CA1114030A - Glow discharge heating apparatus - Google Patents

Abstract

ABSTRACT OF THE DISCLOSURE
The present invention provides an electric shock preventing device for a glow discharge heating apparatus for heating a heated liquid with the generation of heat attendant upon a glow discharge caused between a pair of electrodes through which a heated liquid such as water flows, character-ized in that an insulating pipe is provided on each of an inflow port and an outflow port for the heated liquid on said electrode, said insulated pipe being provided on the extremity portion with a metallic pipe which is connected to ground.

Description

1114~30 BACKGR~UND OF TllE II~VENTION
.__ _ This invention rela~es to a glow discllarge hcating ayparatus for heating a liquid through tlle utilization of a glow discharge ~3stablished between a pa~r of electrodes involved ar~ is a divisional application of our copen~i~g Canadian Patent AEplication Serlal No. 299,801 filed on ~rch 28, 1978.
Japanese laid-open patent application No 106252/1976 describes and claims a glow discharge heating apparatus for heating a liqui(l by utilizing a l~henomenon that a glow disch~rge occurring between a pair of cathode and anode electrodes heats the cathode electrode to an elevated tempera-ture. The glow discharge heating apparatus disclosed in the cited patent application comprises a hollow cylindrical enclosure, a tubular cathode electrode coaxially entended an~l sealed through the enclosure, and having both ends open, a hollow cylindrical anode electrode disposed in the enclo-sure to surround the cathode electrode substantially through-our the length thereof to form an annular discharge gap therebetween, a source of DC voltage connected across the cathode and anode electrodes to cause a glow discharge therebetwcen. The cathode electrode is heated with the glow discharge to directly heat a liquid flowing there-througl~ .
lleating apparatus of this type referred to have instantalleously heated the liquid Wit]l the simple construc-tion and still with the high efficiency. Ilowcver, w]lere high currents are required to establish the glow discharge between the electrodes, it has been difficult to sustain the stabilized glow discharge therebetween. Thele have ~1~4~30 been a fear that the glow discharge will transit to an arc discharge as the case may be. Also the electrodes have been heated to be axially expanded. This might result in a fear that the apparatus is broken.
Further it has been difficult to reliably control the glow discharge because of the absence of a control circuit for starting and extinguishing the glow discharge.
Accordingly it is an object of the present invention to eliminate the disadvantages of the prior art practice as above described by the provision of a new and improved glow discharge heating appratus capable of always sustaining a stabilized glow discharge.
It is another object of the present invention to provide a new and improved glow discharge heating apparatus including means for absorbing thermal strains developed in electrodes thereby to provide a construction difficult to be broken.
It is still another object of the present invention to provide a new and improved glow discharge heating apparatus including a control circuit for easily controlling a qlow discharge occurring across a pair of electrodes involved.
SU~ARY OF THE INVENTION
The present invention provides an electric shock preventing device for a glow discharge heating apparatus for heating a heated liquid with the generation of heat attendant upon a glow discharge caused between a pair of electrodes through which a heated liquid such as water flows, characteri~ed in that an insulating pipe is provided on each of an inflow port and an outflow port for the heated liquid on said electrode, said insulated pipe being provided on the extremity portion with a metallic pipe which is connected to ground.

~114'~3iD

In a preferred embodiment of the present invention the source of voltage may comprise a source of DC voltage and a hollow anode electrode surrounds the middle portion of a hollow cathode electrode to form the predetermined discharge gap therebetween, the cathode electrode forming a flow path for the heated liquid.
In another preferred embodiment of the present invention the source of voltage may comprise a source of AC
voltage and the pair of electrodes are in the form of hollow cylinders having one end closed and substantially identical in shape to each other, the closed ends of the cylindrical electrodes abutting against each other to form the predetermined gap therebetween while flow con~iring means is disposed within each ele~trode to flow the liquid in contact relationship with and along the internal surface thereof.
In order to ensure that the glow discharge is prevented from transiting to an arc discharge, the glow discharge heating apparatus may advantageously include an auxiliary source of voltage for applying across the electrodes ~ ~ 30 I ..
I , a high voltage in excess of a discharge breakdown voltage across the electrodes upon a discharge voltage across the electrodes approaching a glow discharge-hold minimum voltage, to cause a pilot glow discharge therebetween to induce the principal glow discharge.
BRIEF DESCRIPTION OF THE DRAWINGS
The present invention will become more readily apparent fro~ the following detailed description taken in conjunction with the accompanying drawings in which:
Figure 1 is a longitudinal sectional view of a glow discharge heating apparatus constructed in accordance with the principles of the prior art;
Figure 2A is a schematic sectional view of a pair of opposite electrode useful in explaining the glow discharge;
Figure 2B is a graph illustrating a spatial voltage profile exhibited by the arrangement shown in Figure 2A;
Figure 3 is a fragmental schematic plan view illustrating how a quantity of input heat to a cathode ~:
electrode during a glow discharge is measured; -Figure 4 is a graph illustrating ~he results of the measurement shown in Figure 3 with the results of a : corresponding theoretical calculation;
. Figure 5 is a graph illustrating the relationship between a glow discharge voltage and a gap length through which a glow discharge is caused; :~
Figurç 6 is a graph illustrating the relationship between a voltage and a current for the glow discharge;
Figure 7 is a pers~ective view of a modeled ion 1~14Q30 .`

flux usefu] in explaining a quantity of input heat to a cathode e~ectrode resulting from a ~low discharge;
Figure 8 is a graph illustrating the current-to-voltage characteristics of the glow discharge;
Figures 9A and 9B are fragmental schematic plan views of a pair of opposite electrodes useful on explaining the principles of the present inventions;
Figure lOA, lpB and lOC are views similar to Figure 9A or 9B but illustrating typically electrode configurations embodying the principles of the present invention;
Figure 11 is a longitudinal sectional view of one embodiment according to the glow discharge heating apparatus of the present invention;
Figure 12 is a current-to-voltage characteristic curve for a glow discharge caused by the arrangement shown in Figure 11; ~
Figures 13 and 14 are graphs useful in sxplaining the principles of t~e present invention, Figure 15 is a longitudinal sectional view of a : modification of the arrangement shown in Figure 11;
Figures 16 an~. 17 are graphs illustrating the characteristics of the arran~ement shown in Figure 15;
Figure 18 is a longitudinal sectional view of

2~ another modification of the present invention;
Figure 19 is a graph illustrating the characteristic of the arrangement shown in Figure 18;
Figure 20, which appears on the same sheet as Fig. 23, shows a modification of the arrangement ~ 30 .....
shown in Figure 18 wherein Figure 20A is a cross sectional view and Figures 20B and 20C are side elevational views of the lefthand and righthand sides respectively;
Figure 21 is a view similar to Figure 18 but illustrating still another modification of the present invention;
Figure 22 is a view similar to Figure 18 but illustrating a modification of the arrangement shcwn in Figure 21;
Figure 23 is a view similar to Figure 18 but : illustrating another modification of the arrangement shown in Figure 21;
. ~igure 24 is a view similar to Figure 18 but ~ :
illustrating still another modification of the arrangement shown in Figure 21;
Figure 25 is a graph illustrating a leakage current :-calculated with the arrangement shown in Figure 24;
: Figuré 26 ls a graphical representation of voltage : and current waveforms developed in the arrangement of Figure 24 filled with a mixture of helium and hydrogen;
Fîgure 27 is graph illustrating the current-to-vol$age characteristics of glow discharges occurring in the arrangement of Figure 24 filled with mixtures of helium and hydrogen having different proportions thereof;
Figure 28 is a graph illustrating the theoretical relationship between a glow hold minimum voltage and quantity of input heat to an associated electrode resulting from the glow discharge;
.

1~4~30 S`
Figure 2~ is a graph illustrating the relationship between an overlapping area for both electrodes and a pressure of a filling gas;
Figure 30, 31 and 32 is graphs illustrating how S the glow hold minimum voltage is changed with a proportion of mixed gases and a discharge gap-length;
Figure 33 is a graph illustrating the relationship between the glow hold minimum voltage and a peak diccharge current;
Figure 34 is a longitudinal sectional view of a different modification of the present invention including an auxiliary electrode;
Figures 35, 36 and 37 are fragmental perspective views of different modifications of one of the electrodes shown in Figure 34;
Figure 38 is a longitudinal sectional view of modification of Figure 34 along with an associated electric circuit;
Figur~ 39 is a longitudinal sectional view of another modification of the arrangement shown in Figure 34;
Figure 40 is a view similar to Figure 38 but illustrating still another modification of the arrangement shown in Figure 34;
Figure 41 is view similar to Figure 39 but illustrat-ing a different modification of the arrangement shown in Figure 34;
Figure 4~ is a view similar to Figure 39 but illustrating a modification of the arrangement shown in 111~30 ) Figure 41;
Figure 43 is a view similar to Figure 39 but illustrating a modification of the arrangement shown in Figure 40;
S Figure 44 is a view similar to Figure 39 but illustrating another modification of the arrangement shown in Figure 34;
Figure 45 is a view similar to Figure 39 but illustrating a modification of the arrangement shown in Figure 44;
Figure 46 is a diagram of the fundamental used with control circuit the present invention;
Figure 47 is a graph illustrating a voltage and a current waveform developed in the arrangement shown in lS Figure 46;
Figure 48 is a diagram of a control circuit constructed in accordance with the principles of the present invention :
for driving the glow discharge heating apparatus thereof;
Figure 49 is a graph illustrating a voltage and a current waveform developed in the arrangement shown in Figure 48;
Figure 50 is a diagram similar to Figure 48 but illustrating a modification of the arrangement shown in Figure 48;
Figure 51 is a graph similar to Figure 49 but illustrating the arrangement shown in Figure 50;
Figure 52 is a diagram of another control circuit constructed in accordance with the principles of the ~ 433~

¦ present invention and suitable for use with an electrode ¦ structure including an auxiliary elec~rode;
l Figure 53 is a ci.rcuit d;agram similar to Figure 52 ¦ but illustrating a modification of the arrangement shown S ¦ in Figure 52;
¦ Figure 54, which appears on the same sheet as Fig. 56 and 57, ls a graph illustrating voltage waveforms developed at various points in the arrangement shown in Figure 52;
l Figures 55 through 5~ are circuit diagrams similar ¦ to Figure SZ but illustrating different modifications of the arrangement shown in Figure 52;
Figure 59 is a diagram of still another control circuit constructed in accordance with the principles of l the present invention;
lS ¦ Figure 60 is a graph illustrating voltage waveforms developed in the arrangement shown in Figure S9;
Figure 61 is a graphical representation of a Laue . .
plot; .
Figure 62 is a circuit dia~ram similar to Figure 59 but illustrating a modification of the arrangement shown in Figure 59;
Figure 63 is a graph similar to Figure 60 but illustrating.the arrangement shown in Figure 62; ~.
Figure 64 is a sectional view of an embodiment according to the three-phase glow discharge heating apparatus of the present invention and a diagram of a control circuit therefor;
Figure 65 is a graph illustrating various waveforms .

1114~i30 'i``:-developed in the arrangement shown in Figure 64;
Figure 66 is a diagram of the detailes of the control circuit shown in Figure 64;
Figure 67 is a wiring diagram of a modification of the arrangement shown in Figure 66;
Figure 68 is a graph similar to Figure 65 but illustrating the arrangement shown in Figure 67, Figure 69 is a longitudinal sectional view of a modification of the arrangement shown in Figure 44 and a diagram of a control circuit therefor;
Figure 70 is a longitudinal sectional view of a modification of the arrangement shown in Figure 69;
Figure 71 is a view similar to Figure 64 but illustrating a modification of the arrangement shown in Figure 64;
Figure 72 is a view similar to Figure 70 but illustrating a modification of the arrangement shown in Figure 69; and .
. Figure 73 is a longitudinal sectional view of another modification of the arrangement shown in Figure 69.
Throughout the Figures like reference numerals designate the identical or corresponding components.
DESCRIPTION OF T~IE PREFERRED EMBODIMENTS
Referring now to Figure 1 of the drawings, there is illustrated a conventional glow discharge-heating apparatus.
The arrangement illustrated comprises a hollow cylindrical cathode electrode 1) a hollow cylindrical anode electrode 2 surrounding coaxially the cathode electrode 1 to form an "1 1114~3~) ,~.
annular discharge gap 8 therebetwéen with the aid of two electrically insulating spacers 7 in the form of annuli fixedly disposed between both electrodes 1 and 2 adjacent to both end portions of the anode electrode 2, and a cylindrical S enclosure 9 formed of any suitable electrically insulating material such as glass and coaxially housing the electrodes 1 and 2 with the cathode electrode 1 hermetically extending through both ends thereof. A seal fitting 10 is sealed at the outer periphery to one end, in this case, the lefthand end as viewed in Figure 1 of the enclosure 9 and at the inner periphery to the adjacent portion of the cathode electrode 1 while a corrugated seal fitting 11 is sealed at the outer pheriphery to the other end of the envelope 9 and at the inner periphery to the adjacent portion of the cathode electrode 1. The corrugated seal fitting 11 is permitted to be axially contracted and expanded enough to prevent the cathode electrode 1 from damaging due to an axial thermal strains thereof. Thus the envelope 9 and the seal fittings 10 and 11 maintain the discharge gap 8 her~etic.
As shown in Figure 1, the anode electrode 2 includes flared end portions 2g in order to prevent an electric dischaTge from concentrating on the end portions of the anode electrode 2.
A positive terminal 5 connected to the central portion of the anode electrode 2 is e~tended and sealed through the central portion of the cylindrical peripheral wall of the enclosure 9 until it is connected to a positive . . . ,., ~ : -- ~ :

side of a source of DC voltage 3 having a negative side connected through a stabilizing resistor 4 to a negative terminal 6 that is, in turn, connected to one end portion, l in this case, the righthand end portion as viewed in ¦ Figure 1 of the cathode electrode 1.
In the arrangement of Figure 1, a DC voltage is applied across the anode and cathode electrodes 1 and 2 respectively to establish a glow discharge across the l discharge gap g thereby to heat the cathode electrode 1.
¦ Under these circumstances, a liquid to be heated such as water is caused to flow through the interior of the cathode electrode 1 to be directly heated by the heated cathode electrode 1.
l Conventional glow discharge heating apparatus such ¦ as shown in Figure 1 have been enabled to instantaneously heat liquids to be heated, for example, water resulting in heating apparatus simple in construction and still high in efficiency. However, since the apparatus have required the l high current, it has been extremely difficult to stably ¦ sustain the glow discharge across the anode and cathode electrodes. According to circumstances, there hac been a fear that the glow discharge transits to an arc discharge.
Further there has been a fear that, as a result of their ¦ heating, the electrodes are axially expanded leading to ¦ the destruc~ion of the heating apparatus. In addition, ¦ conventional glow discharge heating apparatus have not ¦ been provided with suitable control circuit means for j starting and ceasing the glow discharge with the result that lli~

it has been difficult to reliably control the ~low discharge.
The present invention contemplates to eliminate the disadvantages of and objections to the prior art practice I as above described and characterized by unique means for imparting a positive resistance to the current-to-voltage characteristic of the glow discharge. It ;~as been found that such characteristic is effective for preventing the transit of the glow discharge to an arc discharge.
For a better understanding of the principles of the present invention, the description will r.ow be made in conjunction with the glow discharge, the principles tha, it heats an associated cathode electrode and the current-to-voltage characteristic thereof. I
Figure 2A shows a pair of cathode and anode electrodes 1 and 2 respectively disposed in spaced opposite relationship and a source of DC voltage 3 including a negative side connected to the cathode electrode 1 and a positive side connected to the anode electrode 2 through a stabilizing resistor 4 whereby a glow discharge occurs within a discharge space formed between both electrodes 1 and 2. It is well ]cnown that the discharge space having the glow discharge established therein is divided into a region of ca~hode fall a in which positive ions are enriched, a region of negative glow b forming a thin lumunescent 2S layer, a Faraday dark space c in which no light is emitted, and a positive column do consisting of a plasma including electrons and ions, starting with the side of the cathode electrode 1.

.A I _ Figure 2B shows a spatial voltage profile in the discharge space with the glow discharge established therein.
In Figure 2B, a voltage V is plotted in ordinate against a distance d in abscissa measured from the cathode electrode 1.
From Figure 2B it is seen that the region of cathcde fall has a very large potential-gradient because the presence of a space charge until a cathode fall of potential Vc is reached at the end of the region a spaced from the surface of the cathode electrode 1 by a distance of dc. The voltage reaches a glow voltage Vg on the surface of the anode electrode 2.
By visually observing the glow discharge, it is seen that a boundary between the region of cathode fall a and the region of negative glow b are very distinct but a boundary between the region of negative glow _ and the Faraday dark space c or between the Faraday dark space c and the positive column do is not so distinct.
Also the Faraday dark space c and the positive column ~ are in the so-called plasma state and relatively small in potential gradient. On the other hand, the region of cathode fall a includes positive ions in the form of a beam. As far as the discharge current is concerned, it consists ess,entially of an electron current in each of the Faraday dark space c and positive column do which are in the plasma state and of an ion current in the region of cathode fall a. The region of negative glow _ forms a region of the transition of one to tlle otller of both currents.

