HYBRID IGNITION WITH STRESS-BALANCED COILS
BACKGROUND OF THE INVENTION AND PRIOR ART
The present invention relates to ignition systems for internal combustion (IC) engines, and particularly to high power, high energy ignition simplified by the use of hybrid ignition features with ideal magnetic stress-balanced coils. High energy ignition is essential to the operation of IC engines using difficult-to- ignite mixtures, such as lean mixtures, high exhaust residual or high EGR mixtures, and the more difficult-to-ignite alcohol fuel mixtures. Such mixtures require hundreds of watts of igniting power and fifty millijoules or more of energy versus the ten to thirty watts and millijoules supplied by conventional ignitions. The simplified high power high energy hybrid ignition with stress- balanced coii disclosed herein can deliver the required power and energy with a minimization in the size and cost of parts to make the system practical.
The ignition disclosed is usable in the simpler distributor form or in a distributorless ignition form preferably achieved by the use of a separate leakage inductor disclosed in U.S. patents 5,315,982 and 5,131,376. The high power, high energy, stress-balanced minimum coil size features disclosed are based on Maxwell's equations used in conjunction with the voltage doubling principle disclosed in U.S. Patent No. 4,677,960 and its improvements which were first laid out in U.S. patent 5,315,982. U.S. patents 4,688,538, 4,774,914, 4,841,925, 4,868,730, and 5,207,208 may also be relevant to other features of the invention.
All said above cited patents are of common assignment with this appli¬ cation and all include Dr. M.A.V. Ward as a sole or joint inventor. Reference to the above cited patents is sometimes made by simply listing the last three numerals of the number, as in patents '982, '376, '960, '538, '914, '925, 730, and '208. All are incorporated herein by reference as though set out at length herein.
SUMMARY OF THE INVENTION
A principal object of the present invention is to achieve the high power and high energy ignition needs cited above, i.e. peak power of 100 watts and greater for a sufficient time duration to deliver many tens to a hundred or more millijoules (mj) of total spark energy to the air-fuel mixture to insure the ignition of difficult to ignite mixtures.
A further object of the invention is to use of principles and features of my prior patents cited above with the simplifying new features of the hybrid ignition in the forms disclosed herein accompanied with the concept of stress balance and with the disclosure of actual stress-balanced coils to provide a more simplified, compact, and lower cost effective ignition system able to deliver the required high power and high energy to the mixture.
Another object of the invention is to provide suitable switches for the hybrid ignition circuit and to insure reliable turn-off of the switches which are preferably SCRs which, in this application, do not have a negative bias imposed during turn-off as a result of the unidirectional decaying inductive current. Another object of the invention is to optimize and balance the ignition parts size and cost with the spark discharge size and the spark plug erosion. A consequence of the present hybrid ignition (combined capacitive and inductive system) is the production of a hybrid arc/glow discharge wherein the initial spark of one quarter period is of high frequency (50 to 200 kiloHertz) and high current (2 to 10 amps) followed by a long duration, 0.5 to 5 msec, lower frequency linearly decaying inductive spark of lower spark current which can also be in the ampere range or in the hundred milliampere range of the glow discharge which can provide good quality ignition with a large spark gap of approximately 0.1" or greater while reducing spark plug erosion and spark plug insulator fouling and enhancing combustion reactions through its high spark burning voltage.
The present invention meets the above objects with a system that features a capacitive type ignition system using novel magnetic stress-balanced coils with a high leakage inductance used in conjunction with novel hybrid capacitive/- inductive discharge ignition system for IC engines of the voltage doubling, arc discharge, high power/high efficiency type. The stress balance feature of the coils, i.e. approximately equal maximum coil core magnetic flux density during the peak voltage open circuit and peak current short circuit conditions, is achieved by using closely located side-by-side windings on an E-core (versus concentric windings) in conjunction with other coil and circuit features to achieve the stress-balance. The preferred hybrid ignition feature, characterized by a capacitive first quarter period sinusoidal spark discharge in the ampere range peak spark current followed by a decaying unidirectional inductive spark discharge current of period of order of magnitude of one millisecond, is brought about by including high efficiency high current diodes across the discharge capacitors. Such operation involves the elimination or preferably the relocation of the shunt diodes and/or shunt switches across the main discharge switch Si (i = 1, 2, ..) controlling the ignition coil firing, resulting in a system with essentially fixed spark duration and decaying Kettering type inductive spark for most of the spark duration, but with much higher levels of spaark power, resulting in far greater igniting capability compared to state-of-the-art Kettering ignitions.
Such ignition, designated as "hybrid ignition", i.e. hybrid capacitive and inductive ignition, typically includes a DC to DC converter and controller to charge up one or more discharge capacitors. Preferably the ignition circuit also includes resonating leakage inductor means and stress-balanced coils. Preferably two discharge capacitors and resonating inductors are used comprising a higher and lower discharge frequency circuit separated by an isolation diode to allow for minimum sizing of the compact (preferred stressed-balanced) coils whose high initial open circuit frequency is determined by the higher frequency circuit.
The higher frequency discharge circuit controlling the initial spark discharge may be viewed as an auxiliary discharge circuit to the main lower frequency discharge circuit of the "dual discharge circuit". The spark discharge time of about one millisecond, which is easily varied over a wide range with design, permits simplified spark firing control for the ignition system.
The ignition discharge circuit components are designed according to optimization criteria first disclosed in patents '960 and '982. The basis for the optimization criteria is the solution of coupled differential equations for the circuit voltages which led to the transient voltage doubling formulation first disclosed in patent '960, and the solution of one of Maxwell's equations for the open circuit magnetic flux density in the ignition coil core materials, first disclosed in patent '982. The voltage doubling solution is used as the open circuit high voltage source for generating the peak open circuit magnetic flux density. Taken with the peak short circuit spark firing magnetic flux density, and through experimentation and discovery, the concept of stress balance is disclosed herein from which the design of stress-balanced coils is disclosed herein.
Preferably a flyback type power converter and novel simple controller is employed to provide "soft stall" of the power converter in about 1/4 millisecond following end of the spark discharge current (to insure full recovery of switches Si). In addition, the flyback preferably employs a simple sensor circuit based on sampling the converter discharge current to provide a DC current level for higher power operation. In high speed, i.e. high RPM, IC engine applications the ignition firing (gate) control period Tg is preferably reduced at the higher RPMs through simple circuitry to reflect the reduced spark discharge time constant Tc that occurs due to higher spark dissipation at higher RPM, to thus provide more" time for the power converter to charge up the discharge capacitors.
The energy is preferably delivered by a toroidal gap plug, as disclosed in the prior patents cited, with spark plug tips preferably made of low erosion material such as tungsten-nickel-iron, platinum, etc. The plug tip is well heat- sunk and designed to minimize fouling by keeping the spark discharge away from the plug insulator by recessing the insulator. The recessed insulator may also provide an relatively large spark plug interior combustion volume for further reducing fouling and also for enhancing combustion reactions. Such enhancement can be implemented through electric field enhancement from the high spark burning voltage associated with the glow spark discharge, or through coating of the metallic surfaces of the plug interior combustion volume and the outer plug ends in contact with the engine combustion chamber with catalyst material such as palladium oxide, or by using both electric field and catalyst enhancement.
