US2661426A

US2661426A - Magnetron and circuit therefor

Info

Publication number: US2661426A
Application number: US690135A
Authority: US
Inventors: Clarence W Hansell
Original assignee: RCA Corp
Current assignee: RCA Corp
Priority date: 1942-12-31
Filing date: 1946-08-13
Publication date: 1953-12-01
Anticipated expiration: 1970-12-01

Description

Dec. 1, 1953 c. w. HANSELL MAGNETRON AND CIRCUIT THEREFOR 4 Sheets-Sheet 1 Original Filed Dec. 51, 1942 HIGH POT.

POWER SOURCE RT com .INVENTOR we MEANS 'gfiggm CLARENCE w. HANSELL THE BY A ORNEY OF SHO INCLUD! DULATING 7 SOURCE PU ES FOR M0 FREQUENCY OR TIMING OF THE PULSES Dec. 1, 1953 Original Filed Dec. 31, 1942 HIGH D. C. POWER SOURCE C. W. HANSELL MAGNETRON AND CIRCUIT THEREFOR HIGH POT.

D.C. POWER SOURCE MIT HIGH POT. D.C.

SOURCE TO MAGNETRON 4 Sheets-Sheet 2 SOURCE CONTROL PULSES a0 A I as RF. OUTPUT IN PULSES R sou CE of 11 CONTROL PULSES TO FIL.

HEATING SOURCE FHHW INVENTOR CLARENCE w. HANSELL A RNEY Dec., 1; 1953 A c. w. HANSELL 2,651,426

MAGNETRON AND CIRCUIT THEREFOR Original Filed Dec. 31, 1942 4 Sheets-Sheet 5 RF. OUTPUT T1:' 5 IN PULSES r as |J 12 N 7% 5 HIGH POT. D.c. POWER 1 SOURCE A y A of .coNTRo1. PULI'SES 74 souRcE L TO FIL.

HEATING SOURCE SOURCE SOURCE 34 of -'of CONTROL HIGH PULSES D. c. MODULAT ED INVENTOR CLARENCE W. HANSELL BY u g A RNEY Dec. 1, 1953 c. w. HANSELL 2,661,426

MAGNETRON AND CIRCUIT THEREFOR Original Filed Dec. 51, 1942 4 Sheets-Sheet 4 'I 23 INVENTOR CLARENCE w. HANSELL BY I A ORNEY Patented Dec. 1, 1953 MAGNETRON AND CIRCUIT THEREFOR Clarence W. Hansell, Port Jefferson, N. Y., assignor to Radio Corporation of America, a corporation of Delaware Original application December 31-, 1942, Serial No. 470,768. Divided and this application August 13, 1946, Serial No. 690,135

11 Claims.

The present invention relates to improvements in magnetron oscillation generators and their associated circuits and is a division of my pending application Ser. No. 470,768, filed December 31., 1942, now U. S. Patent 2,409,638, granted October 8, 1946. More particularly, the invention is concerned with an ultra high frequency magnetron to be used for the production of pulses of oscillation, as distinguished fromthe production of oscillations resulting in continuous waves.

In my experiments, I have observed that. magnetrons with small thermionic emission, as low as say fifty milliamperes, will oscillate and can be made to pass currents during oscillation ranging up to, say, fifty amperes, provided the magnetic field and the anode-to-cathode potential are sufficiently high. It is my theory that oscillation produces a rapid growth in cathode emission due to secondary emission from bombardment of the cathode by out-of-phase electrons. By out-ofphase electrons I mean those electrons which absorb energy from the high frequency electric fields set up the magnetron during oscillations as distinguished from those which give up energy to maintain the oscillation. These out-of-phase electrons after one transit out toward the anode and back, strike the cathode with enough energy to produce secondary emission, while in-phase electrons may make 1.5 or more excursions before striking the anode and deliver more power to produce oscillations than the power absorbed by out-of-=phase electrons. Absorption of the outof-phase electrons on the cathode after one excursion is an important factor in electron grouping to enhance the strength of oscillations.

I have also observed that there is a time lag between the application of the high anode-tocathode potential and the beginning of oscillaticn, and that i can control this time lag by controlling the amount of thermionic emission. According to my theory, there is an accumulation of space charge and circulating electron current in the space between the anode and cathode, following the application of the anode to-cathode potential, the initial rate of this accumula tion. is proportional to the electron emission from the cathode. When the space charge and circulating electron current reach a critical value, there occurs a condition of high frequency negative resistance greater than positive resistance in the oscillatory electric circuit of the anode. Oscillations then start just as they do in any negative resistance oscillator. If the anode-to-cathode potential is too low, or if the cathode is a very poor secondary emitter, the oscillations continue only long enough to dispel most of the circulating space charge by causing some electrons to be accelerated and forced back to the cathode while others are slowed down and caused to strike the anode. On the other hand, if the anode potential is high enough, and the cathode is a good secondary emitter, the circulating space charge will be replenished by secondary emission as fast as or faster than it is dissipated and oscillation will continue at great strength so long as the anode potential remains high.