P3~) ¦ Two phenomena developed in the region of cathode ¦ fall a, that is, (1) the mechanism by which the glow discharge is sustained and (2) a phenomenon that the cathode electrode is heated with the glow discharge as well as (3) the current-to-~oltage characteristic of the glow discharge are pertinent to the principles of the present invention and therefore will now be described.
(1) Mechanism of Sustaining Glow Discharge Positive ions present in the region of cathode fall a collide against the surface of the cathode electrode whereupon the cathode electrode 1 emits electrons by means of the action of emitting secondary electrons called the Yi action. The electrons emitted from the cathode electrode (1) collide against neutral atoms or molecules during their movement toward the anode electrode which is accompanied by an ionizing action called the a action with some probability. Electrons and positive ions caused by the~ ionization and co]lision are accelerated ~oward the anode and cathode electrodes respectively by means of the action of an electric field involved. It is noted that the positive ions accelerated with ~he electric ~ield contributes to the yl action.
Here the sustainment of the glow discharge will be somewhat quantitatively described. For example, it is said that, with the cathode electrode 1 formed of nickel, the Yi is approximately equal to 0.01 for slow helium ions having 1 Kev or less. That is, about 100 ions collide against the cathode electrode 1 to emit a single electron .' ' , `~i ~ 1~ 3~:) ~ ..
therefrom.
Also a degree of ionization ~ is a function of the type and pressure of a gas confined in the discharge space and a potential gradient developed therein. Electron-ion pairs formed at a distance x from the cathode electrode 1 are proportional to e~x where e designates the base of Napierian logarithms and therefore increase exponentially with the distance x. Accordingly~ the glow discharge is sustained with a distance and a voltage required for about 100 electron-ion pairs to be formed in the course of movement of a single electrode toward the anode electrode 2. This distance is designated by the distance dc shown in Figure 2B and this voltage substantially corresponds to the voltage Vc. In other words, the glow discharge can be sustained even when the anode electrode 2 has been displaced to its position substantially shcwn by dc in Figure 2A.
This is substantially applicable to electrodes formed of nickel, copper, iron, stainless steel or the like and operatively associated with a gas selected from the group consisting of helium, neon, argon, hydrogen, nitrogen etc.
A more detaile~analysis of the phenomena developed in the vicinity of the cathode electrode 1 teaches that a current density J on the surface of the cathode electrode 1 is expressed by J j+ + j = j+ = KlP ~1) 1~ 3~11 where i+ and i designate densitiés of positive ions and electrons respectively, P a pressure of a discharge gas, and Kl designates a constant determined by both thc type of a cathode material and that of the discharge gas.
S Also the region of cathode fall a has a thickness dc as defined by dcP ~2 (2) where K2 designates a constant dependent upon both the type of the cathode material and that of the discharge gas.
Within the region of normal glow, the cathode fall of "
potential Vc is determined by both the type of the cathode material and that of the discharge gas but scarcely depends upon both a discharge current and the pressure of the discharge gas, The following Table I lists values of the constants Kl and K2 and the ca~hode fall of potential Vc measured within the region of normal glow with different combina-tions of cathode materials and discharge gases with a glow current not higher than 1 amperes and with the discharge gases maintained under the pressure of 50 Torrs or more.
The measured.Kl and ~2 values are expressed in 10 6 ampere per cm2Torr2 and in cm-Torr and the voltage Vc is expressed in volts. Also the current density on the surface of the cathode electrode has been determined by measuring an area of a negative glow b. ProbablyJ the negative glow is very thin so that it is observed like a I .
luminescent film attached to the cathode electrode.
Table I
l ~EASU~ED VALUES OF Kl, K2 and Vc 1 _ Gas He Ne Ar H2 Cathode ~
1 6.0 8 3 27 24 Cu K2 3 0 3.0 0.8 2.0 I
c 150 150 180 290 Kl 8.0 20 32 32 l Ni K2 3 0 4.0 1.5 3.0 I
~- Vc 101 140 185 254 Kl 4.4 4.7 17 30 I
l Mo K2 --- ---- 3 0 0 8 3 0 I Vc 180 175 190 290 l l __ ¦ SUS 2 . 0 8 _~c 119 150 180 2~2 (2) }leating of Cathode Electrode As above described, positive ions present in the ¦ region of cathode fall a collide against the cathode ¦ electrode to cause the Yi action. At that time, the posi-¦ tive ions have surplus kinetic energy that is, in turn spent to heat the cathode electrode 1. Regarding quantities of input and output heat of the cathode electrodes, there are, in addition to collision with the positive ions, heat ~ ~114~30 ~ :

conduction from the plasma portions, exothermic and endothermic effects caused from chemical reactions effected on the surface of the cathode electrode 1 due to the glow l discharge, cooling effects caused from the sputtering on ¦ the cathode electrode and the evaporation of the cathode material etc. However, an extent to which a quantity of heat enters the cathode electrode has not been elucidated until the present.
In order to determine a quantity of input heat to the cathode electrode due to the glow discharge formed between that electrode and an anode electrode, experiments were conducted with a test device schematically shown in Figure 3. As shown in Figure 3, a cathode electrode 1 in the form of a very long circular rod having a radius r of 1.8 mm was disposed to be thermally isolated from the surrounding and opposite to a similar anode electrode 2 to form therebetween a gap having a length d of 4 mm.
Both electrodes were formed of copper and connected across a DC source 3 through a stabilizing resistor 4. Thus a glow discharge g is established across both electrodes 1 and 2 in the atmosphere. Under these circumstance, a radiation thermometer M was used to continuously measure a temperature at a point on the outer surface of the cathode electrode 1 spaced way from the discharge surface thereof by a distance ZO of 3 mm.
The results of the experiments are shown in Figure 4 wherein the temperature in Centigrade is in ordinate against time in seconds in abscissa with a glow current taken as . .' 111~030 ¦ the parameter. In Figure 4, each vertical se~inent designates ¦ a range in which measured values of the temperature are ¦ dispersed and solid curve describes calculated values of the temperature as will be described hereinafter. The refer~nce numerals 111, 112, 113 and 114 mean the tempera-tures measured and calculated with glow currents of 4~0, 250, 200 and 150 milliamperes respectively.
From Figure 4 it has been confirmed that the glow discharge ~ransits to an arc discharge upon the measured temperature approaching 1000C. This will be because an oxide film is formed on the surface of the cathode elec~rode at such a temperature.
It is now assumed that in Figure 3, the cathode electrode l with a radius r has ~he longitudinal axis lying on a z axis and the discharge surface passing through the origin for the z axis and that a quantity of input heat to the cathode electrode 1 is constant per unit area and per unit time. Under the assumed condition, by solving a partial differential equation for conduction of heat referred to the z axis alone and taking account of a radiation lOsâ may be expressed by where ~ designates a thermal diffusibility defined by the square root of the quotient of a thermal conductivity k of the cathode electrode divided by the product of a density p and a heat capacity thereof and ~ is a constant on the ~ ,;
.

i 1114~33~
; .. ....
assumptîon that the radiation loss is a li.near function of a temperature T. By solving the partial equation unclcr the boundary conditions s ~ aZ Iz=o -~ r2k and aZ Iz=co where ~ designates a coefficient of heat input and the initial condition I T(z,o) = T

where To designates room temperature, a solution results in T(z,t) = To ~

F(yl)) - e F(y2) where I: glow current.
In the expression (3) I(yl) and F(y2) are error functions expressed by

3~

(Yl) = ~ ¦; e 2 d~

and 2 ~ 1;2 respectively where rl and Y2 are expressed by Yl = - and Y2 =

respectively. Also a is defined by 2~(Ta3+ToTa2+To~a+To3) pca where E designates an emissivity, a a Stefan-Boltzmann constant and Ta designates the mean value of room tempera-ture and a temperature of the cathode electrode.
The expression ~3) was used to calculate the time dependency of the te~perature rise on the measured point as shown in Figure 3. The results of the calculations are indicated by the solid curves shown in Figure 4.
From Figure 4 it is seen tha~ the measured values of the temperature fairly well coincide Wit]l the calculated values thereof.

~14~33(:~ .

Figure 5 illustrates a glow discharge voltage V in volts plotted in ordinate against a length of a discharge gap in millimeters in abscissa. The voltage V was measured with the electrodes formed of copper and disposed in the S atmosphere. Curves labelled llS, 116, 117 and 118 depict glow currents of 10, 50, lO0 and 400 milliamperes respec-tively.
In Figure 4 it is to be noted that the curves have been drawn by equalling the cathode drop of potential Vc in the atmosphere to a voltage of 285 volts estimated with a null gap length from curves shown in Figure 5.
Also the coefficient of heat input ~ has been determined to cause the calculated valves of the temperature to coincide with the ~easured values thereof shown in Figure 4. The coefficient ~ has been of 1.4.
. Further it is considered that a quantity of heat corresponding to ~.4 iVc per unit area per unit time will result from one portion of heat generated in that portion of t~e glow discharge formed of both the Faraday dark space c and the positive column d except ~or the region of cathode fall a having a thickness dc approximately equal to 2 x 10 3 centimeter.
Figure 6 illustrates a glow voltage Vg in volts plotted in ordinate against a glow current I in milliamperes in abscissa. Curve labelled 119 describes the glow current-to-voltage characteristic exhibited by the arrange-ment of Figure 2. Dotted curve 120 shows the total power consumed by the glow discharge and expressed by IVg while ao broken curve 121 illustrates an electric power entering the cathode electrode and calculated as 1.4 ~Vg. Both the glow voltage and powers in watts are plotted in ordinate against the same glow current in absussa.
S From Figure 6 it is seen that at least 80 % of the total consumed power enters the cathode electrode and that the higher the glow current I the greater the proportion of the power entering the cathode electrode to the total consumed power will be.
Also it is seen that a quantity of input heat q to the cathode electrode 1 per unit area per unit time is give by q= jVc = jVg provided that the spacing d between the cathode and anode electrodes 1 and 2 Tespectively substantially approximates the thickness of the region of cathode fall a ~see Figure 2), that is to say, the glow discharge includes no plasma ? portion. From this it is seen that the smaller the spacing d between the cathode and anode electrodes the larger the proportion of the power entering the cathode electrode to the total consumed power will be.
Figure 7 shows a model for a positive ion flux 2~ striking against the unit area of the surface of the cathode electrode per unit time. In Figure 7, a square prism has a square bottom including each side of 1 centi-meter and contacting the surface of the cathode electrode ¦ I nd a height corresponding t~ the ~el~city Vl cm/sec of ions multiplied by one second. Within the prism, positive ions designated by the symbol "cross in circle" are moved as shown at the arrow ~o strike collide with the cathode S electrode 1. Thus the square prism designates a positive ion flux colliding against the cathode electrode per unit area per unit time and electrical energy of the ion flux results in the quantity of input heat q to the cathode electrode. Since the number of the positive ions is expressed by j/e where a designates the elementary electric charge and since each ion has electrical energy of eVc, the quantity of input heat q is expressed by q = eVc ~ = jVc in watts/cm2.
. .
Thus the model for the positive ion flux also ~ -explains that the quantity of input heat to the cathode electrode is expressed by jVc per unit area per unit time.
From the foregoing it will be understood that the glow discharge established across the cathode and anode electrodes causes the quantity of heat expressed by ~jvc to enter the cathode electrode per unit area per unit time.
Also by de~reasing the spacing between both electrodes to increase the glow current through the spacing, the quantity of input heat to the cathode electrode per unit a~ea per unit time can approximates the product of the current density on the surface of the cathode electrode multiplied by the glow voltage or J-Vg~

Therefore the glow discharge without the positive column can be utilized as a heat source having a high efficiency because almost all heat due to the glow discharge enters the cathode electrode and also as a heat source having a power density variable at will by changing a gas pressure within the spacing between both electrodes because the current density on the surface of the cathode electrode is proportional to the square of the gas pressure.
(3) Current-to-Voltage Characteristic of Glow Discharge ~he current-to-voltage characteristic of the glow discharge will now be described and then the principles of the present invention will be described in detail.
Figure 8 shows the relationship between a current lS and a voltage for the glow discharge. In Figure 8 the axis of abscissas represents a current and the axis of ordinates represents a voltage.
A DC voltage is applied across a cathode and an anode electrode 1 and 2 respectively (see Figure 9A) to render the anode electrode 2 positive with respect to the cathode electrode 1 thereby to cause to a glow discharge thereacross. When a current flowing through both electrodes is increased, a negative glow region b included in the glow discharge spreads in area on the surface of the cathode electrode 1 (see Figures ~A and 9B). This results in a change in current-to-voltage characteristic as shown at solid line N in Figure 8.
However, when the current is quite low, the ~
.
l ~
current-to-voltage characteristic droops as shown by a characteristic portion Nl in Figure 8. A region in which the drooping characteristic Nl appears is called a region l of subnormal glow e.
In a region following the region of subnormal glow e an increase in current causes the voltage to be kept substantially constant as shown by a characteristi.c portion N2 in Figure 8 as long as that the surface of the cathode electrode 1 having the negative glow b caused thereon is smaller in area than the entire surface thereof opposite to the anode electrode 2 as shown in Figure 9A. A region in which the characteristic portion N2 is developed is called a region of normal glow f.
A further increase in current causes an increase in voltage because the negative glow b has covered the entire area of the surface of the cathode electrode 1 opposite to the anode electrode 2 as shown in Figure 9B whereby the negative glow increases in current density. The resulting I-V characteristic is upturned with an increase in current as shown by a characteristic portion N3 in Figure 8. The characteristic portion N3 is called a positive resistance characteristic and a region in which the positive resistance characteTistic N3 appers is called a regi.on of abnormal glow. In that region of abnormal glow q the entire area of the surface of the cathode electrode 1 is covered with the negative glow b (see Figure 9B) with the result that the current is apt to concentrate at the edge portion or the like of the cathode eiectrode 1 and therefore the glow ~ 15.1~3~:) .
I
, I
discharge is easily changed to an arc discharge. As a ¦ result, it is difficult to maintain the glow discharge in ¦ its stable state. The arc discharge appears in a region ¦ h as shown in Figure 8.
¦ With no impedance connected between the cathode and ¦ anode electrodes 1 and 2 respectively and an electric source for supplying an electric power across both electrodes, the source side has the current-to-voltage characteristic l of the constant current type such as shown at horizontal ¦ broken line P in Figure 8. This is because even an increase in current does not cause a voltage drop across an impedance.
Under these circumstance, the glow discharge has l its operating point coinciding with a point Pl where the lS ¦ characteristic P of the source side intersects the charac-teristic N of the glow discharge. However, this operating point Pl is located in the region of abnormal glow g, which is apt to transit to a region of arc discharge h, as above l described. Accordingly it is difficult to maintain the ¦ glow discharge stable in the region of abnormal glow g.
Further it is to be noted that the flat character-istic P of tlle source side can not stably cross the flat characteristic portion N2 of the glow discharge in the ~;
region of normal glow f.
On the other hand, with a resistance R as the impedance connected to the source, an increase in current I
causes an inc~r~ase in voltage drop IR across the resistance.
Thus the source side has the current-to voltage characteristic ~ 30 such as shown at dotted straight line Q in Figure 8 and the glow discharge has its operating point designated by an intersection Ql of the characteristics Q and N. This operating point is located in the region of normal glow f resulting in the stable glow discharge.
Where electrical energy participating in the glow discharge is converted to thermal energy with a very high efficiency, the connection of an impedance to the source as above described forms one of factors of decrcasing an efficiency utilization of electrical energy. For example, the use of a resistor causes a Joules loss and the use of reactor causes a Joules loss of a win~ing involved and an eddy current loss and a hystersis loss of an iron core involved. Since such energy losses scatter as thermal energy, it is possible to recover the thermal energy.
This, of course, deprives the resulting heating device of its convenience and compactness.
From the foregoing it is seen that whether or not an impedance is connected to an electric source retains the abovementioned disadvantages as long as the glow discharge has the current-to-voltage characteristic in the form of a curve such as shown at N in Figure 8.
In order that the glow discharge can be maintained stable even with the flat current-to voltage characteristic of an associated source side such as shown by straight line P in Figure 8 and without an impedance connected to the source, the present invention includes unique means for imparting a positive resistance to the current-to-voltage ~ 30 I
~, I . ..
characteristic of the glow discharge in a different m~nner as compared with conventional abnormal glows.
First it is seen in Figure lOA that surface of l a cathode electrode 1 opposite to an anode electrode 2 has an area made sufficiently larger than that of the anode electrode 2 so as not to impede the spread of a negative glow A b. In other words, the opposite surface area of the anode electrode 2 is limited to a small magnitude with respect to the cathode electrode. ~hus a peripheral edge root bl of the negative glow b lying on the opposite surface of the cathode electrode 1 has a distance to the anode electrode 2 is gradually increased as the negative glow b spreads due to an increase in glow discharge current and therefore a voltage across both electrodes 1 and 2 is gradually raised. Under these circumstances, the glow discharge has the current-to-voltage characteristic such that the voltage increases with the current as shown at broken curve T in Figure 8.
That is, the characteristic is of the positive resistance type.
In this connection, it is to be noted that the positive resistance characteristic developed in the region of abnormal glow g in the prior art practice as shown at ~
curve N3 in Pigure 8 is cause from the fact that the negative glow b has spread over the surface of the cathode electrode and can not any more spread (see Figure 9B~.
Accordingly, such positive resistance characteristic is quite different from that according to the principles of 3 0 ¦~

¦ the present invention. As above described, the negative glow of the present invention is permitted to sufficiently ¦ spread as an increase in current because the active surface l area of the cathode electrode l opposite to the anode S ¦ electrode 2 is sufficiently larger than that of the anode electrode with the result that there is no problem that the ¦ glow discharge is transits to an arc discharge due to the impossibility of spreading the negative glow.
l From the foregoing it is seen that the characteristic ¦ T of the present invention as shown in Figure 8 is developed in the region of normal glow but not in the region of ¦ abnormal glow although it has a positive resistance.
In the present invention, even with an associated l electric source having no impedance connected thereto, ¦ therefore the characteristic T thereof intersects the characteristic of an associated source side at a point Tl ¦ (see Figure 8) where the glow discharge is stablized. It ¦ is to be noted that the point Tl lies in the region of the ¦ normal glow unlike the characteristic N3 of the prior art l practice so that the pres~nt invention does not encounter the problems that the glow discharge transits to an are discharge and so on.
¦ In order to impart a positive resistance to the ¦ current-to-voltage characteristic of the glow discharge ¦ by further increasing the distance between the peripheral ¦ edge bl of the negative glow b on a cathode electrode l and ¦ an associated anode clectrode 2, the cathode electrode l I can be made cylindrical and opposite to the anode electrode 1114~3~
..

2 as shown in Figure lOB. In the arrangement of Figure lOB, the peripheral glow edge bl is located on the peripheral wall surface of the cylindrical cathode electrode 1 at some distance from the end surface thereof. Thus, the g]ow edge ~1 is far spaced away from the anode electrode 2 as compared ~ith the arrangement of Figure lOA, resulting in a satisfactory positive resistance characteristic.
When an AC voltage is applied across the cathode and anode electrodes, either of the electrode becomes alternately a positive electrode so that a glow discharge is caused on the opposite surfaces of both electrodes. With an AC
voltage used, it is desirable that the cathode and anode electrodes are in the form of identical cylinders and oppose to each other as shown in Figure lOC. From Figure lOC it is seen that the peripheral edge bl of the negative glow b on either electrode 1 or 2 is far spaced away from the mating electrode 2 or 1 as in the arrangement of Figure lOB.
From the foregoing it is summarized that the principles of the present invention are to cause an area with which a pair of cathode and anode electrodes are opposite to each other to be smaller than that area of the electrode with which a negative glow is caused.
Referring now to Figure 11, there is illustrated one embodiment according to the glow discharge heating apparatus embodying the principles of the present invention as above described. The arrangement illustrated comprises an electrically insulating enclosure 9 in the form of a -hollow cylinder formed of glass, a cathode electrode 1 in the I ~114~13~
I ., form of a hollow cylinder with both open ends coaxially extending through the enclosure 9 and an anode electrode 2 in the form of a hollow cylinder with both open flare en~s disposed coaxially with the cathode electrode 1 withing ~-the enclosure 9 to form an annular glow discharge gap 8 therebetween. The cathode electrode 1 is extended and sealed through both ends of the enclosure 9 by means of seal fitting 10 and 11 respectively. Thus the enclosure 9 along with the cathode electrode 1 defines an annular space 81 which includes the glow discharge gap 8 and is filled with an electrically dischargeable gas selected from the group consisting of rare gases such as helium, mixtures thereof, for example, a mixture of neon and argon, a mixture of helium and hydrogen etc.
An annular anode terminal 5 is fixedly secured at the inner periphery to the central portion of the outer cylindrical surface of the anode electrode 2 and has a protrusion extended and sealed through the enclosure 9 by having the outer periphery fixed to a seal fitting sealed to adjacent ends of two similar enclosure portions forming the enclosure 9. The anode terminal 5 is connected to a positive side of a source of DC voltage 3 including a negative side connected by a stabilizer 4 to a cathode terminal 6 that is connected to that portion of the cathode electrode 1 dispose~-sutside of the enclosure 9, in this case, adjacent to the seal fitting 11. The stabilizer 4 may be a small capacity reactor or a resistor. If desired, the stabilizer may be omitted.
. .

