US2608654A - Pulse-forming circuit - Google Patents

Description

Patented Aug. 26, 1952 PULSE-FORMING CIRCUIT Jabez C. Street, Belmont, Mass, assignor, by mesne assignments, to the United States of America asrepresented bythe Secretary of the Application March 18, 1943, Serial No. 479,660

26 Claims. 1 This invention relatesto electrical pulse-forming devices and particularly to a circuit for forming a. square :pulse. I

On'eof the objects of the invention is to provide :a circuit for producing a substantially square pulse of a predetermined duration when some condition of the circuitris changed.

Another object of theinvention is to utilize the time delay of tastepnwave .in a transmission linle to produce 'a substantially square voltage pu se.

Still another object of the invention is to pro-.

vide a network by means of which a substantially square pulse is produced bythe mere closing of switch. I

Other objects and objects relating particularly to the manner .of'connecting the various components of the circuit will be apparent as the descrlptionof the invention proceeds.

Illustrative embodiments of the invention are forms of pulse-forming circuits according to this invention.

I have observed that if a continuous transmission linebe charged: to, a-particular voltage and then connected to load having an impedance substantially equal to the nominal value of the characteristic impedance of the line, a shaping action will resultonthe discharge of the line capacity through the. line impedance. This phenomenon may be put to a very useful purpose by means of thearrangement of the circuit of Fig. 1 in which the mere clo ng of a switch will produce a substantially square :voltage wave across an impedance. In this circuit a networkof the uniform laddertype is ;formed including inductances I0, II, along one :sideof the line, and capacitances Il, I8, I9 and 20 connected across the line. "The capacitance I7 is connected betweenthe juncture of the inductances I0 and II tota'wire 2| forming the other side of the circuit;'the capacitance I8 is connected between the Junctureof the inductances II and I2 and the wire 2|; the capacitance I 9 is connected between the juncture of the inductances I2' and I3 and the Wire ZI; while the capacitance .20 is connected betweenthe outer I2 and 13,:connected in series r end of the inductance I3 and the wire-2|. The inductances II-I3 inclusive "are equal in value and the capacitances I'I-I 9 inclusive are equal in value, and the values of the inductances and capacitances are chosen in a manner to be hereinafter described. Theinductance lfl has half the value of the inductances I I-I3, inclusive, and the capacitance 20 has half the value of the capacitances I'I-I9, inclusive, in accordance with the well-known uniform ladder approximation of a transmission line.

.The uniform ladder type of line may be re garded either as a cascaded series of Tsections or as a cascaded series of 1r sections" with appropriate terminations of either the :midshunt or midseries type. Since in the circuit of Fig. l the far end of the line is intended to be open circuited, a midshunt termination is provided (1. e. at the condenser 20). In the circuit of Fig. 1 the other end of the line is provided with a midseries termination (i. e. through the series inductance Ill). The image impedance of the line at the input terminals at this midseries termination therefore correspondsto the characteristic impedance of a T section of the line, sometimes referred to i as the "midseries characteristic impedance of the line. This impedance varies somewhat with frequency, but the nominal value of the characteristic impedanceof the line, which is a sufiiciently close approximation of either the midseries or midshunt impedance at lower frequencies, is a constant with respect to frequency, and is therefore more conveniently used in connection with square pulses, which contain components of a number of frequencies.

The more sections used in the line (which may be regarded as either T or 1r sections) the greater will be the accuracy of the approximation of a rectangular pulse produced by the response of the line, and the dotted line 22 is intended merely to show that other sections may be added between the inductances I0 and I l and between the capacitances i1 and I8.

An impedance 23 is shown connected across the line between the outer end of the inductance I0 and the wire 2!, and this impedance may repre sent the impedance of a load circuit which is to be operated or driven by the pulse. A switching device 25 is connected between the impedance 23 and the inductance Ill and serves merely to open and close the circuit. A triode tube may be used for this purpose, and since it is desirable to have as low an impedance as possible in the switching device, the tube may be of the gas-filled type becoming conductive when the grid 25 thereof is swung in a positive direction above a predetermined critical potential, whereupon the control is taken away from the grid and the tube continues to conduct until the voltage across it falls below a predetermined value. The grid 25 may be given a suitable bias potential from a source, indicated at 26, through a resistance 21. The trigger pulse to close the switch by making the grid 25 positive may be applied to the grid through a condenser 28.

A source of potential is connected across the line in order to charge capacitances "-20 inclusive, and this source, indicated at 29, is shown connected through a high charging resistor 30 to the outer end of the inductance [3 although it might be connected anywhere along the line.

The inductances and capacitances are so chosen that the nominal values of the characteristic impedance of the network will substantially equal the load impedance 23. If desired the network may be made so that the midseries characteristic impedance for the frequency of is equal to the load impedance (6 being the time duration of the pulse in microseconds). This represents only a small variation from the results obtained using the nominal value relation and intended to be understood as included when it is stated that the nominal characteristic impedance is approximately equal to the impedance of the load.

