wireless engineer · 2 the wireless engineer april, 1936 transformers mains transformers and...

WIRELESSENGINEER

NUMBER 151 VOLUME XIII APRIL x936

A JOURNAL OF

RADIO RESEARCHAND

PROGRESS

PUBLISHED BY

ILIFFE & SONS LTD.DORSET HOUSE STAMFORD STREET LONDON

-41

it THE WIRELESS ENGINEER

the buyer throughthe journals he reads

and believes in

&VENOMUSE AND STUDYTRADE AND

iN TECHNICAL JOURNALS

Amateur Photographer &Cinematographer.

Architects' Journal.Architectural Review.Autocar.Automobile Engineer (Incor-

porating Motor Body Build-ing).

Brewers' Journal.British & Empire Confectioner

& Baker.British Baker.British Engineers Export

Journal.British Journal of Photo-

graphy.British Printer.Broadcaster.Bus & Coach.Caterer (The).Chemist & Druggist (The).Cigar & Tobacco World.Confectioners' Union & Ice

Cream & Soda FountainJournal.

Confectionery Journal.Contractors' Record & Muni-

cipal Engineering.

wvrearp:'eV 6

,r4.

April, 1936

FOR TRADE DEVELOPMENTAND PROGRESS

FOR FACTS OF PRODUCTION

AND DISTRIBUTION

FOR THE VERY LATESTNEWS AND INFORMATION

FOP AN INDEX TOGOODS USED AND SOLD

IF you are seeking new markets, or desire to get first-hand infor-mation about a trade or industry, you cannot go to a surer or more

authoritative source than the trade and technical journals serving it.The address of any paper in this list will be furnished on request.

Crushing, Grind-ing, Mining &QuarryingJournal.

Dairyman.Drapers' Organiser - The

Fashion Trade Magazine.Drapers' Record.Dyer, Textile Printer, Bleacher

& Finisher.Economist.Electrical Industries & Invest-

ments.Electrical Review.Electrical Times.Electric Power (Industrial).Electric Vehicles & Batteries.Engineer (The).Engineering.Engineering & Boiler House

Review.Export Trader.Farm, Field & Fireside.Flight.Footwear Organiser & Shoe

& Leather Trades ExportJournal.

Fruit, Flower & VegetableTrades' Journal.

Furnishing Trades' Organiser.-The Magazine for HouseFurnishers.

Gas Engineer.Grocer & Oil Trade Review.Grocers' Gazette & Provision

Trades' News.Grocery & The Provision

Merchant.India Rubber Journal.Ironmonger (The).Journal of Decorative Art.Kinematograph Weekly.Laundry Record & Dyeing &

Cleaning Trades Journal.Licensing World & Licensed

Trade Review.Locomotive, Railway Carriage.

& Wagon Review.Machinery.Meat Trades' Journal.Mechanical World & Engineer-

ing Record.Men's Wear.Mining Journal, Railway &

Commercial Gazette.Model Engineer & Practical

Electrician.Modern Transport.Motor Cycle.Motor Cycle & Cycle Trader.Motor Trader & Review.Motor Transport.Motor World & Industrial

Vehicle Review.

Oil & Colour Trades Journal.Optician.Outfitter.Overseas Engineer.Packing, Packaging & Con-

veying Gazette.Photographic Dealer.Plumbing Trade Journal.Pottery Gazette & Glass Trade

Review.Power Laundry.Rural Electrification& Electro-

Fanning.Specification.Style for Men-The Magazine

for the Man's Shop.Textile Manufacturer.Unit (Domestic).Waste Trade World.Watchmaker & Jeweller,

Silversmith & Optician.Wine Trade Review.Wireless & Gramophone Ex-

port Trader & Buyers' Guide.Wireless & Gramophone

Trader.Wireless Engineer & Experi-

mental Wireless.Wireless World.Yachting World & Motor

Boating Journal.

Official Handbook, giving particulars of over 500 papers, post free 1/-.

PERIODICAL. TRADE PRESS AND WEEKLYNEWSPAPER PROPRIETORS' ASSOCIATION. LTD.

TelephoneCentral 2441 6, Bouverie Street, E.C.4

tti%>kAsSOSIUSsg...

Telegrams:Weneppa, Fleet, London

6V44/04, ,..*NEONA.091tUati, ,9

°

April, 1936 THE WIRELESS ENGINEER I

BEARINGS

25 -IWORM

ADJUSTABLEWORM TENSION

MAINDIAL1801 ARC50 D1VNS

0.064"ALUMINUMPLATES WAXED

ISOLANTITEVIG"

ALUMINUM -ALLOYSPACERS AND TIERODS

- 3/4," ALUMINUMPANEL

3 -POINTSUSPENSION

PHOSPHORBRONZE -TO.BRASS ROTORCONTACT

ISOLANTITETERMINALINSULATOR

BALLBEARINGS

HEAVYALUMINUM -ALLOYCAST FRAME

TABLESUPPORTS

FOR OVER FIFTEEN YEARS our well-known " GENERALRADIO " Type 222 Precision Variable Air Condensers were thestandard by which all other variable condensers were judged.The Type 222 Condenser was extensively copied, throughoutthe civilised world. No better compliment could have been paidto it and no better proof of its superiority could have beendemanded.However, during the past year or two GENERAL RADIOENGINEERS have realised that even the 222 Condenser was atlong last due for a thorough " vetting." After the most carefulresearch, extending over about fifteen months, the new type722 Precision Condensers have been standardised. These nowreplace the type 222, which are no longer sold by us.The Type 722 Condenser is the best in the world, and it sellsat a strictly reasonable price. For complete details pleaserequest our special pamphlet describing the many real advantagesof this new Condenser-advantages which are not offered byany competitive source whatsoever.In these condensers there are 5,000 worm divisions for one-half rotation of the condenser shaft. The worm divisions arelarge and may clearly be set to 1/5th division. This splits upthe half -revolution of the main shaft into zo 000 easily duplicatedpositions. This permits settings to 0.06 micromicrofarad ando.o2 micromicrofarad respectively for the models 722-H and 722-F.Extraordinary stability, positively unattained previously, is anadditional welcome feature.BUT WHY NOT SEND FOR COMPLETE INFORMATIONTO -DAY ?

A NEW

LABORATORY

CONDENSER

FIGURE 1. View of TYPE722.11 Precision Condenserwith cabinet removedshowing constructionaldetails.

FIGURE 2. Panel view of TYPE 722-H Pre-cision Condenser.

PRICE LIST.

722 -HM. Cabinet. 1400 MMF . . £29 .10 .0722-HU. Unmounted. do. . . £23 .10 0722 -FM. Cabinet. 500 MMF. . . £28 . 0 .0722 -FU. Unmounted. do. . . £22 . 0 .0

EXPORT PRICES ON APPLICATION.

CLAUDE LYONS LIDHead Offices: 76, OLDHALL ST., LIVERPOOL, 3.

London: 40, BUCKINGHAM GATE, WESTMINSTER, S.W.1.

Kindly mention The Wireless Engineer" when replying to advertisers.

1

2 THE WIRELESS ENGINEER April, 1936

TRANSFORMERSMains Transformers and Chokes.Intervalve, Microphone andValve - to - Line Transformers

of the High Grade MiniatureTransformers

5

-- 0w

to 8 8

FREQUENCY

Response curve of B.S.R. heavy-dutyoutput transformer. (Reproduced by permission of REPRODUCERS LTD." WIRELESS WORLD" 'Phone: Cradley Heath 6212-13.

'Grams: Electronic, Old Hill. CLAREMONT WORKS, OLD HILL, STAFFS.

Type.

for Microphone Amplifiers, etc.

Special Transformers for Laboratory Work andfor Equipment for Film and Record Recording.

COMPLETE LISTS ON REQUEST.

HETERODYNE OSCILLATORSfor

LABORATORY, OUTDOOR & INSTALLATION WORK

BIRMINGHAM SOUND

TELEVISION DEVELOPMENTSA NEW LARGE DIAMETER TUBE FOR TELEVISION REPRODUCTION

We also supplyHIGH VACUUM

RECTIFIERSGAS -FILLED RELAYSMERCURY VAPOUR

RELAYSfor Television Scanning circuits

The Ediswan Type CH tube hastelevision picture 8" x 6" without diThe screen is of special materialblack -and -white picture.

Specification : IndirectlyHeater Volts - - - 2.0Anode Volts - 3500

a screen diameter of 10" giving astortion.giving a close approximation to a

heated high vacuum tube.1st Anode Volts - - - 1200Sensitivity (mm./V.) - 750/V*

*V= anode volts

List Price 12 . 0 . 0ENTIRELY BRITISH MADE

E D I SWANRADIO VALVESTHE EDISON SWAN ELECTRIC CO LTD Ala CHARING CROSS RD., LONDON. W.C.2

R.H.280

USED BY ALL GOVERNMENT DEPARTMENTS

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

April, 1936 THE WIRELESS ENGINEER

PRECIOUSMETALCONTACTS

FINE SILVER

STANDARD SILVER

IRIDIO PLATINUM

PLATINUM SILVER ALLOY

PLATINUM GOLD SILVER

and other special contact alloys.

The reliability of the precious metals and their alloysas contact media is becoming more and more evidentfrom their increasing application to various types ofelectrical equipment. Johnson, Matthey & Co., Limited,have been specialists in the manufacture of preciousmetal contacts for many years and the knowledge andexperience of their technical experts is at your disposal.

Stocks are held of all the types and sizes of contacts incurrent demand, but a large proportion of those which wesupply are manufactured to conform to particularrequirements and specifications.

JOHNSON; MATTHEY& CO., LIMITED

HEAD OFFICE & REFINERIES:

73/83, HATTON GARDEN, LONDON, E.C.1.

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

(A- 4

SIMPLETHE WIRELESS ENGINEER April, 1936

MEASUREMENrimmiHidimilimilm-iiimmounincpffinimei

of FREQUENCY30-30000 '<cis

0.01 %

CALIBRATIONENTIRELY

INDEPENDENT

OF VALVEREPLACEMENT

AND AMBIENTTEMPERATURE

30

Z

20cc (13 20

CC

0 '".."0.

CC W

-g-w3

10z

ci

ui

ELL-

020 40 0 60 70 Rapid interpolation to 0.001%

is obtained with wavemeter

List No. R.205.

_s. H. W. SULLIVAN LIMITEDLEO STREET LONDON, S.E.I5

List No. R.202.

is

now

possible

by

means

of

SULLIVAN - GRIFFITHS

SUBSTANDARD

DYNATRONWAVEMETER

MINUTES FROM SWITCHING ON

Curve showing the settling down of frequency ofSullivan -Griffiths Wavemeter.

STABILITY 0'001 PER CENT.or much better when used

under good conditions.

RAPID INTERPOLATION TO 0'01%due to strict conformity tolinear law of frequency.

.ter-, acct. -re -Ai


AS

The

WIRELESS ENGINEERA Journal of Radio Research & Progress

EditorHUGH S. POCOCK

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

VOL. XIII. No. 151

APRIL 1936

C 0 NT EN T S

EDITORIAL

VARIABLE SELECTIVITY AND THE I.F. AMPLIFIER. Part II.By W. T. Cocking

175

179

THE ANODE TO ACCELERATING ELECTRODE SPACE INTHERMIONIC VALVES. By J. H. Owen Harries 190

VOLTAGE MEASUREMENTS AT VERY HIGH FREQUENCIES-II (Concluded). By E. C. S. Megaw, B.Sc., -D.I.C. 201

ABSTRACTS AND REFERENCES 205

SOME RECENT PATENTS 232

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.zTelegrams: " Experiwyr Sedist London " Telephone : Waterloo 3333 (5o lines)

Branch OfficesCOVENTRY BIRMINGHAM MANCHESTER GLASGOW

x9 Hertford Street Guildhall Bldgs., Navigation St., 2 26o Deansgate, 3 26B Renfield Street, C.2Telegrams: " Autooar, Coventry" Telegrams: " Autopress, Birmingham" Telegrams: Manchester" Telegrams : Me, Glasgow"Telephone : 5210 Coventry Telephone : Midland 2971 (4 lines) Telephone : Blaektriars 4412 (4 lines) Telephone : Central 4587

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," Dorset House, Stamford Street, London, S.E.i. Especial care should be

taken as to the legibility of MSS. including mathematical work.

ft 6 THE WIRELESS ENGINEER April, 1936

DCONDENSERS SRRESISTANCES

UBILIE. . . . owe their pre-eminence to -day to the factthat for over twenty-fiveyears they have been built toexacting standards and areconsistently right.

DUBILIER manufacture Condensers

for every electrical purpose and

anticipate all modern requirements.

Interesting literature referring to

DUBILIER products is availablegratis on application.

Oil -Immersed Condensers for high fidelityamplifiers and high definition Television.

DUBILIER CONDENSER CO. (1925) LTD., DUCON WORKS, VICTORIA ROAD, N. ACTON, W.!3Kindly mention " The Wireless Engineer" when replying to advertisers.

(7S

WIRELESSENGINEER

Vol.. XIII. APRIL, 1936. No. 151

The Magnetic Recording of Sound

THE magnetic recording of sound is byno means new, but it is only quiterecently that it has been developed

to the stage where it can compete withmechanical and photographic methods. Boththe A.E.G. and the Lorenz Co. have recentlyput magnetic recorders on the market, butthey differ very essentially in the nature ofthe record material. The former use a filmcoated with a deposit of iron, whereas thelatter employ a steel band. The Lorenz Co.*have recently supplied their steel band re-corders to the German Broadcast authoritiesfor outdoor broadcasts, for whidh purposeit is claimed to be specially suitable onaccount of its immunity from the effects ofvibration.

The steel band is 3 mm. wide and o.o8mm. thick (about it inch by 3 mils). This iswound off one drum on to another, andalthough it is possible to use a length suitablefor about 5o minutes, the usual length runsfor 3o minutes. The inner diameter of thedrums is 20 cm. and the outer for a 3o -minute run about 57 cm. The workingspeed is about 1.5 metres per second and thespeed of the drums therefore varies fromabout 5o to 15o revolutions per minute.The maintenance of a constant band speedis ensured by an induction motor of amplepower driving a pulley f over which theband passes under considerable pressure

* See " Die neue Stahlton-Bandmaschine."Lorenz Berichte, Jan., 1936, page 49.

from an endless fabric band shown dottedin Fig. 2, the tension on which can be ad-justed by means of the upper pulley 1. It isvery important to avoid any slack betweenthe drums and the driving pulley ; thisnecessitates very rapid stopping of the drumsif the driving motor stops, and this is not asimple matter in view of the large momentof inertia of the rotating drums when loadedwith the steel band. The leading drummust always tend to run faster than theband speed, whereas the trailing drum mustbe braked sufficiently to maintain sometension on the band. The two drums mustalso be able to reverse, because when arecording is complete it is necessary towind the band back on to the other. drumbefore it can be reproduced. This neces-sitates reversing the driving motor and ex-changing the roles of the two drums, thedriven drum becoming the braked drum andvice versa. How this is done is shown inFig. 3 ; the band a is running on to the leadingdrum b fixed to the shaft k which is coupledto the shaft g by the friction drive f the slipof which allows the drum to rotate at theband speed whatever the speed of g. Thepulley c is driven by the belt d from the maininduction motor, but this pulley is loose onthe shaft which carries the coupling e. Ifthis sliding coupling e is locked to the fixedpart h, the pulley c runs idle and f acts asa friction brake on the drum, whereas if eis locked to the pulley c, g runs faster than

B 2

THE WIRELESS ENGINEER

Fig. i.

k ever runs, and thus maintains tension onthe steel band. In case of a stoppage therapidity with which the trailing drum canbe brought to rest, and thus prevented frompaying out steel band, depends on the

dd

I I '

Fig. 2.

moment of inertia-which depends on thelength of band, i.e., on the length of recordrequired-and on the power which one is

April, 1936

prepared to dissipate in braking. Bothdrums can be removed and new ones fittedin less than a minute.

Between the left-hand drum and thedriving pulley the steel band passes through

/1==IIMMIIMIMINIMIIMIIMIn111111111111

Fig. 3.

the wipe-out and recording apparatus (Fig.5a, and c in Figs. i and 2) and similarlybetween the pulley and the right-hand drum


it passes through duplicate sets of repro-ducing apparatus (Fig. 5b, and d in Figs.i and 2). The principle of these can beseen from Figs. 4 and 6. Two chisel -endedpole pieces are pressed by springs againstopposite sides of the band and the operatingcoils are wound round the pole cores. Thesepole pieces are not exactly opposite eachother but are displaced by an amount a(Fig. 6) which can be adjusted by the micro-meter screws and verniers seen in Fig. 5.The operating flux is that along the bandbetween the pole pieces. The wipe-outmagnets are energised by direct current andremove all trace of previous magnetisationby saturating the band up to S (Fig. 7).On passing on, the value of H not only fallsto zero but is actually reversed, due to thestray field (Fig. 6), thus bringing the mag-netisation back along the hysteresis loop tosome point R,. When the band passesbeyond the influence of this stray field, themagnetisation will follow some such line

Fig. 4.

as R,R,'. The band now passes betweenthe recording pole pieces which are energisedby two currents, one a steady direct currentig which would bring the magnetisation tosome such point as V in the middle of a long

177

linear portion of the curve, the other thecurrent iw to be recorded.

Fig. 6.

In Fig. 7 this current is represented as asine curve causing the point on the curve tomove sinusoidally between V. and Vb.As the band passes out of the field it will beleft with a magnetisation varying sinusoidallybetween Va' and Vb'. The disturbing effectsof the stray fields make it impossible toutilise more than a small portion of theleft-hand side of the hysteresis loop ; boththe direct current and the alternatingcurrent must be kept small.

There are duplicate sets of reproducingapparatus side by side as shown in Fig. 5b.The principle of these is illustrated in Fig. 8where the upper diagram shows the assumedmagnetic state of the band and the lower

(Left) Fig. 5a.-Wipe-out and

recording.

(Right) Fig. 5b.-Reproducing

apparatus.

178 THE WIRELE

one the path of the resulting flux throughthe soft iron cores upon which are the coilsin which an e.m.f. is induced by every

variation of the flux, i.e., by every variationof the magnetic condition of the passingband. At first sight one might expect theinduced e.m.f. to increase with the frequencysince the rate of change of the flux will begreater ; this is true up to a certain point,but for very high frequencies the length ofband corresponding to a wavelength willbecome comparable with the distances aand b in Fig. 8 and the impressed magnetisa-

7654

4.)t. 3

2

10

k 87

0 5(1)

r441)3

O 2

SS ENGINEER April, 1936

tion and the induced e.m.f. fall off. Fig. 9shows the variation of voltage in the repro-ducing coils for a signal current of constant

amplitude and variable frequency. Curvesand 2 refer to different adjustments of the

distance a between the pole pieces of thereproducing apparatus. By suitable designof the amplifier it is possible to correct forthis variation of output and so obtain aconstant output from 70 to 5,50o cyclesper second.

As in every method of sound recording,there is a background of noise ; in this

case it is due to magneticinequalities in the bandand also to the mechanicalvibrations set up in themagnet cores by the fric-tion of the passing bandagainst which they arebeing pressed. Special pre-cautions have been takenin the design and construc-tion of the reproducingapparatus to reduce thisnoise to a minimum.A Lorenz steel -band re-corder mounted in a caris in regular use by theGerman Broadcasting Com-pany.

/\\am

2 3 4 5 6 7 8 910 2 3 4 5 6 7 89103

Cycles Per, (SecondFig. 9.

2 3 4 5 6 7 8910.

G. W. 0. H.

April, 1936 THE WIRELESS ENGINEER 179

Variable Selectivity and the I.F. AmplifierPart II.*-Coupling the I.F.

By W. T.

THE conditions which govern the designof the coupling between the last I.F.valve and the detector differ in two

ways from those which govern the choice ofa coupling between two amplifying valves.In the first place the detector has a compara-tively low input resistance, and, secondly,the voltages involved are much greater.These two factors greatly modify the designof the coupling, for the effective H.F. re-sistance of the circuits employed is increasedby the input resistance of the detector, whilethe high voltages involved sometimes neces-sitate the design being carried out primarilyfor the avoidance of amplitude distortion ;amplification and selectivity then becomesecondary considerations.

B

M0

L - 21

Fig. i3. -A single tuned circuit coupling twovalves can be generalised in the manner shownhere where R' is the input resistance of the

second valve.

There are many different arrangementswhich can be employed, and as a choice willbe dictated by the requirements of the re-ceiver in other respects, it is necessary toconsider them all. The first system em-bodies a single tuned circuit, and the gene-ralised circuit is shown in Fig. 13 in whichRa is the internal A.C. resistance of the lastI.F. valve and R' the detector input resist-ance. By the straightforward applicationof Kirchhoff's laws the following equationcan be developed

62pRico2M1m2

el Zi(Z2Z3 w2M22) Z3co2M12

where Z1= R0 + jcoLi, Ra when Ra>>caL,

(41)

* MS. accepted by the Editor, October, 1935.

Amplifier to the DetectorCocking

Z2 = R + jcuL - jlcuCZ3 = R' + R' when R' >>coL,

Equation (4i) reduces to

e. a, a2z

ei A/rji a22)2 y2Q2) (42)

Where w2M12/RRa

a22= w2M22/RR'Q = wL/Ry = (w2LC - i) = 2n/f when n < < fg = 1.4IRa

n = frequency different from reson-ance

f = resonance frequency.It can readily be seen that the impedance

into which the I.F. valve works, that is theimpedance between AB of Fig. 13, is atresonance the transferred resistance of thefollowing circuits, for the reactance of L1 isusually small enough to be neglected. Theresistance of the tuned circuit . is increasedby the detector loading to R(1 a22) sothat the load RL on the valve is this valueof resistance transferred to the primary, or

co2m12R

R(i + a22)

When y = o, equation (42) reduces to62 ga1a2V Rol?'e, I + a12 a22

and it can be seen that this is a maximumwhen either

al2 a22

or a22 --- I ± a,2

The optima for al and a2 cannot be satis-fied simultaneously, but if either be heldconstant an optimum can be found for theother. In practice, therefore, it is necessaryto fix the value of one arbitrarily, and it ismost convenient to do this by making either

(43)

(44)

Li = L = M, or L2 = L M2,

18o THE WIRELESS ENGINEER April, 1936

for then one circuit disappears and the systemcan be redrawn as in Figs. 14. and 15 respec-tively. In the case of Fig. 15

a12 = w2L2/RRa = RD/Ra

and for Fig. 14a22 = w2L2/RR' = RD/R,

where R, = w2L2/R = wLQ = the dynamicresistance of the tuned circuit alone.

The choice betweenthe two arrange-ments depends onwhich is the greater,the A.C. resistanceof the I.F. valve orthe input resistanceof the detector. IfRa > R' the circuitof Fig. 14 shouldbe adopted, but ifRa < R' then Fig. 15is better, for thiscourse leads to the minimum H.F. resistancefor the tuned circuit assuming that theother coupling is optimum.

When Ra < R', as when a triode is usedfor the I.F. valve,

a22 RD/R,

(ai2),, = I + a22 = + RDIR'and equation (42) reduces to

e2 g1/RaRD(i RD/R')

et 1444 RD/R')2 y2Q9 (45)

g- V RaRDwhen y o . . (46)

2V I + RD/R'

The physical significance of this is morereadily seen if the terms are arranged in theform

Fig. 14.-When R.> R',a single circuit is bestconnected in this manner.

=e X RDR'e, 2 RD ± R'It can be seen that the gain is equal to

half the amplification factor of the valvemultiplied by the turns ratio of an idealtransformer of turns ratio equal to the squareroot of the ratio of secondary to primaryresistance, the primary resistance being thevalve resistance and the secondary thedynamic resistance of the tuned circuit inparallel with the detector input resistance.The dynamic resistance of the tuned circuitrepresents a loss and the gain reaches its

maximum value when RD = co for then theloss is absent.

It is clear that the higher the value of RDthe better from the point of view of amplifica-tion, and this is also true if amplitude dis-tortion be important. Whatever the valueof RD, the load RL on the valve remainsthe same, for

w2M12a12RaR, = R(i a22) a22

= Ra

since a12 = I + a22If the valve will give a certain undistortedoutput when operating with this load agreater detector input is secured by the useof a high value of RD than a low one, for theincrease in amplification does not occur inthe valve but through the reduction of lossesin the coupling system.

The selectivity S is given by the ratioof equations (46) and (45)

= 1/E4(1 RD/R')2 y2Q2](47)2(1 + RD/R')

This equation, however, is of little help indesign for it is usually required to determineQ or RD, the desired values of S and L beingknown. RD is a function of Q so that fordesign purposes it is better to derive therelationship of Q and S in a different manner.

Let the total resistance of the tunedcircuit be

= R w2L2/R' ,0231121Ra

= R(i a12 a22)

= 2R(I a22) since a12 = a22

= 2R + 20,2L2/R'whence R = R,12 ca2L21R,

and Q = wLR,./2 - (02L2/R' (48)

Fig. 15.-When R. < R', the output circuitR' is connected to the tuned circuit directly.

The ratio of response at resonance toresponse at any other frequency for a series

April, 1936 THE WIRELESS ENGINEER 18r

resonant circuit is

S - V[R22 (coL - I/o,C)2]RT

= + y2QT2

where QT = coLIR,

QT ="V/ S1/ 2 - VS2 -y 2n

2nad.and RT COLIQ

f VS2 -combining (48) and (49)I

Q =coL

f V S2- I R'

(49)

(50)

Practically, there is a limit to the selec-tivity obtainable without regeneration, forQ tends to infinity as ntIVS2 - i approachesequality with coLIR'. When n/f1/S2 -< coL IR' , Q is negative and this means thatR must be negative. This can be obtainedby means of regeneration and it should benoted that Q or R negative does not neces-sarily mean QT or R, negative, so that thesystem can remain stable. This conditioncorresponds to the one commonly knownas removing the valve damping with reac-tion.

In design it is necessary to know f, n, S,L, R', and Ra. Equation (5o) gives Q, and

R = coLIQC = ilco2L

v[RR2a

Ri, = coLQ . . . . . (54)If too high a value be chosen for S, Q will

be negative, or positive but impracticablyhigh, and the conditions can then only besatisfied by the use of reaction.

Now in modern practice a triode valve islikely to be used only when the output of atetrode or pentode is inadequate. Theuse of a suitable triode enables a muchgreater undistorted output to be obtained,and this is indeed probably the only justifica-tion for its employment. If it is used forthis end, however, it is important to see thatit is used to the best advantage, and the con-ditions just discussed, although they leadto the greatest amplification, are not the bestfrom the viewpoint of distortion.

The detector is a power operated deviceand the I.F. valve must give out powerrather than voltage. Although maximumpower output is obtained from a valve underthe same conditions as those for maximumvoltage amplification, namely, that RL = Ra,it is well known that the maximum undis-torted output is obtained when 12, > Rafor triodes. The optimum relationshipvaries with different valves, but for smalltriodes it is (usually close to RL = 2Ra.If this relationship is adopted, the maxi-mum undistorted power will be secured inthe load and hence the maximum valueof e2, but e1 will be greater than whenRL = Ra and the amplification correspond-ingly lower. The selectivity will also belower. Where it is necessary to use a triode,however, amplification and selectivity areusually of secondary importance to freedomfrom distortion, so that the condition thatRa = 2Ra is usually the best.

Suitable design equations for this casemay be derived by writing

RL = 2Ra = w2m12iR(1 a22)

whence a12 = 2(1 + a22) = 2(1 RD/R')for a22 = RD/R' as before.

Inserting these values in equationgives

e2 g1/2RaRn(I RD/R')e D')2

1 V[9(1 RIR + y2(22]

g1/2R.RD when y = o (56)3 Vir -F RD/R']

The design equations may be derived inthe same way as before and are :

(42)

(55)

Q = (57)2n coL

3/Vs2 R'

R = coLIQ

Mi= V[2R-a?--Ra (I + RD/R')]

S 1/[9(I RD/R')2 + y2Q2]. .

3(1 + RD/IV)

(58)

(59)

(6o)

The choice of the valve which precedes thecoupling is of considerable importance. Theequations show that whatever the loadimposed upon it, the greatest stage gain issecured from the valve having the highest


value of g-/Ra = /21.0?. The accom-panying Table gives values of g1/Ra fortypical modern triodes, the values being fornormal working voltages and not the makersrating at zero grid bias. It will be seen thatthe MII4I stands out above the others fromthe point of view of amplification. With atriode, however, output is more importantthan amplification and the greatest outputwill be secured from the valve having thegreatest D.C. anode dissipation. A lowresistance valve such as the ML4 would bechosen, therefore, when maximum outputis required. In general, however, the out-put of this valve is unnecessarily large forordinary purposes, and when this is so eithera higher resistance valve can be used or thelow resistance one retained and operated ata lower anode voltage to economise in anodecurrent. The ML4, for instance, consumesabout 20 mA. anode current at 200 volts,and this is higher than can often be allowed.The use of a lower anode voltage only leadsto a saving through the reduction in anodecurrent which it entails and is not in itselfusually of any benefit. It is better, there-fore, to employ a higher resistance valveoperated at the full 200 volts, because theanode dissipation is higher for the sameanode current consumption and the outputis consequently larger. At the same time,higher amplification is likely to be secured.This effect is quite marked in practice.

These remarks about output always holdtrue, but those regarding amplification aretrue only if feed -back effects through thegrid -anode interelectrode valve capacity arenegligible, and they are negligible only ifthe valve capacity be carefully neutralisedor the input circuit between grid and cathodebe of very low impedance. The input im-pedance due' to feed -back effects dependson the ratio of the voltage across AB (Fig.13) to el and not upon the total stage gain.

For maximum input impedance, therefore,the gain in the valve should be as small aspossible and the gain in the coupling ashigh as possible. If it be possible to obtainthe same value of e2/e, with two differentvalves, one of high resistance and the otherof low, the latter should be selected, for ifthe grid -anode capacities be equal, its inputimpedance will be the higher. It is also thebetter valve from the point of view of poweroutput, although it will consume a greateranode current.

Colebrook gives the following equationsfor the input impedance of a valve when theanode load is tuned to resonance.

'g - Ga + G,(6I)R

COacga2 '

Ca' = Cga Ga +g

G,/. . (62)

where Ra'= input resistancet due to feed -

C91= input capacity backCaa= grid -anode capacityGa =GL = x/RL

A tuned coupling between the last andthe penultimate I.F. stages is hardly feasiblefor Ra' will usually be low and negativeat certain frequencies and instability willoccur. The best scheme, if neutralisationis not adopted, is that put forward by Cole-brook and shown here in Fig. 16. A fulldescription of its characteristics appearedin his original paper', and it is sufficient tosay here that if a high gain is to be securedfrom V1 the input capacity C,' of V2 mustbe as high as possible. When V2 is workingunder optimum amplifying conditions Ga

1 A Study of the Possibilities of Radio -FrequencyVoltage Amplification with Screen -grid and withTriode Valves, by F. M. Colebrook, B.Sc., Journ.I.E.E., February, 1934.

TABLE.

Type. P. RaQ

gmA/v.

g 1/1? Cooµ.µF.

()EggX io-u

ML 4 .. .. 12 3,500 3.43 0.2025 7 84MHL 4 .. .. 20 10,000 2.0 0.2 7 140MH 4 .. .. 40 15,000 2.66 0.327 7 280MH 41 8o 18,00o 4.44 0.595 7 56oMHD 4 .. .. 40 20,000 2.0 0.283 3 120


GE and equation (62) becomes= C22(i µ/2) /4.C22/2 .. (63)

When V2 is working under conditions ofmaximum undistorted power output Ga2GE and equation (62) becomes

C9' = C22(i 2/43) 2p,C22/3 .. (64)

In either case it is important that µC92be as small as possible and this is the criterionof a good valve from the point of view ofinput capacity.

Fig. 16.-The Colebrook double -triode amplifier.C1 represents the stray and valve capacities andR,' and C9' are the input resistance and capacity

of V, due to feed -back effects.

Previous conditions give a maximum ofg1/ Ra = 1411/ Ra as desirable. The bestvalve, therefore, is the one which givesmaximum values for both µ/A/R2 andI/p,Cga.

Turning now to the use of tetrode andpentode type valves, Ra is usually very highand with diode and grid detectors Ra isnearly always greater than R'. The con-nections of Fig. 14 are used, therefore, inaccordance with the rule laid down earlier,and the values to be inserted in equation(42) are

a12 RDIRa

(a22)09 (1 a12) RD/Ra)

Equation (42) then becomes

e2 RDR' (I RDIRa)

ei (I RD/Ra) 2y2Q2] (65)

which reduces to

e2 / RDIV

e1 2A/1 RD/Rawhen y = o . . (66)

By a similar process to that adoptedearlier the relevant design equations can be

derived and are as follows :-

S - 1/[4(r RD/Re)2 +Y2Q2] (67)2(1 + RD/Ra)

I(68)Q wL

i1/ S2 - I RaR = coLIQ (69)

C (70)

M2 = V[RT4 RD/Rad (7')+RD =-- caLQ . . .. (72)

At resonance the load ongiven by

R= (021.2

the valve is

RDL R w2m 221R, 2 + RDIRa (73)

In very many cases RD< <Ra and thenRDIRa o and the formulae can be simpli-fied accordingly.

As RE < Ra the value of RE has very littleeffect upon the dynamic characteristics ofthe valve and the dynamic curves are littledifferent from the static. There is conse-quently no value of RE which can normallybe reached which is particularly desirablefrom the point of view of obtaining maximumundistorted output. Distortion in the valveis set rather by the static characteristicsthan by those characteristics in conjunctionwith the load resistance. This being so, itfollows that there is a maximum input tothe valve which is set by the appearance ofdistortion and that this input is almostindependent of RE. For a given value ofR', both RE and e2/e1 increase with an in-crease in the value of RD ; maximum outputis secured, therefore, for those conditionswhich give maximum stage gain.

Even with screened tetrodes and pentodesthe grid -anode valve capacity is not zero,so that the input resistance is still of im-portance. M. O'Connor Horgan2 gives thefollowing equation for the limiting conditionfor stability for a valve with tuned grid andanode circuits.

2 = g(OC90RLRI7

where R9 is the dynamic resistance of thetuned grid circuit. This condition is ob-viously the one where the input resistance

2 " The Grid -Anode Capacity of Valves," byM. O'Connor Horgan, M.Sc., The Wireless Engineer,September, 1934.

THE WIREL

of the valve is negative and has its minimumvalue which is just equal to Rg. Thisequation, therefore, can be used to give theminimum value of Rg', the input resistanceof the valve, by writing it in the form

Rg = - Rg' = 2IgcuCR,,The characteristics of any system are

most easily realised by means of typicalexamples and a common case will accord-ingly be considered. Suppose the valve hasg= 2.7 mA/V., Ra = 1MD, and Cga = 0.0026pi.LF. A common value for R' is 1.25 x 105E2,for a diode detector can hardly have aload resistance higher than o.25MQ ina high quality receiver and the inputresistance of a diode rectifier operating witha large input tends to one-half the D.C. loadresistance. For a frequency of 465 kc/s,L cannot normally be greater than 2,000 µH.

If n = Io4c/s and only 1 db. loss can betolerated at this frequency, S r 1.12 andequations (68) to (72) give Q = 26.9,R=2175, C = 58.8 pg.., RD = 1.57 X 105Q, M2 = 1,920 pH, and Rg' = 1.38 M.Q.

Fig. 17.-A commonly used circuit for theI.F. stage and the detector.

