US2587750A

US2587750A - Vacuum tube oscillator system

Info

Publication number: US2587750A
Application number: US58510A
Authority: US
Inventors: Morrison Montford
Original assignee: Individual
Current assignee: Individual
Priority date: 1948-11-05
Filing date: 1948-11-05
Publication date: 1952-03-04
Anticipated expiration: 1969-03-04

Description

MalCh 4, 1952 M. MORRISON 2,587,750

VACUUM TUBE osoILLA'roR SYSTEM Filed Nov. 5, 1948 5 Sheets-Sheet l March 4, 1952 v M. MORRISON 2,587,750

VACUUM TUBE oscILLAToR` SYSTEM Filed NOV. 5, 1948 5 Sheets-Shes?l 2 l l Jl/L' A FJ l Amir/h I 1 1 l 'M L l l L l l www 4"7"- Y |r W l W Puff Camif/r' n A MIA .t0/MMM;- l l March 4, 1952 M. MORRISON VACUUM TUBE oscILLAToR SYSTEM 5 Sheets-Sheet 3 Filed Nov. 5, 1948 M. MORRISON VACUUM TUBE OSCILLATOR SYSTEM March 4, 1952 F'iled Nov. 5, 1948 5 Sheets-'Sheet 4 Ey. J5

ffy. /4

R bN

March 4, 1952 M. MORRISON 42,587,750

VACUUM TUBE oscILLAToR SYSTEM Filed Nov. 5. 1948 5 sheets-sheet 5 o A j mw xwmw Patented Mar. 4, 1952 UNITED STATES ATENT OFFICE Claims.

The present invention relates generally to electron discharge tube oscillators, in particular to systems of feed-back in such oscillators, and more specifically to the employment of multiple feed-back systems to a single tank circuit.

This application is a continuation in part of application Serial No. 1,595, nled January 10, 1948.

Among the objects of the invention are: to provide an oscillator which may operate at a frequency having a degree of freedom independent of the natural period of the controlling tank circuit; to provide an oscillator which may opcrate over a wide range of frequencies with a fixed LC in the controlling tank circuit; to provide an oscillator which may operate at the maximum degree of controlling-tank-circuit apparatus eiciency for any character of oscillator load; to provide an oscillator which may operate at the maximum degree of controlling-tank-circuit electrical elciency for any character of oscillator load; to provide an oscillator with an indirect inductive reactance stabilized feed-back; to provide an oscillator with a direct condensive reactance stabilized feed-back; to provide multiple feed-back circuits of different reactive characters, permitting variation in oscillator frequency t performance with Xed LC in the tank control circuit; and various and other objects which will be pointed out and obvious to those skilled in the art to which the invention appertains, upon reading the specication hereunder, in connection with the accompanying drawings.

A The nature of the invention resides important- 1y in the employment of oscillator structure and method of oscillator operation, which depends upon a novel discovery in oscillator functioning, orat least in a novel combination of underlying theories, not heretofore collectively consolidated into a working explanation of oscillator operation. y

VIt; is believed that in order to give a full, clear and exact description of the invention, it will be necessary to provide a fuller, clearer and more exact theory of oscillator operation, as applied to the present invention, than is known to the applicant in published texts. The applicant will provide herein such eXtra-conventional theory as is thought to be pertinent.

The invention will be Vmore .fully understood from the following description and eXtra-conventional theory, when read in connection with the accompanying drawings, of which:

Fig. 1V is 'a diagram illustrating the. method employed by the applicant .tov determine .the

phase angle between the voltage and the current, the gure of merit Q, the resonant frequency, the frequency of the natural period or free period, and the frequency under equal reactance or equal susceptance operation, of a tank circuit when referred to herein.

Fig. 2 shows graphs illustrating some 'of the operation of Fig. 1.

Fig. 3 shows a particular oscillator operating under condensive loading and Figs. 4 and 5 are other oscillator circuits which are used in illustrating the' extra-conventional theory given herein;

Figs. 6, 7 and 8 are graphs relating to the oscillators shown in Figs. 3, 4, and 5; and

Fig. 9 is adiagram illustrating the methods of making tests which define certain terms used in the extra-conventional theory presented herein.

Figs. 10 and 12 are simple circuits illustrating and correcting some of the fragmentary theory vectorically represented in Figs. 11 and 13 of the prior art, not heretofore associated with oscillator operation.

Fig. 14 is a diagram of an oscillator circuit having a degree of frequency freedom independent of the LC of the tank circuit controlling the oscillator output.

Fig. 15 is a diagram of an oscillator having a reversed inductively reactive stabilized feed-back, and which oscillator may operate at the resonant frequency as well as at other frequencies of the controlling tank circuit, and Fig. 16 is a vector diagram relating to the operation of Fig. 15.

Fig. 1'7 is a diagram of an oscillator having a multiple or polyphase feed-back, one of which is reversed and one of which is direct. Under multiple feed-back operation, this oscillator may operate at the resonant frequency, as well as at other frequencies, of controlling tank circuit; and Fig. 18 is a vector diagram relating to the operation of Fig. 17.

Fig. 19 is a diagram of an oscillator having multiple feed-back; one feed-back being in phase with the plate voltage and having a magnitude proportional to the amplitude thereof, and one feed-back being in phase with the effective alternating plate current, and having a magnitude proportional to the amplitude thereof. Under multiple feed-back. operation this oscillator may operate at the resonant frequency, as well as at other frequencies, ofthe controlling tank circuit, andFig. 20 is a vector diagram relating to the operation of Fig. 19. y The outstanding literature of the prior art on oscillators is replete with statements which indicate that the authors did not appreciate the operational theory as set forth by the applicant herein.

