US2720611A - Electron tube structure - Google Patents

Description

Oct. 11, 1955 SLOAN ELECTRON TUBE STRUCTURE 2 Sheets-Sheet 1 Filed NOV. 5 1951 H IM-HHI HH R mm H H -HIM-H INVENTOR. DAV/D H. SLOAN A TWP/V516 Oct. 11, 1955 D. H. SLOAN 2,720,611

ELECTRON TUBE STRUCTURE Filed NOV. 5, 1951 2 Sheets-Sheet 2 [l5 I27 27' I23 I13 I l I I FIIE E IN VEN TOR. 0.4 we H. SL omv BY. W W

United States Patent Ofifice 2,720,611 Patented Oct. 11, 1955 ELECTRON TUBE STRUCTURE David H. Sloan, Berkeley, Calif., assignor to Research Corporation, New York, N. Y., a corporation of New York Application November 5, 1951, Serial No. 254,913

16 Claims. (Cl. 3156) This invention relates to electronic tubes and, particularly, to tubes adapted for the generation or amplification of electrical waves in the microwave or centimeter region, specifically tubes of the resnatron type such as are disclosed in the prior Patent No. 2,424,002, granted July 15, 1947, and by various other patents and prior applications of the present inventor, among these being application Serial No. 771,852, filed September 3, 1947, now Patent No. 2,641,734, and application Serial No. 221,017, filed April 14, 1951, now Patent No. 2,653,273.

Tubes of this general type are characterized by the fact that they employ density-modulated electron streams (as distinguished from velocity-modulated streams such as are developed by klystron tubes) developed in cavity resonators tuned substantially to the desired frequencies of operation. Control in such tubes is exercised by a grid structure formed in the wall of the cavity, with a cathode in close proximity to the grid so that the changing shape of the grid-cathode field in the course of the oscillation controls the electron flow. In general, and specifically in the case of the present invention, the anode which receives the electron stream is positioned in a separate cavity tuned to the same frequency as the cathode-grid cavity. In tubes of this character the transit time of the electrons, at least in the anode cavity, is of the order of magnitude of the period of oscillation of the device; the electrons acquire energy by acceleration in a unidirectional field and deliver this energy to the anode cavity by deceleration in the oscillating field set up within the cavity.

Resnatrons have been built using various numbers of electrodes, i. e., as triodes, tetrodes, and pentodes. They are primarily high-power devices, delivering many kilowatts or even megawatts at the extremely high frequencies at which they operate. It has been found that the most advantageous type of structure for many purposes is the tetrode, wherein most of the energy is imparted to the electron stream by an accelerating electrode, preferably biased to the same potential as the anode.

In prior tubes using this construction, two entirely separate cavity-resonator structures have been utilized in order to permit the application of the proper biases to the various electrodes. One of these cavities is the cathodegrid cavity, wherein the cathode and grid operate at the same or relatively slightly dilferent bias potentials. The second cavity has been the accelerator-anode cavity, completely insulated from the first-mentioned cavity, operated at a bias potential several thousand volts higher. The arrangement just described requires that there be a space between the two resonant cavities through which the density modulated electron stream passes. This space possesses the characteristics of and can act as a cavity resonator, and there is therefore a tendency to set up oscillations within it at the frequency to which the two designedly resonant cavities are tuned. To the extent that such oscillations are set up in the intercavity space power is wasted. By detuning the space, and by introducing damping therein, the waste of power can be made relatively small, percentagewise. What power is generated within resonant cavity.

this space must, however, be got rid of in some manner; a part of it will be absorbed as heat in the device, but the most practical method of getting rid of it is to let it radiate into the surrounding space. In tubes of a few kilowatts rating the problems involved in permitting such an escape are usually not serious, but where the device is generating many hundreds, or even several thousands of kilowatts, the radiation of a very small percentage of the total power may cause interference, local heating of surrounding materials, and otherwise cause troubles which are very serious indeed.

Other types of cavity-resonator tubes than the tetrode thus described might, to advantage, use auxiliary electrodes, in or adjacent the electron stream, which for best operation should be operated at high frequency potentials differing from those of the walls of an anode or cathode In the past such tubes have not been constructed because of the structural complications involved in the provision of an additional cavity, whether resonant or non-resonant.

