FR2707053A1 - Voltage converter of the inverter type for supplying an electric load, especially a fluorescent tube - Google Patents

Abstract

The invention relates to converters for converting direct current into alternating current, of the "inverter" type. The converter of the invention comprises a resonant circuit having an inductor (L1) and a capacitor (C). According to one characteristic of the invention, it further includes a second inductor (L2) and means (DC, K1, K2) for connecting the capacitor (C) alternately with the first and the second inductors (L1, L2). This makes it possible particularly to operate at a fixed and high frequency.

Description

 The invention relates to converters of DC voltages in alternating voltages of the "inverter" type, for supplying all electrical loads requiring an alternating current.

The invention is particularly applicable in the inverters for the supply of neon and fluorescent tubes, particularly with cold electrodes, used in particular in motor vehicles to form for example raised brake lights, or multifunction lights, or ceiling lights .

The simplest and least expensive inverter assemblies, in particular for supplying fluorescent tubes, are for the most part inverters with bipolar transistors, of the resonant or self-oscillating type for example.
Figure 1 shows the block diagram of a conventional self-oscillating inverter. A source ST delivers a DC voltage whose positive polarity "+" is connected to the primary winding 1 of a transformer 2.

The other end of the primary is connected to the negative polarity "-" via a transistor I fulfilling a switch function. An electric charge constituted by a fluorescent tube FL is mounted across the secondary winding of the transformer 2. The transistor I is of the bipolar type, its base is connected, on the one hand via a resistor R1, to a special winding 4 of the transformer 2, and on the other hand to the positive polarity "+" by another resistance
Ri.

 When the transistor I is conductive, its base is fed by the special winding 4. When the transformer saturates, the current in the base of the transistor I decreases very rapidly. The voltage at the primary 1, and therefore that applied to the base of the transistor I by the special winding, become negative, and the transistor is blocked. As the transformer restores the energy accumulated during the previous phase to the load, the voltage across the special winding 4 decreases in absolute value. Thanks to the resistor Ri connecting the positive polarity "+" to the base of the transistor, the base-emitter junction of the latter eventually becomes busy again. The transistor tends to become conductive, which, in turn, tends to increase its base current, and so on, this forming a period of the switching frequency according to which the voltage delivered by the source ST is cut off. transistor transistor I.

These converters notably have the following drawbacks:
the switching frequency is variable, it depends inter alia on variations in the impedance of the load and the value of the supply voltage; this frequency variation notably has the effect of prohibiting the improvement of the filtering by the use of tuned filters;
the value of the switching frequency is limited by the losses in the transformer (which saturates and desaturates at each operating period).

It should also be noted that the "hard" switching bipolar transistor also limits the switching frequency.

 The small value of the switching frequency prohibits a real miniaturization of the transformer and line current filtering devices, that is to say the current flowing in the source ST. In addition, the variable character of the switching frequency complicates the filtering and the shielding of the device.

 The present invention aims to avoid the disadvantages and defects mentioned above. To this end, it proposes a converter or inverter assembly of the type comparable to a resonant assembly, but allowing, in a simple manner, to operate at a fixed, high-value switching frequency, and practically independent of the natural frequencies of the circuits formed by the passive elements of the converter.

 The assembly proposed by the invention also makes it possible to reduce losses during switching, and to make inverters delivering very high vacuum voltages which are particularly useful in the case of fluorescent tubes.

 The invention therefore relates to a device for converting a DC voltage into an AC supply voltage applied to an electrical load, comprising an inductor whose first end is connected to a first terminal of a capacitor, characterized in that it comprises a second inductor and in that it further comprises switching means acting cyclically for, firstly during a first phase, applying the DC voltage to a circuit comprising the first inductance and the capacitance arranged in series, and secondly during a second phase, connect in parallel the capacitance and the second inductance.

