US2015082A - Electric wave translating system - Google Patents

Description

Bepi. 24, 1935. s. T. MEYERS i M5932 ELECTRIC WAVE TRANSLATNG SYSTEM Original Filed Deo. 3l, 1930 4 Sheets-Sheet l um um /NVE/v TOR S. 7'. MEVERS JMW A TTORNEY Sept 4, 1935 s. T. MEYERs 2,015,02

' ELECTRIC WAVE TRANSLATING SYSTEM Original Filed Deo. 5l, 1930 -4 Sheets-Sheet 2 -S-jr ummm CUPLANA l? GR/DS um um l l A TTORNEV @pit M, n35., s. T. MEYERS ELECTRIC WAVE TRANSLATNG SYSTEM original Filednec. 31, 1930 4 Sheets-Sheet 3 HHHHH' A TTORNE V Sept. 24, 1935. s T. MEYERS 2,@i58

ELECTRIC WAVE TRANSLATING SYSTEM Original Filed DeC. 3l, 1930 4 Sheets-Sheet 4 COPLANAR GR/DS /NVENTOR 5.7'. ME VERS A T TORNEY Patented Sept. 24, 1935 UNITED STATES vMeur orties Bell Telephone Laboratories,

Incorporated,

New York, N. Y., a corporation of New York Application December 31, l1930, Serial No. 505,802 Renewed December 14, 1933 6 Claims.

This invention relates to Wave translating systems, especially such systems involving electric space discharge devices, as for example electric space discharge amplifiers.

5 I An object of the invention is to control distortion production in said systems, as for instance to reduce asymmetric distortion and/or other distortion produced in vacuum tube amplifiers.

In one specic aspect the invention is a cirlO cuit in which distortion components generated in a vacuum tube amplifying device of high load capacity are so fed back in the circuit as to reduce the amount of distortion produced by the device, and in which, without necessitating increase of l5 the plate voltage or decrease of the power output of the device, a decrease is effected in the variation which the device causes the gain of the circuit to undergo with variation of voltage input to the circuit. In one specific embodiment of the 20 -Tinvention shown in the drawings, the vacuum tube amplifying device is a coplanar grid Vacuum tube with a negative biasing potential on the control grid and a positive biasing potential on the other grid, which is in substantially the same space rei ffl-ation as the control grid with respect to the cathode and the anode or plate. Even for moderate plate voltage, with the positive gri-d the power output of the tube can be high without necessitating driving the control grid positive, and there- 80 fiore without necessitating wide Variation of the input impedance of the tube. Important advantages of this stability of the input impedance are that it makes for gain stability of the circuit and for stability of the phase shifts in the circuit. l-This stability of gain and phase shift are especially important in the case of amplifiers of a type to which the invention is applied in the above mentioned embodiment of the invention, in which distortion components generated in the tube are visolated from the fundamental waves by balance and are so fed back in the circuit as to reduce the amount of distortion produced by the vacuum tube device as referred to above; for, as explained hereinafter, the balance desired in order to isolate the distortion components from the fundamental waves is deleteriously affected by variation of the gain and the phase shifts in the circuit.

In another specific embodiment of the invention which is shown in the drawings, a special gain U corrector circuit is employed which is responsive to gain changes caused by varying input impedance of the vacuum tube device to produce complementary gain changes in an anterior portion of the circuit as explained hereinafter, in order to stabilize the gain. This embodiment may Cil employ a three electrode tube or the coplanar grid tube or any suitable type of tube as the vacuum tube device.

Another specic embodiment of the invention whichis shown in the drawings is a vacuum tube circuit of the type in which distortion components produced by transmission of fundamental waves through an amplifier comprising a vacuum tube device are isolated from the fundamental waves by balance and the isolated distortion components 1o are so fed back in the circuit as to reduce the amount of distortion produced by the device; and in this embodiment the means for isolating the distortion component by balance comprises an auxiliary vacuum tube amplifier for transmitting l5 the fundamental waves substantially without distortion and balancing the Waves so transmitted against waves transmitted through the distoring vacuum tube device to neutralize the fundamental component of the opposed distorted waves from said device and thereby yield the distortion components without fundamental waves. As explained hereinafter, the use of the auxiliary vacuum tube amplifier tends to prevent objectionable regeneration in the circuit and facilitates obtaining correct phase relations for causing the balance to obtain over a considerable frequency range and for considerable variation of transmission level.

Other objects and features of the invention will be apparent from the following description and claims.

Fig. 1 is a circuit diagram of an amplifier circuit embodying a form of the invention including the gain corrector circuit referred to above;

Fig. 2 shows an amplier circuit embodying a form of the invention which employs a coplanar grid tube as referred to above;

Fig. 3 shows an amplier system embodying a form of the invention which employs an auxiliary balancing vacuum tube amplifier as referred to above; and

Fig. e shows a modified form of the system of Fig. 3.

