US2452683A - Sweep circuit - Google Patents

Description

J. W. RIEKE SWEEP CIRCUIT Nov. 2, 1948.

2 Sheets-Sheet 1 Filed March 21, 1944 FIG. IA

TIME

TIME e w i v m L P H\ A" 5 0 M? 1 E 2 G n FIG. 3A

TIME

TIME

Nov. 2, 1948. J. w. RIEKE 2,452,683

SWEEP CIRCUIT Filed March 21, 1944 2 Sheets-Sheet 2 Ha. 5A

Za 1/00 TIME INVENTOR J W R/E/(E AGENT Patented Nov. 2, 1 948 7 i N UNITED STATES PATENT OFFICE SWEEP CIRCUIT John W. Rieke, New York, N. Y., assignor to Bell Telephone Laboratories, Incorporated, New York, N. Y., a corporation of New York Application March 21, 1944, Serial No. 527,457 2 Claims. (Cl. 25036) This invention relates to an improvement in Fig. 3A is a diagram of one form of the circuit sweep circuits producing voltages of saw-tooth embodying negative feedback; wave form, and is particularly useful when it is Figs. 33 and 3C show, respectively, the voltage desired. to obtain a voltage varying over a Wide wave approximated by the circuit of Fig. 3A and range linearly with time to a high degree. the slope of this wave;

The general object of the invention therefore Fig. 4 is a diagram of a circuit electrically is to provide an improved circuit for the generaequivalent to the circuit of Fig. 3A but adapted to tion of saw-tooth voltages. the introduction of further refinement;

The fundamental form of a saw-tooth voltage v Fig. 5A is a circuit including the refinement algenerator includes a sweep condenser which is luded to in connection with Fig. 4; rapidly charged from a source of voltage and then Fig. 5B is a graph of the time variation of voltdischarged at a controllable rate through areage e of Fig. 5A; sistor. Alternatively, the condenser may be Fig. 6 is a circuit for producing a sweep voltcharged at a desired rate and then abruptly disage decreasing uniformly with time. charged. In this simple apparatus the varia- In all figures like elements are represented by tion in condenser voltage is never strictl linear lik numerals and letters. with time, even over a short interval after the Referring first to Fig. 1A, at the time switch 3 beginning of the controlled charge or discharge is closed condenser C has been charged by batcorresponding t a v t change small in tery E to a voltage appearing across the small parison to t maximum v a of t condenserresistance 1'. When switch Sis opened condenser Numerous circuits have been devised to improve (3 begins t charge through resistance R and the linearity of t s voltage and severa o t se tinues to charge until S is again closed. In Fig. are described y Q Puckle in "Time 138 it is shown that the voltage e across condenser Chapter VII, published in 1943 Y Chapman & C rises at first approximately linearly but more Hall, and more slowly as the charge increases. The

The circuit of the present invention enables the rate f Voltage rise from the time h when s is desired linearity to be obtained throughout a toopened to $2 when S is again closed is Shown in tal voltage ha e Which is a large fraction of the Fig. 1C. Condenser C has at any instant between maximum voltage on the sweep condenser and th opening and reclosing of s the voltage this enlargement of the useful range of the volt- 0 t age variation is another object of the invention In the circuit to be described, the optimum use e E 1 (1) is made f known expedients i a novel Omaniin which t is the charging time counted from the zation. By this organization there is made pos- Opening of Switch 8- sible the attainment of substantially perfect lin- If e is a v e used. r-ex pl t p s t on earity of time variation of voltage both for Very the electron spot on the screen of a cathode ray rapid and for very slow rates of voltage change. oscilloscope, successive positions of this spot may Accordingly, another object of the invention is correspond linearly to successive values of e but to provide a swee circuit capable of generating a not so to successive instants of time. There revoltage varying linearly with time over a wide 40 sults aspot displacement error graphically shown range of voltage variation and over a wide range by the curve of Fig. 55 on page 82 of Puckles work of time rates of such variation. Time Bases, above referred to. This curve The invention is to be understood from the shows that the error in spot velocity increases following description read with reference to the much more rapidly than the error in spot posiaccompanying drawings, in which: tion. It is, of course, desired to render negligible Fig. 1A is a diagram of a fundamental fo Of the error in spot velocity, that is, to bring the sweep circuit; curve Figs. 13 and 10 exhibit respectively the variad2 ticn with time of the voltage appearing across the sweep condenser of Fig. 1 as this condenser is being charged, and the variation with time of the a 6 n y t9 a horizontal strr rate of change of this voltage; thus substantially eliminating the error in spot Fig. 2 is a circuit illustrative of the introducposition even cendenser C is h r t a n tion of negative feedback for linearizing the voltsiderable fraction of the voltage of battery E.