One object of the present invention is to provide a tube capable of handling the large powers mentioned and, at the same time, to avoid the dilficulties which have hitherto been inherent in resnatron tubes employing accelerator electrodes. Among other objects of the invention are to provide a tube structure of the resnatron type which employs an accelerating electrode, but wherein only two resonant cavities are employed, with no interspace between them; to provide a tube structure of the character described wherein all of the oscillatory power absorbed from the electron stream appears directly in the anode cavity; to provide a resnatron type electronic tube structure wherein the power wasted through radiation from the tube structure itself is a minimum; to provide an electron tube structure of the character described which may be operated either as an amplifier or as an oscillator and wherein, when employed in the later manner, the anode circuit eflieiency may be made relatively high so that power is delivered into the anode cavity in pulses occurring during the voltage peaks, thus developing the maximum power for a given electron flow; to provide a tube structure whereby an auxiliary electrode may be located within a resonant cavity and there operate at any desired fraction of the oscillating potential of the cavity walls, including substantially zero, and in phase with the potential of either wall as desired; and, in general, to provide an electron tube structure wherein the mechanical and, particularly, the insulation problems are reduced to a minimum.

For the purposes of this specification the invention will be considered as it applies to a tetrode employing an accelerating electrode, the broad principles applying to the use of any auxiliary electrode being readily apparent therefrom.

Considered broadly, the portion of the tube structure with which this invention primarily is concerned comprises a resonant cavity with the electrode structures located at potential loops in opposite walls of the cavity, or, what is the same thing, at opposite poles of maximum electric field concentration across the cavity when the latter is resonating in a mode corresponding to the desired frequency of operation of the device. Means are provided for directing a stream of electrons across the cavity of the resonator between the electrodes and generally parallel to the electric field. Within the cavity, along the path of the electrons therethrough, is located an auxiliary electrode, e. g., an accelerator electrode. The latter may be of a grid or mesh structure, but in tubes for the generation of very high power, to which the invention is primarily directed, it is more practical to form an accelerating electrode as an annulus surrounding the electron stream. In the form of the device wherein the stream is itself annular in form, the accelerating electrode can include an additional annulus substantially coaxial with the outer annulus and just within the path of the electrons. The auxiliary electrode is so mounted within the cavity that its oscillating potential is substantially unaffected by conduction through its supports, but floats between that of the opposing walls of the :cavity in inverseproportion to the capacities of theauxiliary electrode to the walls between which it is located being midway the potential of the walls if these capacities are equal. This resultis achieved by employing the following expedients;

First, the auxiliary electrode is supported by one or more struts .of inherently high inductance, so that large potential differences can be developed across them by relatively small oscillating currents; this means that .the cross sectional area of the struts should be as. small as .is consistent with adequate mechanical :strength. Second, the ,lengthof the strutsshould be such that, viewed from their .points'of attachment, their impedance approaches infinity, i. e. they are anti-resonant. Third, Where possible, the struts should be attached to the cavity wall at a pointsuch-that the energy fed into them from the cavity will reach the accelerating electrode in phase with the energy transmitted thereto through its capacities and there develop the same potential. A separate bias supply for the auxiliary electrode may, if desired, be supplied. Ordinarily, however, whereit is an accelerating electrode, it is, more practical to connect its support directly to the anode structure itself .and operate it at the anode bias potential. The remaining structure of the tube, to which this invention is not specifically directed, may take various forms. In the case of the tetrode here chosen .for illustration of the principles, the complete tube uses but two .cavity resonators having a common wall whereinthe grid structure. is located and which are tuned to oscillate. at. substantially the same frequency. .T he cavities areso proportioned that the cathode and grid are positioned at opposite locations of maximum potential field concentration. in one. cavity, while grid and anode are similarly located in the .second,'all three electrodes being alined.

The accelerating electrode is located in the grid-anode cavity between the two electrode structures. therein. With this arrangement the electron stream enters the grid-anode. cavity directly through the grid, without any intervening. space wherein energy may be wasted. The grid and anode, of course, are operated at widely different D. .C. potentials- The. general methods of construction whereby the: necessary insulation may be. maintained and the. cavity substantially closed, electrically, although the :anode and. .grid' structures are mechanically spaced, are in ac cordance with the disclosures of the prior patent :and ap-- plicationsalready mentioned. a

The .abovemay be .more clearly appreciated by refer ence to thedetailed description of two preferred forms of the, device which follows, taken. in conjunction withthe accompanying drawings wherein:-

Figs. 1 and 1A together comprise an axialsectional view ofanaxial-fiow .resnatronernbodying the invention, the view being broken at the. anode-support end .of Fig.

