THE UNIVERSITY OF MICHIGAN INDUSTRY PROGRAM OF THE COLLEGE OF ENGINEERING THE EFFECT OF THE RATE OF ENERGY INPUT UPON THE MINIMUM SPARK IGNITION ENERGY OF LEAN PROPANE-AIR MIXTURES John C. Steiner A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the University of Michigan Department of Mechanical Engineering 1964 April, 1964 IP-666

ACKNOWLEDGMENTS The author expresses his sincere appreciation to Professor William Mirsky, Chairman of the Doctoral Committee, for his constant encouragement, advice and assistance throughout the course of this investigation. The writer also expresses his gratitude to the members of the Committee, Professor J. A. Bolt, Professor S. W. Churchill, Dr. M. E. Milberg and Professor G. J. Van Wylen. Acknowledgment is due to Professor M. B. Stout of the Electrical Engineering Departmentof the University for his help on several occasions. The financial assistance given by General Motors Corporation, Shell Oil Co., Texaco, Incorporated and the Institute of Science and Technology of the University in the form of Fellowships is gratefully acknowledged. Provision of material and equipment by the Mechanical Engineering Department of the University for this investigation as well as the use of the Computing Center facilities is appreciated. Acknowledgment is also due to the Delco Remy Division of General Motors Corporation for furnishing a special ignition coil. Thanks is due to my wife, Lucy, for preparation of the manuscript. ii

TABLE OF CONTENTS Page ACKNOWLEDGMENTS..................................................... ii LIST OF TABLES.....................o......................... LIST OF ILLUSTRATIONS................................................ i NOMENCLATURE.............................................o o o.... ix I. INTRODUCTION...................................... 1 A. Purpose................................................... B. General..................... O............ 1 II. SURVEY OF PREVIOUS WORK................................. 6 III, THE MECHANISM OF SPARK BREAKDOWN............................ 29 A. General............................................. 29 B. Photoelectric Theory.o........................ 29 C. Townsend Theory.............................. 30 D. Streamer Theory....................................35 Eo Paschen's Law and Similitude.......................... 39 F. Time Lags o..................o.............o..... 41 IV. SURVEY OF IGNITION THEORIES o.............................o45 A. General.............................................. 45 B Thermal Theory.........................o.............. 45 C. Semenov's Self-Ignition Theory.................o. OO 46 D. Fenn's Analysis.......................o........... 50 E. Excess Energy Theory..................oo........o.. 53 F Work of Yango...........................O.......O...... 54 G. Work of Mole........................................ 56 H. Landau's Work.............o.............O............. 57 I. Application to Present Investigation,................... 59 V. EXPERIMENTAL APPARATUS,..o............................. 61 Ao Description of Equipment............................o 61 1. Constant Volume Bomb................................ 61 2. Bomb Heater and Cooler..............................0 64 3o Manifold System........ o................6........ 4. Temperature Measurement Equipment.... 6.............. 68 5. Bomb Pressure Transducer.............................. 68 6. Spark Control Circuit................................. 69 iii

TABLE OF CONTENTS (CONT'D) Page 7. Voltage and Current Measurement........................ 75 8. Photomultiplier Circuit............................... 75 9. Ultraviolet Light Source............................. 76 B. Calibration of Instruments............................... 76 1. Oscilloscope Sweep Rate........................ 76 2. Voltage and Current Measurement........................ 78 3. Pressure Gages................................... e.. 78 4. Pressure Transducer.................................... 79 5. Thermocouple Calibration...........oo.o............ 79 6. Frequency Response of Photomultipliere oo............... 79 VI. EXPERIMENTAL PROCEDURE........................................ 81 A. Fuel-Air Mixing............................................ 81 B. Measurements of Ignition Energy.......................... 83 C. Electrode Spacing...O...o............................... 84 Do Quench Distance........o o..... oo.............................. 84 VII. RESULTS AND DISCUSSION......................................... 86 Ao Results..............o o...ooo....o.....o............... o 86 B. Discussion................ oo o.......................... 87 VIII. CONCLUSIONS AND RECOMMENDATIONS..O....o o...........o........ 115 A. Conclusions........ o o o o...o o...o o... o o. o. o o.... 115 B. Recommendations o.oo.o.o.o e.. ooo.o.o o oo..o. o o........o...o 115 1o Future Investigation..O................................ 115 2, Equipment Modification 00.. 0...0....00.0.....00..00.,o 116 APPENDICES A DERIVATION OF PRE-BREAKDOWN CURRENT EQUATIONS.................. 118 B DETERMINATION OF INTERELECTRODE CAPACITANCE OF A COIL.......... 125 C DESIGN REQUIREMENTS FOR A VOLTAGE DIVIDER...................... 129 D CIRCUIT AND CALIBRATION DATA,..............o..........,..o...o 133 E ORIGINAL DATA.................. o o.................. o. o..... o o o 141 F DATA REDUCTIONo.o....o... e o.oeo.o o.. o o o..e o...o.o oeo 156 BIBLIOGRAPHY,., ooo ooo ooo o oooo o..oo.oe oo o oo o oo oo o ooo.,. o o, ooooo 163 iv

LIST OF TABLES Table Page I Bradford, Finch and Prior Data............................... 11 II Empirical Constants for Energy Equation o......00..0000.00O.oo 99 III Resultso Equivalence Ratio = 0.83, Gap = 0.091 inches....... 107 IV Results. Equivalence Ratio = 0.83, Gap = 0.122 inches...... 108 V Results. Equivalence Ratio = 0.73, Gap = 0.195 inches...... 110 VI Results. Equivalence Ratio = 0.69, Gap = 0o195 inches....... 112 VII Secondary Capacitance Data for Mallory Coil................. 135 VIII Secondary Capacitance Data for Delco Coil.................. 135 IX Spark Control Circuit Data................................... 136 X Counter Calibration Data..................................... 137 XI Oscilloscope Sweep Rate Calibration Data.................... 137 XII Pressure Gage Calibration Data.............................. 138 XIII Fuel-Air Mixing Information............................... 140 XIV Experimental Mixture Data............................. 142 XV Experimental Ignition Data for Equivalence Ratio = 0.83, Gap = 0.091 inches.......................................... 143 XVI Experimental Ignition Data for Equivalence Ratio = 0.83, Gap = 0.122 inches,... o... o......... o o........ o o 145 XVII Experimental Ignition Data for Equivalence Ratio = 0,73, Gap = Oo195 inches.......oo.......................o o......... o..o 148 XVIII Experimental Ignition Data for Equivalence Ratio = 0.69' Gap = 0.254 inches......................................... 151 v

LIST OF ILLUSTRATIONS Figure Page 1ol Conventional ignition circuit................................. 3 1o2 Voltage current and power as a function of time in a conventional ignition spark.................................. 3 2.1 Ignition system used by Lewis and von Elbe................ 13 2.2 Minimum ignition energies for various electrode configurations 13 2.3 Minimum ignition energy EMiN as a function of the percent fuel in air o..................................................... 16 2-4 Quench distance dq as a function of the percent fuel in air... 16 3.1 Circuit for measurement of pre-breakdown currents........... 31 3.2 Pre-breakdown current-voltage characteristics................. 31 3,3 o Electron avalanche......................O O............ O 37 3.4 Formation of a streamer O..........................00000 o o... 37 3.5 Effect of time lag upon breakdown............................ 43 3o6 Increased breakdown voltage due to time lag..8o8,o o.8 43 4o1 Effect of initial gas concentration on self-ignition....o..... 48 4.2 Effect of wall temperature on self-ignition..........o..,... 48 4.3 Criterion for ignition,..o8...................00006000000000000 58 5ol Control panel and associated equipment....................... o 62 5,2 Constant volume bomb...................... o................... 63 5.3 Bomb and rear view of control panel...........................06 63 5.4 Electrode configurations O 0....... o...............09.. 6 0. 65 5.5 Manifold system.......0............................o........ 67 5 6 Spark control circuit Oo oo........... oooo.......o o.o o 70 5.7 Waveshape at point a.............. o......... o......... o o 73 vi

LIST OF ILLUSTRATIONS (CONT'D) Figure Page 508 Bomb with photomultiplier detector installed and 35 mm camera attached........00...... 000........000, 0...... 77 5o9 Photomultiplier circuit..................................,. 77 5.10 Typical current and corresponding light trace..,...,.....,,. 80 5.11 Effect of ultraviolet light upon breakdown voltage......... 80 5.12 Illustration of spark time duration control................. 80 5.13 Photomultiplier calibration trace o....o.... o.............o o 80 7.1 Dependence of quench distance upon equivalence ratio for propane-air mixtures...................................... 88 7.2 Effect of peak power upon minimum ignition energy for 6 less than quench distance..o o o.......... o.......o........oo 89 703 Effect of peak power upon minimum ignition energy, for = 0.83....................oo o,....o.....o o.o............ o 90 704 Effect of peak power upon minimum ignition energy, for = 0.73........................................................ 91 7o5 Effect of peak power upon minimum ignition energy, for = 0.69,oo....... ooooooo ooo. oo ooooooo o. o o o o o o. o o. oo 92 706 Effect of rates of energy input upon minimum ignition energy for 6 less than quench distance............ooo....o...... 93 707 Effect of rates of energy input upon minimum ignition energy for 4 = 0o83 oooof r,,s...................o enryipt o m u i to 94 7~8 Effect of rates of energy input upon minimum ignition energy for 0 = 0o73.........oo o o oo oo......................,..o o 95 7~9 Effect of rates of energy input uppn minimum ignition energy for 0 = 0.69.....00.0.... 0....0.....000...00.00000000o o o6o 96 7o10 Minimum ignition energy vs equivalence ratio for propane-air mixtures using pure capacitance sparks....................o 97 7 11 Determination of v.,...........,.,...,...... o. o 100 7.12 Power vs time for two different sparks....0.................. 102 vii

LIST OF ILLUSTRATIONS (CONT'D) Figure Page 7.13 Spark discharge; 0.122 inch gap, type "B" electrodes....... 106 7,14 Spark discharge; 0.254 inch gap, type "B" electrodes...... 106 7o15 Spark discharge; 0.254 inch gap, type "C" electrodes...,... 106 Aol Pre-breakdown current model................................ 119 B.1 Equivalent coil circuit.,.....,..........,............o.... 125 B.2 Determination of coil capacitance........................ 128 C.1 Voltage divider circuito...o........... o............... 132 Co2 Voltage divider output incompletely compensated for distributed capacitance.................................... 132 D.1 Thermocouple calibration curve......o..o.o................ 134.1l Photographic record of spark voltage and current as a function of time........................o........ 158 Fo2 Computer flow diagram..........,,..,,.................. 159 F 3 Sample computer output.............................. 161 F o4 Sample power vs time computer plot. o................ l.0 162 viii

NOMENCLATURE a Concentration of reactant or fuel a Concentration of active particles or chain carriers ao Concentration of active particles at C1 = 0, page 56 ao Concentration of fuel at source a Initial concentration of particles inside sphere, page 57 ao Value of a as defined on page 54 A Area Ar Constant B Constant C Capacitance C2 Secondary equivalent lumped capacitance in Figure.ol c constant cV Molar specific heat at constant volume cP Molar specific heat at constant pressure cv Constant volume specific heat c Constant pressure specific heat p D Diffusion coefficient for active particles D Constant DL Constant Dp Constant d Quench distance Ea Arrhenius activation energy Ef Final energy in the kernel ix

EH Total energy in high rate spark component EL Total energy in low rate spark component ETMN Minimum ignition energy EMINL Minimum ignition energy for a line source EMINP Minimum ignition energy for a point source E Fraction of condenser energy existing as internal energy in the kernel F Constant F/A Fuel air ratio G Mass flow of mixture toward source g Geometric factor representing the percentage of photons that reach the cathode. H " Heat" h Planck's constant h Average convective heat transfer coefficient Io Steady state primary current i Current i Photocurrent o K Thermal conductivity Ku Thermal conductivity of unburned gases k Second order branching coefficient kb First order branching coefficient kL First order chain breaking coefficient L Inductance L1 Primary inductance MAIR Molecular weight of air MF Molecular weight of fuel x

m Mass Nf Mole fraction of fuel No Mole fraction of oxygen n Total number of moles n Order of a reaction P Pressure Pf Partial pressure of fuel P Total pressure P0 Initial pressure PHA Average rate of energy discharge during high rate component PLA Average rate of energy discharge during low rate component of the spark P Peak power qI Rate of heat generation by reaction qII Rate of heat loss R Resistance R Gas constant R Coil resistance in Figure lol 2 R2 Lead resistance in Figure lo R Effective gap resistance g r Radius of spherical flame volume rl Radius of initial self-propagating flame Su Velocity of unburned gases relative to flame front s Emperical constant Temperature xi

Ta An average temperature between Tb and Tu Tb Self-ignition temperature Tf Adiabatic flame temperature TSLNE Temperature of the line source TSpOINT Temperature of the point source To Initial temperature Tu Temperature of unburned mixture Tw Wall temperature Tb Self-ignition temperature t Time URp Heat of reaction per unit mass TURp Heat of reaction per molecule URp Heat of reaction per mole of fuel URp Heat generation per unit time by each active particle V Volume V Voltage VBR Breakdown voltage V2MAX Maximum secondary voltage Vf Final volume Vs Volume at shock separation v Velocity X Field strength X Reactance Xr Radial component of a space charge field xii

~a Townsend's first ionization coefficient ap Second Townsend coefficient Y Secondary ionization coefficient y Ratio of specific heats, cp/cv AtR Voltage rise time BR Atsp Total spark discharge duration b6 Gap length Fraction of photons which produce electrons at the cathode cathode capable of leaving the surface Ratio of actual air to stoichiometric air 9 Number of photons produced by an electron per centimeter of gap ut Average absorption coefficient of the gas molecule for photons v Constant v Light frequency p Gas density p Density of gas mixture at source pU Density of unburned gases a Constant 1 a2 Constant Equivalence ratio Work function C Reaction rate eI Rate of initiation of chain carriers due to ignition source,2 Constant xiii

I. INTRODUCTION A. Purpose Improved thermal efficiency is realized by increasing the compression ratios of spark ignited internal combustion engines. This involves the ignition and burning of lean fuel-air mixtures. The purpose of this investigation was to determine the significance of the rate of energy input by the spark upon ignition of lean propane-air mixtures. The purpose was to conduct a study in which lean, quiescent propane-air mixtures were ignited in a constant volume bomb under fixed initial conditionso The total energy and the rate of energy transfer to the mixture by each spark were controlled. B. General initiation of a self-propagating reaction in the combustible mixture of an internal combustion engine is most frequently performed by an electric discharge called a sparko The quantity of energy and time of discharge can be controlled with relative ease as opposed to heated surfaces and flames. Two basic types of electrical discharges have been used to initiate combustion: capacitance and inductance sparkso The characteristic difference between these discharges, is the rate at which the energy is released in the electrode gap. Pure capacitance sparks are produced by the discharge of capacitors which are charged to the breakdown voltage of the electrode gap. Pure capacitance implies that the circuit resistance and inductance -1

-2are negligible. The time duration of such a spark is in many cases less than one microsecond for low energies and more than 100 microseconds if the energy is high. The latter long duration is due to non-negligible circuit resistances. The current can reach peak values as high as 300 amps with this type of spark. Therefore, the resultant rate of energy input is high during this type discharge. Inductance sparks are generated by high impedance coils when the circuit is interrupted by opening contacts. The current is low, (i.e., less than 200 milliamps) but the time duration is as long as 2 or 3 milliseconds. Hence, the rate of energy input is low compared to capacitance sparks. The conventional ignition system produces a spark which consists of two components. The first part of the spark is stored in the interelectrode capacitance of the coil in addition to any other secondary lead capacitance. Figure 1.1 illustrates a conventional ignition system circuit for a one cylinder engineo C2 represents the total equivalent lumped capacitance of the secondary circuit and the energy stored in the capacitance at breakdown 2 is equal to 1/2 C2VBR, where VBR is the voltage at the instant spark breakdown occurs, as shown in Figure 1.2. The time for the voltage to rise to the breakdown value for the given electrode spacing and gas is represented as AtBR, or voltage risetime. During the buildup of voltage to the breakdown value, the current in the secondary circuit follows the middle curve which shows current versus time in Figure 1.2 The current during AtBR is required to charge the secondary capacitance, C2 and does not represent current flow in the electrode gap.

-3RI RR2 R CI.1! CI L L -Li -1T ELECTRODES Figure l.1. Conventional Ignition Circuit. 0 _ _ __ _ I I I0 I 0 I 0 I 1 0 I I 0 Time in a Co nventinal Ignition Ci Spark. PP i V TIME iI I - - 1 TIME I I — A~ ~t —1 ------ a TIME Time in a Conventional Ignition Spark.

-4After breakdown, the current in the electrode gap rises to a maximum in 2 to 8 microseconds while the voltage drops off from a to b as shown in Figure 1o2. The current then falls to point c as the voltage levels off between b and c o This initial current pulse is due to the discharge of the secondary capacitance and is limited by the secondary resistance R2 and the effective gap resistance. Actually, the capacitive component is not a pure capacitive discharge. It is a spark component which discharges energy at a high rate. The instantaneous rate of energy input, (i.e., power) is merely the product of voltage and current as illustrated in Figure 1.2. The area under the power curve between a and c is equal to energy in the high rate of capacitive discharge of the spark and is designated, EH Before the high rate energy is completely dissipated, the energy stored in the ignition coil inductance begins to contribute energy but at a low rate. The peak current during the high rate discharge can be as high as 200 to 300 amps, but the maximum current during the low rate component is usually less than 200 milliamps. The area under the power curve of Figure 1.2, between c and d is equal to the energy in the low rate component and is designated, EL o This low rate component is commonly referred to as the inductive component since the energy dissipated during this time was stored in the inductance of the coil. However, the energy EL must be discharged through the coil resistance R~ and the lead resistance R' in addition to the electrode gap resistance.'Thus, it is not a pure inductance sparko The combustible gas

-5located between the electrodes sees only the energy transmitted to it and the rate at which it is received. From this viewpoint, "high rate component" and "low rate component" will be used to refer to the respective components of the spark discharge, Immediately following the high rate component, the current trace shows a unidirectional oscillationo This oscillation is due to coupling between the secondary and primary circuit after the switch S1 opens The total spark duration is designated as Atsp in Figure 1.2. AtR refers to the total time between opening of switch S1 and spark breakdown. The sine wave oscillation of the voltage and current which occurs after the spark is extinguished, is due to the secondary "tank" circuit decay. Control of the rate of energy input to the fuel-air mixture was carried out by adjustment of the secondary resistance and capacitance, The total spark energy was adjusted by control of the total time duration At3p of the electrical discharge, for a given peak rate of energy input.

II. SURVEY OF PREVIOUS WORK A few of the early investigators noted that the electrodes seemed to affect the minimum energy required for ignition. Later it was shown that for a fixed set of initial conditions (i.e., fuel, fuelair ratio, initial pressure and temperature, electrode shape and material), the energy required for ignition decreases to a minimum as the electrode spacing is increased. The electrode gap at which the minimum is reached is defined as the "quench distance". Many early investigations were performed with electrode spacings smaller than the quench distance. Those works which were performed at electrode spacings smaller than the quench distance will be designated as such in the remainder of this section wherever it is significant to the conclusions drawn One of the first methods used todetermine the energy in a (1,2) spark was the calorimeter method. Silsbee, Loeb and Fonseca(l presented results of direct measurement of the total energy supplied to the spark gap by different types of ignition systems operating at various frequencies in an early attempt to obtain design information for new ignition systems. While this work provided information about the energy in the spark, it did not describe the structure of the individual sparks. Wheeler(3) examined the spark structure using a smear camera. He found that the spark consisted of an initial bright spark followed by others of less intensity or by a striated luminous zone which indicated -6

-7an oscillating unidirectional discharge of current. He established that the first bright spark was what is today referred to as the capacitive component and the second spark is commonly referred to as the inductive component. His investigations also showed that the primary current required for ignition decreased to a minimum as the fuel-air ratio increased from a lean mixture to stoichiometrico Further increases in fuel-air ratio required a corresponding increase (4) in primary current for ignition. Wheeler continued his work and found that the above mentioned relation for primary current versus fuel-air ratio was the characteristic shape for most hydrocarbons. That is, the primary current required for ignition is high for lean mixtures, decreases to minimum and increases again as the fuel-air ratio is increased. Further, he noted some cooling effect of the electrodes on the initial combustion processo Taylor-Jones ) et al., presented mathematical solutions of the heat conduction equation starting from either point of spherical sources and applied them to the spark ignition problem. They showed that if the source of ignition is solely regarded as a source of heat, the effectiveness of a given quantity of energy depends upon the way the energy is transferred to a given volume of mixture. They concluded that an instantaneous point source of ignition was the best. Therefore, the capacitive component of the spark would prove to be the most effective, since the rate of energy input is highest in this component. Further work by Taylor-Jones(6) et al. showed that a capacitive spark when discharged between two spherical electrodes passed between the

-8electrodes at the shortest distance. On the otherhand, an inductive spark moved out away from the narrowest gap distance. They surmised that the capacitive spark was in a position relative to the electrodes such that it could transfer heat most readily to the electrodes and less heat to the gas than would the inductive spark. However, they performed calorimeter tests with small spherical carbon electrodes and found that the capacitive sparks transferred more heat to the gas than the inductive sparks transferred with the same primary current. Consequently, they felt that the thermal theory was still valid. (7) Coward and Meiter( determined the volumes of methane-air mixtures that were burned by sparks which were not sufficient to causea self propagating reaction to occur in the mixture. These volumes were roughly proportional to the square of the primary current of the induction coilo In other words, the columns were approximately proportional to the spark energy. They concluded that the spark acted "mainly perhaps entirely as a source for thermal energy in igniting a gas mixture" (8) (9) Finch and Thompson() and also Bradford and Finch on the otherhand, concluded from their works that "ignition is determined by the setting up of a sufficient concentration of suitably activated molecules"o They determined minimum igniting pressures for sparks of different energies and of different current oscillatory frequencies. Their results showed that the total energy and the rate of energy dissipation were unimportant, but the oscillatory frequency of the current was an important factor, In an attempt to shed some light on the reaction mechanism of hydrogen and oxygen, Thompson(lO) and then Linnett, Rayor and Frost(ll)

-9conducted experiments on spark ignition of mixtures of hydrogen and oxygen. The combustible gases were ignited inside 2 to 5 centimeter glass tubes by sparks formed from induction coils, Over a series of runs the spark energy and gap were held constant and the least igniting pressures were determinedfor various mixture ratios. Both investigators found that the partial pressure of the 2H2 + 02 mixture in the gas could be reduced when moderate amounts of inert gas were added at the least igniting pressureo They interpreted this as meaning that the diffusion of chain carriers from the spark volume was inhibited. Linnett(ll) et al., found that larger additions of inert gases cause an increase in the partial pressure of 2H2 + 02 at the least igniting pres(12) sure. Lewis and von Elbe(l2) explain this latter effect as due to thermal factors. The flame temperature is decreased and the thermal dif(13) fusivity of the mixture changes. Linnett and Nutbourne found that the partial pressure of the H2 + 3N20 mixture in the least-igniting pressure mixture was always increased by the addition of inert gases. They concluded that the absence of a branched chain in the H2 + 3N20 reaction was the reason that their results differed from those obtained (12) with the 2H2 + 02 mixtures. Lewis and von Elbe indicate that the results of the preceding experiments are only qualitative and cannot be compared with experiments on minimum ignition energies, because the electrode gap was arbitrarily chosen to be constant. In minimum ignition energy determination, the electrode spacing must be carefully determined to be larger than the flame quenching distance for each fuel-air ratio and initial condition.

-10Morgan ) in his works indicated that the "incendivity" of a capacity spark is greater than that of an inductance spark when each spark dissipates the same energy. He operated a conventional type ignition system and added a variable capacitor to the secondary, in parallel with the electrodes. The primary current was fixed, and when the secondary capacitor was set at zero, the spark (capacitive plus inductive) did not cause ignitiono However, the spark caused ignition if the secondary capacitance was increased while using the same primary currento Therefore, Morgan concluded that the capacitive component was the most effective in igniting a mixtureo It must be pointed out that the experiment was carried out with non-pointed electrodes and was presumably confined to gaps which were smaller than the quench distance, and therefore the results do not establish an absolute trend. Contradictory work was carried out by Bradford, Finch and Prior(l6) In this case, the ignition occurred in a 250 cc glass explosion vesselo The secondary voltage and current in the spark were recorded using a cathode ray-oscillographo A mixture of CO-02-H2 was ignited at a pressure of 100 mm of mercury with a primary current of 3 amps and the spark breakdown voltage was 4000 volts. The total duration for the spark was 2o81 milliseconds. Then, for the same conditions only the total time duration of the spark was reduced to 2.6 milliseconds, the mixture would not ignite until the pressure was increased to 109 mm of mercury. The following table illustrates further the results obtained.

-11TABLE I BRANDFORD, FINCH AND PRIOR(16) DATA Time Duration Energy Pressure at Milliseconds Millijoules Ignition mmHrg. abs. 0079 902 130 0.39 5.2 154 0.20 2.4 195 0.02 0.15 235 Bradford et al., concluded from these results that the inductive component, if not most important, at least was necessary for ignition under the conditions of their experiment. The conclusions presented by Bradford, et alo would seem valid since it was stipulated that the breakdown voltage and thus the capacity component was maintained constant throughout all of the runs. However, Lewis and von Elbe(l2) indicate that they obtained minimum ignition energies for a CH4 + 202 mixture with a pure capacitive spark of 0.04 mjoules at a pressure of 252 mm of mercury. But, Bradford, et al., failed to ignite the same mixture at 263 mm Hg. with a 5,0 mjoule spark which consisted of only the capacitive component. Gap lengths were not reported by Bradford, Finch and Prior. Consequently, the explanation of the extremely large energy value cannot be directly attributed to the quench effect, but indicates that the results are not meaningful.