14~330 ¦¦ In order t~ lacilitate the descript~on of the ¦ present invention, the symbol "S-" designates the entire ¦ area of that portion of the cathode electrode 1 on wh;ch a ¦ glow discharge can be caused whlle the symbol "S+" designates ¦ an area of that portion of the anode electr~de 2 opposing to ¦ the cathode electrode 1 and actually used for the glow dis-¦ charge. Therefore an area labelled S+ is called an "anode ¦ area effective for discharge" or an "effective anode area".
l According to the principles of the present invention ¦ as above described, the anode area S+ effective for discharge ¦ is made smaller than thè cathode discharge area S-.
¦ The operation of the arrangement as shown in Figure 11 will now be described. A DC voltage from the source l ~ is applied across the anode and cathode electrodes 2 and ¦ 1 respectively through the stabilizer 4 to establish a stable glow discharge in the annular discharge gas 8 thereby to heat the cathode electrode 1. Under these circumstances, a fluid to be heated, for example, water flows into the l interior of the cathode electrode 1 as shown at the arrow A
in Figure 11 to absorb heat from ~he cathode electrode 1 to be heated. Then the heated fluid flows out from the cathode ~lectrode 1 as shown at the arrow B in Figure 1.
During the glow discharge; a current and a voltage thereof is illustrated by a characteristic curve shown in Figure 12 wherein the glow discharge current Ig in amperes is plotted in abscissa against the glow discharge voltage ~g in volts in ordinate. The glow discharge voltage Vg may be approximately expressed by ~ 111~ 3~1 Vg = VO + IgR

where VO designates a glow discharge-hold minimum voltage as will be described hereinafter and R designates a slope of the characteristic curve. The slope of the characteristic curve as shown in Figure 12 is called a "positive resistance R".
Referring back to Figure 11, L designates an axial iength of the anode electrode 12 and has been differently changed to vary the effective anode area S+ thereby to obtain the relationship between a ratio of the effective anode area S+ to the cathode discharge area S- and the positive resistance R as sl~own in Figure 13.
In Figure 13, the positive resistance R in ohms is plotted in ordinate against the ratio between both area S+/S- in abscissa. Curves labelled 122, 123 and 124 have been plotted with data measured by filling the interior of the enclosure 9 or the annular space 81 with a gaseous mixture including 70% by volume of helium and 30~ by volume of hydrogen under pressures of 100, 150 and 200 Toors respectively. The gap between both electrodes 1~ and 2 has ~een maintained at a magnitude of 1 mm. Also the vertical segment has the same meaning as that shown in Figure 4.
From Figure 13 it is seen that the positive resist-znce R at the ratio of S+/S- of 0.2 increases to four or four times that at the ratio of 1.
The tendency of the positive resistance characteristic as shown in Figure 13 can be observed with the spacing of ~ 30 I

5 mm between the both electrodes 1 and 2 filled with the dischargeable gas including neon, helium, a mixture of neon and argon, or a mixture of helium and at most 30% by volume of hydrogen under a pressure of 200 Torrs or less.
Also experiments have been conducted with the DC
source 3 having varied regulations of the source voltage.
The results of the experiments are shown in Figure 14 wherein the axis of ordinates represents a regulation of source voltage in percent and the axis of abscissas represents a ratio of the actual discharge current I to a rated discharge current Io in percent. Straight lines labelled 125 and 126 describe the regulations of source voltage with the positive resistance R having value of 1 and 3 ohms respectively.
From Figure 14 it is seen that a variation of 15%
in source voltage gives a current regulation or a ratio of the actual current I to the rated discharge current Io multipled by one hundred in percent ~42~ and +14% with positive resistance R of 1 and 3 ohms respectively. Thus the positive resistance R of 3 ohms renders the glow ? discharge relatively stable.
Further by rendering the positive value R higher, it is possible to control a maximum current for supplying a predetermined~electric power to a small magnitude which is, in turn, advantageous in that the glow discharge apparatus is made compact.
The measure as above described is also applied to constructions in which the AC voltage is applied across the electrodes 1 and 2 to cause the glow discharge thereacross only when the electrode 1 acts as a cathode electrode.
From the foregoing it is seen that the arrangement of Figure 11 eliminates the disadvantages of conventiona]
l glow discharge heating apparatus that the positive resistance ¦ for the glow discharge is low, the glow discharge is moved about on the electrode, the discharge current much changes with a variation in source voltage resulting in the necessity of connecting a stabilizer or like to the source l and so on. Those disadvantages have been caused from the ¦ cathode area substantially equalling the anode area.
Figure 15 shows a modification of the present invention operatively associated with an AC source. The arrangement illustrated comprises an inner electrode 1 in l the form of a hollow cylinder having one end closed, an ¦ cuter electrode 2 in the form of a hollow cylinder having cne end open and disposed coaxially with the inner electrode ¦ 1 so that the closed end portion of the inner electrode 1 is inserted into the opened end portion of the outer electrode l 2 to form an annular discharge gap ~ therebetween.
¦ The inner electrode 1 is coaxially disposed within a tubular glass enclosure 9 to extend beyond both open ends thereof and the open er.d portion of the electrode 1 is rigidly fitted into an annular supporting disc 13 of any l suitable metallic material including an outer periphery ¦ connected to the adjacent end of the enclosure 9 through the ¦ seal fitting 11. The outer electrode 2 has the open end portion extending into the enclosure 9 and supported ~o another annular sup~orting disc 14 of the same material as 1~
~114~30 the disc 13 similarly connected to the other end of the enclosure 9 through another sela fitting 10. In this ~ay the enclosure 9 defines a hermetic space 81 with the supporting discs 13 and 14, the seal fittings 10 and 11, the inner electrode 1 and the outer electrode 2 having the other end closed.
Then a pair of terminals S and 6 is attached to the supporting discs 13 and 14 to connect both electrodes 1 and 2 to an AC source 3] therethrough.
An inflow tube 15 is coaxially disposed within the inner hollow electrode 1 to form an annular passageway therebetween. ~he tube 15 is maintained in place through a closing member 16 rigidly fitted into the open end of the inner electrode 1 and having the tube 15 extending there-through. The inner electrode 1 is provided on the open end portion with an outlet duct 17.
~n the other hand, the outer electrode 2 is double ~alled and provided on the closed end portion of the outer ~all with an inlet duct 18 and that portion thereof adjacent to the supporting disc 14 with an outlet duct 19 communicat-ing with the inlet duct 18 through an annular space defined by the inner and outer walls of the electrode 2. A liquid to be heated, for example, water enters the inlet duct 18 as shown at the arrow A in Figure 15 and thence to the annular space due to the double-walled structure of the outer electrode 2 after which it leave the outlet duct l9. Also water enters the inflow tube 15 as shown at the arrow C in Figure 15 and thence the annular space bet~een the inflow 11 1 ~3 ~ 0 tube 15 and the inner electrode 1. Then the water flows -out from the outlet duct 17 as shown at the arrow D in Figure lS.
It will readily be understood that the space 81 is filled with an easy dischargeable gas as above described in conjunction with Figure 11.
In operation an AC voltage across the source 31 is applied across both electrodes 1 and 2 to cause a glow cischarge mainly in the annular discharge gap 8.
As above described, the inner electrode 1 is i.nserted into the outer electrode 2 to overlap the latter.
This ensures that an area of that portion of one of the electrodes opposite to the other electrode is smaller than an electrode area with which a glow discharge can occur lS between the electrodes 1 and 2. This means that an anode area on the side of that electrode acting as an anode for the glow discharge is always limited. For example, with the dischaTgeable gas maintained under a pressure of about 200 Torrs and with the gap between both electrodes having a length not exceeding 5 millimeters, the limitation of the : anode area results in an indirect limitation of an associated negative glow region and therefore an increase in positive resistance fo~ the glow discharge. That is, the current-to-voltage characteristic of the glow discharge such as shown at curve Tl in Figure 8 has a larger slope whereby the AC
glow discharge can be maintained stable. Accordingly, a stable glow discharge can be sustained even with a high current under a high pressure without the glow discharge changed an arc dis~haree.
Under these circumstances, a either of the inner and outer electrodes 1 and 2 respectively is heated when it acts as the cathode electrode resulting in heating of both electrodes. Thus the fluid such as water flowing in contact with the electrodes is instantaneously heated and the heated f]uid leaves the outlet ducts 17 and 19.
Figure 16 is a characteristic curve illustrating the relationship between the area of one of the electrodes overlapping the other electrode and the positive resistance exhibited by the glow discharge. In Figure 16 the positive resistance R in ohms is plotted in ordinate against a ratio of the overlapping area to the entire area of the electrode acting as the cathode in abscessa. Curves labelled 127, 128 and 129 have been plotted Wit]l the discharge gap 8 having a length not exceeding 5 mm and ~illed with a mixture of helium and hydrogen under pressures of 100, 150 and 200 Torrs respectively. The vertical segment has the same meaning as that shown in Figure 4.
From Figure 16 it is seen that the smaller the overlapping area for both electrodes 1 and 2 the higher the positive resistance will be.
In the arrangement of Figure 15, the inner and outer electrodes 1 and 2 respectively are disposed in coaxial relationship but different in shape from each other. There-fore the curr~n~o-voltage characteristic of the glow discharge is different between the half-cycle of the source 31 having the inner electrode 1 acting as a cathode and that 1~ 3~1 having the outer electrode acting as an anode as shown in Figure 17. In Figure 17, the axis of ordinates repr~sents a discharge voltage V and the axis of abscissas represents a discharge current I. When the inner electrode 1 acts as the cathode, the discharge current I is forwardly and rearwardly changed along a straight line 130 shown in Figure 17 and has a maximum value of Il. In the next succeeding half-cycle the outer electrode 2 takes over the cathode and the current is forwardly and rearwardly changed along a straight line 131 shown in Figure 17. In the latter case, the current has the absolute maximum value I2 different from that of the current I2 flowing in the just preceding half cycle of the sourc-e 31. Both straight lines have the same absolute values of a voltage ~0 at a null current.
Thus the resulting characteristic become unsymmetric to permit a zero-phase sequence component of a current to flow through the AC source 31. This is objectionable to the source 31. Further the inner electrode 1 is free at one end but the outer electrode 2 includes no free end. This results in the occurrence of thermai strains in the outer electrode 2 during the glow discharge.
These objections can be eleminated by still another modification of the present invention shown in Figure 18.
In the arrangement illustrated, a first electrode 1 in the form of a hollow cylinder having one end closed with a flat disc opposes to a second electrode 2 identical to the first electrode to form a discharge gap 8 having a predetermined spacing or gap length of _ between the opposite closed end 11 1114~30 .

surfaces.
A flow confining tube 20 or 21 of the double wall type inserted into the second or first electrode 2 or 1 respectively includes a central tu~ular portion extending on the longitudinal axis of the mating electrode, a radially extended end wall to form ~ predetermined gap between the same and the internal closed end surface of the electrode and a peripheral wall extending in parallel to the internal peripheral surface of the latter to form also a predetermined annular gap therebetween. Each electrode 1 or 2 is provided on the open end portion with an outlet duct 18 or 17 communicating with the flow path formed therein while annular blind cover disc 23 or 22 is rigidly inserted into the annular gap between the peripheral surface of the electrode 1 or 2 and the outerwall of the tube 21 or 20 at the open end. The ~urpose of the flow confining tubes 20 or 21 is to cause a fluid to be heated to enter first the central tubular portion as shown at the arrow A or C in Figure 18 and flow along the internal surface of the mating electrodes at an increased speed to enhance the heat transfer between the fluid and the electrode and also to enable the fluid to be instantaneously heated. The heated fluid then flows out from the outlet duct 18 or 17 as shown at the arrow B or D in Figure 18.
Then the first and second electrodes 1 and 2 respectively are sungly fitted into individual supporting rings 14 and 13 which are hermetically connected to both ends of circular enclosure 9 through annular seal fittings 10 and 11 In this way both electrodes 1 and 2 are supported .

.. . . .
.
' ~ --., ¦ in cantilever manner to the supporting members 14 and 13 and the substantiall portions thereof are coa~ially disposed ¦ within the enclosure 9 to form the space 81 that is then I filled with a dischargeable gas such as previously described.
¦ As in the arrangement shown in Figure 11 or 15, I the AC source 31 is connected across the electrodes 1 and 2 ¦ through the terminals 6 and 5 connected thereto respectively.
¦ In the arrange~ent of ~igure 18 it is noted that ¦ those portions of both electrodes 1 and 2 superposing each ¦ cther as designated by the reference character 1 is made ¦ smaller in area than that portion of each electrode on ¦ ~hich the glow discharge occurs. In the example illustrated ¦ the glow discharge occurs on each of the electrodes 1 and 2 ¦ throughout the surface.
lS ¦ The arrangement of ~igure 18 is characterized in ¦ that the clectrodes 1 and 2 formed to be symmetric abut ¦ a-gainst each other with the predetermined gap 8 formed ¦ therebetween. This results in the symmetric glow discharge - I characteristic as ShOWII in Figure 19. In Figure 19 similar ~20 ¦ to Figure 17, the characteristics 132 and 133 are substantially symmetric and have respective discharge currents Il and I2 ¦ equal in the absolute value to each other.
Also, as the electrodes 1 and 2 supported in I contilever manner to the annular supporting discs 14 and 13 ¦ respectively, the electrodes are prevented from breaking due to thermal stains.
It will readily be understood that the gap 8 betwcen both electrodes 1 and 2 should be dimensioned so that the ll 1114~30 elec rodes are preven-ed from c OD tactin~ each other due to ¦ termal expansions thereof in operation.
¦ As in the arrangement of Figure 15, an AC ~oltage ¦ across the source 3l is applied across the electrodes 1 and S ¦ 2 to cause a glow discharge between the opposite surfaces ¦ thereof while a fluid to be heated enters the interiors ¦ of the electrodes 1 and 2 as shown at the arrows A and C
¦ in Figure 18. Then the fluid flows through spacing formed ¦ between each electrode and the flow confining tube 21 or ¦ 22 to be heated with heat generated on the electrode 1 or 2 ¦ due to the glow discharge. Thereafter the heated fluid ¦ flows out from each outlet duct 19 or 17.
¦ Figure 20 illustrates a modification of the arrange-¦ ment shown in Figure 18. As shown in vertical section in 1~ ¦ Figure 20A, the electrodes 1 and 2 of the identical structure o~poses to and somewhat offset each other to form a predetermined discharge gap 8 therebetween. As seen in side elevational views of Figures 20B and 20C, the electrodes l 1 and 2 are in the form of rectangular boxes and therefore l discharge surfaces thereof are rectangular and flat, Then each electrode is provided on the rear surface with a pair of inlet and outlet tubes.
In other respects, the arrangement is substantially l identical to that shown in Figure 18. The electrodes 1 and ¦ 2 include the discharge surfaces identical in shape to each ¦ other and are of the cantilever type so that the arrangement exhibits the same results as that shown in Figure 18.
In the arrangements of the present invention shown . .
- : ~
.~ -'': ~

14~30 in F ures 15, 18 and Z0 the electrode material and impurities such as metallic oxides included in the electrodes might be scattered in the discharge ~ap during the glow discharge and sticked to that surface portions of the S enclosure 9 facing the electrodes l and 2. ~his sticking of such metallic materials to the enclosure might lead to not only a danger that the seal fitting lO and ll are short-circuit with ea~h other through the sticked materials but also to a fear that, if the scattered impurities agai.n a.dhere to the electrodes that the glow discharge will have transited to an arc discharge.
The present invention a].so contemplates to eliminate t.he danger and fear as above described, by the provision of the arrangement shown in Figure 21. The arrangement lS i.llustrated is different from that shown in Figure 18 only in that in Figure 21 a pair of annular shields 24 and 25 one for each electrode are disposed to surround the mating electrodes and face at least the internal surface portions of the enclosure 9 by having flare ends thereof fixedly secured to the internal surface portions of the enclosure 9 respectively. Each shield 24 or 25 includes the substan-tial portion parallel to the associated electrode and ending short of the adjacent annular supporting disc 13 or l~. The shields 24 and 25 may be of an electrically insulating or conductive material. t In operation when the electrode material and the impurities are emitted from the electrode l or 2 and scattered in the discharge gap, they are sticked to t11at surface of 1~14~3(:) ~ !
each shield 24 or 25 facing the associated electrode and prevented from adhering to that inner surface portion of the enclosure 9 covered with the shield 24 or 25. Also l the shield i5 effective for preventing the scattered ¦ electrode material and impurities from again adhering to the associated electrode.
The arrangement illustrated in Figure 22 is different from that shown in Figure 21 only in that in l Figure 22 a pair of annular electrodes 26 and 27 are buried ¦ in the annular shields 24 and 25 formed of an electrically insulating material respectively. Then a suitable voltage is applied to the annular electrode 26 and 27 whereby the scattered metallic materials are apt to adhere to the l shields 24 and 25.
¦ Figure 23 shows another modification of the arrange-ment illustrated in Figure 21. In Figure 23 the electrodes I -l and 2 are in the form of hollow flat discs and disposed in opposite rela~ionship to form the discharge gap 8 having I a predetermined gap length of d therebetween.
¦ The seal fitting 10 in the form of a short hollow cylinder has one end fixedly secured to the peripheral portion of that surface of the electrode 1 remote from the electrode 2 and the other end in the form of a flange to an l enclosure portion 91 in the form of an annulus. Then an l annular shield disc 28 of electrically insulating material is located between the annular enclosure portion 91 and the peripheral portion of the electrode l b~ having a fitting perpendicular to the same and connected to the outer '\ ..
~, I
peripheral surface of the seal fitting 10. The sealing fitting 11, an enclosure portion 92 and a shield 29 identieal to the components 10, 91 and 28 respectively are l operatively coupled in the same manner to the electrode 2.
¦ A toroidal metallic enclosure portion 93 of double L-shaped cross section is hermetically connected tc the annular enclosure portions 9' and 92 to form a hermetically closed space 81 in the form of a toroid.
As shown in Figure 23, a feed ~ater tube 18 and a drain tube 19 project in spaced relationship from that surface of the electrode 1 remote from the electrode 2 and a pair of deflector or baffle pla~es 30 and 32 are disposed in the interior of the electrode 1 so as to direct a liquid to be heated toward the peripheral portion thereof and enter the fluid into the drain tube 19 after it has flowed along the heated surface of the electrode 1 to be heated. Also a feed water tube 18' and a drain tube 17 similarly project from the electrode 2 and a pair of baffle plates 33 and 34 are similarly disposed within the hollow electrode 2.
If desired, the shield 28 and 29 may be formed of any suitable metallic material. In the latter case, the shields 28 and 29 are suitably insulated from the associated electrodes 1 and 2 respectively.
Further the present invention contemplates to prevent the occurrence of electric shock-accidents through the heated liquid such as water.
The arrangement illustrated in Figure 24 is substan-tially similar to that shown in Figure 22 except for the 1114~)30 '~ ..
¦ provision of means for preventing the user from receiving ¦ electric shocks. As shown in Figure 24, the control tnbular ¦ portion of the flow confining tube 20 or 21 is connected ¦ to an electrically insulating tube 37 or 38 that is, in ¦ turn, connected to metallic inflow tube 41 or 42.
¦ The outlet of the flow confining tube 20 or 21 is connected to connecting tube 35 or 36 subsequently ¦ connected to an electrically insulating tube 39 or 40 that is, in turn, connected to a metallic outflow tube 43 or 44.
¦ The metallic tubes 41 and 43 are electrically connected together to ground as do the metallic tube 42 and 44.
¦ It has been found that an end-to-end distance lp l between the central tubular portion of the flow confining ¦ tube and the inflow tube or between the connecting tube and the outflow tube, that is to say, a length of the insulating portion should be equal to or less than a predetermined magnitude dependent upon a voltage applied across the electrodes, a resistivity of the particular liquid to be heated, a cross sectional area of the tube etc.
The arrangement of Figure 24 is operated as follows:
A switch 45 is closed to apply an AC voltage from the source 31 across the electrodes 1 and 2. This causes the flow confining tubes 2~ and 21, and the connecting tubes 35 and 36 to be put at a certain potential relative to the ground potential. For e~ample~ in glow discharge heating apparatus having a discharge input of about 8 kilowatts, the AC source 31 is required to supply to the heating apparatus an AC

1~14~30 voltage having the effective valué of 200 volts so that the tubes 20, 21, 35 and 36 are put at a voltage having the effective value of 200 volts.
On the other hand, the meta~lic inflow tuhcs 41 and 42 and the mct~llic outf]ow tubes 43 and 44 are connected to ground so that the particular liquid flowing into or out from the extremities thereof is put at a null potential. This ensurcs that electric shock accidents are prevented from occurring throu~h the liquid.
~1Ore specifically, the source voltage is applied across the electrodes 1 and 2 to cause a glow discharge therebetween. Ileat generated during the glow discharge heats the liquid. When the heated liquid flow within the apparatus, the same reaches any of the tubes 41, 42, 43 and 44 where it is put at the ground potential. This ensures that the user is maintained safe.
Under these circumstances the electrodes 1 and 2 rapidly transfers heat to the liquid flowing within the interiors thereof to prevent the electrodes 1 and 2 from effecting an abnormal temperature rise whereby the stable glow discharge is sustained.
llowever, as a potential difference having the effective value of 200 volts occurs between the inflow and outflow tubes 41, 42 and 43, 44 and the confining and connecting t~bes 20, 21 and 35, 36, the insulating tubes 37, 38, 39 and 40 must have a dielectric strength withstanding a voltage having the effective evalue of 200 volts. In this connection, it is required to consider a leakage .