Fig. 1 shows an amplifier circuit of a general 45 type disclosed inBritish Patent 317,005 and in U. S. patent to H. S. Black, No. 2,003,282, June 4l, 1935. In such circuits one portion, which in amplifying Waves gives them an odd number of phase reversals, has the distortion components that are generated in that portion neutralized, and that portion will therefore be referred to as the compensated portion. To effect this neutralization, the distortion components generated in the compensated portion of the amplifier are -f Cri .tubes are shown.

fed in proper amplitude from the output side of that portion to its input side so that after arnplification by that portion they will reappear at its output side in such phase as to neutralize themselves and thus besuppressed.

The odd number of phase reversals in the compensated portion of the amplier can be eected, for example, by having an odd number of amplifying stages in that portion, since the driving voltage generated in a vacuum tube is in phase opposition to the grid voltage which produces it. An amplifier circuit of this general type is shown in Fig. 5 of the above mentioned British Patent 317,005, and of the above mentioned U. S. Patent 2,003,282. However, in that figure the compensated portion of the amplifier includes but one amplifier stage. Fig. 1 of the present application shows a similar amplifier which however, has two additional stages (or a total of three stages) in the compensated portion. The two additional stages are shown in Block B. The reference characters common to the Fig. 5 and the Fig. 1 which are mentioned just above designate the same elements in the two figures. The block B consisting of the two additional stages has been inserted between the condenser l5 and the grid leak resistor i8' of tubes ic, so that the compensated portion of the amplifier contains not only the tubes lo but also tubes lm and in of the two amplifying stages included in block B.

In Fig'. 1, an input transformer impresses Waves from circuit 5 upon the grid of tube la. These waves may be, for example, voice waves to be amplied, or voice modulated carrier waves Vto be amplified and transmitted over carrier wave wire transmission systems or to radio transmitting antenna: or waves received over such systems. The waves impressed on the grid of tube la are amplified by tubes la, lb, lm, In, and lo in cascade connection, the tubes 5c being connected in parallel to each other. The amplified waves from tubes lc are impressed on the outgoing circuit 3 through a Wheatstone bridge circuit which has the space path resistance Ro of tubes lc as one ratio arm of the bridge and has the primary winding of output transformer 2 in the output diagonal of the bridge. Each of the tubes lc has its space path resistance 4R0. The four ratio arms of the bridge are designated by their resistance values 4R0, KRO, KR and R, K being a constant. The driving voltage generated in the space paths of tubes lc may be considered as a generator, in series with resistance R0. The usual plate, filament and grid batteries for the The usual grid battery for tubes Ic is 8c; that for the tube ld is 8d; and that for tube In is 8u. The battery 8c supplies negative biasing potential to the grids of tubes lc through grid leak resistor I8 and choke coil I8 in parallel. The steady plate current for tubes l c passes from the plate battery 5c for those tubes through a choke coil 20, the space discharge paths in the tubes and the resistance KR@ Condensers ll, l2, i3, I4, i6, 36 and 46, and also condensers 9 and 9', are blocking or stopping condensers, of negligibly low reactance for the waves to be amplified.

Across the resistance KR and KR@ in series, is a circuit forming one diagonal of the bridge and comprising a resistance 66 in series with three parallel paths. One of these paths is the blocking condenser S', a resistance :1: (shown adjustable) and the space path resistance R01 of tube Ib. Another, which is of negligibly low `admittance at the frequencies of the waves to be amplified, extends through the stopping condenser 2, input or coupling or grid leak resistor RT for a vacuum tube Id and grid biasing battery 8d for that tube, in series. The third parallel path is an adjustable resistance This diagonal will be referred to as the input diagonal of the bridge or the feed back diagonal; for the driving voltage generated in the space path of tube ib appears in this diagonal and is opposed therein by the driving voltage generated in the space paths (of tubes lo) located in the ratio arm R0 of the bridge, and moreover the voltage across RT which results from this opposition of the two driving voltages is amplified by tube Id which feeds the amplied voltage back to the grid of tube lb through a path comprising condensers 3l and 32, resistance 33 and condenser i4. However, the value of resistances :t and 65 are so adjusted that this opposition of the driving voltages gives a Zero value of resultant voltage across RT as regards the signal wave, or in other words makes point P a null point as regards the signal wave. Consequently, only the distortion components produced in the compensated portion of the amplifier and appearing in the driving voltage generated in tubes I c are fed back to the grid of tube Ib. In other words, the tendency for tubes Ic to feed fundamental or signal waves back to the grid of tube Ib through resistance $5 and tube Id is neutralized by the tendency of tube Ib to feed back to its own grid through a: and tube Id, so that only distortion components (for example modulation products) arising in the compensated portion of the ampliiier and appearing in the driving voltage generated in tubes fc are fed back to the grid of tube Il). These distortion components undergo an odd number of phase reversals in the live tubes Id, lb, im, In and Ic, so that they reappear in reversed phase in the driving voltage in the space 40 path of each of the tubes Ic, and the magnitude of resistance 25 so adjusts the amplitude of the reappearing waves that they balance out or neutralize their originating distortion waves in the driving voltage in the space paths of tubes lc.

Thus distortion in the compensated portion of the amplifier is eliminated.

