Transformer for use in a static inverter

Abstract

A novel transformer is described for use in a static inverter in association with one or two switching semiconductor devices. The transformer produces an output for control of the associated switching device(s) which changes in sense from conduction aiding to conduction inhibiting as a function of the flux level in the transformer core. The invention is applicable to single loop cores, such as are assembled from two "U" cores. Control is effected by a primary and secondary control winding wound through an aperture pair, the aperture pair being oriented for "neutrality" of the second control winding to the main flux. The aperture pair creates a five branch magnetic path which permits optimizing the control voltage applied to the associated semiconductor devices both to enhance the switching efficiency when the switching device is initially turned on and to reduce stresses on the switching device by precluding transformer saturation when the switching device is turned off. With two switching devices, two aperture pairs are normally provided.

Description

as illustrated in Figure, and has two pairs of double apertures, one pair in the upper right side of the core and the other pair in the lower right side of the core using the orientations of FIG. 2. The transformer has a pair of voltage feedback windings 57 and 58 each connected in series with a small resistance across the input junctions of the switching transistors. Current feedback winding pairs 59, 60 and 61, 62 associated with double apertures are also provided. The primary windings of the current feedback windings are connected in series with the respective halves of the center tapped transformer primary and the emitters of the associated power transistor. This connection forces the primary current to pass through the current feedback windings 59 and 61. The second current feedback windings 60 and 62, which are inductively coupled to the control windings 59 and 61, respectively, are each connected with a serial resistance between the emitters and bases of the respective switching transistors. Once oscillation is instituted, the voltage and current feedback produce an alternating switching sequence, and the base current drive applied to the individual switching transistors is initially conduction aiding and then conduction inhibiting as before.

The FIG. 6 circuit commences to oscillate following a starting pulse generated by a circuit consisting of transistor 53, associated resistors and capacitors and windings 63 and 64 associated with one aperture pair. The starting pulse appears in winding 64 and is coupled to the current feedback winding in the same aperture pair. The appropriate switching transistors (51 or 52) begins to conduct as a result. Selective saturation of the magnetic domains (as previously detailed) determines the length of the conduction interval. The turn-off induces an equal magnitude current, but in the opposite direction in the other half of the primary. This current is conducted partially by the flyback diodes, each shown connected between an outer primary terminal and the collector of one of the switching transistors. The remaining oppositely directed current serves to inject charge into the base and results in base to collector reverse conduction. The base potential soon falls below that of the collector, and the transistor conducts normally, setting up the conditions for repeated alternate conductions. During each conduction period, base current is provided via current feedback to insure transistor saturation. For certain loads (resistive loads), commutation conditions are insured without the use of flyback diodes. The reactive current for these loads may be carried without harm by the collector-base junctions.

The voltage feedback provided by winding 57 and 58 is to insure the operation of the inverter under no load conditions. It is possible to provide no load operation by other means such as a large magnetizing current but voltage feedback is preferable since it keeps the current in the transistor to a minimum. The illustrated circuit is suitable for use with a number of loads including fluorescent lights. If used with a gaseous discharge light source, means for supplying a high voltage starting voltage should be provided.

Both the one transistor and the two transistor inverter power circuits which have been described exhibit excellent efficiencies. Known inverters operate with 6 to 8 percent losses. With the use of the two aperture core configurations in which drive is optimized, the system losses are typically reduced a further 1 to 3 percent. Since the transistor losses contribute about one third to the net system losses, the reduction in transistor losses may be in excess of 50%, greatly reducing the internal dissipation.

The transistor control mechanism of the present novel transformer permits lower total circuit dissipation than certain of the cited magnetic state responsive circuits. When a transistor input junction becomes the load to the control winding in the present double aperture configuration, the input junction may be directly coupled across the control winding without the addition of a series voltage dropping resistance to protect the input junction during turn-off and establish the reverse current level.

