THE WIRELESS ENGINEER

THE

WIRELESSENGINEER

ANDEXPERIMENTAL WIRELESS

NUMBER 121 VOLUME X OCTOBER 1933

A JOURNAL OF

RADIO RESEARCHAND

PROGRESS

2EyN6,,,,

PUBLISHED BY

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

October, 1933 ii THE WIRELESS ENGINEER &

RADIIDATA CHA TSA SERIES OF ABACS

providing most of the essential Data required in Receiver Design.By R. T. BEATTY, M.A., B.E., D.Sc.

Reprinted from "The Wireless World."(1930)

" Radio Data Charts" provide designers of wireless apparatus with aready and convenient means of solving problems without havingrecourse to complicated formulae 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.

THE ELEMENTARY PRINCIPLESOF WIRELESS TELEGRAPHY

AND TELEPHONYBy R. D. BANGAY VThird Edition, Revised by 0. F. Baowx, B.Sc.

ICIOR many years this volume has been the standard book of instruction Vr for wireless beginners and students. The progress of wireless during recent vb

years has made necessary a Third Edition, which covers the wholesubject in a clear and simple style. Attention is given to the theoreticalelements of electricity and magnetism, the dynamo and the properties Vof waves.

THIRD Many new chapters have been added and descriptions of circuits havealso been included.

EDITION A leaflet giving full particulars of the volume and a synopsis of chapterswill be sent on request.

Price 7/6 net. By post 8/ -From all leading Booksellers or direct from the Publishers :

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

VV

VV

Kindly mention The Wireless Engineer" when replying to advertisers.

EXPERIMENTAL WIRELESS October 1933

10,000 OHMS TO

10,000 MECOHMSWITH SPEED

AND GREAT ACCURACY.

TYPE 544-AMEGOHM-METER.

NET PRICE

000

E57:10:0CARRIAGE PAID U.K.

DELIVERY EX STOCK.

DIRECT READING.ENTIRELY SELF-CONTAINED.VERY PORTABLE.This is one of a number of new Instruments which we

have recently introduced. This new group of apparatusincludes Type 636-A Wave -Analyser : Type 546-AAudio -Frequency Microvolter : Type 635-A, CheapCathode -Ray Oscillograph Outfit : Type 632-A, Capacity -Bridge for measuring Electrolytic Condensers.

You can easily keep in touch with developments byrequesting us to add your name and address to ourbi-monthly mailing lists. This entails no expense orobligation to make purchases. Such applications arebest directed to our London Offices.

DOES NOTREQUIRE ANY

TECHNICAL EXPERIENCE.

The Type 544-A Megohm-Meter is a Vacuum -TubeBridge Ohmmeter. It is so sensitive that the null -indicator is a high-grade pointer galvanometer. Themain dial is logarithmically scaled and by means of fivemultiplying factors this gives a range of six decades,covering from o.ox Megohm to to,000 Megohms. Ithas a higher range and superior " scale" (44 -inches)than any other form of " Megger," is self-containedand does not depend upon any hand -driven generator.Readings are permanent, not momentary, and veryaccurate.

CLAUDE LYONS, LTD.Head Offices : 76, OLDHALL STREET, LIVERPOOL, 3.LONDON: 40, BUCKINGHAM GATE, WESTMINSTER, S.W.1.

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

October, 1933 112 THE WIRELESS ENGINEER

SULLIVAN - GRIFFITHS INDUCTANCESare essential as Standards and as components inBeat Tone Oscillators, Wavemeters, Generators and, infact, in all circuits where frequency stability is required.

1 iuHto

1 HenrySTANDARDS from £6 each

THE MOST STABLEYET PRODUCED

for proof of this see pages241 and 242 of the Mayissue of this journal.

TEMPERATURECOEFFICIENTLESS

PATENTNo. 326190

SECOND GRADE STANDARDS10 /LH to 30,000 /41-1LOW H.F. LOSSES E1 -5 -0 each.

H. W. SULLIVAN, LTD.,Leo Street, London, S.E.15.

WORLD WIDE SERVICEFactories in London, Erie, Pennsylvania, and Toronto,

Canada, make and distribute Erie Resistors throughoutthe world, a service that has been built on a policy

of precision in manufacturing resistors exclusivelyplus a continual effort to produce and perfect resistors

to meet every exacting purpose and demand. Carefullaboratory tests, watching and checking every manu-

facturing stage, have produced a range of resistors witha record for the smallest percentage of rejects in the

radio industry, and they merit your attention and carefulconsideration. May we send you samples and prices?

FACTORIES at LONDON,ERIE, PENNSYLVANIA &

TORONTO, CANADA ERIE RESISTORSErie Resistor Ltd., Waterloo Rd., Cricklewood, N. W .2. Gladstone 2582

READY SHORTLYMake sure of your copy

This handy little pocket volume is now in its tenth yearof publication. It is a convenient source of informa-tion on different aspects of wireless theory and practice,containing many pages of facts, figures and explanationsto which wireless amateurs make frequent reference.

Bound in leather cloth, pencil inback loop, round corners, gilt edges.

Size 3i by 4} inches.

The

WirelessPRACTL RADIOW

JOICAURNAL orld

DIARYfor 1934

Price 1/6 By Post 1/7From all Booksellers, Stationers and Bookstalls, or direct from the Publishers

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

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

Al)

The

WIRELESS ENGINEERand EXPERIMENTAL WIRELESS

A Journal of Radio Research & ProgressEditor Technical EditorHUGH S. POCOCK Prof. G. W. 0. HOWE D.Sc., M.I.E.E.

VOL. X. No. 121

OCTOBER 1933

CON T EN T S

EDITORIAL 525

APPLICATIONS OF THE DYNATRON By M. G. Scroggie, B.Sc.,A.M.I.E.E. 527

S.G. VALVE AS SUPERHET DETECTOR By C. B. Fisher 541

OLYMPIA, 1933 543

MEASURING RESISTANCES By the late W. A. Barclay, M.A. 552

CORRESPONDENCE 555

ABSTRACTS AND REFERENCES 557

SOME RECENT PATENTS 582

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 (so lines)

Branch OfficesCOVENTRY BIRMINGHAM MANCHESTER GLASGOW

xp Hertford Street Guildhall Bldgs., Navigation St., 2 26o Deansgate, 3 26B Renfield Street, C.2Telegrams: "Cyclist, Coventry" Telegrams: " Antoprem, Birmingham" Telegrams "Aide, Manchester." Telegrams: " Wife, Glaagow "Telephone: 9210 Coventry Telephone: Midland 2970(9 lines) Telephone: Blackfriars 4412(4 lines) Telephone: Central 4857

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, S.E.r.Especial care should be taken as to the legibility of MSS. including mathematical work.

1

Octuber, 1933

Showing how con-denser screws intobase for easy fixing.

¢14

New andIMPROVED

UBILIEPAPER CONDENSERS

Outstanding features -1. True non -inductive 4. New method of fix -

type of construc- ing to chassis.

tion. 5.Aluminiumcontainers withmoulded bakelite

2. Available for work- top of distinctiveing voltages from appearance to300900 D.C. peak. match other com-

ponents.

6. Takes up minimumamount of space onchassis.

3. Adequate factor ofsafety for each.

PRICES : 1 mfd. 2,6 : 2mfd. 3/6 : 4mfd. 7 -

DUBILIER CONDENSER CO. (1925) LTD.DUCON WORKS, VICTORIA ROAD, N. ACTON, W.3.

THE WIRELESS ENGINEER

%%%1OrAZZZAIZZZIKWZMY,WWW.M15101

WIRELESS

IR ViCTIONFINDING

andDIRECTIONAL

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

SECOND EDITION (5927)

This volume deals with the prin-ciples of the subject, and with theconstructional details of directionfinding installations, and includessome information concerning air-craft installation. It describes theprinciples of Direction and PositionFinding in such a way that thesubject may be grasped easily byengineers tackling this field ofwireless work for the first time.Numerous photographs anddiagrams are included.

PRICE21/- net

By post 21/9

From leading booksellers or direct fromthe publishers

ILIFFE & SONS LIMITEDDorset House, Stamford St., London, S .E.1

w.r./.

0,741:41,V.ZZWIZZICKWY.ZW.ZZYMMKindly mention The Wireless Engineer- when replying to advertisers.

52

THE

WIRELESSENGINEER

ANDEXPERIMENTAL WIRELESS

Vol.. X. OCTOBER, 1933 No. 121

EditorialSome New Types of Broadcast Transmitting Aerials

THE August number of the TelefunkenZeitung gives some particulars of twointeresting types of aerial which have

been erected at stations recently equippedby the Company. At Breslau-a 6o kWstation with a wavelength of 325 metres-there is a single self-supporting woodentower, a photograph of which is here repro-duced. The height of the tower is 140metres, and the aerial consists of a wirehanging vertically down the centre of thetower. At the top is a bronze ring io metresdiameter to serve as a terminal capacity ;this suppresses about 4o metres or an eighthof a wavelength of the standing wave on thevertical wire. A current node occurs at aheight of 19 metres and maximum currentat loo metres as shown in the diagram. Thiscurrent distribution gives a polar diagram inthe vertical plane such that practically noenergy is radiated at an angle of 6o to 65degrees to the horizontal, while the hori-zontal radiation is increased by about25 per cent. The radius of the area in whichreception is obtained free from fading hasthus been increased by about 4o per cent,or in other words, the service area has beendoubled.

About 8o metres above the ground a

Diagramshowing

amplitudeof

current.

(Right)The Breslau

self-supportingwooden tower.

October, 1933

switch is inserted in the wire ; this switch,which can be operated from the ground,disconnects the upper part and enables thelower part of the aerial to be used as a

Vienna steel lattice mast.

quarter -wave transmitter with a currentanti -node at the ground level.

An entirely different type of aerial is beinginstalled by the same company at the 15o kWBisamberg station, nine miles north ofVienna. The lattice mast is of steel, coatedwith zinc (not galvanised) ; as can be seenfrom the photograph, its form is that of twosquare pyramids placed base to base. Thetotal height is 130 metres and it is supportedat the midpoint only by four insulated guys.The tower itself constitutes the aerial, thecurrent flowing in the steel and its zinccoating. As the wavelength is 517.2 metres,it will be seen that the height of the toweris almost exactly a quarter of the wave-length. In view of the nature of the soil inthe neighbourhood of the aerial it wasdecided to use an insulated system ofradiating wires as a counter -capacity. Thephotograph shows how the horizontal wiresare kept taut by means of helical springswhich are short-circuited by means ofjumpers to provide a by-pass for the high -frequency current. The hut behind the baseof the tower contains the tuning and couplingelements of the aerial.

Although at present the station is working

526 THE WIRELESS ENGINEER &

with this one tower and radiating uniformlyin all directions, it is intended ultimately,when complete data of the field strength atdifferent distances have been obtained, to

erect another similar towerwith an entirely independ-ent counter -capacity. Thissecond tower will besituated a quarter of awavelength to the east ofthe present tower and willbe tuned to a somewhatlonger wavelength ; it willact as a reflector andcause the energy to beradiated predominantlytowards the west. Thereason for this will beapparent from a studyof the map of Austria,from which it will beseen that the country liesmainly to the west ofVienna. This is evenmore so with regard to thedistribution of German-speaking population.

The wavelength of the Vienna station wastoo long to allow of the use of the anti -fadingdevice employed at the Breslau station.

The horizontal wires are kept taut bymeans of helical springs.

In one case the vertical distribution ofradiation, and in the other the horizontaldistribution, will be affected in a specialmanner and it will be interesting to studythe results obtained in the two cases.

G. W. O. H.

EXPERIMENTAL WIRELESS 527 October, 2933

Applications of the Dynatron *By M. G. Scroggie, B.Sc., A.M.I.E.E.

Summary.-The characteristics and operating conditions of the screen -grid valve used as adynatron are reviewed from a practical point of view. For theory the reader is referred toexisting literature except where it is found necessary to supplement or criticise the treatmentof certain matters of practical importance. The precautions to be observed in obtaining the maximum advantage from the special features of the dynatron are detailed, and it is shownhow it may be applied to receivers, wavemeters, standard signal generators, beat frequencyoscillators, and apparatus for measuring the radio -frequency resistance of oscillatory circuits,the power -factor of dielectrics, and the admittance of chokes and similar components.

THE theory of the dynatron oscillatorhas been dealt with in numerousarticles, mainly of recent date ; for

convenience a bibliography is appended.Here it is proposed to set forth the practicalaspect of the matter, which is of some interestin view of the ubiquity of oscillators inlaboratory work.

The dynatron possesses substantial ad-vantages over other types of oscillator thatare available for the same purposes, and inaddition has applications for which othertypes are not available. For example, atwo -terminal oscillatory circuit, with oneterminal " earthy," can be maintained inoscillation ; there being no necessity fortappings, or auxiliary circuit elements.The oscillations are in general more stable infrequency than those maintained by triodes(17, 19, 25) t and are under more precisecontrol. Details of these and other advan-tages will appear when considering thecharacteristics and applications.

Although a dynatron patent was taken outno less than i8 years ago, it is only com-paratively recently that valve developmenthas made it a really useful tool. While it ispossible to obtain dynatron effects with manytypes of triode, the effect is much moreprominent in the tetrode ; and improvementin valve conductances has still further ex-tended the range of oscillatory circuitresistance that it is possible to neutralise bythe dynatron negative resistance.

Almost any screen -grid tetrode may be usedwith more or less success, and the typicalcharacteristic shown in Fig. i exhibits thecommon features. The anode current atzero anode volts may be quite considerable,

* MS. received by the Editor, February, 1933.t Numbers in brackets refer to bibliography.

amounting to a milliamp or more. As thevoltage is increased the current rises rapidlyand reaches a shafp maximum at about ioor 12 volts, then decreases to form thedynatron slope.

a 4

Zwcc 3w

wcc 2

2-O

-1 1

w

0z 0

-1

pt.,IFYINGRTION

1 0 150

SCREENVOLTAGE Vs

ANODE VOLTS V aFig. 1.

In some valves this portion of the curveis very nearly straight, in others there areoccasionally irregularities, but the typicalcurve is concave viewed from above, andthere is sometimes quite a knee at thepoint P, resulting in exceptionally lownegative resistance between P and the topbend. This is, therefore, a useful region forcoaxing difficult circuits to oscillate.

The anode current is shown to have nega-tive values, and in some high -conductancevalves this forms the major portion of thedynatron curve, while in others negativevalues are nowhere approached. At ananode voltage somewhat lower than the screenvoltage the anode current exhibits the lowerbend, blunter than the upper one ; and themiddle steep upward portion is alwaysreached when anode and screen are at thesame potential. Finally the curve flattensout in the portion generally utilised foramplification.

It is hardly necessary to plot a screenB2

October, 1933

current (Is) curve, as this may be derivedto sufficient accuracy for most purposes byone reading, taken, for example, at thepoint Q where Ia is zero. As Ia +is substantially constant so long as Va onlyis varied, I, is a reflected image of the I acurve.

Fig. 2 shows how the typical character-istic is affected by varying the operatingconditions. Both length and steepness ofthe dynatron curve increase with V8, but apractical limit to this process is set by theconcurrent increase of I e, which has anadverse effect on the life of the valve. Itappears that heavy screen current is par-ticularly damaging to most types, and thoughthe useful limit depends very largely on thedesign of the valve, 5 ma is suggested as alimit for prolonged runs and io mA for brieftests. That these figures may be con-servative is shown by a test on a MazdaAC/S2 valve, which was operated as anoscillator for 2i hours with a screen currentof 20 MA at no volts. Both dynatron andamplification characteristics taken beforeand after the treatment showed no de-terioration.

Increase of Va, the control grid voltage,

dynatron curve (Fig. 2b).It is of interest to note that when I a

is zero or negative the mutual conductanceis zero and negative respectively. Thezero point suggests a useful condition incases where it is desired to use a grid biascontrol without affecting the anode current.

INCREASING Vs

Fig. 2.

INCREASING Vg

Va(b)

vs

Variation of filament or heater currentis similar in effect to variation of Va, butit is generally more satisfactory to usesome other form of control. The thermaltime lag makes this particularly true of


separately heated valves, which, because oftheir greatly superior characteristics, are thetypes mainly in view.

In considering desirable operating con-ditions it will be assumed for the momentthat the various feed voltages are main-tained at the steady values to which they areset.

(a) (b)

Fig. 3.

The simplest connection is that of Fig. 3(a) ;(b) is an alternative where it is desired tobring the oscillatory circuit to the cathodeside of the feed. This circuit should be usedwith caution, as there is a possibility of thechoke oscillating on its own. Oscillation ismaintained when the dynamic resistance ofthe LC circuit is not less numerically than thenegative resistance of the valve path inwhich it is connected.

That simple statement requires someelaboration, for in most treatments of thesubject the negative resistance of the valve,which we will call p, is conveniently assumedto be constant, under any given set of con-ditions, an unpractical assumption whichhas led to great confusion of thought, as willbe shown when discussing the effect of thevalve on the frequency of oscillation.

In practice the initial working point is setto a nearly linear part of the dynatron slope.In general the oscillatory circuit has a greaterpositive dynamic resistance R than the prepresented by the slope at this point, andif p were constant the amplitude of oscillationwould increase without limit -at a rate de-pending on -the difference between R and p.Actually p varies all over the cycle of oscil-lation, and the amplitude adjusts itselfto a value which makes the time -integralof 1.2p, over a cycle, equal to /a2Rt. Withsome valves the range of adjustment of R

EXPERIMENTAL WIRELESS

corresponding to various amplitudes ofoscillation along the " straight " part isvanishingly small, and may be non-existentowing to the numerical value of p tending todecrease as the amplitude increases. Whathappens then is that oscillation once es-tablished travels at least as far as the nearestbend. In other words, the settings of agrid -bias or other control, corresponding tothe starting and stopping of oscillator, donot coincide. This effect is referred to asbacklash.

It is difficult or impossible to obtain con-tinuous adjustment of the amplitude belowthis amount by manual control of Vg,though sometimes irregularities in the char-acteristic permit oscillation at one or morefixed small amplitudes, but the effect is notvery stable. As the difference between Rand the initial value of p is still furtherwidened, by adjustment of Vg or otherwise,oscillation sweeps over the nearer bend inorder to make the average p equal to R.The actual travel cannot be plotted by drawingthe load line of R and noting the points atwhich it intersects the valve curve. Oscillo-graphic tests show that a line joining the twoextremes of oscillation may even show apositive slope. The average along the actualcurve must be negative, however, in orderto maintain the oscillation. Fig. 4 showstypical oscillographcurves of oscillationat three values ofVg, with constant loadand hence constantaverage p. The linesAA, etc., show howerroneous it would beto assume that theyrepresent the aver-age p.

As the current in the oscillatory circuitis much more sinusoidal than I., which is thevalve feed current, and the excursion of V. isproportional to the former, one can see thatthe negative portion of the slope duringwhich the valve feeds power into the oscil-latory circuit is of longer duration than thepositive portion when the valve absorbspower, and hence the balance is maintained.

The situation represented by the curveCC appears to be appalling as regards wave-form, but it is not so bad as it looks, becauseof the fact just mentioned, that the oscil-

0

INITIALSETTING

v a

Fig. 4.

BA

529 October, 1933

latory circuit current, and hence voltageacross it, is much purer ; particularly if the

ratio is small. If this were not so, theviolent flattening of the negative half wavewould be counterbalanced by an excessive,amplitude of the positive half, but as thediagram indicates, the amplitude is nearlyas symmetrical in the over -oscillating adjust-ment as it is in the gentle one.

Nevertheless, CC is not an ideal state ofaffairs when harmonics are objected to, andoscillation should be restricted to the nearlystraight travel.

It has already been explained that inordinary practice this is difficult or im-possible, so in the interests of stability R,or more conveniently p, is adjusted so thatoscillation extends up to and slightly beyondthe upper peak. Though purity of wave-form is thereby sacrificed to some extent,practical tests show that it is a considerableimprovement on the wave -form obtainedfrom a triode oscillator operated undercomparably stable conditions.

The other important feature is thatalthough the amplitudes of harmonics aregreatly dependent on such adjustment, thefundamental is not, and may be relied uponto be constant within fairly close limits. Infact, it may be predetermined to usefulaccuracy without direct measurement, simplyby choosing an appropriate anode voltage.

Control of amplitude, where desired, iscarried out by adjustment of Va, and whereabout 5 per cent. accuracy is sufficient ananode voltmeter may be directly calibratedin oscillation voltage, which is to a greatextent independent of other quantities.Zero corresponds to an anode voltage justto the right of the upper bend, and maximumcorresponds to the centre of the dynatronslope. As V. is further increased theoscillation amplitude diminishes again, butthe first half of the slope forms the moresatisfactory range of adjustment, because ofthe sharper limiting effect of the upper bend.Also because /a is kept away from thenegative region, Is is minimised and greaterpower may be safely applied.

An exception to this useful property occursin valves in which /0 is mainly negative.As the average p cannot be increased by /abecoming negative at the low voltage end,oscillation continues along the voltage axis

October, 1933 530 THE WIRELESS ENGINEER &

and the amplitude increases until the secondbend comes to the rescue. Fig. 5 is a sketchof an oscillograph trace of such a valve, thethick line indicating the extent of the oscil-latory voltage.

Excluding this case ; the desired maximumamplitude being known, the necessary V,

4

INITIALSETTING

Va

Fig. 5.

follows, being approximately 20 volts morethan the double amplitude. Vg is then setto the greatest negative value consistentwith stability of oscillation. In a variable -frequency oscillator a greater margin ofstability is necessary because of the con-sequent variation of R. Assuming con-denser adjustment of frequency, and R being

equal to where r is the resistance of the

inductor, it tends to drop as the capacitanceis increased, unless steps are taken to compen-sate for it. This may be done by designingthe coil to have a resistance sharply risingwith frequency, or conveniently by shuntingit with capacitance and resistance in series(Fig. 6). The values are adjusted experi-mentally to suit the characteristics of thecoil and the permissible minimum capaci-tance.

When it is important that harmonicsshould be excluded, or at any rate the some-what variable harmonic content that isunavoidable even with intelligent manualcontrol of Vg, it is worth while adoptingautomatic amplitude control. This is closelyallied to automatic gain control as used inreceivers, and the numerous published articleson that subject furnish information on thetechnique, which is in most respects appli-cable to oscillators also. As far as the writeris aware, only one article has appeareddealing specifically with oscillator control(27)

In its simplest form, the oscillation voltage

is applied to a diode, and the unidirectionalpotential thus obtainable is used to controlVg. If the whole of the oscillation voltageis so applied, a very steep control of pcan be obtained, so that once the amplitudeof oscillation has built up to the amountsufficient to bring Vg to the point cones-ponding to the p which permits of oscillationat that particular amplitude, any furtherincrease in amplitude brings into play avery large restoring bias. The result isan extraordinary stability of amplitude,which may be made to extend over anydesired part of the slope. If the diode isarranged so that the potential of the anodeis at the starting point of anode current, anyamplitude of oscillation whatever causes acontrolling bias to be applied, and con-sequently the amplitude is kept extremelysmall, and may be made still smaller byapplying an initial bias to the dynatron.Larger amplitudes are obtained by biasing thediode, so that control comes into operationonly when the peak oscillation voltageexceeds the bias. The amplitude thus bearsa simple relation to the diode bias voltage,and by suitable design may be made approxi-mately equal to it, within a few per cent.

The oscillation peak must of courseexceed the diode bias by anamount which (unless amplifiedcontrol is used) cannot be lessthan the controlling bias neces-sary to arrive at equilibrium.Therefore, when it is wished tojudge the amplitude of oscilla-tion by reading the diode biasvoltage it is desirable either tohave an auxiliary manual con-trol to apply approximately thecorrect bias to the dynatron,or to make the diode bias largecompared to the necessary controlling bias.

This method of operation is a close approxi-mation to the ideal linear oscillation, andprovided that the diode load is kept smallthe purity of waveform is exceptionally good.The amplitude may be adjusted very con-veniently over a wide range while preservingthe pure waveform. And it is obvious thatby connecting in series with the diode biasa source of relatively low -frequency alter-nating voltage, very perfect modulation ofthe dynatron oscillation is obtained, and thatthe depth of modulation is equal to the ratio

Fig. 6.

EXPERIMENTAL WIRELESS 531

of peak modulating voltage to steady biasvoltage.

As the tendency is to drive even laboratoryapparatus without batteries, the assumption

Va

(a)Fig. 7.

(b)

as to fixed supply voltages is not alwaysjustifiable. It is possible to arrange afairly low impedance source for supplyingthe screen electrode, and power supplycircuit impedance can be largely eliminatedso far as A.C. is concerned by suitably largeby-pass condensers. The obvious methodof deriving the anode supply is by a voltage -dropping resistor in series, in conjunctionwith a large condenser. Attempts to accom-plish this are rewarded by an entire absenceof results. The reason is clear by drawingthe line representing the feed resistance R,(Fig. 7) from the point S, corresponding tothe screen voltage, to the desired workingpoint P. This line cuts the valve curvein two other places, A and B, and as at Pthe voltage -dropping resistance exceeds thevalve negative resistance it is an unstablestate of affairs, and the working pointpromptly settles down at A or B, dependingon which end it starts from. These beingpronouncedly positive in character, thereis not the faintest hope of dynatron action.

So it is necessary to take the anode supplyfrom a potential divider of resistance con-siderably lower than p. When using verylow p valves, such as the Mazda AC/S2,it is important to exclude from the anodecircuit any coils, such as even a single turnof wire, otherwise undesired oscillation isliable to confuse matters.

As p is the most important characteristicof the valve, it may be well at this point torefer to methods of measurement. Anobvious way is to plot curves and draw atangent at the working point. This is tediousand inaccurate. Another method is to dis-play the curve on the screen of an oscillo-

October, 1933

graph ; expeditious enough if the apparatusis set up, but even more inaccurate. Anexcellent method is that described by E. N.Dingley (i8), the circuit of which is shown inFig. 8. This is a very simple bridge com-prising source and detector, tapped resistorR1 R2, and resistance box R3. When balanceis obtained

R,R,R R3P = R,

Convenient values are R, = 99 and R2 = I.The expression then reduces to p = I00 R3 ±99 and the 99 is often negligible. A decadebox from io-io,000 ohms gives therefore arange of from i,000 'ohms to one megohm.For accurate results the usual precautions inbridge work must be observed. Two factorsthat prevent extreme sharpness of balance are(a) the capacitance of the valve electrodes andleads, which may be neutralised by a mutualinductometer (shown dotted), and (b) thecurvature of the valve characteristic. Thisgives rise to harmonics at balance which areaurally distinguishable if phones are used,and may be rendered negligible by using asmall signal. When properly arranged,balance can be obtained to a few parts inio,boo. The steady voltage drop in theapparatus is negligible.

At very high radio frequencies the dynamicresistance R of oscillatory circuits is low,and difficulty may be experienced in obtainingoscillation (i9). The present writer haspointed out (i9) that the AC/S2 valve isparticularly suitable, and strong oscillationat 6o megahertz has been obtained.

The capacitance of the anode to otherelectrodes should for various reasons be as

R3

R1 R2

tiFig. 8.

0

kLai

small as possible (II, 15, 23). It may beas high as 17 'Lid'. or as low as 2.5 (23).British valves are superior to American inthis respect, that the power factor of theadmittance is low, owing to the anode beingbrought out at the top terminal. They also


appear to be better in oscillating properties(25).

All types seem to he consistent in onerespect, namely their inconsistency. Sampleswhich are closely uniform when used asamplifiers display marked disparity in dyna-tron characteristics. The dynatron effectis a more or less accidental one, and is in factsomething of a blemish for the more commonpurposes of the valve, and so designers do nottry to obtain consistency in this respect.For laboratory purposes valves must in anycase usually be selected, so this drawbackis not so serious as for large scale activities ;but possible applications may make itdesirable for more attention to be devoted tothis matter.

The Table showing the principal char-acteristics of some prominent types of valvesgives the averages of several samples in somecases, but p is liable to vary considerablyfrom the figures quoted. It also varies inany one valve over the slope ; the valuesgiven refer to points near the centre of theslopes, and may be considerably lower nearthe upper bend. As low as 2,000 ohms hasbeen measured. Fig. q shows a typicalcharacteristic. Variable -conductance valvesare generally less suitable than other types,because of the greater range of grid biasrequired to effect a given degree of control.This is particularly so in the case of auto-

(V, = loo)

Turning now to dynatron applications:the first to be described in extended detail

0cc

z

0z

00

500

2

2500z

ONNEENENNEMENNEENEN.MENNEENNEMMENNEN..IMMEMMINIMMONNENENNEMIMENNENEMENEM...MENEM......MEMMHEEMEENENEUNENNEENNEMENNEN..NMENNEMENMENENEENENENNEENNNENNEENNEEN.v.2.

"11

NENE. EN20 40 80 80 100

Va

Fig. 9. -Anode resistance of ACS2(expressed as conductance to avoid

discontinuity).

was that of detector in a heterodyne receiver(3). This paper by A. W. Hull is not so out-of-date as the year of publication mightsuggest, and though heterodyne reception hasbeen in abeyance for some years because

TABLE.

Valve.

Negative Resistanceohms.

1. + 1, mA. I,, startsat

V, =

Ca-,PPR

17(7-0. Vg= -2. 119-0. Vg= -2.Mazda AC/SG .. 13,000 20,000 17.6 II -0.1 17.5AC/S2 .. 6,400 11,400 19 9.8 - 1. 1 11.2

AC/SIVM . . 13,000 18,000 12.3 9.3 -0.4 11.0Mullard S4VB .. 38,000 8o,000 8.z 2.9 -0.5 6.3

S4VA .. 40,000 110,000 9.5 2.7 -1.1 7.0, MM4V 18,000 35,000 13 6.5 - 1.1 7.oPMI2A .. 30,000 mom. 4 1.8 -0.3 5.8

Cossor MSG/LA .. 27,500 40,000 13.3 5 -0.9 13.2VMSG .. 35,000 50,000 13.4 8 -0.5 12.022OVSG . . 14,800 26,000 II 6.3 0 6.9

Vg and Ig = control grid voltage and current respectively.V, and I, = screen grid voltage and current respectively.V, and 1.= anode voltage and current respectively.Ca_, = capacitance from anode to screen.

matic control, and in Fig. 10 the degree of of the focussing of attention on telephonycontrol which V5 has on p is indicated by the rather than C.W., the revival of the super-slope of curves drawn for various valves. heterodyne makes this Io-year-old paper

EXPERIMENTAL WIRELESS .

worth studying. The principal objection tothe dynatron as a single -valve frequencychanger is the inconsistency of its char-acteristics-not an insuperable difficulty,surely. The freedom from interaction be-tween the anode oscillating circuit and thegrid circuit is a very attractive feature forsuperheterodyne purposes, as is also theabsence of oscillator reaction coil.

For the latter reason the dynatron findsan obvious vocation in wavemeter work. Anexisting absorption wavemeter, consisting ofa coil and condenser, may be converted into agenerating wavemeter without the loss ofaccuracy that would result from adaptingto a triode circuit. Tests carried out byMessrs. H. W. Sullivan, Ltd., have establishedthe fact that in suitable screen -grid valvesthe added capacitance is both small andstable and is consistent with the maintenanceof very high precision so that the wavemetererror may be held down to in 20,000. (30)The Mullard S4VA valve is used for this

VB AC S2!SAMPLE 2)

0 1 2 3 4 5 8 7

AC

,SA

/SG

MPLE 1)

A's VARI AS E-14

Vs - looho

8 9 10 11

io5

1o5

io4

4°3NEGATIVE _GRID BIAS IN VOLTS

Fig. 1o.-Slope of curves = dV- . - , which is adp p

measure of control of grid on negative resistance.

purpose and its relatively poor oscillatingpowers are no drawback in such low -lossoscillatory circuits.

533 October, 1933

The differences in capacitance betweenvalve and valve, of the order of i tioF, aretaken up by a fine adjustment calibratedfor a number of spare valves.

Another useful wavemeter produced by thesame firm is due to N. W. McLachlan (8)and theoretical points in connection with ithave recently been dealt with by him (21,23). Both radio and audio -frequency circuitsare connected in series with the anode andwhen suitably arranged a completely modu-lated signal is given. The precision is lower,but for many purposes the advantages ofmodulation are more to be desired. Thematter of modulation will be discussed laterin another connection.

Apart from the convenience of applicationand the stability of interelectrode admittance,the dynatron is superior to the triode oscil-lator as regards independence on accuratesupply voltages (ii, 25). For all except theutmost precision it is possible to dispensewith meters for setting the voltages exactly,so long as oscillation is no more than ade-quately maintained. The frequency ofoscillation with a low -loss circuit is thenextremely close to the limiting value

27TVLC.The advantages of using a dynatron formaintaining a crystal fixed -frequency stan-dard have been investigated in a paper (25)which, among other interesting results, showshow oscillation can be maintained withbattery feed interrupted by a crystal in theanode circuit. The maximum variation offeed voltages over which oscillation ispossible is said to affect the frequency by anamount of the order of one in a million.

The standard signal generator for receiverperformance testing requires an oscillationof stable and controllable amplitude, lowharmonic content; and capable of amplitudemodulation to any desired depth withoutfrequency modulation. The dynatron fulfilsthese requirements with a notable absence ofthe elaborate associated equipment neces-sary for triode oscillators, and is thereforeparticularly suitable for commercial test -gear.

The intensity of the signal is controlledover a wide range by an attenuator, but as itis inconvenient to cover the necessary range(which may be zoo db or more) with closesteps, it is desirable to be able to vary theamplitude of oscillation over a few db without

October, 1933 534

shifting the frequency or waveform appre-ciably. Either the anode voltage control,or the automatic control already described,with voltmeter calibrated direct in db, fills thisneed.

