US2441863A - Electrode for discharge devices - Google Patents

Description

May11s,194zs.v w RZABEL 2,441,863

vELECTRODE FOR DISCHARGE DEVICES Filed March 1o, 1945 /L//5 A 7TH/wn as cathode.

Patented May 18, 1948 ELECTRODE DISCHARGE DEVICES wuuam r. zaten', cleveland Heights, ohio, assignor to General Electric Company, a corporation of New York This invention relates to electric lamps and discharge devices, and is concerned with lament coils and electrodes. It is especially adaptable and advantageous for discharge devices' in which the pressure of the internal atmosphere is low during starting. I have hereinafter explained the invention with reference to lamps or tubes of low pressure type, whichare characterized inr operation by a diffuse discharge that substantially lls the discharge envelope, and particularly long gap, tubular, positive column lamps like ordinary fluorescent lamps and germicidal tubes.

`But it may also be found useful in devices whose internal pressures build up to discharge-constricting values during operation.v 1 f Besides the effects of long use which usually end their life, discharge devices are often subject to a progressive discoloration or darkening of the envelope walls in the neighborhood of the elec-I trodes, commonly known as "end blackening. This may become pronounced long before the -useiul life of the device is over. Though it does not very greatly reduce the light or other useful radiation, it renders the lamp unsightly and is` strongly objected to by users, who tend to regard it as the mark of a defective lamp. The nature and causes of such darkening are not fully understoodoi-.agreed on by those skilled in the art; but,it is conceded to be somehow due to matter reaching the envelope wall from the electrode(s) or the,current lead(s)4. 'Ihe darkening is apt to be worse when lamps areI turned on and oil? very frequently, and it is generally agreed that the causative loss of material from the electrodes or leads occurs very largely during starting of the discharge, and While an electrode is functioning It is Worse when discharge devices are started cold, or without adequate preheat of the electrode(s), though it is also met with in devices started hot, after preheat of the cathode(s) by preliminary currentow. Asl mechanisms any or all of whichmay beV involved in the darkening, there have been suggested: sputtering of electrode material under bombardment of positive ionsin 4the discharge; Asimple vaporization by heat; and transferA of matter from the electrode(s) to .the envelope walls by reaction of the hot cathodel or lead material with small amounts of unwanted gas(es) in the device followedby decomposition of the resulting comon the envelope wall, in a continuous cycle.

ln order to obviate end blackening and the losses v.of electrode material that give rise to it. I employ one or more electrodes of hollow grid-like structure, at least internally thermionic and suitfu sppucaunn Mmn 1o, 194s, serial No. 582,021

' z claims. (o1. 17e- 122) ably proportioned inrelation to the pressure of the `atmosphere in the device during starting.

My electrodes, etc., are effective in lamps started4 cold, as well as inthose started hot, and for operation either on A. C. or on D. C. The type of hollow grid structure herein illustrated consists of a helical coil of main conductor wire, with or without an associated -wire overwind. Important dimensions for correlation with the starting pressure are the internal size or diameter of the electrode structure, and the distanceapart of the grid members or coil turns. Other features are the electrode mass and wire size(s) the disposition of the activating material for the electrode, if any; and the material of the current lead(s).

By suitable internal size of the electrode grid or coil and suitable spacing apart of the grid membersor coil turns, as above mentioned, an internal glow discharge is made to occur within the electrode structure, at such a stage of starting and with such intensity that this internal glow heats the electrode and renders it thermionically emissive earlier and more copiously than would be th'e case with just the usual external' glow. This results in an easier increase in the discharge current, with a smoother transitionvfrom glow discharge to positive column are discharge at a lower peak value of the abnormal cathode drop (as it is called) than would Otherwise be necessary. Accordingly, the energy of the positive ions in the discharge is lower, and does not sufiice to sputter the electrode or leadwire material. Such rapid heating of the electrode necessarily requires a. suitably low electrode mass and wire sizethough it must not, of course, be so small that any electrode or leadwire material will overheat and vaporize, either during starting, or in the subsequent arc discharge operation from a locally heated cathode spot.

