US2250206A

US2250206A - Amplifying system

Info

Publication number: US2250206A
Application number: US366250A
Authority: US
Inventors: Charles S Root
Original assignee: General Electric Co
Current assignee: General Electric Co
Priority date: 1940-11-19
Filing date: 1940-11-19
Publication date: 1941-07-22
Anticipated expiration: 1958-07-22

Description

Jul 22, 1941 Q s, AQT 2,250,206

AMPLIFYING SYSTEM 2 Shegts-Sheet 1 Filed Nov. 19, 1940 RATIO OF APPLIED VOLTAGE AT CENTER FREQUENCY Inventor":

IO -5 0 +5 +10 by KILOCYCLE$ OFF RESONANCE Charles 5 Root,

I .July 22, T941. c. s. ROOT AMPLIFYING SYSTEM Filed Nov. 19, 1940 2 $heets-Sheet 2 LOWER I FREQ.

His Attorney.

FREQ

Inverfitort' "Charles S. Root by H Fig.9.

moucnv CAPACITIVE mnucgvs Fig. 8.

INDUCTIVE Patented July 22, 1941 AMPLIFYING SYSTEM Charles S. Root, Bridgeport, Conn, assignor to General Electric Company, a corporation of New York Application November 19, 1940, Serial No. 366,250

6 Claims.

My invention relates to amplifying systems and more particularly to amplifiers of high frequency potentials. In greater particularity, my invention relates to an improved and simplified arrangement for compensating, in an amplifier of this type, the effect of the retroactive currents which flow through the inherent anode-grid interelectrode capacity of an electron discharge device employed in high frequency signal amplifiers.

This application is a continuation-in-part of my application, Serial No. 223,179, filed August 5, 1938, now Patent 2,226,074, granted December 24, 1940, and assigned to the same assignee as the present invention.

One of the factors which limits the maximum amplification obtainable in a single stage of amplification is the regenerative action inherently present in the amplifier as a result of the interelectrode capacity, particularly the anode-grid capacity, of an electron discharge device employed in the amplifier. This interelectrode capacity serves to couple the output circuit of the discharge device to the input circuit of the device and eventually leads to uncontrollable free oscillation of the amplifier stage when an attempt is made to realize greater amplification beyond a certain point. Where two or more stages of high frequency amplification are used, sufficient amplification can usually be realized Without extending the amplification of any one stage into the regenerative region when a state of free oscillation is likely to occur. Where, however, it is possible to employ only a single stage of amplification, it has been common practice to allow a certain amount of regeneration to exist in order that the amplifier stage may produce a greater output since it has been found exceedingly difficult to reduce the regeneration sufficiently to avoid its undesirable efiects without greatly reducing the amplification in so doing. The reduction of regeneration, where accomplished, has been effected in the past by the use of interstage coupling transformers having a lower voltage step-up ratio or having lower impedance windings for the purpose of providing a poorer impedance matching of the transformer to the amplifier tube, or by biasing the tube to a point on its operating characteristic where the voltage amplification of the tube is greatly reduced. These methods of reducing the regeneration accomplish their desired end only by efiecting such a large reduction in the voltage gain between the input and output circuits of the electron discharge device that the feed-back of energy through the electron discharge device of the amplifier stage becomes negligible. Such methods, however, result in a very inefficient arrangement and a considerable loss of amplification.

The detrimental efiects of regeneration in an amplifier stage are at once evident when it is considered that the presence of regeneration causes first, a pronounced sharpening of the resonance curve of the emplifier, thereby to restrict the range of frequencies which may pass through the amplifier for a given selectivity; secondly, by greatly increasing the difliculty of alignment of intermediate frequency stages of amplification due to a certain amount of interaction between the several electron discharge devices with their interconnecting circuits, a condition which necessitates that all of the intermediate fre-- quency transformer trimmer condensers must be carefully adjusted and readjusted before a final balance of the amplifier stage may be accomplished; and thirdly, by greatly increasing the dimculty of properly aligning an intermediate frequency amplifier for both narrow and broad frequency band-pass operation where the amplifier stage has its frequency response expanded as by any of the several types of expanding systems now in common use.

Furthermore, it is well-known that tuning of a receiver in which regeneration exists is always accompanied by a characteristic swish noise and a high level of hiss background when listening to weak signals. There is also great danger of uncontrollable free oscillation of the amplifier when the amplifier tube (tor an associated component) happens to be at or near the maximum limit of the tolerance range of amplification that is necessary to allow the manufacture of radio tubes and components. Additional trouble occurs in large quantity production of radio receivers due to abnormal differences in gain and selectivity resulting from the large variations in regeneration caused by variations in manufactured parts, as mentioned above.

It is an object of my invention to provide an improved amplifier arrangement in which the regenerative action caused by the grid-anode interelectrode capacity of the amplifying electron discharge device may be completely compensated, even though the effect of such capacity may be small and therefore relatively difiicult to compensate exactly, as where a screen grid type of electron discharge device is employed.

