US2457102A - Spark gap - Google Patents

Description

F. L. JONES Dec. 21,1948.'

SPARK GAP Filed latch 27, 1945 I y: ventor Attorney Patented Dec. 21, 1948 SPARK GAP Frank Llewellyn Jones, Millbank, London, England, assigner to Minister of Supply in His Majestys Government of the United Kingdom of Great Britain and Northern Ireland, London,

England Application March 27, 1945, serial No. 585,165 In Great Britain February 17, 1941 Section 1, Public Law 690, August 8, 1946 Patent expires November 4,1962

2 Claims.

This invention relates to spark gaps of the kind used in high frequency ignition systems of internal combustion engines and in certain radio circuits using rapid spark discharges.

An object of the invention is to provide a spark gap capable of operating at a high rate of sparking for many prolonged periods without substantial change in its electrical characteristics despite changes in external conditions such as barometric pressure, temperature, humidity or incident illumination. It is also an object of the invention that the gap should attain these characteristics within a minimum time of initial forming Another object of the invention is to provide a spark gap which is almost a perfect insulator when the applied E. M. F. is less than a critical value but which rapidly breaks down and becomes a good conductor when the E. M. F. rises above the critical value, the breakdown potential of successive sparks lying within close limits, i. e. having an impulse ratio of nearby unity. Also, the gap must encourage the passage of a spark, as distinct from an arc which might persist for an interval of time sufficient to prevent the gap from reverting to a fully non-conducting condition preparatory to the application of the next pulse of E. M. F.

A specific object of the invention is to provide a spark gap suitable for use in a high frequency ignition system where a condenser, connected inseries with the spark gap and with the primary coil of a high frequency transformer, is charged from an impulsive source of E. M. F., such as a magneto or ignition coil, to a potential diiference in excess of 1000 volts and preferably less than 5000 volts, and then discharged across the gap. For such a system to operate successfully the breakdown of the gap should occur within a very short interval of time, of the order of -5 sec., after the applied E. M. F. has attained the critical value below which the gap is an insulator, and there should be a maximum difference of about 20% between the breakdown potentials of successive sparks.

A further specific object of the invention is to provide a spark gap which is particularly suitable for use in the high frequency ignition circuit of an aircraft in which case reliability becomes paramount and small dmensions are an advantage.

To attain these objects a spark gap constructed according to this invention is enclosed Within a gas-filled and gas-tight envelope in which the electrodes are sealed and rigidly supported. The

2 electrodes are shaped and disposed so that a practically uniform electric eld is maintained throughout a substantial fraction of the spark gap volume, and are made from materials with thermionic work functions greater than 4.0 electron volts and with low erosion properties under the action of the spark discharge. Also the prodf uct of the pressure of the gas filling measured at 15 C. and the distance apart of the electrodes is greater than 5 times the value of the corresponding product at the minimum sparking potential of the gas. When the electrodes consist of two parallel opposed discs the ratio of the diameter of the discs to their distance apart is made not less than 10 and is preferably greater and the distance apart of the electrodes is made not less than approximately 0.2 mm. A hard glass which remains rigid at temperatures up to about 450 C., or quartz, are suitable materials for making the envelope. The gas or gases chosen to ll the envelope are such that any chemical changes produced in the gas itself, or any chemical reaction between the gas and the materials inside the envelope are negligible during the actual operation ofthe device. Suitable gases for lling the envelope are, for example, the monatomic gases, or hydrogen, or nitrogen, and nitrogen is especially suitable'when the improved device is used on an aircraft, while hydrogen is especially suitable when a very highrate of sparking is required. The gas density is determinedby the purpose for which the enclosed spark gap is required, but it should not be less than that corresponding to 20 cm. of mercury at 15 C. and for very short spark breakdown times higher gas densities are necessary. The most suitable gas density when the enclosed spark gap is used in a high frequency ignition system is that corresponding approximately to atmospheric pressure, and a pressure about 5% or 10% less than atmospheric is especially suitable, as then in the process of manufacture the process of sealing of the envelope is thereby simplified, and further, high internal pressures, which tend to break the envelope, are less likely to occur during the operation of the device. In the case of hydrogen or nitrogen the value ofthe aforementioned product which corresponds to the minimum sparking potential is about 1, when the gas pressureis measured in cm. of mercury at 15 C. and the distance apart of the electrodes is measured in mm., so that when hydrogen or nitrogen is used in the improved spark gap the value of the product should not be less than about 5.