According to a feature of the invention, the magnetron, when used with suitable circuits, is made to stop its own oscillations as the potential across the magnetron is lowered by dissipation of energy stored in the input circuit connections. To achieve this result, the magnetron is shunted by a condenser or an artificial line which, in turn, is charged through a suitable impedance, as a result of which oscillations of the magnetron, once started, continue until the condenser or line circuit is discharged down to a potential too low to maintain sufiicient secondary cathode emission. The condenser or line circuit Will then be charged through the series impedance and space charge will accumulate in the space between anode and cathode, and oscillations will start again. This process will repeat itself indefinitely. Thus, so long as the cathode emission is low enough to lengthen the time for re-establishment of circulating space charge to more than the time constant of the charging circuit, the magnetron will pulse itself automatically. Putting it in other words, in order to achieve self-pulsing, the anode potential should be restored to its maximum value before the space charge has been re-established sufficiently to initiate oscillations. The resulting pulses will be of nearly equal energy, and their frequency or repetition rate may be con trolled by controlling the amount of thermionic emission.

Inasmuch as it is not very practical to control and modulate the thermionic emission, it is proposed in accordance with the invention to utilize a controllable electron source for establishing the initial circulating space charge needed to start each pulse.

This feature is quite important when using the magnetron of the invention in pulse communication or in pulse echo systems of the radio locating type, sometimes referred to as obstacle detection systems. Heretofore, in radio locating pulse systems, it has been customary to pulse the magnetron by delivering the whole input power to the magnetron in pulses, which required pulses of potential to be applied between the anode and cathode. Such known systems require a switching arrangement in series with the magnetron which can handle large amounts of power. In the prior art, one type of switching arrangement has required modulator vacuum tubes in series with the magnetrons and a high potential power source. In operation, the control electrodes of the modulator vacuum tubes are pulsed to permit pulses of current to flow through the magnetrons. In this type of modulator, difficulties have arisen because of secondary emission from the control electrodes of the modulator tubes which makes it difficult to interrupt the current at the ends of the pulses. Another type of modulator requires storing energy at high potential and-then discharging the stored energy through a spark gap into the magnetron. Reference is made to U. S. Patents 2,411,140 and 2,500,552, of Nils El Lindenblad, for detailed descriptions and other known systems. By means of my invention, however, I can use a relatively small amount of control energy to make the magnetron as its own modulator, or control switch, and I do this by using a pilot source of control pulses of small pulse energy.

According to another feature of the invention involving the construction of the magnetron, the controllable electron source is, in effect, a priming current which I can turn on and off. A cold cathode as distinguished from a thermionic or heated cathode capable of emitting copious electrons upon bombardment by the priming current is placed physically near the controllable electron source. By modulating the priming current, and turning it on and off, I can control the frequency of the magnetron pulses and start and stop them at will. The length and energy of the pulses is controllable by varying the amount of the dielectric capacity between the anode and cathode of the magnetron, by varying impedances in series with the magnetrons, and to some extent by controlling the potential.

According to a preferred detail feature of the invention, the magnetic field may be tapered somewhat in strength along the axis of the magnetron, with a minimum at the center, so that the circulating space charge tends more nearly to accumulate where it is most needed and not to diifuse out to the end walls. A suitable electric field distribution for aiding in this grouping of the space charge is also desirable.

Among the objects of the invention are: To simplify the design and construction, and reduce the weight, bulk and cost of radio systems employing magnetrons for producing pulses of high frequency energy; to provide a magnetron which requires a relatively small amount of control energy to start oscillations; to provide a novel form of magnetron construction which includes a cold cathode capable of producing copious secondary electrons upon bombardment of primary electrons, and a source of priming current; and to provide a magnetron capable of producing pulses whose frequency or repetition rate can be varied by controlling an electron source which establishes the initial circulating space char e needed to start each pulse.

The following is a detailed description of the invention in conjunction with the drawings, wherein:

Fig. 1 shows, in cross-section, a view of a magnetron oscillator in accordance with one embodiment of the present invention;

Fig. 2 shows the magnetron of Fig. 1 and a circuit associated therewith for causing the production of radio frequency energy in pulses;

Figs. 3, 4 and 5 show other constructional features of magnetrons in accordance with the present invention, together with different kinds of circuits for producing pulses of radio frequency energy;

Fig. la shows an alternative type of artificial line which can be used for that shown in the system of Fig. 4;

Fig. 6 shows in cross-section the essential novel features of construction of a magnetron in accordance with another embodiment of the invention;

Fig. 6a is a sectional view of the magnetron of Fig. 6 along the lines 6a6a; and

Fig. 7 schematically illustrates the magnetron of Fig. 6, together with a circuit arrangement for operating the same.