age wave of Fig. 1B; Referring now to Fig. 2, amplifier it? is introduced into the charging circuit, its input being connected in shunt across terminal ii of resistance R and terminal I2 of battery E remote from junction l3 of battery E and resistance R. Amplifier I is understood to include an odd number of phase reversals between its input terminals l, 3 and its output terminals 2, 3. Ground connection may be made at any of the terminals l, 2 or 3. These numerals are encircled in all figures where they appear. If at any instant charging current i flows in resistance R, the input voltage to amplifier It! is e1=E-2'R resulting in an output voltage e2=/.Le1, where a is the amplification factor of amplifier it. The plates of the condenser C are connected as shown in Fig. 2, namely, one plate to terminal ll of resistance R, the other to output terminal 2 of amplifier ill. The output voltage of amplifier II] is thus, so far as concerns condenser C, in series with that of battery E, and the voltage across condenser C when switch S is opened is given by the formula which may be compred with the voltage of condenser C in Fig. 1A, namely,

z e=E 1-E The linearity improvement due to the negative feedback from amplifier it may be realized by rewriting the above formulae as follows: Without feedback:

we have t =E 1+n( second order distortion terms, factor is required not to exceed de E the time of charge without feedback must be no greater than t=ARC, whereas with feedback this charging time may be t=ARC (1+;r). In the respective cases the maximum value of e is AE from Equation la, but AE (1+ u) from Equation 2a. The allowable change in condenser voltage for a prescribed non-linearity is thus greater in the ratio 1+,u. with negative feedback than with out,

In Fig. 3A vacuum tube V! is used as amplifier it of Fig. 2. Battery !5 supplies 300 volts to anode l6 of VI, of which grid I1 is connected to junction I l of R and C. Cathode l8 of VI, heated by conventional means not shown, is connected to negative terminal I2 of battery E. It will be observed that points 1 and 3 of Fig. 2 correspond respectively to grid I! and to cathode is in Fig. 3A, while the plate of condenser C remote from resistance R is grounded. For voltage variations at anode Hi, this element is grounded through battery l5 and so is effectively connected to condenser C. The ground here corresponds to point 2 of Fig. 2. Such a circuit has been found useful for very rapid sweeps lasting for intervals of the order of microseconds between opening and reclosing of switch S. Actually, switch S for such use is an electronic switch as will later appear.

Sweep voltage is obtainable from the circuit of Fig. 3A either across condenser C or between cathode l8 and ground. At the latter sweep terminals the variable voltage is where e is the condenser voltage given by Equation 2.

Suitable values for the voltages and circuit elements in Fig. 3A are as follows: E=50 volts; R=200,000 ohms; 0: 50 micromicrofarads; =20. RC is here 50 microseconds and from the formula itil di RC the voltage e will rise when S is opened at the rate of about one volt per microsecond from its initial value, namely, the small voltage across 1' when S is closed. Tube V1 is suitably a triode, such as one-half of a GSN'T-GT, while 1' may be a few hundred ohms. Ii switch S is opened for 100 microseconds, e will rise to approximately 100 volts in this interval, falling short by some 4 per cent as shown by Equation 2a. The charging energy is, of course, supplied by battery l5 and E is relied on to fix the direction of flow of current i in resistor R and to control the rate of rise of the voltage across the sweep condenser.

In the circuit of Fig. 4 use is made of the fact last stated. Here negative terminal 12 of battery E is connected to the plate of condenser C remote from ground, positive terminal I3 is connected to grid ll of tube VI. When switch S is closed, the currentv flowing in tube V! biases cathode I8 some I5 volts positive to grid H. The controlling voltage across resistor R. then is the sum of E and the cathode to grid voltage Eg, or E=50+15:'65 volts. On the opening of switch S, the voltage across condenser C rises as in the case of Fig. 3A. In the circuit of Fig. 4 sweep voltages of nearly the same value are obtainable either across the sweep condenser or between cathode l8 and ground.

Neither the circuit of Fig. 3A nor that of Fig. 4 provides absolute linearity. The curves of Figs. 33 and 3C are ideal curves and are approached much more closely when limited am- 5 plification is available if the negative feedback improvement already described is supplemented by a circuit integrating the output sweep voltage.

