US2441567A - Variable frequency oscillator - Google Patents

Description

May 18, 1948. s. DARLINGTON 2,441,567

' VARIABLE FREQUENCY OSCILLATOR Fil ed Feb. 13, 1 945 "r Sheets-Sheet 1 Jul/Euro? SDARL/NGTQN ATTORNEY May 18, 1948.

S. DARLINGTON VARIABLE FREQUENCY OSCILLATOR Filed Feb 15, 1945 v '7 Sheets-Sheet 2 I If? R N/ R: (sue/1,125?

INVENTOR .ijDARL/NG TON A TTORNE Y May 18, 1948. s. DARLINGTON VARIABLE FREQUENCY OSCILLATOR 7 Sheets-Sheet s Filed Feb. 15, 1945 AMA S TAB/LIZER T w w lNl/ENTOR S DARL/GTON Lu (5M ATTORNEVV y 18, s. DARLINGTON I VARIABLE FREQUENCY OSCILLATOR.

Filed Feb. 13, 1 945, 7 Sheets-Sheet 4 I ans/1.12:4

FIG. /0'

.F/al/ INVENTOR S. DARL N 6 TON 8V AMPL/f' If R ATTORNEY May 18, 1948. s. DARLINGTON VARIABLE FREQUENCY OSCILLATOR Filed Feb. 15, 1945 7 Sheets-Sheet 5 FIG. /2

1 STAB/LIZER MPL IFIER 0MP! IFIER amen/25R llL 'INI/ENTOR S. DARL/NGTON A TTORNEY Patented May 18, 1948 UNITED STATES PATENT OFFICE VARIABLE FREQUENCY OSCILLATOR Sidney Darlington, New York, N. Y., assignor to Bell Telephone Laboratories, Incorporated, New York, N. Y., a corporation of New York Application February 13, 1945, Serial No. 577,622

19 Claims. (Cl. 250-36) plitude level of oscillations over the operating range or band of variable frequencies.

One of the objects of this invention is to provide improvements in varying and modulating the frequency and in controlling the amplitude quency modulator may be designed for a high of oscillations in variable frequency oscillation generators.

Another object of this invention is to improve the operation of variable frequency oscillators in respect to frequency and amplitude stability.

In accordance with one feature of this invention, the frequency of oscillations may be varied or modulated with with ease by means of one or more variable loss or variable gain control devices, such as variable potentiometers, resistances, inductances, capacitances, electronic vacuum tubes and similar control devices capable of introducing variable loss or attenuation into a transmission path.

In accordance with another feature of this invention, the amplitude level of oscillations may be automatically stabilized over or throughout a range of operating frequencies by means of nonlinear auxiliary feedback systems which maybe of the variable loss type.

In accordance with this invention, a separate variable resistance or variable potentiometer constituting a variable loss device, or a variable voltage ratio device which may provide a variable loss may be used to control the frequency of oscillations of the circuit, and may be used in place of the usual variable impedance elements such as the resistance, inductance or capacitance elements comprising the usual frequency determining network of the oscillator or amplifier tube system. Such variable loss control of the frequency may be conveniently used to provide a highly stable and continuously variable oscillator for generating variable frequencies which may be in a band below 100 cycles per second for example, or at any higher frequency. The required variable loss may be conveniently obtained by means of a suitable potentiometer such as a precision tapered potentiometer for exampie; or the variable loss may be obtained by electronic means or by attenuators of variable resistance. be used to provide signal-modulated frequency modulation, where, by using a suitable vacuum tube or varistor modulator to introduce the frequency controlling variable loss, a sir'np1efre- Loss control of the frequency also may percentage modulation.

The variable loss principle may be applied. to control not only the frequency of oscillations but also the level amplitude stabilization of oscil lations, a level sensitive variable loss device replacing the conventional level sensitive resistance such as, for example, a tungsten or carbon filament lamp, a thermistor, or a gas-filled tube. A simple and effective means of providing level amplitude stabilization may be obtained with an ordinary amplifier vacuum tube using a special connection.

As applied to variable frequency loss-controlled oscillators, the invention may comprise a suitable amplifier or source of gain, a main feedback path having two parallel transmission paths each contributing to the total feedback, networks in the two feedback transmission paths, means for introducing flat loss or gain into one or both of the two feedback transmission paths, and a stabilizing element. In the two feedback transmission paths, the networks may be such that the corresponding contributions to the total feedback have the following properties: (a) The two contributions to the feedback corresponding to the two parallel transmission paths differ in phase by an amount which is substantially degrees independent of frequency; (b) the amplitudes of said twocontributions vary differently with frequency, the usual difference being a change in the relative levels of the contributions at a rate which may be substantially equal to 12 decibels per octave of frequency change, the above conditions applying to the relative phases and levels of the two contributions to the feedback and not to the actual phase and level of the total feedback; and (c) the two amplitudes are equal at a definite frequency which is substantially the frequency of oscillation. The means for introducing flat loss or gain into one or both of the two parallel transmission paths referred to does so in such a way that the difference between the amplitude levels of the corresponding contributions to the-feedback can be changed at will by amounts which are substantially independent of frequency if levels are measured on a logarithmic or decibel basis, and in such a way as to leave the phases of the two contributions substantially unchanged, the variation in fiat loss controlling the variation in frequency of oscillation of the oscillator by varying the frequency at which the amplitudes of the two contributions to the feedback are equal.

Thestabilizing element referred to may be a level sensitive variable element which may either (a) produce an effective change in one of the networks referred to, or (1)) change the feedback in a subsidiary feedback path. The stabilizing element may comprise a subsidiary or auxiliary feedback path yielding a level controlled variable loss for providing the level amplitude stabilization of the oscillations. The source of gain referred to may be provided with stabilizing elements to meet conditions of high feedback amplifier stability.

The loss-controlled variablefrequency oscill-f lator circuits particularly illustrated in the fig..

ures herein are specific embodiments of a 1110126. general class of bridge type feedback oscillator circuits, and feature the use of variable losses or variable voltage ratios for control or modulationof the frequency, as distinguished from .the usual variable or adjustable bridge arms in the feedback p-circuit bridge ordinarily used heretofore for control of the frequency. The circuits disclosed in this specification also feature the use of "variable losses or variable voltage ratios for level amplitude stabilization of the output voltage of the oscillator circuit. Control of the frequency bym-ea'ns ofvariable'loss devices. and level amplitude stabilization by means of variable loss devices are two independent features, either of "which "may be used with or without the other.

While theuse of variable 'loss or 'variablevoltage ratio network devices to control the frequency of oscillation is disclosed; in connection with'par- .ticular bridge type oscillators, it will be understood that such network devices may be used for other purposes where itmay be desired to introduce'hig'h loss to a frequency which can bevaried at will by merely changing the loss or voltage ratio by means of variable voltage ratio device such as a variable voltage'transformer', a potentiometenor an amplitude modulatorforexample. Such a network may also be used as a variable frequency wave "trap, or in the c-circuit of a feedback'ampIifi-enor'in current analyzers as 'a means 'forproviding a variable narrow band filter;

For .a clearer understanding of the nature of this invention and the additional advantages, features and objectsthereof, referenceis made to the following description taken in connection with the accompanyin drawings, in which like refer-- ence characters represent like or similar parts and in :which:

Fig. .1 'is-a circuit diagram illustrating a variable frequency oscillationgenerator of the bridge stabilized "type employing. a variable loss type of control device in the form. of a variable resistance potentiometer for varying the frequency of oscillations, .and' also employing .a, form .of variable. loss .type .of auxiliary feedback system for automatically stabilizing the amplitude level of oscillations.

Figs. 2 and Bare idealized or generalized circuit diagrams illustrating the principle .of operation of variable loss and voltage ratio devices for controlling thefrequency and the amplitude level of oscillations in variable frequency oscill'ators.

Figs. 4 to, '7 are circuit diagrams illustrating modificationsinvariable potentiometers and variableresistances for controlling thefrequency of oscillations.

:Fi'g, v8::is a circuit diagram similarto the. circuit of 1, but illustrating a modification in the potentiometer controls, and in the amplitude stabilizer control.

Fig.9'is a circuit diagram illustrating a modification comprising a variometer for controlling the frequency of oscillations. 7

Figs. 10 to 15 are circuit diagrams illustrating modifications comprising signal-modulated vac uum tubes for controlling the frequency of oscillations.

Eigslfi and 17 are circuit diagrams illustrating modifications in networks that may be used.

