JULY 1938

PHENOMENA, IN AMPLIFIER VALVES CAUSED BY SECONDARY EMISSION

by J. L. H. JONKER. 537,.533.8 : 621.396.645.3

In amplifier 'Valvessecondaryemission may occur not only from electrodes but also fromnon-conductors which are at a positive potential. In this article seve;_'~lphenomena aredescribed which are connected, with this secondary emission.The secondary emission of the anode has an unfavourable effect on the characteristic oftetrodes. Improvement may be obtained by covering the anode ,with a suitable materialor providing it with ribs, and further by employing an extra electrode and by utilizingthe space charge in a suitable manner. . .The secondary emission of insulators results from the fact that an insulator which is struckby electrons and emits secondary electrons may take on a considerable positive potential.This effect is applied in the fluorescent screen of cathode ray tubes, and mayalso occuras an undesired phenomenon in radio valves where it causes strong distortion and outputdamping, Finally several ways are indicated of oombatting the secondary emission ofinsulators.

When a plate is. struck by electrons which havebeen accelerated by a sufficiently high voltagevsecondary electrons are freed from the plate ashas been described in previous article~ in thisperiodical l- 2)..

In the second article cited it was explained howdeliberate use can be made of the 'phenomenon ofsecondary emission to amplify electron currents;in' this ,.case specially prepared electrodes are usedwhich give a high secondary emission upon bom-bardment with primary electrons. Secondary elec-trons may, however, appear in every amplifiervalve when electrons which are accelerated by thevoltages common in amplifier valves strike a con-ducting or insulating surface. In general the numberof these secondary electrons, even' in the case ofsurfaces which are not specially prepared; may beof the same order of magnitude as that of the pri-mary electrons.

In this article we shall discuss the phenomenawhich occur in an amplifier valve as a result of thissecondary emission, and we shall also study theavailable means of combatting secondary emissionin cases where it is undesired, or at least of prevent-ing its having an unfavourable effect on the action~~~~ .

Secondary emission of conductors

As an example of a conductor on which an un-desired secondary emission may occur we shallconsider the anode of a tetrode, i.e. of a valvewith a cathode, a control ~id, a screen grid andan anode.With such a valve, upon large changes of die

1) H. Bruining, Secondary Electron Emission, Philips technoRev. 3, 80, 1938. -

2) J. L. H. Jonker and M. C. Teves, Technical Applicationsof Secondary Emission, Philips techno Rev. 3, 137, 1938.

anode voltage such as occur in final amplifier valves,the minima of the anode voltage may be consid-erably lower than the screen grid voltage. Thecontinuous line in jig. 1 reprèsents the anode

la\

I· ,....-1/I1 ----,/'

_------.---' ,I_'/ , 1

rj---" 11rj \ 11• 1 I

fj. \ 1 Iti \ , 1. \ I 1

\ I 1

.27007

Fig. 1. -Current-voltage curves of tetrodes._"", --, anode current of a tetrode as a function of the

anode voltage. Vg2 is the screen grid voltage.-.-.-. current-voltage curve when secondary emission is

partially suppressed by means of a layer of carbonand ribs on the anode,

---- current-voltage curve in the absence of secondaryemission. .

current which would flow if no secondary emissionoccurred, as a function of the anode voltage (atconstant screen grid voltage). If, however, second-ary emission does occur the current. varies accord-ing to the dotted line.

The difference between the two curves may beexplained as follows. If the anode voltage is higherthan the screen grid voltage, the secondary electronswhich are freed' from the anode are driven backto the anode by an opposing field. The primary elec-trons, much fewer in number, which strike thescreen grid will also. free secondary electrons. Theseelectrons .are drawn to the anode by the samefield, so .that the anode current is slightly higher

211

Va

212 PHILIPS TECHNICA~ REVIEW

with secondary emission than without it.If the anode voltage becomes lower than the

screen grid voltage, the direction of the field chang-es, and the. secondary electrons from the anodewill be attracted to the screen grid. The anodecurrent therefore decreases suddenly and the screengrid current increases, which results in highdistortions' in the anode alternating current. Themeans available' for oombatting these undesiredphenomena may be divided into two groups:1) attempts may be made to keep the secondary, emission of the anode low,2) attem:ets may be made to make the action of

the secondary electrons harmless by suitableconstruction of the anode or by the use of anextra electrode.

