US1794230A - Apparatus for amplifying electric oscillations - Google Patents

Description

Feb. 24, 1931. 1,794,23e

APPARATUS. FOR AMPLI FYI NG ELECTR I C OS C ILLAT I ON 5 L. M HULL Filegl Jan. -9; 1925 2 Sheets-Sheet l 1.. M. HULL 1,794,230

APPARATUS FOR AMFLIFYING ELECTRIC OSCILLATIONS Feb. 24, 1931.

1 Filed Jan. 9, 925 2 Sheets-Sheet 2 Patented Feb. 24, 1931 UNITED STATES PATENT OFFICE LEWIS M. HULL, OF BOONTON, NEW JERSEY, ASSIGNOR TO RADIO FREQUENCY LABORA TOBIES, INCORPORATED, OF BOONTON,

JERSEY NEW JERSEY, A CORPORATION OF NEW APPARATUS FOR AMPLIFYING ELECTRIC OSGI'LLATIONS Application filed January 9, 1925. Serial No. 1,471.

stage are connected in cascade, the overall voltage amplification of the cascaded system is not equal to the product of the amplification factors of the several stages. This failure to secure the theoretical voltage amplification from a plurality of stages is particularly pro- .nounced at the higher frequencies from 500 to 1500 kilocycles per second but is marked at the intermediate frequencies lying between 30 and 100 kilocycles. If the amplifying tubes in anumber, n, of like cascaded stages were employed with full e'liiciency, the overall voltage am lificat-ion of the system would be numerica ly equal to the voltage amplification per stage raised to the nth power. This I define as true cascade amplification. The amplification obtainable with prior systems and radio receivers has fallen far short of such a figure. With a radiofrequency transformer-coupled amplifier, assembled in the usual fashion and with the tendency towards self-oscillation in individual stages controlled by the introduction of variable losses in the input stage, I have found, experimentally, a voltage amplification' of .100 for a single stage, 200 for two similar stages in cascade and 310 for three similar stages in cascade. In another typical cascade system of the intermediate frequency type, operating at a frequency of 30 kilocycles, the voltage amplification for one stage is of the order of 30, for two stages 450, and for three stages 900. If the amplification for one stage is termed A, the amplification observed at the higher frequencies corresponds respectively, to A A and A for one, two and three stages; and the amplification at intermediate frequencies was A A and A, respectively, for one, two andthree stages. The significance of these figures is not in the absolute values of the amplification, which vary with the types of tubes and coupling units employed, but in the absence of a true cascade effect in the overall amplification of a plurality of stages. Quantitative experiments indicate the absence of this cascade eifect in all ordinary amplifiers, irrespective of the type of audion tube employed.

' An object of the present invention is to provide amplifying apparatus comprising a plurality of cascaded stages and in which the overall amplification increases exponentially with the number of stages. A further object of the invention is to provide a plurality of tuned amplifying stages which are so con- ..structed and arranged that the input reactance of each amplifying stage is lndependent of the value of the variable reactance in the input circuit of the succeeding'stage. A further object is to provide a selective radio receiver having a plurality of tuned amplifying stages adjustable to a common resonance frequency by the manipulation of a single tunlng control. More specifically an object is to provide a selective radio receiver tunable by a single control whose resonance characterlstics are independent of the re actance of the antenna, loop .or collecting structure to which the receiver is connected. More specifically, objects of the invention are to provide certain details of construction and combinations of amplifyin stages or portions thereof which are her nafter described and claimed.

Contrary to the universal opinion, based upon evidence similar to that presented above in quantitative form, that the voltage amphfication of each added radio frequency stage is progressively smaller and that no real advantage is gained by using more than two. or at most three, stages of amplification, I have demonstrated that a number of stages may be so cascaded as to provide an overall voltage amplification which increases exponentially with the number of stages.

