US2477074A

US2477074A - Wide band amplifier coupling circuits

Info

Publication number: US2477074A
Application number: US66741A
Authority: US
Inventors: Frank H Mcintosh
Original assignee: Individual
Current assignee: Individual
Priority date: 1948-12-22
Filing date: 1948-12-22
Publication date: 1949-07-26
Anticipated expiration: 1966-07-26
Also published as: CH296850A; FR1002802A; DE842502C; GB677921A; NL150547B; NL79816C

Description

July 26, 19.49- F. H. M INTOSH WIDE-BAND AHPLIFIER COUPLING CIRCUITS 4 Sheets-Sheet 2 Filed D80. 22, 1948 ,1 N VEN TOR.

FRANK H. Me I NTOSH Ju1y26,1949.' F. H. WINTOSH 2,477,014

WIDE-BAND AMPLIFIER COUPLING CIRCUITS Filed Dec. 22, 1948 4 Sheets-Sheet Z :JINJL 262 27 7 PM- I 14 INVENTOR.

FRANK H. McINTOSH y 1949- v F. H. MGINTOSH 2,477,074

WIDE-BAND AMPLIFIER COUPLING CIRCUITS Filed Dec. 22, 1948 4 Sheets-Sheet s IN V EN TOR.

I FRANK a. McINTOSH July 26, 1949. mm'rosH 2,477,074

WIDE-BAND AMPLIFIER COUPLING CIRCUITS Filed Dec. 22, 14s

4 Sheets-Sheet 4 INVEN TOR.

FRANK H. MC INTOSH Patented July 26, 1949 UNITED STATES PATENT OFFICE WIDE BAND AMPLIFIER COUPLING CIRCUITS.

The present invention relates generally to im- 2 proved audio and video frequency amplifiers, and to transformers utilizable therein, and more particularly to improved class B audio and video frequency electronic amplifiers which introduce extremely slight distortion over a wide band of frequencies, by utilizing output transformers of novel design, connected in novel relation to the electronic tubes of the amplifier.

The class B amplifier is a push-pull amplifier in which the tubes are biased approximately to cut-off. One of the tubes, in the normal system, amplifies the positive half cycles of the signal voltage while the other amplifies the negative half cycles, the output transformer combining the outputs of the two tubes, to reconstruct a replica of the signal voltage.

The frequency limits of the conventional audio or video amplifier depend largely upon the design of the output transformer, loss in amplification at low frequencies resulting from the low incremental inductance of the transformer primary, and falling off at high frequencies resulting from leakage inductance and the various distributed capacities of the transformer.

In order to obtain a good low frequency response the incremental primary inductance of the transformer must be high relative to the plate resistances of the tubes used. The primary winding of the transformer, then, should have a large number of turns. At the same time the resonant frequency of the leakage inductance and secondary capacitance must be beyond the highest frequency desired to be amplified, so that low leakage inductance and shunt capacity is essential, if the frequency response of the transformer is to be extended.

The above requirements are mutually conflicting, in various respects. The size of the core of a transformer, i. e., the total iron utilized, is limited by considerations of cost, space and weight requirements. This in turn fixes the total number of turns allotted to the primary and secondary windings. Decreasing core size and increasing total turns on the primary winding to retain high primary incremental inductance increases leakage inductance and shunt capacity, which in turn, reduces resonant frequency, and hence the high frequency response of the transformer. In practice, leakage inductance is decreased by interleaving primary and secondary windings, but this increases distributed capacity and so tends to neutralize the benefits obtained.

As a further consideration, high permeability cores must be used, to increase primary winding 9 Claims. (Cl. 179-171) impedance. Such cores are adversely affected, in respect to the incremental inductance, by D. C. magnetization. Hence the latter must be avoided. The effect of leakage inductance on class B push-pull amplifiers has been considered in the literature, and attention is directed particularly to an article by A. Pen-Tung Sah, in Proceedings of the I. R. E. for November, 1936. Sah points out particularly the deleterious effects of leakage inductance between primary windings of the output transformer of such an amplifier, first, in causing a decreased output, as frequency increases, and second, in introducing finite time constants into the circuit, thus causing transients which distort the output wave as one of the tubes changes from a conducting condition to a blocking condition, and vice-versa. The latter effect is the basis of great distortion at the higher audio frequenmes.

It is a primary object of the present invention to provide improved push-pull amplifiers having negligible leakage reactance in their output transformers, and hence negligible transient effects during change-over of each tube of the amplifier 25 from conducting to non-conducting condition.

It is an ancillary object of the invention to provide novel push-pull transformers having negligible leakage reactance.

It is a further object of the invention. to provide a push-pull wide band transformer of relativelysimple and economical construction, which eliminates leakage inductance between primary windings of the transformer.

It is another object of the invention to provide an improved push-pull transformer comprising bifilar primary windings, and futher to provide push-pull audio amplifiers capable of employing transformer having bifilar primary windings,

It is, further, an object of the invention to provide a push-pull transformer, having greater coupling between the secondary winding and the primary windings than is available in known designs, without increasing detrimental capacity effects, thereby to improve the frequency response and to enlarge the band width of such transformers when employed in amplifiers.

It is still another object of the invention to provide a push-pull transformer having radically reduced effective distributed capacity across the primary windings, and to provide a push-pull amplifier for effectively utilizing a transformer of this character.

It is a further object of the invention to provide a push-pull transformer of reduced distribpower electronic tubes, wherein is provided means for maintaining the screen grid of each of the tubes at a fixed potential with respect to the associated cathode during operation of the tubes in the amplifier.