A further object of my invention is to neutral- .ize or compensate in a high frequency amplifier a. cathode. l5.

the regenerative effect of the grid-anode interelectrode capacity by an improved arrangement which maintains the cathode of the discharge device at a slight radio frequency potential above ground by action of the output circuit radio frequency currents thereby toprovide a potential in the cathode circuit which may be returned to the input circuit as a degenerative potential.

Another object of my invention is to provide an improved intermediate frequency amplifier whose frequency band characteristic may be expanded over a wide range of expansion values without destroying the symmetry of the response curve of the amplifier aboutthe intermediate frequency. A further object of my invention is to provide an amplifier, particularly of the intermediate frequency type, having greatly improved selectivity and stability characteristics and one whose degree of amplification, or gain, is very materially increased over that obtainable by the prior art amplifier arrangements. 1 Still another and more specific object of my invention is to provide substantially exact neutraliz'ation of undesired regenerative voltages fed back upon the grid of a high frequency amplifier through the grid-anode interelectrode capacity by feeding back voltages, developed across a commoncathode impedance, which are of equal magnitudeand opposite phase with respect to the undesired voltages.

It is also specifically an object of my invention to provide an improved practical amplifier construction of this type wherein substantially exact neutralization may be obtained with inexpensive circuit elements of commercial quality and without recourse to expensive construction, careful shielding or critical adjustments.

The features of my invention which I believe to be novel are set forth with particularity in the appended claims. My invention itself, however, together with further objects and advantages thereof, may best be understood by reference to the following description taken in connection with the accompanying drawings, in which Fig. 1 diagrammatically illustrates an embodiment of my invention; Figs. 2' and .3 are equivalent circuit diagrams, Fig. 4 is a group of vector diagrams and Fig. 5 is a graph pertaining to the operation of my invention; Figs. 6 and '7 diagrammatically illustrate modified forms of my invention; Figs. 8 and 9 a re further diagrams and'Fig. 10 is another g'raph pertaining to the operation of my invention.

Referring now to Fig. 1 of the drawings, my invention is illustrated as embodied in an intermediate frequency amplifier stage of superheterodynera dio receiver. An electron discharge device ll] of the pentode type illustrated, having electrodes including a control grid II, a screen grid- 12, a suppressor grid I3, an anode l4, and Signal-modulated intermediate frequency oscillations are supplied through an input transformer It to the grid II and the cathode l5 .of the device H). The transformer [6 has a primary winding I! coupled to a source of intermediate frequency oscillations, not shown, which may -for example be a converter stage or a prior intermediate frequency amplifier stage in the usual superheterodyne receiver.

The secondary winding I8 .of the input transformer lfi'has one end connected to the grid H and the other end connected to the cathode l5 through a blocking capacitor 1 9 and an inductive impedance element '21]. T-helower end of the impedane element 213 is preferably maintained at a relatively fixed reference potential by a connection to ground, as shown. The primary and secondary windings l1 and I8 are respectively tuned to the frequency of the intermediate frequency oscillations by the small trimmer capacitors 2| and 22. The capacitor 22 is connected directly between the grid II and the cathode 15.

A tertiary winding 23 is provided onthe input transformer M5 for the purpose of varying the frequency band response of the amplifier. When the switch 24 is in the upper position as shown, the winding 23 is excluded from the grid circuit. Under these conditions the windings I! and [8 are so coupled that the amplifier is highly selective. The winding 23 is arranged to overcouple the secondary circuit with the primary circuit of the transformer l 5 when the switch 24 is manually moved from the upper switch contact 25 to the lower switch contact 26. Overcoupling the windings in this manner provides abroad frequency band response with resulting high fidelity operation of the amplifier and minimizes distortion associated with inaccuracies of present-day mechanical tuning of the receiver input circuit, as will readily be understood by those skilled in the art. The introduction of the winding 23 into the tuned secondary circuit in this manner does not appreciably detune this circuit since the inductance of the winding 23 is relatively insignificant in comparison to the much larger inniiuctance of the transformer secondary winding The output circuit of the electron discharge device [0 is connected between the anode l4 and the cathode l5. This includes the primary Winding 28 of an output intermediate frequency transformer 21 and a source of anode potential, not shown, whose positive terminal is connected to the conductor 30 and whose negtaive terminal is connected to ground. The transformer 21 has a secondary winding 29 connected to supply the amplified'intermediate frequency oscillations to a utilization circuit, not shown, which may serially include further stages of amplification, a demodulating device, an amplifier of the modula- 1 I2 througha resistor 34 connected to the positive terminal of the source of anode potential. A bypass capacitor 35 maintains the screen grid at ground potential for currents of intermediate frequency.

