US2155966A - Linear control circuit - Google Patents

Description

pr 25, 39- J. s. CAMPBELL 2,155,966

LINEAR CONTROL C IRCUIT Filed June 7,' 1955 INVENTOR BY Ila/f0 Sea/f Cam 6&9

Y w/why A TTORNEY Patented Apr. 25, 1939 UNITED STATES PATENT OFFICE 8 Claims.

The present invention applies to an electrical network and to means associated in or with the network for maintaining a desirable ratio between the input voltage and the output voltage. That is, there will be a linear relationship between the input voltage and the output voltage. The network then is a circuit which has a linear characteristic, it is a linear circuit.

There are three types of distortion acting to make a circuit non-linear. These are amplitude, frequency and phase distortion.

There have been many attempts made to design and construct a distortionless or linear circuit and these efforts have been concerned with the design of transformers or other coupling impedances, with the design of the geometrical arrangement of the parts of electronic tubes, and with the design of various compensating circuits, such as the push-pull amplifier.

Applicants invention is of the latter, or compensating, type, but possesses the important property of being inherently distortionless as regards amplitude distortion.

In brief, applicants device comprises a network having an input circuit, or impedance, across which an input voltage is developed, an output circuit, or impedance, across which an output voltage is developed, a main valve between said circuits, and a pilot valve for controlling the main valve, which pilot valve is responsive to variations from a desirable constant instantaneous voltage ratio between said input and output voltages, and which acts upon the main valve to restore said desirable ratio.

Applicant anticipates that the greatest field of use for his invention will be in the communication arts and particularly in the radio art where it can be used to advantage in both broadcasting and receiving.

Figure 1 shows the present invention in a simple form using a triode vacuum tube as the main valve and another triode as the pilot valve in a net-work designed to give a single stage of amplification.

Figure 2 shows the present invention again applied to a single stage of amplification with a tetrode as the main valve and a triode as the pilot valve.

Figure 3 shows a further modification of the present invention as it might be applied to a complicated network having several stages of amplification or other modifications. In this modification the tetrode is again used as the main valve and a triode is used as the pilot valve.

In Fig. 1 the input side of the device is connected to a transformer I, having primary coil 2 and secondary coil 3 associated therewith. One end of the primary 2 is connected to the grid, or control, of a triode vacuum tube 5. This tube 5 acts as the main valve of the network and has as its principal parts the grid 6, the cathode I, and the plate 8. The cathode and plate of the tube 5 are in series with the load impedance III, the power supply I2, and the associated resistance I3.

Also connected to the cathode I is one end of the secondary 3. Connected to the other end of the secondary 3 is the positive lead of a battery I4. The triode vacuum tube I5 is used as the pilot valve and has as its principal parts the grid, or control, I6, the cathode I1, and the plate I8. The negative lead of the battery I4 is connected to the grid I6 of the tube I5. Connected in series with the cathode I! and the plate I8 of the triode I5 are the battery 20 and the resistance 2|.

The negative lead of the battery 20 is also connected to one end of the primary 2 of the input transformer I. The cathode ll of the tube I5 is connected to the negative side of the battery I 2.

In a state of balance, when the network is unaffected by an input voltage, the voltage drop across the resistance 2| provides a necessary negative bias on the grid 6 such that the plate current flowing in tube 5 will cause a voltage drop across the resistance l3 of a value which, when combined algebraically with the voltage of the biasing battery It, will maintain the negative bias on the grid l6 at a value such that the plate current of tube I5 will produce a voltage drop across resistance 2| which will provide the necessary negative bias on grid 6 mentioned above.

When the balance of the network is upset by an instantaneous positive signal on the end connected to grid 6 of the input side of the transformer I, the positive aspect of the grid 6 is increased by the signal, and the plate current will increase the voltage drop across the associated resistance I3, which drop is approximated by the increased voltage of the secondary of the transformer l, which in turn will tend to prevent a change in the bias of the grid l6 of the tube l5, and a change in the flow of current in the plate circuit of this tube. The result of a change in the bias of the grid I6 is described in the paragraph next below.