It is still another object of the invention to provide a push-pull amplifier arrangement capable of effectively utilizing a transformer having substantially zero leakage inductance between its primary windings, and which requires but a single anode voltage source for all the tubes of the It is a further object of the invention to provide novel push-pull transformer arrangements which are not bifilarly wound but which have many of the properties of bifilarly wound transformers, and particularly low or negligible leakage inductance between primary windings.

It is still another object of the invention to provide a modulator capable of supplying large amounts of undistorted power for modulating carrier frequency signals.

Briefly described, the various embodiments of the present invention hereinafter described in detail, and illustrated in the drawings, attain the objects of the invention by employing bifilar primary windings in the output transformers to reduce to a negligible value the leakage inductance between these windings. The effect of substantially eliminating leakage inductance between primary windings is radically to reduce transients during current cross-over from one to another of the tubes of a push-pull amplifier, these transients being particularly severe in class B operation. Leakage inductance between the primary windings and the secondary likewise contributes to these transient effects, but in reduced degree. Relating the primary windings in the manner stated inherently enables reduction of leakage inductance between primary windings and the secondary winding.

By proper arrangement and connection of the primary windings in the transformer the equivalent shunt capacity across the primary windings, due to the capacity between windings, which together with leakage reactance and the capacity of the secondary winding determine falling off of response at the higher audio frequencies and the high frequency cut-off point of the amplifier, may be similarly reduced, and the windings may be so related to the electronic tubes of the amplifier that but a single anode power supply is required, and that in certain of the embodiments conventional input circuits may be employed.

The conventional mode of reducing leakage inductance consists of sectionalizing primary and secondary windings and interleaving or interspersing these. This type of construction is expensive, and while it succeeds in reducing leakage inductance, results in increased capacities.

The ,total capacity of the transformer windings may be reduced by avoiding the necessity for interleaving or pi-winding, in accordance with the present invention, in order to reduce leakage inductance. By avoiding the necessity for piwinding, or interleaving, furthermore, the transformer of the present invention may be arranged more compactly than previously known transformers of the same performance, resulting in reduction of iron requirements, and in a, simplified, more economically fabricated core and winding structure.

In the conventional push-pull output transformer for class B amplifiers the primary Windings of the transformer are connected in series between the plates of the electronic tubes of the amplifier. Accordingly, the primary windings being closely coupled, the total impedance of the primary windings is approximately four times the impedance of a single primary winding. In accordance with certain embodiments of the present invention the primary windings of the output transformer are not connected in series with each other between the amplifier tubes, but are connected effectively in parallel. Thereby a reduction in anode terminal to anode terminal impedance of (4), approximately, may be attained. Additionally, each coil, by reason of its bifilar relation to another coil, is for the same length of wire and length of coil of double the number of layers, resulting in a further decrease of shunt capacity. Reduction of anode to anode impedance of the windings, is, therefore, reflected in a corresponding decrease in anode to anode distributed capacity across the windings, and therefore in a radical extension upwards of the cut oif frequency of the amplifier, at its high end. Alternately, more turns may be employed in the primary windings, and the resultant increase of shunt capacity, due to increase in the number of turns, can be tolerated.

Audio amplifiers constructed in accordance with the present invention, and tested for distortion, have shown less than /a% distortion over the band 20 to 20,000 cycles, the conversion efficiency-of the amplifier tubes remaining above 50% over the band, with an essentially flat response over the band of 20 to 200,000 cycles. Nevertheless, transformers constructed in accordance with the present invention inherently cost less to build than do transformers of the highest quality fabricated in accordance with prior art principles, and require less space and weigh less than the latter.

Further, no transformers currently available commercially or known to me are capable of attaining the wide frequency response and low wave form distortion attainable by the present system, regardless of their cost, weight or space.

The above and still further objects, advantages and features of the invention will become apparent upon consideration of the following detailed descriptions of various embodiments of the invention, especially when taken in conjunction with the accompanying drawings, wherein;

Figure 1 is a schematic circuit diagram of an embodiment of the invention wherein is employed a pair of bifilarly wound primary coils in an output transformer, one of the coils being connected in the cathode circuit of a vacuum tube of a pushull amplifier, and the remaining coil connected in the anode circuit of the amplifier;

Figure 2 is a schematic circuit diagram of a further embodiment of the invention wherein is utilized a transformer having two bifilarly wound primary coils, each comprising two windings, each bifilarly wound coil having one of its windings connected in the cathode circuit and the other in the anode circuit of one of the vacuum tubes of the amplifier;

Figure 3 is a schematic circuit diagram illustratin a modification of the system illustrated in Figure 2 of the drawings, wherein the push pull amplifier utilizes vacuum tubes having screen I grids, each screen grid being maintained at a constant diilerence of potential with respect to its associated cathode during operation of the amplifier;

Figure 4 is a schematic circuit diagram of a modification oi the system of Figure 2, wherein controllable degeneration is provided in the amplifier;

Figure is a schematic circuit diagram of a variation of the system of Figure 1 of the drawings, arranged for balanced operation;

' Figure 6 is a schematic circuit diagram of a variation of the embodiment illustrated in Figure 2 of the drawing, wherein the output trans- -former is an auto-transformer;

Figure 7 is a view, showing a transformer having bifllarly wound primary windings for use in, .push-pull amplifiers arranged in accordance with the invention; and,

Figures 8, 9 and represent variations of the unity coupled transformer of Figure 7.

Referring now more particularly to the drawings and having reference particularly to Figure 1 thereof, there is illustrated a push-pull amplifier constructed in accordance with the principles of the present invention and utilizing an output transformer arranged in accordance with the invention.