-12(17) Finch and Sutton verified the theory of the ignition coil through the use of a cathode-ray oscillograph. They confirmed experimentally that the closing of the primary circuit, (ioe., turning the primary current on), during the life of the inductive component stops the secondary discharge. These results enabled them to design a system whereby the duration of the spark could be controlled by control of the time that the primary circuit was open, (i.e., primary current off). Basically, the control of the spark duration controls only the inductive component. The regulation of the capacity component was attained by limiting the secondary current. This limitation was accomplished through the use of a diode which was run under conditions such that the peak current was lower than the normal high capacity component current. Finch and Mole investigated the influence of the inductive component of the spark on ignition in an engine. The initial pressures and temperatures encountered in the engine were substantially (16) higher than those used in the works of Bradford, Finch and Prior Their results indicated that the output, efficiency and speed of the engine remained unchanged when the inductive component time duration was reduced from 2 to 0.5 milliseconds. Therefore, they concluded that in the engine the capacity component is sufficient for ignition. (20,21,22) Lewis, von Elbe, Blanc, and Guest(2021 ) were the first investigators to perform really meaningful experiments to obtain minimum ignition energies as a function of fuel-air ratio. All of their work was performed at an initial pressure of one atmosphere or lower at atmospheric temperature. Their electrical circuit is illustrated in Figure 2.1o The variable air capacitors were charged very

-13llOV.B POWER SUPPLY AIR CONDENSERS VOLTMETER: Tt T BOMB a ADJUSTABLE FLANGED ROTARY ELECTRODES CHARGER Figure 2.1. Ignition System Used by Lewis and von Elbe. {U~ ~ C ~FREE ELECTRODES FREE FLANGED FLANGED 0 dq s b d Emin. ELECTRODE SPACING, (GAP) Figure 2.2. Minimum Ignition Energies for Various Electrode Configurations.

-14slowly by the rotary charger until the breakdown voltage of the spark gap was reached. The voltage at which the spark occurred was observed and the voltage on the capacitors after the spark was recorded. The net energy change in the capacitors was assumed -to be the energy dissipated by the spark. It was felt that resistance losses were negligible since the circuit resistance was less than Ool ohm, and as compared to an effective gap resistance assumed to be much larger. A test of this point was made by the addition of up to 30 ohms of resistance to the circuit before any effect was noted in the energy required. Further tests were run to determine the effect of inductance on the ignition process. The resistance in the previous test was replaced by a helix of heavy wire. This helix did not show any effect of moderate changes of inductance on the minimum ignition energy, It appears that the conclusions that may be drawn from the result are either that the added inductance did not change the spark characteristic or the inductively changed capacitive spark is as effective in the process of ignition as the pure capacitive sparko One of the most important contributions made by Lewis and (14,20,21 23) von Elbe et alo (142 l23) is the determination of quench distances for spark ignition studies. The quench distance is illustrated in Figure 2o2. They found that where free electrodes, as shown in Figure 2,2, are used, the nimum ignition energy goes hrouh a minimum as the a electrode spacing is increased. The quench effect shown by the curve ab for free electrodes was attributed to chain breaking of the initial reactions

-15at the electrodes as well as heat transfer to the electrodes. Proof of this is shown by the curve cb for flanged electrodes. In this case, glass flanges were added to the free electrodeso The effect was to increase the required ignition energy to infinity for electrode spacings less than d, which is defined as the "quench distance". Thus, it is clear that measurement of minimum ignition energies at electrode spacings, 6, less than dq with anything other than flanged electrodes would leadto erroneous results. Therefore, minimum ignition energy in the true sense refers only to the value EM illustrated in Figure 2.2. The flat section bd of Figure 2.2 suggests that most of the energy is transferred to the gas over a small fraction of the gap length. Rise of the curve, d to e is attributable to distribution of the spark energy over a volume which exceeds a critical volume. Lewis and von Elbe (l2 20 2l23) also found that the minimum ignition energy decreases as the initial pressure is increased from 0.2 atmospheres up to 1.00 atmosphere. In addition, the minimum ignition energy follows a U shaped curve as a function of the percent fuel in air as shown in Figure 2.3. This is a characteristic shape for all hydrocarbons. In addition, the quench distance follows an analogous U shape as a function of the percent fuel in air as illustrated in Figure 2.4. (24) Morris() discusses some general relationships applicable to the results obtained by Lewis and von Elbe(20'25) He found that the quench distances are related tothe pressure according to a general equation

-16EMIN. PERCENT FUEL IN AIR Figure 2.3. Minimum Ignition Energy EMIN as a Function of the Percent Fuel in Air. dq PERCENT FUEL IN AIR Figure 2.4. Quench Distance dq as a Function of the Percent Fuel in Air.

-17where is d = c + - (2.1) q P always positive, varying from 0.005 to 0.18 centimeters. Composition changes affect the value of s in an irregular manner. In addition, the minimum ignition energies follow a general relationship = e (2.2) MIN p where e is a constant for the mixture. (26) Belles and Swett ) point out that the quench distances obtained in ignition energy experiments are the same as those obtained in flame quench experiments. In both cases, the quench effect is due to heat transfer from the flame or diffusion of active particles from the flame to the solid surfaces and the surrounding gas. The flame quench experiments involve flashback of flames burning at the port of rectangular burners. The spacing between the long walls of the rectangular burner are varied for a series of flashbacks until the spacing is determined for which the flame will not propagate back into the burner when the fuel-air flow is suddenly stopped~ Investigations (27) (28) carried out by Friedman et al. and Harris et al. both show that the quench distance follows the same "U shape" function as indicated by Figure 2~4. In addition, the quenching distance decreases with increases in temperature and pressure. A comprehensive discussion of the pure capacitive spark is included in the report by Belles and Swetto They indicate that the time duration of the spark can be less than 0,01 microseconds for

-18low energies and as long as 100 microseconds for energies greater than 1 joule. The time duration of the spark is dependent upon the inherent properties of the circuit. Long duration capacitive sparks are obtained by the addition of series resistance to the circuit and when enough resistance is added, the discharge is very similar to that of a pure inductance sparko An extensive experiment on capacitive spark discharges was (29) carried out by Rose and Priede(9) Since no suitable instrument was available at the time, the authors built a high speed oscillograph that had a resolution of one nanosecond.* They recorded the voltage across the spark and the current through the spark as a function of time. Voltage was measured by means of a capacitance potential divider and the current was measured by determination of the voltage across a resistor on the ground side of the circuito For essentially pure capacitive sparks,** the discharge was oscillatory. By plotting the successive positive and negative peaks on semi-log paper versus time, they calculated the resistance of the spark gap from the slope of the curve They found that the gap resistance increased markedly as the system capacitance was decreasedo Also, reduction of the capacitance reduced the maximum value of the current. It was quite common to record peak currents as high as 300 ampso In their work, they found that the gap resistance followed the following emperical equation. 1 nanosecond = 10-3 microseconds System used was very similar to that of Lewis and von Elbe.

-193.4 x 10-2z R = 0.4 e (2.3) where z = )1/2 +R (2.4) When a large series resistance of ilK ohms was introduced, the current was critically damped and unidirectional or aperiodic with a peak value of 0.4 amps. (30) Rose and Priede extended their investigations to determine the effect that electrode configuration had upon the ignition energy and breakdown voltage. They found that the breakdown voltage of a given gap increased when the radius of curvature of the electrode tips were increased. However, addition of a large radius insulting material around a small radius electrode did not change the breakdown voltage. In addition, they found the electrode material affected the energy required for ignition of a given mixture. The minimum energy decreased in the order platinum, aluminum, silver and cadmium. Also, the minimum ignition energy decreases with an increase in the value of the resistance in series with the capacitance spark. Reduction of the minimum ignition energy occurs with reduction of electrode size and reduction of the energy required for a critical area of the electrode material to reach its boiling point. Any decrease in discharge time resulted in a decrease ofthe minimum ignition energy. Clyde Swett(31 32333435) studied the ignition of flowing gases by long duration capacitive sparks, (i.e., up to 24,400 microseconds duration). In these experiments the long duration capacitive

-20discharge was obtained by addition of resistance to the circuit. His results showed that decreasing the spark duration from approximately 25000 microseconds to 125 microseconds decreased the amount of energy required for ignition. In addition, he reported that the energy required for ignition with a spark of short duration (i.e., 1 microsecond) was considerably larger than that required by most of the long duration sparks, (i.e., 600 microseconds). The energy required for ignition decreased and then increased again as the gap was increased and for small diameter electrodes the minimum was not sharply defined. Small diameter electrodes required less energy than large diameter electrodes and the required energy for ignition increased with increased turbulence of the gas stream. In a capacitive spark the energy is transmitted to the gas almost instantaneously. Then the heat is transferred from the gas to (36) the electrodes. Lewis and von Elbe et alo conducted some experiments to determine the heat generation by capacitive sparks and the heat transferred to the electrodes. They mounted electrodes in a small vessel containing a monoatomic gas and recorded the pressure change at constant,volume or the volume change at constant pressure during the ignition process. From these records the spark generated "heat", H, that is present in the gas at any instant was determined as follows: n H = cv (T - To) dn (2.5) For constant pressure, Cp is used. Then from the ideal gas law:

-21n VAP = R f (T - To) dn for constant volume (2.6) 0 and n PAV = R I (T - To) dn for constant pressure (2~7) o where n = total number of moles T = temperature in volume element To = initial temperature c = molar apec heat at constant volume Then if c and cp are constant, the spark generated "heat" is represented by the following for monoatomic gases. H = 1o5 VAP for constant volume (2.8) H = 2o5 PAV for constant pressure (2.9) By plotting the ratio of "heat" to energy supplied by the capacitor as percent, versus the square root of the product of thermal diffusivity and time, the curves were extrapolated back to zero time, in xenon, argon and helium 95 percent of the spark energy is present in the form of, "heat"o (37,38) Olsen et alo studied the capacitive spark using a schlieren technique and used ideal gas assumptions in their solution. They developed an equation to describe the fractional loss of energy by the kernel due to expansion after separation of the initial shock in pure argon. The expression is as follows:

-22Es - Ef (Vf/VS)7 1 - 1 (2.10) (vf/Vs) where Vf = final volume Vs = volume at shock separation Es = fraction of cond. energy existing as internal energy in the kernel Ef = final energy in the kernel Thus the energy in the kernel decreases due to the work that must oe done on the surrounding gas during expansion and the remaining energy in the kernel is the energy available for ignition. Further, their results showed that the energy in the capacitance sparks followed the rules: For argon: 1/2 C V BRo V 3 (2.11) for oxygen: 1/2 C VBR Vf 57 (2.12) where VBR = voltage at breakdown C = capacitance (39) Olsen and associates(39) also investigated the initial flame kernel growth of hydrogen-air and propane-air mixtures using the same equipment as in the preceding discussion. Their observations enabled them to determine the steady state flame velocity for hydrogen and

-23propane mixtures. In addition, they found, that for sparks which had less than the minimum amount of energy, the flame velocity approached zero in less than 200 microseconds while sparks having more than the minimum energy caused the:-'flame velocity to approach the steady state value in approximately 900 microseconds. (40) Interesting results were obtained by Arnold and Sherburne in their investigation of the initial flame kernel growth following ignition by a sparko The observations were made with a schlieren system by photographing the kernel at small time increments following the spark. They found that after a certain fixed time for a given fuel-air ratio that if the kernel had not reached a "critical" radius, the flame would extinguish as it progressed further downstream. Also, the critical radius decreases to a minimum andthen rises again (i.e., U shaped curve) as the equivalence ratio is increased. The minima of these curves agree (20,21) very well with those published by Lewis and von Elbe for minimum ignition energy as a function of equivalence. ratio. Schlieren photographs were used by Kumagai et al. to observe the effect that ultrasonic waves have upon spark ignition and flame propagation. A flame front assumes a ragged form when exposed to ultrasonic waves. It is felt'that this is due to acceleration of the flame. The major effect was considered to be promotion of heat transfer and diffusion of active particles from the burned to the unburned mixture. The waves had no observable:effect on the spark ignition. This suggested that the agitating action of the spark was similar to that resulting from ultrasonic waves.

-24The effect of the molecular structure of hydrocarbons on spark (42) ignition was examined by Calcote, Gregory, Barnett and Gilmero They used a constant volume bomb for their studies and ignited the mixture with pure capacitance sparks. The results indicated that knowledge of the structural configuration of the fuel could enable one to predict the effect of a change in fuel on minimum ignition energy, (43) Metzler measured the minimum spark ignition energies of 12 pure hydrocarbon fuels at reduced pressure and extrapolated the data back to one atmosphere. He ignited the fuels with long duration capacitance sparks. All paraffins and cycloparaffins including the branched members had essentially the same minimum spark-ignition energies. The equivalence ratio at which the minimum spark ignition energy occurred was established by the length of the carbon chain and was relatively insensitive to pressure, He found that the minimum energy was approximately proportional to the inverse of the pressure to the 1.82 power for most fuels and the flame velocity was inversely proportional to the minimum energy to the 0.8 powero And finally, the minimum ignition energy was related to the quenching distance by the following expressiono EMN = 6.36 d179 (2.13) (44) King and Calcote reported the effect of initial temperature on the minimum spark ignition energy using capacitive sparkso They found that the log of the energy was proportional to the inverse of temperature above -30~C as the initial temperature was increased. Also, the minimum ignition energy shifted to leaner mixtureso

-25The effects of radiation at the point of ignition of constant (45) volume combustion were studied by Harnsberger and Van Wylen. Results indicated that the breakdown voltages of the spark gap were reduced by approximately one-half by the presence of a 1.6 curie radiation source at pressures below 60 inches of mercury absolute. Also, the ignition energy requirement can be reduced 30 to 50 percent by the presence of radiation at the ignition point if the electrode spacing is set at greater than the quench distance. Further investigation of the influence of radiation upon constant volume combustion was carried out by Souka. (46 The results indicated that the so-called ignition delay period was unaffected by irradiating the spark gap with alpha radiation. However, a decrease in delay period occurred in the case of ethane. No discernable change was noted in the rate of pressure rise when the spark gap was irradiated. Ethane required less energy for ignition when the electrodes were irradiated but ethylene required more energy for ignition for wide gaps and no change was noted for narrow gaps when the electrodes were irradiated. Furthermore, the energy necessary for ignition of acetylene was increased by irradiation for narrow gaps and decreased for wider gaps. (47) J. S. Clarke indicates that probably the most important property of the spark in an engine is the electrode gap. The electrode gap which is required for stable combustion at idle is larger than that required for combustion at half-load or higher. This is due to the quenching effect that the electrodes cause at the lower pressure

-26idle condition. He also points out that the way in which the spark energy is released in the gap has little effect on the ignition process in an engine. This manner of energy release refers to single or multiple sparks (15) According to Morgan the capacitive or high rate of energy component of the conventional ignition spark is the only part of the spark necessary for ignition. From a design standpoint, it would prove most fruitful to have a large secondary capacitance in the ignition coilo Miller(48) discusses in detail one limitation on the maximum secondary capacitance. For a given primary current, the peak secondary voltage is limited by secondary capacitance according to the following equation, if the energy losses in the secondary circuit can be neglected. VMAx Io L1/L2 (2.14) where L1 = primary inductance Io = primary current before switch is opened, This equation assumes that the primary current can be instantaneously stopped when the switch is opened. Thus, if C2 is large enough, the maximum secondary voltage can be limited to a value smaller than the gap breakdown voltage VBR and no spark will form. A second design problem involves the erosion of the spark gap. The high rate of energy component of the spark is of a short duration, (49) but the current can be very high. Obert indicates that this high

-27current 20 to 200 amps is responsible for a large portion of the spark plug electrode erosion. The erosion can be reduced by reduction of the secondary capacitance and/or addition of a large resistance in series with the electrode gap. (5o,5l,52,53) The Ethyl Corporation conducted a series of studies during the Second World War to evaluate the dependability and durability of aircraft ignition systems with emphasis on electrode wear. (50) In one study,5 various high tension spark generators were evaluated with regard to electrode wear. The magneto systems which provided high energy release caused considerably more erosion than the conventional battery coil systems having equivalent secondary capacitances. Therefore, the conclusion was drawn that the high energy sparks caused more erosion. An auxiliary gap added to a system caused slightly more electrode wear. Addition of 1000 ohms of series resistance to the secondary circuit reduced the electrode wear by 35 to 40 percent. (51) The report(5l) indicates that the high intensity discharge of the secondary capacitance (52) is the major cause of spark plug erosiono Further tests indicated that an increase in the secondary capacitance from 50 picofarads (i.e., 1012 farads) to 95 picofarads increased the electrode wear by 100 percent. (53) Another study indicated that the electrode material affected the spark breakdown voltage in the engine. As the electrical resistivity of the center electrode materials increased,the breakdown voltages decreased, The breakdown voltages decreased as the thermal conductivity of the center electrodes decreasedo Allsop and Guenault(54) discuss the effect that inductance, voltage, frequency of the voltage, rate of separation of electrodes and

-28electrode material have upon ignition of combustible gas mixtures in relation to safety devices. They indicate some of the devides that may be used for prevention of explosions and describe their applicability. (55) Cipriani and Middleton discuss the thermal, ionization and chain theories of spark ignition and their application to automotive ignition systems. They describe how an understanding of the theories lead to larger spark plug gaps so that leaner mixtures could be burned in the spark ignited engineso

III. THE MECHANISM OF SPARK BREAKDOWN A, General (56) The electrical spark is defined by Loeb and Meek as follows: "The spark is an unstable, irreversible and transient phenomenon sometimes marking the transition from one more or less stable condition of current between electrodes in a gas to another more stable one under imposed conditions." In this chapter a brief description is made of the fundamentals of the initial more or less stable conditions leading up to spark breakdown. In particular, a description of the pre-breakdown currents is essential for a thorough understanding of the spark breakdown mechanism. B. Photoelectric Emission Emission of an electron from the cathode in a spark system is most commonly due to the impingment of a light photon of sufficient energy upon the cathode surface. An electron can leave the surface of the cathode if: hv > 0 (31) Electrons close to the surface which have their maximum energy directed normal to the surface leave the cathode at a substantial velocity when the quantum of light energy, hv is larger than the work function, w, These emitted electrons have a velocity given by Einstein's Photoelectric Law(57) 1/2 mv2 = hv - 0 (32) -29

-30Actually, most of the electrons possess less energy than represented by Equation (3.2) because most of the electrons which escape have less than the maximum velocity perpendicular to the surface and also lose energy due to internal encounters other than the gross work function,w so that: () 1/2 mv < hv - (3) The intensity of the incident light which is related to the number of photons does not affect the energy distribution of the escaping electrons, but the number of electrons ejected from the cathode per unit time is (59) directly proportional to the light intensityo C. Townsend Theory Consider two electrodes connected to a variable D.C, voltage supply with an electrometer in series to measure the low pre-breakdown currents as shown in Figure 351. Current is initiated when an electron leaves the cathode and progresses to the anode under the influence of the applied voltageo Each electron emitted by the cathode undergoes an acceleration due to the externally applied voltage. During acceleration, the electron soon collides with a gas molecule. In most cases, this collision is elastic if the collision occurs in the region near the cathode surface where the kinetic energy of the electron is still low. Since the mean free path of an electron between collisions in air at one atmosphere is 5 (60) electron will make x 10 collisions/cm of 5 x 10-5 centimeters, an electron will make 3 x 10 collisions/cm of

-31/ \ CATHODE( ANODE \/ ELECTROMETER Figure 3.1. Circuit for Measurement of Pre-Breakdown Currents. z d^ a(a C VOLTAGE VBR Figure 3.2. Pre-Breakdown Current-Voltage Characteristics.

-32electrode gap, and the net drift velocity just after ejection from (60) the cathode is about one tenth of the random velocity. As a result, most of the emitted electrons suffer collisions which return them to the cathode if the applied voltage is lowo This is called back diffusion." Higher voltages result in electron diffusion to the anode. Electrons which reach the anode are measured as current by the external electrometer shown in Figure 35.1 The current is shown as a function of the applied voltage in Figure 3,2, (6162) At very low voltages, most of the emitted electrons back diffuse with a resultant low current as shown at the left of region a, of Figure 352, An increase in the field strength causes more and more electrons to diffuse to the anode until the saturation current is reached which is equal to the photocurrent, io o The magnitude of the saturation current is dependent upon the light intensity of impinging light. Further increase of the external voltage causes a rise in current above the photocurrent, io 0 Therefore, electrons are being generated in the gas by some other mechanism. The phenomenon is due to inelastic collisions between the electrons and gas molecules, with resultant ionization of some molecules of the gaso This increased current, region b of Figure 3.2 is expressed by the following equation which was originated by Townsend, (59,6 3) i = i ea6 * (3.4) Derivation of this equation my be found in Appendix A Derivation of this equation may be found in Appendix A.

-33a is the first Townsend coefficient and is defined as the number of electrons produced by ionizing collision in the path of a single electron traveling a distance of one centimeter.(59'60,61,63) Equation (3.4) indicates that ee - 1 new electrons reach the anode for each electron emitted by the cathode. This phenomenon is known as an "electron avalanche" and as "gas amplification" in electron tube parlance. Since the photo-current is increased tremendously by the increased field strength, the ionization is sometimes called, field-intensified ionization. Townsend showed experimentally that the first ionization coefficient a, is proportional to the gas density p and the field strength X (59) in the following manner. a = pf (I) (3.5) In addition, his experiments indicated that if the external voltage is raised beyond that of region b, the current increased still more than that represented by Equation (3.4). He concluded that a secondary mechanism was the cause. Townsend felt that the secondary electrons were liberated by positive ion bombardment of the cathode and by collision between positive ions and gas molecules, For secondary emission of electrons due to ionization of gas molecules by positive ions, Townsend developed the following equation.(63) (a-n) e (a-) i = *o (*f5(3.6) p is the second Townsend coefficient and it is related to the gas density and field strength according to the equation Derivation of this equation is left for Appendix A.

-34= pg () (3.7) Under most conditions, the probability is very small that a positive ion will obtain enough energy to ionize a gas molecule 64) An equation similar in form to Equation (3.6) can be derived from secondary emission at the cathode due to positive ion bombardment. (63) aKS i = i - (e) * (3.8) 1 - 7 (ea - 1) if a is substituted for (a-5) and 7 for 5/(Ca-) in Equation (3.6), the Equations (3.6) and (3~8) would be identical More recently, physicists have stated that a third possible process of secondary emission of electrons from the cathode is due to photon impact from excited gas atoms and is expressed as follows: ( C- ) e * ~ = i ~" ~Fry LI.T& (3~9) o (af-) - Og e (o) where e is the number of photons produced by an electron per centimeter of gap, i is the fraction of the photons which produce electrons at the cathode capable of leaving the surface, g is the geometric factor which represents the percentage of photons that reach the cathode, pJ is the average absorption coefficient of the gas molecules for photons. Equation (3~9) resembles Equations (3~6) and (3.8) and since the net effect of the processes are indistinguishable, it is common to express the secondary ionization effects by the coefficient, y. Probably both positive ions and photons cause emission at the cathode. Therefore, Derivation of this equation is left for Appendix Ao

-35Equation (3.8) is useful for determination of breakdown voltages under steady state conditions. Wheatcroft(59) defines electrical breakdown as the point of transition of the gas from insulator to conductor. This causes an abrupt increase in the current. In Figure 3.5 di/dV = o at the end of region c where breakdown occurs. In reality the current is finite and limited to a maximum value by the external circuit parameters. From Equation (3.8) the Townsend criterion for spark breakdown at the end of region c is: 7(e - 1) = 1 (3.10) so that i approaches infinityo 7, is the number of secondary electrons generated per primary electron causing an avalanche in the gap. Since (eab - 1), represents the number of electrons formed in the gap by each electron ejected from the cathode, the Townsend criterion of breakdown guarantees that the discharge is self-sustainingo Equation (3~8) indicates that breakdown is dependent upon a, 7, the type of gas and upon the cathode material (i.e., indirectly through y ), while a and y are both proportional to the ratio of field strength to gas density, Do Streamer Theory (66) (67) In 1940 Meek and Raether independently proposed the streamer theory of spark breakdown. This theory was developed to account for inadequacies in the Townsend mechanism at atmospheric pressure and

-36above. One major problem encountered with the Townsend theory was that time for formation of an avalanche at higher pressures is shorter than the corresponding time it would take for a positive ion to progress across the gap and cause secondary ionization at the cathode. Experimental observations of long sparks indicated a need for a new theory of spark breakdown. Consider an external voltage V applied across the electrodes of Figure 3.1, such that the field strength X is high enough to cause an avalanche which extends to a distance x, across the gap. Electrons travel on the order of 2 x 107 cm/sec as compared to 2 x 105 cm/sec (61) for positive ions. Therefore, the positive ions are essentially stationary with regard to the electrons, and the electrons leave a positive space charge behind as illustrated in Figure 3.3 This space charge distorts the field to which the electrons at the head of the avalanche are subjected. The distortion of the field is the greatest at the head of the avalanche where the ion density is the highest. The conical space charge due to this avalanche induces a radial component Xr to the overall field which enhances the external field X o The electrons at the head of the avalanche move into the anode when the avalanche has crossed the gapo Then the positive space charge is extremely high at the anode and decreased toward the cathode as shown in Figure 3)4ao Photons are produced in the region surrounding the avalanche by the recombination of ions and electrons and by decay of electronically excited molecules. It is assumed that these photons ionize the gas. The resultant electrons initiate secondary avalanches which are directed toward the anode base of

-37J- X - 1HEAD OF AVALANCHE /,f + + +- _ _+ -, CATHODE + + + + ANODE F r +. t. + + ++++ Figure 3.3. Electron Avalanche. \+v+ +-+-+_____ \+ -ri p/ ^V r3T + + A CATHODE B C Figure 3.4. Formation of a Streamer.