The control grid H of the device It has a normal operating bias applied thereto from any suitable unidirectional potential source, not.

shown, having one terminal connected to the conductor 36 and the otherterminal to ground. The control grid I! may additionally have impressed thereon an automatic volume control bias supplied from a demodulator device, not shown, which demodulates the output of the intermediate frequency stage of amplification and supplies a potential to the conductor 36 whose magnitude varies with the average amplitude of the intermediate frequency oscillations. The purpose is to maintain the output of the amplifier substantially constant notwithstandingthe fact that the input signals may vary over a wide range of signal strength as is well understood in the art. The several interelectrode capacities of the electron discharge device are represented in Fig. 1 by the broken lines. The grid to cathode capacity is represented by the capacity CGc, the anode to grid capacity by the capacity Cee and the anode to cathode capacity by the capacity CPc. If the capacity CPG could be made zero, the output circuit of the discharge device l could be completely isolated from the input circuit of the device and the amplifier stage could be operated at its maximum possible amplification without experiencing the detrimental effects of regenerative feedback. However, much the capacity CPG may be reduced by the use of the screen grid i2, there nevertheless exists a small capacity between the anode and grid elements of electron discharge devices customarily used at the present time. The regenerative feed-back of energy through the capacitor Cm, unless compensated by a corresponding degenerative feed-back of energy in the manner of my invention, not only prohibits the attainment of the maximum amplification which is otherwise possible in the amplifier arrangement, but additionally introduces into the operation of the amplifier the detrimental effects of regeneration considered heretofore.

It will be observed that the anode circuit for the device includes the output tank circuit 28, 3i, and the inductive element in series. It is important to note that the anode circuit has in shunt thereto the anode to cathode capacity CPc of the device l0. Therefore, when the primary circuit of the transformer 21 is tuned to resonance at the operating frequency of the amplifier, the net reactance of the tank circuit 28, 3|, alone is inductive since it is deficient in capacity by the small capacity Cpc. The anode circuit may therefore be considered as a voltage divider for the intermediate frequency potentials developed in the anode circuit, having two inductive elements in series.

The inductive impedance 20 is also connected in series with the Winding l8 and the capacitor 22 in the grid tank circuit. The radio frequency potentials developed on impedance 20 by the fiow of anode current therethrough are consequently impressed on the grid ll through the winding l8, which is also highly inductive. As is fully explained in my aforesaid Patent No. 2,226,074, December 24, 1940, and as will be further developed herein, these circuit connections provide a degenerative feed-back to the grid ll. If the magnitude and phase of the potentials which are fed back to the grid II through the impedance 20 are properly adjusted with respect to the regenerative potentials fed back to the grid H from the anode I3 through the inherent capacity CPG, the regenerative potentials are neutralized or compensated.

In the usual design of an amplifier of this type, the windings of the transformers l5 and 21 are constructed so as to have a relatively high ratio of reactance to resistance at the operating frequency. This ratio is generally designated by the symbol Q. Coils of high Q are desirable from the standpoint of improved selectivityand gain in the amplifier, as will be apparent to those skilled in the art without further explanation.

In the practical design of an amplifier built in accordance with the teachings of my Patent No. 2,226,074, December 24, 1940, the resistances present in. the "circuits have heretofore been considered to be negligible, in comparison to the inductive and capacitive reactances at the intermediate frequency, for the ordinary case of an amplifier having tank circuits of reasonably high Q. By making this assumption the practical determination of the circuit constants is simplified considerably since only pure reactances need be considered. That this assumption is not unreasonable has been amply demon-. strated by the fact that amplifiers built in accordance with these teachings have gone into extensive use with very satisfactory results. From this it might be concluded that the higher the Q of the circuit elements the more nearly theoretically perfect neutralization can be approached. Nevertheless, even with the use of relatively expensive coils of high Q for the inductive elements of the circuit branches, the actual results have been found to fall somewhat short of practical realization of the benefits which should be attainable. The reason for this is that even with high Q coils it has heretofore been found necessary to provide a greater degree of neutralization than is theoretically necessary to remove all the eifects of regeneration. This over-neutralization, or degeneration, reduces the amplifier gain too much, so that lesser degeneration is necessary, which in turn increases the diificulty of aligning the tuned circuits and causes dissymmetry of the frequency response curve, especially when band expansion is employed. Thus, many of the deleterious effects of regeneration are not completely eliminated unless an excessive loss of gain is tolerated.

Further investigation has revealed that substantially exact neutralization, together with surprising improvements in amplifier operating characteristics, can be obtained in a practical amplifier construction, if the resistance of these circuits is taken into consideration. It has been discovered that the apparent resistance of the circuit branches is in fact not negligible even though the individual elements may have a reasonably high Q. Surprisingly enough, it has been found, if the resistance is considered, it is not necessary to employ expensive circuit elements of high Q. Rather, practical attainment of these materially improved results requires only that the input and output transformer windings be of commercial quality; and even more unexpected is the fact that certain of the other circuit elements must then necessarily be of inexpensive, low-Q construction. These facts will become apparent from a consideration of the circuit analysis now to be developed in detail.

