worldradiohistory.com · I X X X X X K X X XX X X X X O August, 1933 SERADEX RESISTORS" Wireless...

WIRELESSENGINEER

ANDEXPERIMENTAL WIRELESS

NUMBER 119 VOLUME X AUGUST 1933

A JOURNAL OF

RADIO RESEARCHAND

PROGRESS

PUBLISHED BY

ILIFFE & SONS LTD.DORSET HOUSE STAMFORD STREET LONDON S.E.I

I

XXXXXKXX

XX

X

XXX

O

August, 1933

SERADEXRESISTORS

" Wireless World "reports :-

" Very close agreementwith marked values-excellent value formoney...

Illustrated lists forstamp.

Type M-150 12 wattrating - 8d. each.

Type M-300 3'0 wattrating 1 /3 each.

Also in Glasswound types,prices on application.

From your dealer or

TREVOR PEPPER,

Wake Green Road,

BIRMINGHAM.

ii THE WIRELESS ENGINEER &

JOURNAL OFSCIENTIFIC INSTRUMENTS

CONTENTS OF THE JULY NUMBER.FIELD INSTRUMENTS. By D. C. Gall.STROBOSCOPIC DEMONSTRATION OF TONES. By

W. Anderson, B.Sc.. and H. Lowery, Ph.D., F.Inst. P.THE PERFORMANCE OF RELAYS. By G. A. Tomlinson,

D.Sc.APPLICATION OF COMPENSATED VALVE -VOLT-

METER TO MEASUREMENTS OF GLASS ELECT-RODE POTENTIALS. By A. S. McFarlane, M.A., B.Sc.,M.B.

POLARIZING LOUVRES AND THEIR APPLICATION.By C. F. Smith, F.R.P.S.

A NEW FORM OF BAROSTAT. By F. A. Gould, M.A.,and J. C. Evans, B.Sc.. Ph.D.

A METHOD OF TAKING X-RAY PHOTOGRAPHS OFCRYSTALLINE POWDERS AT THE TEMPERATUREOF LIQUID AIR. By J. A. Santos and J. West.

NEW INSTRUMENTS:A SODIUM LAMP FOR LABORATORY USE.NEW GALVANOMETERS.

LABORATORY AND WORKSHOP NOTESTHE SPECTROSCOPIC EXAMINATION OF THIN

METALLIC FILMS. By A. Harvey, Ph.D., F.Inst.P.,and E. S. Dreblow.

A SIMPLE SUBSTITUTE FOR A MICROMETER EYE-PIECE. By H. E. Watson, D.Sc., F.I.C.

CHECKING THE BALANCE OF A THREE-PHASELOAD. By G. W. Stubbings, B.Sc., A.M.I.E.E.,F. I nst. P.

AWARD OF THE DUDDELL MEDAL, NEW BOOKS,CATALOGUES, NOTES AND COMMENTS.Sing!e copies may be obtained from the publishers, price

2s. 6d. (post free 2s. 8d.). The subscription rate for the yearlyvolume is £.1 10s. net (post free) payable in advance. Sub-scriptions should be sent to the publishers.

CAMBRIDGE UNIVERSITY PRESSFetter Lane, London, E.C.4 England

WIRELESS( DIRECTION FINDINGand DIRECTIONAL RECEPTION

By R. KEEN, B.Eng. (Hons.)

SECOND EDITION: (1927)

This volume deals with the principles of the subject and the con-structional details of direction finding installations, and includes someinformation concerning aircraft installation. It describes the principlesof Direction and Position Finding in such a way that the subjectmay be grasped easily by engineers tackling this field of wireless workfor the first time. Numerous photographs and diagrams are included.

Price 21/- net By post 21/9From leading booksellers or direct from the publishers :

ILIFFE & SONS LIMITED,DORSET HOUSE, STAMFORD STREET, LONDON, S.E.1.

W.E.g.

K

Kindly mention " The Wireless Engineer" when replying to advertisers.

EXPERIMENTAL WIRELESS A August, 1933

THE LAST WORDIN BEAT -FREQUENCY OSCILLATORS.

IMPROVED MODEL, TYPE 613-B.This is a completely portable, battery -operated Beat -Fre-

quency Oscillator which possesses the very desirable features ofexceptionally good waveform, great frequency stability, and alarge, easily -read, directly calibrated and very " open - scale.

The Oscillator has built into it a very critically tuned reedvibrating at 100 cycles. By setting the main scale to 100 cycles,and adjusting the microdial, any slight drift from calibrationmay immediately be corrected. After a preliminary " warmingup," the oscillator remains exceedingly stable under uniformtemperature conditions. The output frequency will remainconstant within a few cycles over a period of several hours.

All Batteries are contained within the cabinet. The Instru-ment, complete with Tubes and Batteries, weighs 45 lbs.Dimensions (cabinet) 22" x 12" x 9" : Panel, 19" x fir.

BRIEF SPECIFICATION :OUTPUT VOLTAGE : Approximately 13 volts.OUTPUT IMPEDANCE : 5,000 ohms.FREQUENCY RANGE : Calibrated between 5 cycles and

10,000 cycles per sec. It will actually deliver power atfrequencies slightly lower and higher. Each instrument isindividually calibrated.

MAIN DIAL : Calibrated over 270°, 6" diameter, giving easily -read scale 14" long. The spacing is approximatelylogarithmic.

METERS: Filament voltmeter : Plate Milliammeter.PRICE : The price is very reasonable.

DELIVERY : Ex -limited U.K. Stocks.

For complete information, see our Catalogue " Pp. 71/72.

CLAUDE LYONS LTD., 76, OLDHALL ST., LIVERPOOL.'PHONE : CENTRAL 4641. 'GRAMS : " MINMETKEM, LIVERPOOL."

LONDON : 40, BUCKINGHAM GATE, WESTMINSTER, S.W.1'PHONE: VICTORIA 7595. 'GRAMS : " MINMETKEM, SOWEST, LONDON."

Kindly mention " The Wireless Engineer" when replying to advertisers. A

August, 1933

Manufacturers ofA 2 THE WIRELESS ENGINEER

STANDARD FREQUENCY EQUIPMENTSwith Crystal or Fork Control.

Accurate measurementsdemand

reliable instruments.

If you requireeither or both of these

write forfull particulars.

WAVEMETERS,

INDUCTANCES and

CONDENSERSin Precision, Laboratory and Commercial grades.

He W. SULLIVAN, LTD.,Leo St., London, S.E.15.

IMPROVEDVOLTAGECO -EFFICIENTDecided improvements in the voltage co-efficient of ERIE RESISTORS have been madeby our Engineering Department. On the higherresistance values the improvement reaches ashigh as so%, making those values far superiorto any units previously offered.

For example, on a No. 4 Resistor of r megohmresistance the drop from zo to zoo volts averaged5t% and on the improved units (as shown onchart) is now about 21%. Again ERIERESISTORS lead with an improved product.

0

1";

Z

3

4,4

05

60

94?5,90 poiea, &or

" 3 3 to

rr

rs

95 .50 95 10 123 15 175 147 823 13 215 30 3135 34 40RESLITANCE 1AI MEGORMS

RIE RESISTORS

ERIESPARK

PLUGSUPPRESSORSeffectively eliminate interferencefrom the Spark Plugs and Distrib-utors. The resistance unit isenclosed in a ceramic tube to pro-tect it from oil, dirt and moisture,and also from mechanical breakage.May we send you Samples and Prices ?ERIE RESISTOR LTD.,Waterloo Rd., Cricklewood,'Phone: Gladstone 2258.

XXXXXXXXXXXXXXXXXXXXXXXM

" Calculation and Measurementof Inductance and Capacity,"

by W. H. Nottage, B.Sc. (1924).The mare generally useful formula, and methodsof measurement of inductance and capacity arehere brought together in a convenient form.224 pages. Over So illustrations and diagrams.Price 7/6 net, by post (home and abroad) 8/- from

ILIFFE AND SONS LTD.,Dorset House, Stamford Street, London, S.E.r,

and leading booksellers. W.E.I2

XXXXXXXXXXXXXXXXXXXXXXXX

PATENTS.

The Proprietors of British PatentsNos. 269,705 and 271,419, relating to" Electric Discharge Devices," aredesirous of entering into negotiationswith interested parties with a view tothe exploitation of the inventions eitherby sale of the Patent Rights or by thegranting of Licences on reasonableterms. All enquiries should be addressedto Wm. Brookes & Son, No. 1, QualityCourt, Chancery Lane, London, W.C.2.

[0274

PATENT AGENTS."A. MATHISEN, B.Sc., PATENTAGENT. Specialist in obtaining patentsfor Television and Radio Inventions.Working drawings, circuit diagramsprepared. Exploitation Advice. FirstAvenue House, High Holborn, London,W.C.I. Holborn 895o." [0268

BOOKS on WIRELESSWrite for complete list toILIFFE & SONS LTD.,

Dorset House, Stamford St.,London, S.E.1 W.E.17

Kindly mention " The Wireless E ngineer" when replying to advertisers.

The

WIRELESS ENGINEERand EXPERIMENTAL WIRELESS

A Journal of Radio Research & ProgressEditorHUGH S. POCOCK

Technical EditorProf. G. W. 0. HOWE D.Sc., M.I.E.E.

VOL. X. No. 119

AUGUST 1933

CON T EN T S

EDITORIAL 409

DISTORTION CANCELLATION IN AUDIO AMPLIFIERS. ByW. BAGGALLY 413

THE ACOUSTICAL PERFORMANCE OF A CONE -TYPE LOUDSPEAKER. By D. A. OLIVER, B.Sc. 420

THE USE OF THE CATHODE-RAY OSCILLOGRAPH AT ULTRA-HIGH FREQUENCIES. By Dr. H. E. HOLLMANN 430

NAVAL DIRECTION -FINDING. ABSTRACT OF A PAPER READBEFORE THE WIRELESS SECTION I.E.E. By MEMBERSOF THE STAFF OF H.M. SIGNAL SCHOOL

ABSTRACTS AND REFERENCES

SOME RECENT PATENTS

434

437

464

Published Monthly on the first of each monthSUBSCRIPTIONS Home and Abroad : One Year, 32/-, 6 Months, 16/-. Single Copies, 2/8 post free

Editorial, Advertising and Publishing OfficesDORSET HOUSE, STAMFORD STREET, LONDON, S.E.

Telegrams : " Experiwyr Watloo London " Telephone : Hop 3333 (50 lines)

Branch OfficesCOVENTRY BIRMINGHAM MANCHESTER GLASGOW

19 Hertford Street Guildhall Bldgs., Navigation St., 2 26o Deansgate, 3 260 Renfield Street, C.2Telegrams: " Cyclist, Coventry" Telegrams: Autopress, Birmingham" Telegrams "Iliffe, Manchester." Telegrams: Iliffe, Glasgow"Telephone: 5210 Coventry Telephone: Midland 2970(9 lines) Telephone : Blackfriars 4412(4 lines) Telephone: Central 4867

The Editor invites the submission of articles with a view to publication. Contributions which arenot exclusive should be so described when submitted. MSS. should be addressed to the Editor," The Wireless Engineer and Experimental Wireless," Dorset House, Stamford Street, London,Especial care should be taken as to the legibility of MSS. including mathematical work.

A3

_'August, 04 THE WIRELESS ENGINEER

The Dubilier TypeBB Condenser is ahigh -efficiency condenser in mouldedbakelite case.Working voltage200 D.C.(peakoTestvoltage 400 D.C.Capacities from.09mfd to 4.0mfd. Prices 9from - -

Only withDUBILIER

can you becertain of

DEPENDABILITYA.8

DUBILIER CONDENSER CO. (1925) LTD.Ducon Works. Victoria Road, North Acton, W. 3

AUTO - TRANSFORMERSwith HIGH OUTPUTS at LOW PRICES

DAVENSET Engineers have produced High Output Transformersat extremely Low Prices. NO SACRIFICE IN QUALITY HASHOWEVER BEEN MADE. DAVENSET add to their compre-hensive range of double -wound transformers, a further series ofAuto -Transformers, designed with Universal Inputs of 200/250volts, 40/100 cycles, and with Outputs of 100/110 volts.

List No. Max. Loading. Retail Price.18 35 Watts 17/619 75 22/620 150 30/.21 350 50/.

DAVENSET Transformers are manufactured to your own specificrequirements, whilst our Technical Department is always at yourservice to advise and suggest the simplest, yet most practicalschemes for overcoming your transformer or rectification problems.

F REEWrite for a copy of "Mains Transformers and PowerSmoothing Chokes for Radio and Industrial Purposes."

PARTRIDGE, WILSON & CO.Specialist manufacturers of Static Rectifiers and Accumulator Charging Plant.

Dept. 32, Davenset Works, Leicester.Scottish Branch: 200, St. Vincent Street, Glasgow, C.2.

[Kindly mention " The Wireless Engineer" when replying to advertisers.

409

THE

WIRELESSENGINEER

ANDEXPERIMENTAL WIRELESS

Vol.. X. AUGUST, 1933. No. 119

EditorialElectromagnetic Induction

LTHOUGH it is now over a hundredyears since electromagnetic inductionwas discovered and systematically

investigated by Faraday, and although thecalculation of the electromotive force in-duced in any given case is now regarded as asuitable exercise for an elementary student,it is surprising how soon one gets into diffi-culties if one tries to get a clear conceptionof the phenomena involved in the simplestcases.

To appreciate this we need only considerthe old question of the cylindrical bar magnetrotated about its axis. Does its magneticfield rotate with it or not ? The magneticcondition of the space surrounding themagnet is undergoing no change whatevereither in magnitude or direction, and thereappears to be no justification for picturingthe space as filled with concrete lines ofmagnetic induction fixed like bristles insome mysterious way to the magnet andforced to rotate with it. In the case of anair -cored solenoid such a conception appearseven more far-fetched.

If one illuminates a table by means of asource of light, the candle -power of which issymmetrical about the vertical axis, will

the illumination of the table rotate when thelamp is rotated ? Or again, if one places auniformly heated disc on the table, thussetting up a symmetrical temperature dis-tribution, will this temperature distributionrotate when the disc is rotated ? Neitherthe illumination nor the temperature at anypoint undergo the slightest change as theresult of the rotation, and the question pro-pounded appears meaningless. That theseare scalar magnitudes whereas the magneticinduction is a vector does not affect thequestion ; what we have said of temperatureis equally true of the vector temperature -gradient.

If the source of light, or heat, or magnetism,is not uniform about the axis of rotation,the result of the rotation might be regardedby a superficial observer as an obvious proofof the rotation of the illumination, or tem-perature, or magnetic induction, but a closerconsideration will show that these conditionsare merely undergoing such changes atevery point that their inequalities of dis-tribution rotate. The illumination, or tem-perature, or magnetic induction at A maynow be equal to what it was a moment agoat B, but it does not follow that the illumina-

August, 1933

tion, or temperature, or magnetic inductionwhich now exists at A is actually that whichwas at B, and that it must necessarily haverotated ; in fact, such a statement appearsmeaningless.

We can see then no escape from theconclusion that the question whether themagnetic field rotates with the magnet ornot can have no answer because it has nomeaning. Were it otherwise the magneticinduction at a point would not be completelydefined by the magnitude and direction .of B.

Let us assume inthe first place that thebar magnet in Fig.is at rest and that themetal disc is rotatingin the direction shown.

The electrons in the disc are moving throughmagnetised space and experience conse-quently an outward radial force crowdingthem towards the edge and charging itnegatively, while leaving the central partcharged positively. If now on rotating themagnet also in the same direction one findsthat the charging of the edge of the disc isless. than it was before one may be temptedto regard this as a proof that the disc and themagnetic field are now both rotating so thattheir relative speed is reduced, notwith-standing what we have said as to the mean-inglessness of the statement.

When the magnet rotates its free elec-trons rotate and they are also in a magneticfield. The electrons of the magnet nowexperience forces tending to remove themfrom some parts of the magnet and crowdthem into others not only on the surfacebut throughout the material.

An ExplanationThe magnet thus becomes charged as

shown in Fig. 2 (a), and these charges producean external electric field which tends toinduce an electric distribution in the discof opposite sign to that of Fig. i. When thedisc is at rest it is thus charged electro-statically by the rotating magnet, and whenthey both rotate in the same direction theelectrostatic forces on the disc are more orless neutralised by the electromagneticforces.

The phenomena can thus be explained bythe movement of electrons in magnetisedspace without doing violence to one's powers

(1s- -N )Fig. 1.

410 THE WIRELESS ENGINEER &

of imagination by endowing with revolutiona condition of space which undergoes nochange.

Those who may have some misgivingsabout this explanation of what occurs whena cylindrical bar magnet is rotated about itsaxis may be somewhat reassured by regardingthe phenomenon from the point of view of anobserver situated on the surface of themagnet about midway between the poles.Let Fig. 2 (b) represent a cross section of themagnet looking towards the north poleand let the observer be at P at the moment.If this observer considers himself to be atrest, he will regard the magnet as revolvingabout him in the direction shown. Belowhim will be a magnetised space through whichthe material of the magnet is moving with avelocity proportional to its distance fromhimself causing the electrons to experiencea force tending to move them away fromhim. He will also notice that the magneticflux through some stationary close curves isincreasing and that through others de-creasing ; we show two closed curves and thedirections of the electromotive forces in-duced in them by these changing fluxes.He will regard the final distribution of elec-trons as due to theresultant action ofthese two phenom-ena. (The electronstend to move inthe opposite direc-tion to the inducedE.M.F.).

The explanation from this point of viewis much more complicated than that givenabove where the axis of the magnet wastaken as the axis of reference.

If the bar magnet is replaced by a solenoidwe get an arrangement similar to that usedin the Lorenz apparatus for the absolutedetermination of resistance, and the in-ductance of electromotive force can beexplained in the same way by the radialforce experienced by the electrons in thedisc due to their rotational movementthrough magnetised space. If the solenoidwere rotated it would also become charged asdid the bar magnet and with the same result.

It would doubtless be possible to explainall these phenomena without the conceptionof a magnetic field, as they are all primarilydue to the relative motion of electrons, but

(a)Fig. 2.

EXPERIMENTAL WIRELESS 411

we do not propose to make such an attempthere.

If brushes are arranged to make contactwith the spindle and edge of the disc, currentcan be maintained through an externalcircuit by this elementary form of homo-polar dynamo, but it seems a very un-necessary and artificial procedure to attemptto explain this current, or the E.M.F. thatproduces it, as due to the magnetic fluxthrough some fictitious circuit undergoing acontinual change, as is sometimes done.

Blondel* replaced the disc in Fig: i by adrum of insulating material on which anever increasing number of turns of insulatedwire were wound as the drum rotated, thewire being unwound from another drumoutside the magnetic field. By means ofbrushes rubbing on a slip ring on each drum,the " circuit " was completed through agalvanometer which, however, gave little orno indication of a current during the windingprocess, although the linkages of the " cir-cuit " were being increased owing to theincreasing number of turns on the drum in themagnetic field. The " circuit " here wasfictitious ; it consisted in part of a con-tinually changing and arbitrary dotted linedrawn on a slip ring, which made it im-possible to apply the conception of linkages.Looked at from the point of view consideredabove there is nothing mysterious about theexperiment ; it is merely an unduly com-plicated and inefficient form of homopolardynamo in which one of the brushes hasbeen replaced by a flexible connection.

Conductor in a Changing FieldUp to the present we have confined

ourselves to cases in which the conductormoves in an unchanging magnetic field.We shall now consider the more general casein which the conductor moves relatively tothe magnet in such a way that it is situatedin a changing field. The explanation given ofthe phenomena will depend on the observer.Whether in Fig. 3 (a) the magnet pole facingus be at rest and the conductor move acrossthe front of it, or the conductor be at restand the pole move across behind it, theresult will be the same ; the terms " moving "and " at rest " are only relative. To anobserver situated on the pole it is at restand the magnetic field is everywhere un-

* Comptes Rendus, Vol. 559, p. 674.

August, 1933

changing ; he sees the conductor movingthrough this field and its electrons con-sequently experiencing an upward forcetending to charge the conductor as shown,the magnitude of the force and chargedepending on the strength of the fieldthrough which it is moving at the moment.If the conductor be replaced by a coil as inFig. 3 (b) the electrons in the two sides willexperience different forces, because they aremoving through fields of different strengthand a current will be set up due to thisdifference.

(b)Fig. 3.

An observer situated on the conductorwould probably explain matters differently ;we say " probably ' because he might beclever enough to see that a great simplificationresulted by assuming that the magnet wasat rest and that he himself was moving, inwhich case his explanation would be the sameas that of the observer on the magnet.Failing this flash of genius he would regardhimself as at rest but surrounded by amagnetic field which was changing frommoment to moment so that the flux throughany closed curve such as those shown inFig. 3 (c) was either increasing or decreasing,thus inducing electromotive forces aroundthese paths which cause the free electronsin the stationary conductor to crowd towardsthe top end. At first sight it might beexpected that these electromotive forceswould cause a similar upward movement ofthe free electrons of the magnet itself, butit must be remembered that to the observeron the stationary conductor the magnet ismoving to the left through magnetised space-it is immaterial that the magnet is itselfproducing this magnetic field, and that itsdistribution in space is changing-and itselectrons consequently experience a down-ward force which counterbalances the up-ward force. Both observers fortunatelyagree as to the facts, i.e., the resultant move-ment of electrons, but their explanations arevery different ; to one of them the sur-

(a)

August, 1933.

rounding space contains an unchangingmagnetic field but is devoid of electric fields,whereas to the other the magnetic field isundergoing great changes and is accom-panied by an electric field.

In the case of a choking coil, induction coil,or transformer, there is no question of themovement of a conductor through magne-tised space, since the observer would natur-ally be at rest relatively to all the windings,and would regard the electromotive forcesas due entirely to variations in the magneticinduction.

To an observer on a pole of a smooth coredynamo or alternator the phenomenon appearssimple ; the induced E.M.F. is due entirelyto the movement of the conductors throughmagnetised space. When we are dealing


with a rotating pole machine, the assumptionby the observer that the pole on which he issituated is at rest involves the further assump-tion that the whole stator is moving bodilyand that the opposite pole is moving atdouble its actual speed, but so long as heconfines his attention to his own pole and theconductors under it he can ignore the restof the machine and leave it to observerson the other poles to make their own assumptions. Strictly speaking, this is not quitecorrect, but we assume, as we did in con-sidering Fig. 3, that the other poles are sofar removed that any electric forces due tothem may be neglected. The case of thetoothed armature is much more complicated,but this is not the place to go into a detaileddiscussion of it. G.W.O.H.

The National Physical LaboratoryWireless Interests at the Annual Visit

THE annual event of an invitation visit to theNational Physical Laboratory took place onTuesday, 27th June, and was largely attended.

Matters of wireless interest were shown in severaldepartments, particularly in the building of thenewly formed Radio Department, including ex-hibits brought from the Radio Research Station atSlough, which, although now included in the RadioDepartment, continues its work at Slough.

Other exhibits comprised various apparatus foruse in connection with ultra short waves, includingapparatus for the generation of waves below

metre, a new type of diode for use as a short-wave electronic oscillator, and apparatus for theinvestigation of the electrical properties of theground at these wavelengths. Soil properties werealso represented by laboratory measurements ofsamples brought from various places. An exhibitin relation to circuital performance was one showingthe use of two triodes as a tuned amplifier (insteadof a single screened -grid valve), with advantagesboth as regards gain and linearity. Other measur-ing apparatus was for determination of totalharmonic content (e.g., of a 1.f. output), and adifferential wire -expansion radio frequency ammeter.

In the adjacent transmitting hut was exhibiteda newly completed transmitting installation de-signed for various purposes of radio research, e.g.,pu1se and frequency -change emissions, as well asfor modulated or C.W. emissions for standardfrequency tests, etc.

An exhibit from Slough was a portable direction -finder for use with either a closed -coil aerial or arotating Adcock -type aerial, along with twocathode-ray oscillograph exhibits, in which theSlough station has specialised. One of these wasan acoustical analogy of the method of echo -

sounding for the determination of the height of theionosphere. The other exhibit was a cathode-ray stroboscopic application for the detailedexamination of a complicated pattern. In additionto these demonstrations there was also exhibited afilm illustrating the work conducted at the Sloughstation. The film first showed a panoramic view ofthe station, followed by (a) the directional receptionof atmospherics and the use of two -station inter-sections to locate thunderstorms ; (b) the receptionof short -duration pulses, showing reflections fromeither or both regions of the ionosphere ; (c) deter-mination of the polarisation of downcoming waves ;(d) measurement of downcoming angle of Trans-atlantic signals.

The Standards division of the Electricity De-partment had a number of typical exhibits relatingto precision measurements. A new exhibit wasthe production of low frequencies (25 to 200cycles) of high accuracy from the standard i,000 -cycle fork, this being demonstrated on a cathode-rayoscillograph. Another new and interesting articlewas a simple method of impedance measurement,using a pointer galvanometer as its detector andrequiring no double adjustments..

The Physics Department had a number of ex-hibits of audio -frequency and acoustic interest.These included open-air acoustic measurementson loud speakers and the absolute measurement ofsound intensities, the measurement and analysis ofnoise, and the delineation of audio -frequencyresponse curves (e.g., of microphones, etc.). Thebasement of the Physics building also had a numberof acoustic exhibits, such as reverberation measure-ments of absorption coefficients, acoustic reflectionsillustrated by a ripple tank, and sound -pulsephotography in relation to building acoustics.

EXPERIMENTAL WIRELESS 413 August, 1933

Distortion Cancellation in Audio Amplifiers*By W. Baggally

SUMMARY. --A method of amplification is described in which the input and outputterminals of an amplifier are connected through a high resistance, the voltage existing betweena point on this resistance and the earthy point being amplified and fed back to the inputcircuit.

It is shown that under suitable conditions all forms of distortion vanish from theoutput circuit.

The mathematical theory is first discussed, the conditions for distortionless workingand for stability being derived.

Experimental results are quoted showing that the residual harmonic distortion in theoutput of an amplifier working at full power may be less than 25 parts in io,000.

Practical details and circuit diagrams are given.

IG. i shows an amplifier of any type,the upper input and output terminalsbeing connected by resistances r and

mr in series (m is an arbitrarily chosennumeric whose value is discussed below) ;Y and mr are, for the sake of analyticalsimplification, assumed to be so large com-pared with the impedances in the input andoutput circuits that there is no appreciablefeed -back of voltage ; further, let the voltagew be assumed to be injected into the inputcircuit of a second amplifier (not shown,assumed to magnify a times without dis-tortion, and hereinafter called the a ampli-fier), the output of which is injected into theinput circuit of the main amplifier as shown,in series with the voltage v which it is desiredto amplify without distortion.

It will be shown that distortion in the aamplifier appears in the main output circuit,but is multiplied by the distortion com-ponent of the main amplifier.

The resultant distortion in the outputcircuit is thus of the third order of magnitudeand may usually be ignored.

If p signifies as usual a small change inthe quantity to which it is prefixed, we have

LE e . dEIde= (Q v+ a Q w)dE/de .. (I)

From circuit considerations we may easilyprove that

w= (me + E)I(m I) .. (2)

w = (m A e E)I(m 1) . . (3)

From this and (I) we have w = Q E(i + mde/dE)/(m 1) (4)

and by substituting in the second form of(1) we get

* MS. received by the Editor, October, 1932.

QE = dE{Qv + aQE(I + mde/dE)/(m ± I)}

de

(5)QE{1 - a(m + dE/de)/(m I)}

= Qv . dE/de .. (6)

dE/deQE= 1- a(m+dElde)1(m +x) AV (7)

dE/deE = oft, .dv (8)-a(m+dElde)1(m+r)Now put

a = (m (9)in (8), and if dE/de is not zero we have

E = - my .. (1o)

We thus have the interesting result thatthe output of the amplifier will not be dis-

Fig. I.

torted in any way, although the amplifieritself may be distorting, since (io) does notinvolve dE/de.

The physical explanation is, of course,that the amplifier distortion is renderedharmless by the input voltage being distortedin an equal and opposite manner, the twoeffects cancelling each other.

Equations (4) and (7) yieldm dE/de

w = m - a(m dElde). Qv.. (ii)

August, 1933

m+ dEldew= m +1- a(m+dElde).dv.. (12)

and putting in the distortionless condition(9) we have

- m(m + dElde).dv= oiv (m + 1)dElde

The condition that the system, which isreally a kind of retroactive amplifier, shallbe stable, will now be investigated.

In equation (5) put dElde = c anda(m + c)I(m+ = f and we have

Z\E =cAv+ f(cAv+ f(cAv +ft. .))) (14)

LEIcAv=i+f+f2+f3+P+.... (15)The condition for stability is that the

series (i5) shall converge ; it does so ifIfI<I.. .. (16)

If we assume that v is sinusoidal, we canexpress phase shift in the amplifier by puttingdE/de = p jq where j = - I, andassuming m to be a real positive quantity weshall have from (r6)

a(m p + jq),I(m + 1 < I .. (17)

This is a scalar expression equivalent to(it 4_ 25)2 ± q2 < (m 4_ 1)2/a2 (18)

If this were an equation, it would be thatof a circle with centre at the point (-m,0)and radius equal to (m r)/a, and since by(9), a = (m + r)/m, the radius will be m ;under these circumstances the origin will lieon the boundary of the circle.

These conditions are summed up in Fig. 2,in which the vector representing the ampli-fication dElde must lie completely withinthe shaded circle at all frequencies if theapparatus is to be stable.

In most systems of intervalve coupling,the amplification approaches zero at zerofrequency and also at very high frequencies,so that to ensure stability it will be necessaryto make a slightly less than (m + r)/m sothat the origin may lie within the circleunder all circumstances.

As an alternative, we may arrange the aamplifier so that the frequency band coveredby it is the same or narrower than that ofthe main amplifier, so that as the vectordElde gets smaller, the circle widens so asto keep the origin within its bounds at theextreme upper and lower frequencies.

.. (13)


In all cases it must be remembered, how-ever, that the a amplifier will be called uponto deal with the harmonics of the inputfrequencies, so that its response must belevel up to at least twice the highest fre-quency for which compensation is 'required,but that if there is no frequency distortionand phase distortion in the main amplifierand the a amplifier is separate from it (theymay be combined as shown later), then thea amplifier may begin to cut off at a fre-quency one octave above the lowest notewhich it is desired to compensate, since underthese conditions it is not called upon to dealwith the fundamental frequencies, but onlythe harmonics.

2m

2m

Fig. 2.

If the real part of dE/de is much greaternumerically than the imaginary part (in awell -designed resistance amplifier this willalways be so over the working range of fre-quency), it is convenient to make m = -pat some mid -point in the frequency band ;in this case w is very nearly zero for smallamplitudes at this frequency, so that the aamplifier does not come into action until themain amplifier distorts ; also the tip of thevector in Fig. 2 is brought to the centre ofthe circle, which gives maximum stabilityand allows of the widest variations in dE/debeing compensated.

Under these conditions we may investigatethe effect of distortion in the a amplifier asfollows.

SupposedElde = -m -g

anda = h+ (m + r)/m,

where g and h are small quantities repre-senting the distortion components and may


be real or complex ; then from (8) we shallhave

E_or 1- (h (m 1)1m)(m -m - g)1(m i)

.dv . . (19)

-m -g

-m. E . dv (20)± mghl(ni + 1)(m g)

. E = op - on(1 - (m i) ) . dv . . (2I)

approximately.Since g and h are both small quantities and

m + i is always greater than 2, the termgh1(m + 1) is extremely small and may bein general ignored.

It is of interest to compare this with thedistortion introduced by a simple un-compensated amplifier under the same con-ditions ; it is readily seen that in this case

dEE or

edv.

. E = ofv - m(i + glm) . dv .. (22)

since e = v and dE/de = -in - g,the distortion being more than i /h times asgreat as in the compensated case.

When it is remembered that the a amplifieris only dealing with the second -order com-ponents in the output circuit of the mainamplifier and only has to magnify them(m i)/m times, the enormous advantageto be obtained from the cancellation becomesevident.

The actual amounts of distortion whichcan be successfully compensated are ofconsiderable interest and may be determinedin the following manner.

The permissible PHASE DISTORTIONcan be read off at once from Fig. z by strikingan arc with centre at the origin and radiusequal to I dE/de I and measuring the anglesubtended at the origin by the segment ofarc lying within the circle ; in particular, ifthe real part of dE/de is equal to - in, theimaginary part may reach the same numericalvalue before instability sets in ; thereforethe maximum phase displacement whichcan be compensated is ch = ± 7r/4.

The amplitude distortion which is capableof being compensated would be limited bythe points at which dE/de = o if the aamplifier could deal with infinite amplitudes(see equation (13) ) ; since it cannot,further investigation is necessary.

August, 1933

If the a amplifier can deliver a voltage uwithout serious distortion, but when u isexceeded the distortion becomes large, thelimiting condition will be reached when

(m -4- dElde)dE/de

. dvof°

which is derived from (i3)From (7) it may be seen that

- dE/deLv - . pe . . (24)

so that (23) may be written

oPmdElde de

= (23)

U (25)

where y is the value of e corresponding to anapplied input voltage v.

It is now necessary to make some assump-tions as to the form of the functionalrelationship of dE/de to e so as to be able toevaluate the integral (25).

The assumptions made should preferablyconform fairly closely to reality and at thesame time render (25) conveniently in-tegrable.

For example, we can stipulate thatdE/de = -m from e = minus infinity toe = e1, then dE/de = k(e - el) -m frome = e1 until dE/de vanishes, after which weare not interested in its further exploits.

Integrating these and equating them whene = e1 (since we are not permitting dis-continuities in the output), we find thatE - me from minus infinity to e1, thenE = ke212 - (ke, + m)e ke1212 from e1

onwards.It is a straight characteristic with a

parabolic top which becomes saturated whene= e, + m1k.

This is probably sufficiently like real lifeover the working range for the presentpurpose , inserting it in (25) we have

ork(e - e1) . de = u . . (26)

since the first part of the integral is zero.The result of the integration is

k- - e,y =u . . (27)m 22 e2

2

or y2 - 2ely e12 - 2mulk = o (28)The relevant solution of this quadratic is

y = e1 -{- '/ 2mulk .. (29)

August, 1933 416 THE WIRELESS ENGINEER &

and finally we have the maximum undis-torted output voltageV = m(y - u) = m(e + 1/2trcu/k - u) (30)if u < 2m/k.

If 2m/k < u, the analysis breaks down.It is seen from the above that in FRE-

QUENCY DISTORTION, dE/de may notrise to a greater value than zm and may notfall below such a value that the output fromthe a amplifier exceeds u.