Now L2 cannot be larger than L if the circuitequations are to apply with any reasonabledegree of accuracy, for with L2 > L the seff-cap.city of L2 will become important.Even if L2 = L, M2 cannot be as great as1,920 µH. when L = 2,000 t.cH., for thiswould demand k = M2IV LL, = 0.96,-adegree of coupling which it seems impossibleto obtain. In this case M2 is so near L.thatthe difficulty can be got over by connectingthe detector across the tuned circuit so thatM2 = L. Assuming optimum values, equa-tion (66) gives e2/e1 = 130.

It is possible, however, to obtain a valueof Q = 100 for reasonably inexpensive coils.If Q = 93.5 (the value used for the three -stage filter of Part I and employed here since

last

ESS ENGINEER April, 1936

the same coils would presumably be usedthroughout a receiver) and L = 2,000 µH.,Ri, = 0.545 MD, R = 62.5 S2, C = 58.8ppF., M2 =. 1,190 µH., Rg' = 0.47 M.S2 ande2/e1 = 440. For n = io4c/s, S = 1.635(4.25 db.).

Fig. r8.-The generalisedequivalent circuit of Fig.17.

e

It may be remarked that although theeffect of the input resistance of an A.V.C.rectifier has not been treated, it may readilybe allowed for by lumping it with the de-tector input resistance ; that is, R' can betaken as being not the detector resistancealone but this resistance in parallel with thatof the A.V.C. rectifier.

Now a single tuned circuit is quite rarelyemployed in the detector coupling of modernsuperheterodynes, and a pair of coupledcircuits is almost universally used, theA.V.C. system being fed from the primaryas shown in Fig. 17. This circuit may begeneralised as in Fig. 18, and it should benoted that if the A.V.C. system be paralleledwith the detector, as is sometimes the case,the following equations may still be usedby writing RAyo = co, and R' = detectorand A.V.C. input resistances in parallel,and eAvu = e2.

By the application of Kirchhoff's laws,the following equations can be derived :

e2 PR'W3M21111/12el Z1 Z2(Z3Z4+W2M22)+Z4w2M2}(74)

w2M12(Z3Z4 + OM 22)8AT° ILCO2M iM 3(Z gZ 4 + OM 22)

el ZAZ2(Z,Z4±(023/22)+ Z4CO2M1 75)=.

±(.02/1/i2(Z3Z4-1-to2M22)

whereZi = R. ± ju,L3 Ra when R, >> cuL3Z2= R,± u.)2M321R, jcuL, - j/wC =

Ri(i a32 jyQ1)

4

April, ;936 THE WIRELE

Z3 = R24 -/COL 2 -j/GUC 2= R 2(I 1..Y Q2)

Z4= R' jeuL4 = R' when R' >> wL4Li= total inductance of first tuned circuit.L2 = total inductance of second tuned

circuit.Then

e 2 gactiaV RaR'81= v[{(1 + a 22) (i

2

a12 + a 32) a2 (76)

y2Q392}2 y2Q12{1 a22

Q2/121(1 + a12 a32))2]

gaaia2VR.R'B (77)

- ga,a 3A/ [RaR, {(I + a 22) 2 + v2() _21""1(78)

BeAre

81

eAr, _.a31 (i a22)2 _4_ y2(222

82 Raa 1/R I2 f, AVC

When y = o, equation (A) reduces toe2 gaa 1(22- VIM?'=81 a2 (1 a22)(1 al2 a32).

The optimum values for the dampingfactors may be found by differentiating withrespect to them and equating to zero and are

(a2) (I + a22)(I a12 a32) (81)

a2(a12)opt. = (I a32) + I + a22

(a22)opt. = I + ai2 a32

From equation (78), (a32)0t. canand is

a2

a2(a32),,pt. = (1 + a12) a22

Equation (82) cannot usually be satisfiedin practice when Ra > RD, for this wouldinvolve L3 > L1 and the self -capacity ofL3 would become important, so that the valveis usually connected across the whole of thefirst circuit together with the A.V.C. system.The detector also is usually connected acrossthe whole of the second tuned circuit as inFig. 17. al, a2, and a3 are then fixed andhave the values

a12 = 0.,2/1/12/RiRa = co2L12/R1RaI= RpilRa .. (85)

= w2M22/R 2R = w21., 22/R2 R,

=12,2112,, .. (86)a32 =w2m32/R1R-Ave = api,_ 2 IR 1'"R

etV0

(79)

(8o)

(82)

.. (83)

be found

.. (84)

=RDIIRAvo (87)

SS ENGINEER 185

Before proceeding further it is necessaryto consider the relative magnitudes of thedamping corresponding to the various factors.RD is usually of the order of 0.1 to 0.5 MD;R' is about 0.125 MD with a good qualitydiode detector, Ra is some I MD and RAveis a variable quantity. With the usualdelayed A.V.C. system, the A.V.C. diode isinoperative for signals below the delayvoltage, so that the input resistance of thediode itself is infinity. As the signal in-creases, the input resistance falls, tendingeventually to a value of one-half the D.C.load resistance. As regards the dampingon the coupling, however, the system as awhole is more complicated than this. Re-ferring to Fig. 19, it is easy to see that whenthe signal is smaller than the delay voltageE the input resistance RATro = r1 r2/(rid-r2)for the reactances of C1 and C2 are, orshould be, negligibly small. With a verylarge input, the diode resistance tends toy,/2, so that

RAVC = [3r lir 2] = r1r21(r1 373)

It might be thought that Rdre could bemade as high as desired merely by usinglarge enough values of ri and 73. This wasthe case in the past but valve makers arenow placing a limit to the maximum re-sistance which mustbe used in the gridcircuit of H.F. pentodetype valves. In thecase of the VMP4G,it is only 2.0 MS2.If one valve of thistype be used, there-fore, yi 72 must notexceed 2.0 MD forthey are in series withthe grid -cathode cir-cuit of all controlledvalves. If more thanone valve be controlled, r, + y2 = 2/m (meg-ohms) where m = number of valves. It isquite common for three valves to be con-trolled and then m = 3 and r1 r2 = o.6MD only.

With a small signal 124 is a maximumwhen r, = r2, but with a very strong signalit is a maximum for r, = 3r2. Thus withthree valves, 71 + Y2 = o.6 MS2, 71 = y2 =0.3 MS2 with a weak signal and RA =-0.16 MS2. With a strong signal, however,

Fig. lg.-The usualA.V.C. circuit in whichthe delay voltage isrepresented by a battery.

r86 THE WIRELESS ENGINEER April, 1936

r1 should equal 3r2 = 0.49 MS2, and thenra = 0.16 MS2 and R," = o.o8 MS2.

As distortion in the valve is most likelywhen it is giving its largest output, thecoupling should be designed for this con-dition and the lower value of RA,o used inthe calculations. It has already beenpointed out by the writer3 that this varia-

tion in R, can leadto distortion so that

01 ideally the A.V.C.diode should be fed

AVC with the aid of anextra amplifying valve

used only for thisC2 purpose. This is, how-

+ ever, too expensive asolution for most pur-poses even althoughit relieves the couplingof the damping of

the A.V.C. system as well as removing onepossibility of serious distortion.

It should be pointed out that a considerableimprovement in the A.V.C. arrangementsis possible if an H.F. choke is interposedbetween the diode and r, and r, as in Fig. 20.If the choke is a good one, RA" = co witha small signal and RA" = r1/2 with a largesignal. With ri+ r2= o.6 M12, ri can be0.5 MS2 with r, = 0.16 MS2 and then RAFT,has the minimum value of 0.25 MS2, aboutthree times as great as before. An alterna-tive arrangement giving the same result isshown in Fig. 21, but it necessitates anextra coupling coil to the tuned primaryand this will introduce losses.

Returning to the coupling, typical valuesfor damping are thus

R. = MS2R' = o.125 MS2RAVO = 0.25 MS2.

The primary circuit damping is due to R.and &VC in parallel and their combinedvalue is some 0.2 MD. Since it is of thesame order as R' it would be easy to equalisethem by modifying the value of ri and thiswould permit considerable simplificationof the equations. Although there is mathe-matically a temptation to do this, there seemsto be no electrical justification for it and the

CHOKE

Fig. 20 -A modifiedA.V.C. circuit having ahigher input resistance.

3 " The Design of A.V.C. Systems," by W. T.Cocking, The Wireless Engineer, August, Septemberand October, 1934.

case is more general if unequal loading ofthe circuits be permitted.

Having fixed convenient arbitrary valuesfor a1, a, and a3, the only remaining couplingfactor is the term a. It can be shown thatwhen a is small the resonance curve has asingle peak, but when it is large two peaks

occur. The frequencies corresponding tothese peaks can be found by differentiatingB (equation 77) with respect to y and equat-ing to zero. This process gives

212py

- i{(I + ae)2(i ae a32)2(2221Q11

Qe. . . . (88)

for the particular case when the primarydamping equals the secondary damping, i.e.,Q, = Q2 and a,' = a12 + a32, this simplifiesto

y = +\/[a3 - (a,' i)2]/Q . . (88a)

The values of y removed from resonancefor which the response is the same as that at

Fig. 21.-Ano her modification of the A.V.C.system giving a higher input resistance than

usual.

resonance may be calculated from (76) and(8o).

2a3422/Q, - {(I + a,2)2± (I + a12 + .232)24222/Q12d

YPB = ±4222

. . (89)

so thatnpg = npV 2 (90)

The formulae given enable the performanceunder any conditions to , be calculated andit is only necessary to deal here with one or

April, 1936 THE WIRELE

two special cases. The first of these is ofgreat importance when variable selectivityis required. It was pointed out in Part Ithat each pair of variably coupled circuitsshould have associated with it a singlecircuit having one-half the Q of those usedfor coupled pairs. It was proposed thatthis third circuit be coupled loosely to thevariably coupled pair so forming a three -circuit intervalve coupling which in itselfhas the required characteristics. Wherevery high selectivity is needed this is anadvantage since it permits a large numberof tuned circuits to be used with a smallnumber of valves. Where a lower standardof selectivity is required or where costdemands the use of fewer circuits, the possi-bility of using a total of only three coupledpairs was mentioned. With this arrange-ment two I.F. valves might be used andone two -circuit variably coupled transformeremployed for the coupling between the fre-quency -changer and the first I.F. valve witha second between the two I.F. valves. Itwas pointed out that to obtain the correctcharacteristics two single circuits havingone half the value of Q should be includedin the amplifier. Now it is possible by care-ful design to use a pair of loosely coupledcircuits between the second I.F. valve andthe detector for this purpose.

The resonance curve obtained from a pairof coupled circuits differs from the squareof that of a single circuit, but it closelyapproaches it if the coupling is made looseenough. It can be seen, therefore, that thecoupling must be loose, but, of course, nolooser than necessary for this would involvea loss of amplification. The effective Qof both circuits must be the same and equalto one-half the Q of the variably coupledcircuits. Consequently, if the coils have thesame Q the damping imposed on each mustbe the same.

Of the many possible methods of obtainingthe correct value of Q, the simplest seems tobe to make each circuit have the same valueof Q which, without external damping, isequal to the Q of the variably coupledcircuits. The damping on each circuit,therefore, must reduce Q to one-half. Nowin general the factors responsible for thedamping are unequal. On the primarycircuit the valve and A.V.C. system dampthe circuit, and on the secondary the

SS ENGINEER 287

detector. As already mentioned, in theprimary circuit it is customary to dispensewith the coupled arrangement of Fig. i8and to connect both valve and A.V.C. systemacross the tuned circuit so that

aiz = &I/R.a32= RDi/RA vc

Now R' is usually smaller than Ra and RAin parallel, so that if L1 = L2 and Q1 = Q2, sothat Ra, = RD2, R' cannot be connecteddirectly across the second tuned circuit forthen the second circuit would be more heavilydamped than the first. There are twopossibilities ; the first is to step-down toR' by adopting either the coupled arrange-ment of Fig. i8 or a tapping on the secondcircuit ; the second is to connect R' acrossthe whole circuit but to reduce L2 so thatRD2 = As regards normal operationthere is no difference between the twopossibilities, but when the load on the secondcircuit is a detector it is better to obtainmatching by adjusting the coil inductanceappropriately than by using a transformertype feed. The reason is that in this circuitthere is not only the intermediate frequencyto be considered but the harmonics of thiswhich are inevitably generated by the de-tector. If a coupled circuit or tapped coilbe used, the impedance presented by theoutput circuit to the detector at the har-monic frequencies is the reactance of L4 orthe reactance of the inductance between theend of the coil and the tapping point. Thisreactance is proportional to frequency sothat it becomes increasingly large as thedegree of harmonic goes up. If matchingis secured by adjusting L2, however, theimpedance presented to the detector is thereactance of the condenser C2, and as thisis inversely proportional to frequency it fallswith an increase in the degree of harmonic.Compared with the other arrangement, theimpedance at any harmonic is much lower.It makes for better detector action, andthere is less chance of feed -back of har-monics to earlier circuits, where they maycause serious trouble, if the output im-pedance of the coupling is low at the har-monic frequencies, so that it is obviouslydesirable to match by varying the coilinductance. Consequently

422 = Rid/R'


Now the total damping of the first circuits represented in the equation by (i a12

a32) and of the second circuit by (i a22).Equality of damping is thus obtained when

(I + ail + a32) = + a22) (9i)It is readily seen from equation (76) that

the a -terms represent the external dampingand that the numeral is due to the circuitresistances themselves. The condition in

1'0

0'1

0'01 -30 -20 -10 0 10 20 30 40

TWO CIR UrSSEPARATED BYA VALVE

2

a=1a -04 -

KILOCYCLES PER SECOND OFF RESONANCE

Fig. 22.-The resonance curves of a pair ofcoupled circuits with various degrees of couplingcompared with the curve (dotted line) of two

circuits separated by a valve.

which the external damping has the valuenecessary to reduce Q to Q12 is thus ob-tained by writing

i = a22 = (a12 a32) . . (92)

The insertion of these values in equation(76) gives, when Q1 = Q2 = Q,

e2 gaaiVRalee, ,V[(4-1- a2 y2Q2)2 16 y2Q9 (93)

which reduces toe2 = gaa,1 Rai?' when y = o . . (94)el 4 + a2

and (a)0.= 2For the particular case under considera-

tion the value of a must be seleCted not foramplification but for the correct resonancecurve. The curves of Fig. 22 have beencalculated from equation (93) and show theresults for three different values of a, com-pared with the response (dotted curve) of

two single circuits which is the requiredresult. As it was found in Part I that thecorrect results are secured with Q = 93.5this value has been used for the coupledcircuits and QT = 46.75 for the singlecircuits, since the damping factors for thecoupled system reduce the effective Q tothe required value of 46.75. It will be seenthat when a = 0.4 the resonance curve isalmost indistinguishable from the ideal.The efficiency, however, is rather low, forFig. 23 shows that a/(4 a2) = 0.095 onlyas compared with the value of 0.25 obtainedwith optimum coupling. The resonancecurve for a = 2, however, differs too widelyfrom the ideal for it to be permissible, butit should be possible to employ a = i with-out the divergencies being too great. Thefactor a/(4 a2) then has the value of 0.2.In other words, the efficiency is 8o per cent.of the maximum. This is quite good andamply justifies the choice of this value of aif the divergencies from the ideal in thematter of the resonance curve can be per-mitted.

The type of response obtained for thissystem is shown in Fig. 24 in which curve Arepresents the combination of two variably -

1'0

10

aFig. 23.-The variation of efficiency with

coupling for a pair of coupled circuits.

coupled two -circuit transformers of the typedescribed in Part I in conjunction with oneloosely coupled transformer having a = i.Curve B represents the ideal case of twoseparate circuits, instead of the looselycoupled pair, in conjunction with the sametwo variably -coupled transformers. The

0'011001

April, 1936 THE NN IRELE

difference between the curves illustrates theprice which must be paid for the convenienceof using a coupled pair instead of separatecircuits. It can be seen that with either

1,000

100

10

l'O

01-30 -20 -10 0 10 20 30

B

A

KILOCYCLES PER SECOND OFF RESONANCE

Fig. 24.-The combination of zwo variably -coupled transformers with two circuits of Q/2is shown here. Curve A represents the ideal casewhen the Q/2 circuits are separated by valves andcurve B that when they are loosely coupled

together.

arrangement the divergencies from a flatresponse in the pass region are about thesame (2 db.), but that with curve A thecentre -peak is so small as to be virtuallynon-existent. The outer peaks are promi-nent instead of being equal to the centre -peak. It is clear, however, that if responseup to ro,000 c/s only be required the couplingcould be reduced somewhat in the variably -coupled circuits and this would tend to makethe peaks in the pass region less prominent,and would give increased efficiency. Alter-natively the coupling of the fixed pair couldbe reduced and the curve would then approxi-mate more closely to (B) but at the expenseof amplification. In the writer's opinion,however, the response given by curve B isquite good enough for high -quality broadcastreception.

SS ENGINEER 189

The next step, therefore, is to develop thedesign equations for the I.F.-detector coup-ling, for the formulae governing the choiceof components in the variably -coupled trans-formers were given in Part I. In the lightof the results just discussed an arbitraryvalue can be fixed for a and a1, a2, and a3are also fixed. Thusa = = 0,2m2/R1R2

a22= a 12 a32 = I =RD2/R' =--- coL 2 QIR'

=RIRad-R,,1IRA=coL1Q(11Ra+TIRA)

L1 RR,L2 R'(Ra RAVC)

Q, R', 'Ra, RA, are known, therefore,L2 = R'/wQ (95)

L2 RaR,L1(96)R (Ra+R,)C2 = 1 iG02/.. 2 (97)C1 = 1/0.)2L1 (98)

R2 = wL2/Q (99)

R1 = coLlIQ (roo)

M = Vk1R210.) (ror)

From (94) e2/e1 = o.2g-1/RDIR' (r02)

For the case first discussed Q = 93.5,R' = 1.25 X 105,Q, Ra = 106D, RAyo = 2.5X 105,Q, f = 4.65 X 165 c/s. Therefore,L2 = 457.5 µH., L1 = 731 µH., C2 = 258/Ltd.., C1 = 161 1142E., R2 = 14.30; R1 =22.952, M = 6.2 ,uH.

In this section the properties of a singletuned circuit coupling between the last I.F.valve and the detector have been fullydiscussed for the cases of a triode and atetrode or pentode I.F. valve and a pair ofcoupled tuned circuits as the coupling hasbeen generally treated, special attentionbeing given to the case where the circuitsmust have particular characteristics to workin conjunction with two -circuit variable -selectivity I.F. transformers. The equationsgiven already enable the performance underany conditions to be calculated and it seemsunnecessary at this stage to treat any otherspecial cases, and this will be reserved forPart III which will deal with the design.ofthe I.F. amplifier as a whole.

THE WIRELESS ENGINEER April, 1936

The Anode to Accelerating Electrode Spacein Thermionic Valves*

By J. H. Owen Harries

PART I1. Introduction.IF the length from the anode to the cathode

in thermionic valves could be greatlyincreased without increasing the voltage

to current ratios above the values found inshort streams (e.g. in triodes and diodes),many advantages would result. Some ofthese advantages appear to have been fairlywell known for many years (Bib. Nos. i and2), but no method of gaining this end appearsto have been published. The advantagesmight be expected to include the prevention ofthe retrograde passage of secondary electronsfrom the anode, and a substantial reductionin interelectrode capacities. Of the mul-titudinous publications on electron jet fre-quency multipliers (e.g. Bib. Nos. 3, 4, 5, 6,7, 8, 9, ro and II) there are few indeed whichfail to call attention to the desirability ofusing long streams, so that the sensitivity todeflection may be reasonably great, but suchlong streams invariably possessed an im-practicably high voltage to current ratio.

2. The Two Principal Parts of a ValveValves of the kind considered in this

paper may be thought of as consisting oftwo parts, the cathode space, from thecathode to the first accelerating electrode,and the anode space, from the first accelera-ting electrode to the anode (Fig. I). Thephysical relationships of the cathode spacehave been analysed to a satisfactory degreeof accuracy (Bib. 12). The only informationthe author was able to find about the anodespace (where the electrons have an initialenergy which is as high or higher than thatcorresponding to the potential of the anode)is concerned with the production of jets ofelectrons, having very high voltage to cur-rent ratios, as in oscillographs ; and, withrespect to short stream tetrodes and pen-

* MS. accepted by the Editor, December, 1935.

todes, statements that the anode must beplaced close to the accelerating electrode toproduce a practicably large current, at asufficiently low voltage. The familiar dyna-tron characteristic will then appear due tothe retrograde passage of secondary electrons." The maximum cur-rent which can flowto the plate getsless as the distancebetween the grid andplate is increased,"is a typical quota-tion from conclu-sions, based uponthe classical solutionsof the basic differen-tial equations only,and arrived at from an analytical con-sideration of the anode space.

3. Method of Attack on the ProblemAn examination of the problem indicated

that there was little hope of arriving at thenecessary information about the anode spaceby endeavouring to obtain a complete solu-tion of the basic differential equations.(Bib. No. 12).

Mathematical analysis is a process of pro-ducing on paper, by a special notation, aworking model of a particular part of theuniverse which it is desired to study. Whena correctly carried out mathematical analysisfails, it is because insufficient data has beenprovided from which to build the model.It is then necessary to make a physical model,in the form of experimental apparatus, andto obtain the missing information from that.The difficulty with experimental apparatuslies in the fact that it is necessary that thephysical model be so designed that othereffects than those which it is actuallyrequired to study are negligible, and thatthose it is required to study are readilyobservable.

ANODE

ANODE SPACEACCELERATINGELECTRODECONTROL GRID

CATHODE SPACE

CATHODE

Fig. I.-The anode andcathode space of a valve.


4. Sliding Anode Tubes

Fig. 2 illustrates the type of valve finallyemployed as an experimental model toinvestigate the problem and to obtain thedata for mathematical analysis. The anodeis arranged to slide in guides, and its positionis adjusted by tilting and tapping the tube.

This rather crude method was the only onefound to be successful in practice.

The essential features of the sliding anodetubes are as follows :-

(a) A focused cathode of ample emission.(b) An accelerating electrode close enough

to the cathode to obtain a space current ofat least 5 to 10 mA. at not less than 200 voltsor so. (To complete the investigations slidinganode valves having more than one accelera-ting electrode are necessary.)

(c) A mesh formation of the acceleratingelectrode so that it does not intercept morethan a very small part of the total spacecurrent. (An important part of the work wasthe realisation of the fact that such meshelectrodes may be made to give entirelysatisfactory results in the production of anelectron jet. Prior " electron gun " accele-rating electrodes are no good for solving theproblem. Their focusing action is un-necessary. They intercept almost all thespace current, or only produce a very lowspace current, and therefore make the anodecurrent too small to be of use.)

(d) An anode which is readily andaccurately adjustable in position from about0.25 cm. to 7 cm. from the nearest accele-rating electrode. The anode must be ofreasonable area, and have a positive electricalcontact to the external circuits.

5. Electrical Characteristics of Sliding AnodeTube

Fig. 3 shows the anode current/anodevoltage characteristics for various distances,in centimetres, between the anode andaccelerating electrode of a typical slidinganode tube.

At short distances the familiar dynatron

characteristic appears. As the distance isincreased, the anode voltage Eb at whichsubstantial saturation of the anode currentoccurs shifts to the left until it reaches aminimum, and the dynatron characteristicdisappears. As the distance is still furtherincreased, Eb becomes greater once more.That the retrograde passage of secondary

Fig. 2.-A type of movableanode discharge tube usedexperimentally to determinethe characteristics of the anode

space.

radiation disappears at extreme distances isnot surprising, but the fact that the primarysaturated anode current is independent ofthe distance, and that the saturation voltageis very low at not very long distances im-mediately outside the dynatron distance, isremarkable. At extremely long distances apeculiar phenomenon appears. Almost noanode current (except that due to strayunfocused electrons) flows until the anodevoltage has risen considerably above zero.At a certain anode voltage Eb, however, theanode current suddenly rises to a saturatedvalue, and remains at substantially thisvalue independently of how much furtherthe anode voltage is raised. The acceleratinggrid current has a characteristic which is thereverse of this, i.e. virtually all the spacecurrent goes to the accelerating grid untilE b is reached, when it drops to a steady value,which is about one -sixth of the anode currentin the case of the tube illustrated in Fig. 2.

E

Eacc.:150V. 0'5

100 200 300 400 500

Ea

Fig. 3.-Anode voltage/anode current charac-teristics of the valve illustrated in Fig. 2.

An important characteristic of the tubeis the " distance curve," connecting thedistance d, between the anode and theaccelerating electrode, and the anode

C 2


saturation voltage Eb. It is illustrated inFig. 4.

Further measurements show the effectupon the position of the minimum in the

distance curve of300 changing the acceler-

ating voltage. The200 distance curve remains

about the same, as re->

w ioo gards the position ofthe minimum, becauselowering the voltage ofa single acceleratingelectrode decreasesthe current and thepotential gradientsimultaneously.

The anode distanceat which the distance

curve is at a minimum has been christenedthe " critical anode distance " (Bib. No. 13).The trough, of the distance curve is quiteflat and if an anode is placed at the preciseminimum, accidental or manufacturing vari-ations in its position will make little differenceto the characteristics. A tetrode valve madewith its anode at this critical distance has acharacteristic curve free of the effects of theretrograde passage of secondary radiation.Provided that the accelerating electrodesubstantially shields the anode from thecathode space, then at low screen voltagesthe critical -distance valve has an extremelyhigh anode differential resistance Ra, i.e. theslopes of the Ealia curves are very small. Thefact that there is always a certain amountof slope is because it is impossible completelyto separate the cathode and anode spaces,and principally because secondary radiationfrom the surface of the accelerating electrodeadds to the anode current. This latter ad -

2 4 6 8

ANODE DISTANCE(0MS)

Fig. 4.-The distancecurve. The critical dis-tance is that at which thecurve is at its minimum.

ANODE

2,,c1 ACCELERATINGELECTRODE

1st ACCELERATINGELECTRODE

CONTROL GRID

CATHODE

Fig. 5.-A long -stream critical -distance valve.

dition is a constant fraction of the anodecurrent. Therefore, the slope increases withanode current.

A wide choice of forms of chafacteristic,particularly as regards the anode differentialresistance, is found according to whetherthe anode is at one side or the other of theminimum of the distance curve. To the lefta low anode differential resistance is pro-duced, and to the right a higher value.

6. A " Long -Stream " ValveThe original problem was solved by the

results of the above experiments. For in-stance, Fig. 5 shows a long -stream valvemA180

120

80

is40

E 0

Eg 11

Eg = -22

0 50 100 150 200 250

EaFig. 6.-Electrical characteristics of the type of

valve illustrated in Fig. 5.

diagrammatically. The control grid andcathode are close together. The firstaccelerating electrode is near to these two.The second accelerating electrode is spacedfrom the first by a distance of about 7 cm.,which is just not too long to cause the streamacross the gap to lose saturation at theworking accelerating electrode potential.This space will be operating under theconditions of the curve 6 in Fig. 3. Thedistance between the second acceleratingelectrode and the anode will be the criticaldistance, of, say, 2 cm. The anode .voltage/anode current characteristics of acritical distance type of valve are shown inFig. 6, and they will be recognised as satis-factory from the power efficiency andamplification standpoint. The breakdownvoltage at the critical anode distance is less,for instance, than the corresponding satura-tion voltage of an ordinary short -streampentode.

Of course, additional accelerating electrodes

300


may be used in this manner to lengthen thestream still further. These, being wellspaced from the preceding electrode, will befound to intercept very little of the totalspace current.

Thus it is possible to design a type of long -stream valve which has characteristics theexcellence of which, from a power handlingor amplification standpoint, is unaffected byextreme length between cathode and anode.

7. AnalysisHaving obtained these results from ex-

perimental models, it was possible to producean analytical interpretation of thephenomena.

It may be shown that both the cathodeand anode spaces may respectively berepresented by a diagram such as Fig. 7,where electrons are emitted from a cathodeK and tend therefore to travel to a positiveanode A across a distance x. The plane ofthe accelerating electrode is considered as acathode with respect to the anode, the poten-tial difference between them being the differ-ence between the accelerating voltage Eace andthe anode voltage Ea. The only differencebetween the anode and cathode spaces, fromthe analytical standpoint, is the differencein the values and distributions of initialvelocities of the emitted electrons.

K Fig. 8 illustrates thegeneral form of the

EMITTEDELECTRONS

A

potential gradientsbetween any cathodeand any anode in avacuum. OE repre-sents the plane of thecathode and DC repre-sents the plane of theanode. V is the

Fig. 7.-The cathode and potential of a pointanode of a high vacuum x distance from the

discharge tube. cathode. The magni-tude of the initial velocities may berepresented on the same scale from 0downwards.

Applying this diagram to the cathodespace we may indicate the maximum emissionvelocity by v0 and assume that a continuousspectrum of initial velocities extended fromv0 to zero, i.e. over the interval v0 to 0 inFig. 8. Only those electrons, the initialenergy of which is greater than the negativepotential dip AB or A'B' or A"B" (which

193

decreases as the anode voltage is raised fromEa,. through E.2 to Ea3) can pass to theanode.

B

Fig. 8. --The potential gradients and conditionsbetween the .cathode and anode illustrated in

Fig. 7.

The basic differential equation to Figs. 7and 8 is:

V d2V 2(V02

2e

m dx2= 67r2i2 . . (I))There is no complete general solution to thisas far as the author is aware. A specificsolution to the case just explained, where theinitial energies extend over a continuousspectrum the maximum value of which isvery small compared with the potential ofthe anode, is the familiar Langmuir 3/2 law.The conditions corresponding to the anodespace cannot, however, have initial velocitieswhich fulfil these conditions. If the accelerat-ing electrode were a perfect uni-potentialplane, then all the electrons issuing into the

194 THE WIRELES

anode space would have one single velocity.A consideration of Fig. 8 will show that the

characteristic corresponding to thesingle velocity case is illustrated in Fig. 9,curve GADE.

Because the accelerating electrode con-sists of spaced wires which are in theneighbourhood of a fairly dense spacecharge, and quite possibly also in theneighbourhood of earthed or negativelycharged metal, the potential of the equivalentcathode falls between the wires, and theinitial energies will therefore vary from avalue tending to equal the potential of theaccelerating electrode, down to a minimum.The resulting emission spectrum extendingfrom vo a short distance towards zero isindicated by kv 0' in Fig. 8. The rise inanode current at saturation point will nolonger be infinitely steep. The result isshown by curve GABE in Fig. 9.

(Later in this paper it will be shown thatthe production of as sharp a knee as possibleto the characteristics in the anode space isdesirable in the design of valves. To attainthis sharp knee, it is necessary to have asnearly as possible a single value of initialenergy, and therefore to have as littlereduction in potential as possible in theneighbourhood of the positive grid immediate-ly preceding the anode.

The above reasoning has neglected theeffect of diffusion due to uneven emissionfrom the accelerating electrode. It will beequivalent to having a number of ideal singlevelocity valves in parallel having slightlydifferent initial energies of emission. Theresult will be to round the curvesat A and B in Fig. 9. Such diffusionmay be neglected without much error whenthe anode distance is considerable, becausethe small change of potential gradient dueto diffusion will be negligible compared withthe large negative dip produced due to thespace charge and the long distance.

In Fig. 9 the dotted curve FHDE repre-sents the result to be expected at a shortanode distance in the absence of diffusion(and secondary radiation), and with aninitial velocity vo of one value only. In thepresence of varying initial velocities and ofdiffusions, the curve will be changed to dottedcurve GIDE.

These results take no account of theemission of secondary electrons. The

S ENGINEER April, 1936

similarity of the general form of curve GABEwith curve 6, Fig. 3, can, however, be noted.The curve marked for 3 cm. distance inFig. 3 will be seen to approximate veryclosely to that in Fig. 9 at.GIDE.

To obtain a complete solution of (I) isnot of great practical importance. Partialsolutions showing the positions of thepotential minimum and so relating is and V,etc., may be found. The saturation currentio is equal to the current to the acceleratingelectrode flowing in the cathode space, lessthat portion intercepted by the acceleratingelectrode. Thus 10 is entirely independentof anode distance d or anode potentialabove saturation. This is an importantconfirmation of the practical results of Fig. 3.(If the screening effect of the acceleratingelectrode is small the current will be added

H

G F

D B

Ea AV

Eat

Ea3

EaFig. 9.-Theoretically deduced anode voltage/anode current characteristics for the anode space.

to by the anode field extending to thecathode space, but this effect is subsidiaryand usually negligible.)8. The Effect of Secondary Electron Emission

Consider the effect of the emission ofsecondary electrons from any metal. (Bib. 14,15, i6 and i7). Fig. io is reproduced fromFarnsworth (Bib. Nos. 14, 15, i6) and showsthat the bulk of secondary electron emissionconsists of a spectrum which is almostentirely of much lower velocity electronsthan the velocity of impact giving rise tothe emission. (Emission and reflection areconsidered as the same thing for the purposeof this analysis.) The primary velocity ofimpact is indicated at vp. /8/4 is the ratiobetween secondary and primary electrons.v, is the velocity of secondary emission.


Consider first the effect of secondaryemission from the accelerating electrodeon the characteristics of the anodespace. The result will be that secondaryelectrons will enter the anode space, and willhave a spectrum of initial energy extendingfrom a value tending to be equal to that ofthe potential of the accelerating electrode,and continuing down to zero. The resultmay be referred to Fig. 8 by imagining thatthe initial spectrum of width ha' extendsfrom v, down to zero. Fig. II OGE andBCD shows GIDE and GABE Fig. 8 redrawnto suit these circumstances. The slope ofthe characteristics above the saturationpoint has been increased and the secondaryelectrons of low velocity have rounded theknees of both the characteristics for short andfor long distances. This makes thesecharacteristics a closer approximation tothose experimentally determined andshown in Fig. 3. Stray electrons anddiffusion will tend to produce the departurefrom zero current shown by the line OE.

The presence of secondary radiation fromthe accelerating electrode constitutes a limita-tion in the design of screened valves. If a veryeffective positive electrostatic screen grid,having a close mesh, is used, the anodedifferential A.C. resistance Ra will fall andthe knees of the Ealia curves become roundedto an undesirable extent. This is the reason

1

Tp

CE5

Vp

24 48 72 98 120 VOLTSvs

Fig. ro.-Velocity spectrum of emission ofsecondary electrons from nickel. The primaryvelocity of impact is indicated at v5. I,/I9 isthe ratio between secondary and primary electrons.

v, is the velocity of secondary emission.

why many short -spaced screened tetrodeshave a very low value of Ra under workingconditions. As secondary radiation is drawn

SS ENGINEER 195

from the positive grid, the total positive gridcurrent is reduced. This result will beat the expense of the anode differentialresistance, which will fall. This agrees withcommon experience in the design of screenedtetrodes of the dynatron type in which theanode field is extremely strong. The reduc-tion in Ra under these circumstances issometimes accepted as a penalty inseparablefrom adequate screening between the anodeand cathode spaces in such tetrodes. A very

to

0

------------------

Ea

Fig. IL-Theoretically deduced anode voltage/anode current curves showing the effects ofsecondary radiation from an accelerating electrode

to an anode.

good positive electrostatic screen will have alarge number of positive wires. Unless agreat deal of secondary radiation is drawnfrom this positive grid, the screen currentwill be impracticably high ; whereas bydrawing sufficient secondary electrons awayit may actually be made negative.