The alternating current circuits of the amplifier of Fig. 1 are represented in simplified form by the equivalent circuit diagram of Fig. 2. The tank circuit impedance between anode l4 and ground is represented by the two circuit branches within the dashed rectangle Z1. The branch comprising the winding 28 has both inductive reactance and resistance at the operating fre quency and is shunted by the branch comprising the capacitive reactance provided by the capacitor 3|, the distributed coil capacity and various stray capacities of the elements and connecting leads with respect to ground. The elements within the rectangle Z1 are thus to be taken to include all the electrical constants, lumped and distributed, which make up the terminal impedance seen looking into the output tank circuit. As previously mentioned, this cireuit has a net inductive reactance at the operating frequency of the amplifier since it is tuned to resonance by the small anode to cathode capacity CPC. As is well known, an LC circuit which is very near resonance becomes, in effect, a very large resistance in series with an approximately equal value of reactance, in this case an inductive reactance. Thus, Z1 has a very large terminal resistance and also an approximately equal value of terminal inductive reactance. Consequently, even though the winding 28 alone may be of fairly high Q, the net impedance Z1 is of relatively low Q because the reactance and resistance are approximately equal and the resistance component thus cannot be considered negligible.

The impedance between cathode I and ground is represented in Fig. 2 by the inductance and resistance within the rectangle Z2. The anode to grid capacity CPG, which is practically pure capacitive reactance, is comprised by Z3. The net terminal impedance of the input winding l8 between grid H and ground, taking into account its inductance, resistance and distributed capacity, comprises inductive reactance and resistance, as indicated within the rectangle Z4. The grid to cathode impedance ZG includes the trimmer capacitor 22 and the inherent grid to cathode capacity C00 in parallel and is practically pure capacitive reactance. Thus, the inductive and capacitive branches of the input tank circuit are comprised by the impedance Z4 and ZG, respectively. The impedances of the capacitors l9 and 33 are so low at the operating frequency that they can be'neglected.

For the purposes of mathematical analysis the equivalent circuit diagram of Fig. 2 can be redrawn in generalized schematic form as the fourterminal network of Fig. 3. This figure represents the impedance network looking from the output circuit of the amplifier to the input circuit. The voltage EP appearing between the anode P and the cathode C is fed back through the impedance network comprised by Z1, Z2, Z3

vector quantity. Otherwise a symbol denotes a scalar quantity.

Expanding Equation 3 and substituting values from Equations 1 and 2:

i (4) Z1Z3 Z 1 G Also:

3 1) 2+( a 2)Z 4-i-EF=0 (5) Substituting values in Equation 5 from Equations 1, 2 and 4 and simplifying, the general equation for feedback voltage Eli is:

In order to satisfy the requirement that Er be equal to zero, set the numerator of Equation 6 equal to zero, from which l+ 3+ 4 (7) Equation '7 expresses, in general terms, the relationship between Z2 and the other network constants which will provide exact neutralization in accordance with my invention. In order to obtain a solution for the particular circuit of Figs. 1 and 2, the next step is to substitute the vector quantities for the various impedances. In this particlar case, these quantities are:

Z =R1+ 7 X (resistive and inductive) 2 R2 +j X 2 (resistive and inductive) 2 j X 3 (capacitive only) Z =R j X 4 (resistive and inductive) Upon substitution of these quantities in Equation and X2 are:

R1+R. +(X1+X.-X3 (8) X2: [(R4X1+R1X4) (R1-i- 4)] 4 1- 4 1) 1-i- 4 a)] (9) l'i' 4) l+ 4 3) and Z4 to the grid G. The condition for neu- Equations 8 and 9 are cumbersome. However,

tralization is that EF, the net resultant feedback tween the grid G and the cathode C, shall be equal to zero for any value of EP at the operating frequency of the amplifier. In other words, for exact neutralization the network of Fig. 3 is an infinite attenuation network having infinite attenuation between input terminals P and C and the output terminals G and C at the frequency to be amplified.

The first step toward determining the constants of the network which will satisfy this condition is to obtain a solution for Er in terms of EP. This solution may readily be obtained by application of Kirchhoifs laws to the network. For this purpose, assume the three clockwise circulating currents I1, I2 and I3, as shown in Fig. 3. In the following three equations, which may be written immediately from an application of Kirchhoifs laws, and in the subsequent equations, it is to be understood that a dot placed over a symbol indicates that it is by makingcertain assumptions, which are those generally true in an amplifier of this type, they can be much simplified. The first of these as,- sllmptions is that the Q of the windings l8 and 28 is reasonably high. Most coils used in amplifiers of this type have a Q of at least 50, and usually higher, which is sufficient to meet this requirement. The second assumption is that the reactance X3 of the anode to grid capacity CPG is very large at the operating frequency as compared to the reactances of the windings I 8 and 28. This is always true, and X3 is especially large when the device NJ is provided with a screen grid. If these assumptions are made, then the following expressions hold good, with negligible error, in a practical embodiment of the amplifier of Fig. 1: I

R4X1+R1X4 R X X R R Xz (.11)

where R1 and X1=the resistance and reactance between anode and ground,

R2 and X2=the resistance and reactance between cathode and ground,

X3=the reactance of the anode to grid capacity CPG, and 7 R4 and X4=the resistance and reactance between grid and ground.