Other features and objects of the invention will be apparent from the f llowing detailed description of preferred embodiments taken in conjunction with the accompanying drawings, in which: BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a circuit diagram, partially block diagram and partially detailed circuitry, of a generic form of the hybrid ignition for multi-cylinder engines with a single discharge circuit and with preferred high leakage coils shown as applied to a distributorless ignition system.
FIG. la is a graph of the primary circuit discharge current flowing as a function of time corresponding to the single discharge circuit of FIG. 1.
FIG. 2 is a circuit drawing of the discharge circuit portion of a preferred distributor version of the ignition of FIG. 1 with the entire leakage inductance integrated into the coil of a preferred stress-balanced coil. FIGS. 2a and 2b are approximately to-scale side and top views of the preferred suess-balanced coil.
FIG. 3 is a circuit drawing of an embodiment of a hybrid, capacitive- inductive ignition with integrated DC to DC converter power supply.
FIG. 4 is a preferred embodiment of the circuit of FIG. 1 which includes two discharge circuits, the lower frequency circuit and the supplementary high frequency circuit for minimizing the size of the coils, and preferred main switches Si and shunt switches SDi, all as mentioned above and showing their preferred location.
FIG. 4a is a graph of the primary circuit discharge current flowing as a function of time corresponding to the dual discharge circuit of FIG. 4.
FIG. 5 is an essentially complete circuit drawing of a preferred embodiment of a distributorless version of the hybrid ignition (using FIG. 4 features) including details of a preferred flyback power supply with further novel features of the power supply and of the overall control and operating system.
FIG. 6a is a schematic cross-section of a preferred moderately high leakage inductance stress-balanced coil with uniform core area; FIG. 6b is a schematic cross-section of a preferred bobbin for the coil of FIG. 6a for winding wire to provide the suitable moderately high leakage inductance.
FIGS. 7a, 7b are side view cross-sections of the ends of the spark plug tips for use in the hybrid ignition, with the plug tips shown pointing vertically upwards versus downwards.
DESCRIPTION OF PREFERRED EMBODIMENTS
FIG. 1 is a circuit diagram, partially block diagram and partially detailed circuitry, of an embodiment of a distributorless version of the hybrid ignition system depicting two of "n" number of parallel cascaded ignition coils Tl, T2, ... Tn, each comprised of a primary winding la, secondary winding lb, and magnetic core 2. The symbol Ti will be used to designate an arbitrary coil, i.e. the "ith" coil, of an arbitrary number of "n" coils. The coils Tl, T2, ... Tn, are part of a spark discharge circuit including a resonating inductor 3 (LeO), energy storage and discharge capacitor 4 (C), main discharge switches SI (5a), S2 (5b), ... Sn, for the coils Tl, T2, ... Tn, which, with the coil primary windings la comprise a primary discharge circuit.
The primary disc; ge circuit preferably includes the generic shunt diodes/switches SDI, SD2, ... SDn, which may comprise shunting switches, e.g. SCRs as shown in FIG. 4, or clamp shunt diodes (as they shall be referred to) as shown in FIG. 2, which have their anodes connected to the anodes of switches Si (assuming they are SCRs as shown in the embodiment of FIG. 4), and have their cathodes connected to include either the primary coil winding la (and its leakage inductance 3a) by connection to point XI, or the resonating inductance 3 by connection to point X2, where the alternative connection to point XI is shown as a dashed line and the connection to point X2 is shown by a dashed-dotted line. If s1 mt diodes/switches SDi are employed then the initial part of the primary discharge current will flow through switches Si and the later part through shunt diodes/switches SDi for greater efficiency.
For each coil there is defined a secondary circuit comprising the coil secondary winding lb and the output capacitance 7 (Cs) and spark gap 8. Discharge capacitor 4 has a high efficiency, high current diode 9 across it defining a part of the hybrid feature of the ignition.
To charge the discharge capacitor 4, a DC to DC converter 10 is provided with its controller 11 and an output diode 12 for charging capacitor 4 to a high positive voltage, typically approximately 400 volts or other convenient voltage determined by the application and by the voltage rating of available components. The term "approximately" as used throughout this application means within plus or minus 25% of the value it specifies. The power converter 10 is connected to a battery 13, although power supply 13 can be any electrical (DC) power supply depending on the application. For simplicity, the automotive application is assumed, employing a 12 volt car battery. Any variety of energy sources 13 and power converter charging systems 10 can be employed in this hybrid ignition.
The circuit operates when a switch Si (one of 5a, 5b, ...) of coil Ti is closed, resulting in the flow of initial open circuit high frequency "fs" current in the primary and secondary circuits which charges the output capacitor 7 at the frequency fs (designated f20 in this case of FIG. 1) to a high peak voltage Vs (of typically 30 to 45 kilovolts (kV) maximum in the present application as applied to passenger car engines or the like) which subsequently breaks down the spark gap 8. Following spark breakdown, the energy stored in capacitor 4 discharges at a frequency flO of typically 5 kHz to 30 kHz determined by the value of capacitance C of capacitor 4 and the total circuit inductance Le, i.e. sum of of inductance LeO of inductor 3 and the primary leakage inductance Lpe of the coil Ti whose switch Si has been activated. flO = l/l2*pi*SQRT(C*Le)] (1) where pi = 3.142, "*" indicates multiplication, and SQRT indicates the square root of the quantity it qualifies, and Le = LeO + Lpe.
Capacitor 4 discharges for a quarter period T10 (TO = 1/flO) as shown in FIG. la until diode 9 becomes forward biased (capacitor 4 is discharged). Thereafter, diode 9 holds capacitor C at essentially zero volts with the stored electrical energy now residing in the resonating inductance LeO (3) and Lpe (3a).
The current Ip in the inductance decays essentially linearly with time with a time period Tc as shown in FIG. la, which is derived to be approximately:
Tc = [Il(0)*Le]/[Ndp+Vr+Vds/N] = [Vc*SQRT(Le*C)]/[Vckt] (2) where 11(0) is the peak current in inductor 3, Vc is the initial voltage on the discharge capacitor 4, Vdp is the total switch (e.g. Si) and diode (e.g. diode 9) voltage drops in the primary circuit, Vr is the sum of the average voltage drops of the resistances of inductor 3 plus that of the coil primary winding la and its secondaiy winding lb reflected into the primary circuit, Vds is the spark discharge voltage drop, Vckt = Vdp+Vr+Vds/N, and N is the coil winding turns ratio N = Ns/Np, where Np and Ns are the primary and secondary winding turns.
For a distributorless ignition application where low leakage, concentrically wound coils are assumed, or for the distributor ignition case of FIG. 2 where the leakage inductance Lpe is entirely in the coil (a form of stress-balanced coil):
Vc = 360 volts, C = 5 microfarads (uF)
Ec = l/2*C*Vc**2 = 325 mj, where Ec is the energy stored on the discharge capacitor 4 and "**" represents exponentiation.
