US2239149A - Electronic amplifier - Google Patents

Description

April 22, 1941. P. "r. FARNSWORTH 2,239,149

ELECTRONIC AMPLIFIER Filed Jan. 5, 1938 4 Sheets-Sheet l A TTE/VUA T/0/v INVEN TORT PH/LO T. FA RNSWOR TH.

BY I E %RNEYS.

April 22, 1941.

.P. T. FARNSWORTH 2,239,149

ELECTRONIC AMPLIFIER Filed Jan. 5, 1938 4 Sheets-Sheet 2 j I I I I I I I INVENTOR PH/LO 7." FA/PNS WORTH.

April 1941- P. T. FARNSWORTH 2,239,149

ELECTRONIC AMPLIFIER Filed Jan. 5, 1938 4 Sheets-Sheet 5 INVENTORT PH/LO 7: FARNSWOPTH u ATTORNEYS.

April 22, 1941.

P. T. FARNSWORTH ELECTRONIC AMPLIFIER FiledJan. 5, 1938 4 Sheets-Sheet 4- JSL m j INVENTORT PH/LO 7T FARNSWOR TH.

A TTORNEYS.

Patented Apr. 22, 1941 ELECTRONIC AMPLIFIER Philo T. Farnsworth, Springfield Township,

Montgomery County, Pa., assignor, by mesne assignments, to Farnsworth Television & Radio Corporation, Dover, DeL, a corporation of Delaware Application January 5, 1938, Serial No. 183,420

21 Claims.

This application relates to electronic amplifiers, and particularly to amplifiers utilizing the principle of multiplication by secondary electron emission.

Amplifiers of the type described, which have become known as electron multipliers or multipactors, have proved to be extremely Valuable in the amplification of photoelectric currents, and have had some application for other uses where it is desired to multiply the entire electronic current flow. Their use in the amplification of weak signals has shown less promise, however, because of the fact that in cases of that character the multiplication is effective on the entire electronic stream, instead of affecting the signal component only. For purposes of signal amplification the attempt has usually been made to use a structure comparable to that in the ordinary thermionic tub-e, wherein a hot electron emitter, serving as a source of electrons, supplies an electronic flow which is controlled by a grid, and that portion of the flow which passes through the grid is multiplied by the secondary emission principle to as large a value as may be desired. Where the incoming signal is weak the percentage modulation of the electron flow is usually very small, and the current multiplied by secondary emission therefore comprises a very large direct-current component upon which is superimposed extremely small alternating orsignal components. Since both components are multiplied equally, there is a very large waste of power in multiplying the unwanted direct-current flow, which "drowns the signal. On the other hand, if the original flow is made very small, it is difficult to obtain adequate grid control. The mutual conductance of the electron gun or equivalent structure which serves as a primary source will, under these circumstances, be very low, and even after the current has been brought up by very large electron multiplications, the over-all mutual conductance or transconductance through the tube will not be greatly in excess of that of a tube of standard construction.

With these facts in mind, the objects of this invention include the provision of an amplifier wherein a relatively large amplified signal current may be obtained, which is not accompanied by a large direct-current component; the provision of an amplifier operating upon the electron multiplication principle, wherein extremely high mutual conductances or transconductances may be obtained; the provision of an electron multiplier amplifier wherein high amplification can be obtained Without extremely high intensity of electron bombardment of the secondary emitting surfaces or (alternatively) the use of large surface areas; and the provision of an apparatus of the character described which is adapted to such purposes as radio-, audio-, and video-frequency amplification in radio receivers for sound or television.

This invention possesses numerous other objects and features of advantage, some of which, together with the foregoing, will be set forth in the following description of specific apparatus embodying and utilizing the novel method. It is therefore to be understood that the method is applicable to other apparatus, and that it is not limited in any way to the apparatus of the present application, as various other apparatus embodiments may be adopted, utilizing the method, within the scope of the appended claims.

Referring to the drawings: 1 Fig. 1 is a view, partly in elevation and partly in section, of a tube embodying this invention, to-

gether with a magnetic coil for focusing or guiding the electronic flow.

Fig. 2 is a cross-sectional view' of the tube of Fig. 1, with a circuit connected thereto shown schematically.

Fig. 3 is an exploded view in modified perspective, showing the three major electrode structures of the tube of Fig. 1, as developed or laid out so as to show the interior of the cylindrical structure.

Fig. 4 is a curve showing the attenuation of the electron flow by various voltages applied to the control elements of the tube of Fig. 1.