Figs. 18 to 20 are generalized circuit diagrams -.illustrating modifications in the over-all circuit system, the system in Fig. 18 corresponding to the .moreispeui'ficfiii-icuit diagrams of Figs. 1 to 17.

Referring to the drawing, Fig. 1 is a circuit diagram;of;a. variableefrequency bridge type oscillacillator'may comprise an amplifier or source of gain A'consisting-of a series of three high gain amplifying vacuum tubes Ti, T2 and T3 provided with suitable "interstage coupling means therebetween. A main-feedback circuit 3 couples the outputl-B of the third vacuum tube Tiiwith the input of the-first-vacumntube Ti and comprises a coupling condenserC'l, and a pair of parallelconnectedtransmission paths G "and 2 which include networks NI and N2, respectively, and variable resistancepotentiometers PI and P2, respectively, both of which have a resistance compo- -nent thereof connected in series circuit relation-withthe-networks Ni and N2, respectively. The variable resistance potentiometer Pi is used only as a-fi'equencyttrimmer, and its loss is thus held within-narrow limits bythe use of series-resistances R1 and R32, "and, a shunting resistance RM. The-main tuning resistance potenti ometer P2 and thetrimmer'potentiometer P! are connected with the main feedback circuit 3. The other end'softhe potentiometer P1 and P2 are connected withth-e ground I. The several ground connections I ,-which are designated in the drawing'by'the conventional 'symbol fora ground connection, maybe-connected together inpractice by means'of'any suitable wire connectors. The output oscillations may "be taken from the output 't'ermiIIaHGa-nd' groundterminal' i orfrom points connected therewith.

In practice, it'is frequently desirable to use an amplifier in'which the phase between the input and opposite circuits is reversed. Such an amplifier A'maybe obtained by employing an odd number of amplifying tube stages, such as the three vacuum tube stages illustrated by the direct-coupled three=stage amplifier tubes T1, T2 and T3 fin Fig; 1'. 'Thetandem connected linear amplifier tubes'il, T2 and T3 of the linear amplifier A, and also the amplitude stabilizer tube T4, maybe foriexamplehigh gain vacuum tubes of the pentode type "having an indirectly heated cathode "I, :acathode lreater filament 8, an input or control grid 9; a screen grid 1-0, a suppressor grid 'il connected'withthe cathode 7, a plate or anode electrode |2,;and amet'al'envelope or tube Whichmay'be connected to ground as illustrated at I3 in Fig. l. The cathode heater filaments 8 of the tubes TI, T2, T3 and T4 may be connected in parallel and supplied with heating current in a known manner from any suitable voltage supply source (not shown).

The first tube TI of the amplifier A may have its cathode electrode 1 and its grid electrode 9 connected to ground I through resistances R8 and RT respectively, and may have its screen grid electrode Ill connected to ground I through a resistance R9. Suitable potentials for the screen grid electrode I and for the plate electrode I2 of the tube TI may be supplied through resistances RID and RI I, respectively, from the positive terminal of a suitable power supply source I5. The interstage coupling between the vacuum tubes Ti and T2 may comprise a resistance RIZ with a grounded series-connected condenser C and a coupling condenser C6. The second vacuum tube T2 of the amplifier A may be provided with a grid resistance RI3, a cathode resistance RM and a screen grid resistance RI5. Suitable potentials for the screen grid electrode It and for the plate electrode I2 of the second amplifier tube T2 may be supplied through resistances RIG and RH, respectively, from the positive termina1 of the power supply source I5. The interstage coupling between the output of the second amplifier tube T2 and the input of the third amplifier tube T3 may include a condenser C'I, a parallel-connected condenser C8 and resistance RI8, a resistance RI9, and a grounded condenser CID, as illustrated in Fig. l. The third tube T3 of the linear amplifier A may be provided with a grid resistance R20 and a cathode resistance R21. Suitable potentials for the grids It and I! and the plate electrode I2 of the tube T3 may be supplied through resistance R22 from the positive terminal of the power supply source I5.

The network N2 as illustrated by the block diaram labeled N2 in Fig. 1, is constructed in the form of a T-network consisting of two series connected resistances R2 and R3 and a single shunt condenser C2 which is conected at one end thereof to the ground I. The network NI as illustrated by the block diagram labeled NI in Fig. 1 is constructed in the form of a T-network consisting of two series condensers C3 and C4 and a. shunt resistance R which is composed of two series resistances R4 and R5, the latter being connected at one end thereof to the ground I. The networks Ni and N2 taken together comprise a double-T or parallel-T network system and taken with the variable resistance potentiometers PI and P2 form two parallel paths 4 and 2, respectively, in the main feedback circuit 3 of the linear amplifier A, the main feedback circuit 3 being connected between the output circuit I6 of the third amplifier tube T3 and the input of the first amplifier tube TI of the amplifier A. The output circuits of the parallel-connected paths 4 and 2 and of the networks N! and N2 are connected with the input of the first amplifier tube TI, and therebetween the resistance R1 is connected to ground I at one end thereof.

The bridge type feedback circuit networks N! and N2 may be practically any type of network system which provides a substantially constant ISO-degree phase difierence and different amplitude characteristics for varying frequency values. Since there are a large number of bridge circuits which provide such transmission char acteristics, the network system NI and N2 may consist of any of a large number of possible bridge circuit arrangements. The particular parallel-T lation depends intimately upon the resistances.

that are used in the networks NI and N2 and hence they are constructed as precision units where precision is desired. At low frequencies, such as, for example, frequencies below cycles per second, the resistance units of the networks Ni and N2 may be made large in order to avoid the use of unduly large precision capacitance devices. Since continuously variable precision resistance units are difficult to realize in practice in large sizes, the large precision resistances of the networks NI and N2 may be fixed or non variable resistances, the frequency of oscillation being adjusted or varied by one of the variable potentiometers PI or P2 which may be made to be of relatively low resistance values and of relatively small size. In the circuit of Fig. l, the variable resistance potentiometer P2 represents a variable loss device which is used to furnish variable control over the frequency of the system. Since high precision in frequency imposes high precision on the resistances and other component parts of the networks NI and N2, where a variable frequency is required, it is often difficult or impractical to obtain high precision in frequency control by means of adjustments in the large precision resistances and other component parts of the networks NI and N2 themselves. For this reason, and for the same precision in frequency, the use of the potentiometer type of frequency control P2 is more easily realized in practice. The impedance of the potentiometer P2 being of relative low value is more easily contructed in adjustable form. Moreover, the potentiometer voltage ratios of the adjustable potentiometer P2 can be normally held to close limits since the voltage ratio thereof depends upon the relative resistances rather than upon the absolute values of resistance. Accordingly, changes in the absolute values of resistance of the potentiometer P2, as caused by changes in temperature, for example, do not change the voltage ratio of the potentiometer P2 except to the extent that such changes in absolute resistance may not be uniform along the potentiometer P2.

While the potentiometer P2 presents a finite impedance to the load connected to the brush 5 thereof, its effect on the frequency stability of the system may be minimized by making the impedance of the resistance potentiometer P2 reasonably small relative to the impedance value of the resistance R2, to which .the brush 5 of the potentiometer P2 is connected. The effect is to add resistance to the resistance R2 by the par-' allel combination of the segments of the potentiometer P2 above and below the brush 5 thereof. To the extent that the effective addition of such resistance to the resistance R2 is constant, it may be compensated for by a reduction in the resistance of R2 itself. The remaining variations in the equivalent resistance of 3.2 as a result of the effect of the resistance of the potentiometer P2 thereon may be taken care of in part by means of the amplitude stabilizer system such as, for example, by the system comprising the auxiliary feedback circuit I3 including the stabilizer S with the tube T4 of Fig.1, andinpart amuse? by propercalibration of the voltage ratio provided bythe resistance arms of the potentiometer P2. If desired, the taper of the resistance values of the potentiometer P2 may be used to obtain a linear variation-of frequency with respect to the shaft position of the potentiometer P2. It will be understood, however, that the adverse effect of. thevariable resistance of the potenti-ometer P2 on the over-all stability of the oscillator system is notlarge when the impedance of the potentiometer P2 is reasonably small relative to the impedance value of resistance R2 of network N2.