Using, as before, the notation dE/de = c,(23) may be at once integrated since c isnow assumed constant with varyingamplitude, thus

v(m+IcI)=Iclu .. (3r)

c I = my/(u - v) . . . . (32)

Therefore the apparatus will not distortor oscillate if

my/(u - v) < I C I < 2M . . (33)

(c is, of course, negative).

The different kinds of distortion have beenconsidered as occurring independently ofeach other, and theabove analysis shouldsuffice to give a goodidea of the order ofmagnitude of the per-missible distortions,but in practice allthese effects will beoccurring simultan-eously and will allmodify each other ; the general case isprobably too complicated to be worth amathematical treatment, and the attemptwill not be made here.

The question now arises : What is thegreatest output voltage we can get withoutdistortion ? Since the amplifiers will pre-sumably be designed with care, so as toreduce phase distortion and frequency dis-tortion to very small values, the answer tothis question is going to be given by equation(30) very nearly, but it must be noted thatany departure of the phase characteristicsand amplitude -frequency characteristics fromthe ideal of straight lines parallel to theaxis of frequency, although not necessarilyinvolving any direct distortion, will, never-theless, cause some reduction in the maximumundistorted output and also some slight

Fig. 3.

increase in the residual harmonic content atlevels other than the maximum, due to thefact that the a amplifier will, under theseconditions, be in action even at smallamplitudes and will therefore contribute itsquota of distortion as shown in equation (21).

The following question is, practically, offundamental importance in the applicationof these methods to speech amplifiers, etc.

Suppose that the amplifiers are high-grade,resistance -coupled instruments, free fromphase distortion and frequency distortion inthemselves ; what effect on the character-istics will be produced by a reactive load,such as a loud speaker, in the output circuitof the main amplifier ?

From Fig. 3, which is the equivalentcircuit of the output stage, we have at oncen = E/E, = (R jX)1(R Ra jX) (34)

I n = V(R2 + X2)/( (R 12,4)2 ± X2) (35)

tan = XRa/(R(R Ra) + X2) .. (36)

where qS is the phase angle of n.By differentiating (36) to X and equating

to zero, we find that has a maximum whenX = R(R Ra) (37)

and substitution of this value back into(36) gives

tan -112,121/12(R R,) .. (38)

which is the worst possible condition forphase distortion.

To get an idea of the way in which thequantities I n I and vary with frequency,assume R fixed whilst X varies from zero toinfinity.

Put R. = sR ; then whenIn +s)tan ck=o

X=R.1/I-I-sInI=I/VI+stan4=-s/2 .Vi+sX= inf. Inl-I tang=o

- (39)If X changes sign, so does ck, but remains

the same numerically.We can make these variations as small as

we please by reducing the value of s, but ifwe go too far in this direction, the load inthe output circuit will become far removedfrom the optimum and the output valve willnot work efficiently.

If s is made about o.3, the load will beonly slightly greater than the optimum formost valves, I n I only changes by ± 0.14 of


the mean value and the maximum phasedisplacement will be ± 0.13 radian.

These changes are well within the capacityof the compensation, and will not be of verymuch consequence in practice, especially asX will not really reach zero or infinity inpractical loads, so that the changes will beeven smaller than the calculated figures.

It is worth noting in this connection thatthe compensation keeps the voltage acrossthe load constant ; that is to say, theamplifier behaves as if the A.C. resistance ofthe output valve were zero ; this may ormay not be advantageous, but it will haveto be taken into account when consideringthe question of the response curve of aloud speaker.

Of course, a series resistance could beused to prevent anything drastic happeningin the bass, with a condenser shunted acrossthe resistance so as not to waste power inthe treble ; it will be noted that the problemis the opposite to that encountered whenusing pentodes, so that we shall have to usethe opposite network to solve it.

When making up practical circuits usingcancellation, the main thing to remember isthat the input and output of the mainamplifier must be in phase -opposition (i.e.,when the input terminal is positive withrespect to the earthy point, the outputterminal must be negative and vice versa),but that the input and output of the aamplifier must be in phase, which limits usto an odd number of valves in the mainamplifier and an even number in the aamplifier ; that is, if resistance amplifierswith common power supply to the valves areused.

If transformers are available which aresufficiently free from distortions of variouskinds, these limitations do not arise, sincewe can get the necessary phase changes byreversing the connections to the transformers.

Of course, considerable liberties may betaken in this direction with the main ampli-fier, since the cancellation will correct allminor faults here, but the conditions to befulfilled by the a amplifier are far morestringent and it is probably not practical tointroduce any commercially existing typesof intervalve transformer into this part ofthe circuit, although it seems probable thata low -ratio transformer following a lowimpedance valve and with a resistive load

August, 1933

across the secondary could be designed tohave the necessary characteristics ; also theuse of transformers in the main amplifierwould allow the cancellation to be combinedwith the method of grid current compensa-tion which was described in a previous paperby the present author in this journal, thusleading to further increases in the undis-torted power output.

Fig. 4.

It is hoped to conduct further researchesalong these lines in the near future.

The simplest circuit of the resistance -coupled type is shown in Fig. 4, in which theoutput valve of a chain is compensated by aseparate two -stage amplifier ; it is usefulfor many purposes in the laboratory, but ishardly a proposition for a broadcast receiver.

However, it serves as an excellent experi-mental circuit with which to test thecapabilities of distortion cancellation, andsome results obtained with it are now given.

A Mazda P650 valve was used as the mainamplifier, with about 200 volts on the anodeand 42 volts grid bias, the anode currentbeing 29 milliamps.

The load consisted of a non -inductiveresistance of 4,500 ohms in series with theinput circuit of the distortion meter, whichwas shunted with non -inductive resistanceso as to have an impedance of 500 ohms,which is almost purely resistive ; the totalload was therefore equivalent to a pureresistance of 5,000 ohms to a sufficientdegree of approximation for the presentpurpose.

The input consisted of carefully purifiedi,000 cycle A.C. from a valve -maintainedtuning fork, the harmonic distortion beingless than one part in a million, the amplitude

August, 1933

being adjusted just below the point at whichgrid current commenced to flow in the P650.

The amplifier adjustments are made asfollows :-The output of the a amplifier isfirst disconnected from the main amplifierinput circuit, the gap in the latter beingbridged, and is connected to a pair of phones.

The grid -anode potentiometer tappingpoint is then adjusted for minimum soundwhen the input voltage is small ; fromequation (12) we now know that under theseconditions m = - dE/de and that conse-quently the a amplifier will not come intoaction until the main amplifier distorts.

The output of the a amplifier is now re-connected to the input circuit and its gainadjusted until the harmonics in the outputare at their minimum value, the inputvoltage having been first increased so as togive a good crop of harmonics before thegain adjustment is made.

If a harmonic measuring apparatus is notavailable, this adjustment can be made byconnecting a potentiometer and phones asshown in Fig. 5 or in some other equivalentmanner.

By adjustment of the slider of thepotentiometer, the fundamental may bepractically balanced out of the phones, whenthe harmonics will become audible and maybe reduced to a minimum by the gainadjustment.

Fig. 5.

These adjustments having been made, thevalve was found to be delivering one wattto the load, the total harmonic distortionbeing 0.25 per cent. ; when the gain of thea amplifier was reduced to zero, thus cuttingout the cancellation without disturbing theother circuit conditions, the total harmonic


distortion rose to 3.4 per cent.: thus thecancellation reduces the harmonics to lessthan 1/13th of the usual value.

When a small moving -coil loud speaker

Fig. 6.

whose reactance and resistance at i,000cycles are 540 ohms and 2,440 ohms, re-spectively, was substituted for the 4,50o ohmresistance in the output circuit, the resultswere substantially the same as in the firstexperiment, thus confirming the conclusionsarrived at theoretically in regard to theeffect of reactive loads.

The percentage harmonic content measuredby the distortion factor meter is the quantityH = E1+ E, E, . . . x loo/E0,

where E0, E1, E,, etc., are the R.M.S. valuesof the total output, first harmonic, secondharmonic, etc., respectively.

The meter consists of a high-pass filterwhich attenuates the L000 cycle funda-mental about no db. but passes the har-monics with only about o.5 db. attenuation,followed by a potential divider calibrated inpercentages, the output of which goes intoan amplifier with an A.C. voltmeter in theoutput circuit.

In taking a measurement the filter is putin and the voltage divider is set to Ioo percent., the gain of the amplifier being adjustedto give a convenient reading on thevoltmeter.

The filter is then cut out and the potentio-meter adjusted so as to give the samereading on the voltmeter.

The setting of the potentiometer then givesthe percentage harmonic distortion as de-fined above, assuming there is no frequencyor waveform error in the amplifier orvoltmeter.


Reverting to the question of circuits forpractical use, in the case of a multi -stageamplifier one may use the early stages tocompensate the output stage ; for example,in a three -stage amplifier the first twostages are used to amplify the incomingvoltage in the usual way, the same twostages being used as the a amplifier forcompensating the output stage, therebysecuring the advantage of a compensatedoutput stage without the use of extravalves.

The arrangement is shown in Fig. 6, thepotentiometer on the extreme right con-trolling the amount of feed -back in the aamplifier without alteration of the totalgain of the complete amplifier.

The resistance of the two potentiometersneeds careful consideration in relation to theother circuit elements, since they con-stitute a high impedance in series with theinput circuit ; capacity effects are to beguarded against here.

IN

G.B.- G.B.- G.B.-

August, 1933

In the case of a well -designed amplifier,however, the percentage harmonic dis-tortion of the output stage will certainly bemuch greater than that introduced by allthe other stages put together, and byeliminating the output stage distortion weshall be getting rid of the bulk of the totaldistortion introduced by the whole system.

The method of compensating a completemulti -stage amplifier by means of an inde-pendent a amplifier is shown in Fig. 7.

The only points of special interest here are,first, the a amplifier output is shownshunted across the main amplifier inputcircuit as an alternative arrangement to theseries connection shown in Figs. 1, 4 and 6,which allows one input terminal to be atearth potential, which is useful in somearrangements, and secondly, the main ampli-fier output circuit is shunted with a highresistance which is tapped down as shownso as to only apply about I/Ioth of theoutput voltage to the balancing potentio-

.111

0OUT

NVVrAnt

Fig. 7.

The theoretical limitation of this circuitis, of course, that it is only the output stage,not the whole amplifier, which is com-pensated, so that the distortion introducedby the first two stages remains withoutreduction.

meter, the object of this arrangement beingto avoid any appreciable feed -back from theoutput to the input of the main amplifierthrough the resistance of this potentiometerwhile keeping the resistance of thepotentiometer within practical limits.


The Acoustical Performance of aCone -type Loudspeaker*

By D. A. Oliver, B.Sc.Communication from the Staff of The Research Laboratories of the General Electric Co.. Ltd.. Wembley. England.

1. IntroductionIN acoustical engineering frequency response

curves of various kinds are now widelyused and they are being relied upon to an

increasing extent. However, to justify theconfidence which is being placed in them itis essential to know the exact conditions oftest and to investigate the limitations on theinterpretation of the results.

So far as the loudspeaker is concerned,the more important points which arise intesting under simplified acoustical conditionshave already been discussed.' This paper isan extension of that discussion, but itattempts primarily to illustrate the recom-mendations already made and to furnish aseries of precision measurements on oneinstrument of the inductor type,23 whenthe boundary conditions are definite and themany variables kept under proper control.

2. Brief Description of EquipmentThe laboratory in which the measurements

about to be described were carried outconsists of two adjacent rooms, the larger ofwhich is i6ft. long by 8ft. high and ioft.wide, covered first with two layers of cabotquilt, having behind the quilt at variousplaces a small air space, and completely linedwith six layers of gamgee tissue or medicalcotton wool. The absorption of low frequencysound energy is believed to be due largely tothe diaphragm action' of the brown paperconfining the eel grass, the paper beingelastically supported but well damped oneither side. The loudspeaker is mounted ona stiff plywood baffle board 2ft. square andgin. in thickness integral with the partitionwall separating the two rooms. The frontof the cone radiates into the heavily dampedroom and the back would radiate into thesmaller adjacent instrument room but for

* MS. received by the Editor, July, 1932.

the fact that a very large tightly fitting boxlined with absorbing material encloses theback of the loudspeaker and nearly absorbsall the sound energy radiated by the backof the vibrating paper diaphragm. It waschecked at several frequencies between widelimits that the effect of this box did notinfluence the pressure measurement in thedamped room by more than a few per cent.,or a fraction of a decibel. The object of thisbox was to prevent the high sensitivitymicrophone amplifier from being affected bysound from the back of the diaphragm.

Elaborate precautions had previously beentaken in the amplifier itself against micro -phonic troubles and adequate checks appliedto verify that the microphone signals werenot being invalidated by spurious e.m.f.sset up by other causes. In most of thetests a direct change -over was made from thecalibrated condenser microphone to astandardising potentiometer immediatelyafter each measurement, thereby eliminatingany special reliance being placed on theamplifying and output recording system.The condenser microphone actually used forthis work was a smaller and still furtherimproved model3 of proven reliability andhad been calibrated against a Rayleigh discin free air. A further check has recentlybeen furnished by calculating its responseon certain assumptions using its initiallydetermined mechanical constants, and goodagreement with the free air calibration hasbeen obtained.

3. Experimental ResultsIn Fig. i are shown the principal dimen-

sions of the inductor dynamic loudspeakercone to which the following data apply. Amore detailed account of the constructionalfeatures will be found elsewhere.3 Thepaper was soft, untreated, and 0.023 cm. inthickness. The velocity of sound in this


paper (1/ p) was found to be 2.3 x robcm/sec.

(a) Axial ResponseIn Fig. 2 (a) is given the final corrected

response curve for the loudspeaker when thesound pressure was measured on the axis ata distance of 6o cm. and the alternatingcurrent was kept constant at all frequencies.This curve plotted in decibels gives, at themeasuring point, the true acoustical powervariation above or below the power level atx,000 cycles per second, as the driving pointimpedance of the medium in free space ora completely damped enclosure is resistiveand constant at all frequencies. A furthercondition is that the radiations from thetwo sides of the vibrating diaphragm are

Fig. s-Section of Inductor Loudspeaker showingprincipal dimensions.

separatedacousticalthem.

20

16

12

8

,1 4

°OV, 4

August, 1933

by an infinite baffle and that nointerference takes place between

8

12

18

20

\.1

IJ(a)

U

(b)

0

91

0

100 1000

FREQUENCY IN CYCLES PER SECOND

Fig. 2.(a) Axial response at constant current of

5.o x so -3 amp. Microphone distance6o cm. P (dyne/cm2/amp) at o db = 1.1 x so3.(S°/pc) x so -7 at 6o cm. at o db = 2.9 X 143-3watticmaiamp.

(b) Current variation with P2 valve fully loaded.o db 8.2 x 10-'amp.

(c) Axial response with P2 valve. (p$/ pc) xat 6o cm. at o db = 5.9 x so -7 watt/cm.2

The constancy of the shape of Fig. 2 (a)when similar measurements are made atdifferent values of constant alternatinginput current and at different axial distancesfrom the plane of the baffle will be discussedlater.

10,000

(b) Current Variation and Valve CouplingWhen an e.m.f. constant at all frequencies

is applied to the grid of the power valvethe current through the loudspeaker isnot necessarily constant with frequency,principally on account of the variation withfrequency of the impedance of the loud-speaker. This variation can be calculatedfrom the circuit constants but it has beenfound more convenient to measure thealternating current flowing through theloudspeaker under constant grid voltageconditions and to express the currentratios in decibels 20 log10 (If/I1o00). Thisgives a curve passing through zero at i,000cycles per second. This procedure does notgive a curve of variation of input electricalpower to the loudspeaker, as that wouldassume a constant loudspeaker input effectiveresistance at all frequencies, but ratherone which, when added to the curve of trueacoustical power variation measured atconstant current, takes account of theactual acoustical power variation due to

August, 1933

the variation with frequency of the currentthrough the loudspeaker when it is con-nected in the output circuit of the powervalve to the grid of which a constant e.m.f.is applied. An underlying assumptionis that the pressure in the sound field andthe input current to the loudspeaker areconnected by a linear relation, the validityof which can be checked experimentallyin a given case.

A curve of current variation is given inFig. 2 (b) and was obtained with an OsramP2 valve fully loaded and coupled to theloudspeaker by a 20 henry choke and 2microfarad condenser, the circuit beingsimilar to Fig. 9 of a previous paper.1

(c) Axial Response Curve at Constant A.C.Grid Voltage

The overall axial response curve Fig. 2 (c)is obtained by the sum of curves (a) and (b)Fig. 2, and curve (c) represents the variationin acoustical power at the measuring pointper unit area of wave front for a constantapplied voltage to the grid of the P2 powervalve. This valve is not necessarily the mostsuitable for use with this loudspeaker,but as complete results with this valvehave been obtained they are given here toillustrate the general method of applyingthe data.(d) Polar Pressure Distribution

At the top of Fig. 2 an indication isgiven of the polar distribution of relativepressure round the source referred to the00. 1.0 08 043 04 02

80'pifistritiiiiitsetVail

110p. ill.1t

, & 0 w asIS 0

7ireinsinti4.48'

\ 4aiAllimilllil WA, //#60

41p 4.111014/1.'44 .... kig't, 2000 , EC,

5 , sAmoiiiposs,..0

\,,sy

02 04 043 08 1-0

40' 30 20 10 0 10' 20' 30' 40'Fig. 3.-Measured polar curves of relative

pressures for cone at several frequencies.

axial value. The actual experimental ob-servations for the loudspeaker in questionare shown in Fig. 3. Each point was referredto a potentiometer calibrating circuit similarin principle to that due to West,6 therebyensuring a degree of consistency difficult toattain in this type of measurement. To

422

1000 ,/ 6E0. I I Iso*

so'

70'

80'

50'

THE WIRELESS ENGINEER &

eliminate errors due to drift in frequency, asmall monitoring source was supplied fromthe heterodyne oscillator, having filteredoutput, and continuous manual adjustmentof frequency was carried out by maintainingslow beats with a series of standard tuningforks. It is seen that the polar pressurecurves are fairly symmetrical and it waschecked that repeating the measurementswith the loudspeaker turned axially through90 deg. gave further curves in close agreementwith those quoted, thereby verifying thatthe polar distribution from the cone hadaxial symmetry to a high order of uniformity.

(e) Polar FactorUnder the test conditions already specified

it has previously been shown' that the totalacoustical power Pa radiated at a frequency!is given by

[Pc]f = 0.13o R2 q2S x ro-7 watts .. (r)

where q is the actual axial pressure (dyne/cm2)at radius R cm. measured from the centreof the front plane of the baffle and S is a polarfactor obtained by applying the Rousseauconstruction to the square of the relativepolar pressure diagram. For hemisphericalradiation S = z and is less than unity whenappreciable directivity is present. In de-ducing the various values of S from theinductor loudspeaker results in Fig. 3 themean pressure -squared curve was foundand then the Rousseau construction applied.Due to the high degree of symmetry in thepolar distribution, these values of S werefound to be in almost exact agreementwith the same construction applied to thesquare of the mean pressure curve, whichis considered a less accurate procedure.

(f) Total Power ResponseFrom equation (I) it is seen that the

variation of q has already been taken intoaccount in the curves of axial pressure (a)and (c), Fig. 2, and therefore to convert thesecurves into curves of total radiated power atthe front of the infinite baffle we have onlyto calculate io logio (S/S1000) to a base offrequency and add the new curve so obtainedto Figs. 2 (a) and 2 (c) to obtain the totalvariation in acoustical power output foreither constant input current or for constantinput voltage to the grid of the power valve.The curve of io log10 (S/S,090) might be


called the " polar factor curve." Theparticular values obtained are shown plottedin Fig. 4 (a), where the general smooth curveconnecting them has been drawn in by achain line. It is realised that if furtherequally reliable measurements were carried

a

-12

-1

2

6

242

20

2

4

0

8

(a)(b

bo

FREQUENCY IN CYCLES

Fig. 4.(a) Polar factor in decibels. S at o db = 0.34.(b) Total power response with P2 valve. P, at

o db = 1.48 x zo-3 watt (fully loaded).

at other frequencies, certain small fluctua-tions in the curve would probably be estab-lished due to the considerable changes thattake place in the mode of vibration of thecone over the working frequency range, butfrom approximate additional polar curvestaken at intermediate frequencies it is notexpected that further lengthy and carefulwork would invalidate continuing the argu-ment on the basis of Fig. 4 (a).

In Fig. 4 (b) is given the variation in totalpower output for the inductor loudspeakercoupled to the P2 output valve, and is thesum of Figs. 2 (c) and 4 (a). This curve onlystrictly applies to the specified conditions oftest, of which the most important is theinfinite baffle condition. The general methodcould doubtless be extended to cope withthe case of a finite baffle, when polar curvesof sound. pressure would be required embrac-ing the speaker or set cabinet and referredto different planes. The source would haveto be centrally situated in absorbing sur-roundings, but these further conditions willnot be elaborated here. However, the actualtest conditions are not so artificial that theycannot be used when subjective aural testsare also desired.

(g) Combined Receiving Set and LoudspeakerCurves

It is convenient to point out at this stage

boo 4.000PER SECOND

August, 2933

that such curves as have already beenobtained can be linked on directly withoverall radio receiving set tests. In Fig. 5 (a)is given an actual curve for an all -mainsreceiving set showing the variation indecibels of the current into the loudspeaker.The high frequency tailing -off is rathersevere for these particular conditions oftest, which were not those of standardpractice. If now this curve is added to thesum of Figs. 2 (a) and 4 (a), we obtain anoverall curve of variation in total usefulsound output power for an infinite baffleand constant radio -frequency dummy -aerialvoltage input, and is primarily given as anexample of the comprehensive flexibility ofthe general recommendations which havebeen put forward.

(h) Further Applications of ResultsThe next step correlating the acoustics of

the loudspeaker and the room in which it isoperated is by no means straightforwardand has to be taken with caution.

Let us consider briefly some of the effectswhich take place when music is reproducedby a loudspeaker in an ordinary room inwhich both reflection and absorption arepresent. Standing waves are produced byprogressive waves travelling simultaneouslyin opposite directions, these being set upin turn by multiple reflections from theboundaries of the enclosure. Some of these

12

8

4

0

:71 4

O8

Pi 12

16

20

24

28

(b)

(a)

100 1000

FREQUENCN IN CYCLES PER SECOND

Fig. 5.(a) Three -valve receiver : loudspeaker current

variation for constant input voltage to aerial.(b) Total power response of receiver : (infinite

baffle).

10p00

standing wave effects are reinforced due tothe many modes of resonant vibration7 intowhich the volume of air in the room can beexcited, and for a completely reflecting room

B2

August, 1933

these resonances are extremely sharp andnumerous, but for ordinary rooms they areappreciably damped. The time requiredfor these resonances to build up to their fullamplitude even then is usually much longerthan the brief period of sustained excitation.In practice, the standing wave pattern ischanging continuously from moment tomoment and measurements of the finestructure of the pressure variations can onlybe made conveniently when using a constantpure tone source. The effects mentionedabove produce large local changes in intensitywhich are further enhanced by the presenceof acoustical foci dependent upon thegeometry of the room. However, as mostforms of reproduction do not involve thesounding of long sustained notes, the de-leterious local effects are not detected by earas readily as might be expected, because thetime intervals involved are usually too shortfor the individual forms of distortion tobuild up to their maxima. Moreover, withbinaural listening an invaluable averagingeffect is obtained. Subjective experiencehas, of course, approved the degree to whichcertain forms of reverberation and localwave distortion are tolerable or even de-sirable, but with such considerations we arenot here concerned.

A further set of complications arises whenthe directional characteristics of the sourceitself are considered and the fact thatthe power radiated by the loudspeakerdepends to some extent upon the magnitudeof the reaction of the enclosure upon itsdiaphragm. This is due to the fact thatthe driving point impedance of the roomexperienced by the speaker diaphragm isdifferent from the impedance imposed uponit in the equivalent infinite medium testcondition, but Crandall° has pointed outthat the effect of this is likely to be aminimum when the power is referred to aconstant force basis, as in this case, ratherthan to a constant amplitude basis. Theseriousness of this change in impedance isdetermined largely by the associated internalmechanical impedance of the driven dia-phragm, being less serious the greater themechanical impedance of the speaker. Acorrection for the effect is, however, expectedto be small in most practical cases.

Some recent observations of the soundpressures in the field of a loudspeaker have


been made by Barnes° and Ballantine,10first in a highly absorbing room and thenin a reflecting one. If the fine structureand general level of the curves are ignoredthe trend of the pressure response in eachcase can be considered the same, althoughin the room having partially absorbingboundaries the variations in pressure fromthe mean are considerable.

There are two steady state cases whichcan be considered. Imagine first that thediaphragm of the loudspeaker is mounted in,and integral with, one wall of a completelyabsorbing room, and arranged so that theback of the diaphragm radiates freelyinto space or its equivalent. The conditionsof use then conform to the conditions oftest and the response curve of Fig. 2 (c)applies directly for all positions on theaxis of the speaker, while the loss in responsefor positions away from the axis is readilyobtained by converting the polar curves ofrelative pressure, Fig. 3, into loss in decibelsfrom the axial value and applying the neces-sary corrections to Fig. 2 (c) . If now weconsider the loudspeaker similarly situatedbut working into a reflecting room anapproximation to the state of affairs canbe made in the following way. If E is themean energy density in the room havingan absorbing power a, and Pa' is the con-stant rate of supply of energy or actualpower radiated to maintain the energydensity E, and c is the velocity of soundin air, then 11

E^4 Pa'ac

. . (2)

or writing this equation in the standarddecibel notation

0'),10 log( io log (Pa51000

-10 logio( a' - (3)

As the values of Pa' are not expected tobe very different from the values of Pameasured in absorbing surroundings anestimate of the variation of mean responsecan be made in the room. Thus Fig. 4 (b)from which the frequency curve of absorbingpower of the room has been subtractedwould give the approximate mean overallresponse in the room for that particular


valve, coupling and speaker for a constantalternating voltage applied 'to the gridof the power valve. A conversion of thefinal characteristic into some scale of loudnessunits on the lines recently indicated"would complete the information. It is, ofcourse, realised, that for a listener in frontof the loudspeaker and close to it, the axialresponse curve will apply to a good degreeof approximation even in a reflecting roomas the direct energy is then large comparedwith the reflected energy. However, ingeneral the effective response curve willlie between the axial response curve (e.g.,Fig. 2 (c)) and that curve corrected for thepolar distribution of the speaker (e.g.,Fig. 4 (b)) and further modified by thefrequency variation in decibels of the absorb-ing power of the room. That these variouscorrections are not very critical by ear isborne out by Capt. West's interestingexperiments," in which various forms ofaxial response curve, with presumably noallowance for polar distribution and roomabsorption, were judged for naturalnessagainst the original, and the vote wasdefinitely in favour of the most generallyflat characteristic. It should be added infairness that relatively large changes in

25

O

S 200

15

5

ci 5

0c

5

-J

100

AXIAL DISTANCE

40- M_ 5ANURVE

501.100

500 1. 1.0001. 2 000 5,0001,

20 40 80 80 100 120 140

FROM FACE OF BAFFLE (CM)

Fig. 6.-Axial sound pressure variation withdistance for constant loudspeaker current.

response were being compared, but neverthe-less the result was encouraging and makesmore exact correlations worth attempting.

4. Additional Experimental Results(a) Variation of Axial Response with

Distance .Previous experiments had shown that

measurements of axial response at different

August, 1933

distances were very similar, but, to obtainprecise information, measurements were madeat various frequencies checked continuouslyby tuning forks. These results are shownin Fig. 6, where 20 logic (p,./po 0) is plottedfor various distances and frequencies. In

18

14

12

10

4

2

0

50 "L/SEC.100 "L/SEC.

500 "LdSEC.. 1,000"U/SEC. 2,000 NdSEC. 5,000 "u/SEC.

eX3

1 2 3 4 5 8 7 8

ALTERNATING CURRENT IN MILLIAMPERES

Fig. 7.-Axial sound pressure variation at 6o cm.with alternating current to loudspeaker.

general, the curves are smooth ones, showingthat standing waves in the room and betweenthe source and the condenser microphoneare small. A mean curve lying betweenthe various sets of results at differentfrequencies has been drawn in as well asthe individual curves for the 5o and 500cycle per second observations which happento show the greatest divergencies.

An inspection of the spread from the meancurve in Fig. 6 indicates that for radialdistances between the limits of 25 and 125cm., the probable maximum uncertaintyto be placed on Fig. 2 (a) and those derivedfrom it, is not greater than ± 2 decibels.

(b) Relation between Axial Sound Pressureand Alternating Current Input

In any high class reproducer the relationbetween the excess pressure set up in thesound field, when reflected waves are absent,and the alternating current input, shouldbe linear, otherwise the shape of the axialresponse curve will vary with the particularvalue of the constant testing input current.

Measurements were carried out to testthe linearity of the inductor loudspeaker,and in Fig. 7 is given a series of curves fordifferent frequencies relating the axial pres-sure at a distance of 6o cm. with the inputalternating current. It is seen that within

August, 1933

the limits of experimental error and for thesomewhat limited range of the independentvariable, the curves can be consideredlinear except for the results at 50 cyclesper second which exhibit distinct non -linearity. However, at this frequency theloudspeaker is cutting off, the pressuresinvolved are very small and the measure-ments are difficult to carry out and thus itis concluded that, for the majority offrequencies, the evidence furnished by Fig. 7is sufficient to show that no serious amplitudedistortion is present and that the shapeof Fig. 2 (a) is independent of the inputcurrent.

Additional measurements were made withthe loudspeaker connected by choke andcondenser in the anode circuit of the P2power valve, and the variation betweenthe a.c. input current to the loudspeakerand the a.c. input grid voltage to the valvewas examined. In this case the linearitywas again checked, the plotted pointsbeing even more consistent among them-selves than those shown in Fig. 7, and there -

426 THE WIRELESS .ENGINEER &

infinite baffle is regarded as the usefulradiation, then the percentage, acousticalefficiency is given by

Pa 0.150R2/32S-5

E113]f = [Z cos 01.X 10 . . (4)

6

Where R = radius at which polar curve ismeasured.

/3 = dyne/cm2/ampere factor on axisat radius R.

S = polar factor.Z cos B = effective resistance between in-

put terminals at a frequency, f.

The efficiency defined in this way forthe inductor loudspeaker has been deter-mined at a few selected frequencies and thevarious quantities are given in Table I.

If we consider the combination of theloudspeaker and amplifier from the pointof view of defining its effective attenuationin the electro-acoustical transmission systemwe can adopt Bostwick's" definition whichhas found favour,'5 and which is the ratio

TABLE I.EFFICIENCIES OF INDUCTOR LOUDSPEAKER.

Radial distance of microphone R = 6o cm.

Frequencycycles/second.- -

fldyne/cm2/

amp.s

0.750 R2g2Sx 10-7

watts/amp.

i for fullyloaded P2

valve : amp.0.150 R2p2S i8

x 10-7 watt.Z cos O. ns% n,%.

50 0.20 X 108 0.94 2.0 12.3 X 10-8 0.30 X 10-8 940 0.22 0.11I00 1.75 X 108 0.83 144 17.6 xfo-3 19.5 X I0-8 980 74.5 7.2500 7.68 X I08 0.34 51 10.9 x ro-3 6.1 X 10-3 1,480 3.4 2.3

1,000 I.fo X 105 0.34 22 8.2 X 108 1.45 X 10-8 2,290 0.95 0.552,000 4.2 X 103 0.103 700 6.o X 20-1 3.6 x 20-8 3,460 2.9 1.35,000 1.15 x Jo' 0.071 4.$ 3.7 x io-3 0.67 x 7o-4 7,370 0.065 0.025

fore it can be rightly concluded that thepressure set up in the sound field is propor-tional to the alternating voltage applied tothe grid of the power stage, provided rectifi-cation and grid current are avoided.

(c) Acoustical EfficiencyThe matter of efficiency has previously

received attention' and at that time it wassuggested that the efficiency of a loudspeakeralone might well be defined on orthodoxlines as the ratio of the total useful acousticaloutput power (Pa) to the total input electricalpower (P8) at the same frequency. If nowthe radiation from the front only of the

Pa 4RaPa= m2,92

Where Pa = measured useful acousticalpower radiated.

Pa = calculated maximum electricaloutput power.

Ra = internal resistance of powervalve at mean bias voltage.

of the acoustical power radiated to themaximum electrical power the amplifiercan deliver into a pure resistance loadmatching the output impedance of theamplifier.

To be more specific we will define theefficiency as

(5)

EXPERIMENTAL WIRELESS

mea

= magnification of power valve.= maximum r.m.s. alternating

grid voltage, taken to bethe same for both loud-speaker and non -reactiveload.

For the P2 valve actually used the measuredconstants yielded a value of

fit2e 2PE =

4Ra= o.27, watt. .. (6)g

where (eg). was 7.o volts for a steadybias voltage of - 10.2 and the correspondingcurrent through the loudspeaker at I,000cycles per second was 8.2 milliamperes,when the valve was fully loaded and gridcurrent was avoided. The calculated valuesof n, are given in the last column of Table Iand are, of course, proportional to Pa.

It is seen that although the absolutevalues of 7)1 are lower than those for theloudspeaker alone, they show less extremevariation. In fact, Fig. 5 (b) showing thevariation of output acoustical power indecibels is really an efficiency (n,) variationcurve and it is necessary only to fix theabsolute value of efficiency at i,000 cyclesper second. In practice the author hasfound this same curve with the levelquoted in acoustical watts output for unitvoltage on the grid of the power valve, witha statement of the maximum a.c. grid swing,to be the most useful curve as such a curvecan then be linked on with the extendedstudy of overall performance outlined inSection 3 (h).

It should be added that the values ofefficiency given in Table I should approxi-mately be doubled if it is assumed that theradiation from the back of the cone is equalto that from the front and that each can beseparately utilised. The values of n3 arechiefly of use in giving a concrete idea of theorder of the individual efficiency of theinductor loudspeaker, when it is studiedseparately as a transducer converting elec-trical into acoustical power.

5. Theoretical ComparisonsIt is of interest to compare the measured

polar distribution (Fig. 3) with the calculateddistribution for a vibrating rigid pistonoperating under similar acoustical conditions.It has been shown by Stenzel" and

427 August, 1933

McLachlan" that, at a radius several timesthe diameter of the source, the polar dis-tribution round a rigid piston working in aneffectively infinite baffle is proportional toJi (x)/x where J, is Bessel's function of thefirst order, and x is equal to kr sin a where

k = w/c = 27r/A, A = wavelength cm.r = radius of piston cm.a = angle between the normal to the

piston through its centre and thepoint considered.

