An Experimental Investigation Into Pre-ignition In The Spark

Transcript

125 An Experimental Investigation into Pre-ignition in the Spark-ignition Engine By D. Downs, B.Sc. (Eng.), A.M.I.Mech.E.*, and J. H. Pigndguy, A.M.I.Mech.E.* Pre-ignition and auto-ignition are uncontrolled ignition of combustible mixture in an engine by a hot surface. Pre-ignition is distinct from “knock” or “detonation” which is caused by the Apid combustion of the last part of the mixture following the initiation of flame by a spark. More work has been carried out on knock than pre-ignition because knock sets the primary limit to the possible power-output and fuel economy of an engine. Owing to increases in engine power in recent years, however, interest in pre-ignition has been revived, and this paper describes an investigation which was primarily directed to finding the pre-ignition ratings of various fuels. Two methods of obtaining pre-ignition ratings have been developed; with one a 500 C.C. singlecylinder water-cooled unit is used, having an electrically heated hot-spot to induce pre-ignition, engine with an air-cooled pre-igniter is used. and with the other a single-cylinder air-cooled air& A rating scale has been constructed in which iso-octane has been given the value of 100, and cumene zero. A large number of pure fuels from the paraffin, naphthene, aromatic, and alcoholgroups have been rated for pre-ignition tendency using these methods. In addition, typical commercial motor-fuels, motor-racing fuels, and water-alcohol injection fluids have been tested. The relationship between pre-ignition tendency and the temperature of the hot-spot with normal running and with pre-ignition conditions has been established. The effects of engine variables, for example, mixture strength, speed, ignition advance, and air and cylinder temperatures, on pre-ignition, have been studied, in addition to the effects of important fuel additives such as tetraethyl lead and the aromatic amines. The additive tests include substances, such as formaldehyde and nitrogen peroxide, which are of interest from the chemical aspect and, based on the results, some information on the fundamental chemical processes involved in hot-spot ignition has been obtained. INTRODUCTION So far as the fuel is concerned, knock or detonation places the greatest restriction on the power output and thermal efficiency of the spark-ignition engine because of the limit which it places on the permissible compression-ratio and boost pressure. For this reason work on the development of high-grade fuels for aircraft and road-vehicle engine use has been concentrated on the evolution of materials having the highest possible anti-knock value. Pre-ignition has been of secondary importance, especially since the development of sparking plugs of high heat-factor and improved cylinder-head cooling arrangements. During the 1939-45 war, however, engine power and operating temperatures were considerably increased, and the lead content of aviation fuel was raised well above the pre-war level. It was not surprising, therefore, that there were aircraft engine failures, usually due to burnt or fractured pistons or bent connecting rods, which appeared to have resulted from pre-ignition rather than knock. A revival of interest in the phenomenon of pre-ignition followed, particularly with respect to the part played by the fuel in determining whether pre-ignition would occur. An investigationwas started at Shoreham towards the middle of 1943 into pre-ignition and the relative pre-ignition tendencies of various fuel blends and components used in aircraft practice. Before then little systematic work on pre-ignition had been carried out and the only published information was that of Serruys (1937)t, who was concerned with the temperature of pre-ignition rather than with pre-ignition itself, and Spencer (1941), who did much valuable work principally with a “single combustion” apparatus. Concurrently with work in Great Britain, investigationswere conducted by the National Advisory Committee for Aeronautics (N.A.C.A.) and other organizations in the United States (Alquist and Male 1944, Biermann and The MS. of this paper w a s first received at the Institution on 21st June 1950. * Research Engineer, Messrs. Ricardo and Company, Engineers (1927), Ltd. t An alphabeticalfist of references is given in the Appendix, p. 139. * Corrington 1942, Corrington and Fisher 1948, Male and Eward 1945, Male 1946) and, since 1945, by the French Institute of Petroleum at Bellevue (Vischnievsky 1947). This paper gives an account of the development of test methods and apparatus for investigating the phenomenon of pre-ignition. The ratings of numerous fuels, including pure a m ponents and commercial blends, have been obtained, and the effects of engine variables and the important fuel anti-knock additives have been determined. Finally, the conclusions of some fundamental work on the nature of the chemical processes involved in hot-spot ignition have been given. PRE-IGNITION A N D AUTO-IGNITION Pre-ignition and auto-ignition are uncontrolled inflammation of combustible mixture in a engine by a hot surface. The hot surface may be the central electrode or insulator of the sparking plug or, in certain circumstances, combustion chamber deposits or the exhaust valve. With sufficientlysevere operating conditions, t h i s surface may become hot enough to ignite the charge independently of the sparking plug. If the hot-spot temperature is not high, there may be insuftiaent time for hot-spot ignition to develop before the whole charge is ignited by the flame from the spark. I n this case no unusual effect would be observed in the operation of the engine and the phenomenon could only be detected by switching off thi ignition when irregular firing would occur for a few strokes before the engine speed fell. This is frequently called “after-firing” and is illustrated in Fig. lu, which shows cylinder-pressure d i a m with various intensities of hot-spot ignition. With further increase in the hot-spot temperature, when the ignition is switched off, the engine w i l l continue to run n o m y as if the charge were still being fired by the spark, indicating that, until the temperature of the hot-spot falls, the miaure is effectively ignited at about the same instant as it would be by the spark. This effect is often called auto-ignition. As the temperature of the hot-spot increases still furthe, it 8 Downloaded from pad.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016 126 EXPERIMENTAL I N V E S T I G A T I O N I N T O P R E - I G N I T I O N I N THE S P A R K - I G N I T I O N ENGINE will eventually fire the charge before the normal spark ignition (pre-ignition), and the effect on the engine will be s w a r to that of over-advancing the spark timing. Inflammation of the charge will probably be more rapid with a hot-spot than with a spark, however, owing to the generally larger area of igniting surface .and greater time during which it is in contact with the mixture. Accordingly, the rate of pressure development in the cylinder may be much greater, resulting in excessive “negative” work before the end of the compression stroke. This would increase The occurrence of pre-ignition depends upon the margin between the temperature to which it is necessary to raise the hot-spot in order to cause pre-ignition, and the temperature at which the hot-spot operates under normal conditions. If these two temperatures are close, the engine is operating near to the pre-ignition limit, but if they are widely separated, the charge will be less liable to pre-ignite. This fact has been utilized in all the methods used for determining the relative pre-ignition tendencies of fuels. PRELIMINARY T E S T S WD LATE T.D.C. ~ D E G EARLY . CRANK ANGLE l\x To enable the maximum possible information to be obtained it was necessary to develop an artificial hot-spot, the temperature of which could be varied independently of the usual controlling engine-factors. This was particularly necessary because the factors which normally promote pre-ignition also promote knock, and to avoid knock, restriction of the investigation to fuels of high anti-knock value might have been necessary. While work on the development of a suitable instrument was in progress, and to enable information to be obtained quickly on the differences between fuels with regard to pre-ignition tendency, and the effect of engine variables on pre-ignition, some work was carried out with a thermocouple sparking plug. This plug, which has a low heat-factor, readily gives rise to pre-ignition under test conditions. The central electrode (nickelsilicon) is drilled to accommodate a fine platinum alloy thermocouple, the hot junction being located within inch of the end of the central electrode. Compensating leads connect the thermocouple to a millivoltmeter. For the preliminary tests, the plug was used as the spark source on the Ricardo E.6 variable-compression engine, and the compression ratio was raised in small increments, the thermocouple temperature and power output being recorded at each step until pre-ignition occurred. This was indicated by an abrupt rise in the reading of the sparking-plug thermocouple and a fall in engine torque. The charge would auto-ignite if the ignition were switched off at about one compression-ratio setting before that at which pre-ignition occurred. BENZENE Fig. 1. Development of Pre-ignition (Napier “Dagger” Unit) a Cylinder-pressure diagrams. b Rate of change of pressure diagrams. - - - - - - - Normal firing. Motoring. Hot-spot ignition. Nos. 1 and 2. After firing. Nos. 3,4, and 5. Auto-ignition. No. 6. Pre-ignition. --__ the rate of heat transfer to the combustion chamber walls. The process can become self-accelerating and the charge ignition could be sufficiently premature to bring a single-cylinder engine t o a halt. With a multi-cylinder engine, one cylinder of which is pre-igniting, the power developed by the cylinders which are not pre-igniting may keep the engine running, and the high temperatures and pressures developed on the compression stroke may cause mechanical failure of the piston or connecting rod. Knock and pre-ignition are different phenomena. Knock is due to the rapid combustion of the last part of the mixture following the initiation of flame by the spark, whereas preignition is ignition of the charge by a hot body before the spark occurs, or in the case of auto-ignition, during the combustion normally initiated by the spark. The flame front produced by a hot body is identical to that produced by the spark (Miller 1947). The two phenomena may be partly interdependent. Severe knock, by increasing the local heat transfer, can form a hot-spot in the combustion chamber which may lead to preignition. Pre-ignition, by giving early development of pressure and high maximum pressures, can lead toknock. Thereis evidence to suggest that the destruction often attributed to knock, for example, fracture or burning of the piston, may be due to pre-ignition itself caused by knock. 5 6 8 7 COMPRESSION RATIO 10 Fig. 2. Pre-ignition Tests with a Thermocouple Plug (E.6 Engine) Test conditions: engine speed, 2,500 r.p.m.; inlet-air temperature, 30 deg. C. ;jacket temperature, 70 deg. C.; maximum-power mixture strength;optimum ignition advance for each compression ratio. Some typical results are shown on Fig. 2. The sparking plug operates at a higher temperature on benzene than on either of the two paraffinic fuels under comparable conditions, and the temperature of the central electrode at which pre-ignition commences is about 100 deg. C. higher in the case of benzene. Under these conditions, the compression ratio for benzene when preignition commences is 6.4/1, compared with 95/1 for iso-octane. With methanol, pre-ignition occurred at the lowest compression ratio of the engine, namely, 5/1, even when the engine was throttled to give as little as 65 lb. per sq. in. brake mean effective pressure. Downloaded from pad.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016 EXPERIMENTAL I N V E S T I G A T I O N I N T O P R E - I G N I T I O N I N T H E S P A R K - I G N I T I O N E N G I N E 127 The effect of mixture strength on pre-ignition was then investigated. Results for representative fuels are shown in Fig. 3. The mixture strength is expressed as a percentage of the stoichiometric mixture. The change in pre-ignition characteristics owing to a change in mixture strength of a given fuel, was comparable with the variation in characteristics between different fuels. The mixture having the greatest tendency to pre-ignite, in these experiments, was between 10 and 15 per cent rich, that is, in the region of maximum-power mixture strength. The thermocouple plug, although useful for giving indications of differences between fuels, had the disadvantage that the tem- Pre-igniter. Various types of electrically heated hot-spot were tested before a suitable design was evolved. The prime requirement was that the temperature of the pre-igniting surface should be measurable. The hot-spot was also required to be’ resistant to erosion and other ill effects of high temperature, so that the recorded temperature would remain consistent in relation to engine conditions. The instrument used was an electrically heated hot-spot consisting of three and a half turns of Nimonic 75 wire. The hotspot was screwed into the combustion chamber diametrically opposite to the sparking plug (Fig. 4). Under normal operating 1.300 A. 1.200 1.100 I .m IONIZATION GAP 900 Km INVERTED PLAN SPARKING PLUG I SECTION AA MIXTURE STRENGTH -PERCENTAGE OF CHEMICALLY CORRECT STRENGTH b Fig. 3. Effect of Mixture Strength on Pre-ignition Tendencies of Various Fuels cE.6 Engine) Thermocouple plug temperature at pre-ignition. b Compression ratio limited by pre-ignition. Test conditions : engine speed, 2,500 r.p.m. ; inlet-air temperature, 54 deg. C.; jacket temperature, 100 deg. C.; optimum ignition advance for each &/fuel ratio and compression ratio. (I perature of the electrode was a fixed function of the engine conditions and so, as a pre-ignition device, it was somewhat inflexible. In addition, under given conditions there was a tendency for the compression ratio, limited by pre-ignition, to increase with engine runnhg time, suggesting that deposits collected in the air annulus surrounding the central electrode causing it to operate cooler under given engine heat-flow conditions. For these reasons an electrically hqated controllable pre-igniter was adopted at this stage. THE E . 6 R A T I N G METHOD Engine. A Ricardo E.6 variable-compression engine was used for these tests. This is a single-cylinder, poppet-valve, four-stroke unit, with a 3-inch bore and 4+inch stroke. A shrouded inlet-valve was fitted to improve the cyclic regularity of combustion. SECTION 88 j POSITION OF PISTON AT T.O.Ci I I Fig. 4. Diagram Showing the Positions of Pre-igniter, Ionization Gap, and Sparking Plug in the E.6 Cylinder Head conditions, and with a low compression ratio (6/1), the temperature of the wire was not sufficiently high to cause either preignition or auto-ignition on most normal fuels. The temperature of the plug was therefore raised and pre-ignition was induced by passing A.C., which was steadily increased, through the wire. The electrical energy (measured with a wattmeter) necessary to cause pre-ignition was taken as a measure of the pre-igniting tendency of the fuel. A high energy indicated a high pre-ignition resistance and wice oersu. I n the early tests the onset of pre-ignition was detected by noting the rapid increase in temperature d the coil, and the fall in torque of the engine. This procedure was dropped, however, as it placed a severe thermal load on the pre-igniter wire. The new method was to place an ionization gap in the combustion chamber in close proximity to the pre-igniter (Fig. 4). A potential difference of 40 volts was placed across the gap and the resultant current was fed to the “Y” axis plates of a cathoderay oscilloscope, a sharp rise in ionization current indicating the arrival of the flame at the gap. Under normal engine conditions, the ionization gap recorded the arrival of the W e , initiated by the spark, at about 10-15 deg. late (Fig. 5a). When the temperature of the pre-igniter was raised sufficiently, the mixture in the vicinity of the filament ignited before the arrival of the normal flame from the sparking plug. This was indicated on the oscillograph by the movement of the ionization signal to an Downloaded from pad.