In Morecrofts book, Principles of Radio Communication, Wiley, 3rd edition, immediately after showing (beginning p. 575) that the grid voltage of an amplifier is in phase with the tube voltage and tube current only under conditions of having a pure resistance load, he, on p. 581, under the heading General analysis of conditions necessary for self-excitation (of oscillators), states; The plate potential and the grid potential both undergo sinusoidal variations of potential in op-V posite phases, and that the relative magnitudes of these two potential variations can be properly adjusted for the tubes being used.y Further,- he bears out this statement in his Fig. 146,*p. 591. This shows that he did not consider his preceding amplifier theory to apply to oscillator operation.

Prince and Vogdes in their book Vacuumy Tubes as Oscillation Generators, publishedy in 1929 by General Electric Company, devote their entire Chapter V to a treatment to show how the grid voltage can be made to operate in phase (opposite) with the plate voltage (on an inductive load) and thereby accomplishing whatv Morecroft set forth, as above related', to the operation of the grid voltage in opposite phase rela,- tion tothe tube voltage. The statements made in their Chapter V show that the authors not only believed that these phase relations could be met (with an inductive load) but attempted to show how it could be accomplished.

Terman in his book Radio Engineering, Mc- Graw-HilL 1937, on page 356, under the heading "67. Frequency and frequency stability of generatedoscillations, states: The alternating current generated by the vacuum tube oscillator has a frequency such that the voltage which the oscillations apply to the grid of the tube is of eX- actly the proper phaseto produce the oscillations that supply the required grid exciting Voltage. This approximates the resonant frequency of the tuned circuit, but This statement must4 mean that the feed-back voltage is applied to the grid and that the resulting frequency approximates the resonant frequency ofthe tuned circuit.

There are many other consistently expressed statements in the prior art literature such as, that an oscillator should operate at the natural period of the tuned circuit and related statements.

The applicant has discovered that:

(a) The operating frequency of' a vacuum tube oscillator bears no necessary direct relation to the resonantr frequency of the tuned circuit.

(b) The natural period of a tuned circuit has no physical significance in a vacuum tube oscillator operation, except as it is related by definition to the figure of merit Q of the circuit.

(c) The grid voltage of the tube can be in opposite phase with the tube voltage and in phase with the tube current', only under conditions when the load is of an effectively resistive character. Y

(d) Under reactive loads of'any character, the grid voltage, the tube voltage and the tube current are always all out of phase, one with the other.

(e) In oscillators, the grid voltage, the plate voltage andl the plate current, have the same relations as given for amplifiers by Morecroft above, for inductive loads, but his amplifier analysis does not apply to capacitive loads, and it is not seen how it correctly represents amplifier operation. That is, the character of the load Xes what the applicant terms the grid operating voltage angle, which is the phase angle at which the grid must operate under the given load. For a reactive load this angle always lies in between the tube voltage and the tube current (plus or minus (f) In an oscillator operating under a reactive load, any attempt to bring the phase of the grid voltage into a phase position different from its naturally assumed phase position, will cause a shift of oscillator frequency to a different frequency, which will cause the grid to assume a new proper out of phase position. The phase of the grid can-not be.corrected on a reactive load to anarbitrary phase position.

(g) In a tuned grid reactively loaded oscillator, the generated frequency is determined by the frequency at which the grid tank circuit must operate to produce the required grid operating voltage angle (G. O. V. A.), between the tube voltage and the grid voltage, which may be lead- ,J ing or lagging depending upon the character of the load, and therefore the oscillator frequency may be below or above the resonant frequency of the grid tank circuit.

(h) The figure-of merit Q of the tank circuit taken at resonance is not the operating Q of the circuit underv reactive load, because the tank circuit does not operate at resonance under these conditions.

(i) The frequency stability ofA a resistance stabilized koscillator is not necessarilyrcritical to the Q ofthe tank circuit, because the frequency stability is dependent upon the G. O. V. A. being maintained at the correct position with variation in the other circuit parameters, rather than. the circuit dissipation.

(i) The variation in the tank circuit Q with different applied tank circuit voltages plays an important role in compensating for variation in the phase angle of the load current with different applied load voltages.

(lc) The grid tank circuit in addition to functioning as an electrical flywheel for the circuit performs the important function of acting as an automatic G. O. V. A. adjuster.

(Z) In a grid tanvc circuit oscillator having an inductively reactive plate load, inductance variations and/or variation in the Q of the load due to any cause whatever, and generally importantly to variations in the applied rload Voltage, cause changes in what the applicant terms the load imposed current angle (L. I. C. A.), For each change in the L. I. C. A., there is required a correct change in the G. O. V. A., to maintain the lsame oscillator operating frequency. Oscillator operating frequency stability is attained when the grid tank circuit (including grid current) automatically adjusts the G. O. V. A. to the correct value .for the new L. I. C. A. at the correct operating frequency.

(m) In a resistance stabilized grid tank circuit oscillator having an inductively yreactive plate load and a single` feed-back circuit, a frequency is best stabilized against plate voltage variations, when the tank circuit is operating at a frequency considerably removed from the resonant value thereof;

(n) In a resistance stabilized grid tank circuit oscillator having an inductivelyv reactive plate load, the most purely sinusoidalplate currentl is attained, when the tank circuit is operating at a of 10,000 ohms.

frequency considerably removed` from the resonant value thereof.

Q (o) When van oscillator tank circuithas an intermittent drain imposed upon it, such as by intermittent grid current, tank vcircuit replenishment current in phase with the intermittent drain, distorts the sinusoidal character of the tank' circuit voltage.' Therefore an oscillator should not 'be operated with thesetwo factors in phase, if sinusoidal operation is desired.

