Untitled record

July 17, 1962 j. v. mcnulty etal 3,045,148

IGNITION SYSTEM WITH TRANSISTOR CONTROL

Filed Dec. 18, 1959

Sheets-Sheet 1

<img file="US3045148A_D0001.tif" />

John VMcNulty Davtd J. Wright By

<img file="US3045148A_D0002.tif" />

Attorney

July 17, 1962 j. v. mcnulty etal 3,045,148

IGNITION SYSTEM WITH TRANSISTOR CONTROL

Filed Dec. 18, 1959

Sheets-Sheet 2

<img file="US3045148A_D0003.tif" />

<img file="US3045148A_D0004.tif" />

<img file="US3045148A_D0005.tif" />

<img file="US3045148A_D0006.tif" />

Attorney

United States Patent Office

3,045,148

Patented July 17, 1962

3,045,148

IGNITION SYSTEM WITH TRANSISTOR CONTROL

John V. McNulty and David J. Wright, Norwich, N.Y., assignors to General Laboratory Associates, Inc., Norwich, N.Y., a corporation of New York

Filed Dec. 18, 1959, Ser. No. 860,486

Claims. (Cl. 315—183)

This invention relates to ignition apparatus of the capacitor discharge type, such as commonly used on jet and rocket engines.

Ignition apparatus for jet and rocket engines and the like is required to produce repeatedly spark discharges characterized by equal , quantities of energy, in order that the conditions for proper ignition at the igniter plug may be consistent from one ignition to the· next. Most such engines are used on aircraft and are supplied with electrical energy from a battery or other source of limited capacity. The terminal voltage available from such a source varies with the age of the battery and with the magnitude of the load represented by other electrical devices which may be energized from the battery concurrently with the ignition apparatus. Other conditions may also affect the voltage available at the source. For example, in one type of installation, the ignition system is required to maintain a substantially constant energy of the spark discharge while the potential of the source varies over a range from 14 to 30 volts.

It is an object of the present invention to provide ignition apparatus producing sparks of substantially constant energy at an igniter over a considerable range of variation in the potential of the source of electrical energy.

Another object of the invention is to provide an improved ignition apparatus of the capacitor discharge type.

A further object of the invention is to provide an improved arrangement for charging a capacitor to predetermined potential.

The foregoing and other objects of the invention are attained in the apparatus described herein, which includes a voltage doubler consisting of an inductance element and a capacitance element connected in series across a source of electrical, energy; means for discharging the capacitor of the voltage doubler in pulses comprising a transformer primary winding and a thyratron semiconductor device connected in series, and means for controlling the triggering potential of the thyratron semiconductor device in response to a control potential varying concurrently with the potential across the capacitance element of the voltage doubler. The secondary winding of the transformer is connected in series with the main ignition capacitor and supplies a pulse charge to that capacitor with each triggering of the thyratron device. The energy stored on the main capacitor is delivered to the igniter whenever the potential across the main capacitor exceeds a predetermined value.

Other objects and advantages of the invention will become apparent from a consideration of the following specification and claims, taken together with the accompanying drawings.

In the drawings:

FIG. 1 is a wiring diagram of one form of ignition apparatus embodying our invention;

FIG. 2 is a graphical illustration of the variation in potential across the voltage doubler capacitor of FIG. 1;

FIG. 3 illustrates a modification of the wiring diagram of FIG. 1;

FIG. 4 is a graphical illustration showing the effect of the modification of FIG. 3 upon the potential variation across the voltage doubler capacitor;

FIGS. 5 to 9 are fragmentary wiring diagrams illustrating further modifications of die circuit of FIG. 1;

FIG. 10 is a complete wiring diagram illustrating a further embodiment of the invention;

FIG. 10A is a wiring diagram illustrating another embodiment of the invention; and

FIG. 11 is a fragmentary wiring diagram illustrating a modification of the circuit of FIG. 10.