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016 128 E X P E R I M E N T A L I N V E S T I G A T I O N I N T O P R E - I G N I T I O N I N 10 deg. early f THE S P A R K - I G N I T I O N E N G I N E 10 deg. early f b d Fig. 5. Ionization Traces at Progressively Increasing Electrical Energy Inputs to the Pre-igniter u Normal combustion. by c, d Auto-ignition of increasing seventy. earlier timing. As more current was supplied to the pre-igniter, ignition of the charge occurred earlier in the cycle and the ionization signal was recorded correspondingly earlier (Fig. 5b, c, and d). The standard degree of pre-ignition adopted for these tests was that which gave an ionization signal 10 deg. early on the compression stroke. For tests in which it was desired to measure the relative normal running and pre-ignition temperatures of the fuels, the preigniter coil was made in two parts, one of alumel and the other of chromel, forming a junction in the middle of the coil. The A.C. component was removed by a filter and the resultant thermal E.M.F. was recorded on a millivoltmeter inserted in the heater circuit. The standard engine conditions used for the fuel rating tests were :Engine speed : 2,500 r.p.m. Compression ratio : 6/1. Ignition advance: optimum, 19 deg. early. 260 watts to induction air at 25 deg. C. (corresponding to an air temperature of 54 deg. C.). Coolant temperature: 70 deg. C. Mixture strength: that giving maximum pre-ignition tendency. ignition resistance near to that of the test fuel and the resulting rating was obtained by interpolation. The information which is available on the fuels used in these tests is shown in Table 2 . The second column of Table 1 shows knock ratings of fuels using the E.6 engine and the power drop highest useful compression ratio (H.U.C.R.) method of test. There is no simple relationship between the pre-ignition and knock tendencies of the fuels. Thus, benzene, which has a higher knock rating than iso-octane, has a much lower pre-ignition rating, and the alcohols have high knock ratings but low pre-ignition ratings. The paraffins, particularly iso-octane, have high pre-ignition ratings while the naphthene, cyclohexane, has a low rating. The behaviour of the aromatics is curious. The simplest, benzene, and the two with long side-chain substitutions, victane and cumene, have low ratings. Toluene, xylene, pseudo-cumene, and mesitylene, on the other hand, have high ratings. Methyl alcohol has a very low rating, lower than cumene, but isopropyl alcohol has a rating number in the region of 40. The Air Ministry 100-130 grade aviation fuel has the fairly high value of 77, probably partly owing to its mainly isoparaf€inic base and partly to the high lead content, namely, 5.47 C.C. tetraethyl lead per imperial gallon (T.E.L. per I.G.). The two 150 grade fuels tested, had ratings of 62 and 51. The lower rating Fuel Ratings Using the E.6 Method. Results of rating tests of the second fuel is possibly due to the presence of monomethyl on both pure substances and typical aircraft-engine blends, are aniline which, later work showed, favoured pre-ignition. shown in Table. 1. The ratings have been expressed with Table 1 shows, by the widely differing tendencies to pre-ignite reference to a scale on which iso-octane (2-, 2-, &rimethyl of the various members of the benzene group, that a high aromatic pentane) is rated at 100, and cumene (isopropyl benzene) at zero. content alone does not necessarily indicate a low pre-ignition Blends of these two fuels have ratings equivalent.to the volume rating. percentage of iso-octane. Thus, the blend 60 per cent iso-octane T o determine the differences which might be expected in the and 40 per cent cumene has a pre-ignition rating of 60. On this pre-ignition performances of typical British motor fuels, tests scale, a high number indicates a good pre-ignition resistance. were carried out on Petroleum Board reference fuels with octane Each test fuel was bracketed with two standard blends of pre- numbers varying from 71 to 87, and fuels of widely different Downloaded from pad.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016 E X P E R I M E N T A L I N V E S T I G A T I O N I N T O P R E - I G N I T I O N I N T H E S P A R K - I G N I T I O N ENGINE 129 composition blended to 80 octane (motor method). T h e composition of these fuels is given in Table 2. T h e pre-ignition ratings, together with the Octane numbers and E.6 H.U.C.R.’sare given in Table 3. Three ratings were obtained for each fuel, and the results were averaged. TABLE 1. hE-IGNITION AND KNOCK RATINGS OF A SELECTION OF FUELS 1 Fuel 1 100 I 10.96 .I 80 I 9.85 Iso-octane (2-,2-,4trimethylpentane) Neohexane Isododecane . . Mixedoctanes Alkylates Hydropetrol . . . . Hydropetrolwith6 C.C. T.E.L. per I.G. Cyclohexane . Benzene . . Toluene Pre-ignition Knock E.6 lE.6H.U.C.R. Xylenes (mixed) 1 66 9.30 72 10.85 77 10.05 56 7.95 75 91 . . . . . Ethyl alcohol (absolute) . . I 10 . . . . 100-130 grade aviation fuel StandardNo.4 . . . . . . 10.30 1 lI - 50 10.00 - . Hydropetrol . . Produced by hydrogenating a gasolene derived by cracking gas-oil . . 25 per cent neohexane, 75 per cent alkylates Standard No. 2 Standard No. 3 ~- 20 per cent benzolized hydropetrol, 30 per cent toluene 1 Experimental 150 grade, fuel 1 35 per cent hydropetrol, 30 per cent octanes, 15 per cent avaro, 10 per cent benzole, 10 per cent cumene Experimental 150 grade, fuel 2 i 85 per cent benzolized hydropetrol, 15 per 1 i cent octanes+2 per cent monomethyl ani- line(M.M.A.)with7.2c.c.T.E.L.perI.G. 100-130 grade aviation fuel Aromatic content 15.2 per cent, lead content 5.47 C.C. T.E.L. per I.G. Base for D.T.D. 230 60 per cent paraffins, 33 per cent naphthenes, 6 per cent aromatics, 1 per cent olehs I . .1 -19 . .I -32 1 62 78 92 1 13.80 1 13.96 I I . c 18.00 15.00 . “Pool” spirit with 3.35 C.C. T.E.L. per I.G. Percentage composition as for P.B.l P.B.4 (motor fuel) . 79 per cent “Pool” spirit, 21 per cent octanes with 3.6 C.C. T.E.L. per I.G. (13 per cent olehs, 9 per cent aromatics, 46 per cent paraffins, 32 per cent naphthenes) 80 octane, blend 1 (motor fuel) 24 per cent olefins, 22 per cent aromatics, 43 per cent paraffins, I 1 per cent naphthenes 80 octane, blend 2 (motor fuel) 57 per cent olehs, 19 per cent aromatics, 12 per cent paraflins, 12 per cent naphthenes 1238 70 - 71 51 62 I 80 octane, blend 3 (motor fuel) 1 per cent olefins, 16 per cent aromatics, 42 per cent paraffins, 41 per cent naphthenes with 1 C.C. T.E.L. per I.G. 80 octane, blend 4 (motor fuel) 24 per cent olefins, 17 per cent aromatics, 35 per cent paraffins, 24 per cent naphthenes with 1-5C.C. T.E.L. Der I.G. 80 octane, blend 5 (motor fuel) 19 per cent olefins, 12 per cent aromatics, 45 per cent paraffins, 24 per cent mphthenes with 2.5 C.C. T.E.L. per I.G. 80 octane, blend 6 (motor fuel) 9 per cent olehs, 29 per cent a r o m t h 30 per cent paraffins, 17 per cent naphthenes, 15 per cent ethyl alcohol 80 octane, blend 7 (motor fuel) 17 per cent olehs, 48 per cent aromatics, 24 per cent parffis, 11 per cent mphthenes 11.42 77 62 . P.B.3 (motor fuel) ~~ .I Experimental 150 grade fuel 2 - I 62 . . 35 per cent paraffins, 34 per cent naphthenes) 15.00 - 4 _____ Experimental 150 grade fuel 1 Base for D.T.D. 230 I -28 . Standard No. 3 100 . Industrial methylated spirits (74 over proof) . Methyl ethyl ketone I . .I loo Pseudo cumene (1-, 2-, benzene) . Isopropyl alcohol Alkylates :::1 : Gctane (iso- and n-butyl benzene) Methylalcohol 11.80 : :1 4 Cumene (isopropyl benzene) Di-isobutylene 1 TABLE 2. COMPOSITION OF THE FUELS ‘Pool” petrol (motor fuel) Racing blend 1 . 1 i , 1 1 i Racing blend 2 Downloaded from pad.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016 15 per cent olefins, 14 per cent aromatics, 40 per cent paraffins, 31 per cent naphthenes 50 per cent aviation gasolene (73 octane number), 50 per cent benzole 30 per cent ethyl alcohol, 40 per cent aviation gasolene (73 octane number), 30 per cent benzole 130 EXPERIMENTAL INVESTIGATION I N T O P R E - I G N I T I O N I N T H E S P A R K - I G N I T I O N ENGINE TABLE 3. PRE-IGNITION AND KNOCK RATINGS OF MOTOR TABLE 2. COMPOSITION OF THE FvELs-continued FUELS Composition Fuel . 75 per cent ethanol, 14 per cent benzole, 5 per cent acetone, 6 per cent water Racing blend 4 . 60 per cent methanol, 10 per cent aviation gasolene (73 octane number), 30 per cent benzole Racing blend 5 . 96 per cent methanol, 3 per cent acetone, 1 per cent castor oil . 80 per cent methanol, 10 per cent aviation gasolene (73 octane number), 10 per cent benzole Racing blend 3 ~~~~ Racing blend 6 Pre-ignition rating I. 1 11 P.B.l . P.B.2 . P.B.3 . P.B.4 . . . . 38 84 per cent methanol, 10 per cent benzole, 5 per cent acetone, 1 per cent castor oil 80octane,blend3 Racing blend 9 . 80 per cent methanol, 6 per cent benzole, 2-5 per cent acetone, 4 per cent water, 4.5 per cent aviation gasolene (87 octane number), 2 per cent nitrobenzene, 1 per cent lubricating oil 1 37 39 38 71 1 6.55 75 7.09 7.72 ~~~~~~~~~~~ I 56 I I 57 I 55 I 56 87 I 9.17 59 59 59 80.6 7.97 -58 ~ 800ctane~blend2 46 46 40 79.7 I 49 50 I 50 I 81.5 I 8.08 32 1 32 1 31 1 32 1 80.4 1 8.43 .I 10 1 9 I 8 1 9 I 70.0 . 38 44 7.82 80 octane, blend 4 80 octane, blend 5 80octane,blend6 75 per cent methanol, 14 per cent benzole, 5 per cent acetone, 6 per cent water 56.5 per cent methanol, 28 per cent ethanol, 4 3 per cent acetone, 7 per cent hydropetrol, 4 per cent water with 1 C.C. T.E.L. per I.G. Match for Gennan Grand Prix fuel (racing fuel) 85 parts methanol, 10 parts acetone, 1 part ether, 2+ parts water, 5 parts nitrobenzene Benzole blend Base petrol spirit with addition of 50 per cent benzene to give octane number of 70 . I 1 I 38 I 39 I 70.0 38 I 1 ~ NOTE.-The test conditions for knock ratings (H.U.C.R.’s) were identical with those for the pre-ignition tests except that the inletair temperature was 35 deg. C. and maximum knock mixturestrength was used. Base petrol spirit with addition of 40 per cent toluene to give octane number of 70 Again there is little relationship between the pre-ignition ratings and the knock ratings of the fuels. T h e ratings of the series P.B.1, P.B.2, and P.B.3 show the tendency of tetraethyl lead to prevent pre-ignition. The high rating of P.B.4 is a reflection of its high iso-octane content. Blends 6 and 7 (alcohol and benzole) have, as expected, low ratings. T h e relatively high rating of the sample of “Pool” sMrit is probably due to the presence of tetraethyl lead. Table 3 indicates the repeatability of pre-ignition ratings with fuels of this type. The repeatability for the fuels listed in Table 1 which had unusual volatility or combustion characteristics was, however, not so good. In Table 4 the pre-ignition ratings of some typical motor racing fuels are compared with their corresponding H.U.C.R. knock ratings. Although all these fuels have low pre-ignition ratings, the figures relate to maximum pre-ignition mixture strength, whereas the mixture generally used in racing practice is considerably richer, when the tendency to pre-ignite is much less. The fuels fall into two well-defined groups, consisting of those containing methanol, which have very low ratings, and those without methanol whose ratings are considerably higher, but not so high as those of fuels such as iso-octane and toluene. Knock Pre-ignition E.6 E.6 H.U.C.R. Fuel Racing blend 5 . . . . . Racing blend 6 . Racing blend 7 . . . Racing blend 1 Racing blend2 Racing blend3 Racing blend 4 Racing blend 8 Racing blend 9 Racing blend 10 Racing blend 11 . . 25 . . . . . . . . Match for German Grand Prix fuel cc I Toluene blend . THE 38 Benzole blend Racing blend 8 Toluene blend 38 1 I 80octane,blend 1 . 37 51 56 per cent ethanol, 3 per cent aviation gasolene (73 octane number), 40 per cent benzole, 1 per cent castor oil Racing blend 11 36 I . . I 1 1 I 1I 1 1 1 Racing blend 7 Racing blend 10 1 1 (motor F+t Second Third Average method) ratmg rating rating DAGGER” RATING METHOD Initially the pre-ignition investigation was mainly concerned with aircraft engine fuels, and, as there was some doubt whether the ratings obtained on the small normally-aspirated E.6 engine would be applicable to aircraft conditions, tests were commenced on a Napier “Dagger” aircraft engine cylinder. I .I . 1.I .1 . 12 0 -lor -168 -139 -137 . 1I . I -100 .. 1 Match for German Grand Prix fuel . . (without nitrobenzene) . 1 9.46 1 1 -77 -112 -96 13.71 1 ~ - - 1 2 6 0 1 I 2 -133 11.62 1356 12.85 12.61 13-36 j [ I ~ 1359 13.90 15.33 154.9 NOTE.-The test conditions for knock ratings were identical with those for the pre-ignition rests except that the inlet-air temperature was 35 deg. C. and maximum knock mixture-strength was used. Downloaded from pad.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016 EXPERIMENTAL I N V E S T I G A T I O N I N T O P R E - I G N I T I O N I N T H E S P A R K - I G N I T I O N ENGINE 131 Engine. A standard Napier “Dagger” piston, cylinder, and towards the exhaust-valve side of the combustion chamber. A valve gear were assembled with a Ricardo crankcase, crankshaft, Standard-Sunbury pressure pick-up was mounted in the and connecting rod. The engine bore is 97 rnm.and the stroke opposite side and was used for the detection and assessment of 95 mm. The compression ratio is 7.1 /I. pre-ignition. At the start of a test the air flow to the pre-igniter was such Pre-igniter. Preliminary tests were carried out using thermo- that pre-ignition was avoided on all fuels. The air flow was then couple plugs and coiled pre-igniters, but these were either in- slowly reduced, and was recorded when the standard degree of sufficientlyrobust or caused pre-ignition without added electrical pre-ignition was obtained. The amount of air necessary to prevent pre-ignition was used as a measure of the pre-ignition energy under supercharged conditions. Electrically heated pre-igniters are inherently unsuitable for tendency of the fuel. A high air-flow indicated a fuel of low preuse in a boosted engine because the temperature of a thermally ignition resistance and vice versa. The air flow was metered by isolated spot in the combustion chamber is suf€icient to cause inserting a restricting orifice in the cooling-air line to the prepre-ignition at a low boost without the addition of electrical igniter plug, and measuring the pressure drop across it. Typical rate of change of pressure diagrams are shown in energy. If, to cool the plug, the heat path to the outside air is improved, the electrical energy is not completely utilized in Fig. lb. The standard intensity of pre-ignition used in these heating the plug, an appreciable proportion being dissipated to tests corresponds with curve No. 3 of Fig. lb. The degree of the outside air. Under these conditions the energy input required pre-ignition was not severe but was deliberately chosen to avoid to cause pre-ignition is an inaccurate guide to the pre-ignition possible damage to the engine and pre-igniter from continuous operation under excessively severe pre-ignition conditions. tendency of the fuel. The standard engine conditions used for fuel rating were :An air-cooled pre-igniter plug will overcome these difliculties as it can be constructed to run at a temperature high enough to Engine speed : 3,000 r.p.m. cause pre-ignition on any fuel without added energy, the temCompression ratio : 7.1/1. perature being controlled by an internal air current. Ignition advance : 56 deg. early. The pre-igniter used with the “Dagger” rating method had a Induction air temperature: 80 deg. C. central tube constructed of Nimonic 75 or 80 material mounted Induction air pressure :35 inches of mercury (abs.). in an adaptor of stainless steel (Fig. 6). An alumel wire was Constant cooling-air supply to cylinder. Mixture strength: that giving maximum pre-ignition tendency. A‘R TPLY Fuel Ratings Using the “Dagger” Method. Table 5 shows the ratings obtained on a selection of fuels using the “Dagger” method. The results are again based on an iso-octane and cumene scale. F7 k Fuel TERMINALS FOR PYROMETER LEADS CHROMEL LEAD Pre-ignition rating Iso-octane (2-, 2-, Ctrimethyl pentane) Mixedoctanes Cydohexane Benzene . Toluene . . . ~ . Standard No. 3 . * S CENTRAL TUBE 75 OR 80) Fig. 6. Air-cooled Pre-igniter welded to the tube in the position of maximum temperature, to form a thermocouple. The temperature of the pre-igniter was raised by contact with the combustion gases and was controlled by a variable air supply conveyed to the inner end of the central tube by hypodermic tubing. The pre-igniter was screwed into a holein the combustion chamber near to the sparking plug and . . . ~~ . I ::I . 