Referring now to Fig. 5A, a pentode V3, suitably a 6AC7, replaces resistance r and switch S between point l2 and ground. Battery E is replaced by the voltage across condenser Co, a voltage of 50 Volts supplied from battery [5 through tap 29 on a voltage divider consisting of resistors 2! and 22 of resistances 100,000 and 20,000 ohms respectively. Condenser Co is charged to such voltage through diode V2 which may be one-half of a 6H6. Plate 23 of tube V2 is connected to tap 2e and cathode 2 5 is connected to one plate of condenser Co and to grid 11 of tube Vl. Resistor 25, preferably of some 2 megohms resistance, shunts condenser C and provides a conductive path over which anode voltage is supplied to anode so of tube V3.

Tube V3 is normally conducting, its cathode 3| being grounded and connected to its suppressor grid 32. The potential of screen grid 33 is controlled through 70,000-ohm resistor 35 from battery and is shunted to ground through 0.1-microfarad condenser 38, which with resistor 35 constitutes a filtering circuit for the voltage of screen 33. A negative voltage pulse np, derived from whatever device is used for timing the start of the sweep voltage, is applied to grid 34 to annul the conductivity of tube Vafor a time interval which may be 100 microseconds from to to two. The anode-cathode path of tube V3 absorbs, while V3 is conducting, a small fraction of the voltage between tap 2a and ground, corresponding to the drop across resistor r in Figs. 1A to 4, inclusive. Switch s, here replaced by tube V3, is abruptly opened at time to and equally abruptly closed at time two.

At the moment of extinction of the conductivity of tube V3, the full voltage from tap 20 is effective to control the charging of condenser C through resistance R. Condenser C0 is preferably of 0.006-microfarad capacitance, while resistor 25 is as stated of 2 megohms resistance so the time constant R2500 is more than 100 times the sweep interval, wherefore the voltage between grid 5? and point I2 is substantially constant. C0, controlled in voltage by diode V2, thus replaces battery E of the preceding Figs. 1A to 4, inclusive. The charging of condenser C begins when tube V3 ceases to conduct and continues to the time of disappearance of the negative voltage pulse on grid 3t. When tube V3 again conducts what small change in the voltage of condenser Co has taken place in the charging interval of condenser C is overcome practically instantly. C0 is charged from diode V2 through the small resistance ofiered by tube V3, and this is a few hundred ohms making the time constant of this path a few microseconds. Cathode is of tube Vi is conductively grounded through resistor R and tube V3 when the latter is conducting, but is ungrounded during the sweep just as in Figs. 3A and 4, except when voltage e is applied across a conductive load. During the sweep, cathode 2 is at a higher potential than anode 23, hence tube V2 is not con-- ducting.

There is added in Fig. 5A the integrating circuit R and C. tween ground and condenser C and resistor R is connected between the junction of these condensers and cathode 53. The sweep voltage from the circuit of Fig. 5A is taken between cathode leiand ground, which again corresponds to point Condenser C is interposed bee in *1+t(r r where 22 is the operator and 4 are the normal mesh equa- Equation 4a relates the voltage of Equations 3 tions, while grid I1,

7'; lg n 1 to the voltage e of cathode 18.

The detailed analysis is omitted here for the sake of brevity but it may be shown that .QtQ. I l l Wig E 1+ D1+ D2+ 5 when D1 and D2 are distortion terms, Dz involving D1 given by which is to be made zero to eliminate completely the first order of distortion in de dt When D1 is made zero by choice of R and C for the preestablished values of R, C and M, D: becomes If R, a and C are already established, and it is desired to make C'=n C, the proper value of R is from Equation 5a.

and

100 1 R (1+5) and R"C"=.13;RC..approximately;

. In Equations 3, and 6 the voltage E is the sum of the voltage E, namely 50 volts, and the voltage of cathode IS with respect to grid ll which is volts in the circuit illustrated in Fig. 5A.

From Equation 5, for an undistorted sweep,

If a sweep of 100 volts is desired, as illustrated in Fig. 5B, in 100 microseconds, one may take 4:20; C'=C=1000 micromicrofarads, R=124,000 ohms, in which case becomes 10 or one volt per microsecond, and t=100 10 seconds the sweep desired. In this case R=4=.75R=590,000 ohms and the second order term becomes one-third of one percent. This is error in the rate of voltage rise at the end of the 100-microsecond interval and according to the customary definition measures the non- 1inearity of the voltage-time curve.