However, the degree to which such neutralization or elimination is realized in the amplifier depends on the accuracy with which the circuit elements are adjusted in such a way that any istortion component fed back from the platelament space path of the last stage lc through tubes Id, lb, lm, ht and Ic, returns to the plate filament space paths of tubes Ic with its original amplitude but with its phase shifted 180 degrees. That is, in order that the distortion voltage shall balance out, the product of the voltage amplifications and diminutions around the circuit should be unity, and the phase shift should be 180 degrees. Since the effectiveness of the circuit thus depends on the production of two voltage waves of equal amplitude and opposite sign, a good balance requires that the phase difference be very close to 180 degrees and that the gain of the circuit change very little after the balance is attained. Consequently, it is desirable to provide accurate compensation for undesirable phase shifts in the amplifier caused, for example, by interstage coupling circuits, and moreover to provide such phase compensation without unduly changing the shape of the gain-frequency characteristic of the amplifier. For example, to compensate for such phase shifts under normal conditions, condensers 3| and 32 can be used (the 75 capacity of condenser 32 preferably being variable and large relative to that of condenser 3 I to facilitate close or fine adjustment of the phase), or any suitable type of phase compensating means desired can be used, as for instance the type disclosed in Nyquist Patent 1,894,322, January 1'7, 1933.

AThe resistance KRO is adjustable, to provide for correcting unbalance of the Wheatstone bridge arising from variations in plate impedance of tube lc caused for example by variations in the power supply voltages for the tube or by substitution of one tube for another. To make the distortion components balance out in the space path of tube lc as explained above, the magnitude of the regenerated distortion components appearing across the feed-back diagonal. of the bridge can be controlled by adjusting the variable resistances 25 in that diagonal, and the condenser 32 can be adjsuted to give the desired phase correction.

If desired, a harmonic analyzer (not shown) can be connected across circuit 3 in order to tell when the best adjustment for Z5 and 32 has been effected.

Since the load or work circuit for the amplifier is comprised in a diagonal of the balanced Wheatstone bridge which is conjugate to the feed-back diagonal, the impedance of the amplifier as seen from the load or work circuit is the same as it would be Without feed back and is independent of the impedance ofthe feed-back diagonal, and the feed-back voltage is Aindependent of the impedance of the work circuit. The system reduces both odd and even order distortion components at the same time, and can stabilize gain and afford high load capacity. v

The system increases the load carrying capacity of electric space discharge tubes (l) not only by attaining an increase in load capacity of very lgreat importance by suppression of distortion components of frequencies other than the fundamental frequencies and thereby permitting the tubes to operate over a larger range of their gridvoltage plate-current characteristics but also (2) by attaining a second increase of very great importance in the load capacity by regeneration of fundamental waves in such a way as to control gain in a desired manner, as for example, to prevent undesired lowering of gain, for the funda- .mental waves.

In connection with this latter increase, it should be noted that the system provides means for correcting for distortion caused by improper degree of amplification of fundamental waves, as for example caused by amplification of a fundamental wave of a given frequency different amounts for different input amplitudes, or as for example caused by amplification of two waves of respectively different fundamental frequencies, different amountsy respectively. For example, if a wave of a given fundamental frequency is amplified in the compensated portion of the amplifier (i. e., in any or all of tubes im, |11l and lc or the circuits associated with the tube or tubes) to a degree less than, say, the normal amplication for that portion of the amplifier, then for that frequency the voltage which tubes lc feed to the resistance RT or cause to appear across the resistance RT tends to be lower than normal, i. e., less than the funda-V mental which is applied across RT from tube Ib` through R01 and m. As a result, the tendency toward lower than normal gain of the system for the fundamental wave of the given frequency is checked. Similarly, if the given frequency is am- ;,plied to a degree greater, instead of less, than normal in the compensated portion of the amplifier, then for that frequency the voltage across RT from tubes l c tends to be higher than the voltage of that frequency which is applied across the resistance RT by tube Ib through R01 and r; and as a result the tendency toward higher than normal gain of the system for the fundamental wave of the given frequency is checked. The system compensates for too low or too high gain for fundamental waves, at the same. time that it suppresses components of frequencies other than fundamental frequencies.

Since the isolatedv distortion components, (0btained across the resistance RT) are amplified separately from the signal in an amplifier shown as comprising a single tube ld (through which the signal components do not pass), and are then fed back to the grid of tube Ib, the amplication of the isolated distortion components can be oontrolled independently of the amplification of the signal components, and can, for example, be made greater than the amplication of the signal components.