~` ~14~30 current f]owin~ to sround through the li~uid, in a~dition to the surface status of the insulating tubes.
In the arrangement of ~igure 24 applied to a water warmer operated with the source voltagc of 200 volts, the same is obli~ated to be provided with a leakage breaker.
Leakage breakers are responsive to the leakage current in excess of the predetermined rnagnitude flowing through the inflow and outflow tubes 41, 42 and 43, 44 to ground to be continuously operated to prevent the source voltage from being applied across the electrodes 1 and 2. Accordingly, it is required to impart to the length Qp of the insulating portion a value sufficient to limit the leakage current to a certain value or less.
Assuming that each of the insulating tubes 37, 38, 39 and 40 has a cross sectional area of flow path designated by S and a ligned to be heated such as water has a resistivity designated by p, the insulating portion presents a resistance Q before the liquid expressed by RQ = p SP ~4~

; Also assuming that each of the insul~ting tubes 37, 38, 39 and 40 has a surface resistance sufficiently large as .
compared with the resistance of the liquid, the leakage current IQ may be expressed by I = RQ = VQ ~ = p - Qp ~ 111~)30 where VQ designates a voltage across tlle liquid located in the insu]ating portion having the length of Q~. Accordingly, the leakag~ current IQ is inversely proportional to the length Qp with the voltage VQ, the cross sectional area S and the S resistivity p remaining unchanged.
Figure 25 a graph illustrating the relationship between the length Q~ of thc insulating portion and the leakage current IQ on the basis of the above two expressions

(4) and (5) and with VQ = 200 volts, S = 0.636 square centimeters (which results from the insulating tubes 37, 38, 39 and 40 havin~ the inside diameter of 9 millimeters) and p = 1300 ohms-centimeter. The resistivity of 1300 ohms-centimeter is a mini~lum value of a resistivity of usable water as determined by the IEC standards. In Figure 25 the leakage current I~ in milliamperes is plotted in ordinate against the length Qp of the insulating portion in centi-meters in abscissa.
Assuming that the particular water warmer is provided with a highly sensitive ]eakage breaker having a rated sensible current of 15 milliamperes, the breaker has a rated inoperative current of 7.5 milliamperes. In order to prevent this leakage breaker from being continuously opeTated due to a leakage current llowing through the insulating portion, the length _~ of the latter is necessarily of at least 13 centimeters with used water having a resist-ivity of 1,300 ohms centimeter as will be seen from the curve of Figure 25. The expression (5) indicates that the length Q~ changes with the leakage current, voltage, the :
. :

~- ~14~3~

cross sectional area of the flow path and resistivity of the liquid. Ilowever, a length of the particular insulating portion can be estimated as above described and in accordance with the rating of a given ]eakage breaker, the source voltage, a resistivity of the particular li~uid and the cross sectional area of the flow path.
In the arrangement of Figure 24, the flow path of the heated liquid has been provided with the insulating tubes having the required length while each of the insulating tubes has been connected at the extremity to the metallic inflow or outflow tube that is connected to ground. Accord ingly, it is ensured that the any electric shock accident can be prevented from occurring through a liquid involved and still one can eliminate the insulating treatment that electrode components are coated with an electrically insulating material. This results in simplified inexpen-sive apparatus and also the heat transfer from the electrode components to the liquid being rapidly effected. Therefore the arrangement of Figure 24 is extremely advantageous in both the heat effeciency and the stability of operation.
Also glow discharge heating apparatus such as shown in Figure 24 can be utilized to instantaneously heat a liquid, for example, water by flowing the water in a flow rate of from 1 to 10 litres per minute t}lrough the intcrior of the electrodes thereby to transfer thermal energy injected into the electrodes to the water. Under these circumstances, water at room temperature must be heated to a temperature of about 80C. This results in the necessity ~ 3 O

of injecting thermal energy of at least 5 kilowatts into the electro~es. This means that, with a power source of AC 200 volts used, the effcctive current of at least 25 ampercs must flow through tl~e clectrodes. If a discharge current becomes high and also if the discharge gap is filled with a gas under an increasing pressure then it is difficult to sustain the flow discharge. For example, the glow dis-charge transits to an arc discharge.
It has becn found that the stable maintenance of the glow discharge is affected by the type of gas filling the discharge space. Also it has been experimentally confirmed that, by filling the discharge space with a mixtuTe of at least helium and hydrogen, the glow discharge can be sustained without the transit to an arc discharge, even with an electric power required for l~eating the p~rticular liquid, that is to say, a discharge current as high as possible.
This will now be described in conjunction with Figure 24. Various experiments were been conducted with the discharge space 81 filled with an inert gas heavier than argon under a pressure ranging from S0 to 200 Torrs. The result of experiments indicates that the glow discharge is difficult to spread and that an increase in glow current causes a contraction of a positive column included in the glow discharge to move the glow discharge about on the electrodes 1 and 2. Thus the glow discharge is put -in its unstable state so that it is apt to transit to an arc discharge. The mean value of the glow current in excess of .' - : . . -..

5 amperes has caused the glow discharge to transit to an arc discharge.
With neon uscd, relatively stable glow discharge has occurred under a gas pressure not higher than 70 Torrs.
Under a gas pressure of 100 Torrs, however, the glow discharge has been relativily stable at the deerctive current up to about 20 amperes. Upon the effective current exceeding 20 ampere, the positive column has been contracted. This migllt cause the glow discharge to transit to an arc discharge.
Further, when an inert gas used has been heavier ; than neon, the scatter from the electrodes 1 and 2 has increased in amounts with the result that the electrodes 1 and 2 are violently consumed while insulating materials such as glass forming the enclosure 9 is sharply deteriorated in electrical insulation because metallic materials scattered from the electrodes 1 and 2 are sticked thereto. As a result, the use~ul life of the glow discharge heating appara-tus has been much reduced.
From the foregoin~ it is summerize~ that, with the arrangement of Figure 24 used as a heating apparatus for instantaneously heating water, it is required to sustain stably the glow discharge under a relatively high pressure of 50 Torrs or more and still at a high current exceding 25 amperes at an AC voltage of 200 volts.
Also from the foregoing it has been found that it is desirable to fill the discharge space 81 with a chemically stab~e, light inert gas and suitable examples of the inert gas involvè helium and hydrogen.

~146~30 In the arrangement of Figure 24, however, it has been seen that, with helium filling the discharge space 81, the flow discharge spreads throughout the surface of the ¦ elcctrodes 1 and 2 at low current because of a small current ¦ density and that electrical energy of the glow discharge entering the electrodes 1 and 2 a~ounts only to about 2 Killowatts. Also in a glow discharge caused in an atmosphcre of helium, its positive cloumn has been con~racted upon a l pressure of helium increasing to 150 I`orrs to increase a ¦ current density for the glow discharge. Thus the glow ¦ discharge has been moved about on the electrodes and become unstable. lhe glow discharge has often transit to an arc discharge.
l On the other hand~ a glow ~ischarge in an atmosphere 1 of hydrogen has made a discharge hold minumum voltage Vo equal to at least 240 volts as shown in Figures 30, 31 and 32 which will be described hereinafter. Therefores it has been difficult to cause a glow discharge having an electric power of 5 kilowatts or more by using an AC source with 200 volts.
It has been found that, in order to manufacture glow discharge heating apparatus requiring at least S kilowatts with an AC voltage of 200 volts, it is optimum to employ a mixture of helium (lle) and hydrogen (~l2) as a filling gas.
When the arrangement of Figure 24 is filled with a mixture of helium and hydrogen under a pressure of 100 Torrs, and applied with an AC voltage of 60 hertzs having a waveform E shown in Figure 26, a glow current flowing ~ )30 .

therethrough is changed in accordance with a proportion of hydrogen to l~elium as shown at current waveorms F, G, H
and I in Figure 26. Figure 26 shows the voltage and current waveforms in one cycle of the source voltage. The current waveforms F, G, ~I and I have been plotted with gaseous mixture including S, 10 30 and 50 gd by volume of hydrogen and the balance, helium respectively.
Also the glow discharge exhibits the current-to-voltage characteristic dependent upon the proportion of the hydrogen ~o the helium as shown in Figure 27 wherein a voltage in volts is plotted in ordinate a~ainst a current I
in amperes in abscissa and like reference characters have been employed to identify the helium-llydroge mixtures identical to those designated in ~igure 26. As shown in Figure 27, each of the current-to-voltage characteristics is substantially rectilinear. By calculating both values of glow voltages S, T, U and W through the extrapolation and slopes of respective characteristic curves, the glow voltage Vg may be approximately expressed by Vg = VO + RI

where VO designates a glow discharge hold minimum voltage designated by S, T, U or W, and R designates the slope of the characteristic ca]led the positive characteristic R.
As well known, the voltage VO is expressed by VO = Em sin ~t where Em designates the peak value thereof and ~ designates an angular frequency of the source voltage. To calculate a ~ 30 ¦ discharge power P form the above expression for VO referring ¦ to Fi~ure 26 gives ¦ P = TR ¦ (Em sin wt - Vo) Em sin wt dt t TR ~ 4 TEm Em sin-lVo 1 ~:' I
l Vo cOs(sin 1 E-m) ~

where T designates a period of the source voltage. The discharge voltage is thermal energy entering the electrodes , 1 and 2 due to the glow discharge.
Assuming that the source voltage has its fre~uency of 60 hertzs and 200 volts or the peak value of Em = ~ 200 ~ 280 volts, its period is of 16.67 milli-seconds and its angular frequency is of 377 radius per second. By using those figures in the expression for the discharge power, the glow discharge hold minimum vcltage VO relates to the positive characteristic R as shown in Figure 28 wherein the positive resistance R in ohms is plotted in ordinate against the glow hold minimum voltage VO in volts in abscissa with the parameter being the discharge power or tllermal ellergy P.

~14 :P30 From the Figure 28 it is seen that, in order to provide the thermal energy not less than 5 kilowatts, the VO and R may lic in a hatched rc~ion as shown in Figure 28 defin~ by a ~ e for thc ]~ower of 5 kilow~tt~, ~nd both coordinate axes.
Also the glow hold minimum voltage VO is determined by a pressure of ~ filling gas and the gap length d between the electrodes 1 and 2 whilA the positive character-istic R is determined by ~he configuration of the electrodes of the overlapping area SO for both electrodes 1 and 2 and the pressure of the filling gas.
By changing a relative diameter M of one to the other of the electrodes 1 and 2 to vary the overlapping area SO therefor and also by changir.g the pressure of the filling gas, the positive characteristic R is varied as shewn in Figure 29 wherein the overlapping area SO in square centimeters is plotted in ordinate against the pressure of the filling gas in rorrs in abscissa with the positive characteristic R variously changed. In Figure 29 solid line indicates measured values and dotted line indicates values estimated from the associated measured values.
From Figure 29 it is seen that, under a gas pressure less than 50 Torrs, a curre]lt density for the glow discharge is low and the supply of a discharge power or a heat input in excess of 5 kilowatts to the electrodes requires an increase in overlapping area S. ~his has encountered the problem in the portability because the electrode area must increases.

~ )30 :

¦ nn the otller h~lnd, a gas pressure in excess of 150 ¦ Torrs causcs the discllar~e input to the electrodes to increase ¦ to at lcast 5 ki]owatts, resultin~ in a glow current of at ¦ least 25 amperes. l3nder these circumstances a positive ¦ column involved is contracted and the particular glow ¦ discharge is moved about on the electrodes. ~his might ¦ sometimes cause the glow disc]lar~s~e to transit to an arc ¦ discharge.
l With the gas l~ressure furtller increased to 200 Torrs ]0 or hi~her, a r)ositive column involved is contracted at a glow current of at least 5 amperes un~il the transit to an arc discharge occurs.
As an examp]e, it is assumed that the glow hold mil)imum voltage VO is impossible to decrease to 176 volts or less. Under the assumed condition, it is seen from Figure 28 that, in order to manufacture glow discharge heating ap~aratus having a discharge input of at least 5 kilowatts, the prcssure of the palticu]ar filling gas, the overla~ping ~rea ~O alld tllc l)ositivc ~l~aractet-istic l~
must lie in the hatched portion shown in Figure 29 as being defined by a pair of vertical broken lines pass;ng through the abscissas of 50 and 150 Torrs respectively and curve labelled R = 2Q
In addition, by changing both the proportion of hydrogen to helium and the gap ]ength d between the electrodes 1 and 2, the ~low hold minimum voltage VO is varied as shown in ~igures 30, 31 and 32 wherein the axis of ordinates represents the pro~ortion of hydrogen to helium in percent ~ 30 and the axis of abscissas rc~r~ent~ the ~ap leng~h d in millimetcrs. Tlle llelillm-hydrogcll mixtl3re is maintained under pressures of 5(), l0n an(] 150 'lorrs in Figurcs 30, 31 and 32 rcsl)e~tivcly. 1ll the~ res curvcs are labcllcd S measured vall~es of the glow hold minimum voltage Vc and for pure hydrogen the measured voltages VO are denoted aside correspondillg dots.
Also the gap length d ]ess than about 0.5 millimeter between both electrodes 1 and (2) has resulted in a danger that both electrodes may contact and shortcircuit each other due to a prcssllle difference betwccn a pressure of the particular heated liquid within either of the electrodes and that of a filling gas involved. On the other hand, an excessively large gap length d between both electrodes lS cause a positive column to constract to move the resulting discharge about on the electrodes until the discharge sometimes transits to an arc dischar~e. This might result in a cause for damaging the electrodes 1 and 2. It has been seen tllat the contraction of the positive column occurs with the gap length d of at least 9, 6 and 3 millimeters under the gas pressures of 50, 100 and 150 Torrs respectively.
With the proportion of hydrogen to he1ium decreased to 2.5% or less, the resulting glow discharge resembles that occurrillg in an atmosphere of pure helium. This has made it difficult to increase ~he discharge input to at least 5 kilowatts. Also as Figure 29 describes that it is difficult to decrease the positive characteristic R to at most 1, 0.5 and 0.3 ohms undcr gas pressures of 50, 1no and 150 Torrs ~1~4~ 0 respectively, it has bcen di~fi~ult to increase the discharge input to at least 5 kilowatts at the glow hold minimnm voltages VO of at ~cast 2]0, 2~0 and 24n volts under the gas pres~ures of 50, 100 al)d lS0 Torrs rcspectively as will readily be undcrstood from the graph shown in ~igure 28.
Further an increase in glow hold minimum voltage VO causes an increase in peak value of the glow current as shown in Figure 33 wherein the peak currcnt for the glow discharge in amperes is plotted in ordinate against the ~low hold minimum voltage VO in volts in abscissa. This has resulted in the disadval-ta~c that the resulting apparatlls should be made larger.
From the foregoing it will readily be understood that the proportion of hydrogcn to helium and gap length d between the electrodcs 1 and 2 are desirably located in dotted closcd areas shown in Figure 30, 31 and 32. ~Iore specifically, the I~roportion of~hydrogen is not less than 2.5% and the gap lcngtIl _ is not less than 0.5 millimeter while the voltage VO has valucs of 210, 2~0 and 240 volts dependent upon the pressure of the filling gas.
While the present ;nvention has been described in conjunction Wit]l an AC sourcc hav jnn a voltage of 200 volts it is to be understood that it is equally applicable to AC -sources having the voltage higher than that of 200 volts, for example, the voltage of 400 volts. In the latter case, the glow current may be low by using a helium-hydrogell mixture including not less than 50% by volume of hydro~en which is effective for increasing the glow hold minim-lm voltage VO

~1~4~30 .

shown at any of the points S~ T, U and W illustrated in Figure 27. This provides a stable glow discharge while being able to decrease the surface area of the electrodes 1 and 2. In addition, wiring leads may be fine. Therefore the resulting apparatus can be made compact.
Examples of the electrode material may involve copper, aluminum, nickel, pure ion, molybdenum, stainless steel, Kovar (Trade mark) etc. used with vacuum tubes or voltage regulator tubes. I~owever, copper is not suitable for use in the present invention because the copper has a high current density for the glow discharge to enhance the sputtering thereby to deteriorate seversely the insulation of associated insulators. Also aluminum is not suitable for used in the present invention because a glow discharge involved transits to an arc discharge with a current as low as one ampere. Therefore suitable examples of the electrode material involve nickel, pure i~ron, molybdenum, stainless steel and Kovar (Trade mark). The electrode used with the present invention has been formed of sheet nickel or stainless steel one millimeter thick.
From the foregoing it is seen that the filling of the discharge space 8 with a mixture including at least helium and hydrogen can eliminate Ihe transit of the glow to an arc discharge and the sputtering with a high discharge current. This gives the result that a stable glow discharge can be sustained. ~he reason for wilich the glow discharge can be prevented from transiting to an arc discharge is to remove oxides on the surface of the electrodes by the hydrogen :

11140.30 included in the filling gaseous mixture.
The use of the helium-hydrogen mixture is also advantageous in that, only by changing the proportion of the hydrogen to the helium, the g~ow hold minimum voltage can be selecte~ at will to control the discharge input to both electrodes involved as d~sired.
Figure 34 shows still another modification of the present invention. The arrangement illustrated is different from that shown in Figure 24 only in that in Figure 34 the opposite surfaces of the electrodes 1 and 2 are corrugated to increase the surface areas of the electrodes and an auxiliary electrode 46 is operatively associated with the gap 8 formed between the electrodes 1 and 2 as will be subsequently described.
In glow discharge heating apparatus having the discharge input of 5 kilowatts, for example, the diameter M of the electrodes 1 and 2 is re~uired to be of at least 80 millimeters and also that of the insulating enclosure 9 is necessarily of at least 100 millimeters. In other words, the larger the diameter of the electrodes the larger the enclosure 9 and therefore the seal fittings 10 and 11 will be. This is attended with the disadvantages that the compo-nents become excessively expcnsive and also a manufacturing cost is increased.
In addition, the opposite surfaces of the electrodes 1 and 2 are can be forced toward each other to be crowned in response to a difference between a pressure within discharge space 81 and a pressure of a hcated liquid within .