The amplifier of Figure 1 is illustrated as employing a pair of triodes i, 2 as amplifying electronic devices, the triodes i and 2 being provided respectively with grid leaks 3 and s, which are connected between the

control electrodes

5 and 6 of the triodes. i and 2, respectively, the midpoint of the grid leak resistors 3 and 4 being grounded via a. bias source 1. Driving potential is applied to the

control electrodes

5 and 6 from sources conventionally illustrated as generators 8, 9, which may be presumed to provide potentials of opposite phases with respect to ground, and of suitable relative magnitudes, the potentials provided by the sources 8, 9 being applied to the

control electrodes

5, 6 via coupling condensers l0 and H respectively, the resistors I2 and i 3 representing the internal resistances of sources 8 and 9, respectively. The bias established by the bias source I may be such as to cause operation of the triodes i and 2 to be either as class A, class AB or class B amplifiers, the significance of the classification being well un derstood in the art, and defined by the Institute of Radio Engineers in its ofilcial definitions. While the circuits and structures oi! the present application have wide utility in amplifiers operating in accordance with any one oi the above mentioned classifications, the invention has primary application to class B amplifiers, and will be described accordingly as utilized in amplifiers of this class, without intending thereby to limit the scope of the invention. For the purpose stated, the bias source 1 will be established to have a value such as to cut on" the plate current of the triodes i and 2, in the absence of signal voltage applied to the grids thereof.

The input circuits of the triodes l and 2 will, accordingly, be seen to be completely conventional and to form essentially no part of the present invention.

A source of anode voltage is is provided, conventionally illustrated as a battery to simplify the drawings. The negative terminal of source It is grounded via the lead l5, and the positive terminal of source is is connected directly via the lead Hi to the anode ll of the triode I, the primary winding l8 of output transforcer T being connected in the cathode lead or the triode l, intermediate the cathode ll thereoi and the negative terminal or the potential source II.' The cathode 22 of the triode 2 is connected directly to ground and a further primary winding 2| 0! the output transformer T is connected between the positive terminal of the potential source I4 and the anode 22 of the triode 2.

The primary windings I! and 2| are wound in biillar manner, or equivalently, as indicated in the schematic circuit diagram, the wires forming one of the windings being immediately adjacent the wires forming the other of the windings so that substantially zero leakage inductance exists as between the windings l8 and 2|.

If it be assumed that a sine wave of potential is applied to the

control electrodes

5, 6 by the sources 8, 9, the positive halt of the sine wave deriving from source 8 eil'ecting current transfer through the triode I, and the positive hall of the sine wave deriving from source 9 eflecting current transfer through the triode 2, it will be apparent that while the positive half of the first mentioned sine wave is applied to the control electrode 5 that the triode 2 is cut oil and that current flow through the primary winding It takes place in the direction of the arrow 23. On the other hand, while the positive half of the second mentioned sine wave is applied to the control electrode 6 of the triode 2, the triode l is cut oil and current flow through the primary winding 2! takes place in the direction of the dotted arrow 24 Accordingly, with respect to the fiux produced in the core 25 of the transformer T, current flow in the windings i8 and 2! is in opposite directions, so that an alternating magnetic flux is set up in th core 25, and an alternating voltage induced in the secondary winding 26 of the transformer T, for application to the load circuit conventionally illustrated as a resistance 21.

By virtue of the close coupling existing between the primary windings l8 and 2|, the close coupling being brought about by the manner of winding the primary windings i8 and 2 l, substantially no leakage reactance will exist between these primary windings, and, accordingly, as explained in the article by Sah, cited hereinbefore, no transient eifects will exist during change over of current carrying function from the triode I to the triode 2, and vice-versa. At the same time, the direction of the voltages E existing across both the windings i8 and 2! are always in identical direction, despite the fact that current ilow in the two windings is in opposite sense because of the fact that the windings conduct in alternation and are closely coupled.

If we assume that the triode 2 is cut oif, and the triode l conducting, for example, the winding it induces in the winding 2! a voltage congruent with its own voltage, and in the same sense in the two windings, the voltage in winding 2 l, however, being incapable of causing current flow in triode 2 because the input voltage applied to grid 6 is now negative in phase and of suificient amplitude with respect to the voltage applied to the anode 22 of triode 2 by winding 2i, to prevent such current flow. Precisely the same argument may be presented when triode 2 is conducting and triode I cut oil.

Furthermore, the

terminals

28 and 29 of the primary windings l8 and 2i are directly connected together via the potential source ll, which may be assumed to have zero impedance, and the total number of turns contained in the windings it cathode lead of the triode 2.

and 2| and are precisely equal. Accordingly, no A. C. potential difference exists between any two adjacent points of the windings I 8 and 2|, so that but slight or zero capacitive currents flow between adjacent turns of the primary windings l3 and 2|. Such currents as do flow tend to maintain the potentials of adjacent points of the two primary windings l8 and 2! identical, and accordingly contribute to the proper functioning of the system.

A condenser C may, if desired, be connected directly from cathode is to anode 22 without al- 'tering the operation of the system essentially, but to assure the equipotential relation between adjacent turns, particularly at the higher frequencies, where some leakage reactance might conceivably be present due to imperfections of the winding spacings.

It will be noted, upon close analysis, that, the triode i being cathode loaded and the triode 2 anode loaded, the former is subject to degeneration and the latter is not so subject. The gains of the triodes l and 2 are not equal, for that reason, and the input signals must be compensated accordingly. This feature of the system of Figure 1 detracts from its utility, in some degree.