The invention will be better understood on reading the following description of some of these embodiments, a description given by way of nonlimiting example with reference to the appended figures among which:
- Figure 1 already described represents a conventional converter;
FIG. 2 is the basic diagram of a voltage converter according to the invention;
FIGS. 3a and 3b are equivalent diagrams of the diagram of FIG. 2 in certain phases of operation;
FIG. 4 is a diagram illustrating the "empty" operation of the converter of the invention
- Figure 5 shows curves that illustrate the operation when the converter is "loaded";
FIG. 6 shows the diagram of a preferred version of the invention using a transformer and a first coupling direction;
FIG. 7 corresponds to the diagram of FIG. 6 but with a second coupling direction;
FIG. 8 represents the diagram of a circuit intended to produce the firing of a thyristor;
FIG. 9 represents another version of the circuit of FIG. 8;
FIG. 10 shows an embodiment of an element of FIGS. 8 and 9.

 FIG. 2 represents the basic diagram of a voltage converter CT according to the invention.

 The converter CT comprises firstly a first self-induction coil or inductance L1 and a capacitor C connected in series with this inductor, and it furthermore comprises a voltage source ST delivering a DC voltage E intended to be applied to the terminals of the circuit constituted by the inductance L1 in series with the capacitor C.

According to a characteristic of the invention, the converter CT further comprises a second self-induction coil or second inductor L2 and, switching means which successively enable the DC voltage to be applied across the terminals of the series circuit L1-C. formed by the first inductor L1 in series with the capacitance
C, then then connect in parallel the capacitor C and the second inductor L2, after having (preferably, but not necessarily) disconnected the series circuit L1-C of the DC voltage E.

 In the nonlimiting example described, the switching means comprise a first and a second switching element K1, K2 controlled by a control device DC. The switching elements K1, K2 perform a switch function. In the example shown, they are constituted by thyristors, but in the spirit of the invention other types of semiconductor switches may be used, for example triacs or bipolar transistors, optionally provided with a appropriate control to enable them to simulate the behavior of thyristors.

 In the nonlimiting example shown in FIG. 1, the first inductance L1 is connected, on the one hand by its first end el, to a first terminal b1 or frame of the capacitor C, the meeting point of L1 and C being spotted P1; and on the other hand by its second end e2 it is connected to the positive polarity (+) of the DC voltage E.

The second inductor L2 is connected on the one hand by a first end el to the point P1, ie at the point common to the first inductance L1 and to the capacitance
C, and secondly its second end e2 is connected to the anode of a thyristor forming the first switching element or first switch K1. The cathode of the first switch K1 is connected to the second terminal b2 or frame of the capacitor C, this meeting constitutes a common point marked P2.

The point P2 common to the capacitor C and the cathode of K1 is connected to the anode of a second thyristor which constitutes the second switching element or second switch K2. The cathode of the second switch K2 is connected to the negative polarity (-) of the DC voltage
E, negative polarity which in the nonlimiting example described is connected to the mass of the converter.

 The triggers G1, G2 respectively of the first and second switches K1, K2 are connected to the control device DC, from which they receive at appropriate times, gate control pulses respectively IG1, IG2.

In these conditions:
when the second switch K2 is closed (by applying a gate pulse IG2 to the second gate G2), the DC voltage E is applied to the series circuit L1-C: a current is established in the branch formed by L1, C, K2, and a voltage Vc is developed across the capacitor C.

 - when then the second switch K2 opens (the DC voltage E is then no longer applied), and that we close the first switch K1 (using a trigger pulse IG1 applied to the first trigger G1): one connects (by K1) the second end e2 of the second inductance L2 to the second terminal b2 of the capacitor C, that is to say that it tends to connect these two elements in parallel; a current i2 is established in the capacitor C and in the second inductance L2, which is due to the energy stored during the previous phase by the capacitor C.

 The voltage Vc developed across the capacitor C is an AC voltage. It is therefore possible to use it directly to supply an electric load such as a FL fluorescent tube. In this case the load FL has one end connected to the first terminal bl of the capacitor C, at the point P1, and its other end is connected to the other terminal b2 of C at the point P2. However, this embodiment is not the preferred version of the invention (for reasons explained below), and to better illustrate the idle operation "(unconnected load) of the basic circuit of FIG. of the load FL is shown in Figure 2 in dashed lines.