As previously stated, the object of this invention is to have the signal voltage drop across the load impedance l5, linearly proportional to the input, or signal voltage drop across the transformer coil 2. The voltage drop in resistance 13 is linearly proportional to the voltage drop in impedance it. Also, the voltage across coil 3 is linearly proportional to the voltage across coil 2, within the limits of transformer construction. Hence, the voltage in impedance is linearly proportional to the voltage in 2 if the voltage drop in the associated resistance I3 is proportional to the signal voltage appearing across the associated coil 3. This latter proportionality is maintained by placing the voltages appearing across impedance 3 and resistance l3 in opposition, and maintaining a constant algebraic sum. This sum is a function of the input and output voltages.

Any differential between the voltage drop in the resistance l3 and the voltage drop in the coil 3, resulting from the instantaneous input signal, will change the bias on the grid iii of the tube l5, which will in turn change the bias on the grid 6 or" the tube 5, allowing a current to flow in the plate circuit of tube 5 such that the change of voltage drop in .the resistance l3 will equal the change of voltage drop in the 'coil 3. Thus any tendency away from a constant instantaneous voltage ratio between the input and output voltages will be corrected and the desired ratio maintained. The resistance of these circuits is greater than that required for critical damping and, hence, no oscillations will be producible and the network will stay in equi-- librium.

Figure 2 is another application of applicant's invention as applied to a single stage of amplification with a tetrode as the main valve and a triode as the pilot valve and with the inherent limitations of a transformer, such as used in Fig. 1 eliminated :by the use of resistances across which the input voltage appears.

In Fig. 2 the input side of the device is connected in series across resistance 3| and its associated resistance 32. One end of the resistance 3| is connected to the grid of the tetrode 35 which has as its main parts the cathode 36, the'primary grid 31, the'plate 38, and the secondary grid 39. Tube 35 acts as the main valve of the network. The value of the resistance 3| is such that the voltage drop across it furnishes the proper variation in grid voltage to actuate the grid 31 of tube 35. The plate and cathode of tube 35 are in series with the load impedance 49, the direct current power supply 4| and the associated resistance 42.

The pilot valve '45 is a triode vacuum tube having as its principal parts the cathode 46, the grid 41, and the plate 48. With an applied signal the grid 41 of the pilot valve is biased by the algebraicisum of the voltage drops in the associated resistances42 and 32 and the directcurrent voltage of the voltage supply 49. The cathode 46 and the plate 48 of the pilot valve are in series with the power supply 4 I and the resistance-5m Thebias'on the secondary grid 33 of the main valve 35 is determined by the voltage of the power supply 4| minus the voltage drop in resistance 50, minus the voltage drop in the resistance 42. The signal voltage developed across resistance 3| is applied to the grid 31 of the tube 35.

The associated resistance 32 is of such a value that the voltage drop across it, due to the signal current, will furnish the necessary variation in grid voltage to actuate tube 45.

The associated resistance 42 is of such a value that the normal plate current of the tube 35 will furnish a voltage drop of the order of the voltage appearing across resistance 32, or that necessary to bias grid 41 of tube 45, over and above the voltage of the battery 49.

Resistance 50 is of such a value that the normal plate current of the tube 45, plus the current of grid 39, will furnish a voltage drop which when subtracted from the voltage of the power supply minus the voltage drop in resistance 42 due to the plate current in tube 35, plus the current of grid 39, will provide the necessary voltage on the secondary grid 39 of tube 35. The equations for this stated mathematically and using vector quantities are as follows:

I45 equals plate current of tube 45 I35 equals plate current of tube 35 E39 equals screen voltage of tube 35 E50 equals (I45 plus 139) R50 E42 equals (I39 plus I35) R42 E39 equals E41 plus E42 plus E50 It should be understood that the essential action here is in the control that current 45 has over voltage 39. Other voltages appearing in the equations are incidental and are not essential to the action of the circuit; indeed, they produce a slight upsetting effect which is compensated for by a modification of I45.

In a state of balance, when the network is unaffected by an input signal, the voltage of the power supply 4| minus the voltage drop across the resistance 53, minus the voltage drop across the associated resistance 42, provides a necessary bias on the secondary grid 39 of tube 35 such that the plate current flowing in the tube 35 will cause a voltage drop in the resistance 42. This voltage drop in 42, when combined with the voltage of voltage source 49, will bias the grid 41 of the pilot tube 45 such that its plate current when flowing through the resistance 50 will provide the necessary bias mentioned above for secondary grid 39 of the main valve 35.