100-130 grade aviation fuel SPOT WELDED THERMOCOUPLE .) 92 / q I.IO0:82 I I PRE-IGNITION TEMPERA B IRE - I YI cd 3 I- 1.000- 100-130 grade aviation fuel with 25 per cent by volume of a 50/50 mixture of methanol 78 100-130 grade aviation fuel with 25 per cent by volume of a 50/50 mixture of methanol and water and 1 per cent by volume of Dromus oil 83 100-130 grade aviation fuel with 25 per cent by volume of a 70/30 mixture of methanol 72 100-130 grade aviation fuel with 25 per cent by volume of a 50/25/25 mixture of water, 88 2 n 5 I-~ and water 4 Some idea of the value of rating numbers in terms of engine variables is given in Table 7, which shows, for a selection of fuels, the induction pressures, limited by pre-ignition, at a constant cooling air flow to the pre-igniter. ATURE and water methanol, and ethanol 7 8 I 1 100-130 grade aviation fuel with 25 per cent by volume of ethanol 100-130 grade aviation fuel with 25 per cent 30 100-130 grade aviation fuel with 25 per cent 100 by volume of methanol by volume of water 100-130 grade aviation fuel alone, as in practice the induction pressure without injection will be lower. No. 2 mixture is that commonly used in Europe, while No. 3 is used in the United States. No. 2 mixture has a slightly higher rating, evidently owing to the presence of 1 per cent Dromus oil TABLE 7. PRE-IGNITION RATING NUMBWS IN TERMS OF INDUCTION PRESSURES Fuel Cumene . Iao-octane . Induction pressure, limited by pre-ignition, inches ofmercury (abs.) 35.0 * 1W130 grade aviation fuel . - .I .I 100-130 grade aviation fuel with 25 per cent by volume of water . . I 38.9 385 44.6 (added as a corrosion inhibitor). The beneficial effect of the Dromus oil is also reflected in the enhanced rating of blend No. 4. An increase in the methanol content of the mixture reduces the pre-ignition rating, and substitution of ethanol for methanol in the mixture raises it. PERCENTAGE OF ISO-OCTANE IN BLEI OF ISO-OCTANE AND CUMENE Fig. 7. Pre-ignition Tests on Blends of Iso-octane and Cumene (E.6 Engine) Standard test conditions. With a two-component fuel blend, for example, iso-octane and cumene, there is a simple relationship between the preignition temperature, the normal running temperature, and the pre-ignition tendency (Fig. 7), but this relationship does not exist for all fuels. It might be expected that the normal running temperature of the hot-spot would be related to the fiame temperature of the fuel. Fig. 8 shows calculated flame-temperatures and experimentally determined normal running temperatures, plotted against the carbon/hydrogen ratio of some pure hydrocarbons. The flame temperature appears to increase in a regular manner with the carbon/hydrogen ratio but, although there is a general tendency for normal running temperature to increase similarly, the scatter of points is much greater than with the flame-temperature curve. Cyclohexane, for instance, has a much higher normal running temperature relative to iso-octane than would be expected from its flame temperature. Methanol, with a normal running temperature of 940 deg. C. and a flame temperature of 2,585 deg. C., is another fuel which appears to give an unexpectedly high hotspot temperature in the engine (although the general level of Downloaded from pad.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016 EXPERIMENTAL I N V E S T I G A T I O N I N T O P R E - I G N I T I O N I N engine temperatures is usually much lower with methanol). The slope of the normal running temperature curve is similar to that of the flame-temperature curve, although the former is a mean and the latter is the maximum cycle-temperature. It appears, therefore, that there is a factor in addition to flame temperature which helps to determine the normal running temperature of the fuel. This factor is probably surface catalysis. T H E S P A R K - I G N I T I O N E N G I N E 133 This would appear to parallel some work of David (1943), who observed that the temperature recorded in the flame gases with a platinum-rhodium wire was higher than that recorded by a quartz-covered wire of the same material. No difference was observed, however, in the case of over-rich mixtures. David attributed this temperature difference to the catalytic effect of the platinum in the recombination of abnormally dissociated products in the flame gases. Coward and Guest (1927) observed higher surface ignition temperature with platinum than with other materials and suggested that the mixture in contact with the platinum was consumed so rapidly that it was incapable of propagating a W e , even though its temperature was higher than the ignition temperature of the mixture. This may explain why, in the engine experiments, although the pre-igniter temperature was higher with platinum, pre-ignition did not always occur at a lower compression-ratio with this material. T H E EFFECT O F E N G I N E VARIABLES O N P R E - I G N I T I O N Increase of induction pressure at constant speed, by increasing either the throttle opening or the boost pressure, promotes preignition by increasing the rate of heat transfer to the hot-spot, PRE-IGNITION TEMPERATURE D-9- 40 ’I I I NORMAL RUNNING TEMPERATURE FUEL MOLECULE CARBON/HYDROGEN RATIO Fig. 8. Variation of Flame and Pre-igniter Temperatures with the Fuel Carbon/Hydrogen Ratio (E.6 Engine) Standard test conditions. Single alumel-chrome1 wire pre-igniter. Calculation: compression ratio 7/1; mixture 12 per cent rich; mixture at inlet valve closure assumed at 100 deg. C.; atmospheric pressure, fuel completely vaporised. Pre-ignition temperature is a rather more complex variable and is related to the chemical properties of the fuel in addition to the physical properties. Insufficient laboratory data on these fuels are available to enable much to be said about the correlation between them. So far as engine pre-ignition temperatures are concerned there are some apparent anomalies. Thus, benzene, which in the E.6 engine was observed to have a generally higher pre-ignition temperature than iso-octane, had a lower preignition temperature than iso-octane in the “Dagger” unit. This again is possible due to a catalytic effect resulting from the generally higher level of pre-igniter temperatures in the E.6 engine. T H E EFFECT OF H O T - S P O T M A T E R I A L During the course of the investigations some interesting results were obtained when using hot-spots made of different materials. Comparative tests on thermocouple sparking plugs fitted with platinum and nickel-silicon central electrodes showed that, at mixture strengths ranging from very weak to about 20 per cent rich, the platinum plug operated at about 100 deg. C. higher than the nickel-silicon plug. With richer mixtures, on the other hand, there was a tendency for the platinum plug temperature to fall to the same level as that of the nickel-silicon plug, and at very rich mixtures, to be lower. Ib WEAK MIXTURE STRE$GTH-PER i0 CENT RICH Fig. 9. Effect of Mixture Strength on the Pre-ignition Tendency of Various Fuels (“Dagger” Unit) Standard test conditions. Fuels: x 100-130 grade. A Cumene. o Benzene. Iso-octane. Downloaded from pad.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016 134 EXPERIMENTAL I N V E S T I G A T I O N I N T O P R E - I G N I T I O N I N T H E S P A R K - I G N I T I O N E N G I N E intake air pressure. Curves of boost pressure, limited by preignition, were obtained at speeds of 1,750-3,750 r.p.m. with leaded iso-octane. Maximum pre-ignition mixture-strength and optimum ignition advance were used in all cases. The results are shown in Fig. 11. The boost pressure, limited and increasing the temperature and pressure of the charge. Increasing the compression ratio has a similar effect. The effect of mixture strength on pre-ignition (“Dagger” unit) is shown in Fig. 9. At the mixture strength givbg maximum pre-ignition, 5-10 per cent richer than the chemically correct mixture, the pre-ignition temperature is, in general, at a minimum and the normal running temperature is at a maximum. : 0 Similar results were obtained on the E.6 engine (Fig. 10). All the fuels were found to have similar characteristics, with regard to the effect of mixture strength on pre-ignition, the maximum pre-ignition mixture-strength occurred with a mixture about 2 1.500 I I I I I I-0 9 1 d YI a 700 6 + CL u1 I- z 17 500 Y 2 n W zz z a 3 i 20 ,sbo I 2,000 z 2,500 3,000 3,500 ENGINE SPEED-R.P.M. Fig. 11. Effect of Engine Speed on Pre-ignition, Limited by Boost Pressure (“Dagger” Unit) Test conditions : inlet-air temperature, 80 deg. C. ;ignition advance, optimum for each engine speed; constant cooling air-flowto cylinder; maximum pre-ignition mixture-strength. Fuel : iso-octane with 4 C.C. T.E.L.per I.G. MIXTURE STRENGTH -PERCENTAGE OF CHEMICALLY CORRECT STRENGTH Fig. 10. Effect of Mixture Strength on the Pre-ignition Tendency of Various Fuels (E.6Engine) Standard test conditions. Fuels: x Standard No. 4. A Toluene. 0 Iso-octane. + Benzene. 0 Cyclohexane. 10-15 per cent rich, that is, at about maximum power mixturestrength. This agrees with earlier findings on the engine, when the variable compression method of test with a thermocouple plug was used. With the possible exception of iso-octane, the pre-ignition temperatures do not appear to be greatly affected by mixture strength. The normal running temperature of the plug is, however, highest at about *maximum-power mixture, thus giving maximum pre-ignition tendency at this mixture-strength. The effect of speed on pre-ignition was determined in the “Dagger” unit by a test in which the heat factor of the preigniter was held constant by an unvarying cooling air-flow, and the onset of pre-ignition was controlled by adjusting the engine by pre-ignition, decreases with increase in speed from 1,750 to between 3,000 and 3,500 r.p.m., and at higher speeds rises again. A similar result was obtained with leaded cumene. The effect of engine speed on pre-ignition is twofold. Firstly, increase in speed decreases the induction time interval for ignition to occur, and consequently necessitates an increase in hot-spot temperature to cause ignition of the mixture ;secondly, in general, increase in speed increases the heat flow to the hotspot, thus raising the normal operating temperature of the hot-spot. Heat flow, however, increases with speed only up to a certain limit, and thereafter decreases owing to reduction in engine volumetric efficiency and other factors. This is shown in Fig. 12, the curves 0f which were obtained under normally aspirated conditions in the absence of both knock and preignition. The pre-igniter temperature peaks between 3,250 and 3,500 r.p.m., the same speed range over which the cylinder temperature becomes a maximum. Even before the peak, the rate of increase with speed has reduced considerably. At all speeds just below that for maximum heat flow therefore, the normal running temperature of the hot-spot increases at a greater rate than the pre-ignition temperature of the mixture, thus increasing the tendency to pre-ignite. At higher speeds, owing to a reduction in the rate of rise of normal running temperature, and eventually to its actual decrease, there is a reduction in the tendency to pre-ignite. Maximum pre-ignition tendency appears to occur at an engine speed just below that for maximum heat flow. Fig. 11 also shows a curve of knock-limited boost against speed, obtained on the “Dagger” unit under conditions similar 8 Downloaded from pad.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016 EXPERIMENTAL INVESTIGATION I N T O P R E - I G N I T I O N I N THE S P A R K - I G N I T I O N E N G I N E 135 m n. I40 m 1 4 y! I30 ’ ‘ 60 I,h f -I 1,500 I I I 2,000 2,500 3,000 ENGINE SPEED-R.P.M. I I 3,500 I 4,000 Fig. 12. “Dagger” Unit Characteristics over a Speed Range Test conditions : inlet-& temperature, 80 deg. C.;ignition advance, optimum for each engine speed; constant cooling air-flow to cylinder; maximum power mixture-strength; normal aspiration. Fuel: iso-octane plus 4 C.C. T.E.L. per I.G. IGNITION ADVANCE--DEGREES Fig. 13. Effect of Ignition Timing on Pre-ignition (“Dagger” Unit) Standard test conditions. Fuels: A Iso-octane. 0 Cumene. to those used in the pre-ignition tests. With knock, the decreasing induction time interval for the knocking reactions to occur, results in the anti-knock effect of increasing speed. Fig. 13 shows the effect of ignition advance on pre-ignition. The tendency to pre-ignite is increased as the ignition is advanced, owing to the rise in normal running temperature resulting from the generally higher temperatures and pressures associated with the more advanced timing. The pre-ignition temperature remains sensibly constant over the ignition-advance range. The effect of the inlet-air temperature on pre-ignition is small (Fig. 14). With iso-octane, a rise in inlet air temperature increases the pre-ignition tendency and with cumene a slight decrease is apparent. Tetraethyl lead appears to be more effective as a preignition prevention agent at the high temperature than at the low. The effect is similar, but not so pronounced, with aniline. Fig. 15 shows that the pre-ignition tendency increases with increase in cylinder temperature. Leaded .fuels were used because clear (unleaded) iso-octane and cumene knocked at cylinder temperatures above 230 deg. C. The increase in preignition tendency appears to be due to a combination of increase in normal running temperature and decrease in pre-ignition temperature, as the temperature of the cylinder is raised. T H E EFFECT OF F U E L A D D I T I V E S O N P R E - I G N I T I O N Experiments with the thermocouple plug showed that the addition of lead to a fuel lowered the normal operating temperature of the plug and raised the compression ratio necessary to cause pre-ignition (Fig. 16). For a given fuel, the effect of tetraethyl lead is approximately constant over a range of mixture Downloaded from pad.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016 136 E X P E R I M E N T A L INVESTIGATION I N T O P R E - I G N I T I O N I N T H E SPARK-IGNITXON ENGINE 1,000 enables the compression ratio, limited by pre-ignition, to be u: increased by 0.7511 whereas the same quantity of tetraethyl lead 2 would enable the knock-limited compression ratio to be raised n I by as much as 311. w 800 5 Fig. 17 shows some lead response curves obtained on the I“Dagger” unit. The curve on cyclohexane was commenced with dw a lead concentration of 4 C.C. T.E.L. per I.G. to avoid knock. 600 E :t: I- Z w& 400 I I 0 3 HEAT TRANSFER -300 5 Y 4 d c 250 ul r -kz--Lu 220 2-40 280 260 CYLINDER TEMPERATURE-0EG.C. 3M) (EXHAUST SIDE) Fig. 15. Effect of Cylinder Temperature on Pre-ignition (“Dagger” Unit) strengths from chemically correct mixture to 45 per cent rich. Table 8 gives the effect of the addition of 4 C.C. T.E.L. per I.G. on the pre-ignition characteristics of a number of fuels, at 3naximum power mixture-strength. The effect of the same Effect of tetra- Effect of tetra- Effect of tetraethyl. lead on ethyl lead. on ethyl lead on pre-igmtion compression normal running temperature, rauo a! temperature, deg. C. pre-igmtion deg. C. Benzene . Iso-octane. Mixed octanes Cyclohexane I . . . .I Standard No. 2 . I -10 -85 -50 -60 -77 1 1 I I +0.25 +0.75 +0.75 +0.75 +1.0 I I I I -77 -27 -54 -70 I2 13 AIR RATIO FUEL IS 14 I6 - Fig. 16. Effect of Lead on Pre-ignition with Iso-octane (E.6 Engine) Fuels : 0 Iso-oaane. x Isosctane plus 4 C.C. T.E.L. per I.G. Test conditions:speed, 2,500 r.p.m. ;inlet-air temperature, 54 deg. C. ; jacket temperature, 100 deg. C.; optimum ignition advance for each &/fuel ratio and each compression ratio. TABLE 9. ~ Fuel !