US3846657A - Fast rise time quenching spark gap - Google Patents

Abstract

Electrically floating wire mesh or perforate plate electrodes are interposed between the two electrodes of a conventional spark gap, with interelectrode spacings from 0.010 to 0.005 inches. Rise times of the order of 2 nanoseconds with repetition rates up to 200 kHz are produced.

Description

United States Patent 1191 Austin [451 Nov. 5, 1974 FAST RISE TIME QUENCHING SPARK GAP OTHER PUBLICATIONS [75] Inventor: William E. Austin, Wayne, Pa.

Morecroft, J. 1-1., Principles of Radio Communica- [73] Ass1gnee. General Electric Company, New on, John Wiley & Sons Inc 1927 York, NY.

22 F1 d: 1 16 May 17 1973 Primary Examiner-James W. Lawrence PP 361,196 Assistant Examiner-Wm. H. Punter Attorney, Agent, or Firm-Henry W. Kaufmann; Allen 52 11.8. C1. 313/195, 313/299 Raymnd Q1115 [51] Int. Cl. H0lj 1/46 [58] Field of Search 313/195, 297, 299, 300,

313/306, 307, D16. 5, 323, 324; 331/127; [57] ABSTRACT Electrically floating wire mesh or perforate plate elec- [56] References Cited trodes are interposed between the two electrodes of a UNITED STATES PATENTS conventional spark gap, with interelectrode spacings from 0.010 to 0.005 inches. Rise times of the order of 12112133 1311333 313L151:1:111::33111111113111133123: iii/i5? 2 with ttptttttttt rates up to 200 kHz 3,349,283 10/1967 Krefft 313/195 P FOREIGN PATENTS O11 APPLICATIONS 2 Claims 2 Drawing Figures 578,664 7/1946 Great Bntain 315/36 V a P07 2 57 /A2 N/f J- JOURCE 1 FAST RISE TIME QUENCIIING SPARK GAP This invention was made in the performance of N000l4-73-C-0042 with the Office of Naval Research of the Department of the Navy.

BACKGROUND OF THE INVENTION 1. Field of the Invention This invention pertains to quenched spark gaps.

2. Description of the Prior Art The quenched spark gap was first applied to overcome certain defects of the straight gap in radio telegraphy. In that application it was used to shock excite an oscillating circuit comprising a charged high-voltage capacitor and an inductor which was either the primary of an oscillation transformer or a part of a singlewinding helix. In either case, the antenna was inductively coupled to that inductor either by the separate secondary of the oscillation transformer or by tapped connections to a de facto secondary formed by part of the helix; and the oscillatory energy in the primary circuit was transferred to the secondary circuit which included the antenna. The energy was desired in the antenna circuit, whence it could be usefully radiated; but if the primary circuit remained closed by the spark gap, energy would be transferrred uselessly back from the secondary circuit to the primary, in accordance with the well-known behavior of coupled systems. Thus it was desirable to provide a spark gap which would open when the current through it first reached the low value which occurred when the energy had been largely transferred to the secondary circuit, and so prevent the reverse flow of energy. It was a rapid deionization that was wanted. The rise time was of less importance because quenched gaps were ordinarily not used for oscillation frequencies above [.5 mHz, and a rise time of 0.l microseconds (which prior art gaps could achieve) was .a negligible fraction of the length of a half cycle. The

usual form of prior art quenched gap comprised a stack of copper disks, silver-plated in their central portions, separated by insulating washers, and held firmly together by a screw clamp. The disks usually protruded radially appreciably beyond the washers to provide flanges for natural or forced air cooling.

More detailed descriptions of the preceding may be found in the technical publications of 1910 to 1920; the first or second editions of Principles of Wireless Telegraphy by .I. H. Morecroft, published by John Wiley and Sons, discuss the behavior of coupled circuits with quenched gaps in some detail.