Assume then that the switch 24 is open, the tube being maintained non-conducting by the negative potential on the grid 25. Energy will then flow into the'capacitances I7, 18, I9 and 25 from the source of potential 29 through the high resistor 35, and when suificient'time has elapsed these capacitances will be charged to a potential which can be assumed to have a value E. At this time the potential across the load impedance 23 will be zero.

In Fig. 2 the voltage across the line is plotted for different positions on the line from the impedance 23 at one end tothe condenser 20 at the other end at a time when the switch '24 is open and the capacitances "-20 have been fully charged. The voltage across the impedance is .zero,but the voltage across the line from a point just to the left of the impedance 23 to the other end of the line has a value of E.-

Now let it be assumed that a short positive trigger pulse swings the grid 25 of the tube 24 sufficiently positive to make the tube conduct. This effective closing of the switch connects the impedance 23 directly across the line. The capacitances H-Zil of the network will then attempt to discharge through the load impedance 23. The

V voltage E of these capacitances will then be distributed between the characteristic impedance of the line and the impedance of the load 23, and, since these impedances are equal, a voltage of will appear across the impedance 23. It will thus be seen that where before there was zero potential across the impedance 23 there has suddenly appeared a voltage When the switch 24 is closed the first condenser starts to discharge through the impedance 23. But the how of current is held back by the time const t of the T section, this time being equal to \/LC' where L is the sum of the two inductances of the T section and C is the value of the capacitance. At the same time the capacitance l8 starts to discharge into the capacitance H, the capacitance I9 into the capacitance l8, and the capacitance 20 into the capacitance E9. The eifect of this action is to produce a negative voltage wave of the value which, beginning at the impedance 23 will travel towards the left along the line.

When this wave has travelled a certain distance along the line the voltage along the line will appear as is indicated in Fig. 3. The voltage across the impedance '23 will still be and this same voltage will appear a certain distance along the line, but the voltage at the left end of the line at the condenser 20 still has a value of E.

Eventually the front of the voltage wave will reach the left end of the line and then the voltage along. the line from the impedance 23 to the condenser 20 will have a value of as shown in Fig. 4. The resistance 30 is high enough so that at this point the line has the effect of being open. Hence, the voltage wave will be reflected from the end of the line'at the condenser 20 and will start back without change of phase toward the impedance 23. As the wave travels it will reduce the voltage to zero, and hence some time later the voltage across the line might be represented as indicated in Fig. 5. Here the voltage wave has been reflected from the end of the line and has moved part way back so that the voltage across the condenser 28 is zero, while there still is a voltage of across the impedance 23. Finally, however, the front of the voltage wave reaches the impedance 23 and the voltage across it is reduced to ZeIO.

It will be notedthat the voltage wave passes through the line twice, first in one direction and then in the other. The time required for it to pass through one section is equal to the /LC and therefore the time, 6, required for it to pass twice through the line will be equal to 2N /LC where N is the number of sections employed. The line shown in Fig. 1 contains an extra halfsection, whether the line is regarded as a cascade of T sections or as a cascade of 1r sections, so that 2N in this case will be an odd number.

When the potential across the impedance 23 drops to zero there will be zero potential across the gas tube 24 and therefore the tube will become non-conducting and will again be placed under the control of its grid 25 which is by this time sufilcientlynegative again to be below the critical potential. The capacitances If, l3, l9 and 20 thereupon :begin to charge again, fed from the source 29 through the high resistance 30.

' The pulse thus produced across the impedance 23 may be represented in Fig. 6. The top of the pulse is somewhat wavy and the number and amplitude of these waves will depend on the number of sections in the circuit. Each section will produce a wave at its resonant frequency and the greater the number of sections the nearer fiat the top of the wave will be. An infinite number of sections will give a perfectly flat-topped pulse. However, practical considerations require that not too many of the sections be used. As many as 40 have been used, but from 4 to 15 will usually be found to be satisfactory; 5 or 6 sections have been used with good results.

The values of the lnductances and the capacitances for the network of the invention for a given pulse width may be determined when the load impedance and the number oisections to be used are known. For example, suppose that a pulse of 1 microsecondwidth is desired from a network consisting of 7% sections with a midseries connection as in Fig. 1 feeding into a loa whose impedance is 750 ohms.

Let L represent the inductance of each of the chokes ll, l2 and i3 and let represent the magnitude of the capacitances l7, I8 and [9. The time, a, required for the wave to pass N sections twice, first in one direction and then in the other, has been given above as:

d=2N /L C Now the nominal value of the characteristic impedance, Z, of a uniform ladder network, will be equal to and it is desired to make this impedance equal to the load impedance which has been assumed to be 750 ohms. The following equations may then be obtained by multiplying by 6:

from which henrys or 50microhenrys.

