US2493269A - High-frequency electrical heating apparatus - Google Patents

Description

Jan; 3, 1950 H. D. SIMPSON 2,493,269

HIGH-FREQUENCY ELECTRICAL HEATING APPARATUS Filed Jan. 25, 1947 is P INVENTOR.

HARRISON D. SIMPSON W ATTORNEY den. 3, 195% UNITED STATES OFFICE HIGH-FREQUENCY ELECTRICAL HEATING APPARATUS 6 Claims.

This invention relates to an oscillator circuit for radio frequency heating wherein vacuum tubes are utilized for 'producing high frequency current, and the circuit to be"later described is especiallya'dapted for use in heating small nonferrous metal articles, .such as are used in the jewelry trade. The use of high frequency heating ofiers 'greatiadva-ntagesinthe jewelry trade, since with its use higher quality-work can be done by unskilled operators-in=much less time than that required in torch brazing.

In the past, several types =of oscillator circuits have been used -for the heating of metals. This development began with the so calle'd bombarder's usedby manufacturers of vacuum tubes for heating the metal elements inside of the tubes for degasifying "during the evacuation process.

Later, the principle was further developed and used in heating iron :andsteel for annealing,

forging, hardening and brazing of these'metals. Ferrous metals draw energy very'readily from a high frequency alternating 'magnetic "field where hysteresis and" resistance losses result in rapid heating.

In applying the"-pr-inciple of high frequency heating to jewelry brazing, however, I found that these prior art methods were not at all suitable. Jewelry metals-such as gold, silver and brass have-a'low electricalresistance and no measurablehysteresis-loss. Hence they do not draw energy from a work coil unless both the field strength (amperes cr'turns) 'and the frequency are very high. T Under these circumstances, the 'work'coil' and" the rest of the tank quency oscillations are obtained by using spark gaps in the secondarycircuit of a high voltage transformer, in combination with capacitors,

.inductances and a work-coil. I'- have found that spark gap circuits do-notoperate efficiently at frequencies much above 200'- kilocycles, whereas,

2 in order to heat thin sections of non-ferrous material, such as wire .050 in diameter, an optimum frequency of approximately 3500 .kilccycles is necessary. .This type of circuit, requiring so many pieces ofapparatus, is not only expensive in the firstsplace, but alsoexpensive to operate, as thespark gapsmust be held-to a very close tolerance and consequently require-.quite frequent dressing. Eurthermore, spark gaps in a circuit of thiskind, operating at about '7 /2 kw., require -wa-ter cooling, which is not re- :quired in mycircuit except in the 'work coil.

In addition, such a spark gap oscillator operates at approximately 8500 volts-R. S. with :a peak voltage reachingsapproximately 12,000 volts. Using .such: a .high voltage has many objectionable featuresaswill be dealt withfurtheron. Furthermore,.I- have found that spark gap oscillators :donot operate at a constant frequency, nor do they .maintain undamped oscillations, and it is common knowledge that they interfere with radio broadcasting and reception. Another objectionable feature is that spark gap oscillators must bev provided with means to tune the work coil to the .generating part of the circuit.

Triode vacuum tubes aresused as oscillators, in other known types of radio frequency generators, which have disadvantages as will be shown.

Broadly speaking, the heating elfect of a work coil, isproportional to the magnetic flux (ampere .rturns) .multiplied by then. C. frequency. In the prior art a highmagnetic flux density was attempted by the following methods:

1. The most common type uses a work coil with a large number 10 to 30) of turns in series with a large tank inductance (5 to 10) microhenries) combined with relatively low frequency (200 to 400 kilocycles). This has the disadvantage of placing the work to be soldered virtually inside of atunnel where it is difiicult to manipulate or even to observe. Also, the low frequency used makes small non-ferrous work very slow in heating. .High frequency (1000 to 2000 kilocycles) cannot be used in this type because of the high total reactance of the tank coil and work coil. This type is not at all suited for hard soldering small non-ferrous articles.

2. An impedance-matching transformer is sometimes used in the above circuit, either of the mutual inductance type in the form of a copper sheet around the tank inductance, or an auto-transformer type, whereby the work coil is taken off of one or two turns of the tank inductance. In these cases, somewhat higher frequencies are used, 1000 to 2000 kilocycles. The disadvantages are as follows:

a. The impedances must be fairly accurately matched (which is not easily accomplished) in order to obtain even fair energy transfer.

. Air core transformers are not efficient at transferring high power at high frequency, due to coupling losses, which average 50%. Iron core transformers cannot be used because of their high impedance and hysteresis losses.

0. The coil constituting the main inductance in this case really has three functions to perform:

1. As a tank inductance.

2. As the primary to a transformer furnishing grid excitation.

3. As the primary of the impedance matching transformer.

As a practical proposition it is impossible to design one coil to carry out all these functions with any degree of efhciency, since each function calls for a coil of different characteristics for optimum results. Hence, some sort of a compromise coil must be employed.

. A typical inductance as above would consist of '7 turns of /4 copper tubing 5" in diameter. This coil will have an inductance of about 5 microhenries, and an inductive reactance of .55 ohms at 1750 kilocycles. The circulating current at 7500 volts will be 1.36 H. F. amperes.