The relatively recent appearance of the laser has created a need for a pulse source having a very fast rise time (since laser efficiency is markedly a function of it) and high repetition rates in order that the average power may be high. For obvious reasons, simple and inexpensive structures are desirablefor economy and, insofar as simplicity implies absence of powerconsuming auxiliaries, for efficiency. Semiconductor, electron tube, and mechanical devices all are lacking in some respect.

SUMMARY OF THE INVENTION:

The present invention may be described as a fixed spark gap in which a plurality of wire screen or foraminate plate electrodes. each insulated and floating electrically, have been inserted between the main end electrodes. The effect of this is to produce rise times comparable to a conventional two electrode spark gap, yet provide quenching as in prior quenching spark gaps so that high repetition rate pulsing can occur. Conventional spark gaps have a limited pulse rate, up to 2 kHz at most, and prior quenching spark gaps have had a slow rise time of the order of 0.1 microseconds. This invention produces rise times of 2 nanoseconds, more or less depending on the peak electrical current in the pulse, yet quenches fast enough to obtain a 200 kHz repetition rate in a relaxation type discharge circuit. This invention, therefore, has the best characteristics of both the conventional spark gaps and prior quenching spark gaps, when fast rise times and high repetition rates are desired. Pulse repetition rates far exceeding 200 kHz can be achieved in circuits not employing RC charging times such as in relaxation discharge circuits.

This should not be confused with the 1.5 mHz oscillatory frequency described with respect to the prior art; the maximum conventional pulse repetition rate of the prior art was in the audio range at 1,000 Hz., by the two halves of each cycle of the output of a 500 Hz. alternator.

BRIEF DESCRIPTION OF THE DRAWINGS:

FIG. 1 represents the preferred embodiment of my invention.

FIG. 2 represents an alternate form of a part of the preferred embodiment.

DESCRIPTION OF THE PREFERRED EMBODIMENT:

FIG. 1 represents, in central section, the preferred embodiment, which is circularly cylindrical in form. Fixed electrodes 1 and 2 are hermetically sealed, preferably by brazing or welding, to sleeves 3 and 4, respectively which are sealed to a ceramic insulating sleeve 5 which forms the housing or envelope of the gap, and also serves to maintain electrodes 1 and 2 fixed in position relative to each other. Electrode l is provided with a drilled path 6 to provide for the introduction of gas, at any desired pressure, into the interior of the housing.

In the gap 7 between the opposed electrodes 1 and 2 there are represented two foraminate electrodes 8 and 9, having a plurality of holes 10 and 11, respectively. For convenience, and to differentiate them readily from electrodes 1 and 2, foraminate electrodes 8 and 9 will be designated as baffles. Insulating washers 12, I3, and 14 serve as spacers to maintain baffles 8 and 9 in desired spacing from each other and from the adjacent electrodes 1 and 2, respectively. In assembly of the device, electrode 2 may be sealed into sleeve 4. Then items 14, 9, 13, 8, 12, and I may be slipped into position in ceramic insulating sleeve 5, and, with electrode I pressed firmly against the stack of other parts, electrode 1 may be sealed to sleeve 3. Thus all the parts indicated are fixed in position; but their fit into ceramic insulating sleeve 5, while adequately close for mechanical stability, is not gas tight, so that introduction of gas through path 6 will fill the entire device.

For complete representation of a use of the device, there are represented a high potential source connected via a resistor 16 to a capacitor 17 and one terminal of a load 18. The other terminal of load 18 is connected to a terminal 19 of electrode 1. A similar terminal 20 of electrode 2 is connected to ground, as is one terminal of capacitor it? and one terminal of high potential source 15. This circuit will be recognized as a conventional relaxation oscillator in which capacitor 17 is charged via resistor 16 until its potential, transmitted via load 18 to terminal 19, causes the gap to discharge capacitor 17 through load 18 until the gap deionizes, whereupon the charging of capacitor 17 begins again and the cycle is repeated.