. l and this .end shown separately in Fig. 1A in {order to the anode structure. respectively. Each of these structures is a composite, but since the general mode of con-. struction .is much the same as that used in the resnatrons described inthe prior applications and patents which have already been mentioned, they will be described generally without going. into detail as .to the method of assembly ofthe parts except. where such detail: is pertinent. to: the inventionrherein covered.

Starting with the cathode structure at the left-hand end or the diagram, a central rod conductor I is surrounded by a tubular column 3, these being the conductors through which current is supplied to a heating coil 5. An insulating seal 7 holds these two conductors in proper relative position.

A slightly dished cathode-9 is supported on the end of the column3. A sleeve 11 surrounds the column 3 and chokes 41, 41' and 41 carries high frequency chokes 13, 13. A flexible diaphragm 15 connects the end of the column '3 with the outer periphery of the choke 15', and the sleeve engages the column frictionally so that, during construction, it can he slid on the column properly to tune the cavity of which the diaphragm forms one wall.

The cathode structure-is supportedfrom the more massive grid structure through a flange 17 secured to and surrounding the column 3 adjacent its outer end, this flange connecting to thegrid structure through a seal ring 19. The seal 19 attaches to a skirt 21 projecting from a;

massive ring 23. This ring surrounds and is attached to a tubular extension 25 projecting, toward the left of the diagram, from-the bottom of a grid cup 27, The grid cavity 29 is formed between the. bottom .of this cup and order to match the impedance of the guide with the im-.

pedance of the cavity. 1

The grid itself comprises a plurality of grid bars .33

positioned across. an axial aperture formed in the bottom of the grid cup 2.7. These bars are .curved so as to follow generally the contour of the dished face of the cathode 9, which they face directly. Due to this formation, electrons liberated from the cathode are focused or converged as they pass through the grid and theelectron stream entering the cavity within the cup is generally a converging cone in form. g

'The grid cup is secured through a pair of heavy flanges 35 and .37 to. a tubular skirt 39 within which areformed, conjointly with the. cup, a plurality of high frequency Surrounding the skirt 39 is a second skirt 43 connecting through a huge 45 with a glass body 47, the right-hand end of which is re-entrant to form a sealed connection 49 with a flange 51. This flange is secured to a column 53 which supports the anode structure 54. The latter is tubular, passing completely through thecolumn 53 and the skirt 39 intothe interior ofthe gridcup. Its inner end is borne by spring fingers55 on the end .of the column 53. The .outer .end connects. through a flange 57 and metal bellows .59 with the-columns?) to form an air-tight seal. The end of. the anode tube S'Aaisclosedhy .a cap-61, which, in. the present.

case, is channelled to form a connection through which cooling liquid may be admitted and circulated through the anode structure .via liquid connections 63'and 65.

The anode may be advanced and retracted by means of v a threaded stud 64, secured to the end of the anode. tube, by means of an internally-threaded worm wheel 67 engaging with a tuning worm 69, these latter port-ions being supported by a cylindrical fitting 71 mounted .to the flange -'57. Useful. power is withdrawn from the cavity through'a' tapered wave guide 8'1.v

'The inner end of the anode tube carries the. anode proper, which is an annulus- 72 on the end of the anode column '54 and provided with a relatively sharp lip 73 facing the grid within the cavity of the grid cup.

The acceleratingfelectrode with the positioning and the anode atria position displaced outwardly of. theanode tuhefromlip 73. Thestrnt 7 7- secured 10km outer edge of flange 76 so as to lie almost entirely outside of the field between the ring 75 and the anode lip 73, but close to the potential loop.

Despite the difference in size the anode cavity is tuned to the same frequency as the cathode cavity, its capacity being smaller although its inductance is greater.

In the description of the prior inventions of this same inventor it has been disclosed how such cavities are eifectively closed and the escape of energy therefrom prevented. This latter is eifected by the combination of approximately quarter-wavelength chokes offering high impedance in the direction away from the cavity and low impedance transversely of the coaxial structure. In particular, in the Sloan Patent No. 2,424,002 already referred to, it has been shown that the transmission line sections going to make up this arrangement need not be of precisely one-quarter wavelength in order to achieve the effects desired and effectively suppress radiation from within the structure.

As a result of the construction of the tube of Fig. 1, when oscillating in its most readily excited or principal mode, the magnetic field is circular and coaxial with the device. The electric field is approximately parallel to the axis in the region between the lip 73 of the anode and the grid, becoming radial between the wall of the grid cup and the anode column. Maxima of electric field occur at the lip 41 of the grid cup and between the grid and the lip 73 of the anode. Such maxima may be termed, for convenience, loops of potential, and the positions of maximum concentration of field on the cavity walls the poles of such potential loops. The tube is designed to operate in accordance with this mode of oscillation. When so oscillating the primary effect of the accelerator electrode, so far as the oscillation within the cavity is concerned, is merely to lower the frequency slightly by increasing the capacity between grid and anode, owing to the thickness of the electrode in this dimension.