each electrode so that the bending of the electrodes increases to ~e proportional to the fourth power of the radium M/2 thereof. Accordingly, an increase in diameter of the electrodes may causes the electrodes 1 and 2 to contact and short circuit each other due to the crowning thereof.
To avoid this objection, the oposite surfaces of the electrodes 1 and 2 have a diametric section of corrugated shape to ir.crease areas of the opposite electrode surfaces with the diameter of the electrodes remaining unchanged.
In the arrangement of Figure 34 each electrode 1 or 2 has the diameter ~ of 52 millimeters and the area of 80 square centimeters of that surface thereof opposite to the other electrode 2 or 1.
As shown in Figure 34, the auxiliary electrode 46 is extended and sealed through the insulating enc],osure 9 so as to center the gap 8 formed between the opposite corrugated surfaces of the electrodes 1 and 2 and to be substantially contacted at the free end by the adjacent portion of the edge of the gap 8.
Then the Ac source 31 is conllected at one end to the electrode terminal 5 through a normally open switch 45 and at the other end directly to tlle electrode terminal 6. The auxiliary electrode 46 is connected to the electrode terminals .=

6 and 5 through respective resistors 47 and 48 and also by a resistor 4~ to one output of an auxiliary source circuit 50.
The auxiliary source circuit 50 includes the other output connected to the electrode terminal 5 and ~here~ore tlle switch 45 and is also connected to the switch 45 through 1114~;~0 another normally open switch 51 and to tl~ other end of the AC source 31. The operation of the abovementioned circuit configuration will be described hereinafter.
With the auxiliary electrode 46 operatively S associated with the discharge gap 8 as in the arrangement of Figure 34, the electrodes 1 and 2 are called hereinafter the "main electrodes" to be distinguished from the auxiliary ele.trode 46.
In the arrangement of Figure 34 a glow discharge is fired between the main electrodes 1 and 2 after which the glow discharge is smoothly spread on the corrugated surfaces la and lb respectively of the main electrodes 1 and 2. Under these circumstances, a high current can enter the opposite corrugated surfaces of the main electrodes 1 and 2 as compared with pairs of discharge electrodes including the opposite flat surfaces. Therefore, the discharge input to the electrodes increased while the voltage across the main electrodes remains unchanged.
As a result, the corrugated surface of the main electrodes permits a decrease in diameter thereof attended with a reduction in diameter of each of the insulating enclosure 9 and the seal fittings 10 and 11. Accordingly a manufacturing cost can be decreased, Also the corrugated surface of the main electrode is effective for preventing Z5 the crowning of the opposite surfaces thereof.
~he o~posite surface la of the main electrode 1 shown in Figure 35 inclu~es a plurality of grooves of rectangular cross section concentrically disposed at substan-111~30 .

tially equal intervals thereon.
Figure 36 shows a plurality of parallel grooves disposed at predetermined intervals on the discharge surface la of the main electrode l.
The discharge surf~ce la of the main electrode l shown in Fi~ure 37 includes a plurality of cylindrical depressions disposed in a predeter~ined pattern thereon.
In the arrangement shown in Figure 38, a pair of flow confining blocks generally designated by the reference numeral 200 and 210 respectively are of ~he same counstance-tion and disposed in place within the main electrodes 2 and 1 to form heating spaces or flow paths 2A and lA for a heated liquid therein respectively. The flow confining block 200 is formed of an electrically insulating material such as a synthetic resinous material and includes a feed water tube 201 and a drain tube 202 formed in parallel relationship on the exposed end surface thereof to be integral therewith and through openings 201a and 202a connected to the tubes 201 and 202 respectively. Then openings 201 and 202a open on that end surface thereof facing the inside of the gap forming surface of the main electrode 2 and a peripheral surface thereof respectively.
The tube 201 and the opening 20a interconnected serves as a feed water tube opening in the flow path 2A while the tube 202 and ~he opening 202a interconnected serves as a drain tube also opening in the flow path 2A.
The flow confining block 210 includes a feed water and a drain tube identical to those as above described in 111~0~0 ::
conjunction with the flow confiring block 200 and designated by like reference numeral identifying the corresponding components of the confining block 200 and added with the numeral 10. For example, the reference numeral 211 designates a feed water tube.
The flow confining blocks 200 and 210 have the exposed end portions screw threaded through the blind cover plate 22 and 23 fixed to the open end portions of the main electrodes 2 and 1 to be flush with the open ends thereof respectively.
In other respects, the arrangement is substantially identical to that shown in Figure 34 except for the omission of the insulating tubes 37, 38, 39 and 40 shown in Figure 34.
In the arrangement of Figure 38, the flow confining blocks 200 and 210 can be removed from the blind cover plates 22 and 23 respectively fo~r the purpose of inspecting or cleaning the internal surfaces of the main electrodes 2 ;-and 1. Therefore the heating efficiency can be always maintained high.
Figure 39 shows modi~ication of the arrangement shown in Figure 15 wherein the user is accessible to the heat transfer surfaces of the main electrodes as in the arrangement of Figure 38 and an auxiliary electrode 46 is operatively associated with the discharge gap 8. As shown in Figure 39 a flow confining tube 200 in the form of a hollow cylinder having both ends open is coaxially disposed within the main electrode 1 to form a flow path for a heated liquid therebetween. The cylindrical tube 200 is screw 3.114U30 ¦ threaded through a screw mcmber 200a rigidly fitted into the ¦ o~en end of the main electrocle 1.
¦ ~Simi]arly ~nother f]ow confining tube 210 in the ¦ form of a ho]low cylinder having one end closed i~ detachably ¦ connected to the maill electrode 2 at the outwardly folded ¦ end through a screw member 210a formed internally with the ¦ tube 21C to form an annular flow ~ath for the heated liquid therebetween.
¦ The flow confining tubes 2~0 and 210 are of an ¦ electrically insulating material such as a synthetic resinous ¦ material.
¦ As in .he arrangement of Figure 38, the flow ¦ confining blocks 200 and 210 can reaidly be removed from ¦ the main electrodes 1 and 2 respectively for purposes of ¦ inspection and cleaning.
¦ In other respects, the arrangement is substantially ¦ similar to that shown in Figure 1~ excep~ing that electric ¦ shock preventing means such as above described in conjunc- ~-l tion with Figure 24 are provided on the feed water and drain ¦ tubes 41, 42 and 43, 44 and the auxiliary electrode 46 is operatively coupled to the gap 8 formed be~ween the main opposite electrodes 1 and 2.
Figure 40 shows a different modification of the l present invention enabled to decrease the dimension of the ¦ electrically insulating enclosure and still increase the diameter of the main electrodes. In the arrangement illustrated a pair of main electrodes 1 and 2 identical to each other are horizontally disposed in opposite relationship I

1114~30 t~ f m a dischar~e gap 8 therebetween. Eacll of the main electrodes 1 or 2 is in the form of a hollo~ cylinder having one end closed and the other end portion lB or 2B reduced in diameter. rhe closed flat ends of both main electrodes 1 and 2 form therebetween the gap 8 having a width or a gap length of d and a dia1neter of M.
Each elcctrode 1 or 2 includes a shoulder connected to an electrically insulating enclosure 9a or 9b in the form of a narrow annulus through a first annular seal fitting lOa or lla. Thus the enclosures 9a or 9_ encircles the reduced diameter end portion lB or 2B of the main electrode 1 or 2.
Then a cylindrical metallic shell 9_ or 9d encircles in spaced relationship the a~j~cent main electrode 1 or 2 and includes a radially inward directed ~lange connected at one end to the enclosure 9a or 9_ through a second annular seal fitting lOb or llb. Both shells 9c and 9d have the other ends abutt;ng against and fixed ~ogether as by welding.
Thus the shells 9c and 9d and the main electrodes 1 and 2 form therebetween an annular discharge space 81 including the gap 8 with the enclosures 9a and 9 the seal fittings 10_ lOb lla and 11_.
The lind cover plate 22 or 2~ is rigidly fitted into the open end of the main electrode 1 or 2. A feed water tube 41 or 42 is extended and sealed through the blind cover plate 22 or ~3 and has an outlet opening substantially flush with the internal surface of the cover plate 22 or 23. Also a drain tube 43 or 44 is extended and sealed through the blind cover plate 22 or 23 and has an end ` 1114~30 .

¦ portion bcnt into an L in order to fill a heating s~ace lA
¦ or 2A formed of the interior of the main electrode 1 or 2 ¦ with a li~uid to be heated. The en~ of the L-sha~ tube 43 ¦ or 44 faces the uppermost portion of the internal surface S ¦ of the main electrode 1 or 2 with a distance Q maintained ¦ therebetween.
¦ Further, the auxiliary electrode 46 and an associated ¦ electric circuit are provide~ in the same manner as above ¦ described in conjunction with ~igure 34.
¦ The main electrodes 1 and 2 may be of any desired ¦ shape other than the cylindrical shape as above described.
As the main electrodes 1 and 2 are of the same structure, the o~eratioll will llOW be described in conjunction l of one of the electrodes, for example, the electrode 1.
A liquid to be heated entcrs the heating space 1 through the feed water tube 42 as sllown at the arrow A in Figure 40 until its liquid surface reaches a level at which the drain tube 44 opens while the liquid is heated by the main electrode 1. Thereafter the heated liquid is exhausted from the space lA through the drain tube 44 as shown at the arrow B in Figure 40. The outflow of the liquid causes a pressure loss across the drain tube 44 permitting the heated liquid charged in the heating space lA to have a pressure higher tl-an the atmospheric pressure. In keeping with this increase in pressure, the surface of the liquid within the heating space lA is forced to be gradually raised beyond the open ond of the drain tube 44 resulting a decrease in volume of a cavity existing in the heating . . . . ..

I 1~14i~,30 ' I .
I space lA.
¦ In this case, the smaller the diameter of the drain ~ tube 44 will be ~hich is accompanied by an increase in ¦ speed of the liquid flowing throu~h the drain tube 44. As ¦ a result, the open end of the drain tube 44 is less in ¦ pressure than the cavity within the heating space lA. This ¦ causes an increase in rate at which the drain tube 44 ¦ sucks up air left within the heatin~ space lA.
l It has been experimentally proved that the distance ¦ QO exceeding 10 millimeters causes the air phase in the heating space lA to be too far spaced from that portion ¦ of the liquid just flowing throu~h the open end of the drain tube 44. Therefore the heating space lA has been difficult to be sufficultly deaerated. This means that the distance QO is r,referably of at most 10 millimeters.
In othcr words, the distance QO is so dimensioned that, even though steam bubbles would be evolved from the liquid being heated within either of the heatin~ spaces lA
and 2A and reach the upper~ost ~ortion of thereof, they can be rapidly exhausted throuxh the drain tube 43 or 44.
After the air has been fully removed from either of the heating spaces lA and 2A as above desdribed, both spaces is entirely filled with the heated li~uid without the steam bubbles accumulated to form a cavity therein.
Otherwise a cavity not filled with tlle hcated liquid is formed within either of the main clectrodes 1 and 2 and therefore that portion thereof ~ontacted by and located adjacent to the cavity excessively rises in tcmperature , ' . . .

resulting in its failure.
The arrangement of Figure 40 is further advantageous in that the insultaing enc]osures decrease in dia~eter and l therefore are easily manufactured with low cost and ~ mechanically strong because the enclosures surround the reduced dia~eter portions of the main electrodes which are encircled by the metallic shells interconnected into a unitary structure to permit a region occupied by the insulating enclosures to be extre~ely decreased. Further the main electrodes are insulated from the shells through the insulating enclosures respectively. Accordingly, the resulting apparatus is easy to be manufactured, inexpensive and robust while having a long useful life.
In the arrangement shown in Figure 41, the insulating enclosure 9 in the form of a hollow cylinder having both ends open includes a pair of upper and lower apertured cover plates 13 and 14 respectively connected to both open ends thereof through annular seal fittings 10 and 11 respectively. A pair of hollow main electrodes 1 and 2 having one end open are vertically disposed in opposite parallel relationship within the enclosure 9 to be staggered longitudinally ¢f the enclosure and form a discharge gap 8 .
in a discharge space 81 defined by the enclosure 9, the seal fittings 10, ll and the cover plates 13 and 14. The main electrodes 1 and 2 have the other open ends fixedly fitted into apertures on the upper and lower cover p'ates 13 and 14 to be flush with the outer surfaces thereof respectively.

- :. : : ,...... . . - , - . . .. . . . . .
'' ~ ' ' ' '' ' . .
: . .. .
.

ll 1114~130 ~he main electrodes 1 and 2 have the o~en ends closed with blind cover plate 23 and 22 having central openings ¦ respectively. Then a L-shaped tube 44 or 41 has one leg I connected to the ccntral opening on the blind cover plate ¦ 23 or 22 and the other leg horizontally extended to form an outflow or an inflow tube.
A feed water tube 42 extends in sealing relationship through the one leg of the outflow tube 44 and into a l heating s~ace lA within the main electrode 1 from above the upper plate l~. Similarly, the drain tube 43 extends through the inflow tube 14 and into a heating space 2A within the main electrode 2 from below the lower plate 14.
As in the arrangement of Figure 40, the drain tube 43 has its open end facing the inside of the closed end of the main electrode 2 through a s~acing QO not greater than 10 millimeters.
As sh~on in Figure 41, the inflow tube 41 has the end cpening in the heating spacè 2A below the inlet of the drain tube 4~ while the feed liquic3 tube 42 has the end opening in the heating space lA below the inlet of the drain tube 44. Therefore the heating spaces lA and 2A can be entirely filled with the heated li~uid as in the arrangement of Figure 40. - .
Further an auxiliary electrode 46 is operatively associated with the discharge gap 8 formed between the main opposite electrodes 1 and 2. If clesired, both main electrodes may be concentrically disposed.
- In the arrangement shown in Figure 42, a seamless ~14~3~1 metallic tube is closely wound into a helix 41a or 42a having the outside diameter substantially equal to the inside diameter of the main electrode 2 or 1. The helix 41a or 42a includes one cnd portion 43 or 42 extending through the central hollow portion thcreof and the other end portion 41 or 44 bent into an L-shape. Both helices 41a and 42a are inserted into the main electrodes 2 and 1 to be brazed or welded to the intcrnal surfaces tllereot' res~ectively for the ~urpose of improving the heat transfer from the ~ating main electrodes thereto. A liquid to be heated enters the helix 41a or 42a through the end portion 41 or 42 and leaves the end portion 43 or 44.
In other respects, the arrangement is identical to that shown in ~igure 41.
Each of the main electrodes 1 or 2 can be prevented from corroding starting with those portions thereof brazed or welded to the helix 42a or 41a because the brazed or welded ~ortions are not directly contacted by the heated liquid flowing through the helix. ~Since the heated liquid flows at a high sl~eed through the helix 41a or 42a, the nuclear ebullition can be prevented and also a pressure loss in the helix is increased to preve!lt steam bubbles from staying in the helix. 'I`his results in the smooth heat transfer from the main electrode to the heated li~uid flowing through thc matjng hc]ix. Thlls thc main electrodes are prevented from excessively rising in surface tem]lerature thereby to sustain stably the glol~ discharge.
The arrangcmerlt shown in ~igure 43 is substantially .

1114~30 similar to that illustrated in Figure 40 ex.epting that, in addition to disposing the main electrodes 1 and 2 vertically, they are in the form o~ square hollow prisms and a tube is closely wound in helix complementary in shape to the interior S of the associated main electrode and fixed thereto.
Each of the arrangements shown in Figures 42 and 43 is characterized in that tube means formed of a good thermally conductive material contacts the internal surface of the mating main electrode to be thermally integral therewith and the heated liquid flows through the tube means. This results in the alleviation of limitations as to the configura-tion of the main electrode while -facilitating the manufacturing of the apparatus and prolonging tl-e useful life.
In the arrangement shown in Figure 44 either of the blind cover plates 22 and 23 is provide~ on that portion diametrically opposite to the normal outlet with an exhaust port that is, in turn, closed with a plug 221 or 231 for example through a screw meahcnism. Further an auxiliary electrode 46 is operatively coupled to the gap formed between the main opposite electrodes 1 and 2 as above described in conjunction with Figure 34.
In other respects, the arrangement is substantially identical to that shown in Figure 24.
The arrangement shown in Figure 45 includes the U-shaped flow path or heating space lA or 2A within the main electrode 1 or 2 and a connecting tube 361 or 351 connected to the heating space IA or lB on the inlet side. Then the connecting tube 361 or 351 is provided with an exhaust `1~ 11.14~30 I

¦ port closed with a detachable plug 231 or 221.
¦ In other respects the arrangement is substantially identical to that shown in ~i~ure 44.
l When each of the arrallgcmcnts shown in ~igures 44 S ¦ and 45 is desired to be out of service for a long time, the plugs 221 and 231 can ~e removed from the associated exhaust ports to drain the liquid out from interior of the main electrodes for the purpose of prcventing the liquid within the main electrode from s~oiling or freezing. Also the useful life can be prolonged.
While the main electrodes have been described as being in the form of hollow cylinders having the same shape and disposed in opposite relationship it is to be understood that the main electrode may be of any other desired shape.
For example, the main electrodes may be in the form of hcllow cylinders disposed in coaxial relationship. It is essential that, in order to empty the interior of the main electrodes, the exhaust port must be provided on the lower portions thereof.
While some of the abovementioned Figur`es, for example, Figure 34 illustrate the control circuit for controlling the glow discharges Figure 46 shows the fundamental circuit configuration of a control circuit for controlling -~
any of the arrangements as above described including no auxiliary electrode. In Figure 46, the arrangement generally designated by the reference numeral 100 comprises a pair of first and second electrodes 1 and 2 respectively disposed in opposite relationship to form therebetween a gap having a 1114S~;~0 ..