The capacitance, C, may then be found by solving for C in the equation:

L we li 5o 1c ss.se farads or Z 750 88.88 micromicrofaracls In accordance with the known properties of uniform ladder networks, the values L and C as above derived correspond respectively to the magnitude of each of the chokes II, I? and I3 and to the magnitude of the condensers ll, 18

Thus:

and I9. The inductance of the choke It] will accordingly be e'qualto too large to prevent the condensers being charged in this particular period of time. However, the resistor 30 must be large enough to prevent a voltage being maintained across the tube 24 sufilcient to maintain its conductivity after the network hasdischarged and also lar e enou h so that the impedance thereby placed across the line is very high compared to thecharacteristic impedance of the line.

. The value .of the resistor 30 may be obtained if the capacitance of the circuit and the repetition frequency of the pulse are known. If a repetition frequency ofv 1000 pulses per second is desired, the charging time, t, of the capacitances may in the case of short pulses be considered to be 1000 microseconds, neglecting the time of discharge. Since the charge on 1a. capacitance charged from discharged condition by the sudden application of a given voltage reaches ap proximately 63% of the full charge of such voltage in the time equal to RC, it can be assumed to be almost entirely charged at 3 times that time, or 330. The charging time, 12, may then be set equal to 3RC (where C equals the total capacitance) and the; equation solved for R,

thus: i v

A midshunt connection-may be used on the line instead of the midserie'sflconnection shown in Fig. l. The line or network then appears as in the circuit of Fig. 7 and may be regarded either Fig. '2, since it has a condenser directly across as a cascade of an integral number of 1r sections or else as a cascade of T sections with a termi nating half section on each end.;'- For the same nominal value 'of the characteristicimpedance of the network, the inductances :31 and the condensers 32 will have the same values as'the inductances ll -l3 and the condensers Il-l 9, respectively, of Fig. l. 'I'heconde'nsers 33 at each end of the line have half thecapacitance of the condensers 32, and are thus equal to the condenser 20 of Fig. l. The midshunt image'impedance of a uniform-ladder network is actually slightly dlfierent from the midseries image impedance of the-same uniform ladder network, but for the practice of this invention it is convenient to employ only the nominal characteristic impedance of the network, which, is the square root of the product 01' the midshunt and the midseries image impedances, and in the case of a purely reactive network of the forms shown in Fig. 1 and Fig. 7 is given by the simple expression a The calculations of the desired inductances and capacitanoes from the desired image impedance, or vice versa, using, the formula for the midshunt or the midseries impedance according to which form of termination is used,.are relatively more complicated, but may be used if an exactmatch is desired. 1 1

The midshunt type oftermination shown in the line terminals, in generalhas the advantage of providing a pulse with a particularly rapid initial rise. The initial rate of rise may be further increased, at the expense of increasing the height of the irregularities at the top of the pulse, especially the overshoot, by increasing the size of the input condenser beyond the theoretically given value; The additional ripples may be of no momentat all if the pulse i'sto be fed to an amplifier which has a limitingaction. It is to be understood that within the scope of this invention the size of the reactive circuit components may be varied to modify the pulse shape in a desired way or to compensate for. the existence of distributed capacitance in the inductance coils, stray capacitance in the wiring, etc. For this purpose it is particularly effective to vary the magnitude of the reactive components at one end or the other of the ladder network. Also, in the case of the circuit of Fig. 1, a modification of this sort may be made by simply adding an adjustable condenser across the input terminals of the reactive network and ad justing it for the best rate of rise consistent with ripple tolerances. The behavior of the circuit can readily be monitored with a cathode ray oscilloscope having a fast sweep.

In the circuit shown in Fig. 7 the voltage for charging the line is connected by the wire 35 to that end of the line which is connected to the switch 24 instead of at the far end of the line as in Fig. 1. Other positions for connecting the charging voltage may, as previously noted, also be used. In the circuit shown in Fig. 7 the charging voltage is connected to the line through a choke 36 instead of through a high resistance 30 as in Fig. 1. The choke 36, like the high resistance 3D has a current limiting function and acts to maintain a voltage drop between the line and the voltage source while the former is being discharged. Charging through a choke as in Fig. '7 has the advantage of avoiding the losses occurring in the resistor 30 of Fig. 1. The choke 36 is preferably designed to resonate at the frequency of charge and discharge with the efiective storage capacitance of the reactive network, substantially the sum of the individual capacitances, so that a maximum charging voltage can be obtained. Normally the device 24 will be operated at a definite recurrence rate for which the choke 36 may thus be designed.

Various type of switching arrangements may be employed in circuits provided according to the present invention. Instead of the gaseous discharge tube 24, for instance, a spark gap, preferably of the rotary type, might be used. One form of circuit for generating electric pulses by means. of a network provided in accordance with the present invention which is adapted to be discharged upon the firing of a spark gap is shown in Fig. 8. In the circuit of Fig. 8 the pulse forming network is shown including the inductances 4|, 42, 43, 44, and 45 and the capacitances 46, 4'], 4B, 49, '50, and The pulse forming line is constructed in the same manner as the line shown in the circuits of Fig. '7. In Fig. 8, instead of being indicated simply by a resistance, the load is represented as a radio transmitting tube 53 having a cathode 54 and an anode 55. In parallel with the load is a choke 56 and a diode 51. The choke 56 and the diode 51 are rendered desirable in this circuit because of the fact that the transmitting tube 53, particularly if the tube 53 is a magnetron type of tube, has an impedance which is slightly capacitive and also varies with applied voltage, generally causing an insufficiently rapid termination of the pulse produced in the circuit. The choke 56 tends to accelerate the termination of the pulse which would otherwise be delayed by the shunt capacitance inherent in the construction of the transmitting tube 53, which would discharge in the wellknown exponential manner. The diode 51, which is oppositely polarized to the transmittingtube '53, serves to prevent the current in the choke 56 from setting up undesired oscillations by short circuiting the choke 56 as soon as the voltage across it becomes negative with respect to that impressed across the choke during the period of the pulse.