--! I II Standard test conditions. A Iso-octane plus 4 C.C. T.E.L. per I.G. Fuels: o Benzene. C OF AT SELECTION OF AMINas ON AN ISOPARAFPINIC FUEL I Dropinrating Aromatic amine Aniline (2+ per cent by volume) ~~ . ~ . ~~~~ Monomethyl aniline (5 per cent by volume) Dimethyl aniline (5 per cent by volume) . Ethyl aniline (5 per cent by volume) . . Xylidines (5 per cent by volume) . .I .I .I .I . I 9.5 ~ 16.0 21.5 26.5 9.5 -66 There is a marked difference between the effects of the aromatic amines on pre-ignition and knock. Aniline and monomethyl aniline are strong anti-knocks but favour pre-ignition. Table 9 shows the effect of a selection of amines in an isoparaffinic fuel (Standard No. 2). The amines had very little effect upon either the normal running temperature or the pre-ignition temperature of the fuel. Downloaded from pad.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016 EXPERIMENTAL INVESTIGATION I N T O P R E - I G N I T I O N I N THE S P A R K - I G N I T I O N ENGINE 137 with fuels which also have a cclow’’temperature peninsula (Downs and Walsh 1949). Correlation between the effect of additives on the knock tendencies of the two types of fuel in the engine and their effect on the “high” and “low” temperature ignition-processes in the laboratory assisted in the establishment of this theory. Townend and Mandlekar (1933) found that the curve of ignition temperature against pressure, for a higher para5nic fuel, showed two main regions. One was a so-called “high” temperature region and, superimposed upon this, was the other, a so-called “low” temperature ignition-peninsula (Fig. 18). The chemical mechanisms assodated with combustion in these two regions appear to be quite different. Thus cool flames and peroxides are associated with ignition in the “low‘y but not in the “high” temperature region. It might be expected that the 600 I I I 2 3 I 4 I I 5 PRESSURE-ATMOSPHERES 1 1 I I 6 7 I8 Fig. 18. Ignition Temperature and Pressure Curve for Iso-octane Miaure :4.7 per Cent iso-octane; 95.3 per cent air. CONCENTRATION OF TETRAETHYL LEAD-C.C. PER IMPERIAL GAL. Fig. 17. Auto-ignition Lead Response (“Dagger” Unit) Standard test conditions. COMBUSTION UNDER PRE-IGNITION CONDITIONS At this stage it was decided to extend the additive investigation to cover other additives which are known to have a significant effect on hydrocarbon combustion, and it was hoped to learn something of the chemical nature of the pre-ignition process. Such an approach has proved valuable in an associated fundamental knock investigationand enabled two chemicalmechanisms whereby knock may occur to be isolated, one with fuels which oxidize only by a “high” temperature process, and the other pre-ignition process, occurring at relatively high temperatures and short induction periods when compared with knock, would occur by a “high” temperature mechanism. A test with formaldehyde on the E.6 engine appeared to confirm this. Formaldehydehas a pro-oxidant effect in the “high” temperature process in the laboratory (Chamberlain and Walsh 1949) and in confirmation it has been found to favour pre-ignition in iso-octane, cumene, benzene, and other fuels, in the engine, the effect being most pronounced with benzene. The results of further tests on additives in the “Dagger” unit are shown in Table 10. Tetraethyl lead is the only additive which resists pre-ignition in all fuels. Methyl iodide resists pre-ignition in the two fuels of low pre-ignition resistance, benzene and cumene, but promotes pre-ignition slightly in toluene and isooctane. Nitrogen peroxide, in agreement with its effect on the high temperature oxidation process, strongly promotes preignition. Organic peroxides, M.M.A.,amyl nitrite, nitro-benzene, and nitro-ethane, all have promoting effects. Acetaldehyde TABLE 10. EFFECT OF ADDITIVES Fuel ON PRE-IGNITION 0 Toluene b Nitro- Acetalde- Methyl Nitrogen hyde iodide tetroxide benzene 2 Benzene RATINGS . . .I Iso-octanewith 4 C.C. T.E.L. per I.G. . I -18 I -21 I -25 1 I -6 I -8 5 1 1.25 -18 I +17 1 -15 -35 -19 -20 -19 -35 -2 -6 -4 -15 Downloaded from pad.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016 !i -19 138 EXPERIMENTAL I N V E S T I G A T I O N I N T O P R E - I G N I T I O N I N T H E S P A R K - I G N I T I O N E N G I N E appears to be without effect. This is unexpected in view of the and this would put the mixture in a condition favourable for strong oxidizing effect of this substance in the laboratory ignition by the hot-spot, Conversely, a large hot-spot, such as an exhaust valve, might heat a volume of gas in its immediate (Chamberlain and Walsh). The effects of these additives on the normal running and vicinity thus predisposing it to ignite spontaneously. T o investigate the relationship, if any, between pre-ignition pre-ignition temperatures of the fuels is shown in Tables 11 and 12. I n most cases the effect of an additive is explicable in and running on, two blends were made to have the same terms of its effect on pre-ignition temperature. The effect on Co-operative Fuel Research (C.F.R.) knock rating but widely normal running temperature is in all cases very small, even with differing pre-ignition ratings, This was achieved by adding tetraethyl lead, which other experiments suggested lowered the 40 per cent toluene and 50 per cent benzene to a base to give, in both cases, a fuel of octane number 70. As expected, the normal running temperature. These tests, while not conclusive, are in accord with the theory benzole blend had a much lower pre-ignition rating than the that pre-ignition occurs by a chemical mechanism similar to toluene blend (Table 3). These fuels were tested for running on tendency in a typical road-vehicle engine and gave similar “high” temperature oxidation in the laboratory. TABLE 11. EFFECT OF ADDITIVES 1 1 1 Additive Fuel RUNNINGTEMPERATURES ON NO+ Temperature changes are given in deg. C. t-Butyl M.M.A. hydroperoxide 2: 1 2 Isooctane Cumene Benzene . . . .I 1 0 1 - :I 1 - 5 1 - 5 1 0 - Nitrogen tetroxide Nitro- Acetalde- Methyl hyde iodide benzene 1 0 0 0 0 0 Tetra- Formaldeethyl hyde lead 1.25 2 1 0.05 0 5 0 0 +I0 -10 0 0 - 15 - 0 TABLE 12. EFFECT OF ADDITIVES ON 1 +I0 +5 0 O I + 5 I - - l a Nitroethane 0 m - I G N I T I O N TEMPERATURES Temperature changes are given in deg. C. Fuel Additive t-Butyl hydroperoxide Percentage mol/mol Benzene . . . Toluene . Iso-octane Cumene - 2 Acetalde- Methyl hyde iodide 5 2 -75 - -80 -20 -20 -35 -35 -35 -35 -60 -70 -60 0 - “RUNNING 1 -45 1 -105 ON” “Running on” is the term applied to the condition in which the engine continues to idle after the throttle has been closed and the ignition switched off. There has been considerable interest in this phenomenon since the 1939-45 war. One suggestion with regard to running on has been that it is related to high-speed auto-ignition from a hot-spot, and another that it is due to spontaneous ignition of the fuel-air mixture. If, under normal running conditions, the temperature of a combustion chamber hot-spot is below that required to give autoignition or pre-ignition, it is possible that, as the ignition is switched off and the speed falls, the increase in the time interval available for ignition would favour hot-spot ignition. I n addition, it might be expected that the hot-spot would require time to cool from its normal operating temperature and so, for a time, would be above the equilibrium temperature for the low speed condition. Spontaneous ignition, on the other hand, would be assisted at the low speed condition by the low rate of air flow through the induction pipe, often in close proximity to the hot exhaust, resulting in a high inlet mixture temperature, and also, in some cases, to a low rate of cooling of the combustion chamber at low speeds owing to poor circulation of the coolant. The two phenomena are related in that some pre-reaction of the mixture undoubtedly o c m s under running on conditions Nitrogen tetroxide 2 Nitroethane 1.25 1 0 -- I -20 0 0 I -35 1 -45 -45 I -50 -70 1 - 1 -40 ~~ -80 results. The toluene blend was rather worse than the benzole. This result suggests that there is little relation between highspeed pre-ignition and running on. CONCLUSION The experimental investigation did not yield much information on the destructive effects of pre-ignition in service, because it was necessary to limit the severity of conditions in order to obtain repeatable observations. Pre-ignition of sufficient severity to stop the engine was frequently encountered on the E.6 engine, and had no serious consequences. The E.6 engine is, however, a sturdy unit of large thermal capacity and would be capable of absorbing a considerable increase of heat flow without iil effect. I n a multi-cylinder engine of small thermal capacity, such as an air-cooled military aircraft unit, the destructive effects of preignition are more obvious (Hundere 1948; Hundere and Burt 1948; Cattaneo and Viscia 1947). Pre-ignition failures often take the form of burning of the piston on one side, caused by a combination of increased heat-flow and blow-by owing to failure of the piston rings to retain the increased pressures. There have been instances of pistons having burned through at the crowns. Photomicrographs show that pre-ignition can fuse the aluminium alloy of the piston, but knock erodes it, that is, physically removes the metal without melting it. The major damage is, therefore, Downloaded from pad.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016 . EXPERIMENTAL I N V E S T I G A T I O N I N T O P R E - I G N I T I O N I N T H E S P A R K - I G N I T I O N E N G I N E brought about by the increased heat flow j the pressures may not be more than 40 per cent higher with pre-ignition if the normal spark-ignition is near the optimum. The most common source of pre-ignition is the sparking plug. Ceramic plugs, which were generally introduced in 1940, are capable of withstanding more arduous Conditions than mica plugs but, particularly in the early days of the war and in the United States, there was trouble owing to insulators cracking. Pieces of the ceramic became trapped in the space behind the electrodes, and these, being thermally isolated, could become very hot and cause pre-ignition. The use of a low-tension surface discharge plug for ignition purposes would eliminate the possibility of pre-ignition from the sparking plug. Their “self-cleaning” action permits a high heatfactor without the risk of oiling at low-duty conditions. Work was carried out on these plugs in Germany during the war (Combined Intelligence Objective Sub-committee 1945) and further experiments are now proceeding in Europe and the United States. If certain difficulties owing to erosion of the electrode material and the insulator can be overcome, these plugs will probably prove useful for both road-vehicle and aircraf3-engine applications. The exhaust valve is another probable source of pre-ignition. Its temperature will, in general, be lower than that of the sparking plug, but its area is greater. Over-heating of the valve as a result of improper seating, and scaling of the crown from oxidation and corrosion, may lead to pre-ignition. Data obtained by the N.A.C.A. have shown that in a particular engine preignition was obtained from the exhaust valve at a temperature of 730 deg. C. when the crown was scaly. With a clean and smooth crown no pre-ignition was obtained at a temperature of 1,000 deg. C. (Sutor, Corrington, and Dudugjian 1947). Combustion chamber deposits are possible sources of hot-spot ignition. Lead compounds may be particularly harmful because they may be deposited under low head-temperature, low power, and weak mixture cruising-conditions. I n these circumstances, if the power is suddenly increased, pre-ignition may occur before the deposits can be burned off. A dZiculty in the diagnosis of pre-ignition failures, owing to combustion chamber deposits, is that they are often purged from the cylinder by the pre-ignition process (Hundere 1948). The differences in pre-ignition ratings between aviation fuels of similar octane numbers blended from normal components are small. Pre-ignition would become of importance only if the use of special blending agents, such as M.M.A. or cumene or injection fluids containing methanol, were contemplated. It is as well, however, to have the pre-ignition resistance of the fuel always in mind and to remember that it is not so closely controlled as the knock rating. In the ordinary road-vehicle engine there would appear to be little danger at the moment with normal petrols. Pre-ignition troubles are more likely to be encountered in motor racing engines, particularly where alcohol blends are used. Even then, pre-ignition is more a problem of correct engine design than fuel quality, and providing steps are taken to ensure that all potential hot-spots are kept cool, no pre-ignition troubles should be experienced. AcknowZe&ements. The investigation described in this paper was carried out in the laboratory of Ricardo and Company Engineers (1927), Ltd., on behalf of the Shell Petroleum Company, Ltd. The authors are indebted to both these companies for permission to publish this paper. APPENDIX REFERENCES ALQUIST, H. E., and MALE, D. W. 1944 N.A.C.A., A.R.R. No. E4125, “Trends in Surface Ignition Temperatures”. 139 BIERMANN, A. E., and CORRINGTON,L. C. 1943 N.A.C.A., A.R.R., “Relation qf Pre-ignition and Knock to Allowable Engine Temperatures”. CATTANEO, A. G., and VISCIA,E. P. 1947 JI. SOC.Automotive Eng., vol. 55, p. 54, “Pre-ignition Piston Failures Diagnosed in Single-cylinder Tests”. CHAMBERLAIN, G. H. N., and W ~ S HA. , D. 1949 Revue de 1’Inst. Franpis du Pet., vol. 4, p. 301, “The Slow Oxidation of Diisopropyl Ether m the Temperature Range 360-460 deg. C.”. COMBINEDINTELLIGENCE OBJECTIVESSUB-COMMITTEB1945 Report No. 33, Appendix No. 38A. CORRINGTON, L. C., and FISHW,W. F. 1948 N.A.C.A. Technical Note No. 1637, “The Effect of Pre-ignition on Cylinder Temperatures”. COWARD,H. F., and GUEST,P. G. ’ 1927 Jl. h e r . Chem. SOC., vol. 49, p. 2479, “Ignition of Natural Gas-air Mixtures by Heated Metal Bars”. DAVID,W. T. 1943 Nature, vol. 152, p. 278, “Temperature of Flame Gases”. DOWNS,D., and WALSH,A. D. 1949 Nature, vol. 163, p. 370, “Knock in Internal Combustion Engines”. HUNDERE,A. 1948 Symposium on Aircraft Reciprocating Engines and Their Fuels (Co-orbating Research Council, Inc., 20 Rockefeller Plaza, New York 20). HUNDERE,A., and BURT,J. A. 1948 Quarterly Trans. SOC. Automotive Eng., vol. 2, p. 546, “Pre-ignition and its Deleterious Effects in Aircraft Engines”. MALE, D. W. 1946 N.A.C.A. Technical Note No. 1131, “The Effect of Engine Variables on the Pre-ignition Limited Performance of Three Fuels”. MALE,D. W., and EWARD,J. C. 1945 N.A.C.A. Report No. 811 and A.R.R. No. E5411, “Pre-ignition Limited Performance of Several Fuels”. MILLER,C. D. 1947 Trans. SOC.Automotive Eng., vol. 1, p. 98, “Roles of Detonation Waves and Auto-ignition in Spark-ignition Engine Knock as Shown by Photographs Taken at 40,000 and 200,000 frames per sec.” N m n , D. M., and TOWNEND, D. T. A. 1938 Science of Petroleum, vol. 4, p. 2958, “Ignition in Gases with Special Reference to Knock Problems”. ROTHROCK, A. M. 1941 Trans. SOC.Automotive Eng., vol. 48, p. 51, “Fuel Rating-Its Relation to Engine Performance”. SERRWS,M. 1937 Publications Scientifiques et Techniques Du Minisdre de l’Air, No. 115, “Auto-ignition in Internal Combustion Engines”. SPENCER, R. C. 1941 N.A.C.A. Report No, 710, “Pre-ifpition Characteristics of Several Fuels under Simulated Engine Conditions”. SUTOR,A. T., CORRINGTON,L. C., and DUDUGJIAN, C. 1947 N.A.C.A. Technical Note No. 1209, ‘‘Operating Temperatures of Sodium-cooled Exhaust Valve as Measured by a Thermocouple”. TOWNEND, D. T. A., and MANDLEKAR, M. R. 1933 Proc. Roy. SOC.,vol. 141, p. 484, “The Influence of Pressure on the SDontaneous Ignition of Inflammable Gas-air Mixtures (Butane-air Mi;rturey’. 