The self -modulating circuit of McLachlanis not sufficiently flexible, but a separateaudio -frequency dynatron makes possible any

Fig. II.

desired depth of modulation without inter-fering with the other adjustments. Fig. IIshows a skeleton circuit of the anode controltype. The two oscillators are run in parallel,with simultaneously variable anode voltage,but whereas the a.f. oscillator is fed straightfrom the supply, the r.f. anode circuitincludes a portion of the a.f. coil La. As theoscillatory voltages across L,. and La areequal, a tapping to give any desired depthof modulation from o to zoo per cent. caneasily be arranged.

The only difficulty is likely to occur whendeep modulation is required and the wholeor greater part of La is in series withThe arrangement is then equivalent to theself -modulating system and the a.f. valveis likely to lose control of the situation.There are various possible methods of dealingwith this ; one is to make the dynamicresistance of the r.f. circuit appreciably higherthan that of the other, so that when the valveis properly adjusted it is incapable of stimu-lating oscillation in the a.f. circuit. Therelatively lower dynamic resistance of the

a.f. circuit should be obtained by a low -C

ratio rather than by high loss resistance, as itis more important to avoid harmonics herethan in the r.f. circuit. This desideratum

THE WIRELESS ENGINEER &

is facilitated in the equipment described byemploying a fixed audio frequency whichpermits close adjustment of oscillation.

These considerations are easily confirmedby connecting one pair of cathode-ray oscil-lograph deflection plates across Li. and theother across the portion of La common toboth valves. Fig. 12 shows the imagescorresponding to various operating con-ditions. When none of La is included, or noa.f. oscillation is taking place, the r.f.oscillation is indicated by a straight line (a) ,the length being a measure of the amplitude.The correctness of the amplitude control, andindependence on other adjustments, is con-firmed. When tapped for 5o per cent.modulation the figure (b) results. The depthof modulation is easily derived, beingCD - ABCD ± AB. With pure linear modulation

the sloping sides are straight and inclined at45°. This is a test of modulation per seand not of a.f. waveform, which has noinfluence on the shape of the figure, thoughit may cause unevenness of illumination.

The right-angled triangle (c) is the sign ofperfect ioo per cent. modulation. (d) in-dicates over -modulation, and is particularlylikely to take place in the self -modulatingcircuit. This condition is also detected bythe sound of the resulting signal.

The dynatron system described yields re-sults as perfect as the oscillograph is capableof indicating. As the amplitude control isoperated both oscillation voltages varyequally and the depth of modulation ispreserved constant. The range of adjustment

(a) (b) (e) (d)

Fig. iz.

(e)

is limited to a quarter of the distance from theupper bend to the lower. By going beyondthe limit a second triangle begins to formand, at the centre of the slope, appears as acomplete image of the first (e). There arethen two complete cycles of modulation toeach one of the a.f. oscillator-a perfectfrequency -doubler.


Obvious extensions of the system, such asthe substitution of an amplifier valve for thea.f. oscillator, to permit of modulation by anexternal source of any kind, require nodetailed description.

If automatic control is applied-and it iswell worth while-there are many possiblevariations of circuit design, and a mass ofdetail is involved which lies outside thescope of this general article. It may bepossible later to give some account of this.

0

Lr

October, 1933 -

method applies a weak and a strong signalto a linear rectifier. Both call for oscillatorsgiving outputs of fairly definite voltageand great constancy of frequency. This is,therefore, another promising field for thedynatron.

An instrument designed to utilise a pairof dynatron oscillators proved very suc-cessful. Each anode circuit included a tunedastatic step-down transformer with a primaryvoltage of about 20 and a screened secondary

R.F.OSCILLATOR

R.F.CONTROL

VALVE

S2

L

.0>

A .F.CONTROL

VALVE

T -r

A.F.

OSC.

Fig. 13.-Modulated oscillator with automatic grid bias controls.

Fig. 13 shows the bare bones of the method.The amplitudes of both r.f. and a.f. oscillationare simultaneously adjusted by the slider S1,which biases the cathodes of both controldiodes positively, and the slider S, controlsthe percentage modulation by tapping off aproportion of the a.f. potential. Filtersystems are included in the grid leads of bothdynatrons, to prevent oscillation -frequencypotentials from being impressed on them,but in the case of the r.f. valve to allowmodulation -frequency potentials to havefull effect.

Particular valves are not given, becausethey are a matter for circuit design, depend-ing on the frequencies of oscillation andother factors.

The companion piece of equipment essen-tial for all performance testing, not only ofradio receivers, but also of lines, amplifiers,filters, etc., is the beat -frequency oscillator.Two systems compete for favour. In theless -known one equal signals are applied to asquare -law detector and a beat -frequency ofgood waveform results. The more popular

Si

with i volt output feeding a 50o -ohm non -reactive resistor. The whole of the potentialacross one resistor and one -tenth that of theother were fed into a power grid detector.The weak signal frequency was fixed at5o kilohertz and the strong signal varied from40 to 50, giving a beat -frequency fromo-Io kh.

This was obviously a case for automaticcontrol, as harmonics are practically elim-inated, and oscillation maintained, regardlessof frequency adjustments or of valve de-terioration. In fact, any dynatron whichwill oscillate at all can be plugged in and willwork always at the optimum adjustment.(See Fig. 14.)

Apart from the screening between thewindings of the tiny r.f. transformers, noscreening of any description and no buffervalves were found necessary-greatly sim-plifying construction-and beat oscillationsof less than one cycle per second wereobtainable. The frequency stability ofthe dynatron is of great value in thisapplication, and the residual drift largely

October, 1933

cancels out because of the symmetricalconstruction.

The utility of the dynatron is not re-stricted to adjustments which cause con-tinuous self -oscillation to be maintained,but may be realised for the purpose ofreducing losses in a tuned circuit (28).It is not always practicable to employ atriode in such cases.

Some considerable attention has beendevoted to the application of the dynatronto the popular problem of measuring r.f.resistance. Colebrook (17) favours thedynatron in the subsidiary role of oscillatorin a method requiring exceptional frequencystability but not otherwise connected withdynatron properties. Pauli (9) appears tohave been the first to describe a methodof measuring positive resistance by equalisingit to the negative resistance of the dynatron,and Iinuma (ii, 13, 15) who worked in-dependently along the same lines hasdeveloped the technique in important re-


equal to the dynamic resistance R of theunknown.

As R7C

r" the effective resistance of

the coil, follows from a knowledge of L andC or L and f. The principal theoreticalerror is that due to the r.f. loss in the valveadmittance. Iinuma has made a detailedinvestigation of this error and found it tobe small. in most cases (z5). Reasonablygood overall accuracy is claimed in spite ofseveral disadvantages which he unnecessarilysuffered. American valves were used, whichas other workers have found (25) are notnearly such good dynatron oscillators ascertain British types. Excessive screencurrents arc required for even moderatelylow negative resistance. The anode isbrought out at the cap along with the otherelectrodes, resulting in larger capacitance andbad power -factor. The method of measuringp was by increments-a much less satisfactoryway than that illustrated in Fig. 8.

.--SAAAAA4-

TFig. 14.

spects. The present writer (24) has suggesteda substitution method for rapidly measuringresistance and reactance, which is capableof useful elaboration.

The method of Iinuma is exceedinglysimple. The oscillatory circuit whose re-sistance is to be measured is connected inseries with the anode, and the grid bias isadjusted to the oscillation point. Thenegative resistance of the valve is mea-sured at this point, and is numerically

TOFILTER

With these improvements this methodmay be expected to compare very favourablyin accuracy with any known methods,and convenience and rapidity are muchgreater.

The dynamic resistance of a tuned circuitcan be measured over a wide band of fre-quencies in a few minutes, and by a suitableswitch -over to the bridge of Fig. 8 errorsof manipulation and calculation are farless likely to arise than in systems of possibly


greater theoretical precision. It will benoted that measurement is not confined toradio frequencies.

Tests made on a number of coils from12 pH to 50,000 pH showed remarkableconsistency of results. Measurements weremade with and without a short piece offine eureka wire in circuit, and the resistanceof this wire, as given by the differencebetween the results of the two tests, wascompared with the measured d.c. value. Theerror, relative to the whole resistance incircuit, was less than i per cent., except atthe extremes of frequency. A Mullard S4VBvalve was used, having a capacitance fromanode to other electrodes of only 5.6 p.i.LFand power -factor of o.oi, and therebyintroducing a loss smaller even than thatof the standard condenser used for tuningthe coils.

A feature of the method which is especiallyvaluable in these days when " potted "coils are the rule is that, unlike other methods,magnetic coupling to the coil under test isnot required. Another feature of greatpractical value is that an oscillatory circuitof the type used for r.f. amplification canbe measured in situ; by rearranging thefeed connections to convert its amplifyingvalve into a dynatron. Condenser, screen,and lead losses are thus included, and theactual amplification can be estimated withaccuracy.

It is important to observe that it is assumedthat the resistance of the valve at the thres-hold of oscillation is the same as that measuredby the bridge. Apart from the possibledifference due to valve dielectric loss alreadymentioned there is an obvious discrepancyif the valve stops oscillating abruptly asdescribed in an earlier section. The secretof accuracy is to adjust the valve to eliminateelectrical back -lash in the oscillation control.The loss due to the valve can be estimatedby paralleling in another valve of the sametype.

This naturally suggests the measurementof power factor and permittivity of dielectricsagainst a standard low loss variable con-denser at any audio or radio frequency.The dielectric sample is clamped betweenmetal plates to form a condenser in parallelwith the standard. A Coil appropriate to thefrequency of the test, and of as high efficiencyas possible, completes the oscillatory circuit.

October, 1933

After a test with the sample in circuit itis removed and the initial frequency re-stored on the standard condenser beforemaking a further resistance test. If C isthe capacitance difference (and thereforethat of the sample) and R, and R2 the twodynamic resistance readings ; neglecting thechange in resistance contributed by thestanddrd condenser, and assuming the powerfactor of the unknown is not very large,then power factor

R, - R,wC R1R2

The usual beat or double beat note method

enables the frequency - to be brought to27r

equality in both tests.The sensitivity of the method, depending

as it does on the difference between R, andR2, is increased by making the capacitance Cof the sample a large proportion of the total.At the same time it is desirable in theinterests of accuracy for the self -capacitanceof the coil to be a small part of the whole,and as it is seldom possible for the capacitanceof the sample to be numerically large, onemust use common sense to judge of the bestcondition of measurement.

A somewhat similar method has beenextensively used by the writer in researchon r.f. chokes. These may be considered asan admittance which is a source of loss toany circuits across which they may beeffectively in parallel. This admittance isconveniently represented by a resistanceshunted by a reactance, but for ease ofmeasurement and for comparison a positiveor negative capacitance was the form inwhich the reactance was actually expressed.It was necessary to test a very large numberof chokes over the whole useful broadcastband of 15o to 15,000 kilohertz at frequentintervals because of the numerous sharpsub -resonances. The apparatus employedwas a dynatron oscillator with grid biascontrol and voltmeter, low -loss coils coveringthe range of frequency, a main tuning con-denser, and a small variable -maximum con-denser on which readings could be taken to0.1 pl.A.F. Curve sheets were made outcovering the whole range of frequency andwith V, as ordinates. A preliminary testwas made to mark off the sheets withcurves corresponding to the critical V, for

October, 1933

just stopping oscillation, with various highnon -reactive resistances, from 20,000 toinfinity, shunted across the tuned circuit.This resulted in a number of roughly hori-zontal lines. Having thus calibrated thevalve, runs could be made on chokes shuntedacross, and the resulting curves plotted onthe prepared sheets indicated the effectiveparallel resistance of the choke over therange of frequency ; and very interestingresults they were. The initial calibrationcurves being not perfectly horizontal, a moreaccurate impression could be gained byreplotting to resistance ordinates, but forgetting a preliminary idea of the performancethis is not necessary.

14

12

14 10

8

4

2

0

2M0IMO

0'5 MO

CC&25M0

100,000 0

50,000 0

25,000 0

0200 300 400 500 600

WAVELENGTH

Fig. 15.

Well -designed chokes maintained a re-sistance of one megohm or more over mostof the range, but some samples descendedwell below 20,000 ohms in places

The capacitance was obtained from theusual substitution readings, and showedobvious relationship to the resistance curves(Fig. 15). Large capacitance and lowresistance ran in accompaniment, and theshape of the capacitance curves at reson-ances was that of a first differential of theresistance curves, for reasons explained inany standard theory. The frequencies werein all cases equalised at the threshold ofoscillation.

The change of frequency resulting fromthe connection of the choke was assumed


to be entirely due to the capacitance thereof.In view of the fact that large changes inresistance were involved, this assumptionis not justified by theory. The usual theoryappears, however, to be at variance withthe facts, as will now be explained.

The usual expression for the frequency ofoscillation is

- 1 V rco 1 - -

A/LC P

where r is the series resistance of the oscillat-ing circuit, and p is, as before, the resistanceof the valve, expressed as a positive quantity.

As -r is normally small compared with unity,P

I (12p)

(Bibliography, 19, 21, etc.)

Now even in cases where p has beenrecognised to be variable over the cycle ofoscillation, results have been deduced onthe tacit assumption that it is constant.Actually the only case that can be expressedin mathematical form is one which is of nopractical interest-an infinite linear valvecharacteristic. For example, some writershave conveyed the impression that it islegitimate to use the above expression forcomputing change in frequency of oscillationdue to varying V, (and hence, we suppose, p)with a circuit of fixed r. It is a fact, notedby Percival (19), that it is possible to obtainvery large shifts of frequency by so doing.

The absurdity of accounting for suchshifts by the assumption of a change in theratio -r , r being a constant property of theoscillatory circuit, and p varied by controlof grid bias or otherwise, is clearly shownby a little substitution.

R, the dynamic resistance of the oscillatorycircuit, must he equal to p, for any differencewould be inconsistent with uniform amplitudeof oscillation. But

So

RrC

r r2C r2

p L 0,2L2=-- m2

where in is the " magnification " of the coil,which is not a function of the valve at all,


and the frequency is therefore constant !The fallacy lies in the failure to recognise pas a periodic variable. It is no use tryingto avoid the difficulty by assuming p to bevery nearly constant over the cycle of smallamplitude oscillation and then applyingit to a case where, due to the large differencebetween the value of p and R, oscillationof the type shown in Fig. 5 takes place !It is impossible to predict the effect onfrequency of altering the dynatron slope ofthe valve, without an appalling problemin integration based on knowledge of thewhole characteristic.

It may perhaps be supposed, however,that the expression for frequency finds alegitimate application where the valve isadjusted to the threshold of oscillation, and

that the correction term2M"-- is negligible

for highly efficient oscillatory circuits butappreciable for inefficient ones: Thus, agood coil has an m of, say, 225, and thecorrection is only ten parts in a million.With a bad coil having an m of 22.5 thecorrection is i,000 parts in a million.

In the choke tests described the equivalentvalue of m was reduced to still smallervalues, and as the change in frequency wasascribed solely to the very small capacitanceof the choke, it appears as if large errors areinvolved.

Assume that no capacitance is added tothe circuit except that due to adjustmentof the measurement condenser necessaryto preserve the frequency constant whena resistance Rx is connected in shunt. Theadded capacitance is - Cx. pi and p2 arethe resistance values before and after. Thefrequency correction term is constant, i.e.:-

r1 r2 Cx2pi 2p2 2C

L L - CxP 1

2 - P22

Cx (PI2 - P22)12p22

Within the bounds of the approximationmade in deriving this result, Cx works outat values of the same order as those ascribedto the capacitances of the chokes, andthus appears to invalidate the method oftest.

October, 1933

Special tests were therefore carried outto shed light on this matter. A dynatronoscillator was set up with a switch -over tothe bridge for measuring p, and smallresistors (Erie watt type) of various valuesshunted across the tuned circuit, the capacit-ance being adjusted by the double -beatmethod to keep the frequency exactly con-stant throughout. In one run a coil ofapproximately 7,15o µ.11 was used at afrequency of r5o,000.

The capacitance of the resistors wasmeasured by comparing one resistor withtwo others in parallel each of double theresistance, and found to be o.8 µp,F. Thevalue of p with no resistor was 217,000 ohms.

Added Capaci-Nominal Capaci- tance Ditto

Shunt p tance Due to Cal-Resist- auti.F to Reduc- culatedance. keep f

constant.tion in p

µp,F.pp.F.

- 217,000 0 - -250,000 I16,000 o.8 0 0.4

50,000 43,90o 1.0 0.2 3.625,000 23,300 1.1 0.3 13r0,000 10,650 1.6 0.8 63

This tends to restore shattered confidence,for it is difficult to see how an effect aslarge as that predicted by theory could havebeen missed. Other tests gave consistentresults. Curiously enough the capacitancerequired to compensate for withdrawingthe various resistors without readjusting V2to the threshold of oscillation was in each caseof the order of that given in the last column.It is not claimed that this is more than acoincidence !

The shift of frequency in practice, then,is comfortably small even over a considerablerange of oscillatory circuit resistance.

ConclusionThe dynatron is, as has been seen, useful.

It might be still more useful if the valvemanufacturers cease to regard it as anaccident.

BIBLIOGRAPHY

1. "The Dynatron " : A. W. Hull, Proc. I.R.E., Feb., 1918,Vbl. 6, pp. 5-35.

2. " An Investigation of the Effects of Electrical Collisions withPlatinum and with Hydrogen " : Horton and Davies,Proc. Roy. Soc., Mar., 1920, Vol. 97, pp. 23-43.

SYNCHRONISINGCOILS


3. " The Dynatron Detector " : Hull, Kennelly and Elder,Proc. I.R.E., October, 1922, Vol. 10, pp. 320,343.

4. " The Numans Oscillator " : K. C. Van Ryn, E.W.& W.E.,Dec., 1924, Vol. 2, pp. 134-136.

5. " The Emission of Secondary Electrons from Nickel " . E. W.B. Gill, Phil. Mag., May, 1923, Vol. 45, pp. 864-878.

6. " Short Electric Waves Obtained by the Use of SecondaryEmission " : Gill and Morrell, Phil. Mag., Feb., 1925, Vol.49, pp. 369-379.

7. " The Secondary Emission from a Nickel Surface due to SlowPositive Ion Bombardment " : A. L. Klein, Phys. Rev.,Dec., 1925, Vol. 26, pp. 800-806.

8. " A New Wavemeter " : Jour. Sc. Inst., Oct., 1929, Vol. 6,pp. 327-328.

9. " Messung Elektrischer Wirkwiderstande mit Hilfe NegativerWiderstande " : H. Pauli, Zeitschr. f. tech. Phys., Dec., 1929,Vol. 10, pp. 592-595.

10. " The Dynatron " : W. H. Newbold, Q.S.T., Feb., 1930, Vol.14, pp. 33-36.

11. " A Method of Measuring the Radio -frequency Resistance ofan Oscillatory Circuit" : H. Iinuma, Proc. I.R.E., Mar.,1930, Vol. 18, pp. 537-543.

12. " Das Gitterdynatron " : Y. Ito, E.N.T., Nov., 1930, Vol. 7,pp. 419-426.

13. " Application of the Dynatron to the Measurement of DieletricLosses " : H. Iinuma, Jour. I.E.E. Japan, Jan., 1931, pp.11-14.

14. " Zur Theorie des Dynatrons " : Y. Ito, E.N.T., Jan., 1931,Vol. 8, pp. 23-30.

15. " Resonant Impedance and Effective Series Resistance ofHigh -frequency Parallel Resonance Circuits " : H. Iinuma,Proc. I.R.E., Mar., 1931, Vol. 19, pp. 467-478.

16. " The High -frequency Resistance of Coils (Novel Methods ofMeasurement) " : A. L. Green, E.W. & W.E., April, 1931,Vol. 8, pp. 183-191.

17. " A New Method of Measurement of Resistance and Reactanceat Radio Frequencies (Appendix IV ; Note) " . Colebrookand Wilmotte, four. I.E.E., April, 1931, Vol. 69, p. 506.

18. " Development of a Circuit for Measuring the Negative Re-sistance of Pliodynatrons " : E. N. Dingley, Jr., Proc.I.R.E., Nov., 1931, Vol. 19, pp. 1948-1950.

19. " The Dynatron Oscillator " : F. M. Colebrook, W.E. & E.W.,Nov., 1931, Vol. 8, pp. 581-584.Correspondence: Oliphant: Basto, Dec., 1931, p. 661.

Black ; Percival, Feb., 1932, p. 77. Megaw, Mar., 1932,p. 152. Basto, April, 1932, p. 213. Groszkowski, Aug.,1932, p. 446. Scroggie, Oct., 1932, p. 574.

20. " Constant -frequency Oscillators " : F. B. Llewellyn, Proc.I.R.E., Dec., 1931, Vol. 19, pp. 2063-2094.(Not specifically with reference to the dynatron.)

21. " On the Influence of Valve Resistance in Oscillation Genera-tors " : N. W. McLachlan, W.E. & E.W., Mar., 1932, Vol.9, pp. 130-135.

22. " An Audio Oscillator of the Dynatron Type " : Don Hale.Rev. Sc. Inst., May, 1932, Vol. 3, pp. 230-234.

23. " On the Frequencies of Double -circuit Screen -grid ValveOscillators " : N. W. McLachlan, W.E. & E.W., Aug., 1932,Vol. 9, pp. 439-444.

24. " A Useful Test Instrument " : M. G. Scroggie, World Radio,Sept. 2nd, 1932, pp. 487-489.

25. " Crystal Control as Applied to the Dynatron Oscillator " :K. A. Mackinnon, Proc. I.R.E., Nov., 1932, Vol. 20, pp.1689-1741.

26. " Frequency Division by Means of a Dynatron Oscillator " :G. Kahan, Wiadomosci i Prace Inst. Radmiech., Warsaw,No. 6, Vol. 3, 1931, pp. 78-79.

27. " An Oscillator having a Linear Operating Characteristic " :L. B. Arguimbau, Proc. I.R.E., Jan., 1933, Vol. 21, pp.14-28.

28. " Voltage Amplification with High Selectivity by Means of theDynatron Circuit " : F. M. Colebrook, W.E. & E.W., Feb.,1933, Vol. 10, pp. 69-73.

29. " The Generation of Polyphase Oscillations with the aid ofDynatron Circuits " : J. Groszkowski, L'Onde Electrique,Feb., 1933, Vol. 12, pp. 85-91.

30. " The Simplification of Accurate Measurement of RadioFrequency " : W. H. F. Griffiths, W.E. & E.W., May andJune, 1933, Vol. 10, pp. 239-247 and 299-306.

31. " Theoretical and Experimental Studies on the FrequencyStabilisation of the Dynatron type Generator " : T. Hayasi,Jour. I.E.E. Japan, May, 1933, Vol. 53 [No. 51, No. 538,English summary, p. 35, Japanese paper, pp. 389-394.

32. "The Interdependence of Frequency Variation and Harmonic.Content, and the Problem of Constant -Frequency Oscil-lators " : J. Groszkowski, Proc. I.R.E., July, 1933, Vol.21, pp. 958-981.

33. Effect of Circuit Parameters of the Constancy of the Fre-quency of a Phodynatron " : W. C. Sears, Physics, July ,1933, pp. 241-245.

34. " The Effect of the Curvature of the Characteristic on theFrequency of the Dynatron Generator " : E. B. Moullin,Jour. I.E.E., Aug., 1933, Vol. 73, pp. 186-195.

35. " The Inter -Electrode Capacitance of the Dynatron " :G. B. Baker, Jour. I.E.E., Aug., 1933, Vol. 73. pp. 196-203.

FR4,11N6WHEEL

KERR Cal,

MIRRORDRUM

MOTOR---.FIXED

RESISTANCE

DI -CONVEX LENS

tAMP TRANSFORMER

B CI( MIRROR

A New Television ReceiverANEW television receiver on the

Baird mirror -drum principle hasrecently been produced by

Bush Radio, and the accompanyingillustration shows the general con-struction of the television projectorunit, the image being reflected from therotating mirror -drum on to a screengin. x 4in. area.

With the present B.B.C. broadcast-ing carried out on 3o line i24 -picturesystem, the mirror -drum method isprobably as satisfactory as any other ;it is only when greater definition isrequired that the advantages ofcathode ray systems are appreciable.

With this new Bush Radio receiverthe programme can be followed withease, and definition is sufficient to makewell-known faces readily recognisable.A feature of the design is the simplicityof the synchronising system. Inoperation it is not unusual to be ableto leave the apparatus without atten-tion to the synchronising controlsthroughout the full period of thepresent B.B.C. half-hour transmissions.


S. -G. Valve as Superhet Detector*By C. B. Fisher

(Engineering Department, Northern Electric Company, Montreal)

THE following notes on the use of ascreen -grid valve as a first, or hetero-dyne, detector in a superheterodyne

receiver are intended as an amplificationof an article by E. C. L. White, entitled" The Screen -Grid Valve as a Frequency -Changer in the Super -Het." which appearedin The Wireless Engineer, November, 1932.

White discusses the circuit in whichthe local oscillations are introduced intothe anode circuit. This is, of course, similarto the usual anode modulation circuit,but the conditions for modulation aredifferent to those with which we are morefamiliar. White points out first that modula-tion can take place even when the anode -voltage, anode -current, and grid -voltageanode -current relations are both linear.Further, from measurements on a particularvalve he shows that maximum efficiency isnot obtained when the valve is biased tothe knee of the grid -voltage anode -currentcurve. The analysis of the general caseestablishes these results, except that modu-lation is shown to be due to the variation ofthe plate resistance with grid -voltage, andnot to variation of p. as stated by White,and shows other desirable operating con-ditions.

Let the variable part of the anode -current,ia, over the operating range, depend in themost general way on the variable part of theanode voltage, p, and the variable partof the control -grid voltage, g.ia = aip + ag .. . . (1)

+ (gap + a4g2 + a5gp

± a6p3 + a7g3 + asp2g + a9pg2

+ autP4 + eine + tai2P3g + ai3P2g2 ± ai4Pe

± The a's are constants for the particular

valve under discussion, and depend ingeneral on the fixed values of grid voltage

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

and anode voltage about which g and p arevariations.

For the sake of ease of manipulation andalso because as we shall see higher orderterms are not important in practical cases,we shall consider only the terms of powerfour or less. Let the voltage introducedin the grid circuit be a sine wave of radianvelocity w, and amplitude A modulatedby a second sine wave of radian velocityw2, the depth of modulation being M. Then

g = A cos wit(r M cos w2t) .. (2)

Let the local oscillator voltage have aradian velocity w3, and amplitude B. Then

p = B cos wit (3)

No general phase relations have beenassumed between the three sine waves, asit can be shown that this is not necessaryto obtain general results. By insertingrelations (2) and (3) in equation (I) andexpanding in first power cosine terms, thecurrent which is passed on to the inter-mediate -frequency amplifier from the detectormay be determined. The only terms ofinterest are those made up of expressionsof the form cos (w, - w, nw2)t where Iris zero or an integer.

These expressions represent the carrierand sidebands at the intermediate frequency.Performing the indicated operations wefind:-is= AB cos (w1 - w3)1[ia, 3I8a12B2

318ai4A2(1 +33122

-2

AB [cos (w, - w3 w2)t

+ cos (w1 - w, - w2)t][1a, 3/8a,2/32

2+3 a14A2(1 11/1-4)1

+ -9-2 a14M2A3B [cos (w1 - 2W2)t3

+ cos (WI - Wg - 2W2)t]

October, 1933

34 a14M3A3B [cos (w1 - 3w2)/o

-+ cos (wi - ws - 3w2)fl (4)The first term represents the carrier at

the intermediate frequency, the secondterm the upper and lower sidebands dueto the modulation frequency, and the thirdand fourth terms represent respectively thesidebands which contain second and thirdharmonics of the modulation frequency.An examination of equation (4) yields thefollowing information :-

I. There is an increase in depth of modula-tion through the first detector when a14 ispositive. This will result in distortionupon demodulation by the second detectorwhen the apparent modulation is greaterthan unity. That is, approximately when

M[la,± 83/8a12B2 ± a,..A2 (1 + M2

4)1

is greater than

- a5 3/8a12B2 3/8a14A2(12

32

Such will be the case only when M isnear unity, as both a and A2 are ordinarilysmall. It is not an important source ofdistortion.

2. There are present in the output, side -bands involving the second and third har-monics of the modulation frequency. Therewill ordinarily he sufficient carrier presentto demodulate these terms, but they willin any case be small as they have as a factorAs, ordinarily small in comparison with AB.

3. It is to be noted that the odd -poweredterms of equation (I) contribute nothing tothe modulation products delivered to theintermediate -frequency amplifier. The termsof power two contribute no distortion, thisis caused by terms of power four or greater.Thus if the valve characteristics are suchthat the family of anode -voltage anode -current curves and also the grid -voltageanode -current curves are all parabolas orcubics, the first detector will be distortion -less. Note also as a matter of interest thatif all the a's except al, a2 and a5 are zero,distortionless modulation will result, eventhough the anode -current anode -voltageand anode -current grid -voltage relationsare then everywhere linear. The character-istic surface becomes a twisted plane. Ifthe characteristic surface is a true planethen a5 = o, and no modulation takes place.


4. Assuming that the shape of the char-acteristic does not change appreciably dueto changing inputs, least distortion willresult with as small a signal voltage, and asgreat a local oscillator voltage as possible.

5. For maximum gain the local oscillatorvoltage should be high, and the grid shouldbe biased to the point at which a5 is amaximum, assuming that a14 is small. For adiscussion of the evaluation of a5 see Petersonand Evans, " Modulation in Vacuum TubesUsed as Amplifier," Bell System TechnicalJournal, July, 1927. Their parameter b11,corresponds to a5 and is shown to beequal to

1R0bEdE or Ro2

Where E, and E, are the anode and gridvoltages respectively, Ia is the anodecurrent, and Ro is the anode impedance,all evaluated at the operating point. Thusthe best condition for modulation does notrequire the amplification factor of the valveto vary, so that the application of the circuitis not limited to screen -grid valves. Theabove analysis however, applies exactlyonly where the load impedance is smallcompared to the anode impedances of thevalve, and this will ordinarily be the caseonly with screen -grid or pentode valves.

These results indicate that the firstdetector can be expected to be free fromdistortion, and provide a means for fmdingthe best operating conditions. The analysis,it should be noted, also applies directly tothe rather important case in which the localoscillations are introduced into the circuitof a second grid of the heterodyne -detectorvalve, instead of the anode. A completeanalysis would include a consideration ofthe effects of the load impedance on theoperation of the heterodyne detector. By acomparison of equation (1) of the presentdiscussion with equation (3) of Petersonand Evans' paper, already mentioned, thecorresponding tube parameters can bewritten down. Then by a substitution intoequations (6) and (4) of their paper, theexpression for the anode current can beobtained, and broken down to give the modu-lation products in the usual manner. Theprocedure for a generalised load impedanceis given by F. B. Llewellyn, " Operation ofThermionic Vacuum Tube Circuits," BellSystem Technical Journal, July, 1926.


Olympia, 1933A Review of Exhibits of Technical Interest

Annual Radio ShowIN this year's receivers the emphasis is on

foreign station reception, as exemplifiedby the use of the superheterodyne prin-

ciple to secure selectivity, automatic volumecontrol to counteract fading, and variousdevices to combat " man-made static."

In the early days of the superheterodyneits chief claim was that of sensitivity ; butthe problem of selectivity has recently be-come so acute that it is this quality which isresponsible for its present popularity. Almostevery receiver with a pretence to long-distance reception is of this type, althoughat least three of them have no I.F. amplifierbut merely I.F. tuned circuits to provide thedesired selectivity. An example of thispractice is to be found in the Kolster BrandesModel 444 A.C. mains superheterodyne,which employs only three valves in thereceiver proper-one of the new H.F.pentodes as combined oscillator and firstdetector, a similar valve as second detectorand a power pentode output valve. It mayseem strange to have associated " man-madestatic " with long-distance, reception in theK.B. range of sets ; but the increasing use ofelectricity for domestic and other purposesis making this type of interference morecommon, and the increasing sensitivity ofthe modern receiver is responsible for makingit more noticeable.

Battery sets have retained their popu-larity, thanks to the new economical outputstages such as Class " B " and quiescentpush-pull, which will give an output of theorder of one watt of undistorted A.C. on fullmodulation, while on a normal programmethe H.T. consumption averages only 7milliamperes at i5o volts.

The majority of receivers are still describedby the number of valves employed, includingthe H.T. rectifier in mains receivers, butE. K. Cole intend to classify theirs accordingto the number of " stages " employed. Onecomplication is that to fall in line withcircuits employing diode detection the triodedetector should be given credit for its twofunctions of rectification and amplification,and classed as two stages. Other difficulties

at the

are special A.V.C. valves, and the oscillatorof the superheterodyne. For, if one is con-sidering the " quantitative " performanceof the receiver, i.e., the number of stationsreceived, it seems reasonable to count onlythose stages which are in the direct chainbetween aerial and loud speaker ; but if oneconsiders it " qualitatively," e.g., freedomfrom fading, reliability of oscillator per-formance, these extra valves must be in-cluded, and in one or two receivers there isa battery -economy rectifier connected to theoutput valve which would demand con-sideration. However so little informationis conveyed by the bare number of " valves "or " stages " that the particular nomen-clature adopted is probably immaterialprovided that a standardised convention isadopted by all manufacturers.