f gap length or a width d and each including an inflow and an outflow tube. Water enters the interior of either electrodes 1 a~d 2 throug]l the inflow tube to be heated and heated water ~cavcs it through the outflow tube.
The source of AC voltage 31 is connected across the electrodes 1 and 2 through a bidirectional triode thyristor 60 with the first electrode 1 connected to ground. The bidirectional triode thyristor is called hereinafter a "Triac" (grade mark). The source 3~ is also connected across a gate circuit 61 through a normally open switch 62.
Then the gate circuit is connected across one electrode and a gate electrode of the Triac 60. The switch 62 is closed to fire a glow discharge between the electrodes 1 and 2 there-by to heat a liquid, for example, water flowing through the interior of each electrode.
The operation of the control circuit shown in ; ~igure 46 will now be described with reference to Figure 47 wherein there are illustrated a voltage waveform V supplied from the source 31 and having a peak value Em and a current waveform V of the glow discharge. As shown in Figure 47, the voltage waveform V in the positive half-cycle of the source gradually increases from its null point until time point tl is reached. At that time voltage reaches a value of a discharge breakdown voltage V5 to fire a glow discharge between the electrodes 1 and 2. At that time point tl a glow current I abruptly flows throu~h the electrodes 1 and 2. The glow current I corresponds to a voltage drop expressed by Vf - VO where VO designates a glow hold minimum voltage 4~30 and ay be expressed by I = (Vf - Vo)/R where ~ designates I a discharge resistance corresponding to a slope of a -¦ current-to-voltage characteristic curve for a glow discharge ¦ as above described in conjunction with Figure 8.
¦ The at time point t2 the voltage V is equal to the ¦ glow hold minimu~l voltage VO after which the glow discharge ¦ is extinguishe~ because the voltage is less than the ¦ voltage VO.
¦ Thereafter the sourc3 31 enters the next succeeding ¦ negative half-cycle of the source in which the process as ¦ above described is repeated to cause a glow discharge between the electrodes 1 and 2. In the arrangement shcwn in Figure ¦ 46 the application of the AC voltage causes the electrodes ¦ l and 2 to act alternately as a cathode and an anode electrode 1 respectively to be heated because the glow discharge heats ¦ that electrode acting as the cathode as above described.
¦ From the foregoing it will ~e seen that, the firing ¦ of the glow discharge at time point tl causes an instantaneous I increase in glow current so that the glow discharge can not ~ spread following this increase in glow current. This results in the tendency to locally concentrate the glow current on the electrode to transit the glow discharge to an arc discharge. T~e arc charge has a fear that it melts the electrode which, in turn, reduces the useful life of the heating apparatus.
Also the glow surrent is initiated to flow through the electrodes l and 2 only upon the source voltage across both electrodes reaching the discharge breakdown voltage Vf ~ ~ 33~) while Vf > VO hilds. ~herefore it is impossible to utilize a time interval during which the source voltage is not less than the glow hold minimum voltages VO as a coduction time resulting in a poor efficiency o~ utilization of the source ~roltage. -Figure 48 shows a control circuit for controlling the glow discharge heating apparatus of the present invention constructed in accordance with the principles thereof. The arrangement illustrated comprises an auxiliary source circuit 61 connected across the source of AC voltage 31 that supplies hC voltage of 200 volts at the co~ercial frequency. The circuit 61 includes a normally open switch 62, a step-up transformer 63 having a primary winding connected across the source 31 through the switch 62 and a secondary winding having one end connected to the electrode 2 through a current limiting resistol 64 and the other end connected to theelectrode 1 and also to ground.
As in the arrangement of Figure 46, the source 31 is connected to the electrode 2 through the Triac 60. The ~20 resistor 64 is connected across a primary winding of an electrically insulating transformer 65 including a secondary winding connected across a pair of AC inputs of a rectifier bridg~ 66. The rectifier bridge 66 include a pair of DC
outputs one of which is connected to the junction of the source 31 and the Triac 60 through a resistor 67 and the other of which is connected to the remaining terminal or a : gate terminal of the Triac 60.
The step-up transformer 63 is designed and -- ~0 11.1~n30 i constructed so that the discharge breakdown voltage Vf is applied across the electrodes l and 2 before time pOillt ~here an instantaneous voltage from the source 31 reaches the glow discharge minimum voltage VO
The operation of the arrangement shown in Figure 48 will now be described with reference to Figure 49 similar to Figure 47. In Figure 49 wherein like reference characters designates the components corresponding to those shown in Figure 47, the switch 62 is closed at time point A to permit the source to apply the source voltage across the primary winding of the transformer 63. At a point B, a secondary or an output voltage from the transformer 63 reaches the discharge breakdown voltage Vf whereupon the gap between the electrodes 1 and 2 is broken down to start an electric discharge therebetween. At that time the output voltage drops to the glow hold minimum voltage VO (see point C, Figure 49) for a glow discharge by means of the current limiting resistor 64. This causes a current 1 on the order of 0.1 ampere to flow through the electrodes l and 2 resulting in a glow discharge occurring across the electrodes l and ~.
That glow discha~ge is called a "pilot glow discharge".
The current for the pilo~ glow discharge causes a voltage drop ,across the current limiting resistor 64 that, in turn, induces a secondary voltage across the transformer 65. The se~ondary voltage from the transformer 65 is applied ~o the gate electrode of the Triac 60 after having been full-wave rectified by the rectifier`bridge 67 to put the Triac 60 in its conducting state. Therefore the source voltage is .3~
, applied across the electrodes 1 and 2. Urder these circumstance, if the pilot glow discharge has not occurred across the electrodes l and 2 then the pilot glow current i coes not flow through the electrodes l and 2 and no voltage is induced across the insulating transformer 65 with the result that the Triac 60 is maintained non-conducting.
This ensures that the application of the high AC voltage across the electrodes 1 and 2 does not results in the GCcUrrence of an arc discharge therebetween unless the pilot glow discharge preliminarily occur across the electrodes 1 and 2.
When the source voltage is applied across the electrodes l and 2 through the now conducting Triac 60 and reaches the glow hold minimum,voltage VO' the principal glow discharge is fired across the electrodes l and 2.
That is, a current I for the principal glow discharge flows through the electrodes 1 and 2. That principal glow discharge current I is extinguished after the source voltage V has again reached the glow hold minimum voltage VO at point E
or time point t2 and therefore the principal glow discharge is extinguished. I~owever it is noted that at point E the voltage VO from the step-up transformer 63 is applied across the electrodes l and 2 through the resistor 64 with the result that the pilot glow discharge is still established.
Then at point F, the output voltage from the step-up transformer 63 becomes also less than the voltage VO to cease the pilot glow discharge.
Then the source 31 enters the next succeeding negative .~ ¦ half-cycle in which the process as above described is repeated.
The concept of the embodiment of the present invention ¦ as shown in Figure 48 is to apply preliminarily a high ¦ ~-oltage across the electrodes by means of the auxiliary ¦ source circuit to cause the preliminary or pilot glow ¦ cischarge thereacross and to smoothly derive the principal ¦ glow discharge from the pilot glow discharge. Therefore ¦ the arrangement of Figure 48 is effective for preventing ¦ the principal glow discharge current from abruptly increasing ¦ resulting an arc discharge as in the arrangement of Figure 46. Further the efficiency of utilization of the source is increased.
The arrangement shcwn in Figure 50 comprises a ¦ reactor 68 connected between the source 31 and the electrode ¦ 2, and an AC pulse generator 69 connected across the source 31 through the normally open switch 62. The pulse generator ~9 includes one output connected by the current limiting resistor 64 to the junction of the reactor 68 and the electrode 2 and the other output connected to the electrode ~20 1 and therefore to ground.
The gap formed between the electrodes 1 and 2 is so dimensioned that the peak voltage Em from the source 31 is prevented from effecting the discharge breakdown of the gap.
As shown in Figure 51 wherein a voltage and a current waveform V and I respectively and a pulse waveform P are :
illustrated, the AC pulse generator 69 generates an AC
pulse voltage P sufficient to reach the discharge breakdown voltage Vf at time point tl where the voltage from the . _ J.30 ¦ source 31 approximately reaches the glow hold minimum voltage ¦ VO The pulse voltage P first effects the discharge break-¦ aown of the gap between the electrodes 1 and 2 followed by ¦ a flow of the principal glow current I through the electrodes.
S ¦ As in the arrangement of Figure 46, the current I
¦ becomes null at time point t2to extinguish the glow discharge after which the process as above described is repeated in the next succeeding negative half-cycle.
l It is noted that the reactor 68 is designed and ¦ constructed so that it present a high impedance to the pulse waveform P but a low impedance to the commercial frequency of the source 31.
I Thus the arrangement of Figure 50 ensures that, ¦ ~ihen the source voltage V is close to the glow hold minimum ¦ voltage VO~ the principal glow discharge is initiated ketween the electrodes 1 and 2 and then the principal glow curr~nt I is smoothly increased without the transit to an arc discharge.
Figure 52 shows a modification of the present invention ;
wherein the pilot glow discharge occurs between the auxiliary electrode and either of the main electrodes prior to the principal glow discharge as above described, for example, in conjunction with Figure 34. In Figure 52, the main and auxiliary electrodes 1, 2 and 46 respectively are schematically shown and may have any of their structures shown in Figure 34 and Figures 38 through 45.
The arrangement illustrated comprises the AC source 31 and an auxiliary source shown as comprising a step-up - ~4 -.

`I 1 1 1 4~ 3 0 ¦ transformer 70 including a primary winding connected across the sourc~ 31 through the normally open switch 51 and a center-tapped secondary winding. The dot convention is l used to identify the polarity of the instantaneous voltage ¦ across the associated winding. The secondary winding includes center tap connected to the auxiliary electrode 46 through a current limiting resistor 71 and a normally open switch 72, and a pair of end terminals connected to the main l electrodes 1 and 2 through individual semiconductor rectifier ¦ diodes 73 and 74 with anode electrodes thereof connected to the main electrodes respectively. The gap formed between the electrodes 1 and Z has a distance or a gap length d satisfying Vf > Em > VO~ where Vf, ~m and VO have been l previously defined.
¦ The switch 51 is closed to apply the AC voltage V
from the AC source 31 across the electrodes 1 and 2 while ¦ the switch 72 is closed to apply a high voltage waveform ¦ from the step-up transformer 70 to the auxiliary electrode ¦ 46. Under these circumstances, when a potential at the main 2~ ¦ electrode 1 is higher than that at the main electrode Z, the ¦ diodes 73 and 74 are turned off and on respectively to cause l a pilot glow discharge between the auxiliary electrode 46 I ¦ acting as an anode and the main electrode 2 acting as a ¦ cathode. On the contrary, when the main electrode 2 is ¦ higher in potential than the main electrode l, the diodes ; ¦ 73 and 74 are turned on and off respectively to cause a pilot glow discharge between the auxiliary electrode 46 acting as anode and the main electrode 1 as the cathode.

41)30 1 j I addition, s the auxiliary electrode 46 has applied ¦ thereto the ~oltage from the center tap on the secondary ¦ transformer 70 winding, the voltage applied across the ¦ auxiliary electrode 46 and the main electrode 1 to cause ¦ the pilot glow discharge therebetween is quite identical to ¦ that applied across the auxiliary electrode 46 and main ¦ electrode 2 to cause ~he pilot discharge therebetween.
¦ Therefore, the transit of the pilot glow discharge- due to ¦ the auxiliary electrode to the principal glow discharge ¦ between the main electrodes 1 and 2 are equally effected ¦ between each of the positive half-cycles and the negative ¦ half-cycle of the source 31.
¦ Further the occurrence of the pilot glow discharge ¦ completes a closed circuit including the diode 73 or 74, ¦ the associated half of the secondary transformer 70 winding, -~
¦ the resistor 71, the closed switch 72 and the pilot glow ¦ discharge between the auxiliary electrode 46 and the main electrode 1 or 2. This prevents the current for the pilot ¦ g-low discharge from entering a circuit with the source 31.
¦ ~he opening of the switch 72 ceases the pilot glow I I discharge from occurring between the auxiliary electrode 46 l and either of the main electrodes 1 and 2. ~hus the : ¦ principal glo~ discharges are not fired in the next succeeding cycle of the source and the cycles following the latter with ¦ the result that the heating operation is not performed. In other words, the ON-OF~ control of the principal glow discharge ¦ can be conducted by turning the pilot discharge on and off.
l It is noted that the pilot glow discharge always 1~'14030 ~`
I
, ¦ occurs between the auxiliary electrode 46 acting as the ¦ ~node and either of the main electrodes 1 and 2 acting as ¦ the cathode so that the auxiliary electrode 46 is not heated.
I This results in the elimination of the necessity of cooling ¦ the auxiliary electrode.
¦ From the foregoing it is seen that, the arrangement ¦ when effecting the ON-OFF control of the heating apparatus ¦ proper of Figure 52 ensures the transit of the glow ¦ discharge by turning the pilot glow discharge on and off.
¦ The arrangement illustrated in Figure 53 is ¦ different from that shown in Figure 52 only in that in Figure 53 a zero-voltage firing circuit is provided to prevent the glow current from bruptly increasing. In Figure 53 a pair of serially connected resistors 75 and 76 ¦ are connected across the AC source 31 through the normally open switch Sl to form a voltage divider, and ~he junction A of both resistor is connected to a resistor 77 subsequently connected to a base resistor 78 that is connected to a base I source VBB. The resistor 76 is connected to ground. The `20 junction B of the resistors 77 and 78 is connected to a base electrode of an NPN transistor 79 including an emitter electrode connected to the reslstor 76 and a collector electrode connected to a DC source Vcc through a collector resistor 80. The transistor 79 has connected across the emitter and base electrodes a semiconductor diode 81 serving to prevent a high reverse voltage from being applied across those electrodes and also connected across the collector and emitter electrodes a differentiating circuit including a ~ 3~) ~ I
¦ capacitor 82 and a resistor 83. The junction of that ¦ collector electrode and the capacitor is designated by the ¦ reference character C and the junction of the capacitor 82 ¦ and the resistor 78 is designated by the reference character ¦ ~ only for purposes of illustration.
The junction D is connected to one AC input to a rectifier bridge 84 including the other AC input connected to the resistor 83. The rectifier bridge 84 incl~des a l pair of DC outputs connected across a resistor 85 that is ¦ connected at one end to a gate electrode of a Triac 87 `
through a normally open switch 86 and at the other end to the primary winding of the transormer 70. The Triac 87 is connected across AC source through the primary transformer . : :.
70 winding and the switch Sl and has connected thereacross a series combination of a capacitor 88 and a resistor 89 serving as an absorber.
The components 75 through 89 as above described ~ form a zero-voltage firing circuit generally designated by : the reference numeral 90.
With the switch 51 closed, an AC voltage developed at the point A is similar to the source voltage and sinusoidal as shown at waveform A in Figure 54. The AC
sinusoidal vo~tage passes through its zero voltage points at time points to, tl and t2 in each cycle of the source 31.
Assuming that the source VBB is at a null potential, a voltage developed at the point B is sinusoidal between time points t and tl or in the positive half-cycle of the source and remains null between time point tl and t2 or in the negative half-cycle thereof by means of the action of the diode 81 as s],own at l~aveform B in Figure 54. Since the transistor 79 is turned on only in response to a voltage applied to the base electrode to render the latter positive with respect to the emitter electrode, the same is in its ON state between time points to and tl and in its OFP
state between time points tl and t2. Accordingly, a voltage developed at the point C is null when the transistor -~
79 is in its ON state and equal to a voltage across the source Vcc also designated by Vcc when it is in its OFF
state as shown at waveform C in Figure 54.
The voltage at the point C is differentiated by the differentiating circuit 82, 83 to produce alternately a negative and a positive pulse at the point D as shown at.
waveform D in Figure 54. Those pulse are rectified by the rectifier bridge 84 to form positive pulses which appear at a point E connected to the switch 84 at time points t and t2 as shown at waveform E on Figure 54.
~ ~ With the switch 86 closed, the pulses shown at ; ~aveform E in Figure 54 are successively applied to the gate electrode of the Triac 87. In other words, gate pulses -are necessarily develo~ed at the gate electrode of the Triac 87 at the zero passage points of the source voltage .-~
~ ~ or at time points to, tl and t2. Thus it is seen that, 25~ ~ ~ ~even though the switch 86 has been closed at any time point, --~
the Triac 87 is brought into its ON state starting with the zero passage point of the source voltage. As a result, a ~;
~pilot~voltage from the eransformer 70 is applled to the -~: :

: ~ . ' : ., ~ . : '-: ` ' '-I
`, I .
' auxiliary electrode 46 starting Wit]l ~he zero passage point ¦ of the source voltage or time point to, tl or t2 with the result that the principal glow current is ~revented from sharply increased. This means that a liquid flowing in heat transfer relationship along the internal surface of ¦ each electrode 1 or 2 is smoothly heated.
¦ The arrangement of Figure 53 is advantageous in ¦~ that a principal glow current is prevented from sharply Il rising at a firing time point and the glow discharge is , prevented from transiting to an arc discharge due to the ¦
local concentration of the current while efficiency of utili- !
i zation of the source voltage is high.
1l If desired, the zero voltage firing circuit 90 may ¦¦ be formed of solid state relays. I
1l In the arrangements shown in Figures 52 and 53 the I ~ -¦ auxiliary source circuit including the step-up transformer !l is formed of components having stray capacitances between I -one another and with respect to ground with the switch 72 I put in its open position. This results in a fear that a potential at the auxiliary electrodes 46 would be raised due to those stray capacitances until a voltage across the auxiliary electrode 46 and either of the main electrodes 1 and 2 exceeds the discharge breakdown voltage across the associated gap. This results in the undesirable occurrence of a glow discharge between the main electrodes 1 and 2 which ; disables the principal glow discharge to be controlled with the pilot glow discharge.
In order to avoid this objection, the arrangement .~

, .: : . .: . - .
: - , . , , : , 1~ 14~330 ill tr~ted in ligur~ 55 includes a pair of dummy resistors ~3 and 94 connected between the diode 73 and the resistor 71 and between the diode 74 and the resistor 71 respecti~ely.
~he resistors 93 and 94 are effective for determining the potential at the auxiliary electrode 46 so as to prevent the voltage across the auxiliary electrodes 46 and either of the main electrodes 1 and 2 from exceeding the discharge breakdown voltage across the gap as above described.
In other respects, the arrangement is identical to that shown in Figure 53 except for the omission of the switch ,2.
The auxiliary electrode 46 is normally positioned ~o be equidistant from both main electrodes l and 2 and l ~herefore the resistors 93 and 94 are equal in magnitude of 1 resistance to each other in order to equal the voltage across t.he auxiliary electrode 46 and the main electrode l to that ¦ across the electrodes 46 and 2 with the switch 62 put in its open position. Even under these circumstances, it is to be understood that the gap length between the auxiliary ¦ electrode 46 and either of the main electrodes 1 and 2, and ¦ the type and pressure of a dischargeable gas should be preliminarily determined so as to prevent the occurrence of a discharge b.reakdown between the auxiliary electrode 46 and : either of the main electrodes 1 and 2 with the switch 62 put in its open positi,on.
: The arrangement illustrated in Figure 56 is different from that shown in Figure 55 only in tha~ in Figure 56 a Triac is suhstituted for the switch 62 in order to permit 13.~3~