Reference is now made to Figure 2 of the drawings, wherein is disclosed a variation of the specific embodiment of my invention illustrated in Figure 1 of the drawings, employing triode vacuum tubes 8 and 2, and having signal input circuits duplicating those disclosed in Figure 1, and described in connection with the description of the circuit connections and operation of the embodiment of my invention there illustrated, except that the signal sources to. and 3a provide signals of identical magnitude.

Whereas in the embodiment of my invention illustrated in Figure l of the drawings a single primary winding is connected in the cathode lead of the triode i, and a single primary winding connected in the anode lead of the triode 2, in the system of Figure 2 a more completely balanced arrangement is provided, wherein the primary circuit of the transformer T comprises four windings, one each in the cathode circuits of the triodes l and 2, and one each in the anode circuits of the triodes l and 2. The negative terminal of the anode supply I4 is again grounded, the positive terminal of one winding 30 being connected via the lead 3i to the anode 22 of the triode 2, and a second winding 33, in series with winding 36, being connected via the lead 34 to the anode ii of the triode 1. Accordingly, the windings 3d and 33 are connected in push-pull relation to the triodes I and 2, and pass currents in alternation if the triodes I and 2 are biased for class-,B operation.

A further winding, 35, is connected in the cathode lead of the triode l, and a winding 35, in series with winding 35, similarly connected in the Accordingly, current flow in the winding 35 takes place in phase with current flow in the winding 33, these current flows being additive in respect to flux production in the core of transformer T. Likewise current flow in the cathode winding 36 is in phase with current flow in the winding 30, and flux production responsive to the current flow in the

windings

30 and 36 is cophasal in the core of the transformer T. Furthermore, th magnitudes of the currents flowing in the

windings

33 and 35 are identical and the magnitudes of the currents flowing in the

windings

30 and 36 are identical,

and all the

windings

30, 33, 33 and 33 are provided with the same number of turns. The terminals 31 and 38 of the windings 33 and 38 are connected together over the extremely low impedance provided by the potential source I4 and. accordingly, may be assumed to be at the same A. C. potential. Th phase of the voltages across the

windings

33 and 36 are identical, for the reasons provided in the explanation of the system of Figure 1, so that voltage correspondence extends along the lengths of the wires forming the

windings

35, 36 to the remote terminals thereof. Due to the fact that voltage correspondence exists between every two adjacent points of the windings 33 and 38, only slight or zero interwinding current flow takes place by reason of capacities existing between the windings. Such current flow as does take place due to capacitive coupling is, moreover, beneficial rather than detrimental because it tends further to eliminate voltage dider= ences between adjacent points of the windings 33, 3d.

The argument presented in the previous paragraph may obviously be duplicated in respect to windings 3t and 35.

Further, the windings 3d and 35 are wound in bifilar or equivalent fashion, as are the

windings

33 and 33, so that substantially no leakage inductance exists among the winding

pairs

33, 33 and 35, 36.

Since substantially no leakage inductance exists between primary windings, the efifect of transients due to leakage inductance, which have been described in the article by Sah, are completely eliminated in amplifiers constructed in accordance with the arrangement of Figure 2 of the drawings. Likewise because interwinding capacity currents are radically reduced, as well as because leakage inductance has been substantially eliminated, and for other reasons above provided, the high frequency resonant point of the transformers is raised by a matter of octaves over the resonant frequency of transformers capable of being constructed at equivalent cost in accordance with prior art principles. The radical reduction in shunt capacities and leakage inductance, furthermore, eliminates the normally expected reduction of response at the higher frequencies, so that the amplifier, taken as a whole, provides an extremely fiat response over a very wide band of frequencies. I

Actual examples of amplifiers comprising the invention illustrated in Figure 3 of the drawings have been constructed and found to produce a fiat response curve over the band 20-200,000 cycles,

. having less than one-half percent distortion, over the range of frequencies 20 cycles to 20,000 cycles, inclusive, the conversion efiiciency of the amplifier tubes remaining above 50% over this band, and the total amount of copper and iron utilized being equal to or less than is employed in high grade transformers of comparable price presently commercially available, the performance of the latter being far inferior.

The embodiment of my invention illustrated in Figure 3 of the darwings is substantially similar to that illustrated in Figure 2 of the drawings, except that the triodes 8-2 are replaced by pentodes dG-Ql, the pentodes being connected in a novel manner to assure high power conversion emciency.

It will be realized that both the amount and the character of the potential difference between the cathode and the screen grid of a pentode, or a tetrode, are parameters which determine in large part the power conversion eficiency oi the tube. Since the cathode potentials of cathode loaded pentode or tetrode tubes vary with respect to ground, it follows that if the screen potentials are fixed with respect to ground the cathode to screen potentials will vary, and, in general, the power conversion efilciency of the tubes will be found to be reduced.

In order to preserve the power conversion eihciency of pentodes and tetrodes when they are connected in cathode loaded circuits, it is usual to connect a condenser between the cathode and the screen grid, and to connect the screen grid tqp itive anode potential through a resistance. The object of circuit arrangements of this character is the maintenance of approximately constant potential of suitable value between the cathode and the screen grid. The total potential difference between the cathode and the screen grid approximates the potential difference between the positive and negative terminals of the power supply for the tube, less the potential drop due to the current flow in the screen dropping resistance and the drop due to D. C. resistance of the cathode load impedance. It accordingly follows that the size of the screen grid dropping resistance varies in inverse proportion to the potential difference between the screen grid and the cathode, as the resistance is varied. However, the minimum size of the screen grid dropping resistance is limited by the amount of loading imposed on the output circuit due to this resistance, so that an ideal solution cannot be realized, and maximum power conversion from pentode and tetrode tubes in cathode loaded circuits likewise cannot be obtained.