From this it will be seen that the plate current of the tube 35 controls the grid voltage of the pilot tube 45 and that the plate current of the tube 45 controls the secondary grid bias of the main valve 35. As the circuits have a resistance greater than that required for critical damping no oscillations will be producible and the network will come in to equilibrium. With these relative values among the several parts of the circuit, the voltage between the cathode 46 and the grid 41 of the tube 45 will be found to quickly revert to a constant value if disturbed. To the extent that the voltage on grid 41 is constant, the ratio between the output and input voltages of the circuit is a constant. 7 7 When this system is in equilibrium as above described the bias on the secondary grid 39 serves as a controlling bias for the plate current flowing in tube 35. If the change'in voltage drop across resistance 32 is not met by a corresponding voltage change in the resistance 42, which is proportional to the drop in40, the bias 'on'the tube 45 will be upset and hence the control on the secondary grid 39 will be upset an amount such that the plate current of tube 35, flowing through the resistance 42 will furnish a change in voltage drop equal to the change in voltage drop in resistance 32. This may be stated mathematically as follows:

Let

E47=the constant grid voltage on grid 41 as mentioned above.

E32=the signal voltage, positive or negative appearing across the input resistance 32.

E42=the voltage developed across the resistance 42, which is equal to this resistance 42 multiplied by the plate current of tube 35.

I35=the plate current of tube 35.

but the output voltage E40 developed across the load impedance 40 is linearly proportional to the plate current, that is:

in which all the quantities are constant except E32 and E40 and which expression is therefore the equation of a straight line, proving that the output and input voltages are linearly proportional.

The operation of this circuit may also be described as follows: What is here said in regard to Fig. 2 applies also to the circuits of Figs. 1 and 3.

With no signal impressed: The d.-c. current in the plate circuit of the main tube produces a voltage drop E42 across resistance 42. The voltage E47 impressed between grid and cathode of the pilot valve is then the difference between battery voltage E49 and this voltage E42. With given d.-c. plate and control grid voltages on the main tube, its plate current is determined by its screen voltage. As is apparent from Fig. 2, the screen grid voltage of the main tube is the battery voltage 4| less the IR drops across resistances 50 and 42. The voltage drop across resistance 50 is determined primarily by the plate current of the pilot valve, which in turn depends upon E47. With given circuit constants, equilibrium is established when the plate current of the main tube, flowing in R42, produces a voltage drop E42 such that the difference between E42 and E49, or E47, acting thru the pilot valve, produces that value of screen voltage which will cause this value of plate current to flow in the main tube. If, for any reason this equilibrium is disturbed, the circuit will immediately react to restore the balance. This action may be shown in the following manner:

Let it first be assumed that an abnormally large plate current is flowing thru the main tube. This, passing thru resistance 42, produces the voltage E42 which tends to make the pilot valve grid more positive. Its plate current increases, producing an increased voltage drop across resistance 50. Now, the voltage applied to the screen grid is always positive and is equal to the battery voltage 4| less the drops across resistances 50 and 42, whence, when the plate current in the pilot valve increases, the screen grid of the main tube becomes less positive, decreasing its plate current.

Conversely, if the plate current of the main tube is less than normal, E42 is too small, the grid of the pilot valve is less positive than it should be, and less'current flows in resistance 50. Thus E50 is reduced and the screen grid 39 becomes more positive, increasing the deficient plate current.

It is thus seen that the circuit is stable for one value of plate current only, and that any disturbance is met by a restoring action.

With signal impressed: The situation can best be seen by considering a given instant. The only difference now between this and the previous case is in the voltages appearing across resistances 3| and 32. That appearing across 3| tends to change the main tube plate current, while that across 32 may be thought of as allowing it to do so, in the sense that the pilot valve would prevent any change in main tube plate current were it not for voltage 32. Let it be assumed that at a given instant, E31 tends to make the grid of the main tube positive, whence E32 tends to make the pilot valves grid negative. The main tube plate current, larger because of E31, places a larger positive voltage upon the grid of the pilot valve; this and E32 cancel and the total pilot valve voltage remains the samei. e., the main tube is allowed to change its plate current in response to the signal. This analysis assumes that the main tube is linear. If it is not, then E42 and E32 cannot exactly cancel, and the pilot valve will change the screen grid voltage of the main tube in the manner described above until they do cancel-that is, until input and output of the main tube are linearly related.