As illustrated inFig; l, the tuning potentiometer P2 has a: resistance composed of a middle portion connected with the adjustable wiper 5, and two end series resistance portions ii and Ma. This construction, when used, insures that the potern tiometer P2 cannot be set-to a value corresponding to a frequency outside the range for which the circuit is designed. Ihis is a desirable fea ture because the average as of the system ordinarily must or should be held Within plus or minus 90 degreesof anegative real over the entire' working range of frequencies. In providing the potentiometer P2 with suitable non adjustable end: resistance portions [4 and I iaas illustrated schematically in Fig. l, the possibility of setting the potentiometer P2 outside the working range of frequencies maybe prevented.

Asillustrated in Fig. 1, an auxiliary feedback circuit I 8 may be provided between the plate outputcircuit l2 of the second amplifier tube T2 and the input circuit of the first amplifier tube T! of the amplifier A in order to stabilize the amplitude level of oscillations generated. As shown in Fig. 1,. the anode E2 of the second amplifier tube T2 of the linear amplifier A is connected through a blocking condenser C9 and a" resistance R24- with the input grid 9 of the pentode T4, and the output or plate electrode I2 of the tube T4 is connected through resistanceRfi'l, resistance and condenser CA with the input grid electrode!) of the first amplifier tube Tl of the amplifier A. The input grid electrode 9' of the tube T4- is connected through resistances R23, R24 and- R25 with the positive (4*) terminal of the supply source iii. A resistance R26 is connected between ground i and a point intermediate the resistances R23 and The middle grid electrode it of the tube This connected through a resistance R28 to ground I, and through a resistance R29 to the positive" terminal of the voltage supply source i5.

Potential for the plate electrode 82 of thetube T4 is'supplied' through a resistance Rial which is connected with the positive terminal of the the latter being suppli dth resistames R25,

and R231 Whenever the'instantaneous value of the voltage at'the'junction I9 is positive, the'grid electrode 9 of the tube T-t draws current and produces a substantial voltage drop across the input resistance- R24. Because: the grid" current andijalsoi the voltagedrop across the resistance R24 increases very rapidly with increasing voltage multiplied by the ratio'of R24 to R2l+R25+R23.

ization of the amplitude of oscillations.

When the amplitude of the signal voltage at the junction l9 begins to exceed this Value of the bias voltage, the voltage of the grid 9 of the stabilizer tube T t swings negative over a portion of each cycle. While such grid voltage is negative, the grid 9 of the tube T4 draws no appreciable current and there is no appreciable corresponding drop across the grid resistance R2 3. Hence, only portions of each cycle of the combination signal and bias voltage at the junction l9 are transmitted and amplified by the tube T4. The

output of the tube T4! is fed into the network N l oi the oscillator through the resistance R4 and opposes the oscillating signal voltage therein thus giving the desired level amplitude stabilization at a level at which the amplitude of the signal voltage at the junction i9 is slightly greater than the bias voltage at the junction l9.

The positive grid vacuum tube T4 of Fig. l accordingly functions to supply pulses of alternatin voltage inopposition to the signal voltage in the network Ni thereby to provide level stabil- At the lower amplitude levels of the output voltage of the amplifier A, the vacuum tube 'I l' due to the steady positive grid bias applied to the grid electrode t thereof from the supply source l5 effectively blocks the oscillations reaching it from the output of the second amplifier tube T2 over the circuit l8 and continues to do so until the amplitude of the oscillations reaches a critical level in amplitude. As the level increases further in amplitude, the tube T l begins to transmit the oscillations to an extent that increases as the level of the amplitude of oscillations increases above the critical value mentioned. Accordingly, the effect on the circuit is of the tube T 3 is to prevent transmission of oscillations at the low levels of amplitude and to transmit oscillations with the increasingly higher values of amplitude applied thereto:

The tube T is fed from the output of the second amplifier tube T2 or from the next to the last tube of the gain circuit A, rather than fromthe output it of the last amplifier tube T3 of the gain circuit A, order to balance against the reversal through the stabilizer tube T l, the similar'reversal in direction of signal that occurs in passing through the last tube T3 of the gain circuit A. A vacuum tube, such as the tubes T3 and T l, reverses the direction of the signal which it transmits. Hence the reversal through the stabilizer tube T4 is balanced by the similar reversal through the last tube T3 of the gain circuit A. The tube T l fixes the amplitude level of oscillations at a value whichis substantially independent of the frequency of oscillations, and accordingly a substantially constant oscillator output voltage level is obtained from the output I6 of the last amplifier tube T3 of the gain circuit A provided the tube T3 has a substantially constant gain over the working range of variable operating frequencies.

frequency due to changes in the setting of the frequency controlling tuning potentiometer P2 even though these changes of frequency setting of the potentiometer P2 call for substantial changes through the tube T4. It is a sensitive system giv- .ing a sharp rise in transmission at a definite amplitude level, and the amplifier tube T4 may be of the same. type as the amplifier tubes Tl, T2

and T3 of the circuit A of the oscillator. While considerable harmonics may be generated relative to the fundamental component transmitted by the tube T4, the amplitude of the harmonics and also of the fundamental itself that enters the circuit over the circuit 18 is relatively small as compared with the amplitude of'the oscillator signal voltage that is fed back over the main feedbackcircuit 3 from the output iii of the last cut-offs, is comparatively simple and operates well even at low levels of voltage applied thereto. This results from the feedback voltage from the output circuit of the oscillator being fed back through the vacuum tube T4 which has a substantial plus bias on the control grid electrode 9 thereof. Where the positive or plus (-1-) bias and the signal are fed to the grid 9 of the tube T4 through the high resistances R23, R24 and R25, the low impedance of the grid 9 at plus voltages operates to prevent any substantial "signal voltage being fed back until the magnitude of the signal voltage becomes large enough to swing the grid 9 negative, above which critical level there will be a sudden increase in transmission level through the vacuum tube T4. Thus the effective stabilizing voltage e is very small in magnitude until the critical level referred to is reached and thereafter it rises very rapidly with further increasing levels of oscillating signal voltage applied to the tube T4. A substantial value of positive voltage bias'may be applied from the supply source [5 to the high voltage side of the grid resistances R23, R24 and R25, without getting excessive positive voltages on the grid 9 itself of the tube T4 as a result of the IR voltage drop that occurs in such grid resistances when the grid 9 of the tube T4 draws current. i a

It will be noted that since the stabilizing tube T i gives a phase reversal, it is fed from the main amplifier A from a different interstage point, and preferably from the first interstage back of the final output tube T3, as illustrated in Fig. 1. It will be noted that variations in the gain a of the high gain linear amplifier A are absorbed by the :non-linear stabilizer circuit 18 comprising the positive grid tube T4without substantially any shifting of the frequency of oscillation provided the phaseshift of the bridge network system NI and N2 is relatively small. Provided that the phase shift of the bridge network Nland N2 referred to can be kept small enough, a rather small value of gain u may be used if the precision required is not too high.

As an illustrative example, a particular oscillator constructed in accordance with the circuit of Fig. 1 and having a variable frequency ranging from 30 to cycles per second was provided with component resistances and condensers having substantially the following'values. In the particular example mentioned, the values for the component resistances expressed in ohms were about: R|=65,000, RZ=200,000, R3=206,000, R4:85,000, R5=12,000, Rl:1,000,000, R8=2,000, R9=1'5,000, Rl0=100,000, BIL-50,000, RI2= 25,000, Ri3:l,000,000, Rl4=2,000, Rl5=15,000, Rl6 =100,000, R" 250,000, Rlfi 10,000,000, Ri9=500,'000, R20=500,000, R2l=800, R22: 10,000, R23=200,000, R24=1,000,000, R25=1,000,- 000, R26=30,000, R21=l,000,000, R28'=15,000, R29=100,000, R30=1,000,000, R3l:500, R32: 720, P|=2,000, and P2=10,000 ohms total, the values of the end resistances being 750 ohms each and the value of the middle resistance covered by the wiper 5 being 8500 ohms. The values for the component condensers expressed in micro- .farads were about: Ci:1.0, 02:.016, C3=.008,

042.008, C5:.005, 06:2.0, Cl=1.0, C8=.012, 09:.1, C|9=.005; and the tubes Tl, T2, T3 and T4 were standard 12SJ7 pentodes having their cathode filaments energized from a cathode heating source of about 12.6 volts and provided with a power supply source l5 of about +250 volts direct current potential.