The network line is charged from a source of high voltage indicated generally at 60 through a choke 6! which functions in a manner similar to the choke 36 of Fig. '7. After the network is charged the spark gap 62 is actuated, as by suitable relative rotation of the electrodes, to permit a spark discharge to occur between the electrodes, thereby permitting the network line to discharge through the transmitting tube 53. The occurrence of the spark discharge at the gap 62 tends to short circuit the power'supply but the choke 6| so limits the current that when the pulse is completed the voltage'across the spark gap drops to such a low level that the spark is interrupted immediately. During the charging of the line, current may flow through the choke 56 and diode 51, but since such current will produce a voltage oppositely directed to that impressed upon, the transmitting tube 53 by the discharge pulse, such voltage will have no eiiect upon the transmitting tube. The circuit of Fig. Smay also be operated with the other side of the load grounded, in which case the circuit would be fully analogous to that of Fig. 7 except for the different representation of the switch and of the load. If the connection between the load and the line is to be grounded and the load is to be a transmitting tube which it is desired to operate with a grounded anode, it will then be necessary to reverse the polarity of the charging voltage, which is to say that the charging voltage should then be a negative voltage with respect to ground. r

Fig. 9 is a circuit for the formation of pulses in accordance with the present invention which may be regarded as one employing a uniform ladder network the ends of which have been. connected together, or it may be regarded as a circuit in which two lines, connected togetherat both ends, are charged and discharged in parallel. The behavior of this circuit may be analyzed as the setting up of two waves in a circular line which travel in opposite directions around the line or it may be regarded as the setting up of a step wave in each line which is reflected at the far end of both lines where they are connected together. In the circuit of Fig. 9 the line inductances are shown at H, 12, 13, I4, 15 and 16, the inductances H and 16 each being half the magnitude of the inductances 12, 13, 14, and 15. The capacitances forming the line appear at BI, 32, 63, 84, and B5 and are all of equal magnitude. The load is shown simply as a resistance 8'! and the switching device is shown as a simple switch 88, it being understood that various suitable forms of switch might be used. Since the load 81, when the switch is closed, will be across both ends of the line (or across both lines if the network is considered as two parallel lines), the impedance of the load 81, for values of the inductances 12-7'5 equal to the inductances H-IS of Fig. 1 and values of the capacitances 81-85 equal to the capacitances l'l-Hl of Fig. 1, should be half the impedance of the load 23 shown in Fig. 1.

The line is charged from a source of high voltage indicated at 96 and the charging is efiected through a choke 9| which corresponds to the choke 6| of Fig. 8. I I

Artificial line networks may be used to produce rectangular pulses in a load upon a switching operation which initiates the charging of the net'- work as well as uponswitching operations which discharge the network in the manner previously described. In the circuits shown in Figs. 1, '7; 8, and 9 rectangular pulseswere formed by closing a switch which permitted the network to discharge through the lead. V explained in connection with Figs, 2, 3, 4, and 5, theswitchingoperation produced a step like voltage wave in the line which traveled down the linerwas reflected from the open circuited end and returned to the begin ning of the line A similarvoltage wave; but an increase of; voltage instead of a decrease of voltage, may be set up in the line by suddenly applying a voltage to the line. ,a I I r Fig. 10 shows a circuit for producing rectangular pulses in a load by suddenly impressin a; voltage upon a; network line.- The voltage is provided by a battery woqrh loadis shown at HM and the pulse forming line comprises the induct ances I02, I03, Mand I05 and the capacitances I06, I01. I08 and I09. 'A two position switch is shown at III]. If the switch is suddenly thrown to its upper position the voltage oithe battery will be applied inseries across the line and the load. If the load IE1 is equal to thei characteristic ill'lpedanc'e of the line (the nominal value bein close enough approximation) the v tage of the battery will divide so that half ofit" appears across the load and half across the line. A voltage wave will then travel down the line and be reflected back andafter. a time interval determined the number of sections of the ne wdrkiaas y the values of the inductance and capacitance elei merits, as explained in connection with Fig. l, the reflected wave will arrive peek at the input and add itself to the voltag' fo'rinerly across the line, sothat'in this case theli'ne will be fully charged: with the full battery voltage across it a d the voltage across the load will drop to zero. During the intervalsre'quired for the voltage wave to travel down th'eline andba'c'k a rectangular pulse of electricity willb'e impresseuubb'ri the load n I. In orderjto repeatthe'operation' it will be nece s s ryf to? discharge the line;- wh h may be done by throwing the switch I IIf'to i lower 'posit'ionnthus lower terminal of switch III) is connected to ground instead'o'f to the line, iiF J; i2 the'line will discharge'throiighthe load, thuspr'o viding a square pulse of a polarityopposite to that of the pulse formedupori' the chargingoftlie" line. This circuit is adaptedfo'r the productionof rectangular pulses of alternatingpolarityQ A Reflections can beset up in a transmissionline by a short-circuited termination as wara pyan open-circuite'd termirl'ejtio'riand a uniform ladder networkmay' be provided which? corresponds to a short-cir'cuited line. A' shortfcir'cuited line is adapted to store energy electromagnetically rather thanelectrostatically; which isto say' that energy isstoredins uchialine by the flow of current" through the' line} or m the case of a-network, by the flow oi current through the inductances; instead by the chargingf the line capacitance. Il a uniformladder network-is pro-- vided with a short -circuited;termination at a midseries connection; it will simulate-a short circuited line just as a uniform ladder network with an s an-arouses midshunt' termination such '25. netwerrs'snowii in Figs]; 1*, :7, 8,-andeafli e her-e s.