The circuit illustrated in Figure 3 of the drawing provides a solution to the problem of attaining maximum power conversion from pentode and tetrode tubes in cathode loaded push-pull amplifier circuits, arranged in accordance with the present invention, the solution consisting in connecting the screen grid 42 of one pentode 40 directly to the anode 43 of the other pentode, and the screen grid 44 of the other pentode 4i directly to the anode 45 of the first pentode 40. The connection of the screen grid 42 to the anode 43 implies connection of the screen grid 42 to the terminal 45 of the output transformer winding 30, which, as has been explained, in connection with the embodiment of my invention illustrated in Figure 2 of the drawings, is maintained at the same A. C. potential as is the point 31 of the winding 35. Since the terminal 31 is always at the cathode potential of the pentode 40, likewise the terminal 45 of the winding 30 is maintained at the same A. C. potential as is the cathode of the pentode 40, the D. C. potential exist 'ing between the two points being, however, that provided by the potential source l4. Accordingly, as the cathode of the pentode l varies in potential, due to the presence in the cathode circuit of the current carrying winding 36, the potential of the screen grid 52 varies in precisely similar manner. The difference in potential is thus maintained constant, thereby maintaining maximum power conversion from the pentode.

A precisely similar explanation may be provided in connection with the pentode 4i, this explanation, however, being sufficiently obvious.

It will be further realized that while I have discosed tubes 40 and 4| as pentodes, that precisely the same principles and mode of operation and circuit connections may be employed in conjunction with the use of tetrodes, including beam power tubes, in the circuit of Figure 3.

Reference is now made to Figure 4 of the drawlngs, which illustrates basically a system of the same character as that illustrated in Figure 2 of the drawings, there being added to the latter, however, controllable degenerative feed-back, still further to reduce the distortion of the amplifier, or in the alternative to necessitate reduced driving signal, as compared with the embodiments of Figures 2 and 3. In the system of Figure 4 of the drawings, controllable degenerative feed-back is derived by connecting across the

primary windings

35 and 35, which are connected in the cathode leads of the tricdes I and 2, a resistor 5D, the latter then having developed across itself a voltage which is a replica of the output voltage available at the output of the transformer T. A pair of variable taps 5i and 52 are provided, taps 5i and 52 being located generally at points equidistantly located with respect to cathodes 2B and I9, respectively. Accordingly, by varying the positions of the taps 5i and 52 the total feed-back voltage to each of triodes I and 2 may be varied. The voltage deriving from the tap 5! is applied to control electrode 8, via a coupling condenser 53, which is connected to one terminal of the secondary winding 55 of an input transformer S, which is supplied with exciting voltage via a primary winding 57 excited from the source of signal voltage A in conventional fashion. The remaining terminal of sec ondary winding 56 is connected to the control electrode 6 of the triode 2. While the control electrode 6 is being raised in potential and the tube 2 is conducting, the potential of the tap 5| decreases in potential due to current flow in resistor 50 in the direction of the arrow I, the cathode 20 of the triode 2 being then at higher potential than is the cathode III of the triode I. In a similar manner, the tap 52 introduces a degenerative voltage into the grid of the triode I via a coupling condenser 531: which is connected in series with one terminal of the secondary winding 58 of the transformer S, the other terminal of winding 58 being connected to the control electrode 5 of the triode I. Grid leaks for triodes I and 2 are provided by resistors 59 and 54, respectively connected in serieswith bias source 55.

It will be clear, then, that, by moving tap Hi to cathode I9, and tap 52 to cathode 2Il, zero degeneration will be introduced into the system, and that degenerative voltages having values as great as twice those normally available in the system of Figure 2 may be made available by establishing tap 5| at cathode I9, and tap 52 at cathode 20.

Reference is now made to Figure 5 of the drawings wherein is illustrated a variation of the system of Figure 1 of the drawings. Specifically, in the system of Figure 1, a primary winding I8 of a transformer T is connectedin series with a cathode circuit of a first triode I and a further primary winding 2|, which is wound in bifilar relation to the primary winding I8, is connected in series with the anode circuit of a further triode 2, so that eifectively the triode I is, cathode load ed, while the triode 2 is plate loaded.

In the system of Figure 5 of the drawings the tube I is cathode loaded by means of the primary winding I8 of transformer T. Input signal from the secondary 60 of an input transformer S is applied between the control electrode 5 of the tube I and the terminal 51 of winding I8 via leads 65 and 65. However, the tube 2 is driven in a difierent manner in Figure 5 than is the triode 2 in Figure 1, the tube 2 being effectively cathode loaded in the system of Figure 5. To this end the primary winding 2| is connected in the anode circuit of the tube 2, in a similar manner to the connecions previously described in conjunction with Figure l of the drawings. A further signal is applied in opposite phase to the first mentioned input signal via a winding 52 connected between the anode 22 of the tube 2 and the controle electrode 6 of the tube 2 via the usual block ing condenser C.

Accordingly, the input circuit of tube l sees two alternative voltages, one originating in the primar winding 63 and which is inductively transferred to the secondary winding 60, the second constituting a degenerative feed-back voltage deriving from the primary winding E8 of the output transformer T by virtue of the connection of the terminal M of the secondary winding 50 via the

lead

65, 66 to the negative terminal 6'5 of the primary winding i 8.