The possibilities of lowering the secondaryemission have already been described in detail inthis periodical (Bruining, loc. cit.).

The lowest secondary emission is obtained froma layer of carbon which is deposited in such a waythat the surface is rough or flaky, so that the second-ary electrons are for the most part captured inthe cavities between the grains of carbon andcannot be drawn away. The secondary emissionwhich has been lowered in this way, however, stillamounts to 30 - 50% of the primary electroncurrent. A complete suppression .of secondaryemission cannot therefore be achieved by thismethod, and in discharge tubes there are a numberof factors aéting which make it difficult to retaineven these low values when the tube is in use. Oneof these factors has already been mentioned in aprevious .article, namely the alteration of the'surface by the deposition of evaporated material,barium and barium oxide, from the hot cathode.The high temperature during the outgassing ofthe electrodes, the 'gases freed' during the processof mänufacture and the getter' employed mayalsobe factors causing a surface modification whichincreasës the values of the secondary emission.

Considering the fact that sup p r eesion ofsecondary electron emission was found to be impos-sible, other methods were sought of neutralizing theeffect of secondary emission. It is possible to dothis to a great extent by giving the anode ribs or'partitions (see fig. 2). The same effect is obtainedin this way as by the use of a porous carbon layer.The secondary electrons freed b'etween the par-titions are to a large extentrecaptured by the par-titions, and thus return to the anode.

In fig. 1 the line drawn thus:'-'-'-' representsthe anode current of/such a valve, which, has inaddition its. anode and screen grid covered with.

Vol. 3, No. 7

a layer of carbon. As was to be expected the second-ary electron current from the anode to the screengrid is considerably reduced but not completelysuppressed. .The remaining irregularity of thecurve proves that the secondary emission is stillappreciable,

-+- g2

-+---1- gl

.2701;;

Fig. 2. Tetrode with ribbed anode: The secondary electronswhich leave the anode are to a large extent recaptured hythe ribs. .

A completely 'effective method of preventing thepassage of secondary electrons from the anodeto the screen grid is the introduetion of an extraelectrode, a so-called suppressor grid, which whenadded to a tetrode converts it into a pentode. Thismethod is used in most radio valves.

How does a suppressor grid work? Its action isbased on the fact that the secondary electrons havefor the greater part relatively low velocities, so thata counter voltage of 20 - 30 volts is sufficient tosend practically all the secondary electrons backto the anode.' .If a fine-meshed grid were placed in front of the.

anode and if it were kept continually. about 20volts lower in potential than the anode, secondaryemission would be suppressed. Practically the. sameeffect can be achieved by placing between screen

!J2 ' a.27014

Fig. 3. Variation of the potential between screen grid andanode in a pentode without space charge (plane parallel.electrodes). .

JULY 1938 PHENOMENA VALVES CAUSED BY SECONDARY EMISSION 213

grid and anode a wide-meshed grid which is con-nected to a point of low potential, for instancethe cathode. By a suitable choice of the dimensionsof the mesh and position of this grid the lowestvoltage along the path of the electrons betweenanode and screen grid can always be kept sufficientlylower than the anode voltage to cause secondaryelectrons to return to the' anode. On the other handthis potential must always be positive even withthe lowest instantaneous anode voltage, since other-

_ wise all electrons will not reach the anode. In fig. 3the variation of potential is given fo~ two values ofthe anode voltage, space charges between screengrid and anode being neglected.

g2 a.27008

Fig. 4,.Variation of potential between screen grid and anodein a tetrode; a' without, and b with space charge, The spacecharge prevents the passage of secondary electrons from anodeto screen grid.