The emission of secondary electrons fromthe anode, as well as from the acceleratingelectrode, tends to reduce the value of Vin the neighbourhood of the anode andalters the potential gradient from thetheoretical shape in the absence of secondaryradiation. At values of Ea greater than theaccelerating voltage Ea, the only result ofthis is to reduce the effective anode potentialacting in the plane of the acceleratingelectrode or equivalent cathode. ProvidedEa is sufficiently great for the secondaryemission to be produced from the anodewhen the anode potential is less than thatof ,the accelerating electrode, then, when itis less, the emission will tend to travel backto the accelerating electrode, and will do sounless the negative space charge potentialdip between the two electrodes is sufficientto prevent this.


(9) The Theory of the Critical DistanceThe minimum found in the distance curve

(Fig. 4) may be explained as follows. Sincethe velocities of almost all secondary elec-trons are less than the velocity correspondingto the anode potential, and therefore lessunder these conditions than the velocity ofthe primary electrons issuing from the planeof the accelerating electrode, it is possibleto find a value of potential gradient sufficientto prevent the retrograde passage of second-ary electrons when Ea is less than Ea, butwhich is not sufficiently great at anodevoltages in excess of a small value, to preventthe passage of primary electrons. From thepreceding analysis, the negative potentialgradient in a given tube is proportional tothe spacing between the anode and accele-rating electrode. Therefore, this explainsthe fact that the experiments with a slidinganode tube resulted in finding a " criticalanode distance " at which the retrogradepassage of secondary radiation is preventedand yet, at this distance, anode currentsaturation occurs at an anode voltage Ebconsiderably less than that of the acceleratingvoltage.

It may be expected that the distributionof space charge will be affected bythe relative configurations of the accele-rating electrode (or equivalent cathode) andanode, and by the degree or otherwise offocusing of the electron stream, due to theeffect of this upon current density and upondiffusion. This is actually found to be thecase. For instance, in the sliding anode

Fig. iz.-A sectional viewof a typical commercialHarries critical - distancevalve. The shaded areaindicates the approximate

path of the discharge.

valve illustrated in Fig. 2 the critical anodedistance is about 3 cm. If an ordinaryfixed anode valve is constructed having atubular section anode, a circular cross-section positive grid and a filament cathodewithin the latter, a very diffuse radiallydirected stream will be produced in whichthe current density is low. The criticalanode diameter, will then be of the order of,say, 5 to 6 cm. or more. For commercial

April, 1936

valves, it is usually necessary so to focusthe stream into a jet that the critical anodediameter is very much smaller than this, soas to go into a reasonably compact bulb.Details of the methods of design in thisrespect are beyond the scope of this paper.Wide bulbs with very large openings toadmit a short electrode assembly of con-siderable diameter are extremely difficult toseal in by machinery.

Fig. 12 shows a section of a typical tetrodehaving a focused stream and its anode atthe critical anode distance. The stream isconfined to the shaded path. Practically thesame result is obtained in the absence ofthose portions of the circular anode fromA to B and from C to D.

PART II1. Ideal Valve Characteristics

Fig. 13 shows the anode voltage/anodecurrent characteristics of an ideal valve.It may be shown that Eb should be as smallas possible, and that the slope of the lines

Ea

Fig. 13.-Theoreticatly perfect anode voltage(anode current characteristics of a multi -grid

valve.

from Eb upwards should be constant. Themore linear the curves are to the right ofthe knee and the lower the value of E b, thelower the distortion.

2. Design of the Anode Space and Examplesof Valves

Referring back to Figs. 9 and II in theprevious part of this paper, it will be seenthat the characteristics illustrated in Fig. 13correspond to those referring to valves inwhich secondary radiation in both directionshas been rendered as low as possible, and thescreening of the anode from the cathodespace is as complete as possible. The effects


of diffusion should also be avoided as far aspossible by widely spacing the anode, andyet not spacing it so, far as to cause Eb tobecome greater than is desirable in any givencircumstances.

SLATS

CATHODE

(a) (b)Fig. 14.-(a) A type of anode design intended toprevent the emission of secondary electrons fromthe anode into the anode space. (b) The anodevoltage/anode current characteristics of this type

of valve.

There are three possible methods of anodespace design, as regards that part of theproblem having to do with preventing theretrograde passage of secondary radiation :-

I. The prevention of emission of secondaryelectrons from the anode surface.

2. The prevention of passage of secondaryradiation across the anode space.

3. A combination of the above.So far as the author is aware T. has

never been successful. 2. is very effective,and is exemplified both in the critical -distance valve and in the pentode valve.3. is also effective, particularly with thecritical -distance valve, though usually itis not worth the trouble, because excellentcharacteristics are obtained by 2. without it.

The prevention of the emission of secondaryelectrons (method 1) was tried by Hull by

80

60

£. 40a

20

0WO 200

Ea= 250V.

DYNAMICCURVE

300 -30 -20 -10

80

60

o

20

0

Ea Eg

Fig. 15.-Typical characteristics of a poweroutput receiving valve of the pentode type. Notecurved top of the dynamic characteristic pro-

ducing distortion of a complex input wave.

E

a

SS ENGINEER 597

chemical treatment of the anode surface(Bib. No. ii). It was only partiallysuccessful. Another suggestion which hasbeen carried out in practice is to arrangethe anode to have radially directed slats orholes therein. It is intended that primaryelectrons will travel into the space betweenthe slats, and that secondary electrons willthen not readily leave the space between theslats and travel back to the acceleratingelectrode. Figs. 14a and 14b show, respec-tively, the section of a slatted anode, andthe type of characteristic usually found withthis kind of valve. It is not very satisfac-tory, because it appears to be quiteimpossible completely to trap the electronsin the spaces between the slats.

Fig. 15 shows the characteristics of apentode. These have a rounded kneecompared with Fig. 13. The reason for this isthat the anode of the pentode is immediatelypreceded by an earthed grid. Electrons inthe neighbourhood of the wires of this gridare reduced in velocity, whilst in between the

Egg -250V.80

60

E

20

40

0

Ea -250V.80

Gni =4-8 mA 80

DYNAMICCURVE 40

20

0100 200 300 -30 -20 -10

Ea Eg

Fig. i6.-Characteristics of commercial Harriesmains type critical -distance valve. Note dynamic

characteristic.

wires the velocity of the stream tends torise. Thus there is a wide spectrum ofinitial velocity from zero upwards. Theanode is close to the equivalent cathode,causing diffusion effects to become important.

Fig. 16 shows the characteristics of aspaced anode power output valve. Thisvalve has been described elsewhere(Bib. i8). It is found to give a substantiallylower distortion level, and greater poweroutput, than the equivalent pentode ofFig. 15. The focusing in this valve has beenarranged so that the anode will go inside areasonable size of bulb. Everything possibleis done to remove negative or earthed metalfrom the neighbourhood of the anode space,and thus, together with the longer anodespace, because of the comparatively small

198 THE W1RELE

width of initial energy spectrum, results inthe sharp knee and linear characteristics,shown.

The rounded knees and S shaped dynamiccharacteristic of the pentode (Fig. 15) is thecause of a very serious type of distortion.This distortion is not indicated by, or pro-portional to, the amplitudes of second andthird harmonic, neglecting phase angle,obtained by conventional distortion measure-ments with a sine wave input. It may beevaluated by measurements on harmonicdistortion using an input consisting of twosimultaneously applied waves, simulating theactual conditions of telephony reception.This distortion is absent in the case of thecharacteristic of Fig. i6. Oscillographicrecords of complex wave distortion havebeen published (Bib. No. DO.

During the past few years the author hasinvestigated the effect of the anode criticaldistance on a large number of control -gridtype valves. It is not possible within thespace of this paper to give more than twoexamples, which are of interest as illustratingmethod 3 of designing the anode space. InFig. 17 two separate anodes are employed,one on each side of the positive grid, and arespaced from it at the critical distance. Theresemblance to Fig. 12 if the portions of theanode from A to B and from C to D areremoved will be noted. The earthed platesE1, E, are too far outside the anode fieldsubstantially to affect the passage of

EARTHED PLATE

ANODE ANODE

--POSITIVEGRID

CATHODE

secondary radiation. The field of theseplates is very slight and similar inits action to a suppressor grid. The

Fig. 17.-An alternative anode design Joy acritical distance valve using earthed side plates.

EARTHED PLATE

CONTROLGRID


result is that the knee of the characteristicsare in general not quite so sharp as those ofa pure spaced anode valve ; but the criticaldistance effect still appears, and the anode isat the critical distance. It is of interestto observe that whenever there is sufficientmetal in the accelerating electrode to give asubstantially positive field and a " singlevelocity " emission " sharp knee " effect,then secondary radiation will travel back tothe accelerating electrode at small anodespacings and the typical distance curveminimum, or " anode critical -distance effect,"will appear.

If a slatted anode is employed at theanode critical distance, the unsatisfactory

FILAMENTACCELERATINGELECTRODE

CONTROLGRID

ANODE

1 '5mm SPACING BETWEENCONTROL GRID &ACCELERATING ELECTRODE

Fig. i8.-71lustration of a type of low capacitycritical distance receiving valve. The capacitycan be reduced further by taking the grid lead out

at the side of the bulb.

nature of the characteristic (Fig. i4b) atthe shorter distance disappears. Providedthe slats are fairly deep it is largelyimmaterial whether the back portion of theanode is present or not.

Extremely low capacity valves may beproduced by taking advantage of the longspacings possible. It is not difficult, by meansof conventional screened grid constructions,only to screen one electrOde, such as theanode, from the control grid ; but electro-static screens, when this is done, increase thecapacities to earth. By means of such aconstruction as that illustrated in Fig. i8 ashort wave valve is produced having ex-tremely low capacities, not only from anodeto control grid, but also from these electrodesto earth.3. Screened Voltage Amplifier Valves

As previously pointed out, ordinarydynatron tetrodes, when well screened to givean anode to control grid capacity as low as0.002 to o.00ipiLF, have undesirably low A.C.anode resistances and curved characteristics.A way of overcoming this is by the introduc-tion of a suppressor grid, but the reduction


in anode field then causes the screen currentto become impracticably great unless thescreening is also reduced. Thus, screened pen-todes have anode to control grid capacitiesnever less than o.004p0F and usually of abouttwice or more this value. On the other hand,their anode A.C. resistances are satisfactorilyhigh, namely, about i,000,000 ohms at 7 mAanode current. A screened critical -distancetype valve has been produced with a closemesh screen, having a screen current of thenormal value, an anode A.C. resistance of1,000,000 ohms, and an anode to control gridcapacity of as low as 0.001

4. ConclusionsFrom the theoretical considerations out-

lined in Part I, and from comparativeexperiments made with some thousands ofvalves over a period of several years (a fewtypical results of which are set out in Part II),it is considered that the correct design of theanode space of a multi -grid valve shouldutilise the anode critical distance and thatthe design of very long stream valves shouldbe in accordance with the methods of spacingaccelerating electrodes described herein.

BIBLIOGRAPHY1. Gill, Philosophical Magazine, May, 1925,

PP 993-1995-2. Tellegen, British Patent, No. 287,985**, 1926.3. Robert Von Lieben, German Patent, No.

179,807.4. Dieckmann & Glage, German Patent, No.

184,710, 1906.5. Rene Zei, British Patent, No. 18,438, 1913.6. Nasarischwily, Ann. de Phys., 1921, Vol. 64,

P 759.7. Gebbert, Jahrboch derdrahtlosen Telegraphic,

1923, VOL 22, p. 107.8. Dieckmann & Glage, Jahrboch derdrahtlosen

Telegraphie, 1922, Vol. 19, p. 194.9. Harries, British Patent, No. 328,680, 1929.

io. Alfven, Zeitschr. f. nochf. Tech., July, 1934,Vol. 38, pp. 27-29.

11. Dieckmann & Gebbert, German Patent, No.375,808, Toth July, 1921.

12. J. J. Thomson, Conduction of ElectricityThrough Gases, Third Edition, 1928, p. 373.

13. Harries, British Patents, 380,429, 385,968, 1931.14. Farnsworth, Physical Review, 1922, VOL 20,

P. 358.15. Farnsworth, Physical Review, 1925, VOL 25,

P. 41.16. Farnsworth, Physical Review, 1926, Vol. 27,

P. 413.17. Hull & Williams, Physical Review, 1926,

Vol. 27, p. 432.18. Harries, Wireless World, 2nd August, 1935.19. Bartlett, Wireless Engineer, Feruary, 1935,

Vol. 12, pp. 70-74.

Osram Valvefor Microphone Amplifiers

THE new Osram MH4o valve is similar to thewell-known MH4 type, but has severalfeatures which particularly adapt it to the

early stages of a microphone amplifier. It is anindirectly heated valve with a heater consuming

r.o ampere at 4.0volts, and it isfitted with a stand-ard 5 -pin base.An automatic biasresistance of r,000ohms is recom-mended, and withits maximum anodepotential of zoovolts, the anodecurrent is 2.7 mA.The optimum loadresistance is 50,000ohms and the valvehas a mutual con-ductance of 2.4mA/V., with anA.C. resistance ofi6,7oo ohms. Thesefigures are, ofcourse, for 100volts anode poten-tial and zero gridbias. The inter -electrode capacitiesare given as : grid -anode 7.3 12,µF.,grid - other elec-trodes 6.o m.LF.,

and anode -other electrod3s 4.0The particular features claimed for the valve, and

the ones which make it so suitable for the earlystages of a high -gain amplifier, are the very lowdegree of microphony and the high insulation of theelectrodes. This is accomplished by the use ofsteatite insulators for the electrode spacers insteadof the usual mica separators.

The valve is priced at 5os. and the makers areThe General Electric Co., Ltd., of Magnet House,Kingsway, London, W.C.2.

Book ReviewsThe Radio Amateur's Handbook, 1938 edition

By the A.R.R.L. Headquarters Staff. 48o pages(including a 96 -page catalogue section) approxi-mately 500 diagrams, illustrations and charts.Published by The American Radio Relay League,West Hartford, Conn. U.S.A. Price $1.15 withpaper cover, or $2.50 with linen cover, post free.

Probably no reference book contains such a wealthof practical data for the short -waves amateur experi-menter as the 1936 edition of The Radio Amateur'sHandbook. It has been revised and brought up-to-date and forms a worthy successor to the twelveprevious editions. Of the several new chapters

200 THE WIRELESS ENGINEER

added, the 3o pages of tabulated matter givingcharacteristics, operating data, and base con-nections of practically every type of Americanvalve likely to be used in amateur short-wavereceivers and transmitters, is a most valuablefeature, and of considerable interest also are the newsections on the ultra -short waves which deal withtransmission and reception on frequencies of 224Mc/s, 112 Mc/s, and 56 Mc/s.

The 1936 edition contains no fewer than 21chapters, and more than half are devoted to essen-tially practical matter, such as the design andconstruction of receivers, transmitters, monitoringapparatus, aerial arrangements, and the layoutof a complete station. There is, also, an extensiveappendix of miscellaneous information.-H B. D.

Elementi di Radiotecnica GeneraleBy Cesare Rimini. Pp. 566+ xx, 383 Figures.

Published by Nicola Zarichelli, Bologna.This is a book eminently suitable for students in

Universities and Technical Colleges. It has asatisfactory air of carefulness and thoroughness,the diagrams are very good, the formulae are setout in a way that will appeal to the teacher andstudent alike. The author is in the Radio Com-munication Department of the University ofBologna, and is evidently experienced in thecareful presentation of radio problems to students.The thirteen chapters and an appendix cover thewhole field of radio communication. The first sixchapters deal with electric circuits, their char-acteristics, etc., including filters and allied problems;the remaining chapters deal with valves, theirtheory and applications, high frequency generators,the radiation and reception of electromagneticwaves, etc. The last forty pages are devoted to amathematical appendix dealing with differentialequations, Fourier's theorem, etc.

The book can be heartily recommended to anyonewho can read Italian.-G.W.O.H.

CorrespondenceJohnson Noise

To the Editor, The Wireless EngineerSIR,-Briefly-I feel that Mr. Robin's experience

is bound up with a thermal variation in the wireof the line (or elsewhere in the circuit). If so itshould easily be demonstrated.

Reference to Planch's radiation function (page.44-45, Tables of Functions, Johnson and Emden)would indicate that " Whereas a body of absolutetemperature emits electromagnetic waves pf allpossible wavelengths, the radiated energy is veryunequally distributed " from which I assume thata variation in temperature would alter the distribu-tion. In the particular case reported this variationmight account for a peak 5 times per second at allthe frequencies examined.

Ware, Herts. GERALD SAYERS.

April, 1936

New Ediswan Output ValveANEW valve de-

signed to give alarge output is being

produced by the EdisonSwan Electric Co., Ltd. ;it is the ES ioo. It hasa filament rated for6 volts at 3 amperes, andin normal operation itshould be A.C. heated ;if a D.C. filament supplybe used, however, theanode return leads shouldbe joined to the positivefilament terminal. Theinternal A.C. resistanceis 1,75o ohms, and theoptimum load impedanceis some 7,000 ohms. Withits rated anode potentialof L000 volts, the anodedissipation is loo wattsand the valve can deliveran output of 3o watts.It has a special 4 -pinbase and is priced atio guineas.

The Industry]MEASURING and Testing Instruments for

the Electrical and Allied Industries " is thetitle of a well -prepared loose-leaf catalogue

issued by the Instrument Department of E. K. Cole,Ltd., Southend-on-Sea. The apparatus introducedso far is all of direct application to radio researchand production ; a standard signal generator, beatfrequency oscillator, variable attenuator, inductancebridge, and an A.F. oscillator are included in theseries. As befits a catalogue of this kind, ampledata for the benefit of engineers is given.

Doolittle & Falknor, Inc., 1306, West 74th Street,Chicago, Illinois, U.S.A., have sent us leafletsdescribing concentric transmission cable designedfor carrying radio frequency power over con-siderable distances. The screened cable is weather-proof, and in the smaller sizes is quite flexible. Afilling of dry nitrogen at slightly above atmosphericpressure is recommended for outdoor installationsas a precaution against condensation. The loss in thei-inch cable is given as o.3 db per L000 feet atL000 kc/s.

The M.S.S. Recording Co., Ltd., 99a, CharingCross Road, London, W.C.2, conducts a servicewhereby B.B.C. transmissions may be recorded ondiscs for the exclusive use of speakers, artistes, etc.,responsible for the item recorded.

Apiil, 1936 THE WIRELESS ENGINEER 201

Voltage Measurements at Very HighFrequencies-II. (Concluded)

By E. C. S. Megaw, B.Sc., D.I.C.

(From the Rebearch Staff of the M.O. Valve Company, Limited, Wembley)

III. Comparison of Peak Voltmeter with aThermal Measurement of Measurement.

(a) Method.Advantage was taken of the special form

of resistance, used as transmission linetermination in the experiments of Section II,to check the " true voltage " derived fromthe peak voltmeter readings against anindependent determination. The arrange-ment of the apparatus is shown diagram-matically in Fig. 17, which should becompared with the photograph in Fig. 1o.The resistance consisted of a film of colloidal

was mounted in a brass screening tubeforming a short length of concentric line ofvery high damping closed at the end remotefrom the oscillator. Calculation shows thatthe skin effect is negligible at frequencies ofthe order of ioo Mc. The input impedanceat the point A could therefore be determinedfrom the measured resistance (at 5o cycles)and the calculated inductance and capacit-ance per unit length, since

Zi= Z, tanh ylwhere 1= length of resistance (= 18.1 cm.),

y = VZY, Zo = -VZ/Yand Z =R, L, and

HEADINDICATOR AI

CHECK

RC

jwL, Y = jwC,being, the resistance, inductance,

and capacitance percm.

The H.F. voltageINLET at the point A wasTHERMOMETER

DOUBLEVOLTMETER

1/determined in thefollowing way.With the 50 -cycle

H.F.

TRANSMISSIONLINE

P

0--0

220 V.50

TERMINATINGRESISTANCE

TESTVOLTMETER

graphite* of the order of o.i mm. thick(resistance about 58 ohms) on the outsideof a thin -walled silica tube, the inside ofwhich was water cooled. t This resistance,which is capable of dissipating over I kW.,

* See H. Higinbotham, Wireless Engineer, 12,

373 (July, 1935)t I am indebted to Mr. B. S. Gossling for suggest-

ing this form of resistance.

WATERINLET

OUTLETTHERMOMETE1

OUTLET

Fig. 17.-Arrangementof apparatus for com-parison between peakvoltmeter and thermalmethod of voltagemeasurement.

plug P removed from the voltmeterhead, the oscillator was switched on andadjusted to give a standard reading on thecheck voltmeter (about 185 v. peak). Thecooling water flow was maintained constantby means of the head indicator and inlettap, and the temperature rise of the waterwas measured after thermal equilibrium hadbeen reached. The test voltmeter readingwas also noted. The oscillator was thenswitched off and the 50 -cycle plug connectedto the line. Blocking condensers at the


oscillator end of the line enabled this to bedone without any alteration to the circuit.The value of the resistance and the tempera-ture rise were then determined for several50 -cycle voltages in the neighbourhood of130 v. R.M.S. Interpolation gave the voltageand resistance corresponding to the tempera-ture rise obtained on H.F., the water. flowbeing constant throughout. The inputimpedance was then calculated for theappropriate values of resistance and fre-quency, giving a result of . the form Z,= A + jB ohms. Now if ki = Zi IR, andk, = AIR,, where Ro is the total resistanceat 5o cycles, it can easily be shown that theratio of H.F. voltage at A to 50 -cycle voltagefor the same temperature rise is h1/A/4

Owing to imperfect matching the voltageat B, the measuring point in the doublevoltmeter head, was slightly different fromthat at A, the beginning of the graphiteresistance. The ratio of the two wasobtained by transferring the test voltmeterfrom A to its normal position B. Since thisratio was not greatly different from unity(about 10 per cent. at zoo Mc.) it could bedetermined with negligible error by usinga diode having small inter -electrode clearancein the test meter. The H.F. voltage at Bis thus determined in terms of the 50 -cyclevoltage.

Thorough precautions were taken to ensureaccuracy. The A.C. instruments were of thedynamometer type accurate to about 0.25 percent., and the electrostatic voltmeters werecalibrated to an accuracy of the same order.The largest experimental inaccuracy (possibly

per cent.) occurred in reading the tempera-ture rise, which was not allowed to exceedabout 20 deg. C. to ensure constancy of thegraphite resistance, which had previouslybeen " aged," during the measurements.The resistance was carefully adjusted andchecked, by means of an electrostatic volt-meter, for uniformity. Its temperaturecoefficient, and the effect of the smalldissipation of heat otherwise than in thecooling water, were eliminated by themethod of measurement. The rate of flowof the cooling water was held constant toless than 0.5 per cent.

(b) Results.Several trial runs were carried out to find

the time necessary to reach thermal equili-

brium (about 5 minutes) and to check theexperimental accuracy. The results wereconsistent to about i per cent. Comparisonsbetween the thermal and peak voltmetermethods of determining the voltage at B(Fig. 17) were carried out at 103 Mc. and35.3 Mc. (see Figs. 13 and 14) giving thefollowing results :

FrequencyVoltage by peak volt-

meter ..Voltage by thermal

methodDifference

103 35.3 Mc -

189.5 186.8 V. peak

172 182 V. peak9.3 2.6%

While the discrepancy between the twomethods is not very large, it is considerablygreater than had been expected and canhardly be explained by experimentalinaccuracies.

There are two other possible causes :(I) there may have been a small differenceof H.F. potential between the outer con-ductor of the line at the measuring pointand at the further end of the resistance, and(2) the thermal method is essentially anR.M.S. measurement and the form factorof the H.F. wave may have differed from 1/2,which value is assumed in the table above.While both these effects may have beenappreciable, it seems likely that the secondwas the more important. The oscillatorused in these tests had been adjusted andaccurately calibrated for other purposes ata considerably higher anode voltage thanthat used in the tests. As it was not per-missible to alter the adjustment, the fre-quency being varied by a single remotecontrol, the load coupling was tighter thanthe optimum at the anode voltage used,particularly at 103 Mc. The " flywheel "effect of the anode oscillatory circuit wouldthereby be reduced with an increase in theflattening of the voltage wave due to griddamping. Owing to the effect of electroninertia the latter would be greater at 103 Mc.than at 35.3 Mc. In addition the CIL ratioin the anode circuit was smaller at the higherfrequency. Direct measurement of the formfactor (e.g. by cathode ray oscillograph)seems hardly possible to the requiredaccuracy without a major investigation intothe measuring apparatus.

The effect of displacement currents in thesilica tube or the cooling water was alsoconsidered as a possible source of error.


A rough estimate of the magnitude of thiseffect indicates that it is unimportant.

While it is believed that the peak voltmetermeasurements are considerably more accuratethan the discrepancies in this comparisonmight suggest, the results illustrate thegreat difficulty of carrying out an absolutestandardisation of voltage at very highfrequencies.

2

2

1

12

4

LOW VOLTAGE TAPPING 15V X 50V.

LANGERANGE

97M

97Mcs. 40M0.

o Mo.

8 12

R.M.S. VOLTAGE FROM PEAKVOLTMETER READING

16

Fig i8.-Calibration curves for triodevoltmeter i (standard valve).

1

8

6

4

2

The apparatus was that previously de-scribed (see Figs. 9, io and 17). As interestlay chiefly in the performance of the triodesat low voltages, use was made of a tapping(removed during the previous tests) on theterminating resistance, since voltages lowerthan about 35 peak could not convenientlybe produced at the voltmeter head. Theratio of this tap at 5o cycles, with an allow-ance for the difference in H.F. voltage be-tween points A and B (Fig. 17), was takenas an approximation. A different resistanceunit was used which was known to be less

50V RANGE 50V.0150V.

f ...to Mo.

TALE TAP. .

LOW V

o

o

20 40 60 80 100

R.M.S. VOLTAGE FROM PEAK VOLTMETER READING

Fig. 09.-Calibration curves for triode voltmetershowing error introduced by using low voltage tap.

IV. Calibration of Other Valve Voltmetersagainst Peak Voltmeter.

A few calibration curves of two triode" anode -bend " voltmeters were taken atfrequencies of 40 Mc. (7.5 m.) and 97 Mc.

m.). It was not possible to arrange formaking these measurements to a high degreeof accuracy but the results serve as a guideto the performance of triode voltmeters atfrequencies above 3o Mc.

Voltmeter i used a valve similar to thestandard ML4 with the grid lead brought outat the top of the bulb ; voltmeter 2 a valveof very small dimensions (" acorn " type),actually a pentode with the screen used asanode. The former was a standard instru-ment intended primarily for broadcast wave-lengths, the latter an experimental one forshort and ultra -short wavelengths.

120

uniform than the one referred to previously.The peak voltmeter diode had a clearanceof 0.12 mm. and the necessary corrections(all small) were applied to its readings. Thetriode voltmeters were connected either tothe tapping or to the voltmeter head by con-centric leads about 14 cm. long for volt-meter i and 8 cm. for voltmeter 2. Thesewere the shortest the form of the apparatuswould allow. An approximate calculatedcorrection for the error due to the impedanceof these leads was applied.

The results are shown in Figs. i8, 19and zo. A form factor of Vi has beenassumed in converting the peak voltmeterreadings to R.M.S. values, the triode volt-meters being originally calibrated in R.M.S.voltage at 5o cycles.

Owing to the uncertainty in determining

204 THE WIRELE

the tap ratio and triode impedance error,the calibration curves do not give an exactaccount of the performance of the triodeitself. However, the presence of electroninertia error may be detected by (I)curvature of the calibration line, and (2)

discrepancy between different voltmeterranges involving different electrode poten-tials. The following conclusions can thus bedrawn about the inertia error :Voltmeter 1 :

40 Mc. The error becomes appreciableat about the following voltages :

15o v. range, 25 v.

5o v. range, io v.15 v. range, 6 v. (?)

97 Mc.

5o v. range, error large at about 15 v.

15 v. range, error much smaller atabout io v., but probably still quitelarge.

Voltmeter 2 : (15 v. range, higher ranges not(available).

40 Mc. No appreciable error down to 6 v.97 Mc. Error appreciable below about io v.An appreciable error may be taken to

mean one of a few per cent. or more. Theslight increase in apparent error at the highervoltages in Fig. 19 may possibly have beendue to a change of triode input impedance oroscillator wave -form with voltage. ,The factthat the upper line is straight, but does notquite pass through the origin when pro-duced, makes the latter explanation moreprobable.

Fig. 19 indicates that the readings at thelow voltage tap are about i8 per cent. highat 40 Mc. with voltmeter i, while Fig. zoshows that they are about 7 per cent. highwith voltmeter 2 at this frequency. Theerror in the estimated tap ratio thereforecannot be more than about 7 per cent. andmay well be less at this frequency. Theremainder of the i8 per cent. is due to theerror in estimating the effect of the connectinglead impedance which is likely to have beengreater at the low voltage tap than at thevoltmeter head. It is clear that in all the97 Mc. readings this effect outweighs thereduction in reading due to electron inertia.

The results bring out another interestingpoint. In voltmeter i the change from the5o v. to the i5 v. range involves a reductionanode voltage from 27o v. to zoo v. accom-


panied by a reduction of the (automatic)negative grid bias roughly in the ratio ofthe range voltages. It is clear from Fig. i8that the grid bias change has a much greatereffect. on the inertia error than the oppositelyacting anode -voltage change as, indeed,

20

15V. RANGELOW VOLTAGE TAPPING

18

97 MG

40 Mo.12

8

4

04

R.M.S. VOLTAGE

12

FROM PEAK

18

VOLTMETER READING

Fig. 2O.-Calibration curves jor triode volt-meter 2 (" acorn" valve).

qualitative theoretical considerations mightlead one to expect.

In conclusion it may be stated that bothtypes of triode voltmeter are serviceable,if not accurate, at frequencies of the orderof 4o Mc. for v9ltages down to 5 or io. Thecalibrations were not accurate enough tobring out the difference in inertia error atthis frequency, but it is not this, but the lowinput impedance, both per se and through itsexaggeration of the effect of the connectingleads, that is the chief disadvantage of theseinstruments as compared with the diodevoltmeter evolved in this investigation, andof the standard as compared with the smalltriode. For very low voltages (of the orderof i v. and less) the only hope for accuratemeasurements seems to be calibration froma diode peak voltmeter by means of anattenuating network capable of withstandinghigh voltages during its calibration. Schemesfor such measurements have been devised andit is hoped shortly to carry them into effect.

The author desires to tender his acknow-ledgment to The General Electric Companyand the Marconi Company, on whose behalfthe work was done which has led to thispublication.


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

of Scientific and Industrial Research

For Me information of new readers it is pointed out that the length of an abstract is generallyno indication of the importance of the work concerned. An important paper in English, in ajournal likely to be readily accessible, may be dealt with by a square -bracketed addition to thetitle, while a paper of similar importance in German or Russian may be given a long abstract.In addition to these factors of difficulty of language and accessibility, the nature of the work has,of course, a great influence on the useful length of its abstract.

205

PAGE PAGE

Propagation of Waves .. 205 Directional Wireless 215Atmospherics and Atmospheric Acoustics and Audio -Frequencies 215

Electricity 206 Phototelegraphy and Television .. 218Properties of Circuits 207 Measurements and Standards .. 221Transmission 208 Subsidiary Apparatus and Materials 225Reception .. 210 Stations, Design and Operation .. 228Aerials and Aerial Systems 212 General Physical Articles .. 228Valves and Thermionies 213 Miscellaneous 229

PROPAGATION OF WAVES1341. MICRO -RAY COMMUNICATION. EFFECT OF

WEATHER ON TRANSMISSION : POSSIBLEFUTURE USES OF WAVELENGTHS BELOW5o cm.-McPherson and Ulbrich. (Elec-trician, 7th Feb. 1936, Vol. 116, pp. 181-182and 184.) Summary of I.E.E. paper andsubsequent Discussion.

1342. SOME EXAMPLES OF THE REFLECTION ANDDIFFRACTION [by Buildings] OF ULTRA -SHORT WAVES OF 3 METRES WAVELENGTH.-J. Obata and Y. Munetomo. ( Journ. I.E.E.Japan, November, 1935, Vol. 55 [No. 11],No. 568, p. 1002: Japanese only.)

1343. TRANSMISSION OF I0 -METRE WAVES DURINGTHE SUMMER OF 1934, and TRANSMISSIONOF 20 -METRE WAVES DURING THE SUMMEROF 1934.-0. Burkard and G. Kunze :G. Kunze. (Sci. Abstracts, Sec. B, 25th Jan.1936, Vol. 39, No. 457, p. 52.)

For the io-m waves, sunspot maxima usuallygave minimum intensity, but atmospheric con-ditions had a greater influence, the best conditionsbeing when the transmitter was in a region of dryair under falling pressure. Fading occurred in51 day periods. For the 20-111 waves the effect ofsunspots was smaller and the 5/ day fading onlysporadic. " Results are taken from experimentsin all parts of the world."1344. THE INTERACTION OF RADIOELECTRIC WAVES

[Results of U.R.S.I. Tests in Feb. andMarch, 1935 : Observations by 3o Physicistsin Various Countries].-B. van der Pol.(L'Onde Elec., December, 1935, Vol. 14,No. 168, pp. 804-808.) See 3749 of 1935.

1345. THE LUXEMBOURG EFFECT [Survey andDiscussion].-D. Graffi. (Alta Frequenza,January, 1936, Vol. 5, No. 1, pp. 42-48.)

A footnote remarks that " recent observations

made by the R.I.E.C. at Leghorn indicate thatthis last rule [that the interfering station shouldbe 250-300 km away along the great circle joiningthe receiver to the wanted station] is not alwaysobeyed, since Luxembourg produced its effect onthe Kattowice (Poland) station."

1346. CONFIRMATION OF COSMIC PHENOMENON[54 -Day Fade -Out], and THE HYDROGENOUTBURST ON THE SUN AND RADIO FADING.-J. H. Dellinger : R. S. Richardson.(Science, 6th Dec. 1935, Vol. 8z, pp. 548-549ioth. Jan. 1936, VOL 83, Supp. pp. 6-7.)See also 1347.

1347. HIGH -FREQUENCY RADIO FADE-OUT ONOCTOBER 24, 1935.-Dellinger. (Journ.Franklin Inst., January, 1936, Vol. 22 I ,No. I, pp. 157-158.)

See also 1346, above ; 1348, below ; and 862 ofMarch. A fade-out of modified form occured on24th Oct. 1935. The obliteration of signals didnot extend to such low frequency values and wasnot so sudden. There was simultaneous sunspotactivity and magnetic disturbance. The high -frequency transmission improved as sunspot activityincreased but was suddenly impaired on 24thOctober, subsequently recovering. The cause ofthe fade-out may be a very sudden eruption on thesun.

1348.

1349.

A NEW SOLAR RADIO DISTURBANCE [54 -DayFade-Out].-J. H. Dellinger. (Electronics,January, 1936, VOL 9, pp. 25 and 34.)Cf. above.

ELECTRIC INFLUENCE OF THE ACTIVE REGIONOF THE Sun [Five Terrestrial ElectricPhenomena give Indications of 27.25 -dayPeriod in Phase with Variation of SunspotActivity].-F. Sanford. (Phys. Review,15th Jan. 1936, Series 2, VOL 49, No. 2.pp. 206-207: abstract only.)


1350. RECENT LARGE SUNSPOTS [Data for Decem-ber, 1935, and January, 1936].-(Nature,1st Feb. 1936, Vol. 137, pp. 183-184: shortnote only.)

1351. DATA ON DAILY NUMBER OF METEORS.-C. C. Wylie. (Sci. News Letter, 25th Jan.1936, Vol. 29, p. 56.)

1352. SEARCHLIGHTS TO EXPLORE THE UPPER AIR.-Tuve, Johnson and Wulf. (Science, 13thDec. 1935, Vol. 82, Supp. p. 13.) See also861 of March.