When the internal coil size, the coil turn spacing, and the wire size are suitably proportioned, and the current leads are made of suitable material, not only is sputtering prevented as just explained, but neither vaporization nor chemical attack of electrode material occurs, and substantial end blackening during useful lamp life is altogether avoided;

, To bring about an internal glow capable of producing the results above stated, both the internal coil size and the coil turn spacing must be within proper limits for the starting pressure in the device. If the coil is too small or too large, the internalglow will not form, or will be unable to heat the coil quickly, or will occur too late in the 3 starting cycle to control the abnormal cathode drop. Or if the coil turn spacing is too great, the internal glow will fail to form or prove too feeble, even though the coil size should-be correct. The interval or spacing between coil turns is extremely critical, in that the allowable maximum spacing is so exceedingly small. This close spacing may also be of advantage after starting has been accomplished, in allowing the cathode spot during ordinary running to spread over more wire turns and the activating material close to them; moreover, adjacent turns mutually heat one another very effectively, so that adequate emission is obtained from the total hot area without overheating any coil turn portion.

While th'e main wire coil in the forms of electrode here illustrated is shown as a simple helix, and the overwind as hugging it in direct contact everywhere, neither need necessarily be the case: i. e., the main wire might itself be coiled and recoiled one or more times, or the overwind wire might surround the main wire quite loosely, or

annees by the line and arrows l-l in Fig. 3; Fig. 5 is a view similar to Fig. 3 showing the other end of the electrode coil; Fig. 6 is a side view of a completed electrode on a still larger scale, partly in longitudinal section; and Fig. 'l is a diagram illustrating the effect of my invention on the starting cycle, the abscissae representing discharge current and the ordinates representing cathode drop.

Figs. 8 and 9 are fragmentary side views showing a wire coil with an overwind, and also illustrating stages in its fabrication; and Figs. 10 and 11 are similar views illustrating asomewhat different overwind.

both. This is exemplified in U. S. Patent No.1

2,306,925 to John O. Aich'er, where a main wire forming a coiled coil carries a loose overwind whose convolutions are considerably larger than the main wire, and either circular or of elongated loop form. By proportionng the Aicher structures as herein indicated with respect to the internal diameter of the major coil (taken to the inside of the overwind on the initial main wire turns), the spaces between major coil turns (taken between the nearest sides of the overwinds on said initial main wire turns), and th'e mass of overwind or other wire that must be heated by the discharge in starting, the known advantages of the Aicher constructions can be combined with those of my present invention.

However the electrode may be constructed, any coating or filling of activating material is preferably enclosed in a basket or "barrel grill embodied in the coil structure. In the forms of cathode here illustrated, the basket for the activating material is formed by the main wire coil, with itsoverwind, if any; in the forms illustrated in the Aicher patent, the basket afforded by the loose overwind or by the initial coiling of the main wire (or both) may sufce, without any need for a charge of activating material inside the final or major coil.

While I have herein illustrated the overwind as of smaller gauge than the main wire, this is not necessarily the case. Also, either the main wire or the overwind (if there is one) may be double or multi-stranded, and the component strands of either may or may not be twisted or Ibraided together. Again, the Wires need not be circular, but may have oval, rectangular, or ribbon-like cross sections.

Various features and advantages of my invention besides those hereinbefore mentioned will appear from the description of species and forms of embodiment, and from the drawing.

In the drawing, Fig. 1 is a. diagrammatic view of an electric discharge lamp in which my invention may be embodied, with suitable circuit connections; Fig. 2 is a tilted side view of an electrode mount suitable for such a lamp, prior to sealing into an envelope; Fig. `3 is a larger fragmentary side view of part of an electrode coil withits charge of activating material and one of its current leads, after it has been subjected to the usual activating treatment incidental to Broadly speaking, the device as illustrated in Fig, 1 may be similar in construction and operation to low pressure positive column fluorescent or germicidal lamps now in general use, comprising a longvtubular vitreous envelope l permeable to the desired lamp radiation, with reentrant tubular vitreous stem flares 2 sealed into its opposite ends, and wires 3. I (one or both serving as current leads) extending through the sealed inner ends of the stem tubes, Fig. 2. At one end of the lamp, at least, an exhaust tube 5 opens through the stem seal and extends out through the stem to its own seal 6. To adapt the device for operation on either D. C. or A. C., the electrodes 1, 1 in the opposite ends of the envelope I may be Just alike, and are here shown as comprising wire coils supported by theirend connections to the wires 3, I, whose ends may be extended to serve as auxiliary anodes. The ion- Y izable discharge atmosphere in the envelope lI may consist of gas or of some vaporizable working substance, such as metal. In low pressure positive column devices, for example, Jinert vrare starting gas like argon may be used at a pressure of some 2 to 6 mm., more or less-say 31A; mm. for a 40 watt fluorescent lamp-and mercury to an amount exceeding what will vaporize during operation, indicated in Fig. 1 by a droplet 8. A suitable pressure of mercury vapor during operation is of the order of 10 microns, more or less.