LeO = 20 microhenries (uH), Lpe = 1 uH; or LeO = 0, Lpe = 21 uH 11(0) = Vc/Ze = Vc/SQRT[(Le)/C) = 175 amps Vdp = 3 volts average, typically between 2 and 5 volts,
Vds/N = 4 volts average, where Vs = 100 volts to 600 volts or greater depending on spark conditions and N is approximately 60 for a discharge voltage Vc of approximately 400 volts (and otherwise inversely proportional to Vc).
T10 = 64 usecs, flO = 16 kHz
Tc = 360*SQRT(21*5)/7 microseconds (usec) = 500 usecs
The spark discharge time of approximately 0.5 millisecond (msec) is in the range of 0.25 to 5 msecs, the preferred spark discharge time for a high speed engine. Typically, Vds is inversely proportional to engine speed, i.e. the spark burning voltage increases with speed induced turbulence.
One can vary the spark discharge time Tc and stored energy Ec by selecting different values of Le and C. For example, doubling Le to 40 uH and C to 10 uF would double Tc, raising it to 1 millisecond and would also double the stress on the magnetic core components of the circuit as will be discussed. The open circuit (high voltage) magnetic flux density B2(x) and the peak closed circuit (spark firing) magnetic flux density B1(0) (or BIO) in the core 2 of a coil Ti are derived from Maxwell's equations and are introduced below in a more generic way that takes into account the case of any number of (in-series) inductances in the circuit from which the definition of "stress-balance" for a coil can be made, where we designate as Lj any circuit inductance and Le the total circuit inductance, where:
Le = Sum of Lj for j = 1 , 2, .. m. For the circuits of FIGS. 1, 2, and 4 respectively:
Le = LeO + Lpe; Le = Lpe; Le = LeO + Lei + Lpe, where Lei is inductance 15 of the auxiliary high frequency circuit of the dual circuit of FIG. 4 which will be disussed presently.
Starting with BlOj, the peak spark firing flux density in inductor Lj: BlOj = lVc/(UF*2*pi*flO*Aj*Nj)]*[Lj/Le] (3) where Aj, Nj are the core area and number of winding turns respectively of the inductor Lj, and UF is the unity factor designed to equal approximately 1 :
UF = 1 + DF (3a)
DF = (N**2)*Cs/C (3b) where DF is called the (voltage) doubling factor and is preferably designed (where practical) to be equal to approximately 0.1.
Designating as B10 the peak flux density in the coil, we obtain: B10T = B10 * ILpe/Le] (3c)
B10 = lVc/(UF*2*pi*fl0*Ap*Np)] (3d) where B10T designates B10 in a circuit that includes other inductances.
The open circuit flux density B2(x) at a time t of x(t) in a coil is: B2(x) = BIO * SQRT(DF/UF)*[Le/Lpe]*[Ap/As]*[x - sin(x)] (4) x(t) = 2*pi*fs*t (4a) fs = flO*SQRT(UF/DF) (4b) where x(t) is the open circuit (high voltage) phase angle, fs is the open circuit frequency, and Ap and As are the core cross-sectional areas on which the primary and secondary windings of f.:rns Np and Ns respectively are wound. If we take DF to have the practical . .__ign value of 0.11, then:
SQRT(DF/UF) = 1/pi, and if we take the preferred value of x, designated as xO, to be 155 degrees (150 to 160 degrees), which we define as the "peak" open circuit magnetic flux density B2(0), or B20, corresponding to 0.95 or [I oos(xO)]/2 of the peak open circuit voltage Vs, then we can write:
B20 = B10 * [Le/Lpe] * lAp/As] * f(x0) (4c) f(x0) = [xO - sin(x0)]/pi, 2/3 < f(x0) < pi/4 (4d) where the two limits represent values of 150 and 160 degrees for xO. We can now define the "stress balance" criteria as: B20 = B10, with B10 satisfying some peak designated value, f(x0) = (Lpe/Le] * [As/Ap] (4e) which is a principal result of this disclosure leading to the development of two types of stress-balanced coils. "Stress balance" assumes the same core material for the core sections on which the primary and secondary windings are wound. For ferrite cores a peak design value of B10 is approximately 4 kiloGauss (kGauss), the core saturation flux density Bsat at approximately 80 degrees C. For the previously defined automotive application with circuit parameters: Vc = 360 volts, C = 5 uF, Le = 20 uH, flO = 16 kHz, B10 = 4 kGauss, a preferred design is a core size of approximately 1 square inch and a number of turns of approximately 12 for a single inductor making up inductance Le.
As already mentioned, for the case of a distributorless ignition with low leakage inductance coils, i.e. Lpe«Le0, the case not of interest since stress balance is not employed by definition, these values, i.e. approximately 1 square inch and 12 turns respectively, are a preferred design for the separate resonating inductor 3 if it were made of ferrite material. This case is also not of interest since it would requires relatively large coils Ti relative to that of FIG. 4 where a dual frequency, i.e. low and high frequency, circuit is used with stress- balanced coils for minimum size of the coils Ti.
However, for the distributor ignition case of FIG. 2 where the circuit inductance is integrated in primary inductor 3a of the coil T the present analysis is of interest. This application represents a form of stress-balancing of the coil in which Lpe = Le, and hence stress-balance is achieved through unequal sizing of the core sections on which the primary and secondary windings are wound, f(x0) = As/Ap, with f(x0) taken to correspond to the 160 degree limit, i.e. f(x0) = pi/4.
FIG. 2 is a circuit drawing of a preferred embodiment of a distributor version of the hybrid ignition with the leakage or resonating inductor 3a integrated into the coil T. The high leakage is achieved by employing a side-by- side winding of the primary la and secondary lb windings as shown in FIG. 2a. Like numerals represent like parts with respect to the earlier figures. In this embodiment two parallel SCRs are shown for the main discharge switch S which can be fired simultaneously or sequentially to reduce the load on the SCRs. Since the leakage (resonating) inductor Lpe (3a) is integrated into the coil and supplies the entire required inductance, no separate resonating inductor Le is required. For shunt diode/switch SD a high efficiency diode 6 is used which shunts the entire primary winding (connection to point XI as per FIG. 1).
