US2673228A - Induction furnace with high-temperature resistor - Google Patents

Description

March 23, 1954 s, KlsTLER 2,673,228

INDUCTION FURNACE WITH HIGH-TEMPERATURE RESISTOR Filed Sept. 15, 1950 Fig.1

2 Sheets-Sheet l March 23, 1954 s, s, K|$TLER 2,673,228

INDUCTION FURNACE WITH HIGH-TEMPERATURE RESISTOR Filed Sept. 15, 1950 2 Sheets-Sheet 2 TEMPERATURE(DEG. K) I000 \2 0O I4 OO IQ OO I6 OO 20%8 AMBlENT/gFEMP VOLTAGE/CENTIMETER looo AMBIENT TEMF. I600 K I l l I 1 l 1 1000 \500 e000 2200 2400 00 0 TEMPERATURE DEG. K)

- 2200 CR|TICAL TEMPERATURE (DEG. K) I600- IO0o 600 l I I I I I I I I I 6 600 1000 I200 I400 1000 \000 AMBIENT TEMPERATURUDEEK) fig 5 4 HEAT WATTS) 3o [Wmtvgtor 5/9MUEL 5T K/STLER 00 00100 I moo l8 0 57%I7FGWM7I7 O0 AMBIENT TEMPERATURE (DEG. K) y Patented Mar. 23, 1954 UNITED STATES PATENT OFFICE INDUCTION FURNA TEMPERATUR 3 Samuel S. Kistler, West Boylston, to Norton Company, Worcester,

Massachus poration of 6 Claims.

The invention relates to high temperature electrical heating resistors.

One object of the invention is to provide a satisfactory high temperature furnace without Figure 6 is a graph of the heat radiated per square centimeter of a zirconia resistor at the critical temperature for different ambient temperatures.

As conducive to a. clearer understanding of certain features of the present invention it is pointed CE WITH HIGH- E RESISTOR Mass., assignor Mass, a coretts Application September 15, 1950, Serial No.

out that while it has long been known that many oxides are electrical conductors at atures that it must be heated by another source of heat before it Will operate. Hence when used strip, with consequent failure. Other difiiculties have heretofore been encountered. One of the objects of this invention is to overcome such disadvantages and difiiculties.

I have found that sintered zirconia blocks give very good results in actual practice.

not found it necessary to cement them together; they can be merely .aid together on any fiat surface.

Supported by the supporting blocks H) is a column of zirconium oxide resistor rings I i which ably not cemented together. The bottom ring i i rests upon the support ng blocks H3 and the top ring ll supports a furnace cap l2 having a sight hole l3. This furnace cap 12 is also preferably made of sintered zirconia but can be made of other refractory material. It is cylindrical with two diameters as shown, the smaller diameter portion extending downward into the space bounded by the hollow cylinder formed of the rings II, for example to a depth of a little over one ring and leaving. a reasonable clearance between itself and the rings, all as clearly illustrated in Figure 1.

Outside of and spaced from the cylinder of rings l I is a hollow cylinder l of refractory material, preferably of sintered alumina although other material could be used. This cylinder I5 has a height the same as that of the cylinder of rings l3. Between the hollow cylinder l5 and the cylinder of rings H, is a packing of zirconia grain l5. Fitting the outside of the hollow cylinder i5 is an asbestos sleeve 11. Outside of and spaced from the asbestos sleeve 11 is a coil N3 of copper tubing which is connected to a high frequency induction heater, for example a fifteen kilowatt, megacycle oscillator. In these inductionheating coils cooling water flows through the tubing. I preferably provide a sintered zirconia. hearth block 20 resting on the supporting blocks Hi. This may be a cylinder of a diameter somewhat less than the inside diameter of the rings H to allow a reasonable clearance as shown.

Especially for t .e resistor rings H and the hearth block 20 and preferably for the supporting blocks 10, the cap l2, and desirably for the zirconia grain, I use stabilized zirconia ZrOz pref erably stabilized with from 3% to 6% lime, CaO, and the material may be made in the following manner. An electric arc furnace of the type disclosed in U. S. Letters Patent No. 775,654 patented November 22, 1904 to Aldus C. Higgins is provided. Furnaces of this type comprising iron shells cooled all over with a cascade of water have been in use practically ever sincethe date of the above patent and are well known to electro-chemists and therefore need not be further described herein. A furnace mixture of zirconia ore, coke, iron, carbon, and lime (CaO) is prepared. Various zirconia ores or partially purifled zirconia powder can be used. In general these are the zircon and baddeleyite ores. "However chemically purified zirconia can equally well be used but is of course more expensive.