Figure 3 is another application of applicants invention as applied to any number of stages of amplification or other operations. The main valve is a tetrode vacuum tube and the pilot valve is a triode vacuum tube.

The input side of the device is connected across resistance 5| and its associated resistance 52 in series. The end of the resistance 5| away from the resistance 52 is connected to the grid 51 of the tetrode 55 which has as its main parts the cathode 55, the primary grid 51, the plate 58, and the secondary grid 59. Tube 55 acts as the main valve of the network. The value of the resistance 5| is such that the voltage drop across it furnishes the proper variation in grid voltage to actuate the grid 51 of the tube 55. The plate and cathode of tube 55 are in series with the input impedance of the network 60, and the direct current power supply 6|.

The pilot valve 65 is a triode vacuum tube having as its principal parts the cathode 66, the grid 61, and the plate 68. The grid of the pilot valve is biased by the algebraic sum of the voltage drops in the resistances 69 and 52. The cathode 66 and the plate 88 of the pilot valve are in series with the power supply BI and the resistance Ill. The bias on the secondary grid 59 of the main valve 55 is determined by the voltage of the power supply 6| minus the voltage drop in the resistance 70. The signal voltage developed across the resistance 5| is applied to the grid 51 of the tube 55. The resistance 52 is of such a value that the voltage drop across it due to the input signal will furnish the necessary variation in tube 65.

The resistance 69 is so connected across the output and is of a value such that the voltage drop across it, due to the output current, is of the same order as the voltage drop across the resistance 52. The resistance 69 may be a single resistance or a compound resistance such as may be had by using the resistance voltage divider 1|.

The resistance voltage divider 1| provides the voltage drop across the resistance 69 of the same order as the voltage drop across the resistance 52. It will thus be seen that the resistance 69 may be dispensed with and is only shown in the present instance as a point of reference for the voltage drop at that place in the circuit where the resistance is placed. The voltage drop across resistance 59 is associated with the output voltage as the voltage drop across the resistance 52 is associated with the input voltage.

The connection between the resistance H and the output should be so made that the ends of resistances 69 and 52 when connected together as shown are of the same polarity at all times.

In a state of balance, when the network is unaffected by an input signal, the voltage of the power supply 6| minusthe voltage drop across the resistance 10 provides the necessary bias on the secondary grid 59 of tube 55 such that the plate current flowing in the tube 55 will produce a voltage drop across the input impedance of the network 69, which will be transmitted through the electrical network 60 and will produce a voltage drop across the resistance A voltage proportional to the voltage across the resistance 1| will be applied to resistance 69 and to the grid 6? of tube 65, acting to establish a plate current in tube 65 such that the voltage drop produced when said plate current flows through resistance 10 is that voltage drop mentioned at the beginning of this paragraph.

When a signal voltage is applied across resistance'5l and its associated resistance 52, the voltage drop across resistance will act to change the plate current of tube 55, and will thus change the output voltage of the network es. This change in'the output voltage will tend to change the bias on the grid 61 of the tube 55 but this tendency will be opposed by the voltage drop in the resistance 52. Any unbalance in this opposition will change the plate current in the tube 65 and hence the plate current in the tube 55. This change in the plate current of the tube 55 will maintain the desired proportionality between the input and the output voltages. It is to be noted that the change in voltage drop in the resistance 69 is in opposition to the change in voltage drop in the resistance 52. That is, the instantaneous positive, or negative, ends are connected together.

By adjustment of voltage 49 in magnitude, poiarity, or both in Fig. 2, the circuit may be made to act as a detector, in which plate current will flow only during the half cycle of the signal voltage in which the grid 31 of the tube 35 is positive. In such a case the impedance 4!] would consist of the usual audio transformer and bypass condenser.