A summary of the main features' and of the operation of the circuit of Fig. l as so far described will now be given. In the output circuit of the last stage is a balanced Wheatstones bridge, the plate resistance of the tubes of this stage forming one of the arms. The output transformer connects to two points of the balanced bridge conjugate to the feed-back diagonal. The bridge voltage across the points connecting to the feed-back diagonal, and also that across resistance 25, is exactly proportional to and in phase with the s'o-called driving voltage in the plate circuit of the last stage. In order to produce a null point at the junction of resistances 68 and 25.l in this diagonal, a second voltage source is introduced in this diagonal, so that a second voltage is applied to resistance 6B, but 40 ijs from an earlier portion of the circuit and in series with a resistance zr. The proper phase relationship may be readily obtained for example, by going back an odd number of stages. If the circuit to which connects is far enough back, 45 f (i. e., if a: is far enough back), then the voltage applied to the null point in series with x will be free from distortion, and across the null point we will have present only what is not wanted in the load or work circuit. The amplifier is a feed-back 50 a amplifier possessing the property that a voltage applied to its input or to the left of :c appears in its output but does not appear at the null point. Accordingly, in feeding back distortion waves, i. e. waves that are not wanted ultimately, such waves are introduced to the left of in order that they will not reappear across the null point and cause the process to be repeated. In the circuit shown, the distortion voltages across the null point are applied to the left of a: through the aid of an auxiliary or pickup tube Id. This means of application is shown merely by way of example, the fundamental requirement being that the feed-back path from the output bridge to :c include an even number of tubes or some unilateral or other suitable circuit arrangement not introducing phase shift. Under these conditions the feed-back may be so adjusted that it produces in the output coil two currents that are exactly out-of-phase Iand equal in amplitude to the distortion currents produced by the compensated portion of the amplier, namely, that portion having the resistance :L' at its input side and having the bridge at its output 75 Cil side. The amplifier is thus a balanced type of feed-back amplifier.

In obtaining curves or measurements of harmonic production, the magnitude of the feedback may be adjusted separately and individually for harmonics of different orders, for example second harmonics and third harmonics. The same balance does not necessarily hold for both second harmonics and third harmonics. This is attributed to the presence of harmonics in the stages preceding zr. While these harmonies do not appear across the null point they do appear in the output transformer of the lamplifier.. The second harmonics in these stages make it necessary to feed back less from the null point than would otherwise be required for a balance, whereas the production of third harmonics in these stages sometimes requires more feed-back than would otherwise be required in order to get complete suppression of third harmonics with tubes of .a given type, for example, Western Electric Company type 10e-D or O tubes, in the output stage. This relationship is not universally true for this type of tube but is random, it being possible to have third harmonics produced by this type of tube reversed 180 from the position just described In such a case, for complete suppression of third harmonics there Would be required, not over compensation in the compensated portion of the amplifier, but, as for the case of complete suppression of second harmonies, under compensation in that portion, inasmuch as the second and third harmonics from the preceding portion would tend to counteract the second and third harmonics, respectively, generated in the compensated portion. In general, harmonic production in the uncompensated portions of the circuit should be maintained small in order to facilitate great suppression. This harmonic production can be maintained small, and the suppression consequently can be very much improved by employing a plurality of stages in the compensated portion of the amplier to reduce the load (grid swing) on the stages preceding Distortion in the preceding stages ahead of then does not become troublesome until after much greater suppressions have been attained. I'he spread between the best balance for second and third harmonics is greatly reduced.

The increase in the number of stages in the compensated portion of the amplifier considerably improves the overall gain regulation of the amplifier, also. To explain this simply, it is convenient to interpret the change in gain as the result of superimposing an eXtra voltage analogous to a modulation, and hence, which can be suppressed like modulation, but at the same time is the same frequency as the fundamental itself. In other words, as indicated above, an excess or shortage of fundamental is self-compensated for in a Way exactly similar to the way in which the various modulation products themselves are suppressed. Assuming that the change in gain of the amplifier before regenertion occurs only in the portion of the amplifier to the right of Rx, gain stabilization with regeneration will be equivalent to the amount of harmonic suppression.

The system of Fig. 1 as so far described differs from the Fig. 4 of the Black application, Serial No. 298,155, and likewise from the Black application, Serial No, 455,239, by including resistance 66. When this resistance is omitted (i. e. short circuited), then, as pointed out ineach of those With the resistance 66 in circuit, the corresponding condition is 1 K (R+Ro)l-Rse =gain of compensated portion of amplifier wherein Rss is the value of the resistance 66. Increasing the gain or number of stages in the compensated portion of the amplifier tends to reduce .r and therefore increases (deleterious) plate circuit modulation in the stage preceding the compensated portion; but this tendency is counteracted by increasing R or resistance 66. While it is desirable to have R large within limits, the use of resistance 66 (of sufficient magnitude) avoids any necessity for making R inconveniently or undesirably large in order to obtain a suitably large value of x. By varying the resistance 25 the magnitude of the distortion components fed back through tube Id can be varied to obtain the desired balance or distortion suppression without changing the null point, i. e. without disturbing the desired condition that the fundamental voltage on the grid of tube Id is (normally) zero.