-38the main avalanche is the space-charge field Xr is of the same order as the external field X o The greatest number of seconsary avalanches feed electrons and positive ions into the head of the main avalancheo A conducting plasma is thus formed as shown in Figure 3,4b,and.this plasma effectively reduces the gap lengtho The positive ions behind the auxiliary avalanches intensify the field at the tip of the plasma. Therefore the plasma, which is called a streamer, moves toward the cathode at an accelerating rate until it reaches the cathode. Whereupon, the gap is bridged by a conducting plasma which is the initial spark channel through which the external circuit energy is discharged. The streamer breakdown criterion as stated by Devins and Sharbaugh (o) is: "When enough positive ion space charge is generated at the original avalanche head so as to cause a space charge field of the same order as the applied field, then the streamer breakdown will occur o In other words, breakdown occurs when Xr is equal to X If the applied external voltage is larger than the minimum breakdown voltage, the main avalanche may form a space charge field as large as the external field before the avalanche traverses the gapo Under these conditions mig-gap streamers have been observed to form. Devins and Sharbaugh(60) indicate that the Townsend mechanism of breakdown sets the low breakdown limit up to one atmosphere and at electrode spacings of one centimeter or lesso However, the streamer theory is applicable for high pressures and larger gaps. Also, the streamer theory seems applicable for overvoltaged gaps at one atmosphere.

-39E, Paschen's Law and Similitude In 1889 Paschen introduced an expression for electrical breakdown which has since been used very extensivelyo Paschen's Law states that in a uniform field the breakdown voltage of a gas depends only upon the product of the gas pressure and the electrode spacing. If temperature variations are considered, the breakdown voltage is dependent upon the product of gas density and spacingo The Townsend theory of spark breakdown established an explanation for the experimental observations of Pascheno The Townsend coefficient, y can be expressed by the following equation: ( = f(x) (3o11) Substitution of Equations (3.11) and (3~5) into Equation (3.10) the criterion for Townsend breakdown gives: f() epf(P) ] = 1 (3.12) Since the breakdown voltage VBR is equal to X6, Equation (3.12) is expressed as follows: vR) F (V \ lj =1 (3013) / L The variables of the equation are only VBR and ps, and explicit solution for VBR would be: VBR = (p) (3.14) which is a mathematical'expression of Paschen's Law.

-40Paschen Law is a particular case of the Similarity Principle. The Principle of Similarity states that if the voltage applied to two different sets of electrodes is the same, the current flow between the electrodes each set will be the same when the gas density is multiplied by a factor K and all linear dimensions are divided by K. Take two plane-parallel plates in a uniform field and increase the gas denisty by K and divide the linear dimensions by K P2 = Kpl and = (3.15) v-1 2 K = K X (3.16) Substituting Equations (3.15) and (3.16) into Equation (3.5) for the first Townsens coefficient and Equation (3.11) for the y coefficient yields: ao X2 f KX! \ al p2 P2,; \ KP9 P1 and 2 = Ol1 (3.17) or a = Kca, (3.18) Also, /x2 / 1 7 = fI - ft-, = 71 (3.19) \ Pl/ Finally, substitution of Equations (3.18) and (3.19) reduces Equation (308) to the following:

-41llK 82 i i....ealK62 (3.20) 2 ~ 1 - Y( e1K 2 -1 61 Recall that 62 is equal to - 1I5!1 i2 = i e (3,21) 10 - y ell- Therefore, the current i2 will be the same for both systems and consequently the breakdown voltage will be the sameo Equations (3o15) reduce to: p262 = P1 (3o22) which is an expression of Paschen's Lawo Fo Time Lags The preceding discussion of conditions before breakdown dealt exclusively with a steady state mechanismo In practice, the time duration of the voltage application plays an important part in determination of the pre-breakdown current and breakdown voltageo A spark does not occur immediately upon application of a sufficiently large voltage to the electrodes. Therefore, there is a time lag before breakdown, This time delay may be divided into two periods: (1) statistical time lag is the time required for one or more electrons emitted by the cathode to acquire a favorable position in the gap for avalanche formation, (2) formative time lag is the time required to develop an avalanche which leads to Townsend breakdown or streamer formation. If the duration of the voltage

-42pulse is less than the time lag (i.e., statistical plus formative) breakdown may not occur even though the peak voltage is equal to the steady state breakdown voltage. This can be illustrated by Figure 3.5 The voltage drops below VBR before the time lag has elapsed. Consider another voltage pulse that continues to rise at the same rate beyond VBR, Figure 3.6. In this case breakdown will occur after the time lag and at the higher voltage V Consequently, a fast rising BR n voltage will result in a higher breakdown voltage for the same conditions. (68) A recent paper by Eason(8) indicates that a faster voltage rise in an automotive ignition system improves the ability to fire fouled plugs. A disadvantage of such a system is the necessary increase in breakdown voltage due to the time lago Fortunately, the time lags encountered in the conventional automotive ignition systems are short enough so that no major difficulty has been noted concerning increased breakdown voltages with decreased risetimes. Examination of secondary voltage traces of an automotive spark indicates a wide variation in breakdown voltage from spark to spark. This variation appears to be statistical. It is just this variation which lead to the term statistical time lag. The statistical time lag may be controlled by control of the rate of electron emission by the cathode in the following ways. 1. Reduction in the work function of the cathode material, i.eo, choice of cathode material, see Section A. 2o Irradiation of the cathode with ultraviolet light of high intensity, see Section A. 3o Irradiation of the gas by radioactive materialso

-43VOR STEADY w STATE C ~ ITmE~~~ ITIME LAO Figure 3.5. Effect of Time Lag Upon Breakdown. STEADY STATE 0 LAG I CAG TIME Figure 3.6. Increased Breakdown Voltage Due to Time Lag.

-44If the cathode is irradiated by ultraviolet light of sufficient energy and intensity, the statistical time lag can be reduced so that the major portion of the time lag is due to the formative lag. This technique is used to separate and measure the time lags. Time lags have been measured that range from less than one microsecond up to 560 microseconds. Irradiation of the gap by a radioactive source emitting for instance alpha particles has been used to reduce the time lag. Alpha particles cause ionization of the gas molecules as opposed to emission of electrons by the cathode. The use of both light and particle radiation are impractical in the automotive ignition system. However, their use is a helpful tool in laboratory studies of spark phenomena,

IV. SURVEY OF IGNITION THEORIES A. General An analysis of the spark ignition process which includes all of the phenomena involved has not been reported in the literature. The ignition process is a combined problem of ionization, diffusion of active particles, chemical reaction, surface quenching and heat transfer. Thus, an analysis must be a simplification of the problem in order to reduce the basic differential equations to a solvable form. B. Thermal Theory (3) In 1920 Wheeler first advocated the thermal theory which he stated as follows, "Ignition depends on the heating of a sufficient volume of the gas to a sufficient temperature." Later Taylor-Jones(5) et al. obtained solutions of the transient heat conduction equation in spherical coordinates. Their results indicated that if the source of energy is strictly thermal, the effectiveness of a given quantity of energy for ignition depends on the manner in which the energy is transferred to the mixture. If a quantity of energy is transferred instantaneously to a given volume of gas, the maximum temperature which the volume attains is higher than if the same quantity.of energy is transferred over a finite time allowing heat loss to occur to the surroundings. It follows that the maximum temperature to which a large volume of gas can be raised by an instantaneous energy source is lower than if the same energy were transferred instantaneously to a smaller volume. Thus, it is concluded that an instantaneous point source of energy is the most effective for ignition. -45

-46C Semenov's Self-Ignition Theory Semenov(69'70) developed an expression for the condition which is required for self ignition to take place in a fuel-air mixture. According to the following procedure he equates the rate of heat generation due to chemical reaction with the rate of heat loss. It is known that a combustible mixture reacts at a slow rate as the temperature increases until a certain ignition temperature is obtained, Then the reaction becomes self-sustaining. The energy liberated by the chemical reaction is expressed as follows Ea V n R T I = Ar V URp an e (4.1) where Ar = constant 3 V = volume of the reacting vessel, cm Up = heat of reaction per molecule, calo/molecule number of molecules a = concentration of reactant, cm3 cmD n = order of the reaction E = arrhenius activation energy, calo/moleo R = gas constant, 1,986 calo/moleaK T = absolute temperature of mixture, OK Heat is transferred from the gas mixture to the vessel walls (ieo,, heat loss) according to the equation: qII = hA(T - Tw) (4,2) Discussion of the reaction rate, order of the reaction and activation energy may be found in References 71 and 72,

-47where h = average convective heat transfer coefficient A = vessel wall area Tw = wall temperature To illustrate the necessary conditions under which self-ignition ofa combustible mixture will occur according to Semenov, it is advantageous to consider the shape of the functions qI and qII with respect to the mixture temperature and their relationships to each othero Consider three different cases of heat liberation qI as shown in Figure 4.lo Cases 1, 2, and 3 represent small, medium and large initial values of the concentration of the reactant a. The heat loss is represented by the straight line qII in Figure 41o. This is the case for one fixed wall temperature Tw o In case 1, qII is lower than qI for mixture temperatures lower than Ta oTherefore, the temperature of the gas will rise to the temperature Ta where qI and qII intersect at point a and are equal. Beyond point a the heat loss to the walls qII is greater than qI and the reaction will cause a return of the mixture temperature to T o At point c, qI and qII again intersect. An unstable condition exists at this pointo Below Tc, qII is greater than qI and the temperature will decrease to Ta Above Tc, qI is greater than qII and a selfpropagating reaction will occur. An external source of energy must be supplied if point c is to be reached. qI is always larger than qII for case 3 representing the highest concentration. Self-ignition will occur in this case. For the intermediate concentration, case 2, curves qI and qII intersect and are tangent at point b which corresponds to temperature T b

-48q qq CASES 3 2 I IC z T T0 Tb TC MIXTURE'w ab TC TEMPERATURE Figure 4.1. Effect of Initial Gas Concentration on Self-Ignition. I 2 3 ~~~~~Cr ~~ C I z (Tw) TaTw (T) Tb Tc MIXTURE TEMPERATURE Figure.2. Effct of Wal Temperature on Self-Ignition

-49This condition defined the critical concentration a for which selfignition will occur with the given wall temperature Tw Figure 4.2 illustrates the effect that wall temperature has upon self-ignition for a single gas concentration. By the same reasoning as.above, self-ignition will not occur for (T ) less than T whereas W w for (Tw)3 greater than Tw, self-ignition will always occur. The condition at point b that qI is tangent to qII defines the ignition temperature T for the given reactant concentration. b At points b of Figures 4.1 and 4.2 the following conditions hold: I = II at T = Tb (4.3) and dT T=T dT (4.4 TT Substitution into Equations 4.3 and 4.4 and then solving simultaneously for Tb gives 4RTw 1+ 1........ (4.5) Tb = Ra ( 2- - E a The smaller. root 4RTb Tb = Ra (4.6) 2E Ea

-50is the solution for the ignition temperature. Equation (4.6) can be expanded in a Maclaurin series and if all terms of higher order than RTw 2 -- 2are neglected the following equation results Ea RT 2 Tb Tw (4.7) E a 2 RTw For (Tb - T) less than, self-ignition is impossible a 2 whereas for (Tb - Tw) greater than RTw whereas for (Tb T) greater than self-ignition would take place. w Ea Do Fenn's Analysis Fenn's(73) approach is similar to that of Semenov.(69) That is, he equates the heat generation to the heat losso He assumes that the temperature of the small spherical volume of gas that surrounds the spark rises to a value such that the heat generation equals or exceeds the heat lost. This condition is analogous to point c of Figure 4,1. The minimum The larger root corresponds to the intersection of qI and qII at a point beyond the inflection point of qI as shown below, -cI and qII T The larger root is therefore not considered because it yields an extremely high temperatureo

-51energy necessary to raise the temperature of the volume of gas to the ignition temperature is related to the smallest self-propagating flame possible at the given initial conditions. By equating the heat loss to the heat generation in this small spherical volume, an expression for the radius of this smallest self-propagating: flame can be found. Heat generation = Heat loss Ea 3~ rn A~ N-N 2 l 2 I(Tf - Tu, URpNfNoP 4e r (4.8) 3 r c r where URP = heat of reaction, cal./mole of fuel Nf = mole fraction of fuel No = mole fraction of oxygen Ar = constant Ea = Arrhenius activation energy, calo/mole R = gas constant, 1,986 calo/mole~K Tf = adiabatic flame temperature, "K Tu = temperature of unburned mixture K = thermal conductivity, calo/cm*K second p = gas density, gm/cm c = constant r = radius of the spherical flame volume According to Fenn, the radius of the small flame volume or reaction zone should be inversely proportional to the square root of the reaction rate. This result is obtained by solving for r in Equation 8) showing that the use of r in the denominator of (4.8) is justified.

-523K (Tf T cArURpNfNoP e e/ ) By assuming that (Tf - Tu) is proportional to NfURp Equation (4.9) reduces to the following where B is a constant into which all other constants from (4.9) are combined. 2-z Ea/R f = B (Nop2)-1 eEa/f (4.10) Once the initial composition and temperature are chosen Tf and therefore r become fixed. The minimum ignition energy is calculated according to the following equation EMiN = D 4 J r3c p(Tf-Tu) (4.11) where cp = mean constant pressure specific heat D = proportionality constant Substitution of Equation (4.10) into Equation (4.11) gives an expression for the minimum ignition energy where a1 combines all of the constants. 3Ea EMIN = alN/ (T- T) e2RTf (4.12) (74,75) Swett 75) modified Fenn's analysis for use with line sources of ignition. In this case, the initial small value of gas surrounding the spark is assumed to be a cylinder of radius r and length d between the electrodes. The resulting equation is as follows for the same assumptions as made by Fenn

-533E EMN = o2N1 (Tf - T) e2RTf (413) where 2 = constant (25) E. Excess Energy Theory Lewis and von Elbe define a quantity called "excess energy" which is equal to the minimum ignition energy. This excess energy is the same energy that is necessary to raise the temperature of the initial spherical volume surrounding the spark to a temperature such that a selfpropagating reaction will occur. That is, the chemical heat generation in the small volume will equal or exceed the heat lost from the volumeo Briefly, the solution involves an evaluation of the energy gain of a mass element of mixture by conduction from the hot gases preceding it as it progresses toward the flame. This same "excess energy" is lost by conduction to the cooler gases following the element as it moves out of the flame. The initial small flame receives this "excess energy" from the spark and it is equal to the minimum ignition energyo The reader is referred to the original work for a complete discussiono Their result is stated below, 2 K u 32 EMN =f - Tu) -2 1 + 1.3 a rl1 1x + —- l [1 r ()/(4.14) 1 + a~00 r1 1-a-,

-54where rl = initial self-propagating flame radius cpPuSu a - Ku a = value of a at r oo 1 cp = constant pressure specific heat, cal/cm3 p = density of the unburned gases, gm/cm3 u Su = velocity of the unburned gases relative to the flame front Ku = thermal conductivity of the unburned gases, cal/cm~K sec. Ta = an average temperature between Tb and Tu Tb = ignition temperature Tu = temperature of the unburned gases Tf = temperature of the burned gases. F. Work of Yang (76) Yang develops equations which predict the minimum ignition energy as a function of the time duration of the energy source. Three source configurations are consideredo They are plane, line and point sourceso He considers that a simple fuel plus oxidant reaction occurs at the surface of each source and then writes the time dependent energy and diffusion equations for a small volume element of the surrounding gas. He assumes that fresh mixture flows toward the source and thus places a mass sink at the source to exhaust the burned gases. Yangs equations for point and line sources are the more appropriate for spark ignition work and they are stated below. URp G(l-a) Kto) EMiN = 16 pocv L 16aon j )l (4.15) LINE ~l6Aap,) ^ \

-554 URpG(l- ao) 3/2 1/Q |Kto 3/2(1-1/Q) EMIN 6= 64PoCvUGl - ) / (4.16) POINT L 64 Ar(apo)n j where & = length of the line source G = mass flow of the mixture toward the source of flame speed S times the density of the gas p gm cm sec a = concentration of the fuel ao = concentration of the fuel at the source URp= heat of reaction, cal/gm A = constant r Po = density of the mixture at the source,, g~ca~~cm K = thermal conductivity, — c cm K sec n = order of the chemical reaction to = time duration of the energy addition Q = constant c = constant volume specific heat cal *v gm The corresponding source temperatures for the line and point sources are given as follows by Yango URpG(l-ao) l/V ( / TS 16 A (.. (4.17) LINE L16 Ar (aPo Kto/ ( - 64 Ar(apo)ff KPOINT 64

-56(77) G. Work of Mole Finch,(8'9) et al. stated the activation theory as, "The necessary prerequisite for ignition of an explosive gaseous mixture is the setting up of a sufficient concentration of suitably activated molecules". Mole assumes that the source of ignition introduces a number of active particles into the small volume of gas surrounding the sourceo Ignition results if this number of particles increases without limito They multiply by chain branching and are lost by chain breaking reaction whereas effects due to temperature change are neglectedo Mole considered a second order chain branching reaction to take place. The resulting equation for the rate of change of concentration of the active particles is d = l + kba2 - kLa (4.19) where a = concentration of active particles or chain carriers 31 = rate of intiation of chain carriers due to ignition source kb second order branching coefficient kL = first order breaking coefficient Ignition occurs if the reaction continues when l1 = 0 e The solution of Equation (4l19) in this case is kL F1 kLt a = - l + l- l e (4~20) kb (kb/kl)ao j kL For a less than - the quantity inside the brackets is greater than one kb kL and the reaction becomes explosive. If ao is greater than -, the reaction stops. Therefore ao has a critical value defined by kb o kb

-57H. Landauls(78) Work A compromise between the Thermal Theory and the Chain Theory (77) set down by Mole is the theory of Landau. Landau assumes that the chemical processes surrounding the source are of a chain branching type and that an initial concentration of active particles is introduced into the small volume surrounding the ignition source. His criterion for ignition is that, "the temperature of the center of the sphere", surrounding the source, "shall never fall". He states and solves time dependent differential equations for the condentration of the particles and the temperature surrounding the source. By applying his criterion for ignition he obtains a relationship between the physical constants for which ignition can take place. His results are best illustrated by Figure 4.3. The quantities A and M are defined as follows: = 4(Tf - T)K (421) URP aorl, r 2 M = k, D (4.22) where Tf = initial temperature inside the small sphere surround the ignition source, ~K Tu = temperature of the unburned mixture surrounding the sphere, ~K cal K = thermal conductivity, cal cm K sec UR= the quantity of heat generated per unit time by each active particle, p arcl particle sec a0 = initial concentration of particles inside the sphere

-58A IGNITION REGION Fig\\\\ \\\\\\\Cr Figure 4.3. Criterion for Ignition.

-59rl = radius of the sphere k1 = first order branching coefficient D = diffusion coefficient for active particles According to Figure 4.3 Landau indicates that if A is of such a value dependent upon M that it lies inside the shaded area for Figure 4.3, ignition will occuro Io Application to Present Investigation The present investigation is primarily concerned with the rate at which the spark energy is discharged into the combustible mixture and the influence this rate has upon the minimum ignition energy. Semenov's(69) approach does not provide an analytical expression to evaluate the ignition energy. However, the conclusions drawn by Taylor-Jones(5) et al. do indicate that an instantaneous point source of ignition is the most effective means of igniting a combustible mixture. In other words, a pure capacitive discharge would be the most effectiveo (76)'The analysis by Yang is applicable to the present investigationo Equations (4.15) and (4.17) are repeated here for convenience. For Line Source: URp G (1 - ao) ] KtO MINL = Pocvl6 Ar (ap)n.15) 1 1 URp G (1 - ao) ~ / " TSL 6 Ar (apo)n Kt(17 It would be advantageous to have an espression for EMIL in terms of the source temperature since for a given to E is a function of T source temperature since for a given to, EMIN is a function of T

-6oSolving explicitly for to in Equation (4.17) and then substituting into Equation (4.15) gives an expression for EMINL as a function of TSL URp G (1 - ao) 1 1 MINL = 16PCv 16 Ar (ap,)n TSL (423) By similar means Equations (4.16) and (4.18) for a point source can be reduced to the following. URp G (1 - ao) 1-2 EMINP = 64pcv 64 Ar (apo)n TSp For a given set of initial condition, (i.e, fuel, fuel-air ratio, pressure, temperature, electrode configuration and spark length 6 or electrode 1-h spacing), all of the terms to the left of T in Equations (4.23) and (4o24) can be combined into constants, DL and Dp o Therefore, MINL = DL 1l (4.25) -SL EMINP - 1 (4 26) EMiNP P DpL SP Yang(76) states that 2 is a constant used as a temperature exponent to approximate the reaction rate in the following manner Ea w = Ar (ap) T T Ar(ap)n (4.27) He indicates that Equation (4.27) is a good approximation over specific temperature ranges and that H is much larger than one. Reference will be made to these results in Chapter VIIo

V. EXPERIMENTAL APPARATUS Ao Description of Equipment Spark ignition of the combustible mixture occurred inside a stainless steel cylindrical constant volume bomb which is shown in Figures 5.2 and 5~35 The spark control panel and peripheral equipment are illustrated in Figure 5.1. The equipment consists of the following nine basic components 1o Constant volume bomb 2. Bomb heater and cooler 3. Manifold system 4o Temperature measurement equipment 5o Bomb pressure transducer 6. Spark control circuit 7. Voltage and current measurement circuit 8. Photomultiplier circuit 9, Ultraviolet light source. lo Constant Volume Bomb The combustion chamber is fitted with two quartz windows, 5 inches in diameter and 1 inch thick for viewing the ignition and combustion processes. The inside diameter of the cylinder is 4 inches, and it is 1-1/4 inches in depth with an outside diameter of 6-1/2 inches. The cylinder and two heads are constructed of 316 stainless steel, and the bomb is mounted on an aluminum pedestal~ Two stainless steel ball valves which are designed for high pressure and high vacuum operation were -61

6&2~~~~~~~~~~~~~ ". i~~:::::; I: -t:..~~~~~~~~~~~~~~~~~~~4 i:::_:'::::ii~~~~~~~ii-iii-iiiii:::i~~~~~~iiiiiii~~~ii~~iiiii4 -~~~~~~~~::::~ ~ ~ ~ ~ ~ ~ ~~~~~~~~~0 CS ~(3 0 41 4)t sif 0~ (2'1

-63~iiiiiiiiiiiiiiiiiii~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~iiiii~~~~~~~~~~~~iiiiii~~~~~~~~~~~~~i~~~~~ii~~~~~~ii~~~~~~iiii~:::iiiiiiii~~,~~..........i....19e g...~-........iii~iiiiii.i i.. -:. ~........................ F.gure 5,3, t Comand~etkar Vf~Me~. ~fof onro }me X............. ~ ~ ~ ~ ~ ~::::::;~.....................................~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~-:::::-::j 1~~~~~~~~~~~~~~~~................ ~,i..................,......... ~ ~I::-:::::I;::.: i_:....................................:: i:~ I~~~~~~~~~~~i:::ii:........ I:: i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~l~~~~~..... ii~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~...... IN~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~I i: ii:,~~~~~~~~~~.......

-64used for charging and exhausting the bomb. Ten 5/16 - 20 NC high strength bolts fasten each head to the cylinder. A 1/8 inch thick, soft rubber gasket functions in much the same manner as an 0-ring to seal the quartz windows against the cylinder. That is, the gasket is compressed to seal, but does not carry the full compressive load exerted by the ten headbolts. Both electrode assemblies which are shown in Figure 5.2 are installed horizontally in the bomb. The assemblies may be easily removed to clean the electrodes and each electrode is replaceable. The three basic type electrodes used in this study are shown in Figure 5.4. The assemblies are sealed with teflon O-rings. The cathode which is located on the left side of the bomb shown in Figure 5.2, is electrically insulated by the long teflon sleeve. This cathode is the high voltage electrode. The adjustable electrode assembly on the right side of the bomb is the anode. It is equipped with a micrometer handle so that the electrode spacing can be accurately seto A pressure transducer is attached to the bomb cylinder through a 14 mm threaded hole located in Figure 5.2 just below the cathode, Temperatures of the combustible gases are measured by a thermocouple which is inserted from the top of the bomb as shown in Figures 5,2 and 5.3~ The wires are sealed by a Conax high pressure connectoro 20 Bomb Heater and Cooler Initial temperature of the mixture before ignition is controlled manually by heating or cooling the outside bomb surface as required. Two hundred ohms of electrical heating element wire are wrapped around each head.

-65I____ 1/32R 5/8 1/46- K \ GLASS DISK Type "A"; Cathode-Aluminum; Anode - 302 S. S. 1/32R' (/8-): (+) 1/4 5/8 3/32 9/32 _ GLASS DISK Type "B"; Cathode-Aluminum; Anode - 302 S. S. (Figure 5.4. Ee c t +1/4 5/8 */8 3/16 9/32 - -1/16 GLASS DISK Type "C"; Cathode-Aluminum; Anode - 302 S. S. Figure 5.I. Electrode Configurations.

