US3242381A - Ballast apparatus for operating fluorescent lamps and electrical coil assemblies therefor - Google Patents

Description

March 22, 1966 P. w. DAVIS, JR 3,242,381

BALLAST APPARATUS FOR OPERATING FLUORESCENT LAMPS AND ELECTRICAL COIL ASSEMBLIES THEREFOR 4 Sheets-Sheet 1 Filed Jan. 2, 1963 INVENTOR.

Pau/ WDawlsJn 7& M

ATTORNEY March 22, 1966 p, w s, JR. 3,242,381

BALLAST APPARATUS FOR OPERATING FLUORESCENT LAMPS AND ELECTRICAL COIL ASSEMBLIES THEREFOR Filed Jan. 2, 1963 4 Sheets-Sheet 2 E lrs 5 45 40 o '0 g T 2 Al ii s 5 47 52- 49 s3 l OVEN f0 I I 55 i j; \A L2 5|- L50 42 L| R I s R W l 46 OSCILLOSCOPE R\ ElE E BEAT FREQUENCY OSCILLATOR 66 R4 0 C YQIVA'AVA'AVA 4 A INVENTOR. 1%41/ MZDaV/s-In ATTORNEY March 22. 1966 P. w. DAVIS. JR 3,242,331

BALLAST APPARATUS FOR OPERATING FLUORESCENT LAMPS Filed Jan. 2; 1963 AND ELECTRICAL COIL ASSEMBLIES THEREFOR 4 Sheets-Sheet 3 o o o 160 2 00 TIME- MINUTES IOO INVENTOR. Pau/ Wfia |//$,Jr: 7% 77am ris 26o TEMPERATURE-CENTIGRADE .OOl

ATTORNEY March 22. 1966 P. w. DAVIS. JR 3,242,331

BALLAST APPARATUS FOR OPERATING FLUORESCENT LAMPS AND ELECTRICAL COIL ASSEMBLIES THEREFOR Filed Jan. 2, 1963 4 Sheets-Sheet A.

INVENTOR. Pau/WDaWSJn BY 7 147 74M ATTORN EY United States Patent 3 242 381 BALLAST APPARATUS oR OPERATING FLUO- RESCENT'LAMPS AND ELECTRICAL COIL AS- SEMBLIES THEREFOR 7 Paul W.'Davis, Jr., Danville, Ill., assignor'to General Electric Company, a corporation of New York Filed Jan. 2, 1963,Ser'. No. 249,099

8 Claims. 1 (Cl. 315-487) This invention relates to ballast apparatus and to improved electrical coil assemblies for use in ballast apparatus for operating fluorescent lamps.

In electrical coils used in ballast transformers layer insulation is generally provided between the layers of conductors. Such convention-a1 coils are wound with aninsulated conductor wire and are not usually completely impregnated in applications where the open circuit voltage is less' than 300 volts. For'example, a representative electrical coil used in a low voltage fluorescent lamp ballast having and open circuit voltage of less than 300 volts is Wound with conductor wire insulated with a thin nylon coating and with layer insulation formed of vegetable parchment orpaper approximately .002 of an inch in thickness. The conductor wire is wound over the layer insulation so that the layer insulation is interleaved between the layers of conductor wire. Thus, the layers ofconductor wire are not contiguous. The voltage between layers in such a typical ballast coil is inthe neighborhood of ten volts and between adjacent turns-the voltage is'abo ut one tenth of'a volt.

The paper layer insulation in such coils minimizes the effect of voltage stress between conductor .wire layers by pioviding'an insulating barrier therebetween. Since electrical coils used in ballasts are Wound on spools withoutrirns, the layer insulation also serves to prevent the end turns of a coil from being mechanically displaced.

The electrical coil or coils used in a ballast transformer o'rreactor are generally disposed on a central winding le'g within a coil receiving Window defined by the center winding leg and side yoke members which provide a return path for the magnetic flux. Generally, the length of the ballast coil dictates the length of the coil receiving windows of'the magnetic core of the ballast transformer since the cross-sectional dimensions are more or less fixed by the requirement that the ballast case not exceed certain specified dimensions in order that it can be mounted in alamp 'fixture.

It will be appreciated that the layer insulation used in conventional ballast coils adds to the over-all coil volume. A coil with layer insulation, as the term is used herein, denotes an electrical coil with flexible sheet insulation, such as paper, interleaved between the layers of conductors. Several types of coils are not wound with layer insulation between the layers of conductor wire. In a precision wound coil, for instance, each layer of turns is formed of consecutively woun'd turns which are accurately positioned to prevent any fall-through of a turn to an adjacent layer. A random wound coil is also wound without layer insulation but, as the term implies, the coil is wound-without any special provision being made for insuring that each turn of the conductor wire will fall in its proper layer. Consequently, in a random wound coil a turn may be displaced one or more layers from its normal layer position or the position it would have occupied if the coil were precision wound.

As comparedwith c-oils not employing layer insulation but having the same number of turns, a conventional ballast coil with'layer insulation will occupy more window space in -a shell-type of ballast transformer. With the dimensional limitations imposed on a ballast, it is necessary that the length of the core be increased to provide the necessary window space to accommodate a longer coil. The use of electrical coils with layer insulation has RF i 1C6 therefore made it necessary to use larger ferromagnetic cores and larger ballast cases. Consequently, ballast coils with layer insulation do not result in the most economical utilization of materials. Also, fluorescent lamp ballasts using electrical coils employing layer insulation are more costlytto manufacture since the layers of magnet wire must be wound over the layers of insulating parchment. Accordingly, there has been a long standing need for ballasts that can employ coils without paper layer insulation.

Since ballasts operating fluorescent lamps arerequired to have an average continuous service life over a period of approximately twelve years, the expected life of a ballast or ballast apparatus is usually determined by temperature accelerated life tests inorder to achieve a reasonably accurate estimate of expected service life in a relatively short period of time. In conducting such tests, coil samples are assembled in ballasts, andthe ballasts are then operated under normal current and voltage'conditions in an elevated ambient'temperature provided by a circulating air oven.

chemical reactions and since the rate of a given chemical reaction can be determined as a function of temperature; it is possible to select elevated temperatures and shortened periods of time to determine the rate at which the reactions will occur at normal operating temperatures and thereby estimate the service life of the ballast. A

conventional coil using a magnet wire coated with nylon and having a layer insulation consisting of vegetable parchment 0.002 inch in thickness has an expected service life of 12 years based on 5000 hours of operation per year at a maximum average coil temperature of degrees centigrade.

By way of comparison, two groups of ballasts employing coils random wound with identical magnet wire' coated with nylon were life tested. Ten of the coils in one group were vacuum impregnated with a mixture of a synthetic fatty acid amide type wax and asphalt.

The coils of the other group were not impregnated."

least 3750 hours of operation at the elevated temperature. Unless the ballast meets this requirement, it cannot be expected to provide an expected service life of 12 years based on 5000 hours of operation per year at a maximum coil temperature of 105 degrees centigrade.

In the past, it was generally believed that a primary factor in the premature failure of coils not employing layer insulation in ballasts undergoing temperature accelerated life tests was internal shorting resulting from This" a copper-to-copper type of contact between turns. copper-to-copper type of contact between turns was also generally believed responsible for coil failures in-ballasts installed in lamp fixtures. Various theories have been advanced to explain the mechanism of this copper-tocopper type of contact. For example, it'has been generally assumed that the copper-to-copper contact in a coil not employing layer insulation is the result of a number of factors, such as cut-through due to plastic flow of the Wire coating at points of maximum mechanical stress or the alignment of the various breaks in the insulating coating on the magnet wire. These cutthroughs and breaks can be variously caused by bending the magnet wire over small radius bobbin corners, by winding friction, by careless handling and other similar causes. A common explanation has been that the breaks in the insulating coating result in a short circuit of the Since the degradation of organic insulating materials can be treated mathematically as a series of coil thereby initiating the degradation processes which prematurely cause the coil to fail.

