US2006731A - Discharge gap - Google Patents

Description

July 2, 1935. F. T. CHESNUT 2,006,731

DISCHARGE GAP Filed June 2, 1932 3 Sheets-Sheet l F. T. CHESNUT July 2, 1935.

D1 SCHARGE GAP Filed June 2, 1932 3 Sheets-Sheet 2 mar July 2, 1935. T CHESNUT 2,006,731

DISCHARGE GAP Filed June 2, 1932 3 Sheets-Sheet 3 j ower HHHIHMHHM 85 Patented July 2, 1935 v UNITED STATES PATENT OFFICE DISCHARGE GAP Frank T. Chesnut, Yardley, Pa... asaignor to Ajax Electrothermic Corporation, Ajax Park, N. 1., a corporation of New Jersey Application June 2, 1932, Serial No. 615,063

13 Claims. (Cl. 250-38) My invention relates to a mercury discharge Figure 1 is a central vertical section of one gap having a solid electrode, for producing high embodiment of my invention. frequency alternating current, with particular Figure 1a is a fragmentary view corresponding application to electric induction furnaces. v to Figure 1, showing the discharge gap tilted from A purpose of my invention is to produce a disthe position illustrated in Figure 1.

charge gap capable of sustained operation at Figure 2 is a front elevation, partly in central high power and preferably also at ni derate voltvertical section, of a discharge gap somewhat e. diiferent from that shown in Figure 1, and also A further purpose is to fill a mercu discharge embodying my invention.

gap with hydrogen at a pressure of 0.7 atmos- Figure 3 is afragmentary right end elevation of 10 phere (room temperature) or above, and to totally the discharge gap shown in Figure 2. seal the gap. Figure 4 is a central vertical section of a still A further purpose is to avoid the need for different form of discharge gap to which my invacuum-tight seals in a mercury discharge gap vention has been applied.

containing hydrogen. Figure 5 is a section upon the line 5--5 of Fig- A further purpose is to avoid the possibility of ure 4. introduction of air into a mercury-hydrogen dis- Figure 6 is a view corresponding to Figure 5 charge gap, thus eliminating danger of an exwith the gap somewhat tilted with respect to the plosion. position of Figure 5.

" A further purpose is to pass an electric cur- Figure I is a curve useful in illustrating the rent through a new mercury discharge gap and principles of my invention.

to clean up the gap by sweeping hydrogen Figure 8 is a diagrammatic illustration of a through the gap, and then, after the cleaning circuit in which the discharge gap of my invenup" process is completed, to totally seal the gap tion may be conveniently used.

retaining an atmosphere of hydrogen in the gap. In the drawings like numerals refer to like A further purpose is to decrease the power parts.

losses of a mercury-hydrogen discharge gap by High frequency electric current is required for requiring the discharge to pass only once from or a number of purposes, such as radio communicato mercury. tion and electric induction heating. For large a A further purpose is to vary the spacing beinstallations, it is possible to obtain high freetween the electrodes of a totally sealed discharge quency current from rotary generators, but for gap without permitting leakage through the gap smaller equipments rotary generators are unwalls. economical. Discharge gaps have frequently been A further purpose is to vary the spacing of used to generate high frequency current at modthe electrodes in a discharge gap having one erate power. 5 liquid electrode by tilting the gap. It has been found in the past that mercury A further purpose is to provide a flexible wall vapor produces a very desirable gap discharge of a discharge gap and to support the electrodes as explained in Edwin F. Northrup U. S..Patent of the gap on opposite ends of the flexible wall, No. 1,594,846, granted August 3, 1926, for Dismoving the electrodes relative to one another by charge gap. The mercury vapor is preferably 40 expanding or contracting the flexible wall. produced from a mercury electrode, which be- A further purpose is to operate a mercury discomes heated by the current. It is important to charge gap having a solid electrode and a merhave at least one solid electrode, as otherwise cury electrode, with a relatively hot band of the arc is likely to pass close to the wall of the 4., vapor above the mercury elect od to Prevent gap, destroying the wall, and also a general bakcondensation of mercury at that point, which ing out will occur in the gap, making the voltage would permit a short circuit discharge through excessive. Therefore, a discharge gap for my the condensate. purpose should contain mercury vapor, and should Further purposes appear in the specification have at least two electrodes, one of which is a m and in the claims. solid electrode and the other of which is prefer- In the drawings I have illustrated several emably a mercury electrode. bodiments of my invention, choosing the forms By experimenting with mercury discharge gaps shown from the standpoints of satisfactory ophaving one solid electrode, I have discovered that eration, simplicity of construction and conventhe character of atmosphere, aside from the merient illustration of the principles involved. cury vapor, within such gaps is a highly important 55 factor in determining the use to which the gap may be put and the satisfaction given by the gap in operation.