1934 Proc. Roy. SOC.,vol. 143, p. 168, “The M u e n c e of Pressure on the Spontaneous Ignition of Inflammable Gas-air Mixtures (Pentane-air kfixture)”. VISCHNIEVS&,R. 1947 Academie des Sciences Comptes Rendus, vol. 225, pp. 565 and 992, “Tendencies of Various Fuels to Self-ignition at a Hot Point”. Downloaded from pad.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016 140 n . U iscus sion in London Dr. E. A. WATSON, O.B.E., M.I.Mech.E., said that although the employed as a pre-ignition detector in the case of the “Dagger” authors stated that with the ordinary road vehicle engine there engine, instead of the ionized gap used with the E.6. He was at the moment little risk of trouble due to pre-ignition, the wondered whether the authors had any preference. The ionized paper was of great value and importance and deserved careful gap was a very simple and effective device which he had found study, particularly by the designers of sparking plugs, some of of great value. He found some difficulty in interpreting Fig. 12, as the whom he hoped were present. He recalled that the phenomenon of auto-ignition, which in meaning of normal running temperature was not clear. He asked its extreme form became pre-ignition, had once formed the whether he was correct in assuming that that was the running standard method of igniting the charge, and many would temperature of the pre-igniter unit with a constant but arbitrary remember the old hot tube ignition on the gas engine, which was cooling air flow so chosen that no pre-ignition occurred in the just the pre-igniter turned inside out, with heat applied to start absence of boost. He was interested in the authors’ reference to surface discharge the process. In some experiments which had been carried out in 1935, plugs, but he thought they were incorrect in ascribing a selfin conjunction with a firm of spark-plug manufacturers, at cleaning action to the plug. The plugs were not self-cleaning. In ilm on their the time of the development of the platinum pointed plug. I n fact, many depended for their action on a carbon f those experiments they had endeavoured to determine the surface. The real point was that a plug of that type must be used relationship between the temperature of the electrode which with a condenser discharge Circuit which was inherently capable and in consequence there was no would give auto-ignition and its total area. They had used a of dealing with heavy plug very similar to the one shown in Fig. 4. Two interesting necessity to run the electrodes hot to burn off any oil. One must, points had emerged, the 6rst being the way in which surface however, remember that, other things being equal, a very cool area affected the auto-ignition temperature. With a small plug reduced the range of mixture strength which could be exposed area, temperatures as high as 1,200 deg. C. were neces- ignited. sary; with a large area 700 deg. C. or less sufficed. That effect was brought out indirectly in the authors’ results. With the Mr. J. G. WITHERS, B.Sc., A.M.I.Mech.E., said that there thermocouple plug, and again with the pre-ignition unit used could surely be few fields in which there were as many indepenon the E.6 engine, electrode temperatures for pre-ignition were dent variables as there were in the study of pre-ignition, and the of the order of 1,000-1,300 deg. C. With the unit used on the authors had provided a mass of experimental data which added “Dagger”, with its much greater surface, temperatures were very very considerably to knowledge of the subject. much lower and were of the order of 550-700 deg. C . a s shown He had one point of criticism, which could be rectified easily in Figs. 10 and 11. While the figures when the paper appeared in the PROCEEDINGS. The second point of interest was that by suitably controlling gave full details of the tests concerned, the tables did not do so; the current to the heater, it was possible to run the engine in a and in reading the paper it was sometimes difficult to be certain perfectly normal manner on auto-ignition. The engine in to which engine the data applied. question had been a small water-cooled stationary one, and no The most important factor affecting pre-ignition seemed to be di5culty had been experienced in making it carry full load the temperature at which the hot spot operated, and that was satisfactorily. Increasing the igniter current was, of course, the shown to be affected markedly by fuel type. The effects of fuel equivalent of advancing the ignition. That experience might, of latent heat and fuel composition on maximum flame temperacourse, set some of the inventors-particularly electronic engi- tures were discussed in the paper, and it was shown that those neers-thinking. It would, on paper, be possible to devise an did not give the whole story. Intensity and duration of flame electrical system responsive to speed and inlet depression which luminosity probably played a large part in determining the would vary the heater current and hence the ignition advance amount of heat received by the hot spot, and that would account in any way desired. One drawback, however, was the life of a for the marked influence of fuel additives in some cases. Further, plug of that type. the radiation of heat from combustion gases was largely a result The relative ease with which pre-ignition was obtained with of the presence of carbon dioxide and water vapour. Thus, if benzene and methanol was, he thought, a surprise to most of oxygen was present in the fuel, as in alcohols and ethers, there them. He remembered, from old motor-cycling days, operating would be more water vapour present in the combustion gases on nearly neat benzol, and although the engine did not knock, which would tend to increase the temperature of the hot spot. it would sometimes thump in a manner strongly indicative of The same argument would apply for water when injected pre-ignition; but it was certainly surprising to note the marked separately. However, it must be remembered that the charge difference between benzene and other fuels shown in Figs. 2 cooling associated with the higher latent heat would operate and 3. in the reverse direction, and that illustration served to show The surprisingly poor behaviour of methanol under those how complicated the subject was. conditions was, however, still more unexpected. No figures He did not agree with the authors that the knock rating had appeared to be given for the plug electrode temperatures with nothing to do with pre-ignition. On the contrary, he believed methanol as fuel, but they would appear to be appreciably lower that pre-ignition was similar in character to detonation, the than with benzene, and the effect was presumably not due to major difference being that the former was the auto-ignition of a excessive electrode temperature. He wondered whether that thin layer adjacent to the hot spot, which started a progressive meant that either the self-ignition temperature or the igniting flame front; and the latter was the auto-ignition of a pocket of gas with the sudden release of energy. energy of methanol was lower than that of normal fuels. The authors ascribed the difference in plug electrode temThe data in Fig. 3 were not inconsistent with the theory that peratures between various fuels and various electrode materials the pre-ignition temperatures were directly related to the knock in part to catalysis and surface combustion. That was doubtless ratings, given as highest useful compression ratios (H.U.C.R.’s) a contributory factor, but it was just possible that some of the in the paper. The reversal of benzene in the “Dagger” series heat received by the electrodes was electrical and came from the of tests was not surprising, as here the conditions were much ignition circuit. Although the heating due to the current was not more severe (using the term in its knock rating sense) than those large, it was not negligible, and its value might depend on a for the E.6 engine, and the E.6 H.U.C.R.’s would no longer be number of factors, including the type of fuel and the ionization appropriate. It was known that benzene gave a lower knock present in the combustion chamber. With regard to ionization, rating than iso-octane when tested by the C.R.C. F-3 method, he would be interested to know why the Sunbury indicator was and F-3 results were probably more appropriate to the “Dagger” r&ge, Downloaded from pad.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016 DISCUSSION O N P R E - I G N I T I O N I N THE SPARK-IGNITION ENGINE 141 tests. It was clear that in most cases pre-igniting conditions test bed experimental conditions the effect seemed to be almost would be more severe than detonating conditions, as both the proportional. speed and charge temperature were likely to be higher in the A second feature that might indicate that spontaneous ignition former case. That would tend to penalize such hydrocarbon was causing the trouble was the very high pressures obtained on types as the aromatics. a running on cycle. In his experience, the pressure was at least Thus, pre-ignition might not be different in character from 50 per cent higher than the normal maximum pressure in the normal detonation. However, the overriding factor seemed to engine, and in addition the maximum pressure was not reached be the normal operating temperature of the hot spot, which early in the cycle but considerably after top dead centre. He depended on many other variables. supposed that was not surprising, as the heat exchange conWith regard to after running, he had a theory that it might be ditions in the charge when the engine was running slowly must associated with high exhaust residual concentrations, as engines be lower. with large valve overlaps seemed to be more prone to it ;and he felt that a multi-cylinder engine might back-charge its own Mr. C. D. CARMICHAEL, B.Sc. (Eng.), A.M.I.Mech.E., said cylinders by a pulse charging type of operation and that exhaust that perhaps the most interesting genera1 conclusion on the residual might tend to reduce pre-igniting temperatures. programme of tests outlined in the paper was the fact that the relative pre-ignition ratings of fuels in the normal aspirated and B.Sc. (Eng.), A.M.I.Mech.E., said that in the supercharged engine were similar, with the exception of Dr. J. H. WEAVING, although, as the authors stated, the automobile engine was not toluene. He wondered whether that could be due to the fact that, very adversely affected by pre-ignition troubles, he did not feel in the case of the supercharged engines, the conditions were of they could afford to ignore the problem or fail to take advantage moderate supercharge only. The test conditions had approxiof the work which the authors had done. He felt that the auto- mated to the cruising output rather than the take-off output, and mobile engine was not very far off the condition of auto-ignizion, very likely that had been in the interests of conserving the life and with increasing compression ratios and the possibilities of of the unit ;but he thought that probably it had had some effect varying fuels, there might well be need for all the information on the actual ratings obtained. He asked the authors whether they had made any attempt to that could be obtained on the subject of fuels. It would be. interesting to apply the fuels to a standard multi- vary the position of the pre-igniter in the cylinder, or whether cylinder engine rather than to a single-cylinder engine. I n the the position had been settled for them by the necessity of using single-cylinder work that had been done, the many variables had existing bosses in the cylinder head. He had been interested to note the low pre-ignition rating of been kept constant, which was, of course, most essential for a systematic study ;but he thought that it would be very useful if monomethyl aniline (M.M.A.), which had been a blending agent the gap between academic work and its practical application in the early type of 100/150grade fuel. Some years ago, when commercial testing this fuel on the Sabre engine, it had been found that could be bridged by applying the results to an actual ’ there was some combustion roughness when running under engine, which the authors might have done. For instance, if the temperature of the auto-ignition device combat power conditions, without there being, however, any was kept constant while using various fuels of various ignition effect on the engine life or the anti-knock response of the fuel. properties, the increased rates of pressure rose and the increased By using an ionization gap apparatus similar to that described, maximum pressures would give another less systematic but incipient pre-ignition had been detected. In later blends of the perhaps more practical rating. Stated in a slightly different fuel without M.M.A. the combustion roughness had entirely way, if the authors could make an engine that would pre-ignite disappeared. He had been surprised to note that the pre-ignition ratings of easily (he thought that an exhaust valve having a very narrow seat would suffer badly from pre-ignition), fuels could be methanol additives was, relatively speaking, low, and he rated in a very practical, though less scientific manner. He felt wondered whether, in the “Dagger” unit, that was due to the that such information would be very valuable in specifying the fact that the induction air temperature was maintained constant various fuels, and it might perhaps give some answer to the at 80 deg. C. for those tests, as also for the test of the base fuel. question why the benzene and alcohol fuels were so bad from In practice a considerable amount of charge cooling occurred with water-methanol injection, with a consequent reduction in that point of view. He had a slight criticism to offer on the question of rating the hot spot temperature. He asked the authors whether the study of pre-combustion fuels by means of measuring the heat transfer. The temperature of the hot spot was presumably the criterion, as the area was reactions in an engine undertaken by Pastell* in the United constant ;but the heat transfer was a function not only of tem- States of America offered any possibilities for exploring the preperature but also of the local velocity that was prevailing. He ignition problem without the complications arising from the existence of the flame front. would like to have the authors’ opinion on that point. He wondered whether the authors had any evidence to The problem of running on was one with which the automobile engineer was not unfamiliar. Most overhead valve engines that indicate that detergent oils, which were now coming into profilled well would run on under test-bench conditions, even if minence in the road vehicle field, had any tendency to promote they did not actually do so on the road. If the engine was run pre-ignition from combustion chamber deposits. He had seen under full load, at high speed, and the throttle closed, the engine it stated that such oils left deposits having high melting points would generally run on. It had been a very great surprise to him which prevented their elimination from the engine cylinder. to read the authors’ conclusion that benzene should be so bad from the auto-ignition point of view as to lead one to anticipate Mr. G. M. BARRETT, M.B.E., T.D., B.Sc., A.M.I.Mech.E., that it would cause running on. That conclusion would suggest said that it was appreciated that running on had a high nuisance that the authors had made a very considerable contribution to value, and it was aggravated by any condition which tended to the theory of running on, namely, by establishing that the cause raise the inlet charge temperature. In order to investigate the was due to the spontaneous ignition of the charge as distinct more fundamental aspects of running on, some work had been from auto-ignition due to a hot spot in the cylinder. done at the Research Laboratories of the Shell Group. Being suspicious of the suggestion that spontaneous ignition A Ricardo E.6 single-cylinder variable-compression engine was the more likely cause, he had tried a benzene petrol mixture had been used under the following basic operating conditions :to see whether that really was so, and he had obtained the Speed : 2,500 r.p.m. same effect as the authors. That would seem to suggest that Coolant outlet temperature : 165 deg. C. auto-ignition was not the trouble as far as running on was conMixture temperature: 40 deg. C. cerned. B.m.e.p.: 95 lb. per sq. in. That conclusion might perhaps be borne out by a few observaCompression ratio : 65/1. tions in experiments which he had done. Running on was very sensitive to octane number. There was no running on with a * PASTELL, D. L. 1950 Quarterly Trans. SOC.Automotive Eng., small multi-cylinder engine with 100 octane fuel, whereas with vol. 4, p. 438, “Pre-combustion Reaction in the Spark-ignition a lower octane fuel the running on was quite considerable. Under Engine”. 