Heretofore, the tests carried out to determine the suitability of coils not employing layer insulation such as are used in ballast transformers were directed towards a determination of the susceptibility of the magnet wire enamel to breaks and cut-throughs. Usually, such tests as abrasion resistance, flexibility, and cut-through temperature tests have been employed. Although coils utilizing conductor wire satisfactorily meeting the requirements of such tests have been tested in ballasts, the ballasts still could not satisfactorily pass temperature accelerated life tests.

From the foregoing considerations, it will be apparent that there is a need for an improved ballast employing electrical coils that do not require the use of layer insulation and that can without such layer insulation satisfactorily meet the requirements of temperature accelerated life tests. Further, it is evident that such an arrangement that does not require layer insulation will result in benefits as reduced ballast size and weight, and in a more economical utilization of materials.

Accordingly, it is a general object of the invention to provide an improved fluorescent lamp ballast employing coils without layer insulation.

Another object of the present invention is to provide an improved coil assembly for use in ballasts for operating fluorescent lamps.

It is another object of the invention to provide an improved coil without layer insulation for use in a shell type of transformer used to operate fluorescent lamps.

It is a further object of the present invention to provide an improved ballast apparatus wherein the size of the ballast is reduced as compared with similar ballasts for operating comparable fluorescent lamps.

A more specific object of the present invention is to provide a new and improved ballast wherein the amount of material required to carry out the voltage transforming and current limiting functions of the ballast is appreciably reduced as compared with similar ballasts used for operating comparable lamps.

In accordance with one form of my invention, I have provided an improved ballast apparatus for operating one or more fluorescent lamps from an alternating power source in which the ballast transformer includes a coil assembly having at least a secondary winding comprised of layers of conductor wire without layer insulation, such as paper, interposed between the layers of conductor wire. The ballast transformer also includes a primary winding inductively coupled with the secondary winding on a magnetic core.

I have found that it is possible to employ an electrical coil in a shell type of ballast transformer for operating fluorescent lamps without need for layer insulation if the electrical parameters of the equivalent circuit of the coil are such that the power dissipated per unit volume of the insulating enamel on the conductor wire of the coil and per unit weight of the metallic conductor of the wire is maintained below a specified limit. Thus, a coil without layer insulation may be used in a ballast for operating fluorescent lamps if the electrical coil is provided with an equivalent circuit resistance R and an operating voltage V, such that the power in watts dissipated in the equivalent circuit divided by the number of cubic centimeters of insulating enamel in the coil and divided by the number of kilograms of metallic conductor in the coil is less than 1.6 at a frequency of 60 cycles per second and at a temperature between 140 and 180 degrees centigrade.

The subject matter which I regard as my invention is set forth in the appended claims. The invention itself, however, together with further objects and advantages thereof may be better understood by referring to the following description taken in connection with the accompanying drawings in which:

FIGURE 1 is an illustration of a coil equivalent circuit;

FIGURE 2 is a sectionalized View showing an idealized conductor wire layer arrangement of a precision wound coil;

FIGURE 3 is a sectionalized view illustrating random wound coils in which the conductor wire has a two layer fall-through or is displaced two layers from its natural position;

FIGURE 4 represents a sectionalized view of an idealized bifilar wound coil illustrating the normal disposition of conductor wires in such a coil;

FIGURE 5 is a schematic circuit diagram of a test circuit used to make dynamic measurements of the electrical performance of the bifilar wound coil samples used to simulate precision and random wound ballast coils;

FIGURE 6 represents a plot of time in minutes versus coil temperature for representative bifilar coils wound with conductor wire coated with various wire enamels;

FIGURE 7 represents a plot of temperature in degrees centrigrade versus power density in the coil dielectric material in watts per gram for bifilar coils with conductor wire coated with various wire enamels;

FIGURE 8 is a schematic circuit diagram of the apparatus used to determine the critical frequency used to compute the equivalent circuit parameter values of the ballast coils;

FIGURE 9 is a diagram illustrating how the reference capacitor used to determine the equivalent capacitance was arranged;

FIGURE 10 is a perspective view of a ballast embody-' ing the improved coil arrangement of one form of the invention with a part of the case and coil cut-away;

FIGURE 11 is a plan view of the ballast transformer shown in FIGURE 10;

FIGURE 12 is a schematic circuit diagram of a ballast apparatus incorporating the ballast transformer shown in FIGURES 10 and 11; and

FIGURE 13 illustrates a sectionalized view of a conventional high voltage ballast coil assembly.

As is shown in FIGURE 1, an electrical coil may be represented by an equivalent circuit consisting of a resistor with an equivalent resistance R, an inductor with equivalent inductance L and a capacitor with an equivalent capacitance C. It will be seen that the resistor, inductor and capacitor are connected in parallel circuit relation with each other. As will hereinafter be more fully explained, the equivalent resistance R at an elevated temperature is a significant parameter in determining the power dissipated in the wire insulating enamel of a coil at the elevated temperature and is an important factor in deter-' insulation when used in fluorescent lamp ballasts. In-

stead I have discovered that a thermal runaway mechanism is a key factor in the failure mechanism of such electrical coils when used in ballasts for operating fluorescent lamps.

In temperature accelerated life tests carried out to investigate the failure mechanism of ballast coils, the coil surface temperatures were measured by means of thermocouples taped to the outside surfaces of the coils. In most cases, I observed that the ultimate failure of the ballast coil without layer insulation was preceded by a period of several hours during which the coil temperatures. steadily rose to levels as high as degrees above the nominal coil temperatures of to degrees centi grade.

Since the ballast coils without layer insulation which were examined for breaks after finishing and the coils with-- out layer insulation which had been heat-aged for peri-- ods longer than the lifetime of similar coils operating at the same or lower temperatures in temperature accel-- erated life tests did not show any evidence of QOPEI- to-coppercontacts', it was considered'improbable that ductor wires positioned side by side in each layer, can

be used to simulate the electrical conditions of a random or precision wound coil. With a bifilar coil simulating a random or precision wound coil, it waspossible to more readily "make various electrical measurements as will now be more fully explained. I,

Referring to FIGURES 2, 3 and 4', I have shown there-.

in a crosssectional view of a representative center turn and its six adjacent'turns of a precision, random and bifilar wound coil. It will be appreciated that in the views shown in FIGURES 2, 3' and 4, the diameter'of the con ductor wire has been greatly enlarged relative to the coil dimensions. Thishas been done to bore clearly show the "relative disposition of the turns In the precision wound coil 10, as shown in FIGURE 2, the center turn 11 is surrounded by and in contact with six other turns 12, 13, 14; 15, .16 and 17, as will be.

seen' intheview of the lower halflof thecoil 10. Two of the turns 12, 13 are inthe inner layer and the two other turns '15', 16*are in the outer layer. The two adjacents turn 14 'and' 17 are disposed inthe same layer as the center turn 11. It'will be appreciated that the two adjacerrt turns14 and 17are disposed in the same layer as the center turn 11.