-66High temperature electrical tape is used for insulation between the heating element wires. The heaters are connected in series with a variac to control the current, and thus the rate at which the bomb and gas are heated. Four turns of 1/8 inch copper tubing are wrapped around each head over the top of the heating element wire. This tubing carries tap water as a coolant. 3. Manifold System The fuel-air mixtures were premixed in three small mixing tanks shown below the bench in Figure 5.1 and shown in the schematic diagram of Figure 5.o5 Each tank is equipped with two valves, one valve is used to throttle the mixture is used in conjunction with the first when scavenging the tank before evacuation. This procedure is described in Chapter VI. Ball valves are used at all points where throttling is not required. Piping is schedule 80, 1/2 inch black iron pipe. This large diameter was chosen to reduce the evacuation time to a minimumo One Cenco vacuum pump is used. The system can be evacuated to less than 100 microns of mercury absolute with this pump. Small absolute pressure measurements are made with a McCleod gage having a scale of 0 to 5000 micronso Pressures from 0 to 67 inches of mercury absolute are measured with a mercury manometer which can be read to the nearest one-hundredth of an incho Higher pressures are measured with either of two bourdon pressure gageso Figure 5o1 shows the location of the gages and the schematic in Figure 5.5 illustrates their location in the manifold system. Both nitrogen and dry-air are passed through Drierite desiccant to assure a negligible content of moisture. The desiccant is contained in two 8 inch lengths of pipe of 1-1/2 and 2 inch diameters respectively.

_EXHAUST MIXTURE TANKS ~~~~~~EXHAUST ~ EXHAUST /^^^\ __-0-~ ~~~~~~~~~~~~~~EXH#AU ST BOMB EXHAUST EXHAUST EXHAUST N~0 ---------— 0 - -0 ------------- cc~~~~~~~~~~~~~~~~~~~~~~O 0-100 PSI w Mc CLEOD z GAGE 0-300 PSI 0 BALL VALVES t gf VACUUM ILr l THROTTLING VALVES 0UM A PRESSURE REGULATORS t cr a z az Figure 5.5. Manifold System.

-68The lengths of pipe are fitted with screens and reducing pipe couplings at each end. They are shown on the right hand side of Figure 5.1 in front of the gas cylinders. The nitrogen and dry-air size 1-A cylinders are equipped with pressure regulators shown in Figures 5.1 and 5.5. 4. Temperature Measurement Equipment Iron-constantan thermocouples are used for all temperature measurements in conjunction with a Minneapolis Honeywell Brown model 156X63P24 continuous reading potentiometer. Temperatures of the combustible gas mixture inside the bomb are measured by a thermocouple constructed of 24 gauge wire. The length of wire exposed to the gas was looped once to increase the length of exposure and thus reduce the error due to conduction along the wireo This thermocouple is double radiation shielded according to the method described in Reference 79. The shield is shown protruding into the bomb from the top in Figure 5.2. Four thermocouple junctions were taped to the outside bomb surface on the top surface of each head and top and bottom of the cylindero Top and bottom refer to the top and bottom of the bomb as it is shown installed in Figure 5.o2 Well-type thermocouples were used to determine the gas temperature inside the mixture tanks. 5~ Bomb Pressure Transducer Time dependent pressure indications were used to determine whether or not ignition occurs inside the bombo A Kistler model 601 piezoelectric transducer is used in conjunction with a Kistler model 568 electrostatic charge amplifier. The amplified signal is displayed on a model 502 Tektronix

-69oscilloscope. In Figure 5o1, the charge amplifier is located between the two oscilloscopes. A water cooled adaptor was used to mount the pickup on the cylinder of the bomb. 6. Spark Control Circuit It is desired to control the time duration of the spark discharge. The method chosen in this investigation is the method first used by Finch (17) and Suttono They showed that the spark can be extinguished at a predetermined time during the discharge by turning the primary current on. Referring again to Figure l.1, a spark is initiated when the primary current is turned off by opening switch S1 i By closing S1 again during the discharge, the primary current is turned back on again. This causes the flux in the coil to change and reverse the voltage and extinguish the spark. Finch and Sutton used two switches connected in parallel where one is open while the other is closed~ They were actuated by a cam as in a conventional distributor. Thus, the spark could be extinguished at set intervals of time after the spark was initiatedO The use of a switch or relay to turn on the primary current poses the difficult problem of switch bounceo Therefore, it was decided to use transistor switching in this investigation. Figure 5,6 shows the circuit used to control the spark duration. The system can also be used to produce a conventional spark of uncontrolled time durationo With switches S6 and S8 in positions 2, switch S7 is used just as switch S of Figure 1lo. The transistor network should be ignored for this description. Switch S4 must be closed for either the conventional or for transistor operation. Closing switch SO energizes relay 10

S7A S.I PULSE CIRCUIT SIA z C 3 RR, S 4 C2~ SA ^I I %R000: I VOLTAGE 1 R 4_______________1 R7 DIVIDER C, I L^' —l e, V — J2 —-' TR. R4 R Di TRIGGER -— |'I" I "'- ~~C2 L /Sz ^_ S 115 V. A.C OSCILLOSCOPE Figure 5.6. Spark Control Circuit.

-71switch S7 and contacts S7A and S7B open. S7A turns the primary current off and S7B reduces the voltage at the oscilloscope trigger to zero and triggers the oscilloscope sweep mechanism to record the spark voltage and current S7B is adjusted so that is opens about 40 microseconds before S7A o Voltage and current measurements are covered in the next sectiono All of the minimum ignition energy data was obtained using the transistor switching circuitryo Switches S6 and S8 must be in position 1 for transistor operation. Transistors Q3A and Q3B functionally replace S o These transistors turn the primary current on and off. 7A The combined base currents of Q3A and Q3B are controlled by transistor Q4 o A monostable multivibrator circuit is used to control transistor Q4 o The reader is referred to References 80, 81, 82, and 83 for a thorough description of multivibrator circuitso Transformer T1 and switch S1 were added to the multivibrator circuit to provide a positive pulse to upset the stable operating point of the circuit as shown in Figure 5~6. The multivibrator and added transformer circuit is referred to as the pulse circuit in the figure. Switch S1 is normally in a closed position. Transistor Q1 is negatively biased and thus turned off. Q2 is based positive and operates in saturation. Point a is at the same voltage as the emitter of Q2 o Opening S2 causes a flux change in. T1 which transmits a positive voltage pulse through diodes D1A and DB to the base of Q1 l1A lB Q1 begins to conduct and its decreasing collector voltage is transmitted through either capacitor C A or CB to the base of Q2 o Thus, the )A. 5~B

-72collector current of Q2 begins to decrease with resultant increase in the base voltage of Q1. This process quickly drives transistor Q1 into saturation and Q2 to cutoff. The base of Q2 is then at a negative voltage. Capacitor C3A or C3B then discharges through R3. The time duration of the output voltage at point a is determined by the time constant of the C3R3 network. The waveshape at point a is shown in Figure 5.7. Normally the waveshape at point a in a multivibrator circuit would be square as indicated in Figure 5.7 by the dashed lines. However, the resistor R8 was added as shown in Figure 5,6 to obtain the waveshape indicated by the solid lines in Figure 5.7. This was necessary to decrease the rate at which Q4 is switched on, This positive voltage at point a and the base of Q4 maintains the transistor Q4 off until the voltage at point a begins to decrease along the line 1-2 of Figure 5.7. This gradual decrease in voltage gradually turns transistor Q4 on. Q4, in turn, controls Q3A and Q3B which are turned back on at a slower rate than they were turned off by Q4 during the initial rise in voltage at point a o The transistors Q3A and Q3B must turn the primary current on at a slower rate than they turn the current off to prevent a rapid change in flux in the ignition coil T2. A rapid change in flux causes the cathode to become positive and anode negative at such a rapid rate that a second spark is discharged in the gapo Zener protection of the transistors Q3A and Q3B follows one of the methods suggested by Larges(84) et alo

-73) VI I 0 l > ". TIME~2 IMFigure 5.7. Waveshape at Point a in Figure Figure 5.7. Waveshape at Point a in Figure 5.6.

-74Batteries V3A and V3B when connected in series with resistors RgA and R9B respectively through switch S12. are used to drive the bases of QA and QB slightly positive when transistor Q4 goes offo This positive voltage speeds up the cutoff of transistors Q3A and 3B which results in a faster rate of voltage rise in the secondary circuit. At larger electrode spacings the breakdown voltage becomes very high. This breakdown voltage can be reduced slightly by switching the batteries out of the base circuits using switch S12 o The breakdown voltage is reduced because of the effect that the statistical time lag has upon breakdown as was discussed in Chapter IIIo The ignition coils used in this investigation were a Mallory model F-12T coil and a special low capacitance coil made for this investigation by Delco-Remy Division of General Motors Corporation. The Mallory coil has a secondary capacitance of 52 picofaradso The Delco coil has a secondary capacitance of only 29 picofaradso Determination of the capac(86) itance of these coils was carried out by a procedure outlined by StoutO This method and the data for the measurements of the Delco and Mallory coils are found in Appendix Bo The circuit and coils are equipped with plugs so that the coils can be interchanged as desiredo Total time duration AtBR + Atsp of Figure 12 can be adjusted from 100 to 1500 microseconds with the circuit shown in conjunction with either the Mallory or Delco coilo The values of all the electrical components in Figure 5~6 are listed in Table IX, Appendix Do Figure 512 shows the voltage and current traces for a series of sparks each having a different time duration.

-757o Voltage and Current Measurement The high voltages necessary to cause spark breakdown are too large to be measured directly by the oscilloscope. Consequently, a voltage divider is necessary to reduce this voltage to a value compatible with the type oscilloscope being usedo Appendix C deals with the necessary design criterion for a voltage divider, A Tektronix P6015 high voltage probe was used in this investigation and is shown in Figure 5.2 connected to the cathodeo It has an attenuation ratio of 1000 to 1 and can be used with peak voltages up to 40,000 voltsO Current measurement is accomplished by recording the voltage drop across the known resistor R in Figure 5~60 17 The voltage and current signals are simultaneously fed to a dual beam model 502 Tektronix oscilloscopeo The two traces are recorded using a Hewlett Packard Oscilloscope camera which is equipped with a Polaroid backo It is necessary to use ASA 3000 speed film to obtain good photographs of voltage and current as a function of time in the sparko The oscilloscope is triggered by the same signal that initiates the operation of the multivibrator circuito 80 Photomultiplier Circuit The bomb was fitted with a 931-A photomultiplier tube to obtain a signal which was proportional to the light intensity of the spark. This was useful in the initial stage of the equipment development to correlate spark light output duration with the voltage and current traceso Also, the photomultiplier was used to examine the spark structureo Figure 508

-76shows the photomultiplier in position on the right side of the bomb, A typical trace which is proportional to the light intensity of the spark versus time is shown in Figure 5o10o The photomultiplier circuit dia(85) gram is shown in Figure 5o9o The circuit used by Saad was modified as required for the present investigationr 9o Ultraviolet Light Source Ultraviolet light is impinged upon the cathode to overcome the work function and reduce the statistical time lag for breakdown as described in Chapter III A General Electric, 100 wAatt, H4AB lamp is used as the source for the ultraviolet lighto The outer safety shell of the lamp is glass and it was removed to increase the intensity of the shorter wavelength lighto The light is focused upon the cathode with a quartz lens. Figures 5~2 and 503 show the light systemo The effect of the ultraviolet light in decreasing variations of the breakdown voltage is shown, in Figure 5ol1 Bo Calibration, of Instruments lo Oscilloscope Sweep Rate The horizontal deflection system calibration is most important to the energy calculations since the energy is obtained by graphically integrating power as a function of timeo An Erie model 400 counter-timer was checked against the National. Bureau of Standards time signal broadcast over station WWVo The counter was setup to count its own crystal oscillator output for 5, 40, and 45 minu.te time intervalso Table X in Appendix D shows the calibration points taken. The maximum error was 0 04 percento

-I 5 5 o 0 0-0 ~...nn t:~ ~ lil~~iii~0 I E~ ~ r~~d I tZ:I$P~~~C s e=,~ ~z o g~s

-78Next, the output from a model 181 Tektronix time marker was compared with the output of the Erie counter and found to be within the accuracy that the oscilloscope can be read, The time marker is a standard instrument sold by Tektronix to be used for oscilloscope calibrationo Finally, the sweep rate of the oscilloscope was calibrated using the time marker. Table XI of Appendix D shows the data and the corrected sweep rates which were used for the energy calculationso 20 Voltage and Current The vertical amplifier deflections were checked be measurement of the voltage of two mercury bias cells. The vertical deflections were within the three percent accuracy stated by the manufacturers of the oscilloscope. Next, the square wave calibrator vertical amplitude was checked with the vertical amplifiers and found to be within the three percent accuracyo The high voltage Tektronix P6015 Probe or voltage divider was frequency compensated and the attenuation was adjusted to 1000 ~ 1 according to the procedure outlined in the probe manualo A discussion of the conditions that must be satisfied for frequency compensation and design of a voltage divider is included in Appendix C. No calibration was necessary for current measurement. Voltage drop was recorded as a function of time across a one percent, 150 ohm carbon resistor which was connected in series with the secondary circuit as illustrated in Figure 5060 35 Pressure Gages The 0-100 and 0-300 psi bourdon gages were calibrated using a dead-weight testero The calibration data and corrections are found in Table XII of Appendix Do The corrections were applied to all of the gage readings made

-794, Pressure Transducer No pressure calibrations were necessary since the pressure time trace was only used to determine whether or not ignition occurred inside the bombo 5. Thermocouple Calibration An iron-constantan thermocouple was calibrated in conjunction with the Brown potentiometer The ice-point and boiling-point of water were usedo The calibration correction curve is shown in Figure D-1 in Appendix Do 60 Frequency Response of Photomultiplier The response of the photomultiplier tube was determined by exposing the tube to light through four tiny pinholes, A rotating disc having six holes was rotated at high speed between the light source and the photomultiplier tube. The output signal was displayed on an oscilloscope and photographed. Calculations were made to determine the frequency response from Figure 5135 The calculations show that the initial rising portion of the wave rises at an equivalent frequency of 4000 cycles per secondo The falling portion has an equivalent frequency of approximately 2000 cycles per secondo This frequency response would be higher if a more intense light source were used with only one small pinholeo

-8oFigture 5.1.0. Typ.cal. Current and Corresponding Light Trace; Upper -current, Lowrer-l..ight. l~I~i 11F-gure 5.31. Effct, of' Ultra-Iolet,lgha Upon Breakdown Voltage; Upper-.UV not Used, LoJe.rUV used. F.igure 512,. I.ll.ustration of Spark Time Duratitlon Control; Upper-...Voltage, Lower -c urroent. Figure 5.:13. Phtoomul]tip.lier CaL 3.bx atiot T grace; Sweep Rate t 500 |sec/cm,

VI. EXPERIMENTAL PROCEDURE A. Fuel-Air Mixing The fuel used in this study was chemically pure, 99.0% propane which was purchased from Matheson Company. Air was also purchased from the Matheson Company in cylinders at 2000 psi. The dry-air has a dew point temperature of -75 Fo Each mixture tank was scavenged with dried nitrogen before evacuation to less than 300 micronso Propane was then introduced until the proper pressure for the desired equivalence ratio was reached. The tank valve was closed and the remainder of the system evacuated. Finally, the mixture tanks were charged with dry-air to the required total pressure for the desired equivalence ratioo Care was taken not to open the mixture tank valve before the charging air pressure was well above the fuel pressure in the tanko Thus, there was little chance for fuel to diffuse out of the tanko The partial pressures were calculated according to the following equations. Propane-air reaction C3H8 + Tr502 + r(5)(3576)N2 - r3C02 + r4H20 + T5(3576)N2 (6.1) where Actual air-fuel ratio Stoichiometric air-fuel ratio The partial pressure ratio of the fuel is expressed as follows: Pf 1 PT = {(6~2) BPT (25o8) + 1 (6.2) -81

-82where Pf = partial pressure of fuel PT = total pressure of fuel plus air A check calculation showed that the assumption of an ideal gas involves negligible error. The equivalence ratio is defined as the ratio of the actual fuel-air ratio to the stoichiometric fuel-air ratio. (F/A)- 1 = 1 (6.3) Volume 54-e76 23.8n Mf 1 (F/A) = = (6.4) Mass 23.8rMar 15.667 air where F/A = fuel-air ratio Mf = molecular weight of fuel Mair = molecular weight of air Therefore, the equivalence ratio is expressed as follows: = 1/A (6,5) A tabulation of the pertinent mixing data is included in Table XIII Appendix Eo This table also includes the required fuel pressures for a total mixing pressure of fuel plus air of 200 psia. Table XIV in Appendix E shows the original data for the mixtures used in this investigation,

-83B, Measurements of Ignition Energy The equipment is designed so that the total ignition energy is easily controlled by means of the time duration of the discharge as discussed in Chapter V. The peak power was maintained nearly constant for a given set of runs while the time duration of the sparks was varied0 With very short duration sparks and a given peak power, ignition does not occur, provided the peak power setting is not too high~ The energy was increased for each consecutive run, by increasing the time duration, until consistent ignition was obtained. The same procedure was followed for three different peak power values at each equivalence ratioo The peak power is dependent upon the breakdown voltage and the peak current as described in Chapter I and shown in Figure 12o An ultraviolet light was used to maintain an approximately constant breakdown voltage. Variation was approximately + 10%, but it was considerably better than when no ultraviolet light was usedo The peak current was set by adjustments of the secondary series resistance and parallel capacitance as shown in Figure 5060 To the left of the large ignition coil in Figure 5.2 one of the ceramic capacitors used for this purpose can be seen. The series resistors are shown in the same Figure 5,2 connected between the high tension terminal of the coil and the cathode on the bomb. Each set of runs was performed with the same electrode spacing and electrodes, Type "B" electrodes were used in those runs where the electrode spacing was larger than the quench distance, whereas type "A" were used for gaps less than the quench distance.

-84For each run, the bomb and associated piping were evacuated to less than 300 microns before being charged with a fuel-air mixture. Correct temperature of the mixture inside the bomb was obtained by manual control of the bomb heater or cooler. When the desired temperature was reached, the bomb pressure was again checked and readjusted if required. Then the charging valve was closed, the cathode exposed to the ultraviolet light, camera shutter opened and the spark fired. The criterion for ignition was whether or not the observed pressure time trace showed a pressure rise. This trace was visually observed and not photographed. In addition, a temperature increase was indicated on the potentiometer if combustion occurred. Regardless of whether or not ignition occurred, the exhaust and inlet valves were opened and the bomb was scavenged with dried nitrogen and the charging process was repeated for the next run. C. Electrode Spacing Zero spacing was obtained by an electrical continuity checko An ohmmeter was connected across the two electrodes and the electrode spacing was reduced to the point where the ohmmeter reading indicated that the electrode tips touched. The corresponding micrometer reading was recorded. All electrode spacings were then set relative to this readingo Each time the electrode assemblies were replaced after cleaning, etc., a new zero reading was made. D. Quench Distance Type "A" electrodes shown in Figure 5.4 were modified slightly for use in the determination of the quench distances. The modification

-85consisted of removing the 1/32 inch radius from the tips of the electrodes so as to make them flush with the glass flanges like the anode of the type "B" electrodes shown in Figure 5.4. The change was necessary because the results obtained with the type "A" electrodes proves to be erroneous. A discussion of the results obtained with the type "A" electrodes modified and unmodified is found in Chapter VIIo Evaluation of the quench distance for a given fuel-air ratio, etc,, involved a series of runs at different electrode spacingso Each run was made with a fresh mixture and the electrode spacing was increased for each run until ignition occurredo Then the electrode spacing was reduced in small increments until the quench distance was defined. The mixtures were exhausted from the bomb after each spark was fired. All of the spark were of the same duration and peak powero

VII. RESULTS AND DISCUSSION Ao Results The electrode spacings used for determining minimum ignition energies in this investigation are shown in Figure 7.1. Quench distances obtained by three other investigations are also shown on the same figure. Minimum ignition energies for propane-air mixtures were obtained for equivalence ratios of 0.69, 0.73 and 0.83. Two electrode spacings were used when determining the ignition energies for an equivalence ratio of 0.835 one smaller and the other larger than the reported quench distance shown in Figure 7.1. The total energy is plotted as a function of the peak power P in Figures 7.2 through 7.5 for each of the P preceding equivalence ratioso The circles are points at which ignition occurred and the squares represent non-ignition points. These same total energies are plotted as a function of PiA/PHA in Figures 7~6 through 79~. PLA is the average power or average rate of energy discharge during the low rate component of the sparko The average power is obtained by dividing the total energy dissipated during the low rate component EL by the time duration of that part of the discharge. PA is the average power or average rate of energy discharge during the high rate component. It is determined in a manner similar to that used for P LA For convenience, a plot of the minimum ignition energies which (21) were reported by Lewis and von Elbe for pure capacitive sparks are shown in Figure 7.10 as a function of equivalence ratio. The dashed portion -86

-87of the curve is an extrapolation of their data to the smallest equivalence ratio used in this study. All of the data reduction calculations were performed on an IBM 7090 digital computer. Details of these calculations are described in Appendix F and the results are shown in Tables III through VI. The original data readings are shown in Tables XV through XVIII. The last three digits of the run number listed in these tables correspond to the mixture number. Data such as equivalence ratio, etc., relating to each mixture number are found in Table XIVo B. Discussion The quench distances reported by References 21, 27, and 28 and shown in Figure 7o1 are seen to approach the same value at the richer mixtures. In the present investigation a quench distance of 0.091 inches was measured at an equivalence ratio of 0.83 using type "A" electrodes. Since this value was smaller by 0,023 inches than that reported by Reference 28, modifications were made on the electrodes to see if better agreement could be obtained. Use of the modified type "A" electrodes gave a 0.120 inch quench distance which was only 0.006 inches larger than the reported value of Reference 28. Therefore, the modified type "A" electrodes were used for the remainder of the quench distances measurements. Disagreement in the values for d exists in the literature for lean mixtures as shown in Figure 7.1. Similar results were obtained in the present investigation, in that the quench distances showed larger scatter at the smaller equivalence ratios. To be assured that tests would

-880.260 ---- 0.240 --- --- --—,~ REF. 28 0.220 Po - I ATM. To, ROOM TEMP -- - ---.... REF. 27 Po = 29.29 in. Hg. 0.200 --- -- Tes75F _1__ —-' —- REF 21 Po= I ATM. e) 0.180 - To ROOM TEMP uj 0.180 Z -- -- \ - 0El ELECTRODE SPACINGS:\ I \ USED IN THIS WORK 0.160 \'- o0.140'' -\|S ~0201 I L - - - z 0.120 ________ w 0.100 0.080 __ 0.0601 I X 0.64 0.68 0.72 0.76 0.80 0.84 0.88 EQUIVALENCE RATIO, < Figure 7.1. Dependence of Quench Distance Upon Equivalence Ratio for Propane-Air Mixtures.

5 --- -- [_ _- _ — _ _ —_-__ _ —-----------— _ —---------- - ---- PROPANE -AIR MIXTURE 4__ _ __ __ __ ______ ___ =< s0.83 T, =79~F co -- -P-, P 30 in.Hg.Abs. g~ 8I I I I I I I I I ~ s~~= 0.091 in. __3 -- -- --- __ ___ __ _ __ __ TYPE "A" ELECTRODES ___ 3 SZ~ ~I~~ 0I~~ I I I oIGNITION L~ ~ _ __ — -- - -- -- -- -- - BNO IGNITION 0 w 2 (C 1 I Q %"I i I B I i I? 2I2 — __ fi -" — 0-__ ___-_ ___ ______ __ __ 0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~0 _ _ _ _ PEAK POWER WATS Figure 7.2. Effect of Peak Power Upon Minimum Ignition Energy for b Itess than Quench Disttance. ( -0 -- ---- o ~ o -- *^ n ~ ^ — - --- -- - O 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 PEAK POWER, WATTS Figure 7.2. Effect of Peak Power Upon Minimum Ignition Energy for 6 Less than Quench Distance.

50 C4.O "' -- -- -- -- -- -- - -- — An _ ___ _ __ ___ __ PROPANE AIR MIXTURE w =0.83 -j 0::)_ __ ______I I I I IT.=79 F 2n3~~~~~~~~~~~~~ l l l l l l P = 30 in. Hg. abs. 0 304____ 060 80 100 120 140 160 18020022 = 0.122 in. P.EA0 PTTYPE "B" ELECTRODES ~~~~Fgr>- 0 IGNITION 0 a —----------— 0 NO IGNITION uJ 2.0 ~ I — 0~:' —-- j.I EB I __ __-_ —_ —--- __ ___ _ __ __ 1.0 PEAK POWER, WATTS Figure 7.3. Effect of Peak Power Upon Minimum Ignition Energy.

50 PROPANE - AIR MIXTURE 4.0 -- I I I I I I I I I I I I = 9 0.73 T. = 79 1F uo ___ o _______ _ P.= 30 in.Hg.abs. r~~~~dS I I~~~~~~~~~~~~ 8 II I I l =0.195 in. -30 I____I____I_______ I_____ _II____ I |TYPE "B" ELECTRODES 3.0 3j' O0 IGNITION I I ____ ___________ I NO IGNITION w 2.0 _ o04 0o 0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 PEAK POWER, WATTS Figure 7.4. Effect of Peak Power Upon Minimum Ignition Energy. I.O r Q ~ -- 0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 PEAK POWER, WATTS Figure 7.4. Effect of Peak Power Upon Minimum Ignition Energy.

5.0 U ~l 0 - 4.00 D 0 o~~~o 0: 2 400 0 00 2 3.0' C-I F 7 E f e c of P ea P w U o I_ 0n r 2.0 __ z -LJ -' I''-* -- I PROPANE- AIR MIXTURE 0 0 1.00.69 0 1.0 T.79 To= 79 OF ____ ____0_____P,_ = 30 in. Hg.abs. 8 = 0.254 in. 0 IGNITION TYPE "B" ELECTRODES 0 NO IGNITION 0 0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 PEAK POWER, WATTS Figure 7.5. Effect of Peak Power Upon Minimum Ignition Energy.