9 Downloaded from pad.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016 DISCUSSION ON P R E - I G N I T I O N I N T H E S P A R K - I G N I T I O N E N G I N E 142 I70 Spark timing: 30 deg. E. &/fuel ratio: 12.5/1. The engine had been run until the required conditions were stabilized and then simultaneously the throttle was closed to the normal idling position, the brake load removed and the ignition switched off. The duration of after run was then measured from the moment at which the engine reached idling speed. Fig. 19 showed the effect of a jacket temperature and com- I I INITIAL JACKET TEMPERATURE-DEGL. Fig. 19. Effect of Jacket Temperatures and Compression Ratio on the Duration of Running On pression ratio on the duration of running on. It would be seen that at each compression ratio there was a limiting coolant temperature, below which the engine would not run on. That was shown more clearly in Fig. 20. An increase in compression ratio from 5.5 to 6-5 was equivalent to a decrease in limiting jacket temperature of 46 deg. C. uI 8 4 I I80 lJ ; 160 n LA d 0 MIXTURE TEMPERATURE-DEG.C. a. E r Fig. 22. Effect of Mixture Temperature on Critical Jacket Temperature ; -< 35 loo I- 120 e U v I50 80 ?, d 5 6 7 8 COMPRESSION RATIO 9 10 3 k 2 2 Fig. 20. Minimum Coolant Temperature for Running On 140 5 2 130 It was soon found that the duration of running on was not a good comparator for measuring the effects of other variables, 2 since the repeatability was poor. However, the limiting jacket i temperature could be repeated within f5 deg. C., and that was adopted as a measure of the running on tendency and was 4 referred to as the critical jacket temperature. Fig. 21 showed the effect of the cooling rate on the duration ofrunning on, using three different cooling rates. For those tests, 2 I*' 8 10 12 14 16 18 AIR/ FUEL nmo 20 Y Fig. 23. Effect of &/Fuel Ratio on Critical Jacket Temperature Downloaded from pad.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016 24 DISCUSSION ON P R E - I G N I T I O N I N THE S P A R K - I G N I T I O N ENGINE 143 after cutting the ignition the coolant circulation had been main- the inlet closed 53 deg. after bottom centre. The second had an tained and passed through an electrical heater which controlled overlap of 30 deg., and the inlet closed 65 deg. after bottom the rate of cooling. For subsequent tests, the cooling rate had been centre. The run on times were 7 min. 17 sec. and 6 min. 10 sec. controlled at 2.6 deg. C. per min. respectively. Presumably with less valve overlap there was less Fig. 22 showed the effect of mixture temperature on critical tendency for material to be drawn back into the cylinder from jacket temperature for two fuels, one being unleaded with an the exhaust manifold. That material would contain fuel which octane number of 70, and the other being leaded with an octane had been subjected to compression without combustion during number of 80. The effect of mixture temperature was more the previous cycle, and would be in an activated state and hence marked with the leaded fuel. more prone to auto-ignition. Also, the later closing of the inlet Fig. 23 showed the effect of the air/fuel ratio on critical valve would reduce the cylinder charge at low speeds and hence jacket temperature. A ratio of about 17/1 showed the greatest lower the compression pressure. In that engine, as in the E.6, tendency towards running on. However, air/fuel ratio had only leaded fuels tended to run on longer than unleaded fuels of the a comparatively minor influence, and the change of ratio from same octane number. 17 to 10 only increased the critical jacket temperature by 5 It was suggested that the reason why running on was more deg. C . prevalent since 1945 than before 1939 might be partly attributable T o examine the effects of sparking plug hardness, five types to seven causes :of plugs, ranging from the soft to very hard, including an aero (1) The use of hot spots was more widespread, and the hot plug having platinum fingers, had been tried, but there had been spots had become hotter. no apparent difference in running on tendency in that engine. (2) Radiator sizes had shrunk at the same time as coolant The effect of fuel quality had been examined using various circulation rates at running speeds had increased, so that there fuels of different types. It had been found that there was an was less reserve to mop up heat at idling when the circulation approximately straight line relationship between octane number rate was low. and critical jacket temperature, although there had been con(3) There was less underbonnet ventilation with modem siderable scatter at the points about that line. For example, at crocodile type bonnets, and therefore the temperature of the the 80 octane number level the spread in critical jacket temair entering the carburettor was higher. perature had been 39 deg. C., which was almost equal to that (4) Interior heaters had become popular, and in order to observed for the change of compression ratio from 5.5 to 6.511. obtain sufficient heat, it was common practice to adjust the The characteristics of the combustion process during running thermostat to a higher jacket temperature. on had been examined with the aid of a cathode-ray oscilloscope (5) The use of anti-freeze mixtures of the glycol or glycerine and an electro-magnetic pick-up. The disconnected sparking type had become more general, and that increased the templug had also been used as an ionization gap. It had been found perature of the liners and cylinder heads, even though the that combustion took place on the average only one in five cycles. coolant temperature and circulation rate were maintained Usually the combustion pressure rise and theionization occurred constant. at approximately 30 deg. after top centre, accompanied by a (6) In the expectation of an eventual increase in the octane normal combustion thump. Occasionally, however, the comlevel above Pool standards, engine designers may have tended mencement of combustion pressure rise had advanced to 10 deg. to increase compression ratio, while at the same time retarded after top centre, and that had been accompanied by an extra ignition advance, thus keeping the anti-knock requirement heavy thump. the same. Running on was independent of spark advance, but It had been found that the maximum rate of pressure rise was strongly influenced by compression ratio. under full load at 2,500 r.p.m. was two and a half times that (7) Pool gasoline was leaded, whereas before 1939 the under running on conditions at 240 r.p.m. The maximum presgreater proportion of gasoline consumed in the United sures had been 975 and 300 Ib. per sq. in. and the b.m.e.p.’s Kingdom had been unleaded. 95 and 6.5 lb. per sq. in. respectively. It was therefore conMr. S. B. BAILEY, M.Sc. (Eng.), A.M.I.Mech.E., said that sidered that the unpleasant mechanical knock associated with running on was due to a moderate rate of pressure rise occurring on p. 128 of the paper the authors had shown how their preat low engine speed. The rate of pressure rise per crankshaft ignition ratings were expressed with reference to a blend of isodegree under running on conditions was four times the rate octane and cumene, the rating being equivalent to the percentage of iso-octane present. That method of fuel rating was easily which was obtained under full load at 2,500 r.p.m. For investigating some of the more practical considerations, understood for ratings between zero and 100, but it was not a modem four-cylinder car engine had been used. I n that stated how the negative ratings, which appeared in Table 4, or engine the duration of run on had been used as a measure of the positive ratings greater than 100, which appeared in Table 5, the running on tendency under various conditions, and the had been derived. Clearly there could not be a negative quantity of iso-octane figures quoted were the average of eight runs. For examining the relative contributions of the four cylinders, the sparking in an iso-octane/cumene mix; nor could there be more than plugs had been used as ionization gaps during run on, and it had 100 per cent iso-octane present. He wondered whether the authors had employed a method been found that No. 2 cylinder was firing for the whole time, the duration being 7 min. 6 sec. ;No. 3 cylinder was firing for similar to that used for grading the anti-knock ratings of fuels 2 minutes ; No. 1 cylinder for 30 seconds ; and No. 4 cylinder superior to iso-octane. If so, he thought they ought to explain not at all. However, in all cylinders fring had been intermittent their method, as there appeared to be no precedent for the use of when it had occurred, and had taken place only about one in five negative fuel ratings. Referring to Table 1, which gave the pre-ignition and knock cycles. The maximum ionization voltage had been reached ratings of a selection of fuels, he observed that there was a large between 30-40 deg. after top centre. In that engine the induction manifold was immediately above number of dashes in places where one would like to see figures. the exhaust manifold, and the two were bolted together at the I n particular, the E.6 highest useful compression ratio was not hot spot joint face. To investigate the effect of a drastic reduction stated for the two experimental blends of 150 grade fuel. He in mixture temperature, an air space of inch had been left asked whether those two missing figures could be supplied, or between the hot spot joint faces. That had had the effect of whether they had been purposely withheld. reducing the run on time from 6 min. 3 sec. down to 40 seconds. Mr. J. B. PERRETT,A.M.I.Mech.E., said that he wished to The corresponding induction temperatures measured at the hot refer briefly to residual exhaust gases, which had been mentioned spot drain bosses had been 125 and 33 deg. C. under full power by Mr. Withers and which might have some real effect on autoconditions, and 88 and 59 deg. C. at the termination of run on. ignition. That modification had naturally made the engine very sluggish There were engines-usually two-stroke engines-which, by in responding to the throttle, and it would have had little their construction and design, must run with a considerable flexibility on the road. proportion of residual exhaust gases. Those engines were very T o study the influence of valve timing, two different cam- subject to auto-ignition, as would be appreciated by anyone shafts had been tried. The first had an overlap of 42 deg., and who had at any time operated a two-stroke engine. + Downloaded from pad.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016 . 144 DISCUSSION ON P R E - I G N I T I O N I N T H E S P A R K - I G N I T I O N ENGINE If that was meant to imply that the rate of initiation and subsequent propagation of the explosion was the same, then he thought the statement needed some qualification; and indeed he felt that the paper itself bore out the fact that there was a difference. The time-lag between the spark and the initiation of the explosion was extremely small; otherwise, the high-speed engine would not be possible. On the other hand, the hot spot ignition would vary according to both temperature and the area of the hot spot. Mr. H. MOORE,M.Sc. Tech. (Manch.), Assoc.I.Mech.E., in That raised the question of the theory of the process of ignition. making a suggestion concerning the mechanism of pre-ignition, While there was abundant support for the purely thermal theory, said that it appeared that there was a spontaneous ignition there were many adherents of the theory that spark-ignited reaction probably in a very localized region very near to the hot engines were assisted by electrical effects. He thought that the spot, and that the actual ignition reaction was only in that region. paper had given support to the latter viewpoint. Moreover, It could be seen from Fig. 1 that the rates of pressure rise were Muller-Huebrand*, in an investigation into the ignition of not very different for the various conditions of pre-ignition. explosive gas mixtures by low voltage sparks, had asserted that Moreover, as stated in the paper, it was known that the rates of an electric spark could not simply be considered as a hot h m e propagation in the cylinder were very similar whether the cylinder, and that the molecules were activated not only by the flame was initiated by a spark or by a hot spot. The authors heat release, but also by the electrical effects. The absence of any had shown in a most interesting way that the conventional appreciable time-lag with the electric spark gave support, he additives had a profound effect on pre-ignition. It was also considered, to the combined electrical and thermal theory for known that additives were almost without effect on normal flame the process of ignition. travel, so that it would appear that they could act only in the Mr. HADLEY asked for more information on the surface initial stages of the pre-ignition. He asked the authors whether the effect of two possible discharge plugs, and wondered what advantages, if any, they variables had been investigated. One was the general volume of had over the others, apart from the advantage of not causing the region that was subjected to heating round the plug. The hot spots; and the nature of the change in the equipment that other was the air movement in the combustion chamber. It was necessary to produce the condenser discharge circuit to seemed to him that that should be of considerable importance which Dr. Watson had referred. in sweeping away the products before they built up to a sponMr. A. PORTER, A.M.I.Mech.E., said that he had been trying taneous ignition. For instance, he asked whether the effects had been compared with the E.6 unit masked and unmasked on its to burn benzole in a converted three-cylinder gas engine, with a bore of 18 inches and a stroke of 22 inches, and he had been inlet valve. He also asked the authors whether they had applied their having trouble. The compression ratio had been varied from sampling technique to a study of the types of reaction which 4-5 to 6, and the mixture strength from theoretical up to about 10 per cent rich. occurred during the initial stages of the pre-ignition reactions. It had been possible to run the engine from 20 minutes to Many interesting points had been made in the discussion, and he wished to make a reference to the point about two-stroke half an hour without pre-ignition, and after that period proengines. He also had considered that matter, and he thought ignition due to overheated exhaust valves had commenced at that probably in the two-stroke engine the compression ratios somewhere about 80 per cent of the stroke. The revolutions had were so low that true spontaneous ignition reactions were not fallen back and the power disappeared. He had med water injection to as much as 10 per cent of the weight of the benzole, likely to take place. and that had brought about a very considerable improvement j