If this adjustment of voltage 49 is so as to tend to make the grid 41 of tube 45 more positive, the plate current of tube 35 will decrease. This is due to a shift in the point of cut-off of plate current, in the conventional grid-voltage plate current characteristic curve for the circuit as a whole, in a direction which will permit only grid voltage to actuate the the more positive parts of the signal voltage cycle to afiect the plate current of tube 35. When this'point of cut-oif of plate current has been made to coincide with the operating point, or value of grid voltage on tube 35 corresponding to zero input signal, plate current will flow only during the positive half cycles of signal voltage and the circuit will become a linear detector. Similarly, if the voltage 49 is adjusted to tend to make the grid 41 more negative, the plate current of tube 35 will increase. At some negative value of the voltage on 41 the plate current of tube 35 will be under the control of the signal voltage when the latter is at its maximum negative value on the grid 31 of tube 35. In this condition the circuit of Fig. 2 is a linear detector. The magnitude of the adjustments here described would depend upon the electrical characteristics of the tubes used.

The circuit of Fig. 3 may be made to correct for distortion in a radio transmitter by means of the following general scheme:

The signal voltage, corresponding to that appearing across resistances 5| and 52, is obtained directly from the input of the first audio amplifier following the microphone 'or record pick-up. The output voltage, corresponding to that appearing across resistance 59, is obtained from the detector described in the foregoing, which receives its signal inductively from the antenna circuit of the transmitter.

The desired constant voltage ratio between the input and output voltages is constant in a manner analogous to the constancy of speed of a mechanical speed governor. The amount of vari ation from constancy to cause a correction depends upon the amount of amplification of the pilot valve, being smallest for the largest amounts of amplification. The pilot valve has been shown as consisting of a single tube giving a single stage of amplification for the control but it mightjust as well be composed of several stages of amplification the resultant of which would be applied to the main valve.

Thismay be done by applying the voltage drops across the resistance 2|, 59 or 19 in Figs. 1, 2, or 3 to the input circuit of an amplifier of one or,

more stages. The output voltage of this amplifier is then applied to the grid 6, 39 or 59 in Figs.

1, 2 or 3, with the same instantaneous polarity r as the voltage across the resistances 2|, 59 or 19 with respect to the grid 6, 39 or 59. In this way a much more powerful correction is attained by virtue of the higher amplification. In the above disclosed modifications applicant has only shown the input signal as appearing across an inductance or a resistance but such might appear across a capacitance or a circuit containing inductance and capacitance together, or all three, as well.

While these circuits are theoretically linear as regards all types of distortion, the degree of linearity realizable is limited by the ability of the tubes to provide the necessary corrective voltages. If the demand upon the tubes exceeds their designed'ability to supply, the circuit is non-linear to the extent of their failure.

From the foregoing it will be perceived that I have invented anelectrical network which has the characteristic that the voltage developed across the input is proportional to the Voltage developed across the output.

I claim:

l. A control circuit comprising: an input cir cuit across which an input voltage is developed,

a main valve having a main valve control divided into a plurality of sub-controls and a main valve output circuit across which an output voltage is developed, a pilot valve having a pilot valve control and a pilot valve output circuit, a coupling between said input circuit and said main valve output circuit such that a function of said voltages may be obtained, means for impressing said function upon said pilot valve control, and means for impressing said input voltage upon one of said main valve sub-controls and for coupling said output circuit of said pilot valve to another of said main valve sub-controls.

2. A control circuit comprising: an input circuit across which an input voltage is developed, an output circuit across which an output voltage is developed, a main valve having a main valve control divided into a plurality of subcontrols and a main valve output circuit, a pilot valve having a pilot valve control and a pilot valve output circuit, a coupling between said input circuit and said output circuit, such that a function of said voltages may be obtained, means for impressing said function upon said pilot valve control, means for impressing said input voltage upon one of said main valve sub-controls and for coupling the output circuit of said pilot valve to another of said main valve sub-controls, and means coupling said main valve output circuit to said output circuit.