The input circuit of tube 2 sees two alternating voltages, one originating in the primary winding 53, which is inductively transferred to the secondary winding 62, the second constituting a degenerative feed-back voltage, deriving from the primary winding 2| of output transformer T, and degenerative by virtue of, the fact that winding M is effectively in the cathode circuit of tube 2, varying the potential of cathode 20 with respect to that of anode 22, in one phase, while the potential of control electrode 6 is being varied in the same phase with respect to anode 22 by the voltage in secondary 62. Looked at in another way, the input voltage for tube 2 consists of the voltages of windings 2i and 62 in series. So, when the positive half cycle of voltage in winding 62 is in the direction of the arrow E, the potential of the control electrode 6 with respect to cathode 20, assuming the latter fixed, increases positively. This results in an increase of plate current, which increases the voltage across coil 2!, resulting in a voltage rise across winding 2! in the direction of arrow E. It will be obvious that, as seen from the grid-cathode circuit of tube 2, voltages E and E are oppositely directed and hence that voltage E is degenerative.

Additionally, if a screen grid 61 is provided in tube I, this may be connected directly to anode 22 of tube 2, and will be maintained at a constant potential with respect to cathode l9 of tube l equal to the voltage of source H, because the same A.-C. voltages exist at all times on anode 22 and on cathode 59. If desired a capacitor, C, may be connected from cathode H to anode 22.

Similarly, a screen grid 68, provided in tube 2, may be connected directly to the positive terminal of source I 4, and will'be maintained at an A.-C. voltage difference from the potential of cathode 20 equal to the voltage of source ll, since no impedance exists between screen grid 68 and cathode 20.

It will be realized that the tubes land 2 in Figure 5 of the drawings may then be triodes,

tetrodes, pentodes, beam power tubes, or the like, as desired, and further that in the various embodiments and examples of my invention, illustrated and described herein, the specific character of the electronic amplifier tubes employed may be selected at will from among the-various types available, i. e., triodes, tetrodes, pentrodes, beam power tubes and the like.

The system of Figure 6 represents a simple variation of the system of Figure 2, demonstrating that, if desired, the bifilar primary windterminals of the voltage supply.

3% ings 85, 38 may be coupled directly to a load, the transformer "r acting as an auto-transformer.

I have disclosed a. variety-of push-pull amphfier systems, employing each a bifllarly wound transformer, and have illustrated in the schematic circuit diagrams, Figures 1-6, conventionally, bifilarly wound transformers. I have, however, recited that transformer primary windings equivalent to bifilar windings may be employed.

Reference is accordingly made to Figures 7-10 inclusive, of the drawings, wherein is illustrated a plurality of difierent transformer winding constructions which may be employed in the circuits of Figures 1-6, inclusive, Figure 7 illustrating a transformer having four primary windings, bi-

filarly wound, and associated with a common secondary winding, Figure 8 illustrating a possible substitute for the system of Figure 7, employing superposed coils in place of bifllarly wound coils,

the transformer of Figure 8 being in some respects equivalent to the transformer of Figure 7, when the windings are properly connected'in a push-pull amplifier arranged in accordance with the invention. Figures 9 and 10 illustrate variants of the transformer of Figure 7 wherein separate layers of a single coil are incorporated by suitable interlayer connections in difi'erent primary windings of a push-pull transformer, providing a true approximate equivalent for a bifilarly wound transformer.

Referring now more specifically to Figure 7 of the drawings, there is illustrated a core E00 of conventional structure having wound thereon a coil i ll! formed of a plurality of bifilarly wound layers 502, the winding commencing at point 963 and terminating at point we. Leads 6%, M6 are brought out from the commencement point N3 of the bifilar winding, to which may be connected 3+ and B terminals of a voltage supply, when the transformer is connected in an amplifier circuit. Similarly from the end of the winding, at point ltd, are brought out two terminals 807, E08, intended for connection, respectively, to the plate or anode P1 of one amplifier tube of a push-pull amplifier, and the cathode C2 of the remaining tube.

A duplicate coil lie is wound on the same core beside the coil till, having terminals ill and i 52 for connection respectively to 3+ and B and terminals M3, M4 for connection respectively to the cathode Ci of the one tube and the anode P2 of the remaining tube;

It will be noted that the respective windings I are wound in opposite winding senses with respect to the core M0, for reasons explained hereinabove, and briefly because the coils ill! and H0 are intended to produce flux in push-pull, or alternately in opposite directions in the core lB-Il.

Secondary windings M5, M6 are superposed on the primary coils wt and iii), respectively, and are shown connected in series by a lead ill, it being understood that parallel connection is equally feasible.

In the broadly or approximately equivalent system of Figure 8, bifilar windings are dispensed with, and four primary windings are provided, numbered i20, i2l, 522 and H23. The superposed windings Ho and i2! are wound in mutually identical sense, and the superposed windings I22 and H3 in identical sense, the latter two oppositely to the first mentioned two windings, and the winding pairs are arranged adjacently on the core. The initial point G24 of winding I20 may be connected to terminal IB- and the terminating point I25 of winding I2I' to terminal 3+, of a plate voltage supply source, by appropriate terminals provided, and the terminal points I24 and I25 being thus Joined by a path of negligible A.-C. impedance remain at-identical A.-C. potential. The terminal point I26 of winding I20 and the initial point I2! of winding I2I are arranged to be in close juxtaposition, and are joined by a condenser Ki, which serves to maintain the points I26 and iZl at identical A.-C. potentials.