To secure this cascade amplification, I have found two definite conditions which must be fulfilled: first, each stage must be a true oneway repeater; second, the interaction between non-adjacent stages due to stray electric and magnetic couplings and to small voltage drops in common battery leads, must be reduced to insignificance. In other words, if each stage is so arranged and adjusted that the input circuit of any iven' audion repeater has no appreciable e ect upon any circuit ahead of this input circuit in the amplifier, and if the current in this input circuit affects all circuits beyond it in the cascade amplifier only through the agency of thermionic space currents in the audion of this stage, then the amplification given by a num-' ber of stages will be numerically substantially equal to the amplification of one stage raised to a power equal to the number of stages.

7 I have also discovered that the maintenance of a true one-way action in each of a series of stages is beneficial in a different way: it renders the efiective input reactance 0 any given stage independent of the constants of each of the succeeding stages, including the output circuit of the particular stage in question, with the result that when the coupling units of each include a tuned circuit, the tuned circuits of all the stages can be made mutually resonant by means of reactances external to the several audions which are all equal. In previous attempts to tune a plurality of similar stages by a single control the progressive differences which may exist in the input reactances of a series of cascaded stages, which differences are present even though the external tuning reactances are equal for all the stages, has apparently been overlooked, with the result that the band of frequencies through which the different stages are even approximately is resonance with each other is extremely limited.

I also gain increased selectivity in a series of cascaded stages owing to the fact that the variation of effective input capacity with frequency which is known to exist when the output circuit of an amplifying stage is allowed to react upon the input circuit of that stage, is absent in an amplifier made up of one-way stages.

For a better understanding of my invention, reference is had to the accompanying drawings, in which,

Fig. 1 is a diagram showing the schematic organization of my invention.

Fig. 2 is a circuit diagram of a one way amplifying stage suitable for use with. my invention. I

Fig. 3 is a broken perspective view of one physicalembodiment of a plurality of cascaded one-way stages.

Fig. 4 is a circuit diagram of a radio receiver including a plurality of cascaded oneway radio frequency amplifying stages, the first or input amplifying stage being adapted for use with an antenna or capacity collecting structure, and

Fig. 5 is a fragmentary diagram of portions of the firststage of a receiver adapted for use with a coil aerial or inductive collecting device.

True unidirectional action of the cascaded stages such as is contemplated in the present invention is not obtained simply by the neutralization of the grid-plate capacit of each tube and the usual shielding of t e stages from each other. As is described in the coending application of Stuart Ballantine,

er. No. 720,708, filed.June 17, 1924, an audion amplifying stageof ordinary type will I usually not be unidirectional for several reasons, as follows:

1. On account of the electrostatic coupling due to the capacity between grid and plate electrodes of the audion and their connections; 2. On account of the flow of ions or electrons from the filament to the grid of the audion;

3. On account of the coupling due to the .flow of the plate (thermionic) current throu h the filament;

4. n account of the charge induced upon the grid electrode by the varying space charge;

5. On account of any incidental coupling between the coils or condensers and their connections on either side of the repeater tube.

For purposes of discussion, I shall separate all amplifyin stages into two general classes, which I de he as follows: I

First, one-Way stages, in which retroactive and interactive couplings due to all five of the above-named causes have been substantially eliminated; second, neutralized stages, in which the coupling of the first type only (electrostatic capacity between grid and plate) has been eliminated. When two neutralized stages are cascaded, if the losses in the interstage coupling transformers are sufliof the couplings used, but true cascade amplification results. The failure of conventional amplifiers comprising two neutralized stages so proportioned as to be non-oscillatory, to eliminate all important retroactive couplings accounts for the relatively high overall amplification in such systems. Although the residual regeneration; present in two neutralized stages may produce an over allamplification for these two which isgreat er than thatwhich would be obtained if the same two stages were converted to the one,

way type, the significant fact is that not even an approxlmatlon of true cascade 'amplifi-f cation is obtained when more stages are added to the neutralized amplifier.