This would not be the case were any material energy to be fed to the accelerating electrode through its supporting strut or struts. As was indicated in the broad description of the invention such supply of energy is prevented by making the impedance of the strut high, as viewed from its point of attachment to the cavity wall. it the strut were attached to a point on the wall that varied in oscillating potential in phase with the accelerating electrode when excited only by its capacities, and to the same degree, when referred to a common datum, the impedance of the strut would be immaterial since no potential difierence would exist across it and hence no currents would be carried by it, irrespective of the impedance of the strut considered by itself. Such an ideal condition is ditficult to attain in a practical type and hence the impedance of the strut itself is made as high as possible. its diameter is made as small as is consistent with its necessary mechanical strength, thus raising its inductance and lowering it capacitance per unit length, so that high oscillating potentials may be developed across it by relatively small currents. This gives a mismatch with the impedance of the cavity wall and with the electrode 76 as well.

Considered as portions of a transmission line system, reflections occur at both ends of the strut 77 and very little power can be transferred through it provided it be properly terminated. For this reason, the strut may be considered as oscillating independently of either of the two structures to which it is connected; i. e., as a dipole.

Such oscillation in the strut implies that it be onehalf wavelength long, and that it connect to the cavity at or very near a potential loop. In the tube of Fig. l the strut is roughly in the form of a partial helix about the tube axis, so that its length closely approximates onehalf wavelength at the mid-frequency of the tuning range of the tube, and it connects as near to the loop of po-.

rrv ID tential at the anode lip as is feasible without materially distorting the field between anode and accelerator.

The object in the design of this tube is to have the potential of the accelerating electrode determined almost entirely by its relative capacities to the grid and anode, the potentials between the accelerator and the other two electrodes being inversely as these capacities. As will be shown hereinafter, in connection with the discussion of tube operation, it is usually desirable that the potential of the accelerator be allowed to swing about its bias potential in phase with the grid, so that one-third, more or less, of 'the potential across the cavity exists between grid and accelerator and the remainder between accelerator and anode. The effective capacity of the accelerator to the grid is therefore made from two to three times as great as that to the anode, the energies transferred through these capacities is in inverse proportion thereto, and the potential of the accelerator will therefore follow the grid. Hence, since the strut is one-half wavelength long, such energy as is fed through it arrives in phase with that from the grid capacity and is simply reflected, having no material effect.

If the accelerator is in the neutral plane or swings with the anode, the result is the same as long as the fields from the anode to the accelerator and grid respectively are in phase. This will be the case as long as the impedance looking into the strut from the cavity is high in comparison to that offered by the capacities of the accelerator to the walls of the cavity. When this condition is met, the effect of the accelerator on the cavity oscillation is almost solely that due to its thickness and the efiect of the struts can be ignored.

If the strut be somewhat less than one-half wavelength, the result is little changed; in effect it borrows enough capacity from the structures to which it is connected to tune it to the half-wave value, reducing the efiective capacity between anode and accelerator to the extent of that borrowed and increasing the oscillatory potential difference between these electrodes correspondingly. Alternatively, the strut may be considered as a small inductive susceptance in parallel with the capacitive susceptance of the accelerator to the anode, and hence subtracted from the latter. A strut slightly over one-half wavelength long would have an opposite effect, increasing the accelerator-anode capacity.

High input impedance to the strut can also be obtained by a quarter-wave strut connected at a potential node. In this case it is current fed and can be considered as one-half of a center-fed dipole. The efiect of minor mis-tuning is therefore precisely the same as in the halfwave case already discussed.

While struts an integral number of quarter-wavelengths long lend themselves most readily to analysis, it should not be assumed that they are the only means of obtaining supports whose theoretical impedance approaches infinity and the actual impedance of which is high. Dipoles may be loaded with either inductance or capacity, and any point of attachment to the cavity wall will have an apparent impedance which will approach zero or infinity at current and potential nodes, and at intermediate points will appear inductive or capacitive depending on the point from which it is viewed. The strut, considered alone, will appear inductive if less than one quarter-Wavelength long, capacitive if over a quarterwavelength and less than one half. It is therefore theoretically possible to find some point of attachment for the strut within the cavity where, viewed from the auxiliary electrode, the impedance of the cavity will so load the strut as to make it anti-resonant, and where this occurs the energy transmitted through the strut Will be a minimum.