A resistance for junction protection with its attendant losses is not required in circuits coupled to the present transformer. In both the conduction aiding and conducting inhibiting modes of operation and until the transistor switch is turned off, the virtual electric generator acting through the second control winding 20 (FIG. 1) remains a current source not capable of generating a stiff inverse voltage across the transistor input junction. After the first branch (44) has saturated, which reduces the forward drive, the second control winding continues to maintain current transformer action. In this state, coupling of the control winding 20 to the primary control winding continues even though reduced and coupling of the control winding to the main flux is affected by the presence of two alternate paths for the main flux. One flux path (40) is not substantially coupled to the second control winding and one path (42) is. The uncoupled path (40) shunts the other path (42) and provides an alternate low reluctance path so that main flux is not forced into branch 42 and the "stiff" voltage transformer action mentioned earlier is avoided. In addition, the flux in branch 42 remains load responsive, being affected by the presence of the input junction, which acting as a generator substains the existent rate of change in flux in the branch and resists entry of the reversely sensed main flux. Since the ferrite in branch 40 has some residual slope in its B/H characteristic, the load impedance presented to the input junction continues to remain large in relation to the small internal resistance of the junction. In consequence, the alternate flux path presented by branch 40 remains available to (and favored for) the main flux and the transformer retains its nature as a current source until the negative drive has removed all stored charge from the transistor input junction, and it is off. With the transistor off, entry of the main flux into branch 42 is not longer resisted. The main flux may now enter branch 42 and does produce inverse voltage at negligible current across the input junction. Had the "virtual electric generator" become a voltage source before transistor turn off, then a series resistor developing several volts and increasing the circuit dissipation several times over that of the actual base junction losses would have been required to protect the transistor during turn-offs. Thus, additional power is saved by the present circuit beyond that attributable to efficient operation of the transistor switch alone.

While the invention is applicable to a variety of ferrite materials, the materials classified as "soft" ferrites are most satisfactory. A suitable material in this class is the Stackpole material 24B. The material has a steeper slope and substantial curvature at lower "B" values, but as B increases, both a curvature remains and an absolute slope remains, though reduced. This property facilitates the function of the transistor input junction in sustaining current generator action and precluding premature voltage generator action. As the second branch 40 is saturating under the growing main flux, by which is meant a continuing reduction in slope of the B/H characteristic is taking place, the two reluctance dependent terms of expression 14 (defining the base current drive) become increasingly negative. This suggests that at some point on the slope of the B/H characteristic, the base drive will become zero, and that beyond this point, the base drive will become negative in correspondence to observation. To achieve an initial regenerative action, a certain slope of the B/H curve is assumed in expression 14. For the final degenerative action, the slope must change enough to increase the two reluctance dependent terms to the point where the full expression becomes negative corresponding to a reversal in drive. The desired change in drive from regenerative to degenerative may therefore take place without precise dependence on the manner in which the B/H curve goes from a steeper slope to a less steep slope. However, as noted earlier, the desired current transformer action is facilitated by the assumption that the B/H curve still retains appreciable slope and so does not present a zero load to the transistor input junction during the interval that it itself is a source of energy. Observation of the constancy of the V_eb voltage shows that soft ferrites generally have the required B/H properties necessary for safe and efficient turn off of conventional junction type power transistors.

In high efficiency circuits, a damping resistor (35, FIG. 1) is generally used to damp any ringing which might occur during switching and prevent accidental retriggering of the transistor. The resistor, which is connected in shunt across the base junction, prevents high frequency energy coupled through the collector-base capacity to the base electrode from causing injection. The resistor can be chosen so that the dissipation is small compared to the base junction dissipation and preserves circuit efficiency.

Transformer-transistor circuits of the nature herein contemplated have an improved stability of power versus temperature. This is true in both triggered and self-oscillating configurations. Both the junction transistor and the ferrite of the core are exposed to the same temperature in conventional packaging. This intimacy is a consequence of the desirability to achieve a minimum size total package and is in part due to the requirement to reduce electro-magnetic interference which increases as lead lengths and component separations increase. With increasing temperature, the "B max" of the ferrite decreases, reducing the volt time area of individual conduction periods. Thus, with increasing temperature and fixed supply voltage, the ferrite action tends to decrease the individual conduction periods. The voltage into which the control winding operates is a diode junction whose voltage decreases with temperature, which tends to increase the individual conduction periods. These two effects are approximately in a ratio of 3 to 2 with the effect of the ferrite dominating, and producing a 3 to 1 improvement in power stability.