The bridge networks NI and N2 of Fig. 1 are represented as being composed of two threeterminal networks N! and N2 connected in parallel. The short-circuit transfer admittances of the networks N2 and Ni may be represented by Y'I'Z and Y"l2, respectively. The short-circuit transfer admittance Yl2 is a function of frequency associated with four terminal networks in general, is frequentlyused in general network theory, and is defined as the ratio of the current I in a short-circuit across the output terminals of the network to the voltage E across the input terminals which produces it. In accordance with a well-known theorem of general net- 'work theory, at any frequency of infinite loss such as 5:0, the transfer admittance of the two parallel-connected networks will satisfy the condition:

' Y'12+Y"'12=o (1) In terms of the reference notations of Fig. 1,

per second.

Assuming that the values of the component resistances and capacitances of the networks NI and N2 are made to satisfy the following restriction:

R2R3 R2 +R3 (3) and assuming that T is used to represent either of these' equated quantities of Equation 3, now

11 assumed to be equal, thesum .of the short-circuit transfer adm'ittances Y'I 2+Y.I 2 maybe written as follows:

Equation 1 and regarding p -w at real frequencies, then provided Equation 3 is satisfied.

If a voltage transforming device such as, for example, the resistance potentiometer P2 of Fig. 1 or the transformer TI of Fig. 2 is introduced at one end of the network-N 2 without changing the network Ni, the short-circuit transfer admittance YI2 of network NI is unchanged and the short-circuit transfer admittance YI2 of network N2 is modified through multiplying itby the voltage ratio K which in the case of a transformer TI of Fig. 2, for example, may be the ratio of the secondary voltage to the primary voltage. If the short-circuit transfer admittance 'YI2 of the network N2 as given in Equation 2 be multiplied by the voltage ratio K, the sum of Y'I2+Y"I2 of Equation 4 is changed to [i= at a 5 R2+R3' Yl2+Yl2- 1+Tp (6) Equating Equation 6 to zero gives:

5:0 at R(R2+R3)C3C4 (7) In accordance with the above equations, a variable [frequency of {3:0 may be obtained by introducing a variable voltage ratio K from a variable voltage device P2, TI, etc., and Without varying any of the resistances or capacitances of the networks NI and N2; This accordingly meets the requirements of a variable frequency bridge oscillator provided a suitable amplitude level sensitive stabilizing device, such as the auxiliary feedback circuit I8, is introduced in order to produce the proper deviations from Equation '7 at other than the level of stable oscillations. Since Equation 7 depends on Equation 3, a level sensitive element may be used which upsets Equation 3 except at the level of stable oscillations.

Fig. 2 is a circuit diagram illustrating the principle of operation of IOSSeCODtI'QH-Gd variable frequency oscillators. The circuit of Fig. 2 is similar to Fig. 1 but illustrates the frequency controlling variable loss Z in the form of an idealized variable voltage ratio transformer TI in place of the resistance potentiometer P2 of Fig. l, the transformer TI being employed to illustrate the principle of variable loss voltage ratio control of the frequency of oscillations. Fig. 2 also illustrates as an alternative arrangement to the stabilizer S of Fig. 1, a level controlling stabilizer S in the form of an idealized variable voltage ratio transformer T2 connected with a network N3 which may be used to illustrate the principle of operation of a level sensitive stabilizer for controlling the amplitude of the output oscillations. The feedback or p circuit in Fig. 2 employs the additional network N3 which is connected in parallel circuit relation with the parallel-connected networks NI and N2.

phase when itis positive.

The resistance R" and the condensers .C"3 and 16"4 comprise the'network N3 .and may be of the same size asthoseof the network NI but they are arranged differently as to which element is connected tozgro'und I and whichis connected to them-put. This makes it possible to use the same type of component elements for both of the networks'NI and N3. The variable voltage ratio e of the variable transformer T2, which is connected with the network N3, is relatively small and the transformer T2 represents means for introducing variable voltage into the circuit comprising the network N3.

The frequencies for 5:0 then correspond to:

where Y I2 is the short-circuit transfer admittance of the additional network N3, and Y'IZ and YI2 are, as before, the short circuit transfer admittances of the networks N2 and NI respectively. Satisfying Equation 3 relating the component elements of networkNZ to those of networks NI and N3, the short-circuit transfer adrnittanc es of such networks are:

R2+R3- l-I-Tp R1C'3 C4p III V Y 1+T wherein K is the voltage ratio of the transformer TI,- 6 is the voltage ratio of the transformer T2, and T represents either of the quantities equated in Equation 3.

Adding the .three values of the preceding equation, solving for the roots, and assuming ,that the voltage ratio ve of transformer T2 is small, gives the complex frequency .of 3:0 as:

In accordance with the above equations the ,8 circuit of "Fig. '2 when restricted by Equation 3, has non-minimum phase when the voltage ratio e of the amplitude level controlling voltage ratio" device T2 is negative and has mini- Hence, it is suitable for a bridge oscillator provided the voltage ratio e depends on voltage level or amplitude in such a way that it isnegative at the low levels and positive at high levels.

The use of both positive and negative values of the stabilizing voltage ratio e is inconvenient whenit' comes to practical embodiments of the variable loss-method of level sensitive amplitude stabilization of volta e. This situation may be avoided; however, by departing slightly from the the tube T4 in Fig. 1, the voltage values should be chosen in such a way that the equivalent values of resistance R2, as determined by the effect of the impedance of the potentiometer P2 thereon, always correspond to departures from "Equation 3 in the right direction in order to obtain the desired effect which is the direction calling for positive values of feedback stabilizer voltage e in the circuit of Figs. 2 and 3.

Fig. 3 is a circuit diagram of a loss-controlled variable frequency oscillator similar to that of Fig. 2 but illustrating a modification in which the network N3 of Fig. 3 has been omitted and the variable ratio transformer T2 has been connected with the network NI to illustrate the principle of providing tion embodied in the more practical feedback stabilization circuit l8 of Fig. 1. The transition from Fig. 2 to Fig. 3 is obvious if the plate-toground and the grid-to-ground terminating impedances of the p-circuit networks be regarded as both zero. For, in Fig. 2 the feed through the level controlling variable loss S is through the resistance R" to a point at which the condensers C3 and C4 are commonly connected,

the condenser C3 being connected to ground I and the condenser C"4 being connected to the grid of the amplifier A; while in Fig. 3 the feed through this variable loss S is through the resistance R to the point at which the condensers C3 and C4 are commonly connected, the condenser C3 being connected to the plate of the amplifier A and the condenser C4 being connected to the input gridof the amplifier A. The efiect will be the same in the two cases of Figs. 2 and 3 if the plate-to-ground impedance is regarded as substantially zero and if the impedance to ground at the output of the level control variable loss S in Fig. 3 is also regarded as substantially zero, so that transmission through the middle network NI is not adversely affected. In practice, the grid-to-ground impedance is relatively unimportant since the circuit is operated at or very close to balance. While the plate-to ground impedance is not zero, it is generally very small compared to the network impedances of the [3 circuit. For example, in the practical example given hereinbefore in connection with Fig. 1, the plate-to-ground impedance is made less than the resistance of the potentiometer P2 which in the example given hereinbefore is.

10,000 ohms, while the network impedance is of the order of 300,000 ohms, so that the plateto-ground impedance, though not zero, is very small compared to the network impedance and the result in practice is not greatly different from that for the assumed zero plate-to-ground impedance.

The frequency stability of the steady state oscillations depends upon the amplitude level V and the real frequency w of the steady state oscillations corresponding to the conditions of p -+1 and the damping constant =zero. One method of realizing frequency stability in an oscillator is to make the gain ,u of the amplifier A large and the loss {3 of the feedback circuit small at the oscillation frequency. distinguishing feature of a stabilized bridge oscillator is a [3 circuit such that fi=0 at some real frequency w=wo, when the amplitude level sensitive element 5 has a particular value corresponding to a particular amplitude level of V=V0. The p circuit is so arranged that it is a minimum phase network when V is greater than V0, but is a non-minimum phase network when V is less than V0. The corresponding ,1 or gain circuit is such that ,u is large at frequency w=wo, and such that oscillations build up when the amplitude level V is less than V0 and damp down when V is greater than V0. For practical purposes, this amounts to designing the cirlevel amplitude stabiliza- The.

.cuit as'for 'a feedback amplifier in which the properties assigned to the p circuit are those assumed by the oscillators circuit under the minimum phase condition of V greater than V0. Any circuit with these properties is in effect a stabilized bridge oscillator.