There are two aspects of the atmosphere in the discharge gap which are important to consider: its nature, and the pressure to which it is subjected.

I have studied discharge gaps with a view to determining what gaseous atmosphere will permit the largest power to be obtained from a given gap without short circuiting. I find that gases vary widely in the power which they will pass through the gap under the same gap conditions, and in the short circuit current which they will permit to pass. Of a large number of gases which I have tried, I find that hydrogen is much superior to all others. The power obtained without short circuiting from a discharge gap operating in hydrogen (in the presence of mercury vapor) is much higher than that obtained from the same gap operating with any other gas or mixture of gases under the same conditions. Nitrogen, oxygen, air, hydrocarbon vapor, the various rare gases, etc., are much inferior to hydrogen for discharge gap use.

Whether the superior quality of hydrogen is related to the fact that it is the lightest gas is a matter of conjecture, but that hydrogen in the presence of mercury vapor acts peculiarly and is superior to all other materials of which I know, for use in a discharge gap atmosphere, has been clearly established by my experiments.

In a mercury discharge gap having a solid electrode andoperating in a hydrogen atmosphere, I find that the pressure of the hydrogen is a very important factor, and that the change of power with hydrogen pressure is abnormal. At low pressures present in a mercury gap operatingunder vacuum conditions, and having hydrogen as a residual gas, the power is low, and a glow discharge fills the entire space around the gap.

As the pressure is increased, the power in creases progressively until a pressure of 0.7 atmosphere (measured at room temperature when the tube is not in operation) is reached, when the rate of change of power with pressure suddenly undergoes an alteration, and above this pressure the power increases but very slightly with further increase in pressure. In other words, from the standpoint of power, 0.7 atmosphere is the "saturation pressure.

In Figure 7 I have shown a curve having power as the ordinate and pressure in the gap measured at room temperature as the abscissa. The curve is not drawn accurately to scale but simply indi-' cates the trend of the relationship between pressure and power as observed in my experiments. At 0.7 atmosphere a point of maximum curvature in the curve exists, indicated by A. Below the point A the power changes very rapidly with change in pressure, while above the point A change in pressure makes relatively little difference in the power.

The explanation of the change in the relationship between power and pressure at 0.7 atmosphere may be that the type of discharge in the gap changes at pressures above this critical point. As a partial explanation of the action of the gap above 0.7 atmosphere, it may be that the mercury vapor acts primarily to conduct the current and permit high power, while the hydrogen actsprimarily to prevent short circuit current.

It is evident from a study of the curve of Figure 7 that high power operation of a mercuryhydrogen discharge gap having a solid electrode must take place at pressures above 0.7 atmosphere. For actual use, I find that one atmosphere is the most satisfactory pressure to employ.