1353. THE IONOSPHERE AND ITS INFLUENCE UPONTHE PROPAGATION OF RADIO WAVES [Sur-vey].-J. P. V. Madsen. (John MurtaghMacrossan Lecture, 1935, University ofQueensland, 43 pp.) Published by Simmons,Limited, Sydney.

1354. ON THE METHOD OF EXPRESSING ANDUTILISING DATA ON RADIATION AND PROPA-GATION [and the Use of the Magnitudes" Forza Cimomotrice " (Cymomotive or" Wave -Producing " Force) and " VirtualDistance "].-L. Sacco and U. Tiberio.(Alta Frequensa, December, 1935, Vol. 4,No. 6, pp. 668-687.)

The need for these two quantities, long employedin the Italian army, is suggested by certain defectsin the ordinary way of representing the character-istics of radio stations, and propagation data,revealed at the Lisbon reunion of the C.C.I.R." Cymomotive force " (alternative terms suggestedare " radiation potential " and " radiomotiveforce ") denotes the capability of a transmitter toproduce a field at a distance in a certain direction ;so that if E is the cymomotive force in volts, Fthe field strength produced, in my/m, at a distanceof D in km, then within certain limits E = FD :for a vertical earthed semi -dipole, of height muchsmaller than the wavelength, E = FD = 3ooVP,where P is the power in kw. Thus such an aerial,radiating r kw with circular emission, wouldpossess a cymomotive force of 30o v in the hori-zontal plane. " Virtual distance " denotes theattenuating property of a given path, and is givenby EIF', where F' is the field strength actuallyproduced ; so that for I kw radiated circularly (asabove) D = 300/F'.

1355. AN IMPROVED PULSE TRANSMITTER [givinga Truly Square -Topped Pulse which can bemade Very Narrow].-G. Millington andS. W. H. W. Falloon. (Marconi Review,Nov./Dec. 1935, No. 57, pp. 12-17.)

With the ordinary Ratcliffe -White circuit (1933Abstracts, p. 495) the h.f. output remains at maxi-mum for a short time and then falls rapidly butnot instantaneously to zero, so that if a very narrowpulse is attempted its top will be so narrow thatthe transmitter will not have time to build up tothe true peak value. The writers therefore employa second thyratron, controlled by a delay circuitfrom the first thyratron, to give a very sharpcut-off.

1356. MEASUREMENT OF THE ATMOSPHERIC OZONEDURING THE POLAR WINTER AT ABISKO[and the Relations between Quantity ofOzone and Horizontal Displacements ofLarge Masses of Air].-Barbier, Chalongeand Vassy. (Journ. de Phys. et le Radium,December, 1935, Vol. 6, pp. 132-133 s.)For a Comptes Rendus Note see 31 of January.

1357. THE OXYGEN AFTERGLOW [Replyto Criticism].-E. M. Stoddart. (Proc. Roy. Soc., SeriesA, 2nd Dec. 1935, Vol. 153, No. 878, pp. 152--157.) See Stoddart, Phil. Mag., 1934, Vol.18, p. 409 ; Rayleigh, Proc. Roy. Soc.,Series A, 1935, Vol. 150, P. 34.

1358. ZEEMAN EFFECT IN THE ATMOSPHERICOXYGEN BANDS : PRODUCTION OF A STRONGMAGNETIC FIELD OVER A LENGTH OF 80 CM.-R. Schmid. (Phys. Review, 1st Feb.1936, Series 2, Vol. 49, No. 3, p. 271.)

1359. THE POSSIBLE IDENTIFICATION OF CERTAINNIGHT -SKY RADIATION WITH THE SCHUMANN-RUNGE BANDS OF THE OXYGEN MOLECULE.-J. Cabannes and J. Dufay. (ComptesRendus, 5th Feb. 1936, Vol. 202, No. 5,PP. 365-367.)

1360. THE EXCITATION OF THE AURORAL GREENLINE BY METASTABLE NITROGEN MOLECULES[Experimental Conditions compared withThose of Night Sky and Aurora Borealis].-J. Kaplan. (Phys. Review, 1st Jan. 1936,Series 2, Vol. 49, No. r, pp. 67-69.)

1361. IONIC DISPERSION [in Electrolytes] IN THEEXTREME INFRA RED [Introduction ofInertia Term into Equations for Infra -RedReflection and Absorption by Thin MetallicFilms].-C. H. Cartwright. (Phys. Review,1st Jan. 1936, Series 2, Vol. 49, No. ,, pp.101-102.)

1362. ABERRATION OF LIGHT AND THE DOPPLEREFFECT [Aberration Phenomena with DoubleStars and Terrestrial Objects indicate that,Velocity of Light depends on Time].-K. Papello. (Zeitschr. f. Physik, No. 7/8,Vol. 98, 1936, pp. 490-495.)

1363. DISPERSION AND RESONANCE PHENOMENAIN THE UPPER LAYERS OF THE GROUND[Natural Vibrations of Ground : Observedand Calculated Dispersion Curves for Ray-leigh Waves].-R. KOhler. (Zeitschr. f. tech.Phys., No. 12, VOL I6, 1935, pp. 597-600.)

ATMOSPHERICS AND ATMOSPHERICELECTRICITY

1364. RADIO NOISES FROM THE GALAXY [due toIonisation of Cold Matter by Diluted StellarRadiation in Interstellar Space].-R. M.Langer. (Phys. Review, 15th Jan. 1936,Series 2, VOL 49, No. 2, pp. 209-210: abstractonly.) See 892 of March : for previous papersby Jansky see 41 of January.


1365. CHANGE OF PATH OF LIGHTNING FLASH [0C-casioned by Induced Charges in Neighbour-hood].-B. Walter. (Ann. der Physik, Series5, Vol. 25, No. 2, 1936, pp. 124-142.)

A further discussion of some photographs oflightning flashes (Figs. 1-6) taken with a movingcamera and already published and discussed bythe writer (Ann. der Physik, Series 4, Vol. Io, 1903,p. 393). The main conclusion is that " the changewith time of the path of a discharge from a thunder-cloud depends to a considerable extent on the numberand nature of the induced charges produced in theneighbourhood of the cloud by its high potential."The path first taken is directed towards one of theinduced charges, but the second stroke may bedirected towards another, e.g. in another cloud,although the first path was already ionised and wouldseem to offer less resistance than a new path. Figs.2 and 3 show cases when both paths end on theearth ; this is explained by the assumption that thefirst path met a region of high conductivity surround-ed by one of low conductivity ; the second pathstrikes a more distant highly -conducting region.Subsequent return of the discharge to its originalpath may occur : there appears to be a high degreeof electrical similarity between clouds holdinginduced charges and the upper layers of the earth'ssurface. See also 452 of February.

1366. THE USE OF ALUMINIUM FOR LIGHTNING -CONDUCTOR LEADS [and the Calculation ofthe Necessary Cross -Section to avoid Fusing].-B. Walter. (Zeitschr. f. tech. Phys., No.Vol. 17, 1936, pp. 17-19.)

1367. LIGHTNING INVESTIGATIONS ON A DISTRIBU-TION SYSTEM [with Interconnection Methodof Protection : Surge Recorder Results].-Halperin and Grosser. (Elec. Engineering,January, 1936, Vol. 55, No. 1, pp. 63-70.)

1368. OSCILLOGRAPHIC STUDY OF OVERVOLTEDDISCHARGES [in Air : Field Emission initiatesDischarge].-J. W. Flowers. (Phys. Review,15th May, 1935, Series 2, Vol. 47, No. I°,pp. 801-802: abstract only.) See also45 of 1935-

1369. A STUDY OF SHORT TIME LAGS IN SPARKS ASA FUNCTION OF OVERVOLTAGE [in Air atAtmospheric Pressure].-R. R. Wilson.(Phys. Review, 15th Jan. 1936, Series 2,Vol. 49, No. 2, p. 210: abstract only.)

1370. EFFECT OF RAINFALL ON IONISATION REGIS-TERED BY RECORDING COSMIC METER WITHToP SHIELD REMOVED [Radioactive Materialsbrought down from Upper Atmosphere].-R. L. Doan. (Phys. Review, 15th Jan. 1936,Series 2, Vol. 49, No. 2, p. 198: abstractonly.)

1371. FURTHER INVESTIGATIONS OF THE ATMOS-PHERIC IONISATION ASSOCIATED WITH RAIN-FALL [Radioactive Matter accumulated inThundercloud brought down by Rain].-G. R. Wait and A. G. McNish. (Phys.Review, 15th Jan. 1936, Series 2, Vol. 49,No. 2, p. 201 : abstract only.) See 1934Abstracts, p. 434.

207

1372. STUDY OF THE EARTH'S MAGNETIC FIELD,ATMOSPHERIC IONISATION AND VERTICALCURRENT AT SCORESBY SOUND DURING THEPOLAR YEAR.-Dauvillier. (Journ. dePhys. et le Radium, January, 1936, Vol. 7,No. 1, pp. 40-48.) For the beginning ofthis paper see 899 of March.

STUDY OF SOME PROBLEMS OF TERRESTRIALMAGNETISM : EXPLANATION OF THE DIURNALVARIATION OF THE FIELD [based on Scoresby

1373.

Sound Observations].-J. P. Rothe. ( Journ.de Phys. et le Radium, January, 1936, Vol.7, No. i, pp. 148-149 D.) For a previouspaper see 1934 Abstracts, p. 32.

1374. A PRACTICAL SYSTEM OF AUTOMATIC RADIOSIGNALS FROM A FREE BALLOON [Radio-meteorograph using 5 m Waves].-Nat.Bureau of Standards. (Journ. FranklinInst., November, 1935, Vol. 220, No. 5, p.659: short note only.)

1375.

PROPERTIES OF CIRCUITSSURVEY OF RECENT RESEARCHES ON NON-LINEAR OSCILLATIONS [Theoretical and Ex-perimental Russian Work : Report toU.R.S.I., London Meeting].-Mandelstam,Papalexi, Andronow, Chaikin and Witt.(Tech. Phys. of USSR, No. 2/3, VOL 2, 1935,pp. 81-134: in French.)

1376. RESONANCE PHENOMENA IN LINEAR SYSTEMSWITH PERIODIC PARAMETERS.-G. Gorelik.(Tech. Phys. of USSR, No. 2/3, Vol. 2, 1935,

French version of the papersdealt with in 2592 of 1935 and 64 of January.

1377. ON A METHOD FOR PARAMETRIC EXCITATION.-J. B. Kobzarev. (Journ. of Tech. Phys.[in Russian], No. 6, Vol. 5, 2935, pp. 958-963.)

In the systems so far investigated, in which thefrequency of oscillations is different from that ofthe applied e.m.f., it is observed that in additionto the variation of the parameters a " direct "action such as a forced variation of the oscillatingcurrent also takes place. In the present paper acase of a " pure " parametric excitation is examined.The circuit under consideration makes use of adynatron oscillator adjusted for operation at thepoint of intersection of the anode-current/anode-voltage curves, i.e. the point for which the anodecurrent is zero, and with the grid bias sufficientto prevent self -oscillations. If an e.m.f. is appliedto the grid of the valve there will be no " direct "action, since no current is flowing through thecircuit ; the slope of the valve characteristic will,however, be varied periodically and it is shownthat if the frequency of the applied e.m.f. is twicethe natural frequency of the circuit, oscillations atthe natural frequency will under certain conditionsarise in the circuit.

A mathematical investigation of the system isgiven including the determination of the amplitudeof the oscillations and of the time required for theirgrowth and decay. Resonance curves obtainedexperimentally are also shown.

D 2


1378. ON THE THEORY OF NON-LINEAR RESONANCE.-J. B. Kobsarev. (Tech. Phys. of USSR,No. I, VOL 2, 1935, pp. 27-42 : in German.)

Author's summary :-By the use of a special,clear method, consisting in the consideration of theaction of an " ideal impulse " on a non-linearsystem and in the representation of the functionin the form of a sum of a continuous succession ofimpulses, general formulae of the first approxima-tion are obtained which are valid not only for thecase of " dividing " resonance [m $ I, n = I ,where min is the ratio of the imposed frequencyto the natural frequency of the non-linear system]but also for " multiplication " resonance [mn * 1 ; there is a printer's error here], " fractional "resonance [m * ,, n * r], and for resonance witha periodic external force of arbitrary form.These formulae contain as dependent variablesthe slowly varying amplitude and phase of thenatural oscillations of the system. By trans-formation of the equations to rectangularco-ordinates it is shown that in the specialcase of " dividing " resonance these equa-tions become the same as those found by Mandel-stam and Papalexi [1932 Abstracts, pp. 279-280].In the same co-ordinates the general stabilitycriteria of the periodic movements are obtained :these are generalisations of the known criteriaobtained by Mandelstam and Papalexi by the useof Poincare-Liapounov methods. The advantageof the method developed in the present paperconsists, in the author's opinion, in the possibilityit presents of a searching investigation of themechanism of the phenomenon. For previouspapers see 929 of March, and 1377, above.1379. ON A CASE OF " PULLING-IN."-J. B.

Kobzarev. (Journ. of Tech. Phys. [in Russian],No. 6, Vol. 5, 1935, pp. 964-974.)

It is known that if two e.m.fs of equal amplitudebut different frequencies are applied to an oscillatoroscillating at a frequency approximately equal tothe arithmetic mean of the applied frequencies," pulling -in " may take place. This is shown bythe fact that out of the three frequencies originallypresent only the frequencies of the applied e.m.fsremain in the circuit.

A detailed mathematical investigation of thisphenomenon, from the point of view of quasi -lineartheory, is presented in this paper for a typical caseof an inductively coupled oscillator. Equationsof the first approximation are derived and con-ditions determined for the stability of the resultantoscillations. The relationship between the ampli-tude of the oscillations and the amount of de -tuning is also investigated and the results are shownin a number of graphs.1380. BUILDING -UP AND BEAT PROCESSES IN

" MITNAHME " [Pull-In].-P. Riazin. (Tech.Phys. of USSR, No. 2/3, VOL 2, 1935, pp.19$-214: in German.) Practically identicalwith the Russian paper dealt with in 3822of 1935.

1381. A NEGATIVE -RESISTANCE OSCILLATOR [Useof Amplifier with Input and Output Ter-minals connected].-N. L. Yates -Fish. (Proc.Phys. Soc., ist Jan. 1936, Vol. 48, Part 1,No. 264. pp. 125-134.)

Author's abstract :-It is shown that an amplifier

with its input and output terminals connectedtogether can behave as a negative resistance. Thisprinciple leads to the design of an oscillator em-ploying two triodes, and having certain advantagesover both the ordinary reaction circuit and thescreen -grid dynatron. Some special wave -formswhich were observed are illustrated. A methodof using the instrument to measure the dynamicresistance of an oscillatory circuit is suggested.1382. [Theory of] DETECTION OF HIGH -FREQUENCY

CURRENT BY THE SHUNTED CONDENSER. INTHE GRID CIRCUIT OF A VALVE.-D.Milossavlievitch. (Comptes Rendus, lothFeb. 1936, Vol. 202, No. 6, pp. 472-474.)

An algebraical expression is found for the gridvoltage when a sinusoidally varying voltage isimpressed on the far side of a shunted condenser inseries with the grid. The expression for the dropin voltage differs from the usual one by a factorof the order of 0.7.

1383. THE FILTER -COUPLED INDUCTIVE GLOW -DISCHARGE OSCILLATOR [Control Circuitcomprising Multisection Filter : Character-istics : Frequency Stability of ResonanceBand].-W. E. Kock. (Phys. Review, 15thJan. 1936, Series 2, VOL 49, No. 2, p. 196 :abstract only.) For previous papers see1906 of 1935.

1384. R.F. TRANSITION LOSSES [due to ImpedanceMis-Matching : Graphical Method of Evalua-ting], and R.F. IMPEDANCE -MATCHING NET-WORKS [with Chart of Pi -Section Networks].-H. A. Wheeler : R. P. Glover. (Elec-tronics, January, 1936, Vol. 9, pp. 26-27and 46 : pp. 28-29.)

1385. " INTRODUCTION TO ELECTRIC TRANSIENTS."-E. B. Kurtz and G. F. Corcoran. (AtPatent Office Library, London : CatalogueNo. 75762.)

TRANSMISSION1386. FUNDAMENTAL CHARACTERISTICS OF ELEC-

TRONIC OSCILLATIONS.-E. C. S. Megaw.(Nature, ist Feb. 1936, Vol. 137, pp. 189-190.)

Experiments are described which show that" electronic oscillations can be produced withoutany external resonant circuit and that they can havea single frequency." The mode of oscillation maybe the same as that reported by Slutzkin and Lel-jakow (1934 Abstracts, p. 33) ; oscillations due to" axial " and " radial ' motion of the electronscould be simultaneously maintained.1387. ON THE THEORY OF GENERATION OF OSCIL-

LATIONS IN LECHER SYSTEMS [Analysis ofWitt's Circuit for Short and Ultra -ShortWaves].-S. Strelkov. (Tech. Phys.- ofUSSR, No. 2/3, VOL 2, 1935, pp. 232-247: inGerman.)

For Witt's paper see 1934 Abstracts, p. 612.It is here assumed that the characteristic impedanceis large compared with the ohmic resistance, andthat the electronic currents in the valve are smallcompared with the capacitive currents (the systemconsists of Lecher wires connected between anodeand grid, the back coupling being provided by thecapacities of the electrodes and of the conductors


connected to them). Conditions for self -excitationare found, and formulae are obtained for theamplitudes of the oscillations ; the stability of theseDscillations is investigated. The possibility of variousoscillation regimes, for fixed parameters, is shown,and for certain special cases the dependence of theseregimes on the length of the Lecher system is given.The far end of the Lecher system is taken as open,short-circuited, or closed by an inductance or acapacity.

Of the various oscillations which are found to bepossible, only some are stable : those with thefrequency of the fundamental are always stable,while those with frequencies of the higher over-tones are only stable when their increment is atleast twice as large as that of the fundamental(A2 > 2 . 01/A8, where 01 is the increment of thefundamental and A the amplitude of the overtone :A 2 is proportional to the increment of the overtone-see eqns. 23 and 24).r 388 . DEVIATION -CONTROL IN MICRO -WAVE GENER-

ATOR VALVES [Theoretical and PreliminaryExperimental Investigation of Recent" Cathode -Ray -Tube " Idea of OscillationMechanism in Barkhausen-Kurz and Mag-netron Oscillators] .-U. Tiberio. (AltaFrequenza, December, 1935, Vol. 4, No. 6,PP. 714-732.)

Recent work (e.g. Carrara's on detection-ro34of 1935 : see also 1879 of 1935) indicates that thenegative resistance of micro -wave oscillators isdue to partial deflection of the emission currentalternately to each of the two electrodes (grid andplate in the B -K oscillator, and the two plates inthe magnetron) under the control of the alternatingp.d. between these. To analyse this hypothesis,the writer imagines a model valve in which thedeflection principle is applied by a control mechan-ism of an oscillographic type. The mode of opera-tion thus appears simple and clear, and allowsvarious deductions to be made regarding impedance,maximum efficiency, etc. The possibility of con-structing actual valves after this model is discussed,and the design of a recent experimental type illus-trated : this has an anode current of 3o mA and aray diameter about x mm. Alfven's early valveon the " transverse field " principle (1931 Abstracts,p. 557) is referred to.1389. ON THE MOTION OF ELECTRONS IN CROSSED

ELECTRIC AND MAGNETIC FIELDS WITHSPACE CHARGE [Physical Meaning of Theory :Magnetic Cut-Off].-L. Tonks. (Physik.Zeitschr. der Sowjetunion, No. 5, Vol. 8,1935, pp. 572-578: in English.)

An examination of Braude's theoretical solution(3437 of 1935) with an exposition of its physicalmeaning. " Braude's conclusion that his solution. . . shows the absence of cut-off by a criticalmagnetic field is shown to be incorrect. The natureof the potential distribution and its relation tocut-off is [here] analysed in detail. A series ofvirtual cathodes of a peculiar type are found, whosespacing is proportional to current. A similar typeof solution must exist in the cylindrical case. Onlythat solution which has no virtual cathode is stableand hence of physical significance, except perhapswhen a cut-off magnetron is generating oscillations.

209

The theory accounts for the decrease in current in amagnetron as cut-off is approached." See 1390,below.

1390. ON THE " CUT-OFF " IN THE PLANE MAGNET-TRON WITH SPACE CHARGE.-S. J. Braude.(Physik. Zeitschr. der Sowjetunion, No. 5,Vol. 8, 1935, pp. 584-586: in English.)Note on above paper, discussing physicaland mathematical truth of equations ob-tained.

1391. THE MOTION OF ELECTRONS IN ELECTRICAND MAGNETIC FIELDS, TAKING INTO CON-SIDERATION THE ACTION OF THE SPACECHARGE.-Braude. (See 1444.)

1392. CORRECTION TO THE PAPER " THE PATHS OFTHE ELECTRONS IN THE MAGNETRON IN-CLUDING THE EFFECT OF SPACE CHARGE. I "[Paths of Electrons in Absence of MagneticField need not be Straight Lines : Depen-dence on Potential Field].-Awender, Thoma,and Tombs. (Zeitschr. f. Physik, No. 7/8,Vol. 98, 1936, pp. 534-535.) See 7o ofJanuary.

1393. THE POSSIBLE USE OF A TEMPERATURE -LIMITED DIODE TO GENERATE NEGATIVE -RESISTANCE OSCILLATIONS AT ULTRA -HIGHFREQUENCIES.-North. (See 1450.)

1394. A NEGATIVE -RESISTANCE OSCILLATOR [Am-plifier with Input and Output Terminalsconnected].-Yates-Fish. (See 1381.)

1395. MODULATION OF FREQUENCY AND THENECESSITY OF TREATMENT AS A NON-LINEARPHENOMENON : EXPERIMENTS WITH A WIEN[Toothed -Wheel] A. F. GENERATOR.-S.Rytov. (Tech. Phys. of USSR, No. 2/3,VOL 2, 1935, pp. 214-231 : in French.)

Previous workers have dealt with frequencymodulation by linear theory. This, the presentwriter urges, must be considered unsatisfactoryfor several reasons (p. 218), all the more so sincefrequency modulation can only be obtained inpractice in a non-linear system such as a triodeoscillator. His treatment by non-linear theoryof such an oscillator, frequency -modulated, leadsto the following results :-The oscillation spectrumis identical with that which linear theory gives,neglecting the regions of instability : regions ofinstability are non-existent (there are no phenomenaanalogous to linear instability) : frequency modula-tion is accompanied by a modulation of amplitude(this is different from the result given by lineartheory). These results apply to practical conditionsonly, e.g. when the oscillator regime is sufficientlyfar from the threshold of oscillation, and the modu-lating frequency is low compared with the carrierfrequency.

The second part of the paper deals with a fre-quency modulation phenomenon observed in atoothed -wheel generator of sinusoidal e.m.fs.The writer finds that besides the principalfrequency of 540 c/s there are often presentsymmetrically placed satellite frequencies of 54o±18, 540±36, and 54°±54, of amplitudes respec-tively 86%, 4o% and ro % of the principal amplitude.Investigation of these shows them to be the result


of frequency modulation produced, not by inequali-ties in the spacing of the rotor teeth, but by slightvariations in speed of rotation. The analysis wascarried out with the help of a special stretched -string resonator : the method may have usefulapplications in the study of periodic inequalitiesin rotation in internal combustion motors, turbines,etc.

1396. FREQUENCY MODULATION EQUIPMENT [NoteModulation for Short -Wave Telegraphy, toreduce Selective Fading : Privacy Modu-lators or " Wobblers " for Secrecy in S.W.Telephony].-E. Green and J. L. Hewitt.(Marconi Review, Nov./Dec. 1935, No. 57,pp. x-ri.) For Roder's paper, on which thefirst section on theory is mainly based, see1932 Abstracts, p. 162.

1397. METHODS FOR THE COMPENSATION OF NON-LINEAR DISTORTION [0/ Grid -ModulatedTransmitters].-A. I. Eilenkrig and E. I.Gorodnichev. (Izvestia Elektroprom. Slab.Toka, No. 12, 1935, pp. 43-54.)

The non -linearity of the modulation characteristicof a radio transmitter using the grid modulationsystem is mainly due to the lower bend of theanode-current/grid-bias characteristic of the modu-lated amplifier valve. In the present paper thefollowing three methods are proposed for correctingthis distortion :

(1) The modulating frequency is appliedsimultaneously to the grids of two consecutivestages of b.f. amplification (Fig. r). The firststage is arranged to operate near the upper bendof the characteristic so that it becomes effectiveonly during the negative swings of the modulatingfrequency, with the result that the grid of thesecond stage receives an additional negativepotential during these swings.

(2) The anode -filament circuit of a com-pensating valve is connected across a resistanceR1 in series with the grid of the modulatedamplifier (Fig. 4). The modulating frequency isapplied to the grids of both the modulatedamplifier and compensating valve (I80° out ofphase in the latter case). The anode currentflows through the compensating valve onlyduring the negative swings of the modulatingfrequency and causes an additional voltage dropin R1 and therefore a decrease in the grid potentialof the modulated amplifier.

(3) A screen grid valve is used for the modu-lated amplifier (Fig. II) and the screen voltageis regulated by means of a compensating valveas in (2). Some indications are given for deter-mining the constants of the circuits and it isstated that circuit (2) has proved so successfulthat it was incorporated in radio broadcastersmanufactured in the U.S.S.R. during 1935.

1398. " QUALITY " : MODIFYING BROADCAST TRANSMISSION TO AID THE LISTENER[Present Demand for 90% Maximum Modu-lation Depth a False Step : Improvement'inReceived Quality if Modulation neverexceeded 60%].-P. P. Eckersley. (WirelessWorld, 7th Feb. 1936, Vol. 38, pp. 133-134.)

April, 1936

1399. GRID -LEAK MODULATION [Good Results withProper Arrangement : Advantages].-O. H.Huston. (QST, February, 1936, Vol. zo,p. 58.)

1400. A METHOD OF TELEGRAPHIC KEYING, ANDOF TELEPHONIC MODULATION, FOR HIGH -POWER RADIOELECTRIC TRANSMITTINGSTATIONS [by Variation of Impedance ofCoupling between Oscillator and AmplifierStages, or between Transmitter and Aerial].-M. Michel. (Ann. des Postes T. et T.,January, 1936, Vol. 25, pp. 47-75.) Givingexcellent signals even on waves down tozo m. The system is in permanent serviceat a 35 kw commercial telegraph station.

140t. BROADCAST TRANSMITTER FEATURES[Cabinet -Style Construction : Accessibility :R.F. Insulators : Fewer Rotating Machines :etc.].-J. P. Taylor. (Electronics, January,1936, Vol. 9, pp. 20-24.)

1402. HIGH -POWER AUDIO TRANSFORMERS [forClass B Amplification : 7.5 and 18o kwTransformers used at WLW].-J. F. Peters.(Elec. Engineering, January, 1936, Vol. 55,No. 1, pp. 34-36.)

RECEPTION1403. UNICONTROL RADIO RECEIVER FOR ULTRA-

HIGH FREQUENCIES USING CONCENTRICLINES AS INTERSTAGE COUPLERS [A/4 Linesgrounded at One End, with Plungers gangedfor Unicontrol : Four -Stage Amplifier withAmplification over i6 per Stage at roo Mc/s,and 2 per Stage at 300 Mc/s].-F. W.Dunmore. ( Journ. of Res. of Nat. Bur. ofSids., December, 1935, Vol. 15, No. 6,pp. 6o9-618.) By running the d.c. platesupply lead down through the centre of theinner concentric tube, only one concentricline is used between each stage, in place ofthe obvious two. See also 1404.

1404. AN UNCONVENTIONAL RECEIVER FOR THEULTRA -HIGH FREQUENCIES; DETAILS OF ANEW DEVELOPMENT FROM THE NATIONALBUREAU OF STANDARDS.-Dunmore. (QST,February, 1936, Vol. 20, pp. 21-23.) Seealso 1403.

1405. ULTRA -SHORT-WAVE SUPERHET FOR A.C.MAINS ; ON Wireless World " SINGLE -SPAN " PRINCIPLE : EXTENSIBLE TO ANALL - WAVE SUPERHET. - G. Faust.(Funktech. Monatshefte, January, 1936,pp. 18-20.)

1406. A TEN -METRE CONVERTER FOR BAND -SWITCHING SUPERHETS WITHOUT 28 MC/SCOVERAGE.-G. Grammer. (QST, February,1936, Vol. 20, pp. 39-41.)

1407. A NOISE -SILENCING I.F. CIRCUIT FOR SUPER -HET RECEIVERS: AN EFFECTIVE METHODOF COPING WITH AUTO -IGNITION AND OTHERELECTRICAL INTERFERENCE IN C.W. ANDTELEPHONE RECEPTION.-J. J. Lamb. (QST,February, 1936, Vol. 20, pp. 11-14 and 38,9o, 92, io6, ro8, no and 112.)

Making use of the short -time nature of the

April, 1936 THE WIRELESS ENGINEER 2 I I

interference (until prolonged in effect by someelement of the receiving circuit such as the loud-speaker) and acting before it reaches the stages mostsusceptible to overloading and cross -modulation,thus differing from other limiting devices. Thenoise peaks extending above the desired signalamplitude are amplified and rectified, and therectified voltage is used to reduce the gain of thefinal i.f. stage, so that the noise kills itself bypartly or wholly blocking this amplifier during thenoise pulse. The time is so short that no audiblegap in the desired signal is discernible. For thenecessary small time constant and freedom frominstability of the silencing action, not only propercircuit constants are required but also valves withsuitable characteristics ; no other valve has beenfound so suitable as the metal 6L7 (cf. 1455) forthe i.f. silencer -amplifier. There is an enthusiasticeditorial note on the performance of the scheme.1408. AUDIO OUTPUT LIMITERS FOR IMPROVING

THE SIGNAL-TO-NOISE RATIO IN RECEPTION.-H. A. Robinson. (QST, February, 1936,Vol. 20, pp. 27-30.)

1409. A DETECTOR CIRCUIT FOR REDUCING NOISEINTERFERENCE IN TELEPHONE RECEPTION[by Balancing -Out All Voltages beyond aCertain Amplitude].-L. E. Thompson.(QST, February, 1936, Vol. 20, pp. 44-45.)For a previous article, on c.w. reception, see2243 of 1935.

1410. THE II -WEEKS' MYSTERY INTERFERENCE[traced to Diathermy (" Inductothenn.")Machines at Athletic Clubs].-(QST, Feb-ruary, 1936, Vol. 20, pp. 9-10: Editorial.)A ' gargly " signal frequently changing itswavelength by a few hundred kilocycles andcausing trouble to innumerable services.Cf. 964 of March.

1411. RADIO -INTERFERENCE : SUPPRESSION METH-ODS FOR [Oil -Burning] WATER -HEATINGPLANTS.-Belling & Lee, Ltd. (Electrician,24th Jan. 1936, Vol. 116, p. 96: summaryonly.)

1412. LUXEMBOURG EFFECT [Various Readers'Results].-(World-Radio, 7th Feb. 1936,Vol. 22, p. 9.) Interference examples inN.W. London, from Luxembourg, Droit-wich, Kalundborg and Radio -Paris, showamong other things that the affected stationswere all above 35o m.

1413. PAPERS ON LUXEMBOURG EFFECT .-Graffi :van der Pol. (See 1344 and 1345.)

1414. SELECTIVITY OF TUNED CIRCUITS [increaseswith Wavelength, Decreased H.F. Resist-ance preponderating over Effect of DecreasedL/G Ratio.]-S. 0. Pearson. (WirelessWorld, 31st Jan. 1936, Vol. 38, pp. 100-102.)

1415. FEEDING [High -Fidelity Resistance -Coupled]PUSH-PULL AMPLIFIERS : METHODS OFPHASE REVERSAL.-W. T. Cocking. (Wire-less World, 7th Feb. 1936, Vol. 38, pp. 126-128.) Including the " paraphas.e " systemand a new system giving high -note accen-tuation without reducing gain at otherfrequencies (see also 983 of March).

1416. DETECTION OF H.F. CURRENT BY THESHUNTED CONDENSER IN THE GRID CIRCUITOF A VALVE.-Milossavfievitch. (See 1382.)

1417. THE PROBLEM OF " DYNAMIC INCREASE "[Contrast Expansion (and Compression) forBroadcasting, Sound Films, and Gramo-phones] : AMPLIFIERS WITH AUTOMATICVARIATION OF AMPLIFICATION [Survey OfSeveral Methods].-H. Lamparter. (Funk -tech. Manatshefte, January, 1936, pp. 13-17.)

With an enthusiastic introduction by Leitlauser.The author concludes that while satisfactory proper-ties have already been obtained, all difficulties havenot yet been eliminated : when this is accomplished,there is no doubt that volume compression at themicrophone end, and expansion at the reproducingend, will come into use. See also 1935 AbstractsIndex, under " Reception "-Volume.1418. STUDY OF LEVEL REGULATORS AND " ANTI -

FADING " DEVICES.-G. Espinasse. (Ann.des Posies T. et T., December, 1935, Vol. 24pp. 1098-1117.)

1419. THE EFFECTIVENESS OF AUTOMATIC VOLUMECONTROL DEVICES.-P. Mandel. (Elec-tronics, January, 1936, Vol. 9, PP. 44-45)Long summary of the French paper dealtwith in 3854 of 1935.

1420. AUTOMATIC AUDIO -FREQUENCY VOLUME CON-TROL DEPENDING ON THE EXTERNAL NOISELEVEL [for Cinemas, Public Address, Restaur-ants, etc.].-H. Boucke. (Funktech. Manats-

, hefte, January, 1936, pp. 31-32.)1421. REMARKS ON AUTOMATIC SENSITIVITY CON-

TROL BY VARIABLE -MU VALVES : ADVANTAGEOF THE USE OF A SPACE -CHARGE -GRIDVALVE.-M. Chauvierre. (L'Onde Elec.,December, 1935, Vol. 14, No. 168, pp.809-820.)

In adopting the variable -mu valve for the auto-matic volume control of broadcast receivers,designers soon realised that valves with trulyparabolic characteristics required considerable gridpotentials-of the order of 4o or even 8o volts ;in these conditions it was impossible to use,unchanged, the component of the current from thedetector ; it was necessary first to amplify it. Thevalve manufacturers, therefore, especially inEurope, brought out types of valve which gavegreat control of volume with only small variationsin grid potential-thus producing valves whichwere indeed of variable amplification but which hadby no means the fundamental properties indicatedby Ballantine, who had stressed the importance ofa curve of the 2nd degree. These AVC valves, on.the contrary, had curves of at least the 3rd degree.In the last year or so the designs have been some-what improved, but the writer's point is that evenvalves with a 2nd degree characteristic are boundto cause pre -detection, which favours all thephenomena of modulation of the incident wave byan undesired frequency and renders the receiver sosusceptible to interference of various kinds that itcannot, as a rule, be worked at its maximumsensitivity.

