WO2010037764A1

WO2010037764A1 - Excitation stage for high frequency generator

Info

Publication number: WO2010037764A1
Application number: PCT/EP2009/062662
Authority: WO
Inventors: Bernard Darges
Original assignee: Thales
Priority date: 2008-09-30
Filing date: 2009-09-30
Publication date: 2010-04-08
Also published as: FR2936666B1; FR2936666A1; EP2335351A1

Abstract

The invention relates to an excitation stage for a solid-state industrial high frequency (IHF) generator. The generator comprises two electronic switches (11, 12) connected in series between two terminals (13, 14) of a DC power supply. To avoid the simultaneous conduction of the two switches (11, 12), the switches (11, 12) are controlled by means of a square function (76, 77) whose edges, for opening and closing the switches (11, 12), are monitored temporally.

Description

Excitation stage for high frequency generator

The invention relates to an excitation stage for an industrial high frequency generator (of acronym HFI) in the solid state. These generators can be used in particular in heating devices by electromagnetic induction, dielectric loss or plasma.

The electromagnetic induction heating consists of producing a conductive part heating by the circulation of currents induced by a magnetic field. This means makes it possible to heat the part in its mass without direct contact with the source of energy. The part to be heated (or induced) is surrounded by at least one current circulation loop (or inductor).

The working frequencies of the generator are between a few tens of kilohertz and a few megahertz. The necessary powers vary from a few kilowatts to more than one megawatt. Electromagnetic induction heating is widely used in industry and science. In industry, it is used in particular in metallurgy to refine metals, to thermally treat metal parts or to produce welded tubes continuously. Heating by dielectric losses consists of producing a heating of an insulating part by causing losses in its mass, from an alternating electric field. The part to be heated forms a poor insulator. It is placed between two conductive plates powered by an alternative source. A capacitor is created whose dielectric is the part to be heated. The generators used generally have higher working frequencies than those of electromagnetic induction heating generators. They can be between a few tens of megahertz and some gigahertz. This heating mode is used in the wood industry for drying or gluing, in the textile industry or in the manufacture or shaping of plastics.

Plasma heating involves ionizing a gaseous medium to transform it into plasma. The kinetic energy of the electrons is transformed into heat. There is a considerable rise in temperature. The room to heat is placed in the plasma. The transformation of the gaseous medium into plasma is obtained by the emission of an antenna. The working frequencies of the generator are between 1 megahertz and a few tens of megahertz. This heating mode is used in many industrial applications such as in particular the fusion of refractory products, chemical synthesis.

Tube power stages are used to provide the large powers required for industrial heating, but advances in the field of power transistors are progressively leading to replacement of the tubes with more flexible solid state power stages. For example, an excitation stage of an inductor based on a class D or E amplifier has been implemented. More precisely, the excitation stage comprises two electronic switches connected in series between two poles of a power supply. continued. An inductor is connected to the common point of the two switches which are alternately closed. Among the commonly used switches are power transistors that work either completely blocked when open or saturated when closed. The internal resistance of a saturated transistor is very low, which makes it possible to obtain very good yields for such stages of excitation and therefore moderate heating of the transistors.

The most important risk of this type of stage is the simultaneous conduction of the two switches. The current passing through them forms a short circuit between the two poles of the DC power supply. This leads instantly to the destruction of the switches, especially when the powers switched by the switches are important. This risk is proven, especially when the switches are controlled by means of square functions in phase opposition, each applied to one of the switches. The closing of one of the switches coinciding with the opening of the other, the risk of simultaneous conduction is important.

To overcome this problem, several solutions have been considered. For example, it is possible to control the switching of transistors by means of sinusoidal signals in phase opposition, a signal for each switch, to slowly reduce the current flowing through one of the switches before opening the other switch. The opening of the second switch is also done gradually. Even if the controls of the two switches are slightly offset because of their manufacturing tolerance, there is no risk of saturating the two switches simultaneously. This solution nevertheless has a risk of simultaneous conduction at high frequency because of the switch tripping time. In addition, this solution reduces the efficiency of the excitation stage due to the opening and the gradual closing of the switches which causes a heating thereof. It is possible to improve this type of control by replacing the sinusoidal signals with trapezoidal signals. The plateau phases of these signals can saturate the switches for part of the time they drive. During this part the efficiency of the floor is improved. There is nevertheless a period during which the switches partially lead. During this period the yield is degraded.