A large gain a leads to a small frequency difference 40-000 and also to a small amplitude level difference VVo. Accordingly, the frequency stability is related to the amplitude level and may be controlled by variation of the a circuit with amplitude level variation. The means used to obtain the required variation of ,B with change in the amplitude level may be any suitable level controlled stabilizing means such as the auxiliary feedback systems S which are particularly described herein in connection with bridge type oscillators although such feedback stabilizing systems S may also be used with other types of oscillators. A specific stabilizing element can usually be used at any of a multiplicity of points in a ,B circuit configuration which is otherwise prescribed. If any point is found such that B=0 at frequency w=w0 when V=Vo, the c circuit can be expected to make the transition from minimum phase to non-minimum phase condition as V passes through V0, provided the variationis made to be in the right direction, namely that the phase is minimum when V is greater than V0, rather than the reverse.

The relation between 5 and V and the control of the phase of is considered as follows. If stabilized oscilla-' tors are considered in general, the methods of obtaining frequency stability depend on the equation:

where 6V is determined by =0, and the derivative P are evaluated as for w=wo. The use of this equation depends on the observation that in practice {3 is very small at the frequency of oscillation. The value of ,8 depends in general on the complex frequency p=iw, and on the amplitude level V of'the generated signal, that is:

where V the amplitude level isa real number. Let the ,6 circuit be such that 5:0 when the real frequency w=w0 and V=Vo. Then for small departures from this condition, 5 may be in Taylors series, and only the first order terms are important:

and

Since in practice the oscillator operates at the frequency for which p,8=1 and since this must she arreal ifrequency, it is necessary thatithe action of tthevstabilizerishall :make =0, 1 and this r'deterrminesifiV. JHen'ee 6p=i '(w-wo), land this with as the change vin frequency from the nominal (6:0) to the actual ([Lfi=1) operating point. Maximum frequencystability requiresthat w-wo shall-ice .zero, and the .closer ito-zero that it can :be'ibroug'ht, the better the oscillator that will result. The expression aboveim'ay be made zero,

or made small, under the following conditions:

(1) The expression :may be made zero :by :making 1 o B I. av

.Since 'V, and hence '5V,-is a'real number, this requires that at; av

shall be a negative real number.

('2) "The expression may be made very small by making very large. This is what is done in a crystal oscillator, or by makingthe reactance-resistance ratio Q of the p circuit high in an inductancecapacitance type oscillator. Ina variable frequency oscillator it is diffi'cult to obtain :a sufficiently high Q to make this above an adequate means of obtaining adequate frequency stability, though in a fixed frequency oscillator it may be suficient. In the resistance-capacitance type of oscillator, the value of is strictly limited in magnitude.

(3) The expression may be made small by making [L large, so that a given percentage change in .will :be asmall absolute change. But this :alone is not adequate. Since 13:1, a large 1. requires :a small ,3, and if the increaseein is accompanied by a decrease in obviously not much is gained. But if [3 is made up of contributions from two parallel paths as occurs in bridge type oscillators, it is possible'to have 1/. very large and ,8 very small while is maintained fairly large.

In the oscillators here considered, is thus made large and is kept from becoming small by using a bridge circuit as .noted above, thus satisfying condition "(3) without violation of" condition (2). .Astinthe 1'6 oscillator described in connection with Fig. 1', it lis found that the phase 'of is then identical with the average "phase of 6,

:as this .must be kept within 190 degrees of a negative. real numberif the oscillator isto have avsingle mode of oscillation at any setting of the potentiometerPZ of Fig. 1. This accordingly approaches a' close realization of condition (1) ,above, and "is thus a means of improving the frequency stability :in View of the simultaneous satisfaction ofconditions (2) and (3) mentioned .above. Thecondi'tion (1) would not be satisfied if theusualresistance stabilizer were used. But 'th'euse'of theloss-controlled type of stabilizer circuit I 8 thus permits better frequency, stability than-that of avoltage-variable resistance stabilizerof the usual type. The matter of frequency stability may be carried further by either of the following methods:

(:a) Designzof the circuit-so that the average 6 over the twhole working range of frequencies is very close to a negati-vereal number. Identity of phase of 9e "ov with average masobtained by the loss-gontrolled stabilizer circuit .18, then substantially satisfies condition (1 :above simultaneously with condition "(3) and. also with condition (2) if sufliciently result :may be obtained independently of the .means for controlling the :frequency. Inaccord- :ancewith conditionil) the phase of the correc- ..ti'on introducediby operation of the voltage stabilizeris, when added to the phase of ,a which is the amplifier part A of the oscillator, equal or nearly equal to degrees, for maximum frequency stability. The circuit illustrated inFig. 1 meets this requirement :near the middle of the frequency range where average 43 is approximately '1 and 'accordinglyimproved perform- :ance is obtained'in "this range. "To illustrate this properly,;stops 5a and 5b are shown in Fig. l to 'c'onflne'the movement of the adjustable brush 5 oi'the potentiometer P2 to this region. This result may also be obtained by designing the circuit -.so :that the phase of average o over the .workingzrangefis very :close to 180 degrees.

Fig.4.isfi simplified circuit diagram of the losscontrolled variable frequency oscillation generator of .Fig. :1, :but illustrates a modification 'in which the two resistance potentiometers PI and P2 may -have :both wipers or brushes 5 on the :same .controlsshaft 6, as in Fig. 8, so that both or 'I-the potentiometers "Pzl xand;-P2 .may be operated rsimultaneouslyfrom .a single control shaft .6 in

order to :control the frequency of oscillation and to'balance outthe efiects of theflnite impedances of the potentiometers PI and P2 on the required amplitude stabilization thereby to reduce the amount of amplitude stabilization required by the stabilizer S. The finite impedances of the potentiometers PI and P2 balance out as regards the real condition for :0 at a real frequency corresponding to Equation (3) when K2=1-K1 and when YI I=Y|I where Y'II and Y"| I are as before, the short-circuit driving point admittances of the networks N2 and NI respectively. The two potentiometers PI and P2 so arranged vary more slowly with frequency and permit a wider spread of frequencies with practical forms of the potentiometers PI and P2.

Fig. 5 is a circuit diagram of a loss-controlled variable frequency oscillation generator similar to those of Figs. 1 and 4 but illustrating a modification in which "a single resistance potentiometer P2 is used to variably control the frequency of the oscillator circuit. While in Fig. 5, the potentiometer P2 is shown connected in circuit 2 with the network N2, alternatively it may be removed and similarly connected in circuit 4 with the other network NI.

Fig. 6 is a. circuit diagram of a loss-controlled variable-frequency oscillation generator similar to those of Figs. 1 and 4 but illustrating a modification thereof in which the frequency of oscillation is variably controlled by means of a pair of variable series resistances 2| and 23, instead of by the pair of potentiometers PI and P2. The variable resistances 2| and 23 of Fig. 6 may be unitarily controlled in opposite directions electrically by a link mechanism 25, in the manner of controlling the two potentiometers PI and P2 of Figs. 4 and 8 by means of the common control shaft 6. Resistances 21 and 23 may be disposed between the circuits 2 and 4 respectively and the ground I.

Fig. 7 is a. circuit diagram of a variable frequency oscillation generator similar to that of Fig. 6 but illustrating a modification thereofin which the frequency of oscillation 'is variably controlled by means of a single variable resistance 2| connected in series circuit relation with one of the networks N2, instead of by the pair of resistances 2| and 23 of Fig. 6 connected in both of the parallel circuits 2 and 4. The circuits of Figs. 6 and 7 accordingly utilize for controlling the frequency either one or two variable series resistances 2| and 23, in place of the potentiometers PI or P2 of Figs. 1, 4 and 5. The variable control resistances 2| and 23 are relatively much smaller in resistance value than the resistances used in the networks NI and N2.

Fig. 8 is a, circuit diagram illustrating modifications in the frequency control and stabilization'of a variable frequency oscillation generator of the bridge stabilized type. As illustrated in Fig. 8, the circuit comprises a r-circuit amplifier or source of gain A, a main feedback circuit .3 interconnecting the output circuit I6 of the amplifier A with the input circuit thereof and comprising two separate parallel-connected transmission paths 4 and 2 which include therein two fixed networks NI and N2, respectively, and two variable simultaneously-controlled potentiometer resistances PI and P2, respectively, which are used for varying the frequency of oscillation, and a modification in an auxiliary feedback circuit I8 which may be used for automatically stabilizing the amplitude level of oscillations generated by any of the oscillations illustrated in this specification. The amplifier A may comprise any suitable source of gain such as a vacuum tube system comprising one or more linear high-gain 1'8 vacuum tubes which function to amplify oscillations in a known manner. As an illustrative example, the amplifier represented by the block diagram A may comprise an amplifier as illustrated by the three-stage vacuum tube system, TI, T2 and T3 of Fig. 1.