The writer urges a return to the old ideal ofrigorously straight characteristics over the working


range, and envisages, as the ideal method of AVC,a hexode made by introducing, between thecathode and the control grid of an existing pentode,a slope -regulating grid. The slope would besteeper the less negative the grid, and could becompletely annulled by a certain (fairly high)negative potential on the grid. He has not beenable to make such a valve, but preliminary tests onexisting hexodes (designed for an entirely differentmode of use), in the space -charge -grid connection,illustrate the possibilities of his method. Thus,starting from the condition of maximum ampli-fication, the variation is at first feeble and thengrows rapid-exactly the contrary to what happenswith an ordinary variable -mu valve arrangement.This -property of the space -charge -grid controlmethod avoids the need for delayed action. More-over the fact that the variable slope is obtainedwithout departing from a first -degree characteristicenables the AVC to be applied also, if desired, tol.f. stages (a very useful possibility for certainpurposes, including automatic contrast control)whereas " the use of a parabolic characteristic inl.f. is absolutely out of order " ; even the carefullydesigned Cossor D.D. Pen. " introduces, mathe-matically, a certain amount of distortion, not agreat deal in actual fact, but troublesome never-theless."

Finally, the writer refers to criticisms by Bruck(Mazda), one being that it is difficult to leave apositive space -charge grid in the immediate neigh-bourhood of an oxide -coated cathode, owing to therapid deterioration of the latter. The writerapproves of the proposal to remedy this by havingtwo grids between the control grid and the cathode,one negative, to which the slope control is applied,and the other positive, to give the electrons thedesired velocity and to reduce the internal resist-ance. He disagrees with Bruck's objection thatsuch a valve as he proposes would possess aninsufficient grid swing.1422. [Glow] DISCHARGE TUBES IN RADIO SETS.-

Heinze and Pohle : Miram. (Electronics`,January, 1936, Vol. 9, pp. 42-43.) BasedOH the papers dealt with in 3855 of 1935and 613 of February.AUTOMATIC SHARP TUNING [" AutomaticTuning Correction "].-R. ,Wigand : MurphyRadio Ltd. (Funktech. Monatshefte,January, 1936, pp. 21-23.) Based on theWireless World description of the MurphyA 28 C receiver (4th Oct. 1935, pp. 38o-382)and Power's description of the device (inset,25th Oct. 1935)BROADCAST RECEPTION IN FLATS [and theReport of the Royal Institute of BritishArchitects].-( World-Radio, 7th Feb. 1936,Vol. 22, p.

BROADCAST AND GRAMOPHONE DISTRIBUTIONIN HOTELS, INNS AND INSTITUTIONS[Waldorf-Astoria and German Systems].-F. Linke. (Funktech. Monatshefte, January,1936, PP. 33-35.)

1426. FRENCH LAWS RELATING TO LICENCES FORBROADCAST RECEIVERS .-(Rev. Gen. del'Elec., ist Feb. 1936, Vol. 39, No. 5,pp. 299-20o.)

1423.

1424.

1425-

1427. CONSIDERATIONS IN THE DESIGN OF A HIGH-FIDELITY RADIO -GRAMOPHONE. - W. J.Brown. ( Journ. I.E.E., February, 1936,Vol. 78, No. 47o, pp. 194-212: Discussionspp. 212-228.) See 3425 of 1935.

1428. THE PHILIPS MULTI -INDUCTANCE 535-A RE-CEIVER [for Short, Medium and Long Waves].-(L' Onde Elec., December, 1935, Vol. 14,No. 168, pp. 834-838.)

1429. THE TELEFUN KEN 586 WLK : A NINE -CIRCUIT SIX -VALVE SUPERHETERODYNE RE-CEIVER FOR A.G. MAIris.-Telefunken Com-pany. (Funktech. Monatshefte, January,1936, pp. 38-40.)

THE THREE -VALVE " SINGLE -SPAN " SUPER -HET : EXPERIENCES AND IMPROVEMENTS.-H. J. Wilhelmy. (Funktech. Monatshefte,January, 1936, pp. 27-30.) See also 1849,2257 and 3428 of 1935.

THE " BY REQUEST " RECEIVER [CrystalReceiver on Modern Lines].-H. F. Smith.(Wireless World, 7th Feb. 1936, Vol. 38,

1430.

1431.

1432.

1433.

1434.

pp. 130-132.)

" FUNKTECHNISCHE SCHALTUNGSSAMMLUNG "[with Circuits, and Data for Testing andMaintenance, of Current German BroadcastReceivers : to be extended Year by Year :Book Review].-E. Schwandt. (E.T.Z.,2nd Jan. 2936, Vol. 57, No. 1, p.

MEASUREMENTS ON RECEIVERS : PRINCIPLESAND UNITS.-A. L. M. Sowerby. (WirelessWorld, 21st Feb. 1936, Vol. 38, pp. 196-198 :to be concluded.)

A RADIO RECEIVER FOR THE PRIVATE PLANE[or as Emergency Receiver for CommercialAircraft : Type 17A].-J. E. Corbin. (BellLab. Record, January, 1936, Vol. 14, No. 5,pp. 161-164.) See also 1465.

AERIALS AND AERIAL SYSTEMS1435. THE BROADCAST ANTENNA [Field Results

with Tower Aerials : Data on TheirEfficiency, Base Voltage, Base Loss :Practical Design Considerations and Cost :Lighting Requirements : Directive Systems :Wire Aerials " definitely Outmoded " : etc.].-A. B. Chamberlain and W. B. Lodge.(Proc. Inst. Rad. Eng., January, 1936,Vol. 24, No. 1, pp. 1-35.)

SOME COMMENTS ON BROADCAST ANTENNAS[Constant Phase and Current preferable toSinusoidal Distribution : Optimum Heightfor Non -Fading Aerial greater than 5/8thsof Wavelength and dependent on Frequencyand Ground Conductivity : Suggested Com-bination of Vertical Constant -Phase -and -Current Aerial with High -Angle SuppressorSystem to eliminate High -Angle Lobe].-R. N. Harmon. (Proc. Inst. Rad. Eng.,January, 1936, Vol. 24, No. pp. 36-47.)See also 3431 of 1935.

1436.


1437. A CRITICAL STUDY OF THE CHARACTERISTICSOF BROADCAST ANTENNAS AS AFFECTED BYANTENNA CURRENT DISTRIBUTION [Analysisof Vertical Wire with Sinusoidal Distribution :with Capacity Hat : " Sectionalised," " Con-stant Current," " Elevated," Franklin -Typeand " Decreased Velocity " Aerials :" Cylinder " Aerial].-G. H. Brown. (Proc.Inst. Rad. Eng., January, 1936, Vol. 24,No. I, pp. 48-81.) Most of the results aresomewhat disappointing : " for heights ofthe order of a half -wavelength, it is hard tofind anything better than a straight verticalwire." The Franklin -type aerial seems themost promising.

1438. RADIATION RESISTANCE OF AERIALS. -G. W. 0. H : Moullin. (Wireless Engineer,February, 1936, Vol. 13, No. 149, pp. 57-58.)Editorial on the paper referred to in 997 ofMarch, and 1439, below.

1439. RADIO RESISTANCE OF [Transmitting andReceiving] AERIALS [Most Economical AerialDesign : Current Distribution not the samein Reception and Transmission].-E. B.Moullin. (Nature, 1st Feb. 1936, Vol. 137,p. 195: short note on recent I.E.E. paper.)See also 1438, above.

1440. ON THE CURRENT DISTRIBUTION IN A LOOPAERIAL [Expression for Amplitude of Cur-rent].-D. Taylor. (Proc. Phys. Soc., ist Jan.1936, Vol. 48, Part i, No. 264, pp. 111-117.)

In continuation of previous work (2279 of 1935),more detailed experiments are here described, andthe theory is extended. An expression for themagnitude of the current at any point in the loopis calculated and tested with a circular loop aerial." The effect of damping is considered, and theprobable value of the attenuation constant iscalculated."

1441.

1442-

1443.

1444.

tit

DIRECTIONAL PROPERTIES OF SHORT-WAVEFRAME AERIALS [Variation of Current withDirection of Propagation and Wavelength :Frame Current Zero in only One Position].-L. S. Palmer. (Nature, 15th Feb. 1936,Vol. 137, p. 278: preliminary letter.)

DISCUSSION ON " THE CURRENT -LOADINGCAPACITY OF EARTH ELECTRODES."-Taylor.( Journ. I.E.E., February, 1936, Vol. 78,No. 470, p. 248.) See 136 of January.VIBRATIONS OF POWER LINES IN A STEADYWIND. III-THE FREE VIBRATIONS OF AHEAVY STRING.-R. Ruedy. (CanadianJourn. of Res., January, 1936, Vol. 14, No. 1,Sec. A, pp. 16-24.) For II see 551 ofFebruary.

VALVES AND THERMIONICSTHE MOTION OF ELECTRONS IN ELECTRICAND MAGNETIC FIELDS, TAKING INTO CON-SIDERATION THE ACTION OF THE SPACECHARGE [Case of the Cylindrical Condenser :Application to Magnetrons].-S. J. Braude.(Physik. Zeitschr. der Sowjetunion, No. 6,Vol. 8, 1935, pp. 667-674: in English.)

The usual method for solving equations (1) . . .

SS ENGINEER 213

(4) leads, as in the plane case, to non-lineardifferential equations of the second and third order.Attempts to find a solution of these equations havenot so far been successful. Let us use the methoddescribed in our paper " [3437 of 1935]. Equation 7is obtained and solved by the method of successiveapproximations. By the first approximation theformula 12' given by Langmuir is obtained ; thesecond gives the wavelength formula A = 13 r50/H,compared with Okabe's A = 13 000/H, whilewith the third approximation this becomesA = 17 950/H. These results are compared withthe experimental results (Table r) of Slutzkin andSteinberg, for magnetrons with diameters ro, 6 and3 mm respectively (1929 Abstracts, p. 326). Theincrease of XII as the anode voltage is increased isexplained by the electric field of the space chargebecoming of negligible effect, at a very high anodepotential, compared with the external electricfield caused by the anode voltage. The data forXH obtained by the second and third approxima-tions, compared with Table 1, show the limits ofthese solutions.

1445.

1446.

1447.

ON THE MOTION OF ELECTRONS IN CROSSEDELECTRIC AND MAGNETIC FIELDS WITHSPACE CHARGE, and ON THE " CUT-OFF " INTHE PLANE MAGNETRON WITH SPACE CHARGE.-Tonks : Braude. (See 1389 and 1390.)" DIE ELEKTRONSTRoMUNG IN DER BREMS-

oHRE " [Electron Flow in the Brake -Field Valve].-F. H. Lange. (Berlin Thesis,1935: at Patent Office Library, London :Cat. No. 75765.)VALVES FOR SHORT [Ultra -Short and Micro-]WAVES.-M. G. Scroggie. (Wireless World,14th Feb. 1936, Vol. 38, pp. 156-158.)Including a comparative table of data of" acorn " valves (triodes and pentodes) andstandard types of American valves, and ofHivac " Midget " and standard types : theX41 triode-hexode is also discussed.

1448. DIODE FREQUENCY CHANGERS [Theoreticaland Experimental Investigation : Com-parison with Other Converters (e.g. Octodes) :No Loss of Gain when Oscillator Harmonicis used for Ultra -Short -Wave Reception :etc.].-M. J. 0. Strutt. (Wireless Engineer,February, 1936, Vol. 13, No. 149, pp. 73-80.)For more mathematical treatments see 1934Abstracts, p. 614.INPUT RESISTANCE OF VACUUM TUBES ASULTRA -HIGH -FREQUENCY AMPLIFIERS.-W. R. Ferris. (Proc. Inst. Rad. Eng.,January, 1936, Vol. 24, No. i, pp. 82-107.)

" The most important effect of the transit timeon tube characteristics proves to be a seriousin -phase or power component of the chargingcurrent to the control grid. This may be expressedas a conductance which will be shown in this paperto vary as the square of the frequency and thus tobecome of enormous importance as the frequencyis increased [g, = Ks.f21-2]. The grid/cathodecapacity varies scarcely at all with frequency inpractical cases. This paper presents a physicalexplanation of the effect, an experimental verifica-tion of the theoretically derived formulas, and

1449.

214 THE \VIRELES

measurements of the magnitude of the effect invarious commercial tubes." The magnitude ofthis input conductance gg is such that it is theprincipal limitation for amplifiers at frequenciesaround ioo Mc/s, and it seriously affects theamplification even at 15 Mc/s. The plate resistanceof s.g. valves is also found to vary with the fre-quency, but constitutes a negligible amount of thetotal loss in the circuit. This work was presented inpart in the papers dealt with in 1934 Abstracts,p. 563, and 1873 of 1935.

1450. ANALYSIS OF THE EFFECTS OF SPACE CHARGEON GRID IMPEDANCE [Extension of ExistingTheory of Transit -Time Phenomena inHigh -Vacuum Diodes to High -AmplificationTriodes and Tetrodes with Parallel PlaneElectrodes].-D. 0. North. (Proc. Inst.Rad. Eng., January, 1936, Vol. 24, No. 1,pp. 108-136.)

The present treatment is limited to triodes withgrounded plate and tetrodes with grounded screengrid. A special section deals with temperature -limited diodes (eqn. 16) : " here we have a curiousreversal of the situation for space -charge -limiteddiodes, for which the inset of h.f. effects is mani-fested by a lowering of the conductance and adecrease in the cold capacity." This equationindicates how a temperature -limited diode canproduce negative -resistance oscillations at veryhigh frequencies. " Contrary to Benham's views,this author feels that a suitable theory of Bark-hausen-Kurz oscillations can be built upon a basisof complete temperature limitation."

1451. THE VARIATION OF INTER -ELECTRODECAPACITY IN THERMIONIC VALVES [Measure-ments by Special Bridge Circuit showCapacity Change between Any Two Elec-trodes to be Not Solely a Function ofMagnitude of Current flowing : etc.: Appli-cation to Constant -Frequency Oscillators].-D. A. Bell. (Marconi Review, Nov./Dec.1935, No. 57, pp. 18-27.)

Among the conclusions reached are the following :for stable -frequency oscillators the proper procedureis to arrange the grid coupling so that the proportionof grid -capacity change transferred to the tunedcircuit balances the anode -capacity change (for thisand another reason, valves of low amplification arelikely to be more suitable than those of very highamplification) : a definite objection to the use ofa dynatron oscillator is suggested (p. 27) : theimportance of avoiding grid current, when measuringinput capacity, is stressed. The reasons for usinga bridge method of measurement are given and themethod described : it is quite practicable tomeasure Miller effect on this bridge.

1452. ON THE FUNCTIONING OF THE TRIODEVALVE (INSTRUCTIONAL NOTES).-U. Ruelle.(Alta Frequenza, December, 1935, Vol. 4,No. 6, pp. 688-713.) A simple approximatetreatment, the triode being considered inconnection with practical circuits (foramplification, detection, retroaction andsuper -regeneration, oscillation and modula-tion) selected to illustrate its properties.


1453. NEW KINDS OF VALVE DESIGN : THE DANISH" RENODE " [and Its German Forerunner] :AMPLIFIER VALVES WITH ELECTRON LENSES :AMPLIFICATION BY SECONDARY EMISSION[Zworykin].-(Funktech. Monatshefte, Jan-uary, 1936, pp. 11-12.)

The German forerunner of the " renode " was anexperimental valve with an electron lens, in whichthe deflecting plates produced a change in the" current sharing " between two adjacent anodes :the " renode " (ioi I of March) has only one anode,the deflecting plates themselves taking up some ofthe electron stream under the action of the signal.A new type of electron -optical amplifier valve isunder developinent in Germany : here the cathodeis enclosed in a control cylinder with two slits whichpass out two sharp electron rays in oppositedirections, each to its own anode. Modulation bythe control cylinder alters only the electron density,not the shape, of the rays. An ideally straightcharacteristic is given. Such valves, " unlike therenode," could be made with current values, slope,etc., corresponding to those of the usual valves :on the other hand they offer the possibility ofobtaining internal resistances down to about200 ohms.

1454. VALVES FOR CAR RECEIVERS, WITH " COPPER-EIFILAR " CATHODES [Heating Watts reducedto 1.5 and 1.75 compared with Usual 2.6,by making use of Small Infra -Red RadiationCoefficient of Copper compared with Nickel].-(Funktech. Monatshefte, January, 1936,pp. 9-11.) The type numerals of thesevalves are followed by " Cu-Bi " (Tele-funken) or " Cu " (Volvo). It is suggestedthat the new type of cathode will be ofimportance in the general field and not onlyfor car receivers.

1455. USING THE 6L7 TO IMPROVE SUPERHETPERFORMANCE : ADAPTING PRESENT RE-CEIVERS TO USE THE NEW METAL -TUBEMIXER [Unlike Most Metal Valves, 6L7 and6H6 (Duo -Diode) are Radically Differentfrom Glass Valves and offer Special Advan-tages].-(QST, February, 1936, Vol. 20,pp. 48-49.) See also end of 1407.

1456. THE 307A POWER PENTODE [particularly forMobile Transmitters : Separate Leads forAll Three Grids].-E. A. Veazie. (Bell Lab.Record, January, 1936, Vol. 14, No. 5,PP. 150-153.)

1457. SOME THERMAL METHODS OF MEASURINGLoss OF POWER IN VACUUM TUBES [and theComparative Merits of Air -Chamber,Thermometer, Bolometer, Thermopile andThermocouple Indicators].-F. P. Cowan.(Review Scient. Instr., January, 1936, Vol. 7,No. I, pp. 13-16.)

" It is important to note that workers on ultra-high frequencies have used thermocouples almostexclusively for their measurements of efficiency, andwithout treatment of the errors due to inducedcurrents. These are large at 16.9 megacycles andmight be larger still at higher frequency."


1458. POTENTIAL RELIEFS OF HIGH -VACUUM ANDGAS -FILLED DETECTORS AND RECTIFIERS.-H. Gottmann. (Funktech. Monatshefte,January, 1936, pp. 5-8.) Extension of workreferred to in 57o and 571 of February.

1459. NOISES OF AMPLIFYING VALVES AND THEIRACCURATE MEASUREMENT [and the Separa-tion of the Different Components].-W.Jacobi and W. S. Pforte. (E.T.Z., 2nd Jan.1936, Vol. 57, No. r, p. 18 : summary only.)

1460. ELECTRIC FORCE versus CENTRIFUGAL FORCE[Effect, on Characteristics, of Revolution ofValve about External Axis].-C. T. Dozier.(Phys. Review, 15th Jan. 1936, Series 2,VOL 49, No. 2, pp. 207-208: abstract only.)

1461. THERMIONIC EMISSION FROM TUNGSTEN ANDTHORIATED TUNGSTEN FILAMENTS.-W. B.Nottingham. (Phys. Review, ist Jan. 1936,Series 2, Vol. 49, No. 1, pp. 78-97.)

From the author's summary :-The electronemission from pure and thoriated tungsten filamentshas been investigated as a function of the appliedpotential over the entire range from a few voltsretarding to I 400 v accelerating, including carefulstudies at zero field. . . . The energy distributionof electrons was found to be deficient in slowelectrons. An empirical reflection factor . . .

represents the observed data for all temperaturesand all states of activation for the thoriatedtungsten. Analysis shows that there is no disagree-ment between the new results and the experimentsof Germer [1925]. Zero -field Richardson plotsshow that the reflection effect alters both theA and b of Richardson's equation. . . . The valueof A indicates a negative temperature coefficientof the work function . . . which has been verifiedby an independent experiment. New data on theelectron emission in accelerating fields are givenfor many states of activation, showing that thereare large deviations from the Schottky mirrorimage theory [1914]. Becker's patch theory [see2656 of 1935] is discussed briefly and a simplerstrip theory is developed which serves to representthe observed data as an empirical result. Thisanalysis shows that the patch theory is not suitableto explain the reflection effect, since it is independentof the state of activation of the filament.1462. A LONG -LIFE ACTIVATED HOT CATHODE

[75 -Micron Tungsten Wire wound on Molyb-denum Core (chemically removed later) :Resulting Wire wound on 450 -MicronTungsten Spiral: Whole dipped in (e.g.)Barium Hydroxide].-Philips Company.(French Pat. 788 959, pub. 21.10.35 : Rev.Gen. de l'Elec., 25th Jan. 1936, Vol. 39,pp. 29-30 D.)

1463.

DIRECTIONAL WIRELESSVISUAL COURSE INDICATOR FOR RADIONAVIGATION.-I. M. Vekslin. (IzvestiaElektroprom. Slab. Toka, No. 12, 1935, pp.54-63.)

A theory of the operation of the Kramar typevisual course indicator (see 1934 Abstracts, pp.443-444) is given, together with an account oftests carried out on a number of models utilising

SS ENGINEER 215

this principle. As a result of this investigation anew output circuit was developed (Fig. 4) which,it is claimed, ensures a smooth movement of thepointer of the indicating instrument. In thiscircuit two detector valves are interposed betweenthe secondary of the output transformer and theindicating instrument, one on each side of theinstrument, which must then be centre -tapped forthe return current. As a practical modificationthe detectors feed into a centre -tapped potentio-meter, across which an ordinary centre -zero milli -voltmeter is connected. It is pointed out that forsatisfactory operation of the system the speed oftransmission of the radio beacon signal should beincreased to at least 25o dash -dots per minute.This speed is not so high as to preclude the use ofaural reception as an auxiliary method of control.

THREE - DIMENSIONAL DIRECTION - FINDER[Linear Aerial perpendicular to Frame :Arrangement gives Zero Deflection whenAerial is directed towards Emitter].-W.Runge : Telefunken. (Hochf:tech. u. Elek:akus., December, 1935, Vol. 46, No. 6, p.215 : German Patent 617 487, dated24.1.1934.)

1465. A RADIO COMPASS FOR AIRCRAFT [for Usein conjunction with Type I7A Receiver].-C. B. Aiken. (Bell Lab. Record, January,1936, Vol. 14, No. 5, pp. 165-168.) Forthe 17A receiver see 1434.

1466. NEW AVIATION RADIO RECEIVES 2 SIGNALSON SAME FREQUENCY [Weather Reportsand Beacon Signals on Same Wave simul-taneously].-(Sci. News Letter, 25th Jan.1935, Vol. 29, p. 52.) No details are given.

1467. DETRIMENTAL EFFECT OF HORIZONTAL POR-TIONS OF LONG LOW LOOP AERIAL (FORAERIAL NAVIGATION) SUPPRESSED BY ONE ORMORE IMPEDANCES SYMMETRICALLY INTRO-DUCED.-Philips Company. (French Pat.789 403, pub. 29.10.1935: Rev. Gen. del'Elec., 25th Jan. 1936, Vol. 39, pp. 30-31 D.)

1468. NEW SYSTEM FOR THE DETECTION OF OB-STACLES AT SEA [Highly Directive Trans-mitter and Receiver for Very Short Waves,mounted as Far Apart as possible on Portand Starboard Bows and Screened fromDirect Action].-Comp. Gen. de T. S. F.(French Pat. 788 795, pub. 16.10.35 : Rev.Gen. de 25th Jan. 1936, Vol. 39, p.29 D.) For a micro -wave " radio feeler "tested on the linear " Normandie," seeWireless World, 8th Nov. 1935, p. 491.

ACOUSTICS AND AUDIO -FREQUENCIES1469. A MAINS -DRIVEN SOUND -MEASURING IN-

STRUMENT WHICH CAN BE CALIBRATED[suitable for Tests on Loudspeakers, Micro-phones, etc.].-F. Dohnal. (Funktech.Monatshefte, January, 1936, pp. 1-5.)

The calibration is performed by excitation, withsuperposed a.c. voltage, of the perforated auxiliaryelectrode in front of the diaphragm of the condensermicrophone. The indicating instrument is of the" profile " (end -on) type, so that readings can bemade, if necessary, through a telescope, thuskeeping the observer out of the sound field.

1464.


1470. DUST FIGURES FORMED BY [Sound Wavesfrom] AN ELECTRIC SPARK [Show Inter-ference : Constant Wavelength of Sound :Patterns regarded as Sound Maps illustrateAcoustic Phenomena, such as DirectiveProperties of Cones for High -FrequencySound].-A. E. Bate. (Proc. Phys. Soc.,1st Jan. 1936, Vol. 48, Part r, No. 264, pp.178-182 : Discussion p. 183.)

1471. PARAMETRIC COUPLING BETWEEN STATIONARYACOUSTIC WAVES [Analogous to the Couplingproducing Interaction between Vertical andAngular Oscillations in Elastic Pendulum :Harmonics and Sub -Harmonics due toAcoustic Properties of Gas].-G. Gorelik.(Tech. Phys. of USSR, No. 2/3, VOL 2, 1935,pp. 248-251: in French.)

1472. TRANSIENT PHENOMENA IN ELECTROACOUSTICTRANSMISSION SYSTEMS [Apparatus forDirect Determination of Audibility andNature of Transients].-W. Biirck, P.Kotowski, and H. Lichte. (Zeitschr. f.tech. Phys., No. 12, VOL 16, 1935, pp. 519-522.)

The apparatus is stated to be particularly suitablefor transients due to variation of attenuation withfrequency. The scheme is shown in Fig. 2. Anaudio -frequency note of constant amplitude isintroduced into an amplifier whose amplificationis varied by an automatic switch working throughan adjustable time circuit ; the sinusoidal noteproduced can thus be switched in with full ampli-tude or with any desired time constant. The noteis received either direct or through the transmissionsystem under investigation. The method ofmeasurement is described (§ 2). A resonant circuitfollowed by a valve was first investigated (§ 3 ;frequency curve and transient time-constant/fre-quency curve Fig. 3). Systems with two or moreresonance peaks are discussed ; Fig. 4 showscurves for an amplifier with 24 resonant circuits,for frequencies corresponding to different positionson the resonance peaks. Echo effects can beimitated by such systems. Fig. 5 shows a fre-quency curve for the whole of a transmissionsystem.

1473. INVESTIGATIONS OF RAPIDLY VARYINGACOUSTIC PHENOMENA [With the OctaveFilter].-F. Trendelenburg and E. Franz.(Zeitschr. f. tech. Phys., No. 12, Vol. 16,1935, pp. 513-516.)

An octave filter was used to separate the com-ponents of different frequencies in the soundsinvestigated : the components in each octave wererecorded with an oscillograph (circuit Fig. I).A number of examples of sounds (vowels, syllables,etc.) with the corresponding oscillograms are given.Transient phenomena for various musical instru-ments are also shown (Figs. 6-8).

1474. THE " AUDIO -FREQUENCY SPECTROMETER,"A FREQUENCY ANALYSER WITH EXTREMELYHIGH ANALYSIS SPEED AND DIRECTLYVISIBLE SPECTRUM.-E. Freystedt. (Zeitschr.f. tech. Phys., No. 12, Vol. 16, 1935, pp.533-539)

This electrical frequency analyser has a large

number of filters connected in parallel (Fig. r).The frequency range under investigation is dividedlogarithmically among them ; their output voltagesare directly visible as an amplitude spectrum on thescreen of a cathode-ray oscillograph. The timerequired for the analysis is reduced to abouto.I sec., the time of the switch -on transient of thenarrowest filter. The construction and action ofthe instrument are given in detail in § 2 (filtercircuit Fig. 2). Photographic records can be madeand the whole instrument (Fig. 4) can be workedoff the mains. Many examples of spectra of noises,vowels, words and musical notes are given.1475. THE FILTER -COUPLED INDUCTIVE GLOW -

DISCHARGE OSCILLATOR.-Kock. (See 1383.)

1476. AN AUDIO -FREQUENCY HETERODYNE OSCIL-LATOR.-M. I. Rodman. (Journ. of Tech.Phys. [in Russian], No. 6, Vol. 5, 1935, pp.1093-1096.) Covering a frequency rangeof from I to 85o cycles ; the change infrequency after 3o minutes of operation isless than I cycle. The harmonic contentof the output is less than 3%.

1477. FEEDING PUSH-PULL AMPLIFIERS : METHODSOF PHASE REVERSAL.-COCkillg. (See 1415.)

1478. A 45 -WATT LOW -LOADING AMPLIFIER [usingDA 3o Power Triodes].-(Television, Novem-ber, 1935, Vol. 8, No. 93, pp. 676, 678, 686and 688.) An AC/Pr, choke -coupled (" whichpartially accounts for the absence of hum "),forms the second stage.OUTPUT TRANSFORMER RESPONSE [for ClassA Power Amplifier : Calculation of Fre-quency Characteristic from Turn -Ratio,Winding Resistances and Leakage Induc-tance : Methods of Reducing the LeakageCoefficient].-F. E. Terman and R. E.Ingebretsen. (Electronics, January, 1936,Vol. 9, pp. 30-32.)

1480. LONG-DISTANCE TELEPHONY : RECENTDEVELOPMENTS AND TREND OF FUTUREADVANCES - EDDY CURRENT LOSSES INSCREENED CONDUCTORS.-A. C. ThRITLiS.(Electrician, 24th Jan., 1936, Vol. 116, pp.91-92: summary of I.E.E. paper.)

1481. SEPARATION OF SOUND RECORD FROM PICTURERECORD IN SOUND FILMS [Post -Synchronisa-tion Methods to avoid Poor Sound Recordingby Portable Apparatus in Open Air : " Uni-kat " Discs : American Use of AcetylCellulose for Discs : etc.].-P. Hatschek.(Funktech. Monatshefte, January, 1936, Supp.PP. 5-6.)

1479.

1482. THE " MECHANICAL KLIRR-FACTOR " [Non -Linear Distortion] IN SOUND -ON -FILM RE-CORDING AND REPRODUCTION [and ItsReduction by Film Stabilising Devices].-P. Hatschek. (Funktech. Monatshefte,January, 1936, Supp. pp. 3-4.)

1483. ON THE SELF OSCILLATIONS OF A FREELYSUPPORTED CIRCULAR PLATE.-G. OStTOB-mov. ( Journ. of Tech. Phys. [in Russian],No. 6, Vol. 5, 1935, pp. 947-957.)

A theoretical investigation of the oscillations of

a uniform isotropic circular plate freely supportedround the edge. This investigation is of practicalvalue since experiments show that sounding boardsof musical instruments oscillate in a similar way.The problem has already been solved by Poissonfor la (Poisson's ratio) = I. In the present paperdifferent methods are used and a general solutionfor various values of IL is found.

In the first part of the paper a differential equationfor the oscillations of the plate is given and methodsare indicated for solving this. A number of tablesand graphs are prepared in order to facilitate thenecessary calculations, and a table is added in whichthe shape of the oscillating plate is shown forvarious values of its mechanical constants. Inthe second part of the paper two important factorsare calculated, one determining the volume dis-placed by the oscillating plate and the other thekinetic energy of the plate.

1484. THE " CELLULOPHONE " [Photocell Organ :the Synthesis of Undamped and DampedSounds].-P. Toulon. (L'Onde Elec.,December, 1935, Vol. 14, No. 268, pp. 822-833.) For previous work see 199 of January.

THE AUDIBILITY OF THE CONTROL PROCESSESIN AUTOMATICALLY CONTROLLED AMPLIFIERSAND FILM PURE -TONE SYSTEMS.-W. BUrck,P. Kotowski and H. Lichte. (Zeitschr. f.tech. Phys., No. 12, VOL 16, 1935, pp. 522-525.)

The type of circuit discussed is shown in Fig. 2,where volume control is applied to the transmissionof a gramophone record. If the control is to beinaudible, its time -constants must satisfy thephysiological conditions imposed by the ear (§ 2,Fig. 3 ; § 3, Fig. 4), which allow o.3 and 5o msecfor the control switch -on and -off time -constantsrespectively. Some technical control methods(e.g. the use of exponential valves in certain circuits)do not permit these values to be reached ; theaudibility of the detector effect (Fig. 5) must thenbe taken into account for control switch -on effects,and the non-linear distortion for the control switch -off time -constant (" klirr " factor Fig. 6).

1486. THE PROBLEM OF CONTRAST EXPANSION[for Broadcasting, Sound Films, and Gramo-phones] .-Lamparter. (See 14 r 7.)

STUDY OF LEVEL REGULATORS AND " ANTI -FADING " DEVICES.-G. Espinasse. (Ann.des Postes T. et T., December, 1935, Vol. 24,pp. 1098-2117.)

1488. AUTOMATIC AUDIO -FREQUENCY VOLUMECONTROL DEPENDING ON THE EXTERNALNOISE LEVEL [for Cinemas, Public Address,Restaurants, etc.].-H. Boucke. (Funktech.Monatshefte, January, 1936, pp. 32-32.)

1489. COMPREHENSIBILITY OF SPEECH IN NOISYRooms [Difficulties of Speaking and Hearingby Telephone : Requirements of Electro-acoustic Transmission System to overcomeThem].-C. A. Hartmann and W. Janovsky.(Zeitschr. f. tech. Phys., No. 12, Vol. 16,1935, pp. 580-584.)

Fig. 1 shows how the intensity of intelligiblespeech increases with the noise in the room where


1485.

1487.

SS ENGINEER 217

it is produced ; Fig. 2, how the intelligibility varieswith speech intensity for various degrees of sur-rounding noise ; Figs. 3 and 4 illustrate the intel-ligibility of reproduced sound as a function of itsstrength in very noisy rooms. Matching the trans-mitted frequency band to the noise is discussed in§ 3, and the effect of non-linear distortion in § 4.Measures for overcoming the noise difficulties mustbe applied at the emitter (§ 5) ; non -carbon micro-phones are preferable. Headphones are recom-mended for the receiving end.

1490. EXPERIMENTS ON THE IMPROVEMENT OFTELEPHONE SYSTEMS FOR NOISY ROOMS[in Ships or Aeroplanes : Microphones andTheir Properties].-K. Kruger and W.Willms. (Zeitschr. f. tech. Phys., No. 12,Vol. 16, 1935, pp. 585-590 : DiscussionP. 590.)

Contact microphones are discussed in § ;

oscillograms of vowels for various positions ofcontact with the speaker's head are given in Figs.1-5. These indicate that the best positions toavoid the transmission of noise are on the cheekor lower jaw. The electrodynamic microphoneused for the measurements is described (Fig. 6) ;it suffers very little from non-linear distortion.Microphones held before the mouth are also dis-cussed (§ 2) ; they give a good ratio of speech tonoise intensity. Screening from extraneous soundat the receiver by headphones, screening caps orcomplete closing of the ear is illustrated by Figs.7, 8. In the discussion a Rochelle salt microphoneis recommended.

1491. APPARENT DURATION OF SOUND PERCEPTIONAND MUSICAL OPTIMUM REVERBERATION[and Application to Broadcasting Studios,etc.].-S. Lifshitz. (Journ. Acous. Soc.Am., January, 1936, Vol. 7, No. 3, pp.213-221.) The full paper, a summary ofwhich was referred to in 3517 of 1935.

1492. THE EFFECT OF DISTANCE IN THE BROAD-CASTING STUDIO [Great Importance of theRatio of Direct and Reverberant Waves,depending on Distance of Source fromMicrophone].-A. V. Rabinovitch. (Journ.Acous. Soc. Am., January, 1936, Vol. 7,No. 3, pp. 299-20.) The great differencein the effects observed over an electricallink and by direct listening is attributed tothe absence, in the former case, of thebinaural effect which ordinarily enables thelistener to isolate, subconsciously, the re-flected waves from the direct. See also1502 of 2935.

2493. ON THE OPTIMUM CONDITIONS IN SOUNDFILM THEATRES.-A. I. Indlin and A. N.Kacherovich. ( Journ. of Tech. Phys. [inRussian], No. 6, Vol. 5, 1935, pp. 928-938.)

An investigation into the relationship betweenthe acoustic properties of a film studio and a filmtheatre, the two being parts of one acoustic system.It is suggested that in order to obtain the optimumconditions in a theatre for all films, certain standardsof reverberation should be adopted for film studios.