If we take 2J1 (x)/x the axial value becomesunity and the values of the function fordifferent values of the angle a are thendirectly the relative pressures at thoseangles. The radius of the piston replacingthe inductor cone has been taken to approxi-mately midway between the edge of thepaper cone and the clamped leather of thesurround, and hence a value of r = 13.o cm.has been adopted. The calculated curves ofrelative polar distribution are shown inFig. 8 for the same frequencies as themeasured curves in Fig. 3.

ace 1.0 08 04 04 Oa 0 091

80'

70'

90.

04 06 08

00

1.01 1 1

50 ^../S C.

100 "./SEC

c.

1000 "../

5000,,/990.

1110 /

70

50'

40 30' 20' 10' 20' 30° 40

Fig. 8.-Calculated polar curves of relativepressures for rigid piston (D = 26 cm.) at

several frequencies.

It is seen that while the agreement is closeat low frequencies, the cone is definitely lessdirectional at high frequencies and that onaccount of the more complicated nature ofthe vibration of the cone there is not theuniform sequence in the curves of relativepolar pressure. Due to cone resonances, thepolar factor and the axial sound pressuresfor the cone are, in general, greater than forthe comparable piston case. Comparisonsbetween the cone and the piston have pre-viously been made by Cosens18 andMcLachlan and Sowter,19 but it is believedthat the data presented here are among thefirst15 to give measured and calculatedquantitative results, especially so far as the


polar factors are concerned, which for thepiston and cone are shown plotted togetherto logarithmic scales in Fig. 9.

In Fig. 9, while it is realised that eachobserved value of S is accurate within closelimits, the general tendency of the points

1'00 907

00

CO

MIIMMEIMMOINSIORIONRINNINIIWINIIIMIIIIIIIIMMINNMINI1 iFIMMOIMIIIMMORMI

ONNINIRR MI

INIMIN,11M1,111`

MENIMMEMENI M111111.1MbAlo0.

0C3 02

ENNIMEMON INIMMINSMIZIINMEMMEMMINIME11.111111.111...11111MINIMEMEMIIII11111111111111M1111111011111MMEMIIIIIII

cc 01 11111111111.11111111111111Vallial00'096.0'07

LL 0.05

e:1003

NIMImismemm 1111.....11111.11111111111IONN=IMERRROIMINI111111011111111111111

M',..C.11111111111

ommumminMINIMM1111.1111.11111

RIBUMBI IIIIMEN!!!INNIAMINIMINI

11111.111-111.111.1111=MEMMOi

UMWMIMIMMENNIIMMEIMMOil MIMMENNII

200 =M111111111111=1111111111111111111111M1111011111112111111111111111111111111111=1111111ON11111

.1111111111111.11111111.11111111111104100 1.000

FREQUENCY IN CYCLES PER SECOND

10,000

Fig. 9.-Polar factor variation with frequency,showing (a) Measured values for the cone, and(b) Calculated values for the rigid piston.

(D = 26 cm.).

has been indicated by the mean line, andwhen further similar results are available forcones of comparable size it will be interestingto see within what limits a typical polarfactor curve can be obtained for loudspeakersof the same general type and size. It maybe mentioned that at no point in Fig. 9 doesthe mean curve differ from the observedpoints by more than 1.5 decibels.

6. Legitimacy of the ObservationsA brief consideration shows that in the

foregoing measurements observations havebeen made over a frequency range such thatat low frequencies the diameter of the sourceis small compared with the wavelength ofthe sound radiated, while at high frequenciesthe reverse holds good. These points arebrought out in Table II and it is seen thatthe ratio of the wavelength A to the diameterof the source D varies by more than no :over the working range of frequencies.

In the last column of Table II the distanceR at which the greater number of observa-tions was made, is expressed in terms of thewavelength A. The ratio of the distance Rto the diameter D is 2.3.

It is customary in the case of a piston toassume that the directional effect is negligiblewhen the wavelength of the sound radiated

is greater than about four or five times thediameter of the source, and thus we canregard the loudspeaker cone as approxi-mating to a point source up to a frequency ofa few hundred cycles per second. Thisstatement is confirmed by the measuredpolar curves, Fig. 3.

Now it has been shown'. that for a pointsource it is valid to regard the acousticalpower as proportional to the square of theobserved sound pressure even at distancessmall compared with the wavelength and,therefore, it is concluded that even whenR/A becomes less than o.r at low frequenciesthe acoustical power can still be taken tovary as the square of the pressure, or theresponse curve to be represented by 20 log10(P/Pi000)

A further point arises at high frequencieswhen the wavelength is small compared withthe diameter of the source. In the case ofa piston there is a series of sound pressuremaxima and minima on the axis, due tosound arriving from different parts of thedisc, up to a distance 2° of approximatelyD2/4A after which the pressure tends to varyinversely as the square of the distance, andaccurately so at large distances. In our

TABLE II.

Frequency. A

cm.AID

D = 26 cm.R/A

R = 6o cm.

5o 68o 26.2 o.o88100 340 13.1 0.176200 170 6.5 0.35500 68 2.6 o.88

t,000 34 1.3 i.762,000 17 0.65 3.55,000 6.8 o.26 8.8

to,000 3.4 0.13 17.6

case D2/.4 A becomes 5o cm. for r = 13 cm.and a wavelength corresponding to a maxi-mum frequency of Io,000 cycles per second.The actual distance of 6o cm. is, therefore,sufficiently great to avoid these special inter-ference effects even at the highest frequencies,although at smaller distances there is no realevidence in Fig. 6 of their presence in thecase of the loudspeaker tested.

Finally, although the measured and calcu-lated curves of polar distribution, Figs. 3 and8, have been made as comparable as possible,it should be pointed out that the calculatedcurves derived from the formula 2J1(x)/x


only refer strictly to the case when the ratioof RID is greater than about 5, but for theobservations in Fig. 3 the ratio of RID is 2.3.

To some extent the question of linearityhas been dealt with in section 4 (b) but to copewith all the volume levels at which the loud -'speaker might be expected to work in prac-tice would make an examination of linearitydesirable over ranges in the variables havingconsiderably greater extreme ratios.

Closely connected with the general ques-tion of linearity is the further one of har-monic distortion. It was checked that fora pure electrical input the acoustical outputwas also pure by ear, and at certain fre-quencies attempts to measure the harmoniccontent were unsuccessful on account of thesmallness of the quantities involved. Suffi-cient evidence was obtained, however, toconfirm that the harmonic distortion was

August, 1933

negligible by ear except at very low fre-quencies and that the measurements at onefrequency were not affected by the integra-tion of the overtones by the condensermicrophone except in a small degree atfrequencies below about zoo cycles persecond.

REFERENCES1. Oliver, E.W. and W.E., VII (1930), 853.2. Dreisbach, Radio Broadcast, June (1929), 94.3. Oliver, W.W., XXIX (1931), 579.4. Davis and Evans, Proc. Roy. Soc. A., 127 (1930), 89.6. Oliver, Jour». Scien. Instr., VII (1930), 113.6. West, W., Journ. I.E.E., 67 (1929), 1137.7. Knudsen. Journ. Acmes. Soc. Am., IV (1932), 20.8. Crandall, " Theory of Vibrating Systems and Sound," 224.9. Barnes, " Acoustical Engineering," West. W. (1932), 109.

10. Ballantine, Journ. Acous. Soc. Am., III (1932), 1.11. Crandall, " Theory of Vibrating Systems and Sound," 210.12. Oliver, Nature, 128 (1931), 288.13. West, A. G. D., Cantor Lectures, Roy. Soc. Arts., Mar., 1931.14. Bostwick, Bell Sys. Tech. Journ., 8 (1929), 135.15. Turner, P. K., Journ. I.E.E., 69 (1931), 591 ; Hughes, 617.16. Stenzel, E.N.T., 4 (1927), 239.17. McLachlan, Proc. Roy. Soc. A., 122 (1929), 604.18. Cosens, E.W. and W.E., VI (1929), 353.19. McLachlan and Sowter, Phil. Mag., XI (1931), 1.20. Wood, A. B., " Sound," 148.

CorrespondenceLetters of technical interest are always welcome. In publishing such communications theEditors do not necessarily endorse any technical or general statements which they may contain.

The Principles of ElectromagnetismTo the Editor, Wireless Engineer

SIR, -Professor Fortescue's letter in your Juneissue suggests that he has not really grasped thenature of the difficulty that he has raised, and thisview is supported by the absence of any answerto my questions.

May I put the matter to him in another way ?There are three experimental relationships which arefundamental in electromagnetics, namely :-

A. The inverse square law of mechanical forcebetween like charges.

B. The inverse square law of mechanical forcebetween like poles.

C. The law of mechanical force between amoving charge and a magnetic pole.

Each of these involves a factor of proportionalityas between the mechanical force produced and theelectrical or magnetic quantities involved. Thefactor in (A) is x/c, in (B) and in (C) we maycall it k. Obviously these three are interdependent,the connection being that Plait = (velocity)2.

All physicists and electrical engineers, exceptProfessor Fortescue, agree in admitting thatthere is as yet no information available which willallow actual dimensions to be allotted to e, /1. and lg.Pending such information, p. convention mustbe adopted. Three such conventions are obviously

possible, namely :-I. To put c = i/v22. To put ti = qv'3. To put k = v.

The first gives rise to the ElectromagneticUnits, the second to the Electrostatic Units and thethird to a system which I think is more rationalthan either, but, as this has never been adopted,it need not concern us here.

If we adopt (i), then B and H are quantitiesof the same kind, and one symbol will suffice.

If we adopt (2), then electric force and electricdisplacement are magnitudes of the same kind andone symbol only is necessary.

Professor Fortescue says definitely that " B andH are the same thing." It follows that the wholeof the dimensions of all the electric and magneticunits are thus settled once and for all, and onlyone system of units is possible. This is splendid;but it is a matter of such paramount importancethat we must be pardoned if, while admitting thatProfessor Fortescue may be right, we hesitate toaccept his dictum until we have access to " thosesecrets of Nature not yet revealed to the restof us "-upon which any such conclusion mustinevitably be based.

WILLIAM CRAMP.The University, Edgbaston, Birmingham.


The Use of the Cathode-ray Oscillographat Ultra -high Frequencies*

By Dr. H. E. Hollmann(Heinrich -Hertz Institute for High -frequency Research, Berlin)

IntroductionIN comparison with oscillographs with

mechanical systems, the cathode-ray oscil-lograph can, of course, be regarded as free

from inertia. But this freedom from inertiais not absolute, since the electrons whichpass through the anode aperture and formthe indicating or recording system of theoscillograph move with a finite velocitydepending on the anode potential. Owingto this fact, the cathode-ray oscillographhas a certain inertia of its own, and this isobviously greater, the longer the path -timeof the electrons compared with the durationof the process which is to be analysed. If,therefore, either the electron speed de-creases or the process to be recorded occursfaster and faster, the inertia of the cathode-ray tube will show itself in altered sensitivityand in distortion of phase. With the usualcommercial tubes working on some hundredsor thousands of volts on the anode, suchdepartures from the normal behaviourmake their appearance at wavelengthsmeasured in decimetres, and require specialmeasures for their suppression. Thesetwo problems, that of the " dynamicsensitivity " at very high frequencies andthat of phase distortion, will now be dealtwith in turn.

I. THE DYNAMIC SENSITIVITY OF THECATHODE-RAY OSCILLOGRAPH

The deflection of the cathode ray in aBraun tube is generally produced by causingthe ray to pass through an electric field, dueto a voltage V, between two deflecting plates.In this way the ray is deflected from itsoriginal path in the manner shown in Fig. 1,a.ccording to the same law as that of thetrajectory, where a horizontally projectedbody describes a parabolic path under theinfluence of gravity. There is one difference

* MS. received by the Editor, July, 1932.

in the case of the Braun tube, namely, thatthe deflection only occurs during the passagebetween the deflecting plates ; after leavingthese, the ray proceeds again in a straightline, on a path inclined to the axis at theangle a, giving a deflection a on the fluor-escent screen. We will define the Sensi-tivity (A) as the tangent of this deflec-tion angle a, or in other words as theratio vv/vo of the velocity component vy,perpendicular to the axis, to the velocity

Y

b

Fig. 1.

component v0 in the axial direction. Thesevalues can be calculated by the simpleadaptation of the horizontal trajectory lawto the case of the cathode-ray tube: thusfor the axial electron motion we have :

x = vo t . . (1)

where vo represents the electron speedcorresponding to the anode potential(v0 =6 . io7A/Ea) and t is the time. In the

y axis, the law of the falling body applies,namely :

= gt2where the " acceleration " is

g = e/m . V lb.

Here e and m are the charge and massof the electron, respectively, and V lb isthe field strength given by the deflectingpotential and the distance between theplates. From equations (I) and (2) we obtainthe trajectory parabola :

= x2002

.. (2)

represented by :

EXPERIMENTAL WIRELESS 431 August, 1933

and from this the static sensitivity :Ao = tan a = dyldx = g xlvo2 .. (3)

From this we see that the static sensitivityis the greater, the longer the deflectionplates (i.e., the greater the value of x) andthe smaller the anode potential (and con-sequently vo) ; also, that A, and consequentlythe screen deflection a of a given tube aredirectly proportional to the " acceleration "g and therefore to the deflecting voltage.If, therefore, an a.c. voltage Vo sin wt is appliedto the plates, the deflection a will exactlyreproduce the voltage curve.

The above results, however, only holdgood so long as the deflecting field is con.stant while the electrons are between thedeflecting plates. If, on the other hand, theperiodic time of the deflecting voltage iscomparable with the time necessary for thepassage past the plates, the transverseaccelerating field can no longer be regardedas quasi -stationary with respect to theelectron beam. We will consider, by theexample shown in Fig. 2, how the conditionsalter in such a case : taking an electronwhich enters the space between the platesjust at the moment when the instantaneousvalue of the deflecting potential is zero.During the passage of the electron, thedeflecting potential passes first through apositive and then through a negative half-period, so that the potential distributionshown on the upper plate in Fig. 2 appliesto this electron..

Fig. 2.

During the positive half -period the electronpath is deflected through the angle a'.The negative half -period deflects the electronthrough an equal but opposite angle a", sothat in the complete period the effects of thetwo half -periods neutralise each other andthe electron emerges from the plates oncemore in its original direction. The displace-ment y can be neglected in comparison with

the screen deflection a. Another point whichcan be seen from Fig. 2 is that the deflectionis greatest when the electron has just passedthrough a half -period : this, then, is thecondition for maximum sensitivity of thetube, whereas when the passage -time isequal to the whole periodic time the sensi-tivity completely vanishes.

y

- --27T

- - -7-___ --1.0(

4

Fig. 3.

wt

-0-

-11/2

-37T/4

The above results can be representedmathematically by combining the temporalvariation of the deflecting field (given bywt) with the spatial variation of the electronspassing between the plates. If we take thedeflecting potential at the moment whenthe electrons enter the plate -gap, i.e., whenx = o, then the spatial distribution is to bedefined by the electron path time 8 = x/vo,or by the phase angle (1) = w8. Hence :

g = e/m .Volb . sin (wt ib), or

= go sin co (t + 8) (4)

From this, by double integration in themanner shown in (I) of the Appendix, weobtain as the equation for the electronmotion :

y = golw2 (w6 cos wt - sin w(t + 8)sin wt) . . (5)

If we then plot y as a function of 8, or ofx = v 0 . 8, with wt as parameter, we obtainpath curves of which examples are givenin Fig. 3. These show that at the pointswhere w8, or wx/vo, = v and 31r, the de-flection angle displays maximum values,while for 2v and 477. it becomes zero.

August, 1933

It is shown in (2) of the Appendix that the" dynamic sensitivity " at a frequency co isgiven by :

.4 , sin (cot 95/2) . sin 95/2/0/2 .. (6)

This equation shows thatthe dynamic sensitivity, andhence also the screen deflectiona, is proportional to sin cot ;that is to say, the tube func- 6

tions without distortion evenat the highest frequencies. Ifwe leave out of considerationthe phase relation betweenthe deflecting potential andthe screen deflection a, wehave, for (cot 0/2) = Tr/2,the maximum dynamic sensi-tivity 4.0 given by : 4

A0. sin 0/2/0/2 .. (7)

This sensitivity, correspondingto the peak voltage Vo, thusdecreases gradually with in-creasing frequency. At definitepoints, where 0 = 2n7r,, itperiodically becomes zero, and reaches at theintermediate points-where 0 = (2n - i) 7r -ever -decreasing positive and negativemaxima: here n represents the successionof whole numbers. For low frequenciesand corresponding small phase angles, A

03

1


at which the tube ceases to function, lie atn . 0.109 cycles per second, corresponding towavelengths 31-6/n (in millimetres).

If we introduce into equation (7) the

- .

...------

'' Aw0=f(Ea)///

A.50cm.1= 2cm.

4

2

0

2

6

becomes equal to A 0.

03

0.

0'5

8400

Ea VOLTS

Fig. 5.

900 1,600 2,500

expression for A, given in equation (3), andfor 0 write cox/v0, we obtain :

A 0 = gox/v02 . sin cox/ 2v box/ 2z,= 2g0/cov0 . sin wx/2z,0 . . .. (8)

In this equation the plate length x onlyoccurs in the argument of the sine function,

so that for a given fre-quency and ray velocitythe optimum plate lengthto give maximum dy-namic sensitivity is givenby :

xopt -= wv,/co . Replacingit by the half -periodcoT/ 2 = coA/ 2c, we have :

x0, = volt . A/2 . . (9)

Thus the plate length

Aceo.f(v)Ea= 1,000 VOLTS

= 2cn.

0

0'5107 2 4 6 8 108 2 4 6 8 10

CYCLES PER SECOND

Fig. 4.

9

Fig. 4 shows the curve representingfunction (7) for an anode potential of I 000volts and a plate length of 2 cm, A, beingtaken as unity. The critical frequencies,

giving maximum dynamicsensitivity corresponds to

2 4 6 8 1010 a half wavelength re-duced in the ratio v0/c,as has already been seenfrom a consideration of

the electron paths in Figs. 2 and 3. Anincrease in the length of the plates beyondthis value merely means a capacitive over-loading of the deflecting potential. This


relation also explains the falling off ofdynamic sensitivity with increasing fre-quency, since it shows that as the wavesgrow shorter so also does the most efficientplate length decrease. For a given lengthof plate and a given frequency, the dynamicsensitivity presents itself as a function of theray velocity or of the anode potential, andyields-according to equation (8) --the curveshown in Fig. 5. The curve displays a seriesof maxima with intermediate zero points.A practical result is that for any tube amaximum of 'sensitivity can be obtained byvarying the anode volts. Nevertheless it ispreferable to employ the higher voltagesin spite of the lower static sensitivityconsequent on their use ; for as the examplein Fig. 5 shows, above i 000 volts a uniformsensitivity can be counted on, without thefear of encountering points of zero sensitivity.

APPENDIX(r) Calculation of electron motion in an alter-

nating accelerating field :g = go sin w(t 8) (4)

The integration over d8 gives the deflection speedin the y-axis :

v, = - go/w . cos cii(t ± 8) +A .. (12)

At the moment of entry of the electrons between theplates, that is when x, and consequently also when8, = o, v, must also equal o :

v. = go/w (cos wt - cos w(t + 8))By integrating again we obtain for the transversemotion :y = go/w (w cos ad - I /w . sin w(t ± 8) + B) (14)

Since here again at the time 8 = o y must alsoequal o, we obtain for the integration constantB = . sin cut, and hence :

y = go/w2. (w8 cos wt - sin (.0(t + 8)+ Sill wt) . . (5 : see p. 431)

(2) The dynamic sensitivity, from equation (i3),is given by :

A., = vilvo = golawo. (cos cot - cos w(t + 8)).This can be written :

A., = 2golun),,, sin w(t + 8/2) sin w8/2 .. (15)

In order to introduce into this the static sensitivityAo, we write (from equation 3)

go = Aoveix A0v0/8,and obtain :A., = Ao sin w(t + 8/2) sin w8/2/w8/2 .. (6a)

Ao sin (wt + 0/2) . sin 0/2/ 0/2.. (6 : see. p. 432)

(The second part of the paper, to bepublished next month, will deal with theproblem of phase distortion and with gas -concentration at very high recording speeds.)

(13)

August, 1933

Book ReviewAlternating Current Electrical Engineering

By Philip Kemp. 595 + xi pp. with 418 Figs.Macmillan and Co. 15s.

In the preface the author says that in preparingthis fourth edition he has expanded the bookto include many of the more important of thelatest developments and that to keep the bookwithin bounds he has had to do some carefulpruning. We must say that what he has donehas been done remarkably well, and the book isone that can be unreservedly recommended toevery student of electrical engineering. The authorhas adhered to principles rather than describedparticular types of machinery, but the applicationsto practice are never lost sight of. On looking throughthe book we have queried two or three minormatters. The opening section of Chapter III isentitled " Back E.M.F. set up due to alternatingcurrent," but there is nothing in the section tosuggest why it should be called a back E.M.F.It is stated correctly that the direction of theE.M.F. is such as to oppose the change of current,so that during the half cycle that the current isdecreasing, the induced E.M.F. is in the samedirection as the current.

It is only when one considers the induced E.M.F.in relation to the applied P.D. that one is justifiedin regarding it as a back E.M.F.

The regulation of alternators is treated veryfully, but we cannot understand Fig. 234, whichpurports to show the variation of synchronousreactance with position. Synchronous reactanceis obtained from the open -circuit and short-circuitcharacteristics, both of which are measured whenthe machine is running at its synchronous speed ;it is a fictitious or equivalent reactance coveringthe combined effects of armature reaction andreactance. It is true that if the machine be stoppedand the reactance of the armature measured bymeans of an A.C. supply, the result obtained willvary with the position of the armature with respectto the poles, but this is not the synchronous reactance.In future editions Fig. 234 might be entitled" variation of stationary reactance with position."

The third point refers to the simple circle diagramof the induction motor, where in Fig. 282 the authormakes the common mistake of showing a constantmagnetising current for all values of stator androtor current. As a matter of fact, as the loadincreases the flux crossing the air -gap decreasesuntil on short-circuit (locked -rotor) about half theflux crosses the gap and the magnetising ampere -turns are consequently about half of those requiredon open circuit. For this reason the author iswrong in saying that " I, is constant in magnitudeso long as 4). and E. are constant." So long asthe stator voltage E. is constant the stator flux G,will be constant, but the fraction of this whichcrosses the air -gap to the rotor will be reduced toabout half, and with it I as one moves roundthe circle. Fifty per cent. is not a negligibleamount and it requires very little modificationof the diagram. We commend this suggestionto the author for use in the fifth edition, whichwe feel sure will soon be required.

G. W. 0. H.


Naval Direction -findingThe Meeting of the I.E.E. Wireless Section on 5th April was devoted to Two Papers on

Naval Direction -finding by Members

The first paper was : " Errors in Direction -finding Calibrations of Steel Ships, due to theShape and Orientation of the Aerial of theTransmitting Station," by Mr. J. F. Coales, B.A.

IN making calibrations of direction -finders onships, in which a transmitting ship steams roundthe vessel to be calibrated, errors were suspected

due to the orientation of the transmitting aerial.These were thought to arise from the horizontalelectric intensity-with which is associated avertical magnetic field-sent out by the horizontalportion of the transmitting aerial. The theoryof the errors so produced is that in one case thevertical magnetic field produces in the effectiveloop of the ship (which loop in the presence of thehorizontal magnetic field alone produces a quad-rantal deviation) a current proportional to a, 1H 2.This current gives rise to a magnetic field in thedirection of propagation of the wave which inturn causes an error e in the observed bearing.

The paper then discusses the simple theory ofthe effect. It can be shown that if the normal tothe effective loop of the ship makes an angle 4i withthe vertical, and its projection on the horizontalplane makes an angle 0 with the direction ofpropagation, then e, the error in the absence of avertical magnetic field H2, is given by

kA to sin 20 sins tfrtan e - approximately2(1kA sin2 sin20)

where A is the effective area of the loop and k aconstant depending on the position of the direction -finding aerial system.

In the case of a vertical magnetic field (Fig.1)* being present as well as the horizontal field H,such that these are in phase and orientated as inFig. I (a), it is easily shown that el, the error of theobserved bearing in this case, is given by

A kw sin tit cos 0 (H, sin ./s sin 0 He costan El= H1+ Aka, sin 1,b sin 0 (Hi sintii sin 0 ± Ha cos 5b)

and in the case shown by Fig. (b), the direction ofpropagation being vertically upwards from thepaper in both cases, the error e2 is given by

Aka, sin tfr cos 61 (II sin d sin B - H.. cos 0)tan ea= Hi+ A ktosin 1,6 sin 0 (Hl in (Pin 9 - Ha cos

In most ships the error when the transmittingstation is right ahead is zero, in which case 9 isalways the relative bearing of the transmittingstation. tis is the angle the axis of the loop makeswith the vertical and so is constant, and we get thatcalibrations in the two cases should be separated byan angle (e1 - e2) with the normal calibration curvelying between them. The angle of separation(el - e2) will depend on cos B and so will be greatestright ahead and astern, and zero on the beams.

* The author's original figure -numbers are adhered to throughoutthese abstracts.

of the Staff of H.M. Signal School

Since the second term of the denominator willusually be small compared with HI, it is seen thatthe value of (el - e2) will be approximately pro-portional to the frequency CU and to the intensityof the vertical magnetic field H2, which latter israpidly attenuated over sea since the correspondingelectric field is horizontal. This gives a possiblemethod -of measuring the attenuation of the hori-zontal electric intensity over sea water.

p.14

Ns

Fig. 1.

This was tested by a transmitting ship steaminground the receiving ship first in a clockwise andthen in a counter -clockwise direction, the trans-mitting ship being a destroyer with L-shaped aerial,the vertical leg being 45 ft. and the roof 125 ft.long, the mean height of the roof being 5o ft. abovethe sea. The results show that if the receiving shipis not in line with the roof of the transmitting aerial,errors of 1° or 2° may arise if the roof is not sym-metrical. The error, however, decreases rapidlywith increasing distance from the transmitter andalso with increasing frequency. The magnitude ofthe error depends on the shape and orientation ofthe transmitting aerial and the shape and orienta-tion of the " effective loop " of the receiving ship.

The paper closes with a mathematical analysisfrom which it is concluded that in big ships thesecalibration errors will be very much more markedand that for such ships it is important that thetransmitting aerial is of such a type that the emittedwave contains no horizontal component of electricforce.

The second paper was : " A Radio CompassDeveloped in H.M. Signal School," by Messrs.C. E. Horton, M.A., and C. Crampton, B.Sc.This describes a wireless direction -finder freefrom ambiguity which has been developed foruse in ships as well as on shore.

THE usual direction -finding practice is to takethe bearing, using a figure -of -eight character-istic and to resolve the ambiguity by switching

in an open aerial and obtaining a heart -shapedcharacteristic approximating sufficiently to a truecardioid to settle the sense of the bearing. This in-volves the provision of a separate " sense " pointerto allow for the displacement of 9o° between the zeroof the figure -of -eight and the zero of the cardioid.It also involves a loss of time, due to the need for


two operations. The direction -finder describedovercomes both of these disabilities by using whatamounts to a distorted figure -of -eight having onesharp zero giving the true bearing and 180° fromthis a blurred zero giving the reciprocal bearing.This is achieved, as shown in Fig. r, by taking asymmetrical figure -of -eight and superposing asmall cardioid under conditions which ensure thatthe minimum of the cardioid coincides with onezero of the figure -of -eight, and that the cardioide.m.f. is in quadrature with that of the figure -of -eight. The cardioid need only be very small inamplitude compared with the main figure -of -eightcharacteristic and need not therefore in itself be aperfect cardioid in order to blur the reciprocalbearing, while leaving the true bearing unaffected.Further, by ensuring that the cardioid is in quadra-ture with the figure -of -eight, it becomes impossiblefor the cardioid to produce an error of bearing, i.e.,the zero of the figure -of -eight cannot be shiftedby the superposition of the cardioid.

If the figure -of -eight response is obtained froma rotating frame F (Fig. 2) the auxiliary cardioidcan be obtained from a smaller frame F2 (at rightangles to and revolving with F1), together with astationary open aerial V. If the cardioid is im-perfect because the response from F2 is not exactlyequal to that from V, then one zero of the figure -of -eight is blurred and the reciprocal bearing is moredefinitely blurred. There is, however, no error of

Reciprocalbearing True bearing

Fig. I. Addition of a cardioid to a figure -of -eightquadrature.

bearing. On the other hand, if the cardioid isimperfect because of a defect of phasing, thenpossibility of an error of bearing arises. But bykeeping the amplitude of the cardioid small com-pared with the amplitude of the figure -of -eight, theerror produced in this way is a small quantity ofthe second order.

The paper then briefly discusses the componentsof e.m.f. active in a direction -finder on ship -board(as given in a previous paper by one of the authors*),and by numerical examples it is shown that thechief features of the new direction -finder can be

* 1931, Vol. 69, p. 623. E.W. & W.E., 1931, Vol.VIII, p. 195.

August, 1933

summarised as follows :-(i) A perfect zero isobtainable without ambiguity of 180°. (2) Thedeviations (when installed on a ship) do not followany very simple law, e.g., quadrantal. (3) The

V

Openaerial

2

2

To H.T. battery

To amplifier

Fig. 2.

amount of correction to be applied to observedbearings has to be determined by calibration, aswith other types of direction -finder on ship -board.(4) The required values of thecoupling M2 (Fig. 2)all lie between two positive upper and lower limits.

The instrument is thus capable of giving accuratebearings free from ambiguity, when allowance ismade for deviation, as in ordinary ship -boardpractice. Electrical methods have been describedin the previous paper by means of which deviationcan be corrected on certain types of ordinarydirection -finders. These can be further appliedto cover deviations of the new system. Alter-natively, with simple rotating -coil direction -findersa common practice is to read from a correction -curve, which can also be modified to include thedeviations of the new system. Details of methodsof correction are further discussed in the presentpaper, which then proceeds to describe a practicalinstallation on the rotating coil system, illustratedin Fig. 4. The loop L is the corrector -loop, placedathwartships. F2 is the main rotating frame coil,forming part of the inductance of the tuned re-ceiving circuit. To obtain the e.m.f. induced byF in the correct phase, this coil is left untuned, butits e.m.f. is amplified by coupling through themutual inductance MI to a two -stage balanced -valve amplifier with resistance -capacity coupling.The output from the amplifier is coupled mag-netically to the receiving circuit, as shown in Fig. 4.By careful construction of special valves and re-sistance -capacity elements in vacuo, it has beenpossible to obtain an amplifier giving appreciablegain over the whole range of frequency covered(the voltages after amplification being comparablein magnitude with those from Fl) and which changesthe phase by only a small angle. The circuitsemployed with the corrector -loop L are similar tothose of F2, similar phasing being required. Inthis case, however, comparatively small e.m.f.s are

August, 1933

required and the resistance -capacity stage isomitted.

A description of sea trials of the new instrumentand of methods of adjustment are lastly given.The sea trials were carried out in the vicinity of theNab Tower in sight of Culver Wireless TelegraphStation, from which all the required transmissionswere made. Correction curves to relative direction -finder bearings are shown in the paper for severaldifferent frequencies, their appearance beinggenerally similar to those of quadrantal -errorcorrection curves.

DiscussionThe discussion which followed the reading of

the papers was relatively short.MR. G. SHEARING stated that the need for the

analysis given in the first paper arose from tbe factthat a second calibration often gave differentresults from the first. In practice, it was importantto get a short -distance calibration correct and inagreement with that of long-distance reception.The radio compass described in the paper had con-siderable advantages in practice. It was simple toadjust on land or on sea and very quick in operation.The trials described in the paper were made undermore exacting conditions than those normal tomercantile marine working.

M2


more rapidly than the author suggested. Asregards the radio compass, he referred to the smallphase -change necessary to give an error of bearing,and to the difficulty of aural determination of therelative change of intensity corresponding to thisphase -change.

DR. SMITH -ROSE enquired as to the possibilityof the radio compass device in application to arotating beacon transmitter. This, on a fixedfrequency, was the most favourable case for makingoptimum adjustments. Mr. Coales' analysis mightbe applied to the case of a flat-topped aerial usedin the reception of the rotating beacon.

MR. MCPHERSON said the compass described wasfor manual operation on a " minimum " signal andenquired as to its operation on " maximum "observation with direct reading.

MR. BAINBRIDGE BELL asked why the " effectiveloop " of the ship was not vertical. As regards theradio compass, how was it ensured that no changeof the sense indication ever occurred in subsequentservice. Why not use a vertical aerial instead ofone which necessitated the transmitting shippointing to the receiver ?

The authors briefly replied to the discussion.Mr. Horton pointed out that the cardioid systemwas untuned and that phase changes due to detunedid not arise. The sense connections were not

0

Fig. 4.-Circuit diagram of radio compass.

DR. C. V. DRYSDALE also referred to the practicalimportance of naval direction -finding, and to theneed for rapid operation. The papers showed thefundamental manner in which the problems hadbeen tackled.

MR. R. H. BARFIELD expressed surprise at theaccuracy of the author's calibrations. He thoughtthe field due to the horizontal roof should disappear

To tuner &amplifier

touched in subsequent operation after beingcorrected by the calibrating officer. Theradio -compass principle could be applied to therotating beacon, although this had not been doneso far.

On the motion of the chairman, Col. A. S. Angwin,D.S.O., the authors were cordially thanked fortheir papers.


Abstracts and ReferencesAugust, 1933

Compiled by the Radio Research Board and reproduced by arrangement with the Departmentof Scientific and Industrial Research.

PAGE PAGEPropagation of Waves .. 437 Directional Wireless .. 451Atmospherics and Atmospheric Acoustics and Audio -Frequencies.. 452

Electricity 440 Phototelegraphy and Television .. 453Properties of Circuits 443 Measurements and Standards .. 456Transmission 443 Subsidiary Apparatus and Materials 457Reception .. 446 Stations, Design and Operation .. 460Aerials and Aerial Systems . 449 General Physical Articles .. .. 461Valves and Thermionics 450 Miscellaneous 461

PROPAGATION OF WAVESINTERACTION BETWEEN RADIO -WAVES ?-B. D. H.

Tellegen. (Nature, loth June, 1933, Vol. 131,P. 840.)

This letter gives a short description of apparentinteraction between waves from the Luxembourghigh -power broadcasting station (wavelength

190 m) and from a series of other stations,observed at Eindhoven. A radio receiver tunedto the other stations shows the modulation of theLuxembourg station in the background. Thephenomenon cannot be due to cross -modulationof the receivers. The stations with which it occursall lie in the direction from Eindhoven to Luxem-bourg and further from Eindhoven than Luxem-bourg. This modulation has also been observedon other stations.