For obvious mechanical reasons the struts should be as short as possible, which limits the choice of points of attachment to the cavity walls. Half-wave or quarterwave struts attached at current or potential nodes are pedance will approach infinity at some specific frequency and where if the-cross section of strut be-small and:

its inductanee high, small-departures frorn'this frequency 'will result -in only slight changes intheoscillafin'g' potential of the" auxiliary electrode, due to capacity laorrowed-fromor lent tothe strut. a V

"The general principles: underlying the design of the struts arethus we'll known, but their application to the pro-design eraspecific tube'may 'be difficult. The cavities employed are of "suchcomplex formthat com putation of the field conformations is not feasible, and the locat'ion of nodes.can be-predetermined only roughly. Moreover, any individual strut'wfl usually be assymetri' cally placed with respect to the-cavity as a Whole, capacity cannot be expressed byany reasonably simple function, and hence its electrical length depart from its physical lengthby an indeterminate amount.

Cold test is therefore preferably relied 'upon in-the final design of tubes embodying this invention. The accelerating electrode is mounted in-the-cavity of a mock-up of the-tube resonator :oninsulating supports, such as silk threads or even toothpicks, and the cavity is excited at a-desiredfrequency from an external source. position is adjusted until the potential swingv ofthe accelerator bears the desired ratio to that of the anodegrid structure. Conductive struts are then inserted, connecting the parts as nearly at the desired points, as de termined by judgment or'roug'h computation, as-poss'ibio, and of approximately the'proper length. The 'efiect upon the resonant frequency of the cavity is then measured. This effect should be substantially If it is not, the length or point of attachment of the-strut is varied until its effect upon'the 'reasonant frequency of the cavity 'isreduced to negligible proportions.

Ideally the strut shouldbe positioned perpendicular to the electric field, i. e.,' in aunipotentia-l surface. Except in cavities that are almost perfectly symmetrical-e. g., a right cylinder-this is impractical. The practical expedient is to make it anti-resonant in Whichcase all points along its length will be very nearly at the potential oft he adjacent space even if the strut be "actually parallel, instead of perpendicular to the field. minimize distortion of thefie'l'd and pick-up of energy from itby the strut. "This condition can ,at best be approximated in practice- Since the tube is tunable, the maximum impedance condition for the strut wm not be met at all ire quencies within the normal tuning range. If the" strut be so dimensioned and positioned that its effect is nil at approximately the ceuter of the range, the capacities etfectivelyiadded :toor subtracted from those of the acceleratorwill be proporti'onatelysmall at any operating frequency; therefore neither'the tuning range nor the potentials appearing'onjthe accelerator will be seriously a'fl'ected. Hence it is not necessary that the condition of zero, e'fi'ect on frequency be met at the precise frequency chosen for test; if the effect is small enough, it will be zero. at v [some closely. adjacent frequency and within tolerance at all frequencies within the tuning range.

There are various combinationsof operating frequency, accelerating, potential and, hence, transit time at which the device may be operated. Two such combinations deserve especial-comments In accordance with the first. the accelerating potential is so adjusted as to make the transit time across the cavity substantially equal to /2 cycleof the resonant frequency, for electrons, entering the cavity at optimum phase. Under these circumstances, such. electrons are delivering energy to the oscillating field duringtheir entire transit, and this might appear to be the optimum condition, of operation. -It results, however, in a'had distribution of electronflow; the variation in electron velocity d'ueto the superposition of the oscillating field on the accelerating field results in a relatively wide disparity in time of arrivalof'electrons at the anode, even though the bias in the cathode-grid cavity be so adjusted that most of the electrons enter the anode cavity at approximately the optimum epoch of the 'cycle,,the instant of reversal of potential within it. The result is that many electrons absorb-energy from the field instead of delivering energy thereto, resulting in relatively low anode 'efiiciency in the device.

Ineifect this mode of operation results in un-bunching of electrons and an effective reduction in the degree of density modulation imparted to the electron stream by the grid. Such 'a result is unfavorable; an opposite and favorable effect can be secured by operating the tube in the manner next to be described.