FIGURE 1

This figure illustrates an ignition system which is sup10 plied with electrical energy from a battery 1. A voltage doubler comprising an inductance element 2 and a capacitance element 3 in series is connected across the terminals of . the battery 1. Across the terminals of the capacitor element 3 is connected a capacitor discharging circuit ni15 eluding a transformer primary winding 4 and an anodecathode path of a thyratron semiconductor device, shown •as a controlled rectifier or thyratron transistor 5. The controlled rectifier '5 has an anode 5a, a gate electrode or control electrode 5b and a cathode 5c. A control po20 tential deriving network including resistors 6 and 7 inseries is connected across the terminals of capacitance element 3. The common junction of the resistors 6 and 7 is connected through a wire 8 to the control electrode 5b.

The primary winding 4 is part of a . transformer 9 25 having a secondary winding 10. A voltage doubler capacitor 11 is connected in series with a diode 12 across the terminals of the secondary winding 10. The diode 12 has its cathode connected to a junction 13, which is common to the capacitor 11. Another diode 14 has its anode 30 connected to the junction 13. A storage capacitor 15 is connected between the cathode of diode 14 and ground. A sealed gap 16 is connected in series with a resistor 17 across the storage capacitor 15. An igniter gap 18 is connected across the resistor 17.

OPERATION OF FIG. 1

Considering the condition in the circuit beginning at •an instant when the capacitor 3 is completely discharged, current flowing through inductance 2 and capacitor 3 will ⁴ cause a potential to build up on the capacitor 3 to substantially twice the potential of the source 1. This phenomenon is well known and is commonly described as a “voltage doubler” action. The potential across capacitor 3 is applied across the terminals of the primary winding 4 and the anode-cathode path of the thyratron device 5 in series.

The thyratron device 5 is shown herein as a semi-conductor device of the type commonly referred to as a con₅θ trolled rectifier or as a thyratron transistor. Such a device has a characteristic that its forward impedance, i.e., the impedance to current flowing from anode-to-cathode, is very high until one of two conditions occur. One of the two conditions is the occurrence of an anode-to-cath_g5 ode current greater than a predetermined breakdown value. The other condition is the occurrence of a current greater than breakdown value between the control electrode and the cathode. Upon the occurrence of either of these two breakdown conditions, the forward impedance of the thyratron device 5 drops to a very low value and ° stays at that low value until the anode-to-cathode current falls to a second value substantially lower than the sustaining value. In the circuit of FIG. 1, device 5 is tripped from its high impedance condition to its low impedance condition when the current between control electrode ⁶⁵ 5b and cathode 5c exceeds a predtermined value. The forward impedance of the device 5 remains at its low value until the anode-to-cathode current is reduced to zero.

•Resistors 6 and 7 form a voltage divider to derive from 70 the potential across the capacitor 3 a proportion of that potential which is applied through wire 8 to control electrode 5b as the control potential. The resistors 6 and 7

3,045,148 are chosen so that the thyratron device 5 is triggered to its low impedance value when the potential across capacitance 3 approaches a value 2E, as shown in FIG. 2, E being the potential of the battery 1.

When the thyratron devie 5 is triggered, a pulse of current flows from capacitance 3 through the primary winding 4, a corresponding pulse being thereby induced in secondary winding 10. The latter pulse charges the capacitor 11. The polarity of the windings 4 and 10 is indicated by the dots in the drawing. The pulse produced in the winding 10 has its positive polarity at the upper terminal of that winding. This pulse flows through capacitor 11, diode 14 and capacitor 15 and tends to charge the capacitors 11 and 15. The pulse does not flow through diode 12, since diode 12 has its high impedance opposed to the pulse.

On the half-waves when the upper terminal of the secondary winding 10 is negative, current flows from that winding through the diode 12 and charges the capacitor 11 with its right-hand terminal positive. The winding 10 and capacitor 11 together act as a voltage doubler, to develop across the capacitor 11 a transient inverse potential equal to twice the applied potential. On the opposite half-waves, current flow through the diode 12 is blocked but current can flow from winding 10 and capacitor 11, which now act as potential sources in series aiding, through diode 14 and capacitor 15, thereby charging capaciotr 15. When the charge on capacitor 15 exceeds a value determined by the breakdown potential of gap 16, that gap breaks down. Thereupon, substantially the full potential of capacitor 15 appears across the resistor 17 and gap 18 in parallel. The gap 18 in turn breaks down, whereupon the capacitor 15 discharges through it.