If we assume that the voltage difference between layers is V any current flowing between the center turn 11? and a'turn, such as .12, 13, 15 or 16, in the adjacent layers will be equal to the voltage difference V divided by the effective resistance R between the center turn 11 and an adjacent turn, It will be understood that the effective resistance R .which is the resistance between adjacent turns, ditfersfrom the equivalent resistance R of the equivalent coil circuit. The equivalent resistance R is.

the pure resistance which when placed electrically in parallel with theequivalent inductance L and equivalent ca-- pacitance C, as shown in FIGURE 1, will form a combined'circuitthat exhibits an electrical behavior essentially similar to that of the coil. The effective resistance R on theother hand, is the resistance in the area of contacfbetween one turn and an adjacent turn.

It was found that in the type of ballast coils which were being, tested, having approximately 100 turns per layer and 13 layers, the voltage difference between adjacent turns in the same layer couldbeig'nored since this voltage diiference was approximately of the voltage difference V between layers Accordingly, for a precision wound coil, the total effective current I between the center turnll and the adjacent turns 12, 13, 1'5 and 16-in the inner and outer layers may be expressed as follows:

4 L I RE where'R is the effective resistance in the area of contact between the center turn and an adjacent turn in an adjoining layer of the coil.

'From an examination of numerous random wound coils wound on a fiyer head winding machine, it was found that the maximum displacement of a wire from its normal position was usually two layers. Thus, the most highly stressed turns in a random wound coil are assumed to be the turns that are displaced two layers from this normal position.

In FIGURE 3, I have illustrated a group of seven turns in an enlarged cross-section of a random; wound coil 20,

in which a center turn 21 is assumed to be displaced inwardly two layers. I have indicated inthe upper sectionalized view of the coil 20 the idealized voltage relationships between the centerturn 21 and six adjacentturns 22, 23, 24, 25, 26 and 27.

Assuming that a potential or voltage difference-V ex'-- ists between the center turn- 21 and the two innerturns 22, 23, the voltage difference between the center turn 21 and one of the adjacent turns 24, 27 in the same layer will be twice the voltage difference between layers or ZV since the center turn is displaced two layers from its normal position. Further, it will be seen that the voltage difference between the center turn 21 and each of the outer adjacent turns 25 and 26 will be approximately equal to three times the" voltage difference between'layers 'or'3V Accordingly, the effective current I between a center turn and the adjacent turns may be expressed as follows:

where R is the effective resistancein the areaof contact between the center turn and an adjacent turn'in an ad'- joining layer of the coil.

I have found that a bifilar coil can be used tosimulate the electrical conditions of a precision or random wound coil by adjusting the voltage difference between the conductor wires. If the bifilar coil 30 of FIGURE 4 is truly precision wound, a center turn 3l may'be' assumed'to be in contact with six adjacent turns 32, 33, 34, 35, 36 and" 37. Two of the turns, adjacent to thecenterturn- 31,

such as turns 33 and 36, are extensions of the center'turn v 31, and the voltage difference between these turns is zero when the applied voltage is impressed between the two windings. The voltage difference between the center turn" 31 and turns 32, 34, 35 and 37 may be assumed to be V Thus, the effective current I maybe expressed as follows:

I RE where R is the effective resistance in'the contact area between a center turn and an adjacent turn of the coil.

If it is desired to have the bifilar coil simulate theelectrical conditions of a precision wound coil, the voltage difference V between the bifilar windings should be made" approximately equal to the voltage difference V between the layers of the precision wound coil. In a typical ballast coil, this voltage difference V was about 10 volts;

Where it is desired to simulate a rand-om woundcoil, the voltage difference V may be taken as being approximately equal to three times the voltage difference V In the random wound bifilar coils tested, as will hereinafter be morefully described, .a voltage of 30 volts r;m.s. or 3 times the normal layer to layer voltage was impressed across the bifilar windings. The bifilar coils used were wound on a winding lathe to the same geometry as the comparable ballast coil. The coil geometry was closely controlled by winding the coils on phenolic bobbins;

To further investigate the mechanism of coil "failure in a ballast transformer, an apparatus'40, as shown inFIG- URE 5', was devised to dynamically determine-the electrical behavior of the bifilar wound 'co-il sample simulating ballast coils under elevated temperature conditions; The-j particular coils tested were random wound. In order to produce internal heating within the bifilar windings L L a transformer T was used. The transformer T include-d a pair of secondary windings S and S inductively coupled with a primary winding P on a magnetic core 41. Since the secondary windings S and S are connected across the bifilar windings L L it will be understood that the insulation impedance of the secondary windings S and S should be on the order of several magnitudes greater than the insulation impedance of the bifilar wind- .provided in the secondary circuits so that the normal ballast operating current density in the copper of the bifilar windings L L could be maintianed. The terminal leads 42 and 43 of the primary winding P were connected to a- 120 volt 60 cycle power supply.

Since the bifilar windings L and L were not disposed on a magnetic core, the voltage required to produce the required current density was about 3 volts. Since this voltage stress is not of the magnitude that would normally be encountered between turns in a ballast coil, a second. transformer circuit including a transformer T and a variable autotransformer T was provided. The transformer T was an isolation transformer having a secondary winding S and a primary winding P inductively coupled on a magnetic core 42. The autotransformer T included an autotransformer winding A and an adjustable tap 45 to provide a variable voltage output across the autotransformer winding A One end of the autotransformer winding A was connected with winding L and the other end was connected with winding L through a resistor R Thus, the voltage supplied across the bifilar windings L and L could be varied to supply a voltage stress that was comparable to the voltage stress normally encountered in a ballast coil.

Leads 46 and 47 were connected to the vertical amplifier terminals 48, 49 of an oscilloscope 50 schematically shown enclosed in the dashed rectangle. These electrical connections were made so that a vertical deflection proportional to the voltage drop across the resistor R or, in other L was produced on the oscilloscope 50. Leads 51 and 52 were connected in circuit with the horizontal amplifier terminal 53 and 54 so that a horizontal deflection proportional to the voltgae drop across the resistor R or, in other words, proportional to the current flow between the bifilar windings L and L was produced. It will be noted that current flowing between the bifilar windings L and L will produce a current flow through the resistor R It was found that with a ohm ressitor R a current of approximately 1X10 amperes in magnitude produced a deflection of 1 centimeter when the one millivolt per centimeter range of the horizontal amplifier of the oscilloscope 50 was employed. The phase angle of the voltage impressed across the bifilar windings L and L and the current between them was measured by using a Hewlett-Packard Webb phase-shift mask placed on the face of the cathode ray tube of the oscilloscope 50. With this device it was possible to read the phase angle within plus or minus 1 degree of arc. An ammeter 55 was connected in circuit with the winding L and also an ammeter 56 was connected in circuit with winding L so that the current flow through the windings could be observed.

Before a bifilar coil containing the windings L and L was placed in the oven 60, a thermocouple was taped to the outer surface of the coil. After the thermocouple was attached, the bifilar coil was placed in a 250 milliliter glass beaker, and the beaker was filled with an asphaltic potting material containing 48 percent by weight of blown petroleum asphalt having a 118 degree centigrade softening point and containing 52 percent by weight of silica. The glass beaker was filled with the asphaltic potting material to a height suflicient to completely cover the coil.

Since in a ballast transformer approximately half of the heat comes from losses in the steel of the magnetic core, this heating effect was simulated by placing the coil in the forced air oven 60. In order to simulate the poor heat transfer away from the interior of a ballast and also to reduce the rapid cooling resulting from air flow in the oven, the beaker containing the bifilar coil was placed in a cardboard container, the glass beaker was then covered with chopped glass rovings to a depth of about A of an inch and several straps of mylar tape were placed across the top of the cardboard container in order to retain the chopped glass rovings.