FIG. 2a is an approximately to-scale side-view o- awing of a preferred stress-balanced coil for the present automotive application with the following assumed, above introduced, approximate values of discharge circuit parameters:
C = 5 uF; Le = 20 uH; Ne = 12; Vc = 350 volts For this case the core dimensions (ferrite core) are given approximately as:
A = 3 inches; B = 2 1/2 inches; Bl = 1 inch; B2 = 1 1/2 inches;
11 = 5/8 inch; 12 = 1 1/8 inch; Wl = 3/8 inch; W2 = 1/2 inch;
Al = A2 = 1 inch Al is a side of a square core and A2 is a diameter of a round post so that the cross-sectional area represented by A2, i.e. As, is pi/4 of that represented by Al, i.e. Ap, satifying the condition of stress balance (with Lpe = Le):
As/Ap = pi/4 = f(- 1 The voltage doubling facto. DF is taken as approximately 0.11 in the analysis of the stress balance equation. This places a constraint on the value of the total output capacitance Cs, which should be no greater than approximately 160 pf for N = 60, C = 5 uF. Such a total capacitance is easily attainable even in the case with long, capacitive spark plug wires or preferred capacitive plugs. A lower value of total output capacitance Cs lowers the value of B20 relative to B10 as can be seen by including the doubling factor DF explicitely in the analysis:
B20 = B10 * ILe/Lpe] * [Ap/As] * f(x0) * SQRT[K,*DF/UF] (5) which leads to the more general stress balance equation: f(x0) = [Lpe/Le] * f As/Ap] * SQRT10.1*UF/DF] (5a)
With regard to the leakage inductance value Lpe it was experimentally determined that a primary winding of twelve primary turns in three layers (FIG. 2a) naturally provides the required leakage inductance of approximately 20 uH. Taken with C = 5 uF, DF = 0.1, C = 5uF, the following values are obtained for the two frequencies and the open and closed circuit magnetic flux densities: f20 = 50 kHz; flO = 16 kHz; B20 = B10 = 4.2 kGauss,
to provide an optimized stress-balaced design of a simple coil design with an essentially square primary core and a round, cylindrical secondary core for easy winding of the approximately sixty times greater turns of the secondary winding (i.e. of approximately 720 turns).
FIG. 2b shows a top view at the interface of the two core sections of FIG. 2a of thickness C equal to Al and depicting the dimensions of the secondary core section in broken curves and the primary core dimensions in solid lines. Note that if a higher voltage or larger discharge capacitor is employed then the core dimensions can be scaled up accordingly. Also, the wider the primary winding channel Wl the smaller die leakage inductance Lpe, the higher the frequencies flO and f20, and the lower the peak magnetic flux densities BIO and B20. That is, this side-by-side winding coil design is stable against adjustments in core area Ap (and As) since changing the core area Ap changes the channel width Wl inversely and the inductance Lpe proportionally in a way to maintain approximately constant peak magnetic flux densities BIO and B20.
Before considering the distributorless hybrid ignition with stress-balanced coils a form of ignition circuit with integrated power supply is disclosed.
FIG. 3 depicts an integrated hybrid capacitive/inductive ignition system (without recharge circuit), where a single inductor 3 is used for both the power converter stage and for the ignition discharge stage as disclosed in the integrated converter patent '208. Like numerals correspond to like parts with respect to earlier figures. By using diode means 9 across the discharge capacitor 4, then the discharge current (and spark) is a DC current as in conventional, low cost Kettering type ignition, except in this case the ignition has much greater power and is more efficient. Such DC ignition does not require a diode in series with the switch 14 as in patent '208 (except possibly for a low voltage Schottky diode) or a shunt diode across the discharge switch Si (SI, S2, shown), shown as an IGBT in this case (which can also be an SCR).
During power converter operation, energy is stored in the inductor 3 when switch 14 is on, and delivered to capacitor 4 through diode 12a when the switch is turned off. During spark firing, switch Si is turned on, and the discharge current rises sinusoidally at a frequency fl' to a maximum in a quarter period (when voltage across capacitor 4 is zero) and then decays exponentially with a time constant L/R, where L is inductance of inductor 3 and R is the equivalent resistance 12b of the discharge circuit, which is typically about 0.2 ohms for a well designed system. Inductance L typically ranges between 20 uH to 200 uH, the smaller value allowing for higher open circuit discharge frequencies, and hence smaller sizes of the coils Tl, T2 Assuming a 400 volt system, then for L and C equal to about 100 uH and 3 uF respectively, the power converter and discharge circuit will operate at a frequency of about 10 kHz. It is emphasized that the ignition circuit disclosed in FIG. 2 is simple and low cost, as is the case of die Kettering ignition, but in this case is much more powerful, efficient, effective, and easier to control, allowing for peak spark currents in the amp range which decay to constitute a hybrid arc/glow spark discharge.
For the present application of the preferred embodiments of FIG. 2 (distributor ignition) and FIG. 4 (distributorless ignition) a flyback type power converter is preferred as will be disclosed with respect to FIG. 5.
FIG. 4 is a partially block, partially circuit drawing of a preferred embodiment of the distributorless ignition of FIG. 1 in which a dual discharge ignition circuit is provided by means of an auxiliary high frequency (HF) circuit comprised of a capacitor 4a (CI), a shunt diode 9a, a resonating inductor 15 (Lei), and an isolation diode 16 isolating the main, lower frequency, discharge circuit of frequency flO as per equation (1) from the high frequency auxiliary discharge circuit of frequency f3 defined as: f3 = l/{2*pi*SQRT[Cl*(Lel+Lpe)]} (6)
Like numerals represent like parts with respect to FIG. 1.
The high frequency auxiliary circuit is particularly simple in that it requires only a diode to isolate it fiom the main discharge circuit. In operation, when switch Si (of SI, S2, ...) is closed, capacitor 4a discharges through inductor 15 and the coil primary winding la at an initial open circuit frequency f4 which is higher than the main discharge open circuit frequency f20. For example, if f3 is approximately two times greater than flO, then so is f4 approximately two times greater than f20 since the open circuit frequencies f4, f20 are related to the closed circuit (spark firing) frequencies in approximately the same way according to: f20 = flO * SQRT[(1 + DF0)/DF0] (7a) f4 = f3 * SQRT[(1 + DFD/DF1J (7b) where DFO, DF1 are the voltage doubling factors given by:
DFO = (N**2)*Cs/C (8a)
DF1 = (N**2)*Cs/Cl (8b)
In this analysis the ignition is operated in the preferred voltage doubling mode wherein the doubling factors are preferably less than 0.2, and preferably approximately 0.1.
The result of a higher open circuit frequency f4 is a proportionally smaller coil Ti core size, all other things being equal, as will be discussed with reference to FIGS. 6a, 6b. Preferred values for the circuit parameters for the main automotive application, assuming (small) ETD 54 cores for coils Ti, are:
C = 3.6 uF; LeO = 30 to 60 uH; Vc = 350 volts
CI = 2.3 uF; Lei + Lpe = 15 to 18 uH where LeO of 60 uH is achieved by winding approximately 20 turns of wire (Litz wire preferred) on a laminated EI-3/4 core (core area 3/4*3/4, or 0.56, square inch) with preferably 7 mil laminations and with the "I" leg absent, i.e. an open core where precautions are taken that no large metallic material are within 1/4" to 1/2" of the open end of the core.