The purpose of the transformer is, of course, to increase the circulating current in the work coil at the expense of lowering the voltage therein and in the leads thereto. This lowering of the voltage brings about serious losses in the entire secondary circuit.

The heating effect in thin sections of nonferrous metals is not good at 1750 kilocycles. If this should be raised to the eflicient range of 3500 kilocycles, the inductive reactance would rise to 110 ohms and the circulating current would be reduced to 68 amperes.

e. Due to the high reactance of this multifunctional. inductance, a high (7500) voltage must be used.

1. With the above type coil the ohmic resistance, due to the A. C. skin-effect is very high. At 3500 kilocycles, it would amount to about .027 ohm. At 136 amperes the PR. loss is .5 kw. in the inductance above or 20% of the entire output.

It is, therefore, the object of my present invention to utilize a circuit in which the objectionable features recited are substantially reduced or done away with entirely. To accomplish this, I eliminate the multifunctional large tank inductance coil entirely and make the work coil itself act as a main inductance. The losses at this point, which occur in the prior art structure, are substantially eliminated.

Thus, it is feasible to reduce the total inductance to about .4 microhenry, whereby the inductive reactance at 3500 kilocycles is only 8.8 ohms. (Compare this with the prior art inductance of 5.0 microhenries and inductive reactance of 110 ohms.) Hence, a circulating ourrent of 400 amperes can be obtained at 3500 volts. (Compared to 34 amperes in the prior art circuits at equal voltage.)

The circulating tank current follows the path denoted by the heavy black lines at 16 and I! in the drawing. This path is far shorter and more direct than any obtainable in prior art circuits. Since circulating currents of several hundred amperes are involved at this point 1 R losses are very important. In this new circuit from 50 to of the PR. loss is eliminated.

With the extra power available, the work does not have to be so closely coupled to the work coil, which permits the work to heat more evenly, and I have found that it is possible to heat small non-ferrous jewelry pieces to a temperature of 1400 F. in far less time (down to one second) than with any other prior art system which I have used.

Efficient grid excitation is obtained in my circuit by using a new special grid transformer which has only this single function to perform. The primary of this grid transformer is connected directly in parallel with the work coil and tank condenser, and since it has an impedance about 10 times as great as the work coil, diverts about 10% of the tank current through itself. Thus, it has only a slight effect on the frequency or circulating current in the work coil. The secondary of this grid transformer can be and is very closely coupled to the primary which furnishes grid excitation in correct phase relationship and of good wave form, which is important in induction heating.

This new circuit has several advantages as follows: Better heating effect is obtained and at lower cost. Simple one, two or three turn work coils can be used without an impedance matching transformer. Transformer coupling losses and the uncertainty of impedance matching are eliminated in the main heating current circuit. I R losses, due to ohmic resistance, are reduced as much as 80% because of the shorter path which the tank current has to follow. The circuit produces undamped oscillations of constant frequency.

The operating voltage is only about one-half as high as that required on other circuits, which is less dangerous to operating personnel. This also makes it possible to use a simple two tube rectifier circuit with or without thyratron phase shift to control output. Commercially available and less expensive plate transformers and triodes can be used. High voltage insulation difficulties such as arc-over, corona discharge and oil breakdown in the tank condenser are virtually eliminated.

The circuit shown in the single figure of the drawing illustrates the improvements that I have made as heretofore pointed out. In the circuit shown, I is a transformer having a primary P and a secondary S for delivering alternating current, from any suitable source such as 210 volt 60 cycle source of supply, to rectifier tubes 2 and 3 which are gas type. The primary P may have an ainmeter A connected in circuit therewith. The cathodes of the tubes 2 and 3 are supplied with current from the secondary 6 of a transformer 4 having a primary 5 connected to the same supply circuit as the primary P of transformer The center tup of winding 5 is joined by a conductor 1 to a radio frequency choke coil 8 to keep oscillations from the high frequency (H. F.) oscillating tube T, out of the source of direct current. The tube T has a plate or anode 9, a grid H1 and cathode llythecathodeis supplied with current from a transformer lihaving a primary i3 and asecondary [4. Also connected to the anode 9 of the tube T is a blocking condenser !5 to keep direct current out of the oscillator circuit which primarily. includes a tank condenser l6 and a heating coil 51, commonly called a work coil, which is connected in parallel with the tank condenser Hi. It will be noted that the, heating coil ll is composed of a small number of turns in the order of 1 to 3, 2 being lustrated. Theadvantages in use of this low number of turns has already been pointed out. I have obtained excellent results in the soldering of non-ferrous articles such as are used in.

jewelry by having a heating coil made of A tubing of good current-conducting material such as copper through which a cooling fluid may be passed. The heating coi I1 is preferably circular in form or it may be somewhat square shaped having its side length from 1" to 1%".

Also connected in parallel with thetank condenser lii and heating coil I! is the primary iii of a grid transformer the secondary iii of which has one end connected by conductor to the grid it of the tube 1 through a parasitic suppressor shown as composed of an inductance 2| and resistance 22. The other end of the secondary i9 is connected by conductor 23 to the cathode of tube T by way of a grid bias resistance 2% and by-pass condenser 25 of well-known characteristics.