. Readings were taken of current, voltage, phase angle and coil temperature over temperatures ranging from degrees to 170 degrees centigrade. In FIGURE 6, I have shown a plot of coil temperature in degrees centigrade versus time in minutes for unimpregnated coils without layer insulation wound with a single film build nylon wire, curve A, and a single film build Formex wire, curve B. It will be seen from the sharp rise in curve A representing the nylon insulated magnet wire that a source of heat is present in the nylon wound bifilar coil that was not present in the coil wound with the Formex wire.

The conductor wire identified herein by the registered trademark Formex employs an insulation coating that is the reaction product of a partially or completely hydrolyzed polymerized vinyl ester and an aldehyde. The insulation is more fully described in US. Letter Patent No. 2,085,995 granted to W. I. Patnode et al. The trademark Formex is used herein for the purpose of conveniently identifying the wire used.

In FIGURE 7, I have illustrated a plot of the power density in the wire coating in watts per gram versus the coil temperature in degrees centigrade as determined from thermocouple readings. In making these plots it was assumed that the measured power was being dissipated in the wire coating. Curve A represents the values obtained for the simulated random wound coil with a nylon wire and curve B represents the values obtained for a simulated random wound coil with the Formex wire corresponding to the curves A and B of FIGURE 6. Curve C represents a plot of the values of power density against temperature for a bifilar coil wound with Formex wire with an asphalt base impregnating material.

The extremely high power densities in the wire enamel coating are believed to be a principal factor in causing the coil failures during accelerated life tests of the ballast transformers using random wound nylon insulated wire. The power apparently dissipated in the wire enamel coatlng raises the coil temperature above normal, and a thermal runaway type of failure mechanism results. As the temperature increases, more power is believed to be dissipated in the wire enamel coating with further increases in the coil temperature until a complete electrical breakdown occurs.

It will be seen from a comparison of curves B and C that the power densities over the to 200 degree temperature range were considerably less in the unimpregnated coil (curve B) than for the coil (curve C) that was impregnated. Thus, it was found that when coils without layer insulation are impregnated with a material, the power. density in the dielectric system may increase sharply with temperature. It is believed that the impregnating material provides additional paths for the flow of current between the turns of the random wound coil thereby increasing the apparent power density in the insulating system.

In the bifilar coils and ballasts tested, the magnet wire had an outside diameter of approximately 0.0143 inch. The average length of the contact with an adjacent wire around the circumference of the conductor wire was about 0.001 inch. Thus, in an unimpregnated random wound coil a turn in contact with six adjacent turns, as shown in FIGURE 3, will make contact along 0.006 inch of its 0.0450 inch circumference. However, in an impregnated random wound coil contact will exist through the impregnant between the total circumference of a turn and the adjacent six turns. Also, the impregnating material connects the pinholes and breaks in the wire covering. It Wlll be apparent, therefore, that the impregnating material provided an increased number of potential paths for.

current flow and introduced additional electrical losses in the coil'circuit.

The foregoing tests in which bifilar coils were used to simulate random wound coils, clearly demonstrated that the electrical characteristics at elevated'temperatures of a ballast coil without-layer insulation were controlling factors in the failure mechanism as contrasted to ballast coils with layer insulation. Accordingly, measurements of the electrical characteristics of' ballast coils without layer insulation at an elevated temperature-were carried out.

In making the measurements :ofthe electrical characteristics-at elevated temperatures, it'was assumedtliat a bal'-- last coil can be represented by aparallel network as shown in FIGURE 1'. It can bereaclily demonstrated 'that-a ballast coilexhibits an electrical behavoir that is typ'ical-ofa parallel RLC network. For example, if the start and finish leads of a ballast coil, before it 'is mountedon'a magnetic core, are connected across an AC. voltage source, as the frequency of the voltage source is increased from a relatively low value, the current through the coil will decrease. Also, the phase angle between the voltage and current'will' decrease.- When acertain critical frequency is reached, the current andvoltage will be in phase and the impedance of the ball'ast'coil will be at its maximum value. Further, it willbenotedlthatcurrent flowing through the coil will be at its minimum value. As the-frequency of theapplied voltage isfurther increasedabove its critical value, it will be found'thatthe phase angle between the current andvoltage Will-increase in magnitude. However, at values of the frequency"above'the'critical value the current leads the applied voltage, whereas: at fre-' quencies below the critical frequency, current-willlag the. applied voltage. Therefore, the electrical behavior of'a ballast coil is similar to that of'ra parallel RLC network.

The manner in which'the equivalent resistance R was determined 'will now'be more fully described. The equivalent resistance Rat a frequency of 60 cycles and at a specific temperature was obtained from the following equation:

where X isthe reactance of the coils equivalent capacitance and D is the dissipation factorof this capacitance. It will be appreciated that=the determination of the value of equivalent resistance R-and the measurements of other equivalent parameters, as described'herein; can be applied to all fluorescent lamp ba-llastjcoils wound'withoutlayer insulation. irrespective of the manner in which the coils are wound ,random, precision or otherwise.

In order to obtain the capacitive reactance X the value. of the'equivalent capacitance C atroom temperature of a ballast coil was .first determined by resonating the coil with a beat frequency oscillator to obtain the critical frequency of the coil. From the value of the critical frequency, the capacitance of the coil wascomputed, and then the value of capacitance. at: the critical :frequency was' converted to a value at 60 cycles per second, as will be hereinafter'more fully explained.=-

It will be understood that the equivalent capacitance of the coil could not be convenientlymeasureddirectly at a frequency of 60' cycles per second; The value of the capacitance at the critical frequency and at room temperature was converted to a value atr 60 cycles per second by multiplying the computed value at the critical frequency by a proportionality factor obtained by making measurements on a comparable reference'capacitor at both the critical'frequency and at 60Icycles .per second, Similarly,

the value of capacitance of the coil at 60 cycles per second' was converted to a value at a desired elevated temperature T by a proportionalityfactor obtained by making measurements of the capacitance of the reference capacitor at room temperature and at the desired temperature T;

Referring nowto FIGURE 8, I have illustrated therein the apparatus 65 used to determine the'critical frequency of a ballast coil 66. It will be seen that the ballast coil 66 to. be tested was connected across the leads 67,- 68"joined to the terminals of a beat, frequency oscillator 69. A frequency counter 70 was also connected across the terminals of the beat frequency oscillator 69 to measure the frequency of the signal. applied across the ballast coil 66. The. ballast coil 66 was mounted on. a ferrite core 72 to increase the inductance of the coil 66 so that the critical frequency of the coil 66 as mounted on the ferrite core 72' would fall within the frequency range of the beat frequency oscillator 69. The ferrite core 72 may, of course, be.

eliminated Where the critical frequency of the coil is within the frequency range'of the beat frequency oscillalator. 69.

Todetermine the critical frequency, the frequency of the signal supplied by the beat frequency oscillator 69 was varied until a minimum voltage was indicated on the scale of the voltmeter 71 connected in shunt with resistor R With a minimum voltage across the resistor R the current flow throughthe resistor R is at a minimum therebyindicating a condition of resonance. The frequency was then measured with the frequency counter 70 by counting oscillations for one second intervals.

The inductanceL of the coil 66 mounted on the ferrite core 72 was measured at 60 cycles per second and at room temperature. It was assumed that this value would not significantly, change with temperature orfre'quency. The measurements were made on a Hays and Maxwell inductance bridge circuit of. a General Radio type 1650-A impedance bridge.