The invention aims to improve the operation of an excitation stage comprising two switches connected in series, in particular by increasing its efficiency while avoiding the simultaneous conduction of the switches. For this purpose, the subject of the invention is an excitation stage for a high frequency generator comprising two electronic switches connected in series between two poles of a DC power supply, a point common to the two switches being connected to a load, characterized in that it further comprises switch control means for opening and closing each switch by following a periodic square function associated with each switch, a rising edge of the function controlling the closing of the corresponding switch and a falling edge of the function controlling the opening of the corresponding switch and in that the control means comprise means for temporally shifting a rising edge of one of the functions with respect to the falling edge of the other function immediately preceding the front amount.

The use of square functions makes it possible to maintain a high efficiency on the excitation stage. Indeed, for each switch, the transition between blocked and saturated states is done as quickly as possible compared to the switch technology used. This limits the maximum duration of partial conduction of the switches, which is a source of yield loss.

The square functions being periodic, the offset between fronts is advantageously a phase shift between fronts. The phase shift is inherently independent of the working frequency of the generator. The only high frequency limit of the generator is related to the cutoff frequency of the switches. This makes it possible to make the offset between fronts independent of the frequency of the square functions. Indeed, in a high frequency generator, the power of the generator can be adjusted by varying the opening width of the electronic switches or by varying the frequency of the square functions. The generator is at variable frequency. A phase shift between fronts of the two square functions therefore makes it possible to ensure that the switches never drive simultaneously even when the operating frequency of the generator is varied. Advantageously, the control means comprise:

A generator of two signals comprising linear portions whose slopes are reversed,

Two limiting circuits with adjustable threshold each receiving one of the signals comprising linear portions, and generating the periodic square function, the offset of the edges being a function of the thresholds of each of the limiting circuits.

It is possible to use sinusoidal signals as a signal with linear portions. In the vicinity of its zero crossing, a sinusoidal signal can be assimilated to a linear portion whose equation is of the type: y = At + B, y representing the signal voltage, t the time, A and B being constants.

To implement this embodiment, it is necessary to master the amplitude of the sinusoidal signals applied to the inputs of each adjustable threshold limiter circuit. Indeed, the constant A and therefore the offset between fronts evolve with this amplitude.

A particular embodiment makes it possible to guard against any variation in the amplitude of the sinusoidal signal. For this purpose, the control means comprise: A generator of two first periodic signals in phase opposition,

Two second limiting circuits each receiving one of the first periodic signals and each delivering a square signal by zero crossing detection of the corresponding first periodic signal,

Two integrating circuits each receiving one of the square signals and each delivering a sawtooth signal as a function of an integration of the corresponding square signal over its period, the sawtooth signals forming the signals comprising linear portions. The zero crossing of the periodic signals in phase opposition does not depend on the amplitude of the two periodic signals. As a result, the slopes of the sawtooth signals are constant and the offset between edges of the square functions no longer depends on the amplitude of the periodic signals.

To simplify the realization of the control means of the switches, the signals comprising linear portions are advantageously in phase opposition. For example, the same threshold can be used for the two limiting circuits with adjustable threshold.

The invention will be better understood and other advantages will appear on reading the detailed description of an embodiment given by way of example, a description illustrated by the attached drawing in which: FIG. 1 represents in schematic form a stage excitation device according to the invention; FIGS. 2a to 2g show in chronogram form the signals present in different subsets of the excitation stage represented in FIG. 1; the time scales are the same for the different figures 2a to 2g. FIGS. 3 to 5 show several examples of measured oscillographs of square functions used for the control of switches of the excitation stage.