Fig. 5 is a graph showing a single cycle of an exciting radio-frequency potential as applied to the electrodes of the tube of Fig. 1, and indicating the phases of such cycle wherein electron multiplication occurs.

Fig. 6 is a view similar to Fig. 3, showing a modified form of electrode structure as applied to the tube of Fig. 1, wherein electrostatic guidance of the electrons is used instead of electro-.

magnetic guidance or focusing.

Fig. 7 is an elevation of the anode structure used with the electrode structure shown in Fig. 6.

Fig. 8 is a view, partly in elevation and partly in section, of a modification of the invention wherein a saturated electron multiplier is used as an initial source of electrons.

Fig. 9 is a schematic view of a static type of electron multiplier embodying the invention.

Considered broadly, the method of this invention comprises subjecting an electron flow to alternate processes of multiplication by secondary emission and attenuation by electron absorption,

the attenuation being preferably (although not necessarily) of approximately the same degree as the multiplication, so that the order of magnitude of the output current does not greatly differ from the initial electron flow, although a very large multiplication may have intervened between input and output; and changing, by means of the signal to be amplified, the extent or degree of one of the two processes. The apparatus specifically described in this application is designed primarily for utilizing the dynamic or alternating-current principal of multiplication and includes means for successively multiplying and for attenuating the electron flow, the attenuation being variable in response to the signal to be amplified, means for guiding the electron flow so that substantially all of the electrons go through the same number of cycles of multiplication and attenuation, and means for applying the exciting oscillation. It will, however, be apparent that the principle or method involved is perfectly general, and is equally as applicable to static or direct-current electron multipliers as it is to multipliers of the alternating-current type.

Considering now in detail the preferred form of structure shown in Figs. 1, 2 and 3, the apparatus comprises an evacuated envelope I, of substantially standard type, which is provided with a central,rod-like anode 2, and a surrounding electrode 3, which may conveniently be a coatin of metal deposited on the inner wall of the tube. The tube is conveniently provided with a base 4 of conventional type, with the usual contact pins 5 which connect to the various electrodes within the envelope.

Surrounding the central anode are three electrodes I, 8 and 9, each of which subtends nearly 120 of arc, and which, when mounted in position, as shown in Figs. 1 and 2, form a cylindrical electrode structure. As previously indicated, these electrodes are mounted and insulated from each other as is well known in the art. Electrodes I and 8are structures whose walls are of imperforate sheet metal, and their interior surfaces are adapted to produce a high degree of secondary emission; e. g., they may be-of caesiated silver. The cylindrical wall of electrode 9, however, is provided with a plurality of apertures,

which are covered with a mesh or grid l0. Each.

of the three electrodes is provided with a plurality of vanes. II, which are so positioned that when the electrodes are assembled, the vanes cooperate to form a helical channel from top to bottom of the cylindrical structure, and sur-.

rounding but slightly spaced from the anode. When potentials are applied to the electrodes as hereinafter described, there is an electrostatic focusing action exercised by these vanes which tends to cause electrons to strike the cylindrical walls of the electron channel. Some electrons will probably strike the vanes, but this merely decreases slightly the effective multiplication per stage, and can readily be compensated by increasing the exciting potentials. The grids I 9 are At the bottom of the cylinder there is also The 5 mounted a disk-shaped mesh shield l2, connected to the anode 2, and beneath the shield is a collector anode 13 which forms the final output electrode. In operation the tube is surrounded by a coil 14, through which current is passed from a source I5 in the proper direction to provide the focusing or guiding field above referred to.

The connections for this tube are shown in Fig. 2. A high positive potential, from a source I1, is applied to the anode 2 through the lead l8, and a somewhat lower potential, conveniently from the same source, is applied to the wall coating 3 through the lead 19.

A suitable generator 20 is connected to supply a radio-frequency potential of approximately 50 volts (although it may, if desired, be increased to several hundred volts) and at a frequency of from forty megacycles to several hundred megacycles to the electrode 1. A radio-frequency choke coil 22 keeps the mean potential of the electrode 1 at ground level.

Electrode 8 is grounded. Electrode 9 is connected through a condenser 24 and inductance coil 25 to ground, these elements being seriestuned to resonance with the exciting frequency from the generator 20, so that electrode 9 is effectively at ground potential so far as this frequency is concerned but is isolated therefrom at other frequencies. It is provided with a negative unidirectional bias from a source 28 through a grid-leak or other suitable impedance 29, and it is to this electrode 9 that the signal to be amplifled is fed through the usual blocking condenser 39. It is to be understood that this method of supplying bias and signal to the control electrode 9 is merely shown as typical, as any of the connections familiar through thermionic tube practice may be used.