(p) Oscillators having inductively reactive plate loads, have an inherent tendency to produce sinusoidal currents and distorted tube voltages. (q) 'oscillators having condensively'reactive plate loads, have an inherent tendency to produce distorted 'current-sand sinusoidal tube voltages. v

(r) Oscillators having condensively reactive plate loads when operating independently. are more diiiicult to stabilize than oscillators having inductively reactive pl-ate loads, but are more sensitive to injected synchronizing currents, than the'latter.

' The above statements'represent some important discovered facts upon which the present disclosure on which the illustrated embodiments of vmy invention depend. Y

In order to establish these discovered facts, the applicant will now treat the extra-conventional theory required to fully understand these statements. n

` To makeit perfectly clear what the applicant means by the terms he uses herein, those terms which are not clearly fixed by universally accepted definitions or by universallyV accepted methods of experimental determination, will be treated in detail.

All of the discussions herein relating to frequency are made', for simplicity, with audio frequencies in mind, and the circuits have all been checked at 600 cycles or around this frequency 'as a center point. All discussions and determinations are made for operating currentor voltage values, and are not made under artificial conditions.` For instance, the Q of an iron-cored coil. not only depends upon the frequency, -but also upon the voltage applied to it, so that the Q thereof made under conditions other than operating conditions, in general does not represent 'the operating Q. While some of such coils show a fairly constant inductance for increasing values of applied voltage or current, none possess a constant Q under such conditions. an example, the inductance of a good iron-cored coil may not increase more than 0.1 for double applied voltage, whereas the Q of this coil may 'drop off` 10% under the same conditions, and a good understanding of this fact is essential to a clear comprehension of the disclosure herein. The method of measuring the phase angle, as

used herein, of a tank circuit is illustrated in 'Fig. l,v by closing the switch |0|. The generator "line voltage is projected upon the oscilloscope screen, through the A input of the electronic switch, this image also, under the conditions ofv test, represents the tank circuit line current. The generator line voltage measurement, which is coupled to the A input of` the electronic switch, i-s made by a resistance drop method, but across a small percentage of the total resistance used. Referring to the gure, the total resistance of the resistor |02, may be of the order of 1,000,000 ohms, and the resistance `across the input of transformer |03, may be of theorder Transformer |03, is preferably (Radio an extremely high-fidelity extremely small type such as go by such trade names as, ouncersr inchers and so forth, and should have an input impedance of the order of 1,000,000 ohms. The fidelity and ph-ase angle vof this coupling can always be checked, by use of the electronic switch projecting the transformer input and output voltages on the oscilloscope-screen, simultaneously. Y The form and phase position of the tankl circult voltage is projected superimposed upon the line voltage image upon the screen, by means of a special A1 amplifier through input "B of the electronic switch. With a precision oscilloscope; both' wave amplitudes and phase displacement can be read with a good degree of accuracy. Y The special A1 amplifier is' constructed to have, at the frequencies employed, negligiblephase difference between the input and output voltages, but without any 'necessarily high amplification.- Such an amplifier may be constructed like a resistancecoupled-outputy design, but with the" coupling capacitor having arelatively large capacity, and the output resistor having an ohmic value relatively-very high to that of the plate resistor? Phase' difference between the input and output can be tested by an electronic switch as related abovef Y l i By varying the frequency of the 40 polegenerator, the familiar resonance' curve, together withA the related curve showing the phase angle between the tank-circuit voltageV and the tank circuit-line current can be accurately determined under conditions representing operating voltage andv current values. Such a set of curves- -is shown in Fig 2, line C. These are the tank cir"- cuit-voltage and current values and phase relations referred to herein. J .There seems to-'be some confusion in-the literature about the definition-of Q-and-the method of determining it. It is usually defined as the ratio V y fi but it isl sometimes also. defined in the same text a-s the ratio volt amperes watts Engineering Handbook, ,1 -lenny, McGraw-Hill, 1941, pp.,1323) which is notthe same expression because the latter The two expressions are only approximately lequal for'l-arge values of Q.

For values o'f Q, where the phase angle between 'the voltage` and current of the coil underv test .can be accurately determined by oscilloscopio indicationcircuit of Fig. l, is employedphy clos'- Ving switch |0|, and opening switch "|04, The Q of the coil alone as defined by the ratio lrapidly even with ,small angular increments, that -accuracy in this range is difficult to attainqay this method. .For these higher values of Q the applicant measures vthe series capacity vrequired to neutralizeetbefinduetance of '.tnecoil-at' the operating -nd frequency,.by closing switch 1.0i, switches H14.; and !0.1.and.measurr.1gthe capacity-required@ condenser l06. .to produce resonancaas indicated @n the oscilloscope screen. Eromg this capacity reading the value Aof -Lis conmutedandl sdeterminedfrom the COlVOlt- 71th thefvalueeof.Lknown Q iS then-.deh d for each or all values of voltage-,and frequency desired- I'e accuracy inaadiierence of phase. angle 4 We measured on the oscilloscope screen, which frequires a xed'positionof the tracesof the were ivrms AIneasurecl .independent f of V'the amplitiidcof the .-synchronizingvoltage, the applicant employs A the method of synchronization .1..is =l is.f1v in latentfno. 2.435.751..

hes'applicant herein distinguishes lbetween the' veteranen of a l parallel resonance `circuit ,underonditions of resonance operation, of

natural period operation, and of'equal rgactance one 'omas Set. forth by. H.J Boyland inExper-imfntal Wireless foi-November `1927. H owever attention -is called to the kfact that this articleffcontains formulas .having .typographical neef-"of the apparatus lshown in Fig. 1. The

method employs the use .of the synchronous con-.- ta'oto'rwhich closes the generator supply voltage .to the L.parallel resonance circuit for a period of time sufficiently long to permit the .circuit .to reach al' steady `state. condition, and -then jdis- 'connects-theisupply voltage and allowsthe cir- .cuitr to oscillate at' .itsjfreej (natural) v period long enough to count the number of natural periods as compared to a denite number of timing Waves thrown on to the oscilloscope screen simultaneously;..by .means .of the electronic switch. This method determines the natural period of the cir.-` cuit under actual conditions of the operation.