For the sake of clarity, the same elements will bear the same references in the different figures. The invention is described in relation to an excitation stage 10 comprising two electronic switches 1 1 and 12 connected in series between two poles 13 and 14 of a DC power supply. For example, field effect transistors can be used as switches. The pole 13 forms the positive pole and the pole 14 forms the negative pole or mass of the power supply. A common point 15 to the two switches 1 1 and 12 is connected to a load 16 comprising for example an inductor 17 and a capacitor 18 connected in series to form a resonant circuit for generating an induced current in a room 19 to be heated. The inductor is connected between the capacitor mass 14. This type of excitation of the load 16 is commonly called half-bridge.

The invention can also be implemented when the excitation stage is in complete bridge often called complete bridge or bridge H. In this bridge the load is connected between two pairs of electronic switches, each pair being connected in series between the poles 13 and 14. The control means of the switches which will be described later are then to double.

The excitation stage 10 comprises control means 25 of the switches 1 1 and 12 to open and close at specific times. One of the aims of the invention is to maintain a high efficiency of the excitation stage 10. For this purpose, the switching means deliver to each switch 11 and 12 a function comprising the steepest edges possible to limit the switching times. switching of the switches 1 1 and 12. The stiffness of the fronts is of course limited by the technology of the components used to achieve the control means 25. Between the fronts, the function follows steps that maintain the corresponding switch is in a boiled state either in a saturated state. More generally, this type of function is called a square function with reference to the stiffness of the fronts and the steps joining the fronts between them.

The control means 25 advantageously comprise means 26 for producing two first periodic signals 27 and 28 in phase opposition. The signals 27 and 28 are advantageously sinusoidal signals because they are easier to generate and easier to convey. The signals 27 and 28 are represented in the form of a timing diagram in FIG. 2a. The time scale is represented on the abscissa and the amplitude of the signals 27 and 28 on the ordinate.

The means 26 comprise for example a transformer 30, a primary winding 31 is excited by a second sinusoidal signal 32 and a secondary winding 33 comprises a midpoint 34 and two end terminals 35 and 36. The signal 27 is for example in phase with the signal 32. The signal 28 is therefore in phase opposition with the signal 32. The sinusoidal signal 27 is delivered between the terminal 35 and the midpoint 34. Similarly, the sinusoidal signal 28 is delivered between the terminal 36 and the point medium 34. The fact of using a passive component such as the transformer 30 to generate the two signals 27 and 28 makes it possible to obtain a constant phase shift, phase opposition in the case shown, even when the frequency or frequency is changed. the amplitude of the sinusoidal signal 32. This frequency is the one at which the switches 11 and 12 switch. It can be slaved to a power measurement consumed by the load 16.

The control means 25 may also comprise two limiting circuits 40 and 41 each receiving one of the sinusoidal signals, respectively 27 and 28. Each limiter circuit 40 and 41 delivers a square signal, respectively 42 and 43, by zero crossing detection of the corresponding sinusoidal signal. 27 or 28. The two square signals 42 and 43 are in phase opposition.

FIG. 2b shows the signal 42 and FIG. 2c the signal 43. The signal 42 has an alternation of high levels 44 corresponding to the positive half-waves of the signal 27 and low levels 45 corresponding to the negative half-waves of the signal 27. and descendants 47 connect levels 44 and 45 to each other. Likewise, the signal 43 has an alternation of high levels 48 corresponding to the positive half-waves of the signal 28 and low levels 49 corresponding to the negative half-waves of the signal 28. Rising edges 50 and falling edges 51 connect the levels 48 and 49 to each other. The edges of the square signals 42 and 43 occur at the time of zero crossings of the sine signals 27 and 28 corresponding. Square signals 27 and 28 are in phase opposition.

Each limiter circuit 40 and 41 comprises, for example, an operational amplifier, respectively 52 and 53, of which an inverting input is connected to ground and a non-inverting input receives the signal. corresponding sine wave 27 or 28. The signals 42 and 43 are delivered to respective outputs of the operational amplifiers 52 and 53.