The network NI transmits the higher values of frequency with a relatively small loss and attenuates the lower values of frequency in the parallel branch 4 of the main feedback circuit 3 and 4. The network N2 transmits the lower values of frequency with a relatively small loss and attenuates the higher values of frequency'in the other parallel branch 2 of the main feedback circuit 2 and 3. The two parallel-connected networks NI and N2taken together represent a frequency determining bridge network for the oscillation generator. The component elements of the network NI and of the'network N2 may be arranged in the form of any bridge suitable for the desired frequency of oscillation. As an illustrative example, the networks NI and N2 may be in theform of known T-networks such as socalled double-T or parallel-T networks. As illustrated in Fig. 8, the networks NI and N2 may consist only of resistances and condensers where it is desired to utilize the lower values of oscillation frequency, and such component resistances and condensers may be disposed in the form of T-networks of any desired form. As particularly shown in Fig. 8, the network NI consists of two series condensers C3 and C4 and two shunt resistances R and R, forming a threeterminal network NI; and the three-terminal network N2 of Fig. 8 consists of three series resistances R2. R3 and R3 and two' shunt condensers C2 and C2; The component elements of the networks NI and N2 may be constructed as fixed elements, as distinguished from mechanically variable or adjustable elements, and being of such fixed construction may be readily constructed in precision form.

In the case of low frequencies, capacitances and resistances only may be employed in the bridge networks NI and N2 which maybe of the type known as the double-T or parallel-T network as described for example in United States Patent 2,341,067 to R. O. Wisedated February 8, 1944'. The double-T network comprises two individual T-networks one comprising series resistances and shunt capacitance, and the other comprising series capacitances and shunt resistance. The two networks are connected in parallel circuit relation with respect to each other in the main feedback circuit 3 of the amplifier A. While particular forms of the networks NI and N2 have been illustrated in Fig. 8, it will be understood that the networks NI and N2 may be of other forms such as for example those illustrated in Fig. 1 and in other figures of this application. The networks NI and N2 may include resistances and condensers only, or they may be more complex and contain known inductances, amplifiers, phase shifters, etc. The component elements of the networks NI and N2 may be chosen in accordance with known design techniques to give the same value of short circuit driving point admittance YI I- and must have suitable values for the short- -eircuit transfer admittance YI2. I

The frequency control is provided by the variable voltage ratio, loss or gain control devices of Fig. 8 and is comprised of the potentiometers PI and P2. Instead of being placed at the left end of the fixed networks NI and N2 as illustrated in Fig. 8, they may be placed in a corresponding potentiozneter P2.

19 sitionat the right thereof, which is at theother or oppositeside or end of the-fixed networks :N'l andN2 as illustrated inFi'g; 20.: The potentiometers PI and P2 as illustrated in Fig. 8 may be constructed of two equal resistances PI and P2 which are connected in the main feedback circuit 3 of the amplifier A,.the potentiometer PI being connected with the network. N1 in' the branch circuit 4 by meansof the adjustable wiper hot the potentiometer Pi, and. the potentiometer P2 beingconnecte-d: with-the: other network N2 in the other parallel branch circuit 2 by means of the corresponding: adjustable :wiper 5 of. theme- The two wipers 50f the two potentiometers Pl' and P2 respectively are insulated electrically withzrespectto each other but are interconnectedmechanically so that a rotation of the potentiometer shaft 6 inonedirection mechanically will introduceresistance values in opposite directions electrically into" the parallel circuits 4 and 2, respectively. Forexample; a clockwise rotation ofthe potentiometer shaft 5 moves the movable wipers B'from the position illustratedin' Fig. 8 and" will increase the resist-- ance above the wiper 50f thepotentiometer Pi that'is introduced intothei circuit 4, andz'simultaneously will decrease thev resistan'ceniabove the wiper 5 of. the potentiOIneter'PZ'that is introduced into the circuit 2, thereby simultaneously increasing the magnitude of'the lossor' attenuation intro'duced into the circuit d'while' decreasing the loss, or increasingthe gain, introduced into the other parallel transmission path'Z." The variable potentiometers Fl and P2 accordingly operate to' introduce variable transmission loss in one of the two paralleltransmis'sion paths Z'and 4 with respect to the other for correspondingly varyingthe frequency ofioscillatiomand the frequenoy of oscillation" maybe selectivelyvaried over a wide range of frequencies merely by adjustment of the potentiometrs' Plxand P2;

It will be noted thatthe'potentiometersPI and P2 are each made up oftwo resistanceanamely, the resistances above'and'below the wipers Sthat are connected to thergroundcohnection I and to the main feedback circuit .3. The absolute values of the over-all resistances of thesp'oten'tiometers PI and P2 haveilittlex effect upon the frequency control provided they are. kept small relative to the valuesof the resistances which comprise the networks NI and N2. Therelative'values of the resistances of thelpotentiometers Pl and'PZ above ahdfbelow'the wiperst5 thereof are however of importance, and may easilybe held to'close limits for precision'frequency control. From a'practical standpoint, variable potentiometers such as the variable potentiometers Pl 1 and P2 whether simultaneously or separately controlled are one of the most precise types of variable elements since their relative rather than absolute values of resistancesmay beused'to make adjustments for control of the frequency of oscillation. The relative values of resistance above and below the wipers 5 may beikeptto closer limits than'absolute values of resistanc'eand hence the potentioiiieter form of resistanceis useful in precision frequency control as compared withseries resistances of thenon-potentiometer form where?- the absolute values of resistance become? of more importance in precise frequency control.

The frequency ofoscillation maybe independent of changes in the absolute resistance values of'the potentiometers- Pi and P2 provided the relative' resistances above and below the wipers 5 thereof remain correct t'o the extent that theadmittanc'es thereof are-relatively greater than the short-circuit driving; point admittances of the networks NI and N2. For example, if the ad.- mittances of the potentiometers Pi and P2? are very large and the short-circuit'driving pointedmittanc'es of the networks NT and N2 aresuificiently-small in comparison, achange of '1" per cent intheabsolute value of the resistances of the potentiometers Pi and P2 will have an effect on the frequency which isvery small in comparison with'l per cent, and this small effect will not-appear if theabsolutevalues of resistances of the two potentiom-eters Pl and'P-Z change by like amounts, as caused by equal temperature changes for example. If desired, one of the resistance-potentiometers PI and P2 may be:omitted or'not used, while the other is used to adjust the frequen-oyof oscillations. For example, the potentiometer P1 corresponding to the network NI may be-omitted as illustrated in Fig. 5, and the voltage ratio of the other potentiometer P2 may be varied approximately as the square of the frequency of oscillation, or if a linearre'lation'of frequency to'potentiometer shaft position 6 of Fig. 8 is desired, this calls for an appropriate shaping for the resistance values and contro mechanismof the potentiometer.

In order to stabilize the amplitude level of oscillations generated in the oscillator system comprising the amplifier A and the main feedbackcircuit 3,, 2 and-4 including the networks Ni and N2 and the variable potentiometers PI and P2, some form of non-linear level sensitive device may be utilized. It is desirable that the level sensitive device be capable of stabilizing the level of the amplitudeof the oscillations generated throughout the operating range or bandof frequencies in a variable frequency oscillator and for thispurpose an auxiliary feedback circuit l8 from the output of the-amplifier A to the input thereof may be provided as illustrated for example in Fig. 1 or in Fig. 8.

Asillustrated in Fig. 8, the auxiliary feedback circuit [8 includes an automatic volume control rectifier 251 which may be of conventional form and. which functions to rectify signal voltage fromthe output circuit to of the amplifier A and supply a proportional direct current control voltage to the grid electrode 9 of a variable transconductance vacuum-tube 22. The tube 22 here'operates as a variable loss device-in accordance'with variations in the amplitude. of oscillations and may be adjusted to vary its transmission loss over a suitable ranget'ostabilize' the level of the amplitude of oscillations over the range of operating frequencies. A: condenser 24 and a resistance 26 may be-usedin connection with the variable loss tube 22. Resistances 2B and 3b may be utilized to provide a fixed transmission loss of suitable value. The resistances 26, 28 and 30 may be adjusted to values that provide an infinite transmission loss for small amplitudes applied thereto from the output N3 of the amplifier A and to values that provide increasingly less loss in transmission for 'higher'amplitudes of voltages. The voltage level stabilizer S of Fig. 8 accordingly may comprise some form of variable gain vac uum tube 22 with an automatic volume control connection 28. The tube 22 may be a non-linear typetriode vacuum tube 22 which is biased below cut-off at low levels of signal voltage and overloaded at high levels of voltage. The resistance of the series resistance 26 and the capacitance of the condenser 26 interconnecting the plate l2 and the grid9' of the anode 2.2v may be chosen 21 to be of suitable values to assist in obtaining a constant output voltage level over the band of operating frequencies.