1494. SOME REMARKS ON THE ACOUSTIC CON-DITIONS IN A FILM STUDIO [and the Effectof Scenery].-A. I. Indlin and A. N.Kacherovich. (Journ. of Tech. Phys. [inRussian], No. 6, Vol. 5, 1935, pp. 939-946-)

An investigation into the effect of scenery onthe acoustic properties of a studio. It is statedthat it is not correct to treat the scenery as anadditional damping factor, but that the studio andthe scenery should be regarded as acousticallycoupled rooms.1495. MODERN SOLUTIONS OF ACOUSTIC PROBLEMS

OF ROOMS AND CONSTRUCTION IN BUILDINGSERECTED FOR BROADCASTING.-H. J. vonBraunmiihl. (Zeitschr. f. tech. Phys., No.12, Vol. 16, 1935, pp. 571-575.)

Some ideas for the attainment of any requireddegree of resonance in broadcasting studios arehere described. Fig. 1 shows the acoustic intensityreached by various forms of performance, whichthe walls between studios must therefore be capableof reducing to an intensity below that of the lowestsound to be transmitted. Fig. 2 shows the pathsby which sound from one studio can reach another ;the best insulation is given by elastic (i.e. non-rigid) connection of the walls. Figs. 3 and 4illustrate how this may be attained for roomsvertically above and below one another (floatingconstruction, on springs) ; photographs of con-structional details are given (Figs. 5-8) and de-scribed. Fig. 9 gives curves of attenuation ob-tained in this way ; Fig. io shows how footstepsare silenced by different methods of floor construc-tion. The remaining figures illustrate the effectsof different wall coverings.1496. THEORY OF THE ABSORPTION OF SOUND BY

WALL COVERINGS.-E. Wintergerst. (Zeit-schr. f. tech. Phys., No. 12, Vol. 16, 1935,PP. 569-571.)

It is shown theoretically that homogeneousmaterials cannot absorb all frequencies equally.An arrangement of a plate of absorbing material,containing slits or holes, at a finite distance from asolid wall (Fig. is found however to give equalhigh absorption of all frequencies over a largerange. The necessary dimensions and propertiesof the arrangement can be determined theoreticallyif the desired frequency curve is given.1497. THE VERTICALLY STRATIFIED WALL AS AN

ACOUSTIC -MECHANICAL CHOKE.-E. Meyer.(E.N.T., December, 1935, Vol. 12, No. 12,pp. 393-400: abstract only in Zeitschr. f.tech. phys., No. 12, VOL 16, 1935, pp. 565 -566.)

The wall and its equivalent electrical circuit(Fig. z) are described in § 1 ; the individual stratacorrespond to the inductances, the air cushionsbetween them to the capacities. Internal frictionis neglected. The voltages at input and outputof the filter circuit correspond to the pressuresbefore and behind the wall respectively. Experi-mental tests of the correctness of the equivalentcircuit are described in § 2. Condenser micro-phones are used to measure the energy density inthe diffuse acoustic fields before and behind thewall ; the acoustic attenuation is then measuredin decibels by the usual formula. The circuit

scheme of the apparatus used is shown in Fig. 2 ;Fig. 3 shows the logarithmic valve voltmeter.Fig. 4 gives a typical record of the attenuation asa function of frequency. The attenuation producedby various types of wall is considered in § 3 (Figs.5, 6) ; the formula given in § 1 for the limitingfrequency is verified. For sufficiently high fre-quencies the wall behaves as if its density wereuniform. The phase velocity of sound propagationis discussed in § 4 and illustrated by Fig. 7 ; thelow frequencies are transmitted more rapidly thanthe high ones.

The limiting frequency is not sharply defined ;this is explained in § 5 as due to transversal reson-ances in the air cushions. Introduction of anabsorbing material in the place of air improves thesharpness of the frequency curve (Fig. 8, whichmay be compared with Fig. 9 for an electricalchoke). A lattice (Fig. io) placed in the air cushiongives the attenuation curves of Fig. r1. Anothermethod of reducing the transversal resonances isto insert absorbing material round the rim of thewall only (Fig. 12). To obtain walls of efficientinsulation but light weight the limiting frequencymust be made as low as possible, either by increasingthe mass of the separate walls or the thickness ofthe air cushions.

1498.

1499.

THEORY OF TRANSMISSION OF PLANE SOUNDWAVES THROUGH MULTIPLE PARTITIONS[explaining the Much Higher Loss thanthrough Corresponding Single Partition].-A. L. Kimball. (Journ. Acous. Soc. Am.,January, 1936, Vol. 7, No. 3, pp. 222-224.)

THE PHYSICAL BASES AND NEW RESULTS OFTHE FIGHT AGAINST NOISE [SummarisingReport].-K. W. Wagner. (Zeitschr. f. tech.Phys., No. 12, Vol. 16, 1935, PP. 544-554.)

Subjects included in this general survey of thepresent position are :-§ 2 : Hearing in the healthyand the deaf. § 3 : Fundamentals of noise measure-ment. §§ 4, 5 : Intensity of sound and the sensa-tion of loudness. § 6 : Intensity of compositesounds. § 7 : Noise meters. § 8 : Traffic noise.§ 9 : Noise in dwelling -houses. § Jo : Absorptionof sound. § 11 : Noise in machines and workshops.

1500. " KLANGE AND GERAUSCHE " [Tones andNoises] and " SOUND."-F. Trendelenburg :F. R. Watson. (Journ. Acous. Soc. Am.January, 1936, Vol. 7, No. 3, p. 231 and232: Book Reviews.)

PHOTOTELEGRAPHY AND TELEVISION1501. A SEALED COLD -CATHODE OSCILLOGRAPH

FOR LOW EXCITING VOLTAGES [8-10 kV :Application to Television].-F. A. Becker.(Arch. f. Elehtrot., 2oth Dec. 1935, Vol. 29,No. 12, pp. 873-876.)

An auxiliary discharge is used to provide ions forthe main discharge so that it starts at a low voltage.The intensity of the light spot can be controlledwithout distortion by inserting a stop before thecathode. Electrostatic lenses are used for beamconcentration (Malsch. & Becker, 290 of 1935).Fig. i shows an oscillograph incorporating thesefeatures ; Figs. 2 and 3 show external photographstaken with it. A recording velocity of 25o km/sec.

1936 THE WIRELESS ENGINEER 219

can be attained. The tube is suitable for televisionreception (intensity characteristic Fig. 4) and as ascanning emitter ; Fig. 7 shows a photograph of afluorescent screen obtained with it. The new fea-tures referred to above may also be used foroscillographs with divided vacuum.

1502. RECENT DEVELOPMENTS OF THE CATHODE-RAY OSCILLOGRAPH [for use with Mains :Methods of Eliminating Mains Disturbances].-J. Dantscher. (Arch. f. Elektrot., lothDec. 1935, Vol. 29, No. 12, pp. 833-841.)

The conditions under which the oscillograph maybe worked off the mains are shortly given in § 2 ;fluctuation of mains and accidental deviationof the spot by the mains connection are the chiefsources of trouble. In § 3 the time -base for usewith mains is discussed (see also 1933 Abstracts,p. 635) ; a screened -grid valve (characteristicsFigs. I, 2) is used as the charging valve to preventmains fluctuations from reaching the electrodes.Fluctuation and stabilisation of mains voltageare dealt with in § 4 and illustrated in Fig. 3 ; thearrangement and screening of the mains trans-former is described in § 5. Fig. 4 shows the circuitof the AEG-oscillograph as built for connection tothe mains. The calibration of the time -base isshortly discussed in § 6.

1503. THE IMPORTANCE OF THE CHOICE OF SUIT-ABLE FLUORESCENT MATERIAL FOR TELE-VISION SCANNING WITH THE CATHODE -RAY -LIGHT SCANNER.-W. Schnabel. (Zeitschr.f. tech. Phys., No. 1, Vol. 17, 1936, pp. 25-27.)

When a film is scanned by the light of the spoton a cathode -ray -tube screen, the fluorescence decayprocess must have ended by the time the cathode rayhas advanced by the amount of the spot diameter ;with a system of 40 000 picture elements and25 frames/sec. this means by the end of 10-6 sec.A screen material already exists with a decay timejust over 10-6 sec. (815 of 1935-see Fig. 24,Schleede's preparation), and the writer has nowexamined whether, with this and other commercialmaterials, the spot -light film -scanning system ispracticable for high -definition television ; using forhis tests a cold -cathode -tube scanner with which,for a pass -band of 700 kc/s, a change from whiteto black produced a voltage difference of 15 volts.

To his surprise he found that the effect of after-glow on the resulting picture was not so marked asmight be expected. Thus with the Schleede pre-paration referred to above the image shown inFig. 2 was obtained, for 65o 000 elements/second :the slight defects in this image are attributed toinadequate sharpness of receiver spot and tomotion of the film during the taking of the photo-graph, as well as to the effect of after -glow : thelens employed produced, moreover, marked dis-tortion at the edges. Even a zinc sulphide materialwith a total decay time of several minutes gave theimage shown in Fig. 4 with 200 000 elements/second, and though this image is obviously un-workable owing to the after -glow effect, it never-theless shows that here again that component ofthe fluorescent light which is active in the scanningprocess has a far more rapid decay than the visiblecomponent. Further investigations are in progress

in the Aachen Institute. Several other suitablematerials have already been found.

1504. THE GOLDMARK ELECTRON -OPTICAL PRO-JECTION SYSTEM, and TRANSMISSION ANDSYNCHRONISATION WITH THE GOLDMARKELECTRON -OPTICAL SYSTEM.-P. C. Gold -mark. (Television, Nov. and Dec. 1935,Vol 8, Nos. 93 and 94, pp. 661-664 and675 : pp. 703-704 and 705.)

A combination of cathode-ray and mirror -drumtechnique. The principle is as follows :-thepicture is scanned in (say) five vertical zones, eachzone being made up of horizontal lines, as long as thewidth of the zone, produced by a cathode-ray tubeof only 1 inch screen diameter whose fluorescentspot moves only in the horizontal direction. A lensprojects the moving spot through a mirror -drumon to the viewing screen. The mirror -drum has atleast as many mirrors as there are zones in thepicture. Each mirror is tilted differently so that therotating drum spreads the moving spot into zonesjoining each other. Thus the very small cathode-ray tube provides the necessary high scanningfrequency, while " the projecting drum enlargesthe image to any desirable size at comparativelysmall loss of light." The intrinsic screen brightnessis calculated and compared with that for ordinarydirect enlargement.

1505. REPORT ON ELECTRON -OPTICAL QUESTIONSCONNECTED WITH TELEVISION.-E. Briicheand W. Schaffemicht. (E.N.T., December,1935, Vol. 12, No. 12, pp. 381-392.)

A review of the part played by electron -optics intelevision (see these Abstracts, passim) ; emphasisis laid on a clear exposition of fundamental prin-ciples rather than on technical details. The sub-divisions are : § 2. The cathode-ray oscillograph andits electrode system. § 3. Its application totelevision. § 4. The " image transformer " (whichtransforms an optical image into an electronimage) as a photographic instrument. § 5. Com-parison of the recent electron photographic methods(at the emitter) of Zworykin (e.g. 1934 Abstracts,p. Jo') and Farnsworth (e.g. 207 of 1935). Acomprehensive list of literature references is given.

1506. ON THE CONCENTRATION OF ELECTRONBEAMS IN GAS -FILLED TUBES.-I. L. Sokol-skaya. (Tech. Phys. of USSR, No. 1, Vol. 3,1936, pp. 28-38: in English.)

It has been observed that while electrons emittedfrom a cathode in a vacuum tube travel towards theanode in a divergent stream, the stream of electronsin a tube filled with gas at a low pressure is con-centrated into a narrow beam and takes the shapeof a standing wave with well-defined nodal points.In this paper an account is given of experimentscarried out with tubes filled with argon and,alternatively, mercury vapour.

The object of the experiments was to determine theeffect on the focal length of the beam (distance be-tween the nodal points) of the following factors :gas pressure, electron -current intensity, velocityof the electrons, and potential acquired by the wallsof the tube. A number of experimental curves areshown and a theoretical interpretation of the resultsobtained is given. One of the conclusions reached is


that the main requirement imposed on these tubeswhen used in television, namely that the size ofthe light spot on the screen must remain constantwhile the intensity of light is varying, cannot be met,since the focal length of the beam varies with theintensity of the electron current.

1507. ELECTRON -OPTICAL BENCH [for investigatingPerformance of Experimental ElectrodeSystems and facilitating Design of Tele-vision Tubes].-G.E.C. Research Labora-tories. (Nature, 15th Feb. 1936, Vol. 137, pp.281-282: short note only from PhysicalSociety Exhibition.)

1508. TYPE 6001 CATHODE-RAY TUBE FOR TELE-visiox.-Mullard Company. (Wireless En-gineer, February, 1936, Vol. 13, No. 149,p. 64 : paragraph only.)

1509. CATHODE-RAY TELEVISION WITH AUTO-MATIC SYNCHRONISATION.-R. Barthelemy.(L'Onde Elec., December, 1935, Vol. 14,No. 168, pp. 794-803.) For a Comptes Ren-dus Note see 2717 of 1935.

1510. SYNCHRONISATION IN TELEVISION, AND RE-SISTANCE -CAPACITY AMPLIFIERS [Survey,with Mathematical Analysis of Amplifiersfor Synchronising or Image Signals].-S.Bertolotti. (Alta Frequenza, January, 1936,Vol. 5, No. 1, pp. 5-41.)

Generation of the synchronising signals both formechanical and electronic scanning, including adisc generator developed by the E.I.A.R. for usewith the latter type of scanning, designed for 18o -line pictures but applicable (without increasingthe diameter) to 24o lines : calculation, for knownvalve data, of resistance -capacity amplifiers, usingthe Poisson probability summation, with oscillo-grams showing various distortions of the synchroni-sing signals resulting from inappropriate values ofthe amplifying circuit constants (e.g. Fig. 21, wherethe effect of an inductance of ro nm in series withthe anode resistance RI, in Fig. 22, is illustrated) :the modem system of superposed synchronisingand image signals, methods of separation at thereceiver, and the synchronisation of the receivingcathode-ray tube (with photographs of results offalse synchronisation). The writer concludes thatif the sight and sound waves are always separatedby a definite frequency difference, it should bepossible to construct a receiver to tune to thecomposite programme by a single adjustment,all other adjustments, including synchronisation,being made once and for all inside the receiver.

1511. TELEVISION : PROGRESS [in 1935] TOWARDS AREGULAR HIGH -DEFINITION SERVICE.-L. E. C. Hughes. (Electrician, 31st Jan.1936, Vol. 116, p. 141.)

1512. HIGH -DEFINITION TELEVISION FOR THEALEXANDRA PALACE : DETAILS OF THESIGNALS RADIATED BY THE BAIRD ANDTHE MARCONI-E.M.I. TRANSMITTERS.-(Tele-vision, November, 1935, Vol. 8, No. 93, pp.631-636 and 688.) See also 239 and 24o ofJanuary.

1513. THE LONDON TELEVISION TRANSMITTER[Description of Marconi-E.M.I. Apparatusat Alexandra Palace].-(Wireless World,31st Jan. 1936, Vol. 38, pp. 103-104.)

1514. THE OPENING CEREMONY FOR THE NEWBERLIN TELEVISION TRANSMITTER, DECEM-BER, 1935.-(Funktech. Monatshefte, January,1936, SuPP pp. 1-2.)

1515. TELEVISION ABROAD [" Mechanical -Optical "versus " Electrical" Scanning in U.S.A. :Plans in Great Britain].-(Funktech. Monats-hefte, January, 1936, SuPP pp. 4-5.)

1516. TELEVISION RECEIVER FOR HIGH DEFINITIONand TELEVISION RECEIVER FE IV [FirstCommercial Type offered by Telefunkenfor Berlin Ultra -Short -Wave 18o -Line Pro-grammes].-H. 0. Roosenstein : P. Besson.(Wireless World, 14th Feb. 1936, Vol. 38,pp. 160-162: L' Onde Elec., December,1935, Vol. 14, No. 168, pp. 783-793.)

1517. THE TELEVISION RECEIVER OF THE FERNSEHA. G.-G. Schubert. (Funktech. Monats-hefte, January, 1936, Supp. pp. 2-3.)

1518. TELEVISION IN NATURAL COLOURS [possibleby Combination of Kerr Cell and Prismswith Photocells of Various Spectral Sen-sitivities ?].-B. Bennett. (Television, De-cember, 1935, Vol. 8, p. 739.)

1519. A CHECK ON QUALITY : A NOTE ON THE[B.B.C.] HIGH -FREQUENCY TEST CARD.-(Television, December, 1935, Vol. 8, No. 94,p. 706.)

1520. MEASURING DELAY ON PICTURE -TRANS-MISSION CIRCUITS.-E. P. Felch. (BellLab. Record, January, 1936, Vol. 14, No. 5,Pp 154-157)

1521. THE EFFECT OF OXYGEN, CARBON DIOXIDE,NITROGEN AND MERCURY ON BARIUM ANDPOTASSIUM PHOTOCELLS.-A. V. Afanasjevaand J. I. Lunkova. ( Journ. of Tech. Phys.[in Russian], No. 6, Vol. 5, 1935, pp. incur006.)

An account of experiments carried out withbarium and potassium photocells in order todetermine the relationship between the photo-electric current and the amount of gas absorbed bythe light-sensitive surface. A number of experi-mental curves are shown and a theoretical interpre-tation of the results obtained is given. It was foundthat when oxygen or mercury vapour act on abarium surface the photoelectric current rises withthe increase of the gas absorption, passes through amaximum, and falls off ; under the influence ofnitrogen the current increases up to the saturationpoint, while carbon dioxide has the opposite effectand causes a decrease in the current. In the caseof a potassium surface the current passes througha maximum with any one of the above gases withthe exception of mercury vapour, the effect of whichis to decrease the current.

741


1522. EFFECT OF OXYGEN UPON THE PHOTOELECTRICTHRESHOLDS OF METALS.-H. C. Rentschlerand D. E. Henry. (Journ. Opt. Soc. Am.,January, 1936, Vol. 26, No. I, pp. 3o-34.)Extension of the work dealt with in 1933Abstracts, p. 169.

1523. PHOTOELECTRIC WORK FUNCTION OF BARIUM[Measured by Fowler's Method of ReflectionCoefficient of Barium deposited on Glass].-R. J. Cashman and N. C. Jamison. (Phys.Review, 15th Jan. 1936, Series 2, Vol. 49,No. 2, pp. 195-196: abstract only.)

1524. THE PHOTOELECTRIC WORK FUNCTION OFCA AND PHOTOEMISSION FROM NONHOMO-GENEOUS SURFACES [measured by Fowler'sMethod with Layers produced by SingleDistillation : Surface rendered Nonhomo-geneous by Heating].-N. C. Jamison andR. J. Cashman. (Phys. Review, 15th Jan.1936, Series 2, Vol. 49, No. 2, p. 2o1:abstract only.)

1525. ON ALKALI FILMS OF ATOMIC THICKNESS ONPLATINUM [Method of DepoFiinn and ExactMeasurement : Measurement of Change ofPhotoelectric Sensitivity of Platinum withIncreasing Thickness of Potassium Film].-H. Mayer. (Zeitschr. f. tech. Phys., No. ii,Vol. 16, 1935, PP. 451-454.)

1526. THE OPTICAL CONSTANTS AND PHOTOELECTRICEMISSION OF POTASSIUM [Confirmation ofIves's Theory].-H. E. Ives and H. B. Briggs.(Science, 6th Dec. 1935, Vol. 82, p. 541 :summary only.)

Optical constants of potassium, recently obtained,strikingly confirm the theory (1932 Abstracts, p.io2) that the photo -emission from thin films on aspecular metallic base is conditioned by the opticalabsorption of the photo -active material. Thesharp maximum of emission in the ultra -violet ispredicted at the right wavelength, and the enormousenhancement of emission, for polarisation withthe electric vector parallel to the plane of incidence," is an immediate consequence of the unusualoptical properties of the alkali metal."1527. CONTRIBUTION TO THE MEASUREMENT OF

SPHERICAL CADMIUM CELLS USED FORREGISTRATION [of Solar and Sky Radiation].-M. Bender. (Physik. Zeitschr., 1st Feb.1936, Vol. 37, No. 3, pp. 107-110.)

Continuation of previous work (see 1934 Abstracts,p. 374, r -h col.-Bender & Kruger). Here thebehaviour of various cells of different spectralsensitivities is described and records of the dailyvariation of ultraviolet solar and sky radiation inthe Black Forest are given (Fig. 1). It is foundthat the cadmium' vapourises, probably owing tothe effect of the radiation. This may be the cause ofthe inconstancy of the cells. Aigon cells do notshow this phenomenon.

1528. BARRIER LAYERS AND PHOTOELECTRICITY[Rectifying Action and Photoelectric Effectincrease with and are connected by ContactResistance].-J. Roulleau. (Comptes Rendus,loth Feb. 1936, Vol. 202, No. 6, pp. 470-472.)

For previous work see 1214 of March. The

221

writer now points out that in a copper -oxiderectifier the photoelectric effect and the rectifyingaction at the contact surface are independent of thespecific resistance of the mass of Cu20. Curves aregiven to demonstrate that they each increase withthe contact resistance, and it is deduced that theyare connected by means of that resistance. Ex-planations are given for the different result obtainedby other authors (see 1933 Abstracts, p. 400-Borissow, S. & W., and 2048, 3583 of 1935-Juse).1529. THE INTERNAL PHOTOELECTRIC EFFECT IN

SEMI -CONDUCTORS [Calculation of ElectronDistribution Function under Illuminationby Monochromatic Light : Calculation ofDember Effect].-H. Frohlich. (Physik.Zeitschr. der Sowjetunion, No. 5, Vol. 8,1935, pp. 501-510: in German.)

1530. THE DEPENDENCE OF SENSITIVITY OF THESELENIUM -SULPHUR RECTIFIER PHOTO-ELECTRIC CELL ON THE OBLIQUITY OF THEINCIDENT LIGHT, AND A METHOD OF COM-PENSATION THEREFOR.-G. P. Barnard.(Proc. Phys. Soc., ist Jan. 1936, Vol. 48,Part I, No. 264, pp. 153-162 : Discussionpp. 162-163.)

" When a restricted portion only of the surfaceof the cell is illuminated, it is found that thesensitivity. increases as the centroid of the illumin-ated area is displaced from the centre of the cell."The method of compensation comprises " the useof a central stop arranged at a suitable distanceabove the surface of the cell." Details of the ap-paratus and method used are given and distributioncurves for cells under various conditions are shown.1531. BECQUEREL EFFECT AND PHOTOCHEMICAL

SENSITIVITY OF SOME FLUORESCENT COLOUR-ING MATTERS [depend on Oxidation and Re-duction].-Cecile Stora. (Comptes Rendus,5th Feb. 1936, Vol. 202, No. 5, pp. 408-410.)See also 1115 of March.

1532. THE YIELD OF QUANTA IN THE FORMATIONOF COLOUR CENTRES IN KBR-CRYSTALS[with Absorption of Light in U -Centres :Temperature Effect : Connection with Sat-uration Value of Stationary PhotoelectricPrimary Currents].-R. Hilsch and R. W.Pohl. (Gottinger Nachrichten, Math.-Phys.Kl., No. 19, Vol. I, 1935, pp. 209-214.)

1533. THE KERR EFFECT OF NITROBENZOL [Im-proved Purification Method and Use ofGlass and Rustless Steel Cell, withoutPacking Material, gives Undistorted andCleaner Interference Pattern, 'Kerr Constant5% Greater and Conductivity over Io TimesLess than Previous Best].-F. Gabler andP. Sokob. (Zeitschr. f. tech. Phys., No. 1,Vol. 17, 1936, pp. 11-17.)

MEASUREMENTS AND STANDARDS1534. MEASUREMENTS OF WAVELENGTHS AND PO-

TENTIAL WITH A LECHER SYSTEM WITHVARYING CHARACTERISTIC IMPEDANCE.-N. N. Malov. (Physik. Zeitschr. der Sow-jetunion, No. 6, Vol. 8, 1935, pp 595-601 :in German.)

A two-part, telescopic Lecher system, with thethermoammeter permanently at the end of the


thinner, movable pair of tubes, has certain ad-vantages such as the elimination of the open-endeffect. Measurements, however, show that thegeometrical length of such a system is always smallerthan a quarter -wavelength. This discrepancy isattributable to the differing characteristic impe-dances of the two parts; and seems likely to invali-date the usual formula for the input voltage (

product of characteristic impedance W and ammetercurrent J). The theory of such a system is thereforeinvestigated, with the result that a factor p is found,less than unity, by which the product must bemultiplied in order to give the input voltage. Theconstitution of p is shown in eqn. 23, and inputvoltages thus measured, for four different Lechersystems, agreed reasonably well with measurementsmade by a special voltmeter (3591 of 1935) ; foraccurate measurements the ammeter correction atvery high frequencies should have been known.

Another equation (15) is derived, for finding therelation between the wavelength and the totallength x of the system : here m = 2u/A and x = xi+x1, and for equation 15 to hold good x,, thelength of the movable part, must be less thane.75x1 if, as is assumed, the ratio Ws/ WI of thecharacteristic impedances of the moving and fixedparts respectively is 1.4 : I. If this ratio is madesmaller the proportion of x2 to xi can be increased.This equation, and also equation 21 for the inputimpedance, is tested for the four systems mentionedby comparison with measured values (dotted curvesin Figs. 3-6) : agreement is very good.1535. AN ABSOLUTE METHOD FOR THE MEASURE-

MENT OF THE EQUIVALENT RESISTANCE OFAN OSCILLATORY CIRCUIT [and Its PossibleApplication to High Resistance Measure-ment at Ultra -High Frequencies].-M. Boella.(Alta Frequenza, December, 1935, Vol. 4,No. 6, pp. 647-656.)

The method consists in observing the capacitychange necessary to reduce the voltage of theresonant circuit to 1/ V2 of its resonance value, thecircuit being inductively coupled to a constant -frequency generator. Possible errors and theiravoidance are discussed. See also 1536, below.1536. MEASUREMENT OF THE EQUIVALENT RESIS-

TANCE OF OSCILLATORY CIRCUITS [NegativeResistance (Dynatron) Method satisfactoryonly up to 3 Mc/s unless with SpeciallyDesigned Tetrode, Short Connections, etc.].-M. Ferrario. (Alta Frequenza, December,1935, Vol. 4, No. 6, pp. 657-667.)

The dynatron-method results were compared'with measurements by the reactance -variationmethod proposed by Boella (1535, above). Above3 Mc/s several phenomena enter into the formermethod and produce values lower than those givenby the latter, the errors increasing with increase'of frequency and also with increase of the equiva-lent resistance : at 18 Mc/s, for a resistance of20 000 ohms, they may be 2o% or more.1537. DETERMINATION OF SMALL TEMPERATURE

COEFFICIENTS [of Coils and Condensers] ATHIGH FREQUENCIES [Automatic Recording].-L. Rohde. (Zeitschr. f. tech. Phys., No. 12,Vol. 16, 1935, pp. 635-637.)

The scheme of this arrangement for recording

temperature variations of coils and condensersat high frequencies is shown in Fig. I and described.An interference method is used ; the frequency ofbeats due to changes in the adjustment is used tomake the record ; the conversion of this frequencyinto a current depending entirely on frequencyand not on amplitude is made by a direct -readingfrequency -meter. The sensitivity can be adjustedto that required for the measurement in progress.The method of measurement is given in § 2 and somerecords are shown (§ 3, Figs. 3, 4).1538. AUTOMATIC ADJUSTMENT OF COMPLEX COM-

PENSATION- AND BRIDGE -CIRCUITS WITHVARYING -PHASE NULL MOTOR [with Ap-plication to Tests of Condensers, Cables andHigh -Voltage Apparatus : Measurement ofImpedances of Coils and Mutual Inductances:Circuit Diagrams].-W. Geyger. (Arch. f.Elektrot., zoth Dec. 1935, Vol. 29, No. 12,pp. 842-85o.)

1539. AN ADJUSTABLE PRECISION STANDARD OFPHASE DIFFERENCE.-G. B. Engelhardt.(Bell Lab. Record, January, 1936, Vol. 14,No. 5, pp. 158-160.)

1540. A TRANSMISSION LINE IMPEDANCE MEASUR-ING SET.-A. L. Drabkin. (Izvestia Elektro-prom. Slab. Toka, No. 12, 1935, pp. 66-72.)

A description of a set developed for the purpose ofmeasuring the various constants of a transmissionline (or aerial) for wavelengths between 13 and ioometres. The set comprises (Fig. 3) a push-pulloscillator inductively coupled to a push-pullamplifier, the output circuit of which is connectedeither to the circuit to be measured or to an equiva-lent circuit. Methods are indicated for measuringthe sending -end resistance and reactance of theline, the resistance per unit length, and the ratioof the velocities of propagation of electromagneticenergy in the air and in the line. Formulae are alsoderived for calculating the characteristic impedanceand the attenuation constant of the line from theabove data plus the capacity per unit length, whichhas to be determined separately.1541. ERRORS OF COMMERCIAL HIGH -FREQUENCY

AMMETERS [10 mA-2o A, at Ultra -ShortWavelengths down to 3 m].-H. Kruse and0. Zinke. (E.T.Z., znd Jan. 1936, Vol. 57,No. 1, p. 12 : summary only.)

Commercial types of thermo-ammeters and hot-wire and hot -strip instruments were checked againstan " optical " ammeter (in which the light radiationfrom a vacuum -enclosed wire, with low skin effectand heated by the current under measurement, wasmeasured by photocell and galvanometer). Thenatural frequencies of all the thermo-ammeterswere found to be in the micro -wave region, so thatthey had no important effect for the wave -rangein question ; this applied also to hot-wire and hot -strip meters with one heating element. Thequestion of the errors due to skin effect wasexamined, for round, strip, and tubular heatingelements. The latter should have their walls asthin. as possible, at any rate thinner than thepenetration depth at the highest frequency involved.In the practically important region of small errorsthe strip was inferior to the wire, but superior in


the region of large errors. The errors calculatedfrom skin -effect considerations agreed well withthe measured over-all, errors, so that it was con-cluded that these thermal instruments had errorsdue only to skin effect, provided they had nopotential with regard to their surroundings (in thetests they were all earthed on one side). " Theseerrors can be reduced, in a wide -range instrument,by the use of tubular elements, and in a narrow -range instrument by the use of high -frequencycurrent transformers."

1542.

1543.

1544.

DISCUSSION ON " A VALVE AMMETER FORMEASURING SMALL ALTERNATING CURRENTSOF RADIO FREQUENCY " [Extension of Fre-quency Range above 10645 c/s by Air -CoredChoke].-Barlow. (Journ. I.E.E., February,1936, Vol. 78, No. 470, p. 193.) See 274 ofJanuary : also Electrician, 31st Jan. 1936,Vol. 116, p. 13o, for paragraph on a com-mercial model.

THE COMPENSATED THERMOCOUPLE AMMETER[Mathematical Analysis of TemperatureDistribution and Heat Flow in ElectricalConductors : Development of the Electro-thermic Ammeter of Thermocouple Type :Application of Theory to Shunt Design].-W. N. Goodwin. (Elec. Engineering, Jan-uary, 1936, Vol. 55, No. 1, pp. 23-33.)

THE DIRECT -CURRENT MEASURING TRANS-FORMER [a New Device giving More AccurateMeasurement of Large Continuous Currents,including Currents with Rapid Variationsand with Superposed A.C.].-O. E. Mike.(E.T.Z., 9th Jan. 1936, Vol. 57, No. 2,PP. 37-39.)

1545. A NEW, HIGH -SENSITIVITY MAINS -DRIVENVALVE VOLTMETER WITH TWO RANGES[covering 0.03-3.0 V.,f1 Input Capacityonly 15 µµF : Input Resistance 5 X io6Ohms : 25 cis -3o Mc/s].-M. von Ardenne.(Elektrot. u. Masch:bau, 19th Jan. 1936,Vol. 54, No. 3, pp. 30-32.)

The second, " insensitive " range is obtained byvoltage -division, without the disadvantages usuallyattributed to this process-namely the reduction ofhigh input resistance (the main advantage of thevalve voltmeter) and non -uniform behaviour todifferent frequencies. The success of the circuitadopted is due to the use, as grid leak, of the twovery high resistances R1 and R2 with the switchingarrangement (using a special low -capacity switch)shown in the diagram. In the " sensitive "position 1 of the switch the voltage to be measuredis applied directly to the grid, and the unwantedcapacity of the grid circuit is not appreciablyincreased by the very small capacity CI (of theorder of 5 cms) or by the switch capacity : more-over, the path to the grid, whose length has apredominant influence on the upper frequencylimit, can be kept very short. In the ' insensitiveposition 2 there is a two -fold voltage division ; forthe lower frequencies an ohmic division (dependingon the ratio of .R2 to R3) and for the higher acapacitive division by the capacities C1 and C2plus the grid capacity. C1 is so chosen that theinput capacity remains unchanged on switching

SS ENGINEER 223

from one range to the other, and C, so chosen thatthe voltage division at high frequencies is the sameas the ohmic voltage division at low frequencies.The new -type valves with separate grid lead -outcan be used to their best advantage, since only thegrid capacity has to be kept small and not, as insome other arrangements, the anode capacity aswell.

1546.

1547.

1548.

1549.

VOLTAGE MEASUREMENT AT VERY HIGHFREQUENCIES-I [Investigation of the SeriesCondenser/Diode Method : Calculation ofLeakage, Impedance, and Electron -InertiaErrors].-E. C. S. Megaw. (WirelessEngineer, February, 1936, Vol. 13, No. 149,pp. 65-72.) The method should give accur-ate measurements up to about ioo MO anduseful estimates up to about 1000 Mc/s(voltages about so v or higher).

AN ELECTROSTATIC HIGH -VOLTAGE VOLT-METER FROM 0.5 TO 35 kV WITH POINTERINDICATION [on Balance Principle].-W.Steubing. (Physik. Zeitschr., ist Jan. 1936,Vol. 37, No. i, pp. 32-35.)A NEW CONSTRUCTION OF THE VACUUMDUANT ELECTROMETER.-G. Hoffmann : B.Zipprich. (Physik. Zeitschr., ist Jan. 1936,Vol. 37, No. I, pp. 35 : 36-38.)

SOME THERMAL METHODS OF MEASURINGLOSS OF POWER IN VACUUM TUBES . -Cowan.(See 1457.)

I 550 . ERROR -FREE THERMAL POWER METERS.-W. Bader. (Arch. f. Elektrot., loth Dec.1935, Vol. 29, No. 12, pp. 809-833.)

A general account, with diagrams, of circuitswhich contain two thermoelements and make powermeasurement without error possible is given in§ II. The properties of these circuits are investi-gated theoretically ; the conditions are found whichgovern the adjustment of the magnitudes of theelements for maximum sensitivity and minimumabsorption of power by the meter itself (§§ III, IV).The theory of the measurement of voltage andcurrent, in addition to that of power, is describedin § V. § VI refers to an arrangement for switch-ing in various measurement ranges ; § VIII givesdata of experimental tests, including one of fre-quency independence at audio frequencies.

1551. THE INFLUENCE OF THE RELATIVE HUMIDITYOF THE AIR ON THE LOSS ANGLE OF INSU-LATING MATERIALS AT HIGH FREQUENCIES[explaining Discrepancies between DifferentMeasurements of Same Material].-H.Schwarz. (E.T.Z., 2nd Jan. 1936, Vol. 57,No. I, pp. 7-9.)