Thequantity of carbon provided in the furnace mixture should be two-thirds of the theoretical quantity of carbon required completely to reduce the silica plus 100% of the theoretical quantity required to reduce all the other oxides (except the zirconia) to metal plus about 20% excess over all of these quantities. This quantity can be varied from the above with no excess to the above or with 40% excess. The reason why only two-thirds of the theoretical quantity of carbon required completely to reduce the silica is provided is that about one-third of the silica is volatilized during the furnacing operation. On the other hand the excess mentioned is provided because some of the coke is used up by combining with oxygen other than that provided by the oxides to be reduced.

The quantity of iron should be enough to form with the silicon that is reduced from silica a ferro-silicon havingv an iron content of from 75% 1.085%. The purpose of the iron is to combine with the silicon to form a ferro -silicon alloy which 'hasa much higher specific: gravity than elementary silicon and therefore'will go to the bottom of the :fumace and, after solidification, form a.

ferro-silicon button containing also other reduction products than can readily be separated from the rest of theingot. The amount of iron to add is enough to make with two-thirds of the silicon present in the ore a ferro-silicon having an iron content of from 75% to minus the amount of iron obtained by the reduction of the iron oxide in the ore to iron and this of course must take into account that a small percentage of iron oxide remains. in the final product.

The quantity of lime as a stabilizing agent to be added should be from 3 to 6% of the amount of ZrOe in the ore. The reason for providing the stabilizing agent in the above percentages is that less will not satisfactorily stabilize the zirconia, and more will form a eutectic thus making the product less refractory. The stabilizing agent in the range given causes the zirconia to crystallize predominantly in the cubic system. but when less of the stabilizing agent is used the crystals are predominantly monoclinic. Ordinary or natural baddeleyite is monoclinic whereas the product of this invention is predominantly cubic. The monoclinic form of zirconia will notwithstand many cycles of heating to over 2000 C. and cooling to C. and even lessertemperature changes may cause cracking or fracturing of the heating element if it is made of monoclinic zirconia. A- zirconia of predominantly cubic crystal form will, however, withstand heat shock for many cycles. When the lime is as much as 6% the crystals are nearly all cubic, when the lime is as low as 2.7% about 35% of the crystals are cubic. Zirconia having 3% to 6% of lime on the ZrOz is referred to as stabilized zirconia.; the expression stabilized means that the zirconia does not have a detrimental volume change at the critical temperature of inversion of baddeleyite, normally about 1000 C. However I do not want to belimited to lime stabilized zirconia nor to the precise furnace operation described herein since magnesium oxide (3% to 6%) is likewise a stabilizing agent and other processes of combining the stabilizing agent with the zirconia can be effectively used. However the best material now known to me for the manufacture of the resistor heating. elements isthat above described.

For the manufacture of the supporting blocks 10, the hearth block 20 and the furnace cap 12 I may take zirconia grain, press it in a mold to produce the desired shape, and then fire it. I may proceed in the same manner to form the rings H and they can be ground to required dimensions after firing. Any suitable temporary binder, for example dextrine, can be used to form the green shapes whichare preferably fired at cone 35. An illustrative example is to wet the grain with a solution of 2% water and 1% dextrine, the percentages'being on the weight of the zirconia.

At ordinary temperatures the hollow cylindrical resistor, preferably made up of the series of superimposed rings H, is non-conductive or insumciently so to respond to the high frequency field created by the induction coil [8 which, as above noted, is connected to any suitable source (not shown) of high frequency energy, preferably a vacuum tube type of oscillator provided as usual with any suitable form of controls for varying at will or regulating the magnitude of the energy supplied to its load and preferably provided, also, with any suitable means for changing at will and for regulating the frequency of its output. Accordingly, the resistor is first heated up to conducting temperature, as by a gas flame or by a temporarily inserted auxiliary electric resistance unit such as a resistor rod of bonded silicon carbide energized from an extraneous appropriate source of current, but preferably its initial or preliminary heating is effected by placing in the furnace a refractory crucible filled with graphite powder and then energizing the induction coil !3 from the high frequency source of energy. The graphite powder being conductive, responds to the rapidly alternating field created by the coil it; being finely divided or powdered, it is, in contrast to a solid cylinder of graphite, a sufiiciently high resistance and the energy dissipated therein in the form of heat is sufficient to heat the oxide resistor H-ll. When the temperature of the latter has risen to around 1400 C., the oxide has a sufiiciently low resistance or a sufficiently high conductivity to absorb energy from the oscillating field of the winding [8 and the crucible can then be lifted out.