The seals through which the connections to 3 the electrodes are brought out through the envelope, which may also be the means by which the electrodes are rigidly supported; should be mechanically strong and capable of working at temperatures up to about 400 C. A suitableform of seal is a tungsten-hard glass seal, for example,

Two practical embodiments of the invention are illustrated in the accompanying drawingin which- Figure l is a longitudinal section of a4 form of the invention in which the electrodes consist of parallel opposed discs and Figure 2 is asimilar view of a modified form of the invention inwhich the electrodes consist of coaxial cylinders. Workable overall dimensions ofV the. device in each case may be about 31/2 in. long by 1 in. diameter. v

In Figure 1 the electrodes I and 2 consist of discs with rounded edges made of a metal having a thermionic work function greater than 410 electron volts, such as tungsten. The diameter of fthe hat parts' of the discs is indicated by the letter D and the overall diameter by DI. The distance between theelectrodes is indicated by the small letter d.'

'Ihefdiscs` are-welded to tungsten rods 3 and 4 respectively, the cross sectional area of which determines the rate of `cooling of the electrodes, andtherodsextend through a hard glass envelope 5. The rods are sealed directly into the envelope the-walls of which are suitably thickened at B and Iso as to increase the lengths or the seal-and give enhanced mechanical support.

After filling withI gas the envelope is perma-` nently sealed atL 8:

In the modification illustrated in Fig. 2 the electrodes consist of-'coaxial cylinders 9 and I0 made of a-metal having athermionic work function greaterrthan 4.0 electron volts and' spaced-y apartby a distance d. The cylinders are supported on metal rods 3 and 4 respectively, which may be similar to the rods 3 and diin Figure 1, and are enclosedin a hard glass envelope'into whichthe rods are sealed `at the thickened parts` G" and 1.

The exactshape of the electrodes-is not important provided'that they are shaped and so disposed that a practically `uniform electric field is maintained over ay substantial portion of the spark gap volumel When, however, the enclosed sparkgap is to bel operatedfrom a source of alternating polarity, such as a-magneto fork instance, then it is essential that the electric'i'leld should be the same at each electrode surface and therefore the electrode arrangement shown in Fig. 1 is to be preferred in such circumstances. If the electric field werenot uniform but were concentrated more strongly at one electrode than at the Other, then the sparking potential would not be the same wheneither electrode is used'as the cathode,` so that when operated from a magneto the breakdown potentials for successive sparks would in general be different, and this would be a disadvantage whenthe-enclosed spark` gap is` used in a-high frequency ignition circuit. Again, under these conditions itis preferable that both electrodes should be made-of the same material so that they have the same thermionic work functions. The reason for this is that'durng `the operation of the spark gap at a high rate of sparking, say in excess-of about'lOO sparks per second, a largerise of temperatureofthe electrodes can occur unless theheat'is'conducted away. Under these conditions considerable th'ermionic emission occurs from the opposed surface of the electrodes. This thermionic emission is of great importance in the operation of the spark gap. Firstly, it can create the supply of electrons in the gap which is necessary for its rapid breakdown when the applied potential exceeds a critical value which is approximately equal to the sparking potential of the gap under a steady applied potential. If the breakdown potential of the spark gap is to be the same when either electrode is used as the cathode, thenthe thermionic emission and the rate of conduction oi heat away must be the same at each electrode. This necessitates a symmetrical arrangement of the electrodes. Secondly, if the supply of electrons is too great and the gap is operating at a very high rate of sparking, then in certain gases the electrical discharge tends to become a practically continuous arc instead of a succession oi distinct sparks. This is a great disadvantage as it renders a high frequency ignition system inoperative, and it also generates a considerable quantity of heat, which may damage the seals and also produce excessive erosion of the electrodes. Excessive therrnionic emission into the gap is reduced by employing electrodes of high thermionic Work function not less than 4.0 electron volts, and by ensuring sufficient conduction of heat away from the electrodes. A degree oi thermionic emission is, however, in some cases an advantage in order to reduce the statistical time lag of sparking and it may be necessary to adjust the cooling of the electrodes to ensure that the total time lag of sparking is less than 10'4 second, and in general the limit of 10-5 second is preferred. Excessive erosion is reduced by employing metals of high boiling point for the electrodes. These Conditions are satisfied by using, for example, metals like tungsten, molybdenum or platinum for the electrode material, and by using, for example metals like 'tungsten for the supports for the electrodes. Tungsten is especially suitable for the electrodes and the supports, since this metal is a good conductor of heat and electricity, and it has a high boiling point probably in excess of 4000" C. Also (Handbook or" Chemistry and Physics, Hodgman, Ohio. 1940, p. 1737), and Becker (Review of Modern Physics 7, 123J 1935) gives its work function as 4.52 electron Volts. Further, since its coefficient `of expansion is approximately equal to that of ywith advantage be made since a hard glass is mechanically strong and also can withstand temperatures up to about 1:00o C. Another advantage of the use of tungsten for the electrode supports is that the expansion of the supports is then approximately the same as that of the glass envelope, so that the gap distance is not greatly altered by any change in operating temperature of the enclosed spark gap. The electrical connections from outside the envelope to the electrodes inside must be of low resistance, since large currents of the order of amperes may be passed lwhen the spark gap breaks down. Also, the connections to the electrodes must be ci adequate size in order to ensure suilcient cooling of the electrodes.