Fig..j2, line A showsthe'appearance of the natural period being measured, and line B, shows y ythe:appearance:off-the timing wave. In practice AlinesrA and B,- are superimposed *for convenience tor 308,.across theN central position of` resistor 3.0 l When switches 306 and 301 are openandswitches 3.02 yand .303 are. closed .thezdiagram will be referred tor-as circuitC. When switches 3.06. and Q01-are closed. andswitches .302 and 3.03 areopen thed-iagram willbe referred to as circuit D.

Fig. 4 shows a push-pull oscilla-tor 4having a resistance stabilized ,feed-back and purey resistive loadingin vthe plate circuits. The udiagram .of this gure .vwillbe ref erredto as` circuitE.

,F ig5 shows .a push-pull. oscillator having a. resistance stabilized feed-,.back; theplatecircuit has a split inductor. 50| .with a center tapped resistor 502. inserted. between the split coils. Y.This part of the diagram will be referred to as .circuit F. The figure also has .a set 1of switches 503.and 504 for introducing parallel variable capacitor 505 acrossfsaidsplit inductor; withswitches503 and 504closedvthe gure vwill bereferred toas circuit G. Y

Thedetailsof Figs 6, 'land 8 will be discussed under the operation of Figs...3, Land 5,.'butibefore these operations are taken up, it is necessary to establishand define certain factors whichenter intosaid operations, `andwhich are ,explained in connection with In Fig. 9, the part of the diagram-ivhichlies entirelyfoutside ofthe dottedareas I'and J, constitutes a circuit lidentical .with circuit IF Vof Fig. 5 .andwill be referred toas circuit Fi,.and while other circuits may be substituted linits place, the illustrated circuit suliices for purposes of illustration. `The circuitiFi of V,theflgure providedwith a set of switches l9D Iv and .$02, which may be thrown .to disconnect the vgrids .of ythe triodes fromthe .oscillatory system andv connect the grids to a variable frequency cententapped source ofvariable .alternating current 903. YThe figure shows anddenesthree methods otmeas.- .urement ,which will bereierred vto hereinafter, namely:

Grid `voltage measurement, which is employed to measure .voltage values and phase positions.

under conditions in .which no linecurrent andl no appreciable. line de-setting is permissible. This method` is Vusedalso for obtaining thephase .angle in counting and -measuring yIf variable ire- :circuit is below the resonancefrequency thereof,

and the frequency at4 which equal reactance is obtained is below the frequency of the natural period, as illustrated along line C, Fig. 2, together with formulas showing how the angles between :fthe ;yoltage;.;and.=. correnti calculated, 'fior the Yfre- .quency employed.

'.Fg-jii3 .showsia :push-pull I .oscillatorv having ==a resistance stabilized: `feed-bach; the plate circuit -has saflcenterftapped Yresistor y301, -one ,set "of fswtohes i302 `and 303 forlintroducing parallel '-variablecapacitors 304 land v'30'5across vthe-ends or resistor 30I,and a second set of switohesg :audaci-'1forintroducinenarallelyariablecapacibetweenthe W'Jltagev and thev ,current in A.a parallel oscillatory system, lunder similar conditions. YThe devicewithin dotted. area H `is aspecial Ai .amplier which is constructed to have negligible Vphase .difference between the input .andoutput tested by means of an electronic-switch and a cathode .ray oscilloscope.

Plate vcurrent measurements (thealternating lcurrent'component thereof) are madel by what may becalled .a resistancedrop method. Re-

.ferring to Fig. 9. the .center tapped. resistorlallll is a highly'accurate,noneinductive resistor which left permanently inthe ,circuit` and has suiicient resistance to give a usable lreading on high-resistanceyoltmeter V. M. The lvoltmeter .resistance shollld'be of ,the .order of 100 times that vof the resistor. Any direct fcurrent presentin the resistoris prevented. from .entering the plate. .cur- .rent nieasurementircuitby capacitor-.305. The

valueof thecurrent in .the resistor is obviously determined by Ithe voltineter. For ytaking oscilloformers sold as high-delity types may not have a zero phase difference between the input and output for the frequency used.

Tube voltage measurements (thealternating current component thereof) are made by a re- .sistance drop method but across a small percentage of the total resistance used, and if the total resistance used effects the circuit constants, `the shunt resistance is either left inthe circuit or an equivalent resistance is substituted, so that the operation of the circuit without the measuring device is the same as when it is in use. Referring to the circuit within the dotted area J,

the total resistance points 901 and 908 may bef of the order of 1,000,000 ohms, the effective rev ,sistance across the input of transformer 909 may be. of the order of 10,000. Transformer 909 is preferably an extremely high-fidelity .extremely small type such as go by such trade names as ouncers, inchers and so forth, and should have an input impedance ofthe order of 1,000,000

ohms. Y

These combinations should always be checked ,for phase difference. With 100 plateA. C. volts obtained w good oscillograph defections are through an electronic switch, with the values given. With the circuit as connected in Fig. 9,

lthe measurements arethose of the circuit as an oscillator. With switches 90| and 902 thrown to Vconnect independent A. C. source 903, the meas,- furements are made on an amplifier and obviously phase angle measurements can be made on various parts of the circuit at any desired frequency. A Certain experimental facts which the applicant has discovered will be established on the operation of Figs. 3, 'l and 5, based upon measurements taken by the circuits explained in Fig. 9.