The output electrode [3 is connected through the usual output impedance 32 to a slightly high er positive than is the anode 2, and the output therefrom is taken off through the coupling condenser 33 or otherwise in accordance with wellknown thermionic practice.

Neglecting, at first, the radio-frequency field applied by the generator 29, it is obvious that an electron liberated from any of the cylindrically segmental electrodes 1, 8 or 9 will tend to be pulled to the central anode. Under the influence of the magnetic field, however, these electrons will be deflected into paths which are approximately epicycloidal, as shown by the dashed lines of Fig. 2, and their transit time along these paths will be primarily dependent upon the anode potential. When the radio-frequency voltage is supplied from the generator 20, electrode 1 is effectively connected to one terminal and electrodes 8 and 9 to the other. Since electrodes 8 and 9 are at the same radiofrequency potential and are effectively grounded at the generator frequency, the velocities and paths of the electrons traveling between electrode 8 and electrode 9 will not be modified by this potential, while the velocities and paths of electrons traveling between 1 and 8 and between 9 and I will be slightly modified by the applied radio-frequency voltage.

The potential of the anode 2 is so adjusted that the transit time of an electron from the surface of one of the three electrodes to the next is approximately one-third of a cycle of the exciting radio-frequency voltage, i. e., so that the descendants of an electron liberated from any one of the surfaces will make a complete circuit and return grids and hence the attenuation.

1 to. the surface of the initial electrode in one radiofrequency cycle.

Consider now an electron liberated from the surface of electrode 1, e. g., a photoelectron. Consider, also, that this electron is liberated at the instant when the electrode 1 is at zero potential, electrodes 8 and 9 also, of course, being at zero potential. Under the influence of the field from the anode, the electron will follow an epicycloidal path, and will tend to strike electrode 8. In the meantime, electrode 8 will have swung positive with respect to electrode 1, as shown in the graph of Fig. 5, in which the potential of electrode 8 with respect to electrode 1 is plotted against time, and the electron will impact upon electrode 8 with an energy corresponding to its initial velocity of emission plus the velocity acquired from the field during the positive swing of electrodes 8 and 9, as is shown by the portion of the graph of Fig, 5 bearing the legend Trip 78. The secondary electrons released by the impact on electrode 8 will then follow a very similar path between electrodes 8 and 9, but since these two electrodes are at the same radio-frequency potential, these secondary electrons will arrive at electrode 9 with only their initial velocity of emission, plus or minus such energy as they acquire from the combined action of the field due to the wall coating 3 of the tube and that due to the bias on the screen structure of the electrode 9, i. e., on the grid H). The device should be so adjusted that the major'portion of the electrons will be either drawn through.

the grid, striking on the electrode 3, or else will strike the wires of the grid, only those electrons having a minimum initial velocity of emission being so repelled by the grid of electrode 9 that they will be brought to rest near the surface of grid I without passing through it. A signal applied to the electrode 9 will vary the potential of the During this second transit period the potential of electrodes 8 and 9 will have swung from positive to negative, as is shown by the portion of Fig. indicated as Trip 8-9.

The few electrons that have been blocked before passing through the grid ID will be reaccelerated by the anode 2 and deflected by the magnetic field to strike electrode 1, but while they are traveling from 9 back to electrode 1, the electrodes will have swung back to zero potential.

Since the electrons left electrode 9 while it was highly negative, however, they will acquire sufiioient velocity to cause multiplication at electrode ll. It will therefore be seen that multiplication occurs at electrodes 1 and 8, but that an attenuation occurs when these electrons approach the grids In. This attenuation may be said to take place by electron absorption. When electrons approach the unapertured portions of electrode 9, however, they are completely screened from the field of electrode 3, and the negative bias is sufiicient to block and reverse the entire flow so that it is neither multiplied nor attenuated. The optimum adjustment of the device is such that the attenuation of the electron stream at the grids, I0 is approximately the reciprocal of the combined multiplications at electrodes 1 and 8.