A further advantage of the use of materials of high work functions inside the envelope is that the enclosed spark gap has no photo-electric properties due to ordinary daylight, so that its characteristics are the same Whether it is operated in the dark or exposed to ordinary daylight.

When the enclosed `spark gap is used in a circuit in which the polarity of .each electrode does not change, a perfectly symmetrical disposition of the electrodes is not then essential, and they may consist of opposed surfaces of different curvature. For instance, as shown in Figure 2, the electrodes may then consist of coaxial cylinders when the radii of curvature of the opposed surfaces are great compared with their distance d apart. In all cases, however, sharp points or edges should be avoided, and the electrodes and their connections must be large enough to ensure adequate cooling of the electrodes and a lowresistance electrical connection to them. l

A practically uniform electric eld in the spark gap is ensured by the electrode arrangements shown in figure 1, i. e. by using as electrodes two parallel opposed discs of large diameter compared .with their distance, apart. The effect of the value ofthe ratio diameter/distance apart on the uniformity of the electri-c field in the gap has been studied by McCallum and Klatzow (Philosophical Magazine 1'7, 291, 1934) who showed that the sparking potential of a gas between parallel discs is only accurately a function of the product of the gas pressure, measured at 15 C., and the distance apart when the ratio exceeds a certain value dependent on the nature of the gas. In neon, for example, the ratio must exceed 13, while inargon it must exceed 40, but in helium it is much longer. Hence, when the electrodes of the improved spark gap `are in the forrn of opposed, parallel discs as shown in Figure 1, the ratio of the diameter D of the flat parallel surfaces to their distance apart d will depend on the gas used to lill the envelope, the ratio being higher when the noble gases neon or argon, for example, are used, than when the diatomic gases such as hydrogen or nitrogen are used. In all cases the ratio should exceed a value of about and preferably be much greater.

The edges of the electrodes should be rounded off with a gentle curvature to avoid sharp points or edges, which should also be avoided on the supports or connections of the electrodes inside the tube and when the edges of the discs are rounded off, the ratio of the overall diameter DI of the electrodes to the distance apart d should then exceed about 15.

The flat parts of the electrode surfaces should be spaced from the envelope at a distance not less than 5 times the distance apart of the electrodes.

The distance apart of the electrodes should not be less than approximately 0.2 mm. when the enclosed spark gap is to be used in high frequency ignition systems, and the gap may be as great as 1.5 mm. when the enclosed spark gap is to be used for other purposes, such as in radio circuits, when a high sparking rate and a breakdown potential as high as 15,000 volts may be required.

To ensure as rapid a breakdown as possible after the applied potential has exceeded a certain critical value it is advisable to have the gas pressure as high as possible provided this does not require a gap distance smaller than approximately 0.2 mm.

As hereinbefore stated the product of the dis-4 tance apart of the electrodes measured in mm. and the gas pressure measured in cm. of mercury at C. should exceed about 5 times the value of the corresponding product at the minimum sparking potential of the gas. For example, the value of this product at the minimum sparking potential for hydrogen with various cathode materials has been found by Llewellyn-Jones and Henderson (Philosophical Magazine 28, DI).v 185, L.