While electrodes 1 and 2 may be of tungsten (or in general of any metal) for better thermal conductivity I prefer copper, to assist in cooling the device. Ceramic insulating sleeve may be of alumina primarily because such sleeves with sleeves 3 and 4 already sealed to them are commercially available. insulating washers l2, l3, and 14 are preferably of boron nitride both because it is readily machined and because it has good thermal conductivity, which facilitates cooling of baffles 8 and 9. Baffles 8 and 9 may be of copper sheet, 0.008 or 0.009 inches thick; the holes 10 and 11 in them may be 0.0l0 to 0.020 inches in diameter. However, baffles 8 and 9 may also be of wire mesh, as represented in FIG. 2, in section, as 8' and 9 respectively.

For low average operating power level, wire mesh is suitable. However, for higher average operating power levels in which cooling of the baffles may present a problem, it is preferable to make the baffles of metal sheet in which a plurality of holes are present, because the effective cross section of metal available for radial heat flow to the edge of the baffle may be greater for a given baffle thickness than is possible with wire mesh in which the thermal contact between overlapping wires is small so that the effective cross section for heat flow is reduced. The baffles 8 and 9 should be made as thin as is consistent with mechanical strength and thermal flow, since the field produced by potential difference between the electrodes 1 and 2 will be reduced within the holes, so that they approximate drift spaces, and the acceleration of charges moving through them will be correspondingly reduced, tending to increase the rise time of the discharge when the gap discharges. It is also desirable to have the holes in successive baffles in alignment (as shown in H6. 1) so that charged particles. whether ions or electrons, may move through a succession of baffles without interruption or increased path length. While this is not readily achievable with wire mesh, plate baffles may simply be drilled in a stack, and mounted in that orientation. This precaution is a refinement whose effect will, of course, be limited by collisons, since the gap of my invention may be pressurized in order to increase its breakdown voltage.

These results were obtained both with 0.004 inch diameter copper wire in a screen of 100 meshes per inch.

and with 0.00l inch diameter nickcl wire in a screen of 400 meshes per inch. The maximum pulse rate was not determined with the closest spacing of 0.005 inch because the heat dissipation caused melting of the baffle screening before an exact figure was determined. It is evident, however, that rise time plus quenching time must have been less than 5 microseconds.

It is hypothesized that the effect of the electrically 3 floating screens in reducing the deionizing time of a gap iwell below that of a gap without the screens results afrom the following mechanism. After the discharge lsubsides only a small potential exists across any spark j gap but the gas is rich in ions and electrons at a highly elevated temperature. A charge moving toward a floating electrode will induce a voltage of like polarity in the screen or floating electrode. This will tend to repel the charge or slow it down. Conversely, a charge moving 5 away from the screen will induce a voltage of opposite l polarity in the screen, hence will attract the charge or slow it down. The voltage that is induced in the screen may be expressed as V (q/C) (dx/dl) volts where q is the magnitude of the charge C is the interelectrode capacitance, and

dx/dl is the path length of the charge relative to the interelectrode spacing 1 ;Note that if the screens are not floating electrically, a

current will flow and the repelling or attracting forces will not be established.

In any event, the charges are less energetic and the '.cross-section for recombination becomes much larger Ethan for the high energy case, and therefore deionization occurs rapidly. During the initiation and maintenance of the discharge, sufficient energy is obtained 5 from the external power supply to overwhelm the small quenching effect.

Tests were conducted on a gap with fixed electrode geometry but a variable gas pressure. It was found that the quenching time, as evidenced by the maximum repetition rate in a relaxation discharge circuit was independent of the gas pressure. This substantiates the above quenching theory.