In the first experiments made with the mercury-hydrogen discharge gap, it was believed that hydrogen was actually consumed in the gap, and therefore hydrogen was passed continuously through the gap from a source of hydrogen under pressure, thus continuously introducing new hydrogen into the gap and sweeping out that which formerly composed the gap atmosphere. My further experiments indicate that the previous deductions are not entirely correct. When a new discharge gap is started for the first time, hydrogen is apparently used in cleaning up" the gap, but after the gap has operated for a while, for example, a few hours, the consumption of hydrogen ceases and further change in the atmosphere is unnecessary.

In starting a new gap I preferably operate it while passing hydrogen through the gap for several hours, until the gap is thoroughly cleaned up and no further hydrogen is consumed. I then seal the gap entirely with the hydrogen atmosphere remaining in it, thus preventing introduction of foreign gas into the gap or escape of hydrogen from the gap.

By thus totally sealing the gap with its hydrogen atmosphere in it, I overcome a number of difliculties observed during experiments with the form of gap in which a hydrogen stream is continuously introduced and withdrawn.

The hydrogen stream gap has inevitably had leakage of air and the air has an unfortunate effect in causing explosions, which necessitates the use of a mica window in the wall of the discharge gap to avoid destruction of the whole apparatus.

No hydrogen tank anddistribution equipment is needed for a sealedgap, such as is required in the stream gap. I can afford to supply hydrogen in much purer condition for a sealed gap than for a stream gap because of the very much smaller quantity of hydrogen needed in a sealed gap. Furthermore, with a scaled gap, I have no loss of time in sweeping out the gap before it is started on every occasion.

While I consider it desirable to sweep out a new gap when it is first started, I find that I may less desirably fill a discharge gap with hydrogen; totally seal the gap and then operate it for the first time, allowing for the use of hydrogen in cleaning up the gap by introducing enough hydrogen so that the pressure will be at least 0.7 atmosphere after the discharge gap has operated for some time.

It will be evident of course that the pressure in the gap will change with the temperature, but, as noted, increase in pressure above 0.7 atmossphere serves merely to increase the power slightly, and therefore a discharge gap having a pressure of 0.7 atmosphere or above at room temperature will inevitably have a suitable pressure at the operating temperature of the discharge gap. When the operating temperature is reached the pressure in the discharge gap will of course be increased due to the increased vapor pressure of the mercury as well as the increased pressure of the hydrogen.

Many features of the construction of the discharge gap itself are matters of individual preferonce but I have illustrated several examples of discharge gaps which I have found to operate satisfactorily.

In Figures 1 and In I show generally diagrammatic illustrations of my preferred form of discharge gap.

The gap comprises two electrodes 2|! and 2|, of which 23 is a mercury electrode and 2| is a solid electrode. The electrodes are totally enclosed to prevent contamination of the atmosphere in the gap. Electric discharge takes place across the gap space 22.

The gap wall comprises an upper metallic truncated conical section 23 welded to the solid electrode 2| at 24, and having a very thin wall at its lower end 23 for convenience in welding it at 23 to the tubular glass wall 21. The truncated conical section 23 is preferably made of copper, which may be conveniently welded to glass. Chrome steel may also be used. Many other types of seal can be used. I have had good results with a packed seal using a plastic cement.

The lower portion of the discharge gap is formed by a lower metallic truncated conical section 23 having a thin wall at its upper end 29 for convenience in welding it to the glass at 30, and

threaded at 3|, near its lower end, to receive a cap 32. The lower truncated conical section 28 is preferably also made of copper, although it may be made of chrome steel. The inside of the cap 32 is preferably filled with gas-tight cement at 33, to prevent leakage through the threads.

A side neck 34 extends from the glass wall 21 of the gap, and is used for introducing the hydrogen into the gap. The side neck 34 is sealed at 35. For sweeping out the gap with hydrogen, I find that only one connection is necessary, as I can successively evacuate the gap and then introduce hydrogen as many times as may be desired to "clean up the gap. Two'connections may, however, be used, and in Figure 2 I show an additional connection 34' sealed at 35'.