192 and 328, 1939) to be about 1.2 and the value of the product corresponding to the minimum sparking potential of nitrogen has been found by Strutt (Philosophical Transactions of the Royal Society A, 193, 377, 1900) to be approximately equal to 0.8, so that when hydrogen is used as the gas filling the value of the product should exceed about 6, and when nitrogen is used the value of the produ-ct should exceed about 4. Consequently when a gap distance of 0.2 mm. is used, vthe pressure of nitrogen at 15J C. in the envelope should exceed 20 cms. of mercury.

The relationships between the gas pressure, p, measured in cm. of mercury at 15 C., and the gap distance d, measured in mm. and the work function w measured in electron Volts may be expressed by the following formulae:

(p) X (d) X (w) C parallel flat discs with a diameter of the flat part of the surface of D cm. then (p) X (d) X (w) v (D/d)=(p) X (w) X (D) (K) where K depends on the gas.

For example, with hydrogen,

(p) (w) (D) 240 and for nitrogen,

(il) (w) (D) During the initial operation of an enclosed spark gap constructed according to the invention a fine deposit is sometimes, but not always, automatically produced onor near the electrodes and walls, probably due to the evolution of absorbed gases, and their reaction with the material of the electrodes when these have not been de-gassed. As the process continues for some time, which may be long, the electrical characteristics of the enclosed spark gap improve, and when the process is completed the characteristics remain steady throughout the life of the spark gap. When the electrodes are made from tungsten this deposit contains oxides of tungsten produced by the sparks in the presence of traces of oxygen and water vapour which are generally present in the enclosed spark gap when it has not been degassed. These traces may be found as impurities, for inst-ance, in the preferred gas filler, or they may have been absorbed by the electrodes and envelope and evolved during the operation of the gap, or they maybe liberated from the electrodes which may have become oxidised during the sealing-in process in the construction of the tube, or they may be liberated by the brazing or joint fixing the electrodes to the supports. These final traces of water vapour and of oxygen and other occluded gases are not easy to remove, and any such removing process would require considerable time. However, in some casesy it is of advantage not to attempt to remove them, but to utilise them in the operation of the gap. In the first place, the presence of water vapour readily assists the formation of oxides, which in some cases may distil off from the electrodes due to the passage of sparks. Secondly, the presence of water vapourassists arc suppression owing to the electron amnity of water molecules. Naturally, the deposits are quickly produced if oxygen, or air are used in the gas filler and traces of water vapour are present.

The deposits on the electrodes can have various eiects on the operation of the tube. For instance, they can raise the thermionic work function of the surface and therefore have the advantage of delaying the transition to a thermionic arc immediately after breakdown with consequent reduction in the rate of liberation of heat from the gap. They also cause successive sparks to pass to dilerent points on the electrodes and so avoid excessive heating of one particular part. In fact an enclosed spark has been operated at 400 sparks per second, and the resulting rise of temperature was found to be about 30 C. to 40 C. Again, the deposits have the Well-known effect of tending to produce constancy in the break-down potential of successive sparks and reducing the statistical time lag, (See for instance H. Paetov, Zeitschrift fr Physik, volume III, p. 770, 1939, also M. J. Druyvesteyn and F. M. Penning, Review of Modern Physics, volume 12, No. 2, p. 117, 1940.)

Further, as the oxides become deposited on the nearby inner walls of the envelope the deposits there can also exert influence on the operation of the spark gap. In the first place there is the advantage that ionising radiations may be produced which tend to reduce the statistical time lag of sparking. (See H. Raether, Zeitschrift fr Physik, volume 110, p. 611, 1938.) However, in the second place, if the deposits are allowed to be too near the region of uniform eld in the gap the statistical lag can be increased, consequently the impulse ratio is raised. To reduce this effect to a minimum the distance of the walls of the envelope from the region of uniform field should, as stated hereinbefore, be five times the length of the discharge gap. Deposits on the walls of the envelope near the gap may help in the dissipation of charges from that region, collected there by diffusion from the ionising path o1' the electrical discharge. Recombination of ions and electrons at the walls can be regulated by the proximity of the walls to the discharge path. (See F. Llewellyn Jones, Philosophical Magazine, volume 15, p. 958, 1933) and this diffusion and recombination can play an important `part in the rapid de-ionising of the gap after each spark has passed. An enclosed spark gap in accordance with the invention using nitrogen and tungsten electrodes has been constructed for which the variation of breakdown potential for successive spark was about 1% to 2%. Once the suitable deposits are formed, their influence remains unaltered during the life of the enclosed spark gap with a suitable gas ller. With nitrogen of commercial purity, for instance, the effect of any chemical action, if any, between the gas and the contents of the envelope is negligible, and no such effect attributable to that cause was observed during the operation of an improved tungsten electrode spark gap for more than 300 hours.