The discharging of the capacitor 5 is repeated each time that the charge on it builds up sufficiently to break down the gap 16.

The circuit goes through a series of pulse producing cycles as described above. Fig. 2 shows at 20 the variation of potential across capacitor 3 during such cycles. During the capacitor discharging phase of each cycle, the inductance in the circuit tends to maintain the flow of current even after the capacitor is discharged, with the result that the capacitor begins to charge in the reverse direction, resulting in the negative excursion of potential appearing at 20b in FIG. 2. Eventually this reverse potential blocks the current flow through the thyratron device 5, which thereupon returns to its high impedance condition and the cycle begins again. During each of these cycles, a pulse of current charges the capacitors 11 and 15. Capacitor 11 cooperates with winding 10 to form a voltage doubler, so that capacitor 15 is charged at a higher potential. As the series of pulse cycles continues, the potential across the capacitor 15 builds up in a seires of steps. The potential across storage capacitor 15 is applied across the sealed gap 16. There being then no current flow through resistor 17, susbtantially the full potential across capacitor 15 is applied across the sealed gap 16. The breakdown potential of the gap 16 is made somewhat higher than the breakdown potential of the igniter gap 18. When the potential of the charge on capacitor 15 exceeds the breakdown potential of the gap 16, the charge on capacitor 15 is discharged through the trigger gap 16 and the igniter gap 18. The gap 18 has a substantially lower impedance than the resistor 17 and takes substantially all the current flowing from the capacitor 15. The breaking down of the gaps 16 and 18 produces a low impedance path to the charge stored on the capacitor 11, and that charge now also flows through the gaps 16 and 18.

It may be seen that the circuit of FIG. 1 provides a series of spark discharges at the igniter 18. Each of those spark discharges is built up by a series of pulses of substantially equal energy applied to the capacitors 11 and 15, so that each spark discharge at the igniter gap 18 has substantially the same energy. Although the potential of the source 1 may vary, the characteristics of the thyratron device 5 do not vary. The breakdown control potential on electrode 5b, at which the thyratron device 5 shifts from its high impedance to its low impedance condition, is always the same, being determined by the characteristics of the device 5 and not by the charge stored on capacitor 3. The magnitude of the current pulses which charge the capacitors 11 and 15 is thereby made independent of the potential of the source. The trigger gap 16 consequently always breaks down after the same number of charging pulses and the charge built up on the capacitors 11 and 15 and discharged through the igniter gap 18 alawys has susbtantially the same energy.

FIGURES 3 AND 4

FIG. 3 illustrates a modification of the circuit of FIG. 1. In accordance with this modification a diode 19 has its anode connected to the cathode of the thyratron device 5, while its cathode is connected to the anode of the thyratron device 5. The presence of the diode 19 limits the reverse polarity potential across the controlled rectifier 5 to the forward impedance drop across diode 19. The operation as modified by diode 19 is illustrated by the curve 20A in FIG. 4. The diode 19 makes the potential from which the capacitor 3 starts to charge more positive. The current for recharging the capacitor 3 from the reverse potential value shown at 205 to a positive value must be supplied by battery 1. When the diode 19 is added as in FIG. 3, this reverse potential is smaller, so that the recharging current is smaller, and hence the circuit losses are lower. The efficiency of the circuit is thereby improved.

The presence of diode 19 also prevents any tendency to build up a potential on the capacitor 3 gradually over several cycles, because of incomplete discharge of the capacitor on each cycle.

