the mid -frequency. Ifthe receiver is detuned, say by an amountx kc/s out of a total bandwidth of ± X kc/s,the impulse may shift the resultant frequencyby this amount plus the depth of modulation,so that the noise level can be related to thesignal level by'the following decibel equation :

[impulsiveof

noise] =' Lmodulated signal[Level of ioo per

F.M F.M.

+ 20 login [1/±ioo X (i8)

where M is the percentage modulation: Thuswith reasonable accuracy of tuning, theimpulsive noise level is below the signal levelin the audio -frequency output, regardless ofthe input signal/noise ratio, so that there isno threshold and F.M. gives a substantialelimination of impulsive noise even on veryweak signals.*

3.4. C.W. Interference.-If the differencebetween the carrier frequency of the desiredsignal and the frequency of the interferingsignal, say B kc/s, is audible, there would bea heterodyne note in A.M. reception. F.M.can then be said to give an improvement

equation, which isvalid for [SIN]> o db. :

zo logio(X/B) P'(19)

The term P', representing the effect of pre -emphasis, depends on both the value of Band the pre -emphasis characteristic, but canbe large. For example, with ioo ,us pre -emphasis and B = 15 kc/s, P' is about20 db., and with a deviation of + 75 kc/sX/B = 5, so that the total improvement is34 db. ; if B is only i kc/s, P' is negligiblebut X/B is 75, giving an improvement of37.5 db.

If B is a supersonic frequency there is noquestion of an improvement due to F.M.,because it could cause no interference in anA.M. receiver. In fact the modulation of the

* Normally, M < 1 oo and x X, so thatlog,o[ro-o xx is usually negative, and the right-hand side of (18) is therefore less than " Level oftoo per cent. modulated signal."

S20

300

40

50

60


desired signal is reduced according to theequation

1 -Level of Level of l[signalLsignal C.W.

- 20 log, 0 [ Va2/ (I7,2 + V 2)].. (2o)

where V and V3 are the amplitudes of theinterfering and desired carriers. If the signalis weaker than the interfering carrier, thenthe predominant effect is that shown byequation (20), both for B audible and super-sonic,, and the signal is rapidly lost as theinterference level rises. The presence of aninterfering carrier also shifts the mean orcarrier frequency of the -resultant ofthe signal and interference, and sois equivalent to detuning the receiverby an amount

v2/ T7,2 + V.2) (2r)

This will increase impulsive noise (ifthe receiver is left tuned to the ori-ginal frequency of the signal) as

May, 1943

where + X is, the frequency change corre-sponding to ioo per cent. modulation of thedesired signal. (It is assumed that [SIN],includes a factor for the percentage modula-tion of the two signals, as well as thecarrier ratio.) For comparison it should benoted that in an A.M. receiver having rectifierdiscrimination (i.e. using linear detectionwhich causes " apparent demodulation of aweak signal by a strong one") we should have

[S//V]Ay. = 2[S/N]it.p.+ 6 (25)

When the interfering signal is amplitudemodulated, therefore, the chief advantage ofusing F.M. communication is in the reduction

0

indicated by adding x' to x in equa- 8tion (Z8).

3.5. Modulated Interference.-The 7-10

response to modulated interferencecarr be a complex problem if theselectivity of the receiver is suchas to cut some of the sidebands of theinterfering signal.6 But if all sidebands ofthe interfering signal are received approxi-mately equally, the undesired frequencymodulation imposed on the desired carrierby the modulated interference can be deducedfrom equation (r3) ; taking A cos wit as thedesired signal and B cos (02t as.the interferingsignal, one represents the effect of amplitudemodulated interference by writing

B24 k sin nt)2}W 0 = + (W2- W1) (A2 + B2 (i + k sin nt)2

. . . . (22)

and frequency -modulated interference bycoo=1'+'[(io2 Sw sin nt)- (01] B2I(A2 + B2)

. . (23)

It can then be shown that for amplitudemodulated interference With B<<A, the A.F.signal/noise ratio is related to the R.F.signal/interference level by the equation

[S IN], 4-- 2[S !N]},+ 20 lOgi 0 [X/ (c02 -to)] . . (24)

NNN-10 0 10 20 30

LEVEL OF INTERFERING CARRIER IN db ABOVE MODULATED CARRIER

Fig. 2.

of heterodyne interference, in accordancewith equation (r9), rather than of interferingmodulation. From equation (23) it can beseen that the variation in the receivedfrequency co, due to the frequency modula-tion aw sin nt of the interfering signal is

B2 Sco sin ntA(WD) = ± B2 . . (26)

If the- wanted and interfering signals bothwork on the same amplitude of frequencyswing, and [SIN]R.F. includes a factor equalto the depth of modulation of the interferingsignal, then for B < A,

[S/N]A.F. -* 2 [S/N]R.r. .. (27)

The threshold of improvement occurs at[S/N]carr, = o db., and below this levelthe interference suppresses the signal. It iscommonly stated that a difference in levelof 6 db. is sufficient to give freedom frominterference, but it is doubtful whether thiscould be called a " high-fidelity " standardof interference, and lo db. seems morereasonable ; the relation between [SIN],".

May, 1943 WIRELESS 239ENGINEER

and [SIN],,,, from equation (26), is plottedin Fig. 2.

3,6. Cross Modulation.-If a frequencymodulated wave is represented by a carrierand series of sidebands, and is then, con-sidered to be applied simultaneously withanother carrier frequency to a non-lineardevice, it will be found that the outputcontains terms whose frequencies correspondto sidebands of the originally unmodulatedcarrier, spaced about it in the same way asthe sidebands of the original F.M. signal, i.e.cross modulation has occurred. But if nowthe carrier which was originally supposed tobe unmodulated, is frequency modulated, weshall have the equivalent of two frequencymodulated signals on a common carrier, andit has been shown above that a difference ofao db. is sufficient to eliminate the weaker.It is therefore only in exceptional circum-stances that cross modulation can be ob-served in a F.M. receiver, since the crossmodulation is usually considerably belowthe level of the direct signal.

3.7. Transmitter gain.-Any gain in fieldstrength from the transmitter will result ina better signal/noise ratio at a given receivingsite, and a transmitter with a given size ofoutput valve or power consumption can givea better field strength with F.M. than withA.M. In an amplitude modulated trans-mitter the peak power output at 10o per cent.modulation rises to four times the carrierpower, so that if the limitation is one of peakpower, the change over to constant ampli-tude F.M. instead of A.M. will allow thecarrier power to be raised four times or 12 db.,but if the limitation is of R.M.S. power, thepermissible increase of power is only 1.5 times,or 3.6 db. This is physically represented bythe fact that in an A.M. transmission (single -tone modulation) the carrier amplitude istwice the sideband amplitude at loo percent. modulation, whereas with F.M. thecarrier amplitude is almost always less thanthe sum of the sideband amplitudes, andvanishes for certain ratios of deviation tofrequency of modulation. There is thusless waste of energy- in the carrier with F.M.than with A.M.

4. Propagation Characteristics and Appli-cations.There are certain limitations on the carrier

frequencies that can be used for wideband

F.M., and this naturally influences the typeof service to which it can be readily applied.If long-distance communication is attempted,it will usually be found that the signal travelsby two or more alternative paths, havingvariable differences in transmission time,which results in " selective fading," i.e.some of the sidebands decrease or vanish atthe same time that others increase ; andF.M. is more severely distorted than A.M. byselective fading 8. 9. This effect may beimagined qualitatively in terms of the largenumber of sidebands required in F.M. torepresent a single -tone 'modulation ; thereis more chance of some of these numeroussidebands being displaced than if there wereonly two sidebands. This suggests, as isverified by mathematical analysis and experi-ment, that the distortion is more severe formodulation frequencies which are smallcompared with the frequency sweep, andtherefore have their energy spread over manysidebands.

The first reaction to this difficulty was towork at carrier frequencks above 4o Mc/s,where propagation is (usually) by direct rayonly. This is quite convenient for broad-casting, and in U.S.A. a substantial numberof commercial F.M. broadcasting stationshad been put into operation in the band from43 to 5o Mc/s by the end of 1941 ; furtherdevelopment of broadcasting was thenstopped by the war, but it had been decidedto make F.M. standard for the sound accom-paniment of television in U.S.A. Thefrequency swing is ± 75 kc/s, and the rangeof modulation frequencies transmittedextends to 15 kc/s, since it is a " high-fidelity " service (see Appendix II). Thenominal service areas of these transmittersare about 7,000 square miles (say 45 milesradius), but some of the transmitters aresituated on mountains, giving an effectiveaerial elevation of the order of 1,200 feet,and good reception at a distance of 200 milesis frequently reported.

F.M. communication on slightly highercarrier frequencies is being used in U.S.A.for mobile communications, e.g. with policecars and repair parties patrolling longelectric power lines. One of the advantagesof F.M, for these applications is the reductionof impulse interference from the ignitionsystem of the vehicle in which the receiveris installed, and (in cars used in urban areas)


from the ignition systems of other traffic.It seems that in urban areas the inter-ference, both from ignition and from electricalmachinery, is predominantly impulsive ; forF.M. has been reported to be superior toA.M. for mobile work even at signal/noiseratios so low that it would be definitelyinferior if the noise were random noise. Itis usual to employ a 5 : i deviation ratio butto restrict the A.F. band to 3 kc/s for thesecommunication services, so that the channeloccupied is only ± 15 kc/s.

Proceeding to still higher frequencies,wide -band F.M. is used with highly direc-tional aerial systems for point-to-pointcommunication. One example is the linkbetween broadcasting studio and trans,mitter ; it is claimed that the F.M. radiolink on a carrier frequency of the order of300 Mc/s is cheaper, quieter, and has agreater A.F. bandwidth than a telephoneline. In this case the advantage of usingF.M. rather than A.M. is probably the reduc-tion of receiver noise (shot and thermal),since external noise is small at such fre-quencies but receiver noise is liable to behigh owing to the comparative inefficiency ofamplifying valves used at these carrierfrequencies.

Wideband F.M. is impracticable on carrierfrequencies much below 4o Mc/s for tworeasons : the distortion with multi -pathtransmission and the lack of sufficient spacein the frequency band to allow 200 kc/s perchannel. But a number of experiments havebeen carried out with F.M. using only arelatively narrow frequency sweep ; thelogical minimum value of sweep is that equalto the highest audio frequency, so that itoccupies approximately the same band widthas an A.M. transmission. Although a fairproportion of the noise -reducing .propertiesare sacrificed by using a narrow frequencysweep, there are still certain advantages :

(i) The noise reduction due to the " tri-angular noise spectrum " remains, andtogether with a suitable degree of pre -emphasis this can contribute a gain of aboutio db. in the ratio of signal to fluctuationnoise.

(ii) The gain on impulsive noise is some-what greater, and in addition there is theautomatic limiting action which prevents anynoise impulse from giving an output exceed -

May, 1943

ing the peak value obtained from a ioo percent. modulated signal.

(iii) There is a reduction of heterodyne andshared -channel interference compared withA.M. working.

The disadvantage to set against thesefeatures is the difficulty with selective fading.Experimental data have been published byCrosby8 for communication over a distanceof 1,15o miles with a carrier frequency of26.3 Mc/s and unity deviation ratio. Threedifferent propagation conditions were ob-served:

(i) On about half the days on whichsignals were received, communication wassatisfactory and free from the effects ofselective fading.

(ii) On one-fourth of the total days, con-ditions were such as to superimpose noise onthe signal.

(iii) On the remaining fourth of the totaldays, there -was selective fading sufficient togive severe distortion of the F.M. signal.

Satisfactory limiting of interference im-pulses due to static (atmospherics) was notedduring these experiments.

The technique of transmission and recep-tion in narrow -band systems tends to differfrom that of wideband frequency modula-tion. Owing to the narrowness of the phaseor frequency swing, the obvious method ofgeneration appears to be to vary the phaseof a constant frequency oscillator, and thismakes it possible to take advantage of thestable carrier generators available, such ascrystal oscillators. In the receiver, theproblem is no longer linearity of the dis-criminator characteristic, but rather thesecuring of sufficient steepness of the ampli-tude/frequency characteristic ; crystal filtercircuits may be used.

AcknowledgmentsThe author wishes to express his thanks

to Mr. L. H. Bedford and to Mr. K. I. Jonesof the Cossor Research Laboratories for theirinterest and encouragement which madepossible this work ; and to the managementof A. C. Cossor Ltd. for permission to publishit.

REFERENCES" Carrier and Side -Frequency Relations with Multi -tone

Frequency or Phase Modulation." R.C.A. Review, Vol. 3, p. 103.1938.

" Non -Linear Distortion, With particular reference to theTheory of Frequency -Modulated Waves." Part I. Phil. Mag.,Vol. 32, p. 265. 1941.

^ir


" N.B.C. Frequency Modulation Field Test." R.C.A. Review,Vol. 5, p. 190. 1040.

" Frequency Modulation Noise Characteristics." Proc. I.R.E.,Vol. 25, p. 472. 1937.

".The Distribution of Amplitude with Time in FluctuationNoise." Proc. I.R.E., Vol. 29, p. 50. 1941.

° " Interference in Relation to Amplitude, Phase and FrequencyModulated Systems." Wireless Engineer, Vol. 18, pp. 6 and 56.1942.

241

" A Method of Reducing Disturbances in Radio Signalling bya System of Frequency Modulation." Proc. Vol. 24, p. 689.1936.

" Observations of F.M. Propagation on 26 Mc/s." Proc.I.R.E., Vol. 29, p. 398. 1941.

° " Frequency Modulation and Distortion." Wireless Engineer,Vol. 7, p. 482. -

.°* " Amplitude, Frequency and Phase -Angle Modulation."Wireless Engineer, Vol. 17, p. 339. 1940.

APPENDIX IThe Sideband Representation of F.M. Waves

In reading various papers, one may find differentcombinations of sine and cosine terms for carrierand sidebands, some of which at first glance do notseem consistent with each other in their sideband

phase relationships ; the four fundamental casesare therefore set out here, and the vector diagramsfor the sidebands shown in the table given at thefoot of the page.

Case A.-Carrier = sin wt, phase modulation = x sin pt.The expression for this case is

E = sin (wt x sin pt) = sin wt cos (x sin pt) ± cos 04 sin (x sin pt)This can be expanded in terms of series of Bessel functions :

E= sin wt {J, (x) ± 2 J (x) COS 2 pt + 2 J (x) cos 4 pa . . .

± cos wt (2 J1 (x) sin pt 2 J3 (x) sin 3 pt . . . }

E = J (x) sin wt.

Z J,+, (x) [sin (w + 2n + ip)t - sin (w - 2n ip)t]

Z JE (x) [sin (w 2np)t + sin (w - 2np)t]I

}

CARRIERPHASE

PHASE -MODULATIONPHASE CARRIER TO SIDEBAND PHASE RELATIONSHIPS AT t=0

CASE A h +sC +5 -SM C 7,

ODD SIDEBANDS EVEN SIDEBANDS

CASE B.

C

+5+SV 7-11-..

1-S

(4n- i) thSIDEBANDS

C 45 -S40 th SIDEBANDS

C

(on+i)thSIDEBANDS

.4 - -4.-+S -S C

(4n-2)th SIDEBANDS

CASE C

+S -

CODD SIDEBANDS

t+St -S

ICEVEN SIDEBANDS.\°°-

CASE D

+S j t+ S

t -S on thSIDEBANDS

1i

IC

IC (on- 2)thI+5 SIDEBANDS

-S

kAC

(4 n-i)thSIDEBANDS

(4n- i)thSIDEBANDS


May, 1943

The three lines of the final expression for E represent carrier, odd sidebands and even sidebandsrespectively.

Case B.-Carrier -= sin cot, phase modulation = x cos pt.E = sin (cot + cos pt) can be developed by the same method to yield E = J0 (x) sin cot

CO

( - r)V2,4.1 (x) [COS (<a + 2n + ip)t + COS (a) - 212 + ip)t]

- j2n (x) [sin (co + 2np)t + sin (co - 2np)t]n

Case C.-Carrier = cos cot, phase modulation = x sin pt.E = cos (cot x sin pt) is equivalent toE = J0(x) cos ad

+ J2n-I-1(X) [cos (to + 212 + ip)t - cos (co - 212 + ip)t]n=o

J 27, (.0 [cos (a) + zip)t + cos (co - 2np)t]

Case D.-Carrier = cos cot, phase modulation = x cos pt.E = cos (cot x cos pt) is equivalent toF = J, (x) cos cot

- j(2+,) (x) [sin (co + zn ip)t ± sin (co - 2n ± ip)t]

c0

I)" J2n (x) [cos (a) + 2np)t + COS (a) -n =

APPENDIX II

Standards of F .11,1 . Broadcasting in the U.S.A.

In licensing commercial F.M. broadcastingstations in the U.S.A. the Federal CommunicationsCommission has made a number of rules coveringboth those standards which must be known beforereceivers can be constructed, and quantitativestandards to ensure that a satisfactory signal canbe received by a properly designed receiver. Thefollowing is the substance of the performancespecification with which stations must comply.

(i) Audio frequency characteristics.(i) Pre -emphasis : the transmitter shall have a

rising frequency characteristic, as produced by acircuit having a time constant of roo microseconds.

(ii) Frequency range: the frequency responseshall be correct within +2 db. from 5o to 15,000 c/s.

(iii) Harmonic distortion : not to exceed 2 percent. R.M.S. at +75 kc/s swing of carrier frequency.

(iv) Noise : spurious frequency modulation mustbe at least 6o db. below ion per cent. frequencymodulation ; amplitude modulation must be atleast 6o db. below carrier.

(2) Radio frequency characteristics.(i) Frequency stability : in the band of fre-

quencies from 42 to 50 Mc/s, the carrier must remainon its allotted frequency to within +2 kc/s.

(ii) Polarisation : F.C.C. allows either horizon-tally or vertically polarised radiation. (In practicehorizontal polarisation seems to be universal.)

(iii) Field strength : field strength is defined interms of the E.M.F. set up in a receiving aerialhaving an effective clearance from ground of 3o feet.

In the service area the E.M.F. in the receiving aerialmust be r,000 icV in urban areas or 5o icIr in ruralareas.

I.E.E. Wireless SectionTHE next meeting of the Section will be on

May 5th when H. J. Finden will deliver a paper on" The Frequency Synthesiser." The last meeting ofthe present session will be on May 11th when aninformal discussion on " Factors Determining theChoice of Carrier Frequencies for an ImprovedTelevision System " will be opened by B. J. Edwards.

"Shot -Effect in Space -Charge -LimitedDiodes "

THE author of the above article, which appearedin the March issue, has asked for the followingcorrection to be inserted.

Equation 17 should read

JP = J.P 2 .J2 + EV1 . 2

akT

I

Paper Goes to WarFivE times as much paper is needed for the

manufacture of various types of ammunition as is re-quired for any of the other purposes for which papermust be found. Waste paper is a weapon of war.


Wireless PatentsA Summary of Recently Accepted Specifications

243

The following abstracts are prepared, with the permission of the Controller of H.M. Stationery Office, fromSpecifications obtainable at the Patent Office, 25, Southampton Buildings, London, W .C.2, price IF each

ACOUSTICS AND AUDIO -FREQUENCY CIRCUITSAND APPARATUS

549 518. -Microphone attachment, say for anoxygen mask, compensated to prevent distortiondue to reflected sound -waves.

Standard Telephones and Cables and J. S. P.Roberton. Application date, 2nd May, 1941.549 701. -Magnetic diaphragm arranged as abalanced armature for telephone transmission andreception without the use of batteries. '

Standard Telephones and Cables and J. S. P.Roberton. Application date, 27th November, 2940.

549 926. -Device for continuously varying thelinear characteristic of an attenuation -equalisingnetwork from a positive to a negative slope.

Standard Telephones and Cables (assignees ofW. R. Lundry). Convention date (U.S.A.), 24thOctober, 2940.

AERIALS AND AERIAL SYSTEMS549- 564. -Impedance -matching device for a frameaerial coupled to a wireless receiver.

Philips Lamps (communicated by N. V. Philips'Gloeilampenfabrieken). Application date, 3rd Ottober,1941.