While the material of the baffles appears not to be critical for deionization, it may be of importance to the start of the discharge. With 0.004 inch copper wire screens, the time jitter of the discharge was marked, exceeding 15 percent of the repetition interval; and the breakdown voltage of the gap varied little from the beginning of operation with continued operation even at high repetition rates. With 0.001 inch nickel wire screening, the time jitter was much less; and the gap breakdown potential decreased by as much as one-half with continuous operation. It is believed that the fine nickel wire became sufficiently hot during the continuous operation so that the field emission produced by the application of potential between electrodes 1 and 2 was markedly increased, and provided rapidly a more abundant supply of electrons to promote breakdown.

Since the gap represented in FIG. 1 is made gas tight by hermetically sealing the various interfaces between its external component parts, it is evident that it may be filled with different gases, and that it may be pressurized.

Tests by filling with nitrogen, hydrogen, and argon gave rise times, respectively, of 1.9, 4, and nanoseconds. These compare with 2 nanoseconds for air. Monatomic gases produce a large proportion of electronelastic collisons, with resulting low electron drift velocities, and so increase the rise time. Polyatomic gases are therefore to be preferred. It is to be noted that while increased pressure increases the gap breakdown voltage, it does not affect the quench or deionization time, and thus is a convenient way of achieving the indicated result.

While the representation of FIG. 1 shows only two baffles, 8 and 9, if a long total gap is required for a high breakdown voltage, the number of baffles should be increased to permit retention even with the longer gap of the small spacing between adjacent electrodes, that is, between the electrodes 1 and 2 and the baffles nearest to them, and between adjacent baffles. For a short gap (e.g. 0.020 inch exclusive of baffle thickness) one will suffice.

The fast rise time for the invention is obtained because upon initiation of the breakdown, energetic electrons and photons may traverse through the porous screens. Thus, rise times comparable to a conventional spark gap are achieved as the charge multiplication (avalanche) on breakdown is present throughout the main gap. Prior quenching spark gaps with solid disc interelectrodes, would spark over sequentially by overvoltaging each successive gap, and thus were slower in total rise time.

Cooling of the gap at high average powers is, of course. a design consideration. My preferred way of metting this requirement is to increase the diameter of the electrodes 1 and 2 and correspondingly of the bafwhich are of metal and extend to the outside so that they may be cooled by various means, increases proportionally to the increase in gap area, and their thermal resistance to proportionally reduced. It would, of course, be possible to alter the design so that the outer portions of the baffles extended through the housing; and the entire device could be immersed in an insulating fluid, such as oil, to prevent external breakdown and convey heat. This is, however, much more complex and I do not now prefer it, although it must be recognized as a possible expedient for particular requirements.

Also pertinent to cooling is the fraction of the baffle area which is perforated. In screening this is of the order of one third. Ideally, the amount of solid baffle material should presumably be as little as is compatible with the baffle producing the effects hypothesized in the preceding; but to carry this to extremes will also reduce the material available for heat flow. The fraction must therefore be a design compromise, depending inter alia upon the magnitude of the cooling problem.

What is claimed is:

l. A fast-rise-time quenched spark gap comprising:

a. a pair of fixed electrodes insulated from and opposed to each other to form a gap between them. and adapted to be connected to a source of potential;

b. at least one foraminate electrode or baffle having its foraminate portion in the gap between the fixed electrodes, spaced uniformly apart from and insulated from all other electrodes,

c. in an atmosphere of gas at a pressure at least approximately 1 atmosphere.

2. The device of claim 1 which comprises a plurality of foraminate electrodes. spaced apart from and insulated from the fixed electrodes and from each other.

Claims (2)

1. A fast-rise-time quenched spark gap comprising: a. a pair of fixed electrodes insulated from and opposed to each other to form a gap between them, and adapted to be connected to a source of potential; b. at least one foraminate electrode or baffle having its foraminate portion in the gap between the fixed electrodes, spaced uniformly apart from and insulated from all other electrodes, c. in an atmosphere of gas at a pressure at least approximately 1 atmosphere.

2. The device of claim 1 which comprises a plurality of foraminate electrodes, spaced apart from and insulated from the fixed electrodes and from each other.

Images