Author's summary :-" Loss factor measure-ments on various insulating materials [includingceramic types, mica, etc. : frequencies 105-107 c/s]give, in the presence of a relative humidity higherthan 5o%, higher values of loss factor than thosegenerally accepted. The additional losses increaseto varying extents with increasing relative humidityand with decrease of frequency. Conductivitymeasurements show that only a small part of theadditional loss can be attributed to surface leakageover the test samples ; it is chiefly a question

E 2

224 THE WIRELES

rather of true dielectric processes. The finish 'of the surface of the sample is also of considerableimportance. The writer urges that when a lossfactor is given the external conditions in which thetest is made should be mentioned." In highrelative humidities, even the time elapsing betweentaking the sample out of the drying oven andcarrying out the measurement has a considerableinfluence on the result.

1552. MEASUREMENTS OF DIELECTRIC Loss FROM5 X 108 TO I x 108 cis.-L. Rohde.(Zeitschr. f. tech. Phys., No. 12, Vol. 16,1935, PP. 637-639.)

The method used is an extension of one pre-viously given for lower frequencies (Rohde &Schwarz, 1934 Abstracts, p. 509). The circuit isshown in Fig. I ; the theory of the method isdiscussed. The sensitivity can be adjusted at will.Results for various materials are tabulated ; solidsshow no difference in behaviour from that at lowerfrequencies, but some liquids, particularly mix-tures, show anomalous behaviour, with absorptionin certain ranges of frequency. Pure organicliquids have sometimes very small loss. A methodof measurement in which the frequency is con-tinuously variable is being developed.

1553. THE ACCURATE MEASUREMENT OF THEBREAKDOWN AND CURRENT/VOLTAGE CHAR-ACTERISTICS OF LIQUID DIELECTRICS WITHDIRECT POTENTIALS [Equipment measuringfrom io-16 to bo-8 Ampere at Voltages0-40 kV].-Baker and Boltz. (Review Scient.Instr., January, 1936, Vol. 7, No. 1, pp.50-52.)

1554. THE VARIATION OF THE DIELECTRIC CON-STANTS OF ANISOTROPIC LIQUIDS WITHFIELD STRENGTH AND FREQUENCY.-W.Kast. (Zeitschr. f. tech. Phys., No. II,Vol. 16,1935, pp. 475-479.)

1555. A METHOD FOR MEASURING HIGH RESIST-ANCE [of Order of pp" ohms : Bridge Methodusing Rapidly Charged -and -Discharged Con-denser as Balancing Resistance].-Van denAkker and G. M. Webb. (Review Scient.Instr., January, 1936, Vol. 7, No. I, pp.44-46.)

1556. [Theoretical] INVESTIGATIONS ON THE NATUR-AL ELASTIC VIBRATIONS OF QUARTZ PLATESEXCITED PIEZOELECTRICALLY.-R. Bech-mann. (Zeitschr. f. tech. Phys., No. 12, Vol.16,1935, pp. 525-528.)

For previous work see 503/4 of 1935. Here themoduli of elasticity, their temperature coefficients,and the piezoelectric moduli of quartz are cal-culated for four plane sections through the elasticwave surface, using the known numerical values ofthe corresponding principal moduli. Three dif-ferent velocities of propagation of elastic waves arefound for each direction, and in general also threenatural elastic vibrations of thin plates, dependingon the orientation of the plates to the wave surface.This surface has two curves along which the tem-perature coefficient of the elastic moduli becomes


zero. Fig I shows curves illustrating the variationof the different quantities discussed.

1557. X-RAY STUDIES OF CRYSTALS VIBRATINGPIEZOELECTRICALLY [DIfIeT0I1Ce in Effectbetween X -Cut and Y -Cut Crystals : Effectof Polishing and Etching : Microphoto-grams].-C. V. Bertsch. (Phys. Review,15th Jan. 1936, Series 2, Vol. 49, No. 2,pp. 128-132.)

1558. PASSAGE OF X-RAYS THROUGH OSCILLATING[Piezoelectric] CRYSTALS [InhomogeneousCrystal Distortion widens Multiply -Re-flected Portion of Beam : Enlargement ofCentral Spot].-R. M. Langer. (Phys.Review, 15th Jan. 1936, Series 2, Vol. 49,No. 2, p. 206: abstract only.) For rele-vant work see Fox & Carr, 1931 Abstracts,p. 57o ; also 3188 of 1935 and 285 of January.

1559. FREQUENCY/THICKNESS CONSTANTS OF THINQUARTZ PLATE CUT PARALLEL TO THEELECTRICAL AXIS, and THIN QUARTZ PLATEOF ZERO TEMPERATURE COEFFICIENT.-H.Yoda. (Journ. I.E.E. Japan, November,1935, Vol. 55 [No. II], No. 568, pp. I000and zoo, : Japanese only.)

1560. THE QUARTZ CLOCKS OF THE PHYSIKALISCH-TECHNISCHE REICHSANSTALT [Recent Ob-servations : Detection of Variation in Earth'sRotation].-Adelsberger. (Zeitschr. V.D.I.,8th Feb. 1936, Vol. 8o, pp. 147-148.)

1561. STANDARD FREQUENCY RADIO BROADCASTINGSERVICE [Details of Times,Modulation, etc.].-Nat. Bureau of Stand-ards. (Journ. Franklin Inst., November,1935, Vol. 220, No. 5, pp. 657-659.)

1562. THE VARIATION OF INTER -ELECTRODECAPACITY IN THERMIONIC VALVES [andConclusions regarding Frequency Stabilisa-tion].-Bell. (See 1451.)

1563. MEASURING INSTRUMENTS : NOTEWORTHYDEVELOPMENTS OF THE PAST YEAR.-A. C.Jolley. (Electrician, 31st Jan. 1936, Vol.116, pp. 130-131.)

1564. THE BALLISTIC GALVANOMETER AND THEMEASUREMENT OF IMPULSES OF EXPON-ENTIAL FORM.-I. Lucchi. (L'Elettrotec.,loth Nov. 1935, Vol. 22, No. 21, pp. 747-750.)

1565. FREQUENCY DIVISION BY THE VIBRATION OFA MECHANICAL OSCILLATING SYSTEM.-Rusakov and Ryabinina. (Tech. Phys. ofUSSR, No. 1, Vol. 2, 1935, pp. 48-60.)German version of the work dealt with in68z of February.

1566. LOW -INERTIA POINTERS FOR MEASURINGINSTRUMENTS [Specially Drawn Thin -WallGlass Tubes : Composite Pointers withMoment of Inertia about 8 Times Smallerthan Aluminium Tube Pointer].-Miehlich.(Zeitschr. f. tech. Phys., No. 1, Vol. 17,1936,PP. 27-31.)


SUBSIDIARY APPARATUS AND MATERIALS1567. DEMONSTRATION OF THE IMAGE ERRORS OF

ELECTRON LENSES WITH IMAGES OF A POINT.-K. Diels and M. Knoll. (Zeitschr. f. tech.Phys., No. 12, VOL I 6,1935, pp. 617-621.)

Authors' summary :-The following errors inelectron -optical images have been produced in theimage of a point, in analogy to the correspondingerrors in ordinary optics : 1. Astigmatism of anobliquely incident beam, investigated with a c -r tubewith a skew electric or magnetic lens and an electronbeam of relatively small diameter (Fig. 1-7).2. Astigmatism of lenses malformed at the side,investigated with a c -r tube with elliptical electricor magnetic lens (Figs. 8,9). Similar errors (due tothe deviating plates) are caused by the electron -optical field of the cylinder lens of the deviatingplates of a c -r tube (Figs. so, II). 3. Apertureerrors of electric tube lenses, investigated byobserving the caustic in the path of the beam of ac -r tube with a lens electrode of diameter smallcompared with that of the beam (Figs. 12-15). 4.Coma, aperture error of obliquely incident beam,investigated with a c -r tube with a skew electrictube lens and a beam of relatively large diameter(Fig. 16). The form of every electron -optical errorshowed good agreement with the correspondingerrors of ordinary optics.

1568. RELATIONS BETWEEN ELECTRON LENS,ELECTRON MIRROR AND CONTROL [of BeamIntensity].-W. Henneberg and A. Reck-nagel. (Zeitschr. f. tech. Phys., No. 12, VOL16, 1935, pp. 621-623.)

Authors' summary :-The single electric lens withnegative lens voltage becomes an electron mirrorwhen the energy of the incident electrons is suf-ficiently small. As the energy decreases, the lensacts successively as a collecting lens, a collectingmirror and a scattering mirror (Fig. I). The pathof the electron beam (Fig. 2) and the curve of focallength of the mirror (Fig. 3) are given for a potentialfield already investigated as a lens. During thechange from lens to mirror, the intensity of thetransmitted beam is controlled.

1569. REPORT ON ELECTRON -OPTICAL QUESTIONSCONNECTED WITH TELEVISION .-BriiChe andSchaffemicht. (See 1505.)

1570. ELECTROSTATIC FOCUSING AT RELATIVISTICSPEEDS [Power Series for Quick Computationof Focal Properties of Common ElectronLenses even when Electron Speed approachesthat of Light].-Hansen and Webster.(Review Scient. Instr., January, 1936, Vol. 7,No. 1, pp. 17-23.). See also 1571.

1571. ION OPTICS OF EQUAL COAXIAL CYLINDERS[Potential Distribution along Axis expressedempirically in Terms of Radius and Separa-tion].-Kirkpatrick and Beckerley. (Re-view Scient. Instr., January, 1936, Vol. 7,No. I, pp. 24-26.) Combined with thetheoretical lens equation of Hansen andWebster (157o), this empirical expression" yields an algebraic formula for lenses ofthis type directly relating object and imagedistances to readily measurable quantities."

1572. ON THE CONCENTRATION OF ELECTRONBEAMS IN GAS -FILLED TUBES. -Sokolskaya.(See 15o6.)

1573. A SEALED COLD -CATHODE OSCILLOGRAPH FORLOW EXCITING VOLTAGES [with AuxiliaryDischarge to provide Ions), and RECENTDEVELOPMENTS OF THE CATHODE-RAY OS-CILLOGRAPH [and the Elimination of MainsDisturbances].-Becker : Dantscher. (See1501 and 1502.)

1574.-A CATHODE-RAY OSCILLOGRAPH FOR THEDIRECT MEASUREMENT OF HIGH -VOLTAGETRANSIENTS [for ioo kV Operating Voltagewithout Use of Potential Divider : withDelayed Impulse Circuits : etc.].-Nuttall.(Journ. I.E.E., February, 1936, Vol. 78,No. 470, pp. 229-234.)

The general features of the tube itself are similarto those of the Burch & Whelpton tube (1933Abstracts, pp. 51 and 339) except for the specialhigh -voltage plates. The variable -capacitancepotential divider shown is used only with the low -voltage plates, when frequent and wide variation ofvoltage of applied impulse is necessary.

1575. CATHODE-RAY RECORDER " REMEMBERS " BYMEANS OF FLUORESCENCE DELAY .-HUll .(Electronics, January, 1936, Vol. 9, p. 36.)See also 715 of February.

1576. THE ELECTRON TELESCOPE [ZWOrykill" Electron Copying Camera " working withInfra -Red, Visible or Ultra -Violet Light :Description : Applications].-Zworykin andMorton. (Electronics, January, 1936, Vol. 9,pp. ro-13.) Based on demonstration toAmerican Association for the Advancementof Science (1092 of March). Among otherthings, the avoidance of " pin -cushion "distortion by the use of a spherically curvedcathode is described.

1577. USE OF A RADIAL -DEFLECTION CATHODE-RAY OSCILLOGRAPH AS A TIME COMPARATOR[with Spiral Time -Base : Time Micro-scope].-Dowling and Bullen. (Nature,15th Feb. 1936, Vol. 137, p. 279.)

1578. A NEW METHOD FOR THE APPLICATION OFLUMINESCENT SCREENS TO GLASS SURFACES[Objections to Liquid Binder and SinteringMethods : New Method using SulphurCoating (from Carbon Disulphide Flame)to pick up Fluorescent Powder and leave itfixed when Sulphur is driven off].-Kohl.(Canadian Journ. of Res., December, 1935,Vol. 13, No. 6, Sec. A, pp. 126-132.) Themechanism may be that the sulphur chargesthe glass surface and causes the binding ofthe powder by electrical forces : or thepowder may be held by surface forces, thesulphur acting as cleaning agent.

1579. THE ABSORPTION OF GASES BY MERCURY[New Results of Interest for VacuumTechnique].-Ptizin. (Tech. Phys. of USSR,No. 1, VOL 2, 1935, pp. 66-78: in English.)


1580. ON THE LIBERATION OF GASES FROM GLASSUNDER THE INFLUENCE OF A HIGH FRE-QUENCY DISCHARGE.-V. I. Romanov andN. A. Tsvetkov. (Journ. of Tech. Phys.[in Russian], No. 6, Vol. 5,1935, pp. 996-999.)

An account of experiments with glow -dischargetubes subjected to the action of h.f. currents.It has been found that under certain ccnditionslarge quantities of CO, and 11, are liberated from theglass walls of the tubes.

1581. OSCILLOGRAPH AMPLIFIERS FOR VERY WIDEFREQUENCY RANGES [0.1 CiS tO 3 X 106c/s and 2 X I 04 c/s (the latter with VoltageAmplification io5)].-von Ardenne. (Wire-less Engineer, February, 1936, Vol. 13, No.149, pp. 59-64.) More extensive account ofthe amplifiers dealt with in 231 of January.

1582. PRODUCTION AND APPLICATIONS OF SHORTCURRENT IMPULSES WITH A VALVE CIRCUIT[Use of Hydrogen Discharge Tube in " Trip "Circuit].-Drewell. (Zeitschr. f. tech. Phys.,No. 12, VOL 16, 1935, pp. 614-617.)

Large condensers at a high voltage can be dis-charged across grid -controlled discharge tubes filledwith hydrogen (" trip " circuit Fig. I). Shortintense pulses of current can thus be produced(circuit Fig. 3) which may be applied to exciteinductors or to feed a mercury lamp for strobo-scopic observations.

1583. THE POSITIVE ION CURRENT TO THE INCAN-DESCENT CATHODE OF A GASEOUS DISCHARGE[Possible Objection to Gvosdover's Methodof Determining the Relation between Elec-tron and Ion Current : Reply].-Druyves-teyn. : Gvosdover. (Physik. Zeitschr. derSowjetunion, No. 5, Vol. 8, 1935, pp. 579-581 : 582-583: in German.) See 3465 of1935.

1584. POTENTIAL RELIEFS OF HIGH -VACUUM ANDGAS -FILLED DETECTORS AND RECTIFIERS .-Gottmann. (Funktech. Monatshefte, Jan-uary, 1936, pp. 5-8.) Extension of workreferred to in 57o and 571 of February.

1585. THE OPERATING CHARACTERISTICS OF SMALLGRID -CONTROLLED HOT -CATHODE ARCS ORTHYRATRONS.-French. (Journ. FranklinInst., January, 1936, VOL 221, No. I,pp. 83-102.)

An investigation of the magnitude and natureof the current in the grid circuit of a thyratronbefore starting. This is found to depend theoret-ically upon the shape of the grid current character-istic before starting, as well as on the grid circuitconstants. The method of measuring the gridcurrent and impedance at the starting point isdescribed. Three types of thyratron were investi-gated ; certain anomalies occurring at low anodepotentials are described in detail. Emphasis isplaced on the significance of the magnitude andsign of the grid impedance. Grid resistances ofroo to 1000 megohms are essential in certain cases.See also Nottingham, 1931 Abstracts, pp. 269 and500.

April, 1936

1586. HIGH -VOLTAGE LOW -CURRENT RECTIFIERTYPE U16, PRIMARILY FOR TELEVISIONCATHODE-RAY TUBES.-Osram. (WirelessEngineer, February, 1936, Vol. 13, No. 149,p. 64 : paragraph only.)

1587. THE ELECTRIC DISCHARGE IN GASES ANDTHE DEBYE-HUCKEL THEORY.-Chiplonkar.(Current Science, Bangalore, January, 1936,Vol. 4, No. 7, p. 481.) " In view of theremarkable success of the theory in the fieldof the kinematics of ions in liquid media,it is not a little surprising that but littleuse of its methods has been made in theanalysis of the phenomena in dischargetubes. . ."

1588. NOMOGRAPHIC REPRESENTATION OF THE SAHAEQUATION [of Thermal Ionisation].-UnsOld.(Zeitschr. f. tech. Phys., No. ii, Vol. 16,1935, pp. 450-461.)

1589. THE COPPER OXIDE RECTIFIER [Formation :Interface Structure : Characteristics : Ther-mal Conductance].-Starr. (Physics, Jan-uary, 1936, Vol. 7, No. 1, pp. 15-19.)

The conditions of formation of the Cu80-rectifierare controlled by the equilibrium diagram of thecomponent substances, which is first discussed.The structure of the interface is investigated withpolarised light. The rectifier characteristics aredetermined by a method free from error due tocurrent heating at the interface. Measurement ofthe thermal conductance of the rectifier showedasymmetry in the direction to be expected fromthe elebtron theory of heat conduction.

1590. SILICATE RESEARCH AND ENGINEERING. -Eitel. (Zeitschr. V.D.I., ilth Jan. 1936,Vol. 8o, pp. 37-41.)

SOME INVESTIGATIONS OF THE TEMPERATUREVARIATIONS OF THE DIELECTRIC PROPERTIESOF TITANIUM- AND TIN -DIOXIDE [used asCeramics in Condenser Construction : Methodof Measurement : Curves for Wavelengths90-1500 m : Impurities cause Dispersionof Dielectric Constant].-Schusterius. (Zeit-schr. f. tech. Phys., No. 12, VOL 16, 1935,pp. 640-642: Discussion p. 642.)

1592. THE PASSAGE OF CURRENT THROUGH THINFILMS OF ALUMINIUM OXIDE IN ELECTRONTUBES. -Kessel. (Zeitschr. f. tech. Phys.,No. I2, VOL 16, 1935, pp. 506-508.)

The A120, -films were produced electrolyticallyand studied in a high vacuum, using an " electron -gas " cathode (Fig. I). The current/voltage char-acteristics were investigated under various con-ditions as regards the adsorbed gas in the film.Curves taken at room temperature are shown inFig. 2 and described ; Fig. 3 shows anode -currentdensity curves (a) with the electron -gas cathodeand (b) in an electrolyte, for comparison of " dry "and " wet " conditions. The results show con-siderable agreement; the passage of currentthrough the film in the two cases seems to bedetermined by conditions in the film itself, itsexact chemical composition and variation with

1591.


time. The various reasons for this conclusion areshortly given.

1593. THE BEHAVIOUR OF ELECTROLYTIC OXIDEFILMS [Hypothesis of Electrolytic Barrier -Layer and Varied Amounts of AdsorbedOxygen].-Stormer. (Zeitschr. f. tech. Phys.,No. 12, Vol. 16, 1935, pp. 508-512: Dis-cussion pp. 512-513.)

A sudden switch -over of the polarity of theelectrodes of an aluminium electrolytic cell showedthat a delay occurred before the conductivityreached its final value (§ 1). The current/voltagecharacteristics of the dry electrolytic oxide filmand the effect on them of various treatments of thecathode were also investigated (§ 2). An explana-tion of the results is found (§ 3) in the hypothesisof an electrolytic barrier -layer and the variationin the quantity of oxygen adsorbed thereon.

1594. ON THE ROTATION OF DIPOLES IN ELASTICAND VISCOUS MEDIA [Application of DipoleTheory to Amorphous Solids : Coupleacting on Rotated Rigid Sphere in ElasticMedium deduced from Stokes' Result forRotation in Viscous Fluid].-Yates-Fish.(Phil. Mag., February, 1936, Series 7, Vol.21, No. 139, pp. 226-233.) See Gemant,2087 of 1935.

1595. ON DIELECTRIC LOSSES IN CERTAIN HETERO-GENEOUS DIELECTRICS.-M. N. Mikhailovand M. M. Stolyarov. (Journ. of Tech.Phys. [in Russian], No. 6, Vol. 5, 1935,pp. 1007-101o.)

An investigation into the losses which occur inheterogeneous dielectrics consisting essentially of aliquid with non -polarised molecules and an in-soluble solid constituent. The systems investigatedwere : talc powder in paraffin oil, rubber in hexane,and paper impregnated with paraffin oil. Theangles of loss of each constituent taken separatelyand of the above systems were measured, atfrequencies of 5o and 106 cis and at temperaturesvarying from -loo to + 50°C. The resultsobtained are shown in a number of curves and atentative explanation is offered of the variouseffects taking place when two dielectrics are com-bined together.

1596. TECHNICAL MASTERY OF THERMAL BREAK-DOWN [of Dielectrics : Physical Expressionof Theory of Thermal Characteristics].-Lueder, Schottky and Spenke. (Naturwiss.,24th Jan. 1936, Vol. 24, No. 4, p. 61.)

1597 THE INFLUENCE OF THE RELATIVE HUMIDITYOF THE, AIR ON THE LOSS ANGLE OF IN-SULATING MATERIALS AT HIGH FREQUENCIES.-Schwarz. (See 1551.)

1598. THE BREAKDOWN OF OIL -IMPREGNATEDCARDBOARD.-Quittner. (Tech. Phys. ofUSSR, No. I, Vol. 2, 1935, pp. 21-23 : inGerman.)

SS ENGINEER 227

1599. NEW PRODUCTIONS OF THE NUREMBERGSCREW FACTORY AND REPETITION TURNERY[Rotary and Fixed Condensers : Tempera-ture -Independent Material with Very SmallLoss Angles : Smoothing Circuit for Tele-vision High -Voltage Potentiometer.]-(Hochf:tech. u. Elek:akus., December, 1935, Vol. 46,No. 6, pp. 215-216: Industry Review.)

1600. ELECTROLYTIC CONDENSERS FOR BROADCASTRECEIVERS [Investigation of Capacity andLoss Angle, Temperature Effect, etc.].-Linder. (E.T.Z., 2nd Jan. 1936, Vol. 57,No. I, p. 15 : summary only.)

The writer's representation of the electrolyticcondenser, and its complex equivalent circuit ofseries -parallel capacities and resistances, explainstwo phenomena : (i) that the layer thickness (andconsequent weight of oxide) calculated from themeasured capacity is too small, and (ii) that con-densers with smaller peak voltages always give alarger loss angle ; this is attributed to a greaterinfluence of the capacity of the " outer " layer(partly permeated by the electrolyte, whereas the" inner " layer is practically free from electrolyte).It also explains the rising capacity with risingtemperature : the loss angle decreases with risingtemperature but later increases slightly on accountof increasing losses in the outer layer (Fig. 3). Asthe frequency increases the loss angle increasessteadily and almost linearly, chiefly under theinfluence of the series resistances in the porous outerlayer and the electrolyte.

The magnitude of the a.c. or superposed d.c.voltage has little immediate effect on the capacity :but over a long period of time the d.c. voltage andthe temperature are important factors. The closerthe impressed voltage comes to the original formingvoltage, the greater will be the slight additionalforming process and the consequent gradualdecrease in capacity (of the order of 5% in a year).The temperature is of importance in that the higherit is, the more inclined is the oxide layer to pass intosolution. But in dry electrolytic condensers thisdissolving process is extremely small, and the writerconsiders that these condensers already form re-liable components for broadcast receivers and,along the lines of future development which hebriefly discusses, should have an increasing fieldof application.

1601. ON THE SETTING OUT OF SCALES AND GRAT-INGS ON THE DISCS AND BARRELS OF MEAS-URING INSTRUMENTS [e.g. Goniometers] BYA PHOTOGRAPHIC METHOD.-F. L. Burmis-trov. (Journ. of Tech. Phys. [in Russian),No. 6, Vol. 5, 1935, pp. 1109-1112.)

A method has been developed, making use ofconvex and concave conical mirrors, by means ofwhich the drawing of a scale wrapped on a cylindercan be reproduced on the outer or inner surface ofanother cylinder, covered with a sensitised layer.For a French version of this paper see Tech. Phys. ofUSSR, No. 1, Vol. 2, 1935, pp. 61-65. The photo-graphic method has always been recognised as thecheapest but hitherto there have been difficultiesin getting precision with curved surfaces.


1602. A RELAY FOR OPERATING TWO CIRCUITSALTERNATELY WITH DELAYED ACTION [De-tailed Description with Diagrams].-Perfect.(Proc. Phys. Soc., 1st Jan. 1936, Vol. 48,Part 1, No. 264, pp. 203-207: demons-tration.)

1603. GRAPHICAL HARMONIC ANALYSIS [GeneralMethod for Waves containing Sine Com-ponents of Odd Harmonics and CosineComponents of Even : by Measurement ofOrdinate Differences between Dynamic Char-acteristic and Straight Line joining Its Ends].-Hutcheson. (Electronics, January, 1936,Vol. 9, pp. 16-18 and 34.)

1604. A TUBE -CONTROLLED MOTOR [" Synchron-ous Motor with Electronic Connectionbetween Field Coils forms Mechanically -Coupled Variable -Frequency Oscillator ofInteresting Uses "].-King. (Electronics,January, 1936, Vol. 9, pp. 14-15.)

1605. POLAROID, A GLASS WITH THE PROPERTY OFPOLARISING LIGHT .-(5Ci. . News Letter, 1stFeb. 1936, p. 77.)

1606. OPTICALLY PERFECT LITHIUM FLUORIDECRYSTALS LARGE ENOUGH FOR LENSES [VeryTransparent to Ultra -Violet and Infra -RedLight].-Stockbarger. (Science, 6th Dec.1935, Vol. 82, Supp. p. to.) See also Elec-tronics, January, 1936, Vol. 9, p. 38.

1607. AN IRON -FREE COIL FOR THE PRODUCTIONOF STRONG MAGNETIC FIELDS [with Water -Cooling System : Constructional Details].-Gerloff and Lowe. (Zeitschr. f. Physik,No. 9/1o, Vol. 98, 1936, pp. 559-560.)

1608. PERMANENT MAGNET MATERIALS [particu-larly the " Oxide " Material].-Williams :Kato and Takei. (Elec. Engineering, Janu-ary, 1936, Vol. 55, No. I, pp. 19-23.) Dis-cussion of nickel -aluminium -iron, cobalt -molybdenum -iron and cobalt -tungsten -ironalloys, and of the non-metallic materialcomposed of cobalt arid iron oxides des-cribed by Kato and Takei (1933 Abstracts,13.579)

1609. FERROMAGNETIC MATERIALS FOR HIGH FRE-QUENCIES [Analysis of Requirements andSurvey of Methods].-Weis. (Arch. f. tech.Mess., October, 1935, pp. T 139-140.)

16,0. ALNICO-NEW MAGNETIC MATERIAL.-Mishima. (Electronics, January, 1936, Vol. 9,p. 4o.) S69 also 741 and 742 of February.

1611. [Reversible and Irreversible] THERMALEFFECTS OF MAGNETISATION [Of Ferro-magnetics].-Okamura. (Nature, 8th Feb.1936, Vol. 137, p. 241 : short note only.)

16,2. PARAMAGNETIC RELAXATION [Heat developedin Alternating Magnetic Field : Large Effectat Low Temperatures : Relaxation Time ofMagnetic Vector].-Gorter. (Nature, IstFeb. 1936, Vol. 137, p. 190.)

STATIONS, DESIGN AND OPERATION1613. DOUBLING THE AVAILABLE RADIO CHANNELS

[Transmission of Two Programmes on OneCarrier by Simultaneous Amplitude andFrequency Modulation : Local SynchronisedOscillator as Only Addition to ReceiverEquipment].-Woodyard. (Proc. Inst. Rad.Eng., January, 1936, Vol. 24, No. I, p. 8:summary only.)

1614. BROADCASTING : PRINCIPAL DEVELOPMENTSIN I935.-Ashbridge. (Electrician, 31st Jan.1936, Vol. 116, p. 140.)

1615. THE CENTRE OF BRITISH BROADCASTING :DROITWICH, NEAR BIRMINGHAM.-Adam.(Genie Civil, 1st Feb. 1936, Vol. 108, pp.toi-to5.)

1616. ORGANISATION AND TECHNIQUE OF THE_

WORLD RADIO LINK OF 27.10.1935 [" YouthSings across the Frontiers "].-Wcil.(T.F.T., November, 1935, Vol. 24, No. II,pp. 289-292.)

1617. NEW 41 -Mc W8XH [New WBEN too -WattUltra -Short -Wave Broadcasting Transmitterwith Flat Characteristic 30-17 000 cis :" Turnstile " Aerial Array : 25-3o Mi'eRange] .-Kingsley. (Electronics, January,1936, Vol. 9, p. 19.) For a previous artic'esee 2456 of 1935. " Does the high-fidelity,high -frequency WBEN transmitter presagea new broadcast service ? "

1618. MICRO -RAY COMMUNICATION. EFFECT OFWEATHER ON TRANSMISSION : POSSIBLEFUTURE USES OF WAVELENGTHS BELOW5o cm.-McPherson and Ullrich. (See 1341.)

1619. RADIO TELEPHONY : NEW LONG AND SHORTDISTANCE LINKS [in 1935 : including Ultra -Short Waves].-Gill. (Electrician, 31st Jan.1936, Vol. 116, pp. I39-140.)

1620. LOCOMOTIVE TO CABOOSE RADIO COMMUNI-CATION [on Ultra -Short Waves : Results ofWestinghouse Tests : Commercial Equip-ment].-Ellis. (Elec. Engineering, January,1936, Vol. 55, No. 1, pp. 109-113.)

GENERAL PHYSICAL ARTICLES1621. THE NATURE OF LIGHT [Photon may consist

of Train of Circular Lines of Force].-J. J. Thomson. (Nature, 8th Feb. 1936,Vol. 137, pp. 232-233.)

1622. THE EQUATIONS OF MOTION [of the LorentzElectron] IN THE NEW ELECTRODYNAMICS.-Chraplywy. (Comptes Rendus, 5th Feb.1936, Vol. 202, No. 5, pp. 396-397.)

1623. ABSOLUTE VALUES OF THE ELECTRON MOBIL-ITY IN HYDROGEN [Electrical Grid -ShutterMethod for Mobility Measurement].-Brad-bury and Nielsen. (Phys. Review, 15th Jan.1936, Series 2, VOL 49, No. 2, p. 205:abstract only.)


1624. INTERRELATIONSHIPS OF e, h/e AND elm[Discussion of Discrepancies : Latest Values].-Birge. (Nature, 1st Feb. 1936, Vol. 137,p. 187: abstract only in Phys. Review,i5th Jan. 1936, Series 2, Vol. 49, No. 2,p. 204.)

1625. NEW INVESTIGATIONS OF LINES OF HELIUMAND HYDROGEN WITH CROSSED ELECTRICAND MAGNETIC FIELDS [Interconnection ofStark and Zeeman Effects].-Steubing.(Zeitschr. f. tech. Phys., No. ix, Vol. 16,1935, PP. 428-431.)

1626. THE INFLUENCE OF A MAGNETIC FIELD ONTHE DIELECTRIC CONSTANTS OF GASEOUSAND LIQUID NITROGEN AND OXYGEN [POS-itive Result with Liquid Oxygen, explainedby Increased Pressure].-Young. (CanadianJourn. of Res., December, 1935, Vol. 13,No. 6, Sec. A, pp. 111-119.)

1627. INVESTIGATION OF THE [Inappreciable] MAG-NETIC CHANGE OF THE DIELECTRIC CONSTANTOF LIQUIDS IN A FIELD OF 20.4 KGAUSS.-Piekara. (Comptes Rendus, 2oth Jan. 1936,Vol. 202, No. 3, pp. 206-207.) See also 342of January.

1628. IONISATION BY NEUTRAL ATOM BEAMS[studied by Balanced Space Charge Method :Energy available for' Ionisation is very nearTrue Ionisation Potential of Gas].-Varney.(Phys. Review, 15th Jan. 1936, Series 2,Vol. 49, No. 2, p. 204: abstract only.)

1629. A PROBLEM IN POTENTIAL THEORY [Methodsfor Calculating Potential Distributions inTwo Regions of Different " Conductivities "separated by Movable Surface of UnknownShape].-Muskat. (Phys. Review, 15th Jan.1936, Series 2, Vol. 49, No. 2, p. 195abstract only.)

MISCELLANEOUS1630. THE LARGE ROOTS OF COS Z = as + C.-

Cooper and Todd. (Phil. Mag., February,1936, Series 7, Vol. 21, No. 139, pp. 249-262.)

1631. NOTES ON THE ASYMPTOTIC EVALUATION OFOPERATIONAL EXPRESSIONS [Extension ofHeaviside's Rule : Use of Watson's Lemma].-Koizumi. (Phil. Mag., February, 1936,Series 7, Vol. 21, No. 139, pp. 265-274.)

1632. INTRODUCTION TO THE APPLICATIONS OF THEHEAVISIDE SYMBOLIC CALCULUS TO THEPROBLEMS OF ELECTROTECHNICS.-Blondel.(Rev. Gen. de l'Elec., 180., 25th Jan. and1st, 8th Feb. 1936, Vol. 39, Nos. 3-6, pp.83-99, 133-146, 179-191, and 219-229.)

1633. FRACTIONAL CALCULUS : FRACTIONAL CON-TOUR INTEGRATION.-Fabian. (Phil. Mag.,February, 1936, Series 7, Vol. 21, No. 139,pp. 274-280.)

1634. BESSEL PRODUCT FUNCTIONS [of Order n :' Formulae : Expansions].-Costello. (Phil.

Mag., February, 1936, Series 7, Vol. 21,No. 139, pp. 308-318.)

229

1635. ON THE COMBINATION OF OBSERVATIONALDATA [Method for Combination of Curvesderived by Different Experimenters intoUnique Curve : Formulae for Calculationof Ordinates].-Levy and Gascoigne. (Proc.Phys. Soc., 1st Jan. 1936, Vol. 48, Part 1,No. 264, pp. 79-84.)

1636. ON THE SIGNIFICANCE OF SLOPES AND OTHERPARAMETERS ESTIMATED BY LEAST SQUARES.-Deming. (Phys. Review, 1st Feb. 1936,Series 2, Vol. 49, No. 3, pp. 243-247.)

1637. PROBABILITY FOR THE DIFFERENTIAL COM-POSITION OF THE ACCIDENTAL ERRORS, ANDITS APPLICATION [to Cross -Talk in Cables].-Goto. (Journ. I.E.E. Japan, November,1935, Vol. 55 [No. r r], No. 568, pp. 990-994English summary pp. 125-126.)

1638. NEW INFRA -RED " EYE " PIERCES THROUGHHAZE AND SMOKE : COMBINED WITH TELE-SCOPE OR MICROSCOPE, WILL FIND MANYSCIENTIFIC AND OTHER USES.-ZWOrykirl.(Sci. News Letter, iith Jan. 1936, Vol. 29,pp. 21-22.) See also 1576.

1639. THE " PETOSCOPE " [Photocell Apparatus fordetecting Moving Objects in Full Daylight :e.g. Aircraft in Flight].-Lancaster : Fitz-gerald. (Wireless World, loth Jan. 1936,Vol. 38, p. 39 : Electronics, October, 1935, Vol.8, pp. 26-29.)

1640. PHOTOTUBES PERFORATE U.S. STAMPS.-Church. (Electronics, September, 1935, Vol.8, pp. 27 and 29.)

1641. THE COLOUR -MATCHING OF TUNGSTEN -FILA-MENT LAMPS BY MEANS OF A SINGLE PHOTO-CELL AND COLOUR FILTERS [Use in CheckingCalibration of Standards of Colour Tempera-ture].-Preston. (Proc. Phys. Soc., 1st Nov.1935, Vol. 47, Part 6, No. 263, pp. 1012-1018.)