549 753. -Electromagnetic -wave horn for radiatingwaves with different planes of polarisation withoutmutual interference in duplex signalling systems.

Marconi's W. T. Co. (assignees of M. Katzin).Convention date (U.S.A.), 29th October, 1940.549 958. -Aerial screening arrangement for varyingthe sharpness with which a navigational course ismarked out by a pair of overlapping beams.

Aga -Baltic Radio Akt. Convention date (Sweden)Loth April, 2940.

RECEIVING CIRCUITS AND APPARATUS549 367. -Mounting and controlling the movableiron cores of high -frequency tuning coils.

Johnson Laboratories, Inc. (assignees of W. H.James). Convention date (U.S.A.), 5th June, 1940.549 459. -Construction of permeability -tuned coils,particularly for producing a constant difference "frequency in superheterodyne sets.

Marconi's W. T. Co. (assignees of W. F. Sandsand P. F. G. Hoist). Convention date (U.S.A.), 27thMay, 5940.549 465. -Receiving circuit of the homodyne typein which a local carrier frequency is utilised to giveincreased selectivity.

Hazeltine Corporation (assignees of N. P. Case).Convention date (U.S.A.), loth July, 1940.549 480. -Device, responsive to a calling -up signal,

for preparing the wireless receiver, say in a policecar, to respond to headquarters messages.

The General Electric Co. ; N. R. Bligh ; D. M.Heller ; and L. C. Stenning. Application date, 6thOctober, 2941.

549 484. -Automatic grid -biasing circuit, utilisinga copper -oxide detector, particularly for modula-tor valves.

Telefon Akt. L. M. Ericsson. Convention date,(Sweden), 26th October, 1938.

549 489. -Super -regenerative circuit including avelocity -modulated oscillator for receiving ultra -short waves.

Akt. Brown, Boverie et Cie. Convention date(Switzerland), 27th February, 2940.549 537. -Means for maintaining a constant outputover a wide frequency range from a heterodyne" mixing " circuit.

Standard Telephones and Cables (assignees ofT. . lSco 4czoe.wski). Convention date (U.S.A.), 3rd

549 769. -Receiver with a tuning system whichscans over a frequency range in which several short-wave transmitters are operating and automaticallymaintains contact with a selected one.

S.IY4o0urt. g. Convention date (U.S.A.), 29th Febru-ary,549 77o. -Method of using a superposed tonefrequency to facilitate the tuning of a televisionor like receiver to one or other of a number of differ-ent transmitters operating within a given frequencyband.

S. IYo0u.ng. Convention date (U.S.A.), 29th Febru-

834 Amplifier circuit in which a filter typeof feed -back circuit is combined with regenerativeand degenerative reaction to reduce noise.

Marconi's W. T. Co. (assignees of N. J. Oman).Convention date (U.S.A.), 29th July, 2940.

TRANSMITTING CIRCUITS AND APPARATUS549 644. -Inductive coupling between a tunedoscillator and an inductive load wherein the powertransferred is maintained constant over a range ofoperating frequencies.

Standard Telephones and Cables and E. C. Wil-loughby. Application date, 27th May, 1941.549 72L -High -frequency transmission line formedin two tandem parts which have different character-istic impedances so that one part may be the moreconveniently matched to a very high -frequencyload.

The General Electric Co. ; M. R. Gavin ; andV. A. Heathcote. Application date, 6th October, 2942.


549 979. -Apparatus for maintaining a constantrelation between the carrier frequency and thespeed of keying in a telegraph transmitter.

Marconi's W. T. Co. and H. J. Wassell. Applica-tion date, 13th June, 1941.

CONSTRUCTION OF ELECTRONIC -DISCHARGEDEVICES

549 301. -Construction of base or holder for valvesof the " acorn " type, to avoid straining the glassseals during insertion or withdrawal of the valve.

Standard Telephones and Cables ; M. M. Levy ;and A. S. Merritt. Application date, 12th May,1941

549 476. -Arrangement for preventing undesiredsecondary -emission effects in the focusing systemof a cathode-ray tube.

Philips Lamps (communicated by N. V. Philips'Gloeilampenfabrieken). Application date, 19thSeptember, 1941.

549 485. -Process for preparing or coating valveelectrodes made of steel in order to reduce secondaryemission and to increase their heat -radiatingefficiency.

Hygrade Sylvania Corporation. Convention date(U.S.A.), 8th December, 1939.

549 676. -Electron multiplier in which secondaryemission is utilised. for generating oscillations ofthe order of L000 to 5,000 kilocycles.

Standard Telephones and Cables and R. M.Barnard. Application date, 30th April, 1941.

549 778. -Low -emission alloy suitable for makinggrids for thermionic valves.

Marconi's W. T. Co. (assignees of E. G. Widell).Convention date (U.S.A.), 31st May, 194o.

549 795. -Construction and assembly of resonantelectrodes as used for " bunching ' or velocity -modulating an electron stream.

Standaid Telephones and Cables (assignees ofC. V. LittOn). Convention date (U.S.A.), 27thAugust, 194o.

549 976.-A frequency -selective amplifier valve inwhich the space charge developed by the inputoscillations applied to one grid induces phase -displaced oscillations on a second grid which, inturn, modifies the output taken from the valve.

Sir L. Sterling. Convention date (U.S.A.), 8thAugust, 1940.

SUBSIDIARY APPARATUS AND MATERIALS549 449 Process for making the powdered mag-netic material used for the cores of variable -perme-ability tuning coils.

Johnson Laboratories, Inc. (assignees of G. Berge).Convention date (U.S.A.), 21st June, 1940.

549 453. -Triggering circuits for the gas -filledvalves used in impulse signalling.

Standard Telephones and Cables ; F. H. Bray ;and L. R. Brown. Application date, 19th May, 1941.

549 499. -Stabilising arrangement for a valveoscillator, particularly when used for generating a

May, 1943

number of different frequencies for multiplex work-ing.

The General Electric Co. ; L. I. Farren ; and R. S.Rivlin.. Application date, 13th August, 1941.

549 594. -Cutting and mounting piezo-electriccrystals to ensure constant frequency and immunityfrom temperature variations.

Standard Telephones and Cables (assignees ofW. P. Mason). Convention date (U.S.A.), 20thSeptember, 1940.

549 623. -Means for coupling a carrier -wavesignalling circuit to a high -power transmission line.

Standard Telephones and Cables (assignees of K. S.Johnson and L. K. Swart). Convention date (U.S.A.),5th October, 1940.549 674. -Inductive " leader gear " system forautomatically steering a road vehicle or othermobile body over a predetermined course, particu-larly under black -out conditions.

C. L. Paulus and R. K. Stout. Convention dates(U.S.A.), 12th April and 19th April, 194o.549 835. -Safety device for minimising the effectof accidental contact with a source of high voltage,such as the supply -leads to a cathode-ray. tube.

E. G. Gage. Application date, 2nd July, 1941.549 88,. -Capacitance device for measuring elec-trically the mechanical displacements due to hydro-static or like pressure.

Philips Lamps (communicated by N. V. Philips'Gloeilampenfabrieken). Application date, 6th June,1941.549 948. -Means for minimising the effect of phasedrift in a power -line signalling system utilising asingle, carrier wave.

Standard Telephones and Cables (assignees ofH. R. Moore). Convention date (U.S.A.), 22ndOctobet, 194o.549 952. -Composition of electrolyte suitable foruse in the dry type of electrolytic condenser.

Standard Telephones and Cables (assignees ofK. G. Compton). Convention date (U.S.A.), 25thMarch, 1941.

The IndustryIn a recent publication (No. Cioz-A) Muirhead

and Co., Ltd., Elmers End, Beckenham, Kent, havecollected together technical information and dataon all the types of resistances and resistance net-works which they can manufacture. These includethe " Munit " series of decade resistances, a numberof precision low resistance slide wire units, andscreened attenuators for A.F. and R.F., the latter

for figures up to zo Mc/s.

The use of latex sleeves for identifying cable endshave many advantages which are set out in a de-scriptive leaflet issued by E. Siegrist, Ltd., 39 BernersStreet, London, W.1.

The Minister of Supply has appointed R. L.Frain, to be Controller of Quartz Crystals, to whomall communications relating to the supply of quartzcrystals should be addressed, at Portland House,Tothill Street, London, S.W.f . Telephone : Abbey7788.


Abstracts and ReferencesCompiled by the Radio Research Board and reproduced by arrangement with the Department

of Scientific and Industrial Research

For the information of new readers it is pointed out that the length of an abstract is notnecessarily an indication of the importance attached to the work concerned. An important paperin English, in a journal likely to be readily accessible, may be dealt with by a square -bracketedaddition to the title, while a paper of similar importance in German or Russian may be given along abstract. In addition to these factors of difficulty of language and accessibility, the natureof the work has, of course, a great influence on the useful length of its abstract.

PAGE PAGE

Propagation of Waves ... 245 Directional Wireless . 253

Atmospherics and Atmospheric Acoustics and Audio -Frequencies... 254

Electricity 248 Phototelegraphy and Television ... 256Properties of Circuits 249 Measurements and Standards ... 257Transmission 251 Subsidiary Apparatus and Materials 259Reception ... 252 Stations, Design and Operation ... 262Aerials and Aerial Systems 252 General Physical Articles ... 262

Valves and Thermionies 252 Miscellaneous 262

PROPAGATION OF WAVES1361. THE RADIATION FIELD OF A VERTICAL

TRANSMITTING DIPOLE OVER STRATIFIEDGROUND. - J. Grosskopf. (Hochf:tech. u.Elek:akus., Nov. 1942, Vol. 6o, No. 5,pp. 136-141.)

" In a series of previous papers (1833 of 1941and 376 & 964 of 1942) it was shown experimentallyand theoretically that in the propagation of electro-magnetic waves it is practically never possibleto reckon on homogeneous ground, and that onthe contrary the ground stratification has a decisiveinfluence. It was shown that the measurementof the Zenneck rotating -field ellipse described bythe electric vector, by means of the dipole measur-ing method, enables a and e to be determined forhomogeneous ground, but that this method appliedto stratified ground yields values for a and c whichhave no direct physical significance and whichare therefore termed the ' effective ' conductivityow and the ' effective ' dielectric constant c,,ff.They bear, obvipusly, a relation to the groundconstants of the individual layers and the thick-nesses of those layers. It was further shown thatfrom the measurement of the frequency character-istics of aot and cof it was possible to obtainknowledge of the ground constants and thick-nesses of the strata.

" The question now arises as to the manner inwhich the ground stratification influences thepropagation attenuation of the field radiated froma transmitter. It is known from the researchesof Sommerfeld, Weyl, and van der Pol & Niessenthat over homogeneous ground the propagation isdetermined by the Sommerfeld function F(w),whose argument [eqn.t] . . . is termed the' numerical distance '. The same attenuation is

245

obtained for equal numerical distances, whateverthe individual values of r, z, e, and a may be.Is there, then, a similar general function applicableto stratified ground, and how is the numericaldistance calculated in this case ? It will beshown that for stratified ground also the propaga-tion is governed by the Sommerfeld functionF(w) and that the corresponding numerical dis-tance w* stands in the closest relation to rye,/and ceff, and therefore to the rotating field presentover stratified ground. Thus the measurementof the rotating -field ellipse by the dipole methodattains an immediate importance also for themeasurement of the attenuation ".

Dealing first with propagation over unstratifiedground, and using Weyl's (1919) method, thewriter arrives at eqn.i9 for F, the propagationlaw of the Hertzian vector already obtained bySommerfeld and others by various ways andapplicable in that form only at the ground andfor small angles of elevation. He then obtainsthe modified eqn. 23 for F', applicable to an arbi-trary angle of elevation (van der Pol & Niessen) :the new numerical distance w' differs from thatfor small angles of elevation only by the replace-ment of the distance r measured at the groundby the distance of the measuring point throughspace, R =V r2 e. For practical applicationseqn. 25 for F', involving the Fresnel reflectioncoefficient for plane waves, is preferable. Passingthen to propagation over stratified ground, thewriter obtains as a first approximation eqn.52for F'*, which differs from eqn.23 only by the factthat in the expression for the numerical distancew'* the complex refraction coefficient of the toplayer, n1, is replaced by the effective refractioncoefficient n1* = n,/tan (8, - j(k) of the stratifiedground : for the significance of tan (S, -j si.) see


eqns. 33/35 : the German characters represent thehyperbolic tangent.

A full discussion of this effective refractioncoefficient and of its determination from thetangent relief diagram has already been given ina previous paper on the propagation of planeground waves over stratified ground (no specialreference is quoted but for previous papers involvingpropagation over stratified ground see 3001 of1942 and the beginning of the present abstract).It was there found that the Zenneck rotating -fieldellipse over stratified ground could be derived fromthat over homogeneous ground by multiplyingthe complex refraction exponent n = j .2alfby the correction factor I/tan (8, - j0). Thiscalculation, derived there for the case of plknewaves, is seen from the above considerations tobe applicable not merely to plane wave fronts(i.e. for large distances from the transmitter)but quite generally. The dependence of the abovecorrection factor on the layer thicknesses, thefrequency, and the ground constants of the indivi-dual layers, is discussed in the paper in questionwith the help of the tangent relief diagram.Whereas on that occasion the course and know-ledge of the function tan (81 - j0) was decisivefor the shape of the rotating -field ellipse, and gavethe possibility of working backwards to obtaininformatidn as to the structure of the stratifiedground by measuring the frequency characteristicof the rotating -field parameters, in the presentcase the function is of importance for the numericaldistance and thus for the attenuation of propaga-tion. The simultaneous appearance of the func-tion in the rotating field of the electric vectorand in the numerical distance makes possible thedirect application of the effective ground con-stants, determined by the dipole method from therotating -field parameters, to the calculation ofthe numerical distance and the attenuation :thus the .dipole procedure for measuring groundconstants is given a further field of applicationand importance.

The tangent relief diagram yields some veryinteresting deductions regarding the influence ofstratification on the numerical distance and thuson the range of the ground wave. Fig. 3, takenfrom the paper in question, shows the tangentrelief diagram with some inserted frequency curvesof the argument 81 - j0. According to whetherthe upper or the lower layer is the better conduct-ing, and to the values of frequency and layerthickness, the absolute value of tan (81 j0) isgreater or less than unity, and the numericaldistance greater or less than that over unstratifiedground. It may happen that a poorly -conductinglayer lying under a well -conducting layer willbe favourable to propagation, and that propagationover a badly conducting ground may be made stillworse by the presence of a well -conducting under -layer. The physical basis for this remarkablebehaviour is to be found in the formation ofreflections at the layer boundaries which, like thereflections at the ionosphere, produce additionalcontributions to the radiation : this interferenceeffect is also evident in the periodic characteristicof the factor tan (SI = j95) with respect to fre-quency and layer thickness.

May, 1943

In the determination of field -strength values forhigh angles of elevation the Hertzian potentialtakes, as has been seen above, the form of eqn. 25(F' = + 91,) (1 - F]), directly involvingthe reflection coefficient for plane waves. But aseqn.36 shows, in the first approximation here usedthe reflection coefficient for stratified ground isalso obtained from that for homogeneous groundby multiplying the complex refractive index ofthe upper layer by the correction factor I /tan (81jqS), as must be the case also from general con-siderations. K. A. Norton has built up a verycomprehensive and useful collection of the expres-sions for all the field components of an arbitrarilypolarised radiator over ground of finite conduc-tivity, from potentials of the form of eqn. 25(14 of 1937 and 32/33 of 1938). " Since his deriva-tions are valid only for homogeneous unstratifiedground, it is .fortunate that the simple correctionof the refractive index just described allows thedirect application of the Norton formulae to strati-fied grounds, which are what must be reckoned within most practical cases of propagation."1362. CORRECTION TO THE PAPER " THE QUESTION

OF PARTIAL REFLECTION AND THE CALCULA-TION OF THE APPARENT HEIGHT OF IONO-SPHERIC LAYERS 13035 of 1939].-K. Rawer.(flochttech. u. Elek:akus., Nov. 1942, Vol.6o, No. 5, p. 141.) Shortened version ofthe correction dealt with in 987 of April.

1363. A SIMPLE METHOD OF DEMONSTRATING THECIRCULAR POLARIZATION OF IONOSPHERIC -ALLY REFLECTED RADIO WAVES [byProduction of Beat Envelopes, takingAdvantage of Early Morning or EveningDoppler Effect to use Ground Wave asHeterodyne].-E. V. Appleton. (Nature,27th Feb. 1943, Vol. 151, No. 3826, p. 25o.)

1364. FURTHER NOTES ON THE ELECTRON DENSITYDISTRIBUTION OF THE UPPER ATMOSPHERE.-0. E. H. Rydbeck. (Phil. Mag., Feb.1943, Vol. 34, No. 229, pp. 130-139.)Calculation of true electron density in theF layer from virtual height data. Forprevious work see 23 of 1941.

1365. ON THE SIGNAL MAXIMUM DURING THESUNSET PERIOD.-S. K. Khastgir & S. N.Mazumdar. (Sci. & Culture [Calcutta],Nov. 1942, Vol. 8, No. 5, pp. 235-236.)

If the explanation of the sunrise and sunsetmaxima in number and field strength of distantatmospherics proposed in another letter in thesame number of this journal (see 1379, below) iscorrect, similar effects in the signal -strengthobservations of a distant transmitter at the timeof ground sunrise and sunset at the receiving sitewould be expected. Accordingly the Calcuttashort-wave station (484o kc/s, A = 61.98 m) waschosen and continuous observations of the signalcurrent in the anode circuit of the detector valveof a three -valve receiver were taken for some daysduring the evening hours, well covering the sunsetperiod. Typical curves for three different daysshow galvanometer deflections due to the signalat short -time intervals during the sunset period.

a


A distinct maximum is evident in each of the threecurves. The computed time is in good agreementwith the observed time of occurrence of the signalmaximum after the ground sunset.

1366. ON THE ATMOSPHERIC ABSORPTION IN THEULTRA -VIOLET [in Connection with theDetermination of the Optical Thicknessof the Atmospheric Ozone in Various Partsof the Spectrum].-A. Vassy. (Ann. dePhysique, July/Sept. 1941, VOL 16, p. 145onwards.)

For recent Comptes Rendus Notes see 633 & 64oof 1942. In the present very long report thewriter investigates the course of the relativeabsorption - coefficients by spectrographic - photo-metric methods, taking as her basis the values ofabsorption in the Huggins bands (300o-3300 AU)given by Ny Tsi Ze. The ozone absorption -coefficients between 4380 and 7585 and between2020 and 217o AU are determined, and the valuesof absorption in the ground layers of the airmeasured between 1898 and 4260 AU. Regionsof transparency and absorption bands of ozoneand oxygen are obtained: the transparency valuesare higher than those given by Graz and Maier -Leibnitz. The paper ends with a calculation ofthe atmospheric transparency for extra -terrestrialsources, and of the penetration depth of sunlightof wavelengths below 3000 AU.

-1367. ATMOSPHERIC ABSORPTION AND THE A-.LAW [Rejection, of Duclaux' Conclusion,from Muller & Kron's Teneriffe Measure-ments, that Rayleigh's Dispersion Law isreplaced by a A-3 Law : A-4 Law confirmedby These Same Measurements and bySmithsonian Institution Results].-J. Dufay.(Journ. de Phys. et le Radium, Series 8,No. 7, Vol. 1, 194o, p. 251 onwards.)

1368. PRELIMINARY RESULTS OF OBSERVATIONSMADE DURING THE TOTAL ECLIPSE OF THESUN ON SEPTEMBER 2ISt, 1941 .- B. G.Fessenkov. (Journ. of Phys. [of USSR],No. 1/2, Vol. 6, 1942, pp. 1-5 : in English.)