By dividing Equation 11 by Equation 10, the following single expression may also be obtained for the practical amplifier of Fig. 1:

where Q1, Q2 and Q4=the ratiosof reactance to resistance of Z1, Z2 and Z4 at the operating frequency.

The vector diagram of Fig. 4 further illustrates graphically the relationship between the voltages in the several amplifier branches for the condition of exact neutralization. Fig. 4a is a vector diagram of the voltages between the anode P and cathode C. It will be seen that the voltage Ep is equal to the vector sum of the voltage drops 1Z1 and 1Z2, from anode to ground and from ground to cathode, respectively. Each of these vectors may further be resolved into the resistance and reactance drops through impedances Z1 and Z2. These vectors also show qualitatively that the Q of Z2 is somewhat less than the Q of Z1. As previously mentioned, the Q of the latter impedance will be quite low since the output tank circuit is very near resonance. Therefore, the Q of the cathode impedance Z2 will also be low.

The vector diagram of Fig. 4?) represents the voltages fed back from the cathode to the grid from the cathode impedance Z2. The voltage drop I Z2 produces the drops IE4 and IX4 through the impedance Z4. Since the grid to cathode impedance CGc is almost purely capacitive, the drop IZG is in phase opposition to the 1X4 drop.

The vector diagram of Fig. 40 represents the voltages fed back from the anode to the grid through the anode to grid impedance Z3. Since the impedances Z3 and Ze are both almost purely capacitive, the voltages 1 3 and Ez across them will be in phase, as shown.

If the circuit constants are selected in accordance with the equations previously developed, then the voltages IZG and 132 are exactly equal and in phase opposition. Figs. 4b and 4c illustrate this condition. The net voltage between the grid and cathode due to EP will therefore be zero.

In a typical amplifier of th type shown in Fig. 1 the intermediate frequency might be equal to 455 kc., the capacity CPG equal to .005 mmf., the anode to cathode capacity to Cpo equal to 5 mmf., the inductance of the windings l8 and 28 each equal to 2.45 mh., and the Q of each winding equal to 100. Using these values merely for illustration, calculations will readily show that R1==about 8,400 ohms X1=about 76,000 ohms Ri about '70 ohms X4=about 7,000 ohms Xz=about 70 megohms.

If these values are substituted in the simplified

practical Equations

10, 11 and 12, it will be found that substantially exact neutralization is achieved when R2 equals 0.9 ohm and X2 equals 7.6 ohms, approximately. Thus, the Q of the impedance 20 will be equal to about 8.4 which is quite low. This is entirely consistent with the preceding analysis since the Q of Z1 is also low, being equal to about 9.0 in this case.

The graph of Fig. 5 shows selectivity curves for a typical superheterodyne receiver whose intermediate frequency amplifier circuits were provided with the neutralizing connections of Fig, 1. When windings I8, 20 and 28 were all designed to have high Q values, th curve A was secured. When the resistance of winding 20 was increased, in accordance with the present invention, so as to secure the proper phase relations between the feedback voltages, curve B resulted. It will be observed that the curve B provides a wider band pass for the same selectivity at 10 kc. from the center frequency, increasing the fidelity of audio reproduction. Stated in the converse, my invention therefore permits the receiver to be designed for greater selectivity with the same band pass characteristics. Furthermore, when the circuit constants were properly selected, it was no longer necessary to over-neutralize to remove all regenerative efiects completely. Consequently, in this particular receiver a gain in sensitivity of about 1.5 times, or per cent, was also achieved. The fact that the receiver fidelity and sensitivity were simultaneously improved to a marked extent for substantially the same selectivity demonstrates concretely the practical merit of my invention.

In my aforesaid Patent No. 2,226,074, December 24, 1940, a modified form of amplifier has been shown in which the common impedance between cathode and ground comprises a self bias resistor and a self bias capacitor in parallel. An amplifier of this form, which may be designed to give substantially exact neutralization in accordance with the principles of my present invention, is illustrated in Fig. 6. In many respects the circuit is similar to that of Fig. 1 and corresponding elements have been designated by corresponding reference numerals. The inductive impedance 2|] is replaced by the resistor 31 and capacitor 38 in parallel. Th s necessitates certain changes in the manner of connecting the input circuit to the cathod I5 if the proper phase relationship between the regenerative and degenerative voltages is to be maintained. In this embodiment the trimmer capacitor 22 is connected between the grid H and ground, and the winding I8 is connected between the grid II and the cathode l5 through a blocking capacitor 40. For reasons that will shortly be apparent it is also generally necessary to include a resistor 39 in series with the capacitor 22, if exact neutralization is to be obtained in accordance with my invention.

The equivalent alternating current circuit diagram for Fig, 6 is illustrated in Fig. 8. This diagram is similar to that of Fig. 2 and corresponding symbols have been used. The cathod to ground impedance Z2 now consists of equivalent series resistance and capacitive reactance. The grid to ground impedance Z4 is also resistive and capacitive and the net grid to cathode impedance Ze is resistive and inductive, since the inductance of the winding 18 predominates over the capacity in shunt thereto.