Midget and Car ReceiversMidget sets are being made by Sunbeam

Electric, Ltd., but the most attractive fea-ture incorporated in these receivers, universaloperation from A.C. or D.C. mains, isavailable also in normal sized receivers madeby Cromwell Ltd. and many other manu-facturers. The midget receiver has beenadapted to the needs of the motorist byfitting it in a weatherproof metal case whichmay be mounted in any convenient position,the controls being extended by Bowden wireto a small panel mounted on the steeringcolumn. H.T. current is usually derivedfrom the car accumulator by means of arotary converter, for which purpose therehas been added to the M.L. series a newtotally enclosed model giving 40 milliamperesat 220 volts. Lissen's preliminary model,however, is a 6 -valve battery superheterodynewith Class " B " output deriving its L.T.supply from the car accumulator, but itsH.T. current from a dry battery. Anotherfeature of this receiver is that it is one of thevery few in the show employing the newpowdered iron cored coils in the I.F. tunedcircuits as well as in the signal -frequencycircuits. Car receivers work from an aerialwire concealed in the roof, and automatic

C 2

October, 1933 544

volume control is essential since localscreening by buildings, etc., is liable to causerapid variations in signal strength while thecar is in motion. A further requirementis the silencing of the car's electrical equip-ment, and amongst the various resistanceand capacity filters there are special heat-proof resistors for fitting at the sparkingplug terminals.

High -frequency AmplificationHistory shows a steady rise in the A.C.

resistance of the high -frequency amplifyingvalve, from the triode of some 30,000 ohmsthat was used in a neutralised circuit to thescreened grid valve averaging a third to ahalf megohm, and now to the screenedpentode whose anode resistance is of theorder of one megohm. Since the dynamicresistance of the ordinary tuned circuit ismuch less than the anode resistance of thescreened grid valve, one would expect thescreened pentode to bring little increase inamplification beyond that due to its slightlybetter mutual conductance. But in anumber of receivers tuned circuits of highdynamic resistance are secured by operating


coils, has a potentiometer ganged to thetuning control which, by varying the biason one of the H.F. valves, keeps the receiver

Ferranti " Arcadia" receiver.

just within the limits of stability throughoutthe wave -band, while several of the smallsuperheterodynes have pre-set regenerationfrom the second detector on the I.F: circuits.When the screened pentode is applied tosuch receivers it should give a twofold gain :for its lower anode to grid capacity permitsthe dynamic resistance of the anode circuit

to be increased without

Lissen car radio receiver and control panel.

the radio -frequency amplifier with a con-trolled amount of regeneration. For ex-ample, the Philips 634 " Superinductance "receiver, in addition to exceptionally good

causing instability, andgiven a sufficiently highvalue of anode load itsgreater amplification factorbecomes useful. Thesevalves are available in boththe normal and the variable -mu types, and the pentodeconstruction enables themto handle without distortiona much larger anode vol-tage swing than wouldbe permissible with atetrode of similar imped-

ance, a feature which is important in receiversdesigned for a very large detector input.The screened pentode with a low voltage onthe screen is an excellent anode bend rectifier.


(The screen voltage required is so low thatin the Atlas A4 receiver the grid bias for the3 -watt power pentode output valve is usedas screen voltage for the pentode detectorpreceding it). Consequently it is frequentlyused as combined oscillator and first detectorin superheterodynes, and with this in viewit is available with a 7 -pin base in whichboth the suppressor grid and the metallisingof the bulb are brought out to independentpins.

The first entirely satisfactory single -valvefrequency changer was the type known inAmerica as the "Pentagrid "* but nowmanufactured by Ferranti who call it the" Heptode." Its electrodes are arranged toact as two valves in series, the oscillatormodulating the flow of electrons into thedetector portion, while the two sets ofelectrodes are screened from each other toprevent interaction between the oscillatorand signal -frequency tuned circuits. A greatadvantage of the Heptode is that the de-

Ekco Model AG 74 in a black bakelite cabinetwith chromium insets.

tector portion has variable -mu character-istics, so that it is the only single -valve fre-quency changer that can be used for volume

* "(The Pentagrid Converter," C. L. Lyons,B.Sc., The Wireless Engineer, July, 5933, P 364.

October, 5933

control. Thus the Ferranti " Arcadia "superheterodynes are fitted with A.V.C.,although they have only two valves before

R1

TO L.F.AMPLIFIER

R2

- H.T.

Fig. r.-Rectification and delayed A.V.C. withtwo " Westectors."

1

the second detector, the Heptode and oneI.F. valve. Until this type of valve is moregenerally available most of the largersuperheterodynes are using separate oscil-lators, and an .unusual example is the CossorA.C. mains receiver, which employs a powervalve (4i MP) as oscillator for the sake ofconstancy of output throughout the wave-band.

Automatic Volume ControlThe simplest form of A.V.C. obtains its

bias voltage by rectifying the carrier voltageat the detector stage, and the valve largelyemployed in this arrangement has been thedouble -diode -triode in which one diode canprovide delayed A.V.C. while the other diodeprovides the A.F. output to feed to the triodepart of the valve. In many popularsuperheterodyne receivers employing thismethod the A.V.C. diode is fed direct fromthe anode of the I.F. valve, i.e., from theprimary of the last I.F. transformer ; thisavoids putting the load of two rectifiers on the.secondary, and presumably aids in balancingprimary and secondary to give a symmetricaltuning curve. The slight loss of selectivitydue to the omission of one tuned circuit isimmaterial to the working of the A.V.C.The objection to this form of A.V.C. is thatthe triode portion of the double -diode-

ill

October, 1933

triode requires only a small A.F. input (say3 volts peak) to feed an output stage of twoor three watts, so that the carrier voltageis strictly limited. But this year a metalrectifier, known as the " Westector," isavailable, using the same copper oxideelements as the Westinghouse power rectifiersbut of such small size that the inter -electrodecapacity no longer prevents its use on radiofrequencies. There are both half -wave andwhole -wave types in two sizes, to withstanda maximum radio -frequency input of either24 or 36 volts peak, and in most cases theaudio -frequency output would be sufficientto feed the output valve, using a step-uptransformer if necessary, without an inter-mediate A.F. valve. The circuit (Fig. 1)shows how A.V.C. can be obtained from thesedetectors by tapping off the voltage at thepoint X due to the rectified carrier ; delayedA.V.C. is obtained with the aid of a second

Underside of Marconi 7 -valve chassis showingsize of " Westectors."


rectifier R2 which has a suitable polarisingpotential from the potentiometer P so thatno current can pass R2 until the potentialat X exceeds a certain value.

The Marconi 7 -valve superheterodyne em-ploys this principle for its second detectorwith delayed A.V.C., though an intermediateA.F. stage is used be-fore the output valve.To illustrate the useof metal rectifiers,Westinghouse ha v e o 0001

published a design foran A.C. mains super-heterodyne in whichthe Westector feedsdirectly a 2i - wattpentode output valve ;on gramophone repro-duction an I.F. valveis used to provide thenecessary A.F. am-plification.

As an alternative to the very large inputdetector there is amplified A.V.C. in whichan additional valve is employed as D.C.amplifier in order to obtain the necessarylarge bias voltage from a small detectorvoltage. An ingenious type of amplifiedA.V.C. for use with triode detectors hasbeen developed by Westinghouse, Theradio -frequency component of anode current,which with an ordinary valve detector isby-passed to earth through a small condenser,is rectified by a Westector and by using thevoltage -doubler circuit shown in Fig. 2a large A.V.C. voltage can be obtained fromthe average detector. The only objectionto this use of the detector as a radio -frequencyamplifier is that the increased radio -frequencyimpedance of the anode circuit will result ina greater negative feed back to the gridof the valve (unless it is of the screenedgrid type) and will increase the total anodevoltage swing. However the Varley andWearite units embodying this principle areundoubtedly the most practicable methodof adding A.V.C. to existing receivers.

A possible objection to A.V.C. was thatsince the receiver automatically attainsits maximum sensitivity in the absenceof any signal, background noises mightbe unpleasantly prominent during tuning.In a number of sets this problem has beenmet by installing a sensitivity limitingdevice which may be set according to localconditions and when desired may be switchedout of action altogether. But in the moreambitious receivers silent tuning is secured

H.F C

mod

A.V.C.

Fig. 2.-AA.V.C. from atriode detector.


by using an intermediate A.F. stage whosenegative bias is controlled so that in theabsence of signals above a certain strengthno anode current can flow. In the Brunswickand R.G.D. models the grid bias is reducedon the arrival of a signal by means of arelay connected to an additional rectifierfed from the I.F. amplifier ; but whatevermethod is adopted the bias must be reducedsuddenly by a considerable amount orelse the A.F. valve would be operated onthe curved portion of its characteristic.Thus in the Kolster Brandes 888 receiver

547

Control panel of R.G.D. Model 1201 showingvisual tuning indicator.

there is a single diode tetrode valve of whichthe tetrode part amplifies the A.V.C. voltagewhile the diode provides positive bias tofree the A.F. valve, but has a suitabledelay potential applied to it so that nothinghappens until the incoming signal is sufficientto free the A.F. valve completely. TheAlba 7 -valve receiver, however, employsno such extra valve but merely a double -diode -triode second detector, of which onediode provides delayed A.V.C. while thesecond gives the A.F. output and the steadyvoltage derived from it controls the biasof the intermediate A.F. valve. It is im-possible to have a delay on this diode,which is also providing the A.F. signals,so there is a special scheme of interactionbetween the anode current of the controlledA.F. valve and the delayed A.V.C. whichhas the desired effect of putting the valvedefinitely in or out of action at a criticalsignal strength. A certain amount ofbacklash is desirable in the operation of allthese systems, so that if a station has once

October, 1933

come in it can be held, even though the signalstrength at the detector may vary slightly.

It is sometimes claimed that by makingtuning more critical these quiet A.V.C.systems render visual tuning indicatorsunnecessary, but the R.G.D. model 1201has both. The most popular type of in-dicator is a meter in the anode circuit ofone of the controlled valves ; as the detectorinput rises A.V.C. increases the negativebias on the radio -frequency valves, so thataccurate tuning produces the minimum anodecurrent. For a direct indication of signalstrength Standard Telephones have a minia-ture cathode ray tube, called the " MicromeshTunograph," whose deflecting electrodesabsorb hardly any power so that they maybe connected across any circuit ; on A.C. aline will appear on the screen whose lengthis proportional tothe amplitude ofthe alternatingpotential ap-plied. (The staticsensitivity of thistube is i cm. for13 volts.)

The use ofA.V.C. has re-sulted in twoconflicting ten-dencies on theA.F. side of thereceiver. On theone hand the de-tectors normallyused, namely,double - diode -triode or Westec-tor, are capableof loading theoutput valve di-rectly, as in theFerranti mainsreceivers, all ofwhich employ a

The " MicromeshTunograph," acathode ray tuning

indicator.

double -diode -triode second detector coupledto a 24 -watt triode output valve. On theother hand silent tuning A.V.C. demandsthe use of an intermediate A.F. stage for

October, 1933 548

noise suppression. A possible solution ofthe problem would appear to be the use ofthe relay type of control to short-circuitthe A.F. signals at some point instead ofcontrolling the bias of an additional valve.

" Iron " Cored Tuning CoilsAnother development in radio -frequency

technique is the introduction of powderediron cores for tuning coils. The first materialto appear commercially was Ferrocart,* inwhich the particles are aligned by a specialmagnetic method and separated into layersby means of paper interleaving. Coils usingthis material are made by Colvern, but bynow almost all coil manufacturers haveproduced their own core material, consistingof some particular magnetic alloy finelypowdered and compressed with special in-sulating and binding materials. These coilsare very compact and suffer less loss whenscreened, thanks to the use of a closedmagnetic circuit in some cases, but it isdoubtful whether any of those produced sofar have a lower resistance than could beobtained from a rather more bulky air -coredcoil. With a high permeability core muchless wire is required for a given inductance,so that many of the medium wavewound with Litz wire. In addition the R.I." Micrion " coils have an adjustable corecontrolled by a screw so that if a single coilin an existing receiver is replaced the Micrioncoil can be adjusted to the same inductanceas the old one ; but in constructing a newreceiver with several ganged circuits it wouldbe preferable to have coils accuratelymatched by the manufacturer. It is interest-ing that Igranic are using their core materialeven down to 13 metres, the advantagesclaimed on short waves being reductionof external field, and better control ofregeneration.

Complete tuning units including a set of" iron " cored coils and a gang condenserare available, and Colvern and Lissen have" iron " cored coils for I.F. amplifiers. Sinceseveral receiver manufacturers are at presentusing air -cored Litz windings for inter-mediates, there should be scope for the newcoils here, particularly as the core losses willbe even lower at the intermediate frequencythan at the higher frequencies.

* Description of Ferrocart in The WirelessEngineer, April, 1933, p. 183.


A sequel to the development of these ironpowder cores is permeability tuning, inwhich the capacity of the tuned circuitremains constant while the inductance isvaried by sliding a core of fairly high per-meability in or out of the coil former. Thedifficulty has been to obtain a sufficientlyhigh ratio of maximum to minimum induct-

Varley permeability tuner.

ance and at the same time to avoid makingthe motion of the core too critical. SovereignProducts have succeeded in producing singletuners, and Varley have an experimental3 -gang model in which there are actuallytwo windings for each medium -wave coillinked together. With this type of tunerone has the advantage of a maximum L/Cratio throughout the wave -band, and thecore losses help to keep the selectivityreasonably constant.

Audio -frequency AmplifiersThe latest luxury is a tone -and -volume

compensator. It is well known that as thevolume of sound is decreased the human earbecomes less sensitive to the extreme highand low frequencies, so that reproductionat a low volume level appears to have anexcess of the middle frequencies ; so in somereceivers an A.F. filter is connected acrosspart of the post -detector volume controlpotentiometer so as to reduce the responseto the middle register when volume is de-creased. Another type of tone correction isillustrated by the Varley " compensatingresistance coupling unit." This has aninductance between the coupling condenserand the following valve which resonateswith the input capacity of the valve at a


frequency of the order of 5,000 c/s and thuscompensates for side -band cutting in theR.F. stages of the receiver.* The cut-offfrequency can be lowered if desired byshunting a minute capacity across the inputof the following valve. In the Marconi 7 -valve receiver there is a tone control whichhas a dearly defined normal position, givingas nearly as possible linear response, and avariable attenuation of either high or lownotes which is worked in conjunction withthe A.F. transformer following the detector.The Marconi 5 -valve receiver, on the otherhand, has fixed tone compensation in theanode circuit of thedetector valve usingthe circuit shown inFig. 3, which producesa cut-off at 5,50o c/sand a bass responsewith a maximum atabout 8o c/s.

The R.G.D. model1201 radiogram is alsonoteworthy, since ituses the paraphase sys-tem of A.F. amplifica-tion which aims atperfect reproduction oftransients by the use Fig. 3.-Tone correctionof resistance coupling in the Marconi 5 -valvethroughout and at the receiver.same time securesthe advantages of push-pull amplification.

The biggest development in A.F. amplifica-tion has been the perfection of H.T. economycircuits for battery receivers. First camethe Quiescent Push -Pull system, which canbe very efficient but necessitates carefulmatching of the valves used. For this pur-pose the Marconi and Osram small pentodesare now graded and marked with a code letterso that matching of any two specimens canbe obtained simply by applying to each thescreen voltage corresponding to its letter.(On Columbia receivers this is known as" Constant Quality Amplification," whileon the Marconi range it is called the" Parallel Conductance Principle.") Themore popular output circuit, however, isthe Class " B " in which a double valve isemployed so that the matching has alreadybeen secured in the manufacture of the valve.

GRIDBIAS

* The Wireless Engineer, July, 1933, p. 370(Correspondence).

October, 1933

Like Q.P.P., the Class " B " stage requiresa high impedance anode load and matchingto the loud speaker is more critical thanwith a single triode. To meet this difficultythe majority of loud -speaker manufacturersare producing a Class " B " stage and movingcoil speaker as a complete unit for con-verting existing receivers, the previous smalloutput valve being over -biased and used asdriver to the Class " B " stage. The Class" B " principle has, of course, restored theintermediate A.F. valve under the title of" driver valve." The most popular valveshere are the small power type, such as theOsram L.P.2, and the L.F. valve such asthe Mullard P.M.2DX, but in the Cossorreceivers the driver valve is a small pentode(22oHPT) which is slightly more economicalin anode current.

Although the Class " B " output stagewas introduced as a means of bringing goodquality reproduction within the capabilitiesof the dry H.T. battery, there are now a largenumber of mains units on the market whichhave been specially designed to give suffi-ciently good regulation for use with Class" B " receivers. The Heayberd model em-ploys a Cossor neon tube regulator, and infavour of this system it must be mentionedthat the regulator tube contributes a greatdeal to the smoothing ; but the majority ofunits rely upon a metal rectifier (the Westing-house H.T.13 has been introduced for thispurpose) and low -resistance smoothing cir-cuits. In theory the ideal source of currentfor these special output stages would be theH.T. accumulator, owing to its very lowinternal resistance ; where reliable chargingfacilities are available for the proper careof the battery, the new and very compact" Block " H.T. accumulator might be aconvenient source of current.

The modern receiver is apt to present anawkward problem when an extension loudspeaker is required ; but for use whereterminals are provided across the secondaryof the output transformer special units withlow -impedance speech coils and no inputtransformer are available. It seems now tobe the general principle for the externalspeaker terminals to be on the output side ofa transformer, and the R.I. Radiogram isexceptionally well equipped in having anentirely separate transformer to feed theexternal speaker.

October, 1933 550

For those who require a larger outputthan is obtainable from the domestic receiver,there is a whole range of Tannoy amplifiersand speakers, including the " Lantern "speaker which is designed to radiate thesound in all directions as uniformly aspossible. The amplifiers all have an inputimpedance of 600 ohms, and the Tannoyradiogram chassis has an output matchedto 600 ohms, so that it can be connectedimmediately to a power amplifier of anydesired capacity.

Short-wave ApparatusThe simplest method of short-wave

reception employs a " converter " for attach-ment to an ordinary broadcast receiver.This usually employs two valves, a screenedgrid valve as " separator " between theaerial and the tuned circuit feeding thesecond valve, which is an autodyne frequencychanger. The " separator " valve avoidstrouble from aerial resonances, and makesthe calibration of the tuning dial possible ;the regeneration ofthe autodyne valvecan be fixed, withpossibly a com-pensating deviceganged to the tun-ing condenser, sothat the receiveris completely con-trolled by a singletuning dial. Thereare one or two all -wave tuning unitsconsisting of a setof coils coveringthe most popularwave - bands be-tween about 15and 2,000 metres

The Baird Kerrcell.

controlled by a selector switch ; but thesewould obviously be very difficult to applyto a modern receiver employing severalganged tuning circuits. For the ultra -shortsthere is an Eddystone 5 -metre transmitterwhich may be suspended in the air to avoidthe use of long aerial feeders.


TelevisionIn television receivers the perforated

scanning disc has now been replaced by themirror drum ; this is natural since in placeof the neon tube which gave a large area

A new type of valve bulb designed to giveincreased rigidity of electrodes.

source of modulated light, a Kerr cell isused which has only a small aperture. TheBaird " Grid Cell Unit " is a convenientmounting carrying a bunched filament lamp,a pair of Nicoll prisms, and a Kerr cell inwhich each electrode consists of a numberof foils, the two sets of foils being inter-leaved like the plates of a condenser. Thiscell, which can be purchased separately andhas a clear aperture o.rin. square, workswith a polarising potential of about 425 voltsand a signal potential of 125 volts peak forfull modulation. The mechanical methodsof scanning seem quite adequate at present,but if ioo line scanning of television imagesis ever adopted in place of the present 3o,the cathode ray tube will undoubtedly bethe ideal receiver, and several firms arepreparing tubes suitable for television.

ValvesA drastic change in valve manufacture

appears in the Marconi and Osram " Catkin "valves, which have no glass bulbs. Theuse of the anode as the envelope of thevalve facilitates cooling, and another featureof these valves is that exceptional rigidityof the electrode system has been securedby the use of insulating distance piecesbetween its elements. Another developmentis the use of the American type S.T.12 bulbwhich is shaped at the top to receive a micabridge. This gives additional support to


the electrodes, but the shape is not sofavourable to the mechanical strength ofthe bulb. To facilitate the dissipation ofheat from the anodes of rectifying and powervalves it is now a common practice toblacken this electrode ; and the next stepis to blacken the inside of the bulb to preventthe glass from reflecting back the radiation.

The new types of valve are the screenedhigh -frequency pentode, the heptode, Class" B," and the various diode -cum -amplifierdetectors. Of the latter the commonestis the indirectly heated double -diode -triode ;but Mullard have also an indirectly heatedsingle-diode-tetrode, while Mazda have inaddition to the mains type a battery double -diode -triode which is used in the Portadynebattery superheterodyne. There are stilllarger pentode output valves this year,the Mullard Pen 4VA, Marconi PT i6 andPT 25, and Mazda AC 2 PEN. The latterhas the exceptionally high mutual con-ductance of 8 milliamps. per volt and givesits 32 -watts A.C. output for so moderatean input that it has also been producedin double -diode form (AC 2 PEN DD) for

Kolster Brandes" Rejectostatic "

units.

use as a combined detector -output valvewhich practically abolishes the A.F. amplifier.

AccessoriesAn important step towards eliminating

electrical interference is to erect the aerial

October, 1933

clear of house wiring, etc., and employ ascreened down -lead. For this purposespecial low -capacity screened wire has beenproduced, but even the best ofcable must intro-duce appreciableloss in a high im-pedance circuit, sothat where verylong screened leadsare required aradio -frequencytransformer shouldbe used at eachend of the lead sothat the input canbe stepped downto match a low -impedance trans-mission line andthen stepped upagain to the re-ceiver input.

" Hivac " quartzoscillator mounted

in vacuo.

screened

Kolster Brandes have now succeeded in pro-ducing transformer units which will functionon both the ordinary broadcast wave -bandswithout switching ; with the aid of these" Rejectostatic" units, as they are called, it ispossible to use hundreds of feet of ordinarylead -covered cable between aerial and receiverwithout the losses becoming prohibitive.

An interesting exhibit by the HighVacuum Valve Co. was of quartz crystalsmounted in vacuo. These can be standardisedif required to an accuracy of i in Io,000 atany frequency between 600 and 8,000 kc/s,and oscillators of lower frequency willshortly be available.

Passing from standards to testing instru-ments, we come to the " Radiolab " Valveand Set Tester, which employs a high-grade moving -coil meter to measure current,voltage, and, with the aid of a small battery,resistance as well. A buzzer providing atest signal on about 450 metres will be avery useful feature when checking receivers,and when dealing with power equipmentA.C. voltages up to i,000 can be measured,using the same meter with a metal rectifier.


Measuring Resistances*By the late W. A. Barclay, M.A.

IT is well known that a high resistancegalvanometer or voltmeter may be usedto measure resistance by noting the

change in the reading when the resistanceis connected in series with the instrument.Let us suppose that the maximum scalereading of the instrument-volts or milli-amps-is the number N, and that the meterresistance is Rm ohms. (In the case of amilliammeter, R,, is, of course, the value ofthe known series resistance which we regardas included in the instrument.) Then if,by means of a battery or other suitabledevice, a steady E.M.F. of E volts is im-pressed upon the terminals of the instru-ment, a current of EIR,,, amps. will flow,causing the needle of the meter to registera value n1 on the scale.

VALUES OF n1

Fig. 1.-Having found the position of the pivotpoint P on the nl - R, network, corresponding valuesof 122 and R are in alignment with it.

Next, let the resistance R, whose value issought, be inserted in the circuit, and letn2 be the new reading on the scale of theinstrument.

Then

and

ni Rm+ Rn2 4, .. (I)

R --= (111 n . R , . . . . (2)n2

* MS. received by_the Editor July, 1932.

The importance of this result lies in thefact that it does not really matter what thereadings n1 and n2 represent ; as we haveseen, it is immaterial whether they are involts or milliamps. ; for our purpose theyare simply scale readings and nothing more.The value of R.,, must, of course, be asaccurately known as possible, and whereexternal series resistances are employedthese should be of " precision " type.

When several values of resistance haveto be estimated by this method., the calcula-tion required by equation (2) becomestedious, and much labour may be saved bythe use of the two " four -variable " align-ment charts now presented. The methodof use will be apparent from the schematicdiagram of Fig. i. The lower left-handportion of the diagram consists of a networkof slanting and vertical lines, correspondingrespectively to values of meter resistanceRm and scale reading n1. Above this is adiagonal scale carrying values of n2, whileto the right is a vertical scale containingvalues of the unknown resistance, R. Theprocedure is to find a point P on the networkcorresponding to the scale reading n, andthe meter resistance Rm. Since n, dependsonly on the meter and the battery, the posi-tion of this point P is thus fixed for allsubsequent work, and can thus be used asa pivot about which the index line con-necting n2 with R may be swung. Theconnection between the unknown resistancevalue R and the second scale reading n2is thus exhibited with great facility. It willbe further evident that if the value of yetanother resistance is required, all that isneedful is the single observation of n2obtained with this resistance in the circuit.The desired value of R is then immediatelyobtained in alignment with n, and the fixedpivot point previously established.

The first Chart (Fig. 2) is designed for amaximum scale reading N = 6, while theresistance values run roughly between I,000and io,000 ohms. It is thus suitable foruse with the average type of medium resist-ance voltmeter. The second Chart (Fig. 3)is intended for use with an instrument of


maximum scale reading N = 25o, while volt. Then the meter resistance will bethe range of resistances catered for is higher, 25o x 32o, or 8o,000 ohms. Then, if asay from 25,00o to 200,000 ohms. battery of 210 volts be employed, the posi-

6

FIG. 2.

-1,00o

-1,100

-1,200

-1,300

-1,400

-1,500

'0

3 <c` -2,0000

-Z500

-3,000

-4,000

-5,000

6 5 4 3

VALUES OF n1

As an example of the use of the latter tion of the pivot point P will correspondChart, let us suppose that we are using a to n1 = 210, R. = 8o,000. If, now, the250 -volt meter rated at 32o ohms per reading n2 consequent upon the insertion

October, 1933

of the unknown resistance is 135, we shallfind on joining P to 135 on the diagonalscale that the required value of R is 44,000ohms.

250

250500,000400,000300,000

200,00

100,00

Ecc 50,0000cs)

40,000

30,000

200

200

160

150


shown upon either of the present Charts.In such a case the remedy is very simple.Select an arbitrary factor k, integral orfractional, such that the value kR, is con -

FIG. 3.

100

100

250 200 150 100

50

50

50

- 25,000

- 30,000

- 40,000

50,000

80,000

100,000

- 150,000

-- 200,000

- 300,000- 400,000- 500,000

VALUES OF n1

It may sometimes occur that the value of veniently found on one or other of thethe meter resistance, R is not conveniently charts. When the usual procedure is carried


out using kRm, the resulting resistance valuefound is kR, and this must therefore bedivided by k in order to obtain the desiredvalue of R. As an illustration, let us sup-pose that our meter resistance was 20,000ohms, while the readings n1 and n2 were 210and 135. By taking k = 4, the pivot pointfor use with Fig. 3 will be the same as thatin the above example, i.e., n1=-- 210,kR, = 8o,000. We must, however, be care-ful to divide the result, 44,000 ohms, bythe same factor, thus obtaining R = 11,000ohms. Furthermore, the scale readings nland n2 may themselves be multiplied by any

October, 1933

convenient factors when this may be neces-sary, so that, e.g., the 6 -unit scale of Fig. 2may be used with instruments having scalesup to 12 units, etc. These circumstances,which follow from equation (2), make themethod one of extraordinary generality.

In conclusion, it is particularly to benoted that the internal resistance of thesource of voltage, E, must be negligiblysmall as compared with the meter resistance.This will generally be the case when drybatteries and accumulators are used inconjunction with instruments of relativelyhigh resistance.

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, The Wireless Engineer

SIR, --Professor Cramp is, I think, mistaken inaccusing me (in his letter in your August issue) ofsaying definitely that B and H are the same thing.All that I maintain is that in so far as all experi-mental laws are concerned they are identical andthat any interpretation of these laws which regardsthese quantities as being fundamentally different,involves conceptions which are additional to theexperiments. I do not even agree that it is per-missible to introduce the factors i/e and i/µ in thestatements of the inverse square laws. For,consider the proof of the inverse square law betweencharges by means of the concentric spheres. Theinverse square law that that experiment proves isthat the mechanical force exerted between twocharges is inversely proportional to the distanceapart, quite irrespective of whether the interveningspace is air or copper or brass ! The result can, ofcourse, be equally well deduced if the factor /e isincluded, but if so, to what does the apply-to the copper, the brass or the air ?

I am sorry I did not mention Professor Cramp byname in my last letter, but my remarks there wereintended to be as much a reply to his criticisms asto those of your other correspondents.

Finally, Sir, since I really must not occupy moreof your space on this subject, let me say that in myview, it is wholly a matter of choice which inter-pretation of these laws is adopted by any individual.Some prefer mathematical symmetry, even if theyhave to introduce additional conceptions to procureit ; some like to think in terms of mechanicalanalogy as Faraday and (according to his pro-fessions) Maxwell did ; others again, prefer thesimplest and most direct interpretation of all, therelevant experiments. Provided that the argumentsare strictly logical these three interpretations mustnecessarily come to the same thing in the end and

no one can say that one view is definitely right andthat another is definitely wrong. It is mainly todogmatic statements of that kind that I object.

C. L. FORTESCUE.City & Guilds (Engineering) College,

London, S.W.7.

High -frequency GeneratorsTo The Editor, The Wireless Engineer

SIR, On page 233 in the April, 1933, issue ofyour journal appeared Patent No. 381087, issuedto Cie. Generale de Telegraphie sans Fil under theconvention date (France) 9th March, 1931, coveringa high -frequency generator operated by making abeam of cathode-ray sweep a circular path witha certain base frequency over the periphery of atoothed disc.

In this connection you may be interested toknow that Mr. Mong-Kang Ts'en, a Fellow of thisInstitute, started in 193o, as a part of his workhere, to develop ultra -high frequency generatorson the principle that split -phase Barkhausen-Kurzoscillations of about one -meter wavelength areconnected to the two pairs of electrodes of acathode-ray tube in order that the beam of cathode-ray sweeps out a circle and falls alternately on theperiphery of a toothed disc and on a metallic ringplaced behind. This was first reported in theSecond Annual Report of Academia Sinica for theyear 5929-1930 (page 85) and again briefly describedin the Third Annual Report (page 73). Theoriginal text was in the Chinese language. AnEnglish version is appended herewith for yourreference.

It is evident that both Mr. Ts'en and the Cie.Generale de Telegraphie sans Fil have independentlyarrived at the same conception of sweeping a beamof cathode-ray over a toothed gear to multiply thefrequency, but as far as we can ascertain fromscientific publications, Mr. Ts'en seems to have

October, 1933

priority in conceiving the idea, working on it andgiving it publication. Experiments are still beingcarried on in the laboratory of this Institute.

Shanghai. S. L. TING,Director.

National Research Institute of Physics,Academia Sinica.

(3) Studies in High -frequency Generation.(Translated from the Chinese text, p. 73,

Third Annual Report of Academia Sinica.)As reported before, the devices used in high -

frequency generation include such as spark gaps,inverse -connected triodes or magnetron. Thewavelengths thus obtained vary from a few metersto a few centimeters. A Fellow of this Institute,Mr. Mong-Kang Ts'en, has proposed a new method,using short-wave magnetic or electric fields to makea beam of electrons emitted by the cathode of acathode-ray tube sweep out a circle, and inter-mittently receiving the beam by a toothed -disc.Theoretically, waves of a few centimeters' lengthcan be generated. Experimental work has alreadybeen started and is still in progress.

(7) Studies in High -frequency Generation.(Translated from the Chinese text, p. 85,

Second Annual Report of Academia Sinica.)The importance of high -frequency currents has

been stated before. They are usually generated bymeans of spark gaps or vacuum tube oscillators ;tube oscillators can ordinarily produce electricwaves of about one -meter wavelength, while four -centimeter waves have been obtained with sparkgaps. Besides studying the limitations of thesetwo methods, a new method of high -frequencygeneration is hereby proposed. Split -phase high -frequency 'oscillations from a tube generator areconnected to the two pairs of electrodes or the twocoils of a cathode-ray tube, making the cathode raysweep out a circle. The electrons in the ray areto be received by two metal discs placed one behindthe other ; a circular ring of holes are made on thefront disc so that the electrons, when meeting theholes, will pass through and will be received by thedisc at the back. These two plates will thus forma condenser, whose potential difference will varyat a frequency equal to that of the original oscilla-tions multiplied by the number of holes in the frontdisc. By means of this method it seems possibleto obtain waves shorter than 4 cm. But whetherthere is any practical difficulty will be decided byexperiment.

Book ReviewElectron Tubes.

By John H. Morecroft, Sc.D. Chapman andHall. 28s.

The applications of the electron tube, in whichterm the author includes photocells, gas -filled valves,cathode ray tubes, and thermionic vacuum tubes,


are by no means confined to radio transmissionand reception, but are finding increasing use inevery branch of industrial activity and scientificresearch. Their widest use is in the field of radio,of course, and it is only natural that the radioengineer should have the greatest knowledge oftheir operation. He is not necessarily the bestperson to judge their usefulness in other fields,however, for his outlook is restricted to the special-ised problems of radio. One of the chief purposesof this book, therefore, is to give a knowledge ofthe properties and applications of electron tubes toengineers in general and not the radio engineer inparticular. To the latter, much of the groundcovered is familiar, but the book is by no meansvalueless. To the former, it should prove in-valuable.

The book opens by discussing the structure ofthe atom, conductors and insulators, and the variousmethods by which electrons may be removed frommatter or the means of obtaining electronicemission. The operation and characteristics ofhot -cathode vacuum tubes, hot -cathode gas -filledtubes, and photoelectric tubes are most adequatelydealt with in succeeding chapters, while sufficientspace is given to such special devices as cold -cathode tubes, the Magnetron, photo -conductivecells, selenium cells, and photo -voltaic cells.