Throughout the figures of the drawing, the same reference numerals designate the same or like parts.

The oscillation generator shown in Fig. 1 comprises an envelope 1, made of any suitable material such as copper, containing within it a hollow cylindrical non-thermionic or cold cathode 2, a cylindrical control electrode 3 located along the axis of the magnetron, a pair of thermionic cathodes t located between the control electrode 3 and the cold cathode 2, and a cylindrical anode structure 5 having an even number of protruding anode portions 6 which are substantially or effectively spaced from one another by one-half wavelength and which bend inwardly toward the cathode, more or less in the manner shown in Fig. 1. This type of anode, which is a preferred type, though not essential in the practice of the invention, is of the type generally shown and described in my United States Patent 2,217,745, granted October 15, 1940. A field coil 1, which may or may not employ iron to aid its effect surrounds the envelope and functions to produce an intense but constant magnetic field which has flux lines running through the envelope in a direction more or less parallel to the axis of the cold cathode so as to influence the movement of the electrons emanating therefrom.

The envelope I is evacuated in the manner of any vacuum tube. The cold cathode 2 is apertured at diametrically opposite points 8 to permit electrons emanating from the hot cathodes d to enter the space between the cold cathode 2 and the anode 6 under conditions described hereafter. The cold cathode 2 is made from a metallic material and is of the type whose exterior surface is capable of emitting copious electrons when bombarded by primary electrons emanating from the hot cathodes. The cold cathode is preferably made of some light metal, such as aluminum or an alloy of light metals which emit secondary electrons easily. Alternatively, the cold cathode may be of metal, coated with oxides of earth metals such as barium and strontium oxides formed by the reduction of the carbonates in vacuum, which oxide coatings I have found to be sufiiciently good secondary emitters for the purpose of the invention. The apertures 5 in the cold cathode 2 may be in the form of slots extending nearly the entire length of the cathode, in which case the control elec trode 3 will extend the entire length of the cathode. It is only necessary for the control electrode 3 to have substantially the same or somewhat greater length than the length of the apertures 8.

The hot cathodes or filaments 4, 4 and the control electrode 3 are so positioned that when the control electrode has a positive potential with respect to the filaments ll, l, electrons will leave the hot filaments and circulate out into the magnetic field, passing through, the slots 8 in the cold cathode. The positive potential on the control electrode 3 will draw electrons from the filaments toward it, but the magnetic field will bend the electron paths and cause them to pass out through the slots and into the space between the cold cathode and the anode; in the manner shown by the dotted lines, having arrowsthereon to indicate the direction of electron motion. When the control electrode 3 .is at zero or negative potential relative to the filaments, substantially no electrons will pass through the slots in the cold cathode. In the operation of the magnetron of Fig. 1, it (while the control electrode 3 is zero or negative relative to the fila-v ments t) a large potential from a suitable source is applied between the. anode B and the cold cathode 2, substantially no current will flow due to the absence of emission from the cold cathode 2. No current will flow through the slots to bombard the cold cathode due to the fact that there is nearly complete shielding of the fila-- ments 5 from the electric, field produced by the anode to-cold cathode potential, under the above condition. In this situation, there will be substantially no accumulation of circulating electron space charge between the cold cathode and the anode. However, if the control electrode 3 is made to be positive relative to the filaments l, electrons will pass through the cold cathode slots 3, and a circulating electron space charge will begin to accumulate in the space between the anode and the cold cathode. As this circulating space charge grows, and if the anode-tocathode potential is high enough, though of a direct current character, a point will be reached where oscillations start. The cold cathode 2 will then be bombarded by out-of-phase electrons, and emission will grow rapidly due to secondary emission; as a result of which a large anode-to-cathode current will flow. This flow of current may discharge a condenser or line circuit connected betweenthe anode and cathode in the manner described more in detail later, and reduce the anode-to-cathode potential, thus stopping the oscillations again.

If the. control electrode 3 is kept positive relative to the filaments 4 and a dielectric capacity between anodes iiand cold cathode 2 is charged through an impedance, the magnetron will act in a manner quite similar to a Thyratron gaseous discharge tube, operating in pulses and at a rate which may be determined by the time constant of the power supply circuit or by the rate of growth of circulating space charge, depending upon which is quicker. Alternatively, the control electrode may be rendered negative when the oscillation and anode current start, and not made positive until a pulse of anode current and oscillation is desired. Thus, by means of a relatively small potential change and very little energy applied to the control electrode, I can control the timing and rate of pulses of radio frequency energy obtainable from the magnetron and stop and start them as desired, using the magnetron itself as a modulator as well as an oscillator. This is also described later in connection with the circuits of Figs. 2 to 4, inclusive.