The terminal point I26 may be connected to cathode C2 and the terminal Ifil to anode PI.

The coil I22 may be similarly arranged, terminal I 28 being connected to 13-, terminal I29 of coil I23 to 3+, and terminals H and I3I joined by a condenser K2, so that the terminals of pair I28, I29 and the terminals of pair I30, I3i are at identical A.-C. potentials. Terminal I30 may be connected to cathode CI and terminal I3i'to anode P2, of the tubes of the amplifieremploying the transformer. The secondary winding i232 may be arranged as in the embodiment of Figure 7 of the drawings.

It will be realized that leakage inductance, in the case of the embodiment of my invention illustrated in Figure 8, will be greater than in the case of the embodiment of Figure 7. However, the transformer of Figure 8 may conceivably be more economically constructed than the transformer of Figure 7, and may prove desirable for that reason, despite its relatively poorer performance.

In Figure 9 of the drawings is illustrated a further modification of the system of Figure 7, wherein the effect of a bifilar coil is attained by winding the respective primary windings which are desired to be unity coupled, in successive layers, and joining the layers thereafter by means of suitable leads. Having particular reference to a transformer suitable for use in the amplifier system of Figure 2, for example, the winding 29 may comprise the winding layers I4I, I42, I43, I44, I45, I46, etc., and the winding 30 the alternate layers I41, I46, I49, I50, II

The terminal point of layer I4I may be connected 1 to B- and its other end point joined by lead I to an adjacent end point of layer I42, the layers I and I42 being wound in the same direction and current in each turn of both layers Ill and I42 flowing in the same sense, to produce mutually additive flux in the core. The process of layer interconnection is continued to the end of the winding, the winding layers I45, I42, I43 being thus connected in series. The alternate layers, I41, I48, I49 are likewise connected in mutual series relation by leads IN, and a secondary winding I63 may be superposed on the primary windings, in conventional fashion. The terminal points of the outermost pair of adjacent winding layers may then be brought out to anode PI and cathode C2, respectively.

A similar pair of primary windings, I54 and I65 may be provided on the core, adjacent to the

primary windings

29 and 30, for connection to the anode P2 and the cathode CI, and with the

windings

29 and 30 associated a further secondary winding I64, connected in series with secondary winding I63, it being understood that parallel connection is equally feasible.

It will be realized, since the windings are adjacent in alternate layers, and since the starting points of initial layers HI and I41 are adjacent and interconnected by a path. of low impedance provided by the voltage source 3+ and B-, that the potentials of adjacent turns of each pair of layers is ideally at identical A.-C. potential. To compensate for any departures from ideal conditions, brought about by winding irregularities and the like, I may interconnect the ends of the windings by means of a large condenser K, which establishes the ends of the windings at identical A.-C. potential.

Figiu'e 10 illulstrates a winding sequence which approaches that of the sequence provided in the embodiment of Figure 9 of the drawings, the winding being laid in successive layers, 170 which are left mutually unconnected when the coil is wound. The first. fourth, fifth, eighth, ninth, twelfth .layers are connected in series by leads i6I to provide one winding; the second, third, sixth, seventh, tenth, eleventh layers are connected in series by leads I62 to provide the other winding. The initial points of the first and second layer may be connected respectively to the 23+ and B- terminals of a voltage supply, and the two outermost windings (the eleventh and twelfth layers of a twelve layer winding, for example), connected to the cathode, C2, of one vacuum tube and the anode, Pi, of a further vacuum tube of a push-pull amplifier arranged in accordance with the invention. The latter two terminals may be connected across a condenser K to assure that the same A.-C. potential exists at these terminals, as in the transformer arrangements of Figures 8 and 9, inclusive. The upper windings may be duplicated to provide two pairs of biiilarly wound equivalents.

It will further be realized, while the transformers illustrated in Figures 7, 9 and 10 approach relatively closely to the ideal, or bifilarly wound transformer, that the embodiment of Figure 8 is at best a very rough approximation, and, while operative, operates but imperfectly in circuits arranged in accordance with the invention, and is not recommended except in cases where other considerations than excellence of performance are primary.

It will further be realized that further variants of the transformers illustrated in Figures 7, 9 and 10 may be resorted to without departing from the true scope and spirit of the invention,

which requires the provision of unity coupled transformers, for best performance, and which mat, employ any type of unity coupled transformers having the requisite windings, and which are known or which may become known to the art.

Consideration of the system of Figures 2, 3, 4 and 6 of the drawings Will render evident that each of the tubes of the amplifiers or modulators disclosed is operated with negative feedback, since' in each case a primary winding of an output transformer is connected in a cathode lead of a tube and the grid-cathode or input circuit is connected across the winding. Each of the tubes is, however, also plate loaded so that part of the output of each tube derives from its plate circuit, and part from the cathode circuit. Thereby, I provide a push-pull amplifier which possesses the advantages of both a plate loaded and acathode loaded system, simultaneously, in addition to the other advantages previously disclosed, a feature which has not previously been attainable in push-pull amplifiers useful for wide band amplification of audio or video signals, or the like, to my knowledge.

While I have described various modifications of amplifiers, or modulators, arranged to employ to advantage output transformers having bifilarly id wound primary windings, further modifications may be devised, and re-arrangements and modifications of the modifications illustrated and described, resorted to, without departing from the true spirit and scope of the inventions, as defined in the appended claims.

In particular it will be realized that the various embodiments of my invention herein disclosed have particular application, though not exclusive application, to class AB and class B power amplifiers, and their variants, and hence to utilization as modulators in various systems of modulation, and particularly to class B plate modulators.