The rest of the book is devoted to the applicationsof these devices, and a lengthy explanation is givenof the operation of the valve, using the word in itstrue sense of a one-way device or rectifier and notwith the popular meaning of any electron tube.The index discicoes that no fewer than six chaptersare given to the triode vacuum tube as a powerconverter, a low -frequency amplifier, a radio -frequency amplifier, an oscillator, a modulator,and a detector, but a closer examination of the bookreveals that tetrodes and pentodes are also dis-cussed in this section. The value of the electrontube in measurement work is not overlooked, andthe last chapter gives a resume of special applica-tions of electronic devices to clocks, television,stroboscopes, and time -delay relays.

From the point of view of the radio engineer, thechapters on the applications of the triode are themost interesting, for they are not confined entirelyto the valve, but deal also with the design of suitablecircuits for use with it. The actual design of theassociated components, of course, is not touchedupon, but formulae are given which enable therequired circuit values to be calculated. In mostcases, however, the formulae appear more for thepurpose of explaining some particular effect thanas an aid to design. The book is not " popular,"therefore, for in places it is highly technical andmathematical ; it should prove 'readily intelligibleto any well -trained engineer, however, with nospecial knowledge of radio technique.

The book is well bound and unusually free fromerrors, and it is profusely illustrated, for the 452pages contain 537 illustrations of which the majorityare experimental curves showing the action ofelectron tubes. There is a good index.

W. T. C.


Abstracts and References

October, 1933

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

PAGE PAGE

Propagation of Waves .. 557 Directional Wireless 570Atmospherics and Atmospheric Acoustics and Audio -Frequencies.. 570

Electricity 561 Phototelegraphy and Television .. 572Properties of Circuits 562 Measurements and Standards .. 574Transmission 563 Subsidiary Apparatus and Materials 576Reception .. 566 Stations, Design and Operation .. 579Aerials and Aerial Systems 567 General Physical Articles .. 579Valves and Thermionics .567 Miscellaneous 580

PROPAGATION OF WAVESVERSUCHE UBER BETRIEBSVERFAHREN FUR LANG-

STRECKEN-FLUGFUNKVERKEHR (Field Testson Radio Communication over Long -Dis-tance Aircraft Routes [including the Re-duction of Short -Wave Fading by " Alter-nate -Aerial " Transmission or Reception]).-P. von Handel, K. Kruger and H. Plendl.(Hochf:tech. u. Elek:akus., July, I933, Vol.42, No. a, pp. II -2o.)

Short -distance tests by the DVL showed asatisfactory decrease of fading by the method ofalternate transmission from differently polarisedaerials (Abstracts, 1932, p. 156). The present paperdescribes long-distance tests from a ship sailingfrom Wilhelmshaven to the Dogger Bank and thecoast of Norway, reception being at Berlin (max.distance 1200 km). Not only was the alternate -aerial system tested, but experiments with periodicfrequency -varying transmissions and short -pulsesignals were carried out to investigate the selectivepolarisation changes occurring over long distances.

Dealing first with the alternate -aerial system,the writers point out that the note -frequencyalternation necessary for actual service was notused in these tests, since oscillographic observationof the signals from the two aerials separately wasdesired ; the one aerial therefore transmitted ashort dash and the other a longer dash, alternatingat a rate of 3 times a second. Another changefrom the short -distance tests was that the twotransmitting aerials were a symmetrical horizontaldipole and a vertical aerial, in place of the twocrossed horizontal aerials. Reception at Berlinwas with two receivers, connected to crossed hori-zontal aerials or to one horizontal aerial and avertical aerial. It would appear that while thealternate -aerial system at the transmitting endfrequently failed to compensate deep fading, andwhile the same thing occurred at the receiving end,these failures only very occasionally occurredsimultaneously at both ends. The combinationof the two systems should therefore give a markedimprovement.

Among the various results, it was found that thehorizontal transmitting aerial generally gavedeeper fading than the vertical ; in the long, narrowNorwegian fjords the horizontal transmitting aerial

gave stronger field strengths at the receiver thanthe vertical, while on the open sea the verticalaerial, with its better height above the ship's deck,showed its superiority.

It is also mentioned that the latest DVL testshave shown that the alternate -aerial system isquite applicable to telephony as well as to tele-graphy : for satisfactory telephony the switchingfrequency must be increased to twice the highestfrequency which has to be transmitted. Appliedto telephony the system has a marked effect indiminishing deep amplitude fading, and also de-creases the distortion produced by selective polarisa-tion changes. " Favourable for combined fadingcompensation (amplitude- and selective -fading)is also the use of the rotating -field principle " [seeend of abstract, July, pp. 383-384]. Finally, thealternate -aerial system can be applied to spaced,instead of differently polarised, aerials ; this planhowever would be difficult to carry out for mobilestations.

The methods just described lead to an extensivecompensation of fading troubles due to temporalvariations of polarisation through the action of theearth's magnetic field. The paper then deals withthe frequency -varying and pulse transmissionsundertaken to investigate selective fading pheno-mena and to devise ways of combating them. Inthe frequency -varying tests the ship wavelengthof 53 m was periodically changed, its frequencybeing raised by I 200 or 80o c/s in about sec. andthen diminished by a like amount in the same time.At regular intervals the horizontal and verticaltransmitting aerials were alternated. Receptionwas on an east -west horizontal dipole and simul-taneously on a vertical or a north -south horizontalaerial, the oscillograms of Fig. 6 showing therecords from the E -W aerial at the top of the stripand from the vertical or N -S aerial at the bottom.The sharply angular curves running along belowthe strip indicate the frequency -variation cycle.

In the pulse transmissions a rotating disc andphotocell method was used, giving pulses of about0.7 msec. at a rate of about 13 per sec. Reception,as before, was on two aerials simultaneously ; trans-mission, was from the horizontal dipole. Specimenoscillograms are shown in Fig. 7. One result " canhardly be explained otherwise than by wide

October, 1933

changes of the angle of the downcoming rayoccurring within a period of io minutes."

INVESTIGATION OF SELECTIVE POLARISATIONCHANGES AND FADING, BY FREQUENCY -CHANGE AND PULSE TRANSMISSIONS.-vonHandel, Kruger and Plendl. (See precedingabstract.)

CHARACTERISTICS OF ELECTROMAGNETIC RADIATIONFROM AIRCRAFT IN FLIGHT.-T. C. Rivesand J. C. Coe. (Trans. Am.I.E.E., Dec.,1932, Vol. 51, pp. 981-986.)

The full paper, a summary of which was dealtwith in 1932 Abstracts, pp. 575-576.

INTERACTION OF RADIO WAVES fin the PropagatingMedium] .-Tellegen . (World -Radio, lithAugust, 1933, Vol. 17, No. 42o, p. 165.)

A short article commenting on Tellegen's recentobservations (August Abstracts, p. 437) and theirpossible significance.

SOME CHARACTERISTICS OF FIVE -METRE [Ultra -Short -Wave] TRANSMISSION.-J. D. Kraus.(Summary only in Proc. Inst. Rad. Eng.,August, 1933, Vol. 21, No. 8, pp. 1071-1072.)

With a transmitting output of about one watt." The effects of hills and valleys were very promin-ently indicated upon the field -strength maps. Anantenna system giving a vertically polarised wavegave best results, and apparently regardless ofhow the wave leaves the antenna, it tends tobecome vertically polarised at a distance. Resonantreceiving antenna systems gave best results, andhigher voltages were obtained with long antennasworking at harmonics. Phase -shifting circuitslocated at correct points in the resonant systemalso proved valuable."

ULTRA -SHORT-WAVE TESTS FROM SUMMIT OFMOUNT SNOWDON [Five -Metre Signals toHoddesdon, Herts, 200 Miles away].-O'Hefferman and Myatt. (Daily Mail,25th August, 1933, p. 7 : paragraph only).

EXPERIMENTAL RESEARCHES ON THE TOTAL RE-FLECTION OF HERMAN WAVES.-G. Beauvais.(L'Onde Elec., May and June, 1933, Vol. 12,Nos. 137 and 138, pp. 213-235 and 273-294.)

Continuation and end of the paper referred to inAbstracts, September, p. 494, 1-h column. Seealso August, p. 44o, r -h column.

EXPERIENCES SUR LA PROPAGATION DES ONDESTRES COURTES DANS LES TUNNELS (Experi-ments on the Propagation of Ultra -Shortand Micro -Waves in Tunnels [the Shafts andGalleries of Mines and a Concreted Tunnel]).-A. Arenberg and W. Peicikov. (L'OndeElec., May, 1933, Vol. 12, No. 137, pp. 250-261.)

The main tests were with wavelengths of 400-500cm, but 33 -cm waves were also used. Not onlystrong absorption by the walls, but also complexinterference effects, render propagation very poor(cf. Ritz, 1929 Abstracts, p. 15o, 1-h col.). Thewaves behave as if " clinging " to the surfaces

558 THE WIRELESS ENGINEER e.c.

of separation of two media, suggesting that thePoynting vector has everywhere a componentparallel to the axis of the corridor : this wouldaccount for the decrease in possible distancebetween transmitter and receiver caused by anintersecting gallery.

MITTEILUNG USER DIE VERSUCHE IN DEN PUNK-WAH oHLEN (Report on the Tests in thePunkwa Caves and Tunnels).-V. Fritsch.(Hochf :tech. u. Elek:akus., June, 1933,Vol. 41, No. 6, p. 218.)

The results of the earlier tests (1932 Abstracts,pp. 397-398) led to a survey of that part of themountain range by radio signal -strength measure-ments, with a view to tracing as yet undiscoveredtunnels. Already a cavern 150 m x 20 m hasbeen charted, and confirmed by actual exploration.The tests are being continued.

CHARACTERISTICS OF THE IONOSPHERE.- J . P.Schafer and W. M. Goodall. (Nature,3rd June, 1933, Vol. 131, p. 804.)

The writers have observed echoes from theionosphere of wireless pulses emitted vertically,and summarise their results in this letter. BetweenE region, at a height of 100-120 km, and F region,at a height of 19o-300 km, they find an intermediatereflecting region, which they refer to as the Mregion, at an average height of about 15o km.This was observed on a number of occasionsbefore noon during December, 1932, and Jan. andFeb., 1933. It can be observed only so long asits maximum of ionic density exceeds that of theE region below it.

They also find that ionisation in the F regionhas a step -like structure during daylight hours,with abrupt ionic density gradients at heights near200, 240 and 28o km. Erratic variations in ionisa-tion of F region have been found during daylighthours in the past winter (1932-1933). " Theabrupt changes found in the F region are not nor-mally present in the E region."

FINE -STRUCTURE OF THE IONOSPHERE.-E. V.Appleton. (Nature, 17th June, 1933, Vol.131, pp. 872-873.)

The maximum ionisation content of the variousupper atmospheric regions has been measured atthe Radio Research Station, Slough (SeptemberAbstracts, p. 494), by finding the critical penetra-tion frequencies for the ordinary ray of the twomagneto -ionic components into which the incidentwireless beam is resolved by the action of theearth's magnetic field. Curves have been obtainedshowing the relation between the equivalent pathP1 of the atmospheric waves and their frequency f ;such a curve is shown in this letter and deductionsfrom the observations are made and comparedwith those of Schafer and Goodall (see precedingabstract).

Evidence of a region of electrification inter-mediate between the two main regions E and Fis found, particularly in the early morning hours[Schafer and Goodall observe it usually beforenoon]. Evidence of the existence of a ledge onregion F has also been found. From the (PI, f)curve shown, the maximum ionisation contents in


electrons per cc are deduced as follows : region E,1.8 x io5 ; intermediate region, 2.5 x io6 ; regionF ledge, 3.8 x 105 ; region F main, 6.1 x ro5.A curve showing qualitatively the most frequentdaytime variation of ionisation with height isshown ; it has maxima Eli and VI and ashelf V. " The British series of observationssuggests therefore that there are four main com-ponents in the ionosphere caused by the influenceof ultra -violet light from the sun. It is temptingto associate the four components with the fourionisation potentials of oxygen and nitrogen atomsand molecules."

FINE -STRUCTURE OF THE IONOSPHERE. -J. A.Ratcliffe and E. L. C. White. (Nature,17th June, 1933, Vol. 131, p. 873.)

The writers have made automatic records ofionospheric reflections (September Abstracts, p. 495)and have recently found indications of the iono-spheric reflecting region between E and F regionsfound by Schafer and Goodall and by Appleton(see preceding abstracts). A record taken on awavelength of 15o m is given, showing jumps ofthe two magneto -ionic components from reflectionat F region to reflection at the intermediate regionand then to reflection at E region, and alsosimultaneous reflections from the two lower regions." The fact that the jump from the F region to theintermediate region takes place at different timesfor the right- and the left-handed componentsshows that the intermediate region is mainlycomposed of electrons."

THE IONOSPHERE. -R. A. Watson Watt. (Nature,1st July, 1933, Vol. 132, pp. 13-17.)

A note on our present knowledge of and recentwork on the constitution of the ionosphere (cf.three preceding abstracts) and on the RoyalSociety discussion on June 22. A diagram is givenof the track of the corpuscular eclipse of Aug. 21,1933, for an assumed corpuscular velocity of 1 000km/sec., computed by J. C. P. Miller.

THE EFFECT OF THE EARTH'S MAGNETIC FIELDON THE PROPAGATION OF SHORT WIRELESSWAVES. -J. A. Ratcliffe and E. L. C. White.(Phil. Mag., July, 1933, Series 7,, Vol. 16,No. 103, pp. 125-144.)

For abstracts of papers referring to this subjectsee Abstracts, 1931, p. 143 (Hartree) ; 1932, pp. 155,400 and 575 (Appleton and Builder : White :Schafer and Goodall) ; January, pp. 3o (Appleton-two) and 29-30 (T. L. Eckersley : Appleton andRatcliffe) ; and September, p. 495 (Ratcliffe andWhite). The present paper describes " the con-struction and use of a circularly polarised receiverwhich makes it possible to pick out the right- andleft-handed circularly polarised components for awave incident vertically " [cf. last reference].Effects found in continuous records of reflectionsfrom E and F regions are explained on the basisof the magneto -ionic theory, and it appears that" electrons are responsible for the deviation of thewaves in both these regions." A discussion is alsogiven of simultaneous reflections from the Eand F regions and the polarisation of daytimesignals.

October, 1933

THE STATE OF POLARISATION OF DOWNCOMINGWIRELESS WAVES OF MEDIUM LENGTH [andthe Cause of Rapid Fading]. -J. A. Ratcliffeand F. W. G. White. (Phil. Mag., Aug.,1933, Supp. No., Series 7, Vol. 16, No. 105,PP. 423-440.)

The method adopted in the experiments describedin this paper has already been referred to in 1932Abstracts, p. 276, 1-h col. A suppressed ground -ray aerial system is used to observe the waves andis essentially the same as that previously used byAppleton and Ratcliffe (Proc. Roy. Soc., A, 1927,Vol. 115, p. 315). The adjustments required areinvestigated theoretically ; the writers also findthat the effect of the ground may be completelyneglected in the range of wavelengths underconsideration (too -500 m). Down -coming wavesfrom some of the B.B.C. transmitting stations arefound, in general, to be elliptically polarised withleft-handed sense of rotation. The observationsare found to agree with conclusions drawn fromthe Appleton-Hartree formula (see, for example,January Abstracts, p. 3o, Appleton). Observedvariations of the state of polarisation suggest thatrapid fading is due to an interference mechanismrather than to absorption effects.IONISATION DENSITY AND CRITICAL FREQUENCY. -

L. Tonks. (Nature, 15th July, 1933, Vol.132, p. xox.)

This letter remarks upon the introduction of theLorentz polarisation term in' recent work on thetheoretical relation between ionisation density inthe upper atmosphere and critical frequency[e.g., Abstracts, 1929, p. 441, and 1931, pp. 143and 432 (Hartree) ; 1932, p. 575 (Appleton andNaismith) ; and May, p. 262 (Appleton andBuilder)] and states that such a " polarisation "force can only exist " when there is some detailedarrangement of the negative with respect to thepositive charges " [in Hartree's work, referred toabove, it is, however, shown that the polarisationterm is valid for a medium whose ionisation con-sists of discrete free electrons or positively chargedions].

RECORDING WIRELESS ECHOES AT THE TRANS-MITTING STATION. -I. Ranzi. (Nature, 29thJuly, 1933, Vol. 132, p. 174.)

The method described by the writer has alreadybeen referred to in 1932 Abstracts, p. 155. Noincrease of echo amplitude was obtained by in-creasing the distance between emitting and re-ceiving stations up to 500 m.MEASUREMENTS OF ATTENUATION, FADING, AND

INTERFERENCE [by Atmospherics] in SOUTH-EASTERN AUSTRALIA, AT 200 KILOCYCLESPER SECOND. -G. H. Munro and A. L.Green. (Journ. Inst. Engineers, Australia,June, 1933, Vol. 5, No. 6, pp. 193-199).

" The importance of the lower frequencies forbroadcasting depends on the greater range of thedirect ground ray ; hence it is of much interest tocheck by experiment the theoretical formulaewithin the maximum useful range of the ground rayat these frequencies. From the point of view ofproviding a broadcasting service it is also necessaryto know something of the fading which may becaused by the indirect ray from the transmitter,

D 2

October, 1933 56o

and of the interference due to fulgurs " [a name,with abbreviation " F's," suggested for replacing" atmospherics " and " X's " when the disturbancesare due to lightning discharges, " static " beingreserved for man-made electrical interference andfor that due to discharges from aerials, etc., causedby high potential gradients, charged rain, andpossibly dust storms].

As regards the direct ray attenuation for bothinland and coastal districts, the following groundconductivities, in e.m. units, were found :-alongcoast, path mainly over sea, 10-13 to 5 x 10-13 ;over flat country, io-13 ; fiat wooded country,5 x io-14 ; undulating wooded country, io-14;country including mountain slopes, 5 x

With regard to fading : for coastal districts,where much of the path is over sea, the 5o% fadingring (according to the U.I.R. Madrid Conferencethe service area of a broadcasting station can besaid to be terminated where 5o% fading is en-countered) is not closer than 55o km to the trans-mitter ; for inland districts, where the interveningcountry is undulating and wooded, it is first en-countered at about 250-350 km from the trans-mitter, or at 200-250 km if some of the path isover mountain slopes.

Average values of the indirect ray were of theorder of o.o5 mv/m for i kw radiated, the greatestvalues being slightly above 0.1 my/in (T. L.Eckersley's figures for the European 550-I 500 kc/sband being much higher, of the order of o.3 my/m).

The general noise level, due to indirect -raypropagation of atmospherics from great distances,was found not to exceed 0.015 my/m, so that " abroadcasting service, at 200 kc/s, of 1.5 mv/mwill be entirely free from atmospheric interference,it being understood that purely local thunderstormsare excepted." Special tests were made to translatethe estimates of integrated " fulgur " level to abasis of interference with broadcast reception, andit was found that with the carrier field intensityabout io times that of the integrated disturbancelevel, speech was over 9o% intelligible but musicwas marred (" classical music : owing to the highcontinuity of dance music a much greater intensityof noise is bearable.") A ratio of ioo to I wasfound to be desirable for perfect reception of music.

The field intensity set was substantially of thesame design as that described by Green and Wood(see under " Measurements and Standards ") buta Cambridge thread recorder was substituted for thefaster -moving Einthoven photographic equipment,and a vertical aerial was used instead of the loop.PROPAGATION OF WAVES OF 150 TO 2 000 KILO-

CYCLES PER SECOND (2 000 tO 150 Metres)AT DISTANCES BETWEEN 50 AND 2 000KILOMETERS [Madrid Sub -Committee Graphsof Average Data].-B. van der Pol, T. L.Eckersley, J. H. Dellinger and P. le Cor-beiller. (Proc. Inst. Rad. Eng., July, 1933,Vol. 21, No. 7, pp. 996-1002.)

See also Abstracts, August, p. 439, r -h column ;also September, p. 518, 1-h column.

PROPAGATION OF WAVES : WORK OF SPECIALCOMMITTEE AT LUCERNE [Qualifying andAmplifying Madrid Committee's Reports].(World -Radio, 18th August, 1933, Vol. 17,No. 421, pp. 200 and 198.)


A METHOD OF CALCULATION OF FIELD STRENGTHSIN HIGH -FREQUENCY [Short -Wave] RADIOTRANSMISSION.-S. Namba and T. Tsukada.(Proc. Inst. Rad. Eng., July, 1933, Vol. 2 X,No. 7, pp. 1003-1028.)

Continuation of the work dealt with in MayAbstracts, p. 265, r -h col. The theory is summarisedand applied to practical cases. The Tokio-CapeTown and Tokio-Melbourne circuits are takenas examples, 15- and io-Mc/s diurnal transmissioncharacteristics being computed and plotted for theformer and zo- and 2o-Mcis for the latter, thepower being taken as r kw in all cases and thefield strength plotted in db above 1 µv/m. As a thirdexample signals from Bolinas received at Tokio arecalculated on the assumption of a radiated powerof 20 kw and an aerial -system gain of 20 db."The calculated result is shown in Fig. 17, whichcoincides fairly well with our observations."Finally the problem of how the transmitter outputand the form of the aerials must be chosen to give24 -hours' high-speed communication on the Tokio-Melbourne service is considered. The writersconclude : " The proposed method is of greatuse not only for the projection of a new communica-tion service, but also for an investigation of h.f.propagation phenomena."

FIELD -STRENGTH MEASURING SET FOR GROUNDAND SKY WAVES, AND SOME RESULTS OFPRACTICAL TESTS.-Green and Wood. (Seeunder " Measurements and Standards.")

DISTANT TELEVISION RECEPTION : MULTIPLEIMAGES DUE TO PROPAGATION EFFECTS.-Bodroux and Rivault. (See under " Photo -telegraphy and Television.")

THE TEMPERATURE OF THE UPPER ATMOSPHEREDEDUCED FROM DAY/NIGHT CHANGE OFHEIGHT OF ABSORPTION LAYER FOR AURORALLIGHT.-G. Angenheister. (Sci. Abstracts,Sec. A, May, 1933, Vol. 36, No. 425, p. 45o.)

METEORS AND THE 80-90 KM LAYER OF THEEARTH'S ATMOSPHERE.-Malzev. (See under" Atmospherics and Atmospheric Elec-tricity.")

VERTICAL DISTRIBUTION OF OZONE IN THE ATMO-SPHERE.-F. W. P. G6tz, G. M. B. Dobsonand A. R. Meetham. (Nature, igth Aug.,1933, Vol. 132, p. 281.)

A new method is being used to determine notonly the centre of gravity of the layer of atmosphericozone but also the general character of its verticaldistribution. The method is " based on observationsof the spectrum of the light received from theclear blue zenith sky as the sun is rising or setting."The average height of the centre of gravity atArosa now appears to be about 20 km, and thereseems to be an appreciable amount of ozone inthe lower layers.

STUDY OF ATMOSPHERIC OZONE BY A RAPIDMETHOD OF VISUAL PHOTOMETRY.-J.Gauzit. (Comptes Rendus, loth July, 1933,Vol. 197, No. 2, pp. 178-180.)

The' values found during the first six months of


1933 for the thickness of the ozone layer are veryconsiderably higher than those found in the sameperiod in 1931. This large variation in the spaceof 2 years is comparable with that which occurredin 1912-1914.

THE PHOTOMETRY OF SOLAR ECLIPSE PHENOMENA.-C. H. Sharp and others. (Journ. Opt.Soc. Am., July, 1933, Vol. 23, No. 7, pp.234-245.)

A NEW METHOD OF CALCULATING THE HYPO -CENTRAL DEPTH. -P. Caloi. (La RicercaScient., 30th April, 1933, 4th year, Vol. r,No. 8, pp. 508-510

PERIODIC COMPONENTS IN LOVE WAVES [from the1922 Formosa Earthquake Records]. -Y.Labrouste. (Comptes Rendus, 24th July,1933, Vol. 197, No. 4, PP 344-346.)

REFRACTION OF INFRA -RED RAYS [Ships detected6 Miles beyond Horizon].-Macneil. (Seeabstract under " Miscellaneous.")

ELLIPSOIDAL FUNCTIONS AND THEIR APPLICATIONTO SOME WAVE PROBLEMS. -Hanson. (Seeunder " Miscellaneous.")

ATMOSPHERICS AND ATMOSPHERICELECTRICITY

INFLUENCE OF GEOPHYSICAL FACTORS ON THEFREQUENCY OF LIGHTNING STROKES ON ANAREA. -L. N. Bogolavlensky. (Nature,25th July, 1933, Vol. 132, p. 99.)

This letter describes an investigation whichshowed that " the points of the lightning strokesalways correspond to the places with increasedintensity of the [radioactive] earth radiation " andthat the latter is connected with a layer of rock ofconsiderable conductivity, lying sometimes at aconsiderable depth below the surface. For aprevious paper by the same writer see Abstracts,1931, P. 494, r -h col., and for other papers on thesame subject see Jan., Feb. and March Abstracts,pp. 32, 93 and 151, all r -h columns ; also 1932,p. 577, 1-h column.

RADIO WAVES FROM OUTSIDE THE SOLAR SYSTEM[Direction of Arrival of " Hiss Type "Atmospherics]. -K. G. Jansky. (Nature,8th July, 1933, Vol. 132, p. 66.)

This preliminary letter refers to a continuationof experiments on the direction of arrival of high -frequency atmospherics, already referred to inAugust Abstracts, p. 441, r -h col. The horizontalcomponent of the direction of arrival of " hisstype " atmospherics was there found to changenearly 36o° in 24 hrs and the source of the dis-turbance was assumed to be associated with thesun, but, from records taken for more than a year,it is now thought that the source of noise is locatedin a region stationary with respect to the stars,with co-ordinates approximating to a right ascensionof 18 hours and a declination of - ro°.

October, 1933

ATMOSPHERICS [" Fulgurs "] IN SOUTH-EASTERNAUSTRALIA, MEASURED ON A FREQUENCY OF200 KILOCYCLES PER SECOND. -Munro andGreen. (See abstract under " Propagationof Waves.")

METEORS AND THE 80-90 KM LAYER OF THEEARTH'S ATMOSPHERE. -V. Malzev. (Nature,22nd July, 1933, Vol. 132, p. 137.)

The writer draws attention to the fact that themaxima of brightness of meteor paths are foundat a nearly constant height of 85-90 km.

LOW AURORAS AND TERRESTRIAL DISCHARGES. -C. S. Beals. (Nature, I 2th Aug., 1933,Vol. 132, p. 245.)

" Reports have been received from observers innorthern Canada, who have detected the odour ofozone during auroral displays . . . accompaniedby sounds." The sounds and low-level displaysappear to be due to electrostatic discharges near theearth's surface.

MOTION OF ELECTRONS IN HETEROGENEOUS MAG-NETIC FIELDS. -H. Ertel. (Gerlands Beitr.z. Geophys., No. 2, Vol. 38, 1933, pp. 142-146.)

SECULAR CHANGES OF THE MAGNETIC ELEMENTS,OTTAWA, 1500-1930.-W. H. Herbert.(Canadian Journ. of Res., July, 1933, Vol. 9,No. r, pp. 94-96.)

THE [Continual] FLUCTUATIONS OF THE EARTH'SELECTRIC FIELD [not only due to LocalIrregular Disturbances]. -A. Duperier andG. Collado. (Comptes Rendus, 31st July,1933, Vol. 197, No. 5, pp. 422-423.)

[Establishment of an Important Relation between]THE ELECTRICAL POTENTIAL GRADIENT ANDATMOSPHERIC PRESSURE. -R. Guizonnier.(Comptes Rendus, 17th July, 1933, Vol. 197,No. 3, pp. 265-266.)

VARIATIONS IN THE SMALL -ION CONTENT OF THEATMOSPHERE, AND THEIR CAUSES. -G. R.Wait. (Journ. Franklin Inst., Aug., 1933,Vol. 216, No. 2, pp. 147-155.)

PENETRATING RADIATION FROM THUNDERCLOUDS.-J. E. I. Cairns. (Nature, 29th July, 1933,Vol. 132, p. 174.)

The writer has used ionisation methods to detectthe penetrating radiation from thunderclouds ; hisresults agree with those of Schonland and Viljoen(August Abstracts, p. 442, 1-h column).

AIRPLANE COSMIC -RAY INTENSITY MEASUREMENTS.-L. M. Mott -Smith and L. G. Howell.(Phys. Review, xst July, 1933, Series 2,Vol. 44, No. I, pp. 4-11.)

An abstract of this paper was dealt with in JuneAbstracts, p. 322, r -h column.

COMPARISON OF THE ANGULAR DISTRIBUTIONS OFTHE COSMIC RADIATION AT ELEVATIONS6 28o FT AND 620 FT. -T. H. Johnson.(Journ. Franklin Inst., Aug., 1933, Vol. 216,No. 2, pp. 259-267.) See July Abstracts,p. 388, r -h column.

October, 1933

ANGULAR DISTRIBUTION OF LOW ENERGY COSMICRADIATION AND INTERPRETATION OFANGULAR DISTRIBUTION CURVES. -T. H.Johnson and E. C. Stevenson. (Journ.Franklin Inst., Aug., 1933, Vol. 216, No. 2,pp. 267-27o.) See July Abstracts, p. 388,r -h column.

THE ASYMMETRY OF THE COSMIC RADIATION ATSWARTHMORE.-T. H. Johnson and E. C.Stevenson. (Phys. Review, 15th July, 1933,Series 2, Vol. 44, No. 2, pp. 125-126: Journ.Franklin Inst., Aug., 1933, Vol. 216, No. 2,pp. 276-280.)

THE INTERPRETATION OF THE AZIMUTHAL EFFECTOF COSMIC RADIATION. -M. S. Vallarta.(Phys. Review, ist July, 1933, Series 2,Vol. 44, No. x, pp. 1-3.)

This note gives a discussion based upon thetheory of Lemaitre and Vallarta (Abstracts, March,p. 152, 1-h col. ; April, p. 209, r -h col. ; alsoJohnson, August, pp. 442-443, and Alvarez andCompton, p. 442, r -h col.) which indicates that apredominant part of cosmic radiation consists ofpositive particles.

AZIMUTHAL INVESTIGATION OF COSMIC RADIATION.-S. A. Korff. (Phys. Review, 15th July,1933, Series 2, Vol. 44, No. 2, p. 13o :abstract only.)

THE DISTRIBUTION OF THE IONISING PARTICLESOF THE PENETRATING [Cosmic] RADIATIONIN RELATION TO THE MAGNETIC MERIDIAN.-J. P. T. Viljoen and B. F. J. Schonland.(Phil. Mag., Aug., 1933, Supp. No., Series 7,Vol. 16, No. 1o5, pp. 449-456.)

This paper describes an experimental investiga-tion of the direction of arrival of ionising particlesassociated with the penetrating radiation at sea -level in Cape Town. No significant differencebetween north and south directions, or betweeneast and west directions, is found ; there aresmall defects from east and west as compared withnorth and south. The conclusion is drawn that" any charged particles of extra -terrestrial originare accompanied by a larger number of secondaryparticles generated in the atmosphere by someradiation which is not affected by the earth'smagnetic field. The north-east difference suggeststhat the majority of the primary particles arepositively charged."

LATITUDE EFFECT OF COSMIC RADIATION. -H.Hoerlin. (Nature, 8th July, 1933, Vol. 132,p. 61.)

This preliminary letter describes measurementsof the intensity of cosmic radiation in differentgeomagnetic latitudes, made on a voyage fromthe Strait of Magellan to Hamburg. A graph isgiven showing the measured intensity in volts perhour as a function of the geomagnetic latitude.The intensity is found to increase by about 52 percent. when passing from the equatorial region to55° N geomagnetic latitude. The observationsagree in general with those of Clay (Abstracts,January, p. 34), and A. H. Compton (June, p. 322 :both I -h columns). The northern and southern


parts of the curve are found not to be symmetricalwith respect to either the geomagnetic or thegeographical equator.

INTERACTION BETWEEN COSMIC RAYS AND MATTER.-B. Rossi. (Nature, 29th July, 1933,Vol. 132, pp. 173-174.)

This letter describes results obtained by themethod of triple coincidences in the investigationof the frequency of occurrence of secondary" showers " generated by cosmic rays passingthrough matter.

ON THE NUCLEAR DISINTEGRATION PRODUCEDBY THE PENETRATING RADIATION. -B. Rossi.(La Ricerca Scient., 15/31st May, 1933,4th Year, Vol. 1, No. 9/1o, p. 586.)

PROPERTIES OF CIRCUITS"OBER DIE RESONANZFREQUENZEN ZWEIER GEKOP-

PELTER SCHWINGUNGSKREISE (The Reson-ance Frequencies of Two Coupled OscillatoryCircuits [and a New Method of FrequencyAnalysis]). -W. Fehr and K. Kreielsheimer.(Zeitschr. f. tech. Phys., No. 8, Vol. 14,1933,

Follow.306-308.)

ing on the work of Kossel (1931 Abstracts,p. 378) the writers discuss an electrical circuitwhich is clearly analogous to Kossel's pendulums.Two closed oscillatory circuits I and II are con-nected by a coupling link forming part of neithercircuit (in Fig. x the lower ends of the two in-ductances are connected by a straight lead, theupper ends by an inductance L0; the inductancesof the two circuits are spaced so as to have nomutual magnetic coupling) and are excited by athird circuit containing two exciting coils, onecoupled to the inductance of I and the other to theinductance of II. The conditions resulting fromI and II being excited in the same phase, and fromtheir being excited in phase opposition, are in-vestigated.