A layer of luminescent material or phosphor 9 excitable by the radiation from the discharge is shown as internally coated on the glass envelope wall I throughout its length.

Figs. 3, 4, 5, and 6 illustrate a thermidnic electrode 1 as consisting of a simple helical wire coilwhich is characterized by the close spacing of the coil turns or convolutions. The coil 1 may be of tungsten, tantalum, or any other refractory or otherwise suitable metal. The tubular grid formed by the coil convolutions 1 aifords room for an ample supply of activating material I0, such as a mixture of alkaline earth oxides including those of barium and strontium. The mass of activating material I0 is securely held by the closely spaced Wire turns 1, while amply Fig. 3, -with the wire coil in section as indicated exposed between them; and the convolutions 1 key into the electro-insulative activating mass III and are thus anchored and insulated against contact or short circuiting, while freely exposed at their own outer sides. The coil 1 electrically shields material I0 from the discharge during running. Straight end legs II of the wire 1 projecting beyond the ends of the coil may be clamped in the hooks I2, I2 of the wires 3, 4. As shown, the activating material I Ii does not extend quite the full length of the coil 1, so that the last few coil turns at each end are empty and open; also. the mass of, activating material I0 may be hollowed out (approximately as indicated at I3) .at least for some distanceinward from each end, or even throughout itslength. The .spaces between the coil ends and the sides of the clamphooks I2, I2y permit free access of the discharge into 'the coil 1, and linto any cavity(ies) I3, during y starting, as described hereinafter.

Illustrative circuit connections for cold-starting operation of the lamp are shown in Fig. 1 as including a high-leakage-reactance step-up autotransformer I4 whose primary connection is to an A. C. power supply circuit I5, while its secondary circuit I6 forms the discharge circuit connected across the electrodes 1, 1. Properly se lected, such a transformer I4 gives a suillclently high voltage for starting the discharge, and a suitably lower operating voltage when discharge current is drawn from it after startingv has been accomplished.

In the cold-starting of a fluorescent or germicidal lamp Vequipped ,with ordinary coil electrode(s), a diifuse glow discharge surrounds the Y cathode(s) while the current is increasing according to the thinner curve in Fig. 7, throughout both the period of stable moderate normal cathode drop (e. g., 90 volts) from A to B and that of rising abnormal cathode drop from B lto C. As or after the peak C (e, g., 250 volts) is passed and while the current is still increasing during the falling oil of the cathode drop from C to D or thereabout, transition from starting to ordinary running occurs: the glow around the cathode goes out, and the positive column arc discharge becomes established, taking 0E from a small highly heated spot on the cathode that provides abundant thermionic emission. The cathode drop during running has the relatively low value indicated at D (e. g., 17 volts).

With my novel electrode(s) 1, another phenomenon supervenes. Soon after the external glow has appeared around the cathode, an intense internal glow appears in its unfilled end turns and any internal cavity or cavities I3, representing a direct increase in the discharge current, and rapidly heating the cathode to temperatures of greatly increased thermionic emission that result in cumulative further increases of discharge current and of electrode heating. In the case illustrated in Fig. 7. this internal glow starts substantially at the transition from normal to abnormal cathode drop, on the portion of the curve from B to G. As represented in Fig. 7 by a heavy line leaving the thinner curve at G (e. g., at 170 volts), the appearance of the internal glow is quickly marked by a cessation of the rise'in the abnormal cathode drop, and even by a rather abrupt decrease of this drop .to H, after which it changes more gradually as the current increases during the remainder of the usual abnormal cathode` drop period, until the heavy curve approaches the thinner one and rejoins it atlI. It is to be remarked that the external glow, which began before the internal glow, may continue awhile after the latter appears, but tends to go out before the ultimate transition from glow tov arc discharge. It is of no consequence if the activating material I0 should extend out ytoo close to the ends of the coil 1, or should not be hollowed out as at I3, because the starting discharge will quickly dig out the superfluousend material and thus provide adequate hollowing and open coil turns.