" Contrary to expectations it was found, how-ever, that the aurora circles . . . were very weak. . . This weak development . . . is a characteris-tic peculiarity of the 1941 eclipse and needs inter-preting." Photographs make it clear " that thechromosphere in the equatorial regions of the sunextends to great heights and is much brighter thanat the poles ". The 1941 corona shows a " lackof symmetry between the two sides of the equator,and in this it differs sharply from the typical lookof the corona when the sunspots are at a minimum ".On 16th September an " unusual group of spotspassed through the centre of the sun's disc. On18th September at 6.3o, when night had alreadylong fallen at Alma-Ata, bright aurora polarissuddenly broke across the sky . . . At 6.4o thephenomenon had almost disappeared. At about9 thefe was a renewal of the phenomenon but lessintense. We noticed a third flare-up at aboutThe next day the usual morning radio broadcaston the short wave was completely upset. It isinteresting to notice that this strong disturbance,

247

obviously connected with the above -mentionedgroup of sunspots, influenced the luminosity ofthe night sky for several days ", increasing itfrom the normal value between I and 1.5 milli-stilbs to 2.3 millistilbs immediately after theionospheric disturbance and then during thefollowing nights letting it drop slowly to its normalvalue by about 25th September. It is difficultto say definitely whether this sunspot group alsoinfluenced the corona, but " it is of interest thatthe chaotic disturbances in the equatorial regionsof the corona were on the same side of the equatoras this group of spots, which at the moment ofthe eclipse almost reached the edge of the disc . . .".Very good corona photographs were taken, andthe Abas-Tuman expedition obtained excellentmaterial for determining the degree of polarisationin' different spectral rays : radiometric measure-ments showed (in agreement with Nikonov) that" in the infra -red rays in the region about 1-2there is a noticeable excess of radiation ".

1369. THE SYSTEMATIC MOVEMENT OF SUNSPOTSIN HELIOGRAPHIC LATITUDES [StatisticalInvestigation of Greenwich PhotographicResults, Recurring Groups only].-J. Tuo-minen. (Physik. Berichte, ist Sept. 1942,Vol. 23, No. 17, p. 1676: summary ofFinnish paper.)

Leading to the " law " that sunspots have ageneral flow, those between the parallelsmoving towards the equator, while those outsidethis region move towards the poles: The speedof these last seems to be the greater, the further theyare from the equator.

1370. ON THE NATURE OF THE FACULAE ON THESOLAR DISC : II-PHOTOMETRY OF THEREGIONS OF FACULAE.-P. ten Biuggencate.(Physih. Berichte, 1St Aug. 1942, Vol. 23,No. 15, p. 1531 : summary from Zeitschr.f. Astrophys., No. 3, Vol. 21, 1942, p. 162onwards.) For previous work see 998 ofApril.

1371. TENTATIVE THEORY OF SOLAR PROMINENCES.-H. Aliven. (Arkiv for Mat., Astron., ochFysik [Stockholm], No. 2o, Vol. 27A, 1941.)

For other recent work by the same writer see3501 & 3504 of 1942 and 13 of January. In thepresent paper he.explains the solar prominences onthe hypothesis that the sun is to be regarded as agood electrical conductor up to a definite limitinglayer, beyond which it must be looked upon as avacuum where only free electrons or ions with longfree paths can transport energy. In that case,

. electrical potential differences can be produced onlythrough movements in a magnetic field. In theouter region an electric current, in the presence ofmagnetic fields, can flow practically only along themagnetic lines of force. For an eddy motion abouta magnetic pole (i.e. a sunspot) such a mechanism,on plausible assumptions of values, would yieldon the sun's surface potential differences of the orderof io7 volts, which would -draw after themselves,in the outer space, discharges along the magneticlines of force. The calculated field lines show asimilarity to the path curves of the prominences.


1372. RELATION OF THE COSMIC RADIATION TOGEOMAGNETIC AND HELIOPHYSICAL ACTIVI-TIES.-J. W. Broxon. (Phys. Review,istl15th Dec. 1942, Vol. 62, No. 11/12,pp. 508-522.)

1373. ON THE THEORY OF COSMIC -RAY SHOWERS[Further Contributions to the FluctuationProblem].-W. T. Scott & G. E. Uhlen-beck. (Plays. Review, istli5th Dec. 1942,Vol. 62, No. 1142, pp. 497-508.)

AN EXCEPTIONAL INCREASE OF COSMICRAYS ? [in mid -August, 1942].-A. Duperier.(Nature, 13th March 1943, Vol. 151 , pp. 308and 309.) For previous work see 2279 of1942.

1375. PROPAGATION METHODS OF RADIO PROSPECT-ING : B-ABOVE-GROUND METHODS [in-cluding Those of Grosskopf & Vogt and ofBurstyn].-V. Fritsch. (Arch. f Tech.Messen, June 1941, Part 120, V65-21,_Sheets T8o-81.)

1376. ATMOSPHERIC VERTICAL MOVEMENTS.-M.Robitzsch. (Meteorol. Zeitschr., No. 2,VOL 59, 1942, p. 52 onwards.) For a longsummary see Physik. Berichte, 15th Aug.1942, Vol. 23, No. 16, pp. 1604-1605.

DETERMINATION OF THE NUMBER OF DROPSPER UNIT VOLUME AND THE RELATIVEHUMIPITY IN CLOUDS [95% in Stable Clouds,sonfetimes sinking to 75% in UnstableClouds, but reaching Saturation in Cloudswith Big Drops of Radius greater than 84-J. Bricard. - (Comptes- Rendus [Paris],No. 9, VOL 214, 1942, p. 439 onwards.)For. a summary see Physik. Berichte, 15thAug. 1942, Vol. 23, No. 16, p. 1608.

1378. " EINFTIYHRUNG IN DIE GROSSWETTERFOR-scHuNG " [Second & Improved Edition :Book Review].-F. Baur. (Zeitschr. f. tech.Phys., No. 9, Vol. 23, 1942, p. 243.) Forreference to Baur's methods see 1897 of1942.

1374.

1377.

ATMOSPHERICS AND ATMOSPHERICELECTRICITY

1379. ON THE OBSERVED MAXIMA IN NUMBER ANDFIELD STRENGTH OF DISTANT ATMOSPHERICSDURING SUNRISE AND SUNSET PERIODS.-S. R. Khastgir. (Sci. csy Culture [Calcutta],Nov. 1942, Vol. 8, No. 5, pp. 233-235.)

A general explanation is given of the observedmaxima in number and field strength of the distantatmospherics before the ground sunrise and afterthe ground sunset in terms of the changes in theionospheric conditions during the transition timefrom day to night or vice versa. " It is now wellknown that the E -layer ionisation gradually fallsduring the night and attains a minimum value in thesmall hours of the morning. The ionisation thenbegins to increase and the increase is rather rapidduring the sunrise period. Before the ionisationminimum is attained, it is evident there is a gradualdecrease in attenuation due to gradually diminishing

May, 1943

collision frequency in the layer, causing thereby agradual increase in the intensity of the down -comingwaves, until there is a peak when the E -layerionisation is minimum. After the ionisation mini-mum, the ionisation rapidly increases during thesunrise period. This would cause a large fall in theintensity of the down -coming waves due to twocauses, viz., (1) higher attenuation due to largercollision frequency in the layer, and (2) larger devia-tion of the rays due to higher electron density. Amaximum is therefore expected during the sunriseperiod. Before the sunset time, too, it is knownthat there is a gradual decrease in the E -layerionisation. Thus there would be a gradual rise inthe intensity of the down -coming waves due to acontinuous fall in ionospheric absorption or attenua-tion. At the sunset time when the ionising solarrays are suddenly withdrawn, there would be asudden and a perceptible decrease in electrondensity of the layer, so that the down -comingwaves originally coming from the distant source ofatmospherics would be much less deviated and wouldnecessarily fail to reach the receiving point. Atthis point of the withdrawal of the ionising solarrays, the intensity would therefore fall in spite ofsmall attenuation. Subsequently, however, thegradually decreasing attenuation would assertitself and the intensity would increase. Thisexplains the sunset maximum." See also 1365,above.1380. THE ACTION OF LIGHTNING CURRENTS OF

LONG DURATION ON LIGHTNING PROTECTORS.OF THE VARIABLE - RESISTANCE TYPE[Theoretical Treatment leading to Approxi-mate Formulae : Importance of ThermalCapacity : etc.].-S. Szpor. (Bull. Assoc.suisse des Elec., 13th Jan. 1943, Vol. 34,No. I, pp. 15-21 : in French.)

1381. " MECHANISM OF THE ELECTRIC SPARE "[Book Review].-L. B. Loeb & T. M. Meek.(Nature, 13th Feb. 1943, Vol. 151, No. 3824,p. 178.) Referred to in 968 of 1942.

1382. WAITING FOR LIGHTNING [Measurements OfCrest Value of Lightning Current enteringBuried Telephone Cables : Steepness ofWave Front : Quantity of Electricity inLightning Surge and Crest Voltage acrossInsulation of Buried Cable Conductors].-T. J. Mahoney. (Bell Lab. Record, Dec.1942, VOL 21, No. 4, pp. 86-9o.)

1383. LOCATION OF THUNDERSTORMS BY RADIOMETHODS [Atmospherics associated withLightning Strokes to Ground : LightningStrokes as Cause of Breakdown of OverheadPower Lines : Warning obtained of Ap-proaching Thunderstorms].-J. S. Forrest.(Nature, 6th March 1943, Vol. 151, pp. 285-286 : summary only.)

1384. METEOROLOGICAL PHENOMENA ON THEOCEAN, AND THE EARTH'S ELECTRIC FIELD[Estimate of Influence of " Waterfall "Electricity at Surface of Sea (due to Windsand especially Cyclones) on the Atmos-pheric -Electric Field : contributes to SpaceCharge of Air Layers next to Surface, & 46


to Ionisation : Possible Importance inMaintenance of Earth's Negative Charge].-G. Aliverti & G. Lovera. (Physik. Berichte,15th Aug. 1842 Vol. 23, No. 16, p. 1599.)

PROPERTIES OF CIRCUITS

1385. AMPLIFIER PROBLEMS [Investigation of NoiseLevel due to Thermal Motion of Electrons :Amplification of Wide Frequency Bandswith Resistance -Coupled & Carrier -Fre-quency Amplifiers].-E. Baldinger. (Bull.Assoc. suisse des Elec., 3rd June 1942,Vol. 33, No. I I, p. 310 onwards : in German.)

1386. " RADIO RECEIVER DESIGN : PART I-RADIO-FREQUENCY AMPLIFICATION AND DE-TECTION " [Book Review].-Sturley. (See1409.)

1387. AMPLIFICATION OF DIRECT VOLTAGES WITHA HIGH -FREQUENCY PUSH-PULL CIRCUIT[for the Measurement of Small Voltages].-F. ' Kerkhof. (Zeitschr. f. tech., Phys., No.1o, Vol. 23, 1942, pp. 267-269.)

In a previous paper (2982 of 1942) the writerdescribed a method of avoiding the various dis-advantages of resistance -coupled amplifiers for d.c.voltages, by conversion of the latter into audio -frequency voltages with the help of a push-pullmixing stage. The present note describes howa similar arrangement can be used with the additionof a special compensating method, for a h.f. carrier(actually moo kc/s) in place of the 310 c/s carrierpreviously employed. This a.f. carrier had theadvantage that the output indicating instrumentcould take the form of a loudspeaker or simple a.c.meter, without any need for rectification. Ithad also, however, the objection that any un-wanted harmonic components were very diffiCultto compensate. Filtering by h.f. resonant circuitsis much easier, and above all the use of higherfrequencies makes the harmonics fundamentallymore easy to suppress (as a simple considerationof the logarithmic decrement will show) providedthat the ratio of inductance to ohmic resistance inthe coil is maintained constant.

Circuits were developed for valves of the A seriesand of the E series : the one here illustrated anddiscussed uses two AK2 octodes and an AF7pentode. By using octodes or multiple valves inthe push-pull stage the need for a separate oscillatoris avoided : the oscillator is obtained by joiningthe grids G1 and G2 of the two pentodes : its fre-quency is about I000 kc/s, with an amplitude of8 v,if. The d.c. voltage to be measured is appliedbetween the two control grids G4 at E1 and E,.The resistances between these points can be of sucha value that (as an alternative) an input stage canbe formed from an electrometer valve T113 withgrid resistance up to low ohms : in this case theanode of. the T113 is connected to El, the space -charge grid to E2, and the necessary workingvoltage is provided for the electrometer valve byintroducing a larger cathode resistance in the push-pull stage. The anodes of the octodes are linkedby a push-pull resonance transformer with the samenumber of turns in the two primaries and the

249

secondary, all the windings being wound to haveas little capacitance as possible. The secondary isconnected to an AF7 amplifying stage to which acathode-ray oscillograph or a rectifying stage canbe connected through a simple resonance trans-former. All supplies can be derived from a stab-ilised mains unit. The d.c. voltages for the gridsG2, G3+5 of the two push-pull valves are providedby an ordinary voltage -dividing circuit with bridg-ing condensers : for the amplifying AF7 a " sliding '

screen -grid voltage UO2 is supplied, but the well-known circuits for this purpose are omitted fromthe diagram for simplicity's sake.

However carefully the circuit was designed, itwas found that with the desired compensation to(at the least) 5 x Io5 v certain balancing diffi-culties presented themselves in the mixing stage,due principally to capacitive retroaction of theh.f.-carrying anode leads onto the control grids GIof the oscillator system. A convenient compensa-tion was finally obtained by taking off a small h.f.voltage from the push-pull transformer by meansof an additional winding of 3 turns only, andapplying it to GI, its fine adjustment being carriedout by a yoltage-dividing circuit with a trimmercondenser T and a rotating condenser C, in serieswith the three turns. By this means the high -frequency bridge shown can be balanced muchmore easily than the previous a.f. type, down to aresidual voltage corresponding to an input voltage -difference of about 5 x ro-5 v. The amplificationof the two stages shown in Fig. I is then about2 X 104, given by the ratio of the effective h.f.voltage appearing at A (terminal of the secondaryof the output transformer of the amplifier AF7)to the voltage change between E1 and E2.To prevent the appearance of h.f. voltages whichwould upset the balance and which readily occurfrom unavoidable capacitive couplings betweencircuit components (carrying large h.f. voltages)and sensitive grids, care must be taken to buildthe apparatus very rigidly so that, for instance,capacities between leads are very constant. Onthis depends the constancy of the zero point in thepresence of vibration, and the difficulty is onedisadvantage of the apparatus in comparision withthe a.f. bridge previously described.

1388. COMPARISON OF VOLTAGE- AND CURRENT -FEEDBACK AMPLIFIERS.-E. H. Schulz. (Proc.LR.E., Jan. 1943, Vol. 31, No. I, pp. 25-28.)

Author's summary :-" This paper points out thedifferences between an amplifier with voltage feed-back and one with current feedback. The effect ofvariations of amplifier constants on output voltageand current is decreased by either type of negativefeedback. Voltage feedback decreases the effectof load impedance on output voltage, and currentfeedback decreases the effect of load impedanceon load current. Voltage feedback increasesthe damping of a loudspeaker and improves itsresponse. A table is also given to assist in thechoice -of type and amount of feedback to be usedin a given application."

1389. RESEARCH ON A DOUBLE -VALVE -RECTIFICA-TION SYSTEM : THE OPERATING CURRENT AS AFUNCTION OF THE TIME, MEAN VALUES OFCURRENT, AND THE WAVINESS [Analysis of

1


Circuit for Supply of Direct Current fromA.C. Mains, for Relays, etc.].-C. Weisglass.(Physik. Berichte, 15th Aug. 1942, Vol. 23,No. 16, pp. 1568-1569: summary of anIstanbul University paper.)

1390. IMPEDANCE OF SOME SIMPLE ELECTRICALCIRCUITS [with Reference Sheets to aid inAnalysis of More Complicated Circuits].-B. Dudley. (Electronics, Dec. 1942, Vol 13,No. 12, pp. 75-77.)

1391. THE MUTUAL INDUCTANCE OF COAXIALCIRCULAR RINGS IN THE PRESENCE OF APERMEABLE CORE . -H. Buchholz. (Zeitschr.f. tech. Phys., No. 9, VOL 23, 1942, pp.221-234.)

The mutual inductance of parallel and coaxialrings has hitherto been calculated always on theassumption that the rings were in a homogeneousmedium incapable of magnetic polarisation : thecoefficient of mutual inductance is then dependentonly on the radii of the rings and their spacing.But if a coaxial permeable core is present, themagnetic coupling of the rings is also a function ofthe core diameter and of the magnetic propertiesof the core material, and the question arises to whatextent the mutual inductance in such a case varieslinearly with the core permeability, and how thevarious dimensions of the arrangement influencethe deviations from the linear law. The analyticaltreatment of this problem 'is very difficult, at anyrate as regards the numerical working out, and itis advisable not to tackle it directly for the case ofa single pair of rings, but to approach it by wayof the much more convenient solution for a numberof current rings equally spaced along an infinitelylong permeable core of circular cross-section.Further, the mathematical treatment of the per-meability of magnetisable materials is complicatedas a rule by the variation of the permeability withfield strength, according to a law which is toocomplicated to adapt itself to calculation. Thesecomplications are avoided in the present workby the assumption that the core is of somemodern compressed -powder or other material forwhich the dependence of permeability on fieldstrength is greatly reduced.

Author's summary :-" The question investi-gated is the magnetic coupling of coaxial currentrings in the presence of a common permeable coreof circular cross section. The basic problem con-sists in the calculation of the magnetic field of asingle ring in the presence of such a core. Thisfundamental solution is then extended step bystep. It is first adapted to the case where theexciting current ring is a ribbon -shaped conductor,and the final case considered is that where a largenumber of equidistant ,current rings of quasi -linear or ribbon construction excite the field. Inthe limiting case of an infinite number of such rings,the field becomes a strictly periodic structure andat the same time lends itself most readily tocalculation. For the magnetic coupling betweensuch an infinite system of current rings and a singleexternal current ring a formula is obtained [eqn.4.6, p. 232] which is suitable for numerical calcula--tion; Among other things it shows under what

May, 1943

external conditions a large deviation from thelinear law of increase of coupling with permeabilitycan occur."

1392. REMARK ON THE INVESTIGATION BY H.STEYSKAL, BERLIN, ON " A SPECIAL TYPEOP CONCENTRIC LINE AS ULTRA -SHORT-WAVE RESONATOR," and A SUPPLEMENTTO THE ABOVE REMARK. -F. Borgnis :H. Steykal. (Zeitschr. f. tech. Phys., No. 9,Vol. 23, 1942, pp. 240-241 : p. 241.)

Steyskal's paper (2969 of 1942) deals with afinite length of concentric -tube conductor open atboth ends and provided with a radial conductingjoining -wall. The oscillation mode considered isone which shows no electric field in the axialdirection (magnetic type) and whose field com-ponents are independent of the z-coordinate. " Thepresent note gives some remarks on the oscillationmodes possible in such a resonator, which shouldbe of interest as an extension of Steyskal's work."The case, in point is similar to that of the staticmagnetic field of a cylindrical coil : for an infinitelylong coil the inside field may be considered asuniform and can be calculated easily, but for afinite length of coil the H lines are closed in openspace, and the distortion of the internal field ismore and more marked the shorter the coil.Similarly, in the case considered the H lffies will beclosed in the external space : this involves adefinite participation by the latter in the oscillationregime. Apart from the radiative damping, thiswill bring with it an alteration to the resonancewavelength. Steyskal's equations 20 & 22 showthat in certain cases the resonator can be regardedas a Lecher system, short-circuited at both endsand bent into a circle, with a mean length (r1 r2)

= m . A/2 : in this case the equation for theresonance wavelength is directly interpretable onphysical lines. In other cases however (smallvalues of //A and large values of p = r,/r1) largedeviations from the theoretical resonance wave-length must be expected. Investigation to find amode of oscillation possible for the short resonatorunder consideration, such that no serious participa-tion by the external space can occur, leads to theconclusion that only an electric type of wave canfulfil the Conditions : the natural wavelength ofthe fundamental is found to be

. A = 2/1/(Ii(P - 1)?1}2 (0)2,or, when 1 is large, A/2 (r2 - r1). This oscillationregime is the only one for the system in questionin which the oscillation is practically concentratedin the interior of the resonator : small edge distor-tions are naturally to be expected, of the sameorder of magnitude as in a concentric open n . A/2Lecher system."