Preferably, the high frequency inductance (Lel+Lpe) is obtained by using a stress-balanced coil with leakage inductance Lpe approximately 0.6 of the total inductance (Lpe+Lel). That is, turning to the stress balance equation (5a), and taking As equals Ap (the coil core on which the two windings are wound is uniform as in standard cores), we obtain for the stress balance condition:
Lpe/(Lpe + Lei) = f(x0)/SQRTf0.1 *UF/DF] Assuming N = 60 and Cs = 60 pF, then DF = 0.09, UF = 1.09, leading to:
Lpe/(Lpe + Lei) = f(x0)/1.10 and taking the lower limit of f(x0), i.e. 2/3, we obtain:
Lpe/(Lpe + Lei) = 0.6
The other condition for stress balance is that the values B20, B10, which we will designate as B40, B30 respectively for the present preferred high frequency auxiliary circuit, are equal to 4 kGauss, e.g. 4.2 kGauss. These are satisfied if Lpe is made to equal to 10 uH for a primary turns Np equal to 11 turns for an assumed ETD 54 core, where the term "equal to" as used in this context means within plus or minus 10% of the value it specifies. These special values of Lpe and Np, and low value of Cs, are achieved by the coil designs of FIGS. 6a, 6b and will be disclosed there. The inductor 15 (Lei) provides the remaining small inductance of 6.5 uH required according to the above equation.
For the case where it is desirable to have a long spark firing period Tc an inductance LeO of approximately 1.6 milliHenry (mH) can be used which is obtained by winding approximately 100 turns of 16 to 18 gauge wire, e.g. 17 AWG magnet wire, on an open El- 3/4 laminated core. For a voltage Vd = 10 volts, where Vd = Vdp + Vds/N, and an assumed coil resistance Rcoil of coil LeO of approximately 0.2 ohms, comprising essentially the entire circuit resistance, the spark firing time constant Tc is approximately 2.4 msecs, using the more exact version of equation (2) given by:
Tc = (ln[I(0) + (Vd/Rcoil)]/(Vd/Rcoil)} * LeO/Rcoil (9)
In this application, the spark discharge current is made up of two distinct components as shown in FIG. 4a as the dashed curve, an initial arc component of approximately 2 amps peak current and of short duration of approximately 0.3 msecs, and a glow discharge spark current of approximately 250 ma peak with the long duration or discharge time constant Tc of approximately 2.4 msec. A longer time duration is achievable by increasing the inductance Le according to equation (9), e.g. by winding 160 turns of 18 AWG magnet wire on an EI-3/4 open laminated core an inductance of approximately 4 mH is obtained and a time constant of approximately 3.6 msecs. Such a design (of long or even longer time constant Tc) can be of particular use for ignition systems of slow speed stationary gas engines.
Returning to FIG. 4 to complete the analysis of the circuit operation, it is noted that SCRs 6a, 6b, ... , are employed for the shunt switches SDI, SD2, ... , for the automotive case where the coils Ti are located close to each other, e.g. adjacent to each other, and the cathodes of the shunt switches are returned to the preferred location X2 which includes the main inductor 3 in the circuit. Note that for the case where inductance 3 is large, i.e. of the order of magnitude of 2 mH, i.e. between 0.2 and 20 mH, it may be simpler to return the cathodes of the shunt switches/diodes to location XI, i.e. across the primary windings of the coils, and use diodes instead of SCRs, such as high efficiency Motorola MR2406 diodes having a low forward drop of approximately one volt at a high current of 50 amps or greater. This would be of particular interest where the coils Ti must be remote from each other as in large stationary gas engines.
When the ignition is fired, say switch SI of coil Tl is turned on, and the discharge capacitors 4 and 4a complete their quarter cycle discharge and the electrical discharge energy is stored in the form of current in inductors 3 (LeO), 15 (Lei), and in the primary leakage inductance 3a (Lpe), the current can flow
either in a circuit including the capacitor shunt diodes 9 and 9a with the switch SI (5a), or in a circuit excluding switch SI (5a) which is replaced by the corresponding shunt switch SDI (6a). The path including the shunt switch SDI is the preferred path since it represents only one forward drop versus two drops for the path including the capacitor shunt diodes 4, 4a and control switch SI (SCR 5a shown). Essentially all the inductive current goes through the shunt switches SDi and very little through the SCR switches Si, reducing the electrical dissipation on the SCR switches Si and making for a more efficient operation. Note that a diode 6aa (6bb for SCRs 5b,6b) may be required between the triggers of the two SCRs SI (5a) and SDI (6a), i.e. with its anode and cathode connected to the triggers of SCR switches SDI and SI respectively as shown, to insure that the shunt SCR switch of a non-firing SCR (say SD2) is not inadvertently turned on during the firing of another SCR pair, i.e SI and SDI.
The preferred packaging of parts for the hybrid ignition is in two boxes: 1) a power box including the power converter 10 and controller 11, and 2) the coil assembly comprised of the inductors 3 and 15, the low and high frequency capacitors 4 and 4a and their shunt diodes 9 and 9a and the isolation diode 16, and the coils Ti with their switches Si and clamp shunt switches SDi, where i=l, 2, 3, ... . Preferably, coils Ti are stress-balanced coils as shown in FIGS. 6a, 6b. For the distributor ignition (FIG. 2) a single enclosure containing the entire ignition is both practical and preferred.
Note that in the analyses disclosed, the voltage doubling criterion was assumed. An alternative way to invoke the voltage doubling criterion is through a frequency criterion of requiring that the open circuit frequencies f20, f4 be approximately three times (between 2.25 and 3.75 times) greater than their closed circuit spark firing frequencies flO and f3 respectively, i.e.
SQRTf(l + DF)/DFJ = 3 where DF is a generalized voltage doubling factor.
FIG. 5 is a more complete more detailed circuit drawing of a preferred embodiment of a hybrid ignition system which includes control features for both the ignition and a preferred flyback power converter circuit shown. Like numerals correspond to like parts with respect to the earlier figures.
The ignition system shown is of the hybrid ignition type disclosed in FIG. 4. The power converter is comprised of circuit block 10 and the controller of circuit block 11. The trigger circuit assumes an arbitrary trigger input conditioner 19a, and the phase input assumes a phase conditioner 19b for providing cam- based phasing signals for the present case shown of a disuibutorless ignition. Circuits suitable for 19a and 19b for providing conditioning of the signals are known to those versed in the art.
The flyback power converter 10 includes input filter capacitor means 20 (for minimizing stray inductance), flyback transformer 21, which for a 100 watt application can be designed around an ETD-39 gapped core of 1 1/4 square cm core area, with 5 to 8 primary turns, with about 12 uH primary inductance, and a turns ratio of approximately 12 (assuming the main switching transistor 22 is a high efficiency 60 volt FET and the output voltage is approximately 360 volts, the preferred automotive case, and assuming the peak current through the transformer 21 primary winding is approximately 20 amps). The drive for FET switch 22 is provided by a transistor pair 23 and 24 to provide turn-on of FET 22 when a positive voltage is supplied to the bases of transistors 23, 24 to charge the gate of FET 22 through resistor 23a, and turn off FET 22 when the bases are pulled low and the FET gate is rapidly discharged (for minimum switching loss) through diode 24a and transistor 24.
For the snubber of the flyback a snubber capacitor 25 and diode 26 is used as is known to those versed in the ait, except in this case a lossless (or low loss) snubber circuit is provided by means of transistor switch 27 and inductor 28 and diode 29. Zener 27a (e.g. a 20 volt zener) and divider resistors 27b, 27c
control operation of FET 27. Preferably capacitor 25 is about 0.1 uF and inductor 28 is about 50 uH, where the term "about" as used throughout this application means between one half and twice the value it specifies.