SUR LA VITESSE APPARENTE DES ONDES RADIO-ELECTRIQUES COURTES (The ApparentVelocity of Short Radioelectric Waves).-N. Stoyko and R. Jouaust. (ComptesRendus, ist May, 1933, Vol. 196, No. 18,pp. 1291-1292.)

Hitherto the apparent velocity of short waveshas only been determined by comparison with theknown apparent velocity of long waves ; the resultobtained was 272 700 ± 5 700 km/sec. The writers,by the reception, at Paris and Buenos Ayressimultaneously, of time signals from Pontoise(28.35 m) and Monte Grande (15.35 m), havenow made direct measurements giving a value of269 700 ± 4 400 km/sec. Assuming a true velocityequal to that of light, and an emission angle givinga minimum number of reflections (1931 Abstracts,p. 434) between earth and a layer at 30o km,this corresponds to 8 reflections between Paris andBuenos Ayres and an angle of 19°.8 between wave -direction and horizon.

By an identical method they have also measuredthe apparent velocity of long waves, with a resultingvalue of 245 000 km/sec. This difference betweenthe apparent velocities of long and short wavesmay be explained by the long waves travelling,rather like a sound wave in a tube, between theionised layer and the earth : owing to the differentelectrical resistances of the two conductors thewave front undergoes a spreading out, and theelectric field at the receiver does not suddenly

take up the value corresponding to the strengthof the signal but builds up gradually until it isstrong enough to act upon the receiver : thisinstant is taken to be the " arrival " of the signal.

ANOMALIES IN THE PROPAGATION OF SHORT RADIO -ELECTRIC WAVES [Abnormally Long PathTime for Annapolis Time Signals receivedin France].-N. Stoyko and R. Jouaust.(Comptes Rendus, 22nd May, 1933, Vol. 196,No. 21, pp. 1583-1585.)

The results of five months' observations on these37.36 m signals indicate delays with two pointsof convergence, dl = 0.00676 0.00078 sec. and

= 0.02943 ± 0.00107 sec. The Annapolis beamis directed towards the west : owing to the earth'scurvature it is possible for a ray to penetrate theionised layer that it suffers a strongdiffusion which sends a part of the energy back-wards (cf. Eckersley, Abstracts, 1929, p. 625,1-h col.). This diffusion may occur at the firstcontact with the ionosphere or after one or twonormal reflections : di would correspond to afirst contact, 42 to a diffusion after two normalreflections. Assuming for the angle between thedirection of propagation and the horizon the valuededuced as in the paper dealt with in the precedingabstract, the layer heights calculated from theobserved delays are 400 ± 50 km for z11, and340 ± 15 km for 42.

The distance of the first diffusion point fromAnnapolis, along the great circle, is about goo km.This is the region of the Great Lakes. The seconddiffusion point is at a distance of 4 000 km fromAnnapolis, corresponding to the coast of the PacificOcean. " Can these anomalous dispersions perhapsbe explained by the fact that above regionscorresponding to discontinuities on the surfaceof the earth there exist distortions in the higherionised layers ? At any rate such results as theseshow that very serious errors may result, in certaincases, from the use of short waves for the determina-tion of longitude."

THE REFLECTING LAYERS OF THE UPPER ATMOS-PHERE [and Changes in Barometric Pressure].-R. C. Colwell and I. 0. Myers. (Phys.Review, ist May, 1933, Series 2, Vol. 43,No. g, pp. 774-775: abstract only.)

See June Abstracts, p. 321, 1-h column. The

August, 1933

writers have made, at West Virginia University,simultaneous observations on short-wave propaga-tion from W8XK (6 140 kc/s) and long -wavepropagation from KDKA (98o kc/s), both of whichstations are located at Saxonburg, Pa. Theyplotted the fading curves obtained from U.S.Weather Bureau maps and found at once that" the signal strength of the long waves variesafter nightfall according to the change in barometricpressure." The short waves have the samecharacteristic curve whatever the weather. " TheF (Appleton) layer covers the' sky in a more orless uniform sheet, while the E (Kennelly -Heaviside)layer is concentrated in the regions of low pressureand is most active in the eastern half of the cyclone."

DER EINFLUSS DES WETTERS AUF DIE AUSBREI-TUNG DER RADIOWELLEN (The Influence ofthe Weather on the Propagation of RadioWaves).-J. Fuchs. (Funkmagazin, June,1933, Vol. 6, pp. 363-370.)

Further development of the work dealt with in1930 Abstracts, p. 152, 1-h col. As an exceptionto the statement that processes in the tropospherecannot directly affect the propagation of spacewaves, owing to the absence of sufficiently highionisation, the writer cites the ionisation in theneighbourhood of a thunder -cloud just before theoccurrence of a lightning flash : he mentions Stoye'sobservations of d.f. errors setting in just before aflash and disappearing the moment the flashoccurred, and also the recent observations on theincrease of cosmic -ray intensities below thunder-clouds (e.g., Schonland and Viljoen, under " Atmos-pherics "). The ion densities produced by normalcold and warm fronts are comparatively seldom highenough to affect propagation : usually, when suchfronts produce changes in signal strength, it is theground wave which is affected as a result of con-ductivity changes due to rain, etc.

The latter part of the paper deals with the in-fluence of the stratosphere, discussing whetherchanges occur there (besides the well-known dailyand yearly periods, magnetic disturbances, etc.)which are related to the conditions in the tropo-sphere. The formation of ozone at about 6o kmand its descent to a height of from zo to 3o km, andthe fairly definite relation between the ozonecontent in this lower layer and the atmosphericpressure at the earth's surface, are quoted. More-over, the writer showed in 1928 that the propa-gation of short waves over distances of the order. of7 000 km was connected with the atmosphericpressures at the surface along the path, althoughthe active regions for the short waves are far abovethe- ozone layer. The writer inclines to Earner'sview of convective (vertical) and advective (hori-zontal) air currents in the intermediate regions,simultaneously producing pressure changes below,at the earth's surface, and risings and fallings of theair masses above ; thus causing as secondaryphenomena the variations in reflection and re-fraction which affect propagation. Such currentswould be produced, for example, by the heating ofthe ozone layer at 20-30 km. The correlationbetween weather and propagation is likely to giveinconsistent results owing to the effects varyingaccording to the height at which the disturbingaction occurs : if it occurs in the troposphere the


relation is simple, a low-pressure system producingan increase, and a high-pressure system a decrease,of the skip zone. The writer's 1928 results were anexample of this, the path lying over the N. Atlantic,where the troposphere is strongly disturbed underthe influence of the Gulf Stream.DOUBLE -REFRACTION EFFECTS IN THE KENNELLY-

HEAVISIDE LAYERS.-H. R. Mimno andP. H. Wang. (Phys. Review, ist May, 1933,Series 2, Vol. 43, No. 9, pp. 769-770.)

A preliminary note on work on the continuousrecording of echoes from the ionosphere on a fre-quency of 4 095 kc, with the receiver distant about14 km to the south of the transmitter. Magneto -ionic split echoes have been observed in agreementwith the results obtained by other workers (cf.Abstracts, 1932, pp. 275-276, Rukop and Wolf ;May, p. 263, Wolf ; Jan., pp. 29-3o, Eckersley, andAppleton and Ratcliffe). During disturbed con-ditions the curves of effective height on the recordsare often found to intersect at an angle. Theapparent presence of a third component in the firstorder reflection has been noted. A record of theeffective height is given showing violently dis-turbed conditions. The writers have also notedstrong multiple reflections from E 'layer occurringseveral hours before sunrise. They have alreadyused and discarded the synchronised " grid -blocking " transmitter for sending short pulsesdescribed by Herd (May Abstracts, pp. 263-264).STUDIES OF THE IONOSPHERE AND THEIR APPLI-

CATION TO RADIO TRANSMISSION [Measure-ments of Critical Frequencies : the F Regioncomposed of Two Layers by Day : SkipDistances as Absorption, not Penetration,Phenomena].-S. S.and D. M. Stuart. (Proc. Inst. Rad. Eng.,June, 1933, Vol. 21, No. 6, pp. 737-738abstract of I.R.E. Convention paper.)

Bureau of Standards results using the pulsemethod on a number of wavelengths in quicksuccession (see next abstract). " It is found thatthe E -layer critical frequency [c.f.] shows a regulardiurnal variation and a slight seasonal variation innoon maximum, indicating that the sun is the chiefsource of ionising forces in the day. The seasonalvariation of ion content is about 25% of the maxi-mum. In addition, sporadic increases of E -layerionisation to abnormally high values occur at ran-dom, and most frequently at night from someionising force apparently independent of directradiation from the sun. Comparison of terrestrialmagnetic data and electrical disturbances such asthunderstorms show no apparent relation to suchirregularities so far observed.

" The F region is found to be composed of twodefinite strata or layers in the daytime as evidencedby a second critical frequency. This c.f. denoted asFl c.f. is approximately 10oo kc higher than theE c.f., and is found to vary diurnally and seasonallyin phase with the E c.f. Heights for the Fl layercan only be estimated as 185-190 km and may besomewhat in error. The boundary between the F1(lower F) and F2 (higher F) [layers] becomesindistinct in midwinter and at night. Certainchanges near the F1 c.f. are compared with terres-trial magnetic data.

" The F2 region is found to be composed of one


or more irregular and changing strata withoutregular seasonal or diurnal characteristics. TheF2 c.f. is found to have less regular characteristicsthan the E or F1 c.f., but shows definite diurnalcharacteristics which change with season, the maxi-mum occurring near noon in the winter and aftersunset in the summer. The maximum value of thisc.f. is higher during the winter noon than during thesummer noon. This and other evidence shownindicates that the F2 c.f. does not measure maximumF2 -layer ionisa tion but is limited by total absorptionrather than penetration. . . . From the evidence it isindicated that daytime and possibly night skipdistances are absorption phenomena, as suggested byT. L. Eckersley, rather than penetration pheno-mena. . . . An estimate of daytime ionisation. . . indicates an ion content in excess of 2 X toeelectrons/cc for the F2 layer."NOTE ON A MULTIFREQUENCY AUTOMATIC RE-

CORDER OF KENNELLY-HEAVISIDE LAYERHEIGHT [and the Existence of Three Layers].-T. R. Gilliland. (Ibid., pp. 759-760.)

See preceding abstract. The transmitter andreceiver are automatically varied in frequency from2 500 to 4 400 kc/s at a uniform rate of 200 kcper minute. The virtual height is recorded by anoscillographic recorder. " In the daytime duringthe period of these tests three strata were usuallyindicated. As a rule the E layer with a virtualheight of about 12o km is found to return energyfor frequencies below 30o kc. Between 3 000and 3 600 kc reflections are likely to come fromthe F1 region with a virtual height of about200 to 240 km, while above 3 600 kc the F2 regionat 280 km and higher returns energy. Thesevalues vary considerably from day to day.

" Of particular interest is the character of thechange observed when, passing from one stratumto another as the frequency is increased. Althoughat times, when passing from E to F1, reflectionsmay drop out completely for a short interval,frequently the curve is continuous and the timeretardation will reach a high value just before theappearance of the F1 reflection. When passingfrom F1 to F2 the virtual height frequently reachesBoo or goo km. As evening approaches, reflectionsno longer come from the E layer and the longretardation between F1 and F2 becomes less pro-nounced. By sunset the curve is almost straightand there is little change of height with frequency.Later at night the highest frequencies cease to bereturned and long retardations again occur. Thephenomenon of double refraction is in evidence atthis time."A BALANCED RECEIVING CIRCUIT FOR KENNELLY-

HEAVISIDE LAYER OBSERVATIONS [Avoid-ance of Paralysis by Direct Radiation].-H. R. Mimno and P. H. Wang. (Phys.Review, ist May, 1933, Series 2, Vol. 43,No. 9, p. 774: abstract only.)

In experiments on reflection from the ionosphereat vertical incidence, the direct radiation from theemitting aerial is likely to paralyse a sensitivereceiving set. The writers have constructed theemitter and receiver as " a duplex r.f. networkwhich can respond to the reflected wave withoutany interference from the direct radiation."Possible methods are the use of the emitting -set

August, 1933

modulator to vary the receiving -set gain, or thebalancing of the r.f. voltage induced in the re-ceiving aerial by an equal and opposite voltageobtained directly from the emitter coils. Thewriters have used the second method and obtainedsuccessful results with the receiving apparatusabout four metres from the emitter ; several typesof receiving aerials have been employed directlyunder the emitting aerial.ATTENUATION OF OVER -LAND [and Over Land -

and -Water Combinations] RADIO TRANS-MISSION IN THE FREQUENCY RANGE 1.5 TO3.5 MEGACYCLES PER SECOND.-C. N.Anderson. (Proc. Inst. Rad. Eng., June,1933, Vol. 21, No. 6, pp. 758-759: abstractof I.R.E. Convention paper.)

" . . It was not, however, until quite recentlythat it was generally appreciated that the effect ofthe ground per mile varies with distance from thetransmitter, and that the effect is different for agiven amount of land when the transmission isentirely over land, or when the land is adjacentto either the radio transmitter or receiver, orintermediate (non -adjacent). . . It is the pur-pose of this paper to present various data in the1.5- to 3.5 -megacycle range which have beenaccumulated during the past few years chiefly inconnection with various site surveys, and to showthe relation of these data to a more generalisedpicture of over -land attenuation." The generalisa-tions will be given chiefly in the form of curves,which will include some for entirely over -landtransmission on other frequencies, especially inthe broadcast band. " When plotted with distanceon a logarithmic scale, the curves of apparentabsorption are, in general, quite similar for thevarious frequencies, and beyond a critical distance,which varies approximately with the square ofthe frequency, the absorption per mile variesapproximately inversely with its distance from thetransmitter."MESURES DE RAYONNEMENT ELECTROMAGNETIQUE

DES ANTENNES (Measurements of the Elec-tromagnetic Radiation from Aerials [Field -Strength Measuring Set for Motor -CarTransport : Charts of Poste-Parisien andRadio -Paris : Results over Sea, with Com-parison of Austin Formula and MadridConference Curves : P. P. Eckersley'sIndirect Night -Field Rule]).-P. David.(Rev. Gen. de l'Elec., 13th May, 1933, Vol.33, No. 19, pp. 623-63o.)

ELECTROMAGNETIC FIELD AT A DISTANCE FROM ATRANSMITTER [General Discussion : Watson,Austin -Cohen, and Eckersley Formulae :Madrid Conference Sub -Committee Re-ports].-(Journ. Telegraphique, Dec., 1932,Vol. 56, pp. 344-348). See also Abstracts,February, pp. 113-1 14 (Eckersley) andJune, p. 319, r -h column.

LA DECHARGE A HAUTE FREQUENCE DANSLES GAZ (The High -Frequency Dischargein Gases [and the Effects of a MagneticField]).-Th. V. Jonescu and Irene Mihul.(Comptes Rendus, 1st May, 1933, Vol. 196,No. 18, pp. 1292-1294.)

" Researches on the h.f. discharge in gases have

SUR

C2

August, 1933

shown that the potential necessary to sustain thedischarge passes through a minimum for frequenciesof the order of 108 c/s and pressures of the orderof io-a mm Hg. It may be smaller than theionisation potential. This fact has not so farreceived a satisfactory explanation. It has beenshown [Abstracts, February, p. 93, 1-h col.] thatthe electrons of the ionised gases have a naturalfrequency of vibration of the order of ro8 c/s.Electric fields of frequencies in this neighbourhoodcan communicate to these electrons an energygreater than that necessary to ionise the gas, andthus the discharge is sustained. Starting withthese considerations we have begun to study thedischarge in gases in a magnetic field."

Two regimes appear in these tests (where wave-lengths of 4.7-8.7 m were used) according to thestrength of the exciting h.f. field. In the firstregime (weak field) it is quite clear that the magneticfield can diminish the potential necessary for thedischarge, but the most interesting point is that asthe magnetic field is increased from zero theluminosity first increases slowly, and then decreasessharply and disappears when H = 22 gauss(A = 4.7 m) or 12 gauss (A = 8.7 m) : these are thevalues for which the free electrons come intoresonance with the external electrical field anddescribe spirals, increasing their energy at eachturn. Like the high -velocity electrons in thecathode space of a Geissler tube, they do notexcite the tube to luminescence. Even at thehighest pressures of the total range from somethousandths to some tenths of a millimetre ofmercury, the moment when the free electrons passthrough resonance can be clearly observed.

On the other hand, in the second regime (strongerh.f. fields) no such variations are found : every-thing occurs as if no free electrons were present.This regime presents some analogy to that ofmetallic conductivity.

PROPRIETES DES GAZ IONISES DANS LES CHAMPSDE HAUTE FREQUENCE (Properties of IonisedGases in High Frequency Fields). -C.Gutton. (L'Onde Elec., Feb., 1933 [publica-tion delayed], Vol. 12, No. 134, pp. 61-70.)

The object of this paper is to remark that theeffect of the passage of electromagnetic wavesthrough ionised gases can take very varied formsaccording to the boundary conditions imposedupon the gas. The writer considers the case ofplane simple harmonic waves passing through anionised gas ; the effects of collisional friction andof the external magnetic field are neglected. Hestarts from Maxwell's equations and finds theknown formula k' = k - .47rNe2lmw2 [usual nota-tion] given by Eccles for the dielectric constantwhen the medium is unlimited and there is noionisation by collision. When however the mediumis confined between two parallel planes, normal tothe direction of the electric vector and at a distancefrom one another small compared to the wave-length, the displacement of the electrons causessurface charges of opposite sign on the planes sothat there is a mechanical force on the electronsand the dielectric constant assumes the formk' = k 47rNe2[(47rNe2 - mco2). Eccles's formulais then only applicable if N is very small ; there isa resonance phenomenon for 4./rNe2 - m.02 = o.


If the medium is contained in a sphere of radiussmall compared with the wavelength, the resonancefrequency col is given by 47rNe2 - 3m.012 = o.Such resonance frequencies have been observed byH. Gutton (Abstracts, 193o, p. 267, 1-h column) .If the ionised gas is contained in a long narrowtube, a disturbance in the electron distributionwill be propagated along the tube with velocityas VNIm, where a is a function of the diameter ofthe tube. This has been found experimentallyby M. Chenot (1931, p. 261). The plasma -electronfrequency of Tonks and Langmuir (1929, pp.273, 512, 576 ; 193o, p. 175 ; 1931, pp. 374, 490 ;1932, pp. 88, 600) is also cited. The writer remarksthat in a non -uniformly ionised gas containingelectron clouds, such as the upper atmospherepresumably is, it is difficult to know what will bethe applicable form of the dielectric constant.

LA PROPAGATION DES ONDES ELECTROMAGNETIQUES-EXPOSE DES CONNAISSANCES ACQUISES ;SYNTHESE DES IDEES ET DES THEORIES(The Propagation of Electromagnetic Waves-Account of the Facts already Acquired ;Synthesis of Ideas and Theories). -P. Labbat.(Book Review by Mesny, in L'Onde Elec.,Feb., 1933 [publication delayed], Vol. 12,No. 134, p. 6A.)

PROPAGATION OF HIGH -FREQUENCY CURRENTSALONG OVERHEAD POWER LINES AFFECTEDBY SHORT-CIRCUITS OR FAULTS IN INSU-LATION. -J. Fallou. (Rev. Gen. de l'Elec.,8th April, 1933, Vol. 33, No. 14, PP. 449-450)

THE DAMPING OF SURGES ON H.T. LINES [HigherValues than those calculated from SkinEffect, and their Causes]. -E. Flegler andJ. ROhrig. (Archly f. Elektrot., 1st June,1933, Vol. 27, No. 6, pp. 413-422.)

RADIOMETER FOR HERTZIAN WAVES, AND ITSAPPLICATION TO THE STUDY OF TOTALREFLECTION. -G. A. Beauvais. (Rev. Gen.de l'Elec., 29th April, 1933, Vol. 33, No. 17,P. 553.)

Summary of a French Physical Society's paperbased on the work dealt with in Abstracts, 1932,PP. 474-475, and March, p. 150, ,1-h column.

VARIATIONS OF PHASE BY REFLECTION AT VERYTHIN METALLIC LAYERS. - P. Rouard.(Comptes Rendus, 2211d May, 1933, Vol. 196,No. 21, pp. 1592-1594.)

Continuation of the work dealt with in AprilAbstracts, p. 208, 1-h column.

ATMOSPHERICS AND ATMOSPHERICELECTRICITY

SUR LA VARIATION RAPIDE DES ATMOSPHERIQUESAU LEVER DU SOLEIL (On the Rapid Variationof Atmospherics at Sunrise [and Lugeon's" Sounding " of the Upper Atmosphere]). -R. Bureau : Lugeon. (Comptes Rendus,8th May 1933, Vol. 196, No. 19, pp. 1426-1428.)

" The rapid decrease of atmospherics observed


at sunrise on waves of the order of 3o kc/s (ro 000 m)is a consequence of the violent changes imposedby the solar radiation on the propagation of wavescoming from distant sources. I have given anexperimental proof of this by comparing the atmos-pherics curves of Paris, Tunis and Rabat, which arefrequently identical during the night hours " [Ab-stracts, 1932, p. 518, r -h column]. ApplyingLugeon's method to these curves [1931, pp. 493-494]the writer obtains 15 to 20 km as the height ofthe reflecting layer. This far too low estimate isdue, he considers, to Lugeon imagining that thesudden change occurs when the first rays of the sunreach the layer at the zenith of the receivingstation. There is no justification for this assump-tion ; the critical point is more likely to be wherereflection takes place, so that the moment at whichthe change occurs depends not only on the layerheight but also on the position of the sources andthe number of reflections between layer and earth.

An upper limit to the layer height may, however,be calculated on the assumption that the energyarrives horizontally, from directions ascertainedby d.f. observations. " A number of heightsthus calculated give reasonable results, a factwhich would indicate that the energy of the atmos-pherics is localised to the neighbourhood of thehorizontal plane." D.f. records at St. Cyr confirmhis suggestion : at the end of the night the sourcesof atmospherics are often spread out from S to W,and the disappearance at sunrise takes placesuccessively, beginning at the S and finishing,about 45 minutes later, at the W. " If the solaraction took place at the zenith, the disappearanceswould be simultaneous. When the morningdecrease occurs in successive steps, the reasonshould more often be sought in the simultaneousexistence of sources in different azimuths than inthe successive action of several ionised regions."

NOISE INTENSITY DUE TO ATMOSPHERICS, MEASUREDAT DIFFERENT PARTS OF THE BROADCASTSPECTRUM.-J. F. Byrne. (Rad. Engineering,March, 1933, Vol. 13, No. 3, p. ro.)

" The data show that the noise intensity at thelow -frequency end of the broadcast spectrum isin the neighbourhood of 4 or 5 times as greatas at the upper end of the spectrum in daylighthours, while the night-time ratio is approximately24 to I. It is interesting to note that the noiseenergy received at night is about 2 50o timesthat received on the average quiet.day. In additionlocal storms may cause an increase in noise intensityin a ratio as high as roo coo to r, in a period of afew hours." For other work by Byrne on thecomparison of the different parts of the broadcastspectrum, see April Abstracts, p. 229.

PROBABILITY IDEAS APPLIED TO RADIO COM-MUNICATION : THE QUESTION OF ATMOS-PHERICS.-Deloraine. (See reference under" Transmission.")

A RECORDING FREQUENCY COUNTER DEPENDING ONTHE TIME CONSTANT OF A CIRCUIT [Counterfor Atmospherics,Tachometer, etc.].-Lugeonand Gurtzman. (See under " SubsidiaryApparatus and Materials.")

August, 1933

ELECTRICAL DISTURBANCES OF EXTRATERRESTRIALORIGIN.-K. G. Jansky. (Proc. Inst. Rad.Eng., June, 1933, Vol. 21, No. 6, p. 758:abstract of I.R.E. Convention paper.) Onthe Group 3 type of disturbance discussedin the writer's paper dealt with in AprilAbstracts, p. 208.

VERSUCH EINER THEORIE DER BLITZSAULE (Attemptat a Theory of the Lightning Flash).-F.011endorff. (Archiv f. Elektrot., 15th March,1933, Vol. 27, No. 3, pp. 169-184.)

This paper is a preliminary attempt to develop aquantitative theory of an almost stationary light-ning column, under various simplifying assumptions.Calculations are given of the temperature of theelectrons, the current density and the current in theflash. The connection between the potentialgradient and the current in the flash gives qualitativeevidence as to the temporal course of the flash, and anexplanation is found for partial discharges.

ZUR VORWACHSGESCHWINDIGKEIT DES BLITZES (Onthe Velocity of the Extension of a Light-ning Flash).-H. Jehle. (Zeitschr. f. Physik,1933, Vol. 82, No. 11/12, pp. 785-793.)

Author's summary :-This theoretical paper givescalculations for the velocity of extension of a light-ning flash, (r) by calculating the velocity of theelectrons or positive ions at a suitable place, (2) fromelectrostatic considerations. Two simple continuityconditions for the onset of lightning are derived,from which the current and charge distribution maybe deduced and which give an explanation of balllightning.

INFLUENCE OF THE GEOLOGICAL COMPOSITION OFTHE GROUND ON THE IONISATION OF THE AIRAND ON THE INCIDENCE OF LIGHTNING.-C.Dauzere. (Rev. Gen. de 11th March,1933, Vol. 33, No. ro, pp. 297-298.) Forprevious papers see Abstracts, 1930, p. 503,r -h col., and 1931, p. 553, 1-h col.

A DESTRUCTIVE LIGHTNING FLASH [DescriptiveAccount].-C. V. Boys. (Nature, 27th May,1933, Vol. 131, pp. 765-766.)

PROTECTION AGAINST LIGHTNING BY THE " DISTANTCONDUCTOR."-B. Walter. (Zeitschr. f.tech. Phys., No. 6, Vol. 14,1933, p. 254.)

Dannmeyer and Will during the war worked outa similar method to that of the writer (MayAbstracts, p. 268, r -h col.), in special connectionwith the protection of ammunition stores in thefield.

SUR L'INTERVALLE DE TEMPS ENTRE LES PHENO..MANES SOLAIRES ET LES PERTURBATIONSMAGNETIQUES TERRESTRES (The Time In-terval between Solar Phenomena and Ter-restrial Magnetic Disturbantes).-Ch. Maur-ain . (Comptes Rendus, 24th April, 1933,Vol. 196, No. 17, pp. 1182-1186.)

A survey of the writer's own work (by thestatistical method for the years 1883-1923, and bythe individual method for 193o-1932) and that ofothers. His statistical results showed that themaximum solar activity took place on the average

August, 1933 442

days before the magnetic disturbances : for theyears of minimum solar activity the average in-terval was about 3f days. His individual resultsgave an average interval of 2k days : certainparticularly marked solar phenomena gave valuesof about 40 hours. He mentions, as a unique case,the 1872 observation of a " solar paroxysm "(near the edge of the sun) accompanied by simul-taneous violent magnetic variations.VARIATION OF THE HORIZONTAL COMPONENT

AROUND DAYS OF MAGNETIC CALM. -L. Eble.(Comptes Rendus, 8th May, 1933, Vol. 196,No. 19, pp. 1429-1431.)

Giving observations leading to the hypothesisthat the currents of the upper atmosphere, whosesudden variations are the cause of disturbances,produce with unequal velocities two secondaryeffects : a certain magnetic agitation precedingthe principal phase of the disturbance, and a rapiddiminution of the horizontal component. It maybe supposed that the first effect comes directly fromthe external currents, while the second results froman induction in the body of the earth.ELEKTRONENBOMBARDEMENT ALS FAKTOR BEI AT-

MOSPHARISCHEN ERSCHEINUNGEN (Elec-tronic Bombardment as a Factor in Atmos-pheric Phenomena [Blue Colour of Sky, Un-polarised Light from Night Sky Clouds, etc.]).-W. M. Cohn. (Physik. Ber., 15th Feb.,1933, Vol. 14, No. 4, pp. 331-332.)

POLARISATION OF THE SOLAR CORONA DURING THETOTAL ECLIPSE [Almost Constant for AllWavelengths: the Corona behaves as aGas of Free Electrons]. -J. Dufay and H.Grouiller. (Comptes Rendus, 22nd May, 1933,Vol. 196, No. 21, pp. 1575-1576.)

ELECTRICAL PROPERTIES OF THE ATMOSPHERE INTHE NEIGHBOURHOOD OF FLYING AIRCRAFT[Marked Stratification of Atmospheric SpaceCharge at and just above Cloud Level : Pos-sibility of using Electrical Phenomena as anaid to Height Estimation in Blind Flying]. -D. C. Rose. (National Research Council,Canada, Fifteenth Annual Report, 1931-1932,p. 52.)

MEASUREMENT OF THE ELECTRICAL CONDUCTIVITYOF THE AIR BY A NULL METHOD. -O. Thel-lier. (Comptes Rendus, 29th May, 1933, Vol.196, No. 22, pp. 1684-1686.)

THE INFLUENCE OF THUNDER STORMS ON RADIOPROPAGATION. -Fuchs. (See abstract under" Propagation of Waves.")

ON A PENETRATING RADIATION FROM THUNDER-CLOUDS. -B. F. J. Schonland and J. P. T.Viljoen. (Proc. Roy. Soc., 3rd May, 1933,Vol. 140, No. A 841, pp. 314-333.)

For a letter on these experiments see JanuaryAbstracts, p. 33.

FLUCTUATIONS OF COSMIC -RAY IONISATION. -J. W.Broxon, G. T. Merideth and L. Strait. (Phys.Review, 1st May, 1933, Series 2, Vol. 43, No.9, PP. 687-694.)

The writers have determined by a series ofaccurate measurements and corrections that a


decrease of 2.1% in the cosmic -ray ionisationcorresponds to an increase of 1 cm in the barometriccolumn. No correlation was established with otheratmosphericdata.

COSMIC -RAY INTENSITIES IN THE STRATOSPHERE. -I. S. Bowen and R. A. Millikan. (Phys.Review, 1st May, 1933, Series 2, Vol. 43,No. 9, pp. 695-700.)

The writers have made flights corresponding tothe barometric pressures 79 mm, 32 mm and 16 mmof mercury and obtain the same shape of ionisation -altitude curve, concave downward at the top, aswas found by Regener and by Piccard (see JanuaryAbstracts, pp. 33 and 33-34).

POSITIVE AND NEGATIVE ELECTRONS APPARENTLYPRODUCED IN PAIRS [Support to Millikan'sPhoton Theory of Cosmic Rays]. -R. M.Langer : Anderson. (Sci. News Letter, 27thMay, 1933, Vol. 23, No. 633, p. 332.)

DIE ABSORPTIONSKURVE DER ULTRASTRAHLUNGUND IHRE DEUTUNG (The Absorption Curveof Cosmic Radiation and Its Meaning). -E. Regener. (Physik. Zeitschr., 15th April,1933, Vol. 34, No. 8, pp. 306-323.)

This paper describes measurements of theintensity of cosmic radiation in Lake Constanceand in the stratosphere. The writer's final pictureof cosmic radiation is as follows :-There are fivedifferent hard components with discrete absorptioncoefficients at intervals between I and 4o. Thehard components H2 and H1 are regarded as waveradiation, coming from the dissociation of heliumand protons. The chief intensity of the soft com-ponents is in a component W2, also of undular nature,corresponding to the formation of helium. Thereseems to be no reason for regarding these or theother components as corpuscular by nature. Theymay be accompanied by corpuscular secondaryradiation. The variation with latitude (Clay, 1931Abstracts, p. 264 ; April, p. 209 ; also Compton,February, p. 94) is not regarded as a proof of thecorpuscular nature of the rays.

A POSITIVELY CHARGED COMPONENT OF COSMICRAYS. -L. Alvarez and A. H. Compton.(Phys. Review, 15th May, 5933, Series 2,Vol. 43, No. io, pp. 835-836.)

The writers find experimentally a preponderanceof rays from the West which " seems necessarilyto imply the existence of a positively chargedcomponent of the cosmic rays . . . the differencein counts in the East and West directions is of theorder of magnitude to be expected due to thedeflection of the particles by the earth's magneticfield." See also next abstract.

THE AZIMUTHAL ASYMMETRY OF THE COSMICRADIATION. -T. H. Johnson. (Phys. Re-view, 15th May , 1933, Series 2, Vol. 43, No. 10,PP. 834-835)

The results of the experiments shortly describedhere, in which the relative intensities of cosmicrays coming from the West and East directionsrespectively were measured, show that the Westintensity is greater than that of the East at anglesbetween 3o° and 65° from the zenith. The results


" accord with the I,emaitre-Vallarta theory [AprilAbstracts, p. 209, r -h col.] and show that theprincipal corpuscular component of the cosmicradiation is positively charged." See also Alvarezand Compton, above.

PROPERTIES OF CIRCUITSTRANSFORMATOREN MIT VERANDERLICHER KOP-

PLUNG ([Air -Cored] Transformers with Vari-able Coupling).-E. Siegel: (Hochf : tech. u.Elek : akus., May, 1933, Vol. 41, No. 5,pp. 167-176.)

Author's summary :-" It is shown that thebehaviour of a transformer with capacities inprimary and secondary circuits, under conditionsof varying coupling, can be represented by a circlediagram. This diagram is investigated for certainimportant special cases [transformer with zerocapacity in both circuits ; with tuned secondary ;with tuned primary and secondary ; with tunedprimary ;] and formulae are developed for the mostimportant working conditions.

Further, the behaviour of the transformer withcapacities in primary and secondary circuits, orwith a capacity in either circuit, under conditionsof varying coupling and of excitation by a varyingfrequency, is investigated, and the most importantworking conditions for these cases are set out ana-lytically and graphically." The general theoryshows that the efficiency of the transformer isindependent of k (representing the tuning of theprimary) being dependent only on p (that of thesecondary) and on the coupling factor. To obtainthe greatest efficiency the secondary tuning shouldbe complete : when this is the case the efficiencyincreases with increasing tightness of coupling.If, together with the greatest efficiency, a powerfactor cos = 1 is required, the primary tuningalso should be complete. Maximum energy trans-ference occurs with cos 44= 1 and a primary copperloss of 5o% ; but in practice this involves too highlosses.

Similar investigations of the conditions formaximum efficiency, etc., are given for the case ofexcitation by a varying frequency.

IDEAL TRANSFORMER, IDEAL SYNAPTER [dealingalso with Phase] AND PERFECT TRANSDUCER.-P. C. Vandewiele. (Rev. Gen. de l'Elec.,15th April, 1933, Vol. 33, No. 15, pp. 481-495.)

THE PRODUCTION OF NEGATIVE CONDUCTANCES BYMEANS OF RETROACTIVE COUPLINGS.-Kautter. (See under " Reception.")