Fig. 2 schematically shows a complete circuit arrangement utilizing the magnetron of Fig. 1 for producing pulses of radio frequency energy. The hot cathodes or filaments 4 are supplied with filament heating current over leads 9, which connect to opposite terminals of the secondary winding of an ordinary sixty cycle power transformer Hi. In shunt to or across the cold cathode 2 and the anode 5, there is provided a charging circuit including a storage condenser 13, a charging reactor or choke coil! 4 and a high potential direct current power source l2. A relatively small cushioning choke II is provided between the charging reactor 14 and the cold cathode. Source 12 supplies a substantially constant current through the choke coil I t to the storage condenser l3, and this condenser 13 is discharged through the magnetron solely during pulses of oscillation, at which time the cold cathode 2 has high sec ondary emission. Choke coil IT aids in starting oscillation and in obtaining flat top pulses in the output of the magnetron by virtue of the potential drop in the choke coil, due to rate of change of current therethrough. In order to obtain output pulses of high frequency energy from the oscillator, there is provided a loop l5, one end of which is directly connected to the anode (as shown) and the other end of which extends outwardly through a concentric line [6 for utilization by suitable apparatus, such as an antenna.

In order to control the initiation of circulatin space charge and oscillation (that is to start the pulses) there is provided a source H of control pulses which is coupled (through transformer It) between the control electrode 3 and the hot cathodes i, as shown. Source H is, in effect, a keyer of very short pulses of moderate power and potential which are applied to the control electrode 3 and are sufficient to cause electron current from the hot cathodes d to flow through the slots in the cold cathode to initiate a growth of circulating space charge and oscillations. Source I I (which is a pilot of relatively small pulse energy) 1 starts the pulse While the magnetron with its circuit l2, l3, I l makes and breaks the circuit to start and stop the main power pulse. In other words, I employ a small amount of control energy from H to make the magnetron act as its own switch.

In the operation of the system of Fig. 2, the

pulse voltage from source I will prime the magnetron to cause it to begin oscillations. This occurs because the electrons from the hot cathodes or filaments l flowing through the slots of the cold cathode 2 will produce a growing circulating space charge inside the anode structure until oscillations start. These oscillations start by virtue of the negative resistance. Once the oscillations start, the electrons in the space charge are replenished as fast as they are used up, by virtue of they bombardment of the exterior surface or" the cold cathode by out-of-phase electrons and the consequent production of second ary emission. These oscillations continue until condenser I3 is discharged below a critical poten tial, at which time the bombardment of the cold cathode by the out-of-phase electrons does not take place with sufiicient energy to replenish the circulating space charge as fast as it is used up. The oscillations then stop suddenly. The time between the initiation of oscillations and the cessation of oscillations constitutes the duration of one pulse of radio frequency current as taken out from loop 55. After the cessation of oscillations,

accrues condenser [3 will be recharged from source 12 through choke coil id. Source [2 has a magnitude of voltage necessary to cause the magnetron to oscillate efficiently by virtue of the cold cathode emission phenomenon. Source ll initiates each pulse and thus determines when the oscillations start, and controls the time of the initiation of the pulse and the rate of the pulses. In practice, the source H may generate pulses of 1000 volts (by way of example) of extremely small current, at a rate of a few cycles per second up to 20,000 or 30,000 cycles per second, depending upon the type of detection or communication system associated with the circuit of Fig. 2.

If the system of Fig. 2 is designed for use in a pulse echo system (sometimes known as an obstacle detection system of the type employed now for military purposes), source H can generate pulses anywhere in the range from 120 to 4000 or 5000 cycles per second, whereas if the system of Fig. 2 is to be used for telephone communication, the pulse rate of this source might be 20,000 or 30,000 cycles per second, in which case this pulse rate or the pulse timing might be frequency modulated by voice currents. Where the system of Fig. 2 is used for telephone communication purposes, the pulse rate of source Ii must be higher than the audio frequency range utilized for communication. Source H can be a small synchronous motor driving a commutator doing the pulse or it may be a vacuum tube pulser which can be modulated in frequency or timing. As for high potential direct current power source l2, considering present types of magnetrons, this source should have voltage between 10,000 and 50,000 volts, depending upon the particular design of the magnetron. The magnetron itself may generate oscillations having a frequency anywhere in the range from 300 to 30,000 megacycles, more or less.