The writers then show how the properties ofsuch an arrangement enable it to be used for fre-quency analysis or for measuring non-linear dis-tortion. Thus if the frequency mixture to beinvestigated is applied to I and II in the samephase, resonance with the fundamental can befound in both circuits and a measure of the ampli-tude of the fundamental can be obtained. If then,without changing the tuning of I and II, the fre-quency mixture is applied in phase opposition,resonance points for the higher (or lower) partialfrequencies will be found, depending on the valueof the coupling inductance L0 (or coupling capacityGo).

Another method, which gives a simultaneousmeasurement of fundamental and partial ampli-tudes, is to use a one-sided application of the mixedfrequencies ; i.e. to couple the exciting circuitto one circuit only (Fig. 2) and to lead the currentsfrom the two points x and 2 to a differential meter.FREQUENCY OF OSCILLATIONS IN TWO INDUCTIVELY

COUPLED CIRCUITS [One being in a ValveAnode Circuit : the Theoretical Existenceof Three Frequencies]. -M. Rousseau. (Sci.Abstracts, Sec.B, May, 1933, Vol. 36, No.425, P. 317-)


SOME GENERAL RESONANCE RELATIONS AND ADISCUSSION OF THEVENIN'S [Pollard's]THEOREM.-J. G. Brainerd. (Proc. Inst.Rad. Eng., July, 1933, Vol. 21, No. 7, pp.1050-1054.)

Author's summary :-It is shown that themaximum possible current which can be obtainedthrough an impedance joining two points of alinear network is in general greater than the short-circuit current between those points ; the maximumpossible voltage is in general greater than the open -circuit voltage ; the ratio of open -circuit voltageto short-circuit current is equal to the ratio of themaximum possible voltage to the maximum possiblecurrent, and both are equal to the magnitude ofthe input impedance of the network, etc. Thecondition for maximum voltage across a branch ofa linear network is derived.

As a preliminary, a form of Thevenin's theoremdifferent from the usual one and of greater useful-ness in the analytical solution of some circuitproblems is obtained.

OSCILLATORY CIRCUITS [Mathematical Investiga-tion of Toroidal Inductance connected toCondenser].-F. Dacos. (Sci. Abstracts,Sec.B , June, 1933, Vol. 36, No. 426, p.394.)

A NOTE ON NONLINEARITY IN TRANSDUCERS USEDIN COMMUNICATION [and' Its Correction bya Series Compensating Transducer].-P.Caporale. (Proc. Inst. Rad. Eng., July,1933, Vol. 21, No. 7, pp. 1029-1038.)

Author's summary :-" The characteristic of anonlinear transducer can usually be expanded ina power series, when it is not simply a single termraised to some power. For those cases where thecharacteristic may be so expressed, a general lawor equation is derived for the characteristic of acompensating transducer, that is, a transducerwhich in combination with the first transducer willproduce over-all linearity. It is shown, moreover,that when the series for the distorting transduceris not sufficiently rapid in its convergence, it be-comes impractical to realise the necessary com-pensating transducer. A simple case of an actualcompensating transducer used to correct a square -law characteristic is briefly described."

As an interesting case of a compensating trans-ducer for correcting amplitude distortion, referenceis made to Massa's use of an impedance decreasingwith current (actually composed of Thyrite-Abstracts, 1930, p. 332, 1-h col.) in the plate circuitof the distorting valve (July, p. 397, 1-h col.).

EXTENSION OF MAXWELL'S RULE FOR ANALYSINGELECTRICAL NETWORKS.-Y. H. Ku. (Sci.Abstracts, Sec.A , June, 1933, Vol. 36,No. 426, p. 621.)

A SIMPLIFIED FREQUENCY -DIVIDING CIRCUIT [usingOne Pentode Valye].-Andrew. (See under" Subsidiary Apparatus and Materials.")

DIODE DETECTION ANALYSIS.-Kilgour and Glessner.(See under " Reception.")

October, 1933

TRANSMISSIONDIE ULTRADYNAMISCHE SCHWINGUNGSANFACHUNG

DURCH RUCKKOPPLUNG (The UltradynamicExcitation of Oscillation by Retroaction[" Inversions " Theory of Ultra -Short WaveGeneration by Normal and Retarding -FieldCircuits]).-H. E. Hollmann. (Hochf :tech.u. Elek:akus., July, 1933, Vol. 42, No. r,pp. 32-33 : long summary only.)

The phenomena occurring in the generation ofu.s.w. oscillations in triodes are attributed to thephase displacement between valve current andcontrol potential, caused by the finite electronpath -time. The phase angle thus produced isgiven by eiS , and can -obviously be made large bychoosing a very high frequency cu or a long path -time S. The term " ultradynamic characteristic "is given to the characteristic curves for ia/e.outr,plotted at these ultra -high frequencies and analogousto Lissajous' figures in that at all phase anglescorresponding to uneven multiples of Tr (4' =(2n - r)n) the ultradynamic characteristics becomeinverted, while at even multiples of /T they againagree with the original static characteristic ;in the intermediate regions they take on a cyclic -elliptic form.

The summary continues : " For the more precisetheoretical treatment of the problem, taking intoaccount the displacement current, the space -charge -controlled triode is traced back to theequivalent circuit of a diode with emission control.From this there result certain deviations from thefundamental ideas of ultradynamic inversions.Thus the inversions maxima denoted by the serialorder number n do not proceed simply with n,but have their phase angle represented by tan 4 = 4.A further effect of the displacement current isthat the anode a.c. current gradually sinks (in closeapproximation with /0) as a result of the space -charge maxima collecting between the electrodesmore and more abundantly as the serial ordernumber increases. This effect shows itself in adecreasing slope of the ultradynamic curves, whichexplains the increasing weakness of excitation atthe higher frequencies."

The second part of the paper deals with thegeneration of ultra -high frequencies by retroactionin an ordinary triode with positive anode. Allnormal retroaction circuits (in u.s.w. work the three-point connection is the most common) give arising characteristic. From this it follows that inall ' back -inversion " zones, in which the ultra -dynamic characteristics are already ascending, theexcitation of the corresponding high frequenciesmust be possible. Equally, however, it would beexpected that in the interlying " inversion "zones a positive excitation would occur if only thecurve -inversion were compensated by the vectorialreversal of the external back -coupling. In practice such" inversions oscillations," following each other in aseries governed by the formula A2 E. = constina,have not so far been obtained; but observationshave been reported from other sources which canonly be interpreted as inversions phenomena.On the other hand it has been possible, by measuringthe excitation of a normal three-point circuit,to show the existence of several positive and negative

October, '933

zones ; with constant frequency the ultradynamicphase displacement was gradually increased byvariation of the path -time through change ofaccelerating potential. The dotted curve of Fig.shows such a measurement ; the abscissae representthe anode voltage and the ordinates the excitationon an arbitrary scale.

The third part of the paper deals with the oscilla-tion zones in the retarding -field method. Herethe case of pure emission -control, assumed in thetheory, actually exists. Consideration of thegeneral laws of such oscillation excitation showsthat the external back -coupling in retarding -fieldcircuits must be vectorially inverted comparedwith the normal valve circuit. Applied to theexcitation of ultra -high frequencies, this givesthe law that the excitation zones of the retarding -potential valve must lie just between those of thenormal triode. Complete experimental confirmationof this law was found, as is shown by the continuous -line curve in Fig. 1, exactly opposed in phase to thedotted retroaction -valve curve already mentioned.

In retarding -field circuits, as in retroaction -triodes(see above), the inversion of characteristic can becompensated by a reversal of the external back -coupling. If the normal triode is represented by-R, the retarding -field valve by + R and theexternal coupling by ± K, the sign of the excitationA is given by the relation sign A = sign R.K (- I)"."The generator arrangements commonly used foru.s.w. production have several degrees of freedom,so that the retroaction -coupling can be, positiveor negative. In practice, therefore, oscillationsoccur in zones of all orders, obeying the generallaw n2 1t Eg = const. On the exact " inversionstheory " this formula requires correction, in thatthe series of whole numbers n = I, 2, 3 . . . mustbe replaced by the values N given by the inversionscondition tan 4, = cs.

With a specially designed valve, whose grid is aspiral freely oscillating as a dipole, a wave of18 cm can be obtained with a grid voltage ofonly 128 v. This corresponds to the ninthinversion. " The development of an arrangementwith only one degree of freedom led to a generatorwith 5 watts output on a 50 -cm wave. Furtherconfirmation of the inversions theory was givenwhen the external back -coupling was reversed,the wave shortening itself by 3o% as a result of itsshift into a higher inversion zone."

The paper ends by considering the " inversions "theory in connection with other views on themechanism of u.s.w. generation, with the resultthat all theories based on the Barkhausen pictureof " electron swings " are to be regarded asspecial cases of the general theory.

SLIGHTLY DAMPED ULTRA -SHORT ELECTRIC WAVES[below so cm : Experimental Method ofGeneration].-K. Zakrzewski and M.Miesowicz. (Sci. Abstracts, Sec. A, June,1933, Vol. 36, No. 426, p. 654; Physik.Ber., 15th June, 1933, Vol. 14, No. 12, p.939)

" The waves radiated by the oscillator are con-ducted through a metal tube 5o cm long and ofsquare cross-section with 3.7 -cm sides. The wavesemerging from the tube have different wavelengths


and smaller decrement as compared with thosedirectly radiated froxn the exciter. A decrementof 0.19, for a 6.6 -cm wavelength, was attained."

THE GENERATION OF MICRO -WAVES WITH SPECIALVALVES IN A RETARDING -FIELD CONNECTION.-Gossel : Kohl. (See abstract under " Valvesand Thermionics.")

A GRAPHICAL METHOD FOR DETERMINING THETRANSIT TIMES OF ELECTRONS IN A THREE -ELECTRODE VALVE UNDER CONDITIONS OFSPACE -CHARGE LIMITATION.-J. S. McPetrie.(Phil. Mag., Aug., 1933, Series 7, Vol. 16,No. 104, pp. 284-293.)

THE INTERDEPENDENCE OF FREQUENCY VARIATIONAND HARMONIC CONTENT, AND THE PROBLEMOF CONSTANT -FREQUENCY OSCILLATORS.-J. Groszkowski. (Proc. Inst. Rad. Eng.,July, 1933, Vol. 21, No. 7, pp. 958-981.)

English version of the Polish paper dealt with inJuly Abstracts, pp. 39o-391. The writer's methodof establishing the interdependence is first appliedto dynatron circuits and then generalised so as torefer to triode circuits with retroaction. It isshown that the ways of proceeding towards theestablishment of constant -frequency generators are :(1) operation without harmonics, i.e. close to thecritical state. But to maintain the critical statethe operative conditions must be kept constant-which is already equivalent to maintaining aconstant frequency ; the only interest in this plantherefore lies in automatic methods. (2) Theremoval of the harmonics from the oscillatorycircuit, the simplest by a number ofseries filters tuned to the corresponding harmonics.In practice it is sufficient to use a few filters for thelowest harmonics only (k = 2, 3, 4, 5) ; the higherones are shunted sufficiently by the capacity of theoscillatory circuit (this explains the fact thatgenerators whose circuits have a large capacity areless subject to frequency variations due to changesin the operative conditions). (3) The equalisationof the imaginary energy distribution of the har-monics. Under certain circuit conditions it isshown that the second harmonic reduces thefrequency, while all higher harmonics raise it.In such a system the action of the higher harmonicscan be compensated by the action of the lowerones, and this fact allows a considerably betterfrequency constancy to be obtained with the helpof a few filters only than can be expected from the" shunt -trapping " action of method (2) alone.By the suitable detuning of one filter, the com-pensation of those harmonics which have not beenremoved by the filters can be attained.

THE EFFECT OF THE CURVATURE OF THE CHAR-ACTERISTIC ON THE FREQUENCY OF THEDYNATRON GENERATOR [Frequency Changesgreater than can be accounted for by Curva-ture alone].-E. B. Moullin. (Journ. I.E.E.,August, 1933, Vol.- 73, No. 440, pp. 186-'95.)

From the author's summary :-" The analysisshows that the fundamental frequency is affectedby the phase of the higher -harmonic components


of anode potential and that the effect would be zeroif the impedance of the circuit could be madepurely resistive to all the higher -harmonic currents.A network is described which fulfils closely therequired conditions. It is found that the changeof frequency which accompanies the change ofoutput is much greater than can be accounted forby the curvature of the characteristic, and henceit is presumed there must be some other cause ofinstability inherent in the valve. It is suggested[see Baker, below] that the electrode capacitancemay depend on the condition of the space charge."See also Groszkowski, above ; also Hayasi, Sep-tember Abstracts, p, 502, r -h column. Withregard to the 1923 work of Appleton and Greaves,the writer remarks that his method presents theanswer directly in an easier form and makes clearthe mechanism involved, thus leading at once tothe special form of circuit which renders the effectof curvature innocuous.

THE INTER -ELECTRODE CAPACITANCE OF THEDYNATRON, WITH SPECIAL REFERENCE TOTHE FREQUENCY STABILITY OF THE DYNA-TRON GENERATOR.-G. B. Baker. (Journ.I.E.E., August, 1933, Vol. 73, No. 440,pp. 196-203.)

From author's summary :-" The changes offrequency are such as to indicate that the anode -grid capacitance is a function of the anode potentialand is independent of the parameters of the oscilla-ting circuit. The same changes of capacitance arefound to occur when the circuit is not oscillatingand the valve is used to reduce the power factorof the circuit. The dependence of capacitance onspace charge is discussed briefly and outlinedanalytically. The dynatron characteristic is thenanalysed and the probable form of the secondaryemission-current/anode-potential curve is deduced.This deduced curve indicates the presence of spacecharge in the operating conditions where the changesof frequency are found to occur." See Moullin,above.

SUPPRESSION OF TRANSMITTER HARMONICS.-C. G.Dietsch. (Electronics, June, 1933, Vol. 6,No. 6, pp. 167-169.)

THE CRYSTAL DRIVE OF THE EXPERIMENTALSHORT-WAVE BROADCASTING STATION G5SW.-T. D. Parkin. (Marconi Review, May/June, 1933, No. 42, pp. r-ro.)

APPLICATION OF TRANSFORMER -COUPLED MODU-LATORS [and the Higher Efficiency and De-creased Distortion of Class B Amplification].-J. A. Hutcheson. (Proc. Inst. Rad. Eng.,July, 1933, Vol. 21, No. 7, PP. 944-957.)

Author's summary :-A brief résumé of Heising'smodulator theory is presented together with adiscussion of the class B operation of tubes in a push-pull audio -frequency circuit. Several cases ofcommercial applications of class B audio amplifiersare mentioned. Several general problems involvedin the use of class B audio amplifier are discussed,with the conclusion that such an amplifier will pro-duce more audio power for a given tube complement,

October, 1933

at higher efficiency, and with less distortion, thanamplifiers previously used in commercial appli-cations.

RADIOTELEPHONIE A BANDE LATERALE UNIQUE(Single Sideband Radiotelephony [Theoryand Methods, including Balanced Modu-lation by Diodes]).-M. Kolesnikov. (L'OndeElec., May, 1933, Vol. 12, No. 137, pp.237-249.)

A NEW METHOD OF OBTAINING FREQUENCYMODULATION [by Varying Resistance throughUse of R.F. Transmission Line].-E. D.Scott : Eastman. (Summary only inProc. Inst. Rad. Eng., August, 1933, Vol. 21,No. 8, p. 1074)

The two chief drawbacks of frequency modulationare the difficulty of producing linear modulationby means of a variable reactance, and of finding asatisfactory type of detector. The writer developsequations showing that very satisfactory linearfrequency modulation can be obtained with atransmission line one -eighth of a wavelength long,with a terminal resistance about equal to thecharacteristic impedance of the line : this wasdemonstrated experimentally. He also mentionssome work [no details are given here] which hehopes will solve the second difficulty.

FREQUENCY MODULATION AND THE EFFECTS OF APERIODIC CAPACITY VARIATION IN A NON -DISSIPATIVE OSCILLATORY CIRCUIT.-W. L.Barrow. (Proc. Inst. Rad. Eng., August,1933, Vol. 21, No. 8, pp. 1182-1202.)

Author's summary :-Certain fundamental char-acteristics of the theory of frequency modulation forarbitrarily large degrees of modulation and un-restricted modulation frequencies are developedfrom the differential equation for a dissipationlesscircuit with fixed inductance and variable capaci-tance. The several modes of modulating thefrequency are discussed and classified ; it is shownthat they give the same results only when theamount of modulation is very small. The case of" inverse capacity modulation " is then treated indetail. This treatment discloses the possibility ofunstable oscillations occurring with certain valuesof the parameters ; the nature and physical sig-nificance of these unstable oscillations are deter-mined, and it is explained why they are notordinarily observable in radio -frequency modula-tion or in the warble -tone generator. The frequencyspectrum of the stable oscillations is found, and ameans of calculating the amplitude given. Forcertain adjustments of the circuit the oscillationsmay be represented by a true Fourier series, whilein general this is not the case. Frequency modula-tion in radio -telephony, the warble tone, and thespecial case where the natural period of the un-modulated circuit and the frequency of modulationare of comparable magnitude, represent succes-sively more complicated cases of the same generalphenomena ; the latter is of special interest. Thenature of the phenomena accompanying otherthan a sinusoidal inverse capacity variation ismentioned.

October, 1933

RECEPTIONAN ANALYSIS OF POWER [Grid] DETECTION [and the

Use of a High -Impedance Choke in place ofthe Grid Leak and of a Tungsten Lamp asPlate Series Resistance].-R. de Cola.(Proc. Inst. Rad. Eng., July, 1933, Vol. 21,No. 7, pp. 984-989.)

" The reasons for overloading of detectors asoutput devices is discussed. It is shown that ingrid -circuit rectification replacing the grid leakwith a high -impedance choke [about too henrys]extends the overload point considerably. Usinga 247 -type pentode, 800 milliwatts are obtained atthe output circuit with 7 % maximum distortion."The writer considers that although his results wereobtained by running the valve at space currentssomewhat higher than the values recommended bythe makers, this should not seriously limit its life ;the total plate dissipation is only slightly higher thanthat specified. The audio output is probably morethan the available baffling area in a small set canutilise, and " certainly much more than the averagehome can generally use." Also, a power detectorof this type can supply sufficient power to a class Bamplifier without the necessity of an intermediate" driver " valve.

Before arriving at the use of the high -impedancechoke, the writer experimented with the employ-ment of a small tungsten -filament 7.5 w lamp as aseries resistance, in the plate circuit, which wouldhave comparatively large resistance to protect thevalve from abnormal plate current and yet, underan impressed carrier upon the grid, would auto-matically decrease in resistance to a value pro-portional to the carrier and thus quite accuratelyand automatically control the plate current andvalve characteristics. The improved form of theplate current thus obtained is seen in Fig. 2. In-cidentally it is mentioned that the presence of sucha lamp gives a good indication of resonance " andcan be used to advantage . . . in place of a tuningmeter."


DIODE DETECTION ANALYSIS.-C. E. Kilgour andJ. M. Glessner. (Proc. Inst. Rad. Eng.,July, 1933, Vol. 21, No. 7, pp. 930-943.)

Authors' summary :-This paper gives thecurrent and output relations for the linear andsquare -law diode detector in terms of the detectionefficiency and the detector and load resistances.The square -law case is shown to yield resultssimilar to the linear for large input potentials andreasonable values of load. Employing the recti-fication-current/input-voltage characteristic, thegraphical method is used to show that serious dis-tortion is developed for deep modulation wheneveran appreciable audio load is shunted across thediode resistor. It is found that M = D(A - i),where M is the modulation ratio at which cut-offbegins, D the detection efficiency, and A the ratio ofthe audio impedance to the direct -current value ofthe diode load. The value of M is shown for variousload conditions in both the linear and square -lawcases. The paper shows that the use of a smallcondenser across the load decreases the detectionefficiency but increases the input impedance. Theconnection to the driving circuit of a second diode forthe development of automatic volume control

potential is shown to cause distortion when such atube is used with a delay bias. Mention is made ofthe importance of considering the impedance of thedriving circuit, and attention is called to the failureof any of the methods of analysis to consider allof the factors involved.

H.F. AMPLIFIER WITH HIGH AMPLIFYING POWER[Reduction of Anode -Circuit Damping in aS.G. Stage without Alteration of Tuning,by a 2 -Stage Resistance -Capacity Cascadewith Output capacitively coupled to AnodeCircuit]. (Hochf:tech. u. Elek:akus., July,1933, Vol. 42, No. 1, p. 37: German Pat.No. 571 326, pub. 27th Feb., 1933, vonArdenne.)

A SUPERHETERODYNE RECEIVER WITH SEVERALFREQUENCY CHANGES IN THE I.F. CIRCUIT.(Hochf:tech. u. Elek:akus., July, 1933, Vol.42, No. I, p. 37 : German Pat. No. 570 682,pub. 18th Feb., 1933, Aigner.)

The first i.f. is led through an amplifier to afrequency -doubling detector. The new frequency2n1 and a fixed second heterodyning frequency N2give a sum- or difference -frequency n2 = N2 ± 2n1which is sent through the same amplifier, doubledby the same detector, and then combined with N2again. The resulting frequency n, = N2 ± 2n2passes to the output detector.

A LONG -WAVE SINGLE-SIDEBAND TELEPHONY RE-CEIVER FOR TRANSATLANTIC WORKING :PART I.-C. J. W. Hill and H. Page. (Mar-coni Review, May/June, 1933, No. 42, pp.13-26.)

A VARIABLE WAVELENGTH AERIAL [Aerial Wirewinding on Metal Drum : Improved Short -Wave Reception by Adjusting to OptimumNatural Wavelength].-(World-Radio, 18thAug., 1933, Vol. 17, No. 421, p. 205.)

AUTOMATIC REGENERATION [One Adjustment givesSuitable Regeneration Constant over WholeTuning Scalel.-R. W. Tanner. (Rad.Engineering, June, 1933, Vol. 13, No. 6, p.io.)

" Automatic regeneration has always been thehope of engineers and experimenters. . . . It isnow a possibility through a development recentlycompleted. . . . Considerable work is being donewith this new idea, applying it to various circuitsand with different tubes. . . ." The usual grid coiland the retroaction coil are both tuned to resonanceby a two -gang condenser : the coupling between themis both magnetic and capacitive. The formercoupling tends to make the plate feed -back currentsstronger at the low -capacity end of the scale, thelatter at the high -capacity end. By a suitablecombination the regeneration is made constantover the whole scale.

ADJUSTMENT OF BAND BREADTH IN A BAND-PASSFILTER BY VARIABLE -COUPLING COILS TEND-ING TO NEUTRALISE THE FIXED COUPLINGS.(Hochf:tech. u. Elek:akus., July, 1933, Vol.42, No. p. 37 : German Pat. No. 572 073,pub. loth March, 1933, Meissner.)


REFLEX CIRCUIT CONSIDERATIONS .- . M. Stinch-field and 0. H. Schade. (Electronics, June,1933, Vol. 6, No. 6, pp. 153-155.) See also" Trends . . .", under " Miscellaneous."

TRENDS IN RADIO DESIGN AND MANUFACTURING.(See under " Miscellaneous.")

CUTTING OUT DISTORTED RECEPTION OF OVER -WEAK SIGNALS IN RECEIVERS WITH A.V.C.,WITH HELP OF THYRATRON CONTROLLED BYA.V.C. RECTIFIER. (Hochf:tech. u. Elek:akus.,July, 1933, Vol. 42, No. i, p. 37: GermanPat. No. 572 273, pub. 13th March, 1933,R.C.A.)

AUTOMATIC VOLUME CONTROL FOR SELECTIVEFADING BY SUB DIVISION OF FREQUENCYBAND INTO SEVERAL COMPONENTS EACHWITH ITS OWN CONTROL FREQUENCY. (Hochf:tech. u. Elek:akus., July, 1933, Vol. 42, No. I,p. 37 : German Pat. No. 571 237, pub. 25thFeb., 1933, Siemens & Halske and Sedl-mayer.)

AUTOMATIC VOLUME CONTROL AND FADING COM-PENSATION [Survey of New Methods and

Wigand. (Radio, B., F.fiir Alle, August, 1933, pp. 368-378.)

THE " PEOPLE'S RECEIVER." (Radio, B., F. furAlle, August, 1933, PP 341-343.)

On the cheap German receiver dealt with inSeptember Abstracts, p. 505, 1-h col. For furthercomments (including the effect of this receiver onthe position of other single -circuit receivers) seeibid., pp. 337 and 338.

NEW PENTAGRID TUBES AND COIL -SWITCHING INTHE AMATEUR -BAND SUPERHET.-H. B.Allen. (QST, August, 1933, Vol. 17, No.8, pp. 12-17 and 72, 74.)

AN UNUSUAL 56 -Mc [Ultra -Short -Wave] SUPER -REGENERATIVE RECEIVER : DETAILS OF APORTABLE SET WITH SELF -QUENCHING DE-TECT OR .-J. G. Haydock. (QST, July,1933, Vol. 17, No. 7, pp. 14-16.)

RADIO DISTRIBUTION SYSTEM FOR APARTMENTBUILDINGS. -C. F. Boeck : Bell Labora-tories. (Rad. Engineering, June, 1933, Vol.13, No. 6, pp. 20-21.)

AERIALS AND AERIAL SYSTEMSTHEORIE UND PRAKTISCHE ANWENDUNG DER GE-

RICHTETEN STRAHLUNG (Theory and PracticalApplication of Directed Radiation [Com-prehensive Survey of Beam Aerial Theory :Calculation of Radiation Diagram of ZeesenShort -Wave Broadcasting Aerial. Part I.]).-W. Ochmann and M. Rein. (Hochf:tech.u. Elek:akus., July 1933, Vol. 42, No. I,pp. 27-32.)

With 53 literature references. For the Zeesenaerial which is taken as an example for calculation,see 1932 Abstracts, pp. 525-526.

567 October, 1933

OPERATION OF THE NEW [" Concentrator " Direc-tive Broadcasting] ANTENNA AT KYW.(Rad. Engineering, June, 1933, Vol. 13,No. 6, p. io.)

See August Abstracts, p. 449, 1-h col. SinceJanuary, 1933, the " exciter antenna " has beenworked as a quarter -wave aerial instead of half -wave, rendering it little affected by weather con-ditions and capable of being tuned and adjusted atthe base of the mast, and avoiding induction intothe wooden mast itself. Cf. also TelefunkenCompany, April Abstracts, p. 215, 1-h column, andcross references (Wilmotte).

" THE CALCULATION OF THE RADIATION CHAR-ACTERISTICS AND RADIATION RESISTANCES OFAERIAL SYSTEMS " CORRECTION OF A FOR-MULA.-Bechmann. (Hochf:tech. u. Elek:akus., June, 1933, Vol. 41, No. 6, p. 219.)See 1931 Abstracts, p. 155, r -h column.

FIELD TESTS ON RADIO COMMUNICATION OVERLONG-DISTANCE AIRCRAFT ROUTES [in-cluding the Reduction of Short -WaveFading by Alternate -Aerial Transmissionor Reception]. -von Handel, Kruger andPlendl. (See under " Propagation of Waves.")

THE PICKARD DIPOLE SYSTEM FOR ULTRA -SHORT-WAVES, USING UNTUNED FEEDER WITHSPECIAL COUPLING TRANSFORMER. -G. W.Pickard. (See Leonard and Hadlock, QST,August, 1933, Vol. 17, No. 8, pp. 23-25.)

For the transformer, heavy insulated wire iswound into a 3 -turn flat spiral, a tap being taken atthe last (outside) turn. Winding is continued in-wards to give a second, similar spiral, with a tap atthe inside, and still further (outwards) for a thirdspiral. The three coils are bound together side byside ; the dipole is connected to the two tappingsand the feeder to the two ends.

CABLE -TYPE H.F. FEEDERS [Buried, for Trans-mitting and Receiving Aerials]. -L. Lambin.(Sci. Abstracts, Sec. B, June, 1933, Vol. 36,No. 426, p. 39o.)

TWISTED -PAIR FEEDERS FOR THE TRANSMITTINGANTENNA. -G. Grammer. (QST, July, 1933,Vol. 17, No. 7, pp. 17-20.)

A VARIABLE WAVELENGTH AERIAL [Aerial Wirewinding on Metal Drum : Improved Short -Wave Reception by Adjusting to OptimumNatural Wavelength] .-(W orld-Radio, 18thAug., 1933, Vol. 17, No. 421, p. 205.)

BUCKLING LOADS IN LATTICE MASTS, PARTICULARLYIN HIGH RADIO TOWERS. -L. Ffippl. (Sci.Abstracts, Sec. B, June, 1933, Vol. 36, No. 426,P. 390.)

VALVES AND THERMIONICSMESSUNGEN AN ULT RAK URZW ELLEN R oHREN (Mea-

surements on Valves for Generating Ultra -Short Waves [Micro -Waves down to 5 cm :including a Special Duplex Valve]). -E.Gossel. (Hochf:tech. u. Elek:akus., July,1933, Vol. 42, No. 1, pp. i-io.)

A continuation of the work of Hornung and of

October, 1933 568

Collenbusch (Abstracts, 1929, p. 269 ; 1932, pp.403-404) using valves of Kohl's type with specialspiral grid (1929, p. 508 ; 1930, p. 41-two ; alsoAugust, pp. 445 and 450-two, 1-h cols.) connectedin a retarding -field circuit. Among the resultsobtained are the following :-(i) the grid circuit(formed of the spiral grid and its supports-a bow,or in one type an identical spiral with connectinglead to its mid -point) has a determining influenceon the frequency, and its mechanical and electricalproperties govern the merit and output of suchu.s.w. valves. (2) The grid/anode Lecher systemforms a circuit which, according to its tuning, canextract energy from the oscillation circuit. Nomarked effect on the wavelength of the oscillationcircuit, by the grid/anode system, could be found.(3) While all the various circuits act as extractorsof energy and by their tuning greatly influence theoutput, the coupling of the heating circuit to theoscillation circuit has an especially importantinfluence on the energy available for radiation.(4) The position of the zones of oscillation depends onthe ratio of grid -radius to anode -radius. (5) Therelation between oscillation energy and emissioncurrent i, to the positive grid shows a resonanceeffect which changes with the grid voltage. (6)The sharpness of resonance in the curves showingthe influence of the anode potential on the oscil-latory energy (Fig. 19) decreases with increasedemission current i,. (7) Preliminary experimentswith a special valve type UK 416 (Figs. 20 and 21)with two complete electrode systems galvanicallycoupled but led out separately (except the gridpotential lead), indicated that the energy conditionsare distinctly more favourable than in a singlevalve ; but a close study of such a duplex valve willbe necessary in order to obtain a really satis-factory energy increase.

RAUMLADUNGSSTROME BEI HOHEN FREQUENZEN(Space -Charge Currents at High [and Ultra -High] Frequencies.-C. Protze. (Hochf:tech.u. Elek:akus., July, 1933, Vol. 42, No. r,pp. 20-21.)

At low or medium frequencies the convectioncurrent in the space -charge zone at any momentcan be ascertained with sufficient accuracy from theLangmuir-Schottky theory or from the staticanode -current characteristics. The paper gives theresults of an investigation into the changes pro-duced by increasing the frequency, tests being car-ried out on triodes, and triodes connected as diodes,at wavelengths of 3 000 m and in the range 125to 6.1 m. For the sake of uniform accuracy ofmeasurement at all the various frequencies, thequantity measured was the rectified current ; un-fortunately, owing to the integrate character of thiscurrent, only limited conclusions as to the currentstrength at any moment can be drawn. The sinu-soidal peak voltages were measured by a com-pensation method.

Preliminary tests having shown that the currentcurve for 3 000 m could be regarded as identicalwith that derived from the static characteristic, thecurves at the higher frequencies are all comparedwith the 3 000 m curve, and in Figs. 5 and 6 theordinates represent the ratio of the current at thehigh frequency to the corresponding value at the


standard frequency ; the abscissae, as in the previouscurves, representing the peak voltages. The formerdiagram shows especially clearly the effect ofincreasing frequency, the curves displaying anincreasing tendency to a " hump," with a maximumaround 75 v (for the particular diode concerned)where the ratio of the currents rises to over 1.6 for the6.1-m wave, compared with less than 1.3 for the 75-mwave. This effect is explained by considering thediagram of potential distribution in a diode (Fig.7) ; at high frequencies a displacement current issuperposed on the convection current, increasingin effect as the frequency is raised. The result isthat the sagging curve due to the convection currentis progressively tautened, the potential minimum infront of the cathode decreases, and the current in-creases.

A THEORY OF AVAILABLE OUTPUT AND OPTIMUMOPERATING CONDITIONS FOR TRIODEVALVES.-M. V. Callendar. (Proc. Inst.Rad. Eng., July, 1933, Vol. 21, No. 7, pp.909-929)

From the Lissen laboratories. Author's sum-mary :-" The output characteristics of triode andpentode valves have been investigated in the pastby many workers, using either the static method ofplotting a whole family of curves, or the dynamicmethod of measurement with a harmonic analyser :these have not, however, led to many simplifyinggeneralisations, and are cumbersome and difficultmethods if they are to be applied to each individualvalve. In this paper, the form of the triode curvesis first investigated experimentally, and the allow-able limits of dynamic swing thus determined for anygiven per cent. harmonic : on this basis, a series ofexpressions are mathematically developed givingthe required output characteristics in terms of aneasily obtained valve constant (i.e., the valvealternating -current resistance at V = roo, o ;this is the standard figure quoted by English manu-facturers). Curves are given for output and correctplate current with various values of resistive loadand of per cent. harmonic limit, with or withoutconditions of limited anode dissipation. The caseof the practical load with low direct -current re-sistance is examined, and the results are checkedwith a harmonic analyser ; the paper concludeswith a few practical rules. It is hoped in a futurepaper to extend the analysis to cover push-pullcircuits, the pentode valve, and the practicalinductive load."