The presence of the space charge has the'result that the line representing the variation inpotential between two flat electrodes is no longerstraight, but curves. Fig. 4b shows the variation ofpotential between the screen grid and the anodeas it would be with a large anode current (large

--+-If-g2---I-II--gt

27015

•Fig. 5. In order to prevent the trànsition of secondary electronsfrom anode to screen grid, the potential in the space betweenthese electrodes must be lowered. This is achieved here bymeans of the auxiliary electredes It which are kept at cathodepotentlal. At the same time the transition time of the electronsand therefore also the space charge is increased, whereby thepotential is still 'further lowered.

space charge) a low anode voltage, and no sup-. pressor grid.It may be seen that with a large current between

screen grid and anode a potential minimum mayoccur here also, and may be able of preventing thepassage of secondary electrons between anode andscreen grid.A practical type of construction which makes

use of this effect was obtained by combining theaction of the spa~e charge with that of an extraelectrode, consisting here of two plates connectedwith the cathode (fig. 5) .. This construction issimpler rhan that with a suppressor grid and itentire~y suppresses 'the transition of secondaryelectrons down to an anode voltage of about' 30V, asmay be seen fromfig. 6 in which the variation oftheanode current is given as a function of the anodevoltage. The anode current remains fairly constantupon decrease of the anode voltage 'to about 30V,and then decreases suddenly with a sharp bend.

la(mA) .t50

.27016

5

(1//

Vg2=100V

'VtOO

oo 20 40 60 . 80 Va(V) tOO

Fig. 6. Ia- Va characteristic of a tetrode with auxiliary electrodeas in fig. 5.

Secondary emission of insulators

Let. us consider an insulating part of a radiovalve which is struck by electrons, for instance aninsulating support. This component, which mayeasily have a large secondary emission factor,will be connected to the different electrodes overthe poorly conducting insulation. A given surfaceelement of the insulating support may now he 'considered' as a secondary emitting plate P (seefig. 7), which is conne~ted over a high resistance Rwith a type of average voltage Vo of the electrodes,which will for instance be somewhat lower than the

. anode voltage. This scheme corresponds exactlywith that. of a dynatron (see article quoted in foot-note 2), and the same is 'true of the current lp ofthe' secondary emitting plate as a function of itsvoltage. The characteristic is given in fig. Ba. With

,

214 PHILIPS TECHNICAL REVIEW Vol. 3, No. 7

increasing voltage lp at first increases, reaches amaximum and then decreases (it may even become .negative), because the coefficient of secondary

R

Vo Va.2700!J

Fig. 7. Equivalent circuit for an insulating part of a radiovalve which is struck by electrons.

emission increases so much in a given range ofvoltages that the number of secondary electronsincreases more rapidly than the number of primaryelectrons. As soon as the plate voltage is higher than<the anode voltage Va, however, the secondaryelectrons can no longer leave the plate and thecurrent suddenly increases rapidly to the valuewhich it would have in the absence of secondaryemission.What potential will the secondary emitting plate

assume in a stationary state? From fig. 7 the fol-lowing relation may be seen to' hold between the, '.current lp and the voltage Vp on the plate:

Vp = Vo - IpR.'

This relation is represented in fig. 8 (straight

Jp

hII

_III/I//I ..1

............. ----_/ .

I" ...

.27010

Fig. 8. Curve a: current lp of a plate with a high secondaryemission which is bombarded with electrons as a functionof the plate voltage. With a lower secondary emissio:Uthecurrent will vary according to the dotted-line curve.. Curve b: voltage. Vp of the same plate as a function of thecurrent when the plate is connected with a source of directvoltage Vo over a high resistance. .The points of intersection of the lines a and b determine

the stationary states. The adjustment has two values, givenby pand q. Point r is an unstable state. In state p the potential~f thè 'plate is higher than the voltage VOo ,

line b), together with the characteristic a which alsogives a relation between lp and Vp- The stationarystates are given by the points of intersection P» 'q and r. The intersection q always occurs: the inter-section pand r will not occur if the secondaryemission is so low that the characteristic has forexample the form of the dotted line curve. We see

. that the secondary emitting plate can adjust itselfto different potentials when the secondary emissionis high. The lowest potential (q) corresponds approx-imately to that of thè cathode, the highest (p) to .that of the anode; the adjustment r is not stable.