Mr. L. GRIFFITHS, M.I.Mech.E., said that it appeared to him but the engine was not yet running correctly. He did not know whether those facts gave any further informathat the cathode-ray oscillograph and the ionization method of signifying pre-ignition was not only reliable, but would give a tion on the effect of engine size on pre-ignition. higher precision than other methods. * MULLER-HLLL~BIWND, D.1948 Engineers’Digest (BritishEdition), It was stated on p. 126 of the paper that the flame front pro- vol. 9, p. 50, “The Ignition of Explosive G a s Mixtures by Low-voltage duced by a hot body was identical to that produced by the spark. Sparks”. The point he wanted to make was that those engines were not at all subject to running on. He therefore thought that there was a case there for saying that running on was in no way connected with auto-ignition. It might be that there were factors which he had not taken into consideration, but there was the fact that on an engine which was very subject to auto-ignition, there was absolutely no running on. Discussion in Coventrv Dr. J. N. H. TAIT,M.Eng. (L’pool), M.I.Mech.E., said that There seemed under those conditions, little chance of any hot pre-ignition was not a trouble with which he had had much spot in the combustion chamber causing running on. The conexperience, except when sparking plugs ran at too high a tem- clusion he came to was that ignition was due to spontaneous perature. He had, however, some experience with “running on”, combustion of the mixture. He found subsequently that particularly on a small four-cylinder engine during its develop- various things which would reduce the temperature of the mixment stage. That engine was very prone to running on. If it was ture at the end of compression would also reduce the running run at maximum power at about 4,000 r.p.m., and the throttle on time. For example, reducing the induction pipe hot spot then closed to a position which would normally run the engine temperature, and therefore the mixture temperature, reducing at about 1,000 r.p.m. and the ignition was switched off, the the cooling water temperature, or the compression ratio slightly, engine would run on for several minutes. That was a severe con- had the same effect. The thing which had the most effect was dition for investigation. He noticed that during that time the reducing the valve overlap, and the conclusion there was that induction pipe which was heated by contact with the exhaust the mixture of the charge with products of combustion was manifold got hotter because the gas speed was reduced and heat reduced, which in turn reduced the compression temperature. soaked through from the exhaust manifold. When that engine Reducing the valve overlap had also another benefit; it enabled had stopped running on, it was possible to start without switch- a slower idling speed to be used, which, in itself, reduced ing on again. He found that if the hot spot on the induction pipe running on. The combination of those changes solved the was heated by means of a blow lamp, one could start the engine running on problem. At the time he thought that the trouble after it had been stopped for an hour without switching on. was due more to spontaneous ignition of the charge, than to a Downloaded from pad.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016 DISCUSSION O N P R E - I G N I T I O N I N THE S P A R K - I G N I T I O N E N G I N E 145 hot spot in the combustion chamber. When he read the paper It was always possible to cut out pre-ignition entirely by he wondered whether the temperature of the artificial hot spot, choosing the correct grade of plug. That covered many different which the authors had introduced, would give some indication types of engine with compression ratios up to 1011 unsuperof the spontaneous combustion temperature of mixture. He charged and boosts up to 20 lb. per sq. in. gauge with 6 5 / 1 comthought that the mixture in contact with that hot spot, would pression ratio. rise to the hot spot temperature and that would give some The internal cooling of alcohol fuels appeared to keep piston indication of the spontaneous ignition temperature. head and valve temperatures adequately low, but very much He was surprised to see in the “Dagger” unit that the pre- harder sparking plugs had to be used on the same engine when igniter temperatures were in the region of 500-700 deg. C., as a change was made from petrol-benzol to alcohol mixtures. shown in Fig. 10, whereas in the E.6 unit, the results of which One particular engine would invariably crack cylinder heads were shown in Fig. 11, the pre-igniter temperatures were more on petrol-benzol, but never on alcohol blends. in the region of 1,100-1,300 deg. C. He noticed too that the Since the sparking plug electrode was the hottest part of the “Dagger” unit required an ignition advance of 56 deg., whereas combustion chamber, it would appear that a much cooler plug in the E.6 unit only 19 deg. was stated. As ignition advance was required to keep the temperatures below the pre-ignition sometimes gave an indication of the amount of turbulence in an temperature for alcohol mixtures. engine, he queried whether it was possible that the “Dagger” In general, the tendency for pre-ignition increased with unit had less turbulence, and whether, therefore, a lower pre- r.p.m. up to the peak of the power curve, and the cooler the igniter temperature gave pre-ignition, as the gas was not swept plug the higher the speed that could be reached before the onset so rapidly past the pre-igniter unit. of pre-ignition. He was very interested in Fig. 19 which gave ignition temperatures of an iso-octane mixture, and he noticed that reactions Mr. L. H. DAWTREY, M.I.Mech.E., said that he was most could take place at quite low temperatures in the order of 300 deg. C. That, compared with recent work carried out in interested in what the authors said about running on as his the United States of America, showed that chemical reaction experience was identical. The majority of his firm’s products have been side valve engines, and running on trouble was not took place during the compression stroke. experienced with them except under very arduous conditions, such as when the engine was used in a car with very low axle B.A. (Oxon.), A.M.I.Mech.E., Mr. A. N. L. WCLACHLAN, thought that running on was a phenomenon which was practically ratio, which was heavily loaded and driven hard. It was true, unknown before the 1939-45 war, and since compression ratios in his experience, that the overhead valve engine had been more had changed little since then, it was presumably connected with prone to it. Having experienced that and having run a prototype fuel quality. His experience was that side valve engines were not motor car for 10,000 miles with no running on, he thought he prone to running on, but overhead valve engines were. He asked was lucky and that the new engine was free from that trouble, whether the authors considered that due to surface volume ratio but after stripping the engine for examination and rebuilding, it never ceased to run on. That showed how elusive it was to or to greater turbulence in the engine. It also appeared to be a low-temperature effect since, under trace that fault. He was quite sure running on was largely connected with temidle conditions, exhaust valve temperatures could hardly exceed 300 deg. C., and sparking plugs would be relatively cool, whilst perature of induction pipe. With development of the tliermomaximum compression pressures would be of the order of 60 lb. static hot spot manifold and using insulating washers, which per sq. in. or 4 atmospheres. He wondered whether that was were too thin, between the inlet manifold and the exhaust, the connected with cool flame combustion, and whether the authors temperature of the manifold was too high and that simultaneously could give any suggestions as to what did occur to cause self- introduced quite pronounced knocking and running on. The two came together. T o cut down idle speed to the minimum was ignition under such conditions. He had recently been engaged in an investigation into that essential when one had an engine inclined to run on. His recollection was that when using 80 octane fuel no running on problem and the following facts had emerged :was experienced. He was sure the authors would begin to feel (1) Running on was very much affected by atmospheric that running on was of more importance to the industry than humidity, it was quite difficult to produce it on a really damp pre-ignition. day and in that respect it showed a similarity to detonation. There existed one well-known type of engine which in(2) It was much less sensitive to induction pipe hot spotting corporated a solenoid to cut off the slow-running mixture when than to combustion chamber temperature. the ignition was switched off, thus preventing m g - o n . The (3) A very slow idle was beneficial, which would seem to slow running mixture was important and a distorted carburettor conflict to some extent with para. 2 of the section on running could lead to running-on. on, p. 138. (4) On a particular engine, although running on could be Mr. W. M. HEW, M.I.Mech.E., discussed the question of produced with cylinder head water temperatures as low as 140 deg. F. (60 deg. C.) with the thermostat in position, it running-on in relation to cylinder head shape. He said that in could not be produced up to 200 deg. F. (94 deg. C.) with the his experience he found that the lozenge head was more prone thermostat removed. That suggested that sluggish water flow, to that trouble than any other type of head, and believed that it due to the thermostat restriction, produced stagnant areas and was due to the lack of turbulence which existed in the lozenge head. high local temperatures in the combustion chamber. Experiments had shown that by making various modifications ( 5 ) Back pressure also appeared to aggravate it-presumably to the shape of the head and the directional flow of the inlet due to high temperature residual gasses. (6) Increasing the by-pass channel area had a beneficial ports, a certain improvement could be obtained, but of all the modifications tried none could guarantee a 100 per cent cure effect confirming the above. (7) Addition of tetraethyl lead to “Pool” petrol up to when production tolerances were taken into account. Experiments in that direction were carried out with various mixture 2 cu. cm. per gal. had no significant effect on running on. The technique used to produce running on was to run the and water temperatures, and, although those points again had engine for 1 minute at 1,500 r.p.m. light, then allow it to idle, an influence, none of them singly could effect a complete cure. With the hemispherical type of head, however, it was possible followed by an immediate switch off. I n the worst conditions running on continued for up to 15 seconds, accompanied by to run at least half a compression ratio higher than with a lozengeviolent detonations. He welcomed the authors’ views on those shaped head without encountering the slightest tendency for running-on to occur. It was suggested that that was probably points. Referring to true pre-ignition, experience with a large number due to the greater turbuience and shorter flame travel. of racing engines, both before the 1939-45 war and since, using alcohol blends in general very similar to blends 4,6, and 8 tested M.I.Mech.E., confirmed that with aircraft iMr. C. H. FISHER, by the authors, had convinced him that the sparking plug was engines it was part of the pilot’s drill to put the control into the usual initiator of pre-ignition. “idle cut-off’ position before switching off the engine. Downloaded from pad.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016 DISCUSSION ON P R E - I G N I T I O N I N THE SPARK-IGNITION ENGINE With regard to the authors’ remarks on heavy flywheels, he did not appreciate before. Mr. Heynes mentioned that running had observed years ago how well engines idled in the region of on appeared to be confined to four-cylinder engines, and he did 200 r.p.m., but that was no doubt connected also with the very not understand why running on was not experienced on larger simple valve timings of that period. The authors had also or smaller engines. He noted in Table 6 the effect of various methanol and water commented on rich mixtures on idle, but he thought that the cars mentioned by other speakers were nearly all experimental percentages on the pre-ignition reading, and wondered whether cars where the idle would be carefully adjusted, or at any rate, he was to assume from that that the use of methanol water more so than with the average motor car. The authors were injection for take off conditions, at least for military aircraft, was aware that idle mixtures were considerably richer than that liable to encourage dangerous conditions of pre-ignition, or giving normal idle, but in the ordinary way the mixture ratio whether they would be of such short duration that there could required for idle was of the order of 0-08to 0.09 fuel/air ratio. not be any serious mechanical damage during the take off or He did not consider that engines of vehicles being run under climbing conditions, when the engine was working at its hardest. close supervision by a motor-car firm would be maladjusted to He had also heard of exhaust valve failures on engines, and asked the extent of bringing on the phenomenon of running on by itself. whether there would be any connexion between pre-ignition and exhaust valve failures, or whether that was more likely to be due Mr. E. W. COY, B.Sc. (Lond.), B.A. (Cantab.), M.I.Mech.E., to the high lead concentration which arose in aircraft fuels. said that there was one point connected with running on that he He asked for further information on Dromus oil. 146 Discussion in Glasgow Mr. E. B. STEAD,A.M.I.Mech.E., said that the paper gave a glimpse of the research that had been carried out on a feature of an internal-combustion engine, which did not necessarily give any trouble, but was studied in order that everything could be learnt about it. He was reminded, when looking at the fuel tables, of the Report by the Fuel Research Committee, published in the Proceedings of the Institution of Automobile Engineers, and against each fuel was the knock highest useful compression ratio. That term, at one time, usually applied to a particular engine, but it had now been superseded by octane rating. It was interesting to observe that it was still in use and co-related to the octane number. He was impressed by the simplicity of the methods used, especially the method of obtaining the variable and controllable hot spot, and he wondered whether, under varying conditions of hot spot, they had had any trouble with overheating exhaust valves. He was surprised to note, in regard to thermocouple sparkplugs, that maximum electrode temperature occurred at maximum power mixture strength. I n experiments with which he himself had been familiar, where thermocouple plugs were used, the highest temperature usually occurred at the chemically correct mixture. He wondered whether it was a feature of that particular type of engine which the authors used in their test. He said that his worst personal experience of pre-ignition was the result of a change from 80 octane petrol to 71. Most engines were forced to operate on that low octane fuel after the 1939-45 war. The usual procedure, on such occasions, if two gaskets were not fitted, was to retard the ignition, but the efficiency of the engine dropped if that course was taken. Another common occurrence, dealt with briefly by the authors, was the phenomena of running on. That tendency was more pronounced in the larger petrol engines, especially the size now supplanted by the Diesel. He recalled one engine of that type which ran on, switched off, for 5 minutes, and sometimes, before stopping, reversed for an equally alarming period. The engine improved after running on full power, but when it was started cold some time later, run for a short period and then switched off, it failed to stop. That engine was later found to have burnt inlet valves, and he wondered whether that gave the authors any indication to what may have been the cause. Mr. G. H. LEE,B.A., A.M.I.Mech.E., said that, in Table 1, there appeared to be no relationship between pre-ignition and octane rating. The authors’ octane rating was based on H.U.C.R., and he wondered whether the C.F.R. motor or Research method might show a closer connexion with the tendency to pre-ignite. The authors stated that the same factors, which influenced detonation, also influenced pre-ignition. He agreed that detonation, although it should be avoided for other reasons, did not produce h a r m effects mechanically, but that pre-ignition, which might result from severe detonation, usually resulted in severe damage. M.I.Mech.E., asked whether, with petrols Mr. S.WIGHTMAN, of a known octane number, say, 71, one should expect those petrols to have the same anti-knock rating and pre-ignition rating, or whether they varied widely. The authors mentioned that the octane rating usually determined the other ratings. He asked the authors for further information on low-tension surface discharge ignition plugs. Mr. D. L. CAMPBELL, B.Sc., A.M.I.Mech.E., referred to the problem of running 6n and said that a six-cylinder engine of the commercial type of 120 h.p. was bad in that respect, but the fourcylinder model was quite good. He asked the authors to enlighten him on that point as he could not understand why the four-cylinder engine should be better. Mr. J. M. FORBES, M.I.Mech.E., asked, in the case when an engine was running normally, whether a particular cylinder would pre-ignite and hence spoil the performance of the engine. He assumed, that when the engine was actually running, the same defect occurred on the road. Mr. J. A. KEMP,M.I.Mech.E., would liked to have heard more of the influence of turbulence and swirl on pre-ignition in general and wondered whether the authors had investigated in any manner the effects of petrol injection on pre-ignition and whether turbulence improved or spoilt the ignition performance. Other speakers asked whether the E.6 H.U.C;R. figures quoted in Table 1 were in any way affected by cylinder capacity, and whether it was a fact that with a smaller capacity one could increase the compression ratio. Downloaded from pad.