THE REPRESENTATION OF [Single -Peak] RESONANCECURVES BY " SELECTIVITY INDEXES."-Kafka. (See under " Reception.")

CRITICAL REVIEWS OF BOOKS AND PAPERS ONELECTRIC FILTER CIRCUITS.-P. David :Cauer : Guillemin : Jaumann : Strutt :Beatty. (L'Onde Elec., March, 1933, Vol 12,No. 135, pp. 12-14 A.)

The writer deals first with Cauer's book(Abstracts, 1932, pp. 5o and 537, r -h columns) ;then with Guillemin's paper (Jan., p. 52, 1-h col.)which he describes as a good introduction to thestudy of Cauer's work ; and then with Jaumann's

August, 1933

paper on the multiple bridge filter (1932, pp. 595-596: see also April, p. 228, r -h col.) of which heremarks that while these filters may be considered asspecial cases of Cauer's filters, the simplicity oftheir construction and of their calculation justifiesthe direct study given by Jaumann. The last twopapers dealt with are those of Strutt (1932, p. 402)and Beatty (February, p. 98, 1-h column).

CONTRIBUTION A L'ETUDE DE LA STABILITE DESOSCILLATIONS DE COUPLAGE (Contributionto the Study of the Stability of Oscillationsin Coupled Circuits [e.g. in OscillatingCircuit coupled to Wavemeter]).-J. Mercier.(L'Onde Elec., Feb., 1933 [publication de-layed], Vol. 12, No. 134, pp. 93-112.)

A theoretical and experimental investigation ofthe conditions under which zones of discontinuity,breaking -off of oscillation, etc., are liable to occur.Applied to the case of the wavemeter, the papershows that the resonance and absorption methodsare equivalent except for certain conveniences inthe latter : that close coupling is fatal to accuratemeasurement : and that suitable conditions foraccurate measurement are indicated when it isfound that the result is not altered by a distinctincrease in coupling.

THE MECHANISM OF THE PRODUCTION OF OSCILLA-TIONS.-Le Corbeiller. (See under " Trans-mission.")

VON DER GEGENSEITIGKEIT DER WIRKUNGVON VERANDERUNGEN (SPERRWIRKUNGEN)BEI LINEAREN SYSTEMEN (Law of Reciprocityfor Changes involving the Removal of OneDegree of Freedom in Linear Systems).-A. Bloch. (Zeitschr. f. tech. Phys., No. 6,Vol. 14, 1933, pp. 234-239.)

THE RELATION BETWEEN THE DISTURBANCE CUR-RENT AND THE DEGREE OF ASYMMETRY INLINES SUBJECT TO INDUCED NOISE.-K.Schotte. (E.N.T., May, 1933, Vol. 1o,No. 5, pp. 223-225.)

TRANSMISSIONELEKTRONENSCHWINGUNGEN IM HOCHVAKUUM

(Electronic Oscillations in a High Vacuum[Electron Flow equivalent to NegativeResistance for a Certain Frequency Band]).-J. Muller. (Hochf : tech. u. Elek : akus.,May, 1933, Vol. 41, No. 5, pp. 156-167.)

A theoretical investigation of the electron flowbetween two infinite parallel plane electrodes,taking into account the effect of space charge andlimiting the treatment to those cases where at eachpoint in the space only electrons of the same speedoccur. It is shown, by considering the processesinvolved in the superposition of a small alternatingvoltage on the steady d.c. potential, that owing tothe electron inertia such an electron stream actsas a negative resistance for frequencies within acertain band, and the question is examined whetherthis negative resistance is large enough to neutralisethe damping of an oscillatory circuit.

In Fig. 10 a quasi -stationary oscillatory circuit isshown, consisting of a plane -electrode valve P andan inductance L ; F is a blocking condenser acting

SATZ

August, 1933. 444

as a short circuit for the r.f. oscillations and allowingthe supply of anode potential through the chokecoils D. The equivalent circuit is shown in Fig. I I.The conditions for the setting up of oscillations arefound : the frequency is determined by the externalcircuit, the only effect of the electrons being thesmall change which their presence causes in thenormal capacity of the valve. If C and R are theequivalent capacity and negative resistance of thevalve, W the ohmic resistance of the externalcircuit, and n the circuit frequency, the conditionfor the setting up of oscillations is W/,L < I tan p I .

The maximum value for tan p lies in the space -charge region when the product of the path time rand the circuit frequency ti is vr -= 2.377r ; itsvalue then is 0.074. In the saturation region itsvalue is smaller.

The paper ends with an example of a circuitsuitable for generating oscillations in this manner.Here two anodes, one on either side of the cathode,are connected by a circular loop : the d.c. supplygoes to the cathode and to the middle point of thisloop, so that the two anodes are at the same potentialand the cathode lies at a potential node of the r.f.circuit. Numerical values are found for thevarious components in order to generate a r -metrewave, and it is shown that the loss angle of theoscillatory circuit is to times smaller than themaximum possible value of tan p (see above). Toobtain the optimum condition nr = 2.37 it for thisi-metre wave, 7 must be 3.95 x io-a sec., and toobtain this an anode potential of 164 v is required.From energy considerations it is found that for theproduction of a given frequency the greatest amountof energy is obtained by the use of high currentdensities, while if the current density is limited thegreatest energy is given by the use of longer waves.

ELECTRON CONDUCTION IN THERMIONIC VALVES[and the Generation of Ultra -High Fre-quencies in Diodes].-W. E. Benham :Potapenko. (Elec. Communication, April,1933, Vol. I1, No. 4, pp. 223-225.)

" In a recent paper [1931 Abstracts, p. 212] thedielectric constant and conductance of an electronatmosphere between the parallel electrodes of aLangmuir diode were evaluated as functions ofpT Fig. I shows that the dielectricconstant e has a value less than unity for fre-quencies lying between o and 9/271T. Taking zo-gsec. as a typical value for T, we see at once that thefrequency range for which e < 1 embraces nearlyall the frequencies so far obtained using thermionicvalves (an angular frequency of nearly 1010 may beobtained under favourable conditions). Fig. r mayalso be read as the value for e for different valuesof T, p being held constant. Read in this way itwill be seen that as pT increases from zero up toabout z or 2, the value of E remains sensibly con-stant at the value o.6, which means that in thiscase the value of E is constant over the spaceincluded between the cathode (T = o) to the anode(T = Td). This surprising result that (for a rangeof frequencies of practical interest) the dielectricconstant of all points in the inter -electrode spaceis 3/5 must constitute the first known case of a' dielectric' which is distinctly non -homogeneousin constitution and yet homogeneous in respect ofdielectric constant. Although on the above con -


siderations e is constant, the displacement currentnevertheless varies from point to point. The spacevariation of displacement current and of potentialcombine to effect complete neutralisation of thespace variation of dielectric constant in a Langmuirdiode." The writer's tests and the work ofBergmann and During (Abstracts, 1929, p. 440) arementioned.

By interpreting Fig. 2 (the relative change inconductance for values of pT between o and 22) inthe same way, it is seen that the conductivity forexceedingly high frequencies is alternately positiveand negative at a number of regions between theplates. If the anode lies in any one of these regions,the diode as a whole exhibits negative conductance :or for any given anode voltage there are ranges offrequency for which this occurs. Potapenko'sexperimental " normal " and " dwarf " waves(1932, p. 582, 1-h col.) are discussed in this con-nection, and the fair constancy of the last columnof Table I shows that these waves have frequencieswhose relation to one another is not very differentfrom the ratios determined by taking the minimaof Fig. 2. " A significant feature of the theory, incontrast to previous conceptions, is that thegeneration of oscillations in an external circuit cantake place with amplitudes of electron motion sosmall that if the motion in transit of the electronsunder the influence of steady potentials could befollowed with the eye, the superimposed oscillatorymotion would escape notice altogether."

RESONANCE IN THREE -ELECTRODE VALVES.-E. W. B. Gill and R. H. Donaldson. (Phil.Meg., June, 1933, Series 7, Vol. 15, No. 102,pp. 1177-1181.)

The writers have studied a three -electrode valvein which the grid was kept at a positive potential Vand the anode at a slightly negative potential v withrespect to the filament. A constant alternatinge.m.f., whose periodic time was the same as that ofthe free oscillations of the electrons in the valve,was superposed on the space between filament andanode ; a resonance curve was found for observa-tions of the direct current reaching the anode whenthe grid potential was varied ; a large valve wasused and, with impressed e.m.f. of wavelengthto.or metres, resonance was obtained for a gridvoltage V of 24 volts. This was independent offilament emission : AW was also found to beconstant, within experimental error, A being thewavelength of the oscillation at which resonanceoccurs. Since resonance is practically independentof filament emission, the experimental results arenot in agreement with recent theories of Bark-hausen-Kurz oscillation based on the considerationof the normal modes of free oscillation of the systemof space - charges in the valve (cf. Rostagni,Abstracts, 1932, p. 282 ; also January, p. 37, 1-hcolumn).

THE USE OF MAGNETIC FIELDS FOR THE PRODUC-TION OF ULTRA -SHORT WAVES. MAGNETRONTRANSMITTRS.-M. Ponte. (Journ. de Phys.et le Red., No. 12, Vol. 3, 1932, pp. 183-184.)See also July Abstracts, p. 390, r -h col.,Soc. Franc. Rad : elec. ; also Rev. Gen. del'Elec., 4th March, 1933, Vol. 33, No. 9,p. 278.


THE USE OF INSULATED REFLECTION PLATES,SLIDING ON PARALLEL -WIRE SUPPLY LEADS,IN PLACE OF CHOKING COILS IN ULTRA -SHORT -WAVE WORK.-(Hochf:tech. U.Elek:akus., March, 1933, Vol. 41, No. 3,p. : German Pat. No. 54o 338, pub.14th Dec., 1932, Kohl.)

FREQUENCY MULTIPLICATION IN ULTRA -SHORTWAVES BY SUPERPOSITION ON RETARD-ING FIELD POTENTIAL. (Hochf:iech. U.Elek:akus., May, 1933, Vol. 41, No. 5,p. 188: German. Pat. No. 568 768, pub.31st Jan., 1933, Gerhard.)

" Taking advantage of the relation betweenretarding -electrode current (and r.f. power) andretarding -electrode potential, shown in Fig. 5, afrequency multiplication can be obtained bysuperposing the a.c. potential to be multiplied onthe retarding -electrode potential in such a way thatboth the rising and falling parts of the character-istic are worked on. The frequency -multipliedpotential is taken from the grid circuit."FREQUENCY MULTIPLICATION BY SUPERPOSITION OF

Two OSCILLATIONS EACH MODULATED ATTHE OSCILLATING FREQUENCY, ONE INPHASE AND THE OTHER IN OPPOSED PHASE.(Hochf:tech. u. Elek:akus., Feb., 1933,Vol. 41, No. 2, p. 71 : German Pat. No.555 952, pub. 17th Aug., 1932, Wassermann.)

INDUCTIVE COUPLING ARRANGEMENT FOR SEVERALULTRA -SHORT-WAVE VALVES. (Hock...tech.U. Elek:akus., Feb., 1933, Vol. 41, No. 2,p. 71 : German Pat. No. 56o 129, pub.28th Sept., 1932, Gerhard.)

See Fig. 4. Each anode has a radiator attachedwhich is coupled to a half -wavelength portion of azig-zag (Greek pattern) feeder.

GRADUATING TO OSCILLATOR -AMPLIFIER TRANS-MITTERS ON 56 Mc [and the Prevention ofFrequency Modulation and Drift in theUltra -Short -Wave Band]. -D. A. Griffin.(QST, May, 1933, Vol. 17, No. 5, pp. 21-22and 68.)

L'OBTENTION DES OSCILLATIONS POLYPHASEES AL'AIDE DES SYSTEMES DYNATRONIQUES (TheGeneration of Polyphase Oscillations withthe aid of Dynatron Circuits). -J. Grosz-kowski. (L'Onde Elec., Feb., 1933 [pub-lication delayed], Vol. 12, No. 134, pp. 85-91.)

" The polyphase electrical oscillations firstobtained by Mesny with the aid of triodes withreaction, and later treated theoretically by Arenberg[193x Abstracts, pp. 438 and 497-498], can also beproduced in dynatron systems. The latter, thanksto the absence of reaction, are simpler and also moreinteresting owing to the possibility of obtainingtwo modes of functioning. . . . The negativeresistances of the dynatrons may be arranged in' star ' or in mesh ' connection ; moreover, themutual combination of the system of circuits withthe dynatron system may be star and star, meshand mesh, or star and mesh. . . . In view of theease of energy supply from a common source, andalso of the simplicity of construction of the circuitelements, the star arrangements of circuits and ofdynatrons are more convenient and more interest -

August, '933

ing. The mesh arrangements may in fact bereduced to a certain number of single-phase circuitsjoined among themselves at certain peaks ; whereasin the star arrangements the circuits are mutuallycoupled, a fact which exerts an effect on the fre-quency of the system and on its condition for themaintenance of oscillation."

The writer therefore confines himself to dealingwith the various types of star connection -inductances in star connection, capacities in starconnection, and the case (usual in practice) wherethe inductances in star connection have self -capacities. The frequency equations and thelimiting conditions are derived. He ends by givingsome experimental results, using Philips triodesworking as dynatrons. The wavelength was of theorder of 2 000 metres.LE MECANISME DE LA PRODUCTIONS DES OSCILLA-

TIONS (The Mechanism of the Production ofOscillations). -Ph. Le Corbeiller. (L'OndeElec., March, 1933, Vol. 12, No. 135,pp. 116-148.)

Author's summary :-" Two general types ofself -sustained oscillating systems are first studied.The first type consists of an irreversible relay whoseoutput is coupled to the input through a passivequadripole. The second type is represented by aconductor whose characteristic presents a descendingarc, connected to the supply source and to a passivebipole. Each of these two types can be set up intwo different manners, corresponding to each other' by duality' [i.e. according to the ' principle ofduality' based on Maxwell's relation betweenelectrostatic and electromagnetic energy : thus theseries connection of L, C and R of the dipole inFig. 9B, p. 132, corresponds by duality' with theparallel connection of the dipole in Fig. 9A, and thecurrents of the one are given by the same differentialsystem as the potentials of the other]. A non-lineardifferential equation is derived corresponding to avariable electrical quantity of each of the twotypes. The differential equation of the second typeis a particular case of that of the first type.

" The paper then deals with the simplestsymmetrical oscillations obtained from electricalor mechanical self -sustained oscillators : the' series ' oscillations are given by the periodicsolution of the equation z" - (Z' z'3/3) z = o,and the' shunt' oscillations by the periodic solutionof the equation y - e(I - y2)y' y = o. Theyhave opposed properties and correspond to eachother by duality. " The work of Lienard onclosed integral curves, or " limiting cycles," ofwhich use is made in the paper, is referred to in1928 Abstracts, p. 469, r -h column.MAGNETOELECTRIC RADIATIONS FROM RING -SHAPED

OR SPIRAL OSCILLATORS. -K. F. Lind/Tian.(Physik. Ber., 15th April, 1933, Vol. 14,No. 8, p. 600.)

The writer gives the name " magnetoelectric " tothe radiations produced in such oscillators by thelines of magnetic induction whose changes generatelines of electrical induction at right angles. Suchradiations are less intense than those from acorresponding rod insulator, but can be detectedat distances of some wavelengths. If, in a closediron ring, sufficiently strong magnetic currents aregenerated by the rapid variation of a magnetic

August, 1933 446

induction in the ring, Maxwell showed that anelectrical field of the second type must be formed.The writer believes that he has demonstrated theexistence of such a field in the space enclosed in alarge iron ring, by the use of a sensitive thermo-electric detector. When the iron ring was replacedby one of copper, the effect was absent.GRAPHISCHE BEHANDLUNG VON MODULATIONS-

PROBLEMEN (The Graphical Treatment ofModulation Problems [including Applica-tions to the Transmission of Carrier and Side -Bands from Separate Aerials, and toCommon -Wave Broadcasting]). -H. Roder.(E.N.T., May, 1933, Vol. to, No. 5, pp. 225-229.)

For papers in English on the author's graphicalmethod see April Abstracts, pp. 210 (r -h col.) and212-213. The present article deals separately withthe two applications mentioned. Three alternativeways of transmitting the carrier and side -bandsfrom separate aerials are shown in Fig. 4 ; of these,only the right-hand arrangement, in which thecarrier is transmitted from the central aerial andone side -band from each of the outer aerials, isfree both from linear and non-linear distortion,although its practical carrying out involves somedifficulties.TIME DELAY EFFECTS IN SYNCHRONOUS [C011111:1011 -

Wave] BROADCASTING. -Aiken. (See under" Stations, Design and Operation.")

SINGLE SIDE -BAND TELEPHONY ON SHORT WAVES.-E. M. Deloraine : Reeves. (Rev. Gen. del'Elec., 15th April, 1933, Vol. 33, No. 15,pp. 476-477 : summary only.)

ARRANGEMENT FOR SINGLE SIDE -BAND MODULA,TION.-(Hochf:tech. u. Elek:akus., March,1933, Vol. 41, No. 3, p. I io : German Pat.No. 56o 226, pub. 29th Sept., 1932, vonPlebanski.)

THE BALANCING AND STABILISING OF HIGH -FREQUENCY AMPLIFIERS, WITH SPECIALREFERENCE TO POWER AMPLIFIERS FORRADIO TRANSMITTERS [particularly for NavalShip Requirements]. -W. Ure, E. J. Grainger,and H. R. Cantelo. (Wireless Engineer,May, 1933, Vol. to, No. 116, pp. 259-261.)

Long abstract of an I.E.E. paper and Discussion.Among the points discussed are :-the fact that avalve with its associated circuits may be neutralisedfor the frequency at which it is designed to workis no guarantee that it cannot oscillate at a differentfrequency (even the push-pull circuit is capableof generating spurious oscillations) : " hot " and" cold " balance methods : naval ship require-ments : design of coils for h.f. master -oscillator,to give constant inductance over temperaturechanges of about 40° C.ELIMINATION OF FADING COMPENSATING DEVICES

BY CHOICE OF MODULATION CURVE OFTRANSMITTER, AND RECTIFYING CURVE OFRECEIVER, AS IDENTICAL EXPONENTIALFUNCTIONS.-(Hochttech. u. Elek:akus., Feb.,1933, Vol. 41, No. 2, p. 71 : German Pat.No. 56o 227, pub. 29th Sept., 1932,Roosenstein.)

By this arrangement " the ratio of the maximum


to the minimum amplitude of the rectified currentis determined only by the degree of modulation,assumed to be constant, and is independent offading."

GETTING QUALITY PERFORMANCE WITH CLASS BMODULATION : PRACTICAL DESIGN ANDOPERATING DATA FOR BEST TUBE COMBINA-TIONS. -A. A. Collins. (QST, May, 1933,Vol. 17, No. 5, pp. 12-17 and 57.)

A SENSITIVE TUNING INDICATOR [Dry -PlateRectifier Disc and Milliammeter shuntedby Condenser]. -J. D. Blitch. (QST, May,1933, Vol. 17, No. 5, p. 2o.)

RECEPTIONEMPFANGSVERSUCHE MIT DEM KRISTALLDETEKTOR

IN CHINA (Reception in China on a CrystalDetector [Nanking at 500 km by day :Manila at 2 50o km by night]). -F. Oster.(Radio, B., F. fur Alle, May, 1933, p. 212.)

CUTTING THE COST OF SINGLE -SIGNAL RECEPTION:CONVERTING THE T. R. F. REGENERATIVERECEIVER TO AN S.S. SUPERHET.-J. J.Lamb. (QST, April, 1933, Vol. 17, No. 4,pp. 8-12 and 58-66.)

A short summary of what is implied by " single -signal selectivity " is given on the last two pages.For previous papers see June Abstracts, p. 327,1-h column.

SECOND CHANNEL SUPPRESSION. -R. I. Kinross.(Wireless World, 9th June, 1933, Vol. 32,PP. 416-417.)

There is a tendency at the present time to cutdown the number of signal -frequency circuits in asuperheterodyne, but unless special precautionsare taken, second channel interference is likelyto prove troublesome. The author describes anew input filter, combining the virtues of inductiveand capacitive coupling, which satisfactorilyminimises the effect of " image " frequencies.

THE INTRODUCTION, INTO AN EXISTING RECEIVER,OF EXPONENTIAL VALVES IN THE PLACE OFORDINARY SCREEN -GRID VALVES. -(Radio,B., F. fur Alle, May, 1933, pp. 215-219.)

THE TWO -UNIT PORTABLE. -H. F. Smith. (WirelessWorld, 23rd June, 1933, Vol. 32, pp. 440-443.)

A four -valve portable receiver for the amateurconstructor, designed, for easy transport, in theform of two units which can be quickly connectedtogether.

CHECKING THE PERFORMANCE OF THE SUPER-HETERODYNE FIRST DETECTOR [and theChoice of Valve Types]. -A. L. Chaney.(QST, May, 1933, Vol. 17, No. 5, pp. 34-35and 7o, 72.)

A SUPERHETERODYNE RECEIVER USING THE NEW" FADING HEXODE " AND " MIXING HEX -ODE " VALVES.-Wigand : TelefunkenCompany. (See abstract under " Valvesand Thermionics.")


ULTRA -SHORT-WAVE Two. -H. B. Dent. (WirelessWorld, 16th June, 1933, Vol. 32, pp. 422-424)

A two -valve ultra -short-wave receiver for theamateur constructor. The waveband covered isfrom 5 to 7 metres. The super -regenerativeprinciple is employed.

IMPROVING THE 56 -MC RECEIVER : CONSTRUCTIONALDETAILS OF TWO NEW SETS FOR THE ULTRA-HIGH FREQUENCIES. -C. F. Hadlock. (QST ,May, 1933, Vol. 17, No. 5, pp. 23-26.)

AN INTER -CARRIER NOISE SUPPRESSION SYSTEM[using the Wunderlich Valve of a New Type,with Small Shielded Extra Anode]. -N. E.Wunderlich. (Rad. Engineering, March,1933, Vol. 13, No. 3, pp. 7-9 and 13.)

" A simplified and improved method for silenttuning. An exceptional 7 -tube superhet withamplified AVC and noise suppression." See alsoa short article on p. 13 on the use of a standardWunderlich " A " valve, in the absence of the new" B " type.RADIO RECEIVER SITUATION FOR 1933 [including

" Colour " Visual Indication of Noise,Tone and Volume Control].-(Rad. Engineer-ing, May, 1933, Vol. 13, No. 5, pp. 12-13.)

CONTROLS FOR 1933 RECEIVERS [particularlyMoulded Carbon Variable Resistances andTheir Uses]. -N. E. Wunderlich. (Rad.Engineering, May, 1933, Vol. 13, No. 5,pp. 17-19.)

AUTOMATIC VOLUME CONTROL FOR RADIO RE-CEIVERS [Survey of American Developments].-C. B. Fisher. (Wireless Engineer, May,1933, Vol. Io, No. 116, pp. 248-254.)

CORRECTED AVC.-Cossor Laboratories. (Wire-less World, and June, 1933, Vol. 32, pp.386-388.)

Although highly satisfactory results can beobtained by ordinary AVC systems in which ther.f. valves are controlled, theoretical perfection isimpossible, for the controlling bias is dependent onthe detector input. In the recently developeddouble -diode pentode valve, which has variable -mu characteristics, the AVC is taken both forwardsand backwards to the a.f. and r.f. stages, respec-tively. As a result it is possible to obtain almostperfect AVC, since both pre- and post -detectorvolume control is obtained.

BACK TO QUALITY IN RADIO RECEIVERS ! [WithCurves showing Wide Difference in ToneFidelity between Large and Small Re-ceivers]. -(Electronics, April, 1933, Vol. 6,No. 4, pp. 106-107.)

RADIO AT THE LEIPZIG SPRING FAIR. -(Radio,B., F. fur Alle, May, 1933, pp. 230-234.)

THE USE OF INSTRUMENTS IN RADIO RECEIVERMANUFACTURING. -J. H. Miller. (Rad.Engineering, May, 1933, Vol. 13, No. 5,pp. 20-21.)

August, 1933

RADIO RECEIVER DESIGN [The Equivalent CircuitsUseful in Broadcast Receiver Design]. -J. E. Smith. (Rad. Engineering, March,1933, Vol. 13, No. 3, pp. 20-22.)

DIE HERSTELLUNG VON NEGATIVEN LEITWERTENMIT HILFE VON R6CKKOPPLUNGSSCHAL-TUNGEN. I. (The Production of NegativeConductances by means of RetroactiveCouplings [and the Calculation of Retro-active Receiver Circuits]). -W. Kautter.(E.N.T., May, 1933, Vol. io, No. 5, pp.199-214.)

" In a previous paper [1931 Abstracts, pp.495-496] it was shown that, by the use of thecurrent -source equivalent diagram of the aerial,the complete input circuit of an ordinary retro-action -less receiver can be represented by a parallelconnection of positive and negative susceptances[imaginary components of admittances] and posi-tive conductances, fed by the aerial with a certainpractically constant current. The potential changesat the first grid due to adjustment of the tuningcan be represented by the corresponding changesin the resultant admittance." The calculation ofthe selectivity and potential transformation (ratioof grid a.c. and aerial potentials) is unaffected,owing to the small frequency range involved, bythe frequency dependence of the input current.

The object of the present paper is to extend themethod to the case of a receiver with retroaction,by the introduction of a " negative conductance,"which fortunately is easy to calculate and haspractically no dependence on frequency. Theretroaction is accompanied also by an unwanted,but unimportant, wattless component which showsitself in a slight de -tuning of the grid circuit. Thevariation with amplitude of the retroaction -equiva-lent " negative conductance " has a definite con-nection with the characteristic curves of the valves.After applying his method to capacitive andinductive retroactive couplings, the writer con-siders certain special circuits, such as retroactionmethods which do not cause de -tuning, and retro-action over several stages.INTERFERENCE WITH BROADCAST RECEPTION. -

(Rev. Gen. de l'Elec., 6th May, 1933, Vol. 33,No. 18, p. 143 D : summary of NationalElectric Light Association Publication No.33, Jan., 1933, pp. 1-3.)

" This report has been issued by the Joint Com-mittee created by the National Electric LightAssociation, the National Electrical IVIanufacturersAssociation and the Radio Manufacturers Associa-tion. . . . The recommendations given in the con-clusions of this report are as follows : increase of thepower of the transmitting stations ; the study ofreceivers in order to make them as insensitive aspossible to parasitic fields without sacrificing theirother qualities ; installation of receivers so that thecoupling between aerial and the neighbouring circuitsliable to create parasitic fields may be very loose ;the fitting, in special cases, of filters to apparatusproducing well recognised interference."INTERFERENCE WITH BROADCAST RECEPTION.-

G.W.O.H. (Wireless Engineer, May, 1933,Vol. io, No. 116, pp. 237-238.)

Editorial on the need for legislation ensuring that

August, 1933 448

electrical apparatus should not be offered for sale ifit cannot be operated without causing interference,and that the operation of apparatus in such a wayas to produce interference should be an offence.A recent law promulgated in Roumania is discussed.The future need for the compulsory suppression ofmotor -car ignition interference, for the sake ofultra -short-wave services, is pointed out.

THE INSTALLATION OF AERIALS WITH SCREENEDDOWN-LEADS.-Wigand. (See under" Aerials and Aerial Systems.")

ANTENNA TRANSMISSION LINE SYSTEMS FOR RADIORECEPTION [Elimination of Man -Made Staticby Suitable Location of Aerial and the Useof a Shielded Lead -In]. -Brigham. (Ibid.)

NOISE INTENSITY IN RADIO RECEPTION. -Byrne.(See abstract under " Atmospherics andAtmospheric Electricity.")

DECOUPLING AND INSTABILITY [Production ofMotor -Boating by Incorrect Design ofDecoupling Circuits]. -S. J. Preston. (WorldRadio, 9th June 1933, Vol. 16, No. 411, p.772.)

WHY DOES THE LOW -FREQUENCY AMPLIFIERHOWL ? TESTS ON THE STATIC SCREENINGOF A.F. TRANSFORMERS. -H. Piesch. (Radio,B., F. fur Alle, March, 1933, pp. 97-101.)

THE PROBLEM OF STABILITY [111 Radio Receivers]. -L. Chretien. (Rev. Gin. de l'Elec., 29thApril, 1933, Vol. 33, No. 17, p. 132 D :summary only.)

TUNING DISCONTINUITIES OF A ROTATING VARIO-METER SMOOTHED OUT BY COMBINATION WITHA ROTATING CONDENSER WITH WAVE -FORMEDGES. - (Hochf:tech. u. Elek:akus., Feb.,1933, Vol. 4,, No. 2, p. 72 : German Pat.No. 559 542, Telefunken.)

SUPPRESSION OF MAINS BUM : THE USE OF ANAUXILIARY VALVE IN OPPOSITION TO OUTPUTVALVE : OF A BRIDGE CIRCUIT : OF RECTI-FIERS IN OPPOSITION ACROSS HEATINGCIRCUIT.-(Hochf:tech. u. Elek:akus., March,1933, Vol. 41, No. 3, pp. III -112: GermanPat. Nos. 562 640, 562 305 and 562 304,pub. 27th, 24th and 24th Oct., 1932, Siemens-Schukert, Telephonfabrik Berliner, R.C.A.)

IMPROVED SMOOTHING IN MAINS RECEIVERS BYCONNECTING VALVES AS SERIES ELEMENTSOF ANODE -CURRENT FILTER CHAIN.-(Hochf:tech. U. Elek:akus., Feb., 1933, Vol. 41, No.2, p. 72 : German Pat. No. 559 640, pub.22nd Sept., 1932, Siiddeutsche Tel. Kab. u.Drahtwerke.)

ETUDE DU CIRCUIT D'ENTREE D'UN POSTE RE-CEPTEUR RADIOPHONIQUE (Study of theInput Circuit of a Radiotelephonic Re-ceiver [and the Predominant Importance ofthe H.F. Losses]).-Mezey. (L'Onde Elec.,March, 1933, Vol. 12, No. 135, pp. 149-16o.)

A large number of receivers, particularly Ameri-


can, use a pre -selector circuit consisting of twoinductively coupled tuned circuits, the two variablecondensers being ganged. The coupling of the firstcircuit to the aerial circuit is sometimes inductive,sometimes capacitive, and is loose in either case.The writer deals with the inductively coupled type,and examines the forms of the resonance curves ofsuch a band filter with a view to showing how theyvary with the carrier wavelength and how tocalculate the circuit constants to give the bestresults over a range such as 200-60o m. The first partof the paper gives the solution of the general equa-tions in a simplified form ; the second deals with a.numerical example and shows the importance of ther.f. losses. The discussion of these losses is basedon the writer's tests on duplicates of Americancoils : one of these had a resistance of ro ohms at400 m, which he considers too high to give any goodresult in a pre -selector circuit. The resistance waschiefly due to skin effect, and by using a wire of 3oinsulated strands he reduced it to 5.8 ohms. Heconcludes that though many makers have abandonedthe inductive coupling between the two circuits, inorder to render the filter independent of frequency,and have adopted resistance and capacity couplings,the r.f. losses and theft' variation with frequencyare of such predominant importance that improve-ments in this direction will furnish the true solution.

THE USE OF MULTI -WAVE CIRCUITS IN GRID ANDANODE CIRCUITS OF A RECEIVER WITH RETRO-ACTION, TO GIVE RECTANGULAR RESONANCECURVE.-(Hochf:tech. u. Elek:akus., Feb.,1933, Vol. 41, No. 2, p. 72 : German Pat. No.558 300, pub. 6th Sept., 1932, Lorenz.)

DIE KENNZEICHNUNG VON RESONANZKURVEN DURCHSELEKTIVITATSZAHLEN (The Representationof [Single -Peak] Resonance Curves by" Selectivity Indexes "). - H. Kafka.(Hochf:tech. u. Elek:akus., May, 1933, Vol. 41,No. 5, pp. 176-184.)

Taking first the case of a series connection ofresistance r, inductance L, and capacity C, of whichL is the variable component, and making use of aresonance curve in the form of a special circle diagram(Fig. I), the writer examines how the value of tanchanges as the resonance point is passed through.Here represents the phase position of the vectorOP : the relation of the special circle diagram to theordinary resonance curve is shown by the fact thatthe latter is obtained if the length OP is plotted asordinate along a horizontal L scale. The resonanceis the sharper the larger 9S is for a given change± da with respect to the resonance value A =(here A is the ratio L/Lr of a value of the inductanceL to its resonance value The relation betweentan ¢ and Xis tan # = (i - A), which canbe represented by a straight line whose slope is ameasure of the size of ck corresponding to a givenchange ± dA with respect to A= I.

On this basis, therefore, the writer suggests thata numerical measure of the selectivity with varyinginductance is conveniently given by the absolutevalue SL of the ordinate to the tan # straight line,corresponding to a change d A = I% withrespect to the resonance value A = I. Thusthe formula for this " selectivity index " is

0.01 lb tan OMAI= o.or caL,.1r.


Dealing next with the more practically importantcase of constant r and L and variable C, the writershows that here the corresponding connectionbetween tan .rk and y (y = C/C,) is not a straightline but a curve. This non-linear variation of tanhas the result that the resonance curve for variableC is not strictly symmetrical as in the case ofvariable L. But to obtain a simple measure ofselectivity the curve may be replaced, in theneighbourhood of its resonance point, by the tangentat that point. Then the selectivity index S. isdefined as the absolute value of the ordinatecorresponding to a change Lly = % drawn tothe tangent to the tan ¢ curve at its resonancepoint y = 1. In this case S, = 0.01 cuL/r.

The very important case is then treated wherer, L and C are all constant and the frequency fvaries (r is taken as constant near a given resonancefrequency, in spite of its general variation withhigh frequencies). Here again the connectionbetween tan 9i. and n (77 = flf,) is representedby a curve which must be replaced by its resonance -point tangent, giving a selectivity index St= 0.01 . 20.41.fr. Thus the selectivity index forvariable frequency is twice as great as that forvariable inductance or capacity. This point isdiscussed on p. 179, r -h column.

The writer applies his methods to a number ofpractical circuits, including frame aerials, a rejectorcircuit with series input resistance, and a series oftwo high -frequency stages with rejector -circuitcoupling. He ends with a discussion of theselectivity relations in multi -circuit receivers forthe broadcast range of wavelengths, where mattersare complicated by the large variations withfrequency of the r.f. resistance in the various coils.The case of resonance curves with two maxima willbe dealt with in a later paper.

AERIALS AND AERIAL SYSTEMSTHE CONCENTRATOR ANTENNA [giving Decreased

Space Wave and Increased Efficiency atthe Shorter Broadcast Wavelengths]. -Westinghouse Company. (Rad. Engineering,March, 1933, Vol. 13, No. 3, p. o.)