The substantial lowering of the peak of the abnormal cathode drop' phase of the cathode cycle (e. g., from 250 volts to 170 volts) that is produced by my electrode(s) 1 as described sufces to obvlate sputtering of the electrodes, owing to the lower energy that this lower voltwhich may arise in the age can give the positive ions when they strike the electrode surfaces after repeated atomic collisions in the discharge. With low current densities in the discharge and a pressure in the envelope I giving a thick dark space at the cathode, this is possible even though the cathode drop greatly exceeds the disintegration voltage for the gas or vapor and electrode material(s) involved. Moreover, any particles sputtered from the internal cathode surfaces are mainly harmless as regards the envelope I', since they only strike the closely spaced coil turns 1. f

The relation of the internal .glow to the coil size and coil turn spacing mustbe understood from the nature of the glow discharge, which comprises what is known as a cathode sheath or Crookes dark space next to the emissive cathode surface, and outside of this an outer luminous cathodic glow or "negative glow, in whichy the heat of the discharger is mainly generated. The thickness of the cathode sheath or dark space diminishes as the pressure in the tube I is raised, or as the current density increases, and vice-versa. When an emissive wire is helically coiled in large, widely spaced turns, the surrounding cathode sheath and glow become a helical tube around the wire. If the coil turn spacing is reduced, the cathode sheaths and glows ,of individual turns eventually merge into tubular sheath-and-glow layers surrounding and lining the coil as such. Under this condition, the internal glow is`much more eilective in heating the electrode than is the external glow; for many of the internally emitted electrons travel back and forth inside the coll until all their energy is dissipated there, whereas the externally emitted electrons are free to move olf too far from the electrode to afford it much heat. The larger the coil, the more of it there is to be heated by its internal glow, and to dissipate the heat; and vice-versa. As the coil size is reduced, however, it eventually becomes too small to accommodate the necessary thickness of the internal cathode sheath for a given gas pressure and current density, and it then becomes impossibler to have any internal glow except at a higher current density. Accordingly, the internal glow may startconsiderably after the transition from normal to abnormal cathode drop, and even at or beyond the peak C in Fig. 7, or may not occur at all.

The energy of the internal glow resides in the electrons and positive ions inside the coil, and

in the heat of the gas there. As electrons, ions,y

and heat all tend to escape between the coi-l turns 1, the intensity of the internal glow and the consequent heating and thermionic emission of the coil increase as the coll turn spacing is reduced, and'vice-versa. Therefore the coil turn spacing should be as small as consists with avoiding short-circuiting contacts vbetween coil turns,

manufacture of the`e1ectrode. l

' For a starting gas pressure of 31%; mm. such as above mentioned, satisfactory starting without sputtering can be hadwithelectrode coils 1 ranging in internal diameter from about 15 mils or less up to about 40 mils or more. I at ,f present prefer about 25 to 30 mils for currents l, up to about l/2 ampere. The corresponding coil )turn spacing should preferably be. as small as is practicable in manufacture. In one specific example hereinafter, which yhas given very good results, the spacing is about 1 mil; 3 mils has also given good results in cathodes of coiled coi-l type, including theoverwound forms disclosed in Patent No. 2,306,925 above cited, and mils is considered practicable; while 8 mils is definitely too large. There is some reason to believe that coiled coils may allow of larger coil turn spacings than do simple coils with or without tight overwinds, becauseof the greater length of the narrow passages between the larger sized major convolutions. As regards lower and higher starting gas pressures even outside the 2-6 mm. range mentioned above for low pressure positive column devices, suillce to say that both internal'coil size and coil turn spacing should vary inversely with the pressure. Accordingly, the invention is practicaoble with argon as the starting gas at a pressure under substantially 25 mm. of mercury, and with other starting' gases at corresponding pressures. However, I generally prefer a more limited pressure rangel corresponding to substantially 1 to 10 mm. for argon.

It is to be remarked that the incidence of the out-side glow relative to the starting cycle, Fig. '1, depends on the internal size of the coil 1. because of the fact that the thickness of the Crookes dark space inside the coil ,decreases as the current density there increases, and viceversa. It the coil 1 is too small, an internal glow inside it may be impossible on any current value reached before the peak of the abnormal cathode drop, or the glow may occur too late to lower the peak enough to do any good. If the coil is too large, the heating eiect of theintern-al glow may be inadequate to heat the large mass of the coil to temperatures of adequate emission. The wire size being controlled by the necessity that the arc discharge shall bring the coil to a tempera-v ture affording adequate emission, but not high enough to evaporate off the activating materialas is well understood in the art-it follows that enlargement of the coil size necessarily increases its mass and radiating surface per turn, and usually its total mass as well. Fortunately, this limitation on the wire size lies within the range of what can be heated quickly and adequately by the internal glow in a coil of a size such as above indicated." y

The length of the coil 1 axially is mainly controlled by the consideration that it must hold enough activating material I0 to give the lamp a commercially satisfactory useful life. In the case of lamps intended to be used on hot-starting circuits that preheat the cathods, as well as on cold starting circuits, thecoil length may also be iniluenced by the desire that its resistance should afford a voltage drop adequate for arcing across the cathode(s) after suiilcient preheat by current flow through them, in accordance with usual hot-starting practice.