Referring to Fig. 4, if the voltage values and.;-

phase angle are taken of the parallel oscillatory circuit alone of circuit E, as a function of the applied frequency at a constant effective current, there is obtained the familiar resonance-voltage curve 10| and its phase angle curve 102, with reference to said current. If said oscillatory circuit is tuned to resonance at say X cycles independently and then put into circuit E, and the feed-back resistor properly adjusted, circuit E can be made to oscillate at X cycles or the resonant frequency of said oscillatory circuit.

However if the same oscillatory circuit with its X cycle resonant tuning and with the same adjustment of feed-back resistors, is substituted in circuit F and the proper measurements made,

it will be found that circuit F does not oscillate at X cycles, but at some higher frequency 80|, Fig. 8, and with a lower output with the same plate impedance, because it will be found that the grid voltage is lower.

Then ifrswitches 503 and 504 are closed with ,low capacity adjustment ofV capacitor 505, it will be found .that by increasing said capacity the frequency and grid voltage can be brought to the values of circuit E. Further adjustments Vof said capacity will lower the grid voltage.

Now if the same oscillatory circuit with its X Y cycle resonant tuning and with the same adjustment of feed-back resistors is substituted in either circuit AC or D, say in'circuit I3. it will be 10 found that circuit D does not oscillate at X cycles, but at some lower frequency 60|, Fig. 6, and with a lower output with the same plate impedance, because it'will be found that the grid voltage is lower.

It willbe found that other factors, like amount of feed-back resistance, Q of oscillatory circuit at the frequency and voltage employed, and other factors which are beyond the scope essential to this disclosure, also affect the oscillator frequency.

From the above tests it has been discovered that the circuits C, D, E, F and G, can be made to oscillate at the same frequency by employing different LC values in the parallel oscillatory circuits.

If the parallel oscillatory circuit of circuit F is set so that circuit F oscillates at any 600 cycles, it will be found that the oscillatory circuit voltage values and their 'phase relation to the tube voltage (oscillatory circuit current) as a function of the circuit frequency, is represented by Fig. 8, in which the frequency corresponding to the line 80|, represents 600 cycles. Assume the phase lead shown to be say 60, which is a practical value. This means that the grid voltage lags the tube voltage by 60 and that the oscillatory circuit is not operating at the resonant point, but in the steep region of the high-frequency side of the resonance curve.

By increasing the inductance of coil that is by increasing the phase angle between the A. C. tube current and the A. C. tube voltage, the line 80| moves to thevright and the operating frequency is increased, and by *decreasing vthe inductance of coil 50|, the operating frequency is decreased. A

As it has been indicated, circuits C, D and G, can be adjusted so that the grid voltage leads the A. C. tube voltage at the operating frequency and say this frequency is 600 cycles and is represented by line Fig. 6. Obviously increasing the shunt capacity of these circuits, decreases the frequency thereof and decreasing the capacity increases the frequency, Within proper operating limits.

With these experimental phenomena disclosed, the applicant will establish a working theory for the making of oscillators in accordance with his invention, by the application of the analysis of amplifier operation.

In dealing with oscillators having an inductive load, use will be made of Morecrofts analysis of inductively loaded plate circuit amplifiers (in his. above-referred to book, page 575), but the analysis given by him for capacitively loaded amplifiers is discarded, because it cannot be validly applied, as willbe more fully pointed ou hereinafter. Y Y `While the Morecroft analysis used herein is not a mathematically exact theory, even for amplifiers, it does provide a practical Working theory, at least for audio frequencies, for the employment of the present invention, in the making of oscillators having inductive plate load.

A mathematically exacttheory.ofoscillators made in accordance with the teachings of the present invention, as determined by settingl up differential equations forthe operation of the tube with its connected circuits, -and obtaining solutions for them -for a particular set of conditions, by numerical methods, leads to difficulties. which prevent .such solutions from helpfulgin the making of oscillators.. a ,However-the.deferment ,of the .there .being given herein, is suliciently accurate to enable one skilled in the art, to make and use-oscillators` employing circuit elements having individually fixed circuit parameters, whichelements when properly combined into oscillatorecircuits, result'in oscillator operation ata predetermined frequency without circuit tuning, which is broadly new, to the best knowledge obtained by the applicant.

Referring to Fig. l0, this lwill beY recognized by those skilled in the art, as a circuit which may be operated either asV an amplifier or as an oscillator depending upo-n whether the switch 00! is closed upon contact i002 or upon |003. Switch |001 -is rst closed upon contact `0132 and amplifier operation is obtained from alterna-ting voltage source i004, and the resulting circuit is taken tooperate in accordance with the assumed theory and analysis of Morecroft.

Fig. 11 is taken substantially from the Morecroft book and representsthe approximate analysis, of the above circuit on'the assumptions made by him in arriving at his analysis. The vector notations are conventional and will be understood by those skilled in the art. For the basic assumptions andy development of the theory, those skilled in the art are referred to the book. It isl to be noted that the voltages and currents treated, are the alternating lcomponents of the voltages only.