Fig. 3 shows a modification of the system of Fig. 2. In Fig. 3, the magnetron differs somewhat from that shown in Figs. 1 and 2 in the absence of a control electrode. Instead of the control electrode described in connection with Figs. 1 and 2, Fig. 3 employs a hot cathode 4'. The non-thermionic or cold cathode 2 has certain ones of its edges curved slightly inwardly to provide a target area for the electrons emanating from the hot cathode in order to produce secondary emission from these curved ends. Control potential from source H is now applied between the cold cathode 2' and the hot cathode 4', as shown. It should be noted that the electrons from the hot cathode i first strike the curved ends of the slots of the cold cathode to produce secondary electrons which then emerge from the slots to be added to the circulating space charge. The electrons emerging from the slots of the cold cathode 2' will, of course, strike the exterior surface of the cold cathode to produce additional secondary electrons. It has not been deemed necessary to show the heater circuit for the hot cathode in the interest of simplification of the drawing. The elements of the sytsem of Fig. 3 which are the same as the elements of the system of Fig. 2 have been given the same reference numerals, while the elements of Fig. 3 which are equivalent in purpose or structure to those of Fig. 2 have been given the same reference numerals with a prime designation. The operation of the system of Fig. 3, except for the difference mentioned above, is the same as that of Fig. 2 and will not be repeated.

Fig. 4 shows another embodiment of the invention wherein a magnetron and the associated circuit are slightly different from the magnetron and circuits of Figs. 2 and 3. In Fig. 4, the magnetron is shown along a section parallel to the axis, rather than perpendicular to the axis. The cold cathode in Fig. 4 is represented by the reference number 2", while the hot cathode is represented by the reference 4". The cold cathode of Fig. 4 is made up of a cylinder, one end of which is closed and connected to the source of control pulses H, as shown, and the other end of which is open to permit the emergence of the leads from the hot cathode i" to the filament heating transformer 10. The slot 8 in the cold cathode is circumferential instead of parallel to the axis, as hereinbefore described. This slot 8 may approximate one-third of the circumference of the cold cathode 2". It should be understood that there may also be other slots in the cold cathode. The anode 5 is substantially the same in construction as the same numbered elements in Figs. 2 and 3. The magnetic field in Fig. 4 is produced by a pair of pole pieces marked N and S, representing north and south. These pole pieces are connected together by a yoke which is surrounded by a coil 20, in turn energized from a direct current source 2i through a variable resistor 22. The arrangement of the magnetic field is such that it is somewhat tapered in strength along the axis of the magnetron with a minimum at the center so that the circulating space charge tends to more nearly accumulate where it is wanted and not to diffuse out to the end walls. Putting it in other words, the tapering magnetic field is such that electrons tend to concentrate in a plane at right angles to the axis of the tube located at the center. Because the intensity of the magnetic field is a minimum at the center, the circulating electrons have a tendency to drift toward this minimum field location. Suitable glass seals 23 serve to provide a vacuum tight enclosure for the magnetron. The leads from the filament heating transformer it! to the hot cathode and the lead from the source II to the cold cathode, as well as the lead from the output circuit to the loop l5, enter the interior of the magnetron through these glass seals. A metallic can-like arrangement 2 provides an envelope for the magnetron. An artificial line 25 serves the same purpose as condenser E3 of Figs. 2 and 3 but gives a modified wave form of input potential and current for the magnetron. The system of Fig. 4 has the advantage of providing a more nearly rectangular wave form for the pulses obtainable from the output loop I5.

An alternative arrangement for the line 25 of Fig. 4 is shown in Fig. 4a. It consists of a series of circuits each containing inductance and capacity in parallel. These taper in size as one goes from the source i2 to the magnetron. This arrangement, I believe, is called a Guillemin line,

after Prof. Guillemin, its inventor.

Fig. 5 is another embodiment of the invention and again shows the magnetron in cross-section in a plane passing through the axis of the tube.

The magnetic field has been shown in Fig. 5 for the sake of simplicity of the drawing, as taking the form shown in Fig. 4, although if desired it may take the form shown in Figs. 2 and 3. The non-thermionic or cold cathode of Fig. 5 is shown as a rod or hollow metallic cylinder 26, while the hot cathode is at one end and designated as 27. This hot cathode may be of the indirectly heated type as shown and serve to supply the priming electrons for bombarding the cold cathode 26.

Although the hot or thermionic cathode of Fig. can emit continuously, movement of electrons from the hot cathode to the control electrode can be prevented or reduced sufiicently by making the thermionic cathode sufficiently positive with respect to the cold cathode. It takes only moderate potentials, as compared with the anode-to-cold cathode potential, to control electron motions in directions parallel to the magnetic field. The operation of the system of Fig. 5, except for the difference just po nted out, is substantially the same as that described above in connection with Figs. 1 and 2.

My construction of the magnetron which employs the use of a cold cathode and a thermionic cathode, to the latter of which a controllable potential can be applied, has the practical advantage that during the building of the tube there can be appl ed temporarily much more than nor-- mal potential between the thermionic and cold cathodes, thus heating the cold cathode high enough temporarily to activate it. This cold cathode may be oxide coated, similar to the hot cathodes in the known types of magnetrons. The magnetron of the invention has the further advantage that the equ pment needed with it is much less expensive, much less bulky and much lighter than many prior art alternatives used in pulsing transmitters. I might thus more readily use the magnetron of the present invention in portable and mobile military equipment.