For the convenience of those wishing' to practice my invention, an example is given of appropriate proportions and procedure for an electrode for a 40-watt fluorescent lamp having a 48-inch long envelope tube I of 11/2 inch diameter, and intended to be operated with a current of about 0.41 ampere. However, these particulars are not to be understood as limiting or denning the invention in its broader aspects.

The coil 1 may be made of round non-sag tungsten wire of 3.5 mil diameter wound about 220 T. P. I. (tu'rns per inch) on a 30 mil diameter round iron mandrel (not shown), each coil consisting of 108 turns and having at each end an additional uncoiled length or leg II about 3 mm. long for mounting it, extending lengthwise of the coil. After heat treating or annealing at some l300 C. in moistened hydrogen to set the coil on the mandrel, and dissolving out the latter, the coil may be found to have an increased pitch corresponding to less than 220 T. P. I. This coil 1 is mounted by clamping in mount hooks I2, I2 spaced 13.5 mm. apart, after which the coil may be stretched by permanently spreading the clamps I2, I2 further apart, to a distance of about 15.8 to 16.3 mm. This gives a spacing of substantially 1 mil between adjacent coil turns 1.

The coil 1 may then be charged with activating material such as a mixture of powdered alkaline earth metal carbonatos in a binder of nitrocellulose lacquer and diluent, if needed. This may conveniently be applied by dipping the coil 1 between the clamps I2, I2 into the liquid mix in a spoon (not shownl-i'or example, as described in U. S. PatentNo. 2,363,055 to Flaws and in my U. S. application Serial No. 548,852, led August 10, 1944, which issued June 17, 1947, as U. S. Patent No. 2,422,45'7leaving a few turns at each end oi the coil empty, as suggested in Figs. 2, 3, 5, and 6. After drying out of the activating material in the coil 1, either naturally, or at a low temperature ,of some C., the mount (Fig. 2) is ready to be sealed into the ends of an envelope tube Las indicated in Fig. 1. D'uring the exhaustion and processing of the lamp, the electrode(s) 1 are treated and activated in the usual way, to break down the metal compounds of the mixture to oxides, render the product electron-emissive, and "degas" the electrode(s). After this, the lamp is exhausted, charged with mercury and starting gas,

and sealed off, as indicated at 6 in Fig. l.

To enable the coilsy 1 to be filled more uniformly and completely with activating material, the liquid mix used may be thicker than that commonly employed for commercial coiled coil electrodes, about of the consistency of very thick cream. I generally prefer to fill the cross-section of the coil 1 asl completely as possible; but in the drying out of the mix in the coil and in the processing and conversion to oxides, cavities or "pipes as indicated in Fig. 6 develop in the ends of the activating mass I0, and may connect together as a bore clear through the center of the mass. I do not aim to embed the coil 1 completely in the mass I0, or'even to coat its convolutions completely, but rather to leave their outer sides exposed as shown in Figs.r3, 5, and 6.

The inlead wires 3, I are shown in Fig. 2 as of the usual type, comprising short sections I1 of seal wire (such as that known commercially as Dumet) butt welded to outer and inner sections of more ordinary metal.- Instead oi iron or even nickel or alloys thereof, I prefer to make the inner lead sections (which extend inside the lamp and form the clamps I2, I2 and the auxiliary anodes) of copper, deoxidized and, ii desired, lightly nickel-plated to obviate or minimize oxidation in sealing them into the glass. as at the stems 2, 2. I find that with proper sized inner leads of this metal, the envelope I remains substantially free of discoloration or end blackening throughout the useful life of the lamp. This is the case even though some carbonate in the emission mix should not be completely broken down during the exhaust processing, etc., but remain to break down during the use of the lamp and yield carbon dioxide, which is reduced to carbon monoxide by the hot metal parts in the lamp. The results are best when the deoxidized copper inner lead sections are left unplated and in this condition sealed into the glass with such care as to avoid any material dition.

. 9 oxidation or absorption of gases during the sealing. For a 40-watt fluorescent lamp such as about referred to, copper leads of some 25 to 30 mil diameter are suitable.