Steyskal supplements this note by some .practicaldata regarding the magnitude of the effects dueto the participation by the external space pointedout by Borgnis. For four different experimentalmodels, with air filling for both the internal andexternal spaces, the departure zIA of the observedresonance wavelength from the values given byformulae 19 or 20 of Steyskal's paper was less than1% for 1/A = 0.36 and 0.32, and about 2% for//A = 0.074 and 0.064 (p being 1.5 in all cases).Hence if the ratio //A is not too small the effect of


the external space may be neglected withoutappreciable loss of accuracy.

THE NATURAL ELECTROMAGNETIC OSCILLA-TIONS OF CAVITIES.-M. j011gUet. (Rev.Gen. de l'Elec., June 1942, Vol. 51, No. 6,p. 318 onwards.)

RADIO DATA CHARTS : 5 [" Q " of Quarter -Wavelength Resonant Line]. -J. McG.Sowerby. (Wireless World, March 1943,Vol. 49, No. 3, pp: 72-74.)

1395. THE CAPACITY EFFECTS IN SOLENOIDS [andthe Calculation of the Self -Capacitance by aNew Formula].-A. C. Lopez. (Formationy Documentation professional [Madrid],July/Aug. 1942, Vol. 1, No. 4, pp. 404-408.)

Breit's simple formula G.= o.44R is not satis-factory in practice, and Palermo attributes this factto the neglect of two fundamental quantities, thewire diameter d and the spacing between turns, S.He obtains for the capacitance between two turnsthe expression Co = arD/3 .6 cosh-'Sld pF, and arguesthat as the coil contains N turns, the coil capaci-tance must be N times larger, but that as thepotential between turns is r/N of the total, the coilcapacitance is equal to that of one complete turn.Sacco (2119 of 1941) starts from the same hypothesisthat the only effect to be considered is that of thecapacitance between adjacent turns, but differsfrom Palermo by concluding that the capacitanceC, of the whole solenoid is (n - r)/n2 times thecapacitance Co of one turn.

The present writer's treatment, to decide whetherthe factor by which Palermo's single -turn expressionmust be multiplied to give the capacitance for thewhole coil is the unity of Palermo, the (n - r)/n2of Sacco, or some other value, leads to the resultthat for S/d = 2, for instance, the following valuesapply to this factor for coils of 2, 3, 5, and 7 turnsrespectively (Sacco's corresponding values are givenin brackets) :-0.25 (0.25), 0.30 (0.22), 0.33 (0.16),and 0.34 (0.12). Thus for a seven -turn coil the self -capacitance (for S/d = 2, as before) would be0.34 x D13.6 cosh-1.S/d = o.44R, thuS agreeing withBreit's simple formula in this particular case.Among the various references listed at the end ofhis paper is the Hamburger -Miller article dealt within 82o of 1941, where the failure of the Palermoformula as applied to thick -wire coils for ultra -highfrequencies is mentioned.

1396. THE INFLUENCE OF LOSSES ON THE PROPERTIESOF ELECTRICAL NETWORKS [Treatment, byTwo Methods, of Low -Pass Filters of Basic &Transformed Types and Band -Pass Filter ofBasic Type].-J. H. Schouten & J. W. Kliite.(Philips Tech. Rundschau, May 1942, Vol. 7,No. 5, p. 138 onwards.)

1397. PERFORMANCE CURVES FOR M -DERIVEDFILTERS.-W. J. Cunningham. (Journ.Applied Phvs., Dec. 1942, Vol. 13, No. 52,pp. 768-772.)

Author's summary :-" For a composite filtermade up of several m -derived sections, the greatestdiscrimination will be obtained if the values of thedesign parameter m for the various parts of thefilter are chosen so that successive minima of inser-

1393.

1394.

251

tion loss in the attenuation band are all the same.Many different low-pass structures were set up andadjusted experimentally to show this type of per-formance. The necessary values of m and theresulting discrimination for each network have beenplotted on curves, which give information about thesharpness of cut-off. These curves are useful forthe most economical design of low-pass, or high-pas's,composite filters to meet definite requirements."

1398. A BRIDGING FILTER FOR OPEN -WIRE LINES.-E. A. Schramm. (Bell Lab. Record, Dec.1942, Vol. 21, No. 4, pp. 93-96.)

FILTERS FOR CARRIER -FREQUENCY TELE-PHONE SYSTEMS [and the Spectral Distribu-tion of Speech Intensities, MicrophoneSensitivity, Distribution of Attenuationbetween the Various Filters, Practical FilterDesign, etc.].-T. J. Weijers. (Philips Tech.Rundschau, April 1942, Vol. 7, No. 4, p. 104onwards.)

1400. TRANSITRON OSCILLATORS [Wide Rang andHigh, Frequency Stability with UntappedCoils].-A. G. Chambers. (Wireless World,March 1943, Vol. 49, No. 3, pp. 86-87.) Forother papers on the " transitron " oscillatorsee, for example, 3557/8 of 1942.

1399.

1401. SOME USEFUL CIRCUITS EMPLOYING THYRA-TRONS AND IGNITRONS.-A. J. Maddock. (Journ. of Scient. Instr., March 1943, Vol. 20,No. 3, pp. 37-46.)

Circuits are described for several typical examplesof applications. These are :-(r) relays for amplifi-cation of power as in temperature control, countingof physical phenomena, etc ; (2) instantaneousswitches for control of power as in pulsecathode-ray tube time bases, etc ; (3) current andvoltage regulators ; (4) commutating devices.

TRANSMISSION

1402. THE " TRAFFIC COP " TRANSMITTER [Com-pact and Economical -zoo -Watt Transmitter].-P. J. Palmer. (OST, Jan. 1943, Vol. 27,No. 1, pp. 38-41 and 9o, 92.)

A FREQUENCY - MODULATED RESISTANCE -CAPACITANCE OSCILLATOR [AlternativeMethod to Armstrong & Reactance -TubeMethods, giving Wide Frequency Deviationdirectly].-C. K. Chang. (Proc. I.R.E.,Jan. 1943, Vol. 31, No. I, pp. 22-25.) Aresistance element in an R -C -tuned oscillatoris replaced by the output resistance of avariable -mu valve.

1403.

1404.

1405.

COUPLED RESONANT CIRCUITS FOR TRANS-MITTERS [Design of Interstage CouplingUnits].-N. I. Korman. (Proc.' I.R.E.,Jan. 1943, Vol. 31, No. I, pp. 28-35.)

AMPLITUDE MODULATION UP TO DATE[Improving the Efficiency of Low -PoweredR.T. Transmitters].-O. J. Russell. (Wire-less World, March 1943, Vol. 49, No. 3,pp. 64-67.)


1406. FREQUENCY MODULATION : III [InterferenceSuppression, the Limiter, and the CaptureEffect].-C. Tibbs. (Wireless World, March1943, Vol. 49, 1`:To. 3, pp. 82-85.)

1407. BACK -HEATING OF MAGNETRON FILAMENTBY IONS OR OUT,OF-PHASE ELECTRONSCOUNTERED BY CURRENT LIMITER, SUCH ASDISCHARGE TUBE WITH TUNGSTEN CATHODEOR SPACE -CHARGE GRID, IN ANODE D.C.SUPPLY CIRCUIT.-F. Hilister. (Hochf:tech.u. Elek:akus., Nov. 1942, Vol. 6o, No. 5,pp. 144-145.) Telefunken patent, D.R.P.72o 62o.

1408. FACTORS DETERMINING THE INTENSITY OFOSCILLATIONS IN THE PLASMA OF GASEOUSDISCHARGE.-A. A. Sluzkin & A. P. May-danov. (Journ. of Phys. [of USSR], 1942,Vol. 6, No. 1/2, pp. 7-14: in English.)

Authors' summary (I) The effect of theshape of electrodes upon the intensity of oscilla-tions occurring in the plasma of a gaseous dischargehas been studied. It is shown that the maximumintensity is obtained with a cylindrical anode anda cathode in the form of a straight-line filamentplaced along the axis of the anode. (2) The effectof the anode current, anode voltage, and pressureupon the intensity of the oscillations was studied.The existence of optimum waves giving maximumintensity of oscillations is shown. (3) The depen-dence of the optimum wavelength is shown to beproportional to the anode diameter. (4) Certainconsiderations concerning the possible mechanismof self -excitation of oscillations in the plasma aresuggested.

RECEPTION1409. " RADIO RECEIVER DESIGN : PART I-

RADIO - FREQUENCY AMPLIFICATION ANDDETECTION " [Book Review].-K. R. Sturley.(Electrician, 26th Feb. 1943, Vol. 130,No. 3378, p. 22o.) A detailed study ofthe theoretical and practical bases.

1410. A SIMPLE BAND -SPREAD METHOD for SHORT-WAVE RECEPTION [Use of Usual TuningCondenser with Added Series & ParallelFixed Condensers : Simple MechanicalConstruction, No. Additional Knob, Speci-ally Low Noise Level, & Absence of AcousticRetroaction].-C. J. van Loon. (PhilipsTech. Rundschau, No. 9, Vol. 6, 1941p. 269 onwards.)

1411. METHOD FOR THE DEMODULATION OF FRE-QUENCY -MODULATED OSCILLATIONS [Circuitembodying Push -Pull Connection of TwoGas- or Vapour -Filled Valves controlledby a " Commutating " Condenser whichExtinguishes One Valve when the OtherIgnites]. - G. R. Clark. (Hochf:tech. u.Elek:akus., Nov. 1942, Vol. 6o, No. 5,p. 146.) A.E.G. patent, D.R.P. 720 781.

1412. FREQUENCY MODULATION : III [InterferenceSuppression, the Limiter, and the CaptureEffect].. --C. Tibbs. (Wireless World, March1943, Vol. 49, No. 3, pp. 82-85.)

May, 1943

1413. CORRECTIONS TO " DIODE AS A FREQUENCYCHANGER " [1080 of April].-Colebrook &Aston. (Wireless Engineer, March 1943,Vol. 20, No. 234, p. 126.)

AERIALS AND AERIAL SYSTEMS

1414. THE RADIATION FIELD OF A VERTICALTRANSMITTING DIPOLE OVER STRATIFIEDGROUND.-Grosskopf. (See 1361.)

" RHOMBIC ANTENNA DESIGN " [Book Re-view].-A. E. Harper. (Proc. I.R.E., Jan.1943, Vol. 31, No. I, p. 43.)

1416. CORRECTIONS TO " THEORY OF ANTENNASOF ARBITRARY SIZE AND SHAPE."-S. A.Schelkunoff. (Proc. I.R.E., Jan. 1943, Vol.

"31, No. I, p. 38.) See 1049 of 1942.

1417. AIRCRAFT ANTENNA CHARACTERISTICS [WithCharts for Reactance & Radiation -Resis-tance Estimation (e.g. for Design of DummyAerial)].-P. J. Holmes. (Electronics, Dec.1942, Vol. 15, No. 12, pp. 46-48.)

1415.

VALVES AND THERMIONICS

1418. RADIO VALVES [Notice of the Issue of a" War Emergency British Standard Codeof Practice relating to the Use of RadioValves in Equipment "].-(Electrician, 26thFeb. 1943, Vol. 13o, No. 3378, p. 23o.)

1419. [ULTRA-] " HIGH -FREQUENCY THERMIONICTUBES " [Book Review].-A. F. Harvey.(Electrician, 26th Feb. 1943, Vol. 130,No. 3378, p. 22o.)

1420. THE STANDARDISATION OF BROADCAST RE-CEIVING VALVES [American & EuropeanClassifications]. - M. Adam. (Formaciony Documentacion profesional [Madrid], May/June 1942, Vol. r, No. 3, pp. 308-309:summary, from Genie Civil, 12thi19thApril 1941.)

1421. NEW TUBE TYPES [Characteristics of aCathode -Ray Tube, a Pentode, a TwinTriode, and a Full -Wave Rectifier].-(Radio [New York], Dec. 1942, No. 275, pp. 20-21.)

1422'. ULTRA -SHORT-WAVE TRANSMITTING VALVEWITH CENTRE -FED FILAMENT, GRID, ANDANODE EACH BUILT AS AN OPEN DIPOLE.-H. E. Hollmann. (Hochf:tech. u.Nov. 1942, Vol. 6o, No. 5, p. 144.) D.R.P.72o 303.

1423. ELECTRON -TUBE TERMINOLOGY [First Pub-lished Record of Word " Electron ".(Stoney,1891) : Its Use as an Adjective, and Occa-sions where " Electronic ' should be em-ployed : " Electronics " & Its Use : EarlyTube Names : Suggestions for ProperNaming, including the Westinghouse-Gen-eral Electric Agreement]. - W. C. White.(Electronics, Dec. 1942, Vol. 15, No. 12,pp. 42-45 and rm.)


1424. ULTRA -HIGH -FREQUENCY MAGNETRON OS-CILLATORS [Brief Account of Types ofMagnetrons].-E. M. Noll. (Radio [NewYork], Dec. 1942, No. 275, pp. 12-13.)

1425. PHYSICAL ASPECT OF THE PRINCIPAL LIMITSTO THE FUNCTIONING OF CONVENTIONALVALVES AT ULTRA -HIGH FREQUENCIES [13 -Page Survey].-R. Warnecke. (Rev. Gen.de April 1941, p. 231 onwards.)

1426. THORIUM DETECTOR [Which automaticallySorts Filament Wire to determine whetherIt is made of Pure Tungsten or containsThorium : Arc & Spectroscope Combination].- Westinghouse Company. (Electronics,Dec. 1942, Vol. 15, No. 12, p. 82.) Cf.786 of March.

1427. SHOT EFFECT IN SPACE -CHARGE -LIMITEDDIODES [Discussion of Space -Charge Reduc-tion of Shot Effect in Planar Diodes].-M. Surdin. (Wireless Engineer, March 1941VOL 20, No. 234, pp. 127-130.) For aletter on the effect in temperature -limiteddiodes see 242o of 1941.

1428. GRAPHICAL DETERMINATION OF POWER -AMPLIFIER PERFORMANCE [Complete Per-formance of Class B & C Power -Amplifiersobtained Quickly from Static Character-istics by use of Plastic Calculating Device].-R. I. Sarbacher. (Electronics, Dec. 1942,Vol. 15, No. 12, pp. 52-56 and 158.)

1429. SECONDARY ELECTRON EMISSION FROMMETALS WITH A LOW WORK FUNCTION[Experiments to clear up ContradictoryViews on the Small Output].-H. Bruining.(Physica, No. 1o, Vol. 8, 1941, p. 1161onwards.)

Measurements were made on compact aluminiumand beryllium plates whose oxide skins wereremoved by sputtering. The removal diminishedthe s.e. output considerably. In the case ofaluminium the output curve of the oxide -free filmwas identical with that of a film deposited byvaporisation. With beryllium the oxide skincould not be removed completely, and the emissionwas therefore rather higher than that of a vaporisedlayer.1430. VELOCITY -MODULATION CURRENTS [in Mathe-

matical Treatment of Bunching in Klystron].-D. L. Webster. (Journ. Applied Phys.,Dec. 1942, Vol. 13, No. 112, pp. 786-787:letter expanding part of paper dealt within 395o of 1939.)

DIRECTIONAL WIRELESS

1431. ELIMINATION OF QUADRANTAL ERROR INLOOP -AERIAL DIRECTION FINDERS MOUNTEDIN SWAYING VEHICLES.-F. Stein & E.Marre. (Hochf :tech. u. Elek:akus., Nov. 1942,Vol. 6o, No. 5, p. 146.)

Telefunken patent, D.R.P. 720 746. To com-pensate for re -radiation effects, the axis of rotationof a frame aerial may be so inclined to the polarisa-tion direction of the incoming waves that the

253

resulting decrease in sensitivity just counteractsthe field -strength increase due to the re -radiation.To make this plan still serve in the case of swayingvehicles, it is proposed to fit jat the back of thevehicle) two frames R, and R2 rotated by a commondrive, connected in series or parallel to the receiver,and inclined in opposite sense to the direction ofpolarisation at an angle y = 2 arc tan's/sinwhere fma= is the maximum value of the quadrantalerror.1432. METHOD FOR THE RECEPTION OF PERIODIC

IMPULSES [in Pulse Direction Finding].-G. Ulbricht. (Hochf:tech. u. Elek:akus., Nov.1942, Vol. 6o, No. 5, p. 148.)

Telefunken patent, D.R.P. 72o 842. In orderto separate the ground, and space waves, thetransmitter sends out oscillations which whenbriefly interrupted renew themselves in correctphase : they are, in fact, coherent. The receiverpicks out the coherent oscillations carried by theground wave, in preference to the incoherentspace -wave oscillations, because the beat fre-quency produced by heterodyning gives only forthe coherent ground -wave impulse a differencebetween beat and impulse frequencies which canbe separated out by filtering. See also 72o of 1942.

1433. ARRANGEMENT FOR THE DETECTION OF ANOBJECT MOVING IN AN ELECTROMAGNETICRADIATION FIELD.-E. Gerhard. (Hochf:tech. u. Elek:akus., Nov. 1942, Vol. 6o, No. 5,

147.)Telefunken patent, D.R.P. 72o 352. Part of

the energy from the unmodulated transmitter" 1 " (Fig. 15) arrives at the receiver " 3 " afterreflection at the object " 2," by the path 5-6 :another part arrives directly by the path 8-9,which includes the re -radiator 10 " modulatedby a modulator " 7," and produces interferenceat the receiver. By this method of modulationwith an easily amplified frequency, small frequencyfluctations at the transmitter are rendered harm-less and the full energy of the main beam 5 canbe utilised.1434. DIRECTION -FINDING SYSTEM USING A CAPA-

CITIVE GONIOMETER AND A CATHODE-RAY OSCILLOGRAPH. - W. Hasselbeck.(Hochf:tech. u. Elek:akus., Nov. 1942, Vol.6o, No. 5, p. 148.) Telefunken patent,D.R.P. 729 855.

1435. AERIAL AND FEEDER SYSTEMS FOR THEPRODUCTION OF LANDING BEAMS [C. LorenzPatents].-J. Goldmann : R. von Ottenthal.(Hochf:tech. u. Elek:akus., Nov. 1942, Vol.6o, No. 5, p. 147 : p. 147.) D.R.P. 72o 837& 72o 745.

1436. BLIND LANDING SYSTEM GIVING CONTINUALINFORMATION AS TO DISTANCE FROM LAND-ING POINT. - Y. Rocard. (Hochf:tech. u.Elek:akus., Nov. 1942, Vol. 6o, No. 5,p. 147.) D.R.P. 72o 748.

1437. MULTI -BEAM ULTRA -SHORT-WAVE LANDINGSYSTEM GIVING HEIGHT INDICATIONS.-F. Nebel. (Hochf:tech. u. Elek:akus., Nov.1942, Vol. 6o, No. 5, p. 148.) Telefunkenpatent, D.R.P. 72o 785.

1442.

1443.

1444

1445

1446.

1447.

1448.


1438. COMMERCIAL AIRCRAFT AIDS [C . A.A. Inno-vations, including Radio -Range Monitorgiving Automatic Warning of 3° Drift].-(Electronics, Dec. 1942, Vol. 15, No. 12,pp. 86 and 88.)

ACOUSTICS AND AUDIO -FREQUENCIES

1439. CONTEMPORARY PROBLEMS IN TELEVISIONSOUND.-Townsend. (See 1462.)

1440. ON THE MUTUAL SUPERPOSITION OF PLANEUNDAMPED PRESSURE WAVES OF LARGEAMPLITUDE IN GASES.-H. Pfriem. (Akust.Zeitschr., March 1942, Vol. 7, No. 2, pp. 56-65.)

1441. THE PERCEPTION OF NOTE PITCH . . F.Schouten. (Akust. Zeitschr., Jan. 1942,Vol. 7, No. 1, pp. 33-35.) A long summaryof the Philips paper, an Italian abstract ofwhich was dealt with in 1425 of 1942.