Using the measured value of the inductance L obtained from'the impedanc'ebridge' measurements, the capacitance C at thecriticalfrequency was computed from the following equation for the critical frequency in cycles per second; f r

The procedure followed to convert the values of the capacitance C at the critical frequency to the values at 60 cycles per second will now be more fully described. As is shown in FIGURE 9, a reference capacitor was formed by bringing out a start lead 61 and a finish lead 62 from the outer turn layer of coil 66. The finish lead 62 was soldered to the start lead 61. A start lead 63 and a finish lead. 64 were brought out from a layer adjacent to the outer layer and' were connected as shown. Since the capacitance between the two turn layers'of the reference capacitor changes in frequency in substantially the same manner as the equivalent capacitance of the coil 66, it was assumed that the values at 60 cycles per second and at the critical frequency are proportional. of this reference capacitor was measured at room temperature and at the critical frequency f of the ballast coil and at 'a'frequency of 60 cycles pe'r secondwiththe General Radio impedance bridge used to measure the inductance L, using the'beat frequency oscillator 69 as a source of voltage for the impedance-bridge. bridge was used -to measure the capacitance of the reference capacitor at 60 cycles per second with the reference? capacitor being maintained at the desired elevated tem-' perature T. Further, the dissipation factor D at the desired elevated temperature T was measured with the impedance bridge;

To obtain the value of the equivalent capacitance of coil 66 at 60 cycles per second at room temperature, the value-of the equivalent coil capacitance at the critical frequency was multiplied by the ratio of the capacitance of the reference capacitor at 60 cycles to the capacitance of the reference capacitor measured at the critical frequency. The room temperature value of the equivalent capacitance of the coil 66 was converted to the value corresponding to the desired elevated temperature T by multiplying the room temperature value by the ratio of the capacitance of the reference capacitor at the desired temperature to the capacitance of the reference capacitor at room temperature. The equivalent capacitive reactance X was then computed from the following relationship:

1 X c where: fis the frequency in cycles per second, and

The capacitance Also, the impedance 1.1 C is the equivalent capacitance of the ballast coil (at 60 cycles per second) in farads at the desired temperature T.

The value of the equivalent resistance R was determined from the value of the dissipation factor D that was obtained simultaneously from the measurements at 60 cycles per second of the capacitive value of the reference capacitor at the desired temperature T. When measurements are made with the parallel capacitance comparison bridge circuit of a General Radio impedance bridge, the values of the dissipation factor D can be obtained directly. Since the dissipation factor is a quantity that is independent ofthe size of a capacitor and since the change in the capacitance of the reference capacitor was essentially proportional to changes in the capacitance of the ballast coil 66 with changes in frequency, the values of the dissipation factor D observed for the reference capacitor can be used to obtain equivalent resistance R of the ballast coil 66. The values of the equivalent resistance R were computed from the following equation:

where:

X is the equivalent capacitive reactance of the coil at 60 cycles per secondand at the desired temperature T, and

D is the dissipation factor at 60 cycles per second for the reference capacitor at the desired temperature T.

Having found the value of the equivalent resistance R corresponding to a frequency of 60 cycles per second and a temperature T, the power in watts dissipated in the wire enamel of the coil was computed from the equation,

V2 P R where: V is the operating voltage of the coil, and R is the equivalent circuit resistance of the coil corresponding to a frequency of 60 cycles per second and a temperature T.

. The volume of the wire enamel used in the coil 66 and the volume of the copper conductor were determined. The weight of the copper in kilograms was computed from the volume. The power was then expressed in watts per cubic centimeter of wire enamel per kilogram of the copper conductor, and this value is referred to in this specification as the thermal runaway factor. I have found that this factor provides a basis for predicting whether a ballast coil will fail as a result of a thermal runaway condition developing within the coil.

- For the convenience of those desiring to practice the present invention, the following specific components were used in the apparatus 65 shown in FIGURE 8 to determine the critical frequency of the ballast coils tested:

Voltmeter 71 Hewlett Packard Model 400-A, Vacuum Tube Voltmeter, Serial Number 14,572. Beat Frequency Oscillator 69 Bruel and Kjoer Beat Frequency Oscillator type 1014, Serial No. 22,687. Beckman Model 5230BP Serial No. 386.

Frequency Counter 70 ured value of the inductance L was found to be 0.268

henry with a quality factor of 250. The value of selfcapacitance C was then calculated from the value of the resonant frequency f and the inductance L as follows:

(21rf L (21rX21,4-5O) X.268

C,,=2.06 10 farads The coil 66 was cut, and the leads 61, 62, 63 and 64 were brought out and connected as shown in FIGURE 9 to provide the portion of the coil 66, which was used as the reference capacitor. Using the impedance bridge, measurements of the self-capacitance C if the reference capacitor were made at the resonant frequency of 21,500 cycles per second, at 60 cycles per second, at room temperature (25 degrees centigrade) and at a temperature of degrees centigrade. The capacitance of the reference capacitor at the resonant frequency and at room temperature was found to be 1.40 10- farads. At a frequency of 60 cycles per second and at room temperature, the measured value was l.48 l0- farads. The value of the total equivalent circuit capacitance C was converted to a value at 60 cycles per second as follows:

A measurement was made of the capacitance of the reference capacitor at 60 cycles an a temperature of 170 degrees centigrade. This value was found to be 2.64 10- farads. The dissipation factor D at 60 cycles and at a temperature of 170 degrees centigrade was found to be 5.67. The total equivalent circuit capacitance at 60 cycles per second and room temperature was converted to the value C at a temperature of 170 degrees as follows:

equivalent circuit of coil 66 was determined from the following equation:

To determine the volume of wire enamel in the coil 66, the diameter of the enameled wire was first measured and was found to be 0.1071 centimeter. The wire enamel on a portion of the wire was removed with a chemical stripper, and the diameter of the bare wire was found to be 0.1023 centimeter. The total length of the wire was determined by multiplying the mean length of a turn (17.0 centimeters) by the number of turns (402) of the coil 66. The total length of wire was 6834 centimeters. The volume of wire enamel (5.0 cubic centimeters) was found by subtracting the volume of copper (56.2 cubic centimeters) from the volume of the enameled wire (61.2 cubic centimeters). The weight of the copper conductor was determined by multiplying the volume of copper by the density in grams per cubic centimeter (8.92). The Weight of the copper in the coil 66 was 0.500 kilogram.

r 13 The thermal runaway factor for'coil 66 was 0.0001491 watt dissipated in the equivalent circuit per cubic centimeterlof wire enamel per kilogram of copper.

InfTable I, I have listed the insulating systems used, and the thermal runaway factor and the coil failures resulting from thermal runaway based on temperature accelerated life tests of a group of ballasts using the coil constructions as indicated in Table I.

elongated central winding leg 88 on which the coil assemblies 84, 85 are disposed and a pair of side yoke members 89, 90 which abut the opposite sides of the ends of the central winding leg 88 to form a closed magnetic circuit. The highreactance of the transformer'87 is provided'to some extent by the distributed leakage of the magnetic flux between the elongated central winding leg 88 and the side yoke members 89, "90.

Table I Results of Thermal Thermal Coil Insulating System Runaway Tests Runaway Factor Seeondary Nylon Wire Enamel Impregnated with Thermal runaway 6.0

' Wax-Asphalt Mixture, 1,304 turns eviden 0.0320 dia. wire. Do; NylonWire Enamel Impregnated with .do 16. 4

Wax-Asphalt Mixture, 1,271 turns 00142 dia. wire. p

Do Nylon Wire Enamel No Impregnant, do 1.64

1,271 turns 0.0142 dia. wire. Do NylonWire Enamel, No Impregnant, No thermal run- 0.6 1,304 turns 0.0320 dia. wire. away.