During the stage of preparatory heating, as by the inserted crucible, care should be taken, as by controlling or regulating the frequency or energy output of the oscillator, to achieve relatively slow or gradual rise in temperature of the resistor H-H in order to effect as uniform heating as possible throughout its mass. In this manner each transverse section of the hollow tube-like resistor |lll, Whether subdivided into rings II or not, has substantially the same resistance, and as heating progresses with such controlled uniformity, decreases in resistance or increases in conductivity of the several sections take place at the same rate so as to maintain uniformity throughout. Each transverse section, looped by the oscillating field of the coil Iii, can then respond substantially in the same measure or degree. Likewise, the dielectric losses in each section should be the same and increase at the same rate.

Illustrative dimensions of the zirconia heater tube ll-H are a length of 8 inches, an outside diameter of 5 inches, and an inside diameter of 3.5 inches, giving a wall thickness of 0.75 inch. Where, as is preferred, the stabilized zirconia tube l|-ll is physically subdivided into transverse sections or rings l|-l I, etc., as shown in Figure 1, an illustrative number of rings for the just stated dimensions is eight, giving each an axial length of one inch. These rings rest upon one another, and preferably their upper and lower faces are nicely plane, so that successive rings make good thermal contact with each other; illustratively, these faces may be ground fiat. Also, as above noted, the rings ii need not be secured one to the other, as by cement, and preferably they simply rest in flatwise engagement with each other. Illustratively, the coil i8 may be of %-inch copper tubing and may comprise about five and a half or six turns.

The dimensional relations between diameter and wall thickness of the zirconia resistor tube H-H are preferably predetermined according to certain features of my invention so that the wall thickness of the zirconia tube is not too thick relative to the diameter of the tube and meets the criterion that Wa is not less than 2 where ID is the inside diameter of the tube and Wu is the wall thickness. This preferred aspect of my invention will presently be further amplified. It is'a factor. that contributes toward overcoming limitations and difiiculties heretofore encountered and toward achieving a number of practical advantages and results here tofore impossible so far as I am aware.

In order better to understand certain features of my invention, it wil be assumed, in the following preliminary discussion, that the dividing planes between rings H, il, H, etc., are imaginary and that we are considering a single unitary hollow or tube-like oxide resistor. Let it be further assumed that this unitary resistor tube is energized by an alternating current (say, 60 cycles) through electrodes making uniform surface contact therewith at its upper and lower end faces whereby substantial uniformity of current flow per unit area of the cross section of the resistor tube is substantially achieved. Such a resistor tube, being a hollow cylinder, may be regarded for analytical purposes as comprising a number of individual longitudinal resistor elements arranged in a circle and in intimate side wall contact with each other and each carrying the same amount of heating current therethrough, and if the resistor tubes are dimensionally within the above criterion and certain preferred controls effected, it is possible to achieve operation of such a resistor and of the furnace without materially disturbing the just described uniformity of current flow through these imagi nary longitudinal subdivisions or elements of the hollow cylinder.

If a constant potential is applied to the opposite ends of such a tubular resistor and it is heated by some suitable means until the resistor becomes conducting, the conditions are in general unstable and, without more, the temperature will increase until the element is destroyed. Qn the other hand, if the applied voltage is so controlled that the temperature does not run away, the likelihood is that one side of the tube will become warmer than the other parts, thus becoming a better conductor with increase in current therethrough, which in turn leads to still higher temperature until virtually all of the current is channeling down one side of the tube, which usually results in breakage. It is this sort of thing that has been one of a number of major factors heretofore limiting the uses of conductive oxides and preventing employment in heating elements of large surface area.