The analysis previously developed for the general network of Fig. 3 applies equally well to this amplifier circuit and Equation 7 expresses the The Q of the cathode impedance Z2 is expressed by It will be observed that

Equations

13, 14 and 15 are in the same general form as

Equations

10, 11 and 12 except for changes of sign. This makes one important practical difference. From Equation 13 it will be seen that R4X1 must be at least equal to R1X4. Otherwise R2 will be a negative resistance, which is not physically realizable. Expressed another way, the ratio of X1 to R1, or Q1, be equal to or greater than the ratio of Xrto or Q4. This same fact is also apparent from inspection of the denominator of Equation 15. Q1 is quite low, since the output circuit comprised by Z1 is very near resonance, as previously explained. Since the trimmer capacity 22, which provides almost all of the reactive component of Z4, is of high Q, i. e., of very low resistance, the external resistance 39 must generally be inserted to make Q4 at least as low as Q1.

l urther analysis will show that the resistance of resistor 39 may necessarily be of considerable magnitude for exact neutralization in an amplifierhavin'g typical circuit constants. Since this resistance is included directly in the grid tan-k circuit, the grid circuit is damped and neutralization is secured only at the expense. of loss of gain and selectivity in the ordinary case.

The circuit of Fig. '7 overcomes these disadvantages and still permits substantially exact neutrali'zati'on with the resistor 31 and the capacitor 38 in the cathode circuit. The circuit of Fig. '7 differs from that of Fig. 6 in that the lower end of the winding I8 and the lower terminal of the trimmer capacitor 22 are both connected to, ground through the blocking capacitor I9. A separate inductive impedance 4| completes the grid tank'circuit from .the grid I l to the cathode [5 through the blocking capacitor 48.

The primary winding 11 may be coupledrto the winding l8, as shown, or optionally coupled to the winding'dl Also, since'the lower ends of both windings I 8 and M are isolated from the cathode circuit for direct current by the blocking capacitors l9 and 40, A. V. C. potentials may be supplied to the grid H over the conductor 36, or optionally over the conductor 3 as shown in broken lines.

1 59 represents the equivalent alternating current circuit for Fig. '7. It is electrically the substantial equivalent ofFig. 8. The closed tank circuit I8, 22 of Fig. '7 forms the capacitive branch Z of the input network in this case. It is "tuned to resonance at the operating frequency of the amplifier by the inductive reactance.

of the impedance 4| which forms the inductive branch Ze If the impedance 4! has large inductanca its reactance at the operating frequency will be large and the

tank circuit

18, 22

will be very near to resonance. Therefore, the tank circuit I8, 22, will be capacitive, because it lacks the inductance 4|, and highly resistive, because it is so near resonance. Thus, for the same reasons previously outlined for the case of Z1, the net terminal impedance Z4 between grid and ground will have a relatively low Q, because the resistance is approximately equal to the reactance, which is capacitive in this case. Equation 7 and the simplified

practical Equations

13, 14 and 15 can therefore be satisfied in this circuit so as to make R2 positive without actually damping the grid tank circuit by the addition of external resistance. The proper phase shift through Z4 for exact neutralization is thus secured without adverse effects upon the selectivity and gain of the amplifier.

It will be observed from an inspectiton of Figs. 2, 8 and 9, that Z1 is always inductive since it is tuned to resonance at the operating frequency of the system by the anode to cathode capacity CH). The impedances Z4 and Zc are always of opposite sign, since Z4 is included in one of the two parallel branches between the grid and cathode of the amplifier, and Z6 is included in the opposite branch, the two branches being tuned to resonance at the operating frequency, Finally, it will be noted that the common cathode impedance Z2 must be of the same sign as Z4 if the proper phase relations for neutralization are to be secured. If Z2 and Z4 are both inductive,

Equations

10, 11, and 12 can be utilized for the practical case where the simplifying assumptions introduce negligible error. If these impedances are both capacitive, then

Equations

13, 14, and 15 should be used for practical design calculations. In either case Equation '7 expresses, in general terms, the exact relationships of circuit constants in terms of the impedance vectors involved.