It will be seen from the drawing that the secondary winding I9 is arranged, preferably turn-for-turn, concentric with the turns of the primary 88 but is spaced a suitable distance within the inner periphery of the primary l8. I have secured excellent results by making the primary of the grid transformer with 10 turns of copper tubing the internal diameter of the coil being approximately 4", while the secondary may be made of 10 or 12 turns of solid copper or silver wire having an outside diameter of 3%".

In order to better illustrate the advantages of the means described for heating metal parts, especially parts of non-ferrous material, I set forth the following data: In use of an oscillator circuit of the spark gap kind first referred to herein, when the input was 34 amperes at 208 volts. giving approximately 7.1 k. v. a. and holding the circuit closed to the heating coil for two seconds, a thermocouple connected to a pyrometer, gave a reading of 1800 F. Utilizing the circuit arrangement where the heating coil is in series with a primary of the transformer and with an input of 32 amperes at 208 volts or 6.65 k. v. a. in two seconds the pyrometer read 1900 F. With my improved circuit as shown and described herein, and with an input of 23 amperes at 208 volts 01' 4.7 k. v. a., in two seconds the pyrometer read 2660 F. These readings taken with the thermocouple in the same place within the heating coil and the same pyrometer show a great saving in power as well as a great saving in the time of heating the pieces that are to be soldered together.

While it is old in the art to use high frequency alternating current for heating purposes, it is only after long study and many tests that I was able to get the results as above set forth. Because of the elusive and unpredictable nature of high frequency phenomena, it was difficult to find the correct proportions for the grid transformer coils. Ordinary high frequency calculations were tried vention is the savingin the-costof .the apparatus itself. Becauseit'operates at lower voltage, uses low costcommercial components, such as are made in quantities for amateur broadcasting, and because it has fewer: and simpler parts, the total cost. is only one third to one-half of that of available apparatus of equal power. This saving in cost is important because each jewelry manufacturer can advantageously use a considerable number of complete high frequency generatingunits, since a' separate unit is required for each brazing operator.

Having thus described my invention, what I claimis:

1. Meansfor heating'metal parts including a H. F. oscillator tube with means for feeding direct current to the plate of the'tube. a tank condenser connected in the output circuit of the tube, a heating coil connected in parallel with the tank condenser and composed of few turns to facilitate the handling operation of the parts to be inserted into the space within said turns for heating them, a transformer having its primary connected in parallel with the tank condenser and heating coil and its secondary located within the magnetic field of the primary and connected to the grid of the tube and to the filament of said tube with a grid bias.

2. Electrical means for heating metal parts, including a high frequency oscillator tube and means for heating its filaments, means for feeding current to the plate of said oscillator tube, a blocking condenser connected to said plate, a tank condenser connected to and between said blocking condenser and the filament of said tube, a heating coil of few turns connected in parallel with said tank condenser, both said heating coil and said tank condenser being connected between the plate of said tube and ground, and a grid transformer having a primary of few turns connected in parallel with the tank condenser and the heating coil and having a secondary also of few turns connected to the grid of said tube and through a grid resistor and a grid condenser to the filament and to ground.

3. Means for heating metal parts, including a high frequency triode tube and means for heating its filament, means for supplying high voltage direct current to the plate of said tube through a radio frequency choke, a blocking condenser to keep high voltage direct current away from the oscillator circuit, a heating coil of few turns, a tank condenser in parallel with said heating coil, both said heating coil and said tank condenser being connected between the plate of the triode tube and ground, a grid transformer whose primary is connected in parallel with the tank condenser and whose secondary is connected to the grid of the triode and to the filament of said triode with a grid bias.

4. Means for heating metal parts as defined in claim 1, wherein said primary is made of good conducting material and has turns in number so its inductance as compared with the inductance of the heating coil will be such that less than twenty-five percent (25%) of the tank current will be diverted to the primary, and said secondary comprises turns nearly the same in number as the primary and is positioned within and in closely spaced relation to the inner surface of the primary.

5. Electrical means for heating metal parts as defined in claim 2, wherein said primary of the grid transformer has approximately eight (8) turns of copper tubing of a diameter of the order of three-sixteenth inch and the secondary comprises approximately ten (10) turns of wire of a diameter of the order of one-sixteenth inch.

6. Means for heating metal parts including a H. F. oscillator tube with means for feeding current to the plate of the tube, a tank condenser connected in the output circuit of the tube, a heating coil connected in parallel with the tank condenser and composed of a few turns to facilitate the handling operation of the parts to be inserted into the space within said turns for heating them, a transformer having a primary of few turns connected in parallel with the tank condenser and heating coil and a secondary also of few turns closely coupled to the primary and connected to the grid of the tube to excite said grid and produce a high circulating tank current and consequent high heating effect, with relatively low voltage.

HARRISON D. SIMPSON.

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

UNITED STATES PATENTS velopment Associates, 125 East 46th Street, New York 17, New York.

Images