The control means 25 may also comprise two integrator circuits 56 and 57 each receiving one of the square signals, respectively 42 and 43, and each delivering a sawtooth signal respectively 58 and 59 depending on an integration of the corresponding square signal 42 or 43 over his period.

FIG. 2d represents the signal 58 and FIG. 2e the signal 59. The signal 58 has an alternation of increasing linear portions 60, when the signal 42 is at its low level 45, and decreasing 61 when the signal 42 is at its level 44. Similarly, the signal 59 has an alternation of increasing linear portions 62 when the signal 43 is at its low 49 and decreasing 63 when the signal 43 is at its high level 48. The sawtooth signals 58 and 59 are in opposition of phase. Each integrator circuit 56 and 57 comprises for example an operational amplifier, respectively 66 and 67, a non-inverting input is connected to ground and an inverting input receives the corresponding square signal 42 and 43 via a resistor, respectively 68 and 69. The signals 58 and 59 are delivered to respective outputs of the operational amplifiers 66 and 67. These outputs are connected to the respective inverting inputs via a capacitor, respectively 70 and 71. The slopes of the linear portions 60 to 63 are defined by the values of the resistors 68 and 69 and the capacitors 70 and 71. The control means 25 may also comprise two limit circuits with adjustable thresholds 74 and 75 each receiving one of the sawtooth signals, respectively 58 and 59, and generating a periodic square function, respectively 76 and 77 applied to the switches, respectively 1 1 and 12, possibly through a preamplifier, respectively 78 and 79. The functions 76 and 77 obtained from the adjustable threshold limiting circuits 74 and 75 are represented respectively in FIGS. 2f and 2g.

Each square function 76 and 77 includes rising and falling edges. By convention, it is considered that a rising edge of the function in question makes it possible to close the corresponding switch and that a falling edge opens the switch. This convention is taken to facilitate the understanding of the invention. It is of course possible to choose the opposite convention. Similarly, it is possible to invert the inverting and non-inverting inputs of the different operational amplifiers shown in FIG. 1. These inversions will simply have the effect of inverting the signs of the various signals represented without departing from the scope of the invention. .

The control means 25 make it possible to shift a rising edge of one of the functions 76 and 77 with respect to the falling edge of the other function immediately preceding the rising edge.

In the example shown in FIG. 1, the offset of the fronts is a function of the thresholds of each of the limiting circuits 74 and 75. Advantageously, the offset is adjustable. This adjustment can be done separately for each of the two functions 76 and 77. Each limiter circuit 74 and 75 comprises for example an operational amplifier, respectively 80 and 81, of which an inverting input receives the corresponding sawtooth signal 58 or 59 and whose a non-inverting input receives a threshold voltage, respectively 82 and 83. The threshold voltage 82 or 83 is for example obtained by means of a variable resistor, respectively 84 and 85 connected between a positive voltage and ground. An adjustable output of each variable resistor 84 or 85 is connected to the corresponding inverting input. The adjustment of the offset, and more precisely of the pulse width, is for example obtained by adjusting the variable resistors and more generally by varying the threshold voltages applied to the inverting inputs of the operational amplifiers 80 and 81.

FIG. 2f represents the signal 76 and FIG. 2g the signal 77. When a decreasing linear portion 61 of the signal 58 takes the threshold value 82, a falling edge 86 intervenes in the function 76 and when an increasing linear portion 60 of the signal 58 takes the threshold value 82, a rising edge 88 intervenes in the function 76. Similarly, when a decreasing linear portion 63 of the signal 59 takes the threshold value 83, a falling edge 89 intervenes in the function 77 and when an increasing linear portion 62 of the signal 59 takes the threshold value 83, a rising edge 91 occurs in the function 77. FIGS. 3 to 5 show several examples of measured oscillographs of square functions 76 and 77. Indices have been given to the marks 76 and 77 to distinguish the curves of the different figures. In FIG. 3, a rising edge 91 ₃ of the function 77 ₃ and a falling edge 86 ₃ of the function 76 ₃ are phase shifted by 20 °. Between these two fronts, the two functions 76 ₃ and 77 ₃ are both low, thus preventing the two switches 11 and 12 from opening simultaneously.