FIGURE 5

This circuit is modified from that of FIG. 1 by the substitution of a saturable core transformer 21 in the control potential deriving network, in place of the resistors 6 and 7 of FIG. 1. Transformer 21 has a primary winding 22 and a secondary winding 23. Primary winding 22 is connected between the ungrounded terminal 24 of capacitor 3 and the control electrode 5b. Secondary winding 23 is connected between the anode of the diode 19 and ground.

OPERATION OF FIG. 5

During charging of the capacitor 3 the secondary winding 23 is substantially open circuited, due to the diode 19, which has its high impedance in series with winding 23. The current flowing through winding 22 and control electrode 5b increases as the charge on capacitor 3 increases, finally saturating the core of transformer 21, whereupon a pulse of current flows through control electrode 5b, setting off a discharge of capacitor 3 through the thyratron device 5. When the thyratron device 5 breaks down to its low impedance value, the capacitor 3 is discharged and the potential across it reverses as the magnetic field of the inductance 2 collapses. This reverse potential sends a high current through the secondary winding 23 and diode 19, resetting the core of transformer 21 by saturating it in the opposite sense, thereby restoring primary winding 22 to its high impedance condition.

FIGURE 6

In this modification of the invention, the potential of the control electrode 5b is derived by the use of a semiconductor device 25 of the type known as a double-base diode. The double-base diode 25 has two base electrodes 25α and 25b and a control electrode 25c. Base electrode 25α is connected to junction 24. Base electrode 25b is connected to control electrode 5b. Control electrode 25c is connected to the common terminal of a time constant

3,046,148 •1 ( + network including a resistor 26 and a capacitor 27. A switch 28 connects the time constant network in series between the junction 24 and ground. The switch 28 is movable from the position shown to a second position in which the time constant network is connected across g a battery 29.

OPERATION OF FIG. 6

The timing between the start of a cycle at zero potential across capacitor 3 and the breaking down of the thyratron device 5 is controlled in FIG. 6 by the time ¹⁰characteristics of the network 26, 27 rather than by the time characteristics of the capacitor 3 itself. When the switch 28 is in its right hand position, as shown, the time characteristics of the network 26, 27 are superimposed on the time characteristics of the voltage doubler, i.e., ¹⁵the potential applied to the network 26, 27 is supplied from the capacitor 3. When the switch 28 is thrown to its left-hand position, the triggering time of the thyratron device 5 is determined only by the time characteristics of the network 26, 27 and the potential of the bat- ²⁰tery 29. If this battery is provided to supply this network only, then its characteristics may be much more closely controlled and remain much more stable than the characteristics of the main battery 1 which supplies energy for the trigger gap and perhaps for other load devices. ²⁵

The characteristics of the double base diode 25 are such that its impedance is high until such time as a predetermined potential is applied to control electrode 25c, at which time the impedance between the control electrode 25c and base 25b drops to a very low value. ³θ Capacitor 27 then discharges through this low impedance, producing an output pulse.

FIGURE 7

This figure illustrates a modification of the circuit of FIG. 6 and shows a different arrangement for supplying potential to the terminals of the time constant network 26, 27. In FIG. 7, the battery 1 is used to .supply current through a resistor 31 and a reverse biased diode 4,0 32. The diode 32 is of the Zener voltage type and has a high impedance in its reverse direction until a predetermined potential is exceeded. The impedance then becomes very low and effectively fixes the potential at that predetermined value. The potential across the diode 32 and hence across the time constant network is thereby fixed at a very definite value.

The ideal operation in circuits of the type disclosed herein is to maintain the power delivered to capacitor 15 substantially constant, regardless of changes in the poten- ₅₀tial of the battery 1. An increase in the potential of battery 1 tends to increase the energy output per pulse of the capacitor 3. A compensating effect may be provided by arranging ithe triggering system which times the pulses from the capacitor 3 so that an increase in the battery 55 potential tends to decrease the pulse rate. If the decrease in the pulse rate exactly balances the increase in the energy per pulse, then a substantially constant power output may be produced, even though the battery potential changes. 60

In the circuit of FIG. 7, the battery potential is applied across the base electrodes 25α and 25b of the double base diode 25. An increase in this potential requires a higher potential on the control electrode 25c to trip the double base diode to its low impedance value. The 65 potential of control electrode 25c is determined by the time constant network 26, 27, which is now supplied by the constant potential across diode 32. Capacitor 27 therefore charges at the same rate, regardless of changes in the potential of battery 1. However, since an increase 70 in the battery potential requires a higher potential at electrode 25c to trip the double base diode, a longer charging time for capacitor 27 is required before the diode 25 trips.