The "idle" operation of the CT converter basic circuit shown in FIG. 2 is illustrated with reference to FIGS. 3a and 3b.
FIG. 3a is the equivalent diagram of the circuit of FIG. 2 in a first phase PH1 of the operation in which the second switch K2 is in the "closed" state. It is assumed that initially, the alternating voltage
Vc = 0; it = 0; i2 = 0. In this first phase, the representation of the circuit is limited to the voltage source ST, to the first inductance L1, and to the capacitance
C at the terminals of which develops the AC voltage Vc.

 This phase starts when sending a trigger pulse IG2 to the second switch K2.

E = Vc + L1. (Dil / dt); and, C. (dVc / dt) = it;
hence L1.C (d2 il / dt2) + il = 0;
from where he = A. cos wt + B.sin wt, where w is the pulsation, w = 1 / LC; he (0) = 0: and Vc (0) = 0; hence A = 0, B. = E / Lw; Vc = E (1-cos wt).

 Phase 1 ends spontaneously at a date or time tl such that wtl = ir; (Ir = pi). At this date, the second switch K2 is blocked (because of the change to the value 0 of the current il), ie goes to the "open" state.

 FIG. 3b shows the equivalent circuit of the converter CT in a second phase during which the first switch K1 is put in the "on" state, from the sending of a trigger pulse IG1 to the first gate G1. The elements to be considered are then the capacitance C and the second inductance L1 in which a current i2 is established.

 Vc (0) = 2E (at the end of the first phase PH1 which preceded).

 L2 (di2 / dt) = Vc, and i2 = -C (dVc / dt), hence i2 = A'cos wt + B'sin wt.

i2 (0) = 0, where A '= 0,
and hence B '= Vc (0) / Lw = 2E / Lw.

 Vc = 2E. cos wt.

 Phase 2 ends when current i2 vanishes. At this moment the first switch is blocked, that is to say goes to the "open" state. The duration of this second phase PH2 is s / w.

 A third phase PH3 following the second phase explained above corresponds to the setting in the enclosed state of the second switch K2. It is similar to the first phase PH1, with Vc (0) = -2E.

 he = (3E / Lw) sin wt.

 Vc = E - 3E cos wt.

A fourth phase PH4 following this third phase corresponds to the "closed" state of the first switch K1. It is similar to the second phase
PH2, with Vc (0) = 4E.

 i2 = (4E / Lw) sin wt.

 Vc = 4E cos wt.

 etc. Thus for example for an nth phase k similar to the first phase PH1: il = kE / Lw; and for a phase similar to the second phase PH2 of the same rank k, Vc (0) = 2kE.

 It should be noted that between the phases, it is possible to provide dead times Tm, during which the switches K1, K2 are then both blocked.

 Thus one can consider that a period of operation comprises two successive main phases: the first phase PH1 which corresponds to the conduction of the second switch K2, and the second phase corresponds to the conduction of the first switch K1.

 FIG. 4 contains curves which illustrate the appearance of currents and voltages developed during the various phases mentioned above.

 The first phase PH1 starts at time t0, where also begin the first current it in the series circuit L1-C (represented by a curve 8 in solid line), and secondly the alternating voltage Vc (represented by a curve 10) developed at the terminals of the capacitor C. The end of this phase occurs at time t1 which also marks the end of the current it. At time t1 the alternating voltage Vc has a value 2E (that is to say twice that of the DC voltage delivered by the source ST).

 The second phase PH2 then begins at a time t2 at the same time as the second current i2 (represented by a curve 9 in dashed lines). The time interval between t1 and t2 represents a previously mentioned dead time Tm, during which the alternating voltage Vc retains the same value. After passing through a maximum, the current i2 returns to a value 0 at a time t3 which marks the end of the second phase PH2. The AC voltage has a value -2E.

 The next phase PH3 starts at time t4 at the same time as the first current it. The time interval between t3 and t4 represents a new dead time Tm, during which the alternating voltage Vc keeps the same value. After passing through a maximum, the current goes back to 0 at a time t5 which marks the end of this phase PH3. The AC voltage has a value 4E.