10 10 will simulate an open-circuitedline, The uniform ladder network adaptedtosimulate a shorte circuited line may also be constructed by useof the principle of duality,whereby the desired network is constructed from the corresponding open-circuited line ladder network by replacin shunt capacitances by series inductances of corresponding values and by replacing-series inductances by shunt capacitances of correspond-1 ing values. I r A A circuit employing a uniform ladder network which simulates a short-'circuited transmission line is shown in Fig. ii. A source of voltage is indicated generally by battery 12%. The

switching device is indicated simply by a the switch I2 I. The load is a transmitting tube I22 which is so polarized that when the switch I2 I ls closed it willnot be energized by the battery IZ Il; The cholze I23 performs the sanie function as the choke 55; A resistance I24 may be provided, in series with it to limitthe currentthrough the chock I23 when the switch IZI is closed. In addition the resistor I24 may act todanipthe oscillations occurring at the end of pulses produced in the load and this damping is for many purposes sufficient to permit the dispensing with a shunt diode such as the diode 510i Fig. 8. H The pulse" forming line comprisesth e inductances I30, I3I, i32, I33, and l 34 andthe capacitances I35, I31, 138,- and I323. The inductlT-IQ'ofFig. 1. Tire inductance I34 has half the I value of each of theother network inductances and likewise the capacitance; I 35 has, half the value o any of theot hernetWOrk capacitances.

The network is thus a" uniform ladder network; wi e. d h ntpierm nati n attheioadend o the line and a short-circuited midseries'termina- W en helsw ich isi b e i; area since" through the inductances I30 I'3'I I32, I33 and 3 a thi cu rent i bu d; uni c.... s ead ale dg ei linsueq t e chera te sticsoit 1 p wersus lv tin Bis-p1 thebatt y andthe e sl eee flihs g swh shzr r y 'de the nduct: ance. The flow of current through these inductances will set upLamagnetic field about each nde mcaw b 1. 1: w ener y Wi '1i. be stor d; j s s g l r reu sprevieuslv d sc h de er was stored in the electrostatic; field of; the con.

denser Whenthe switch 2I issuddenly opened the current flowing from-the battery through the inductances and back to the battery will be interrupted {Ihemagnetic-fieldzof the ,inductances will begin tdcpllapsegthus tending to maintain the current and a voltage will appear across the load of a polarity oppositeto'thatof the battery.

If the load I22 has an impedance equal .to the image impedance on thelinepwhich is approiib matelythe f nominal characteristic: impedance; current will flow through theload which is half of that which; formerly flowedathroughthein ductances. This current" will flow for thetime required for the: step wave pr'oducedflb'y" the switching=- operation to travel downfito the." shortcireuited end of "the: line and bereflectedba'ckto' the input-end of the line's Thetrprogressnf the current wavegin the line 'will': be" similar tottlie" progress 'ofthe voltage waveinthelineoi Fig: 1

which was represented graphically in Figs. 2, 3, 4, and 5. In the case of Fig. 11, however, the source of current after the opening of the switch I2I is not the'discharge of condensers but the collapsing of the magnetic field of the coil. The condensers cooperate with the coil to produce a rectangular pulse type of discharge just as in the arrangement of Fig. 1 the coils cooperated with the condensers so that the line discharged in a manner adapted to produce a square pulse in the pro-per type of load.

Of the various illustrative circuits herein described for the production of rectangular pulses in response to a switching operation in accordance with the present invention, the arrangement of Figs. 1, 7 and 8 are preferred. In a typical circuit similar to that shown in Fig. 7, which has been used'for producing pulses, the impedance of the load and the nominal characteristic impedance of the network were each equal to 700 ohms. The capacitance of each condenser except the two terminal condensers was 100 micromicrofarads, the terminal condensers having a capacitance of 50 micromicrofarads. The inductances have a value each of 0.49 microhenry. Seven seca tions were provided, each with a time delay of 0.7 times 10- seconds. This gave a total time for the width of the pulse equal to 9.8 times 10 seconds or 0.098 microsecond.