SUBJECTIVE MEASUREMENT OF THE QUALITYOF TELEPHONIC LINKS: II-OBSERVATIONSOF REPETITIONS : TECHNICAL PROCEDURE.-Panzerbieter & Rechten. (Bull. Assoc.suisse des Elec., 13th Jan. 1943, Vol. 34,No. 1, pp. 22-24.) French version of theGerman paper referred to in 1141 of April.For Part I of the German paper see 3023 of1942.REMARKS ON THE " OELSNER " BUILDINGCONSTRUCTION [Review of W. Oelsner's" The Carrying Out of Modern HouseConstruction as a Mechanically Non -RigidSystem "].-G. Hofbauer : Oelsner. (Akust.Zeitschr., May 1942, Vol. 7, No. 3, pp. I xi-rr5.) See also 1448, below.

. SELECTED PROBLEMS IN ARCHITECTURALACOUSTICS [Undesirable Acolistic Effectsand Remedial Measures].-M. Rettinger.(Proc. I.R.E., Jan. 1943, Vol. 31, No. r ,pp. 18-22.)

. ABSORPTION OF SOUND BY POROUS MA-TERIALS I & II.-J. van den Eijk & C.Zwikker. (Akust. Zeitschr., Jan. 1942, Vol.7, No.' r, pp. 40-42.) Long summary of apaper in Physica, Nos. 2 & 5, VOL 8, 1941,PP. 149 & 469 onwards.

THEORY OF SOUND EXCLUSION BY THINWALLS FOR OBLIQUE INCIDENCE. - L.Cremer. (Akust. Zeitschr., May 1942, Vol.7, No. 3, pp. 81-104.)

THE IMPROVEMENT OF AUDIBILITY BY SOUNDREFLECTORS [Experiments on Improvementof Acoustically Bad Lecture Theatre ofTrondheim Technical College].-R. Berg &J. Holtsmark. (Akust. -Zeitschr., May 1942,Vol. 7, No. 3, pp. 119-120: summary only.)

RECENT IMPROVEMENTS IN THE TECHNIQUEOF THE ACOUSTICS OF ROOMS [particularlythe Oelsner Technique, including the Useof Plate Resonators with Electronic Controlof Tension].-M. Nuovo : Oelsner. (Forma -

1449.

May, 1943

cion y Documentation profesional [Madrid],March/April 1942, Vol. I, No. 2, p. 201 :summary, from L'Elettrotecnica, May 1941.)

THE SOUND -EXCLUDING ACTION OF WALLSON SPEECH AND Music [and the Failure ofthe Usual Method of Specifying Its Valueto give Direct Practical Conclusions as to theNoise -Level Reduction in Given Conditions :New Method enabling Noise -Reducing Actionfor a Given Noise Spectrum to be expressedin Phons.].-J. Capek. (Akust. Zeitschr.,July 1942, Vol. 7, No. 4, pp. 152-156.)

1450. THE PROPAGATION OF SUPERSONIC WAVES INWIRES.-E. Czerlinsky. (Akust. Zeitschr.,Jan. 1942, Vol. 7, No. 1, pp. 12-17.)

" Up to the present, no information has beenavailable as to the frequency -dependence of thepropagation velocity of supersonic waves in longwires : only concerning the very large ranges ofsupersonic signals along wires are measurementsknown (Sokolov, 2258 of 1936 [" theoretical range10 000 km for one watt "] : Kuntze, 3472 of 1936).Author's summary :-" The working out of thefrequency equation derived from the oscillationequation for infinitely long cylinders yields arepresentation of the dispersion curves for super-sonic waves along wires. In the frequency regionwhere the wavelength is of the order of the wirediameter, the propagation velocity falls steadilywith increasing frequency; for higher frequenciesthe velocity is practically independent of frequency.Dead zones do not occur. The simple calculationof two points characterising the shape of the dis-persion curve is given. Comparison with measure-ments on fifteen cylinders of different materials anddiameter shows the agreement of these investiga-tions with the above calculations."

1451. A SUPERSONIC -WAVE TRANSMITTER DRIVENBY AN ARC GENERATOR.-H. H. Zschirnt.(Hochf:tech. u. Elek:akus., Nov. 1942, Vol.6o, No. 5, pp. 126-135.)

The two chief methods of generating supersonicwaves are the particularly simple magnetostrictionoscillator, for frequencies from 20 to about 5o kc/s,and the piezo-quartz oscillator for higher fre-quencies. Unlike the latter device (which has lowmechanical and electrical losses and converts nearlythe whole of the energy supplied to it into oscillatingenergy) the magnetostriction oscillator has seriouslosses due to internal friction, eddy currents, andhysteresis. Since such an oscillator is usuallyrequired to deliver considerable supersonic power,its excitation needs a fairly large input, and valvegenerators of some kilowatts are neither cheap norsimple. These facts led to the idea that a Poulsenarc generator might be used as exciter : on the faceof it, it provides the simplest device for producinglarge h.f. outputs with a good efficiency, thoughwith the defects of rather poor constancy of fre-quency and amplitude.

Experiments were therefore made with a Poulsenarc in a coal -gas atmosphere, with a water-cooledcopper anode and a rotating carbon, a transversemagnetic field, and a 700 v d.c. generator. as supply.The magnetostriction, oscillator in the first seriesof tests was a commercial type (for submarine


signalling) built up of nickel stampings with tworectangular slots through which the thick rubber -covered exciting turns were wound (Fig. 2) : itsnatural frequency was 20.8 kc/s, varying slightlywith temperature or with change of biasing current.It was water-cooled, the top only of the core beingin contact with the atmosphere. Its position inthe complete arc circuit is shqwn at " Sch " inFig. i.

In developing methods for measuring the super-sonic amplitudes and frequencies, it was decidedthat the, applications of such a magnetostrictionoscillator did not require very great frequencyor amplitude constancy, and that very brief fluc-tuations of amplitude, or indeed short completebreaks in the oscillations, would not be seriousprovided they lasted no longer than about 1/Toothof a second. It was therefore necessary to look foran amplitude -recording device with sufficientsensitivity and a response time of less than f x oothof a second. Tests with thermophones showed that

- these had too great inertia even when very thinwires were used. Finally, a torsional sound -pressure' recorder was developed (similar to theRayleigh disc but with the plate comparable insize with the wavelength and suspended asym-metrically) on the lines of Fig. 3, where a 20 X 20 x0.1 mm mica plate is cemented at one edge to astretehed wire and damped by a drop of castor oilabove and below it, close to its suspension. Theplate is silvered on one side so as to act as a reflectorin the optical system of Fig. 3b, in which the illumi-nation of a photocell is varied, as the mica disc israised by the sound pressure coming from below it,by the action of the wedge-shaped aperture in thestop B. The photocurrent is amplified in a d.c.amplifier and recorded on a loop oscillograph. Anadjustable reflector R above the disc can be used toincrease the deflectiun, a maximum being obtainedwhen R is a whole multiple of a half -wavelengthaway from the oscillator and the mica plate liesin a pressure antinode of the standing waves.Fig. 6 shows the circuit developed for recording thefrequency and amplitude fluctuations of the Poulsenarc oscillations, and the oscillograms of Fig. 9 showhow the arc, when the magnetostriction generatoris put out of action, behaves as regards current(trace 1) and frequency (trace 2) : the former ispractically a straight line, while the latter showscompletely irregular fluctuations between 20.85and 20.44 kc/s. Very different is the frequencyoscillogram of Fig. ro : the interaction between theoscillator and the Poulsen arc circuit brings aboutan emphasis on the natural frequency of theformer (20.99 kc/s) and such fluctuations as thereare show a drawn-out saw -tooth course about thisfrequency. The lower trace is that of the super-sonic intensity as measured with the mica -discdevice : it shows a practically steady value withwith small dips corresponding to the steep sides ofthe frequency saw -tooth trace. Both traces canbe considerably improved (Fig. II) by measuresdescribed in section iv which make the fullest useof the stabilising effect of the magnetostrictionoscillator and ensure the optimum conditions (re-duced magnetic field, at the cost of some efficiency,thinner carbon, etc.). Preliminary results, inter-rupted by the war, showed that a much greater con-stancy of frequency and amplitude could be attained

255

for smaller supersonic powers by the use of a" Poulsen test lamp " arc generator (without amagnetic field) combined with an air-cooled nickel -tube oscillator: no fluctuations could be detectedwith the instruments used, apart from the slightmodulation due to the brushes of the d.c. generator.

Finally, it was found that by replacing the mag-netostriction oscillator by an equivalent electricalcircuit (Fig. 13) the same stabilising interaction withthe arc circuit was obtained : Fig. 14 shows thefrequency trace for the greatest output (thick carbon,strong magnetic field) and Fig. 15 that for the great-est constancy (thin carbon, weak field).

1452. SUPERSONIC WAVES: THEIR EFFECTS ANDEMPLOYMENT IN VARIOUS FIELDS OF MODERNTECHNIQUE [including Purification of Water,Stroboscopy, Television, etc.].--. Oggioni.(L'Elettrotecnica, 25th Aug. 194T, p. 410 on-wards : summary in Formation y Documen-tacion profesional, Sept./Oct. 1942, Vol. i,No. 5, p. 543.)

1453. A NEW SONIC AND SUPERSONIC TRANSMITTERFOR THE PRODUCTION OF STRONG INTENSITIESIN GASES.-L. Ehret 8r H. Hahnemann.(Zeitschr. f. tech. Phys., No. io, Vol. 23, 1942,pp. 24.5-266.)

The development of supersonic applications inindustry such as the preparation of the finestcolloids, the coagulation of aerosols, the supersonictesting of materials, and so on, as well as the medicalapplication of the biological effects of supersonicwaves, is continually impeded by the present lackof a high -intensity sound generator of reasonabledesign, reliable working, and satisfactory efficiency.Piezoelectric and magnetostrictive radiators delivertheir high intensities (up to to w/cm2) only to liquidmedia (Bergmann's book is cited here) : gaseousmedia take up only a 2.5 x ro4th part of the inten-sity radiated in the liquid. Hartmann's air:jetgenerator (see, for example, 4304 of 1940 and backreference [and, for later work, 1118 of 1941]) isparticularly suitable for. gases, but has the greatdisadvantage that the generated sound field liesactually in the gas current which produces it, whichin many applications is out of the question or at leasttroublesome.

The writers, therefore, working at the MotorResearch Institute of the Hermann Goring AircraftResearch Establishment, Brunswick, have modifiedthe air -jet generator so as to separate the soundfield from the flow of air. Two possible ways ofdoing this are illustrated in Fig. 3. A cylindricalring cc could be fixed just in front of the entranceto the Hartmann resonator b (for the relation of bto the compressed -air nozzle D see Fig. i) : it wouldbe designed so that its radial natural frequency wasequal to the Hartmann frequency, and would beexcited by the Riemann shock -waves formed a fewmillimetres from the resonator by the main andresonator beams. This plan has certain disadvan-tages, including the fact that the sound field ob-tained is not plane but approximately cylindrical.The second method was therefore adopted, in whichthe base of the resonator is replaced by a diaphragmaa, with a transverse natural frequency equal tothe Hartmann frequency, close to the open back ofthe resonator b. With this " aerodynamic plate


oscillator " intensities of the order of I w/cm2, freefrom the stream of air, were obtained. 'Variouspossible ways of improving the efficiency are dis-cussed at considerable length, and it is concludedthat the perfected apparatus would be useful alsofor the excitation of liquids, as a substitute for thepiezoelectric generator : preliminary tests with theunimproved model have confirmed this. A longsection (pp. 253-261) describes measuring methodsof various kinds developed in the course of theinvestigations : they assist in the rapid and accurateadjustment of the apparatus to give the maximumradiation.

1454. " PHYSIK UND TECHNIK DES TONFILMS[Book Review].-H. Lichte & A. Narath.(fount. Applied Phys., Dec. 1942, Vol. 13,No. 12, p. 746.) A German review wasrefetred to in 725 of 1942.

1455, MEASUREMENTS IN CONNECTION WITH THETHEORIES OF THE NATURAL VIBRATIONSOF HOLLOW CIRCULAR RINGS OF ARBITRARYWALL THICKNESS [Magnetically ExcitedSteel Rings, at Frequencies up to ioo kcis].-W. Kuhl. (Akust. Zeitschr., July 1942,Vol. 7, No. 4, pp. 125-152.) Cf. Federhofer,1184 of April.

1456. METERS FOR D.I.N. SOUND LEVELS : COM-MITTEE'S RECOMMENDATIONS.-(Akust. Zeit-schr., July 1942, Vol. 7, No. 4, pp. 156-159.)For an instrument on these lines see Kalden,1153 of April.

1457. THE RECORDING OF RAPIDLY VARYING PRO-CESSES WITH THE NEUMANN ATTENUATIONRECORDER [on the Magnetic Rotating -Disc -&-Fork Principle : Application to Re-verberation -Time & Noise -Level Recording :etc.].-M. Gosewinkel. (Akust. Zeitschr.,May 1942, Vol. 7, No. 3, pp. 104-III.)Cf. 799 of March.

1458. SIMPLE HARMONIC WAVE ANALYSER [Ampli-fier-Filter-Rectifier Instrument for DirectReading of Amount of Second- & Third -Harmonic Distortion, for Rapid Testing &Inspection of A.F. Devices].-R. F. Thomson.(Electronics, Dec. 1942, .Vol. 15, No. 12,pp. 61-62 and 155.)

1459. SOME DETAILS IN DIRECTIONAL DISCRIMINA-TION IN HEARING [Researches with " Arti-ficial Head " : Discrimination between Fore& Aft Sounds (and also between Sourcesabove & below Horizontal Plane) attributedto Small Involuntary Movements of Head,Not to Asymmetry of Head due to Shell ofEar].-K. de Boer & A. Th. van Urk.(Philips Tech. Rundschau, No. 12, Vol. 6,1941, p. "363 onwards.) For previous worksee 2741 of 1941.

1460. ELECTRONIC MUSICAL INSTRUMENTS AND THEPROGRESS OF THE PIPELESS ORGAN.-L.Vellard. (Rev. Gin. de d'Elec., April 1941,p. 214 onwards.) A fourteen -page survey.

May, 1943

1461. RECEIVER FOR SOUND SIGNALS TO GIVEWARNING TO MOTOR VEHICLE BEING OVER-TAKEN BY ANOTHER [Maximum Sensitivityselected so as to allow for Doppler Effect,assuming Speed Difference of 20-40 km perHour].-W. M. Hahnemann. (Akust. Zeit-schr., July 194.2, Vol. 7, No. 4, p. I59")Lorenz patent, D.R.P. 694 508.

PHOTOTELEGRAPHY AND TELEVISION

1462. CONTEMPORARY PROBLEMS IN TELEVISIONSOUND.-C. L. Townsend. (Proc. I.R.E.,Jan. 1943, Vol. 31, No. I, pp. 3-7.)

Author's summary :-" The present rapid develop-ment of television is introducing new problems insound pickup and operation. As the art progresses,engineering tools and methods must not only keeppace with, but generally anticipate, the needs ofthe program -producing staff in the production ofmore and more intimate material. The nature ofthe acoustic problems so raised, and their solutions,are treated in this paper. New tools necessary toproper operation and the methods of their employ-ment are discussed. For a better understanding oftelevision requirements the methods normallyemployed in motion pictures and standard radiobroadcasting are compared with those in use in thepresent television studio. Some indications as towhat may be required in the near future are dis-cussed and possible developments suitable for suchuse are described."

1463. AUTOMATIC FREQUENCY AND PHASE CONTROLOF SYNCHRONISATION IN TELEVISION RE-CEIVERS.-K. R. Wendt & G. L. Fredendall.(Proc. I.R.E., Jan. 1943, Vol. 31, No.PP. 7-15.)

A summary was dealt with in 829 of March.From the authors' summary : - " One of theproblems in the reception of television images is toprovide satisfactory synchronisation in the presenceof noise. . . . This paper describes a synchronisingmeans at the receiver that employs a new principlein the field of synchronisation. ' The principle isautomatic frequency and phase control of the saw -tooth scanning voltages. . . . Consideration of thisnew development indicates that its use would resultin several improvements in television service :(i) when severe noise conditions occur, an improvedpicture is obtainable at points within the presentservice area ; (2) under such noise conditions, theuseful service area is extended ; (3) the maximumresolution permitted by a television channel isrealisable at locations having low field strengths.It is expected that these improved results will beattained without increase in the cost of the tele-vision receiver."

1464. AN EXPERIMENTAL TELEVISION SYSTEM [for114 Mc/s Band : Semi -Portable : Experiencewith Type 1847 Two -Inch Iconoscope (in-cluding Effect of Excessive Capacitance toGround, and the Desirability of a Shading -Voltage Generator) : etc.].-R. Mautner &F. Somers. (Electronics, Dec. 1942, Vol. 15,No. I2, pp. 68-71 and 17o ..I73.)

4

May. 1943 WIRELESSENGINEER

1465. A 9 -KILOWATT TELEVISION TRANSMITTERFOR EXPERIMENTAL PURPOSES [with Par-ticular Attention to the Output Stage &Modulator].-M. van der Beek. (PhilipsTech. Rundschau, May 1942, Vol. 7, No. 5,p. 129 onwards.)

1466. THE OBTAINING OF HIGH VOLTAGE FROM THELINE -SWEEP UNIT : II [Variation of theDerived Voltage with Load : Relationbetween High Voltage & Angle of " Kipp " :Analysis, including Form of Deflecting Coils,Inductance of Deflecting System, ResultantInductance of " Kipp " Transformer loadedby Deflecting Coils, and Dependence of CoilLength on Deflection Angle & Ray Length].-H. Bahring. (Fernseh-A .G. Hausmitteil-ungen, No. 3, Vol. 2, 1941, p. 84 onwards.)For part I see 446 of 1941.

MEASUREMENTS AND STANDARDS

1467. STANDARDS AND STANDARDISATION [At-tempt to Define ' Accuracy " as applied toStandards of Self -Inductance, Capacitance,and Resistance].-W. H. F. Griffiths. (Wire-less Engineer, March 1943, Vol. 20, No. 234,pp. 109-126.)

1468. A PUSH-PULL ELECTROMETER -VALVE POTEN-TIOMETER [for Use with a Glass Electrode inpH Measurement].-C. T. Abichandani &S. K. K. Jatkar. (Quarterly Journ. IndianInst. Science, Oct./Dec. 1940, Vol. 3, No. 4,PP. 91-97.)

A d.c. amplifier having extremely low input -gridcurrent, comprising two electrometer triodes inpush-pull, followed by a pair of high -mutual -con-ductance triodes. Overall sensitivity o.8 µA/mv,which could be increased to roo µA/my withadequate screening. Zero drift 0.2µA per hour atthe lower sensitivity.

1469. RECTIFIER MILLIAMMETERS [Satisfactory In-strument for indicating Small Out -of -BalanceCurrents.]-T. A. Ledward. (Electrician,5th March 1943, Vol. 13o, No. 3379, p. 239.)

1470. AN ELECTROSTATIC VOLTMETER WITH LIGHT -SPOT INDICATION [" Universal Electrometer,"for D.C. & A.C. up to 6 Mc/s : in Plastic Caseabout 8" x 9" x 5" : Quadrant - Electro-meter Movement with Stretched -RibbonSuspension : Wide Range].-SiemensHalske A.G. (Arch. f. Tech. Messen, May 1941,Part 119, J762-3, Sheet F3.)

1471. ON THE DEVELOPMENT OF A NEW CURRENTAND VOLTAGE MEASURING INSTRUMENT FORULTRA -HIGH FREQUENCIES [Readings Inde-pendent of Frequency from Zero up to aboutroo Mc/s] BASED ON THE OPTICAL STAVE OR" SCHLIEREN " METHOD.-W. FringS. (1Iochf:tech. u. Elek:akus., Nov. 1942, Vol. 6o, No. 5,pp. 117-125.)