Consider now an electron stream which is subjected to 11. cycles of alternate multiplication and attenuation, the coefficient of multiplication in each stage being K, and the coefficient of attenuation in each stage being A. It will be noted that A as here defined is a quantity of the same nature as K, being merely a, multiplication of find that the rate of change of In with varying attenuation is (11,, n nIK A 1 (2) Fig. 4 shows approximately the variation of attenuation coefficient with blocking voltage, assuming secondary emission from caesiated surfaces and primary electron velocities of the order required to give 3 to 4 secondary electrons per primary. Since such a curve isclependent upon both the material and the heat treatment of the surface, this curve cannot be relied upon except for the specific conditions under which the data for .it were taken, but it shows that in this case, at least, over 99 per cent of the electrons have initial velocities of less than a volt, and that the steepness of the curve increases as the attenuation coefficient becomes smaller. With such surfaces as are here considered these conditions are generally true, and since such curves go through practically their entire range ofattenuation between 0 and 1 with approximately a one-volt change of grid potential, we can take as numerically a first approximation of the mutual conductance of the tube.

This would indicate that the greater we make both A and the input current I, the greater will be the mutual conductance. This would lead us back to the simple multiplier, with unmanageable direct-current components, and what we are really concerned with is to provide a definite mean output current, and to combine the greatest mutual conductance with a high degree of modulation of this current. We may therefore adopt some suitable value for the mean, or unmodulated output current, and call it I, and by substituting in Equation 1, find the corresponding value of input current in terms of A, as follows:

K' A" and substituting this value in Equation 2,

dI 'nI This shows that the smaller we make'A, and the larger we make the input current, the greater will be the mutual conductance of a tube having a definite output capacitance.

It is highly desirable, however, to hold the currents within the tube Within reasonably narrow limits, i. e., not to permit the electron stream to vary by a factor greatly in excess of 100, purely for structural and operational reasons. Under this limitation we may make I not greater than 100 I (if we attenuate before any multiplication), or equal to I if we multiply by 100 before the first attenuation. Considering the latter condition first, and assuming a mean output level of 200 microamperes, K=100, and 12:4, then A=0.0-1 and din and since we see from Fig. 4 that the slope of the attenuation curve for values of A in the neighborhood of 0.01 is at least double the mean slope,

the mutual conductance of the tube when thus operated would be not less than 0.16 mho.

If we should so modify the tube as to make the first impact of the electrons an attenuating one, (thus adding an additional attenuating stage), and make 1:100 I, A would still be 0.01, but the value of d1. dA

would become 2O0 10' l00 (.0l) =0.l0

and the mutual conductance of the device would rise to'0.2 mho.

Similarly, it can be shown that four stages of attenuation with I=100 I will give a mutual contluctance of about 0.5 mho., while with the same number of stages, and I reduced to 0.01 I, the mutual conductance drops to 0.05 mho. This indicates that a change in input current of 10,000 fold only changes the mutual conductance by a factor of 10, where the output conditions are maintained constant.

The value of 100 here assumed for K is quite moderate, since there are four impacts between attenuations and it is relatively easy to obtain multiplications of from 3 to 4 per impact, which would make K from 81 to 256. Higher values of K will of course give greater mutual conductance, since A may be made higher.

The limitations on the device are that space charges must not be allowed to accumulate along the electron paths, which consideration limits the amount of multiplication between attenuations, and on the other hand, the currents must be sufficiently high so that shot effect does not become important. This limits the degree of attenuation which may be used. As a working rule, the output current should be kept within one or two orders of magnitude of the input current for best results.

There is also a signal frequency limitation on the device, as it starts to become ineffective when the period of the signal approaches the transit time of the stream through the entire structure. Leakage of electrons between impacting stages also makes the tube sluggish and limits amplification. It will be seen, however, that with currents as small as 200 microamperes, unity gain can be obtained into output impedances of the order of ohms, while it is possible to work the tube into impedances of the same order as are used in ordinary thermionic practice, so that if proper precautions are used in shielding the apparatus, extremely high gains may be obtained. A 20.000-ohm output, for example, will yield a voltage gain of 3,000 or more.

Figs. 6 and 7 show an electrode structure wherein the guidance of the electrons down their helical path is electrostatic instead of electromagnetic. In this case the anode 2' is provided with equidistant vanes 40 of mesh material, while each of the electrodes 1', 8 and 9' is also provided with a continuous mesh shield 42, which, when the electrodes are in place, parallels the mesh vanes 40 on the input side of the segmental electrodes. Such an arrangement of grids, connected otherwise in the same manner as the tube shown in Figs. 1 and 2, will serve to guide the electrons down through the helical path in the same manner as will the magnetic coil.