It will now be evident that I have accomplished by my invention the very complete and effective control of the magnitude of the regenerativeaction experienced in the prior art amplifier arrangernents. The advantages which follow from r the use of my invention are manifold. Fig. 10'is a graph illustrating, in a manner well known in the art, the frequency response of an intermediate frequency stage of amplification. The curve (1 represents the frequency response of a highly selective amplifier which, being selective, may completely fail to pass the higher audio frequencies. It has become customary to provide in high fidelity radio receivers or in mechanically tuned receivers some form of frequency response expanding arrangement whereby the frequency response of the amplifier may be expanded in a manner to allow the passage of the higher audio frequencies to obtain high fidelity reproduction of the modulated intermediate frequency oscillation whenever receiving conditions render the use of a highly selective arrangement unnecessary. One such arrangement, comprising the tertiary winding 23 and associated switching means, has been illustrated and described herein. These receivers are aligned in the narrow band-pass position to have the highly selective frequency response represented by the curve a. When, however, the frequency response is expanded in these receivers, the prior art amplifier arrangements no longer have a symmetrical response curve about the intermediate frequency (represented by the vertical line of Fig 10) but have, by virtue of the regenerative action present in the prior art amplifiers, a distorted response curve with an abnormal double hump which maybe represented by the broken line b of Fig. 10. The peak at c is higher than at d because the grid to anode capacity (Cm in Figs. 1, 6 and 7) produces greater regenerative feed-back at higher frequencies. This causes considerable distortion of the audio frequencies, thereby to produce a highly undesirable condition of operation. In addition, the operator is more apt to tune the receiver to the peak than to the intermediate frequency, causing further distortion. My invention, by eliminating the effect of the regenerative action, not only renders the response curve of the amplifier symmetrical about the intermediate frequency throughout a large range of values of frequency expansion, but also flattens off the top of the response curve by reducing any double hump which would otherwise be present (according to the well-known tendency of degeneration to flatten off a frequency response characteristic of an amplifier), giving a frequency response curve as shown by the full line e in Fig. 10.

The neutralization of the regenerative currents in the amplifier allows the realization of the highest possible amplification or gain for each stage of amplification since the amplifier no longer has a tendency to break into free oscillation. The gain or amplification obtainable in an amplifier arrangement embodying my invention is. limited only by the design of the amplifier input and output transformers and by the maximum amplification obtainable with the particular electron discharge device used.

My invention has the further advantage that the degenerative voltage not only originates in but is utilized in each individual amplifier stage and, therefore, the degenerative circuit does not include interstage coupling transformers Whose value of magnetic coupling may be changed by expansion of the frequency band response. The neutralization of the regeneration effected within a single stage has an important advantage in that it is unnecessary to consider and to cope with phase shifts between the regenerative and degenerative currents where the degenerative voltage originates in an amplifier stage following or preceding the neutralized amplifier stage. A phase shift of this nature may perhaps be corrected for a given frequency band response of the amplifier at a given amplifier output, but the correction is improper for other values of frequency band response or for higher or lower amplification since changes in either the frequency response or the amplification results in corresponding changes in the phase and magnitude between regenerative and degenerative currents. For high fidelity reproduction, this requires that the intermediate frequency stages of amplification in the prior art arrangements be realigned for each and every value of frequency response expansion. The operating conditions of an amplifier embodying my invention do not affect the symmetry of the frequency band response about the intermediate frequency since the frequency response is unaffected throughout a large range of values of frequency expansion and is independent of the power output of the amplifier.

My amplifier arrangement has another important advantage. In the design of the prior art high gain single stage amplifiers which, at best, include as much regeneration due to grid to anode interelectrode capacity as can safely be tolerated, it becomes necessary to resort to great expense and trouble in reducing all other stray or circuit regeneration to an absolute minimum,

often at a sacrifice of flexibility of receiver layout and even of receiver performance. With my invention, such elaborate precautions are no longer necessary since slight amounts of stray regeneration can be tolerated, and my invention may be utilized to compensate both the additional regeneration and that introduced by the grid to plate capacity. This, of course, requires the furnishing of degenerative currents in excess of those which would normally be required.

Previous neutralizing circuits have involved the added cost and complexity of an additional neutralizing winding or critically positioned tap on the transformer and a small, critical neutralizing capacitor whose leads were at a high radio frequency potential above ground and therefore inherently non-stable. By contrast, my invention does not require any additional components over and above those required for prior art amplifiers but uses stable, low-potential cathode circuit components.

While I have shown particular embodiments of my invention, it will of course be understood that I do not wish to be limited thereto since various modifications may be made, and I contemplate by the appended claims to cover any such modifications as fall within the true spirit and scope of my invention.

What I claim as new and desire to secure by Letters Patent of the United States is:

1. In combination, an amplifier having a cathode, a grid and an anode, an output circuit having one terminal connected to said anode and the other terminal connected to said cathode through an impendance element, said output circuit having a net terminal impedance Z1 and said element having an impedance Zz at the operating frequency of said amplifier, each of said impedances having a relatively low ratio of reactance to resistance at said frequency, and an input circuit having a capacitive branch and an inductive branch, one of said branches being connected between said grid and said cathode, the other of said branches being connected between said grid and the point between said output circuit and said impedance element, said latter branch having a net terminal impedance Z4 at said operating frequency," the relationship between said impedances and the impedance Z3 due to the capacity between said anode and said grid being substantially expressed by the equation:

2. In a high frequency amplifying system, an electron discharge device having an anode, a cathode and a grid, an output circuit having a net terminal inductive reactance X1 and a resistance R1 at the operating frequency of said system, an impedance having a reactance X2 of one sign and a resistance R2 at said frequency, said output circuit having one terminal connected to said anode and the other terminal thereof connected to said cathode through said impedance, and an input circuit having an inductive branch and a capacitive branch, one of said branches having a terminal reactance X4 of the same sign as said impedance and a resistance R4, said one branch being connected between said grid and the point between said output circuit and said impedance, the other of said branches being connected between said grid and said cathode, the relationship between said impedance, said circuits and the inherent anode to: grid-capacitive reactance X: of said device being substantially expressed by the equations:

3. In combination, an amplifier having an anode,.a cathodesand a grid, an'output circuit comprising an inductive branch and a capacitive branch: and havinga. net terminal inductive re- ELCtELI'ICQXl and a resistance'Ri at the operating frequency of said amplifier, an impedance having an inductive reactance. X2 and a resistance R2 at said frequency, said output circuit and said impedance being serially connected between said anode and said cathode, and an input circuit comprising a second capacitive branch connected between said grid and cathode and asecnd inductive branch having reactance X4 and resistance R4 at said frequency connected between said grid and the point between said output circuit and said impedance, said impedance being between said point and said cathode, both of said inductive branches having a relatively high ratio of reactance to resistance at said frequency, the relationship between said impedance, said circuits and the inherent anode to grid capacitive reactance X3 of said amplifier at said frequency being substantially expressed by the equations:

4. In combination, an amplifier having an anode, a cathode and a grid, an output circuit comprising an inductive branch and a capacitive branch and having a net terminal inductive reactance X1 and a resistance R1 at the operating frequency of said amplifier, an impedance having capacitive reactance X2 and a resistance R2 at said frequency, said output circuit and said impedance being serially connected between said anode and said cathode, and an input circuit comprising a second inductive branch connected betweensaid grid and said cathode and a second capacitive branch having capacitive reactance X4 and a resistance R4 at said frequency connected between said grid and the point between said output circuit and said impedance, said impedance being between said point and said cathode,

bothof said inductive branches having a relatively high ratio of reactance to resistance at said frequency, the relationship between said impedance, 'said circuits and the inherent anode to grid capacitive reactance X3 of said amplifier at said frequency being substantially expressed by the equations:

: RX-RX /R2=' 4 1X31 4 and.-

5. A tuned high frequency amplifying system comprising in combination, an electron discharge amplifier having electrodes including an anode, acathode and a control grid, said electrodes having inherent capacity therebetween, a tunable output circuit comprising an inductance and gether with said third inductance, and means to capacitance connected in parallel between saidanode and ground, said circuit having a net terminal inductive reactance X1 and a resistance R1 when tuned to the operating frequency of said amplifier in conjunction with the inherent anode to cathode capacity, an inductive impedance having a reactance X2 and a resistance R2 connected between said cathode and ground, a tunable input circuit comprising a second inductance connected between said grid and ground and a second capacitance connected between said grid and cathode, said second inductance having a reactance X4 and a resistance R; when said input circuit is tuned to said operating frequency, and means to impress potentials of said operating frequency upon said input circuit, said inductances individually having relatively high ratios of reactance to resistance at said operating frequency, the relationship between said reactances, resistances, and the reactance X3 due to the inherent capacity between said anode and said grid being substantially expressed by the equations:

whereby potential variations at said frequency between said anode and cathode produce substantially no potential variations between said grid and cathode.

6. A tuned high frequency amplifying system comprising, in combination, an electron discharge amplifier having electrodes including an anode, a cathode and a control grid, said electrodes having inherent capacity therebetween, a tunable output circuit comprising an inductance and capacitance connected in parallel between said anode and ground, saidcircuit having a net terminal inductive reactance X1 and a resistance R1 when tuned to the operating frequency of said amplifier in conjunction with the inherent anode to cathode capacity, a capacitive impedance having a reactance X2 and a resistance R2 connected between said cathode and ground, a tunable input circuit comprising a second inductance and a second capacitance connected in parallel between said grid and ground and a third inductance connected between said grid and said cathode, said second inductance and second capacitance having a net terminal capacitive reactance X4 and a resistance R4 when tuned to the operating frequency of said amplifier toimpress potentials of said operating frequency upon said input circuit, all said inductances individually having relatively high ratios of reactance to resistance at said operating frequency, the relationship between said reactances, resistances and the reactance X3 due to the inherent capacity between said anode and grid being substantially expressed by the equations:

CHARLES S. RQOT.

CERTIFICATE OF CORRECTION.

Patent No. 2,2 0,206. Y Y Jul 22, 19m.

- CHARLES s. ROOT. g

It is hereby certified that error appears in theprinted specification of the above numbered patent requiring correction as follows: Pagel, second column, linelO, for "amplifier" re ad -aJnplifier--; page 2, second col. umn, line 59, for "negtaive" read nega,tive-; page 14,-, first column, line 55, for the word 'impedance" read -impedances-; and second column, line 15, Equation ii, for "E read line 58,,for "particlar" read -par ticularpage 5, first column, line 65, strike out "to" after "capacity";

O page 6, second column, 'line l8, for "inspectiton" read --inspection--; page Y, second column, line 55, claim 1, for impndance read --impedance and that the said Letters Patent shouldbe read with this correction therein that the same magi conform to the record of the case in the Patent Office.

Signed and sealed this 16th day of September, A. D. 19m.

, Henry Van Arsdale, (Seal) Acting Commissioner of Patents.