The evolution of the usual absorbed gases by the electrodes and envelope, or the presence of traces of oxygen and water vapour in the gas ller renders the production of the deposits of oxides automatic during the initial stages of operation of the enclosed spark gap, but the process may take considerable time, and this would be a disadvantage in large scale production. This time can be considerably reduced, or even eliminated, by first depositing a fine layer of the appropriate oxides over the electrodes during the assembly of the enclosed spark gap, and before nally sealing off the envelope. During the passage of the first few sparks across the gap with tungsten electrodes, for instance, some of these deposits are vaporised and condensed on the en velope.

The effect on the sparking potential of the presence of mercury deposited on the cathode surface has been investigated by Llewellyn Jones and Galloway (Proceeding of the Physical Society, 50, 207, 1938),and the effect of the presence of sodium deposited on the cathode has been investigated by Ehrenkranz (Physical Review 55, 219, 1939), and in both cases the sparking potential with a cold cathode was considerably reduced due to the lowering of the work function of the electrode. Even when the electrodes are hot, as would be the case when the enclosed spark gap is operating at a high rate of sparking, the presence of the metallic vapours in high concentration tends to introduce variations in the thermionic work functions of the electrodes and also in the electrical properties of the gas, such as the ionisation coefficients for example, and consequently in the value of the breakdown potential.

A further disadvantage of the use of any metallic vapours in the enclosed spark gap is due to the variation of density of the metallic vapours caused by the operation of the enclosed spark at widely diierent temperatures. For instance, if the enclosed spark gap were fitted to an aircraft the temperature of the atmosphere surrounding the envelope would undergo large variations due to the operation of the aircraft at different altitudes and under different atmospheric conditions.

The nature and pressure of the gas filling is determined by the purpose for which the enclosed spark gap is to be used. When the en closed spark gap is toibe used in a high irequency ignition system and the breakdown potential of the spark gap lies within the approxi mate limits 1000 volts and 500() volts, the gas used must be such that the gap distance under these conditions must exceed 0.2 mm. when the gas pressure measured at 15o C. is greater than 20 cm. of mercury and preferably less than about 76 cm. of mercury. Also, the gas must allow separate sparks to pass when the rate of sparking does not exceed 400 to 800 sparks per second depending on the type of engine. Consequently, it follows that the deionsation of the gap must be so complete after the initial applied pulse of electromotive force which produces a spark has passed, that the discharge does not persist until the next spark produced by the next pulse of electromotive force from the magneto, or ignition coil or other means which produces the potential necessary to break down the spark gap. If the gas atoms or molecules have metastable states, and if the gas contains traces of impurities and if the potential energy of the metastable levels is lgreater `than the impurity, the ionisation may persist in the gap after the applied electromotive force has been reduced to zero, due to the process of ionisation by collisions of the second kind. A method of reducing the persistence of ionisation in gases with high metastable energies is the well known one of destroying the metastable atoms, where possible, by lprocesses which do not lead to ionisation, and such processes can occur, for example, when the potential energyof the metastable atom is less than the ionisation potential of an impurity molecule with which it collides. Further, `the rate of de-ionisation ofthe gap after a spark has passed depends on the coeflicients of recombination and of diiusion of the electrons and the positive and negative ions in the gas used. These coeilicients, generally, are higher in a` light gas than in a heavier gas, consequently shorter deionisation times are to be expected in hydrogen than in nitrogen, so that when very high rates of sparking are required it would be of advantage to use hydrogen as the gas lling. However, when the enclosed spark gap is used in a high frequency ignition system on an aircraft, the use of nitrogen as a filling'. is preferable. This is due to the fact that the sparking potential of nitrogen is not very different -fromthat of air under similar conditions. Consequently, if nitrogen at a pressure of about '70 cm. of mercury were used in the enclosed spark gap, the breakdownv potential of the gap would not rbe greatly altered at low altitudes if by any chance the envelope. were damaged so as to admit air which would 'contaminate the nitrogen and also would equalise the internal gas pressure with that of the surrounding atmosphere, provided the position of the electrodes was unaltered. Such a form of damage would be a crack in the envelope or a failure of a metal-glass seal, thoughsuch damage is not likely to occur during the normal operation of the tube. Consequentlmthe sparking plugs in the engine would continue to ire in spite of the damaged envelope. On the other hand, if the damage to the envelope occurred at yhigh altitudes when the atmospheric pressure was greatly reduced, say to about 0.1 of the pressure at sea level, the breakdown potential of the enclosed spark gap would tend to be so low that insufficient electromotive force would be generated at the sparkingy plugs and inisring consequently occur. However, when the aircraft had descended to the lower altitudes, regular rlng would again be produced in the engine when the former breakdown potential of the enclosed spark gap was restored by the higher atmospheric pressure at the lower altitude near the ground. In this way the chance of an aircraft crashing due to an engine mis-fire `produced by such damage to the envelope' of the i enclosed spark gap could be minimised.