The values of inductance heretofore given refer to the inductance of the components in service position. This value may be affected by coupling between the coils, the presence of ferromagnetic materials near the coils and other similar factors. In compact equipment it is difiicult to predict the effect of stray coupling and it may be convenient after building a first approximation to a line by means of uniform ladder network design, to compensate for coupling effects and other disturbing factors by adopting an experimental approach and perfecting the simultation of the behavior of a line by the network and its associated circuit through adjustment 'of the values of the various reactive components, checking each adjustment by a monitoring operation, such as viewing the pulse shape produced on a rapid-sweep oscilloscope.

Since the uniform ladder networks of the type here involved simulate the properties of a transmission line by approximation only, especially in the case of a network of a small number of sections, and since in many types of apparatus there will be some practically unavoidable deviation from the network design on account of coupling between coils, great precision in the provision of the exact values of inductance and capacitance theoretically indicated in accordance with the foregoing explanations is not generally to be sought in practice. Fortunately, pulse shapes of an adequate rectangularity for practical purposes are obtainable by networks that when constructed only roughly approximate the theoretical artificial line design as herein described.

It is possible to use other types of network of a general class known as artificial lines which are adapted to simulate a transmission line of the desired length and characteristic impedance. Use of an ordinary transmissian line is impractical because of the length that would be required even for short pulses. The uniform ladder network provides a compact and reasonably simple structure for incorporation into a pulseforming apparatus, but other types of artificial line networks may also be found suitable. For instance the transmission line of desired characteristics may be simulated by a network aswell as by a ladder type of network I I t 7 It will be seen from the above explanationand description that I have provided circuits'which will produce a voltage pulse by simply closing a switch and that this pulse can be given any desired time interval or width by the selection of the values of inductances and capacitances and the number of sections used inthe network. For

best results it is desired that the load impedance match the characteristic impedanceof'the network. Mismatching will produce differences in the pulse. For instance if the'load impedance is greater than the characteristic impedance of the network, the resulting pulse may he stepped, while if it is less than thecharacteristic impedance of the network, the'eiiectmightloeto narrow the pulse.

What I claim and desire to secure by Letters Patentis: 7 1

1. A pulse-forming circuit comprising, in combination a reactive network approximately simulating a substantially dissipationless transmission line having a characteristic impedance substantialiy equal to the impedance of theload. with which the network is to be used, said network being adapted to have said load connected across one end thereof, means to charge said line to a predetermined voltage at a relatively slow rate, and switching means to connect said load across the end of said line.

2. A pulse-forming circuit comprising, in, combination, a reactive ladder network approximately simulating a substantially dissipationless transmission line havingv a characteristic impedance substantially equal to the impedance of the load with which the network is to'be used, said network being adapted tohave said load connected across one end thereof. the inductance and capacitance of the elements of said network being such as to permit a wave propagation along said line at a predetermined velocity, means to charge said line to a predetermined voltage at relatively slow rate, and switching means to connect said load across the end of said line.

'3. A pulse-forming apparatus comprising, in combination, a reactive network approximately simulating a transmission line having a characteristic impedancesubstantially equal to the impedance of the loadwith which the networkis to be used and adapted to have said load connected across one end thereof, means to charge the capacitance of said network to" a predetermined voltage at a relatively slow rate, and switching means to connect said load across the end of said network and automatically to disconnect said load from said networkafter the capacitance of said network is discharged.

4. A pulse-forming circuit comprising, in combination, a reactive. network of properties approximately similar to those of 'a substantially dissipationless transmission line having a characteristic impedance substantially equal to the impedance of the load with which the network is to be used and adapted to have said load connected across one end thereof, means to charge said network to a predetermined voltage at a relatively slow rate, a gas tube connected to the load end of said line and adapted to be connected in series with said load, and means to cause said gas tube to become conductive.-

5. A circuit adapted to generate a substantally rectangular pulse in responseto a switching operation which circuit includes a reactive network,

a source of electric eiiergy for feeding electric energy to said reactive network and switching means adapted to produce a sudden change of energy transfer relation between said source and said network and adapted to connect a load in circuit with said network during a predetermined time following the time of said sudden change, said network being such as substantially to simulate a dissipationless transmission line of a characteristic impedance approximately equal to the impedance of said load.

6. A circuit adapted to generate a substantially rectangular pulse in response to a switching operation which circuit includes a two terminal reactive network, a source of electric energy for electrically charging said reactive network and switching means adapted to produce a sudden change of potential across the terminals of said network and adapted to connect a load in circuit with said network at the time of said sudden change and for a period thereafter, said network being such as substantially to simulate a dissipationless transmission line open circuited at the end farthest from the said terminals of said network and having a characteristic impedance approximately equal to the impedance of said load.