 The following PH4 phase begins at time t6 at the same time as the second current i2. The time interval between t5 and t6 represents a new dead time Tm, during which the alternating voltage Vc keeps the same value. After passing through a maximum, the current i2 goes back to 0 at a time t7 which marks the end of phase PH4. The AC voltage has a value of -4E.

 Phase PH5 starts at time t8 at the same time as the first current it. The time interval between t7 and t8 represents a new dead time Tm, during which the alternating voltage Vc keeps the same value. After passing through a maximum, the current goes back to 0 at a time t9 which marks the end of this phase PH5. The alternating voltage has a value 6E.

 The next phase PH6 then starts at a time t10 together with the second current i2. The time interval between t9 and t10 represents a dead time Tm, during which the alternating voltage Vc keeps the same value. After passing through a maximum, the second current i2 returns to a value 0 at a time t11 which marks the end of phase PH6. The AC voltage has a value -6E.

 The operation of the basic circuit as illustrated by the curves of FIG. 5 corresponds to an undefined regime, and of course the growth of the alternating voltage Vc observed in FIG. 4 does not occur in the case of steady state operation, and charged.

 On the other hand, it can be noted that the phases or sequences of operation PH1 to PH6 and the following ones (not shown), each correspond to a transition of the alternating voltage Vc: from negative to positive for the phases PH1, PH3, PH5, and of positive to negative for the phases PH2, PH4, PH6, and that the different phase relationships between the currents I1, I2 and the alternating voltage Vc that appear in FIG. 4 are found in the steady state in the different embodiments of FIG. invention.

 The basic circuit shown in FIG. 2 makes it possible to produce very high voltages (theoretically infinite) by resonance, without imposing any relationship between the operating frequency of the control circuit and the natural frequencies of the passive elements (L1, L2, C). of the CT converter: no agreement is necessary.

 In the explanations of operation given above, the two inductances L1, L2 are assumed to have equal values, but the operation is not compromised if these inductances have different values. It should be noted finally that the switching constraints of the thyristors only impose the following condition: (n / w)>> trr (trr is the inverse recovery time); this in order not to have excessive reverse conduction, because the reverse conduction of the thyristors K1, K2 tends to discharge the capacitor C, the peak voltage then increases less rapidly than in the absence of reverse conduction. The tq time of defusing can be respected thanks to the dead time.

 FIG. 5 shows the appearance of the currents I1, I2 and of the alternating voltage Vc, in the case of operation of the base circuit with a load FL connected in parallel with the capacitor C as indicated in FIG. 2.

 The curves are marked 8 'for the first current 11, 9' for the second current i2 and 10 'for the alternating voltage Vc. It is observed that with the exception of a slight decrease during the dead times Tm, the peak value of the alternating voltage Vc is constant from one cycle to another.

In the configuration of the basic circuit of FIG. 2, the converter of the invention makes it possible to obtain a current in the load FL approaching
a sinusoid, in this case the dead times Tm are small, T = 0, T being the duty cycle.

a square signal, then T>> o / w), and
T>>RU.C; where Ru is the impedance of the FL load, and C is the capacitance.

 - Or for example the shape of a square signal distorted by differentiation, and other intermediate forms are possible.

 In the case of a current approaching a square signal, the proper pulsation of the L-C circuit conditions only the edges of the pseudo square signal and is constrained only by the condition on Trr mentioned above. It is therefore possible to accept a higher pulsation than that generally encountered in thyristor converters, which allows advantageous dimensioning of inductances L1, L2 and capacitance C.

 The embodiment shown in Figure 2 may however have a disadvantage if it is necessary to achieve high voltages because these voltages being present across the capacitor or capacitor C, this component may be expensive and bulky.

 It is also possible to directly connect the load FL to the terminals of one or other of the two inductors L1, L2. But this assembly is not the most interesting, because the voltage across each of these inductances is zero for at least half of the operating period. On the other hand, if the load FL is placed at the terminals of the two inductors L1, L2 (those being connected in series) and if, by appropriate control of the switches K1, K2, the dead times Tm are avoided, this load is obtained at the terminals of this load a voltage likely to constitute a supply voltage. This latter arrangement is illustrated in dashed lines in FIG. 2, with the load (marked FL ') connected to the second end e2 of the first inductor L1 and to the second end e2 of the second inductor L2. However, this arrangement can also have the drawback mentioned above about high voltages, because the peak voltages across the load and across the capacitor C remain close.