Contact is made to the mercury electrode desirably by soldering a connection to the cap 32 at 36, and contact is made to the solid electrode 2| by soldering to it at 31.

The solid electrode 2| is preferably formed of copper, with or without a tip of some other metal, as later explained. It has an interior passage 38 for a cooling medium such as water, which is introduced through an inlet connection at 39, passing down through a pipe 40, and leaves the electrode through an outlet pipe 4|. The outlet and inlet may if desired be reversed.

The space 42 within the discharge gap is filled with hydrogen and with mercury vapor at a pressureexceeding 0.7 atmosphere when the gap is at room temperature. At operating temperature the pressure will be somewhat higher. Operating temperature may be, for example, between room temperature and C. in the mercury electrode, although the operating temperature may rise higher. I prefer to maintain the mercury electrode below 60 C. The gap is entirely sealed so that gas cannot enter the gap from the outside, and hydrogen and mercury vapor in the gap cannot escape. Under these conditions, the possibility of an explosion due to the hydrogen is greatly reduced.

The tube is not a vacuum tube, and the construction is greatly simplified because vacuum conditions do not exist. The seals at various points throughout the tube need only be gastight, and need not be vacuum-tight; a very important difference in commercial operation. In large apparatus, it is almost impossible to produce a vacuum-tight seal. For example, in making large mercury arc rectiflers, the search for a vacuum-tight seal has been practically given up, and all such rectifiers are supplied with a vacuum pump to care for the inevitable leakage. With gas-tight seals the situation is diilerent, however, and it is possible to obtain a gas-tight seal even for large size apparatus.

Considerable heat is developed during operation of my gap, and mercury is vaporized and passes upward from the mercury electrode 20 into the space 42 around the solid electrode. It is desirable to cool the discharge gap to prevent undue rise in temperature, and therefore I immerse the entire gap in a suitable cooling medium, as for example, oil 43 contained in a'tank 44.

Due to the cooling, condensation of mercury vapor takes place within the tube. I find that there should be a hot band in the tube a short distance above the mercury electrode, to prevent condensation of mercury on the tube wall 21, since the glass wall acts as an electrical insulator and condensation at this point causes a short circuit discharge through the condensate. The entire wall 21 need not be hot, provided an annular band is hot enough to prevent the formation of a film of condensate extending from the top to the bottom of the gap. In order to maintain this hot band around the electrically insulating wall 21, I preferentially cool other parts of the gap. The mercury electrode is cooled by a coil 45 having inlet and outlet connections at 46, and containing a suitable cooling medium, such as water or an expanding vapor or gas. The top of the gap is cooled by a cooling coil 41 having inlet and outlet connections at 46', and desirably cooled in the same manner as the coil 45. The solid electrode 2| is cooled as previously explained, and a considerable part of the condensation takes place on the sides of the solid electrode. These sides should preferably be straight to prevent erosion.

A band around the insulating wall 21 is thus relatively hot, and condensation takes place at other parts of the gap rather than at that point. I find that it is not harmful to have condensate run down the wall 21, as long as a continuous film does not form.

In order to vary the power of the discharge gap, it is important to be able to change the distance between the electrodes. This has been accomplished in the past, as for example, by fastening one electrode to the wall of the tube by a screw thread, and screwing the electrode in or out. This form of adjustment is altogether unsuited to the present use because the thread must be adjustable and so cannot be cemented, and leakage inevitably takes place between the sliding surfaces of the thread.

Efforts have also been made to adjust the spacing between the electrodes by pumping mercury into or out of the gap, but this is unsatisfactory because it complicates the apparatus, opens the possibility that mercury will leak through the pump and cause leakage of gas into or from the gap, and causes pressure changes in the gap.

I have discovered that the spacing between the mercury and the solid electrode may bevaried while maintaining the system entirely closed, without sliding joints of any kind, and without change in the quantity of mercury in the gap.