A feature of the circuit of Fig. 5A is the connection of diode V2 in such wise that a constant charge is maintained on condenser Co unafiected by the average charge of condenser C. I am aware that diodes have previously been used to supply constant potential to a condenser located as is Co in the circuit of Fig. 5A but in the circuits so far known the junction 4! of R and C is connected directly to grid l1 and condenser Co is inserted between R and cathode H8. The diode connection used in the present invention provides for the rapid restoration or any voltage lost by Co in the sweep, even if there is a variation in the repetition rate grid 34 of tube V3, so that for each sweep the voltage on condenser Co shall be the same. In the circuit of Fig. 5A. when the negative signal passes from grid 34 of tube V3, the diode charges C0 through the small resistance of tube V3. If this resistance is 1000 ohms and the capacity of Co is .006 microfarad, the time constant of 6 microseconds determines the recharging of Co whereas in some heretofore known circuits Co is charged through a large resistance connection between cathode I3 and ground, and its normal voltage may not be fully restored before the appearance of a succeeding negative pulse on grid 34.

The combination of an integrating circuit with negative feedback is most important where the available amplification is limited. When ,u is sufficiently large, the integrating circuits RC (Fig. 5A) may be omitted. This is done in the circuit of Fig. 6 where tubes V1 and V5 constitute a direct current series amplifier succeeded by cathode follower tube V4. Cathode $5 of tube V4 is connected to ground through resistor 45, suitably of about 100,000 ohms resistance. Ground here is at point 3; condenser C is connected between point 2 (cathode 35 of tube V4) and point i (grid I! of tube V1); and the sweep voltage is taken across resistor 48. As will appear, this sweep voltage may be regarded as a rising negative voltage of the negative voltage pulse on 8 progressively canceling an initial charge or condenser C.

In a particular application it was desired to cause the charge of condenser C to decrease from 200 to 100 volts in an interval varying from 100 to 400 seconds. With R=3 megohms, 0:4 microfarads and E=3 volts, there is a cancellation of the initial charge of condenser C at the rate of one volt in four seconds. It has been found possible with the circuit of Fig. 6 to obtain a sweep of 100 volts in 400 seconds with a distortion of less than one per cent at the end of the sweep interval. It will be seen from Equation 2b that to reduce the first order distortion term to seventenths of one per cent with RC=12 and t=400 seconds, the value of a required is 5000. For 15:12, with =5000, the 100-volt change across resistor 46 occupies 100 seconds and the distortion in becomes one-sixth of one per cent.

In Fi 6 tubes V1 and V5 are two triodes, suitably the two parts of a GSL'IGT constituting with tube V4 a direct current series amplifier. Voltage to anode 50 of tube V4 is supplied directly from battery It. Anode [6 of tube V1 is connected directly to cathode 5i of tube V5. Of the latter tube grid 52 is made positive with respect to cathode 5! by battery 53 which may conveniently be volts derived through a voltage divider from battery I5, and anode 54 is supplied from battery l5 through resistor 38. From the end of resistor 38 remote from battery l5 anode 56 is connected through resistor 56 to grid 49 of tube V4, a 6SN7.

Condenser 57 is in circuit between the ground and the junction of resistors t8 and 56 and with resistor 56 serves to prevent oscillations of voltage on grid 69. Resistor 50 is of the order of ohms while condenser 57 is about .00025 microfarad capacity. Through resistor d8 of about 10 megohms resistance, voltage is supplied from battery I5 to anode 5 3 of V5 and thus from cathode 5! to anode iii of V1. Cathode 48 of V1 is grounded through resistors 58 and 59 in series- Across these resistors between cathode i8 and ground is impressed a voltage derived through resistor 60 from battery 15 to provide a suitable biasing voltage for grid l'l. Resistor 58 is made variable for this adjustment. Between ground and grid I I are connected battery E and resistor R in series as in Fig. 1. Resistor 5| of about 80,000-ohms resistance provides positiv feedback from cathode 45 of V4 to the junction of resistors 58 and 59 which are conveniently 200 and 100 ohms respectively.