7. An electric circuit adapted to generate a substantially rectangular pulse in response to a switching operation which circuit includes a uniform ladder network adapted to approximately simulate a resonant and substantially dissipationless transmission line of a characteristic impedance approximately equal to the impedance of a, load in which said pulses are desired to be produced,- a source of electric energy for feeding energy to said network and switching means adapted to produce a sudden change in the electrical condition of said network and adapted to connect the said load in circuit with said network for a predetermined time including and following said sudden change.

8. An electric circuit adapted to generate a substantially rectangular pulse in response to a switching operation which circuit includes a reactive uniform ladder network composed of series-connected inductances and shunt capacitances, one end of which network has a substantially reactive termination and the other end of which, having two terminals,- is connected in ciredit with a switching means, a source of electric energy for feeding energy to said network and a load in which said pulses are desired to be produced, the nominal value of the characteristic impedance of said network being approximately equal to the impedance of said load, said switching means being adapted to produce a sudden change in the electrical condition of said network and adapted to connect the said load in energy transferring relation with said network for a predetermined time including and following said sudden change.

9. An electric circuit adapted to generate a substantially rectangular pulse in response to a switching operation which circuit includes a reactive uniform ladder network consisting of series inductance's and shunt capacitances having an open circuited midshunt termination at one end and, at the other end, being connected through a midseries termination to a switching means, a source of electric voltage and a load, in a manner adapted to be controlled by said switching means, said switching means being adapted to produce a sudden change in the voltage across the said midseries termination or said network 14 and being further adapted to connect the said load in energy transferring relation with said network for a predetermined time including and following said suddenchange, and said network having a nominal characteristic impedance approximately equal to the impedance of said load.

10. An electric circuit adapted to generate a substantially rectangular pulse in response to a switching operation which circuit includes a reactive ladder network consisting of series inductances and shunt capacitances, having an open circuited midshunt termination at one end and, at the other end, being connected through a midshunt termination to a switching means,a source of electric voltage and a load in a manner, adapted to be controlled by said switching means. said switching means being adapted to produce a sudden change in the voltage across the said midshunt termination of said network and being further adapted to connect the said load in energy transferring relation with said network for a predetermined time including and following said sudden change, and said network having a nominal characteristic impedance approxi-- mately equal to the impedance of said load.

11. An electric circuit adapted to generate a substantially rectangular pulse of a predetermined length 6 in response to a switching operation which circuit includes a reactive uniform ladder network having series inductances and shunt capacitances a n d a number of sections N such that 6=2N /LC', said inductances and capacitances being also such that the nominal value of the characteristic impedance of said network is approximately equal to the impedance of a load in which it is desired to produce the said pulses, said uniform ladder network having at one end a termination, which is substantially purely reactive and at the other end being connected in circuit with a switching means, a source of electric energy and said load in a manner adapted to be controlled. by said switching means, said switching means being adapted to produce a sudden change in the electrical condition of said network and adapted to connect the said load in energy transferring relation with said network at the time of said sudden change and for at least the time period 6 thereafter.

12. A circuit adapted to generate a substantially rectangular pulse in response to a switching operation which circuit includes a reactive twoterminal network, a source of electric otential for electrostatically charging said reactive network through a current limiting series high impedance, switching means adapted to connect a load in circuit with said network and thereby to produce a sudden discharge of said network through said load, said network. being such as substantially to simulate a dissipationless trans mission line open circuited at the end farthest from said terminals of said network and having a characteristic impedance approximately equal to the impedance of said load.

13. A pulse forming circuit comprising, in combination, a network simulating a substantially dissipationless transmission line, a load element having an impedance substantially equal to the characteristic impedance of said network, an energy source connected to said network, and switching means for connecting said load element across said network in energy transferring relation therewith, whereby a substantially rectangular pulse appears in said load element for a time duration dependent upon predetermined characteristics of said network.

'14. Apparatus for producing electrical pulses and applying them to a load, comprising a time delay network which is terminated at one end to make it reflecting and which is terminated at its other end in a substantially reflectionless manner by said load, means for electrically charging said network, and means forcausing said network to be discharged periodically through said load whereby there is applied to said load a pulse having a duration equal to the amount a wave is delayed in traveling the length of said delay network and back. i

15. The methodof utilizing a time delay networkfor producing and applying to a load high voltage impulses of short duration at time intervals long. in relation to said short duration, comprising charging said network and periodically discharging said network into said load substantially without reflection 'at the load end of said network but with reflection at the other end of said network.

16. Apparatus for applying to a load high voltage impulses of short duration at time intervals long in relation to said short duration, comprising at least .one electric discharge tube arranged so as normally to be non-conducting, a time delay network connected in the output circuit of said valve, said time delay network being terminated at its end remote from said tube to make it reflecting, and terminated atits other end in a substantially refiectionless manner by said load, and means for periodically charging and dischargingsaid delay network, said last means comprisingmeans for making said discharge tube periodicallyconducting.