1642. A PHOTOELECTRIC COLORIMETER WITH LOGA-RITHMIC RESPONSE.-Mfiller and Kinney.(Journ. Opt. Soc. Am., October, 1935, Vol.25, No. ro, pp. 342-346.)

1643. APPARATUS FOR DARKFIELD PHOTOMETRYAND DENSITOMETRY.-Williams and Scott.(fount. Opt. Soc. Am., October, 1935, Vol.25, No. 10, pp. 347-349.)AN ELECTROPHOTOMETER WITH BARRIER -LAYER CELLS FOR THE PRACTICAL MEASURE-MENT OF OPACITY.-Meunier. (ComptesRendus, 23rd Dec. 1935, Vol. 201, No. 26,

1644.

1645.

1646.

PP. 1371-1373.)THE EMPLOYMENT OF " PHOTOEMISSIVECELLS IN PHOTOMETRY.-Fleury and Boutry.(Rev. GIn. de l'Elec., 7th Sept. 1935, Vol. 38,No. ro, pp. 323-333.)THE SELECTIVITY OF PHOTOMETRIC IN-TEGRATORS, WITH PARTICULAR REFERENCETO THE PHOTOMETRY OF GASEOUS DISCHARGETUBES [Theoretical Treatment of Errors inDetermination of Luminous Outputs : Ex-perimental Verification].-Buckley. (Phil.Mag., November, 1935, Series 7, Vol. 20,No. 135, pp. 745-76o.)


1647. THE PHOTO-RELAYS.-Mikhailov. (Elek-tritchestvo, No. 19,1935, pp. 23-32.)

1648. PHOTOCELL MONITORING OF A CHEMICALMATURING PROCESS IN EXPLOSIVE CHEMI-CALS.-Zesch. (E.T.Z., 6th Feb. 1936, Vol.57, No. 6, p. 155.)

1649. THE SPECTRAL EMISSION REGION OF CHEMI-CAL REACTIONS [Measured with SuitableSensitive Photocathodes : Region 2000-2200 A].-Audubert. (Comptes Rendus, 1661.13th Jan. 1936, VOL 202, No. 2, pp. 131-133.) See also 1934 Abstracts, p. 54, and1310 and 1997 of 1935.

1650. PHOTOELECTRIC COLOUR -MEASURING DEVICEUSED FOR PROCESS CONTROL [Sugar and OilRefining, Heat Treatment, etc.].-(Elec-tronics, January, 1936, Vol. 9, pp. 36-37.)

1651. A NEW HIGH-SPEED HIGH -SENSITIVITYPHOTOELECTRIC POTENTIOMETER [for theMeasurement, Recording and Controllingof Physical Quantities].-Gilbert. (ReviewScient. Instr., January, 1936, Vol. 7, No. I,PP 41-44.)

1652. PHOTOCELL DETECTOR [registering Presenceof Candle at 7 miles without Telescope] ToBE USED WITH NEW 200 -INCH TELESCOPEMIRROR.-(Electronics, January, 1936, Vol.9, P. 40.)

1653. PHOTOELECTRIC HALF -SHADOW METHOD FORTHE DETERMINATION OF ROTATORY DIS-PERSION.-Schtinrock and Einspom. (Physik.Zeitschr., 1st Jan. 1936, Vol. 37, No. 1, pp.

1654. PHOTOCELL CONTROLS [Special Number ofElektrotechnische Zeitschrift].-(E.T.Z., 6thFeb. 1936, Vol. 57, No. 6, pp. 137-154.)

Including papers by Sewig (fundamental con-siderations) : Tuczek (some applications, includinghigh -sensitivity methods) : Kluge (latest develop-ments of alkali cells) : Seidelbach (control of sound -film amplifiers) : Johannsen (latest alkali cells andtheir conquest of the prejudice against photoelectricdevices in rolling mills) : Richter (photocells as" twilight switches " and photocells in the serviceof water works).1655. " DIE PHOTOELEMENTE UND IHRE ANWEN-

DUNG " [Book Review].-Lange. (Zeitschr. f.tech. Phys., No. 1, Vol. 17,1936, pp. 31-32.)

1656. HIGH -FREQUENCY METHOD OF REGISTERINGTHE EQUILIBRIUM POSITION OF THE BEAM OFA MICROBALANCE.-Santo Rini. (ComptesRendus, 30th Dec. 1935, VOL 201, No. 27,pp. 1468-1470.) For previous work seeAbstracts, 1932, p. 361 : also 1934, P. 54.

1657. THE SPEED -TIME ELECTROGRAPH [for Testson Electric Railway Motors, etc.].-Crom-well. (Elec. Engineering, September, 1935,Vol. 54, No. 9, pp. 923-930.)

1658. INSTRUMENT FOR THE LOCATION OF FAULTSON OVERHEAD LINES [High -Frequency Methodwithout Use of Cathode -Ray Oscillograph].-Allen and Gross. (E.T.Z., 14th Nov. 1935,Vol. 56, No. 46, pp. 1255-1256: summaryonly.)

April, 1936

1659. TRAFFIC CONTROL SYSTEMS [and the Electro-matic and Autoflex Timing Circuits].-Vause. (P.O. Elec. Eng. Journ., January,1936, Vol. 28, Part 4, pp. 301-309.)

1660. THE DRYING OF WOOD BY ULTRA -HIGH -FREQUENCY OSCILLATIONS [Russian Work].-Frolov. (Zeitschr. V.D.I., 21st Dec. 1935,Vol. 79, No. 51, p. 1531 : summary only.)

HOMOGENEOUS VULCANISATION [by H. F.Furnace and Short -Wave Electric Field].-Leduc. (Recherches et Inventions, January,1936, pp. 20-25.)

1662. INDUSTRIAL HIGH -FREQUENCY POWER [7-x0kc/s, for Induction Furnaces, Welding, etc. :Design of 20 kW Generator].-Noble.(Electronics, October, 1935, Vol. 8, pp. 22-24.)

1663. RADIO WAVES ACT AS TAMPER -PROOF BUR-GLAR ALARM [Micro -Waves lace Room byRepeated Reflection from Wall to Wall:Possible Application to Aircraft Detection].-(Sci. News Letter, 4th Jan. 1936, Vol. 29,p.

1664. NEW RADIO DEVICE [Microphone/AmplifierCombination] DETECTS LEAKS IN WATERMAINS.-Pies. (Sci. News Letter, 28th Dec.1935, Vol. 28, p. 407)

1665. RADIOACTIVITY BY BOMBARDMENT [and theApparatus used by Lawrence and Livingston].-Livingood. (Electronics, November, 1935,Vol. 8, pp. 7-9 and 6o.)

1666. ULTRA -SPEED IN MOTION PICTURES [3 000Pictures/Sec., with Electronic Timer :analyses High -Speed Industrial Processes,etc.].-(Electronics, December, 1935, Vol. 8,pp. 6-10.)

1667. ELECTRONICS IN OIL [Applications to Finding,Refining, Analysing, etc.].-(Electronics,December, 1935, Vol. 8, pp. 15-17.)

1668. INDUSTRIAL APPLICATIONS OF THE GRID -CONTROLLED CURRENT DIRECTOR [Photo-electric Counting of Objects : GalvanometerProtection : Safety Device for Looms :Loss -Free Regulation of A.C. : Control ofX -Ray Equipments : Motor Speed Regu-lation].-Hauffe. (Zeitschr. V.D.I., 7thDec. 1935, Vol. 79, No. 49, pp. 1475-1478.)

1669. " INDUSTRIAL ELECTRONICS " [Book Re-view].-Gulliksen and Vedder. (Electronics,December, 1935, Vol. 8, p. 35.)

1670. MONITORING A WELDER [Diode Circuitmeasuring Total Heat Energy of Weld].-RCA Radiotron. (Electronics, January,1936, Vol. 9, p. 38.)

1671. EXPERIMENTS ON THE FREQUENCY MODULA-TION OF A WIEN [Toothed -Wheel] GENERA-TOR : APPLICATION TO STUDY OF PERIODICSPEED VARIATIONS IN MOTORS.-Rytov.(See 1395.)


1672. ELECTRO-MAGNETIC TESTS FOR WORK -HARDENING STEEL WIRES [from Used SteelWire Ropes : Magnetostriction and TravellingSurge Methods].-Wall. (Engineer, 13thSept. 1935, Vol. 16o, pp. 260-263.)

1673. A ELECTROMAGNETIC FATIGUE TESTER [ex-cited by Variable -Frequency Oscillator].-(Engineering, nth Oct. 1935, p. 406.)

1674. SOME DYNAMICMETHODS FOR DETERMINATIONOF YOUNG'S MODULUS [particularly anElectrostatic Method].-Ide. (Review Scient.Instr., October, 1935, Vol. 6, No. 1o, pp.296-298.)

1675. [Measurements Of] INFRA -RED RADIATIONFROM OTTO CYCLE ENGINE EXPLOSIONS.-Nat. Bureau of Standards. (Jour*. FranklinInst., January, 1936, Vol. 221, No. I, pp.158-159: short note only.)

1676. A PIEZOELECTRIC INDICATOR FOR HIGHSPEED INTERNAL-COMBUSTION MOTORS[Complete Equipment with Cathode -RayOscillograph and Camera].-Jungnickel.(Zeitschr. V.D.I., 18th Jan. 1936, Vol. 8o,pp. 8o -8i.)

1677. THE STANDARD-SUNBURY ENGINE INDICATOR[Design of Engine Pick -Up Unit, andElectrical Amplifying and IntegratingCircuits].-Beale and Stansfield. (Engineer,27th Dec. 1935, Vol. i6o, pp. 667-669.)

1678. CATHODE-RAY OSCILLOGRAPH ENGINE INDI-CATOR [and the Distortion produced byPhotography of Curved Image].-Dodds.(Engineering, 8th Nov. 1935, p. 508.)

1679. ELECTRICAL INDICATOR FOR HIGH-SPEEDINTERNAL-COMBUSTION ENGINES [Three -Plate Condenser Principle].-Fieber. (Zeit-schr. V.D.I., 9th Nov. 1935, Vol. 79, No. 45,pp. 1368-1369.) For Reisch's " potential -dividing " system for ultra -micrometers, onwhich this indicator is based, see 1932 Ab-stracts, pp. 53 and 109.

1680. THE MEASUREMENT. OF VIBRATION.-Sell.(Zeitschr. V.D.I., i6th Nov. 1935, Vol. 79,No. 46, pp. 1401-1402.) Siemens & Halskeapparatus.

1681. AN ELECTRODYNAMIC VIBRATION METER [forMechanical Vibrations] AND ITS APPLICATIONTO THE INVESTIGATION OF VIBRATIONS OFBUILDINGS [with Circuits of Amplifier andT -Filter Sections for Measurement of Ac-celeration and Movement : Oscillograms ofSwitch -On Transients : Frequencies belowloo cis].-Meyer and Bohm. (E.N.T.,December, 1935, VOL 12, No. 12, pp. 404-414 : abstract only in Zeitschr. f. tech. Phys.,No. 12, VOL 16, 1935, pp. 567-568.)

1682. ON THE MEASUREMENT OF WEAK MECHANICALOSCILLATIONS [by Movement of One Mirrorof Optical Interferometer].-Seleznev andKnopfer. ( Journ. of Tech. Phys. [in Russian],No 6, Vol. 5, 1935, PP. 1113.)

1694.

231

1683. SEISMIC PROSPECTING METHOD USES ELEC-TRONIC RECORDERS.-(Electronics, January,1936, Vol. 9, p. 4o.)

1684. EXPERIMENTS ON CONDUCTING LAMINAE INPERIODIC MAGNETIC FIELDS [SecondaryMagnetic Field increases with Dimensions ofConducting Circular Disc, Conductivity,Frequency : Phase Changes : Application toLocation of Mineral Lodes].-Bruckshaw.(Proc. Phys. Soc., 1st Jan. 1936, Vol. 48,Part r, No. 264, pp. 63-74.)

1685. AN ELECTROVIBRATOR WITH COMBINED [LOWand] HIGH FREQUENCY [Nerves resonate toUltra -Short Waves].-Boursin. (ComptesRendus, 6th Jan. 1936, Vol. 202, No. 1, pp.95-96)

1686. DISCHARGE TUBE CIRCUIT USED IN STUDYOF BRAIN [Steep -Wave -Front Impulses in-duced in Coil in Ape's Skull].-Chaffee andLight. (Electronics, October, 1935, Vol. 8,P. 37.)

1687. COSMIC RAYS AND THE ORIGIN OF SPECIES[Mutations might be caused by HighEnergy Particles striking Chromosomes :Great Variety of Crop Plants at HighAltitudes].-Thomas. (Nature, i1th and18th Jan. 1936, Vol. 137, pp. 51-53 : 97-98.)

1688. OSCILLATIONS DUE TO CORONA DISCHARGES ONWIRES SUBJECTED TO ALTERNATING POTEN-TIALS [One Type resembling RelaxationOscillations : Effect of Cosmic Rays : etc.].-Tykociner & others. (Univ. of IllinoisBull., 17th. Sept. 1935, Vol. 33, No. 3,53 pp.)

1689. THERMAL MEASUREMENTS IN SHORT-WAVEDIATHERMY [satisfactorily made by BenzeneThermometers in Quartz].-Bessemans,Rutgers and van Thielen. (Comptes Rendus,13th Jan. 1936, VOL 202, No. 2, pp. 157-159.)

1690. METHODS OF ENHANCING X-RAY ACTION [byRadiothermy with Ultra -High Frequencies :Possible Use of Steel Needles heated byRadiothermic Induction, to cause TissueNecrosis].-Delario. (Sci. Abstracts, Sec. A,25th Jan. 1936, Vol. 39, No. 457, pp. 76-77.)

1691. THE APPARATUS USED FOR MEDICAL APPLICA-TIONS OF ELECTRICITY.-Leraoine. (GenieCivil, 8th and 25th Feb. 1936, Vol. 108,pp. 131-133 and 157-159.)

1692. AN APPARATUS FOR RECORDING SYSTOLICBLOOD PRESSURE [using PiezoelectricMicrophone].-Omberg. (Review Scient.Instr., January, 1936, Vol. 7, No. 1, pp.33-34.)

1693. A RECORDING ELECTROCARDIOGRAPH [Pre -Amplification by D.C. or Very Low FrequencyA.C. Amplifier, Modulation by i 000 c/sNote, then A.F. Amplification].-Portier.(Genie Civil, 1st Feb. 1936, Vol. 108, p. 119 :summary only.)1935-IN REVIEW : 1936-PROSPECTS.-(Electronics, January, 1936, Vol. 9, pp. 7-9and 33.)


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.A selection of abstracts from patents issued in the U.S.A. is also included, and these bear a seven -figureserial number.

TRANSMISSION CIRCUITS AND APPARATUS437 671. -Oscillation -generator on which the gridand anode potentials are maintained substantiallyin exact anti -phase in order to stabilise frequency.

Marconi's W. T. Co. and E. B. Moullin. Appli-cation date 4th May, 1934.439 747. -Wired " broadcasting " system usinghigh -frequency carrier -currents, so as to providealternative programmes.

P. P. Eckersley and R. E. H. Carpenter. Applica-tion date 26th June, 1934.440 238. -Feed -line for ultra -short waves made upof a number of repeated units which are, at leastapproximately, tuned to the working frequency.

N. V. Meaf. Convention date (Germany) 27thApril, 1934.2 000 584. -Frequency modulating system in whichthe applied signal varies the resistance of a valveshunted across a part of the main oscillatory circuit.

C. Fichandler.2 000 68. -Stabilising the frequency of the oscil-lations generated in a circuit of the multivibratortype.

V. J. Andrew (assignor to Westinghouse Electricand Manufacturing Co.).2 012 o18.-Piezo-electric master -oscillator unit de-signed to be immersed in an oil -tank for coolingpurposes.

M. Osnos (assignor to Telefunken Co.). Con-vention date (Germany) 14th January, 1931.

RECEPTION CIRCUITS AND APPARATUS437 442. -Wireless receiver fitted with fine-tuningmeans which automatically comes into operationwhen the controls have been roughly tuned to adesired station.

H. Jackson. Application date 24th February,1934-

437 460. -Single -valve super -regenerative ampli-fier with electron coupling between the quenchingand signal frequencies.

D. W. Pugh and Baird Television.437 305.-Homodyne receiver with means forvarying the degree of modulation of the receivedsignal.

Marconi's W.T. Co. and G. M. Wright. Applica-tion date 28th April, 1934.

437 569. -Arrangement of the illuminating lampfor the tuning dial of a wireless receiver.

E. K. Cole. Convention date (Sweden) 26thApril, 1934.437 643. -Detector valve with two grids fed inphase -opposition so as to balance out the carrier -wave component, particularly in television systems.

Radio Akt. D. S. Loewe. Convention date (Ger-many) 2nd March, 1933.

438 oio.-Tuning dial in which the receivedstation is indicated by light of a different colour orintensity from the general illumination.

N. V. Philips. Convention date (Germany) 21stJuly, 1934-

438 o14. -Wireless receiver fitted with two loud-speakers, one or other of which is brought intocircuit according to the setting of the volume control.

E. K. Cole (Aga Baltic Radio). Application date25th June, 1935.

438 670. -Suppressing inter -station " noise " intuning without using an extra valve for thispurpose.

E. K. Cole and others. Application date 25th June,1934-

438 797 --Wireless receiver with adjustable band-pass selector interlocked with the tuning control.

Hazeltine Corporation (assignees of H. A. Wheelerand N. P. Case). Convention date (U.S.A.) 7thNovember, 1933.

439 o47.-Heptode valve used in a reflex circuit,as a tetrode high -frequency stage and a triode low -frequency amplifier.

Marconi's W.T. Co. ; N. M. Rust ; and O.' E.Keall. Application date 28th May, 1934.

439 288.-Superhet circuit particularly suitable for" midget " sets.

Marconi's W.T. Co. (assignees of R. M. Smith).Convention date (U.S.A.) 3oth June, 1933.440 451. -Direct -current or zero -frequency amplifierdesigned to suppress fortuitous voltage -variationsin the grid, filament, or anode supply.

G. Krawinkel. Convention date (Germany) 25thOctober, 1933.

44o 607. -Tuning indicator designed to display alarge number of stations within a limited space.

J. A. Wiegand and P. R. Weenink. Conventiondate (Holland) 8th May, 1934-

989 394. -Balanced Wheatstone bridge arrange-ment for preventing the D.C. component in theoutput circuit of a valve amplifier from passingthrough the windings of inter -stage couplingtransformers or chokes, or through the windingsof a loud -speaker.

TV. Aull.

I 998 479 Circuit for automatically matchingthe two valves of a push-pull amplifier.

R. 0. Wise (assignor to Bell Telephone Labora-tories).

2 015 191. -Wave -band switching system with aminimum number of tapped coupling and reactioncoils.

R. M. Planck (assignor to Radio Corporation ofAmerica).


2 000 677. -All -wave type of receiver fitted withinterchangeable plug-in inductance coils, arrangedin units for changing over from one wave -band toanother.

B. Trevor (assignor to Radio Corporation ofAmerica). Application date, 12th February, 1932.

VALVES AND THERMIONICS436 533.-" Renode " valve with means for con-centrating the electron stream and subsequentlycontrolling its passage through perforated elec-trodes.

C. A. S. Jensen. Convention date (Denmark)12th April, 1933.

436 6o6. -Electrode assembly in a thermionicvalve designed to minimise any looseness createdby such vibration as may arise in transport.

E. Y. Robinson and another and AssociatedElectrical Industries. Application date 19th June,1934.

436 636.-Thermionic-valve base designed to per-mit the soldering of the electrode leads in a plungebath.

Delta Co. Convention date (Switzerland) 5th May,1934-

437 726. -Multi -grid frequency -changing valveutilising electron -coupling.

Marconi's W.T. Co. (assignees of J. C. Smith).Convention date (U.S.A.) 31st January, 1933.

437 594. -Rigid electrode -assembly forming theelectron gun of a cathode-ray tube.

L. F. Broadway and W. F. Tedham. Applicationdate 26th March, 1934.

437 624. -Cathode-ray tube in which the electrodesare correctly spaced apart by means of a number ofinsulated distance pieces.

General Electric Co., L. C. Jesty and G. W. Seager.Application date 17th August, 1934.

437 924. -Valve for ultra short-wave working inwhich the grid lead is taken to the centre of thatelectrode through an opening formed in the anode.

Radio Akt. D. S. Loewe. Convention date (Ger-many) 13th March, 1934.

438 343. -Indirectly -heated cathode constructionfor a cathode-ray tube designed to minimise heat -losses.

British Thomson -Houston Co. Convention date(Germany) 21st July, 1934.

439 851. -Electrode assembly for high-poweredvalves designed to maintain uniform spacing andto be proof against mechanical shocks.

Standard Telephones (assignees of J. E. Clark andV. L. Ronci). Convention date (U.S.A.) 1st Novem-ber, 1934.

439 879. -Gas -filled amplifier or detector in whichthe quantitative relation between input and outputis determined by the grid -cathode spacing and bythe use of an ionizable metallic vapour containingcadmium and mercury.

S. Ruben. Convention date (U.S.A.) loth December,1933.

233

44o 146. -Short-wave valve in which the cathodeconsists of two parallel bands or strips with aninterposed space -charge grid.

N. V. Meaf. Convention dates (Germany) 9th and13th January, 1934-

I 991 387. -Modulating valve fitted with controlelectrode comprising two sets of parallel plateswhich deflect the electron stream away from theanode.

R. Gunn.

DIRECTIONAL WIRELESS2 002 181. -Method of marking out a predeterminedroute by overlapping radio beams, both of whichare radiated from dipole aerials set slightly out ofthe focal line of a common reflector.

W. Ilberg (assignor to Telefunken Co.).

ACOUSTICS AND AUDIO FREQUENCY CIRCUITSAND APPARATUS

437 973. -Portable public-address unit in which avibration -absorbing medium is interposed betweenthe microphone and loud -speaker.

A. C. Snell. Application date 21st June, 1934.

438 824. -Arrangement of nested loud -speakersdesigned either for dual or independent operation.

W. T. Tennant (communicated by the Rola Co.).Application date 5th October, 1934.

439 561. -Loud -speaker with a two -coil drive,only one of the coils responding to the lowerfrequencies.

Marconi's W.T. Co. (assignees of A. S. Ringel).Convention date (U.S.A.) 17th March, 1933.

440 612. -Class B push-pull amplifier arranged tooffset the effect of variations in the plate supplyvoltage.

Telefunken Co. Convention date (Germany) 29thMay, 1934.2 oo8 7c1. -Sound reproducing system in which.the range of volume is expanded automatically byvarying the gain of the amplifiers in step with theinstantaneous intensity of the applied signals.

J. H. Hammond.2 ca 6 402. -Push-pull amplifier operated with afixed positive grid bias, substantially half -way tomaximum plate current, so as to increase thepower -sensitivity of the circuit.

C. Travis (assignor to Radio Corporation ofAmerica).

2 017 5 15 .-Shunt volume control with automatictone compensation for the pick-up of a radiogram.

C. M. Sinnett (assignor to Radio Corporation ofAmerica).

TELEVISION AND PHOTOTELEGRAPHY436 809. -Television system in which series ofsynchronising impulses are sandwiched between thepicture signals and in which' the first impulse ismade greater in amplitude than any others in thesame series.

T. M. Constable and Baird Television. Appli-cation dates ist May and 31st August, 1934.

234 THE WIRELE

437 02I. -Television system in which the pro-duction of a kinema film or record forms an inter-mediate stage in the process of transmission.

J. C. Wilson and Baird Television. Applicationdate 23rd January, 1934.437 34o.-Electro-optical system for synchronisingtelevision signals.

- P. V. Reveley and Baird Television. Applicationdate 26th June, 1935.437 6o2. -Television scanning system in which thesource of light is an electric discharge moving alongthe track of a pair of " rail " electrodes.

Communications Patents Inc. Convention date(U.S.A.) 21st rJune, 1933.437 656. -Construction of television scanning -discwith particular reference to the formation of thelight -apertures.

Electrical Research Products (assignees of H. E.Ives). Convention date (U.S.A.) 24th May, 1933.437 731.-Superhet circuit for the simultaneousreception of sound and television signals.

Electric &' Musical Industries and C. S. Agate.Application date 5th April, 1934--437 988.-Light-polariser, as used in television, ofthe type in which the unwanted ray is suppressedby total internal reflection.

P. V. Reveley and Baird Television. Applicationdate 28th August, 1934.438 386. -Means for adjusting the terminal posi-tions of the electron beam on the fluorescent screenof a cathode-ray tube when used for televisionscanning.

General Electric Co. and L. C. Jesty. Applicationdate 24th July, 1934.438 424. -Television system in which a rotatingring of cylindrical lenses co-operate with a fixedprism to scan a continuously -moving cinema film.

Scophony, Ltd. ; J. H. Jeffree ; and G. Wikken-hauser. Application date 18th April, 1934.439 2 I. -Interlaced scanning system in which abeam of light is swept by a rotating mirror over aset of inclined mirrors and then on the screen.

C. 0. Brown. Application date 29th May, 1934.439 164. -Reproducing the background or averageintensities of a televised scene as well as the instan-taneous light and shade values.

Radio Akt. D. S. Loewe. Convention date(Germany) 25th February, 1933.439 225. -Interleaved scanning system in whichthe frame -repetition frequency is an exact self -multiple of the line -scanning frequency.

Electrical Research Products (assignees of F. Gray).Convention date (U.S.A.) 19th May, 1934.439 236. -Scanning system based on the use of alight -cell containing a transparent liquid which issubjected to high -frequency mechanical vibrationsas the light -ray to be modulated passes through.

Scophony Ltd. and J. H. Jeffree. Applicationdate 3rd March, 1934.439 414. -Electrostatic lens system for focusingthe electron beam in a cathode-ray tube.

Fernseh Akt. Convention date (Germany) 13thJune, 1934.


439 737. -Electron -optical system applied to acathode-ray tube with the object of enlarging thenormal size of a televised picture.

L. Schiff. Application date 9th April, 1934.439 813. -Time -base circuit for television withmagnetic synchronising control of the gas -filleddischarge valves.

Marconi's W.T. Co. and R. J. Kemp. Applica-tion date 14th June, 1934.439 990. -Electrode arrangement for preventingundesired variations in the size of the spot pro-jected on to the screen in a cathode ray tube ofthe gas -filled type.

Telefunken Co. Convention date (Germany) zndAugust, 5933.439 994. -Framing the picture by moving theelectron stream into proper phase on the fluorescentscreen of a cathode-ray television receiver.

T. E. Bray and Baird Television. Applicationdate 24th August, 1934.440 386. -Cathode-ray viewing -screens on whichthe picture is reproduced partly by incandescenceand partly by fluorescence or phosphorescence.

J. L. Baird and Baird Television. Applicationdate 13th October, 1934

SUBSIDIARY APPARATUS AND MATERIALS436 407. -Method of cutting a piezo-electric crystalso that it has a zero temperature coefficient.

Telefunken Co. Convention date (Germany) 21stJuly, 1933.

1934437

crystal cut so as to havea zero temperature -coefficient and no fortuitousmode of vibration.

Standard Telephones. Application date 27th April,

437 605.-Electro-dynamic microphone in whichthe moving coil is supported wholly by the dia-phragm and is capable of movement in more thanone direction.

S. L. Price and J. E. G. Parritt. Application date27th April, 1934.

984 312. -Photo -electric cell with a controlelectrode connected to a valve amplifier which feedsback energy in phase with the photo -electricemission.

A. J. McMaster (assignor to G -M Laboratories).2 000 026. -Valve -maintained magnetostrictive os-cillator made of an alloy of cobalt chromium,nickel and iron so as to have a constant temperaturecoefficient.

J. M. Ide.2 014 786. -Photo -electric cell cf the gas -filled typewhich is operated at maximum anode -cathodevoltage and is coupled to a valve amplifier whichautomatically prevents the applied voltage fromrising beyond the " glow " point.

F. H. Shepard (assignor to Radio Corporation ofAmerica).2 015 836.-Piezo-electric crystal mounting inwhich one of the electrodes is curved or shaped soas to prevent the formation of standing air -waveswhich tend to damp the crystal vibrations.

R. Bechmann and others (assignors to the Tele-funken Co.).

April. 1036 THE WIRELESS ENGINEER

INDUCTANCE BRIDGiRANGE 10 MICROHENRIES TO 100 HENRIES, ACCURACY 0.25%.

The Inductance Bridge, Type 1-A, isreading audio -frequency instrumentthe above range.The bridge ratio is unity for all rangesthe input impedance is low.

It is designed for rack mounting, but canbe supplied as a bench instrument.Fully described in Bulletin No. B.270-C.

a directcovering

and

also

MUIRHEAD & CO., LTD.MAKERS OF PRECISION INSTRUMENTS FOR OVER 50 YEARS.

ELMERS END KENT.TELEPHONE : BECKENHAM 0041-2.


A zi THE WIRELESS ENGINEER April, 1936

A MAINS -OPERATED IMPEDANCE COMPARISON BRIDGE-

CompletelySelf - containedand Portable-Type TF.202A.

FEATURESMeasures directly the percentage differ-

ence and relative power factors of any twosimilar reactances.

Incorporates 10 kcs. Oscillator and selec-tive visual balance indicator.Range 20 p.Hy to 0.1- Hy.

20 mmfds. to 1 mfd. t30.40 ohms. to 0.1 megohms.Power factor 121%. NETT.

E. K. Cole, Ltd., Government Contract

Precision bridgefor the rapidcomparison of in-ductors,capacitors

and resistors.

Precision electrical measuring instrumentsfor Research and Industry.

STANDARD INSTRUMENTS.

Standard Signal Generator Special AmplifiersBeat Frequency Oscillator Impedance BridgesDistortion Factor Meter Valve VoltmetersOutput Meters Field Strength Measuring Setsinductance Bridges Noise Meters

Capacity Bridgesand allied apparatus designed for special requirements.

and Instrument Dept., Southend - on -Sea.

Second Edition-

RADIOCompletely Revised

ATA CHARTSA SERIES OF ABACS

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

Published from the Offices of THE WIRELESS WORLD "

RADIO DATA CHARTS-now issued in a completely revised (second) edition-provide designers of wireless apparatus with a ready, convenient means of solvingall the more familiar problems connected with the design of modern radio apparatuswithout having recourse to complicated formulae and mathematics.

In order to keep abreast of the great advances in wireless communication since thefirst edition of " Radio Data Charts " appeared, obsolete abacs have been omitted inthis second edition and important fresh material added. The book is now completelyup to date.

By the use of the new " Radio Data Charts," such abstruse problems as the design oftuning coils are solved almost as easily as the simple applications of Ohm's Law.

Price 4'6 net By post 4'10(37 CHARTS and 46 Diagrams)

From all leading booksellers or direct from the Offices of

"THE WIRELESS WORLD," Dorset House, Stamford Street, London, S.E.1


April, ;6 THE WIRELESS ENGINEER

RADIO RECEIVERMEASUREMENTS

MIFFS

By Roy M. Barnard, B.Sc., A.M.I.R.E.

Although primarily designed for the benefit of the radio service engineer,this concise handbook is also of practical value to the amateur experimenter.

It describes the methods of measuring receiver performance and providesprovisional standards as a basis for judging performance. Measurements ofsensitivity, selectivity and fidelity are explained at length and the interpretationin estimating receiver performance is carefully set out.

Details are given of up-to-date methods of receiver testing with full descriptionsof commercial signal generators and their application to the adjustment ofmodern superheterodynes and "straight" receivers.

Complete with fifty-three illustrationsand diagrams, summaries of method,four appendices and a general index.

PRICE 4/6 net By post 419From dl hooks:11er. or direct from the Publtshert

& SONS LTD., DORSET HOUSE, STAMFORD STREET, LONDON, S.E. tW.R.3

PIEZO QUARTZCRYSTALS

for all applications.Price List of Commercial,Broadcast, and Amateur Trans-mission Crystals, free on request

THE QUARTZ CRYSTAL CO., LTD.,63 & TI, Kingston Rd., NEW MALDEN, Surrey.

Telaphoue-BW01. 0334.

Publishers'

BINDINGCASES

in Grey Clothfor

VOLUME XIIJan. -Dec., 1935

of

WIRELESSENGINEERPrice 3/- By post 3/3Round Volume Complete, £2

By Post £2.1 .0

Earlier Volumes can be supplied.Write for particulars to the

Publishers

ILIFFE & SONS LTD.Dorset House, Stamford St., London, S.E.I

W.B.x6.

SITUATIONS VACANT.Designers and Assistant Designers forRadio Equipment and Rectifiers re-quired by large Firm not far fromLondon. Applicants should state age,qualifications, experience and salaryrequired. University qualifications anadvantage. Box No. 7112, c/o " TheWireless Engineer." 1315

SITUATIONS WANTED.Highly Qualified Technician, withvaluable experience in BroadcastReceiver Design, would like to changehis post to one with ConsultingEngineers. Box No. 7003, c/o " TheWireless Engineer." [314

FINANCIAL, PARTNERSHIPSAdvertiser wishes to invest up to£250,000 in an established businesshaving a good record of profits.Any business with profits of not lessthan 12,000 per annum or a businesswhere profits have grown progressivelyfor the last 3 or 4 years would beconsidered.Management would not be required-to the contrary the Advertiser wouldprefer the existing management toremain if possible.Principals only are requested to com-municate direct giving brief details asto profits, nature of business, scope fordevelopments and price required toBox No. 7211, c/o " The WirelessEngineer." [316

AMPLIFIERSAND ALL SOUND REPRODUCINGEQUIPMENT FOR ALL PURPOSES.Individually made not mass produced.

Pull &toils or deraonstra/iora from

TA N N OY PRODUCTS(Guy R. Fountain. Ltd.).

Cantebury Grove, WEST NORWOOD,London, 8.E.27

WIRELESSDIRECTION

FINDING& Directional Reception

By R. KEEN, B.Eng.Second Edition (1927)

This volume describes the principles of Directionand Position Finding, and is invaluable to wireless.engineerstackling this held of work for the first timesWell illustrated from photographs and diagrams.

Price 21f- By post 21/9From learliag booksellers or direct from

ILIFFE & SONS LTD., w.a.aDorset House, Stamford St., London, S.E.I


iv THE WIRELESS ENGINEER April, 1936

The

THEPRACTICAL

RADIO & TELEVISION JOURNAL

orldWhat is new in the world of radio and television,both in theory and practice, is recorded week byweek in " The Wireless World." It contains thevery latest information on set construction andincludes articles on broadcasting and othertopics. " The Wireless World " is essentiallythe wireless newspaper for the keen amateur.

SPECIAL FEATURES ARE:Theoretical Articles discussing new principles ; DetailedTechnical Reviews of Commercial Radio Receivers ;Designs for Experimental Receivers specially preparedto illustrate the practical application of new principles

in valves, components or circuit arrangements.

EVERY FRIDAY 4d.Subscription:

Home and Canada, El 1 8:Other countries El 3 10 per annum,

post free.

ILWFX & SONS LTD., DORSET HOUSE, STAMFORD ST.. LONDON, S.E.1

w.E.4

Printed in England for the Publishers, Ihrryz & Soria LTD., Dorset House, Stamford Street, London, 8.E.1.,by The Cornwall Press Ltd., Paris Garden. Stamford Street, London,

wireless engineer · 2 the wireless engineer april, 1936 transformers mains transformers and...

Documents