In these figures the attenuating stages have been shown as differently positioned from those in the form first shown, since it will be seen from the nature of the equations above given that the amplification obtainable is dependent only on the total attenuation, and not on the position of the attenuation in the electron path. The latter, as has already been suggested, is determined only by the requirement that the currents be kept within convenient limits of magnitude.

From the foregoing it will be seen that the highest mutual conductances will be obtainable with high input current and small coefiicient of attenuation. It will also be apparent that the use of photoelectrons to initiate the electron flow is very far from being the most desirable method of operation, although this was the source used in the first practical test of the device.

It would obviously be possible to use a thermo-emissive filament, or an electron gun, as an initial source of flow. It is preferable, however, to use an electron multiplier as the initial source, as is shown in Fig. 8.

In this case, a chamber is mounted at the top of the multiplier helix, comprising two semicylindrical plates 5| and 52, these plates being connected to electrodes 1 and 8', respectively. The cylinder formed by the two plates is slightly larger than that formed by electrodes 1', 8' and 9, so that the transit time of an electron across them will approximate a half cycle of the exciting frequency instead of one-third of a cycle. The anode 2' extends upward axially of this auxiliary chamber, and the chamber itself forms a multiplier of the type disclosed in the copending application of Philo T. Farnsworth, Serial No. 10,604, now United States Patent No. 2,143,- 262 issued January 10, 1939. The upper stage of the multiplier helix projects slightly into the chamber.

It has been found that if a multiplier of the semi-cylindrical plate type be excited by the proper frequency, multiplication will occur until a space charge equilibrium is reached. If the parts are arranged as here shown, there will be a constant spill-over of this space charge to the first stage of the helical multiplier, and this spillover constitutes the input current of the device.

Experience has shown that there is always a suflicient supply of electrons to start the multiplying process, even with materials that are only slightly electron-emissive, and in the dark. Hence it is not necessary to supply any filamental or other source of electrons where this type of device is used. It is also possible, by connecting a semi-cylindrical plate as shown in my United States Patent No. 2,071,516, to use this upper multiplying compartment as the driving oscillator for the device, in which case the external source 20 may be omitted.

The principles here described can also be applied to electron multipliers of the static type, as has already been suggested. It is preferable to use the dynamic type of multiplier, since .the circuit connections are simpler, the potentials which need to be supplied are lower, and the devices thus far constructed are more eflicient. In Fig. 9, however, is shown schematically a statictype multiplier which may be used for this purpose.

In this embodiment of the invention, the evacuated envelope 60 encloses a filament 62 which serves as the original source of current, and which is shielded so that the current flow is afiected only by the screen 63, which is electrically connected to the first multiplying electrode 64. This multiplying electrode is in the 'trode 16.

form of a rectangular box, having its input side protected by the screen 63 and having an open output side. Such multiplying electrodes are described in the copending application of B. C. Gardner, Serial No. 159,492, now United States Patent No. 2,200,166 issued May 7, 1940. The first, and all of the succeeding electrodes to be zero, but slightly on the positive side, there will be some slight emission of secondary electrons with a secondary-to-primary electron ratio greatly less than unity and which varies directly with the positive potential on the electrode. It will be seen that a solid electrode so biased would 1 meet all of the requirements for an attenuating hereinafter described are shown as obtainingof screen or grid formation. Electiiode 69 is biased as to be slightly negative to screen 61, while electrode 68 is positive thereto, so that the emitted electrons may themselves cause the emission of secondary electrons.

Following electrode 68 is a series of three simple multiplying electrodes 10, I2 and 13, each of these electrodes being similar in character to electrode 64, and each being arranged to receive the secondary electrons emitted from the preceding electrode, and being maintained at a potential more positive than the preceding one to accomplish this.

Following electrode 13 is a second attenuating stage, comprising an accelerating electrode 14 and attenuating screen or grid 15, after which the electrons enter the next multiplying elec- The potentials of the screens 14 and 15 bear the same relation to the electrode 13 as the potentials of the screens 61 and 69 do to the electrode 64.

Following the attenuating stage last described, the figure is interrupted to indicate that the multiplying and attenuating stages may be repeated in as long an amplifying sequence as may be desired. Following the final multiplying stage, the output stage of the tube comprises a final shield electrode 18 followed by an output anode 19, these being held at potentials increasingly positive to the final multiplying electrode 80.