17. Apparatusforproducing electrical pulses and applying them to a load, comprising a time delay network which is terminated at one end to make it reflecting and which is terminated at its other end in a substantially refiectionless manner by said load, means for electrically charging said network, an electric discharge tube connected in series with said network and said load, and means for causing said network to be discharged periodically through said discharge tube and said load wherebythere is applied to said load a pulse having a duration equal to the amount a wave is delayed in travelling the length of said delay network and back.

18. The invention according to claim 17 wherein said discharge tube is a vapor tube having a control grid.

19. The invention according to claim 17 wherein said network is charged at its reflecting end through a choke coil having a high impedance at the repetition rate of said pulses.

20. In a pulse transmission system, a load, and means for supplying said load with pulses of voltage of constant amplitude, said means including an electron discharge device having a cathode, a control electrode and another electrode, said device normally being in a non-conducting condition, an energy storing circuit connected to said other electrode of said device, a circuit for storing a charge on said energy storing circuit, a connection from said cathode to said load, and means for supplying to said control electrode a voltage pulse of sufiicient magnitude and polarity for rendering said' electron discharge device momentarily conductive, whereby the stored charge in said energy storing circuit discharges through said device.

21. In a radio system having a line of predetermined length and associated therewith a space path, the method of producing a pulse which comprisesgradually storing a charge at a predetermined rate on said line of predetermined length, causing a discharge across said space path from one terminal of said line after the stored charge reaches a certain value, initiating at the beginning of said discharge a traveling wave on said line from said one terminal which travels down the length of said line, reflecting said wave back along said lin from the other terminal thereof in such sense and magnitude as to completely and instantaneously extinguish said discharge, by reducing the discharge sustaining voltage to a value less than necessary to maintain ionization of the space path.

22. In a radio system having a line of predetermined length and associated therewith a space path, the method of producing periodically repeated equal time duration pulses which comprises gradually storing a charge at a predetermined rate on said line of predetermined length, causing a discharge across said space path from one terminal of said line after the stored charge reaches a certain value, initiating at the beginning of said discharge a traveling Wave on said line from said one terminal which travels down the length of said line, and reflecting said wave back along said line from the other terminal thereof in such sense and magnitude as to completely and instantaneously extinguish said discharge by reducing the discharge sustaining voltage to a value less than necessary to maintain ionization of the space path, and repeating the foregoing steps.

23. In a radio system having a line of predetermined length and associated therewith a space path, the method of producing a pulse which comprises gradually storing a charge at a predetermined rate on said line of predetermined length, causing a discharge across said space path from one terminal of said line after the stored charge reaches a certain value, utilizing said discharge to produce oscillationsof ultra high frequency for a duration equal to the time of said discharge, initiating at the beginning of 7 said discharge a traveling wave on said line from said one terminal which travels down the length of said line, and reflecting said wave back along said line from the other terminal thereof in such sense and magnitude as to completely and instantaneously extinguish said discharge by reducing the discharge sustaining voltage to a value less than necessary to maintain ionization of the space path.

24. In a radio system having a line of predetermined length and associated therewith a space path, the method of producing periodically repeated equal time duration pulses which comprises gradually storing a charge at a predetermined rate on said line of predetermined length, causing a discharge across said space path from one terminal of said line after the stored charge reaches a certain value, utilizing said discharge to produce an electron current flow for a duration equal to the time of said discharge, initiating at the beginning of said discharge a traveling wave on saidline from said one terminal which travels down the length of said line, and reflecting said wave back along said line from the other terminal thereof in such sense and magnitude as to completely and instantaneously extinguish said discharge by reducing the discharge sustaining voltage to a value less than necessary to maintain ionization of the space path, and repeating the foregoing cycle of operations.

25. In a radio system having a line of predetermined length, the method of producing a pulse which comprises gradually storing a charge at a predetermined rate on said line of predetermined length, causing a discharge from one terminal of said line after the stored charge reaches a certain value, initiating at the beginning of said discharge a traveling wave on said line from said one terminal which travels down the length of said line, utilizing said discharge to produce high frequency oscillations, and terminating the oscillations immediately upon the arrival of said traveling wave at the other terminal of said line.

26. A radio system including means for storing up energy, said system having a transmitter arranged to transmit periodically pulses representative of the stored energy for time periods short compared to the time intervals between transmitted pulses, said means including an energy storing circuit in the form of a length of transmission of delay line having such parameters 18 that the time it takes a wave to travel the efl'ective length of said line is equal to one-half the duration of the pulses generated.

JABEZ C. STREET.

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

UNITED STATES PATENTS Number Name Date 2,188,970 Wilson Feb. 6, 1940 2,255,839 Wilson Sept. 16, 1941 2,265,996 Blumlein Dec. 16, 1941 2,390,659 Morrison Dec. 11, 1945 2,394,389 Lord Feb. 5, 1946 2,405,069 Tonks July 30, 1946 2,408,824 Varela Oct. 8, 1946 2,411,140 Lindenblad Nov. 12, 1946

Images