The power converter controller 11 is based on charging a timing capacitor 30 (of capacitance Ct) from the output voltage Vc through a resistor 34 (resistor Re) which provides a required decreasing charging time with increasing output voltage, defining an off-time Toff. The on-time Ton represents the discharging of capacitor Ct through a resistor 31 (of value Rb) connected to the output of a comparator 32 through an isolation diode 31a. The charging capacitor 30 is connected to the inverting input of a comparator 32, and the non-inverting input has a reference voltage obtained from the divider resistors 33a, 33b, 33c which make the non-inverting input flip between l/3*Vcc and approximately 2/3*Vcc depending on whether the comparator output is low or high. The normally high output of the comparator timer 32 corresponds to the off-time (Toff) versus the usual on-time (Ton).
The comparator (32) timer oscillator circuit, i.e. the "Timer", can be designed to turn main switch 22 on and off with our without a DC current. Preferably, operation with DC is employed which is set by a sensor circuit comprised of the NPN sensor transistor 18, resistor 17 connected between the secondary winding of transformer 21 and ground, and temperature regulating thermistor 17a connected across resistor 17. The base of sensor transistor 17 is at ground and its emitter at the high side of resistor 17 of value about 0.5 ohms for the present application so that the transistor switches when its base-emitter voltage rises above 0.62 volts and the current through the secondary is above the threshold current Ith (of approximately 1.2 amps in this case). The collector of sensor transistor 17 is tied to the low side of the off-time resistor Re (34) to divert timing capacitor 30 charging current when the sensor current rises above Ith to increase the off-time and stabilize operation.
The timing (charging) resistor 34 is connected at one end to the voltage node Vc and at the other end through a shunting zener diode 35 and small resistor 36 which shunt the other timing resistor 31 of resistance Rb. In operation, capacitor 30 is charged by voltage Vc through resistors 34 and 36 representing the off-time Toff, to raise the capacitor 30 voltage from l/3*Vcc to approximately 2/3*Vcc. The "Timer" then switches, and capacitor 30 discharges (with on-time Ton) through resistor Rb to l/3*Vcc.
Zener diode 39 is a voltage limiting diode of approximately 9 volts zener voltage which, in addition to providing over-voltage protection, provides a high battery voltage shut-off of the timer oscillator and of the boost converter.
Since die "Timer" is operated in a reverse mode, an inverting output circuit is required, comprised of a NPN transistor 40 with its emitter to ground and its collector connected through pull-up resistor 41a to Vcc and its base connected to the comparator 32 output through a base resistor 41b. A base emitter resistor 41c is also included and output of comparator 32 is connected to Vcc via pull-up resistor 4 Id. Transistor 40 inverts die comparator oscillator timer output node and supplies current to the driver transistors 23, 24 of main FET switch 22. In this way, the boost converter is provided with the required "on-time" drive for say 15 amps average current, and with the required "off- time" drive as a function of the output voltage (and peak secondary current).
In this controller operation if Vc falls below approximately 2/3*Vcc the charging capacitor 36 can never charge up and the output stays low to provide a built-in low output voltage shut-off. For power converter start-up following spark discharge, the discharge capacitor 4 is charged in less than a millisecond to above 2/3*Vcc from the supply Vcc through resistor 42 and transistor 42a controllable by the ignition firing to keep it off during spark firing through turn- on of shut-off transistor 47 pulling base resistor 42b to ground. Shut-off tran¬ sistor 47 holds timing capacitor 30 low through diode 46 during spark firing.
Power converter tum-on is also speeded up by partial charging of timing capacitor 30 directly through hysteresis resistor 37 (and diode 37a) and resistor 43, connected to Vcc, of value about one half of die charging resistor Re.
Regulation of the output voltage Vc is controlled by comparator 38 whose inverting input is connected to a voltage divider made up of resistor 44 (e.g. 360 kOhms for 360 volts output) and resistor 45 (e.g. 5 kOhms for a 5 volt reference Vref on the non-inverting input of the comparator 38).
In FIG. 5 is shown an ignition trigger conditioner 19a which can be designed by those versed in the art to control ignition firing by converting any of a number of possible ignition trigger input signals into a well defined short trigger pulse which, in this case, is applied to the base of an NPN transistor 50 whose collector is connected to Vreg and whose emitter is connected to capacitor 51 (Csig) shunted by a timing resistor 52 (Rsig) defining a decay time constant Tsig. The emitter is connected to the inverting input of comparator 53 and the collector to a circuit block 54 designed to provide a variable signal with engine speed to the non-inverting input of comparator 53. The output 55 of the comparator is normally high through connection to Vcc through pull-up resistor 56 with resistor 57 acting as a hysteresis resistor. When an input trigger is received at trigger conditioner 19a, comparator output 55 is pulled low to GO, designated as the spark trigger "gate" GO, and modulated (reduced) with increasing engine speed by circuit 54, ranging from several msecs at low speeds to about 1 msec or less at high speed for typical automotive applications.
The comparator output 55 is indirectly connected to the power converter controller timing capacitor 30 through connection to base of shut-off transistor 47 to turn-off the power converter when the ignition is firing (output 55 is low), a preferred operating condition for the hybrid ignition which does not employ recharging of the discharge capacitors during ignition firing.
Finally, output 55 of comparator 53 is connected to an inverting stage 60 whose output G is connected to the bases of drive transistors 63 (NPN transistor with collector to Vcc) and 64 (PNP transistor with collector to ground). The emitters of transistors 63 and 64 are inter-connected and represent the drivers for the ignition triggering SCRs 5a, ... , and shunt switch SCRs 6a, ... , etc.
For a distributor system the drivers would be connected to a capacitor 65a (capacitance of order of magnitude 1 uF), producing a positive pulse to the trigger of the SCR on spark firing (beginning of gate G) and a negative pulse at the end of gate G. For a distributorless ignition (case shown) there is included steering (FET) switches 66a, 66b whose gates are connected to the outputs of a spark steering counter 67 and whose inputs are connected to the comparator output 55 (GO) and to the phasing signal output of comparator 58 (connected to Vcc via a pull-up resistor 59) which resets the counter when a signal is received from the output of comparator 58. At every firing cycle of all the engine cylinders, a phase signal is received at the reset pin (RST) of counter 67 which resets the outputs to begin another complete firing cycle of a multi-cylinder engine employing a distributorless ignition.
While the initial high frequency discharge circuit (supplementing the main discharge circuit of the dual discharge circuit) was shown with reference to the distributorless ignition, it can also be included in the distributor ignition. Other combinations of high and low frequency circuits are possible to achieve a good balance between small core size, efficient circuit operation, and acceptable spark energy delivery. Also, other control strategies are possible, including, for example, using multiple paralleled SCRs with a distributor ignition which are triggered from a counter either individually or in pairs for application where unusually high energy and power is being delivered at a rapid firing rate as in high speed racing engines.