5.2' i' i' i|....-. 4.8 PROPANE -AIR MIXTURE 4.4 --— __ 4=0.83 To = 79~F 4.0 P s= 30 in.Hg. Abs. 8 = 0.091 in. 3.6. TYPE "A" ELECTRODES U) LW 3.2 " 0 IGNITION J3 _ _ _ / NO IGNITION ~ 2.8 - A REF. 21 2 _2.4 0 _ _ -J 2.A0 l. 2.0 -0 —-- Z 1.6 1.2 -- / o 0.8 _. 0.4 - __ —--- 0 0.04 0.08 0.12 0.16 0.20 0.24 0.28 0.32 036 0.40 P /P LA HA Figure 7.6. Effect of Rates of Energy Input Upon Minimum Ignition Energy for 5 Less than Quench Distance.

5.2' I --- ---' F T I 4.8 PROPANE-AIR MIXTURE 4.4 _________ 0.83 To= 790F 4.0.. R =30 in. Hg. Abs. 8 0.122 in. g___3.6__ ______ _ ______ _______ TYPE "B" ELECTRODES Cu 3.2 - 0 IGNITION _ 0 NO IGNITION e2.8 A A REF 21 -i 2.4 _ — >- 2.0' 0 0 W E 0- -2 -— " - - I.l -3 zE- 013 0.4 0.4 0 0.04 0.08 0.12 0.16 0.20 0.24 0.28 0.32 0.36 0.40 P / P LA HA Figure 7.7. Effect of Rates of Energy Input Upon Minimum Ignition Energy.

5.2 4S 4.4 __ _PROPANE AIR MIXTURE = = 0.73 4.0 aT = 79 ~F P = 30 in. Hg. abs. 3.6 8 = 0.195 in. 0 3.2 __| TYPE "B" ELECTRODES _ 3. -L —2-i 0 IGNITION -j o __ _0 NO IGNITION X 2.8 AZ REF. 21 0 0 2.4. P ZHA -Figur7.8 Effe 0 t ~ 0 0.04 0.08 0.12 0.16 0.20 0.24 0.28 0.32 0.36 0.40 Figure 7.8. Effect of Rates of Energy Input Upon Minimum Ignition Energy. Figure 7.8. Effect of Rates of Energy InpUt Upon Minimum Ignition Energy.

5.2 i T 1 T f IF i 0 El 4.8 / 0 4.4' -' 0 4.0 _ ___ 3.6 0 0 2.8 _ 2. 4_0 U_ _ _ _ __ _ 2.4..ofl~ ~~-0 —. 0' 1~ a 0 El' 2.0 0/ = ~ ~ ~ iu3 7.9 e f w PROFANE-AIR MIXTURE 00 0 0T 79 F 1.2 0E lP =30 in.Hg.Abs. 0-I~~~~ 00=0.254 in. 0.8 ___ TYPE B ELECTRODES 0 IGNITION 0.4 l- NO IGNITION A RER 21 O0 L I I 1 I 0.04 0.08 0.12 0.16 0.20 0.24 0.28 02 0.36 0.40 P /P LA HA Figure 7.9. Effect of Rates of Energy Input Upon Minimum Ignition Energy.

-974.0 —| 3.8 3.6 ---- -3-. T.To= ROOM TEMP. 3.2 ---— _ P.o I ATM. REF. 21 3.0 - - r EXTRAPOLATION 2.8, 2.6 -- ------- - w =O \ 0 2.4 2 -\ 2 2.0.. 2,3 1.8 1.6.4xtures z 1.2 = 1.0 Z 0.8 0.6 0.4 - 0.2 0.64 0.68 0.72 0.76 Os0 0.84 0.88 Figure 7.10. Minimum Ignition Energy vs a for Propane-Air Mixtures.

-98be run with gaps larger than the quench distance, larger values than those reported by Harris(28) were used, Type "B" electrodes were used for ignition energy measuremnets where the electrode spacings were larger than the quench distance. These electrodes were used because modified type "A" electrodes require large breakdown voltages causing improper operation of the control circuit with no spark formation. An aluminum cathode was used in all cases to reduce the work function of the cathode surfaceo This is discussed in Chapter III. The upper dashed curves in Figures 7.2 through 7.9 were drawn through the non-ignition points representing the largest energies while the lower dashed curves were drawn through the ignition points having the smallest energieso The area between the two curves is a transition region between non-ignition and ignition. Spark energies which are smaller than the values defined by the lower curves would not possess enough energy to cause ignition. On the other hand, energies which fall above the upper curves would be expected to cause ignition. Therefore, the upper dashed curves in Figures 7.2 through 7,5 show the minimum ignition energy as a function of peak power Pp In Figures 7.6 through 7.9 the upper dashed curve shows the minimum ignition energy as a function of the ratio PLA/PHA o Consider the curves for minimum ignition energy versus peak power shown in Figures 7.2 through 7.~5 In each figure, the minimum ignition energy decreases as the peak power is increased. It was found that this relationship can be expressed by an empirical equation of the following form:

-991 V EMN = F or E P F (7.1) p v MIN p P where F = constant v = constant If a log-log plot of EM versus P is a straight line, the slope of MIN P this line defines the value of the constant v. The minimum ignition energy curves of Figures 7.2 through 7.5 are replotted on a log-log plot in Figure 7o11, Straight lines closely approximate the data and the values of the slopes v are listed in Table II. The corresponding values of F are also shown in the table. Each F was calculated according to Equation (7.1) using the graphically determined constant v The negative slope v is largest for the case where the electrode spacing is smaller than the quench distance. For electrode spacings larger than the quench distance, the slope v increases as the equivalence ratio is decreased, In other words, the peak power has a greater effect upon the minimum ignition energy at smaller equivalence ratios. It appears that this is due to greater quenching of the initial flame volume by the mixture surrounding it which may be due to chain breaking rather than thermal processes. TABLE II EMPIRICAL CONSTANTS FOR ENERGY EQUATION 0 b dq F v 0,83 0o091 0,114 52.1.738 o.83 0.122 0o114 6.61 348 0.73 0,195 0,183 15.1.427 0,69 0,254 0.242 37.5.463

-100100.0 = 0.69 10.0 -- - = 0.254 in.: 0.73 { l 3-S = 0.195 in._'' 1.0 1N 0.83 = 0.091 in. 10 100 1000 Pp, WATTS Figure 7.11. Determination of v

-101It would be of interest to see if any relationship exists between Equation (7.1) and Equations (4.25) and (4.26) of Chapter IV. Recall that Equations (4.25) and (4.26) were derived for the case where heat is transferred to the mixture from a source at temperature Ts This temperature is proportional to the rate of heat transfer q. Since q is really a power term, the relationship between Equation (7.1) and Equation (4o25) or (4.26) is easily seen. The relevant difference between q and Pp is that q is constant over a period of time while Pp is an instantaneous peak value. Another difference exists between Equations (4.25) and (7.1) Yang indicates that (f-l) should be large if the approximation in Equation (4.27) is to holdo The corresponding values of v found in this investigation are all smaller than unity. Nevertheless, the basic forms of the equations are the same. The shapes of the instantaneous power versus time curves for two sparks may be different and yet the peak powers Pp may be the same in each caseo Consider the two curves shown in Figure 7o12. The area under curve b is larger than the area under curve a. If the energy for curve b is just adequate for ignition of a given mixture, the energy of curve a would be inadequate. The overall time duration of spark a would have to be increased as shown by the dashed curve to cause ignition. Thus, it can be seen that the peak power Pp does not thoroughly describe the spark requirement for ignition. To alleviate this, a plot of the total energy versus the ratio PLA/PHA was chosen where PLA and PHA were described in Section A. The two sparks in

-102PP,c iA PA)bI PLA). _r^-^. TIME Figure 7.12. Power vs Time for Two Different Sparks.

-103Figure 7.12 would have nearly the same PHA, but the respective values of PLA would be different. Consequently, PLA/PHA would be different for the two sparks. The results plotted in Figures 7.6 through 7.9 show that the minimum ignition energy decreases as the ratio PLA/PHA decreaseso At PLA/PHA = 0, corresponding to pure capacitive sparks, the data extrapolates to values close to those reported by Lewis and von Elbe(21) for capacitive sparks. Similar comparisons at all equivalence ratios show that the largest difference if for an equivalence ratio of 0.73 illustrated in Figure 7.8, At all equivalence ratios, the width of the transition region increases with decrease in the peak power or increase in the ratio PLA/PHA o In other words, the transition region increases in width as the rate of energy input decreases. In addition, the band is wider for the leaner mixtures. Ignition seemed to occur at random within this region. Lewis and von Elbe (l221) also noticed a randomness of ignition with hydrocarbons such as propane and higher. In addition, at the leaner mixtures they also noticed this randomness of ignition. There are several possible reasons for this, Lewis and von Elbe indicate that the cause may be due to the higher breakdown voltages encountered with leaner mixtureso That is, the quench distance increases with decrease in equivalence ratio whereas according to Paschen's Law as discussed in Chapter III the breakdown voltage increases. The higher resultant currents may cause a net mass flow between the electrodes due to electron wind resulting in a contraction of the reacting volume of gas after the discharge. Irregular movement of the initial flame kernel

-104toward the anode, a possible source of variation in the quench effect, (12) has been photographed by Lewis and von Elbe. They also suggest that because of the non-uniform potential distribution between the electrodes during breakdown., that the energy discharged in the spark may not be uniformly distributedo Variations from spark to spark would result in irregular ignition patterns. The present investigation has shown that at electrode spacings of 0.122 inches or less the discharge path of the spark is nearly the same for a series of sparks and the breakdown voltage is more uniform than for larger electrode spacingso Figure 7o13 shows a typical photograph of a spark at an electrode spacing of 0.122 inches. At an electrode spacing of 0.254 inches, extreme differences occur in the discharge path from spark to sparko This is illustrated by Figures 7.14 and 7515o Certainly, the distribution of energy along the discharge path is different for each of these sparks. This may be a reason for the randomness of ignition at the larger electrode spacings shown in Figures 7~5 and 7~9~ However, no correlation has as yet been made between the spark path and ignition. Another phenomenon was noticed in the photographs of the sparkso At large electrode spacings the spark discharge did not always start at the tip of the cathode. In many cases it started on the side of the cathode about 1/16 inch back from the tip. This could also cause random distributions of the energy between sparks. Type "C" electrodes were used in several runs to see if the irregular spark path and discharge from the side of the cathode might be improved with a pointed cathodeo No difference was noted.

-105Determination of the exact cause for the spread in the data would require a very extensive study involving a systematic variation of all of the parameters. Even with the wide transition band, the results show that increasing the rate of energy input results in a decrease of the minimum ignition energy.

0 0 io m C - A O ~0 _0.0.....o 0 -------—............................... —.

-107TABLE III. RESULTS Propane-Air Mixture )=0.83, To=79 F, P =30 in. Hg. Abs., 6 =0.091 in. Type "A" Electrodes Rn P LA Run ET Pp No. (Milloules) Watts HA Inition 225001.923 133.3.102 No 212001.881 110.8.083 No 213001 1.474 118.8.127 No 214001 1.703 107.7.134 No 224001 1.732 113.2.096 No 218001 2.016 111.0.101 Yes 217001 2.354 133.3.082 Yes 226001.964 66.7 146 No 227001 1.596 62.7.233 No 228001 1.943 60.2.187 No 229001 2.159 60.2.172 No 249001 2.070 57.0.156 No 250001 2.398 58.3.178 Yes 245001 2.416 60.0.177 Yes 241001 2.088 61.8.153 Yes 230001 2.793 60.2.193 Yes 237001 1.982 59.5.152 No 247001 2.086 59.5.164 Yes 232001 2.277 66.5.154 Yes 238001 2.327 59.4.175 No 258001.794 193.3.059 No 254001.923 224.0.044 No 263001 1.448 190.7.125 Yes 260001 1.451 183.3.104 Yes 251001 1.431 190.0.117 Yes 264001 1.340 210.0.09 Yes 266001.873 206.7.071 No 268001.807 206.7.077 Yes 255001.843 200.0.093 No 256001.760 220.0.075 Yes 267001 1.210 260.3.085 Yes 253001.878 220.0.055 Yes 259001 1.211 221.0 072 Yes 261001 1.314 210.0. 01 Yes 257001.942 215.3.064 Yes 262001 1.419 253.3.067 Yes 265001 1.37 220.0.088 Yes

-108TABLE IV. RESULTS Propane-Air Mixture g=0.83, T 79 F, P=30 in. Hg. Abs.,6=.122 Type "A" Electrodes Run ET Pp p No. (Millijoules) Watts HA Ignition 224005.877 108.8.146 No 225005 1.134 100.4.191 No 226007 1.457 113.5.201 Yes 227005 1.243 100.0.208 Yes 228005 1.026 101.0.172 No 229005 1.075 112.5.137 No 230005 1.041 95.5.344 No 231005.789 91.9.2 3 No 231105.768 95.2.147 No 232005 1.265 141.2.152 No 233005 1.360 95.2.177 No 234005 1.448 105.0.162 Yes 235005 1.192 98.7.178 Yes 236005 1.869 121.1.199 Yes 23700$5 1.986 L 996.0O.234 Yes 205005 1.673 60.5.299 No 206005 1.653 59.5.292 Yes 207005 1.285 57.8.269 No 209005.864 50.4.246 No 211005 1.096 53.0.254 No 212005.977 52,3.236 No 213005 1.379 61.6.256 Yes 214005 1.331 55.7.271 No 216005 1.517 59.0.259 Yes 217005 1.757 56.0.268 Yes 218005 1.926 52.9.336 Yes 219005.881 52.0.266 No 220005 1.358 56.5.305 No 221005 1.374.494.323 No 223005.643 83.0.133 No 225005.937 151.0.132 No 256005 1.062 158.7.126 Yes 257005.961 158.7.128 No 258005.918 189.0.079 Yes 25900o5.952 158.4.121 No 261005 1.082 177.5.136 Yes 263005.893 175.5.125 No 264005 1.046 201.7.137 No

-109TABLE IV. (Contid) Run ET p No. (Milloules) Watts HA gnition 265005 1.132 200.5.141 Yes 266005 1.307 219.7.126 Yes 267005 1.306 170.2.147 Yes 10005.... ~ 4T-7 — ---- Yes 181005 1.214 67.5.216 Yes 183005.784 80.0.188 No 184005.832 72.8.187 No 186005.587 57.5.197 No 188005.569 69.1.130 No 189005.655 105.6.075 No 190005 1.309 75.6.223 Yes 191005 1.155 70.5.206 Yes 192005.837 71.4.176 No 193005.999 78.4.173 No 194005 1.025 96..147 No 195005 1.023 102.0.160 No 196005 1.112 65.0.213 No 197005 1.392 104.5.246 No 198005 1.330 80.3.221 No 199005 1.416 76.5.221 Yes 200005 1.283 99.6.155 No 201005 1.758 87.7.229 Yes 202005 1.609 105.6.161 Yes 203005 2.388 89.8.321 Yes

-110TABLE V. REULTS Pr8pane-Air Mixture p=0.73, To=79 F, P=30 in. Hg. Abs., 6=0.195 Run T P LA No. (Milli oules) Watts HA Ignition 10014 1.832 71.9.306 No 11014 1.808 63.8.326 No 12014 1.825 62.0.44 Yes 13014 1.653 60.6.2 Yes 14014 1.787 61.3.337 Yes 15014 1.688 72.6.269 Yes 16014 1.697 67.9.257 Yes 18014 1.661 70.6.312 No 19014 1.561 71.9.275 No 20014 1.190 78.4.219 No 21014.934 67.2.175 No 22014 1.017 65.3.234 No 24014 1.018 66.0.197 Yes 28014 2.559 65.3.369 No 29014 2.344 71.9.322 Yes 0014.0714No 33014.97 112.2.152 No 34014 1.139 136.8.132 No 35014 1.333 110.9.191 No 36014 1.189 172.0.124 No 37014 1.592 152.4.142 No 38014 1.267 110.9.158 Yes 39014 2.086 139.2.161 Yes 40014 1.416 142.8.117 Yes 41014 1.646 121.6.186 No 42014 1.530 116.4.152 No 43014 1.886 108.0.218 No 44014 1.990 125.7.198 Yes 45014 2.039 128.1.207 Yes 46014 2.35 137.1.193 Yes 47014 2.598 131.5.218 Yes 48014 1.459 124.8.178 No 49014 1.775 128.0.187 Yes 50014 1.141 105.0.149 No 4610'14 _ 2.835 _ 101.6 _.113 es No 7014 2.58 -6 1981.23014 Yes 2014 2.329 90.0.373 No 3014 1.866 98.1.306 No

-111TABLE V. (Conttd) RunTpE P pLA Run ET p Fp t No. (Miliule) Watts HA Ignition 4014 1.808 99.7.301 No 6014 2.187 108.8.230 Yes 7014 2.189 92.4.250 Yes 8014 1.48060.7.327 No 9014 1.58 64.5.339 No 25014 1.531 65.3.259 No 54014.800 165.3.065 N 55014.651 162.7.062 No 56014 1.037 185.8.105 No 57014 1.112 192.7.087 No 59014 1.399 202.4.104 Yes 60014 1.564 189.3.131 Yes 61014 1.469 127.5.182 Yes 62014 2.250 186.7.175 Yes 63014 1.526 165.0.156 No 64014 1.604 186.7.120 Yes 65014 1.19 195.5.110 No 66014.824 173.3.073 No

-112TABLE VI. RESULTS gropane-Air Mixture 0=.69, T =79 F, P=30 in. Hg. Abs., 6 =.254 Type "A"l Electrodes Run ET Pp No. (Milijoules Watts HA Ignition 31017 1.965 82.6.232 No 32017 2.251 100.5.210 No 33017 2.176 105.6.217 No 34017 2.305 97.6 264 No 35017 2.590 102.7.225 No 36017 2.874 86.5.264 No 37017 3.343 111.9.188 Yes 38017 2.908 145.6.167 No 39017 2.850 167.2.1242 No 40017 3.070 173.1.141 No 41017 4.392 133.3.224 Yes 42017 3.38 167.0. 143 No 43017 3.374 171.9.162 No 44017 3.319 162.2.146 No 45017 4.262 122.7.223 Yes 46017 4.210 120.0.229 No 48017 4.132 117.3.214 Yes 49017 3.741 168.5.140 Yes 50017 3.516 171.6.121 No 51017 3.544 175.4.121 Yes 52017 3.510 170.1 115 Yes 53017 3.763 170.1.119 Yes 54017 3.356 165.7.140 No 55017 2.928 170.1.142 No 56017 2.919 162.8.152 No 57017 2.402 138.6.208 No 59017 3.944 173.3.107 Yes 60017 1.023 229.5.057 No 61017 1.155 245.7.048 No 62017 1.384 283.3.068 No 63017 1.595 215.6.099 No 64017 2.088 329.3.079 No 65017 2.357 306.0.100 No 66017 2.456 307.2.118 Yes 67017 2.352 333.2.106 No 68017 2.393 236.1.135 Yes

-113TABLE VI. (Cont Id) Run ETp No. (Milli joules) Watts HA Ignition 69017 3.224 306.0.172 Yes 70017 2.657 272.3.129 Yes 71017 3.135 312.6.120 Yes 72017 3.309 290.1.112 Yes 73017 3.5 44 313.5.102 Yes 74017 3.638 266.0.121 Yes 75017 4.438 321.0.146 Yes 76017 2.552 329.9.098 No.77017 2.025 93.8.249 No 78017 2.070 107.0.238 No 79017 2.160 103.0.245 No g0017 2.270 85.0.276 No 81017 2.540 107.4.236No 82017 2.660 100.0.227 No 83017 3.20 106.0.220 Yes 84017 3.070 106.0.203 Yes 35017 3.150 103.0.232 Yes 86017 3.316 77.4 -299 Yes 87017 3.370 74.7.330 Yes 88017 3.37 75.0.347 Yes 89017 3.214 81.8.320 Yes 90017 2.525 105.7.233 Yes 91017 2.268 107.1.228 Yes 92017 2.280 107.1.235 Yes 93017 2.060 104.1.235 No 94017 2.038 84.3.320 No 1017 4980 95.4.171 Yes 2017 4.660 86..195 No 3017 6.743 107.3.242 Yes 4017 4.652 74.8.223 Yes 5017 4.929 79.3.200 Yes 6017 4.943 88.0.200 Yes 7017 4.961 76.9.249 No 8017 4.318 77.3.228 No 9017 4.387 93.00.171 No 10017 3.975 94.70.153 Yes 110174 066 95.20.152 Yes 12017 4.499 78.70.234 Yes 13017 3.823 91.3.242 No 15017 4.323 83.8.24Yes 16017 3.536 100.3.1 7 No 14017 3.6814 105.6.200 Yes 17017 2.759 80.0.268 Yes 19017 3.197 74.0.316 Yes

-114TABLE VI. (Cont d) Run ET LA No. (Mlllijoules) Watts HA Igition 20017 2.528 106.2.235 Yes 21017 1.309 93.9.136 No 22017 1.264 99.7.156 No 23017 1.00 95.1.201 No 24017.862 67.2.191 No 26017.742 94.4.08 No 27017.955 73.5.182 No 28017.541 76.3.083 No 30017 1.676 104.3.177 No

VIII. CONCLUSIONS AND RECOMMENDATIONS Ao Conclusions 1o For the initial mixture conditions used in this study, minimum ignition energy decreases by approximately 40 percent as the rate of energy input increases by approximately 100 percent. 2, The minimum ignition energy can be correlated with the peak power according to the semi-empirical equation EMN P v= F 3. The value of v increases with decrease in equivalence ratio. 4o The transition band between non-ignition and ignition becomes wider as the equivalence ratio decreases. 5o The transition band decreases in width as the rate of energy input increases. 6o Even under carefully controlled conditions in the constant volume bomb, variations in energy and rates of energy input requirements are appreciable. 7. Information derived from a plot of PLA/PHA versus total energy would be indicative of the adequacy of a conventional engine ignition system for lean mixtures. 80 A plot of PLA/PHA versus total energy can be used to obtain the minimum ignition energy for a pure capacitive spark. Bo Recommendations Li Future Investigation a, To determine the cause of the wide transition band, a complete study involving changes of all parameters should be undertakeno -115

-116Some of the more important parameters appear to be; fuel type, electrode shape, electrode material and electrode temperature bo It appears that experimental data on quench distances at high pressures is lacking in the literature. Therefore, quench distances for higher pressures (i.e., up to 150 psi) would be a contribution. Co The effect of the electrode flange size upon quenching should be investigated, do Effects of the rate of energy input upon the minimum ignition energy should be studied at higher pressures and temperatures. e. A study of the influence of diluents upon the relationship between the rate of energy input and minimum ignition energy should be made. f. An exploratory investigation of the energy distribution and molecular makeup of the spark during ignition appears to be a fruitful area of investigation. One possibility would be the use of emission spectroscopy with photomultiplier detectors, 2o Equipment Modification a. It appears that a decrease in the data reduction time can be accomplished through the use of a diode multiplier. Thiis device could be used for multiplication of the voltage and current during the discharge to produce a power versus time

-117trace on the oscilloscope. Then the area under the trace on the photographic record could be measured with a planimeter to obtain the energy. b. Improved data during the initial phase of the discharge could be obtained if two oscilloscopes are used simultaneously. One oscilloscope sweep rate could be set at a fast rate to record the initial phase of the high rate component and the other at a slow rate to record the low rate componento c. Use of a storage type oscilloscope would reduce the quantity of photographic records during equipment development and data operation.

APPENDIX A DERIVATION OF PRE-BREAKDOWN CURRENT EQUATIONS 1. Derivation of Equation (3.4), from Reference 59 Assume that a discharge is taking place at a constant applied voltage in Figure Alo Consider the differential element dx o Electrons will generate n odx electrons and ions per second while passing through dxo dn_ = on.dx (A-l) and n = no when x = O Similarly dn+ = -on-dx (A-2) where n = 0 at x = Integrating Equations (A-l) and (A-2) and multiplying by the electron charge and electrode area gives i_ = io e (A-3) i, = io (e -e ) (A-4) The total current is expressed as follows: i = i_ + i+ (A-5) -118

-119n. ^\ n+.dn. nr1dn, CATHODE ANODE dx -Fige A.1. e-Breakown Cel Figure A.1. Pre-Breakdown Current Model.