Covering the same ground as the paper by Malsch& Frings dealt with in 243o of 1942, but in a fullerway. Thus the d.c. amplifier developed for thedifferential photocell is described fully and its

257

working discussed. Currents down to about ro MAcan be measured with an accuracy within 5%. Theagreement between results obtained with thisinstrument and those given by Braune's instrument(see abstract cited above) and by a " home-made "thermojunction, and the disagreement with thosegiven by a Hartmann & Braun thermojunctionmeter, led to the conclusion that at 3 m the latterinstrument was incorrect. Enquiries on this pointto the makers led to a reply from J. Fischer sayingthat a series of their thermo-converters had justbeen tested at the P.T.R. with the result that the5o mA and loo mA converters were found to havea 4.5%. error at 3 m, while the ro mA model showedabout a 5% error at 6 m : " at metric wavelengths,conditions in the converters below the 5o mA sizeare rather difficult from the h.f. viewpoint, in so faras the hot wire is not stretched straight but iscoiled." The particular H. & B. instrument inquestion had a full-scale reading for 20 MA. Thewriter points out the special suitability of the" Schlieren " instrument for testing the u.h.f.errors of commercial meters.

1472. MAGNETIC ZERO -CURRENT AMPLIFIERS [forMeasuring & Control Purposes].-W. Geyger.(Arch. f. Tech. Messen, June 1941, Part 12o,Z634-2, Sheet T91.) For other papers see1134 of 1942 and back references.

1473. AMPLIFICATION OF DIRECT VOLTAGES WITHA HIGH -FREQUENCY PUSH-PULL CIRCUIT.-,Kerkhof. (See 1387.)

1474. MEASUREMENTS IN CONNECTION WITH THETHEORIES OF THE NATURAL VIBRATIONS OFHOLLOW CIRCULAR RINGS OF ARBITRARYWALL THICKNESS.-Kuhl. (See 1455.)

1475. ON THE MEASUREMENT OF THE Loss ANGLEOF INSULATING MATERIALS [with Descriptionof Equipment & Technique at the RadioLaboratory of the Italian Ministry of Aero-nautics].-F. Bocci. (Formation y Docu-mentacion profesional [Madrid], May/June1942, Vol. r, No. 3, p. 290 : summary, fromL'Elettrotecnica, loth & 25th July 1941.)

1476. NOTES ON BRIDGE BALANCING [Brief Accountof Types of Bridges for measuring Impe-dances].-C. F. Nordica. (Radio [New York],Dec. 1942, No. 275, pp. 9-11 and 46, 47.)

1477. THE MUTUAL INDUCTANCE OF COAXIALCIRCULAR RINGS IN THE PRESENCE OF APERMEABLE CORE.-Buchholz. (See 1391.)

1478. THE CAPACITY EFFECT IN SOLENOIDS [and aNew Method of calculating the Self -Capaci-tance of a Multi -Turn Coil].-Lopez. (See

1395.)

1479. AUTOMATIC CIRCUIT TESTING.-Bissmire &Davies. (See 1537.)

1480. COUNTER TUBES AND THEIR USE IN MEASUR-ING TECHNIQUE.-H. Neuert. (Arch. f. Tech.Messen, June 1941, Part 120, J076-1, SheetsT85-86.) For later papers see 2918 of 1942.


1481. PRECISION TIME CONTROL [Amplified Tuning-Fork -Output Power -Supply avoids Syn-chronous -Clock Variations caused by War-time Overloading of A.C. Mains : Accurateto One -Third Second per Day].-Nat. Broad-casting Company. (Electronics, Dec. 1942,

. Vol. 15, No. 12, PP. 49 and 156, 157.)

1482. A FREQUENCY -MODULATED RESISTANCE-CA-PACITANCE OSCILLATOR.-Chang. (See 1403.)

1483. A NEW FREQUENCY -DIVIDER FOIL OBTAININGREFERENCE FREQUENCIES.-F. R. Stansel.(Bell Lab. Record, Dec. 1942, Vol. 21, No. 4,PP. 97-99.)

1484. A METHOD FOR THE EXACT REGULATION OFFIXED ROTATIONAL SPEEDS OF AN AXLE.-F. Kirschstein. (Zeitschr. f. tech. Phys.,No. 9, Vol. 23, 1942, pp. 234-240.)

In measuring technique it is often necessary to drivesome rotating apparatus at various exact speeds :at each speed the r.p.m. must be. adjusted to thepredetermined value with a definite absolute accu-racy and must be held constant over'a certain timeof test. If the speeds involved differ greatly amongthemselves the only type of driving motor which canbe considered is a d.c. motor with its field excitedby a fixed current and its armature supplied from aspecial d.c. generator with adjustable excitation(Leonhard connection). The speed of the drivingmotor is then proportional to the voltage deliveredby the d.c. generator, and thus can be regulated overa wide range through the excitation of the latter.Such an arrangement, however, adjusted to a'defi-nite number of r.p.m. and then left to itself, willshow after only a short time a departure from theprescribed speed as a result of variable frictionalresistances, mains fluctuations, warming -up effects infield and armature windings, and so on : and sooneror later it will be necessary to restore the wantedspeed by a readjustment. This readjustment mustbe made with the help of some form of tachometerand is liable to be not very precise because of thelimited accuracy of these instruments. The paperdescribes a regulating method suitable for thosecases which demand an exact setting and mainten-ance of rotational speed.

The arrangement is shown schematically inFig. r, where " 2 " is the driving motor with itsaxle prolonged in one direction to form the drivingshaft " r " and in the other to terminate in themagnetic -brake disc " ro." The shaft " r " carriesa toothed tone -wheel " 5 " which generates in thetelephone pick-up coil " 6 " a voltage of frequencyproportional to the r.p.m. of the shaft : this voltageis taken to a phase -circuit " 7 " where it is com-bined with a voltage of frequency f,In given bythe action of a frequency -dividing circuit " 4 "on the fundamental frequency f, of a precisiongenerator " 3 " which may be of the tuning -forkhummer type. Fig. 2a shows how the circuit" 7 " adds together the control voltage Ust offrequency f,In and the tone -wheel voltage Ur,and with the help of a rectifier G and resistance R.converts the resultant voltage U, into a d.c. regu-lating voltage and (through an amplifier valve V)causes a corresponding regulating d.c. to flowthrough the winding of the magnet " 8 " which forms

May, 1943

with the copper or aluminium disc " z o " an eddy -current brake on the driving motor. On the assump-tion of synchronism, so that the two alternatingvoltages U,t and UT are of precisely the samefrequency, Fig. 2b gives the vector diagram for thesevoltages and the resultant, U.: from this if is seenhow a rise in axle speed would cause the vectorUT to rotate in the direction reducing the angle(about 9o°) of its lag behind U,,, with a resultingincrease in U, and in the current through themagnet winding, and a consequent braking of thedriving motor. The reverse action is produced byfall in the axle speed, so that speed constancy isobtained : it is not necessary to adjust the speedaccurately so that the two frequencies are preciselyequal, only so that suitably slow beats occurbetween U" and UT, after which the synchronisa-tion occurs automatically.

The above arrangement is satisfactory for balanc-ing out small fluctuations of the driving moment :such, for instance, as might occur if the motor wereleft to itself for a few minutes. If it is desired tocompensate for larger fluctuations, occurring gradu-ally, the additional device included in Fig. 3 may beused. Here the amplifier valve V of the originalarrangement has added to it in parallel a second,similar valve V', to whose grid the regulating d.c.is taken through an RC section, and whose anodecircuit contains two relays, R, and R,. The anodecircuit of the valve V is shunted by two resistancesin series, of which Rb is set by hand and R. isadjusted by a series motor M with two oppositely -wound field coils F1 and F,, so that the motor runsforwards or backwards according to whether thearmature is supplied through F1 or F,. The relayR1 closes the contact r1 putting F1 in circuit, whilethe relay R, opens the contact r, which normallykeeps F, in circuit : shunt resistances are adjustedso that RI closes and opens for larger currents thanR,, so that (in fact) R1 opens in the presence of acurrent larger than that required to close R,. Themode of action of this arrangement (depending onthe fact that the current through V, and conse-quently through V', depends on the value of thedriving moment obtaining at any particularinstant) is described on p. 237. The result is thatthe equipment can be run for hours on end at therequired speed without any readjustment, providedonly that no sudden large fluctuations occur forthe two controls, being as it were " in series,"havea fairly large time constant. Testing techniqueis described in section iv : in the actual case of aturntable revolving at loo r.p.m., the estimatedfluctuations were ro-3 of the mean r.p.m., with aperiod of a few seconds, and 5 x io-4 of the meanr.p.m., with a period of one rotation of the turn-table.

In a discussion of the choice of a frequency -dividing device, an arrangement due to Well-schmied (to be published in the near future) ismentioned as very satisfactory : this consists oftwo " kipp " circuits in series, of which the first issynchronised by the control frequency f, and thesecond by the frequency given by the first.Withan input frequency of ro kc/s the arrange-ment gives 23 available frequencies between rooand 50o c/s by the rotation of a single selectorswitch.


SUBSIDIARY APPARATUS AND MATERIALS

1485. A METHOD FOR THE EXACT REGULATION OFFIXED ROTATIONAL SPEEDS OF AN A XLE . -Kirschstein. (See 1484.)

1486. RADIO BIBLIOGRAPHY : REMOTE CONTROL[and List of Previous Subjects].-Retten-meyer. (Radio [New York]; Dec. 1942, No.275, pp. 26 and 29..34.) For a previousset see 455 of February.

1487. SOME USEFUL CIRCUITS EMPLOYING THYRA-TRONS AND IGN IT RONS.-Maddock. (See1401.)

1488. LOCALISATION OF THE DISCHARGE IN GEIGER-MULLER COUNTERS . -Wilkening & Kaune.(Phys. Review, istli3th Dec. 1942, Vol. 62,NO. I I / 1 2, pp. 534-537.)

1489. ELECTRONIC VOLTAGE REGULATOR FOR PRE-CISE DENSITOMETRIC MEASUREMENTS.-Di,e-tert & Hasler. (Journ. opt. Soc. Am., Jan.1943, Vol. 33, No. I, pp. 45-49.)

A number of simple voltage regulators are avail-able which claim to limit the output variation toone volt for an input variation from 90 to 13o volts,but the regulation for variation of input frequencyis by no means so good. A summary of the methodhere described was dealt with in 878 of March.

1490. THE STATISTICAL AVERAGE IN THE MEASURE-MENT AND USE OF DRY -PLATE RECTIFIERPLATES : I & II.-Maier. (Arch. f. Tech.Messen, May & June 1941, Parts 119 &120, Z52-3 & 4, Sheets T76-77 & 89-9o.)

1491. FACTORS DETERMINING THE INTENSITY OFOSCILLATIONS IN THE PLASMA OF GASEOUSDISCHARGE . -Sluzkin & Maydanov. (See1408.)

1492. THE RECORDING OF RAPIDLY VARYING PRO-CESSES WITH THE NEUMANN ATTENUATIONRECORDER.-Gosewinkel. (See 1457.)

THE METHOD OF CHROMATOGRAPHIC ABSORP-TION IN PHOSPHORESCENCE CHEMISTRY [forPurification of the Basic Material in thePreparation of Sulphide Phosphors], andON THE FINE STRUCTURE OF ALKALINE -EARTH SULPHIDE PHOSPHORS . -Tiede &Schikore : Tiede. (Physilt. Berichte, 15thSept. 1942, Vol. 23, No. 18, pp. 172o-1721 :p. 1721: summaries only.)TWO NEW HOT -FILAMENT RESISTANCES.-Minter. (Journ. Applied Phys., Jan 1943,Vol. 14, No. 1, pp. 49-5o.)

Author's summary :-" This report gives theresults of attempts to produce a gaseous atmospherein which the thermal conductivity increases withincreasing ambient temperature much more rapidlythan the Normal, thereby causing a hot filamentmounted in such an atmosphere to behave as if ithad a negative temperature coefficient of resistance.There is also described a means of producing a gase-ous atmosphere the thermal conductivity of whichdecreases very rapidly as the ambient temperatureis increased, so that a hot filament mounted in suchan atmosphere behaves as though it had a positive

1493.

1494.

259

temperature coefficient of resistance three to fourtimes as great as normal." The two resistancesdescribed should be useful in many laboratorycircuits, provided there is enough current flowingto heat the filament to a sufficiently high tempera-ture above surroundings.

1495. STANDARDISATION OF FIXED RESISTORS[Agreement on Preferred Values and Toler-ances].-(Wireless World, March 1943, Vol.49, No. 3, p. 71.) The result of consultationbetween the Interservice (Communications)Components Committee and others.

1496. REMARKS ON THE USE OF " SIMMER " RINGSAS VACUUM - TIGHT GLANDS. - Raether.(Zeitschr. f. tech. Phys., No. io, Vol. 23,1942, pp. 266-267.)

" In many laboratories use is now being made of' Simmer. rings ' (made by Carl Freudenberg,Simmerwerk, Weinheim/Bergstr.) where some kindof movement inside a vacuum apparatus is required,such as the rotation or longitudinal adjustmentof a spindle. Such a ' Simmer ' ring is shownschematically in Fig. i. In the type most generallyused the ring is composed of a cylindrical outersheath continuing inwards to form a sleeve or cuffwhich makes the air=tight sliding joint : bothparts are made of ' Simrit ', a rubber -like material.The outer sheath is stiffened by a metal ring[helping to press it outwards against the fixed tubeor housing, to make a stationary vacuum -tightjoint]. The. vacuum-tightness of the cuff is broughtabout with the help of its ' vacuum lip ', whosediameter is rather smaller than that of the spindle :it also is made of ' Simrit ', and thanks to itsflexibility it clings to the spindle and prevents theentrance of air. With the help of a spiral -springring or a rubber ring (shown in Fig. r as press -ring ') the lip is given an additional pressure againstthe spindle, the contact surface being about 0.5 mmlong. The ' Simmer ring ' is pressed into thehousing and sits firmly in it, since the outer sheathis a ' press fit ' to the housing diameter specifiedin the catalogue. Since the vacuum -tightnessthus obtained is questionable, we insert a ring ofApiezon wax in the seating before introducing thering.

The simplification of the vacuum gland forrotatable and displaceable spindles compared withprevious arrangements is obvious. To find outthe limits of usefulness of these gland's we havecarried out the following orienting measurements ",with a polished steel (or glass) spindle of 8 mmdiameter, freed from grease (as was also the " lip "of the ring) with benzine. The conditions ofvacuum were such that within four hours thepressure rose by about 3 x io-4 mm Hg whetherthe ring was in place or was replaced by a glassplate, so that this leakage had nothing to do withthe ring. A single slow rotation of the de -greasedspindle caused a barely measurable rise in pressure,ten whole turns made it rise by about 4 x 10-4 mmHg. A careful and slow displacement of thespindle through 5 cm raised the pressure by5 x ro-4 mm. Converted to atmospheric pres-sure, an increase of i x ro-4 mm representedabout half a cubic centimetre of air.

With a good pump, such a leakage is of little

26o WIRELESSENGINEER

importance as a rule, particularly since in manycases the spindle would be moved only for a momentand would be stationary during an actual test.But in certain circumstances, as in a vaporisingapparatus with hot metallic foils, it is desirableto reduce the entrance of air as much as possible.The writer therefore repeated the measurementsafter moistening the spindle with Apiezon oil B(high -vacuum pump oil) : the pressure rises weresmaller by at least one order of magnitude, tenrotations causing no change that could be read onthe McLeod gauge, and displacements of a fewcentimetres a barely measurable rise of a few

0-5 mm Hg. The use of the oil is not onlyharmless but also lasting : Simrit is unaffected byoil. In some Cases the writer filled the part abovethe lip with oil, in others a moistening with oilwas all that was required. Since these rings aresupplied in diameters up to 285 mm they can beemployed for expansion chambers in addition totheir many other uses. Altogether, the writerbelieves that they will soon become indispensable.

1497. AN ELECTRICAL CONTACT TESTING MACHINE.-Suggs. (A .S.T.M. Bulletin, Dec. 1942,No. 119, pp. 25-30.)

Consists principally of three units of mechanismfor opening and closing electrical contacts, withequipment for measuring their operating character-istics. Three pairs of contacts can be fitted andcan be tested simultaneously.

1498. PROTECTIVE RELAYS [Notes on New BritishStandard Specification]. -Casson. (Elec.Review, 12th March 1943, Vol. 132, No.3407, PP. 346-348)

1499. " BOUNCELESS BALL " [Steel Ball Half -Filled with Metallic Powder] POINTS WAYTO PROGRESS IN ELECTRICAL RELAYS.-(Scient. American, Nov. 1942, Vol. 167,No. 5, p. 207.)

1500. INVESTIGATION OF CAR CONDENSERS [Russian& Foreign : Dependence of Capacitance -Variation, with Temperature, on Nature ofPaper -Impregnating Dielectric (Castor Oilor Halowax) : of Loss -Angle Variation,with Temperature, on Nature of Contactsto Aluminium Foil : Exclusion of Moisture :etc.] . - Renhe & Schljachter. (Physik.Berichte, 1st Sept. 1942, Vol. 23, No.17,pp. 1637-1638: summary of Russian paper.)

1501. THE MUTUAL INDUCTANCE OF COAXIALCIRCULAR RINGS IN THE PRESENCE OF APERMEABLE CORE.-Buchholz. (See 1391.)

1502. NEW STACKPOLE IRON CORES [Note on NewProduct].-(Radio [New York], Dec. 1942,No. 275, J. 49-50.)

1503. THE PERMEABILITY OF IRON AT RADIOFREQUENCIES.-Smith & Dickey. (Phys.Review, istlx 5th Dec. 1942, Vol. 62, No.II/12, p. 561 : summary only.)

][504. A PERMALLOY MAGNETOMETER.--JOhlISOIL(Phys. Review, zst/i5th Dec. 1942, Vol. 62,No. x1/12, p. 561: summary only.)

May, 1943

1505. HAMMERED SOLDERING - IRON TIPS LASTLONGER [Technique evolved for HotterIrons now necessary with Tin -EconomisingSolders] .-(Electronics, Dec. 1942, Vol. 15,No. 12, p. 84.)

1506. CONTRIBUTION TO THE INDIRECT BRAZINGOF ALUMINIUM [Use of Autogal-C Paste &Silumin Filings].-Maier. (Physik. Berichte,15th Aug. 1942, Vol. 23, No. 16, pp. 1588-1589.)

1507. RESEARCH ON A DOUBLE -VALVE -RECTIFICA-TION SYSTEM [for Supply of Direct Currentto Relays, etc.].-.-Weisglass. (See 1389.)

1508. EXPERIMENT ON THE ELECTROSTATIC RE-PULSION BETWEEN PARALLEL ELECTRONRAYS. - Stehberger. (Ann. der Physik,No. 5, Vol. 41, 1942, p. 325 onwards.)

Of the two ways in which a mutual repulsionof the electrons in an electron beam can manifestitself-a longitudinal effect (space -charge actionin a valve) and a transverse effect (in the long raysof electron optics)-the latter has had hardly anyquantitative investigation. The writer has there-fore carried out measurements with two " spiral "electron beams, such as he used in previous experi-ments (see 1154 of 1938), running parallel and closeto each other. The maximum deflection wasalways less than 1 mm. The measured valuescorresponded quite well with theory : a discrepancywhich appeared to increase with increasing longi-tudinal magnetic field was attributed to a smallretroactive component of the spiral beam.

1509. THE ELECTROSTATIC FOCUSING OF HIGH-SPEED ION AND ELECTRON BEAMS.-CraggS.(Journ. Applied Phys., Dec. 1942, Vol. 13,No. 12, pp. 772-780.)

An account is given of experiments carried outin preparation for the development of an apparatusfor artificial atomic disintegrations (see also 2797and 3356 of 1942). The problems involved are(i) the focusing of a beam of 10-30 kv hydrogenions from a canal -ray ion source, and (2) thebehaviour of a cascade acceleration tube of theCoolidge type. The general properties of electro-static focusing systems are treated briefly andreferences are given to other investigations in thisfield.1510. CATHODE-RAY TUBE FOR VERY HIGH FRE-

QUENCIES, WITH DEFLECTING ELECTRODESFORMED OF STRETCHED CONDUCTORS ONEITHER SIDE OF RAY, WITH TERMINATINGRESISTANCE AT FAR END TO ELIMINATEREFLECTION. - Graffunder. (Hochf:tech. u.Elek:akus., Nov. 1942, Vol. 6o, No. 5, pp.145-146.) Telefunken patent, D.R.P.72o 754.