According to certain features of my invention, these obstacles are overcome. I have investiated the factors involved therein and as of aid in describing how I overcome the above obstacles, let it first be assumed that an elongated oxide resistor element, sufficiently small in diameter to heat uniformly throughout the cross section (say, on the order of 2 millimeters) is enclosed in a furnace maintained at a constant temperature To. If a small current is passed through the element, heating it to some temperature T above that of the furnace, it Will lose heat to the furnace through convection and radiation. At least. as a first approximation, one can reasonably represent the heat loss by the equation in which H represents the heat lost per square centimeter in watts, k is the heat transfer 00- efiicient through convection, and k is the heat transfer coefiicient through radiation. Temperatures are on the Kelvin scale. As a first approximation, I have taken the values of k and k as 0.002 and 1.75 times 10 watts. For practical purposes, at temperatures of interest in this art, the first term in the equation can .be meglected. In calculating k it was assumed that the emissivity of zirconia is 30% of black body. This figure is basedupon measurements made in a laboratory.

Accordingly, the heat lost from the heatin element at any temperature above the furnace temperature can now be calculated. Knowing the resistance of the oxide at any temperature, the E. M. F. required to supply the heat lost can be calculated. Such calculations have been made for a stabilized zirconia tube of large bore with a wall thickness of 2 millimeters (in effect representinga series of small-cross-sectioned elementsarranged in a circle), and the results :are

plotted in Figure 4 for two furnace temperatures, 11000 K. (abscissae at the top) and l600 K. (abscissae at the bottom). It is assumed, for the moment and for purposes of this analysis,

thatthis very thin-walled tube remains at uniform temperature throughout its cross section.

The specificresistance of lime-stabilized zirconia has been measured over a wide temperature range, and while there are variations due to impurities, porosity, and grain size distribution, it can be represented reasonably well by the equation o T sea (2) Log R= maximum, although the one for To1000 K. rises much more steeply and attains its maximum much sooner than the other. At any temperature and constant voltage to the left of the maximum point the element will be stable, any fluctuation causing a higher temperature resulting in more energy being radiated than is generated, and the temperature will fall back to the equilibriumposition. On the other hand, any potential and temperature to the right of the maximum point will lead to eventual run-away conditions, since any accidental displacement of the temperature upwards will cause a greater generation of heat than can be radiated and the temperature will rise at a progressively faster rate until fusion occurs.

Furthermore, the element is permanently unstable at any voltage above the maximum point. In this connection it should be pointed out that the voltage at the maximum point decreases as To in increased, so that a potential that is safe at a low furnace temperature may become excessive as thefurnace heats up.

The relation between the critical temperature (i.-e., the temperature at maximum voltage) and the'furnace temperature is given in Figure 5 and the following equation Inthis equation, 13 comes from the variant of Equation 2,

RzAe (4) in which, for stabilized zirconia, A:0.00152, and B:13,'700. Note that A, which contains such factors as porosity, grain size, shape of cross section of the element, and magnitude of the specific resistance, does not appear in Equation 3. It is inconsequential, therefore, from a stability standpoint, how the heating element is formed as long as the activation energy of the conductionprocess is'not affected.

Figure v5 shows that at low temperaturethe critical temperature is very little above the 1ambient temperature, thus providing only a narrow regionof stability, while at high furnace temperatures the stability limits are very wide. At furnace temperatures above 1800 K. instability is no longer possible with a voltage below that necessary to melt the zirconia element.

In Figure 6 is plotted the heat radiated froma square centimeter of zirconia at the critical temperature as a function of ambient temperature. It shows more strikingly than does Figure 5 that at low furnace temperatures the safe radiation from a heater is very low, while with high temperatures the performance per square centimeter can be large. This indicates that a large radiating surface per unit of heat output is very desirable.

Thus critical temperature and safe radiation desiderata are determinable for the oxide resistor whose wall thickness should not be great enough to permit detrimental or excessive radial temperature gradients, and in illustrating ,thisaspect of my invention, specific criteria are above illustratively set forth for a stabilized zirconia element. No such radial gradients are present when, as has been assumed above for purposes of analysis, the assumed heating element is a solid cylindrical element so small in diameter and hence so small in cross section that its temperature is uniform throughout its cross section. Nor are they present in the large-bore thin-walled tube used in some of i the above calculations. The latter, a large-bore thin-walled tubular heating element, is in effect an assembly of smallcross-sectioned parallel elements equally "and uniformly energized electrically, and the conditions for stability within that assembly so energized are the same as those described above. Moreover, the thin-walled tube used inthese calculations also meets the criterion for wall thickness earlier above set forth; the wall thickness is not great enough to permit excessive radial temperature gradients. The conditions'for stability apply equally well to the stability of a multiplicity of identical heating elements connected in parallel in a furnace. It has been amply confirmed in experimental furnaces.