The " exciter " is a copper rod 204 feet high ;the " concentrator " is another copper rod about150 feet high, about 25o feet from the " exciter "and adjusted to resonate with KYW's frequencyof r 020 kc/s. Under each rod is a buried earthof copper sheet with eight radial strips. " Theconcentrator tends to bend down the radio wavescoming from the exciter, flattening them so thatthey are intensified over the useful area of thestation."

FADING ELIMINATED BY HIGH AERIAL SUPPORTEDBY BALLOON [" Blimp Antenna"] . (Electronics,April, 1933, Vol. 6, No. 4, p. 87 : News item.)

" To eliminate troublesome fading, KDKAengineers obtain baby blimp to hold antenna aloftto a vertical height of 1500 ft." See also Rad.Engineering, April, 1933, Vol. 13, No. 4, p. 14,where it is stated that these experiments " mayhave considerable effect upon the future efficiencyof broadcasting stations . . . An experimentalhalf -wave antenna, about 50o ft. long, now trails

500 ft. in the air. A light aluminium wire,

August, 1933

serving as a combination guy and feed wire, servesthe dual purpose of restraining the ' captive aerial 'from free flight and ' feeds ' it radio programmes."

INVESTIGATING THE DIRECTIVE PROPERTIES OF ANAMATEUR ANTENNA [Long L -Aerial for ShortWaves]. -S. L. Seaton. (Physik. Ber.,April, 1933, Vol. 14, No. 7, pp. 514-515.)

Working on a frequency of 14 285 kc/s, thishorizontal wire aerial had a length of 3 A and aheight of A/2. It showed a strong and sharplydefined maximum in the usual direction, with4 weak subsidiary maxima.

HORIZONTAL DIRECTIVE SHORT-WAVE ANTENNA.-(Hochf : tech. u. Elek akus., March, 1933,Vol. 41, No. 3, p. 112: German Pat. No.562 306, pub. 24th Oct., 1932, U.S.S.R.Elektrot. Vereinigung.)

" A symmetrical dipole with two radiators of0.6-0.65 A gives in the main direction twice asmuch energy as a half -wave radiator."

J ONCTION ENTRE DEUX LIGNES D' IMPEDANCESCARACTERISTIQUES DIFFERENTES (The Join-ing of Two Lines of Different CharacteristicImpedance [and the Calculation of theDissymmetric Quadripole for Matching aTwo -Wire and a Concentric Line]). -J. C.Pomey. (Rev. Gen. de l'Elec., 18th March,1933, Vol. 33, No. II, pp. 331-335.)

REFLECTORS FOR ULTRA -SHORT WAVES.-(Hochf :tech. u. Elek : akus., May, 1933, Vol. 41,No. 5, p. 188: German Pat. No. 568 015, pub.12th Jan., 1933, Telefunken Company.) .

See Fig. 4. The radiator o at the focus of aparabolic mirror has its direct radiation screenedby the small spherical concave mirror b. ACcord-ing to the invention, a portion of the para-bolic mirror around its vertex, about equal inaperture to the concave mirror, is removed andreplaced by a spherical surface either inside orat the back of the parabolic mirror, in order toprevent the emission, in an incorrect direction, ofrays reflected back from b on to that part of theparabolic mirror.

ANTENNA TRANSMISSION LINE SYSTEMS FOR RADIORECEPTION [Elimination of Man -MadeStatic by Suitable Location of Aerial andthe Use of a Shielded Lead -In]. -C. E.Brigham. (Rad. Engineering, March andApril, 1933, Vol. 13, Nos. 3 and 4, pp. 11-12and 21-22.)

The Radio Club paper referred to in JuneAbstracts, p. 325, r -h column.

DER BAU VON ANTENNEN MIT ABGESCHIRMTERZULEITUNG (The Installation of Aerialswith Screened Down -Leads). -R. Wigand.(Radio, B., F. fur Alle, May, 1933, pp. 193-198.)

Based on the use of the shielded " Kapa " cable,made by the Vacha Cable Works, with air insula-tion. The writer's remark that a lead of 15 mlength would have a capacity of only about 400 cmscalls for the editorial comment that the " only "seems rather out of place, since such a capacityis large for modern receivers.

August, 1933 450

VALVES AND THERMIONICSVACUUM TUBES FOR USE AT EXTREMELY HIGH

FREQUENCIES [Very Small Triodes and S.G.Valves for Reception of Ultra -Short Wavesdown to 6o Centimetres with ConventionaLCircuits, or for Conventional OscillatorCircuit for 3o Centimetre Waves.]. -B. J.Thompson and G. M. Rose, Jr. (Proc. Inst.Rad. Eng., June, 1933, Vol. 21, No. 6,pp. 755-756: abstract of I.R.E. Conventionpaper.)

VACUUM TUBE CHARACTERISTICS IN THE POSITIVEGRID REGION BY AN OSCILLOGRAPHICMETHOD. -H. N. Kozanowski and I. E.Mouromtseff. (Ibid., p. 756.)

VALVES FOR ULTRA -SHORT WAVES : REDUCTIONOF INDUCTANCE OF GRID ELECTRODES BYCOMBINATION OF TWO OR MORE HELICES INOPPOSED SENSE.-(Hochf:tech. u. Elek:akus.,May, 1933, Vol. 41, No. 5, p. 188: GermanPat. No. 567 56o, pub. 5th Jan., 1933, Kohl.)

VALVES FOR ULTRA -SHORT WAVES: PREVENTINGDEFORMATION OF GRID ELECTRODE BYINSULATING STRIP MADE FAST TO NODALPOINTS OF SPIRAL.-(Ibid., p. 188: GermanPat. No. 567 903, pub. L1th Jan., 1933,Kohl.)

ELECTRON CONDUCTION IN THERMIONIC VALVES.-Benham. (See under " Transmission.")

THE HEXODE, A NEW GERMAN VALVE. -R. Wigand:Telefunken Company. (Radio, B., F. fur/We, June, 1933, pp. 279-283.)

Two types are described, one with indirectlyheated cathode, control grid, screen grid, secondcontrol grid, second screen grid, and anode ; theother with indirectly heated cathode, control grid,screen grid, perforated anode, second control grid,and anode. The first type is called a " fadinghexode," the second a " mixing hexode." Of theformer the writer says : " For a change of potentialof grid 1, curves are obtained resembling those ofthe exponential valves. With the latter, however,the potential change necessary to produce a verylarge variation in amplification is itself very large(about 40 v)," whereas the hexode gives a regulationof at least 000 for potential changes of the orderof 10-15 V.

In connection with the " mixing hexode " thewriter points out the desirability of a mixing valvein which the straight part of the characteristic canbe worked on, without rectification, and in whichit is easy to make the oscillator portion free fromharmonics. As an example of the use of the twotypes, the diagram (Fig. 337) is given of a 5 -valvesuperheterodyne with a fading hexode V1 regulatedby the " binode " V, (combination of half -waverectifier, from which the volume -control potentialis taken, and 1.f. amplifier leading to the outputvalve V5), followed by a mixing hexode V2 in theNumans connection, and a second fading hexode V,not connected to the AVC. " Contrary to otherreports the hexodes are a purely German invention ;they will be put on the market by Telefunken and


Valvo at this year's Radio Show and will then beobtainable from the shops."

THE HEXODE TUBE [" Emission Valve Modulator"].-Wheeler. (Rad. Engineering, March, 1933,Vol. 13, No. 3, p. 19.) See also next abstract,and Wigand, above.

THE HEXODE VACUUM TUBE : EMISSION -VALVEMECHANISM MAKES PRACTICABLE OSCIL-LATION, MODULATION, HIGH AMPLIFICATIONAND GRID -BIAS CONTROL OF AMPLIFICATION.-H. A. Wheeler. (Rad. Engineering, April,1933, Vol. 13, No. 4, pp. 12-14.) See alsoJune Abstracts, p. 328, and above.

CONSIDERATIONS ON DETECTOR -OUTPUT TUBE SYS-TEMS [and the Possibilities of the Two -ValveSuperheterodyne]. -J. R. Nelson. (Elec-tronics, April, 1933, Vol. 6, No. 4, pp. 94-95and 98.)

The possible arrival of a two -valve superhetero-dyne receiver, using one of the recently announcedoscillator -detector valves and a possible detector -output valve, was referred to in June Abstracts,p. 326, r -h col., and the present article deals withsome of the possible types of combination valvesfor this latter service. The writer considers firstthe twin control -grid pentode ; an experimentalvalve was made up to see how much improvement ofperformance over an ordinary pentode could beobtained, and a considerable improvement wasfound, though the voltage sensitivity as a detectorwas rather poor and it was impossible to load up theoutput -valve section when the degree of modulationwas low. The simplest arrangement, a combinationdiode and pentode amplifier, should have about thesame sensitivity as that of the twin control -gridpentode. A combination triode and pentode amplifierwould have considerably better voltage sensitivity,but the writer concludes that the only true solutionas regards both output and sensitivity would be thediode, triode, pentode combination, whose com-plicated structure (with 8, or preferably 9,- con-nections) would make it only problematically eco-nomical compared with two standard valves.

STILL MORE TUBES [including " Pentagrid Con-verters" and " Duo -Diodes" with Pentodes inOutput Portion].-(QST, May, 1933, Vol.17, No. 5, pp. 3o and 68, 7o.)

DETECTOR TUBE PERFORMANCE CURVES [Diode;Conventional Detectors ; Duplex Diode -Triodes ; Types 75 and 2A6, with High -MuTriode Section ; Duplex Triode -Pentodes]. -J. R. Nelson. (Rad. Engineering, April 1933,Vol. 13, No. 4, pp. 15 and 23.)

THE 77 AS A [Self] BIASED DETECTOR WITH I00VOLTS PLATE SUPPLY [for Small UniversalA.C.-D.C. Receivers].-(Rad. Engineering,April, 1933, Vol. 13, No. 4, p. 18.)

NEW TUBE TYPE DESIGNATIONS [Valve -NumberingSystem adopted by American Radio Manu-facturers Association].-(QST, May, 1933,Vol. 17, No. 5, p. 28.)


VALVE DATA DIAGRAMS: CORRESPONDENCE. -F. P. Basto : I. Koga : G.W.O.H. (Wire-less Engineer, May, 5933, Vol. ro, No. 116,pp. 257-258.)

Letters on the Editorials referred to in Abstracts,1931, p. 441, r -h col., and March, p. 161, 1-h col.Basto points out that the logarithmic diagram canbe improved according to the following considera-tions : the maximum power output per peak voltsquared is given by P/E2 = Q/8, where Q is the" Giite " (= µG,), so that the scale of Q may givean idea of the comparative sensitivity of identicalpower valves : if this power scale is plotted withthe divisions proportional to log Q, there resultsa very useful scale in bels, or in db, easy to dividebecause it has equal divisions. Also, for theclassification of h.f. valves, another scale, usefulfor determining the maximum voltage gain obtain-able with a valve followed by a suitable trans-former, may be plotted over the Q axis.

Koga describes, and illustrates by a specimendiagram, another method using ordinary rect-angular ruled paper. This has the advantage offacilitating the interpolation of numerical valuesand the preparation of the diagram sheet, but thedisadvantages that the " Giite " is not given con-veniently and that the scale of the co-ordinatesmust be given in multiple to meet the large rangeof characteristic values.

TWO PROCESSES FOR THE FORMATION OF OXIDE -COATED CATHODES. -E. Bluhm : T6dt.(E.T.Z., 18th May, 1933, Vol. 54, No. 20,PP 477-478.)

SURFACES FROM WHICH COLD EMISSION CURRENTSAPPEAR ONLY AT VERY HIGH FIELDGRADIENTS. -C. C. Chambers. (Phys.Review, 1st May, 1933, Series 2, Vol. 43,No. 9, pp. 768-769.)

The surfaces referred to were those of thoriatedtungsten filaments after various heat treatments.

OUT -GASSING ELECTRODES BY MEANS OF A SHORTINDUCTION WINDING MOVED TO-AND-FROIN AN AXIAL DIRECTION.-(Hochptech. u.Elek:akus., May, 1933, Vol. 41, No. 5, p. 188 :German Pat. No. 566 467, pub. 21st Dec.,1932, Telefunken Company.)

NEW TUBES : NEW TUBE MATERIALS [particularlySvea Metal]. -(Electronics, April, 1933, Vol.6, No. 4, p. 93.)

APPLICATION OF GRAPHITE AS AN ANODE MATERIALTO HIGH -VACUUM TRANSMITTING TUBES. -E. E. Spitzer. (Proc. Inst. Rad. Eng., June,1933, Vol. 21, No. 6, pp. 756-757 : abstractof I.R.E. Convention paper.)

THE ROLE OF VACUUM TUBES IN A TUBE FACTORY[Measuring Low Mutual Conductances :Automatic Control of Cathode Spray Gun :Protective Circuits for Meters, etc.]. -W. P.Koechel. (Electronics, May, 1933, Vol. 6,No. 5, pp. 121-123.)

August, 1933

DIRECTIONAL WIRELESSTHE AUTOMATIC COMMUNICATION TO A MOVING

CRAFT OF ITS BEARINGS TAKEN BY TWOFIXED STATIONS. -(Hock Mech. u. Elek: akus.,Feb., 1933, Vol. 41, No. 2, p. 73 : GermanPat. No. 557 257, pub. loth Aug., 1932 ,Csejkovits and Dieckmann.)

" The signals from the craft are received by twoautomatic [direct -reading] direction finders atdifferent points, each of which transmits the bearingby the simultaneous modulation, with two differentaudio -frequencies, of its own transmitter. Thismodulation takes place in such a way that thedepth of modulation on the one frequency repre-sents the angle between zero of the indicating instru-ment and the actual indication, while the depth ofmodulation on the second frequency representsthe angle between the indication and the scalemaximum. In the receiver on board the movingcraft the two audio -frequencies are separated byfilters and led to a quotient -showing instrumentwhich controls the motion of a mechanical oroptical pointer over a chart. The position is givenby the intersection of two such pointers, eachcontrolled by one of the two fixed transmitters,either through two separate receivers or by thealternate switching -over of a common receiverto the two transmitters."

THE " IRON RING QUOTIENT METER."-Geyger.See reference on p. 457, r -h column.

DISTANCE DETERMINATION BY ELECTROMAGNETICWAVES. - (Hochf:tech. u. Elek:akus., Feb.,1933, Vol. 41, No. 2, p. 73 : German Pat.No. 545 168, pub. 3oth Sept., 1932,Weigl.)

" A short-wave radiation modulated with a longerwave is received by a second station and simul-taneously re -transmitted, being then received bythe originating station and again re -transmitted,with the result that a resonance effect occurs whenthe length of the modulating wave is equal to thedistance between the two stations or a multiplethereof." Cf. Koulikoff and Chilowsky, 1929.Abstracts, p. 168.

ULTRA -SHORT-WAVE BEACON [with GeneratingCircuit alternating between Positions oneither Side of Focus of Parabolic Reflector].- (Hochf:tech. u. Elek:akus., May, 1933,Vol. 41, No. 5, p. 188: German Pat. No.567 640, pub. 6th Jan., 1933, TelefunkenCompany and Ludenia.)

See Fig. 2. A receiver on the middle line m getsequally strong signals from the two positions ;away from m the signal from one position pre-ponderates.

RADIO SYSTEM FOR BLIND LANDING OF AIRCRAFT[Runway Localising Beacon on 300 kc/s,Two Marker Beacons on ro Mc/s withDifferent Modulation Frequencies, and aLanding Beam on roo Mc/s].-(Bur. ofStds. Tech. News Bull., April, 1933, No. 192,PP. 39-40.)

August, 1933

ACOUSTICS AND AUDIO -FREQUENCIESELECTRO-MECHANICAL OSCILLATOR PARTICULARLY

APPLICABLE TO ELECTRICAL MUSICAL INSTRU-MENTS [Loud Speaker with " PreponderentInertia " acting as Large Capacity variableby regulating Field Current or by switching -in Different Moving -Coil Windings]. -(Rev.Gen. de l'Elec., 29th April, '933, Vol. 33,No. 17, pp. 134-135 D : French Pat. No.742 079, pub. 27th Feb., 1933, Michaud.)

BACK TO QUALITY IN RADIO RECEIVERS 1 [WithCurves showing Wide Difference in ToneFidelity between Large and Small Re-ceivers]. -(Electronics, April, 1933, Vol. 6,No. 4, pp. 106-107.)

HORNLESS LARGE -SURFACE LOUD SPEAKER DIS-PENSING WITH BAFFLE AND CABINET, OUTEREDGE OF DIAPHRAGM CARRIED BACK TOWALL OF RooM.-(Hochf:tech. u. Elek:akus.,May, '933, Vol. 41, No. 5, p. 188: GermanPat. No. 568 988, pub. 14th Jan., 1933,Teuke.)

A NEW CONE LOUD SPEAKER FOR HIGH FIDELITYSOUND REPRODUCTION [Uniform Responsefrom 8o to zo 000 c/s by Voice Coil andCylinder " segregated into Masses andCompliances " and connected to " SuitablyCorrugated Cone "].-H. F. Olson. (Proc.Inst. Rad. Eng., June, 1933, Vol. 21, No. 6,pp. 760-761: abstract of I.R.E. Conventionpaper.)

CONE LOUD SPEAKER WITH CONE SPACE TAKENUP BY INDEPENDENT FILLER [e.g., WoodenCone] WITH SOUND CHANNELS LEADINGTO FLAT SURFACE: TO PREVENT INTER-FERENCE EFFECTS DUE TO DEPTH OF CONEDIAPHRAGM AND TO IMPROVE FREQUENCYCHARACTERISTIC.-(Hochpech. u. Elek:akus.,Feb., 1933, Vol. 41, No. 2, pp. 73-74German Pat. Nos. 555 543 and 559 259,pub. 23rd July and 11th October, 1932,AEG and Stenzel.)

ARRANGEMENT FOR IN -PHASE TRANSMISSION OFFORCES TO DIFFERENT PARTS OF A SURFACE,PARTICULARLY LOUD SPEAKER DIAPHRAGMS[Use of Links of Materials with Differ-ent Propagation Velocities].-(Ibid., p. 74:German Pat. No. 557 016, pub. 17th Aug.,1932, Telefunken and Hahnle.)

REFLECTION OF A SOUND WAVE FROM A CIRCULARPLATE [Experimental and Theoretical].M. Sasao. (Sci. Abstracts, Sec. A, Feb.,1933, Vol. 36, No. 422, p. 168.)

" Near the plate . . . the envelope of standingwaves, if the wavelength is smaller than the radiusof the plate, has a maximum and a minimum, andthe interval between the maxima of standing wavesis greater than half a wavelength. These factswere confirmed by theory."

RELATED [Vibration] FIGURES ON CIRCULAR PLATES.-R. C. Colwell. (Phys. Review, ist May,1933, Series 2, Vol. 43, No, 9, p. 782:abstract only.)


VAST PUBLIC ADDRESS SYSTEM AT THIRTY-FIRSTINTERNATIONAL EUCHARISTIC CONGRESS,DUBLIN, JUNE, 1932.-W. L. McPherson.(Elec. Communication, April, 1933, Vol. II,No. 4, pp. 208-222.)

ANALYSIS OF THE WAVE FORM IN SOUND -ON -FILMRECORDING BY THE INTENSITY PROCESS. -0. Sandvik and V. C. Hall. (Physik. Ber.,1st May, 1933, Vol. 14, No. 9, p. 675.)

MINIATURE TALKIES. -B. R. Davies and E. A.Buckland. (Wireless World, 2nd June, 1933,Vol. 32, p. 389.)

A description is given of a 16 -mm equipmentusing the sound -on -film process.

THE ADVANTAGES OF THE LIGHT -WEIGHT CRYSTALMICROPHONE FOR SOUND PICTURE RE-CORDING [Carried on Fishing Rod]. -(Elec-tronics, April, 1933, Vol. 6, No. 4, p. 102paragraph only.)

AKUSTISKA PROBLEM VID STUDIOLOKALER FORRUNDRADIO (The Acoustic Problem inBroadcasting Studios [Convergent Walls,Stockholm and Malm8 Studios, etc.]). -S. Lemoine. (Teknisk Tidskr., June, 1933,Vol..63, No. 22, pp. 81-87.)

SCHALL- UND SCHWINGUNGSDAMMUNG IM HOCHBAU(The Deadening of Sound and Vibration inHigh Buildings). -W. Genest. (Zeitschr.V.D.I., 22nd April, 1933, Vol. 77, No. 16,pp. 427-428.)

TRANSMIT " AUDITORY PERSPECTIVE " IN MUSIC :BELL LABORATORIES COOPERATE WITH DR.STOKOWSKI'S ORCHESTRA IN 16 000 -CYCLEFIDELITY AT TEN -FOLD VOLUME. -(Elec-tronics, May, 1933, Vol. 6, No. 5, pp.118-120.) See also July Abstracts, pp.396 and 396-397.

THE DESIGN OF FILTERS FOR CARRIER PROGRAMMECIRCUITS [and the Use of Quartz Crystals inobtaining the Necessary Side -Band Sup-pression]. -F. Ralph. (Elec. Communica-tion, April, 1933, Vol. I1, No. 4, pp. 204-207.)

A NEW CRITERION OF CIRCUIT PERFORMANCE [NewPerformance " Unit " obtained by DividingInterval between " No Performance " and" Perfect Performance " into ioo EqualParts : Effects of any Particular TechniqueEliminated]. -J. Collard. (Elec. Engineer-ing, April, 1933, Vol. II, No. 4, pp. 226-233.)

THE ADVANTAGES AND DISADVANTAGES OF THEARTICULATION, INTELLIGIBILITY, AND RE-PETITION RATE TECHNIQUES. -Collard. (Seepreceding reference.)

SCALES OF LOUDNESS. -B. G. Churcher and A. J.King. (Nature, 27th May, 1933, Vol. 131,p. 76o.)

A preliminary letter giving a table of averagefigures for the relation between decibels above the


threshold of audibility at 800 c/s and loudnessdeduced from 2 : I and 4 : r estimates of intensitymade by 3o different individuals.

453

COSTRUZIONE E PROVE DI GENERATORI A FREQUENZAACUSTICA (Construction and Tests of Audio -frequency Generators). -S. Rosani. (L'Elet-trotec., 15th Feb., 1933, Vol. 20, No. 5,pp. 102-104.)

AUDIO -FREQUENCY CONSTANTS OF CIRCUITS ANDTELEPHONE LINES [by Mallett's Methods,for Frequencies 3-12 kc/s].-S. P. Chakra-varti. (J014.71. I.E.E., May, 1933, Vol. 72,No. 437, pp 416-422.)

EMPFINDLICHE RAYLEIGH-SCHEIBEN FUR SCHALL-MESSUNGEN (Sensitive Rayleigh Discs forAcoustic Measurements). -W. Grosser.(Archiv f. Elektrot., 2nd May, 1933, Vol. 27,No. 5, pp. 329-334.)

Long rectangular discs are recommended insteadof the circular ones hitherto used. The tuningmoment exerted by sound waves on such a discis calculated.

SUPERSONIC WAVES : EXPRESSIONS FOR ENERGYTRANSMISSION AND DECAY OF AMPLITUDEWITH DISTANCE DUE TO VISCOSITY AND HEATCONDUCTION. -P.. Biquard. (Sci. Abstracts,Sec.A, Jan., 1933, Vol. 36, No. 421, p. 78.)

APPLICATIONS OF SUPERSONICS, INCLUDING DEPTHSOUNDING AND SUBMARINE SIGNALLING. -C. Florisson. and F. Vecchiacchi. (Sci.Abstracts, Sec.B, Jan., 1933, Vol. 36, No.421, pp. 67-68: Internat. Elec. CongressPaper.)

UNSUCCESSFUL TESTS ON THE TRANSMISSION OF ANULTRASONIC BEAM THROUGH AIR. -G. S.Field. (National Research Council, Canada,Fifteenth Annual Report, 1931-1932, p. 55.)

'OBER EINIGE VERSUCHE ZUR BEUGUNG DES LICHTESAN ULTRASCHALLWELLEN (Some Experi-ments on the Diffraction of Light by Ultra-sonic Waves [in Liquids]). -R. Bar andE. Meyer. (Physik. Zeitschr., 15th May,1933, Vol. 34, No. It), pp. 393-396.) Seealso Debye, March Abstracts, p. 167, r -hcolumn.

PHOTOTELEGRAPHY AND TELEVISION

PHOTOTELEGRAPHY OF HALF -TONE PICTURES BYMULTIPLE MODULATION OF A SHORT WAVE.-(Hochf:tech. u. Elek:akus., March, 1933,Vol. 41, No. 3, p. 112: German Pat.No. 563 op, pub. 3rd Nov., 1932, Telefunkenand Schrater.)

Each group of tone values has a modulationfrequency allotted to it, which is set in action whenthe light spot of an oscillograph falls on the par-ticular one of a row of photocells. The oscillographdeflection is controlled by the intensity of the lightfalling on the scanning photocell.

August, 1933

DIE ERSTEN UKW-FERNSEH-EMPFANGER AUF DEMMARKT (The First Ultra -Short -Wave Tele-vision Receiver on the Market [Te-Ka-DeReceiver]). -F. Noack. (Radio, B., F.fur Alle, March, 1933, pp. 101-103.)

THE ICONOSCOPE-A MODERN VERSION OF THEELECTRIC EYE [Photo -Sensitive Surfacescanned by Cathode -Ray Beam]. -V. K.Zworykin. (Proc. Inst. Rad. Eng., June,1933, Vol. 21, No. 6, p. 755 : abstractof I.R.E. Coilvention paper.)

DIE VERZERRUNGEN DURCH DIE RAUMLADUNG INDER BRAUNSCHEN ROHRE (Distortions causedby the Space Charge in Cathode -RayTubes [and the Use of a Bent Tube]). -E. Hudec. (E.N.T., May, 5933, Vol. io,No. 5, pp. 215-22o.)

The movement of the light spot in cathode-raytelevision is controlled by two trip voltages, arapid one for the " line " deflection and a slow onefor the " picture " deflection. The return -strokeof the line deflection is invisible on account ofits speed and the consequent feeble luminosity ;the return -stroke of the " picture " deflection israpid in comparison with its forward motion,but is slow enough to be of the same order as theforward motion of the quick " line " deflection,and if allowed to remain in the picture is conspicuousas a bright line down one edge. This can beavoided by taking it outside the frame by makingthe long synchronising impulse for the " picture "deflection act also on the trip potential for the" line " deflection.

After this preliminary discussion the writergoes on to describe the occurrence of a central' cross," brighter than the rest of the picture

and stretching from side to side and from topto bottom, which becomes more and more prominentas the working ray current of a gas -filled tubeincreases from o.o5 to 0.2 mA (Fig. 3, a to c, wherethe current is varied by increasing the Wehneltcylinder potential). The appearance of this " cross "is accompanied by an increase in the thickness of,and a decrease in the spacing between, the linesin the middle of the picture, so that apart from thedefect of the cross itself the centre of the picturesuffers distortion. Such distortion can also occurwhen the tube is used as an oscillograph, butthe trouble is not so marked in this case becausea less bright spot, and consequently a smallerray current, can be used. The writer discussesthe causes of the " cross " and the distortion :the changes in the line spacing and thickness aredue to the space charge between the deflectingplates ; the ions produce a subsidiary field whichdistOrts the main field in the manner shown inFig. 5. The greater the potentials causing themain field, the smaller the effect of the ion field :it is when the deflecting potentials are small thatthe thickening, etc. of the lines occur. Thisaccounts for the formation of the central " cross "of Fig. 3, discussed above ; for in the tube employedthe ray was deflected in both directions, so that thedeflecting potential was zero in the middle. Toavoid these troubles, therefore, each deflectionmust be in one direction only, so that the zero

August, i933 454

potential is at the extreme edge of the picture :this means that the position of rest of the spot mustbe at one corner of the picture.

In order to obtain this result while still makingfull use of the screen area, the German Post Officehave developed a bent container, the conicalpart of which is at a slight angle to the shortcylindrical neck containing the electrode system.This arrangement brings the spot, in its central," rest " position, close to the corner of the screen(Figs. 8, 9 and to). Fig. ii shows the good resultobtained.

Another way of bringing the distortion to theedge of the picture is to employ a pair of auxiliarydeflecting plates, at an angle to both the mainpairs and carrying a constant d.c. potential whichbrings the ray from its central position S to a" corner " position S' (Fig. is). But this planhas certain defects which decrease the sensitivity,and better results are obtained with two pairs ofauxiliary plates. In Fig. 14, obtained with thisarrangement, the return stroke of the slow trippotential is seen crossing the picture. This can besuppressed as in the previous records, but the simplestplan is to darken the picture at the transmitterduring this instant.

THE THUN " LINE CONTROL " SYSTEM [VariableSpeed Scanning]. -(German Pat. No. 563 294,pub. 3rd Nov., 1932, Thun.) See 1932Abstracts, pp. 44, r -h column ; also January,pp 46-47)

SCANNING DISC WITH TWO -TURN SPIRAL, COM-BINED WITH HALF -SPEED DISC WITH SPIRALWINDOW : SMALL DIAMETERS WITHOUTLOSS OF BRIGHTNESS. -(Rev. Gen. de

xst April, 1933, Vol. 33, No. 13,p. IOI D : French Pat. No. 739 644, pub.14th Jan., 1933, Barthelemy.)

PHOTOCELL AND AMPLIFIER COMBINATION WITHVERY SHORT OR No CONNECTING LEADS,BY USE OF VALVE WITH EXTERNAL CONTROLELECTRODE.-(Hochf : tech. u. Elek : akus.,May, i933, Vol. 41, No. 5, p. 188: GermanPat. No. 568 093, pub. 14th Jan., 1933.Tonbild-Syndikat.)

PHOTOCELLS AND THEIR APPLICATIONS. -V. K.Zworykin and E. D. Wilson. (Book Reviewin Review Scient. Instr., May, 1933, Vol. 4,No. 5, p. 316.)

THE [Constant Quantic] EFFICIENCY OF FLUORES-CENCE OF SODIUM SALICYLATE [in connectionwith the Construction of a PhotoelectricCell for Ultra -Violet Light]. -P. Dubouloz.(Comptes Rendus, 24th April, 1933, Vol. 196,No. 17, pp. 1221-1222.) See 1932 Abstracts,p. 422, r -h column.

THE USE OF COLOUR FILTERS, CONTROLLING PARTONLY OF SENSITIVE SURFACE, TO MAKESPECTRAL SENSITIVITY OF PHOTOELECTRICCELL SIMILAR TO THAT OF THE EYE. -A.Dresler. (E.T.Z., 18th May, 1933, Vol. 54,No. 20, pp. 476-477: summary only.)


EXPERIMENTS ON THE SUITABILITY OF SOMERECTIFIER PHOTO CELLS FOR THE MEASURE-MENT OF DAYLIGHT [Auger, Serpidox,Siemens and Halske, Westinghouse, Berg-mann Selenium, etc.]. -H. H. Poole andW. R. G. Atkins. (Electrician, 9th June,1933, Vol. 110, No. 287i, pp. 737-738.)

NON -ADDITIVE EFFECT OF DIFFERENT RADIA-TIONS ON COPPER OXIDE PHOTOCELLS. -C. Lapique. (Comptes Rendus, ist May,1933, Vol. 196, No. 18, pp. 1301-1303.)

Leading from the researches on the sensitivityof these cells (1931 Abstracts, p. 62,, r -h col.)the tests here described show that if a radiation ofwavelength greater than 65o mµ, acting alone,produces a current iI, and another radiation ofwavelength between 450 and 65o mau, also actingalone, produces a current i2, the two radiationsacting together produce a current greater thani2 : there is a positive supplementary effecteven if i1 is negative (owing to the inverse photo-electric effect which occurs in that part of thespectrum). This supplementary sensitivity variesconsiderably in magnitude from one cell to another,but not in characteristic form nor in its positionin the spectrum. In particular, it begins markedlyat about 45o my., just where the ordinary sensitivitycurve always shows a bend which appears to indicatethe superposition of two photoelectric effects, theone general and the other more selective. Appa-rently the more selective effect is amplified by thepresence of the longer wave, while the other effectis not altered. The clearly selective characterof the supplementary sensitivity shows that itcannot be attributed to a simple variation of theinternal resistance of the cell.

UBER. DAS VERHALTEN VON SPERRSCHICHT-PHOTO-ZELLEN BEI BELICHTUNG MIT RoNTGEN-STRAHLEN (The Behaviour of Barrier -LayerPhotocells illuminated by X -Radiation). -K. Scharf and 0. Weinbaum. (Physik.Zeitschr., ist April, 1933, Vol. 34, No. 7,p. 283.)

A note on experiments, the full paper on whichhas already been referred to in May Abstracts,pp. 277-278.

REVERSAL OF CURRENT IN RECTIFIER PHOTO-CELLS. -H. H. Poole and W. R. G. Atkins :Guild. (Nature, i5th April, 1933, Vol. 131,pp 547-548.)

A reply to the letter by Guild dealt with in MayAbstracts, p. 277, r -h column.

A NEW TYPE OF PHOTOELECTRIC EFFECT INCUPROUS OXIDE IN A MAGNETIC FIELD. -I. Kikoin and M. Noskow. (Nature, aothMay, 1933, Vol. 131, pp. 725-726.)

A preliminary account of an effect of lightincident perpendicularly on a plate immersed inliquid air, placed in a magnetic field parallel to itsplane ; an e.m.f. is produced in the plate per-pendicular to the direction of the light and of themagnetic field. Its magnitude is proportional tothe field strength, up to fields of about 2 500 gauss,and the direction of supposed flow of the negative


electrons coincides with the direction of the lightbeam. The effect disappears completely whenred light is used and is not found at room tem-peratures.

THE DISCONTINUITIES OF POTENTIAL AT THECONTACT OF A SEMI -CONDUCTOR AND AMETALLIC ELECTRODE. -G. Dechene.(Comptes Rendus, 22nd May, 1933, Vol. 196,No. 21, pp. 1577-1579.)

THE RESPONSE OF BARRIER -LAYER PHOTO -CELLSTO X -RAYS. -P. R. Gleason. (Phys. Review,1st May, 1933, Series 2, Vol. 43, No. 9, pp.775-776: abstract only.)

PHOTOELECTRIC EFFECT IN CUPRITE SINGLECRYSTALS. -R. Deaglio : Dember. (ComptesRendus, ist May, 1933, Vol. 196, No. 18,pp. 1303-1305.)