While the transformer herein described has been illustrated in use in two kinds of static inverters, it is also applicable to blocking oscillators and other forms of static inverters.

Claims

1. A transformer having

(1) a core of substantially linear magnetic material for main flux pursuing a closed magnetic path, a pair of apertures in a localized region of the core, a first aperture being arranged before a second aperture along said magnetic path, said first aperture dividing the magnetic path into a first and a second branch, and said second aperture dividing the magnetic path into a third and fourth branch, with a fifth branch being formed between said apertures, the first and fourth branches forming a first diagonal pair of branches and the second and third branches forming a second diagonal pair of branches,

(2) a primary winding encircling said magnetic core for generating a main flux in said closed magnetic path when a voltage is applied thereto,

(3) a first control winding, serially coupled with said primary winding, encircling said fifth branch for generating a circulating flux forming two counter-rotating loops, one around each aperture, when current is supplied thereto, the fluxes in said two loops combining additively in said fifth branch, the flux in one of the loops combining with said main flux additively in each of said first diagonal pair of branches, and substractively in each of said second diagonal pair of branches, predisposing a branch in said first diagonal pair of saturate first as energization increases, and

(4) a second control winding encircling said fifth branch for deriving an electrical quantity whose sign reverses as a function of the magnetic state of said core.

2. A transformer as in claim 1 wherein

said apertures are serially arranged along said magnetic path placing said fifth branch orthogonal to said main flux to reduce the tendency of main flux to flow into said fifth branch and to be coupled to said second control winding.

3. A transformer as set forth in claim 1 wherein

the product of the reluctances of said first diagonal pair of branches equals the product of the reluctances of said second diagonal pair of branches to reduce the tendency of main flux to be coupled to said second

control winding. 4. A transformer as The combination set forth in claim 3 20 wherein

the reluctances of said first and second branch are equal and the

reluctances of said third and fourth branches are equal. 5. A transformer as The combination as set forth in claim 3 20 wherein

the reluctances of said first and second branches are equal and the reluctances of said third and fourth branches are equal and larger than the reluctances of said first and second branches, predisposing said fourth branch in said first diagonal pair to saturate first.

6. A transformer as set forth in claim 5 wherein

said first aperture is of lesser diameter than said second aperture. 7. A transformer as The combination set forth in claim 6 20 wherein

the reluctance of said fifth branch is less than that of said other branches to preclude saturation thereof prior to said first and fourth branches.

8. A transformer as set forth in claim 2 wherein

said first and second control windings are of a few turns and are closely coupled in the absence of selective saturation to achieve substantial current transformer action.

9. A transformer as set forth in claim 8 wherein

said first control winding has maximum core coupling to said second control winding in the absence of saturation, said coupling being reduced as each

branch saturates. 10. In combination,

(1) a core of substantially linear magnetic material for main flux persing persuing a closed magnetic path, a pair of apertures in a localized region of the core, a first aperture being arranged before a second aperture along said magnetic path, said first aperture dividing the magnetic path into a first and second branch, and the second aperture dividing the magnetic path into a third and fourth branch, with a fifth branch being formed between said apertures, the first and fourth branches forming a first diagonal pair of branches and the second and third branches forming a second diagonal pair of branches,

(2) a primary winding encircling said magnetic core for generating a main flux in said closed magnetic path when an alternating voltage is applied thereto,

(3) a first control winding encircling said fifth branch for generating a circulating flux forming two counterrotating loops, one around each aperture, when an alternating current is supplied thereto, the flux in said two loops combining additively in said fifth branch,

(4) means for energizing said primary winding and said first control winding with alternating quantities having a suitable fixed relative phase so that, generating a combined flux in which the flux in one of the loops combines additively with said main flux additively in each of said first diagonal pair of branches and subtractively in each of said second diagonal pair of branches, predisposing a branch in said first diagonal pair to saturate first as the energization increases, said energization being increased until saturation is approached, and

(5) (4) a second control winding encircling said fifth branch for deriving an electrical quantity whose sign reverses as a function of the magnetic state of said core, and

(5) a junction transistor switching device for applying alternating current to said primary winding and to said first control winding, said device exhibiting appreciable stored charge, and (6) means coupling said second control winding across the input junction of said transistor for applying a current to said switching device in a sense aiding normal conduction in the absence of saturation, and in a sense inhibiting conduction when saturation occurs.