The pulse rate is thereby reduced to compensate for 75 the increased energy per pulse due to the increased battery potential.

FIGURES

This figure shows a modification of the circuit of FIG. 7, in which a resistor 33 is added between the Zener diode 32 and the terminals of the time constant network. This circuit, by virtue of the potential drop due to current flowing in resistor 33, varies the potential applied to the time constant network as a function of the battery potential. In other words, an increase in the battery potential increases the potential applied to the time constant network and consequently makes the capacitor 27 charge to.its tripping value in a shorter time.

It has been found that the circuit of FIG. 7 tends to overcompensate for an increase in battery potential. In other words, as the battery potential increases, it slows down the pulse rate so much that the power output is actually decreased. In order to correct that unbalance, the resistor 33 has been added in the circuit of FIG. 8. Because of the resistor 33, the potential across the time constant network 26, 27 is not fixed, but increases with an increase in the potential of battery 1, due to the potential drop across resistor 33. By properly balancing the potential drop across resistor 33 with respect to the potential across battery 1, the circuit of FIG. 8 may be made to compensate for the changes in the battery potential, so as to maintain a close control of the power output, at least over limited ranges of variation in the potential of battery 1.

As a further alternative, the voltage dividers of FIGS. 7 and 8 may be connected across capacitor 3 instead of across battery 1.

FIGURE 9

This circuit illustrates a modification of the circuit of FIG. 1, in which the voltage divider network 6, 7 is replaced by a Zener voltage diode 34, connected between the control electrode 5c and junction 24. The operation is the same as the circuit of FIG. 1, except that the application of the breakdown control current to electrode 25c is determined by the characteristics of the Zener diode 34.

FIGURE 10

This circuit is considerably modified from the previous circuits, particularly in the potential supply for the time constant network Which controls the discharge times, the connections of the diode 19, and in the circuitry connected to the secondary winding 10.

The potential supply for the time constant network 26, 27 is obtained through a filter including an inductance element 35 and a capacitance element 36 connected in parallel with the voltage doubler 2, 3 but otherwise independent of it. The potential appearing across capacitor 36 is applied to a voltage divider network including resistor 37 and resistor 38 and a Zener voltage diode 39 similar to the network shown in FIG. 8 above. The. voltage divider network 37, 38, 39 is utilized to control the variation of output power with changes in the battery voltage. The variation may be controlled over a wide range, using the principles discussed above in connection with FIG. 8. The potential across the resistor 38 and diode 39 in series is applied, across the time constant network 26, 27. The junction between inductance 35 and capacitor 36 is also connected through a resistor 40 to the base 25α of the double base diode 25. By virtue of these arrangements, the time intervals at which the thyratron device 5 is triggered are determined by the time constant network and are not affected by transient conditions existing in the main spark energy supply circuit. The elements used in the control voltage deriving network, may, therefore, be more precisely selected and controlled as to their impedance values, since they do not have to carry the heavier currents required in the main energy supply circuit.

Diode 19, instead of being connected directly in parallel with the anode cathode path of the thyratron device 5,

3,045,148 is connected in series with a high Q, low loss inductance 41. The branch circuit including diode 19 and inductance 41 is connected in parallel with the branch circuit including the anode-cathode path of thyratron device 5 and the primary winding 4. By virtue of this arrangement, a very low impedance path is provided to discharge the reverse potential on capacitor 3, so that the losses in the circuit are decreased. The utilization of the energy from the capacitor 3 is thereby made considerably more efficient.