When switch S is closed battery I5 charges condenser C through the small resistance r and the low resistance to ground of resistors 58 and 59 and the grid-cathode path of V1, grid I! being now positive to cathode H3. The voltage now applied to charge condenser C is that across resistance i6, which is suitably chosen twice 1'. Condenser C is thus charged to 200 volts, positive toward point 2 and negative toward point 1, grid I1 becoming slightly positive with respect to cathode I8. The time constant RC is of the order of seconds for which reason the specifically described arrangement of Fig. 6 is adapted to slowly as well as to rapidly repeated operation. On the opening of switch S grid H tends to become negative to cathode i8, and battery E takes control of the circuit to discharge condenser C at a rate determined by the voltage drop across R. As the charge on condenser C begins to decrease the potential at point 2 begins also to fall. This results from a minute potential rise at grid I! of V1, which change is reversed at anode l6 wherefore the potential at cathode of V5 falls. Since grid 52 is maintained by battery 53 at a fixed voltage with respect to ground the fall of potential at cathode 5| is reflected as a fall also at anode 54. This potential fall at anode 54 appears also in grid 49 of V4, a cathode follower tube of which cathode 45 follows closely in potential grid 49. The fall in potential in cathode 45 instigated by the rise of potential at grid I 7 is returned through condenser C to oppose the change in potential at grid IT.

The negative feedback from cathode 45 to grid 11 thus checks nearly completely the tendenc of grid I! to rise in voltage. The amplification involved is that of the direct current amplifier consisting of tubes V1 and V5 in tandem and to raise this factor to the value #:5000 positive feedback is provided through resistor 6| from cathode 45 of V4 to cathode 18 of V1. Point 1 therefore remains at substantially constant potential, that of cathode I 8, and E being constant there is a constant voltage drop across resistor R through which a constant current flows cancelling linearly with time the original charge on condenser C. Point 2 progressively falls in potential as C is discharged and this falling potential is the decreasing sweep voltage taken on across resistor 46. In practice, it has been found that the potential at grid I! actually changes about .02 volt during a sweep over a range of 100 volts.

In such a sweep the first order distortion term from Equation 2b is t RC(1 +11) For t=100 seconds, RC=12 and =5000, the distortion is .17 per cent. With the specified circuit elements of Fig. 7 the initial closing of switch S charges condenser C through grid I! to about 200 volts and this voltage is reduced to 100 volts at a rate of decrease determined by the voltage of battery E. It is obvious that a faster cancellation of the voltage on condenser C may be obtained by the proper selection of R, C and E.

What is claimed is:

1. A circuit for the production of a sweep voltage varying linearly with time at a rate determined by a unidirectional control voltage comprising a first, a second and a thirdvthermionic vacuum tube having each at least a cathode, a control grid and an anode, a first resistance connected between ground and the cathode of the first tube, a second resistance connected between ground and the cathode of the third tube, a direct connection between the anode of the first tube and the cathode of the second tube, a conductive connection between the anode of the second tube and the control grid of the third tube, power supply for said tubes including a source of unidirectional voltage conductively connected negatively to ground and positively to the cathode of the first tube and to the anodes of the second and third tubes, a sweep condenser connected between the cathode of the third tube and the control grid of the first tube, a source of unidirectional control voltage connected negatively to ground and positively through a third resistance to the control grid of the first tube, a fourth resistance connected between the cathode of the third tube and a point on the first resistance and a switching circuit connected between the cathode and the anode of the third tube operable directly to charge initially the sweep condenser and thereafter reversely to enable said condenser to be discharged at a rate determined by the magnitude of the control voltage, thereby providing across the second resistance a voltage decreasing substantially linearly with time.

2. A circuit for the production of a sweep voltage varying linearly with time at a rate determined by a, unidirectional control voltage comprising a first, a second and a third thermionic vacuum tube having each at least a cathode, a control grid and an anode, a first resistance connected between ground and the cathode of the first tube, a second resistance connected between ground and the cathode of the third tube, a direct connection between the anode of the first tube and the cathode of the second tube, a conductive connection between the anode of the second tube and the control grid of the third tube, power supply for said tubes including a source of unidirectional voltage conductively connected negatively to ground and positively to the oathode of the first tube to the anodes of the second and third tubes, a sweep condenser connected between the cathode of the third tube and the con trol grid of the first tube, a source of unidirectional control voltage connected negatively to ground and positivel through a third resistance to the control grid of the first tube and a switching circuit connected between the cathode and the anode of the third tube operable directly to charge initially the sweep condenser and thereafter reversely to enable said condenser to be discharged at a rate determined by the magnitude of the control voltage, thereby providing across the second resistance a voltage decreasing substantiall linearly with time.

JOHN W. RIEKE.

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

UNITED STATES PATENTS Number Name Date 2,237,425 Geiger et al Apr. 8, 1941 2,251,973 Beale et al. Aug. 12, 1941

Images