DIE GRAPHISCHE KLASSIFIKATION DER ELEKTRON-ENROHREN (The Graphical Classification ofValves [Valve Diagrams including " Giite "]).-S. J. Zilitinkewitsch. (Physik. Zeitschr.der Sowietunion, No. 6, Vol. 3, 1933, pp.606-616.)

See also Gundlach, January Abstracts, p. 41,1-h col. Author's summary :-" Starting with twogenerally known main equations linking the fourprincipal valve parameters, R, = i.t/S and Gthe writer makes use of the fact that these can berepresented in logarithmic form by a rectangulardouble -co-ordinate system with a mutual inclinationof 45°. In this way a network diagram is obtained,formed of the following four axes : the ' amplification


axis' (p.), the ' resistance axis' (R,), the ' mutualconductance axis' (S) and the Giite axis' (G).Each point of this network corresponds to a definitecombination of characteristic valve parameters,which allows these valves to be arranged con-veniently in the network.

" Such a classification greatly simplifies thecomparison of one valve with another and theirchoice for any particular practical purpose. Thewriter points out that the progress in valve designis linked up with the position along the ' Giiteaxis' in the direction of higher values of G."The specimen diagram given includes transmittingand receiving valves of prominent types made inRussia, America and Germany.

VACUUM TUBE CHARACTERISTICS IN THE POSITIVEGRID REGION BY AN OSCILLOGRAPHICMETHOD.-H. N. Kozanowski and I. E.Mouromtseff. (Proc. Inst. Rad. Eng., August,1933, Vol. 21, No. 8, pp. 1082-1096.)

Authors' summary :-A method determining" complete " plate and grid characteristics ofvacuum tubes in the positive grid region by means ofoscillographic recording has been developed. Acondenser of high capacity furnishes a single pulseof grid excitation which can be made to cover theentire region from any desired positive grid voltageto zero. Owing to the rapidity of this excitation,instantaneous power input to the tube of twentyto thirty times nominal rating has been recordedwithout danger to the tubes.

Several typical complete charts of plate and gridcharacteristics obtained by this method for anexperimental tube of the so-called " 50 -watt"type are given. The experimental procedure inobtaining these characteristics is discussed indetail. Also the circuit for obtaining oscillo-graphically the highly important " composite diodeline," with E, = E is described. The inadequacyof the usual logarithmic extrapolation of zero -gridcharacteristics into the positive grid region is dis-cussed.

The complete plate and grid current charts,which can be obtained accurately only by anoscillographic method, are practically indis-pensable in precalculation of class B and class Cperformance. The method has been successfullyused in studying the characteristics of the smallestand the largest existing tubes in this country.

COEFFICIENT OF AMPLIFICATION OF TRIODES [Dis-cussion of Chief Methods of Measurement].-C. Dei. (Sci. Abstracts, Sec. B, June, .1933,Vol. 36, No. 426, p. 396.)

" Discusses the three chief methods . . . thoseof Miller, Appleton, and lo Surdo. The last method,devised in 1931, places a diode in parallel with theelectrostatic voltmeter in the plate circuit of thetriode to keep the plate current steady."

NEW VALVES AT THE BERLIN RADIO EXHIBITION :BINODES AND 9 -WATT PENTODES. (Radio,B., F. fur Alle, August, 1933, pp. 343-347.)

Both ordinary and screen -grid " binodes " aredescribed and their circuits explained. The pentodesdiscussed are the directly and indirectly heated type,both having the feature that the screen -grid potential

October, 1933

is the same as the anode potential, which simplifiesthe design of the set.

" ELECTROMETER " VALVES versus ORDINARY TRI-ODES IN VALVE VOLTMETERS WITH HIGHINPUT IMPEDANCE.-McFarlane. (See ab-stract under " Measurements and Stan-dards.")

VALVES FOR SOUND PICTURE SYSTEMS.-Kelley.(See abstract under " Acoustics and Audio -frequencies.")

A MEASURING SYSTEM FOR MICROPHONIC NOISECURRENTS.-Kelley. (Ibid.)

A STUDY OF HUM GENERATION IN VACUUM TUBESAS AFFECTED BY HEATER DESIGN.-J. J.Glauber and A. G. Campbell. (Rad. Engin-eering, June, 1933, Vol. 13, No. 6, pp.12-14 and 17.)

The paper concludes : " The hum in both typesof tubes employing the double -hairpin constructionfor 6.3 v operation is less than for all other types ofheaters examined. . . . It may be concluded thatall new tubes could and should be designed aroundthis type of heater, except in special cases."

EMISSION FROM OXIDE -COATED CATHODES. II.-M. Benjamin and H. P. Rooksby. (Phil.Mag., Aug., 1933, Supp. No., Series 7, Vol.16, No. 105, pp. 519-525.)

For reference to Part I see June Abstracts, p.329, r -h col. The writers find that in the case of acathode activated with Bao/sro, " the constituentoxides can exist in all states ranging from a simplemixture to a homogeneous solid solution ; andthere exists a definite relationship between thephysical state of the oxides and the thermionicemission obtained from them."

THE PROCESSING OF THORIATED TUNGSTEN FILA-MENTS.-M. A. Ausman. (Rad. Engineering,June, 1933, Vol. 13, No. 6, pp. 15-17.)

SECONDARY EMISSION OF ELECTRONS FROM MOLYB-DENUM.-P. L. Copeland. (Journ. FranklinInst., May, 1933, Vol. 215, No. 5, pp. 593-598.)

An account of an experimental determination ofthe number of secondary electrons to primaryelectrons of various energies incident on a molyb-denum plate.

OBSERVATIONS OF A BARIUM CATHODE, SPUTTEREDON NICKEL, WITH AN ELECTRON MICROSCOPE.-E. Briiche and H. Johannson. (Zeitschr.f. Physik, 1933, Vol. 84, No. 1/2, pp. 56-58.)

THE APPLICATION OF GRAPHITE AS AN ANODEMATERIAL TO HIGH -VACUUM TRANSMITTINGTUBES.-E. E. Spitzer. (Proc. Inst. Rad.Eng., August, 1933, Vol. 21, No. 8, pp.1075-1081.)

The higher radiation emissivity gives lowerglass temperatures and therefore less danger ofglass electrolysis and strain cracking. Anodewarping is avoided, giving much greater electrical

October, 1933

uniformity. With proper manufacturing methods,there is no sacrifice in valve life when graphite issubstituted for molybdenum.

A LIFE -TEST POWER SUPPLY UTILISING THYRATRONRECTIFIERS.-H. W. Lord. (Proc. Inst.Rad. Eng., August, 1933, Vol. 21, No. 8,pp. 1097-1102.)

Author's summary :-Thyratron rectifiers forsupplying high -voltage direct current to radiotransmitting tube life -test racks are superior tomotor -generator sets where quietness, flexibility,low operating cost, and safe operation over longintervals of time are desirable.

An installation for supplying typical voltages,in conjunction with usual forms of electric powersupplies, is described. This consists of two high -power rectifiers for 425 and 1000 volts d.c. platesupplies and a low -power 125 -volt d.c. rectifier forbias voltage. A control circuit provides completeprotection against faults detrimental to tubeoperating conditions such as low filament voltage,low bias, and resumption of power after failure,ensuring a maximum of life -testing hours available,consistent with safe operation.

DIRECTIONAL WIRELESSTHE CATHODE-RAY COMPASS.-W. Ende and M. H.

Gloeckner A.E.G. (Zeitschr. f. Flug-technik u. Motorluftschiffahrt, 28th Oct.,1932, Vol. 23, No. 20, pp. 603-609.)

Including some constructional details of thecompass dealt with in 1931 Abstracts, p. 271, r -hcol.-two, Salmony Briiche. The systematic in-terpretation of the varying indications is discussed,and the influence of the nearness of the magneto.A test calibration is compared with that of amagnetic compass. For a special A.E.G. low -voltage cathode-ray tube see 1932 Abstracts, p.653, 1-h col., and for Schwerin's " electronic com-pass " ibid., p. 35o, r -h col.

EIN NEUES STROBOSKOPISCHES EIGENPEILGERAT(A New Stroboscopic Direction Finder [foruse on Moving Craft, particularly Aircraft]).-R. Hardy. (Hock!:tech. u. Elek:akus.,July, 1933, Vol. 42, No. I, pp. 35-36.)

Long summary of a L'Aeronautique paper on theinstrument dealt with in April Abstracts, p. 216,r -h col. An accuracy of one -sixth of a degree isclaimed.

RADIO SYSTEM FOR LANDING AIRCRAFT DURINGFOG : INSTALLATION AT NEWARK AIRPORT.-H. Diamond. (Electronics, June, 1933,Vol. 6, No. 6, pp. 158-161.) See alsoSeptember Abstracts, p. 5o8, 1-h column,and below.

RADIO SYSTEM FOR BLIND LANDING OF AIRCRAFT.-Bureau of Standards. (Journ. FranklinInst., May, 1933, Vol. 215, No. 5, pp. 599-6o1.)

A note on, practical tests of the radio system forthe blind landing of aircraft, carried out at Newark,N.J. See also Diamond, above.


ACOUSTICS AND AUDIO -FREQUENCIES

MESSUNG MECHANISCHER WIDERSTANDE BEIDREHBEWEGUNGEN (The Measurement ofMechanical Resistances to Rotary Motion[and an Electrical Method of producingMechanical Resistances of Any DesiredValue and Kind]).-E. Paolini. (Zeitschr.f. tech. Phys., No. 8, Vol. 14, 1933, pp.332-335.)

After discussing the growing importance inacoustics of mechanical filters (e.g. in gramophonetechnique and sound analysis) and mentioningthe various ways of obtaining dissipative resistance(vanes in oil or glycerine ; rubber rods ; eddycurrent braking, etc.) the writer describes his methodof measuring such resistances, defined as the ratioof torque to rate of angular displacement, for afrequency range of about 20-100 c/s. Two rigidlyconnected rotating -coil systems are used, oneacting as the torque producer while the other,by the voltage generated in it, measures the rateof angular displacement produced by the torque ;the voltage is determined with the help of a Larsencompensator and an amplifier, dry -plate rectifier,and d.c. meter combination. The torque -producingsystem is fed by a heterodyne note -generator.Specimen measurements of a vane in oil and in aspecial very viscous fluid are shown, and the feebledissipative resistance yielded by eddy -currentdevices is mentioned. This leads to an outlineof the writer's method of obtaining large andadjustable resistances by the amplification, correctas to phase, of the potentials generated in a short-circuited coil and the application, with the rightpolarity, of the amplified potentials to a coil systemrigidly connected to the first coil. Such an arrange-ment is very suitable for use as the terminal resis-tance for a mechanical filter system, withoutwhich a smooth curve in the pass -band zone ofthe system cannot be obtained, as is shown (Fig. 7)by the curves of transmission equivalents with andwithout a terminal resistance.

USE OF CATHODE-RAY OSCILLOGRAPH FOR OBTAIN-ING FREQUENCY -RESPONSE CURVES [of Loud-Speakers].-A. H. Davis. (Phil. Mag.,Aug., 1933, Supp. No., Series 7, Vol. 16,No. 105, pp. 408-410.)

In this circuit an oscillator of variable frequencyand constant voltage, supplies the loudspeakerwith an e.m.f. of constant magnitude but variablefrequency ; the acoustical output of the loud-speaker is measured by a microphone circuit,uniformly responsive at all frequencies, whoserectified output is applied to one pair of plates ofa cathode-ray oscillograph, giving a verticaldeflection. " The same oscillator supplies currentto a weighting network so adjusted that the output,when rectified and smoothed, is closely related tothe logarithm of the frequency of the excitingcurrent." This output is applied to the other pairof plates of the oscillograph, so that the horizontaldeflection of the spot is determined by the frequencyof the note concerned. As the frequency is variedthe spot traces a curve giving the relation betweenthe pitch of the note and the loudspeaker out-put.


VOLUME CONTROL BY POTENTIAL DIVIDER CON-SISTING OF A SERIES OF MUTUALLY DE -COUPLED INDUCTANCES EACH WITH THESAME TIME CONSTANT AS THE PICK-UP.-(H ochf :tech. u. Elek:akus., July, 1933,Vol. 42, No. I, p. 37: German Pat. No.570 703, pub. 18th July, 1933, Bethenod.)

ECOUTEUR TELEPHONIQUE ELECTROSTATIQUE (Elec-trostatic Telephone Receiver).-G. Longo.(L'Onde Elec., May, 1933, Vol. 12, No. 137,pp. 262-272.)

The writer first enumerates the various factorswhich decrease the effectiveness of electrostaticloudspeakers-the necessarily thin diaphragm (forlightness) and consequent tight stretching (forstable equilibrium) : the need for damping (toreduce the natural frequencies of the diaphragm)and the small gap (causing non-linear distortionfor any but very small amplitudes). " Theseobjections, which would substantially be found inany electrostatic telephone receiver constructedon the same principles, seem to be avoided almostcompletely in the proposed system, which shouldgive a response curve practically rectilinear andhorizontal over the whole range of frequenciesgenerally desired to be reproduced."

A number of small (i.i X 1.2 cm) thin metallic(aluminium) plates, each rectangular and with aside lug for connection, are arranged in a pile asin a multi -plate condenser, each being separatedfrom its neighbour by a U-shaped spacing pieceof mica, e.g., oriented alternately so that successiveinter -plate airgaps are open to the air at oppositeedges of the pile. By suitable location of the lugs,the plates are connected in 3 groups, r, 5, 9 . . .

connected to a constant -{-ve potential, 3, 7, II . . .

to a constant - ye potential, and the even numbersconnected to the varying potential. The pile ismounted in the telephone receiver case so that oneset of openings faces the ear, the other set havingfree access to the air. The behaviour of such areceiver is examined quantitatively for telephony.For c.w. telegraphy it can be used without detec-tion, by modulating the polarising voltage at asuitable audio -frequency; but here of course theadvantages of fidelity over the whole frequencyrange are not made use of, and a tuned single -diaphragm design would probably be more efficient.

ELLIPSOIDAL FUNCTIONS AND THEIR APPLICATIONTO SOME WAVE PROBLEMS [including Trans-mission of Waves through a Circular Aper-ture and Scattering of Waves by a CircularDisc].-Hanson. (See under " Miscel-laneous.")

VACUUM TUBE AND PHOTOELECTRIC TUBE DEVELOP-MENTS FOR SOUND PICTURE SYSTEMS.-M. J. Kelley. (Bell Tel. System MonographB-694, 21 pp.)

Including a description of the 262A, an indirectly -heated triode making the use of a.c. possible inamplifiers with flat frequency characteristics andover-all gains of the order of roo db. A measuringsystem for microphonic noise currents is described ;also the Western Electric 263A rectifier for mainssupply of the sound lamp (usually fed from accumu-lators) and valve filaments of reproducing systems.

October, 1933

The high -sensitivity caesium -oxygen -silver photo-electric cell type 3A is also discussed (see also 1932Abstracts, pp. 592 and 648).

A THEORY OF AVAILABLE OUTPUT AND OPTIMUMOPERATING CONDITIONS FOR TRIODE VALVES.-Callendar. (See under " Valves andThermionics.")

DETECTORS AS OUTPUT VALVES GIVING 800 !MN' :POWER GRID DETECTION WITH HIGH -IMPEDANCE CHOKE IN PLACE OF GRID LEAK.-de Cola. (See abstract under " Recep-tion.")

DAS GRUNDPROBLEM DER ELEKTROAKUSTIK (TheFundamental Problem of Electro-Acoustics[the Conditions for Fidelity of Reproductionof the Original Sound]).-A. Koerts. (Hochf:tech. u. Elek:akus., June, 1933, Vol. 41,No. 6, pp. 214-218.)

The writer evolves a general theory by whichthe reproduction of any sound phenomenon by anytransmission system can be represented. Everysystem possesses a characteristic function, com-bining all its properties, which can only in isolatedcases be given clearly by calculation but whichcan always be determined experimentally by thecathode-ray oscillograph. For any " complete "sound the conditions for subjective identity be-tween primary and reproduced sound are the sameas for objective identity ; but for sounds repre-sented by trigonometrical series with slowlychanging coefficients, subjective fidelity dependson a quantity [see text below equation 12] which isclosely connected with the ordinary frequencycharacteristic. In this latter class come humanspeech and all music.

Sounds which last a very short time, on the otherhand, can be reproduced with subjective fidelityif the dampings of the natural movements of thetransmitting system are greater than those of the" physical " (as contrasted to the " physiological ")ear. In a later paper the writer hopes to discussthe various practical transmission systems in thelight of his theory.

A NOTE ON NONLINEARITY IN TRANSDUCERS USEDIN COMMUNICATION [and Its Correction bya Series Compensating Transducer].-Caporale. (See under " Properties of Cir-cuits.")

AMPLITUDE DISTORTION COMPENSATED BY IM-PEDANCE DECREASING WITH CURRENT[Thyrite] IN PLATE CIRCUIT OF DISTORTINGVALVE.-Massa. (See preceding reference.)

BESTIMMUNG DES SCHLUCKGRADES BEI SCHRAGEMSCHALLEINFALL MIT HILFE STEHENDERWELLEN (Measurement of the AcousticAbsorption Percentage at Oblique Incidence,with the Help of Standing Waves).-L.Cremer. (E.N.T., July, 1933, Vol. ro, No.7, pp. 302-315.)

For normal incidence, the " absorption per-centage " of a material (a = I - R2, where R isthe ratio of reflected to incident amplitude) can

October, 1933 572

conveniently be measured by the standing -waveprocedure using a Kundt's tube ; for obliqueincidence, however, the method fails owing to itsessentially one-dimensional nature. But the useof standing waves is still practicable if the tube isabandoned and the exploration carried out in freespace. This entails the use of a large expanse ofhomogeneous wall of the material investigated, anequally homogeneous plane wave -train which can bedirected on to the wall at any desired angle and canproduce an interference field, a very small receiverby which this field can be explored, and aboveall an absence of other wave -trains encroachingon the field and upsetting the interference patterns.

The writer describes an equipment for carryingout such measurements. The last requirement -the screening of the explored field from disturbance-is fulfilled by an elaborate system of rock -woolmats. As the mobile suspended microphone movesfrom point to point of the field, the correspondingpressure amplitudes are recorded on a strip movingabout 5 cm per minute ; the microphone explora-tions are controlled by hand in time with a metro-nome. The equipment not only enables the varia-tion of the absorption percentage with the angleof incidence to be determined, but also allows ageneral investigation of standing waves in spaceto be carried out. Various tests on commercialinsulating substances, both rigid (e.g. celotex) andsoft porous materials such as the rock -wool matsused for the screening, are described. With theformer the absorption percentage increases as theincidence becomes more oblique ; with the latterit decreases.VARIATION OF ACOUSTICAL ABSORPTION COEFFI-

CIENTS OF POROUS BODIES WITH THE ANGLEOF INCIDENCE AND THE FREQUENCY. -V. Kiihl and E. Meyer. (Sci. Abstracts,Sec.A, June, 1933, Vol. 36, No. 426, pp. 618-619.)

THE RESONANCE FREQUENCIES OF TWO COUPLEDOSCILLATORY CIRCUITS [and a New Methodof Frequency Analysis and Non -LinearDistortion Measurement]. -Fehr and Kreiel-sheimer. (See under " Properties of Circuits.")

A NEW MAINS -OPERATED AUDIO -FREQUENCYOSCILLATOR [20-10 000 c/s Range, withGood Stability and Waveform]. -C. H. W.Brookes -Smith. (Journ. Scient. Instr.,Aug., 1933, Vol. 10, No. 8, pp. 251-255.)

PROBLEM RELATING TO THE THEORY OF THE RAY -LEIGH DISC. -M. Kotani. (Sci. Abstracts,Sec.A, June, 1933, Vol. 46, No. 426; p. 617.)

MEASUREMENTS OF AMPLITUDE AND PHASE OF THEOCTAVE IN A TUNING FORK [QuantitativeConfirmation of Helmholtz's Theory dfCombination Tones : Mechanism of theFormation of the Octave]. -B. Derjaguin.(Physik. Zeitschr. der Sowietunion, No. 6,Vol. 3,1933, pp. 574-585.)

STROBOSCOPIC DEMONSTRATION OF TONES [Modifi-cation of Scripture's " Strobilion "].-W.Anderson and H. Lowery : Scripture.(Journ. Scient. Instr., July, 1933, Vol. I 0,No. 7, pp. 203-204.)


ON THE PROPERTIES OF JAPANESE VOWELS ANDCONSONANTS [Determination of the " For-manten " from Oscillograms].-J. Obataand T. Tesima. (Japanese Journ. of Physics,Vol. 8 [1932], pp. 1-24, with 20 Figures.)

ZUR SCHALLEISTUNG EINES FLUGELS (The SoundOutput of a Grand Piano). -E. Ltibcke andK. H. Wernicke. (Hochf:tech. u. Elek:akus.,June, 1933, Vol. 41, No. 6, pp. 212-214.)

The output powers for single notes of a " concertgrand " were measured in a hall. The resultingspectrum is discussed, together with reasons forthe differing powers on the various notes. Whenthe tests were repeated in an entirely differenttype of room, the spectra agreed quite well (at leastwith the piano open) so that the method employedshould be useful in practice.

RANGE OF SOUND INTENSITY IN MUSIC. -K. W.Wagner. (Sci. Abstracts, Sec.A, June, 1933,Vol. 36, No. 426, pp. 619-620.) See alsoApril Abstracts, pp. 218-219.

ACOUSTIC METHOD OF DISTANCE MEASUREMENT. -H. Yagi and S. Matusuo. (Sci. Abstracts,Sec.B, June, 1933, Vol. 36, No. 426, p. 397.)

" The frequency of an emitted sound -wave ismade to vary linearly with time. A receiver picksup the out -going wave and also the wave reflectedfrom a wall, the distance of which is derivable fromthe beat note produced."

INFLUENCE OF SUPERSONIC WAVES ON CHEMICALPROCESSES. -H. Beuthe. (Sci. Abstracts,Sec. A, May, 1933, Vol. 36, No. 425, p. 489.)

'OBER EINIGE BEOBACHTUNGEN AN SCHWINGENDENPIEZOQUARZEN UND IHREM SCHALLFELD(Some Observations on Oscillating Piezo-quarz Crystals and their Acoustic Field). -K. Bucks and H. Muller. (Zeitschr. f.Physik, 1933, Vol. 84, No. 1/2, pp. 75-86.)

VOICE TRANSMISSION ON A BEAM OF LIGHT [De-monstration Apparatus]. -R. R. Ramsey.(Science, 4th Aug., 1933, Vol. 78, No. 2014,p. 105.)

PHOTOTELEGRAPHY AND TELEVISIONSUR QUELQUES RECEPTIONS LOINTAINES DES

EMISSIONS DE RADIO -VISION DE LONDRES(Some Distant Receptions of TelevisionTransmissions from London [and the Occur-rence of Multiple Images]). -D. Bodroux andR. Rivault. (Comptes Rendus, 17th July,1933, Vol. 197, No. 3, pp. 231-232.)

To decide whether the frequent occurrence ofmultiple (5 or 6) images was due to propagationphenomena or to reflection effects in the amplifier, aspecial receiver was constructed with a pass -bandof 25-30 kc/s and without any iron -cored com-ponents. This gave much finer images but did noteliminate the multiple effects. The size and bril-liancy of the pictures were progressively increasedfirst by replacing the original Nipkow disc by aBrillouin lens disc and a " point " neon lamp(yielding a satisfactorily bright image 14 X 35 cm)and then by replacing the neon lamp by a Kerr cell


and a low -voltage lamp with 100 -watt filament(giving a good white -light image 25 x 58 cm).The writers thus had at their disposal either alarge, sufficiently bright image or a very brightsmall image, and were therefore in a position tophotograph the various results in order to measurethe observed delays. Unluckily these measure-ments have been prevented, up to the present, byinterference from the Frankfort Station, which since8th April has continued its programmes up tomidnight.

NEW OPTICAL ASSEMBLY FOR TELEVISION PRO-JECTION RECEIVERS [Kerr Cell Equipmentusing both Ordinary and ExtraordinaryRays : Overall Efficiency Greatly Increased].-N. Levin. (Marconi Review, May/June,1933, No. 42, pp. I 1-12.)

POLARISING LOUVRES AND THEIR APPLICATION[to replace Nicol's Prisms]. -C. F. Smith.(JMW. Scient. Instr., July, 1933, Vol. io,No. 7, pp. 212-214.)

THE ICONOSCOPE [Three Million Photocells scannedby Cathode Ray].-Zworykin. (World -Radio, 4th Aug., 1933, Vol. 17, No. 419, p.135.)

Paragraph giving some details of the " electriceye " television transmitter referred to in AugustAbstracts, p. 453, r -h column.

GOOD MATCHING CONDITIONS FOR LOW -RESISTANCEBARRIER -LAYER PHOTOCELL BY USE OFMAGNETRON IN FIRST AMPLIFIER STAGE.(Hochf:tech. u. Elek:akus., July, 1933, Vol.42, No. I, p. 37 : German Pat. No. 570 599,pub. 17th Feb., 1933, Zeiss Ikon.)

COMPARISON OF PHOTO -EMISSIVE, PHOTO -CON-DUCTIVE AND PHOTO -VOLTAIC CELLS. -A. J.McMaster. (Summary in Electronics, June,1933, Vol. 6, No. 6, p. 171.)

CAESIUM -OXYGEN -SILVER PHOTOELECTRIC CELL OFHIGH SENSITIVITY, FOR SOUND PICTURESYSTEMS. -Kelley. (See abstract under" Acoustics and Audio -frequencies.")

THERMISCHE UND PHOTOELEKTRISCHE EMISSIONVON CASIUM-CASIUMOXYDKATHODEN UNDDEREN BEEINFLUSSUNG DURCH EINBAU VONCASIUMATOMEN IN DAS DIELEKTRI KUM(Thermal and Photoelectric Emission fromCaesium -Caesium Oxide Cathodes and theInfluence Exerted on Emission by Inclusionof Caesium Atoms in the Dielectric [i.e. inthe Oxide Film : Both Emissions Improved]).-J. H. de Boer and M. C. Teves. (Zeitschr.f. Physik, 1933, Vol. 83, No. 7/8, pp. 521-533.)

PHOTOELECTRIC EFFECT OF CAESIUM VAPOUR. -J. Kunz. (Phys. Review, 15th June, 1933,Series 2, Vol. 43, p. 1052 : abstract only.)

INTERNAL PHOTOELECTRIC ABSORPTION IN HALIDECRYSTALS . -R. W. Gurney. (Proc. Roy. Soc.,3rd July, 1933, Vol. 141, No. A 843, pp. 209-2 I 5.)

573 October, '933

THE CRYSTAL PHOTOEFFECT .-A. Joffe and A. F.Joffe. (Nature, 29th July, 1933, Vol. 132,pp. 168-169.)

The writers of this letter describe experimentstesting the fundamental equations of the diffusiontheory of the photoelectromotive force. They alsoconclude that the carriers of the photocurrent aredifferent from the heat electrons. For recentpapers on the subject see Abstracts, 1932, p. 232(Dember) ; July, p. 400 (Teichmann) ; August, p.455 (Joffe and Joffe) ; all 1-h columns. AlsoDeaglio, below.

PIIOTOELEKTRISCHER EFFEKT BEI MONOKRI STAL-LEN VON CUPRIT (Photoelectric Effect inSingle Crystals of Cuprite).-R. Deaglio :Dember. (Zeitschr. f. Physik, 1933, Vol. 83,No. 3/4, pp. 179-183.) See also AugustAbstracts, p. 455, 1-h column.

THE PHOTOELECTRIC EFFECT OF CRYSTALS OFARGENTITE, PROUSTITE AND PYRARGYRITE.-G. Athanasiu. (Comptes Rendus, 3rdJuly, 1933, Vol. 197, No. 1, pp. 42-44.)

WINKELABHANGIGKEIT DES PHOTOEFFEKTES ANISOLATOREN IM POLARISIERTEN LICHT(Angular Variation of the PhotoelectricEffect on Insulators in Polarised Light). -S. M. Spies. (Zeitschr. f. Physik, 1933,Vol. 82, No. 5/6, pp. 355-371.)

The experiments described in this paper confirmand extend results obtained by Wehnelt and Sch-merwitz (193o Abstracts, p. 401, r -h column).

REMARK ON THE PAPER BY Q. MAJ ORANA : " ONA NEW PHOTOELECTRIC EXPERIMENT." -A. Etzrodt : Majorana. (Physik. Zeitschr.,15th April, 1933, Vol. 34, No. 8, pp. 338-34o.)

For references to the paper referred to and otherssee March Abstracts, p. 170. The writer of thisnote attributes Majorana's results to the effectof heat radiation on the electrical resistance ofthe conducting metal, rather than to a photoelectriceffect, and describes experiments in support ofhis view. On p. 34o Majorana replies that he doesnot consider that the experiments of Etzrodt excludethe possibility of the photoelectric effect beingpresent.

VEER DIE BESTIMMUNG DER LICHTELEKTRISCHENGRENZWELLENLANGE AM RHENIUM (TheDetermination of the Limiting Photo-electric . Wavelength for Rhenium). -A.Engelmann. (Ann. der Physik, 1933, Series 5,Vol. 17, No. 2, pp. 185-208.)

DIE OPTISCHE ABSORPTION AN LICHTELEKTRISCHZWEIFACH SELEKTIV WIRKENDEN OBER..FLACHENSCHICHTEN [The Optical Absorptionat Surface Films with Doubly SelectivePhotoelectric Effect). -W. Kluge. (Physik.Zeitschr., 15th June, 1933, Vol. 34, No. 12,PP. 465-472.)

THEORY OF THE COMPLEX PHOTOELECTRIC EFFECTS[Liberation of More than One Electron by aPhoton falling on a Multi -Electron Atom]. -L. Goldstein. (Comptes Rendus, 24th July,1933, Vol. 197, No. 4, pp. 304-306.)

October, 1933

EXTENSION OF FOWLER'S THEORY OF PHOTO-ELECTRIC SENSITIVITY AS A FUNCTION OFTEMPERATURE.-A. T. Waterman and C. L.Henshaw. (Phys. Review, Ist July, 1933,Series 2, Vol. 44, No. I, pp. 59-6o.)

MEASUREMENTS AND STANDARDSDIE MESSUNG HOCHFREQUENTER SPANNUNGEN

MIT DEM FADENELEKTROMETER (The Measure-ment of [Ultra-] High -Frequency Potentialswith the Thread Electrometer).-G. Ul-bricht. (Hochf:tech. u. Elek:akus., July,1933, Vol. 42, No. I, PP. 21-24.)

The writer points out that the thread electro-meter, a particularly suitable and simple instrument,is far too little used. For medium potentials(20-30o v), such as are encountered in small valvetransmitters, the two -thread type is preferable ;for small and for high potentials the single -threadtype, without polarising voltage.

As regards its use at very high frequencies,the fact that it has a certain capacity may leadto errors ; the writer therefore investigates thispoint. He derives, with the help of the telegraphyequations, a formula showing the variation withfrequency of the potential acting at any point onthe filament, and particularly at the importantmid -point, whose movement is observed ; thusthe frequency can be calculated at which, for agiven thread, a fixed percentage error is produced.Applied to a certain double -thread instrument witha 6 -cm Wollaston thread 41» thick, this gives thecurve of Fig. 3, which shows that the instrumentcan be used down to a wavelength of 2.7 m if anerror of 1% is tolerated. Care, however, must betaken that with high voltages at such wavelengthsthe thread is not burnt out by the capacity currentsflowing in the electrometer. The use of short,thick threads should allow still lower wavelengthsto be dealt with, but a limit is set by the capacityat the leading -in insulators. Fig. 6 shows thecalculated behaviour of a 4 -cm thread of thickness81.L (capable of standing a capacity current of5o mA such as might be produced by a voltage of200 v at a wavelength of 1.86 m) from which itappears that even at a wavelength of 15o cm, theerror would be only I %. The short, thick filament,however, would naturally decrease the sensitivityof the instrument. See also next abstract.

DAS ZWEIFADENELEKTROMETER ALS WECHSEL-SPANNUNGSMESSER FUR HOHE FREQUENZEN(The Double -Thread Electrometer as anA.C. Voltmeter at High [and Ultra -High]Frequencies).-H. F. Nissen. (Ibid., pp.24-26.)

See also Ulbricht, above. The present writerrecommends the two -thread type exclusively, forfrequencies of io7 cis upwards, on account of itsfreedom from asymmetrically arranged electrodes." At very high frequencies the double -threadelectrometer shows certain errors, which mustbe considered in designing the instrument and whichlimit the frequency range in which it can be em-ployed." Dealing in succession with the variouscauses of these errors, he begins with the voltagedrop along the thread : the latter, for the sakeof sensitivity, must be very thin, but as the voltage -drop error increases with the square of the frequency


it is necessary to keep the resistance low. Thisis also necessary to prevent burning -out. Thinsilver or aluminium strips are suitable.

Another component of the error at very highfrequencies is that due to the electrodynamicforces between the threads, caused by displace-ment currents. Since these forces increase withthe square of the frequency, while the electro-static forces are independent of frequency, thiserror, like the voltage -drop error, increases withthe square of the frequency ; it varies very littlewith the voltage. Another source of error is thealteration of capacity with the deflection ; sincethis change can be measured a correction can bemade for it ; but in, certain cases (e.g. if the changehas a marked effect on the resonance curve of anoscillatory circuit) this effect makes the electro-meter unusable.

The potential rise due to a long connection tothe instrument is considered to be negligibleprovided the length, including the electrometersystem, does not exceed 0.05 A. Finally, influencestending to impair the accuracy of observations areexamined : to reduce vibration a suitable mountingand the choice of a thread with the least possiblemass are necessary. Elastic hysteresis woulddiminish the accuracy, but in the tests mentionedbelow no sign of any error due to this was found.