What I claim and desire to secure by Letters Patent oi the United States is:

l. A push-pull wide band audio frequency amplifier, comprising, a first electronic amplifier tube having a first anode, cathode and control electrade, a second electronic amplifier tube having a second anode, cathode and control electrode, a source of anode voltage having a positive and a negative terminal, a magnetic core, first and second primary output transformer windings of substantially equal inductance and having each a high impedance at said audio frequencies, arranged in bifllar relation about said core, means vfor connecting said first winding between said negative terminal and said first cathode, means for connecting said second Winding between said positive terminal and said second anode, third and fourth primary output windings of substantially equal inductance and having each a high impedance at said audio frequencies arranged in bifilar relation about said core, means connecting said third winding between said negative terminal and said second cathode, means for connecting said fourth winding between said positive terminal and said first anode, a secondary winding coupled substantially equally to said first, second, third and fourth windings, and a push-pull input circuit for said first and second electronic amplifier tubes, responsive to a wide band audio source for driving said control electrodes with oppositely phased wide band audio voltages.

2. The combination in accordance with claim 1 which includes means for biasing said amplifier tubes for class B operation.

3. The combination in accordance with claim 1 wherein said first amplifier tube comprises a first screen grid, and wherein saidsecond amplifier tube comprises a second screen grid, means for connecting said first screen grid directly to said second anode and means for connecting said second screen grid directly to said first anode.

4. A push-pull wide band audio frequency amplifier, comprising, a first electronic amplifier tube having a first anode, cathode and control electrode, a second electronic amplifier tube having a second anode, cathode and control electrode, a source of anode potential having a negative and a positive terminal, an output transformer having a magnetic core, multiturn primary windings of substantially equal inductance and having high impedance at audio frequencies linking said core and coupled to said first and second tubes, said primary windings comprising at least two' closely coupled windings wound' in identical winding sense, means for connecting one of said windings between said negative terminal and said first cathode, means for connecting the other of said windings between said positive terminal and said second anode, said primary windings comprising at least two further closely coupled multiturn windings of substantially equal inductance and having high impedance at audio frequencies linking said core and wound in identical winding sense, opposite to said first mentioned winding sense, means for connecting one of said further windings between said positive terminal and said first anode, means for connecting the other of said further windings between said negative terminal and said second cathode, and a push-pull wide band input circuit for applying a wide band of audio frequencies to said control electrodes in push-pull relation, and means for biasing said electronic amplifier tubes for anode current fiow in at least one of said tubes at all times in response to said signals.

5. I'he combination in accordance with claim 4 wherein said first and second electronic amplifier tubes have a first and a second screen grid, respectively, and wherein is provided means for maintaining constant potential between said first screen grid and said first cathode and between said second screen grid and said second cathode, during operation of said amplifier.

6. The combination in accordance with claim 4 wherein said means for maintaining constant potential between said first screen grid and said first cathode comprises a direct current connection between said first screen grid and said second anode, and wherein said means for maintaining constant potential between said second screen grid and said second cathode comprises a direct current connection between said second screen grid and said first anode.

7. A wide band amplifier, comprising, a first amplifier tube having a first cathode circuit and a first anode circuit, a second amplifier tube having a second cathode circuit and a second anode circuit, an output transformer having a magnetic core, a first pair of unity coupled primary windlugs and a second pair of unity coupled primary windings, both linking said core, means for connecting one of said first pair of windings in said first anode circuit and the other of said first pair of windings in said second cathode circuit, means for connecting one of said second pair of primary windings in said second anode circuit and the other of said second pair of primary windings in said first cathode circuit, a load circuit coupled substantially equally to all said primary windings, means for biasing said amplifier tubes to provide current fiow in at least one of said tubes in response to any finite signal, and a wide band input circuit connected in push-pull relation to said control electrodes for applying said wide band of signals thereto,

8. An amplifier for amplifying a wide band of signals with essentially fiat response, comprising, a first electronic amplifier tube having a first anode, cathode and control electrode, a second electronic amplifier tube having a second anode, cathode and control electrode, a source of anode voltage having a positive and a negative termi- 112.1, a magnetic core, first and second primary output windings-arranged in unity coupled relation about said core, each of said windings having an impedance at the low end of said band, which is of the same order of magnitude as the internal resistance of one of said tubes, means for connecting said first winding between said negative terminal and said first cathode, means for connecting said second winding between said positive terminal and said second anode, third and fourth primary output windings arranged in unity coupled relation about said core, said third and fourth primary output windings each substantially duplicating an impedance one of said first and second primary windings, means connecting said third winding between said negative terminal and said second cathode, means for connecting said fourth winding between said positive terminal and said first anode, an untuned load circuit coupled to said first, second, third and fourth windings equally, an input circuit coupled in push-pull relation to said first and second control electrodes for applying to said control electrodes in push-pull relation said wide band of signals, and means for biasing said amplifier tubes for operation with anode current flowing in at least one of said amplifier tubes at all times in response to said signals.

9. The combination in accordance with claim 8 wherein each of said first and second amplifier tubes comprises a further control electrode.

18 means for connecting said further control electrode of said first amplifier tube to said second anode over a path oi negligible impedance, means for connecting said further control electrode of said second amplifier tube to said first anode over a path of negligible impedance.

FRANK H. MCINTOSH.

REFERENCES crrEn The following references are of record in the file of this patent:

UNITED STATES PATENTS 15 Number Name Date 1,791,236 Drake Feb. 3, 1931 2,187,782 Hardwicl: et a1. Jan. 23, 1940