The second mode of operation involves the adjustment of the accelerating potential so. that the transit time is one whole cycle, divided substantially one-half between grid and accelerator, the other half between accelerator and anode. Under these circumstances, optimum-phase electrons are accelerated by and therefore absorb energy from both the constant held and the oscillating field during As has already been stated, the oscillating potential between grid and accelerator is preferably one-third to onehalf that between accelerator and anode. The energy of any electron. is directly proportional to the potential through which it falls; therefore the energy absorbed from the oscillating field will be from one-half to onethird that delivered back to that field during the second part of the journey and the net energy delivered to the field will be from one-half to two-thirds that so delivered in the latter portion of the trip. For maximum efficiency this net energy should be equal to that absorbed from the bias field, and the electrons should arrive at the anode at relatively low velocity.

Such an optimum condition cannot be met in practice for all electrons; there must be some distribution in time of entry into the cavity, the amounts of energy absorbed from :and delivered to the field will vary with phase, while the D. C. energy remains a constant. Where the average transit time is a full cycle, however, the velocity modulation imparted to the originally density-modulated stream by the accelerating electrode results in a bunching of the electron flow, so that many more of the electrons therein make their transit during optimum phase conditions and energy is delivered in short. pulses at the most eflicient epoch of the cycle. Such conditions of operation are somewhat analogous. totype C operation of ordinary low frequency triodes, where high anode efiiciency is secured by similarly pulsing .an oscillating circuit at optimum phase.

In order to achieve such operation, the spacing of the accelerator must be so proportioned as properly to divide the transit time between the electrodes, taking into account the varying acceleration in the grid-accelerator space and deceleration in the accelerator-anode space. This spacing must be reconciled with the necessary differences in capacities between the accelerator and the opposed cavity walls. The latter, however, are controllable separately from the spacing by varying the size and .position of the flange 7 6 and the conformation of the cavity walls opposed to the accelerator. In the tube of Fig. 1' the accelerator and anode face each other edge to edge; 7

grid "and accelerator more nearlyflat to fianwhich' gives the desired capacity difl'erence. I It should not be inferred from what has been said howter. The two modes are discussed hereas representative of two special cases and not as the only modes of operation of which the device is capable.

In order to give some idea of the approximate dimensions and power capabilities of tubes of the character here described, it may be stated that that illustrated in Fig. 1 has an over-all length of 30" and is capable of delivering an average power output of 20 kilowatts, delivered in megawatt pulses.

The tube illustrated in Fig. 2 is capable of delivering power at a still higher rate and is shown to illustrate the application of the same general principles to resnatrons of the annular cathode type, the general principles of which are set forth in copending application Serial No. 771,852, above referred to. This particular tube is designed for zero grid bias operation so that no chokes or insulators are provided between the grid and the cathode. Without going into the details of constructions, since the latter are not pertinent to the invention herein specifically claimed, the cathode 161, like that of the tube first described, is of the so-called Phillips type, formed of sintered tungsten powder impregnated with a barium oxide and capable of high-density electron emission. It is heated by electronic bombardment from a small filament 103 of the ordinary thoriated tungsten type. The cathode is annular in form, and, as already indicated, is electrically connected to the grid so as to operate without grid bias. The input to the cathode-grid cavity is through a coaxial transmission line sleeve 105' and a central conductor 107 which also carries a supply for cooling fluid. This transmission line may be fed by wave guide 109 and the whole arrangement is so tuned that the cathode 161 and the grid 111 which faces it are at the potential loop nearest the shorted periphery of the cathode-grid cavity.

An annular anode 113, having a deep electron receiving slot 115 formed therein, is mounted coaxially with the cathode and grid so as to receive the annular electron stream therefrom. A plurality of annular cups 117, 118, 119 and 120 surround the anode structure, which is water cooled through channels 121, and connects with a tuning mechanism, indicated at 123, for varying the capacity of the anode cavity.

The anode cavity may be considered as terminating in potential node at the bottom of the cup 117. Viewed from Within the cavity a current node occurs substantially at the lip of the cup one quarter-wavelengti1 from the bottom, and the outer conductor can therefore be opened at this point, by the space between the cup and the outer wall, without affecting the mode of oscillation or permitting the escape of large amounts of energy. Such energy as does escape through this space is attenuated by the successive cups 118, 119 and 120, as is described in the above mentioned patent to Sloan and Marshall.

The accelerating electrode comprises two annuli 125, 125, mounted within and without the annular path of the electron stream, on struts 127 and 127' respectively. A potential node occurs substantially where the outer wall of the cavity starts to flare, and struts 127 connect to the anode structure at the potential node just discussed, although, owing to the complex shape of the cavity, this is experimentally determined. They are substantially a quarter-wavelength long and their size is such that they offer a large impedance mismatch. Hence they transfer little energy to the accelerating electrode, even though the tube be tuned to a frequency which will displace the nodes slightly from the position of attachment of the strut. At the mid-frequency of oscillation of the device, however, the ideal conditions of nodal attachment are substantially met in this particular design.