Do Formex Wire Enamel, Wax-Asphalt ,do '0. 01

'Irnpregnant, 1,271 turns 0.0142 die. wire.

I have found that a ballast coil for operating fluorescent lamps and Wound without layer insulation will not fail prematurely by'a thermal runaway mechanism if the thermal runaway factor, based on an equivalent resistance R'determined at a frequency of 60 cycles per second and a temperature between 140 and 180 degrees centigrade, was less than 1.6. g

In determining whether the ballast coil would fail as a result of a thermal runaway condition, the coil was operated in a ballast electrically loaded with appropriate lamps for the particular ballast in an elevated ambient temperature. The ambient temperature of the oven wasadjusted to produce the desired average ballast coil temperatures as measured by the resistance rise of the coils. The time to failure at theelevated temperature was the time at which a fuse in the ballast primary rated at approximately 1.5 times the normal current failed or the ballast failed to operate the lamps.

Referring to FIGURES 10, 11 and 12, I have illustrated therein the improved ballast coil arrangement of the invention embodied in a ballast apparatus 80 for operating apair of fluorescent lamps. In the perspective view of FIGURE 10, I have illustrated the apparatus 80 with a portion of the case 81 and potting material 82 cutaway to show the internal disposition of the components, the electrical connections having been omitted. The connec- ,tions are shown schematically in the circuit diagram of FIGURE 12.

As will be seen in the cutaway portion of the primary coil assembly 84 shown in FIGURE 10, no paper layer insulation was used. Also, the high'voltage or secondary coil assembly 85 was wound without layer insulation. In the schematic circuit diagram of FIGURE 12 the equivalent resistance R and the equivalent capacitance C of the primary winding or coil P are represented by a resistor and'a capacitor shown in dashed out-line. The equivalent inductance L is identified with the primary winding P Similarly, the corresponding parameters of the secondary winding or coil S are identified by the reference letters R C and L In accordance with the invention, the coils P and S were wound without paper layer insulation. The parameters of the equivalent coil circuit, R C were such that values of the equivalent resistance R provided a thermal runaway factor that was less than 1.6. A transformer 87 of the shell type was used to perform the voltage transforming and current limiting functions of the ballast apparatus 8 0. As will be seen in FIGURE 11, the transformer 87 includes the magnetic core 101 formed of an Magnetic shunting pieces 91 may be provided bet-ween thecoil'assemblies '84, to'pro'vide a flux leakage path between the coils'84, 85 if required. It will be understood that depending upon the design "of the transformer, the flux leak-age path or shunts may be formed either through nonmagnetic materials, such as air, or through magnetic material such as by insertable shunts or by shunt legs formed on the side yoke members'89 and 90.

The magnetic core 1010f the ballast transformer 87 is formed from a stack of l-aminations' made ofmagnetic material. The laminations of the winding leg .88 are stacked, and the coil assemblies .84, 85 are placed in as-' sem'ble-d relation on the center winding leg 88. The stacks of laminations which form the yoke members 89, 90 are assembled with the center winding leg to form 'the magnetic core101 and are held in assembled relation by me-ans of clamps92 and 93 or may otherwise securely be heldtog ether by other suitable means.

As will be seen in the cutaway portion of the coil assembly 84 as shown in FIGURE 10, the conductor wire 86 is wound on a paper spool 95. If desired, molded phenolic bobbins may be used. in place of the paper spool 95. Since coils P and S were wound on a spool 95 that did not have a rim to support the end turns, paper ground insulation 96 was provided to prevent the end turns from contacting the core laminations and grounding. A paper wrapper 97 was provided on the outside of the coil assembly 84 to prevent the outer turn layer from being damaged during handling and assembling and also to insulate the outer turn layer. A similar wrapper 98 was provided for the coil assembly 85. It will be seen in FIGURE 10 that a two-section capacitor 98 is encased in the potting material 82 and includes the series capacitor C and the starting capacitor C of the ballast apparatus 80.

In FIGURE 12, I have illustrated a schematic circuit diagram showing the electrical connections for the ballast apparatus 80 of FIGURE 10. The ballast transformer 87 must not only provide the requisite starting and operating voltages for a pair of fluorescent lamps 1, 2 but must also provide the ballasting action required to limit the current during the lamp arcing period. The ballasting action is necessary because the lamp resistance characteristic-possesses a negative slope.

The ballast apparatus 80 as shown in FIGURE 12, includes the high reactance transformer 87, the series capacitor C and the starting capacitor C Lamps 1 and 2 are positioned in proximity to a grounded conductive plate 100 so that the cathodes of the lamps 1 and 2 are 15- capacitively coupled with the conductive plate 100 to facilitate starting. In most applications, the lighting fix- .ture in which lamps 1 and 2 are mounted serves as the conductive plate 100.

In the schematic circuit diagram of FIGURE 12, it will be noted that the high reactance transformer 87 includes the magnetic core 101, the primary winding P the secondary winding S cathode heating windings H H H and the magnetic shunts 91. Secondary winding S is included in the coil assembly 85 shown in FIGURE 11. The cathode heating windings H H and H are wound over the primary winding P and are included in the coil assembly 84 shown in FIGURE 11. If desired, the cathode heating windings H H and H may be wound over a wrapper of paper insulation.

A pair of input terminal leads 102, 103 are provided for connection to a suitable alternating power source. When the ballast apparatus 80 is energized, the cathode heating windings H H H supply the cathodes of lamps 1 and 2 with heating current. Cathode heating windings H and H are connected in circuit with the lamps 1, 2 by output lead 104 and leads 105, 106, 107. Cathode heating winding H which is an extension of the primary winding P is connected in circuit with lamp 1 by output lead 108 and lead 109.

When the input terminal leads 102, 103 are connected across an alternating power source, the open circuit voltage developed across the, primary winding P and .secondary winding S is initially applied across lamp 1 because of the shunting action of the starting capacitor C Further, during the open circuitcondition, to aid starting lamp 1 the combined voltage across the primary winding P and secondary winding 8,, is applied between a cathode of lamp 1 and theconductive plate 100. After lamp 1 is started, it conducts current, and current flows through the starting capacitor C .Since lamp 2 is also disposed in close proximity to the conductive plate 100, a starting aid potential is also applied to a cathode of lamp 2 because of its capacitive coupling with the conductive plate 100.

When both lamps 1 and 2 have started, the impedance of the starting capacitor C relative to the lamp impedance is such that there is n-oappreciable current flow through the capacitor C Further, because of the impedance in .the capacitive coupling between lamps -1, 2 and the conductive plate 100, no significant current flow occurs between the lamps 1, 2 and the conductive plate 100 after the lamps are started.

By way of a more specific exemplification of the invention, a ballast apparatus was constructed for starting and operating two 96 PG 17 power groove fluorescent lamps. The primary winding P was wound with 344 turns of 0.0508 inch copper wire, and the secondary winding was wound with 1300 turns of 0.0320 inch copper wire. The equivalent circuit parameters at 60 cycles and a temperature of 180 degrees centigrade of the secondary coil assembly 84, were as follows:

Equivalent capacitance C 1.73 lmicrofarads Equivalent resistance R 3.5 X ohms For the secondary coil assembly the thermal runaway factor was 0.0019. The ballast apparatus employing coils with these equivalent circuit values satisfactorily passed temperature accelerated life tests with no evidence of thermal runaway. It was possible to contain the ballast in a case having the following outer dimensions: 2% x 3%" x 13". A comparable ballast of the prior art would be housed in a case having the following outer dimensions: 2 /8" X 3% x 18". Thus, it will be seen that significant reduction of 5 inches or approximately 28 percent of the length of the ballast was achieved by the practice of the'present invention.