In the operation of amplifiers, it is to be noted that with inductive plate load, the grid voltage Eg, does not and cannot b-emade to operate in phase with the tube voltage or in phase with the plate current. The grid voltage Eg (reversed in phase) must and does always lie,4A in phase relatiom between the tubevoltageEp'and tube current lp. This is setforth asV amplifier operation only and nowhere in the prior lart Idoes the applicant nd this analysis developed for oscillator theory and further, all oscillator theory which has come to the attention of the applicant, is inconsistent with this analysis when applied to oscillator theory and in some cases entirely contrary thereto.

If the amplifier of Fig. 10 resulting from closing switch |001 upon Contact |002, operates in accordance with Fig. 11, by changing the'source of grid voltage will not change the Vnecessary phase relation established between the, tube voltage Ep, the grid voltage Eg andthe ,plate current lp. Therefore ifswitch ll, is closed upon contact i003; forming an oscillator ofthe circuits of Fig. these saidphase relations must be maintained if the oscillator frequency is to be thesame'as that'under amplifier operation.

Since thephase relation between the grid voltage Eg and the tube voltage Ep, must remain as shown in Fig. 1l, the tank circuit I005'of iFig. 10, must supply this difference ofphase position of these two voltages. In the figure, the tank circuit current has the phase position of the tube voltage Epand the tank circuit voltage mustlag the tank circuit current, by the exact value which provides the grid voltage Eg, `as called for in Fig.

Cil

l2 cillate at a frequency higher thanV the-resonant frequency of the tank circuit, as described as an experimental result in connection with Figs. 5 and 8.

It is entirely impractical as well asimpossible to develop, in a patent application specification. this theory and analysis for all types of oscillator circuits, and all the discussion herein is directed to the types of circuits shown in Figs. 3, 4, and 5. It is believed that disclosing a good teaching of the theory applied to these types of circuits, will provide those skilled in the art Vwith su'icient knowledgeof the subject to enable them to apply and embody the principles in other types of circuits.

It is obvious, to those skilled in the art, that under the above teaching the grid voltage phase angle cannot be corrected to any angle different from the G. O. V. A. which is determined by the L. I. C. A., without changing the frequency of oscillations, and further under inductive plate loading, the G. O. V. A. cannot be brought into phase with the tube voltage or tube current under any conditions of operation.

Also the natural period of the tank circuit is not directly involved in the frequency of operation of such an oscillator, since the natural period is the period at which the tank circuit operates when freej that is when itis not being driven, whereas in oscillator operation the tank circuit is driven at the period which provides the necessary G. O. V. A., and there is no necessary direct relation between the two periods. The Q of a tank circuit influences the free period thereof `as a separated consideration, whereas the Q of a tank circuitas it functions in an operative element in an oscillator inuences the difference between the free period and the driven .period ,of the tank circuit of such an oscillator.

Referring to Figs. 5 and 8, the higher theQ of the tank circuit, the clcser will be the operating frequency to the resonant frequency of the tank circuit. However good stable operation can be obtained with an operating Q (Q of tank circuit takenunder operating conditions) as low as 5, and probably lower. Operating Q is not to be confused with'Q taken at resonance frequency of the tank circuit, as this condition often indicates a higher value. The operating Q of the tank circuit in Fig. 10, must also includethe grid losses.

Attention is directed to the factthat in anoscillator, having an inductive plate load. the-tube reaction is in series .with the load inductance and therefore the load inductance functions characteristically toi reduce harmonics in the tube circuit current, any harmonics present appearing inv the voltage. This accounts for thefact that such oscillators usually producea tube current having a good sine wave-form, and a tube voltage having a considerable harmonic content in the vwave-form thereof, mostly even harmonics introduced by the tube characteristics. It is to be borne in mind that-in any circuit having the property of producing harmonics,Y` they -canhbe reduced or leliminated in the currentl or in 4the voltage of the circuit, but not in both. The inductively loaded oscillator reduces them inthe current, and their reduction in capacitively load oscillators will be treated later on herein.

Now concerning this disclosed oscillator theory as applied to capacitively loaded oscillators. the amplifier analysis as ,given by Morecroft above, cannot 4be validly applied, because his analysis.

V13 as applied to amplifiers, as understood by the applicant is incomplete, faulty and incorrect.

Morecroft does not illustrate the amplifier circuit from which his oscillograms of an amplifier having a capacitive load were taken. His vector diagram of his capacitively loaded amplifier, is a diagram of a load circuit having only a resistance in series with a capacitance, corresponding with his vector diagram of his inductively loaded amplifier circuit, in which he does have only resistance in series with inductance. But the plate circuit of an amplifier (and of an oscillator as well) is a direct current circuit as well as an alternating current, and circuits having,V

fler operation having a capacitive load and zero resistance could not have been made with the two circuit elements in series.

The applicant discloses hereunder an amplifier circuit having a capacitive load, and a correct vector diagram therefor and how the analysis of these may be'applied to oscillator operation.

in the resistance branch and the current in the capacitance branch, of the parallel load of the' tube. When dealing only with the alternating current components of the currents and voltages as heretofore premised herein, the same is tre for the alternating current components thereon' Referring to Fig. l2, switch |20| is closed upon f contact |202, and the circuit operates as an amplifier from a grid voltage source |204. Referring to Fig. 13, the angle of the load reactioniis seen to be the reciprocal of the value for the series circuit, given by Morecroft above, and the z.

angles between the grid voltage and tube voltage and tube current are very different values from those given by him for capacitive loads .(see his page 579). A

Referring to Fig. 12, the switch |20|, is closed upon contact |203, under which condition theresultant circuit operates as an oscillator andthe tank circuit |205, provides the G. O. V. Auasfjdetermined by the vector diagram of Fig. 13. g

The corresponding circuits of Fig. 3 operate similarly and an illustrative graph of the G. O. V. A. with the corresponding tube voltage angles are shown in Fig. 6. Yj

Referring to'Fig. l2, it will be seen, thatuthe tank circuit current has the phase of the tube voltage Ep, andreferring to Fig. 13, that thetank circuit voltage must l'eadthe tank circuit crrent to provide the G.' O. V. A., necessary for oscillator operation. This explains the experimentally discovered operation of Fig.` 3, as illustrated in the graph of Fig. 6, in that the oscillator operates at a frequency lower than'the resonant frequency of the tank circuit.