 FIG. 6 represents the diagram of a preferred embodiment of the invention, making it possible to solve the problem posed by the high voltages, thanks to the use of a transformer TR. The circuit diagram of the base circuit is identical to that shown in Figure 2, but in this version of one invention, the two inductors L1, L2 constitute two primary windings of the same transformer TR. The difference between the diagram presented in FIG. 2 and that of this preferred version of the invention resides in the fact that the load FL is fed through the transformer TR. In this preferred version shown in FIG. 6, the magnetic coupling direction (symbolized in FIG. 6 by circles marked Sm) of the two inductances L1, L2 or primary windings is the same for these two windings.

 The transformer TR therefore comprises a secondary winding ES, at the terminals of which the AC supply voltage VA is developed, and at the terminals of which the load FL is connected.

 It is thus possible to obtain a supply voltage VA of high value, and thanks to the transformation ratio n2 / nl of the transformer, to keep at the alternating voltage Vc across the capacitor C a sufficiently low value to allow the use of an ordinary capacitor (n2 being the number of turns of the secondary winding ES, and nl the number of turns of each of the primary windings L1, L2).

By way of nonlimiting example, it can be indicated that very satisfactory results have been obtained under the following operating conditions:
- DC voltage E = 12 V;
L1, L2 = 2 mH; C = 0.1 mF;
- n2 / nl = 4 (L1 and L4 have in this example the same number of turns);
the load FL is a neon tube of the type with cold electrodes, diameter and length respectively of 5 mm and 300 mm, having an equivalent impedance of the order of 45 kW;
- operating frequency 22 kHz;
power delivered to the 3.8 W tube (the pulses IG1, IG2 for the control of the gates G1, G2 being delivered so that the dead times have a low duration, practically negligible).

 In this configuration, the converter CT delivers to its load FL an alternative supply voltage VA to an operating frequency which is twice the natural frequency of the circuit L1-C, L2-C. The operation is practically insensitive to the simultaneous conduction of the two switches or thyristors K1, K2, and can tolerate dead times Tm (longer or shorter depending on the nature of the FL load).

 Indeed an advantage of this configuration is on the one hand, that when sending a trigger pulse IG1, IG2 to one of the switches K1, K2 before blocking the other switch, this switch that receives the pulse is polarized in such a way that it can be initiated, and that, on the other hand, the initiation of this switch causes the other switch to be blocked; this type of operation is exercised both in the case of sending a gate pulse to the first switch K1 that second switch K2.

It should be noted, however, that the configuration shown in FIG. 6 may in certain cases have a drawback, which lies in the presence of a certain harmonic content of the voltage applied to the load.
FL.

FIG. 7 shows the diagram of another version of the invention, different from that represented in FIG. 6 solely by the magnetic coupling directions of the two inductors L1, L2, which inductors constitute two primary windings of the transformer.
TR.

 In the nonlimiting example described, the second direction of coupling of the inductances L1, L2 in this version of the invention is symbolized in FIG. 7 by the fact that circles marked Sm, Sm 'are arranged near the windings L1. , L2, and that the circle Sm 'is disposed near the second winding L2 in a manner opposite to that whose circle Sm is disposed on the same second winding or inductance L2 in the case of Figure 6; consequently, in the example of FIG. 7, the windings L1, L2 are connected so as to have opposite coupling directions.

 Compared to the previous configuration shown in FIG. 6, the main change induced by the coupling direction change is in the more favorable form of the current generated in the load. Indeed, this second coupling direction makes it possible to obtain substantially sinusoidal signals, ie a substantially sinusoidal alternating supply voltage, which minimizes the electromagnetic pollution generated by the harmonics of the current flowing in the load.

 It should be noted, however, that in this variant of the converter 1 of the invention, unlike the previous example described with reference to Figure 6, when one of the switches K1, K2 is conductive, the other can be initiated without a blockage of the first occurs, so that both switches or thyristors can drive simultaneously. Under these conditions, the two primary windings L1, L2 of the transformer TR connected in series constitute a short circuit; this amounts to putting the DC voltage source ST short circuit, and leads in particular to the destruction of the thyristors.