It will be appreciated that to start power groove fluorescent lamps at a temperature between -20 and 90 degrees Farenheit, a ballast. must provide a peak starting aid voltage of 575 volts and R.M.S. voltage of 500 volts. It will be noted that even though shorter lamps are used, the same peak starting aid voltage must be provided by the ballast since the peak voltage requirement of a fluorescent lamp depends upon the temperature rating, if rcliable starting is required at a particular temperature. Therefore, it will be appreciated that the ballast transformers must supply appreciable voltages in order to start and operate fluorescent lamps at lower temperature ratings. Despite the appreciable voltage requirement imposed upon the coils of a ballast transformer, it is possible with the present invention to use coils in such ballast transformers in which the layers of conductor wire are contiguous or in other words, the layers of conductor wire are not wound over layers of insulation.

Turning now to FIGURE 13, I have illustrated therein a magnified sectional view of a portion of a conventional high voltage coil as seen under a microscope. The insulated conductor wire turns are disposed in layers and paper layer insulation 111 is interposed between the turn layers. It will be apparent that the layer insulation 111 causes an appreciable displacement between the turn layers 110. This results in a relatively larger coil assembly for a given number of turns as compared with a coil wound with no layer insulation where the turn layers are contiguous.

Heretofore, it has been generally considered necessary to provide paper insulation between the layers of turns of a ballast coil to prevent early coil failures. Attempts made in the past to employ high voltage coils without paper layer insulation were not successful because the ballasts frequently failed, as a. result of a thermal runaway condition developing in the ballast coils.

It will be appreciated that in the typical ballast coil assemblies presently used to operate the smaller fluorescent lamps, such as the 40 watt rapid start lamps, the open circuit voltage required is less than 300 volts R.M.S. The. coil assemblies in the ballasts utilize a thin film nylon wire wound over a 0.002 inch vegetable parchment layer insulation. These coils are not usually impregnated but are dipped in a wax and asphalt mixture. The voltage between turns in the same layer is about 0.10 volt and the voltage between layers is about 10 volts.

In view of the low voltage and voltage stresses present in such coils, the failures of coil assemblies without layer insulation in ballast transformers have been attributed to various mechanisms involving copper-to-copper contacts in the finished coils. However, I have discovered that a primary mechanism of failure in ballasts using coils -without layer insulation is essentially an electrical phenomenon. Further, I have found that if electrical parameters of the equivalent coil circuit, such as the equivalent resistance, and the equivalent capacitive reactance as determined at elevated temperatures are maintained within certain limits to maintain the power dissipated in the wire enamel below a specified level as I have herein set forth, coils without layer insulation may be used in ballasts for operating fluorescent lamps and will not fail as a result of a thermal runaway condition. The coils may be random precision or universal wound as may be desired in a particular application, so long as the criteria for the electrical parameters in accordance with my invention are met. In other words, the thermal runaway factor or the power dissipated in the wire enamel in watts per cubic centimeter of the wire enamel per kilogram of the metallic conductor must be less than 1.6 based on an equivalent resistance of the coil determined at a frequency of 60 cycles per second and a temperature range between and degrees centigrade.

An important advantage of the arrangement of the invention is that appreciable reductions in the size, weight and amount of materials used in a ballast apparatus are made possible. Further, as compared with conventional ballasts, the improved arrangement has, resulted in significant reduction in the noise level of the ballast.

It will be apparent to those skilled in the art, that there are many different types of insulating systems that may be used in the practice of my invention. Accordingly, it it not intended to limit my invention to the specific exemplifications which I have herein described byway of illustration of the invention. While the present invention has been described by reference to specific exemplification thereof, it is to be understood that other modifications may be made by those skilled in the art without actually departing from the invention. It is, therefore, intended in the appended claims to cover all such equivalent variations that come within the true spirit and scope of the invention.

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

1. A ballast apparatus for operating a fluorescent lamp from an alternating power source, said ballast apparatus comprising: a ballast transformer having a magnetic core with an elongated winding leg and side yoke members, at least a pair of coil assemblies disposed on said Winding leg, one of said coils including at least a primary winding adapted for connection to the power source and the other of said coil assemblies including at least a secondary winding, said secondary winding and said primary winding being comprised of turns of conductor wire formed of a metallic conductor coated with an insulating enamel, at least said secondary winding being comprised of turns of contiguous layers of conductor wire, means including output leads for connection with the fluorescent lamp and for applying at least the voltage of said secondary winding across said output leads, said coil assembly with said secondary winding having an equivalent resistance R and an operating voltage V such that the power V /R in watts dissipated in the equivalent circuit divided by the number of cubic centimeters of the insulating enamel in the coil and divided by the number of kilograms of metallic conductor in the coil is less than 1.6, said value of the equivalent resistance R being determined at a frequency of 60 cycles per second and a temperature between 140 and 180 degrees centigrade.

2. A ballast apparatus for starting and operating fluorescent lamps from an alternating power source, said ballast apparatus comprising: a high reactance transformer having a magnetic core formed with an elongated winding leg and side yoke members, said side yoke members and said winding leg defining coil receiving windows, at least a pair of coils mounted on said winding leg and disposed within said coil receiving windows, one of said coils including at least a primary winding and the other of said coils including at least a secondary winding, said primary winding and secondary winding being inductively coupled on said magnetic core, means including output leads for connection to the lamps whereby the voltage across at least said secondary winding is supplied at said output leads for operating said lamps, a pair of input leads for connection with the power source, said primary winding being connected in circuit with said input leads, said secondary winding and said primary winding being comprised of turns of conductor wire formed of a metallic conductor insulated with resin, at least said secondary winding being impregnated and wound to provide turns arranged in contiguous layers, said coil including said secondary winding having an equivalent resistance R, and an operating voltage V such that the power V /R in watts dissipated in the equivalent circuit divided by the number of cubic centimeters of enamel in the coil and divided by the number of kilograms of metallic conductor in the coil is less than 1.6, said value of the equivalent resistance R being determined at a frequency of 60 cycles per second and a temperature between 140 and 180 degrees centigrade.

3. A ballast apparatus for operating at least one fluorescent lamp from an alternating power source, said ballast apparatus comprising: a ballast transformer having a magnetic core, a primary winding and a secondary winding inductively coupled with the primary winding on said magnetic core, a pair of input leads for connection to the alternating power source, said primary winding being connected across said input leads, and means including output leads for supplying the output of the ballast transformer across said at least one electric discharge lamp, said primary winding being included in a first impregnated coil wound with contiguous layers of conductor wire, said secondary winding being included in a second impregnated coil wound with contiguous layers of conductor wire, said conductor wire being formed of a copper conductor covered with an insulating enamel, each of said coils having an equivalent resistance R and an operating voltage V such that the power V /R in watts dissipated in the equivalent circuit divided by the number of cubic centimeters of enamel in the coil and divided by the number of kilograms of metallic conductor in the coil is less than 1.6, said value of the equivalent resistance R being determined at a frequency of 60 cycles per second and a temperature between and degrees centigrade.