It will be seen in ,capacitively loaded oscillators that the tube reaction is partly in series withlthe ytube characteristics.

14 load condenser and therefore the load condenser functions characteristically to induce harmonics inthe tube circuit to appear in the tubel current, and therefore reduce them in the tube voltage. This accounts for the fact that such oscillators usually produce a tube voltage having a good sine wave-form, and a tube current having consider'able harmonic content in the wave-form thereof, mostly even harmonics introduced by the If the oscillator of Fig. l2 has switch |206 opened, all the angles of the diagram of Fig. 13 disappear and all voltages and current are thrown into phase. This explains the operation of the oscillator of Fig. 4, as experimentally determined by the graphs shown in Fig. '7.

With the experimental disclosures and explanatory operating theory developed above, and with the aid of the measurement directions given herein, the applicant now shows how his discoveries can be embodied in some other new and useful devices. l

Fig. 14 shows an oscillator in which the tank circuit has a fixed LC, and it is caused to operate over a wide range of different frequencies, by in'- terposing, in the oscillator feed-back circuit, a feed-back phase shifting means.

The circuit enclosed within the dotted area J, is an oscillator of the F type shown in Fig. 5, when switches 40| and |402 are closed respectively upon contacts |403 and |404. As a specific` example, switches |40| and |402 are closed, the tank circuit |405 and the feed-back resistors are adjusted to give stable operation at 600 cycles. The G. O. V. A. is determined by the method illustrated in Fig. 9. This oscillatorwith the same feed-back current and the same G. O. V. A. will oscillate at 600 cycles regardless of the source of the feed-back current. Also this oscillator operates at other frequencies for other phase positions of the G. O. V. A. as shown experimentally in Fig. 8 and analytically in Fig. 1l, so that interposing a phase shifting in the feed-back circuit of the oscillator enclosed within the dotted area J, will result in an adjustable frequency oscillator having a fixed LC in the tank circuit.

Referring back to Fig. 14, the circuit enclosed withinthe dotted area K, is an amplifier having a resistance-capacitance output and a phase shifting means in the input circuit; In'the form illustrated this phase shifting device is a parallel inductance-capacitance circuit having an adjustable LC, but any other suitable phase shifting means may be used. j

Switches |40| and |402, are opened and switches |409 and |4|0 are closed respectively upon contacts |4|| and |4|2. The condenser I4| 3 is selected so that at its middle point, the phase angle of the output of the amplifier is eX- actly in phase with the oscillator tube voltage, then the'feed-back angle to the oscillator tank circuit through the amplifier, is the same as when switches |40| and |402 were closed, hence the oscillator operates at 600 cycles, as it did originally. Adjustment of 'condenser |4|3, causes the oscillator to run faster or slower, depending upon whether the G. O, V. A. is increased or decreased,r as illustrated in Fig. 8. f

vIt has been stated (F. lTermanjResistance stabilized oscillators, Electronics, July 1933) that in oscillators which he describes, which are type Vthis type can be constructed with more inductanoe than resistance nin the .feed-back circuit, :and .op-

v-erate with 'lgood stability.

Fig. l shows an oscillator embodying the vap- `plicants discovery and linvention which'has a large inductance in the feed-back circuit. The `oscillator is constructed ifor 600 cycle operation as follows:

' vFirst the circuit has the .feed-back circuit open fand the Vvtank circuit removed. The G. O. V. A. for the L. I. C. A. determined by the method Ishown in Fig. 9. The tank circuit is separately adjusted to kresonance by the method shown in Fig. .1. The tank circuit is then properly connected in the 4grid circuit as shown in Fig., 15.

The amount of feed-back current is calculated from the tubes and circuit employed, then the inductanoe of `reactors |502 `and |503, and the resistance of resistors Y|504 and |505,1are cal- A`culated from vthe well known` vector Lrelation shown'in Fig. 16, to provide the correct G. O. V. A. vand when the feed-back .circuit is connected up as shown, the oscillator operates .at 600 cycles.

.or`very close thereto, depending upon the accuracy of vthe work performed. Thus is produced aninductively loaded oscillator, with a tank cur- ...rent 'operating at resonance, and Withglargely inductive impedances inthe feed-back circuit.

When properly constructed, the oscillator shows good frequency stability.

With the teaching given in connection with Fig. further disclosure can be made with reference to Fig. 14. `In the disclosure given for the construction of oscillators in accordance with Fig. 14,.directions were given to set the tank circuit |405, for proper noperation at the given frequency Withifeed-.backresistors |406 and |401 incircuit. Attention Wasdirected to the fact that the circuit enclosed within dotted varea K, constitutes aphase shifting circuit replacing these feed-back resistors .under conditions of .adjustable frequency operation.