The systems of Figs. 1 to 5, inclusive, are well suited for telephone and like types of pulse communication in which the pulse length is held constant and the pulse frequency or pulse timing is varied in response to modulation. From a practical standpoint, the magnetron of Figs. 1 to 5 is not so well suited for modulation which requires varying the length of the pulses.

Figs. 6 and 6a illustrate a magnetron in accordance with another embodiment of the invention, and Fig. 7 shows this new type of magnetron in connection with a circuit arrangement, as a result of which the pulse oscillations can be both controllably started and stopped, even though a substantially constant direct current potential is maintained between the anode and the cold cathode. This magnetron is therefore suitable for pulse length modulation in addition to the other types of modulation.

eferring to Fig. 6 in more detail, it have shown a magnetron having a hollow col cathode of the type generally illustrat n 4, except that this cold cathode is provided with a pin. rality of circumferential slots The col-cl cathode til accommodates in its interior a hot cathode 3! having leads which extend externally of the magnetron through a glass sea-l The cold cathode is provided with a metallic tube ti which terminates in a disc-like terminal '33. Tube 3'! shields the heater leads for the hot cathode 3!. There is provided a strong atrial magnetic field for producing lines of flux extending parallel to the axis of the cold cathode, and this accomplished by means of afield coil Electron absorber electrodes are provided at opposite ends of the cold cathode, in order to stop the oscillations in a manner to described later in connection with Fig. '7. These ahsorher electrodes are supported by metal rods The metal support rods are positioned in places of balanced high frequency field and so are very little coupled to the elfective oscillating circuit. A suitable evacuated metallic can-dike arrange ment 2G constitutes the envelope of th device and glass seals 23 serve to provide a vacuum tight enclosure for the elements within the can 25. The usual output loop 15 is shown for deriving high frequency oscillations from the magnetron.

Fig. 6a is a cross-section of the magnetron of Fig. 6 along the lines 6c-6a. The magnetron of Figs. 6 and 6a is shown in connection with a complete circuit arrangement in Fig. '7. In Fig. '1 there is provided a source ll of control pulses of a relatively low power which may be modulated in length, frequency or timing by the use of conventional vacuum tubes and circuits. Source 5: is coupled to the two electron absorber electrodes 33 through a coupling transformer it. It should be noted that both absorber electrodes are connected together by means of supports 30 A source [2 of high direct current potential is shown coupled across the cold cathode fit and the can 24. A condenser 36 in series with the high impedance smoothing reactor it is shown shunted across the source l2. Condenser is a. large smoothing condenser which prevents substantial potential drop across itself in response to pulse currents into the magnetron. It is much larger than the storage condensers of the other figures.

In the operation of Fig. 7, the hat cathode ill held at a positive or a relatively low negative potential with respect to the cold cathode so that, as a result of this potential and the pres ence of a strong axial magnetic field produced by field coil 40, substantially no electrons emitted by the hot cathode 3! pass out through the slots 32 of the cold cathode 353. However, by pulsing the hot cathode 3i suiiiciently negative with respect to the cold cathode 3d, electrons from the hot cathode will pass out from the slots of the cold cathode and will cause secondary emission from the cold cathode and accumulation of rotating space charge between the anode ii and the cold cathode 3!. If he anode 5 is sufliciently positive with respect to the cold cathode so, but not more positive than the magnetron cut-off potential for the magnetic field strength used, oscillations will start as soon as sufficiently large circulating space charge is accumulated and these oscillations will continue as long as the anode potential remains high enough.

The starting of oscillations in the system of Fig. '7 will cause a rapid increase in anode tm cold cathode direct current, which in turn may cause a decrease in anode-to-cold cathode potential due to reactance l'l. decrease in potential is an aid to growth of total cold. cathode emission due to secondary emission, by causing out-of-phase or wrongly timed electrons to strike the cold cathode with greater energy. It is im' portant to control the amount of the potential drop, by adjusting the value of reactance ii, to prevent the potential from falling too low; otherwise emission may fail again and cause oscillations to stop too soon, after which the oscillations may start again.

To prevent the starting and stopping of oscillations from causing undesirable amplitude and frequency modulation of the radio frequency ou put current, which will throw the energy out over a very wide frequency band at the expense of decreased energy in the desired frequency band. it is important that the total eifective series pedance in the direct current input circuit to the magnetron be kept low but not nearly zero. If pulses effectively one microsecond long with peak currents of about 30 amperes about 18,000 volts are desired, then I have found, under one set of conditions, that the direct current input circuit may contain not more than about 1000 microhenrys of inductive reactance, and better results seem to be obtained with about 200 microhenrys of reactance, provided the direct current supply potential and magnetic field strength are properly coordinated. That is, for those conditions, reactance i? should have a value less than 1000 microhenrys. These figures are given for a particular magnetron desired to operate at about 13,000 gausses magnetic field with an initial peak potential up to about 25,000 volts and during the main body of the pulse of about 18,000 volts. In this magnetron, the cathode diameter was about one-quarter of an inch, while the anode had an inside diameter of about one inch.