The circuit through the secondary winding 1ft includes a diode 42, a capacitor 43 and a resistor 44 in series. Diode 42 acts as a half-wave rectifier, to determine the polarity of the charge on capacitor 43. An igniter gap 45 of the semiconductor type is connected across the resistor 44. A trigger gap 46 is connected between ground and the common junction of diode 42 and capacitor 43.

The energy to be discharged at the gap 45 is stored on the capacitor 43. When the potential on capacitor 43 exceeds the breakdown potential of the sealed gap 46, the energy on the capacitor 43 is discharged through a circuit consisting only of that capacitor and the gaps 45 and 46.

FIGURE 10A

The circuit shown in this figure is based on that in FIG. 10, but has been improved by the addition of several elements. This is the presently preferred embodiment of the invention.

The elements added in this figure include a diode 70, a transformer 71 having a primary winding 72 and a secondary winding 73, a diode 74, a capacitor 75, a capacitor 76, a diode 77, and a resistor 78.

Diode 70 is effective to hold the charge on capacitor 3 if it reaches full charge before the triggering pulse is applied to the thyratron device 5.

Transformer 71 is a step-up transformer, and is effective, at lower voltages to improve the tripping or starting characteristics. That is to say, it permits the circuit to trigger the transistor at a lower voltage of battery 1.

Diode 74 prevents reverse potential due to overshoot in transformer 71 from reaching the control electrode 5b of thyratron device 5.

Capacitor 75 provides a minimum capacitive load for thyratron device 5, and thereby prevents certain undesirable conditions which might otherwise occur in the case of an open circuit or high resistance load on the secondary winding 1ft. Such an open circuit condition would increase the impedance of primary winding 4 to the capacitor discharge current to a very high value, which would tend to delay the capacitor discharge and spread out the charging pulse. The capacitor 65 establishes a maximum impedance limit on the primary winding 4, and ensures that capacitor 3 will discharge on each pulse.

Capacitor 76 and diode 77 cooperate with secondary winding 10 to form a voltage doubler, functioning in a manner generally similar to the capacitor 11 and diode 12 of FIG. 1.

Resistor 78 is provided to protect diodes 77 and 42 from overcurrents in the forward direction which might occur during oscillatory discharges through the gap 45.

FIGURE 11

The circuit illustrated in FIG. 11 shows a different form of mechanism for controlling the potential supplied to the control electrode 5b of the controlled rectifier 5. This circuit is in other respects similar to that shown and described in FIG. 10. Those elements which correspond, both as to structure and function, to their counterparts in FIG. 10, have been given the same reference numerals.

The circuitry for supplying a potential to control electrode 5b includes a first voltage divider connected across the terminals of the battery 1 and traceable through a resistor 46α, a resistor 47, a capacitor 48, and a transformer winding 49 to ground at 50. A capacitor 51 is connected in parallel with resistor 46α. A second voltage divider is also connected across the battery 1 and includes resistors 52 and 53 in series. A capacitor 54 is connected across the resistor 53.

A transistor 55 has an emitter electrode 55e, a base electrode 55b, and a collector electrode 55c. In the present circuit, the base serves as the input electrode, the emitter as the output electrode, and the collector as the common electrode. Collector 55c is connected to the junction 56 between resistors 46 and 47. Base electrode 55b is connected to the opposite terminal of resistor 47. Emitter electrode 55c is connected through a secondary winding 57 of a transformer 58 to the common junction 59 of resistors 52 and 53. Transformer 58 has an output winding 6ft having one terminal connected to ground and the other connected through a wire 61 to the control electrode 5b. A Zener diode 62 is connected between junction 56 and ground.

FIGURE 11—OPERATION

It is desired to keep the power output to the igniter 45 at a substantially constant value, regardless of changes in the potential of battery 1. An increase in the potential of battery 1 tends to increase the energy output per pulse of the capacitor 3. In order to maintain the total power to the igniter 45 substantially constant, the triggering system which times the pulses from the capacitor 3 is made to respond to an increase in the battery potential in such a manner as to decrease the pulse rate. The increase in energy per pulse is compensated by the decrease in pulse rate, resulting in a substantially constant power output.