The stacked-ring zirconia heater tube il-H above described meets the criterion of wall thickness above set forth, and were the current to flow into and out of its ends at uniform current density throughout its end faces or surfaces, after preliminary heating has made the heater tube sufficiently conductive, the flow of current from one end to the other wouldtake place at substantially uniform density throughout any cross section taken between the ends of the heater tube. It functions in effect like a set of small-cross-sectioned individual elements arranged integrally in a circle, and because the above relationship of wall thickness to inside diameter is not departed from, if any oneof these imaginary elements or lengthwise subdivisions of the zirconia tube heats up beyond the others, it can and does lose more heat by radiation than it gains, and thus the conditions that are otherwise conducive to channeling are opposed. 'Thus continued heating up of the heater tube is facilitated, as is also subsequent operation of the furnace at the desired ultimate furnace temperature.

While the above mathematical derivationsof the conditions for stability "apply to a resistor radiating to furnacewalls, the same conclusions can be reached where the .outside of the tubular resistor is insulated .and radiation is across the space in the interior.

This action of continued heating up after the tube becomes conducting will thus also be seen to contribute toward maintaining uniform current density in a tube so energized, and thus uniform heating up is more reliably and more quickly attained. The action just above described might be said to achieve, under otherwise stable conditions, substantially automatic maintenance of temperature equilibrium throughout the mass of the heater, and a similar substantially automatic action takes place during the run of the furnace at its ultimate furnace temperature at which treatment of articles or materials is desired to take place.

It is preferred to supply energy to the hollow cylindrical element of the furnace under conditions of constant current control, and where the energy is supplied to its ends by suitable electrodes, this may be done as in my copending applications, Serial No. 184,926 and Serial No. 184,927, filed of even date herewith, and to which reference may be made, where I described a. suitable current regulator of the supply system to 9 the desired current value, the regulator being manually set to function at the selected standard or value of current to be kept constant. As there described, it is possible, once the heate has been heated to become sufficiently conductive, to set the regulator at a. current value that will ultimately give the desired operating furnace temperature upon completion of the heating up by the current itself, for the above-described dimensional criterion or characteristic of the zirconia tubular heater and the limitation of current or amperage effected by the current regulator coact to achieve prevention of conditions of instability such as those described above in connection with Figure 4. coactions and by setting the current regulator to successively different standards of operation, it is also possible to supply current to the heater tube at a different value during the heating-up period than is supplied during the continued operation of the furnace at its desired temperature, so long as the change in value is effected with due regard to the conditions of stability and of instability that exist, respectively, to the left and to the right of the maximum points of the graphs, such as those of Figure 4, for the heater element. Because of the Wide range of stability that exists to the left of the knee in the graphs, preciseness and closeness of control can be departed from and energization, throughout that range of stability, could even be effected at higher current values or at varying (e. g., increasing) current values, and in this latter connection the energy could be applied in successive stages at successively higher values so long as the relatively high critical voltages (to the left of the maximum point in the graphs) are not reached or exceeded for any current value employed, and thus channeling is opposed further during the heating-up period. But substantial precision and closeness of control of'constancy of current should be instituted just before or when the critical temperature is reached, for thereafter, as above explained and demonstrated, conditions of instability can be brought about at potentials and temperatures to the right of the maximum point in the corresponding graph like that of Figure 4. In the latter range (to the right), because the resistance of the oxide element continues to decrease, at given constant current flow Because of such 10 is accompanied by continual decrease in voltage, and the standard of current regulation can be raised to achieve higher temperatures in the furnace, but always within the just described voltage and temperature limits.