 The control device DC may be constituted in different ways in themselves conventional, that is why in Figures 2, 6 and 7 it is represented in the form of a functional block. It is known that the gate control G1, G2 of a thyristor such as the switches K1, K2 consists of applying to the gate a positive voltage signal with respect to the cathode of the thyristor considered. Thus the control device DC could comprise different circuits (not shown), such as for example a control logic circuit delivering trigger pulses IG1, IG2, at the appropriate frequency, via a simple impedance amplifier .

 The trigger pulses IG2 intended for controlling the second switch K2 may for example be referenced to a ground, to which in the example is also connected the negative polarity (-) of the DC voltage E and thus also the cathode of K2.

In this case, the control of the first trigger G1 belonging to the first switch K1 is a little less simple to accomplish, because during operation, the cathode of the first switch K1 is at a potential that varies with respect to the mass depending on whether the second K2 switch is blocked or passing.

The trigger switch G1 of the first switch K1 can, however, be realized without difficulty, because many circuits are known that allow to control a thyristor whose cathode is not referenced to ground. Some of these circuits are very simple and are made using a transformer. In the latter case, the gate control pulse IG1 of the first thyristor or switch K1 would be delivered by a secondary transformer winding (not shown) contained in the control device DC, winding to which the first thyristor K1 would be connected by two links. as shown in FIG. 2, the first link arriving at the last trigger G1 remains to be controlled by the control device
DC.

 FIG. 8 represents a diagram of the converter of the invention reproducing exactly the diagram of FIG. 6 with the exception of the gate control G1 of the first thyristor K1. In the diagram of FIG. 8, the control of the trigger G2 is obtained by means of a so-called trigger circuit CCG, which can either be included in the control device DC, or, advantageously, be arranged in the closer to K2 given the small number of its components.

 The trigger circuit CCG in its simplest version of FIG. 8 comprises a resistor Rg and a diode Dg: one end of the resistor Rg is connected to the cathode of the second thyristor K2 (and therefore to ground), and its another end is connected to the anode of the diode Dg; the cathode of the diode Dg is connected to the gate G1 of the first thyristor K1.

In this configuration, the potential of the cathode of the first thyristor K1 being negative with respect to the mass (and thus with respect to the cathode of K2), at the end of a phase (first phase PH1) during which the second thyristor K2 was conducting, a current which flows in the diode Dg, the resistor Rg and the trigger G1, shortly after the blocking of the second thyristor
K2, performs the priming of the first thyristor K1 and the conduction of the latter at the appropriate time. The diode
The function of Dg is to protect the trunk junction of K1 against inverse voltages which may appear in the second part of a phase (second phase PH2, after cancellation of Vc) in which the first switch or thyristor K1 is conducting.

 FIG. 9 shows another version of the gate circuit CCG, based on the fact that when the second thyristor K2 is conducting, the gate current of the first thyristor K1 is zero. The potential of the cathode of the latter passes abruptly from 0 V. to a negative value during the blocking of the second thyristor K2.

 A simple current pulse being necessary to turn on the first thyristor K1, the gate G1 of the latter is connected to ground with a capacitor Cg, with a resistor Rg in series with the capacitor. In addition, a diode Dg is connected by its cathode to the gate G1 of K1, and by its anode at the anode of K2.

 The diode Dg allows the capacitor Cg to be discharged and limits the cathode-gate reverse voltages supported by K1.

 The value of the resistor Rg must be chosen so as to enable the first thyristor K1 to be turned on at the end of the first operating period (regime not established); the values of the alternating voltage Vc then being the lowest, the minimum gate current required for the initiation of K1 will be reached for all subsequent periods. The value of the current obtained is generally relatively low, so that in steady state the peak value of the trigger current can become excessive. It is possible in this case to replace this resistance Rg by a pseudo current source.