The superiority of copper arises, as I believe, from the fact that it does not react with carbon monoxide, as does nickel or iron. In the case of electrode coil which differs from that shown in l Figs. 2 to 6 in that the once-coiled main wire 1 carries an overwind of finer wire I8. Both these wires may be used as received from the wireworks, without any preliminary cleaning or treatment. The main wire 1 is run through an ordinary, standard wire-winding machine (not shown), and overwound with the wire I8. The wire I8 'is helically wound on the then straight wire 1 with its turns spaced apart an amount of the order of its own thickness, or even comparable to that of the wire 1. As shown in Fig.

8, the thickness of the wire I8 is a substantial minor fraction of that of the wire 1, about onethird. Thereafter, the main wire 1 with its overwind I8 is helically wound on a mandrel wire I9, Fig. 9, considerably larger than the wires 1, I8, which may also -be conveniently done on an ordinary standard coiling machine. The spacing of the convolutions of the main wire 1 is much less than is usual for cathode coils, that is, of the order of the thickness of the wire 1, or even less. After winding, the coil may be heat treated, freed of its mandrel I9, again heat treated if required, and charged with activating material, etc., etc., as already described in connection with the simple coil of Figs. 2-6. Illustrative particulars suitable for a coil of this sort for a 40-watt uorescent lamp such as above referred to are as follows:

The wires 1 and I8 may be of 3.0 and 0.7 mil diameters, and the mandrel I9 may be iron or low carbon steel of 22.65 mil diameter. The

- wire I8 may be wound 293 T. P. I. on the wire 1, and this composite of wires 1, I8 may be uncoiled end legs II, II each 3 mm. long in ad- After' heat-treating, removal of the mandrel I9, and mounting between clamps 13 mm. apart, the coil may be stretched by spreading the clamps to 15 mm. apart, before charging with activating material, etc.

The variation illustrated in Figs. 10 and 11 differs from that of Figs. 8 and 9 in the wider spacing of the turns of the overwind wire I8,

- which is of the order of some twenty-five times its own thickness, or six times that of the wire 1. Illustrative particulars of a coil for a 40- watt fluorescent lamp such as above referred to are as follows:

The wires 1 and I8 may be of 3 and 0.7 mil sizes, and mandrel I8 of 22.65 mil size. The winding of wire I8 on wire 1 may be 54 T. P. I.. and that of wires 1 and I8 together onmandrel I8 may be 222 'I'. P. I. The coil length for an electrode may be 10 with additional end legs I I, II

as before. and the mounting, stretching. etc..

may also be as above indicated in connection with Figs. 8 and 9. l

What I claim as new and desire to secure by Letters Patent of the United States is:

1. In an electric discharge device the combination comprising an enclosing envelope, an ionizable atmosphere within said envelope comprising an inert startinggas having a pressure of about 3.5 mm. and mercury vapor at an operating pressure of about 10 microns, a thermionic electrode comprising a substantially helical wire coil structure embodying a helically wound wire of approximately 3 mils diameter wound to have an internal coil diameter of about 40 mils and an over-wound 0.7 mil diameter wire around the first-mentioned wire, activating alkaline earth oxide enclosed by said coil structure on the interior surface thereof, the internal coil size and the space between coil turns of the first-mentioned wire being 5 mils or less and correlated with the gas pressure to produce within said coil during starting, an internal glow constituting a merger of the negative glows associated with the individual coil turns for heating the thermionic electrode to a temperature capable of increasing the current sufficiently to produce transition from internal glow to an external arc discharge supporting a discharge current of l/2 ampere or less without transition through as great an abnormal cathode drop as would otherwise accompany transition from aglow discharge to an arc discharge.

2. In an electric discharge device the com-` bination comprising an enclosing envelope, an ionizable atmosphere within said envelope comprising an inert starting gas having a pressure of about 3.5 mm. and mercury vapor at an operating pressure of about 10 microns, a thermionic electrode comprising a substantially helical wire coil structure embodying a helically wound wire of approximately 3 mils diameter wound to have an internal coil diameter of about 40 mils and an over-wound 0.7 mil diameter wire around the first-mentioned wire. activating alkaline earth oxide enclosed by said coil structure on the interior surface thereof, the internal coil size and the space between coil turns of the first-mentioned wire being 5 mils or less and correlatedv with the gas pressure to produce within said coil during starting, an internal glow constituting a merger of the negative glows associated with the individual coil turns for heating the thermionic electrode to a temperature capable of increasing the current sumciently to produce transitionfrom internal glow to an external arc discharge.

1 WILLIAM P. IZABEL.

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

UNITED STATES PATENTS

Images