It is according to the principles and within the methods of energization just described that the furnace structure of Figure l is operated, and by manipulating or setting the various regulating controls of the high-frequency oscillator, continued heating up of the zirconia tube element f fl i is effected within the limits of the factors of stability to the left of the maximum point, in effect maintaining the effective induced voltage below the critical limit; as earlier pointed out, the heating-up process should also be effected slowly in order thereby, also, to maintain uniform rise in temperature throughout the mass of the oxide resistor element. In this manner channeling during heating up can be better guarded against. In assembling the induction coil 18 relative to the oxide heater tube ll-f I, both being of circular cross section, they are positioned (the coil if; being supported by any suitable means, not shown) so that they are coaxial and so that the one is substantially uniformly distributed lengthwise relative to the other, at least throughout the effective length of the interior furnace space, and the winding positioned relatively close to the heater tube (about in the relationship shown in Figure 1 for the above given dimensions of the heater tube), so as to provide an appropriate coupling between the two to transfer energy from the coil to the conductive resistor element, having due regard for the need of heat insulation between the two as is provided by the parts I6, l5, and ll of Figure 1.

The operation of this furnace has been based upon the assumption that uniform current flows axially through the walls of the heating cylinder;

the same conclusions can be reached when, as in this case, the current flows tangentially around the cylinder.

Nevertheless, during the heating-up process there may be a tendency for one transverse section of the tubular oxide element to heat up more, or more rapidly, than another. There are some differences in the inductive effects on the different rings, differences in heating loss, viz., the end rings lose heat more rapidly than the center, and it is at times difiicult during the early heating-up period to apply the alternating magnetic field to such intensity as to raise the temperature but still not to exceed the critical conditions for any of the rings. On account of these causes and possibly some variations in resistance from ring to ring, heating during the early period may be non-uniform. Whatever the cause, I have found that the effect is to break the oxide tubular element were it to be made of a single piece throughout its length, fracture occurring usually in the general direction lengthwise of the tube. appar- I ently because one transverse portion expands more rapidly than the other; breakdown of furnace operation may result.

But I have also found that by physically and actually sectionalizing the oxide heater tube ll--H into transverse rings or sections II, II, I I, etc., I am enabled to achieve efficient continued heating up and subsequent operation of the furnace at the desired ultimate furnace temperature. With such a succession of rings yieldably held in face-to-face and end-to-end contact, as by stacking one upon the other, each is free to expand individually of the other and 11 the, operational integrity of the physically sub dividedoxide heater tube is maintained. I have; found,v as a result, that as the temperature rises during the completion of heating up, any uneven heating of the above-mentioned kind becomes less and less evident, thus, showing that the action of the physical subdivision into rings progressively counteracts more and. more the causesthat give rise to such uneven heating, and by-thetime the selected furnace operating temperature is reached, which illustratively may be 2009 C. or even up to: 2600 C., the tubular wall structure. shows a remarkable uniformity of temperature throughout. And it so continues during the. continuedfurnace run.

Moreover, I, have found that these same beneficial advantages and results take place even if, duringthe heating-up process, one or more of therings ll become cracked. Not only does the laminated or built-up oxide. heater tube remainphysically intact and serves, as in the illustrative embodiment of Figure 1, also as the furnace chamber to receive, on the hearthblock 20, productsto be treated, but also no detrimental. heating action is preceptible with respect toany cracked ringlor rings. This probably is because, even if circulatory currents were to be lessened or halted by the crack through the ring, eddy currents and some dielectric heat losses wouldstill: continueto be effective throughout the masaof the ring, or perhaps also by the fact that the circuit of the crackedring is maintained closed by the adjacent face-to-face contacting ring or rings. Moreover, it appears clear, from tests, that a. crack or cracks in any ring do not interrupt the path. of-fiow of current around and within the ring, because, at the high temperatures involved, there is aheavy electron emission from theoxide and any gap that results from the crack,

is highly conductive. Moreover, also, any such crack is usually very small and, hence, the gap very narrow.

By the way of illustration, with the above-described furnace structure, a temperature of 2170? C. was attained with an energy of about 5 kw.. input from the above-described 1 5 kw. 1 megacycle oscillator. In the range of such high temperatures, operation is very smooth and there isno sign of instability.