151 I. ELECTROSTATIC ELECTRON MICROSCOPY : I.-Bachman & Ramo. ( Journ. AppliedPhys., Jan. 1943, Vol. 24, No: 1, pp. 8-18.)

Investigations to develop a simplified, practicalmicroscope yielding magnified images of transparentspecimens. The first part deals with the generalproblem of design, including the electron gun andimaging lenses.


1512. MICROSCOPES IMPROVED ENeW PortableModels make Electron Microscope availablefor Wider Usefulness in Small Laboratoriesand War Industries].-(Sci. News Letter, 12thDec. 1942, Vol. 42, No. 24, pp. 374-375.)

1513. SURFACE -LEAKAGE PHENOMENA IN PHENOLICRESIN PLASTICS.-Esch. (Physik. Berichte,ist Aug. 1942, Vol. 23, No. 15, p. 1517:summary, from Kunststoffe, No. 3, Vol. 32,1942, p. 81.)

Beginning with an extract from the reports ofStager and his colleagues (see 3144 [and 3521] of1941) dealing with investigations of structure andsurface leakage, the writer shows that results withphenolic, cresolic, and aniline resins (by the usualdropping method with electrolytes containing aNekal addition, in the presence of an electric field)indicate that the chemical influence of the electro-lyte plays a dominating role. Thus salts of thealkaline elements are particularly effective in intro-ducing surface leakage, the chlorides being moreactive than the sulphates. Sodium sulphate isdistinguished by a particularly small temperaturedependence in the formation of leakage paths.

1514. SURFACE -LEAKAGE CURRENTS ON SYNTHETICMATERIALS. - Vieweg & KlingelhOffer.(Physik. Berichte, 1st Aug. 1942, Vol. 23,No. 15, p. 1517: summary, from Kunststoffe,No. 3, Vol. 32, 1942, pp. 77 onwards.)

A new method is described of measuring the sur-face conductivity of materials, which is so importantin the development of synthetic insulators : pre-vious Methods using two knife-edges only measurecurrents which are the sum of the surface -leakageand penetrating currents and are not, therefore,satisfactory in the knowledge they provide. In

a circular disc of the material iscovered on both sides with a metallic coating, thaton one side being divided into a number of ringsby concentric spaces. The conductivity is meas-ured for various sizes of electrode, by connectingall the rings together for the first measurementand then disconnecting ring after ring,, starting atthe centre. The conductivity is 'plotted as afunction of the surface, and extrapolation to zerosurface gives the pure surface conductivity at theperiphery of the sample. As a precaution, allspecimen discs are made with the same length ofperiphery.

Results are given for " Type S " material atvarious degrees of humidity and various tempera-tures, and also under the action of ammonia. Theinfluence of water -vapour pressure on the surfaceconductivity is explained by the hypothesis ofconduction paths due to absorption by the saltsexisting in the surface : these paths unite, at highhumidities, and form closed circuits. For previouswork by these writers see 2815 of 1942.

1515. DUCTILITY OF FOILS OF SYNTHETICMATERIALS [Important in the Insulationof Wires, particularly of Those of FineDiameter]. - Fischer & Witte. (Physik.Berichte, 1St Aug. 1942, Vol. 23, No. 15,pp. 1517-1518 : summary, from Kunststoffe,No. 4, Vol. 32, 1942, p. r10 onwards.)

Even polystyrol is found to be suitable for the

261

insulation of wires, in spite of its small breakingelongation : apparently the application to the wirecauses a marked increase in ductility. The presentwork investigates this point by experiments onstyroflex foils gummed to carrier foils, after it hadbeen shown that the mere treatment of the styroflexwith the gum had no influence on the ductility.Measurements on foils of styroflex (polystyrol),luvitherm (polyvinyl chloride), triafol N (celluloseacetate), and lyafol (polyamide) gdve in all casesductilities for the combined foils which werehigher than that for polystyrol alone: in thecase of triafol and styroflex, higher than that of thecarrier foil. The writers conclude that the breakingstretch of the foils is very greatly reduced by theeffects of inhomogeneities, variations in thickness,and non -uniform loading : when these are reduced(as happens when the foils are stuck onto a carrierfoil) the ductility is considerably improved.

1516. HIGH - FREQUENCY ENERGY LOSSES INSOLUTIONS CONTAINING MACROMOLECULES[Measurements by Malsch ThermometricMethod at Frequencies up to 6o Mc/s].-Conner. (Physik. Berichte, 1st Sept. 1942,Vol. 23, No. 17, p. 1640.)

1517. ELASTIC AFTER-EFFECTS AND DIELECTRICABSORPTION IN GLASS [and Similar HighlyViscous FluidS such as Ebonite : Investigationof the Mechanism]. -Taylor. (Physik. Ber-ichte, 1st Sept. 1942, Vol. 23, No. 17, p. 164o :summary, from Journ. Applied Phys., No.IO, Vol. 12, 1941, p. 753 onwards.)

1518. DIELECTRIC MEASUREMENTS ON PLASTICISEDPOLYVINYL CHLORIDE WITH EXTERNAL ANDINTERNAL PLASTICISER.-Wfirstlin. (Physik.Berichte, 1St Sept. 1942, Vol. 23, No. 17,p. 1639.)

1519. DIELECTRIC MEASUREMENTS ON BITUMENAND ITS DERIVATIVES : II. -Walther.(Physik. Berichte, 1St Sept. 1942, Vol. 23,No. 17, p. 1638: summary, from Kolloid-Zeitschr., No. 2, Vol. 99, 1942, p. 129 on-wards.)

1520. ALTERNATIVE MATERIALS : DEVELOPMENTSOF ELECTRICAL INTEREST FOR PANELS ANDSCREENING. -(Electrician, 12th March 1943,Vol. 13o, No. 338o, p. 27o.).

Metrelite boards are made by a process in whichfinely divided metals, zinc or aluminium, aresprayed on to synthetic resin bonded boards orsimilar materials. The electromagnetic screeningproperties of a standard sheet are equal to thoseof tinfoil of 0.2 mm thickness.

1521. ENAMELLED WIRE [Exceptionally Good Pro-perties of " Thermex " and " Formex "Enamel].-(Elec. Review, 26th Feb. 1943,Vol. 132, No. 3405, p. 29o.)

1522. BRITISH PLASTICS [Report of Speech at-British Plastics Federation Luncheon]. -Mohr. (Electrician, 26th Feb. 1943, Vol.13o, No. 3378,. p. 211.)


1523. FIBRE GLASS AND ITS APPLICATIONS INELECTRICAL TECHNIQUE [Comparative Dataof " Vetrotex," Cotton, & Silk : Use ofFibre Glass as Substitute for Mica, etc.].-Esme. (Physik. Berichte, 15th Aug. 1942,Vol. 23, No. 16, p. 1584: summary ofFrench paper.)

1524. ON THE MECHANISM OF FILM -FORMATIONFROM EMULSIONS OF SOME HIGH -POLYMERCOMPOUNDS [particularly Polychlorvinyl].-Wojutzki & Dsj adel. (Physik. Berichte,1st Aug. 1942, Vol. 23, No. 15, p. 1518:summary of Russian paper.)

1525. MEASUREMENT OF THE FILM THICKNESS OFANODIC COATINGS ON ALUivIINIUM ANDALUMINIUM ALLOYS [Method depending onSpecial Solvent which Dissolves the OxideSkin without Damage to the Metal].-Wied-erholt & others. (Physik. Berichte, 15thAug. 1942, Vol. 23, No. 16, p. 1586.)

STATIONS, DESIGN AND OPERATION1526. SUBJECTIVE MEASUREMENT OF THE QUALITY

OF TELEPHONIC LINKS : II-OBSERVATIONSOF REPETITIONS : TECHNICAL PROCEDURE.-Panzerbieter & Rechten. (See 1442.)

1527. A 112 - MC/S TRANSMITTER - RECEIVER.-Lynch. (QST, Jan. 1943, Vol. 27, No.PP. 30-33.)

GENERAL PHYSICAL ARTICLES

1528. A PROBLEM IN ELECTROSTATICS.-Thomas.(Phys. Review, 1st/15th Dec. 1942, Vol. 62,No. 11/12, p. 561 : summary only.)

1529. RATIONAL ELECTRODYNAMICS : I - THELIMITATIONS OF CLASSICAL ELECTROMAG-NETISM : II-THE IDEAS OF KINEMATICALRELATIVITY.-Milne. (Phil. Mag., Feb.1943, Vol, 34, No. 229, pp. 73-82, 82-101.)

The ideas developed " constitute the extensionto the domain of electromagnetism of the methodswhich I have been developing since 1932 in therealms of kinematics, dynamics, and gravitation."1530. FUNDAMENTAL ELECTROMAGNETIC CON-

CEPTS AND THEIR DIMENSIONS [Editorial].-G.W.O.H. (Wireless Engineer, March1943, Vol. 20, No. 234, pp. 105-109.)

1531. THE MOTION OF THE MESON IN A HOMO-GENEOUS MAGNETIC FIELD [TheoreticalTreatment leading to Expressions for Energyof Meson showing Its Possession of a MagneticMoment of Magnitude which (unlike That ofan Electron) is Independent of the Velocity].-Galanin. (Jason. of Phys. [of USSR],No. 1/2, Vol. 6,1942, pp. 27-34 : in German.)For a further paper, on electron and mesonspin, see pp. 35-47.

1532. THEORETICAL PHYSICS IN U.S.S.R.DURING THE PAST TWENTY-FIVE YEARS.-Ivanenko. (Nature, 13th March 1943, Vol.151, pp. 293-294.) Cf. 1309 of April.

May, 1943

1533. THE EXTREME PROPERTIES OF MATTER [49thJames Forrest Lecture to Institution ofCivil Engineers : Subjects discussed includeElasticity, Plastics, Fluids, Gases, Weight,Electrical Properties, and Magnetism].-Darwin. (Engineering, 12th March 1943.Vol. 155, pp. 136---137: 19th March, PP. 144-145.)

ON THE THEORY OF COSMIC -RAY SHOWERS[Further Contributions to the FluctuationProblem].-Scott & Uhlenbeck. (Phys.Review, 1st/15th Dec.1942_, Vol. 62, No. IT /i2,PP. 497-508.)

1535. RELATION OF THE COSMIC RADIATION TOGEOMAGNETIC AND HELIOPHYSICAL ACTIVI-TIES.-Broxon. (Phys. Review, ist115thDec. 1942, Vol. 62, No. 11/12, pp. 508-522.)

1534.

MISCELLANEOUS

1536. AMPLIFICATION OF DIRECT VOLTAGES WITHA HIGH -FREQUENCY PUSH-PULL CIRCUIT.-Kerkhof. (See 1387.)

1537. AUTOMATIC CIRCUIT CHECKING [Design ofApparatus for Increasing Speed and Relia-bility].-Bissmire & Davies. (Wireless World,March 1943, Vol. 49, No. 3, pp. 68-71.)

1538. RADIO BIBLIOGRAPHY : REMOTE CONTROL[and List of Previous Subjects].-Retten-meyer. (Radio [New York], Dec. 1942, No.275, pp. 26 and 29 ..34.) For a previous setsee 455 of February.

1539. AMPLIFIERS AND MAINS UNITS FOR THESUPPLY OF COUNTER TUBES [including theVarious Components of the Over -All Resolv-ing Power of the Equipment, Rossi -TypeCoincidence Amplifiers, Principles of HexodeMixing, etc].-Rehbein. (Physik. Berichte,15th Aug. 1942, Vol. 23, No. 16, p. 1547summary only.)

1540. SUPER -FREQUENCY DEVICES IN VIBRATIONMEASUREMENT [Short Survey of Pick -Upswith Natural Frequencies Higher than Thatof Vibration to be Measured : includingthe A.E.G. Rochelle -Salt Pick -Up, Meister'sCapacitive Type, etc : with LiteratureReferences].-Meister. (Arch. f. Tech. Messen,June 1941, Part 120, ' V172 -I, Sheet T78.)

1541. MAGNETIC ZERO -CURRENT AMPLIFIERS [forMeasuring & Control Purposes].-Geyger.(Arch. f. Tech. Messen, June 1941, Part 120,Z634-2, Sheet T91.) For other papers see.1134 of 1942 and back reference.

1542. FILTERS FOR CARRIER -FREQUENCY TELE-PHONE SYSTEMS.-Weijers. (See 1399.)

1543. OPPORTUNITIES FOR ELECTRONICS IN INDUS-TRIAL TEMPERATURE INSTRUMENTATION.-Behar. (Electronics, Dec. 1942, Vol. 15,No. 12, pp. 72-74 and 163.. 166.) Seealso p. 41.


1544 " ARZTLIC,HE ELEKTROKARDIOGRAPHIE " 1554.[Book Review].-Holzer & Polzer. (Elektrot.u. Masch:bau, 5th June 1942, Vol. 6o,No. 23/24, p. 26o.)

1545.

1546.

ELECTRONIC COUNTER FOR RAPID IMPULSES[primarily Action -Potential " Spikes "].-Wellman & Raeder. (Electronics, Oct. 1942,VOL 15, No. 10, pp. 74 and 140..142.)

THE USE OF THE PHILIPS PRESSURE INDICA-TOR IN GEOPHYSICAL PROSPECTING BYTHE RADIO METHOD.-Fritsch & Forejt.(Zeitschr. f. Geophys., No. 5/6, Vol. 17,1942, p. 217 onwards.) The method wasdealt with in a paper in another journal(3167 of 1942) ; the present paper is con-cerned with the geophysical applicationand the carrying out of the tests.

1547. THE . DYNAMIC MEASUREMENT OF THEMODULUS OF ELASTICITY OF SYNTHETICMATERIALS [using Mechanico - PneumaticNote Generator & Electro-Magnetic Vibra-tion Detector to find Resonance Point].-Frolich. (Physik. Berichte, 15th July 1942,Vol. 23, No. 14, p. 1447 : summary ofa Kunststoffe paper.)

1548. VIBRATION TABLE FOR DYNAMIC TESTS INTHE NOTE -FREQUENCY RANGE [for Investi-gations on the Frequency Characteristicsof Vibrometers : Sinusoidal Vibrations upto so kc/s]. - Meister. (Akust. Zeitschr.,No. 2, Vol. 7, 1942, pp. 51-56.)

1549. ELECTRICAL - RESISTANCE PROPERTIES OFDILUTE BINARY ALLOYS OF COPPER, SILVER,AND GOLD [including a Silver-ManganeseAlloy very suitable for the Measurementof Pressure].-Linde. (Zeitschr. f. Instr:kunde, Oct. 1941, Vol. 61, No. ICI, pp. 353-355 : summary only.)

1550. AN INSTRUMENT FOR MEASURING SURFACEROUGHNESS [using Rochelle -Salt Pick -Up].-Gravley. (Electronics, Nov. 1942, Vol. 15,No. II, pp. 70-73.)

1551. MEASURING INSTRUMENTS FOR STATIC EX-PANSION MEASUREMENTS : III - INSTRU-MENTS WITH ELECTRICAL INDICATION [Typesusing Barrier -Layer Photocell, Inductance -Change, Capacitance -Change (Too Sensitiveto Disturbances) & Steel -Wire (Maihak)Principles].-Lehr. (Arch. f. Tech. Messen,Oct. 1942, Part 136, V9 r 22-7, SheetsTio2-1o3.)

1552. ENERGY - STORAGE WELDING CONTROLS :PART 5 [Magnetic & Electrostatic StorageSystems to reduce Peak Power Demand onMains, for Resistance Welding of Non -Ferrous Metals]. - Rogers. (Electronics,Dec. 1942, Vol. 15, No. 12, pp. 63-67 and174, 175.)

1553. RADIO -FREQUENCY HEATING AIDS TIN -CANMAKERS [in Their Electroplating Method ofEconomising in Tin].-(Electronics, Dec.1942, Vol. 15, No. 12, p. 86.)

263

AVOIDING PATENT PITFALLS [particularlythe Necessity of keeping Records of Inven-tion & All Work relating Thereto].-Wild.(Electronics, Dec. 1942, Vol. 15, No. 'I2,pp. 78 and 159

1555. FROM THE PHYSICS AND TECHNIQUE OF THESHORTEST RADIO WAVES [Circuit Compo-nents, Valves, Velocity -Modulation Devices].-Tank. (Bull. Assoc. suisse des Elec.,3rd June 1942, Vol. 33, No. II, p. 315 on-wards.) A ten -page survey.

1336. PRECISION TIME CONTROL [for SynchronousClocks]. - Nat. Broadcasting Company.(See 1481.)

1557. PAPER ON THE EMPLOYMENT OF SUPERSONICWAVES IN MODERN TECHNIQUE.-OggiOni.(See 1452.)

1338. RESULTS WITH GEOELECTRICAL POLARISA-TION MEASUREMENTS [Method of Prospect-ing based on the Forces of Polarisation setup by A.C. or Periodic D.C., and alteringthe Electrical Properties : Special Advan-tages]. - (Physik. Berichte, 15thAug. 1942, Vol. 23, No. 16, p. ,1612: sum-mary, frOm Zeitschr. f. Geophys., No. 7/8,Vol. 16.)

1559. PROPAGATION METHODS OF RADIO PROSPECT-ING : B-ABOVE-GROUND METHODS [in-cluding Those of Grosskopf & Vogt and ofBurstyn].-Fritsch. (Arch. f. Tech. Messen,June 1941, Part 120, V65 -2I, Sheets T8o-

/1560. NEW DEVELOPMENTS IN THE ANALYSIS OFMACHINE VIBRATIONS [Short Survey ofVarious Methods (including Piezoelectric)leading to Inertilless System of LightBeam, Rigidly Attached Mirror, & Photocell/Amplifier/Cathode - Ray - Oscillograph Com-bination]. - De Gregorio. (Formation yDocumentation profesional [Madrid], Jan./Feb. 1942, Vol. r, No. I, pp. 46-47: sum-mary, from L'Energia Elettrica.)

1561. " QUARZDRUCKMESSGERATE HOMER EIGEN-FREQUENZ " [Pressure -Measuring Devices (forMotor Research) of High Natural Frequency :Vibration Properties, and Prevention ofTroubles due to Inertia : Book Review].-Gohlke. (Akust. Zeitschr., July 1942, Vol. 7,No. 4, pp. 168-17o.)

1562. A METHOD FOR THE EXACT REGULATION OFFIXED ROTATIONAL SPEEDS OF AN AXLE.-Kirschstein. (See 1484.)

1563. THE ARABIC TELEGRAPH ALPHABET. -Worrell. (QST, Jan. 1943, Vol. 27, No. 1,P. 34)

1564. NEW ENEMY -SPONSORED EUROPEAN POSTSAND TELECOMMUNICATIONS UNION.-(Elec.Review, I 2th March 1943, Vol. 132, No.3407, p. 362.)


1565. GENERATORS FOR SHORT-WAVE THERAPY[Choice of Frequency & Power : GeneratorCircuits : Matching to the Patient Circuit :Details of Actual Equipment : etc.].-J. Fransen & J. M. Ledeboer. (PhilipsTech. Rundschau, May 1942, Vol. 7, No. 5,p. 147 onwards.)

1566. THE STATISTICAL AVERAGE IN THE MEASURE-- MENT AND USE OF DRY -PLATE -RECTIFIER

PLATES.-Maier. (Arch. f. Tech. Messen,May & June 1941, Parts 119 & 120, Z52-3& 4, Sheets T76-77 & 89-90.)

1567. THE USE OF SECONDARY ELECTRON EMIS-SION TO OBTAIN TRIGGER OR RELAY ACTION.-Skellett. (fawn. Applied Phys.,' Aug.1942, Vol. 13, No. 8, pp. 519-523.) Asummary was dealt with in 676 of February.

1568. ELECTRICAL REMOTE CONTROL [TypicalApplications, Circuit Initiating Methodsand End-Actions]..-Dory & Galton. (Elec-tronics, Nov. 1942, Vol. 15, No. II, pp. 60-64and 173, 174.) Continued in Dec. issue,pp. 57-60 and 167..169.