A typical one-way amplifying stage is shown dia rammatically in Fig. 2. The input trans ormer IJ IJg is tuned by a condenser C in the usual manner. The audion T comprises a three electrode amplifying tube having capacity C between the grid and plate P. The output circuit of the stage includes the primary L of the transformer L L the secondary of which is tuned by the condenser 0 these forming the input circuit of the succeeding stage. The input terminals of the stage, i. e., of L 6 are not connected across the input terminals of the audion T, but the low oint, Y, of the secondary of the input trans ormer is connected to a filament lead F through-a coil L which is coupled to L and a grid biasing battery C. By low point of a circuit element or impedance I mean that point of the element which is at the lowest alternating current potential, or, in an audion circuit, most closely approaches the alternating current potential of the audion filament. The A and B batteries and the rheostat B may be of the usual forms. 7

As has been described in the above mentioned Ballantine application, this arrangement of input and output networks constitutes a bridge having the capacity arm C... and a second capacity arm C between the grid and filament terminals which second arm may comprise the capacity between the electrodes or their connecting wires, or an external condenser. The points G, P and F, are three corners of the bridge system and the fourth point is the point X in the output inductance L Due to the coupling between L and L the point Y in the input circuit is at the same A. C. voltage as some point X of the output circuit, and the input of the stage is effectively across G and X, which are at opposite corners of the bridge. By adjusting C to provide the proper capacity and thus balancing the bridge, the reaction of the output upon the input circuit through C... is revented since the output is across P and the other corners of the bridge. Theoretically the bridge can be balanced independent of frequency with any value of coupling between the coils L and L since a decrease in this coupling, with a given pair of coils, is entirely compensated in obtaining a balance by an increase in the value of capacity 0,. I have found in one specific practical embodimentof my invention using this type of circuit with an audion of the type known commercially as UV-201A that a suitable value for the coefficient of coupling between coils L and L is K=0.25, where K is defined by the relation:

I K =M45 1/ 4 The numerical ratio between inductances L and L should be in this case: L4/L =4 to 4.5.. A bridge balance may be obtained with a capacity at C ofthe order of 40micro-rnicrofarads. If this balance is obtained in the absence of thermionic space currents Within the audion then the electrostatic coupling through C only is eliminated and the stage falls in the class of neutralized stages. My invention is not, of course, limited to these or any other specific constants. p

In order to develop this system further, into a one-way stage, the other causes of retroactive currents are eliminated as follows:

The grid-filament current,which manifestsv terminals. Incidental'couplings between the input and output circuits are eliminated by shielding around the coils, condensers, and high-potential connecting wires.

The general technique of radio and audiofrequency shielding is Well known in the art, as is the-practice of shielding certain parts of amplifiers from each other. I have found, however, certain special types and arrange ments of shields to be peculiarly adapted for the provision of one-way action in this particular circuit. An individual shield for each transformer and its associated leads is absolutely essential for one-way action. It i: also desirable to shield the tuning condensers from each other and from the audions and associated parts. The preferred form of shielding is indicated diagrammatically by dotted lines in Fi 2, which lines'define foux enclosed spaces which are located, respectively, the input transformer L L input tuning condenser C the output transformer L L,,, which comprises the input transformer of the succeed ing stage, and the tuning condenser U for tuning L t These shields may be formed of-highly conducting metal, of a thickness which depends upon the constants of the metal. With coupling transformers having straight solenoidal windings, positioned about 5 inches between centers, sheet copper or sheet brass about 0.02 inches thick is suitable.

Apractical physical embodiment of tht principal parts of a stage and of the shields is illustrated in Fig. 3, in which view the several parts are identified by the same reference characters as are used in the diagram of Fig. 2. The several shields S S S and B S S and 3,. within I S takethe form of closed rectangular boxes or cubicles within which the elementsare placed.