Aside from the connections to the voltage divider 65, as already described, the circuits for this tube comprise a filament circuit, shown as supplied by a battery 82, an output circuit comprising an output impedance 83 and the usual output blocking condenser 84, and an input circuit which is connected to the attenuating electrodes 69, 15, etc. This latter circuit comprises choke coils 85 and 81, through which the bias is supplied to the attenuating electrodes, and coupling condensers 88 and 89, through which the input signal received through the blocking condenser 90 is applied to the attenuating grids. The action of these grids is similar to the action of the grids 10 in the dynamic type of the device already described, and the same general considerations apply throughout with this device as apply to the dynamic type.

It will be understood that the structures shown in this application are merely illustrative of many in which the principles herein set forth may be applied. The method of attenuation here described in detail is one which is desirable because of its extreme sensitivity, but various others will at once suggest themselves to those skilled in the art. If, for example, the apertures are omitted entirely from the attenuating electrode 9, and this electrode be biasednearly to electrode, although the control would not be as fine as that previously described.

Furthermore, it is not necessary that the control signal be applied to the attenuating electrode, since it will be seen that the quantities K and A enter into the equations for mutual conductance in precisely the same manner, and hence if the attenuation be held constant and the multiplication be changed in response to the signal, exactly the same results will obtain. It is, however, more difficult to obtain reasonably large changes in multiplication than it is in attenuation, and hence in the practical device the attenuation is the quantity which will usually be chosen as that to be varied by the signal.

Other rather obvious modifications are to omit the grids ill in electrode 9, and control the attenuation merely by the potential on the edges of the apertures. In this case, of course, the positive bias on electrode 3 can be greatly reduced, as when the grids are used the bias must be quite high in order to draw the desired number of electrons through, especially if the grids are operated sufficiently far negative toprevent grid current. important itwould also be possible to apply the control potential to the electrode 3, but in this case the latter should be placed in closer proximity to the apertures in electrode 9. In this case also the grids should be omitted.

One other modification should be considered. If a single attenuating stage be used, immediately following the initial electron source, and the attenuation in this stage be made very large, i. e., the attenuation coeflicient be made even smaller than 0.01, great sensitivity may be combined with high mutual conductance and small direct-current component by mere multiplication in a few succeeding stages.

Thus, by making A=0.00l, and keeping I at'200 microamperes, as in the former examples, we find by substituting in Equation 4:

. L0 dA and since the steepness of the attenuation curve increases as A approaches zero, this would give a mutual conductance of at least 0.5, and probably considerably more. The only difiiculty with such a device is that the current through the tube varies so widely in difierent portions of the electron channel that it becomes difficult to reconcile the two desiderata of no space charge effect and minimum shot effect, but for special purposes this could be taken care of by careful design. Such a tube has the decided advantage that its frequency limitations are not as severe as in the case of multiple attenuations, as all of the control is exercised at a single space phase of the electron transit, and subsequent multiplication does not affect the degree of control.

The logical development of this idea is that the multiplication itself is only a convenient and not a necessary step in the process, as it will be seen that the multiplication constant K does not enter into Equation 4. Hence we might make the last impact the only attenuating one in a tube of the general type shown in Fig. 1, and the Were the latter consideration not remainder of the, stages would then becomea mere source ofelectrons, and the. equivalent of a,

thermionic .cathode.

The tube as a whole would then bearastrong resemblance to the ordinary thermionic tube, comprising an electron source, or cathode, a control element, and a co1lector, the, shield I2 not being strictly necessary any more than .it is in the ordinary thermionic type.

ts action, however, is quite different from the ordinary device. In the latter, certain electrons are retarded by the grid and subtracted from the main stream, which is used to carry the signal. In; my device the main stream is discarded, and it is only the retarded electrons which are, used to carrythe signal.

Under, ordinary circumstances there would be no advantage in using the device last described, sincethe direct-current component is usually eliminated by a transformer or resistance-capacitance coupling between tubes, It might, however, be useful in certain forms .of direct-current amplifier, and it is. described here both for the sake of completeness, and also to bring out clearlythe difference between this and prior art devices.

Iclaim:

1. The method of signal amplification which comprises the steps of causing the initiation of aspace flow of electrons, causing said flow to be subjected to alternate processes of attenuation by electron absorption and multiplication by secondary emission, and applying the signal to vary one of said processes.