In tubes of this type an additional problem enters with respect to the number and disposal of the struts. One of the primary reasons for exciting the accelerating electrode capacitively from the walls of the cavity is that this leads to its operation at substantially uniform potential around its entire circumference. To the extent that energy is supplied to it through the struts this uniformity is upset;

this is bound to happen when the tuning of the cavity is off of the anti-resonant frequency of the struts and the latter borrow or lend capacity to the accelerator. When it occurs circulating currents flow around the accelerator rings with consequent waste of energy.

As long as the differences of potential around the accelerator rings are relatively small in comparison with the potentials of the accelerator with respect to the cavity walls this condition may be tolerated, as resulting in merely a minor loss. If the potential differences around the rings become greater the currents rise, causing greater losses. Moreover, the potential differences may become great enough to disturb the uniformity of the electron stream, and under certain circumstances this can increase the original potential differences, so that the effect becomes cumulative.

This problem does not arise in tubes of the type shown in Fig. l, as the size of the accelerator electrode is so small in comparison to the wavelengths generated by the tube that no serious differences of potential can be set up around its periphery. In tubes of the type of Fig. 2 it may become serious if precautions are not taken to avoid it.

It is obvious that the worst condition from this point of view would be that arising if the fundamental resonant frequency of an accelerator ring coincided with the operating frequency of the tube. This is unlikely to occur; with the accelerator ring spaced radially approximately a half-wavelength from the tube axis their circumferential length will almost certainly be much greater than one waveiength even in complex cavities. It would be possible for the rings to be so proportioned that they could resonate on harmonics of their fundamental frequencies if these harmonics were approximately the operating frequency of the tube, but fortunately the mechanical features of the design are more likely to bring the operating frequency between the third and fourth harmonics of the resonant frequency of the outer ring and between the second and third harmonics of that of the inner one of the accelerating electrode; care in design can insure that harmonic resonance is absent.

The parasitic oscillations which may be excited in the rings are therefore unlikely to be of resonant character. None the less their magnitude can be influenced to a very high degree by any factor which varies the circumferential impedance of the rings as viewed from the points of attachment of the struts. One such factor is the angular spacing of the struts; another is the characteristic impedance of the accelerator rings to circulating currents. The two factors are interdependent, but for the purpose of the present invention the former is the more important. With the particular tube illustrated in Fig. 2 the use of six struts for the outer accelerator ring and four for the inner reduced the circulating currents to negligably small values; the use of eight struts for each ring resulted in large circulating currents. Since any factor changing the impedance of the rings may also change the optimum number of struts the latter is best determined by cold test as above described, but if a number of struts dictated by mechanical consideration results in large parasitic currents of this character it is almost always possible to so load the rings, and thus change their characteristic impedance, as to reduce such currents to tolerable proportions.

Although the tube of Fig. 2 is considerably shorter than that shown in Pig. 1, being approximately 20 inches long as actually constructed, it is capable of much higher output; approximately 20 megawatts peak power.

In the tube first described the grid and anode are located on the axis where a potential loop must form in any mode of oscillation where the magnetic field is circumferential; The tube of Fig. 2 has grid and anode displaced radialiy from the axes, and in order that it may operate sarisfactorily, the cavity must oscillate in accordance with a more complex mode than that utilized in operating the tube of Fig. 1. For the purpose of this specification,

modes of oscillation where the magnetic fieldsare all circular, andthe nodes and loops 'in'the electric field therefore are also circular, will be referred to as principal modes'of oscillation, to'dist'inguishsuch modes from those having nodes and loops angularly'disposed' around the circumference.

It will be understood that the invention as specifically described in connection With the two forms of resnatron illustrated herewith are merely examples, of many forms in which the invention can be utilized. Several of the resnatrons' shown in the prior patents and applications listed herein can obviously be modified toemploy the principles herein set forth and many additional designs can be evolved to place an auxiliary electrode within an otherwise necessary cavity' and avoid an nntuned intermediate cavity or space through which power may be Wasted. It is desired, therefore, to protect the invention as broadly as is possible within the scope of the following claims.