In Fig. 3, the input stage is shown at the right of the first complete amplifying stage,

it being understood that additional and s1m1- larly constructed amplifying stages will ordinarily be arranged to the left of these stages. The particular arrangement illustrated is intended for use in a cabinet of the usual type and with the aligned ends of the boxes adjacent the usual vertical panel. The fixed inductances are arranged at the lower part of the assembly, the several cylindrical transformers being arranged'with their axes horizontal and normal to the ends of the shields S 8,, which will be adjacent the panel. The shields S S. which house the tuning condensers C C are shorter than the boxes S S and are arranged above the latter at their forward ends. The socket for audion T, the battery leads and the condensers C C are positioned'on or above'the shelf which is provided by the rear portion of the shield boxes S S it being obvious that these various parts are insulated from the shields. An additional shield S is provided,

this taking the form of a flat plate which is' positioned above and parallel to the upper surfaces of the shields S S4 and which hasspaced openings above the several tube sockets. v

The second of the two requirements for cascade amplification, viz: isolation of the several stages, is accomplished with respect to parasitic couplings by the shieldswhich surround the coupling units, since these shields serve to isolate the units individually.

With certain arrangements, I have found it advisable also to enclose the audion tubes in individual shields S... In addition to this, it is necessary to-bring to ground potential all low voltage battery leads by the liberal use of high-capacity by-pass condensers.

. It will be noted that the tuning condensers C and C 'are shielded from their respective inductances, L and L and that the wiring within a stage 15 positloned within the shields which house the parts of the stage. The tube socket and battery Wires are located outside of the boxes but the leads from the plate and grid terminals of the socket pass directly into the boxes S S and the exposed portions of these leads are'kept as short as is possible. i

I have enumerated all of the features necessary to provide one-way characteristics for.

the amplifying stage of Figs. 2 and except that one which concerns the retroactlve coupling due to the space charge in the audion.

As pointed out in the copendi-ng Ballanti'ne application, 720,708, the electric flux due to this mobile space charge which constitutes the thermionic plate current of the audion, manifests itself as a change 1n the effective grid-plate capacity of the audion. When the filament of the audion is heated and the steady plate voltage turned on the effective capacity. between grid and plate is different from the electrostatic capacity between these electrodes. I wish to emphasize that this is no minor, second-order effect. In an ordinary receiving audion, I have detected a change in the efi'ective grid plate capacity when the filament was turned on, which amounted to of the original electrostatic capacity between electrodes.

Thus the old methods of securing a balance with a cold filament give spurious results, and do not produce the desired one-Way effect even though the circuit purported to be balanced includes all physical elements necessary for securing .an accurate balance. ,An'

"having more than two cascaded stages will be apparent from a consideration of Fig. 3, and Fig. 4 is a circuit diagram of an amplifier, specifically a radio receiver, having three stages in addition to the first amplifying or input stage, and the audion detector. The several tubes of the four amplifying stages are indicated by numerals 1 to 4, respectively, and the detector audion by numeral 5. The input or first amplifying stage differs from the succeeding stages as the reactance of the first input transformer will be influenced by the properties of the collecting structure with which the receiver is associated. In addition to the input transformer L L and tuning condenser C an adjustable reactor, as the condenser C is provided in series with the in- V cascading a plurality of stages, common bat-' teries are used, and additional by-pass condensers of large capacity C and C are provided across the B and C batteries, respectively. Separate by-pass condensers C are provided for each stage and individual balanc ingcondensers C are of course necessary, the condensers C being once adjusted and preferably clamped in proper position to balance the separate stages.

The several tuning condensers C are arranged for operation b which as indicated in ig. 3, comprises. a

a single control,

shaft arranged parallel to the front of the set, suitable insulation being provided between the sections of the shaft to which the several sets of movable plates are attached. Due to the fact that the effective input reactances of stages between terminals G and Y, note Fig. 2, are equal when the couplingtransformers have primary and secondary inductances, respectively, equal to each other, the one-Way stages including audions 2, 3 and 4 may be simultaneously tuned to resonance by similar variable condensers G which are so coupled as to move in unison. Although the effective input reactance of the input audion 1 is afi'ectedby the particular current collecting structure which may he used, the effect of the arbitrary value of such a structure can be eliminated by a suitable adjustment of the compensating condenser C As the input stage is also a one-way repeater, the adjustment of C is substantially independent of frequency and in actual practice no variation in the setting of C can be observed when the latter is adjusted, for a given antenna, for best reception of the highest or the lowest frequencieswithin the broadcast range. The four amplifying stages may be concomitantly tuned to resonance by a single control, even though the tuning of the individual stages be extremely sharp, an operation which has not been possible with previous constructions which include a plurality of sharply tuned amplifying stages. In prior constructions using a single control, no account has been taken of the fact that even though the constants of all external reactances are equal, a progressive difference may exist between the effective input reactance of each amplifying tube, with the result that there has been only one or two settings of the single control at which all stages were'on the resonance peak. At other settings or frequencies all but one of the stages have been slightly off the resonance peak. The present invention avoids this difficulty since the input stage and all subsequent stages will be tuned to the resonance peak when the stages are one-way amplifiers and the external reactances of the several stages are identical.