Once oscillations are started, in the system of Fig. 7, if the anode potential remains high enough, the oscillations will continue indefinitely. I have provided a means to stop oscillations comprising a pair of electron absorber electrodes 33. During oscillation, these absorber electrodes are maintained at the same or a more negative potential than the cold cathode 30, so that almost no electrons are absorbed by them from the circulating space charge. However, if oscillations are once started, and it is desired to stop them again, I propose pulsing the absorber electrodes 33 to a positive potential with respect to the cold cathode 31! by means of source II. By pulsing the absorber electrodes, they will exert a component of force upon the electrons in a direction parallel to the magnetic field and this force will greatly reduce the space charge by absorption of circulating electrons and by reduction of bombardment of the cold cathode. At the same t me, secondary emission from the absorber electrodes 33, by wrong timing and unsuitable dimensions for aiding oscillations, will throw electron loading on the oscillation circuit, thereby tending to stop oscillation. Since, after oscillations have been started, the margin of excess secondary emission from the cold cathode 30 beyond that required to maintain oscillations may be made quite small; the disturbance to the space charge produced by positive potential applied by source H to the absorber electrodes 33 need not be very great to stop oscillations. The stopping of oscillations, like the starting, tends to be regenerative or self-helping. Absorption of space charge by the absorber electrodes, by reducing the anode-to-cold cathode current, tends to cause the anode potential to rise. A rising anode potential reduces the energy of electrons striking the cold cathode and this also tends to reduce secondary emission. Explained in another way, the input impedance of the magnetron tends toward zero or even a negative alternating current impedance, so that it tends toward starting and stopping itself in pulses, as a result of which the amount of control energy from source i l to cause either starting or stopping of oscillations may be made very small.

In summing up the description of the operation of the system of Fig. 7, it may be said that the start of a direct current or rectangular wave control current pulse from source H momentarily forces the inner hot cathode to be negative with respect to the cold cathode, as a result of which electrons pass out through the slots in the cold cathode and cause circulating space charge to accumulate. The magnetron oscillations thus start and the magnetron passes heavy input and output power. The end of a control current pulse forces the absorber electrodes to be positive with respect to the cold cathode, as a result of which the space charge is thrown out of the active volume of electrons and the circuits are loaded and oscillations stopped. Thus, oscillations last as long as the length of the control pulses from source H. These control pulses are of relatively low power and may be modulated in length, frequency or timing by the use of conventional vacuum tube circuits already developed for the control of magnetron pulses, in which modulator tubes are used in series with the magnetron.

Although the magnetron of the present inven tion has been illustrated particularly with respect to a scalloped type of anode, of the kind generally described in my Patent 2,217,745, it should be clearly understood that the invention is not limited to this construction of anode since any suitable anode structure can be used, provided the growth of total emission from the cold cathode due to secondary emission can take place as a result of oscillation. By way of example,

.- reference is made to my application Serial No.

470,438, filed December 29, 1942, now U. S. Patent 2,424,886, granted July 29, 1947 for an alternative anode structure arrangement which is an improvement upon the anode structure here shown in that it has only one resonant frequency and is therefore proof against oscillation on undesired or spurious frequencies.

What is claimed is:

1. A magnetron oscillator having a hollow anode within a housing, a cold cathode located along the axis of said oscillator within said anode and capable of emitting secondary electrons upon bombardment by electrons, and a thermionic cathode also located on the axis of said oscillator in the same straight line with said cold cathode but physically spaced therefrom, and means adjacent said housing for producing a constant magnetic field having flux lines extending near-- ly parallel to said axis and of such intensity as to force the electrons emanating from said hot cathode into the anode-cathode space to produce a circulating electron current, said thermionic cathode comprising a hollow cylindrical surface having heating means therewithin.

2. Electron discharge device apparatus comprising a hollow cold cathode whose outer surface is capable of emitting secondary electrons upon bombardment by electrons, an anode surrounding said cold cathode, a thermionic cathode in the interior of said cold cathode, said cold cathode having at least one aperture for permitting electrons emitted by said thermionic cathode to escape into the space between said cold cathode and anode, means for producing a constant magnetic field whose flux lines extend nearly parallel to said cold cathode, a pair of electron absorber electrodes placed at opposite ends of said cold cathode, and a source of pulses.

connected to said absorber electrodes.