Experiments on the internal photoelectric effect(" crystal photoeffect " : Dember, 1932 Abstracts,pp. 232 and 291, 1-h columns) in crystals of cuprite,leading to the conclusion that Dember's inter-pretation on electronic lines is wrong and that thephenomenon is due to the property possessed bycuprite of acquiring an electrolytic conductivity onillumination.

tTBER DEN PHOTOEFFEKT IN CUPRITKRISTALLEN(The Photoelectric Effect in Cuprite Cry-stals). -Anna Joffe and A. Joffe. (Zeitschr.f. Physik, 1933, Vol. 82, No. II/12, pp.754-758-)

In the experiments of which this is a preliminarydescription it was found that the potential distribu-tion in an illuminated crystal does not agree withthe diffusion theory ; nor does the variation ofphoto -electromotive force agree with the intensityof the light. The phenomena can, however, allbe explained on the assumption that the photo-electric current is carried by electrons of otherenergy levels than those forming the current in theabsence of illumination.

VBER DEN LICHTELEKTRISCHEN EFFEKT IN BESON-DERER AKTIVER SCHICHT DER KARBORUND-KRISTALLE (The Photoelectric Effect in aParticular Active Layer of the CarborundumCrystal). -O. W. Lossew. (Physik. Zeitschr.,15th May, 1933, Vol. 34, No. Jo, pp. 397-403.)

A continuation of work already referred to in1932 Abstracts, p. 108, r -h column.

PAPERS ON PHOTOVOLTAIC CELLS OF SALTS OFCOPPER. -R. Audubert. (Comptes Rendus,8th and 22nd May, 1933, Vol. 196, Nos. 19and 21, pp. 1386-1,387 and 1588-1590.)

Researches giving quantitative confirmation ofthe writer's theory of photovoltaic actions (Ab-stracts, May, p. 278, 1-h column ; also 1932, p.470, r -h column).

THE INTERNAL PHOTOELECTRIC EFFECT IN LIQUIDDIELECTRICS. -G. Liandrat. (ComptesRendus, 8th May, 1933, Vol. 196, No. 19,pp. 1385-1386.)

August, 1933

THE ROLE OF COLOURING MATTER IN PHOTOVOLTAIC PHENOMENA. -C. Stora. (Rev.Gen. de 27th May, 1933, Vol. 33,No. 21, p. 698: long summary.)

THE PHOTOELECTRIC SENSITIVITY OF MAGNESIUM. -R. J. Cashman and W. S Huxford. (Phys.Review, 15th May, 1933, Series 2, Vol. 43,No. io, pp. 811-818.)

PHOTOELECTRIC PROPERTIES OF MAGNESIUM. -G. Dejardin and R. Schwegler. (ComptesRendus, 22nd May, 1933, Vol. 196, No. 21,pp. 1585-1587.)

Including criticisms of the results of de Laszlo(1932 Abstracts, p. 470).ELEKTRONENBEUGUNG UND LICHTELEKTRISCHE

WIRKUNG AN ALKALIMETALLOBERFLACHEN(Electron Diffraction and PhotoelectricEffect at Alkali Metal Surfaces). -W. Klugeand E. Rupp. (Zeitschr. f. Physik, 1933,Vol. 82, No. 9/1o, pp. 568-583.) Continu-ation of work referred to in 1932 Abstracts,p. 649, 1-h col. See also pp. 232-233.

VARIATION OF PHOTOELECTRIC EFFICIENCY WITHWORK FUNCTION IN THE EXTREME ULTRA -VIOLET. -C. Kenty. (Phys. Review, 1stMay, 1933, Series 2, Vol. 43, No. 9, p. 776abstract only.)

THEORY OF THE ENERGY DISTRIBUTION OF PHOTO-ELECTRONS. -L. A. Du Bridge. (Phys.Review, 1st May, 1933, Series 2, Vol. 43,No. 9, pp. 727-741.)

UNTERSUCHUNG fiBER DEN INNEREN LICHTELEK-TRISCHEN EFFEKT AN METALLEN (Investiga-tions of the Internal Photoelectric Effectin Metals). -R. Schulze. (Physik. Zeitschr.,1st May, 1933, Vol. 34, No. 9, pp. 381-385.)

The internal photoelectric effect for variousmetals in the thinnest possible film was measured ;measurements with varying directions of incidenceof light also gave increased current in the directionof the light.

EIN BEITRAG ZUR BESTIMMUNG DES REFLEXIONS-VERMOGENS VON METALLEN IM SICHTBARENUND IM ULTRAVIOLETT (Contribution tothe [Experimental] Determination of theReflection Coefficients of Metals for Visibleand Ultra -Violet Light [including Photo-electric Method for Ultra -Violet Light]). -K. von Fragstein. (Ann. der Physik, May,1933, Series 5, Vol. 17, No. 1, pp. I-21.)

ENERGY DISTRIBUTION OF PHOTOELECTRONS FROMZINC SURFACES. -N. E. Bradbury. (Phys.Review, ,5th March, 1933, Series 2, Vol. 43,No. 6, p. 502: abstract only.)

'OBER DIE KLASSISCHE VERTEILUNG DER ELEK-T RONEN IM PHOTOELEKTRISCHEN EFFEKTDER RONTGENSTRAHLEN (On the ClassicalDistribution of Electrons in the Photo-electric Effect of X -Rays [TheoreticalInvestigation]). -J. Kunz. (Physik. Zeit-schr., ist March, 1933, Vol. 34, No. 5, pp.219-220.)

D 2

1

August, 1933 456

ERHALTUNG DER ENERGIE UND DES IMPULSES IMPHOTOELEKTRISCHEN EFFEKT (Conservationof Energy and Impulse in the PhotoelectricEffect [Theoretical Investigation]). -J. Kunz.(Physik. Zeitschr., ist March, 1933, Vol. 34,No. 5, pp. 218-219.)

LICHTELEKTRISCHE LEITUNG UND ABSORPTIONDER LENARDPHOSPHORE IM ROTEN UNDULTRAROTEN SPEKTRALGEBIET (PhotoelectricConduction and Absorption in LenardPhosphorescent Materials [e.g. casei, cas,zns with or without fluxes] in the Red andInfra -Red Regions of the Spectrum). -L.Weber. (Ann. der Physik, 1933, Series 5,Vol. x6, No. 7, pp. 821-843.)

DER EINFLUSS DER MATERIE AUF LANGSAMSTEELEKTRONEN NACH LICHTELEKTRISCHENUNTERSUCHUNGEN (The Influence of Matteron Very Slow Electrons, investigated Photo-electrically). -G. Lang. (Ann. der Physik,1933, Series 5, Vol. 16, No. 7, pp. 781-792.)

MEASUREMENTS AND STANDARDSON SOME NEW FORMULAE FOR THE CALCULATION OF

SELF AND MUTUAL INDUCTION OF COAXIALCIRCULAR COILS IN TERMS OF ARITHMETICO-GEOMETRICAL SCALES. -L. V. King. (Phil.Mag., June, 1933, Series 7, Vol. 15, No.102, pp. 1097-1114.)

The writer uses well-known results from thetheory of gravitational attractions to obtain highlyconvergent expressions for the self and mutual in-duction of various arrangements oS coaxial circularcoils and develops useful computational forms fordirect accurate calculations without reference totable% of elliptic integrals.

ON THE USE OF ARITHMETICO-GEOMETRIC MEANSERIES FOR THE CALCULATION OF ELLIPTICINTEGRALS, WITH SPECIAL REFERENCE TOTHE CALCULATION OF INDUCTION. -F. W.Grover. (Phil. Mag., June, 1933, Series 7,Vol. 15, No. 102, pp. 1115-1133.)

This paper illustrates the use for numerical cal-culation of the formulae developed in the paperreferred to above.

THE SELF AND MUTUAL INDUCTANCES OF LONGCONCENTRIC MULTI -LAYER SOLENOIDS. -P.Janet. (Rev. Gin. de l'Elec., 13th May, 1933,Vol. 33, No. 19, pp. 619-623.)

" The expression given in a classic work andreproduced in a certain number of formulae for theself inductance of a long multi -layer solenoid beingerroneous, I think it useful to give here the exactvalue " : for this the writer derives the formula

L = 2/3 7,2 lab 3a2),

where n is the total number of turns and a and b arethe internal and external radii of the winding.Dealing similarly with the various cases of two con-centric solenoids, he finds formulae for the mutualinductances, and shows that the erroneous self-inductance formula referred to above

(L = 413 . n'n2// (b2 ab a2))was due to its derivation from the mutual -inductance


formula for concentric solenoids one inside the other,having no part in common ; it should have beenderived from the case where one winding encroacheson, or is entirely embedded in, the other winding.

INTERNATIONAL ACTION ON UNITS IN ELECTRICITYAND LIGHT. -Bureau of Standards. (Journ.Franklin Inst., April, 1933, Vol. 215, No. 4,PP. 471-476.)

This note gives the formal recommendations on.electrical units and standards and on photometryto be submitted to the International Committee on.Weights and Measures.

DIMENSIONS OF FUNDAMENTAL UNITS. -F. M.Denton. (Nature, 22nd April, 1933, Vol. 131,P. 585.)

This letter answers various recent letters on thesubject in the same journal, and gives argumentsfor regarding " magnetic pole " as a unit derivedfrom ' electric charge " rather than as an inde-pendent unit.

A NEW SYSTEM OF ELECTRIC AND MAGNETICUNITS, AND THEIR DIMENSIONS. -H. Bucher.(Rev. Gen. de l'Elec., loth May, 1933, Vol.33, No. 20, pp. 665-666.)

A NEW EXPOSITION OF ELECTROMAGNETIC PHENO-MENA: THE USELESSNESS OF THE IDEA OFMAGNETIC MASS. -N. Vasilesco-Carpen.(Rev. Gen. de l'Elec., 6th May, 1933, Vol. 33,No. 18, pp. 585-586.)

DUE CONFRONT! SIGNIFICATIVI SULL' USO DELLEUNITA ASSOLUTE M.K.S.1? (Two Significant.Comparisons of the Use of the AbsoluteUnit M.K.S.Q [and of the orthodox C.G.S.Unit : Calculations of the Capacity of theEarth and the Inductance of a Coil]). -G. Giorgi. (L'Elettrotec., 25th Feb., 1933,Vol. 20, No. 6, p. 125.)

THE SIMPLIFICATION OF ACCURATE MEASUREMENTOF RADIO -FREQUENCY. -W. H. F. Griffiths.(Wireless Engineer, May and June, 1933,Vol. 1o, Nos. 116 and 117, pp. 239-247 and299-306.)

The development of harmonic wavemeters : avariable frequency resonant circuit of great stabilityand calibration permanence (with thermal ex-pansion compensation -1930 Abstracts, pp. 52-53 and r x2) : the development of a simple dynatron.oscillating wavemeter : the selection of valves andoperating conditions : the effect of the valve uponfrequency, etc.

PRACTICAL MEASUREMENT OF ULTRA -SHORT WAVES.-C. C. Whitehead. (World -Radio, and and.9th June, 1933, Vol. 16, Nos. 410 and 411,pp. 747 and 774-775 : to be continued.)

A WAVEMETER WITH ALTERNATIVE CLOSE ANDOPEN SCALES. -Radio Research Station,Slough. (Wireless Engineer, May, 1933,Vol. xo, No. 116, pp. 255-257.)

Authors' summary :-" An absorption wave -meter is described which can be used (a) with a


wide frequency scale, (b) with a restricted range andcorrespondingly greater openness of scale. Observa-tions are made on the calibration of such wave -meters, and their use in connection with receivingapparatus. The use of the principle in receivertuning is noted." The normal, wide -range con-dition is " opened up " by the introduction of twofixed condensers, one in series with a tuning con-denser and the other across the series of two thusformed, and therefore also across the fixed induct-ance. When a grid -current meter is not used, the" double beat " method is employed (Colebrook,1932 Abstracts, p. 175, 1-h column). Cf. andcontrast Rohde and Schwarz, Jan. Abstracts, p.49, r -h column.

CONTRIBUTION TO THE STUDY OF THE STABILITYOF OSCILLATIONS IN COUPLED CIRCUITS[e.g. in Oscillating Circuit coupled to Wave -meter]. -Mercier. (See under " Propertiesof Circuits.")

NEW RADIO TEST OSCILLATOR [With High Fre-quency Stability due to " Electron -Coupled "Circuit].-(Rad. Engineering, May, 1933,Vol. 13, No. 5, p. 26.)

CRYSTAL -CONTROLLED CLOCK WITH QUARTZ VIBRAT-ING AT SUCH A LOW RATE AS TO REQUIRENo FREQUENCY REDUCTION.-S011enberger.(Sci. News Letter, 27th May, 5933, Vol. 23,No. 633, p. 325.)

A STANDARD OF RADIO FREQUENCY [Quartz -Driven Clock : the Necessity of devisinga Ball Thrust Bearing to replace the JewelBearing]. -W. A. Steel. (National ResearchCouncil, Canada, Fifteenth Annual Report,1931-1932, p 42.)

" The accuracy of the frequency of the crystalhas at all times been better than one part permillion and has been held to within 3 parts in iomillion for periods of a month at a time."

POSSIBLE ERRORS IN LONGITUDE DETERMINATIONSDEPENDING ON THE APPARENT PROPAGATIONVELOCITY OF SHORT WAVES.-Stoyko andJouaust. (See abstracts under " Propaga-tion of Waves.")

A MULTIPLE INTERFEROMETER FOR ANALYSINGTHE VIBRATIONS OF A QUARTZ PLATE. -H. Osterberg. (Phys. Review, 15th May,1933, Series 2, Vol. 43, No. 1o, pp. 819-829.)

" This interferometer combines in a convenientform six interferometers for measuring the relativemotion of any two plane surfaces of the vibratingplate."

HIGH -FREQUENCY MEASUREMENTS. -A. Hund.(Book Review in Review Scient. Instr., May,1933, Vol. 4, No. 5, pp. 316-317.)

FIELD -STRENGTH MEASURING SET FOR MOTOR-CAR TRANSPORT, AND RESULTS OBTAINED.-David. (See reference under " Propaga-tion of Waves.")

August, 1933

THE MEASUREMENT OF THE POWER FACTOR ANDCAPACITANCE OF A CONDENSER BY COM-PARISON WITH A MUTUAL INDUCTANCE [usingI 000 c/s Mains and giving Great Accuracy].-A. W. Smith. (Review Scient. Instr., May,1933, Vol. 4, No. 5, pp. 280-284.)

The secondary of the mutual inductance is in thetelephone bridge circuit and carries zero current.A Wagner ground circuit gives complete silence atthe balance point. The final balance is obtainedby adjusting a small variable self-inductance inseries with the primary of the mutual inductance :for a given capacitance and frequency the valueof this self-inductance is directly proportional tothe power factor of the condenser.

SOME IMPORTANT APPLICATIONS OF THE A.C.POTENTIOMETER WITH CAPACITIVE POTENTIALDIVISION [and the Measurement of the LossAngles of Dielectrics at High Voltages]. -A. M. Angelini. (L'Elettrotec., 5th Feb., 1933,Vol. 20, No. 4, pp. 77-81.)

A METHOD OF AUTOMATIC COMPENSATION OFPOTENTIALS AND ITS APPLICATION TODETERMINING THE POTENTIAL DISTRIBU-TION IN L.F. ELECTRIC FIELDS. -S. Duni-kowski. (Rev. Gen. de l'Elec., 3rd June,1933, Vol. 33, No. 22, pp. 719-727.)

DEGREE OF MODULATION MEASURED BY TWO -VALVEVOLTMETERS, ONE WITH PARABOLIC CHAR-ACTERISTIC AND THE OTHER ACTING ASLINEAR RECTIFIER.-(Hochttech. u. Elek:akus., March, 1933, Vol. 45, No. 3, pp.io-I : German Pat. No. 561190, pub.

Irth Oct., 1932, Wigge.)

DETERMINATION OF THE CHEMICAL COMPOSITIONOF LIQUIDS BY ELECTRICAL RESISTANCEMEASUREMENTS [Using the Geyger " IronRing Quotient Meter "3.-W. Geyger. (Zeit-schr. V.D.I., 13th May, 1933, Vol. 77, No. 19,pp. 507-508.

TEILWEISE AUFHEBUNG DER MITTELWERTBILDUNGBEI MESSINSTRUMENTEN (The PartialElimination of the Averaging Effect[" Smoothing " due to Inertia] of MeasuringInstruments). -E. Gschwind and W. Loh-mann. (Zeitschr. f. tech. Phys., No. 4, vol.14, 1933, pp. I69-172.)

ASTATIC PRECISION METERS FOR CURRENT, VOLTAGEAND POWER. -Siemens & Halske. (E.T.Z.,I 8th May, 1933, Vol. 54, No. 20, pp. 475-476.)

SUBSIDIARY APPARATUS AND MATERIALSUN FREQUENCEMETRE ENREGISTREUR A CONSTANTE

DE TEMPS (A Recording Frequency Counterdepending on the Time Constant of a Circuit[Counter for Atmospherics, Tachometer, etc.])-J. Lugeon and J. Gurtzman. (L'Onde Elec.,Feb., 1933 [publication delayed], Vol. 12, No.134, pp. 71-83.)

The process whose frequency is to be measured ismade to work a relay with front and back contacts.


With the armature in repose the back contactconnects a condenser C1 to a source of d.c. potentialV ; the arrival of an impulse first breaks this con-tact and then makes the front contact, which con-nects the charged condenser across a second, muchlarger, condenser C2, shunted by a recording milli -ammeter of resistance R. The values for the com-ponents of an equipment for recording the averagenumber of atmospherics per minute are as follows :-

V = io V, Cl = 50 p,F, C2 = 40 000µF, R= goo ohms.

The bulk of the paper is taken up with the mathe-matical proof that the milliammeter reading is asimple function of the frequency of the relaymovements. A specimen record of atmospherics isgiven, below a simultaneous record taken on themore complex " atmoradiograph " with Richardcinemograph, previously used (1930 Abstracts,pp. 502-503.)

HIGH-SPEED CATHODE-RAY OSCILLOGRAPHIC RE-CORDING ON COMMERCIAL PHOTOGRAPHICPAPER. -I. Stekolnikoff : Forster. (Zeit-schr. f. tech. Phys., No. 6, Vol. 14, 1933,13.354.)

Referring to Forster's paper dealt with in 1932Abstracts, p. 596, 1-h cok, the writer points out thatwith his oscillograph described in Westnik Elek-trotechniki, 1931, p. 43 [see also 1932 Abstracts,p. 277, 1-h column] he obtained recording speeds upto 5 000 km/sec. on commercial photographicpaper.

EIN KATHODENSTRAHLOSZILLOGRAPH ZUR UN MIT-TELBAREN AUFNAHME SEHR HONER GLEICH-UND WECHSEL-SPANNUNGEN (A Cathode -Ray Oscillograph for Direct Photographyof Very High Direct and AlternatingVoltages [of the order of Too kv]).-M.Messner. (Archiv f. Elektrot., 2nd May,1933, Vol. 27, No. 5, pp. 335-340.)

A new deflecting system, contained in a specialhigh -voltage chamber, is described.

RECENT DEVELOPMENTS IN CATHODE-RAY TUBESAND ASSOCIATED APPARATUS [includingCathautograph and Cathode -Ray Compass].-A. B. Du Mont. (Rad. Engineering,March and April, 1933, Vol. 13, Nos. 3and 4, pp. 15-19 and 19-20 and 23.)

NEW CATHODE-RAY TUBE SCREEN MATERIAL. -Du Mont Laboratories. (Electronics, April,1933, Vol. 6, No. 4, p. 113.)

" . . . produces a spot intensity five times asbrilliant as any screen previously used . . . Aunique feature . . . is the ability to retain forwell over a minute any wave or figure applied toit when used in a darkened room or hood. Thisfeature, however, in no way affects the use of thetube for ordinary oscillograph use or photographicrecording, because of the large difference betweenthe spot intensity and the after -glow." No detailsare given.

LIAISON BETWEEN THE TWO GENERAL METHODS OFPREPARING PHOSPHORESCENT ZINC SULPHIDE.-R. Coustal. (Comptes Rendus, 1st May,1933, Vol. 196, No. 18, pp. 1306-1307.)

THE LUMINESCENCE OF ALKALINE EARTH TUNGS-TATES CONTAINING LEAD. -F. E. Swindells.(Journ. Opt. Soc. Am., April, 1933, Vol. 23,No. 4, pp. 129-132.)

THE PREPARATION OF METALLIC TUNGSTATES[and the Advantage of the Presence of anAlkaline Chloride]. -A. Karl. (ComptesRendus, 8th May, 1933, Vol. 196, No. 19,pp. 1403-1404.)

THE EFFICIENCY OF FLUORESCENCE OF SODIUMSALICYLATE.-P. Doubouloz. (ComptesRendus, 24th April, 1933, Vol. 196, No. 17,pp. 1221-1222.)

ZUR THEORIE DER GASKONZENTRATION VON ELEK-TRONENSTRAHLEN (Theory of the Gas Con-centration of Electron Beams). -0. Scherzer.(Zeitschr. f. Physik., 1933, Vol. 82, No. II/12,pp. 697-708.)

Author's summary :-This paper contains amathematical investigation of stationary conditionsof gas -concentrated electron beams under simplifyingassumptions. It is found that, if the distributionof electrons over the cross-section of the beam isgiven, the density distribution of the ions whichconcentrate the beam is also given. In the specialcase of the beam with nodes, the electron and iondensities are constant over the whole cross-section ;the equation connecting them is given. The distancebetween two nodes is found to be always greaterthan w/ep 'm_/m+ where ip is the probabilityof ionisation and m_/m+ the ratio electron mass/ionmass. It is shown how a layer of positive chargein a plasma surrounds itself with a region of negativespace -charge.

DISTORTIONS DUE TO SPACE CHARGE IN CATHODE-RAY TUBES.-Hudec. (See under " Photo -telegraphy and Television.")

'OBER DIE OPTIK DER BRAUNSCHEN NIEDERSPAN-NUNGSRoHRE (The Optics of the Braun Low -Voltage [Cathode -Ray] Tube). -E. Briiche.(Archiv f. Elektrot., loth April, 1933, Vol. 27,No. 4, pp. 266-274)

Author's summary :-The electrode system of alow voltage oscillograph, working with gas con-centration, must be regarded as an arrangementsimilar to the electron microscope. The enlargedimage of the cathode produced by this system onthe fluorescent screen is diminished to an illuminatedpoint by the effect of the optics of gas concentration.This way of regarding a Braun low -voltage tube isdemonstrated by experiments with an A.E.G.oscillograph, and a number of conclusions aredrawn which may be experimentally confirmedand are of interest in the construction of oscillo-graphs.

ZUR THEORIE DER ELEKTRONENOPTIK (On theTheory of Electron Optics). -J. Picht :Scherzer. (Zeitschr. f. tech. Phys., No. 6,Vol. 14,1933, pp. 239-241.)

Showing that the writer's formula for the focallength of an " electrical lens " (Abstracts, April,p. 224, r -h col.) is identical with that found inde-pendently by Scherzer (ibid., same col.). Both


formulae are second approximations, and the writernow derives a closer approximation for focal lengthand for the position of the principal points.

ELECTRON OPTICS: ON ELECTRON BEAMS INHIGH VACUUM. -C. J. Davisson : V. K.Zworykin. (Phys. Review, 1st May, 1933,Series 2, Vol. 43, No. 9, pp. 777 and 778-779 : abstracts only of Surveys.)

SYMPOSIUM ON ELECTRON OPTICS.-Davisson,Farnsworth, Swami and Danforth, Zworykin.(Journ. Opt. Soc. Am., May, 1933, Vol. 23,No. 5, pp. 190-192: summaries only.)

THE EFFECTIVENESS OF PHOTOGRAPHIC LAYERS INSPECTROGRAPHY OF ALL WAVELENTHS [in-cluding Cathode Rays : a Survey]. -J. Eggert. (Zeitschr. f. tech. Phys., No. 5,Vol. 14,1933, pp. 177-186.)

VAPOUR PRESSURE OF VACUUM CEMENTS. -R. M.Zabel. (Review Scient. Instr., April, 1933,Vol. 4, No. 4, pp. 233-234.)

NOTES ON THE MANUFACTURE OF MERCURY VAPOURRECTIFIER TUBES. -P. G. Weiller. (Elec-tronics, April, 1933, Vol. 6, No. 4, pp. 99-

oi . )The coating process : use of netting as filament

base : relation between gas pressure and life :" long life, the reward of many items."

THE IMPORTANCE OF TEMPERATURE REGULATIONIN THE USE OF MERCURY VAPOUR RECTI-FIERS, AND A DEVICE FOR THIS PURPOSE. -C. Stansbury and G. C. Brown. (Electronics,April, 1933, Vol. 6, No. 4, p. 104.)

THE SINGLE -TUBE THYRATRON INVERTER [andIts Applications, particularly as Generatorof Saw -Tooth Voltage for Linear Time Baseor Spot -Welding Control]. -0. W. Livingstonand H. W. Lord. (Electronics, April, 1933,Vol. 6, No. 4, pp. 96-98.)

CALCULATIONS ON THYRATRONS [on Basis ofRichardson Equation: Desirable RelationsFound]. -J. Frenkel. (Sci. Abstracts, Sec.B, March, 1933, Vol. 36, No. 423, p. 2oo.)

THE TEMPORAL COURSE OF THE IGNITION OF IONICVALVES [Thyratrons].-Klemperer. (Archivf. Elektrot., 2nd May, 1933, Vol. 27, pp.322-328.)

DIE STROM- UND SPANN UNGSVERHALTNISSE DERGITTERGESTEUERTEN GLEICHRICHTER (TheCurrent and Voltage Ratios of the Grid -Controlled Rectifiers). -K. Mfiller-Liibeckand E. Uhlmann. (Archiv f. Elektrot., 2ndMay, 1933, Vol. 27, pp. 347-373.)

GAS AND VAPOUR TUBE MULTIPLIERS FOR INDI-CATOR USE [The Use of Grid -ControlledRectifiers for Amplification of Small D.C.Voltages, as from Thermocouple Pyro-meters]. -P. G. Weiller. (Rad. Engineering,April, 1933, Vol. 13, No. 4, pp. 16-17.)

August, 1933

THREE-PHASE TRANSFORMER CONNECTIONS ANDTHEIR APPLICATION TO HIGH -VOLTAGERECTIFYING CIRCUITS. -J. B. Epperson.(Rad. Engineering, May, 1933, Vol. 13,No. 5, pp. 14-16.)

TESTS ON THE NEW ARC -IN -AIR RECTIFIER FORVERY HIGH VOLTAGES AND POWERS. -E. Marx. (E.T.Z., 27th April, 1933, Vol. 54,No. 17, PP 396-397)

Highly successful tests of a large rectifier con-structed on the principles outlined in Marx's paper(1932 Abstracts, p. 597, r -h column).

SUR UN NOUVEAU DATECTEUR (A New Detector[Insensitive Galena Crystal treated withSulphur, with Copper Electrode]). -J. Cayrel.(Comptes Rendus, 24th April, 1933, Vol. 196,No. 17, pp. 1216-1218.)

The so-called " insensitive " galena crystals giverectification in the opposite sense to the " sensitive "crystals. If covered with quasi -colloidal sulphur(by electrolysis of sulphuretted hydrogen) andused with a noble metal as electrode, they giveabout the same order of rectification as before thesulphur treatment, but with much greater stability :with oxidisable electrodes they behave practicallyas they did before the treatment : but with copperelectrodes (which give poor results with the un-treated crystals) the rectification is so much im-proved that it surpasses, for small amplitudes, theresults given by the most sensitive galena detectors.The effect is due to the formation, at the contact,of copper sulphide. This cold sulphur treatment isquite different from Florisson's sulphur vapourtreatment, which converts " insensitive " into" sensitive " galena and thus changes the sense ofrectification.DYNAMOS FOR RADIO TRANSMITTING STATIONS. -

H. Hutt : Brown-Boveri Company. (Rev.Gen. de l'Elec., 15th April, 1933, Vol. 33,No. 15, p. 119 D : summary only.)

AN IMPORTANT IMPROVEMENT IN THE CONSTRUC-TION OF THE LEAD ACCUMULATOR [theFery-Carbone Non -Sulphating Cell]. -Ch.Fery. (Rev. GM. de l'Elec., 11th March,1933, Vol. 33, No. 10, pp. 313-318.) See1931 Abstracts, p. 453, r -h column.

IODIDE CELLS -AN INVESTIGATION OF CADMIUMIODIDE AND ZINC IODIDE TYPES. -A. E.Salazar. (Electrician, ,9th May, 1933, Vol.

io, No. 2868, p. 65o.)H.T. FOR THE CAR SET. -A. Dinsdale. (Wireless

World, ,9th May, 1933, Vol. 32, pp. 352-353)

The tuned vibratory transformer used duringthe war to obtain a high voltage from an 1.t. accumu-lator is well known. This article describes a greatlyimproved version of this device which has beendeveloped in the U.S.A. for supplying h.t. to aradio receiver installed in a car, using the existingcar accumulator as the source of 1.t.

THE DEVELOPMENT OF A SOUND -ISOLATING BASEFOR MOTORS. -E. H. Hull. (Gen. Elec.Review, May, 1933, Vol. 36, No. 5, pp. 223-227.)

August, 1933 460

THE USE OF TWO -GRID VALVES FOR THE AMPLI-FICATION OF CONTINUOUS CURRENTS. -P.Donzelot and J. Divoux. (Comptes Rendus,22nd May, 1933, Vol. 196, No. 21, pp. 1579-1581.)

Instability due to variations of the anode batteryvoltage has already been dealt with (Feb. Abstracts,p. 113, 1-h col.). The writers now show how theuse of two -grid valves overcomes the instabilitydue to variations in the filament and grid batteries.

THE AUTOMATIC TRANSMISSION BY RADIOTELE-GRAPHY OF THE READING OF AN INSTRUMENTBY TWO AUDIO -FREQUENCY MODULATIONSOF VARYING DEPTH .-Csejkovits andDieckmann. (See abstract under " Direc-tional Wireless.")

SPECIAL DESIGN OF H.F. COILS [for Master -Oscilla-tor] TO GIVE CONSTANT INDUCTANCE OVERTEMPERATURE RANGE OF 40° C.-Ure,Grainger and Cantelo. (See abstract under" Transmission.")

[Air -Cored] TRANSFORMERS WITH VARIABLE COUP-LING [Theoretical Investigation]. -Siegel.(See under " Properties of Circuits.")

CONTROLS FOR 1933 RECEIVERS [particularlyMoulded Carbon Variable Resistances andTheir Uses]. -N. E. Wunderlich. (Rad.Engineering, May, 1933, Vol. 13, No. 5,pp. 17-19.)

VDE REGULATIONS FOR CONDENSERS FOR BROAD-CAST RECEIVERS AND INTERFERENCE PRE-VENTI N.-Verband Deutscher Elektro-techniker. (E.T.Z., 18th May, 1933, Vol. 54,No. 20, pp. 484-485.)

RECENT DEVELOPMENTS IN MICA CO NDEN SERS.-A. E. Thiessen. (General Radio Experi-menter, Jan., 1933, Vol. 7, pp. 1-4.)

'CYBER EINEN PLATTENKONDEN SATO R MIT NEGATIVERKAPAZITAT (A Plate Condenser with Nega-tive Capacity). -H. Benndorf. (Zeitschr. f.Physik, 1933, Vol. 82, No. 5/6, pp. 397-403)

The writer has defined electrostatic capacity asC = deldV, where V is the potential difference ande the charge on the plates; he finds that it canbecome negative when the plates of a condenser arein elastic connection with one another.A NEW COMPRESSED INSULATING MATERIAL [Syn-

thetic Resin and Linen Aggregate]. -H.Rommler Company. (E.T.Z., 18th May,1933, Vol. 54, No. 20, pp. 478-479.)

" CALIT " AND " CALAN " TWO NEW CERAMICINSULATING MATERIALS. (Zeitschr. V.D.I.,6th May, 1933, Vol. 77, No. x8, pp. 483-484.)See also July Abstracts, p. 402, r -h column.

ON THE PERMEABILITY OF IRON AT ULTRA -RADIOFREQUENCIES. -W. Arkadiew. (Phys. Review,15th April, 1933, Series, 2, Vol. 43, No. 8,pp. 671-672.)

This note compares results due to Hoag and Jones(April Abstracts, p. 226, r -h col.) with previouswork of the writer. The frequencies used areabove 3 X ios c/s.


DAUERMAGNETE (Permanent Magnets). -W. Elen-baas. (Zeitschr. f. tech. Phys., No. 5, Vol. 14,1933, pp. 191-197.)

From the Philips' Company laboratories : anew treatment leading to simpler formulae thanthose usually given. The demagnetisation curveis replaced by an ellipse.VDE RULES FOR DESIGN AND TESTING OF SMALL

AND LOW -VOLTAGE TRANSFORMERS.-Ver-band Deutscher Elektrotechniker. (E.T.Z.,23rd March, 1933, Vol. 54, No. 12, pp. 293-294.) Continued : see June Abstracts, p.340, 1-h column.

CRITICAL REVIEWS OF BOOKS AND PAPERS ON ELEC-TRIC FILTERS. -P. David. (See under" Properties of Circuits.")

THE NEVSON TANGRAPH RULE .-(Journ. Scient.Instr., April, 1933, Vol. To, No. 4, pp.119-12o.)

A ROUND CHART INDICATING RE CORDER. -Leedsand Northrup Company. (Ibid., p. 119.)

A DRAWING INSTRUMENT FOR LARGE RADIUSARCS. -E. Simeon. (Journ. Scient. Instr.,May, 1933, Vol. Jo, No. 5, pp. 157-158.)

AN ELECTRICAL CALCULATING MACHINE . -R. R. M.Mallock. (Proc. Roy. Soc., May, 1933, Vol.140, No. A 841, PP. 457-483)

This machine was devised for solving sets oflinear algebraic simultaneous equations involvinga large number of unknowns ; it consists of a setof a.c. transformers each with a number of windings.The unknowns are proportional to the fluxes in thetransforthers and the coefficients to the numbers ofturns in the windings. The present model can dealwith to equations and the errors in the roots havebeen reduced to less than o.1 % of the largest root.

STATIONS, DESIGN AND OPERATIONINTERNATIONAL RADIOTELEGRAPHIC ANDTELEGRAPHIC CONFERENCE IN MADRID,1932: J Reyval. (Rev. Gin. de l'Elec.,20th May, 1933, Vol. 33, No. 2o, pp. 667-673 . )

INTERNATIONAL ORGANISATION OF RADIOCOMMUNICATIONS : ORGANISATIONS, CON-VENTIONS AND INTERNATIONAL REGULA-TIONS . -M. Adam. (Rev. Gin. de l'Elec.,3rd June, 1933, Vol. 33, No. 22, pp. 733-741.)