In FIG. 4, a rising edge 91 ₄ of the function 77 ₄ and a falling edge 86 ₄ of the function 76 ₄ are phase shifted by 45 °.

In Figures 3 and 4, the opening times or angles of the two switches 1 1 and 12 are adjustable simultaneously. This corresponds to the same threshold voltage used for the two limiting circuits with adjustable thresholds 74 and 75. For example, for FIG. 3, the threshold voltages 82 and 83 have the same value u1 and for FIG. thresholds 82 and 83 have the same value u2.

In Figure 5, with respect to an origin of time when both switches are open to one and would close to each other simultaneously, which corresponds to zero threshold voltages, the phase shift function 76 ₅ closing the corresponding switch of 20 ° and the function 77 ₅ expands the closing of the corresponding switch by 45 °. In Figure 5, the opening times or angles of the two switches 1 1 and 12 are different. This corresponds to different threshold voltages 82 and 83 used for the two limit circuits with adjustable thresholds 74 and 75.

Claims

1. An excitation stage for a high frequency generator comprising two electronic switches (1 1, 12) connected in series between two poles of a DC power supply (13, 14), a common point (15) to the two switches (1 1, 12) being connected to a load (1 6), characterized in that it further comprises control means (25) for the switches (1 1, 12) for opening and closing each switch (1 1, 12 ) by following a periodic square function (76, 77) associated with each switch (1 1, 12), a rising edge (88, 91) of the function (76, 77) controlling the closing of the switch (1 1, 12) and a falling edge (86, 89) of the function (76, 77) controlling the opening of the corresponding switch (1 1, 12) and that the control means (25) comprise means ( 74, 75) for temporally phase shifting a rising edge (88, 91) of one of the functions (76, 77) with respect to the falling edge (86, 89) of the function (76, 77) immediately preceding the rising edge (88, 91).

2. excitation stage according to claim 1, characterized in that the phase shift is adjustable.

3. Excitation stage according to one of the preceding claims, characterized in that the two square functions (76, 77) define regularly spaced and alternating pulses.

4. Excitation stage according to claim 3, characterized in that the width of the pulses is adjustable.

5. Excitation stage according to one of the preceding claims, characterized in that the control means (25) comprise:

A generator (26, 40, 41, 56, 57) of two signals (58, 59) comprising linear portions (60, 61, 62, 63) whose slopes are reversed,

Two adjustable threshold limiting circuits (74, 75) each receiving one of the signals comprising linear portions (60, 61, 62, 63), and generating the periodic square function (76, 77), the edge shift (86, 88, 89, 91) being a function of the thresholds of each of the limiting circuits (74, 75).

6. An excitation stage according to claim 5, characterized in that the signals (58, 59) comprising linear portions (60, 61, 62, 63) are in phase opposition.

7. Excitation stage according to claim 6, characterized in that the control means (25) comprise:

Means (26) for producing two first periodic signals (27, 28) in phase opposition, two second limiting circuits (40, 41) each receiving one of the first periodic signals (27, 28) and each delivering a square signal (42, 43) by zero crossing detection of the corresponding first periodic signal (27, 28),

Two integrating circuits (56, 57) each receiving one of the square signals (42, 43) and each delivering a sawtooth signal (58, 59) according to an integration of the corresponding square signal (42, 43) on its period, the sawtooth signals (58, 59) forming the signals having linear portions (58, 59).

8. excitation stage according to claim 7, characterized in that the means (26) for producing two first periodic signals (27, 28) in opposite phase comprise a transformer (30) of which a primary winding (31) is excited by a sinusoidal signal (32) and a secondary winding (33) comprises a midpoint (34) and two end terminals (35, 36), the first two periodic signals (27, 28) being respectively delivered between one of the extreme terminals (35, 36) and the midpoint (34).