2. The method of signal amplification which comprises the steps of causing the initiation of a spaced fiow of electrons, causing said flow to be subjected to alternate processes of attenuation by electron absorption and multiplication by secondary emission, and applying the signal to vary the degree of attenuation.

3. The method of signal amplification which comprises the steps of causing the initiation of aspace fiow of electrons, causing said-flow to be subjected to successive cycles of alternate processes of electron multiplication and attenuation by electron absorption, and applying the signal to vary the degree to which one of said processes is effective.

4. The method of signal amplification which comprises the steps of causing the initiation of a space flow of electrons, causing said fiow to be subjected to successive cycles of alternate processes of'electron multiplication and attenuation by electron absorption, and applying the signal to vary the degree to which one of said processes is effective.

5. An amplifier comprising means for initiating an electron flow, means for subjecting said flow to successive stages of electron multiplication, means for subjecting said fiow to successive stages of attenuation by electron absorption, said stages of multiplication and attenuation being interspersed, and means for controlling the effectiveness of stages of like character in accordance with a signal.

6. An amplifier comprising means for initiating an electron flow, means for subjecting said flow to successive stages of electron multiplication, means for subjecting said flow .to successive stages of attenuation by electron absorption, said stages of multiplication and attenuation being interspersed, and means for con-trolling the degree of attenuation in accordance with a signal.

n m l fie r s. mean o n n a ing an, electronfiow, means for subjecting said fiowto successive stages of electron multiplication, means interspersed with the stages ofsaid multiplication for attenuating said fiow byelectron absorption to. a mean value difiering from thevalue of theinitial fiow by not more than two. orders of magnitude, and meansresponsive to a signal to be amplified for varying thev effectivenessof like stages.

8. Anamplifier comprising means, for initiating an electron flow, means for subjecting said now. to, successive, stages of electron multiplica- .tion, means interspersed with the stages of said multiplication for attenuating said now, by electronabsorption to a mean value differing from the value. of the initial flow by not more than two orders. of. magnitude, and means responsive to a signal to be amplified for varying the degree of attenuation produced in. each of the attenuating stages.

9. An electronic amplifier comprising an evacuated envelope, a source of electrons withinsaid envelope, a plurality of electrodes within said envelopeat least one of whicch has a surface capable of. emittingssecondary electrons. at a ra-v tip to impacting electrons materially greater than unity and anotherof which has a surface capable of absorbingthe major portion of an electron streamudirected against it, means, fordirecting an electron stream comprising electrons emitted from said. source and descendants thereof generated by secondary emission successively against saidplurality of electrodes, meansfor impressing a signal voltage. onsaid absorbing electrode to vary .itsdegree of absorption, and ,means for collecting the electrons comprisingsaid stream after attenuation thereof. by said absorbing electrode to provide an amplified output signal current.

10.. An. electronic amplifier comprising an evacuated envelope, asource of electrons within said envelope, a plurality of electrodes within said envelope. at least one of which has a surface capable of emitting secondary electrons. at a ratiouto impacting electrons materially greater than. unity and. another of which has a surface capable of absorbing the major portion of an electron stream directed against it, means for directing .an electron stream comprising electrons emitted from said-source and descendants thereof generated by secondary emission successively againstsaidplurality of electrodes a predetermined number of times, means for impressing a signalvoltage on said absorbing electrode to vary its degreeof absorption, and means for collecting the electrons comprising said stream after attenuation thereof by said absorbing electrode.

11. In an amplifying device operated by secondary electron emission, means for initiating an electron space current, means for augmenting said current by secondary electron emission, means for-attenuating the previously augmented current to approximately its original order of magnitude by electron absorption, and means responsive to an applied signal for varying the degree of absorption.

12. An electron multiplieramplifier comprising which descendant electrons from saidfiow. are. driven to causean increased fiow, a plurality of stages interspersed between'cer-tain of said multiplier stages, each comprising meansfor attenuating said flow by electron absorption to approximately its original order of magnitude, and means operative in response to a signal on a plurality of the stages of like function for varying the degree to which it modifies the flow.