Hence, when tted to an aircraft, the gas filling for the improved enclosed spark gap should, as a safety measure, preferably be nitrogen at a pressure within the limits of 50 cm. to 100 em. of mercury at 15 C., but a pressure of about '70 cm. of mercury is preferred. However, under conditions when the life of the enclosed spark gap under operating conditions need not be so long as it would be if nitrogen were used as the gas lling, then air could be used at pressures a little less than normal atmospheric. This would be the case when the rate of sparking is very low, such as when the enclosed spark gap is used as a lightning arrestor.

If the enclosed spark gap is used in a high frequency ignition system on an aircraft, for instance, it is advisable that the risk of a misre in the spark plugs of the engine should be reduced as much as possible. Such a misflre would occur, for instance, if the breakdown potential of the enclosed spark gap were momentarily raised due to some cause so as to exceed the E. M. F. applied to the gap. Again, if the breakdown potential of the enclosed spark gap varied from spark to spark in l0 akf'sccessionf'ofbreakdown within limits, for example, of about 10%, due to some cause, and if the applied E. M. F. fell within those limits there would then be risk of a misre. This risk can be reduced by employing two or more improved spark gaps in parallel when the spark gaps have electrical characteristics, including breakdown potential, impulse ratio,'and gap voltage-time curves, for instance which Vare as far as possible the same for all the gaps. This `is due to the fact that when the breakdown' potentials for successive sparks vary within limits of say, about 10%, measurements of those breakdown potentials showthat themost probable value of the breakdown potential is rather less than the arithmetic mean of the maximum and minimum values of the breakdown potential.v In other words, if, for example, two practically identicalenclosed spark gaps, in accordance with the invention, each with an impulse ratio of say 1.1, were 'used' in parallel the eiective impulse ratio ofthepairwould be about 1.05, because the probability of the occurrence of an impulse ratio of 1.1 maybe neglected. Similarly, if the impulse ratio of each gap were about 1.05, then the impulse ratio of the pair used in parallel would be reduced to about`1.025 for the same reason. If, one gap failed to breakdown the other gap would be unaffected, andk would vstill allow the system dependent on the gap to operate. f The success of such a combination of spark gaps is due to the fact that each gap has a (breakdown potentialrate of sparking) characteristic which rises slightly as the rate `of sparking increases from aboutl 40 sparksf 'per 'secondtoabout 300 sparksper second, and is also dueto the factthat an improved enclosed spark gap can be constructed to have impulse ratios as low as about 1.02 and alsofas-hlgh as about'1.2; When -using gaps in parallel it is not advisable to use gapst which have impulse ratios very near unity, andigood results have been obtained when using .two gaps each'with an impulse ratio as low as about 1.05.

The gaps used in any such combination replace the' use lof a' .single enclosed spark gap and the combination may consist of two or more complete and separate gaps, oralternatively they Ymay consist of two or more pairs of electrodes in one envelopebutv the use of distinct andseparate enclosed gaps is preferred, since then one: gap can be destroyed vor removed without upsetting the operation ofthe,systemyprovided that the electrodes of the gap holden. are not short circuited. Another advantage of the use of such a combination is that a pair of gaps, for instance, can operate at a rate of sparking twice as great as that at which a single gap can operate satisfactorily. This is due to the fact that if one gap of such a pair commences to operate at a very great number of sparks per second then its breakdown potential tends to rise slightly. This will -then increase the probability of the other gap in the pair breaking down which consequently will reduce the sparking rate of the first gap. In that way there is a tendency to equalise the sparking rate for the two gaps in a pair.