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016 147 Authors ’ Reply Pre-ignition. Mr. D. DOWNS and Mr. J. H. PIGN~GUY wrote in reply to Dr. Watson’s comments on the low pre-ignition rating of methanol that such evidence as they were able to obtain showed that the sparking-plug electrode and pre-igniter temperatures with that fuel were appreciably higher than those obtained with a normal fuel, in spite of the fact that the general engine temperatures were reduced by the high latent heat of evaporation of the methanol. Some figures obtained on a small air-cooled engine running at 3,000 r.p.m. are shown in Fig. 24. Those observations were made in the absence of auto-ignition and therefore point to some form of catalytic action at the electrode since the maximum flame temperature of methanol was slightly lower than that of the particular hydrocarbon fuel used. I10 2 ‘W 9 d 0 u1 m5Q -I I 4 y! f 80 70 oxidation process occurring after a relatively short induction period. Knock arose from a lower temperature high pressure process occurring in a body of gas which had been considerably pre-reacted before ignition. The chemical nature of the autoigdions was very different in the two cases as was fully discussed in the paper and in a later paper by Downs and Wheeler (1951-52)*. Mr. Carmichael was correct in assuming that the “Dagger” engine was only moderately supercharged for the sake of the durability of the unit. The load and speed were nevertheless appreciably higher than in the E.6, and, in addition, the engine was air-cooled, and employed a totally different form of preigniter. The agreement between the results on the two test units was sufficiently good to justify confidence in the validity of the general conclusions as to the relative merits of the fuels, at any rate at the mixture strength employed (that was, that giving maximum pre-ignition tendency). In all the “Dagger” tests the intake air temperature was maintained constant at 80 deg. C. Any advantage arising from charge cooling by the use of alcohol would therefore be directly obtained. In reply to Mr. Bailey, the pre-ignition ratings outside the limits of the reference fuels employed were obtained by simple extrapolation. That was found to be reasonably reliable and each result was checked several times. The H.U.C.R.’s of the 150 Grade fuels were in the region of 13 : 1. More recently tests have been made in the E.6 engine on a number of gaseous fuels. Both the pre-ignition and knock values have been obtained, and the results are shown in Table 13. 8w u : 4 7M: “I Rating Fuel d 3 4 I g6M: P ~ I Pre-igfition E6 W Methane ;5@ d 2 Propane Iso-butene . . Town’s gas . Iso-butane 401 ~~~~ Fig. 24. Figures Obtained on a Small Air-cooled Engine Running at 3,000 r.p.m. : :. Ij . . ~ 3:’ . 115 ~~ An ionization gap could not be used as a pre-ignition detector in the “Dagger” unit because the pre-igniter was placed near to the sparking-plug in order to make it run as hot as possible. In those circumstances there would be little difference in the time of flame arrival at an ionization gap, whether the flame originated at the spark plug or at the pre-igniter, and that would make it impossible to detect the early stages of auto-ignition by those means. An increase in the rate of change of pressure consequent upon pre-ignition was found to be easily observable and proved to be a very satisfactory alternative to the ionization method. With regard to the meaning of the n o d running temperature referred to in Fig. 12, that was, as Dr. Watson assumed, the temperature of the pre-igniter with a constant, but arbitrary, cooling-air flow sufficient to prevent pre-ignition in the absence of boost. While the authors agreed with Mr. Withers that pre-ignition and knock were both due to auto-ignition of some part of the fuel/& charge, the essential difference between them lay in the conditions of temperature, pressure, and time under which they occurred. Pre-ignition was a high temperature, low pressure Carbon monoxide Hydrogen . . Downloaded from pad.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016 I 1 1 Knock E.6H.U.C.R. 17.6 12.1 104 121 10.2 -56 -86 11.4 - 8.7 18.1 148 AUTHORS’ REPLY O N P R E - I G N I T I O N I N THE S P A R K - I G N I T I O N E N G I N E combustion chamber, there appeared a prospect of cooling specific areas by means of directional spray. That was done in German military aircraft engines (Daimler-Benz and B.M.W.) during the 193945 war, and was believed to have been effective in reducing pre-ignition. Replying to Mr. Moore, no investigation into the general volume of heating around the pre-igniter was made, neither were the effects of turbulence examined. There was, however, appreciable swirl in the E.6 engine arising from the use of a shrouded inlet valve which was used in order to improve the cyclic regularity of combustion during the experiments. NO work was carried out with the gas sampling valve for the purpose of examining the gas reactions leading to pre-ignition. In answering Mr. Grifliths, the authors’ reference to the nature of the flame front was intended to imply that once ignition had occurred, combustion proceeded in a normal way. That was evidenced by the fact that when the engine was auto-igniting the electrical ignition could often be switched off without any change in speed or load being observable. The rate of initiation of combustion could possibly be lower in these circumstances, but since a greater quantity of mixture would be involved owing to the larger area of the pre-igniter the final engine result would be similar to that with electrical ignition. When the pre-igniter temperature was high the rate of propagation was extremely rapid and if not checked would bring on pre-ignition of damaging severity. Mr. Porter’s experience would appear to indicate that the pre-ignition observed on his engine was due to local overheating and was reduced by the cooling effect of the injected water. It seems unlikely that engine size per se could be responsible for the trouble, but more probably it was due to operating and design factors. For example, the centre of an uncooled piston crown, particularly if made of cast iron, could reach a very high temperature. Exhaust valves and spark plugs could also cause trouble if not sufficiently cool. Carbon deposits could also give rise to pre-ignition. Mr. Maclachlan’s experience of pre-ignition in high-duty racing engines provides valuable confirmation of the necessity for careful selection of sparking-plugs. The authors are glad to have his agreement on their observations on the opposing effects of alcohol on sparking-plug and general engine temperatures. His confirmation of the effect of engine speed on pre-ignition was reassuring in that laboratory observations carried out at relatively low speeds fell in line with practical experience at high engine speeds. Mr. Coy asked whether the use of methanol/water blends used for extra power output at take-off on aircraft engines were liable to increase the risk of pre-ignition. The practice was now a well established one and the authors were not aware of any trouble arising from its use. The authors have no experience of any connexion between pre-ignition and exhaust valve failures. It was, however, not unlikely that an exhaust valve already “blowing”, and running at an unusually high temperature, could give rise to pre-ignition. It was known that the high T.E.L. concentration in aircraft fuels was responsible for a certain amount of exhaust valve trouble. Dr. Weaving’s comment on the validity of rating fuels by measuring the heat transfer may be answered by saying that all fuels were compared under the same engine conditions, so that the velocity of gases past the pre-igniter would be the same in each case. Replying to Mr. Wightman’s question, it is unlikely that the current motor fuels of a nombal70 octane number would vary very widely in either knocking tendency or pre-ignition rating. In answering the point raised by Mr. Lee, it was significant that a range of fuels tested by three distinct methods (two in the E.6 and one in the “Dagger”) gave substantially the same relative pre-ignition ratings (the one exception was toluene) which showed no relationship whatever to their H.U.C.R. ratings or octane numbers. It might be possible slightly to modify that conclusion in certain cases by an experiment such as he suggests, but it was extremely doubtful in view of the weight of evidence that any relationship could be established. In regard to Mr. Forbes’ question, if pre-ignition was occurring in one cylinder of a multi-cylinder engine the performance could be seriously affected. In the case of a single-cylinder engine such as fitted to a motor-cycle, pre-ignition could cause a sudden stoppage very similar to engine seizure, due to “negative work” on the piston during the compression stroke. The h u e n c e of petrol injekion on pre-ignition had not been studied by the authors. In the case of direct injection into the Running-on. There was considerable support for the idea that running-on was spontaneous ignition of the charge. That it was not directly associated with a hot spot in the combustion chamber was shown in the experiment described by Dr. Tait, when, with engine cold, he heated the induction inanifold hotspot with a blow-lamp and produced running-on immediately on starting. The interesting contributions of Mr. Barrett and Mr. Maclachlan formed a valuable summary of what appeared to be the main factors controlling the phenomenon. The very great influence of mixture temperature indicated clearly the directionin which efforts to overcome the complaint should be turned. That applied to measures both outside and inside the combustion chamber. Desirable features appeared to be a minimum amount of hot-spotting in the induction manifold, with as cool an air intake as possible, combined with good circulation of water around the combustion chambers at low engine speeds, and a general reduction in coolant temperature. Thst latter was, apparently, often controlled at a high figure for the purpose of car body heating. A suitable alternative there could be to utilize the exhaust heat. That would reduce under-bonnet temperature, and would also obviate the need for unnecessarily high coolant temperatures. Mr. Barrett, Dr. Tait, and Mr. Withers had all commented on the adverse effect of overlap valve timing, and Mr. Maclachlan had observed that excessive back-pressure in the exhaust system could be a contributory cause of running-on. Mr. Coy’s reference to the fact that four-cylinder engines seemed more prone to running-on than are six-cylinder engines, must be set against the observation of Mr. Campbell who had found the reverse to be the case, albeit in an engine presumably rather larger than those Mr. Coy had in mind. Mr. Stead’s remarks suggested that in the engine referred to, badly seating inlet valves could, by causing blowing back into the inlet manifold, result in a critical rise in temperature of the ingoing mixture. Although it was shown that there was some relationship between the octane number of the fuel and the running-on tendency, it would seem from all the evidence, that the latter was largely an engine problem. The influence of compression ratio should not be overlooked, since current ratios were now nearly 1 unit greater than in 1939, while present-day fuels were unavoidably rather lower in octane number than those obtainable at that date. Mr. Perrett’s experience that two-cycle engines are immune from running-on was of interest, particularly as in the case of the three-port type an appreciable degree of heating of the mixture must occur as a result of its passage through the crankcase. As Mr. Moore had pointed out, the compression ratios in those engines were usually low and that would reduce the chance of a spontaneous ignition. TABLE 14 Cylinder diamerer,/ inches 1 2.75 i I H.U,C.R. 7.9 i 7.5 5.5 I 6.2 8.5 I 5.4 3.5 I Surface Discharge Plugs. The authors were grateful for Dr. Watson’s clarification of their redarks on the surface - discharge m- - e of d U g . In reply to Mr-. Gadley and Mr. Wightman, so far as the Downloaded from pad.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016 AUTHORS’ R E P L Y O N P R E - I G N I T I O N I N T H E S P A R K - I G N I T I O N E N G I N E 149 authors are aware the above type of igniter was proposed for give trouble due to normal fouling. For that reason a certain high-altitude aircraft piston engines because the conventional amount of interest in surface discharge plugs had been created high-tension systems gave nouble due to flash-over at the low where engines were installed in inaccessible positions and normal atmospheric pressures obtaining at altitudes around 40,000 feet. spark plug maintenance could not be carried out. Some general information on that form of low-tension ignition One way of avoiding that was to pressurize the magneto, and another was to introduce a low-tension ignition system. The may be obtained in an article by W. Beye Smits (1951)*. latter involved the use of a condenser discharge of low voltage, Effect of Engine Size on H.U.C.R. The compression ratio at across a flush-surfaced spark plug. The discharge path was of low resistance and the deposition of fuel and oil deposits on the which knock occurred would increase as the cylinder size was surface was advantageous in maintaining satisfactory working reduced, other things being equal. Figures showing the infiuence of cylinder diameter on the H.U.C.R. of a given fuel are shown conditions for the low-tension discharge. An advantage of the latter feature was that the plugs would not in Table 14. * SWTS, W.BEYE1951 De Ingenieur,vol. 63, No. 3. Downloaded from pad.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016