With reference to Fig. r15, it was shown that .the Ytank circuit |50| can be set to resonance `at the frequency vof operation, when the feedback circuit has a proper phase shifting means inserted into it. Since circuit enclosedin dotted .area K isa vphase shifting means, the tank circuit |405 of Fig. 14 can be set also to the resonance at the frequency of operation, and the proper phase shift `of the feed-back circuit can -be obtained by adjustment .of condenser 4'|3, ,or ofthe LC ofthe circuit including ,condenser |413. Further, since .it isseen that the G. vO.'V. A. vrequired for any L. I. C. A. is a `Vvector result of the phase .angle of the tank circuit |405,v and of the .relative phase angle of the feed-back circuit, the LC of the tank circuit |405 can have a [wide Vrangeof/different valuesfor the same oscillator operating frequency, ifmeansgisemployed V1in the feed-back-circuit to cause'th'e proper phase angle .feed-back current to the tank circuit for the operating frequency desired.

.The applicant will now. give direction`s 'for the embodiment of his discovery and invention in oscillators bythe employement ofpolyphase feedlback circuits. Referring toiFigl'l, ,there is shown lfeed-back circuitcomprising condensers |102 .and

|103, in series with resistors |104 and |105. This --circuit supplies feed-backlcurrent in phase with alle mure. inthe .circuit .L1-n n.. :is to were at resonance, 'the 'etank circuit 1f eedback :current must :be in phase `with the G. O. V. A. .This .is

'-Iaccomplished' in -this embodiment ,by a .direct feed-back circuit, comprising, in Fig. 17, resistors.

feed-.backlcurrent and the proper G. O. V. A. to lcause the desired frequency of operation of the oscillator.

' While `only two phases 4 of feed-back .current :is shown, it will .be appreciated by -those skilled in the art `of polyphase currents, that any number `of properly chosen :phases V.may be vemployed to produce a single phasecurrent of .a desired phase .relation to the .reference phase of thecircuit In view .of the foregoing disclosures, Vit will `be seen that the LC of the tank circuit |1.0|, .does not have to be set to ,the resonant operating frequency Aof the oscillator, but may .be set at different other LC values for the same operating frequency of the oscillator, if the polyphase feedback `circuits are adjusted .to provide the required G. O. V. A.,for the load employed at .the frequency vof operation.

A further methodand structure, forsupplying polyphase feed-back .circuits to the tank circuit .as compared to the totalI load voltage, andthe secondary there of is generally shunted by a condenser |906, and .has the feed-back Aresistors |901 and |908, asa load therefor. This current ,transformer is constructed and vadjusted such that the feed-back load current is n'phase with, but not Ynecessarily so,.` the alternating current component of the plate load current. The phase lposition of this second ,phase of feed-back curvrrent yis identied in the vector vdiagram of Fig.

20, and its vector composition to give the required G. O. V. A., for .thedesired operation is conventionally analyzed in theFig 20.

It will be appreciated by those skilledr'in the ,f art that, in general, the operation of the tank circuit at its resonance frequency, provides a more Y.economical use ofthefLC of that circuit and and this isanimportant item'snce the tank circuit usually constituted an vexpensive element in the cost of oscillators.

All of the circuits disclosed herein vhave .been actually constructed vand operated as described herein at 60,0 cycles as afrequency center point and thereforeany differences of opinion vas to alternate explanations of ,the operation .does not effect .the validityl of vthe disclosures.

f .Haiing n taught herein,- the nature of vmy .dis-

Acovery andinvention, other andv further embodiments will be obvious tofthose skilled in the art.

What I 4claim is:V 1. In-.an electrical oscillation system including an electron discharge tubehaving a plate, a Acontrolgrid anda-cathode; QalOadcircuit, ei-discharge control .circuit .including gsaid -gridf-saidlcathode l? and a tank circuit, and a feed-back circuit from said load circuit to said tank circuit including electrical-circuit phase-shifting means varying the operating frequency of said system.

2. In an electrical oscillation system including an electron discharge tube having a plate, a control grid and a cathode; a load circuit including a reactive member, a cathode-grid circuit including a tank circuit set to an operating voltagecurrent phase-difference having an angular value different from the grid operating voltage angle required for the load imposed current angle of said system, and a feed-back circuit from said load circuit to said tank circuit including phase shifting circuit means compensating for said different Value.

3. In an electrical oscillation system including an electron discharge tube having a plate, a control grid and a cathode; a load circuit, a cathodegrid circuit including a tank circuit, and a feedback circuit from said load circuit to said tank circuit including electronic-tube phase-shifting means varying the operating frequency of said system by angular variation of the grid operating voltage angle.

4. In an electrical oscillation system including an electron discharge tube having a plate, a control grid and a cathode; a load circuit, a cathodegrid circuit including a tank circuit, and a feedback circuit from said load circuit to said tank circuit including an electronic tube amplier having a variable LC tank circuit in the cathodegrid circuit thereof.

18 5. In an electrical oscillation system including an electron discharge tube having a plate, av control grid and a cathode; a load circuit including an inductive reactor, a cathode-grid circuit including a tank circuit set to resonance at the operating frequency of said system, and a feed-back circuit from said load circuit to said tank circuit including phase shifting circuit means causing current fed to said tank circuit to ow at a phase angle equal to the grid operating voltage angle required for said operating frequency of said system.

MONTFORD MORRISON.

REFERENCES CITED The following references are of record in the le of this patent:

UNITED STATES PATENTS Number Name Date Re.21,807 Lindenblad May 20, 1941 2,076,264 Chireix et al Apr. 6, 1937 2,162,470 Kautter June 13, 1939 2,346,331 Roberts Apr. 11, 1944 2,389,025 Campbell Nov. 13, 1945 2,415,868 Clark Feb. 18, 1947 2,421,725 Stewart June 3. 1947 2,445,811 Varian July 27, 1948 2,482,766 Hansen et al. Sept. 27, 1949 2,506,329 Ames, Jr May 2, 1950