13. An electronic ampifier comprising an evacuated envelope, an electrode assembly within said envelope comprising a plurality of cylindrically segmental electrodes, said electrodes being provided with inwardly projecting vanes cooperat ing to form a helical electron channel within said assembly, said electrodes including at least one having a surface capable of emitting secondary electrons at a ratio to impacting electrons materially greater than unity and at least one other having a surface capable of absorbing electron flow directed thereagainst, a central anode extending substantially axially through said structure, means for applying a positive potential to said anode to accelerate electrons from adjacent the surfaces of any of the electrodes of said structure, means for guiding such electrons to cause a majority thereof to miss said anode and impact an adjacent electrode of said structure, means for applying a radio-frequency potential between the previously identified electrodes of said structure, said potential being timed with the transit time of electrons between said electrodes as determined by the anode potential and said guiding means so that electrons emitted from said secondary-emitting surface and descendants thereof will arrive at said absorbing surface with very small energy derived from said radio-frequency potential while electrons unabsorbed by said absorbing surface will arrive at said secondary-emitting surface with sufficient energy to cause material secondary emission therefrom, means for applying a signal potential to the absorbing electrode to change its degree of absorption, and means for collecting elec-- trons discharged at the end of said helical channel.

14. An electronic amplifier as described in claim 13, having a work circuit connected to the collecting means.-

15. An electronic amplifier comprising a plurality of cylindrically segmental electrodes each having inwardly projecting vanes therein cooperating to form a helical path within the cylinder formed by said electrodes, means for guiding electrons released within said helical path along said path and against an adjacent one of said electrodes, means on certain of said electrodes for augmenting electron flow in said path by secondary electron emission, apertured portions on another of said electrodes for attenuating electron flow in said path by diversion of electrons therefrom, means for applying a signal potential to said last-identified electrode to vary the proportion of the electron fiow diverted thereby, means for collecting electrons emerging from the end of said path, and a work circuit connected to said collecting means.

16. An electronic amplifier comprising three electrodes cooperating to form a longitudinally divided cylindrical structure, a high-frequency oscillator, connections between said oscillator and said electrodes to vary the potential of one of said electrodes with respect to that of the other two, means for applying a signal potential to one of said other two electrodes, means for guiding an electron from adjacent the surface of any one of said electrodes to the surface of an adjacent one thereof in a time substantially equal to one-third the period of said oscillator, means for limiting the number of impacts made by the descendants of an initial electron upon said electrodes, and means for collecting such descendant electrons after said limited number of impacts.

17. An electron tube comprising a plurality of electron multiplying electrodes, a plurality of electron absorbing electrodes, an anode structure positioned to guide electrons from a first of said multiplying electrodes to a first of said absorbing electrodes and from a second of said attenuating electrodes to a second of said multiplyin electrodes, connections for supplying a signal potential to said absorbing electrodes, an initial sourceof electrons, and means for collecting electrons descended from said source after repeated multiplications and attenuations.

18. An electron tube comprising three electrodes forming a trichotomous cylinder, two of said electrodes having electron multiplying surfaces and each of said electrodes having inwardly projecting vanes thereon cooperating to form a helical channel, a central anode rod mounted axially of said cylinder, a plurality of apertured portions formed in the third of said trichotomic' electrodes, a collector electrode mounted to receive electrons passing through said apertured portions, a source of electrons adjacent one end of said helical channel, and an output electrode mounted adjacent the other end of said channel.

19. An electron tube comprising three electrodes forming a trichotomous cylinder, two of said electrodes having electron multiplying surfaces and each of said electrodes having inwardly projecting vanes thereon cooperating to form a helical channel, a central anode rod mounted axially of said cylinder, screen vanes connected to said anode rod and projecting radially between said electrodes, a guide screen connected to each of said trichotomic electrodes and extending parallel to said anode vanes, a plurality of apertured portions formed in the third of said trichotomic electrodes, a collector electrode mounted to receive electrons passing through said apertured portions, a source of electrons adjacent one end of said helical channel, and an output electrode mounted adjacent the other end of said channel.

20. An amplifier comprising three cylindrically arranged electrodes, at least two of which have surfaces capable of electron multiplication by secondary emission, a ground connection for one of said multiplying electrodes, a high-frequency generator connected to apply a potential to the other of said multiplying electrodes, and circuit connections for maintaining the third of said electrodes at ground potential with respect to the generator frequency and for applying a signal potential above ground thereto.

21. An-amplifier comprising means for initiating a first stream of electrons, means responsive to said first electron stream for producing a second stream of electrons by secondary electron emission, means for attenuating by electron absorption the magnitude of said second electron stream in accordance with the magnitude of a signal current supplied to said attenuating device, and means for collecting the unabsorbed component of said second electron stream to provide an amplified output signal current.

PI'HLO T. FARNSWORTH.

Images