3. A control circuit comprising: an input circuit having: an input impedance across which an input voltage is developed, and an associated input impedance which is part of said input impedance and across which associated input impedance is developed a voltage drop proportional to said input voltage; a main valve having: a main valve control divided into a plurality of subcontrols, an output impedance across which an output voltage is developed, and an associated output impedance which is part of said output impedance and across which associated output impedance is developed a voltage drop proportional to said output voltage; a pilot valve having a pilot valve control and a pilot valve output circuit; a coupling between said associated impedances and said pilot valve control such that a function of said voltages may be obtained and impressed upon said pilot valve control; and means for impressing said input voltage upon one of said main valve sub-controls and for coupling the output circuit of said pilot Valve to another of said main valve sub-controls.

4. A control circuit comprising: an input circuit across which an input signal voltage is developed, a main valve having a main valve control divided into a plurality of sub-controls and a main valve output circuit across which an output signal voltage is developed, a pilot valve having a pilot valve control and a pilot valve output circuit, a coupling between portions of said input circuit and portions of said main valve output circuit such that for a desired linear relationship between input signal voltage and output signal voltage there will be in a portion of said coupling zero signal voltage; means connecting said portion to said pilot valve control; and means for impressing said input signal voltage upon one of said main valve sub-controls and for coupling said output circuit of said pilot valve to another of said main valve sub-controls.

5. A control circuit comprising: an input circuit having: an input impedance across which an input signal voltage is developed, and an associated input impedance which is part of said input impedance and across which associated input impedance is developed a voltage drop proportional to said input signal voltage; a main valve having: a main valve control divided into a plurality of sub-controls and a main valve output circuit having: an output impedance across which an output signal voltage is developed, and an associated output impedance which is part of said output impedance and across which associated output impedance is developed a voltage drop proportional to said output voltage; a pilot valve having a pilot valve control and a pilot valve output circuit; a coupling between said associated impedances such that and said associated impedances being so porportioned that for a desired linear relationship between input signal voltage and output signal voltage there will be in a portion of said coupling zero signal voltage; means connecting said portion to said pilot valve control; and means for impressing said input signal voltage upon one of said main valve sub-controls and for coupling said output circuit of said pilot valve to another of said main valve sub-controls.

6. A control circuit comprising: an input circuit across which an input voltage is developed, a main valve having a main valve control divided into a plurality of sub-controls and a main valve output circuit across which an output voltage is developed, a pilot valve having a pilot valve grid and a pilot valve output circuit, a coupling between said input circuit and said main valve output circuit such that a function of said voltages may be obtained, means for impressing said function upon said pilot valve grid, means biasing one of said main valve sub-controls so that said main valve will respond only to a signal of a given sign, and means for impressing said input voltage upon one of said main valve subcontrols and for coupling said output circuit of said pilot valve to another 01' said main valve sub-controls.

7. A control circuit comprising: an input circuit across which an input voltage is developed, a main valve having a main valve control divided into a plurality of sub-controls and a main valve output circuit across which an output voltage is developed, a pilot valve having a pilot valve grid and a pilot valve output circuit, a coupling between said input circuit and said main valve output circuit such that a function of said voltages may be obtained, means for impressing said function upon said pilot valve grid, means biasing said pilot valve grid so that said main valve will respond only to a signal of a given sign, and means for impressing said input voltage upon one of said main valve sub-controls and for coupling said output circuit of said pilot valve to another of said main valve sub-controls.

8. A control circuit comprising: an input circuit having: an input impedance across which an input signal voltage is developed, and an associated input impedance which is part of said input impedance and across which associated input impedance is developed a voltage drop proportional to said input signal voltage; a main valve having: a main valve control divided into a plurality of sub-controls and a main valve output circuit having: an output impedance across which an output signal voltage is developed, and an associated output impedance which is part of said output impedance and across which associated output impedance is developed a voltage drop proportional to said output voltage; a pilot valve having a pilot valve control and a pilot valve output circuit; a coupling between said asso that the output voltage of said main valve will sociated impedances such that and said assobe zero for any desired value of signal voltage on ciated impedances being so proportioned that for said main valve control; and means for impressa desired linear relations 'p betweeninput signal ing said input signal voltage upon one of said voltage and output signal voltage there will be in main valve sub-controls and for coupling said a portion of said coupling zero signal voltage; output circuit of said pilot valve to another of means connecting said portion to said pilot valve said main valve sub-controls.

control; means biasing said pilot valve control JOHN SCOTI CAMPBELL.

Images