-120Therefore: i = i e a (3.4) 2, Derivation of Equation (3.6) from Reference 59 According to Townsend, each electron produces a ions and electrons per centimeter of pathlength. Each positive ion produces B ions and electrons per centimeter of pathlength. Therefore, under steady conditions of Figure A.1 the following equations apply. dn_ = (o_ + pn+)dx (A-6) or dn. - = on_ + pn+ dx also dn+ d- = On + fPn+ (A-7) dx Therefore: dn dn+ + = 0 dx dx Since n = n_ + n+ dn = (-)n+ n (A-8) (c- )nn + Pn dx Upon integration of Equation (A-8): n = _ _ n + De(-) (A-9) - aHere again n = no at x = 0

-121Therefore: D = n + - n (A-10) ~ a~Also since n = n_+n+ n+ n - D e(n-)x (A-ll) and n+ = 0 at x = Thus: D e(-)5 = _ —- n (A-12) Substitution of Equation (A-10) into Equation (A-12) and multiplying by the electron charge and electrode area yields Equation (3.6) i = e io() (a: f) (3.6) 0a-P e(a-C)6 3. Derivation of Equation (3.8) from Reference 49 Current at the anode is expressed as: i = (io + i ) e06 (A-13) where io is the photocurrent and i+ is the number of electrons ejected from the cathode due to positive ion bombardment and is expressed as follows: i = 7[i - (io + i)] (A-14) y is the number of secondary electrons generated per primary electron which causes an avalanche in the gap,

-122Elimination of i+ in Equation (A-13) and (A-14) gives: ea6 i = i (3.8) 1 - y(e6 - 1) 4. Derivation of Equation (3.9) from References 64 and 65 The total number of electrons emitted at the cathode is: n' = no + TZ (A-15) where no is the number of electrons emitted per second due to photons impinging from outside the system. Z is the number of photons generated in the gas by electrons. r is the fraction of photons which produce electrons which leave the cathode. Therefore, the number of photons produced in an element dx at a distance x from the cathode and available at the cathode is: dz = (n' + nc) gG e" dx (A-16) where nc is the number of new electrons created by collision between the cathode and element dx. G is the number of photons produced by an electron per centimeter of path length in the field direction. p. is the average absorption coefficient of the photons by the gas. g is the geometric factor which represents the fraction of photons which reach the cathode. The number of new electrons produced in the element dx by collision is: dnc = (nO + nc) adx (A-17)

-123where a is the number of ions produced per centimeter of travel, and nc = n at x = 0 By integration of Equation (A-17) the following result is obtained: nc = n (ex - 1) (A-18) Then (a-.)x dz = ntg 9 e dx (A-19) In this case Z = 0 at x = 0 Upon integration of Equation (A-19), the number of photons generated in the gas by electrons is expressed as follows: Z = ne (a-) 1 (A-20) By substitution of Equation (A-20) into Equation (A-15), the total number of electrons emitted at the cathode is expressed by the following: = no (A-21) 1 _ g-lg [e(-a4)x - 1 The number of electrons which reach the element dx at steady state are: n = n' + no (A-22) Substitution of Equation (A-18) into Equation (A-22) gives: n = n1 eCx (A-23)

-124Finally by substitution of (A-21) into (A-23) gives: (a- )e= _ n = no (3o9) (a-1) - l9g [e(t-4)xj which is Equation (3.9) when n and no are both multiplied by the electron charge and electrode areao

APPENDIX B DETERMINATION OF INTERELECTRODE CAPACITANCE OF A COIL Stout(86) discusses a technique which he feels is the most reliable for the determination of the capacitance of a coil such as is shown in Figure B.1. C L R Figure B.1. Equivalent Circuit of a Coil. It can be shown that the following equation applies to this circuit. _ = 1 _2C (B-l) Le L where Le = equivalent inductance L = zero frequency inductance Cl = 2Trf f = frequency C = capacitance The procedure for obtaining C is to measure Le as a function of CD using an alternating current bridge. Then plot L eL versus cu and the slope of the resultant straight line is equal to -125

-126the interelectrode capacitance of the coil. The zero frequency inductance is determined by the L intercept of the straight line at zero frequency. Le Tables VII and VIII respectively include the data taken for the secondaries of the Mallory and Delco coils. Figure B 2 is a plot of the data. The results are as follows along with resistances of the secondary coils. Mallory: C = 53.5 pf; L = 55 henry 0 R = 8600 ohms Delco: C = 29o1 pf; Lo = 53.6 henry R = 9900 ohms The following is a derivation of Equation (3-i)o. The equivalent impedance is 1 1 1 -jXc+R+jXL + = (B-2) Ze R+jXL -jXc jXc(R+jXL) which can be reduced to the following R+j LL(l-wLC)-R R2C Ze wflS2R2c2+(l-L2LC)2 B ) Then taking the imaginary part for the capacitive reactance Xe -CD L(1- 2LC)-R2C e D2R2C2+ (l-c2LC)2 Since R2C K< L, neglect R2C terms then Le (1l-cLC) Xe - 7I i(B- 5)

-127And L Le. ^ -..L (B-6) 1 - c2 LC Finally 1 1 2C. ( Le L (B-)

2.4 2.0 -- 1.6 ___ ___ ______ DELCO r 0 X10 Figure B'2. DeterMALLORY 0.4 0 4 8 12 16 20 24 28 32 36 2 -7 CW xlO Figure B,.2. Determination of Coil Capacitance.

APPENDIX C DESIGN REQUIREMENT FOR A VOLTAGE DIVIDER Reference 87 describes a method of construction of a voltage divider for high voltages. If the voltage divider is to operate properly at high frequencies, the divider must be frequency compensated. The following equation must be satisfied if a voltage divider is properly frequency compensated where the distributed capacitance C' of the high voltage section is neglected. R1C1 = ReCe (C-1) where Re = equivalent resistance of the low voltage section Ce = equivalent capacitance of the low voltage section Derivation of Equation (C-l) follows. The capacitive reactance is expressed by the following equation c = 12 (C-2) c 2irfc Therefore: Xlc = 2fc-3) X (c-4) 2c- 2rtfc2 Xoc = 2tfc (c-5) -129

-130The equivalent impedances are 1 1 + (c-6) Z1 R1 Xlc 1 1 j - = (C-?) Z R2 + (c-7) Z2 R2 X2c 1 1 + - (c-8) Zo Ro XOC and 1 1 1 1 1 -+ z+ j ( —+ R-) (c-9) Ze R2 Ro X2c Xoc Therefore: RoR2 (C-10) Re (c-) Ro + R2 and ec 2Tf(C2+C) (-l or Ce = Co + C2 (C-12) The current in both parts of the voltage divider must have the same magnitude and phase angle if the output is to be a fixed proportion of the input signal. The necessary condition is that the phase angles in the two parts of the circuit be equal. That is, tan e = tan Xec (c-13) R1 Re Substitution of Equation (C-ll) and (C-10) into (C-13) and solving:

-131R1 = eC = R2R (C2 + Co) (c-14) 11 e e R2+R0 Therefore, the capacitor C2 is made adjustable for frequency compensation. Correct adjustment of C2 is accomplished by applying a square wave signal to the input, viewing the oscilloscope output, and then adjusting until the output waveform is a square wave. In voltage dividers which have large attenuation ratios the resistor R1 will have an appreciable amount of stray capacitance C1 distributed along its length as shown in Figure C.1. The resultant output waveform after compensation is like that shown in Figure C.2 due to the distributed capacitance for which compensation has not been made. Compensation for this stray capacitance is a most difficult problem and involves a trial and error procedure. Usually addition of a compensation capacitance between R and point a as shown by 1 C1 in Figure C.1 will yield satisfactory resultso This final compensation should cause the output waveform to be a perfect square wave.

-132I c RaI c' I R2 C, I I -- I — Figure C.. Voltage Divider Circuit. Figure C.2. Voltage Divider Output Incompletely Compensated for Distributed Capacitance.

APPENDIX D CIRCUIT AND CALIBRATION DATA -133

240-. —. 160 w I1. 120_ UJ w H 40 / _ ca__ _ __ __r _ IRON-CONISTANTAN f 0 80 80 THERMOCOUPLE) IN CONJUNCTION WITH BROWN POTENTIOMETER 40 0 40 80 120 160 200 240 TEMPERATURE READING, F Figure D.1. Thermocouple Calibration Curve.

-135TABLE VII SECONDARY CAPACITANCE DATA FOR MALLORY COIL Model F-12T Coil, Primary Open f. ao2 Le 1/L. cps X10-3 X10-7 (henry) XIO02 400 2.51.63 55.8 1.79 1000 6.38 4.07 64.0 1.56 1400 8.80 7.75 70.0 1.43 1800 11.30 12.80 87.1 1.15 2000 12.60 15.90 102.0.98 TABLE VIII SECONDARY CAPACITANCE DATA FOR DELCO COIL Model 115251 Coil, Primary Open f Le 1/Le cps X10-3 X10-7 (henry) X10O2 500 3.15.982 55.2 1.81 700 4.40 1.940 55.5 1.80 900 5.66 3.200 56.5 1.77 1100 6.91 4.780 57.8 1.73 1300 8.18 6.690 60.2 1.66 1500 9.43 8.900 63.1 1.58 1700 10.70 11.450 66.4 1.51 1900 11.95 14.300 69.0 1.45 2100 13.20 17.430 73.7 1.36 2300 14.50 21.000 79.2 1.26 2500 15.72 24.700 86.9 1.16 2700 17.00 28.900 97.3 1.03 2900 18.24 33.30 112.2 0.89

-136TABLE IX SPARK CONTROL CIRCUIT DATA RA = 50, 5w C1 = 0o0054f RB = 130Q, lOw C2 = 0.Ol4f R2 = 820Q, 0.5w C3A l= 1o0f R3 = 0-5K, 10 turn precision pot. C3B = 2.0tf R4 = 5K, 0.5w C4 = 1.0l f R5 = 60, lw C5 = variable capacitance R6 = 50Q, 0.5w D1A = DlB = Type lN100 R7 = 1M, 0.5w QT Q2 = ype 2N1481 R8 = 3.3K, 0.5w Q3A 3= = Type 2N1099 R9A = 60o, lOw 4 = Type 2N1183B R 6n= 602, lOw Z = Type 1N3008B 9B ROA = 10B = 30, lOw T = Type TA-7 (Stancor) Rll = 680Q, 0.5w V = 12 volts R12 = 2.7, 2w V = 6 volts R13 = o-6Q, lOw V3a 3b = 3 volts R14 = 0-4Q 125w V = 6 volts 14 4 R15= 0-4Q, 125w R16 = variable resistance R17 = 1502, + 1%, lw

-137TABLE X COUNTER CALIBRATION DATA ERIE MODEL 400 COUNTER Count ing Time Interval Counter Correct Percentage (Minutes) Reading Reading Error 40 2399X105 2400X105 0.04 40 2400X105 2400X105 0 45 26998X104 27000X104 0.0074 3000X104 3000X104 0 5 29987X103 30000X103 0.04 5 30005X103 30000.016 5 29999X103 30000X103 0.003 TABLE XI. OSCILLOSCOPE SWEEP RATE DATA Time Oscilloscope Actual Actual Marker Sweep Rate Horizontal Corrected Output Setting Deflection Sweep Rate 1 msec./cm. 1 msec./cm. 1.00 cm. 1. msec./cm. 1 msec./cm. 0.5 msec./cm. 1.87 cm. 53.5 psec./cm. 1 msec./cm. 0.2 msec./cm. 4.85 cm. 20.5 pLsec./cm. 100 ssec./cm. 100,sec./cm. 1.00 cm. 100.0 vsec./cm. 100 Vsec./cm. 50 Lsec./cm.X2 Mag. 1.87 cm. 53.5 psec./cm. 100 lsec./cm. 50 psec./cm. 3.73 cm. 21.1 osec./cm. 100 psec./cm. 20 jLsec./cm. 4.86 cm. 20.5 psec./cm. 10 ksec./cm. 10 ~sec./cm. 1.00 cm. 10.0 ~sec./cm. 10 psec./cm. 20 isec./cm.X2 Mag. 0.96 cm. 10.4 psec./cm.

-138TABLE XII. PRESSURE GAGE CALIBRATION DATA Gage age Average Reading Reading Gage Average Correct Psi(Pressure Psi(Pressure Reading Deviation psi Decreasing) Increasing) psi psi (Sp) C. L. 01, U. S. Gauge Co., 10861, Supergauge 0-100 psi, 1 psi increments 5.0 5.2 5.0 5.1 + 001 10.0 10.0 9.7 9.85 - 0 15 15.0 15.2 14.6 14.90 - 0.10 20.0 20.2 19.7 19.95 - 0.05 25.0 25.2 24.5 24. 85 - 0.15 30.0 30.3 29.7 30.00 0.00 3I.0 35.3 34.7 35.00 0 00 40.0 40.4 39.7 40.05 + 0.05 45.o 45.4 44.3 45. to + 0.10 50.0 50.3 49.3 50,05 + o.05 j5.0 55.4 54.8 5.o10 + 0,10 60.0 60.3 59.' 60.05 + 0.05 65.0 65.2 64.8 65.00 0 00 70.0 70.2 69.8 70.00 0.00 75.0 75.3 74.8 75.05 + 0.05 80.0 80.2 79.8 80.00 0.00 85.0 85.1 84.8 84. - o05 90. C 90.0 39. 89 90 - 0. 10 9.0 95.2 94.3 95.00.o 0.00 100.0 99.9 99.7 99.30 - 0.20 C. L. 02, Lonergan Maximoni Gauge, 0-300 psi, S psi /iv, 15.0 16.5 16.0 16.25 + 1.25 20.0 21.5 21.0 21.25 + 1.25 25.0 26.0 25.5 25.75 + 0.75 30.0 31.0 30.0 30.50 + 0.50 35.0 35.5 34.5 35.00 0.00 40.0 40.0 39.0 39.75 - 0.25 45.0 45.o 44.0 4.5 - o.50 50o.o 49.0 49. 48.25 - 0.75 55.0 54 54.0 4.o -.00oo 60.0 590 58.58.75 - 1.25 65.0 64.5 64.0 64.25 - 0.75 70.0 69.0 68.0 68.50 - 1.50 75.0 74.00 74.0 7.00 - 1.00 80.0 78.5 78.5 78.50 - 1.50 85.0 83.5 83.0 83.25 - 1.75 90.0 88.0 88.5 88425 - 1.75 95.0 93.0 93.5 93.25 - 1.75

-139TABLE XII. (Cont d) Gage Gage Avrerage-' Reading Reading Gage Average Correct Psi(Pressure Psi(Pressure Reading Deviation psi _ Decreasing.) Increasing) p p si (AP) 100.0 98.0 98.0 98.00 - 2.00 105.0o 102.5 12.5 102.50 - 2.50 115.0 112.5 1122.5 11250 - 2.50 120.0 117.5 117.5 117.50 - 2.50 125.0 122.5 123*0 122.75 - 2.25 130. 0 1270 127.5 127.25 - 2.75 135.0 132.0 132.0 132.00 - 3.00 140.0 137.0 137.0 137.00 - 3.00 14.5.0 142. 142.0 142.00 - 3.00 150.0 147.0 147.0 147.00 - 3.00 155.0 152.0 152.0 152.00 - 3.00 160.0 157.5 157.5 157.50 - 2.50 165.0 162.0 162.5 162.25 - 2.75 170.0 167.0 167.5 167.25 - 2.75 175*0 172.0 172.0 172.00 - 3.00 180.0 177.0 177.0 177.00 - 3.00 185.0 1820 182.0o 182.00 - 3.00 190.0 187.0 187.0 187.00 - 3.00 195.0 192.0 192.0 92.00 - 3.00 200o0 970 9719700 197.00- 3.00 205.0 202.5 202.5 202.50 - 2.50 210.0 207.5 207.5 207.50 - 2.50 215.02 2125 212.5 212.50 - 2.50 220.0 217.5 217.0 217.25 - 2.50 225.0 223.0 2230 223.00 - 2.00 230.0 228.0 228.0 228.00 - 2.00 235.0 233.0 233.0 233.00 - 2.00 240.0 238.0 238.0 238.00 - 2.00 245.0 243.0 243.0 243.00 - 2.00 250.0 248.0 248.0 248.00 - 2.00 255.0 253.0 253.0 253.00 - 2.00 260.0 258.0 258.0 258.00 - 2.00 265.0 263.5 263.5 263.50 1.50 270.0 268.5 268.5 268.50 - 1.50 275.0 274.0 274.0 274.00 1.00 280.0 279.0 279.0 279,00 - 1.00 285.0 2840 240 284.00 - 1.00 290.0 289.5 289.5 289.50 - 050 295.0 295.0 295.0 295.00 0.00

TABLE XIII. FUEL-AIR MIXING INFORMATION * Lean Limit Ref. 26 /Pf Pf //in.Hg. A/F A/F F/A F/A r in for ji ___ (By Vol. (ByWt.) (By Vol.) P(B Wt,)?ercent T_ PT=200psia 1.00 23.80 15.66.0420.0639 100.0.0403 16.46 0.98 24.30 15.98.0411.0626 102.0.0395 16.10 0.96 24.80 16.30.0403.0613 104.1.0388 15.84 0.94 25.30 16.65.0395.0600 106.3.0380 15.50 0.92 25.86 17.00.0386.0588 108.7.0372 15.18 0.90 26.45 17.40.0378.0575 111.0.0365 14.90 0.88 27.04 17.78.0370.0563 113.6.0357 14.57 0. 86 27.68 18.20.0361.0550 116.2.0349 14.25 0.84 28.35 18.63.0353.0536 119.0.0341 13.92 0.82 29.00 19.10.0345.0523 122.0.0333 13.60 0.80 29.75 19.57.0336.0511 125.0.0325 13.26 0.78 30.52 20.05.0328.0499 128.2.0318 13.00 0.76 31.32 20.60.0319.0485 131.6.0310 12.66 0.74 32.20 21.15.0311.0473 135.0.0302 12.33 0.72 33.05 21.73.0325.0460 139.0.0293 11.96 0.70 34.00 22.35.0294.0448 143.0.0285 11.63 0.68 35.00 23.00.0286.0435 147.0.0278 11.36 0.66 36.05 23.70.0278.0422 151.5.0270 11.00 0.64 37.20 24.45.0269.0409 156.2.0262 10.70 0.62 38.40 25.25.0260.0396 161.2.0254 10.37 0.60 39.70 26.10.0252.0383 166.6.0246 10.05 0.58 41.00 27.00.0244.0370 172.4.0238 9.72 0.56 42.50 27.95.0235.0358 178.6.0230 9.39.54.06 2900.022.0 13..0222 9.08. 3. 20 29.10.0226 34 16.0.0221 90 0. 52 45 O 30.C-01-.021b 0-333. 192.3.0214. 74

APPENDIX E ORIGINAL DATA 1411

TABLE XIV. EXPERIMENTAL MIXTURE DATA ^.. i i ^ k k r-** -- ~> <D oo +4-3 4 )4 4zO rl F~~r k~C~ ad r - l ^bO * 0 r * H H 0 O O k a P r4 F-4 i"k i * D CJ <PbO~ X o o H 0 0. 0;4 "H Erip 0 05r4F aS 0 e- 0 0 H -40 ~~~ o 0 H~~~~~~~~p E E E-4 PiOP 1 A 0.84 29.29 71 29.17 1)4.33 14.88 200. 125 67.0 69.5 202.5 216.8.0336 0.83 5 A 0.84 28.95 72 28.83 14.12 13.92 135.9 115 70.0 72.2 188.9 203.0.0336 0.83 14 A 0.74 29.09 82 28.84 14. 14 12.33 185.9 200 79.6 82.4 188.9 203.0.02970.73 17 A 0.70 29.28 75 29.15 14.3 11.63 185.7 50 73.2 75.6 188.7 203.0.0281 0.69

TABLE XV. EXPERIMENTAL IGNITION DATA s0.83 6= 0.091 inches Barometer ScoPe Setti s Tante ----- iivyePo T 6 3w. a'teConstant ~Run p ~Evac. in. T I Kvy. sec v. Settings IgniNo. in.H. F Press. HE. F Amps. cm. cm. cm.' I t C tion Mallory Coil + 174 K ohms and 50 Pf added through 77 K ohms 226001 29.55 72 130 30 80 5.0 5.0 53.5 5.0 0.5 1.0 No 227001 29.55 72 130 30 80 5.0 5.0 53.5 5.0 0.75 10 No 228001 29.55 72 105 30 80 5.0 5.0 100.0 5.0 1.00 1.0 No 229001 29.55 72 120 30 80 5.0 5.0 1000 5.0 1.25 10 No 230001 29.55 72 110o 30 8o 5.0 5.0 0loo0o0 5.0 1.50 1,0 Yes 232001 29.55 72 110 30 80 5.0 5.0 100.0 5.0 1.35 1.0 Yes 237001 29.55 72 130 30 80 5.0 5.0 100.0 5.0 1.30 1.0 No 238001 29.55 72 130 30 80 5.0 5.0 100 0 5.0 1.35 10 No 241001 29.55 72 130 30 80 5.0 5.0 100.0 5.0 1.40 1.0 Yes 245001 29.51 76 150 30 80 5.0 5.0 100.0 5.0 1.35 1.0 Yes 247001 29.51 76 150 30 80 5.0 5.0 10 5.0 1.30 1.0 Yes 249001 29.51 76 135 30 80 5.0 5.0 100.0 5.0 1.30 1.0 No 250001 29.51 76 140 30 80 5.0 5.0 100.0 5.0 1.35 1. 0 Yes Mallory Coil + 174 K ohm and 50 Pf added through 43.5 K ohm 212001 29.54 72 110 30 80 5.0 5.0 53.5 10.0 0.5 1.0 No 213001 29.54 72 120 30 80 5.0 5*0 53.5 10.0 0.75 1o No 214001 29.52 72 110 30 80 5.0 5.0 100.0 10.0 1.00 10 No 216001 29.52 72 130 30 80 5.0 5.0 100.0 10. 1.25 1.0 Yes 217001 29.52 72 125 30 30 5.0 5.0 100.0 10.0 1.25 1.0 Yes 218001 29.52 72 120 30 80 5.0 5.0 100.0 10. 1.125 1 0 Yes 224001 29.52 72 125 30 80 5.0 5.0 100.0 10.0 1.06 1.0 No 225001 29.52 72 125 30 80 5.0 5.0 100.0 10.0 0.30 1.0 No

TABLE XV. (Conttd)'s'Scope Set t igs Time Barometer T VConstant Run Temp. Evac. I ~r Sus~~te~T onstant Run Temp- Evac. in. o Kv. U sec. v. Settings IgniNo. _ in.Hg. F Press. Hg. F Amps. cm. cm. cm. R C tion Mallory Coil + 174 K ohm and 50 Pf added through 18 K ohm 251001 29.41 76 130 30 80 50 5.0 53.5 10.0 0.5 1.0 Yes 253001 29.41 76 140 30 80 5.0 50 53.5 10.0 0.3 1.0 Yes 254001 29.51 76 140 30 80 5.0 5.0 53.5 10.0 0.25 1.0 No 255001 29.41 76 140 30 80 5.0 5.0 53.5 10.0 0.30 1.0 No 256001 29.41 76 145 30 80 5.0 5.0 53.5 10.0 0.30 1.0 Yes 257001 29.41 76 150 30 80 5.0 5.0 53.5 10.0 0.20 1.0 es 258001 29.41 76 150 30 80 5.0 5.0 53.5 10.0 0.18 1.0 No H 259001 29.41 76 150 30 80 5.0 50 53.5 10.0 030 1.0 Yes 260001 29.41 76 150 30 80 5.0 5.0 53.5 10.0 0.40 1.0 Yes 261001 29.41 76 150 30 80 5.0 5.0 53.5 10.0 0.35 1.0 Yes 263001 29.40 76 150 30 80 5.0 5.0 53.5 10.0 0.3 1.0 Yes 264001 29.40 76 150 30 80 5.0 5.0 53.5 10.0 0.30 1.0 Yes 266001 29.40 76 155 30 80 5.0 5.0 53.5 10.0 0.30 1.0 No 267001 29.40 76 150 30 80 5.0 5.0 53.5 10.0 0.35 1.0 Yes 268001 29.40 76 150 30 80 5.0 5 53.5 10.0 0.18 1.0 Yes

TABLE XVI. EXPERIMENTAL IGNITION DATA o =0.83; 6= 0.122 inches Barormeter Scope Settins Time ----- mp T Sw.Rate~I Constant Run ~pEvac. in 0 o 0o Ky pxsec. v. Settings IgniNo. in.H g. F Press. Hg F Amps. cm. cm. cm. R C tion Mallory Coil + 174 K ohm 180005 29.04 73 135 30 81 5.0 2.0 100.0 20 0.4 1.0 Yes 181005 29.04 73 130 30 80 5.0 2.0 53.5 2.0 0.3 10 Yes 182005 29.04 73 13$ 30 80.0 2.0 20.5 20 0.2 10 No 183005 29.05 7 135 30 80 5.0 2.0 20,5 2.0 0.2 10 No 184005 29.05 75 140 30 80 5.0 2.0 20.5 2.0 0.2 1.0 No 186005 29.05 74 150 30 80 5.0 2.0 20. 2.0 0.15 1.0 No 188005 29.05 74 135 30 80 5.0 2.0 20.5 2.0 0.15 1.0 No 189005 29.05 74 135 30 80 5.0 2.0 20.5 2.0 0.2 1.0 No 190005 29.05 75 140 30 80 5.0 2.0 20,5 2.0 0.3 1.0 Yes 191005 29.05 75 150 30 80 5.0 2.0 20.5 2.0 0.2 1.0 Yes 192005 29.05 75 135 30 80 5.0 2.0 20.5 20 0.22 1.0 No 193005 29.05 75 145 30 80 5.0 2.0 20,5 20 0.23 1.0 No 194005 29.05 75 150 30 80 5.0 2.0 20.5 2.0 0.24 1.0 No 195005 29.05 7 30 80 5.0 2.0 20.5 20 0.25 1.0 No 196005 29.05 75 150 30 80 5.0 2.0 20.5 2.0 0.27 1.0 No 197005 29.13 75 150 30 80 5.0 2.0 20.5 2.0 0.30 1.0 No 198005 29.13 75 145 30 80 5.0 2.0 20.5 2.0 0.35 10 No 199005 29.13 75 140 30 80 5.0 2.0 20.5 2.0 0.4 1.0 Yes 200005 29.13 75 135 30 80 5.0 2.0 20.5 20 0.3 10 No 201005 29.13 75 135 30 80 5.0 2.0 20,5 20 0.5 1.0 Yes 202005 29.13 75 135 30 80 5.0 2.0 20.5 2.0 0.4 1.0 Yes 203005 29.13 75 135 30 80 5.0 2.0 53.5 2.0 0.6 10 Yes