The paper ends with a description of a specialinstrument (resembling a Wulf electrometer) andits testing. The silver strips are lengthened down-wards by non -conducting threads connected to anelastic bow of quartz; this plan not only increasesthe sensitivity but also allows the microscopeobjective to be brought closer to the unchargedthreads than it could be to the strips. The varioustypes of error were found, by calculation from thedata of this instrument, to be as follows :-noappreciable error due to voltage -drop can occuruntil wavelengths very much shorter than 6 mare reached ; electrodynamic forces can cause a

% error at a wavelength of 90 cm ; the capacitychange for a deflection of 0.7 mm amounts to0.04 µµF ; the error due to potential rise alongthe connecting lead would reach about 1% at awavelength of 1.5 m.

Comparing the thread electrometer with otherpotential measuring methods for high frequencies,the writer says that valve -voltmeters, Kerr effectdevices and cathode-ray tubes all fail, or requireinvolved corrections, at wavelengths below about3o m. Rohde's voltmeter (1931 Abstracts, p. 393)seems the only rival, but takes longer to use.ARRANGEMENT FOR MEASURING THE AMPLIFYING

POWER OF H.F. AMPLIFIERS.-(Hochttech.U. Elek:akus., July, 1933, Vol. 42, No. 1,p. 37 : German Pat. No. 57o 683, pub.18 Feb., 1933, Telefunken and Runge.)

The output is connected to the input throughan attenuator in the form of a variable inductivecoupling, and the amplification is measured by thecoupling necessary to produce oscillation.

A FIELD -INTENSITY SET [for Field -StrengthMeasurements on both Ground and SkyWaves, and Some Results].-A. L. Greenand H. B. Wood. (Journ. Inst. Engineers,Australia, Jan., 1933, Vol. 5, No., I, 8 pp.)

Authors' summary :-" Following the Kennelly-


Heaviside layer investigations at Jervis Bay,distant 140 km from the sending station, itbecame desirable to continue the work at muchshorter distances from the transmitter, of theorder of 20 km. This called for short investigationof sky wave and ground wave field -intensities, forwhich apparatus was developed.

" The paper describes a field -intensity set forboth ground and sky waves. The developmentof a vacuum -tube millivoltmeter enabled thecomponents of the artificial signal injectionapparatus to be calibrated directly ; the accuracyof the complete assembly therefore depends onlyon the precise measurement of the geometricaldimensions of the loop aerial, and on the calibrationof the millivoltmeter. The use of a substantiallylinear vacuum -tube voltmeter, following the ampli-fier in the receiver, greatly facilitated the measure-ment of low field -intensities.

" A few results are given of practical tests withthe apparatus in measuring field -intensities ofboth ground and sky waves, from which the con-clusion is drawn that a suitable set of conditionsfor subsequent Kennelly -Heaviside layer studiesis the combination of a distance between senderand receiver of about 25 km, with a transmissionfrequency of about 1500 kilocycles." See alsoMunro and Green, under " Propagation of Waves."

APPLICATION OF COMPENSATED VALVE -VOLTMETERTO MEASUREMENTS OF GLASS ELECTRODEPOTENTIALS [without Use of " Electro-meter " Valves].-A. S. McFarlane. (Journ.Sclera. Inst., July, 1933, Vol. io, No. 7,pp. 208-212.)

The writer's voltmeter (Abstracts, 1932, pp. 650-651, and July, p. 401, r -h col.) does not employ" electrometer " valves but ordinary triodes. Thevoltage sensitivity is considerably greater thanthat of electrometer valve instruments, and theinstrument is more sensitive, and has a morestable zero, than other instruments using ordinaryvalves. Its maximum grid -input impedance is5 000 megohms.

With regard to electrometer valves the writerconsiders that the transference of the grid to theopposite side of the filament from the anode canhave little effect on Metcalf and Thompson's " sixfactors " (1931, p. 98, r -h col.) and that the highgrid impedance is largely to be attributed to thehigh degree of evacuation and to the fact that anodevoltages are used which are below the criticalionisation potential of the residual gas. It issignificant that in the Metcalf and Thompson valve(grid impedance 1016 ohms) no such alterationis made.

THE DIFFERENT OSCILLATION REGIMES OF APARALLELEPIPED OF QUARTZ [and theCause of Rotation of a Pivoted Crystal].-A. de Gramont. (Comptes Rendus, loth July,1933, Vol. 197, No. 2, pp. 101-103.)

The writer considers a parallelepiped withsurfaces parallel to the electrical, optical andmechanical axes, the dimensions along these axesbeing e, o and m respectively ; the crystal is setin vibration along the mechanical axis, f being thefrequency obtained. The coefficient f x m isdenoted by K,.

October, 1933

The experimental curve shows that so long as mis large compared with o, K,,, is practically constant,so that f is inversely proportional to the length m :tests on a large number of crystals give an averagevalue K. = 2 706, near to the value given byHund. If, other things being constant, m isgradually decreased, Km begins to decrease whenm/o is about 2 ; when m/o = 0.52, the limit ofoscillation along the mechanical axis is reached,Km being then just above i 600. A further diminu-tion of m produces at first a quite unstable regime ;then, for m/o = 0.36. a new regime begins withthe oscillations now along the electrical axis.The coefficient K, = f x e tends, as m decreases,towards a limiting value (round 2 800) at m/o= 0.14. " We have proceeded no further : thecrystal, having then hardly r mm along m, isdifficult to maintain vertical between the plates."

For such a crystal the velocity of propagationalong e, calculated from the frequency, was5 504 m/s, agreeing with Cady's anticipated valueof 5 500 m/s. Variations of thickness along ehave little effect. In crystals of dimensionscorresponding to the zone of discontinuity (i.e.the unstable regime for values of m/o between0.52 and 0.36) the simultaneous oscillations alongtwo orthogonal axes produce considerable mechan-ical tensions ; their study is a delicate mattersince they frequently shatter the crystal. Theresulting electric charges are responsible for themovements of rotation described elsewhere(September Abstracts, p. 518, r -h column). Thefact that these rotations are not of aerodynamicorigin is proved by exhausting the space aroundthe rotating crystal ; at first the motion slows downowing to the ionising of the gas, luminous feathersappearing ; then, as the vacuum increases, itspeeds up and becomes faster than in open air.In air the electric wind due to the electric chargesmay add its action to that of the latter and thusmask the true cause of the phenomenon.

'CHER EINIGE BEOBACHTUNGEN AN SCHWINGENDENPIEZOQUARZEN UND IHREM SCHALLFELD(Some Observations on Oscillating Piezo-quartz Crystals and their Acoustic Field).-K. Bucks and H. Miiller. (Zeitschr. f.Physik, 1933, Vol. 84, No. 1/2, pp. 75-864

THE CONVERSE PIEZOELECTRIC EFFECT IN MIXEDCRYSTALS ISOMORPHOUS WITH ROCHELLESALT.-S. Bloomenthal. (Physics, May,1933, Vol. 4, No. 5, pp. 172-177.)

To explain the peculiar behaviour of Rochellesalt near 25°C, several experimenters have postu-lated a critical temperature for this dielectricanalogous to the Curie point for ferromagneticmaterials. The writer obtains confirmatory evi-dence by showing that the critical temperatureshifts with the introduction of isomorphous im-purities, just as the Curie point shifts when theferromagnetic material is alloyed with a secondelement.

ELECTRICAL PROPERTIES OF ISOMORPHOUS CRYSTALSOF ROCHELLE SALT AND SODIUM AMMONIUMTARTRATE.-M. Eremejew and B. Kurt-schatow. (Physik. Zeitschr. der Sowjetunion,No. 3, Vol. 3, 1933, pp. 304-320.)

E2

October, 1933

THE LOWER CURIE POINT IN ROCHELLE SALT. -B. Kurtschatow and I. Kurtschatow. (Ibid.,PP. 321-334.)

ON THE DETERMINATION OF SOME OF THE ELASTICCONSTANTS OF ROCHELLE SALT [Piezo-electric Resonators] BY A DYNAMICALMETHOD. -R. M. Davies. (Phil. Mag., July1933, Series 7, Vol. 16, No. 103, pp. 97-124.)

MODES OF VIBRATION OF PIEZOELECTRIC CRYSTALS.-N. H. Williams. (Proc. Inst. Rad. Eng.,July, 1933, Vol. 21, No. 7, pp. 99o-995.)

Author's summary :-(a) A crystal is made tooscillate by subjecting it to the action of an airwave produced at a considerable distance from thecrystal by a jet of air escaping from a small tube.The piezoelectric charge developed on the surfaceof the crystal as a result of the oscillation is mappedout by means of a tuned amplifier, and the vibrationis analysed into a fundamental and many overtones.For crystals that are long in one dimension, theovertones are nearly exact harmonics.

(b) Exciting the crystal electrically by meansof two tubes and especially designed electrodes,any harmonic up to the tenth can be produced.The crystal usually has only one mode of vibrationfor each pair of electrodes.MULTIVIBRATOR CIRCUIT WITH TWO TRIODES, OR

ONE TETRODE, WITH PIEZOELECTRICCRYSTAL REPLACING A CONDENSER.-(Hochttech. u. Elek:akus., July, 1933, Vol. 42,No. i, p. 36: German Pat. No. 571235,pub. 25th Feb., 1933, Telefunken.)

" A frequency is generated for which the crystalpossesses a minimum impedance, so that electrode -and lead -capacities have no influence on the fre-quency."MEASUREMENTS OF THE LOGARITHMIC DECRE-

MENTS OF MAGNETOSTRICTION RESONATORS :SECOND COMMUNICATION. -E. A. Kopilo-witsch. (Physik. Zeitschr. der Sowietunion,No. 6, Vol. 3, 1933, pp. 561-566.)

Continuation of the work dealt with in 1932Abstracts, p. 355, r -h col. The resonators in-vestigated were of nickel, elinvar and a nickel -ironalloy. The frequency employed was about 8o kc/s.Among other points the need of a good fixingmethod, in order to obtain a small decrement, isestablished.

THE MICROMETER FREQUENCY METER. -G. F.Lampkin. (QST, July, 1933, Vol. 17, No. 7,pp. 10-13.)

ROBOT CLOCK ANNOUNCES TIME FROM THE PARISOBSERVATORY. -E. Esclangon Belin.(Science, 26th May, 1933, Vol. 77, No. 2004,Supp. pp. 11-12.)

" In clear, human tones the exact time, correctto a small fraction of a second," will be heard byanyone ringing up a certain telephone number inParis.

TABLES FOR THE CALCULATION OF THE MUTUALINDUCTANCE OF ANY TWO COAXIAL SINGLE -LAYER COILS. -F. W. Grover. (Proc. Inst.Rad. Eng., July, 1933, Vol. 21, No. 7, pp.1039-1049.)


" The tables and formulae here presented enablethe mutual conductance of any two coaxial solenoidswhatever to be calculated from a single formula.Examples make clear that a five -figure accuracymay be attained with concentric coils and evenwith poorly coupled coils the error does not exceeda few parts in ten thousand."

THE FUNDAMENTAL AND HARMONIC FREQUENCIESOF MULTI -LAYER INDUCTANCE COILS. -A. Korn. (Zeitschr. f. angewandte Math. u.Mech., Feb., 1933, Vol. 13, pp. 32-35.)

THE CALCULATION OF IRON -FREE CHOKING COILS[Simplified Method allowing for HeatingEffects]. -R. Edler. (Bull. Assoc. Suisse d.Elec., 31st March, 1933, Vol. 24, pp. 149-157.)

SKIN EFFECT IN A STRATIFIED CYLINDER OFCIRCULAR SECTION. -J. Fischer and M. J. 0.Strutt. (Hochf:tech. u. Elek:akus., June,1933, Vol. 41, No. 6, p. 219.)

The simultaneous and independent researchesof the two writers (March and May Abstracts,pp. 171 and 269) are compared, and for the sakeof those who wish to employ the various formulaethe reason for any apparent differences are explained.

ENGINEERING CALCULATION OF INDUCTANCE ANDREACTANCE [at Low Frequencies]] FOR REC-TANGULAR BAR CONDUCTORS. -0. R. Schurig.(Gen. Elec. Review, May, 1933, Vol. 36, No.5, pp. 228-231.)

FIRST COMPARISONS OF THE NATIONAL STANDARDSOF ELECTRICAL RESISTANCE, CARRIED OUT ATTHE INTERNATIONAL BUREAU OF WEIGHTSAND MEASUREMENTS. -A. PE rard and M.Romanowski. (Comptes Rendus, ist May,1933, Vol. 196, No. 18, pp. 1288-1291.)

NOTE ON AN IMPROVED FORM OF AN INDUCTIONCOIL BUZZER [as A.C. Source for ResistanceMeasurements using Special A.C. Galvano-meter as Null Indicator]. -H. Mukherjee.(Physik. Ber., 15th June, 1933, Vol. 14,No. 12, p. 92o.)

SUBSIDIARY APPARATUS AND MATERIALSDIE ROCKWIRKUNG METALLISCHER SPULENKAPSELN

AUF VERLUSTE, INDUKTIVITAT UND AUSSEN-FELD EINER SPULE (The Influence of aMetallic Cover on the Losses, Inductanceand External Field of a Coil). -H. Kaden.(E.N.T., July, 1933, Vol. io, No. 7, pp. 277-284.)

Further development of the work dealt with inAbstracts, 1932, p. 654, 1-h col. Methods of calcula-tion depending on treating the cover as the secondarywinding of a transformer (May, p. 281, Hillers) canonly give qualitative results, since they neglectthe influence of current and flux skin effect, soimportant at high frequencies. For obtaining exactresults applicable at all frequencies, even thehighest, it is necessary to integrate the Maxwelliandifferential equations of the electromagnetic fieldfor the whole space. In his previous paper the


writer showed that the important factors in thescreening effect of a metallic cover, besides thefrequency and the constants of the material, weresimply the wall -thickness and a single furtherquantity defining the external dimensions, theactual shape of the cover playing a subordinaterole. In his present paper he therefore simplifiesthe problem by replacing the practical form ofcover by a hollow sphere of the same wall -thicknesswith a diameter which is the geometrical mean ofthe cover measurements along the three co-ordinates. A further simplification is given by theassumption of quasi -stationary conditions : " thedimensions of the coil with its cover, even at highfrequencies, is small compared with the wavelengthof the field."

It is found that the screening effect of the coveragainst the field of the coil inside it is exactly equalto its screening of the space inside against anexternal field. Therefore the mutual screening oftwo coils, each with a cover, is equal to the productof the screenings of the two covers taken separately.An important point is the calculation of the effectof the cover on the inductance and phase angleof its coil. This is possible with the help of theformulae derived, which cover all the variablefactors -frequency, constants of the material,dimensions of cover and of coil. They are validat " all " frequencies. Among other results it isshown that the losses in the cover decrease as thefourth power of the reciprocal cover -diameter. Anumerical example is given of a cylindrical coilwith a cylindrical manganin cover ; the volume ofthe coil is 34 cma, that of the equivalent sphericalshell 270 cma. By the application of formula 9the ratio of the inductance with cover to thatwithout cover is found to be o.88 ; from formula 8the phase angle due to the cover is 7 X io-3.The corresponding measured values are 0.87 and6.3 x 10-3, which show good agreement.

THE ESSENTIALS OF WIRELESS COMPONENTS.PART I : HIGH -FREQUENCY CHOKES. -M. G.Scroggie. (World -Radio, II Aug., 1933,Vol. 17, No. 420, pp. 163 and 165.)

A NEW TYPE OF IMPEDANCE UNIT [Compact andCheap Filter Unit from Quarter -WavelengthLine of Twisted Pairs of Conductors Encasedin Shielding, Coiled Up and Inserted inCan]. -F. S. Allen. (Summary only inProc. Inst. Rad. Eng., August, 1933, Vol. 21,No. 8, p. 1074.)

THE CONDUCTIVITY OF COMPRESSED MERCURICSULPHATE [with Zinc and Platinum Elec-trodes : Rectifying Action]. -W. Scislowski.(Physik. Ber., 15 June, 1933, Vol. 14,No. 12, p. 933.)

THE IGNITRON-A NEW CONTROLLED RECTIFIER[Arc started by Spark between Carborundum(e.g.) and Mercury, during Conducting Half -Cycle : Output varied by Phase -Shift ofControl Voltage]. -D. D. Knowles : West-inghouse Company. (Electronics, June, 1933,Vol. 6, No. 6, pp. 164-166.)

577 October, 1933

PHYSICAL PROCESSES IN LOW-PRESSURE MERCURY -VAPOUR ARCS [Mercury -Vapour Rectifiers].-E. Liibcke. (Zeitschr. f. tech. Phys.,No. 8, Vol. 14, 1933, pp. 297-304.)

THE USE OF THYRATRONS AS RELAYS IN HEATERSETS [for Crystal Incubators : the Use ofAnode Voltage Supply for Grid Control,after Rectification, thus eliminating D.C.Control Battery]. -(Marconi Review, May/June, 1933, No. 42, p. 27.) .

A LIFE -TEST POWER SUPPLY UTILISING THYRATRONRECTIFIERS. -H. W. Lord. (Proc. Inst.Rad. Eng., August, 1933, Vol. 21, No. 8,pp. 1007-1102.)

DER STROMRICHTER BEI KAPAZITIVER LAST (TheCurrent Rectifier with Capacitative Load).-H. Klemperer and K. Strobl. (Archiv f.Elektrot., ist July, 1933, Vol. 27, No. 7,pp. 485-488.)

UREA DIE ZIJNIMEDINGUNG VON NIEDERVOLT-BOGEN UND GITTERGESTEUERTEN NIEDER-VOLTBOGEN. I. (On the Sparking Conditionof Low Voltage Arcs and Grid -ControlledLow Voltage Arcs [Thyratrons]. I ).-F.Klaiber. (Physik. Zeitschr., ist June, 1933,Vol. 34, No. II, PP. 441-447.)

VOLTAGE STABILISATION : THE MECHANISM OFVOLTAGE REGULATION BY MEANS OF GASDISCHARGE TUBES. -F. M. Colebrook.(World -Radio, 23rd and 3oth June, 1933,Vol. 16, Nos. 413 and 414, pp. 829 and862-863.)

THE PROVISION OF SEVERAL DIFFERENT ANODEVOLTAGES BY THE " STABILIVOLT " GLOW -DISCHARGE TUBE. -(Hoch f: u. Elek:akus.,April, 1933, Vol. 41, No. 4, p. 150: GermanPat. No. 565125, pub. 26th Nov., 1932,Koros.) See April Abstracts, p. 226, 1-hcolumn.

NOTE ON THERMAL INSTABILITY IN CYLINDRICALCONDENSERS DUE TO DIELECTRIC LOSSES[Calculation of Electric Intensity givingInstability]. -B. L. Goodlet. (Proc. Camb.Phil. Soc., 3oth July, 1933, Vol. 29, Part 3,PP. 401-404.)

JEER EINE UNIPOLARE FORM DER GLEITKORONA(A Unipolar Form of " Gliding Corona "[a Kind of Glow Discharge : at Voltagesdown to 200 V : and a Criticism of Gemant's" Electrophotography " of the Porous Struc-ture of Insulating Materials]). -A. Coehnand W. Ziegler. (Zeitschr. f. tech. Phys.,No. 6, Vol. 14, 1933, pp. 246-2494 ForGemant's paper see 1931 Abstracts, p. 396,1-h column.

DIRECT PROOF OF THE SPACE CHARGE IN OILS. -A. Gemant. (E.T.Z., 18th May, 1933,Vol. 54, No. 20, pp. 468-469.)

1

October, 1933

IS THE BREAKDOWN OF INSULATING LIQUIDS AHEAT PROCESS ?-F. Koppelmann. (Archief. Elektrot., 1st June, 1933, Vol. 27, No. 6,PP 448-453)

Further development of the work referred to inAbstracts, 1931, p. III, and 1932, p. 47, r -hcolumns.

DIE ELEKTRISCHE FESTIGKEIT FLUSSIGER ISOLIER-STOFFE BEI HOHEN FREQUENZEN (TheElectrical Strength of Liquid InsulatingMaterials [Oil and Xylol] at High Frequencies[and the Thermal Breakdown Effect]). -W. Schlegelmilch. (Physik. Zeitschr.,ist July, 1933, Vol. 34, No. 13, pp. 497-507.)

THE THEORY OF GLOW DISCHARGE VALVE TUBES[for Protecting Devices : Possibility ofUsing Greater Electrode Gaps]. -L. Heer.(Archie f. Elektrot., i5th March, 1933, Vol. 27,pp. 196-205.)

" WAX -OPERATED SWITCHES " AND METALLISEDWAX STRIPS AS FUSES FOR INSTRUMENTS,FRACTIONAL-H.P. MOTORS, ETC. -A. G.Bullen. (Sci. Abstracts, Sec.B, May, 1933,Vol. 36, No. 425, pp. 281-282.)

THE PERFORMANCE OF RELAYS [as regardsVariability in Time of Operation : and aReed -Armature Relay with Lag not exceed-ing o.i Millisecond]. -G. A. Tomlinson.(Journ. Scient. Instr., July, 1933, Vol.No. 7, pp. 204-208.)

A SIMPLIFIED FREQUENCY -DIVIDING CIRCUIT [usingOne Pentode Valve : for Frequency -Measuring Equipments]. -V. J. Andrew.(Proc. Inst. Rad. Eng., July, 1933, Vol. 21,No. 7, pp. 982-983.)

The control grid and screen grid are used togetheras grid and plate of a normal triode, and are con-nected to a piezo-oscillator circuit. The suppressor(collector) grid and the plate are used as grid andplate of a second triode forming a 1.f., self-excitedoscillator. " Since the electron current whichoperates the outer triode has passed through theinner triode, it has been modulated by the frequencyof the piezoelectric oscillator. This electroncoupling is sufficient to make the 1.f. oscillator' lock -in ' at a sub -harmonic of the crystal fre-quency." The arrangement has been found towork excellently when division by a small integer(say not over 4) is desired.

A NEW METHOD OF FREQUENCY ANALYSIS APPLIC-ABLE TO ACOUSTIC AND RADIO FREQUENCIES.-Fehr and Kreielsheimer. (See abstractunder " Properties of Circuits.")

ON ELECTRON OPTICS [Survey of the Magnetic andElectrostatic Focusing of High -VacuumElectron Beams]. -V. K. Zworykin. (Journ.Franklin Inst., May, 1933, Vol. 215, No. 5,PP. 535-555.)

SIMULTANEOUS TRACES WITH CATHODE-RAYOSCILLOGRAPH [with Single Ray : the Useof a Synchronised Rotary Switch]. -C. Bradner Brown. (Electronics, June, 1933,Vol. 6, No. 6, p. 170.)


USE OF CATHODE-RAY OSCILLOGRAPH FOR OBTAIN-ING FREQUENCY -RESPONSE CouvEs.-Davis.(See under " Acoustics and Audio -frequencies.")

SPECIAL LENS AND OSCILLOGRAM CAMERA FORCATHODE-RAY TUBES .-M. von Ardenne.(Ilochflech. u. Elek:akus., July, 1933, Vol. 42,No. 1, pp. 37-39.)

Designed with a view to a great reduction of thecost of the auxiliary apparatus of a c.r. oscillograph.The lens gives just the right definition for thepurpose ; by dispensing with unnecessarily perfectdefinition a reasonable price is possible in spiteof the high light admission and other desirablequalities. The usual expensive drum -recordingequipment, with its attendant dark room, isreplaced by a spring -driven slider device, in whichthe speed with which the plate or film is projectedacross the image slit is regulated (up to a maximumof 4-5 m/s) by the tension of the spring. Thisspeed attains its maximum before the slit isreached ; the brake is applied after it is passed,so that during exposure the speed is constant.It is indicated on the record by the trace given bya small lamp lit off the a.c. network.

ELEKTRONENMIKROSCOPISCHE BEOBACHTUNGEN ANDER BARIUM-AUFDAMPFKATHODE (Observa-tions of a Barium Cathode, sputtered onNickel, with an Electron Microscope). -E. Brfiche and H. Johanson. (Zeitschr.f. Physik, 1933, Vol. 84, No. 1/2, pp. 56-58.)

THE VARIATION OF THE INTENSITY OFFLUORESCENCEWITH THE WAVELENGTH OF THE EXCITINGRADIATION [and the Effect of the Solvent]. -W. Fabrikant. (Physik. Zeitschr. derSowietunion, No. 6, Vol. 3, 1933, pp. 567-573.)

SYNCHRONISATION OF MOTORS: THE PREVENTIONOF HUNTING [using an Elastically -ConnectedFlywheel,a Photoelectric Cell and Thyratrons].-Lichine : Beghin. (Comptes Rendus, 17thJuly, 1933, Vol. 197, No. 3, pp. 223-224.)

KURVEN ZUR BERECHNUNG DER PERMEABILITATUND DER VERLUSTE IM BLECH (Curves forthe Determination of the Permeability of,and Losses in, Iron Sheet). -W. Arkadiew.(E.N.T., May, 1933, Vol. 10, No. 5, pp. 220-222.) See also June Abstracts, p. 34o,1-h col.

MAGNETIC PROPERTIES AND CHEMICAL LINKAGEIN ALLOYS. -J. Dorfman. (Physik. Zeitschr.der Sowietunion, No. 4, Vol. 3, 1933, pp.399-417: in German.)

THE MAGNETIC PROPERTIES OF ELECTROLYTICIRON [in Very Thin Layers]. -K. W. Gri-gorow. (Ibid., pp. 418-420: in German.)

A COMPACT [Oil -Cooled] ELECTROMAGNET FORGENERAL PURPOSES [giving 32 500 Gauss].-L. F. Bates and B. J. Lloyd -Evans.(Proc. Phys. Soc., 1st May, 1933, Vol. 45,Part 3, No. 248, PP. 425-433.)


A PERMANENT OXIDE MAGNET AND ITS CHAR-ACTERISTICS [Specially High Coercivity :Poles located at Any Point : SpecificGravity Very Low : etc.] -Y. Kato and T.Takei. (Journ. I.E.E. Japan, May, 1933,Vol. 53 [No. 5], No. 538: English summaryp. 37, Japanese paper pp. 408-412.)

579

THE PROBLEM OF UNIPOLAR INDUCTION AND THEELECTRICAL SURFACE CHARGES ON ROTATINGMAGNETS. -L. Slepian. (Physik. Zeitschr. derSowietunion, No. 5, Vol. 3, 1933, pp. 469-486 : in German.)

THE " FER-X " [Iron -Cored] H.F. Transformers. -G. Budich Company. (Hochf:tech. u. Elek:akus., July, 1933, Vol. 42, No. I, p. 39.)

A MAGNETIC FLUX RECORDER. -O. von Auwers.(Zeitschr. f. tech. Phys., No. 8, Vol. 14,1933, pp. 316-319.)

COMPOSITION OF GRIDS FOR POSITIVE PLATES OFSTORAGE BATTERIES AS A FACTOR INFLUENC-ING THE SULPHATION OF NEGATIVE PLATES[and the Effect of the Presence of Antimony].-Vinal, Craig and Snyder. (Bur. of Stds.Journ. of Res., June, 1933. Vol. Jo, No. 6,pp. 795-808.)

A RESISTANCE -COUPLED [Two -Stage] AMPLIFIERFOR MEASURING IONISATION CURRENTS [downto 10-13 Ampere]. -L. F. Curtiss. (Bur.of Stds. Journ. of Res., May, 1933, Vol. Ico,No. 5, pp. 583-589.)

ELECTRO-MECHANICAL DETERMINANT MACHINE. -A. Weygandt. (Sci. Abstracts, Sec.A, June,1933, Vol. 36, No. 426, pp. 544-545.)

A CALCULATING RULE OF HIGHLY ELASTIC RUBBER.(Electronics, May, 1933, Vol. 6, No. 5, p.143.)

A NEW COUPLING FOR ELECTRIC DRIVE, WITHOVERLOAD RELEASE. -AEG. (E.T.Z., 15thJune, 1933, Vol. 54, No. 24, p. 584.)

ON THE DRIVING MECHANISM OF RECORDINGAPPARATUS. -J. Baltzer. (E.T.Z., 20thApril, 1933, Vol. 54, No. 16, pp. 377-379.)

SPECIAL FLEXIBLE SHAFT FOR REMOTE CONTROL. -(Electronics, June, 1933, Vol. 6, No. 6, p. 174.)

STATIONS, DESIGN AND OPERATIONTHE LUCERNE FIGHT OVER THE RADIO WAVES. -

(Radio, B., F. fur Alle, August, 1933, pp.347-351.)

THE PROSPECTS OF LONG -WAVE BROADCASTINGIN SOUTH-EASTERN AUSTRALIA.-MBDIO andGreen. (See abstract under " Propagationof Waves.")

A METHOD OF CALCULATION OF FIELD STRENGTHSIN HIGH -FREQUENCY [Short -Wave] RADIOTRANSMISSION.-Namba and Tsukada. (Seeunder " Propagation of Waves.")

October, 1933

LAND -LINE CIRCUITS FOR BROADCASTING : EQUAL-ISATION OF LAND -LINE CIRCUITS. (World -Radio, 4th and Iith Aug., 1933, Vol. 17,Nos. 419 and 42o, pp. 131 & 133 and167 & 165.)

GENERAL PHYSICAL ARTICLESTHE COMPLEXITY OF THE PROTON AND THE MASS

OF THE NEUTRON [Proton composed ofNeutron and Positive Electron]. -I. Curieand F. Joliot. (Comptes Rendus, i7th July,1933, Vol. 197, No. 3, pp. 237-238.)

ELECTROSTATIC DEVIATION AND SPECIFIC CHARGEOF THE POSITIVE ELECTRON. -J. Thibaud.(Comptes Rendus, 7th Aug., 1933, Vol. 197,No. 6, PP. 447-448.)

QUANTUM THEORY OF INELASTIC COLLISIONS OFELECTRONS. -L. Goldstein. (Ann. de Phy-sique, April, 1933, Vol. 19, pp. 305-420.)

QUANTUM THEORY OF DISPERSION (CONTINUED).PARTS VI AND VII. -G. Breit. (Reviews ofMod. Phys., April, 1933, Vol. 5, No. 2, pp.91-140.)

RITZ'S ELECTRODYNAMIC THEORY. -W. Hovgaard.(Sci. Abstracts, Sec.A, March, 1933, Vol. 36,No. 423, P. 304.)

NOTES ON SOME ELECTRONIC PROPERTIES OF CON-DUCTORS AND INSULATORS. -R. H. Fowler.(Proc. Roy. Soc., 3rd July, 1933, Vol. 141,No. A 843, pp. 56-71.)

" The paper is a critical survey of certain currentviews on the nature of crystalline insulators and theirelectronic properties."

AN ELEMENTARY THEORY OF ELECTRONIC SEMI-CONDUCTORS, AND SOME OF THEIR POSSIBLEPROPERTIES. -R. H. Fowler. (Proc. Roy.Soc., ist June, 1933, Vol. 140, No. A 842,pp. 505-522.)

THE ELECTRONIC CONDUCTIVITY OF THE COPPEROXIDES. II. -M. Le Blanc, H. Sachse andH. SchOpel. (Ann. der Physik, 1933, Series5, Vol. 17, No. 3, PP. 334-344)

PHASE BOUNDARY POTENTIALS OF ADSORBED FILMSON METALS. PARTS I, II AND III. -Whalleyand Rideal : Jacobs and Whalley. (Proc.Roy. Soc., 1st June, 1933, Vol. 140, No.A 842, PP 484-504.)

SUPERCONDUCTIVITY OF TIN AT RADIO FREQUEN-CIES. -F. B. Silsbee, R. B. Scott, F. G.Brickwedde and J. W. Cook. (Phys. Review,15th June, 1933, Series 2, Vol. 43, pp.1050-1051 : abstract only.)

ON THE INFLUENCE OF A TRANSVERSE MAGNETICFIELD UPON THE RESISTANCE OF. LIQUIDMETALS. -I. Fakidow and I. Kikoin. (Phy-sik. Zeitschr. der Sowietunion, No. 4, Vol. 3,1933, pp. 381-392 : in English.)

CORONA AND SPARK DISCHARGE IN GASES. -J. D.Stephenson. (Journ. I.E.E., July, 1933,Vol. 73, No. 439, pp. 69-82.)

October, r933

THE INFLUENCE OF CIRCUIT CONSTANTS ANDFREQUENCY ON CHARACTERISTICS OF GASDISCHARGES AT LOW PRESSURES (PART VI).-Y. Asami. (Journ. I.E.E. Japan, June,1933, Vol. 53 [No. 6], No. 539: Englishsummary pp. 46-48.)

MEASUREMENT OF THE TOWNSEND COEFFICIENTSFOR IONISATION BY COLLISION. -F. H.Sanders. (Phys. Review, 15th July, 1933,Series 2, Vol. 44, No. 2, p. 129: abstract only.)

RECOMBINATION RADIATION IN THE CAESIUMPOSITIVE COLUMN. -F. L. Mohler. (Bur.of Stds. Journ. of Res., June, 1933, Vol. ro,No. 6, pp. 771-780.)

MISCELLANEOUSELLIPSOIDAL FUNCTIONS AND THEIR APPLICATION

TO SOME WAVE PROBLEMS. -E. T. Hanson.(Phil. Trans. Roy. Soc. Lond., A, Vol. 232,No. 713, pp. 223-283.)

Results of importance in the theory of ellipsoidalfunctions are obtained in this paper, including theexpansion of an arbitrary function in a series ofellipsoidal functions and an expression suitable torepresent a divergent wave. Among the applica-tions considered are the scattering of plane wavesincident normally on a thin circular disc and thetransmission of waves through a circular aperture.