Fig. 9. Construction of a cathode ray tube. The collectingelectrodeas has the function of providing that the fluorescentscreen shall assume a positive potential upon being bombardedwith electrons, and in addition of capturing the secondaryelectrons.

Let us assume for example that the state ofthe plate was originally given by the workingpoint r, and that the voltage Vp has increasedslightly due to a fluctuation in the secondaryemission. The result is that the supply of electronsto the plate (Iprimary ~ [secondary) decreases morerapidly (curve a) than the loss of electrons over theleakage resistance R (curve b). The plate will there-fore assume a positive charge, and as a result itspotential will increase further until the stationarystate p is reached.

I

The fact that an insulator can reach a highpotential and give a high secondary emission uponbeing bombarded by electrons, sometimes leads toundesired phenomena in radio valves. Before goinginto this matter we shall show how this same phe-nomenon may be put to practical use in cathoderay tubes.. The fluorescent screen of cathode ray tubes con-sists of an insula~ing material and must be at a

JULY 1938 PHENOMENA VALVES CAUSED BY SECONDARY EMISSION 215

high potential III order that the electrons maystrike it with sufficient kinetic energy to producefluorescence. If the secondary emission of the screenwere lower than the primary electron current, thescreen would assume a negative charge and thehigh potential could not continue to exist. Thescreen must therefore have a high secondary emis-sion. Furthermore, care must be taken to ensurethat, upon switching on, the screen can reach a poten-tial higher than r in fig. 8, otherwise it would notadjust itself to the working point p but to point q.

Infig. 9 it is shown how the construction may becarried out practically. Close to the screen theinner wall of the tube is covered by an electrodewhich is raised to a high potential. This electrodehas a two-fold purpose: first that of providing thatthe potentialof the screen shall be such thatupon switching on a sufficiently high secondaryemission will occur, and second that of being ableto draw the secondary electrons away rapidly.

To

Vpq

Fig. 10. Current from an insulating component in a radiovalve as a function of the potentialof that part; a at a highanode voltage, b at a low anode voltage.

As was stated above, the secondary emission ofnon-conductors mayalso occur as an undesiredeffect on the insulators and glass walls of radiovalves. The phenomena may lead in two ways todisturbances: with power amplifier valves distor-tion occurs, with high-frequency amplifier valvesan ex tra d ampin g occurs in the output circuit.With power amplifier valves where the alter-

nating anode voltage has a large amplitude, thelp - Vp characteristic of an insulator in the neigh-bourhood of the anode will vary very much. Incertain cases the characteristic at the highest anodevoltages may be such that a working point p ofhigh potential exists, while this is not the case atthe lowest anode voltage, because the secondaryelectrons from the insulator are not drawn awaysufficiently rapidly (v. fig. 10). In such a case itmay happen that the potentialof the insulatorupon rise and fall of the anode voltage jumpsup and down between a high and a low value.Upon each jump in potential a charge may be

induced on the modulating grid which causestransients in the anode current v. (fig. 11).

In high-frequency valves, where it is desirablethat the anode current shall change little with the

Z7ZBZ

Fig. 11. Variation of the anode current for a power amplifiervalve with sinusoidal grid voltage. The curve exhibits anumber of irregularities due to a sudden jump in thepotentialof an insulating support or of the glass wall.

anode voltage (i.e. that the internal resistanceshall be high), difficulties may be experienced dueto secondary electrons which are emitted from aninsulator and reach the anode. Considering the factthat the insulator itself can conduct no current,this extra current is equal to the primary electroncurrent to the insulator. This primary current is in

a bFig. 12 a) Power amplifier valve in which the mica sup-porting plates and the glass wall are sbielded against bom-bardment with electrons by enclosing caps at the ends of thesystem. b) In an amplifier valve witb a gauze anode the glasswall is exposed to a strong bombardment by electrons. Inorder to avoid this the anode is surrounded by a gauze cagewhich is connected to the_cathode.