TABLE XVI, (ContId) - Scpe S~~~~~~~e~~Ft~~~tdffs~ins LTme BarometerScpe SetnTim pV Sw.Rate I C ons tant Run Temp. Evac. in. T o Ky sec. V. Settings IgniNo, in.Hg. F Press. Hg. F Amps. cm. cm. cm. R C tion Mallory Coil + 256 K ohm * 25 Pf added through 125 K ohm 204005 29.29 74 120 30 80 5.0 2.0 20.5 2.0 0. 1.0 Yes 205005 29.29 74 125 30 o 5.0 2.0 20.5 2.0 0.4 1.0 No 206005 29.29 74 125 30 80 5.0 2.0 20.5 20 04 1.0 Yes 207005 29.29 74 130 30 80 5.0 2.0 20.5 2.0 03 10 No 209005 29.29 74 135 30 80 5.0 2.0 20.5 20 0.2 1.0 No 211005 29.29 74 160 30 80 5.0 2.0 20.5 2.0 025 1.0 No 212005 29.29 77 145 30 80 5.o0 2.0 20.5 2.0 0.25 10 No 213005 29.42 69 9 30 80o 5.0 2.0 20.5 2.0 0.35 10 Yes 214005 29.42 7~ 98 30 80 5.0 2.0 20.5 2.0 0.3 1.0 No 215005 29.42 70 115 30 80 5.0 2.0 20.5 2.0 0.4 10 No 216005 29.42 70 115 30 30 5.0 2.0 20.5 2.0 0.4 1.0 Yes 217005 29.42 70 125 30 80 5.0 2.0 20.5 2.0 045 10 Yes 214300 29.42 70 120 30 80 5.0 2.0 53.3 20 055 1 0 Yes 219005 29.41 73 120 30 80 5.0 2.0 20.5 2.0 0.2 1.0 No 220005 29.41 73 125 30 80 5.0 2.0 20.5 2.0 0.3 1.0 No 221005 29.41 73 125 30 80 5.0 2.0 20.5 20 0.35 1.0 No 223005 29.40 74 140 30 80 5.0 2.0o 20.5 2.0 18 1.0 NoMallory Coil + 256 K ohm and 50 Pf added through 125 K ohm 224005 29.40 74 135 30 80 5.0 2.0 20.5 20 02 1.0 No 225005 29.40 74 135 30 81 5.0 2.0 20.5 2.0 0.25 10 No 226005 29.40 74 135 30 80. 2.0 20.5 2.0 0.30 1.0 Yes 227005 29.40 74 140 30 80 5.0 2.0 20.5 2.0 025 1.0 Yes 228005 29.40 74 150 30 79 5.0 2.0 20.5 20 0.22 1.0 No 229005 29.40 74 150 30 80 5.0 2.0 20.5 2.0 0.22 1.0 No

TABLE XVI. (ContId) _' " "' - ~'~~Scope Settinas Time Baro met~e r Barometer-Sw Rat e " I C ons t ant Temp. T IS Run Evac, in 0 o Kv ec. V Settings Igni0 0 No. in.Hg. F Press. Hg FF AmpAs.o cm.___cM.' CM.'_SR C tion Mallory Coil + 256 K ohm and $0 Pf added through 125 K ohm 23000$ 29.40 74 140 30 80 5.0 2.0 20.5 2.0 0.20 1.0 No 231005 29.40 74 140 30 80 5.0 2.0 20.5 2.0 0.18 10 No 231105 29.40 74 140 30 80 5.0 2.0 20.5 2.0 0.18 10 No 232005 29.40 74 145 30 80 5.0 2.0 20.5 2.0 0.27 10 No 233005 29.40 74 145 30 80 5.0 2.0 20.5 2.0 0.27 1.0 No 234005 29,40 77 150 30 80 5.0 2.0 20.5 2.0 0.30 10 Yes 235005 29.40 77 150 30 80 5.0 2.0 20.5 2.0 0.35 1.0 Yes 236005 29.40 77 150 30 80 5.0 2.0 20.5 2.0 04 1.0 Yes 237005 29.40 77 155 30 80 5.0 2.0 53.5 2.0 0.5 10 Yes Mallorv Coil + 256 K ohm and 2$ Pf added through 25,5 K ohm 255005 29.55 70 100 30 80 5.0 2.0 20.5 5.0 0.36 1.0 No 256005 29.55 78 140 30 80 5.0 2.0 20.5 5.0 040 1.0 Yes 257005 29.55 78 150 30 80 5.0 2.0 20.5 5.0 0.36 1.0 No 258005 29.55 78 155 30 80 5.0 2.0 20.5 5.0 0.38 1.0 Yes 259005 29.55 78 155 30 80 5.0 2.0 20.5 5.0 0.36 1.0 No 261005 29.55 80 170 30 80 5.0 2.0 20.5 5.0 0.40 1.0 Yes 263005 29.52 82 195 30 80 5.0 2.0 20.5 5.0 0.40 1.0 No 264005 29.48 82 200 30 80 5.0 2.0 20.5 $.0 0.43 1.0 No 265005 29.48 82 200 30 80 5.0 2.0 20.5 5.0 045 1.0 Yes 266005 29.48 82 200 30 80.0 2.0 20.5 5.0 0.50 1.0 Yes 267005 29.48 82 200 30 80 5.0 2.0 20.5 5.0 0.50 1.0 Yes

TABLE XVII. EXPERIMENTAL IGNITION DATA _ _= 0.73; 6 = 0.195 inches Zarormeter Scope Settings Time Teoa. p. teT V Sw.Rate I Constant Temp T Run o Evac. in. o I Kv U sec., v. Settings IgniNo,_ n. g. F Press.H. _ F Ais. cm, c- cm. cm R C _ton Delco Coil + 249.5 K ohm 10014 29.20 75 120 30 80 4.6 5.0 100.0 5.0 1.0 1.0 Yes 20014 29.20 75 125 30 80 4.6 5.0 53.5 5.0 0. 1.0 No 30014 29.20 79 130 30 80 6 2.53.5 2.0 05 1.0 No 4001L4 29.20 79 130 30 80 4.6 2.0 53.5 2.0 0.5 1.0 No 60014 29.20 79 140 30 80 4.6 2.0 53.5 2.0 0.70 1.0 Yes 70014 29.20 79 130 30 80 4.6 2.0 53.5 2.0 0.70 1.0 Yes 80014 29.23 81 150 30 80 4.6 2.0 53.5 2.0 0.70 1.0 No 90014 29.23 81 135 30 80 4.6 2.0 53.5 2.0 0.70 1.0 No 100014 29.23 81 75 30 80.60 2.0 3.5 2. 0.8 1.0 No 110014 29.23 81 100 30 80 4.6 2.0 53.5 2.0 0.8 1.0 No 120014 29.23 81 l^ O 30 80 4.6 2.0 53.5 2.0 0.85 1.0 Yes 130LOo 2.?2 81 145 30 80 46 2.0 53. 2.0 0.85 1.0 Yes 140014 29.23 31 120 30 80 4.6 2.0 53.5 2.0 0.85 1.0 Yes 150014 29.23 81 145 30 80 ['.6 2.0 53.5 2.0 0.80 1.o Yes 160014 29.23 81 125 30 80 4.6 2.0 53.5 2.0 80 1.0 Yes 180014 29.23 81 150 30 80 4.6 2.0 53 5 2.0 0.75 1.0 No 190014 29.23 81 130 30 80 [.6 2.0 53.5 2.0 0.75 1.0 No 200014 29.23 81 170 30 80 4.6 2.0 53.5 2.0 0.60 1.0 No 210014 29.23 81 125 30 80 4.6 2.0 53.5 2.0 0.5 1.0 No?2n00n! 29.23 81 150 30 80 4.6 2.0 535 2.0 0.5 1.0 No 240014 29.23 81 160 30 80 [.6 2.0 53.5 2.0 0.8 1.0 Yes 250014 29.23 81 150 30 80 4.6 2.0 53.5 2.0 0.9 1.0 es 280014 29.23 81 130 30 80 4.6 2.0 53.5 2.0 1.0 1.0 No

TABLE XVII. (Cont d) Barometer S.cope Settings- Time T ~: TeImp- Ia V Sw.Rate I Constant Run mpEvac in. T I Kv sec. v. Settings IgniNo. in.Hg. OF Press. Hg. oF Amps. c. cm. cm. R C tion Delco Coil + 249.5 K ohm 290014 29.23 81 140 30 80 4.6 2.0 53.5 2.0 1.0 1.0 Yes 300014 29.23 81 175 30 80 4.6 2.0 53.5 2.0 0.8 1.0 No Delco Coil + 249.5 K ohm and 25 Pf added through 118 K ohm 320014 29.23 76 105 30 80 5.0 2.0 20.5 2.0 0. 1.0 No 330014 29.23 76 140 30 80 5.0 2.0 20.5 2.0 0.5 1.0 No 340014 29.23 76 120 30 80 5.o 2.0 20. 2.0 0.5 1.0 No 350014 29.23 76 140 30 80 5.0 2.0 20.5 2.0 0.6 1.0 No 360014 29.23 76 100 30 80 5.0 2.0 20.5 2.0 0.6 1.0 No 370014 29.23 76 130 30 80 5.0 2.0 53.5 2.0 0.7 1.0 No 380014 29.23 76 125 30 80 5.0 2.0 53.5 2.0 0.8 1.0 Yes 390014 29.23 76 105 30 80 5. 2.0 53.5 2.0 0.7 1.0 Yes 400014 29.23 76 130 30 80 5.0 2.0 53.5 2.0 0.6 1.0 Yes 410014 29.23 76 125 30 80 5.0 2.0 53.5 2.0 0.6 1.0 No 420014 29.23 75 135 30 80 5.0 2.0 53.5 2.0 0.55 1.0 No 430014 29.23 75 105 30 80 5.0 2.0 53.5 2.0 0.8 1.0 No 440014 29.23 75 90 30 80 5.0 2.0 53.5 2.0 0.8 1.0 Yes 450014 29.23 75 85 30 80 5.0 2.0 53.5 2.0 0.9 1.0 Yes 460014 29.23 75 100 30 80 5.0 2.0 53.5 2.0 1.0 1.0 Yes 470014 29.23 75 110 30 80 5.0 2.0 53.5 2.0 1.1 1.0 Yes 480014 29.23 75 80 30 80 5.0 2.0 53.5 2.0 0.7 1.0 No 490014 29.23 75 80 30 80 5.0 2.0 53.5 2.0 0.8 1.0 Yes 50001) 29.23 75 160 30 80 5.0 2.0 53.5 20 0.5 1.0 No 510014 29.23 72 85 30 80 5.0 2.0 535 2.0 0.4 1.0 No

TABLE XVII (Cont d) Barometer Scope SettingsTime Ba o Temp,, ^T w.Kaie ~ Ionstant Run pEvac. in. o o Ky p. sec, v. Settigs IgniNo. in.Hg. F Press. H g. F Amps. cm, cm, cm. R C tion Deloo Coil & 249.5 K ohm and 25 pf added through 52 K ohm 540014 29.23 72 95 30 80 5.0 2.0 205 5.0 0.40 1.0 No 550014 29.24 76 160 30 80 5.0 2.0 20.5 5.0 0.40 1.0 No 560014 29.24 76 90 30 80 5.0 2.0 20.5 5.0 0.5 1.0 No 570014 29.24 76 95 30 80 5.0 2.0 20.5 5.0 0.5 1.0 No 580014 29.24 76 100 30 80 5.0 2.0 20.5 5.0 0.6 1.0 No 590014 29.24 74 105 30 80 5.0 2.0 20.5 5.0 0.6 1.0 Yes 600014 29.24 74 70 30 80 5.0 2.0 53.5 5.0 0.7 1.0 Yes H 610014 29.24 74 100 30 80 5.0 2.0 53.5 5.0 0.8 1.0 Yes o 620014 29.24 74 75 30 80 5.0 2.0 53.5 5.0 0.9 1.0 Yes 63001114 29.24 74 120 30 80 5.0 2.0 53.5 5.0 0.7 1.0 No 640014 29.24 74 90 30 80 5.0 2.0 53.5 5.0 0.65 1.0 Yes 650014 29.24 74 80 30 80 5.0 2.0 20.5 5.0 0.55 1.0 No 660014 29.24 74 95 30 80 5.0 2.0 20.5 5.0 0.45 1.0o No

TABLE XVIII. EXPERIMENTAL IGNITION DATA =0.69; 6 0.254 inches B- arortrScope SettingsTime TempI TeV Sw.Rate IConstant Run pEvac. In. o o Ky p sec. v. Settings IgniNo. in.Hg. F Press. Hg.* OF Amps. cm. cm. cm. _ R_ C tion Delco Coil + 249.5 K ohm 10017 29.42 80 55 30 80 5.0 2.0 100.0 5.0 3.0 1.0 Yes 20017 29.42 0 60 30 80 5.0 2.0 100.0 2.0 3.0 10 No 30017 29.42 80 70 30 80 5.0 2.0 100.0 2.0 3.5 1.0 Yes L0017 29.42 80 73 30 80 5.0 2.0 100.0 2.0 3.5 1.0 Yes 50017 29.42 80 70 30 80 5.0 2.0 100.0 2.0 3.0 1.0 Yes 60017 29.42 80 70 30 80 5.0 2.0 100.0 2.0 2.5 10 Yes 70017 29.42 80 75 30 80 5.0 2.0 100.0 2.0 2.0 1.0 No 80017 29.42 80 85 30 80 5.0 2.0 100.0 2.0 2.0 1.0 No 90017 29.42 80 85 30 80 5.0 2.0 100.0 2.0 2.25 10 No 100017 29.42 80 85 30 80 5.0 2.0 100.0 2.0 2.25 1.0 Yes 110017 29.42 80 85 30 80 5.0 2.0 100.0 2.0 2.25 1.0 Yes 120017 29.41 85 90 30 80 5.0 2.0 100.0 2.0 2.00 10 Yes 130017 2941 85 90 30 80 5.0 2.0 100.0 2.0 1.50 10 No 140017 29.41 85 90 30 80 5.0 2.0 100.0 2.0 1.50 1.0 No 150017 29.41 85 95 30 80 5.0 2.0 100.0 2.0 1.75 1.0 Yes 160017 29.41 85 95 30 80 5.0 2.0 100.0 2.0 1.50 1.0 Yes 170017 2941 85 95 30 80 5.0 2.0 100.0 2.0 1.25 1.0 Yes 190017 29.41 85 95 30 80 5.0 2.0 53.5 2.0 1.25 1.0 Yes 200017 29.41 85 100 30 80 5.0 2.0 53.5 2.0 1.0 1.0 Yes 210017 2941 85 95 30 80 5.00.0 53.5 2.0 0.5 10 No 220017 29.41 85 100 30 80 5.0 2.0 20.5 2.0 0.5 1.0 No 230017 29.41 85 95 30 80 5.0 2.0 20.5 2.0 0.5 1.0 No 240017 29.41 82 95 30 80 5.0 2.0 20.5 2.0 04 1.0 No

TABLE XVIII. (Contid) ~,,,, ~-+.-~~' ""' Scope Settings' Time Barome t e r V T Sw RateW- Constant Run Tmp* Evac in*o Kv sec. vSettings gniRun Evc in 10 V No. in.Hg. F Press. H g F Amps. cm. cm. cm. R C tion Delco Coil + 249.5 K ohm 260017 29.41 82 85 30 80 5.0 2.0 20.5 2.0 0.4 1.0 No 270017 29.41 82 80 30 80 5.0 2.0 20.5 2.0 04 1.0 No 280017 29.41 82 90 30 80 5.0 2.0 20.5 2.0 0.3 1.0 No 300017 29.41 80 90 30 80 5.0 2.0 53.5 2.0 0.6 10 No 310017 29.41 80 85 30 80 5.0 2.0 53.5 2.0 0.7 1.0 No 320017 29.41 80 70 30 80 5.0 2.0 53.5 2.0 0.8 1.0 No 330017 29.41 80 70 30 80 5.0 2.0 53.5 2.0 08 1.0 No 340017 29.41 80 100 30 80 5.0 2.0 53.5 2.0 08 1.0 No 350017 29.41 80 65 30 80 5.0 2.0 53.5 2.0 10 10 No 360017 29.41 80 80 30 80 5.0 2.0 53.5 2.0 1.10 1.0 No 370017 29.41 80 90 30 80 5.0 2.0 53.5 2.0 120 1.0 Yes Delco Coil + 249.5 K ohm and 25 Pf added through 149.5 K ohm 380017 29.41 77 85 30 80 5.0 2.0 53.5 2.0 1.20 1.0 No 390017 29.41 77 70 30 80 5.0 2.0 53.5 2.0 1.30 10 No 400017 29.41 77 70 30 80 5.0 2.0 53.5 2.0 1.30 10 No 410017 29.41 77 70 30 80 5.0 2.0 53.5 2.0 1.40 10 Yes 420017 29.41 77 65 30 80 5.0 2.0 53.5 2.0 1.50 1.0 No 430017 29.41 77 65 30 80 5.0 2.0 53.5 2.0 1.50 1.0 No 440017 29.41 77 65 30 80 5.0 2.0 53.5 2.0 1.60 10 No 450017 29.41 77 60 30 80 5.0 2.0 53.5 2.0 1.70 1.0 Yes 460017 29.41 77 70 30 80 5.0 2.0 53.5 2.0 1.70 1.0 No 480017 29.41 77 80 30 80 5.0 2.0 53.5 2.0 1.80 10 Yes 490017 29.41 79 75 30 80 5.0 2.0 53.5 2.0 1.90 10 Yes 500017 29.41 79 70 30 80 5.0 2.0 53.5 2.0 2.00 1.0 No

TABLE XVIII. (Cont d) Barometer Scope Settin s Time, - oTemp. T I V Sw.Rate I Constant Run emp- Evac. in. o o Kv j sec. v. Settings IgniNo. in.Hg..F Press. Hg. ~ F Amps. cm. _ cm. _ cm. R C c tion Delco Coil + 249.5 K ohm and 25 Pf added through 149.5 K ohm 510017 29.41 79 70 30 80 5.0 2.0 53.5 2.0 2.0 1.0 Yes 520017 29.41 79 70 30 80 5. 2.0 53.5 2.0 2.0 1.0 Yes 530017 29.41 79 80 30 80 5.0 2.0 53.5 2.0 2.2 10 Yes 540017 29.41 79 100 30 80 5.0.0 53.5 2.0 1.8 10 No 5.5017 29. 11 79 90 30 80 5.0 2.0 53.5 2.0 1.4 1.0 No 560017 29.47 80 70 30 80 5.0 2.0 53.5 2.0 1.4 1.0 No 570017 29.47 80 80 30 80 5.0 2.0 53.5 2.0 1.0 1.0 No 590017 29.47 80 100 30 80 5.0 2.0 53.5 2.0 2.4 1.0 Yes Delco Coil + 249.5 K ohm and 25 Pf added through 49.5 K ohm r 600017 29.43 81 95 30 80 5.0 2.0 20.5 5.0 0.5 1.0 No 610017 29.43 81 100 30 80 5.0 2.0 20.5 5.0 0.5 1.0 No 620017 29.43 81 125 30 80 5.0 2.0 20.5 5.0 0.6 1.0 No 630017 29.43 81 100 30 80 5.0 2.0 20.5 5.0 0.7 1.0 No 640017 29.43 81 100 30 80 5.0 2.0 20.5 5.0 0.8 1.0 No 650017 29.43 81 90 30 80 5.0 2.0 53.5 5.0 0.9 1.0 No 660017 29.43 81 95 30 80 5.0 2.0 53.5 5.0 1 1.0 0 Yes 670017 29.43 81 90 30 80 5.0 2.0 53.5 5.0 0.9 1.0 No 680017 29.43 81 85 30 80 5.0 2.0 53.5 5.0 1.0 10 Yes 690017 29.4,3 81 100 30 80 5.0 2.0 53.5 5.0 1.1 1.0 Yes 700017 29.43 81 85 30 80 5.0 2.0 53.5 5.0 1.2 1.0 Yes 710017 29.43 81 100 30 80 5.0 2.0 53.5 5.0 1.3 1.0 Yes 720017 29.43 81 100 30 80 5.0 2.0 53.5 5.0 1.4 10 Yes 730017 29.43 81 90 30 8 5 20 53.5 5.0 1.5 1.0 Ye 740017 29.43 81 90 30 80 5.0 2.0 53.5 5.0 1.6 1.0 Yes

TABLE XVIII (Cont d) Barometercope Settingsme Bar Tomet T Sw.Rate I Constant Run p Evac. in., O o Kv sec. v. Settings IgniNo. in.Hg. F Press. Hg. F Amps. cm. cm. cm. R C tion Delco Coil + 249.5 K ohm and 25 Pf added through 49.5 K ohm 750017 29.43 81 100 30 80 5.0 2.0 53.5 5.0 1.7 1.0 Yes 760017 29.43 81 100 30 80 5.0 2.0 53.5 5.0 0.95 1.0 No..._....__.......Delco Coil + 249*5 K ohm 770017 29.43 81 90 30 80 5.0 2.0 20.5 2.0 0.7 1.0 No 780017 29.43 81 100 30 80 5.0 2.0 20.5 2.0 0.8 1.0 No 790017 29.43 81 90 30 80 5.0 2.0 20.5 2.0 0.9 1.0 No 800017 29.43 81 105 30 80 5.0 2.0 20.5 2.0 10 10 1o - 810017 29.43 81 90 30 80 5.0 2.0 53.5 2.0 1.1 1.0 No 820017 29.43 81 120 30 80 5.0 2.0 53.5 2.0 1.3 1.0 No 830017 29.43 81 85 30 80 5.0 2.0 53.5 2.0 1.5 1.0 Yes 840017 29.43 81 100 30 80 5.0 2.0 53.5 2.0 1.5 1.0 Yes 850017 29.43 81 95 30 80 5.0 2.0 53.5 2.0 1.4 10 Yes 860017 29.43 81 100 30 80 5.0 2.0 53.5 2.0 1.35 1.0 Yes 870017 29.43 81 105 30 80 5.0 2.0 53.5 2.0 1.3 1.0 Yes 880017 29.43 81 90 30 80 5.0 2.0 53.5 2.0 1.2 1.0 Yes 890017 29.43 81 95 30 80 5.0 2.0 53.5 2.0 1.1 1.0 Yes 900017 29.43 80 105 30 80 5.0 2.0 53.5 2.0 1.0 1.0 Yes 910017 29.43 80 90 30 80 5.0 50 3.5 2.0 0.9 1.0 Yes 920017 29.43 80 95 30 80 5.0 2.0 53.5 2.0 0.9 1.0 Yes 930017 29.43 80 95 30 80 5.0 2.0 53.5 2.0 0.8 1.0 No 940017 29.43 80 105 30 80 5.0 2.0 53.5 2.0 0.7 1.0 No

APPENDIX F DATA REDUCTION Data for each spark consists of the photograph of the voltage and current traces and the corresponding initial temperature and pressure readings, etco, which are listed in Tables XIV through XVIII. A typical photograph is shown in Figure F.lo The traces shown on the photograph were broken up into small time increments such that the curves could be closely approximated by straight segments. Values of voltage, current and time were tabulated at each of the time increments. The data was then punched on IBM cards for computer analysis. The computer solution follows the procedure outlined by the flow diagram shown in Figure F.2. Samples of the computer output are shown in Figures F,3 and F.4o The last column on the right of Figure F.3 is the list of the cumulative energy values for each time increment up to the total. The transition point between the EH and EL components of the spark was chosen at the knee of the curve shown in Figure Fo4o The corresponding value of PLA/PHA is found in the second set of tabulated values in Figure F.3o The following symbols are used in the computer program and in the flow diagram A = voltage divider attenuation ratio AVHP(J) = PHA AVHP(J) = PLA B = constant required in calculation in amperes from centimeters -155

-156BAR = barometer reading CA(J) = amplitude of current trace in centimeters at DELTA(J) CAP capacitance setting in pulse circuit CBAR = corrected barometer C(J) = current in milliamperes CS = oscilloscope setting for current DELTA(J) = time increment in centimeters measured from point of breakdown DELT(J) = time interval between each point in centimeters E (J) = energy in millijoules for each time interval T (J) GAP = electrode spacing K = a constant used in calculation of corrected pressure M = constant referring to total number of elements the voltage and current traces were broken up into PC = primary current P (J) = power in milliwatts PRESS 1 = initial pressure in psig or inches of mercury absolute PRESS 2 = initial pressure converted to inches of mercury absolute R = resistance setting in pulse circuit RATIO = LA/PA RN = run number SRT = oscilloscope setting for sweep rate, in psec/cm SUM(J) = cummulative energy for all T(J) before it T 1 = barometer temperature TA(J) = same as DELTA(J) in real time units of microseconds TEMP 1 initial temperature

-157T (J) = same as DELT(J) in real time units of microseconds V (J) = actual voltage VS = oscilloscope setting for voltage

Figure F.I Photographic Record of Spark Volago Ft.greCurrent Fs a Funecord of Tpime.rge an Current as a (Fabun.ction of P Tl...,.