In Fig. 5, I have illustrated a portion of the input stage of an amplifier designed for use with a coil aerial; the equalizer which permits the simultaneous tuning of the input and subsequent stages takes the form of a variable inductance or variomet'er, L which is so connected that loop aerial, A and the equalizer L are in series across the tuning condenser, C

In its broader aspects, the present invention comprises the cascading of a-plurality of one-way repeaters. Contrary to previously accepted theories,.I have found that the elimination of such incidental feed-back as has increased the amplification in one or two stages, has not resulted in lessened ampli- In this modification fication for a plurality of stages but has had the surprising result of securing an actual amplification equal to the maximum which was theoretically obtainable with a given number of similar stages. I have built cascaded amplifiers having four resonant oneway stages operating at frequencies between 500 and 1500 kc. in which each stage was constructed and adjusted in accordance with my invention as described herein. The amplification for a single stage was 7, and the observed overall amplification of the four stages was, within the limits of experimental error. equal to 7 or 2401.

I have found moreover that a one-way amplifier including an arbitrary number of stages and constructed as an isolated shielded unit in accordance with my invention can be used in connection with a separate retroactive amplifying step which yields an amplification of 50 to alone, without impairing the efficiency of either. The retroactive amplifying step may be an ordinary tuned stage in which incidental regenerative currents are controlled by a variable loss device in the input circuit. The one-way amplifying unit should be shielded from the retroactive unit and for this purpose shielding in addition to that necessary to produce onewayaction may be required. The combination yields an overall amplification equal to that'obtaincd from my amplifier alone, multiplied by that obtained from the retroactive stage alone. Such an advantageous combination is not possible with an ordinary am plifier comprising neutralized stages. in which electrical stimuli from the retroactive stage react upon and through the various stages of the other amplifier. It will thus be seen that my invention offers possibilities for enormous amplification, both when used alone and when used in connection with other amplifying equipment.

It is apparent that the invention is not limited to the specific circuit details and arrangements since other means are now known or may be devised for eliminating some of the causes of retroaction in a stage or between stages.

I claim:

1. In a radio receiver of the tuned radio frequency amplifier type, the combination with a plurality of non-regenerative cascaded vacuum stages each of which has a tuned input circuit including a fixed and a variable reactor for tumng the stage to reso-' nance at the frequency of the signal to be received, the variable reactor of each stage comprising rotor and stator elementsof such character that the increments of reactance are equal for equal adjustments of the rotor with respect to the stator, and means mechanically connecting said rotors to insure that all of said rotors shall partake of equal extent of movement in adjustment from resonance. at one fre uency to resonance at v another frequency; a coupling system for tran'sferrin signal energy from a collector structure 0 arbitrary reactance to the tuned input circuit of the first of said stages, said coupling system including a reactor for connection in series with the collector structure, the said reactor of said coupling system being of the same type as said variable reactors of said tuned circuits and adjustable to a single definite value when the receiver is associated with any particular collector structure to insure that the tuned input circuit of said first stage will remain continuously in resonance with a succeeding stage as t e variable reactors of both stages are simultaneously adjusted by said connecting means to tune said stages over a band of frequencies. v

2. The invention as set forth in the next preceding claim, wherein said variable reactors and the said reactor of said coupling system are condensers.

In testimony whereof, I aflix my signature.

LEWIS M. HULL.

Images