I claim: I

1. An electron tube structure comprising a cavity resonator, an electrode structure positioned within said resoof the oscillating electric field between the walls thereof when said resonator is oscillating at a principal mode, means for initiating an electron stream across the cavity of said resonator substantially along the lines of said field concentration, an auxiliary electrode positioned intermediate the walls of said cavity in line with said electron stream and apertured to permt the passage of electrons therethrough, and conductive means for supporting said auxiliary electrode, said conductive means being proportioned to have a high impedance at said principal mode of oscillation in comparison with the capacitive impedance of sad auxiliary electrode to the cavity walls between which. it lies, whereby the oscillatory potential of said auxiliary electrode with respect to said walls is determined primarily by its capacities thereto.

2. A structure as defined in claim 1 wherein said conductive means comprises at least one strut connected to a wall of said cavity and dimensioned to offer a mismatch of impedance with respectto the wall whereto it is connected' sufiicicnt to cause reflection of a major portion of oscillating energy carried by said strut.

" 3. A structure as defined in claim 1 wherein said conduc tive means comprises at least one strut substantially an integral number of quarter wavelengths at said principal mode of oscillation in length and connected to the inner wall of the cavity of said resonator'at a position of maximum impedance.

4. A structure in accordance with claim 3 wherein the length of said strut is substantially two quarter wavelengths and its connection within said resonator is substantially at a point of maximum concentration of electric field.

5. A. structure in accordance with claim 3- wherein the said length of said strut is substantially one quarter wavelength and its connection within said resonator is substantially at a point of minimum concentration of electric field.

6. An electron tube structure comprising a cavity resonator, an anode structure within said resonator substantially at a pole of maximum concentration of the oscillating electric field between the walls thereof when said resonator is oscillating at a principal mode, means for initiating an electron streamacross the cavity of said resonator substantially along the lines of said field concentration, an

accelerating electrode within said. cavity and having an opening therein positioned to permit the passage of said electron stream therethrough, and conductive supporting means connecting said accelerating electrode to the walls of said cavity and dimensioned to present an impedance high in comparison to the capacitive impedance between 1' 2 said accelerating electrode and the walls of the'cavity' at the frequency of said mode.

7. An-electron tube in accordance with claim'6 wherein saidconnecting means comprises at least one supp'ort- 7 ing strut connectedto said accelerating electrode and to the interior of said resonator, the cross-sectional area vide an impedance mismatch at the points of connection na'tor substantially at a pole of'rn'a'xi'mum concentration of said strut.

8. An electron tube in accordance with claim 6wherein said accelerating electrode comprises a conducting annulus surrounding the path of said stream of electrons. 9. An electron tube comprising a cavity resonator, an

anode structure and a grid structure positioned in opposite walls of said cavity resonator substantially at the poles of a maximum of electric field therein when oscillating at a principal mode, a cathode positioned-externally of said cavity resonator to direct a stream of electrons through said grid structure toward saidanode structure, an electrode so positioned within said cavity and between said poles as to accelerate said'stream, and connections for applying an accelerating potential to said electrode.

10. An electron tube in accordance with claim 9 in cluding a second cavity resonator having in common withsaid first mentioned cavity resonator the wall including said grid structure, said grid structure and said cathode being located substantially at opposite poles of maximum of electric field within said second cavity resonator whenoscillating at the same frequency as said first mentioned cavity resonator.

11. An electron tube in accordance with claim 9 wherein said electrode comprises an annulus surrounding the path of said electron stream.

12. An electron tube in accordance with claim 9 wherein said connections comprise at least one supporting strut for said electrode, said strut connecting to the interior of said resonator and mismatched in impedance therewith.

13. An electron 'tube in accordance with claim 9 where'- in said anode and grid structures and said cathode are coaxial and annular in form, said resonant cavity is substantially symmetrical with respect to the axis of said structures, and said electrode comprises coaxial annuli surrounding and surrounded by, respectively, the path of said electron stream.

14. An electron tube structure comprising a cavity resonator, an anode structure and a gridstructure within said resonator cavity on opposite sides thereof, an'electrode mounted within said cavity between said structures and conductive supporting means for said electrode connected tothe wall of said cavity resonator and mismatched therewith in impedance whereby the alternating potential of said electrode is determined primarily by the relative capacities thereof with respect to said anode and grid structures.

15. An electron tube structure in accordance with claim 14 wherein the capacity of said electrode with respect to said grid structure is materially greater than its capacity with respect to said anode structure.

16. An electron tube structure in accordance with claim 14 wherein the capacity of said electrode with respect to said grid structure is from two to three times as great as its capacity with respect to said anode structure.

References Cited in the file of this patent UNITED STATES PATENTS Lehmann Mar. 1, 1949

Images