SIX MONTHS OF EMPIRE B ROAD CASTIN G. -N.Ashbridge. (World -Radio, 19th May, 1933,Vol. 16, No. 408, pp. 677-678 and 680.)

LA T.S.F. DANS L' AA RONALTTI QUE (Radio Com-munication in Aviation [a Survey]). -P.Brenot. (Rev. Gin. de l'Elec., 27th May,1933, Vol. 33, No. 21, pp. 704-710.)

TIME DELAY EFFECTS IN SYNCHRONOUS [COMM:LOB -Wave] BROADCASTING. -C. B. Aiken. (Elec-tronics, May, 1933, Vol. 6, No. 5, pp. 124-126.)

A short mathematical treatment followed by anaccount of experimental investigations, the firstof which was to confirm that when M =m and13 = o there is no distortion but only variations in

THE

THE


intensity of the resultant signal as K or y is varied(cf. July Abstracts, p. 392) ; the other experimentsyielded curves showing the " tolerable carrierratio " as a function of y and as a function of thetime delay when y is maintained within ± 1 % of180°. At this value of y (i.e., carriers in phase oppo-sition) the distortion is of a very unpleasant non-linear character, while when y is near o° the distortionis largely a matter of impaired fidelity and is notnearly so unpleasant. Other conclusions, in agree-ment with theory, are reached by these tests. Thusa relative time delay of only 5o microseconds (etherpath difference of 9.8 miles) between two otherwiseidentical transmissions will produce a distortion,when the carriers are out of phase, of sufficientmagnitude to require a carrier ratio of about 6 dbfor its elimination.

FIELD -STRENGTH CHARTS OF POSTE-PARISIEN ANDRADro-PARIS.--David. (See reference under" Propagation of Waves.")

FIVE METRE WORK FOR AMATEURS. -H. L. O'Hef-fernan and S. G. Morgan. (Wireless World,26th May, 1933, Vol. 32, pp. 368-369.)

GENERAL PHYSICAL ARTICLESELECTRICAL CONDUCTIVITY OF IONISED GAS IN

THE PRESENCE OF A MAGNETIC FIELD. -T. C. Cowling. (Sci. Abstracts, Sec.A, March1933, Vol. 36, No. 423, p. 289.)

THE FIELD DISTORTION OF A PLANE SPARK GAP,CROSSED AT CONSTANT VOLTAGE BY ANIONISING ELECTRON LAYER [TheoreticalInvestigation]. -J. J. Sammer. (Zeitschr. f.Physik, 1933, Vol. 81, No. 7/8, pp. 440-444.)

ON THE MOBILITY OF ELECTRICAL CHARGE CARRIERSUNDER THE SIMULTANEOUS INFLUENCE OFSTRONG ELECTRIC AND MAGNETIC FIELDS[Theoretical Investigation]. -G. A. Kugler,F. 011endorff and A. Roggendorf. (Zeitschr.f.Physik, 1933, Vol. 81, No. 11/12, pp. 733-744.)

HIGH -FREQUENCY ELECTRIC DISCHARGE IN GASES.-J. C. Wilson. (Nature, 15th April, 1933,Vol. 131, pp. 546-547.)

This letter describes curious streamer -like dis-charges obtained with a range of frequency fromIr. i to 21.4 Mc/s in the capacitive electrodelessdischarge.

MOVEMENT OF CHARGED PARTICLES IN A FIELDOF SUDDENLY CHANGING POTENTIAL. -E. Persico. (Sci. Abstracts, Sec.A, March,1933, Vol. 36, No. 423, p. 300.)

OBSERVATION WITH THE EYE OF A SCHROT EFFECTFOR PHOTONS. -R. R. Barnes and M. Czemy.(Sci. Abstracts, Sec.A, March, 1933, Vol. 36,No. 423, p. 273.)

ELECTRON THEORY OF METALS [MathematicalPaper on Present State : includingThermionic and " Cold ". Emission]. -R.Peierls. (Sci. Abstracts, Sec.A, Feb., 1933,Vol. 36, No. 422, p. 177.)

August, 1933

THE QUANTUM THEORY OF SEMI -CONDUCTORS [aSurvey]. -E. L. Hill. (Review Scient.Instr., April, 1933, Vol. 4, No. 4, pp. 186-188.)

ON THE [Quantum Mechanics] THEORY OF ELEC-TRONIC SEMI -CONDUCTORS. -M. Bronstein.(Physik. Zeitschr. der Sowjetunion, No. r,Vol. 2, 1932, pp. 28-45.)

COURSE OF CURRENT AND VOLTAGE IN INSULATINGMATERIALS WITH ONE OR SEVERAL MOVABLETHIN LAYERS OF CHARGE. -W. 0. Schumann.(Archly f. Elektrot., 25th March and lothApril, 1933, Vol. 27, Nos. 3 and 4, pp. 155-168and 241-253.) See also May Abstracts, p.282, 1-h column, 3 references.

POSITIVE AND NEGATIVE ELECTRONS APPARENTLYPRODUCED IN PAIRS [Support to Millikart'sPhoton Theory of Cosmic Rays]. --R. M.Langer : Anderson. (Sci. News Letter, 27thMay, 1933, Vol. 23, No. 633, p. 332.)

THE INTERACTION OF CHARGES IN DIRAC'S THEORY[Interaction of Two Electrons in ThreeDimensions]. -K. Nikolsky. (Ibid., No. 6,Vol. 2, 1932, pp. 447-452.)

MISCELLANEOUSLAMA, MATHIEU AND RELATED FUNCTIONS IN

PHYSICS AND ENGINEERING [including Pro-pagation of Electromagnetic and AcousticWaves]. -M. J. 0. Strutt. (Book Review inL'Onde Elec., Feb., 1933 [publication de-layed], VOL 12, No. 134, p. 6A.)

TANGENTIAL POLAR COORDINATES OR POLAR CO-ORDINATES, AND THEIR APPLICATIONS [TOProblems of Vibratory Motion in Acoustics,Electricity, etc.]. -M. Jacob. (Rev. Gen. del'Elec., ist April, 1933, Vol. 33, No. 13, pp.403-411.)

SUCCESSIVE APPROXIMATIONS [to the Solution ofLinear Differential Equations] BY THERAYLEIGH-RITZ VARIATION METHOD. -J. K.L. MacDonald. (Phys. Review, 15th May,1933, Series 2, VOL 43, No. io, pp. 83o-833.)

A SIMPLE METHOD FOR THE NUMERICAL SOLUTIONOF DIFFERENTIAL EQUATIONS : NOTE ONERROR AND ITS AVOIDANCE. -W. G. Bickley.(Phil. Mag., June, 1933, Series 7, Vol. 15,No. 102, pp. 1174-1177.)

KURVENTAFEL ZUM AUSWERTEN VON MESSUNGENKOMPLEXER GRoSSEN (Curve Sheet forWorking out Measurements of ComplexQuantities). -P. C. Hermann. (Archiv f.Elektrot., loth April, 1933, Vol. 27, No. 4,pp . 283-286.)

The sheet of curves given enables the sine orcosine of the angle between two vectors to be read iftheir components are given.

ELECTRO-PHYSICS [Review of Progress]. -N. R.Campbell. (Journ. I.E.E., Feb., 1933,Vol. 72, No. 434, pp. 153-159.)

High-speed particles : thermionic amplifiers :

'August, 1933

photoelectric cells : electron microscope : quantumtheory (and the return of intuitive reasoning, as inGurney's application to the phenomena at anelectrolytic cathode, using only the simplest mathe-matics) : electron theory of conduction : ther-mionics and photoelectricity : radioactivity : dis-charge through gases (" the conception of positiveions causing the emission of electrons from thecathode ' by impact ' has become meaningless ") :penetrating rays (" the possibility of a terrestrialsource, once almost abandoned, is now being re-considered. It is at least certain that the far-reaching speculations on cosmology and the originof matter, to which one phase of the study of theserays gave rise, are quite unfounded.")

THE PRINCIPLES OF ELECTROMAGNETISM. -E. B.Moullin : G.W.O.H. (Book Review inWireless Engineer, Feb., 1933, Vol. so, No.113, pp. 61-64.) See also ibid., March, 1933,pp. 146-147, for a letter from Fortescue, andApril, pp. 179-182 and 202-204, for replies.


ON WRITING TECHNICAL ARTICLES. -J. Williamson.(British Radio Annual, 1932, pp. 64-74.)

In the section on " the English " of an article, thewriter says : " For accuracy, continually be tryingto read wrong meanings, even foolish or pig-headedones, out of everything you write ; and change thewording or the sentence structure until those arerooted out. . . . Recognise punctuation as a neces-sary part of sentence structure, rather alike to thebracketing of an algebraic expression. It dependsless on breath -taking than on logic."

RADIO AND ELECTRICAL NOMENCLATURE [a Pleafor a General " Cleaning Up "].-M. G.Scroggie. (Wireless Engineer, March, 1933,Vol. so, No. 114, pp. 145-146.)

RADIO ENGINEERING HANDBOOK. -Edited by K.Henney. (Book Review in Electronics,March, 1933, p. 77.)

THE PRODUCTION OF 4 800 000 VOLT HYDROGENIONS. -M. S. Livingston and E. 0. Lawrence.(Phys. Review, 1st Feb., 1933, Series 2,Vol. 43, No. 3, p. 212.)

THE ELECTROSTATIC PRODUCTION OF HIGH VOLTAGEFOR NUCLEAR INVESTIGATIONS. -R. J. Vande Graaff, K. T. Compton and L. C. Van Atta.(Phys. Review, ist Feb., 1933, Series 2,Vol. 43, No. 3, PP 149-157.)

MUTUAL IMPEDANCE OF GROUNDED WIRES FORHORIZONTALLY STRATIFIED TWO -LAYEREARTH. -J. Riordan and E. D. Sunde.(Bell S. Tech. Journ., April, 1933, Vol. I2,No. 2, pp. 162-177.)

The full paper, a preliminary account of whichwas referred to in 1931 Abstracts, p. 457, r -hcolumn.

MUTUAL INDUCTION BETWEEN POWER AND TELE-PHONE LINES UNDER TRANSIENT CONDITIONS.-W. G. Radley and H. J. Josephs. (Journ.I.E.E., March, 1933, Vol. 72, No. 435, pp.259-264.)

CONTROL OF WOOD -PULP COOKING BY PHOTOCELL.-Navarre Paper Mills. (Electronics, March,1933, P. 75.)

MEASUREMENT OF VERY WEAK LUMINOUSFLUX BY MEANS OF A PHOTOELECTRIC CELL[Increased Sensitivity by using a VacuumTube with Filament and Two HighlyInsulated Electrodes as a Variable HighResistance]. -E. Gambetta. (ComptesRendus, 27th March, 1933, Vol. 196, No. 13,PP. 906-908.)

THE

THE LIMITING MAGNITUDE OBSERVABLE WITH APHOTOELECTRIC STELLAR PHOTOMETER. THEAPPLICATION OF A THERMIONIC AMPLIFIERTO THE PHOTOMETRY OF STARS. -SinclairSmith : A. E. Whitford. (Summaries inReview Scient. Instr., April, 1933, Vol. 4,No. 4, p. 240.)

A NEW SYNCHRONISING EQUIPMENT FOR HANDOR AUTOMATIC CONTROL [for H.T. Networks :the Synchronoscope, using a PhotoelectricCell]. -H. Heyne. (E.T.Z., 6th April, 1933,Vol. 54, No. 14, PP 321-324.)

EXACT MEASUREMENT OF PHOTOGRAPHICDENSITIES [by Photoelectric Methods]. -P. Fleury and G. A. Boutry. (ComptesRendus, 3rd April, 1933, Vol. 196, No. 14,pp. 1013-1015.)

A PRECISION DENSITOMETER USING A PHOTO-ELECTRIC CELL. -G. A. Boutry. (Ibid.,loth April, 1933, Vol. 196, No. 15, pp. Ica -z Ica.)

A PHOTOELECTRIC INSTRUMENT FOR COMPARINGTHE STRENGTHS OF COLOURED SOLUTIONS.-E. W. H. Selwyn : Kodak Company.(Journ. Scient. Instr., April, 1933, Vol. 1o,No. 4, pp. 116-118.)

SONIC NEPHELOMETER [for Comparison of Lighttransmitted by Translucent Materials, Photo-cell and Loud Speaker being used to detect" Flicker "].-J. S. Wilson. (Journ. Scient.Instr., April, 1933, Vol. 1o, No. 4, pp. 97-105.)

A PRACTICAL PHOTOELECTRIC REFLECTION METER[using the Bergmann Selenium Barrier -Layer Photocell]. -L. Bergmann. (Zeitschr.f. tech. Phys., No. 4, Vol. 14, 1933, pp. 157-158.)

THE

REMOTE INDICATION AND CONTROL DEVICES. -G. F. Tagg. (Journ. Scient. Instr., March, A1933, Vol. io, pp. 65-71.)

THE MEASUREMENT OF TELEGRAPH DISTORTION. -V. J. Terry. (Elec. Communication, April,1933, Vol. I1, No. 4, pp. 197-203.)

DETERMINING THE TRANSMISSION EFFICIENCY OFTELEGRAPH CIRCUITS. -E. H. Jolley. (P.O.Elec. Eng. Journ., April, 1933, Vol. 26, PartI, pp. 1-7.)


PHOTOELECTRIC PHOTOMETER TESTS AIR PURITY.-(Electronics, Feb., 1933, p. 45.)

PHOTOCELL SIGNALS WHEN COFFEE IS ROASTED TOCORRECT SHADE.-(Ibid., p. 45.)

ROTATIONS OF FRICTIONLESS STEAM OR GAS METERCOUNTED BY PHOTOELECTRIC CELL. -(Elec-tronics, Feb., 1933, p. 44.)

PHOTOELECTRIC CONTROL OF PAPER CUTTER FORPAPER BAG MAKING. -D. R. Shoults. (Elec.World, 22nd October, 1932, Vol. I 00 , pp.562-563.)

PHOTO -ELECTRIC CELLS: THEIR PROPERTIES ANDUSES. II. SOME APPLICATIONS. -L. G.Stoodley. (World Power, May, 1933, Vol.19, No. 133, pp. 287-292.) See June Ab-stracts, p. 344, r -h column, for Part I.

REPRODUCTION OF PRINT FOR THE BLIND [FivePoints of Light reflected successively on toPhotocell and actuating Solenoids con-trolling Five Points touching the Finger]. -F. Allen. (National Research Council,Canada, Fifteenth Annual Report, 1931-1932, p. 229.)

OPTICAL TELEPHONY TO A SHIP BELOW THE HORI-ZON -(Electronics, April, 1933, Vol. 6,No. 4, p. 87 : News item.)

" A great Atlantic liner talks with a ship belowthe horizon, by means of a modulated light -beamplaying on a distant cloud -bank."

LOUD SPEAKERS SUMMON PHYSICIANS IN THENEW YORK HOSPITAL [" Doctors' PagingService "].-(Bell Lab. Record, April, 1933,Vol. r1, No. 8, pp. 241-243)

AN ELECTRICAL CALCULATING MACHINE [for solvingSets of Linear Simultaneous Equationsinvolving a Large Number of Unknowns].-Mallock. (See under " Subsidiary Apparatusand Materials.")

SUGGESTED UNIT OF VIBRATION AMPLITUDE [the" Pal," corresponding to the Phon inAcoustics]. -W. Zeller. (Zeitschr. V.D.I.,25th March, 1933, Vol. 77, No. 12, p. 323.)

OPTICAL MANOGRAPH WITH LOW INERTIA [forStudy of Internal Combustion Motors]. -M. Serruys. (Comptes Rendus, loth April,1933, Vol. 196, No. 15, pp. 1083-1085.)

[Variation of] CONDUCTIVITY OF NACL CRYSTALSUNDER PRESSURE . -Z. Gyulai. (Sci. Ab-stracts, Sec. A, Feb., 1933, Vol. 36, No. 422,p. 178.)

AN AMPLIFYING LEVER FOR THE MEASUREMENTOF SMALL DISPLACEMENTS [Such as theDeformations of Piezoelectric Plates]. -G. A. Fink and G. W. Fox. (Review Scient.Instr., May, 1933, Vol. 4, No. 5, pp. 276-279.)

August, 1933

PIEZOELECTRIC MEASUREMENT OF PRESSURES [e.gin Large Guns, or Very Low Pressures usuallymeasured by McLeod Gauge] AND WEIGHTS[Deflections of One Millionth of an Incheasily Measured]. -H. C. Weber. (Elec-tronics, April, 1933, Vol. 6, No. 4, p. 103.)

SHORT-WAVE TREATMENT OF NERVOUS PARALYSIS.-D. Kellner. (World -Radio, 21st April,1933, Vol. x6, No. 404, p. 539.)

EXPERIMENTAL RESEARCHES ON THE EFFECTS OFULTRA -SHORT WAVES ON THE FUNCTIONINGOF THE CEREBELLUM. -P. A. Gemelli. (LaRicerca Scient., 31st March, 1933, pp.373-374.)

MEASUREMENTS OF THE TEMPERATURE IN THEINTERIOR OF AN EGG IN THE ELECTRICFIELD OF ULTRA -SHORT WAVES [Supportof Theory of Specific Action as opposed toPurely Thermal]. --A. Jellinek. (ComptesRendus, loth April, 2933, Vol. 196, No. 15,pp 1149-1150.) Continuation of the workreferred to in 1931 Abstracts, p. 169, r -hcolumn.

KILLING WEEVILS IN WHEAT BY [Ultra-] SHORTWAVE . -Westinghouse Company. (Elec-tronics, March, 2933, p. 75.) See alsoFebruary Abstracts, p. 117, left-handcolumn, and below.

[Ultra -Short] RADIO WAVES KILL INSECT PESTS[in Wheat, Seeds, Tobacco, Vegetables,etc.]. -J. H. Davis. (Scientific American,May, 1933, Vol. 148, No. 5, pp. 272-273.)

Using 20 kw 6 -metre waves, art exposure of 6seconds was sufficient to exterminate eggs, larvaeand adult insects in wheat, with no damage to thewheat itself. " The plant has been in operationfor a year and many kinds of infested materialshave been successfully treated . ." See alsopreceding abstract.

RADIOACTIVE PHENOMENA OF THE SECOND ORDERAND OF ARTIFICIAL ORIGIN [Highly Ab-sorbable Radiations from " ResistanceCells." Wavelengths between Some Dozensand Some Hundreds of Angstrom Units :Electrons of the Peripheral Orbits of theAtom involved]. -G. Reboul. (spurn. dePhys. et le Rad., Feb., 1933, Vol. 4, Series 7,No. 2, pp. 73-89.) Continuation of the workreferred to in 1932 Abstracts, p. 541, r -hcolumn.

ON THE PROBABLE EMISSION OF A RADIATION OFSMALL PENETRATING POWER FROM CERTAINMETALS. -J. Reboul. (Comptes Rendus,22nd May, 1933, Vol. 196, No. 2 I , pp.1596-1598.)

THE E.M.Fs DEVELOPED BY MAN IN CONTACT WITHA METALLIC CONDUCTOR [and Their DiurnalVariation]. -F. Vies and others (ComptesRendus, 27th March, 1933, Vol. 196, No. 13,pp. 965-967.) For another paper by Vieson an allied subject, see March Abstracts,p. 174, r -h column.


Some Recent PatentsThe 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.

COLOUR TELEVISIONConvention date (Germany), 3oth May, 1931.

No. 387206The viewing -screen of a cathode-ray tube is

coated with three differently -coloured fluorescentsalts, for instance red, blue, and green in themanner of a lineated grid, one fundamental colourbeing allotted to each line in succession. Thescanning ray may be kept horizontal and made totravel three times faster than is normally necessary,or it may be applied transversely to the grid andthe intensity -variation trebled.

Patent issued to Siemens Schuckertwerke A.G.

AERIALSApplication date, 9th June, 1932. No. 387215.An indoor aerial is made by enclosing a metal

wire or ribbon between two strips of insulatingmaterial, the combination being coated on one sidewith a non-drying adhesive, so that it can be appliedto the wall of a room. The other side of the stripis suitably coloured or decorated to simulate theeffect of a dado. Tags or eyelets are provided totake the connecting -lead to the wireless set.

Patent issued to C. S. Garland.

LOUD SPEAKERSApplication date, loth May, 1931. No. 385603The resonance frequency of the movement of a

loud speaker is varied by adjusting the geometricalrelations of the magnetic circuit. For instance in

No. 385603

an " inductor " type of instrument the pole -pieces P are splii as at S and Si so that the distanceY between their centres can be varied. Similarly

the distance X between the masses M, MI ischanged by sliding them along the armature bar.This also alters the effective gap between the opposedpoles.

Patent issued to The General Electric Co., Ltd.

SIGNALLING SYSTEMS

Convention date (U.S.A.), 2nd September, 1930.No. 386356

The invention utilises the fact that when anegatively charged electrode is placed in a stronglyionised gas at low pressure it becomes covered with

No. 386356.

a sheath of positive ions. The electrode and sheathin effect form a condenser the capacity value ofwhich is, however, independent of the relativeposition of the opposed " plates," though it can becontrolled by varying either the degree of ionisationof the gas, or the potential of the electrode withrespect to the gas. As shown the valve V is filledwith mercury vapour, and the grid is negativelybiased, so that it acquires a sheath of positive ionsin the manner described. The valve is insertedin shunt with the main oscillatory circuit LC of aconstant -frequency generator 0 stabilised by aquartz crystal Q. Modulating signals from amicrophone M are applied between the grid andfilament of the valve V, and alter the effectivecapacity of the sheath by varying the voltage acrossit. This in turn causes a frequency -modulation ofthe outgoing carrier wave from the generator 0.

Patent issued to British Thomson Houston Co.,Ltd.


PICTURE -TRANSMISSION SYSTEMSConvention date (France), 3oth June,

No. 385536Relates to systems of the kind in which the

different light -and -shade values of the picture arefirst translated into signals of constant intensity

465

1931.

No. 385536.

which vary in length so that the time factor controlsthe intensity of light reproduced at the receivingend. Any " fading " that may occur in transitdoes not then produce " differential " effects inreception. As shown, light from a constant sourceS passes through an optical system 0 on to a mirrorM, which is vibrated by a current depending uponthe light -and -shade values of the picture. Thisthrows the transmitted ray up and down relativelyto a screen P having a flared aperture and anassociated disc D with radial slots. When the rayfollows the upper lateral track marked I only ashort impulse reaches the photo -electric cell C,whereas when it follows the lower track 2 the im-pulse is more prolonged.

Patent issued to Etablissements E. Belin.

GRAMOPHONE PICK-UPSConvention date (Switzerland), zznd April, 1931.

No. 384531In order to secure accurate tracking along the

whole face of the record, the axis of rotation ofthe needle is inclined at an angle instead of beingperpendicular to the plane of the magnetic flux.Instead of inclining the whole of the pick-up, theair -gap is cut at an angle with the pole -pieces, andthe armature (which is of square section) is fittedinside and packed with rubber. The arrangementallows the pick-up to be conveniently replaced bya sound -box without requiring any modificationof the usual opening in the sound -arm.

Patent issued to E. Paillard et Cie Soc. Anon.

Application date, 4th November, 1931. No. 385702To minimise distortion arising from (a) mech-

anical inertia, (b) the electric " lag " due to the coil,and (c) the uneven magnetic distribution from thepole -pieces, the upper end of the pick-up needle does

August, 1933

not carry the usual winding but extends inside aclosed vessel containing a conducting fluid. Hereit vibrates between two adjacent electrodes so asto alter the effective resistance of a circuit inaccordance with the sound -track variations on therecord.

Patent issued to A. M. Low.

Application date, 4th May, 1932. No. 386507The needle N is mounted in a block B on the

lower pole -piece P, and the armature A is in theform of a light, stiff lever extending at right -anglesand having an end -piece Ai which vibrates betweenthe bifurcated parts of the upper pole -piece Ptto vary the flux passing between them. The arma-ture is pivoted at M, a spring S giving the requiredrestoring force. When in use the needle is heldin position by the normal pressure from the record,no clamping -screw being necessary. When thepick-up is lifted off the record, the needle is stillretained in situ by magnetic attraction. Thearmature is so light that its natural frequency liesabove the useful audible range.

Patent issued to The General Electric Co., Ltd.,and F. H. Brittain.

No. 386507.

TELEVISION SYSTEMSConvention date (U.S.A.), 24th April, 1931.

No. 386183The incoming signals are recorded by means of a

spark discharge which etches a permanent image onthe prepared surface of a photographic film. Afterthe film has been suitably fixed, it is passed before aprojector in the ordinary way and reproduced on alarge screen. The method of recording removessome of the present limitations as regards illumina-tion and therefore enlarges the size of televisedscenes. Means are provided for compensatingfor the " lag " introduced by the necessary treat-ment of the film, so as to allow " sound " effects tobe accurately synchronised.

Patent issued to American Television Labora-tories Inc.


CATHODE-RAY TUBESConvention date (Germany), August 21st, 1931.

No. 386706The filament of a cathode-ray tube, when heated

directly by alternating current, creates a magneticfield which tends to deflect the ray out of thestraight. To prevent this the filament is twistedinto the shape of a figure eight, open at the bottom ;or it may be formed as two co -axial coils oppositelywound ; the object in both cases being to reduce theresultant field to zero.

Patent issued to Lignose Hiirfilm System Breu-sing G.m.b.h.

THERMIONIC AMPLIFIERSConvention date (U.S.A.), 21st May, 1931.

No. 386500When an amplifier is operated so that the output

is more than say 5o% of its commercial rating, itdevelops a " drooping " characteristic (causedchiefly by the falling -off in effective resistance dueto grid rectification) and a consequent tendency todistortion. According to the invention an aux-iliary valve is used to feed additional energy to thecommon output circuit, as and when required torestore a strictly straight-line response. TheFigure shows the arrangement applied to a trans-mitting arrangement comprising a high -frequencysource S feeding two S.G. amplifiers V, VI. Modu-lating signals are applied from a microphone Mthrough an amplifier A. The screening grid of thecompensating valve VI is biased through a potentio-meter P, so that it does not come into operation

No. 386500.

until the output from the valve V commences to" droop," whereupon it contributes the requiredextra energy to the common output circuit LCfeeding the power amplifier Ai.

Patent issued to Marconis Wireless Telegraph Co.,Ltd.

REDUCING INTERFERENCE.Application date, 21st October, r931. No. 38705oIn order to cut out a definite disturbing fre-

quency, such as an interfering station on the high -frequency side, or the audio -frequency of a Morse

No. 38705o.

transmitter or a heterodyne whistle on the low -frequency side, a rejector circuit L, C is insertedin the cathode lead of an indirectly -heated valve V,shown in the Figure as a low -frequency amplifier.A resistance R supplies grid bias.

Patent issued to W. T. Sanderson and R. Black-burn.

CATHODE-RAY TELEVISIONConvention date (Germany), 27th March, 1931.

No. 387536For scanning, two " tilting " or quick -return

oscillations of constant amplitude are applied tothe deflecting plates of a cathode-ray tube. Theviewing screen at the end of the tube is formed bya coating of potassium silicate, or water -glass,sprayed evenly with cadmium tungstate. This isstated to give a point intensity of illumination offrom 2 to 4 candle power, with practically no" time-lag." In transmission the luminous areaproduced on the fluorescent screen is concentratedby a lens system on the object to be televised, thereflected ray (or the transmitted ray in the case ofa photographic film) being directed on to a numberof photo -electric cells.

Patent issued to M. von Ardenne.

BROADCAST RELAY SYSTEMSApplication date, 3rd December, 1931. No. 386764

Relates to wired -wireless relays for redistributingBroadcast programmes in low -frequency form.In order to shut out extraneous disturbances, due toelectric trolley buses, etc., chokes are inserted inseries in the distributing network. Each choke con-sists of a pair of tightly coupled coils, preferablywound together of bifilar wire, so that the flux ispractically neutralised for signal currents whichflow through the coils in series ; the chokes howeveroffer a high impedance to disturbing currents sincethese are usually induced in parallel.

Patent issued to Siemens Bros. & Co., Ltd.

EXPERIMENTAL WIRELESS A5 August, 1933

for the PRODUCTIONTESTING of

INDUCTANCESMODERN

radio assemblydemands the most

accurate matching ofinductances.

The MUIRHEAD Inductometeris a precision instrumentrobustly built for generallaboratory and works use.The MUIRHEAD Inductometeris fully described in Bulletin283-B.

MUI RHEADMAKERS of PRECISION INSTRUMENTS for OVER 50 YEARS

ELMERS END KENTKindly mention The Wireless Engineer- when replying to advertisers.

August, 1933 A 6 THE WIRELESS ENGINEER &

PAPER CONDENSERS. TERMINAL TYPES

Mfd. Type 50:61 Type 80/81 Type 101 Type 121

0.10.250.51

23456

10

s. d.--2 42 63 65 05 67 38 6

14 0

s. d.2 02 42 63 04 06 07 09 0

10 617 6

s. d.--5 06 09 0-

17 622 025 0-

s. d.--7 08 6

13 0-25 031 037 6-

PAPER CONDENSERS. SOLDERING TAG TYPES

Mfd. Type 65 Type 84 1 Type 87

0.10.250.51

234568

10

s. d.1 81101112 02 8-5 0-7 09 0

11 6

s. d.2 02 22 42 93 9-6 9-

10 013 016 0

s. d.2 22 42 63 04 0-7 3----

MICA CONDENSERS

Mfd. Type M S.P. Type Type 34

.00005

.0001/3

.0004,5

.001,'4

.005/6

.01

s. d.0 80 80 91 01 62 0

s. d.-2 02 02 63 0-

s. d.1 31 31 31 62 03 0

ELECTROLYTIC CONDENSERS

Mfd.Type 802Aqueous

Type 801Aqueous

Type 902Dry

847

s. d.6 05 0-

s. d.--6 0

s. d.6 6--

Here are the current prices of some of the more popular of theT.C.C. Condenser Range. Ask for complete list.

T.C.C.ALL- BRITISH

CONDENSERS

SAFETYcosts but little extraIT is worth something to know that the condensers

you use will function as "per specification "-thatthey will not fail and jeopardise other components incircuit. That is to know true safety-and T.C.C.Condensers are your safeguard. Twenty-seven yearsreputation is at stake with every T.C.C. sold, andrepeated test assures that the reputation is maintained

With their cost so very little more it is worthbuying T.C.C. and to be sure.

See the Complete Rangeon Stand 98 Radiolympia

The Telegraph Condenser Co. Limited, Wales Farm Rd., N. Acton,London, W.3.

Here is illustrated Type 141 Condenserfor maximum working voltage of 2,500peak value (tested 5,000 V.D.C.) Price28/-. Made also in 2 and 4 mfds.Prices 50/- and 97/6 respectively.

Kindly mention " The Wireless Engineer" when replying to advertisers.0 3102

EXPERIMENTAL WIRELESS iii August, 1933

ADIO DATA CA SERIES OF ABACS

ARTSproviding most of the essential Data required in Receiver Design.

By R. T. BEATTY, M.A., B.E., D.Sc.Reprinted from " The Wireless World."

(3930)

" Radio Data Charts" provide designers of wireless apparatus with aready and convenient means of solving problems without havingrecourse to complicated formula and mathematics.

By the use of the charts it is possible to tackle all the more familiarproblems in radio receiver design ; such as, for example, finding therelationship between inductance capacity and frequency, and workingout the design of high frequency transformers. All keen experimenterswill appreciate this helpful book.

Price 4/6 net. By post 4/10.(39 CHARTS and more than 5o Diagrams.)

From all leading booksellers or direct from the Publishers.Published from the Offices of " THE WIRELESS WORLD."

ILIFFE & SONS LTD., Dorset House, Stamford St., London, S.E.1.

Publishers'

BINDINGCASES

in Grey Clothfor Vol. IX of

TheWIRELESSENGINEERPrice 3'- By post 3'3

[ Volume IX, Bound Complete, 1

£2. By post £2 :1 : 0Earlier Volumes can besupplied. Writefor

particulars to the publishers.

ILIFFE & SONS LIMITED,Dorset House, Stamford Street, London, S.E.x

FIFTY YEARS ofELECTRICITYMemories of an Electrical Engineer

By Sir Ambrose FLEMING, M.A.,D.Sc., F.R.S.

(1921)The name of Sir Ambrose Fleming givesto this book an authority which thoseinterested in the subject will appreciate.The author has a word to say upon everyphase of applied electricity, and besidesgiving a critical account of the activitiesof the past fifty years, he indicates thetrend of present research and suggests thecourse of future progress and development.Every student should make a point ofobtaining a copy of what may be regardedas a classic of technical literature.

z

F1F TY YEARScoy Crown 4to. 37x Pages. in Plates.

Cloth Bound.f ELECTR1... ----'

PRICE 151- NETBy Post 15/9

From all leading Booksellers or direct.... from the Publishers:

ILIFFE & SONS LIMITED,Dorset House, Stamford St.,

LONDON, S.E.I. W.E,6.

Kindly mention " The Wireless Engineer" when replying to advertisers.

111

August, 1933 iv THE WIRELESS ENGINEER1

"THE WIRELESS WORLD"The weekly journal which keeps its readersthoroughly in touch with all that is happeningin wireless. It covers every wireless interestand appeals to every wireless amateur.

It is in every sense a wireless newspaper, forit records and reviews prevailing theory andcurrent practice in design, and also gives,week by week, broadcasting and othergeneral news.Special features are : Theoretical Articlesdiscussing new principles: Detailed TechnicalReviews of Commercial Radio Receivers :Designs for Experimental Receivers speciallyprepared to illustrate the practical applica-tion of new principles in valves, components,or circuit arrangements: The completeweekly programmes of all ForeignStations.

WirelessWorld

EVERY FRIDAY 4dSubscriptions:

Home & Canada, £1.1.8, other CountriesAbroad, £1.3.10 per annum, post tree.

ILIFFE & SONS LIMITED DORSET HOUSE STAMFORD STREET LONDON S.E.t

Printed for the Publishers, ILIFFE & SONS LTD., Dorset House, Stamford Street, London, S.E.i,by The Cornwall Press Ltd., 1-6, Paris Garden, Stamford Street, London. 8.E.1.

Sole Agents for Australasia-Gordon & Gotch (Australasia), Ltd..Sole Agents for South Africa-Central News Agency, Ltd.

worldradiohistory.com · I X X X X X K X X XX X X X X O August, 1933 SERADEX RESISTORS" Wireless...

Documents

Transcript of worldradiohistory.com · I X X X X X K X X XX X X X X O August, 1933 SERADEX RESISTORS" Wireless...