(^ r-~~ READ: PRINT: RNTEMPI,PRESS2,CBAR, READ: M,._.I ABC IGAPPCCAPRVSCSSRT DELTA(J)......DELTA(M) (i |VA(J)...........VA(M) CA(J).......... CA(M) READ ___ COMMENT DO J=1, M-1 IREAD: KRNTEMPI, DELT(J+1) = DELTA(J+1) - DELTA(J) DELT(1) =DELTA(l) PRESSI,BARJGAP;,TI,9PC, F-] CAP;REVS ICS,SRT BAROMETER CORRECTION I |. c6 (oBAR-30- )X (c8-c4) | cc c6+ C7 5I =1.0 I PRINT: MINIMUM SPARK IGNI- I TION ENERGY CALCU- C8=.99+TEx.0028 C7 0083 + C5 x 000'8 LATION Ic8 =.+TEx.0028 PRINT: BAR:30. C5 = (BAR-29.) c6 = c2+C5 x (c4-c2) COMMENT I 1. I r —--------— J t "I I I I |TE 8 TI 6 C| 4.096+TEx 00265 PRESS2 = ( D + CBAR)/30.0 I __i__ I 1 0.09+TE x.0025 BR:29. C3 = Cl + (BAR.0 )x(C2-Cl) (I -— | \-. -,'0< > I / — -— \> i -------' -------- I ^-\ ( K:l 1) PRESS2 = PRESS1/30. | I__C3=C1 cC = C3 +.00 —- CBAR = BAR - CC Figure F.2. Computer Flow Diagram.

(7)- T(J) = SRT xDELT(J) — TA(J) = SRTxDELTA(J) — V(J) = AxVSxVA(J) - DO J=1 M ----- p(J) = V(J) x C(J) C( —-- C(J) = B x CS x CA(J) P(J)~~~~~~~~~~pW+P (J+l) )x(J6xl SUM(l) - 0.0 -E(J+1) Jt 2.0 BO PRINT: PLOT: jPRINT TA(J)..........TA(M) - SUM(J+)= SUM(J) E(J+) COLUMN VERSUS TITLES PW(J)..........PW(M) PRINT: -SUIMJ(J) = SUM(J) x -10'3- DELTA(J).......DELTA(M) PRINT: _ _ TA(J).............TA(M) TA(J)..........TA(M) (J) = P(J) x 10 i DELT(J)........ DELT(M) AVHP(J)......AVHP(M) i T(J)...............T(M) AVLP(J)......AVLP(M) TASEC(J) TA(J)xlO-6 ~ VA(J).............VA(M) RATIO(J)....RATIO(M) V(J)...............V( CA(J).............CA(M) C(J)............... C(M AVHP(J) = SUMJ(J)/TASEC(J) p(J)..............P(M) PRINTF F (n SuM(Ji)...........SM(M) COLUMN TITLE DO J=2,M-1 |BOW(J) = SUMJ(M)-SUMJ(J) RATIO(J) = AVLP(J)/AVHP(J) < —--- AVLP(J) = BOW(J)/(TASEC(M)-TASEC(J) Figure F.2. (Continued).

-161-..................;MINIKUM. SPARK.GlNIT.L.OL-6N.RC.Y-CALCULATIOK —..- --—..- ______.... 7/24/63,DELCO+249.5,25PF 149.5KtER=.69, RUN NO 56017 INITIAL TEMPERATURET 79.OF INITIAL PRESSURE,ATM.- 1.0 CORRECTED OAROMETER 29.271N.HG. ELECTRODE GAP-.2541N. PRI. CURRENT-5.0ANPS. CAPACITY-1.0 R-1.40 OSC. SETTINCS VOLTAG6E 2 V/CM. CURRENT- 2 V/CM. SWEEP RATE,MICROSEC./CM.- 54 TIME TIHE TIME TIfE VOLTAGE CURRENT INSTANTANEOUS V S VES OF NET ENERGY FOR INCREFENT INCREMENT INTERVAL INTERVAL INCRERENT INCREMENT CURRENT VOLTAGE POWER TIME INCREMENT: IN,CM. IN,MICROSEC. IN,CM. INtMICROSEC. INCM. IN,CM. IN,MILLIAMP IN,VOLT IN,MILLIWATT IN,MILLIJOULES ---— j..-I —---— _. -- A OL --------— P — %p A —-6 —---- 6^0- --------------- 32 ------------ As 7-,, 1 2AQ NA - I-416 -2 A --------— 0 — _.05 2.67.05 2.67 1.10 5.55 74.00 2200.0 162808.13.291.11 5.8e.06 3.21 1.02 3.22 42.94 2040.0 87588.38.693.22 11.77.11 5.88.65 2.00 26.67 1300.0 34668.4C 1.052.-. —, -,^2Jt 14 —---------..J,19.2^2-...._.t..-_ 7_____ _ —,t.,A _ _.., 2.. --------— A —---------.,.._ —------— l.. 192..i-2-.7_-9.9_.1.2... -.58 31.03.22 11.77.65.95 12'67 1300.0 16467.49 1.467._ _ 4._.. U...8....-.__ - -...2.,8 —.-.. - - — 15 4Q.....L —- -9 —--—.....L-tJ.l. —----— J —L,6_.J_ —-. 2.52 134.82 1.72 92.02.75.43 5.73 1500.0 8600.43' 2.694 3.50 187.25.98 52.43.60.00.00 1200.0.00 2.919. TiE-EA&-E —-—............. —--—...E. A-........... E —--------------------------------—.................................... —----------------------------------------—... —------------------------ INCRENENT HIGH RArE LOW RATE IN.MICR OSEC POWER IN..PAGS. IP.ERI.L MAI Q _._ __,.. —- - ---—. — - - 2.67 108.71 14.24.131 5.88 117.70 12.28.104 11.77 89.42 10.64.119.1 2g _.__......._5._._... —. -— __... 11. L52-__ —____ 31.03 47.28 9.30.197 _ _ 420 3 -_ 22_____ _ __. 2________L22...__....._________ ______ —----— _ 134.82 19.98 4.30.215.FIGURF F.3 SAM__ LE QMPU TE JJ.TPUT -_ ____.'

-162-. __ —- - - - -- - - - -- - - - -- - - - -- - - - --- - - - D m_ — - - - — P- - - - - — V S- -- - - - - - - RUN NO. 56017 Q.7OQ f —---------------------------- ----------------------- I I I I I I I I I I I --------- ------— jL —----------- -------------. —------- -- -.. —-----—, —-----------—.I —----------— j ------- I —------------- - - --. - L I I I I I I I I I I I ------------ -----------— I —----------—.. —---------------- ------------------------------------------------------- --------— I —-------- -- I I I I I *I I I I I I I I I I_______II___________II_________I________I__..____I__. ______I1________ I I I I I I I I I I i N III —---------- I —------------------ I - ------- - - --- ----- ------------ I I I --------- ~ AI I I I I I I I I I I I It I I I I I I I I I I A —------------- -------- ------------- -------------- -------— I —---------------------------------------------------- -- - I- - -- r I I I: I I I I I I I I A *1I - 1. II..1 - I N I I I I I I I I I I N I I: r.1.1 1 __ ___ _ _ __,,,,, L,,,,,___ _ _ _ __ _ _I _ _ _ _ _ _ _ _ _ _ T ii i I I I I I I I I.,,,,,,,,d_,,,,*1 _I.- __.1 -._,I,_.. - - - L.. -.1..1 I,,, I,_ N If I I I I I I I I I -E_______11 1 __I_______I _______ _______I1_____I____ I___I ________ _______I._______I1_____I_ E I I I I I I I I V.1..-.I +-: — ---— +-, —,-I —-+-1_- I-.-., —R I I I I I I I I I I I I 11 I I I I I I I I I.IL..IQ-aQ.QQ ~ - -- -- - -- - -- -- - - --— ~ ~ ~-~ -~ -~- ~-. -- " ~- +^ — _.-+. w i i r r i i I i I I A I I I I I I I I I I I I I I: I I I I I I I'R """'~' I ""'"' -— + —-— j -— +- ^ ---— + —--- ^- - -^- - -- - -^- - - -~I___- ---------— _ —— _ —------—. —-— __ —— _ —. ___ —-.-I —I —-—.__-..J —----------— I —---------— I - --------- ------------— _ I I I I I I I I I I I l ~~~~ ^^^ —-— __,,,,_,I,,,,j_-,,X N 8.00 --— ~ —- ---- --- ----------— _ —— l —-—, —----— L ——. —----— I —— _ —----— l —-----— J- _, - - _ -— l —-_ —----------- -- A I I I I I I I I I I J --- - -_ I - I- I I I I I I -------- -^ -. - ^ -------- --- ---------— I —---------- ^- - -- -- --- - — __ ---— _ —.I__.-___ —--- ----- — ^ ----— ^ - - ^ -- ---- - ----— ^ --- ------ ---- -- ^ - --- --- -^ -- _ ----- - _ —-___ —— _ —__ —_ __ --— __-_ _______ —_ ___ —__-_* —tQQQ.Q:z: -— 4 -~~ —--- ----- ~- 4 —---------- — ~ —-~ - __nI I I Io I I I I I I I I I I - I I I I I I I I I I I I I I I I I I ^Q,J- O - -- ----- — ~+ —-----— + —-----— + —-----— + —---- -------— + —---------------— + —-------- ------ L 0.-_ —000O ~ CC.C..._-4 C. — 4 —-. -----— 4~.-+.- ~ +~0,_p..._~0.0 ----- ---------------------------------------------------- - - - - 0.000 2C.CCO 4C.COC 60.000 80.000 100.000 120.000 140.000 160.000 18C.COO 200.000 TIHE IN NICROSEC. --— 4 — FURE. ----— AMPL_ -POWERV-IME.QMPUT PL —---------- ----- ----------- — P WER vs. IME COM UT PL O

BIBLIOGRAPHY 1. Silsbee, F. B., Loeb, L. B., and Fonseca, E, L., "Part I. Method of Measuring Heat Energy of Ignition Sparks" Fifth Annual Report of NACA, Report No. 56, (1919), 163. 2. Silsbee, F, B., and Fonseca, E. L., "Part II. Measurement of Heat Energy per Spark of Various Ignition Systems,"? Fifth Annual Report of NACA, Report No, 56, (1919), 169. 3. Wheeler, R. V., "The Ignition of Gases. Part I. Ignition by Impulsive Electrical Discharge. Mixtures of Methane and Air' J. Chem. Soc., 117, Part 2, (1920), 903. 4. Wheeler, R. V., "The Ignition of Gases. Part III. Ignition by Impulsive Electrical Discharge. Mixtures of Paraffins with Air," J. Chem. Soc., 125, Part 2, (1924), 1858. 5. Taylor-Jones, E., Morgan, J. D., Wheeler, R. V., "On the Form of the Temperature Wave Spreading by Conduction from Point and Spherical Sources; with a Suggested Application to the Problem of Spark Ignition," Lond. Phil. Mag., 43, (1922), 359. 6. Taylor-Jones, E., "Spark Ignition," Lond. Phil. Mag., Ser, 7, 6, N. 40, (Dec. 1928), 7. Coward, H. F., and Meiter, E, G., "Chemical Action in the Electric Spark Discharge," J. Am. Chem. Soc., 49, (1927), 396. 8, Finch, G, I., and Thompson, H. W,, "The Effect of Frequency on the Condensed Discharge Ignition of Carbonic Oxide-Air Detonating Gas," Proc. Royal Soc., A, 134, (1931), 343, 9. Bradford, B. W., and Finch, G, I., "The Mechanism of Ignition by Electric Discharges," Chem, Rev,, 21, No, 2, (Oct. 1937)o 10. Thompson, H, W., "The Explosive Combination of Hydrogen and Oxygen - The Function of Walls in Gaseous Reactions," Faraday Soc. Trans., 28, (1932), 299o 11. Linnett, J, W., Raynor, E, J., and Frost, W. E., "The Mechanism of Spark Ignition," Faraday Soc. Trans., 41, (March, 1945), 487. 12, Lewis, B., and von Elbe, G., Combustion, Flames and Explosions of Gases, New York: Academic Press, Inc., 1951. 13. Linnett, J. WL, and Nutbourne, D. M., "The Spark Ignition of Nitrous Oxide-Hydrogen Mixtures," Third Symposium on Combustion, Flame and Explosion Phenomena, Baltimore: Williams and. Wilkins, (1949), 336, -163

-16414. Morgan, J. D., "The Ignition of Explosive Gases'by Electric Sparks," Jo Chemo Soc. Trans., 115, (1919), 94. 15. Morgan, J Do, Principles of Ignition, London: Sir Isaac Pitman and Sons, Ltd., 1942. 16. Bradford, Bo Wo, Finch, G. I., and Prior, Ao M., "The Coil Ignition of Some Explosive Gaseous Mixtures," J. Chemo Soc., Paper 63, (1933), 227. 17. Finch, G, I., and Sutton, R. W., "The Control of Ignition - Coil Discharge Characteristics," Proc. Royal Soc Lond., 45, (1933), 288. 18. Jost, Wo, Explosion and Combustion Processes in Gases, New York: McGraw-Hill Book Co,o Inc,, 1946o 19o Finch, G. I., and Mole G,, "Thle Mechanism of Electrical Ignition," Proc. Insto Auto. Eng., 29, (1934), 71. 20. Blanc, Mo V., Guest, P. Go, von Elbe, G., and Lewis, B., "Ignition of Gas Mixtures by Electric Sparkso Io Minimum Ignition Energies and Quenching Distances of Mixtures of Methane, Oxygen, and Inert Gases," Jo Chemo Phys., 15, No. 11, (Nov. 1947), 797. 21, Blanc, Mo V., Guest, Po G., von Elbe, G,, and Lewis, B., "Ignition of Explosive Gas Mixtures by Electric Sparks. III. Minimum Ignition Energies and Quenching Distances of Mixtures of Hydrocarbons and Ether with Oxygen and Inert Gases," Third Symposium on Combustion and Flame and Explosion Phenomena, Baltimore: The Williams and Wilkins Coo, (1948), 363. 22, Guest, Po Go, "Apparatus for Determining Minimum Energies for ElectricSpark Ignition of Flammable Gases and Vapors," U.S Buro Mines Repto Invest., 3753, (1944)o 23. Lewis, Bo, and von Elbe, G., "Igniition and Flame Stabilization in Gases," ASME Trans., 70, (1948), 307, 24. Morris, H., "Ignition of Gas Mixtures by Electric Sparks," Third Symposium on Combustion, Flame and Explosion Phenomena, Baltimore: Williams and Wilkins, (1949), 361, 25, Lewis, B., and von Elbe, G, "Ignition of Explosive Gas Mixtures by Electric Sparkso IIo Theory of the Propagation of Flame from an Instantaneous Point Source of Ignition," J. Chem. Phys., 15, No, 11, (Novo 1947), 8030 26o Belles, Fo Eo, and Swett, Co Co, "Ignition and Flammability of Hydrocarbon Fuels," Basic Considerations in the Combustion of Hydrocarbon Fuel with Air, NACA Report 1300, (1959), 83.

-16527. Friedman, R., and Johnston, W. C., "The Wall Quenching of Laminar Propane Flames as a Function of Pressure Temperature, and Air-Fuel," J. Apple Physo, 21, No. 8, (Aug., 1950), 791. 28. Harris, Mo E., Grumer, J,, von Elbe, Go, and Lewis, Bo, "Burning Velocities, Quenching, and Stability Data on Nonturbulent Flames of Methane and Propane with Oxygen and Nitrogen-Application of Theory of Ignition, Quenching, and Stabilization of Flames of Propane and Air," Third Symposium on Combustion, Flame and Explosion Phenomena, Baltimore: The Williams and Wilkins Coo, (191), 80. 29. Priede, To, and Rose, H. E., "An Investigation of the Characteristics of Spark Discharges as Employed in Ignition Experiments," Seventh Symposium (International) on Combustion, London: Butterworths Scientific Publications, (1959), 454o 30. Priede, To, and Rose, H. E., "Ignition Phenomena in Hydrogen-Air Mixtures," Seventh Symposium (International) on Combustion, London: Butterworths Scientific Publications, (1959), 436. 31. Swett, C. C,, Spark Ignition of Flowing Gases, Part 1o Energies to Ignite Propane-Air Mixtures in Pressure Range of 2 to 4 Inches of Mercury Absolute, NACA RM-E9E17, (Augo, 1949). 32. Swett, C. C., Spark Ignition of Flowing Gases, Part 2, Effect of Electrode Parameters on Energy Required to Ignite a Propane-Air Mixture, NACA RM-E51J12, (Dec., 1951)o 33. Swett, C. C., and Donlon, R, H., Spark Ignition of Flowing Gases, Part 3, Effect of Turbulence Promoter on Energy Required to Ignite a Propane-Air Mixture,NACA RM-E52J28, (Jan. 1953)o 34 Swett, C. Co, "Effect of Gas Stream Parameters on the Energy and Power Dissipated in a Spark and on Ignition," Third Symposium on Combustion and Flame and Explosion Phenomena, Baltimore: The Williams and Wilkins Co., (1949), 3540 35o Swett, C, C,, "Spark Ignition of Flowing Gases Using Long Duration Discharges," Sixth Symposium (International) on Combustion, New York: Reinhold Publishing Corp., (1956), 523 36. Guest, P. G., Lewis, B., Roth, W., and von Elbe, Go, "Heat Generation by Electric Sparks and Rate of Heat Loss to the Spark Electrodes," J. Chem. Phys., 19, No, 12, (Dec. 1951), 1530. 37. Olsen, Ho Lo, Gayhart, Eo L., and Edmonsen, Ro B., "Application of Ideal Gas Theory to the Gaseous Expansion from an Electric Spark," J. Appl. Phys.,, 25, No, 8, (Aug.,, 1954), 1008.

-16638. Olsen, H. L., Gayhart, E, L., and Edmonsen, R. B., "Microchronometric Schlieren Study of Gaseous Expansion from an Electric Spark," Jo Apple Physo, 23, No. 10, (Oct. 1952), 1157. 39. Olsen, H. L., Gayhart, Eo Lo and Edmonson, Ro Bo, "Propagation of Incipient Spark-Ignited Flames in Hydrogen-Air Mixtures," Fourth Symposium on Combustion, Baltimore: Williams and Wilkins, (1953)9 144. 40. Arnold, J. S., and Sherburne, R, K., "Obervations of the Ignition and Incipient Flame Growth in Hydrocarbon-Air Mixtures," Fourth Symposium on Combustion, Baltimore: Williams and Wilkins, (1953), 1390 41. Kumagai, Seuchiro, Tadami Sakai, and Itsuro Kimura,"Effect of Ultrasonic Waves on Flame Propagation and Spark Ignition," Fourth Symposium on Combustion, Baltimore: Williams and Wilkins, (1953) 48. 42. Calcote, H. F., Gregory, C. Ao, Barnett, C. Mo, and Gilmer, R, B,, "Spark Ignition. Effect of Molecular Structure," Indo Engro Chem, 44, (1952), 2656, 43, Metzler, A, J., Minimum Spark-Ignition Energies of 12 Pure Fuels at Atmospheric and Reduced Pressure, NACA RM-E53H31, (0cto 1953) 44. King. I. R,, and Calcote, H. F., "Effect of Initial Temperature on Minimum Spark-Ignition Energy," J. Chemo Phys., (Letters to the Editor), 23, No, 12, (Dec, 1955), 2444, 45. Harnsberger, Do Ao and Van Wylen, G. J., Effects of Radiation at Ignition Point of Constant-Volume Combustion. Ann Arbor: University of Michigan College of Engineering Industry Program, IP-171, 1956. 46. Souka, Ao Fo., Influence of Alpha-Radiation on Some Aspects of Constant Volume Combustion, PhoDo Dissertation9 University of Michigan, 1957. 47. Clarke, J. So, "Initiation and Some Controlling Parameters of Combustion in the Automobile Engine," SAE Trans., (1962), 240o 48. Miller, W., "Fast Break Induction,'9 SAE, Paper 652D, 1963. 49. Obert, E. F., Internal Combustion Engines Analysis and Practice, Scranton, Pa.: International Textbook Co, 1950, 50. Ethyl Corporation. Effect of Ignition Source on Spark Plug Electrode Erosion, Report No. A. R. 102, prepared by Ethyl Corporation, Detroit, Dec. 1943. 51. Ethyl Corporation. Effect of Internal Resistors on Spark Plug Electrode Erosion, Report No. Ao R, 101, prepared by Ethyl Corporation, Detroit, Dec. 1943o

-16752. Ethyl Corporation. Effect of Low and High Values of Capacitance on Spark Plug Electrode Erosion, Report Noo A. R. 100, prepared by Ethyl Corporation, Detroit, Dec., 1943. 53. Ethyl Corporation. The Effect of Different Center Electrode Materials on Sparking Voltage During Actual Engine Operation, Report No, A. Ro 139, prepared by Ethyl Corporation, Detroit, Apro, 1945. 54, Allsop, G., and Guenault, E. M., "The Incendivity of Electric Sparks in Relation to the Characteristics of the Circuit," Third Symposium on Combustion, Flame and Explosion Phenomena, Baltimore, (1949), 341. 55. Cipriani, C., and Middleton, Lo H., "Ignition Theory Uncovers Electrical Design Facts," SAE Journal, 56, No, 10, (Oct., 1948), 47. 56, Loeb, L. B., and Meek, JO M., The Mechanism of the Electric Spark, Stanford University, Califo: Stanford University Press, 1941o 57. Loeb, Lo Bo, Fundamentals of Electricity and Magnetism, New York: Dover Publication, Inco (Unabridged and corrected republication of 3rd edition punlished by John Wiley and Sons, Inc., 1947), 1961. 58o Dow, Wo G., Fundamentals of Engineering Electronics, New York: John Wiley and Sons, Inc., 1937o 59. Wheatcroft, Eo Lo E., Gaseous Electrical Conductors, England: Oxford at the Clarendon Press, 1938. 60o Devins, Jo C,, and Sharbaugh, Ao Ho, "The Fundamental Nature of Electric Breakdown," Electro-Tech., (Febo, 1961), 103o 61o Meek, Jo Mo, and Craggs, Jo Do, Electrical Breakdown of Gases, Great Britain: Oxford at the Clarendon Press, 19530 62, Cobine, Jo Do, Gaseous Conductors, (Unabridged republication of the 1st edition with corrections by Author 1941), New York: Dover Publications, Inc,, 1958, 63. Townsend, Jo S., Electricity in Gases, London and New York: Oxford at the Clarendon Press, 1915. 64. Loeb, Lo B., Fundamental Processes of Electrical Discharge in Gases, New York: John Wiley and Sons, Inc., 19390 65, Loeb, L. B., Basic Processes of Gaseous Electronics, Los Angeles and Berkeley: University of California Press, 1960, 66. Meek, J. Mo, "A Theory of Spark Discharge," Phys. Rev., 57, (1940), 722, cited in Reference 610

-16867. Raether, Ho, Arch. Electrotech., 34, (1940), 49, cited in Reference 61, 68, Eason, W. R., "Voltage Risetime - A New Ignition Criterion?", SAE, Paper 652C, Jan, 1963. 69~ Semenov, No No, Thermal Theory of Combustion and Explosion, NACA Tech. Memo No. 1024, (1942) from Progress of Physical Science (UoSoS.R.) Volo 23, No. 3, (1940) 70. Vulis, L. A,, Thermal Regimes of Combustion, New York: McGraw-Hill Book Co., Inc., (Translated by M. D7 Friedman and G. C. Williams), 1961. 71. Frost, Ao Ao, and Pearson, Ro G,, Kinetics and Mechanism, 2nd Ed., New York: John Wiley and Sons, IncO., 1961. 72. Benson, S. W., The Foundations of Chemical Kinetics, New York: McGraw-Hill Book Co,, Inc., 1960. 730 Tenn, J. B., "Lean Flammability Limit and Minimum Spark Ignition Energy," Indo and Engr. Chem,, 43, (July-Dec., 1951), 2865. 74. Swett, C. Co, Spark Ignition of Flowing Gases, Part IV. Theory of Ignition in Nonturbulent and Turbulent Flow Using Long-Duration Discharges, NACA RM-E-54F29a, Augo, 1954o 75. Swett, C. Co, Spark Ignition of Flowing Gases, Part V. Application of Fuel-Air Ratio and Initial Temperature Data to Ignition Theory, NACA RM-E55I16, Nov. 1955. -76. Yang, Co Ho, "Theory of Ignition and Auto Ignition," Combustion and Flame, 6, (Deco, 1962), 215, 77. Mole, G,, "The Ignition of Explosive Gases," Proc. Phys. Soc., 48, No, 61, (Octo, 1936), 857 78. Landau, H. G., "The Ignition of Gases by Local Sources," Chem, Rev,, 21, No. 2, (July, 1937), 245. 79o McAdams, W. Ho, Heat Transmission, New York: McGraw-Hill Book Co,, Inc., 1954. 80. Radio Corporation of America, RoCoAo Transistor Manual, 1962. 81o Shea, Ro Fo, (editor), Transistor Circuit Engineering, New York: John Wiley and Sons, Inc,, 1957o 82, Strauss, L., Wave Generation and Shaping, New York: McGraw-Hill Book Co., Inc., 1960o

-16983. Basic Theory and Application of Transistors, New York: Dover Publication, 1963 84. Larges, Eo R., Frank, R. K., and Schneider, A., "Operating the Ignition Transistor within Safe Limits," SAE, Paper 652B, Jan, 1963 85. Saad, M. A., Evaporation and Combustion of Single Fuel Droplets in a Hot Atmosphere, Ph.D. Dissertation, University of Michigan, IP-140, 1956. 86. Stout, M. B., Basic Electrical Measurements, Second Edition, New Jersey: Prentice-Hall, Inc, 1960. 87. Orr, L. W., Roberts, Go A., Schulte, Ho T., Jr., Butler, T. W. Jr., and Otterman, J, Research and Development of Improved Ignition Systems, ERI Report 2370:2-P-3, University of Michigan, May 1956.

UNIVERSITY OF MICHIGAN II3 91111111 11111 11111111115 03526 1 3 9015 03526 7551