11. The combination as set forth in claim 10

having in addition thereto a resistive load coupled to said second control winding,

saturation of said first diagonal pair of branches tending to force main flux into said fifth region and reverse the polarity of the electrical

quantity coupled to said resistive load.

12. In combination with a transformer as in claim 10,

a junction transistor switching device for applying alternating current to said primary winding and to said first control winding, said device exhibiting appreciable stored charge, and

means coupling said second control winding across the input junction of said transistor for applying a current to said switching device in a sense aiding normal conduction in the absence of saturation, and in a sense

inhibiting transistor conduction when saturation occurs. 13. The combination set forth in claim 12 10 wherein said transformer core material has a B max which decreases with increasing temperature tending to reduce the volt time area of each conduction period while the voltage of the transistor input junction to which the transformer second control winding is connected decreases with increasing temperature so as to increase each conduction period, said connection reducing the effect of temperature upon

the output power. 14. The combination set forth in claim 12 13 wherein

said coupling means couples said second control winding directly across said input junction to provide a serial path of low resistance to reduce circuit dissipation, said second control winding not appearing as a

voltage source until said transistor is substantially non-conductive. 15. The combination set forth in claim 14 wherein

the reluctances of said first and second branches are substantially equal and the reluctances of said third and fourth branches are equal,

said input junction sustanining a constant voltage drop across said second control winding and forcing a constant rate of change of flux in said fifth branch so long as said transistor is conductive, saturation of said first diagonal pair of branches increasing the reluctance coupling said second control winding to said first control winding and causing a reversal in polarity of said applied current until stored charge is

removed from said switching device and conduction is terminated. 16. The combination set forth in claim 14 wherein

the reluctances of said first and second branches are equal and the reluctances of said third and fourth branches are equal and larger than the reluctances of said first and second branches, predisposing said fourth branch to saturate first and said first branch to saturate second, saturation of said fourth branch increasing the reluctance coupling said second control winding to said first control winding and reducing the rate of increase of applied current; saturation of said first branch causing a reversal in slope of said applied current and a reversal in current, said reversals continuing until stored charge is removed from said switching

device and conduction terminated.

17. A transformer having

(1) a core of substantially linear magnetic material having a closed magnetic path, a first pair of apertures in the core in a first localized region of the core, and a second aperture pair in a second localized region of the core,

one aperture of each said pair dividing the magnetic path into a first and a second branch and the other aperture of that pair dividing the magnetic path into a third and fourth branch, with a fifth branch being formed between said apertures, the first and fourth branches forming a first diagonal pair of branches and the second and third branches forming a second diagonal pair of branches to form a set of five branches in each region,

(2) a pair of primary windings encircling the magnetic path for generating a main flux in said core when current is supplied thereto,

(3) a pair of first control windings each encircling the fifth branch of each region for generating a circulating flux forming two counterrotating loops around said apertures when current is supplied thereto, the fluxes in said two loops combining additively in said fifth branch, the flux in one of the loops combining with said main flux additively in each of one diagonal pair of branches and subtractively in each of the other diagonal pair of branches, predisposing a branch in said first diagonal pair of saturate first as energization increases,

(4) a pair of second control windings, each encircling the fifth branch of said region for deriving an electrical quantity whose sign reverses as a function of the magnetic state of said core.

18. A transformer as in claim 17 wherein

the apertures of each pair are serially arranged along said magnetic path, placing said fifth branch orthogonal to said main flux to reduce the tendency of the main flux to flow into said fifth branch and to be coupled

to said second control winding. 19. The combination set forth in claim 10 wherein said fixed relative phase is achieved by the connection of said first control winding in series with said primary winding. 20. The combination set forth in claim 19 wherein

said apertures are serially arranged along said magnetic path placing said fifth branch orthogonal to said main flux, and

the product of the reluctances of said first diagonal pair of branches equals the product of the reluctances of said second diagonal pair of branches to reduce the tendency of main flux to be coupled to said second control winding.