The, above temperature of 2170 C., attained with an input of, approximately kw., was at a frequency of l0.rnegacycles and was achieved with the existing coupling coefficient between the above-described and dimensioned furnace structure of Figure 1. Higher power inputs and hence higher temperatures are achievable by improving the coupling coefficient of above or by balancing of the, circuit with respect to the load. The optimum frequency at the just-mentioned temperature is much below megacycles and is about 3 megacycles, and I have found that the applied frequency can be rather widely removed from the optimum and effective functioning of the furnace still achieved. I have found that the type and character of coupling effected as above. described in the structure of Figure 1, apparently cause the coupling to act in the high- ,temperature or furnace operating range to decrease energy transfer from the induction coil it as the temperature and conductivity of the zirconiahresistor element increase, and because the, induced current increases, with reduced resistance the inductive reaction. upon the induction coil l8,in,eifect loading it, serves to diminish,

voltage. as the current increases and thus main.-

I2 tain the operation well withinzthe. temperature and voltage limits of stability as; explained above with respect to. the region to the right ofvthe maximum point in the graphs of. Figure 4;

Thus: it=will beseen that, by my invention, many thoroughly practical advantages are successfully achieved. Various risks, damage, losses, and limitations imposed: or caused by channelingv can be successfully lessened. and controlled stability achieved according to theprinciples of my invention. While I have described my invention in connection withthe use of zirconia resistor ele:- ments, and more particularly stabilized zirconiaa heater elements, I have done so; because ofithe: over-all superiority and many novel advantages that I am enabled thereby to achieve, but Igdo: not wish to, be limitedthereto except assuch limitation maybe expressed inone-or moreof the claims. Many of the advantages and various of the results of my invention may be achieved by utilizing other known refractory oxides that are conducting at high enough temperatures. Of course, elements; of expense or cost of some of these will enter as a consideration in theirpractical application. Also, theparticular characteristics of any such other refractory oxide might make it more suitable forv some particular purpose; for example, for very high temperatures, thoria may beemployed in preference to. zirconiav because, for example, it has a higher melting point. Whatever the selected refractory oxide, stability of operation may be achieved according; to the principles fully explained and illustrated above inconnection with the use of stabilized zirconia, as, for example, by shaping it in the, form of a tube in which the relation: between the inside diameter and wall thickness is such that detrimental or excessive radial temperature gradients. are not produced, andsupplying elec.- trical energy thereto so controlled that ambient temperature and resistor temperature are maintained within thoseratios that mean stability, as. is illustratively explained above in connection withstabilized zirconia tubular resistors.

It will also: be seen that some known disadvantages or limitations, met with heretofore are overcome, according to my invention,,by the use. of zirconia tubes instead of solid rods, othersby, the use of stabilized zirconia, others. by the. use of laminatedor sectionalized oxide tubes (like the rings H, II,- etc.), still others by an interrelated ring sections andinduction heater coil (like. coil. l8), and still others by constant-current supply or current-limiting regulation, all as illustratively above described.

It will thus be seen that there has beenprovided by this invention a high-temperature resister, in which the various, objects hereinabove set forth, together with many thoroughly practi-- cal advantages,v aresuccessfully achieved. As

many possible, embodiments may be made of the invention andras many, changes might be made in the embodiment above set forth; it is to be understood that all matter hereinbefore set forth.

or, shown in the accompanyingqdrawlngs isto be;

interpreted as illustrative-and not in a limiting;

sense.

I, claim:

1. A high temperature electric" furnace: comprising; a: refractory enclosure, an: induction coil.

shaped resistor elements within said enclosure and substantially coaxial with said induction coil and made of a refractory oxide that is selected from the group consisting of zirconia and thoria, stabilized, containing from 3% to 6% of alkaline oxide selected from the group consisting of lime and magnesia and that has to be heated to be" come significantly conductive and that has a negative temperature ooefficient of resistance, said plurality of ring-shaped resistor elements as a whole forming a hollow cylinder that has a ratio of surface area to cross-sectional area greater than that at which, for a given uniform temperature at which it is conductive, heat is gen-- erated internally thereof, in response to induced current flow therein, faster than the cylinder can lose heat through its surface, and means for raising the temperature of said ring-shaped resistor elements to become inductively affected by the field produced by said induction coil, said induction coil being adapted to induce substantially constant value of current in said refractory oxide secondary at relatively low voltage within the voltage factor of stability of the refractory oxide for the temperature of continued furnace operation.