4. A ballast apparatus for operating fluorescent lamps from an alternating power source, said apparatus comprising: a high reactance transformer having a shell type magnetic core, a primary winding and a secondary winding inductively coupled with said primary winding on said magnetic core, a pair of input leads for connection with the alternating source, said primary winding being connected across said input leads, means including a starting capacitor and output leads for supplying the output of the apparatus to. said lamps and including electrical leads for connecting said starting capacitor in shunt with one of said lamps, a series capacitor connected in series circuit relation with said secondary Winding, said circuit means including connections for placing said series capacitor in series circuit relation with said lamps, said primary winding being wound in a coil formed of contiguous layers of conductor wire and disposed on said magnetic core, said secondary winding being wound in a coil formed of contiguous layers of conductor wire and disposed on said magnetic core, said conductor wire being formed of a metallic conductor coated with insulation, and means for insulating said coils from said magnetic core, each of said coils having an equivalent resistance R and an operating voltage V such that the power V /R in watts dissipated inv the equivalent circuit divided by the number of cubic centimeters of conductor insulation in the coil and divided by the number of kilograms of metallic conductor in the coil is less than 1.6, said value of the equivalent resistance R being determined before said coil is assembled on said magnetic core at a frequency of 60 cycles per second and a temperature between 140 and 180 degrees centigrade.

5. A ballast apparatus for operating at least one fluorescent lamp from an alternating power source, said ballast apparatus comprising: a high reactance ballast transformer having a shell type magnetic core, a primary winding and a high reactance secondary winding inductively coupled therewith on said magnetic core, said primary winding being adapted for connection across the alternating current power supply, circuit means including electrical leads for supplying the output of the apparatus to said at least one lamp, said primary and secondary windings being wound with conductor wire to form coils without layer insulation, said conductor wire being comprised of a metallic conductor insulated with a resin enamel, each of said coils being impregnated with an insulating material and having an equivalent resistance R and an operating voltage V such that the power V /R in watts dissipated in the equivalent circuit divided by the number of cubic centimeters of the insulating enamel in the coil and divided by the number of kilograms of metallic conductor in the coil is less than 1.6, said value of the equivalent resistance R being determined before said coil is assembled on said magnetic core at a frequency of 60 cycles per second and a temperature between 140 and 180 degrees centigrade.

6. A coil assembly for use in a ballast transformer having a shell type magnetic core and adapted for operating fluorescent lamps, said coil assembly comprising: a spool of insulating material formed with an axial opening adapted for mounting on the magnetic core of the ballast transformer and at least one winding formed of a plurality of turns of conductor wire wound on said spool without layer insulation and impregnated with an insulating material, said conductor wire being formed of a metallic conductor insulated with a resin enamel, said coil having an equivalent resistance R and an operating voltage V such that the power V /R in watts dissipated in the equivalent circuit divided by the number of cubic centimeters of the insulating enamel in the coil and divided by the number of kilograms of metallic conductor in the coil is less than 1.6, said value of the equivalent resistance R being determined at a frequency of 60 cycles per second and a temperature between 140 and 180 degrees centigrade.

7. A coil assembly for a ballast transformer adapted for operating one or more fluorescent lamps, said coil assembly comprising: at least one winding formed of a plurality of turns of conductor wire, said conductor wire being formed of a copper conductor insulated with a resin enamel, said winding being random wound to form a coil with an axial opening and impregnated with a resin, an insulating means disposed within said axial opening to insulate the coil from the magnetic core of the ballast transformer, said coil having an equivalent resistance R and an operating voltage V such that the power V /R in watts dissipated in the equivalent circuit divided by the number of cubic centimeters of the insulating enamel in the coil and divided'by the number of kilograms of metallic conductor in the coil is less than 1.6, said value of the equivalent resistance R being determined at a frequency of cycles per second and a temperature between and degrees centigrade.

8. A coil' assembly for mounting on a winding leg of shell type magnetic core of a ballast transformer for operating fluorescent lamps comprising: a spool of insulating material formed with an axial opening and adapted for mounting on the winding leg of the magnetic core, a coil including at least one Winding comprised of a plurality of turns of conductor wire wound on said spool without layer insulation, said conductor wire being comprised of a copper conductor and coating of insulating enamel, said coil being impregnated with a resinous material and having an equivalent resistance R and an operating voltage V such that the power V /R in watts dissipated in the equivalent circuit divided by the number of cubic centimeters of the insulating enamel in the coil and divided by the number of kilograms of metallic conductor in the coil is less than 1.6, said value of the equivalent resistance R being determined at a frequency of 60 cycles per second and a temperature between 140 and 180 degrees centigrade.

References Cited by the Examiner UNITED STATES PATENTS 2,971,124 2/1961 Feinberg et a1 315257 X 3,141,112 7/1964 Eppert 315-257 X 3,200,290 8/1965 Moerkans 31599 GEORGE N. WESTBY, Primary Examiner.

C. R. CAMPBELL, Assistant Examiner.

Claims (1)

1. 4. A BALLAST APPARATUS FOR OPERATING FLUORESCENT LAMPS FROM AN ALTERNATING POWER SOURCE, SAID APPARATUS COMPRISING: A HIGH REACTANCE TRANSFORMER HAVING A SHELL TYPE MAGNETIC CORE, A PRIMARY WINDING AND A SECONDARY WINDING INDUCTIVELY COUPLED WITH SAID PRIMARY WINDING ON SAID MAGNETIC CORE, A PAIR OF INPUT LEADS FOR CONNECTION WITH THE ALTERNATING SOURCE, SAID PRIMARY WINDING BEING CONNECTED ACROSS SAID INPUT LEADS, MEANS INCLUDING A STARTING CAPACITOR AND OUTPUT LEADS FOR SUPPLYING THE OUTPUT OF THE APPARATUS TO SAID LAMPS AND INCLUDING ELECTRICAL LEADS FOR CONNECTING SAID STARTING CAPACITOR IN SHUNT WITH ONE OF SAID LAMPS, A SERIES CAPACITOR CONNECTED IN SERIES CIRCUIT RELATION WITH SAID SECONDARY WINDING, SAID CIRCUIT MEANS INCLUDING CONNECTIONS OF PLACING SAID SERIES CAPACITOR IN SERIES CIRCUIT RELATION WITH SAID LAMPS, SAID PRIMARY WINDING BEING WOUND IN A COIL FORMED OF CONTIGUOUS LAYERS OF CONDUCTOR WIRE AND DISPOSED ON SAID MAGNETIC CORE, SAID SECONDARY WINDING BEING WOUND IN A COIL FORMED OF CONTIGUOUS LAYERS OF CONDUCTOR WIRE AND DISPOSED ON SAID MAGNETIC CORE, SAID CONDUCTOR WIRE BEING FORMED OF A METALLIC CONDUCTOR COATED WITH INSULATION, AND MEANS FOR INSULATING SAID COILS FROM SAID MAGNETIC CORE, EACH OF SAID COILS HAVING AN EQUIVALENT RESISTANCE R AND AN OPERATING VOLTAGE V SUCH THAT THE POWER V2/R IN WATTS DISSIPATED IN THE EQUIVALENT CIRCUIT DIVIDED BY THE NUMBER OF CUBIC CENTIMETERS OF CONDUCTOR INSULATION IN THE COIL AND DIVIDED BY THE NUMBER OF KILOGRAMS OF METALLIC CONDUCTOR IN THE COIL IS LESS THAN 1.6, SAID VALUE OF THE EQUIVALENT RESISTANCE R BEING DETERMINED BEFORE SAID COIL IS ASSEMBLED ON SAID MAGNETIC CORE AT A FREQUENCY OF 60 CYCLES PER SECOND AND A TEMPERATURE BETWEEN 140 AND 180 DEGREES CENTIGRADE.

Images