As already discussed, for the distributor ignition, a single enclosure containing the entire ignition is both practical and preferred. For the distributorless version there are several possibilities including having a coil assembly with the coils Ti and switches Si contained within it, and the discharge capacitors and resonating inductors nearby in a separate enclosure which may or may not include the rest of the ignition system.
In the disclosure of the ignition system of FIG. 4 designs were presented in which preferred stress-balanced coils were employed. Such preferred stress- balanced coils are disclosed with reference to FIGS. 6a and 6b.
FIG. 6a depicts an approximately to-scale partial schematic side-view drawing of a preferred stress-balanced coil 71 for the preferred dual discharge distributorless ignition of FIG. 4 (with coils Ti, i=l,2,3, ... ,) employing a high leakage inductance ferrite E-core structure designed to have the required stress balance feature of equal open circuit peak magnetic flux density B40 and closed circuit flux density B30, equal to the saturation flux density Bsat of the core material 2 of, for example, 4.2 kGauss as disclosed with reference to FIG. 4.
For the present application is used a preferred more standard E-core such as the newly introduced ETD-54 core, with a center post diameter A0 of 0.75" (cross-sectional area Ap of 0.43 square inches or 2.8 square cms), a winding length "1" of approximately 1.5 inches and a window width "w" of approximately 0.43 inches, with preferred number of primary turns Np in the range of 10 to 12 and with a preferred turns ratio N of approximately 60. For an assumed separate leakage inductance 15 of value Lei of 5 to 8 uH a side-by-side primary la and secondary winding lb must be used to obtain significant leakage inductance, i.e. higher leakage inductance Lpe than Lei as derived in the stress balance criteria with reference to FIG. 4. A preferred doubling factor DF of approximately 0.1 was assumed to provide an open circuit frequency f4 approximately three times the short circuit frequency f3.
It is found that for the stress balance criteria, a side-by-side winding in which the secondary winding lb is the conventional layer wound structure provides too high a leakage inductance and hence too high a short circuit flux density B30. That is, for a prefeπed 11 turns primary winding la the leakage inductance Lpe is 14 to 15 uH, which for a preferred capacitance CI of 2.3 uF, inductance Lei of 6.5 uH, and primary voltage of 350 volts gives a peak closed circuit flux density B30 of 6 kGauss, forty percent higher than the maximum allowable of 4.2 kGauss for ferrite cores. In addition, the conventional layer winding provides a relatively high coil output capacitance (Cs)coil.
This problem was resolved through the use of the compartmentalized secondary winding lb of a bobbin 72, shown at approximately twice scale in FIG.όb, in which the separation 73a (thickness "tO") between the primary and secondary winding is minimum, i.e. approximately 0.030 inches or less, versus approximately 0.20 inches. This reduces the leakage inductance Lpe from 14.5 uH to 11 uH, nearly the required forty percent to make for an ideal and practical design. Hence the use of the compartmentalized winding with a primary winding la of preferred 11 turns of Litz wire of approximately 0.10 inches diameter packed into three or four layers of primary compartment 74 about three times the size of the preferred six secondary winding compartments 74a to 74f which preferably use 30 gauge (28 to 32 gauge) heavy insulated magnet wire, with the whole structure encapsulated to withstand high voltage. In the preferred embodiment shown, the secondary compartment length "dl" is approximately 0.125" and the separations 73b to 73f have a progressively reduced thickness tl to t5 of approximately 0.060 to 0.030 inches as do the turns per compartment Ni (i=l to 6) which progressively drop from, for example, 160 to 60 in decrements of twenty turns per compartment for a preferred total number of secondary turns Ns of 660 for the preferred turns ratio N of 60 for 11 primary turns Np. Having the highest number of turns Nl adjacent to the primary winding la helps further
reduce the primary leakage inductance to the required 10 uH for primary turns Np of 11. The bottom layer 75 of thicknesses tij, i.e. tOl, tl2, ... t56, between the bottom of the secondary compartments and the inner diameter of the bobbin (the core surface) are also tapered, increasing from approximately 0.04" (tOl) to approximately 0.10" (t56) to accommodate the progressively higher voltages of the secondary winding lb which has its low voltage end 76 at the compartment 74a adjacent to the primary winding 1 a and its high voltage end 77 (connected to the high voltage tower 8a, FIG. 6a) at the last compartment 74f.
The secondary winding is shown as cross-hatched layers where the progressively reduced windings Nl, N2, N3 increase the margins between the top of the winding 78 in each compartment and the magnetic core surface to accommodate the progressively increasing voltage, with the last compartment 74f of turns N6 having margins of approximately 0.15" between the winding top surface 78 and the inner core sidewall 2a (FIG. 6a) and the inner core top wall 2b (FIG. 6a) to accommodate the preferred peak secondary voltage of approxim¬ ately 36 kVolts. For me dimensions and core disclosed, the outside diameter of the bobbin 72 is approximately 1 1/2", which for an inner hole diameter of 3/4" makes the fin dimensions approximately 3/8" less the thickness tij.
For the six compartments shown and given dimensions and wire size the peak voltage between turns (the first and last turns of two consecutive layers in a given compartment) is approximately 1000 volts for 36 kVolt operation, or approximately 250 volts per mil for quad-coated magnet wire of thickness approximately 0.002 inches (2 mil thick). Employing more compartments, e.g. seven compartments of 0.10 inch size will reduce the electrical stress, which may be preferred although proper encapsulation of the entire coil should insure adequate voltage piotection for six secondary compartments. The bobbin is slotted for communication between compartments as is well known to those versed in compartment windings, and for improved encapsulation.
Other advantages of the compartment winding is low secondary winding capacitance to provide a high open circuit frequency of 80 to 100 kHz for the preferred circuit and coil parameters disclosed above and in the disclosure of FIG. 4 to satisfy the stress balance criteria. Also, the bobbin structure 72 of the compartment winding simplifies large-scale manufacture (winding) of the coil.
In the preferred embodiment of the disclosure of FIG. 4 employing coils of FIGS. 6a and 6b, resonating inductances 3 is preferably approximately 50 uH for me prefeπed 350 volts Vc operating primary voltage, which with the other specified parameters produce peak short circuit primary Ipl and secondary currents Is of 120 amps and 2 amps respectively, where Ipl is the higher frequency peak current through discharge of capacitor CI. The turns ratio N is varied depending on application but preferably not more than by 25% to maintain other required relationships. The cores for inductance LI (15) can be a standard E625 ferrite gapped core; the core for inductance L0 (3) is preferably an open laminated core as already disclosed (an EI-3/4 being preferred).
In the application where the coils must be remote and where no high frequency resonating inductor 15 is employed, a larger leakage inductance 3a, 3b, ..., may be desired, e.g. 12 to 20 uH, but with the required lower ratio of leakage inductance Lpe to the number of primary turns Np of approximately one uH per turn as obtained with the coil designs of FIGS. 6a, 6b. To accommodate this, a somewhat larger core area (larger center post diameter A0) is required. For example, using the recently introduced ETD-59 core with core area of 3.65 cm square, a preferred design is one capacitor CI per coil equal to approxi¬ mately 2.5 uF and a primary turns Np of 12 or 13 for a leakage inductance of 13 or 15 uH respectively. The turns ratio N would preferably be higher, e.g. 66, for a higher 42 kV peak output required in the typical stationary gas engine application where such remote coils may be required.