A further advantage of the use of a combination of gaps in the manner indicated above is that, for a given rate of sparking required by the system utilising the gaps, the life of the combination of enclosed spark gaps is greater than that of a single gap. Again, another advantage of the use of a combination of enclosed spark gaps is that overvoltages considerably in exces of the breakdown potential of the gaps may be applied to the combination with safety under circumstances 1.1 which would `tend to zproduce considerable Aheat and consequentdamageto a gap. acting alone.

If an enclosed spark gap constructedaszhereinbefore described had lanenvelope `with overall dmensions of about -3" x t" x $5", lthen'the use of a combination of four such gaps wouldoccupy a rectangularvolumenotgreatly exceeding about 3" `x 1 x 1". Such a combination would have a fourfold increase in safety factor, an effective breakdown voltage -variation of about one quarter f that of the single gap, and also .the ability to handle a spark .frequency of the ordervof `1600 sparks per second. Alternatively, an extremely long `life is to be expected :from thecombination at lower spark frequencies.

The use of sparkgaps in parallel has another advantage in that it introduces the possibility of obtaining an effective spark gap which,-when used in `an ignition system on laircraftfor example, is protected against a failure of any of=^the separate gaps to breakdown, and also against a ycrack in the envelope of any gap of the combination, provided `one enclosed sparkfgap ,is still intact. A cracked `envelope can `equalise the pressure .inside the gap with that of the atmosphere. It :is only necessaryto use anenclosed sparkf gap in which the nature of the gas filler, anci'its density, and the gap distance are :so arranged `zthat, `when filled `Withatmospheric air at thedensity appropriate to the highest operational `altitude considered, thebreakdown potential is :either equal to or higher than,zthat of theother gaps in the combination, which `gaps then control'the sparking at all altitudes. Consequently, misiiring due either to a'cracked envelope or :failuretobreakdown in any1of the constituent gaps .isfthereby avoided.

When nttedto an aircraft Athe overall .dimensions of the enclosed spark gapmust be such that the minimum distance `between the `exposed metal connections *at theiseals, orany'other conductors connected to them, must ybe Asuch .that spark over between them does 'notioccur outside the envelope whenthe enclosed spark gapis used at the highest operationalraltitude, `or at a reduced atmospheric pressure surrounding the envelope and connections.

The improved enclosed spark gap constructed as described above may be designedto operate at breakdown'voltagesln excess 'of 5000 voltsby adjusting `thegap distance and/or the `density of the gas 1111er to have theappropriate value. In

.order to produce the higher `breakdown potentials an increase in the'gas density is, vin many cases, preferred, lwhen practicable, to an increase in the gap distance, as this increases the value of thefelectric intensity at the electrode surfaces, and it also helps to maintain the dimensions of the electrodes reasonably small.

4I claim:

1. A spark `gap device comprising electrodes, arranged in parallel spaced relationshipsuch that a substantially uniform electric field is creatable between them, made of materials having thermionic Work `functions greater than 4.0 electron volts, at least `one electrode surface having deposited thereon an `oxide of the material of the electrodeto enable the gapto operate at a high sparking rate, the electrodes being rigidly mounted and spaced apart a distance greater than approximately 0.2 mm. within a sealed envelope filled with nitrogen at a pressure of be tween 50 and 100 cm. mercury at 15 C.

2. A sparkr gap device comprising two electrodes consisting of coaxial, telescopically arranged cylinders made of materials having thermionlc work functions greater than `1.0 electron volts, at least one electrode surface having deposited thereonan oxide of the material `of the electrode to enablevthe gap `to operate at a high sparking rate, `the electrodes being vrigidly mounted `and spaced apart a distance Adwithin a sealed envelope filled with gas at a pressure -p such that theproduct pd is greater than flve times the product p d which corresponds to the minimum sparking potential of the gas to prevent deleterious arc formation.

FRANK LLEWELLYN JONES.

REFERENCES CITED The following references are of record in the le `of this patent:

UNITED ySTATES PATENTS Number Y Name Date 571,109 Culgan Nov. 10, 1896 1,144,029 Creighton June 22, 1915 1,271,794 Stevenson M July 9, 1918 1,929,661 Von Wedel Oct. 10, 1933 1,933,329 Hull Oct. 31, 1933 1,965,584 Foulke July 10, 1934 .2,300,931 Kalischer Nov. 3, 1942 2,354,786 Wall Aug. 1, 1944

Images