1569. TEMPERATURE MEASUREMENT AND CONTROLBY ELECTRONICS [General Account with aDescription of Several Electronic Instru-ments].-Walsh. (Electronics, Oct. 1942,Vol. 15, No. 1o, pp. 56-61 and 94..98.)

1570. A NEW METHOD OF MEASURING THE VELO-CITY OF A GASEOUS JET WITH STEADYFLOW [Electric Circuit with Two Gridsacross the Flow, and a Condenser whosePotential is Measured by Sensitive Instru-ment].-Yadoff. (Physik. Berichte, 1st July1942, Vol. 23, No. 23, p. 1296: summary,from Comptes Rendus [Paris], No. 3, Vol.214,1942, p. 102 onwards.)

1571. ON THE PRODUCTION, PREVENTION, AND DAN-GEROUSNESS OF ELECTROSTATIC CHARGES [inIndustrial Processes : including Data onMinimum Spark Energies for Ignition ofVarious Vapour-Air Mixtures]. - Nitka.(Physik. Berichte, 15th June 1942, Vol. 23,No. 12, pp. 1234-1235.)

1572. ELECTRIFICATION BY BUBBLING : IV - EF-FECT OF VISCOSITY AND OF TEMPERATURE[Further Development of Researches onElectrification by passing Gases throughLiquids].-Lovera & Pochettino. (Physik.Berichte, 15th July 1942, Vol. 23, No. 14,pp. 1423-1424: summary of Italian paper.)

1573. APPLICATIONS OF CATHODE-RAY TUBES [inUltra -High -Frequency Technique : Survey].-Dudley. (Electronics, Oct. 1942, Vol. 15.No. io, pp. 49-52 and 154, 155.)

1574. THE CATHODE-RAY OSCILLOGRAPH IN POLARO-GRAPHY [Application of Electronic Methodsto Chemical Analysis].-Jones. (ElectronicEngrg, Feb. 1943, Vol. 15, No. 18o, pp.367-371.)

1575.

May, 1943

MEASUREMENT OF THE MOISTURE CONTENTOF SALTS WITH HIGH ELECTRICAL CON-DUCTIVITY, BY AN ULTRA -HIGH -FREQUENCYMETHOD, SUITABLE FOR RAPID ANALYSIS,BASED ON THE DEPENDENCE OF DIELECTRICCONSTANT ON MOISTURE CONTENT.-Worasko & others. (Physik. Berichte, 1stJuly 1942, Vol. 23, No. 13, pp. 1302-1303 :summary of a Russian paper.)

1576. RADIO -FREQUENCY HEATING SPEEDS PLY-WOOD BONDING. [by Extending the Use ofHot Press Methods to Thick Plywood].-(Electronics, Nov. 1942, Vol. 15, No. II,p. 79.) Cf. 975 of March.

1577. THE USE OF WOOD, DRIED IN A HIGH -FREQUENCY (PARTICULARLY AN ULTRA -HIGH -FREQUENCY) FIELD, FOR ELECTRICAL INSULA-TION. - Siemens - Schuckert. (Zeitschr. f.Fernmeldetech., 16th Feb. 1942, Vol. 23,No. 2, p. 28.) D.R.P. 704 667.

1578. ON THE INFLUENCE OF SOME ELECTRICALPARAMETERS ON THE CHEMICAL REACTIONOF ETHYLENE IN A HIGH -FREQUENCYCORONA DISCHARGE [converting into Acety-lene : Absence of Dependence on Wave-length, between 190 & 500 m : Dependenceon Current Strength (with Empirical For-mula)].-Eiduss & Netschafewa. (Physik.Berichte, 15th June 1942, Vol. 23, No. 12,p. 1241.).

1579. A SUMMARY OF X-RAY SATELLITES.-Hirsh.(Reviews of Modern Phys., Jan. 1942, Vol. 4,No. 1. pp. 45-54.)

1580. X -RAY -TUBE CATHODES [Effect of DesignDetails on Performance of Tube].-Long.(Electronic Eng:g, Feb. 1943, Vol. 15, No. 18o,pp. 364-366.)

1581. THE BETATRON [Possible Use of New PowerfulX -Ray for Medical Purposes].-(Science,4th Dec. 1942, Vol. g6, No. 2501, p. 10.)For the Betatron see 153 of January.

i582. A NEW SONIC AND SUPERSONIC TRANS-MITTER FOR THE PRODUCTION OF HIGHINTENSITIES IN GASES [and Liquids].-Ehret & Hahnemann. (See 1453.)

1583. LOCALISATION OF THE DISCHARGE IN GEIGER-MOLLER COUNTERS.-Wilkening & Kaune.(Phys. Review, 'St/15th Dec. 1942, VOL 62,No. 11/12, pp. 534-537.)

1584. COUNTER TUBES AND THEIR USE IN MEAS-URING TECHNIQUE.-Neuert. (Arch. f. Tech.Messen, June 1941, Part 120, Jo76-r, SheetsT85-86.) For later papers see 2918 of1942.

1585. RESEARCHES ON THE CHEMICAL REACTIONSOF ELECTRICAL DISCHARGES : XXVII-NOTE ON THE PRODUCTION OF HYDRAZINEBY MEANS OF THE ARC, OF HIGH OR LOWFREQUENCY, IN NITROGEN - HYDROGENMIXTURES OR AMMONIA.-Briner & Hoefer.(Physik. Berichte, 15th Aug. 1942, Vol. 23,No. 16, p. 1564: summary, from Helvet.Chim. Acta, 1942.)

rt


1586. ON THE FUNCTIONING OF AN X-RAY TUBESUBMITTED TO VOLTAGE PULSES [for VeryShort Exposures : the Magnitude of thePulse Currents : Equipment : the Practica-bility of making Stroboscopic Records ofPeriodic Processes]. - Herreng. ,(ComptesRendus [Paris], No. 9, Vol. 214, 1942, p. 421onwards : summary in Physik. Berichte,15th Aug. 1942, Vol. 23, No. 16, p. 1570.)

1587. RADIATION MEASUREMENT IN THE ULTRA-VIOLET : I-GENERAL CONSIDERATIONS ANDMEANS OF MEASUREMENT [Thermal, Photo-electric, Photochemical, Photographic,Fluorescence Methods : for Wavelengths200-400 Nanotnetres (i nm=-- to -9 m) : etc.].Seitz & Meyer. (Arch. f. Tech. Messen, Oct.1942, Part 136, J323-1, Sheets Tio8-109.)

1588. CONTRIBUTION TO OUR KNOWLEDGE OF THEVARIATION OF TRANSPARENCY OF COLLOIDALIRON IN A MAGNETIC FIELD.-Frongia &Agus. (See 1173 of April.)

1589. ON AN OPTICAL EFFECT IN [Flowing] COL-LOIDS, and ON THE TYNDALL LIGHT OFCOLLOIDS IN A MAGNETIC FIELD [Experi-ments with Colloidal Silver, Bravais' Iron,etc : Effects due to Imperfect Sphericity ofParticles].-Petralia. (Physik. Berichte, 1stAug. 1942, Vol. 23, No. 15, p. 1484 : pp.1484-1485: summaries of Italian papers.)

1590. ZEISS T -OBJECTIVE [with Reflection -Reduc-ing Films : Design Calculations, & Advan-tages additional to Increased Transmission].-Leistner. (Physik. Berichte, 15th July1942, Vol. 23, No. 14, p. 1427: summary,from Photogr. u. Forsch.)

159 I . TELESCOPES FOR MEASURING INSTRUMENTS[Combination of Refracting & ReflectingTelescope Systems to give a " Lens -Re-flector Telescope of Very Short Length butLarge Aperture].-Wild. (Zeitschr. f.kunde, Oct. 1941, Vol. 61, No. to, pp. 355-356 : summary only.)

1592. THE DETERMINATION OF AN UNKNOWNFREQUENCY FROM A PHOTOGRAPHIC RECORD.-Scott. (Electronic Eng:g, Nov. 1942,Vol. 15, No. 177, p. 243.)

1593. " SICHTBEOBACHTUNGEN VOA/ METEOROLO-GISCHEN STANDPUNKT " [Book Review :Visual Ranges & the Effects of Moisture,Dust, Mist, etc.].-Lohle. (Zeitschr, f. tech.Phys., No. 9, Vol. 23, 1942, p. 242.)

Referring to the chapter on telephotographyand infra -red photography, the reviewer (Gaertner)disagrees with the author's contention that theincreased ranges obtained by infra -red photographyare chiefly due to the " chlorophyll " effect-anincreased contrast between the green of foliageand its surroundings : for his own personal observa-tions have proved the better haze -penetrating powerof the infra -red rays in the case of objects havingno foliage near them.

265

1594. A PHOTOELECTRIC DENSITOMETER [primarilyto ensure Uniformity of Gray -Scale Densitiesof Colour -Photography Negatives].-Smith.(Electronics, Dec. 1942, Vol. 15, No. 12,pp. 79 and 82.)

1595. MEASURING FAINT LIGHT [Application ofSelenium Cells]. - Baker. (Elec. Review,12th March 1943, Vol. 132, No. 3407, PP.343-345)

1596. THE PRESENT PRACTICAL RANGES OFAPPLICATION OF PHOTOELECTRIC DEVICES[Photoelectric Cells & Multipliers, Photo-electric Resistances, & Dry -Plate Photo-cells : Life, and Loading Limit : MinimumIllumination : Frequency CharacteristicsTemperature Limits]. - Kluge. (Arch. f.Tec)I. Messen, May 1941, Part x19,Sheets T74-75.)

1597. ELECTRONIC CONTROL FOR ONE-WAY TUNNEL[Traffic Lights controlled by PhotoelectricCounter, Ventilation by Carbon -MonoxideElectronic Equipment].-(Electronics, Dec.1942, Vol. 15, No. 12, p. 82.)

1598. PHOTOELECTRIC CONTROL OF COAL LARRY[for Even Distribution of Fuel to Hopperfeeding Chain -Grate Boiler].-(Electronics,Dec. 1942, Vol. 15, No. /2, pp. 84 and 86.)

1599. A PHOTOMETRIC METHOD OF MEASURINGTHE SURFACE QUALITY OF FINISHED PRO-Ducrs.-Heyes & Lueg. (Physik. Berichte,ist Aug. 1942, Vol. 23, No. 15, p. 1521 :summary, from Werkstattstech., No. 5/6,Vol. 36, 1942, p. 98 onwards.)

1600. NEW MEASUREMENTS OF THE VISUAL CON-TRAST THRESHOLD [for Field Densities betweento & ro-7 Stilb : Comparison of Valuesfor Monocular and Binocular Vision : etc.].-Siedentopf. (Physik. Berichte, ,5th Aug.2942, Vol. 23, No. 16, p. 159o.)

1602. THREE-DIMENSIONAL SEEING. [" Stereo -Stro-boscope", primarily to assist in Choice ofGround for Forced landing : Combination ofTwo Telescopes alternately blocked & openedby Perforated Rotating Disc : Base Length of3 m for Flying Speed of 36o km/hour].-Bartow Beacons, Inc. (Physik. Berichte,IS/ Aug. 1942, Vol. 23, No. 15, pp. 1465-1466: summary, from Flight, 1942.)

1602. POST-WAR INDUSTRIAL RESEARCH [Editorial].-(Engineering, 19th Feb. 1943, Vol. 155,pp. 151-152.)

1603. THE PROFESSION OF ENGINEER : THE DIFFER-ENT INTELLECTUAL CAPACITIES AND MORALQUALITIES WHICH FORM THE GOOD EN-GINEER, and REMARKS AND CONTRIBUTIONSFROM PRACTICAL EXPERIENCE.-Silberet :Schiesser. (Bull. Assoc. suisse des Elec.,13th Jan. 2943, Vol. 34, No. I, pp. 1-7in French : pp. 7-9 : in German.)


1604. SCIENTIFIC RESEARCH AND DEVELOPMENT[Research in War -Time : Comparison ofOrganisation in United Kingdom, U.S.A.,and U.S.S.R.].-(Nature, loth Feb. 1943,Vol. 151, No. 3825, pp. 203-207.)

1605. THE FULLER UTILISATION OF SCIENTIFICRESOURCES FOR TOTAL WAR.-Rosebury.(Science, 25th Dec. 1942, Vol. 96, No: 2504,PP. 571-575.)

1606. PROGRESS IN ELECTRICAL RESEARCH [Reviewof 22nd Annual of the British Elec-trical & Allied Industries Research Associa-tion].-(Engineering, 19th Feb. 1943, Vol.155, pp. 156-157.)

1607. " ELECTRICAL TECHNOLOGY FOR TELE-COMMUNICATIONS " [Book Review].-Date.(Elec. Review, I 2th March 1943, Vol. 132,No. 3407, p. 348.)

1608. " THE RADIO AMATEUR'S HANDBOOK, SPE-CIAL DEFENSE EDITION " [Book Review].-American Radio Relay League. (Proc.I.R.E., Jan. 1943, Vol. 31, No. I, p. 43.)For another review see 3444 of 1942.

1609. " RADIO ENGINEERING " [Book Review].-R. C. Norris. (Elec. Review, 5th March 1943,Vol. 132, No. 3406, p. 33i.)

1610. " ELEMENTARY ELECTRICITY FOR RADIOSTUDENTS " [Book Review].-Flood. (Wire-less World, March 1943, Vol. 49, No. 3,

P. 71.)161i. " ELECTRICAL FUNDAMENTALS OF COMMUNI-

CATION (QST,Jan. 1943, Vol. 27, No. 1, p. 64.)

1612. " ELECTRICITY AND MAGNETISM 's [BookReview].-Gilbert. (Journ. Applied Phys.,Dec. 1942, Vol. 13, No. - 12, pp. 745 and746.)

1613. " EXPERIMENTAL RADIO ENGINEERING " [2ndEdition : Book Review].-Rapson, assistedby Ackermann. (Electrician, 26th Feb.1943, Vol. 130, No. 3378, p. 220.) Describesexperiments suitable for college laboratories.

1614. ELECTRON - TUBE TERMINOLOGY. - White.(See 1423.)

1615. " MITTEILUNGEN AUS DER FORSCHUNGSAN-STALT DER DEUTSCHEN REICHSPOST (RPF) :BAND VII, 1942 " [Book Review].-(Hochf:tech. u. Elek:akus., Nov. 1942, Vol. 6o, No.5, p. 148.)

For the twelve papers here listed see (1) 2963 o:1941 ; (2) 3091 of 1941 ; (3) 356 & 926 of 1942 ;(4) 376 of 1942 ; (5) 1926 of 1942 ; (6) 1953 of 1942 ;(7) 1927 & 3526 of 1942 ; (8) 3552 of 1942 ; (9) 964of 1942 ; (io) " Technical Applications of aGround -Conductivity Meter " ; this is probablythe short T.F.T. paper mentioned at the beginningof the long abstract 2955 (see also 2956) of 1942 ;II) 3586 of 1942 ; and (12) 1065 of 1942.

May, 1943

1616. " A SHORT COURSE IN TENSOR ANALYSIS FORELECTRICAL. ENGINEERS " [Book Review].-Kron. (Journ. Applied Phys., Jan. 1943,Vol. 14, No. I, p. 36.)

1617. " MATHEMATICS FOR ELECTRICIANS 'ANDRADIOMEN " [Book Review].-Cooke. (Proc.I.R.E., Jan. 1943, Vol. 31, No. I, p. 41.) ,

1618. THE USE OF A MECHANICAL SYNTHESISER TOSOLVE TRIGONOMETRIC AND CERTAIN TYPESOF TRANSCENDENTAL EQUATIONS AND FORTHE DOUBLE SUMMATIONS INVOLVED INPATTERSON CONTOURS.-Leroy Brown &Wheeler. (Journ. Applied Phys., Jan. 1943,Vol. 14, No. I, pp. 3o-35.)

Authors' summary :-"A method is shown where-by equations in trigonometric form and types oftranscendental equations, which can be expandedinto rapidly converging series, may be solved by amechanical and graphical process. A mechanicaland graphical method is also shown which effects thedouble summation of Fourier series used in thePatterson method for the determination of inter-atomic distances in crystals. Graphs are given toillustrate each of the processes discussed.

1619. A NEW ANALYTICAL METHOD OF SOLVINGVAN DER POL'S AND CERTAIN RELATEDTYPES OF NON-LINEAR DIFFERENTIAL EQUA-TIONS, HOMOGENEOUS AND NON -HOMO -GENEOUS.-Shohat. (Journ. Applied Phys.,Jan. 1943, Vol. 14, No. 1, pp. 40-48.)

Author's summary :-" A simple analytical treat-ment is developed for the differential equation(d'u/dt')-e (1 - u2) (duldt)-Fu=o (van der Pol), inorder to study the behaviour of its solution assumedbounded in (o, 00 ). It is shown, without any furtherassumption, that, if u (t) can be approximatedclosely, for t> o, by a simple oscillation with ampli-tude 2. This is a more precise form of a statementdue to van der Pol. It is further shown that u'(t) +

t> o, o< e ' I. The method and resultsare then extended to the more general equation(d2u/dt2)-cF(u)(duldt)-Fu=o, in particular, forF(u)=I-u*, ---eau-u', a=constant, also to thenon -homogeneous equation (Puldt2)--EF(u)(duldt)-Fu=a given function of t. The analytical resultsobtained in this paper show a remarkable agreementwith those obtained for the same equations bymechanical means (on the differential analyser)".

1620. APPLICATION OF VECTOR ALGEBRA [Element-ary Article on Application of Vector Methodsto Radio Problems]. - Rensselaer. (Radio[New York], Dec. 1942, No. 275, pp. 14 -16 -and 44, 45.)

GOODS FOR EXPORTThe fact that goods made of raw ma-terials in short supply owing to warconditions are advertised in this journalshould not be taken as an indicationthat they are necessarily available for

export.


iir

..KA

ETALASTIK BUSHES are typical examples of the scientific application

of Rubber -to -Metal Weld.

The rubber, being welded to the metal is used to the maximum efficiency,

and the whole volume can be stressed in every conceivable combination of

torsion, shear, compression or tension.

Thus it is that higher loads can be applied or smaller bushes employed.

The range of METALASTIK Rubber -to -Metal Weld Bushes is practically

unlimited and covers such widely differing loads and frequencies as are

called for in Automobiles, Ships, Railway Engines and Rolling Stock,

Machine Tools, and in Agricultural, Textile and Domestic Machinery.

METALASTIK LTD.METALASTIkRUBE .E R -TO -Mt -

CONSULT US ON ALL YOUR VIBRATION PROBLEMS

LEICESTER

iv WIRELESSENGINEER

Cox iliell?//1IN COMMUNICATIONS

To -day, whereverBullers Low LossCeramics are used, re-ception is clearer andbetter than ever before.

Years of research in Bullers Labora-

tories have developed these materials to such

a high level of efficiency that Dielectric Loss

is virtually eliminated. Bullers CeramiCs are

now playing a vital part in maintaining com-

munications under all conditions.

BULLERS LIMITEDTHE HALL, OATLANDS DRIVE, WEYBRIDGE, SURREY

Telephone : Walton -on -Thames 2451Manchester Office : 196 Deansgate, Manchester

BullersLOW LOSS

May, /943

Madein ThreePrincipal

Materials:

FREQUELEX. An InsulatingMaterial of Low Dielectric Loss.

For Coil Formers, Aerial Insulators,Valve Holders, etc.

PERMALEX. A Highmittivity Material. For thestruction of Condensers of

smallest possible dimensions.

Per-con -

the

TEMPLEX. A Condensermaterial of medium permittivity.For the construction of Condensershaving a constant capacity at all

temperatures.

CERAMICS

Printed in England for the Publishers, ILIFFE & SONS LTD., Dorset House, Stamford Street, London, 8.E.1,by The Cornwall Press Ltd., Paris Garden, Stamford Street, London, S.E.1.

WIRELESS ENGINEER - WorldRadioHistory.Com · 2019. 7. 17. · RADIO ENGINEERING LABORATORIES INC....

Documents

Transcript of WIRELESS ENGINEER - WorldRadioHistory.Com · 2019. 7. 17. · RADIO ENGINEERING LABORATORIES INC....