There are a number of ways in which the gap spacing may be changed while maintaining the tube entirely sealed, and I am accordingly illustrating several, in order to suggest that this function may be accomplished in various ways.

In Figures 1 and 1a I show a plan for adjusting the gap spacing by tilting the gap. In Figure 1 the gap is vertical and the gap spacing at 22 is relatively wide, corresponding to a relatively high power. Figure la shows the discharge gap tilted bodily, as by rocking about a ball 48 supporting the-gap in a socket 49, decreasing the discharge distance to a new value 22' corresponding to a lower power in the tube. In Figure 1a the discharge will be largely concentrated at the edge 50 of the solid electrode 2|. It will be evident that the discharge space may be further decreased by further tilting the gap, from a maximum when the tube is vertical to a minimum when the tube is rocked far to one side. In a gap of the type of Figure 1, the rocking may take place in any direction, thus avoiding excessive wear on one portion of the electrode.

It should be noted that the lower surface of the solid electrode 2| has a substantial area, thus spacing the edges considerably and facilitating movement of the upper surface of the mercury electrode toward the solid electrode when the gap is tilted.

It is possible to move the solid electrode 2| toward the mercury electrode 20 without providing any sliding joint whatever in the wall of the gap and without changing the volume of mercury. In Figures 2 and 3 I illustrate one way of doing this, by the use of an expansible and contractible wall at 5|, commonly called a sylphon bellows-or other metal bellows.

The gap of Figures 2 and 3 is intended to be used in the tank of Figure 1 in place of the tube there shown. The construction of the lower portion of the tube wall and the general construction of the solid electrode 2| in Figures 2 and 3 are the same as of Figures 1 and la, except that the solid electrode is provided with a tungsten tip at 52.

The upper portion of the solid electrode 2| has welded to it at 53 a metallic collar 54, which is in turn welded at 55 to the upper end of the sylphon" element 56. The upper truncated conical section 23 is welded at 51 to the lower end of the sylphon element.

The solid electrode 2| is therefore able to move down by compressing the sylphon and to move up by expanding the sylphon. Upward and downward movement are made possible by a hand wheel 58 on a shaft 59 supported in bearings 60 in brackets 6| which are welded at 62 to the upper truncated conical section of the gap. The shaft 59 carries a pinion 63 which engages a rack 64 secured to the side of the solid electrode 2|. The rear surface of the solid electrode is guided by a roller 65 on a shaft 66 turning in bearings 61 on the brackets 6|.

In Figures 4, 5 and 6 I illustrate still a different way of varying the discharge gap spacing. The form .of Figures 4 to 6 is a modification of Figures 1 and 1a.

In Figures 4 to 6 the solid electrode 2| has an eccentric end 68 provided with an arcuate outer surface 69 which varies in its distance from the mercury electrode 20 depending upon the angular position of the gap.

The space 22 ,between the electrodes varies as the gap is tilted. For example, in Figures 4 and 5 the portion of the eccentric end 68 of the solid electrode 2| which extends farthest from the axis is down, and the gap space 22 is minimal. In Figure 6, the discharge gap has been tilted (or rotated, which amounts to the same thing in a circular tube), increasing the space 22 due to the angularshifting of the eccentric end 68.

The gap is desirably tilted by rotating it in bearings 10 by a gear Aside from the slightly different means for varying the gap spacing, the form oI-Figures 4 to 6 is very similar to that of Figures land 1a and to that of Figures 2 and 3. The entire gap of Figures 4 to 6 is intended to be immersed in oil as in the other forms, but will remain substantially horizontal rather than substantially vertical. The gap of Figures 4 to 6 has a glass wall 21' joined at each end by a copper neck 12 or 13, which is sealed to the glass at 14 and internally threaded at 15 to receive an externally threaded plug 16 or Tl. Each of the plugs 18 and TI has a small opening 18 or 19 passing through it, filled with cement 80. The openings 5 18 and I9 serve to introduce the hydrogen atmosphere into the discharge gap. The solid electrode 2| extends through and makes a tight fit with the plug I6 and is cooled as previously described. Electrical connection is made to the solid electrode 2|, as by soldering a wire to it at 31.