216I

PHILIPS TECHNICAL REVIEW Vol. 3, ,No. 7

AN APPARATUS FOR THE MEASUREMENT OF SCANNING SPEEDS OF CATHODERAY TUBES

general very greatly influenced by the potentialof the insulator which will change about equallywith the anode voltage. The secondary, electroncurrent which reaches the anode will therefore alsochange sharply with the anode voltage, and thismeans that the internal resistance, becomes con-siderably lower and the amplification of the high-frequency stage diminishes.

In order to avoid, these harmful effects hereagain several methods may be chosen. When, forexample, the disturbing phenomenum is -due toinsulating supports within the system, the mostpractical solution will be to cover them with asubstance whose secondary emission is less thanunity (for instance ~ layer of carbon). If the dan-ger spot is the inner side of the bulb, the jumpin potential can also be avoided byearthing thissurface through a capacity. In order ~o do this theoutside of the bulb may be covered with a layer

of metal which is kept.at a low potential (that ofthe cathode for instance). Finally lthe charging ofthe wall of the bulb may be avoided by entirely pre-venting electrons from leaving the electrode systemand reaching the glass wall. This ;may be achievedby means of shields. Fig. J2a shows a power am-plifier 'valve with a massive ail.od~ whose ends areclosed by caps which make it impossible for elec-trons to leave the anode. If an anode in the formof a gauze is used as is often necessary in order toimprove the heat radiation of the componentswithin the anode, this shielding must also. beextended behind. the anode. For this purpose awire gauze cage at, cathode potential is suitable.Electrons are unable to pass through this cage to

, the outside (fig. 12b). Shields with a low potentialmayalso be introduced in front of the muchused mica supports at the extremities of the systemin order to prevent their being struck by electrons.

by L. BLOK. 621.317.087: 621.317.755: 711.3

A simple apparatus is described for the determination of the greatest scanning speedof cathode ray tubes which can be photographed. 'The light spot on the screen is allowedto describe a logarithmic spiral, at every point of which the speed can easily he determined.The point ofleast intensity on the spiral which is still sufficiently visible in the photographdetermines the maximum scanning speed required. -

If, with the help of a cathode ray tube, phenom-ena are investigated which' are not periodic orstationary, but only occur once, it will in generalbe 'found practical to record the, phenomenonphotographically. In the most commonly occurringcases the details cannot be adequately studied byvisual observation.

As examples of phenomena which are investigatedin this way we may mention the switching on oftransformers, break-down in the testing of insu-lating materials, back-surge in rectifiers, atmosphericdischarges, etc.

In the photographic recording of such a phenom-cnon the blackening of the light-sensitive plate

I will depend on: 'a) the brightness B of the cathode ray light spot,h.) the diameter of the spot D,c) the velocity v at which the spot moves,d) the optical enlargement N,e) the light sensitivity 1] of the photographic

material.

A detailed treatment of the relation betweenblackening and the factors a) to e) has alreadybeen published in this periodical,1), together witha discussion of measurements relating to the max-imum scanning speed. As a control in the manufac-ture of cathode ray tubes it is desirable to be ableto determine this scanning speed' in a simple way.In the following a description is given of a meas-uring apparatus solved for this purpose.

Principle

In order to determine the scanning speed bymeans of a single photograph it is necessary toallow the spot to describe a path on which thevelocity changes continuously', while at every pointof the path the velocity must be known with suf-ficient accuracy. One may then determine on thephotograph the point where the path of the spot

1) J. F. H. Cu st er s, The recording of rapidly occurringphenomena with the aid of the cathode ray tube and thecamera, Philips techno Rev. 2, 148, 1937. .

- . ,,/

)/