¦ the ON-OFF operation to be repeatedly performed with a high ¦ frequency. As shown in Figure 56, a Triac or a bidirectional triode thyristor 95 is located in place of the switch 62 l shown in Figure 55. rhe Triac 95 includes a gate circuit ¦ 95 connected to a gate electrode thereof to deliver trigger l signals to the gate electrode to turn the Triac 95 on and off ¦ and a series combination of a capacitor 97 and a resistor 98 serving as an absorber.
If desired, the Triac 95 may be included in t]le zero voltage firing circuit 90.
When the pilot glow discharge has the discharge breakdown characteristic with a fairly long time delay, the pilot glow discharge may be fired at time point where the source voltage approaches its peak value provided that the lS Triac 95 has flowing therethrough a current an excess of its holding current. This is attended with the occurrence of the principal glow discharge having a sharply rising current. A current for this glow discharge may sharply rise. In this case, a negative glow included in the principal discharge can not spread following an increase in current to locally concentrate the current resulting in a danger that the glow discharge transits to an arc discharge.
ln order to avoid this danger, it is necessary to determine magnitudes of resistances 93 and 94 and an impedance on the primary side of the step-up transformer 70 enough to prevents a flow of current through the Triac 95 inexcess of its holding current.
In the arrangement illustrated in Figure 57 an 11.l~030 Il `~, I .
electronic switch 98 such as a thyristor with a trigger circuit 99 is connected between the resistor 71 and the junction of dummy resistors 93 and 44 as shown in Figure 57.
~hen a voltage drop across the serially connected resistors 93 and 94 decrease to some extent, and when the electronic switch 98 is put in its ON state by the trigger circuit 99, a current flowing through the electronic switch 98 may exceed its holding current even in the absence of a pilot glow discharge. Under these circumstances, if the pilot glow discharge has the discharge breakdown characteristic with a long time delay, there is a danger tha~ the resulting glow discharge transits to an arc discharge as above described. In order to avoi~ this danger, the resistors 93 and 94 are required to high somewhat in resistance.
Alternatively the electronic switch 98 with its trigger circuit 99 may be connected between the junction of the dummy resistors 93 and 94 and the auxiliary electrode 46 as shown in Figure 58. In these case, the resistors 93 l and 94 are not particularly subjected to limitations as to ¦ their resistances unless a voltage across the auxiliary electrode 46 and either of the main electrodes l and 2 is reduced.
The arrangements shown in Figures 55 through 58 l ensure that the principal glow discharge is controlled with ¦ the pilot glow discharge. This is because, the dummy resistors prevent the potential at the auxiliary electrode from floating by means of stray capacitances as above described in conjunction with Figures 52 and 53 and the lll tO3~1 !!
¦ like in the absence of the voltage applied to the auxiliary electrode.
¦ The arrangement illustrated in Figure S9 comprises I an electrically isolating transformer 141 including a ¦ primary winding connected across the AC source 31 and a secondary winding connected across a series combination of a ¦ rectifying dic~e 142, a current limiting resistor 143 and capacitor 144, and an NPN transistor 149 inclu-ding an , emitter electrode connected to one side of the capacitor 1l 144 and a collector electrode connected to the other side of I¦ the capacitor 144 through a semiconductor diode 146 for !1 absorbing back pulses. ~he transistor 145 includes a base li electrode connected to a gate circuit 149 also connected to ¦I the emi~ter electrode thereof to turn the transistor 145 on 1 and off.
The components 141 through 146 form a high voltage ,, pulse generator circuit generally designated by the reference i, numeral 140 with a step-up pulse transformer 147 which Il includes a primary winding connected across the diode 146 ¦1 and a secondary winding connected to a semiconductor diode !1 148 for shaping a pulse waveform.
¦ As in the arrangement of Figure 57, the diode 148 is connected to the resistor 71 subse~uently connected to the auxiliary electrode 4h through the thyristor 98 which is turned on and off by a trigger circuit 99. Further the serially connected dummy resistors g3 and 94 are connected across the main electrodes 1 and 2 also through the switch 51 across the AC source with the junction of bcth resistors ~14~30 ¦! _ ~ j ¦I connected to the auxiliary electrode 46.
The operation of the arrangement shown in Figure 59 will now be described with reference to Figure 60 wherein there are illustrated a voltage waveform V across the main electrodes 1 and 2 and a no-load voltage waveform VN at ¦ the auxiliary electrode 46. With the main electrode 1 disposed oppositely to the main electrode 2 to form there-between a predetermined gap fulfilling the relationship ~I that the discharge breakdown voltage Vf for the gap is ; higher than the peak value Em of the source voltage under the predetermined discharge conditions, the switch 51 is closed to apply the AC voltage across bcth electrodes 1 and from the source 31. Also the source 31 charges the capacitor Il 144 with the polarity illustrated through the transfor~er 141,¦
ll the diode 142 and the resistor 143. Then gate and trigger ¦ circuits 149 and 99 respectively apply simultaneously I respective gate signals to the transistor 145 and the ¦ thyristor 99 to turn them on. The turn-on of the transistor 149 causes the charged capacitor 144 to discharge through the primary winding of the pulse transformer 147 and the now ~ -conducting transistor 145. As a result, a pulse voltage stepped up by the pulse transformer 147 is supplied from the secondary winding thereof through the diode 148, the resistor 71 and the now conducting thyristor 98 to the auxiliary electrode 46. It is noted that the circuits 149 and 99 generate the respective pulses before the voltage across the maln electrode 1 and 2 reaches the discharge breakdown voltage VO. As shown in Figure 60, the circuits 149 and ; .
99 generate the pulses at time point t2 before time point to ~here the source voltage reaches the discharge breakdown voltage VO in each positive half cycle thereof and the ~ pulses terminates short after time point to. That is each : 5 pulse has a predetermined pulse width a little longer than a time interval between time points t2 and to. Each pulse ~, is shown at waveform VN in Figure 60 as being superposed on that portion of the source voltage divided by the resistors 93 and 94, assuming that both resistors are equal in magnitude of resistance to each other. In the next succeeding negative cycle of the source voltage the pulse is similarly developed at time point t3 before time point tl where the voltage across the main electrodes 1 and 2 reaches the negative value -VO of the discharge breakdown ~ voltage and terminates short after time point tl to have the same pulse width as that appearing in the positive half-cycle of the source voltage.
In the arrangement of Figure 59 it is required to cause a pilot glow discharge before time point to or tl by ~ applying the pulse waveform VN to the auxiliary electrode ~6 as above described. Also it is required to selec~ the pulse width so as to effect surely the discharge breakdown of the gap between the auxiliary electrode and either of Il the main electrodes 1 and within the duration of the ~¦ associated pulse.
; ~ In general, a time delay is caused after the voltage has been applied across discharge gaps and until the discharge breakdown is accomplished therebetween. It is well I
,~ I .
known that this time delay is equal to the sum of a time interval between the application o~ the voltage across discharge gap and the appearance of a first electron resulting in the initiation of development of the electron avalanche and another time interval between the initiation of development of an electron avalanche and the completion of a stead-state discharge. The first mentional time interval is called a statistic delay and the latt~r is called a formation delay. The statistic delay is over-poweringly long.
Assuming that a voltage applied across the particular ~ischarge gap has the peak value higller that a voltage effecting the DC breakdown of the discharge gap, steped voltages are applied across the discharge gap nO times.
Assuming that, among them the n applications of the voltage has time delays not shorter than T and (n + an) applications , thereof has time delays not shorter than (T + aT), '.'' an = -AnaT

~ holds where A designates a constant. Thus ': ~ .
'~ n nOe .
2~5 is fulfilled by the statistic delay. The above expression may be plotted into a straight line with the axes of ordinates and abscissas representing the n and T respectively in a semilogarithmic scale~ A graphic representation thus ., ::

: . . . .. , , . . ... ..
: . ' . . . . . :
, ~ .

1~
~ 330 ' ~ I . ''.
plotted is called a Laue plot.
¦Figure 61 shows on example of the Laue plot. In ¦ Figure 61 an extremity of an auxiliary electrode having a ¦ diameter of 3 millimeters is located at an edge of a gap ¦ of 3 millimeters formed between a pair of main opposite ¦ electrodes to form a spacing of about 1 millimeter between ¦ the extremity of the auxiliary electrode and either of the main electrodes. The gap was filled with a discharge gap ¦ formed of a mixture including 89% by volume of helium and ¦ 11% by volume of hydrogen under a pressure of 100 Torrs.
[n Figure 61 the reference numerals 150, 151, 152 and 153 ¦ depict the source voltages having the peak values of 600, ¦ ~00, 1000 and 1200 volts respectively. From a stepped ¦ curve lS2, for example, it is seen that for the peak source lS ¦ value of 1000 volts the time interval between the t2 and to ¦ or between the t3 and tl (see Figure 60) must be of at ¦ least 250 microseconds. Also the auxiliary source for the pilot glow discharge should have a current capacity of at l least about 10 milliamperes in order to transit smoothly - 20 ¦ the pilot glow discharge to the principal glow discharge.
By taking account of a tir,le delay with which the discharge gap is brown down with the pulse voltage of the voltage wavef~rm NN shown in Figure 60, the waveform UN is given a pulse width or a duration defined by the time intervalc I ranging from time point t2 or t3 to time point to or tl l _ _ _ _ respectively while the current capacity of the auxiliary ¦ source is determined as required for transiting the pilot l glow discharge to the principal glow discharge and the I

.
i. ~ -¦ pulse voltage delays rapidly at and after time point ~ or tl.
This measure ensures that the pilot glow discharge is always ¦ caused prior to time point to or tl and the principal ¦ discharge current surely rises at time point to or tl.
¦ After the principal glow discharge has been caused ¦ between the main electrodes 1 and 2, discharge energy from I the principal glow discharge as thermal energy alternately ¦ enters the main e]ectrodes 1 and 2 with result that a ¦ liquid flowing in contact relationship through either of the ¦ main electrodes is instantaneously heated.
The arrangement of Figure 59 is advantageous in that the principal discharge current smooth]y rises to cause the deve~bpm-entof a negative glow involved to satisfactorily follow up a change in discharge current thereby to prevent ~15 ¦ the local concentration of the current without the glow ¦ discharge transiting to an arc discharge while the efficiency of utilization of the source voltage. This is because the , ¦ auxiliary electrode is adapted to be applied with a pulse ¦ voltage that rises before time point where a voltage applied , 20 ¦ across the main electrodes reaches a glow hold minimum voltage across the main electrodes thereby to fire always ¦ ~ the pilot glow discharge before that time poin~ and rapidly ¦ falls to its null value at and after said time point. Also the use of the pulse waveform is effective for decreasing ~2S ¦ the power capacity of the auxiliary source and therefore ¦ reducing a dimension and a cost thereof.
Figure 62 shows a modification of the arrangement shown in Figure 59. The arrangement illustrated comprises a , I . i ~ ' ' ' .:
;
,.. . .. ' - ' ' ' : ';- .:'.'':.: ,'',: ' , :

. :~
pair of electrically isolating transformers 141 and 155 including a common iron core and a common primary winding connected across the AC source 31 through the normally o~en switch 51, the high voltage pulse generator circuit 140 as above described in conjunction with Figure S9 connected to t.he transformer 141, and a current supply circuit generally Gesignated by the reference numeral 154 and connected across the transformer 155.
The current supply circuit 154 includes a center-~10 t:apped secondary winding of the transformer 155, and a pair c)f se~iconductor diodes 156 and 157. The diode 156 is c.onnected at the anode electrode to one side of the source 31 through the switch 51 and therefore the main electrode 1 while diode 157 is connected at the anode electrode to the ~ther side of the source 31 and therefore the main electrode 2 that is, in turn, connected to ground. lhe center tap on the secondary transformer 155 wi.nding is connected to the outp~t of the pulse generator circuit 140 or the junction of the diode 148 and the current limiting resistor 71. ..
1 In other respects, the arrangement is identical to that shown in Figure 59. The dot convention is used to identify the polarity of the instantaneous voltage developed across the associated transformer winding. - --The current supply circuit 155 is operative to : 25 full-wave rectify an AC voltage induced across the secondary transformer 155 winding and supply a current due to the full-wave rectified voltage to the auxiliary electrode 46 throu~h the resistor 71 and the thyristor 9~ with the pulse ~` .

. .
: ' - ' .: , .

114~30 ~ .
r I .
voltage from the pulse generator circuit 140.
¦ In the arrangement of Figure 62, the discharge gap ¦ hetween the main electrodes 1 and 2 has been dimensioned as above described in conjunction with Figure 59 and the ¦ switch 51 is closed to supply the source voltage across the .
¦ main electrodes 1 and 2. The source voltage is a commercial ¦ hC voltage having a frequency of 60 hertzs as shown at dotted ¦ waveform V in Figure 63 wherei.n its cycle has a duration of ¦ i6.7 milliseconds.
¦ The pulse generator circuit 140 generates a high ¦ ~oltage pulse in each of the half-cycles of the source voltage in the same manner as above described in conjunction ~ ¦ with Figure 59. After having shaped by the diode 148, the I ¦ high volts pulse is developed on the resistor 71 and . 15 ¦ superposed on the full-wave rectified voltage from the I ¦ current supply circuit 154 also applied to the resistor 71 : :
¦ as shown at voltage waveform VN in Figure 62. Then pulse ¦ voltage VN superposed on the voltage from the current supply .
¦ circuit 154 is supplied to the auxiary electrode 46 through ? ¦ the conducting ~hyristor 98.
¦ From Figure 63 it is seen that the voltage waveform I ¦ ~N includes the full-wave rectified component having a ¦ relative volt~ge to the main electrode 2 equal to a voltage ~ -. ¦ VOP for the pilot glow discharge at time point t6 in the I positive half-cycle of the source voltage an also a relative voltage to the main electrode 1 equal to that voltage VOP
¦ at time point t7 in the negative half-cycle thereof. Time ¦ points t6 and t7 are ahead of time points to and tl ~ I

4~30 I
I
respectively where the source voltage is equal to the glow hold minimum voltage VO.
¦ With the main electrode 1 higher in potential than ¦ the main electrode 2, the diodes 156 is in its OFF state ¦ while the diode 157 is in its ON state tending to cause a ¦ pilot glow discharge between the auxiliary electrode 46 ¦ and the main electrode 2. On the contrary, with the main -¦ electrode 1 less in potential than the main electrode 27 the ¦ diodes 156 and 157 are turned on and off respectively.
¦ This tends to cause a pilot glow discharge between the auxiliary electrode 46 and the main electrode 1. In each ¦ case, the voltage across the auxiliary and main electrode 46 and 1 respectively is equal to that across the auxiliary and main electrode 46 and 2 respectively so that a current for the pilot glow discharge remain unchanged. With the .' auxiliary electrode 46 equidistant from the main electrodes 1 and 2, the transit of the pilot glow discharge to the ~' principal glow discharge between the main electrodes 1 and 2 ; is accomplished in the similar manner in both cases.
20~ ~ The voltage waveform VN also includes a pulse waveform component from the pulse generator circuit 140 rising at time point t2 or t3 behind time point t6 or t2 and falling at time point t4 ahead of time point to or tl.
The pulse waveform component results from a gate pulse P i 2~5 ~ ~ from either of the gate and trigger circuits 149 and 9g rising and falling simultaneously with the rise and fall of the associated pulse component. The pulse waveform component is required to have a pulse width sufficient to effect the :' :~ ~
~ - 102 -,.: ~
'' -- .

1~4030 I .............................................................
,.
`j I .
discharge breakdown of the gap beiween the allxiliary electrode 46 and either of the main electrodes 1 and 2. It is to be noted that it is not required to cause time point .4 or tS to coincide with time point t2 or tl resnectively as in the arrangement of Figure 59 and that the pulse width may be sufficiently shorter than that reauired for the latter. In addition, the discharge breakdown scarcely ' .equires a current resulting in the pulse generator circuit 140 reducing sufficiently in power cap~city.
The gate pulse from each of the gate and trigger , , circuits 149 and 99 should have a rise time fulfilling the Collowing requirements: ~he gate pulse P1 should rises at -time point t2 or t3 required to be behind time t6 or t7 respectively while the pilot glow discharge should be caused not later than time point to or tl. Otherwise the principal discharge current is too sharply raised to cause the spread of the particular,negative glow to follow this rise in , current resulting in a danger that the current is locally concentrated on either of the main electrode to permit the glow discharge to transit to an arc discharge. Also the source voltage can be utilized only with a low efficiency.
~hus time point _4 or t5 should be ahead of time point t or tl respectively. n With the gate ~ulse Pl generated to fulfill the requirements as above described, the pilo~ glow dischar'ge is always caused ahead of time to or tl in the positive or :
negative half-cycle of the source voltage and the glow discharge current through the main electrodes 1 and 2 smoothly ' ' .': ', ~ ; - 103 -.,".~.

.. . . .
'- : . ~ ' , .: .
:' ' , ' ,: . ' , :
.

ll 1~1~030 ris s ~t and after t~me point to or tl in the pos;tive or negative half-cycle of the source voltage. Accord;ngly, ¦ the principal glow discharge is established resu]ting in the ¦ .nstantaneous heating of the particular liquid contacted I by either of the main electrodes 1 and 2.
¦ Further it is required to make the peak voltage value of the sinusoidal component of the vo]tage waveform ~N less than the discharge breakdown voltage for the gap ¦ between the auxiliary electrode 46 and either o~ the main ¦ electrodes 1 and 2 thereby to prevent the pilot glow I discharge from firing with the sinusoidal component.
¦ hlternatively, it is required to impart a high value to each ¦ of the resistance 93 or 94 to prevent the voltage wave~orm ~N from being applied to the auxiliary electrode 46 in the lS absence of the gate pulse Pl and to prevent a current flowing through the thyristor 98 via the resistors 93 and 94 from exceeding the holding current thereof when thc pilot glow discharge is not fired. Also the diodes 156 and 157 must have such reverse voltage withstanding charac~eristic that both diodes are not broken down with the high voltage pulses ~enerated by the pulse generator circuit 14~.
If desired, the pulse generator circuit may utilize a peak transformer.
The arrangement of Figure 62 is advantageous in that the pulse generator circuit can reduce in power capacity resulting in the provision of an auxiliary source circuit easy to be manufactured and inexpensive. ~his is because the pulse generator circuit for effecting the , , -10~-¦ discharge breakdown of the pilot glow discharge gap is ¦ separated from the circuit for supplying current to the main ¦ electrodes after this discharge breakdown.
¦ Figure 64 shows a different modification of the S ¦ present invention driven by a three-phase AC source. The ¦ arrangement illustrated comprises three main electrodes ¦ lU, lV and lW radially disposed by having their longitudinal ¦ axes arranged at equal angular intervals of 120 degrees.
¦ The main electrodes are in the form of hollow cylinders ¦ ~laving one end closed into a crown shape that, in turn, ¦ faces the remaining closed ends of the same shape. The ¦ main electrode lU, lV and lW include the other end portions rigidly fitted into respective annular supporting members 1 14U, 14V and 14W interconnected through enclosure portions 9 I formed of an electrically insulating material such as glass ¦ polcelain or the like and seal fittings lOU, lOV and lOW
connected to both adjacent suppQrting members and the adjacent ¦ edges of the enclosure portions 9 to define a hermetic l discharge space. The other ends of each electrode lU, lV or 1 lW is closed with a blind cover plate 23U, 23V or 23W having I an inflow tube 42U, 42V or 42W and an outflow tube 44U, 44V
¦ or 44W is extended and sealed therethrough.
I Three auxiliary electrodes 46U, 46V and 46W are ¦ radially extended and sealed throu~h the enclosure portions 9 ~5 ¦ respectively to be equidistant from the adjacent main electrodes and includes end portions bent toward the associatec main electrodes to form very narrow gaps therebetween. For example, the auxiliary electrode 46U is radially extended and .

. . . ~ . . .

~ 4Q30 -¦ sealed through the enclosure portion 9 disposed between the ¦ main electrodes lU and lV and includes the end portion bent ¦ toward the main electrode lU so as to cause a pilot glow ¦ discharge, Each of the auxiliary electrodes is coated with ¦ the same electrically insulating material as the enclosure :
: ¦ portion 9 except for both the end facing the associated ¦ main electrode and that portion externally protruding from ¦ the mating enclosure portion 9.
¦ A three-phase source is represented by source ¦ terminals U, V and W which are connected to annular electrode . ¦ terminals 6U, 6V and 6W fitted onto those portions of the main electrodes lU, lV and lW disposed externally of the ` ¦ enclosure portions 9 respectively. Each of the auxiliary ¦ electrodes is connected to the electrode terminals disposed 1 on the adjacent main electrodes through individual dummy resistors. For example, the auxiliary electrode 46U is ~' I connected to the electrode terminal 6U of the main cylinder lU through the dummy resistor 47U on the one hand and to the . I electrode terminal 6V of the main electrode lV through the .~; 20 ¦ dummy resistor 48U.
The auxiliary electrode 46U is also connected by a current limiting resistor 49U to an auxiliary source circuit 50 also connected to the electrode terminal 6U. The ¦ auxiliary source circuit 50 is further connected across the 2s~ ¦ source terminals U and V through a normally open switch : ¦ 51U connected to the source terminal V.
¦ A circuit identical to that above described is ; provided for each of the remaining main electrodes and the ~ .
`~ - 106 -,, .. ~ .. .... .. . .

, . :...: .. ..

-~1 1~ 30 auxiliary electrode operatively associated therewith and includes the components identical to those above described.
Therefore the identical components are designated by like reference nu~erals suffixed with the reference character U, V or W i~entifying the mating source terminal or the phase of the three-phase source.
The operation of the arrangement shown in Figure 64 will now be described with reference to Figure 65 wherein there are illustrated voltage and current waveforms developed at various points in the arrangement of Figure 64 with a voltage Vu applied to the main electrode lU being selected as a reference.
While a liquid to be heated is flowing through the interior of each main electrode via the associated inflow tube and leaves the mating outflow tube a three phase voltage is applied to the main electrodes lU, lV and lW through the source terminals U, V and W and all the switches 51U, 51V
and 51W put in their closed position. At time point tl short before a voltage (see waveform Vv, Figure 65) applied across the main electrodes lU and lV reaches a glow hold .--minimum voltage V0, a high voltage pulse ~see waveform Puo, Figure 65~ from the auxiliary source circuit 50U is applied to the auxiliary electrode 46U to cause a pilot glow discharge across the narrow gap between the auxiliary electrode and main electrodes 46U and lU respectively with the main electrode lU acting as a cathode. This pilot glow discharge is caused with a low current, and upont time point D being reached, it instantaneously induces a glow ., ."