Electrical contact is made to the mercury electrode by a pipe 8| desirably of copper, passing through the plug I1 and serving also to cool the mercury electrode, since in the form of Figures 4 to 6,'the mercury electrode is in contact only with the glass wall of the tube, and not with the highly heat conducting copper and closures. The

inlet and outlet connections to pass a cooling medium through the .pipe 8| are provided at 82 and 83.

The end plugs 16 and 11 are packed by gastight gaskets 84 desirably consisting of lead or some similar soft metal.

It will be evident that in all of the forms shown, the gap space may be varied withina wide range without using any sliding connection, likely to leak and contaminate the gas atmosphere and without changing the amount of mercury in the gap. In my preferred form shown in Figures 1 and la, and also in the form of Figures 4 to 6, the gap spacing is changed by tilting the gap, while in the form of Figures 2 and 3, the solid electrode is moved with respect to. the mercury electrode, but the gap walls are entirely sealed at all times.

This avoids the possibility of contamination of hydrogen with air, with possible harmful effects upon the gap operation, for example due to explosions.

By comparative tests, I find that, for a given power, a mercury-hydrogen gap at a. pressure above 0.7 atmosphere operates with about onethird less gap loss where the current jumps only once through the mercury-hydrogen atmosphere in passing through the gap than when the current makes two such jumps. In other words, I find that a single mercury-hydrogen gap at the pressure range mentioned is much more efiicient than a multiple mercury-hydrogen gap at .the same pressure range. The reason for this remarkable fact appears to be that the loss voltage drop at each electrode surface is high compared to the loss voltage drop in the space between the electrodes due to the space charge at the electrode surfaces. By making the discharge traverse only one gap space, instead of two, I eliminate the loss voltage drops at two electrode surfaces, and increase the efiiciency.

My discharge gap is primarily intended for use in electric induction furnaces, to provide the high frequency oscillating current necessary for this use; It operates at a relatively low voltage, for example, 7000 volts. which is moderate compared induction furnace 89 through capacitors 90 and 9|. A discharge gap 92 of the type described herein is shunted across the electric induction furnace 89 between the transformer secondary l8 and the capacitors 90 and Si.

The discharge gap 92 has only one solid electrode 2| and a mercury electrode 20, so that the current jumps only once through the mercury-hydrogen atmosphere, and the gap losses are at a minimum.

In view of my invention and disclosure variations and modifications to meet individual whim or particular need will doubtless become evident to others skilled in the art, to obtain part or all of the benefits of my invention without copying the structure shown, and I, therefore, claim all such in so far as they fall within the reasonable spirit and scope of my invention.

Having thus described my invention, what I claim as new and desire to secure by Letters Patcut is:

1. In a discharge gap comprising two electrodes, at least one of which is mercury, and a hermetically sealed electrically insulating wall between said electrodes, the method of preventing short circuiting of said gap which consists in permitting said insulating wall to attain a temperature, by direct radiation of an arc, above which a substantial amount of mercury will not condense, and maintaining the temperature of other parts of the gap below the temperature of the electrically insulating wall to cause condensation of the mercury vapor to take place at other points in the gap rather than on the electrically insulating wall.

2. The method of producing a mercury discharge gap having a mercury electrode and a solid electrode and capable of operating under constant conditions, which consists in initially operating the gap with a hydrogen atmosphere, in changing the hydrogen atmosphere in the gap after the initial operation and in sealing up in the gap for subsequent operation a hydrogen atmosphere different from that in the gap when it was first operated and maintained at a pressure in excess of 0.7 atmosphere at room temperature.