With the circuit of Fig. 1 as so far described, as the load (current or power output, or grid swing) is increased to the value at which the grids in the output stage are driven positive, there is a sudden drop in the suppression of the distortion or crosstalk currents. This may be attributed to two causes. One is the change in phase-shaft with the stabilizing action of the circuit. Analysis of the circuit shows that there are three round trip paths or loops around which regeneration may take place. Each one of these paths has phaseshift. One path or loop, in particular, which is in the corrected or compensated portion of the amplifier changes when the grids of the output stage are positive. The opposing regeneration in this path tends to decrease. When this path changes the other two change simultaneously. Therefore, all three paths changing cause an aggregate change in phase shift which is sufficient to destroy the good balance obtained before the grids went positive. The second reason for the sudden decrease in suppression is the change in amplitude. This inherently tends to accompany a change in phase, but aside from this, even if the reactive phase shift in the circuit be zero, the amplitude will change for still another reason. Regeneration in the corrected or compensated portion of the amplifier destroys the balance or null point if it is varied or changed. Hence, even if the phase be entirely corrected, because of amplitude balance it will also still be necessary to eliminate this change in regeneration entirely or else make it sufficiently small.

A separate gain corrector, shown in block C in Fig. 1, may be used to compensate for or eliminate the gain change caused by the change in grid impedance of the last stage when the grids of that stage go positive. It comprises a vacuum tube amplifier G fed from tube |11, through stopping condenser 10 and resistor 1l, and fed from tube Im through stopping condenser 36 and variable resistance 12, and feeding back to the grid of tube Im through resistance 13 and stopping condenser I6. The grid battery 8u for tube In is shown as serving also as the grid battery for tube G. This battery 811. supplies grid biasing potential for tube G through a resistance 14, and supplies grid biasing potential for tube I n through resistances 14 and 12 in series. Plate current for tube G is supplied through choke coil 15.

Regeneration around'the path through tubes I m and G would tend to increase the gain of tube Im, (since the number ofl tubes or phase reversals around the path is even); whereas regeneration around the path through tubes im, In and G would tend to decrease the gain of tubes Im and In (since the number of tubes or phase reversals around the path is odd) However, resistance 12 is so adjusted, for example, that the voltages applied to the grid of tube G from tubes In` and Im, which are 180 dfferentin phase, Aare equal in magnitude and therefore cause the grid of tube G, i. e. the point p, to be a null point or point of zero A. C. potential when the grid impedance of tubes le is normal. Under these conditions the tube G feeds nothing back to tube im. However, when the grids of tubes lc are driven positive by the signal, and the grid impedance of tubes Ic therefore decreases, departing from its normal value, then a decrease in the output voltage of tube In tends to result but the tendency is checked because the voltage decrease causes the voltage applied' to the grid of tube G from tube Im to overpower the voltage applied to the grid of tube G from tube In and thereby, provided the value of resistance 14 bears the proper relation to the value of resistance 12, result in such feedback action or regeneration around the path through'tubes Im and G as to increase the gain of tube Im sufficiently to substantially counteract or neutralize the gain decrease of tube In and thereby stabilize the overall gain of tubes Im, In and Ic.

The ratio of '14 to 72 is governed by how much the amplication from the plate of tube Im `through G and back to the plate of Im, (which varies with different circuit designs) must be reduced to make it equal to unity. I n other words, 14 is similar to 25 in function.V The value of 12 is determined in the same manner in which a: is determined. The values of 12, 1I and 14 should be such as not to place too low impedance across the plate circuit of tube Im or tube In. summarizing the operation of the gain corrector C, the adjustment of resistance 12 is such that when the grid impedance of'tubes Ic is normal the grid of tube G, i. e. the point P, is a null point. Feed-back action which occurs in the path through tubes Im and G, thev path through tubes Im, In and G, and the path through tube In and resistances 1I and 12, is such feed-back operation as will change to stabilize the overall gain of tubes Im, In and lc when the grid impedance of tubes lc changes.

By switch 61, grid battery c can be connected in circuit in place of the battery 8c, so that tubes Ic will be given, instead of the negative4 grid bias, a positive grid bias suicient to maintain the grid potential of these tubes always positive and cause them to operate on a straighter `portion of their static characteristics of Vinput voltage versus output current. Such operation with positive instead of negative grid bias results in less variation of gain caused by changing grid impedance; because the average impedance is nearly constant due to some secondary emission in the grid circuit which tends to prevent the grid current versus grid voltage curve from rising to its normal value. The impedance at any point is given by the reciprocal of the slope of this curve. For such operation with positive instead of negative grid bias, variation of plate impedance with load, i. e. with grid swing, is also 5 less, because the operating point is on a steeper portion of the static characteristic. Further, in such operation with positive instead of negative grid bias, for a given power output the space current is less. Moreover, at least in some cases, harmonic production in the tubes of the last stage is less than for negative grid bias. This is an added advantage in that the requirements on the suppressor are reduced.

When the grid potential of the last stage is maintained always positive the tube or tubes used in that stage preferably have high lament emissicn so that there is at all times an abundant supply of electrons. Such a tube has a long straight characteristic in the region of positive grid potentials before the saturation point is reached. An example of such a tube is the Western Electric Company type 102-D or V tube. A special tube construction which has been used for operation with the grid always positive 25.

is a Western Electric Company type 205-D. or E tube with the grid of the V tube. This makes a tube with a high amplification factor and with which a large space current at a low plate voltage can be obtained which varies substantially linearly with input voltage over a large range of positive grid potentials.