~ 1 ~14~30 ¦ discharge between the main electrodes lU and lV with the ¦ electrode lU acting as a cathode. The latter discharge ¦ spreads through the surface of both main electrodes lU and ¦ lV and is sustained after time point D.
I Then when a voltage (see waveform Vw, Figure 65) ¦ applied across the main electrodes lU and lW exceeds the ¦ glow hold minimum voltage V0 at time point E, the glow I discharge developed between the main electrodes lU and lV
¦ plays a role of the pilot glow discharge to cause a glow I discharge between the main e~ectrodes lU and lW at and ¦ after that time point with the maln electrode lU acting as a cathode.
¦ At time point F voltaee across the main electrodes ¦ lV and lW is equal to the voltage V0 but no discharge is ¦ caused between those main electrodes because of the absence ¦ of a pilot glow discharge with the main electrode lV acting ¦ as a cathode. Therefore a high voltage pulse tsee waveform ¦ PvO, Figure 65) from the auxiliàry source circuit 50V is I applied to the auxiliary electrode 46V at time point t2 I short ahead of time point F to cause a pilot glow discharge , I between the auxiliary and main e]ectrodes 4~V and lV
¦ respectively. That pllot glow discharge similarly causes a glow discharge between the main electrodes lV and lW at I and after time point F with the main electrode lV acting as ¦ a cathode.
I When time point G is reached, the voltage Vv across ¦ the main electrodes lU and lV is e~ual to the voltage V0 and : I the glow discharge caused between the main electrode lV
I ~ ~
I

I .`

.
.

4~3~
.,,, . .
acting as the cathode and the main electrode lW plays a role of a pilot glow discharge. This causes a glow discharge between the main electrode IV acting as a cathode and the main electrode lW at and after time point G.
Similarly, since the voltage Vw across the main electrodes lW and lU exceeds the voltage VO at time point H, a high voltage pulse (see waveform PwO, Figure 65) from the auxiliary source ~ircuit 50W has been preliminarily applied to the auxiliary electrode 46W at time point t3 short ahead of time point H to cause a pilot glow discharge between the auxiliary electrode 461~ and the main electrode ~ , acting as a cathode. The pilot glow discharge between the auxiliary and main elect~ode 46W and lW respectively transits to a glow discharge caused between the main electrode lW
acting as a cathode and the main electrode lU at and after , ,,,,-time point H.
Then at time point I the voltage Vw across the main electrodes lV and lW exceeds the voltage VO so that the ' glow discharge between the main electrodes lW and lU serves as a pilot glow discharge to cause,a glow discharge between ' , the main electrode lW acting as a cathode and the main electrode lU until one cycle of the source voltage is completed.
Thereafter the pTocess as above described is ~5 repeated to cause repeatedly glow discharge between pairs of the main electrodes. When acting as the cathode, the ' main electrodes successively heat the liquid therein.
From the foregoing it will readily be understood :~

.

ll 1~4~30 that the gate pulses are repeatedly applied to the auxiliary electrodes 46U, 46V and 46W at time points t defined by t = tl + nT, t = t2 ' nT and t = t3 + nT

respectively where T designates a period of the three-phase source voltage and n indicate any positive integer including ~ zero.
: In Figure 65 solid current waveform IU designates a glow discharge currents with the main electrode lU
acting the cathode, dotted current waveform Iy those with ; the main electrode iV acting as cathode and broken current waveform lW designates the glow discharge current with the main electrode lW acting as the cathode. The reference Uo~ PVo and PwO designate no-load pulse forms which or change to the actual pulse waveforms Pu~ PV
and PW respectively after the associated pilot glow discharges have been fired.
Also it is noted that Figure 65 illustrates the ~0 waveforms developed during a time interval equal to twice the period T of the source voltage Vv applied across the main electrodes lU and lV and that the polarity of the current waveforms have not been considered.
From the foregoing it will readily be understood that the glow discharge has a time period equal to three times that provided by single-phase system and therefore three-phase apparatus tripple in p~wer capacity single-phase apparatus.
~ ' . .

. .
.

1 1~14~30 In the arrangement of Figure 63 the auxiliary electrode is disposed between each pair of adjacent main electrodes for the purpose of controlling thermal energy entering each of the main electrodes. Ho~ever it is included in the scope of the present invention to replace the auxiliary electrode by a bidirectional triode thyristor serially connected to each of the main electrode to control thermal energy entered thereinto through the ON-OFF
operations of the thyristors.
The arrangement illustrated in Figure 66 is different from that shown in Figure 64 only in that in Figure 66 a combination of a pulse transformer 70U, 70V or 70W and a high voltage pulse generator circuit 140U, 140V -or 140W is substituted for each auxiliary source circuit. ~
The combination of the pulse transformer and pulse generator :
circuit may be identical to the pulse generator circuit 140 ,~ shown in Figure 59.
Also the main and auxiliary electrodes are -schematically illustrated in Figure 66 and may ~e similar Z0 ~ ~ to those shown in Figure 64 and the resistors 48U, 48V and 48W are omitted.
Figure 67 shows another modification of the arrange-ment shown in Figure 66. In the arrangement illustrated, ,~ the electrically isolating transformer 7~ includes a primary 25~ winding Wl connected across the source terminals U and V
through the swltches 51 and a pair of secondary windings W2 and W3 connected respectively across a high voltage ¦¦ pulse ge rator circuit 140 such as above described in ~ I -111- I
: ,~
'-- , . - -' - ' :

conjunction with Figure S~ and a gate circuit 161. The pulse generator circuit 140 includes one output connected to the source terminal U and the other output connected to anode electrodes of thyristors Su, Sv and Sw through the common current limiting resistor 49. The thyristors Su, Sv and Sw include cathode electrodes connected to the auxiliary electrodes 46U, 46V and 46W respectively. The gate circuit 161 is connected to the thyristors Su, Sv and SW to control the firing thereof.
In other respects, the arrangement is substantially identical to that shown in Figure 66.
Figure 69 illustrates voltage and current waveforms developed at various points in the arrangement shown in Figure 67. From the comparison of Figure 68 with Figure 65 it is seen that voltage and current waveforms shown on the upper portion of Figure 68 are substantially similar to .
those illustrated in Figure 65 and pulse waveforms Po are substituted for the pulse waveforms PU-Puo, PV-Pvo and ~ PW-Pwo shown in Figure 65. Thus like reference characters ; 20 have been employed to identify the waveforms corresponding to those illustrated in Figure 65. Thus the arrangement is substantially identical in opeTation to that shown in Figure 66.
As seen in Figure 68, the gate circuit 161 applies a gate pulse (see waveform Gu) across the gate and cathode electrodes of the thyristor Su short before the high voltage pulse tsee waveform Po frnm the pulse generator 140 is supplied to the auxiliary electrode 46U to bring it in ., , . : .
- : , 4~3~) ~ l .' '': I .
its conducting state and then the ~ulse Po is supplied tO
¦ the auxiliary electrode 46U through the resistor 49 and the ¦ now conducting thyristor Su. This is true in the case of -¦ the remaining pulses P passin~ through the respective ¦ thyristors Sv and Sw.
¦ Each of the gate pulses shown at waveforms Gu, GV
¦ and GW in Figure 68 should have a pulse width sufficient ¦ to ensure that a pilot glow discharge is fired between the ¦ associated auxiliary and main electrodes such as shown by ¦ 46U and lU and transits to the principal glow discharge I caused between the mating main electrodes such as shown by ¦ lU and lV. That is, the gate pulse should be at least I sustained until time point is reached where the associated ¦ source voltage, for example, the voltage Vv exceeds the ; 15 ¦ glow minimum voltage VO If the pilot glow discharge ~ I causes a current flowing through the associated thyristor i~ ¦ to exceed its holding current then the gate pulse may continue until the pilot glow discharge is fired.
. ; The arrangement of Figure 67 is advantageous over ~ that shown in Figure 66 in that the resulting circuit is ~ ~ simple, small-sized and inexpensive because of the provision 3~ of a single high voltage pulse generator circuit.
In the preferred embodiments of the present invention, the main electrodes and associated components, such as the ~3~ ~ . 25~ flow confining tubes, the connecting tubes, the inflow and outflow tube, the blind cover plates shown, for example, in Figure 24 are formed of metallic material and put in ~- contact with a heated liquid that is electrolytic. This :~:

:~ .' ...................... ' ' .
.

,11 may result in a fear that those metallic components are corroded with the heated liquid and reduced in useful life.
Particularly the main electrodes and those tubes dir~ctly connected thereto have high probabilities of electrolytic corrosion because the source voltage is directly applied across the main electrodes while the inflow and outflow tubes are connected to grouned thereby to permit currents to flow in to the main electrodes and those tubes through the heated electrolytic liquid.
The arrangement shown in Figure 69 includes corrosion preventing electrodes for pre~enting metallic components from corroding as above described. In the arrangement illustrated a corrosion preventing electrode 161 or 162 is electrically insulatingly extended and sealed through that wall portion of the flow confining tube 20 or 21 facing the inside of the gap forming surface of the main electrode 2 or 1, that is, each of the opposite surfaces of both main electrodes with an electrically insulating holder 163 or 164 hermetically interposed therebetween. The electrode protrudes into the flow path for the heated liquid. The anticorrosive electrode may be formed of platinum, carbon, triiron tetroxide (Fe3O4? or the like. The a DC source 165 or 166 is connected across the corrosion preventing electrode 161 or 162 and the electrode terminal 5 or 6 thereby to supply to the electrode 161 or 162 a voltage higher than the voltage across the main electrodes. To this end. Each of the DC source 165 or 166 includes a negative side connected to the associated electrode terminal 5 or 6.
::
'' .
- 114 - ~
'.

1~ 31~
..
. ' :, Then the electrode terminals 5 and 6 are connected to a control circuit identical to that shown in ~igure 34.
In other respects, the arrangement is ;dentical to that shown in Figure 24 except for the provision of the S auxiliary electrode 46 but the main electrode 1, in this case, made of stainless steel or the like, the flow confining tube 21, the blind cover plate 23, the connecting tube 36, the insulating tubes 38 and 40, the inflow tube 42 and the outflow tube 44 form an assembly prevented from corroding and generally designated by the reference numeral 167. Also, the similar components 2, 20, 22, 35, 37, 39, 41 and 43 form another assembly prevented from corroding and generally designated by the reference numeral 168. The main electrode 2 is also made of stainless steel.
lS The corrosion of the main electrodes and others is called the electrolytic corrosion resulting from a flow of current therethrough via a heated electrolytic liquid that is caused from the dissolution of materials forming the main electrode and others into an electrolyte such as water. In the arrangement of Figure 69, the DC sources 165 and 166 are adapted to apply to the respective corrosion preventing electrodes 161 and 162 voltages higher that the voltage applied across the associated main electrodes. Thus the corrosion preventing electrodes 161 and 162 provide the ~5 so-called scapegoat electrodes. That is, the material or materials forming the scapegoat electrode is or are dissolved ;~
into an electrolyte such as water thereby to prevent the materials forming the assemblies 167 and 168 form dissolving :

Ll ;111~13(1 ,, . ~:
into the heated liquid resulting in no corrosion occurring.
The DC sources 165 and 166 may be omitted by forming the corrosion preventing electrode of a metallic material less in corrosion potential and more easily ionized than the material of the main electrode. For e~ample, with the main electrodes 1 and 2 formed of stainless steel, magnesium, zinc, aluminum, etc. are optimum for forming the corrosion ~reventing electrode.
Also the DC source may be repaced by any suitable source for supplying a DC voltage.
Figure 70 shows corrosion preventing electrodes rovided on the arrangement shown in Figure 39. In Figure 70 the corrosion preventing electrodes 161 and 162 are provided on the exposed portion of the feed water tube 20 disposed within the main electrode 2 and on the outer wall of the main electrode 1 respectively in the same manner as above described in conjunction wit~ Figure 69.
Then the corrosion preventing electrodes 161 and 162 ~ fire connected to terminals d and e subsequently connected, 20~ for example, to the DC sources 165 and 166 (see ~igure 69) respectively. Also terminals a and b connected to the electrode terminals 5 and 6 respectively are connected across the AC.source 31 shown in Figure 69 while a terminal C connected to the auxiliary electrode 46 is connected to teh auxiliary source circuit 50 also shown in Figure 69.
~; Figure 71 shows anticorrosive electrodes provided ~, ~ on the arrangement illustrated in Pigure 64. As shown, an ¦~ anticor sive electrode 161U, 161V or 161W is electrically . '~
:: ~:~
1, , . . - . . .
'' . ' .'' ' ,' ~ ..
.

..
'"; , lnsulatingly extended and sealed through the feed water ~ube 42U, 42V or 42W operatively coupled with each main electrode lU, lV or lW with an electrically insulating holder 164U, 164V or 174W interposed therebetween.
Figure 72 shows a separate modification of the present invention wherein a temperature of a heated liquid is measured. In the arrangement illustrated, a temperature sensor 169 such as a thermistor is electrically insulatingly extended and sealed through that portion of a flow confining .ube 20 facing the peripheral wall of the main electrode 2 with an electrically insulating holder 170 interposed therebetween.
The temperature sensor 169 may entirely covered with a electrically insulating material in accordance l~ith the particular electric field established in the vicinity .' thereof. ..
~: In other respects, the arrangement is substantially :~
-dentical to that shown in Figure 24 except for the provision of the auxiliary electrode 46. .:
:~ The electrode terminals 5 and 6 and the auxiliary :.
electrode 46 are connected to a control circuit identical to that shown in Figure 57 except for the omission of the zero volt fir.ing circuit 90. The temperature sensor 169 -:
includes an output connected to the trigger circuit 99 for .. :
. 2~5~ the thyristor 98.
'~ In operation the temperature sensor 169 senses '~ the temperature of the heated li~uid and feeds a . ..
measured temperature signal to the trigger circuit 99.

~:; - 117 -::

.

More specifically~ with the temperature sensor 169 for~ed of a thermistor or a temperature measuring resi.stor, a resistance thereof is changed with a temperature so that a signal representative of a change in resistance is applied to the trigger circuit 99. Alternatively, with the tempera-ture sensor 169 formed of a thermocou~le, it responds to the temperature of the heated liqui.d to change in thermo-electromotive force thereof. This change in thermoelectro-motive force is signalled to the trigger circuit 99.
If it is desired to control the heated liquid to a predetermined fixed temperature then the actual tempera-t.ure measured by the temperature sensor 169 is compared ith an output therefrom at a predetermined temperature as : -the reference. With the actual temperature higher than the predetermined temperature, the trigger circuit 99 zpplies no trigger signal to the thyristor 98. Otherwise, :
the trigger circuit 99 delivers the trigger signal to the thyristor 98. Then the thyristor 98 is correspondingly ~:
: turned on and off to fire and extingui.sh a pilot glow dis-~harge thereby to effect the ON-OFF control of a glow discharge between the main electrodes 1 and 2.
Under these circumstances, some time goes until ~ heat from either of the main electrodes 1 and 2 acting as ..
; the heating surface is transferred to the heated liquid.
This results in a time delay with which the temperature of the heated liquid is controlled. Therefore, the temperature sensor 163 has preferably a sensor end located as near to the heating surface of the associated main electrode as - 118 - ~ .

,~ ,., , ,. .

, 1114:'~30 p~s ble.
With the temperature of the heated temperature controlled according to a predetermined program, the function of effecting such control may be incorporated into the S trigger circuit 99 and the thyristor 98 is operated in the ON-OFF control mode and in accordance with the output signal from the temperature sensor 169.
The temperature sensor 169 may be used wi~h the control of the glow discharge effected by a control such as a thyristor connected in series to the particular glow ~ discharge heating apparatus in a circuit with an electric: source circuit for the apparatus.
The arrangement illustrated in Figure 73 is different ~rom that shown in Figure 72 only in that in Figure 73 a lS bidirectional triode thyristor or a Triac is provided to c:ontrol the glow discharge as in the arrangement of Figure .
: 48. In Figure 73 a thyristor 172 is connected at the anode electrode to one of the DC output terminals of the rectifier bridge 66 and at the cathode electrode to the Triac 60. The resistor 67 is connected to the Triac 60 ~: at the gate electrode but not to one main electrode thereof.
Then the thyristor 172 has the cathode and gate electrodes ~:~ connected across a trigger circuit 173 subsequently connected 3 to the temperature sensor 172.
~ZS In other respects, the control circuit is substan-~! ; tially identical to that shown in Figure 48. I~owever the : dot convention is used to identifv the polarity of the . instantaneous voltage across the associated transformer ' : .

'., ~,,, ll ~ 30 `' .. .
wlndlng .
The electrically isolating transformer 65 is o~era-tive to adapt a potential difference developed across the resistor 64 to a voltage re~uired for the Triac 60 to be fired.
With the switch 51 put in its closed position, the step-up transformer 70 applies a high AC voltage across the electrodes 1 and 2 resulting in the discharge breakdown cccurring therebetween. ~his cause a potential difference across the current limiting resistor 64 whereby a potential difference appears across the resistor 67 through the transformer 65 and the rectifier bridge 66. At that time, the trigger circuit 173 is actuated to put the thyristor 172 in its ON state to cause a trigger signal to be applied ; -`
to the gate electrode of the Triac 60 to turn it on. There-fore the AC source across the source 31 is supplied across the main electrodes 1 and 2 through the now conducting thyristor 60 to cause a glow discharge therebetween.
Under these circumstances, the temperature sensor ;20 ~ 169 senses a temperature of a heated liquid involved and -~
~ ~ feeds signal for the sensed temperature to the trigger i- circuit to control the glow discharge between the ` electrodes 1 and 2.
' From the foregoing it is seen that in the arrange-,~! 25 ~.ents shown in Figures 72 and 73 the temperature of the heated liquid sensed by the temperature sensor is fed to the auxiliary source circuit for controlling the glow dis-charge used between the main electrodes resulting in the . . .
: : - : . : . ., , : ,:

~ 03 ! 1.
easy, reliable temperature control of the heated liquid.
¦ While the present invention has been illustrated ¦ and described in conjunction with various preferred embodi-¦ ments thereof it is to be understood that numerous changes ¦ and modifications may be resorted to without departing ¦ from the spirlt and scope of the present invention. For I example, the embodiments of the present invention illustrated ¦ and described in conjunction with the single-phase source I may readily be modified to be driven by the three-phase ¦ source. Similarly, the embodiments illustrated and described ¦~ -I in conjunction with the three-phase source may readily be ¦ suited for use with polyphase sources having m phases ¦ where _ is greater than three (3). In the latter case, an ¦ m-phase AC voltages is applied to m main electrodes to cause 1 successively glow discharges between the pairs thereof.
The resulting power capacity is equal to m times that I provided by single-phase apparatus leading to inexpensive I ¦ structures. --, I ~

¦ . .. ~
r~ ;l .

~ I - 121 -:' I .
.' I . , , , ~ -: . , -

Claims

THE EMBODIMENTS OF THE INVENTION IN WHICH AN EXCLUSIVE
PROPERTY OR PRIVILEGE IS CLAIMED ARE DEFINED AS FOLLOWS:

1. An electric shock preventing device for a glow discharge heating apparatus for heating a heated liquid with the generation of heat attendant upon a glow discharge caused between a pair of electrodes through which a heated liquid such as water flows, characterized in that an insulating pipe is provided on each of an inflow port and an outflow port for the heated liquid on said electrode, said insulated pipe being provided on the extremity portion with a metallic pipe which is connected to ground.