3. In combination, electron discharge device apparatus comprising a hollow cold cathode.

whose outer surface is capable of emitting secondary electrons upon bombardment by electrons, an anode surrounding said cold cathode,

a thermionic cathode in the interior of said cold cathode, said cold cathode having at least one aperture for permitting electrons emitted by said thermionic cathode to escape into the space be-.

tween said cold cathode and anode, means for producing a magnetic field whose flux lines ertend nearly parallel to said cold cathode, a pair of electron absorber electrodes placed at opposite ends of said cold cathode, a direct connection of low resistance between said absorber electrodes, and circuit means for pulsing said absorber electrodes to a positive potential relative to said cold cathode.

4. A cold cathode magnetron oscillator comprising an anode and a non-thermionic secondary emissive cathode in the center of said anode, means to supply a priming electron current for bombarding such secondary emissive cathode to start oscillation, and means including a surface within said oscillator to absorb circulating electron space charge to stop oscillation, said means also including a circuit for applying a momentary charge on said surface which is positive relative to said cathode.

5. Means to stop oscillations of a cold cathode magnetron oscillator comprising a cathode having ends, electrodes oppositely disposed with respect to the ends of said cathode, and means for applying to said electrodes a controllable potential positive relative to said cathode whereby said electrodes absorb at least a portion of the circulating space charge from the magnetron to cause oscillation to cease.

6. A magnetron electron discharge device comprising a resonant anode structure having a plurality of electron target portions protruding inwardly toward the center, a cold cathode capable primary electrons also located along said axis but to one side of and in axial juxtaposition with said cold cathode and adapted to be heated by a source of externally located electric-a1 energy, and means in proximity to said anode structure for producing a magnetic field having flux lines extending parallel to said axis.

7. A magnetron electron discharge device comprising a resonant anode structure having a plurality of electron target portions protruding inwardly toward the center, a cold cathode capable of emitting secondary electrons upon bombardment by primary electrons and extending along the axis of said device and located in the center of said resonant anode structure, a source of primary electrons also located along said axis but to one side of said cold cathode and adapted to be heated by a source of externally located electrical energy, and means for producing a magnetic field having flux lines extending parallel to said axis, an energy storage circuit coupled to said cold cathode, a source of direct current potential connected across said storage circuit, and a source of control pulses coupled between said cold cathode and said source of primary electrons.

8. A magnetron oscillator having a generally annular anode, a secondarily emissive cathode disposed axially with respect to said anode, a thermionic auxiliary emitter and a control electrode adjacent said auxiliary emitter and means for applying a control potential to said control electrode of such magnitude and polarity as to control, by virtue of the voltage of said control electrode relative to said auxiliary emitter, the

1 1 transit of electrons emitted from said auxiliary emitter towards the anode-cathode space.

9. A circuit for producing intermittent pulses of ultra-high-frequency oscillations including a magnetron tube having an annular multicavity anode and a compound cathode comprising a secondarily emissive cathode and a relatively small thermionic auxiliary emitter, means for preventing thermionic emission of said emitter from exciting said tube in the absence of a trigger pulse applied to said circuit, a reactive pulseforming network adapted to discharge upon initiation of oscillations in said tube by emission from said auxiliary emitter with the formation of a substantally rectangular electrical pulse,

means including a source of high voltage for impressing a relatively high potential between said cathode and said anode and for charging said reactive pulse-forming network, and current limiting means connected in series with said source of high voltage adapted to permit the discharge of said network as aforesaid and to reduce the said voltage between said anode and said cathode at the end of such discharge to a value low enough to interrupt oscillations of said tube.

10. A circuit for producing intermittent pulses of ultra-high-frequency oscillations including a magnetron tube having an annular multicavity anode and a compound cathode comprising a secondarily emissive cathode and a relatively small thermionic auxiliary emitter, a reactive pulse-forming network adapted to discharge upon initiation of oscillations in said tube with the formation of a substantially rectangular electrical pulse, means including a source of high voltage for impressing a relative high potential between said cathode and said anode and for charging said network connected between said cathode and said anode, means for efiectively maintaining said auxiliary emitter at a voltage positive with respect to said secondarily emissive cathode and negative with respect to said anode and means associated therewith for periodically lowering the positive potential of said auxiliary emitter, and current limiting means connected in series with said source of high voltage and adapted to reduce the said voltage between said anode and said cathode for a predetermined time after the discharge of said network to a value low enough to interrupt oscillations of said tube.

11. A magnetron oscillator tube having a generally annular anode and a compound cathode positioned substantially coaxially therewith, said cathode comprising a large secondary emissive portion and a small thermionic portion comprising a short overall length of tubular surface positioned in axial juxtaposition with said large portion, and means for heating said surface independently of said large portion.

CLARENCE W. HANSELL.

References Cited in the file of this patent UNITED STATES PATENTS Number Name Date 2,151,766 I-Iollmann Mar. 28, 1939 2,400,770 Mouromtseff et a1. May 21, 1946 2,409,038 Hansell Oct. 8, 1946 2,411,601 Spencer Nov. 26, 1946