2. A high temperature electric furnace comprising an induction coil adapted to be energized from a source of high-frequency alternating current, and a substantially non-channeling refractory oxide multi-turn secondary for said coil in the form of a plurality of individually unitary and continuous contacting ring-shaped pieces of a refractory oxide having a negative temperature coefiicient and having contacting end surfaces each of relatively large area compared to the radial cross-sectional area of the ring-shaped piece, each being of the same diameter, inside and outside, and the ratio of the inside diameter of said pieces to the wall thickness thereof in a radial direction being not less than two, said ringshaped pieces being arranged coaxially and surrounded by said induction coil and with end surfaces of adjacent pieces in conductive contact whereby, upon radial fracture of any one ringshaped piece, its adjacent conductively-contacting ring-shaped piece maintains continuity of conductivity of the fractured piece in circumferential direction.

3. A high temperature electric furnace comprising an induction coil adapted to be energized from a source of high-frequency alternating current, and a substantially non-channeling refractory oxide multi-turn secondary for said coil in the form of a plurality of individually unitary and continuous contacting ring-shaped elements made of a refractory oxide selected from the group consisting of zirconia and thoria, stabilized, containing from 3% to 6% of alkaline oxide selected from the group consisting of lime and magnesia, said ring-shaped elements having end surfaces of a shape adapted to engage with the end surface of an adjacent ring-shaped element for conductive contact therewith, said ringshaped elements being of the same diameter, inside and outside, and each having a ratio of surface area to cross-sectional area greater than that at which, for a given uniform temperature at which they are conductive, heat is generated therein, in response to electric current flow, faster than heat can be lost through said surface area.

4. A substantially non-channeling multi-turn refractory oxide secondary for an induction coil high temperature electric furnace, said secondary consisting of a plurality of indvidually unitary and continuous ring-shaped elements made of a refractory oxide selected from the group consisting of zirconia and thoria, stabilized, containing from 3% to 6% of alkaline oxide selected from the group consisting of lime and magnesia, said ring-shaped elements having end surfaces of a shape adapted to engage with the end surface of an adjacent ring-shaped element for conductive contact therewith, said ring-shaped elements being of the same diameter, inside and outside, and each having a ratio of surface area to crosssectional area greater than that at which, for

a given uniform temperature at which they are conductive, heat is generated therein, in response to electric current flow, faster than heat can be lost through said surface area.

5. A resistor for a high temperature electric furnace comprising a plurality of individually unitary and continuous ring-shaped resistor elements each made of a refractory oxide that comprises principally the reaction product under heat of zirconia with lime of a quantity on the order of from 3% to 6% of the amount of the zirconia, and that has to be heated to become significantly conductive and each having annular end faces for making surface contact with each other when arranged coaxially, said resistor elements having a ratio of surface area to cross sectional area greater than that at which, for a given uniform temperature at which they are conductive, heat is generated therein, in response to electric current flow, faster than heat can be lost through said surface area.

6. A resistor for a high temperature electric furnace comprising a plurality of individually unitary and continuous ring-shaped resistor elements each made of a refractory oxide that is selected from the group consisting of zirconia and thoria, stabilized, containing from 3% to 6% of alkaline oxide selected from the group consisting of lime and magnesia, and that has to be heated to become significantly conductive and each having annular end faces for making surface contact with each other when arranged coaxially, said resistor elements having a ratio of surface area to cross sectional area greater than that at which, for a given uniform temperature at which they are conductive, heat is generated therein, in response to electric current flow, faster than heat can be lost through said surface area.

SAMUEL S. KISTLER. References Cited in the file of this patent UNITED STATES PATENTS Number Name Date 652,640 Potter June 26, 1900 684,296 Nernst et a1. Oct. 8, 1901 775,654 Higgins Nov. 22, 1904 847,003 Von Iscewsky Mar. 12, 1907 1,091,808 Calhane Mar. 31, 1914 1,132,684 Queneau Mar. 23, 1915 1,378,189 Northrup May 1'7, 1921 1,438,936 Elmer Dec. 12, 1922 1,456,891 Little May 29, 1923 1,470,195 De Roiboul Oct. 9, 1923 1,572,881 Brace Feb. 16, 1926 1,799,102 Kelley Mar. 31, 1931 1,830,481 Northrup Nov. 3, 1931 1,851,984 Rennerfeld Apr. 5, 1932 2,231,723 Jung et al Feb. 11, 1941 2,291,532 Clark July 28, 1942 2,516,570 Hartwig et a1. July 25, 1950 OTHER REFERENCES Materials and Methods, vol. 33, No. 3, March 1951, page 81.

Images