FIG. 10 represents a diagram of a pseudo current source Rg 'that can replace the resistor
Rg, both in the assembly of Figure 8 than in that of Figure 9. The pseudo source Rg 'comprises a zener diode DZ and a transistor Q1 NPN type in the example, the collector is connected to ground. The cathode of the zener diode DZ is connected to the base of Q1 and a bias resistor Rp whose other end is connected to ground. A resistor Rc is connected at one end to the emitter of Q1, and at the other end to the anode of the zener DZ; the meeting point of the zener diode DZ and the resistor Rc is marked P3 and this point P3 represents the same end of the resistor Rg as that which in FIGS. 8 and 9 is also marked P3.

 The value of the current supplied by the generator or pseudo current source Rg 'is limited to VZ / Rc (where VZ is the voltage across the zener diode DZ); the resistor Rp serves to bias the zener diode DZ.

 Of course, the control of the first thyristor K1 can be carried out according to any of the embodiments described above, and can be applied to the different versions of the invention shown in Figures 2, 6 and 7.

Claims (16)

 1) Device for converting a DC voltage (E) into an AC supply voltage (Vc, VA) applied to an electric load (FL), comprising an inductance (L1) whose first end (el) is connected to a first terminal (bl) of a capacitance (C), characterized in that it comprises a second inductor (L2) and in that it furthermore comprises switching means (DC, K1, K2) acting cyclically to, firstly during a first phase (PH1), apply the DC voltage (E) to a circuit having the first inductance (L1) and capacitance (C) arranged in series, and secondly during a second phase , connect in parallel the capacitance (C) and the second inductance (L2).  2) Converter device according to claim 1, characterized in that the electric charge (FL) is a fluorescent tube.  3) Converter device according to one of the preceding claims, characterized in that the switching means (DC, K1, K2) comprise at least two switching elements or switches (K1, K2) and a control device (DC) controlling cyclically the "conducting" state of at least one of the switches.  4) Converter device according to claim 3, characterized in that at least one of the switches (K1, K2) is a thyristor.  5) converter device according to any one of claims 3 or 4, characterized in that a first end (el) of the second inductor (L2) is connected to the meeting point (P1) of the capacitance (C) and the first inductor (L1), the second end (e2) of the second inductor (L2) being connected to the second terminal (b2) of the capacitor (C) via a first switch (K1), said second terminal (b2) being on the other hand connected to one (-) of the polarities (+, -) of the DC voltage (E) via a second switch (K2), the second end (e2) the first inductance (L1) being connected to the other polarity (+) of the DC voltage (E).  6) Converter device according to one of the preceding claims, characterized in that the electric load (FL) is connected in parallel with the capacitance (C).  7) converter device according to claim 5, characterized in that the first and the second ends of the load (FL) are respectively connected to the second end (e2) of the first inductor (L1) and the second end (e2) the second inductor (L2).  8) converter device according to any one of claims 1 to 5, characterized in that the two inductors (L1, L2) constitute primary windings of a transformer (TR) whose secondary winding (ES) delivers a voltage of power supply (VA) applied to the load (FL).  9) Converter device according to claim 8, characterized in that the two primary windings (L1, L2) have the same direction of magnetic coupling.  10) converter device according to claim 8, characterized in that the two primary windings (L1, L2) have opposite magnetic coupling directions with respect to each other.  11) Converter device according to one of the preceding claims, characterized in that the first switch (K1) is a thyristor whose cathode is at a floating potential, and in that it comprises second means (Rg, Dg, DZ ) for controlling the conduction of the first switch (K1) according to potentials present at the second terminal (b2) of the capacitance (C).  12) Converter device according to claim 11, characterized in that the trigger (G1) of the first switch (K1) is connected to one of the polarities (-, +) of the DC voltage (E) by a supply resistor (Rg).  13) Converter device according to claim 12, characterized in that a diode (Dg) is connected in series with the supply resistor (Rg).  14) converter device according to claim 12, characterized in that a capacitor (Cg) is arranged in series with the supply resistor (Rg).  15) Converter device according to claim 14, characterized in that a diode (Dg) connects said gate (G1) to the second terminal (b2) of the capacitance (C).  16) converter device according to any one of claims 12, 13, 14, characterized in that the supply resistor (Rg) is constituted by a current generator (Rg ').