When the grid voltage of a vacuum tube sweeps over a portion of the tube characteristic including both the negative and positive regions of grid potential the input impedance changes between a very high impedance and a low impedance. Therefore, since the range of negative grid potentials for which plate current will flow is a relative small range, this change in the input impedance will occur if the grid is operated with a negative bias, when a large grid swing is employed. This change in impedance will introduce distortion. This change in impedance' and this distortion are avoided by maintaining the grid always positive. Maintaining the grid always positive avoids also the limitation on grid swing that arises in the case of negative grid bias because of the fact that a vacuum tube characteristic has considerable curvature as the region of zero current is approached, and the consequent fact that when negative grid bias is used distortion will occur unless the input voltage range is coni-med to considerably less than the total range of negative grid potentials for which plate current will flow (with any given plate potential employed). When a tube is operated with the grid always at positive potential, the input impedance is not high as for negative grid potentials, but'is a comparatively low resistance 60 and therefore the circuit of the tube becomes in effect a power operated system as distinguished from a voltage operated system.

Fig. 2 shows a circuit like that of Fig. 1 except that the gain corrector C is omitted and the last stage employs, instead of four singlegrid tubes, an electric space discharge tube CP having coplanar grids, such for example as the coplanar grid tube disclosed by H. A. Pidgeon and J O. McNally in Patent 1,923,686, August 22, 1933 or in the Proceedings of the Institute of Radio Engineers, volume 18, pages 266 to 293, February,f1930. The use of the coplanar grid tube avoids the variation of grid or input impedance of the last stage which caused the gain change for which the gain corrector C of Fig. 1 was designed to compensate, and has other important advantages. The coplanar grid tube has two grids, each active elementary area on either grid being close to a corresponding active elementary area on the other grid and being at substantially the same location as that correspondthe grids is made positive by battery 8.

Ving area with respect to the cathode and the anode or plate. By way of example, each grid may have its lateral wires lie in the same plane or cylindrical surface as the lateral wires of the other grid and alternate with them. One of The other is the control grid, to which the signal is applied, and which is given a negative biasing potential by battery 8c.

The grids function alike as far as their action upon the plate circuit is concerned, and at the same time are independent and function independently of one another. Therefore, one can be maintained positive and the other negative, so that the same operating point can be reached as would be used for a three element tube with its grid positive, while at the same time there is available a negative grid which can be used for driving, i. e. as a control grid. The use of the coplanar grid tube not only avoids the variation of grid or input impedance of the last stage of the amplifier circuit, but moreover affords a high value of that impedance, in the interest of low power losses in the circuit. It also avoids occurrence of any objectionable local oscillations or other deleterious effects that tend to result from secondary electron emission or negative resistance consequent to positive potentials on control grids.

If desired, as for example when the grid swing for the last stage may be so large as to drive the control grid positive even in the case of the coplanar grid tube, the gain corrector C of Fig. l can also be employed in Fig. 2. In other words, a coplanar grid tube or coplanar grid tubes can be employed in the last stage of Fig. 1 as one is employed in the last stage of Fig. 2.

The system in Fig. 3 is an amplifier which is a modification of that shown in Fig. 2, especially in that Fig. 3 employs an amplifier auxiliary to the signal amplifier for producing the null point P. Fig. 3 omits the tube lb of Fig. 2 and includes in the compensated portion of the system not only the tubes lm, in and CP of Fig. 2, but also tubes 2m, 3m and 4m in cascade connection with each other between the tubes im and In. The auxiliary amplifier comprises tubes Bim, 82m, 83m, 84m and Bln, all in cascade connection.

The null point P, or point normally at zero fundamental potential, is produced in general as in Fig. 2, by feeding to the grid leak resistance RT of the tube Id, (i. e., to the grid of the tube ld or the point P) two signal or fundamental waves of equal amplitude but opposite phase, one from the bridge in the output circuit of tube-CP and the other from the plate of tube la (or the grid of tube Im, i. e. the input to the compensated portion of the amplifier or system). The path for feeding the fundamental wave to point P from the output bridge is through variable resistance 56', corresponding to resistance 66 of Fig. 2. Also, the path for feeding the fundamental wave of equal amplitude but opposite phase to point P from the plate of tube la includes a resistance 9: corresponding to the resistance :c of Fig. 2; but it further includes the auxiliary amplifier, interposed between the resistance z and the point P so as to amplify the potential delivered by resistance zc before applying that potential to point P.

As in the system of Fig. 2, only modulation components or distortion components produced in the compensated portion of the amplier or system appear at point P and these components are fed back or returned to the plate of tube CP through a path comprising tube Id and the compensated portion of the amplifier or system. In Fig.. 3 this path includes a variable phase cor- 10 recting condenser 32' and a variableI attenuating or amplitude adjusting resistance 33 connecting the plate of tube Id to the grid of tube im, as the plate of tube id in Fig. 2 is connected to the grid of tube I m through the phase correcting condensers 3l and 32 and the resistance 33. The distortion components return to the plate of tube CP with the same amplitude that they originally had at that point, but with their phase reversed, so that they balance themselves out or neutralize themselves at that point and consequently do not appear in the load or work circuit.