It is to be noted that as an alternative preferred embodiment to the distributor ignition design of FIG. 2, a smaller E-core than that of FIGS. 2a, 2b, may be employed, having a uniform center winding post, e.g. an ETD-59 core or larger, according to the designs of FIGS. 6a, 6b, but with a separate resonating inductor 3 of small inductance less than the coil leakage inductance 3a (Lpe) to satisfy the stress balance criteria.
With regard to applications of FIG. 4 and 5 to large stationary gas engines running at constant speed, e.g. 300 RPM, it is noted that it is a relatively simple matter, known to those versed in the art, to place time delay circuits between the output of a firing sequencer device such as a counter 67, FIG. 5, to achieve variability in the firing of the switches SI, S2, ... , of the coil discharge circuits relative to the trigger input signal that is received. In this way, variations in the burn or combustion time in different engine cylinders that often exist with large stationary gas engines can be accommodated by varying the time delays of the various cylinders to achieve peak pressure at equal piston positions with respect to engine top-center.
The high peak secondary spark currents in the ampere range require high erosion resistant and fouling resistant plugs as depicted in FIGS. 7a, 7b.
FIG.7a depicts an approximately three times scale drawing of a side-view end section of a prefeπed toroidal or circular gap plug which may be useful with the present hybrid ignition. The plug has a threaded shell 80, center conductor 81, insulator 82, and a firing end 83 with upward and outwardly extended electrodes 84 which form an extended gap with the shell edge 85 (at an angle theta with the vertical of preferably 15 to 75 degrees) which is less than the gap formed with the horizontal, e.g. a 0.10" versus 0.12" gap as shown. This design reduces both erosion of the plug and fouling of the insulator end surface 86 by keeping the spark 87 away from it. For a 14 mm plug, the shell inner diameter is preferably a large 0.40" to accommodate the 0.12" gap shown between the
insulator surface 86 and the inner surface 88 of the spark plug shell 80, making for a thin spark shell of only approximately 0.050" versus the conventional 0.1". The insulator end section 89 of the insulator 82 is extended far enough into the spark plug shell to define a large air volume 90 of sufficient length to minimize the possibility of tracking of the spark along the surface 91 due to fouling.
FIG. 7b depicts and alternative design of the plug of FIG. 7a with like numerals corresponding to like parts with respect to FIG. 7a. In this design, the end electrodes coπespond more closely to a standard plug, the extended electrodes 84a comprise the tip of extended sections 80a, 80b of the shell 80 versus the center conductor 81 with the shell tip 84a forming a spark gap with an end section 85a of the end button 83a of the center electrode 81. The shell extensions 80a, 80b can be multiple electrodes or a continuous circular surface whose tip 84a makes an angle theta of preferably 15 to 75 degrees with the vertical. The actual angle theta selected will depend on the application. If the shell tip 84a defines a continuous essentially circular surface then preferably opening means are provided in the side of the shell extensions 80a to allow air- fuel mixture to flow through the plug end from side to side.
In the design of FIG. 7b (which can also be applied to FIG. 7a) is shown a recessed insulator section 82, 89, 91 which foπns an unusually large air volume 90 of about 0.3" length and of radius approximately 0.15" (which can be larger for larger spark plugs such as 18 mm or larger plugs used in stationary gas engines). The insulator is shown extending from the larger diameter shell section 92 to achieve maximum recessing. Besides minimizing possibilities for fouling, the larger volume 90 defines a combustion volume which can be coated with a catalyst material 93 such as palladium oxide on all the inner metallic surfaces defining the large volume 90, i.e. surfaces 81a and 88, as well as the outer extended electrode surfaces 80a and 80b as shown to enhance the combustion reactions.
Besides, or along with, catalyst combustion enhancement within the spark plug, electric field enhancement can be employed by using the hybrid ignition with dual cischarge circuits (FIG. 4) and selecting the main low frequency inductanc. L0 to be about 2 mH. This results in the maintenance of the maximum oltage across the volume 90 between the center conductor surface 81a and the inner shell surface 88 during the second low frequency spark discharge stage covering the spark current range of 50 ma to 200 ma (where the spark burning voltage Vds is maximum) for the major part of the spark duration. The voltage Vds is maximum for a spark current around 100 ma, in the range of 600 to 1,500 volts depending upon sparking conditions (gap size and mixture flow through gap), producing an electric field E of 1,600 to 4,000 volts per cm (for the radial dimension of 0.15" shown) which can enhance combustion reactions in the volume 90, especially near the inner electrode surface 81a.
There are many changes that can be made in the preferred embodiments disclosed within the inventive principles disclosed herein for the hybrid ignition with stressed-balanced coils. Certain features disclosed need not be limited to the hybrid ignition, the preferred ignition and subject of this patent application.
For example, if instead of a standard ETD-54 core for the preferred coil of FIGS. 6a, 6b, a similar core is employed with a wider window width "w" of say 0.5" and quad or heavier insulating coated wire for the secondary winding, then the window length "1" could be reduced to say approximately 1.25" and the bobbin designed with fewer compartments, e.g. four compartments, while still preserving the essential features of approximately 10 uH primary leakage inductance Lpe for 11 turns of primary winding Np to satisfy the stress balance criteria (open and closed circuit peak magnetic flux densities of comparable value near the saturation value Bsat). Also, other core materials could be used with different magnetic saturation values Bsat and designs accordingly modified to adhere to the inventive principles disclosed herein.
As another case, one can have a lower energy distributorless ignition, hybrid or not, in which only one capacitor means of value, say, 2.5 uF of 400 volt rating is used with no resonating inductors but larger coils with ETD-59 cores with side-by-side windings as in FIG. 6a with approximately 12 turns of primary winding, which would lead to a particularly compact coil assembly of the coils Ti, the capacitor means, switches Si, and clamp shunt diodes SDi.
A further case is to employ a single component for the dual switches Si and SDi since they share a common anode and triggers. The semiconductor model for an SCR is a four layered semiconductor structure P1-N1-P2-N2, where P and N indicate P-type (positive) and N-type (negative) semiconductor material and where the anode connection is made to layer PI, the trigger to layer P2, and the cathode to layer N2. In place of the two sets of four layered structures for SCRs Si, SDi, one can design a single four layered structure P1-N1-P2-N2/N3, where the two common connections (of SCR switches Si, SDi) of the two anodes and triggers are made to layers PI and P2 respectively, and in place of a single layer N2 two paralleled separated layers of N-type material are used to make up the two separate cathode connections N2 and N3. We call this device a CDRCAT device, for "Controlled Dual Rectifier Common Anode Trigger".
It is therefore particularly emphasized with regard to the present invention, that since certain changes may be made in the above apparatus and method without departing from the scope of the invention herein disclosed, it is intended that all matter contained in the above description, or shown in the accompanying drawings, shall be interpreted in an illustrative and not limiting sense.