A portion of the output of the amplifier A which is fed through the feedback circuit lead I8 reaches two parallel circuits namely, the circuit through the resistance 28, and the circuit through the automatic volume control rectifier 20 which may, as an illustrative example, be arranged as shown in Fig. 8. With nosignal voltage applied to the lead IS, the negative voltage applied through the resistance Rg biases the vacuum tube VT to cut-off so that plate current does not flow, and the direct current potential of the lead wire 9a, is then equal to the negative voltage ap plied at the bottom end of the resistance Re. When signal voltage is applied through the lead I8 and the condenser Cg, plate current will flow during a portion of each cycle when the grid 9 of the tube VT is driven off, and the flow of this current through the resistance Rc will raise the potential of the oathode I and of the lead 9a above its no-signal value. The condenser Cc serves to by-pass the alternating current portion of the plate current so that only the direct current potential of the lead So will vary as the signal voltage is changed. The resistance Ri is used to prevent the condenser Cc from shorting out the alternating current signal voltage applied to the grid 9 of the tube 22, while the direct current potential applied to the lead 9a and at the cathode 'I of the tube VT is used as the grid bias of the tube 2.2. The circuit is so proportioned that the grid 9 of the tube VT is never positive, as this would short down the transmission through the path in parallel with the tube VT which is the path commencing with the resistance 28.

The signal voltage is attenuated in the resistances 28 and 3E) and then divides, a part flowing through the resistance 26 to the input of the amplifier A while another part reaches the grid 9 of the tube 22 throughthe condenser 24. The voltage applied at the bottom of the resistance R0 is sufficient to bias the tube 22 nearly to cutoff so that the signal current flowing in the plate circuit of the tube 22 in response to the signal voltage applied to the grid 9 through the condenser 24 is equal to that reaching theoutput lead I8 through the resistance 26. Since the signal through the tube 22 is reversed in phase, these currents through the condenser 24 and the resistance 26 may be made to counteract each other so that no signal is returned to the input of the amplifierA. By' somewhat over-biasing;

the tube VT, this condition may be maintained until the output voltage of the amplifier A reaches a predetermined magnitude. As the output voltage of the amplifier A further increases, the transmission through the resistor 26 will remain constant at a constant transmission loss, but the action of the tube VT raises the grid potential at 90. continuously andaccordingly the loss through the tube 22 continually decreases. The signal from the plate circuit of the tube'22' thus more than balances that from the resistance 26, and the signal returned to the input of the amplifier A increases more rapidly than changes in the output voltage of amplifier A. The phase of the returned signal is opposite to that of the output of the amplifier A because of the reversal or turnover in the tube 22, and hence is correct to stabilize the output of the amplifierAt With increasing output of the amplifier A, the negative direct current voltage of the "grid 9 of the tube to a potential above cut-.

22 gradually decreases, and the following characteristics gradually increase with increase of output ofthe amplifier A, namely, the signal through the plate I2 of the tube 22 which is of negative phase, and the signal through the resistor 26 which is of opposite or positive phase. The difference of the last two characteristics results in the signal returned to the input of the amplifier A which gradually increases with increase in the output of the amplifier A.

The use of the resistor 26 and the delayed automatic volume control 20 as obtained by overbiasing the tube V'T insures that this stabilizer path I8 does not give a positive feedback which may or might permit undesired oscillations to build up and interfere with the desired oscillation controlled by the main feedback path 2, 3 and 4. The use of the automatic volume control rectifier 22in Fig. 8 requires a condenser C0 or equivalent in order to get the'r'ectifier characteristic. "A change too quick in frequency will upset the stabilization until the transient in the Rc-Cc circuit has died down. The circuts I8 of Figs. 1, 2 and 3 do not require a rectifier and hence are more.

suitable where very high speed variations in frequency' occur. A

When the networks NI and N2 are built to high precision and the gain of the amplifier circuit A is very high, the auxiliary feedback stabilizing system I8 of Fig. 8 is readily made to feed back only at high amplitude levels. To reduce the introduction of harmonics in undesirable quantities, the sensitive automatic volume control rectifier 20 and the high loss device in the form of the'series resistance 28 and the shunt resistance 30 in front of the tube plitude stabilizer I8 may be made more or less slow-acting to a degree that is dependent upon the constancy of the level of oscillations, the rate at which the frequency of oscillation is to be changed and the precision to which the network system NI and N2 is built. If the network system NI and N2 were in perfect adjustment, the feedback through the variable stabilizer circuit I8 at equilibrium is independent of frequency change to the extent that it is made independent of frequency change. In practice, the stabilizer will reach equilibrium corresponding to the actual conditions. In general, only a small contribution from the amplitude stabilizer circuit I8 willbe needed to establish any real frequency of infinite loss, if the potentiometer resistances PI and P2 are made small relative to those of the networks NI and N2 or stated differently if the transfer admittance of the potentiometers PI and P2 is networks NI and N2.

An important feature of this invention is that the frequency ofv oscillation'may be varied, not by mechanically changing or varying in the conventional manner the values of resistance and capacitance of the networks NI and N2, but rather by varying the resistance and voltage ratios of the potentiometers PI and P2. Thus the resistances and capacitances or other component elements of thenetworks NI and NZ-per se may be fixed elements in so far as variation in the variable frequency of oscillation is concerned, and the frequency of oscillation may be varied merely by varying the resistances of the potentiometers PI and P2 which are relatively small resistances compared with those of the network-s NI and N2. The ease with which the frequency of oscillations may be controlled by the variable potentiometers PI and P2 or 22 may be used. The amother variable loss control devices is an important feature of this invention. The frequency of oscillation is varied by supplying similar: phased inputs to the. two fixed networks NI and N2, adding or subtracting the outputs thereof, andsetting the phase of the resultant output thereof by the relative magnitudes of the inputs to the networks NI and N2. Thus, by using the two potentiometers PI and P2 to control the two inputs to the respective networks N I andNZ in opposite directions, a change in phase of. the-resultant outputs at a given frequency will occur with a change'in rotation of the wipers 5 of the potentiometers PI and P2. Such change in phase will be little or large in the output depending upon the actual phase differences at the output of the networks NI and N2, and the frequency of oscillation will change until the resultant phase is that necessary for oscillation.

In general, high frequency stability may be obtained with bridge type of oscillators which use a so-called a and '5 loop circuit, the a circuit comprising an amplifier or source of gain A having over-all input and output circuits which are interconnected by means of a suitable network system to form the main feedback {3 circuit 3. The-,3 circuit also includes some form of nonlinear voltage'variable stabilizing" element which might be a conventional thermistor, or tungsten or carbon lamp, butwhich preferably is an auxiliaryfeedback circuit I8 which automatically stabilizes the amplitude level of oscillations and is so arranged by proper adjustment of the amplitude level stabilizerS that it gives infinite loss or attenuation" in the B circuit at or very near to theoscillation frequency. At a point of infinite loss 3:0. It is known that the natural frequency modes of a simple feedback amplifier are at frequencies of ,u/5'=+1, which transposed gives 1 c= There will be such a frequency close to any frequency of :0. If the amplifier gain is large at that frequency, then a small value of ,8 times the large value; of a is equal to unity. Thus with a high gain -circuit A, the feedback bridge typeoscillator will oscillate at a frequency close to that of 5:0 with the proper adjustment of the stabilizer circuit I8 which automatically adjusts itself to the proper value by the amplitude level of theoscillations. The bridge type oscillator maybe any such oscillator embodying the principle of using a point where 5:0 on proper adjustment of the amplitude level stabilizers.