In Fig. 3, as in Fig. 2, there should be an odd number of phase reversals, (obtained, for example, by an odd number of amplifying stages) in the regenerative path from the plate of the last stage (of the compensated portion of the amplifier) through the feed-back path to the grid of the first stage of the compensated portion of the amplifier and through the compensated portion back to the plate of the last stage.

It has been pointed out above that when a null point P is produced by the system of Fig. 2 there are three regenerative paths, one being the loop in the compensated portion of the amplifier, and that regeneration in this loop resulting from production of the null point in this manner is detrimental. In the circuit of Fig. 3 the (unilateral) auxiliary amplifier used in producing the null point P eliminates this regeneration in this loop and facilitates maintaining the null point over a wide band of frequencies. Considerable feed-back has been used in such an amplifier without occurrence of singing. The load on the auxiliary amplifier was very light (note the large number of stages or considerable gain in the compensated portion of the system before the last stage) so that the auxiliary amplifier did not itself introduce appreciable harmonics. The cir=50 cuit using the auxiliary amplifier for production of the null point showed an improved stability, and less sudden increase of harmonic production as a function of power output. The uncorrected distortion voltages were made very small, and the adjustments made for maximum suppression of second harmonics served for third harmonics also. The phase shift of the auxiliary or null point amplifier and the compensated portion of the system approached one another and the phase correcting condenser 32 was common to both. When gain stabilization did take place the change in phase shift was not as great as for the case of null point production without use of the auxiliary amplifier. These factors were of assistance in eliminating sudden decrease of distortion suppression with load increase.

The coplanar tube CP has its control grid maintained at negative potential by grid battery 8c, through choke coil I8". This tube has its 70 other grid maintained at positive potential by grid battery 8. Tubes 2m, 3m, 4m, In, 82m, 83m, 84m and Bln, have grid leak resistors 85, connected to grid batteries 8 or 811.. The grid leak resistance for tube lm includes the resistance V ar' in series with the grid circuit resistor 85 of.

tube 8lm. A plate battery 6 supplies space current for tube In (which may be, for example, a

AWestern Electric Company type lOl-D or L tube), through a frequency selective circuit comprising an inductance 86 and a capacity 88 in addition to a plate inductance 90 for the tube. Plate batteries 6 supply space current for all of the tubes except I n and CP through resistancecapacity filters which are individual to the tubes and each of which comprises a resistor 81 and a condenser 88 in addition to a plate resistor 89 for its tube. These tubes may well be Western Electric type Z39-A tubes, for example. With this type of tube the plate current is small and therefore the resistance 89 through which it is supplied can be ordinary lavite resistances di rectly conected to the plates without lead wires, in order that the interstage grid-to-filament capacity may be small. If desired, any suitable phase correcting or compensating means may be employed, as for example the above mentioned phase compensating means of the Nyquist Patent 1,894,322. In any case it is of assistance to have the total capacity shunted across the input of each stage small, and to that end it is helpful to have the capacity to ground of each of the individual elements of the interstage coupling 'circuits a minimum. Therefore, it is advantageous to have the physical dimensions of these elements small, as for example, to have the tubes and sockets and the coupling condensers of small dimensions and to have the grid and plate wiring eliminated as far as feasible.

If desired, as for example to ease the requirements on the filters for the plate current supply to the tubes, the capacity of one or more of the interstage coupling condensers 3B in Fig. 3 or in any of the gures of the drawings can be made sufficiently small to assist in maintaining the amplifier gains at low frequencies below the singing point, instead of being given the large capacity usually given condensers which are to function merely as blocking or interstage coupling condensers. The small capacity can counteract singing tendency at frequencies below the utilized frequency band, and yet need not be so small as to cause objectionable phase vshift within or above that band. When such a condenser of small capacity is used it is preferably one in the compensated portion of the amplifier, as compensation is then afforded for its phase shift.

'I'he system of Fig. 4 is an amplifier which is a modification of the amplifier of Fig. 3, and which is more simple. Fig. 4 omits the tubes la, 4m, 82m, 83m, 84m, and Id of Fig. 3, and has tubes I'm, Z'm, 3m, I'n, CP, 8l'm and 8F11, which correspond respectively to tubes Im, 2m, 3m, In, CP, Slm and Bln of Fig. 3 but are all shown as heater type tubes (for example Western Electric Company type 24'7-A tubes) except the coplanar tube CP. The filament current for indirectly heating the cathode elements is supplied from a voltage source shown as a filament heating battery 90. Battery B supplies space current for tube CP through a choke coil 2IJ' and the primary winding of transformer 2 in series. The grid of the first stage of the auxiliary amplifier is fed from the grid of the first stage of the compensated portion of the system directly, instead of through a resistance such as x of Fig. 3. That' is, the resistance a," is omitted. A grid leak resistor 85 forms the entire grid leak resistance for tube lm, whereas the grid leak resistance for tube im of Fig. 3 comprises the resistance :c