3. The method of producing a discharge gap having a mercury electrode and a solid electrode and capable of operating at substantial power and relatively constant voltage, which consists in introducing a hydrogen atmosphere into the gap, in passing current between the electrodes while the hydrogen atmosphere is in the gap, in subsequently changing the hydrogen atmosphere in the gap repeatedly to clean-up the gap until consumption of'hydrogen ceases, passing current between the electrodes while successive hydrogen atmospheres are in the gap, and in sealing in the gap a hydrogen atmosphere at a pressure exceeding 0.7 atmosphere at room temperature.

4. A discharge gap comprising a solid electrode and a mercury electrode spaced apart, means for totally sealing the space between said electrodes comprising an insulating wall and means for cooling said electrodes and sealing means below the vaporization point of mercury while permitting said insulating wall to attain a temperature, by direct radiation from the arc, sui'ncient to prevent a substantial amount or mercury vapor from condensing thereon.

5. In a discharge gap, a mercury electrode containing an unchangeable quantity oi mercury, a solid electrode having an end of a plane and sub stantial area, the spacing oi the edges of said end being large relatively to the distance between said mercury and said solid electrode, a

gap wall forming a totally sealed space around the electrodes and means ror tilting the gap so that the gap between the electrodes can be varied while the gap is in operation.

6. In a discharge gap. a mercury electrode, a solid electrode having an end of substantial area above and in one position, parallel to the mercury electrode, the spacing of the edges 01' said end being large relatively to the distance between said mercury and said solid electrode, spaced and insulated from the mercury electrode, a wall surrounding the electrodes and totally sealing the gap and means for tilting the gap during operation.

7. In a discharge gap, a partially liquid electrode, a solid electrode having an end of a plane and substantial area, the spacing of the edges of said end being large relatively to the distance between said mercury and said solid electrode, a gap wall forming a hermetically sealed space around the electrodes, hydrogen at a pressure exceeding 0.7 atmosphere at room temperature in said space and means for tilting the gap during,

operation to vary the spacing between the electrodes.

8. In a discharge gap having a central axis, a mercury electrode, a solid electrode having a cylindrical surface eccentric with respect to the central axis above the mercury electrode, the spacing of the edges of said end being large relatively to the distance between said mercury and said solid electrode, spaced and insulated from the mercury electrode, and walls surrounding the electrodes and forming a totally sealed gap space, the distance between the electrodes being adapted to change when the gap is rotated about said central axis and means for rotating the gap.

9. In a discharge gap having a central axis, a mercury electrode, a solid electrode above the mercury electrode and spaced and insulated from the mercury electrode, having a surface eccentric with respect to the central axis which is at variant distances from the mercury electrode in different angular positions about said axis of the discharge gap, walls forming. a totally sealed space around the electrodes and hydrogen at a pressure exceeding 0.7 atmosphere at room temperatures and means for rotating the gap about its axis.

10. In a discharge gap, relatively spaced electrodes insulated from one another, walls surrounding the electrodes and forming a totally closed system, a portion of the walls being flexible to permit relative movement of the electrodes and one oi. the electrodes moving with the flexible portion of the walls without impairing the enclosure, and means for relatively moving the electrodes.

11. In a discharge gap, a mercury electrode, a solid electrode above the mercury electrode and spaced and insulated from it, walls forming a totally closed system around the electrodes, a portion of the walls being flexible to allow movement of the solid electrode with respect to the mercury electrode and the solid electrode moving with the flexible portion of the walls without impairing the enclosure, and means for moving the solid electrode toward and away from the mercury electrode.

12. In a discharge gap, a mercury electrode, a solid electrode having an end oi! a plane and substantial area above the mercury electrode and spaced from the mercury electrode, insulating walls forming a totally sealed space around the gap, means for cooling the electrodes and sealing junctures with the walls below the vaporization point 01' mercury permitting a band of said insulating wall to become heated, during operation, by direct radiation from the are above FRANK T. CHESNUT.

Images