EXTREME WERTE VON VEKTOREN EINER ORTS-KURVE (Limiting Values of Vectors of a LocusCurve [Application to the Simple Calculationof Maxima and Minima, in place of Differ-ential Calculus]) .-H. Kind. (E.N.T., July,1933, Vol. Jo, No. 7, pp. 285-286.)

As an example of the method, which postpones tothe last, or avoids completely, the involved treat-ment of real and imaginary quantities, the writertakes the simplified equivalent diagram of a trans-former and finds the frequency at which the im-pedance vector has the smallest phase -displacement,and how this frequency varies with the load.MECHANICAL MODELS FOR ELECTRICAL PROCESSES. -

R. Weller. (Radio, B., F. far Alle, May, I933,pp. 199-211.)

THE WORK OF THE INSTITUTE FOR THE APPLICA-TIONS OF CALCULATION [including Problemsconnected with Oscillation Transformers,Thermionic Currents, etc.].-(L'Elettrotec.,5th May, 1933, Vol. 20, No. 13, pp. 285-293.)

ROYAL SOCIETY CONVERSAZIONE-RADIO ANDOTHER EXHIBITS. -(Electrician, 26th May,1933, Vol. 110, No. 2869, pp. 671-672.)

PUBLICITY AND THE PHYSICIST. -R. D. Potter.(Review Sclera. Instr., May, 1933, Vol. 4,No. 5, pp. 261-263.)

SURVEY OF THE MOST IMPORTANT DEVELOPMENTSIN RADIO ENGINEERING AND ACOUSTICS INI Q32.-Elektrotechnischer Verein. (E.T.Z.,18th May, 1933, Vol. 54, No. 20, pp. 480-483.)

SOME ELECTRICAL METHODS OF REMOTE INDICA-TION. -C. Midworth and G. F. Tagg.(Journ. I.E.E., July, 1933, Vol. 73, No. 439,PP. 33-48.)

58o THE WIRELESS ENGINEER &

THE RADIO CORPORATION OF AMERICA AND ITSRAMIFICATIONS. -H. Winkler. (E.T.Z.,15th June, 1933, Vol. 54, No. 24, pp. 585-587.)

ENGINEERING ORGANISATION. -R. F. Shea. (Rad.Engineering, June, 1933, Vol. 13, No. 6, pp.18-19 and 21.)

TRENDS IN RADIO DESIGN AND MANUFACTURING[Binaural Transmission : Hill -and -DaleGramophone Recording : Reflex Circuits :etc.]. -(Electronics, June, 1933, Vol. 6, No.6, pp. 151-152 and 161.)

" The future may see radios sold in double units,the tuning mechanism housed in a small . . .

cabinet, an external loud speaker with sufficientbaffle to provide good bass reception." The recentdemonstration of binaural transmission has shownthat this " was not an idle dream." " Those who longfor high-class home entertainment will find furtherencouragement in the possibility that the long-playing vertically cut records of Dr. Frederick(Bell Laboratories) may be available to the public "[hill -and -dale recording, ro 000 c/s or more withwide range of volume : Abstracts, 1932, p. 'or ;May, p. 275 ; June, p. 33z ; all r -h cols.]. " Thelong -dead reflex is being revamped [see Stinchfieldand Schade, under ' Reception '] . . . Togetherwith the electron -coupled oscillator, the reflexedtube will permit a 7 -tube set to be compressed into4 tubes. Even the power tube may be worked twicein some sets, once at i.f ." " First announcement ofthe Osram Catkin all -metal receiving tubes inEngland must be taken seriously in this country. . . "" There is a strong possibility that the rapid in-crease in tube types may cease as a result of anotherEnglish tube development which appears to simplifyappreciably tube construction. . . ."

WHAT THE INDUSTRY BRINGS TO THE BERLINRADIO EXHIBITION. -(Radio, B., F. far Alle,August, 1933, pp. 338-341.)

SHORT-WAVE WIRELESS COMMUNICATION. -A. W.Ladner and C. R. Stoner. (Book Review inProc. Inst. Rad. Eng., July, 1933, Vol. 21,No. 7, p. 1054.)

DIRECT DETERMINATION OF THE MEAN PRESSUREIN HEAT ENGINES [Varying Curvature ofThick Metallic Membrane alters Divergenceof Light Beam reflected on to Photocell-Microammeter Combination]. -A. Labarthe.(Comptes Rendus, 7th Aug., 1933, Vol. 197,No. 6, pp. 440-442.)

THE USE OF PIEZOELECTRIC CRYSTALS IN RESONANTOSCILLATION FOR THE MEASUREMENT OFPRESSURES [New Piezoelectric Dynamo-meter]. -A. Guerbilsky. (Comptes Rendus,31st July, 1933, Vol. 197, No. 5, pp. 399-401.)

Ordinary piezoelectric dynamometers requireextraordinarily careful insulation when constantor slowly varying pressures are to be measured. Byusing the crystal in a state of resonant oscillation,in an absorption circuit similar to that of the writer'spiezoelectric microphone (Sept. Abstracts, p. 508,r -h col.) a very practical arrangement is obtained,for pressures of several kilogrammes or some tens ofgrammes.

1

of


AN IMPROVED PHOTOELECTRIC RECORDER. -C. W.La Pierre. (Gen. Elec. Review, June, 1933,Vol. 36, No. 6, pp. 271-274.) For previouspapers see 1932 Abstracts, p. 539, r -hcolumn.

THE POSOPHOTOMETER [Photographic ExposureMeter using Photoelectric Cell].-Etablisse-ments Filmograph. (Rev. Gen. de29th April, 1933, Vol. 33, No. 17, p. 538.)In an article on the French Physical Society'sExhibition.

PHOTOCELL CONTROL AGAINST PAPER BREAKAGEIN NEWSPAPER PRINTING [with Precautionsagainst Variable Factors]. -(Electronics,April, 1933, Vol. 6, No. 4, p. Io2.) See alsoMay Abstracts, p. 288, r -h column.

PHOTOELECTRIC CELLS IN PHOTOMETRY. -J. W. T.Walsh. (Illuminating Engineer, March,1933, Vol. 26, pp. 64-71.)

PHOTOELECTRIC CELLS IN INVESTIGATIONS OF THERANGE OF LIGHTHOUSES WITH REVOLVINGBEAMS. -A. Blondel. (Rev. Gen. de l'Elec.,loth June, 1933, Vol. 33, No. 23, pp. 747-758.)

MAKING MEASUREMENTS WITH THE PHOTOCELL. -R. C. Walker. (Wireless World, 23rd June,1933, Vol. 32, PP. 445-446.)

This article outlines the circuits used in colori-meters, optical density meters and similar devices,and describes the Electrometer Triode.

PHOTOELECTRIC CONTROL OF WELDING, RIVET,AND VALVE -STEM HEATERS : OF AMOUNTOF CHLORINE INTRODUCED INTO WATERSUPPLY. -(Electronics, April, 1933, Vol. 6,No. 4, p. 103.)

AN INTEGRATING [Photoelectric] PHOTOMETER FORX-RAY CRYSTAL ANALYSIS. -B. WheelerRobinson. (Journ. Scient. Instr., Aug.,1933, Vol. 1o, No. 8, pp. 233-242.)

LAUNDRY CONVEYOR OPERATED BY PHOTOCELL.-[Electronics, June, 1933, Vol. 6, No. 6, p.162.)

COLOUR MATCHING IN AUTOMOBILE PLANT [byPhotocells : for Parts finished in DifferentPlaces or Other Factories].-(Ibid., p. 162.)

PHOTOCELL ENSURES PERFECT JIG-SAWS.-(Ibid.,p. 163.)

REPORT ON THE TESTS IN THE PUNKWA CAVES ANDTUNNELS [Charted by Radio Field -StrengthMeasurements]. -Fritsch. (See under " Pro-pagation of Waves.")

THE APPLICATION OF RADIO TECHNIQUE TO THETESTING OF MATERIALS : THE MEASURE-MENT OF THE INTERNAL FRICTION OF METALS.-L. Polotowsky. (Physik. Zeitschr. derSowietunion, No. 6, Vol. 3, 1933, pp.555-560.)

October, '933

THE USE OF A CONDENSER -RESISTANCE COM-BINATION AS TIMING CIRCUIT FOR GRID -CONTROLLED RECTIFIERS USED IN WELDING.-(Electronics, April, 1933, Vol. 6, No. 4,p. 505.)

THE RUE 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.)

VALVE METHOD OF MEASURING THE MOISTURECONTENT OF WHEAT. -H. E. Hartig andB. Sullivan. (Electronics, April, 1933, Vol.6, No. 4, p. Ito : summary only.)

STUDY OF THE WING BEATS OF A BIRD OR INSECTBY THE USE OF A HOT WIRE AND CATHODE-RAY OSCILLOGRAPH.-A. Magnan and C.Magnan. (Comptes Rendus, 8th May, 1933,Vol. 196, No. 19, pp. 1369-1372.)

BRUSH DISCHARGE IN AIR AFFECTS ENGINE PER-FORMANCE. -(Bur. of Stds. Tech. News Bull.,April, 1933, No. 192, p. 4o.) See also JulyAbstracts, p. 405, r -h column.

MOSQUITOES ATTRACTED BY ULTRA -VIOLET LAMPSAND THEN TRAPPED OR ELECTROCUTED. -(World -Radio, 19th May, 1933, Vol. 16, No.408, p. 674.) See also acoustic method,February Abstracts, p. 117, r -h column.

INVESTIGATIONS OF THE TRANSPARENCY TO LIGHTOF FOG.-Bom, Dziobek and Wolff.(Zeitschr. f. tech. Phys., No. 7, Vol. 14,1933,pp. 289-293.)

INFRA -RED FOG -EYE PICKS UP SHIPS BELOWHORIZON [Refraction of Infra -Red Rays :Ships detected 6 Miles beyond Horizon :Naval Aeroplane detected at 44 Miles andCourse easily followed]. -P. H. Macneil.(Electronics, June, 1933, Vol. 6, No. 6, pp.162-163.)

SOME REMARKS ON THE QUESTION OF THE RESIST-ANCE OF THE BODY TO HIGH -FREQUENCYCURRENTS. -N. N. Malov. (Hochf:tech. u.Elek:akus., April, 1933, Vol. 45, No. 4,pp. 138-141.)

Further results in the work dealt with in 1932Abstracts, pp. 540-541. The effect of certainillnesses and of radioactive baths, in changing the" normal " resistance of the human body, isreferred to.

A METHOD FOR THE INVESTIGATION OF THE IM-PEDANCE OF THE HUMAN BODY TO ANALTERNATING CURRENT. -M. A. B. Brazier.(Journ. I.E.E., August, 1933, Vol. 73, No.440, PP. 204-209.)

THE ROTATION OF A QUARTZ CRYSTAL. -de Gra-mont. (See abstract under " Measurementsand Standards.")


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 each.

COMBINED TRANSMITTER AND RECEIVERConvention date (U.S.A.), 4th March, 1931.

No. 388213A short-wave set is arranged to generate strong

oscillations which are used (a) for transmission and(b) also to assist reception by super -regenerativeaction. The valves V, VI are arranged in push-pullacross the input circuit L, C, which is back -coupledto a similar output circuit Li, Ct. A leaky gridsystem GL is adjusted to " quench " the oscillationsat supersonic frequency. A choke L2 in the platesupply circuit forms one arm of a Wheatstonebridge. Modulated signals are applied across adiagonal of the bridge from a microphone M,whilst received signals are taken off from the otherdiagonal to an amplifier V2 and headphones.

No. 388213

photographic film is moved intermittently, andcertain of the mirrors on the scanning -drum areblackened so that no reflection occurs during theactual movement of the film.

Patent issued to The Gramophone Co., Ltd., andG. E. Condliffe.

VISUAL TUNING INDICATORSApplication date, 15th December, 1931. No. 388516

A visual indicator, particularly adapted for setsfurnished with automatic volume control, consistsof a glow -discharge lamp inserted across the anodeand cathode of the detector valve, and connectedin series with a resistance of, say 20,000 ohms to theH.T. supply. The presence of an incoming signal

The grid condenser may be disconnected duringperiods of transmission by means of a switch S.

Patent issued to Marconi's Wireless TelegraphCo., Ltd.

TELEVISION SYSTEMSApplication date, 18th August, 1931. No. 388404

The picture, preferably a kinematographic film,is scanned by a drum fitted with tangential mirrors,which throw a series of images over several, saythree, scanning -apertures, all co-operating with asingle photo -electric cell. Successive images aredisplaced relatively, in a direction at right -angles totheir traverse, so that each aperture serves in turnto scan different portions of the picture. The

reduces the voltage drop across the anode resistanceuntil the lamp " strikes." The cathode of thelamp is a relatively long rod so that variations inthe discharge current are indicated by the relativelength of glow. Preferably the lamp has anauxiliary anode which is connected to the fullH.T. voltage through a high resistance so as tomaintain a " priming " discharge.

Patent issued to The General Electric Co., Ltd.,N. R. Bligh and N. L. Harris.

EXPERIMENTAL WIRELESS 58.3

SCREEN -GRID AMPLIFIERSConvention date (Germany), 29th April, 1931.

No. 388600In a resistance -coupled screen -grid amplifier,

reaction is obtained by means of a coil inserted inthe screen -grid lead and back -coupled to the inputcoil supplying the control grid. The degree ofreaction is controlled either by varying the couplingbetween the coils or by varying the biasing voltageapplied to the screening -grid. The arrangement isparticularly suitable for circuits using valves ofthe multi -stage type.

Patent issued to S. Loewe and P. Kapteyn.

WAVE -CHANGE SWITCHINGConvention date (Denmark), 3rd September, 1931.

No. 388632A rotatable drum -switch is employed for wave-

band changing and is provided with surface contactswhich serve the purpose of connecting the varioustuning -coils either in series, parallel, or series -parallel as required. All the windings are incircuit at each setting, i.e., for long, medium, orshort-wave working.

Patent issued to F. M. Frederiksen.

GAS -FILLED AMPLIFIERSConvention date (Germany), 22nd May, 193o.

No. 38873oRelates to gas -filled amplifiers designed to use

from 15 to zo volts on the main anode with a heatedcathode, or a higher anode voltage with a coldcathode. According to the invention the inputcircuit L feeding the grid G is coupled at Li to an

LOUDSPEAKER (3-9-°"1"

No. 38873o

auxiliary anode A which serves to divert someof the main discharge -stream between the cathodeK and the main anode A in rhythm with the input.

Patent issued to Dubilier Condenser Co. (1925),Ltd.

TUNING DIALSConvention date (Germany), 18th December, 1931.

No. 388938Means are provided to throw the names of broad-

casting stations covering three different wavebandson to a common projection screen, according to theposition of the wave -change switch and the setting of

October, 1933

the tuning condenser. The tuning -knob K, whichcontrols the ganged condensers C, CI, rotates a discD on which the various stations (or alternately thedegree divisions) are marked on three concentriccircles, I, 2, 3. The wave -change switch, as it is

No. 388938

operated, simultaneously moves a rod R carrying alamp P into alignment with the correspondingmarkings 1, 2, 3. A system of lenses L i, L2, L3is arranged to project the name of the particularstation being received on to a viewing -screen Slocated on the front panel of the receiver.

Patent issued to Ideal Werke Akt. far drahtloseTelephonie.

SHORT-WAVE RECEIVERSApplication date, 29th August, 1931. No. 388755

For the reception of the waves below ten metres,the anode is maintained at cathode potential whilstthe grid is biased positively. The anode and gridleads are fitted with telescoping inductances whichare connected at the other end across a tuningcondenser. The valve is intended to be operated ata point just short of self -oscillation, reaction beingcontrolled by means of a filament rheostat. Dioderectification occurs between the cathode and posi-tively -charged grid.

Patent issued to Marconi's Wireless TelegraphCo., Ltd., and E. W. B. Gill.

Convention date (U.S.A .), 16th January, 1931.No. 389225

For receiving wavelengths of the order of 10metres, a pair of valves are arranged in push-pullacross tuned input and output circuits, symmetri-cally arranged, the whole system being maintainedin strong oscillations at the carrier frequency. Anadjustable grid -leak is inserted in parallel with bothvalves, and draws considerable current. This servesperiodically to " quench " the local oscillations andproduce a super -regenerative effect. The rectifiedsignals are fed to a low -frequency amplifier througha coil connecting a centre tapping on the commonoutput coil to the H.T. supply.


1

October, 1933

RADIO -RECEIVING CIRCUITSConvention date (U.S.A.), 31st May, 1930.

No. 388732The circuit is characterised by the omission of any

standard type of rectifier, the high -frequency signalsafter passing through at least three stages of H.F.amplification being applied directly to an electro-static speaker. Since the force applied to a speakerof this type is at all times equal to the square of theapplied E.M.F., the effective pressure is invariablyin the same direction and has an average amplitudecorresponding to the envelope of the orthodox curveof rectification.

Patent issued to Dubilier Condenser Co, (1925),Ltd.

PENTODE MODULATORSApplication date, 29th August, 1931. No. 388756

A valve of the pentode type is used as a modu-lator. The H.F. carrier frequency is appliedbetween the innermost grid and cathode, and themicrophone current between the middle grid andcathode. The outermost grid, instead of beingconnected directly to the cathode, is suitably biasedto offset the tendency to develop a " falling "characteristic.

Patent issued to Marconi's Wireless TelegraphCo., Ltd., and H. J. Round.

DIRECTION -FINDING RECEIVERSApplication date, 4th September, 1931. No. 388816

A factor which tends to blur the theoreticalzero -point in D.F. reception is inherent " pick-up "by the tuning -coils. The invention consists in

applying, at onereceiving circuit,

No. 388816

or more suitable points in thea balancing component of such


magnitude and phase as to eliminate undesiredsignals due to this effect. The ordinary rotatingframe aerial A is coupled to the first H.F. amplifierV, the anode circuit of which is coupled at L to thenext amplifier Vr. The output circuit also com-prises a coil Li which is so mounted that both thedegree and sense of its coupling to a third coil L2can be varied at will. The desired balancingcomponent is derived from an auxiliary aerial Athrough an amplifier V2. In operation the D.F. setis first carefully tuned to the desired signal. Theframe aerial A is next disconnected at the pointsmarked X, and the coils Li, L2 are adjusted untilany inherent pick-up signal disappears. If theframe aerial A is now again connected in circuit, aclear-cut zero is obtained.

Patent issued to Marconi's Wireless TelegraphCo., Ltd., and F. Woods.

DRY -CONTACT RECTIFIERSApplication date, 8th September, 1931. No. 389096.

In order to prevent accidental short-circuitingbetween the metal bodies of the oxide -coatedrectifying discs r through the interposed soft -metal

7

No. 389096

washers 2, particularly at the inner edges where thewashers may come into contact with the uncoated(unoxidised) portions of the discs, auxiliary washers3 with intumed ends 7 are inserted, as shown,between the rectifying elements x and the leadwashers 2. The usual cooling -fins are shown at 4.

Patent issued to R. G. Sell and The WestinghouseBrake and Saxby Signal Co., Ltd.

SUPERHETERODYNE RECEIVERSConvention date (U.S.A.), 13th February, 1931.

No. 389026A superheterodyne circuit with automatic volume

control is described in detail, the radio -frequencycircuits being made sharply selective whilst theintermediate frequency stages are deliberatelygiven a broad resonance curve, in order to coverthe essential side -bands and at the same time totolerate a moderate amount of " mistracking "between the H.F. and local -oscillator tuning. Thefirst two valve -stages in the I.F. amplifier arecoupled together by a doubly -tuned and tightly -coupled transformer, whilst the second stageconsists of a singly -tuned transformer adjusted tohave a sharp resonance peak located between thedouble humps of the resonance curve of the firstI.F. coupling.

Patent issued to Hazeltine Corporation.


MAGNETIC -WIRE SOUND -RECORDERSApplication date, 4th September, 1931. No. 38882!

Instead of passing the tape or wire of a magneticsound -recorder through the gap between two electro-magnets of the open type, it is taken between twoprojecting pole -pieces of a circular or other closed-circuit core, so that the recording -tape shunts amagnetic flux closed on itself. This avoids over -saturation of the iron parts, and consequent dis-tortion. In addition the mechanical force on therecord is comparatively small, so that undesirablevibrations and their resultant magnetic reactionsare reduced to a minimum.

Patent issued to G. Liguori.

WIRED WIRELESS SYSTEMSApplication date, 13th January, 1932. No. 388901

Audio -frequency currents as used in L.F. broad-cast rediffusion are supplied at low voltage (forinstance a maximum of 4 volts) over an A.C.distribution network carrying a normal load of 220volts in single or multi -phase. Harmonic filtercircuits are provided, and comprise couplingsanti -resonant at power frequency, in series withcouplings resonant at power frequency. The in-ductances of the anti -resonant couplings may beprovided with secondary windings which are ener-gised from the power supply.

Patent issued to E. Hallowell and NorthernUtilities Trust, Ltd.

LOUD SPEAKERSConvention date (U.S.A.), 4th October, 193o.

No. 389839In moving -coil speakers with a direct -acting

diaphragm, the latter is usually designed to act as apiston over the lower range of frequencies up to say80o cycles, and as a wave -type radiator for thehigher frequencies. Usually the mass reactance ofthe vibrating system is the predominating im-pedance, resulting in a " mass controlled " device.In such an arrangement the efficiency is inherentlylow, most of the applied force being expended inaccelerating the vibrating system, and but little inthe useful work of producing sound. According tothe invention the diaphragm is given the form of thesurface of revolution of a straight or curved line,and the scalar magnitude of the mechanical im-pedance of the driving coil and its input circuitis made equal to the scalar impedance of the vibrat-ing parts, at a frequency between the lowest fre-quency at which the diaphragm operates as a pistonand that at which it breaks up into a complex orwave motion. The arrangement is stated to im-prove the overall efficiency, and also to suppress themain diaphragm " resonance " effect.

Patent issued to Electrical Research ProductsInc.

SHORT-WAVE GENERATORSApplication date, 21st January, 1932. No. 389233

A 3 -metre generator of the magnetron type com-prises a split -anode surrounding an incandescentcathode, the two electrodes being subjected to a

October, 1933

strong magnetic field. The parallel leads from eachhalf of the anode are bridged, inside the valve, by amovable cross -piece to form a tuned Lecher circuit,which is completely enclosed in the glass bulb. Thebridge -piece is supported on an extended glassfork, which is notched at intervals to allow thebridge to be moved from point to point to vary thetuning. In operation the valve is mounted hori-zontally so that the bridge -piece is brought closeto, without actually touching the anode leads, soas to form a capacity shunt.

Patent issued to The M -O Valve Co., Ltd., andE. C. S. Megaw.

Convention date (Germany), 11th July, 1931.No. 389346

A valve oscillator of the Barkhausen-Kurz type,for generating wavelengths measured in centimetres,is tuned by means of an extended Lecher circuitwhich is characterised by the provision of variableshunt capacities at each of the several voltagenodes. The capacities may be formed by suitablybending the Lecher wires at the points in question.

Patent issued to Telefunken Ges. fur drahtloseTelegraphie m.b.h.

SHORT-WAVE AERIALSApplication date, lith September, 1931. No. 389142

A short-wave directional aerial consists of one ormore wires or tubes 1, 2, 3, 4, bent into parabolicform, and mounted on an insulating stand S, asshown. Each of the main wires is fitted with aseries of transverse radiators, after the fashion of a

herring - bone.The transverseradiators m a ybe fitted withadjustable end -pieces for tuningpurposes, orsmall end -discsmay be providedto increase thecapacity - coup -ling between oneparabolic aerialand another.The system maybe directly ener-gised, or it mayserve as a re-flector for asingle energisedoscillator loca-ted at the focalpoint. Steam orhot water maybe circulatedthrough themain tubes whenit is necessaryto dry them out

No. 389142

or to thaw off frost or snow. "Patent issued to Marconi's Wireless Telegraph Co.,

Ltd., and G. A. Mathieu.

October, 1933

SUPER -REGENERATIVE RECEIVERSConvention date (U.S.A.), 12th June, 1931.

No. 389326Two super -regenerative amplifiers are coupled in

direct cascade, a common " quenching " frequencybeing applied to both stages simultaneously. Thequenching oscillations from a generator 0 are fedto a coil L tuned to the same frequency by a fixedcondenser C, which is connected in parallel throughleads X, Y with a similar fixed condenser CI in theinput of the first stage. The coils L2, L, L3 in the


Patent issued to Telefunken Ges. f tir drahtloseTelegraphie ni.b.h.

PHOTO -ELECTRIC CELLSConvention date (Germany), 11th December, 193o.

No. 389198In order to increase the response of a photo-

electric cell to a given light stimulus, the cell is madein two divisions. One part contains the sensitivematerial in a high vacuum whilst the other is gas -filled. The two chambers or divisions are separated

7-02

11111111-0 PHONES

z -a

0

No. 389326

input circuit of the second stage are tuned to thesignal frequency by a variable condenser C2, whilethe split coils of the first input are similarly tunedby a condenser C3. The output coil L4 of thesecond stage is back -coupled to the input coils tomaintain the system in self -oscillation, a similarback -coupling being effective between the outputand input coils of the first stage. Both stagesoperate on the push-pull principle, the grids of eachpair of valves being " quenched " alternately.


D.F. INSTALLATIONSConvention date (Germany), 26th July, 1931.

No. 389351Relates to aircraft apparatus of the type in which

the critical bearing is determined by switching twoseparate aerials alternately into circuit in different" senses." In flight it is desirable to maintain theaerial in fixed orientation with the fore-and-aftline of the vessel, so as to minimise wind resistanceand facilitate stream -lining. The usual rotatableframe aerial is accordingly replaced by a fixed one,and means are provided to apply a compensatingfactor, when, for instance, the machine is beingflown so as to compensate for wind -drift. This maybe effected either by inserting suitable resistances inthe coupling -circuits, or by varying the relativeduration of the switching periods as between one" sense " and the other.

by a thin metal foil, which is ordinarily gas -tightbut is capable of being penetrated by high -velocityelectrons. In operation the electrons emitted fromthe photo -sensitive material are accelerated by ahighly -positive anode, and pass through the foilmembrane into the gas -filled chamber, where theyionise the gas molecules and liberate a copioussupply of secondary electrons, thereby augmentingthe effective output from the cell.

Patent issued to Telefunken Ges. fiir drahtloseTelegraphie m.b.h.

SUPERHETERODYNE CIRCUITSApplication date, 17th September, 1931. No. 389505

Relates to a superheterodyne receiver fitted witha simple switching scheme to render it suitable forreceiving signals on wavelengths between 200-2000metres and also between 10-70 metres. For normalreception the first valve operates as a high -frequencyamplifier, the second as a first detector, and thethird as an intermediate -frequency amplifier, aseparate local oscillator being provided. Whenswitched over to the short-wave band, the firstvalve operates as a combined first detector andoscillator, whilst the second becomes the first I.F.amplifier. Simultaneously the short-wave coilshort-circuits the normal input coil, and the gridof the local oscillator valve is connected to groundto prevent heterodyning.

Patent issued to Kolster Brandes, Ltd., and W. S.Percival.

EXPERIMENTAL WIRELESS A5 October, 1933

PRECISIONLABORATORY APPARATUS

by

The Type 11

Variable AirCondenser isfully describedin our BulletinB -231-A.

TYPE 11

VARIABLE

AIR

CONDENSER

41The Muirhead Type 11 variableair condenser is a precisioninstrument of particularly

robust construction, capable ofwithstanding constant use withoutchange of calibration.

Quartz insulation ensuresminimum electrical losses andmaximum mechanical rigidity.

Worm drive of special designprovides subdivision to 1 partin 5,000.

MUIRHEADMAKERS of PRECISION INSTRUMENTS for OVER 50 YEARS

ELMERS END - KENTKindly mention The Wireless Engineer when replying to advertisers.

October, 1933 k6 THE WIRELESS ENGINEER &

The TYPE 87 Tested to1,000 v. D.C. Normalworking voltage 450.To withstand surges of

650 volts.

H 400-10

0 5 q IS 33 30TIME IN SECONDS

The above graph shows vol age overload when anA .C. Receiver is switched on.

SPECIALLY BUILT TOWITHSTAND HIGH SURGE

THE condensers in any mains operated gear,at the moment of switching on are subjected

to a heavy rise in voltage-in many cases as

much as 600 v. This overload can bedisastrous-and costly. Guard against it byuse of T.C.C. Type 87 Condensers-they arebuilt to withstand loads of up to 65o v. UseT.C.C. and be certain.

CapacityMids.

DimensionsPrice

Height Width Thickness

Inches Inches Inches s. d.0'1 1 i3.- If it 2 20'25 24 11 lir 2 40'5 2} 1} Ar 2 6110 24 2 ii. 3 02'0 24 2 14 4 04'0 2} 2 2} 7 3

VOLTAGES

T.C.C.ALL -BRITISH

CONDENSERSTELEGRAPH CONDENSER CO. LTD., Wales Farm Road, N. Acton, London, W. 3

3507Kindly mention " The Wireless Engineer" when replying to advertisers.

EXPERIMENTAL WIRELESS iii October, 1933

SERADEXRESISTORS

Wireless World "reports :-

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

Illustrated lists of

other types.

Type G-125 1 f wattsrating, from 9d. each.

Accuracy ± 2ro.

Wire -wound.

Colour -coded.

From your dealer or

TREVOR PEPPER,

Wake Green Road,

BIRMINGHAM.

JOURNAL OFSCIENTIFIC INSTRUMENTSproduced by the Institute of Physics with the co-operation of the National Physical Laboratory. TheEnglish Journal devoted to scientific instruments,their principles, construction and use. Essential todesigners, makers and users.

CONTENTS OF THE AUGUST NUMBER.AN INTEGRATING PHOTOMETER FOR X-RAY

CRYSTAL ANALYSIS. By B. Wheeler Robinson, M.A.,Ph.D.

A GAS ANALYSIS APPARATUS. By Dr. Rodolfo Mar aria.A NEW FORM OF INSULATION FOR THE INTRO-

DUCTION OF ELECTRICAL CONNEXIONS INTOHIGH PRESSURE VESSELS. By H. J. Welbergen.

A MOLYBDENUM -RESISTANCE FURNACE OF NEWDESIGN. By G. Burrows and F. L. Clark, M.Sc.

NEW INSTRUMENTS:A NEW MAINS OPERATED AUDIO -FREQUENCY

OSCILLATOR. By C. H. W. Brookes -Smith.A USEFUL MAGNIFIER.

LABORATORY AND WORKSHOP NOTES:ELECTRO-PLATING COPPER ON MANGANIN. By

C. R. Cosens, M.A.A SIMPLE METHOD OF FITTING FINE CROSS -

WIRES IN AN OPTICAL INSTRUMENT. By D. G.Drummond, M.Sc.

CORRESPONDENCE, NEW BOOKS, CATALOGUES,NOTES AND COMMENTS, INSTITUTE NOTES.Single copies may be obtained from the publishers, price

2s. 6d. (post free 2a. 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

VOLUME IXof

'WE

WIRELESSENGINEER

ANDEXPSRIMENTAL 'FIRELESS

Bound in Grey Cloth

Price, Complete, £2 .0 . 0

By post £2 . 1 . 0

Earlier Volumes can be supplied

Write for particulars to thePublishers

ILIFFE & SONS LTD.Dorset House, Stamford St., London, S.E.1

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

xxxxxxxxxxxxxxxxxxxxxxxxxx xCalculation xx xX

andxxMeasurement x

x xxof Inductancex

xx xx & Capacity x

xxx BY W. H. NOTTAGE, B.Sc. xx

X X(1924)X The more generally useful XX Xx formula and methods of XX measurement of inductance and xX capacity are here brought X

Ktogether in a convenient form. xX

x 224 pages. Over 8o illustra- XX tions and diagrams. XX Xxx Price 7/6. By Post 8/-. xxX ILIFFE AND SONS LTD., XX XDorset House, Stamford Street, London, xXx S.E.r, and leading booksellers. w.E. 12 XXXXXXXXXXXXXXXXXMXXXXXXN

AMPLIFIERS 1

AND ALL SOUND RE-PRODUCING EQUIPMENTFOR ALL PURPOSES.

INDIVIDUALLY MADENOT MASS PRODUCED.

Full details or demonstration from

TANNOYPRODUCTS

(Guy R. Fountain Ltd.),Canterbury Grove,

WEST NORWOOD,London, 8.E.27.

THE THERMIONIC VALVE ANDITS DEVELOPMENT IN RADIO

TELEGRAPHY.By Sir Ambrose Fleming

M.A., D.Sc., F.R.S., M.I.E.E.(19124)

A valuable book for all experimentersPrice 15/- net. By post 15/9

ILIFFE Sr SONS LTD.Dorset House, Stamford Street, London,

S.E.I. W. E .7

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

w MEWOctober, 1933

and appeals to every wireless amateur.in wireless. It covers every wireless interest

"THE WIRELESS WORLD"The weekly journal which keeps its readersthoroughly in touch with all that is happening

iv THE WIRELESS ENGINEER

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.

The

WirelessWorld

EVERY FRIDAY 4dSubscriptions:

Home & Canada, 41.1 .8, other CountriesAbroad, 41.3.10 per annum, post tree.

ILIFFE & SONS LIMITED DORSET HOUSE STAMFORD STREET LONDON SILT

Printed for the Publishers, ILIFFE & SONS LTD., Dorset House, Stamford Street, London, S.E.by The Cornwall Press Ltd., 1.6, Paris Garden, Stamford Street, London, S.E.1.

Sole Agents for Australasia-Gordon & Gotch (Australasia), Ltd.Sole Agents for South Africa-Central News Agency, Ltd.

Documents

THE WIRELESS ENGINEER