in series with a'resistor 85. The null pointP is fed from the output bridge through a resistance 66 corresponding to the resistance 66 of Fig. 3; but the connection through which the null point P feeds distortion components to the grid of the first stage of the compensated portion of the amplifier omits the phase correcting condenser 32 and attenuating resistance 33 of Fig. 3. However, phase correction can be obtained by a variable condenser 9". In Fig. 4 the 10 plate resistor for the first stage of the auxiliary amplifier, instead of being a simple resistance such as 89 in Fig. 3, is a potentiometer 89 for varying the gain of the auxiliary amplifier. The potentiometer is so adjusted as to give amplitude 15 balance between part of the signal or fundamental output oi" the compensated portion of the system fed to point P through resistance 66 and part of the input amplified by the auxiliary amplifier, and consequently the signal or funda- 20 mental potential at point P should be zero when the phase shift through the main amplifier or compensated portion of the system is an odd multiple of 180 and the phase shift through the auxiliary amplifier is an even multiple of 180. 25 The distortion voltage then appearing at point P is fed back through the main amplifier or compensated portion of the system so that it balances itself out or neutralizes itself at the plate of the last stage and leaves the output to the 3o load or work circuit free from distortion. The operation of the system is similar to the operation of the system of Fig. 8, and Will be apparent from the explanation above of the operation in Fig. 3, without further description.

For the sake of simplicity the invention has been explained above with reference especially to pure resistance impedances Where impedances have been described as external or ratio arms of the Wheatstone bridge. However, the invention 40 is not limited to the case in which the bridge, if used, has its ratio arms resistances. They, as Well as the load, may be impedances of any character, (proper provision being made, of course, for the necessary supply of steady potential to the plates and grids of the tubes). Where Ru has been treated as constituting one arm of the bridge, the plate-filament capacity is so small that its reactance at frequencies of the order of those of the waves to be amplified is so great compared to the impedance Ro as to be negligible.

What is claimed is:

l. A circuit comprising a vacuum tube device, means for transmitting to said device fundamental wa-ves which produce distortion components in said device, a vacuum tube amplifier for transmitting the fundamental waves substantially without distortion and opposing the waves so transmitted against waves transmitted through said device, said amplifier having such transmission efficiency and phase shift that the fundamental waves transmitted therethrough neutralize the fundamental component of the opposed distorted Waves from said device and thereby yield the distortion components without fundamental waves, and means for causing these isolated distortion components to reduce the amplitude of distortion components delivered by said circuit.

2. A circuit comprising a vacuum tube device, means for transmitting to said device fundamental waves which produce distortion components in said device, a unilateral wave transmission path for transmitting the fundamental Waves substantially Without distortion and opposing the waves so transmitted against waves transmitted through said device, said path having such transmission eiciency and phase shift that the fundamental waves transmitted therethrough neutralize the fundamental component of the opposed distorted waves from said device and thereby yield the distortion components without fundamental waves, and means for causing these isolated distortion components to be so regenerated in said circuit as to reduce the amplitude of distortion components delivered by said circuit.

3. A circuit comprising a vacuum tube device, means for transmitting to said device fundamental waves which produce distortion components in said device, a unilateral wave transmission path for transmitting the fundamental Waves substantially without distortion and opposing the waves so transmitted against waves transmitted through said device, said path having such transmission efficiency and phase shift that the fundamental waves transmitted therethrough neutralize the fundamental component of the opposed distorted Waves from said device and thereby yield the distortion components without fundamental waves, and means for causing these isolated distortion components to be so regenerated in said circuit as to control the amplitude of distortion components delivered by said circuit.

4. A Wave translating system comprising a source of fundamental Waves, a wave path associated therewith and containing means amplifying said waves and increasing their energy and distorting them, a unilateral Wave transmission path associated with said source for transmitv ting the fundamental waves substantially without distortion' and opposing the waves so transmitted against waves transmitted through said means, said unilateral path having such transmission efliciency and phase shift that the fundamental Waves transmitted therethrough neutralize the tortion and opposing the waves so transmittedl against waves transmitted through said means, said unilateral path having such transmission efficiency and phase shift that the fundamental waves transmitted therethrough neutralize the fundamental component of the opposed distort--4 ed waves from said means and thereby yield the distortion components Without fundamental waves, and means so feeding back these isolated distortion components to said source that their amplification in said first means reduces the distortion produced by said first means.

6. A wave translating circuit comprising an electric space discharge device having an anode, a cathode and two grids therebetween, said grids being located principally in the same surface withf respect to said anode and said cathode, means for supplying a steady negative bias potential to one of said grids, means for supplying a steady positive bias potential to the other of said grids,

means supplying to said wave translating cir- T cuit signals causing potential swing of said one grid that generates distortion in said device, and means so feeding back said distortion from the output of said device to its input as to reduce said distortion.

STANLEY T. MEYERS.

Images