Animportant characteristic of the b-circult is the location of the infinite loss positions. It

isconvenient to assume zero termination at'both ends of the network or ,8-circuit. This is be cause the short-circuit admittance YIZ of parallel-connected networks Ni and N2 is the sum of the corresponding short-circuit transfer admittances of the separate networks NI and N2, as is known. In accordance with conventional usage,-

YI I is the short-circuit driving point admittance of'the associated network with its output terminals shorted, YIZ is the short-circuit transfer admittance of the associated network with its output terminals shorted, and Y is a function of p wherep equals in; and of 1220:2200 where we is an assigned fixed frequency inradians per second. The driving point admittances YI i on short circuit arethe same'for networks N2 and Ni in the circuit of Fig. 8 while their corresponding transfer admittances YIZ are of the form 22 1 K) I I P 2 0.. where Fromthese expressionsv it follows that when the variable loss amplitude stabilizing system I8 is set for infinite loss, there will be a real frequency w of infinite loss for the complete {3 network at As the gain a of the amplifier A approaches infinity or is high, the singing frequency will approach this frequency 0:. When the variable loss network system is set a little oii infinite loss setting, the natural mode becomes'complex and of the form ui -iw, which corresponds to a mode of oscillation of the form-e sin (wt-+0). The damping value of a will be positive which represents a decreasing oscillation level when the variable loss gives transmission in phase, and negative representing an increasing oscillation level when the variable loss gives transmission with reversal in phase. Thus in order to stabilize the level ofv oscillations, the automatic volume control of the variable lossamplitude stabilizer S should ideally be such that low amplitude levels give transmission with reversal in phase leading to increasing levels, while high amplitude levels give direct transmission'leading to decreasing amplitude levels. The amplitude stabilizer" S of the auxiliary feedback circuit iii of Fig. 8 conformswith the latter condition.

Fig. 9 is a circuit diagram of a variable frequency oscillator illustrating one of several ways in which a variometer or variocoupler VC maybe used to variably control the frequency of oscillation. As illustrated in Fig. 9, the variometer V0 is disposed in the output of the networksNl and N2 and may consist'of inductance windings BI, 33 and 35. The windings 3| audit may be the fixed crossed primary windings, and the winding 35 may be the rotatable secondary winding of the variometer VC. The'coupling between the windings SI and 35 maybe a maximumv when the coupling between the windings 33 and 35 iszero; and vice versa. By variation in the rotatable sec-- ondary winding 35, the frequency of oscillations may be varied.

It will be understood that in loss-controlled variable frequency oscillators, the variable loss may be controlled not only by variable potentiometers or variable series resistances butalternatively, the loss may be variably controlled by variable inductancessuch asthe variable voltage ratio variometer; VC' of Fig. 9' which like the when the wiper-posi.

flable amplitude modulator of jtude type or otherwise.

electrode [2 .of the amplitude modulator M potentiometer may be used as a means of producing a variable voltage ratio to obtain a variable frequency of oscillation. In addition, a variety of other variable loss type means, as described hereinafter, may be used for producing the variable loss or voltage ratio required to obtain a variable frequency of oscillation. As an ,example, vacuum tube circuits may be arranged in such a way that their through gains or losses may be varied by varying auxiliary control vo1tages applied thereto in order to obtain variations in the oscillator frequency. This method as more fully described hereinafter offers a convenient means of obtaining frequency modulation of an oscillator or other circuit.

Figs. 10, 11, 12 and. 13 are circuit diagrams illustrating examples of systems for converting amplitude modulators into frequency modulators that are relatively simple in construction in com parison with conventional frequency modulation systems using any but very small percentage modulations. In the circuitsillustrated in Figs. to 13, variations in or modulation of the frequency of oscillations are obtained by varying the gains or losses in the circuit by means of auxiliary control voltages which may be supplied from an amplitude modulator connected therewith.

Fig. 10 is a circuit diagram illustrating a losscontrolled frequency modulated oscillator in which the loss is variably controlled and the frequency is correspondingly modulated by means ofan electronic vacuum tube or other suitable amplitude type modulator M controlled by variable signal control voltages applied by a suitable signal source 36 to the amplitude modulator M. Except for the substitution of the amplitude modulator M in place of the potentiometer P2 of Fig. 5, the circuit of Fig. 10 is similar to that of Fig. 5, the corresponding parts being given like reference characters. In the loss-controlled oscillator circuit of Fig. 5, the loss is variably controlled by a potentiometer P2 which may be used as a means for producing a variable voltage ratio in order to control the frequency of oscillations. In the loss-controlled oscillator of Fig. 10, the loss is variably controlled by the modulator M by varying the auxiliary signal'control voltage applied thereto from the signal modulated volt- ;age source 36 in order to correspondingly modulate the frequency of oscillations. The system illustrated in Fig. 10 accordingly offers a simple and convenient means for obtaining frequency modulation of relatively large percentage modu- 'l ation, and for converting amplitude modulators :into frequency modulators.

The modulator M of Fig. ,10 may be any suitthe linear amph- If a known square law type of amplitude modulator M relating its volt- .age ratio K .to the frequency of oscillations in accordance with Equation 7 be used in the cir- ..,cuit of Fig. 10,, linear frequency modulation thereof maybe obtained. As illustrated in Fig.

:10, the input connections to the amplitude modulator M are conventional. The amplitude mod- ;ulated signal is applied 1mm the signal source $36 to the input grid ll! of the amplitude modula- .tor M. The carrier frequency from the principal feedback circuit 3 is applied to the input grid 9 of the amplitude modulator M. The modulated :carrier output is taken from the output plate and applied to the input of the network N2. The frequency modulated output signal may be taken from the output terminals l and it which are connected with the output circuit of the amplifier ,wcircuitA. The level sensitive amplitude stabilizer S may be in accordance with that of Figs. 1 or 8.

Fig. 11 is a circuit diagram illustrating a losscontrolled frequency-modulated oscillator which is similar to that of Fig. 10 except for using two amplitude modulators MI and M2 which may be linear type electronic amplitude modulator tubes of conventional design. By using the two amplitude modulators Ml and M2'connected as illustrated in Figs. '11 and 14, linear frequency modulation may be obtained with the conventional linear type of amplitude modulators. As shown in Figs. lland 14, the two amplitude modulators Ml and M2 may be connected in tandem, the carrier frequency from the main feedback circuit 3 being applied to the input grid 9 of the first modulator tube MI and the output of the first modulator tube Ml being applied to the input grid 9 of the second modulator tube M2. The modulated carrier output from the second modulator tube M2 is applied to the input of the network N2. The amplitude modulated signal from the signal source 36 is applied to the grid circuits IU of both of the amplitude modulator tubes MI and M2.

Fig. 12 is a circuitdiagram illustrating a losscontrolled frequency modulated oscillator which is similar to that of Fig. 11 except for the connetcion arrangement of the two amplitude modulators MI and M2, and the addition of a three- Winding transformer 31 having a primary winding Li and two series-connected secondary windings L2 and L3 which are connected in series aiding relation. By using two amplitude modulators MI and M2 connected as illustrated in Fig. 12, fairly good linear frequency modulation may be obtained even when using linear type ampli-- tude modulator tubes MI and M2, and fairly large percentages of modulation may be obtained. With the two amplitude modulators MI and M2 connected in the parallel arms of the two net- 'works NI and N2 respectively, the carrier frequency from the main feedback circuit 3 is simultaneously applied to the inputs of the amplitude modulators MI and M2, and the modulated carrier outputs thereof are applied to the networks N l and N2 respectively. The amplitude modulated signal from the signal source 36 and the transformer 31 is applied to input circuits of both of the modulatorsMl and M2.

Fig. 13 is a circuit diagram illustrating a freulator tubes MI and. M2.

of Fig. 12 except for illustrating a particular type of amplitude modulating devices MI and M2 each in the form of a conventional multigrid mixer tube,'which may be of the typedisclosed for example in United States Patent No. 1,896,780, issued February 7, 1933, to F. B. Llewellyn. As illustrated in Fig. 13, the'carrier frequency from the main feedback circuit 3 may be applied to the inner grid 9 of each of the two modulator tubes MI and M2, and the amplitude modulated signal from the signal s0urce36 and from the end windings L2. and L3 of the transformer 31 may be applied to the other grids'l fl of the mod- The modulated carrier outputs from the plate electrodes I2 of the modulator tubes MI and N2 are applied to the networks Nl'and N2 respectively. while particular types of modulator tubes MI and M2 and circuit connections have been illustrated in Fig. 13, it will be understood that other circuit connections and other types of amplitude modulat-

Images