The circuit, including transistor 55, transformer 58, and related elements may be described as a blocking oscillator. It operates to apply periodically to the control electrode 5b potentials which are effective to trigger the controlled rectifier 5 to its low impedance value, thereby producing an output pulse through the transformer 9.

The transistor 55 is shown as an NPN type, so that it is held off by an emitter potential more positive than the base potential, and is turned on by an emitter potential more negative than the base potential.

The Zener diode 62 fixes the potential at junction 56 and collector 55c with respect to ground. The impedances of resistors 52 and 53 are selected so that when the transistor is not conducting the junction 59 is negative with respect to junction 56.

When power is first applied to the circuit, emitter 55e is substantially at the potential of junction 59, since there is then no current flow through or potential induced in the winding 57. Base 55b is connected through the uncharged capacitor 48 to ground, there being substantially no potential across winding 49. Base 55b is therefore more negative than emitter 55e, and the transistor is off. The capacitor 48 immediately starts to charge through resistors 46 and 47, and its terminal nearest the resistor 47 swings in a positive sense, eventually becoming more positive than the potential of emitter 55e, whereupon the transistor starts to conduct.

As the current flow through emitter 55e increases, it passes through primary winding 57, inducing a potential in secondary winding 49 of a polarity tending to charge capacitor 48 reversely, i.e., with its lower terminal positive. This charging current flows through base 55b and tends to drive the transistor to conduct more strongly. Finally, the charge on capacitor 48 reaches a condition of balance with the potential across secondary winding 49, and the charging current stops. The potential stored on capacitor 48 is then effective to swing base 55b in a negative sense, thereby cutting off the transistor 55. Current then stops flowing through winding 57. The charge on capacitor 48 then holds the transistor off until that charge is dissipated by current supplied through resistors 46 and 47. The cycle then repeats. The pulse of current in winding 57 induces a potential in winding 60, where it is effective to control the thyratron device 5.

If the potential of battery 1 increases, the potential of

3,045,148 junction 59 swings more positive, while the potential of junction 56 remains fixed at a value more positive than junction 59. The emitter potential, when the transistor is off, is substantially the same as that of junction 56. Hence, in order to turn the transistor on, the capcitor 48 must charge to a more positive potential to make the base 55b more positive than emitter 55e. This charging of capacitor 48 to a higher potential takes a longer time, with a consequent decrease in the rate of supply of tripping pulses to the thyratron device 5. The increase in pulse energy is compensated by the decrease in the pulse rate, so that the power output remains substantially constant.

While I have described the operation of the invention, with respect to its constant power output characteristics, in terms of an increase in battery potential, it should be apparent that a decrease in battery potential produces an analogous but reverse operation, with a compensating increase in the pulse frequency, 'and a similar ultimate result, i.e., constant power output spark energy at the igniter.

Resistor 46α is provided to limit the current flow through the diode 62.

Capacitor 51 provides a low impedance path to alternating current, so that resistor 46α does not limit the operation of the blocking oscillator. Capacitor 54 provides a similar alternating current by-pass around resistor 53, so that during pulsing of the blocking oscillator, resistors 46 and 53 are by-passed, and the full battery potential is effective between the emitter and collector of the transistor.

The following table shows a suggested set of values which will work in the circuit of FIG. 11. Obviously, the invention is not limited to any of these values.

Table

Resistor 46α_______________________ohms 500

Resistor 47_________________________do____4000

Capacitor 48__________________________mfd 0.1

Resistor 52________________________ohms__ 10,000

Resistor 53----,______________________do. 1,000

It should be understood that the circuits shown and described maintain the power output to capacitor 15 substantially constant only over a limited range of variation of the potential of battery 1. Given a particular range of source potential, however, it is easy to design a circuit following the invention which will hold the power output constant.

While we have shown and described certain preferred embodiments of our invention, other modifications thereof will readily occur to those skilled in the art, and we therefore intend our invention to be limited only by the appended claims.